Advances in Streptomyces coelicolor Genetics - NCBI

BACTERIOLOGICAL REVIEWS, Sept. 1973, p. 371-405Copyright ( 1973 American Society for Microbiology

Vol. 37, No. 3Printed in U.S.A.

Advances in Streptomyces coelicolor GeneticsDAVID A. HOPWOOD, KEITH F. CHATER, JOHN E. DOWDING' AND ALAN VIVIAN2

John Innes Institute, Norwich NOR 70F, England

INTRODUCTION ................................................................ 371SOME TECHNICAL ADVANCES ........ ........................................ 373Mutagenic Procedures........................................................ 373Mutant Isolation ........................................................... 373Crossing Techniques ....................................................... 374Phage Methods.............................................................. 376Growth of Liquid Cultures for Biochemical Studies.376Culture Preservation......................................................... 376

THE LINKAGE MAP AND MARKERS ....... ................................... 377THE FERTILITY SYSTEM ...................................................... 377

First Indications of the Existence of a Fertility System 377

Characterization of the NF, UF, and IF Fertility Types .... ................... 377New Donor Types............................................................ 386Role of the Plasmid SCPI in Fertility ...... .................................. 389Transfer of SCPI from A3(2) to Other Strains ..... ........................... 391Comparison of Gene Transfer in S. coelicolor with That in E. coli.391

ANALYSIS OF DIFFERENTIATION ........ .................................... 393Phases of Colonial Development ........ ..................................... 393Characterization on Morphological Mutants ..... ............................. 393Biochemical and Physiological Approaches to the Study of Differentiation in S.

coelicolor ................................................................. 395

GENETIC ASPECTS OF PHAGES ......... ...................................... 395Isolation and Characterization of Phages..................................... 395

Phage Mutants and Recombination ....... ................................... 397Lysogeny ................................................................... 397Transduction................................................................ 399

RADIOBIOLOGY ................................................................. 399Ultraviolet-Sensitive Mutants ........ ...................................... 399Photoreactivation ........................................................... 401

ANALYSIS OF BIOSYNTHETIC AND DISSIMILATORY PATHWAYS ... ...... 401Tryptophan Synthesis....................................................... 401Aromatic Amino Acid Synthesis ........ ..................................... 401Histidine Synthesis........................................................... 401Dissimilatory Pathways..................................................... 401Carbon Dioxide Mutants..................................................... 401

CONCLUDING REMARKS....................................................... 402ACKNOWLEDGMENTS ........................................................ 402LITERATURE CITED............................................................. 402

INTRODUCTIONSince the last detailed reviews of the method-

ology and results of genetic studies with Strep-tomyces coelicolor (51, 87), a number of ad-vances in our understanding of genetic phenom-ena in this organism have been made. Mostrecent research has been in three areas: thefertility system; colonial differentiation; andthe interactions between bacteriophages and

I Present address: Department of Biochemistry,University of Wisconsin, Madison, Wisconsin 53706.

2 Present address: School of Biological Sciences,Thames Polytechnic, London SE18 6PF, England.

streptomycetes. There have been some radiobi-ological studies and preliminary investigationsof the genetic and biochemical organization of afew pathways. A number of improvements inthe techniques of mutagenesis, mutant isola-tion, and genetic analysis have also been made.The previous review (51) describes most of thebasic methods of genetic analysis that are stillin routine use in this laboratory. The practicaldetails will not be repeated here; thus theprevious article (51) should be consulted inconjunction with the present review by anyonewishing to become familiar with the practicalmanipulation of the organism. However, it may

371

HOPWOOD ET AL.

be useful to recapitulate here the main featuresof the genetic system of the organism as abackground to the description of the fertilitysystem which is introduced historically in alater section.

Genetic transfer in S. coelicolor occurs by aprocess which is described as "conjugation"because it results in the transfer of long seg-ments of linked genes; but this does not neces-sarily imply a close structural analogy withconjugation as it occurs in gram-negative eu-bacteria. However the process resembles that inEscherichia coli and its relatives in leading tothe production of merozygotes, each containinga complete (circular) chromosome from oneparent and a fragment of variable length fromthe other. In experimental genetic studies,crosses are made by mixing spores of twogenetically marked strains on solid media. Mat-ing occurs between the hyphae produced aftergermination of the spores, and the results of themating process are studied by analyzing asample of progeny spores (or occasionally ofhyphal fragments produced by mechanical dis-ruption of the growing mat of mycelium) laterproduced on the mixed mycelial culture. Thespores themselves do not mate so that, if matingin liquid media were to be employed, as isalmost invariably the case in gram-negativeeubacteria, the difficulties of handling the heter-ogeneous mycelial aggregates that result fromspore germination would have to be solved, andthis has not so far been seriously attempted.The spores harvested from the crosses include

recombinants and, usually, members of bothparental classes produced asexually. In mostcases the latter are in a large majority, and it isfeasible to study the recombinants only byselecting various genotypic classes on media onwhich neither parental type can develop. Incertain combinations of mating types, however,recombinants are very frequent, and may evenmake up virtually the entire spore progeny ofthe cross; in these circumstances it is useful torecover a sample of the total progeny as colonieson nonselective medium.Much of the previous review (51) was devoted

to a consideration, in formal genetic terms, ofthe events following merozygote formation, anda hypothesis was developed, which has not beendisproved, that two alternative pathways areopen to an individual merozygote, depending onwhether even or odd numbers of crossoversoccur between the chromosome fragment andthe region of the whole circular chromosomewith which it is homologous. Even-numberedcrossovers generate circular, haploid recombi-nant genomes directly, whereas odd-numbered

crossovers lead to terminally redundant, andtherefore partially heterozygous, genomes.Subsequently, further opportunities for crossingover occur within the duplicated regions of suchgenomes, leading finally to the production ofhaploid recombinants. However, when a hetero-zygous genome is included in a spore thatdevelops into a colony on selective medium, thecolony contains the still heterozygous descend-ants of the original genome, and later a varietyof haploid recombinants arising from them bycrossing over at various points within the dip-loid region; such colonies are called heteroclones(59, 91), and these were discussed in detail inthe earlier review (51).

In the section of the present review dealingwith the fertility system, the characteristics ofthe merozygotes in various combinations offertility types will be described, and the fertilitytypes will be interpreted in terms of the pres-ence, in autonomous or integrated state, or theabsence, of a plasmid (SCP1) which, possibly inconjunction with other as yet unidentified fac-tors, plays a role as a sex factor in S. coelicolor.It will emerge that there are some analogiesbetween the fertility system of S. coelicolor andthe well-known system of Escherichia coli K-12,but there are also differences, which justify theuse of different terms, initial fertility (IF),ultrafertility (UF), and normal fertility (NF), todesignate the fertility types of S. coelicolor.

In the last few years some of the techniquesdeveloped with S. coelicolor A3(2) have beenextended to other actinomycetes. In most casesS. coelicolor has turned out to be a useful modelfor studies on other strains, in at least five ofwhich some linkage data are now available: awild-type (2) and a high-yielding oxytetracy-cline-producing indust-rial strain (37) of S. rimo-sus, a strain of S. bikiniensis producing zor-bonomycin (20), a melanin-producing wild-typeof a strain of Streptomyces glaucescens (R.Baumann, R. Hutter, and D. A. Hopwood,manuscript submitted for publication), a strainof Streptomyces olivaceus (72, 73); and the rif-amycin-producing Nocardia mediterranei (T.Schupp, personal communication). In all fivespecies a circular linkage map has been con-structed, and in S. glaucescens, S. rimosus, S.olivaceus, and S. bikiniensis the arrangementof markers on the map shows a strong overallsimilarity with that in S. coelicolor A3(2) (2. 37,73; R. Baumann, R. Hutter, and D. A. Hop-wood, manuscript submitted for publication).In the strain of S. rimosus studied by Alacevici(2), and in S. glaucescens (R. Baumann, R.Hutter, and D. A. Hopwood, manuscript sub-mitted for publication), heteroclones of the S.

BACTERIOL. REV.372

STREPTOMYCES COELICOLOR GENETICS

coelicolor type (51, 59, 91) were found, and thegeneral characteristics of crosses thus seemedto be the same as those of the early crosses in S.coelicolor. Heteroclones were not sought in thestrain of S. rimosus studied by Hopwood andFriend (37), nor in S. bikiniensis (20), and theywould not have been expected in the selectiveconditions employed; however, a differencefrom S. coelicolor A3(2) was the frequent occur-rence of heterokaryons (defined as coloniesgrowing on selective media, but segregating outonly parental genotypes on subculture) in thecrosses of the industrial S. rimosus strain. InThermoactinomyces vulgaris, application ofsome of the procedures of Streptomyces ge-netics led not to a further example of conjuga-tion, but to the first unequivocal demonstrationin an actinomycete of transformation (54, 61).Genetic studies on actinomycetes other thanS. coelicolor have recently been reviewed byHopwood (52).

SOME TECHNICAL ADVANCES

Mutagenic ProceduresMutagenic treatments in S. coelicolor have

been greatly improved since 1967. The followingprocedures are currently used.

Ultraviolet light. Ultraviolet (UV) irradia-tion involves no special features. However, be-cause S. coelicolor A3(2) and its descendants donot show any significant photoreactivation (seesection on Radiobiology), all operations can becarried out under normal illumination. Thedose for 1% survival varies from about 2,000 to4,000 ergs/mm2 depending on the strain (41).N - methyl - N' - nitro - N - nitrosoguanidine

(NTG). The spores of most streptomycetes aremuch more resistant to NTG than the cells ofother bacteria, and so efficient mutagenesis ofspores is achieved only at rather high NTGconcentrations and at alkaline pH, whichcauses rapid breakdown of the NTG to theeffective mutagen, diazomethane (28).

In a very intense treatment, incubation for 2 hin 0.05 M tris(hydroxymethyl)aminomethane-maleic acid buffer at pH 9 containing 3 mg ofNTG per ml is used; a milder treatment mightinvolve 1 mg/ml for 1 h.Near-ultraviolet light in the presence of

8-methoxypsoralen. The procedure wasadapted with little change (98) from that de-scribed for Aspergillus nidulans (86). Althoughdifferent wild-type streptomycetes vary mark-edly in sensitivity, many differently markedderivatives of S. coelicolor A3(2) have (with afew exceptions) shown a reproducible sensitiv-ity: survival of 1 to 5% after 10 min of irradia-

tion under our standard conditions (98; V.Najfeld, personal communication). This treat-ment should give 3 to 5% auxotrophs amongsurvivors. If a more intense treatment is re-quired, irradiation can be continued; after 20min, survival of A3(2) in one experiment was0.001%, with 9% auxotrophs (98). 8-Methoxy-psoralen plus near-ultraviolet light has the ad-vantage for a general-purpose mutagen of effi-cient mutagenesis without, apparently, causingclustered mutations.Comutation. Following the discovery that

NTG tends to cause groups of clustered muta-tions in E. coli, probably as a result of itspreferential action at the chromosomal replica-tion fork (39), Randazzo et al. (80) investigatedthe possibility that S. coelicolor clones selected,after mutagenesis, for reversion of a markermight show an enhanced frequency of forwardmutations at nearby loci. They found a highfrequency of forward mutations at sites closelylinked to hisA among NTG-induced hisA +revertants. This discovery opens the way to theselection of new mutations in particular regionsof the linkage map.

Mutant IsolationSome additional classes of mutants have been

isolated and studied since 1967: morphologicalmutants with defects in sporulation, mutantsresistant to rifampin or to phages, and thoseunable to utilize carbon sources.The main new class of morphological mutants

includes those with a white colony phenotype.Since the normal spore color in S. coelicolor ispale grey, it is possible to recognize visually, bytheir white colonies, mutants (called whi) defec-tive in sporulation (17, 60; A. C. McVittie,manuscript submitted for publication), even onplates carrying many hundreds of colonies.The isolation of induced rifampin-resistant

(K. F. Chater, manuscript submitted for publi-cation) and streptomycin-resistant mutants(D. A. Hopwood and H. M. Wright, unpublisheddata) has emphasized the need to allow a periodof growth after mutagenesis and before platingon medium containing the inhibitor, to allow forthe occurrence of segregational delay and/orphenotypic lag if induced drug-resistant mu-tants are to be recovered. This is, of course, notso easily arranged in a mycelial organism as in aunicellular one, but it can be done in a number ofways. A period of incubation of 6 h or more inliquid medium may be allowed before plating onagar containing the drug. If excessively clumpedgrowth occurs, this can be dispersed by briefsonic treatment before plating. Alternatively,

VOL. 37, 1973 373

HOPWOOD ET AL.

mutagenized spores may be plated directly on

cellophane disks laid on agar medium lackingantibiotic, and the disks transferred, after over-

night incubation, to fresh medium containingthe antibiotic. Not only was the yield of rif-ampin-resistant mutants greatly increased bythis last procedure, but it had the added ad-vantage of ensuring the independence of muta-tions in different colonies isolated from thesame series of platings (K. F. Chater, manu-

script submitted for publication).Dowding (J. E. Dowding, Ph.D. thesis, Univ.

of East Anglia, Norwich, England, 1972) foundthat mutants resistant to lysis by a virulentphage, VP11, could not be isolated by directplating in the presence of the virus. This re-

sulted from the very frequent occurrence ofphenotypically resistant, but genotypically sen-

sitive, growth after incubation long enough forresistant mutants to develop into visible colo-nies. The problem was overcome by platingspores with VP11 on minimal medium (supple-mented with 4 mM calcium nitrate), collectingthe progeny spores after 5 to 6 days, andrepeating the plating with phage five more

times. On testing single colonies developingafter this procedure, 1 out of 120 was found to begenotypically resistant. In this way a mutation,pel-1 (phage eleven resistance), was obtainedwhose segregation in crosses could be readilyfollowed (Fig. 1).

Strain A3(2) and its derivatives grow consid-erably on agar media containing salts, but withno added carbon source (96). This property ofagar utilization made it difficult to recognizemutants unable to grow on particular carbonsources by simple plate tests. Recently, byplating mutagenized spores on minimal me-

dium containing glucose and replica plating tomedium without glucose, it has been possible toisolate mutants that fail to grow without an

added carbon source, and the mutations in twoof these have been mapped to locus agaA (D. A.Hopwood and H. M. Wright,, unpublisheddata). By using strains carrying aga mutations,the ability of the wild type to grow on variouscarbon sources could be assessed, and mutantsdefective in the utilization of various sugarscould then be isolated (K. F. Chater and F.Flury, unpublished data).

Crossing TechniquesNormal crossing. Crosses are performed es-

sentially as described previously (51) by mixingspores of complementary parental strains on

solid media. The only modifications apply to"ultrafertile" crosses (see under The Fertility

System) which require a reasonable balance ofthe two parents in the cross and an adequateconcentration of each in the inoculum to giverise rapidly to confluent growth on the crossingmedium. With well-sporulating parent strains itis enough to suspend spores by gently rubbing asmall area of a culture with a loop containing adrop of water and mixing a loopful of eachparent carefully on a moist slant of completemedium. When one or both parents sporulatepoorly, material from one or more slants aresuspended, centrifuged, and resuspended in asmall volume of water to obtain a sufficientlydense inoculum. Some of our ultrafertilecrosses, and others in which reproducible esti-mates of recombination frequencies weresought, have been made on plates rather than intubes because of the greater ease with whichuniform balanced crosses could be obtained;however, for most purposes crosses on slants arecompletely satisfactory.

Plate-crossing. Because of its compact modeof growth, S. coelicolor replicates very cleanlyby velvet pads, and patches retain the same sizeand shape through several cycles of'growth andreplication. This property is exploited in theprocedure known as "plate crossing." This tech-nique, first introduced by Sermonti and Cas-ciano (89), is invaluable in allowing the per-formance of large numbers of crosses quicklywhen only limited information is required fromeach cross. It is particularly valuable in assess-ing the fertility of large numbers of colonies orstrains when mated with a common testerstrain. The colonies from a suspension to betested are grown on a nonselective medium, orpreexisting strains are streaked in patches onsuch a medium. When sporulating, these cul-tures are replica plated with velvet to plates ofcomplete medium on which "lawns" of 107 to109 spores of the tester have been spread. Theresulting "plate crosses" are incubated for a fewdays until they sporulate, when they can bereplicated to one or a series of selective mediaon which various classes of recombinant prog-eny are recovered.The technique of plate crossing is suitable for

rapid genetic studies of large numbers of pheno-typically similar strains. An orderly array of upto 20 patches of the strains is plate crossed witha tester strain, and the progeny are replicated tomedium selecting a suitable population of re-combinants. Classification of this population inrespect to the mutation(s) under test isachieved by visual inspection or by furtherreplication.Such techniques have been useful in the

374 BACTERIOL. REV.


FIG. 1. Phenotypic and genotypic resistance to phage VP11. Segregants from a cross of a strain carrying pel(phage eleven resistance) with a pel+ strain were inoculated on a master plate (top left) which was replicaplated to new plates without VP11 (top right) and with VP11 (bottom). The bottom left-hand picture shows aclear segregation of pel and pel+ after 18-h incubation of the replica. By 42 h (bottom right) a considerablegrowth of phenotypically resistant colonies from the genotypically sensitive pel+ patches is seen.

identification of recombinant strains carrying awhiB mutation in addition to a whiG or Hmutation (K. F. Chater, unpublished data) andin the rapid identification of rifampin-resist-ance (rif) mutations differing from the commonrifA class in not being closely linked to the strAlocus (K. F. Chater, manuscript submitted forpublication). A further adaptation of the tech-nique was in a rapid complementation test foruvs mutants (43). Judicious choice of fertilitytype of the strains to be tested was important inall of these experiments to ensure suitableyields of recombinant colonies.The cellophane transfer technique. The

cellophane transfer technique (88), described in

detail by Sermonti (87), has been applied bothin S. coelicolor (88) and in S. rimosus (2). Amixed spore suspension of two parents is platedon cellophane disks over complete medium at adensity just sufficient to give confluent growthafter incubation for 24 h. One parent is auxo-trophic and streptomycin-resistant and theother streptomycin-sensitive, so that when,after 24 h of incubation, the cellophane istransferred to selective minimal medium con-taining streptomycin, further parental growth isprevented. During the next few days minutetufts appear and can be isolated individuallyand analyzed in the same way as heteroclones(59). Probably they are not individual hetero-

375VOL. 37, 1973

HOPWOOD ET AL.

clones, but localized areas in which merozygoteshave been produced, singly or in small groups,and have given rise to recombinant progeny,doubtless including heteroclones (51). It seemsto us that this disadvantage renders the tech-nique of dubious value in most studies wherethe objective is to select and analyze hetero-clones; however, it can contribute useful infor-mation when growth of the mixed culture is tobe interrupted at particular time intervals (88).

Phage MethodsMethods for studying phages of S. coelicolor

were described by Lomovskaya et al. (68, 69)and Dowding (33).Media. Lomovskaya et al. (69) used a pep-

tone broth for propagation and a corn steepliquor agar for assay of OC31. Dowding (33)found that nutrient broth and nutrient agarsupplemented with glucose and calcium ionswere satisfactory for phage propagation, dilu-tion, and assay and adopted them because oftheir simplicity. Nutrient broth (Difco) is sup-plemented after autoclaving with glucose to0.5% and calcium nitrate to 4 mM. The mediumis solidified with 1.5% or with 0.5% Oxoid agarno. 3 for use as base layers and top layers,respectively.

Lysates. Lysates of the virulent phage VP11have been prepared both from confluently lysedplates and in shaking flasks (33). Confluent-plate lysates yielded approximately 108 plaque-forming units (PFU) per ml and shaking flasklysates 2 to 4 x 109 PFU per ml. The temperatephage VP5 could not be propagated in shakingflasks (J. E. Dowding, Ph.D. thesis, Univ. ofEast Anglia, Norwich, England, 1972), and ly-sates were prepared only by the confluent-platemethod (J. E. Dowding and D. A. Hopwood, J.Gen. Microbiol., in press).Phage mutagenesis. Dowding (33; J. E.

Dowding, Ph.D. thesis, Univ. of East Anglia,Norwich, England, 1972) found the method ofStudier (97) for mutagenesis of intracellularphage by NTG to be very effective for theinduction in VP11 of temperature-sensitive (upto 14% of survivors) and plaque morphologymutants (up to 4%). Free VP5 was effectivelymutagenized with hydroxylamine (J. E. Dowd-ing, Ph.D. thesis, Univ. of East Anglia, Nor-wich, England, 1972; J. E. Dowding and D. A.Hopwood, J. Gen. Microbiol., in press). In theconditions employed, plaque mutants increasedup to at least 8 h, when survival was approxi-mately 1%.

Growth of Liquid Cultures for BiochemicalStudies

S. coelicolor, like many other streptomycetes,grows as large granular masses of mycelium inshaken liquid culture. They are formed by theaggregation of hyphae from a large number ofgermlings soon after germination has takenplace. Consequently, it is difficult to disrupt thecells for biochemical studies, although com-paratively intense treatment with ultrasound(19; P. P. Engel, personal communication),Virtis homogenization (82), or grinding withalumina yields extracts suitable for enzymeassays. It has been found that the addition ofhigh concentrations of sucrose and MgCl2 to thegrowth medium results in liquid cultures thatare much less aggregated than in conventionalmedia, and the mycelium is much more easilylysed by lysozyme (K. F. Chater, manuscriptsubmitted for publication; T. R. Danford and J.I. Frea, Bacteriol. Proc., p. 37-38, 1970; M. E.Townsend, personal communication). We rou-tinely use nutrient broth (Difco) to which isadded 34% sucrose, 1% MgCl2, and 0.5% glu-cose.

Culture PreservationPermanent stock cultures are kept as lyophils

prepared on filter paper strips by a simpleprocedure (53). Unfortunately, the silica gelmethod (76, 77), which has the advantage that aculture can be grown from a few granules of thegel while the rest is stored for future use, wasfound not to be reliable for S. coelicolor strains(53).

In genetic work with streptomycetes, particu-lar strains from the culture collection are con-stantly being required when particular combi-nations of markers are needed in a cross or fortesting the fertility type of groups of strains.Starting from a lyophil or an old stock slant itmay take 7 to 10 days before a freshly clonedculture can be prepared for use. Moreover, it issometimes useful to go back to a cross to obtainfurther recombinants, perhaps imposing a newselection that was not employed in the originalplating. Thus the ability to store spores in goodcondition ready for immediate use is invaluable.Storage of spore suspensions in 20% glycerol at-20 C has recently been found to result in nodetectable loss of viability over a 15-week periodduring which the suspensions were nine timesthawed for sampling and frozen again (K. F.Chater, unpublished data).

376 BACTERIOL. REV.


THE LINKAGE MAP AND MARKERSThe current linkage map of S. coelicolor

A3(2) is shown in Fig. 2 and a list of markers isshown in Table 1. Additional details of themorphological mutations (bld and whi) are

described in the section on Analysis of Differen-tiation and of the uvs mutations under Radiobi-ology.

THE FERTILITY SYSTEM

First Indications of the Existence of aFertility System

The wild-type strain A3(2) of S. coelicolorwas a derivative of Waksman's strain 3443 (96).A3(2) was a single spore isolated by Erikson (35)with the aid of a micromanipulator from a

culture obtained by her from Stanier. All of ourstrains have been derived either by mutation or

recombination from A3(2), which was obtainedby D. A. Hopwood from D. Erikson in February1955 as an old culture preserved in sterilizedsoil.

Recombination between genetically markedderivatives of this culture was first shown byHopwood (45). In these early crosses the fre-quencies of complementary recombinant geno-

types were equal, indicating statistically equalcontributions of genetic material by the twoparents to the progeny, but not necessarily toeach zygote (46). Assuming incomplete zygotes,as we now know to be the case, each parent hadan equal probability of acting either as "donor"of fragments or as "recipient", supplying itswhole genome to the zygote. Later crosses,involving additional mutants and recombi-nants, tended to show unequal frequencies ofcomplementary genotypes (48, 58). Some ofthese crosses also showed considerably greaterrecombinant frequencies than those found inthe early crosses (D. A. Hopwood, Ph.D. thesis,Cambridge Univ., Cambridge, England, 1958).These findings were noted at the time, but theirsignificance became apparent only later (102).The first attempts to investigate a possible

fertility system were reported by Sermonti andCasciano (89), who were working with a set ofstrains derived from our culture collection.They classified certain of these strains as R- (R= recombination) because they were interster-ile, or nearly so, whereas other strains were

designated as R+ because they were fertile witheach other and with the R- strains. A fertilityfactor, R, was postulated to be present in R+and absent from R- strains, but its transmissa-bility from R+ to R- and its state in the R+

strains were unclear. Sermonti and Cascianofound evidence that, in an R+ x R- cross,nonselected alleles of the R- parent tended tobe inherited preferentially by the progeny, sug-gesting that the R+ parent acted as a donor ofan incomplete genome to the R- strain; how-ever, because knowledge of the linkage map wasincomplete at that time, a clear picture of theconstitution of the zygotes in the crosses, andtherefore the relative roles of the two parents,could not be deduced.

In subsequent papers from the same labora-tory, a closer analogy between R+ and R- S.coelicolor strains and F+ and F- cultures of E.coli K-12 was drawn (87). The chief new evi-dence contributing to this interpretation camefrom the results of applying the newly devisedcellophane transfer technique to the study ofthe products of R+ x R- crosses (88). It wasfound that, when the R- parent was resistant tostreptomycin and the R+ parent was sensitive,the longer the period of growth of a cross beforetransfer to selective medium containing strep-tomycin, the greater the proportion of tufts thatcontained the R+ as well as the R- allele at aparticular locus. The results were interpreted asindicating a progressive transfer of chromosomefrom the R+ (donor) to the R- (recipient) strainin a way reminiscent of that in E. coli K-12.

Characterization of the NF, UF, and IFFertility Types

In an attempt to clarify the fertility situationin S. coelicolor, Hopwood et al. (56, 57) set outto isolate strains of altered fertility after muta-genic treatments, rather than relying on sponta-neous variations that might have arisen bychance in the stock culture collection. Becauseour stocks had not shown any evidence ofintersterile strains, two strains were chosen atrandom in respect of fertility; that is simplybecause they were complementary in respect ofa convenient set of markers. One was irradiatedwith UV and plated, and when the colonies werewell grown they were replicated to completemedium spread with a dense spore suspensionof the complementary strain. The resultingplate crosses were subsequently replicated toselective medium to reveal the fertility of eachcolony of the irradiated strain when crossedwith the common tester strain; thus, the tech-nique was the same as that used by Sermontiand Casciano (89), except that a large popula-tion of new colonies, rather than a small groupof preexisting strains, was tested. It had alreadybeen shown by Hopwood (51) that each zygote

377VOL. 37, 1973

HOPWOOD ET AL. BACTERIOL. REV.

-17 |f~B |C 1-53 |D |E A F | 5 H In

NF8 t I °W9|oli VI VI |II |V I II INFo

4e~~~~~~~~~

41~~~~~~~~~~~1

FIG. 2. Genetic map of S. coelicolor. Outer circle: linkage map of all genetic markers mapped to date. Forlocus symbols, see Table 1. The order of bracketed markers, and the location of markers outside broken lineswith respect to those inside the lines, is unknown. Numerals inside the outer circle are allele numbers oftemperature-sensitive mutants (50), those in brackets having since been lost. Inner circle: linkage map ofsporulation mutations (whi) and reference markers used to map them. Within each box is given a capital letterindicating the whi locus designation, a Roman numeral indicating the predominant phenotype of mutants ofthat locus, and a number of dots indicating the number of mutations identified for the locus. Two or more

adjacent dots indicate mutations which may have originated from the same clone.

in a "normal" cross contained a complete chro-mosome from one parent and a fragment fromthe other, each parent contributing the frag-ment to some zygotes and the complete chromo-some to others. In looking for strains of altered

sexual capabilities, the disposition of the se-

lected alleles was such that the irradiated strainwas obliged to contribute three markers atwidely separated loci to the progeny, whereasthe tester strain contributed a single marker.

378


TABLE 1. List of markers

Locus Allelesa | Characteristics [ References"

acr-9, [50]acr-3ade-3, 7,22, [14], [15], h3ade-h2ade-f2, [v10]aga-7amm-1, 2,3,4,5arg-1

arg-t74, [2]arg-4

arg-v8

aro-flaro-wlaro-1, 2ath-2, 8, 10,

h3, h4, h511, 12, hl, h2,

bld-1 [S48]

tps-56tps-59tps-91cdx-vk62cys-15, [17], [19], [20]cys-4, 6, 22, 23, 24cys-3, [14]cys-5, 7, 9. 11, 13, 18,

[21]cys-24, vidap-1glu-vigua-i, hlhis-1, 120, 132

', [16],

his-2, 11, 12

his-9, 10, 14, 119, 127, 128

his-3, 4,15,121 [h2]

his-6, [5]

his-116, 123

his-8, 13, 117

his-129

ile-hlilv-lilv-v2

Resistant to acriflavineResistant to acriflavineRequirement for purinesRequirement for adenineRequirement for purinesUnable to utilize agar as carbon sourceUnable to utilize nitrateRequirement for arginine, citrulline or

ornithineRequirement for arginineRequirement for arginine, citrulline or

ornithineRequirement for arginine, citrulline or

ornithineRequirement for aromatic amino acidsRequirement for aromatic amino acidsRequirement for aromatic amino acidsRequirement for purines plus thiamine

Attachment site for phage OC31Attachment site for phage VP5Lacks macroscopically visible aerial my-

celium (bald)Requirement for "casein hydrolysate"Requirement for "casein hydrolysate"Requirement for "casein hydrolysate"Requirement for carbon dioxideRequirement for cysteineRequirement for cysteine or S2O0Requirement for cysteine or S.0O or S204Requirement for cysteine or S.O. or S504

or S,0.Requirement for cysteine or S20ORequirement for diaminopimelic acidRequirement for glutamic acidRequirement for guanineRequirement for histidine (histidinol de-

hydrogenase)Requirement for histidine or histidinol

(imidazole glycerol phosphate dehy-drase)

Requirement for histidine or histidinol (?isomerase, amidotransferase, or cyclase)

Requirement for histidine or histidinol(histidinol phosphate phosphatase)

Requirement for histidine or histidinol orpurines (PR-AMP 1, 6-cyclohydrolase)


Requirement for histidine or histidinol(imidazole acetol phosphate transami-nase)


Requirement for isoleucineRequirement for isoleucine plus valineRequirement for isoleucine plus valine

I.

499049,5851AB

acrAacrBadeAadeBadeCagaAammAargA

argBargC

argD

aroAaroBaroCathA

att-C31att-VP5bldA

cas-icas-2cas-3cdxAcysAcysBcysCcysD

cysEdapAgluAguaAhisA

hisB

hisC

hisD

hisE

hisF

hisG

hisI

ileAilvAilvB

4946, 48

5051

C

ADD49

68B51

50505010149494949

51511014949, 58, 78, 81

58,78,81

49, 58, 78, 81

49,58,78,81

49

78,81

78,81

78,81

E51C

a Brackets indicate uncertainty.b Personal communications as follows: (A) E. J. Friend; (B) D. A. Hopwood and H. M. Wright; (C) A. Vivian;

(D) P. P. Engel; (E) R. J. Harold; (F) P. J. Hartley; (G) J. E. Dowding; and (H) K. F. Chater.

I

.L

VOL. 37, 1973 379

HOPWOOD ET AL.

TABLE 1-ContinuedL

Locus |AlleleSa Characteristics References"

leu-1, [3], [4]

leu-5, t92

lys-imet-2met-3, 4,5,6

mth-i

mth-2

nic-i, [2], [3], [6], [hl]nic-ul [h2]pab-ipdx-vlpel-iphe-ipro-1, 3, [2], hl, h2red-i

rib-hlrib-irif-i

rif-37rif-d6ser-lser-2spc-lstr-istr-1ithi-i

thi-3, [7], [9], [h2]thi-tiO6, [2], [4], [5], [6],

[hl]thi-8

thr-ithr-2thy-itrp-w7, w15

trp-1, vl, 2,[? 8]

e3, ell, ei3, el4

trp-elO

trp-e2, e6

tyr-t98

tyr-iura-lura-2

Requirement for leucine or a-ketoisoca-proic acid

Requirement for leucine or a-ketoisoca-proic acid

Requirement for lysineRequirement for methionineRequirement for methionine or homocys-

teineRequirement for methionine plus threo-

nineRequirement for methionine plus threo-

nine; or homoserineAttachement site of SCP1 in NF fertility

typeRequirement for nicotinamideRequirement for nicotinamideRequirement for p-aminobenzoic acidRequirement for pyridoxineResistant to phage VP11Requirement for phenylalanineRequirement for prolineRed pigment instead of blue/red indicatorpigment

Requirement for riboflavinRequirement for riboflavinResistant to rifampin (RNA polymer-

ase)Resistant to rifampin (permeability?)Resistant to rifampin (permeability?)Requirement for serine or glycineRequirement for serine or glycineResistant to spectinomycinResistant to streptomycin (high level)Resistant to streptomycin (low level)Requirement for thiamine; does not cross-

feed thiBRequirement for thiamine or thiazoleRequirement for thiamine; crossfeed thiB

Requirement for thiamine; crossfeedsthiB

Requirement for threonineRequirement for threonine or serineRequirement for thymine (leaky)Requirement for tryptophan or indole;

accumulation of indole glycerol (? tryp-tophan synthase a subunit)

Requirement for tryptophan; crossfeed in-dole utilizers (? tryptophan synthase,.subunit)

Requirement for tryptophan or [1-(indol-3'yl) glycerol 3-phosphate synthase]

Requirement for tryptophan or indole;[N-(5'-phosphoribosyl) anthranilateisomerase]

Requirement for tyrosine or phenylala-nine

Requirement for tyrosineRequirement for uracilRequirement for uracil

leuA

leuB

lysAmetAmetB

mthA

mthB

NF

nicAnicBpabApdxApelpheAproAredA

ribAribBrifA

rifBrifCserAserBspcAstrAstrBthiA

thiBthiC

thiD

thrAthrBthyAtrpA

trpB

trpC

trpD

tyrA

tyrBuraAuraB

49

50

5146, 5849

49

49

57, 102

4910151101G46, 50, 584949

B51H

HH49495146, 49, 58F49

4949

50

5151F34, D

34, D

34, D

34, D

50

5149, 5851

380 BACTERIOL. REV.


TABLE 1-Continued

Locus Allelesa Characteristics References5

uraC ura-3 Requirement for uracil 51uraD aur-1, (arg-3, 7) Requirement for uracil plus arginine 51ureA ure-1, [3], [20] Urease-negative 49uvsA uvs-2, etc. UV sensitive; Hcr- for phage VP11 33, 41, 42uvsB vus-6, [21] UV sensitive; Hcr+ for phage VP11 33, 41, 42uvsC uvs-7, etc. UV sensitive; Hcr- for phage VP11 33, 41, 42uvsD uvs-1, etc. UV sensitive; Hcr- for phage VP11 33, 41, 42uvsE uvs-13 UV sensitive 41, 42uvsF uvs-25 Enhances UV sensitivity of mutations in 41, 42

uvsA, uvsC, and uvsD (other uvs locinot tested)

whiA whi-72 White colony; sporulation defective class 17H

whiB whi-70 White colony; sporulation defective class 17I

whiC whi-193 White colony; sporulation defective class I 17whiD whi-16 White colony; sporulation defective class 17

VIwhiE whi-107 White colony; sporulation defective class 17

VIwhiF whi-99 White colony; sporulation defective class 17

VwhiG whi-71 White colony; sporulation defective class I 17whiH whi-119 White colony; sporulation defective class 17

IIIwhiI whi-6 White colony; sporulation defective class 17

IVwhi-53 - White colony; oligosporogenous 17, H

Thus, any colony with an enhanced capabilityof acting as recipient of chromosome fragmentswould give an increased frequency of recombi-nants, whereas a colony with enhanced abilityto donate chromosome fragments, and a re-duced tendency to act as recipient, would givefewer recombinants. In the event variants of theformer type were obtained (Fig. 3), and with ahigh and reproducible frequency; they proved tobe so fertile (up to 100% of all the sporesproduced on the mixed cultures were recombi-nants), not only with the strain used as tester intheir isolation, but with every one of seven otherrandomly chosen strains from the culture collec-tion, that they were called UF for "ultrafertil-ity." All other strains were distinguished, atthat time, as NF ("normal fertility").The characteristics of NF x UF crosses were

described by Hopwood et al. (57). It was de-duced that the NF strain contributed a popula-tion of fragments, but not a random population,to the effective zygotes (that is those able toyield haploid recombinant progeny), whereasthe UF strain contributed a whole chromosometo each zygote. All the fragments from the NFparent carried the 9 o'clock region of its genome;in other words this region was obligate in theeffective zygotes of such crosses. Furthermore,

this region was obligatorily inherited by allviable haploid spore progeny from the cross.Because all the progeny were of NF fertility, itwas postulated that the difference between NFand UF fertility might be controlled by adifference between the NF and UF chromo-somes mapping at the 9 o'clock position. Analternative possibility of a cytoplasmic elementin NF strains absent from UF strains andtransferred with the same 100% efficiency as thechromosome in NF x UF matings could not, atthat time, be excluded, but detailed analysis ofthe frequencies of different classes of progenydid not support the hypothesis of transfer ofsuch an element, without transfer of any chro-mosomal material, to some progeny. Progenyclones of a genotype identical to that of the UFparent except for their NF property (i.e., theclass which theoretically could have ariseneither by "infectious conversion" or by chromo-somal recombination leading to incorporation ofa locus determining fertility type, but of noother segregating marker) had the frequencyexpected for a chromosomal recombinant re-ceiving the implicated region of the NF parent.Thus there was no evidence for infectious con-version of UF to NF.

Figure 4 shows the average contribution of

381VOL. 37, 1973

HOPWOOD ET AL.

FIG. 3. Isolation of UF variants from an IF culture. The latter was irradiated with UV to approximately 1%survival and plated (top plate). The resulting colonies were replica plated to a lawn of an NF tester strain whichwas incubated to produce the "plate cross" (lower left plate). This in turn was replicated to medium selectingrecombinants (lower right plate). Most of the original colonies are still IF and yield a few recombinants (white)on a grey outline of leaky parental growth, but UF colonies (arrows) yield dense patches of recombinants.

different NF markers to the haploid progeny ofNF x UF matings, based on many crosses ineach of which between four and eight markerswere segregating. The bidirectional gradient ofNF marker frequencies, falling in both clock-wise and anticlockwise directions from the 9o'clock position, is characteristic of suchcrosses. This gradient was interpreted (57) tomean that the chance of inclusion of an NFmarker on a donated fragment decreases withincreasing distance from the 9 o'clock region, inboth directions, and it was shown that bothends of the population of fragments were in factvariable in position. Moreover, it is clear thatthe inheritance by haploid progeny of particularNF markers at loci in the bottom half of themap is essentially independent of the inherit-ance of a particular NF marker from the tophalf of the map (Table 2); in other words, we are

probably dealing with a single population offragments whose average midpoint is at the 9o'clock position. By the alternative hypothesis,that there are two populations both starting at 9o'clock and extending for variable distanceseither clockwise or anticlockwise, every zygotewould have to receive a fragment of each type,and the proximal (9 o'clock) end of both frag-ments would have to be obligatorily inherited.

Although NF strains donate this nonrandompopulation of fragments to the effective zygotesin matings with UF strains, the same NF strainsdonate a random population of fragments to thezygotes when crossed with other NF strains.Moreover, in such NF x NF matings eachparent provides the fragment for some zygotes,thus presumably acting as donor, and the com-plete chromosome for the remainder, presum-ably acting as recipient. This feature was dis-

382 BACTERIOL. REV.


8,< hisAargA-' r

metB ~ 00i0"I. -- 0-4

13 0413 "- --4 - _2

FIG. 4. Composite diagram showing the NF allelefrequency at each of a number of loci among therecombinant progeny of many NF x UF crosses. Theinner circle represents the complete UFgenome of thezygotes, and the outer arc indicates the chromosomefragments, always including the 9 o'clock region of theNFgenome and of variable length, as indicated by thedashed lines. (From ref. 57, with later additions.)

cussed in detail by Hopwood (51) by using datafrom crosses later shown to involve pairs of NFstrains. By carrying out a total nonselectiveanalysis of such a cross, Vivian and Hopwood(102) showed that the combined frequency of allrecombinant classes in NF x NF matings couldreach about 1% of all spores produced on themixed culture.

Although the first eight strains chosen as

mates for the UF strains isolated by Hopwood etal. (57) were all of NF fertility, it soon becameapparent that the culture collection in factcontained strains of at least two fertility types:NF and IF. The latter was named "initialfertility" because the wild-type A3(2) belongs tothis class. It turned out that, by chance, thestrain that had been chosen for the isolationexperiments yielding the first UF strains was IFand that this fertility type, unlike the NF, givesrise to UF strains with appreciable frequency.At this point, it may be helpful to summarize

the main features of the three fertility types inTable 3 before continuing with a description ofthe historical development of the subject.A UF strain is a very efficient tester for

distinguishing NF from IF strains, because incontrast to NF x UF crosses which yield onlyrecombinants, IF x UF crosses yield recombi-

nants at a frequency not usually exceeding one

per 104 spores produced in the cross, or of theminority parental type in cases when one parentis in large excess over the other (see below). Itwas thus a simple matter to determine thefertility of many of the mutant and recombi-nant derivatives of A3(2) isolated in the earlydays of S. coelicolor genetics. Vivian and Hop-wood (102) showed that all mutant derivativesof A3(2) thus far tested resembled it in being IF,and so did the earliest recombinants. However,a recombinant strain which had been isolatedfrom among the progeny of one of the earlycrosses turned out to be NF; whether the changeto NF had occurred as a result of the cross couldnot be determined, because one of the parents ofthe cross was no longer available for testing, anda change from IF to NF might already haveoccurred in this parent. From this recombinantstrain, the NF character had been inherited bythe majority of recombinants isolated fromsubsequent crosses. Nevertheless, some IF x

TABLE 2. Inheritance of markers byhaploid progeny of an NF x UP

crossa

nicA nicA+ nicA nicA+

uraA 344 147 pheA+ 304 138uraA+ 61 59 pheA 101 68

a Data from Table 4 of ref. 57. Numbers representthe frequencies of marker combinations in a sample ofprogeny recovered on nonselective medium.

VOL. 37, 1973 383

TABLE 3. Summary of the properties of the three main fertility types

Fertility type and status of plasmid SCP1Property

IF NF UF

Status of Plasmid SCP 1 Autono- Integrated at 9 o'clock Absentmous region

IF "Random" NF donates fragments cen- Probably heterogeneoustered on 9 o'clock; NF and

ConstitutionofIF 9 o'clock regions segre-

Constitution ofcrs F gate to haploid progenymerozygotes in crosses NF "Random" NF donates fragments cen-with each fertility type tered on 9 o'clock; all hap-

loid progeny inherit NF 9o'clock region

iUF "Random"

Approximate average ' IF 0.01% 10% 0.01%proportions ofrecombinants among NF 1% 100%total spores in crosseswith each fertility type UF 0.001%

Fertility type(s) of JIF IF IF and NF IFarecombinants in crosses NF NF NFwith each fertility type UF UF

a In this case, all progeny (not only recombinants) are converted to IF by plasmid transfer.

NF crosses were found that had yielded both NFand IF descendants. This indication that theNF-IF difference segregated in crosses was con-firmed, and it was shown that the difference, insuch crosses, is chromosomally determined byalternatives at the 9 o'clock position. Any possi-ble infectious conversion of fertility of oneparent by the other in the absence of genetransfer would have been detected, but was notfound (102).

Vivian and Hopwood (102) showed that thezygotes of NF x IF crosses resembled those ofNF x UF crosses; the NF strain acted as adonor of fragments centered on the 9 o'clockregion. The words "polarized" and "asymmet-ric" were used to describe this situation, theformer referring to the fact that one parentfunctions as donor and the other as recipient(89), and the latter to the fact that the frag-ments represent nonrandom regions of thedonor chromosome. However, from the NF x IFzygotes, in contrast to those in NF x UFcrosses, there was no obligate inheritance byhaploid progeny of any region of either genome;thus NF and IF segregated among the progeny.NF x IF matings were also considerably lessfertile than NF x UF matings, yielding up toabout 10% recombinants among all the sporesderived from the crosses.

IF x IF crosses, like NF x NF and UF x UFcrosses, yield mixed ("random") populations of

zygotes, with each parent acting alternativelyas donor and recipient. This accounts for theobservation referred to at the beginning of thissection, that in the early crosses, which musthave been of IF x IF type, complementaryrecombinant genotypes usually had equal, ornearly equal, frequencies. The later departurefrom this state of affairs, also referred to above,must have coincided with the origin of the firstNF strains. As it happened, Sermonti receivedsome strains of each type in 1959 and 1960 andit was possible, by tracing the origin of some ofhis strains from our culture collection, to equatethe IF and NF strains with the R- and R+strains, respectively. Because R- originally gaverise to R+, an analogy of R+ S. coelicolor with F+E. coli, as suggested by Sermonti (87), wasevidently incorrect.The next step was to study the inheritance of

fertility type in IF x UF crosses, and it wasshown by Vivian (100) that all of the sporeprogeny of such crosses, including those of eachparental genotype in respect of characteristicsother than fertility type, were IF. The max-imum frequency of chromosomal recombinantsin such crosses had already been shown (102)not usually to exceed about 10-4. Thus, in an IFx UF cross, the determinant of the fertilitydifference was transferred at a frequency some10,000 times higher than that of chromosomalrecombination: in other words a true infectious

384 HOPWOOD ET AL. BACTERIOL. REV.


conversion had been found. The putative plas-mid involved was called SCP1 (S. coelicolorplasmid 1). Its determination of the differencebetween IF and UF strains explained, byplasmid loss, the high spontaneous fre-quency (approximately to 0.03 to 0.3%; ref. 102)with which UF strains arise in IF cultures, afrequency that was markedly increased by ir-radiation, but not by chemical mutagenesis.Culotta and Puglia (25) confirmed the inde-pendent inheritance of the IF determinant andchromosomal markers in IF x UF crosses, andPuglia et al. (79) confirmed the infectiousconversion of UF to IF. Thus the IF strainsdesignated R- by Sermonti and Casciano (89)and postulated to lack a sex factor present in R+

strains, turned out instead to harbor an autono-mous plasmid and in this respect to be compa-rable with F+ strains of enteric bacteria.

Evidently, NF strains could not harbor thesame plasmid in the same autonomous state asin IF strains. That the plasmid was still presentin NF strains was revealed through the discov-ery by Vivian (100) that SCP1 confers on itshost strain the ability to excrete, and to beresistant to, a diffusible inhibitor of aerialgrowth of UF strains. Because NF strains re-semble IF strains in respect to the inhibitor(Fig. 5), they must also harbor SCP1 and, if it isnot autonomous, it must be associated in someway with the chromosome. The 9 o'clock posi-tion is clearly the region of association, because

FIG. 5. Growth reactions in pairwise crosses of IF, UF, and NF strains. The center plate carries patches of IF,UF, and NF strains, each bearing the marker pheAl. This plate was replica plated to lawns of IF, UF, and NFstrains carrying the marker uraAl to produce the "plate crosses" placed around the central plate. Note theinhibition of aerial mycelium production of the UF culture by IF and NF patches and resistance of the IF andNF cultures to inhibition. (From ref. 52).

VOL. 37, 1973 385

HOPWOOD ET AL.

this region of the NF genome is always inheritedwith the NF character in NF x IF and NF x UFcrosses (see above).Thus, in the sense only that we are dealing

with strains having a chromosomally attachedplasmid (NF), an autonomous plasmid (IF) andno plasmid (UF), the sexual types of S.coelicolor A3(2) may be compared with the Hfr,F+, and F- strains, respectively, of E. coli K-12.We will discuss some of the obvious differencesbetween the two systems at the end of thissection (see also ref. 95). Meanwhile, we shouldnote in particular that the NF type is clearly nota close analogue of Hfr strains, because itgives a bidirectional gradient of donor markerfrequencies in crosses with UF recipients; andall progeny of such crosses are donors, instead ofnearly all being recipients as in E. coli.

New Donor Types

In order to explain the origin of all NF strainsin our culture collection it has been necessary topostulate the spontaneous production of NFfrom IF only once, in or immediately precedinga cross made in 1956 (102). In an attempt toisolate further NF types, or any other kinds ofdonors that might be produced from an IFstrain, Vivian and Hopwood (103) carried outthe procedure that had led to the isolation of UFvariants from IF cultures (see above), but thistime the tester strain was UF instead of NF. Inthis way a strong visual selection was made forcolonies within the IF population that coulddonate efficiently a particular allele, in this casepabA+, to the UF culture. It was found thatdonors could indeed be recognized among the IFcolonies. However, the new donors differed,usually in more than one way, from NF strains.

In the first experiments reported (103), 0.2%of the survivors of UV irradiation contained adonor clone that could be isolated. However,this is not a good estimate because of the factthat the great majority of the new donor typesare unstable, usually markedly so. This meansthat they appear first not as whole colonies inthe IF population, but as sectors of variableextent which have to be picked and streaked anumber of times before they can be obtained inreasonably pure form. Many small sectors maybe missed on the primary isolation plates or lostbefore they can be purified. Even after exhaus-tive purification, many of the new donors are sounstable that a plating from a carefully clonedculture may yield a majority of nondonor colo-nies spontaneously, and an even higher propor-tion after UV irradiation. Only a small minorityof the new donors (1 out of 23 in the sampledescribed in ref. 103) are stable.

A second feature of the new donors was theirwhite colony phenotype (Fig. 6a). The strainshad no morphological abnormality in sporematuration observable by light microscopy;that is, they could be compared with class VIwhite mutants according to the scheme ofChater (17) (see section on Analysis of Dif-ferentiation). The significance of the frequentassociation between ability to act as an efficientdonor of chromosome segments and white col-ony phenotype in these strains is not known. Itis not an obligate association, as shown by thefrequent production of donors of normal greyphenotype from one of the white donor strains(A607); these in turn frequently gave whitedonors again (Fig. 6d) (103).The most constant and novel feature of the

new donors was their pattern of marker dona-tion which turned out to result in aunidirectional gradient of marker frequency inthe haploid progeny of crosses with UF strains.A particular donor region, different for differentnew donors, had a very high frequency ofinheritance and the frequencies of donor mark-ers fell gradually with increasing distance in onedirection from this region, whereas there was avery abrupt cut-off of inheritance even of veryclosely linked donor markers in the oppositedirection (Fig. 7). Thus, the merozygotes in newdonor x UF crosses were postulated to contain apopulation of fragments with one end constant(that carrying the maximally donated region)and the other variable in position.Any hypothesis on the nature of the new

donors has to take account of the followingfacts: (i) these donors still contain SCP1, be-cause they inhibit the aerial development of UFstrains and are resistant to inhibition by IFstrains; (ii) the frequency of inheritance of somedonor markers in crosses with UF strains is ashigh as in NF x UF crosses; (iii) most of thedonors are unstable, often very unstable, pro-ducing numerous UF variants and also some IFvariants; (iv) the progeny of crosses of unstablewhite colony donors with UF strains have thesame unstable donor state and white colonyphenotype as the parent donor; on the otherhand, those few donors that are stable give riseto stable donor progeny; and (v) bidirectionalNF donors arise occasionally from the unidirec-tional donors, either from the donor culturealone or among their progeny when crossed withUF strains; conceivably, all bidirectional donorsisolated from the unidirectional donors in factarise by donor x UF mating, and this may occurwithin a single donor culture, because there isalways a proportion of UF individuals withinsuch a culture.The extreme instability of some of the new

386 BACTERIOL. REV.


FIG. 6. (a) Colonies of white unstable donor strain A607producing numerous grey UF sectors. (b) Plate-crossof plate (a) on a lawn of a UF strain; note inhibition of aerial mycelium production on the background UF cul-ture by white colonies and sectors and lack of inhibition by grey (UP) sectors. (c) Replica from plate (b) to selec-tive medium; note recombinant production by white (donor) colonies and sectors and lack of recombinantproduction by grey (UF) sectors. (d) A slow-growing grey donor derivative ofA607 giving rise to numerous faster-growing sectors of white donor strains.

donors, leading to loss of ability to donatechromosomal markers efficiently, usually si-multaneously with loss of SCP1, to yield UFvariants suggests an autonomous state for theplasmid in the unstable donors. However, theultrafertility of the crosses with UF recipientsargues for an easy association with the chromo-some. We therefore proposed (103) that the un-

stable donors harbored an autonomous SCP1that had incorporated a piece of chromosome af-ter the style of F prime strains of E. coli K-12(1, 44, 63). This piece would represent a differentregion of the chromosome in different newdonors, because two classes, donating a differ-ent marker with the highest frequency, werefound among the 23 strains in the original

VOL. 37, 1973 387

HOPWOOD ET AL.

2 _ - - _5 -o je argA "

proA arA %PryOA/l

strA

FIG. 7. Composite diagram showing the approxi-mate percentage frequencies of donor markers at eachof a number of loci among the recombinant progeny ofnew donor x UF crosses. The inner circle representsthe complete UF genome, and the outer arc indicatesthe chromosome fragments donated by a pabA + donor(A607) with one constant end and one variable(dashed line).

sample: they were called "pabA+ donors" and"uraA + donors" after the marker donated withthe highest frequency. Having, on this hypothe-sis, an appreciable region of homology with thechromosome, the plasmid might be expected tointeract readily with it during or before itstransfer to a UF recipient, possibly by "donorcrossing over" (11) or perhaps by the so farpoorly understood mechanism by which the Ffactor of E. coli promotes chromosome transferby an F+ culture (36). A transient interactionwas suggested by the finding that the progeny ofnew donor x UF crosses resembled the donorparent in their donation properties, while beingrecombinant in respect of normal markers. Theinstability of long F' strains of E. coli K-12 hasbeen observed, but even so the extreme instabil-ity of some of the S. coelicolor donors is remark-able on the proposed model. The production ofIF strains from the new donors would, by thishypothesis, perhaps occur by internal crossingover within the modified plasmid, resulting inthe elimination of the inserted piece of chromo-some.

This hypothesis is not applicable to the stabletypes of new donors, and probably these resultfrom a more permanent association between theplasmid and the chromosome, as in NF strains,but presumably in a different configuration. A

straightforward integration of the Hfr type is apossibility. A satisfactorily simple model toexplain the behavior of NF strains is stillelusive. The finding that the derivation of NFfrom IF strains probably involves two steps mayindicate that the mechanism of chromosomemobilization in NF strains is more complexthan in the unidirectional donors.

Since the isolation of the first group of newdonors (103), an assortment of other donor typeshas been isolated from various IF strains (D. A.Hopwood, A. Vivian, and H. M. Wright, unpub-lished data). In general terms these may also beexplicable by interactions between SCP1 andvarious regions of the chromosome. However,one recently isolated donor behaved in a muchmore readily interpretable fashion, and indeedthis strain is clearly analogous with an F' strainof E. coli K-12, because it harbors an SCP1plasmid that has incorporated a recognizedchromosomal marker. The strain has beencalled an SCP1' strain (D. A. Hopwood and H.M. Wright, manuscript submitted for publica-tion), and the particular substituted plasmidwas designated SCP1-cysB, because it carriesthe chromosomal cysB locus. The strain wasevidently a secondary (merodiploid) SCP1'strain, rather than a primary (haploid) strain,because the plasmid was easily lost from it,giving viable UF variants.

Part of the evidence for the SCP1' nature ofthis strain was the finding that, whenever theplasmid was transferred to a cysB UF recipient,as occurred with the same high frequency astransfer of the wild-type SCP1 plasmid from IFto UF, the progeny were converted to phenotyp-ically Cys+, but genotypically heterozygous(cysB/cysB+), strains that could lose the SCP1-cysB plasmid to revert to cysB UF derivatives.It was also shown that the SCP1-cysB plasmidcould be transferred to another wild-type strain,1326, to which the wild-type SCP1 had alreadybeen shown to be transferable (see below; D. A.Hopwood and H. M. Wright, J. Gen. Microbiol,in press).Markers not borne on the SCP1-cysB plasmid

were transferred by the SCP1' strain to UFrecipients with frequencies below 10- 2, andsometimes as low as 10-5 Thus, for somemarkers, they overlapped with frequencies in IFx UF crosses (see below), and it was difficult toanalyze any specific pattern of marker transferattributable to the SCP1-cysB plasmid over andabove that due to properties of the normal SCP1plasmid. The pattern was clearly not the sameas that in the new donors, which were postu-lated to be SCP1' strains harboring plasmidsthat happened to carry unmarked regions of the

388 BACTERIOL. REV.


chromosome. Possibly the chromosomal inser-tion in the SCP1-cysB plasmid, which repre-

sented less than 2% of the linkage map, was tooshort for efficient "donor crossing over" tooccur. However, much more needs to be learnedabout such strains before a definitive interpre-tation of marker transfer of nonplasmid-linkedgenes is possible.

Role of the Plasmid SCP1 in Fertility

UF x UF crosses, in which SCP1 is absentfrom both mates, are far from sterile (102).Table 4 illustrates a range of recombinationfrequencies in such crosses. Moreover the gene

transfer in UF x UF crosses shows no evidenceof occurring by a mechanism basically differentfrom that operating in crosses involving IF or

NF strains, where SCP1 is present; on thecontrary, the hallmark of a conjugationsystem-the inheritance by the progeny of longregions of linked markers-is present in UF x

UF crosses. (Dowding and Hopwood [J. E.Dowding and D. A. Hopwood, J. Gen. Micro-biol., in press] used a UF x UF cross to map thesite of attachment of the prophage of VP5 to thechromosome: see below.) Thus SCP1 is clearlynot necessary for conjugation in S. coelicolor,and we should bear in mind that not all of thefeatures of the conjugation observed in crosses

involving IF or NF strains need necessarily bedue to SCP1. In any case, an additional system

controlling conjugation in UF x UF crosses

evidently awaits elucidation.IF x UF crosses are not always significantly

more fertile than UF x UF crosses (102). It isnot easy to obtain meaningful estimates ofrecombination in crosses of either type, becausedifferences in the vigor of asexual reproductionof the various strains lead to widely differentvalues for the proportion of recombinant sporesin the cross, even if these are expressed in termsof the minority parent as in Table 4; differentselections in the same cross may also lead todiverse estimates. However, a general tendencyfor IF x UF crosses to be more fertile than UF x

UF crosses is apparent (Table 4). Moreover,comparison of the patterns of marker inheri-tance in pairs of IF x UF crosses involving thesame groups of markers in the two parents, butwith the fertility reversed, shows that SCP1does indeed play a role in gene transfer, even

when we are not dealing with strains in which itis demonstrably associated with the chromo-some. In Fig. 8a the "inner" parent is IF, and itis seen to act preferentially as a donor, becausethe frequency of IF markers at loci such as cys,far from the selected marker (pro+), are low. InFig. 8b the "outer" parent is IF, and here thereis a reverse polarization with a strong tendencyfor the inheritance of only a limited region of theIF chromosome around the selected marker(str).

TABLE 4. Recombination frequencies in some UF x UF and IF x UF crosses

UF parent 1 UF parent 2 Recombinants

Genotype No.a Genotype No. Selected markers |qu-eny

PheAl 1.9 x 10' hisAl uraAl strAl 9.1 x 106 phe+ his' ura+ 20pheAl 9.8 x 106 hisAl uraAl strAl 1.4 x 10' phe+ his+ ura+ 39proAl argAl cysDl8pheAl strAl 1.3 x 106 ade-vlO uraAl 4.5 x 10' pro+ str 6proAl argA l cysDl8pheAIstrAI 4.8 x 106 ade-vlOuraAI 1.1 x 10' pro+ str 1

proAIargAl cysD18pheAI strAI 1.2 x 10' hisC9 1.4 x 10{ phe+ his+ 0

proAI argAl cysD18 pheAl strA | 2.0 x 10' hisC9 6.5 x 106 {pher+ s+ 94

IF parent 1 UF parent 2 Recombinants

pheAl 1.0 x 106 hisAl uraAl strAl 2.0 x 10. phe+ his+ ura+ 10hisAl uraAl strAl 8.2 x 106 pheAl 7.8 x 106 phe+ his+ ura+ 87proAl argAl cysDl8pheAl strAl 7.0 x 10' ade-vlO uraAl 2.3 x 109 pro+ str 77ade-vOuraAl 1.0 x 109 proAl argAl cysDl8pheAl strAl 2.5 x 10' pro+ str 19

hisC9 8.0 x 106 proAl argAI cysD18pheAl strAl 5.0 x 106{ pro+ str130

hipo+st 1,50

proAl argAl cysDl8pheAl strAl1 6.0 x 10' hisC9 6.2 x lO{ phe+ his+ 1,500

aPer ml of suspension from the cross.' Per 106 spores of minority parent.

VOL. 37, 1973 389

HOPWOOD ET AL.

4argA

strA strA100. 100

FIG. 8. Allele frequencies (percentage) among selected recombinants (triangles indicate the selected alleles)in crosses of IF and UF strains: IF (inner) x UF (outer), UF (inner) x IF (outer). Both crosses involved strainsbearing the markers ade-v1O uraAl and argAl cysD18 strAl pheAl proAl.

As we have seen in the previous section, IFcultures contain a low frequency of donor clonesable to transfer particular chromosomal regionsefficiently. These presumably contribute to themarker transfer in IF x UF crosses. Thisinterpretation is supported by the results of anexperiment in which the recombinant progenyof an IF x UF cross were fertility tested by theplate-crossing technique (D. A. Hopwood andH. M. Wright, unpublished data). It was foundthat some of the progeny could be classified asdonors, with a higher fertility against a UFtester strain than that shown by IF strains,while the remainder of the progeny were indis-tinguishable in this test from typical IF cul-tures. The two classes presumably corre-sponded to two possible modes of recombinantformation in IF x UF crosses.At the upper extreme of the fertility scale, the

NF x UF cross, it is clear that the presence ofSCP1 associated with the 9 o'clock region of theNF genome leads to high frequency transfer ofchromosome fragments carrying this region.However, it is not excluded that other classes offragments might be transferred, but give rise tononeffective zygotes. The suppression of UFaerial mycelium by the diffusible product ofSCP1 may explain the absence of UF sporeprogeny and conceivably, therefore, the ob-served obligate inheritance of SCP1, and henceof its chromosomal attachment point. Hopwoodet al. (57) found a considerable proportion of the

UF parent in the young substrate mycelium ofNF x UF crosses, but not in the spores ofmature cultures; however, a careful searchamong the earliest recombinants formed in suchcrosses, and recovered in mycelial fragmentsfrom immature crosses, failed to reveal any ofUF fertility (D. A. Hopwood and H. M. Wright,unpublished data), tending to argue against theidea that UF recombinants arise, but are simplyprevented by the SCP1 product from beingrecovered. Sermonti, Puglia, and Ficarra (92),in a detailed study of the time course ofrecombinant production in NF x UF crosses,also failed to find an appreciable frequency ofUF recombinants even at the earliest samplingtimes, although the occurrence of rare UFrecombinants at all sampling times, includingthose when mature sporulating cultures wereinvolved, was recorded. In any case the reasonfor the failure of the NF parent to contributespores to the mixed culture is unclear.

Vivian and Hopwood (103) found evidencethat SCP1, carrying with it the region of thechromosome with which it is postulated to betransiently associated in the unstable newdonor strains, is inherited obligatorily by theprogeny of crosses of these donors with UFstrains. Probably the same mechanism operatesin these ultrafertile crosses as in NF x UFcrosses.The role of SCP1 in controlling fertility in S.

coelicolor will doubtless be illuminated when

390 BACTERIOL. REV.

I


mutations affecting particular plasmid func-tions are isolated and studied, as in the case ofthe F factor of E. coli K-12 (114). Suggestiveevidence is already available to indicate that itwill be possible to analyze the group of pheno-typic characters controlled by SCP1, whichincludes at least the following: production of thediffusible inhibitor ofUF strains; determinationof resistance to the inhibitor; and the variousaspects of plasmid maintenance and transfer.Although most of the new donors already iso-lated still possess the first two properties (103),some that we have recently isolated fail toinhibit UF cultures, although they are stillresistant to inhibition by IF strains, whereasothers fail to inhibit and are themselves sensi-tive to inhibition (D. A. Hopwood and H. M.Wright, unpublished data). If all of the donorsturn out to represent various kinds of substitu-tion of pieces of chromosome into the plasmid,those altered in production or response to theinhibitor may be interpreted as having sufferedvarious losses of plasmid functions by deletionsof varying extent. Conceivably, the extremeinstability of some of the donor strains mayresult from loss of genes concerned in plasmidmaintenance.

Transfer of SCP1 from A3(2) to Other StrainsIt is characteristic of plasmids of gram-nega-

tive bacteria that they can be transferred byconjugation between strains that are not closelyrelated: such transfer has been observed amongmembers of the genera Escherichia, Salmo-nella, Proteus, Pasteurella, Shigella, Serratia,Erwinia, Pseudomonas, Rhizobium, Agrobac-terium, and Klebsiella (4, 19, 27, 30, 66).Because nothing is known of the morphologicalbasis of conjugation in S. coelicolor or otherstreptomycetes, we cannot say, a priori,whether conjugation between disparate strep-tomycetes is likely to occur. In studies ofgram-negative eubacteria, great use has beenmade of selective plasmid-linked markers todetect plasmid transfer, even when it occurs ata low frequency: prototrophic markers in thecase of substituted F factors (F's) or, moregenerally, drug resistances in the case of Rfactors. In S. coelicolor it is now possible to usea prototrophic plasmid-linked marker in anSCP1' strain to detect plasmid transfer betweendifferent wild types (D. A. Hopwood and H. M.Wright, manuscript submitted for publication).Earlier, the inhibitor of UF growth determinedby SCP1 (100) served to show that SCP1 can betransferred in a cross from IF strains of A3(2) toanother wild-type strain of the S. coelicolor

species group, strain 1326, converting it from astrain inhibited by A3(2) strains of IF fertilitytype to one resistant to inhibition and itselfcapable of inhibiting A3(2) strains of UF type(D. A. Hopwood and H. M. Wright, J. Gen.Microbiol., in press). One to three percent of1326 spores harvested from such a cross hadreceived the plasmid, as determined by a visualprocedure (Fig. 9). By this procedure, we haveso far failed to detect transfer of SCP1 fromA3(2) to strains less closely related to it than isstrain 1326, but the use of selective proceduresmay soon detect transfer at low frequency tosuch strains, just as selection allowed the dem-onstration of transfer of SCP1-cysB back from1326 to A3(2) at a frequency of about 10-6 (D. A.Hopwood and H. M. Wright, manuscript sub-mitted for publication).

Comparison of Gene Transfer in S.coelicolor with That in E. coli

The wild-type S. coelicolor A3(2) resemblesthe original strain of E. coli K-12 in harboringan autonomous plasmid that can be lost, canacquire an insertion of a chromosomal region, orcan associate itself with the chromosome. Thus,the IF type of S. coelicolor may be comparedwith the F+ form of E. coli K-12. In IF x UFcrosses, as in F+ x F- matings, it would appearthat some gene transfer may be due to donorclones within the IF culture, involving more orless permanent interaction between SCP1 andthe chromosome, while some may not. Thissituation, in S. coelicolor, has not been studiedin any detail, whereas in E. coli it has received agood deal of attention (26, 36) without a com-pletely satisfactory explanation emerging.Apart from these simple resemblances there

are several differences between the two systems.One area of difference concerns the UF strainsof S. coelicolor, lacking the known plasmidSCP1. These, unlike F- strains of E. coli, arefertile, albeit at a relatively low level, whencrossed with each other. This may mean that asecond plasmid promoting conjugation andgene transfer, and present in all of our strains,remains to be identified or, alternatively, thatrecombination can occur in the absence of aplasmid. It may be that SCP1 itself has nothingto do with the formation of conjugation unionsbetween the cells, a function that might becontrolled by genes of a second plasmid or bychromosomal genes. However, SCP1 is trans-ferred very efficiently in IF x UF matings (100)and in association with pieces of donor chromo-some in NF x UF (57) or new donor x UF

391VOL. 37, 1973

HOPWOOD ET AL.

FIG. 9. Colonies of prototrophic strain 1326 converted to IF by growth in mixed culture with an auxotrophicIF derivative of A3(2). The products of the mixed culture were plated with an excess of pure strain 1326 onminimal medium, selecting against the A3(2) auxotroph. The IF-converted colonies are surrounded by a zone ofinhibition of aerial mycelium production on the unconverted strain 1326; the unconverted colonies are not seensince they merge with the background growth of strain 1326. (D. A. Hopwood and H. M. Wright, J. Gen.Microbiol., in press).

crosses (103), and this suggests that it is respon-sible for its own transfer, rather than dependenton an association with some other unidentifiedtransmissible plasmid, unless this were an ex-tremely efficient association.

Another area of difference from E. coli K-12concerns the behavior of strains capable ofdonating segments of chromosome. Hfr x Hfrmatings display a very low fertility unless one isdeliberately grown under physiological condi-tions leading to a female "phenocopy" (67), andthen the cross is obviously polarized in favor ofthe phenocopy acting as recipient. In S.coelicolor, NF x NF crosses are very fertile, andthis is not due merely to spontaneous UFphenocopies within the cultures, because thepattern of marker inheritance reveals the trans-fer of random fragments to the zygotes, ratherthan the special populations of fragments, cen-tered on the 9 o'clock region, involved in NF xUF crosses. NF x IF crosses, also, are veryfertile in contrast to Hfr x F+ matings. Otherdifferences, already discussed, concern the do-nation of fragments extending symmetrically onboth sides of the point of interaction of theplasmid with the chromosome in NF strains andthe inheritance of donor status by the recombi-

nant progeny. Certain of the stable new donorstrains resemble the Hfr model more closely inthat the donated fragments probably have oneconstant end, but again the recombinant prog-eny are donors rather than recipients.

In conclusion, we may say that the finding ofa conjugation system, at least in part plasmidcontrolled, in a complex, differentiated, gram-positive bacterium, S. coelicolor, is of great in-terest in indicating the widespread importanceof plasmids as genetic determinants in bacteria.These studies have added the streptomycetes tothe gram-negative enteric bacteria, pseudomo-nads, and their relatives and the gram-positivestaphylococci as examples of plasmid-carryingbacteria, although it should perhaps be notedthat to date there is only genetic evidence forthe existence of a plasmid in S. coelicolor (orany other streptomycete). There is, however, noreason to suppose that physical or biochemicalevidence, or both, will not be forthcoming.Additional work will be required to elucidatethe morphological basis of conjugation in S.coelicolor. Sermonti, Puglia, and Ficarra (92)have attempted to analyze the mating processin NF x UF crosses by blending the mixedcultures after different periods of incubation

392 BACTERIOL. REV.


and determining the relative proportions ofvarious classes of recombinant progeny amongthe plating units obtained. They found evidencethat the proportion of recombinants inheritingdonor markers some distance from the 9 o'clockregion tended to increase with time, butwhether this finding is to be interpreted asindicating a basic similarity between the mat-ing and genetic transfer processes of S.coelicolor and E. coli remains to be seen.

ANALYSIS OF DIFFERENTIATION

Phases of Colonial DevelopmentThe mature colony of S. coelicolor on agar

media is in effect a multicellular, differentiatedorganism comprising: the substrate mycelium,which has a purely vegetative function, growingon and into the medium; the aerial mycelium,whose function is the production of spores; andthe spores themselves, which represent thedistributive phase of the organism and arerelatively resistant to desiccation (47, 109).After spore germination, the early growth of thecolony is entirely vegetative, and young coloniesare translucent and "bald" (i.e., devoid of aerialmycelium). The formation of aerial branches ismacroscopically visible as an opaque, chalkylayer on the surface of the colony, which thenturns grey as the aerial hyphae form sporechains.

Ultrastructural studies show that aerial hy-phal cells, unlike those of the substrate myce-lium, possess a distinctive superficial layer, the"fibrous sheath" (55, 113), the function of whichis not known, but which persists even on de-tached mature spores. Comparable superficiallayers are also found on many other strep-tomycetes (110, 111, 115).

Aerial hyphae form spore chains through asequence of events involving coiling of thehyphal tip, subdivision of the long apical cell bythe simultaneous, double annular ingrowth of anumber of closely spaced sporulation septa, androunding up of the newly formed spore unitsaccompanied by thickening of the spore wall(112; A. C. McVittie, manuscript submitted forpublication). Spores generally contain a singlegene complement. It should be emphasized thatthe structure and mode of formation of thesespores is completely different from that in-volved in endospore formation in, for example,Bacillus, Clostridium (84), Thermoactinomyces(23, 24, 32, 75), and other bacteria of doubtfulrelationship (93).

Thus, S. coelicolor has macroscopic featureswhich facilitate the detection of morphological

mutants, microscopic features to help in under-standing the nature of mutational blocks, andan excellent recombination system permittinggenetic analysis of mutants. These facilitieshave been exploited in studies of white (whi)and, to a lesser extent, bald (bld) colony mu-tants (17, 18, 60; A. C. McVittie, manuscriptsubmitted for publication).

Characterization of Morphological MutantsPhase contrast microscope observation of the

aerial mycelium of more than 50 whi mutantshas led to the identification of six broad pheno-typic classes (Fig. 10) ranging from phenotype I,in which the aerial hyphae are "fibrous" andshow neither helical coilings nor any sign ofspore formation, to phenotype VI, in whichnumerous spores sometimes indistinguishablefrom wild-type spores by presently availablecriteria, are formed (17; A. C. McVittie, manu-script submitted for publication). Thus the useof whi mutants will help to identify a sequenceof events in sporulation, each of which occursonly upon accomplishment of earlier events. Itis, however, possible that there are aspects ofsporulation that are not dependent upon thissequence and which will not, therefore, bedetected by the use of whi mutants. Alternativemethods of isolating morphogenetic mutantswould be very desirable.

Genetic mapping of the whi mutations has sofar revealed nine distinct map locations (Fig. 2)(17, 60). Several mutations have been found atmany of these locations, probably indicatingthat few loci remain to be identified. Mutationsat each locus generally result in the samephenotype. (An exception to this is whi-99(phenotype V), which is located very close to thewhiG locus (phenotype I) and which may repre-sent a 10th locus, whiF (17)). Thus, it isgenerally possible to select representative mu-tants of each locus for further study. In theabsence of meaningful complementation anddominance tests, which have not yet beendevised for whi mutants, the possibility thatsome whi loci are gene clusters has not beeneliminated, although the frequencies of recom-bination between mutations mapping togetherhave suggested that whiA, E, G, H, and I aresingle genes (17 and K. F. Chater, unpublisheddata).Knowing the map locations and phenotypes

of various whi mutations, we can attempt toanswer the question of the sequence of geneexpression (where the phenotypes leave this indoubt) by examining the phenotypes of appro-priate double whi mutants. To date, whiG

393VOL. 37, 1973

HOPWOOD ET AL.

5%

S

FIG. 10. Phase-contrast photomicrographs of aerial growth of whi mutants. (1) whiG, phenotype I; (2) whiA,phenotype II; (3) whiH, phenotype III; (4) whil, phenotype IV; (5) whiF, phenotype V; (6) whiD, phenotype VI(in this example the spores are abnormally spherical due to a defect in spore wall synthesis); (7) wild-type. Barmarkers indicate 10,im.

394 BACTERIOL. REV.


(phenotype I) and H (phenotype III) have beenfound to be epistatic to whiA and B (bothphenotype II) and to whiI (phenotype IV), andwhiG is almost certainly epistatic to whiH (K.F. Chater, unpublished data). Morphologicalstudies (17, 60; A. C. McVittie, manuscriptsubmitted for publication) have shown thatwhiA, B, G, and H mutants are all unable tomake sporulation septa, while in whiI mu-tants these septa are abnormally spaced; thesingle whiD mutant has thin-walled spores; thesingle whiF mutant makes spores that fail toround up; and whiE mutants make morphologi-cally apparently wild-type spores. Combiningthe genetic and morphological information, thesequence of events given in Fig. 11 is indicated.The place in this scheme of two mutants,whiC193 and whi-53, which are both type I, butoligosporogenous (17; A. C. McVittie, manu-script submitted for publication), is uncertain.

Only a small number of bld mutants (Fig. 12)have been isolated; one of these (bldAl) islocated between ilvA and pel, apparently nearto the whiB locus, whereas others have beenlocated, more approximately, in the 1, 6, and 9o'clock regions (D. A. Hopwood, K. F. Chater,and E. N. Lawson, unpublished data).

Biochemical and Physiological Approachesto the Study of Differentiation in S. coelicolor

Clearly, there must be chemical signals in-volved in triggering key events in differentiation(particularly in aerial mycelium formation),and these signals must be transmitted to thetranscription and/or translation machinery ofthe cell. We have so far only begun attempts toanalyze the role of transcriptional control indifferentiation, and our efforts have been con-fined to preliminary in vitro studies of deoxyri-bonucleic acid (DNA)-dependent ribonucleicacid (RNA) polymerase and to the isolation ofrifampin-resistant mutants. Three resistanceloci have been mapped, and mutations in onlyone (rifA) appear to decrease the sensitivity ofRNA polymerase to rifampin in crude extracts(K. F. Chater, manuscript submitted for publi-cation). At this stage, none of about 100 rifAmutants isolated has been found to possessmorphological abnormalities attributable to therifA mutation. This contrasts with the situationin B. subtilis, where Sonenshein and Losick (94)and Doi et al. (31) had no difficulty in findingrifampin resistant mutants altered in sporula-tion.Our studies of the enzyme itself have not yet

proceeded beyond preliminary characterization,

but several interesting features of the enzyme,not connected with differentiation, haveemerged (K. F. Chater, manuscript submittedfor publication and unpublished data). Likethat of Bacillus subtilis, the enzyme is highlysensitive to KCl; it behaves like that of otherbacteria (e.g., E. coli; ref. 10) in its sedimenta-tion characteristics in low and high salt glycerolgradients; it is scarecely inhibited at all byrifampin at 1 Ag/ml and only about 70% at 20,ug/ml, in contrast to the high sensitivity of theenzymes of all other bacteria examined (104);and it is unusually sensitive to streptolydigin(50% inhibition at about 0.75 ,g/ml) comparedwith other bacterial RNA polymerases (e.g., E.coli where about 5 ,ug/ml gives 50% inhibition;ref. 14). We have found that the enzyme ofStreptomyces albus strain CMI 52766, on theother hand, is sensitive in crude extracts torifampin concentrations as low as 0.01 ,ug/ml,but is also highly sensitive to streptolydigin(K. F. Chater, manuscript submitted for publi-cation).A particularly interesting possibility opened

up by advances in streptomyces phage tech-niques (see below) is that phages may be foundwhose DNA is transcribed only by RNA polym-erase from vegetative hyphae, and not by thatfrom sporulating hyphae. This would provide anelegant test of RNA polymerase template speci-ficity analogous to that described by Losick andSonenshein (71) for B. subtilis, where vegeta-tive, but not sporulation, RNA polymerasetranscribes phage qe DNA. At present, thisapproach is hampered by the difficulty of ob-taining either phage DNA or sporulating hy-phae in useful quantities.

GENETIC ASPECTS OF PHAGES

Isolation and Characterization of PhagesAlthough bacteriophages attacking strep-

tomycetes have been known for more than 35years (107, 108), it is only very recently thatphages for S. coelicolor A3(2) have been iso-lated. They have come from two sources.Dowding (J. E. Dowding, Ph.D. thesis, Univ.

of East Anglia, Norwich, England, 1972) iso-lated 28 phages from soil samples either byplating membrane filtrates of soil suspensionsdirectly in top layers with the host or afterovernight enrichment of the soil sample byincubation in nutrient broth with host spores,followed by membrane filtration and plating asbefore. The specific enrichment method provedmore reliable, resulting in the isolation of

395VOL. 37, 1973

HOPWOOD ET AL.

whiF

rounding up

whiE

grey pignt,matI& rbeese

sporuationseptation

whiAwhiBwhil

freespore

0

- ni& outgrowth

completion ofcoilng

whiH

substratemycelium

growth of WIdAaerc hyae

initiation ofcoiling

whiGwhic?

FIG. 11. Diagrammatic representation of the genetic control of morphogenesis in S. coelicolor. Thedesignations of genes thought to be implicated in each step are given outside the arrows. The diagramincorporates information obtained from genetic mapping tests of epistasis and light and electron microscopicstudies (17, 60; A. C. McVittie, manuscript submitted for publication; K. F. Chater, unpublished data).

phages from 25 out of 28 soil samples. Theplaque morphologies of this group of phages fellinto four types (Fig. 13). The group of phageswith type IV plaques may perhaps be hetero-geneous, the small size of the plaques simplyreflecting the fact that optimal conditions forplaque development by these phages have notyet been defined. A representative phage withtype III plaque morphology (large, clearplaques) is the virulent phage VP11 (33),whereas the temperate phage VP5 (J. E. Dowd-ing and D. A. Hopwood, J. Gen. Microbiol., inpress) is an example of a phage with type IIplaque morphology, the plaques having a turbidmargin. During prolonged incubation the

plaques of VP5 increase in size and come tocontain concentric rings of surviving and lysedhost (Fig. 14), an appearance that has beeninterpreted in terms of changes in the multiplic-ity of infection and consequently in the relativeprobabilities of the lytic cycle and of lysogeny(J. E. Dowding and D. A. Hopwood, J. Gen.Microbiol., in press).The second source of a phage for strain A3(2)

is the strain itself. Lomovskaya et al. (68)showed that A3(2) carries a defective prophage,but nevertheless releases phage at a very lowfrequency, presumably as a result of mutationof the prophage to a nondefective form, followedby spontaneous induction. The phage so pro-

396 BACTERIOL. REV.


FIG. 12. Colonies of wild-type and bld (bald) colony mutants; the latter are distinguished by their smooth,shiny appearance. (x5.5).

duced, OC31, lysogenized an indicator strainand also A3(2), which had been cured of thedefective prophage by UV irradiation.Dowding (J. E. Dowding, Ph.D. thesis, Univ.

of East Anglia, Norwich, England, 1972) ex-amined six of his phages by negative staining inthe electron microscope and found that they allhad long, striated, flexible tails, thus belongingto group B of Bradley's morphological classifi-cation (8). The virulent phage VP11 and thetemperate phage VP5 had slightly differenthead and tail dimensions (33; J. E. Dowding andD. A. Hopwood, J. Gen. Microbiol., in press)(Fig. 15). No icosahedral RNA phages or othersex-specific phages have yet been described forS. coelicolor or any other streptomycete, and itmay be that they do not exist. Those associatedwith E. coli (22), Pseudomonas (9), andCaulobacter (85) have all been shown to adsorbto pili, and so far these have not been demon-strated in Streptomyces.

Phage Mutants and RecombinationLomovskaya et al. (69) isolated clear-plaque

mutants of the temperate phage 0C31, as wellas mutants able to grow at 37 C, a temperaturetoo high for reproduction of the wild-type

phage. By mixedly infecting the host withphages differing in these two characteristics, itwas demonstrated that phage recombinationoccurred.

Clear-plaque mutants of phage VP5 were alsoisolated (J. E. Dowding and D. A. Hopwood, J.Gen. Microbiol., in press), spontaneously orafter mutagenesis of free phage by UV orhydroxylamine; they have been used in a searchfor nonsense suppressors (V. Najfeld, personalcommunication) but so far not in recombinationexperiments. In the virulent phage VP11,Dowding (33) obtained a high yield of tempera-ture-sensitive mutants and plaque morphologymutants after mutagenesis of intracellularphage by NTG (97) and used them to demon-strate phage recombination. A reproduciblevalue of 5 to 6% wild-type recombinants wasobtained in crosses between a pair of tempera-ture-sensitive mutants.

LysogenyProbably because of their mycelial growth

habit, which results in a culture containing cellsof diverse morphological and physiologicalstate, some unfavorable for phage multiplica-tion, streptomycetes are liable to show the

VOL. 37, 1973 397

HOPWOOD ET AL.

FIG. 13. Four plaque types among phages capable of growth on S. coelicolorA3(2). I, VP9, II, VP5; III, VPJ1;IV, VP19. (From J. E. Dowding, Ph.D. thesis, Univ. of East Anglia, Norwich, England, 1972.)

phenomenon of "pseudolysogeny" (106). A viru-lent phage usually lyses only a part of a culturebefore phenotypically resistant cells arise, andthese can be propagated indefinitely, along withthe phage, superficially mimicking true ly-sogeny unless the culture is rigorously cloned,preferably under conditions unfavorable tophage adsorption, to free it from trivially carriedphage.True lysogeny has, however, been demon-

strated in several streptomycetes (69, 105, 106;J. E. Dowding and D. A. Hopwood, J. Gen.Microbiol., in press); the temperate phages

involved, with the exception of VP5 and twoother closely related phages, were derived froma lysogenic culture shown to carry a prophageby the production of plaques on a sensitiveindicator strain.The two phages OC31 and VP5 are heteroim-

mune as shown by the fact that A3(2) issensitive to VP5, but resistant to OC31, unlesscured of the defective OC31 prophage, and bythe finding that an indicator strain sensitive toOC31 is still sensitive to it when lysogenized byVP5 (J. E. Dowding and D. A. Hopwood, J.Gen. Microbiol., in press). Mapping experi-

398 BACTERIOL. REV.


ments, in which defective lysogens of 4C31 (68)or VP5 (J. E. Dowding and D. A. Hopwood, J.Gen. Microbiol, in press) were crossed withnonlysogenic strains, have revealed a chromo-somal attachment site for each prophage (Fig.2). They turn out to have interesting locations,because certainly VP5, and probably pC31,integrate in the long "silent region" of thelinkage map (50) which contains no normalmarkers, raising the possibility that the geneticlength of this region may in part be due toenhanced crossing over through some kind ofsite-specific recombination involving prophageattachment sites (J. E. Dowding and D. A.Hopwood, J. Gen. Microbiol., in press).

TransductionThere is still no unambiguous demonstration

of transduction in a streptomycete; the firstreport of transduction in S. olivaceus (3) has notled to a reproducible system, and preliminaryattempts to demonstrate transduction by VP5in S. coelicolor A3(2) gave negative results (J.E. Dowding, Ph.D. thesis, Univ. of East Anglia,Norwich, England, 1972). However, now thatseveral temperate phages for this strain have

been isolated, there is a good chance thattransduction will become a reality.

RADIOBIOLOGY

Ultraviolet-Sensitive MutantsNinety-four ultraviolet-sensitive mutants,

isolated by a simple replica-plating technique,have served to identify six loci concerned withthe repair of UV damage in S. coelicolor A3(2)(41-43). Three of these, uvsC, A, and D, areclosely clustered, two others, uvsB and E, arediametrically situated with respect to the firstcluster, and the sixth, uvsF, is not closely linkedto either of the two clusters.The functions of these - loci have not been

defined. However, mutations in some of themcould be differentiated phenotypically. uvsA, B,C, and D mutants had survival curves with ashoulder and an exponential region as in thewild type, but showing a sensitivity increasedby a factor of 7 to 20; the two uvsB mutationsappeared to confer somewhat less sensitivitythan most uvsA, C, and D mutations. Muta-tions in uvsE differed from the others in abol-ishing the shoulder of the wild-type curve butretaining a similar slope in the exponential part

4~~~

FIG. 14. Three-day-old plaques of the temperate phage VP5 on S. coelicolor A3(2) showing a clear, lysedcenter and ring of lysogenised surviving growth surrounded by a further ring of lysis. (x5).

VOL. 37, 1973 399

HOPWOOD ET AL.

FIG. 15. Electron micrographs of virulent phage VPJ1 (left) and temperate phage VP5 (right) negativelystained with neutral potassium phosphotungstate. Marker bar represents 100 nm (from ref. 33; J. E. Dowdingand D. A. Hopwood, J. Gen. Microbiol., in press).

of the curve, which in the case of the uvsEmutants began from the origin. The singlemutation characterizing uvsF resulted in noincrease of UV sensitivity unless combined witha uvsA, C, or D mutation, when the doublemutant had a greatly increased sensitivity com-pared with the single uvsA, C, or D mutants.The incidence of mutations in the different

loci was far from random; there were only twouvsB mutations among the first 23 mutantsisolated (41, 42) and none among the next 71mutants (43). Thus, other loci affecting UVsensitivity could well await discovery. In partic-ular, no mutants with a recognizable effect onrecombination or on X-ray sensitivity have yetbeen found.

It seems likely, particularly in view of thefinding that the uvsA, C, and D loci control host

cell reactivation of UV-irradiated phage VP11(33), that these three genes are concerned withthe excision repair well characterized in otherbacteria and would therefore be the counter-parts of the uvr loci of E. coli K-12 (62). Haroldand Hopwood (41) concluded, from a considera-tion of the survival curves of double mutants(uvsA uvsB, uvsC uvsB, uvsD uvsB), that uvsBmight also be concerned with part of the samesystem. However, the finding that a uvsB muta-tion had no effect on host-cell reactivation ofUV-irradiated phage VP11 (33) makes this lesslikely. In formal terms, the uvsE locus can besaid to be responsible for a repair systemsaturable at low UV doses, hence the lack of theshoulder in the survival curves of uvsE mutants,whereas the uvsF locus can be postulated to becapable of repairing a proportion of the damage

400 BACTERIOL. REV.


also reparable by the uvsA, C, and D system(41); this locus also had no detectable effect onrepair of phage VP11, even in the presence of auvsD mutation (33).

PhotoreactivationPhotoreactivation was first discovered in

Streptomyces griseus (65). Hence, it came as asurprise that S. coelicolor A3(2) is incapable ofenzymatic photoreactivation (41). This is not acharacteristic of all strains of S. coelicolor,however, some of which show a classical photo-reactivation with a dose reduction factor ofabout 0.5 (41).

ANALYSIS OF BIOSYNTHETIC ANDDISSIMILATORY PATHWAYS

What little is known of biosynthetic pathwaysin S. coelicolor depends almost wholly on theresponses of biochemical mutants to biosyn-thetic intermediates and on genetic mappingand complementation. Specific studies of bio-synthetic pathways have been initiated forhistidine, tryptophan, and the aromatic aminoacids.

Tryptophan SynthesisEngel (34 and personal communication) has

studied 28 tryptophan auxotrophs (trp) and hasrecognized at least four loci on the basis ofphenotypes, including enzyme defects (Table 2),and genetic mapping. Two clusters of genes areindicated, trpA and B mapping between hisCand ammA, and trpC and D mapping betweenrifB and thiC. This situation contrasts with thatprevailing in E. coli (116), Salmonellatyphimurium (8), and B. subtilis (13), wherethe trp genes are arranged in a single contiguouscluster, and perhaps resembles that in Pseudo-mona putida, where trp structural genes mapin at least three regions (40). No trp mutants ofS. coelicolor have been isolated that respondto anthranilic acid, although Watkins (C. A.Watkins, M.Sc. thesis, Virginia PolytechnicInstitute and State University, 1972) observedthat the tryptophan requirement of aromaticamino acid auxotrophs was satisfied by thiscompound.

Aromatic Amino Acid SynthesisThe pathway of aromatic amino acid biosyn-

thesis in cell-free extracts of a strain of S.coelicolor different from A3(2) resembled thatin other bacteria (6), although there is evidencethat in several streptomycetes, the first enzymeis unusual in being feedback inhibited only bytryptophan (38, 64). Watkins (C. A. Watkins,

M. Sc. thesis, Virginia Polytechnic Instituteand State University, 1972) studied five mu-tants (aro) having multiple aromatic aminoacid requirements. One aro mutation was lo-cated between proA and hisC, defining the aroBgene. The aroB mutant required tryptophanplus tyrosine plus phenylalanine plus p-amino-benzoic acid. The other four aro mutations,which had less extensive requirements, werelocated between argA and hisE and formed twocomplementation groups defining the aroA andC genes. None of the mutants responded toshikimic acid.

Histidine SynthesisThe major recent advance reported in this

research has been the detection (12) of a cis-dominant deletion mutation of part of the majorhis gene cluster, which causes constitutive syn-thesis of at least two enzymes specified by genesof the cluster, but does not affect expression ofthe hisD gene, which is located far from thecluster (81).

Dissimilatory PathwaysDissimilation in S. coelicolor has been almost

completely ignored. Recently, however, Saba-ter, Sebastian, and Asencio (82, 83) have charac-terized constitutive glucokinase and induciblemannokinase and fructokinase activities in anindependently isolated strain of S.violaceoruber. It would be of great interest torepeat these results in S. coelicolor A3(2), inwhich a genetical approach could also be em-ployed, particularly now that agar nonutilizing(aga) mutants are available, from which theisolation of strains defective in the use ofvarious carbon sources has been simple (K. F.Chater and F. Flury, unpublished data).

Carbon Dioxide MutantsThe occurrence of mutants of various micro-

organisms auxotrophic for carbon dioxide hasbeen reviewed by Charles and Broadbent (15)and Charles and Roberts (16). In most cases themutants respond to particular alternative nutri-ents, sometimes already explicable in terms ofknown biosynthetic pathways, and most of thecarbon dioxide mutants of S. coelicolor isolatedby Vivian and Charles (101) were of this type.However, one mutant, defining a locus cdxA,appeared to have a carbon dioxide requirementfor which no alternative nutrient was found. Inview of the possible role of carbon dioxide as anenvironmental factor influencing colonial devel-opment in microorganisms (5, 70), a deeperstudy of the response of mutant and wild-type

VOL. 37, 1973 401

HOPWOOD ET AL.

S. coelicolor to carbon dioxide appears to beworthwhile.

CONCLUDING REMARKSFrom this review it is apparent that, in spite

of progress in a number of directions, there aremajor areas where little or no research has beendone and which could profitably be explored.One is in biochemical genetics, where the timeis surely ripe for thorough investigations ofbiochemical pathways. Already there is evi-dence of different groupings in S. coelicolor ofgenes concerned with particular pathways ascompared with other bacteria. Regulation ofgene expression and enzyme activity in a dif-ferentiated soil bacterium would undoubtedlyprovide informative comparisons with entericbacteria, pseudomonads, and bacilli. Some mu-tants are already mapped and available inmany biosynthetic pathways, and a similarsituation for catabolic pathways will soon exist.A second general area where research is

needed is in central aspects of molecular biol-ogy, such as DNA replication, transcription,and translation. Such research would obviouslycomplement the genetic studies on phages,plasmids, and differentiation which are beingactively pursued in our laboratory. The inte-grated picture of the molecular biology, bio-chemistry, and genetics of S. coelicolor thatwould emerge from these studies would not onlybe valuable in terms of comparative biology,but also in the context of the large-scale produc-tion of antibiotics and other substances bystreptomycetes; the biosynthesis of antibiotics(21, 29, 74, 99) and its regulation could bestudied more incisively, and methods of trans-ferring the genetic determinants for high yieldinto other strains or even species having supe-rior cultural characteristics might well be devel-oped through the agency of plasmids or phages.It is significant that genetic studies of ac-tinomycetes other than S. coelicolor are prolif-erating (52) and that much more interest inrecombination studies is currently being shownby industrial laboratories, a sure sign that somesuccessful applications of recombination instrain improvement are anticipated.

ACKNOWLEDGMENTSIt is a pleasure to thank Helen M. Wright for her

valuable collaboration and assistance with much ofthe work described in this review. We should also liketo thank Judith M. Humphries and Helen Tovell forskilled assistance with the studies of differentiationand fertility, respectively. We are grateful to Anne C.McVittie, Paulinus P. Engel, and Thomas Schupp fordetails of some of their unpublished experiments.

LITERATURE CITED1. Adelberg, E. A., and S. N. Burns. 1960. Genetic

variation in the sex factor of Escherichia coli.J. Bacteriol. 79:321-330.

2. Alaievic, M. 1973. Genetics of tetracycline-pro-ducing streptomycetes, p. 59-70. In Z. Vafiek,Z. HogtAlek, and J. Cudlin (ed.). Genetics ofindustrial microorganisms. Academia,Prague.

3. Alikhanian, S. I., T. S. Ilyina, and N. D.Lomovskaya. 1960. Transduction in ac-tinomycetes. Nature (London) 188:245-246.

4. Baron, L. S., P. Gemski, E. M. Johnson, and J.A. Wohlhieter. 1968. Intergeneric bacterialmating. Bacteriol. Rev. 32:362-369.

5. Bartnicki-Garcia, S., and W. J. Nickerson. 1962.Induction of yeast-like development in Mucorby carbon dioxide. J. Bacteriol. 84:829-840.

6. Berlyn, M. B., and N. H. Giles. 1969. Organiza-tion of enzymes in the polyaromatic syntheticpathway: separability in bacteria. J. Bacteriol.99:222-230.

7. Blume, A. J., and E. Balbinder. 1966. Thetryptophan operon of Salmonella typhimu-rium. Fine structure analysis by deletion map-ping and abortive transduction. Genetics53:577-592.

8. Bradley, D. E. 1967. Ultrastructure of bacterio-phages and bacteriocins. Bacteriol. Rev.31:230-314.

9. Bradley, D. E. 1972. Shortening ofPseudomonasaeruginosa pili after RNA-phage adsorption.J. Gen. Microbiol. 72:303-319.

10. Burgess, R. R. 1969. A new method for the largescale purification of Escherichia coli deox-yribonucleic acid-dependent ribonucleic acidpolymerase. J. Biol. Chem. 244:6160-6167.

11. Campbell, A. 1962. Episomes. Advan. Genet.11:101-145.

12. Carere, A., S. Russi, M. Bignami, and G. Ser-monti. 1973. An operon for histidine biosyn-thesis in Streptomyces coelicolor I. Geneticevidence. Mol. Gen. Genet. 123:219-224.

13. Carlton, B. C., and D. D. Whitt. 1969. Theisolation and genetic characterization of mu-tants of the tryptophan system of Bacillussubtilis. Genetics 62:445-460.

14. Cassani, G., R. R. Burgess, H. M. Goodman, andL. Gold. 1971. Inhibition of RNA polymeraseby streptolydigin. Nature N. Biol.230:197-200.

15. Charles, H. P., and J. A. Broadbent. 1964.Carbon dioxide mutants in Neurospora. Na-ture (London) 201:1004-1006.

16. Charles, H. P., and G. A. Roberts. 1968. Carbondioxide as a growth factor for mutants ofEscherichia coli. J. Gen Microbiol.51:211-224.

17. Chater, K. F. 1972. A morphological and geneticmapping study of white colony mutants ofStreptomyces coelicolor. J. Gen. Microbiol.72:9-28.

18. Chater, K. F., and D. A. Hopwood. 1973. Dif-ferentiation in Actinomycetes, p. 143-160. In

402 BACTERIOL. REV.


J. M. Ashworth and J. E. Smith (ed.), Mi-crobial differentiation: 23rd Symposium of theSociety for General Microbiology. CambridgeUniv. Press, Cambridge, England.

19. Chatterjee, A. K., and M. P. Starr. 1972. Trans-fer among Erwinia spp. and other enterobac-teria of antibiotic resistance carried on Rfactors. J. Bacteriol. 112:574-584.

20. Coats, J. H., and J. Roeser. 1971. Geneticrecombination in Streptomyces bikiniensisvar. zorbonensis. J. Bacteriol. 105:880-885.

21. Corcoran, J. W. 1973. Genetic considerationsand erythromycin production by Strep-tomyces ethrythreus, p. 339-351. In Z. Vanik,Z. Hogtalek, and J. Cudlin (ed.), Genetics ofindustrial microorganisms. Academia,Prague.

22. Crawford, E. M., and R. F. Gesteland. 1964. Theadsorption of bacteriophage R-17. Virology22:165-167.

23. Cross, T., F. L. Davies, and P. D. Walker. 1971.Thermoactinomyces vulgaris. I. Fine structureof the developing endospores, p. 175-180. In A.N. Barker, G. W. Gould, and J. Wolf (ed.),Spore research-1971. Academic Press Inc.,London.

24. Cross, T., P. D. Walker, and G. W. Gould. 1968.Thermophilic actinomycetes producing resist-ant endospores. Nature (London)220:352-354.

25. Culotta, S., and M. Puglia. 1972. La natura"episomica" del fattore per la fertility inStreptomyces coelicolor. Ann. Ist. Super.SanitA 8:81-87.

26. Curtiss, R. 1969. Bacterial conjugation. Annu.Rev. Microbiol. 23:69-136.

27. Datta, N., R. W. Hedges, E. J. Shaw, R. B.Sykes, and M. H. Richmond. 1971. Propertiesof an R factor from Pseudomonas aeruginosa.J. Bacteriol. 108:1244-1249.

28. Delia, V., D. A. Hopwood, and E. J. Friend.1970. Mutagenesis by N-methyl-N'-nitro-N-nitrosoguanidine (NTG) in Streptomyces co-elicolor. Mutat. Res. 9:167-182.

29. Demain, A. L., and E. Inamine. 1967. Biochem-istry and regulation of streptomycin and man-nosidostreptomycinase (a-D-mannosidase)formation. Bacteriol. Rev. 34:1-19.

30. Dixon, R. A., and J. R. Postgate. 1972. Genetictransfer of nitrogen fixation from Klebsiellapneumoniae to Escherichia coli. Nature (Lon-don) 237:102-103.

31. Doi, R. H., L. R. Brown, D. Rogers, and Y. Hsu.1970. B. subtilis mutant altered in sporemorphology and in RNA polymerase activity.Proc. Nat. Acad. Sci. U.S.A. 66:404-410.

32. Dorokhova, I. A., N. S. Agre, L. V. Kalakoutskii,and N. A. Krasilnikov. 1968. Fine structure ofspores in a thermophilic actinomycete, Mi-cromonospora vulgaris. J. Gen. Microbiol.14:295-303.

33. Dowding, J. E. 1973. Characterisation of a bac-teriophage virulent for Streptomyces co-elicolor A3(2). J. Gen. Microbiol. 76:163-176.

34. Engel, P. P. 1973. Genetic control of tryptophan

biosynthesis in Streptomyces coelicolor, p.125-147. In Z. Vanek, Z. Hostalek, and J.Cudlin (ed.), Genetic of industrial microorga-nisms. Academia, Prague.

35. Erikson, D. 1955. Loss of aerial mycelium andother changes in streptomycete developmentdue to physical variations of cultural condi-tions. J. Gen. Microbiol. 13:136-148.

36. Evanchik, Z., K. A. Stacey, and W. Hayes. 1969.Ultraviolet induction of chromosome transferby autonomous sex factors in Escherichia coli.J. Gen. Microbiol. 56:1-14.

37. Friend, E. J., and D. A. Hopwood. 1971. Thelinkage map of Streptomyces rimosus. J. Gen.Microbiol. 68:187-197.

38. Gorisch, H., and F. Lingens. 1971. 3-Deoxy-D-arabino-heptulosonate-7-phosphate synthaseof Streptomyces aureofaciens Tu 24. II. Re-pression and inhibition by tryptophan andtryptophan analogues. Biochim. Biophys.Acta 242:630-636.

39. Guerola, N., J. L. Ingraham, and E. Cerda-Olmedo. 1971. Induction of closely linkedmultiple mutations by nitrosoguanidine. Na-ture N. Biol. 230:122-125.

40. Gunsalus, I. C., C. F. Gunsalus, A. M. Chakrab-arty, S. Sikes, and I. P. Crawford. 1968. Finestructure mapping of the tryptophan genes inPseudomonas putida. Genetics 60:419-435.

41. Harold, R. J., and D. A. Hopwood. 1970. Ultravi-olet-sensitive mutants of Streptomyces co-elicolor. I. Phenotypic characterisation. Mu-tat. Res. 10:427-438.

42. Harold, R. J., and D. A. Hopwood. 1970. Ultravi-olet-sensitive mutants of Streptomyces co-elicolor. II. Genetics. Mutat. Res. 10:439-448.

43. Harold, R. J., and D. A. Hopwood. 1972. A rapidmethod for complementation testing of ul-traviolet-sensitive (uvs) mutants of Strep-tomyces coelicolor. Mutat. Res. 16:27-34.

44. Hirota, Y., and P. H. A. Sneath. 1961. F' and Fmediated transduction in Escherichia coliK-12. Jap. J. Genet. 36:307-318.

45. Hopwood, D. A. 1957. Genetic recombination inStreptomyces coelicolor. J. Gen. Microbiol.16:ii-iii.

46. Hopwood, D. A. 1959. Linkage and the mecha-nism of recombination in Streptomyces co-elicolor. Ann. N. Y. Acad. Sci. 81:887-898.

47. Hopwood, D. A. 1960. Phase-contrast observa-tions on Streptomyces coelicolor. J. Gen. Mi-crobiol. 22:295-302.

48. Hopwood, D. A. 1961. Genetic recombination inStreptomyces. Sci. Rep. Ist. Super. SanitaRome 1:463-466.

49. Hopwood, D. A. 1965. New data on the linkagemap of Streptomyces coelicolor. Genet. Res.6:248-262.

50. Hopwood, D. A. 1966. Non-random location oftemperature-sensitive mutants on the linkagemap of Streptomyces coelicolor. Genetics54:1169-1176.

51. Hopwood, D. A. 1967. Genetic analysis andgenome structure in Streptomyces coelicolor.Bacteriol. Rev. 31:373-403.

VOL. 37, 1973 403

HOPWOOD ET AL.

52. Hopwood, D. A. 1972. Genetirs of the ac-tinomycetales, p. 131-153. In G. Sykes and F.A. Skinner (ed.), Actinomycetales: character-istics and practical importance. AcademicPress Inc., London.

53. Hopwood, D. A., and H. M. Ferguson. 1969. Arapid method for lyophilizing streptomycescultures. J. Appl. Bacteriol. 32:434-436.

54. Hopwood, D. A., and H. M. Ferguson. 1970.Genetic recombination in a thermophilic ac-tinomycete, Thermoactinomyces vulgaris. J.Gen. Microbiol. 63:133-136.

55. Hopwood, D. A., and A. M. Glauert. 1961.Electron microscope observations on the sur-face structures of Streptomyces violaceoru-ber. J. Gen. Microbiol. 25:325-330.

56. Hopwood, D. A., R. J. Harold, A. Vivian, and H.M. Ferguson. 1968. Non-selective geneticanalysis in Streptomyces coelicolor. Heredity23:628.

57. Hopwood, D. A., R. J. Harold, A. Vivian, and H.M. Ferguson. 1969. A new kind of fertilityvariant in Streptomyces coelicolor. Genetics62:461-477.

58. Hopwood, D. A., and G. Sermonti. 1962. Thegenetics of Streptomyces coelicolor. Advan.Genet. 11:273-342.

59. Hopwood, D. A., G. Sermonti, and I. Spada-Ser-monti. 1963. Heterozygous clones in Strep-tomyces coelicolor. J. Gen. Microbiol.30:249-260.

60. Hopwood, D. A., H. Wildermuth, and H. M.Palmer. 1970. Mutants of Streptomyces co-elicolor defective in sporulation. J. Gen. Mi-crobiol. 61:397-408.

61. Hopwood, D. A., and H. M. Wright. 1972.Transformation in Thermoactinomyces vulga-ris. J. Gen. Microbiol. 71:383-398.

62. Howard-Flanders, P., R. P. Boyce, and J. B.Sedeford. 1966. Three loci of Escherichia coliK-12 that control the excision of pyrimidinedimers and certain other mutagen productsfrom DNA. Genetics 53:1119-1136.

63. Jacob, F., and E. A. Adelberg. 1959. Transfert decaracteres genetiques par incorporation aufacteur sexuel d'Escherichia coli. C. R. Acad.Sci. 249:189-191.

64. Jensen, R. A., D. S. Nasser, and E. W. Nester.1967. Comparative control of a branch-pointenzyme in microorganisms. J. Bacteriol.94:1582-1593.

65. Kelner, A. 1949. Effect of visible light on therecovery of Streptomyces griseus conidia fromultraviolet irradiation injury. Proc. Nat. Acad.Sci. U.S.A. 35:73-79.

66. Lawton, W. D., B. C. Morris, and T. W. Bur-rows. 1968. Gene transfer in strains of Pas-teurella pseudotuberculosis. J. Gen. Micro-biol. 52:25-34.

67. Lederberg, J., L. L. Cavalli, and E. M. Leder-berg. 1952. Sex compatibility in Escherichiacoli. Genetics 37:720-730.

68. Lomovskaya, N. D., L. K. Emeljanova, and S. I.

Alikhanian. 1971. The genetic location of pro-phage on the chromosome of Streptomycescoelicolor. Genetics 68:341-347.

69. Lomovskaya, N. D., N. M. Mkrtumian, N. L.Gostimskaya, and V. N. Danilenko. 1972.Characterization of temperate actinophage4C31 isolated from Streptomyces coelicolorA3(2). J. Virol. 9:258-262.

70. Loomis, W. F. 1961. Feedback factors affectingsexual differentiation in Hydra littoralis, p.337. In H. M. Lenhoff and W. F. Loomis (ed.),The biology of hydra and some other coelen-terates. Univ. of Miami Press, Coral Gables,Florida.

71. Losick, R., and A. L. Sonenshein. 1969. Changein the template specificity ofRNA polymeraseduring sporulation of Bacillus subtilis. Nature(London) 224:35-37.

72. Matselyukh, B. P., M. E. Podgorskaya, A. S.Stenko, and V. Y. Lavrinchuk. 1972.Chromosome mapping in Actinomyces oliva-ceus VKX by means of heteroclone analysis.Genetika 8:123-130.

73. Matselyukh, B. P., M. E. Podgorskaya, A. S.Stenko, and V. Y. Lavrinchuk. 1973. Ac-tinomyces olivaceus VKX-a new object foractinomycete genetics. Mikrobiolo. Zh. (Kiev)35:26-36.

74. McCormick, J. R. D. 1969. Point-blocked mu-tants and the biogenesis of tetracyclines, p.163-176. In Symposium on Genetics andBreeding of Streptomyces. Yugoslav Academyof Sciences and Arts, Zagreb.

75. McVittie, A. C., H. Wildermuth, and D. A.Hopwood. 1972. Fine structure and surfacetopography of endospores of Thermoac-tinomyces vulgaris. J. Gen. Microbiol.71:367-381.

76. Ogata, W. N. 1962. Preservation of Neurosporastock cultures with anhydrous silica gel. Neu-rospora Newslett. 1:13.

77. Perkins, D. D. 1962. Preservation of Neurosporastock cultures with anhydrous silica gel. Can.J. Microbiol. 8:591-594.

78. Piperno, R., A. Carere, and G. Sermonti. 1966.Polarized recombination in Streptomyces co-elicolor. Ann. Ist. Super. Sanita 2:393-407.

79. Puglia, A. M., I. Spada-Sermonti, S. Basile, F.Misuraca, and G. Sermonti. 1973. Infectioustransfer of a fertility factor in Streptomycescoelicolor. Mutat. Res. 21:107-113.

80. Randazzo, R., G. Sermonti, A. Carere, and M.Bignomi. 1973. Comutation in Streptomyces.J. Bacteriol. 113:500-501.

81. Russi, S., A. Carere, A. Siracusano, and A. Bal-lio. 1973. An operon for histidine biosynthesisin Streptomyces coelicolor II. Biochemical evi-dence. Mol. Gen. Genet. 123:225-232.

82. Sabater, B., J. Sebastian, and C. Asencio. 1972.Identification and properties of an induciblemannokinase from Streptomyces violaceo-ruber. Biochim. Biophys. Acta 284:406-413.

83. Sabater, B., J. Sebastian, and C. Asencio. 1972.

404 BACTERIOL. REV.


Identification and properties of an inducibleand highly specific fructokinase from Strep-tomyces violaceoruber. Biochim. Biophys.Acta 284:414-420.

84. Schaeffer, P. 1969. Sporulation and the produc-tion of antibiotics, exoenzymes, and exotox-ins. Bacteriol. Rev. 33:48-71.

85. Schmidt, J. M. 1966. Observations on the ad-sorption of Caulobacter bacteriophages con-taining ribonucleic acid. J. Gen. Microbiol.45:347-353.

86. Scott, T., and B. R. Alderson. 1971. The photo-sensitizing effect of 8-methoxypsoralen on theinactivation and mutation of Aspergillus co-

nidia by near ultraviolet light. Mutat. Res. 12:29-34.

87. Sermonti, G. 1969. Genetics of antibiotic-pro-ducing microorganisms. Wiley-Interscience,London.

88. Sermonti, G., M. Bandiera, and I. Spada-Ser-monti. 1966. New approach to the genetics ofStreptomyces coelicolor. J. Bacteriol. 91:384-392.

89. Sermonti, G., and S. Casciano. 1963. Sexualpolarity in Streptomyces coelicolor. J. Gen.Microbiol. 33:293-301.

90. Sermonti, G., and D. A. Hopwood. 1964. Geneticrecombination in Streptomyces, p. 223-251. InI. C. Gunsalus and R. Y. Stanier (ed.), Thebacteria, vol. 5. Academic Press Inc., NewYork.

91. Sermonti, G., A. Mancinelli, and I. Spada-Ser-monti. 1960. Heterogeneous clones ("hetero-clones") in Streptomyces coelicolor A3(2).Genetics 45:669-672.

92. Sermonti, G., A. M. Puglia, and G. Ficarra.1971. The time course of recombinant produc-tion in Streptomyces coelicolor. Genet. Res.18: 133-145.

93. Slepecky, R. A. 1972. Ecology of bacterial spore-formers, p. 297-313. In H. 0. Halvorson, R.Hanson, and L. L. Campbell (ed.), Spores V.American Society for Microbiology, Washing-ton, D.C.

94. Sonenshein, A. L., and R. Losick. 1970. RNApolymerase mutants blocked in sporulation.Nature (London) 227:906-909.

95. Spada-Sermonti, I., and G. Sermonti. 1973.Fertilization in Streptomyces coelicolor, p.47-57. In Z. Vanek, Z. Hostalek, and J. Cudlin(ed.), Genetics of industrial microorganisms.Academia, Prague.

96. Stanier, R. Y. 1942. Agar-decomposing strains ofthe Actinomyces coelicolor species group. J.Bacteriol. 44:555-570.

97. Studier, F. W. 1969. The genetics and physiologyof bacteriophage T7. Virology 39:562-574.

98. Townsend, M. E., H. M. Wright, and D. A.Hopwood. 1971. Efficient mutagenesis by nearultraviolet light in the presence of 8-methox-ypsoralen in Streptomyces. J. Appl. Bacteriol.34:799-801.

99. Vanek, Z., and Z. Hostalek. 1972. Some aspects

of the genetic control of the biosynthesis ofchlortetracycline. Postepy Hig. Med. Dosw.25:445-467.

100. Vivian, A. 1971. Genetic control of fertility inStreptomyces coelicolor A3(2): plasmid in-volvement in the interconversion of UF and IFstrains. J. Gen. Microbiol. 69:353-364.

101. Vivian, A., and H. P. Charles. 1970. The occur-rence and genetics of some CO2 mutants inStreptomyces coelicolor. J. Gen. Microbiol.61:263-271.

102. Vivian, A., and D. A. Hopwood. 1970. Geneticcontrol of fertility in Streptomyces coelicolorA3(2): the IF fertility type. J. Gen. Microbiol.64: 101-117.

103. Vivian, A., and D. A. Hopwood. 1973. Geneticcontrol of fertility in Streptomyces coelicolorA3(2): new kinds of donor strains. J. Gen.Microbiol. 76:147-162.

104. Wehli, W., and M. Staehelin. 1971. Actions ofthe rifamycins. Bacteriol. Rev. 35:290-309.

105. Welsch, M. 1956. Evidence for the occurrence oftrue lysogeny among actinomycetes. Virology2:703-704.

106. Welsch, M. 1959. Lysogenicity in strep-tomycetes. Ann. N. Y. Acad. Sci. 81:974-993.

107. Welsch, M. 1969. Biology of actinophages, p.43-62. In Symposium on genetics and breed-ing of Streptomyces, Dubrovnik, 1969. Yugo-slav Academy of Sciences and Arts, Zagreb.

108. Wieringa, K. T., and G. L. W. Wiebols. 1936. Deaarcappelschurft en de heterolyse des schurf-tparasiet. Tijdschr. Plantenziekten 42:235-240.

109. Wildermuth, H. 1970. Development and organi-zation of the aerial mycelium in Streptomycescoelicolor. J. Gen. Microbiol. 60:43-50.

110. Wildermuth, H. 1972. The surface structure ofspores and aerial hyphae in Streptomycesviridochromogenes. Arch. Mikrobiol. 81:309-320.

111. Wildermuth, H. 1972. Morphological surfacecharacteristics of Streptomyces glaucescensand S. acrimycini, two streptomycetes with"hairy" spores. Arch. Mikrobiol. 81:321-332.

112. Wildermuth, H., and D. A. Hopwood. 1970. Sep-tation during sporulation in Streptomycescoelicolor. J. Gen. Microbiol. 60:51-59.

113. Wildermuth, H., E. Wehrli, and R. W. Horne.1971. The surface structure of spores andaerial mycelium in Streptomyces coelicolor. J.Ultrastruct. Res. 35:168-180.

114. Willetts, N. 1972. The genetics of transmissibleplasmids. Annu. Rev. Genet. 6:257-268.

115. Williams, S. T., R. M. Bradshaw, J. W. Caster-ton, and A. Forge. 1972. Fine structure of thespore sheath of some Streptomyces species. J.Gen. Microbiol. 72:249-258.

116. Yanofsky, C., and E. S. Lennox. 1959. Transduc-tion and recombination study of linkage rela-tionships among the genes controlling trypto-phan synthesis in Escherichia coli. Virology8:425-447.

VOL. 37, 1973 405

Advances in Streptomyces coelicolor Genetics - NCBI

Documents