journal of biological rhythms€¦ · journal of biological rhythms official publication of the...

83
Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL Pebbles of Truth 515 Martin Zatz FEATURE Review 516 Clockless Yeast and the Gears of the Clock: How Do They Mesh? Ruben Baler ARTICLES Resetting of the Circadian Clock by Phytochromes 523 and Cryptochromes in Arabidopsis Marcelo J. Yanovsky, M. Agustina Mazzella, Garry C. Whitelam, and Jorge J. Casal Distinct Pharmacological Mechanisms Leading to c-fos 531 Gene Expression in the Fetal Suprachiasmatic Nucleus Lauren P. Shearman and David R. Weaver Daily Novel Wheel Running Reorganizes and 541 Splits Hamster Circadian Activity Rhythms Michael R. Gorman and Theresa M. Lee Temporal Reorganization of the Suprachiasmatic Nuclei 552 in Hamsters with Split Circadian Rhythms Michael R. Gorman, Steven M. Yellon, and Theresa M. Lee Light-Induced Resetting of the Circadian Pacemaker: 564 Quantitative Analysis of Transient versus Steady-State Phase Shifts Kazuto Watanabe, Tom Deboer, and Johanna H. Meijer Temperature Cycles Induce a Bimodal Activity Pattern in Ruin Lizards: 574 Masking or Clock-Controlled Event? A Seasonal Problem Augusto Foà and Cristiano Bertolucci LETTER Persistence of Masking Responses to Light in Mice Lacking Rods and Cones 585 N. Mrosovsky, Robert J. Lucas, and Russell G. Foster MEETING Eighth Meeting of the Society for Research on Biological Rythms 588 Index 589

Upload: others

Post on 30-Jul-2020

1 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Journal of Biological RhythmsOfficial Publication of the Society for Research on Biological Rhythms

Volume 16, Issue 6 December 2001

EDITORIAL Pebbles of Truth 515Martin Zatz

FEATURE

Review 516Clockless Yeast and the Gears of the Clock: How Do They Mesh?Ruben Baler

ARTICLES Resetting of the Circadian Clock by Phytochromes 523and Cryptochromes in ArabidopsisMarcelo J. Yanovsky, M. Agustina Mazzella, Garry C. Whitelam, and Jorge J. Casal

Distinct Pharmacological Mechanisms Leading to c-fos 531Gene Expression in the Fetal Suprachiasmatic NucleusLauren P. Shearman and David R. Weaver

Daily Novel Wheel Running Reorganizes and 541Splits Hamster Circadian Activity RhythmsMichael R. Gorman and Theresa M. Lee

Temporal Reorganization of the Suprachiasmatic Nuclei 552in Hamsters with Split Circadian RhythmsMichael R. Gorman, Steven M. Yellon, and Theresa M. Lee

Light-Induced Resetting of the Circadian Pacemaker: 564Quantitative Analysis of Transient versus Steady-State Phase ShiftsKazuto Watanabe, Tom Deboer, and Johanna H. Meijer

Temperature Cycles Induce a Bimodal Activity Pattern in Ruin Lizards: 574Masking or Clock-Controlled Event? A Seasonal ProblemAugusto Foà and Cristiano Bertolucci

LETTER Persistence of Masking Responses to Light in Mice Lacking Rods and Cones 585N. Mrosovsky, Robert J. Lucas, and Russell G. Foster

MEETING Eighth Meeting of the Society for Research on Biological Rythms 588

Index 589

Page 2: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Journal of Biological RhythmsOfficial Publication of the Society for Research on Biological Rhythms

EDITOR-IN-CHIEF

Martin Zatz

EDITORIAL BOARD

ADVISORY BOARD

FEATURES EDITORS

Larry MorinSUNY, Stony Brook

Anna Wirz-JusticeUniversity of Basel

ASSOCIATE EDITORS

Michael HastingsUniversity of Cambridge

Ken-Ichi HonmaHokkaido Univ School Medicine

Michael YoungRockefeller University

Josephine ArendtUniversity of Surrey

Charles A. CzeislerHarvard Medical School

Serge DaanUniversity of Groningen

Terry PageVanderbilt University

Ueli SchiblerUniversity of Geneva

William J. SchwartzU Mass Medical School

Deborah Bell-PedersenTexas A&M University

Gene D. BlockUniversity of Virginia

Vincent M. CassoneTexas A&M University

Jay C. DunlapDartmouth Medical School

Russell G. FosterImperial College of Science

Albert GoldbeterUniversity of Brussels

Carla B. GreenUniversity of Virginia

Paul E. HardinUniversity of Houston

William J. M. HrusheskyAlbany VA Medical Center

Helena IllnerovaCzech Academy of Sciences

Carl JohnsonVanderbilt University

Steve KayThe Scripps Research Institute

Jennifer J. LorosDartmouth Medical School

Robert Y. MooreUniversity of Pittsburgh

Steven M. ReppertHarvard Medical School

Till RoennebergUniversity of Munich

Mark D. RollagUniformed Services University

Benjamin RusakDalhousie University

Rae SilverColumbia University

Kathleen King SiwickiSwarthmore College

Fred W. TurekNorthwestern University

David R. WeaverHarvard Medical School

Irving ZuckerUC, Berkeley

Page 3: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Journal of Biological RhythmsOfficial Publication of the Society for Research on Biological Rhythms

The JOURNAL OF BIOLOGICAL RHYTHMS (JBR) publishes original, full-length reports in English of empirical inves-tigations into all aspects of biological rhythmicity. Of particular interest are submissions that focus on rhythms related tothe major environmental cycles, including daily (circadian) rhythms, tidal rhythms, annual rhythms (including photo-periodism), and biological rhythms that interact with those rhythms influenced by the environment. Studies using ge-netic, biochemical, physiological, behavioral, and modeling approaches to understand the nature, mechanisms, andfunctions of biological rhythms in all species are welcome. A major objective of JBR is to serve as a vehicle for transmit-ting information about biological rhythms in plants and animals, as well as those studying human rhythms in both theclinical and real world setting.

Preliminary or incomplete studies will not be considered. Research reported in the journal must meet the higheststandards of experimental design and data analysis. Opinion papers and reviews of significant, timely issues will also beconsidered.

JOURNAL OF BIOLOGICAL RHYTHMS (ISSN 0748-7304) is published bimonthly in February, April, June, August,October, and December by Sage Science Press, an imprint of Sage Publications, 2455 Teller Road, Thousand Oaks, CA91320; telephone (800) 818-SAGE (7243) and (805) 499-9774; fax/order line (805) 375-1700; e-mail [email protected];http://www.sagepub. com. Copyright © 2001 by Sage Publications. All rights reserved. No portion of the contents maybe reproduced in any form without written permission of the publisher.

Disclaimer: The authors, editors, and publisher will not accept any legal responsibility for any errors or omissions thatmay be made in this publication. The publisher makes no warranty, express or implied, with respect to the material con-tained herein.

Subscriptions: Annual subscription rates for institutions and individuals are based on the current frequency. Pricesquoted are in U.S. dollars and are subject to change without notice. Canadian subscribers add 7% GST (and HST asappropriate). Outside U.S. subscription rates include shipping via air-speeded delivery. Institutions: $550 (withinthe U.S.) / $574 (outside the U.S.) / single issue: $105 (worldwide). Individuals: $130 (within the U.S.) / $154 (out-side the U.S.) / single issue: $33 (worldwide). Orders by MasterCard or Visa can be placed by phone at (805) 499-9774;fax/order line (805) 375-1700. Payment must be made in U.S. dollars.

Indexed in Academic Search, Agricola, Biological Abstracts, Biosciences Citation Index, BIOSIS Previews, Chemical Ab-stracts, Comprehensive MEDLINE with FullTEXT, Corporate ResourceNET, Current Citations Express, Current Contents

®:

Life Sciences, Elsevier BIOBASE/Current Awareness in Biological Sciences, Ergonomics Abstracts, Health SourcePlus, Index Medicus, INSPEC, International Aerospace Abstracts, ISI Basic Science Index, MasterFILE FullTEXT,MEDLINE, Neurosciences Abstracts, Periodical Abstracts, Psychological Abstracts, PsycINFO, PsycLIT, ScienceCitation Index

®, Science Citation Index Expanded, and TOPICsearch.

Back issues: Information about availability and prices of back issues may be obtained from the publisher’s order depart-ment (address below). Single-issue orders for 5 or more copies will receive a special adoption discount. Contact the orderdepartment for details.

Inquiries: Address all correspondence and permissions requests to Sage Publications, 2455 Teller Road, Thousand Oaks,California 91320; e-mail [email protected]; http://www.sagepub.com.

Authorization to photocopy material for internal or personal use under circumstances not falling within the fair use pro-visions of the Copyright Act is granted by Sage Publications to libraries and other users registered with the CopyrightClearance Center Transactional Reporting Service, provided that the fee of $.50 per copy plus $.10 per copy page is paiddirectly to the CCC, 21 Congress Street, Salem, MA 01970. The identification code for JOURNAL OF BIOLOGICALRHYTHMS is 0748-7304/2001/$.50 +.10.

Advertising: Current rates and specifications may be obtained by writing to the Advertising Manager at the ThousandOaks office (address above).

Claims: Claims for undelivered copies must be made no later than 6 months following month of publication. The pub-lisher will supply missing copies when losses have been sustained in transit and when the reserve stock will permit.

Changes of address: Please inform the publisher at least 6 weeks prior to move. Enclose present mailing label withchange of address.

Effective with the 1986 volume, this journal is printed on acid-free paper.

Cover photo: Plamen D. Penev, Northwestern University

For Sage Science Press: Ben Sztajnkrycer, Production EditorEric Law, Copy EditorPaul Doebler, DesignerKatinka Baltazar, Designer

PRINTED IN THE UNITED STATES OF AMERICA. COPYRIGHT © 2001 SAGE PUBLICATIONS

Page 4: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

JOURNAL OF BIOLOGICAL RHYTHMS / December 2001Zatz / EDITORIAL

EDITORIAL

Pebbles of Truth

Science has gotten such a good reputation foranswering questions that just about everybody claimsthe adjective “scientific” for what they say. An impec-cable scientific approach is, however, useless for mostof life’s important questions like “Wherein lies theGood?” “Why me?” “Hold ‘em or fold ‘em?” “Shallwe send troops?” Scientists are no better than any-body else at making most personal and political deci-sions and can be a real pain when it comes to provid-ing clear answers to simple questions—especially iftheir defensive scientific cloaking device is turned on.The way we scientists deal with questions andanswers often frustrates the people who consult usand support us.

There are two reasons for this. First, we frequentlydon’t accept the question. Many of the biggest, mosturgent, or most important questions are concernedwith what should be, and science addresses only whatis. As Richard Feynman explained, “The question:‘Should I do this?’—whether you want something tohappen or not—must lie outside of science.” Second,our tendency, even when the questions posed are sci-entific, is to refuse to answer until we’re good and ready.Unlike policy makers or executives or police officersor editors, we need not (and often refuse to) come to afast conclusion. We claim the privilege of uncertaintylong after others have made up their minds. Accord-ing to Feynman, a scientist is never certain. When astatement is made, the question is not whether it istrue or false but rather how likely it is to be true orfalse. There is no certainty; even our best answers areat least a little provisional. This chronic hedging ofours is a remarkable trait and a precious privilege—rare across history and geography—the freedom todoubt and to declare “I don’t know” publicly.

Ever doubtful, wary of conclusions, even wary offacts, we parse the truth of statements ever so fine:

How true is it? Is it the whole truth? Is it entirely true orjust partially true? Is it strictly true, necessarily true,generally true, often true, true under certain circum-stances? Is it conditionally true, likely true, possiblytrue? We thereby bypass some of the deeper, moreintractable, issues of truth and causality and compen-sate with the benefits of open-mindedness, disinterest(not fooling ourselves), and small hard truths.

The ability to declare a question presently unan-swerable, no matter how important, and to acceptinterim and partial truths without commitment, isperhaps the greatest strength of science and a hall-mark of its different worldview. We have had the priv-ilege, so far, of choosing our questions. Althougheveryone wants answers to big questions, we usuallyprefer to settle for results that clearly answer a smallquestion over results that merely bear on a big ques-tion. What a peculiar way of getting answers normalscience has: nibbling at a problem, not trying to swal-low it whole. Yet, by an invisible hand, it seems to endup giving us a better grasp of truth and causality afterall.

Isaac Newton said, “I do not know what I mayappear to the world; but to myself I seem to have beenonly like a boy playing on the seashore, and divertingmyself in now and then finding a smoother pebble or aprettier shell than ordinary, whilst the great ocean oftruth lay all undiscovered before me.” This prettyscene embodies several of the ideals of science: mod-esty, curiosity, and wonder. We have found treasureson the beach: shiny shells and pebbles—what starsand people and firefly flashes are made of. The shellsand pebbles add up and tell us about the sea. Each ofus gets to place some on the pile.

Martin ZatzEditor

515

JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 515© 2001 Sage Publications

Page 5: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

JOURNAL OF BIOLOGICAL RHYTHMS / December 2001Baler / THE YEAST HYBRID SYSTEM

REVIEW

Clockless Yeast and the Gearsof the Clock: How Do They Mesh?

Ruben Baler1

Unit on Temporal Gene Expression, Laboratory of Cellular and Molecular Regulation,National Institute of Mental Health, National Institutes of Health, Bethesda, MD 20892, USA

Abstract In spite of its apparent weakness as a clock model, the budding yeasthas spawned a technique that has revolutionized our ability to study specificprotein-protein interactions like those at the core of the molecular timekeepingmechanisms. Here, the author will summarize the evolution, power, and limita-tions of this technique and highlight its potential and actual contributions to thefield of chronobiology.

Key words yeast two hybrid, circadian interactome, clock genes

Specific interactions between proteins form thebasis of almost every aspect in a cell’s life. These inter-actions create a dynamic and tightly regulated com-munications network and weave a complex connec-tivity map from which cell phenotypes emerge. It ishardly surprising that the identification, characteriza-tion, modification, and exploitation of all these spe-cific contacts would constitute such major focal pointsin modern biological research.

The study of protein-protein interactions waslargely confined, until 1989, to biochemical techniquessuch as crosslinking, co-precipitation, and fraction-ation. However, these techniques are limited by thefact that the interaction of interest is either initiated ormeasured outside of a living cell. This situation wasdramatically changed with the introduction of agenetics-based strategy (Fields and Song, 1989),referred to as the yeast two-hybrid system (YTH), thatmade it possible, for the first time, to investigate pro-tein-protein interactions inside a living cell.

The continued application of the YTH and relatedprotocols led to the description of thousands of bind-ing partners. This new information contributed greatlyto the identification of novel genes (Boulton et al.,2001), the dissection of complex signal transduction

networks (Hartwell et al., 1999), and the assignment oflikely roles to functionally uncharacterized proteins(Brent and Finley, 1997; Uetz et al., 2000).

THE YTH SYSTEM:A CATALYST OF DISCOVERY AND

SELF-IMPROVEMENT

YTH relies on the modular nature of transcriptionactivating factors (Sadowski et al., 1988), which tendto bind to the control region of a gene with one domainwhile activating the transcriptional machinery withanother (Fig. 1A). The strategy that makes YTH possi-ble is based on the separation of the DNA binding andthe activation domains of a known transcription fac-tor, which are then fused to putative binding partners(Fig. 1B). The two resulting fusion proteins are notexpected to activate transcription on their own. How-ever, after their simultaneous expression into a suit-able yeast reporter strain, a successful interaction willreestablish the physical link between the DNA bind-ing and activation domains (Fig. 1C). This event, evenif transient or of low affinity, can be recorded if theyeast strain carries an easily detectable (e.g., colori-

516

1. To whom all correspondence should be addressed: Building 36, Room 2A-09, National Institutes of Health, Bethesda, MD20892, USA; e-mail: [email protected].

JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 516-522© 2001 Sage Publications

Page 6: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

metric, such as lacZ) or selectable (e.g., prototrophic,such as his3) reporter gene. Typically, one of the bind-ing partners is a known entity and referred to as the“bait” (b) because it is used to “catch” an unknown“prey” (p), via more or less stable interactions.

The remarkably high sensitivity of YTH screens,however, comes at the expense of a relatively highnumber of false positives, a persistent concern forusers of this technique. Another disturbing fact relatesto the predicted interactions that are not detected bythis technique, the so-called false negatives. Indeed,less than 15% of previously established binding part-ners in yeast have been rediscovered by supposedlycomprehensive screens (Hazbun and Fields, 2001). Asa consequence, researchers are constantly looking forthe perfect balance between increased stringency andmaximal coverage. At the present time, YTH tech-niques can suggest potential binding partners, butindependent confirmation by additional methods isrequired to demonstrate that the interaction is specificand physiologically relevant.

These clear disadvantages notwithstanding, theYTH strategy enjoys a high level of success and accep-tance, possibly due to its remarkable ability to evolverapidly in response to the frequently encountered tech-nical obstacles (Finley and Brent, 1996; Brachmannand Boeke, 1997; Fashena et al., 2000). For example,as already mentioned, a significant fraction of YTH-derived findings represent biologically irrelevantinteractions. Commonly isolated interactors, such asheat shock, ribosomal, or proteasome-related pro-teins, can interact with approximately one-third ofrandomly chosen baits.

Hence, much of the initial efforts to improve theYTH protocol focused on assessing and increasing thespecificity of the observed interaction. An early stepforward in this direction was the swapping of the baitand prey moieties between the DNA binding andtransactivation domains of the reporting factor (Duet al., 1996). This domain swap was initially designedto bypass the false positives that result from the fusionof transcriptional activators onto the DNA bindingdomain, but it quickly became a standard specificity

Baler / THE YEAST HYBRID SYSTEM 517

Reporter gene

b

A

B

C

p

pb

Figure 1. The yeast two-hybrid system. Transcription factors are often composed of separate domains (A) for DNA recognition (grabbinghand) and transcription activation (pointing finger). This characteristic allowed the physical separation of these domains onto whichunknown protein moieties can be grafted. If these moieties, usually referred to as “bait” (b) and “prey” (p), are indifferent to each other, thereporter gene will remain silent (B). If, on the other hand, a specific (or nonspecific) interaction between them occurs, the event will bedetected through reporter gene activation (C). This simple configuration represents the basic idea behind all the yeast hybrid protocols thatallow detection of protein interactions under many different conditions.

Page 7: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

control. Later, the use of carefully designed doublebait systems significantly increased specificity throughthe introduction of a second bait vector carrying a dif-ferent DNA binding domain fused to a nonspecificinteractor (Serebriiskii et al., 1999). The extent towhich this alternative pathway also becomes acti-vated by the same prey provides a good indication ofthe level of nonspecific interaction.

The most significant advance in the specificityfront, however, has been the inclusion of multiple andindependent reporter systems. Thus, true bindingpartners have to activate several different genes beforebeing considered for further analysis. This strategydramatically increased the stringency of the screenand, as a result, the capacity of YTH to expose candi-dates worth pursuing.

Another limitation of the original YTH method liesin the nature of the exclusively transcription-basedreadout. This feature excludes a significant number ofpotential interactors that are either membrane bound,unable to access the nuclear compartment in an activeform, or independently active on transcription. Overtime, several variations were developed, either inmammalian or yeast cells, that address these problemsby probing interactions at their natural site in the cell(Johnsson and Varshavsky, 1994), in the cytosol (Rossiet al., 2000), or in association with the plasma mem-brane (Johnsson and Varshavsky, 1994; Aronheim et al.,1997; Isakoff et al., 1998; Medici et al., 1997).

In addition, a given interaction might occur only asa consequence of unique posttranslational modifica-tions. Absence of a differentially expressed modifyingsystem would prevent the detection of such an inter-action, a limitation of significant consequences whenwe consider a recent report regarding the dispropor-tionate representation of proteins involved in signaltransduction among different species (Chervitz et al.,1998). If the missing link is known, it is possible toco-express it during the interaction screen (Kochanet al., 2000); this approach appears to be particularlyuseful when tyrosine kinases are involved (Chervitzet al., 1998). Similarly, some interactions that involveternary complexes are unstable if one of the compo-nents is missing. After the arrival of the three-hybridsystem (Zhang and Lautar, 1996; Kochan et al., 2000;Brachmann and Boeke, 1997), formation of three-com-ponent complexes has been achieved repeatedly, thusallowing us to address the issue of how larger proteinstructures (or even scaffolds) are put together. Predict-ably, the development of reliable yeast hybrid (YH)

screens that target quaternary complexes is justaround the corner (Pause et al., 1999).

Finally, YH screens have not remained confined tothe realm of interactions between proteins. In theso-called RNA three-hybrid system (Putz et al., 1996),interactions between proteins and RNAmolecules canalso be investigated. Meanwhile, a one-hybrid proto-col (Inouye et al., 1994) has been developed to identifyDNA binding proteins using specific cis-acting ele-ments as interaction targets. On the other hand,reverse two- and one-hybrid systems (Vidal et al.,1996) offer the possibility to screen for mutations orsmall ligands that disrupt a particular interaction,adding a new dimension to the analysis of structure-function relationships and the high-throughputsearch for compounds with potentially useful phar-macological properties.

BEYOND YTH

Most classic biochemical techniques have beenused to corroborate an interaction first suggested by ayeast hybrid search. In the context of new technolo-gies, however, a group of methods referred to as reso-nance energy transfer (RET; Li et al., 2001) is emergingas one of the most powerful tools for confirming a sus-pected interaction in vivo. RET is based on the interac-tion of two energetically linked luminescent (LRET),or fluorescent (FRET) probes (fused to interacting pro-teins). When these moieties are brought into proxim-ity, energy resonance causes either quenching or exci-tation and the concomitant change in emission spectrathat can be monitored (see De Angelis, 1999, for a com-prehensive review on the major applications of FRET).The demonstration that the cyanobacteria circadianprotein KaiB can form homodimers, using biolumi-nescence RET (BRET; Xu et al., 1999), was among theearliest successful applications of this approach. Thestudy of conformational changes in proteins (Wanget al., 2001) or the kinetics of proteasome targeted deg-radation (Tung et al., 2000) are just two additional ex-amples of the exciting possibilities that these techniquesoffer for the real-time monitoring of interactive events.

The arrival of the proteomics age has brought alongthe intriguing possibility of a genome-wide applica-tion of the YTH and derived approaches (Uetz et al.,2000; Legrain and Selig, 2000; Ito et al., 2001). Severalconsortia are combining the power of robotics, multi-cloning, and high throughput screening to lay out

518 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 8: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

(more or less) comprehensive protein interactionmaps. It is hoped that the painstaking construction ofthese “interactomes” will usher in a new era ofproteomic databases. By “mining” these resources, wemight be able to generate sound hypotheses regardingthe possible role of uncharacterized proteins, oncetheir likely location within such a map is identified. Itis important to bear in mind, however, that two earlyattempts at defining a comprehensive catalogue of allpossible interactions in the yeast system yielded dis-appointingly few overlapping hits (Uetz et al., 2000;Ito et al., 2001). This result indicates that current tech-niques have not yet reached the level of saturationrequired to detect every possible interaction (Hazbunand Fields, 2001) and that the type of data obtaineddepends strongly on the details of a particular screendesign.

DISSECTION OF GENERAL ANDCIRCADIAN INTERACTION NETWORKS

Most signal transduction pathways use a standardapproach to deliver information into the nucleus.Although different pathways display a wide range ofvariations in the details, they operate in three concep-tually distinct domains designed to control the all-important crossing of the nuclear/cytoplasmic barrierby factors that affect transcription. In the first domain(domain I = input), the primary information, fromchannel traffic, receptor activation, intracellular vari-ables, and so on, is received by the proper sensor andconveyed through second messengers onto regula-tory proteins that control the fate of a nucleus-bounddevice (domain II = processing). The translocation,final destination, activity, and partners of this nucleardevice(s) determine the nature of the genetic response(domain III = output). Predictably, a dynamic balancebetween the rescue and degradation of these nucleus-bound proteins, and their regulators, is a recurringtheme in a large number of signaling pathways. Thiscontrol mechanism is particularly common in the sec-ond domain of this simplified model, where the stabil-ity and fate of key factors is determined, to a largeextent, by complex phosphorylation/dephosphorylationcascades and selective interactions.

The levels of organization defined above can beeasily recognized in many well-characterized path-ways, a few of which are outlined in Table 1. The firstpoint of this exercise is to emphasize that while infor-

mation in a cell is handled through “traditional”channels, distinguishing features within individualpathways are conferred by a rich tapestry of highlyspecific protein-protein contacts. A real understand-ing of any particular pathway (or circuit) will dependon a comprehensive description of the sequence, loca-tion, and timing of every specific interaction.

The second point, of course, is that the biologicalclock (see Lowrey and Takahashi, 2000; Chang andReppert, 2001, for recent reviews) can be construed asthe signal transduction pathway that keeps track ofinternal time. As such, the underlying transcriptional/translational mechanism that supports circadianrhythmicity can be fitted onto a similar model, once werecognize that it represents a special case in which, inaddition to outside cues, key variables along the (path)way can feed into itself to generate a self-sustainingloop (Shearman et al., 2000).

Consequently, the explosive progress in molecularcircadian biology in recent years can be largely accred-ited to the discovery of clock-relevant interactions(many through YTH screens) and the tacit effort todescribe a “circadian interactome.”

Consider the heterodimerization of PER (Period)and TIM (Timeless) in Drosophila (Sehgal et al., 1994;Vosshall et al., 1994, Gekakis et al., 1995); the light-dependent sequestration of dTIM by dCRY (Cerianiet al., 1999); the combinatorial capacity of the threemammalian PER proteins to interact among them-selves (Zylka et al., 1998) and with the CRYs(Cryptochromes) (Kume et al., 1999); the interaction,discovered through a YTH screen, between the appar-ently clock gene–specific transcription factors BMAL1and CLOCK (Gekakis et al., 1998) and their trans-criptional repression by a PER/CRY complex (Kumeet al., 1999). Each example turned out to be a pivotalassociation for the establishment of a circadian clock.More recently, we learned of the ability of CaseinKinase Iε/δ to recognize PER 1 and 2 as phosphoryl-ation substrates, thereby regulating their susceptibil-ity to proteosomal degradation (Suri et al., 2000;Camacho et al., 2001; Vielhaber et al., 2000). This find-ing retraced the previously reported light- inducedtyrosine phosphorylation of dTIM by DOUBLETIME,an event that targets TIM for ubiquitination andproteosomal degradation (Naidoo et al., 1999).

Similarly, the key scaffolding role of thenonrhythmic protein WC-2 is of particular interest inthis context because it seems to control the formationof a ternary complex as part of the timing mechanism

Baler / THE YEAST HYBRID SYSTEM 519

Page 9: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

in Neurospora (Collett et al., 2001). This hypothesis hasnow been confirmed by independent biochemicalmethods (Denault et al., 2001).

By the same token, it can be argued that the piecesstill missing in the circadian puzzle represent uniqueopportunities to find novel interactions that mightaffect the phase and output of the circadian clock. Itseems reasonable to predict that these unknown inter-actions are still many and likely to play importantroles in the compartments described above.

First, we do not yet fully understand the nature ofthe interactions underlying circadian photoreceptivesignaling. ZEITLUPE (ZTL), for example, is anArabidopsis protein that exerts profound effects onclock-controlled processes consistent with a role inlight input to the clock (Somers et al., 2000), possiblyjust downstream of converging photoreceptive path-ways (Millar, 2000). The role of ZTL within or aroundthe clock is far from clear; however, the presence ofF-box and PAS domains suggests likely classes of ZTLbinding partners belonging to the degradation andtranscription control pathways, respectively.

Second, we still have a far from complete descrip-tion of the components responsible for the specificmodification, degradation, and translocation eventsthat control the level and subcellular localization of allknown clock proteins in different species. One wouldassume that the turnover of specific clock compo-

nents should be a carefully monitored parameter, anassumption amply supported by the arrhythmicity ofDOUBLETIME mutant flies in which the hypo-phosphorylation of TIM (Naidoo et al., 1999) and PER(Price et al., 1998) results in their increased stability.Similarly, the tightly regulated nucleocytoplasmicshuttling of proteins like PER, TIM, and CRY is likelyto involve a number of specific (and not so specific)auxiliary factors that need to be identified. A similarcase can be made regarding the pathways and factorsinvolved in controlling the intranuclear traffickingand subnuclear localization of circadian transcriptionfactors. There is a growing number of examples inwhich these processes play an important regulatoryrole in determining the outcome of a specific trans-duction program (Stein et al., 2000). Similar mecha-nisms are likely to play a role in circadian molecularroutines.

Third, we are quite in the dark regarding the spe-cific interactions that must occur around circadianDNA consensus elements, such as E-boxes (orGATA-boxes in Neurospora), to induce selective tran-scription of clock-controlled genes. As a matter of fact,we do not even know (although we might suspect thisto be the case) whether clock components do contrib-ute to other pathways, as part of different protein con-stellations, in either oscillatory or linearly responsivenetworks. The one-hybrid system and its variations

520 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Table 1. Many different signal transduction pathways utilize common themes of protein interactions to process information and reprogramgene expression. The mechanisms involved operate in three domains to ultimately control the activity of nucleus-bound factors. Circadianclocks are no exception.

Pathway Domain I (Input) Domain II (Processing) Domain III (Output) Reference

Heat shock (HS)/ Abnormal protein HS factor trimerization HS gene activation, protection Baler et al., 1996; Soncinstress response production and nuclear translocation. from stress et al., 2000

Modulated by calcium-activated kinases

Nuclear hormone Ligand-mediated Receptor release from HSP90 Hormone-responsive gene Miyata et al., 1997receptors activation and nuclear translocation. activation, regulation of

Modulated by Tyr and reproductive (and other)CKII kinases tissues

Immune, anti Extracellular stimulation Inactive NFκB subfamily Activation of anti-apoptotic Zandi et al., 1997cell-death. (NFκB) through IL-1, TNF, members released from and immune response genes

and other cytokines IκB cytoplasmic inhibitors, such as IL-2, TNFα, GM-CSFnuclear translocation.Modulated by IKK kinases

Tumor suppressor Cell cycle, DNA Reduced affinity for Mdm2 Regulation of genes involved Maya et al., 2001protein (p53) damage, apoptosis (a p53 E3 ligase), nuclear in cell cycle arrest, apoptosis,

translocation. Modulated DNA damage repairby ATM kinase

Circadian clock Photic input, phase Per/Tim, Per/Cry Activation/suppression of Shearman et al., 2000;of the cellular clock heterodimerization, nuclear clock and clock-controlled Camacho et al., 2001(i.e., relative level translocation. Modulated genesof state variables) by CKIδ/ε

Page 10: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

(Wilson et al., 1991) might prove useful in this context,since they provide a means for the identification oftransacting factors and cis-acting regulatory elementpairs.

It is interesting to consider that the apparentabsence of a typical circadian clock mechanism in theyeast might have resulted in a particularly well-suitedenvironment for the search of clock-associated mole-cules through the various YH systems described. Wecan expect that the wider and deeper use of these tech-niques will illuminate new areas of the molecularchronobiology field, as the mapping of all possibleinteractions yields ever-denser networks of interact-ing proteins. I like to think that this illumination has instore for us many advances, few delays, and one ortwo paradigm shifts.

REFERENCES

Aronheim A, Zandi E, Hennemann H, Elledge S, andKarin M (1997) Isolation of an AP-1 repressor by a novelmethod for detecting protein-protein interactions. MolCell Biol 17:3094-3102.

Baler R, Zou J, and Voellmy R (1996) Evidence for a role ofHsp70 in the regulation of the heat shock response inmammalian cells. Cell Stress Chaperones 1:33-39.

Boulton S, Vincent S, and Vidal M (2001) Use of pro-tein-interaction maps to formulate biological questions.Curr Opin Chem Biol 5:57-62.

Brachmann R and Boeke J (1997) Tag games in yeast: Thetwo-hybrid system and beyond. Curr Opin Biotechnol8:561-568.

Brent R and Finley R (1997) Understanding gene and allelefunction with two-hybrid methods. Annu Rev Genet31:663-704.

Camacho F, Cilio M, Guo Y, Virshup DM, Patel K, KhorkovaO, Styren S, Morse B, Yao Z, and Keesler G (2001) Humancasein kinase I δ phosphorylation of human circadianclock proteins period 1 and 2. FEBS Lett 489:159-165.

Ceriani M, Darlington T, Staknis D, Mas P, Petti A, Weitz C,and Kay S (1999) Light-dependent sequestration ofTIMELESS by CRYPTOCHROME. Science 285:553-556.

Chang D and Reppert S (2001) The circadian clocks of miceand men. Neuron 29:555-558.

Chervitz S, Aravind L, Sherlock G, Ball C, Koonin E,Dwight S, Harris M, Dolinski K, Mohr S, Smith T, Weng S,Cherry J, and Botstein D (1998) Comparison of the com-plete protein sets of worm and yeast: Orthology anddivergence. Science 282:2022-2028.

Collett M, Dunlap J, and Loros J (2001) Circadian clock-spe-cific roles for the light response protein WHITECOLLAR-2. Mol Cell Biol 21:2619-2628.

De Angelis DA (1999) Why FRET over genomics? PhysiolGenom 1:93-99.

Denault D, Loros J, and Dunlap J (2001) WC-2 mediatesWC-1-FRQ interaction within the PAS protein-linked cir-cadian feedback loop of Neurospora. EMBO J 20:109-117.

Du W, Vidal M, Xie J, and Dyson N (1996) RBF, a novelRB-related gene that regulates E2F activity and interactswith cyclin E in Drosophila. Genes Dev 10:1206-1218.

Fashena S, Serebriiskii I, and Golemis E (2000) The contin-ued evolution of two-hybrid screening approaches inyeast: How to outwit different preys with different baits.Gene 250:1-14.

Fields S and Song O (1989) A novel genetic system to detectprotein-protein interactions. Nature 340:245-246.

Finley R and Brent R. (1996) Two-hybrid analysis of geneticregulatory networks. Proc Natl Acad Sci U S A 91:12980-12984.

Gekakis N, Saez L, Delahaye-Brown A, Myers M, Sehgal A,Young M, and Weitz C (1995) Isolation of timeless by PERprotein interaction: Defective interaction between time-less protein and long-period mutant PERL. Science270:811-815.

Gekakis N, Staknis D, Nguyen H, Davis F, Wilsbacher L,King D, Takahashi J, and Weitz C (1998) Role of theCLOCK protein in the mammalian circadian mechanism.Science 280:1564-1569.

Hartwell L, Hopfield J, Leibler S, and Murray A (1999) Frommolecular to modular cell biology. Nature 402:C47-C52.

Hazbun T and Fields S (2001) Networking proteins in yeast.Proc Natl Acad Sci U S A 98:4277-4278.

Inouye C, Remondelli P, Karin M, and Elledge S (1994) Isola-tion of a cDNA encoding a metal response element bind-ing protein using a novel expression cloning procedure:The one hybrid system. DNA Cell Biol 13:731-742.

Isakoff S, Cardozo T, Andreev J, Li Z, Ferguson K, AbagyanR, Lemmon M, Aronheim A, and Skolnik E (1998) Identi-fication and analysis of PH domain-containing targets ofphosphatidylinositol 3-kinase using a novel in vivo assayin yeast. EMBO J 17:5374-5387.

Ito T, Chiba T, Ozawa R, Yoshida M, Hattori M, and Sakaki Y(2001) A comprehensive two-hybrid analysis to explorethe yeast protein interactome. Proc Natl Acad Sci U S A98:4569-4574.

Johnsson N and Varshavsky A (1994) Split ubiquitin as asensor of protein interactions in vivo. Proc Natl Acad SciU S A 91:10340-10344.

Kochan J, Volpers C, and Osborne M (2000) The yeast tribidsystem: cDNA expression cloning of protein interactionsdependent on posttranslational modifications. MethodsEnzymol 328:111-127.

Kume K, Zylka M, Sriram S, Shearman L, Weaver D, Jin X,Maywood E, Hastings M, and Reppert S (1999) mCRY1and mCRY2 are essential components of the negativelimb of the circadian clock feedback loop. Cell 98:193-205.

Legrain P and Selig L (2000) Genome-wide protein interac-tion maps using two-hybrid systems. FEBS Lett480:32-36.

Li H, Ng E, Lee S, Kotaka M, Tsui S, Lee C, Fung K, andWaye M (2001) Protein-protein interaction of FHL3 withFHL2 and visualization of their interaction by green fluo-

Baler / THE YEAST HYBRID SYSTEM 521

Page 11: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

rescent proteins (GFP) two-fusion fluorescence reso-nance energy transfer (FRET). J Cell Biochem 80:293-303.

Lowrey P and Takahashi J (2000) Genetics of the mammaliancircadian system: Photic entrainment, circadian pace-maker mechanisms, and posttranslational regulation.Annu Rev Genet 34:533-562.

Maya R, Balass M, Kim S, Shkedy D, Leal J, Shifman O, MoasM, Buschmann T, Ronai Z, Shiloh Y, Kastan M, Katzir E,and Oren M (2001) ATM-dependent phosphorylation ofMdm2 on serine 395: Role in p53 activation by DNAdam-age. Genes Dev 15:1067-1077.

Medici R, Bianchi E, Di Segni G, and Tocchini-Valentini G(1997) Efficient signal transduction by a chimericyeast-mammalian G protein alpha subunit Gpa1-Gsalpha covalently fused to the yeast receptor Ste2. EMBO J16:7241-7249.

Millar A(2000) Clock proteins: Turned over after hours. CurrBiol 10:R529-R531.

Miyata Y, Chambraud B, Radanyi C, Leclerc J, Lebeau M,Renoir J, Shirai R, Catelli M, Yahara I, and Baulieu E(1997) Phosphorylation of the immunosuppressantFK506-binding protein FKBP52 by casein kinase II: Regu-lation of HSP90-binding activity of FKBP52. Proc NatlAcad Sci U S A 94:14500-14505.

Naidoo N, Song W, Hunter-Ensor M, and Sehgal A (1999) Arole for the proteasome in the light response of the time-less clock protein. Science 285:1737-1741.

Pause A, Peterson B, Schaffar G, Stearman R, and Klausner R(1999) Studying interactions of four proteins in the yeasttwo-hybrid system: Structural resemblance of thepVHL/elongin BC/hCUL-2 complex with the ubiquitinligase complex SKP1/cullin/F-box protein. Proc NatlAcad Sci U S A 96:9533-9538.

Price J, Blau J, Rothenfluh A, Abodeely M, Kloss B, andYoung M (1998) Double-time is a novel Drosophila clockgene that regulates PERIOD protein accumulation. Cell94:83-95.

Putz U, Skehel P, and Kuhl D (1996) A tri-hybrid system forthe analysis and detection of RNA–protein interactions.Nucleic Acids Res 24:4838-4840.

Rossi F, Blakely B, Charlton C, and Blau H (2000) Monitoringprotein-protein interactions in live mammalian cells bybeta-galactosidase complementation. Methods Enzymol328:231-251.

Sadowski S, Ma J, Triezenberg S, and Ptashne M (1988)Gal4-VP16 is an unusually potent transcriptional activa-tor. Nature 335:245-249.

Sehgal A, Price J, Man B, and Young M (1994) Loss of circa-dian behavioral rhythms and per RNA oscillations in theDrosophila mutant timeless. Science 263:1603-1606.

Serebriiskii I, Khazak V, and Golemis E (1999) A two-hybriddual bait system to discriminate specificity of proteininteractions. J Biol Chem 274:17080-17087.

Shearman L, Sriram S, Weaver D, Maywood E, Chaves I,Zheng B, Kume K, Lee C, van der Horst G, Hastings M,and Reppert S (2000) Interacting molecular loops in themammalian circadian clock. Science 288:1013-1019.

Somers D, Schultz T, Milnamow M, and Kay S (2000)ZEITLUPE, a novel clock associated PAS protein fromArabidopsis. Cell 101:319-329.

Soncin F, Asea A, Zhang X, Stevenson M, and Calderwood S(2000) Role of calcium activated kinases and phos-phatases in heat shock factor-1 activation. Int J Mol Med6:705-710.

Stein G, van Wijnen A, Stein J, Lian J, Montecino M, Choi J,Zaidi K, and Javed (2000) Intranuclear trafficking of tran-scription factors: Implications for biological control.J Cell Sci 113:2527-2533.

Suri V, Hall J, and Rosbash M (2000) Two novel doubletimemutants alter circadian properties and eliminate thedelay between RNA and protein in Drosophila. J Neurosci20:7547-7555.

Tung C, Mahmood U, Bredow S, and Weissleder R (2000) Invivo imaging of proteolytic enzyme activity using anovel molecular reporter. Cancer Res 60:4953-4958.

Uetz P, Giot L, Cagney G, Mansfield TA, Judson R, Knight JR,Lockshon D, Narayan V, Srinivasan M, Pochart P,Qureshi-Emili A, and Li Y (2000) Acomprehensive analy-sis of protein-protein interactions in Saccharomycescerevisiae. Nature 403:623-627.

Vidal M, Brachmann R, Fattaey A, Harlow E, and Boeke J(1996) Reverse two-hybrid and one-hybrid systems todetect dissociation of protein-protein and DNA-proteininteractions. Proc Natl Acad Sci U S A 93:10315-10320.

Vielhaber E, Eide E, Rivers A, Gao Z, and Virshup D (2000)Nuclear entry of the circadian regulator mPER1 is con-trolled by mammalian casein kinase I epsilon. Mol CellBiol 20:4888-4899.

Vosshall L, Price J, Sehgal A, Saez L, and Young M (1994)Block in nuclear localization of period protein by a sec-ond clock mutation, timeless. Science 263:1606-1609.

Wang Y, Wang G, O’Kane D, and Szalay A (2001) A study ofprotein-protein interactions in living cells using lumines-cence resonance energy transfer (LRET) from Renillaluciferase to Aequorea GFP. Mol Gen Genet 264:578-587.

Wilson T, Fahrner T, Johnston M, and Milbrandt J (1991)Identification of the DNA binding site for NGFI-B bygenetic selection in yeast. Science 252:1296-1300.

Xu Y, Piston D, and Johnson C (1999) Abioluminescence res-onance energy transfer (BRET) system: Application tointeracting circadian clock proteins. Proc Natl Acad SciU S A 96:151-156.

Zandi E, Rothwarf D, Delhase M, Hayakawa M, and KarinM (1997) The I kappaβ kinase complex (IKK) containstwo kinase subunits, IKKα and IKKβ, necessary forI kappaβ phosphorylation and NF-kappaβ activation.Cell 91:243-252.

Zhang J and Lautar S (1996) A yeast three-hybrid method toclone ternary protein complex components. AnalBiochem 242:68-72.

Zylka M, Shearman L, Levine J, Jin X, Weaver D, andReppert S (1998) Molecular analysis of mammalian time-less. Neuron 21:1115-1122.

522 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 12: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

JOURNAL OF BIOLOGICAL RHYTHMS / December 2001Yanovsky et al. / PHYTOCHROMES AND CRYPTOCHROMES IN ARABIDOPSIS

Resetting of the Circadian Clock by Phytochromesand Cryptochromes in Arabidopsis

Marcelo J. Yanovsky,*,1 M. Agustina Mazzella,* Garry C. Whitelam,† and Jorge J. Casal*,2

*IFEVA, Faculty of Agronomy, University of Buenos Aires, Avenida San Martín 4453,1417-Buenos Aires, Argentina, †Department of Biology, University of Leicester,

University Road, Leicester LE1 7RH, United Kingdom

Abstract The authors sought to investigate the role of phytochromes A and B(phyA and phyB) and cryptochromes 1 and 2 (cry1 and cry2) in the synchroniza-tion of the leaf position rhythm in Arabidopsis thaliana. The seedlings were trans-ferred from white light–dark cycles to free-running conditions with or withoutexposure to a light treatment during the final hours of the last dark period. Thephase advance caused by a far-red light treatment was absent in the phyAmutant,deficient in the fhy1 and fhy3 mutants involved in phyA signaling, and normal inthe cry1 and cry1 cry2 mutants. The phase shift caused by blue light was normal inthe cry2 mutant; reduced in the phyA, cry1, phyA cry1, and cry1 cry2 mutants; andabolished in the phyA cry1 cry2 triple mutant. The phase shift caused by red lightwas partially retained by the phyA phyB double mutant. The authors concludethat cry1 and cry2 participate as photoreceptors in the blue light input to theclock but are not required for the phyA-mediated effects on the phase of the circa-dian rhythm of leaf position. The signaling proteins FHY1 and FHY3 are sharedby phyA-mediated photomorphogenesis and phyA input to the clock.

Key words Arabidopsis, circadian rhythms, cryptochrome, leaf movement, light input,phytochrome

Light is acknowledged as the most important envi-ronmental cue involved in resetting the circadianclocks. Our understanding of the light input to theclock has improved very significantly in recent years.The discovery of cryptochromes first in plants (cry1,cry2; Ahmad and Cashmore, 1993; Guo et al., 1999),then in mammals (mcry1, mcry2; Hsu et al., 1996; Todoet al., 1996, Kobayashi et al., 1998), insects (Stanewskyet al., 1998), and so on, accounts for a significant deal ofthis advance. Cryptochromes have sequence homo-logy to distinct types of photolyases in plants and ani-mals and, for this reason, are believed to be the prod-uct of independent evolution events (Cashmore et al.,1999). Cryptochromes lack photolyase activity and

possess a distinctive C-terminal extension. In Droso-phila, cryptochrome appears to be the only photo-receptor involved in light input to the clock, and resid-ual effects of light on rhythmic behavior are consideredto be the indirect consequence of vision effects onbehavioral rhythms (Emery et al., 2000). In Drosophila,cryptochrome physically interacts with clock compo-nents (Ceriani et al., 1999). In mammals, the situationhas been more controversial. Cryptochromes werefirst proposed as photoreceptors involved in the lightinput to the circadian clock (Miyamoto and Sancar,1998). Although the observation that the mcry2mutant of mice shows reduced sensitivity to inductionof mPER1 (Thresher et al., 1998) is in favor of the latter

523

1. Current address: Department of Cell Biology, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA92037, USA.2. To whom all correspondence should be addressed: e-mail: [email protected].

JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 523-530© 2001 Sage Publications

Page 13: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

view, the idea lost support because mcry2 displayedincreased and not reduced sensitivity to light resettingof the clock (Thresher et al., 1998). The observationthat mcry1 mcry2 double mutants were arrythmic indarkness gave credit to a role of cryptochromes as cen-tral components of the mammalian clock (Van derHorst et al., 1999). More recently, experiments withmutant mice lacking mcry1, mcry2 as well as rods andmost cones have suggested a redundant role ofcryptochromes and opsins in the light input to theclock (Selby et al., 2000). Thus, cryptochromes inmammals would have a dual role as photoreceptorsand as central components of the clock. In plants, cry1and cry2 are clearly important for normal control ofgrowth and development by blue light (Lin, 2000).Mutations on these genes increase the period of rhyth-mic expression of a photosynthetic gene under certainfluence rates of blue light (Devlin and Kay, 2000;Somers et al., 1998). The double cry1 cry2 mutantretains robust rhythmicity in Arabidopsis, indicatingthat in contrast to the situation in mammals, crypto-chromes are not essential components of the clock(Yanovsky et al., 2000b; Devlin and Kay, 2000). Inplants, phytochromes A, B, D, and E (phyA, phyB,phyD, and phyE) are also involved in the input to theclock (Somers et al., 1998; Yanovsky et al., 2000a;Devlin and Kay, 2000). phyA is a light-modulatedkinase (Fankhauser et al., 1999; Yeh and Lagarias,1998). phyA and phyB are present in the cytoplasm indark-grown seedlings and migrate to the nucleusupon light activation (Kircher et al., 1999; Yamaguchiet al., 1999), where they interact with factors that bindto DNA (Martinez-García et al . , 2000) . Abacteriophytochrome is involved in resetting the cir-cadian clock in Synechococcus elongatus (Schmitz et al.,2000).

The identification of cryptochromes as circadianphotopigments requires a demonstration that thefunction of cryptochromes is directly related to theirlight-absorbing properties (Lucas and Foster, 1999).Selective spectral effects on biological rhythms havenot been demonstrated for cryptochromes. In plants,the effects of cry1 and cry2 on photomorphogenesisare observed under UV-A/blue light but do notextend to the red/far-red region of the spectrum.Phytochromes, in contrast, operate predominantly inthe red/far-red wavebands and to a lesser degree inthe blue region (Casal and Mazzella, 1998; Neff andChory, 1998), where the pigment shows a secondarypeak of absorption. Devlin and Kay (2000) recentlyobserved that the absence of both cry1 and cry2 causes

an increase in the length of the period not only underblue light but also under red light, indicating that cry-ptochromes are necessary for phyA signaling to theclock but not for phyAsignaling during photomorpho-genesis. In the present work, we use a different proto-col (Yanovsky et al., 2000a) to quantify the phase shiftof the circadian rhythm of Arabidopsis leaf position inresponse to different light qualities. The results showthat phyA-mediated resetting of the circadian clock byfar-red light is unaffected by the cry1 and cry2 muta-tions and impaired by the photomorphogenicmutants fhy1 and fhy3.

MATERIALS AND METHODS

Plants of Arabidopsis thaliana of the ecotypeLandsberg erecta or of the phyA-201 (Reed et al., 1994),fhy1, fhy3 (Whitelam et al., 1993), cry1-1 (Ahmad andCashmore, 1993), phyA-201 phyB-1 (Mazzella et al.,1997), cry1-1 cry2 (where cry2 is the fha-1 allele; Guoet al., 1999), phyA-201 cry1-1, phyA-201 cry1-1 cry2(Mazzella and Casal, 2001) were grown in 5 cm3 potscontaining a soil/sand medium. After sowing, thepots were incubated for 3 days in darkness at 7 °C,exposed to a saturating pulse of red light (30 min, 30µmol.m-2.s-1) to induce germination, and allowed togerminate for 24 h in darkness at 25 °C. The seedlingswere grown at 20 °C under a 12 h white light/12 h darkphotoperiod provided by fluorescent lamps (80 µmol.m-2.s-1).

After 2 weeks under 12-h photoperiods, during thelast 5 h of the dark period prior to transfer to free-run-ning conditions, the seedlings were either exposed tothe light treatment (far-red, red, or blue light) orremained as dark controls. The far-red light treatment(100 µmol.m-2.s-1) was provided by 60-Watt incandes-cent lamps in combination with a water filter, six blueacrylic filters (Paolini 2031), and two red acetate filters(La Casa del Acetato, Buenos Aires, Argentina). Bluelight (20 µmol.m-2.s-1) was provided by fluorescentlamps (Philips, TLF 40W/54) in combination with ablue acetate filter. Red light (40 µmol.m-2.s-1) was pro-vided by red fluorescent tubes (Philips, 40/15).

After transfer to free running conditions (20 °C,continuous white light 10 µmol.m-2.s-1), the position ofthe first pair of leaves was recorded every 2 h with adigital video camera (Quick Cam, Connectix Corpora-tion, San Mateo, CA, USA), and the angle between thepetioles was measured using the Scion Image analysisprogram (Scion Corporation, Frederick, MD, USA).

524 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 14: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Data were fitted, according to Millar et al. (1995), to theequation: L(t) = (co + c1t) + (a0 – a1t) sin [2π/T (t – φ)],where L is the leaf angle, co is the estimated value of theleaf angle at t = 0, c1 is an estimate of the linear rate ofchange in leaf angle, a0 is the estimated value of theamplitude at t = 0, a1 is an estimate of the linear changein cycling amplitude, t is time, T is the period estimate,and φ is the estimate of the phase at t = 0. The first 6 h ofmeasurements were not included among the dataused to estimate the parameters of the above equationto avoid effects produced by shifting the plants fromcolored to dim white light. The phases were calculatedby linear regression through the peaks in the days fol-lowing the light pulse, extrapolated back to the day ofthe pulse (Kondo et al., 1991; Roenneberg and Deng,1997). Phase shifts were calculated as the differencebetween the phases of light-treated and dark-controlplants. Data are means of at least six different plants,from three independent experiments. Period lengthand phase shift data are presented with their standarderrors.

RESULTS

Seedlings of Arabidopsis were grown under dailywhite light–dark cycles. Five hours before the end ofthe last dark period, different groups were exposedto far-red, red, or blue light while control seedlingsremained in darkness. At the end of the treatments (orobjective night in the controls), all the seedlings weretransferred to free-running conditions, that is, con-stant low fluence white light and constant tempera-ture (Fig. 1A). In control seedlings, the leaf tips movedupward, reaching the most erect position at the begin-ning of the subjective night. This was followed by theopposite movement leading to the least erect positionat the beginning of the subjective day (Fig. 1B). Thisrhythm continued for several days under free-runningconditions. The period of the rhythm was 25.6 h(Table 1). Exposure to 5 h of far-red, red, or blue lighthad no effects on the length of the period but caused apersistent phase advance of the circadian rhythm, thatis, each phase of the rhythm occurred at an earlier timepoint (Fig. 1B, Table 1).

The phyA, cry1, and cry1 cry2 mutants showed nor-mal rhythmicity when transferred to free-runningconditions without a selective light treatment (Fig. 2,Table 1). The phyA mutant showed no phase advancein response to the far-red light treatment and only par-tial phase advance (compared to the wild type) in

response to blue light. The cry1 and cry1 cry2 mutantsshowed a normal phase advance in response to far-redlight and partial phase advance by blue light (Fig. 2,Table 1). This indicates that under the present experi-mental conditions, the effects of cry1 and cry2 were

Yanovsky et al. / PHYTOCHROMES AND CRYPTOCHROMES IN ARABIDOPSIS 525

Figure 1. Far-red, red, or blue light reset the rhythm of leaf move-ment in Arabidopsis thaliana. The seedlings were grown underday-night cycles, given far-red light (100 µmol.m-2.s-1), red light(40 µmol.m-2.s-1), blue light (20 µmol.m-2.s-1), or no light treat-ment during the last 5 h of the final night and transferred to free-running conditions, that is, continuous white light and constanttemperature (20 °C). A. Experimental protocol. B. Leaf angle.

Page 15: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

wavelength selective (i.e., observed under blue lightbut not under far-red light).

The residual phase advance observed in the cry1and phyA mutants in response to blue light suggestedthat different photoreceptors could be redundantlyinvolved in the resetting of the circadian clock underthese conditions. To investigate this issue in furtherdetail, the phase shift was analyzed in multiplemutants. The dark controls of the cry2, cry1 cry2, phyAcry1, phyA cry2, and phyA cry1 cry2 mutants retainednormal rhythmicity (Fig. 3, Table 1). The phase shiftinduced by blue light in the cry1 cry2 and phyA cry1double mutants was reduced compared with the wildtype and similar to that observed in the cry1 singlemutant. Although the cry2 mutation did not reducethe phase shift produced in response to blue light inthe wild-type background (cf. cry2 and wild type) orcry1 mutant background (cf. cry1 cry2 and cry1), it dideliminate the residual phase shift observed in the cry1phyA mutant (cf. cry1 cry2 phyA and cry1 phyA)(Table 1). This indicates that phyA, cry1, and cry2 par-

ticipate as partially redundant components of the bluelight input to the clock.

The fhy1 and fhy3 mutants show impaired seedlingde-etiolation under continuous far-red light and havebeen implicated in selective branches of phyA signal-ing in photomorphogenesis (Barnes et al., 1996;Cerdán et al., 1999; Yanovsky et al., 2000c). To investi-gate whether these mutants also affect phyAsignalingto the clock, the seedlings were transferred to free-run-ning conditions with or without exposure to far-redlight during the final 5 h of the night. Both fhy1 andfhy3 showed normal rhythmicity but no phaseadvance of the circadian clock in response to far-redlight (Fig. 4, Table 1).

The effect of red light on the phase of the circadianclock was only partially reduced by the absence ofphyA and phyB (Fig. 5, Table 1). The analysis of seed-ling morphology (hypocotyl length, cotyledon un-folding) in response to different wavebands comparedwith darkness indicates that the phyA mutant is blindto far-red light (Whitelam et al., 1993), the phyA cry1cry2 mutant is blind to blue light (data not shown), andthe phyA phyB mutant is blind to red light (Mazzellaet al., 1997). Thus, additional red light photoreceptorscontrol the phase of the circadian clock compared withseedling morphology during de-etiolation.

DISCUSSION

cry1 and cry2 are known to affect the period of theclock under blue light (Devlin and Kay, 2000; Somerset al., 1998). Since the double cry1 cry2 mutant retainsrobust rhythmicity, cry1 and cry2 cannot be essentialcomponents of the clock (Devlin and Kay, 2000;Yanovsky et al., 2000b). The results presented here go astep further by showing that cry1 and cry2 are specifi-cally involved in the phase response to blue light(Table 1). This observation provides a link between therole of cry1 and cry2 in the control of circadianrhythms and their spectral properties and supportsthe role of cryptochromes as circadian photoreceptorsin plants.

Resetting of the circadian rhythm of leaf movementby far-red light is mediated by phyA (Yanovsky et al.,2000a). This action of phyA was unaffected by the cry1and cry2 mutations (Table 1, Fig. 2) and abolished bythe fhy1 and fhy3 mutations (Fig. 4). An effect of phyAwas also observed under blue light in the cry1, cry2,and cry1 cry2 backgrounds (Table 1). Under fluences ofred light lower than 5 µmol•m-2 s-1, the period of the

526 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Table 1. Period of the leaf movement rhythm in Arabidopsisthaliana under free-running conditions and phase shift caused by 5 hof far-red (100 µmol.m-2.s-1), red (40 µmol.m-2.s-1), or blue light (20µmol.m-2.s-1) compared with the controls that remained in dark-ness. The seedlings were grown under day-night cycles and giventhe light treatments during the last 5 h of the last night before trans-fer to free-running conditions.

Light Pulse Genotype Period (h) Phase-Shift (h)

None WT 25.6 ± 0.4phyA 26.1 ± 0.4cry1 25.6 ± 0.5cry2 25.4 ± 0.5fhy1 25.6 ± 0.4fhy3 24.7 ± 0.4

phyA phyB 24.7 ± 0.3cry1cry2 26.6 ± 0.4

phyA cry2 25.7 ± 0.5phyA cry1 25.8 ± 0.5

phyA cry1 cry2 26.5 ± 0.6Far-red WT 24.8 ± 0.4 5.7 ± 0.6

phyA 25.6 ± 0.4 0.6 ± 0.5cry1 24.4 ± 0.5 6.1 ± 0.6

cry1cry2 26.3 ± 0.5 5.6 ± 0.6fhy1 25.1 ± 0.5 0.6 ± 0.4fhy3 24.6 ± 0.6 0.9 ± 0.5

Blue WT 25.8 ± 0.5 8.6 ± 0.6phyA 25.6 ± 0.3 5.1 ± 0.6cry1 24.4 ± 0.5 4.6 ± 0.5cry2 24.6 ± 0.6 8.0 ± 0.6

cry1 cry2 28.1 ± 0.6 5.1 ± 0.7phyA cry2 24.9 ± 0.5 5.9 ± 0.6phyA cry1 25.8 ± 0.4 3.7 ± 0.5

phyA cry1 cry2 26.2 ± 0.6 0.4 ± 0.5Red WT 26.2 ± 0.6 3.9 ± 0.5

phyAphyB 24.8 ± 0.4 2.2 ± 0.5

Page 16: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

rhythmic bioluminiscence driven by the luciferacegene under the control of a light-harvesting complexgene promoter (CAB:LUC) is extended by the phyAmutation (Somers et al., 1998). A similar effect iscaused by the cry1 mutation (Devlin and Kay, 2000).

Since phyA and cry1 mutants also cause extended peri-ods over the same range of fluence rates of white lightand the phyA cry1 double mutant shows a deficiencythat is not larger than that observed in any of the singlemutant parental lines, cry1 appears to be necessary for

Yanovsky et al. / PHYTOCHROMES AND CRYPTOCHROMES IN ARABIDOPSIS 527

Figure 2. Far-red light resets the rhythm of leaf movement in the cry1 mutant but not in the phyA mutant, whereas blue light resets therhythm in both mutants. The seedlings were grown under day-night cycles, given far-red light (100 µmol.m-2.s-1), blue light (20µmol.m-2.s-1), or no light treatment during the last 5 h of the final night and transferred to free-running conditions.

Figure 3. Blue light resets the rhythm of leaf movement in cry1 cry2 and phyA cry1 but not in phyA cry1 cry2. The seedlings were grownunder day-night cycles, given blue light (20 µmol.m-2.s-1) or no light treatment during the last 5 h of the final night and transferred tofree-running conditions.

Page 17: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

phyA signaling to the clock (Devlin and Kay, 2000).Our data show that cry1 is not universally required forphyA signaling to the clock.

Both sets of data can be reconciled by consideringthat phyA signals via two different pathways that canbe dissected photobiologically (Casal et al., 2000),

genetically (Yanovsky et al., 1997; Yanovsky et al.,2000c), and molecularly (Cerdán et al., 2000). Theso-called very-low-fluence response (VLFR) pathwayoperates under red or far-red light and saturates withinfrequent (hourly) light pulses. The high-irradianceresponse (HIR) pathway requires sustained (continu-ous or very frequently pulsed) far-red light. phyArequires cry1 under red light (Devlin and Kay, 2000),that is, under conditions where phyA operates via theVLFR pathway. Hourly pulses of far-red light (insteadof continuous far-red light) were fully ineffective toreset the circadian rhythm of leaf movement inArabidopsis (data not shown), indicating that phyAoperated via the HIR pathway. This is consistent withthe dramatic failure to reset the clock in response tofar-red light observed in fhy3, a mutant impaired inHIR but retaining VLFR (Yanovsky et al., 2000c). Inconclusion, we propose that cry1 is required for phyAsignaling to the clock in the VLFR but not in the HIRmode. This working hypothesis should be tested in asystem where both VLFR and HIR operate.

Little is known about the mechanisms by whichphytochromes and cryptochromes signal to the clock.In Drosophila, for instance, cryptochrome shows directinteraction with clock components (Ceriani et al.,1999). The observation that phyArequires cry1 for sig-naling to the clock but not for photomorphogenesis(Devlin and Kay, 2000) suggests that some elementscould be different between both processes. The failurein photomorphogenesis and far-red-mediated phase

528 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 4. Far-red light fails to reset the rhythm of leaf movement in the fhy1 and fhy3 mutants involved in phyA signaling. The seedlingswere grown under day-night cycles, given far-red light (100 µmol.m-2.s-1) or no light treatment during the last 5 h of the final night andtransferred to free-running conditions.

Figure 5. Red light resets the rhythm of leaf movement in thephyA phyB double mutant. The seedlings were grown underday-night cycles, given red light (40 µmol.m-2.s-1) or no light treat-ment during the last 5 h of the final night and transferred tofree-running conditions.

Page 18: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

shifting observed in fhy1 and fhy3 demonstrates that,at least in the HIR mode, some signaling elements areshared between the two processes. fhy1 and fhy3 arepredicted to operate downstream phyA in the light-signaling cascade. This argues against a direct actionof phyA on clock components as observed for crypto-chrome in Drosophila. The effects of fhy1 and fhy3 areconsistent with the earlier conclusion that phyA,although affecting the response to several wavebands,is not an essential component of the clock itself.

The phyA phyB cry1 cry2 quadruple mutant is virtu-ally blind to white light for de-etiolation but retainsrobust circadian rhythmicity implicating otherphotoreceptor(s) involved in the entrainment of thecircadian clock (Yanovsky et al., 2000b). The observa-tions that the phyA cry1 cry2 triple mutant shows nophase shift in response to blue light whereas the phyAphyB double mutant retains a small but significantphase shift under red light points toward a red-lightphotoreceptor as the pigment involved in changingthe phase of the rhythm in response to white light inthe quadruple mutant. The small effect of red lightcould accumulate over several days. Since the periodof the circadian rhythm of CAB::LUC bioluminiscenceis extended in the phyA phyB phyD and phyA phyB phyEtriple mutants compared with the phyA phyB doublemutant (Devlin and Kay, 2000), phyD and phyE aregood candidates. phyD and phyE have very weakeffects during de-etiolation (Aukerman et al., 1997;Devlin et al., 1998), and this favors the idea that likeflies and mammals, plants could also have photo-receptors predominantly or exclusively involved incircadian rhythms (Yanovsky et al., 2000a).

ACKNOWLEDGMENTS

This work was supported by grants from FONCYT(BID 1201/OC-AR - PICT 06739) and FundaciónAntorchas (A-13622/1-40) to JJC.

REFERENCES

Ahmad M and Cashmore AR (1993) HY4 gene of Arabidopsisthaliana encodes a protein with characteristics of ablue-light photoreceptor. Nature 366:162-166.

Aukerman MJ, Hirschfeld M, Wester L, Weaver M, Clack T,Amasino, RM, and Sharrock RA (1997) A deletion in thePHYD gene of the Arabidopsis Wassilewskija ecotypedefines a role for phytochrome D in red/far-red lightsensing. Plant Cell 9:1317-1326.

Barnes SA, Quaggio RB, Whitelam GC, and Chua N-H(1996) fhy1 defines a branch point in phytochrome A sig-nal transduction pathways of gene expression. Plant J10:1155-1161.

Casal JJ and Mazzella MA (1998) Conditional synergismbetween cryptochrome 1 and phytochrome B is shown bythe analysis of phyA, phyB and hy4 simple, double andtriple mutants in Arabidopsis. Plant Physiol 118:19-25.

Casal JJ, Yanovsky MJ, and Luppi JP (2000) Two photo-biological pathways of phytochrome A activity, only oneof which shows dominant negative suppression byphytochrome B. Photochem Photobiol 71:481-486.

Cashmore AR, Jarillo JA, Wu Y-J, and Liu D (1999) Crypto-chromes: Blue light receptors for plants and animals. Sci-ence 284:760-765.

Cerdán PD, Staneloni RJ, Ortega J, Bunge MM, Rodriguez-Batiller MJ, Sánchez RA, and Casal JJ (2000) Sustained butnot transient phytochrome Asignaling targets a region ofa Lhcb1*2 promoter that is not necessary for phytochromeB action. Plant Cell 12:1203-1211.

Cerdán PD, Yanovsky MJ, Reymundo FC, Nagatani A,Staneloni RJ, Whitelam GC, and Casal JJ (1999) Regula-tion of phytochrome B signaling by phytochrome A andFHY1 in Arabidopsis thaliana. Plant J 18:499-507.

Ceriani MF, Darlington TK, Staknis D, Mas P, Weitz CJ, andKay SA (1999) Light-dependent sequestration of TIME-LESS by CRYPTOCHROME. Science 285: 1599-1603.

Devlin PF and Kay SA (2000) Cryptochromes are requiredfor phytochrome signaling to the circadian clock but notfor rhythmicity. Plant Cell 12:2499-2509.

Devlin PF, Patel RS, and Whitelam GC (1998) PhytochromeE influences internode elongation and flowering time inArabidopsis. Plant Cell 10:1479-1487.

Emery P, Staneswsky R, Hall J, and Rosbash M (2000)Drosophila cryptochromes: A unique circadian-rhythmphotoreceptor. Nature 404:456-457.

Fankhauser C, Yeh KC, Lagarias JC, Zhang H, Elich TD, andChory J (1999) PKS1, a substrate phosphorylated by phyto-chrome that modulates light signaling in Arabidopsis. Sci-ence 284:1539-1541.

Guo H, Duong H, Ma N, and Lin C (1999) The Arabidopsisblue-light receptor cryptochrome 2 is a nuclear proteinregulated by a blue-light dependent post-transcriptionalmechanism. Plant J 19:279-289.

Hsu DS, Zhao XD, Zhao SY, Kazantsev A, Wang RP, Todo T,Wei YF, and Sancar A (1996) Putative human blue-lightphotoreceptor hCRY1 and hCRY2 are flavoproteins.Biochem 35:13871-13877.

Kircher S, Kozma-Bognar L, Kim L, Adam E, Harter K,Schäfer E, and Nagy F (1999) Light-quality dependentnuclear import of the plant photoreceptors phytochromeA and B. Plant Cell 11:1445-1456.

Kobayashi K, Kanno S, Smit B, Van der Horst GTJ, Takao M,and Ysui, S (1998) Characterization of photolyase/blue-light receptor homologs in mouse and humans cells.Nucleic Acid Res 26:5086-5092.

Kondo T, Johnson CH, and Hastings JW (1991) Action spec-trum for resetting the circadian phototaxis rhythm in theCW15 strain of Chlamydomonas: I. Cells in darkness.Plant Physiol 95:197-205.

Yanovsky et al. / PHYTOCHROMES AND CRYPTOCHROMES IN ARABIDOPSIS 529

Page 19: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Lin C (2000) Photoreceptors and regulation of floweringtime. Plant Physiol 123:39-50.

Lucas RJ and Foster RG (1999) Something to cry about? CurrBiol 9:214-217.

Martinez-García JF, Huq E, and Quail PH (2000) Direct tar-geting of light signals to a promoter element-bound tran-scription factor. Science 288:859-863.

Mazzella MA, Alconada Magliano TM, and Casal JJ (1997)Dual effect of phytochrome A on hypocotyl growth undercontinuous red light. Plant Cell Environ 20: 261-267.

Mazzella MA and Casal JJ (2001) Interactive signalling byphytochromes and cryptochromes generates de-etiolationhomeostasis in Arabidopsis thaliana. Plant Cell Environ24:155-162.

Millar AJ, Straume M, Chory J, Chua NH, and Kay SA(1995)The regulation of circadian period by phototransductionpathways in Arabidopsis. Science 267:1163-1166.

Miyamoto Y and Sancar A (1998) Vitamin B-2-basedblue-light photoreceptors in the retinohypothalamictract as the photoactive pigments for setting the circadianclock in mammals. Proc Natl Acad Sci U S A95:6097-6102.

Neff M and Chory J (1998) Genetic interactions betweenphytochrome A, phytochrome B and cryptochrome 1during Arabidopsis development. Plant Physiol 118:27-36.

Reed JW, Nagatani A, Elich TD, Fagan M, and Chory J (1994)Phytochrome A and phytochrome B have overlappingbut distinct functions in Arabidopsis development. PlantPhysiol 104:1139-1149.

Roenneberg T and Deng T-S (1997) Photobiology of theGonyaulax circadian system: I. Different phase responsecurves for red and blue light. Planta 202:494-501.

Schmitz O, Katayama M, Willams SB, Kondo T, and GoldenSS (2000) CikA, a bacteriophytochrome that resets thecyanobacterial circadian clock. Science 289:765-767.

Selby CP, Thompson C, Schmitz TM, Van Gelder RN, andSancar A (2000) Functional redundancy of cryptochromesand classical photoreceptors for nonvisual ocularphotoreception in mice. Proc Natl Acad Sci U S A 97:14697-14702.

Somers DE, Devlin PF, and Kay SA(1998) Phytochromes andcryptochromes in the entrainment of the Arabidopsis cir-cadian clock. Science 282:1488-1490.

Stanewsky R, Kaneko M, Emery P, Beretta B, Wager-Smith K,Kay SA, Rosbash M, and Hall JC (1998) The cryb mutation

identifies cryptochrome as a circadian photoreceptor inDrosophila. Cell 95:681-692.

Thresher RJ, Vitaterna MH, and Miyamoto Y (1998) Role ofmouse cryptochrome blue light photoreceptor in Droso-phila. Cell 95:681-692.

Todo T, Ryo K, Yamamoto H, Toh T, Inui H, Ayaki T, andNomura M (1996) Similarity among the Drosophila (6-4)photolyase, a human photolyase homolog, and the DNAphotolyase-blue-light photoreceptor family. Science 272:109-112.

Van der Horst GT, Muijtjens M, Kobayashi K, Takano R,Kanno S, Takao M, de Wit J, Verkerk A, Eker AP, vanLeenen D, Nuijs R, Bootsma D, Hoeijmakers JH, andYasui A (1999) Mammalian Cry1 and Cry2 are essentialfor maintenance of circadian rhythms. Nature 398:627-630.

Whitelam GC, Johnson E, Peng J, Carol P, Anderson ML,Cowl JS, and Harberd NP (1993) Phytochrome A nullmutants of Arabidopsis display a wild-type phenotype inwhite light. Plant Cell 5:757-768.

Yamaguchi R, Nakamura M, Mochizuki N, Kay SA, andNagatani A (1999) Light-dependent translocation of aphytochrome B-GFP fusion protein to the nucleus intransgenic Arabidopsis. J Cell Biol 145:437-445

Yanovsky MJ, Casal JJ, and Luppi JP (1997) The VLF loci,polymorphic between ecotypes Landsberg erecta andColumbia dissect two branches of phytochrome Asignal-ling pathways that correspond to the very-low fluenceand high-irradiance responses of phytochrome. Plant J12:659-667.

Yanovsky MJ, Izaguirre M, Wagmaister JA, Gatz C, JacksonSD, Thomas B, and Casal JJ (2000a) Phytochrome Aresetsthe circadian clock and delays tuber formation underlong days in potato. Plant J 23:223-232.

Yanovsky MJ, Mazzella MA, and Casal JJ (2000b) A quadru-ple photoreceptor mutant still keeps track of time. Cur-rent Biol 10:1013-1015.

Yanovsky MJ, Whitelam GC, and Casal JJ (2000c) fhy3-1retains inductive responses of phytochrome A. PlantPhysiol 123:235-242.

Yeh K-C and Lagarias C (1998) Eukariotic phytochromes:Light-regulated serine/threonine protein kinases withhistidine kinase ancestry. Proc Natl Acad Sci U S A 95:13976-13981.

530 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 20: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

JOURNAL OF BIOLOGICAL RHYTHMS / December 2001Shearman and Weaver / DISTINCT PHARMACOLOGICAL MECHANISMS

Distinct Pharmacological MechanismsLeading to c-fos Gene Expression inthe Fetal Suprachiasmatic Nucleus

Lauren P. Shearman1 and David R. Weaver2

Laboratory of Developmental Chronobiology, MassGeneral Hospital for Children,Massachusetts General Hospital, Boston, MA 02114, USA, and Department of Pediatrics,

Harvard Medical School, Boston, MA 02115, USA

Abstract Maternal treatment with cocaine or a D1-dopamine receptor agonistinduces c-fos gene expression in the fetal suprachiasmatic nuclei (SCN). Othertreatments that induce c-fos expression in the fetal SCN include caffeine and nico-tine. In the current article, the authors assessed whether these different pharma-cological treatments activate c-fos expression by a common neurochemical mech-anism. The results indicate the presence of at least two distinct pharmacologicalpathways to c-fos expression in the fetal rat SCN. Previous studies demonstratethat prenatal activation of dopamine receptors affects the developing circadiansystem. The present work shows that stimulant drugs influence the fetal brainthrough multiple transmitter systems and further suggests that there may bemultiple pathways leading to entrainment of the fetal biological clock.

Key words circadian rhythms, D1-dopamine receptor, gene expression, entrainment, SKF38393, caffeine, nicotine

The suprachiasmatic nuclei (SCN) contain a biolog-ical clock that regulates circadian rhythmicity in mam-mals (Klein et al., 1991; Weaver, 1998). The biologicalclock within the SCN is oscillating prior to birth inrodents, and the timing of the fetal clock is set(entrained) prior to birth (Davis, 1997; Reppert andWeaver, 1991). It appears that redundant signals fromthe mother are normally involved in entraining thefetal SCN (Reppert and Weaver, 1991; Viswanathanet al., 1994; Viswanathan and Davis, 1997).

Prenatal administration of melatonin and theD1-dopamine receptor agonist, SKF 38393, can set thephase of the fetal biological clock (Davis andMannion, 1988; Viswanathan et al., 1994; Viswanathanand Davis, 1997). Both D1 receptors and melatoninreceptors are expressed in the developing rodent SCN

(Bender et al., 1997; Carlson et al., 1991; Duffield et al.,1999; Shearman et al., 1997; Strother et al., 1998a;Weaver et al., 1992). Activation of these receptor types,which have opposite effects on cyclic AMP levels, setsthe phase of the fetal clock to opposite phases in ham-sters (Viswanathan and Davis, 1997). Prenatal entrain-ment by these pharmacological treatments demon-strates that functional receptors are present in the fetalbrain. The correlation of entrainment with c-fos geneinduction in the fetal SCN following dopaminergicdrug treatment suggests that induction of c-fos geneexpression is a useful marker for functionally relevantstimulation of the SCN. It is important to note, how-ever, that failure of a treatment to induce c-fos geneexpression does not necessarily indicate insensitivityto this input. For example, melatonin entrains the fetal

531

1. Current address: Department of Animal Pharmacology, Merck Research Laboratories, P.O. Box 2000, Rahway, NJ 07065.2. To whom all correspondence should be addressed: David R. Weaver, Ph.D., Department of Neurobiology, University of Mas-sachusetts Medical School, 55 Lake Avenue North, Worcester, MA 01655; e-mail: [email protected].

JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 531-540© 2001 Sage Publications

Page 21: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

SCN without induction of c-fos expression (Viswanathanand Davis, 1997).

Nicotine also has been reported to induce c-fos geneexpression in the fetal SCN (Clegg et al., 1995; O’Haraet al., 1998; see O’Hara et al., 1999, for review). Theeffects of nicotine on c-fos mRNA levels in the rat SCNare restricted to the perinatal period (O’Hara et al.,1999). Prenatal nicotine treatment leads to increasedexpression of other immediate early genes, includingjunB (O’Hara et al., 1999). In these respects, prenataltreatment with nicotine is similar to prenatal treat-ment with SKF 38393, which has a transient, perinatalperiod of efficacy and leads to induction of severalimmediate early genes (Weaver and Reppert, 1995;Weaver et al., 1995). The possibility that prenatal nico-tine treatment alters entrainment or other aspects ofdeveloping circadian function has not been examined.In adult rodents, however, nicotine treatment cancause phase shifts (Ferguson et al., 1999; O’Hara et al.,1999; Trachsel et al., 1995), and there are nicotinic com-ponents to light-induced c-fos expression and light-induced phase shifts (Keefe et al., 1987; O’Hara et al.,1998; Zhang et al., 1993).

We are interested in identifying agents that inducec-fos gene expression in the fetal SCN, as these agentsmay reveal receptor-mediated mechanisms for influ-encing the SCN relevant to prenatal entrainment. Caf-feine treatment induces c-fos gene expression in adultrodents that is limited to the striatum (Nakajima et al.,1989). Our interest in the functional development ofadenosine receptors (Shearman and Weaver, 1997;Weaver, 1993, 1996) led us to perform preliminarystudies on the effects of caffeine on c-fos gene expres-sion in fetal brain. To our surprise, maternal caffeinetreatment caused a widespread increase in c-fosexpression in the fetal brain, most notably in the SCN.The anatomical pattern of c-fos expression in the fetalbrain after maternal caffeine treatment did not closelymatch the distribution of expression of either A1 or A2A

adenosine receptor subtypes, but it did resemble thepattern observed after maternal treatment with SKF38393 (Shearman et al., 1997; Fig. 1). The patterns ofc-fos gene expression after SKF 38393 and caffeine,while similar, are distinct from the restricted pattern ofc-fos gene expression in the fetal brain after maternaltreatment with nicotine (Clegg et al., 1995). Theseobservations suggested to us that caffeine and SKF38393 may share a convergent neurochemical mecha-nism for induction of c-fos expression, while the effectsof nicotine may be independent. In the present report,a pharmacological approach was used to identify the

receptor types leading to c-fos gene expression in thefetal brain. The results indicate that there are multipleand distinct pharmacological mechanisms leading toc-fos gene expression in the fetal SCN.

MATERIALS AND METHODS

Animals and Drug Treatments

Timed-pregnant Sprague-Dawley rats (MBM:VAF,Zivic Miller Laboratories, Zelienople, PA, USA) werehoused in a centralized vivarium and were exposed toa light-dark cycle consisting of 12 h light:12 h dark,with lights on at 0600 h EST (LD). During the darkphase of the cycle, animals were exposed to red light.Gestational day (GD) 0 is defined as beginning themorning after overnight pairing resulting in a spermplug. Rats were studied in the afternoon, 6 to 10 h afterlights-on, on GD 20.

Each pregnant rat received two intraperitonealinjections, 30 min apart. The first injection was apotential antagonist or vehicle. The antagonists usedwere the D1-dopamine receptor antagonist SCH 23390(0.5 mg/kg), the nicotinic cholinergic antagonistmecamylamine (Meca) (5 mg/kg), and the NMDAreceptor antagonist MK 801 (0.5 mg/kg). The secondinjection was an agent to induce c-fos gene expression,SKF 38393 (10 mg/kg), caffeine (100 mg/ kg), nicotine(free base, 1 mg/kg), or vehicle. Limitations on thenumber of animals that could be treated on 1 day andthe number of sections that could be processedtogether for in situ hybridization dictated that threeseparate experiments be conducted, with each experi-ment consisting of all four pretreatments in combina-tion with two posttreatments (one inducing drug andits vehicle). Doses were selected on the basis of prelim-inary dose-response studies (caffeine; see alsoNakajima et al., 1989) or on the basis of responsesreported in the literature (Clegg et al., 1995; Weaver andClemens, 1987; Weaver et al., 1992). In preliminarystudies, we noted that the noncompetitive N-methyl-D-aspartate (NMDA) receptor antagonist, MK 801,was not well tolerated by fetal rats at the doses mostfrequently used in the literature (1.0-3.0 mg.kg;Kennaway and Moyer, 1999; Nakazato et al., 1998;Svenningsson et al., 1996), and so we used a lowerdose (0.5 mg/kg). Injections were given at a volume of2 mL/kg. Dams were killed by decapitation 30 minafter the second injection. Fetuses were then removedfrom the uterus and decapitated. Fetal heads were

532 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 22: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

frozen in cooled 2-methybutane (–20 °C) and stored at–80 °C. Animal studies were reviewed and approvedby the Massachusetts General Hospital Subcommitteeon Research Animal Care.

In Situ Hybridization

Sections (15 microns) were cut on a Bright-Hackercryostat at –20 °C and thaw-mounted onto slidescoated with Vectabond (Vector Labs, Burlingame, CA,USA). In situ hybridization was used to detect c-fosmRNA. Antisense and sense (control) cRNA probeswere produced from linearized plasmid DNA by invitro transcription in the presence of 35S-alpha- thio-UTP (1100-1300 Ci/mmol, NEN). Probes were puri-fied by extraction and ethanol precipitation prior touse. The coding region of the mouse c-fos cDNA inpGEM3 (plasmid MUSFOS3, kindly provided by Dr.Michael E. Greenberg) was used as the template in thetranscription reaction, as previously described

(Weaver et al., 1992). Prehybridization, hybridization,and wash conditions were as previously described(Weaver, 1993). Film autoradiograms were generatedby apposition of slides to Kodak SB-5 film for 10 to 15days. Radioactive standards (14C, 20-micron thickness;American Radiolabeled Chemical, St. Louis, MO,USA) were included on each film.

Data Analysis

Optical density measurements were performedusing a computer-based image analysis system as pre-viously described (Weaver, 1993). In situ hybridiza-tion results are presented as SCN relative optical den-sity (OD) values, defined as the OD of the SCNdivided by the OD of the adjacent hypothalamus inthe same section. Relative OD values from the two tothree sections having the most intense SCN hybridiza-tion signal were averaged to give the relative ODvalue for each fetus. To determine whether

Shearman and Weaver / DISTINCT PHARMACOLOGICAL MECHANISMS 533

Figure 1. c-fos gene expression in fetal rat brain after administration of (A) SKF 38393, (B) caffeine, or (C) vehicle. Sections shown in panelsA, B, and C were hybridized with an antisense probe to detect c-fos mRNA. Panel D represents a section adjacent to the section shown in Bbut hybridized with a sense strand (control) probe to illustrate nonspecific hybridization. Abbreviations: CP = caudate-putamen (striatum);DE = dorsal endopiriform nucleus; SCN = suprachiasmatic nuclei.

Page 23: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

pretreatments had a statistically significant impact onc-fos gene expression, data (OD values) were analyzedby one-way analysis of variance and Dunnett’s testusing Statview version 1.03 on a Macintosh computer.Pairwise comparisons were performed using Stu-dent’s t test. Statistical significance was set at p < 0.05.

For production of the figures, images were cap-tured using a Polaroid DMC IE digital camera andimported into Photoshop 5.0 for assembly of compos-ite images.

Chemicals and Drugs

SKF 38393, SCH 23390, MK 801, and Meca wereobtained from Research Biochemicals, Inc., division ofSigma-Aldrich, Inc. (Natick, MA, USA). Caffeine andnicotine (free base) were purchased from SigmaChemical Co. (St. Louis, MO, USA). General lab chem-icals were purchased from Sigma or Fisher Scientific(Springfield, NJ, USA).

RESULTS

Maternal treatment with caffeine (100 mg/kg) orSKF 38393 (10 mg/kg) significantly increased c-fosgene expression in the fetal SCN (Fig. 1). The two drugtreatments produced a grossly similar, widespreadpattern of c-fos expression in the fetal brain, extendingbeyond the SCN and including the striatum and otherregions (Fig. 1 and data not shown). The similarity inpattern between fetuses exposed to the dopaminergicagonist and to caffeine suggested that there might be asimilar neurochemical mechanisms, for example, thatthe effects of caffeine might be exerted through indi-rect effects on dopamine systems in the fetal brain(Ferre et al., 1992; Fredholm, 1995; Garrett andHoltzman, 1994). To assess this possibility, pregnantdams were pretreated with antagonists prior to receiv-ing a stimulant drug.

The antagonists used had differential effects onstimulant-induced c-fos expression in the SCN (Figs. 2,3). Pretreatment with the D1-dopamine receptorantagonist, SCH 23390, prevented the response to SKF38393 (Figs. 2E, 3A). Pretreatment with the nicotiniccholinergic receptor antagonist Meca or the NMDAreceptor antagonist MK 801 did not affect the SCNresponse to SKF 38393 (Figs. 2H, 2K, 3A). MK 801 pre-treatment did, however, markedly reduce the induc-tion of c-fos expression in the lateral portion of the fetalstriatum (Fig. 2K).

Caffeine-induced c-fos expression in the SCN (Fig. 2C)was not prevented by pretreatment with SCH 23390(Figs. 2F, 3B). In fact, SCN c-fos mRNA levels wereactually higher following treatment with SCH 23390plus caffeine than after treatment with vehicle pluscaffeine (Fig. 3A; p < 0.05, Dunnett’s test). This increasemay be due in part to an increase in c-fos expressiondue to treatment with SCH 23390. Remarkably, how-ever, the induction of c-fos in the striatum was com-pletely prevented by the D1 receptor antagonist (com-pare Figs. 2C and 2F). Both Meca and MK 801 reducedthe amplitude of the SCN response to caffeine (Fig. 2I,2L; Fig. 3). MK 801 also reduced the striatal response tocaffeine (Fig. 2L).

Experiments were also conducted to assess thepharmacological characteristics of c-fos induction inresponse to nicotine treatment (Figs. 4, 5). Preliminaryexperiments demonstrated that nicotine bitartrate didnot produce a substantial c-fos response in the fetalSCN when administered at 2.9 mg/kg (equivalent to1 mg/kg nicotine). In the free-base form, however, nic-otine did cause a modest increase in c-fos expression inthe fetal SCN, consistent with previous reports (Clegget al., 1995; O’Hara et al., 1998; O’Hara et al., 1999).Nicotine also induced c-fos expression in the supra-optic and paraventricular nuclei of the hypothalamus(Fig. 4 and data not shown; see also O’Hara et al., 1999),but not in the striatum, giving an anatomical patternclearly distinct from that seen following maternaltreatment with caffeine or SKF 38393. Pretreatmentwith each of the antagonists, SCH 23390, Meca, or MK801, significantly reduced the effect of nicotine on SCNc-fos mRNA levels (Fig. 5 ; Dunnett’s test). SCH 23390and mecamylamine each blocked the effect of nico-tine, while MK 801 reduced the amplitude of responsebut did not prevent it (Fig. 5; t tests).

DISCUSSION

The results of these experiments indicate the exis-tence of multiple and distinct pharmacological mech-anisms for regulating c-fos gene expression in the fetalSCN. In the adult SCN, major regulators of c-fosinclude glutamate and serotonergic agonists (Ebling,1996; Kennaway and Moyer, 1999), while in the fetalSCN, major regulators of c-fos expression includedopamine receptor activation and an unidentifiedpathway activated by caffeine. Remarkably, none ofthe treatments used to induce c-fos expression in thefetal SCN in this study causes a detectable increase in

534 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 24: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

c-fos expression in the adult SCN (Clegg et al., 1995;Nakajima et al., 1989, unpublished data; Weaver andReppert, 1995; Weaver et al., 1992). The mechanismsregulating c-fos expression in the brain appear to varydepending on the pharmacological stimulus, theregion, and the developmental stage examined, andcoincident stimulation of multiple neurotransmittersystems may play a more important role in someregions (e.g., striatum) than in others (e.g., SCN, seebelow).

The dopaminergic agonist, SKF 38393, acts throughD1 receptors to induce c-fos gene expression in the fetal

SCN. The D1-receptor antagonist, SCH 23390, blockedthe induction of c-fos by SKF 38393. The present resultsare consistent with previous results in mice with tar-geted disruption of the D1-dopamine receptor geneand previous pharmacological data, indicating thatdopaminergic agents induce c-fos expression in theSCN via activation of D1-dopamine receptors (Benderet al., 1997; Weaver et al., 1992). In contrast, inductionof c-fos expression in the fetal SCN by caffeine was notdisrupted by SCH 23390, indicating the existence of anadditional, D1-receptor-independent pathway lead-

Shearman and Weaver / DISTINCT PHARMACOLOGICAL MECHANISMS 535

Figure 2. SKF 38393 and caffeine induce c-fos expression by distinct pharmacological mechanisms. Representative autoradio- grams illus-trate c-fos gene expression in fetal brain after various drug treatment combinations. Pregnant rats were pretreated with vehicle (A, B, C),SCH 23390 (D, E, F), Meca (G, H, I), or MK 801 (J, K, L). Thirty minutes later, they received vehicle (A, D, G, J), SKF 38393 (B, E, H, K), or caf-feine (C, F, I, L).

Page 25: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

ing to c-fos gene expression (see also Bender et al.,1997).

The mechanism by which caffeine induced c-fosexpression in the fetal SCN is unclear. Caffeine is anadenosine receptor antagonist (Fredholm, 1995; Nehliget al., 1992), but A1 and A2A adenosine receptor mRNAsare not expressed at detectable levels in the fetal SCN(Weaver, 1993, 1996). Furthermore, the induction ofc-fos expression in the adult striatum occurs only atpharmacological levels of caffeine and does notappear to be due to antagonism of specific adenosinereceptor subtypes, suggesting a more complicatedmechanism (Fredholm, 1995; Johanssen et al., 1992;Nakajima et al., 1989). Notably, the mechanism bywhich caffeine stimulates c-fos gene expression in thefetal striatum is pharmacologically distinct from themechanism by which caffeine leads to c-fos in the fetalSCN. Pretreatment with the D1-dopamine receptorantagonist, SCH 23390, prevented induction of c-fosexpression in the striatum but not in the SCN.D1-dopamine receptors are present in both the SCNand striatum, and dopaminergic innervation to bothareas has been demonstrated during late fetal life(Bender et al., 1997; Duffield et al., 1999; Shearmanet al., 1997; Strother et al., 1998a). The presence of D2-dopamine receptors in the fetal striatum and theirabsence in the SCN (Schambra et al., 1994; Weaveret al., 1992) raises the possibility that synergistic inter-actions between the D1 and D2 receptor subtypes occurin the striatum and that these interactions may con-tribute to regulation of c-fos expression in fetalstriatum. In the fetal SCN, the effects of caffeine onc-fos expression appear to be independent of D1-dopa-mine receptor stimulation.

In the adult brain, glutamate receptor activationplays a major role in the regulation of c-fos expression(Chaudhuri, 1997). Glutamate regulates c-fos expres-sion in the SCN and is a principal mediator conveyingphotic information from the retina to the SCN forentrainment (Ebling, 1996; Mintz et al., 1999). In theadult striatum, induction of c-fos expression byD1-dopamine-receptor activation appears to requirecoactivation of NMDA receptors (Nakazato et al.,1998). Similarly, in the present study, the NMDAreceptor antagonist MK801 prevented SKF 38393–induced c-fos expression in the lateral striatum. In con-trast, the induction of c-fos expression in the fetal SCNby SKF 38393 treatment was remarkably unaffected bypretreatment with MK 801. Thus, synergism betweenD1-dopamine and NMDA receptors is apparentlyrequired in the adult and fetal striatum, but not in thefetal SCN.

536 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 3. Quantitative assessment of antagonist effects ondrug-induced c-fos gene expression. A. SKF 38393. B. Caffeine.Values represent the mean ± SEM of 7 to 11 brains per group. Sam-ple sizes are shown within each bar. Values are expressed as rela-tive optical density (OD) (OD SCN/OD adjacent hypothalamus).Pretreatments were given 30 min before the treatment. Brainswere collected an additional 30 min later, frozen, and processed todetect c-fos RNA. Statistical comparisons key: *p < 0.05; NS = notstatistically significant (p > 0.05) (Student’s t tests). # = signifi-cantly different from control group that received vehicle pretreat-ment followed by SKF 38393 (panel A) or caffeine (panel B) byDunnett’s test (p < 0.05).

Page 26: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Dopaminergic, cholinergic, and glutaminergicmechanisms each appear to contribute to the fetalresponse to nicotine. There may be serial activation ofseveral receptor types in generating the response tonicotine, making it susceptible to disruption at each ofseveral points, or responsive cells may require coinci-dent stimulation by several receptor subtypes. Nico-tine may influence c-fos expression by acting as anindirect agonist, stimulating release of monoaminesand/ or glutamate through a presynaptic mechanism.

Indeed, nicotinic induction of Fos in the adult brain isdisrupted by NMDA receptor antagonists and D1

dopamine receptor antagonists (Kiba and Jayaraman,1994). Data from adult SCN slices indicate that nico-tine affects the SCN by action directly within thenucleus, however (Trachsel et al., 1995). Furthermore,nicotinic receptor subunit gene expression has beendocumented in the fetal SCN (O’Hara et al., 1999), pro-viding a substrate for direct action of nicotine in thefetal SCN. These lines of evidence indicate that while

Shearman and Weaver / DISTINCT PHARMACOLOGICAL MECHANISMS 537

Figure 4. Induction of c-fos gene expression by nicotine. Representative autoradiograms illustrate c-fos gene expression in fetal brain aftervarious drug treatment combinations. Pregnant rats were pretreated with vehicle (A, B), SCH 23390 (C, D), Meca (E, F), or MK 801 (G, H).Thirty minutes later, they received vehicle (A, C, E, G), or nicotine (B, D, F, H; 1 mg/kg).

Page 27: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

the mechanism of action of nicotine is pharmacologi-cally complex, this complexity may occur within theanatomical confines of the SCN.

Our studies were conducted during the subjectiveday on GD 20. Previous studies indicate that the fetalSCN expresses detectable but nonrhythmic (“basal”)levels of c-fos mRNA(Viswanathan et al., 1994; Weaveret al., 1992). Nevertheless, there is abundant evidencethat the circadian oscillator in the SCN is functioningat this age (Reppert and Weaver, 1991), raising the pos-sibility that the responses observed were influencedby the time of day of study. In the fetal SCN, c-fos geneexpression is induced equally well following adminis-tration of dopaminergic drugs during the subjectiveday and subjective night (Viswanathan et al., 1994;Weaver et al., 1992). This is in contrast to the pro-nounced “gating” of photic induction of cfos, junB, andmPer gene expression in the adult, in which responsesare restricted to subjective night. While it is possiblethat the basal neurochemical tone to the fetal SCN var-ies over the course of the day, it seems unlikely that thepharmacological mechanisms leading to drug-

induced cfos expression would vary as a function oftime of day. Thus, we would not expect differentresults had the studies been conducted at night. Theresults might have been different, however, had weconducted them at a different developmental age.Rhythmicity of basal cfos levels appears in thedorsomedial SCN shortly after birth. Detection ofdrug-induced cfos during the daytime would havebeen more difficult during the postnatal period. Fur-thermore, the loss of the cfos responses to drugs duringthe early postnatal period limits the developmentalwindow during which these studies could be con-ducted (Weaver and Reppert, 1995).

Pharmacological activation of dopamine receptorsin the fetal SCN can influence the development andfunction of the circadian timing system. Asingle injec-tion of SKF 38393 late in gestation is sufficient toentrain fetal Syrian hamsters (Viswanathan andDavis, 1997). Sensitivity to SKF 38393 begins duringthe prenatal period and continues into the neonatalperiod (Grosse and Davis, 1999; Viswanathan et al.,1994). Entrainment to SKF 38393 may result frominduction of Per1 gene expression; Per1 gene expres-sion is increased in the fetal rat SCN by maternal treat-ment with SKF 38393 (Shearman and Weaver, unpub-lished data). Prenatal stimulation of dopaminereceptors also alters the development of SCN respon-siveness to light at night in both rats and hamsters(Ferguson & Kennaway, 2000; Ferguson et al., 2000;Strother et al., 1998b).

Prenatal exposure to caffeine or nicotine may alsoalter the circadian timing system. There is limited evi-dence to suggest that theophylline, a methylxanthinewith pharmacological effects similar to caffeine, caninfluence the circadian timing system in adult rodents(Ehret et al., 1975). While much of the current researchon caffeine relates to its ability to promote arousal andreduce fatigue (Wright et al., 1997), the possibility thatcaffeine may have effects mediated by the SCN shouldbe considered. The absence of detectable c-fos expres-sion in the adult SCN after high-dose caffeine treat-ment (Nakajima et al., 1989; our unpublished data)does not rule out an effect of caffeine on the circadiantiming system, as c-fos gene expression is not necessar-ily induced by phase-shifting stimuli (Colwell et al.,1993; Kumar et al., 1997; Rea et al., 1993; Viswanathanand Davis, 1997).

The present results support the concept that multi-ple, functional neurotransmitter systems are active inthe fetal brain, and suggest that exposure to widelyused stimulant drugs during pregnancy may influ-

538 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 5. Quantitative assessment of antagonist effects on nico-tine-induced c-fos gene expression. Values represent the mean± SEM of 5 to 10 brains per group, expressed as relative opticaldensity (OD) (OD SCN/OD adjacent hypothalamus). Samplesizes are shown within each bar. Pretreatment drugs, doses, andtiming as described in the legends to Figures 2 and 4. Pairwisecomparisons using Student’s t test are indicated by brackets (*p0.05; NS indicates not significant, p 0.05). # = significantly differ-ent from control group that received vehicle pretreatment fol-lowed by nicotine (Dunnett’s test, p < 0.05).

Page 28: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

ence brain development and function. Indeed, inges-tion of caffeine and nicotine during development havebeen shown to alter neurochemical development andcan have long-lasting neurobehavioral effects (Adenet al., 2000; Etzel and Guillet, 1994; Fisher and Guillet,1997; Grimm and Frieder, 1988; Nehlig and Debry,1994; Ribary and Lichtensteiger, 1989; Slotkin, 1998;Sobotka, 1989). Among the effects of prenatal expo-sure to common neuromodulatory drugs may bealtered development of the circadian timing system.

ACKNOWLEDGMENTS

This work was supported by NIH grants HD14427and HD29470. LPS was supported in part by a fellow-ship from the Program for Training in Sleep, Circadianand Respiratory Neurobiology at Harvard MedicalSchool (HL07901).

REFERENCES

Aden U, Herlenius E, Tang L-Q, and Fredholm BB (2000)Maternal caffeine has minor effects on adenosine recep-tor ontogeny in the rat brain. Pediatr Res 48:177-183.

Bender M, Drago J, and Rivkees SA (1997) D1 receptorsmediate dopamine action in the fetal suprachiasmaticnuclei: Studies of mice with targeted deletion of the D1dopamine receptor gene. Mol Brain Res 49:271-277.

Carlson LL, Weaver DR, and Reppert SM (1991) Melatoninreceptors and signal transduction during developmentin Siberian hamsters (Phodopus sungorus). Dev Brain Res59:83-88.

Chaudhuri A(1997) Neural activity mapping with inducibletranscription factors. Neuroreport 8(13):iii-vii.

Clegg DA, O’Hara BF, Heller HC, and Kilduff TS (1995) Nic-otine administration differentially affects gene expres-sion in the maternal and fetal circadian clock. Dev BrainRes 85:46-54.

Colwell CS, Kaufman CM, and Menaker M (1993)Phase-shifting mechanisms in the mammalian circadiansystem: New light on the carbachol paradox. J Neurosci13:1454-1459.

Davis FC (1997) Melatonin: Role in development. J BiolRhythms 12:498-508.

Davis FC and Mannion J (1988) Entrainment of hamster pupcircadian rhythms by prenatal melatonin injections to themother. Am J Physiol 255:R439-R448.

Duffield GE, McNulty S, and Ebling FJ (1999) Anatomicaland functional characterisation of a dopaminergic sys-tem in the suprachiasmatic nucleus of the neonatal Sibe-rian hamster. J Comp Neurol 408:73-96.

Ebling FJ (1996) The role of glutamate in the photic regula-tion of the suprachiasmatic nucleus. Prog Neurobiol50:109-132.

Ehret CF, Potter VR, and Dobra KW (1975) Chronotypicaction of theophylline and of pentobarbital as circadianzeitgebers in the rat. Science 188:1212-1215.

Etzel BA and Guillet R (1994) Effects of neonatal exposure tocaffeine on adenosine A1 receptor ontogeny usingautoradiography. Dev Brain Res 82:223-230.

Ferguson, SA and Kennaway DJ (2000) Prenatal exposureto SKF-38393 alters the response to light of adult rats.NeuroReport11: 1539-1541.

Ferguson SA, Kennaway DJ, and Moyer RA (1999) Nicotinephase shifts the 6-sulfatoxymelatonin rhythm andinduces c-Fos in the SCN of rats. Brain Res Bull 48:527-538.

Ferguson SA, Rowe SA, Krupa M, and Kennaway DJ (2000)Prenatal exposure to the dopamine agonist SKF 38393disrupts the timing of the initial response of thesuprachiasmatic nucleus to light. Brain Res 858:284-289.

Ferre S, Fuxe K, von Euler G, Johansson B, and Fredholm BB(1992) Adenosine-dopamine interactions in the brain.Neuroscience 51:501-512.

Fisher S and Guillet R (1997) Neonatal caffeine alters passiveavoidance retention in rats in an age- and gender-relatedmanner. Dev Brain Res 98:145-149.

Fredholm BB (1995) Adenosine, adenosine receptors and theactions of caffeine. Pharmacol Toxicol 76:93-101.

Garrett BE and Holtzman SG (1994) D1 and D2 dopaminereceptor antagonists block caffeine-induced stimulationof locomotor activity in rats. Pharmacol Biochem Behav47:89-94.

Grimm VE and Frieder B (1988) Prenatal caffeine causeslong lasting behavioral and neurochemical changes. Int JNeurosci 41:15-28.

Grosse J and Davis FC (1999) Transient entrainment of a cir-cadian pacemaker during development by dopaminergicactivation in Syrian hamsters. Brain Res Bull 48:185-194.

Johansson B, Svenningson P, Aden U, Lindstrom K, andFredholm BB (1992) Evidence that the increase in striatalc-fos following acute high-dose caffeine is not due todirect interaction with striatal adenosine receptors. ActaPhysiol Scand 146:539-541.

Keefe DL, Earnest DJ, Nelson D, Takahashi JS, and Turek FW(1987) A cholinergic antagonist, mecamylamine, blocksthe phase-shifting effects of light on the circadian rhythmof locomotor activity in the golden hamster. Brain Res403:308-312.

Kennaway DJ and Moyer RW (1999) MK-801 administrationblocks the effects of a 5HT2A/2C agonist on melatoninrhythmicity and c-fos induction in the suprachiasmaticnucleus. Brain Res 845:102-106.

Kiba H and Jayaraman A (1994) Nicotine induced c-fosexpression in the striatum is mediated mostly by dopa-mine D1 receptor and is dependent on NMDA stimula-tion. Mol Brain Res 23:1-13.

Klein DC, Moore RY, and Reppert SM (Eds) (1991)Suprachiasmatic Nucleus: The Mind’s Clock, Oxford Uni-versity Press, New York.

Kumar V, Goguen DM, Guido ME, and Rusak B (1997)Melatonin does not influence the expression of c-fos in thesuprachiasmatic nucleus of rats and hamsters. Mol BrainRes 52:242-248.

Shearman and Weaver / DISTINCT PHARMACOLOGICAL MECHANISMS 539

Page 29: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Mintz EM, Marvel CL, Gillespie CF, Price KM, and AlbersHE (1999) Activation of NMDA receptors in the supra-chiasmatic nucleus produces light-like phase shifts of thecircadian clock in vivo. J Neurosci 19:5124-5130.

Nakajima T, Daval JL, Morgan PF, Post RM, and MarangosPJ (1989) Adenosinergic modulation of caffeine-inducedc-fos mRNA expression in mouse brain. Brain Res501:307-314.

Nakazato E, Ohno M, and Watanabe S (1998) MK-801reverses Fos expression induced by the full dopamine D1receptor agonist SKF-82958 in the rat striatum. Eur JPharmacol 342:209-212.

Nehlig A, Daval J-L, and Gérard D (1992) Caffeine and thecentral nervous system: Mechanisms of action, biochemi-cal, metabolic and psychostimulant effects. Brain Res Rev17:137-170.

Nehlig Aand Debry G (1994) Consequences on the newbornof chronic maternal consumption of coffee during gesta-tion and lactation: A review. J Am Coll Nutr 13:6-21.

O’Hara BF, Edgar DM, Cao VH, Wiler SW, Heller HC,Kilduff TS, and Miller JD (1998) Nicotine and nicotinicreceptors in the circadian system. Psychoneuroendocri-nology 23:161-173.

O’Hara BF, Macdonald E, Clegg D, Wiler SW, Andretic R,Cao VH, Miller JD, Heller HC, and Kilduff TS (1999)Developmental changes in nicotinic receptor mRNAsand responses to nicotine in the suprachiasmatic nucleusand other brain regions. Mol Brain Res 66:71-82.

Rea MA, Michel AM, and Lutton LM (1993) Is Fos expres-sion necessary and sufficient to mediate light-inducedphase advances of the suprachiasmatic circadian oscilla-tor? J Biol Rhythms 8:S59-S64.

Reppert SM and Weaver DR (1991) A biological clock isoscillating in the fetal suprachiasmatic nuclei. InSuprachiasmatic Nucleus: The Mind’s Clock, DC Klein, RYMoore, and SM Reppert, eds, pp 405-418, Oxford Univer-sity Press, New York.

Ribary U and Lichtensteiger W (1989) Effects of acute andchronic prenatal nicotine treatment on central catechola-mine systems of male and female rat fetuses and off-spring. J Pharmacol Exp Ther 248:786-792.

Schambra UB, Duncan GE, Breese GR, Fornaretto MG,Caron MG, and Fremeau RT Jr (1994) Ontogeny of D1Aand D2 dopamine receptor subtypes in rat brain using insitu hybridization and receptor binding. Neuroscience62:65-85.

Shearman LP and Weaver DR (1997) [125I]4-aminobenzyl-5′-N-methylcarboxamido-adenosine (125l)AB-MECA)labels multiple adenosine receptor subtypes in rat brain.Brain Res 745:10-20.

Shearman LP, Zeitzer J, and Weaver DR (1997) Widespreadexpression of functional D1-dopamine receptors in fetalrat brain. Dev Brain Res 102:105-115.

Slotkin TA (1998) Fetal nicotine or cocaine exposure: Whichone is worse? J Pharmacol Exp Ther 285:931-945.

Sobotka TJ (1989) Neurobehavioral effects of prenatal caf-feine. Ann N Y Acad Sci 562:327-339.

Strother WN, Norman AB, and Lehman MN (1998a)D1-dopamine receptor binding and tyrosine-hydroxy-lase immunoreactivity in the fetal and neonatal hamstersuprachiasmatic nuclei. Dev Brain Res 106:137-144.

Strother WN, Vorhees CV, and Lehman MN (1998b)Long-term effects of early cocaine exposure on the lightresponsiveness of the adult circadian timing system.Neurotoxicol Teratol 20:555-564.

Svenningsson P, Johansson B, and Fredholm BB (1996) Caf-feine-induced expression of c-fos mRNA and NGFI-AmRNA in caudate putamen and in nucleus accumbensare differentially affected by the N-methyl-D-aspartatereceptor antagonist MK-801. Mol Brain Res 35:183-189.

Trachsel L, Heller HC, and Miller JD (1995) Nicotinephase-advances the circadian neuronal activity rhythmin rat suprachiasmatic nuclei explants. Neuroscience65:797-803.

Viswanathan N and Davis FC (1997) Single prenatal injec-tions of melatonin or the D1-dopamine receptor agonistSKF 38393 to pregnant hamsters sets the offsprings’ circa-dian phases 180° apart. J Comp Physiol [A] 180:339-346.

Viswanathan N, Weaver DR, Reppert SM, and Davis FC(1994) Entrainment of the fetal hamster circadian pace-maker by prenatal injections of the dopamine agonistSKF 38393. J Neurosci 14:5393-5398.

Weaver DR (1993) A2a adenosine receptor gene expression indeveloping rat brain. Mol Brain Res 20:313-327.

Weaver DR (1996) A1-adenosine receptor gene expression infetal rat brain. Dev Brain Res 94:205-223.

Weaver DR (1998) Suprachiasmatic nucleus: A25-year retro-spective. J Biol Rhythms 13:92-104.

Weaver DR and Clemens LG (1987) Nicotinic cholinergicinfluences on sexual receptivity in female rats. PharmacolBiochem Behav 26:393-400.

Weaver DR and Reppert SM (1995) Definition of the devel-opmental transition from dopaminergic to photic regula-tion of c-fos gene expression in the rat suprachiasmaticnucleus. Mol Brain Res 33:136-148.

Weaver DR, Rivkees SA, and Reppert SM (1992) D1-dopaminereceptors activate c-fos expression in the fetal biologicalclock. Proc Natl Acad Sci U S A 89:9201-9204.

Weaver DR, Roca AL, and Reppert SM (1995) c-fos and jun-BmRNAs are transiently expressed in fetal rodentsuprachiasmatic nuclei following dopaminergic stimula-tion. Dev Brain Res 85:293-297.

Wright KP Jr, Badia P, Myers BL, and Plenzler SC (1997)Combination of bright light and caffeine as a counter-measure for impaired alertness and performance duringextended sleep deprivation. J Sleep Res 6:26-35.

Zhang Y, Zee PC, Kirby JD, Takahashi JS, and Turek FW(1993) A cholinergic antagonist, mecamylamine, blockslight-induced Fos immunoreactivity in specific regionsof the hamster suprachiasmatic nucleus. Brain Res615:107-112.

540 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 30: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

JOURNAL OF BIOLOGICAL RHYTHMS / December 2001Gorman and Lee / SPLIT ACTIVITY RHYTHMS IN HAMSTERS

Daily Novel Wheel Running Reorganizes andSplits Hamster Circadian Activity Rhythms

Michael R. Gorman1 and Theresa M. LeeDepartment of Psychology and Reproductive Sciences Program,

University of Michigan, Ann Arbor, MI 48109-1109, USA

Abstract The phenomenon of splitting of locomotor activity rhythms in con-stant light has implied that the mammalian circadian pacemaker is composed ofmultiple interacting circadian oscillators. Exposure of male Syrian hamsters tonovel running wheels also induces splitting in some reports, although novelwheel running (NWR) is better known for its effects on altering circadian phaseand the length of the free-running period. In three experiments, the authors con-firm and extend earlier reports of split rhythms induced by NWR. Male Syrianhamsters, entrained to LD 14:10, were transferred for 6 to 11 consecutive days todarkened novel Wahmann wheels at ZT 4 and were returned to their home cagesat ZT 9. All hamsters ran robustly in the novel wheels. NWR caused a markedreorganization of home cage wheel-running behavior: Activity onsets delayedprogressively with each additional day of NWR. After 11 days, activity onset inthe nighttime scotophase was delayed by 7 h and disappeared completely in2 hamsters (Experiment 1). After 6 to 7 days of NWR (Experiment 2), activityonset delayed by 5 h. Transfer of hamsters to constant darkness (DD) after 7 daysof NWR revealed clearly split activity rhythms: The delayed nighttime activitybout was clearly identifiable and characterized by a short duration. A secondbout associated with the former time of NWR was equally distinct and exhibiteda similarly short duration. These components rejoined after 3 to 5 days in DDaccomplished via delays and advances of the nighttime and afternoon compo-nents, respectively. The final experiment established that rejoining of activitycomponents could be prevented by perpetuating the light-dark:light-dark cycleused to induce split rhythms. The data suggest that NWR causes selective phaseshifting of some circadian oscillators and that component oscillators interactstrongly in constant darkness.

Key words splitting, oscillator interaction, coupling, nonphotic

A multioscillator basis for mammalian circadianrhythms has been adduced through studies ofphotoperiodic control of activity duration (α), internaldesynchronization, splitting, and most recently, invitro electrical recordings of single SCN cells (Aschoff,1965; Elliott and Tamarkin, 1994; Gorman et al., 1997;

Illnerova, 1991; Liu et al., 1997; Pittendrigh and Daan,1976). Each set of studies reinforces the idea thatcoherent circadian rhythms are generated from theinteraction of coupled constituent oscillators with arange of free-running periods, τ. Although significantadvances have been made in clarifying the neuroana-

541

1. To whom all correspondence should be addressed: Department of Psychology, University of California, San Diego, La Jolla,CA 92093-0109; e-mail: [email protected].

JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 541-551© 2001 Sage Publications

Page 31: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

tomical and physiological substrates for rhythm gen-eration and entrainment, the formal properties ofoscillator interaction have received less sustainedattention.

Amajor exception to this generalization is the studyof split locomotor activity rhythms first reported inthe arctic ground squirrel, Spermophilus undulatus(Pittendrigh, 1960) and elaborated further in studiesof Syrian hamsters, Mesocricetus auratus (Pittendrighand Daan, 1976). After prolonged (e.g., 60 days) expo-sure to constant light (LL), locomotor activity rhythmsof some individuals dissociate into two componentsthat free-run initially with different frequencies. Whenthe two split activity components adopt an antiphaserelationship (180 degrees apart), they free-run with acommon frequency greater than that measured justprior to splitting. A comparable phenomenon isobtained in a day-active species exposed to low levelsof light intensity (Hoffmann, 1971).

Exposure to constant lighting conditions is not theonly manipulation capable of splitting mammaliancircadian rhythms. Although not discussed in the text,Bruce’s (1960) study of frequency demultiplicationincludes a single actogram of a hamster maintained inshort cycles of 2 h light, 4.5 h dark (LD 2:4.5). In thisrecord, two activity components 180 degrees apartwere apparent for approximately 7 days before one ofthese components disappeared. Mrosovsky and Janik(1993) reorganized the activity rhythms of hamstersmaintained in LD 14:10 by exposing them each after-noon to 3-h pulses of novel wheel running (NWR) inthe dark (beginning 7 h before normal lights-off).When NWR was discontinued and hamsters were leftin their home cages in constant darkness (DD), loco-motor activity rhythms were split into two compo-nents that rejoined after 3 to 5 days, although this pat-tern was not equally clear in all records shown (e.g.,#3802 in their Fig. 2). Nighttime activity onsets in LD14:10 were also phase-delayed by several hours dur-ing NWR. Sinclair and Mistlberger (1997), using a dif-ferent strain of hamster and a slightly modified proto-col, found less compelling evidence of splitting after17 days of NWR, although nighttime activity onsetwas delayed in some animals. Using the hamsterstrain employed by Mrosovsky and Janik (1993) anda modification of their experimental protocol, wehere describe marked reorganizations of locomotoractivity rhythms induced by three regimens of dailyNWR.

MATERIALS AND METHODS

Animals and Husbandry

For all experiments, a subset of the same 24 maleSyrian hamsters that were used in a separatelyreported study published in this issue (HsdHan:AURA; Harlan, Indianapolis, IN, USA) (Gorman et al.,2001 [this issue]), 5 to 6 weeks of age at acquisition,were housed with Sani-Chip bedding in polypropy-lene cages (48 × 27 × 20 cm) equipped with Nalgene(d = 34 cm) running wheels (Fisher Scientific, Pitts-burgh, PA, USA). Food (Purina Rodent Chow #5001,St. Louis, MO, USA) and water were available ad libi-tum. Syrian hamsters were entrained to LD 14:10(lights on 0500-1900 h; approximately 100 lux) for 3weeks before an initial regimen of daily NWR wasinitiated.

Novel Wheel Running

Following entrainment to LD 14:10 (lights off = ZT12), hamsters were transferred within the same roomto Wahmann wheels (d = 34 cm) 0 to 15 min beforelights were extinguished at 1100 h (ZT 4) EST. At 1600h (ZT 9), the lights were turned on and hamsters werereturned to home cages in the light over the next 15 min.Thus, during NWR, animals were exposed to anLDLD 6:3:5:10 light schedule. On one day, the dark-ened hamster room was entered through a light lock athourly intervals from 1200 to 1600 h to record thenumber of novel wheel revolutions with the aid of adim red light.

Analysis

Wheel-running activity in the home cage was mon-itored by Dataquest III software (Mini-mitter, SunRiver, OR, USA) and compiled into 10-min bins. Whilein the novel wheels, activity patterns were not moni-tored, but the total number of wheel revolutions afterthe 5-h interval was recorded manually. Data analyseswere carried out with Excel (Microsoft, Bellevue, WA,USA) and ClockLab software (Actimetrics, Evanston,IL, USA).

Activity onset was defined as the first time point ina scotophase in which a hamster ran more than 20 rev-olutions in a 10-min interval followed immediately by

542 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 32: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

an additional 10-min interval with more than 20 wheelrevolutions. Activity offset was defined as the lasttime point in a scotophase that the animal ran morethan 20 revolutions and that was preceded immedi-ately by a similar 10-min interval of activity. Activityduration (α) was calculated as the interval betweenactivity onset and activity offset. An interval of inac-tivity was calculated as the difference between activityoffset and the subsequent activity onset. The circadianperiod of activity onsets either in constant conditions(τ) or while exposed to a light-dark cycle (τ*) was esti-mated with linear regression by determining slope ofactivity onsets over 4- to 7-day intervals. The phaseangle of entrainment was determined from the aver-age value predicted by the regression line and wasexpressed in relation either to the time of lights-off(ψlights-off) or lights-on (ψlights-on). When activity compo-nents were split, circadian parameters were calculatedseparately for activity bouts corresponding to theoriginal 10-h dark period (i.e., the nighttime, n, activ-ity bout) and to the 5-h interval of NWR (i.e., the after-noon, a, activity bout). The phase angle between com-ponents was defined as the difference between theirrespective activity onsets (ψn–a).

Experiment 1

Hamsters, 8 to 9 weeks of age, previously entrainedto LD 14:10, were exposed to NWR in LDLD 6:3:5:10(n = 20). After 11 days of these treatments, hamstersremained undisturbed in their home cages for 2 addi-tional days under the same light conditions.

Experiment 2

Because 11 days of NWR in Experiment 1 phase-delayed nighttime activity onset more than expectedon the basis of published studies, we next assessedwhether more evenly split activity would be obtainedafter fewer days of NWR. Hamsters from Experiment 1,12 to 13 weeks of age, were re-entrained to LD 14:10 for14 days and exposed to NWR under LDLD 6:3:5:10 for7 days (n = 19). Four additional hamsters, with identi-cal light histories but no previous running-wheelexposure, were equipped with Nalgene wheels. In thisexperiment, these control hamsters were exposed toLDLD 6:5:3:10 without NWR. All hamsters were mini-mally disturbed during days of NWR except for a sin-gle retro-orbital bleeding conducted on the final dayas part of another study. After the final day of NWR, 9of the 19 hamsters, randomly selected, remained in

this study and were exposed to constant darkness(DD) initiated during the subsequent 10-h scotophase(i.e., the lights remained off at 0500 h). Data from theremaining 10 hamsters are reported here only throughthe final day of NWR, after which they received a dif-ferent light treatment described in a separate study(Gorman et al., 2001). Periods of the free-runningrhythms of activity onset were calculated for each ofthe 9 hamsters during days 1 to 4 and 8 to 11 of DD.

Experiment 3

Because a distinctly and evenly split home cagerunning rhythm was obtained in Experiment 2, weasked whether these hamsters could be entrained tothe LDLD cycle in effect during NWR. The same ham-sters (n = 20 including 1 former control hamster fromExperiment 2), 30 to 31 weeks of age, were re-entrained to LD 14:10 and treated as described inExperiment 2 except that they were not bled. After6 days of NWR in LDLD 6:5:3:10, hamsters remainedin their home cages for 11 days on the same LDLDcycle described above. Two hamsters with no priorNWR exposure (controls from Experiment 2) wereexposed to identical light conditions but were nottransferred to novel wheels.

Analyses of activity onsets were performed usingdata from the last 7 days of exposure to LD 14:10 priorto NWR and the first 7 days of continuous home cageexposure to LDLD 6:5:3:10 after NWR.

Statistical tests (all two-tailed where applicable)were performed with Statview 5.0 software (SAS Insti-tute, Cary, NC, USA).

RESULTS

Experiment 1

Hamsters transferred to novel wheels exhibitedrobust wheel-running (mean = 8186 ± 160 revolutions/5 h, range = 6817-9210, n = 20), with no significantchange in amount of wheel running over the 11 days ofthe experiment (p > 0.70, repeated measures ANOVA).When measured on Day 2 of NWR, the number ofwheel revolutions varied over time (p < 0.001), withsignificant monotonic increases (p < 0.05) over the first4 h and a decrease from the 4th to the 5th hour (p < 0.05).

All 20 hamsters showed a marked reorganization ofnighttime activity during NWR characterized by pro-gressive delays in the onset of home cage activity

Gorman and Lee / SPLIT ACTIVITY RHYTHMS IN HAMSTERS 543

Page 33: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

(Figs. 1, 2). In 2 hamsters (e.g., Fig. 1B), activity onsetsdelayed so far as to eliminate all nighttime activity onthe last 1 to 2 days of NWR. When left in the home cagein the LDLD cycle, 19 out of 20 hamsters showed spon-taneous activity in the afternoon dark period, and 17out of 20 hamsters showed activity in both the after-noon and nighttime scotophases (Fig. 1). The 1 ham-ster that did not run in the afternoon dark phase wasexceptional in having the smallest delay of nighttimeactivity onset. On each of the final 2 days when ham-sters remained in the home cage on the LDLD cycle, adisproportionate amount of running activity occurredin the afternoon scotophase (65% ± 5%, 72% ± 5%,respectively, n = 20).

Experiment 2

As in Experiment 1, NWR was observed to berobust in the entire cohort of animals tested (mean =8153 ± 220 revolutions/5 h, range = 6,127-10,044, n =19). After 7 days in novel wheels, nighttime activityonset was delayed approximately midway throughthe scotophase (Figs. 2, 3).

Upon release into DD, 8 out of 9 hamsters showedtwo distinct (i.e., split) activity components roughlycoincident with the prior time of nighttime runningand novel wheel exposure, respectively (Fig. 3). In DD,nighttime activity onsets occurred progressively laterin the first 3 days of DD (Table 1), whereas the after-noon component was neither markedly advanced nordelayed. Thus, the phase angle between components(Ψn–a) and interval of inactivity initially separating thenighttime and afternoon bouts in DD rapidly dimin-ished. In the first 3 days of DD, the two activity boutscontained comparable amounts of activity, and boutswere characterized by short αs. A redistribution ofactivity from the afternoon activity component to thenighttime component was commonly seen on the 3rdto 5th day in DD (Fig. 3). The free-running rhythmderived from onsets of the nighttime componentshortened significantly after 7 days in DD (Table 1,Fig. 3). Acomparable analysis of the afternoon compo-nent was not undertaken, because it seldom remaineddistinct for more than 3 days in DD. Abimodal activitypattern persisted beyond the fifth day of DD, after thedisappearance of a robust, clearly distinguishableafternoon component. However, this bimodalityappeared to be indistinguishable from that character-istic of unsplit hamsters with comparable α in DD. Inother words, by this time the bimodal pattern does notsuggest persistent splitting. A quantitative descrip-

tion of nighttime and afternoon activity componentsin DD is presented in Table 1.

Control animals exposed to identical LDLD condi-tions showed no reorganization of nighttime locomo-tor activity patterns or splitting in DD (data notshown).

Experiment 3

As in Experiments 1 and 2, transfer to novel wheelselicited running in the entire sample of experimentalhamsters (mean = 7631 ± 268 revolutions per 5 h;range = 4036-9197, n = 20), with no significant changesin activity over the 6 days of NWR. Running inducedprogressive delays in nighttime activity onset. After6 days of NWR, nighttime activity onset occurredapproximately 6 h after lights-out. Compared to sub-stantial further delays observed in identical light con-ditions in Experiment 1, discontinuation of NWR after6 days largely prevented further delays in nighttimeactivity onset (Figs. 2, 4).

Hamsters (18 out of 20) exposed to NWR adoptedsimilar novel entrainment patterns in the home cageunder LDLD: Locomotor activity was distributed intotwo roughly equal components corresponding to thetwo daily scotophases (i.e., splitting occurred; Fig. 4).Of the 2 nonsplitters, 1 showed the least amount ofactivity in the novel wheels (4036 rev/5 h) whereas theother showed typical activity levels (7820 rev/5 h).Neither of the 2 control hamsters exposed to this sameLDLD cycle, without NWR, adopted this entrainmentpattern (data not shown). This split pattern of locomo-tor activity was sustained for a minimum of 7 days inall 18 animals and for the duration of the experiment(11 days) for 15 of these hamsters.

Quality of entrainment was assessed by examiningwhether the slope of the best-fitting regression linethrough activity onsets differed significantly from 24 h(p < 0.05). Under baseline entrainment conditions, allbut 2 animals yielded regression lines not significantlydifferent from 24 h, indicating that they were wellentrained by the 24-h LD cycle. Slopes of these 2 ham-sters, moreover, deviated only slightly from 24 h (0.04and 0.05 h/day, respectively). Likewise, after NWRthe afternoon activity component of split hamsterswas well entrained under LDLD with only 3 out of 18split hamsters producing activity onsets with slopessignificantly different from 24 h. Activity onsets wereless well entrained for the nighttime component of thesplit rhythms. The majority of hamsters exhibited τ*ssignificantly greater than 24 h. Only 3 hamsters had

544 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 34: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

τ*s not significantly different from 24 h. Finally, qual-ity of entrainment was further assessed by quantify-ing the sum of squared residuals of actual onsets fromthe best-fit regression line. Variability of both splitcomponents was greater than for the baseline condi-

tion (Table 2; p < 0.001), but the two components didnot differ from one another (p > 0.05).

Phase angle to lights-off (ψlights-off) differed signifi-cantly from the unsplit to the split state and betweenthe two split components (Table 2). ψlights-off was signifi-cantly less negative in the baseline condition prior toNWR than in either the afternoon (p < 0.05) or night-time (p < 0.001) activity components in LDLD. In thesplit condition, ψlights-off was far more negative for thenighttime activity component than for the afternooncomponent. Relative to lights-on (ψlights-on), the phaseangle of the afternoon activity component was greaterthan that of the nighttime activity component (p < 0.05;Table 2). Phase angle of the two split components rela-tive to each other (ψn–a) varied from 8.40 to 12.24 h(mean = 10.5 ± 0.22 h). Total activity was nearlyequally distributed between nighttime (55% ± 2%)and afternoon (45% ± 2%) scotophases.

DISCUSSION

In three separate experiments, daily NWR mark-edly reorganized the locomotor activity rhythms ofmale Syrian hamsters maintained in an LD cycle.Nighttime activity onset was progressively delayedwith subsequent days of NWR: Whereas nighttime

Gorman and Lee / SPLIT ACTIVITY RHYTHMS IN HAMSTERS 545

Figure 1. Representative double-plotted actograms of hamsters initially entrained to LD 14:10 and later exposed to daily afternoon novelwheel running (NWR). Data are unclipped and scaled between zero and the hamster’s maximum activity count in the interval depicted.Dark rectangles above actograms represent the 10-h scotophase maintained throughout the experiment. Hatched rectangles above and onthe right side of actograms represent NWR treatment paired with darkness. This second daily scotophase was maintained for 2 final dayswhen hamsters remained in their home cage.

Figure 2. Mean ± SEM activity onsets (clock hour) of hamsters(n = 18-20) exposed to successive days of novel wheel running(NWR) in Experiments 1-3. Onsets for 2 days prior to NWR aredesignated “Pre-NWR.” Hamsters were exposed to 11, 7, and 6days of NWR in Experiments 1-3, respectively. In Experiment 3,activity onsets for 5 days after NWR was discontinued are plottedfor comparison with values in Experiment 1. The arrow indicateswhen the Experiment 3 hamsters remained thereafter in the homecage without further NWR.

Page 35: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

activity disappeared entirely in some hamsters after11 days of NWR, more modest delays were observedafter 6 to 7 days of NWR. This latter condition wasassociated with distinctly split activity rhythms thatrejoined after several days of DD. Perpetuation of theLDLD cycle, however, allowed the split rhythms to besustained for at least an additional 11 days in the homecage. In the absence of NWR, exposure to the LDLDcycle had no marked effect on nighttime locomotoractivity rhythms and yielded no evidence of splitting.As suggested previously (Mrosovsky and Janik, 1993),afternoon NWR phase-shifts some component circa-dian oscillators, which thereafter give rise to theexpression of a new activity bout in the afternoon dark

period. Subsequently, when the system is released intoDD, the two bouts fuse or rejoin under the influence ofstrong oscillator interactions, but alternatively may beeffectively entrained by an LDLD cycle. NWR cantherefore override typical entrainment patterns estab-lished in an LDLD cycle and reorganize activity into asecond stable configuration.

It is not clear why others have failed to replicate theinduction of splitting with afternoon NWR (Sinclairand Mistlberger, 1997) and why NWR induced largerphase-delays of home cage activity onset in this studythan in others (Mrosovsky and Janik, 1993). We usedthe same hamster supplier as the original report, incontrast to the study with largely negative effects.

546 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 3. Representative actograms of hamsters from Experiment 2. The asterisk on the actogram indicates beginning of exposure to con-stant darkness (DD). Slanted lines on right side of the actogram represent least-squares regression lines for nighttime activity onsets onDays 1-4 and 8-11 of exposure to DD. Other conventions as in Figure 1. NWR = novel wheel running.

Table 1. Circadian rhythm patterns of split hamsters (n = 8) transferred to DD in Experiment 2. Noted are significant differences over time(repeated measures ANOVA, two-tailed).

Day 1 Day 2 Day 3

Phase angle between components (Ψn–a)(h) 11.6 ± 0.3 10.5 ± 0.3 9.4 ± 0.5 p < 0.005Inactive interval (h) 8.9 ± 0.4 7.8 ± 0.5 5.9 ± 0.5 p < 0.0001

Nighttime boutActivity onset (h) 1.02 ± 0.18 1.92 ± 0.20 3.21 ± 0.25 p < 0.0001Bout duration (α)(h) 2.70 ± 0.18 2.68 ± 0.33 3.45 ± 0.82 nsWheel revolutions 749 ± 61 913 ± 124 1190 ± 274 ns

Afternoon boutActivity onset (h) 12.60 ± 0.34 12.40 ± 0.36 12.56 ± 0.52 nsBout duration (α)(h) 2.68 ± 0.40 3.25 ± 0.42 2.55 ± 0.45 nsWheel revolutions 778 ± 110 1031 ± 132 851 ± 135 p < 0.01

τ( ) . . . .hdays 1 - 4 days 8 - 11

25 09 0 15 24 58 0 06± ± p < 0.01

Page 36: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Minor differences in intrinsic periods, propensity torun in novel wheels, or oscillator coupling may distin-guish splitting and nonsplitting strains. We also usedlonger exposures to NWR (5 h vs. 3 h) than used in pre-vious studies. Notably, running in novel wheels wasmost intense during the 4th hour of exposure. Regard-less of differences between studies, the significance ofthis experimental paradigm is as a probe of specificaspects of oscillator function, which likely differ quan-titatively rather than qualitatively among differenthamster strains and experimental conditions.

Various formal mechanisms may account for theprogressive delays in nighttime activity onset duringsuccessive days of NWR. Amore negative phase angleof entrainment as was obtained in each experimentcan result from a lengthening of τ. Alternatively, eachday of NWR may induce a single phase-delay in activ-ity onset without lengthening τ. In prior studiesemploying a single bout of NWR in early subjectiveafternoon of hamsters in DD, however, activity onsetwas advanced rather than delayed, and τ was length-

ened (Mrosovsky, 1993; Weisgerber et al., 1997). More-over, at the conclusion of NWR in the present study,the nighttime activity component free-ran in DD witha long τ. Together, these results suggest an enduringeffect on τ as opposed to a transient (e.g., phase shift)effect of NWR on the circadian pacemaker.

Two factors may dictate the pattern of rejoiningobserved in DD, which in all cases was achieved viareduction of the inactive interval following nighttimeactivity and preceding afternoon activity. First, inde-pendent of any oscillator interactions, the two compo-nents may have different intrinsic free-running peri-ods, which would favor rejoining. The large negativephase angle of the nighttime component and the rela-tively small negative value for the afternoon compo-nent suggest free-running periods, which are >24 and<24 h for the nighttime and afternoon components,respectively. Alternatively, coupling interactionsbetween oscillator components may favor theobserved pattern of rejoining regardless of the periodsof the free-running rhythms. That is, the split statemay be intrinsically unstable, and oscillators mayinteract in DD to establish a limited range of phaseangles with respect to each other. Oscillator interac-tions have been invoked to understand the limiteddecompression of α, which may be obtained in DD inunsplit hamsters. The pattern exhibited in this experi-ment is consistent with oscillators recoupling by theshortest possible route (i.e., reducing the shorter of thetwo respective phase angles, Ψn–a versus Ψa–n betweenthem), although this proposition cannot be evaluated

Gorman and Lee / SPLIT ACTIVITY RHYTHMS IN HAMSTERS 547

Figure 4. Representative actograms of hamsters from Experiment 3. Conventions as in Figures 1 and 3. NWR = novel wheel running.

Table 2. Entrainment parameters of hamsters expressed in base-line LD 14:10 and following splitting under the experimental LDLD6:5:3:10 cycle of Experiment 3.

Baseline Nighttime Afternoon N

τ* (h) 24.00 ± 0.01 24.16 ± 0.04 23.99 ± 0.02 18SS residuals 1.49 × 10–4 7.90 × 10–4 7.65 × 10–4 18Ψlights-off (h) –0.83 ± 0.02 –6.79 ± 0.22 –1.26 ± 0.05 18Ψlights-on (h) 9.17 ± 0.02 3.21 ± 0.22 3.74 ± 0.05 18

Page 37: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

against alternatives with the present data set. A role ofoscillator interactions is further suggested by themarked change in τ measured from the nighttimecomponent after several days in DD, when splitting ispresumably ended.

The results of Experiment 3 complement those ofBoulos and Morin (1985) who, with daily dark pulses,entrained activity rhythms split by LL. In that study,one component roughly coincided with a daily 2-hdark phase, while the second activity component per-sisted 8 to 12 h out of phase with the dark-entrainedcomponent. In the current study, it appears that thesplit nighttime activity component, which free-runs inDD with τ > 24 h, may be entrained solely by thephase-advancing effects of light onset at ZT 22. In con-trast, the split afternoon activity component may beentrained by either phase-delaying effects of lightprior to lights-off at ZT 4, by phase-advancing effectsof lights-on at ZT 9, or by both. In hamsters split by LL,each component of the activity rhythm expresses aPRC to dark pulses with defined regions of delays andadvances (Boulos and Rusak, 1982).

How does NWR split circadian activity rhythms?Convergent evidence from cellular, physiological,behavioral, and mathematical paradigms (e.g.,Illnerova, 1991; Liu et al., 1997; Pittendrigh and Daan,1976; Enright, 1980) points to the following model ofthe circadian pacemaker: Overt circadian rhythmsreflect the output of multiple circadian oscillators thatconstitute a coupled dual oscillatory system, whichmay be functionally described in terms of evening andmorning oscillators (Fig. 5A). A functional eveningoscillator results from the coupling of oscillators withrelatively short τs, and as such, its overall τ is < 24 h,whereas a functional morning oscillator is derivedfrom coupling of longer period constituent oscillatorswith τ > 24 h. As one effect of NWR at ZT 4 appears tobe a marked lengthening of the period of the nighttimeactivity component (Mrosovsky, 1993; Weisgerberet al., 1997), and because NWR delayed nighttimeactivity onset (present studies), we hypothesize thatearly afternoon NWR preferentially lengthens theperiod of the oscillators underlying nighttime activity,perhaps by uncoupling some of the short-period com-ponent “evening” oscillators from the coupled oscilla-tor network that generates normal nighttime activity(Fig. 5A). In DD, the larger coupled system mighttherefore free-run under the influence of its remainingcoupled constituent oscillators, of which those withlonger intrinsic τs predominate. Under entraining LDconditions, one would expect a more negative phase

angle as a result of the oscillator’s lengthened τ.Whether NWR selectively uncouples short-periodoscillators because of a particular anatomical relation-ship (e.g., such oscillators receive neuropeptide Y pro-jections) or a temporal relationship (e.g., a particularphase angle between NWR and short-period oscilla-tors) is entirely unknown.

Additionally, NWR apparently phase-shifts con-stituent oscillators just as it phase-shifts withoutovertly splitting the pacemaker in two other para-digms: After an 8-h phase-advance of the LD cycle,hamsters running in novel wheels during the new ZT13-16 re-entrained within one to two cycles, whereasnonrunning controls required several days (Mrosovskyand Salmon, 1987). NWR of sufficient duration begin-ning at ZT 5, moreover, induced rapid phase shifts inexcess of 8 h in some hamsters transferred to DD(Gannon and Rea, 1995). The present paradigm like-wise induces phase shifts, albeit of only a fraction ofthe oscillators formerly generating the nighttimeactivity. The present data strongly suggest that succes-sive days of NWR recruit cohorts of oscillators toexpress their subjective night in the afternoonscotophase. After 6 to 7 days of NWR, activity is nearlyequally divided between nighttime and afternoonscotophases, whereas the afternoon scotophase con-tains disproportionate activity (and in some cases all)after 11 days of NWR. We suggest that a single day ofNWR produces a large phase shift of a small fraction ofcomponent oscillators. With additional days of NWR,a threshold fraction of oscillators may be phase-shifted to generate an activity component in the after-noon scotophase. The progressive delays in nighttimeactivity onset are consistent with this model.

In contrast to other paradigms used (Gannon andRea, 1995; Mrosovsky and Salmon, 1987), the presenceof an LDLD cycle with a short (5 h) second scotophasemay prevent the entire complement of oscillators frombeing phase-shifted to express subjective night in theafternoon. Moreover, in the absence of further NWR,the light pulses that bracket the afternoon scotophaseimpede the recoupling of constituent oscillators backinto the unsplit state (Experiment 3), which so readilyoccurs in DD. Notably, when the intervening lightintervals were very short as in skeleton photoperiods,daily NWR from ZT 5 to ZT 8 induced complete inver-sion of activity rhythms to what was previously sub-jective day (Sinclair and Mistlberger, 1997). Similarly,11 days of NWR apparently also shifted the entireoscillatory system in a few hamsters of Experiment 1.Thus, these two factors—a titratable shifting of oscilla-

548 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 38: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

tors and countering of oscillator tendencies to rejoin—may facilitate induction and maintenance of stable,split activity rhythms under an LDLD cycle and 1 ormore days of NWR.

Beyond the significance of NWR-induced splittingnoted previously (Mrosovsky and Janik, 1993), thedemonstration that some oscillatory components canbe dissociated from others and rephased with respectto the nighttime activity component may provideinsight into the mechanism of phase-shifting effects ofNWR, or indeed, of any photic or nonphotic zeitgeber.For example, a single bout of NWR around ZT 5 pro-duces large phase advances and lengthening of τ insubsequent DD (Mrosovsky, 1991; Reebs andMrosovsky, 1989a, 1989b; Weisgerber et al., 1997). A

single bout of NWR is not sufficient to induce measur-able splitting under LDLD conditions (MR Gorman,unpublished observations), but if it selectively phase-shifts a small fraction of circadian oscillators (Fig. 5B),overt circadian rhythms might be altered as a conse-quence of recoupling dynamics among constituentoscillators. Unfortunately, little is known about theseprocesses, except that darkness favors recoupling inLL-split hamsters (Earnest and Turek, 1982) and inter-vening light pulses appear to minimize recoupling inNWR-induced splitting (Experiment 3). Actograms ofNWR-split hamsters suggest that recoupling may beaccompanied by abrupt changes in phase, with therejoining activity bout typically expressed in an inter-

Gorman and Lee / SPLIT ACTIVITY RHYTHMS IN HAMSTERS 549

Figure 5. Formal model of multioscillator basis of novel wheel running (NWR) effects on circadian rhythms. In all panels, the inverted Ushape represents subjective night of hypothetical individual circadian oscillators. For clarity, the remaining phases of each oscillator arenot depicted. In A, the integrated rhythm reflects a distribution through the scotophase of the subjective nights of individual oscillators.Those with short and long periods express subjective night early and late in the scotophase, respectively. The oscillator denoted with adashed line will be phase-shifted by NWR. After a single day of NWR from ZT 4 to ZT 9, oscillators with short periods undergo large phaseshifts to roughly the same circadian phase as NWR. Light pulses bracketing NWR preclude oscillator recoupling for reasons that are not yetclear. Subsequent activity onset is delayed, and τ is lengthened (τ* > τ). Additional days of NWR phase-shift additional cohorts of oscilla-tors such that approximately half are phase-shifted after 6 to 7 days but all or nearly all are shifted after 11 or more days. B. Application ofthe splitting model to understand effect of a single day of NWR applied in DD. In the absence of a light pulse following NWR, thephase-shifted oscillator pulls (advances) the nighttime component that follows. Until full recoupling is achieved, the nighttime activitycomponent may continue to express a lengthened τ.

Page 39: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

mediate phase between the former two components(Mrosovsky and Janik, 1993). If no light follows a boutof NWR, strong oscillator interactions may result inrapid recoupling via reduction of the shorter phaseangle between component oscillators. This wouldadvance rather than delay the main activity onset(Fig. 5B). Consistent with this general model, lightpulses shortly after a single day of NWR, such as thosethat impeded oscillator recoupling in the presentstudy, greatly attenuated the phase-advancing effectsof NWR (Mrosovsky, 1991).

An understanding of oscillator-oscillator interac-tions has lagged behind our knowledge of other fea-tures of circadian rhythms, although physiologicaldata suggest that such interactions must be central toan understanding of the pacemaker. Only a fraction ofSCN cells, for instance, receive direct retinal or IGLprojections, and yet both of these pathways are capa-ble of phase-shifting, presumably, the entire SCN. Anyzeitgeber, therefore, first shifts a subpopulation ofSCN cells that receives direct projections from thetime-giving entrainment mechanism. These cells inturn interact with the greater complement of SCN cellsto arrive at a steady-state phase shift. Under routineconditions, this selective shifting and oscillator inter-action may happen in a cycle or even more rapidly.The use of LDLD cycles in the present paradigm, incontrast, facilitates a temporal dissociation of theseprocesses by impeding the recoupling process.

ACKNOWLEDGMENTS

We thank Jim Donner for excellent animal care andJeff Elliott for helpful comments on an earlier draft ofthis manuscript. This research was supported byNHLBI grant HL61667 to TML and NICHD-48640 toMRG.

REFERENCES

Aschoff J (1965) Circadian rhythms in man. Science148:1427-1432.

Boulos Z and Morin LP (1985) Entrainment of split circadianactivity rhythms in hamsters. J Biol Rhythms 1:1-15.

Boulos Z and Rusak B (1982) Phase-response curves and thedual-oscillator model of circadian pacemakers. In Verte-brate Circadian Systems, J Aschoff, S Daan, and G Groos,eds, pp 215-223, Springer-Verlag, Berlin.

Bruce VG (1960) Environmental entrainment of circadianrhythms. Cold Spring Harb Symp Quant Biol 25:29-48.

Earnest DJ and Turek FW (1982) Splitting of the circadianrhythm of activity in hamsters: Effects of exposure to con-stant darkness and subsequent re-exposure to constantlight. J Comp Physiol [A] 145:405-411.

Elliott JA and Tamarkin L (1994) Complex circadian regula-tion of pineal melatonin and wheel-running in Syrianhamsters. J Comp Physiol [A] 174:469-484.

Enright JT (1980) Timing of Sleep and Wakefulness, Springer-Verlag, New York.

Gannon RL and Rea MA (1995) Twelve-hour phase shifts ofhamster circadian rhythms elicited by voluntary wheelrunning. J Biol Rhythms 10:196-210.

Gorman MR, Freeman DA, and Zucker I (1997) Photo-periodism in hamsters: Abrupt versus gradual changesin day length differentially entrain morning and eveningcircadian oscillators. J Biol Rhythms 12:122-135.

Gorman MR, Yellon SM, and Lee TM (2001) Temporal reor-ganization of the suprachiasmatic nuclei in hamsterswith split circadian rhythms. J Biol Rhythms 16:552-563.

Hoffmann K (1971) Splitting of the circadian rhythm as afunction of light intensity. In Biochronometry, M Menaker,ed, pp 134-146, National Academy of Sciences, Washing-ton, DC.

Illnerova H (1991) The suprachiasmatic nucleus and rhyth-mic pineal melatonin production. In SuprachiasmaticNucleus: The Mind’s Clock, DC Klein, RY Moore, and SMReppert, eds, pp 197-216, Oxford University Press, NewYork.

Liu C, Weaver DR, Strogatz SH, and Reppert SM (1997) Cel-lular construction of a circadian clock: Period determina-tion in the suprachiasmatic nuclei. Cell 91:855-860.

Mrosovsky N (1991) Double-pulse experiments withnonphotic and photic phase-shifting stimuli. J BiolRhythms 6:167-179.

Mrosovsky N (1993) τ changes after single nonphotic events.Chronobiol Int 10:271-276.

Mrosovsky N and Janik DS (1993) Behavioral decoupling ofcircadian rhythms. J Biol Rhythms 8:57-65.

Mrosovsky N and Salmon PA (1987) A behavioural methodfor accelerating re-entrainment of rhythms to newlight-dark cycles. Nature 330:372-373.

Pittendrigh CS (1960) Circadian rhythms and the circadianorganization of living systems. Cold Spring Harb SympQuant Biol 25:159-184.

Pittendrigh CS and Daan S (1976) A functional analysis ofcircadian pacemakers in nocturnal rodents: V. Pacemakerstructure: A clock for all seasons. J Comp Physiol [A]106:333-355.

Reebs SG and Mrosovsky N (1989a) Effects of induced wheelrunning on the circadian activity rhythms of Syrian ham-sters: Entrainment and phase response curve. J BiolRhythms 4:39-48.

Reebs SG and Mrosovsky N (1989b) Large phase-shifts ofcircadian rhythms caused by induced running in are-entrainment paradigm: The role of pulse duration andlight. J Comp Physiol [A] 165:819-825.

550 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 40: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Sinclair SV and Mistlberger RE (1997) Scheduled activityreorganizes circadian phase of Syrian hamsters underfull and skeleton photoperiods. Behav Brain Res87:127-137.

Weisgerber D, Redlin U, and Mrosovsky N (1997) Lengthen-ing of circadian period in hamsters by novelty-inducedwheel running. Physiol Behav 62:759-765.

Gorman and Lee / SPLIT ACTIVITY RHYTHMS IN HAMSTERS 551

Page 41: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

JOURNAL OF BIOLOGICAL RHYTHMS / December 2001Gorman et al. / SPLIT CIRCADIAN RHYTHMS IN HAMSTERS

Temporal Reorganization of the SuprachiasmaticNuclei in Hamsters with Split Circadian Rhythms

Michael R. Gorman,*,1 Steven M. Yellon,† and Theresa M. Lee**Department of Psychology and Reproductive Sciences Program,

University of Michigan, Ann Arbor, MI 48109-1109, USA, †Center for Perinatal Biology,Loma Linda University School of Medicine, Loma Linda, CA 92350, USA

Abstract A dual oscillator basis for mammalian circadian rhythms is suggestedby the splitting of activity rhythms into two components in constant light and bythe photoperiodic control of pineal melatonin secretion and phase-resettingeffects of light. Because splitting and photoperiodism depend on incompatibleenvironmental conditions, however, these literatures have remained distinct.The refinement of a procedure for splitting hamster rhythms in a 24-h light-dark:light-dark cycle has enabled the authors to assess the ability of each of twocircadian oscillators to initiate melatonin secretion and to respond to light pulseswith behavioral phase shifting and induction of Fos-immunoreactivity in thesuprachiasmatic nuclei (SCN). Hamsters exposed to a regimen of afternoonnovel wheel running (NWR) split their circadian rhythms into two distinct com-ponents, dividing their activity between the latter half of the night and the after-noon dark period previously associated with NWR. Plasma melatonin concen-trations were elevated during both activity bouts of split hamsters but were notelevated during the afternoon period in unsplit controls. Light pulses deliveredduring either the nighttime or afternoon activity bout caused that activity com-ponent to phase-delay on subsequent days and induced robust expression ofFos-immunoreactivity in the SCN. Light pulses during intervening periods oflocomotor inactivity were ineffective. The authors propose that NWR splits thecircadian pacemaker into two distinct oscillatory components separated byapproximately 180 degrees, with each expressing a short subjective night.

Key words novel wheel running, melatonin, c-Fos, phase-shift, oscillator interaction

The suprachiasmatic nucleus (SCN) of the hypo-thalamus is the principal circadian pacemaker inmammals (Weaver, 1998). Although individual cul-tured SCN cells express circadian rhythms with vary-ing free-running periods and phases (Herzog et al.,1997; Liu et al., 1997; Welsh et al., 1995), in vivo cellularactivity from the SCN is largely synchronized(Bouskila and Dudek, 1993). Thus, the question of

oscillator-oscillator interactions would appear to be acentral issue in understanding the mammalian pace-maker. A dual-oscillator basis for mammalian circa-dian rhythms was posited following the observationthat prolonged exposure to constant light (LL)induced locomotor activity patterns of hamsters andother species to diverge into two components(Pittendrigh and Daan, 1976). The SCN of split ham-

552

1. To whom all correspondence should be addressed: Department of Psychology, University of California, San Diego, La Jolla,CA 92093-0109; e-mail: [email protected].

JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 552-563© 2001 Sage Publications

Page 42: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

sters express comparable split rhythms in multiunitelectrical activity, suggesting that splitting is intrinsicto the principal circadian pacemaker (Mason, 1991).

Studies of pineal melatonin synthesis and photo-periodic regulation of activity duration (α) likewisepoint to a dual oscillator model of circadian rhythms(Elliott and Tamarkin, 1994; Gorman et al., 1997;Illnerova, 1991; Pittendrigh and Daan, 1976). An eve-ning oscillator (E), entrained by evening light, is pur-ported to initiate nocturnal melatonin secretion andlocomotor activity, whereas a morning oscillator (M),entrained by morning light, mediates the expressionof these functions near the end of night. In longer daylengths of spring and summer, E and M are entrainedin a phase relation that generates a short subjectivenight, while short day lengths of fall and winter allowthe phase angle between E and M to increase to gener-ate a long subjective night. Parallel photoperiodicmodulation of light-induced phase-resetting of loco-motor rhythms, as well as of expression of the imme-diate early gene c-Fos in the SCN, suggests that com-ponent oscillators are localized within the SCN (Elliottand Kripke, 1998; Pittendrigh et al., 1984; Sumováet al., 1995), although identification of two distinctoscillators has been lacking. Unfortunately, the cleartemporal resolution of distinct oscillators achieved inLL-induced splitting is not seen in the photo-periodism literature, where, with rare exceptions(Jagota et al., 2000), evening and morning oscillatorfunctions may greatly overlap. Conversely, notwith-standing the theoretical milestone it represents, thesplitting paradigm has been of limited utility for prob-ing the nature of underlying oscillators largelybecause its requirement for LL complicates applica-tion of the bulk of analytical techniques fruitfullyemployed in photoperiodism research. Specifically,LL masks locomotor activity, inhibits melatonin secre-tion, and precludes assessment of acute effects of lightpulses on pacemaker function. Thus, these two theo-retically important literatures have remained largelyseparate.

Mrosovsky and Janik (1993) suggested that LL isnot the only stimulus capable of splitting rhythms inthe hamster: Repeated afternoon exposures to 3 h ofnovel wheel running (NWR) delayed the onset ofnighttime wheel running and, in some cases, led to atransient splitting of locomotor activity rhythmswhen hamsters were later transferred to constantdarkness (DD). Modifying this protocol, Gorman andLee (2001 [this issue]) recently demonstrated that

activity rhythms could be reliably split and that com-ponents of the split rhythms could be entrained to alight-dark:light-dark (LDLD) cycle.

Entrainment of NWR-induced split activity rhythmsto an LDLD cycle indirectly suggests that each of thetwo oscillators mediating the split rhythms can be sep-arately phase-shifted by light. To assess this directly,and to rule out the possibility that extra-SCN oscilla-tors are recruited to mediate one or both activity com-ponents (Honma et al., 1989; Stephan et al., 1979), wecharacterized the effects of light pulses on locomotoractivity rhythms and on induction of Fos immuno-reactivity in the SCN during each of the two split activ-ity bouts induced by NWR. Additionally, we pre-sented light pulses during the inactive periodsbetween the activity components to exclude the possi-bility that α was simply lengthened to incorporateboth activity components. Finally, models of photo-periodic control of melatonin secretion have positedthe existence of distinct evening and morning oscilla-tors timing the initiation and termination, respec-tively, of the nightly pattern of elevated pinealmelatonin secretion (Illnerova, 1991). Because lightacutely suppresses melatonin secretion, it has notbeen possible to evaluate the role of component oscil-lators split in LL. We therefore assessed whether eachof the two split components was capable of initiatingmelatonin secretion.

MATERIALS AND METHODS

Animals and Husbandry

For all experiments except one, the same 24 maleSyrian hamsters (Mesocricetus auratus, HsdHan:AURA, Harlan, Indianapolis, IN, USA), 5 to 6 weeks ofage at acquisition, were housed with Sani-Chip bed-ding in polypropylene cages (48 × 27 × 20 cm)equipped with Nalgene (d = 34 cm) running wheels(Fisher Scientific, Pittsburgh, PA, USA). Food (PurinaRodent Chow #5001, St. Louis, MO, USA) and waterwere available ad libitum. These hamsters are thesame as used in another separately reported studypublished in this issue (Gorman and Lee, 2001). Eigh-teen additional hamsters from the same supplier andhoused similarly were used only for the final study ofFos-immunoreactivity in the SCN. Hamsters werehoused for 2 to 3 weeks in a 14 h light, 10 h dark condi-tion (LD 14:10; lights on 0500-1900 h) before each regi-

Gorman et al. / SPLIT CIRCADIAN RHYTHMS IN HAMSTERS 553

Page 43: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

men of daily NWR was used to split activity rhythms.Room illumination at the level of the cage lid variedfrom 100 to 300 lux.

Split activity rhythms were induced by scheduledexposures to NWR as previously described (Gormanand Lee, 2001). Briefly, after entrainment to LD 14:10,20 hamsters were transferred daily for 6 to 10 dayswithin the same room to Wahmann wheels 0 to 15 minbefore lights were extinguished at 1100 h (ZT 4) EST.At 1600 h (ZT 9), the lights were turned on and ham-sters were returned to home cages in the light over thenext 15 min. With additional days of NWR, onset ofnighttime locomotor activity in the home cage is pro-gressively delayed. The number of days of NWRwithin each experimental run was adjusted so thatnighttime activity onset occurred approximately mid-way through the 10-h nighttime scotophase. After thefinal day of NWR, hamsters were left in their homecages and exposed to DD by leaving lights off after thesubsequent 10-h nighttime scotophase. Nonsplittingcontrol hamsters (n = 4) housed in the same roomexperienced identical lighting conditions but were notexposed to NWR and thus never split their activityrhythms.

Wheel-running activity in the home cage was mon-itored by Dataquest III software (Mini-mitter, SunRiver, OR, USA) and compiled into 10-min bins.Activity onset (CT 12) was defined as the first bin ineach activity bout with wheel revolutions of morethan 20, and that was immediately followed by a sec-ond interval exceeding this threshold. While in thenovel wheels, activity patterns were not monitored,but the total number of wheel revolutions after the 5-hinterval was recorded manually. Data analyses werecarried out with Excel (Microsoft, Seattle, WA, USA)and ClockLab software (Actimetrics, Evanston, IL,USA).

Light Pulse–InducedPhase Shifts of Activity

Changes in the phase response to light followingNWR-induced splitting were assessed in threesequential experimental iterations (begun at approxi-mately 12, 18, and 24 weeks of age, respectively). Thesame cohort of hamsters was repeatedly firstentrained to LD 14:10, then exposed to a regimen ofNWR to split activity, and then transferred to DD(Aschoff Type II methodology) where they receivedeither a 15-min light pulse or a sham pulse (~450 lux atthe level of the cage lid).

Afternoon Light Pulse (CT 13-a)

In the first run, lights remained off after the night-time scotophase following the final day of NWR(Fig. 1A). Activity rhythms were monitored remotelyfor onset of wheel running the following afternoon.One hour after onset of this afternoon component ofthe split activity rhythm, 9 randomly selected ham-sters were moved in their home cages to an adjoiningroom where they experienced the 15-min light pulse.Unpulsed split control animals (n = 10) were similarlyjostled in their home cages but remained in the sameroom in darkness. Following the light pulse, animalswere returned to the housing room and remainedundisturbed for 2 weeks.

Evening Light Pulse betweenSplit Activity Components

Hamsters were next re-entrained to LD 14:10 for2 weeks and then exposed to an additional regimen ofNWR. Following the final day of NWR, lights wereturned off 2 h prematurely (1700 EST) to minimizepotential phase-shifting effects of the entrainingphotoperiod and to unmask activity onset in unsplitcontrol hamsters. Of the split hamsters, 10 were ran-domly selected and pulsed with light for 15 min begin-ning approximately 1 h after unsplit animals begantheir normal nighttime activity (n = 4). The remainingsplit hamsters (n = 9) served as unpulsed controls: theywere similarly jostled at the designated time but notexposed to light.

Nighttime Light Pulse (CT 13-n)

Hamsters were again re-entrained to LD 14:10 for2 weeks and then induced to split their rhythms withNWR. Following the final day of NWR, hamstersremained in their home cages, and lights remained offafter the subsequent nighttime scotophase. Activitywas monitored to verify that hamsters exhibited anighttime activity bout followed by an interval ofinactivity and then an afternoon activity bout. Ham-sters were pulsed with light for 15 min (n = 9) or shampulsed (n = 8) as described above beginning 1 h afterthe next nighttime activity onset.

Analyses of Phase-Shift Data

Because split circadian rhythms are unstable in DD(i.e., components re-fuse rapidly and τ and phase are

554 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 44: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

in flux), establishment of a prepulse baseline and cal-culation of a light-induced phase shift in an individualanimal are highly problematic. Therefore, phase shiftswere assessed by comparing light-pulsed andsham-pulsed animals in a between-subjects design.For each animal, nighttime and afternoon activityonset of the day preceding the light-pulse orsham-pulse were compared with the correspondingactivity onsets on the 2 days following the stimulus(see Fig. 2). Timing of the light pulse was such that wecould not obtain a prepulse afternoon activity onset touse as a phase reference in the second experimentalrun. Hence, phase shifts of this component were notdetermined. Moreover, activity onsets of the night-time component were always unambiguous, but as

the experiment progressed, fragmentation of theactivity rhythms occasionally precluded identifica-tion of a single clear afternoon activity component. Insuch cases, this component was not analyzed. Last, thefree-running period of the nocturnal activity compo-nent was computed using least squares regressionover the 4 days following the pulse or sham pulse.Because the afternoon component typically rejoinedwith the nighttime component within this interval, acomparable analysis of the free-running period of theafternoon component was not attempted.

Melatonin in Circulation

To assess whether melatonin secretion patternswere altered by NWR-induced splitting, hamsterswere lightly anesthetized with methoxyfluranevapors (Metofane) and retro-orbitally bled in dark-ness with the aid of a dim red light (< 1 lux). Bloodsamples were collected into ethylene diaminetetraacetic acid–treated tubes. Plasma was harvestedafter centrifugation at 5000 rpm for 20 min and wasstored at –70 °C until assay. Plasma samples werethawed and extracted with dichloromethane, andmelatonin concentrations were determined in a singleassay as previously described (Bae et al., 1999). Theintra-assay coefficient of variation was 14%, and assaysensitivity was 17 pg/ml. Samples below the limit ofdetectability were assigned values of 17 pg/ml forpurposes of graphing and statistics. One to 3 days priorto each of the three runs of NWR described above,blood samples were collected at ZT 16, ZT 20, and ZT21 from pseudo-randomly selected unsplit hamsters(n = 3-5/time point). After the 6th day of NWR of eachexperimental run, additional samples were collectedat these same times or at ZT 9 from hamsters that werethen split (n = 3-6/time point; Fig. 1B). Amongunsplit control animals that remained at home, sam-pling was performed at ZT 9 on the 6th day of theafternoon dark pulse of each run (Fig. 1B). No hamstercontributed more than one sample at any given timepoint, except non-NWR controls at ZT 9. As melatoninconcentrations were always undetectable at ZT 9 forthese 4 hamsters, only one determination from eachhamster was considered in statistical analyses.

SCN Fos-Immunoreactivity

Finally, we assessed whether the temporal patternof light-induced Fos expression in the SCN was splitfollowing exposure to daily NWR. Following the

Gorman et al. / SPLIT CIRCADIAN RHYTHMS IN HAMSTERS 555

Figure 1. Schematic representation of times of experimentalmanipulations in hamsters with split activity rhythms and inunsplit controls. Each horizontal bar represents a 24-h period,with dark rectangles reflecting times of lights-off. Hatched rect-angles indicate times of novel wheel running (NWR) in darkness.One to 2 days following the final day of NWR are plotted below.Superimposed white dumbbell shapes represent characteristictimes of wheel-running activity in DD. Asterisks indicate time oflight pulse/sham pulse for determination of behavioralphase-shifting (A), blood sampling (B), or brain collection forSCN Fos-ir (C). In (A), the gray bar indicates that lights were extin-guished 2 h early for assessment of the earliest light pulses.

Page 45: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

behavioral phase-shifting experiment, hamsters (now38 weeks of age) were re-entrained to LD 14:10 andexposed to 6 days of NWR. Areplicate experiment wassubsequently performed with 18 additional hamstersat 12 weeks of age. For both cohorts, hamstersremained at home and lights remained off after thenight that followed the 6th day of NWR (Fig. 1C).Hamsters with split activity patterns were randomly

assigned to receive a 15-min light pulse or sham pulseat one of three time points (Fig. 1C; n = 3-6/group): 1 hafter the subsequent nighttime activity onset (CT13-n), during the morning inactive period (break) fol-lowing nighttime activity (2-4 h after nighttime activ-ity offset), and 1 h after the subsequent afternoonactivity onset (CT 13-a). Unsplit control hamsters werealso left in DD and were exposed to light or sham-

556 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 2. Representative doubled-plotted actograms (left) of hamsters split by novel wheel running (NWR), transferred to DD and pulsedwith light (A) or sham-pulsed (B) 1 h after the afternoon activity onset. Data are unclipped and are scaled between zero and maximum revo-lutions for that hamster. The light-dark cycle prior to transfer to DD is represented above the actogram, with hatched areas representingtimes of NWR in darkness. Days of NWR are represented by a similar rectangle superimposed upon the right half of the actogram. AfterNWR, hamsters entered DD beginning during the nighttime scotophase. The first 3 days of DD exposure are enlarged and singly plotted(right) to illustrate method of phase-shift calculation. Double-headed arrows represent Day 1 and Day 2 phase shifts of each component.Onsets of nighttime and afternoon activity on the 1st day in DD served as reference points from which to measure phase shifts. The differ-ence between initial and subsequent activity onsets was measured for each component (net phase shift).

Page 46: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

pulsed the next afternoon at the same time when splithamsters were given their light pulses. Additionalunsplit control hamsters received light pulses or shampulses 1 h after the following nighttime activity onset.

Hamsters were injected with a lethal dose ofNembutal 55 min after the light pulse or sham pulseand perfused intracardially with 40 to 60 ml 0.1Mphosphate buffered saline (PBS; pH 7.5) followed by100 to 150 ml 4% paraformaldehyde in PBS. Brainswere postfixed in paraformaldehyde at room temper-ature for 2 h and transferred to 20% sucrose/PBS.Serial coronal slices were cut at 40 µm and every fourthsection processed for immunocytochemistry. Free-floating sections were incubated successively in 0.1%hydrogen peroxide; 1:1000 anti-fos rabbit IgG (sc-52,Santa Cruz, CA, USA) in PBS with 0.4% Triton- X(Fisher) and 4% normal goat serum for 24 h at 4 °C;1:200 biotinylated goat anti-rabbit secondary anti-body (Vector Labs, Burlingame, CA, USA) for 1 h atroom temperature; ABC reagent (Vector) for 1 h; and0.1% diaminobenzidine (DAB) with 0.02% peroxidefor 2 min. Sections were mounted on gelatin-coatedslides. The most densely stained section was selectedby an observer blind to experimental treatment, andthe number of Fos-positive cells in the SCN wascounted.

Statistical tests (all two-tailed where applicable)were performed with Statview 5.0 software (SAS Insti-tute, Cary, NC, USA).

RESULTS

NWR and Splitting

In each induction of splitting by NWR, virtually theentire sample (>85%) of hamsters ran robustly in thenovel wheels. A subset of the activity data duringNWR has been analyzed in detail (Gorman and Lee,2001) and thus is not reported here. Figure 2 depictshamsters split by NWR, released into DD, and pulsedwith light (A) or sham-pulsed (B).

LIGHT-PULSE INDUCEDPHASE SHIFTS OF ACTIVITY

Afternoon Light Pulse (CT 13-a)

The afternoon activity component of pulsed ham-sters was phase-delayed relative to that of unpulsed

controls on Day 1 (p < 0.05; Figs. 2A, 3A), with a trendin the same direction on Day 2 (p < 0.10). The nighttimeactivity component, which was not itself pulsed, wassignificantly phase-advanced by this afternoon lightpulse (p < 0.05 for both days, Fig. 3A). The free-runningperiod of the nighttime component was unaffected bythe light pulse (mean τ ± SE: 25.08 ± 0.12 vs. 24.80± 0.18 h for unpulsed and pulsed animals, respec-tively; p > 0.20).

Evening Light Pulse betweenSplit Activity Components

Early evening light pulses (matched to CT 13 ofunsplit control hamsters) had no significant effect onthe nighttime activity component of split hamsters(Fig. 3B). The free-running period of this componentalso was not affected by the light pulse (25.11 ± 0.14 vs.25.37 ± 0.20 h for unpulsed and pulsed animals,respectively; p > 0.30).

Nighttime Light Pulse (CT 13-n)

On Days 1 and 2, the nighttime activity componentof pulsed hamsters was significantly phase-delayedby the 15-min light pulse (p < 0.005, p < 0.01, respec-tively, Fig. 3C), compared with sham light pulse con-trols. On Days 1 and 2, afternoon activity in the splitrhythm component was not significantly advanced bylight pulses (p > 0.25; Fig. 3C). The free-running periodof the nighttime activity component was unaffectedby the light pulse (mean τ ± SE: 24.68 ± 0.04 vs. 24.55 ±0.11 h for unpulsed and pulsed hamsters, respectively;p > 0.28).

Melatonin in Circulation

Prior to NWR, plasma melatonin concentrationswere detectable in only a minority (4/12) of unsplithamsters at ZT 16 but had increased above thresholdin 9 of 10 animals by ZT 20 (Fig. 4A). One hour later(ZT 21), only 2 out of 12 remained above detectablelimits. In contrast, plasma melatonin concentrations ofhamsters with split activity patterns in no case (0 outof 10) exceeded the minimum detectable level at ZT16, a significantly smaller proportion than among con-trols at this time (χ2 = 4.1; p < 0.05; Fig. 4B). By ZT 21,however, plasma melatonin had increased in 6 out of10 hamsters, a proportion significantly greater thanamong hamsters not exposed to NWR (χ2 = 4.4; p <

Gorman et al. / SPLIT CIRCADIAN RHYTHMS IN HAMSTERS 557

Page 47: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

558 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 3. Phase-shifting responses to 15-min light pulses delivered at three phases of the circadian cycle of split hamsters. In each panel,the mean (± SE) change in nighttime and afternoon activity onset relative to the first value in DD is depicted for light-pulsed (open bars)and unpulsed (sham) hamsters (filled bars). The net phase shift is indicated above each pair. Sample size is represented below or abovebars (pulsed/unpulsed). Asterisks denote significant differences between pulsed and unpulsed animals (p < 0.05; Mann-Whitney U;two-tailed). (A) Day 1 and Day 2 phase shifts after light pulse or sham pulse delivered 1 h after onset of the afternoon activity component(CT 13-a). (B) Phase shifts after pulse or sham pulse during the evening inactive period (see text). Design limitations prevented collection ofa prepulse reference value for the afternoon activity component. (C) Phase shifts after pulse or sham pulse delivered 1 h after the nighttimeactivity onset (CT 13-n).

Page 48: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

0.05), but not different from controls at ZT 20. More-over, after 6 days of NWR, nearly all NWR-split ham-sters (17 out of 18) had elevated concentrations ofmelatonin in circulation at ZT 9. This proportion wassignificantly greater than in non-NWR controlsexposed to similar light conditions where noneyielded detectable melatonin titers at ZT 9 (χ2 = 16.6;p < 0.001).

SCN Fos-Immunoreactivity

In hamsters with unsplit activity rhythms, the15-min light pulse after nighttime activity onsetinduced significant Fos expression, compared withthat in unpulsed controls (p < 0.05; Mann-Whitney U;Figs. 5A, 6A). A light pulse coinciding with the previ-ous afternoon dark period had no stimulatory effecton the number of Fos+ cells in these unsplit hamsters(p > 0.50; Figs. 5B, 6A). Hamsters split by NWR, in con-trast, showed two periods of increased Fos expression(Figs. 5C, 5E, 6B): Light pulses increased Fos-ir afterthe nighttime activity onset (p < 0.05) and after after-noon activity onset (p < 0.05). Moreover, light adminis-tered between these two time points had no effect onthe number of Fos+ cells (p > 0.40; Figs. 5D, 6B).Unpulsed controls at every time point tested—whether split or unsplit—showed minimal Fosexpression (Figs. 5F, 6). In groups where Fos wasrobustly induced by light (split hamsters in the nightand the afternoon and unsplit hamsters in the night),there were no differences in the number of Fos+ cellscounted (p > 0.05). Fos expression was symmetric withrespect to the left and right SCN and was concentratedin the ventrolateral SCN in all groups.

DISCUSSION

As in previous experiments, daily exposure toNWR in an LDLD cycle induced split activity rhythmsthat free-ran and recoupled in DD (Mrosovsky andJanik, 1993). Circadian mechanisms underlying eachbout responded similarly to timed light pulses as mea-sured by behavioral phase-shifts or induction of Fos-irin the SCN, and both activity components wereaccompanied by elevated plasma melatonin concen-trations. The results suggest that NWR temporallyreorganizes circadian oscillators within the SCN intotwo distinct components. We propose that each com-ponent oscillator or group of oscillators contains a rel-atively short subjective night characterized by loco-motor activity, elevated melatonin secretion, and lightresponsiveness.

Conceivably, NWR might have re-entrained the cir-cadian system to generate a long subjective night thatspanned the original night and the afternoonscotophase paired with NWR. If one of the two

Gorman et al. / SPLIT CIRCADIAN RHYTHMS IN HAMSTERS 559

Figure 4. Plasma melatonin concentrations (mean ± SE) of ham-sters with (A) unsplit or (B) split activity rhythms. Sample sizeand proportion of hamsters with detectable melatonin concentra-tions are given above histogram. Asterisks denote significant dif-ferences (p < 0.05) in the proportion of split and unsplit hamsterswith elevated melatonin concentrations at that sampling interval.

Page 49: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

photophases separating dark periods masked loco-motor activity, the rhythm might only appear split. Inthe present study, light pulses had marked effectswhen delivered during either activity component butdid not induce Fos expression in the SCN during themorning inactive period or shift activity rhythms inthe early evening inactive interval. Limited resourcesprecluded assessment of possible light-induced phaseshifts during the morning rest period or of SCN Fos-irexpression in the early evening. In unsplit hamsters,however, Fos induction by light is restricted to subjec-tive night and temporally correlates closely with peri-ods of behavioral phase shifting (Kornhauser et al.,1990; Sumová et al., 1995; Travnickova et al., 1996).This close relationship, moreover, is maintained invarious photoperiods that alter α. Despite assessmentby different methodologies, each inactive period thusappears to represent a dead zone with respect to lightresponsiveness. These dead zones separate intervalsof light responsiveness as measured jointly by behav-ioral phase-shifting and Fos-ir induction.

Elevated melatonin in circulation during both thesubjective afternoon and nighttime bouts of activity inhamsters with split activity rhythms supports thehypothesis that the principal circadian pacemaker is

split by NWR. Among unsplit hamsters, in contrast,plasma melatonin concentrations were elevated in amonophasic pattern. Four hours into the night (ZT 16),4 out of 12 hamsters had detectable concentrations ofmelatonin in circulation, consistent with onset of therising phase in pineal melatonin production in thisspecies (Elliott and Tamarkin, 1994; Goldman et al.,1981). A nighttime elevation in plasma melatonin atZT 20 was followed by decreases below detectable lev-els that were manifest by some hamsters at ZT 21 andall hamsters at ZT 9. If NWR delayed the circadianpacemaker as suggested by nighttime activity onsets(Gorman and Lee, 2001) in split hamsters, no elevationof melatonin concentrations at ZT 16 would beexpected; importantly, none was found. Increasedmelatonin secretion clearly occurred later during thenighttime bout of activity. However, the morningdecline in circadian melatonin production, evident inthe majority of unsplit hamsters at ZT 21, was notobserved in split animals, providing further evidenceof a delayed nighttime oscillator. Most important,plasma melatonin concentrations were elevated at theend of the afternoon scotophase paired with NWR.Among hamsters exposed to light-dark cycles, dark-ness in the afternoon to our knowledge has never been

560 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 5. Representative photomicrographs of Fos-immunostained sections from unsplit (A, B, F) and split (C-E) hamsters pulsed orsham-pulsed with light during the nighttime active phase (A, C), the morning inactive period (D), and the afternoon (B, E, F). For the lattertime point, split hamsters were active, but unsplit hamsters were inactive.

Page 50: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

reported to induce melatonin secretion, and no eleva-tion was apparent in nonrunning controls. Thus, thesedata suggest that the pattern of melatonin, like that ofactivity, is split by NWR into two components.

The ability of both nighttime and afternoon lightpulses to induce Fos expression in the SCN discountsthe possibility that split rhythms reflect recruitment of

extra-SCN oscillators that might mediate one of thetwo activity components. Although their neural sub-strates are unknown, extra-SCN oscillators are suffi-cient to generate circadian rhythms in a variety ofexperimental paradigms (Honma et al., 1989; Stephanet al., 1979). It is not feasible to observe SCN Fosexpression in a single hamster during each activitybout, but the uniformity of expression of Fos in theSCN following the afternoon or nighttime light pulseestablishes that the principal circadian pacemakeritself responds to light during both activity compo-nents. In LL-induced splitting, rhythms of Fos andclock gene expression are out of phase in the left andright SCN (de la Iglesia et al., 2000), although otherinvestigators discerned no left/right differences inelectrophysiological activity in the SCN of LL-induced split hamsters (Zlomanczuk et al., 1991). Thepresent Fos data do not suggest an anatomical orphysiological distinction between the two oscillators.In the present study, all sections revealed symmetricalinduction of Fos in the two SCN, indicating that bothleft and right are expressing a subjective night simul-taneously rather than alternately. Within each SCN,the pattern of Fos induction was also similar betweenthe two bouts, with expression concentrated in theventrolateral SCN. How each SCN becomes Fos-inducible twice daily is not clear. At the tissue level oforganization, the oscillators may be two distinct, butspatially intermixed, cell populations, each respon-sive to light during either the afternoon or nighttimedark period, but not both. Alternatively, a single pop-ulation of SCN cells may be induced to express Fostwice daily, perhaps because individual cells in thispopulation contain two oscillators within them, orbecause they are simultaneously clock-controlled bytwo other cell populations with rhythms out of phase.

Between-group comparisons demonstrated robustbehavioral phase shifts after light pulses. Phase shiftscalculated in this fashion might result either from dis-crete phase shifts or from changes in free-runningperiod (e.g., a lengthening of τ after a light pulsewould appear as Day 1 and Day 2 phase-delays by thismethod). The latter possibility is discounted, how-ever, by the analysis of free-running period in the daysfollowing the light pulse. After all pulses, the free-run-ning period was not significantly altered; and in thecase of the nighttime pulses, the trend was toward ashortening of τ, which would act counter to thereported phase delays. Thus, phase shifts of the activ-ity components were not secondary to changes in τ butinstead reflect discrete phase shifts.

Gorman et al. / SPLIT CIRCADIAN RHYTHMS IN HAMSTERS 561

Figure 6. Number of Fos-ir cells/SCN section of (A) unsplit and(B) split hamsters pulsed with light (open bars) or sham-pulsed(filled bars) at various times. Sample size is indicated in or abovebars. Asterisks denote significant increases (p < 0.05; Mann-Whit-ney U) in the number of Fos-ir cells relative to unpulsed controlsat that time point.

Page 51: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

As in LL-induced split animals probed with darkpulses (Boulos and Rusak, 1982; Lees et al., 1983), bothNWR-split activity components responded with simi-lar phase shifts to light pulses 1 h after their respectiveactivity onsets. Afternoon light pulses additionallyinduced phase advances of the unpulsed nighttimecomponent, but pulses in the early nighttimescotophase did not shift rhythms; nor did light pulsesin the morning inactive period lead to induction ofFos-ir in the SCN, as is generally a prerequisite forlight-induced phase shifts. How is this pattern ofresponsiveness to be understood? One possibility isthat the light-pulse PRC for the nighttime componentdoes indeed contain two light-responsive zones sepa-rated by two dead zones. Alternatively, each compo-nent may be directly light responsive only during asingle short subjective night associated with only oneof the two scotophases. The greater fraction of eachoscillation may be a dead zone. The observed phaseadvances of the nighttime activity component follow-ing light at CT 13-a, however, could be a secondaryconsequence of interactions with the afternoon com-ponent, which is directly phase-delayed by light.When light phase-delays the afternoon activity com-ponent by 90 min as in the current study, the phaseangle between the nighttime and afternoon compo-nents is altered. If the interaction between the twooscillators depends on their phase relation as com-monly presumed, the phase shift of the unpulsed com-ponent might reflect this altered interaction with thepulsed component rather than a direct action of light,per se. Future work can develop and test this model ofoscillator interactions.

Multioscillatory models of circadian rhythms com-monly distinguish between E and M oscillators (Daanand Berde, 1978; Elliott and Tamarkin, 1994; Gormanet al., 1997; Illnerova and Vanacek, 1985; Pittendrighand Daan, 1976; Sumová et al., 1995). It is not possibleto attribute the respective activity components gener-ated here to hypothesized evening and morning oscil-lators. Regardless of the relation to E and M, however,the circadian timekeeping system can be resolved intotwo seemingly similar oscillators, both of which canbe phase-delayed by light and both of which can drivemelatonin secretion within a single 24-h period. Fur-ther probing of these two oscillators will complementin vitro work to characterize the suboscillatory basis ofthe mammalian pacemaker.

ACKNOWLEDGMENTS

We are grateful to Jim Donner for excellent animalcare and Jeff Elliott for helpful comments on an earlierdraft of this manuscript. This research was supportedby NHLBI grant HL61667 to TML and NICHD-48640to MRG.

REFERENCES

Bae HH, Mangels RA, Cho BS, Dark J, Yellon SM, and Zucker I(1999) Ventromedial hypothalamic mediation of photo-periodic gonadal responses in male Syrian hamsters. JBiol Rhythms 14:391-401.

Boulos Z and Rusak B (1982) Phase-response curves and thedual-oscillator model of circadian pacemakers. In Verte-brate Circadian Systems, J Aschoff, S Daan, and G Groos,eds, pp 215-223, Springer-Verlag, Berlin.

Bouskila Y and Dudek FE (1993) Neuronal synchronizationwithout calcium-dependent synaptic transmission in thehypothalamus. Proc Natl Acad Sci U S A 90:3207-3210.

Daan S and Berde C (1978) Two coupled oscillators: Simula-tions of the circadian pacemaker in mammalian activityrhythms. J Theor Biol 70:297-313.

de la Iglesia HO, Meyer J, Carpino A Jr, and Schwartz WJ(2000) Antiphase oscillation of the left and rightsuprachiasmatic nuclei. Science 290:799-801.

Elliott JA and Kripke DF (1998) Photoperiodic regulation oflight induced phase shifts and PRC amplitude in wildtype male and female Syrian hamsters. Soc Res BiolRhythm Abstr 6:62.

Elliott JA and Tamarkin L (1994) Complex circadian regula-tion of pineal melatonin and wheel-running in Syrianhamsters. J Comp Physiol [A] 174:469-484.

Goldman BD, Hall V, Hollister C, Reppert S, Roychoud-hury P, Yellon S, and Tamarkin L (1981) Diurnal changesin pineal melatonin content in four rodent species: Rela-tionship to photoperiodism. Biol Reprod 24:778-783.

Gorman MR, Freeman DA, and Zucker I (1997) Photo-periodism in hamsters: Abrupt versus gradual changesin day length differentially entrain morning and eveningcircadian oscillators. J Biol Rhythms 12:122-135.

Gorman MR and Lee TM (2001) Daily novel wheel runningreorganizes and splits hamster circadian activityrhythms. J Biol Rhythms 16:540-549.

Herzog ED, Geusz ME, Khalsa SBS, Straume M, and BlockGD (1997) Circadian rhythms in mouse suprachiasmaticnucleus explants on multimicroelectrode plates. BrainRes 757:285-290.

Honma S, Honma K, and Hiroshige T (1989) Methamphet-amine induced locomotor rhythm entrains to restricteddaily feeding in SCN lesioned rats. Physiol Behav45:1057-1065.

Illnerova H (1991) The suprachiasmatic nucleus and rhyth-mic pineal melatonin production. In SuprachiasmaticNucleus: The Mind’s Clock, DC Klein, RY Moore, and SMReppert, eds, pp 197-216, Oxford University Press, NewYork.

562 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 52: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Illnerova H and Vanacek J (1985) Complex control of the cir-cadian rhythm in pineal melatonin production. In ThePineal Gland: Current State of Pineal Research, B Mess,C Ruzsas, L Tima, and P Pevet, eds, pp 137-153, NewYork, Elsevier.

Jagota A, de la Iglesia HO, and Schwartz WJ (2000) Morningand evening circadian oscillations in the suprachiasmaticnucleus in vitro. Nat Neurosci 3:372-376.

Kornhauser JM, Nelson DE, Mayo KE, and Takahashi JS(1990) Photic and circadian regulation of c-fos geneexpression in the hamster suprachiasmatic nucleus. Neu-ron 5:127-134.

Lees JG, Hallonquist JD, and Mrosovsky N (1983) Differen-tial effects of dark pulses on the two components of splitcircadian activity rhythms in golden hamsters. J CompPhysiol [A] 153:123-132.

Liu C, Weaver DR, Strogatz SH, and Reppert SM (1997) Cel-lular construction of a circadian clock: Period determina-tion in the suprachiasmatic nuclei. Cell 91:855-860.

Mason R (1991) The effects of continuous light exposure onSyrian hamster suprachiasmatic (SCN) neuronal dis-charge activity in vitro. Neurosci Lett 123:160-163.

Mrosovsky N and Janik DS (1993) Behavioral decoupling ofcircadian rhythms. J Biol Rhythms 8:57-65.

Pittendrigh CS and Daan S (1976) A functional analysis ofcircadian pacemakers in nocturnal rodents: V. Pacemaker

structure: A clock for all seasons. J Comp Physiol [A] 106:333-355.

Pittendrigh CS, Elliott JA, and Takamura T (1984) The circa-dian component in photoperiodic induction. CIBAFound Symp 104:26-47.

Stephan FK, Swann JM, and Sisk CL (1979) Entrainment ofcircadian rhythms by feeding schedules in rats withsuprachiasmatic lesions. Behav Neural Biol 25:545-554.

Sumová A, Trávnícková Z, Peters R, Schwartz WJ, andIllnerová H (1995) The rat suprachiasmatic nucleus is aclock for all seasons. Proc Natl Acad Sci U S A 92:7754-7758.

Travnickova Z, Sumova A, Peters R, Schwartz WJ, andIllnerova H (1996) Photoperiod-dependent correlationbetween light-induced SCN c-fos expression and reset-ting of circadian phase. Am J Physiol 271:R825-R831.

Weaver DR (1998) The suprachiasmatic nucleus: A 25-yearretrospective. J Biol Rhythms 13:100-112.

Welsh DK, Logothetis DE, Meister M, and Reppert SM(1995) Individual neurons dissociated from rat supra-chiasmatic nucleus express independently phased circa-dian firing rhythms. Neuron 14:697-706.

Zlomanczuk P, Margraf RR, and Lynch GR (1991) In vitroelectrical activity in the suprachiasmatic nucleus follow-ing splitting and masking of wheel-running behavior.Brain Res 559:94-99.

Gorman et al. / SPLIT CIRCADIAN RHYTHMS IN HAMSTERS 563

Page 53: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

JOURNAL OF BIOLOGICAL RHYTHMS / December 2001Watanabe et al. / TRANSIENT VERSUS STEADY-STATE PHASE SHIFTS

Light-Induced Resetting of the CircadianPacemaker: Quantitative Analysis of

Transient versus Steady-State Phase Shifts

Kazuto Watanabe,* Tom Deboer,† and Johanna H. Meijer†,1

*Department of Physiology, Dokkyo University School of Medicine, 321-02 Mibu, Japan,†Department of Physiology, Leiden University Medical Center, Leiden, the Netherlands

Abstract The suprachiasmatic nuclei of the hypothalamus contain the major cir-cadian pacemaker in mammals, driving circadian rhythms in behavioral andphysiological functions. This circadian pacemaker’s responsiveness to lightallows synchronization to the light-dark cycle. Phase shifting by light ofteninvolves several transient cycles in which the behavioral activity rhythm gradu-ally shifts to its steady-state position. In this article, the authors investigate inSyrian hamsters whether a phase-advancing light pulse results in immediateshifts of the PRC at the next circadian cycle. In a first series of experiments, theauthors aimed a light pulse at CT 19 to induce a phase advance. It appeared thatthe steady-state phase advances were highly correlated with activity onset in thefirst and second transient cycle. This enabled them to make a reliable estimate ofthe steady-state phase shift induced by a phase-advancing light pulse on thebasis of activity onset in the first transient cycle. In the next series of experiments,they presented a light pulse at CT 19, which was followed by a second light pulseaimed at the delay zone of the PRC on the next circadian cycle. The immediateand steady-state phase delays induced by the second light pulse were comparedwith data from a third experiment in which animals received a phase-delayinglight pulse only. The authors observed that the waveform of the phase-delay partof the PRC (CT 12-16) obtained in Experiment 2 was virtually identical to thephase-delay part of the PRC for a single light pulse (obtained in Experiment 3).This finding allowed for a quantitative assessment of the data. The analysis indi-cates that the delay part of the PRC—between CT 12 and CT 16—is rapidly resetfollowing a light pulse at CT 19. These findings complement earlier findings inthe hamster showing that after a light pulse at CT 19, the phase-advancing part ofthe PRC is immediately shifted. Together, the data indicate that the basis forphase advancing involves rapid resetting of both advance and delay compo-nents of the PRC.

Key words circadian rhythms, suprachiasmatic nucleus, entrainment, phase shift, tran-sients, phase response curve, evening/morning oscillators

An endogenous circadian pacemaker in thesuprachiasmatic nuclei (SCN) determines the timing

of many behavioral and physiological functions(Meijer and Rietveld, 1989). For proper timing, this

564

1. To whom all correspondence should be addressed: Department of Physiology, LUMC, P.O. Box 9604, 2300 RC Leiden, theNetherlands.

JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 564-573© 2001 Sage Publications

Page 54: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

pacemaker is responsive to the environmental light-dark cycle. Light presented during the beginning ofthe subjective dark period induces phase delays of thecircadian activity rhythm, whereas light presentedtoward the end of the subjective dark period producesphase advances. The effects of light on the circadianactivity rhythm can be described by a phase responsecurve (PRC). In a PRC, the magnitude and direction ofphase shifts in activity onset are plotted as a functionof the circadian time of pulse application (Daan andPittendrigh, 1976).

After a light pulse, a phase shift is often not com-pleted within the first circadian cycle but insteadgrows over the course of several days. Such circadiancycles are called transient cycles. They are most pro-nounced following phase-advancing light pulses, butthey occur after delaying pulses as well (Pittendrighet al., 1958). It is not clear whether activity onset dur-ing the transient cycles reflects the true position of theunderlying pacemaker. One possibility is that thepacemaker itself requires several days to complete thephase shift. Alternatively, the pacemaker may shiftimmediately to its final position but the activityrhythm requires several days to become fully synchro-nized with the new phase of the pacemaker.

On the molecular level, it has been shown that theexpression of the mammalian homologs of the insectperiod gene, mper1 and mper2, peak 1 h and 2 h afterapplication of a light pulse, respectively (Albrechtet al., 1997; Shearman et al., 1997; Shigeyoshi et al.,1997). It is clear that resetting mechanisms are inducedvery quickly and increases in the putative proteinproducts of mper may be the cause of resetting to a newphase. The finding that inhibition of mper1 expressionblocks the phase shifts induced by light and glutamatepulses supports the notion that mper is required forphase shifting by light (Akiyama et al., 1999). Itappears therefore that resetting of the clock is estab-lished within a few hours. To understand the discrep-ancy between the molecular biology and the behaviorof the overt rhythm, it is important to get insight intothe position of the pacemaker during the transientcycles in vivo. For this purpose, it is necessary to deter-mine the position of the PRC during the transientcycle, assuming that the position of the PRC providesa true reflection of the phase of the pacemaker.

One way to investigate the position of the PRC is byapplying a second light pulse during the first transientcycle after a light pulse and to investigate the phase-shifting effect of the two light pulses on the final phaseof the activity rhythm. Two-pulse experiments have

been performed in invertebrates such as Drosophila(Pittendrigh, 1979) and Neurospora (Crosthwaite et al.,1995), in mice (Sharma and Chandrashekaran, 2000),and in Syrian hamsters (Elliot and Pittendrigh 1996;Best et al., 1999). In the hamster, it has been establishedthat the phase-advance part of the PRC shifts within afew hours after a light pulse at CT 18 by applying a sec-ond light pulse 1 to 2 h after the first pulse (Best et al.,1999). It was also shown that the phase-delay partshifts within a few hours after a light pulse at CT 13 byapplying a second pulse 1 to 2 h after this pulse (Bestet al., 1999). However, it has not been investigatedwhether the phase-delay part of the PRC shifts imme-diately after a phase-advancing light pulse. This mat-ter is of great importance in view of recent findings,indicating that the pacemaker is composed of distinctgenetic components that exhibit different responsive-ness to phase-advancing and phase-delaying lightpulses (see Daan et al., 2001).

We investigated the responsiveness of the phase-delay part of the PRC during a transient cycle that wasinduced by a phase-advancing pulse. To this purpose,we applied a second light pulse shortly after activityonset at the first transient cycle. By comparing theeffect of a single light pulse with the effect of the dou-ble light pulse, we could investigate the phase-shiftingeffect of the second light pulse. From the magnitude ofthe phase shift, it can be estimated what the position ofthe circadian clock was at the time point of applicationof the second light pulse. To optimize this estimate, wedetermined the relation between activity onset duringthe first transient cycles and the final steady-statephase after application of a single phase-advancinglight pulse. This enabled us to make a reliable estimateof the steady-state phase shift after the first activityonset.

METHODS

This study was performed on 60 male Syrian ham-sters (Mesocricetus auratus, Harlan/CPB, Zeist, theNetherlands) aged 2 months at the start of the experi-ment. The animals were individually housed in cages(36.5 × 25.0 × 16.0 cm) with a running wheel (diameter26 cm) and were kept in a sound-attenuating, venti-lated room at a temperature of 23 °C. Food (Hopefarms B.V. the Netherlands) and water were continu-ously available. Running-wheel activity was recordedper minute to determine the animals’ circadian activ-ity rhythm. In all the experiments, light pulses (15

Watanabe et al. / TRANSIENT VERSUS STEADY-STATE PHASE SHIFTS 565

Page 55: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

min) were presented to groups of animals, and theprotocol was repeated three times. Positioning of thelight sources on the wall behind every single cageensured similar light levels (100-120 lux) for all ani-mals. During presentation of the light pulses, the ani-mals remained in their home cages. All animalsreceived all three treatments, and treatments werepresented in a randomized order. At least 1 monthelapsed between application of subsequent lightpulses. Data were excluded from the analysis whenthe actogram of running-wheel activity did not allowfor unambiguous determination of activity onsets.

Experiment 1

The animals were entrained to LD 14:10 for at least7 days before they were released into constant dark-ness (DD). After 7 days in DD (Day 0), the animalsreceived a light pulse between CT 17.77 and CT 21.90.After the light pulse, the animals were kept in DD for14 days. Lines were eye-fitted through activity onsetsbefore and after the light pulse. The first transientcycles after the light pulse were excluded from the fit.Steady-state phase shifts were determined by measur-ing the difference between the fitted lines extrapolatedto the first cycle after the light pulse (Day 1, Fig. 1A). Inaddition, the immediate phase shift on the first (Day 1)and second circadian cycle (Day 2) after the light pulsewere determined. The immediate phase shift wasdefined as the difference between the time of observedactivity onset and the time as predicted by the linethrough activity onsets before the light pulse.

A strong correlation between the immediate phaseshift on Day 1 and the steady-state phase shift wasfound. Moreover, we found a strong correlationbetween the immediate shift on Days 1 and 2 (for fur-ther details, see Results and Fig. 3 B,C). These correla-tions were used for the analysis of Experiment 2.

Experiment 2

Alight pulse was applied between CT 17.75 and CT21.71 according to the protocol as in Experiment 1. Inthis experiment, the light pulse was followed by a sec-ond light pulse given 0.01 to 4.31 h after activity onseton the first transient cycle (Day 1, Fig. 1B). After thesecond light pulse, the animals were kept in DD for14 days. The immediate phase shift induced after thefirst and second light pulses and the steady-statephase shift induced by the two light pulses weredetermined.

We considered that the steady-state phase shiftinduced by the second light pulse is equal to the differ-ence between the steady-state phase shift induced bythe first light pulse and the steady-state phase shiftinduced by the two pulses together. To estimate thesteady-state phase shift induced by the first light

566 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 1. Experimental protocol (see Methods for explanation ofthe protocol). A. Experiment 1: A light pulse was applied betweenCT 17.77 and CT 21.90 on Day 0. The steady-state phase shift (∆φst,A1 – C1), and the immediate phase shift on Day 1 (∆φim(1), A1 – B1)and Day 2 (∆φim(2), A2 – B2) were measured. B. Experiment 2: Alight pulse was applied between CT 17.75 and CT 21.71 on Day 0and a second pulse was given 0.01 to 4.31 h after activity onset onDay 1. Steady-state phase shifts (∆φst, E1 – C1) and immediatephase shifts (∆φim, D2 – B2) induced by the second light pulse weremeasured. C. Experiment 3: A light pulse was applied betweenCT 10.80 and CT 15.43 on Day 1. Steady-state phase shifts (∆φst,C1 – A1) and immediate phase shifts (∆φim, B2 – A2) were measured.

Page 56: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

pulse, the strong correlation between the immediatephase shift and the steady-state phase shift obtainedin Experiment 1 was used. In other words, the imme-diate phase shift on Day 1 was determined and thesteady-state phase shift (A1 – C1) for that particular ani-mal was estimated on the basis of the regression lineobtained in Experiment 1 (Fig. 3B). To determine thesteady-state phase shift induced by the two lightpulses, the difference between the steady-state activ-ity onset lines before and after the two light pulses wascalculated (A1 – E1). The steady-state phase shiftinduced by the second light pulse only is then (A1 –E1) – (A1 – C1) = C1 – E1.

Similarly, the immediate phase shift induced by thesecond light pulse was estimated by measuring thedifference between the time of observed activity onseton Day 2 (B2) and the time of activity onset on Day 2when the second light pulse would not have beengiven (D2). This latter value was predicted from thestrong correlation between immediate shifts on Day 1and Day 2 obtained in Experiment 1.

Experiment 3

A light pulse was presented according to the proto-col as in Experiment 1. The light pulse was appliedbetween CT 10.80 and 15.43 on Day 1. After the lightpulse, the animals were kept in DD for 14 days. Thesteady-state phase shift on the day of the light pulse(Day 1) was determined by measuring the differencebetween the fitted lines before and after the light pulse(Fig. 1C). The immediate phase shift on the first cycleafter the light pulse (Day 2) was measured as the dif-ference between the time of observed activity onsetand the time predicted by the line through activityonsets preceding the light pulse. ANOVAand post hoct tests served to compare Experiment 2 and Experi-ment 3.

RESULTS

Experiment 1

In Experiment 1, 93 phase advances were analyzed(Fig. 2A). These light pulses were presented betweenCT 17.77 and CT 21.90 (mean = 19.07). The effect of thelight pulse on the steady-state phase shift, ∆φst, and onthe immediate phase shift on the first cycle after thelight pulse, ∆φim(1), are summarized in Figure 3A. Themagnitudes of the steady-state phase shift and the im-

mediate phase shift were 1.42 ± 0.44 h and 0.64 ± 0.24 h,respectively (mean ± SD). Although the same animalsreceived light pulses several times, the differencebetween interindividual and intraindividual varia-tions was not significant for the magnitudes of bothsteady-state and immediate phase shifts. Therefore,phase shifts obtained from the same animal were treat-ed as independent values in all of the experiments.

The results of this experiment are important for theprotocol of the next experiment. As is evident fromFigure 3A, a rather large range of phase shifts can beobtained at each circadian time. However, when plot-ting the steady-state phase shift, ∆φst, as a function ofthe magnitude of the immediate shift on Day 1, ∆φim(1)

(Fig. 3B), a strong relation is observed between imme-diate and steady-state phase shift, which can bedescribed as follows:

∆φst = 1.462 × ∆φim(1) + 0.482 (r = 0.79, p < 1*10–20).

This regression line was used to predict for each ani-mal the steady-state phase shift from the immediateshift on Day 1 in Experiment 2.

Moreover, a strong correlation between the magni-tude of the immediate shift on Day 1, ∆φim(1), and theimmediate shift on Day 2, ∆φim(2), was found (Fig. 3C).This correlation (r = 0.89, p < 1*10–20) can be describedby the following function:

Watanabe et al. / TRANSIENT VERSUS STEADY-STATE PHASE SHIFTS 567

Figure 2. Typical examples of activity records of all three experi-mental conditions in the same animal. The activity record isdouble-plotted to enable visualization of the activity rhythms.Arrows and stars indicate day and time of light pulse application.Panel A shows a phase advance after application of a light pulse atCT 21. Panel B shows a phase shift after application of a lightpulse CT 21.5 followed by a light pulse applied 2.75 h after activityonset on the next circadian cycle. Panel C shows a phase delayafter application of a light pulse at CT 14.

Page 57: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

∆φim(2) = 1.325 × ∆φim(1) + 0.150.

This regression line was used in Experiment 2 to pre-dict for each animal the time of activity onset on Day 2in the case that only the first light pulse would havebeen applied and to measure the difference with theactivity onset on Day 2.

Experiment 2

In the double pulse experiment, 85 activity recordswere obtained that allowed for an unambiguous anal-ysis (Fig. 2B). The first light pulse fell between CT17.75 and CT 21.71 (mean = 19.48), inducing a phaseadvance on Day 1 in all cases (range from 0.24 to 1.26 h;mean = 0.76 h). The second pulse was given 0.01 to4.31 h (mean = 1.61 h) after activity onset on Day 1.Those cases where the second light pulse fell beforeactivity onset were excluded from the analysis sinceno estimate of activity onset on Day 1 could be made.In most cases, the second light pulse induced a phasedelay but a few advances were also observed.

Experiment 3

Eighty-five clear phase-shifts were obtained fromlight pulses that fell between CT 10.80 and CT 15.43(mean = 12.95) and were used to describe the phase-delay part of the PRC (Fig. 2C). The maximum delaywas obtained around CT 13 (Fig. 4). Mean hourly val-ues for the immediate and steady-state phase shiftwere calculated between CT 11 and CT 16 and werecompared with the data from Experiment 2.

Comparison of Experiments 2 and 3

In Experiment 2, the second light pulse was given atthe beginning of the subjective night. The phase shiftinduced by this second light pulse was compared withthe shift induced by a single light pulse applied atcomparable phases (CT) in Experiment 3. The circa-dian time of application of the second light pulse isunknown. However, there are two predictions for thecircadian time of the second light pulse. (1) If the overtrhythm is the manifestation of the underlying oscilla-tor during transient cycles, the time of the activityonset on Day 1 is equal to CT 12. (2) If the steady-statephase shift reflects the real position of the oscillator,the extrapolated steady-state phase shift on Day 1 isequal to CT 12.

568 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 3. Immediate and steady-state phase shifts induced by aphase-advancing light pulse. A. The phase shifts are plotted as afunction of the time of light pulse application. Dots indicatesteady-state phase shifts, circles immediate shifts. B. Thesteady-state shift is plotted as a function of the immediate shift onDay 1. C. The immediate phase shift on Day 2 is plotted as a func-tion of the immediate shift on Day 1.

Page 58: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

The two possible phase delays obtained in Experi-ment 2 with the second light pulse were calculatedand compared with the data obtained in Experiment 3(Fig. 4). The analysis indicates that the steady-statephase shift in Experiment 2 did not differ significantlyfrom the PRC obtained in Experiment 3 when suppos-ing that the pacemaker shifted to its steady-state posi-tion within one circadian cycle (Fig. 4, Prediction 2). Incontrast, significant differences were obtained at CT13 and CT 14 between the steady-state phase shifts inExperiment 2 and Experiment 3 when supposing thatthe first transient indicates the phase of the pacemaker(Prediction 1). Also, the two predicted PRCs differedsignificantly at CT 13 and CT 14 (Fig. 4).

DISCUSSION

The question posed in this article is whether tran-sient cycles accompanying a light-induced phaseadvance reflect the circadian time of the pacemaker orwhether the pacemaker shifts immediately to itssteady-state position. We specifically addressed thequestion of whether the phase-delay part of the PRC

shifts immediately after application of a phase-advancing light pulse. To answer this question, wedetermined the position of the PRC at the time ofapplication of the second light pulse (Experiment 2)with the delay part of the PRC induced by a singlelight pulse (Experiment 3). The results are in accor-dance with previous studies (Best et al., 1999) and addto this that (a) the phase-delay part of the pacemakershifts within one circadian cycle to its new steady-stateposition after a phase-advancing light pulse, (b) themagnitude of a transient shift predicts the phase of thenew steady-state position, and (c) the waveform andamplitude of the phase-delay part of the PRC do notchange after a phase-advancing light pulse.

PRC Amplitude and Waveform Stability

Amplitude

The question arises whether the properties of thepacemaker change during transient cycles. In fact, thedata can only be interpreted on the premise thatthe pacemaker responds in its usual way to light dur-ing the transient cycles. Only then do the results fromExperiment 3 form a reliable prediction for the pace-maker’s responsiveness to the second light pulse.

Previous double-pulse experiments in inverte-brates (Drosophila, Pittendrigh, 1979; Neurospora,Crosthwaite et al., 1995), and the Syrian hamster (Bestet al., 1999) were performed to investigate if there is arefractory period in the pacemaker after application ofa light pulse and to determine how long this refractoryperiod lasts (i.e., how fast the pacemaker can react to asecond light pulse). These data show that the pace-maker is capable of reacting to a new light pulsewithin a few hours. In other studies, however, it wasshown that mice and hamsters are significantly lessresponsive to a second light pulse when appliedwithin 4 h after the first light pulse (Khammanivongand Nelson, 2000; Nelson and Takahashi, 1999). More-over, Khammanivong and Nelson (2000) indicatedthat responses are not even back to normal in the nextcircadian cycle (about 70% of expected shift). Ourexperiments show that the amplitude of the phasedelay is back to normal 17 to 21 h after a phase-advanc-ing light pulse. This raises the possibility that respon-siveness to light is decreased 24 h after a light pulse butis unchanged 17 to 21 h after a light pulse. In otherwords, the part of the PRC that received the first lightpulse is still affected by it after 24 h, whereas otherparts of the PRC may be unaffected.

Watanabe et al. / TRANSIENT VERSUS STEADY-STATE PHASE SHIFTS 569

Figure 4. Comparison of phase shifts induced by the singledelaying light pulse and those induced by the second of the dou-ble light pulse. Hourly values (± SEM) of steady-state phase shiftsinduced by the second light pulse are plotted against CT of thesecond light pulse in “steady-state phase” (triangles) or that in“transient phase” (circles) together with the phase shift inducedby the single light pulse (squares). Asterisks indicate significantdifferences between the conditions (p < 0.05, two-tailed t test aftersignificant ANOVA for factor “condition” over CT 13 to CT 15).

Page 59: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Waveform

The waveform of the PRC appeared unaltered onthe first cycle after the light pulse between CT 12 andCT 16. We have no knowledge of the phase responseproperties at other circadian times. Our data are con-sistent with unpublished data of Elliot and Pittend-righ (1996; Pittendrigh, 1981, Fig. 9). With respect tothe phasedelay part of the PRC, they showed thatbetween CT 12 and CT 16, the PRC had shifted to thesteady-state position within one circadian cycle. How-ever, they showed that the onset of the phase-delaypart of the PRC (about CT 11) had not shifted. As a con-sequence, the phase-delay part of their PRC was com-pressed. We did not investigate responsiveness beforeCT 12 because we administered our second light pulseafter activity onset of the animal to obtain an accurateprediction for the induced phase shifts.

Predicting Steady-State Phase Shiftfrom the First Transient Cycle

The present analysis shows that there is a strongrelationship between the position of the activity onsetduring the first and second transient cycles and thesteady-state phase shift of activity onset (Fig. 3). Thisenabled us to make a very reliable estimate of thesteady-state phase shift induced by the first light pulsein Experiment 2 on the basis of the first activity onset.The strong correlation indicates that the daily shifts inactivity onsets during transient cycles are regulatedwith great precision and reflect the magnitude of thesteady-state shift. Thus, the immediate shift is a fixedfraction of the steady-state shift. This indicates by itselfthat during the first transient cycle the steady-statephase shift is already determined. This is consistentwith our conclusion that the pacemaker has reachedits steady-state position on the first transient cycle.

This result is suggestive for an interaction betweenthe endogenous pacemaker and a secondary down-stream oscillatory system, either inside or outside theSCN, of which the kinetics can be described with greatmathematical precision. We will discuss possible sec-ondary systems in the next section.

Immediate Resetting

We showed that after a phase-advancing lightpulse, the phase-delay part of the PRC shifts withinone circadian cycle to the new steady-state positionwhile the overt rhythm displays transients. If the pace-

maker shifts immediately, the question arises, Whatprocess induces transient behavior? Transients inDrosophila have been attributed to the existence of twocoupled oscillators of which one is light sensitive,whereas the other is temperature sensitive (Pittendrighet al., 1958). In mammals, two mutually coupled circa-dian oscillators within the central pacemaker havebeen proposed (Pittendrigh and Daan, 1976). Oneoscillator is thought to lock onto the evening light (theE oscillator) and controls activity onset in nocturnalanimals, while the other locks onto the morning light(the M oscillator) and controls activity offset. It hasbeen suggested that the differential shift in onset andoffset after a phase-advancing light pulse causes dif-ferent shifts of E and M, with the M oscillator shiftingimmediately to its new steady-state position and Eshifting more slowly (Honma et al., 1985; Meijer andDevries, 1995; Elliot and Tamarkin, 1994; Pittendrighand Daan, 1976).

Accumulating evidence at a number of researchlevels supports the proposition, initially put forwardby Pittendrigh and Daan (1976), that the circadianpacemaker is composed of different oscillators. Daanet al. (2001) summarized these results elegantly in aconceptual framework. The E and M components arethought to result from a double set of circadian genes.The M oscillator consists of per1 and cry1 and is accel-erated by light, the E oscillator of cry2 and per2 and isdecelerated by light. Evidence in favor comes, forinstance, from Albrecht et al. (2001), who demon-strated that per1 knockout mice lost the capacity torespond with advances to a light pulse, whereas per2knockouts no longer respond with phase delays (seeDaan et al., 2001, for more details). An alternativemodel was proposed by Hastings (commentary toDaan et al., 2001). E and M could reflect per1/per2expression (M) versus cry1/cry2 expression (E). Thismodel is supported by differential peak times of thepers and crys under different photoperiods.

The transients that are observed after a phase-advancing light pulse are attributed to an immediateshift of M (the oscillator that responds with advancesto light) and a delayed shift of E, as a consequence ofcoupling forces. This explanation would also be con-sistent with Jagota et al. (2000), who demonstratedtwo peaks in multiunit activity in the SCN slice prepa-ration. One peak occurred at the onset of dawn (possi-bly the M oscillator) and responded to glutamate withan immediate advance, while the other componentdid not respond. The evening rise and morningdecline of melatonin show similar differences in their

570 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 60: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

responsiveness to light. A phase-advancing light pulseresults in an immediate advance of the melatonindecline and in a delayed advancing shift in melatoninrise (Elliot and Tamarkin, 1994; Illnerova, 1991). Wehave summarized the responses to phase-advancinglight pulses in Table 1.

Best et al. (1999) demonstrated that the phase-advancing part of the PRC is reset within a few hoursby light at CT 19. Our results demonstrate that withinone circadian cycle light presentation at CT 19 alsoresults in advances of the delay part of the PRC, atleast of the delay part between CT 12 and CT 16. If thedelay part of the PRC shows rapid resetting, similar tothe advance part of the PRC, transients in the onset ofactivity cannot readily be explained on the basis of dif-ferential shifts of the E and M component when it isassumed that E is represented by the delay and M bythe advance portion of the PRC.

Although our results do not follow directly fromthe proposed model of Daan et al. (2001), they are notnecessarily in conflict with it either. An immediateshift of the M component (per1/cry1) in response tolight at CT 19 may result in an immediate shift of the Ecomponent (per2/cry2) as a consequence of strongcoupling forces between the two. The same reasoningis applicable to the model of Hastings (2001), but nowwith a different set of genes; an immediate shift of M(per1/per2) may result in a shift of E (cry1/cry2) withinone cycle as a consequence of coupling. However, irre-spective of the set of genes that may underlie E and M,our results suggest rapid resetting of both the delayand advance portions of the PRC.

A different viewpoint arises when data from intactanimals are compared with data obtained in slicepreparations. Phase resetting of the pacemaker hasbeen studied in vitro by recording circadian rhythmsin sampled single unit activity from SCN brain slicepreparations. This activity displays a circadianrhythm, and the time of peak activity is used as aphase marker (Prosser and Gillette, 1989; Prosser,1998). SCN brain slices can be kept in good conditionand recorded from for two or three cycles. Phase shiftsare commonly induced between the first and secondcycle. Phase shift experiments in vitro have indicatedthat phase shifts in the SCN slice are immediate andstable (McArthur et al., 1991; Prosser, 1998; Watanabeet al., 2000) also when glutamate is administered atCT 19 (Ding et al., 1994). In these preparations, mostdownstream processes are eliminated because theslice is disconnected from most of its input and out-put. This indicates that transients could be the result of

secondary processes, which are downstream from thepacemaker, outside the SCN.

The latter offers an alternative explanation for theoccurrence of transients. It has been proposed thatthere might be secondary oscillatory systems (slaveoscillators) regulating the coupling of the activityrhythm with the circadian pacemaker (Pittendrighand Daan, 1976; Takamure et al., 1991; Yamazaki et al.,2000). These secondary oscillators may require morethan one circadian cycle to achieve complete re-entrainment. This hypothesis is supported by the find-ing that the velocity of re-entrainment of different cir-cadian rhythms (urinary ion secretion, adrenocorticallevels, core body temperature) after crossing severaltime zones is not the same (Klein et al., 1972;Moore-Ede et al., 1982; Van Cauter and Turek, 1986).Recently it was proposed that resetting of circadian

Watanabe et al. / TRANSIENT VERSUS STEADY-STATE PHASE SHIFTS 571

Table 1. Phase shifts induced by a phase-advancing light pulse atCT 19. The table illustrates clear asymmetry in the effect of light onbehavioral activity and melatonin (NAT) rhythms and on multiunitactivity (MUA). Immediate shifts were obtained in activity offset,melatonin offset, and morning component of MUA. No immediateshifts were obtained in activity onset, melatonin onset, and in theevening component of MUA. No asymmetry exists for the shifts inadvancing and delaying parts of the PRC, as both shift immediately.The shift of the rhythm in mPer is investigated in response to a 6-hadvance of the light-dark cycle (and not in response to a short lightpulse at CT 19) mPer1 shifts immediately in response to this shiftedcycle. It is unknown whether mPer2 rhythms are immediately resetby phase-advancing stimuli.

Response to AdvancingLight Pulse at CT 19

Immediate No ImmediateShift Shift

Behavioral activity rhythm offsetElliot and Tamarkin, 1994Meijer and De Vries, 1995 x

Behavioral activity rhythm onsetElliot and Tamarkin, 1994De Vries and Meijer, 1995 x

Melatonin (NAT) offsetElliot and Tamarkin, 1994Illnerova, 1991 x

Melatonin (NAT) onsetElliot and Tamarkin, 1994Illnerova, 1991 x

Multiunit activity peak dawn (M)Jagota et al., 2000 x

Multiunit activity peak dusk (E)Jagota et al., 2000 x

Advance part PRCBest et al., 1999 x

Delay part PRCThis study, 2001 x

mPer1 rhythmYamazaki et al., 2000 x

mPer2 rhythm ? ?

Page 61: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

time in peripheral tissue occurs via glucocorticoid sig-naling (Balsalobre et al., 2000), indicating thatglucocorticoid is one of the possible candidatesentraining secondary oscillators to the circadian pace-maker. Together the data indicate the existence of vari-ous time lags from the circadian pacemaker to eachsecondary oscillatory system.

The present experiments were not aimed to indi-cate where transients originate and leave open thepossibility that behavioral transients are attributableto downstream processes, outside the SCN. Theresults have demonstrated that at least part of thepacemaker has shifted fully and within one circadiancycle in response to light. It is possible that the firstpart of the delaying area of the PRC (before CT 12)shows compression in response to a phase-advancinglight pulse (see Pittendrigh, 1981) and that thisaccounts for transients in activity onset. This wouldlead to the conclusion that part of the delay curveshifted immediately, and part of it did not. The ques-tion of whether the pacemaker shows immediateresetting in response to light pulses should then bereformulated into a more specified question: Whichparts of the pacemaker show immediate resetting andwhich parts do not? It may appear necessary to notonly distinguish between the advance and delay areasof the PRC but even within these areas.

REFERENCES

Akiyama M, Kouzu Y, Takahashi S, Wakamatsu H, Moriya T,Maetani M, Watanabe S, Tei H, Sakaki Y, and Shibata S(1999) Inhibition of light- or glutamate-induced mPer1expression represses the phase shift into the mouse circa-dian locomotor and suprachiasmatic firing rhythms. JNeurosci 19:1115-1121.

Albrecht U, Sun ZS, Eichele G, and Lee CC (1997) Adifferen-tial response of two putative mammalian circadian regu-lators, mper1 and mper2, to light. Cell 91:1055-1064.

Albrecht U, Zheng B, Larkin D, Sun ZS, and Lee CC (2001)mPer1 and mPer2 are essential for normal resetting of thecircadian clock. J Biol Rhythms 16:100-104.

Balsalobre A, Brown S, Marcacci L, Tronche F, Kellendonk C,Reichardt HM, Schütz G, and Schibler U (2000) Resettingof circadian time in peripheral tissues by glucocorticoidsignaling. Science 289:2344-2347.

Best JD, Maywood ES, Smith KL, and Hastings MH (1999)Rapid resetting of the mammalian circadian clock. JNeurosci 19:828-835.

Crosthwaite SK, Loros JJ, and Dunlap JC (1995) Light-induced resetting of a circadian clock is mediated by arapid increase in frequency. Cell 81:1003-1012.

Daan S, Albrecht U, van der Horst GTJ, Illnerova H,Roenneberg T, Wehr TA, and Schwartz WJ (2001) Assem-

bling a clock for all seasons: Are there M and E oscillatorsin the genes? J Biol Rhythms 16:105-116.

Daan S and Pittendrigh CS (1976) A functional analysis ofcircadian pacemakers in nocturnal rodents: II. The vari-ability of phase response curves. J Comp Physiol [A]106:255-266.

Ding JM, Chen D, Weber ET, Faiman LE, Rea MA, andGillette MU (1994) Resetting the biological clock: Media-tion of nocturnal circadian shifts by glutamate and NO.Science 266:1713-1717.

Elliot JA and Pittendrigh CS (1996) Time course of hamsterclock resetting following single light pulses. Fifth meet-ing of the Society for Research on Biological Rhythms[abstr] 157:105.

Elliot JA and Tamarkin L (1994) Complex circadian regula-tion of pineal melatonin and wheelrunning in Syrianhamsters. J Comp Physiol [A] 174:437-447.

Hastings M (2001) Commentary: Modeling the molecularcalendar. J Biol Rhythms 16:117-123.

Honma K, Honma S, and Hiroshige T (1985) Responsecurve, free-running period, and activity time in circadianlocomotor rhythm of rats. Jpn J Physiol 35:643-658.

Illnerova H (1991) The suprachiasmatic nucleus and rhyth-mic pineal melatonin production. In SuprachiasmaticNucleus. The Mind’s Clock, DC Klein, RY Moore, and SMReppert, eds, pp 197-216, Oxford University Press, NewYork.

Jagota A, de la Iglesia HO, and Schwartz WJ (2000) Morningand evening circadian oscillations in the suprachiasmaticnucleus in vitro. Nat Neurosci 3:372-376.

Khammanivong A and Nelson DE (2000) Light pulses sup-press responsiveness within the mouse photic entrain-ment pathway. J Biol Rhythms 15:393-405.

Klein KE, Wegmann HM, and Hunt BI (1972) Desynchron-ization of body temperature and performance circadianrhythm as a result of outgoing and homegoingtransmeridian flight. Aerosp Med 43: 119-132.

McArthur AJ, Gillette MU, and Prosser RA(1991) Melatonindirectly resets the rat suprachiasmatic circadian clock invitro. Brain Res 565:158-161.

Meijer JH and Devries MJ (1995) Light induced phase shiftsin onset and offset of running-wheel activity in the Syrianhamster. J Biol Rhythms 10:4-16.

Meijer JH and Rietveld WJ (1989) Neurophysiology of thesuprachiasmatic circadian pacemaker in rodents. PhysiolRev 69:671-707.

Moore-Ede MC, Sultzman FM, and Fuller CA (1982) TheClocks That Time Us, Harvard University Press, Cam-bridge, MA.

Nelson DE and Takahashi JS (1999) Integration and satura-tion within the circadian photic entrainment pathway ofhamsters. Am J Physiol 277:R1351-R1361.

Pittendrigh CS (1979) Some functional aspects of circadianpacemakers. In Biological Rhythms and Their Central Mech-anism, M Suda, O Hayaishi, and H Nakagawa, eds, pp 3-15, Elsevier/North-Holland, Amsterdam.

Pittendrigh CS (1981) Circadian systems: Entrainment. InHandbook of Behavioural Neurobiology, Vol. 4, BiologicalRhythms, J Aschoff, ed, pp 95-1241, Plenum Press, NewYork.

572 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 62: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Pittendrigh CS, Bruce VG, and Kaus P (1958) On the signifi-cance of transients in daily rhythms. Proc Natl Acad SciU S A 44:965-997.

Pittendrigh CS and Daan S (1976) A functional analysis ofcircadian pacemakers in nocturnal rodents: V. Pacemakerstructure: A clock for all seasons. J Comp Physiol106:333-355.

Prosser RA (1998) In vitro circadian rhythms of the mamma-lian suprachiasmatic nuclei: Comparison of multi-unitand single-unit neuronal activity recordings. J BiolRhythms 13:30-38.

Prosser RA and Gillette MU (1989) The mammalian circa-dian clock in the suprachiasmatic nuclei is reset in vitroby cAMP. J Neurosci 9:1073-1081.

Sharma VK and Chandrashekaran MK (2000) Probing thecircadian oscillator of a mammal by two-pulse perturba-tions. Chronobiol Int 17:129-136.

Shearman LP, Zylka MJ, Weaver DR, Kolakowski LF Jr, andReppert SM (1997) Two period homologs: Circadianexpression and photic regulation in the suprachiasmaticnuclei. Neuron 19:1261-1269.

Shigeyoshi Y, Taguchi K, Yamamoto S, Takekida S, Yan L,Tei H, Moriya T, Shibata S, Loros JJ, Dunlap JC, andOkamura H (1997) Light-induced resetting of a mamma-lian circadian clock is associated with rapid induction ofthe mPer1 transcript. Cell 91:1043-1053.

Takamure M, Murakami N, Takahashi K, Kuroda H, andEtoh T (1991) Rapid reentrainment of the circadian clockitself, but not the measurable activity rhythm, to a newlight-dark cycle in the rat. Physiol Behav 50:443-449.

Van Cauter E and Turek FW (1986) Depression: Adisorder oftime keeping? Perspect Biol Med 29:510-519.

Watanabe K, Vanacek J, and Yamaoka S (2000) In vitroentrainment of the circadian rhythm of vasopressin-releasing cell in suprachiasmatic nucleus by vasoactiveintestinal polypeptide. Brain Res 877:361-366.

Yamazaki S, Numano R, Abe M, Hida A, Takahashi R, UedaM, Block GD, Sakaki Y, Menaker M and Tei H (2000)Resetting central and peripheral circadian oscillators intransgenic rats. Science 288:682-685.

Watanabe et al. / TRANSIENT VERSUS STEADY-STATE PHASE SHIFTS 573

Page 63: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

JOURNAL OF BIOLOGICAL RHYTHMS / December 2001Foà and Bertolucci / TEMPERATURE AND BIMODAL PATTERN IN LIZARDS

Temperature Cycles Induce a BimodalActivity Pattern in Ruin Lizards: Masking orClock-Controlled Event? A Seasonal Problem

Augusto Foà1 and Cristiano BertolucciDipartimento di Biologia, Università di Ferrara, via L. Borsari 46 I-44100, Ferrara, Italy

Abstract The daily locomotor activity pattern of Ruin lizards in the field is mainlyunimodal, except for summer months when soil temperatures exceed 40 °C to42 °C around midday. In such a situation, lizards reduce their locomotor activityaround midday to avoid overheating, and thus their daily activity patternbecomes bimodal. The bimodal pattern expressed in the field is usually retainedin the free-running rhythm under constant temperature and DD for a couple ofweeks, after which the bimodal pattern changes into a unimodal pattern. In thepresent study, the authors examined whether 24-h temperature cycles (TCs)would change lizard activity from a unimodal to a bimodal pattern. Administra-tion of TCs to unimodal lizards free-running in DD is able to entrain locomotorrhythms and to induce a bimodal pattern both in summer and autumn-winter.There are, however, striking seasonal differences in the effectiveness with whichTCs achieve bimodality: (a) Numbers of lizards rendered bimodal are signifi-cantly higher in summer than in autumn-winter; (b) TCs require less time toachieve bimodality in summer than autumn-winter; (c) bimodality is retainedas an aftereffect in the postentrainment free-run in summer, but not in autumn-winter; (d) TCs change activity duration in summer, but not in autumn-winter.All this demonstrates the existence of seasonal changes in responsiveness of thecircadian oscillators controlling activity to the external factors inducingbimodality. Oscillators’ responsiveness is high in summer, when bimodality isthe survival strategy of Ruin lizards to avoid overheating around midday inopen fields, and low in autumn-winter, when bimodality has no recognizableadaptive significance.

Key words circadian rhythms, locomotor activity, entrainment, temperature cycles,bimodal activity pattern, lizards

Seasonal changes in the daily pattern of locomotoractivity in the field were reported in most diurnalLacertid lizards from southern Europe (Bowker, 1986;Foà et al., 1992; Henle, 1988; Pough and Busack, 1978;Tosini et al., 1992; Van Damme et al., 1990). Generally,the daily pattern of locomotor activity changes from

unimodal in spring (with only one activity peak perday) to bimodal in summer (with two well-separatedactivity peaks per day), becoming unimodal again inautumn. Since lizards are ectotherms, which can beactive only within a limited range of body tempera-tures, seasonal changes in locomotor patterns have

574

1. To whom all correspondence should be addressed: Dipartimento di Biologia, Università di Ferrara, via L. Borsari 46 I-44100,Ferrara, Italy; e-mail: [email protected].

JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 574-584© 2001 Sage Publications

Page 64: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

been regarded more or less explicitly as a direct behav-ioral response of these animals to related changes insolar radiation and ambient temperature (Avery, 1980;Ouboter, 1981). Recent studies in our model animal,the Ruin lizard (Podarcis sicula), showed that a greatdeal of these activity changes are controlled by endog-enous temporal programs (Bertolucci et al., 1999; Foàet al., 1994). Short-term experiments (20-30 days), inwhich Ruin lizards collected in different months weretested immediately after capture under constant tem-perature (29 °C) and darkness (DD), showed that theactivity pattern typical of each season (bimodal/unimodal) is retained in the lizard circadian locomo-tor rhythm (Foà et al., 1994). Furthermore, while thebimodal locomotor pattern expressed by the lizards inconstant conditions between June and August is typi-cally associated with a short free-running period (τ)and a long circadian activity (α), the unimodal loco-motor pattern expressed in the remaining months istypically associated with long τ and short α (Foà et al.,1994). To test whether such seasonal changes in circa-dian locomotor rhythms were driven by a circannualclock, the locomotor activity of Ruin lizards wasrecorded in long-term experiments (over 12-15months) under constant temperature (29 °C) and DD.The results demonstrated the existence of circannualcycles of both τ and α and, at the same time, theabsence of a circannual cycle of activity pattern frombimodal to unimodal and vice versa (Bertolucci et al.,1999). The longest τ along its circannual cycle wasassociated with short α, and the shortest τ in the samecycle with long α, so that the pattern of mutual associ-ation between τ and α was found to be the same as inshort-term experiments (Bertolucci et al., 1999; Foàet al., 1994). Most lizards, however, stayed unimodalall the time, showing that seasonally changing envi-ronmental factors (such as, for instance, photo-thermoperiodic conditions) are involved in the induc-tion of a bimodal activity pattern (Bertolucci et al.,1999). In other reptiles, such as the Namib desert dunelizard (Aporosaura anchietae) and the garter snake(Thamnophis radix), the locomotor activity pattern wasshown to be switched between unimodal and bimodalpatterns, respectively, by lowering or raising the ambi-ent temperature (Heckrotte, 1962, 1975; Holm, 1973;Underwood, 1992). This suggests that the expectedseasonal changes of locomotor activity pattern (fromunimodal to bimodal and vice versa) can be achievedin lizards by means of appropriate manipulations ofambient temperature levels and/or administration of

temperature cycles in the laboratory. Hoffmann (1968)showed that 24-h temperature cycles (TCs) of lowamplitude (0.9-3 °C) are capable of entraining circa-dian locomotor rhythms of Ruin lizards kept in con-stant lighting conditions.

The present study carried out completely in DDexamined whether the administration of 24-h TCs iscapable of changing a unimodal activity pattern into abimodal activity pattern. When Ruin lizards weretested in summer, TCs were found to induce a bimodalpattern. To establish whether effectiveness of TCs inachieving bimodality changes depending on season,we compared the locomotor behavior of lizards col-lected and exposed to the same TCs in different sea-sons. We expected effectiveness to be maximal in thesummer, as Ruin lizards usually express a bimodalactivity pattern in summer but not at other times of theyear (Foà et al., 1992, 1994). Studies focused on the roleof the pineal and other clock components in establish-ing and/or maintaining a bimodal activity pattern inRuin lizards have been reported previously (Innocentiet al., 1994, 1996; Tosini et al., 2001).

MATERIALS AND METHODS

Animals andLocomotor Recording

Ruin lizards, Podarcis sicula campestris De Betta 1857(adult males only, 6.5-8 cm snout-vent length) fromthe area of Ferrara (Italy) were used. Lizards were col-lected in the field in groups of 8 to 14 individuals inJune and November 1999 and June 2000. After cap-ture, each seasonal group of lizards was carried to thelab and immediately put into individual tilt-cages (30× 15 × 11 cm) for locomotor recording. Tilt-cages wereplaced inside thermally programmable environmen-tal chambers and connected to a computer-based dataacquisition system (DataQuest III, MiniMitter, Sun-river, OR, USA) for monitoring locomotor activity.Food (Tenebrio molitor larvae) and water were suppliedtwice a week. During experiments, lizards were keptunder DD and either constant temperature (30 ± 0.2 °C)or 14:10-h TCs. Experiments used three different TCs.TC1: 30:27 °C (30 °C from 0800 to 2200 h; 27 °C from 2200to 0800 h); TC2: 30:28.3 °C (30 °C from 0800 to 2200 h;28.3 °C from 2200 to 0800 h); and TC3: 30:28.3 °C (30 °Cfrom 1400 to 0400 h; 28.3 °C from 0400 to 1400 h).

Foà and Bertolucci / TEMPERATURE AND BIMODAL PATTERN IN LIZARDS 575

Page 65: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Experimental Design

Summer 1999 (June 4 to July 27):Pilot Experiment

Lizards (n = 8) were allowed to free-run in DD andconstant temperature (29 °C) for 3 weeks and thenwere subjected to the TC1 for 25 days.

Autumn-Winter 1999 to 2000(November 17 to March 8)

Lizards (n = 14) were allowed to free-run in DD andconstant temperature (29 °C) for 2 weeks and thenwere subjected to three TCs: TC1 for 28 days, TC2 for 22days, and TC3 for 22 days. After these periods of time,locomotor activity in constant temperature (29 °C)was recorded for a further 2 to 3 weeks.

Summer 2000 (June 19 to September 9)

Lizards (n = 14) were allowed to free-run in DD andconstant temperature (29 °C) for 2 weeks and thenwere subjected to two TCs: TC1 for 23 days and TC2 for17 days. After these periods of time, locomotor activ-ity in constant temperature (29 °C) was recorded for afurther 4 weeks.

Data Evaluations

For locomotor rhythms free-running in constanttemperature and DD, τ and α were estimated by theeye-fitting method (Pittendrigh and Daan, 1976a). τwas also measured by means of χ2 periodogram analy-sis (Sokolove and Bushell, 1978). Bimodality andunimodality of the locomotor pattern were establishedby visual inspection: The bimodal pattern is character-ized by the existence of two daily peaks of locomotoractivity regularly repeated in subsequent days, whilethe unimodal pattern is characterized by the existenceof one peak of activity each day. In presence of abimodal activity pattern, either duration of circadian(α) or duration of entrained activity includes durationof the pause between the two daily activity peaks.

RESULTS

Before administration of TCs, lizards were free-running in constant temperature and DD. The meanvalue of τ in summer was significantly shorter than the

mean value of τ in autumn-winter (p < 0.05), and themean value of α in summer was significantly longerthan the mean value of α in autumn-winter (p < 0.001)(Fig. 1).

Locomotor rhythms of all lizards kept in DDentrained to the TC1 of 3 °C in amplitude (Figs. 2-5).Entrainment continued after reduction of the TCamplitude to 1.7 °C (TC2 and TC3; Figs. 2-3, 5) in allcases. The entrained activity phase always fell withinthe 30 °C portion of the TC. When, in the final part ofexperiments, lizards were exposed again to constanttemperature, activity onsets in the postentrainmentfree-running rhythm always extrapolated back to thetime of activity onsets during entrainment. In winter, a6-h shift from TC2 to TC3 induced shifts of the activityonsets, which resulted, after several transient cycles,in entrainment to the new schedule (Fig. 5). Hence,14:10-h TCs actually entrained the activity rhythmand did not merely cause masking of the underlyingoscillation. No seasonal differences in all these aspectsof entrainment were found. In most lizards, activityonsets delayed about 1 h the daily onsets of high(30 °C) temperature phase of TC. No differences in ψeither among TCs or among seasons were found.

Administration of the TC1 to lizards expressing aunimodal circadian locomotor activity pattern underconstant temperature and DD was capable of induc-ing the appearance of a bimodal pattern in all seasons(Table 1, Figs. 2-3, 5). Bimodality always persistedafter reduction of amplitude of the TC from 3 °C to 1.7°C (TC2 and TC3 cycle; representative examples inFigs. 2-3, 5). However, the percentage of lizards ren-dered bimodal by TC1 is significantly higher in sum-mer than in autumn-winter (Fisher exact test, p < 0.04;Fig. 6A). Furthermore, the time elapsed betweenadministration of TC1 and appearance of bimodality issignificantly shorter in summer than in autumn-win-ter (Kruskal-Wallis one-way ANOVA, H = 4.47, p <0.04; Fig. 6B). Administration of TC1 also increases sig-nificantly the duration of daily activity in summer(paired Student t test, p < 0.01) but leaves the durationof daily activity unchanged in autumn-winter (pairedStudent t test, p > 0.10). As expected, from the verybeginning of locomotor recording in summer, severallizards already expressed a bimodal pattern in con-stant temperature. In all cases, the administration ofTC1 lengthened the rest interval between the two dailypeaks of activity that characterize the bimodal statefrom 2 to 3.1 h, enhancing in that way bimodality(paired Student t test, p < 0.002; representative exam-ples in Fig. 4).

576 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 66: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Administration of constant temperature led all liz-ards to free-run (Figs. 3, 5). While in autumn-winter nobimodal lizard maintains bimodality after release inconstant temperature, in summer several lizards do so(Table 1 and Figs. 3, 5). The seasonal difference is sta-tistically significant (p < 0.006, Fisher exact test). Fur-thermore, the time elapsed between administration ofconstant temperature and reappearance of unimodalityis significantly shorter in autumn-winter than in sum-mer (Kruskal-Wallis one-way ANOVA, H = 7.17, p <0.02; Fig. 6B). Several summer lizards were stillbimodal 3 to 4 weeks after release in constant tempera-ture (Table 1 and Fig. 3).

DISCUSSION

Behavioral observations of Ruin lizards in their nat-ural environment showed that the daily locomotorpattern of focal animals in the field is mainlyunimodal, except for those summer months in whichsoil temperatures exceed 40 to 42 °C around the mid-dle of day. In such situations, lizards reduce their loco-motor activity dramatically around midday, retreat-ing into shade or burrows to avoid overheating, andthus exhibiting the markedly bimodal activity patternso typical of summer (Foà et al., 1992, 1994; Tosiniet al., 1992). Once carried to the laboratory and placedin constant temperature (29 °C) and DD, summer liz-ards mainly retained their endogenous, free-running

rhythm under constant conditions, the bimodal pat-tern previously expressed in the field (Foà at al., 1994).This demonstrates that bimodality is not merely adirect behavioral reaction of these ectotherm animalsto the extremely high levels of soil temperaturesaround midday in summer but is a season-dependentstate of the circadian pacemaker that has evolved as anadaptation to high temperatures predictably occur-ring at that time of day. Long-term experiments fur-ther revealed that the retained bimodal pattern disap-pears—that is, the pattern becomes unimodal—in thecourse of the first 2 months of locomotor recordingunder constant temperature (29 °C) and DD, andbimodality never reappears (Bertolucci et al., 1999).Hence, bimodality of Ruin lizards typically shows allrecognized properties of aftereffects on the circadianpacemaker: (a) Its appearance in constant conditionsis dictated by previous exposure to specific environ-mental stimuli, such as, for instance, summerphoto-thermoperiodic conditions; (b) once estab-lished, it persists for several weeks in constant condi-tions, after which, (c) it decays to a different state:unimodality. The present study examined whetherchanges in ambient temperature—which were admin-istered in the form of 24-h TCs—were the specificenvironmental stimuli capable of inducing bimodalactivity patterns at the expected time of the year. Firstof all, our results show that circadian locomotorrhythms of Ruin lizards become entrained to 24-h TCsof low (3-1.7 °C) amplitude. This confirms the results

Foà and Bertolucci / TEMPERATURE AND BIMODAL PATTERN IN LIZARDS 577

τ

24.1

24.4

24.7

Autumn-Winter Summer

hour

s

α

7

10

13

Autumn-Winter Summer

hour

s

Figure 1. Estimates of τ and α of locomotor rhythms of Ruin lizards free-running in constant temperature (29 °C) and DD before startingtemperature cycle administration. τ in autumn-winter was significantly longer than in summer, and α in autumn-winter was significantlyshorter than in summer.

Page 67: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

of the pioneering work done by Hoffmann in Ruin liz-ards, which for the first time indicated that low-ampli-tude TCs are capable of entraining circadian behav-

ioral rhythms in ectotherm vertebrates (Hoffmann,1968). More important, administration of 24-h TCs inthe laboratory to unimodal lizards collected in the

578 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 2. Locomotor records of lizards free-running in constant temperature and DD, which were subjected to a 24-h temperature cycle(TC) 30:27 °C, and then to a 24-h TC 30:28.3 °C. Each horizontal line is a record of 1 day’s activity, and consecutive days are mounted onebelow the other. Rectangles encompass the 30 °C phase during the whole period of time of TCs administration. Arrowhead indicates theday on which the amplitude of TC was reduced from 3 °C to 1.7 °C. Records are representative examples of the fact that TC administration insummer almost immediately induces a change from unimodal activity pattern to bimodal activity pattern, while TC administration inautumn-winter does not change the pattern in 50% of the lizards.

Page 68: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Foà and Bertolucci / TEMPERATURE AND BIMODAL PATTERN IN LIZARDS 579

Figure 3. Behavioral effects of temperature cycles (TCs) in summer. Brackets (on the right) delimit sections of records where the activitypattern was bimodal. Records are representative examples of the fact that TC administration entrained circadian locomotor rhythms andinduced a fast change of locomotor activity pattern from unimodal to bimodal. Furthermore, after final release in constant temperature, liz-ards maintained bimodality. Further information in Figure 2.

Page 69: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

field during late spring to early summer was found toinduce rapid appearance of a bimodal activity patternin most animals (Table 1 and Figs. 2-3).

As expected, in summer several lizards werealready bimodal in constant temperature before start-ing with TCs administration (example in Fig. 4). TCsfurther enhanced bimodality by lengthening thepause between the two daily peaks of the bimodalactivity pattern of these lizards. Most summer lizardswere still bimodal 3 to 4 weeks after cessation of theTCs (after final release in constant temperature), andafter this period of time, bimodality disappeared inabout 50% of the animals. These aspects of the dataconfirm our assumption that bimodality is retained asan aftereffect in the postentrainment free-runningrhythm in constant conditions. To our surprise, 24-hTCs were capable of inducing bimodality even inautumn-winter, seasons in which changing fromunimodal to bimodal activity pattern is not of appar-ent adaptive significance (Foà et al., 1994). It must be

pointed out, however, that all lizards that becamebimodal in autumn-winter returned to beingunimodal immediately after cessation of TCs (Fig. 5).

Thus, differently from summer, in autumn-winterbimodality is not retained in the postentrainmentfree-running rhythm in constant conditions: It is notan aftereffect. Instead of resulting from TCs affectingthe circadian oscillators that control activity, the tran-sition from a unimodal to a bimodal pattern inautumn-winter seems to result from TCs directlyaffecting activity, therefore bypassing the circadianclock. In other words, bimodality achieved inautumn-winter looks like a phenomenon of masking(Underwood, 1992, p. 249). For the sake of clarity, itseems useful to underline that the case of maskingproposed here exclusively concerns achievement ofbimodality in autumn-winter. In fact, “true” entrain-ment (and not masking) of circadian locomotorrhythms of lizards to the administered 24-h TCs hasoccurred in all seasons, as postentrainment free-runsin constant temperature have verified.

Changes in locomotor pattern in response to ambi-ent temperature manipulations in the laboratory havebeen reported previously in other reptiles. Forinstance, in the Namib desert dune lizard (Aporosauraanchietae) kept under constant conditions and the gar-ter snake (Thamnophis radix) exposed to an LD cycle,the locomotor activity pattern was shown to beswitched between unimodal and bimodal patterns,respectively, by lowering or raising the ambient tem-perature. It is unclear, however, whether the reportedeffects were season-dependent (Heckrotte, 1962, 1975;Holm, 1973). Noteworthy, Underwood (1992) pro-posed that each activity peak of the bimodal pattern oflizards and snakes may be the overt expression of oneof the two (or sets of) circadian oscillators described inPittendrigh and Daan’s (1976b) model. Underwoodfurther stated that since a rise in ambient temperatureswitches activity from a unimodal to a bimodal pat-tern and increases the duration of daily activity, tem-perature as well as photoperiod may be able to changephase relationship between the two oscillators con-trolling activity. Because the model proposes that thetime interval between activity onset and activity offseteach day would be a measure of the phase relationshipbetween the two oscillators, changes in that phaserelationship would be expected to cause a change induration of activity. According to this view, it is impor-tant that in Ruin lizards the bimodal pattern achievedin response to TCs is associated with a prominentincrease in duration of daily activity in summer—thus

580 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 4. Representative example of summer lizards exhibiting abimodal activity pattern already before administration of temper-ature cycles (TCs). Dotted lines delimit the pause between the twodaily peaks of activity. Record is a representative example of thefact that TC administration in summer-bimodal lizards length-ened the interruption between the two daily peaks of activity thatcharacterize the bimodal state.

Page 70: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Foà and Bertolucci / TEMPERATURE AND BIMODAL PATTERN IN LIZARDS 581

Figure 5. Effects of temperature cycles (TCs) in autumn-winter. Records are representative examples of the fact that TC administrationentrained circadian locomotor rhythms. However, it took several days for the TC to induce a bimodal pattern. Furthermore, differentlyfrom summer, autumn-winter lizards returned unimodal immediately after release in constant temperature. Other information in Figure 3.Finally, 50% of autumn-winter lizards remained unimodal after TC administration, as it is shown in Figure 2.

Page 71: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

potentially reflecting a change in phase relationshipbetween circadian oscillators controlling activity—while the bimodal pattern achieved in autumn-winteris not associated with a change in activity duration.

Furthermore, there are striking seasonal differ-ences in the effectiveness with which TCs achievebimodality: (a) The percentage of lizards that TCshave rendered bimodal is significantly higher in sum-mer than in autumn-winter; (b) the time required forthe first administered TC to achieve a bimodal patternis significantly less in summer than autumn-winter.According to Pittendrigh and Daan’s (1976b) model,such differences in effectiveness may depend on sea-sonal differences in the mutual phase relationshipbetween the two oscillators, that is reflected in the sea-sonally different associated values of τ and α of the

overt rhythm. With regard to the present results, suchseasonal differences are expressed in the free-runningrhythm before TC administration: in summer a short τassociated with a long α, and in autumn-winter, a longτ associated with a short α (Fig. 1). Because previousexperiments in constant conditions made it clear that ashort τ together with a long α typically brings about abimodal pattern, the summer effectiveness of TCs inachieving bimodality shown here is likely to dependon the expression in the free-running rhythm of a shortτ together with a long α, facilitating a change in pat-tern from unimodal to bimodal as soon as Ruin lizardsperceived the TC (Foà et al., 1994). The long τ togetherwith the short α of autumn-winter reflect a phase rela-tionship between constituent oscillators that presum-ably hampers a change in pattern from unimodal tobimodal, thus reducing autumn-winter effectivenessof TCs in achieving bimodality. When, in spite of allthat, TCs administered in autumn-winter were suc-cessful in changing the pattern from unimodal tobimodal, they clearly did it by bypassing the circadianclock, as bimodality did not entail a change in activityduration and, moreover, was not retained as an after-effect in the postentrainment free-run.

Unlike bimodality, effectiveness of 24-h TCs inentrainment of locomotor rhythms does not changewith season. This was somehow expected. In fact,Ruin lizards are diurnal, ecthoterm animals whoselocomotor activity is necessarily limited by tempera-ture. In such a situation, the circadian clock of lizardshas to entrain to the TC of the external day, to force thelocomotor activity it drives into the range of the ambi-ent temperatures that are most suitable for locomotorperformances, thus optimizing daily time-budget forbiologically significant behaviors through the entireyear. Again, no seasonal differences in any aspect ofentrainment were found. Both summer and autumn-winter entrained activity fell in the 30 °C phase and(entrained) rest in the 27 °C (or 28.3 °C) phase of TCs.This confirms previous work indicating that Ruin liz-ards that are held on a thermal gradient in DD sponta-neously select warm temperature zones of the gradi-ent during activity and cooler temperature zonesduring rest (Innocenti et al., 1993). On the other hand,field studies in Ruin lizards showed the existence ofseasonal differences in body temperatures recordedduring activity phase, with a mean temperature of29 °C in autumn-winter and a mean of 32.7 °C in latespring–summer (Tosini et al., 1992). It looks like thetemperature of 30 °C we chose in the present experi-

582 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

0

5

10

15

20

Autumn-Winter Summer Autumn-Winter Summer

n=7 n=12 n=7 n =7

from unimodal to bimodal from bimodal to unimodalB

0

25

50

75

100

Autumn-Winter Summer

%

A

days

Figure 6. Seasonal differences in effectiveness of temperaturecycles (TCs) in the induction of a bimodal activity pattern. (A) Per-centage of lizards rendered bimodal by TCs is significantlyhigher in summer than in autumn-winter. (B) Number of daysbetween administration of TCs and appearance of bimodality(left), and between final release in constant temperature andappearance of unimodality (right). Appearance of a bimodal pat-tern is faster in summer than in autumn-winter. Furthermore, thetime elapsed between final release in constant temperature andreappearance of unimodality is significantly shorter inautumn-winter than in summer. There are also summer lizardsthat remained bimodal at the end of the experiments (Table 1).

Page 72: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

ments is a reasonable compromise between those mea-sured in the field and therefore suitable for activity ineach season, while 27 °C seems less suitable for activity.

CONCLUSIONS

Although long-term records in constant conditionsdid not find circannual changes between bimodal andunimodal patterns, the same experiments did revealthe existence of circannual changes of both τ and α oflocomotor rhythms with, moreover, some phases ofthe circannual cycle characterized by long τ associatedwith short α (potentially conducive to unimodality)and other phases characterized by short τ associatedwith long α (potentially conducive to bimodality)(Bertolucci et al., 1999). The circannual data made itclear, on one hand, that environmental factors are nec-essary to obtain a change from a unimodal to abimodal pattern, and, on the other hand, that there aremonths in which the state of the system, expressed inthe association between short τ and long α of its overtrhythm, facilitates achievement of a bimodal pattern.The present results show that a change in ambienttemperature—in the form of a 24-h TC—is the appro-priate stimulus to induce bimodality, and they con-firm the existence of pronounced changes in respon-siveness of the circadian oscillators controllingactivity to the external factors inducing bimodality,with high responsiveness in summer and low respon-siveness in autumn-winter. The data further suggestthe possibility that a circannual rhythm in responsive-ness of the circadian system to the bimodality-stimu-lating effects of TCs parallel the previously demon-strated circannual rhythms of τ and α or, alternatively,the circannual rhythm in responsiveness may bemerely a consequence of those of τ and α.

ACKNOWLEDGMENTS

This work was supported by research grants of theUniversità di Ferrara and of the Italian Ministero dell’Università e della Ricerca Scientifica e Tecnologica(COFIN 1999).

REFERENCES

Avery RA (1980) Ecophysiology and behaviour of lacertidlizards—towards a synoptic model. Proc Euro HerpSymp CWLP, pp 71-73.

Bertolucci C, Leorati M, Innocenti A, and Foà A (1999)Circannual variations of lizard circadian activity rhythmsin constant darkness. Behav Ecol Sociobiol 46:200-209.

Bowker RG (1986) Patterns of thermoregulation in Podarcishispanica (Lacertilia: Lacertidae). In Studies in Herpetology,Z Rocek, ed, pp 621-626, Proc Euro Herp Meeting,Prague.

Foà A, Minutini L, and Innocenti A (1992) Melatonin: A cou-pling device between oscillators in the circadian systemof the ruin lizard Podarcis sicula. Comp Biochem Physiol103A:719-723.

Foà A, Monteforti G, Minutini L, Innocenti A, Quaglieri C,and Flamini M (1994) Seasonal changes of locomotoractivity patterns in ruin lizards Podarcis sicula: I. Endoge-nous control by the circadian system. Behav EcolSociobiol 34:227-274.

Heckrotte C (1962) The effect of the environmental factors inthe locomotory activity of the plains garter snake(Thammophis radix radix). Anim Behav 10:193-207.

Heckrotte C (1975) Temperature and light effects on the cir-cadian rhythm and locomotor activity of the plains gartersnake (Thammophis radix radix). J Interdiscipl Cycle Res 6:279-290.

Henle K (1988) Dynamics and ecology of three Yugoslavianpopulations of the Italian wall lizard (Podarcis siculacampestris de Betta) (Reptilia, Lacertidae). Zool Anz 220:33-48.

Hoffmann K (1968) Synchronisation der circadianenAktivitaetsperiodik von Eidechsen durch Temperatur-

Foà and Bertolucci / TEMPERATURE AND BIMODAL PATTERN IN LIZARDS 583

Table 1. Summary of the results. All numbers of unimodal versus bimodal lizards in each stage of experiments in different seasons.

T Cost Temperature Cycles T Cost

n Unimodal Bimodal Unimodal Bimodal Unimodal Bimodal

Summer 1999 8 6 2* 1 7 — —Autumn-Winter 1999-2000 14 14 0 7 7 14 0Summer 2000 14 9 5* 1 13 7 7

*Summer lizards that already expressed a bimodal pattern in constant temperature.

Page 73: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

zyclen verschiedener Amplitude. Z vergl Physiol 58:225-228.

Holm E (1973) The influence of constant temperatures uponthe circadian rhythm of the Namib desert dune lizardAporosaura anchietae Bocage. Madoqua 2:33-41.

Innocenti A, Bertolucci C, Minutini L, and Foà A (1996) Sea-sonal variations of pineal involvement in the circadianorganization of ruin lizards Podarcis sicula. J Exp Biol199:1189-1194.

Innocenti A, Minutini L, and Foà A(1993) The pineal and cir-cadian rhythms of temperature selection and locomotionin lizards. Physiol Behav 53:911-915.

Innocenti A, Minutini L, and Foà A (1994) Seasonal changesof locomotor activity patterns in ruin lizards Podarcissicula: II. Involvement of the pineal. Behav Ecol Sociobiol35:27-32.

Ouboter PE (1981) The ecology of the island-lizard Podarcissicula salfii: Correlation of microdistribution with vegeta-tion coverage, thermal environment and food-size.Amphibia-Reptilia 2:243-257.

Pittendrigh CS and Daan S (1976a) A functional analysis ofcircadian pacemaker in nocturnal rodents: I. The stabilityand lability of spontaneous frequency. J Comp Physiol106:223-252.

Pittendrigh CS and Daan S (1976b) A functional analysis ofcircadian pacemaker in nocturnal rodents: V. Pacemakersstructure: A clock for all seasons. J Comp Physiol106:333-355.

Pough FH and Busack SD (1978) Metabolism and activity ofthe spanish fringe-toed lizard (Lacertidae:Acanthodactylus erythrurus). J Thermal Biol 3:203-205.

Sokolove PG and Bushell WN (1978) The chi squareperiodogram: Its utility for analysis of circadian rhythms.J Theor Biol 72:131-160.

Tosini G, Bertolucci C, and Foà A (2001) The circadian sys-tem of reptiles: A true multioscillatory and multiphoto-receptive system. Behav Physiol 72(4):461-471.

Tosini G, Foà A, and Avery RA (1992) Body temperaturesand exposure to sunshine of ruin lizards Podarcis sicula incentral Italy. Amphibia-Reptilia 13:169-175.

Underwood H (1992) Endogenous rhythms. In Biology ofReptilia, C Gans, ed, Vol 18, pp 229-229, University of Chi-cago Press, Chicago, IL.

Van Damme R, Bauwens D, Castilla AM, and Verheyen RF(1990) Comparative thermal ecology of the sympatric liz-ards Podarcis tiliguerta and Podarcis sicula. Acta Oecol11:503-512.

584 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 74: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

JOURNAL OF BIOLOGICAL RHYTHMS / December 2001Mrosovsky et al. / IRRADIANCE DETECTION

LETTER

Persistence of Masking Responses to Lightin Mice Lacking Rods and Cones

N. Mrosovsky,*,1 Robert J. Lucas,† and Russell G. Foster†

*Departments of Zoology, Physiology and Psychology, University of Toronto,Ontario, Canada M5S 3G5, †Department of Integrative and Molecular Neuroscience,Imperial College School of Medicine, Charing Cross Hospital, London, UK W6 8RF

Key words cones, irradiance, masking, phase shifting, retinal degeneration, rods

Overt rhythms of locomotor behavior are a result oftwo complementary processes. The first mechanism isthe synchronization of an endogenous clock, which inturn directs the animal to be active in the day or thenight; this is called entrainment. The second mecha-nism involves an acute response to light, which inhib-its or promotes activity, depending on whether thespecies is nocturnal or diurnal; this is called masking(Aschoff, 1960; review in Mrosovsky, 1999). The factthat masking can occur in hamsters with lesions of theSCN (Redlin and Mrosovsky, 1999) and in mice lack-ing cryptochromes (van der Horst et al., 1999;Vitaterna et al., 1999), that is, in animals lacking anypersistent circadian rhythm to be entrained, is furtherevidence that entrainment and masking are behavior-ally and physiologically different responses to light.Although masking may be of considerable impor-tance in influencing the amplitude of overt rhythms, ithas received much less study than has entrainment.Indeed, the main concern of rhythms researchers hasbeen to exclude masking as much as possible fromtheir experiments.

At the receptor level at least, however, masking andentrainment may have considerable commonalities,in that they are both spared in retinally degenerate(rd/rd) mutant mice (Foster et al., 1991; Mrosovsky,1994). Eventually, rd/rd mice come to lose all their rods,and many cones degenerate secondarily (Carter-Dawson et al., 1978; Foster et al., 1991). Nevertheless, afew cones remain, and it is possible that such residualcones might be enough to mediate responses to light.

This has been ruled out for phase shifting, which hasbeen shown to persist even in mice genetically modi-fied to lack both rods and cones (Freedman et al.,1999). The current experiment was undertaken todetermine if masking was also spared in such rodlessconeless mice.

METHODS

C3H/He wild-type and rodless coneless (rd/rd cl)mice were bred in London as previously described(Lucas et al., 1999) and sent to Toronto for testing.Assessment of inhibition of locomotion by light fol-lowed previous procedures (Mrosovsky et al., 1999).Tests began when the mice were approximately 4 monthsold. Male mice of both genotypes (wild-types n = 8,rd/rd cl n = 5) were housed individually in cages 44 × 23× 20 cm fitted with running wheels (17.5 cm d). Revo-lutions were monitored with Dataquest III hardwareand software (Sunriver, OR, USA). Prior to the start ofthe masking tests, the animals were entrained to a 16:8h light-dark (LD) cycle for 17 days; the entraining lightprovided approximately 1300 lux in the cages, as mea-sured with a ISO-TECH ILM350 meter.

To examine the acute effects of light on wheel run-ning, an additional set of lights were used. These werefluorescent tubes (Sylvania Octron 32 watt 4100 K) fit-ted with a broad-spectrum green filter (RoscoSupergel filter #89, maximum transmission around520 nm, half bandwidth approximately 60 nm; Rosco,

585

1. To whom all correspondence should be addressed: Department of Zoology, 25 Harbord Street, University of Toronto,Ontario, Canada M5S 3G5; e-mail: [email protected].

JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 585-588© 2001 Sage Publications

Page 75: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Stamford, CT., U.S.A.). Irradiance, as measured with aHagner E2X luxmeter, was ca 120 lux in the cages below.To reduce this illumination for subsequent tests, neutraldensity filters (Rosco Cinegel), calibrated in photo-graphic stops, were added as needed; illumination atthree, six, and nine stops, respectively, was approxi-mately 12, 1.5, and 0.25 lux. Tests with 1-h light pulses,starting 2 h after dark onset, were spaced not less than3 days apart and were conducted in the followingorder for all animals: 0, 3, 6, 9, 7.5, 4.5, 1.5, 1, 2, 0.5stops. The number of wheel revolutions made duringa light pulse was expressed as a percentage of thenumber made by the same animal during the samehour on the previous day when there was no light. Foradditional details, see Mrosovsky et al. (1999).

RESULTS

Both wild-type and rd/rd cl mice inhibited theirwheel running on exposure to the broad-spectrumgreen light presented in the early night (Fig. 1). At thehighest irradiance studied here, wheel running wasreduced to about 10% of normal levels in both geno-types (means ± SEMs: 12.8 ± 4.6% wild types, 10.7 ±9.3% rd/rd cl). In both genotypes, the effects wereirradiance dependent but there were no significantdifferences between the genotypes (two-way ANOVAp > 0.1).

DISCUSSION

It is evident that, as well as phase shifting, rodlessconeless mice retain an acute masking response tolight and that no impairment of this acute response ispresent.

Both masking and entrainment serve the functionof confining an animal’s activity to a daytime or night-time niche. Both masking and entrainment requireonly detection of overall illumination, not ofspatio-temporal differences in light (Mrosovsky,1994). Indeed, to determine whether it is night or day,a system that integrates illumination over space andtime is desirable. For it to be day, it should be light inall directions and this input should be maintained.Both masking and entrainment are spared in micelacking rods and cones and so presumably depend onsome novel photoreceptor type elsewhere in the eye.

Given these similarities, it would be most parsimo-nious to suppose there is a single novel receptor type

that projects to various brain areas involved in differ-ent responses. If there is a single receptor type for thesedifferent responses to light, whether or not it had beencalled a circadian receptor in the past, it should be con-ceptualized more as a general irradiance detector. Ithas been demonstrated that besides unimpairedphase shifting and masking, rd/rd mice spend moretime in the dark side of a differentially illuminated box(Mrosovsky and Hampton, 1997). It has long beenknown that they retain pupillary responses (Keeler,1927) and suppression of melatonin by light pulses(Goto and Ebihara, 1990). Recently it has been shownthat pupillary contraction and melatonin suppressionare also present in rd/rd cl mice (Lucas et al., 1999,2001). Thus, there are a variety of responses to lightthat do not require rods or cones.

Of course it remains possible that there are severaldifferent receptor types subserving phase shifting,masking, dark preferences, melatonin suppression,and pupillary responses, perhaps with some func-tional redundancy (Selby et al., 2000). While beingopen-minded on this possibility, it seems more likelythat photoreception for resetting circadian rhythms is

586 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 1. Photic inhibition of locomotor activity in wild-type androdless coneless (rd/rd cl) mice. Top: actograms showing wheelrunning over 2 days. The number of revolutions per 10 min areplotted in 15 quantiles on the y axis, with each quantile represent-ing 26 revolutions. The open and solid bar shows the light-darkcycle used to entrain the mice. The arrowheads show the onsetand offset times of the 1-h light pulses (ca 120 lux in this example)used to study masking. Bottom: Means ± SEMs for the suppres-sion of locomotor activity by light pulses of decreasing irradiance.Some symbols have been slightly shifted on the x axis for clarity.

Page 76: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

an aspect of a more general irradiance detection sys-tem and that the hunt for the circadian receptor willturn up a receptor type with a much wider functionalrole.

ACKNOWLEDGMENTS

We thank Peggy Salmon for help. Support camefrom the Canadian Institutes of Health Research andBiotechnology and Biological Sciences ResearchCouncil.

REFERENCES

Aschoff J (1960) Exogenous and endogenous components incircadian rhythms. In Biological Clocks, J Aschoff, ed,pp 11-27, Cold Spring Harb Symp Quant Biol, ColdSpring Harbor, New York.

Carter-Dawson LD, LaVail MM, and Sidman RL (1978) Dif-ferential effect of the rd mutation on rods and cones in themouse retina. Invest Opthalmol Visual Sci 17:489-498.

Foster RG, Provencio I, Hudson D, Fiske S, De Grip W, andMenaker M (1991) Circadian photoreception in theretinally degenerate mouse (rd/rd). J Comp Physiol [A]169:39-50.

Freedman MS, Lucas RJ, Soni B, von Schantz M, Muñoz M,David-Gray Z, and Foster R (1999) Regulation of mam-malian circadian behavior by non-rod, non-cone, ocularphotoreceptors. Science 284:502-504.

Goto M and Ebihara S (1990) The influence of different lightintensities on pineal melatonin content in the retinaldegenerate C3H mouse and the normal CBA mouse.Neurosci Lett 108:267-272.

Keeler CE (1927) Iris movements in blind mice. Am J Physiol81:107-112.

Lucas RJ, Douglas RH, and Foster RG (2001) Characterisa-tion of an ocular photopigment capable of drivingpupillary constriction in mice. Nature Neurosci 4:621-626.

Lucas RJ, Freedman MS, Muñoz M, Garcia-Fernández J-M,and Foster RG (1999) Regulation of the mammalianpineal by non-rod, non-cone, ocular photoreceptors. Sci-ence 284:505-507.

Mrosovsky N (1994) In praise of masking: Behaviouralresponses of retinally degenerate mice to dim light.Chronobiol Int 11:343-348.

Mrosovsky N (1999) Masking: History, definitions, and mea-surement. Chronobiol Int 16:415-429.

Mrosovsky N, Foster RG, and Salmon PA (1999) Thresholdsfor masking responses to light in three strains of retinallydegenerate mice. J Comp Physiol [A] 184:423-428.

Mrosovsky N and Hampton RR (1997) Spatial responses tolight in mice with severe retinal degeneration. NeurosciLett 222:204-206.

Redlin U and Mrosovsky N (1999) Masking by light in ham-sters with SCN lesions. J Comp Physiol [A] 184:439-448.

Selby CP, Thompson C, Schmitz TM, Van Gelder RN, andSancar A (2000) Functional redundancy of crypto-chromes and classical photoreceptors for nonvisual ocu-lar photoreception in mice. Proc Natl Acad Sci U S A97:14697-14702.

van der Horst GTJ, Muijtjens M, Kobayashi K, Takano R,Kanno S, Takao M, de Wit J, Verkerk A, Eker APM, vanLeenen D, Buijs R, Bootsma D, Hoeijmakers JHJ, andYasui A (1999) Mammalian Cry1 and Cry2 are essentialfor maintenance of circadian rhythms. Nature 398:627-630.

Vitaterna MH, Selby CP, Todo T, Niwa H, Thompson C,Fruechte EM, Hitomi K, Thresher RJ, Ishikawa T,Miyazaki J, Takahashi JS, and Sancar A(1999) Differentialregulation of mammalian Period genes and circadianrhythmicity by cryptochromes 1 and 2. Proc Natl AcadSci U S A 96:12114-12119.

Mrosovsky et al. / IRRADIANCE DETECTION 587

Page 77: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

588

SRBR 2002

The Eighth Meeting of theSociety for Research on Biological Rhythms

(SRBR)will take place on May 22-26, 2002

at Amelia Island Plantation, Jacksonville, FL.

A joint one day meeting with theSleep Research Society

(SRS)will take place on May 22, 2002.

Log in at http://www.srbr.org/for details as they emerge.

Page 78: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

JOURNAL OF BIOLOGICAL RHYTHMS / December 2001INDEX

INDEXTo

JOURNAL OF BIOLOGICAL RHYTHMSVolume 16

Number 1 (February 2001) pp. 1-96Number 2 (April 2001) pp. 97-192Number 3 (June 2001) pp. 193-280Number 4 (August 2001) pp. 281-432Number 5 (October 2001) pp. 433-512Number 6 (December 2001) pp. 513-596

Authors:ALBRECHT, URS, BINHAI ZHENG, DAVID LARKIN,

ZHONG SHENG SUN, and CHENG CHI LEE, “mPer1and mPer2 Are Essential for Normal Resetting of the Cir-cadian Clock,” 100.

ALILA-JOHANSSON, AINO, LEA ERIKSSON, TIMOSOVERI, and MAIJA-LIISA LAAKSO, “Seasonal Varia-tion in Endogenous Serum Melatonin Profiles in Goats: ADifference between Spring and Fall?” 254.

ARANDA, A., J. A. MADRID, and F. J. SÁNCHEZ-VÁZQUEZ, “Influence of Light on Feeding AnticipatoryActivity in Goldfish,” 50.

ARANDA, A., see Sánchez-Vázquez, F. J.ARENDT, J., see Sopowski, M. J.BADURA, LORI L., see Imundo, J.BALER, RUBEN, “Clockless Yeast and the Gears of the

Clock: How Do They Mesh?” [Review], 516.BALL, GREGORY F., Dawson, A.BARTNESS, T. J., C. K. SONG, AND G. E. DEMAS, “SCN

Efferents to Peripheral Tissues: Implications for Biologi-cal Rhythms” [Review], 196.

BENHABEROU-BRUN, DALILA, see Dumont, M.BENTLEY, GEORGE E., Dawson, A.BERTOLUCCI, CRISTIANO, see Foà, A.BIELEFELD, ERIC, see Imundo, J.BIOULAC, BERNARD, see Taillard, J.BLOCH, GUY, DAN P. TOMA, and GENE E. ROBINSON,

“Behavioral Rhythmicity, Age, Division of Labor and pe-riod Expression in the Honey Bee Brain,” 444.

BULT, ABEL, see Mahoney, M.CAPODICE, CAMALA E., see Weaver, D. R.CARD, J. PATRICK, see Hannibal, J.CARRÉ, ISABELLE A., “Day-Length Perception and the

Photoperiodic Regulation of Flowering in Arabidopsis,”415.

CASAL, JORGE J., see Yanovsky, M. J.CHASTANG, JEAN-FRANÇOIS, see Taillard, J.

CHEMINEAU, PHILIPPE, Malpaux, B.DAAN, S., U. ALBRECHT, G.T.J. VAN DER HORST, H.

ILLNEROVÁ, T. ROENNEBERG, T. A. WEHR, and W. J.SCHWARTZ, “Assembling a Clock for All Seasons: AreThere M and E Oscillators in the Genes?” [Conjecture],105.

DAAN, SERGE, “A Parallactic View” [Response], 124.DE LA IGLESIA, HORACIO, see Schwartz, W. J.DAWSON, ALISTAIR, VERDUN M. KING, GEORGE E.

BENTLEY, and GREGORY F. BALL, “Photoperiodic Con-trol of Seasonality in Birds,” 365.

DEBOER, TOM, see Watanabe, K.DEMAS, G. E., see Bartness, T. J.DIEFENBACH, KONSTANZE, see Taillard, J.DITTAMI, JOHN P., see Millesi, E.DODGE, JAMES, see Imundo, J.DRAZEN, DEBORAH L., see Kriegsfeld, L. J.DUDEK, F. EDWARD, see Smith, B. N.DUMONT, MARIE, DALILA BENHABEROU-BRUN, and

JEAN PAQUET, “Profile of 24-h Light Exposure and Cir-cadian Phase of Melatonin Secretion in Night Workers,”502.

EDGAR, DALE M., see Kas, M.J.H.EDELSTEIN, K., see Mrosovsky, N.ERIKSSON, LEA, see Alila-Johansson, A.FAHRENKRUG, JAN, see Hannibal, J.FIEDER, MARTIN, see Millesi, E.FOÀ, AUGUSTO, and CRISTIANO BERTOLUCCI, “Tem-

perature Cycles Induce a Bimodal Activity Pattern inRuin Lizards: Masking or Clock-Controlled Event? ASeasonal Problem,” 574.

FOSTER, RUSSELL G., see Mrosovsky, N.FOLLETT, BRIAN, Hastings, M.FREE,AM. DAVID A., see Larkin, J. E.GANNON, ROBERT L., “5HT7 Receptors in the Rodent

Suprachiasmatic Nucleus” [Review], 19.

589

JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 589-593© 2001 Sage Publications

Page 79: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

GARCIA, CÉLIA R. S., REGINA P. MARKUS, and LUCI-ANA MADEIRA, “Tertian and Quartan Fevers: Tempo-ral Regulation in Malarial Infection” [Review], 436.

GIEBULTOWICZ, JADWIGA M., see Ivanchenko, M.GOLDMAN, BRUCE D., “Mammalian Photoperiodic Sys-

tem: Formal Properties and Neuroendocrine Mecha-nisms of Photoperiodic Time Measurement,” 283.

GORMAN, MICHAEL R. and THERESA M. LEE, “DailyNovel Wheel Running Reorganizes and Splits HamsterCircadian Activity Rhythms,” 541.

GORMAN, MICHAEL R., STEVEN M. YELLON andTHERESA M. LEE, “Temporal Reorganization of theSuprachiasmatic Nuclei in Hamsters with Split CircadianRhythms,” 552.

GOVERNALE, MARCIA M. and THERESA M. LEE, “Olfac-tory Cues Accelerate Reentrainment Following PhaseShifts and Entrain Free-Running Rhythms in FemaleOctodon degus (Rodentia),” 489.

GUBIK, BETTY, see Smale, L.HAMPTON, S. M., see Sopowski, M. J.HANNIBAL, JENS, NIELS VRANG, J. PATRICK CARD,

and JAN FAHRENKRUG, “Light-Dependent Inductionof cfos during Subjective Day and Night in PACAP-Con-taining Ganglion Cells of the Retinohypothalamic Tract,”457.

HARRINGTON, MARY E., “Feedback” [Poem], 277.HASTINGS, J. WOODLAND, “The Colin S. Pittendrigh Lec-

ture: Fifty Years of Fun” [Lecture], 5.HASTINGS, M. H., see Mrosovsky, N.HASTINGS, MICHAEL, “Modeling the Molecular Calen-

dar” [Commentary], 117.HASTINGS, MICHAEL, and BRIAN FOLLETT, “Toward a

Molecular Biological Calendar?” 424.HAZLERIGG, D. G., P. J. MORGAN, and S. MESSAGER,

“Decoding Photoperiodic Time and Melatonin in Mam-mals: What Can We Learn from the Pars tuberalis?” 326.

HORTON, TERESA H., and STEVEN M. YELLON, “Aging,Reproduction, and the Melatonin Rhythm in the SiberianHamster,” 243.

ILLNEROVÁ, H., see Albrecht, U.ILLNEROVÁ, HELENA, see Schwartz, W. J.IMUNDO, JASON, ERIC BIELEFELD, JAMES DODGE, and

LORI L. BADURA, “Relationship between Norepine-phrine Release in the Hypothalamic Paraventricular Nu-cleus and Circulating Prolactin Levels in the SiberianHamster: Role of Photoperiod and the Pineal Gland,”173.

IVANCHENKO, MARIA, RALF STANEWSKY, andJADWIGAM. GIEBULTOWICZ, “Circadian Photorecep-tion in Drosophila: Functions of Cryptochrome in Pe-ripheral and Central Clocks,” 205.

KAS, MARTIEN J. H., and DALE M. EDGAR, “ScheduledVoluntary Wheel Running Activity Modulates Free-Run-ning Circadian Body Temperature Rhythms in Octodondegus,” 66.

KENDALL, ADAM R., ALFRED J. LEWY, and ROBERT L.SACK, “Effects of Aging on the Intrinsic Circadian Pe-riod of Totally Blind Humans,” 87.

KENNAWAY, DAVID J., see Stepien, J. M.KING, VERDUN M., Dawson, A.KORF, HORST-WERNER, see Stehle, J. H.KRIEGSFELD, LANCE J., DEBORAH L. DRAZEN, and

RANDY J. NELSON, “Circadian Organization in MaleMice Lacking the Gene for Endothelial Nitric OxideSynthase (eNOS–/–),” 142.

KYRIACOU, C. P., see Tauber, E.LAAKSO, MAIJA-LIISA, see Alila-Johansson, A.LARKIN, DAVID, see Albrecht, U.LARKIN, JENNIE E., DAVID A. FREEMAN, and IRVING

ZUCKER, “Low-Ambient Temperature AcceleratesShort-Day Responses in Siberian Hamsters by AlteringResponsiveness to Melatonin,” 76.

LEE, CHENG CHI, see Albrecht, U.LEE, THERESA M., see Governale, M. M.LEE, THERESA M., see Gorman, M. R.LEWY, ALFRED J., see Kendall, A. R.LUCAS, ROBERT J., see Mrosovsky, N.MADEIRA, LUCIANA, see Garcia, C.R.S.MADRID, J. A., see Aranda, A.MADRID, J. A. see Sánchez-Vázquez, F. J.MAHONEY, MEGAN, ABEL BULT, and LAURA SMALE,

“Phase Response Curve and Light-Induced Fos Expres-sion in the Suprachiasmatic Nucleus and Adjacent Hypo-thalamus of Arvicanthis niloticus,” 149.

MALPAUX, BENOÎT, MARTINE MIGAUD, HÉLÈNETRICOIRE, and PHILIPPE CHEMINEAU, “Biology ofMammalian Photoperiodism and the Critical Role of thePineal Gland and Melatonin,” 336.

MARKUS, REGINA P., see Garcia, C.R.S.MAYWOOD, E. S., see Mrosovsky, N.MAZZELLA, M. AGUSTINA, see Yanovsky, M. J.MCELHINNY, TERESA, see Smale, L.MEIJER, JOHANNA H., see Watanabe, K.MIGAUD, MARTINE, see Malpaux, B.MILLESI, EVA, HERMANN PROSSINGER, JOHN P.

DITTAMI, and MARTIN FIEDER, “Hibernation Effectson Memory in European Ground Squirrels (Spermophiluscitellus),” 264.

MORGAN, L., see Sopowski, M. J.MORGAN, P. J., see Hazlerigg, D. G.MROSOVSKY, N., K. EDELSTEIN, M. H. HASTINGS, and

E. S. MAYWOOD, “Cycle of period Gene Expression in aDiurnal Mammal (Spermophilus tridecemlineatus): Impli-cations for Nonphotic Phase Shifting,” 471.

MROSOVSKY, N., ROBERT J. LUCAS, and RUSSELL G.FOSTER “Persistence of Masking Mice Lacking Rods andCones,” 585.

NELSON, RANDY J., see Kriegsfeld, L. J.NELSON, RANDY J., see Prendergast, B. J.NELSON, RANDY J., see Young, K. A.

590 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 80: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

NIXON, JOSHUA, see Smale, L.PAQUET, JEAN, see Dumont, M.PETRI, BERNHARD, and MONIKA STENGL, “Phase Re-

sponse Curves of a Molecular Model Oscillator: Implica-tions for Mutual Coupling of Paired Oscillators,” 125.

PHILIP, PIERRE, see Taillard, J.PICKARD, GARY E. see Smith, B. N.PRENDERGAST, BRIAN J., STEVEN M. YELLON, LONG T.

TRAN, and RANDY J. NELSON, “Photoperiod Modu-lates the Inhibitory Effect of in vitro Melatonin on Lym-phocyte Proliferation in Female Siberian Hamsters,” 224.

PROSSINGER, HERMANN, see Millesi, E.PUPIQUE, M., “How to Fix the ‘Review Process’” [Letter],

191.RIBEIRO, D.C.O., see Sopowski, M. J.ROENNEBERG, T., see Albrecht, U.ROENNEBERG, TILL, and MARTHA MERROW, “Season-

ality and Photoperiodism in Fungi,” 403.ROBINSON, GENE E., see Bloch, G.ROSE, SANDRA, see Smale, L.SACK, ROBERT L., see Kendall, A. R.SALDANHA, COLIN J., ANN-JUDITH SILVERMAN, and

RAE SILVER, “Direct Innervation of GnRH Neurons byEncephalic Photoreceptors in Birds,” 39.

SÁNCHEZ-VÁZQUEZ, F. J., A. ARANDA, and J. A. MA-DRID, “Differential Effects of Meal Size and Food EnergyDensity on Feeding Entrainment in Goldfish,” 58.

SÁNCHEZ-VÁZQUEZ, F. J., see Aranda, A.SCHOMERUS, CHRISTOF, see Stehle, J. H.SCHWARTZ, W. J., see Wehr, T. A.SCHWARTZ, W. J., see Albrecht, U.SCHWARTZ, WILLIAM J., HORACIO DE LA IGLESIA,

PIOTR ZLOMANCZUK, and HELENAILLNEROVÁ,“Encoding Le Quattro Stagioni within the MammalianBrain: Photoperiodic Orchestration through theSuprachiasmatic Nucleus,” 302.

SHEARMAN, LAUREN P. and DAVID R. WEAVER, “Dis-tinct Pharmacological Mechanisms Leading to c-fos GeneExpression in the Fetal Suprachiasmatic Nucleus,” 531.

SILVER, RAE, see Saldanha, C. J.SILVERMAN, ANN-JUDITH, see Saldanha, C. J.SMALE, LAURA, TERESA MCELHINNY, JOSHUA

NIXON, BETTY GUBIK, and SANDRA ROSE, “Patternsof Wheel Running Are Related to Fos Expression in

Neuropeptide Y–Containing Neurons in the IntergeniculateLeaflet of Arvicanthis niloticus,” 163.

SMALE, LAURA, see Mahoney, M.SMITH, BRET N., PATRICIA J. SOLLARS, F. EDWARD

DUDEK, and GARY E. PICKARD, “Serotonergic Modu-lation of Retinal Input to the Mouse Suprachiasmatic Nu-cleus Mediated by 5-HT1B and 5-HT7 Receptors,” 25.

SOLLARS, PATRICIA J., see Smith, B. N.SONG, C. K, see Bartness, T. J.SOPOWSKI, M. J., S. M. HAMPTON, D.C.O. RIBEIRO, L.

MORGAN, and J. ARENDT, “Postprandial Triacylglyc-

erol Responses in Simulated Night and Day Shift: GenderDifferences,” 272.

SOVERI, TIMO, see Alila-Johansson, A.STANEWSKY, RALF, see Ivanchenko, M.STEHLE, JÖRG H., CHARLOTTE VON GALL, CHRISTOF

SCHOMERUS, and HORST-WERNER KORF, “Of Ro-dents and Ungulates and Melatonin: Creating a UniformCode for Darkness by Different Signaling Mechanisms,”312.

STENGL, MONIKA, see Petri, B.STEPIEN, JACQUELINE M., and DAVID J. KENNAWAY,

“Phase Response Relationships between Light Pulsesand the Melatonin Rhythm in Rats,” 234.

SUN, ZHONG SHENG, see Albrecht, U.TAILLARD, JACQUES, PIERRE PHILIP, JEAN-FRANÇOIS

CHASTANG, KONSTANZE DIEFENBACH, BERNARDBIOULAC, “Is Self-Reported Morbidity Related to theCircadian Clock?” 183.

TAUBER, E., and C. P. KYRIACOU, “Insect Photoperiodismand Circadian Clocks: Models and Mechanisms,” 381.

TOMA, DAN P., see Bloch, G.TRAN, LONG T., see Prendergast, B. J.TRICOIRE, HÉLÈNE, Malpaux, B.VAN DER HORST, G.T.J., see Albrecht, U.VON GALL, CHARLOTTE, see Stehle, J. H.VRANG, NIELS, see Hannibal, J.WAYNE, NANCY L., “Regulation of Seasonal Reproduction

in Mollusks,” 391.WATANABE, KAZUTO, TOM DEBOER and JOHANNA H.

MEIJER, “Light-Induced Resetting of the Circadian Pace-maker: Quantitative Analysis of Transient versusSteady-State Phase Shifts,” 564.

WEAVER, DAVID R., see Shearman, L. P.WEAVER, DAVID R., and CAMALA E. CAPODICE, “Post-

mortem Stability of Melatonin Receptor Binding andClock-Relevant mRNAs in Mouse Suprachiasmatic Nu-cleus,” 216.

WEHR, T. A., see Albrecht, U.WEHR, T. A., and W. J. SCHWARTZ, “Assembling a Clock

for All Seasons: Are There M and E Oscillators in theGenes?” [Conjecture], 105.

WEHR, THOMAS A., “Photoperiodism in Humans andOther Primates: Evidence and Implications,” 348.

WHITELAM, GARRY C., see Yanovsky, M. J.YANOVSKY, MARCELO J., M. AGUSTINA MAZZELLA,

GARRY C. WHITELAM, and JORGE J. CASAL, “Re-setting of the Circadian Clock by Phytochromes andCryptochromes in Arabidopsis,” 523.

YELLON, STEVEN M., see Gorman, Michael R.YELLON, STEVEN M., see Horton, T. H.YELLON, STEVEN M., see Prendergast, B. J.YOUNG, KELLY A., BARRY R. ZIRKIN, and RANDY J.

NELSON, “Testicular Apoptosis Is Down-Regulatedduring Spontaneous Recrudescence in White-FootedMice (Peromyscus leucopus),” 479.

INDEX 591

Page 81: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

ZATZ, MARTIN, “Editorial: Pebbles of Truth,” 515.ZATZ, MARTIN, “Editoral: On Telling It Like It Was,” 195.ZATZ, MARTIN, “Editorial: Show Me the Data!” 99.ZATZ, MARTIN, “Editorial: What Do Reviewers Really

Want?” 3.ZATZ, MARTIN, “Editorial: Yes Sir, That’s My Data!” 435.ZHENG, BINHAI, see Albrecht, U.ZIRKIN, BARRY R., see Young, K. A.ZLOMANCZUK, PIOTR, see Schwartz, W. J.ZUCKER, IRVING, see Larkin, J. E.

Articles:“Aging, Reproduction, and the Melatonin Rhythm in the Si-

berian Hamster,” Horton and Yellon, 243.“Behavioral Rhythmicity, Age, Division of Labor and period

Expression in the Honey Bee Brain,” Bloch et al., 444.“Biology of Mammalian Photoperiodism and the Critical

Role of the Pineal Gland and Melatonin,” Malpaux et al.,336.

“Circadian Organization in Male Mice Lacking the Gene forEndothelial Nitric Oxide Synthase (eNOS–/–),”Kriegsfeld et al., 142.

“Circadian Photoreception in Drosophila: Functions ofCryptochrome in Peripheral and Central Clocks,”Ivanchencko et al., 205.

“Cycle of period Gene Expression in a Diurnal Mammal(Spermophilus tridecemlineatus): Implications forNonphotic Phase Shifting,” Mrosovsky et al., 471.

“Daily Novel Wheel Running Reorganizes and Splits Ham-ster Circadian Activity Rhythms,” Gorman and Lee, 541.

“Day-Length Perception and the Photoperiodic Regulationof Flowering in Arabidopsis,” Carré, 415.

“Decoding Photoperiodic Time and Melatonin in Mammals:What Can We Learn from the Pars tuberalis?” Hazlerigget al., 326.

“Differential Effects of Meal Size and Food Energy Densityon Feeding Entrainment in Goldfish,” Sánchez-Vázquezet al., 58.

“Direct Innervation of GnRH Neurons by EncephalicPhotoreceptors in Birds,” Saldanha et al., 39.

“Distinct Pharmacological Mechanisms Leading to c-fosGene Expression in the Fetal Suprachiasmatic Nucleus,”Shearman and Weaver, 531.

“Editorial: Pebbles of Truth,” Zatz, 515.“Editoral: On Telling It Like It Was,” Zatz, 195.“Editorial: Show Me the Data!” Zatz, 99.“Editorial: What Do Reviewers Really Want?” Zatz, 3.“Editorial: Yes Sir, That’s My Data!” Zatz, 453.“Effects of Aging on the Intrinsic Circadian Period of Totally

Blind Humans,” Kendall et al., 87.“Encoding Le Quattro Stagioni within the Mammalian Brain:

Photoperiodic Orchestration through the SuprachiasmaticNucleus,” Schwarz et al., 302.

“Hibernation Effects on Memory in European GroundSquirrels (Spermophilus citellus),” Millesi et al., 264.

“Influence of Light on Feeding Anticipatory Activity inGoldfish,” Aranda et al., 50.

“Insect Photoperiodism and Circadian Clocks: Models andMechanisms,” Tauber and Kyriacou, 381.

“Is Self-Reported Morbidity Related to the CircadianClock?” Taillard et al., 183.

“Light-Dependent Induction of cfos during Subjective Dayand Night in PACAP-Containing Ganglion Cells of theRetinohypothalamic Tract,” Hannibal et al., 457.

“Light-Induced Resetting of the Circadian Pacemaker:Quantitative Analysis of Transient versus Steady-StatePhase Shifts,” Watanabe, Deboer and Meijer, 564.

“Low-Ambient Temperature Accelerates Short-Day Re-sponses in Siberian Hamsters by Altering Responsive-ness to Melatonin,” Larkin et al., 76.

“Mammalian Photoperiodic System: Formal Properties andNeuroendocrine Mechanisms of Photoperiodic TimeMeasurement,” Goldman, 283.

“mPer1 and mPer2 Are Essential for Normal Resetting of theCircadian Clock,” Albrecht et al., 100.

“Of Rodents and Ungulates and Melatonin: Creating a Uni-form Code for Darkness by Different Signaling Mecha-nisms,” Stehle et al., 312.

“Olfactory Cues Accelerate Reentrainment Following PhaseShifts and Entrain Free-Running Rhythms in FemaleOctodon degus (Rodentia),” Governale and Lee, 489.

“Patterns of Wheel Running Are Related to Fos Expressionin Neuropeptide Y–Containing Neurons in theIntergeniculate Leaflet of Arvicanthis niloticus,” Smaleet al., 163.

“Phase Response Curve and Light-Induced Fos Expressionin the Suprachiasmatic Nucleus and Adjacent Hypothal-amus of Arvicanthis niloticus,” Mahoney et al., 149.

“Phase Response Curves of a Molecular Model Oscillator:Implications for Mutual Coupling of Paired Oscillators,”Petri and Stengle, 125.

“Phase Response Relationships between Light Pulses andthe Melatonin Rhythm in Rats,” Stepien and Kennaway,234.

“Photoperiod Modulates the Inhibitory Effect of in vitroMelatonin on Lymphocyte Proliferation in Female Sibe-rian Hamsters,” Prendergast et al., 224.

“Photoperiodic Control of Seasonality in Birds,” Dawsonet al., 365.

“Photoperiodism in Humans and Other Primates: Evidenceand Implications,” Wehr, 348.

“Postmortem Stability of Melatonin Receptor Binding andClock-Relevant mRNAs in Mouse Suprachiasmatic Nu-cleus,” Weaver and Capodice, 216.

“Postprandial Triacylglycerol Responses in SimulatedNight and Day Shift: Gender Differences,” Sopowskiet al., 272.

“Profile of 24-h Light Exposure and Circadian Phase ofMelatonin Secretion in Night Workers,” Dumont et al.,502.

“Regulation of Seasonal Reproduction in Mollusks,”Wayne, 391.

“Relationship between Norepinephrine Release in the Hy-pothalamic Paraventricular Nucleus and Circulating

592 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Page 82: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Prolactin Levels in the Siberian Hamster: Role ofPhotoperiod and the Pineal Gland,” Imundo et al., 173.

“Resetting of the Circadian Clock by Phytochromes andCryptochromes in Arabidopsis,” Yanovsky, Mazzella,Whitelam, and Casal, 523.

“Scheduled Voluntary Wheel Running Activity ModulatesFree-Running Circadian Body Temperature Rhythms inOctodon degus,” Kas and Edgar, 66.

“Seasonal Variation in Endogenous Serum Melatonin Pro-files in Goats: A Difference between Spring and Fall?”Alila-Johansson et al., 254.

“Seasonality and Photoperiodism in Fungi,” Roennebergand Merrow, 403.

“Serotonergic Modulation of Retinal Input to the MouseSuprachiasmatic Nucleus Mediated by 5-HT1B and 5-HT7

Receptors,” Smith et al., 25.“Temperature Cycles Induce a Bimodal Activity Pattern in

Ruin Lizards: Masking or Clock-Controlled Event? ASeasonal Problem,” Foà and Bertolucci, 574.

“Temporal Reorganization of the Suprachiasmatic Nuclei inHamsters with Split Circadian Rhythms,” Gorman,Yellon and Lee, 552.

“Testicular Apoptosis Is Down-Regulated during Spontane-ous Recrudescence in White-Footed Mice (Peromyscusleucopus),” Young et al., 479.

Commentary:“Modeling the Molecular Calendar,” Hastings, M., 117.Conjecture: “Assembling a Clock for All Seasons: Are There

M and E Oscillators in the Genes?” Daan et al., 105.

Lectures:“The Colin S. Pittendrigh Lecture: Fifty Years of Fun,”

Hastings, 5.

Letters:“How to Fix the ‘Review Process,’” Pupique, 191.“Persistence of Masking Responses to Light in Mice Lakcing

Rods and Cones,” Mrosovsky, Lucas and Foster, 585.

Poem:“Feedback,” Harrington, 277.

Reflection:“Toward a Molecular Biological Calendar?” Hastings and

Follett, 424.

Response:“A Parallactic View,” Daan, 124.

Reviews:“Clockless Yeast and the Gears of the Clock: How Do They

Mesh?” Baler, 516.“5HT7 Receptors in the Rodent Suprachiasmatic Nucleus,”

Gannon, 19.“SCN Efferents to Peripheral Tissues: Implications for Bio-

logical Rhythms,” Bartness et al., 196.“Tertian and Quartan Fevers: Temporal Regulation in Ma-

larial Infection,” Garcia et al., 436.

INDEX 593

Page 83: Journal of Biological Rhythms€¦ · Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms Volume 16, Issue 6 December 2001 EDITORIAL

Journal of Biological RhythmsOfficial Publication of the Society

for Research on Biological Rhythms

INSTRUCTIONS TO AUTHORS

Four legible, double-spaced copies of each paper (includingillustrations) should be submitted to the Editor. There are nopublication charges except for circumstances requiring specialprinting, color reproduction, or unusual length and number ofillustrations; in these cases, the publisher will provide cost in-formation before the paper is accepted.

Manuscripts should be written clearly and concisely andshould conform generally to the Council of Biology Editors StyleManual. Manuscripts should consist of the following:

1. A title page, including the authors’ names, the title and ashort running title, and the institution(s) (with all words spelledout in full) from which the paper emanates. If current addressesare different, then these should be indicated in a footnote. Alsoinclude the name, mailing address, phone and fax numbers, ande-mail address of the person to whom correspondence andproofs should be sent. Indicate the number of pages, illustra-tions, and tables.

2. Asuccinct Abstract, not exceeding 350 words, in a form ac-ceptable to most abstracting services, followed by six to eightkey words.

3. An Introduction that provides a brief review of relevantbackground material and indicates the purpose of the study.

4. AMaterials and Methods section that provides sufficient in-formation for qualified investigators to reproduce the work insimilar fashion. Reference to published procedures by appro-priate succinct summary and citation is encouraged but shouldnot replace adequate methodological description. The Journal ofBiological Rhythms endorses the Declaration of Helsinki and theU.S. National Institutes of Health Guide for the Care and Use ofLaboratory Animals. It will be assumed that authors have consci-entiously followed principles and practices in accord with thesedocuments in experiments involving human subjects and ex-perimental animals.

5. A Results section that provides a concise description of thefindings with appropriate reference to illustrations and tables.

6. A Discussion section that includes a summary of the mainfindings, their relation to other published work, and a statementof their significance.

7. References should be listed sequentially in alphabeticalorder according to the name of the first author with (a) a full listof authors, (b) date, (c) full title of the paper, (d) journal titles ab-breviated as per Index Medicus, (e) volume number, and (f) firstand last pages. Only papers published or in press may be in-cluded in the reference list. Papers should be cited in the text byauthor(s) and date.

Examples

Aschoff J (1965) Response curves in circadian periodicity. InCircadian Clocks, J Aschoff, ed, pp 95-111, North-Holland, Am-sterdam.

Pittendrigh CS and Daan S (1976) A functional analysis ofcircadian pacemakers in nocturnal rodents. I. The stability andlability of spontaneous frequency. J Comp Physiol [A] 106:223-252.

Richter CP (1965) Biological Clocks in Medicine and Psychiatry,Charles C Thomas, Springfield, IL.

8. Tables and illustrations should be submitted in a formsuitable for direct reproduction when possible. In preparationof figures (glossy prints, crisp laser prints, or hand-drawngraphs), authors should consider page format and letter size af-ter final reproduction. Transfer (rub-on) type should beavoided. An acetate overlay containing type on a clear back-ground is recommended for lettering and calibration bars ap-plied to halftone illustrations. Each illustration should bemarked on the reverse side with its number, name(s) ofauthor(s), orientation, and instructions for reduction when ap-propriate. Copies of the original set can be used for review pur-poses but should be of suitable quality to enable the referees toevaluate them. Place each table on a separate sheet, following thefigure legends. Number consecutively with Arabic numerals.Charges for color figures are $800.00 for the first figure and$200.00 for each additional figure.

9. Alist of figure captions should follow the references. Abrieftitle and description of each illustration should be included. Theseshould be typed consecutively on the same page(s).

10. Abbreviations should be introduced in parentheses afterthe first occurrence of the term being abbreviated. Use standardmetric units wherever possible.

Submission of a manuscript implies assurance that no simi-lar paper has been submitted for publication elsewhere. Con-currence of all listed authors is assumed for all papers to bepublished. Manuscripts are evaluated on their merit by the edi-tors with the assistance of selected referees. Copyright is vestedin the publisher.

Mail submissions toMartin Zatz, Ph.D., M.D.Editor, Journal of Biological RhythmsPMB 3364938 Hampden LaneBethesda, MD 20814 - 2914tel: 301 - 656 - 4655e-mail: [email protected]