the peptide antibiotics ofbacillus: chemistry, biogenesis...

BACTmaioLoGIcAL Rzvizws, June 1977, p. 449-474Copyright C 1977 American Society for Microbiology

Vol. 41, No. 2Printed in U.S.A.

The Peptide Antibiotics of Bacillus: Chemistry, Biogenesis,and Possible Functions

EDWARD KATZ* AND ARNOLD L. DEMAINDepartment of Microbiology, Georgetown University Schools ofMedicine and Dentistry, Washington, D.C.20007,* and Department ofNutrition and Food Science, Massachusetts Institute of Technology, Cambridge,

Massachusetts 02139

INTRODUCTION............ 449PROPERTIES OF PEPTIDE ANTIBIOTICS.449FORMATION OF ANTIBIOTICS IN RELATION TO GROWTH ...... ............ 450CELL-FREE SYNTHESIS OF PEPTIDE ANTIBIOTICS ...... .................. 451ENZYMATIC SYNTHESIS OF GRAMICIDIN S ........ ......................... 451ENZYMATIC FORMATION OF TYROCIDINE ........ ......................... 453LINEAR GRAMICIDINS ...................... ................................. 456BACITRACINS ................................................................ 457EDEINE...................................................................... 458MYCOBACILLIN ................. ............................................ 459POLYMYXIN AND COLISTIN ............... .................................. 460SPECIFICITY OF INCORPORATION ........... ............................... 461RACEMIZATION AND THE SYNTHESIS OF D-AMINO ACIDS.................. 463MUTANT STUDIES........................................................... 464POSSIBLE FUNCTIONS OF PEPTIDE ANTIBIOTICS IN THE PRODUCING OR-

GANISM .............................................................. 465CONCLUDING REMARKS ................... .................................. 468ADDENDUM IN PROOF .................... .................................. 469LITERATURE CITED......................................................... 469

INTRODUCTIONIn a recent review, Berdy stated that the

number of antibiotics produced by members ofthe genus Bacillus was 167 (8). Only a few ofthe total number elaborated are listed in Table1. As is generally recognized, these antibioticsare mainly polypeptides. Ofthis total, 66 differ-ent peptide antibiotics are elaborated by strainsofBacillus subtilis and 23 are products ofBacil-lus brevis. Polymyxin and the closely relatedcolistin, bacitracin, the tyrothricin complex(linear gramicidin plus tyrocidine), and grami-cidin S have been used, to some extent, forantibacterial therapy. Most ofthe peptide anti-biotics produced by bacilli are active againstgram-positive bacteria; however, compoundssuch as polymyxin, colistin, and circulin exhibitactivity almost exclusively upon gram-negativeforms, whereas bacillomycin, mycobacillin, andfungistatin are effective agents against moldsand yeasts. Although bacilli mainly synthesizepeptides, one should not lose sight of the factthat antibiotics belonging to other chemicalclasses are also produced by these microorga-nisms (e.g., butirosin [2], an aminoglycosidederived from Bacillus circulans and proticin[111], a phosphorus-containing triene elabo-rated by Bacillus licheniformis var. mesenteri-cus).

PROPERTIES OF PEPTIDEANTIBIOTICS

A number of reviews relating to peptide anti-biotics has appeared in the literature (14, 15,55, 71, 72, 80, 109, 117, 132). As a consequence,the properties of peptide antibiotics will be de-scribed only briefly below.

(i) Peptide antibiotics are generally muchsmaller than proteins. Their molecular weightsrange from 270 (bacilysin) to about 4,500 (lich-eniformin).

(ii) Generally, a family of closely related pep-tides rather than a single substance is producedby an organism. The members may differ fromeach other by one or, at most, a few amino acidresidues. In the case of B. brevis, both thelinear gramicidins (A, B, and C) and the cyclictyrocidines (A, B, and C) are formed simultane-ously. The linear gramicidins (pentadecapep-tides) and the tyrocidines (decapeptides) differmarkedly in chemical composition, yet withineach family the principal variation resides inthe aromatic amino acid substitutions of themolecule (34, 73, 114, 131, 134, 135, 136, 137).

(iii) Most of the peptide antibiotics formed bythese organisms are composed entirely ofamino acids, whereas others may contain aminoacids plus other constituents. For example, ed-eine A contains the base spermidine in addition

449

on April 16, 2018 by guest

http://mm

br.asm.org/

Dow

nloaded from

http://mmbr.asm.org/

450 KATZ AND DEMAIN

TABLE 1. Some antibiotics elaborated by species ofthe genus Bacillus

Species

Bacillus brevis

Bacillus subtilis

Bacillus pumilis

Bacillus mesentericus

Bacillus thiaminolyticus

Bacillus licheniformis

Bacillus polymyxa

Bacillus circulans

Bacillus laterosporus

Bacillus cereus

Antibiotic

Gramicidin STyrocidineLinear gramicidinBrevinEdeineEseineBresseineBrevistin

MycobacillinSubtilinBacilysinBacillomycinFungistatinBulbiforminBacillinSubsporinBacillocinMycosubtilinFungocinIturinNeocidinEumycinMicrococcin PPumilinTetain

Esperin

Octopytin (Thianosine)Baciphelacin

BacitracinLicheniforminProticin

PolymyxinColistinGatavalinJolipeptin

ButirosinCirculinPolypeptinEM-49Xylostatin

LaterosporamineLaterosporin

BiocerinCerexinThiocillin

to five amino acid constituents (48). Poly-myxins possess 6-methyloctanoic acid or 6-methylheptanoic acid as a fatty acid residuealong with a number of amino acid constituents(71, 163).

(iv) Frequently, peptide antibiotics containamino acids which are unique and are notfound in proteins (14, 15, 71). 1)-Amino acids,basic amino acids (ornithine, diaminobutyricacid), ,-amino acids, dehydroamino acids (de-hydroalanine), and sulfur-containing aminoacids (lanthionine) are often present. In con-trast to peptides synthesized by fungi and acti-nomycetes, peptides from bacilli do not containN-methylamino acid residues.

(v) Most of the peptides are cyclic structures;however, a few are linear (e.g., edeine, lineargramicidins). Besides the cyclic nature of a mol-ecule, there may be unusual linkages or ar-rangements of the amino acids in the antibiotic.Bacitracin provides an example, for it possessesa thiazoline ring (derived from the condensa-tion of cysteine and isoleucine), a cyclic hexa-peptide, as well as an amide bond between thef3-carboxyl group of aspartic acid and the epsi-lon amino group of lysine (1, 167).

(vi) Peptide antibiotics are generally resist-ant to hydrolysis by peptidases and proteases ofanimal and plant origin (14, 15, 71, 72, 117),although a few are susceptible to enzymaticattack (polymyxin B by ficin and papain [116];edeines A and B by carboxypeptidase [49]; baci-lysin by leucine aminopeptidase and Pronase[124]; gramicidin S by subtilopeptidase A [174]).

FORMATION OF ANTIBIOTICS INRELATION TO GROWTH

A number of investigations have shown thatthe synthesis of peptide antibiotics is initiatedafter the organism has passed the rapid growthphase. This was noted with gramicidin S (160),tyrocidine (36, 90), polymyxin (116), edeine(82), bacitracin (11), mycobacillin (5), and baci-lysin (124, 126). Tomino et al. (160) demon-strated that gramicidin S and the enzymes nec-essary for its synthesis were produced by B.brevis during the late logarithmic phase ofgrowth (107, 160). The enzymatic activity forantibiotic synthesis then disappeared within afew hours. Similar results were observed withtyrocidine synthetase (36). The reason for theloss ofenzyme activity has not been establishedconclusively. Recently, Lee et al. (90) observedthat the tyrocidine-forming enzymes disap-peared from the soluble portion of the cell andbecame associated with the forespore mem-brane during stage IV of sporulation. However,even this membrane-bound activity disap-peared soon thereafter. Demain et al. (25) found

BACTERIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from


PEPTIDE ANTIBIOTICS OF BACILLUS 451

that the disappearance of gramicidin S synthe-tases in B. brevis is due to an oxygen-dependentinactivation.Although antibiotic formation usually fol-

lows logarithmic growth (presumably due tosome type of repression of antibiotic synthe-tases in the growth phase), this is not univer-sally observed. It is clear that antibiotics aresometimes produced during growth and thatboth genetic and nutritional modifications canshift the time of antibiotic synthesis in relationto the growth phase (23). It is quite possiblethat peptide antibiotic formation is controlledby carbon and nitrogen catabolite repression oris under growth rate control (106); manipula-tions which affect these controls would be ex-pected to modify the temporal relationship be-tween antibiotic synthesis and growth. In thisregard, production of bacitracin by B. licheni-formis parallels growth in a chemically definedmedium (47, 152). However, Haavik (41, 42)reported that when glucose was added to a de-fined medium, bacitracin synthesis was in-hibited during the first few hours ofgrowth. Heinitially concluded that this delay in antibioticproduction was not due to carbon cataboliterepression but to the low pH created by theorganism's metabolism. His conclusion wasbased on the reversal of the glucose effect byadded CaCO3 (41). However, he later (42) foundthat the depression of bacitracin synthesis wasnot due to pH per se but to the acetic andpyruvic acids produced from the glucose, i.e,these acids were only inhibitory when undisso-ciated and thus neutralization eliminated theirinhibition. Lowering pH with HCl or phthalicacid did not inhibit bacitracin formation. Itshould be emphasized that this work does noteliminate the possibility that repression ofbaci-tracin synthetase is due to the undissociatedorganic acids produced by glucose catabolism.

In addition to bacitracin, the production oftyrocidine (79), linear gramicidins (79), and po-lymyxin E (colistin) (28) has also been observedto parallel active growth of the producing ba-cilli under certain conditions. Moreover, grami-cidin S (13) and its synthetases (106) can besynthesized by exponentially growing cells incontinuous culture. Thus, it is clear that pep-tide antibiotics can be synthesized during ac-tive growth. Whether they function in the pro-ducing cells' metabolism at some stage duringgrowth, as some investigators (43, 52, 70) haveconsidered, will be discussed in a later sectionof this review.CELL-FREE SYNTHESIS OF PEPTIDE

ANTIBIOTICSThe cell-free synthesis of gramicidin S, tyro-

cidine, linear gramicidin, edeine, bacitracin,

colistin, and mycobacillin has been reviewed(15, 55, 72, 80, 117). These biosynthetic systemsare quite distinct from those synthesizing pro-teins. The requirements for antibiotic synthesisare fairly similar and generally include therequisite amino acids of the peptide antibioticin question, adenosine 5'-triphosphate (ATP),Mg2+ ion, a reducing agent, and the particle-free supernatant fluid. Many ofthe synthetaseshave been extensively purified by conventionalprocedures, and considerable success has beenachieved in resolving the polyenzyme com-plexes by column chromatography (diethyl-aminoethyl [DEAE]-cellulose, hydroxylapatite,Sephadex G-200). The peptide-synthesizing sys-tems have proven to be insensitive to chloram-phenicol, puromycin, deoxyribonuclease, andribonuclease (RNase), and it is quite clear thattransfer ribonucleic acid (tRNA), messengerRNA (mRNA), and ribosomes are not directlyinvolved in the biosynthetic process. The con-tributions of several laboratories have led toour present understanding of the mechanism ofbiosynthesis of gramicidin S and tyrocidine (9,12, 35-38, 65, 75, 81, 87, 88, 94, 95, 97, 113, 119,133, 160, 162, 175). Far less is known, however,regarding the cell-free synthesis of other pep-tide antibiotics produced by bacilli.

ENZYMATIC SYNTHESIS OFGRAMICIDIN S

Gramicidin S (Fig. 1), formed by certainstrains ofB. brevis, is a cyclic decapeptide con-sisting of two identical pentapeptides, D-phen-ylalanyl-L-prolyl-L-valyl-L-ornithinyl-L-leucine(15, 55, 71). The presence of -phenylalanineand L-ornithine, constituents which are notfound in proteins, is noteworthy.

In the formation of gramicidin S in vivo, the

L-LeuL-Orn D-Phe/ -

L-VaI L-Pro

L-Pro

D-Phe

L-LeuFIG. 1. Gramicidin S.

L-Val/

L-Orn

VOL. 41, 1977


http://mm

br.asm.org/

Dow

nloaded from


452 KATZ AND DEMAIN

first amino acid added is phenylalanine; in vi-tro it can be replaced by tyrosine, but not bytryptophan (128). Phenylalanine (or the tyro-sine replacing it) is incorporated as the 1-aminoacid. An enzyme (designated as fraction II orthe light enzyme) with a molecular weight of100,000 (Table 2) activates L-phenylalanine (orD-phenylalanine) and racemizes it (37, 38, 160,171, 172). Activation (Fig. 2) is an ATP-depend-ent process and, as observed in protein synthe-sis, there is formation ofa phenylalanyl adenyl-ate-enzyme complex (37, 75, 160, 171, 172). Theremaining four amino acids, L-proline, L-va-line, L-ornithine, and iLleucine, are activatedby a second protein designated fraction I orheavy enzyme (Table 2) which has a molecularweight of 280,000 (37, 75, 160). In addition to theamino acids, ATP and Mg2+ ion are require-ments for the activation step. Each of the fouramino acids is activated independently onseparate sites of the polyenzyme, also beingbound as aminoacyl adenylates (Fig. 2). Theactivation process, as measured by the ATP-inorganic pyrophosphate (PP1) exchange reac-tion, appears to be relatively specific since onlyamino acids chemically related to those ingramicidin S are activated. In contrast to the

ribosomal system, acylation of tRNA does nottake place during activation with either thelight or the heavy enzyme (80, 87, 94, 97).Both enzymes have been shown to catalyze

not only an ATP-PPj exchange reaction, but anATP-adenosine 5'-monophosphate (AMP) ex-change as well with the appropriate amino acidsubstrates. This is due to the presence of sec-ondary acceptors of the bound amino acids onthe enzymes (37, 38). The aminoacyl moiety istransferred from the adenylate site to an SHgroup on the enzyme, yielding a covalentlybound thioester-linked amino acid. Thus, theheavy enzyme contains four activating en-zymes organized into a tightly associated mul-tienzyme complex; the binding of the aminoacid substrates to this complex exhibits a 1:1:1:1stoichiometry (75). Both the adenylate and SHsites will be occupied under saturating condi-tions, because once the amino acid is trans-ferred to a thiol group, the adenylate site is freeto accept another molecule of activated aminoacid (75, 94, 97).

Several experiments have provided supportfor the view that covalent thioester bonds areformed between the individual amino acids andthe specific enzymes involved (32, 38, 88, 129).

TABLE 2. Features ofgramicidin S and tyrocidine synthetases4'-Phos-

Enzyme Fraction Nomenclature Amino acid activated phopan- Mol wttetheine

Gramicidin S syn- II Light enzyme L- and D-phenylalanine 0 100,000thetase

I Heavy enzyme L-Proline, L-valine, L-ornithine, L-leucine 1 280,000Tyrocidine syn- I Light enzyme L- and D-phenylalanine 0 100,000

thetaseII Intermediate en- L-Proline, L- and D-phenylalanine (trypto- 1 230,000

zyme phan)III Heavy enzyme L-Asparagine, L-glutamine, iphenylalanine 1 460,000

(tryptophan, tyrosine)L-valine, L-ornithine, L-leucine

ACTIVATION

(1) SHaa + ATP _ g

ESH..... .aa - AMP + PP1

(2) ESH E S-aa......aa - AMP .L + AMP

E5 ~+aa + ATP MS-aa.........aa-AMP + PPi

FIG. 2. Proposed mechanism of activation of amino acids in gramicidin S. (Reprinted with permissionofF. Lipmann.)

BACTZRIOL. Rzv.


http://mm

br.asm.org/

Dow

nloaded from


VOL. 41, 1977

(i) After an incubation, an activated aminoacid-enzyme complex can be recovered by Seph-adex G-50 chromatography; precipitation ofsuch a complex with trichloroacetic acid re-leases all of the bound adenylate, but only one-half of the bound amino acid is liberated. (ii)Dilute alkali (at pH 9 to 10) or treatment withmercuric salts at neutral pH brings about aquantitative discharge of the remaining aminoacid from the trichloroacetic acid precipitate.(iii) Hydroxylamine treatment of the precipi-tate will also release the amino acid as thehydroxamate. (iv) Treatment of the precipitatewith sodium borohydride reductively cleavesthe bond with release of the amino acid as thecorresponding amino alcohol. (v) Oxidativecleavage with performic acid also will liberatethe amino acid. Thus, the behavior shown bythe covalently bound amino acid to acid, base,mercuric salts, etc., comprises all ofthe charac-teristic properties of a thioester linkage.

Initiation of peptide bond synthesis. It isnow well established that free peptide interme-diates are not participants and that an enzyme-bound, activated D-phenylalanine residue is thestarting point for gramicidin S formation (32,38, 53, 98). The direction of peptide chaingrowth is from the N-terminal to the C-termi-nal end as in protein synthesis, and only en-zyme-bound intermediates are involved. Theseconclusions were based on in vivo (53, 159) andin vitro (16, 38, 75) experiments. During grami-cidin S synthesis, the amino acids are addedsingly and in a fixed order until a pentapeptideis formed; omission ofone amino acid interruptsthe elongation process.

For peptide synthesis to occur, the charged oractivated enzymes have to be combined (75).The first peptide bond involves thioester-boundD-phenylalanine on the light enzyme andthioester-linked proline on the heavy enzyme(74, 81, 87, 94, 95, 97). The carboxyl-activated D-phenylalanine is transferred to the imino groupof proline on the heavy enzyme (Fig. 3). Inturn, the thioester-linked dipeptide is trans-ferred to the next thioester-linked amino acid,L-valine, to form a tripeptide. All peptides re-main attached to the enzyme until leucine, thelast amino acid in the sequence, is added (Fig.4). Leucine addition brings about a rapidrelease of cyclic gramicidin S from the polyen-zyme. The mechanism of the cyclization step isnot well understood. Lipmann's group proposedthat the cyclic decapeptide of gramicidin S mayarise by an antiparallel doubling reaction be-tween two carboxyl-activated pentapeptideunits presumably attached to different heavyenzyme components (94, 97) (Fig. 4). Some evi-dence, however, suggests an intramolecular cy-


Initiation

[- S-PheNH2 + -S -Pro NH

E -SH + E s - Pro- Phe NH2FIG. 3. Initiation reaction for gramicidin S for-

mation. (Reprinted with permission ofF. Lipmann.)

clization (87, 171). Stoll et al. (153) reportedthat, after the first pentapeptide chain isformed, it is transferred to a holding site (-SHgroup) on the heavy enzyme. When the secondpentapeptidyl unit is synthesized subsequently,the two pentapeptide chains attached to thesame heavy enzyme cyclize by head-to-tail con-densation to yield a molecule of gramicidin S(Fig. 5).4'-Phosphopantetheine, bound to the heavy

enzyme of gramicidin S synthetase, mediatesthe biosynthesis of the tri-, tetra-, and penta-peptides in alternative transthiolation andtranspeptidation reactions (38, 39, 76, 77, 87,94, 95, 97, 98). 4'-Phosphopantetheine is con-ceived to function as a swinging arm in which italternates as acceptor of the growing peptidechain and donor of the peptide to the nextthioester-linked amino acid in sequence (87, 95,119). These reactions continue until the tetra-peptide is linked to the terminal amino acid,leucine, to complete the nascent pentapeptidechain. Lipmann has postulated that the thioes-ter-linked polymerization of polypeptides mayhave evolved from thioester-linked fatty acidsynthesis (94, 96).

ENZYMATIC FORMATION OFTYROCIDINE

The mechanism of biosynthesis of tyrocidineis similar to that of gramicidin S (80, 94, 95, 97),and only a brief discussion will be presentedhere. Tyrocidines A, B, and C are cyclic deca-peptides (Fig. 6) (34, 71, 73, 114, 131) producedby the same strains of B. brevis that elaboratethe linear gramicidins. This strain cannot pro-duce the cyclic gramicidin S. Although they arecyclic decapeptides like gramicidin S, tyroci-dines possess a somewhat different sequence ofamino acids. The biosynthesis of tyrocidine alsobegins with D-phenylalanine and continueswith proline (33, 35, 36, 89, 91, 94, 95, 97, 119,128). However, five different amino acids, L-phenylalanine, D-phenylalanine, L-asparagine,L-glutamine, and I-tyrosine, follow. After tyro-


http://mm

br.asm.org/

Dow

nloaded from


454 KATZ AND DEMAIN

Elongation and Cyclization

I

-s PRO-PHE

-s '-'VAL-PRO-PHE

-S VORN-VAL-PRO-PHE

-s (LEU-ORN-VAL-PRO-PHE

PHE-PRO-VAL-ORN-LEU- S- 1U'I

FIG. 4. Proposed mechanism ofelongation and cyclization for gramicidin S biosynthesis. (Reprinted withpermission ofF. Lipmann.)

FIG. 5. Model for the role ofenzyme I in gramicidin S biosynthesis. Positions 1, 2, 3, and 4 are occupied bythioester-linked proline, valine, ornithine, and leucine, respectively. Position 5 is envisioned as the thiol"waiting" site for the pentapeptide. The zigzag line represents the pantetheine arm with its terminal SHgroup.(A) Growth of nascent peptide chain. (B) Head-to-tail cyclization reaction involving two identical pentapep-tides. (Reprinted with permission of S. G. Laland.)

sine, the sequence of gramicidin S is seen againwith L-valine, L-ornithine, and then ileucinebeing the last amino acid added.

Purification of the tyrocidine-synthesizingsystem yields three complementary enzymes: alight fraction (molecular weight 100,000), anintermediate one (molecular weight 230,000)and a heavy enzyme (molecular weight 440,000)(22, 36, 69, 91, 127). The amino acids activatedby each of these fractions are shown in Table 2.The light enzyme activates and racemizesphenylalanine (tyrosine or tryptophan can re-

place it) (97, 127). The intermediate enzymeactivates proline, phenylalanine, and phenylal-anine in positions 2, 3, and 4, respectively.Phenylalanine at site 4 is also racemized on theenzyme. The heavy enzyme catalyzes the acti-vation of the remaining six amino acids. If onedivides the molecular weight of the heavy en-zyme of the gramicidin S synthetase complex orthe intermediate or heavy enzyme of tyrocidinesynthetase by the number of amino acids acti-vated, one obtains a value of 70,000 to 75,000daltons per amino acid activated and thioesteri-

IBACTECRIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from


PEPTIDE ANTIBIOTICS OF BACILLUS

fled (89, 91, 95). The intermediate and heavyenzymes contain 1 mol each of 4'-phosphopan-tetheine bound per mol of enzyme protein. Intyrocidine synthesis, 4'-phosphopantetheinealso functions in the alternative transthiolationand transpeptidation reactions during nascent

peptide elongation (76, 77, 89, 128). Figure 7schematically depicts the mechanism of initia-tion, elongation, and cyclization ofthe decapep-tide of tyrocidine. Initiation of chain growth issimilar to that seen in gramicidin S biosyn-thesis. Addition of amino acids occurs singly,

L-Pro --

A

L-Phe

D-PheQ GD-I

L- Le u ® @ L-

L-Orn®) L-(

L-V\a L-Tyr

Ph

G I

-- 1- T r p

e D-Ph4

sfn L-A'

1 L-Glr

- L-Ty r

B

*-s L-Trpi - T r peD-Trr

sn L-A'

1 L-GIn

- L-Ty r

C

--sL-T r p\-T r pD-Trp

sn L-Asn

L-G In/

DL-T r p

DFIG. 6. Amino acid sequence of tyrocidines A, B. C, and D. The left half sequences, omitted in the

diagrams of tyrocidines B, C, and D, are identical to that of tyrocidine A.

Sr3 D- Phe

SX Ze Pro

0{)t)@Phe

0O @I~D-Phe(50(S) @ ~Asn

> O (S) @ ~Gln

[c b ) @ ~~~Phe

ofQ ) @ ~~~~~VaI}( }(>O ) @ ~~OrnCO 0000~~~~~~~~~~A-I Leu

FIG. 7. Summary of reactions involved in tyrocidine biosynthesis. The cross-hatched areas indicate indi-vidual sites for binding ofthe aminoacyl adenylates, each next to thiol groups responsible for thioester bindingof amino acids and peptides as shown. Phosphopantetheine arm is omitted. (Reprinted with permission ofF. Lipmann.)

455VOL. 41, 1977


http://mm

br.asm.org/

Dow

nloaded from


456 KATZ AND DEMAIN

with a tetrapeptide being formed on the inter-mediate enzyme. The tetrapeptide is thentransferred to the heavy enzyme where the re-maining six amino acids are added sequentiallyto complete the peptide chain (127, 128). Cycli-zation of the tyrocidine decapeptide appears tobe a relatively slow reaction, the linear deca-peptide having been isolated from the heavyenzyme and characterized. Cyclization involvesthe activated enzyme-bound thioester-linkedcarboxyl group of leucine and the free aminogroup of D-phenylalanine. As with gramicidin Ssynthesis in vitro, omission ofone amino acid inthe sequence halts further chain elongation.Disaggregation of the intermediate and

heavy polyenzymes into subunits (89, 91, 95).Each ofthe polyenzyme components ofthe tyro-cidine-synthesizing system appears to consist ofsubunits and contains 1 mol of covalentlybound 4'-phosphopantetheine per mol of po-lyenzyme. On the basis of dissociation experi-ments, Lee et al. (89, 91) confirmed that theintermediate enzyme comprises three, and theheavy enzyme comprises six, amino acid-acti-vating subunits of the 70,000-dalton size. Thesubunits retained the ability to catalyze ATP-PPj exchanges and the thioesterification reac-tions with the appropriate amino acid sub-strates, but were unable to carry out the polym-erization step. Upon dissociation, both polyen-zymes also yielded a pantetheine-containingprotein fraction with a molecular weight of17,000 (77, 89, 91, 95). Evidence that 4'-phos-phopantetheine participates in the polymeriza-

tion reaction was provided also by the dissocia-tion experiments. After incubation of lysates ofB. brevis or purified polyenzymes with aminoacid substrates, the polyenzymes (from thecrude or purified enzyme system) were de-graded into subunits and subsequently puri-fied. In both cases, a product containing bothpantothenic acid and a nascent tyrocidine pep-tide was recovered (89, 95). These studies dem-onstrate that the intermediate and heavy en-zymes of tyrocidine synthetase consist of spe-cific amino acid-activating subunits (molecularweight 70,000) and a pantetheine-bearing pro-tein of 17,000 daltons which functions in thebiosynthesis of the nascent tyrocidine peptidechain (77).

LINEAR GRAMICIDINSThe gramicidins are linear pentadecapep-

tides produced by the same strain of B. brevisthat elaborates the tyrocidines (Fig. 8). Grami-cidins A, B, and C are a mixture of valine andisoleucine gramicidins (134, 137). They differonly in the aromatic amino acid (tryptophan,tyrosine, or phenylalanine) incorporated at po-sition 11 of the peptide. The amino terminal L-valine (or L-isoleucine) is formylated and thecarboxyl terminal tryptophan is linked to theamino group of ethanolamine. It should benoted that the peptide molecule consists almostentirely of alternating L and 1-amino acid resi-dues.Bauer et al. (7) were able to achieve a partial

biosynthesis of linear gramicidin, using a solu-

COMPOSITION OF GRAMICIDIN

Positions

Compd. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

I Valine-gramicidin A:

HCO-L-Val-Gly- L -Ala-D-LeU-L-Ala-D-Val-L-Val-D-Val- L-Trp-D-LeU-L-Trp-D-LeU- L-Trp-D-LeU-L-Trp-NHCH2CH20H

11 Isoleucine-gramicidin A:

HCO-L-lIeu-Gly-L-Ala-D-Leu-L-Ala-D-Val- L-Val-D-Val-L-Trp-D-Leu- L-Trp-D-Leu- I-Trp-D-Leu-L-Trp-NHCH2CH20H

III Valine-gramicidin B:HCO-L-Val-Gly- L-Ala- D-Leu-L-Ala-D-Val-L-Val-D-Val- L-Trp-D-Leu-L-Phe-o-Leu- -Trp- D-Leu-L-Trp-NHCH2CH20H

IV Isoleucine-gramicidin B:

HCO-L-Ileu-Gly-L-Ala-D-Leu-L-Ala-D-Val- L-Val-o-Val- L-Trp-D-Leu-L-Phe-D-Leu-L -Trp- D-Leu-i-Trp-NHCH2CH20H

V Valine-gramicidin C:

HCO-L -Val-Gly- L -Ala-D-LeU- L-Ala-D-Val- L-Val-D-Val- L-Trp-D-LeU- L-Tyr-D-LeU- L Trp-D-Leu- L-Trp-NHCH2CH20H

VI Isoleucine-gramicidin C:

HCO-L-IIeu-GIy-L-AIa-D-Leu-L-AIa-D-Val-L-Val-D-Val-L-Trp-D-LeU-L-Tyr-D-LeU-L -Trp- D-Leu-L-Trp-NHCH2CH20H

FIG. 8. Composition of linear gramicidins A, B, and C.

BACTZRIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from



ble enzyme system obtained from extracts ofB.brevis that had been purified through ammo-nium sulfate fractionation and Sephadex G-200column chromatography. The location of thelinear gramicidin-synthesizing enzymes wasfollowed by their amino acid-dependent ATP-32pp1 exchange activities. The fractions thatactivated the amino acids incorporated into lin-ear gramicidin were devoid ofamino acid-tRNAligase activity. However, due to its instability,further resolution of the linear gramicidin syn-thetase into complementary enzyme fractionswas not attempted. Despite these difficulties,the enzyme system catalyzed the synthesis ofan enzyme-bound pentadecapeptide. Chemicalethanolaminolysis liberated the peptide fromthe enzyme, and subsequent chemical formyla-tion of the N-terminal valine yielded a com-pound that appeared to be chromatographicallyidentical to authentic linear gramicidin.Bauer et al. (7) noted that enzymatic formyla-tion of the N-terminal valine could only beachieved with crude extracts derived from B.brevis. Unfortunately, attempts to add ethanol-amine to the carboxyl terminal tryptophan en-zymatically to complete the biosynthesis of lin-ear gramicidin were unsuccessful.

All the amino acid precursors of linear gram-icidin can be incorporated into a protein-boundthioester-linked peptide which proved to be sta-ble to acid, but liberated by alkali (pH 11) orperoxidation (7). The stoichiometry of theamino acids incorporated into the peptide asdeduced from double-labeling experiments was1 Gly:2 Ala:4 Leu:4 Val:4 Phe in agreementwith the stoichiometry of the amino acids inlinear gramicidin. Alanine was found only inthe L-configuration, and valine was presentequally as D and L, whereas leucine was only

C2H5

\ ~~~~SOH-CH

CH3 NH2

found in the 1-configuration, which correlatesrather well with the optical configuration oftheamino acids in linear gramicidin. These resultsclearly suggest that a pentadecapeptide whichremains thioester-linked to the enzyme is bio-synthesized. The peptide may be released en-zymatically by aminoethanolysis. The stageduring biosynthesis at which formylation ofN-terminal valine occurs remains uncertain.However, Bauer et al. (7) suggested thatformylation (presumably requiring tetrahydro-folate) most likely is carried out after synthesisof the polypeptide is completed. The participa-tion of 4'-phosphopantetheine as a cofactor inlinear gramicidin formation is anticipated, butthis point has not been established to date.

BACITRACINSThe bacitracins are synthesized by certain

strains ofB. licheniformis (167). Although sev-eral bacitracins have been described in the lit-erature, the most completely investigated isbacitracin A (Fig. 9), a dodecapeptide with fourof the amino acids (glutamic, ornithine, phen-ylalanine, and asparagine) in the )-configura-tion (1, 22, 99, 100). As already noted, the mole-cule contains a cyclic hexapeptide and a thiazo-line ring structure.

Several laboratories have reported the cell-free synthesis of bacitracin (30, 31, 56-58, 118,148, 149). Purified bacitracin synthetase (mo-lecular weight 800,000) has been resolved intothree complementary fractions by means ofSepharose affinity column chromatography (30)or a combination of hydroxyapatite columnchromatography and sucrose density gradientcentrifugation (56). It appears that all threefractions are required for maximal bacitracinsynthesis; a combination of any two compo-

Wv .L-L~U

L-His D-AspNH/~~~~~~~~D-Phe L-Asp

I 16 2 0

L- I leu L-Lys- L-I le

D-OrnFIG. 9. Bacitracin A.

VOL. 41, 1977


http://mm

br.asm.org/

Dow

nloaded from


458 KATZ AND DEMAIN

nents is less than 25% as active as the threefractions together. The enzyme in peak A (des-ignated component I) has a molecular weight of200,000 and activates L-isoleucine, L-cysteine,L-leucine, and L-glutamic acid. A second frac-tion, B (component II), has a similar molecularweight (210,000) but only activates L-lysine andL-ornithine, whereas the third fraction, C (com-ponent III), possesses a molecular weight of380,000 and activates L-phenylalanine, L-his-tidine, L-aspartic acid, L-asparagine, and L-iso-leucine (or L-valine). Amino acid activation re-quires ATP and Mg2+, and it has been notedthat aminoacyl adenylates are formed; subse-quently, the amino acids are thioester-boundto the enzyme before polymerization occurs onthe polyenzyme. Froyshov and Laland (31) firstpresented evidence that pantothenic acid ispresent in bacitracin synthetase and concludedthat 4'-phosphopantetheine is bound to thesynthetase. Recently, Ishihara et al. (56) ex-tended this observation and reported that eachofthe three components may contain one equiv-alent of covalently bound 4'-phosphopante-theine. These components are relatively impure(40 to 70%), so that the results should be con-sidered as somewhat preliminary in nature.

Earlier studies with intact cells, protoplastpreparations, and in vitro systems indicatedthat 1)-amino acids could not replace their re-spective L-enantiomorphs as precursors for bac-itracin formation (11, 21, 56, 57, 118, 151). How-ever, studies with the partially purified enzymepreparation revealed that the L-enantiomorphof glutamic acid, phenylalanine, and aspara-gine can be replaced, to a considerable extent,by the 1)-isomers (glutamic acid, 27%; phenylal-anine, 79%; asparagine, 81%) for antibiotic syn-thesis (31). By contrast, D-ornithine could notsubstitute for its L-isomer. It will be of consider-able interest to learn the detailed mechanism ofbacitracin biosynthesis as there are obvious dif-ferences as well as similarities between thestructure of this molecule and those of gramici-din S and tyrocidine.

EDEINEThe edeines (Fig. 10) are strongly basic, lin-

ear oligopeptide antibiotics produced by B.brevis strain Vm4 (48, 50, 82, 125). The amino

acid composition is unusual; besides glycine,the molecule consists of isoserine, f3-tyrosine,a,,8-diaminopropionic acid, and 2,6-diamino-7-hydroxyazaleic acid. The glycine residue in ed-eine A is attached to the basic constituent,spermidine. In the case of edeine B, the base isguanylspermidine (N-guanyl-N'-(3-aminopro-pyl)-1,4-diaminobutane). The structure of ed-eine D is similar to edeine A with f3-phenyl-,/-alanine substituting for f-tyrosine (169).Although a number of papers have appeared

on the cell-free synthesis of edeine (82, 84-86),the mechanism of its biosynthesis is not fullyunderstood. Polyenzyme fractions that are spe-cific for the activation and polymerization ofthe constituent amino acids ofedeine have beenisolated from B. brevis Vm4 by Kurylo-Bo-rowska (85). The size and number of fractionsthat are resolved by DEAE-cellulose columnchromatography and Sephadex G-200 sievingappear to vary with the age of the culture andmethod of extract preparation (e.g., lysozymetreatment, grinding). However, only two po-lyenzyme fractions (designated A and B) areneeded for edeine formation in the presence ofconstituent amino acids, spermidine (or guan-ylspermidine), ATP, and Mg2+. Fraction C,which may be a disaggregation product of themore complex polyenzyme B fraction, enhancesthe formation of edeines A and B. The molecu-lar weight of these fractions is reported to be210,000 (fraction A), 180,000 (fraction B), and100,000 (fraction C) (79). Fraction A activatesonly f8-tyrosine; fraction B catalyzes the activa-tion of the other four constituent amino acids ofedeine (isoserine, 2,6-diamino-7-hydroxyazaleicacid, a,,8-diaminopropionic acid, and glycine).Fraction C activates isoserine and diaminopro-pionic acid as well as certain nonconstituentamino acids such as L-a- and L-I3-alanine and L-leucine. Fractions A and B also catalyze ATP-AMP exchanges, suggestive ofsecondary accep-tor sites for the amino acids on the polyenzymecomplexes. Fractions A and B were shown tocontain covalently bound pantetheine. There isno information concerning the sequence of addi-tion of amino acids during nascent peptide syn-thesis, nor are there any data regarding themechanism of attachment of the base to glycineduring edeine formation. The large size of frac-

E I N E

1-TYR-r-SER-DAPA-DAHAA-GLY-(GUANYL) SPERMI DINEFIG. 10. Structure of edeines A and B. Abbreviations: S-TYR, ,3tyrosine; /3-SER, ,&serine (isoserine);

DAPA, diaminopropionic acid; DAHAA, diaminohydroxyazelaic acid; GLY, glycine. In edeine B, guanyl-spermidine substitutes for spermidine present in edeine A.

BACTERIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from


VOL. 41, 1977

tion A is surprising in light of the fact that itonly activates f8-tyrosine. A separate enzyme,tyrosine a,,3-amino mutase (molecular weight75,000) catalyzes the conversion of L-a-tyrosineto L,-(-tyrosine (84).

MYCOBACILLIN

Mycobacillin (Fig. 11) is a cyclic peptide anti-biotic that contains 13 residues of 7 differentamino acids (6, 103, 144). The sequence andlinkages of the amino acids in the moleculewere recently elucidated (103, 144). Myco-bacillin lacks free amino groups but con-tains two free a-carboxyl groups. All of theaspartic acid residues of the antibiotic havebeen shown to be in alpha-peptide linkage; theglutamic acid residues (1-isomers), however,are gamma-linked to the adjacent amino acid.Sengupta et al. (144) demonstrated that both L-aspartic acid (position 5) and 1)-aspartic acid(positions 2, 8, 11, and 13) are present.A 10,000 x g supernatant fraction obtained

from the protoplast lysate ofB. subtilis B3 was


reported to catalyze the synthesis of the anti-biotic (145). The process required ATP and wasinsensitive to the action of RNase and inhibi-tors of protein synthesis such as streptomycin,chloramphenicol, and puromycin.There appear to be certain features which

distinguish mycobacillin formation from thebiosynthesis of other peptide antibiotics (gram-icidin S, tyrocidine, bacitracin, edeine). For ex-ample, the activation step may involve an ATP-inorganic orthophosphate (Pi) exchange reac-tion. Sengupta and Boge noted that such anexchange occurred with irproline, a constituentamino acid of mycobacillin (146). This exchangewas further stimulated by the other amino acidcomponents of the antibiotic. Omission of cer-tain of the constituent amino acids (D-asparticacid, D-glutamic acid, L-tyrosine, and L-serine)diminished the stimulation observed. Pyro-phosphate, an inhibitor of gramicidin S forma-tion, stimulated mycobacillin synthesis in vitro.It was also reported that several peptides weresynthesized as intermediates in the biosynthe-sis of the antibiotic in vitro (146). Peptides

a

D-Aspartic acid L-Proline

L-Alanine D-Asp

al

D-Aspartic acid

D-Glutamic acid

a\

D-G

artic acid

;lutamic acid

-Tyrosine

|artic acidi-Leuci ne L-Aspai

D-Aspartic acid L-Tyrosine

FIG. 11. Structure of mycobacillin.


http://mm

br.asm.org/

Dow

nloaded from


460 KATZ AND DEMAIN

containing two, three, four, five, and sixdifferent amino acids were formed after in-cubation with proline plus one to six additionalamino acid components of the mycobacillinmolecule. No peptide containing all seven dif-ferent mycobacillin amino acids was observed,nor was any straight-chain tridecapeptide ob-tained, suggesting that cyclization may occurimmediately after such a peptide is synthe-sized. Characterization of the peptides revealedthat each compound had proline as its N-termi-nal residue and, depending on the peptide inquestion, aspartic acid, glutamic acid, or tyro-sine as its C-terminal amino acid. The peptideswere di-, tri-, hexa-, octa-, and undecapeptides;however, the total sequence as well as the ster-eochemistry of the constituent amino acids inthese peptides was not established. The partici-pation of such compounds as intermediates inthe biosynthesis of mycobacillin is certainlysuggested by the data presented, yet very littleis known concerning the detailed mechanism ofsynthesis of this antifungal antibiotic, particu-larly with regard to the substrate and cofactorrequirements, the nature of the activation, ini-tiation, elongation reactions, and the cycliza-tion of the nascent linear mycobacillin peptide.It would be of interest to know whether, duringantibiotic synthesis, the peptide intermediatesare enzyme-bound or free. Also, the nature andproperties of the polyenzyme systems that cata-lyze the polymerization reactions are not wellknown. The answer to these as well as otherquestions pertaining to the biosynthesis of my-cobacillin must await additional purification ofthe enzyme system(s) involved.

POLYMYXIN AND COLISTINPolymyxin is produced by strains of Bacillus

polymyxa and the related colistins are elabo-rated by Bacillus colistinus Koyama (115, 155-157, 163). Structurally similar to the poly-myxins and colistins are the circulins (115).These antibiotics are branched, cyclic decapep-tides linked to a fatty acid residue. For exam-ple, 6-methyloctanoic acid or isooctanoic acid,present in colistins A and B, respectively, ispresent in a-amide linkage with the terminala,y-diaminobutyric acid residue (155, 157). Thestructure of colistin B is presented in Fig. 12. Itis of interest to point out that 6-methyloctanoicacid and isooctanoic acid are synthesized fromisoleucine and valine, respectively (59). Twocompounds, N-6-methyloctanoyl-a,-diamino-butyric acid and Na-isooctanoyl-a,y-diamino-butyric acid, formed by colistin-producing cells,have been implicated as intermediates in thebiosynthesis of colistins A and B, respectively

L-Leu

D-Leu

L - DA B - N H

-*i-DAB-NH2y

L-DAB-NH2y

12r L-Thr

\. (y)-DAB

(a)4t

L -DAB -NH2Y

L-T h r

L -DAB -NH2Y

IOAFIG. 12. Structure of colistin B. Abbreviations:

IOA, isooctanoic acid; DAB, diaminobutyric acid;Thr, threonine; Leu, leucine. In colistin A, the fattyacid present is 6-methyloctanoic acid.

(60). The cell-free synthesis of colistins A and Bhas been achieved using soluble extracts ob-tained from B. colistinus and an incubationmixture consisting ofNl-6-methyloctanoyl-a,.y-diaminobutyric acid (and Na-isooctanoyl-a,y-diaminobutyric acid), leucine, a,,y-diaminobu-tyric acid, [14C]threonine, ATP, and Mg2+ (61,62). The radioactive product formed was identi-fied as colistin by paper chromatography. Thesestudies suggest that the fatty acyldiaminobu-tyrate intermediate may initiate colistin bio-synthesis. As observed earlier in the case ofpolymyxin formation in vivo (115), synthesis ofcolistin in vitro was insensitive to chloram-phenicol and puromycin (62). Enzyme activityfor colistin synthesis was observed chiefly inthe soluble fraction (100,000 x g supernatantfluid) of the cell. Additional evidence that thepolymerization process is distinct from proteinsynthesis was the finding that antibiotic syn-thesis was not inhibited by RNase (62).

BACTERIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from



An activating enzyme which may function inpolymyxin biosynthesis was partially purifiedfrom sonic extracts of the polymyxin B-produc-ing organism, B. polymyxa 2459 (66). The en-

zyme catalyzes an L-ay-diaminobutyrate-de-pendent exchange ofATP with PP1. The specificactivity of this enzyme increased approxi-mately 10-fold during vegetative growth, whichcorrelated with the appearance of the antibioticin the growth medium. .-a,y-Diaminobutyrateactivation was also noted in extracts ofB. poly-myxa ATCC 10401 and B. circulans 14040, pro-

ducers of polymyxin D and circulin, respec-

tively. By contrast, enzyme activity was absentfrom extracts prepared from several polymyxin-negative and asporogenic mutants derived fromB. polymyxa. Although these data certainlysuggest that the activating enzyme plays a

role in polymyxin, circulin, and possibly colis-tin formation, its direct participation has notbeen established. No system for the cell-freesynthesis of polymyxin has been described, noris there any evidence that the diaminobutyrate-activating activity is present in extracts of thecolistin producer, B. colistinus.

SPECIFICITY OF INCORPORATION

In vivo (102, 112, 130, 168) as well as in vitro(30, 33, 34, 37, 58, 63, 93, 97, 120, 128, 153)experiments have revealed that structurallysimilar amino acids and amino acid analoguescan replace certain of the amino acid residues ofthe peptide chain during antibiotic synthesis.These studies have demonstrated that anti-biotic formation is catalyzed by enzymes with

rather broad specificities. Multiple forms of anantibiotic are generally produced simultane-ously by an organism. These often differ only inthe aromatic amino acid or branched-chainamino acid residue that is present at a givensite in the peptide. Supplementation with oneof these amino acids in vivo or in vitro can

direct the biosynthetic process so that the anti-biotic(s) formed preferentially will be that richin the amino acid added exogenously. For ex-

ample, tyrocidines A, B, and C (differing solelyin the number of tyrosine, tryptophan andphenylalanine residues present) are producedin vivo in the ratio of 1:3:7 (102). When trypto-phan is added to the medium, B. brevis synthe-sizes principally tyrocidine C and a new tyroci-dine, designated D, both of which are rich intryptophan; neither tyrocidine A or B (phenyl-alanine-rich) is produced in significantamounts under these conditions. If phenylala-nine is supplied instead of tryptophan, there issynthesis of tyrocidines A and B without forma-tion of the C and D components. Provision ofboth aromatic amino acids leads to the synthe-sis of tyrocidines A, B, C, and D simultane-ously. Amino acid substitutions have also beennoted in experiments with the purified gramici-din S and tyrocidine synthetases (33, 34, 37, 63,93, 97, 120, 128, 153). Thus, tyrosine can replacephenylalanine (position 1) in both antibiotics,whereas tryptophan can substitute for phenyl-alanine or tyrosine only during formation ofthetyrocidine peptides in vitro (34, 128). Additionalexamples are listed in Tables 3 and 4. In some

instances there may be a replacement only atthe activation step (i.e., in the formation of the

TABLE 3. Amino acid and analogue substitutions observed with growing cells

Antibiotic Amino acid or analogue Comment Referenceadded to medium

Gramicidin S DL-Thienylalanine Replaces D-phenylalanine 168DL-Norleucine Replaces L-leucine to a small extent 168DL-F-phenylalanine Replaces D-phenylalanine to a small ex- 168

tent

Tyrocidine No addition Tyrocidines A, B, and C formed in ratio of 102, 1301:3:7

L-Phenylalanine Tyrocidine A synthesized almost exclu- 102, 130(D-phenylalanine) sively at expense ofB and C; tryptophan

replacedL-Tryptophan (i-trypto- Tyrocidine C and D produced; little or no A 102, 130phan) or B formed; phenylalanine and tyrosine

replacedL-Phenylalanine plus L- Tyrocidines A, B, C, and D synthesized 102, 130tryptophan

DL-Thienylalanine Substitutes for D-phenylalanine 112L-Alloisoleucine, L-iso- Substitutes for L-leucine and L-valine 112, 130

leucineL-Pipecolic acid Replaces proline to a slight extent 112

VOL. 41, 1977


http://mm

br.asm.org/

Dow

nloaded from


TABLE 4. Amino acid and analogue substitutions observed with cell-free systems

Antibiotic Amino acid replaced Analogue exchange Antibiotic synthesisn (%) 1

I echnga Atiioi 1efrecGramicidin S L-Phenylalanine

L-Phenylalanine

L-Phenylalanine

L-Prolineb c

L-Valine

L-Ornithine

L-Leucine

L-Phenylalanine

L-Phenylalanine

Aromatic aminoacidsd

D-PhenylalanineL-TyrosineL-Tryptophanp-F-L-phenylalanineDL-Threo-/,-phenylserineDL-,3-Thienylalanine

L-TryptophanL-Tyrosine

D-PhenylalanineL-Tyrosine

Hydroxy-L-prolineL-Azetidine-2-carboxylic

acidThiazolidine-4-carbox-

ylic acidL-Pipecolic acidTrans-3-methylprolineTrans-4-fluoroproline

D-ValineL-IsoleucineD-IsoleucineL-AlloisoleucineL-Norvaline

L-LysineL-ArginineD-Ornithine

D-LeucineL-IsoleucineD-IsoleucineL-AlloisoleucineL-Norleucine

L-TryptophanL-Tyrosine

DL-Threo-/3-phenylserineDL-,3-2-ThienylalanineDL-p-F-PhenylalanineL-TryptophanDL-5-MethyltryptophanDL-p-F-tryptophanStandard incubationmixture with L-trypto-phan (100 ,umol)Standard incubationmixture with L-trypto-phan (1/7 above)Standard incubationmixture minus L-trypto-phanStandard incubationmixture minus L-tyro-sineStandard incubationmixture minus L-trypto-phan and L-tyrosine

1001003543956989

10071103100

064

<1

13271

1001, <1

189

1732100

22

<1100

19, <120, 16101923, 4

1001028250

2997723

100105

10

940, 5

25100

0108

1000

Replaces proline

1000

480

58105100

00

100

0390

38221003595

1001

3519875233

Tyrocidine D is majorcomponent; small amountof tyrocidine CTyrocidines A, B, and C;smaller amount of D pro-ducedTyrocidine A formed

Tyrocidine D formed

Tyrocidine E synthesized(4-phenylalanine residues;no tyrosine or tryptophan)

74, 97

128

37

93, 97153

93, 97

93, 97

74, 93,97

128

34

34

462

Gramicidin S

Gramicidin S

Gramicidin S

Gramicidin S

Tyrocidine


http://mm

br.asm.org/

Dow

nloaded from



TABLz 4-ContinuedATP-PPj

Antibiotic Amino acid replaced Analogue exchan- Antibiotic synthesis' () Reference~~~~gea

Tyrocidine L-Leucine Isoleucine Isoleucyl-tyrocidines 34L-Leucine 100 100

DL-Norleucine 73 11Tyrocidine L-Ornithine Lysine Lysyltyrocidines 34

L-Ornithine 100 100L-Diaminopimelic acid 215 0L-Diaminobutyric acid 5 4

Tyrocidine L-Proline 100 100 34Hydroxy-L-proline 7 8

Tyrocidine i-Valine 100 100 34DL-Norvaline 90 5

aR 100% is the enzyme activity or antibiotic synthesis obtained with the natural amino acid.I Trans-3-methylproline inhibits L-proline-dependent ATP-PP1 exchange by 29%, whereas under similar

conditions trans-4-fluoroproline inhibits the reaction by 93% (93).c The ATP-PP1 exchange reaction with the Jramino acid was inhibited by the corresponding D-enantio-

morph as follows: D-valine (80%), D-ornithine (82%), D-leucine (86%), and D-proline (83%) (93).d But see Rao and Hall (120) for strain differences (B. brevis ATCC 8185 versus 10068) with respect to

effect of aromatic amino acids on cell-free synthesis of tyrocidines A, B, C, and D.

aminoacyl adenylate), suggesting that a greaterdegree of specificity exists at the point of inter-action between the aminoacyl adenylate andthe thiol group on the enzyme (97).

RACEMIZATION AND THE SYNTHESISOF D-AMINO ACIDS

n-Amino acids have been found as constitu-ents of microbial cell walls (142), peptidolipids(3), capsules (161), toxins (46), and antibiotics(14, 20). Racemases for various amino acidshave been observed in microorganisms (110);however, the biosynthetic mechanism for the D-amino acid residues in most of these peptidemetabolites remains unclear. In some cases,both the L- and D-enantiomorph of an aminoacid can be utilized for synthesis of the peptide-bound n-amino acid (71, 160). In other in-stances, there is evidence to suggest that onlythe L-isomer is used for the biogenesis of theamino acid in the peptide structure (11, 26,151). Troy recently reported that a particulatefraction from an encapsulated strain of B. ii-cheniformis catalyzed the polymerization of L-glutamic ascid (but not the 1)-isomer) to form ahigh-molecular-weight poly(a)-D-glutamic acid(161). It has also been observed that the >-amino acid may actually inhibit peptide syn-thesis, such inhibition being reversed by the irenantiomorph (26).

Several hypotheses have been advanced toexplain the biogenetic origin of 1)-amino acidsin microbial peptides. Mauger (108) has postu-lated that 1)-amino acids in antibiotics areformed from L-amino acids after incorporationof the latter into stereochemically labile inter-

mediates such as cyclic dipeptides (e.g., diketo-piperazines). Bycroft (20) proposed that a com-bined (presumably enzyme-bound) form of adehydroamino acid derived from the corre-sponding L-amino acid might be reduced ster-eospecifically in vivo to the 1-isomer duringantibiotic formation. These hypotheses havenot yet been substantiated experimentally. Infact, studies by Huang et al. (54) and Mason etal. (105) indicate that a dehydroamino acid doesnot participate as an intermediate in the inver-sion of i-valine to its D-enantiomorph duringthe respective syntheses of penicillin and acti-nomycin. By contrast, it now seems certainwith respect to tyrocidine and gramicidin Sthat, after the activation step, racemization ofphenylalanine takes place on an enzyme-bound, thioester-linked intermediate (38, 81,97, 128). Purified light enzyme obtained fromeither the gramicidin S or tyrocidine synthe-tase and the intermediate enzyme (tyrocidinesynthetase) catalyze the racemization reactionwith ATP and ir or D-phenylalanine as sub-strate (38, 128, 160). The phenylalanine residueat position 1 of the gramicidin S or tyrocidinemolecule, as well as the phenylalanine at posi-tion 4 of the tyrocidine peptide, is incorporatedas the 1-amino acid. Recently, Lipmann (95)observed that racemization ofphenylalanine la-beled with tritium at the a-carbon takes placewith complete loss of hydrogen. Details of thisexperiment have not been published, so themechanism of the reaction remains somewhatobscure. Likewise, the cell-free synthesis of n-amino acids present in other peptide antibioticsproduced by bacilli is not well understood. Ra-

VOL. 41, 1977


http://mm

br.asm.org/

Dow

nloaded from


464 KATZ AND DEMAIN

cemization of amino acids may proceed via an

analogous mechanism (38, 95).

MUTANT STUDIESMutants have frequently provided a useful

means to investigate biosynthetic mechanismsand to determine the genetic control of metabo-lite formation. In this regard, mutants havebeen obtained that are unable to synthesize a

number of peptide antibiotics (e.g., gramicidinS [B. brevis] [64, 68, 81, 147], bacitracin [B.licheniformis] [43, 44], polymyxin [B. poly-myxa] [66], and mycobacillin [B. subtilis][121]). Certain of these mutants appear to havedefects in structural genes, whereas others maybe affected in regulatory sites of the genome

which control antibiotic synthetase formation.Kurahashi et al. (68, 81) and Saito and co-

workers (64, 147) characterized several classesof mutants of B. brevis which are unable tosynthesize gramicidin S. A summary of certainof these mutants is presented in Table 5. Mu-tants of group I were shown to lack the lightenzyme, but possessed a normal heavy enzyme.

Group II mutants had a normal light enzyme,but did not contain an L-proline-, L-valine-, L-ornithine-, and L-leucine-activating enzyme ac-tivity. Group III mutants lacked both the heavyand light enzyme fractions. As might be ex-pected, cell-free extracts of these mutants were

unable to synthesize gramicidin S and could notform D-phenylalanyl-L-prolyl diketopiperazine(i.e., could not initiate peptide synthesis on theheavy enzyme). On the basis of these results,Kambe et al. (68) suggested that different typesof mutations may occur, i.e., mutations affect-ing the regulatory mechanism controlling bothlight and heavy enzyme synthesis as well as

those mutations affecting the regulatory or

structural genes of the light or heavy enzyme,individually.

Other types of gramicidin S-less mutantshave been described (68, 147). A fourth groupappears to have a normal phenylalanine-acti-vating enzyme, but an incomplete or defectiveenzyme I complex. Characterization of mutantsin the fourth group has revealed that the heavyenzyme complex is missing one specific activat-ing enzyme (e.g., for L-proline, L-valine, or L-leucine), but is still able to activate the remain-ing three amino acids. The fifth group of mu-tants retained all of the amino acid-activatingenzymes needed for gramicidin S formation,but still cannot synthesize gramicidin S. Thedefect in mutants of groups IV and V may bedue to (i) the absence ofa subunit that catalyzesthe activation of a specific amino acid, (ii) adefect only in the secondary amino acid accep-tor site, i.e., the thioester-binding site, or (iii) aconformational change [possibly due to (i)] inthe heavy or light enzyme that results in re-duced complex formation between the twopolyenzyme components. It has also been pro-posed that other mutational events may in-fluence the interaction and association of thesubunits of the heavy enzyme or the formationof the phosphopantetheine arm (68).Haavik and Froyshov (43) recently isolated a

mutant, obtained from B. licheniformis AL,that was unable to produce bacitracin as a re-sult of a defective enzyme complex. As de-scribed earlier (30, 56), the parental strain pro-duces a bacitracin synthetase that can be sepa-rated into three complementary fractions, A, B,and C; all three components are needed forbacitracin synthesis. Only two fractions (ACoand Bo) were obtained from the mutant, B.licheniformis SB319 (43). Component ACo acti-vated the same amino acids as components Aand C from the parent. Bo activated the sameamino acids as the parental B fraction. How-ever, there were marked differences seen in the

TABLE 5. Properties ofsome gramicidin S-less mutants ofBacillus brevisAmino acid-activating enzyme for: Ability to form:

Mutant classm Phenylala- GramicidinMcnine iL-Proline L-Valine L-Ornithine L-Leucine S

DKpa

None, wild type + + + + + + +I - + + + + _ _

II + - - - - - -

III - - - - - - -

IV + - + + + - -+ + + + ++ + + + --++ + + +___

V + + + + + - ++ ++d+p+

ai DKP, i-Phenylalanyl-i>prolyl diketopiperazine.

BACTECRIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from



degree of the ATP-PP, exchange reaction withthe enzyme fractions derived from the producerand nonproducer cells. Mixing experiments re-vealed that enzyme component B0 could cata-lyze the synthesis of bacitracin with compo-nents A and C; by contrast, component ACO,when used with component B, was unable tosynthesize the antibiotic. The evidence, there-fore, suggests that the bacitracin synthetaseobtained from the nonproducer was defectiveand that the lesion may involve the activationsites for the amino acids. The affinity betweenthe enzyme components may also be affected.

POSSIBLE FUNCTIONS OF PEPTIDEANTIBIOTICS IN THE PRODUCING

ORGANISMThe function of antibiotics in the metabolism

of the producing organism has been the subjectof considerable speculation and discussion (17,18, 164, 166, 170). However, there has beenlittle understanding with respect to the role, ifany, that they actually do play. Several pro-posed functions for antibiotics are no longeraccepted as likely candidates, whereas othersare still under consideration in several labora-tories. Those that have been discarded involveantibiotics as evolutionary relics, waste prod-ucts of cellular metabolism, reserve food mate-rials, spore coat components, or breakdownproducts derived from cellular macromolecules(166). Still under current consideration (40) isthe possibility that antibiotics function to killor inhibit the growth of other organisms innature, thereby providing a competitive advan-tage to the producing species (17). Although oneusually thinks of this competition as betweenbacterial species A versus bacterial species B orbacterium versus fungus, it could involve bac-teria versus amoebae (S. H. Hutner, per-sonal communication) since these protozoa usebacteria as sources of food (150). Another possi-bility is that the competition exists betweenstrains of the antibiotic-producing species itself(101; E. Rosenberg, personal communication).A further variation of the competition hypothe-sis involves the excretion of the antibiotic dur-ing spore germination in order to eliminatecompetitors in the immediate environment ofthe germinating spore. A completely differentconcept, proposed by Bu'Lock (18), is that anti-biotic synthesis is a means of keeping the cellu-lar machinery in working order during the timewhen cell growth is not possible due to unfavor-able conditions. Although this is an interestingspeculation, no data exist in support of themaintenance hypothesis. Additional hy-potheses currently being debated state that

synthesis of an antibiotic (or other secondarymetabolites) is a method of avoiding cell deathdue to unbalanced growth (166), or that it pro-vides a mechanism of detoxification (170). Withrespect to the unbalanced growth hypothesis, itis assumed that certain primary metabolites(e.g., amino acids or nucleotides) are overpro-duced, when conditions for exponential growthare no longer favorable, due to ineffective con-trol systems in the producing organism. Thesemetabolites are thought to be converted to anti-biotics which are released from the cell. Thedetoxification hypothesis is slightly different,proposing that certain toxic antimetabolites aresynthesized when exponential growth termi-nates, e.g., analogues of nucleosides, aminoacids, or amino sugars. Conversion to the anti-biotic is thought to provide a method of detoxifi-cation. Inherent in both the unbalanced growthand detoxification proposals is the assumptionthat the antibiotic is not toxic to the producingorganism (170). There exists no evidence insupport of these two hypotheses. Indeed, mostantibiotics are more toxic to the producing or-ganism than are their precursors (24). Otherpossibilities for antibiotic function include (i)facilitation of metal transport into cells (43) and(ii) autoinhibition of germination (29). The lat-ter is suggested by the following observations:many microorganisms produce self-inhibitors ofgermination (154); various antibiotic producerscontain the antibiotic in their spores (27, 29,67).Much of the criticism against the competition

hypothesis is based on the inability of certaininvestigators to detect antibiotic production inunsterilized soil which has not been supple-mented with organic nutrients (164). The sametype of criticism has also been used to supportthe viewpoint that antibiotics are biological ac-cidents which are of no direct benefit to the cell.However, it is known that soil which containsseeds and decaying vegetable matter and therhizosphere support antibiotic producion (17,19, 40, 78).Of the various functions postulated for anti-

biotics, the one which has received the mostattention in recent years is the view that anti-biotics are important compounds in cellular dif-ferentiation, i.e., the transition from vegetativecells to spores (52, 55, 132, 139). Numerous anti-biotics are elaborated by microorganisms thatundergo sporulation. Due to the fact that thevast majority produced by bacilli are peptidic innature, these peptides have been singled out asplaying a key role in the termination of vegeta-tive growth, allowing sporulation to take place.The following factors make this hypothesis at-tractive (24, 45). (i) Practically all sporulating

VOL. 41, 1977 465


http://mm

br.asm.org/

Dow

nloaded from


466 KATZ AND DEMAIN

microorganisms produce antibiotics. (ii) Al-though it was once believed that the antibioticselaborated by an organism have no effect on theproducing cell, it is now well established thatthey are inhibitory to vegetative growth at con-centrations at which they are generally pro-duced by sporulating cultures (65, 123, 138,152). Peptide antibiotics produced by bacillispecifically inhibit important cellular proc-esses, e.g., deoxyribonucleic acid (DNA) syn-thesis, cell wall synthesis, membrane function,and structure (52, 132). (iii) Production of pep-tide antibiotics usually begins at the late-loga-rithmic phase of growth and continues duringthe early stages of the sporulation process inbacilli (5, 10, 11, 36, 82, 115, 138, 160). (iv)Sporulation and antibiotic synthesis are in-duced by depletion of some essential nutrient.(v) There are genetic links between the synthe-sis of antibiotics and the formation of spores(24, 45, 139). For example, antibiotic formationby certain bacilli is an early event beginning atstage I of sporulation. Mutants that are asporo-genous (and map at different loci) are almostalways antibiotic negative and, when so, areblocked very early (66, 68, 143). Revertants,transductants, and transformants of stage 0 as-porogenous mutants, restored in their ability tosynthesize antibiotic, also regain the ability tosporulate (4, 51, 140). Also, conditional asporo-genous mutants fail to produce antibiotic at thenonpermissive temperature (92, 158). (vi) Phys-iological correlations also favor a relationshipbetween the production of an antibiotic andspore formation. As an example, inhibitors ofsporulation also inhibit antibiotic synthesis (10,65, 115). Furthermore, both processes are re-pressed by glucose (10, 139), and concentrationsof manganese ion of at least two orders of mag-nitude higher than that required for normalcellular growth are needed for sporulation andantibiotic synthesis by certain species ofBacil-lus (165).

Because of the above observations, severalinvestigators have concluded that antibioticproduction is required for sporulation. Hodgson(52) speculated that peptide antibiotics mightbe used in several ways by an organism duringthe process of sporulation as modifiers of thecell membrane, e.g., as detergents disruptingstructural components, or as ion carriers, mod-ifying permeability properties. Hodgson arguedthat, by selectively functioning at certainstages, the sporulation process would be able toproceed normally. Sadoff (132) pointed out thatthe peptide antibiotics produced by bacilli ex-hibit effects upon membrane synthesis andfunction, cell wall synthesis, and nucleic acidsynthesis, which are important processes forsporulation. He contended that the antibiotics,

BACTERIOL. REV.

when produced by an organism, may act asselective modifiers of cell function-i.e., re-pressing or inhibiting vegetative cell macromo-lecular synthesis, but permitting endospores tobe formed.Although the above observations and conclu-

sions point to an intimate relationship betweenantibiotic formation and sporulation, by nomeans do they prove that the antibiotic is nec-essary for sporulation. Furthermore, there aredata which fail to support this hypothesis. Forexample, there are conditions in which in-creased sporulation is accompanied by a de-crease in antibiotic production (143) and viceversa (11). Recent nutritional studies on grami-cidin S formation by B. brevis revealed a disso-ciation between sporulation and antibiotic for-mation (25). Media used at the beginning ofthe investigation supported sporulation butyielded little to no gramicidin S. As more suita-ble media were devised for antibiotic produc-tion, sporulation became markedly poorer. An-other indication of a physiological dissociationbetween gramicidin S formation and sporula-tion arose from chemostat experiments inwhich nutrient limitation was used to studyformation of the light and heavy gramicidin Ssynthetases (106). The enzymes were producedat intermediate growth rates under carbon, ni-trogen, sulfur, or phosphorus limitation. Al-though carbon limitation yielded the poorestsynthetase activity, it was the only conditionunder which significant amounts ofspores wereformed. Also, the growth rate favoring sporula-tion was considerably lower than that whichwas optimal for gramicidin S synthetase forma-tion.The most damaging evidence to the hypothe-

sis involving antibiotics in sporulation is theexistence of mutants which are antibiotic nega-tive but can still sporulate. For example,Kambe et al. isolated gramicidin S-negativemutants ofB. brevis with genetic blocks in theformation of the heavy synthetase, the lightsynthetase, or both (68). Upon examining thesporulating abilities of the gramicidin S-nega-tive mutants, Kambe et al. (68) found them allto be oligosporogenous. However, recent stud-ies on mutant hh, reportedly containing a spe-cific defect in the leucine-activating function ofthe heavy synthetase (68), revealed that thisgramicidin S-negative mutant sporulates nor-mally (25).The ability ofpeptide antibiotic-negative mu-

tants to sporulate is also seen with other bacilli.Adenine-requiring mutants of B. subtilis B3,which fail to produce intra- or extracellularmycobacillin, sporulate and germinate verywell, the spores containing normal concentra-tions of dipicolinate (121). Similarly, B. licheni-


http://mm

br.asm.org/

Dow

nloaded from



formis mutant SB 319 produces no detectableintra- or extracellular bacitracin, yet it sporu-lates normally (44). Both parent and mutanthave the same growth rate, morphology, andbiochemical characteristics, but the mutant hasa defective bacitracin synthetase complex (43).Much of the enthusiasm in favor ofa function

of antibiotics in sporulation derives from thework of Sarkar and Paulus (138). They noted aspecific inhibition of RNA synthesis by the ty-rothricin complex during growth of the produc-ing culture, B. brevis ATCC 8185. Moreover,purified B. brevis RNA polymerase was alsoinhibited by the complex. Both components ofthe tyrothricin complex (linear gramicidinand tyrocidine) but not gramicidin S, producedby another strain of B. brevis, were shown toexhibit similar inhibitory activity, indicatingthe specificity of the effect. They also observedthat the cessation of exponential vegetativegrowth (to) was accompanied by a sharp declineof net RNA synthesis. Tyrothricin synthesisbegan at to; the antibiotic complex remainedlargely associated with the cell material andreached a maximum level 2 h after exponentialgrowth (t2). Sarkar and Paulus (138) advancedthe view that peptide antibiotics regulate genetranscription during the transition from vege-tative growth to sporulation by selectively ter-minating the expression of vegetative genes.Unfortunately, they erroneously claimed thatall peptide antibiotics were produced by sporu-lating microorganisms, a statement that ig-nores the production of comirin by Pseudomo-nas antimycetica, micrococcin P by Micrococ-cus sp., and nisin by Streptococcus lactis andStreptococcus cremoris. It is even more discon-certing to read (65, 141) that all antibiotics aremade by sporulating cultures!Kleinkauf and co-workers (122, 123, 141) also

proposed a role for tyrocidine and linear grami-cidin in B. brevis sporulation. Their data sug-gest opposing effects ofthe two peptides in vivo.In vitro, tyrocidine inhibited transcription byforming a complex with DNA. Although lineargramidicin also inhibited transcription, it didnot complex with DNA. Of particular interestwas the ability of linear gramicidin to abolishthe action of tyrocidine, allowing RNA synthe-sis to proceed. Tyrocidine also inhibited in vivotranscription in B. brevis cells; in the presenceof linear gramicidin, however, a higher concen-tration of tyrocidine was required to attaincomplete inhibition. To explain these data, Ris-tow et al. (122) proposed that in vivo a complexmay be formed between linear gramicidin andtyrocidine which would have less affinity forDNA than tyrocidine alone. They argued thatperhaps both peptides are required to ensurethat not all of the genes are turned off at the

end of vegetative growth so that some can betranscribed during sporulation. What is dis-turbing about this hypothesis are their datashowing that, although linear gramicidin re-versed the effect of tyrocidine on growth and invivo RNA synthesis, it did not reverse tyroci-dine's inhibition of sporulation; in fact, in thepresence of the antibiotic mixture (i.e., in theproportion in which they are made as the ty-rothricin complex), sporulation was completelyinhibited. To us, this means that the inhibitionofRNA synthesis has nothing to do with sporu-lation. The main conclusions that one can de-rive from these studies is that both tyrocidineand linear gramicidin inhibit B. brevis RNApolymerase. However, this is not specific norsurprising since this group (141) has shown thattyrocidine also inhibits E. coli RNA polymer-ase, and the effect on B. brevis polymerase isalso exerted when calf thymus DNA is substi-tuted for B. brevis DNA as template.

All of the experimental data described to datecould fit in well with the hypothesis that sporu-lation and antibiotic formation, although inde-pendent phenomena, are regulated by a com-mon (or a similar) regulatory mechanism.Many ofthe mutants used in obtaining the datasuggesting a role of antibiotics in sporulationwere pleiotropic and probably blocked in regu-latory genes. An attempt to demonstrate afunction for the antibiotic in sporogenesis byadding crude antibiotic to pleiotropically aspo-rogenous and antibiotic-negative Bacillus mu-tants failed to induce sporulation (139). As sug-gested by Schmitt and Freese (143), the criticalexperiment involving the isolation ofan asporo-genous mutant that can sporulate only uponaddition of pure antibiotic has never been re-ported. The closest to such a finding was theclaim ofJayaraman and Kannan (65) that addi-tion of tyrothricin to the sporogenic B. brevisATCC 8185 culture inhibited growth but en-hanced sporulation. However, Ristow et al.(123) failed to confirm this result. Jayaramanand Kannan (65) also reported that addition ofpolymyxin to its producer, B. polymyxa, in-hibited growth and stimulated sporulation.However, the significance of this observation isobscured by their finding that the antibioticalso inhibited glucose uptake. Thus, sporula-tion could have been enhanced merely byrelease of sporulation from catabolite repres-sion. When Dleucine, an inhibitor of poly-myxin synthesis (but not growth), was suppliedto B. polymyxa, sporulation was strikingly in-hibited. L-Leucine reversed the effect of n-leu-cine on both antibiotic production and sporula-tion. However, since ileucine is needed forboth polymyxin synthesis and for protein syn-thesis during sporulation, the effect of D>leucine

VOL. 41, 1977


http://mm

br.asm.org/

Dow

nloaded from


468 KATZ AND DEMAIN

can be explained by an inhibitory effect on twoindependent phenomena.An oft-quoted and misquoted (65, 122, 123,

141) finding in support of a sporulation functionfor antibiotics are the data of Yoshida et al.(173), which revealed that addition ofactinomy-cin to young cultures of the producer organism(Streptomyces antibioticus) markedly inhibitedprotein synthesis, whereas later addition dur-ing actinomycin production had no effect. It isclaimed by these investigators (65, 122, 123,141) that, since the antibiotic exerts its effectduring growth but not during the stationaryphase, it plays an important role in the transi-tion between a growing and a differentiatingcell. However, the subsequent finding (104)that S. antibioticus becomes impermeable toexogenous actinomycin when the organism en-ters the stationary phase has apparently beenmissed by those who quote the original observa-tion and surely invalidates their claim.Some recent experiments by Lee et al. (90)

have provided an additional, although as-yet-unexplained, facet to the present controversy.It has been known that the enzymes for tyroci-dine synthesis suddenly appear at the end ofthe exponential growth phase; only during ashort period at the onset of the stationary phasehas it been possible to obtain active solubleenzyme extracts. Later in the stationary phase,only particulate enzymes are recovered. Bio-chemical investigations carried out concomi-tantly with electron microscopy have shownmost of the tyrocidine-forming activity becomesassociated with forespores in B. brevis. Thesignificance of the migration of antibiotic-syn-thesizing activity from the soluble portion ofthe cell into the forespores is not clear. Perhapsit merely represents a nonspecific trapping ofvegetative cell components that are truly unes-sential to the sporulation process or to the sporeitself. The authors (90) are properly cautious intheir interpretation of these results, statingthat their data are "consistent with the as-sumption of an effect of the antibiotic facilitat-ing the transition from the vegetative phaseinto sporulation. However, this laboratory hasnot yet probed into the causative connectionbetween antibiotic and spore formation."

It should be clear to the reader that the mys-tery of antibiotic function has not been solved.We feel, however, that peptide antibiotics musthave some role in the survival of the producingorganism. It is inconceivable to us that theintricate reaction sequences of antibiotic bio-synthesis would have been retained in naturewithout benefit to the organism. It is oftenargued that the strain degeneration problem inthe antibiotic industry speaks against a sur-vival function. However, this argument is irrel-

evant since the amounts produced' in commer-cial antibiotic fermentations are many orders ofmagnitude greater than that needed for an or-ganism's survival. In fact, the abnormal degreeof synthesis observed in industrial fermenta-tions is probably detrimental to the well-beingof the organism and it is logical to assume thatthe culture would revert to a lower-producing,more rapidly growing form which would be ca-pable of more balanced metabolic activity.Whatever the true function of antibiotics,

many mechanisms exist whereby organismscan protect themselves from the antibioticsthey elaborate (24). Permeability changes, com-partmentalization, and the presence of an inac-tive form of the antibiotic intracellularly mayall play a role in preventing self-annihilation.In support of the idea of "a built-in survivalkit," Kurylo-Borowska (83) demonstrated in B.brevis Vm4 that edeine B synthesized in vivo ispresent in an inactive form covalently bound tothe edeine-synthesizing polyenzymes and asso-ciated with a complex of cytoplasmic membraneand DNA. Free edeine is not found in the cellsand the polyenzyme-membrane-DNA-boundedeine neither exhibits antimicrobial activitynor influences DNA synthesis. Biological activ-ity is seen only when the antibiotic is releasedfrom the cell or from the complex in vitro.

CONCLUDING REMARKSIt is now evident that the biosynthesis of

peptide antibiotics by members of the genusBacillus is directed by multienzyme thiotem-plates. Amino acids employed as substrates forantibiotic synthesis are activated by ATP toform enzyme-bound aminoacyl adenylates. Theaminoacyl moiety is then transferred to second-ary acceptors, i.e., to specific SH groups, toform covalent thioester bonds. Antibiotic syn-thesis involves polyenzyme complexes that ex-hibit a broader specificity than that shown byribosomal protein-synthesizing systems. tRNA,mRNA, and ribosomes are not directly in-volved. In antibiotic synthesis, the polyenzymecomplex serves as template for the proper se-quencing of the amino acids in the peptide. Theprocesses of initiation, elongation, and termi-nation are inherent properties of the polyen-zyme systems. The energy derived from thethioester bonds provides the driving force forthe peptidation reactions. As observed in thegrowth of ribosomal polypeptide chains, anti-biotic formation is initiated from the N-termi-nal position and ends at the carboxyl terminus.In the case of antibiotics such as gramicidin Sand tyrocidine, cyclization is the final step be-fore release of the antibiotic from the enzyme.This reaction, as well as the mechanism ofamino acid racemization, is not well under-

BACTERIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from



stood. 4'-Phosphopantetheine plays an impor-tant role as coenzyme in the process of peptidechain elongation, mediating alternative trans-thiolation and transpeptidation reactions.

Little progress has been made in elucidatingthe function(s) of antibiotics in the producingorganism. We believe that they indeed have afunction(s) and that they are made in nature,especially in microenvironments containing or-ganic matter. It is doubtful that all antibioticshave the same general function; we certainlyaccept the fact that different classes of proteinshave different functions.The view that peptide antibiotics function in

cellular differentiation (i.e., sporulation) is notsupported by data obtained with producers ofgramicidin S, mycobacillin, and bacitracin. Inthese cases, nonproducing mutants are capableof sporulation. Of course, not enough work hasbeen done to eliminate the possibility thatother peptide antibiotics are involved in sporu-lation.

If the antibiotic is not an obligatory part ofthe sporulation process, then why are antibioticproduction and sporulation so often related toone another? We propose that they are regu-lated by the same or similar control mecha-nisms and, furthermore, that there are reasonswhy it is beneficial for the sporulating cell toproduce an antibiotic. For example, (i) the anti-biotic could be packaged in the Bacillus sporeand excreted during germination to inhibit orkill competitors, thereby providing an environ-ment favorable for the development of the ger-minating spore; or (ii) the antibiotic could bepackaged in the spore and inhibit germinationuntil environmental conditions become morefavorable. It is hoped that the availability ofnonproducing Bacillus mutants will now allowus to approach the problem of antibiotic func-tion in a more conclusive manner.

ACKNOWLEDGMENTSE. K. was supported by Public Health Service

grant CA 06926 from the National Cancer Instituteand a grant-in-aid from the Schering Corp. Supportfor the antibiotic studies of A. L. D. was receivedfrom National Science Foundation grant GI-34284and from Bristol-Myers Co.

ADDENDUM IN PROOFAkers et al. (lb) recently reported that two en-

zyme fractions (derived from B. brevis) catalyzedthe synthesis of gramicidin A to the heptapeptidestage. By contrast, Akashi et al. (la) noted that thesynthesis of formylvaline and formylvalylglycinecould be achieved employing a partially purifiedenzyme system from B. brevis ATCC 8185. Theenzyme preparation (component I) catalyzed anATP-PP, exchange reaction that was dependentsolely on valine and glycine. The molar ratio ofvaline to glycine bound by component I was ap-

proximately 1. Evidence was also obtained thatthe amino acids were bound to component I by thio-ester linkage and that formylvaline and formyl-valylglycine were synthesized (in the presence of aformyltetrahydrofolate-synthesizing system, ATP,valine, glycine, and an "AS45" fraction from B.brevis) as enzyme-bound intermediates. The authors(la) proposed the following biosynthetic sequencefor gramicidin A formation: (i) initiation via theformylation ofvaline bound to component I, followedby the synthesis of formylvalylglycine; (ii) peptideelongation through transfer of formylvalylglycine toa second enzyme component that activates andbinds alanine and leucine; (iii) further elongation ofthe nascent peptide, which may require additionalenzyme components. The biochemical mechanismfor linking ethanolamine to the C-terminal trypto-phan residue of gramicidin A still remains undeter-mined.

LITERATURE CITED1. Abraham, E. F. 1957. Biochemistry of some

peptide and steroid antibiotics. J. Wiley andSons, Inc., New York.

la.Akashi, K., K. Kubota, and K. Kurahashi.1977. Biosynthesis of enzyme-bound formyl-valine and formylvalylglycine. J. Biochem.81:269-272.

lb.Akers, H., S. G. Lee, and F. Lipmann. 1976.Fed. Proc. 35:1533.

2. Anderson, L. E., G. L. Coffey, G. D. Senos, M.A. Underhill, D. L. Vogler, and J. Ehrlich.1972. Butirosin, a new aminoglycoside anti-biotic complex. Bacterial origin and somemicrobiological properties. Antimicrob.Agents Chemother. 2:79-83.

3. Asselineau, J. 1966. The bacterial lipids.Hermann, Paris.

4. Balassa, G. M., H. Ionesco, and P. Schaeffer.1963. Preuve genetique d'une relation entrela production d'un antibiotique par Bacillussubtilis et sa sporulation. C. R. Acad. Sci.Paris 257:986-988.

5. Banerjee, A. B., and S. K. Bose. 1964. Biosyn-thesis of mycobacillin, a new antifungal pep-tide. I. Role of nucleic acid. J. Bacteriol.87:1397-1406. -

6. Banerjee, A. B., S. K. Majumdar, and S. K.Bose. 1967. Mycobacillin, p. 271-275. In D.Gottlieb and P. D. Shaw (ed.), Antibiotics.Springer-Verlag, Berlin.

7. Bauer, K., R. Roskoski, Jr., H. Kleinkauf, andF. Lipmann. 1972. Synthesis of a lineargramicidin by a combination of biosyntheticand organic methods. Biochemistry 11:3266-3271.

8. Berdy, J. 1974. Recent developments of anti-biotic research and classification of antibiot-ics according to chemical structure. Adv.Appl. Microbiol. 18:309-406.

9. Berg, T. L., L. 0. Froholm, and S. G. Laland.1965. The biosynthesis of gramicidin S in acell-free system. Biochem. J. 96:43-52.

10. Bernlohr, R. W., and G. D. Novelli. 1959. Anti-biotic production as a function of spore for-mation in Bacillus licheniformis. Nature

469VOL. 41, 1977


http://mm

br.asm.org/

Dow

nloaded from


470 KATZ AND DEMAIN

(London) 184:1256-1257.11. Bernlohr, R. W., and G. D. Novelli. 1963. Baci-

tracin biosynthesis and spore formation: thephysiological role of an antibiotic. Arch. Bio-chem. Biophys. 103:94-104.

12. Bhagavan, N. V., P. M. Rao, L. W. Pollard, R.K. Rao, T. Winnick, and J. B. Hall. 1966.The biosynthesis of gramicidin S. A restudy.Biochemistry 5:3844-3850.

13. Blanch, H. W., and P. L. Rogers. 1971. Pro-duction of gramicidin S in batch and continu-ous culture. Biotechnol. Bioeng. 13:843-864.

14. Bodanszky, M., and D. Perlman. 1964. Arepeptide antibiotics small proteins? Nature(London) 204:840-844.

15. Bodanszky, M., and D. Perlman. 1969. Peptideantibiotics. Science 163:352-358.

16. Bredesen, J. E., T. L. Zimmer, and S. G. La-land. 1969. Starting point and direction ofbiosynthesis of gramicidin S. FEBS Lett.3:169-172.

17. Brian, P. W. 1957. The ecological significanceof antibiotic production, p. 168-188. In R. E.0. Williams and C. C. Spicer (ed.), SeventhSymposium of The Society of General Micro-biologists. Cambridge University Press,Cambridge.

18. Bu'Lock, J. D. 1961. Intermediary metabolismand antibiotic synthesis. Adv. Appl. Micro-biol. 3:293-342.

19. Burkholder, P. R. 1959. Antibiotics. Science129:1457-1465.

20. Bycroft, B. W. 1969. Structural relationshipsin microbial peptides. Nature (London)224:595-597.

21. Cornell, N., and J. E. Snoke. 1964. Biosyn-thesis of bacitracin and protein. Biochem.Biophys. Acta 91:533-536.

22. Craig, L. C., W. Hausmann, and J. R. Weisi-ger. 1962. The qualitative and quantitativecontent of bacitracin A. J. Biol. Chem.199:865-871.

23. Demain, A. L. 1972. Cellular and environmen-tal factors affecting the synthesis and excre-tion of metabolites. J. Appl. Chem. Biotech-nol. 22:345-362.

24. Demain, A. L. 1974. How do antibiotic-pro-ducing microorganisms avoid suicide? Ann.N. Y. Acad. Sci. 235:601-612.

25. Demain, A. L., J. M. Piret, T. E. Friebel, E. J.Vandamme, and C. C. Matteo. 1976. Studieson Bacillus brevis directed towards the cell-free synthesis of gramicidin S, p. 437-443. InD. Schlessinger (ed.), Microbiology -1976.American Society for Microbiology, Wash-ington, D.C.

26. DiGirolamo, M., 0. Ciferri, A. B. DiGirolamo,and A. Albertini. 1964. Effect of D-leucine onthe biosynthesis of polymixin D. J. Biol.Chem. 239:502-507.

27. Egorov, N. S., A. N. Vypiyach, and G. G. Zhar-ikova. 1970. Free amino acids and the grami-cidin S content during the germination ofspores of the R-variant of Bacillus brevis.Mikrobiologiya 39:331-336.

28. Emilianowicz-Czerska, W., M. Szczepan-

BACTERIOL. REV.

kowska, H. Niemczyk, and M. Wokowicz.1973. Studies on biosynthesis of polymyxin E(Colistin). II. Biosynthesis ofpolymyxin E bymutant m-5 and d-13 ofBacillus polymyxa K-1. Acta Microbiol. Pol. 5:29-34.

29. Ensign, J. C. 1976. Properties and germinationof Streptomyces spores and a suggestion forfunction of antibiotics, p. 531-533. In D.Schlessinger (ed.) Microbiology -1976.American Society for Microbiology, Wash-ington, D.C.

30. Froyshov, 0. 1974. Bacitracin biosynthesis bythree complementary fractions from Bacilluslicheniformis. FEBS Lett. 44: 75-78.

31. Froyshov, O., and S. G. Laland. 1974. On thebiosynthesis of bacitracin by a soluble en-zyme complex from Bacillus licheniformis.Eur. J. Biochem. 46:235-242.

32. Froyshov, 0., T. L. Zimmer, and S. G. Laland.1970. The nature of the enzyme-bound inter-mediates in gramicidin S biosynthesis.FEBS Lett. 7:68-71.

33. Fujikawa, K., Y. Sakamato, and K. Kura-hashi. 1971. Biosynthesis of tyrocidine by acell-free enzyme system of Bacillus brevisATCC 8185. III. Further purification of com-ponents I and II and their functions in tyroci-dine synthesis. J. Biochem. 69:869-879.

34. Fujikawa, K., Y. Sakamato, T. Suzuki, and K.Kurahashi. 1968. Biosynthesis of tyrocidineby a cell-free enzyme system of Bacillusbrevis ATCC 8185. II. Amino acid substitu-tion in tyrocidine. Biochim. Biophys. Acta169:520-533.

35. Fujikawa, K., T. Suzuki, and K. Kurahashi.1966. Incorporation of L-leucine-14C into ty-rocidine by a cell-free preparation ofBacillusbrevis Dubos strain. J. Biochem. 60:216-218.

36. Fujikawa, K., T. Suzuki, and K. Kurahashi.1968. Biosynthesis of tyrocidine by a cell-freeenzyme system ofBacillus brevis ATCC 8185.I. Preparation of partially purified enzymesystem and its properties. Biochim. Biophys.Acta 161:232-246.

37. Gevers, W., H. Kleinkauf, and F. Lipmann.1968. The activation of amino acids for bio-synthesis of gramicidin S. Proc. Natl. Acad.Sci. U.S.A. 60:269-276.

38. Gevers, W., H. Kleinkauf, and F. Lipmann.1969. Peptidyl transfers in gramicidin S bio-synthesis from enzyme-bound thioester in-termediates. Proc. Natl. Acad. Sci. U.S.A.63:1335-1342.

39. Gilhuus-Moe, C. C., T. Kristensen, J. E. Bre-desen, T. L. Zimmer, and S. G. Laland. 1970.The presence and possible role of phospho-pantethenic acid in gramicidin S synthetase.FEBS Lett. 7:287-290.

40. Gottlieb, D. 1976. The production and role ofantibiotics in soil. J. Antibiot. 29:987-1000.

41. Haavik, H. I. 1974. Studies on the formation ofbacitracin by Bacillus licheniformis: effect ofglucose. J. Gen. Microbiol. 81:383-390.

42. Haavik, H. I. 1974. Studies on the formation ofbacitracin by Bacillus licheniformis: role ofcatabolite repression and organic acids. J.


http://mm

br.asm.org/

Dow

nloaded from


VOL. 41, 1977

Gen. Microbiol. 84:321-326.43. Haavik, H. I., and 0. Froyshov. 1975. Function

of peptide antibiotics in producer organisms.Nature (London) 254:79-82.

44. Haavik, H. I., and S. Thomassen. 1973. A baci-tracin-negative mutant of Bacillus licheni-formis which is able to sporulate. J. Gen.Microbiol. 76:451-454.

45. Hanson, R. S., J. A. Peterson, and A. A. Yous-ten. 1970. Unique biochemical events in bac-terial sporulation. Annu. Rev. Microbiol.24:53-90.

46. Hatfield, G. M. 1975. Toxins of higher fungi.Lloydia 38:36-55.

47. Hendlin, D. 1949. The nutritional require-ments of a bacitracin-producing strain ofBa-cillus subtilis. Arch. Biochem. Biophys.24:435-446.

48. Hettinger, T. P., and L. C. Craig. 1968. Ed-eine. II. The composition of the antibioticpeptide edeine A. Biochemistry 7:4147-4152.

49. Hettinger, T. P., and L. C. Craig. 1970. Ed-eine. IV. Structures ofthe antibiotic peptidesedeines A, and B,. Biochemistry 9:1224-1232.

50. Hettinger, T. P., Z. Kurylo-Borowska, and L.C. Craig. 1968. Edeine. III. Composition ofthe antibiotic peptide edeine B. Biochemis-try 7:4153-4160.

51. Hoch, J. A., and J. Spizizen. 1969. Geneticcontrol ofsome early events in sporulation ofBacillus subtilis 168. In L. L. Campbell.(ed.), Spores IV. American Society for Micro-biology, Bethesda, Md.

52. Hodgson, B. 1970. Possible roles for antibioticsand other biologically active peptides at spe-cific stages during sporulation ofBacillaceae.J. Theor. Biol. 30:111-119.

53. Holm, H., L. 0. Froholm, and S. G. Laland.1966. Isolation of a peptide conjugate withthe sequence phe-pro-val-orn from a cell-freesystem producing gramicidin S. Biochim.Biophys. Acta 115:361-370.

54. Huang, F. C., J. A. Chan, C. J. Sih, P. Faw-cett, and E. P. Abraham. 1975. The nonparti-cipation of a,,B-dehydrovalinyl intermedi-ates in the formation of8-(L-a-aminoadipyl)-L-cysteinyl-D-valine. J. Am. Chem. Soc.97:3858-3859.

55. Hurst, A. 1969. Biosynthesis of polypeptide an-tibiotics, p. 167-182. In G. W. Gould and A.Hurst (ed.), The bacterial spore, chapter 5.Academic Press Inc., London.

56. Ishihara, H., Y. Endo, S. Abe, and K. Shi-mura. 1975. The presence of 4'-phosphopan-tetheine in the bacitracin synthetase. FEBSLett. 50:43-46.

57. Ishihara, H., T. Sasaki, and K. Shimura. 1968.Biosynthesis of bacitracin. II. Incorporationof '4C-labeled amino acids into bacitracin bya cell-free preparation from Bacillus licheni-formis. Biochim. Biophys. Acta 166:496-504.

58. Ishihara, H., and K. Shimura. 1974. Biosyn-thesis of bacitracin. III. Partial purificationof a bacitracin-synthesizing enzyme systemfrom Bacillus licheniformis. Biochim. Bio-phys. Acta 338:588-600.


59. Ito, M., K. Aida, and T. Uemura. 1969. Studieson the bacterial formation of peptide anti-biotic, colistin. II. On the biosynthesis of 6-methyloctanoic acid and isooctanoic acids.Agric. Biol. Chem. 33:262-269.

60. Ito, M., K. Aida, and T. Uemura. 1970. Isola-tion of a new intermediary substance consist-ing of colistin fatty acid and L-a,y-diamino-butyric acid from colistin-producing cells ofBacillus colistinus Koyama. Agric. Biol.Chem. 34:476-479.

61. Ito, M., K. Aida, and T. Uemura. 1970. Biosyn-thesis of colistin by Bacillus colistinus Koy-ama. Biochim. Biophys. Acta 213:244-247.

62. Ito, M., Y. Koyama, K. Aida, and T. Uemura.1970. Biosynthesis of colistin by a cell-freesystem Bacillus colistinus Koyama.Biochim. Biophys. Acta 215:418-420.

63. Itoh, H., M. Yamada, S. Tomino, and K. Kura-hashi. 1968. The role of two complementaryfractions of gramicidin S synthesizing en-zyme system. J. Biochem. 64:259-261.

64. 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.

65. Jayaraman, K., and R. Kannan. 1972. The roleof polypeptide antibiotics on the bacterialdifferentiation. Biochem. Biophys. Res.Commun. 48:1235-1239.

66. Jayaraman, K., J. Monteal, and H. Paulus.1969. L-a,-y-diaminobutyrate-activating en-zyme from Bacillus polymyxa. Properties anddistribution. Biochim. Biophys. Acta185:447-457.

67. Kalakoutskii, L. V., and N. S. Agre. 1976.Comparative aspects of development and dif-ferentiation in actinomycetes. Bacteriol.Rev. 40:469-524.

68. Kambe, M., Y. Imae, and K. Kurahashi. 1974.Biochemical studies on gramicidin S non-producing mutants of Bacillus brevis ATCC9999. J. Biochem. 75:481-493.

69. Kambe, M., Y. Sakamoto, and K. Kurahashi.1971. Biosynthesis of tyrocidine by a cell-freeenzyme system ofBacillus brevis ATCC 8185.IV. Further separation of component II intotwo fractions. J. Biochem. 69:1131-1133.

70. Katz, E. 1967. Biogenesis of the actinomycins,p. 276-341. In D. Gottlieb and P. D. Shaw(ed.), Antibiotics. Springer-Verlag, Berlin.

71. Katz, E. 1968. Peptide antibiotics: proteinsthat never grew up? Lloydia 31:364-394.

72. Katz, E. 1971. Biosynthesis ofpolypeptide anti-biotics. Pure Appl. Chem. 28:551-570.

73. King, T. P., and L. C. Craig. 1955. The chemis-try of tyrocidine V. The amino acid sequenceof tyrocidine B. J. Am. Chem. Soc. 77:6627-6631.

74. Kleinkauf, H., and W. Gevers. 1969. Non-ribo-somal polypeptide synthesis: the biosyn-thesis of a cyclic peptide antibiotic, gramici-din S. Cold Spring Harbor Symp. Quant.Biol. 34:805-813.

75. Kleinkauf, H., W. Gevers, and F. Lipmann.


http://mm

br.asm.org/

Dow

nloaded from


472 KATZ AND DEMAIN

1969. Interrelation between activation andpolymerization in gramicidin S biosynthesis.Proc. Natl. Acad. Sci. U.S.A. 62:226-233.

76. Kleinkauf, H., W. Gevers, R. Roskoski, Jr.,and F. Lipmann. 1970. Enzyme-bound phos-phopantetheine in tyrocidine biosynthesis.Biochem. Biophys. Res. Commun. 41:1218-1222.

77. Kleinkauf, H., R. Roskoski, Jr. and F. Lip-mann. 1971. Pantetheine-linked peptide in-termediates in gramicidin S and tyrocidinebiosynthesis. Proc. Natl. Acad. Sci. U.S.A.68:2069-2072.

78. Krasil'nikov, N. A. 1958. In Soil microorga-nisms and higher plants. Academy of ScienceU.S.S.R., Moscow.

79. Krogvik, R. 1973. Ph.D. thesis, University ofOslo, Oslo, Norway.

80. Kurahashi, K. 1974. Biosynthesis of small pep-tides. Annu. Rev. Biochem. 43:445-459.

81. Kurahashi, K., M. Yamada, K. Mori, K. Fuji-kawa, M. Kambe, Y. Imae, E. Sato, H. Tak-ahashi, and Y. Sakamoto. 1969. Biosyn-thesis of cyclic oligopeptide. Cold SpringHarbor Symp. Quant. Biol. 34:815-826.

82. Kurylo-Borowska, Z. 1967. Edeine, p. 342-352.In D. Gottlieb and P. D. Shaw (ed.), Antibiot-ics, vol. 2. Springer-Verlag, Berlin.

83. Kurylo-Borowska, Z. 1975. Biosynthesis of ed- ,

eine. II. Localization of edeine synthetasewithin Bacillus brevis Vm4. Biochim. Bio-phys. Acta 399:31-41.

84. Kurylo-Borowska, Z., and T. Abramsky. 1972.Biosynthesis of 3-tyrosine. Biochim. Bio-phys. Acta 264:1-10.

85. Kurylo-Borowska, Z., and J. Sedkowska. 1974.Biosynthesis of edeine. Fractionation andcharacterization of enzymes responsible forbiosynthesis of edeine A and B. Biochim.Biophys. Acta 351:42-56.

86. Kurylo-Borowska, Z., and E. L. Tatum. 1966.Biosynthesis of edeine by Bacillus brevisVm4 in vivo and in vitro. Biochim. Biophys.Acta 113:206-209.

87. Laland, S. G., 0. Froyshov, C. Gilhuus-Moe,and T. L. Zimmer. 1972. Gramicidin S syn-thetase, an enzyme with an unusually largenumber of catalytic functions. Nature (Lon-don) New Biol. 239:43-44.

88. Laland, S. G., and T. L. Zimmer. 1973. Theprotein template mechanism of synthesis forthe peptide antibiotics produced by Bacillusbrevis. Essays Biochem. 9:31-57.

89. Lee, S. G., and F. Lipmann. 1974. Isolation of apeptidyl-pantetheine-protein from tyroci-dine-synthesizing polyenzymes. Proc. Natl.Acad. Sci. U.S.A. 71:607-611.

90. Lee, S. G., V. Littau, and F. Lipmann. 1975.The relation between sporulation and theinduction of antibiotic synthesis and ofamino acid uptake in Bacillus brevis. J. CellBiol. 66:233-242.

91. Lee, S. G., R. Roskoski, Jr., K. Bauer, and F.Lipmann. 1973. Purification of the polyen-zymes responsible for tyrocidine synthesisand their dissociation into subunits. Bio-

BACTERIOL. REV.

chemistry 12:398-405.92. Leighton, T. J., P. K. Freese, R. H. Doi, R. A.

J. Warren, and R. A. Kelln. 1972. Initiationof sporulation in Bacillus subtilis: require-ment for serine protease activity and ribonu-cleic acid polymerase modification, p. 238-246. In H. 0. Halvorson, R. Hanson, and L.L. Campbell (ed.), Spores V. American Soci-ety for Microbiology, Washington, D. C.

93. Leung, D. C., and R. Baxter. 1972. Substrate-derived reversible and irreversible inhibitorsof the multi-enzyme I of gramicidin S biosyn-thesis. Biochim. Biophys. Acta 279:34-47.

94. Lipmann, F. 1971. Attempts to map a processevolution of peptide biosynthesis. Science173:875-884.

95. Lipmann, F. 1973. Nonribosomal polypeptidesynthesis on polyenzyme templates. Acc.Chem. Res. 6:361-367.

96. Lipmann, F. 1975. The search for remnants ofearly evolution in present day metabolism.Biosystems 6:234-238.

97. Lipmann, F., W. Gevers, H. Kleinkauf, and R.Roskoski, Jr. 1971. Polypeptide synthesis onprotein templates: the enzymatic synthesisof gramicidin S and tyrocidine. Adv. En-zymol. 35:1-34.

98. Ljones, T. L., 0. Vaege, T. L. Zimmer, L. 0.Froholm, and S. G. Laland. 1968. The pres-ence of protein-bound intermediates in thebiosynthesis of gramicidin S. FEBS Lett.1:339-342.

99. Lockhart, I. M., and E. P. Abraham. 1954. Theamino acid sequence in bacitracin A. Bio-chem. J. 58:633-647.

100. Lockhart, I. M., E. P. Abraham, and G. G. F.Newton. 1955. The N-terminal and sulphur-containing residues of bacitracin A. Bio-chem. J. 61:534-544.

101. Luria, S. 1973. Colicins, p. 293-320. In L. Leive(ed.), Bacterial membranes and walls. Mar-cel Dekker, Inc., New York.

102. Mach, B., and E. L. Tatum. 1964. Environ-mental control of amino acid substitutions inthe biosynthesis of the antibiotic polypeptidetyrocidine. Proc. Natl. Acad. Sci. U.S.A.52:876-884.

103. Majumdar, S. K., and S. K. Bose. 1960. Aminoacid sequence in mycobacillin. Biochem. J.74:596-599.

104. Marshall, R., B. Redfield, E. Katz, and H.Weissbach. 1968. Changes in phenoxazinonesynthetase activity during the growth cycleof Streptomyces antibioticus. Arch. Biochem.Biophys. 123:317-323.

105. Mason, K. T., G. J. Shaw, and E. Katz. 1977.Arch. Biochem. Biophys., in press.

106. Matteo, C. C., C. L. Cooney, and A. L. De-main. 1976. Production of gramicidin S syn-thetase by Bacillus brevis in continuous cul-ture. J. Gen. Microbiol. 96:415-422.

107. Matteo, C. C., M. Glade, A. Tanaka, J. Piret,and A. L. Demain. 1975. Microbiologicalstudies on the formation of gramicidin S syn-thetases. Biotechnol. Bioeng. 17:129-142.

108. Mauger, A. B. 1968. Peptide antibiotic biosyn-


http://mm

br.asm.org/

Dow

nloaded from


VOL. 41, 1977

thesis: a new approach. Experientia 24:1068-1072.

109. Maier, W., and D. Groger. 1972. Biosyntheseder peptid antibiotica. Pharmazie 27:491-505.

110. Meister, A. 1965. Biochemistry of the aminoacids, p. 369-375. Academic Press Inc., NewYork.

111. Nesemann, G., P. Praeve, D. Sukatsch, and L.Vertesy. 1972. Polyene antibiotic from bacte-ria. Naturwissenschaften 59:81-82.

112. Okuda, K., G. C. Edwards, and T. Winnick.1963. Biosynthesis of gramicidin and tyroci-dine in the Dubos strain ofBacillus brevis 1.Experiments with growing cells. J. Bacte-riol. 85:329-338.

113. Otani, S., T. Yamanoi, Y. Saito, and S. Otani.1966. Fractionation of an enzyme system re-sponsible for gramicidin S biosynthesis. Bio-chem. Biophys. Res. Commun. 25:590-594.

114. Paladini, A., and L. C. Craig. 1954. The chem-istry of tyrocidine. III. The structure of tyro-cidine A. J. Am. Chem. Soc. 76:688-694.

115. Paulus, H. 1967. Polymyxins, p. 254-267. In D.Gottlieb and P. D. Shaw (ed.), Antibiotics,vol. 2. Springer-Verlag, Berlin.

116. Paulus, H., and E. Gray. 1964. The biosyn-thesis of polymyxin B by growing cultures ofBacillus polymyxa. J. Biol. Chem. 239:865-871.

117. Perlman, D., and M. Bodanszky. 1971. Biosyn-thesis of peptide antibiotics. Annu. Rev. Bio-chem. 40:449-464.

118. Pfaender, P., D. Specht, G. Heinrich, E.Schwarz, E. Kuhnle, and M. M. Simlot.1973. Enzymes of Bacillus licheniformis inthe biosynthesis of bacitracin A. FEBS Lett.32:100-104.

119. Rao, R. K., N. V. Bhagavan, K. R. Rao, and J.B. Hall. 1968. Peptide synthesis in cell-freeextracts of Bacillus brevis 8185. Biochemis-try 7:3072-3077.

120. Rao, K., and J. B. Hall. 1970. Biosynthesis oftyrocidines by extracts of two strains of Ba-cillus brevis. Proc. Soc. Exp. Biol. Med.135:345-347.

121. Ray, B., and S. K. Bose. 1971. Polypeptide-antibiotic negative spore former mutants ofBacillus subtilis. J. Gen. Appl. Microbiol.17:491-498.

122. Ristow, H., B. Schazschneider, K. Bauer, andH. Kleinkauf. 1975. Tyrocidine and the lin-ear gramicidin. Do these peptide antibioticsplay an antagonistic regulative role in sporu-lation? Biochim. Biophys. Acta 390:246-252.

123. Ristow, H., B. Schazschneider, and H. Klein-kauf. 1975. Effects of the peptide antibioticstyrocidine and the linear gramicidin on RNAsynthesis and sporulation of Bacillus brevis.Biochem. Biophys. Res. Commun. 63:1085-1092.

124. Rogers, H. J., N. Lomakina, and E. P. Abra-ham. 1965. Observations on the structure ofbacilysin. Biochem. J. 97:579-586.

125. Roncari, G., Z. Kurylo-Borowska, and L. C.Craig. 1966. On the chemical nature of the


antibiotic edeine. Biochemistry 5:2153-2159.126. Roscoe, J., and E. P. Abraham. 1966. Experi-

ments relating to the biosynthesis of bacily-sin. Biochem. J. 99:793-800.

127. Roskoski, R., Jr., W. Gevers, H. Kleinkauf,and F. Lipmann. 1970. Tyrocidine biosyn-thesis by three complementary fractionsfrom Bacillus brevis. (ATCC8185). Biochem-istry 9:4839-4845.

128. Roskoski, R., Jr., H. Kleinkauf, W. Gevers,and F. Lipmann. 1970. Isolation of enzyme-bound peptide intermediates in tyrocidinebiosynthesis. Biochemistry 9:4846-4851.

129. Roskoski, R., Jr., G. Ryan, H. Kleinkauf, W.Gevers, and F. Lipmann. 1971. Polypeptidebiosynthesis from thioesters of amino acids.Arch. Biochem. Biophys. 143:485-492.

130. Ruttenberg, M. A., and B. Mach. 1966. Studieson amino acid substitution in the biosyn-thesis of the antibiotic polypeptide tyroci-dine. Biochemistry 5:2864-2869.

131. Ruttenberg, M. A., T. P. King, and L. C.Craig. 1965. The chemistry of tyrocidine. VI.The amino acid sequence of tyrocidine C.Biochemistry 4:11-18.

132. Sadoff, H. L. 1972. Sporulation antibiotics ofBacillus species, p. 157-166. In H. 0. Halvor-son, R. Hansen, and L. L. Campbell (ed.),Spores V. American Society for Microbiol-ogy, Washington, D.C.

133. Saito, Y., S. Otani, and S. Otani. 1970. Biosyn-thesis of gramicidin S. Adv. Enzymol.33:337-380.

134. Sarges, R., and B. Witkop. 1965. GramicidinA. V. The structure of valine and isoleu-cine-gramicidin A. J. Am. Chem. Soc.87:2011-2020.

135. Sarges, R., and B. Witkop. 1965. GramicidinA. VI. The synthesis of valine and isoleu-cine-gramicidin A. J. Am. Chem. Soc.87:2020-2027.

136. Sarges, R., and B. Witkop. 1965. Gramicidin.VII. The structure of valine and isoleucine-gramicidin B. J. Am. Chem. Soc. 87:2027-2030.

137. Sarges, R., and B. Witkop. 1965. Gramicidin.VIII. The structure of valine and isoleu-cine -gramicidin C. Biochemistry 4:2491-2494.

138. Sarkar, N., and H. Paulus. 1972. Function ofpeptide antibiotics in sporulation. Nature(London) New Biol. 239:228-230.

139. Schaeffer, P. 1969. Sporulation and the pro-duction of antibiotics, exoenzymes, and exo-toxins. Bacteriol. Rev. 33:48-71.

140. Schaeffer, P., H. lonesco, A. Ryter, and G.Balassa. 1963. La sporulation de Bacillussubtilis: etude genetique et physiologique.Colloq. Int. C. N. R. S. 124:553-563.

141. Schazschneider, B., H. Ristow, and H. Klein-kauf. 1974. Interaction between the anti-biotic tyrocidine and DNA in vitro. Nature(London) 249:757-759.

142. Schleifer, K. H., and 0. Kandler. 1972. Pepti-doglycan types of bacterial cell walls andtheir taxonomic implications. Bacteriol.


http://mm

br.asm.org/

Dow

nloaded from


474 KATZ AND DEMAIN

Rev. 36:407-477.143. Schmitt, R., and E; Freese. 1968. Curing of a

sporulation mutant and antibiotic activity ofBacillus subtilis. J. Bacteriol. 96:1255-1265.

144. Sengupta, S., A. B. Banerjee, and S. K. Bose.1971. y-Glutamyl and D- or L-peptide link-ages in mycobacillin, a cyclic peptide anti-biotic. Biochem. J. 121:839-846.

145. Sengupta, S., and S. K. Bose. 1971. Propertiesand localization of mycobacillin-synthesizingenzyme system in Bacillus subtilis B3.Biochim. Biophys. Acta 237:120-122.

146. Sengupta, S., and S. K. Bose. 1972. Peptidesfrom a mycobacillin-synthesizing cell-freesystem. Biochem. J. 128:47-52.

147. 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.

148. Shimura, K., T. Sasaki, and K. Sugawara.1964. Biosynthesis of bacitracin I. Formationof bacitracin by a subcellular fraction ofBa-cillus licheniformis. Biochim. Biophys. Acta86:46-55.

149. Simlot, M. M., P. Pfaender, and D. Specht.1973. Synthesis of antibiotics by enzymesfrom altered growth conditions by Bacilluslicheniformis. FEBS Lett. 35:231-235.

150. Singh, B. N. 1973. Current status of the prob-lem of exogenous and endogenous amoe-biasis. J. Sci. Ind. Res. 32:399-432.

151. Snoke, J. E. 1960. Formation of bacitracin byprotoplasts of Bacillus licheniformis. J. Bac-teriol. 81:986-989.

152. Snoke, J. E., and N. Cornell. 1965. Protoplastlysis and inhibition of growth of Bacilluslicheniformis by bacitracin. J. Bacteriol.89:415-420.

153. Stoll, E., 0. Froyshov, H. Holm, T. L. Zim-mer, and S. G. Laland. 1970. On the mecha-nism of gramicidin S formation from inter-mediate peptides. FEBS Lett. 11:348-352.

154. Sussman, A. S., and H. A. Douthit. 1973. Dor-mancy in microbial spores. Annu. Rev. PlantPhysiol. 24:311-352.

155. Suzuki, T., and K. Fujikawa. 1964. Studies onthe chemical structure of colistin. IV. Chem-ical structure of colistin B. J. Biochem.56:182-189.

156. Suzuki, T., K. Hayashi, K. Fujikawa, and K.Tsukamoto. 1965. The chemical structure ofpolymyxin E: the identities of polymyxin E,with colistin A and of polymyxin E2 withcolistin B. J. Biochem. 57:226-227.

157. Suzuki, T., H. Inouye, K. Fujikawa, and Y.Suketa. 1963. Studies on the chemical struc-ture of colistin. I. Fractionation, molecularweight determination, amino acid and fattyacid composition. J. Biochem. 54:25-33.

158. Szulmajster, J., C. Bonamy, and J. Laporte.1970. Isolation and properties of a tempera-ture-sensitive sporulation mutant ofBacillussubtilis. J. Bacteriol. 101:1027-1037.

159. Tomino, S., and K. Kurahashi. 1964. Enzymic

BACTERIOL. REV.

synthesis of D-phenylalanyl-L-prolyl-L-va-line, a peptide sequence present in gramici-din S. Biochem. Biophys. Res. Commun.17:288-293.

160. Tomino, S., M. Yamada, H. Itoh, and K. Kura-hashi. 1967. Cell-free synthesis of gramicidinS. Biochemistry 6:2552-2560.

161. Troy, F. A. 1973. Chemistry and biosynthesisof the poly (y-D-glutamyl) capsule in Bacil-lus licheniformis. I. Properties of the mem-brane-mediated biosynthetic reaction. J.Biol. Chem. 248:305-315.

162. Tsuji, T. 1966. Biosynthesis of gramicidin S. V.Enzymatic formation of several small pep-tides related to gramicidin S biosynthesis. J.Osaka Med. Center 15:1-9.

163. Vogler, K. and R. 0. Studer. 1966. The chemis-try of the polymyxin antibiotics. Experientia22:345-354.

164. Waksman, S. A. 1961. The role of antibiotics innature. Perspect. Biol. Med. 4:271-287.

165. Weinberg, E. D. 1970. Biosynthesis of second-ary metabolites: roles of trace metals. Adv.Microbial Physiol. 4:1-44.

166. Weinberg, E. D. 1971. Secondary metabolism:raison d'etre. Perspect. Biol. Med. 14:565-577.

167. Weinberg, E. D. 1967. Bacitracin, gramicidinand tyrocidine, p. 240-253. In D. Gottlieb andP. D. Shaw (ed.), Antibiotics. Springer-Ver-lag, Berlin.

168. Winnick, R. E., and T. Winnick. 1961. Biosyn-thesis of gramicidin S. II. Incorporation ex-periments with labeled amino acid analogs,and the amino acid activation process.Biochim. Biophys. Acta 53:461-468.

169. Wojciechowska, H., J. Ciarkowski, H.Chmara, and E. Borowski. 1972. The anti-biotic edeine. IX. The isolation and composi-tion of edeine D. Experientia 28:1423-1424.

170. Woodruff, H. B. 1966. The physiology of anti-biotic production: the role of the producingorganism p. 22-46. In B. A. Newton and P. E.Reynolds (ed.), Sixteenth Symposium of theSociety of General Microbiologists. Cam-bridge University Press, Cambridge.

171. Yamada, M., and K. Kurahashi. 1968. Adeno-sine triphosphate and pyrophosphate de-pendent phenylalanine racemase of Bacillusbrevis Nagano. J. Biochem. 63:59-69.

172. Yamada, M., S. Tomino, and K. Kurahashi.1964. Conversion of L-phenylalanine to D-phenylalanine by cell-free extracts of Bacil-lus brevis. J. Biochem. 56:616-618.

173. Yoshida, T., H. Weissbach, and E. Katz. 1966.Inhibitory effect of actinomycin upon theproducing organism. Arch. Biochem. Bio-phys. 114:252-255.

174. Yukioka, M., Y. Saito, and S. Otani. 1966.Enzymatic hydrolysis of gramicidin S. J.Biochem. 60:295-302.

175. Yukioka, M., Y. Tsukamoto, Y. Saito, T.Tsuji, S. Otani, and S. Otani. 1965. Biosyn-thesis of gramicidin S by a cell-free system ofBacillus brevis. Biochem. Biophys. Res.Commun. 19:204-208.


http://mm

br.asm.org/

Dow

nloaded from


the peptide antibiotics ofbacillus: chemistry, biogenesis...

Documents