ANwICROBIAL AGzNTs AND CHIMOTHURAPY, Feb. 1973, p. 181-187Copyright © 1973 American S3ociety for Microbiology

Vol. 3, No. 2Printed in U.S.A.

DD-Carboxypeptidase-Transpeptidase and Killing Siteof -Lactam Antibiotics in Streptomyces Strains

R39 R61, and KllJEAN DUSART, ALBERTO MARQUET, JEAN-MARIE GHUYSEN, JEAN-MARIE FRIRE,RAMON MORENO, MELINA LEYH-BOUILLE, KENNETH JOHNSON,1 CHANTAL LUCCHI,

HAROLD R. PERKINS, AND MANUEL NIETOService de Microbiologie, Facultd de Mddecine, Institut de Botanique, Universitt de Lihge, Sart-Tilman,

4000 Libge, Belgium; National Institute for Medical Research, Mill Hill, London, England;and Centro de Inve8tigaciones Biologicas, Instituto de Biologia Celular, Madrid, Spain

Received for publication 4 October 1972

Additional evidence is given that in Streptomyces strains R39, R61, and Kll thesame enzyme performs DD-carboxypeptidase and transpeptidase activities and thatthis enzyme is the killing site of #-lactam antibiotics. With strain R61, it was foundthat the exocellular enzyme has a sensitivity towards some antibiotics differentfrom that of the membrane-bound enzyme. Under the growth conditions used inthe present investigations, ,B-lactamase activity was not involved in susceptibilityto fl-lactam antibiotics.

The susceptibility of bacteria to penicillinsand cephalosporins is a complex phenomenon.It depends upon (i) the sensitivity of the targetitself, i.e., the membrane-bound transpeptidasewhich cross-links the peptides of the nascentpeptidoglycans (14, 15), (ii) the fi-lactamaseactivity of the bacteria (2), and (iii) the per-meability of the bacterial envelope to the anti-biotics and the nonspecific fixation of the anti-biotic by cell-envelope components other thanthe transpeptidase (12). During growth, Strep-tomyces strains R39, R61, and Kll excrete$-lactamases and DD-carboxypeptidases-trans-peptidases. The efficiency and specificity profileof the ,B-lactamases of strains Kll and R39towards gS-lactam antibiotics has been described(7). The DD-carboxypeptidases-transpeptidasesexcreted by strains R39, R61, and Kll wereshowin to function as hydrolytic enzymes (car-boxypeptidases) or synthesizing enzymes (trans-peptidases) depending upon the availability ofnucleophilic groups (water or NHrR). Theseenzymes had a considerable specificity in theirsubstrate requirements both for donor peptidesand for acceptor amino groups, strongly sug-gesting that they were the membrane-boundtranspeptidases which had undergone solubili-zationi (8, 9, 11).In the experimenits described in this paper, we1 Permanent address: Division of Biological Sciences,

National Resarh Council of Canada, Ottawa, Ontario,Canada.

studied the in vivo susceptibility of Strepto-myces strains to #-lactam antibiotics as relatedto their P-lactamase activity and the propertiesof their exocellular DD-carboxypeptidases-trans-peptidases and membrane-bound transpepti-dases.

MATERIALS AND METHODSAntibiotics. 6-Aminopenicillanic acid, benzyl-

penicillin, substituted bensylpenicillins, phenoxy-methylpenicillin, and various cephalosporins wereused (see Fig. 1 in the accompanying paper [71).Culture media. Peptone-oxoid medium and

glycerol-casein medium have been describedpreviously (8). Agar-CGY medium contained,per liter of final volume: agar, 20 g; CasaminoAcids (Difco), 3 g; glycerol, 30 ml; and yeast ex-tract (Difoo), 3 g. Agar-KC medium contained,per liter of final volume: agar, 20 g; partiallyhydrolyzed keratin from hens' feathers, 0.25%(w/v); partially hydrolyzed casein, 0.25% (w/v);NaCl, 0.5 g; CaCO,, 0.1 g; MgS0O47H,O, 0.1 g;K2HP04, 1 g; and FeSO4-7H20, 0.1 g. Hydrolyzedkeratin and casein were prepared as describedpreviously (4). Agar-APG medium contained,per liter of final volume: agar, 20 g; asparagine,0.5 g; peptone, 0.5 g; glucose, 10 g; and K2HP0,,0.5 g.

Streptomyces strains. Streptomyces strainsR39, R61, and Kll were used. R61 and Kll weremaintained on slants of agar-APG, and R39 wasmaintained on slants of agar-KC. Viable countswere determined as follows. Conidia were col-lected from slants with the help of a 1:10,000

181

ANTIMICROB. AG. CHEMOTHER.

dilution of sodium dodecyl sulfate (6). Afterthorough shaking, the mycelial debris was re-moved from the suspensions by filtration on What-man no. 2 paper. The conidia suspensions werediluted to contain 100 to 200 "colony-formingunits" per ml. Viable counts were made by spread-ing 1 ml of the suspensions on agar-CGY plates.Incubation was at 30 C.#3-Lactamase activity of isolated Strepto-

myces colonies. The enzyme activity of isolatedcolonies was estimated by using the starch-iodinetest of Foley and Perret (3), modified as follows.About 5 to 10 conidia (colony-forming units)were spread on GA-6 filters (porosity, 0.45 gm;diameter, 47 mm; Gelman Instrument Co.), de-posited on agar-CGY plates. Incubation at 30 Cwas maintained until the colonies were 2 mm indiameter. The filters were then transferred to afreshly prepared starch-iodine indicator agarcontaining: agar, 1.8%/o (w/v); iodine, 0.0127%(w/v); KI, 0.0254% (w/v); soluble starch, 0.3%(w/v); sodium phosphate buffer (pH 7.4), 10 mm;and benzylpenicillin, 200 gg/ml. j-Lactamaseactivity provokes decolorization around thecolonies. The diameters of decolorized areas weremeasured after increasing times at 30 C (up to 90min). When penicillin was not present in thestarch-iodine agar, decolorization around thecolonies did not occur under the above conditions.Exocellular DD - carboxypeptidases - trans -

peptidases. The purified, exocellular DD-carboxy-peptidases - transpeptidases from Streptomycesstrains R39, R61, and Kll have been described(8, 9). One enzyme unit catalyzes the hydrolysisof 1 nanoequivalent of D-alanyl-D-alanine linkageper hr at 37 C when the enzyme is exposed toNe,N, - diacetyl - L - lysyl - D - alanyl - D - alanineat concentrations 10 times the Km values, in theabsence of any acceptor amino group. The Kmvalues are 0.8 mm for the R139 enzyme and 12mm for R61 and Kll enzymes.Carboxypeptidase activity of exocellular

enzymes and its inhibition by ,B-lactam anti-biotics. N,N' - diacetyl - L - lysyl - D - alanyl -D-alanine (1.23 mM, final concentration) was in-cubated with the enzymes for 2 hr at 37 C, in0.03 M K2HPO4, in the absence and in the presenceof various concentrations of antibiotics. Theamount of D-alanine liberated was estimated bythe fluorodinitrobenzene technique (5). In theabsence of antibiotic, the yield of hydrolysis was90%0 of the theoretical value.Transpeptidase activity of exocellular en-

zymes and its inhibition by 6-lactam anti-biotics. The conditions of the tests were as de-scribed above for the determination of carboxy-peptidase activity except that 14C-glycine wasadded to the reaction mixtures. The molar ratiosof 14C-glycine (i.e., the acceptor) to Na,NE-diacetyl - L - lysyl - D - alanyl - D - alanine (i.e.,the donor) were 10:1 for enzymes from R61 andKll and 20:1 for enzyme from R39. The amiiouintof NNa,N-diacetyl-L-lysyl-D-alanyl transferredfrom the tripeptide donor to "4C-glycine (i.e.,

the yield of transpeptidation) was determined asdescribed previously (11). The transpeptidationproduct, N-,N-diacetyl-L4ysyl-D-alanyl-"'C-gly-cine, was separated from the excess of "4C-glycineby paper electrophoresis at pH 5.6, and theamount of radioactivity incorporated was meas-ured with a Packard Tri-Carb liquid scintillationspectrometer. In the absence of antibiotic, theyield of transpeptidation was 40%/, expre#sed asconversion of the donor tripeptide.Membrane - bound transpeptidase from

Streptomyces R61. Membranes were preparedessentially as described by Munoz et al. (10).Streptomyces R61 was grown (volume of inoculum,5%) for 48 hr at 29 C, with vigorous shaking,in 1-liter flasks containing 500 ml of peptone-Oxoid medium. This medium was used rather thanglycerol-casein medium because less carboxy-peptidase-transpeptidase is excreted by the grow-ing bacteria in peptone-Oxoid medium than inglycerol-casein medium. The mycelium (from 2.5liters of culture) was harvested, washed fourtimes with distilled water by decantation, re-suspended in water, homogenized with a PotterElvehjem tissue homogenizer, and centrifuged at10,000 X g for 20 min. The pellet was washed twiceby centrifugation (10,000 X g, 20 min) with 0.1 mtris (hydroxymethyl)aminomethane (Tris)-hydro-chloride buffer, pH 7.5, and finally resus-pended in 100 ml of the same buffer supplementedwith 0.05 M MgC12 and 25 g of sucrose. Afterhomogenization as above, 250 mg of lysozymewas added, and the suspension was kept at 4 Cfor 20 hr. The pellet obtained after centrifugationat 12,000 X g for 15 min was suspended in 100 mlof 0.05 M Tris-hydrochloride buffer, pH 7.5, con-taining 0.005 M MgCl2 and 75 mg of deoxyribo-nuclease I (pancreatic). The suspension was keptat room temperature until complete lysis occurred(usually 16 hr). The lysate was centrifuged at1,100 X g for 10 min; the pellet was resuspendedin 100 ml of 0.05 M Tris-hydrochloride buffer(pH 7.5) plus 0.005 M MgCIs and centrifugedagain at 1,100 X g for 10 min. The supernatantfluids of the two last centrifugations were com-binied and centrifuged at 37,000 X g for 30 min,yielding the membrane preparation. Membraneswere suspended in a volume of 0.03 M KHPO4such that the suspension contained about 25 mgof protein per ml. The membrane suspensionscould be stored at 4 C, in the presence of thymol,for 6 weeks without loss of transpeptidase activ-ity. The same technique as above was applied toStreptomyces R139, but the membranes did notexhibit transpeptidase or carboxypeptidase activ-ities.

Activity of the Streptomyces R61 mem-brane-bound transpeptidase and its inhibi-tion by,B-lactam antibiotics. NaN-diacetyl-Ix-lysyl-D-alanyl-D-alanine (1.35 mm) and 14C-glycyl-glycine (13.5 mm) were incubated for 16 hrat 37 C, with 500 pg (in protein) of R61 membrane,in a final volume of 35 pliters of 0.016 M K2HPO4containing 0.3% (v/v) Triton X-100 (final con-

182 DUSART ET AL.

KILLING SITE OF ,3-LACTAM ANTIBIOTICS 183

centrations). Incubations were carried out in theabsence and in the presence of various concen-trations of antibiotics. The transpeptidationproduct, Na,NE - diacetyl - L - lysyl - D - alanyl-14C-glycyl-glycine, was separated from the ex-cess of "4C-glycyl-glycine by paper electrophore-sis at pH 5.6 for 2 hr at 30 v/cm, and the amountof radioactivity incorporated was measured asdescribed above. In the absence of antibiotic,the yield of transpeptidation (expressed as con-version of the donor tripeptide) was 30 to 40%,depending upon the membrane preparation.Transpeptidation reactions with acceptors otherthan glycyl-glycine were also carried out underthe same conditions. When no radioactive accep-tor was available, the radioactive tripeptide4C-diacetyl-L-lysyl-D-alanyl-D-alanine was used,and the radioactive transpeptidation product4C-diacetyl-L-lysyl-D-alanyl-acceptor was sepa-rated by electrophoresis at pH 6.45 for 4 hr at 60v/cm and estimated (Perkins et al., Biochem.J., in press).

RESULTSEffect of B-lactam antibiotics on growth

of Streptomyces (LDso values). As for manyother bacteria, the susceptibility of Streptomycesto ,-lactam antibiotics is critically dependentupon the size of the cell inoculum. To ensuremaximal reproducibility in the quantitative esti-mates, we measured susceptibility by observingthe effect of the presence of various concentra-tions of antibiotics on the ability of conidia toform "single-cell" colonies. Colonies were counteddaily until maximal values were obtained (5 to 8days, depending upon the strain, the antibiotic,and its concentration). In the absence of anti-biotic, control plates contained about 100 to

200 colonies. The LD50 values (Table 1 to 3)were determined graphically as the concentra-tions of antibiotics allowing 50% cell survival.The data show that the antibiotic susceptibilityof strains R61 and Kl 1 was almost the same.These strains were less susceptible than strainR39, except to 6-aminopenicillanic acid andcephalosporin C. Great variations were observeddepending upon the antibiotic.

Effect of ,6-lactam antibiotics on activityof the exocellular carboxypeptidases-trans-peptidases of Streptomyces (ID50 values,soluble enzyme). The effect of the antibioticson carboxypeptidase activity was expressed asthe concentration which decreased by 50% theamount of D-alanine liberated from Na,NI-diacetyl-L-lysyl-D-alanyl-D-alanine. Similarly, theeffect of antibiotics on transpeptidase activitywas expressed as the concentration which re-duced by 50% the amount of transpeptidationproducts. The ID50 values are also reported inTables 1-3. These data show that inhibition ofthe transpeptidase activity of the exocellularenzymes always occurred at those concentrationsof antibiotics which inhibited the carboxypep-tidase activity. When the whole series of anti-biotics is taken into account, there appears to beno correlation at all between the in vivo LD50values and the in vitro ID60 values for the exo-cellular enzymes. The in vivo susceptibility ofthe Streptomyces strains may be either similarto, or considerably higher or lower than, thein vitro sensitivity of the exocellular enzymes.Membrane-bound carboxypeptidase-trans-

speptidase activity of Streptomyces R61.Since the catalytic properties of an enzyme

TABLE 1. Effect ofJ-lactam^ antibiotics otn Streplornyces strain R890

ID50Antibiotic LDro IDso (avg)/LD5o

Carboxy- Trans-peptidase peptidase Avg

6-Aminopenicillanic acid. 5,100 500 650 575 0.113Penicillin G ............. 70 3.8 7.1 5.4 0.077Penicillin V ............. 140 3.5 4.6 4.0 0.029Ampicillin .............. 240 3.25 3.6 3.4 0.014Carbenicillin ............ 830 195 200 197 0.24Oxacillin ................ 710 12 33 22 0.031Cloxacillin .............. 530 29 46 37 0.070Methicillin .............. 730 42 53 47 0.064Cephalosporin C.... 9,000 4.7 6.0 5.3 0.00059Cephaloglycine. 120 5.4 7.4 6.4 0.053Cephalexin .............. 15 26 17 21 1.4Cephalothin ............. 90 2.9 5.5 4.2 0.047

Comparison between the lethal action (LD5o values) and the inhibitioni of the exocellular DD-car-boxypeptidase-transpeptidase activity (ID5o values). Values for LD60 and IDro represent 108 X molarconcentration.

VOI,. 3, 1973

ANTIMICROB. AG. CHEMOTHER.

TABLE 2. Effect of P-lactam antibiotics on Streptomyces strains Kll and R61a

Antibiotic

6-Aminopenicillanicacid................

Penicillin G..........Penicillin V..........Ampicillin............Carbenicillin .........Oxacillin .............Cloxacillin.Methicillin...........Cephalosporin C......Cephaloglycine .......Cephalexin ...........Cephalothin..........

LD50

K1l

1,160390390670

3,0803,3002,6303,1805,6201,630510590

R61

1,850560510540

3,3104,7105,0404,41011,2202,8602,0601,400

ID50

Carboxypeptidase

Kll

135,00015

1103,300400

2,1003,50015,000

515,00050,000

62

R61

130,00012

1903,000640

1,9007,00014,000

466,60060,000

. 80

Transpeptidase

Kll

135

3,4004,70017,500

R61

130,00015

2,200900

427,600

50,00080

ID5o/LD0ob

K11 R61

1160.0380.314.90.130.831.65.10.00913.1980.105

700.0240.374.80.230.401.43.20.00392.5

270.057

a Comparison between the lethal action (LD5o values) and the inhibition of the exocellular DD-car-boxypeptidase-transpeptidase activity (ID5o values). Values for LD5o and ID5o represent 108 X molarconcentration.

b When the data are available, the IDro values used in this ratio are the average values betweencarboxypeptidase and transpeptidase. In the other cases, the carboxypeptidase values alone were used.

TABLE 3. Effect of ,-lactam antibiotics onStreptomyces strain R61a

ID6. formembrane- ID50/Antibiotics LDso bound LD6.

trans-peptidase

6-Aminopenicillanicacid ............... 1,850 8,000 4.3

Penicillin G.......... 500 215 0.38Penicillin V .......... 510 1,850 3.6Ampicillin ........... 540 800 1.5Carbenicillin ......... 3,310 3,400 1.0Oxacillin............. 4,710 8,800 1.9Cloxacillin ........... 5,040 3,900 0.78Methicillin ........... 4,410 3,500 0.79Cephalosporin C ..... 11,220 51,000 4.5Cephaloglycine...... 2,860 6,200 2.2Cephalexin ........... 2,060 6,000 2.9Cephalothin.......... 1,400 7,000 5.0

a Comparison between the lethal action (LD5ovalues) and the inhibition of the transpeptidaseactivity of the membrane-bound enzyme. Valuesfor LD5o and IDso represent 108 X molar concen-tration.

may depend uponi its localizatioin, whether itoccurs extracellularily or is membrane-bound,the carboxypel)tidase-transl)eptidase activitiesexhibited by the plasma membranes preparedfrom strain R61 (Materials and Methods) were

tested under a variety of experimental conditions.Figure 1 shows the effect of increasing the molarratio of acceptor to donor from 1:1 to 20:1 onthe amount of Na, N1-diacetyl-L1ySyl-D-alanyl-glycyl-glycine (transpeptidation product) formedfrom N,NI-diacetyl-LIysyI-D-alanyl-D-alanine(donor, at a constant concentration of 1.68 mM)and '4C-glycyl-glycine (acceptor) by the R61membrane-bound enzyme. Incubation was for16 hr at 37 C with R61 membranes (500 ug ofprotein) in a final volume of 37 uliters of 0.017 MK2HPO4 containing 0.27% Triton X-100. At amolar ratio of acceptor to donor of 10:1, theyield of transpeptidation was almost maximal(35%, expressed as conversion of the tripeptidedonor) and was proportional to the amount ofmembrane present in the reaction mixture. TritonX-100 (0.27%) was not necessary for the trans-peptidation reaction to occur and did not solu-bilize the enzyme activity which remained sedi-mentable (40,000 X g, 30 min). However, thepresence of the detergent in the reaction mixturefacilitated the separation of the transpeptidationproduct by paper electrophoresis (Materialsand Methods). Table 4 shows that the speci-ficity l)rofile of the membrane-bound transpep-tidase for various acceptors (with Na,NE-diacetyl-IAysyl-D-alanyl-D-alanine as donor) resemblesthat of the exocellular carboxypeptidase-trans-peptidase enzyme. Glycine, D-alanimie, glycyl-glycine, glycyl-L-alanine, racemic diaminoadipic

184 DUSART ET AL.

KILLING SITE OF 13-LACTAM ANTIBIOTICS 185

L0-

20

Donor N N!dlAc-L-Lys-D-AIa-D-AIa1.68 mM

ratio Acceptor- 8/ ~~~~~Donor I

GlyGly Acceptor Concentration mM

10 20 30FIG. 1. Effect of increasing ratios of acceptor to

donor on transpeptidation by the membrane-boundenzyme from Streptomyces R61. The reaction mix-tures contained, in a final volume of 37 pliters:membranes, 500 pg of protein; Na,N-diacetyl-L-lysyl-D-alanyl-D-alanine, 1.68 mM; glycyl-glycine,from 1.68 to 88.6 mM; Triton X-100, 0.27%; andK2HPO4, 0.017 M. Incubation uw for 16 hr at 37 C.Transpeptidation is expressed as conversion of thetripeptide donor.

acid, and 2-amino-D-galacturonic acid are goodacceptors for both exocellular and membrane-bound enzymes. Glycyl-D-alanine is a poor ac-ceptor, and L-alanine, L-alanyl-L-alanine, andD-alanyl-D-alanine are not acceptors at all.However, in spite of this similarity between thesubstrate requirements, the relative yields oftranspeptidation with the various acceptors arenot identical for the two enzyme preparations.Under the same conditions as above, and even

in the absence of acceptor, the R61 membranepreparation did not hydrolyze the tripeptidedonor into D-alanine and Na,N'-diacetyl-L-lysyl-D-alanine. This lack of carboxypeptidaseactivity is no doubt related to the fact that thepurified exocellular DD-carbox.ypeptidases-trans-peptidases from Streptomyces R61 and R39can be forced to work preferentially as trans-peptidase rather than as carboxypeptidase byproperly adjusting the conditions of the invitro assays (Ghuysen et al., Frere et al., inpreparation).

Effect of ft-lactam antibiotics on the trans-peptidase activity of the membrane-boundenzyme from Streptomyces R61. In markedcontrast to what was observed with the exocellu-lar enzyme, there was a high correlation be-tween the ID50 values of the membrane-boundenzyme and the in vivo LD50 values of strainR61 (Table 3). In comparing Tables 2 and 3,one should note that, in those instances whenthe correlation between the ID50 values of theexocellular enzyme and the LD60 values is good(in the range of 0.5 to 5.0), the correlationbetween the corresponding ID50 values for the

membrane-bound enzyme and the LD 0 valuesremains of the same order of magnitude. How-ever, in the cases where the ID50 values for thesoluble enzyme do not correlate with the LDr0values, the corresponding ID50 values for themembrane-bound enzyme do correlate well withthe LD60 values.g-Lactamase activity of Streptomyces

strains R39, R61, and KII. The f-lactamaseactivity of Streptomyces colonies on benzyl-penicillin was estimated under growth conditionssimilar to those used for the estimations of thein vivo LD50 values (Materials and Methods).The tests were performed on isolated colonies ofidentical size by the starch-iodine method. Thediameters of the decolorized zones around thecolonies were 0, 2.5, and 4 mm, respectively, forstrains R61, R39, and Kl1 after 10 min at

TABLE 4. Profiles of the membrane-bound enzymeand exocellular enzyme from Streptomyces strainR61 for acceptors in transpeptidation reactions

with N , N-diacetyl-L-lysyl-D-alanyl-D-alanine as donora

Acceptor

Glycine.....D-Alanine..L-Alanine........Glycyl-glycine...

Glycyl-L-alanine.

Glycyl-D-alanine.D-Alanyl-D-

alanine ........

L-Alanyl-L-

alanine ........

2-Amino-3-galac-turanic acid...

Racemic diami-noadipic acid..

Membrane-bound enzyme

Ac-ceptor

todonorratio

10:110:110:110:12:110:110:1

10:1

10:1

2:1

2:1

Yield oftrans-pepti-dation

18250

2617112

0

0

12

9

Exocellularenzyme

AcceptorYield of

to donor trans-ratio pepti-

dation

1:11:1

10:11:1

1:11:1

1:1

1:1

5:1

5:1

1480

18

254

0

0

18

54

a Yield of transpeptidation is expressed as per-centage of conversion of the tripeptide donor. Forthe experiments carried out with the membrane-bound enzyme, the tripeptide donor was notradioactive and the acceptors were 'IC-labeled(glycine, D-alanine, L-alanine, glycyl-glycine) or,conversely, the tripeptide donor was "4C-labeledon the acetyl groups and the acceptors were notradioactive. Conditions used with membrane-bound enzyme are given in the text; those forexocellular enzyme are given in reference 12.

O._

a'i

I~-

VOL. 3, 1973

ANTIMICROB. AG. CHEMOTHERI.

30 C; after 90 min of incubation, the corre-sponding diameters were 5, 9, and 10 mm. Thus,strain Kll has a higher fl-lactamase activitythan strain R61, although these two strains areequally susceptible to the antibiotics. Similarly,strain R39 has a higher 0-lactamase activitythan strain R61, although strain R39 is moresusceptible to the antibiotics than strain R61.

If 0-lactamase activity is involved in the rela-tive susceptibility of bacteria to ,-lactam anti-biotics, the susceptibility of a given strain to aseries of antibiotics should bear an inverse re-lationship to the catalytic efficiency and speci-ficity profile of the ,B-lactamase for the sameantibiotics. Comparison of the LD50 values ofthe antibiotics for Streptomyces Kll and R39with the profiles and efficiencies of the correspond-ing fl-lactamases (see Table 2 in the accom-panying paper [7]) clearly confirms that thesusceptibility of these Streptomyces strains to,-lactam antibiotics is not related to theirf3-lactamase activities, at least under the growthconditions used for the tests.

DISCUSSIONA complete lack of correlation of the in vivo

susceptibility of the Streptomyces strains to#-lactam antibiotics, the #-lactamase activityof the strains, and the efficiency and specificityprofiles of their ,-lactamases was observed.#-Lactamase activity thus appears to be ir-relevant to antibiotic susceptibility, at leastunder the conditions used for the determinationof the ability of conidia to form colonies in thepresence of antibiotics. Although for manybacteria there is a high correlation between peni-cillinase production and penicillin resistance,there are also numerous examples in the litera-ture which are not consistent with the idea thatpenicillinase is the prime factor involved in re-sistance (13).

Irrespective of the structure of the P-lactamantibiotics used, inhibition of the transpeptidaseactivity of the three exocellular enzymes fromStreptomyces R39, R61, and Kll always oc-curred at those antibiotic concentrations thatinhibited carboxypeptidase activity. This pro-vides additional support for the idea previouslyproposed that in Streptomyces sp. the same en-zyme performs both carboxypeptidase andtranspeptidase activities.The exocellular enzymes from Streptmoyces

R61 and Kll are indistinguishable from eachother but differ greatly from the exocellularenzyme from Streptomyces R39 with respect tosubstrate requirements for peptide donor and

acceptor, to Km and VmaX values in carboxy-peptidase assays, and to sensitivity to ,-lactamantibiotics. Parallel to this, the in vivo LD50values of the ,8-lactam antibiotics for strainsR61 and Kll are very similar, and these valuesdiffer greatly from those for strain R39. How-ever, with the three Streptomyces strains studied,the lethal concentrations of some antibiotics arevery different from those that inhibit the invitro activity of the exocellular enzymes. Theexperimenits carried out with the membrane-bound enzyme from Streptomyces R61 providethe explanation of these discrepancies. Theseexperiments show that the sensitivity of theR61 exocellular enzyme towards some anti-biotics may be very different from that of thecorresponding membrane-bound enzyme. Anincreased resistance of the membrane-bound en-zyme, compared to that of the exocellular en-zyme, can be easily explained, since nonspecificbinding sites for these antibiotics are likely tooccur on the membrane. An increased sensitivityof the membrane-bound enzyme is more difficultto interpret. It suggests conformational changesof the enzyme as the result of its integration inthe lipophilic environment of the membrane.Integration of the enzyme into the membrane,however, does not seem to modify its generalspecificity profile for peptide acceptors.The present work suggests that in Strepto-

myces sp. carboxypeptidase and transpeptidaseactivities are performed by the same enzyme,and this enzyme, when integrated into the mem-brane, seems to be the killing site for ,-lactamantibiotics. Contrary to what has been observedin Streptomyces sp., it has been proposed byBlumberg and Strominger (1) that in BaciUlussubtilis the killing site, i.e., the transpeptidase,is distinct from the carboxypeptidase. Theirconclusion was mainly based on the fact that6-aminopenicillanic acid inactivated at least95% of the carboxypeptidase activity (toluene-treated cells) at nonlethal concentrations (liquidcultures) and, conversely, that cephalothin waslethal at concentrations that did not inactivatethe enzyme. One should note, however, that asimilar conclusion would have been drawn inthe case of Streptomyce8 R61 if, for example,onie had limited the investigations to the simplecomparison between the LD50 values for 6-amino-penicillanic acid and cephalosporin C and thecorresponding ID50 values of the exocellularcarboxypeptidase-transpeptidase. Indeed, 6-ami-nopenicillanic acid is 70 times more active onthe soluble enzyme than in vivo and, conversely,cephalosporin C is 200 times less active on thesoluble enzyme than in vivo.

186 DUSART ET AL.

KILLING SITE OF j-LACTAM ANTIBIOTICS 187

ACKNOWLEDGMENTS

This work was supported by grants from the Fondsde la Recherche Fondamentale Collective, Brussels(no. 1000), and the Institut pour l'Encouragement de laRecherche Scientifique dans l'Industrie et l'Agriculture,Brussels (no. 1699 and 2013), to J. M. Ghuysen, by anE.M.B.O. fellowship to Alberto Marquet, and by aN.A.T.O. fellowship to Kenneth Johnson. Jean-MarieFrere is charg6 de recherchesou, Fonds National dela Recherche Scientifique, Brussels, Belgium.

LITERATURE CITED

1. Blumberg, P. M.. and J. L. Strominger. 1971. In-activation of D-alanine carboxypeptidase bypenicillins and cephalosporins is not lethal inBacilus 8ubtili. Proc. Nat. Acad. Sci. U.S.A.68:2814-2817.

2. Citri, N. 1971. Penicillinase and other ,-lactamases,p. 23-41. In P. D. Boyer (ed.), The enzymes,vol. 4. Academic Press Inc., New York.

3. Foley, J. M., and C. J. Perret. 1962. Screeningbacterial colonies for penicillinase production.Nature (London) 195:287-288.

4. Ghuysen, J. M. 1957. Activit6s bact6riolytiques del'actinomyc6tine de Streptomycees albus G. Arch.Int. Physiol. Biochem. 65:173-305.

5. Ghuysen, J, M., D. J. Tipper, and J. L. Strominger.1968. Enzymes that degrade bacterial cell walls,p. 685-699. In S. P. Colowick and N. 0. Kaplan(ed.), Methods in enzymology, vol. 8. AcademicPress Inc., New York.

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

7. Johnson, K., J. Dusart, J. N. Campbell, andJ.-M. Ghuysen. 1973. Exocellular P-lactamases

of Streptomyces strains albus G and strains R39and Kl1. Antimicrob. Ag. Chemother. 3:289-298.

8. Leyh-Bouille, M., J. Coyette, J. M. Ghuysen, J.Idczak, H. R. Perkins, and M. Nieto. 1971.Penicillin-sensitive DD-carboxypeptidase fromStreptomyces strain R61. Biochemistry 10:2163-2170.

9. Leyh-Bouille, M., M. Nakel, J. M. Frere, K.Johnson, J. M. Ghuysen, M. Nieto, and H. R.Perkins. 1972. Penicillin-sensitive DD-carboxy-peptidases from Streptomyees strains R39 andKll. Biochemistry 11:1290-1298.

10. Munoz, E., A. Marquet, V. Larraga, and J. Coyette.1972. Isolation, partial characterization of thecytoplasmic membrane fraction of Streptomycesalbus G and DD-carboxypeptidase localization.Arch. Mikrobiol. 81:273-288.

1l. Pollock, J. J., J. M. Ghuysen, R. Linder, M. R. J.Salton, H. R. Perkins, M. Nieto, M. Leyh-Bouille, J. M. Frere, and K. Johnson. 1972.Transpeptidase activity of Streptomyces D-alanyl-D carboxypeptidases. Proc. Nat. Acad.Sci. U.S.A. 69:662-668.

12. Rogers, H. J. 1967. The inhibition of mucopeptidesynthesis by benzylpenicillin in relation to ir-reversible fixation of the antibiotic by Staphylo-cocci. Biochem. J. 103.90-102.

13. Saz, A. K. 1970. An introspective view of penicil-linase. J. Cell. Physiol. 76:397-403.

14. Tipper, D. J., and J. L. Strominger. 1965. Mech-anism of action of penicillins: a proposal based ontheir structural similarity to acyl-D-alanyl-D-alanine. Proc. Nat. Acad. Sci. U.S.A. 54:1133-1141.

15. Wise, E. M., Jr., and J. T. Park. 1965. Penicillin:its basic site of action as an inhibitor of a pep-tide crosslinking reaction in cell wall mucopep-tide synthesis. Proc. Nat. Acad. Sci. U.S.A.54:75-81.

VOL. 3, 1973