THE EFFECT OF AMINO ACID ANALOGUES ON GROWTH AND PROTEINSYNTHESIS IN MICROORGANISMS

M. H. RICHMONDBacterial Physiology Division, National Institute for Medical Research, Mill Hill, London, England

I. Introduction............................................................................ 398II. Chemical Structure of Analogues........................................................ 398

III. Effect of Amino Acid Analogues on Growth of Bacteria.................................. 403IV. Effect of Amino Acid Analogues on Mechanism of Protein Synthesis and Properties of

Proteins................................................................................ 404A. Effect on Assimilation and Supply of Amino Acids ................................... 405B. Effect on Process of Protein Synthesis ............................................... 408C. Incorporation of Amino Acid Analogues into Cell Proteins ........................... 411

V. Effect of Amino Acid Analogues on Synthesis of Bacteriophage ................. 414VI. Conclusions ............................................................................ 415

VII. Acknowledgments....................................................................... 417VIII. Literature Cited........................................................................ 417

I. INTRODUCTION

Since the work of Woods (92) on the mecha-nism of action of sulfonamides, it has frequentlybeen observed that inhibitors of biological in-terest are structural analogues of some essentialmetabolite. Such inhibitors often compete withthe natural metabolite at the active center ofan enzyme, and this holds true for the usefulstructural analogues of the 20 natural aminoacids found to occur universally in proteins. Inthis case, however, there is an additional dimen-sion to the problem because certain analoguesare so similar to the natural compounds thatthey become incorporated into proteins in theplace of the natural amino acids. The proteinsformed in this way are sometimes altered in theirspecific enzyme activity, and since some of theseproteins will themselves be the enzymes whichhandle the analogue as a substrate, the uptakeof an effective analogue into a cell may havefar-reaching and complex effects, leading finallyto complete growth inhibition.

In practice, many amino acid analogues havebeen found to inhibit the growth of microor-ganisms. First, therefore, the effect of differentanalogues on growth will be examined to see ifany common patterns of inhibition emerge. Then,the effect of these analogues on the mechanismof protein synthesis and the properties of pro-teins synthesized in the presence of the analogueswill be considered. This evidence allows sometentative conclusions to be drawn concerning:(i) how some analogues exert their growth in-

hibitory effect, (ii) the factors determining aminoacid incorporation into proteins, and (iii) theeffects of replacement of a natural amino acidby an analogue on the enzymic and immuno-logical properties of certain proteins. Evidenceobtained with mammalian systems has onlybeen considered in the review where it bearsdirectly on the processes in microorganisms, andno attempt has been made to discuss in detaileffects of amino acid analogues on the forma-tion of virus particles. This subject has not yetbeen reviewed, but a number of papers on thistopic have recently appeared (1, 2, 80, 89, 96).A number of structural analogues of amino

acids not found in proteins have been synthe-sized and found to be antimetabolites. Theyinclude, for example, structural analogues of the3-alanine portion of pantothenic acid, but theyhave not been considered further (93).A preliminary section is included to consider

the analogues that have inhibitory properties,to decide what chemical changes can lead to theformation of potent analogues.

It should, perhaps, be pointed out that twocompounds of similar shape are correctly called"isosteres," but in this review the term "ana-logue" has been employed throughout since itis common usage.

II. CHEMICAL STRUCTURE OF ANALOGUES

Two criteria have to be fulfilled for an "un-natural" amino acid to act as an effective anti-metabolite.

398

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

1962] AMINO ACID ANALOGUES AND GROWTH AND PROTEIN SYNTHESIS

TABLE 1. Amino acid analogues of biochemical interest grouped according to the change in structure thatleads to antimetabolite properties

Alteration Analogue Natural amino acid Key references

-F for -H

-0- for -CH2-

-S- for CH2--CH2-for -S--0- for -5-Se- for -S--CH= for -CH2--CH(OH)- for -CH2-Increase in chain length by

-CH2-

-NH2 for -OH-Cl for -CH3-N= for -CH== in ring sys-tems

-CH3 for -H in a phenyl ring

Change in the nature of rings

Miscellaneous

p-Fluorophenylalanineo-Fluorophenylalaninem-Fluorophenylalanine3-Fluorotyrosine5-Fluorotryptophan6-FluorotryptophanCanavanineO-MethylthreonineS- (2-Aminoethyl)cysteineNorleucineMethoxinine (O-methylhomoserine)SelenomethionineMethallylglycinej3-PhenylserineEthionineHomoarginineco-Methyllysinep-Aminophenylalaninea-Amino-,6-chlorobutyric acidAzatryptophanTryptazan4-Methyltryptophan5-Methyltryptophan6-Methyltryptophana. Thienyl for phenyl

B-2-Thienylalaninej8-3-Thienylalanine

b. Benzothienyl for indolyl,f-3-Benzothienylalanine

c. Furyl for phenyl,5-2-Furylalanine,B-3-Furylalanine

d. Thiazole for imidazole2-Thiazolealanine

3- (Aminoethyl)cyclohexaneglycineCyclohexaneglycine,B- (Cyclopentane) alaninef- (1 -Cyclopentene) alanine

PhenylalaninePhenylalaninePhenylalanineTyrosineTryptophanTryptophanArginineIsoleucineLysineMethionineMethionineMethionineLeucinePhenylalanineMethionineArginineLysineTyrosineValineTryptophanTryptophanTryptophanTryptophanTryptophan

PhenylalaninePhenylalanine

Tryptophan

PhenylalaninePhenylalanine

HistidineLysineIsoleucineLeucinePhenylalanine

1) The molecule must have a similar shapeand size to the naturally occurring molecule.

2) If the substituent group is one that can

ionize, it must either produce the same type ofion (e.g., N (-NH)-NH3+ and

-{-N-C(=NH)NH3+) or be un-ionized atphysiological pH values.

In practice, all the useful amino acid analogueshave been close structural isomers of naturallyoccuring amino acids. Apart from a report (66)that D-norleucine inhibits the growth of Escher-ichia coli, all amino acid analogues have the L

configuration and compete with the natural

L-amino acids. Table 1 shows the compoundsconsidered in this review grouped according tothe change in structure that has produced an

active antimetabolite. Most frequently, fourtypes of change seem to have been effective:(i) replacement of hydrogen in a ring system byfluorine, (ii) replacement of phenyl by some otherresonant ring structure, (iii) replacement of one

type of heterocyclic ring by another, and (iv)replacement of a residue from the backbone ofan amino acid by another with similar size andshape.

Replacement of hydrogen in a ring system by

(17, 64, 65)(63)(58)(12)(8, 9)(69)(75)(71)(72)(23)(79)(82, 86)(28)(47, 57)(95)(88)(54)(10)(11)(69)(13)(85)(59)(52)

(65)(28)

(37)

(28)(28)

(60)(26)(33-36, 43,51, 68)

399

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

M. H. RICHMOND

TABLE 2. Comparison of size and shape of analogous groups in amino acid analogues and theirnatural competitors

Group Compound distance deangle group involved(Angstrom) dps (Angstrom)

Phenyl C-F o, m, and p-Fluorophenylalanine; fluoro- 1.3 Planar 0.64*tryptophan; 3-fluorotyrosine

Phenyl C-H Phenylalanine, tryptophan, tyrosine 1.08 Planar 0.37*

C-C-C Norleucine 1.54 108 1.84C-S-C Methionine, S- (aminoethyl)cysteine 1.81 105 1.02C-Se-C Selenomethionine 1.98 98 1.13C-O-C Methoxinine, O-methylthreonine 1.43 108 0.6C-CH CH- Methallylglycine 1.33 108 1.84

Phenyl C-OH Tyrosine 1.36 PlanarPhenyl C-NH2 p-Aminophenylalanine 1.47 Planar

C-C-CH3 Valine 1.54 108 1.86C-C-Cl a-Amino-1-chlorobutyric acid 1.76 110 1.05

* Van der Waals radii: given by Bergmann (8). Other data from "Interatomic Distances" (46).

phenyl-

\Zl

thienyl

FIG. 1. Phenyl (C0H5-), thienyl (C4H3S-),and furyl (C4H30-) drawn to scale to show relativebond lengths and angles. All radicals are planarand symmetrical about the dashed lines (see refer-ence (46)).

Bond lengths (A)

Phenyl: a = b= c= 1.40.Thienyl: a = 1.44, b = 1.35, c = 1.74.Furyl: a = 1.46, b = 1.35, c = 1.46.

AnglesPhenyl: Zab = Lbc = Zcc = 120°.Thienyl: Lab = 1130, Zbc= 112°, Lcc = 910.Furyl: Lab = 1060, Zbc = 1110, Lcc = 1070.

Dimensions (A):

a = 1.33b= 1.33c= 1.47d = 1.54e = 2.47

Zcd = 1100

a = 1.33b = 1.33c = 1.38d = 1.43e = 2.32

Lcd = 111 i 30

fluorine. Although the fluorine atom is appreci-ably larger than hydrogen, the aromatic C-Fand C-H bonds are very similar in length (Table2). In addition fluorine in such a combination isvery unreactive (44), has similar electronic prop-erties to hydrogen, and, like hydrogen, producesminimal disturbance of the electronic resonancesof the phenyl radical.So far this transposition has produced o-, m-

and p-fluorophenylalanine (phenylalanine ana-logues), 3-fluorotyrosine (tyrosine analogue),and 5- and 6-fluorotryptophan (tryptophan ana-logues). Bergmann has recently discussed thesynthesis of other fluorine-containing antimetabo-lites (8). Substitution of hydrogen by halogensother than fluorine rarely leads to potent ana-logues. However, replacement of -CH3 by -Clor -Br is sometimes effective (11).

Replacement of phenyl by another resonant ringsystem. The replacement of the phenyl residueof phenylalanine by thienyl- or furyl- has led tothe formation of active antimetabolites. How-ever, the thiophen and furan rings differ fromphenyl since the C-C bond length varies de-pending on the distance of the bond from theO or S atom in the molecule (Fig. 1). The slightlydifferent potency of (8-2- and fl-3-thienyl or furylalanines (27) probably reflects the fact that sub-stitution in the 2 position gives a slightly closersimilarity to a substituted phenyl residue thansubstitution at the 3 position. The thiophen

400 [VOLF. 26

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

1962] AMINO ACID ANALOGUES AND GROWTH AND PROTEIN SYNTHESIS

analogue of tryptophan (3-(benzothienyl)alanine)is an analogue of doubtful potency (5, 37).

Replacement of one type of heterocyclic ring byanother. Moyed (60) has recently shown thatreplacement of inlidazole by thiazole producesan effective histidine analogue. This compoundis a potent inhibitor of the histidine biosyntheticpathway, and its effect is reversed by L-histidine.Unfortunately the relative size and shape of theimidazole and thiazole radicals are not accuratelyknown.

Replacement of carbon in a ring system bynitrogen often produces active antimetabolites,and has led to the design of the tryptophananalogues a-amino-f3-(3-indazole)propionic acid(tryptazan) and 7-azatryptophan. No crystal-lographic data for these compounds exist, butin pyridine the analogous system =C-N=C-has interatomic distances of 1.37 A and 1200which is close to that found for =C-C=C-in benzene derivatives, and clearly the replace-ment of carbon by nitrogen in 7-azatryptophanwill have little effect on shape of the moleculewhen compared with tryptophan. The same isalmost certainly true for the introduction of thesecond nitrogen into the five-membered ring oftryptazan. As the replacement of nitrogen forcarbon in the six-membered ring of tryptophanis so effective, it is surprising that fl-2(pyridyl)alanine is such a poor analogue of phenylalanine(51).Replacement of a residue from an amino acid

backbone. Changes of this type involve the re-placement of -5- by -CH2- (or vice versa),-CH2- by -0-, -5- by -0- and -5-by -Se-. Table 2 shows the interatomic dis-tances, bond angles, and atomic radii associated

,I, NH2a I,.C=NHbI?>NH\

C\I \

c>CH2 edI/

'CH2

H2

CH (NH2)

COOH

Arginine

,//NH2a I>C=NH

b I

>NHe

c',i "e

"CH2'

CH2

CII (NH2)

COOH

Canavanine

FIG. 2. Structures of arginine and canavanine

with changes of this kind. It will be seen thatcertain of these changes introduce a considerablemodification in the structure of the moleculeswhen compared with the changes consideredabove; thus the w-methyl of norleucine will lie2.49 A from the 6-carbon of the backbone,whereas the distance is 2.87 A in methionine,a shortening of 13.2%. Further, when one con-siders that selenomethionine can presumablyreplace at the same sites as norleucine, the dif-ference in that measurement becomes 2.49 A to2.99 A, or d10%.An interesting compound that falls into this

group is canavanine (Fig. 2). Here the replace-ment of -C- by -0- leads to a shorteningfrom 2.47 A in arginine to 2.32 A in canavanine,although the bond angle is scarcely affected. Afurther effect of the substitution is to lower thedissociation constant of the guanidino groupfrom pH 12.5 in arginine to about pH 8.0 incanavanine. It is surprising, therefore, thatcanavanine is capable of taking part in the ma-jority of metabolic interconversions undergoneby arginine (48). Despite the fall in dissociationconstant, however, the majority of canavaninemolecules will have an ionized guanidino groupat physiological pH values, and from this pointof view be indistinguishable from arginine.Another example in which two compounds with

very different substituent groups can act as ana-logues when un-ionized is shown by p-amino-phenylalanine and the reversal of growth in-hibition caused by this compound with tyrosine.The shape and size of the -NH2 and -OHgroups are similar, and both groups can be ionized(Table 2). The hydroxyl of tyrosine forms RO-and H+ with an acid dissociation constant ofabout 109, whereas the -NH2 of p-amino-phenylalanine has a constant of about 4 X 10-(unpublished observations), and forms RNH3+.This means that, at physiological pH values,both amino acids are present in the un-ionizedform. The relative shape and size of substituentsin other useful analogues is summarized in Table2.

Size limits for analogues of inditidual aminoacids. Although many amino acid analogues havenow been examined, it is not possible to gener-alize about changes that will produce activeanalogues of any specific amino acid. The systemsresponsible for handling the natural amino acidsseem to vary greatly in the divergence from the

401

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

M. H. RICHMOND

normal structure that they will tolerate in theirsubstrates. Under certain circumstances en-zymes normally specific for methionine (R * CH2 .S *Me) will act on methoxinine (R 1CH2*0. Me),selenomethionine (R-CH2- Se . Me), norleucine(R*CH2.CH2. Me) or ethionine (R*CH2 * S -

CH2 *Me). On the other hand, organisms donot seem to mistake glycine for alanine or leucinefor isoleucine or valine to any great extent-allof which changes are of the same order of sizeas the changes possible for methionine. In cer-tain cases the structure required in an aminoacid antimetabolite seems to be extremely ex-

NH2

CH2

CH2/ \

OH2 O1H2CH2

0H2(NH2)

COOH3-(Aminomethyl)-cyclohexaneglycine

OCH2\OH2 CH2

OH2 CH2

OH

H2N COOH

Isoleucine Cyclohexaneglycine

FIG. 3. Comparison of structures of isoleucine andcyclohexaneglycine.

CHI-CH2CH2 CH2 CH3 CH3.\CH \CH/

OH2 OH2

CH CH

H2N/ C~OOH H2N COOH

,0-(Cyclopentane)alanine Leucine

CHCH2-CH2 CH CH

CH2 \bH OlH

\CH/

OH2 CH2I

OH CH

H2N COOH H2N COOH

,B-(1-Cyclopentene)alanine Phenylalanine

FIG. 4. Comparison of f-(cyclopentane)alaninewith leucine and ,#-(1-cyclopentene)alanine withphenylalanine.

FIG. 5. Comparison of structures of 3- (amino-methyl)cyclohexaneglycine and lysine.

acting as far as the relative positions of certainatoms are concerned, whereas the over-all sizeof the molecule is relatively unimportant. Hard-ing and Shive (43) have synthesized a range ofantimetabolites which illustrate this point well.These workers first synthesized cyclopentane-glycine, which was found to inhibit growth ofE. coli. This compound has a puckered ring sys-tem, and in the actual molecules the carbon atoms1-5 and 8 in cyclohexaneglycine have a veryclose spatial similarity to the carbon atoms 1-6of isoleucine (Fig. 3). Growth tests showed thatisoleucine reversed the growth inhibition due tothe antimetabolite. In a similar way f3-(cyclo-pentane)alanine (which has a puckered ringsystem) was found to be a leucine analogue,whereas f3-(1-cyclopentene)alanine has a planarring and resembled phenylalanine (Fig. 4) (68).Synthesis of further cyclohexane and cyclo-hexene derivatives of alanine and glycine led toantimetabolites specific for leucine, isoleucine,and valine. In the above examples the relativepositions of the carbon atoms in the antimetabo-lites are very critical, whereas the over-all sizeand shape of the molecules are quite dissimilar(33-36). These principles of design have ledto 3-(aminomethyl)cyclohexaneglycine (a lysineanalogue), in which the relative positions of thea-carboxyl, a-amino, and co-amino groups areclose to those found in lysine. The shape of themolecule is, however, quite different (Fig. 5)(25, 26).Table 1 shows that effective antimetabolites

have been made only for the larger amino acids.To date no effective analogues of glycine, serine,

NH2

OH2

CH2

UH2

CH2

CH(NH2)

OOOH

Lysine

CH3

OH2 CH3

CH

CH

H21 \COOH

402 [VOL. 26

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

1962] AMINO ACID ANALOGUES AND GROWTH AND PROTEIN SYNTHESIS

cysteine, threonine, alanine, aspartic acid, glu-tamic acid, or proline have been reported, andit is tempting to suggest that the biological sitesthat deal with these amino acids are more ex-acting to molecular shape and size than thosehandling the larger amino acids. Probably thisis not so, for changes in the structure of the smallamino acids (unless the replacing group is iden-tical in shape and size with the natural com-pound) will produce a relatively larger changein over-all shape and size than in a large aminoacid. Furthermore, certain of the changes usedto produce antimetabolites from the larger aminoacids cannot be used with the smaller, since theresulting compounds are too unstable, for ex-ample, fluoroglycine or the introduction of sulfurinto the backbone chain of threonine or glutamicacid in the place of -CH2-. The absence ofeffective proline analogues is puzzling, but azeti-dine-2-carboxylic acid might prove to be in-teresting in this respect (39).1

III. EFFECT OF AMINO ACID ANALOGUES ONGROWTH OF BACTERIA

Two points will be considered in this section.First, what are the characteristic effects of aminoacid analogues on the growth of microorganismsand secondly, is it possible to deduce anythingabout the function of an analogue from the typeof inhibition observed?The majority of amino acid analogues, whether

designed as antimicrobial, antiviral, or antitumoragents, are screened routinely for possible anti-metabolite activity against bacteria, and a massof reports exists, therefore, on the effect of ana-logues on bacterial growth. Usually E. coli orsome organism with a well characterized aminoacid requirement such as Leuconostoc mesenter-oides P60 is inoculated into a simple syntheticmedium containing suitable concentrations ofthe compound under test, and the total yield ofbacteria obtained after overnight incubation ismeasured. It is not possible to deduce much aboutthe detailed effect of an amino acid on growthfrom experiments of this kind since no idea ofthe time sequence of inhibition can be obtained.Also, growth in the test could be due to the selec-tion of resistant forms (3, 59, 60). These criticismsapply even more strongly to tests in which the

1 Tests have now shown that azetidine-2-car-boxylic acid replaces proline in protein from E.coli and certain higher plants (39a).

suspected analogue is mixed with varying quan-tities of the natural amino acid in an attemptto show competition.The first careful work on the growth kinetics

following addition of amino acid analogues tobacterial cultures was carried out by Cohen andMunier (17, 64, 65). Addition of 4- or 5-methyl-tryptophan, p-fluorophenylalanine, (3-2-thienyl-alanine, norleucine, or ,3-phenylserine to culturesof E. coli growing exponentially in a simple syn-thetic medium resulted in an immediate changein growth rate. The growth curve obtained wasdescribed as "linear," that is, the mass of theculture measured turbidimetrically increased ata linear rate from the point of addition of theanalogue, whereas the control culture continuedto increase exponentially. Linear growth con-tinued for periods up to 5 hr and was associatedwith an increase of optical density of 3- to 4-fold.A similar type of inhibition was found on addingethionine to cultures of E. coli (40).

Cultures of Bacillus cereus, on the other hand,showed little change in growth rate (measuredturbidimetrically) for about 20 min after additionof p-fluorophenylalanine. Thereafter growthslowed, and eventually ceased about 4 hr afteraddition of the analogue (76). Addition of 7-azatryptophan to E. coli B caused an immediatefall in growth rate, but the growth curve remainedexponential. In this case the growth curve wasmeasured both turbidimetrically and by colonycounts of viable organisms (69).The implications of linear growth are rather

interesting. Linear increase in mass of individualcells between divisions followed by cell divisionto two parts at intervals so that the mean cellsize remains constant results in an exponentialincrease in cell mass in the culture, i.e., an ex-ponential growth rate. A continued linear growthrate could be achieved in a number of ways. Forexample, it would occur if one of the daughtercells formed on division ceased to grow, whereasthe other continued to grow at full rate. Experi-ments on clones of organisms growing in thepresence of p-fluorophenylalanine have shownthat cell division continues after addition of theanalogue, and that both daughter cells continueto reproduce for at least 90 min. If both thedaughter cells are equally potent, a linear growthrate must imply that on division both daughtercells continue to grow, but only at half the rateof the parent. Cohen and Munier (17) suggest

403

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

M. H. RICHMOND

that such a type of growth could arise if additionof an analogue caused an immediate halt in for-mation of an enzyme whose function immediatelybecame growth limiting. Such a situation wouldindeed lead to linear growth, but it is difficultto see how this could continue for very long with-out some imbalance in cell metabolism leadingto a fall in growth rate. In practice a linear growthrate of precisely the type suggested by Munierand Cohen does not occur, since the slope of thecurves obtained after addition of p-fluorophenyl-alanine or thienylalanine (16, 17, 83) is less thanwould be expected from the exponential rate ofthe control culture. If the growth rate and themass of an exponentially growing culture areknown at the point of addition of the analogue,the expected slope of the subsequent lineargrowth curve can be calculated accurately. Therates published for p-fluorophenylalanine- andthienylalanine-grown cells are between 10 and30% less than would be expected from the ex-ponential rate at the point of addition of theanalogues.

In view of these considerations it seemed im-portant to determine whether addition of p-fluorophenylalanine did indeed lead to true lineargrowth or only to an approximately linear rate.The analogue was therefore added to culturesof E. coli growing exponentially in a simple syn-thetic medium. The growth curve was followedby (i) turbidimetry, (ii) colony counts of viableorganisms, and (iii) cell number. In additionsamples were removed at intervals and celllength was determined by direct measurement offixed and stained preparations (20). Preliminaryresults suggest that addition of the analoguecauses an immediate onset of a growth rate be-tween exponential and truly arithmetic whenmeasured turbidimetrically, but that this ratebecomes progressively slower as incubation con-tinues. Colony counts, dry weight, and total cellnumber increased in parallel with turbidity, andalthough cell division continued, the mean sizeof the cells increased. These results are verysimilar to those reported for yeast (16), but ratherdifferent from those with E. coli (65).

Quite apart from determining whether additionof an analogue leads to perfect "linear" growthor not, the growth curve obtained after additionof an antimetabolite to a culture growing ex-ponentially in a single synthetic medium can,in practice, suggest how the analogue may act.

If growth is immediately inhibited completely,yet the inhibition is reversible by the naturalamino acid, it is likely that the antimetaboliteinhibits some essential step in the biosyntheticpathway of the natural amino acid. In this caseit is most unlikely to be incorporated into cellprotein. On the other hand, onset of "linear"growth (either immediately or within a genera-tion under current cultural conditions), followed4 to 5 hr later by complete inhibition, is charac-teristic of all the analogues known to replacenatural amino acids in proteins. Thus, from thedata published by Moyed (60), it is likely that2-thiazolealanine is incorporated into the pro-teins of E. coli and that it partially inhibits thebiosynthesis of histidine. Similarly, cyclohexene-alanine has a growth-inhibitory effect on E. colisimilar to that of p-fluorophenylalanine, and maybe a further example of a structural analogue ofphenylalanine that can replace the natural aminoacid in protein (33). The inhibitory action of alluseful analogues is immediately and completelyreversed by the natural structural analogue, butnot by any other amino acid (except where freemetabolic interconversion to the natural com-petitor is possible). Thus, L-p-fluorophenylala-nine inhibition is immediately reversed by L-phenylalanine and phenylpyruvic acid, but notby tyrosine. Nonspecific reversals such as thoseobtained by Bergmann, Sicher, and Volcani (10)for a number of phenylalanine derivatives sug-gest that the inhibitions observed in these caseswere not due to any of the mechanisms consideredbelow. Of course, the effect of an analogue ongrowth rate only gives an indication of the mecha-nism of action; so far, however, all analoguesuseful in studies on bacterial metabolism haveshown some effect on growth within a generation,provided the natural competitor was absent.This last point brings up another limitation ofthe method, since only relatively few bacterialspecies will grow in defined media and allow ef-fective growth tests.

IV. EFFECT OF AMINO ACID ANALOGUES ONMECHANISM OF PROTEIN SYNTHESIS AND

PROPERTIES OF PROTEINS

In practice the majority of amino acid ana-logues inhibit growth by interfering with eitherprotein synthesis or the function of the proteinssynthesized in the presence of the analogues. Inthe first case they prevent or reduce the rate of

404 [VOL. 26

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

1962] AMINO ACID ANALOGUES AND GROWTH AND PROTEIN SYNTHESIS

formation of normal proteins; in the second, theyreplace natural amino acids in proteins, and thismay, secondarily, affect protein synthesis by al-tering the properties of the enzymes involved.These two processes will be dealt with separately,although certain compounds will be consideredin both sections.

Effect of analogues on protein synthesis: relationto growth. An early paper by Halvorson andSpiegelman (42) was interpreted, and has beenwidely reported, as showing that a number ofanalogues (e.g., ethionine and o-, m-, and p-

fluorophenylalanine) completely inhibited pro-

tein synthesis in yeast. The experiments were

carried out in media containing all the require-ments for growth except a nitrogen source. Underthese conditions yeast is able to synthesize a

limited amount of protein from the free aminoacid pool contained in the organisms, and it wassuggested that protein synthesis could be meas-

ured by following the disappearance of aminoacids from the pool. Addition of the analoguesled to an immediate complete inhibition of thedisappearance of pool amino acids, and this was

interpreted as a complete inhibition of proteinsynthesis.

Subsequent work by Cohen and Munier (17)has shown that addition of p-fluorophenylalanineto exponentially growing cultures of E. coli slowsbut does not stop protein synthesis. When thequantity of protein synthesized was plotted as a

differential plot (increase in amount of proteinvs. increase in bacterial dry weight) it was foundthat the analogue had no preferential inhibitoryeffect on protein synthesis; that is, growth (asmeasured turbidimetrically) and protein syn-

thesis were inhibited to the same extent. In viewof these results, Cohen et al. (16) have reinvesti-gated the effect of p-fluorophenylalanine on

protein synthesis in yeast. Addition of the ana-

logue to exponentially growing cultures led im-mediately to "linear" growth and had no effecton the differential rate of protein synthesis. The1952 experiments were therefore reinterpreted:the nondisappearance of amino acids from thepool of yeast was due to an increased flow ofamino acids to the pool rather than to a specificinhibition of protein synthesis. This reinterpreta-tion was supported by experiments showing an

increased rate of protein turnover in the presence

of p-fluorophenylalanine (16). On the basis ofthese and other experiments it is now well es-

tablished that certain analogues (e.g., o-, m-,and p-fluorophenylalanine, ethionine, thienyl-alanine, norleucine, canavanine, p-aminophenyl-alanine, selenomethionine, and tryptazan) causepartial growth inhibition but do not have a pref-erential inhibitory effect on protein synthesis.This is not true of 5-methyltryptophan. Thiscompound causes an immediate complete in-hibition of protein synthesis, and has therefore asimilar effect to that suggested by Halvorson andSpiegelman for all analogues.

In the following sections an attempt will bemade to locate the step at which the analoguescause a partial or complete inhibition in the rateof protein synthesis and of growth.

A. Effect on Assimilation and Supply ofAmino Acids

Protein synthesis involves two main processes.First, the provision of the amino acids necessaryto make the proteins, and, secondly, the processof ordering the amino acids to form specific pro-teins. In most bacteria some amino acids aretaken up from the surrounding environment andsome are synthesized within the cell from pre-cursors, which in turn have been assimilated fromoutside the cell. Both processes are susceptibleat various stages to inhibition by amino acidanalogues.

It is now fairly well established that the aminoacids in the "pool" in bacteria are in rapid equi-librium with the protein-synthesizing sites inthe cell (14). Any inhibition by analogues of theassimilation of essential amino acids into thepool will therefore interfere with protein syn-thesis. Cohen and Rickenberg (15) have studiedthis problem by incubating E. coli in the presenceof an amino acid-C14 until the pool is maximallylabeled, and then transferring the organisms tomedia containing a variety of single nonradio-active amino acids. When the second amino acidis the nonradioactive isotope of the first and atequimolar concentration, the level of radioac-tivity in the pool falls rapidly to a value of about30% of the maximal level characteristic of thatamino acid. In practice it was found that L-i50-leucine, L-norleucine, and -leucine were capableof exchanging with valine-C14 in the pool butmethionine, proline, and phenylalanine were not.Similarly, L-phenylalanine-C'4 could be displacedby p-fluorophenylalanine (if present at 20 timesthe concentration of phenylalanine) but not by

405

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

M. H. RICHMOND

phenylserine or isoleucine, and L-methionine-C'4could be displaced by L-norleucine but not byD-norleucine, D-methionine, proline, or phenyl-alanine. These experiments were interpreted asshowing that there are structural requirementsinvolved in holding an amino acid in the pool andfor an exogenous amino acid to displace thisamino acid from the pool. Furthermore, certainanalogues could pass freely into the cell, probablyusing the same catalytic sites as the naturallyoccurring structural isomers. It is interesting thatL-norleucine will displace both valine and me-thionine from the pool, yet methionine will notdisplace valine, nor valine displace methionine.This means, presumably, that norleucine is suf-ficiently similar to both valine and methionineto react on the same catalytic sites as those aminoacids, whereas valine and methionine are toodissimilar to have an overlapping specificity.The use of the membrane filter technique has

allowed Cowie (21) to study the assimilation ofp-fluorophenylalanine-C14 into the pool of E. coliand the effect of phenylalanine on this process.Radioactive p-fluorophenylalanine flowed rapidlyinto the pool, and thence into protein, when nophenylalanine was present in the growth medium.When equimolar concentrations of phenylalanineand p-fluorophenylalanine were present, thephenylalanine passed rapidly to the pool whilethe p-fluoro analogue remained in the medium.Only when the p-fluorophenylalanine concentra-tion in the medium was four times that of thephenylalanine, did any analogue pass into thepool. In a similar way the phenylalanine passedfrom the pool to the cell protein preferentially.It follows almost certainly that phenylalanineand the p-fluoro analogue use the same routeinto the pool. Further, these results explain whyrelatively low exogenous levels of phenylalaninewill completely protect cultures from the in-hibitory effects of p-fluorophenylalanine. Similarexperiments carried out with B. cereus and Sal-monella typhimurium have shown that phenyl-alanine was used in preference to p-fluorophenyl-alanine in both cases. Quantitatively S.typhimurium was similar to E. coli but with B.cereus p-fluorophenylalanine was used signifi-cantly only when the concentration of analoguein the medium rose to about ten times the phenyl-alanine concentration ((76, 77) and unpublishedobservations). Unfortunately, similar experimentshave not been carried out with other analogues,

but the kinetics of inhibition found with ,-2-thienylalanine and the reversal by phenylalaninesuggest that this analogue behaves in the sameway as p-fluorophenylalanine (49).Dunn (29, 30) and co-workers (31, 32) have

studied the effect of a range of di- and tripep-tides containing ,B-2-thienylalanine on E. coli.These compounds are as effective growth in-hibitors as thienylalanine itself, and the inhibitionis reversed by phenylalanine. However, the ki-netics of reversal suggest that thienylalanine istaken up into the cell by a different route fromthat used by the peptides. This deduction is onlytentative since there is no conclusive evidencethat the peptides are not hydrolyzed to the freeamino acids before exerting their inhibitory effect.

In summary, therefore, it seems likely thatanalogues are taken into the cell by the sameroute as their natural competitors, and that thetransport system discriminates against the un-natural compound. No analogues have showninhibitory properties in broken cell preparationsmade from bacteria and yet been inactive againstintact organisms because of their inability to passthe permeability barriers of the cell, a situationwhich is of course theoretically possible.

Analogues as inhibitors of endogenous synthesis.Recent work on the regulation of the endogenoussupply of amino acids has shown that the rate ofsynthesis of the amino acids may be controlledby two processes known as feedback inhibitionand repression. Feedback inhibition implies thatthe compound formed by the biosynthetic path-way will inhibit the functioning of the enzymesof the pathway if it is allowed to accumulateeither in the cells or the growth medium. There isno change in the quantity of the biosyntheticenzymes that are synthesized; it is their functionthat is inhibited. Repression, on the other hand,is a process whereby accumulation of the productof a biosynthetic pathway leads to a block in theformation of the biosynthetic enzymes. In practiceboth inhibition and repression by analogues havebeen reported. It is difficult to see, however, howrepression by an analogue can be proved unlessthe enzymes in question are purified so that theactual quantity of enzyme protein that is syn-thesized can be measured, because the incorpora-tion of analogues into enzyme proteins is knownin some cases to affect specific enzyme activity.Some of the earliest work in which amino acid

analogues were used concerned the inhibition of

406 [VOL. 26

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

19621 AMINO ACID ANALOGUES AND GROWTH AND PROTEIN SYNTHESIS

COOH

HO O.POH

Shikimic acid-5-phosphate

+ glutamic acid 1 anthranhlic acid

5-methyltiyptophan6-fluorotryptophan

ØOOH

NH2

Anthranilic acid

+ HO.P.P.ribose.P

5-phosphoribosyl-1-pyrophosphate

Ø> COOH 0

NH-CH CH2O.P

CH-CHOH OH

N-o-Carboxyphenyl-D-ribosylamine-l-phosphate

COOH4.C2HO*C*CHOH*CHOH.CH20P 4. -COn CHOH.CHOH.CH20,P

N NH H

5- ) OHC.CHOH.CH20P + >INH

Indole glycerol-3-phosphate

-~ CH2CH(NH2)COOH± serine 6.

(tryptophan-"'

desmolase Hor synthetase)

Glyceraldehyde-3-phosphate

Indole Indoleacrylic4-methyltryptophan6-methyltryptophan

FIG. 6. Tryptophan biosynthetic pathway

tryptophan synthetase, an enzyme involved inthe biosynthesis of tryptophan. This enzyme

catalyzes the condensation of indole and serineto form tryptophan (Fig. 6), and Fildes (38)showed that in intact organisms it is competi-tively inhibited by indoleacrylic acid. Subse-quently the isomeric 2-, 4-, 5-, 6-, and 7-methyl-tryptophans were tested on the same enzyme:2-methlytryptophan was inactive, whereas theothers were inhibitory in the order 4 > 5 >6 > 7. 6-Methyltryptophan has subsequentlybeen shown to be a genuine repressor of a num-

ber of enzymes of the tryptophan biosyntheticpathway (52), whereas the 4-methyl derivativewas confirmed as a strong inhibitor of purifiedtryptophan synthetase (85). Pardee and Prestidge(70) have reported that E. coli mutant 19-2 (atryptophan requirer) releases anthranilic acidinto the medium and that this process is blockedby 7-azatryptophan. It seems likely, therefore,

that 7-azatryptophan blocks an enzymic stepprior to the formation of anthranilic acid.

Other tryptophan analogues inhibit at otherstages of tryptophan biosynthesis (Fig. 6). Moyedand Friedman (62) report that 6-fluorotryp-tophan inhibits the condensation of anthranilicacid with phosphoribosylpyrophosphate, whereas5-methyltryptophan inhibits the conversion ofshikimic acid-5-phosphate + glutamic acid toanthranilic acid (59). Bergmann (7) and co-

workers (9) have found that 5-fluorotryptophanblocks the conversion of anthranilic acid to in-dole. Probably this analogue acts on the same

enzyme as 6-fluorotryptophan (62), but it is notpossible to be certain since the experiments were

carried out on whole organisms, unlike the ex-

periments with the 6-fluoro analogue, where a

partly purified enzyme system was used.Munier and Cohen (65) have tried to measure

the effect of p-fluorophenylalanine and p3-2-

[Anthranilic deoxyribonucleotide]

Tryptophan

407

2.

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

M. H. RICHMOND

thienylalanine on the biosynthesis of phenylala-nine, and of norleucine on methionine synthesis.When p-fluorophenylalanine or fl-2-thienylala-nine are present in the culture medium thereseems to be a decrease in the total amount ofphenylalanine synthesized by the organism. Asthe authors point out, these experiments are notconclusive since only a single sample (taken at anunspecified time after addition of the analogues)was analyzed, and decomposition of amino acidin the growth medium was not measured. Withthese provisos the results are consistent witheither inhibition or repression of the biosyntheticpathway. Inhibition seems the more likely, sincerepression of enzymes of the phenylalanine path-way must be restricted to the last stages of syn-thesis. The synthetic route to this amino acidhas many steps in common with tyrosine, andneither p-fluorophenylalanine nor thienylalaninehave any inhibitory effect on tyrosine synthesis(65). Under similar conditions norleucine has noeffect on the biosynthesis of methionine althoughit does have a profound effect on the incorpora-tion of methionine into protein. In this case theanalogue increased enormously the amount ofmethionine released into the medium (65).

Recently the histidine biosynthetic pathwayhas been shown to be inhibited by 2-thiazole-alanine (60). This compound inhibited the en-zymic formation of "compound III" (61) from

ribose-5-phosphate and adenosine triphosphate(ATP).In summary, therefore, it is clear that certain

analogues, on gaining entry to the cell, inhibitthe endogenous supply of the natural competitor.As will be seen later, this greatly facilitates theincorporation of the analogue into proteins byincreasing the relative concentration of analogueto competitor that is supplied to the protein-synthesizing enzymes.

B. Effect on Process of Protein Synthesis

As has been mentioned previously, analogueseither interfere with the normal process of order-ing the amino acids in the peptide chains of theprotein, or the proteins that contain the analoguein place of a natural amino acid are less effective,whether in a structural or catalytic role. Theformation and function of abnormal proteins willbe dealt with later; at this stage, discussion willbe confined to the effect of analogues on the in-dividual steps of the synthetic process, and forthis purpose an outline of the steps involved maybe helpful (Fig. 7).The reports of many workers with mammalian

cells suggest that one site at which the aminoacids are arranged into peptide chains is themicrosome. In bacteria the situation is less clear-cut, since to date no synthesis of a specific en-zyme protein has been shown unequivocally tooccur on any of the many different sized ribo-

ATP

,\Amino acids V PP

amino acidadenylates \

specific ribosonacceptor + nasceiRNA

(SRNA) A

AMP

tinformation from DNA

by way of "messenger" RNA

FIG. 7. Postulated mechanism for the insertion offree amino acids into peptide chains of proteins. A TP,adenosine triphosphate; PP, pyrophosphate; AMP, adenosine monophosphate; RNA, ribonucleic acid;SRNA, soluble RNA; DNA, deoxyribonucleic acid.

nal RNAunt prote

RNA? + "messenger" RNA

408 [VOL. 26

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

1962] AMINO ACID ANALOGUES AND GROWTH AND PROTEIN SYNTHESIS

nucleoprotein particles found in bacteria (45).However, the experiments reported by severalgroups of workers (55, 78) make it likely thatthe 70S ribosome is the structure in bacteriaanalogous to the microsome of mammalian tis-sues. Two problems exist, therefore. First, how isthe genetic information contained in the deoxy-ribonucleic acid (DNA) of the organism trans-ferred to the particles in which the synthesis ofthe peptide chains occurs; and secondly, whatchanges do the amino acids undergo before theyare incorporated into the peptide chains of theindividual proteins? The early steps in the trans-fer of genetic information to the ribosome (ormicrosome) do not concern us directly here,since the reactions involved are, as yet, onlytentative, and the effect of amino acid analogueson them is undetermined. The changes undergoneby the amino acids, on the other hand, are rele-vant.The first step in the reaction sequence is the

"activation" of the amino acids to form aminoacid adenylates (45, 84). The process is catalyzedby "activating enzymes" present in the "pH 5enzyme fraction" of bacterial cells. Since thereaction proceeds according to the followingequation:

ATP + amino acid M___amino acid adenylate + pyrophosphate,

the reaction may be followed experimentallyeither by measuring the formation of the ade-nylate or by "pyrophosphate exchange" (24).The second method is more commonly used, butmay give rise to misleading results since pyro-

phosphate exchange per se is not ncessarilyevidence for the formation of an amino acidadenylate.

Next, the amino acid residue of the adenylateis transferred to a low molecular weight ribo-nucleic acid (RNA) specific for the amino acidresidue being incorporated. These low molecularweight RNA molecules are usually called theacceptor or transfer RNA, and are part of the"soluble" RNA (SRNA). From the transferRNA, the amino acid residue is probably passedto its specific place on the RNA of the ribosome(or microsome) with the liberation of the trans-fer RNA, which can now accept another activatedamino acid residue. It is on the RNA of the ribo-some, it is thought, that the formation of thepeptide bonds occurs, with the liberation of the

TABLE 3. Activation of structural analogues ofnatural amino acids by a preparation from

Escherichia coli*

Pyrophosphate exchangeAmino acid (nwnole P~3n incor-porated into ATP: hr:

mg protein X 1I"

L-Valine 8.0D-Valine 0DL-Norvaline 6.7a-Amino-fl-chlorobutyric 7.5

L-Methionine 4.6D-Methionine 0DL-Selenomethionine 4.8DL-Ethionine 1.5DL-Norleucine 0.8

L-Phenylalanine 1.1DL-p-Fluorophenylalanine 0.1

L-Tryptophan 7.0DL-5-Methyltryptophan 0

* Results recalculated from Nisman and Hirsch(67). Activation measured by "pyrophosphateexchange. "

intact peptide chain. It seems probable that theamino acids carried on the specific transfer RNAare inserted into their correct position in thepeptide chain by a sequence of determinantscontained in the base sequence of the "messen-ger" RNA on the ribosome.

Effect on Amino Acid Activation. Nisman andHirsch (67) have shown that norvaline, a-amino-,B-chlorobutyric acid, p-fluorophenylalanine,selenomethionine, ethionine, and norleucine areactivated by extracts from E. coli (Table 3).Both norvaline and a-amino-g-chlorobutyric acidare activated at approximately the same rate asL-valine, and selenomethionine as rapidly asmethionine. Ethionine and norleucine stimulateto 30 and 15% of the methionine rate, and p-fluorophenylalanine only reaches 10% that ofphenylalanine. 5-Methyltryptophan is not ac-tivated. Since p-fluorophenylalanine will replacethe majority of phenylalanine in the proteins ofE. coli under conditions where the rate of proteinsynthesis is about half that found in the absenceof the analogue (65), the rate of activation of ananalogue has little relation to its rate of incorpora-tion into protein. For example, extracts of aphenylalanine-requiring mutant of E. coli (strain

409

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

M. H. RICHMOND

C2) activated thienylalanine at the same rate asphenylalanine, yet thienylalanine could not sup-port RNA and protein synthesis in the absenceof phenylalanine (91). Equimolar quantities ofphenylalanine and thienylalanine were activatedat the same rate as either amino acid alone (meas-ured by the hydroxamate method), yet no in-formation was given about the relative amountsof the two amino acids converted to the ade-nylate. Experiments with the purified trypto-phan-activating enzyme from pancreas (81) haveshown that azatryptophan, tryptazan, and 6-and 5-fluorotryptophan can be activated by thetryptophan-activating enzyme, but that 5-methyltryptophan, 6-methyltryptophan, and 6-methyltryptazan are not. The formation of ade-nylates was measured both by the hydroxamatetests and by pyrophosphate exchange. Further-more, azatryptophan hydroxamate was actuallyisolated from the reaction products. Though it istrue that all the amino acids incorporated intoproteins (including analogues) have now beenshown to undergo an activation step, it is notpossible to argue that the activation of aminoacid means that an amino acid is incorporatedinto protein (81).

Sharon and Lipmann's work with the pan-creatic enzyme does show, however, that certainanalogues can inhibit competitively the action ofthe activating enzyme. Thus, 5- and 6-methyl-tryptophan at 200 times the concentration oftryptophan inhibit activation of the naturalamino acid by about 70%. The physiologicalsignificance of this is rather doubtful since a rela-tive concentration of this order within the cellcould only be achieved if the analogue, but noneof the natural amino acid, was present in thegrowth medium.

Recently Berg (6) and Bergmann et al. (11)have shown that certain amino-acid-activat-ing enzymes are also capable of transferring theamino acids to either the 2'- or 3'-hydroxyl group(which group is uncertain) of a terminal ade-nylic acid in the specific acceptor RNA. Theyhave called the enzymes RNA synthetases. Thevalyl RNA synthetase was found to transferone of the structural isomers of a-amino-0-chlorobutyric acid at the same rate as L-valine,although the Km of the enzyme for the analoguewas 3.3 times as great. The other isomer, theallo-a-amino-,f-chlorobutyric acid, was muchless active. Presumably this must mean that the

active sites in the enzymes responsible for han-dling valine can distinguish between the twomethyl groups in the 4,4 positions of valine.The kinetics were determined by "pyrophosphateexchange," and there is, therfore, no evidencethat the a-amino-fl-chlorobutyric acid actuallyforms a stable complex with the acceptor RNA.The effect of this analogue has also been in-vestigated on hemoglobin synthesis in reticulo-cytes (73). Addition of valine-C14 to a partiallypurified hemoglobin-synthesizing system led toa flow of valine-C14 to the microsomal particles,to soluble proteins, and to hemoglobin. If thesystem was incubated with a-amino-,3-chloro-butyric acid, hemoglobin synthesis was blocked.However, if small quantities of valine-C'4 wereadded after prior incubation with the analogue,valine-C14 was found in the microsomal particles,but no valine-C'4 could be detected in hemoglobin,even if precautions were taken (e.g., addition of"carrier" hemoglobin) to trap all the hemoglobinsynthesized. Addition of larger quantities ofvaline led to restoration of hemoglobin synthesis.Complete reversal of inhibition was obtained byan analogue to valine molar ratio of 2:1. Similarresults with lysine-C'4 make it unlikely that theanalogue is blocking only the incorporation ofvaline into protein. Taken in conjunction with theresults obtained with E. coli (11), these datamake it seem likely that a-amino-,8-chlorobutyricacid is activated, passed to the acceptor RNA,and thence to the microsome (or ribosome).Whether it is true that no hemoglobin is formed,or whether some abnormal protein is synthesizedwhich cannot be detected in the normal hemo-globin detection system, cannot yet be decided.

It follows from the experiments discussed inthis section that the majority of analogues canbe activated to a greater or lesser extent by thesystems active on the natural amino acids. Mostexperiments have been carried out in vitro withbroken cell preparations and in the absence ofexogenous natural competitors. Clearly this is anartificial situation which is unlikely to decidewhether competitive inhibition of amino acidactivation by an analogue can cause growthinhibition. In the one case in which an activatingenzyme is inhibited by an analogue (5-methyl-and 6-methyltryptophan) under physiologicalconditions, relatively enormous quantities of theanalogue are required. There is not enoughevidence to show whether analogues can inhibit

410 [VOL. 26

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

1962] AMINO ACID ANALOGUES AND GROWTH AND PROTEIN SYNTHESIS

growth by blocking the transfer of activatedamino acids to the acceptor RNA or by orderingthe amino acids on the microsome (or ribosome).The results obtained with a-amino-f3-chloro-butyric acid on hemoglobin synthesis suggestthat this compound reaches the acceptor RNAbut is not incorporated into protein.

C. Incorporation of Amino Acid Analogues intoCell Proteins

The majority of useful amino acid analoguesinhibit protein synthesis in bacteria to someextent, but a 3- to 4-fold increase in cell proteinfrom the point of addition of the inhibitor isfrequently found. Analysis of the amino acidcontent of the protein formed in this way oftenshows that the analogue amino acid has beenincorporated in place of the naturally occurringcompetitor. To obtain formal proof of the in-corporation of an analogue into protein it isnecessary to isolate from the cell protein peptidescontaining the analogue. Usually it is sufficientto show that the analogue is present in a proteinfraction so that neither the -GOOH nor the-NH2 group of the analogue are accessible toreagents without hydrolysis. The presence offree -NH2 groups can be shown by reactionwith fluorodinitrobenzene followed by isolationof the N-dinitrophenyl derivative of the analogue(75). The presence of free -COOH groups canbe detected by treating the protein before hy-drolysis with ninhydrin. If the analogue is in-volved in a peptide bond, it will not be accessibleto ninhydrin, and should be recoverable afterhydrolysis.The incorporation of an unnatural amino acid

into the total mixed proteins of a tissue was firstshown by Levine and Tarver (53) using ethyl-C14-labeled ethionine. Although the ethionine wasnot reisolated from the protein in these experi-ments, subsequent work with Tetrahymenapyriformis confirmed that ethionine-C'4 wasincorporated and ethionine-C'4 peptides couldbe isolated (41). It was not possible, however,to decide whether the ethionine incorporationoccurred at the expense of the methionine nor-mally present in the protein.

Subsequently it has been shown that norleucine(65); tryptazan (13); 7-azatryptophan (69); o-,m-, and p-fluorophenylalanine (63, 65); canava-nine (50, 75); ,B-2-thienylalanine (65); seleno-methionine (86); and p-aminophenylalanine

(unpublished observations) are incorporated intothe protein of bacteria, and there is strong cir-cumstantial evidence that ,B-phenylserine is alsoincorporated (47). It is certain that incorpora-tion of norleucine occurs at the expense of methi-onine, and not leucine or isoleucine, and p-fluoro-phenylalanine and thienylalanine at the expenseof phenylalanine, but not tyrosine. Incorporationof p-fluorophenylalanine or thienylalanine hasno effect on the incorporation of any aminoacid in the protein other than phenylalanine,and the composition of the protein synthesizedin the presence of the analogue is thereforeunchanged except for the phenylalanine-p-fluorophenylalanine or phenylalanine-thienylala-nine transposition (65, 77). Although a numberof analogues are incorporated, the level that theyreach in the cell protein varies markedly. Munierand Cohen (65) report that p-fluorophenylalanineor thienylalanine replaces phenylalanine by about75% in the total mixed cell protein of E. coli,and a similar degree of replacement was found inB. cereus (77). In addition, norleucine and ethio-nine replace 50% of the methionine of E. coli,whereas 7-azatryptophan only reaches about 30%of the tryptophan normally found in E. coliproteins. The tyrosine analogue, p-aminophenyl-alanine, only reaches a maximal level of about2 to 5% of the tyrosine content of protein.Canavanine replaces arginine (at least partially)in the Walker carcinoma system, since additionof canavanine leads to a fall in the argininecontent of protein (50), but whether the uptakeof canavanine exactly compensates for the fallin arginine in this system or bacteria is not known(75).

It is possible that selenomethionine can replacemethionine completely (22). The analogue willallow exponential growth of a methionine-re-quiring mutant of E. coli in the absence of me-thionine, although the doubling time of theculture is 75 min rather than 55 which is achievednormally. Since the mutant can use neitherinorganic sulfate nor cysteine as a source ofmethionine, it follows that the analogue mustreplace all the methionine residues in the protein.Furthermore growth can continue in the pres-ence of selenomethionine alone for up to 100generations under conditions in which there islittle chance of carry-over of methionine beingresponsible for the growth. An unexplainedfinding is that organisms growing exponentially

411

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

M. H. RICHMOND

in selenomethionine will not grow when trans-ferred to solid media containing selenomethionineto satisfy their methionine requirement.

Addition of p-fluorophenylalanine to B. cereusresults in replacement of 75% of the phenyl-alanine of the protein, and this occurs at thesame rate in the total cell protein, the totalexocellular protein, and the purified exopenicil-linase synthesized by the organism (77). It seemsvirtually certain, therefore, that p-fluorophenyl-alanine replaces phenylalanine to about the sameextent throughout all the proteins of the cell. InE. coli, Cowie and his collaborators (23) haveshown that either norleucine or p-fluorophenyl-alanine show an approximately equal degree ofreplacement in a number of crude protein frac-tions. The organisms were grown in radioactiveanalogue, and the soluble protein was chromato-graphed on diethylaminoethyl (DEAE)-cellulose.The specific analogue concentration (micromolesof analogue incorporated per milligram of protein)was determined in certain fractions from thecolumn, and no significant variation was foundfor either compound. These results also suggestan equal degree of replacement in all bacterialproteins. However, this is not absolutely certainsince the specific analogue concentration wasdetermined only at a relatively few points in theelution diagram, and the experimental errorsinvolved do not allow calculations to be madewith an accuracy of better than t20%. Further-more, the soluble protein may not be representa-tive of the total cell protein.Yoshida and Yamasaki (95) have studied the

problem of replacement of amino acids by ana-logues in purified a-amylase from Bacillus subtilis.This molecule normally contains four methionineresidues. The synthesis of ca-amylase was carriedout in a medium containing ethionine-S35 and thepurified enzyme was digested with trypsin. Inthis way Yoshida was able to isolate three of thefour expected methionine and ethionine-contain-ing peptides and measure the methionine toethionine ratio. The ratios were: peptide 1,0.64:0.36; peptide 2, 0.61:0.39; peptide 3, 0.60:0.4 against a ratio in the total mixed protein fromthe cell of 0.59:0.41. These results suggest randomreplacement. However, Yoshida was unable tofind the fourth peptide, and this inevitablyleaves open the possibility that the replacementis not completely random.

Incorporation of analogues into enzyme proteins.It is not possible to deduce much about the effect

of analogues on specific enzyme activity unlessone is in a position to compare enzyme activityand quantity of enzyme protein directly. Variousanalogues have been reported to affect the ac-tivity of enzymes in microbial cultures. How-ever, only in a few cases has care been taken tosee whether the inhibitory effect is specific forthe enzyme and not a reflexion of the generalinhibitory effect that most analogues have onprotein synthesis. To establish a prima-facie casethat an analogue has a specific effect on anenzyme it is necessary to show that the differen-tial rate (increase in enzyme activity vs. increasein culture density) of appearance of enzymeactivity is lowered. Even an effect of this kindmust be interpreted cautiously, for an artificiallimitation of the growth rate of E. coli can leadto differential effects on the rate of enzymesynthesis (56). In this case inhibition was causedby increased "glucose repression" following the"sparing" of glucose that occurs when growthis limited artificially. Thus the addition of ananalogue may lead to an effect on the differentialrate of enzyme synthesis because of increasedglucose repression rather than any effect on thespecific enzyme activity of the protein synthe-sized in the presence of the analogue. Azatryp-tophan almost certainly causes inhibition by a"sparing" action on glucose, since the analoguecompletely inhibited ,B-galactosidase synthesis inE. coli, whereas aspartyl-carbamyl transferasewas inhibited 50% and D-serine deaminase wasunaffected (70). For this reason, too, the experi-ments reported by Munier and Cohen (65) witha range of analogues are subject to some ambi-guity. In these experiments the rate of oxygenuptake was to measure the quantity of respiratoryenzymes present in the culture, and addition ofanalogues such as f3-thienylalanine, 5-methyl-tryptophan, norleucine, and p-fluorophenylala-nine led to impaired oxygen consumption. Thiscould indicate that the analogue inhibited thefurther synthesis of the respiratory enzymes to agreater or lesser extent. The suggestion that thiswas caused by the incorporation of analoguesinto the enzymes resulting in inactive proteinsis, however, less likely than that "glucose repres-sion" of the enzymes becomes more marked afteraddition of the analogues. Similarly, the effectsof analogues on ,B-galactosidase (17) and ,B-glucuronidase (83) synthesis are difficult tointerpret, paricularly since H. V. Rickenberg(personal communication) has recently suggested

412 [VOL. 26

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

19621 AMINO ACID ANALOGUES AND GROWTH AND PROTEIN SYNTHESIS

that the fl-galactosidase synthesihpresence of /3-thienylalanine is less Econditions of assay than the nornSome direct method of measuring

of enzyme protein present is theref(before deductions can be made aboutanalogues on specific enzyme activithe effect of analogues has been measbacterial enzymes, the exopenicillinsized by B. cereus (76, 77) and thea-amylase of B. subtilis strain H (9

Since ca-amylase from this strainprepared from organisms in the staticof growth, it is not possible to obtaihtial plot of appearance of enzyme adpresence of p-fluorophenylalanine tmeaningful. However, Yoshida (94)amylase from cultures grown in 72of DL-p-fluorophenylalanine per ml,mined the degree of replacement of pIby the analogue, together with the ereplacement on specific enzyme acenzyme synthesized in the lower corof the analogue contained about 2 p-flalanine molecules per molecule of ersynthesized in the higher concentratieThe specific enzyme activities of the Iwere 85% and 70% of normal, respecincorporation of analogue occurred at

Lia

z

z

cr

I

Q.

aTR

100

90

80

70

60

50

40

30

20

I0

/0

- if

-0go IX I I I

30 60MIN.

90

FIG. 8. Percentage of total phenylala7in exopenicillinase replaced by p-ialanine in the course of incubation inof the analogue. Normal exopenicillino7 phenylalanine residues. Symbols 0 a

separate experiments.

zed in thestable underaal enzyme.the amountore essentialthe effect ofty. To dateured on twoLase synthe-- exocellular4).is normallyonary phasen a differen-tivity in thehat is verypurified a-and 150 Agand deter-

lenylalanineaffect of this

100

80I

60

I-

cU, 4C4at0z

e 20

p

00

0

20 40 60 80 100/ PHE. REPLACE.

FIG. 9. Effect of replacement of phenylalanine(PHE.) by the p-fluoro derivative on the specificenzyme activity of exopenicillinase.

tivity. The of phenylalanine only, but it was impossible toncentrations decide whether replacement was random oruorophenyl- confined to certain residues. It was not possibleazyme; that to separate any enzyme with impaired specificon, about 4. activity from the complete mixture, althoughpreparations clearly some such protein must be present. Inter-ctively. The pretation of the results is made more difficultthe expense by the possibility that the active enzyme molecule

arises by the addition of a peptide to a previouslysynthesized amylase precursor.The exopenicillinase, isolated from B. cerews

grown in an amino acid medium + p-fluoro-,o * phenylalanine, was found to contain a maximum-0 \ of about 75% of the normal phenylalanine

0 residues replaced by the p-fluoro derivative. Thekinetics of replacement following the addition ofthe analogue are shown in Fig. 8. The normalexopenicillinase of B. cereus contains 7 phenylala-nine residues, so in this experiment an averageof 5to 6 out of the 7 residues seem to be replaced.In fact, replacement may be complete in theprotein synthesized at the latter stages of incuba-ion, since (unlike the incorporation of p-fluoro-phenylalanine into the total protein of E. coli

1o1To (65)), the degree of replacement in B. cereus isnot constant with time, but increases. In these

nine (PHE.) experiments the specific enzyme activity of theYun hen(PHE.) exopenicillinase fell to about 50% of normal underthe presenyce conditions where about 75% of the phenylalaninease contains had been replaced. The results presented in Fig. 9nd 0 denote show that there is fair agreement between the

degree of replacement and the net level of specific

413

11

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

M. H. RICHMOND

enzyme activity of the mixture of enzyme mole-cules. As with the a-amylase system, there is noevidence as to whether replacement is random ornot, but at least one type of enzyme moleculemust be synthesized under these conditions withan altered specific enzyme activity. Attempts toseparate this molecular type from the mixture ofenzyme proteins obtained in this way have, sofar, been unsuccessful (77). The immunologicalproperties of exopenicillinase containing p-fluorophenylalanine showed that the reaction ofthe enzyme with normal antiserum was altered(77).

In both the bacterial enzymes studied so far,incorporation of p-fluorophenylalanine led toimpaired specific enzyme activity. Vaughan andSteinberg (87) have followed the incorporationof the same analogue into the lysozyme andovalbumin synthesized by a hen oviduct prepara-tion. The lysozyme showed no fall in specificenzyme activity. Recently Westhead and Boyer(90) prepared aldolase and glyceraldehyde-3-phosphate dehydrogenase with normal specificenzyme activities from rabbits which have beenfed on diets rich in p-fluorophenylalanine. Thetwo enzymes contained 25% and 16% of theirphenylalanine replaced by the p-fluoro analogue,respectively. Incorporation of the analogue doesnot, therefore, always lead to lowered specificactivity. The impaired specific enzyme activitywith a-amylase and exopenicillinase could bebecause phenylalanine is involved either directlyin the active center or in controlling the configura-tion of the protein near the active center. Theeffect with a-amylase is particularly dramatic,since an average of only 4 out of the 18 residuesis replaced. Unfortunately, no studies on theenzyme kinetics of normal and analogue-con-taining enzymes have yet been reported. Suchstudies might throw some light on changes at theactive center.

V. EFFECT OF AMINO ACID ANALOGUES ONSYNTHESIS OF BACTERIOPHAGE

Various amino acid analogues have beenshown to block the maturation of viruses withincells, but in this review only the effect of ana-logues on the development of bacteriophage will beconsidered. 5-Methyltryptophan completelyblocks the appearance of mature T2 or T4 phagein E. coli (18, 19). It was suggested (4) that theanalogue blocked tryptophan synthesis by the

bacteria, which were then unable to supply anessential constituent of the phage. Subsequentwork has shown that this interpretation is prob-ably correct (see earlier).

Pardee and Prestidge (70) have shown thatazatryptophan inhibits the multiplication ofphage T2r in E. coli. At 2 to 5 ,ug of analogue perml, less than 20% of the infective centers weremaintained, but it was not possible to blockphage infection completely, even at high analogueconcentrations. The inhibition of phage produc-tion was not caused by nonrelease of otherwisemature phage. If azatryptophan was added to aculture within 10 min of infection and the ana-logue then reversed by tryptophan at variousintervals, no mature phage was found on furtherincubation, although this treatment reversed thegrowth inhibition of the receptor cells. Duringthe first 10 min after infection, azatryptophanhad little effect on protein synthesis (65% ofnormal), but DNA synthesis was inhibited by75%. Since the bacteriophage DNA contains5-methylcytosine, it is possible to show that after16 min in the presence of the analogue, phageDNA is less than 30% of the level in the unin-hibited system. The proteins whose synthesis isnecessary for the formation and release of maturephage can be classed in three main groups: (i)those necessary for the development of the phage,(ii) those which are to become part of the struc-ture of phage, and (iii) those involved in lysis ofthe cell to liberate the phage. Pardee andPrestidge suggest that azatryptophan interfereswith the synthesis of proteins of the first class,and that it may be the synthesis of 5-methyl-cytosine which is blocked. Since the total bacterial+ phage protein at this point contains about0.5% azatryptophan, they suggest that the in-hibition is caused by the incorporation of theanalogue into the proteins concerned. Investiga-tion of the phage structural proteins showed thatfew viable phages were synthesized in the pres-ence of the analogue, and no intact T2 phagecould be seen on electron micrographs. Themajority of the phage DNA is liberated from thecells without its protein envelope, whose syn-thesis is presumably disrupted by the analogue.The purified phage proteins obtained in thepresence of the analogue will not kill or adsorb tobacteria. Titration of the abnormal protein withanti-T2 serum showed that only 20% of the

414 [VOL. 26

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

1962] AMINO ACID ANALOGUES AND GROWTH AND PROTEIN SYNTHESIS

protein was precipitable, whereas with normal T2protein the reaction went virtually to completion.

It seems likely from these experiments, there-fore, that azatryptophan blocks phage formationby interfering primarily with the synthesis andfunction of proteins and thereby with the syn-thesis of phage DNA.

VI. CONCLUSIONSThe foregoing sections have dealt primarily

with the experimental facts about the effects ofanalogues on growth and protein synthesis inbacteria. From these results it is possible to drawsome tentative general conclusions about theways in which bacteria respond to a structuralanalogue of a natural amino acid. The responsedepends largely on three factors:

1) The extent to which the analogue satisfiesthe substrate specificity of the enzymes whichnormally handle the natural amino acid.

2) The availability of the natural amino acid.3) The extent to which the incorporation of

the analogue into proteins in the place of thenatural amino acid modifies the structure of theproteins and the enzymic activity of enzymes.

Clearly the vast majority of compounds whichare formally amino acids (i.e., having the generalformula R-CH(NH2)-COOH) have no effects onbacteria, certainly not those of the type charac-teristic of the useful structural analogues. This is,presumably, because the bacteria are imper-meable to such compounds.

If an amino acid analogue does have a suffi-ciently similar shape, size, and ionic configurationto the natural compound to use a common trans-port system, then in most cases the analogue willsatisfy the structural requirements of the otherenzymes involved in using the amino acids forprotein synthesis to some extent. In the case ofanalogues that are incorporated into protein,there is nothing to suggest that the active centersof the enzymes involved in protein synthesisbecome more exacting toward their substratesthe later they lie in the reaction sequence leadingto peptide bond formation. Certain amino-acid-activating enzymes, however, activate structuralanalogues at rates of the same order as thoseachieved with the "natural" substrates, yet theanalogues are incorporated into proteins onlyat 1% or less of the rate that this implies. Clearly,in this case some reaction with a higher specificitymust lie between activation and the formation

of the intact protein. Work with a-amino-f3-chlorobutyric acid suggests that the limiting stepis the one involving actual peptide formation(73). The methyl-substituted tryptophans arealso exceptions. These compounds do pass intothe cell, but act as repressors or inhibitors oftryptophan biosynthesis. The compounds are notactivated or incorporated into protein, and it isdoubtful even whether they pass into the cellby the same route as tryptophan. There is everyreason to believe, however, that norleucine;selenomethionine; o-, m-, and p-fluorophenyl-alanine; thienylalanine; azatryptophan; tryp-tazan; and ethionine are taken up into the cellby the route used by the natural amino acid andare then incorporated into proteins by the en-zymes normally used for that purpose by thenatural compound. In all the cases that have beentested so far, amino acid analogues are found toenter proteins by replacement of naturally oc-curring structural analogue, but no other aminoacid. This suggests that analogues are selectedfor insertion into specific positions in the cellproteins by the determinants used for the naturalamino acids.

In practice this means that when an aminoacid analogue which satisfies the structural re-quirements for active transport into the cell isadded to a bacterial culture (which lacks ex-ogenous competitor), the compound will be in-corporated randomly in the place of the naturalamino acid in all residues in all proteins syn-thesized after addition of the analogue. Further-more, unless there is any of the natural aminoacid present free in the organism, replacementby the analogue will be complete. (See, forexample, the effect of canavanine on a strain ofStaphylococcus aureus exacting as to arginine (74),and of selenomethionine on a methionine-re-quiring mutant of E. coli (23).)

If the natural amino acid is present in theculture, either in the medium or the free aminoacid pool, competition will occur with the ana-logue for the active centers of the enzymes in-volved, and the quantity of the analogue flowingto protein will be reduced. Often it is found thatthe natural amino acid is handled preferentiallyby the transport system when a mixture ofstructural analogues is present in the growthmedium.

If the natural amino acid is not present in thegrowth medium, the analogue flows unhindered

415

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

M. H. RICHMOND

into the cell. This means that the enzymesresponsible for synthesis of protein will be pre-sented with a mixture of analogue and naturalamino acid at a ratio determined by the rate ofinflow of analogue through the transport system,the rate of production of the natural amino acidby the biosynthetic system, and the rate of lossof the biosynthetic compound from the freeamino acid pool to the medium. If the analoguehas no effect on the supply of endogenouslysynthesized competitor, the enzymes of the cellwill continue to be fed with a mixture of structuralanalogues, and the degree of replacement in theprotein will be determined by the relative Kmvalues of the enzymes for the amino acids.Maximal replacement by an analogue should beexpected, therefore, when the external analogueconcentration is sufficient to saturate the freeamino acid pool. Ethionine is an analogue thatprobably acts in this way since it does not seemto inhibit or repress methionine biosynthesis.A number of analogues, however, (e.g., p-

fluorophenylalanine, thienylalanine, 6-fluorotryp-tophan) do inhibit the endogenous synthesis ofthe natural amino acid either by feedback inhibi-tion or repression. In this case, as soon as theanalogue enters the free amino acid pool of thecell it will compete with the natural amino acidfor the active centers of the enzymes involvedin protein synthesis, but also will start to blockthe flow of the endogenously synthesized naturalamino acid. As time goes on (and providing thatthe level of analogue in the free amino acid poolis maintained by an adequate external concen-tration), the ratio of analogue to natural aminoacid presented to the enzymes will move in favorof the analogue, and this will result in a largerreplacement in protein.The kinetics characteristic of this process

depend upon whether the block in endogenoussupply is caused by feedback inhibition of anenzyme in the biosynthetic pathway or repressionin the formation of all the enzymes of biosyntheticpathway. If inhibition occurs and is complete,the endogenous supply of amino acid will behalted virtually instantaneously, and competitionwith the analogue will last only as long as thelevel of the natural compound in the free aminoacid pool lasts. This will be shown by an almostimmediate and maximal flow of analogue to theproteins and an immediate effect on growth. Onthe other hand, if repression occurs (without any

inhibition), the rate of endogenous synthesis maybe as much as 50% of normal after one generation,and the ratio of analogue to natural amino acidfed to the enzymes will change relatively slowly.This will be reflected in an increasing rate ofincorporation of analogue into protein and adelayed effect on growth. It is possible that thedifferences between the effects of p-fluorophenyl-alanine on growth and protein synthesis in E.coli and B. cereus (see earlier) could be explainedon the basis of an inhibition of endogenous supplyin E. coli but repression in B. cereus.Replacement of a natural amino acid by an

analogue in an enzyme can affect specific enzymeactivity, and this may have a number of repercus-sions on the response of organisms to amino acidanalogues. First (since analogues are thought tobe incorporated at random into all sites) replace-ment will occur in the proteins responsible forthe biosynthesis of all amino acids. If such en-zymes have a lowered specific activity, thisprocess could further reduce the supply of en-dogenously synthesized competitor. On the otherhand, this process may equally cause a loweringof the specific activity of the enzymes responsiblefor protein synthesis, and this would affect anatural amino acid and an analogue equally.The over-all effect would probably be a net lower-ing in the rate of total protein synthesis. Only ifa low degree of incorporation into an enzymecarrying out a normally rate-limiting reactioncauses complete loss of enzyme activity will therate of growth and protein synthesis immediatelybecome linear with time. Cohen and Munier(17, 65) have suggested that this does occur whenE. coli is treated with (3-thienylalanine.

Effect of incorporation of analogues on physico-chemical properties and activity of enzymes. Sincean analogue must satisfy the specificity of theactive sites of several enzymes to be incorporatedsuccessfully into enzyme proteins, it must be aclose structural analogue of an amino acid nor-mally found in proteins. It is unlikely, therefore,that an effective analogue will differ in structuresufficiently from the normal compound to havemuch effect on the physicochemical properties ofproteins into which it is incorporated. This isparticularly so since changes of this sort in theprotein would only occur with changes in theionic, stereochemical, or hydrogen-bonding prop-erties of an analogue, and these are just the

416 [VOL. 26

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

1962] AMINO ACID ANALOGUES AND GROWTH AND PROTEIN SYNTHESIS

sort of changes that would be likely to make theanalogue unacceptable to the cell.

It is less possible to be dogmatic about theeffect of analogues on specific enzyme activitysince the way in which an active center operatesis so little understood. So far exopenicillinase anda-amylase have been shown to be affected by in-corporation of analogues, whereas alkaline phos-phatase ofE. coli (77a), lysozyme from henoviductpreparations (87), and aldolase and glyceralde-hyde-3-phosphate dehydrogenase from rabbitmuscle were unaffected (90). In general it ispossible to suggest tentatively that the effect ofan analogue on the activity of an enzyme maydepend upon the following factors: (i) the natureof the analogue and the degree to which it differsfrom the natural amino acid, (ii) whether aresidue implicated in the active center is replaced(There is no evidence yet to suggest that such aresidue would be less accessible to replacementthan any other in the protein.), and (iii) whetherany residues are replaced which are involved inmaintaining the tertiary structure of the enzyme.Too little is yet known about the mechanism ofenzyme action to make application of thesefactors more than tentative.

Future prospects for use of analogues. In thepast the use of amino acid analogues has notconferred any particular advantage over the useof the natural amino acids in elucidating thesteps in biochemical reactions, particularlyprotein synthesis. Usually the processes havebeen worked out with natural amino acids andthen the tolerances allowable in the varioussteps have been tested by analogues.The discovery that analogues can be incor-

porated into individual enzyme proteins hasopened up the possibility of using these com-pounds to study the mechanism of action of theactive center. It would be particularly interestingin this connection to develop an analogue whichcould be incorporated into a protein in vivo andthen allowed to react in various ways after isola-tion and purification of the proteins. Such com-pounds as p-aminophenylalanine could be idealfor such purposes. A similar approach could alsoyield valuable information about the mechanismof enzyme-antibody interactions, since the reac-tion with the analogues in situ could be used tocomplement the studies made by treating theamino acids occurring naturally in proteins with

various reagents and studying the effect onenzyme neutralization.

VII. ACKNOWLEDGMENTS

I should like to thank M. R. Pollock, J. Man-delstam, and J. F. Collin for much helpfuldiscussion during the preparation of this review.

VIII. LITERATURE CITED

1. ACKERMAN, W. W., A. ROBSON, AND H. KUNTZ.1953. Growth characteristics of poliovirusin Hela cell cultures. Lack of parallelismin cellular injury and virus increase. J.Exptl. Med. 100:437-450.

2. ACKERMAN, W. W., AND H. F. MAASAB. 1955.Growth characteristics of influenza virus.Biochemical differentiation of stages ofdevelopment. II. J. Exptl. Med. 102:393-402.

3. ADELBERG, E. A. 1958. Selection of bacterialmutants which excrete antagonists of anti-metabolites. J. Bacteriol. 76:326.

4. ANDERSON, T. F. 1945. The activity of a bac-teriostatic substance in the reaction betweenbacterial virus and host. Science 101:565-566.

5. AVAKIAN, S., J. Moss, AND G. J. MARTIN.1948. Synthesis and microbiological prop-erties of ,- (2-benzothienyl)-a-amino-pro-pionic acid. J. Am. Chem. Soc. 70:3075-3076.

6. BERG, P. 1961. Specificity in protein synthesis.Ann. Rev. Biochem. 30:293-324.

7. BERGMANN, E. D. 1954. 5-Fluoro-tryptophanand its effect on growth of Escherichia coli.Koninkl. Ned. Akad. Wetenschap. Ser. C.57:8.

8. BERGMANN, E. 1961. Synthetic and chemicalstudies on organic fluorine compounds.Bull. Research Council Israel. Sect. A.10:1-16.

9. BERGMANN, E. D., E. ESCHINAZI, S. SICHER,AND B. E. VOLCANI. 1952. The action of 5-fluorotryptophan on Eacherichia coli. Bull.Research Soc. Israel 2:308-309.

10. BERGMANN, E., S. SICHER, AND B. E. VOLCANI.1953. Action of substituted phenylalanineson Escherichia coli. Biochem. J. 54:1-13.

11. BERGMANN, F. H., P. BERG, AND M. DIECK-MANN. 1961. The enzymic synthesis of aminoacyl derivatives of ribonucleic acid. II.The preparation of leucyl-, valyl-, isoleucyl-,and methionyl ribonucleic acid synthetasesfrom Escherichia coli. J. Biol. Chem. 236:1735-1740.

12. BOYER, P. D., R. J. EVANS, AND P. H. PHIL-

417

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

48M. H. RICHMOND

LIPS. 1941. The effect of certain fluorinatedphenylalanine and tyrosine derivatives onthe rat. J. Pharmacol. 73:176-185.

13. BRAWERMAN, G., AND M. Y6AS. 1957. In-corporation of amino acid analog tryptazaninto the protein of Escherichia coli. Arch.Biochem. Biophys. 68:112-117.

14. BRITTEN, R. J., AND F. T. MCCLURE. 1962.The amino acid pool in Escherichia coli.Bacteriol. Rev. 26:292-335.

15. COHEN, G. N., AND H. V. RICKENBERG. 1956.Concentration spdcifique reversible desamino acides chez Escherichia coli. Ann.inst. Pasteur 91:693-720.

16. COHEN, G. N., H. 0. HALVORSON, AND S.SPIEGELMAN. 1958. Effects of p-fluoro-phenylalanine on the growth and physiologyof yeast, p. 100. In Microsomal particles andprotein synthesis. Pergamon Press, London.

17. COHEN, G. N., AND R. MUNIER. 1959. Effetsdes analogues structuraux d'aminoacides surla croissance, la synthbse de protdines et lasynthbse d'enzymes chez Escherichia coli.Biochim. Biophys. Acta 31:347-356.

18. COHEN, S. S., AND T. F. ANDERSON. 1946.Chemical studies on host-virus interactions.The chemical simulation of the interferencephenomenon by 5-methyl tryptophane. J.Exptl. Med. 84:525-532.

19. COHEN, S. S., AND D. FOWLER. 1947. Chemicalstudies on host-virus interactions. J. Exptl.Med. 85:771-784.

20. COLLINS, J. F., AND M. H. RICHMOND. 1962.Growth of Bacillus cereus between divisions.J. Gen. Microbiol. 28:15-33.

21. COWIE, D. 1959. In Carnegie Institution ofWashington Year Book. v. 58, p. 294.

22. COWIE, D. B., AND G. N. COHEN. 1957. Bio-synthesis by Escherichia coli of active al-tered proteins containing selenium insteadof sulphur. Biochim. Biophys. Acta 26:252-261.

23. COWIE, D. B., G. N. COHEN, E. T. BOLTON,AND H. DE ROBICHON-SZULMAJSTER. 1959.Amino acid analog incorporation into bac-terial proteins. Biochim. Biophys. Acta34:39-46.

24. DAVIE, E. W., V. V. KONINGSBERGER, ANDF. LIPMANN. 1956. The isolation of trytophanactivating enzymes from liver. Arch. Bio-chem. Biophys. 65:21-38.

25. DAVIS, A. L., J. M. RAVEL, C. G. SKINNER,AND W. SHIVE. 1958. Synthesis and biologicalactivity of some diamino acid analogs. Arch.Biochem. Biophys. 76:139-147.

26. DAVIS, A. L., C. G. SKINNER, AND W. SHIVE.1960. 3-Amino-methylcyclohexaneglycine, a

lysine analog. Arch. Biochem. Biophys.87:88-92.

27. DENNIS, R. L., W. J. PLANT, C. G. SKINNER,G. L. SUTHERLAND, AND W. SHIVE. 1955.Synthesis of 2-cyclopentene-1-glycine, aninhibiting amino acid analog. J. Am. Chem.Soc. 77:2362-2364.

28. DITTMER, K. 1950. The structural basis ofsome amino acid antagonists and theirmicrobiological properties. Ann. N. Y.Acad. Sci. 52:1274-1302.

29. DUNN, F. W. 1958. Inhibition studies withB3-2-thienylalanine, f3-1-naphthylalanine, p-tolyalanine and their peptides. J. Biol.Chem. 233:411-414.

30. DUNN, F. W. 1959. Preparation and micro-biological properties of tripeptides of j3-2-thienylalanine. J. Biol. Chem. 234:802-808.

31. DUNN, F. W., AND K. DITTMER. 1951. Thesynthesis and microbiological properties ofsome peptide analogues. J. Biol. Chem.188:263-272.

32. DUNN, F. W., J. M. RAVEL, AND W. SHIVE.1956. Inhibition studies with peptides ofthienylalanine and phenylalanine. J. Biol.Chem. 219:809-821.

33. EDELSON, J., P. R. PAL, C. G. SKINNER, ANDW. SHIVE. 1957. Synthesis of 2-cyclohexene-1-glycine and 1-cyclohexene-l-alanine, in-hibiting amino acid analogs. J. Am. Chem.Soc. 79:5209-5212.

34. EDELSON, J., J. D. FISSEKIS, C. G. SKINNER,AND W. SHIVE. 1958. 3-Cyclohexene-l-glycine, an isoleucine antagonist. J. Am.Chem. Soc. 80:2698-2700.

35. EDELSON, J., C. G. SKINNER, J. M. RAVEL,AND W. SHIVE. 1959. 3-Cyclohexene-1-DL-alanine, an analog of leucine. Arch. Biochem.Biophys. 80:416.

36. EDELSON, J., C. G. SKINNER, J. M. RAVEL,AND W. SHIVE. 1959. Structural specificityof some amino acid antagonists. J. Am.Chem. Soc. 81:5150-5153.

37. ELLIOTT, D. F., AND C. R. HARINGTON. 1949.Bacteriostasis in the amino acid series. II.Further studies with alanine derivatives.J. Chem. Soc. p. 1374-1378.

38. FILDES, P. 1941. Inhibition of bacterial growthby indole acrylic acid and its relation totryptophan. Brit. J. Exptl. Pathol. 22:293-298.

39. FOWDEN, L. 1962. The non-protein aminoacids of plants. Endeavour 21:35-42.

39a.FOWDEN L., AND M. H. RICHMOND. 1962. Bio-chim. Biophys. Acta (to be published).

40. GROS, F., AND F. GROS. 1958. RMle des acidesamines dans la synthbse des acides nucl6-

418 [VOL. 26

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

19621 AMINO ACID ANALOGUES AND GROWTH AND PROTEIN SYNTHESIS

iques chez Escherichia coli. Exptl. CellResearch 14:104-131.

41. GROSS, D., AND H. TARVER. 1955. Studies onethionine. Incorporation of ethionine intoproteins of Tetrahymena. J. Biol. Chem.217:167-182.

42. HALVORSON, H. O., AND S. SPIEGELMAN.1952. The inhibition of enzyme formationby amino acid analogues. J. Bacteriol. 64:207-220.

43. HARDING, W. M., AND W. SHIVE. 1954. Cyclo-pentaneglycine, an inhibitory analogue ofisoleucine. J. Biol. Chem. 206:401-410.

44. HASZELDINE, R. N., AND A. G. SHARPE. 1951.Fluorine and its compounds. Methuen, &Co., Inc., London.

45. HOAGLAND, M. B. 1960. The relationship ofnucleic acid and protein synthesis as re-vealed by cell-free systems, p. 349-408. InThe nucleic acids. v. III. Academic Press,Inc., New York.

46. Interatomic Distances. 1958. Tables of inter-atomic distances and configurations inmolecules and ions. Chemical SocietySpecial Publication No. 11, London.

47. JANA6EK, J., J. CHALOUPKA, K. VERES, ANDM. HAVRANEK. 1960. Regulation of proteinsynthesis in Escherichia coli by amino acidanalogues. Folia Microbiol. 5:207-216.

48. KALYANKAR, G. D., M. IKAWA, AND E. E.SNELL. 1958. The enzymic cleavage of cana-vanine to homoserine and hydroxyguani-dine. J. Biol. Chem. 233:1175-1178.

49. KERRIDGE, D. 1960. Effect of inhibitors on theformation of flagella by Salmonella typhi-murium. J. Gen. Microbiol. 23:519-538.

50. KRUSE, P. F., JR., P. B. WHITE, H. A. CARTER,AND T. A. McCoy. 1959. Incorporation ofcanavanine in protein of Walker carcino-sarcoma 256 cells cultured in vitro. CancerResearch 19:122-125.

51. LANSFORD, E. M., AND W. SHIVE. 1952. Growthinhibition by some ,-arylalanines. Arch.Biochem. Biophys. 38:347-350.

52. LESTER, G., AND C. YANOFSKY. 1961. Influenceof 3-methyl anthranilic acid and anthranilicacid on the formation of tryptophan synthe-tase in Escherichia coli. J. Bacteriol. 81:81-90.

53. LEVINE, M., AND H. TARVER. 1951. Incorpora-tion of ethionine into rat proteins. J. Biol.Chem. 192:835-850.

54. McLAREN, A. D., AND C. A. KNIGHT. 1953.The response of Leuconostoc masenteroidesP60 to some compounds related to lysine.J. Biol. Chem. 204:417-421.

55. MCQUILLEN, K., R. B. ROBERTS, AND R. J.

BRITTEN. 1959. Synthesis of nascent proteinby ribosomes in Escherichia coli. Proc. Natl.Acad. Sci. U. S. 45:1337-1347.

56. MANDELSTAM, J. 1962. The repression of con-stitutive ,P-galactosidase in Escherichia coliby glucose and other carbon sources. Bio-chem. J. 82:489-493.

57. MILLER, D. A., AND S. SIMMONDS. 1957.Phenylalanine and tyrosine metabolism inEscherichia coli K12. Science 126:445-446.

58. MITCHELL, H. K., AND C. NIEMANN. 1947.The competitive inhibition of the metabo-lism of a-amino acids by their halogenatedanalogs. J. Am. Chem. Soc. 69:1232.

59. MOYED, H. S. 1960. False feedback inhibition.Inhibition of tryptophan biosynthesis by5-methyltryptophan. J. Biol. Chem. 235:1098-1102.

60. MOYED, H. S. 1961. Interference with the feed-back control of histidine biosynthesis. J.Biol. Chem. 236:2261-2267.

61. MOYED, H. S., AND B. MAGASANIK. 1959. Therole of purines in histidine biosynthesis.J. Am. Chem. Soc. 79:4812-4813.

62. MOYED, H. S., AND M. FRIEDMANN. 1959.Interference with feedback control. Amechanism of antimetabolite action. Science129:968-969.

63. MUNIER, R. 1959. Substitution totale de laphenylalanine par l'o- ou la m-fluorophenyl-alanine dans les proteines d'Escherichia coli.Compt. rend. 248:1870-1873.

64. MUNIER, R., AND G. N. COHEN. 1956. In-corporation d'analogues structuraux d'ami-noacides dans les proteines bactdriennes.Biochim. Biophys. Acta 21:592-593.

65. MUNIER, R., AND G. N. COHEN. 1959. In-corporation d'analogues structuraux d'ami-noacides dans les proteines bacteriennes aucours de leur synthbse in vivo. Biochim.Biophys. Acta 31:378-390.

66. NICOLLE, J., AND J. WALLE. 1961. Action desantipodes optiques de la norleucine sur lacroissance de Salmonella typhimurium etd'Escherichia coli. Compt. rend. 252:3339-3340.

67. NISMAN, B., AND M-L. HIRSCH. 1958. Etude deI'activation et de l'incorporation des acidesaminds par des fractions enzymatiquesd'Escherichia coli. Ann. inst. Pasteur95:615-634.

68. PAL, P. R., C. G. SKINNER, R. L. DENNIS,AND W. SHIVE. 1956. Cyclopentane alanineand 1-cyclopentene-1-alanine, inhibitinganalogues of leucine and phenylalanine.J. Am. Chem. Soc. 78:5116-5118.

69. PARDEE, A. B., AND L. S. PRESTIDGE. 1958.

419

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from

M. H. RICHMOND

Effects of azatryptophan on bacterial en-

zymes and bacteriophage. Biochim. Bio-phys. Acta 27:330-344.

70. PARDEE, A. B., AND L. PRESTIDGE. 1959. Onthe nature of the repressor of f3-galactosidasesynthesis in Escherichia coli. Biochim.Biophys. Acta 36:545-547.

71. RABINOVITZ, M., M. E. OLSON, AND D. N.GREENBERG. 1955. Steric relationship be-tween threonine and isoleucine as indicatedby an antimetabolite study. J. Am. Chem.Soc. 77:3109-3111.

72. RABINOVITZ, M., AND R. K. TUvE. 1959. Anti-metabolite activity of lysine analogues inlysine incorporation into rat bone marrow

protein in vitro. Proc. Soc. Exptl. Biol.Med. 100:222-224.

73. RABINOVITZ, M., AND H. MCGRATH. 1961.Protein synthesis by rabbit reticulocytes.II. Interruption of the pathway of hemo-globin synthesis by a valine analog. J. Biol.Chem. 234:2091-2095.

74. RICHMOND, M. H. 1959. The differential effectof arginine and canavanine on growth andenzyme formation in Staphylococcus aureus

524SC. Biochem. J. 73:155-167.75. RICHMOND, M. H. 1959. Incorporation of

canavanine by Staphylococcus aureus 524SC.Biochem. J. 73:261-264.

76. RICHMOND, M. H. 1960. Immunological prop-erties of exopenicillinase synthesized byBacillus cereus 569/H in the presence ofamino acid analogues. Biochem. J. 77:112-121.

77. RICHMOND, M. H. 1960. Incorporation ofDL-,3-(p-fluorophenyl)-/6-14C-alanine into exo-

penicillinase by Bacillus cereus 569/H. Bio-chem. J. 77:121-135.

77a.RIcHMoND, M. 1962. Biochem. J. 84:110P.78. ROBERTS, R. B. 1958. Microsomal particles

and protein synthesis. Pergamon Press,London.

79. ROBLIN, R. 0., JR., J. 0. LAMPEN, J. P.ENGLISH, Q. P. COLE, AND J. R. VAUGHAN,JR. 1945. Studies in chemotherapy. VIII.Methionine and purine antagonists and theirrelation to sulfonamides. J. Am. Chem. Soc.67:290-294.

80. SCHULTISSEK, C., AND R. ROTT. 1961. Influ-ence of p-fluorophenylalanine on the pro-duction of viral RNA and on the utiliza-bility of viral protein during the multiplica-tion of fowl plague virus. Nature 191:1023-1024.

81. SHARON, N., AND F. LIPMANN. 1957. Reac-tivity of analogs with pancreatic trypto-phan-activating enzyme. Arch. Biochem.Biophys. 69:219-227.

82. SHRIFT, A. 1961. Selenium metabolism inplants and animals. Federation Proc. 20:695-727.

83. STOEBER, F. 1961. fPtudes des proprietes et dela biosynthese de la glucuronidase et de laglucuronide-permease chez Escherichia coli.These. C.D.U. & S.E.D.E.S., Paris.

84. Symposium. 1958. Symposium on amino acidactivation. Proc. Natl. Acad. Sci. U. S.44:67-131.

85. TRUDINGER, P. A., AND G. N. COHEN. 1956.The effect of 4-methyl-tryptophan on growthand enzyme systems of Escherichia coli.Biochem. J. 62:488-491.

86. TUVE, T., AND H. WILLIAMS. 1957. Identifica-tion of selenomethionine in the proteins ofEscherichia coli employing the chromato-graphic "fingerprint" technique. J. Am.Chem. Soc. 79:5830-5831.

87. VAUGHAN, M., AND D. STEINBERG. 1960. Bio-synthetic incorporation of fluorophenylala-nine into crystalline proteins. Biochim.Biophys. Acta 40:230-236.

88. WALKER, J. B. 1955. Homoarginine inhibitionof Escherichia coli B. J. Biol. Chem. 212:617-621.

89. WECKER, E., AND E. SCHONNE. 1961. Inhibitionof viral RNA synthesis by p-fluorophenyl-alanine. Proc. Natl. Acad. Sci. U. S. 47:278-282.

90. WESTHEAD, E. W., AND P. D. BOYER. 1961.The incorporation of p-fluorophenylalanineinto some rabbit enzymes and other pro-

teins. Biochim. Biophys. Acta 54:145-156.91. WOLFE, A. D., AND F. E. HAHN. 1960. Dis-

crepancy between thienylalanine activationand protein synthesis in bacteria. Biochim.Biophys. Acta 41:545-547.

92. WOODS, D. D. 1940. The relation of p-aminobenzoic acid to the mechanism of action ofsulphanilamide. Brit. J. Exptl. Pathol.21:74-90.

93. WOOLLEY, D. W. 1952. A study of antimetabo-lites, p. 47. John Wiley & Sons, Inc., NewYork.

94. YOSHIDA, A. 1960. Incorporation of p-fluoro-phenylalanine into a-amylase of Bacillussubtilis. Biochim. Biophys. Acta 41:98-104.

95. YOSHIDA, A., AND M. YAMASAKI. 1959. Studieson the mechanism of protein synthesis.Incorporation of ethionine into a-amylaseof Bacillus subtilis. Biochim. Biophys. Acta34:158-164.

96. ZIMMERMANN, T., AND W. SCHXFFER. 1960.Effect of p-fluorophenylalanine on fowlplague virus multiplication. Virology 11:676-698.

420 [VOL. 26

on April 16, 2021 by guest

http://mm

br.asm.org/

Dow

nloaded from