Proc. Nat. Acad. Sci. USAVol. 71, No. 12, pp. 5103-5108, December 1974Symposium

Synthesis of Messenger RNA and Chromosome Structure in theCellular Slime Mold*

[Dictyostelium discoideum/RNA and protein synthesis/poly(A)J

HARVEY F. LODISH, ALLAN JACOBSONt, RICHARD FIRTELI, TOM. ALTON, AND JESSICA TUCHMAN§

Department of Biology, Massachusetts Institute of Technology, Cambridge, Mass. 02139

ABSTRACT This paper summarizes our knowledge ofthe structure and biosynthesis of messenger RNA in theslime mold Dictyostelium discoideum, the arrangement ofDNA sequences in the Dictyostelium chromosome, andthe changes in the pattern of predominant polypeptidessynthesized during Dictyostelium development.

Introduction

In the presence of a food source, amoebae of the cellular slimemold, Dictyostelium discoideum, will grow exponentially as ahomogeneous population of morphologically identical cells.When such amoebae are removed from nutrients and de-posited on the surface of a filter impregnated with a buffer,they synchronously undergo a series of morphologically de-fined events: groups of approximately 105 cells aggregate toform migrating pseudoplasmodia which subsequently differ-entiate into fruiting bodies containing spore and stalk cells(1-3).Biochemical studies of this simple differentiating system

have been facilitated by two major attributes of the organism:First, Dictyostelium is easy to grow; thus, large quantities ofsynchronous cells can be obtained from any stage (4, 5).Second, Dictyostelium is genetically simple: It is haploid andcontains only 0.05 pg of nuclear DNA per cell (6). WhenDictyostelium nuclear DNA is analyzed by DNA-DNA re-naturation kinetics, 30% of the nucleotide sequences arefound to be repetitive, present in 100-300 copies per genome.The remaining 70% renature with kinetics characteristic ofsequences present only once per genome. The complexity ofthe nonreiterated DNA is approximately only seven times thegenome size of the bacterium Escherichia coli (6).

It has been possible to demonstrate that the morphologicalchanges of the life cycle are accompanied by alterations in thepatterns of both RNA and protein synthesis. RNA -DNAhybridization saturation experiments show that during thefirst 20 hr of morphogenesis at least 28% of the nonrepetitiveDNA is transcribed. Additional hybridization experimentsindicate that there are also changes in the portion of the non-reiterated DNA which is transcribed at different morpho-logical stages (7). Further, at specific stages during morpho-

* This paper was presented at the Annual Meeting of the NationalAcademy of Sciences in Washington, D.C. on April 22-24, 1974.It was presented in A Symposium on Organization and Transcrip-tion of the Eukaryotic Genome.t Present address: Department of Microbiology, University ofMassachusetts Medical School, Worcester, Mass. 01605.1 Present address: Department of Biology, University of Cali-fornia, San Diego, La Jolla, Calif. 92037.§ Present address: Environment Sub-Committee, Committee ofInterior and Insular Affairs, U.S. House of Representatives,Washington, D.C.

genesis the specific activities of certain enzymes increase con-siderably and then remain constant or decrease. Experimentswith protein synthesis inhibitors suggest that the increase inenzyme activity is dependent on concomitant protein syn-thesis (2, 3, 8). This point has been firmly established for theenzyme UDP-glucose pyrophosphorylase (EC 2.7.7.9; UTP:a-D-glucose-1-phosphate uridyltransferase) (9, 10). Recentstudies in which a combination of RNA synthesis inhibitorsare used, suggest, in contrast to earlier results (2), that there isonly a short lag between synthesis and translation of mRNAsfor "developmentally regulated" enzymes (11).

Electrophoretic Characterization of Protein Synthesi. Sincethe synthesis of none of the developmentally regulated en-zymes studied represented more than 1% of the cells' totalprotein synthesis at any instant, we decided to characterizethe predominant species of polypeptides which are synthe-

DEVELOPMENTAL CHANGES IN PROTEIN SYNTHESIS

N gr WD O CQ IT

V

> 0 NS\q 0 Nj

I

- ..b

B<.- A

- N N

w 0 0_ - cl

.. ;,-1 v-i

_ v~~~~~

FIG. 1. Radioautogram of polyacrylamide gel analysis ofprotein synthesis during Dictyostelium growth and development.Cells were labeled for 2 hr with [14S]methionine as described inref. 12, harvested, lysed, and electrophoresed on 5.5% poly-acrylamide gels that contain 0.1% sodium dodecyl sulfate. Thedried gels were exposed to Kodak Royal Blue X-Ray Film for 8days. The morphological stages corresponding to the period oflabeling are shown schematically below each gel. Aggregation -iscomplete by 12 hr under these conditions. Arrowheads on the leftof a gel denote bands which increase markedly in intensity at theindicated time; arrowtails on the right of a gel denote the markeddecrease in intensity of a given band. Band A is another pro-tein synthesized in early development, and may be myosin.Band B is the putative Dictyostelium actin. Veg refers to grow-ing, vegetative cells; numbers refer to the time in hours after initi-ation of development.

5103

5104 Symposium: Lodish et al.

STRUCTURE AND TRANSCRIPTION OF DICTYOSTEUUM DNA

R SC A25 R' SC' A25I I IlI I I I

T25

R" SC" A25II

T25

I TRANSCRIPTION

R SC A25X

IIR SC A2XpA1

I TRANSPORT TOCYTOPLASM

r SC IA25XJpAlw

FIG. 2. A model of Dictyostelium nuclear DNA. The primary genetic unit in Dictyostelium contains a repetitive DNA sequence (R)that averages 300 to 350 nucleotides at the 5' end; a sequence of nonreiterated or single copy (SC) DNA of 1100 to 1200 nucleotides; anda sequence of 25 adenylic acid residues (A25) at the 3' end. This unit is transcribed by RNA polymerase and then poly(A) of 100 to 150residues is added after transcription. Before transport of the heterogeneous RNA to the cytoplasm, the majority of the 5' repetitive se-quences are removed, leaving a short repetitive sequence. The evidence for the presence of a portion of the repetitive sequence transcript(r) on the majority of mRNA molecules is given in ref. 13. Since the poly(A)25 and poly(A)soo are separable on polyacrylamide gels afterdigestion with RNase T1 and RNase A, at least one other nucleotide (X) must be present between the 3' end of the short poly(A)25 andthe 5' end of the larger poly(A)10o.

IJ 5 10 15 20 25 30CUofo Fraction numberFIG. 3. Sucrose gradient analysis of nuclear and cytoplasmic

heterogeneous RNA. Top panel (Total RNA radioactivity):*-4, pulse-labeled nuclear RNA; O-O, cytoplasmic RNA;

, absorbance (260 nm). Bottom panel [RNA radioactivitybound to poly(U) filters]: * -*, pulse-labeled nuclear RNA;0-O, cytoplasmic RNA. To label cytoplasmic RNA, wetreated log phase vegetative cells with 3 ug/ml of actinomycin Dand added 32PO4 (2 mCi/ml) 15 min later. After an additional 15min, the cells were diluted with 4 volumes of medium containing125 ,g/ml of actinomycin D and 200 Ag/ml of daunomycin tostop all RNA synthesis (11). After an additional incubation of 60min the cells were harvested. The cells were fractionated and the,RNA extracted from the cytoplasm. Since 3 ug/ml of actino-mycin D suppresses rRNA synthesis but does not affect labelingof nuclear or cytoplasmic poly(A)-containing RNA (17), theuse of this concentration permitted the cytoplasmic poly(A)-containing RNA to be visualized more readily. To obtain pulse-labeled nuclear RNA, we harvested log phase vegetative cellsand resuspended them in fresh medium that contained 200pCi/ml of [3H]uracil. After 2 min, the cells were harvested,

sized at different developmental stages. Fig. 1 shows a radio-autogram of a polyacrylamide gel analysis of protein syn-thesized during 2-hr periods throughout development. Asindicated by the arrowheads and arrowtails, a large number ofpredominant polypeptide species are made at discrete times.An unexpected dividend of this work was the discovery that asingle polypeptide species, band B, comprises 25% of the pro-tein synthesized during the first 8 hr of development-thepre-aggregation stage. This protein is produced in at least5-fold lower amounts (relative to total synthesis) duringvegetative growth, and during the later post-aggregationdevelopmental stages. We recently showed that this protein,of molecular weight 47,000, is Dictyostelium actin (12).

mRNA Structure and Metabolism. We presently have afairly detailed understanding of the synthesis and processingof total Dictyostelium mRNA (see Fig. 2 and ref. 13). Ourdata is drawn from the average behavior of all mRNAs andmust be extended to specific mRNA species, such as that foractin. Nevertheless, the scheme outlined in Fig. 2 representsour present knowledge of the structure and transcription of atleast half of the Dictyostelium genome. The remainder of thisreview summarizes our evidence supporting this scheme.

fractionated, and RNA extracted from the purified nuclei. Thecytoplasmic and nuclear RNAs were mixed and run on 36 ml of15-30% sodium dodecyl sulfate sucrose gradients. Centrifu-gation was done in the Beckman SW 27 rotor for 20 hr at 26,000rpm, 220. The gradients were collected from the bottom througha Gilford recording spectrophotometer and the A260 was moni-tored. An aliquot was counted directly and another was dilutedwith one volume of 0.2 M Na2HPO4-NaHPO4, 1% sodiumdodecyl sulfate (pH 6.8) and passed through poly(U) filters.The absorbance profile is from the cytoplasmic RNA anddoes not indicate the relative amount of nuclear RNA present.Centrifugation is from right to left; 28S and 18S rRNA are atfractions 9 and 16, respectively. The 32p radioactivity at the topof the gradient (top panel) is not in RNA, since the materialremains precipitable by trichloroacetic acid after extensivealkaline digestion (17).

Proc. Nat. Acad. Sci. USA 71 (1974)

Symposium: Dictyostelium Messenger RNA 5105

10 8 6 4 2 0

CM (TOP)

FIG. 4. Cell-free translation of Dictyostelium RNA containingand lacking poly(A) sequences. One milligram of cytoplasmicRNA from vegetative Dictyostelium was dissolved in bindingbuffer (0.5 M NaCl, 0.01 M Tris-HCl, pH 7.5; 0.5% sodiumdodecyl sulfate) and passed through a 0.5-ml column of oligo(dT)-cellulose (Collaborative Research, Inc.; Waltham, Mass.).The eluate from this column was passed through a second column,and the eluate from this through a third column. Material whichdid not bind to any of these columns was considered nonbindingRNA. RNA adsorbed to each of these columns was eluted in a

low-salt buffer (0.01 M Tris -HCl, pH 7.5; 0.05% sodium dodecylsulfate). All RNA samples were precipitated with ethanol anddissolved in 0.06 ml of H20. Ten microliters of each RNA were

added to a 50-,ul cell-free reaction that contained wheat germ

extract and ['4S]methionine (27, 28). Incubation was for 2 hr at22.50. The protein products were analyzed on 10% polyacryl-amide gels containing 0.1% sodium dodecyl sulfate. Shown are

scans of the radioautograms of the dried gels. Wheat germ re-

actions with no added RNA incorporated less than 1% of theradioactivity as compared to samples with RNA; the radioauto-graphs showed no bands and the profile is not shown here. Theamount of any of these RNAs used was not sufficient to saturatethe cell-free system, and the amount of radioactivity incorporated

As in metazoan cells (reviewed in refs. 14 and 15), themessenger RNA of Dictyostelium contains poly(adenylic acid)[poly(A) ] sequences (16). Thus, messenger RNA can be sep-arated from ribosomal and tRNA by chromatography oncolumns of poly(U)-Sepharose or oligo(dT)-cellulose; over90% of the rapidly labeled heterodisperse cytoplasmic RNAthat sediments between 12 and 16 S will bind to immobilizedpoly(U) (Fig. 3) and is, therefore, associated with at least onesequence of poly(A) (16, 17). Since over 90% of this labeledpoly(A)-containing RNA is polysomal (17), we consider it tobe mRNA.We have shown that Dictyostelium mRNA can be trans-

lated faithfully in cell-free extracts of rabbit reticulocytes orwheat germ, yielding protein products expected for the stagefrom which the mRNA was isolated (T. Alton and H. F.Lodish, submitted for publication). As an example, Fig.4 shows that poly(A)-containing RNA isolated from thecytoplasm of developing cells of age 2 hr directs the synthesispredominantly of a single polypeptide, which comigrates onpolyacrylamide gel electrophoresis with Dictyostelium actin.Analysis of tryptic peptides of this protein shows that it is in-deed actin (Alton and Lodish, unpublished). As shown by thedashed lines it appears that other authentic Dictyosteliumproteins are also made in the cell-free system, but we have notyet confirmed their identity by peptide mapping. A largefraction of messenger RNA activity does not stick to oligo-(dT)-cellulose, even after several passages, and presumablydoes not contain a poly(A) sequence (Fig. 4). This RNA alsodirects synthesis of actin and of the other major Dictyo-stelium polypeptides whose synthesis is directed by poly(A)-containing RNA. We suspect that these mRNAs are derivedfrom poly(A)-containing RNAs whose poly(A) sequences be-come short with age (18).

Synthesis ofmRNA

Unlike messenger RNA of metazoan cells, mRNA in Dictyo-stelium is synthesized as a precursor molecule only 20%longer than is the mRNA itself. This precursor also containsa poly(A) sequence (13, 17).

In the experiment depicted in Fig. 3, vegetative Dictyo-stelium amoebae were labeled for 2 min with ['Hjuracil; RNAwas extracted from nuclei and mixed with 82P-labeled cyto-plasmic RNA. The 8H-labeled nuclear RNA exhibits a largepeak of material sedimenting at 32 S, smaller peaks at 22 S518 S, and 4 S, and some heterodisperse material sedimentingbetween 6 S and 18 S. The labeled 32S species is a precursor ofrRNA (see ref. 19); probably this is the case as well for alarge fraction of the labeled 22-24S RNA. In marked con-trast to the case in metazoan cells, little radioactive RNAsediments faster than the 32S rRNA precursor.Of the labeled nuclear RNA that sediments between 6 and

18 S, over 60% adsorbs to poly(U) filters, indicating that itwas proportional to the amount of RNA added at least up to theamount used here. a) Cell-free product directed by RNA whichdid not bind to oligo(dT)-cellulose. b) Cell-free product directedby RNA which bound to the third oligo(dT)-cellulose column.c) Cell-free product directed by RNA which bound to the secondcolumn. d) Cell-free product directed by RNA which bound tothe first column. e) Profile of proteins synthesized by vegetativecells during a 2-hr pulse with [3Slmethionine.Aliquots of 5 41 were taken to determine the incorporation of

radioactivity into protein. a) 187,000 cpm, b) 26,000 cpm, c)53,000 cpm, d) 218,000 cpm. No added RNA: 23,000 cpm.

Proc. Nat. Acad. Sci. USA 71 (1974)

Proc. Nat. Acad. Sci. USA 71 (1974)

I

w 3zz

I

,

x

E0

40w

30 Zzw

20

10 :

10

x

E

5 0

IN

0

x

E0

cs

20 40 60 80 100 120 140 20 40 60 80 100 120

(E MM E)* MMe

FIG. 5. Analysis of Dictyostelium poly(A) and poly(dT) on polyacrylamide gels. A. Poly(A) in RNA. Nuclear and cytoplasmic hetero-

geneous RNAs labeled with [14C]adenine and [3H]adenine, respectively, were isolated by poly(U)-Sepharose chromatography. After

digestion with RNase A and RNase T1, poly(A) was purified by oligo(dT)-cellulose chromatography and subjected to electrophoresis in

10% polyacrylamide gels as described in ref. 23. BPB is the position of the bromphenol blue tracking dye; 5 S and 4 S are the positions

of the respective RNAs run on parallel gels. On these gels a sample of reovirus oligo(A) of about 18 bases in length migrated to 114 mm

(data not shown). Note that the scale for 14C (nuclear RNA) is broken. HnRNA is heterogeneous RNA. B. Poly(dT) in DNA. Cells of

Dictyostelium discoideum were labeled with 50 mCi Of 32PO4. Nuclear DNA was purified by the method of Firtel and Bonner (6) and chemi-

cally depurinated with diphenylamine and formic acid as described by Burton (24). After depurination, the DNA was extracted with

ether three times and then chromatographed on poly(A)-Sepharose. Conditions for poly(A)-Sepharose chromatography are precisely the

same as those used for poly(U)-Sepharose chromatography (17). Poly(dT) residues, bound to, and eluted from, poly(A)-Sepharose were

precipitated with ethanol and then electrophoresed on a 10% gel.

must contain at least one poly(A) sequence. The peak ofnuclear poly(A)-containing RNA sediments at 15-16 S, or

about 2S units faster than the peak of cytoplasmic poly(A)-containing RNA. As determined by sedimentation in gradientscontaining 99% dimethyl sulfoxide, the average molecularweight of the nuclear poly(A)-containing RNA is 500,000;this is 20-25% larger than the mRNA, which has an average

molecular weight of 400,000 (17).In mammalian cells, over 90% of nuclear RNA labeled

during a short pulse of radioactivity is degraded within thenucleus, and is not transferred to the cytoplasm (14, 15).By contrast, pulse-labeling experiments in Dictyosteliumshowed that over 70% of the nuclear poly(A)-containingRNA is transferred to polysomes and, hence, is a materialprecursor for mRNA (17). This point is substantiated byhybridization experiments in which DNA complementary tomRNA (cDNA) was used; the cDNA was synthesized by re-

verse transcriptase with oligo(dT) as primer (13, 20).When labeled mRNA is hybridized to an excess of cDNA

complementary to mRNA, over 90% of the RNA radio-activity is incorporated into an RNA -DNA hybrid and henceis resistant to ribonucleases; this result indicates that the anti-message DNA contains sequences complementary to over 90%of the mRNA species.

When labeled nuclear poly(A)-containing RNA is hy-bridized to anti-message DNA, 70-75% of the RNA radio-activity was incorporated into the nuclease-resistant form(13). This result indicates that at least 70% of the sequences

in nuclear RNA are in fact found in cytoplasmic mRNA and,hence, that nuclear poly(A)-containing RNA is an informa-tional percursor to mRNA. About 30% of nuclear poly(A)-containing RNA is not found in mRNA, and is presumablydestroyed within the nucleus.

Studies measuring the rate of hybridization of RNA to a

vast excess of denatured Dictyostelium DNA showed that over

90% of sequences in mRNA are complementary to, and thustranscribed from, DNA sequences present in only one copy

per genome. By contrast, the kinetics of hybridization of pulse-

labeled poly(A)-containing RNA isolated from nuclei of vege-

tative cells indicated that about 25-28% of the sequences

are complementary to reiterated DNA, and the remainder tosingle copy DNA. Further studies suggested that the re-

iterated transcripts are found predominantly at the 5' end ofnuclear poly(A)-containing RNA (17). Taken together withthe fact that most of the nuclear poly(A)-containing RNA is

conserved and is transported into the cytoplasm, the presentresults suggest that it is mainly transcripts of reiterated DNApresent-about 300 nucleotides long-in the nuclear RNA

molecules which are lost preferentially at some stage before

the RNA leaves the nucleus (13, 17).

Poly(A) in messenger RNA

With the exception of histone mRNA, which contains no

poly(A) (21), each mRNA in mammalian cells contains a

single sequence of poly(A) of about 150 bases in length (14,15). By contrast, Dictyostelium mRNA contains two types of

poly(A) sequences-one migrating on polyacrylamide gelsbetween 4S and 5S RNA and having a size of about 100

nucleotides, and one having a size of about 25 nucleotides

(Fig. 5) (22, 23). Nucleotide analysis of the putative poly-(A)loo and poly(A)25 labeled with 32p showed that they indeed

were poly(A) sequences (23).From the average molecular weight of mRNA (400,000)

and from the fraction of 32P-labeled material in mRNA which

is poly(A), we calculate that, on the average, each molecule of

mRNA contains one sequence of poly(A) of length 25, and

A 5S 4S BPB

I I I

f 1 ~~~~~~~~~~11

mRNA

HnRNA 0'

I tb

I

5106 Symposium: Lodish et al.

cSymposium: Dictyostelium Messenger RNA 5107

one of length about 100. Additional experiments showed thatat least half of the mRNA molecules contained both poly(A)sequences; the poly(A)1o0 is located, as in mammalian cells,at the 3' end, and the poly(A)25 sequence is located within 150nucleotides of the 3' end, separated from the poly(A)1oo byabout 10 to 40 nucleotides (23).

Pulse-labeled nuclear poly(A)-containing RNA also con-tains two classes of poly(A) sequences (Fig. 5). One class isabout 25 nucleotides long-precisely the same size as inmRNA. The other class is very heterogeneous and containssome sequences migrating at about 6 S [poly(A),5o] which arelarger than is found in mRNA, and also material smaller thanthat isolated from mRNA; possibly the intermediate (2-4S)sized material represents nascent, growing poly(A) chains.Based on calculations similar to those for mRNA, we concludethat each molecule of nuclear poly(A)-containing RNA (iso-lated from pulse-labeled cells) contains one sequence of poly-(A)25 while only one molecule in three contains a longer poly-(A) sequence (13, 23).Since all of the nuclear poly(A)-containing RNA molecules

are precursors to cytoplasmic mRNA molecules which con-tain both a poly(A)2z and poly(A)1o0 sequence, it would appearthat the poly(A)1oo sequence would be added after transcrip-tion of the RNA, as in mammalian cells. As is shown later,the poly(A)25 sequences are encoded in the nuclear DNA andare transcribed by RNA polymerase.The experiment in Fig. 6 shows that the poly(A)100 se-

quences are added just before the RNA exits from the nucleus.Following addition of [3H]adenosine to vegetative cells,radioactive poly(A)1oo sequences, attached to mRNA, appearin the cytoplasm after only a 2-min lag period. By contrast,labeled poly(A)25 sequences and labeled nonpoly(A) segmentsOL mRNA are found in the cytoplasm only after a 4-min lag.These results mean that it takes 4 min for a mRNA precursormolecule, containing a poly(A)25 sequence, to be processed andbe transported from the nucleus. The poly(A)1oo_1ro sequenceis added during the final 2-min period in the nucleus, perhapscoincidentally with the loss of the repetitive transcript at the5' end.Poly(dT)25 sequences in Dictyostelium DNADictyostelium DNA contains sufficient poly(dT) sequences ofapproximate size to encode the poly(A)25 sequences in mRNA(13, 23). To demonstrate this, 2P-labeled Dictyostelium DNAwas depurinated (24) and poly(dT) sequences were purifiedby poly(A)-Sepharose chromatography. In this fraction was0.3% of the DNA radioactivity. Fig. 5 shows that thesepresumptive poly(dT) sequences migrate with virtually thesame mobility as do the poly(A)25 sequences in mRNA. Thismaterial was further characterized by several techniques andfound to contain over 95% T residues (23). Furthermore, de-natured Dictyostelium DNA was hybridized to an excess of['H]poly(A). Poly(A) hybridized 0.33% of the nuclear DNA.The DNA ['H]poly(rA) hybrid showed a very broad thermalmelting profile with a Tm of 450 (13). This is 170 below thevalue of 620 obtained for long hybrids of poly(dT) . poly(rA)and is consistent with the notion that the sequences of hybridspoly(dT) - poly(A) are very short. This result, in itself,however would be consistent with mismatching. Since theamount of poly(dT) in Dityostelium DNA was the same(0.3%) whether determined by depurination (Fig. 5) orhybridization (13), we believe there is no extensive mismatch-ing in the poly(A) - poly(dT) hybrids.

O

_4 _

x

Io

0-joI_ o

. x

9,E

-a

TIME (MINUTES)

FIG. 6. Kinetics of appearance of poly(A)100 and poly(A)25 inthe cytoplasm. Vegetative cells were labeled with ['H]adenosine;at various times cytoplasmic RNA was isolated by the proteinaseK method (17). Poly(A)-containing RNA was purified by chro-matography on poly(U)-Sepharose. Total poly(A) was determinedby the fraction of the [3H] adenosine label which was resistant toRNases A and Ti; the fraction of poly(A) radioactivity whichrepresents poly(A),oo and poly(A)25 sequences was determined byelectrophoresis on 10% polyacrylamide gels as in Fig. 5A. Non-poly(A) sequences were the fraction of 3H-radioactivity whichwas sensitive to the RNases. All samples were normalized for agiven amount of rRNA present in the extracted samples. 0, totalpoly(A)-containing RNA; A, nonpoly(A) sequences; 0, poly-(A),oo; A, poly(A)25.From the above data and the nuclear genome size of

Dictyostelium (30 to 35 X 109 daltons, ref. 6), we calculatethat 3.6 X 105 dTMP residues per genome are in poly(dT)sequences. This is equivalent to 15,000 sequences of poly-(dT)25 per genome.The Dictyostelium genome could encode about 30,000 pro-

teins of average size 40,000 molecular weight. Since previouswork indicated that about half of the DNA is transcribed intomRNA (7), it appears that there is, on the average, onepoly(dT)25 sequence in the Dictyostelium genome per mRNAencoded. Additional hybridization experiments showed thatthe poly(dT) sequences are interspersed throughout thegenome, and not clustered in one region (23). Thus, it appearsthat most DNA sequences coding for mRNA terminate witha poly(dT) sequence, as is indicated schematically in Fig. 2.

Chromosome structure

The present results show clearly that in at least one eukaryoticprotist, Dictyostelium, the nuclear precursor of mRNA is onlyabout 20% larger than the mRNA itself. Our results stronglysuggest that the nuclear poly(A)-containing RNA of size500,000 daltons (mRNA precursor) is a primary transcriptionproduct, and is not derived by processing from a precursorwhich is significantly larger. We have not been able to recoverthe 5' triphosphate terminus of the putative primary tran-

Proc. Nat. Acad. Sci. USA 71 (1974)

Proc. Nat. Acad. Sci. USA 71 (1974)

script. However, when mRNA precursor is synthesized in iso-lated nuclei, under conditions where no degradation of RNAis detectable, the in vitro synthesized product is identical insize to its in vivo counterpart (22).The majority of nuclear poly(A)-containing RNA is ap-

parently conserved during processing and appears in the cyto-plasm as mRNA. About 75-80% of the sequences in nuclearpoly(A)-containing RNA are conserved and transported intothe cytoplasm, whereas the remaining material never leavesthe nucleus and is presumably destroyed.Our results indicate that about 25% of nuclear poly(A)-

containing RNA is transcribed from reiterated DNA. Further-more, the 5' ends of nuclear poly(A)-containing RNA areenriched in reiterated transcripts (17). Our results are mostconsistent with the notion that, before the bulk of nuclearpoly(A)-containing RNA is transported into the cytoplasm,about 300 nucleotides are removed from the 5' end. Thecleaved portion is greatly enriched in repetitive transcripts.The mechanism of synthesis of mRNA in Dictyostelium is

summarized in the model in Fig. 2. As synthesized initially,the mRNA contains a sequence of about 25 adenylic acidresidues near the 3' end. It is very likely that this poly(A)25sequence is encoded by the Dictyostelium DNA. Subsequent totranscription, a larger sequence of poly(A) is added at the 3'end. This poly(A)1oo is not encoded by the DNA, but thepoly(A)25 is most likely transcribed from the DNA by RNApolymerase. The poly(A)1o0 may be added coincident with theloss of the repetitive sequence at the 5' end, but the exactsequence and timing of poly(A) addition, loss of repetitivetranscripts, and transport of mRNA into the cytoplasm is notknown.

Previous hybridization experiments showed that over 56%of the single copy DNA is transcribed at some time duringvegetative growth or during development up to culmination(20 hr) (7). Our results showed that approximately 80% of thepulse-labeled nonribosomal RNA contains at least one poly-(A) sequence. We assume that a majority of single copy DNAsequences, themselves present in whole cell RNA, are alsotranscribed into RNA molecules which contain a poly(A)sequence.

Thus, all of the transcripts of single copy DNA contain, inthe same molecule, transcripts of repetitive DNA, with themajority of the repetitive sequences at the 5' end of themolecule.We conclude that a large fraction of the repetitive and

single copy DNA sequences are interspersed throughout thegenome. This is also illustrated by the model shown in Fig. 2.Repetitive sequences (R) of about 300 to 350 nucleotides areassociated with single copy (SC) sequences of average 1000 to1100 nucleotides. We believe that the 15,000 sequences ofpoly(dT)25 in the genome correspond to the 3' end of thetranscribed mRNA and that they may well represent a tran-scription termination signal.The unit RSC A25 is the primary transcription unit in

Dictyostelium (13, 23). We believe that since less than 100%of the genome has been found to be transcribed, there is a

spacer unit in between the primary transcription units, al-though we have no direct evidence for it. This chromosomemodel is very similar to the known size and organization of theinterspersed repetitive and single copy sequences in Xenopus(25) and is consistent with a similar analysis of the Dictyo-stelium genome (Firtel, unpublished data).

Since, by definition, each repetitive sequence R is present(on the average) in about 100 copies per genome, it is reason-able to assume that the same (or very similar) sequence mustbe found associated with about 100 single copy sequences, ifthey are all interspersed. We can speculate that these repeti-tive sequences are important in gene regulation and perhapsare analogous to promotor or operator sequences in pro-karyotes. A regulatory protein which binds to one such re-iterated sequence could be involved in turning on or off thetranscription of about 100 genes during differentiation. Sucha function of the reiterated DNA was first proposed byBritten and Davidson (26).

This research was supported by Grants GB-29308 and GB-42597 from the National Science Foundation. A.J. was a fellowof the Jane Coffin Childs Memorial Fund for Medical Research.R.F. was a fellow of the Helen Hay Whitney Foundation. T.A.is supported by a predoctoral fellowship from the NationalScience Foundation. H.F.L. is a recipient of Research CareerDevelopment Award 1 K04 GM-50175 from the U.S. NationalInstitutes of Health. We thank Mr. Ed. Loechler and Ms. LolleyGee for expert technical assistance.

1. Bonner, J. T. (1967) The Cellular Slime Molds (PrincetonUniversity Press, Princeton, N.J.), 2nd Ed.

2. Sussman, M. & Sussman, R. P. (1969) Symp. Soc. Gen.Microbiol. 19, 403-435.

3. Newell, P. C. (1971) Essays Biochem. 7, 87-126.4. Cocucci, S. M. & Sussman, M. (1970) J. Cell Biol. 45, 399-

407.5. Watts, D. J. & Ashworth, J. M. (1970) Biochem. J. 119,

171-174.6. Firtel, R. A. & Bonner, J. (1972) J. Mol. Biol. 66, 339-361.7. Firtel, R. A. (1972) J. Mol. Biol. 66, 363-377.8. Roth, R., Ashworth, J. M. & Sussman, M. (1968) Proc. Nat.

Acad. Sci. USA 59, 1234-1242.9. Franke, J. & Sussman, M. (1971) J. Biol. Chem. 246, 6381-

6388.10. Franke, J. & Sussman, M. (1973) J. Mol. Biol. 81, 173-185.11. Firtel, R. A., Baxter, L. & Lodish, H. F. (1973) J. Mol. Biol.

79, 315-327.12. Tuchman, J., Alton, T. & Lodish, H. F. (1974) Develop.

Biol., in press.13. Lodish, H. F., Firtel, R. & Jacobson, A. (1973) Cold Spring

Harbor Symp. Quant. Biol. 38, 899-914.14. Weinberg, R. A. (1973) Annu. Rev. Biochem. 42, 329-354.15. Darnell, J. E., Jelinek, W. R. & Molloy, G. R. (1973)

Science 181, 1215-1218.16. Firtel, R. A., Jacobson, A. & Lodish, H. F. (1972) Nature

New Biol. 239, 225-228.17. Firtel, R. A. & Lodish, H. F. (1973) J. Mol. Biol. 79, 295-

314.18. Sheiness, D. & Darnell, J. E. (1973) Nature New Biol. 241,

265-268.19. Iwabuchi, M., Mizukami, G. & Sameshima, M. (1971)

Biochim. Biophy8. Acta 228, 693-700.20. Verma, I., Firtel, R., Lodish, H. F. & Baltimore, D. (1974)

Biochemistry, in press.21. Adesnik, M. & Darnell, J. (1972) J. Mol. Biol. 62, 397-408.22. Jacobson, A., Firtel, R. A. & Lodish, H. F. (1974) J. Mol.

Biol. 82, 213-230.23. Jacobson, A., Firtel, R. & Lodish, H. F. (1974) Proc. Nat.

Acad. Sci. USA, in press.24. Burton, K. (1967) in Methods in Enzymology, eds. Grossman,

L. & Moldave, K. (Academic Press, New York), Vol. 12A,p. 222.

25. Davidson, E. H., Hough, B. R., Amenson, C. S. & Britten,R. J. (1973) J. Mol. Biol. 77, 1-23.

26. Britten, R. J. & Davidson, E. H. (1969) Science 165, 349-354.

27. Roberts, B. & Paterson, B. (1973) Proc. Nat. Acad. Sci.USA 70, 2330-2334.

28. Morrison, T., Stampfer, M., Baltimore, D. & Lodish, H. F.(1974) J. Virol. 13, 62-72.

5108 Symposium: Lodish et al.