JOURNAL OF BACrERIOLOGY, May 1970, p. 476-482Copyright © 1970 American Society for Microbiology

Vol. 102, No. 2Printed in U.S.A.

Relative Affinity of Vancomycin and Ristocetin forCell Walls and Uridine Diphosphate-N-

Acetylmuramyl PentapeptideGARY K. BEST, MIRIAM K. GRASTIE, AND ROBERT D. McCONNELL

Department of Microbiology, Medical College of Georgia, Augusta, Georgia 30902

Received for publication 25 February 1970

A determination of the relative affinity of vancomycin and ristocetin for isolatedcell walls and for a peptidoglycan precursor was made. These antibiotics had pre-

viously been shown to adsorb to cell walls and to complex with peptides containinga D-alanyl-D-alanine C-terminus. By using 14C-uridine diphosphate (UDP)-N-acetylmuramyl pentapeptide, it was shown that the complex which is formed be-tween this peptidoglycan precursor and either vancomycin or ristocetin does notpreclude adsorption of the antibiotics to cell walls of Micrococcus lysodeikticus.Complex formation between ristocetin and UDP-N-acetylmuramyl pentapeptidewas assured by differential absorption spectra. However, when the complex was

mixed with cell walls, the antibiotic was sedimented with the walls, and the radio-activity remained in the supernatant solution. This indication that ristocetin andvancomycin have a greater affinity for walls than for UDP-N-acetylmuramyl penta-peptide and that the complex per se does not bind to cell walls suggests that adsorp-tion of these antibiotics to cell walls is probably responsible for the inhibition ofpeptidoglycan synthesis. This proposal is strengthened by the observation that com-plexed antibiotic is no less inhibitory for growth of Bacillus subtilis than free van-comycin or ristocetin.

Vancomycin and ristocetin are closely relatedantibiotic inhibitors of peptidoglycan synthesisby gram-positive organisms (2, 7, 15). There haverecently been two similar mechanisms advancedto account for the inhibition of cell wall synthesisby these antibiotics. The first, the mechanism ofBest and Durham (3), was based on the demon-strated adsorption of vancomycin and ristocetinto isolated cell walls of Bacillus subtilis. Theypresented evidence to suggest that ionic linkageswere involved in the adsorption and suggestedthat a physical or steric interference with peptido-glycan polymerization resulted from those anti-biotic molecules adsorbed proximal to the site(s)of addition of the disaccharide pentapeptideprecursor to the existing polymer (3).An alternative mechanism of action of these

antibiotics was recently proposed by Perkins (10).By using differential ultraviolet absorptionspectra, Perkins demonstrated that both vanco-mycin and ristocetin formed a stable complex ofunknown nature with not only the uridine diphos-phate (UDP)-N-acetylmuramyl pentapeptide pre-cursor for peptidoglycan synthesis, but also withseveral peptides. In each case, complex formation

was shown to require an acyl-D-alanyl-D-alanineC-terminal sequence. Since the required aminoacid sequence has only been found in associationwith the peptidoglycan precursor, Perkins con-sidered the specificity exhibited by vancomycinand ristocetin toward this peptide to be improba-ble if it was not related to the inhibition of cellwall synthesis.

Since two logical mechanisms of inhibition ofpeptidoglycan synthesis by vancomycin andristocetin exist, this study was undertaken todetermine whether the antibiotic complex betweeninsoluble cell wall or soluble peptidoglycan pre-cursor is more relevant to the biological activity.

MATERIALS AND METHODSMicroorganisms. Micrococcus lysodeikticus ATCC

4698 and Staphylococcus aureus strain Copenhagenwere obtained from F. C. Neuhaus of NorthwesternUniversity. B. subtilis W-23 was used for the growthinhibition studies.

Isolation of cell walls from Micrococcus lysodeikti-cus. Cell walls from M. lysodeikticus were preparedwith cells grown at 37 C in a medium consisting (perliter) of 10 g of Yeast Extract (Difco), 10 g of trypti-case soy broth, 5 g of peptone, 5 g of K2HPO4, and

476

VOL. 102, 1970 RELATIVE AFFINITY OF VANCOMYCIN AND RISTOCETIN

10 g of glucose was sterilized separately and addedto the cooled medium.

After 12 to 14 hr growth, the micrococci were har-vested by continuous-flow centrifugation at 25,000 Xg. The cell pellets were washed twice with distilledwater by centrifugation and broken with a Mickledisintegrator by using an equivalent weight of cells andglass beads (0.13 mm in diameter). Approximately75 to 80% cell breakage was obtained in 10 min bythis technique (as estimated by microscopic observa-tion). The walls were recovered as a white layer onthe unbroken cell pellet after centrifugation at 32,000X g for 20 min. The resulting walls were washedtwice with cold, distilled water and purified by incu-bation overnight at 37 C in 0.025 M tris(hydroxy-methyl)aminomethane (Tris)-hydrochloride buffer(pH 7.5) containing trypsin (100 ,g/ml) and ribo-nuclease (10,ug/ml). A thin layer of toluene was usedto retard contamination of the walls during this in-cubation. After enzyme treatment, the walls weresuccessively washed with 25-ml volumes of cold,distilled water until the A280 of the wash supernatantfraction was less than 0.05.The purity of these preparations was ascertained

by paper chromatography after acid hydrolysis (6 NHCI, 100 C, 18 hr). After removing the HCl by re-peatedly drying the hydrolysate on a steam table, theamino acids were separated on Whatman no. 1 filterpaper by using butanol-acetic acid-water (4:1:1upper phase) by descending chromatography. Sig-nificant amounts of only alanine, lysine, glutamicacid, and glycine were detected with ninhydrin.

Adsorption studies. The adsorption of vancomycinand ristocetin to cell walls was demonstrated by theprocedure of Best and Durham (3). Freshly preparedcell walls were suspended in distilled water to thedesired level by using a standard curve relating theA540 to dry weight. Adsorption was attained by incu-bating 2 mg of cell walls with 750 jug of antibiotic at25 C in a total liquid volume of 3 ml. Cell walls andbound antibiotic were sedimented by centrifugation(12,000 X g for 10 min), and the amount of antibioticadsorbed was ascertained by measuring the decreasein absorbance of the supernatant solution at 280 nmand relating this reading to a standard absorbancecurve for the antibiotic. All spectrophotometricreadings were made by using cuvettes with a 1-cmlight path.

Extraction and isolation of UDP-N-acetylmuramylpentapeptide from S. aureus. Four 1-liter flasks con-taining tryptose broth (500 ml) were inoculated toan A540 value of 0.10 with S. aureus from a 12-hrtryptose agar slant. Cell growth at 37 C was followedspectrophotometrically at 540 nm. When the ab-sorbance reached 0.6 to 0.7 [0.2 to 0.3 mg (dry weight)per ml], the cells were harvested by continuous-flowcentrifugation and washed once with 0.10 M potas-sium phosphate buffer (pH 7.0). The cells were thenresuspended in 500 ml of a medium containing (perliter): K2HPO4, 14 g; KH2PO4, 6 g; (NH4)2SO4, 2 g;glucose, 10 g; DL-alanine, D-glutamic acid, L-lysine,glycine, and penicillin, 200 mg each; and 1 ml of amineral salts solution containing 5% MgSO4, 0.1%MnSO4, and 1.0% FeCI3. Labeled nucleotide peptide

was prepared by adding 50 Mc of 14C-L-lysine (spe-cific activity 220 mc/mmole, uniformly labeled;New England Nuclear Corp.) to this incubation mix-ture.

After incubation at 37 C in the above medium for 2hr, the cells were collected by centrifugation (12,000 Xg for 10 min) and extracted by continuously stirringat 4 C with 50 ml of trichloroacetic acid (10%, w/v).After 1 hr, the suspension was centrifuged, and theyellow supernatant solution was recovered. The cellswere again suspended in 50 ml of 10% trichloroaceticacid and extracted as before. The combined superna-tant solutions were then extracted 4 to 5 times withequal volumes of cold, anhydrous ether to remove thetrichloroacetic acid. The resulting pale yellow liquidwas taken to near dryness by rotary evaporation at30 C and was then dissolved in 1 to 2 ml of distilledwater.The extracted nucleotides were applied to a Sepha-

dex G-25 column (50 by 3 cm) and eluted with dis-tilled water. Fractions (5 ml) were assayed spectro-photometrically at 260 nm and 0.1-ml portions werehydrolyzed with 0.1 N HCI (100 C for 100 min), neu-tralized with NaOH, and analyzed for N-acetyl-amino sugar content by the method of Reissig et al.(11). A single, broad peak of N-acetylamino sugarwas pooled and taken to near dryness by rotary evapo-ration at 30 C. This material was applied to anotherSephadex G-25 column (60 by 1.5 cm), rinsed ontothe bed with 2.0 ml of distilled water, and eluted withdistilled water. Fractions (5 ml) were again collectedand assayed as before. The first and largest peak con-taining N-acetylamino sugar was pooled and freeze-dried. The identity of this nucleotide peptide wasshown to be UDP-N-acetylmuramyl-L-alanyl-D-isoglutaminyl-L-lysyl-D-alanyl-D-alanine (UDP-Mur-NAc-pentapeptide) by automatic amino acid analysis,N-acetyl-amino sugar analysis (11), phosphate analy-sis (5), and by ultraviolet absorption ratios (9).

Radioactivity measurements. All radioactivity meas-urements were made using a Packard Tri-Carb liquidscintillation spectrometer. Samples (0.1 ml) weremixed with 10 ml of scintillation fluid composed of1,4-bis-[2-(4-methyl-5-phenyl-oxazyl)] benzene, 0.2 g;diphenyloxazole, 4.0 g; absolute ethanol, 400 ml;and toluene, 600 ml.

Differential absorption measurements. The complexbetween vancomycin or ristocetin and UDP-Mur-NAc-pentapeptide was detected by the procedure ofPerkins (10). One cuvette containing 2.5 ml of theantibiotic and another containing 2.5 ml of distilledwater was placed into the reference beam. Identicalcuvettes were placed in the sample beam. UDP-MurNAc-pentapeptide (0.5 ml) was added to thecuvette containing water in the reference beam and tothe cuvette containing antibiotic in the sample beam.The final concentration of antibiotic and nucleotidepeptide was 0.16 mm in the respective cuvettes. Water(0.5 ml) was added to the cuvette in the referencebeam which contained only antibiotic to correct fordilution. Differential spectra were obtained by scan-ning absorbance in the range of 320 to 250 nm byusing a Cary model 14 recording spectrophotometer.

Calculations of antibiotic concentrations in these

477

BEST, GRASTIE, AND MCCONNELL

experiments assumed a molecular weight of 4,000 forristocetin and 1,600 for vancomycin, since these were

the values used by Perkins (10). However, the truemolecular weight of vancomycin is probably 3,200(6).Growth studies. The effect of nucleotide pentapep-

tide on growth inhibition by vancomycin and risto-cetin was determined with B. subtilis W-23 growingin a glucose salts medium (2). Cells from a 12-hrglucose salts slant were grown to an A540 value of 0.5to 0.6 [0.4 to 0.5 mg (dry weight) per ml] at 37 C inthe liquid glucose salts medium. Five milliliters ofthese actively growing cells was then used to inocu-lated 125-ml flasks containing glucose salts medium(20 ml), antibiotic, nucleotide peptide as indicated,and distilled water to give a final liquid volume of25 ml. The antibiotic concentration was adjusted tothat level which gave complete inhibition of growthwithout lysis of the cells. An initial A540 value of0.15 [0.1 mg (dry weight) per ml] was obtained ineach flask, and growth was followed turbidimetricallyat 540 nm at 30-min intervals. All flasks were shakenin a water bath at 37 C during growth measurements.

RESULTS

Adsorption of vancomycin and ristocetin toMicrococcus Iysodeikticus cell walls. Figure 1 de-picts the adsorption of ristocetin B to isolatedcell walls of M. lysodeikticus. The amount ofantibiotic bound to the walls increases progres-sively with increasing cell wall mass until 90 to95% of the ristocetin is bound. At the optimumwall-ristocetin ratio, about 200 ,ug of antibiotic isbound per mg (dry weight) of cell wall.

Antibiotic complexes with UDP-MurNAc-pentapeptide. The reported complexes betweenaw

z 1000coa 90

Z80-70

60OP 60-

z 50-

I-40-z

w

O 3.0

w

20

1 2 3 4 5 6 7 8 9

CELL WALL MASS (mg dry wt.)FIG. 1. Adsorption of ristocetin B to cell walls from

M. lysodeikticus. The procedure is described in thetext.

either vancomycin or ristocetin and nucleotidepentapeptide have been confirmed by using dif-ferential absorption spectra (10). As shown inFig. 2, both antibiotics give a distinctive absorp-tion spectrum in the presence of UDP-MurNAc-pentapeptide. Vancomycin and ristocetin alonegive an absorption peak at about 280 nm, whereasthe vancomycin complex gives an adsorptionminimum at 282 nm and the ristocetin complexminimum occurs at 287.5 nm.

The maximum complex formation, as indi-cated by the differential spectra, was obtainedwith a 1 :1 ratio of vancomycin to nucleotide pep-tide as reported by Perkins (10). This result wasonly obtained, however, when it was assumedthat the molecular weight of vancomycin is 3,200.An increase in the proportion of either componentbeyond this ratio did not result in an additionalamount of complex formation. With ristocetin,

+.I 2-

+.Os -

, +.04-

z

0

a .00-0

4o

4 - .04-

-.08.

+ .02 -

w .00-z

4 - D2-00o -.04 -

m4 -.06-4

- .08-

- .10-

- .12-

- .14-

a

I Il270 280 290 300 310 320

b

I I I I I_

280 290 300 310 320

WAVELENGTH (n m)FIG. 2. Differential absorption spectra obtained

with vancomycin (a) and ristocetin B (b) in the presenceof UDP MurNAc pentapeptide. The procedure is de-scribed in the text.

478 J. BACrERIOL.

I

VOL. 102, 1970 RELATIVE AFFINITY OF VANCOMYCIN AND RISTOCETIN

a 2:1 proportion of antibiotic to UDP-MurNAc-pentapeptide resulted in a slight increase (about10%) in the apparent amount of complex formed.The difference spectra in Fig. 2 were obtained byusing 0.16 mm UDP-MurNAc-pentapeptide,0.32 mm vancomycin (assumed molecular weightof 1,600), and 0.16 mm ristocetin (assumed mo-lecular weight of 4,000).

Effect of UDP-N-acetylmuramyl pentapeptideon ristocetin B and vancomycin adsorption to cellwalls. Since both antibiotics and the nucleotidepeptide absorb ultraviolet light at 280 nm, theeffect of the latter on antibiotic adsorption to cellwalls could not be accurately measured by usingonly the standard spectrophotometric assay ofBest and Durham (3). An indication of the rela-tive affinity of the antibiotics for cell walls andfor the peptidoglycan precursor was obtained,however, by using this assay in combination with"C-labeled nucleotide pentapeptide. The resultspresented in Table 1 depict the effect of two levelsof nucleotide pentapeptide on the adsorption of

TABLE 1. Effect of uridine diphosphate (UDP)-MurNAc-pentapeptide on ristocetin B adsorption

to cell walls of Micrococcus lysodeikticus

Radio-activity

Concn Absorb- (countsSamnple Addition8 'Umo- ance per minno. le) (280 per ml ofle) nm) super-

natantfluid)

1 Ristocetin B 0.62 0.362 14C-UDP-Mur- 0.62 0.89 9,994

NAc-pentapep-tide

3 14C-UDP-Mur- 1.24 1.65 19,810NAc-pentapep-tide

4 Walls + risto- 0.62 0.06cetin B

5 Walls + risto- 0.62 1.07 9,860cetin B + 14C.UDP-MurNAc-pentapeptide 0.62

6 Walls + risto- 0.62 1.76 19,010cetin B + 14C- 1.24UDP-MurNAc-pentapeptide

a Ristocetin B and 14C-UDP-MurNAc-penta-peptide, at the indicated concentrations, werebrought to a final liquid volume of 3.0 ml withdistilled water. In those tubes containing cellwalls (2 mg), the contents were centrifuged at10,000 X g for 10 min before absorbance andradioactivity measurements. The antibiotic andnucleotide peptide were mixed together sepa-rately before their addition to tubes containingcell walls in samples 5 and 6.

vancomycin and ristocetin to M. lysodeikticuscell walls. A comparison of the A2N readings indi-cated that adsorption of ristocetin in samples 5and 6 was comparable to that observed in theabsence of added nucleotide peptide. These datademonstrate that the ristocetin-nucleotide pep-tide complex does not bind to walls as a unit,since the radioactivity in samples 5 and 6 was notsignificantly decreased relative to that in samples2 and 3. Therefore, ristocetin has a greater affinityfor intact cell walls than for the peptidoglycanprecursor.

Ristocetin and the labeled nucleotide peptidewere mixed before incubation with the cell wallsin the experiment described in Table 1. To directlydetermine the fate of the antibiotic-nucleotidepeptide complex in the presence of cell walls, thetwo components were mixed in the cuvette asdescribed previously, and complex formationwas confirmed by the resulting differential absorp-tion spectrum. This complex was then quantita-tively mixed with sedimented cell walls. After 10min at 25 C, the suspension was centrifuged toremove the cell walls and any bound material.The resulting supernatant solution was recoveredquantitatively, and a 0.1-ml amount was removedfor radioactivity measurement. The remainder ofthe supernatant solution was returned to the cu-vette to assess the effect of incubation with cellwalls on the previous differential absorptionspectrum. The excessive absorbance by thereference beam relative to the sample beam as-sured that adsorption of antibiotic had occurred,and the radioactivity measurements (Table 2)

TABLE 2. Stability of the 14C-UDP-MurNAc-pentapeptide-ristocetin B complex in thepresence of cell walls of Micrococcus

lysodeikticus

RadioactivityOSample Sample contents (counts per min

supernatantfluid)

1 14C-UDP-MurNAc- 3,440pentapeptide (0.16mmolar)

2 Ristocetin B + 14C_ 3,600UDP-MurNAc-pen-tapeptide (0.16mmolar each)

a Ristocetin B and 14C-nucleotide pentapeptidewere mixed, and the complex formation wasconfirmed by the resulting differential absorptionspectrum as shown in Fig. 2. The complex wasmixed with M. lysodeikticus cell walls (2 mg).After removing the walls by centrifugation, thesupematant solution was quantitatively recovered,and the radioactivity was measured.

479

BEST, GRASTIE, AND MCCONNELL

revealed that all the radioactivity remained inthe supernatant solution after adsorption. Identi-cal results were obtained by using either vancomy-cin or ristocetin B. Thus, both antibiotics have agreater affinity for cell walls than for UDP-MurNAc-pentapeptide, and there is no indica-tion that the complex of the latter and antibioticbinds to cell walls.

Effect of complexed antibiotic on growth ofBacillus subtilis. If the mechanism of action ofvancomycin and ristocetin depends on the forma-tion of a stable complex with the lipid-bounddisaccharide pentapeptide as proposed byPerkins (10), prior incubation of the antibiotics,with sufficient peptidoglycan precursor to assuremaximum complex formation, would be expectedto at least reduce the biological activity of theinhibitors. However, if the inhibition of peptido-glycan synthesis depends on the adsorption of anantibiotic molecule at a cell wall site proximal tothe site of addition of the disaccharide pentapep-tide precursor, the previously demonstrated affin-ity of these antibiotics for cell walls relative toUDP-MurNAc-pentapeptide should result in es-sentially complete growth inhibition, in spite ofan excess of exogenous nucleotide pentapeptide.The results presented in Fig. 3 show the effect

of UDP-MurNAc-pentapeptide on growth in-hibition by vancomycin or ristocetin B. The levelsof both antibiotics were carefully selected to givea complete inhibition of growth. Even slightlyhigher concentrations of either antibiotic resultedin lysis and a 10% reduction in the concentrationof either vancomycin or ristocetin allowed someincrease in turbidity. As indicated, nucleotidepentapeptide, even at an eightfold excess overthat required for maximum complex formation,did not affect the inhibition of growth of B.subtilis by ristocetin B or vancomycin. This resultwas not altered by premixing the antibiotics withall levels of nucleotide pentapeptide before theiraddition to the growth medium.

DISCUSSION

Two mechanisms of action for vancomycinand ristocetin inhibition of peptidoglycan synthe-sis have been proposed. The first, by Best andDurham (3), was based on the observed adsorp-tion of these antibiotics to whole cells and iso-lated cell walls. According to their proposal,those molecules of antibiotic which are bound atsites of active wall synthesis should block theaddition of the lipid-bound disaccharide penta-peptide to the growing peptidoglycan polymer.Support for the significance of adsorbed anti-biotic was provided by Sinha and Neuhaus (14)who observed that isolated cell walls of M.

E 0.40-0In 0.30-w

maz 0.20-4

e0.6-

Z 0.4-0

In

ma14

a 0.6-

4 .40.

0

b

a A -&.

I I I I15 30 45 60

TIME (MINUTES)FIG. 3. Effect of UDP-MurNAc-pentapeptide on

the inhibition ofgrowth ofB. subtilis by vancomycin (a)and ristocetin B (b). Actively growing cells were inocu-lated into flasks containing glucose salts (2) and thefollowing concentrations of antibiotic and UDP-MurNAc-pentapeptide: 0, no antibiotic, with or withoutUDP-MurNAc-pentapeptide; 0, vancomycin (0.4lAg/ml; 4 ,ug per mg (dry weight) ofcells) or ristocetin B(0.3 ug/ml; 3 ,ug per mg (dry weight) ofcells); A, anti-biotic concentrations listed above and UDP-MurNAc-pentapeptide. The same growth inhibition was observedwhen the antibiotic to nucleotide pentapeptide molarproportions were 1:1, 1:4, or 1:8.

lysodeikticus reversed the inhibition of cell-freepeptidoglycan synthesis by vancomycin andristocetin. These results established that theantibiotics have a greater affinity for walls thanfor the other components in the cell-free prep-arations.The latest mechanism of action which has been

proposed holds that the complex which is formedbetween these inhibitors and the pentapeptideprecursors having a D-alanyl-D-alanine C-termi-nus could block synthesis of peptidoglycan bypreventing the enzymatic transfer of the disac-charide peptide from the lipid carrier to the grow-ing peptidoglycan chain (10).Both of the proposed mechanisms of action

would affect the same step in the biosyntheticsequence, and both would affect the same reac-

480 J. BACrERIOL.

VOL. 102,1970 RELATIVE AFFINITY OF VANCOMYCIN AND RISTOCETIN

tion defined by Anderson et al. as being mostsensitive to these antibiotics, i.e., the transfer oflipid-bound disaccharide pentapeptide to an ac-ceptor which would be the existing peptidoglycanpolymer in vivo (1). The evidence presented inthis paper would appear to favor a mechanism ofinhibition based on adsorption. First, bothvancomycin and ristocetin have a greater affinityfor walls than for UDP-MurNAc-pentapeptide.In addition, complexed antibiotic is no less in-hibitory than free antibiotic. Finally, the excep-tional affinity of vancomycin and ristocetin forisolated cell walls makes it difficult to rationalizethe penetration of significant amounts of anti-biotic through existing walls to the membrane-bound peptidoglycan precursor. As pointed outpreviously, one mg (dry weight) of B. subtiliscell walls can adsorb 750 ,g of vancomycin (3).Thus, at reasonable levels of inhibitor, adsorp-tion to the external cell wall would certainly pre-cede and likely preclude appreciable penetrationof either antibiotic to the membrane.

In choosing between the two mechanisms ofaction, there is an interesting paradox to con-sider with regard to adsorption and complexformation chemistry. Both processes appear toinvolve an interaction between the antibioticsand carboxyl groups. The involvement of thesegroups was initially demonstrated by esterifyingcell walls with diazomethane, a treatment whichcompletely prevented the adsorption of vancomy-cin (3). This observation and the fact that diva-lent cations reduce adsorption and alleviategrowth inhibition by vancomycin indicated thationic linkages were involved in the adsorptionphenomenon.As demonstrated by Perkins (10), complex

formation specifically involves the D-alanyl-D-alanine carboxyl terminus on the peptide. Al-though the chemical nature of this complex is notknown, esterification of the terminal carboxylgroup also eliminates complex formation (10).Magnesium, at 0.01 M, has no effect on the dif-ferential spectrum obtained with vancomycinand nucleotide peptide (Best and McConnell,unpublished data). These results suggest that al-though both adsorption to walls and complexformation involves carboxyl groups, differentassociations could be involved in the two proc-esses. In addition, the more attractive mecha-nism of action (adsorption to cell walls) impliesa lack of specificity which is disturbingly apparentin the alternative proposal. A large number of in-nocuous binding events to several possible groupson the cell wall surface must be acknowledged.Only those adsorption events in the vicinity ofactive cell wall synthesis can be considered ef-fective. This situation is contrasted with the

specificity apparent in the complex formed be-tween these antibiotics and peptides containing aD-alanyl-D-alanine C-terminus. As pointed outby Perkins (10), it is difficult to reconcile thespecificity of vancomycin and ristocetin towardthese peptides without invoking significance tothe resulting complex.

It would be tempting to compromise the twotheories and suggest that an interaction betweenantibiotic and either cell wall or wall precursorcould be involved in the eventual inhibition ofpeptidoglycan synthesis. However, the alleviationof growth inhibition by divalent cations can notbe accommodated to the complex formationproposal as yet. Since such cations both alleviategrowth inhibition and reduce adsorption to iso-lated cell walls (2-4, 14) any mechanism of actionfor vancomycin and ristocetin must account forthis fact. If adsorption to cell wall is consideredfortuitous and inconsequential, and if divalentcations have no effect on complex formation, thenthese cations should enhance rather than decreasethe activity of these antibiotics. This would beexpected since the cations would decrease adsorp-tion of the antibiotics to the wall and more mole-cules would be available to participate in com-plex formation.Although it would be difficult to prove that the

adsorption of vancomycin and ristocetin to cellwalls is not responsible for the inhibition ofpeptidoglycan synthesis, additional informationon the significance of complex formation withprecursor peptides might be obtained with cell-free peptidoglycan-synthesizing systems. If com-plex formation is responsible for inhibition, agiven level of this inhibition might be reversed byincreasing the amount of UDP-MurNAc-penta-peptide relative to either antibiotic in the incuba-tion mixture. However, if complexed antibioticis as inhibitory in the cell-free system as in thegrowth studies reported here, then support forthe adsorption concept would be indicated.

ACKNOWLEDGMENTS

This investigation was supported by a grant from the Brown-Hazen fund of the Research Corporation to the senior authorand by training grant A100248 from the National Institute ofAllergy and Infectious Diseases.We are indebted to John Howard for his cooperation with the

spectrophotometric analyses and to Robert Crounse for per-forming the automatic amino acid analyses.

LITERATURE CITED

1. Anderson, J. S., M. Matsuhashi, M. A. Haskin, and J. L.Strominger. 1965. Lipid-phosphoacetylmuramyl-penta-peptide and lipid-phosphodisaccharide-pentapeptide: pre-sumed membrane transport intermediates in cell wallsynthesis. Proc. Nat. Acad. Sci. U.S.A. 53:881-889.

2. Best, G. K., and N. N. Durham. 1964. Effect of vancomycinon Bacillus subtilis. Arch. Biochem. Biophys. 105:120-125.

3. Best, G. K., and N. N. Durham. 1965. Vancomycin adsorp-

481

BEST, GRASTIE, AND McCONNELL

tion to Bacillus subtilis cell walls. Arch. Biochem. Biophys.111:685-691.

4. Best, G. K., and N. N. Durham. 1966. Adsorption of theristocetins to Bacillus subtilis cell walls. Antimicrob. AgentsChemother.-1965, p. 334-338.

5. Chen, P. S., T. V. Toribara, and H. Warner. 1956. Micro-determination of phosphate. Anal. Chem. 28:1756-1760.

6. Higgins, H. M., W. H. Harrison, G. M. Wild, H. R. Bungay,and M. H. McCormick. Vancomycin: a new antibiotic.VI. Purification and properties of vancomycin. AntibioticsAnnual-1956/1957, p. 906-914.

7. Jordan, D. C. 1961. Effect of vancomycin on the synthesis ofthe cell wall mucopeptide of Staphylococcus aureus. Bio-chem. Biophys. Res. Commun. 6:167-170.

8. Jordan, D. C., and H. D. C. Mallory. 1965. Site of action ofvancomycin on Staphylococcus aureus. Antimicrob. AgentsChemother.-1964, p. 489-494.

9. Mandelstam, P. R. Loercher, and J. L. Strominger. 1962. Auridine diphosphoacetylmuramyl hexapeptide from penicil-

lin-treated Streptococcus faecalis. J. Biol. Chem. 237:2683-2688.

10. Perkins, H. R. 1969. Specificity of combination between muco-peptide precursors and vancomycin or ristocetin. Biochem.J. 111:195-206.

11. Reissig, J. L., J. L. Strominger, and L. F. Leloir. 1955. A modi-fied colorimetric method for the estimation of N-acetyl-amino sugars. J. Biol. Chem. 217:959-966.

12. Reynolds, P. E. 1961. Studies on the mode of action of vanco-mycin. Biochim. Biophys. Acta 52:403-405.

13. Reynolds, P. E. 1966. Antibiotics affecting cell wall synthesis.Symp. Soc. Gen. Micro. 16:47-57.

14. Sinha, R. K., and F. C. Neuhaus. 1968. Reversal of vancomy-cin inhibition of peptidoglycan synthesis by cell walls. J.Bacteriol. 96:347-382.

15. Wallas, C. H., and J. L. Strominger. 1963. Ristocetins, inhibi-tors of cell wall synthesis in Staphylococcus aureus. J. Biol.Chem. 238:2264-2266.

482 J. BACTERIOL.