JOURNAL OF BACTERIOLOGY, Aug. 1980, p. 981-9880021-9193/80/08-0981/08$02.00/0

Vol. 143, No. 2

Approach to Recognition of Regulatory Mutants ofCyanobacteria

GERALDINE HALL, ' MARYANN B. FLICK,' AND ROY A. JENSEN2*

Department of Biological Sciences' and Center for Somatic-cell Genetics and Biochemistry, 2 StateUniversity of New York at Binghamton, Binghamton, New York 13901

Antimetabolite analogs of essential amino acids are useful as selective agentsfor isolation of regulatory mutants of cyanobacteria, although we observed strikingmicrobiological differences from other widely used eubacterial systems. Regula-tory mutants shown to overproduce and excrete tryptophan, phenylalanine,tyrosine, methionine, or arginine were isolated from four cyanobacteria: Ana-baena sp. 29151, Synechococcus sp. 602, Synechococcus sp. AN Tx2O, andSynechocystis sp. 29108. Surprisingly, regulatory-mutant colonies did not supporta halo of cross-fed wild-type growth on selective medium. Since regulatorymutants were shown to excrete substantial levels of amino acids, it was deducedthat poor cross-feeding must reflect a generally low nutritional responsiveness ofthe cyanobacterial background. This conclusion was confirmed by results whichshowed that regulatory-mutant cells of cyanobacteria dispersed among wild-typepopulations of Bacillus subtilis did produce halo colonies on solid analog-con-taining medium. Cross-feeding between one cyanobacterial pair (a phenylalanineexcretor and a phenylalanine auxotroph) was successfully demonstrated in theabsence of the analog under conditions in which relatively large masses of eachcell population type were spread near one another on agar plates. These resultssuggest that amino acid excreted by regulatory mutants of cyanobacteria on

analog-containing selective medium is transported into nearby wild-type cells tooinefficiently to overcome the antimetabolite effects of the analog, thereby failingto generate halos of physiologically resistant background cells. Consistent withthis interpretation was the finding that thepheAl auxotroph from Synechococcussp. 602 exhibited a linearly proportional dependence of growth rate upon exoge-nous concentration of L-phenylalanine (below 20 PM). Wild-type B. subtilis servesas a convenient and sensitive test lawn for screening obvious regulatory mutantsfrom among collections of analog-resistant cyanobacterial mutants. AppropriateB. subtilis auxotrophs can be used as convenient indicator strains for the identi-fication of regulatory mutants in cyanobacteria through the observation of syn-trophic growth responses.

Biochemical perturbations produced by regu-latory mutations in bacteria have offered ahighly successful approach to the understandingof regulatory mechanisms, enzyme structure,and biosynthetic pathways (6, 8, 16) and to theelucidation of interactions between biosyntheticpathways (10). The ability of analogs to mimicnatural end products has been a familiar basisfor the selection of regulatory mutants recog-nized as a fraction of analog-resistant individualsin a sensitive population. Analog-resistant mu-tants which are defective in the regulation ofamino acid biosynthesis have been used to iden-tify those enzyme activities subject to feedbackinhibition and enzyme repression. Regulatorymutants are commonly sorted from other classesof resistant mutants by selection ofhalo colonies.The latter colonial phenotype results from over-production and excretion of a natural end prod-

uct which then physiologically relieves the an-timetabolite effect, stimulating growth of a back-ground halo of wild-type cells around mutantcolonies on the original selection plates (1, 5).The identity of an excreted end product canquickly be established by the ability to cross-feed a particular auxotrophic strain (1). Pub-lished descriptions of regulatory mutants in cy-anobacteria are virtually nil. An isolated exam-ple was with Anacystis nidulans mutants inwhich relatively modest quantitative differencesin the sensitivity of 3-deoxy-D-arabinoheptulo-sonate-7-phosphate synthase to feedback inhi-bition by L-tyrosine were observed (14). Thesemutants were not characterized as excretors ofaromatic end products.Our initial experiments, in which sensitive

wild-type cyanobacteria were plated on analog-containing medium, produced surprising results.

981

on April 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

982 HALL, FLICK, AND JENSEN

Although analog-resistant colonies were readilyobtained as spontaneous mutants, colonies sur-rounded by halos of cross-fed background cellswere never seen. In organisms such as Bacillussubtilis or Escherichia coli, excreting regulatorymutants are readily discerned from other classesof resistant mutants, (e.g., transport-deficientmutants) on the original selection plate by ob-serving the cross- feeding of sensitive back-ground cells, manifested as a gradient of growtharound the resistant colony (halo).A priori, this could mean that regulation of

biosynthetic pathways in cyanobacteria is sounspectacular that loss of regulation in mutantsgenerates only minor quantitative increases inpathway flow, that analog selection for someunknown reason is unsuitable for recognition ofregulatory mutants, or that regulatory mutantsare in fact selected by analog resistance but donot cross-feed background cells because of feebletransport of excreted molecules into analog-sen-sitive background cells. The last explanation issupported by our results.

MATERIALS AND METHODS

Microorganisms. All cyanobacterial strains usedexcept Anabaena sp. are unicellular organisms whichreadily plate out as colony-forming units on solidmedium. Synechocystis sp. ATCC 29108 and Ana-baena sp. ATCC 29151 were obtained from the Amer-ican Type Culture Collection, Rockville, Md. Synecho-coccus sp. 602 and Synechococcus sp. 602 pheAl, aphenylalanine auxotroph, were generous gifts of S. V.Shestakov. Synechococcus sp. AN Tx2O (commonlydenoted as Anacystis nidulans in the literature) isATCC 27144. All wild-type cyanobacterial strains wereroutinely maintained on CglO medium (20) solidifiedwith 1.4% (wt/vol) agar (Difco) sterilized separately.Analog-resistant mutants were maintained on CglOmedium supplemented with 100lOg of the appropriateanalog per ml.The prototrophic B. subtilis strain NP1 is a deriv-

ative of strain 168 (15). The auxotrophic B. subtilisstrains used, all derived from strain 168, and theiramino acid requirements were: E-7, tryptophan; E-34,tyrosine; E-62, phenylalanine; A-140, histidine; A-87,methionine; and A-54, arginine. B. subtilis F-2 is aconstitutive regulatory mutant which is resistant to 5-methyltryptophan and is a heavy excretor of trypto-phan (11).

Selection of mutants resistant to amino acidanalogs. Spontaneous analog-resistant mutants wereselected by either (i) placing a few crystals of analogin the middle of CglO agar medium spread with a lawnof 106 to 108 cells or (ii) spreading 106 to 108 cells on aplate of CglO medium containing 100 ,ug of analog perml. Plates were incubated at 34°C under constantillumination. Resistant colonies appeared within 7 to10 days and were then purified by single-colony isola-tion on analog-containing medium. Isolated single col-onies were checked for phenotypic stability and excre-tion.

Determination of auxotrophic growth re-sponse to L-phenylalanine. Cells of Synechococcussp. 602 pheAl were grown in Dm medium (19) supple-mented with 0.5 mM L-phenylalanine. A 10-ml sampleof actively dividing cells (optical density at 500 nm,0.6) was washed twice with 10-ml portions of sterileunsupplemented Dm medium via cell capture on a0.45-,um membrane filter (Millipore Corp.). After beingwashed, the cells were suspended in 10 ml of sterilemedium. A 1.0-ml sample of washed cells was addedto each of seven 75-ml test tubes containing 50 ml ofDm medium supplemented with various amounts offilter-sterilized L-phenylalanine. The growth vesselswere incubated in a 34°C water bath illuminated withfluorescent lights. Care was taken to ensure that alltubes received equal light exposure. Growth vesselswere aerated with a 2% (vol/vol) mixture of C02-air.Samples were aseptically removed at periodic inter-vals, and the percent transmittance values were deter-mined with a Bausch & Lomb Spectronic 20 colorim-eter at both 500 and 600 nm. The values recorded wereconverted to optical density units. Growth curves ob-tained by using values at 500 and 600 nm were com-parable. At these wavelengths, the contribution of thepigments is less than 10% of total absorbancy. Sampleshaving optical densities greater than 0.4 were appro-priately diluted to maintain the range in which opticaldensity is proportional to turbidity.The wild type grows exponentially, and accurate

doubling times are conveniently obtained in the opti-cal density range of 0.1 to 1.0. At turbidity valuesabove 1.0, shading effects complicate growth rate de-terminations. Growth rates of phenylalanine auxo-troph Synechococcus sp. 602 pheAl were determinedthrough the 40- to 88-h growth period after the initiallag phase and were calculated as an increment inoptical density per hour rather than a doubling timebecause of the linear nature of the growth curve (seeFig. 3).

Routine identification of regulatory mutants.A biologically sensitive test to identify amino acidexcretors in cyanobacterial strains is a heterologoussystem employing a confluent lawn of B. subtilis. B.subtilis grows readily on CglO medium in the lightwhen 0.5% glucose is present as a carbon source.Mixtures of prototrophic populations of cyanobacteriaand B. subtilis were compatible during growth. Insome cases, auxotrophic lawns of B. subtilis were used'for detection of halo growth responses. In other cases,wild-type (prototrophic) lawns sensitive to the analogpresent in the medium were used to detect potentialhalo responses.

Analytical techniques. The assay of culture fil-trates for protein was done by the method of Bradford(3). Quantitative assays for excreted aromatic aminoacids were carried out as follows. Actively growingcultures (optical density at 500 nm, 1.0 to 4.5), incu-bated in liquid CglO medium at 34°C and aerated witha 2% (vol/vol) C02-air mixture, were centrifuged at10,000 x g for 5 min in a Sorvall model RC 2-Bcentrifuge. Portions of the resulting supernatants werefiltered through a 0.45-um Millipore filter, and thefiltrate was immediately frozen at -15°C until assayed.Phenylalanine was measured by the fluorescence assayof McCaman and Robins as modified by Wong et al.

J. BACTERIOL.

on April 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

REGULATORY MUTANTS OF CYANOBACTERIA 983

(21), except that 50 ,ul of 25% trichloroacetic acid wasadded to a 200-pl sample to reduce excessive sampledilution, 1.0 ml of copper reagent was added, andrelative fluorescence was measured at an emissionwavelength of 470 nm. Tyrosine was determined bythe fluorescence assay of Wong et al. (21), except that50 ,ul of 20% trichloroacetic acid was added to a 200-pd sample and 200 ,ul of a-nitroso-fB-naphthol reagentwas added to 50 ,ul of trichloroacetic acid supernatant.Tryptophan was measured by the fluorescent decayassay of Guilbault and Froelich (9). All fluorescencemeasurements were performed with an Aminco-Bow-man spectrophotofluorometer.

Biochemicals. All compounds used were reagentgrade or of the highest purity available. All analogswere obtained from Sigma Chemical Co., except 3-methylhistidine (Calbiochem) and N-acetylleucine(Schwartz/Mann).

RESULTS

Isolation of regulatory mutants. Four cy-anobacterial species were screened for sensitiv-ity to a number of amino acid analogs. Analogsof arginine, methionine, phenylalanine, trypto-phan, and tyrosine inhibited growth of one ormore strains sufficiently to allow the selection ofpotential regulatory mutants, expected to pos-sess analog resistance phenotypes. The mostuseful selective agents found are listed in Table1. Other analogs tested but found to have littleor no inhibitory effect on any of the speciesscreened include 2-methylleucine, N-methylleu-cine, a-methylmethionine, norleucine, norvaline,a-methyl-m-tyrosine, 5-methyltryptophan, andD-tyrosine. 3-Methylhistidine and 3-aminotyro-sine were found to inhibit growth of some of thestrains tested, but spontaneous resistant mu-tants were not recovered.Recognition ofregulatory mutants. It was

immediately apparent that although various an-

alog-resistant colonies were easily obtained,none of them supported a halo of cross-fed back-ground growth. This is illustrated by the pho-tograph in Fig. 1A showing resistant colonies ofSynechocystis sp. 29108 arising on 4-fluorophen-

ylalanine-containing medium. Even when resist-ant mutants were purified and spread at greatercell mass onto wild-type lawns on 4-fluorophen-ylalanine agar, cross-fed halos did not arise (Fig.1B). On the other hand, many of the resistantmutants initially selected (circled in Fig. 1A)proved to be phenylalanine excretors since eachcross-fed a lawn of the auxotrophic Synechococ-cus sp. 602pheAl on minimal medium (Fig. 1C).Control platings established the inability ofwild-type Synechococcus sp. 29108 to cross-feedSynechococcus sp. 602 pheAl even after pro-longed incubation.

Similar experiments to recognize the possibleexcretion of amino acids by other analog-resist-ant mutants were not initially feasible due to thelack of appropriate cyanobacterial auxotrophs(12). However, B. subtilis NP1 was found to becapable of growth on CglO medium supple-mented with glucose, and auxotrophic strainsderived from B. subtilis NP1 were suitable forrapid identification of cyanobacterial regulatorymutants (i.e., wild-type cyanobacteria did notfeed B. subtilis auxotrophs, whereas analog-re-sistant cyanobacterial excretors produced typi-cal halos within background lawns of the appro-priate auxotrophs). For example, the heterolo-gous combination of cyanobacterial tryptophanexcretor and B. subtilis tryptophan auxotrophexhibited cross-feeding (Fig. 2A). The same cy-anobacterial tryptophan excretors successfullycross-fed in heterologous combination with theB. subtilis wild type on 6-fluorotryptophan (Fig.2B). Similarly, the heterologous combination ofcyanobacterial phenylalanine excretor and wild-type B. subtilis on 4-fluorophenylalanine ex-

hibited cross-feeding (Fig. 2C) in contrast to thenegative results obtained in Fig. 1B with thehomologous combination of cyanobacteria.When cyanobacterial excretors of tryptophan(Fig. 2D) or of phenylalanine (Fig. 2E and F)were spread onto analog-containing agar spreadwith heavy cyanobacterial lawns, no cross-feed-

TABLE 1. Selection of analog-resistant mutants of cyanobacteriaRegulatory mutants'/total resistant mutants

Arnaog Anabaena Synechocystis p. Synecho- Synechococcus sp T

sp. 29151 29108 COCCUS Sp. AN Tx2O Total602

Canavanine NIb 2/4 0/0 0/0 2/4Ethionine 6/15 No inhibition 2/7 0/2 8/242-Fluorophenylalanine 6/6 1/7 No inhibition 7/133-Fluorophenylalanine 3/3 0/0 17/19 20/224-Fluorophenylalanine 2/2 10/10 0/0 0/0 12/126-Fluorotryptophan 8/10 0/0 2/17 0/11 10/384-Aminophenylalanine 0/10 0/10

a Defined as the ability to carry out syntrophic cross-feeding of appropriate auxotrophs of B. subtilis.b NI, no inhibition.

VOL. 143, 1980

on April 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

984 HALL, FLICK, AND JENSEN J. BACTERIOL.

*. 4.^

** C

C. * * *A.

* C

A * B CFIG. 1. Examination of phenylalanine cross-feeding in homologous combinations of cyanobacteria. (A)

Photograph of resistant colonies selected from a confluent lawn of Synechocystis sp. 29108 cultured on 4-fluorophenylalanine-containing medium. Circled colonies were isolated, purified, and tested for ability tocross-feed a wild-type lawn of Synechocystis sp. on analog-containing medium (B) a lawn ofphenylalanineauxotroph pheAl on CglO medium (C).

L da~~~~~

I.... .......

.......~~ ~ ~ ~ -'-- ,.

FIG. 2. Comparison ofcross-feeding results in homologous and heterologous combinations ofcyanobacterudand B. subtilis. Populations of the following were spread with an inoculating loop onto an agar surfacepreviously spread with the indieated confluent lawn: (a) B. subtilis tryptophan excretor F-2, (b) wild-typeAnabaena sp. 29151, (c) Anabaena sp. tr.yptophan excretor 6FT2, (d) Anabaena sp. tryptophan excretor 6FT1,(e) Synechocystis sp. phenylalanine excretor 4FP30, (f) Anabaena sp. phenylalanine excretor 4FPI, (g)Anabaena sp.phenylalanine excretor4FP2, and (h) Synechocystis sp.phenylalanine excretor4FP3l. Confluentlawns corresponding with each panel shown are: (A) B. subtilis tryptophan auxotroph E-7 on Cgl1O minimalmedium, (B) B. subtilis wild-type NPI on 6-fluorotryptophan medium, (C) B. subtilis wild-type NP) on 4-fluorophenylalanine medium, (D) Anabaena sp. 29151 wild type on 6-fluorotryptophan medium, (E) Anabaenasp. 29151 wild type on 4-fluorophenylalanine medium, and (iF) Synechocystis sp. 29108 on 4-fluorophenylalaninemedium.

on April 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

REGULATORY MUTANTS OF CYANOBACTERIA 985

ing between these homologous-combinationpairs was observed.The approach of using heterologous combi-

nations of B. subtilis auxotrophic lawns andpotential cyanobacterial amino acid excretorswas exploited to screen for various classes ofregulatory mutants. A very large fraction of theanalog-resistant mutants isolated in four cyano-bacterial strains were regulatory mutants (Table1). Only obvious excretors were scored as regu-latory mutants, and it is possible that some ofthe nonexcreting analog-resistant mutants werealso regulatory mutants.Amino acid excretion by regulatory mu-

tants. Quantitative measurements ofamino acidexcretion were made in seven mutants repre-senting four organisms (Table 2). Strains weregrown in liquid culture, and the extracellulargrowth fluid was assayed as described above.Growth supernatants did not contain detectablelevels of protein. The wild-type organisms testedaccumulated little or no phenylalanine in the

TABLE 2. Extracellular excretion of amino acids byregulatory mutants of cyanobacteria

Amino acidexcretiona

Cyanobacterial Analog resistancespecies phenotype Phen- Tryp

nine tophan

Anabaena sp. 29151 Sensitiveb 19 0Mutant 4FP1 4-Fluorophenyl- 64

alanineMutant 4FP2 4-Fluorophenyl- 39

alanineMutant 6FT2 6-Fluorotryptophan 122

Synechocystis sp. Sensitive 0 129108c

Mutant 4FP30 4-Fluorophenyl- 134 19alanine

Mutant 4FP31 4-Fluorophenyl- 133 21alanine

Synechococcus sp. Sensitive 5602

Mutant 2FP40 2-Fluorophenyl- 26alanine

Synechococcus AN Sensitive 4Tx20

Mutant 3FP70 3-Fluorophenyl- 125alanine

aExpressed as micromolar concentration per unit ofgrowthmass (optical density at 500 nm).

bThe top line of data shown for each of the four speciesstudied corresponds to the wild-type parent (sensitive togrowth inhibition by the analog compounds indicated) fromwhich the mutants indicated were derived by selection forresistance to inhibitory effects of the indicated analog.

'The excretion of tyrosine in Synechocystis sp. 29108 wasless than 1 ,uM, whereas that of 4FP30 and 4FP31 was 13 and8 ,uM, respectively.

growth medium, compared with the elevatedlevels found in regulatory mutants resistant tothe indicated fluoro derivatives of phenylala-nine. The most dramatic instance was with 4-fluorophenylalanine-resistant mutants of Syne-chocystis sp. 29108. The two mutants, 4FP30and 4FP31, also overproduce tryptophan andtyrosine and have been shown to possess a mu-tant 3-deoxy-D-arabinoheptulosonate-7-phos-phate synthase that is desensitized to allostericcontrol (Hall and Jensen, unpublished data).Mutant 6FT2, a tryptophan excretor from An-abaena sp. 29151, illustrates another obviousexample of a regulatory mutation.Nature of analog inhibition of growth.

Since homologous combinations of cyanobacter-ial pairs cross-fed in the absence of an analog(excretors and auxotrophic lawns) but not in thepresence of an analog (excretors and proto-trophic lawns on analog-containing medium),the possibility that analog exposure was bacte-ricidal was tested. Lawns of Synechocystis sp.29108 and Anabaena sp. 29151 were spread toconfluence on CglO medium containing 100 ,tgof 4-fluorophenylalanine per ml, perhaps themost inhibitory analog tested. The agar plateswere incubated at 340C under constant illumi-nation. Each day over a period of 1 week, 0.1 mlof a 20mM mixture of the three aromatic aminoacids was added to a well cut in the agar mediumwith a cork borer. In both species, continuedviability after 4 days of analog inhibition wasdemonstrated by observation of good growtharound each amino acid-containing well. Hence,the primary analog effect appears to be bacteri-ostatic, as in other bacteria. Generally, recoveryof viable cyanobacteria was minimal after 7 daysof analog exposure.Relationship of amino acid transport.

When single cells of the phenylanine-excretingmutants were plated on lawns of phenylalanineauxotroph Synechococcus sp. 602 pheAl, cross-feeding was not apparent until 4 to 6 days afterappearance of the excreting mutant colonies.This delayed response suggested feeble trans-port. Accordingly, it seemed likely that a lack ofactive transport, or a very low level of transport,might explain the relatively poor cross-feedingresults with homologous combinations of cyano-bacteria on analog-containing medium. Thiswould be consistent with the finding of back-ground halos of B. subtilis, which has an excel-lent transport system (2, 7), around cyanobac-terial regulatory mutants on analog containingmedium, in contrast to the failure ofhomologouscyanobacterial combinations to cross-feed.When the transport of L-["4C]phenylalanine intocells of Agmenellum quadruplicatum or Syne-

VOL. 143, 1980

on April 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

986 HALL, FLICK, AND JENSEN

chococcus sp. AN Tx2O was compared with B.subtilis under the conditions specified byD'Ambrosio et al (7), transport rates in cyano-bacteria were indeed found to be roughly twoorders of magnitude less than those of B. sub-tilis. At 50 ,uM L-['4C]phenylalanine, 0.2 nmol ofamino acid was transported in 5 min per mg ofcell mass in Synechococcus sp. AN Tx2O com-pared with 29 nmol/mg per 5 min in B. subtilis.At 5 ,uM L-['4C]phenylalanine the transport ratein B. subtilis was 8 nmol/mg per 5 min comparedwith a rate of <0.05 nmol/mg per 5 min inSynechococcus sp. AN Tx2O.

If amino acid transport were sufficiently defi-cient, one might predict a direct dependence ofgrowth rate of cyanobacterial auxotroph uponexogenous amino acid present at low concentra-tions. In organisms with an active transportmechanism, the uptake of low concentrations ofamino acids required by auxotrophs is adequateto maintain the wild-type growth rate, even in arange that distinctly limits the ultimate massyield. The family of growth curves shown in Fig.3 reveals that differing rates of growth wereindeed obtained at different exogenous concen-trations of L-phenylalanine supplied in thegrowth medium to a cyanobacterial phenylala-nine auxotroph. It was convenient to use linearcoordinates to represent growth rates at initialL-phenylalanine concentrations lower than 25uM since the kinetics of growth in this concen-tration range approximate linearity. At concen-trations of L-phenylalanine greater than 100mM, growth was exponential and equaled the 9-h doubling time of the wild type. Figure 4 is aplot of the data points obtained when growthrate (determined from the slope of each curve inFig. 3) is plotted as a function of the concentra-tion of exogenous L-phenylalanine initially sup-plied. A linear proportionality of growth rateand L-phenylalanine concentration was observedbelow 15 ,uM. In E. coli or B. subtilis, a compa-rable manipulation of growth rate through ad-justment of exogenous L-phenylalanine can onlybe achieved in the chemostat.

DISCUSSIONAnalogs of metabolites such as amino acids

inhibit microbial growth by acting as end prod-uct mimics. Antimetabolite effects are readilyreversed in the presence of the natural end prod-uct (1). Thus, regulatory mutants which over-produce end product are selected as resistantcolonies on analog-containing medium. One ex-pects that wild-type cells contiguous to a regu-latory mutant colony will be phenotypically re-sistant because of the presence of excreted endproduct. Halo formation around colonies of an-

J. BACTERIOL.

20pM

0.9

14jMM0.8

0.7

E

20.6 100MM

phA> safucino inrasn cocntaionm o

0.5w

4 6 MMC.) 0.4

0

0.3

0.2

2;MM

0.1

020 40 60 s0 100 120

HOURS

FIG. 3. Growth response of Synechococcus 8p. 602pheAl as a function of increasing concentration-s ofexogenous L-phenylalanine. Within the range of op-tical density shown, the parental wild ty)pe growsexponentially at a doubling time of 9 h.

uLI 16'4c 12

80CC0

5 10 15 20

,uM L-PHENYLALANINEFIG. 4. Proportional relationship of growth rate

and exogenous L -phenylalanine concentration. The:growth rates on the ordinate scale are expressed asincrement ofoptical density at 5MX nmper hour x 103.

alog-resistant mutants has apparently neverbeen described in cyanobacteria. This couldmean that regulatory mechanisms (and there-fore the possibility of genetic deregulation) arerather undeveloped or even nonexistent formany biosynthetic pathways in cyanobacteria.Repression control of biosynthetic pathways incyanobacteria does indeed appear to be minimal(4) as in fact is the case with many other nonen-teric bacteria (17). However, allosteric control of

I I I I

on April 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

REGULATORY MUTANTS OF CYANOBACTERIA 987

regulatory enzymes has been found in most, ifnot all, biosynthetic pathways where it has beensought.

In this study analog-resistant mutants of cy-anobacteria were shown to excrete levels of endproduct that were qualitatively similar to mu-tants isolated in other procaryotes such as B.subtilis. On original selection medium the for-mer mutants do not cross-feed the sensitivebackground. However, the ability of cyanobac-terial regulatory mutants to cross-feed wild-typelawns of analog-sensitive B. subtilis on analog-containing medium eliminates the possibilitythat the presence of analog somehow diminishesexcretion of end product in cyanobacteria. Theunexpected obstacle to rapid identification ofexcreting mutants, i.e., lack of wild-type growthhalos on selection plates, was overcome by theuse of B. subtilis tester lawns which proved tobe growth compatible with the cyanobacterialstrains utilized. It is likely that other species ofbacteria capable of growth on minimal salts me-dium supplemented with glucose, such as E. coli,will also prove to be growth compatible withspecies of cyanobacteria, thus making availablea wide array of auxotrophic indicator strains forthe identification of regulatory mutants in cy-anobacteria. Many of the regulatory mutantsproved to be stable through several transfers onnonselective medium and so should provide use-ful genetic markers.The results of cross-feeding experiments with

a variety of B. subtilis-cyanobacterial heterolo-gous combinations in comparison with homolo-gous combinations reveal that cyanobacteriallawns respond poorly to excreted end product.Most likely this is the consequence of weaktransport capability in cyanobacteria, shown tobe qualitatively poor relative to B. subtilis. Thelinear dependence of the growth rate of Syne-chococcus sp. 602 pheAl upon the exogenousconcentration of L-phenylalanine further sug-gests the absence of active transport for L-phenylalanine. In various species of autotrophicbacteria (including Synechococcus sp. AN Tx2O)the rate of amino acid assimilation was found tovary proportionally with exogenous concentra-tion, a finding interpreted to reflect generallypoor transport of amino acids (17). The absenceof sugar permeases has been postulated to ex-plain the failure of obligately photoautotrophiccyanobacteria to grow in the dark at the expenseof various sugars, including glucose (18). A com-parative study of glucose utilization by Apha-nocapsa sp. 6714, a facultative photo- and che-moheterotroph, with glucose utilization by twoobligate photoautotrophs, Synechococcus sp.6301 and Aphanocapsa sp. 6308, indicated that

the failure of the latter two organisms to growwith glucose was due to the absence of an effec-tive glucose permease (13). It would not be sur-prising, therefore, to find that many, if not all,cyanobacteria lack permeases for amino acids.

In summary, one can envision the followingscenario of events during the emergence of acyanobacterial regulatory-mutant colony on an-alog-containing selective medium. Initially, verylittle excreted metabolite is elaborated becauseof the small cell population within the growingclone. Nearby wild-type cells must be able totransport via a high-affinity system to scavengethe low concentrations of metabolite present.The metabolite analog, present at high concen-tration, will most likely exacerbate transportproblems in a competitive fashion. Relativelyslow growth of cyanobacteria will result ingreater opportunity for diffusion of excreted me-tabolite through the agar relative to generationtime, thereby effectively diluting the concentra-tion of metabolite encountered by backgroundcells. In contrast, the efficient high-affinitytransport of amino acids such as phenylalaninein B. subtilis provides a sensitive response toexcreted metabolites even under plating condi-tions in which a single cell generates a colonysupporting a cross-fed halo.

ACKNOWLEDGMENTSThis investigation was supported in part by Public Health

Service research grant AM-19447 from the National Institutesof Health and in part by funds from the Center for Somatic-cell Genetics and Biochemistry.We gratefully acknowledge the help of Joan Markiewicz

with the illustrative art work.

LITERATURE CITED1. Adelberg, E. A. 1958. Selection of bacterial mutants

which excrete antagonists of antimetabolites. J. Bacte-riol. 76:326.

2. Bernlohr, R. W. 1967. Changes in amino acid permeationduring sporulation. J. Bacteriol. 93:1031-1044.

3. Bradford, M. M. 1976. A rapid and sensitive method forthe quantitation of microgram quantities of proteinutilizing the principle of protein-dye binding. Anal. Bio-chem. 72:248-254.

4. Carr, N. G. 1973. Metabolic control and autotrophicphysiology, p. 39-65. In N. G. Carr and B. A. Whitton(ed.), The biology of blue-green algae. Blackwell Sci-entific Publications, Oxford.

5. Champney, W. S., and R. A. Jensen. 1969. D-Tyrosineas a metabolic inhibitor of Bacillus subtilis. J. Bacte-riol. 98:205-214.

6. Cohen, G. N. 1965. Regulation of enzyme activity inmicroorganisms. Annu. Rev. Microbiol. 95:105-126.

7. D'Ambrosio, S. M., G. I. Glover, S. 0. Nelson, and R.A. Jensen. 1973. Specificity of the tyrosine-phenylala-nine transport system in Bacillus subtilis. J. Bacteriol.115:673-681.

8. Fowden, L., D. Lewis, and H. Tristram. 1967. Toxicamino acids: their action as antimetabolites, p. 89-163.In F. F. Nord (ed.), Advances in enzymology, vol. 29.John Wiley & Sons, Inc., New York.

9. Guilbault, G. G., and P. M. Froelich. 1973. Rapid,

VOL. 143, 1980

on April 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

988 HALL, FLICK, AND JENSEN

sensitive and selective assay for tryptophan in serum.

Clin. Chem. 19:112-113.10. Jensen, R. A. 1969. Metabolic interlock: regulatory inter-

actions exerted between biochemical pathways. J. Biol.Chem. 244:2816-2823.

11. Kane, J. F., and R. A. Jensen. 1970. Metabolic interlock:the influence of histidine on tryptophan biosynthesis inBacillus subtilis. J. Biol. Chem. 245:2384-2390.

12. Ladha, J. K., and H. D. Kumar. 1978. Genetics of blue-green algae. Biol. Rev. 53:355-386.

13. Pelroy, R. A., R. Rippka, and R. Y. Stanier. 1972.Metabolism of glucose by unicellular blue-green algae.Arch. Microbiol. 87:303-322.

14. Phares, W., and L. F. Chapman. 1975. Anacystis nid-ulans mutants resistant to aromatic amino acid ana-

logues. J. Bacteriol. 122:943-948.15. Rebello, J. L., and R. A. Jensen. 1970. Metabolic inter-

lock: the multimetabolite control of prephenate dehy-dratase activity in Bacillus subtilis. J. Biol. Chem. 245:

3738-3744.16. Richmond, M. H. 1962. The effect of amino-acid ana-

logues on growth and protein synthesis in microorgan-isms. Bacteriol. Rev. 26:398-420.

17. Smith, A. J., and D. S. Hoare. 1977. Specialist photo-trophs, lithotrophs, and methylotrophs: a unity amonga diversity of procaryotes? Bacteriol. Rev. 41:419-448.

18. Stanier, R. Y. 1973. Autotrophy and heterotrophy inunicellular blue3green algae, p. 501-518. In N. G. Carrand B. A. Whitton (ed.), The biology of blue-greenalgae. Blackwell Scientific Publications, Oxford.

19. Van Baalen, C. 1965. Quantitative surface plating ofcoccoid blue-green algae. J. Phycol. 1:19-22.

20. Van Baalen, C. 1967. Further observations on growth ofsingle cells of coccoid blue-green algae. J. Phycol. 3:154-157.

21. Wong, P. W. K., M. E. O'Flynn, and T. Inouye. 1964.Micromethods for measuring phenylalanine and tyro-sine in serum. Clin. Chem. 10:1098-1104.

J. BACTERIOL.

on April 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from