BACTrzwOLOGICAL Rzvmws, Sept. 1977, p. 595-635 Vol. 41, No. 3 Copyright 0 1977 American Society for Microbiology Printed in U.S.A. Genetics of Antibiotic Production D. A. HOPWOOD* AND M. J. MERRICKt John Innes Institute, Colney Lane, Norwich NR4 7UH, England INTRODUCTION .............................................................. 596 INCIDENCE OF ANTIBIOTIC PRODUCTION ............ .. ................... 596 General . ..................................................................... 596 Fungi........................................................................ 597 Actinomycetes ............................................................... 597 Eubacteria .................................................................. 598 GENETIC SYSTEMS AVAILABLE IN ANTIBIOTIC-PRODUCING MICROOR- GANISMS ............................................................... 598 General . .................................................................... 598 Fungi . ....................................................................... 599 Actinomycetes ............................................................... 601 Transformation in Thermoactinomyces ......... ............................. 601 Conjugation in other genera ................................................ 601 Conjugation and its consequences in Streptomyces ....... .................... 602 Linkage analysis in Streptomyces ........... ................................ 602 Heteroclones in Streptomyces ............ .................................. 604 Heterokaryons in Streptomyces ........... .................................. 604 Plasmids in Streptomyces ............... ................................... 605 Interspecific recombination in Streptomyces ........ ......................... 605 Recombination in N. mediterranei .......... ................................ 605 Recombination in Micromonospora .......... ................................ 605 Recombination by protoplast fusion ......................................... 605 Eubacteria . ................................................................. 605 SOME REMARKS ON THE PHYSIOLOGY OF ANTIBIOTIC PRODUCTION ... 606 Primary and Secondary Metabolites .......... ................................ 606 Regulation of Secondary Metabolic Pathways .................................. 606 Pleiotropic Effects of Mutations on Antibiotic Production ........ .............. 607 The Adaptive Significance of Antibiotics ................ ...................... 607 "Low Specificity" of Secondary Metabolic Enzymes ........ ................... 607 Chemical Classes of Antibiotics ............................................... 608 Why "Biogenesis"? . ................................................. 608 MUTATIONAL STUDIES OF ANTIBIOTIC SYNTHESIS ......... .............. 608 Fungi....................................................................... 608 P. chrysogenum ..................... ............................ 608 A. nidulans ................................................ 610 C. acremonium ................... .............................. 611 P. griseofulvum ..................... ............................ 613 P. patulum ............. .................................... 614 0. mucida ............. .................................... 614 Actinomycetes .................................................. 614 Tetracyclines.............................................................. 614 Macrolides ............. .................................... 615 Aminocyclitols (aminoglycosides) . .......................................... 615 Novobiocin ............. .................................... 616 Rifamycins .................................................. 616 Actinorhodin ................ ................................. 616 Zorbamycin, etc. .................................................. 617 Plasmid involvement in antibiotic synthesis ........... ...................... 617 Eubacteria ........... ...................................... 619 Peptide antibiotics of bacilli ................................................ 619 Butirosin ................................................... 619 Nisin .................................................. 619 Prodigiosin .................................................. 620 QUANTITATIVE GENETICS OF ANTIBIOTIC PRODUCTION ...... ........... 620 General Principles of Polygenic Control of Quantitative Characters ..... ....... 620 t Present address: ARC Unit of Nitrogen Fixation, The Chemical Laboratory, The University of Sussex, Brighton, BN1 9QJ, England. 595 on June 20, 2020 by guest http://mmbr.asm.org/ Downloaded from
Transcript
BACTrzwOLOGICAL Rzvmws, Sept. 1977, p. 595-635 Vol. 41, No. 3Copyright 0 1977 American Society for Microbiology Printed in U.S.A.
Genetics of Antibiotic ProductionD. A. HOPWOOD* AND M. J. MERRICKt
John Innes Institute, Colney Lane, Norwich NR4 7UH, England
INTRODUCTION .............................................................. 596INCIDENCE OF ANTIBIOTIC PRODUCTION ............ .. ................... 596General...................................................................... 596Fungi........................................................................ 597Actinomycetes ............................................................... 597Eubacteria .................................................................. 598
GENETIC SYSTEMS AVAILABLE IN ANTIBIOTIC-PRODUCING MICROOR-GANISMS ............................................................... 598
General..................................................................... 598Fungi........................................................................ 599Actinomycetes ............................................................... 601Transformation in Thermoactinomyces ......... ............................. 601Conjugation in other genera ................................................ 601Conjugation and its consequences in Streptomyces ....... .................... 602Linkage analysis in Streptomyces ........... ................................ 602Heteroclones in Streptomyces ............ .................................. 604Heterokaryons in Streptomyces ........... .................................. 604Plasmids in Streptomyces ............... ................................... 605Interspecific recombination in Streptomyces ........ ......................... 605Recombination in N. mediterranei .......... ................................ 605Recombination in Micromonospora .......... ................................ 605Recombination by protoplast fusion......................................... 605
Eubacteria.................................................................. 605SOME REMARKS ON THE PHYSIOLOGY OF ANTIBIOTIC PRODUCTION ... 606Primary and Secondary Metabolites .......... ................................ 606Regulation of Secondary Metabolic Pathways .................................. 606Pleiotropic Effects of Mutations on Antibiotic Production ........ .............. 607The Adaptive Significance of Antibiotics ................ ...................... 607"Low Specificity" of Secondary Metabolic Enzymes ........ ................... 607Chemical Classes of Antibiotics ............................................... 608Why "Biogenesis"?.................................................. 608
QUANTITATIVE GENETICS OF ANTIBIOTIC PRODUCTION ...... ........... 620General Principles of Polygenic Control of Quantitative Characters ..... ....... 620
t Present address: ARC Unit of Nitrogen Fixation, The Chemical Laboratory, The University of Sussex, Brighton,BN1 9QJ, England.
Application of Biometrical Methods to Haploid and Nonmeiotic Systems ........ 620Nonmeiotic systems ........................................................ 621
Biometrical Genetics of Penicillin Production in A. nidulans ................... 621Quantitative Mutations Affecting Penicillin Yield ........ ..................... 623P. chrysogenum ......................................................... 624A. nidulans ......................................................... 624
CONCLUSION.......................................................... 624The Application of Genetic Knowledge ............. ........................... 624
LITERATURE CITED......................................................... 625
INTRODUCTIONThe explosion of microbial genetics over the
last 30 years coincided with an equally spectac-ular increase in the production of antibiotics.Yet knowledge of the genetics of antibiotic pro-duction, and even of the biosynthetic pathwaysof antibiotics, is still disproportionately small.In part this arises from the fact that, aside frommutagenesis, which has been employed on ahuge scale, genetic approaches to the improve-ment of strains for industrial antibiotic produc-tion have been largely ignored (236). Importantfundamental genetic knowledge has, therefore,not been gained from applied microbiology as ithas from plant breeding (204). Conversely, aca-demic geneticists have not chosen antibioticbiosyntheses as models for studies of gene-en-zyme relationships or of the regulation of bio-chemical pathways. Perhaps there was a feel-ing that no principles would emerge that couldnot be better revealed by the intensive study ofsynthesis or catabolism pathways of moleculessuch as amino acids or sugars in species ofEscherichia, Salmonella, Neurospora, Asper-gillus, or Saccharomyces with their highly de-veloped experimental genetics. This concentra-tion of effort was highly successful. However, amore diversified effort is now appropriate. Forexample, enteric bacteria inhabiting the hu-man gut are clearly not completely representa-tive of bacteria as a whole in their metabolicregulation, and much can be learned from otherkinds of bacteria, such as actinomycetes or ba-cilli. Moreover, regulatory systems that ensurea constant supply of amino acids for proteinsynthesis during steady growth may differ fromthose controlling the synthesis of antibiotics,which typically appear at defined stages in thelife cycle.There are problems in studying the genetics
of antibiotic production, but they are not insur-mountable. Very few antibiotic-producing mi-croorganisms have well-developed genetic sys-tems, but most belong to genera with a geneti-cally well-known member; thus, genetic analy-sis could probably be extended to them. More-over, three 'academic" organisms with well-de-veloped genetic systems and representing thethree major groups of antibiotic producers, As-
pergillus nidulans, Streptomyces coelicolor,and Bacillus subtilis, are known to produceantibiotics; these antibiotics are, therefore,open to genetic investigation. The enzymes ofantibiotic synthesis may occur in low concen-trations, be part of multienzyme complexes, orbe difficult to isolate for other reasons. How-ever, this is often true of biosynthetic enzymesfor primary metabolites; such enzymes haveoften been purified or studied in other waysonce their role was identified by the isolationof a suitable mutant. Finally, the synthesisof most antibiotics is not easily studied in thesteady-state conditions of an exponentiallygrowing culture of a unicellular microorga-nism, since they are associated with differen-tiation, and this is undoubtedly an experimen-tal obstacle to the study of the regulation ofantibiotic synthesis.A start has been made in the study of the
control of antibiotic production in several di-verse experimental systems, filamentous fungi,actinomycetes, and sporulating eubacteria. So-phisticated chemical, biochemical, and geneticapproaches have been used, but the most com-plete biochemical studies have not coincidedwith the most extensive genetics, to the detri-ment ofboth disciplines. We hope that this arti-cle will lead to an increased awareness of thebiologically and industrially interesting prob-lems ripe for solution in this area by the com-bined use of biochemical and genetic tech-niques.
INCIDENCE OF ANTIBIOTICPRODUCTION
General
Many groups of organisms produce sub-stances toxic to microorganisms: for example,invertebrates, algae, and higher plants (27).However, to be reasonably coherent, we shalllimit discussion to the three groups whosemembers produce antibiotics with some com-mercial or significant experimental applica-tion: the eucaryotic fungi and the procaryoticactinomycetes and eubacteria (27, 231). More-over, although some definitions of the wordantibiotic would include bacteriocins (38, 244,
277), we shall exclude them and regard an anti-biotic as a compound produced by stepwise bio-synthesis, rather than by the template proc-esses of protein synthesis, and inhibitory toorganisms outside the immediate relatives ofthe producing microbe.
FungiAlthough the first important antibiotic, peni-
cillin, came from a filamentous fungus, fewerthan a quarter of the more than 3,000 antibiot-ics since described are produced by fungi. TheAspergillaceae, including Penicilium and As-pergillus, produe the great majority of impor-tant flmgal antibiotics, although recently theMoniliales have yielded most of the new com-pounds (27). Only 10 natural fungal antibioticshave been made commercially: penicillins G. V,and 0, cephalosporin C (CPC), griseofulvin,
fumagillin, fusidic acid, siccanin, variotin, andxanthocillin; of these only grisefulvin, fusidicacid, and the A-lactam antibiotics are signifi-cant clinically.
The development of the semisynthetic p-lac-tam antibiotics, with improved therapeuticproperties, has maintained the penicillins andcephalosporins as one of the most widely usedgroups of antibiotics. More than 20 semisyn-thetic penicillins have found a place in chemo-therapy. They are usually made by chemical
addition of a new side chain to 6-aminopenicil-lanic acid (6APA), originally isolated from fer-mentations of Penicillium chrysogenum, butnow produced by chemical or enzymatic hydroly-sis of benzylpeniicillin enicillin G). 7-Amino-cephaloporanic acid (7-ACA) is made chemi-cally from CPC; some eight semisynthetic ceph-alosporins are in clinical use, all chemical mod-
ifications of7-ACA. In contrast to the formationof the classical penicillins (G, 0, and V), theaddition ofside chain precursors to the mediumdoes not influence the antibiotics. formed byCephaloaporium acremonium: they are invari-ably CPC and penicillin N (PCN).
It is interesting that the four main fungalantibiotics are all produced by more than onegenus. Penicillin, first described inPenicilliumnotatum, is made commercially by strains ofPenicillium chrysogenum. Penicillins havebeen isolated from several species ofAspergil-l8s and Cephaloporium and at least six other
genera (159), and PCN has also been isolatedfrom streptomycetes (219). CPC, produced in-dustrially by C. acremonium, is also made byspecies ofEmericellopsis and several other gen-era (159). Several new frlactam derivativescome from Streptomyces species (219, 275) anddeacetoxycephalosporin C from Streptomyces
species and a variety offungi (122). Fusidic acidwas first isolated from Fu8idium coccineum(109), a fungus closely related to the cephalos-poria, and later from Mucor remannianus (290)and Isaria kogana (123). Griseofulvin was orig-inally found in Penicillium griseofulvum andsubsequently in about 12 otherPenicillium spe-cies including P. janczewski and P. patulum;mutant strains of P. patulum are used to pro-duce it commercially. A completely unrelatedorganism, Khuskia oryzae, also makes griseo-fulvin (108).Of the rem commercial fungal anti-
biotics, fumagillin is made by Aspergillus fu-migatus (31), siccanin by Helminthosporiumsiccans (124), variotin by Paecilomyces varioti(279), and xanthocillin by P. notatum (1).
ActinomycetesThis group is unique in the number of anti-
biotics made from it, and in the diversity ofthese antibiotics' chemical structures andmodes of action. Perlman (231) lists nearly 70that are produced commercially. There is acompilation of actinomycete antibiotics (287),which is being supplemented by monthly infor-mation in the same format in the Journal ofAntibiotics (Japan). Berdy (27) counts morethan 1,950 streptomycete compounds, includ-ing the important chloramphenicol, erythromy-cin, kanamycin, neomycin, novobiocin, strepto-mycin, and tetracyclines, out of2,080 actinomy-cete antibiotics, but this may reflect the ease ofisolation of streptomycetes. More recently, No-cardia species (rifamycin, ristocetin) and Mi-cromonospora species (gentamicin, sisomicin)have emerged as producers of important anti-biotics, and attention is turning to genera suchas Actinoplanes (23, 60, 61), Streptosporan-gium (161, 278), Streptoverticillium (311), andThermoactinomyces (233). With increasingknowledge of actinomycete biology (64) and therealization that many organisms remain to beisolated by suitable selective procedures, thelist of actinomycete antibiotics will doubtlesscontinue to grow.
Semisynthetic actinomycete antibiotics areproportionally less important than in the fim-gal case but are nevertheless significant in ab-solute terms: for example, derivatives of kana-mycin (e.g., amikacin), rifamycin (rifampin),lincomycin (clindamycin), and various tetracy-clines.Recent taxonomy suggests that the genus
Nocardia is heterogeneous, and genetic studies(see below) support this view. N. mediterranei,the rifamycin producer, resembles Strepto-myces species in its genetics, whereas N. ery-
thropolis and some related strains share ge-netic features with Mycobacterium smegmatis.This agrees with the idea that N. mediterraneiis close to Streptomyces (it was originallynamed S. mediterranei), whereas N. erythro-polis is closer to Mycobacterium, forming partof the so-called Mycobacterium rodochrouscomplex (37). It will, therefore, be interestingto investigate other antibiotic-producing No-cardia strains by recent taxonomic criteria, in-cluding the presence of characteristic lipidscalled mycolic acids found in mycobacteria andthe true nocardiae but not in the so-called no-cardiae represented by N. mediterranei (34,215). N. lurida, the ristocetin producer, lackssuch lipids, as does the producer of nocardin, N.autotrophica (110). However, caution is needed:the nocardin producer was originally named N.coeliaca, authentic strains of which possess thecharacteristic lipids. Taxonomic and geneticstudies must be done on strains that actuallyproduce antibiotics. Such studies have not beenmade on strains producing the other nocardiaantibiotics listed by Umezawa (287), althoughstrains of some of the relevant species (N. for-mica, N. gardneri) appear to lack mycolicacids, and none of these species is known topossess such compounds. Thus there is the in-teresting possibility that most, if not all, of thenocardiae that produce antibiotics belong to thestreptomyces-like organisms. This would haveimportant implications for genetic investiga-tions of antibiotic production.
Eubacteria
Berdy (27) credits the eubacteria with about360 antibiotics scattered through many taxo-nomic groups. However, bacilli produce nearlyhalf the total and pseudomonads nearly aquarter. The genus Bacillus is by far the mostsignificant commercially, nearly all the inter-esting eubacterial antibiotics being cyclic orlinear peptides produced by members of thisgenus (250): B. licheniformis (bacitracin), B.brevis (gramicidin S and linear gramicidin Atyrocidine, edeines), B. polymyxa (polymyxin),B. colistinus (colistin), B. subtilis (mycobacil-lin). Another Bacillus product of interest isbutirosin (B. circulans), the only aminocyclitolknown to be produced by a non-actinomycete(55). Of the many Pseudomonas antibiotics,only pyocyanine (P. aeruginosa [183]) and pyr-rolnitrin (P. aureofaciens) have some commer-cial interest. The cyclic polypeptide nisin isproduced by Streptococcus lactis. Other eubac-terial antibiotics appear to have neither com-mercial nor genetic relevance at the presenttime. However, perhaps it is worth noting two
recently described myxobacterial antibiotics(232, 289) because of current interest in thedevelopmental genetics of myxobacteria.
GENETIC SYSTEMS AVAILABLE INANTIBIOTIC-PRODUCING
MICROORGANISMSGeneral
Even without genetic analysis, biochemicalstudies of mutants help to explain the biosyn-thesis of antibiotics. However, genetics is thencontributing a small fraction of its analyticalpotential. For example, evidence that a mutantdiffers from its progenitor by a single mutationnormally depends on observing no other pheno-types besides those ofthe mutant and wild typein a suitable cross; genetic analysis, embodyingsome kind of recombination test, is needed todetermine the arrangement of the genes in-volved in antibiotic synthesis (whether they arescattered or clustered, chromosomally or plas-mid borne, etc.); a genetic complementationtest is the best criterion for classifying mutantsbefore embarking on biochemical studies ofrep-resentative examples; dominance or gene dos-age tests help to determine the roles of regula-tory genes; and studies of genetic fine structurecould illuminate structure-function relation-ships of individual proteins or multienzymecomplexes involved in antibiotic biosynthesis.A perfect genetic system would embody allthese tests and possibly others. Very few orga-nisms have been developed as ideal genetic sub-jects, and, as far as antibiotic producers areconcerned, the fungus Aspergillus nidulans,the actinomycete Streptomyces coelicolor, andthe eubacterium Bacillus subtilis, all "aca-demic" organisms known to produce antibiot-ics, come nearest to the ideal though still fall-ing short of it. However, systems of gene ex-change are widespread in the microbial groupsthat produce antibiotics, so the potential forgenetic analysis is enormous.Eucaryotes (here represented by the fungi)
and procaryotes (actinomycetes and eubacteria)differ significantly in their genetics. The deoxy-ribonucleic acid (DNA) of eucaryotes occurs inseveral separate chromosomes together withstructural and regulatory (histone) proteins;complex nuclear division mechanisms (mitosisand meiosis) ensure the exact partitioning ofgenes to daughter nuclei and progeny; and thechromosomes are retained in a discrete regionofthe cell by a nuclear membrane. The chromo-some of procaryotes, on the other hand, is asingle circular DNA molecule with few, if any,protein molecules permanently associated withit, and there is no nuclear membrane. Procar-
yotes possess plasmids (86, 213)-circular DNAmolecules considerably smaller than the chro-mosome and representing genes dispensable, atleast under certain conditions, to the organismcarrying them. Such genes determine charac-ters-sex, antibiotic production or resistance,catabolic proficiency, pathogenicity, etc. -thatare important for the evolutionary versatility ofthe population but are carried by only a propor-tion of its members. Plasmids will have manyapplications in the experimental manipulationof antibiotic-producing organisms (138). Eu-caryotes usually, though not always, have reg-ular sexual cycles involving fusion of wholenuclei and subsequent reassortment of chromo-somes from the fusion nucleus; on the otherhand, procaryotes indulge in several processes,transformation, transduction, and conjugation,with the same genetic consequences as sexualreproduction-the creation of new combina-tions of genes-but differing markedly from it.Probably because of differences in chromosomestructure, eucaryote chromosomes readily un-dergo rearrangement ofsegments to give inver-sions and interchanges, whereas those of pro-caryotes do not. The resulting conservation ofgene arrangements in procaryotes has the use-ful consequence that crosses between strainsrepresenting divergent lines of selection from acommon ancestor are not hindered by inhomol-ogy between chromosomes. Finally, althoughmany details of the storage, transmission, andexpression of genetic information (DNA repli-cation, transcription, and translation) are thesame in procaryotes and eucaryotes, there arealso differences, particularly in the signals ini-tiating and terminating these processes (175,176, 297). These differences present barriers,currently under attack in many laboratories, tothe useful transfer ofgenes between eucaryotesand procaryotes.
FungiRecombination has been described in six spe-
cies of antibiotic-producing fungi, all members
of the Plectomycetales (Table 1). Only three ofthese species, A. nidulans, Emericellopsis sal-mosynnemata, and Emericellopsis terricola,have a conventional sexual cycle. All are homo-thallic, so that many ofthe perithecia (correctlytermed cleistothecia) arise by self-fertilization.The asci in individual perithecia are generallyall selfed or all crossed; hence, a single perithe-cium can usually be diagnosed as of crossedorigin by analyzing a sample of its ascospores,most conveniently by the use of conidial colormutations in one or both parents (238). Discreteasci containing eight ascospores are presentonly in immature perithecia; at maturity, theascus wall breaks down so that orthodox tetradanalysis is not normally carried out in thesefungi, although it can be done. Analysis of ran-dom spores from a single hybrid peritheciumallows the detection of linkage betweenmarkers and of polygenic segregation and epi-stasis. Fine mapping can be done by selectiveplating of random spores.
Several types of genetic analysis -tests ofdominance and complementation, centromeremapping, and detection of linkage groups-areroutinely carried out by "parasexual" analysiseven in organisms with a regular sexual cycle.It is the only system available in the imperfectfungi P. chrysogenum, P. patulum, and C.acremonium. The life cycle of all these fungi ispredominantly haploid, with occasional hetero-karyon formation at frequencies varying signif-icantly between species. From these heterokar-yons occasional diploid nuclei may be selectedas first demonstrated by Roper (248) inA. nidu-lans.
Roper's technique, applicable to any moldwith uninucleate conidia, has been successfullyapplied to several fungi. A prototrophic hetero-karyon is formed on minimal medium from twonutritionally different auxotrophs. All haploidconidia formed by the heterokaryon are of oneor the other auxotrophic type and fail to growon minimal medium. However, a diploid conid-ium resulting from fusion of unlike nuclei is
TABLE 1. Antibiotic-producing fungi in which genetic recombination has been demonstrated
Species Antibiotic Type of recombination ReferencesAspergillus nidulans Penicillin G Sexual, parasexual 238,127,128Cephalosporium acremonium CPC Parasexual 223
prototrophic; in A. nidulans, plating of conidiafrom a heterokaryon yields diploids at a fre-quency of 10-6 to 10-7.
Diploid strains of A. nidulans can be distin-guished from haploids by their larger conidialdiameter (238), but this is not always a reliableindicator of ploidy in P. chrysogenum (191, 239)or P. patulum (49). The introduction of majormutations into strains can have considerableeffects on conidial size (84), and, therefore, co-nidial diameter and DNA content should becorrelated for accurate ploidy determination,and diploid strains should always be comparedwith their immediate parents.
Genetic analysis by the parasexual cycle de-pends on the isolation from heterozygous dip-loids of two kinds of vegetative segregants. (i)Diploids homozygous at one or more loci thatwere originally heterozygous arise by mitoticcrossing-over between chromatids of homolo-gous chromosomes, followed by appropriatesegregation of chromatids at the next nucleardivision. Homozygosity occurs for all loci distalto the point of crossing-over on that chromo-some arm, and so analysis of such diploids al-lows ordering of markers on the same chromo-some arm, estimation of relative mitotic cross-over frequencies between them, and mapping ofthe centromere ifmarkers are available on bothchromosome arms. (ii) Haploids (and aneu-ploids that usually tend to be selected against)arise by nondisjunction at diploid mitosis withrandom assortment of chromosomes seldom ac-companied by recombination within linkagegroups. Hence, lack of segregation of two lociafter haploidization indicates that these loci areprobably on the same chromosome.
Either heterokaryons or heterozygous dip-loids may be used for dominance and comple-mentation tests, although the irregular nuclearconstitution ofheterokaryons may preclude rig-orous tests, and, therefore, heterozygous dip-loids are probably preferable. For further de-tails of genetic analysis by the sexual and par-asexual cycles, see Fincham and Day (91) andHopwood (135).The parasexual cycle occurs in P. chryso-
genum, C. acremonium, and P. patulum, andproblems ofgenetic analysis have been found inall three species. However, in many caseswhere the parasexual cycle has been examinedin an imperfect fungus, the parent strains havebeen "improved" mutants selected after muta-genic treatments. Hence, some difficulties maywell be due to mutations or chromosome aber-rations induced during these treatments. Cer-tain problems, however, may be due to herita-ble characteristics such as the ability to form
heterokaryons, which, for example, variesmarkedly among aspergilli. Hence, whereasheterokaryons in A. nidulans are invariably"forced" by the use of auxotrophic markers, inAspergillus amstelodami heterokaryons (as-sessed by formation of mixed conidial heads)readily arise between prototrophic strains (C.E. Caten, personal communication). To estab-lish heterokaryons in the penicillia and cephal-osporia, one must usually mix conidia of thetwo auxotrophic parents on supplemented min-imal medium (223) or complete medium (191)and allow growth for 7 to 10 days, after whichthe mycelial mat is broken up and plated on orin minimal agar medium.Such problems of poor heterokaryon forma-
tion may well be overcome by the use ofinducedprotoplast fusion, which has recently been dem-onstrated in A. nidulans, Aspergillus niger, P.chrysogenum, P. notatum, P. patulum, and C.acremonium (13, 90) and also between Penicil-hum roquefortii and P. chrysogenum (12).
Diploids have proved particularly difficult toisolate in C. acremonium (223), P. patulum(49), and E. salmosynnemata (87). In C. acre-monium and P. patulum, frequent small-spored prototrophic colonies arose as sectorsfrom heterokaryons or by plating spores fromheterokaryons. The prototrophs could not beinduced to segregate by treatment with p-fluorophenylalanine (PFA), an effective hap-loidization agent in fungi (177). It seems likelythat unstable diploid clones had arisen withinthe heterokaryotic colony and that the proto-trophic colonies were the products of mitoticsegregation and haploidization in these dip-loids. Variations in the "classical" parasexualcycle as found in A. nidulans or P. chryso-genum may well be found as more imperfectfungi are studied. An extreme example occursin Humicola species (66) where heterozygousdiploids were isolated directly from mixedcultures of auxotrophs without any detectableintervening heterokaryotic state.When diploids were eventually isolated in C.
acremonium (223) and P. patulum (49), theywere extremely stable and did not segregatespontaneously, although haploidization wasachieved by treatment with PFA. The haploidsegregants in C. acremonium (223) showed ap-parent random assortment of markers, al-though the recombinant double auxotrophs oc-curred at a very low frequency, which was at-tributed to poor germination. In P. patulum(49), haploid segregants occurred very rarely,and, therefore, no genetic analysis was carriedout. Recombinant genotypes were obtainedfrom a presumptive heterozygous diploid in E.
salmosynnemata (87), but the nonrecovery ofsome recombinant types and the lack of conid-ial measurements makes it impossible to besure if haploidization had occurred.The parasexual cycle has been applied more
extensively in P. chrysogenum (20, 21, 83, 189-193) where diploids can be obtained at frequen-cies of 10f to 10-7 and show a low frequency ofspontaneous segregation.
In early studies with "improved titer" strainsof P. chrysogenum (83, 192, 193, 263), mosthaploid segregants from diploids were of one orthe other parental genotype, a phenomenontermed "parental genome segregation" (192).This was probably due to the parent haploidsdiffering in chromosomal rearrangements suchas reciprocal translocations, which preventedrandom chromosome assortment. By usingclosely related 'sister" strains as parents, thesebarriers to recombination were overcome, andrandom assortment of "haploidization groups"(presumably chromosomes) was achieved (20).Since spontaneous mitotic recombination and
haploidization are both rare inP. chrysogenum(20), it is convenient to induce segregation withchemical and physical agents. Nitrous acid, ni-trogen mustard, ultraviolet light (UV), and Xrays all significantly increased segregationfrom a heterozygous diploid, although effects onmitotic crossing-over and on haploidizationcould not be clearly distinguished (186). UV,
ethylenimine, nitrous acid, and 5-fluorouracilwere tested for their ability to increase mitoticcrossing-over specifically, and ethylenimineand 5-fluorouracil both increased it more than15-fold (K. B. Morrison and C. Ball, Abstr. 2ndInt. Symp. Genet. Ind. Microorg. 1974, p. 83).PFA increased haploidization in P. chryso-genum and also significantly reduced selectionagainst certain alleles (20). Other chemicalagents such as benlate (121) are also good hap-loidizing agents; benlate was more effectivethan PFA for haploidization in P. chrysogenum(G. F. St. L. Edwards, I. D. Normansell, and G.Holt, Aspergillus Newsletter 12, 1975, p. 15).The antibiotic griseofulvin induced haploids inCoprinus lagopus (222) and may also be effec-tive in other fingi. When haploidization hasbeen achieved, other chemicals such as N-glyceryl polyfingin can be used to select thehaploids (18).
ActinomycetesMost actinomycetes produce antibiotics. We
shall, therefore, review the genetics of thegroup, since interesting compounds may wellbe discovered in the relatives of antibiotic non-
producers that have been studied genetically.
Recombination is very widespread, havingbeen reported in the genera Streptomyces, No-cardia, Micromonospora, Mycobacterium, andThermoactinomyces. We are not aware of ge-netic studies in Actinoplanes, Streptosporan-gium, or Streptoverticillium, but increased in-terest in antibiotics produced by these orga-nisms may lead to genetic studies.
In mesophilic organisms of the genera Strep-tomyces, Nocardia, and Micromonospora,there is no firm evidence for transduction ortransformation; some preliminary reports oftransformation in Streptomyces (13) and one oftransduction (10) have not been confirmed orextended. With the exceptions of transforma-tion in Thermoactinomyces vulgaris (see be-low) and of some limited studies oftransductionin Mycobacterium species, known genetic inter-actions in actinomycetes are currently confinedto naturally occurring conjugation phenomenaand to artificially induced protoplast fusion(146a). However, very few serious attemptshave been made to develop systems of geneticanalysis for streptomycetes based on transduc-tion or transformation, and the existence ofsuch systems is, therefore, not precluded.Transfection of protoplasts by bacteriophageDNA and conditions for regenerating proto-plasts into viable mycelia have been achieved(224, 226, 227), so that useful systems for intro-ducing biologically active DNA into theseorganisms may be potentially available.Transformation in Thermoactinomyces.
Typical transformation occurred in the thermo-philic Thermoactinomyces vulgaris (144), withtransformation frequencies per marker up to10-3 when donor DNA was added to growingrecipient cultures, or when donor and recipientwere grown together. Recipients became com-petent at a particular stage in mycelial growth,possibly associated with the onset of sporula-tion. More than halfofa sample of 20 wild typeswere competent (139, 144); of the rest, a smallminority were unable to act as DNA donors (orrecipients) in mixed-growing cultures, and theyexcreted deoxyribonuclease in amounts proba-bly sufficient to abolish transformation (139). Itis suggestive that typical transformation occursin T. vulgaris with an endospore structure (65)and a DNA base composition (273) similar tobacilli; it supports the view that T. vulgarisand its relatives are distinct evolutionarilyfrom most other groups of actinomycetes.Conjugation in other genera. In mesophilic
actinomycetes, the morphological basis of geneexchange is unknown, but conjugation is indi-cated by two observations: contact between via-ble (and, in practice, growing) cells or mycelia
is required for gene exchange, and the geneticconsequences are those characteristic of conju-gation, since groups of linked markers fromeach parent are typically inherited together.The involvement of plasmid sex factors in S.coelicolor A3(2) (see below) may imply a closersimilarity to conjugation in gram-negative eu-bacteria than might be expected from the taxo-nomic gulf between the two groups. Heterokar-yons, which are common in some streptomy-cetes, perhaps arise by a different mechanism(see below).There is at least one significant difference
between conjugation in Streptomyces species,Micromonospora species, and Nocardia medi-terranei on the one hand and other Nocardiaspecies (N. erythropolis, N. canicruria, and N.restrictus) and Mycobacterium smegmatis onthe other: each examined strain of the lattergroup was self-sterile but fertile in crosses withother strains (N. erythropolis x N. canicruria[2], N. canicruria x N. restrictus [C. Vezina,Abstr. 1st, Int. Symp. Genet. Ind. Microorg.1970, p. 167], and interstrain crosses of M.smegmatis [284]). Genetic analyses of inter-strain crosses may have complicated interpre-tation of the Nocardia data (2). A divergence ofgenetic behavior within the genus Nocardiareflects the idea (see above) that N. erythro-polis, N. canicruria, and N. restrictus repre-sent part of a 'cluster" of strains within "Myco-bacterium rodochrous" (64), whereas N. medi-terranei is taxonomically close to Streptomyces .Since the self-sterile strains are not known toproduce important antibiotics, we shall not con-sider their genetics further.Conjugation and its consequences in Strep-
tomyces. Table 2 lists genetic phenomena inStreptomyces other than S. coelicolor A3(2).There are many reports of recombination, butin some cases, unfortunately including pro-ducers of four major antibiotics -chlortetracy-cline (S. aureofaciens), erythromycin (S. ery-threus), neomycin (S. fradiae), and streptomy-cin (S. griseus) -the data did little more thanestablish its existence. Another early report, inS. griseoflavus, included the intriguing possi-bility of a regular diploid or dikaryotic state ofthe genes in spores in contrast to the haploidyof most species, but these studies, also, havenot been extended. For one ofthe other species,S. olivaceus, a series of results resembling thoseof the earlier studies in S. coelicolor A3(2) hasappeared, mainly in Ukrainian, but with a re-cent English summary (205). This included theremarkable claim that the hairy projections onthe spore surface, interpreted by others as inert
appendages (306), are sexual organs analogousto the pili of enteric bacteria. A priori, matingbetween spores seems unlikely; moreover itwas excluded in S. coelicolor A3(2), whosespores in any case are smooth (306); gene trans-fer occurred only after incubation of spores longenough for germination (136).
S. coelicolor A3(2) is the strain in which thegreat majority of genetics has been done (140,264) and the only one in which plasmid fertilityfactors have been recognized (29, 140, 145, 240,293). It probably provides a good model for otherstrains, even though some differences, at leastwith respect to heterokaryon and heterocloneformation, are already apparent betweenstrains (see below). The marked similarities ofgene sequence on the circular linkage maps ofthe species so far examined may in any caseindicate that they are rather closely related. Inview of the production of two antibiotics by S.coelicolor A3(2), one (actinorhodin) chromo-somally determined (308) and the other (methy-lenomycin) by plasmid-borne genes (164, 307),this organism offers good possibilities for thestudy ofthe genetic control of antibiotic synthe-sis.Linkage analysis in Streptomyces. The com-
monest system of linkage analysis in Strepto-myces is the selection of haploid recombinantsfrom mixed cultures of genetically marked par-ents allowed to sporulate by growth on nonse-lective solid medium for several days. Nonse-lected marker segregation in samples of theselected progeny allows very efficient grossmapping since the linkage maps of the speciesso far studied are short; hence, all chromosomalmarkers show linkage.A general mapping procedure for prelimi-
nary study of a new strain was described byHopwood (130, 134, 135). It was applied success-fully in S. bikiniensis (57), S. rimosus (96), S.glaucescens (25), Streptomyces sp. 3022a (possi-bly S. venezuelae) (94), and S. acrimycini (D.A. Hopwood and H. M. Wright, unpublisheddata). The products of crosses between parentswho each carry two selectable alleles wereplated on four differently supplemented media;the data immediately indicated circular link-age of the four markers, providing the basis forfurther analysis by particular selections as out-lined below.An interesting alternative approach, adopted
in S. venezuelae to overcome the problem ofreversion of selective alleles in poorly fertilecrosses (3), was to disregard all recombinantsdiffering from a parent by a single marker(since they might have included revertants),
and to choose the circular sequence of markersthat minimized quadruple (or sextuple) cross-overs. The approach was valid because the ex-cluded classes could arise by the simplest cross-over pattern, that is, a double crossover on acircular arrangement, irrespective of markersequence, and, therefore, did not contribute se-quence information.Once the positions of some markers are
known, an efficient approach is selection foralleles at two loci on opposite sides of the link-age map with one or more nonselected markersof known map location in each of the two arcsbetween. The data are analyzed for linkage bya two-stage process, considering first the alleleratios and second the segregation of the un-known locus with respect to possible neighbor-ing loci (133, 135). This approach was used inS.rimosus (96), S. glaucescens (25), and S. oliva-ceus (205) in addition to S. coelicolor. For finermapping, the most efficient procedure is usu-ally selection for recombination between out-side markers (139).
In S. coelicolor A3(2), recombination incrosses between certain mixtures of fertilitytypes with respect to the SCP1 plasmid is sofrequent (up to 100% of progeny) that no selec-tion is necessary; samples of total progeny canbe analyzed for linkage.
In S. coelicolor A3(2) (267, 268) and S. oliva-ceus (205) merozygotes became progressivelymore complete as incubation time of the crossincreased, over periods of many hours, and at-tempts were made to interrupt the process ex-perimentally. However, there appears to be noexample so far oflinkage analysis by this proce-dure.
Heteroclones in Streptomyces. Heteroclonesare colonies developing on selective mediumfrom partially diploid units and containingmixtures of recombinant and parental geno-types. In their simplest form, their foundinggenome could arise by single crossing-overwithin a merozygote between a circular chro-mosome of one parent and a linear fragmentfrom the other to generate a terminally redun-dant genome (133). Crossing-over at variouspositions within the duplicated segment duringsubsequent growth of the colony generates afamily of recombinant spores. Colonies derivedby nonselective plating ofthese provide data forthe calculation of map intervals, informationnot easily obtained from haploid analysis. Thismethod established the spacing of markers onthe linkage maps of S. coelicolor A3(2) (131-133, 141), S. rimosus (5, 6), and S. olivaceus(205).Whether or not the model outlined above
explains the origin of all heteroclones has beenquestioned (264). Certainly not all segregationdata fit the model, especially in heteroclonesselected to be heterozygous over long regions(264), but distortions in the frequencies of re-combinant classes due to segregation and com-petition within the growing colony explainmany of the apparent discrepancies (134).The original method for heteroclone selection
(266) required two closely linked markers, onein each parent. A cellophane transfer methodthat renders this unnecessary was later devisedfor S. coelicolor A3(2) (265), and adopted for S.rimosus (5, 6).
In S. achromogenes var. rubradiris, littleprogress could be made in mapping by haploidselection, nor could typical heteroclones of theS. coelicolor type be recovered (56). Evidence ofgene sequence was based primarily on the pat-terns of heterozygosity of groups of markers inthe atypical heteroclones that arose.
Heteroclones, by developing on selective me-dia, provide dominance and complementationtests for nutritional and resistance or sensitiv-ity mutations (119, 120, 141), but whether or notthey can be harnessed to the study of antibioticgenes remains to be seen; probably plasmidprimes will be more useful (see below).Heterokaryons in Streptomyces. Colonies
conforming to the operational definition of fun-gal heterokaryons occur in streptomycetes (39);they grow on selective medium but constantlysegregate both parental genotypes. In somestrains, no true recombinants were found (39,40), but in others, heterokaryons occurred invarying proportions alongside recombinants:for example, in S. rimosus (96) and S. glauces-cens (25). In S. rimosus, heterokaryons werefound only when the selective markers wererather leaky (96); growth on the selective plateswas presumably needed for their formation. InS. coelicolor A3(2), heterokaryons occur veryrarely among samples of selected recombinants(141) but were recovered in large numbers bytransferring mixed growth of complementaryauxotrophs growing on cellophane from nonse-lective to selective medium (249, 264).
In some cases, heterokaryons hindered ge-netic analysis by obscuring the rarer recombi-nants that provided mapping information; in-deed, it is conceivable that rare recombinantsmight have been overlooked in those strains inwhich exclusively heterokaryons were de-tected. However, heterokaryons can also beuseful analytically. Their formation on selec-tive medium provided a complementation testfor auxotrophs in S. coelicolor A3(2) (249) andallowed recovery of newly induced auxotrophic
mutations in comutation studies in S. coelico-lor A3(2) and S. rimosus (51). In S. scabies,heterokaryons between melanin-producing andmelanin-nonproducing strains segregated (al-most) exclusively melanin-producing clones,providing evidence for extrachromosomal de-termination of melanin production (114). Nowthat plasmid involvement in the control of var-ious characters in Streptomyces is being estab-lished (see below), it will be interesting to see ifheterokaryons will provide a clear test for plas-mid involvement.The nature of heterokaryons in these procar-
yotes is an intriguing problem that has notbeen seriously studied. The challenge is toimagine a mecanism allowing the mainte-nance of two complete parental genomes in acommon cytoplasm without frequent opportuni-ties for recombination. Whether or not hetero-karyons and the merozygotes that lead to re-combination arise by different conjugationroutes is also not clear. Intuitively this seems
more likely, and the balance of evidence proba-bly points in this direction, but there is noconclusive proof (136, 264).Plasmids in Streptomyces. S. coelicolor
A3(2) is the only strain in which plasmids havebeen definitely implicated in .gene exchange
(29, 140, 145), but, in view of the similaritiesbetween streptomycetes, it seems likely thatthey will be found in other species. The SCP1plasmid, which also determines methylenomy-cin synthesis, is responsible for a limitedamount of chromosome transfer when in theautonomous state, SCP1+ x SCP1- crossesbeing approximately 10 times more fertile thanSCP1- x SCP1- crosses. Recently a second sexfactor, SCP2, has been revealed (29), responsi-ble for nearly all of the recombination, thatoccurs at a frequency of 10O to 10 in SCP1- x
SCP1- crosses. SCP1 can integrate into thechromosome in various ways (145) and is then avery efficient agent for the transfer of chromo-somal segments to SCP1- strains with nearly100% efficiency. It can also acquire chromo-somal insertions to give the SCP1' state (144-146). In principle, SCP1' strains (or SCP2' ifthese occur) provide the best hope for quantita-tive complementation and dominance tests, buta systematic method for their isolation has notyet been devised. DNA corresponding to SCP1has not been isolated, precluding in vitro ma-
nipulation of the plasmid, but SCP2 has beenisolated as a plasmid of molecular weight 18 x106 to 20 x 106 (29, 257).In other streptomycetes, genetically defined
plasmids have been implicated in the produc-tion of melanin (26, 114) or antibiotics (see be-
low), but there is no proof that they are sexfactors.
Interspecific recombination in Strepto-myces. Some empirical attempts to select inter-specific recombinants have succeeded (4, 7, 182,235). However, imperfect homology between ge-nomes is likely to limit such recombination,and the best hope for interspecific gene transferis probably the development ofsuitable plasmidvectors.Recombination in N. mediterranei. Analy-
sis of haploid recombinants in the rifamycin-producing Nocardia mediterranei gave resultsstrikingly similar to those in streptomycetes,extending even to a resemblance in gene se-quence on the circular linkage map (259). Al-though the proportion of mixed-recombinantcolonies was higher than in typical strepto-mycete crosses, presumably because the plat-ing units were often multinucleate mycelialfragments, this proved no serious obstacle togenetic analysis.Recombination in Micromonospora. Recom-
bination studies in Micromonospora (28) estab-lished a conjugation type of gene exchange inself-fertile strains of M. chalcea, M. echinos-pora, and M. purpurea but did not extend tolinkage analysis.Recombination by protoplast fusion. It has
recently been shown (146a) that very efficientrecombination occurs in several streptomyceteswhen protoplasts are artificially fused by poly-ethylene glycol treatment and allowed to re-generate on nonselective medium. This tech-nique promises to be extremely useful for thedevelopment of recombination studies in newspecies, probably being capable of extension tomembers of other genera and perhaps to inter-specific crosses.
EubacteriaBacillus and Pseudomonas are the signifi-
cant genera of eubacterial antibiotic producers.Apparently, 66 antibiotics are produced bystrains ofBacillus subtilis (27), two closely re-lated strains of which have been extensivelyinvestigated genetically. A major thrust hasbeen the analysis of sporulation (253), the ge-netics of antibiotic production having beenstudied only for its possible relationship to spor-ulation: most mutants blocked early in sporula-tion fail to produce antibiotic(s), but it is un-likely that the genes identified by such muta-tions are directly concerned in antibiotic syn-thesis. Strains producing the important Bacil-lus antibiotics, several ofthem studied in detailbiochemically (see below), have not been devel-oped as experimental genetic systems.
B. subtilis genetics depends on transforma-tion and generalized transduction. Transforma-tion and transduction, by phage SP10, are welladapted to mapping over comparatively shortmap intervals (about 1% of the whole genome),but for the large phage PBS1, the proportion isas high as 5 to 8%. A combination of transfor-mational and transductional approaches has al-lowed a complete circular linkage map to bedefined (174). A system allowing dominanceand complementation testing, lacking for manyyears in B. subtilis, has recently become avail-able for some chromosomal regions; it involvesa special class of partially diploid strains (15).There are several useful reviews of techniquesin B. subtilis genetics (309, 311).Transformation has been studied in certain
other bacilli, including strains of Bacillus li-cheniformis and Bacillus pumilus (309), butapparently not in relation to antibiotic synthe-sis.Pseudomonas, the second most prolific genus
of eubacterial antibiotic producers, is also agroup in which genetic studies are fast develop-ing. However, once again, the strains studiedgenetically (126, 274) have not been investi-gated from the standpoint of antibiotic synthe-sis.
Certain strains of Pseudomonas aeruginosahave a well developed conjugation system,whereas general transduction has been used forgenetic analysis inP. aeruginosa andP. putida(126, 274). A particular interest of Pseudomo-nas genetics is the existence of plasmids, manyof them self-transmissible, that control cata-bolic functions (99). In view of the importanceof such plasmids in pseudomonads and theimplication of plasmids in antibiotic synthesisin streptomycetes (see below), it would be inter-esting to investigate the possibility that someof the Pseudomonas antibiotics are plasmiddetermined.Transformation has been studied in Strepto-
coccus (77), but chromosomal recombinationhas not been related to studies of the putativenisin plasmid (see below).The vast majority of genetic work on eubac-
teria, of course, concerns the Enterobacteria-ceae, exemplified particularly by Escherichiacoli and Salmonella typhimurium. This groupproduces some antibiotics (27) but none of im-portance, and no genetic studies appear to havebeen made on them.
SOME REMARKS ON THE PHYSIOLOGYOF ANTIBIOTIC PRODUCTION
Several concepts are relevant to antibioticproduction that do not arise in studies of the
genetics of intermediary metabolism. Theyhave important bearings on the genetics of an-tibiotic production, particularly on its regula-tion and on the consequences of mutationalinterruption of biosynthetic pathways.
Primary and Secondary MetabolitesAlthough the boundary between the two
areas of metabolism is imprecise, it seems use-ful to distinguish "primary" from "secondary"metabolism (42, 43, 286). The first concerns thesynthesis of materials essential for the growthof organisms: the components of proteins, nu-cleic acids, carbohydrates, lipids, coenzymes,etc.; and the pathways are likely, and haveindeed often been shown, to be very similarover broad groups of microorganisms; the genesinvolved have been conserved over long periodsof evolution. On the other hand, materials suchas antibiotics, which are produced only by cer-tain groups of microbes and are often differenteven between apparently closely relatedstrains, are known as secondary metabolites, atleast in cases where their role in the life of theproducing organism is open to argument. Notonly is the genetic information for their synthe-sis confined to particular narrow taxonomicgroups, but it is also usually, though not al-ways, expressed only at defined stages in thelife cycle of the organism -or of a batch culturein a fermentor. The growth of such a culture isoften divided into an early "trophophase," whencellular or mycelial growth occurs through theoperation of primary metabolism, and a later"idiophase," when growth slows down or ceasesand when secondary metabolic pathways startto operate.
Regulation of Secondary Metabolic PathwaysUnderstanding of the genetic basis of the
regulation of primary metabolic pathways ismost advanced in certain eubacteria, notablyE. coli. Rather precise models describe specificcontrols for transcription of particular operonsor for the activity of particular enzymes as wellas for more generalized controls operatingthrough catabolite repression via the concen-trations of effector molecules such as cyclic nu-cleotides (245). The study of isolated mutationswas the single most powerful tool needed todescribe such control systems. In fungi muchless is known, but comparable mechanisms in-volving specific regulator genes and general-ized control (through carbon or nitrogen catabo-lite repression) are being revealed by the studyof mutants (62, 302).For antibiotics, evidence is accumulating for
induction or derepression, at the onset of idi-ophase, of genes specifying the biosynthetic en-
zymes, and attempts are being made to recog-nize those effector molecules or signals that aredirectly involved, whether these include inor-ganic phosphate, adenosine 5'-triphosphate,energy charge, cyclic nucleotides, or others(252, 298). However, it is difficult to see howsuch control systems can be understood withoutthe isolation of mutations specifically modify-ing or abolishing them. A start has been madein the study of the regulation of penicillin andcephalosporin synthesis (see below), but muchmore genetic analysis of regulatory mutants isneeded, particularly a study of their dominancerelations, etc., in order to build satisfyingmodels of regulation.
Pleiotropic Effects of Mutations onAntibiotic Production
In any mutational study of antibiotic-produc-ing organisms, mutations are encountered thatreduce or abolish production but whose effect isreasonably interpreted as indirect and nonspe-cific.To carry out a genetic analysis of antibiotic
production, genetic markers are required. Aux-otrophic mutations are often used, and thesecommonly depress titer significantly. Bonner(33) found only 4% of 400 monoauxotrophs ofP.notatum defective in penicillin production, butthese results appear to be rather atypical. Mac-donald et al. (189) found the majority of aminoacid and vitamin auxotrophs of P. chryso-genum to have yields less than halfthat oftheirparents. In Emericellopsis species, auxotrophicmarkers usually caused a significant reductionin or complete loss of antibiotic production (87).Pleiotropic depression of antibiotic titer by aux-otrophic mutations seems also to be common instreptomycetes (81, 214).Morphological mutations, also, tend to have
pleoitropic effects on antibiotic production. InB. subtilis, 20% of sporulation mutants showedno antibiotic activity (19), and all these wereblocked at stage 0, suggesting an associationbetween antibiotic production and an early stepin sporulation. In S. coelicolor, production ofactinorhodin and methylenomycin is preventedby mutation at the bidA, B, and D loci, all ofwhich prevent normal aerial mycelium devel-opment (211); on the other hand, the variants ofStreptomyces alboniger that had lost the abilityto produce aerial mycelium, and which mayhave arisen by plasmid loss, still produced pur-omycin (243).One of the most detailed studies of a relation-
ship between antibiotic synthesis and morpho-logical differentiation concerns C. acremonium(221). In submerged culture, four morphologicalforms occur: hyphae, arthrospores, conidia, and
germlings. The phase of hyphal differentiationinto arthrospores coincides with the maximumrate of 4-lactam antibiotic synthesis (272), and,when arthrospores were enriched by densitycentrifugation, they were found to have 40%greater antibiotic-producing activity than anyother morphological cell type (221). Further-more, in a series of mutants, each with anincreased potential to produce 84-lactam anti-biotics, differentiation into arthrospores was di-rectly proportional to the increased antibiotictiter (221). Since a nonantibiotic-producing mu-tant readily differentiated into arthrospores,antibiotic synthesis was not a prerequisite forcellular differentiation; this is the case also inS. coelicolor, in which most nonproducing mu-tants of actinorhodin and methylenomycin dif-ferentiate normally, and in B. licheniformis, inwhich a bacitracin-negative mutant sporulatednormally (117).
The Adaptative Significance of AntibioticsThere has been much speculation on this
topic, and diametrically opposed conclusionshave been reached (71). That antibiotics are infact produced in nature, albeit in low concen-trations, seems beyond doubt (305). That theirproduction is adaptive, at least in the majorityof cases, is a priori almost certain in view ofthedifficulty of finding adaptively neutral charac-ters in serious studies of population genetics ofhigher organisms. Increased knowledge of thegenetic control of antibiotic production couldundoubtedly illuminate discussions of theirnatural role. For example, the finding (210)that the majority wild strains of A. nidulansproduced a similar titer of penicillin but thatthe titer-demonstrating genes were different indifferent strains provides strong support for theadaptive nature, not only of the capacity toproduce an antibiotic, but of a particular levelof antibiotic synthesis.
"Low Specificity" of Secondary MetabolicEnzymes
The typical result of mutational loss of anenzyme of a primary biosynthetic pathway isinterruption of the pathway with accumulationofthe intermediary metabolite immediately be-fore the block; rather rarely is this metabolizedby a side pathway to a "shunt" product. Thisfollows from the fact that most pathways ofprimary metabolism are catalyzed by enzymesof high specificity for their normal substrate. Aseries of modifications to a carbon skeleton, forexample, must occur in a precise sequence ifthey are to take place. In contrast, enzymescatalyzing steps of secondary metabolism aremuch less sensitive to variations in the substit-
uent functions of a molecule. This means thatnot all traffic along a pathway is going by thesame route from starting material to end prod-uct; the different functions may be carried outin various sequences so that the system resem-bles a metabolic "grid" rather than a linearpathway (43, 44). In wild-type organisms, var-ious combinations ofthe functions may occur togive a "family" or "complex" of related antibioticmolecules, whereas in mutants lacking a path-way enzyme, simple accumulation of an inter-mediate will not always occur; often enzymesnormally functioning later in the sequence willstill operate with the production of a shuntmetabolite. One of the best examples is in thetetracycline pathway (see below). These factorscomplicate deductions about biosynthetic path-ways made from the nature of mutant metabo-lites or from the pattern of cosynthesis (thesecretion ofan antibiotic precursor by a blockedmutant that can be converted to antibiotic by amutant blocked at an earlier step in the path-way) between pairs of blocked mutants. On theother hand, they greatly widen the scope for theproduction of unnatural antibiotics by muta-tional biosynthesis (see below).
Chemical Classes of Antibiotics
At first sight, antibiotics, particularly thoseproduced by actinomycetes, have a bewilderingarray of chemical structures. However, bio-chemical studies indicate considerably moresimilarity in their biosynthetic routes than thediversity of their structures might suggest. Aparticularly important unifying hypothesis isthe polyketide concept (30, 286). The idea is ofaclose analogy with fatty acid synthesis in whicha multienzyme "fatty acid synthetase" cata-lyzes the sequential addition, associated withdecarboxylation, of 2-carbon units from malo-nyl coenzyme A (CoA) "extender" units to a 2-carbon acetyl CoA "primer," the growing car-bon chain remaining enzyme bound until fullygrown. In the case of some actinomycete anti-biotics, the primer is propionyl CoA and theextender is methylmalonyl CoA, so that 3-car-bon units are added, leading to the presence ofmethyl groups attached to the growing polyke-tide chain; various other variations are alsopossible. By cyclization reactions, the polyke-tide is stabilized, and only at this stage are freebiosynthetic intermediates of the final anti-biotic produced. Enzymological studies appearto be most advanced for 6-methylsalicylic acidsynthetase from P. patulum (75), which is in-volved in the synthesis ofthe antibiotic patulin,but evidence for the existence of multienzymepolyketide synthetases, susceptible, like fatty
acid synthetases, to inhibition by the antibioticcerulenin (228), is available for several of theimportant actinomycete antibiotics, includingtetracyclines, erythromycins, and rifamycins(see below). Not all antibiotics, of course, arisein this way. Many are put together from moie-ties derived from typical primary metabolitessuch as amino acids or sugars: the 3-lactamantibiotics characteristic of fungi and the largeclass of aminocyclitol actinomycete antibioticsare examples. Still another completely differ-ent biosynthetic mode is the nonribosomal pep-tide assembly process characteristic of manyantibiotics of bacilli.
Why "Biogenesis"?Many biochemical papers refer to the biosyn-
thesis of antibiotics as "biogenesis." This seemsto be an unnecessary term with almost mysticovertones. A more serious criticism applies toits adjectival form "biogenetic," which impliesthat studies described by this term have a ge-netic content. What is wrong with the termsbiosynthesis and biosynthetic?
MUTATIONAL STUDIES OF ANTIBIOTICSYNTHESIS
FungiP. chrysogenum. In 1947 Bonner (33), noting
that 25% oflysine auxotrophs produced no peni-cillin, predicted a common precursor of penicil-lin and lysine. Ten years later (68), lysine wasfound to be a potent inhibitor of penicillin syn-thesis; a precursor of lysine biosynthesis infungi, a-aminoadipic acid, not only reversedthe inhibition due to lysine but stimulated pen-icillin synthesis in its absence (273). In themeantime, a noncyclic tripeptide containingresidues of a-aminoadipic acid, cysteine, andvaline was isolated (14), giving rise to the tri-peptide theory of penicillin biosynthesis.
Genetic evidence for a-aminoadipic acid asan essential intermediate in penicillin synthe-sis was obtained (111, 220). Lysine auxotrophsfell into two phenotypic classes with regard topenicillin synthesis: mutants blocked before a-aminoadipic acid could make penicillin in thepresence of lysine only if a-aminoadipic acidwas added, but those blocked after a-aminoadi-pic acid synthesized antibiotic when grownwith lysine alone. Lysine exerts end productcontrol on a-aminoadipic acid formation, and,hence, decreases the amount available for peni-cillin biosynthesis (111). These results and la-beling experiments appear to confirm that botha-aminoadipic acid and the tripeptide L-(a-ami-noadipyl)-L-cysteinyl-D-valine are penicillin
precursors (89), but the final stages of the path-way are unknown.Two possible precursors of benzylpenicillin
(penicillin G) have been isolated from fermen-tations devoid of side chain precursors: isopeni-cillin N (94) and the penicillin nucleus, 6-APA(24). Most evidence favors isopenicillin N as theprecursor ofthe penicillins produced byP. chry-sogenum (70), and 6-APA may be a shunt me-tabolite derived by deacylation of isopenicillinN in the absence of a side chain precursor. Theterminal reaction of penicillin G synthesis istherefore likely to be an exchange of the L-aminoadipic acid side chain of isopenicillin Nfor phenylacetic acid from phenylacetyl CoA.An acyltransferase that could catalyze such areaction occurs in crude extracts, and this en-zyme or multienzyme complex may catalyze atleast five other related reactions (70). The pos-tulated pathway of penicillin G synthesis isshown in Fig. 1.In spite of the attention paid to the parasex-
ual cycle in P. chrysogenum, it is remarkablethat very little use has been made of blockedmutants to study penicillin biosynthesis. Someearly studies of such npe mutants were de-scribed by Sermonti (48, 260). A more detailedcomplementation analysis was reported by Holtet al. (129). A sample ofmutants producing lessthan 10% of the parental titer were selectedfrom strain NRRL 1951. Five complementationgroups (V, W, X, Y, and Z) were identified with
1, 2, 2, 9, and 1 representatives, respectively(187). Mutants in four of the five groups weremorphologically normal, but the group Z mu-tant had a reduced growth rate in surface cul-ture. Parasexual analysis assigned members ofgroups W, X, Y, and Z to the same haploidiza-tion group (chromosome) and the group V mu-tant to a separate chromosome.
Preliminary characterization of a number ofnpe mutants (220) used a lysine auxotrophblocked after a-aminoadipic acid. This strainsynthesized penicillin when grown in the pres-ence of lysine; when 14C-labeled a-aminoadipicacid was supplied, label was incorporated intoseven intracellular compounds including a-aminoadipyl-cysteinyl-valine and isopenicillinN. Eleven npe mutants selected from thisauxotroph were examined for their ability toform the labeled intermediates. A complex pat-tern was found, and the eleven mutants couldbe grouped into nine types. Only one compoundwas common to all the mutants, and the resultsdid not unambiguously suggest an order forbiosynthesis of the seven compounds. Analysisof the biosynthetic pathway may, however,have been confounded by the production ofshunt metabolites as has been found for cepha-losporin nonproducing mutants (see below).Regulation of penicillin biosynthesis. Regu-
lation of three primary pathways, for lysine,cysteine, and valine, influences directly thesynthesis of penicillin. Regulation of these
pathways has been studied in two ways: byselecting known types of mutation in the path-way and characterizing their effects on anti-biotic production; or by comparing strains se-lected for increased titer with their parentstrain and determining whether or not altera-tions had been selected in the primary path-ways.Demain (69) postulated a branched pathway
for lysine and penicillin (Fig. 1) and suggestedthat the inhibitory effect of lysine on penicillinbiosynthesis was due to negative feedback con-trol of an early enzyme. Lysine feedback in-hibits homocitrate synthase, an early enzymeof lysine biosynthesis, in yeast (195) and inNeurospora crassa (125). A similar inhibitionwas demonstrated in P. chrysogenum by De-main and Masurekar (73), who isolated ninelysine auxotrophs of the Wis54-1255 strain, oneof which was a leaky mutant that could besatisfied by a-aminoadipic acid. It accumulatedhomocitrate, and this accumulation was in-hibited 50% by 0.3 mM lysine. Such feedbackinhibition by lysine occurred in both the growthand the penicillin production phases and sug-gests that penicillin synthesis could be ex-tremely sensitive to lysine present during pro-duction. Homocitrate synthase activity in crudeextracts of the early lysine auxotroph wasfound, somewhat surprisingly, to be insensitiveto lysine inhibition (202) for unknown reasons.
Inhibition of penicillin production under fer-mentation conditions requires higher exoge-nous concentrations (10 to 20 mM) of lysinethan those in a commercial medium. However,endogenous production of lysine might signifi-cantly limit penicillin production. This possibil-ity was examined by selecting lysine-overpro-ducing mutants by their resistance to a lysineanalog; they showed significantly reduced peni-cillin formation (201). It was suggested thatthese mutants were derepressed at a point afterthe branch point in the lysine pathway, result-ing in overproduction of lysine and so an under-production of penicillin. These results indicatean intimate intracellular relationship betweenthe biosynthesis of lysine and penicillin andsuggest that mutants desensitized to lysine in-hibition ofhomocitrate synthase might be supe-rior producers of penicillin.The apparent regulatory control of penicillin
production by lysine led Pirt (234) to postulatethat if a lysine auxotroph blocked after a-ami-noadipate were fed lysine at a low level, thenthe feedback control could be overcome andpenicillin production increased. Three lysineauxotrophs of Wis54-1255 blocked after a-ami-noadipic acid were selected (229); two (H and
13a) were blocked between saccharopine andlysine and showed very low saccharopine reduc-tase activity, and the other (45) showed no a-aminoadipic acid reductase activity (Fig. 1). Allthese blocks caused a large decrease in penicil-lin synthesis, and only in strain H could this beovercome by feeding lysine. However, maxi-mum penicillin production by strain H requiredlysine feed concentrations 180% in excess ofthat required for growth, and even then thetiter was less then that of the parent with nolysine feed. The virtually complete failure ofstrain 45 (blocked in a-aminoadipic acid reduc-tase) to produce penicillin suggested that thebranch point between lysine and penicillinpathways may be adenyl a-aminoadipic acidrather than a-aminoadipic acid (Fig. 1).A second aspect ofprimary metabolism influ-
encing penicillin synthesis is sulfur metabo-lism. Studies with 35S showed that sulfur forpenicillin comes efficiently from sulfate via thesulfate reduction pathway but can also be de-rived by reverse transsulfuration from methio-nine. Tardrew and Johnson (281) examined theuptake of inorganic sulfate and the excretion oforganic sulfate during the penicillin-producingphase in a number of wild-type and improvedstrains. The high-yielding mutants took upmore inorganic sulfur from the medium thanthe wild type, and the concentration of inor-ganic sulfur in the mycelium of one mutant(Wis5l-20F3) was at least twice that in its an-cestral strain NRRL 1951-B25. This increasedcapacity for sulfur uptake was apparently in-troduced between strain X1612 and strainQ176, which was obtained from X1612 by UVirradiation, and it may be that this change inmetabolism is primarily responsible for thedoubling in titer between X1612 and Q176.The third amino acid involved in the biosyn-
thesis of penicillin is valine, which is synthe-sised by the same pathway in fungi and bacte-ria. In P. chrysogenum Q176, the first enzymein the valine pathway, acetohydroxy acid syn-thetase, is sensitive to feedback inhibition byvaline (112). In a high-titer derivative of Q176,the enzyme had lost one ofthe two binding sitesfor valine, allowing increased formation of theamino acid. The high-titer strain also producedmore enzyme than Q176.A. nidulans. Although it is not used commer-
cially, the availability of sexual and parasexualcycles in Aspergillus nidulans and its exten-sive use in genetic studies (155, 237, 238) haveprompted several workers to use it as a modelfor genetic analysis of penicillin production.Mutants were classified as "penicillinless"
parental level (129). Twenty-eight such mu-tants were selected after treatment with a vari-ety of mutagens; 11 had a significantly lowercolony radial growth rate than the parentstrain, and 8 showed changes in mycelial orconidial pigmentation. In at least one (npeC007), such effects appeared to be due to pleio-tropy of the npe mutation. A heterokaryon test(151) between npe+ and npe- strains showedthat the npe character was under nuclear con-trol and all the mutations were recessive indiploids. The 28 mutations fell into at least fivecomplementation groups, one of them (npeA)containing 20 mutations (187). The npeA locuswas mapped by parasexual and sexual analysis(129) to chromosome VI, probably distal to sbA.As in P. chrysogenum, biochemical analysis ofthese mutants will be of considerable use inelucidating the pathway and control of penicil-lin biosynthesis, while application of geneticsshould be much more straightforward.
C. acremonium. Cephalosporium (andEmericellopsis) species produce two (-lactamheterocyclic antibiotics with n-a-aminoadipicacid as an acyl side chain and either 6-APA(PCN) or 7-ACA (CPC) as nuclei. Little or nofree 6-APA and no 7-ACA is formed during aCephalosporium fermentation, and the addi-tion ofside chain precursors has no influence onthe antibiotics formed, which are always PCNor CPC.Cephalosporia produce a tripeptide [i.(a-
aminoadipyl)-i-cysteinyl-n-valine] similar tothat produced by penicillia, although the intra-cellular levels are much lower. Two tetrapep-tides containing a-aminoadipic acid, cysteine,valine, and glycine or a-aminoadipic acid, cys-teine, 3-hydroxyvaline, and glycine have alsobeen isolated from mycelia of C. acremonium(181), but their involvement in CPC and PCNbiosynthesis is not understood.Biosynthesis of the tripeptide. Genetic stud-
ies have clearly indicated involvement of thetripeptide in antibiotic synthesis. Lemke andNash (173a) obtained seven mutants that couldnot synthesize ,B-lactam antibiotics. Six ofthemwere prototrophic and fell into two classes. Twomutants (peptide-) did not incorporate labelfrom valine or a-aminoadipic acid into the cys-teine-containing peptides, whereas the remain-ing four (peptide+) did so. Mutants of the firstclass were presumably blocked in the synthesisof the peptides, and those of the second classwere blocked in the conversion of the peptidesinto antibiotics. Mutants of these same twophenotypic classes were also isolated by others(241; T. Kanzaki, personal communication).Pairs of inactive mutants did not cosynthesize
antibiotics when grown in mixed culture, but aheterokaryon between two peptide- mutants,although rather unstable, synthesized anti-biotic at 10% of the parental level (220). Thisobservation is ofparticular interest since it sug-gests that at least two genes may be involved inthe synthesis ofthe tripeptide from its constitu-ent amino acids, although no intermediateshave been identified. Heterokaryons betweenpeptide- and peptide+ mutants produce anti-biotic (S. Queener and T. Kanzaki, personalcommunications), and in one heterokaryon be-tween two peptide+ mutants no antibiotic wasproduced (Kanzaki, personal communication).The seventh mutant isolated by Lemke and
Nash (173a) was a lysine auxotroph blocked ina step after a-aminoadipic acid. The mutantsynthesized trace amounts of antibiotic- andcysteine-containing peptides in the presence ofa-aminoadipic acid and lysine but not with ly-sine alone. This result is surprising since onewould expect such a mutant to produce PCNand CPC when grown on lysine alone. Essen-tially similar results were reported by Nueschet al. (Abstr. 4th Int. Ferm. Symp., Kyoto 1972,p. 228) for lysine auxotrophs blocked before orafter a-aminoadipic acid. The lack of antibioticproduction is apparently due to extreme sensi-tivity of the mutant to lysine feedback inhibi-tion (70).Biosynthesis of PCN. Virtually nothing is
known about subsequent reactions in penicillinN (PCN) synthesis, but it seems unlikely that6-APA is involved since only trace amounts arefound in mycelium grown on complex media.Lemke and Nash (173a) obtained a mutant withincreased ability to produce 6-APA; it producedpenicillin N but no cephalosporin C.Biosynthesis of CPC. Cephalosporium spe-
cies produce two major compounds in additionto CPC, deacetylcephalosporin C (DCPC) anddeacetoxycephalosporin C (DOCPC) (104, 122).The biosynthetic relationships between thesecompounds have been demonstrated very clearlyby the use of blocked mutants (CPC-).
All CPC- mutants (102, 104, 105, 179, 241)fall into three classes, accumulating PCN,DOCPC, and DCPC; PCN and DOCPC; or PCNonly. This is consistent with a linear biosyn-thetic pathway: PCN -+ DOCPC -- CPC. Analternative branched pathway (Fig. 2) withPCN and CPC as end products (102, 241) is alsoconsistent with the observed accumulation pat-terns but predicts the existence of mutants(CPC+ PCN-) producing CPC but not PCN, andno such mutants have been reported. A test forlabeling of CPC by [14C]PCN (272) failed be-cause the penicillin was not taken up by the
FIG. 2. Hypothetical branched pathway for biosynthesis of PCN and CPC (244) and suggested shuntmetabolites (159, 285).
cells. Recently, however, cephalosporin synthe-sis by cell-free extracts of C. acremonium wasclaimed to be markedly stimulated by PCN butnot by penicillin G or 6-APA (166). These re-sults, taken with the absence of CPC+ PCN-mutants, suggest that PCN may indeed be anintermediate in cephalosporin synthesis.Further evidence for the final step of the
proposed pathway was provided by the demon-stration ofan acetyl CoA:DCPC acetyl transfer-ase that converts DCPC to CPC in CPC+ strains(104, 179) but was lacking in four out of fiveCPC- mutants (104), thus accounting for DCPCaccumulation. An enzyme catalyzing the hy-
droxylation of DOCPC to DCPC has also beenreported (102, 179).
Several apparent shunt metabolites havealso been found in CPC- mutants (Fig. 2).Those blocked before DOCPC, some of whichalso produced no PCN, accumulated the dimerof I-(a-aminoadipyl)-L-cysteinyl-D-valine andthe disulfide ofthe tripeptide and methanethiol(159). Some of the mutants accumulatingDOCPC and PCN produced traces of CPC andaccumulated a new compound identified as N-acetyl DOCPC (285). Mutants accumulatingDCPC also produced a second compound (C-2)believed to be derived from DCPC (101), and in
a potent CPC-producing mutant a new com-pound (F-i) was identified, apparently derivedfrom CPC (160). Hence, CPC provides anotherclear example ofthe accumulation ofshunt me-tabolites in mutants blocked in the biosyntheticpathway of a secondary metabolite.Regulation of CPC biosynthesis. CPC and
PCN biosynthesis is substantially stimulatedby methionine, especially the -isomer, but theexact mecanism is uncertain. Methionine do-nates sulfur efficiently to CPC (50, 223), andsulfur incorporation is thought to occur via thereverse tasfuration pathway (Fig. 3).Whether or not this is the only role for methio-nine has, however, been questioned for a num-ber of reasons. First, the postulated intermedi-ates between methionine and CPC (homocys-teine, cstathionine, and cysteine) will not re-place methionine (74, 223). Second, the nonsul-fur analog of methionine, norleucine, will re-place methionine (78). Third, prototrophictrains produce moderate levels of CPC in adefined medium with sulfate as sole surfacesource. Drew and Demain (80) asked whethersuch antibiotic formation involved methionine,or whether it was due to passage of sulfur di-rectly from sulfate via cysteine to CPC (Fig. 3).The introduction of a mutation in the transsul-furation pathway between cystathionine andhomocysteine eliminated antibiotic production,even in the presence of excess sulfate, thusimplying an obligatory role of endogenous me-thionine in the regulation of cephalosporin bio-synthesis.This conclusion was reinforced by the isola-
tion of two further mutants (79). A mutantso-
4N.\, 274-1
'IV\ cystein
11-8\`11-~~ cystathiom
transulphuratin
cephalosporin C2'
I,-
I'l
Reverse tranasulphuraticn
homocysteine
E- adenosyl~hcuocysteine
rdeylmethiane
Fxo. 3. Probable pathway ofsulfur metabolism inCephalosporium acremonium, (79) showingpositionsofblocks introduced by mutation in strains 274-1 and11-8.
(274-1) blocked between sulfate and cysteine(Fig. 3) grew well on both methionine and cys-teine, but antibiotic production on cysteine wasonly 30% of that on methionine. Thus, even inthe absence ofany possible competition by inor-ganic sulfate, only methionine was effective instimulating antibiotic synthesis. A double mu-tant (114) blocked not only in the pathwayfrom sulfate to cysteine but also in the pathwayfrom cysteine to methionine (Fig. 3) producedvirtually no CPC when grown in excess cys-teine (plus a low level of methioninecessaryto support growth), but supplementation byeither methionine or norleucine very efficientlystimulated cephalosporin synthesis. Such ge-netic and biochemical analysis has thereforeprovided strong evidence that methionine stim-ulation of CPC production is due to a role ofmethionine other than that of sulfur donation,but there is as yet little indication of what thisrole may be.The improvement of CPC production by
strain selection has undoubtedly been due, inpart, to selection of regulatory mutations bothin the CPC pathway and in related primarypathways. Queener et al. (242) showed that in anumber of high-yielding mutants of C. acre-monium M8650 the level ofglutamate dehydro-genase activity is four or five times higher thanin the parent strain, and it seems likely thataltered regulation of glutamate dehydrogenasesynthesis has increased the supply of -glu-tamic acid and hence of a-amino nitrogen forthe three constituent amino acids of CPC.
In the fifth of the CPC- mutants isolated byFujisawa et al. (104; see above), CPC was syn-thesized and then enzymatically hydrolyzed toDCPC by an extracellular CPC acetylhydrolase(CAH) that was not detected in the parentstrain or in the other four mutants (103). Asimilar CAH activity was demonstrated by A.Hinnen and J. Nfiesch (personal communica-tion) in a high-titer derivative of M8650. Thisactivity appeared after 120 h and was paralleledby exhaustion of glucose. They suggested thatsynthesis of CAH may be regulated by carboncatabolite repression, and they also reportedthat mutants with highly reduced CAH activitycan be selected that then retain a hightiter ofCPC under conditions of glucose starvation.P. griseofulvum. The only description of the
use of an induced mutation to alter the anti-biotic produced by a fungus concerns the pro-duction of the 2'-ethoxy analog of griseofulvin(150). Griseofulvin contains three methylgroups derived from choline, and, on the basisof work with tetracyclines (82), it was antici-pated that ethoxy analogs of griseofulvin could
be produced under appropriate fermentationconditions. Six methionine auxotrophs of astrain of Penicillium griseofulvum were ob-tained, three of which grew to a varied extentwith cystine and three of which could utilizesodium thiosulfate. Ethionine was highly toxicto the parent strain of P. griseofulvum and toone of the auxotrophs. Four other auxotrophsshowed reduced- growth on ethionine (200 ,uglml), and one was resistant. Three of the auxo-trophs were shown to produce a new antibiotic,the 2'-ethoxy analog of griseofulvin, whengrown in ethionine-supplemented fermentationbroths. This compound had been considered ashaving possible agricultural application (65),but it was abandoned because of poor transloca-tion in plants.P. patulum. P. patulum produces the anti-
biotic patulin as an extracellular end productvia the classical aromatic polyketide intermedi-ate, 6-methylsalicylic acid (93, 216). The majorpathway of patulin biosynthesis in P. patulum(synonymous with Penicillium urticae andPenicillium flexuosum and closely related to P.griseofulvum) is part of a proposed matrix ofreactions involving up to 16 metabolites (44)and is one of the most extensively studied sec-ondary metabolic pathways in fungi. Althoughpatulin is of no commercial interest, owing toits high toxicity to animals (271), the presenceof a parasexual cycle inP. patulum (49) and theextensive biochemical studies available makegenetic analysis of this system an attractiveproposition as a model secondary metabolic ma-trix.
0. mucida. This basidiomycete, Oudeman-siella mucida, produces an antifungal anti-biotic, mucidin (217). Nonproducing mutantswere isolated after treatment with N-methyl-N'-nitro-N-nitrosoguanidine, and genetic anal-ysis indicated that each of the mutants differedfrom the producing monokaryons in a singlegene (M. Semerdiieva, personal communica-tion).
mycete antibiotics have been made, and inmany of the strains some antibiotic-nonproduc-ing mutants were probably isolated. However,in few cases have sufficiently large series ofmutants been studied to make a crucial contri-bution to the clarification of biosynthetic path-ways, and in even fewer cases have the mu-tants been analyzed genetically. We shall re-view those cases where blocked mutants havebeen important for deducing steps in the bio-synthesis of important antibiotics or where re-combination studies have been made.
Tetracyclines. There are two main producersof tetracycline antibiotics: Streptomyces aureo-faciens, wild types ofwhich make mainly chlor-tetracycline, and Streptomyces rimosus, theproducer of oxytetracycline. Mutants of S. au-reofaciens failing to chlorinate the molecule inposition 7 accumulate tetracycline. Oxytetracy-cline differs from tetracycline by hydroxylationat position 5.The similarity between the antibiotics of the
two strains suggests a largely common biosyn-thetic pathway, so that a study of gene-enzymerelationships for one would probably provide agood model for both. However, biochemicalstudies, aided by a rather complete set ofblocked mutants, are most advanced in S. au-reofaciens, in which recombination studies arelittle developed, whereas in S. rimosus geneticstudies have not been complemented by bio-chemical characterization of the mutants.
Tetracycline biosynthesis provides a good ex-ample of the polyketide route. McCormick (206)presented a pathway of chlortetracycline syn-thesis by S. aureofaciens, the early steps some-what speculative, in which one molecule ofmalonamyl CoA and eight of malonyl CoA arecondensed to a nonaketide that is cyclized togive a ring system for subsequent substitutionat specific positions. A complete set of blockedmutants for the reactions involved in the poly-nuclear part of the pathway was crucial in de-fining the steps. These mutants were used inpairwise cosynthesis tests, and the precursors,or more often shunt products, that they accu-mulated could be identified. Extensive biosyn-thetic studies were also made by the Praguegroup (147), who have emphasized particularlythe series of up to 72 compounds (the shuntproducts of McCormick) that could arise wheneach of 11 biosynthetic steps, after the forma-tion of a hypothetical tricyclic nonaketide, isblocked.Recombination was first described in S. au-
reofaciens by Alikhanian and Borisova (9) andhas been studied more recently by the Praguegroup (32). Unfortunately, no linkage map isavailable, and the locations of the pathwaygenes for chlortetracycline are unknown, al-though some clustering was suggested by a fail-ure to obtain antibiotic-producing recombi-nants by selecting prototrophs from crosses ofauxotrophs carrying mutations blocking thepathway at different steps.The genetics of S. rimosus is much further
advanced, but genetic determination of the bio-synthetic pathway to oxytetracycline is still in-completely known. No data on the mapping ofblocked mutants in the strain studied by Friendand Hopwood (96) have yet been published. In
the strain studied in Zagreb, blocked mutantsbelonging to at least four cosynthesis groupswere mapped to one region of the circular link-age map (5), but no results of fine mappinghave been reported. In a third strain, mutantsbelonging to nine cosynthesis groups weremapped in two clusters (35), but, in the absenceof a general linkage map of the strain, therelative spacing of these two groups was un-known. They fell between two streptomycin re-sistance loci, but the argument that these cor-responded to strA and strB in S. coelicolorA3(2), and were therefore fairly close, needsconfirmation. We can probably conclude fromall these studies that there is significant clus-tering ofthe pathway genes for tetracyclines onthe chromosomes of the producing organisms,but there is no information on the precise spac-ing ofthe genes, nor any details ofgene-enzymerelationships. The possible involvement of aplasmid in oxytetracycline production is dis-cussed below.
Macrolides. Erythromycins. There are fourmain erythromycins, diglycosides of 14-mem-bered lactones (59, 194). All four have the samebasic sugar, desosamine; they differ in the neu-tral sugar and at position 12 in the lactone ring.Erythromycins A (the clinically important com-pound) and C have a 12-hydroxyl, whereas Band the recently discovered D do not; the neu-tral sugar is cladinose in A and B, and myca-rose (lacking a methylether of cladinose) in Cand D.The lactone skeleton of erythromycins is ap-
parently derived from a polyketide chain syn-thesized from propionate units (291) by succes-sive condensation of six methylmalonyl CoAunits onto a propionyl CoA primer, and theproperties of a multienzyme "lactone synthe-tase" involved are being studied. The lactone soformed is apparently 6-deoxyerythronolide B,which is hydroxylated at position 6 to yield thelactone found in erythromycins B and D (ery-thronolide B). Attachment of mycarose and,subsequently, desosamine to the lactone wouldyield erythromycin D, postulated to be the par-ent member of the erythromycin family. Thiscan suffer two substitutions in either order. 0-methylation ofthe mycarose oferythromycin Dby a S-adenosyl methionine-dependent trans-methylase would yield B, and hydroxylation ofits lactone would then give A. Alternatively,hydroxylation of the lactone of D would give Cfollowed by O-methylation of mycarose to giveA (194).Biochemical studies of blocked mutants were
instrumental in deducing certain features ofthis scheme, notably that 6-deoxyerythronolideB is a precursor of erythronolide B (198) and
that mycarose is then added to this lactone toyield a neutral monoglycoside (199). A pre-sumed earlier blocked mutant accumulated alactone that appeared to be a shunt compound,possibly arising through lack of a gene productforming part of the presumed multienzyme lac-tone-synthesizing complex (197). A further useof blocked mutants was the identification of afifth erythromycin, E, formed when blockedmutants unable to make the erythromycinswere fed with erythromycin A (196). E is appar-ently derived from A by a change in the linkageof the cladinose to the lactone ring. Althoughrecombination has twice been reported inStreptomyces erythreus (67, 162), it has notbeen extended to linkage analysis, and blockedmutants have not been analyzed genetically.Information on the location of biosyntheticgenes is therefore completely lacking.Platenomycins. Twenty-four blocked mu-
tants ofStreptomyces platensis subspecies mal-vinus, the producer of these basic 16-memberedmacrolides, were classified into seven cosyn-thesis groups; one group contained non-consyn-thesizing mutants, and two contained mutantsthat acted only as secretors (106). These twoclasses were found to secrete, respectively,monoglycosides, lacking the mycarose moietyof the platenomycins, and lactones (plateno-lides), lacking both this sugar and the mycami-nose moiety. The terminal part of the plateno-mycin synthesis pathway was, therefore, pro-posed (107) to proceed from the platenolides bysuccessive addition of the two sugar moieties,which form a disaccharide linked at one pointto the lactone instead of being independentlyattached to it as in erythromycin. No studieshave been reported on other classes of blockedmutants, and no genetic analysis has been de-scribed; thus the locations of the genes con-cerned are unknown.Turimycin. Possible plasmid involvement in
the synthesis of this macrolide by Streptomyceshygroscopicus is discussed below.Aminocyclitols (aminoglycosides). Com-
pounds of this class consist of two to five moie-ties joined by glycosidic linkages; one moiety isan aminocyclitol (cyclohexane with amino andalcohol substitutions), and the others aresugars. In most commercial members of theclass (neomycin, kanamycin, paromomycin, ri-bostamycin, tobramycin, destomycin, hygro-mycin B, lividomycin, validamycin, gentami-cin, sisomicin), the aminocyclitol is deoxystrep-tamine, but in streptomycin it is streptidine, inspectinomycin, actinamine, and in bluensomy-cin, bluensidine. There is a recent comprehen-sive review of the biosynthesis of the group(246) as well as earlier reviews of streptomycin
biosynthesis (72, 295, 296).The vast majority of biosynthetic studies
have used the approach of feeding radioactivelylabeled or 13C-enriched precursors to biosyn-thetically proficient strains, and there havealso been some enzymatic studies. With vary-ing degrees of certainty, the various aminocy-clitol and sugar moieties have been shown to bederived from glucose, and a few ofthe pathwayshave been deduced in considerable detail, nota-bly for streptidine in S. griseus. Details of thelinking of the various moieties are few; proba-bly dihydrostreptose is added to streptidine-6-phosphate to form a disaccharide intermediatein streptomycin synthesis (165).
Blocked mutants have contributed extremelylittle to pathway clarification, since significantseries of such mutants do not appear to havebeen isolated. Study of a mannosidostreptomy-cin-negative mutant of S. griseus indicatedthat mannosidostreptomycin is not an essentialintermediate in streptomycin synthesis (72),and a few tentative conclusions on the joiningsequence of some antibiotic moieties have beendeduced by feeding blocked mutants with indi-vidual moieties or linked pairs or trios of them(246, 256).Although recombination was reported early
in S. griseus (streptomycin) and Streptomycesfradiae (neomycin) (40), no linkage studieshave been made nor have crosses involvingblocked mutants been described. The same istrue of recombination studies in strains of Mi-cromonospora echinospora and M. purpureamaking gentamicins (28). In fact, there appearsto be no published information about the loca-tion of any gene concerned with the biosyn-thesis of any aminocyclitol antibiotic.Mutational biosynthesis. The aminocyclitol
antibiotics, built from separate moieties, lendthemselves to a technique first described bySchier et al. (255) and later designated "muta-tional biosynthesis" (218). Mutants blocked inthe synthesis of one moiety (in this case theaminocyclitol) are fed with analogs of that moi-ety, which are incorporated by the mutant intounnatural antibiotics. For S. fradiae, Strepto-myces rimosus forma paromomycinus, Strepto-myces kanamyceticus, Streptomyces ribosidifi-cus, and Micromonospora inyoensis, makingneomycin, paromomycin, kanamycin, ribosta-mycin, and sisomicin, respectively, mutantsblocked in deoxystreptamine biosynthesis wereused (167, 254, 255, 283); for S. griseus, makingstreptomycin, streptidine-negative mutantswere fed (218). The analogs used, with varyingdegrees of success, were streptamine, epistrep-tamine, and several others.
BACTERIOL. REv.
Novobiocin. This antibiotic consists of threemoieties, rings A, B, and C, which are a substi-tuted benzoic acid, a substituted aminocou-marin, and a substituted sugar, carbamoylno-viose, respectively. Blocked mutants have sofar made only a modest contribution to explain-ing the pathway of novobiocin synthesis byStreptomyces niveus. Mutants confirmed theroles of carbamylation and methylation of thenoviose sugar moiety (168), and another mu-tant, blocked in the pathway from tyrosine toring B, indicated the sequence of assembly ofthe rings, C being added only after A and B arejoined (0. K. Sebek, Abstr. 2nd Int. Cong. Ge-net. Ind. Microorg. 1974, p. 14). This mutantwas also used in mutational biosynthesis toproduce a novobiocin analog with a methylgroup of the aminocoumarin replaced by chlo-rine.Rifamycins. Rifamycins are the most impor-
tant members of a class of antibiotics namedansamycins; they contain an aliphatic "ansa"chain bridging an aromatic chromophore. Therifamycins are a family of six or more com-pounds produced by N. mediterranei. Labelingexperiments, more recently with 13C-enrichedprecursors, have largely established the originof the carbon skeleton of the rifamycins (172,300). A 7-carbon aromatic ring would initiatesynthesis of a polyketide chain derived fromtwo malonates and eight methylmalonates; thispolyketide would become the ansa chain and,together with the original aromatic ring, thechromophore. The product of this synthetic se-quence should have a continuous carbon skele-ton, yet the various rifamycins produced by thewild-type organism have an ether linkage in-terrupting the chain between carbons 12 and29. A key finding was therefore the isolation,from a blocked mutant, of a novel compound,rifamycin W, lacking the ether linkage (301).This was shown to be converted to natural rifa-mycins and is, therefore, presumably a precur-sor of them, confirming the hypothesis that theether linkage is introduced by interrupting theansa chain after completion of the rifamycincarbon skeleton. Blocked mutants were alsoinvaluable in elucidating a sequence from rifa-mycin W through SV (or S) to B (172).A good system of genetic analysis is available
in the rifamycin-producing strain of N. medi-terranei (258, 259), but, unfortunately, no re-sults of crosses involving blocked mutants havebeen published; thus, we are completely in thedark as to the genetic control of rifamycin bio-synthesis.Actinorhodin. This pigmented antibiotic is a
naphthoquinone half molecules (41). A priori itwas likely to be synthesised by a polyketideroute, and this has been made even more likelyby establishment of such an origin for thenanaomycins ofStreptomyces rosa var. notoen-sis (280), which closely resemble the half mole-cule of actinorhodin.
Actinorhodin was first studied in S. coeli-color strains of the type renamed Streptomycesviolaceoruber (170), whereas strains ofS. coeli-color (Muller) produce a completely differentblue pigment, amylocyanin, and a different(heptaene) antibiotic (H. J. Kutzner et al.,Abstr. 5th Int. Ferment. Symp. 1976, p. 235).Genetic interest in actinorhodin arises from thefact that it is produced in large quantities by S.coelicolor A3(2) (also a strain of the S. violaceo-ruber type; 308). A set of five blocked mutantswas mapped to the chromosome of the strainpresumptively in a gene cluster (308). Thesestudies are being continued by B. A. M. Rudd,who has exploited the acid-base indicator prop-erties of actinorhodin to isolate, by inspection ofcolonies, a large series of blocked mutants.These fall (B. A. M. Rudd, personal communi-cation) into at least seven phenotypic classes onthe basis of differences in precursor or shuntpigments and of cosynthesis tests, and all mapto a region of the chromosome between theclosely linked loci hisD andguaA. Recent isola-tion of an SCP1' strain for this region (Rudd,personal communication) is allowing perform-ance of dominance and complementation tests,while fine mapping ofrepresentative mutationsshould establish the sequence of genes in thecluster. This system, because of the well-devel-oped experimental genetics of S. coelicolorA3(2), promises to be an excellent model foranalysis of the genetic control of synthesis of apolyketide antibiotic. The strong evidence ofgene clustering, taken with the less completeevidence for the tetracyclines, suggests thatclustering may be a feature ofgenes controllingthe biosynthesis of many antibiotics of thisclass.Zorbamycin, etc. Streptomyces bikiniensis
var. zorbonensis was found to produce severalantibiotics, the structures of which appear notto have been published. Three mutations weremapped to a circular chromosomal linkagegroup (57): zorD lacked the ability to synthesizezorbamycin and zorbonomycins B and C,whereas zorA and zorB were also blocked in theproduction of some other antibiotics. In view ofthe scanty information available, few conclu-sions can be drawn from these results; conceiva-bly zorA and zorB were pleiotropic mutationsnot directly involved in antibiotic synthesis
Uust as most bald mutations of S. coelicolorA3(2) fail to produce the chemically and geneti-cally unrelated antibiotics methylenomycinand actinorhodin].Plasmid involvement in antibiotic synthe-
sis. Methylenomycin in S. coelicolor. The anti-microbial activity of this material was first de-scribed by Vivian (293); strains of S. coelicolorA3(2) known by genetic evidence to contain asex factor (SCP1) inhibited the development ofstrains that had lost the plasmid. The inhibitorwas later shown to be a low-molecular-weight,broad-spectrum material-a true antibiotic(164)-which was identified as methylenomy-cin A (307), a compound meanwhile describedin another streptomycete (118). In S. coelicolorA3(2), the ability to produce methylenomycinA, and to be resistant to it, is completely corre-lated with the presence of SCP1, and this is alsotrue of strains of Streptomyces lividans andStreptomyces parvulus, which lack SCP1 but towhich it can be transferred by mating. Moredirect evidence that genes coding for biosyn-thetic enzymes are SCP1 linked comes from theisolation of presumed point mutations (mmy)that fail to produce the antibiotic. All 16 mmymutations so far isolated were SCP1 linked(163). The number of genes represented is un-certain because a reliable complementation testfor mutations on the low-copy-number SCP1plasmid is not yet available, whereas cosyn-thetic reactions between the mutants were rela-tively few, with only one mutant acting as aconverter. From cosynthesis data and someother phenotypic differences between the mu-tants, they were tentatively assigned to fivegroups; two were represented by single muta-tions, suggesting that further groups remain tobe recognized.Chemical studies on the biosynthetic path-
way of methylenomycin A are so far lacking,and the mmy mutations are only now underbiochemical investigation (U. Hornemann andD. A. Hopwood, unpublished data). However,an immediate precursor of methylenomycin Ais apparently a compound postulated on massspectral evidence to be the "des-epoxy" equiva-lent of the antibiotic (307) and is now identifiedas such (Hornemann and Hopwood, unpub-lished data). This compound is found, alongwith methylenomycin A, in SCP1-containingcultures of S. coelicolor A3(2), S. lividans, andS. parvulus (163, 307) and is converted to meth-ylenomycin A by an mmy mutant of S. coeli-color A3(2) but not by SCP1- strains (Horne-mann and Hopwood, unpublished data).
It will be very interesting to know the de-tailed genetic determination of methylenomy-
cin biosynthesis by S. coelicolor A3(2), in par-ticular the number of biosynthetic steps in-volved and at what point the pathway divergesfrom primary metabolic reactions determinedby chromosomal genes. This may be the firstexample, in any organism, of a comparativelyextensive anabolic pathway controlled by plas-mid-borne genes. It would also be extremelyinteresting to know if the corresponding genesare plasmid linked in other methylenomycin-producing streptomycetes; the strain describedby Haneishi et al. (118) has apparently not beenstudied from the genetic point of view.Chloramphenicol in S. venezuelae. Evidence
for plasmid involvement in chloramphenicolbiosynthesis was the failure of chloramphenicol-nonproducing variants (cpp), which arose withhigh frequency after acridine orange treat-ment, to map to a chromosomal linkage groupthat satisfactorily accommodated all auxo-trophic mutations (3). The evidence is certainlysuggestive, although the mapping procedurewas unconventional because recombination fre-quencies were too low to distinguish recombi-nants differing from the parents by singlemarkers from reverse mutations. No physicalevidence for the putative plasmid has been pub-lished, and little information on its role inchloramphenicol biosynthesis is available.There is considerable biochemical evidence fora pathway of synthesis ofchloramphenicol fromshikimic acid (292). Mutants lacking an en-zyme, arylamine synthetase, postulated to actearly in the pathway, have been described (154)in a culture, Streptomyces sp. 3022a, presumedto be a strain of S. venezuelae (L. C. Vining,personal communication). A linkage map forthis strain has been published (94), but no re-sults of crosses involving cpp mutants areavailable. In the S. Venezuela strain, aryla-mine synthetase was lacking in one of the cppmutants presumed to arise by plasmid loss (M.Okanishi, personal communication), but thisresult is, of course, equally compatible with theputative plasmid carrying regulatory ratherthan structural genes for chloramphenicol bio-synthesis.Turimycin in S. hygroscopicus. Evidence for
plasmid involvement in the biosynthesis of thismacrolide was the frequent origin of nonpro-ducing variants, especially after acridine orangeor ethidium bromide treatment (157). Incrosses, the antibiotic production character wastransferred to nonproducing variants at a muchhigher frequency than that of chromosomal re-combination (D. Noack, personal communica-tion). There is some physical evidence for theputative plasmid: antibiotic-producing cultures
yielded different DNA profiles in gradient cen-trifugation compared with nonproducing var-iants (Noack, personal communication).Kasugamycin and aureothricin in S. kasu-
gaensis. A brief description of the frequent ori-gin of strains that failed to produce one or otherof these antibiotics after acridine orange orhigh-temperature treatment (225) provided thefirst suggestion that plasmids might be in-volved in the biosynthesis of streptomycete an-tibiotics, but, unfortunately, no genetic orphysical evidence for the putative plasmid(s)has so far been published. Some DNA differ-ences between the wild type and the nonproduc-ing variants have been found (Okanishi, per-sonal communication), and a full description ofthese results is awaited.
Oxytetracycline in S. rimosus. Althoughmost classes of oxytetracycline-nonproducing(otc) mutations described by Boronin and co-workers mapped to a linkage group presumablyrepresenting the chromosome (see above), therewas a class of nonproducers, also sensitive tothe antibiotic, that arose with high frequencyafter acridine orange treatment (36) andshowed evidence of nonlinkage with chromo-somal markers (A. M. Boronin, A. N. Boriso-glebskaya, and L. G. Sadovnikova, Abstr. 2ndInt. Symp. Genet. Ind. Microorg. 1974, p. 103).Further evidence that such strains arose byplasmid loss is needed, as is clarification oftheir biochemical lesion. It seems possible thatall the structural genes for the biosyntheticpathway are chromosomal but that an extra-chromosomal element is involved in the regula-tion of antibiotic biosynthesis or possibly ex-port, a situation that may have a counterpartin the tyrosinase system of Streptomyces glau-cescens (26). It is suggestive that a class of otcvariant corresponding to the presumed extra-chromosomal class was not found in anotherstrain ofS. rimosus by a second group of work-ers (J. Pigac, M. Vesligaj, and V. Delic, Abstr.2nd Int. Symp. Genet. Ind. Microorg. 1974, p.105).A cautionary note. Not all phenotypes aris-
ing too frequently to be easily reconciled withgene mutation and failing to map to the chro-mosome need necessarily be due to plasmidloss. In S. coelicolor A3(2), chloramphenicolsensitivity (Cmls) had such characteristics, butthe powerful selection available for reversion tochloramphenicol resistance (Cmlr) allowed anoscillation between Cmls and Cmlr to be demon-strated with no permanent loss of genetic infor-mation (95). Moreover, in contrast to the expec-tation for a self-transmissible plasmid, therewas no infectious transfer of Cmlr into Cml5
strains in crosses at frequencies greatly in ex-cess of chromosomal marker transfer. This ex-ample, which is perhaps interpretable in termsof some kind of transposition of genetic infor-mation rather than its permanent loss, servesas a reminder that acceptable genetic and/orphysical evidence is needed before a case ofunstable antibiotic production is interpreted asnecessarily due to plasmid loss.
EubacteriaPeptide antibiotics of bacilli. Biochemical
research has led to an understanding of thesynthesis ofseveral ofthese antibiotics, notablythe cyclic decapeptides gramicidin S and tyroci-dine of B. brevis (171, 180) and the cyclic do-decapeptide bacitracin A of B. licheniformis(99). Nonribosomal peptide bond formation iscatalyzed by an antibiotic "synthetase" consist-ing of two or three subunits: "light" and "heavy"for gramicidin S (which may be considered asimpler molecule than the other two becauseit consists of two identical pentapeptides joinedhead to tail); "light," "intermediate," and"heavy" for tyrocidine, and A, B, and C forbacitracin. Peptide chain synthesis proceeds bysequential addition of amino acid residues to agrowing chain, initiated at the N-terminal end.Sequence information is provided by the bind-ing, through covalent linkage to enzyme -SHgroups (thioester linkage), of each amino acidresidue at a spatially distinct site on one oranother of the enzyme subunits. Each subunitapparently contains a separate protein for eachamino acid bound - one for the smallest sub-units of gramicidin S and tyrocidine synthe-tases, which bind only phenylalanine, six forthe largest component of tyrocidine synthetase,which binds six amino acids, etc. The growingpeptide chain is handed on from one site to theamino acid bound at the next on the same or anadjacent subunit as appropriate. For certainphenylalanine residues of the two B. brevisantibiotics, racemization to the 1-isomer alsooccurs at the binding site. Few details of cycli-zation of the completed peptides of tyrocidineand bacitracin and of the head-to-tail joining ofthe two pentapeptides of gramicidin S appear tobe available. Nor is there information on thesynthesis and attachment ofthe thiazoline ringpresent in bacitracin A.These systems should provide fascinating ob-
jects for biochemical-genetic studies. The mini-mum number of gene products can be esti-mated. For example, for tyrocidine the threesubunits contain one, three, and six proteinsthat bind amino acids, and in addition the twolarger subunits have a small binding protein
for phosphopantotheine, which plays a key rolein peptide bond formation at each step in thesynthesis (171, 180) (this is probably also truefor bacitracin [247]), making about 12 geneproducts. Biochemical analysis of blocked mu-tants is proceeding, particularly for gramicidinS. Two Japanese groups have described seriesof such mutants. One group (158) classified fivemutants into three functional classes by in vi-tro complementation using preparations of thetwo fractions of the synthetase. (Their fraction Iis the "heavy" subunit referred to above, whichactivates four amino acids; fraction II is the"light" subunit, which activates and racemizesphenylalanine.) The three classes of mutantslacked the activity of one, the other, or bothfractions. The other group (149) classified 20mutants into three similar classes; later (269)these 20 mutants were tested for the ability oftheir fractions I and II to activate proline, va-line, ornithine, and leucine (fraction I) or phen-ylalanine (fraction II), and five classes werefound. In one, activation of all the amino acidsoccurred, whereas in another none was acti-vated. A third class failed to activate phenylal-anine and presumably had a lesion in fractionII. The other two classes were defective in frac-tion I function, failing to activate proline, va-line, ornithine, and leucine or alternativelyfailing to activate only one of them (proline,valine, or leucine); this latter group presum-ably represented at least three classes of lesionin three different proteins of the fraction I sub-unit. The possibility- of analyzing mutants ofthe cyclic peptide antibiotics functionally bythese in vitro techniques makes the systemvery attractive, and it is disappointing that nosystem of genetic analysis is available in any ofthe producing strains.
Butirosin. This is a mixture oftwo aminogly-cosides produced by some strains of Bacilluscirculars. The technique of mutational biosyn-thesis (see above) was used with mutantsblocked in the synthesis of the deoxystrepta-mine moiety (55, 282) by production of unnatu-ral antibiotics containing streptamine, strepti-dine, or dideoxystreptamine. No genetic studieson the mutants have been reported. They wouldbe of interest; could it be that the unique pro-duction of aminoglycosides by certain strainsonly of this eubacterial species indicates plas-mid involvement?
Nisin. This linear polypeptide of 34 aminoacids (116) is produced by Streptococcus lactisstrains. Circumstantial evidence for plasmidinvolvement was the finding that certainstrains gave up to 1% nisin-nonproducing cul-tures spontaneously and at very high frequency
after acridine orange, ethidium bromide, orhigh-temperature treatment (179). Certain ofthe nisin-negative derivatives lacked plasmidDNA present in their parent strains, but othersdid not (100). Thus, it is very likely that autono-mous plasmids are involved in nisin productionat least in some strains of S. lactis, and theresults of further studies are eagerly awaited.
Prodigiosin. This pigment made by Serratiamarcescens is probably the only antibiotic pro-duced by a member of the Enterobacteriaceaethat has been studied extensively (304). Largenumbers of blocked mutants were isolated, andtheir study provided evidence for several fea-tures of the postulated biosynthetic pathwayincluding its synthesis from two heterocyclicmoieties, since reciprocal cosynthesis was ob-served between pairs of mutants, one blockedin the synthesis of each presumptive moiety.No recombination studies have, however, beenmade.
QUANTITATIVE GENETICS OFANTIBIOTIC PRODUCTION
General Principles of Polygenic Control ofQuantitative Characters
Many characters in all organisms show con-tinuous variation; individuals from naturalpopulations, progeny from a cross, or survivorsof mutagenesis often form a continuous rangeof types inseparable into discrete classes. Suchcontinuous variation is considered to resultfrom the joint action of a number of "poly-genes," each having a small effect in relation tothe total genotypic variation. When these ge-netic effects are combined with the blurringeffect of environmental influences, differencesdue to individual genes cannot be discerned bymere inspection of the population; but poly-genes can be seen to segregate and obey thesame Mendelian laws as the more readily iden-tifiable 'major" genes when analyzed by thetechniques of"biometrical genetics." This is notan alternative to Mendelian genetics but a de-velopment of it that aims to determine thekinds ofgene action and interaction involved inquantitative variation and to predict the out-come of future generations and the conse-quences of selection (85, 203, 204).
Biometrical genetics differs from Mendeliangenetics in its analytical methods. Each gener-ation comprises a continuous range of pheno-types, and, therefore, the properties of genera-tions and the relationships between them aredescribed statistically by means and variancesas well as by correlations and covariances, re-spectively. By comparing these statistics withmodel expectations, the observed means and
correlations (first-degree statistics) and vari-ances and covariances (second-degree statistics)can be interpreted in terms of and partitionedinto effects due to additive gene action, domi-nance, epistasis, linkage, environment, and in-teractions between genotype and environment.
Application of Biometrical Methods toHaploid and Nonmeiotic Systems
Microbial genetics have largely dealt withmajor gene differences where the inheritance ofcharacters can be unambiguously analyzedqualitatively. Complications due to quantita-tive variation have been minimized by workingmainly with closely related strains. The micro-bial geneticist investigating antibiotic produc-tion can usually choose neither the organismnot the character for study and often cannotafford to ignore variation that is not discreteand that occurs in unrelated strains. Further-more, antibiotic titer is often very sensitive toenvironmental variables so that even if theunderlying genetic determination is relativelysimple, the inherent discontinuities may be ob-scured by environmentally induced variation.Therefore, particularly when concerned withstrain improvement for antibiotic production,biometrical genetics has a part to play alongwith the more widely recognized techniques ofmutagenesis, recombination, and biochemicalgenetics.Very little biometrical genetics has been
done with microorganisms despite their manytechnical advantages. This may be due to thehistorical development of biometrical geneticsin relation to plant and animal breeding and,hence, to diploid meiotic systems. An early at-tempt to analyze the polygenic nature of mul-tistep chloramphenicol resistance in E. coli (54)has not been followed up, although geneticanalysis in this organism is now much morerefined. Recently, however, microorganismshave been used in the assessment of certainbiometrical techniques (97, 98, 153), and thepossibility of using large populations and ofstrictly defining environmental variables areimportant advantages.The use of biometrical techniques to analyze
quantitative variation in microorganisms wasconsidered in detail by Caten and Jinks (53).The classical biometrical models for generationmeans and variances require complete diploidformation, unbiased segregation, and recoveryof all possible genotypes in proportions deter-mined by their linkage relationships; these re-quirements are clearly not satisfied in manymicrobial systems. However, classical biomet-rical methods can easily be applied to those eu-
caryotic microorganisms with a meiotic system,and the predominantly haploid life cycle simpli-fies analysis considerably since the complica-tions of dominance do not occur. The applica-tion of biometrical techniques to a meiotic hap-loid system is exemplified below: the inherit-ance of penicillin production in Aspergillusnidulans.Nonmeiotic systems. In all nonmeiotic sys-
tems, selection must be used to separate recom-binant progeny from parentals. Normal bio-metrical methods do not apply because esti-mates such as those ofthe genetic component ofvariation are biased because all progeny carrya selected portion of the genome and are not arandom sample of all possible genotypes. Com-parisons of sexual and parasexual progeny inA. nidulans (53) showed considerable segrega-tion and reassociation of polygenes during theparasexual cycle, but although differences be-tween the two types of progeny sample wereobserved these may well have been due to selec-tive bias in the parasexual analysis. Such biascould possibly be overcome by using more thanone type of selection and reconstituting a ran-dom sample of genotypes from the resultantprogeny.Although nonmeiotic systems are not amena-
ble to typical biometrical analysis, they couldlend themselves to the study of quantitativevariation by "chromosome assay" techniques.This involves the substitution of part of thegenome, e.g., a chromosome or chromosomesegment, from a selected strain into a testerstrain while the rest of the genome is eitherheld constant or allowed to vary randomly. Ifsuch substitution has a significant effect, thena gene or genes affecting the character are lo-cated in the tested region (202). A linkage mapand some suitable selective markers are essen-tial, and interpretation of the results may becomplicated if the markers have pleiotropic ef-fects on the character as is often found for anti-biotic titer (89, 208). Nevertheless, many as-pects of nonmeiotic systems are advantageousfor this type of analysis. For example, the rar-ity of mitotic crossing-over in the parasexualcycle (20, 178, 237) means that whole chromo-somes could be assayed by haploidization anal-ysis without any significant intrachromosomalrecombination. Such analysis could be particu-larly useful in strains in which antibiotic titerhad been increased by mutagenesis, but its usein the study of natural quantitative variationin Aspergillus species has, unfortunately, beenprevented by the heterokaryon incompatibilityof unrelated isolates (52).There are plenty ofexamples from strain im-
provement programs with antibiotic-producing
actinomycetes of the stepwise increase in anti-biotic titer over many rounds of mutation andselection (8). This in itself confirms the poly-genic control of antibiotic titer. There are alsosome data from Alikhanian's group on the re-sults of crosses between S. rimosus strains thatindicate considerable variation in titer amongthe progeny of crosses even between closelyrelated strains and more when wider crosseswere made (8). However, the data were notanalyzed by the techniques of biometrical ge-netics; nor was the segregation of individualgenes affecting titer recognized (except thepleiotropic effects of the markers). These stud-ies will therefore not be considered further.
Biometrical Genetics of PenicillinProduction in A. Nidulans
Wild isolates ofA. nidulans have been usedin several studies ofnatural quantitative varia-tion (46, 47, 152), and examples have beentaken from studies of quantitative variation inpenicillin production to demonstrate the type ofinformation obtainable by biometrical analysis(for a more detailed account see references 208-210, 212). Holt and Macdonald (128) demon-strated that in three crosses between A. nidu-lans strains (one between two high-titer wildtypes and two between a wild type and a high-titer mutant) recombinants with substantiallyincreased penicillin titers could be obtained.The distribution of titers among the recombi-nants indicated that penicillin production wasprobably under polygenic control, and thisprompted a more detailed study of the amountof variation for penicillin titer present in natu-ral populations. A random sample of wild iso-lates had a range of titers from 0.0 to 14.4 U/mlwith a mean of 8.0 + 0.4 U/ml (212). Withsuitable experimental designs and methods ofanalysis, the total phenotypic variation in anexperiment may be partitioned into environ-mental (aE2) and genetic (OrG2) components (53).About 60% of the observed variation amongwild isolates was attributable to genetic differ-ences between the isolates (212), indicatingthat there is a significant amount of heritablevariation for penicillin titer in natural popula-tions ofA. nidulans.
In A. nidulans, heterokaryon formation be-tween pairs of isolates is restricted to membersof each heterokaryon compatibility (h-c) group(115), but sexual outcrossing is possible be-tween all wild isolates to a greater or lesserextent regardless of h-c status (45). Thirty-oneof the isolates tested had been allocated to h-cgroups, and 67% of the variation in yield be-tween isolates was due to differences betweenh-c groups (212).
To confirm genetic control of titer differencesbetween wild isolates, crosses were made be-tween representative isolates, and the titers ofa sample of single ascospore cultures from eachcross were determined. Such a cross can be usedto investigate a number of factors, discussed indetail in reference 53. Estimates of heritabilityvalues (which indicate the proportion of thetotal observed variation fixable by selection[531) varied between 0.0 and 0.77 in sevencrosses between wild isolates (212). Of particu-lar interest was the observation that crossesbetween isolates from different h-c groups had amuch higher mean value for aG2 (7.72) thancrosses between isolates from the same h-cgroup (3.46). This reinforced the conclusions(47, 152) that "isolates from the same h-c groupshave similar genotypes and aided the choice ofparental isolates for crosses to generate maxi-mal genetic variation among the progeny.The data can also be used to determine
whether the segregating penicillin genes actadditively or nonadditively. In a diploid orga-nism, nonadditive variation can be due toeither allelic interaction (dominance) or nonal-lelic interaction (epistasis), but in haploids allnonadditive variation must be due to epistasis.In A. nidulans, the gene action was predomi-nantly additive, giving a symmetrical distribu-tion of the progeny titers about a value close tothe parental mean (212).Knowledge of the type of gene action is par-
ticularly important when choosing a crossingscheme to select for increased titer. Withmainly additive effects, it is simply necessaryto combine the maximum number of "increas-ing" alleles into a single genotype, which can bereadily achieved by a program of line selection(208, 209). If, instead, much ofthe gene action isepistatic, opportunity for free recombinationshould be provided from the outset of the pro-gram by maintaining a single breeding popula-tion on which selection is practiced.The influence of genetic markers on anti-
biotic production can also be examined biomet-rically. In crosses with A. nidulans, progenywith yellow conidia were often found to have alower mean titer than those with green, and itwas shown that pleiotropy of the y mutationrather than linkage of this allele to genes fordecreased titer was the causative factor (208).On the other hand, linkage of polygenes toknown chromosomal markers has been demon-strated by biometrical analysis in Schizophyl-lum commune (58), Neurospora crassa (230),and Aspergillus amstelodami (53).Having shown that significant genetic varia-
tion can be released by crosses, one can attempt
to recombine such variation to generate strainswith improved titers. In A. nidulans, four se-lection lines were initiated, each from a crossbetween two wild isolates. Each parental strainwas chosen for high titer and for heterokaryonincompatibility with the other seven strains soas to maximize the available genetic variationin the initial gene pool. During selection forthree to five generations, the mean progenytiter approximately doubled in each line. Theavailable genetic variation was significantlyreduced in all four lines by generation 4 or 5(208). Such a response is expected in a haploidorganism when, as in this case, a scheme ofsibling mating is used. But the expectation wasthat if different genes for increased titer hadbeen accumulated in the different selectionlines, further genetic variation should havebeen releasable by crossing selected strainsfrom different lines.
This was confirmed by crosses between repre-sentatives from three of the A. nidulans selec-tion lines, and in each case a substantialamount of genetic variation, equal to that inthe original crosses between wild isolates, wasreleased (209). Moreover, the gene action ininterline crosses remained additive so that newhigh-titer genotypes were generated. It wasthen possible to use further rounds of crossingand selection to obtain further increases in ti-ter. By introducing new sources of genetic vari-ation when necessary, such stepwise increasesin titer might be continued over several cycles.The biometrical methods applied here to a
system of crossing and selection based on natu-ral variation are equally applicable to onebased on mutationally induced variation, andsuch an analysis has recently been carried out(I. Simpson and C. E. Caten, personal commu-nication). Mutation and selection were used toincrease penicillin production by A. nidulansin two independent selection lines, one treatedwith 8-methoxypsoralen plus near-UV and theother with ethylmethanesulfonate. Six cycles ofselection resulted in increases in titer from 5 to15 U/ml and from 7 to 18 U/ml. The inducedtiter-increasing mutations either were reces-sive or showed ambidirectional dominance indiploids. Biometrical analysis suggested thatthe increased titers were probably due to muta-tions in several genes and that the mutations inthe two selection lines were not allelic. Crossesbetween the selection lines showed slight epi-stasis, with a progeny mean slightly lower thanthe midparental value, but further cycles ofselection and hybridization initiated from thiscross produced recombinant strains with im-proved titers over 20 U/ml.
Hence, in this system there appear to be a noinherent differences between naturally occur-ring and mutagenically induced variation, andboth can be successfully analyzed by biometri-cal techniques. Even in schemes where hybridi-zation is not employed, the use of biometricalmethods to monitor available genetic variationwould be of considerable value both in compar-ing the efficiency of different mutagenic treat-ments and in assessing the amount of variationavailable for selection.
Quantitative Mutations Affecting PenicillinYield
All commerical strains of P. chrysogenumderive from the Wisconsin family, the result ofa vast mutation program at the Botany Depart-ment of Wisconsin University between 1946and 1956, which gave at least a fivefold im-provement in penicillin titer (16). Further largeyield increments occurred later in commercialstrain development programs (83). The geneticchanges that led to these increments have beenstudied extensively.Heterokaryons between high- and low-titer
strains produced segregants in which penicillintiter was associated with parental chromosomalmarkers (190, 262), indicating the yield muta-tions to be nuclear rather than cytoplasmic.Diploids between high-titer strains and theirprogenitors invariably had titers similar tothose of the low-yielding strain (261). Thus pos-itive mutations, including such significantevents as that differentiating strains WisQ176and X1612, were all recessive. Diploids be-tween a wild-type strain (NRRL 1951) and twoderivatives selected after a number of cycles ofmutagenesis with several mutagens to produceabout 3,000 U more than the wild type all hadtiters similar to or only slightly greater thanthe wild type. Hence, again, the positive muta-tions were almost completely recessive (190).
Diploids between strains carrying independ-ent positive mutations usually had yields equalor inferior to that of the less productive parent(192, 263, 264). If all the mutations had beenrecessive and nonallelic, the diploids shouldhave had a yield similar to that of the commonancestor; thus, many of the mutations wereallelic (192). However, one diploid betweenstrains Wis49-133 and Wis50-1247 had a consist-ently higher yield than the component strains(262). This result is, of course, difficult to ex-plain unless dominant mutations for increasedyield were involved, and a satisfactory explana-tion has not been put forward.Large variations in titer occurred among dip-
loids between the same two high-titer parents
(184, 192). The yields of 68 diploids, all synthe-sized from the same two parents, yielding 3,000and 3,500 U, fell into two distinct groups; 44yielded around 600 U/ml and 24 about 3,000 U/ml. Two explanations were proposed. In one(184), the high-titer diploids were normal andthose of low titer arose by dominant mutationsreducing yield and conferring a selective ad-vantage on the resultant low-titer strains. Thishypothesis predicted that variability shouldalso arise in diploids homozygous for penicillinmutations, but a study of such diploids (184)was inconclusive. In an alternative hypothesis(264), the high-titer diploids arose from those oflow titer by segregation, which would have hadto be transitory and had to occur soon afterdiploidization since isolated low-yielding dip-loids did not spontaneously segregate high-titerdiploids. Low-titer diploids did produce high-titer, first-order haploid segregants (184). Thisinterpretation requires some lack of allelismbetween positive mutations. Conclusive proofofeither hypothesis would be difficult to obtain.Any unusual segregation data may reflect
that many ifnot all improved strains may carrychromosome aberrations arising from the mu-tagenic treatments used in selection programs.Such aberrations have not been conclusivelydemonstrated, but several pieces of evidenceindicate that they are common in improvedstrains. Genetic analysis in A. nidulans dem-onstrated that chromosome aberrations can beinduced at a significant frequency by UV andX-irradiation (17, 156), and the spontaneousproduction of a variety of morphological types(termed "population patterns" [276]) by manyimproved strains ofP. chrysogenum is reminis-cent of the morphological effects ofchromosometranslocations or duplications in A. nidulans(156, 288). Diploids between pairs of high-yield-ing strains invariably produced a majority ofhaploid segregants ofone or other parental phe-notype (192, 193). This phenomenon, termed"parental genome segregation," (192) is proba-bly due to differences in chromosome structurebetween the parent strains (185).
Haploid segregants from a heterozygous dip-loid never had a significantly higher titer thanboth parent strains (192). However, haploidswith titers significantly lower than both par-ents often arose presumptively by recombina-tion of deleterious recessive mutations fromboth parents (263).The studies described so far dealt with
strains carrying one or more unidentified muta-tions causing increased penicillin yield. Morerecently, experiments have been carried outwith P. chrysogenum andA. nidulans in which
single titer-increasing mutations were mapped,and the effects of recombining such mutationsinto a single strain were studied.P. chrysogenum. Some of the earlier prob-
lems of parasexual analysis in P. chryso-genum-parental genome segregation and pos-sible selection against certain segregants (185,192, 193, 262)- were overcome by using sisterstrains differing in only a few mutational steps(20). After haploidization by PFA treatment, 21independently isolated spore color and auxo-trophic markers were allocated to one or otherof three "haploidization groups." Strains carry-ing these markers were then used as parents inwhich single-step titer increases were inducedby UV. The parents had titers of ca. 3,000 U/ml,and the mutations caused increases of 500 to2,000 U/ml. Attempts were made to allocatethese mutations to haploidization groups bycrossing with other marked strains. Manysegregants had titers considerably lower thanthe parents. Free recombination was observedbetween haploidization groups, but this recom-bination apparently disrupted the balancedpolygenic systems responsible for the hightiters in the parent strains. Unfortunatelythese effects, combined with the very low num-bers of segregants scored (sometimes less than10), made the unambiguous allocation of titer-increasing mutations to haploidization groupsrather difficult.On the basis of tentative allocations, strains
carrying different mutations were crossed in anattempt to obtain segregants carrying both mu-tations (21, 22); two crosses produced such se-gregants with significantly increased titers.The effects of the two mutations were com-pletely additive, a promising result for the fu-ture application of recombination to titer im-provement. Segregants with much lower titersalso arose, but the authors did not comment onthe stability of the high-titer recombinants.A. nidulans. Three mutations causing in-
creased penicillin titer in A. nidulans, penAl,penB2, and penC3, had titers of 20, 12, and 20U/ml, respectively, compared with 6 U/ml forthe parental strain (NRRL 194) (127, 188). Eachmutation was mapped to its chromosome byparasexual. haploidization analysis (207):penAl, penB2, and penC3 to chromosomesVIII, III, and IV, respectively. They weremapped relative to other loci on the chromo-some by cleistothecial analysis (76). PenAl wastentatively placed equidistant between chaAand nirA at the distal end of the right arm ofchromosome VIII. Strains carryingpenB2 weremorphologically abnormal, and genetic analy-sis indicated that this might have been a pleio-tropic effect of penB2, which was located 13
map units from moC on the left arm of chromo-some III. Strains carrying penC3 were appar-ently infertile, but hybrid perithecia were ob-tained by crossing with a strain carrying themutation sgpC. Crosses between strains bear-ing allelic sgp mutations are sexually sterile,but those between nonallelic sgp mutations arefertile (148), and penC3 appears, therefore, tomimic the behavior of sgp mutations. By cross-ing a strain carrying penC3 with one carryingan sgp mutation, fertility was achieved, andpenC3 was shown to be linked to pyroA on theright arm of chromosome IV.The titer of heterozygous diploids between
each of the three mutations and a wild-typestrain suggested that penAl was recessive,penB2 dominant, and penC3 semidominant(76). Recessivity of penAl was confirmed in apenAl penB2 (trans) heterozygous diploid,which had a titer similar to that of the penB2parent. A penB2 penC3 (trans) heterozygousdiploid also had the same titer as the penB2parent, and no additive effect of the two muta-tions was found. The titers of haploid recombi-nants carrying two ofthe three mutations in allthree combinations indicated that penAl wasepistatic to penB2 and penC3 and that penC3was epistatic to penB2. The lack of additivitybetween titer-increasing mutations in thisstudy contrasts with the marked additivityfound in P. chrysogenum (22) and in studies ofnatural variation in titer in A. nidulans (seeabove).
CONCLUSIONThe Application of Genetic Knowledge
This review has highlighted the discrepancybetween the amount of knowledge of the ge-netic control of antibiotic production so far ob-tained and the potential for obtaining suchknowledge. A similar gulf separates the poten-tial for harnessing of genetics in the develop-ment of new industrial strains and the use thathas been made of genetics in such endeavors.The paucity of knowledge accounts, in part, forthe dearth of applications. However, especiallywith the opening of the era of genetic engineer-ing, strain improvement will increasingly de-mand answers to questions about the number,arrangement, and roles of genes determiningantibiotic production if the new techniques areto be used effectively (138). We can expect,therefore, a big increase in research in this areain the near future.
ACKNOWLEDGMENTSWe are grateful to all those, mentioned in the
text, who kindly supplied us with their unpublisheddata. We should particularly like to acknowledge
Georgina Moore for expert secretarial assistance inthe preparation of this article.
LITERATURE CITED
1. Achenbach, H. 1967. Xanthocillin, p. 26-28. InD. Gottlieb and P. D. Shaw (ed.), Antibiot-ics: biosynthesis, vol. 2. Springer-Verlag,Berlin, Heidelberg, and New York.
2. Adams, J. N., and G. H. Brownell. 1976. Ge-netic studies in Nocardia erythropolis, p.285-309. In M. Goodfellow, G. H. Brownell,and J. A. Serrano (ed.), The biology of thenocardiae. Academic Press, London, NewYork, and San Francisco.
3. Akagawa, H., M. Okanishi, and H. Umezawa.1975. A plasmid involved in chloramphenicolproduction in Streptomyces venezuelae: evi-dence from genetic mapping. J. Gen. Micro-biol. 90:336-346.
4. Alacevic, M. 1963. Interspecific recombinationin Streptomyces. Nature (London) 197:1323.
5. Alakevic, M. 1976. Recent advances in Strepto-myces rimosus genetics, p. 513-519. In K. D.Macdonald (ed.), Second International Sym-posium on the Genetics of Industrial Micro-organisms. Academic Press, London, NewYork, and San Francisco.
6. Alacevic, M., M. Straiek-Vesligaj, and G. Ser-monti. 1973. The circular linkage map ofStreptomyces rimosus. J. Gen. Microbiol.77:173-185.
7. Alakevic, M., D. Vlasic, and I. Spada-Ser-monti. 1966. Interspecific recombinationamong antibiotic-producing streptomycetes,p. 720-722. In M. Herold and D. Zdenek (ed.),Antibiotics: advances in research, produc-tion and clinical use. Butterworths, London.
8. Alikhanian, S. 1970. Applied aspects of micro-bial genetics. Curr. Top. Microbiol. and Im-munol. 53:91-148.
9. Alikhanian, S. I., and L. N. Borisova. 1961.Recombination in Actinomyces aureofaciens.J. Gen. Microbiol. 26:19-28.
10. Alikhanian, S. I., T. S. Ilyina, and N. D. Lo-movskaya. 1960. Transduction in actinomy-cetes. Nature (London) 188:245-246.
11. Alikhanian, S. I., N. D. Lomovskaya, and V.N. Danilenko. 1976. Suppressor-sensitivemutations of Streptomyces coelicolor A3(2)and actinophage qC31, p. 595-606. In K. D.Macdonald (ed.), Second International Sym-posium on the Genetics of Industrial Micro-organisms. Academic Press, London, NewYork, and San Francisco.
12. Anne, J., H. Eyssen, and P. De Somer. 1976.Somatic hybridisation of Penicillium roque-fortii with P. chrysogenum after protoplastfusion. Nature (London) 262: 719-721.
13. Anne, J., and J. F. Peberdy. 1976. Inducedfusion of fungal protoplasts following treat-ment with polyethylene glycol. J. Gen. Mi-crobiol. 92:413-417.
14. Arnstein, H. R. V., D. Morris, and E. J. Toms.1959. Isolation of a tripeptide containing a-
aminoadipic acid from the mycelium ofPeni-
cillium chrysogenum and its possible signifi-cance in penicillin biosynthesis. Biochim.Biophys. Acta 35:561-562.
15. Audit, C., and C. Anagnostopoulos. 1973. Ge-netic studies relating to the production oftransformed clones diploid in the tryptophanregion of the Bacillus subtilis genome. J.Bacteriol. 114:18-27.
16. Backus, M. P., and J. F. Stauffer. 1955. Theproduction and selection of a family ofstrains in Penicillium chrysogenum. Mycolo-gia 47:429-463.
17. Bainbridge, B. W., and J. A. Roper. 1966. Ob-servations on the effects of chromosome du-plication in Aspergillus nidulans. J. Gen.Microbiol. 42:417-424.
18. Bal, J., E. Bartnik, B. Goryluk, and N. J.Pieniazek. 1975. An easy way of obtainingAspergillus nidulans haploids in the para-sexual cycle using N-glycosyl polifungin. Ge-net. Res. 25:249-252.
19. Balassa, G. 1969. Biochemical genetics of bac-terial sporulation. I. Unidirectional pleio-tropic interactions among genes controllingsporulation in Bacillus subtilis. Mol. Gen.Genet. 104:73-103.
20. Ball, C. 1971. Haploidisation analysis in Peni-cillium chrysogenum. J. Gen. Microbiol.66:63-69.
21. Ball, C. 1973. Improvement of penicillin pro-ductivity in Penicillium chrysogenum by re-combination, p. 227-237. In Z. Vanek, Z.Hoiailek, and J. Cudlin (ed.), Genetics ofindustrial microorganisms: actinomycetesand fungi. Academia, Prague.
22. Ball, C., and J. L. Azevedo. 1976. Genetic in-stability in parasexual fungi, p. 243-251. InK. D. Macdonald (ed.), Second InternationalSymposium on the Genetics of Industrial Mi-croorganisms. Academic Press, London, NewYork, and San Francisco.
23. Bardone, M. R., E. Martinelli, L. F. Zerilli,and C. Coronelli. 1974. Structure determi-nation of purpuromycin, a new antibiotic.Tetrahedron 30:2747-2754.
24. Batchelor, F. R., F. P. Doyle, J. H. C. Nayler,and G. N. Rolinson. 1959. Synthesis of peni-cillin: 6-aminopenicillanic acid in penicillinfermentation. Nature (London) 183:257-258.
25. Baumann, R., R. Hotter, and D. A. Hopwood.1974. Genetic analysis in a melanin-produc-ing streptomycete, Streptomyces glauces-cens. J. Gen. Microbiol. 81:463-474.
26. Baumann, R., and H. P. Kocher. 1976. Ge-netics of Streptomyces glaucescens and regu-lation of melanin production, p. 535-552. InK. D. Macdonald (ed.), Second InternationalSymposium on the Genetics of Industrial Mi-croorganisms. Academic Press, London, NewYork, and San Francisco.
27. Berdy, J. 1974. Recent developments of anti-biotic research and classification of antibiot-ics according to chemical structure. Adv.Appl. Microbiol. 18:309-406.
28. Beretta, M., M. Betti, and M. Polsinelli. 1971.Genetic recombination in micromonospora.
J. Bacteriol. 107:415-419.29. Bibb, M. J., R. F. Freeman, and D. A. Hop-
wood. 1977. Physical and genetical charac-terisation of a second sex factor, SCP2, forStreptomyces coelicolor. Mol. Gen. Genet.154:155-166.
30. Birch, A. J. 1967. Biosynthesis of polyketidesand related compounds. Science 156:202-206.
31. Birch, A. J. 1967. Fumagillin, p. 152-153. In D.Gottlieb and P. D. Shaw (ed.), Antibiotics:biosynthesis, vol. 2. Springer-Verlag, Ber-lin, Heidelberg, and New York.
32. Blumauerova, M., 2. Hodialek, and Z. Vankk.1972. Biosynthesis of tetracyclines: problemsand perspectives in genetic analysis, p. 223-232. In G. Terni (ed.), Fermentation technol-ogy today. Osaka Society of FermentationTechnology, Osaka.
33. Bonner, D. 1947. Studies on the biosynthesis ofpenicillin. Arch. Biochem. 13:1-9.
34. Bordet, C., M. Karahjoli, 0. Gateau, and G.Michel. 1972. Cell walls of nocardiae andrelated actinomycetes: identification of thegenus Nocardia by cell wall analysis. Int. J.Syst. Bacteriol. 22:251-259.
35. Boronin, A. M., and S. Z. Mindlin. 1971. Ge-netic analysis of Actinomyces rimosus mu-tants with impaired synthesis of antibiotic(in Russian). Genetika 7:125-131.
36. Boronin, A. M., and L. G. Sadovnikova. 1972.Elimination by acridine dyes of oxytetra-cycline resistance inActinomyces rimosus (inRussian) Genetika 8:174-176.
37. Bousfield, I. J., and M. Goodfellow. 1976. The"Rhodochrous" complex and its relationshipswith allied taxa, p. 39-65. In M. Goodfellow,G. H. Brownell, and J. A. Serrano (ed.), Thebiology of the nocardiae. Academic Press,London, New York, and San Francisco.
38. Bradley, D. E. 1967. Ultrastructure of bacteri-ophages and bacteriocins. Bacteriol. Rev.31:230-314.
39. Bradley, S. G., and J. Lederberg. 1956. Hetero-karyosis in Streptomyces. J. Bacteriol.72:219-225.
40. Braendle, D. H., and W. Szybalski. 1959. Het-erokaryotic compatibility' metabolic cooper-ation, and genic recombination in Strepto-myces. Ann. N.Y. Acad. Sci. 81:824-851.
41. Brockmann, H., A. Zeeck, K. van der Merwe,and W. Muller. 1966. Ober Actinomyceten-farbstoffe, VIII. Die Konstitution des Anti-norhodins. Liebigs Ann. Chem. 698:209-229.
42. Bu'lock, J. D. 1961. Intermediary metabolismand antibiotic synthesis. Adv. Appl. Micro-biol. 3:293-342.
43. Bu'lock, J. D. 1975. Secondary metabolism infungi and its relationships to growth anddevelopment, p. 33-58. In J. E. Smith, andD. R. Berry (ed.), The filamentous fungi:industrial mycology, vol. 1. Edward Arnold,London.
44. Bu'lock, J. D., D. Hamilton, M. A. Hulme, A.J. Powell, H. M. Smalley, D. Shepherd, andG. N. Smith. 1965. Metabolic development
and secondary biosynthesis in Penicilliumurticae. Can. J. Microbiol. 11:765-778.
45. Butcher, A. C. 1968. The relationship betweensexual outcrossing and heterokaryon incom-patibility in Aspergillus nidulans. Heredity23:443-452.
46. Butcher, A. C. 1969. Non-allelic interactionsand genetic isolation in wild populations ofAspergillus nidulans. Heredity 24:621-631.
47. Butcher, A. C., J. Croft, and M. Grindle. 1972.Use of genotype-environmental interactionanalysis in the study of natural populationsof Aspergillus nidulans. Heredity 29:263-283.
48. Caglioti, M. T., and G. Sermonti. 1956. A studyof the genetics of penicillin-producing capac-ity in Penicillium chrysogenum. J. Gen. Mi-crobiol. 14:38-46.
49. Calam, C. T., L. B. Daglish, and W. S. Gait-skell. 1973. Hybridisation experiments withPenicillium patulum and Fusarium monili-forme, p. 265-282. In Z. Vanek, Z. Hostalek,and J. Cudlin (ed.), Genetics of industrialmicroorganisms: actinomycetes and fungi.Academia, Prague.
50. Caltrider, P. G., and H. F. Niss. 1966. Role ofmethionine in cephalosporin synthesis.Appl. Microbiol. 14:746-753.
51. Carere, A., and R. Randazzo. 1976. Co-muta-tion in Streptomyces, p. 573-581. In K. D.Macdonald (ed.), Second International Sym-posium on the Genetics of Industrial Micro-organisms. Academic Press, London, NewYork, and San Francisco.
52. Caten, C. E., and J. L. Jinks. 1966. Heterokar-yosis: its significance in wild homothallic as-comycetes and fungi imperfecti. Trans. Br.Mycol. Soc. 49:81-93.
53. Caten, C. E., and J. L. Jinks. 1976. Quantita-tive genetics, p. 93-111. In K. D. Macdonald(ed.), Second International Symposium onthe Genetics of Industrial Microorganisms.Academic Press, London, New York, andSan Francisco.
54. Cavalli, L. L., and G. A. Maccacaro. 1952.Polygenic inheritance of drug-resistance inthe bacterium Escherichia coli. Heredity6:311-331.
55. Claridge, C. A., J. A. Bush, M. D. Defuria, andK. E. Price. 1974. Fermentation and muta-tion studies with a butirosin-producingstrain of Bacillus circulans. Dev. Ind. Mi-crobiol. 15:102-113.
56. Coats, J. H. 1976. Genetic recombination inStreptomyces achromogenes var. rubradiris,p. 521-530. In K. D. Macdonald (ed.), SecondInternational Symposium on the Genetics ofIndustrial Microorganisms. Academic Press,London, New York, and San Francisco.
57. Coats, J. H., and J. Roeser. 1971. Genetic re-combination in Streptomyces bikiniensis var.zorbonensis. J. Bacteriol. 105:880-885.
58. Connolly, V., and G. Simchen. 1968. Linkageto the incompatibility factors and mainte-nance of genetic variation in selection lines
59. Corcoran, J. W. 1974. Lipid and macrolide lac-tone biosynthesis in Streptomyces erythreus.Dev. Ind. Microbiol. 15:93-100.
60. Coronelli, C., G. Tamoni, and G. C. Lancini.1976. Gardimycin, a new antibiotic from Ac-tinoplanes. II. Isolation and preliminarycharacterisation. J. Antibiot. 29:507-510.
61. Coronelli, C., R. J. White, G. C. Lancini, andF. Parenti. 1975. Lipiarmycin, a new anti-biotic from Actinoplanes. HI. Isolation, chem-ical, biological and biochemical characterisa-tion. J. Antibiot. 28:253-259.
62. Cove, D. J. 1976. The control of catabolism inAspergillus nidulans, p. 407-418. In K. D.Macdonald (ed.), Second International Sym-posium on the Genetics of Industrial Micro-organisms. Academic Press, London, NewYork, and San Francisco.
63. Cross, T., and R. W. Atwell. 1975. Actinomy-cete spores, p. 3-14. In H. 0. Halvorson, R.Hanson, and L. L. Campbell (ed.), Spores VI.American Society for Microbiology, Wash-ington, D.C.
64. Cross, T., and M. Goodfellow. 1973. Taxonomyand classification of the actinomycetes, p.11-112. In G. Sykes, and F. A. Skinner (ed.),Actinomycetales: characteristics and practi-cal importance. Academic Press, London.
65. Crowdy, S. H., J. F. Grove, and P. McClosky.1959. The translocation of antibiotics inhigher plants. IV. Systemic fungicidal activ-ity and chemical structure in griseofulvinrelatives. Biochem. J. 72:241-249.
66. de Bertoldi, M., and C. E. Caten. 1975. Isola-tion and haploidisation of heterozygous dip-loid strains in species of Humicola. J. Gen.Microbiol. 91:63-73.
67. Delic, V. 1965. Genetic recombination in Strep-tomyces erythreus (in Serbo-Croat). Mikro-biologija 2:153-156.
68. Demain, A. L. 1957. Inhibition of penicillinformation by lysine. Arch. Biochem. Bio-phys. 67:244-245.
69. Demain, A. L. 1966. Industrial fermentationsand their relation to regulatory mechanisms.Adv. Appl. Microbiol. 8:1-27.
70. Demain, A. L. 1974. Biochemistry of penicillinand cephalosporin fermentations. Lloydia37:147-167.
71. Demain, A. L. 1974. How do antibiotic produc-ing microorganisms avoid suicide? Ann.N.Y. Acad. Sci. 235:601-602.
72. Demain, A. L., and E. Inamine. 1970. Bio-chemistry and regulation of streptomycinand mannosidostreptomycinase (a-D-man-nosidase) formation. Bacteriol. Rev. 34:1-19.
73. Demain, A. L., and P. S. Masurekar. 1974.Lysine inhibition of in vivo homocitrate syn-thesis in Penicillium chrysogenum. J. Gen.Microbiol. 82:143-151.
74. Dennen, D. W., and D. D. Carver. 1969. Sulfa-tase regulation and antibiotic synthesis inCephalosporium acremonium. Can. J. Micro-
biol. 15:175-181.75. Dimroth, P., E. Ringelmann, and F. Lynen.
1976. 6-Methylsalicylic acid synthetase fromPenicillium patulum. Eur. J. Biochem.68:591-596.
76. Ditchburn, P., G. Holt and K. D. Macdonald.1976. The genetic location of mutations in-creasing penicillin yield in Aspergillus nidu-lans, p. 213-227. In K. D. Macdonald (ed.),Second International Symposium on the Ge-netics of Industrial Microorganisms. Aca-demic Press, London, New York, and SanFrancisco.
77. Dobrzanski, W. T., and H. Osowiecki. 1967.Isolation and some properties of the compe-tence factor from group H streptococcusstrain CHALLIS. J. Gen. Microbiol. 48:299-304.
78. Drew, S. W., and A. L. Demain. 1973. Methio-nine control of cephalosporin C formation.Biotech. Bioeng. 15:743-754.
79. Drew, S. W., and A. L. Demain. 1975. Produc-tion of cephalosporin C by single and doublesulfur auxotrophic mutants of Cephalospo-rium acremonium. Antimicrob. AgentsChemother. 8:5-10.
80. Drew, S. W., and A. L. Demain. 1975. Theobligatory role of methionine in the conver-sion of sulfate to cephalosporin C. Eur. J.Appl. Microbiol. 1:121-128.
81. Dulaney, E. L. 1969. Variation in selected pop-ulations of tetracycline producers, p. 93-120.In G. Sermonti and M. Alacevic (ed.), Ge-netics and breeding of streptomyces. Yugo-slav Academy of Sciences and Arts, Zagreb.
82. Dulaney, E. L., I. Putter, D. Drescher, L.Chaiet, W. J. Miller, F. J. Wolf, and D.Hendlin. 1962. Transethylation in antibioticbiosynthesis. I. An ethyl homolog of oxytet-racycline. Biochim. Biophys. Acta 60:447-449.
83. Elander, R. P. 1967. Enhanced penicillin bio-synthesis in mutant and recombinant strainsof Penicillium chrysogenum. Abh. Dtsch.Akad. Wiss. Berlin Kl. Med. 2:403-423.
84. Elander, R. P., M. A. Espenshade, S. G. Pa-thak, and C. H. Pan. 1973. The use of para-sexual genetics in an industrial strain im-provement program with Penicillium chryso-genum, p. 239-253. In Z. Vanek, Z. Hostalek,and J. Cudlin (ed.), Genetics of industrialmicroorganisms: actinomycetes and fungi.Academia, Prague.
85. Falconer, D. S. 1960. Quantitative genetics.Oliver and Boyd, Edinburgh and London.
86. Falkow, S. 1975. Infectious multiple drug re-sistance. Pion Ltd., London.
87. Fantini, A. A. 1962. Genetics and antibioticproduction of Emericellopsis species. Ge-netics 47:161-177.
88. Fantini, A. A., and L. S. Olive. 1960. Sexualrecombination in a homothallic, antibioticproducing fungus. Science 132:1670.
89. Fawcett, P. A., J. J. Usher, and E. P. Abra-ham. 1976. Aspects of cephalosporin and pen-
icillin biosynthesis, p. 129-138. In K. D. Mac-donald (ed.), Second International Sympo-sium on the Genetics of Industrial Microor-ganisms. Academic Press, London, NewYork, and San Francisco.
90. Ferenczy, L., F. Kevei, M. Szegedi, A. Frank6,and I. Rojik. 1976. Factors affecting high-frequency fungal protoplast fusion. Exper-ientia 32:1156-1158.
91. Fincham, J. R. S., and P. R. Day. 1971. Fungalgenetics, 3rd ed. Blackwell Scientific Publi-cations, Oxford and Edinburgh.
92. Flynn, E. H., M. H. McCormick, M. C.Stamper, H. de Valeria, and C. W. Godzeski.1962. A new natural penicillin from Penicil-lium chryosgenum. J. Am. Chem. Soc.84:4594-4595.
93. Forrester, P. I., and G. M. Gaucher. 1972.Conversion of 6-methylsalicylic acid into pa-tulin by Penicillium urticae. Biochemistry11:1102-1107.
94. Francis, M. M., R. Cella, and L. C. Vining.1975. Genetic recombination in a chloram-phenicol-producing strain of Streptomycesspecies 3022A. Can. J. Microbiol. 21:1151-1159.
95. Freeman, R. F., M. J. Bibb, and D. A. Hop-wood. 1977. Chloramphenicol acetyltransfer-ase-independent chloramphenicol resistancein Streptomyces coelicolor A3(2). J. Gen. Mi-crobiol. 98:453-465.
96. Friend, E. J., and D. A. Hopwood. 1971. Thelinkage map of Streptomyces rimosus. J.Gen. Microbiol. 68:187-197.
97. Fripp, Y. J. 1972. Genotype-environmental in-teractions in Schizophyllum commune. II.Assessing the environment. Heredity28:223-238.
98. Fripp, Y. J., and C. E. Caten. 1971. Genotype-environmental interactions in Schizophyl-lum commune. I. Analysis and character.Heredity 27:393-407.
99. Froyshov, 0. 1975. Enzyme-bound intermedi-ates in the biosynthesis of bacitracin. Eur. J.Biochem. 59:201-206.
100. Fuchs, P. G., J. Zajdel, and W. T. Dobrzan-ski. 1975. Possible plasmid nature of the de-terminant for production of the antibioticnisin in some strains of Streptomyces lactis.J. Gen. Microbiol. 88:189-192.
101. Fujisawa, Y., and T. Kanzaki. 1975. Occur-rence of a new cephalosporoate in a culturebroth of a Cephalosporium acremonium mu-tant. J. Antibiot. 28:372-378.
102. Fujisawa, Y., K. Kitano, and T. Kanzaki. 1975.Accumulation of deacetoxycephalosporin Cby a deacetylcephalosporin C negative mu.tant of Cephalosporium acremonium. Agric.Biol. Chem. 39:2049-2055.
103. Fujisawa, Y., H. Shirafuji, and T. Kanzaki.1975. Deacetylcephalosporin C formation bycephalosporin C acetylhydrolase induced in aCephalosporium acremonium mutant. Agric.Biol. Chem. 39:1303-1309.
104. Fujisawa, Y., H. Shirafuji, M. Kida, K. Nara,M. Yoneda, and T. Kanzaki. 1973. New find-
ings on cephalosporin C biosynthesis. Na-ture (London) New Biol. 246:154-155.
105. Fujisawa, Y., H. Shirafuji, M. Kida, K. Nara,M. Yoneda, and T. Kanzaki. 1975. Accumu-lation of deacetylcephalosporin C by cephalo-sporin C negative mutants of Cephalospo-rium acremonium. Agric. Biol. Chem.39:1295-1301.
106. Furumai, T., and M. Suzuki. 1975. Studieson the biosynthesis of basic 16-memberedmacrolide antibiotics, platenomycins. I. Se-lection of and cosynthesis by non-platenomy-cin-producing mutants. J. Antibiot. 28:770-774.
107. Furumai, T., K. Takeda, and M. Suzuki. 1975.Studies on the biosynthesis of basic 16-mem-bered macrolide antibiotics, platenomycins.IV. Biosynthesis of platenomycins. J. Anti-biot. 28:789-797.
108. Furuya, K., R. Enokita, and M. Shirasaka.1967. Antibiotics from fungi. II. A new gri-seofulvin producer, Nigrospora oryzae.Annu. Rep. Sankyo Res. Lab. 19:91-95.
109. Godtfredsen, W. O., S. Jahnsen, H. Lorck, K.Roholt, and L. Tybring. 1962. Fusidic acid: anew antibiotic. Nature (London) 193:987.
110. Gordon, R. E., D. A. Barnett, J. E. Hander-han, and C. H. Pang. 1974. Nocardia coe-liaca, Nocardia autotrophica and the nocar-din strain. Int. J. Syst. Bacteriol. 24:54-63.
111. Goulden, S. A., and F. W. Chattaway. 1968.Lysine control of a-aminoadipate and peni-cillin synthesis in Penicillium chrysogenum.Biochem. J. 110:55P.
112. Goulden, S. A., and F. W. Chattaway. 1969.End-product control of acetohydroxyacidsynthetase by valine in Penicillium chryso-genum Q176 and a high penicillin-yieldingmutant. J. Gen. Microbiol. 59:111-118.
113. Gregory, K. F., and J. C. C. Huang. 1964.Tyrosinase inheritance in Streptomyces sca-bies. I. Genetic recombination. J. Bacteriol.87:1281-1286.
114. Gregory, K. F., and J. C. C. Huang. 1964.Tyrosinase inheritance in Streptomyces sca-bies. II. Induction of tyrosinase deficiency byacridine dyes. J. Bacteriol. 87:1287-1294.
115. Grindle, M. 1963. Heterokaryon compatibilityof unrelated strains in the Aspergillus nidu-lans group. Heredity 18:191-204.
116. Gross, E., and J. L. Morell. 1971. The structureof nisin. J. Am. Chem. Soc. 93:4634-4635.
117. Haavik, H., and 0. Froyshov. 1975. Function ofpeptide antibiotics in producer organisms.Nature (London) 254:79-81.
118. Haneishi, T., A. Terahara, M. Arai, T. Hata,and C. Tamura. 1974. New antibiotics, meth-ylenomycins A & B. Structures of methylen-omycin A & B. J. Antibiot. 27:393-399.
119. Harold, R. J., and D. A. Hopwood. 1970. Ultra-violet-sensitive mutants of Streptomyces coe-licolor. II. Genetics. Mutat. Res. 10:439-448.
120. Harold, R. J., and D. A. Hopwood. 1972. Arapid method for complementation testing ofultraviolet-sensitive (uvs) mutants of Strep-tomyces coelicolor. Mutat. Res. 16:27-34.
121. Hastie, A. C. 1970. Benlate-induced instabilityof Aspergillus diploids. Nature (London)226:771.
122. Higgens, C. E., R. L. Hamill, T. H. Sands, M.M. Hoehn, N. E. Davis, R. Nagarajan, andL. D. Boeck. 1974. The occurrence of deace-toxycephalosporin C in fungi and streptomy-cetes. J. Antibiot. 27:298-300.
123. Hikino, H., Y. Asada, S. Arihara, and T. Take-moto. 1972. Fusidic acid, a steroidal anti-biotic from Isaria kogana. Chem. Pharm.Bull. 20:1067-1069.
124. Hirai, K., S. Nozoe, K. Tsuda, Y. litaka, K.Ishibashi, and M. Shirasaka. 1967. Thestructure of siccanin. Tetrahedron Lett.23:2177-2179.
125. Hogg, R. W., and H. P. Broquist. 1968. Homo-citrate formation in Neurospora crassa. Rela-tion to lysine biosynthesis. J. Biol. Chem.243:1839-1845.
126. Holloway, B. W. 1975. Genetic organization inPseudomonas, p. 133-161. In P. H. Clarkeand M. H. Richmond (ed.), Genetics and bio-chemistry of pseudomonas. Wiley, London.
127. Holt, G., and K. D. Macdonald. 1968. Penicil-lin production and its mode of inheritance inAspergillus nidulans. Antonie Van Leeu-wenhoek J. Microbiol. Serol. 34:409-416.
128. Holt, G., and K. D. Macdonald. 1968. Isolationof strains with increased penicillin yieldafter hybridisation in Aspergillus nidulans.Nature (London) 219:636-637.
129. Holt, G., G. F. St. L. Edwards, and K. D.Macdonald. 1976. The genetics of mutantsimpaired in the biosynthesis of penicillin, p.199-211. In K. D. Macdonald (ed.), SecondInternational Symposium on the Genetics ofIndustrial Microorganisms. Academic Press,London, New York, and San Francisco.
130. Hopwood, D. A. 1959. Linkage and the mecha-nism of recombination in Streptomyces coeli-color. Ann. N.Y. Acad. Sci. 81:887-898.
131. Hopwood, D. A. 1965. New data on the linkagemap of Streptomyces coelicolor. Genet. Res.6:248-262.
132. Hopwood, D. A. 1966. Lack of constant genomeends in Streptomyces coelicolor. Genetics54:1177-1184.
133. Hopwood, D. A. 1967. Genetic analysis andgenome structure in Streptomyces coelicolor.Bacteriol. Rev. 31:373-403.
134. Hopwood, D. A. 1969. Genome topology andmapping in Streptomyces coelicolor, p. 5-18.In G. Sermonti and M. Alacevi6 (ed.), Ge-netics and breeding in streptomyces. Yugo-slav Academy of Sciences and Arts, Zagreb.
135. Hopwood, D. A. 1972. Genetic analysis in mi-croorganisms, p. 29-159. In J. R. Norris andD. W. Ribbons (ed.), Methods in microbiol-ogy, vol. 7B. Academic Press, London.
136. Hopwood, D. A. 1973. Developments in actino-mycete genetics, p. 21-46. In Z. Vanek, Z.Hosftalek, and J. Cudlin (ed.), Genetics ofindustrial microorganisms: actinomycetesand fungi. Academia, Prague.
137. Hopwood, D. A. 1973. Genetics of the actino-
mycetales, p. 131-153. In G. Sykes and F. A.Skinner (ed.), Actinomycetales: characteris-tics and practical importance. AcademicPress, London.
138. Hopwood, D. A. 1977. Genetic recombinationand strain improvement. Dev. Ind. Micro-biol. 18:111-123.
139. Hopwood, D. A., and K. F. Chater. 1974. Strep-tomyces coelicolor, p. 237-255. In R. C. King(ed.), Handbook of genetics, vol. 1. PlenumPress, New York and London.
140. Hopwood, D. A., K. F. Chater, J. E. Dowding,and A. Vivian. 1973. Advances in Strepto-myces coelicolor genetics. Bacteriol Rev.37:371-405.
141. Hopwood, D. A., and G. Sermonti. 1962. Thegenetics of Streptomyces coelicolor. Adv. Ge-net. 11:273-342.
142. Hopwood, D. A., and H. M. Wright. 1972.Transformation in Thermoactinomyces vul-garis. J. Gen. Microbiol. 71:383-398.
143. Hopwood, D. A., and H. M. Wright. 1973.Transfer of a plasmid between streptomycesspecies. J. Gen. Microbiol. 77:187-195.
144. Hopwood, D. A., and H. M. Wright. 1973. Aplasmid ofStreptomyces coelicolor carrying achromosomal locus and its inter-specifictransfer. J. Gen. Microbiol. 79:331-342.
145. Hopwood, D. A., and H. M. Wright. 1976. In-teractions ofthe plasmid SCP1 with the chro-mosome of Streptomyces coelicolor A3(2), p.607-619. In K. D. Macdonald (ed.), SecondInternational Symposium on the Genetics ofIndustrial Microorganisms. Academic Press,London, New York, and San Francisco.
146. Hopwood, D. A., and H. M. Wright. 1976. Ge-netic studies on SCP1-prime strains of Strep-tomyces coelicolor A3(2). J. Gen. Microbiol.95:107-120.
146a.Hopwood, D. A., H. M. Wright, M. J. Bibb,and S. N. Cohen. 1977. Genetic recombina-tion through protoplast fusion in Strepto-myces. Nature (London) 268:171-174.
147. Hodialek, Z., M. Blumauerova, and Z. Vanek.1974. Genetic problems of the biosynthesis oftetracycline antibiotics, p. 13-67. In T. K.Ghose, A. Fiechter, and N. Blakebrough(ed.), Advances in biochemical engineering3. Springer-Verlag, Berlin, Heidelberg, andNew York.
148. Houghton, J. A. 1970. A new class of slow-growing non-perithecial mutants of Asper-gillus nidulans. Genet. Res. 16:285-292.
149. Iwaki, M., K. Shimura, M. Kanda, E. Kaji,and Y. Saito. 1972. Some mutants ofBacillusbrevis deficient in gramicidin S formation.Biochem. Biophys. Res. Commun. 48:113-118.
150. Jackson, M., E. L. Dulaney, I. Putter, H. M.Shafer, F. J. Wolf, and H. B. Woodruff.1962. Transethylation in anitbiotic biosyn-thesis II. Production of the 2'-ethoxy ana-logue of griseofulvin by biosynthesis.Biochim. Biophys. Acta 62:616-619.
151. Jinks, J. L. 1954. Somatic selection in fungi.Nature (London) 174:409-410.
152. Jinks, J. L., C. E. Caten, G. Simchen, and J.H. Croft. 1966. Heterokaryon incompatibil-ity and variation in wild populations of As-pergillus nidulans. Heredity 21:227-239.
153. Jinks, J. L., and V. Connolly. 1975. Determi-nation of the environmental sensitivity ofselection lines by the selection environ-ments. Heredity 34:401-406.
154. Jones, A., and D. W. S. Westlake. 1974. Regu-lation of chloramphenicol synthesis in Strep-tomyces sp. 3022a. Properties of arylaminesynthetase, an enzyme involved in antibioticbiosynthesis. Can. J. Microbiol. 20:1599-1611.
155. Kafer, E. 1958. An eight chromosome map ofAspergillus nidulans. Adv. Genet. 9:105-145.
156. Kafer, E. 1965. Origins of translocations inAspergillus nidulans. Genetics 52:217-232.
157. Kahler, R., and D. Noack. 1974. Action of acri-dine orange and ethidium bromide on growthand antibiotic activity of Streptomyces hy-groscopicus JA 6599. Z. Allg. Mikrobiol.14:529-533.
158. Kambe, M., Y. Imae, and K. Kurahashi. 1974.Biochemical studies on gramicidin S non-producing mutants of Bacillus brevis ATCC9999. J. Biochem. (Tokyo) 75:481-493.
159. Kanzaki, T., and Y. Fujisawa. 1975. Recentprogress in cephalosporin fermentation. J.Takeda Res. Lab. 34:324-349.
160. Kanzaki, T., T. Fukita, H. Shirafuji, Y. Fujis-awa, and K. Kitano. 1974. Occurrence of a 3-methylthiomethylcephem derivative in aculture broth of Cephalosporium mutant. J.Antibiot. 27:361-362.
161. Kawamoto, I., S. Takasawa, R. Okashi, M.Kohakura, I. Takahashi, and T. Nara. 1975.A new antibiotic victomycin (XK 49-1-B-2). I.Taxonomy and production of the producingorganism. J. Antibiot. 28:358-365.
162. Khua-16, L. 1962. Hybridisation in Actino-myces erythreus. (in Russian). Mikrobiolo-giya 31:61-65.
165. Kirby, R., and D. A. Hopwood. 1977. Geneticdetermination of methylenomycin synthesisby the SCP1 plasmid of Streptomyces coelico-lor A3(2). J. Gen. Microbiol. 98:239-252.
166. Kirby, R., L. F. Wright, and D. A. Hopwood.1975. Plasmid-determined antibiotic synthe-sis and resistance in Streptomyces coelicolor.Nature (London) 254:265-267.
167. Kohsaka, M., and A. L. Demain. 1976. Conver-sion of penicillin N to cephalosporin(s) bycell-free extracts of Cephalosporium acre-monium. Biochem. Biophys. Res. Commun.70:465-473.
168. Kojima, M., and A. Satoh. 1973. Microbialsemi-synthesis of aminoglycosidic antibioticsby mutants of S. ribosificus and S. kanamy-ceticus. J. Antibiot. 26:784-786.
169. Kominek, L. A., and 0. K. Sebek. 1974. Bio-synthesis of novobiocin and related antibiot-ics. Dev. Ind. Microbiol. 15:60-69.
170. Kozak, W., M. Rajchert-Trzpil, and W. T. Do-brzanski. 1974. The effect of proflavin, ethid-
ium bromide and an elevated temperature onthe appearance of nisin-negative clones innisin-producing strains of Streptococcus lac-tis. J. Gen. Microbiol. 83:295-302.
171. Kutzner, H. J., and S. A. Waksman. 1959.Streptomyces coelicolor Muller and Strepto-myces violaceoruber Waksman and Curtis,two distinctly different organisms. J. Bacte-riol. 78:528-538.
172. Laland, S. G., and T. Zimmer. 1973. The pro-tein thiotemplate mechanism of synthesis forthe peptide antibiotics produced by Bacillusbrevis. Essays Biochem. 9:31-57.
173. Lancini, G. C., and R. J. White. 1976. Rifamy-cin biosynthesis, p. 139-153. In K. D. Mac-donald (ed.), Second International Sympo-sium on the Genetics of Industrial Microor-ganisms. Academic Press, London, NewYork, and San Francisco.
173a. Lemke, P. A., and C. H. Nash. 1972. Muta-tions that affect antibiotic synthesis byCephalosporium acremonium. Can. J. Micro-biol. 18:255-259.
174. Lepesant-Kejzlarova, J., J.-A. Lepesant, J.Walle, A. Billault, and R. Dedonder. 1975.Revision of the linkage map of Bacillus sub-tilis 168: indications for circularity of thechromosome. J. Bacteriol. 121:823-834.
175. Lewin, B. 1974. Gene expression. 1. Bacterialgenomes. John Wiley & Sons, London andNew York.
176. Lewin, B. 1974. Gene expression. 2. Eucaryoticchromosomes. John Wiley & Sons, Londonand New York.
177. Lhoas, P. 1961. Mitotic haploidization by treat-ment ofAspergillus niger diploids with para-fluorophenylalanine. Nature (London)190:744.
178. Lhoas, P. 1967. Genetic analysis by means ofthe parasexual cycle in Aspergillus niger.Genet. Res. 10:45-61.
179. Liersch, M., J. Nuesch, and H. J. Treichler.1976. Final steps in the biosynthesis of ceph-alosporin C, p. 179-195. In K. D. Macdonald(ed.), Second International Symposium onthe Genetics of Industrial Microorganisms.Academic Press, London, New York, andSan Francisco.
180. Lipmann, F. 1973. Nonribosomal polypeptidesynthesis on polyenzyme templates. Acc.Chem. Res. 6:361-367.
181. Loder, P. B., and E. P. Abraham. 1971. Isola-tion and nature of intracellular peptidesfrom a cephalosporin C-producing Cephalo-sporium sp. Biochem. J. 123:471-476.
182. Lomovskaya, N. D., T. A. Voeykova, and N.M. Mkrtumian. 1977. Construction and prop-erties of hybrids obtained in interspecificcrosses between Streptomyces coelicolorA3(2) and Streptomyces griseus Kr.15. J.Gen. Microbiol. 98:187-198.
183. MacDonald, J. C. 1967. Pyocyanine, p. 52-65.In D. Gottlieb and P. D. Shaw (ed.). Anti-biotics: biosynthesis, vol. 2. Springer-Ver-lag, Berlin, Heidelberg, and New York.
184. Macdonald, K. D. 1966. Differences in diploidssynthesised between the same parentalstrains ofPenicillium chrysogenum. Antonievan Leeuwenhoek J. Microbiol. Serol.32:431-441.
185. Macdonald, K. D. 1968. The persistence of pa-rental genome segregation in Penicilliumchrysogenum after nitrogen mustard treat-ment. Mutat. Res. 5:302-305.
186. Macdonald, K. D. 1971. Segregants from a het-erozygous diploid of Penicillium chryso-genum following different physical andchemical treatments. J. Gen. Microbiol.67:247-250.
187. Macdonald, K. D., and G. Holt. 1976. Geneticsof biosynthesis and overproduction of penicil-lin. Sci. Prog. (London) 63:547-573.
188. Macdonald, K. D., G. Holt, and P. Ditchburn.1972. The genetics of penicillin production, p.251-257. Proceedings of the Fourth Interna-tional Fermentation Symposium: Fermenta-tion Technology Today. Society for Fermen-tation Technology, Osaka.
189. Macdonald, K. D., J. M. Hutchinson, and W.A. Gillett. 1963. Isolation of auxotrophs ofPenicillium chrysogenum and their penicil-lin yields. J. Gen. Microbiol. 33:365-374.
190. Macdonald, K. D., J. M. Hutchinson, and W.A. Gillett. 1963. Heterokaryon studies andthe genetic control of penicillin and chryso-genin production in Penicillium chryso-genum. J. Gen. Microbiol. 33:375-383.
191. Macdonald, K. D., J. M. Hutchinson, and W.A. Gillett. 1963. Formation and segregationof heterozygous diploids between a wild-typestrain and derivatives ofhigh penicillin yieldin Penicillium chrysogenum. J. Gen. Micro-biol. 33:385-394.
192. Macdonald, K. D., J. M. Hutchinson, and W.A. Gillett. 1964. Properties of heterozygousdiploids between strains ofPenicillium chry-sogenum selected for high penicillin yield.Antonie van Leeuwenhoek J. Microbiol.Serol. 30:209-224.
193. Macdonald, K. D., J. M. Hutchinson, and W.A. Gillett. 1965. Heterozygous diploids ofPenicillium chrysogenum and their segrega-tion patterns. Genetica 36:378-397.
194. Majer, J., J. R. Martin, R. S. Egan, and J. W.Corcoran. 1977. Antibiotic glycosides. VIII.Erythromycin D, a new macrolide antibiotic.J. Am. Chem. Soc. 99:1620-1622.
195. Maragoudakis, M. E., H. Holmes, and M.Strassman. 1967. Control of lysine biosyn-thesis in yeast by a feedback mechanism. J.Bacteriol. 93:1677-1680.
196. Martin, J. R., R. S. Egan, A. W. Goldstein,and P. Collum. 1975. Extension of the eryth-romycin biosynthetic pathway. Isolation andstructure of erythromycin E. Tetrahedron31:1985-1989.
197. Martin, J. R., and T. J. Perun. 1968. Studies onthe biosynthesis of the erythromycins. III.Isolation and structure of 5-deoxy-5-oxoery-thronolide B, a shunt metabolite of erythro-
mycin biosynthesis. Biochemistry 7:1728-1733.
198. Martin, J. R., T. J. Perun, and R. L. Girolami.1966. Studies on the biosynthesis of theerythromycins. I. Isolation and structure ofan intermediate glycoside, 3-a-L-mycarosyl-erythronolide B. Biochemistry 5:2852-2856.
199. Martin, J. R., and W. Rosenbrook. 1967. Stud-ies on the biosynthesis of the erythromycins.II. Isolation and structure of a biosyntheticintermediate, 6-deoxyerythronolide B. Bio-chemistry 6:435-440.
200. Masurekar, P. S., and A. L. Demain. 1974.Insensitivity of homocitrate synthase in ex-tracts of Penicillium chrysogenum to feed-back inhibition by lysine. Appl. Microbiol.28:265-270.
201. Masurekar, P. S., and A. L. Demain. 1974.Impaired penicillin production in lysine reg-ulatory mutants of Penicillium chryso-genum. Antimicrob. Agents Chemother.6:366-368.
202. Mather, K. 1942. The balance of polygenic com-binations. J. Genet. 43:309-336.
203. Mather, K. 1949. Biometrical genetics, 1st ed.Methuen, London.
204. Mather, K., and J. L. Jinks. 1971. Biometricalgenetics, 2nd ed. Chapman and Hall, Lon-don.
205. Matselyukh, B. P. 1976. Structure and func-tion of the Actinomyces olivaceus genome, p.553-563. In K. D. Macdonald (ed.), SecondInternational Symposium on the Genetics ofIndustrial Microorganisms. Academic Press,London, New York, and San Francisco.
206. McCormick, J. R. D. 1969. Point-blocked mu-tants and the biogenesis of tetracyclines, p.163-176. In G. Sermonti and M. Alacevic(ed.), Genetics and breeding of streptomyces.Yugoslav Academy of Sciences and Arts, Za-greb.
207. McCully, K. S., and E. Forbes. 1965. The use ofp-fluorophenylalanine with 'master strains'of Aspergillus nidulans. Genet. Res. 6:352-359.
208. Merrick, M. J. 1975. Hybridisation and selec-tion for increased penicillin titre in wild-typeisolates of Aspergillus nidulans. J. Gen. Mi-crobiol. 91:278-286.
209. Merrick, M. J. 1975. The inheritance of penicil-lin titre in crosses between lines ofAspergil-lus nidulans selected for increased produc-tivity. J. Gen. Microbiol. 91:287-294.
210. Merrick, M. J. 1976. Hybridization and selec-tion for penicillin production in Aspergillusnidulans -a biometrical approach to strainimprovement, p. 229-242. In K. D. Macdon-ald (ed.), Second International Symposiumon the Genetics of Industrial Microorga-nisms. Academic Press, London, New York,and San Francisco.
211. Merrick, M. J. 1976. A morphological and ge-netic mapping study of bald colony mutantsof Streptomyces coelicolor. J. Gen. Microbiol.96:299-315.
212. Merrick, M. J., and C. E. Caten. 1975. Theinheritance of penicillin titre in wild-typeisolates of Aspergillus nidulans. J. Gen. Mi-crobiol. 86:283-293.
213. Meynell, G. G. 1972. Bacterial plasmids. Mac-millan, London.
214. Mindlin, S. Z. 1969. Genetic recombination inthe actinomycete breeding, p. 147-159. In G.Sermonti and M. Alacevi6 (ed.), Geneticsand breeding of streptomyces. YugoslavAcademy of Sciences and Arts, Zagreb.
215. Miniken, D. E., and M. Goodfellow. 1976.Lipid composition in the classification andidentification of nocardiae and related taxa,p. 160-219. In M. Goodfellow, G. H. Brow-nell, and J. A. Serrano (ed.), The biology ofthe nocardiae. Academic Press, London,New York, and San Francisco.
216. Murphy, G., and F. Lynen. 1975. Patulin bio-synthesis: the metabolism of m-hydroxyben-zyl alcohol and m-hydroxybenzaldehyde byparticulate preparations from Penicilliumpatulum. Eur. J. Biochem. 58:467-475.
217. Musilek, V., J. Cerri, V. gagek, M. Semerd-zieva, and M. Vondratek. 1969. Antifungalantibiotic of the basidiomycete Oudeman-ciella mucida. I. Isolation and cultivation ofa producing strain. Folia Microbiol. (Prague)14:377-387.
218. Nagaoka, K., and A. L. Demain. 1975. Muta-tional biosynthesis of a new antibiotic, strep-tomutin A, by an idiotroph of Streptomycesgriseus. J. Antibiot. 28:627-635.
219. Nagarajan, R., L. D. Boeck, M. Gorman, R. L.Hamill, C. E. Higgens, M. N. Hoehn, W. M.Stark, and J. G. Whitney. 1971. f8-lactamantibiotics from streptomyces. J. Am. Chem.Soc. 93:2308-2310.
220. Nash, C. H.,N. de la Higuera, N. Neuss, and P.A. Lemke. 1974. Application of biochemicalgenetics to the biosynthesis of j8-lactam anti-biotics. Dev. Ind. Microbiol. 15:114-132.
221. Nash, C. H., and F. M. Huber. 1971. Antibioticsynthesis and morphological differentiationof Cephalosporium acremonium. Appl. Mi-crobiol. 22:6-10.
222. North, J. 1976. The effect of griseofulvin ondiploid strains of Coprinus lagopus. J. Gen.Microbiol. 98:529-534.
223. Nuesch, J., H. J. Treichler, and M. Liersch.1973. The biosynthesis of cephalosporin C, p.309-334. In Z. Vanek, Z. Hostalek, and J.Cudlin (ed.), Genetics of industrial microor-ganisms: actinomycetes and fungi. Acade-mia, Prague.
224. Okanishi, M., K. Hamana, and H. Umezawa.1968. Factors affecting infection of proto-plasts with deoxyribonucleic acid of actino-phage PK-66. J. Virol. 2:686-691.
225. Okanishi, M., T. Ohta, and H. Umezawa. 1970.Possible control of formation of aerial myce-lium and antibiotic production in Strepto-myces by episomic factors. J. Antibiot. 23:45-47.
226. Okanishi, M., K. Suzuki, and H. Umezawa.
1974. Formation and reversion of streptomy-cete protoplasts: cultural condition and mor-phological study. J. Gen. Microbiol. 80:389-400.
227. Okanishi, M., R. Utahara, and Y. Okami.1966. Infection of the protoplasts of Strepto-myces kanamyceticus with deoxyribonucleicacid preparation from actinophage PK-66. J.Bacteriol. 92:1850-1852.
228. Omura, S. 1976. The antibiotic cerulenin, anovel tool for biochemistry as an inhibitor offatty acid synthesis. Bacteriol. Rev. 40:681-697.
229. O'Sullivan, C. Y., and S. J. Pirt. 1973. Penicil-lin production by lysine auxotrophs of Peni-cillium chrysogenum. J. Gen. Microbiol.76:65-75.
230. Papa, K. E. 1971. Growth rate in Neurosporacrassa: linkage of polygenes. Genetica42:181-186.
231. Perlman, D. 1974. Prospects for the fermenta-tion industries, 1974-1983. Chem. Technol.4:210-216.
232. Peterson, E. A., D. C. Gillespie, and F. D.Cook. 1966. A wide-spectrum antibiotic pro-duced by a species of sorangium. Can. J.Microbiol. 12:221-230.
233. Pirali, G., S. Somma, G. C. Lancini, and F.Sala. 1974. Inhibition of peptide chain initia-tion in Escherichia coli by thermorubin.Biochim. Biophys. Acta 366:310-318.
234. Pirt, S. J. 1969. Microbial growth and productformation. Symp. Soc. Gen. Microbiol.19:199-221.
235. Polsinelli, M., and M. Beretta. 1966. Geneticrecombination in crosses between Strepto-myces aureofaciens and Streptomyces rimo-sus. J. Bacteriol. 91:63-68.
236. Pontecorvo, G. 1976. Presidential address, p.1-4. In K. D. Macdonald (ed.), Second Inter-national Symposium on the Genetics of In-dustrial Microorganisms. Academic Press,London, New York, and San Francisco.
237. Pontecorvo, G., and E. Kafer. 1958. Geneticanalysis based on mitotic recombination.Adv. Genet. 9:71-104.
238. Pontecorvo, G., J. A. Roper, L. M. Hemmons,K. D. Macdonald, and A. Bufton. 1953. Thegenetics of Aspergillus nidulans. Adv. Ge-net. 5:141-253.
239. Pontecorvo, G., and G. Sermonti. 1954. Para-sexual recombination in Penicillium chryso-genum. J. Gen. Microbiol. 11:94-104.
240. Puglia, A. M., I. Spada-Sermonti, S. Basile, F.Misuraca, and G. Sermonti. 1973. Infectioustransfer of a fertility factor in Streptomycescoelicolor. Genet. Res. 21:107-113.
241. Queener, S. W., J. J. Capone, A. B. Radue, andR. Nagarajan. 1974. Synthesis of deacetoxy-cephalosporin C by a mutant of Cephalospo-rium acremonium. Antimicrob. AgentsChemother. 6:334-337.
242. Queener, S. W., J. McDermott, and A. B.Radue. 1975. Glutamate dehydrogenase spe-cific activity and cephalosporin C synthesis
in the M8650 series of Cephalosporium acre-monium mutants. Antimicrob. AgentsChemother. 7:646-651.
243. Redshaw, P. A., P. A. McCann, L. Sankaran,and B. M. Pogell. 1976. Control of differen-tiation in streptomycetes: involvement of ex-trachromosomal deoxyribonucleic acid andglucose repression in aerial mycelia develop-ment. J. Bacteriol. 125:698-705.
244. Reeves, P. 1972. The bacteriocins. Chapmanand Hall Ltd., London.
245. Rickenberg, H. V. 1974. Cyclic AMP in prokar-yotes. Annu. Rev. Microbiol. 28:353-369.
246. Rinehart, K. L., and R. M. Stroshane. 1976.Biosynthesis of aminocyclitol antibiotics. J.Antibiot. 29:319-353.
247. Roland, I., 0. Froyshov, and S. G. Laland.1975. On the presence of pantothenic acid inthe three complementary enzymes of bacitra-cin synthetase. FEBS Lett. 60:305-308.
248. Roper, J. A. 1952. Production of heterozygousdiploids in filamentous fungi. Experientia8:14-15.
249. Russi, S., A. Carere, B. Fratello, and V. Khou-dokormoff. 1966. Caratterizzazione biochim-ica di alcuni mutanti di Streptomyces coelico-lor richiedenti istidina. Ann. Ist. Super. San-ita 2:506-522.
250. Sadoff, H. L. 1972. Sporulation antibiotics ofBacillus species, p. 157-166. In H. 0. Halvor-son, R. Hanson, and L. L. Campbell (ed.),Spores V. American Society for Microbiol-ogy, Washington, D.C.
251. Saito, H., and Y. Ikeda. 1959. Cytogeneticstudies on Streptomyces griseoflavus. Ann.N.Y. Acad. Sci. 81:862-878.
252. Sankaran, L., and B. M. Pogell. 1975. Biosyn-thesis of puromycin in Streptomyces alboni-ger: regulation and properties of 0-demethyl-puromycin 0-methyltransferase. Antimi-crob. Agent Chemother. 8:721-732.
253. Schaeffer, P. 1969. Sporulation and the pro-duction of antibiotics, exoenzymes and exo-toxins. Bacteriol. Rev. 33:48-71.
254. Schier, W. T., S. Ogawa, M. Hichens, and K.L. Rinehart. 1973. Chemistry and biochemis-try of the neomycins. XVII. Bioconversion ofaminocyclitols to aminocyclitol antibiotics.J. Antibiot. 26:551-561.
255. Schier, W. T., K. L. Rinehart, and D. Gottlieb.1969. Preparation of four new antibioticsfrom a mutant of Streptomyces fradiae. Proc.Natl. Acad. Sci. U.S.A. 63:198-204.
256. Schier, W. T., P. C. Schaefer, D. Gottlieb, andK. L. Rinehart. 1974. Use of mutants in thestudy of aminocyclitol antibiotic biosyn-thesis and the preparation of the hybrimycinC complex. Biochemistry 13:5073-5078.
257. Schrempf, H., H. Bujard, D. A. Hopwood, andW. Goebel. 1975. Isolation of covalentlyclosed circular deoxyribonucleic acid fromStreptomyces coelicolor A3(2). J. Bacteriol.121:416-421.
258. Schupp, T. 1976. Genetic analysis in Nocardiamediterranei, p. 531-533. In K. D. Macdonald
(ed.), Second International Symposium onthe Genetics of Industrial Microorganisms.Academic Press, London, New York, andSan Francisco.
259. Schupp, T., R. Hatter, and D. A. Hopwood.1975. Genetic recombination in Nocardiamediterranei. J. Bacteriol. 121:128-136.
260. Sermonti, G. 1956. Complementary geneswhich affect penicillin yields. J. Gen. Micro-biol. 15:599-608.
261. Sermonti, G. 1957. Produzione di penicillina dadiploidi eterozigoti di Penicillium chryso-genum. Conv. Genetica 1956 Ricerca Sci.27:93-100.
262. Sermonti, G. 1959. Genetics of penicillin pro-duction. Ann. N.Y. Acad. Sci. 81:950-973.
263. Sermonti, G. 1961. The parasexual cycle inPenicillium chrysogenum and its applicationto the production of penicillin. Sci. Rep. Ist.Super. Sanita. 1:441.
264. Sermonti, G. 1969. Genetics of antibiotic-pro-ducing microorganisms. Wiley-Interscience,London.
265. Sermonti, G., M. Bandiera, and I. Spada-Ser-monti. 1966. New approach to the genetics ofStreptomyces coelicolor. J. Bacteriol. 91:384-392.
266. Sermonti, G., A. Mancinelli, and I. Spada-Sermonti. 1960. Heterogeneous clones ("het-eroclones") in Streptomyces coelicolor A3(2).Genetics 45:669-672.
267. Sermonti, G., and A. M. Puglia. 1976. Progres-sive fertilization in Streptomyces coelicolor,p. 565-572. In K. D. Macdonald (ed.), SecondInternational Symposium on the Genetics ofIndustrial Microorganisms. Academic Press,London, New York, and San Francisco.
268. Sermonti, G., A. M. Puglia, and G. Ficarra.1971. The time course of recombinant produc-tion in Streptomyces coelicolor. Genet. Res.18:133-144.
269. Shimura, K., M. Iwaki, M. Kanda, K. Hori, E.Kaji, S. Hasegawa, and Y. Saito. 1974. Onthe enzyme system obtained from some mu-tants of Bacillus brevis deficient in gramici-din S formation. Biochim. Biophys. Acta338:577-587.
270. Silvestri, L. G. 1970. The evolution of the ther-mophilic Actinomycetales: an apparent evo-lutionary paradox, p. 239-243. In H. Prauser(ed.), The Actinomycetales, the Jena Inter-national Symposium on Taxonomy, 1968.Gustav Fischer Verlag, Jena.
271. Singh, J. 1967. Patulin, p. 621-630. In D. Got-tlieb and P. D. Shaw (ed.), Antibiotics, vol.1. Springer-Verlag, Berlin, Heidelberg, andNew York.
272. Smith, B., S. C. Warren, G. G. F. Newton, andE. P. Abraham. 1967. Biosynthesis of peni-cillin N and cephalosporin C. Biochem. J.103:877-890.
273. Somerson, N. L., A. L. Demain, and T. D.Nunheimer. 1961. Reversal of lysine inhibi-tion of penicillin production by a-aminoadi-pic acid. Arch. Biochem. Biophys. 93:238.
274. Stanisich, V. A., and M. H. Richmond. 1975.Gene transfer in the genus Pseudomonas, p.163-190. In P. H. Clarke and M. H. Rich-mond (ed.), Genetics and biochemistry ofpseudomonas. John Wiley & Sons, Londonand New York.
275. Stapley, E. O., M. Jackson, S. Hernandez, S.B. Zimmerman, S. A. Currie, S. Mochales,J. M. Mata, H. B. Woodruff, and D. Hen-dlin. 1972. Cephamycins, a new family of 38-lactam antibiotics. I. Production by actino-mycetes, including Streptomyces lactamdur-ans sp. n. Antimicrob. Agents Chemother.2:122-131.
276. Stauffer, J. F., and M. P. Backus. 1954. Spon-taneous and induced variation in selectedstocks of the Penicillium chrysogenum se-ries. Ann. N.Y. Acad. Sci. 60:35-50.
277. Tagg, J. R., A. S. Dajoni, and L. W. Wanna-maker. 1976. Bacteriocins of gram-positivebacteria. Bacteriol. Rev. 40:722-756.
278. Takasawa, S., I. Kawamoto, I. Takahashi, M.Kohakura, R. Okachi, S. Sato, M. Yamom-oto, T. Sato, and T. Nara. 1975. PlatomycinsA and B. I. Taxonomy of the producing strainand production, isolation and biologicalproperties of platomycins. J. Antibiot.28:656-661.
279. Tanaka, N. 1967. Variotin, p. 216-221. In D.Gottlieb and P. D. Shaw (ed.), Antibiotics:biosynthesis, vol. 2. Springer-Verlag, Ber-lin, Heidelberg, and New York.
280. Tanaka, H., Y. Koyama, T. Nagai, H. Mar-umo, and S. Omura. 1975. Nanaomycins,new antibiotics produced by a strain ofStrep-tomyces. II. Structure and biosynthesis. J.Antibiot. 28:868-875.
281. Tardrew, P. L., and M. J. Johnson. 1958. Sul-fate utilization by penicillin-producing mu-tants of Penicillium chrysogenum. J. Bacte-riol. 76:400-405.
282. Taylor, H. D., and H. Schmitz. 1976. Antibiot-ics derived from a mutant of Bacillus circu-lans. J. Antibiot. 29:532-535.
283. Testa, R. T., G. H. Wagman, P. J. L. Daniels,and M. J. Weinstein. 1974. Mutamicins: bio-synthetically created new sisomicin ana-logues. J. Antibiot. 27:917-921.
284. Tokunaga, T., Y. Mizuguchi, and K. Suga.1973. Genetic recombination in mycobacte-ria. J. Bacteriol. 113:1104-1111.
285. Traxler, P. H., J. Treichler, and J. Nuesch.1975. Synthesis of N-acetyldeacetoxycephal-osporin C by a mutant of Cephalosporiumacremonium. J. Antibiot. 28:605-606.
286. Turner, W. B. 1971. Fungal metabolites. Aca-demic Press, London and New York.
287. Umezawa, H. 1967. Index of antibiotics fromactinomycetes. Pennsylvania State Univer-sity Press, University Park.
288. Upshall, A., and E. Kafer. 1974. Detection andidentification of translocations by increasedspecific nondisjunction in Aspergillus nidu-lans. Genetics 76:19-31.
289. Vaks, B., A. Zuckerberg, and E. Rosenberg.
1974. Purification and partial characterisa-tion of an antibiotic produced by Myxococcusxanthus. Can. J. Microbiol. 20:155-161.
290. Vanderhaeghe, H., P. Van Dijck, and P. deSomer. 1965. Identity of ramycin with fusidicacid. Nature (London) 205:710-711.
291. Vanek, Z., and J. Majer. 1967. Macrolide anti-biotics, p. 154-188. In D. Gottlieb and P. D.Shaw (ed.), Antibiotics: biosynthesis, vol. 2.Springer-Verlag, New York.
292. Vining, L. C., V. S. Malik, and D. W. S.Westlake. 1968. Biosynthesis of chloram-phenicol. Lloydia 31:355-363.
293. Vivian, A. 1971. Genetic control of fertility inStreptomyces coelicolor A3(2): plasmid in-volvement in the interconversion of UF andIF strains. J. Gen. Microbiol. 69:353-364.
294. Vladimirov, A. V. 1966. Genetic recombinationin Actinomyces antibioticus producing olean-domycin (in Russian). Antibiotiki (Moscow)2:117.
295. Walker, J. B. 1971. Enzymatic reactions in-volved in streptomycin biosynthesis and me-tabolism. Lloydia 34:363-371.
296. Walker, J. B. 1967. Enzymatic studies on thebiosynthesis of streptomycin and related an-tibiotics. Dev. Ind. Microbiol. 8:109-113.
297. Watson, J. D. 1976. Molecular biology of thegene, 3rd ed. W. A. Benjamin, Inc., MenloPark, Calif.
298. Weinberg, E. D. 1974. Secondary metabolism:control by temperature and inorganic phos-phate. Dev. Ind. Microbiol. 15:70-81.
299. Wheelis, M. L. 1975. The genetics of dissimila-tory pathways in Pseudomonas. Annu. Rev.Microbiol. 29:505-523.
300. White, R. J., E. Martinelli, G. G. Gallo, G.Lancini, and P. Beynon. 1973. Rifamycinbiosynthesis studied with 13C-enriched pre-cursors and carbon magnetic resonance. Na-ture (London) 243:273-277.
301. White, R. J., E. Martinelli, and G. Lancini.1974. Ansamycin biogenesis: studies on anovel rifamycin isolated from a mutantstrain of Nocardia mediterranei. Proc. Natl.Acad. Sci. U.S.A. 71:3260-3264.
302. Wiame, J. M., and E. L. Dubois. 1976. Theregulation of enzyme synthesis in argininemetabolism of Saccharomyces cerevisiae, p.391-406. In K. D. Macdonald (ed.), SecondInternational Symposium on the Genetics ofIndustrial Microorganisms. Academic Press,London, New York, and San Francisco.
303. Wildermuth, H. 1970. Surface structure ofstreptomycete spores as revealed by negativestaining and freeze-etching. J. Bacteriol.101:318-322.
304. Williams, R. P., and W. R. Hearn. 1967. Pro-digiosin, p. 410-432. In D. Gottlieb and P. D.Shaw (ed.), Antibiotics: biosynthesis, vol. 2.Springer-Verlag, Berlin, Heidelberg, andNew York.
305. Williams, S. T., and M. R. Khan. Antibiotics-a soil microbiologist's view. Postepy Hig.Med. Dosw. 28:395-408.
306. Williams, S. T., G. P. Sharples, and R. M.Bradshaw. 1973. The fine structure of theactinomycetales, p. 113-130. In G. Sykes andF. A. Skinner (ed.), Actinomycetales:characteristics and practical importance. Ac-ademic Press, London, and New York.
307. Wright, L. F., and D. A. Hopwood. 1976. Iden-tification of the antibiotic determined by theSCP1 plasmid of S. coelicolor A3(2). J. Gen.Microbiol. 95:96-106.
308. Wright, L. F., and D. A. Hopwood. 1976. Acti-norhodin is a chromosomally determined an-tibiotic in Streptomyces coelicolor A3(2). J.Gen. Microbiol. 96:289-297.
309. Young, F. E., and G. A. Wilson. 1972. Genetics
of Bacillus subtilis and other gram-positivesporulating bacilli, p. 77-106. In H. 0. Hal-vorson, R. Hanson, and L. L. Campbell(ed.), Spores V. American Society for Micro-biology, Washington, D.C.
310. Young, F. E., and G. A. Wilson. 1974. Bacillussubtilis, p. 69-114. In R. C. King (ed.), Hand-book of genetics, vol. 1. Plenum Press, NewYork and London.
311. Zhukova, R. A., L. N. Lanskaya, and T. G.Gavryuchenkova. 1973. Lethal and selectiveeffect of polygenic antibiotics levonin andmycoheptin on their producing organisms (inRussian). Genetika 9:69-77.
