O Acetyl Migration as the Mechanism for 0 Acetylation of ...

Vol. 173, No. 14

JOURNAL OF BACTERIOLOGY, JUlY 1991, p. 4318-4324

0021-9193/91/144318-07$02.00/0 Copyright C 1991, American Society for Microbiology

Evidence for N- >O Acetyl Migration as the Mechanism for 0 Acetylation of Peptidoglycan in Proteus mirabilis CLAUDE DUPONTt AND ANTHONY J. CLARKE*

Guelph-Waterloo Centre for Graduate Work in Chemistry, Department of Microbiology, University of Guelph, Guelph, Ontario NIG 2WI, Canada Received 22 March 1991/Accepted 14 May 1991

0-acetylated peptidoglycan was purified from Proteus mirabilis grown in the presence of specifically radiolabelled glucosamine derivatives, and the migration of the radiolabel was monitored. Mild-base hydrolysis of the isolated peptidoglycan (to release ester-linked acetate) from cells grown in the presence of 40 ,uM [acetyl-3HJN-acetyl-D-glucosamine resulted in the release of [3H]acetate, as detected by high-pressure liquid chromatography. The inclusion of either acetate, pyruvate, or acetyl phosphate, each at 1 mM final concentration, did not result in a diminution of mild-base-released [31H]acetate levels. No such release of [3H] acetate was observed with peptidoglycan isolated from either Escherichia coli incubated with the same radiolabel or P. mirabilis grown with [1,6-31H1N-acetyl-D-glucosamine or D-[1-_4C]glucosamine. These observations support a hypothesis that 0 acetylation occurs by N--O acetyl transfer within the sacculus. A decrease in [3H]acetate release by mild-base hydrolysis was observed with the peptidoglycan of P. mirabilis cultures incubated in the presence of antagonists of peptidoglycan biosynthesis, penicillin G and D-cycloserine. The absence of free-amino sugars in the peptidoglycan of P. mirabUis but the detection of glucosamine in spent culture broths implies that NO-> transacetylation is intimately associated with peptidoglycan turnover.

0-acetylated PG was first observed independently by two different groups 30 years ago (1, 7), and the biological significance of this modification was discerned soon after (6). Whereas its role in resistance to muramidases and the consequences of such have been well documented, very little is known concerning the biosynthetic process involved with the 0 acetylation of PG. There is evidence to suggest that 0 acetylation occurs after nascent PG strands have been both attached and cross-linked to the preexisting sacculus (10, 18, 26-28, 39). By pulse-chasing radiolabelled N-acetylD-glucosamine (GlcNAc) into the PG of P. mirabilis, Gmeiner and Kroll (18) revealed that only non-O-acetylated PG subunits are incorporated into the growing polymer. It has also been observed that the 0 acetylation of PG continues in a cell-free biosynthetic system of P. mirabilis PG (30). Lear and Perkins (26-28), using a similar experimental strategy, obtained analogous results with N. gonorrhoeae. Indeed, searches for lipid (bactoprenyl)-linked N-acetylglucosaminyl-N,O-diacetylmuramyl pentapeptide precursors in either the cytoplasm or the cytoplasmic membrane proved futile (27, 39). These observations suggest that an acetyl transferase is present external to the cytoplasm. Comparison of the penicillin-binding protein (PBP) profiles of both penicillin-resistant and penicillin-sensitive strains of N. gonorrhoeae with the extent of PG 0 acetylation led Dougherty to implicate PBP 2 as having a role in 0 acetylation (9, 11). However, it is likely that the reduced levels of PG 0 acetylation in penicillin G-treated cells are consequential of incomplete cross-linking caused by the inhibition of transglycosidase or transpeptidase, possibly PBP 2. Nothing else is known about the putative acetyltransferase(s), and the source of acetate is unknown. Through in vivo labelling experiments with P. mirabilis, we provide evidence to suggest that acetate is transferred from the N-2 position of either N-acetylglucosaminyl or N-acetylmuramyl residues to the C-6 hydroxyl group of the latter.

Following the lysis of invading microorganisms, mammalian tissues clear cellular debris, including peptidoglycan (PG), very rapidly through the action of the hydrolytic enzymes of either phagocytic cells or serum. However, in some instances, large-molecular-weight fragments of PG have been observed to persist and circulate in the host organism. These large PG fragments have been shown to induce diverse pathobiological and pathophysiological effects, many of which have been recently reviewed (36). Some of these include the induction of slow-wave sleep, complement activation, pyrogenicity, modulation of blastogenesis, and arthritogenicity. In vivo studies with the pathogens Staphylococcus aureus and Neisseria gonorrhoeae indicated that the phenomenon of PG persistence is directly attributable to the presence of numerous O-acetyl substituents on the glycan backbone (5, 15, 24, 34, 40). 0-acetylated PG has been observed in a number of bacteria, including many pathogenic species, both gram positive (e.g., S. aureus [17, 41]) and gram negative (e.g., N. gonorrhoeae [40] and Proteus mirabilis [14]). 0 acetylation of PG occurs at the C-6 hydroxyl group of N-acetylmuramyl residues, producing the corresponding 2,6-diacetylmuramyl derivative. This modification to PG inhibits the hydrolytic activity of lysozymes (reference 13 and references therein), presumably through steric hindrance since the C-6 hydroxyl moieties of the substrate directly participate in its binding to the active-site cleft of the enzyme (3, 23). The PG of some species of eubacteria has been reported to be 0 acetylated up to 70%, thereby conferring both intrinsic and complete resistance to lysozyme hydrolysis (30).

Corresponding author. t Present address: Department of Chemistry, Carlsberg Laboratorium, Valby, Copenhagen DK 2500, Denmark. *

4318

PEPTIDOGLYCAN 0 ACETYLATION IN P. MIRABILIS

VOL. 173, 1991

MATERIALS AND METHODS Bacterial strains. P. mirabilis 19 was kindly provided by J. Gmeiner, Technische Hochschule Darmstadt, Germany; a stock laboratory strain was obtained from H. Perkins, Liverpool, United Kingdom. These strains have been extensively characterized (4, 29, 31). Other P. mirabilis strains (ATCC 12453 and ATCC 7002) and P. vulgaris ATCC 13315 were purchased from the American Type Culture Collection, Rockville, Md. Escherichia coli K-12(CSH4) was provided by J. Wood, University of Guelph. All bacteria were maintained on nutrient agar slants at 4°C. Growth conditions. For radiolabelling experiments, P. mirabilis strains were cultivated in aerated nutrient broth supplemented with 5.0 g of NaCl, 4.5 g of Na2HPO4, and 3.6 g of glucose per liter on a rotary shaker at 200 rpm and 37°C. In some cases, the culture broths were made to 20 to 40 puM GlcNAc. The mineral salts medium described by Grabow and Smit (21), supplemented with 3.6 g of glucose, vitamin B12, and nicotinic acid per liter, was used for both PG turnover and some radiolabelling experiments. Radioactive labelling. Aliquots (5 ml) of cultures (50 ml) grown overnight in either nutrient broth or mineral salts medium were used to inoculate 95 ml of the same medium in a 250-ml Erlenmeyer flask and incubated at 37°C in a New Brunswick G24 environmental incubator with shaking at 200 rpm. Growth was monitored at frequent intervals with a Spectronic model 70 spectrophotometer at 578 nm. At absorbance values of 0.25 to 0.30 (ca. 275 mg- liter-', dry weight), 10 to 20 ,uCi (0.37 MBq) of either [acetyl3H]GlcNAc (10 Ci mmol-1; 370 GBq- mmol-1) or [1,63H]GlcNAc (34.2 Ci * mmol-1; 1.27 TBq * mmol-1) was added, and the cultures were further incubated at 37°C. In some experiments, the culture medium was supplemented with either 1 mM (final concentration) pyruvate, acetate, or acetyl phosphate. Preparation of insoluble peptidoglycan. When the cultures attained an A578 of 0.8, the bacteria were cooled to 2°C in an ice-alcohol bath. Cells were harvested by centrifugation (Sorvall RC-2B; DuPont Sorvall) at 9,000 x g for 15 min at 4°C. The cells were washed twice with 10 mM sodium phosphate buffer (pH 6.5) and then resuspended in 10 ml of Milli-Q water on ice. Insoluble PG was extracted from whole cells by the procedure of Hoyle and Beveridge (22). The cell suspensions were added dropwise to an equal volume of 8% sodium dodecyl sulfate (SDS) buffered with 10 mM sodium phosphate (pH 6.5) at 100°C. After all of the cells were added, the SDS extraction at 100°C was continued for 3 h under reflux. After cooling to ambient temperature and stirring overnight, the SDS-insoluble material was collected by centrifugation [Beckman L5-50; Beckman Instruments (Canada)] at 160,000 x g for 60 min at 20°C. The pellet was washed thrice by centrifugation with 10 mM sodium phosphate buffer (pH 6.5) and resuspended in 10 mM Tris HCI buffer (pH 6.8). The crude peptidoglycan preparation was treated with pronase (preheated for 2 h at 60°C) at 0.20 mg- ml-' (final concentration) for 1 h at 60°C. The PG was reextracted with 4% SDS at 100°C for 1 h. The insoluble material was collected and washed by centrifugation thrice as described above. The purified PG was resuspended in Milli-Q water and lyophilized. 2-Keto-3-deoxyoctonic acid measurements and amino acid analysis of this purified material indicated it to be free of both lipopolysaccharides (LPS) and lipoproteins, respectively. Release and quantitation of acetate. Radiolabelled PG in water containing 0.1% sodium azide was evenly suspended -

4319

by brief (three bursts of 20 s each) low-power sonication (Bransonic 12 sonicator bath; 50/60 Hz; Braun) to a concentration of 2 to 5 mg- ml-'. Aliquots (200 p.l) of these suspensions were incubated overnight at ambient temperature with an equal volume of either 40 mM NaOH or 40 mM sodium phosphate buffer (pH 6.5). The PG was collected by centrifugation at 160,000 x g for 15 min at 20°C in a Beckman Airfuge, and the pellet was washed once by centrifugation with 200 pl of 0.1% sodium azide. The supernatants were pooled and filtered through a Millipore HA 0.45-p.m-pore-size membrane (Millipore Ltd., Mississauga, Ontario, Canada). Released acetate was detected and quantitated by high-pressure liquid chromatography (HPLC), using an Aminex HPX-87H (Bio-Rad) organic acid column (7.8 by 300 mm) as previously described (13). Fractions (0.5 ml) of the column effluent were collected and, together with a 100-p.l aliquot of the original supernatant, were separately counted for levels of radioactivity by liquid scintillation counting. The PG recovered from the pellets of the Airfuge centrifugation was solubilized by treatment with 5 to 10 U of mutanolysin in 10 mM Tris HCI (pH 6.80) containing 8 mM MgCl2 at 37°C with gentle agitation for 2 h prior to the quantitation of radioactivity remaining bound. Radiolabelling in the presence of antibiotics. Cultures of P. mirabilis 19 were grown in minimum medium (100 ml), supplemented with 30 mM glucose and 40 p.M GlcNAc, for 2 h at 37-C (A578 of 0.15 to 0.2), and then 20 ,uCi of [acetyl-3H]GlcNAc was added. After a 2-h incubation in the presence of radiolabel, sterilized penicillin G, chloramphenicol, and D-cycloserine were introduced into the separate cultures at final concentrations of 10, 5.0, and 2.5 mg * ml-', respectively. After a 2-h incubation, now in the presence of both antibiotic and radiolabel, the cultures were rapidly chilled in an ice bath and PG was isolated as described above. Aliquots (2 ml) of cultures were harvested both immediately before antibiotic introduction and at the end of the incubation period and mixed with an equal volume of ice-cold 10% (wt/vol) trichloroacetic acid (TCA) solution. The samples were well mixed and kept at 0°C for at least 1 h. The TCA precipitates were filtered over a glass fiber membranes (GF/F; Whatman Ltd., Maidstone, United Kingdom) and washed with 10 ml of ice-cold 5% (wt/vol) TCA. The membranes were dried (60°C, 2 h), transferred to glass scintillation vials, and counted for radioactivity following the addition of 5 ml of scintillation cocktail. Detection of free amines in peptidoglycan and turnover products. P. mirabilis 19 was cultured at 37°C with aeration in the minimal salts medium of Grabow and Smit (21) containing glucose as the carbon and energy source. Cultures (100 ml), incubated at 37°C until growth achieved an A578 of 0.80, were cooled to 2°C by swirling the flasks in an ice-ethanol bath, and the cells were separated from the spent broth by centrifugation at 9,000 x g and 4°C. The culture supernatant was filtered through 0.45-p.m-pore-size cellulose-acetate membranes (Millipore) to remove any residual cells and then lyophilized while the PG of the collected cells was extracted and prepared as described above. Samples of both the purified PG and lyophilized culture supernatant were suspended in Milli-Q water, and the pH was adjusted to 10.4 with the addition of NaOH. These samples were then treated with 10 mM (final concentration) 5-dimethylaminonaphthalene-l-sulfonyl * Cl (dansyl- Cl)-acetone solution at 70°C for 1 h. The reaction solutions were lyophilized, and following acid hydrolysis in either 4 or 6 M HCI (100°C, 18 h), the dansylated samples were subjected to amino acid analysis by reverse-phase HPLC as described by Negro et

4320

J. BACTERIOL.

DUPONT AND CLARKE

al. (33). Prior to dansylation, some samples of culture supernatant were suspended in 50 mM Tris * HCl buffer (pH 8.0) containing 0.2 ,uM pyridoxal phosphate and 50 mM a-ketoglutarate and treated with an aliquot (200 ,ul) of glutamate-pyruvate transaminase (0.5 mg * ml-'). After 2 h of incubation at 37°C with shaking, the pH of the solution was raised to 10.2 to 10.6 with sodium hydroxide and dansylation was performed as described above. Standards consisting of dansylated and didansylated diaminopimelic acid [DNS-DAP and (DNS)2-DAP], dansylated muramic acid (DNS-Mur), and dansylated glucosamine (DNS-GlcN) were prepared as described above and purified by HPLC, using a Beckman Ultrasphere ODS column (5 ,m; 4.6 by 150 mm) (33). Analytical methods. PG concentrations were determined by amino acid analysis, using a Beckman Gold amino acid analyzer with postcolumn ninhydrin detection [Beckman Instruments (Canada)]. Samples of PG (0.15 mg) were hydrolyzed in vacuo with 4 M HCl at 110°C for 16 h. Acetate quantitation was also performed by HPLC on a Beckman system consisting of two model 110B pumps, a model 167 dual-channel rapid-scanning UV-visible detector, a model 406 analog interface, an IBM-XT computer controller with Beckman Gold chromatography software, and a Bio-Rad HPLC column heater as previously described (13). Radioactivity was measured with a Packard Tri-Carb 2000 liquid scintillation counter (Canberra-Packard Canada, Mississauga, Ontario, Canada) with Ecolume (ICN Biomedicals) or Liquiscint (National Diagnostics, Manville, N.J.) as the scintillation cocktail. 2-Keto-3-deoxyoctonic acid was determined by the periodic acid-thiobarbituric acid method as described by Karkhanis et al. (25). Enzymes and biochemicals. GlcNAc, acetyl phosphate, D-cycloserine, dansyl. Cl, dansylated alanine (DNS-Ala), DNS-Glu, D-glucosamine HCl, mutanolysin, pyruvic acid, and SDS were purchased from Sigma Chemical Co., St. Louis, Mo. Boehringer Mannheim Canada, Laval, Quebec, Canada, supplied the pronase. Chloramphenicol was obtained from BDH Chemicals, Toronto, Ontario, Canada. Penicillin G was a product of Ayerst Laboratories, and nicotinic acid was obtained from Nutritional Biochemicals Corp. Ecolume, vitamin B12, [1-14C]GlcN (specific activity, 45 mCi mmoF-1; lot 3157143) and [acety1-3H]GlcNAc (specific activity, 10 Ci mmol-1; lots 3305141, 3609124, and 4021162) were purchased from ICN Biomedicals Canada, Ltd., Montreal, Quebec, Canada. [1,6-3H]GlcNAc (specific activity, 34.2 Ci- mmol-1; lot 2482-084) was provided by New England Nuclear Research Products, DuPont Canada, Dorval, Quebec, Canada. -

RESULTS Radiolabelling of peptidoglycan. Initial experiments investigating the mechanism of 0 acetylation pertained to the culture of both P. mirabilis Perkins and E. coli K-12(CSH4) in nutrient broth supplemented with 20 ,uM tritiated GlcNAc, with the radiolabel confined to specific sites in the carbohydrate. With [acetyl-3H]GlcNAc, 14% 2.2% of the total added radioactivity was incorporated into the cells, and 27% + 4.1% of this label (representing 3.7% of total added label) was confined to insoluble PG. Identical levels of radiolabel incorporation into P. mirabilis were documented by Martin and Gmeiner (30). LPS-free PG isolated and purified from such radiolabelled cells was subsequently incubated overnight with either 40 mM NaOH or 40 mM sodium phosphate buffer (pH 6.5). Following the removal of

a

I

I 0.

b

I

E0. I , .

0

.

.ia

.

10 20 Time (min)

30

FIG. 1. Release of [3H]acetate from the isolated peptidoglycan of P. mirabilis grown in the presence of [acety1-3H]GlcNAc as detected by HPLC. Purified peptidoglycan (225 ,ug) was incubated in 400 ,ll of 40 mM NaOH (a) or 40 mM sodium phosphate buffer (pH 6.5) (b) at ambient temperature for 18 h, and insoluble PG was removed by ultracentrifugation. Aliquots (200 ,ul) were injected onto an Aminex HPX-87H organic acid analysis column at 45°C with H2SO4 serving as the eluant at 0.6 ml * h-1. Column effluent was monitored at 210 nm, and fractions (0.5 ml) were collected and counted for radioactivity. The retention time of acetic acid is 18.9 min. The solid bars represent 0.01 absorbance units and 200 cpm, respectively.

insoluble PG by ultracentrifugation, the supernatants were subjected to organic acid analysis by HPLC as previously described (13). Base treatment of PG isolated from P. mirabilis grown in the presence of [acety1-3H]GlcNAc led to the release of tritiated acetate (Fig. la). No such release of either radiolabel or acetate was observed from similar samples of PG treated with the phosphate buffer (Fig. lb). Approximately 23% of the incorporated radioactivity was released as acetate from the insoluble PG after mild-base hydrolysis (Table 1), while no significant differences in the amounts of released radioactive acetate were observed between labelled PGs isolated from cultures of P. mirabilis Perkins grown in either nutrient broth or minimal salts medium. Acetate was also detected from base-treated PG isolated from the same microorganism but incubated in medium containing either [1-14C]GlcN or [1,6-3H]GlcNAc. However, the base-hydrolyzed acetate was not found to be radioactive in these cases; the radiolabel was confined to either C-1 or both C-1 and C-6, respectively, of the glucosamine (Table 1). Control experiments conducted with non-O-acetylated E. coli K-12(CSH4) indicated that this transfer was unique to P.

VOL. 173,

1991

PEPTIDOGLYCAN

TABLE 1. Release of radiolabelled acetate from P. mirabilis Perkins peptidoglycan Radiolabelled PG precursor

%o

[acety1-3H]GlcNAcd [acetyl-3H]GIcNAcf [1,6-3H]GlcNAc [1-14C]GIcN

3.54 (2.98)e 3.12 5.36 13.9 (1.20)

ACETYLATION IN P. MIRABILIS

4321

TABLE 2. Recovery of [3H]acetate from different sources of PG radiolabelled with [acety1-3H]GlcNAc

Radiolabelled acetate released after incubation witha:

Bufferb

0

Source of [3H]PG

0 acetylation

Base' n

%

n

8 1 1 2

22.9 (8.25) 22.7 (1.51) 5.89 (0.225) 14.2 (1.15)

8

2 2 3

Isolated and purified PG (225 bLg) was incubated with either buffer or base overnight at ambient temperature (23°C). b 40 mM sodium phosphate buffer (pH 6.50). c 40 mM (final concentration) sodium hydroxide. d Bacteria cultured in saline nutrient broth. e Values in parentheses represent standard deviations of n samples. f Bacteria cultured in the minimal salts medium of Grabow and Smit (21). a

mirabilis. Neither acetate nor radiolabel was detected from base-treated PG isolated from E. coli cells that had been incubated with any of the radiolabelled precursors (Table 2). Further studies with [acetyl-3H]GlcNAc that involved a number of bacterial species and strains were carried out. The results of incubation of the various bacterial strains in nutrient broth supplemented with 40 ,uM [acetyl-3H]GlcNAc are summarized in Table 2. In all cases, no acetate was detected when supernatants from buffer incubations were injected onto the HPLC column, and none of the radioactivity released was found to elute at the retention time of acetate. However, with the Proteus strains, significantly higher amounts of radioactivity were detected in the supernatants of mild-base incubations, and the levels varied with the particular strains. When these supernatants were subjected to HPLC analysis, acetate was detected in each case, and all radioactivity coeluted with authentic acetate. These data confirm that the release of radioactive acetate from specifically labelled and purified PG is correlated with the species of bacteria possessing 0-acetylated PG. Effects of antibiotics. The effects of various antibiotics on the extent of base-labile radioactive acetate released from P. mirabilis 19 cells incubated in the presence of [acetyl3H]GlcNAc are listed in Table 3. The addition of each of the antibiotics, at levels slightly lower than their respective MICs, to cultures preincubated with [acetyl-3H]GlcNAc for 2 h drastically reduced the further incorporation of the radiolabelled precursor into the PG sacculus. The two antibiotics having a primary action on PG biosynthesis, penicillin G and D-cycloserine, also inhibited the transfer of labelled acetate from GlcNAc to a base-labile position. Penicillin G has previously been shown to block PG 0 acetylation in S. aureus (37), N. gonorrhoeae (4), and P. mirabilis (19). In contrast, the protein synthesis inhibitor chloramphenicol, while affecting the incorporation of [acetyl-3H]GlcNAc into the PG sacculus, did not significantly alter the percentage of radioactive base-labile acetate in comparison with the control culture which lacked antibiotic addition. This finding is also consistent with observations made with S. aureus (24) and N. gonorrhoeae (35). Source of the acetyl moiety for peptidoglycan 0 acetylation. The observations described above suggest that radioactive acetate derived from [acetyl-3H]GlcNAc was specifically transferred in some manner to a base-labile position on PG. In an attempt to show that radioactive acetate is not released through the catabolism of [acetyl-3H]GlcNAc and subsequently transferred to the C-6 of muramyl residues after

mirabilis 19 mirabilis Perkins mirabilis 7002 mirabilis 12453 vulgaris 13315 E. coli K-12(CSH4)e P. P. P. P. P.

E. coli K-12(CHS4f

52.8 40.2 43.6 24.1 9.65 0.0

0.0

% of radiolabelled acetate released with: Bufferb Base'

0.820 (0.0992)d 3.54 (2.98) 6.25 2.52 0.730 (0.0542)

4.81 (2.99) 0.542

23.1 (1.07) 22.9 (8.25) 24.5 32.0 9.22 (0.119) 4.94 (3.27) 0.751

n

4 8 1 1 2 3 1

Data obtained from reference 13. 20 mM sodium phosphate buffer (pH 6.50), overnight at ambient temperature (23°C). C 20 mM (final concentration) sodium hydroxide, overnight at ambient temperature (23°C). d Values in parentheses represent standard deviations of n samples. e Bacteria cultured in saline nutrient broth. f Bacteria cultured in minimal salts medium of Grabow and Smith (21). a

b

prior activation, experiments using P. mirabilis 19 labeled with this radiolabel in the presence of other potential sources of acetate were carried out. Mild-base hydrolysis of PG isolated from P. mirabilis 19 grown in the presence of 40 ,uM [acetyl-3HJGlcNAc resulted in the release of 23% + 0.32% [3H]acetate. The inclusion of sodium acetate, sodium pyruvate, or acetyl phosphate, each at 1 mM final concentration, in the culture medium did not decrease the level of [acetyl3H]GlcNAc incorporation into the PG sacculus; moreover, no diminution of the level of radioactive acetate release (21% ± 2.9% for acetate; 25% + 1.1% for pyruvate; 22% ± 1.7% for acetyl phosphate) was observed following the mild-base hydrolysis of the respective PG isolates. Detection of free amines. If an N->O acetyl migration is occurring within the PG sacculus, then free amines of glucosaminyl or muramyl residues should be either present in PG itself or released as by-products of PG turnover. To test for the presence of free amines, the purified PG and concentrated spent broth of both P. mirabilis 19 and E. coli K-12(CSH4) cultures were dansylated by incubation with dansyl- Cl at 70°C for 1 h and then acid hydrolyzed in 4 M HCl at 100°C for 18 h. Representative results obtained by reverse-phase HPLC analysis of the dansylated and hydrolyzed adducts are shown in Fig. 2a. No DNS-Mur was evident in any of the samples tested, and DNS-GlcN was not detected in the PG preparations isolated from either P. mirabilis or E. coli. Evidence for the presence of DNS-GlcN in spent broth samples of P. mirabilis cultures was observed, but the DNS-GIcN eluted from the column as a shoulder to DNS-Ala due to the high concentration of the latter in the spent broth. To prevent this coelution of the two dansylated derivatives, most of the alanine present was converted to pyruvate by treatment of the spent broth samples with glutamate-pyruvate transaminase for 2 h at 37°C prior to dansylation. A clear separation of the two components was subsequently achieved by this transaminase pretreatment (Fig. 2b) and allowed for the quantitation of the DNS-GlcN contents. Treatment with the transaminase also seemed to substantially diminish the amount of DNS-DAP detected. It is conceivable that a peptidase(s) was present in the commercial transaminase preparation which hydrolyzed the soluble peptidoglycan peptides and the resulting (DNS)2-DAP was subsequently lost under the large ammonia peak (Fig. 2b). Regardless, the levels of glucosamine released by P. mirabilis 19 detected by this dansylation procedure (Table 4)

4322

J. BACTERIOL.

DUPONT AND CLARKE

TABLE 3. Effects of antibiotics on the transfer of [3H]acetate in the peptidoglycan of P. mirabilis 19 labelled with [acety1-3H]GlcNAc

[acetyl-3H]GlcNAc incorporated intob: . . . TCA-insoluble cell wall material Antibiotica Before addition After addition

None Chloramphenicol Penicillin G

D-Cycloserine

(cpm)

(cpm)

Increase (%)

1,426 1,361 1,430 1,372

2,638 1,548 1,630 1,847

85 13 14 35

Base-released [3H]acetate from SDS-insoluble PGc

SDS-insoluble PG (cpm mg of PG-1)

12,317 6,366 6,173 8,622

cpm. mg of PG-1

%d

5,224 2,324

42.4 36.5 10.1 13.6

623

1,172

a Antibiotic concentrations were 10.0, 5.0, and 2.5 mg- ml-' for penicillin G, chloramphenicol, and D-cycloserine, respectively. b Cells were cultured in the presence of 40 ,uM [acetyl-3H]GlcNAc for 2 h before the addition of antibiotics. TCA-insoluble cell wall material and SDS-insoluble PG were prepared as described in Materials and Methods. c 40 mM (final concentration) sodium hydroxide, overnight at ambient temperature (23). An average of 1.39% of the [3H]acetate was released from PG samples incubated in 40 mM sodium phosphate buffer (pH 6.50). d Calculated as the percentage of base-released counts relative to those of SDS-insoluble PG isolated from cells following 2 h of incubation in the presence of antibiotics.

are comparable to the concentrations of diaminopimelic acid observed in samples of the spent broth not treated with transaminase, indicating that glucosamine is indeed an authentic product of PG turnover in this microorganism. No DNS-GlcN was observed in samples of either fresh medium or E. coli spent broth that had been treated in a similar way, suggesting a correlation between its release from a bacterial species and 0 acetylation.

0.15 E

DISCUSSION We have presented evidence to suggest that 0 acetylation of PG arises via an N-+O acetyl migration during PG turnover. Several lines of evidence strongly suggest that the radioactive acetate released from hydroxyl positions is transferred directly from the N-2 position of N-acetylglucosaminyl or N-acetylmuramyl residues and not from an intermediate precursor molecule derived through the catabolism of the supplemented [acety1-3H]GlcNAc. (i) Metabolic studies with members of the family Enterobacteraceae have shown that in the presence of glucose, GlcNAc is transported into the cells by the phosphoenolpyruvate-dependent phosphotransferase system and directly incorporated into wall layers, as has been shown for LPS and PG (8). This has been shown to be the case for P. mirabilis (30) and was corroborated in this study by the finding that radioactive acetate is not released from base-treated PG prepared from cells incubated with either [1,6-3H]GlcNAc or [1-'4C]GlcN. Moreover, the presence of acetate, acetyl phosphate, or pyruvate (other potential sources of transferred acetate) did not compete with [acety1-3H]GlcNAc in the 0-acetylation process. (ii) The inhibition of [3H]acetate transfer by antagonists of PG biosynthesis suggests the prerequisite incorporation of labelled subunits into PG prior to transfer of labelled acetate. Under the experimental conditions used, cells were preincubated with [acety1-3H]GlcNAc for 2 h prior to the addition of the various antibiotics. This would have provided enough time for a hypothetical catabolism of GlcNAc in the cytoplasm and transfer of [3H]acetate to the sacculus, but such did not apparently occur. Instead, the subsequent administration of penicillin G, which blocks PG transglycosylation and transpeptidation reactions, among others, or D-cycloserine, an inhibitor of both alanine racemase and D-alanyl-D-alanine synthetase, were found to block the transfer of [3H]acetate. The fact that these inhibitors of PG maturation not only prevent the incorporation or further processing of [acety1-3H]GlcNAc into the preexisting PG sacculus but also limit [3H]acetate transfer implies that

a

0.20

Dap /0 2 3

10

L()

c\J 0.05

GIcN

4-a~~~~~~~~l (D 0.00

.

..

.....

co0.20b

-~0.15 0.10 0.05

0.00. 0 10 20 30 40 50

.....

60 70

Time (min) FIG. 2. Detection of free amines in the spent culture medium of

P. mirabilis. Lyophilized preparations of spent culture medium were dansylated and hydrolyzed as described in Materials and Methods. Aliquots (25 lil) of the hydrolyzed preparations were injected onto a Beckman ODS Ultrasphere reverse-phase HPLC column (5 ,um; 4.6 by 150 mm) previously equilibrated in 14% acetonitrile in 25 mM sodium phosphate-25 mM sodium acetate (pH 6.8). The column was eluted with a gradient of acetonitrile (14 to 22% over 5 min, 22 to 34% over 10 min, and finally 34 to 62% over 22 min), and the effluent was monitored at 254 nm. (a) Dansylated and hydrolyzed spent culture medium; (b) the same sample of spent culture medium but treated with glutamate-pyruvate transaminase for 2 h at 37°C prior to dansylation and hydrolysis. Retention times for DNS-Glu, DNSGlcN, DNS-Ala, DNS-DAP, DNS-NH2, and (DNS)2-DAP are 30.1, 36.7, 38.1, 51.2, 58.1, and 58.6 min, respectively.

VOL. 173, 1991

PEPTIDOGLYCAN 0 ACETYLATION IN P. MIRABILIS

4323

TABLE 4. Amino acid and amino sugar composition of P. mirabilis spent culture medium Sample treatment

Unhydrolyzed Hydrolyzed 4 M HCI 6 M HCl

DNS-Ala

13.9

14.33 (1.75) 13.27 (1.25)

Dansylated peptidoglycan componenta (nmol- mg of spent culture medium-') DNS-Glu DNS-GlcN DNS-DAP

ND

0.0960 (0.031) ND

2.93

2.59 (0.245) 3.75

(DNS)2-DAP 0.222

2.78 (0.0695) 2.32 (0.156)

0.231 (0.0178) 0.92 (0.0304)

a Lyophilized spent culture supernatant was dansylated prior to hydrolysis in HCl at 110°C for 16 h. The dansylated amines were separated and quantitated by reverse-phase HPLC (see Materials and Methods). ND, Not determined. Values in parentheses are standard deviations.

the processes are intimately linked. (iii) Preliminary studies concerning the development of an in vitro PG biosynthesis system have indicated that radioactive acetate may be recovered from base-treated PG that was synthesized by the incubation of cell membranes with UDP-[acetyl-1-'4C]Nacetyl-D-glucosamine (data not shown). In view of these observations, it seems unlikely that the added [acetyl3H]GlcNAc was catabolized in the cytoplasm to produce a unique source of radioactive acetate. Assuming an even distribution of [3H]acetyl from [acetyl3H]GlcNAc among the PG acetyls, each disaccharide subunit should contain one unit (N linked) as part of GlcNAc, a second unit (N linked) as part of MurNAc, and 0.53 unit (O linked) attached to MurNAc (in the case of P. mirabilis 19 with 52.8% 0 acetylation). Hence, the O-acetyl content as a proportion of the whole would theoretically be 0.53/2.53 or 20.9%, which is lower than the observed value of 23.1%. The observed superfluity of the radiolabelled O-acetyl contents for each strain (Table 2) is probably a consequence of the experimental design, since nonsynchronized cultures were used and the dynamic nature of the PG sacculus may preclude the possibility of acquiring unequivocal data. Previous studies conducted by others on P. mirabilis, N. gonorrhoeae, and S. aureus have suggested that the process of PG 0 acetylation occurs within the PG sacculus and external to the cytoplasm (10, 18, 26-28, 39). In view of the absence of readily available activated acetate (e.g., acetyl coenzyme A or acetyl phosphate) in the milieu external to the cytoplasm, it appears that Proteus spp. have developed a method for utilizing the conserved bond energies stored within the PG sacculus to achieve 0 acetylation in a manner analogous to transpeptidation. The fact that free amino sugars are not detected in the PG sacculus but are observed in the spent culture medium as turnover products indicates that the processes of PG turnover and 0 acetylation are associated. While there is a dearth of information concerning the details of PG turnover in P. mirabilis, extensive studies have been pursued with E. coli and members of the genus Bacillus (for a recent review, see reference 12). With E. coli, components of the PG polymer, including oligopeptides which comprise the peptide moiety of the muramyl side chain, are indeed recovered by the cell for reutilization after their hydrolysis from the sacculus (20). This feature of recyclization of PG components thus casts doubts on the reported rates of turnover, since most studies were based on the recovery of radiolabelled PG components from spent culture media following the incubation of prelabelled bacteria. For this reason, no attempt has been made here to correlate the rates of glucosamine appearance in spent cultures with PG 0 acetylation. However, the observed occurrence of recyclization further lends support to the hypothesis of N->O acetyl transfer during PG turnover. The phenomenon of N->O acyl transfer has been widely

investigated since 1924, when Bergman and Mickeley demonstrated the migration of acyl groups from N-acylamino alcohol to form an O-acyl analog under the influence of anhydrous acidic reagents (2). It was shown by White (42) and later by Fodor and Otvos (16) that the N-acetyl groups in derivatives of GlcNAc easily migrate under the influence of acid to the oxygen atom at C-3. Examination of the thermodynamics involved in this migration supports our postulate for the presence of a PG N,O-acetyltransferase. The calculated average free energies of formation for amide and ester bonds, 36 and 17 kJ- mol1, respectively (43), indicate that sufficient energy is theoretically available upon release of acetate from the N-acetyl moieties of PG for the N--O transfer. Furthermore, a decrease in overall activation free energy for trypsin-catalyzed reactions of approximately 21 kJ. mol' is calculated from kinetic data (32) for the change from amide (peptide) to ester bond hydrolysis. The overall activation free energy is calculated as AAG' = Rlln[(Vmax/ KM)estelJ(Vmax/KM)amide], where (Vmax/KM)ester and (Vmax/ KM)amide are the averages of the steady-state parameters (Vmax and KM) for the trypsin-catalyzed hydrolysis of selected esters and amide (peptide) bonds, respectively. The values of Vmax and KM used were those reported (32) for the substrates Ala-Ala-Lys-O-Me, Leu-Lys-0-Me, and PheLys-O-Me for ester hydrolysis and Z-Lys-Ala-Ala, Z-LysPhe-Ala, and Z-Lys-Leu-Ala for amide hydrolysis (where Z and Me are N-carbobenzoxy and methyl ester, respectively). Finally, the existence of an N-arylhydroxamic acid N,Oacetyltransferase (EC 2.3.1.56), which transfers the N-acetyl group of some aromatic acetylhydroxamates to the 0 position of some aromatic hydroxylamines, has recently been proven to exist as a unique and independent enzyme in a variety of mammalian tissues (38). Hence, it appears feasible that N-*O acetyl transfer may occur in PG, especially if it is intimately coupled to the release of glucosamine during the process of turnover. We are currently developing a cell-free system for the 0 acetylation of PG so as to isolate and characterize the putative enzymatic system involved. ACKNOWLEDGMENTS These studies were supported by operating grants to A.J.C. from both the Natural Sciences and Engineering Research Council of Canada (NSERC) and the Medical Research Council and by an NSERC postgraduate scholarship to C.D. REFERENCES 1. Abrams, A. 1958. O-acetyl groups in the cell wall of Streptococcus faecalis. J. Biol. Chem. 230:949-959. 2. Bergman, M., and A. Mickeley. 1924. Rearrangements of peptide-like substances. III. Derivatives of dl-serine. Z. Physiol. Chem. 140:128-145. 3. Blake, C. C. F., L. N. Johnson, G. A. Mair, A. T. C. North, D. C. Phillips, and V. R. Sarma. 1967. Crystallographic studies of the activity of hen egg-white lysozyme. Proc. R. Soc. London Ser. B Biol. Sci. 167:378-388.

4324

DUPONT AND CLARKE

4. Blundell, J. K., and H. R. Perkins. 1981. Effects of ,B-lactam antibiotics on peptidoglycan synthesis in growing Neisseria gonorrhoeae, including changes in the degree of O-acetylation. J. Bacteriol. 147:633-641. 5. Blundell, J. K., G. J. Smith, and H. R. Perkins. 1980. The peptidoglycan of Neisseria gonorrhoeae: O-acetyl groups and lysozyme sensitivity. FEMS Microbiol. Lett. 9:259-261. 6. Brumfitt, W. 1959. The mechanism of development of resistance to lysozyme by some Gram-positive bacteria and its results. Br. J. Exp. Pathol. 40:441-451. 7. Brumfitt, W., A. C. Wardlaw, and J. T. Park. 1958. Development of lysozyme resistance in Micrococcus lysodeikticus and its association with an increased O-acetyl content of the cell wall. Nature (London) 181:1783-1784. 8. Dobrogosz, W. J. 1968. N-acetylglucosamine assimilation in Escherichia coli and its relation to catabolite repression. J. Bacteriol. 95:585-591. 9. Dougherty, T. J. 1983. Peptidoglycan synthesis in Neisseria gonorrhoeae strains sensitive and intrinsically resistant to P-lactam antibiotics. J. Bacteriol. 153:429-435. 10. Dougherty, T. J. 1983. Synthesis and modification of the peptidoglycan in Neisseria gonorrhoeae. FEMS Microbiol. Lett. 17:51-53. 11. Dougherty, T. J. 1985. Involvement of a change in penicillin target and peptidoglycan structure in low-level resistance to P-lactam antibiotics in Neisseria gonorrhoeae. Antimicrob. Agents Chemother. 28:90-95. 12. Doyle, R. J., J. Chaloupka, and V. Vinter. 1989. Turnover of cell walls in microorganisms. Microbiol. Rev. 52:554-567. 13. Dupont, C., and A. J. Clarke. 1991. Dependence of lysozymecatalysed solubilization of Proteus mirabilis peptidoglycan on the extent of O-acetylation. Eur. J. Biochem. 195:763-769. 14. Fleck, J., M. Mock, R. Minck, and J. M. Ghuysen. 1971. The cell envelope of P. vulgaris P18. Isolation and characterization of the peptidoglycan component. Biochim. Biophys. Acta 233: 489-503. 15. Fleming, T. J., D. E. Wallsmith, and R. S. Rosenthal. 1986. Arthropathic properties of gonococcal peptidoglycan fragments: implications for the pathogenesis of disseminated gonococcal disease. Infect. Immun. 52:600-608. 16. Fodor, G., and L. Otvos. 1956. Die Raumstruktur der Aminodeoxyzucker. II. Mitteil: Die Konstellation des D-Glucoamins. Chem. Ber. 89:701-708. 17. Ghuysen, J. M., and J. L. Strominger. 1963. Structure of the cell wall of S. aureus strain Copenhagen. II. Separation and structure of disaccharides. Biochemistry 2:1119-1125. 18. Gmeiner, J., and H. P. Kroll. 1981. Murein biosynthesis and O-acetylation of N-acetylmuramic acid during the cell division cycle of Proteus mirabilis. Eur. J. Biochem. 177:171-177. 19. Gmeiner, J., and E. Sarnow. 1987. Murein biosynthesis in synchronized cells of Proteus mirabilis. Quantitative analysis of O-acetylated murein subunits and of chain terminators incorporated into the sacculus during the cell cycle. Eur. J. Biochem. 163:389-395. 20. Goodell, E. W., and C. F. Higgins. 1987. Uptake of cell wall peptides by Salmonella typhimurium and Escherichia coli. J. Bacteriol. 169:3861-3865. 21. Grabow, W. 0. K., and J. A. Smit. 1967. Methionine synthesis in Proteus mirabilis. J. Gen. Microbiol. 46:47-57. 22. Hoyle, B. D., and T. Beveridge. 1983. Metal binding by the peptidoglycan sacculus of Escherichia coli K-12. Can. J. Microbiol. 30:204-211. 23. Imoto, T., L. N. Johnson, A. T. C. North, D. C. Phillips, and J. A. Rupley. 1972. Vertebrate lysozymes, p. 666-868. In P. D. Boyer (ed.), The enzymes, 3rd ed., vol. 7. Academic Press, Inc., New York. 24. Johannsen, L., H. Labischinski, P. Burghaus, and P. Giesbrecht. 1983. Acetylation in different phases of growth of staphylococci and their relation to cell wall degradability by lysozyme, p. 261-266. In R. Hakenbeck, J. Holtje, and H. Labischinski (ed.), The target of penicillin: the murein sacculus of bacterial cell

J. BACTERIOL.

25.

26. 27.

28.

29. 30. 31. 32. 33.

34.

35.

36.

37.

walls; architecture and growth. Walter de Gruyter and Co., New York. Karkhanis, Y. D., J. Y. Zeltner, J. J. Jackson, and D. J. Carlo. 1978. A new and improved microassay to determine 2-keto-3deoxyoctonate in lipopolysaccharide of gram-negative bacteria. Anal. Biochem. 85:595-601. Lear, A. L., and H. R. Perkins. 1983. Degrees of O-acetylation and cross-linking of the peptidoglycan of Neisseria gonorrhoeae during growth. J. Gen. Microbiol. 129:885-888. Lear, A. L., and H. R. Perkins. 1986. O-acetylation of peptidoglycan in Neisseria gonorrhoeae. Investigation of lipid-linked intermediates and glycan chains newly incorporated into the cell wall. J. Gen. Microbiol. 132:2413-2420. Lear, A. L., and H. R. Perkins. 1987. Progress of O-acetylation and cross-linking of peptidoglycan in Neisseria gonorrhoeae grown in the presence of penicillin. J. Gen. Microbiol. 133:17431750. Martin, H. H. 1984. In vitro synthesis of peptidoglycan by spheroplasts of Proteus mirabilis grown in the presence of penicillin. Arch. Microbiol. 139:371-375. Martin, H. H., and J. Gmeiner. 1979. Modification of peptidoglycan structure by penicillin-action in cell walls of Proteus mirabilis. Eur. J. Biochem. 95:487-495. Martin, H. H., W. Schilf, and P. Gruss. 1975. D-Alanyl-Dalanine carboxypeptidase in the bacterial form and L-form of Proteus mirabilis. Eur. J. Biochem. 55:465-473. Morihara, K., and T. Oka. 1973. Effect of secondary interaction on the enzymatic activity of trypsin-like enzymes from Streptomyces. Arch. Biochem. Biophys. 156:764-771. Negro, A., S. Garbisa, L. Gotte, and M. Spina. 1987. The use of reverse-phase high-performance liquid chromatography and precolumn derivatization with DANSYL chloride for quantitation of specific amino acids in collagen and elastin. Anal. Biochem. 160:39-46. Rosenthal, R. S., W. J. Folkening, W. J. Miller, and S. G. Swim. 1983. Resistance of O-acetylated gonococcal peptidoglycan to human peptidoglycan-degrading enzymes. Infect. Immun. 40: 903-911. Rosenthal, R. S., M. A. Gfell, and W. J. Folkening. 1985. Influence of protein synthesis inhibitors on regulation of extent of O-acetylation of gonococcal peptidoglycan. Infect. Immun. 49:7-13. Seidl, P. H., and K. H. Schleifer (ed.). 1985. Biological properties of peptidoglycan. Walter de Gruyter and Co., New York. Sidow, T., L. Johannsen, and H. Labischinski. 1990. Penicillininduced changes in the cell wall composition of Staphylococcus aureus before the onset of bacteriolysis. Arch. Microbiol.

154:73-81. 38. Smith, T. J., and P. E. Hanna. 1986. N-acetyltransferase multiplicity and the bioactivation of N-arylhydroxamic acids by hamster hepatic and intestinal enzymes. Carcinogenesis 7:697702. 39. Snowden, M. A., H. R. Perkins, A. W. Wyke, M. V. Hayes, and J. B. Ward. 1989. Cross-linking and O-acetylation of newly synthesized peptidoglycan in Staphylococcus aureus H. J. Gen. Microbiol. 135:3015-3022. 40. Swim, S. G., M. A. Gfell, C. E. Wilde III, and R. S. Rosenthal. 1983. Strain distribution in extents of lysozyme resistance and O-acetylation of gonococcal peptidoglycan determined by highperformance liquid chromatography. Infect. Immun. 42:446452. 41. Tipper, D. J., J. M. Ghuysen, and J. L. Strominger. 1965. Structure of cell wall of Staphylococcus aureus strain Copenhagen. III. Further studies of the disaccharides. Biochemistry 4:468-473. 42. White, T. 1938. Amino sugars. I. A case of acyl migration. J. Chem. Soc. 1938:1498-1500. 43. Wilhoit, R. C. 1976. Thermodynamic properties of biochemical substances, p. 33-81. In H. D. Brown (ed.), Biochemical microcalorimetry. Academic Press, New York.