Mutant of Staphylococcus aureus - Journal of Bacteriology - American ...

JOURNAL OF BACTERIOLOGY, Aug. 1992, p. 4952-4959

Vol. 174, No. 15

0021-9193/92/154952-08$02.00/0 Copyright © 1992, American Society for Microbiology

Isolation and Characterization of a Tn551-Autolysis Mutant of Staphylococcus aureus TADAHIRO OSHIDA AND ALEXANDER TOMASZ*

The Rockefeller University, 1230 York Avenue, New York, New York 10021 Received 27 January 1992/Accepted 17 May 1992

A Lyt- mutant with reduced autolytic activity was isolated after TnSSI mutagenesis of the methicillinsusceptible Staphylococcus aureus laboratory strain RN450. The Lyt- phenotype could be transferred back into the parent and into a variety of other S. aureus strains by transduction of the transposon marker. Southern analysis has located the Tn551 insert to a 3.2-kb HindIII DNA fragment on the SmaI B fragment of the staphylococcal chromosome. The Lyt- phenotype included reduced rates of cell wall turnover and autolysis induced by detergent or methicillin treatment; however, the rate of methicillin-induced killing was not affected. Peptidoglycans prepared from the parental and mutant cells showed identical muropeptide compositions, as resolved by a high-resolution high-pressure liquid chromatography technique. On the other hand, LiCl extracts of the mutant cells contained reduced amounts of total protein and lower specific cell wall-degrading activity compared with those of extracts of parental cells. The profile of bacteriolytic enzymes as detected by sodium dodecyl sulfate-polyacrylamide gel electrophoresis showed multiple band differences between mutant and parental cells; a major lytic band with properties characteristic of the staphylococcal endo-13-N-acetylglucosaminidase was completely absent from the Lyt- cells. The Lyt- phenotype transduced into a series of methicillin-resistant strains of both homogeneous and heterogeneous phenotypes caused only a modest decrease in the level of methicillin resistance, as determined by population analysis.

physiological function(s) of autolytic enzymes in this bacterium.

Autolytic enzymes, i.e., enzymes that break covalent bonds in the bacterial cell wall, are widespread among bacteria (7, 16). The activities of at least some of these enzymes are clearly involved with phenomena such as bacterial autolysis (induced by antibiotics or adverse physiological conditions) and cell wall turnover. Although speculations concerning the involvement of autolysins in cell wall enlargement, cell division, and/or morphogenetic processes have been repeatedly considered, the physiological function(s) of autolysins for the bacterium remains to be elucidated (21). Recent improvements in the techniques for detection of autolysins indicate that bacteria may contain multiple and, in some cases, a surprisingly large number of autolytic enzymes (11, 18), making the task of identification of physiological functions more complex. Staphylococcus aureus contains at least three kinds of catalytically distinct autolytic enzymes: an amidase, an N-acetyl endoglucosaminidase, and an endopeptidase (8, 17, 19, 20). Molecular cloning of the glucosaminidase (1) and amidase (9) has been reported. Autolytic mutants of staphylococci have also been described (3, 10), and analyses of such mutants have yielded much interesting information about the control of these enzymes and their involvement with cell separation and wall turnover. Nevertheless, all of the mutants were generated by chemical mutagenesis, making a precise association of mutational site and phenotype difficult. In this communication we describe the isolation and characterization of an autolytic mutant of S. aureus in which the lytic defect is clearly associated with the transposon (TnS51) insertional site. The easy selection for the TnS51 marker should allow the introduction of the lytic defect into various genetic backgrounds and thus help to elucidate the

*

MATERIALS AND METHODS Strains and growth conditions. The S. aureus strains used are listed in Table 1. Bacteria were grown in tryptic soy broth (TSB; Difco) with aeration at 37°C and plated on tryptic soy agar (TSA; Difco). Transposon mutants and transductants were grown in TSB containing 10 ,ug of erythromycin per ml in overnight cultures. For each experiment, the overnight culture was diluted 1,000-fold into prewarmed TSB without erythromycin to allow exponential growth. Growth was monitored by the increase in optical density at 620 nm (OD620) in a spectrophotometer (Pharmacia LKB, Piscataway, N.J.) in cuvettes with a 1-cm path

length. Selection of TnSS5 mutants. TnS51 was inserted into the chromosome as described previously (13). In short, RN2906, harboring the thermosensitive plasmid pRN3208 carrying Tn551 with the erythromycin resistance determinant, was plated at different cell concentrations on TSA containing 100 ,ug of erythromycin per ml and incubated at 43°C for 48 h. The approximate frequency of insertion was 6 x 10-4. Colonies were replica plated onto TSA containing 0.25 mM CdNO3 and incubated at 43°C for 16 h to eliminate bacteria that have retained the whole plasmid (about 1% of all erythromycin-resistant colonies that grew at 43°C). Colonies that failed to grow in the presence of cadmium nitrate (indicating the loss of pRN3208) were selected. Each selected colony was grown in TSB containing 10 ,ug of erythromycin per ml at 37°C and used to screen for Lyt- mutants. Screening for Lyt- mutants. Plates used for screening contained 10 ml of TSA bottom agar plus 3 ml of halfstrength (0.5x) TSA containing heat-inactivated S. aureus RN450 cells (3 mg/ml) and about 200 CFU of TnSSI-mutagenized bacteria to be screened. An additional 3 ml of 0.5 x

Corresponding author. 4952

4953

TnS51-AUTOLYSIS MUTANT OF S. AUREUS

VOL. 174, 1992

TABLE 1. S. aureus strains used

Susceptibility

Strain

to methicillina

Relevant

properties and/or

Laboratory strain S (1.5) S RN450(pRN3208) Lytic mutant (TnS5) of RN450 S RUSALl Backcross of RUSALl into RN450 S RUSAL2 RN2677 S Laboratory strain R (6) mec transformant of RN2677 RUSAll R (100) 870b 2491b R (100) 301b R (800) R (1,600) DU4916b R (800-1,600) coLo R (3) Auxiliary TN551 mutant of COL RUSAL0 RN2677d RUSAL3C S R (6) RUSA11d RUSAL4c 870d R (50) RUSAL5C R (50) 2491" RUSAL6c R (400) RUSAL7C 201d R (1,600) DU4916a" RUSAL8c R (100) RUSAL9c coLo R (800) COLd RUSAL16' R (800) RUSAL11c coLo R (800) RUSAL12b COLC R (400) RUSAL16e COLU R (50) COLf RUSAL17e R (12) COLf RUSAL18e R (400) CoLf RUSAL19e R (800) COLf RUSAL20e MIC resistant. within indicate Numbers R, (micrograms per milliliter) for the cell majority (22). susceptible; parentheses aS,

RN450 RN2906

Source

or

reference

genetic cross

Richard Novick Richard Novick This study This study Richard Novick 12

5 This This This This

study study

study

study

This study

This study This study This study This study This study This study This study This study

This study This study

dClinical MRSA isolate. Genetic outcross of RUSALl into the indicated recipient.

c

d

Recipient in genetic

outcross.

Genetic backcross of RUSAL9 into the indicated recipient. f Recipient in genetic backcross. e

TSA was used to cover the surface of the agar plates. The plates were incubated at 37°C for 4 days. Uniform zones of clearing, approximately 5 mm in diameter, appeared around colonies of the parental strain, RN450. Colonies that exhibited smaller lysis zones were picked and examined. A single colony with a barely visible lysis zone around it was designated Lyt- mutant RUSALl and was used as the source of the TnS51-inactivated autolytic determinant(s) in all of the

experiments. Transduction. The method of Pattee (15) was used in a slightly modified form. A phage 80a lysate was prepared from donor strains and used to infect recipient cells. Transductants were selected in plates that contained erythromycin (30 ,ug/ml) in the 10-ml (total volume) bottom layer of agar; no drug was added to the middle (20-ml) layer or to the top (3-ml) layer; the top layer contained the bacteria. MIC determination and population analysis. The MIC was determined by agar dilution as follows. Overnight cultures were diluted and spread on TSA plates containing different concentrations of methicillin. Colonies were counted after incubation at 37°C for 2 days. Population analysis was performed as previously described (22). Autolysis assay. Triton X-100-stimulated autolysis was measured as described previously (6). Cells were grown exponentially to an OD620 of 0.3. The culture was quickly chilled, and the cells were harvested by centrifugation (10,000 x g, 4°C, 5 min). Pellets were washed once with cold distilled water and suspended to an OD620 of 1.0 in 50 mM glycine-0.01% Triton X-100 buffer (pH 8.0). Autolysis was measured during incubation with aeration at 37°C as a

decrease in OD620 with a model 340 spectrophotometer

(Sequoia-Turner Corp., Mountainview, Calif.).

Cell wall turnover. Cells were labeled for three to six

generations (up to an OD620 of 0.2 to 0.3) in 10 ml of TSB containing [3H]N-acetylglucosamine (18.5 KBq/ml, 3.7 GBq/ mmol) with aeration at 37°C. After labeling, the culture was centrifuged at room temperature for 5 min at 10,000 x g. The pellet was resuspended to an OD620 of 0.02 in 50 ml of prewarmed fresh TSB containing cold 10 mM N-acetylglucosamine. Samples of 2 ml were withdrawn for the measurement of OD620, and the radioactivity remaining in the cell wall was determined as previously described (6). Preparation of labeled cell wall and assay of autolytic activity. [3H]N-acetylglucosamine (37 KBq/ml; specific activity, 214 GBp/mmol) was added to exponentially growing cells. The cells were grown to an OD620 of 0.5 for three generations and then harvested. Walls were prepared as described previously (4). To extract crude autolytic enzyme, cells were grown to an OD620 of 1.0, chilled quickly, and centrifuged (10,000 x g, 4°C, 5 min). The cells were washed twice with cold distilled water and suspended in 3 M LiCl (0.6 ml/g of wet cells) for 30 min. The suspension was centrifuged, and the activity of autolytic enzyme in the extract (supernatant) was measured by its ability to solubilize radioactively labeled cell walls. The 3H-labeled cell walls (0.2 mg/ml) were mixed with the extract and 70 mM sodium phosphate buffer (pH 7.0, 0.5-ml total volume) and incubated at 37°C. At intervals, 100-,ul samples were withdrawn, mixed with 20 ,ul of formaldehyde (34%), and chilled on ice. After centrifugation at 4°C, radioactivity in 100 plI of

4954

OSHIDA AND TOMASZ

J. BACTERIOL.

1 2A Kb

Kb

388P

388'+388*

48.5-P

48.5.

FIG. 1. Localization of the Lyt- TnSSl insert on the staphylococcal chromosome by pulsed-field electrophoresis. SmnaI digests of chromosomal DNA isolated from the parental strain RN450, the Lyt- mutant, and several Lytv transductants were analyzed by pulsed-field electrophoresis. Bands (a) were probed with radioactively labeled pRT1, a plasmid carrying an Xbal1-Hpa I fragment of TnSSJ (21) (b). Lanes: strain RN450 (A); LytF mutant RUSALl (B); Lyt- transductant RUSAL2 (C); strain RN2677 (D) and its LytF transductant RUSAL3 (E); RUSAll (pI258), a mec transformant of RN2677 that carries plasmid pI258 with Tn5Sl (F); RUSAL4, a Lyt- transductant of RUSAll (G); MRSA strain DU4916 (H) and its LytF transductant RUSAL8 (I); MRSA strain COL (J) and two of its Lyt- transductants, RUSAL11 (K) and RUSAL12 (L). Lanes marked 1 and 2, contain molecular size markers. Preparations in lanes A, D, H, and J contained no TnSSJ and therefore did not react with the radioactive pRT1 probe.

supernatant was counted in Ready Safe scintillation cocktail (Beckman). The activities of samples were calculated as percentages of the total radioactive label incorporated into the cell wall. Reference samples representing 100% of the incorporated radioactive label were digested completely with an excess amount of the extract from the staphylococcal strain RN450, and radioactivity was measured without centrifugation after formaldehyde was added. The protein concentrations of cell extracts were determined by the Coomassie blue assay (2). Bacteriolytic enzyme profiles on sodium dodecyl sulfatepolyacrylamide gels. Bacteriolytic enzyme profiles were analyzed as described by LeClerc and Asselin (11). The gels contained heat-inactivated cells (1 mg of dry cells per ml) of either S. aureus RN450 or Micrococcus luteus ATCC 4698 (Sigma). After electrophoresis, the gels were washed for 1 h with several changes of distilled water and incubated in 0.1 M sodium phosphate buffer (pH 7.0) at 37°C for 24 h. The gels were stained in 0.1% methylene blue for 3 min and then destained extensively with distilled water overnight. Other methods. Cell wall peptidoglycan was prepared and enzymatic cell wall hydrolysates were analyzed with reversed-phase high-pressure liquid chromatography as described previously (4). RESULTS Isolation of lytic mutant RUSALl and transduction of the Lyt- phenotype with the TnSSI marker. After insertion of TnSSJ into the chromosome of S. aureus RN450, 2,060 erythromycin resistant colonies capable of growing at 43°C were obtained. Bacterial colonies were examined for autolytic activity in plates containing heat-inactivated S. aureus cells. Only 1 colony out of 2,060 examined exhibited a clearing zone that was substantially smaller than those of parental (RN450) colonies. This Lyt- colony was designated RUSALl. Another property of RUSALI cultures was that

they autolyzed more slowly than did cultures of RN450 during treatment with methicillin (10 ,ug/ml) (data not shown). The TnSSJ marker of RUSALl was backcrossed into the parent (RN450) by transduction. Backcrosses were obtained with a frequency of 3.3 x 10'. Each of the three transductants examined showed the same Lyt- phenotype (small lysis zone around colonies and slow methicillin-induced autolysis) as RUSALl. These findings indicated that the Lyt- phenotype of mutant RUSALl was caused by the TnSSJ insertion into the chromosome. One of the Lyttransductants, designated RUSAL2, was used in most of the physiological studies. Localization of the insertional site. The chromosomal location of the TnSSl insert in RUSALl and its several transductants in a variety of S. aureus recipients was tested by pulsed-field electrophoresis of SmaI chromosomal fragments and probing with radioactively labeled Tn551 (Fig. 1). The location of TnSSI in RUSALl and RUSAL2 was also examined by conventional Southern analysis with TnSSI as a probe. The probe hybridized with DNA fragments of RUSALl and RUSAL2 in the same manner; no signal was detected with RN450. From the estimated lengths of the three TnS51-hybridizing fragments produced and from the known position of the two HindIII sites in Tn917, which has a sequence highly similar to that of TnS5J, the insertion site of TnSSI could be located within a 3.2-kb HindIII fragment that was contained within a 17-kb EcoRI fragment (Fig. 2). Mechanism of lysis defect and cell wall composition. The TnS51-interrupted genetic determinant may be involved with the synthesis of cell wall muropeptides, causing the formation of an abnormal (autolysin-resistant) cell wall. Peptidoglycans prepared from the parental strain RN450 and from the isogenic Lyt- transductant RUSAL2 were purified, and the enzymatic hydrolysates containing disaccharide peptides were analyzed by a recently developed high-resolution tech-

VOL. 174, 1992


B

A

Kb

4955

+

23.1 9.4-0 6.6-0

4.44-o -

4-205 4-116 4-97

-466

2.3-o

2.000

4-45 FIG. 2. Southern analysis of TnSSl insertion in S. aureus RUSALl and its backcross, RUSAL2. Total DNAs from RN450, RUSAL1, and RUSAL2 were digested with HindIII and/or EcoRI and probed with pRT1 (see Fig. 1 legend). The XbaI-HpaI fragment has two restriction sites for HindIll and no site for EcoRI. Lanes: HindIlI digestion of RN450 (A), RUSALl (B), and RUSAL2 (C); EcoRI digestion of RN450 (D), RUSALl (E), and RUSAL2 (F); HindIII-EcoRI double digestion of RN450 (G), RUSALl (H), and RUSAL2 (I); molecular weight markers (lambda DNA digested with HindIII) (X).

nique (4). No differences in the muropeptide compositions of parental and mutant peptidoglycans were detected (Fig. 3). Profile of bacteriolytic enzymes. Zymograms of sodium dodecyl sulfate extracts prepared from RN450 and RUSAL2 showed marked and multiple differences. With heat-inactivated S. aureus cells in the gel, several lytic bands were fainter in the RUSAL2 extracts and one particular band appeared to be more intense than the bands of corresponding mobility in the parental extracts. With heat-inactivated M. luteus cells as a substrate in the gels, a single intense lytic band of RN450 extracts was completely missing from the extracts of the lytic mutant RUSAL2 (Fig. 4). The electro-

-*29 FIG. 4. Bacteriolytic enzyme profiles of S. aureus RN450 and RUSAL12. SDS-polyacrylamide gels containing heat-inactivated cells of S. aureus (A) or M. luteus (B) were prepared, and SDS cell extracts (13 ,ug of protein each) of RN450 (+) and RUSAL2 (-) were electrophoresed as described in Materials and Methods. After electrophoresis, the gels were washed with distilled water, incubated in 0.1 M phosphate buffer (pH 7.0) at 37°C for 24 h, and then assayed for enzyme activity.

phoretic mobility and the lack of lytic activity with S. aureus walls are consistent with this band representing the 51-kDa endo-o-N-acetylglucosaminidase described by Sugai et al. (18). Similar observations were made with LiCl extracts of the bacteria (data not shown). Solubilization of radiolabeled cell wall by crude autolytic extracts. To compare the activities of autolytic enzymes of RUSAL2 with those of RN450, LiCl cell extracts and radiolabeled cell walls were prepared from both strains. The amount of protein in the LiCl extract from RUSAL2 (1.2 mg/g of dry cells) was much less than the amount of protein recovered in the extract of RN450 (3.4 mg/g of dry cells). This indicated that the amount of protein noncovalently

Et ! ,,,L~~~~

80

RN450

RUSAL2

60 _

c

40-

200

0 FIG.

3.

50 Separation

of

Time

(min)

100

150

muropeptides by reversed-phase high-

pressure liquid chromatography in strain RN450 (A) and the LytTn551 mutant RUSAL2 (B). Peptidoglycan was obtained and digested with a muramidase. Separation was done as described previously (4).

1

2 0 Time (h)

1

2

FIG. 5. Solubilization of labeled cell wall by LiCl cell extracts of S. aureus RN450 or RUSAL2. Various amounts of crude bacteriolytic enzyme extracts from S. aureus RN450 or RUSAL2 were incubated with cell wall labeled with [3H]N-acetylglucosamine prepared from RN450. Symbols: 8 (-), 4 (A), 2 (-), 1 (0J), 0.5 (A), and 0.25 (0) ,ug of protein per ml. Samples were withdrawn at intervals, and the amount of radioactivity released into the supernatant was determined.

4956

J. BACTERIOL.

OSHIDA AND TOMASZ

800

600 20 0 S E 400 m

0 1 2 3

5

24

Time (h) FIG. 6. Release of reducing sugars and free amino groups during incubation of the cell wall of S. aureus RN450 with LiCl cell extracts of strain RN450 or RUSAL2. Cell walls (1 mg/ml) were suspended in 20 mM sodium phosphate buffer (pH 7.0) and incubated with LiCl extracts (10 F±g of protein per ml) from S. aureus RN450 (A, 0) or RUSAL2 (A, 0). Samples were removed at intervals and boiled for 5 min to stop the reaction. The concentration of reducing sugars (0, *) and free amino groups (A, A) in the soluble fraction were determined as described previously (7, 14).

bound to the cell wall was three times less in RUSAL2 than in RN450. The extract of RN450 hydrolyzed the cell walls of RN450 and RUSAL2 with similar rates at all protein concentrations tested. The specific wall-hydrolyzing activity per milligram of protein in the LiCl extracts of RUSAL2 was about one-third of that of parental extracts (Fig. 5). Since the total amount of protein in the mutant extracts was already about three times less (per milligram of cell weight) than that in strain RN450, the autolytic activity of the mutant cells appears to be about nine times lower than that of the parental cells. The amounts of reducing sugars and free amino groups released during hydrolysis of S. aureus cell walls in the RN450 and RUSAL2 LiCl cell extracts were compared (Fig. 6). The cell extract of RN450 released N-terminal amino groups about four times as fast as the RUSAL2 extract did. An even more pronounced difference was observed in the release of reducing sugars: glucosidase activity was virtually absent from RUSAL2 extracts. Effect of the Lyt- mutation on some properties of the bacteria. Growing cultures of RN450 and RUSA12 were

compared for their rates of growth and cell wall turnover, antibiotic (methicillin)-induced loss of viability, and rates of autolysis induced by treatments with detergent (Triton X-100) or methicillin. RN450 and RUSAL2 each grew with a doubling time of 26 min in TSB at 37°C. We observed substantial differences in the cell wall turnover rates of various Lyt+ strains of S. aureus and their respective isogenic Lyt- derivatives generated via transduction with the Tn551-inactivated lyt gene(s) from RUSAL1; the degree of suppression of turnover rates clearly depended on the genetic background (Fig. 7). Introduction of the Lyt- marker also suppressed methicillin-induced autolysis; however, the rates of loss of viability during methicillin treatment of the isogenic pairs of strains were similar (Fig. 8). The Lyt- transductants all showed greatly reduced rates of autolysis induced by Triton X-100 (Fig. 9). The transposon-inactivated Lyt- determinant was transduced from the TnS5l mutant RUSALl into a number of methicillin-resistant S. aureus (MRSA) clinical isolates with different (homogeneous and heterogeneous) modes of expression of antibiotic resistance. The recipient strains showed widely different rates of methicillin-induced autolysis (the extremes being the very slow-lysing strain DU4916 and the fast-lysing strain COL), whereas all of the Lyttransductant derivatives of these MRSA strains showed the reduced autolysis rate characteristic of RUSALl (data not shown). Figure 10 compares the methicillin resistance levels (population analysis profiles) of the various staphylococcal strains used as recipients in the genetic crosses and their Lyt- derivatives. Introduction of the lytic defect had no detectable effect on the MIC values for MRSA strain DU4916 and the two methicillin-susceptible strains and only caused a modest shift in the direction of lower resistance in most of the other MRSA isolates tested. In only one of six Lyt- transductants of strain COL (RUSAL9) purified was there a more substantial (16-fold) decrease in the methicillin MIC. However, when RUSAL9 was used as a source of transducing lysate, which was then backcrossed into strain COL, most of the transductants again showed only a modest reduction of methicillin resistance (from the MIC for COL of 1,600 to 800 or 400 ,ug of methicillin per ml), and only 2 of the 10 transductants tested showed more substantial decreases in methicillin MICs (100 and 12.5 p,g/ml, respectively).

100

110 0

:S 6

8

0

No. of generation FIG. 7. Suppression of cell wall turnover in the Lyt- strains of S. aureus. (A) Symbols: parent strain RN450 (0) and Lyt- transductant RUSAL2 (0). (B) Symbols: Lyt+ recipient MRSA strain COL (0) and its Lyt- transductants RUSAL9 (0), RUSAL10 (A), and RUSAL11 (l); RUSA10, a TnS51-induced cell wall mutant of strain COL (A).


VOL. 174, 1992

FIG. 8. Suppression of methicillin-induced autolysis in the Lytstrain of S. aureus. (A) Methicillin was added to exponentially growing cultures of the parental strain RN450 and its Lyt- transductant RUSAL2 at time zero, and the turbidity (OD620) of the cultures was monitored at intervals. (B) At the same intervals, samples were withdrawn, diluted, and plated on TSA for determination of viability. Symbols: control (0) and cultures incubated with 0.5 (A), 1.0 (E), 2.0 (-), 5.0 (A), and 10 (O) ,ug of methicillin per ml.

A RUSAL2

RN450

4957

1 0

(0 0

0

DISCUSSION

.1

.01

i

0

.2

4 401 0 Time (h)

2

3

The Lyt- mutant RUSALl appears to be the first TnSSlinduced autolytic mutant of S. aureus. The successful transduction of the Lyt- phenotype via selection for the TnSSI marker into a variety of staphylococcal backgrounds indicates that the phenotype of this mutant is caused by the insertional inactivation of a chromosomal determinant. As to the nature of the inactivated gene, we can only speculate. Unlike the recently described auxiliary TnSSl mutants, in which autolysis and cell wall turnover were suppressed (6, 12), the Lyt- transposon mutant described here had no effect on the muropeptide composition of the cells. However, one should note that the analytical method used (4) would not resolve differences in other aspects of wall structure (e.g., degree of 0 acetylation, glycan chain length, etc.) that may influence autolysis. Nevertheless, the data together suggest that the autolytic phenotype may be related to some change(s) in the complement of autolytic enzymes that are detectable by the method of LeClerc and Asselin (11, 18). The most striking of the multiple autolytic band differences between the parental and isogenic Lyt- cells was the virtually complete absence from the mutant extract of a band that most likely represents the endo N-acetylglucosaminidase (19). The absence of this band also explains the lack of production of reducing groups from cell walls incubated with crude autolysin extracts from the mutant cells. However, the inactivated determinant is more likely to be a regulatory function than a structural gene for an autolysin, since the

4

B 9 8

=7

U-

06 4 3 Time (h)

(n80] 0 6601 t400

cu20-

0

60

120O

180

0

60

120

180

Time (min) FIG. 9. Suppression of detergent-induced autolysis in the Lyt- mutant of S. aureus. Bacterial strains were grown and labeled biosynthetically with [3H]-N-acetylglucosamine in their cell walls. Autolysis was induced by Triton X-100, and the rate of release of the radioactive label was measured. (A) Symbols: parent strain RN450 (0), Lyt- transductant RUSAL2 (-). (B) Symbols: Lyt' MRSA strain COL (0); Lyt- transductants RUSAL9 (-) and RUSAL10 (A); RUSA10; a Tn551-induced cell wall mutant of COL (A).

4958

J. BACTE RIOL.

OSHIDA AND TOMASZ

9 -0-° RN450 -* RUSAL2

7

-

- -

RN2677 RUSAL3

5 , . .

1

--- RUSAl 1 -.- RUSAL4

. . . . .l.l I. I. ...L. .j,I.I I

I

I

I

9

LL 7

0 0 3

°

2491 301 COL -U9 870 - -RUSAL6- - RUSAL7 -.-RUSAL5

7 5 4e DU4916 * RUSAL8

3 .

o

,

...LQ ,3L L

,,,,

CO RUSAL9 -

~

~

~O

00 cn0 C0 0o 00 0 0 0 -8

Methicillin concentration (pg/ml) FIG. 10. Expression of methicillin resistance in Lyt- transductans of MRSA isolates. Overnight cultures of bacteria were plated at different cell concentrations on TSA containing serial dilutions of methicillin and incubated at 37°C for 48 h before colonies were counted. Symbols: parental strains (0), Lyt- transductants (@). The lower right panel shows population analysis profiles of the following genetic backcrosses with transducing lysates of RUSAL9 and strain COL as the recipient: COL (0); backcrosses RUSAL16 (-), RUSAL17 (A), RUSAL18 (A), RUSAL9 (O), and RUSAL20 (-).

band differences between parental and mutant extracts were clearly multiple. The substantial (threefold) decrease in the amounts of total cell wall-bound protein in RUSAL2 (as compared with those in the parental cells) suggests that the inactivated function is involved with the transport of autolysins to the outer surface of the bacterium. A major advantage of the Tn551-linked Lyt- phenotype is in the relative ease with which the inactivated autolytic phenotype may be introduced into various genetic backgrounds for testing the physiological effects of the autolytic system. As expected, the Tn551-linked Lyt- marker resulted in the suppression of both detergent- and antibioticinduced autolysis and cell wall turnover in all of the genetic backgrounds tested. On the other hand, inactivation of autolysis and suppression of cell wall turnover in MRSA isolates had only a minor effect on the level of methicillin resistance in most of the transductants. This finding is of importance, since other Tn5SI mutants (MRSA auxiliary

mutants [22]) were shown to cause dramatic parallel decrease both in the methicillin MIC and in the wall turnover rate of the highly and homogeneously resistant strain COL (6) and similar findings were reported for the so-called femA mutants of another MRSA strain (12). Furthermore, the reduced MIC and wall turnover cotransduced in the backcrosses (6). On the basis of this finding, it was proposed that cell wall turnover has a direct functional role in the expression of high-level antibiotic resistance in staphylococci, in a functional analogy with DNA excision repair, by providing a mechanism for the removal of structurally abnormal (and potentially lethal) pieces of cell wall material that may incorporate into the peptidoglycan of the cells during exposure to beta-lactam antibiotics (6). A recent biochemical analysis of these auxiliary mutants showed that these bacteria produced peptidoglycan of drastically altered composition (5) and that the reduced rate of wall turnover in these mutants was most likely a conse-


VOL. 174, 1992

quence of this change in the chemical structure of peptidoglycan, making these polymers poor substrates for the enzymes catalyzing cell wall turnover; this suggestion had been made earlier (12). If the reduction of antibiotic resistance in the auxiliary mutants were, indeed, related to their gross slower rate of wall turnover, then a direct suppression of this process (e.g., in the autolysin-defective TnS5I mutant) would also be expected to reduce the level of antibiotic resistance in MRSA. The availability of the TnS51-linked autolytic mutant has allowed us to test this hypothesis. Transduction of the Lyt- phenotype caused only a minor decrease in the level of methicillin resistance, whereas cell wall turnover rates and autolytic rates were drastically reduced. Southern analysis of RUSAL9 (the Lyt- transductant of MRSA strain COL) and its backcrosses into COL (constructs RUSAL9-6 through 9-10; Fig. 10) showed that TnSSl was retained in its original position in all of these transductants, irrespective of whether they had lower methicillin MICs. These findings indicate that even a massive reduction in cell wall turnover need not necessarily bring about reduction in methicillin resistance. It is conceivable that the particular murein hydrolase needed to remove the fraudulent (antibiotic-induced) pieces of cell wall remains intact in the Lyt- Tn551 mutant. Alternatively, the findings may imply that the drastic reduction in methicillin resistance level observed in the auxiliary mutants was caused not by the slowdown in wall turnover but rather by some other biosynthetic misfunction related to the abnormal muropeptide composition of the mutant cell walls. The relative ease with which the TnS51-linked Lyt- phenotype can be introduced into various staphylococcal backgrounds should make the RUSALl mutation a useful tool for testing the role of the staphylococcal autolytic system in other phenomena. ACKNOWLEDGMENT These investigations received partial support from Public Health Service grant R01 AI16794 from the National Institutes of Health.

REFERENCES 1. Biavasco, F., C. Pruzzo, and C. Thomas. 1988. Cloning and expression of the Staphylococcus aureus glucosaminidase in Escherichia coli. FEMS Microbiol. Lett. 49:137-142. 2. Bradford, M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principles of protein-dye binding. Anal. Biochem. 72:248-254. 3. Chatterjee, A. N., W. Wong, F. E. Young, and R. W. Gilpin. 1976. Isolation and characterization of a mutant of Staphylococcus aureus deficient in autolytic activity. J. Bacteriol. 125:961967. 4. De Jonge, B. L. M., Y.-S. Chang, D. Gage, and A. Tomasz. 1992. Peptidoglycan composition of a highly methicillin resistant Staphylococcus aureus strain: the role of penicillin binding protein 2A. J. Biol. Chem., in press.

4959

5. De Jonge, B. L. M., Y.-S. Chang, D. Gage, and A. Tomasz. 1992. Peptidoglycan composition in heterogeneous Tn551 mutants of a methicillin-resistant Staphylococcus aureus strain. J. Biol. Chem., in press. 6. De Jonge, B., H. de Lencastre, and A. Tomasz. 1991. Suppression of autolysis and cell wall turnover in heterogeneous Tn551 mutants of a methicillin-resistant Staphylococcus aureus strain. J. Bacteriol. 173:1105-1110. 7. Ghuysen, J.-M., D. J. Tipper, and J. L. Strominger. 1966. Enzymes that degrade bacterial cell walls. Methods Enzymol. 8:685-699. 8. Huff, E. C., C. S. Silverman, N. J. Adams, and W. S. Awkard. 1970. Extracellular cell wall lytic enzyme from Staphylococcus aureus: purification and partial characterization. J. Bacteriol. 103:761-769. 9. Jayaswal, R. K., Y.-I. Lee, and B. J. Wilkinson. 1990. Cloning and expression of a Staphylococcus aureus gene encoding a peptidoglycan hydrolase activity. J. Bacteriol. 172:5783-5788. 10. Koyama, T., M. Yamada, and M. Matsuhashi. 1977. Formation of regular packets of Staphylococcus aureus cells. J. Bacteriol.

129:1518-1523. 11. LeClerc, D., and A. Asselin. 1989. Detection of bacterial cell wall hydrolases after denaturing polyacrylamide gel electrophoresis. Can. J. Microbiol. 35:749-753. 12. Maidhof, H., B. Reinicke, P. Blumel, B. Berger-Bachi, and H. Labischinsld. 1991. femA, which encodes a factor essential for expression of methicillin resistance, affects glycine content of peptidoglycan in methicillin-resistant and methicillin-susceptible Staphylococcus aureus strains. J. Bacteriol. 173:3507-3513. 13. Murakami, K., and A. Tomasz. 1989. Involvement of multiple genetic determinants in high-level methicillin resistance in Staphylococcus aureus. J. Bacteriol. 171:874-879. 14. Park, J. T., and M. J. Johnson. 1949. A submicrodetermination of glucos. J. Biol. Chem. 181:149-151. 15. Pattee, P. A. 1987. Staphylococcus aureus. Genet. Maps 4:148154. 16. Perkins, H. R. 1980. The bacterial autolysins, p. 437-456. In H. J. Rogers, H. R. Perkins, and J. B. Ward (ed.), Microbial cell walls and membranes. Chapman and Hall, Ltd., London. 17. Singer, H. J., E. M. Wise, Jr., and J. T. Park 1972. Properties and purification of N-acetylmuramyl-L-alanine amidase from Staphylococcus aureus H. J. Bacteriol. 112:932-939. 18. Sugai, M., T. Aldyama, H. Komatsuzawa, Y. Miyake, and H. Suginaka. 1990. Characterization of sodium dodecyl sulfatestable Staphylococcus aureus bacteriolytic enzymes by polyacrylamide gel electrophoresis. J. Bacteriol. 172:6494-6498. 19. Sugai, M., H. Koike, Y.-M. Hong, Y. Miyake, R. Nogami, and H. Suginaka. 1989. Purification of a 51 kDa endo-o-N-acetylglucosaminidase from Staphylococcus aureus. FEMS Microbiol. Lett. 61:267-272. 20. Tipper, D. J. 1969. Mechanism of autolysis of isolated cell walls of Staphylococcus aureus. J. Bacteriol. 97:837-847. 21. Tomasz, A. 1984. Building and breaking of bonds in the cell wall of bacteria-the role for autolysins, p. 3-12. In C. Nombela (ed.), Microbial cell wall synthesis and autolysis. Elsevier/ North-Holland Publishing Co., Amsterdam. 22. Tomasz, A. 1990. Auxiliary genes assisting in the expression of methicillin resistance in Staphylococcus aureus, p. 565-583. In R. P. Novick (ed.), Molecular biology of the staphylococci. VCH Publishers, Inc., New York.