Vol. 137, No. 1JOURNAL OF BACTERIOLOGY, Jan. 1979, p. 295-3000021-9193/79/01-0295/06$02.00/0

Rifampin-Resistant Mutants of Myxococcus xanthusDefective in Development

KENNETH RUDD AND DAVID R. ZUSMAN*

Department ofBacteriology and Immunology, University of California, Berkeley, California 94720

Received for publication 6 September 1978

Rifampin, an antibiotic which is known to bind to and inhibit RNA polymerase,was used to probe the molecular regulation of development in Myxococcusxanthus. Rifampin-resistant mutants were screened for defects in fruiting-bodyformation. About 20% of the isolates in the initial screenings showed major defectsin developmental aggregation or sporulation. Eleven independent mutants withwild-type growth rates and stable phenotypes were analyzed by transduction. Inthese strains, the rifampin-resistant and nonfruiting phenotypes showed cotrans-duction frequencies equal to or greater than 99.0 to 99.9%. The RNA polymeraseactivities were resistant to rifampin in vitro, indicating that the RNA polymeraseis altered in these strains. Although their fruiting phenotypes are heterogeneous,these strains can be divided into two classes based on the level of aggregation.The results suggest that RNA polymerase plays a significant role in the regulationof development in M. xanthus since mutations which cause no apparent changesin vegetative growth result in striking defects in fruiting-body formation.

Myxococcus xanthus is a gram-negative bac-terium which grows vegetatively in soils on de-caying organic material, on the bark of livingtrees, or by preying upon other microorganisms(7). When nutrients are depleted from a solidculture medium, cells aggregate to form mounds.Within the mounds, the rod-shaped vegetativecells convert to round or ovoid spores. Moundsof mature myxospores are referred to as fruitingbodies,We have recently investigated protein synthe-

sis during aggregation and fruiting-body forma-tion in M. xanthus and have found that manyproteins showed significant changes in theiramounts and patterns of synthesis during devel-opment (M. Inouye, S. Inouye, apd D. R. Zus-man, Dev. Biol., in press). These results suggesta complex but precise program of gene expres-sion. It is likely that the synthesis of at leastsome of these proteins is regulated at the tran-scriptional level. To study this possibility, wesearched for developmental mutants with de-fects in RNA polymerase. RNA polymerase mu-tations are easily obtained in many bacteria byselecting for resistance to the antibiotic rifampin(6, 10, 11, 15), a specific inhibitor which is knownto bind to the ,B subunit of RNA polymerase (11,14). We found that a significant class of theserifampin-resistant mutants in M. xanthus were,in fact, defective in fruiting-body formation; inat least 11 of these mutants the two phenotypesare probably caused by a single mutation. Theheterogeneous terminal phenotypes observed inthese mutants suggests that transcriptional spec-

ificities may be altered in several different waysduring development.

MATERIALS AND METHODSBacterial and bacteriophage strains. Three dif-

ferent strains that are "wild type" for fruiting-bodyformation were used: DZ2 (5), DZF1 (FB) obtainedfrom D. Kaiser, who obtained it from M. Dworkin, andDZF6 (4). DZ1 (18), a nonfruiting strain, was used asa phage indicator. The phage strain MX4ts27htfhfrm(4) was used for all tranaductions.Media and cultural conditions. Bacteria were

routinely grown in CT broth (4), CYE broth (4), orTYE broth (CYE broth which contains Trypticase[BBL] instead of Casitone [Difco]). Cultures weregrown at 300C with gyratory shaking at 175 to 200rpm. Fruiting-body formation was induced by inocu-lating cells on coli agar (plastes containing 8 x 108cells of autoclaved Escherichia coli HfrH per ml, 0.1%MgS04 *7H20, and 1.6% Difco agar) or on CF agar (2,9).

Isolation of rifampin-resistant mutants. Spon-taneous rifampin-resistant (Rif') mutants were ob-tained by plating rifampin-sensitive strains on CF, CT,CYE, or TYE agar containing rifampin (25 ,ug/ml;Sigma Chemical Co.). Single colonies were repurifiedon CT, CYE, or TYE agar without rifampin. Whenthe use of independent cultures is indicated, eachculture was grown from a single colony to midlogphase in the absence of rifampin and then plated onmedia containing rifampin (25,ug/ml).Phenotypic analysis of Rifr mutants. Rifampin-

resistant strains were grown on CT, CYE, orTYE agarfor 3 to 6 days and then transferred by toothpick tocoli agar to test for fruiting-body formation. After atleast 10 days at 300C, fruiting phenotypes were ex-

295

on April 6, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

296 RUDD AND ZUSMAN

amined under a dissecting microscope. Strains whichfailed to form fruiting bodies on three subsequent testswere considered stable mutants. Fruiting phenotypeswere analyzed further by harvesting exponentiallygrowing cells (-1.2 x 109 cells per ml) by centrifugation(10 min at 7,000 x g at 4°C) and resuspending thepellets in TM buffer [0.01 M tris(hydroxy-methyl)aminomethane buffer, pH 7.6, containing 0.01M MgCl2] at low (1.6 x 109 cells per ml) or high (8 x109 cells per ml) density. The cells were then spottedon CF agar and incubated at 30°C. Fruiting pheno-types were observed and photographed after 5 days,using a dissecting microscope equipped with a Polaroidcamera.Growth rates in TYE broth were determined by

measuring turbidity, using a Klett-Summerson pho-toelectric colorimeter. Spore samples were preparedby harvesting cells after 5 days on CF agar. The cellswere resuspended in TM buffer, heated at 55°C for 1h, and sonically oscillated for 1 min. (The probe wassterilized in chloroforn and rinsed in 95% ethanolbetween samples to prevent cross-contamination.)This procedure reduced viability of vegetative cells by>109-fold but did not affect spore viability. Sampleswere plated on CYE or TYE agar, using a CF agaroverlay. Sporulation frequencies were normalized to1.0 for wild type.The effect of wild-type cells on the sporulation of

mutant strains was determined by mixing wild-typeand mutant cells in a 1:1 ratio and spotting at highdensity on CF agar. Spore counts were determined onplates containing no rifampin and plates containing 10,ug of rifampin per ml.

Glycerol inductions were performed by adding glyc-erol (0.5 M) to exponential-phase cultures (8).Bacteriophage growth and transductions.

Stock lysates of bacteriophage were prepared by in-fecting DZ1 at a multiplicity of 1 in CYE broth,incubating at 280C overnight, treating with chloro-form, and clearing lysates by centrifugation. DZ1 wasused as an indicator for all phage lysates.

Transducing phage lysates were prepared by inoc-ulating CT plates with donor bacteria and bacterio-phage at a multiplicity of infection of 0.005. Plateswere incubated for 2 to 4 days at 28°C and harvestedby leaching overnight in 5 ml of 0.1% NaCl per plate.Bacteriophage yields were between 109 and 1010 phageper ml.

Transductions were carried out by mixing phageand recipient bacteria (multiplicity of infection, 2 to10) and plating directly on CF agar at 35°C. After 24h of outgrowth, rifampin was added in an agar overlayto a final concentration of 10 Ag/ml. This procedureensured that all transductants were independentlyderived. DZF1 was the recipient strain in all transduc-tion experiments.

In vitro assay ofDNA-dependent RNA polym-erase activity. The procedure of Burgess (3) wasused to assay in vitro DNA-dependent RNA polym-erase activities, with modifications as noted. BuffersG and A contained 1.0mM ethylenedinitrilotetraaceticacid and 1.0 mM phenylmethanesulfonyl fluoride toinhibit protease activity; the pH of buffer G was 7.5,and buffer A contained 10% glycerol. Cells were grownin TYE broth at 30°C to 1.0 x 109 cells per ml. The

cells were centrifuged at 7,000 x g for 10 min, washedonce in buffer G, and frozen as pellets at -76°C.Thawed cells were resuspended in 2 ml of buffer G perg (wet weight) and sonically oscillated for a total of 3min in 30-s bursts. The extracts were maintained below4°C until assayed. Three volumes of buffer G wereadded to the sonic extracts, and solid ammoniumsulfate was added to bring the final concentration to25% saturation. The extracts were centrifuged at27,000 rpm for 2 h (using a Spinco no. 30 rotor) toremove ribosomes and cell debris. The supernatantwas brought to 65% saturated ammonium sulfate,stirred for 45 min, and centrifuged at 40,000 x g for 25min. The top of the pellet was washed three times andresuspended in a minimal volume of buffer A. Theextracts were dialyzed for 4 h against buffer A andstored in 50% glycerol at -20°C. Protein concentrationwas determined by the method of Lowry et al. (12).The RNA polymerase assay mixture of Burgess was

modified to contain 1.0 mM ethylenedinitrilotetraa-cetic acid, 1.0 mM phenylmethylsulfonyl fluoride, and40 Ag of salmon testes DNA per ml. Newly synthesizedRNA was labeled with [3H]UTP and [3H]CTP (Amer-sham Corp., 53 mCi/mmol) at 35°C, and 100-pl sam-ples were trichloroacetic acid precipitated on What-mann 3MM filters, using a batch-washing procedure.Filters were washed four times in cold trichloroaceticacid, twice in acetone, and once in ethyl ether. Sampleswere air dried and counted in a liquid scintillationcounter. One unit of activity incorporates 1 pmol ofCMP or UMP into trichloroacetic acid-precipitablematerial per min per ml.

RESULTS

Isolation of rifampin-resistant mutantsdefective in fruiting-body formation. M.xanthus was extremely sensitive to rifampin; theminimum inhibitory concentration was only 0.2,ug/ml, as determined by cell viability on CYEagar containing various drug concentrations.Therefore, spontaneous mutants resistant torifampin (Rif') were selected by simply platingcultures of strains DZ2, DZF1, and DZF6 ongrowth media containing 25 ug of rifampin perml, a concentration 100 times higher than theminiimum inhibitory concentration and there-fore sufficiently high to reduce the probabilityof selecting for permeability mutations. The Rif r

mutants appeared at a frequency of 1O-7 to 10-8.In an initial experiment, 1325 Rif' colonies de-rived from two independent cultures of strainDZ2 were isolated and then tested for theirability to fruit on coli agar. Of these, 84 wereunable to forn fruiting bodies on the first screen-ing. In a control experiment, 1,000 rifampin-sen-sitive colonies were tested for their ability toform fruiting bodies, and all did. These resultssuggested that nonfruiting mutants (Fru-) maybe found as a subclass among Rif r mutants.

Since these Fru- mutants were not derivedfrom independent cultures, they could not be

J. BACTrERIOL.

on April 6, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

RIFr DEVELOPMENTAL MUTANTS OF M. XANTHUS 297

used to determine the frequency of Rifr Fru-mutations. In Fig. 1, 64 independent cultures ofM. xanthus were used to select Rifr colonies.Eighteen of these colonies from each independ-ent culture were tested for their ability to formfruiting bodies. The results showed significantdifferences between the independent cultures.For example, in several cultures none of the 18colonies were defective in fruiting; in others, asmany as 15 of 18 colonies were Fru-. Theseresults probably reflect the large numbers ofsiblings present in the independent cultures sothat the earliest spontaneous mutant to Rifr

predominates. In Fig. 1, the average spontaneousfrequency of Fru- phenotypes among many in-dependent cultures was about 20% of the initialRif r isolates.

It should be noted that about half of the initialRif r Fru- isolates showed Rif r Fru+ phenotypeson subsequent subculturing and retesting (eachstrain was subcultured on CYE agar and retestedfor fruiting on coli agar at least three times). Inall cases, the Fru- phenotype but not the Rifrphenotype reverted. These results suggest eitherthat the original double phenotype (Rifr Fru&)was due to two separate mutations or that theFru- phenotype (but not the Rifr phenotype)was suppressed by a second site mutation in the"revertants."Genetic analysis of the Rifr Fru- mu-

tants. Sixty-seven stable RifFr Fru- strains wereinfected with bacteriophage MX4ts27htfhrm, atemperature-sensitive, high-frequency general-ized transducing phage with wide host range forstrains of M. xanthus (4). The phage lysateswere used to transduce a "wild-type" strain,DZF1, to RifrF. The transductants were thentested for their ability to forin fruiting bodies oncoli agar. About one-half of the strains usedshowed very high cotransduction of the riffru-allele(s); the remaining strains showed no co-transduction or, in a few cases, low cotransduc-tion. Since many of these crosses were not iso-genic, it is possible that the original Rifr Fru-phenotypes were dependent upon the particulargenetic backgrounds of the initial Rifr isolates.In other cases where the crosses were isogenic,it is clear that the original strains, althoughspontaneously induced mutants, were actuallydouble mutants.The strains which showed very high cotrans-

duction frequencies were studied further. ElevenDZF1 transductants from independently derivedmutants were used to prepare the phage lysates,and their mutations were transduced again intoDZF1 (Table 1). The transductants were platedin a CF agar overlay. After 24 h of outgrowth at350C, a second overlay containing rifampin wasplaced above the microcolonies. Thus, in each

n20

15

10

0-0.1 0.1-0.2 0.2-0.3 0.3-0.4 0.4-0.5 0.5-0.6 0.6-0.7 0.7-0.8

Frequency of Fru- Phenotype in RifR IsolatesFIG. 1. Distribution of Rif' Fru- phenotypes in

independent cultures. Sixty-four independent culturesof M. xanthus strain DZF6 were concentrated andplated on CF plates containing rifampin (25 pg/ml).After 10 days of incubation at 30°C, 18 Rif' colonieswere isolated, and their fruiting phenotypes weredetermined on coli agar. The histogram shows thefrequency ofRif' Fru- phenotypes observed in the 18colonies.

experiment, the Rif F clones are independenttransductants; furthermore, the medium is afruiting medium so that the Fru+ phenotype canbe screened directly. In this manner, over 900independent transductants were analyzed foreach mutant. In each case the cotransductionfrequencies were 299%. The resolution of thetransduction data was limited by the number oftransductants screened and by the low but finitelevel ofnew spontaneous mutations to Rifr (10-7to 10-8). The transduction frequency of the riflocus was 100- to 1,000-fold higher than thisspontaneous rate. The high cotransduction fre-quencies in these strains support the hypothesisthat the Rif r and Fru- phenotypes are the resultof a single mutation.Phenotypic analysis of the Rif Fru mu-

tants. Figure 2 shows the morphologicalchanges which occur when the wild-type strainDZF1 is spotted at high density on CF agar. Thespots undergo a series of transitions which in-clude: smooth spots (0 h), wrinkled aggregates(4 to 8 h), translucent mounds (8 to 15 h), andmorphologically mature fruiting bodies (28 to 32h). All the Rifr Fru- mutants analyzed appearto show terminal morphological phenotypeswhich roughly correspond to intermediate stagesof wild-type development. We therefore dividedthe mutants into two groups on the basis of theirterminal phenotypes: the class I mutants appearto be blocked early. They show very little ten-dency to aggregate at high density (8 x 10' cellsper ml) but some tendency to aggregate intoswirls or flat mounds at low density (1.6 x 109

VOL. 137, 1979

on April 6, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

TABLE 1. Phenotypic and genetic analysis of independent Rifr Fru- mutants

Sporulation frequency N

Growith rateco. off tdepand CtnducPhenotypic classm Strain one Plus wild (% wild type) ent

te

tion (%)

I. Low-level DZF1020 10-2_10-4 1.0 100 1,024 299.0aggregation DZF3042 10-3-10-5 1.0 100 1,669 299.9

DZF3030 10-I_10-2 10-l 100 1,193 >99.9DZF3039 <10-7 lo-, 100 985 299.0DZF3028 10-7 10-2 100 939 299.0DZF1782 <10-7 lo-, 100 908 299.0DZF3041 10-6 10-6 100 916 >99.0

II. Translucent DZF3031 10-2_10-3 10lo 95 1,585 ; 99.9mounds DZF3044 10-2 lo-1 90 1,903 299.9

DZF3045 10-3-10-5 1.0 90 1,181 299.5DZF3043 10-7 lo-1 100 1,375 299.9

a Phenotypes were determined by spotting cells at high density on CF agar and incubating at 300C for 5 days.b Sporulation frequencies were determined from the frequency of spores found in the 5-day-old spots described

above. It should be noted that sporulation frequencies are normalized to 1.0 for wild type (DZF1).'Growth rates were determined in TYE broth at 300C.d Independent Rifr transductants were selected as described in the text.

cJ fvpe ........X

..~~~~~~~~~~~~~~~~~~~~r.,

7kF1782o.x 3;^C'43iXC.44

..............~~~~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~~~~..

FIG. 2. Morphology of wild type and Rif' Fru- mutants on a fruiting medium. Wild type (DZFI) and Rif'Fru- mutants (DZF1782 and DZF3043) were concentrated and spotted at high density on CFagar as describedin the text. The spot diameter is 5 mm.

cells per ml). The class II mutants form tightaggregates (translucent mounds) at both highand low density. Many of the mutants withineach class show small but distinctive morpho-logical differences.The ability of the Rifr Fru- mutants to

sporulate was also examined. All mutants couldbe induced to form glycerol spores by the addi-tion of 0.5 M glycerol to exponential-phase cul-tures. Under these conditions, Dworkin and Gib-son (8) showed that vegetative cells synchro-nously convert to spherical sporelike cells inliquid broth in the absence of aggregation. Incontrast, all of the mutants showed low (<10-2)or very low (<10-7) levels of sporulation onfruiting medium (Table 1). The small number of

spores obtained with the mutants gave rise tomutant clones when plated on growth mediumand therefore were not revertants.Hagen et al. (9) found that many nonfruiting

mutants, when mixed with wild-type cells, canbe stimulated to sporulate, although usually ata reduced level. We therefore mixed equal num-bers of RfrFru- mutants with DZF1 and spot-ted the cells on fruiting medium. After 5 days,the cells were removed from the plates andanalyzed for Rffr spores. Table 1 shows that 10of 11 strains tested in this manner could bestimulated to sporulate by wild-type cells. Inseveral strains (DZF3039, DZF1782, andDZF3043), sporulation was 106-fold higher thanthat obtained by the mutant strain alone. These

298 RUDD AND ZUSMAN J. BACTERIOL.

on April 6, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

RIFr DEVELOPMENTAL MUTANTS OF M. XANTHUS 299

results shave threason cthe absTable

strains:cases tkvalue idDZF1.ments inormal:the "yelthe mutinterferIn vil

mutantxanthusto be s4rifampirwere stiassaymnj

,ug of rifexperimof DZF:rifampirDZF303The extextensivsibility tificationtracts wity of Dmutatioto be caerase itsbility or

6

5

-, 4to0

E 3

2

FIG. 3

from wilkprepared(DZF303scribed i

pin addo(A) comj

added.-

;uggest that most of the Riff Fru- strainsto ennuritv tn annrmlut& h-ili fnrr arninm

DISCUSSIONIC ullpaulL . ae.U. some We have isolated a collection of Rif r mutantslo not receive the signal to sporulate m which are defective in fruiting-body formation.mnce of wild-type cells.wlhaedfclemfIlm-bd omto1nce o ilde te The mutants appear to be altered in their RNAin richmedu the browth). In most polymerases since in vitro studies showed rifam-inrichmedum(TY,o pin-resistant RNA polymerase activity but noie growth rates were 3.75 h at 3000, a detectable detoxification activity. Since the RifrWentical to the parental wild-type strain, and Fru- phenotypes cotransduce at very highWTe estimate the error in these measure- frequency (in some strains 299.9%), it is likelyat about 5%. These strains alsoshow...it b 5 %. T he s st rains nl s o w sin le u tio nfers ru resist-

mlow"tphse variant form(DZ7) Therefore ance and inpaired ability to forn fruiting bodies.

atow "-nha v ar a ntfo r (17). T he re fore, The data presented, however, cannot rule out

*tion the possibility that two very closely linked mu-e With vegetativeftnmctions.tr veeaier .tations arose spontaneously in each ofthe strainsto. RNA polymerase assays of t.h examined. Reversion analysis would be useful int8 Thne RNA polymerase activity Of M-. teghnn h rgmn notntls DZF1 was assayed in extracts and found strengthening the argument. Unfortunately, we

ensitive (50% inhibition) to 0.02 ug of have so far been unsuccessful in the selection of

n per ml. The strains isted in Table 1 fru' revertants... m. It was curious that many of our initial isolatesadied in vitro by preparing extracts and of spontaneous rifr mutants clearly contained at

them in the presence and absence of 4 least two mutations: one responsible for the drugfampin per ml. The results of a typical resistance and another one for the defect inLent are shown in Fig. 3. Whereas extracts fruiting-body formation. One possible explana-1 were almost completely inhibited by ifr

n, extracts of the Rifr Fru- mutant tion for. thirs finding is that the initial rlf muta-19 we xr ac ts o th RhF mrut. tion might confer a selective disadvantage on

9 er c m p et l r s to t . cell, allowing tation

,racts were rendered DNA dependent by arise during the growth of a single clone thatre sonic treatment. To eliminate the pos- * * .that the mutants have acquired a detox- might affect both fruitng and vegetative growthactivity, rifampin from the mutant ex-

rtuaey,teese condaryimuttions'ras tested on the RNA polymerase activ- readmuyabesseparaedbyotrnsduction.ZF1 and found to be active. Thus, the The mutants studed show normal veetatve

SW-EW*fr * a * * s s ^ s growth rates and morphology but exhibit seriousins ofR in the strains tested are lkely defects in their ability to form fruiting bodies.tused by alterations in the RNA polym- The mutants can be divided into two phenotypic.elfrather than in changed cell permea- classes on the basis of their progress in theinmcreased detoxificationactivity..ddevelopment pathway. The most likely explan-

tion of these findings is that mutations in the

WildtylpeS. if mutont 8

RNA polymerase molecule are responsible forchanging the transcriptional specificities of the

.-Is enzyme. The alterations must be subtle sincethe detectable vegetative functions remain un-

- 12 changed. However, developmental transcription3 may be seriously impaired. This could occur by

9 altering the ability of the enzyme to bind a

i. regulatory factor or, alternatively, by directly/6/changing its ability to recognize certain pro->-~/3 moter sites. The fact that at least two pheno-

typic classes of Rifr Fru- mutants were found5 15 20 S 10 5 20 suggests that the transcriptional specificities

Time (min) may be altered more than once during develop-ment. Altered transcriptional specificities have

1. In vitro RNApolymerase assays ofextracts been reported during phage development (1, 16)

d type and a Rifr Fru mutant. Extracts were and have been suggested during sporulation of

from wild-type (DZFI) and mutant Bacillus subtilis (13). Changes in transcriptional9) vegetative cultures and assayed as de- sccitiesduring developmentrof .rxathusn the text. Symbols: (0) complete, no rifam- specificities duerng development of M. xanthus

'ed; (0) complete plus rifampin (4 pg/ml); can best be determined by in vitro studies on

,lete minus salmon testes DNA, no rifampin defined templates. This work is currently inprogress.

A.

i-

VOL. 137, 1979

on April 6, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

300 RUDD AND ZUSMAN

The developmental defects of the Rif mu-tants of M. xanthus differ significantly fromthose found in B. subtilis (6, 10, 15). The mu-tants-described here appear to be blocked pri-marily in developmental aggregation or cell-cellcommunication, and only as a consequence, insporulation. Presumably, the cells do notsporulate because they never receive the aggre-gation-dependent signal for sporulation. Thisconclusion is supported by the very dramatic(106-fold) higher sporulation frequency observedfor several mutants when they are stimulated bywild-type cells. Furthermore, all of the mutantswere able to form heat-resistant and sonic treat-ment-resistant, sporelike cells in the presence ofglycerol. In contrast, the Rifr developmentalmutants isolated in B. subtilis appear to bedefective in sporulation. The coupling of sporu-lation to cell-cell communication in M. xanthusis an interesting area for future research.

ACKNOWLEDGMENTSThis work was supported by grants from the Public Health

Service National Institute of General Medical Sciences (GM20609) and the National Science Foundation (PCM 75-09231A01) One of us (K.R.) was a Public Health Research Predoc-toral Trainee (5-T32-GM07232).We thank C. Morrison for critical reading of the manu-

scipt.

LITERATURE CITED1. Bolle, A., R. H. Epstein, and E. P. Geiduwchek. 1968.

Transcription during bacteriophage T4 development:synthesis and relative stability of early and late RNA.J. Mol. Biol. 31:326-348.

2. Bretacher, A. P., and D. Kaiser. 1978. Nutrition ofMyxococcus xanthus, a fruiting myxobacterium. J. Bac-teriol. 133:763-768.

3. Burgess, R. R. 1969. A new method for the large scalepurification of E. coli DNA-dependent RNA polymer-ase. J. Biol. Chem. 244:6160-6168.

4. Campos, J. M., J. Geisselsoder, and D. R. Zusman1978. Isolation of bacteriophage MX4, a generalizedtransducing phage for Myxococcus xanthus. J. Mol.

Biol. 119:167-178.5. Campos, J. KL, and D. R. Zusman. 1975. Regulation of

development in Myxococcus xanthus: effect of 3':5'-cyclic AMP, ADP, and nutrition. Proc. Natl. Acad. Sci.U.S.A. 72:518-522.

6. Doi, RI H., L RI Brown, G. Rogers, and Y. Hsu 1970.Bacillus subtilis mutant altered in spore morphologyand in RNA polymerase activity. Proc. Natl. Acad. Sci.U.S.A. 66:404 410.

7. Dworkin, K. 1972. The myxobacteria: new direction instudies of procaryotic development. Crit. Rev. Micro-biol. 2:435-452.

8. Dworklin, K, and S. M. Gibson. 1964. A system forstudying microbial morphogenesis: rapid formation ofmicrocysts in Myxococcus xanthus. Science 146:243-244.

9. Hagen, D. C., A. P. Bretacher, and D. Kaiser. 1978.Synergism between morphogenetic mutants of Myxo-coccus xanthu8. Dev. Biol. 64:284-296.

10. Leighton, T. J. 1973. An RNA polymerase mutationcausing temperature-sensitive sporulation in Bacilldusubtilis. Proc. Natl. Acad. Sci. U.S.A. 70:1179-1183.

11. LAnn, T. G., R. Losick, and A. L Sonenshein. 1975.Rifampin resistance of Bacillus subtilis altering theelectrophoretic mobility of the beta subunit of ribonu-cleic acid polymerase. J. Bacteriol. 122:1387-1390.

12. Lowry, 0. H., N. J. Rosbrough, A. L Farr, and RI J.RandaLl 1951. Protein measurement with the Folinphenol reagent. J. Biol. Chem. 193:265-275.

13. Pero, J., J. Nelson, and R. Losck. 1975. In vitro and invivo transciption by vegetative and sporlating Bacil-lus subtilu, p. 202-212. In P. Gerhardt, R N. Costilow,and H. L Sadoff (ed.), Spores VI. American Society forMicrobiology, Washington, D.C.

14. Rabussay, D., and W. Zilg. 1969. A rifampicin ristantRNA-polymerase from E. coli altered in the ,B subunit.FEBS Lett. 5:104-106.

15. Sonenshein, A. L, and R. Losdck. 1970. RNA polym-erase mutants blocked in sporulation. Nature (London)227:906-909.

16. Taklngton, C., and J. Pero. 1977. Restriction fragmentanalysis of the temporal program ofbacteriophage SPOltranscription and its control by phage-modified RNApolymerases. Virology 83:366-379.

17. Wireman, J. W., and KL DworkLn. 1975. Morphogenesisand developmental interactions in myxobacteria. Sci-ence 189:516-523.

18. Zusman, D. RI, D. K Krotolci, and KL Cumaky. 1978.Chromosome replication in Myxococcus xanthus. J.Bacteriol. 138:122-129.

J. BACTERIOL.

on April 6, 2018 by guest

http://jb.asm.org/

Dow

nloaded from