INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Pneumococci differ from other gram-positive bacteria in that their lipoteichoic acid (LTA) and wall teichoic acid (TA) have the same chain structure which is, moreover, unusually complex (Fig. 1): glycerophosphate is replaced by ribitol phosphate (7), and between the ribitol phosphate residues a tetrasaccharide is intercalated (23). It contains D-glucose, 2-acetamido-4-amino-2,4,6-trideoxy-D-galactose (AATGal), and two N-acetyl-D-galactosaminyl residues, one or both of which carry a phosphocholine residue at O-6 (references 3 and 12 and this report).

View larger version (22K): [in this window] [in a new window]   FIG. 1.   Pneumococcal TA and LTA. As shown, in strain R6 most of the repeats carry two phosphocholine residues each, at O-6 of the N-acetyl-D-galactosaminyl residues (3, 12). In strain Rx1 and Rx1/AL, most repeats contain one phosphocholine residue (this report) attached to O-6 of the non-ribitol-linked galactosaminyl residue (14).

Pneumococci are not able to synthesize the choline required for the synthesis of these substituents. Moreover, choline is an essential growth factor (2, 30) but can be substituted in this function by nutritional ethanolamine (EA) (38). Phosphoethanolamine is incorporated into LTA and TA in place of phosphocholine (14), but it cannot replace phosphocholine functionally. Phosphocholine-substituted LTA serves to anchor pneumococcal surface protein A (PspA) to the outer layer of the cytoplasmic membrane, with choline-mediated interaction between membrane-associated LTA and the C-terminal repeat region of PspA. In EA-grown bacteria, PspA is no longer retained and is released into the surrounding medium (45). Phosphocholine substituents also play an essential role for the activity of the major pneumococcal autolysin, an N-acetylmuramyl-L-alanine amidase (38). This protein possesses a choline-binding C-terminal domain that is essential for activity but, unlike PspA, is not essential for retention on the pneumococcal cell surface (16, 32). Binding of phosphocholine-substituted LTA to this domain results in potent inhibition of the amidase (21). The inhibitory property is dependent on the micellar structure of LTA (13) and lost by deacylation (5). Phosphocholine-substituted LTA may also participate in the transport of the amidase through the cytoplasmic membrane from the cytosol (5), the location of its synthesis (15). It additionally effects the conversion of the inactive E form of the enzyme into the active C form (5). This conversion is likewise effected by the choline residues of cell wall-linked TA (33, 39). Furthermore, binding of the amidase to the choline residues of TA is prerequisite for the hydrolysis of cell walls by the enzyme (18, 22). It should be noted that the amidase is not essential for growth. Though the enzyme is completely inactive in EA grown cells, the growth rate is not affected. However, cell separation is impaired, and there is a loss of stationary-phase and penicillin-induced cell lysis (38, 40), as well as a loss of genetic transformation (38). After insertional inactivation of the autolysin gene (lytA), the autolysin-deficient mutants (Lyt) grew normally (31) and did not even show impeded cell separation (41).

In this report, we describe a mutant which acquired the ability to grow in the absence of choline and analogs. Except for the observation that [³H]choline-substituted LTA is not a precursor of [³H]choline-substituted TA (6), nothing is known about the biosyntheses of pneumococcal LTA and TA and the stage of biosynthesis at which phosphocholine is incorporated. Since the absence of choline incorporation might affect the structure of LTA and TA as well as the composition of cell walls, we included relevant analyses in our study.

(A preliminary report of this work was presented in an overview on pneumococcal LTA and TA at the International Meeting on the Molecular Biology of Streptococcus pneumoniae and Its Diseases, Oeiras, Portugal, September 24 to 29, 1996 [10].)

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Materials. CPL, a muramidase from the pneumococcal bacteriophage Cp-1 (17), and pneumococcal N-acetylmuramyl-L-alanine amidase (15, 33), purified as described in the references cited, were provided by Ernesto Garcia (Centro de Investigaciones Biológicas, Madrid, Spain). Cellosyl from Streptomyces coelicolor was provided by Stefan Müllner (Höchst AG, Frankfurt am Main, Germany). Other enzymes and coenzymes were from Boehringer Mannheim GmbH (Mannheim, Germany) or Sigma-Aldrich Chemie (Deisenhofen, Germany). LTA- and TA-containing cell walls of S. pneumoniae R6 were from previous work (3, 12). Muramic acid-6-phosphate was prepared from Staphylococcus aureus cell walls as described elsewhere (25).

Bacterial strains, growth, and transformation. S. pneumoniae Rx1 (36) and its derivative JY2190 were grown at 37°C in either chemically defined medium (CDM) (42), prepared by JRH Biosciences (Denver, Pa.), or Todd-Hewitt broth supplemented with 0.5% yeast extract (THY; Difco Laboratories, Detroit, Mich.) or on blood agar base no. 2 (BAP; Difco) containing 3% sheep blood. Where indicated, CDM was supplemented with EA (0.02%) or choline (0.0005%). For chemical analyses of the parent Rx1, the nonlytic derivative Rx1/AL (4), provided by James Paton (Adelaide Children's Hospital, North Adelaide, South Australia, Australia), was grown at 37°C in a complex medium (10 g of meat extract, 20 g of casein peptone, 5 g of yeast extract, 2 g of NaCl, 2 g of K₂HPO₄, 4 g of glucose, 1 liter of H₂O). Bacteria were stored as frozen stocks at 85°C in growth medium containing 10% glycerol. Transformation of S. pneumoniae was performed as previously described (43). DNA was obtained by phenol-chloroform extraction when the mutant strain was grown in CDM or by standard deoxycholate lysis (43) when it was grown in choline-containing medium. Rx1 recipients transformed by JY2190 DNA were initially selected for by the ability to grow in liquid CDM lacking choline. Isolated colonies obtained from samples plated on blood agar medium were then tested for the ability to grow in the absence of choline.

Immunoblot analyses. For immunodot blot analyses of PspA and autolysin, cell lysates and culture supernatant fluids (clarified through 0.45-µm-pore-size low-protein-binding filters) were prepared as previously described (44, 45). Samples from mid-exponential-phase cultures grown to equivalent optical densities (ODs) were spotted on a nitrocellulose membrane and developed as previously described for Western blots (44, 45). Equivalent amounts of all samples, representing 1 µl of unconcentrated culture, were examined. Monoclonal antibodies Xi136 (PspA specific [26]) and 140.1C2 (phosphocholine specific) were provided by David Briles (University of Alabama at Birmingham). Autolysin-specific polyclonal antiserum (4) was provided by James Paton.

Microscopy. Bacteria, grown to mid-exponential phase in the indicated media, were observed under phase contrast. Photomicrographs represent a final magnification of ×862. For electron microscopy, samples were centrifuged, washed once in 1/20 volume of phosphate-buffered saline (PBS), fixed in 1/20 volume of 1% glutaraldehyde-PBS, and then resuspended in 1/5 volume of PBS. Electron micrographs represent a final magnification of ×50,000.

Extraction and purification of LTA and cell walls. LTA was extracted from late-exponential-phase bacteria as described previously (3) except that the bacteria were disintegrated with glass beads in a Braun disintegrator (11). Briefly, LTA and lipids were extracted from broken cells with a Bligh-Dyer monophasic system and separated from cell walls by centrifugation. To the supernatant fluid, CHCl₃ and H₂O were added to achieve phase separation. The methanol-H₂O layer was dialyzed, freeze-dried, and for purification of LTA, subjected to hydrophobic interaction chromatography on octyl-Sepharose (3). For analysis of TA, cell walls were prepared from disintegrated bacteria and purified as previously described (20) by hot sodium dodecyl sulfate extraction, digestion with nucleases and trypsin, a second sodium dodecyl sulfate extraction, and several washing steps, including one in 5.8 M LiCl. After washing with H₂O, cell walls were freeze-dried.

Analytical procedures. Carbohydrate (28), choline (1), D-glucose (24), glycerol (27), hexosamine (37), periodate (8), formic acid (34), and phosphate (35) were measured as described in each reference. Ribitol and anhydroribitol were quantified as acetates by gas-liquid chromatography (internal standard, mannitol) (3). Galactosamine, glucosamine, muramic acid, muramitol, quinovosamine, and EA were identified and quantified (internal standard, taurine) as fluorescent fluoren-9-yl-methoxycarbonyl (Fmoc) derivatives by reverse-phase high-pressure liquid chromatography (HPLC) (12). Tetrasaccharide ribitol and the anomeric forms of the tetrasaccharide released by HF from pneumococcal LTAs and TAs (see below) were separated as Fmoc derivatives by reverse-phase HPLC using the recently described elution program (12) with a modified flow rate (1 ml/min).

Periodate oxidation. Cell walls from the mutant strain were solubilized by treatment with cellosyl (see below), and the phosphate-containing products were isolated by column chromatography on DEAE-Sephadex A-25 which was eluted with water, followed by a linear gradient of 0 to 0.3 M NH₄HCO₃ (pH 8.2). After desalting by passage through a column of Sephadex G-15, the phosphate-containing products (20 µmol of phosphate) were oxidized in 0.2 M NaIO₄ (1.55 ml) containing 0.1 M sodium acetate (pH 4.7) at room temperature in the dark. The oxidation was monitored photometrically (8).

Amidase activity. The activity of pneumococcal amidase was assayed in a standard mixture consisting of 1.5 mg of cell wall (0.5 µmol of N-acetylmuramic acid) and 50 mM HEPES (pH 7.0) containing 0.04% NaN₃ (1 ml). The arbitrary unit was v₀ = 0.3 (E₅₇₈ · min¹).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Generation and initial characterization of the mutant strain. Mutant JY2190 was isolated by serial passage of strain Rx1 in CDM containing decreasing concentrations of EA with each passage. Initially, approximately 10⁶ Rx1 cells were inoculated into 2 ml of CDM containing 200 µg of EA/ml, and the culture was grown for 12 h at 37°C and then diluted 100-fold into CDM containing the same concentration of EA. Following five 12-h passages in CDM containing 200 µg of EA/ml, similar passages were performed in successively lower concentrations of EA, i.e., 20, 2, 0.2, and then 0 µg/ml. The mutant obtained in this manner was capable of growth in CDM without added choline, EA, or other choline analogs, whereas the parent Rx1 was not.

View larger version (24K): [in this window] [in a new window]   FIG. 2.   Growth of parent strain Rx1 and mutant JY2190 in CDM containing no choline (x) or choline at a concentration of 5 µg/ml (), 2 µg/ml (), 1 µg/ml (), or 0.5 µg/ml ().

View larger version (44K): [in this window] [in a new window]   FIG. 3.   Reactivity with S. pneumoniae-specific antibodies. Culture supernatants (S) and cell lysates (L) were reacted with antibodies specific for PspA, autolysin (LytA), C-polysaccharide (C-PS), and phosphocholine (PC). Bacteria were grown in CDM containing choline (5 µg/ml), EA (200 µg/ml), or no analog, as indicated. Lanes: 1, Rx1 plus choline; 2, JY2190 plus choline; 3, JY2190, no choline or analogs; 4, Rx1 plus EA; 5, JY2190 plus EA.

View larger version (90K): [in this window] [in a new window]   FIG. 4.   Photomicrographs of JY2190 grown in CDM. Cultures were grown in the absence of choline or analogs (A), the presence of EA (200 µg/ml) (B), and the presence of a low concentration (5 µg/ml) of choline (C).

View larger version (58K): [in this window] [in a new window]   FIG. 5.   Electron micrographs of Rx1 and JY2190 grown in CDM plus choline or EA. (A) Rx1 plus choline (5 µg/ml) (JY2190 appeared identical and is not shown); (B) JY2190 with no choline or analogs; (C) JY2190 plus EA (200 µg/ml).

Isolation and analysis of LTA and TA. Estimates of phosphate in the cell fractions of JY2190 and Rx1/AL suggested that the cellular contents of LTA and TA were similar and that LTA and TA were present in ratios between 0.06:1 and 0.2:1, respectively.

View this table: [in this window] [in a new window]   TABLE 1.   Compositions of the LTAs and TAs from S. pneumoniae JY2190, Rx1/AL, and R6a

Fingerprint identification of the repeat. In LTA and TA, the glycosidic bond of the -N-acetylgalactosaminyl residue to the ribitol moiety proved extraordinarily susceptible to acid hydrolysis (3, 12, 23), so that depending on the duration of hydrolysis with HF, increasing amounts of the reducing tetrasaccharide and equimolar amounts of ribitol were formed. In this work, we took advantage of this acid lability for a fingerprint identification of the repeats (Fig. 6). The HF hydrolysis products were labeled by precolumn derivatization with 9-fluorenylmethyl-chloroformate at the 4-amino group of the AATGal residue, and the Fmoc derivatives were separated by reverse-phase HPLC. The teichoic acids tested displayed the same fingerprint pattern, which, together with the identical compositions given in Table 1, suggests that all of them contain the same tetrasaccharide ribitol repeat unit. That the sugar sequence, the anomeric bonds, and the glycosidic substitutions are actually identical was shown by nuclear magnetic resonance, fast atom bombardment-mass spectroscopy (MS), and MS-MS analyses (14).

View larger version (25K): [in this window] [in a new window]   FIG. 6.   Hydrolysis of the chains of various LTAs and TAs with HF and identification as Fmoc derivatives by reverse-phase HPLC of the released tetrasaccharide ribitol (single peak) and the anomeric forms of the concomitantly released ,-tetrasaccharide (double peak).

Connection of the repeats. Nonsubstituted TA-containing cell walls from the mutant were hydrolyzed with cellosyl, a muramidase, in order to obtain low-M_r products. The released TA-containing muropeptides were separated from nonsubstituted muropeptides by column chromatography on DEAE-Sephadex A-25 (not shown) and after desalting oxidized with periodate. Although by cleavage with cellosyl oxidizable N-acetylglucosaminyl residues are formed, their contribution to periodate consumption was estimated to be less than 10%. After complete oxidation (6 h), 4.3 mol of IO₄ had been consumed per mol of repeat phosphate, with the concomitant formation of 1.9 mol of formic acid. Since formic acid can arise only from C-3 of a nonsubstituted or O-6-substituted hexopyranosyl moiety and from C-3 of a 1,5-substituted ribitol moiety, these results show that, like in LTA and TA of strain R6, the repeats are connected by phosphodiester bonds between O-5 of the ribitol moieties and O-6 of the adjacent glucosyl residues (Fig. 1). The consumption of 4 mol equivalents of periodate additionally provides evidence that the N-acetylgalactosaminyl residues are not oxidizable and therefore glycosidically substituted at O-4 and/or O-3.

Structure of the glyceroglycolipid anchor and number of repeats in LTA. The glycolipid moieties of the LTA from the mutant and strain Rx1/AL were isolated after hydrolysis by HF (3). On TLC, the native, N-acetylated, and deacylated forms displayed the same mobility as the respective forms of the glycolipid moiety of the LTA from S. pneumoniae R6 (data not shown). Moreover, the Fmoc derivatives of the deacylated glycolipids displayed identical retention times on reverse-phase HPLC. These observations suggest that in all three strains, the LTAs possess Glc(1-3)AATGal(1-3)Glc(1-3)acyl₂Gro as lipid anchor (Fig. 1). For chain length determination, the LTAs were deacylated (0.1 M NaOH, 37°C, 30 min), hydrolyzed in 2 M HCl (100°C, 2.5 h), and analyzed for glycerol. The number of repeats per glycolipid anchor, given by the molar ratio of intrachain phosphate to glycerol (Table 1), was approximately five for the LTA of the mutant, compared to seven and eight for the LTAs of strain Rx1/AL and R6, respectively.

Cell wall composition. As shown in Table 2, there was no significant difference between the cell walls of the mutant and strain Rx1/AL. On a weight basis, the components of TA and peptidoglycan were present in essentially identical amounts. However, the phosphate content in the cell wall of the mutant was half that of strain Rx1/AL, which reflects the nonsubstituted and phosphocholine-substituted repeats in the TA of these strains. The molar ratio of intrachain phosphate to muramic acid-6-phosphate suggests six and eight repeats per chain for the TAs from the mutant and Rx1/AL strains, respectively.

View this table: [in this window] [in a new window]   TABLE 2.   Components of purified cell walls from the mutant JY2190 and strain Rx1/ALa

Behavior of cell wall hydrolases against phosphocholine-free cell walls. N-Acetylmuramyl-L-alanine amidase, the major autolytic enzyme of pneumococci, requires phosphocholine residues in cell wall-linked TA for activity; cell walls containing phosphoethanolamine are not hydrolyzed (22). In accordance with this specific need, the cell walls of the mutant resisted cleavage by the amidase, whereas cell walls of strain Rx1/AL were readily hydrolyzed (Fig. 7). After growth in the presence of choline, the cell walls of the mutant became susceptible to hydrolysis with the amidase, which solubilized more than 80% of the cell wall phosphorus (data not shown).

View larger version (19K): [in this window] [in a new window]   FIG. 7.   Treatment of cell walls from mutant JY2190 () and strain Rx1/AL () with pneumococcal N-acetylmuramyl-L-alanine amidase. Cell walls (1.5 mg in 1 ml of 50 mM HEPES [pH 7.0] containing 0.04% NaN3) were incubated at room temperature with amidase (, 30 µg; , 3 µg). Hydrolysis was monitored by OD578. For preparation of cell walls, the bacteria were grown as indicated in Table 2, footnote a.

View larger version (21K): [in this window] [in a new window]   FIG. 8.   Cleavage of cell walls from mutant JY2190 () and strain Rx1/AL () with CPL (A) and cellosyl (B). Cell walls (4 mg) were incubated with CPL (22 µg) and cellosyl (32 µg) in 1 ml of 20 mM ammonium acetate (pH 6.0) at room temperature. Cleavage of the mutant cell walls (a) was accelerated by doubling the enzyme concentration (b). For overnight incubation, 0.05% (mass/vol) NaN3 was added. Hydrolysis was monitored by OD578. For preparation of cell walls, the bacteria were grown as indicated in Table 2, footnote a.

Penicillin tolerance of the mutant strain. When cells of the mutant strain growing in CDM were treated with 0.01 µg of penicillin per ml during exponential growth phase, they continued growing at the same rate as the untreated control (Fig. 9A). At penicillin concentrations of 0.1 and 1.0 µg/ml, growth ceased but no lysis was seen. When, however, the mutant had been grown for 4 h in the presence of 0.1% choline before penicillin was added, antibiotic concentrations of 0.1 and 1.0 µg/ml led to cell lysis (Fig. 9B). The response to penicillin of the choline-grown mutant was the same as that of the parent strain (Fig. 9C): 0.01 µg of the antibiotic per ml did not affect the growth rate; 0.1 and 1.0 µg/ml evoked immediate cell lysis.

View larger version (20K): [in this window] [in a new window]   FIG. 9.   Response to penicillin of mutant JY2190 (A and B) and of parent strain Rx1 (C). Penicillin was added at time zero. Bacteria were grown in CDM in the absence of choline and analogs (A) and in the presence of 0.1% choline (B and C). Numbers on the graphs are penicillin concentrations (micrograms/milliliter). Growth and lysis were monitored by OD578.

Incorporation of choline. Since some of the preceding observations suggested incorporation of choline into LTA and TA, JY2190 was grown in choline-containing CDM (5 mg/liter), and LTA and TA were isolated as described above. Choline was released from both polymers by HF. Analysis revealed a molar ratio of choline to phosphorus of 0.45, which compares with 0.54 in Rx1/AL. These data let us calculate that in JY2190, on average 82% of the repeats are substituted with one phosphocholine residue, the rest being unsubstituted, whereas in Rx1/AL, on average 83% of the repeats are substituted with one phosphocholine residue and the rest are substituted with two.

Transformation of the mutant in the absence of choline. Using our standard transformation medium (THY-0.2% bovine serum albumin [BSA]-0.01% CaCl₂ [43]), we did not observe transformation of the mutant. However, when JY2190 was grown in CDM containing BSA and CaCl₂, transformation was observed whether or not choline was present in the medium (Table 3). JY2190 transformed at a lower efficiency than Rx1, and peak competence occurred at a lower cell density under both conditions (Table 3). In addition, the mutant grew more slowly and achieved a lower maximum cell density when cultured in the absence of choline in the CDM transformation medium than when cultured in the presence of choline in this medium. In the latter case, growth appeared identical to that of the parent Rx1 (data not shown).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In this report, we describe the isolation of a pneumococcal mutant which has acquired the capability to grow in the absence of choline and analogs. Isolation and analysis of LTA and TA from the mutant grown in CDM revealed both polymers to be free of phosphocholine and other phosphorylated amino alcohols. The proved absence of phosphoethanolamine precludes mutationally induced, endogenous synthesis of EA, which would render growth independent of nutritional choline. The choline-free cell walls of the mutant were totally resistant against the action of pneumococcal autolysin, which has a specific requirement for phosphocholine residues on cell wall-linked TA (22). In vivo, the inactivity of the autolytic enzyme led to impaired cell separation at the end of cell division and rendered the cells resistant to stationary-phase and penicillin-induced lysis. When JY2190 was grown in the presence of choline, LTA and TA were substituted with phosphocholine to about 80% of the phosphocholine substitution in Rx1/AL. The walls of these cells were susceptible to hydrolysis with autolysin, the cells separated normally, penicillin-induced lysis was restored, and PspA was retained at the cell surface. In contrast to penicillin-induced cell lysis, stationary-phase lysis was not restored. Since both kinds of lysis are effected by autolysin, the reason for this distinction may reside in the triggering of lysis rather than in the lysis step itself. Other deviations from the wild type were poor reactivity of cell lysates with monoclonal antibody specific for phosphocholine and the release of larger amounts of material that was reactive with antibody to the TA (C-polysaccharide). For unknown reasons, the material released from JY2190 and Rx1 was only weakly reactive with antibody specific for phosphocholine.

The results of compositional analysis along with the novel fingerprint identification let us propose that the repeats of LTA and TA from the mutant and from strain Rx1/AL are identical to each other and to the structures previously reported for the LTA and TA from S. pneumoniae R6 (3, 12). There was also no change in the linkage of the repeats by phosphodiester bonds between O-5 of the ribitol and O-6 of the adjacent glucosyl moieties. Moreover, the LTAs from the three strains possess the same glycolipid anchor and have similar lengths of between five and seven repeats. The cell wall of the mutant appeared also not to be affected: as indicated by the contents of the constituents, the cellular amounts of peptidoglycan and TA were similar in the mutant and Rx1/AL. The phosphate content of the cell walls of the mutant was half that of strain Rx1/AL, consistent with the absence of choline phosphate from TA. Also similar in both strains were the lengths of TA chains, which varied between six and eight repeats. Since the choline-free LTA and TA of the mutant were in all structural details identical to their choline-containing counterparts in Rx1 and R6, it is unlikely that these structures are altered when nutritional choline is incorporated.

The ability of S. pneumoniae to undergo genetic transformation was lost when choline was replaced in the cell wall by EA (38). That transformability cannot be dependent on choline itself is indicated by the present observation that the mutant retained this capability even when grown in the absence of choline. Compared to Rx1, transformation was less frequent and occurred at an earlier stage of growth, but these differences also existed during growth in the presence of choline.

The metabolic background responsible for the change from a strict requirement for nutritional choline or its analogs to the ability to dispense with such compounds is not yet understood. A hint may come from in vitro experiments that provided evidence that the synthesis of peptidoglycan, which is the basis of cell growth, was inhibited by choline deprivation (9). On the basis of this observation, a metabolic interdependence of TA and peptidoglycan metabolism was suggested: the two polymers may be synthesized in linkage to polyprenol phosphate, and only completed TA, substituted with phosphocholine or phosphoethanolamine, is transferred to peptidoglycan. Accordingly, polyprenol phosphate-linked TA lacking these substituents would not be transferred but would trap polyprenol phosphate and render it unavailable for peptidoglycan synthesis (9). If this hypothesis proves correct, the nutritional requirement of choline may reside in a recognition site for phosphoamino alcohols on TA, by which the activity of the TA-transferase is regulated. Accordingly, the mutation may cause the activity of the transferase to become independent of this regulation, making a transfer of nonsubstituted TA to cell walls possible. Further genetic and biochemical characterizations of the mutant are under way and are expected to provide insights into the mechanisms involved in synthesis of the TAs and peptidoglycan of S. pneumoniae.