Identification and Characterization of a Novel Polysaccharide Deacetylase C (PdaC) from Bacillus subtilis*

Background: Peptidoglycan modification is a very important process that bacteria use to adjust to various environmental conditions. Results: B. subtilis pdaC was associated with lysozyme sensitivity. Surprisingly PdaC is able to deacetylate N-acetylmuramic acid but not N-acetylglucosamine in peptidoglycan. But chitin oligomers were deacetylated by PdaC. Conclusion: PdaC is a unique enzyme exhibiting two different deacetylase activities. Significance: Novel deacetylase is characterized. Cell wall metabolism and cell wall modification are very important processes that bacteria use to adjust to various environmental conditions. One of the main modifications is deacetylation of peptidoglycan. The polysaccharide deacetylase homologue, Bacillus subtilis YjeA (renamed PdaC), was characterized and found to be a unique deacetylase. The pdaC deletion mutant was sensitive to lysozyme treatment, indicating that PdaC acts as a deacetylase. The purified recombinant and truncated PdaC from Escherichia coli deacetylated B. subtilis peptidoglycan and its polymer, (-GlcNAc-MurNAc[-l-Ala-d-Glu]-)n. Surprisingly, RP-HPLC and ESI-MS/MS analyses showed that the enzyme deacetylates N-acetylmuramic acid (MurNAc) not GlcNAc from the polymer. Contrary to Streptococcus pneumoniae PgdA, which shows high amino acid sequence similarity with PdaC and is a zinc-dependent GlcNAc deacetylase toward peptidoglycan, there was less dependence on zinc ion for deacetylation of peptidoglycan by PdaC than other metal ions (Mn2+, Mg2+, Ca2+). The kinetic values of the activity toward B. subtilis peptidoglycan were Km = 4.8 mm and kcat = 0.32 s−1. PdaC also deacetylated N-acetylglucosamine (GlcNAc) oligomers with a Km = 12.3 mm and kcat = 0.24 s−1 toward GlcNAc4. Therefore, PdaC has GlcNAc deacetylase activity toward GlcNAc oligomers and MurNAc deacetylase activity toward B. subtilis peptidoglycan.

acts as a protector toward various environmental conditions. Peptidoglycan can be modified by several cell wall hydrolases and deacetylases (2).
Only one MurNAc deacetylase (Bacilllus subtilis PdaA) has been characterized (10). Previously, our group demonstrated that this deacetylase acts as a MurNAc deacetylase toward spore peptidoglycan to produce a muramic-␦-lactam structure in vivo (10), and also that it is active toward a GlcNAc-MurNAc polymer in vitro (11). Blair and van Aalten (12) published the crystal structure of PdaA. However, no other MurNAc deacetylases have been identified. Therefore, we have further investigated deacetylase activity to better understand the processes of GlcNAc and/or MurNAc deacetylation.
We have identified a unique polysaccharide deacetylase gene, yjeA (renamed pdaC). It is known that this gene is regulated by an essential two-component system, YycFG (13), which is associated with cell division (14 -16). We demonstrated that the pdaC mutant is sensitive to lysozyme treatment and that PdaC acts as a GlcNAc deacetylase toward chitin oligomers and as a MurNAc deacetylase toward B. subtilis peptidoglycan.

EXPERIMENTAL PROCEDURES
Construction of Deacetylase Mutants-The strains, plasmids, and primers used in this study are shown in supplemental Tables S1 and S2. The pdaC gene fragment was amplified by PCR with YJEA-SD and YJEA-RV primers, digested with SalI and MunI, and then ligated to the SalI-EcoRI site of pBlue-scriptII SK(ϩ), resulting in pBLJEA. A plasmid, pDG1515, was digested with XbaI and SalI, and the fragment containing a tetracycline cassette was blunted and then ligated to the EcoRV site of pBluescriptII SK(ϩ), resulting in pBLTC. The plasmid, pBLTC, was digested with HindIII and the fragment containing the tetracycline cassette was ligated into the HindIII site of pBLJEA, resulting in pBLJEATC.
The yheN gene fragment was amplified by PCR with YHEN-SD and YHEN-RV primers, digested with BamHI and HindIII, and then ligated to the corresponding sites of pGEM3Zf(ϩ), resulting in pGMHEN. A plasmid, pBCATERV, containing a chloramphenicol cassette, was digested with SmaI and HincII, and the fragment containing the cassette was ligated in the Eco47III site of pGMHEN, resulting in pGMHENCM.
The yxkH gene fragment amplified by PCR with YXKH-SD and YXKH-RV primers was blunted and phosphorylated with a BKL kit (Takara) and then ligated to the EcoRV site of pBlue-scriptII SK(ϩ), resulting in pBLXKH. To eliminate the ClaI site of pBLXKH, the plasmid was digested with HindIII and HincII, blunted, and then self-ligated, resulting in pBL⌬XKH. A plasmid, pDG792 containing a kanamycin cassette, was digested with ClaI and the fragment containing the cassette was ligated to the corresponding site of pBL⌬XKH, resulting in pBL⌬XKHKM. B. subtilis 168 was transformed with the linearized pBLJEATC, pGMHENCM, and pBL⌬XKHKM, resulting in JEATdd (pdaC::tet), HENCdd (yheN::cat), and XKHKdd (yxkH::km) strains, respectively.
The truncated ylxY gene fragment amplified by PCR with YLXY-HF and YLXY-BR primers was digested with HindIII and BamHI and ligated to the corresponding sites of pMUTIN4, resulting in pM4⌬LXY. B. subtilis 168 was transformed with the concatenated pM4⌬LXY derived from Escherichia coli C600 (pM4⌬LXY) by single crossing over recombination, resulting in the YLXYd strain (ylxY::erm).
Construction of a Conditional Null Mutant of pdaC, PdaCp-A fragment containing the SD sequence of pdaC and truncated pdaC was amplified by PCR with PC-SD-HinF and PC-D-BamR primers. The amplified fragment was digested with HindIII and BamHI and ligated to the corresponding sites of pMUTIN4, resulting in pM4SD-PdaC. After the concatenated pM4SD-PdaC derived from E. coli C600 (pM4SD-PdaC) strain was purified, B. subtilis 168 was transformed with the plasmid by single crossing over recombination, resulting in the PdaCp strain (P spac -pdaC). PdaC was expressed in the strain by isopropyl ␤-D-thiogalactopyranoside (IPTG) 3 addition.
Construction of a Plasmid, pQE30⌬YjeA for Overexpressing Truncated PdaC-The truncated pdaC gene fragment amplified by PCR with yjeAϩ87F and yjeAϩ1383R primers was digested with SphI and SalI, and then ligated to the corresponding sites of pQE-30, resulting in pQE30⌬YjeA. This plasmid was utilized to overexpress truncated PdaC (from amino acids 30 to 467).
Transformation of E. coli and B. subtilis-E. coli and B. subtilis transformations were performed as described by Sambrook et al. (17) and Anagnostopoulos and Spizizen (18), respectively.
SDS-PAGE and Purification of Cell Wall and Peptidoglycan from B. subtilis-SDS-PAGE was performed as described by Sambrook et al. (17). B. subtilis was grown at early stationary phase, and the cells were utilized for preparation of cell wall and peptidoglycan as described previously (19 -21). Purified peptidoglycan was N-acetylated as described previously (11).
Overexpression and Purification of Truncated PdaC-Truncated PdaC (30 -467 amino acids) was overexpressed in E. coli JM109 harboring pQE30⌬YjeA. The strain was incubated at 37°C in LB medium containing 100 g/ml of ampicillin. When the absorbance reached 0.5 at 600 nm, 1 mM IPTG (final concentration) was added to the culture, and then the cells were further incubated for 1 h. The cells were harvested and suspended in 10 mM imidazole NPB buffer (10 mM imidazole, 1 M NaCl, 20 mM sodium phosphate (pH 7.4)), and then disrupted by sonication. After centrifugation, the supernatant was utilized for purification of the protein with a HiTrap Chelating HP column (GE Healthcare), according to the manufacturer's instructions. The purified truncated PdaC was dialyzed against 20 mM phosphate buffer (pH 7.4).
For the purpose of kinetics for peptidoglycan, the molarity was calculated based on the assumption that peptidoglycan consists of a repeating unit of GlcNAc-MurNAc-L-Ala-D-Glu-A 2 pm (M r 868.8), because that is the main unit of B. subtilis peptidoglycan (24). As a result, we defined that 2.5 mg/ml of peptidoglycan is equal to 2.88 mM. For kinetics analysis of GlcNAc 4 , 2 mg/ml of GlcNAc 4 (2.41 mM) was used. For the assay 188.7 nM PdaC was also utilized.
For the other substrates derived from B. subtilis peptidoglycan, peptidoglycan (final concentration, 2.5 mg/ml) was suspended in 50 mM HEPES buffer (pH 7.0) and then digested without or with 12.5 g/ml (final concentration) of cell wall hydrolases, an L-alanine amidase (CwlH) (22), LD-endopeptidase (CwlK) (23), and DL-endopeptidase (LytF (CwlE)) (20) at 37°C overnight. After the sample had been boiled for 10 min to denature the hydrolases and centrifuged, the supernatant was collected. Ten g/ml of PdaC (final concentration) and 50 mM HEPES buffer (pH 7.0) containing 5 mM MnCl 2 were added to the supernatant, followed by incubation for 4 h at 37°C.
For determination of a divalent cation effect, PdaC was dialyzed against 20 mM EDTA (pH 7.4) twice and then against 20 mM sodium phosphate buffer (pH 7.4) ("EDTA-treated PdaC"). On the other hand, PdaC was dialyzed against 20 mM sodium phosphate buffer (pH 7.4) (no EDTA treatment; "Native PdaC"). Two mg/ml of GlcNAc 4 or 2.5 mg/ml of peptidoglycan (final concentration) in 50 mM HEPES (pH 7.0) containing 5 mM MnCl 2 , ZnCl 2 , MgCl 2 , NiSO 4 , or CaCl 2 , 5 M ZnCl 2 , or 50 nM ZnCl 2 was digested by 10 g/ml of PdaC (final concentration) for 1, 2, 3, and 4 h at 37°C. Five mM CoCl 2 was also utilized for analysis of deacetylase activity with an F-kit. However, Co 2ϩ interfered with the assay kit (drastic increase in absorbance without any acetic acid because of changing solution color from crystal clear to yellow during the assay), thus the released acetic acid was not able to be determined in the presence of Co 2ϩ . After the reaction by PdaC had been performed, all samples were boiled for 10 min and centrifuged, and then the supernatant was utilized for measurement of deacetylase activity with an F-kit.
Purification of Glycan Strands Consisting of (-GlcNAc-MurNAc[-L-Ala-D-Glu]-) n Polymer-Purification of glycan strands from B. subtilis peptidoglycan was performed as described previously (11). After 1 mg of peptidoglycan had been digested with 5 g of LytF (CwlE) in 20 mM HEPES buffer (pH 7.0) at 37°C for 14 h and then boiled for 10 min, it was centrifuged and the supernatant was collected. The supernatant components (containing the glycan strands and peptides) were separated by size exclusion chromatography. After the purified glycan strands (-GlcNAc-MurNAc[-L-Ala-D-Glu]-) n had been freeze-dried, they were N-acetylated as described previously (11).

Deacetylation of Glycan Strands Containing L-Ala-D-Glu Side Chains (-GlcNAc-MurNAc[-L-Ala-D-Glu]-) n with PdaC-
Two mg of purified glycan strands containing L-Ala-D-Glu side chains were deacetylated at 37°C overnight with (deacetylated sample) or without (non-deacetylated sample) 20 g of PdaC in 50 mM HEPES buffer (pH 7.0) containing 5 mM MnCl 2 . To further deacetylate the strands, 20 g of PdaC was added to the deacetylated sample, and then incubation was performed at 37°C for 4 h. The samples were digested by 10 g of the N-terminal domain of CwlT, which is a muramidase (25), at 37°C for 4 h, followed by boiling. Borate buffer (pH 9.0) (final concen-tration, 0.5 M) was added to the samples, and then the reducing ends of MurNAc were reduced with NaBH 4 as described previously (11). The reduced samples were separated by RP-HPLC with a Symmetry Shield RP18 column (Waters) (flow rate, 1 ml/min; monitoring wavelength, 202 nm; column oven temperature, 40°C; Shimadzu LC-10AD HPLC system). Elution buffer A contained 0.05% TFA, and buffer B contained 0.05% TFA with 40% CH 3 CN. Elution was performed for 10 min with buffer A only with an isocratic gradient and then for 60 min with a linear gradient of buffer B (from 0 to 50%).
Purification of Reduced Tetrasaccharide Dipeptides (4S2P)-To purify reduced tetrasaccharide dipeptide, GlcNAc-MurNAc(-L-Ala-D-Glu)-GlcNAc-MurNAcr(-L-Ala-D-Glu), purified glycan strands containing L-Ala-D-Glu side chains were digested with a muramidase, CwlT, and then the sample was reduced and separated by RP-HPLC as described above. Determination of DNase Activity of PdaC-Five-hundred ng of pBR322 plasmid DNA was digested with 2 g of PdaC or DNase I (TaKaRa) in reaction buffer (50 mM Tris-HCl (pH 7.5), 100 mM NaCl, 10 mM MgCl 2 ) at 37°C for 1, 5, 10, 20, 30, 60, 90, and 120 min. After the reaction was complete, the sample was boiled for 10 min to inactivate the enzyme. The sample was centrifuged and then the supernatant was applied to a 0.7% agarose gel.

B. subtilis pdaC Mutant Is Sensitive toward Lysozyme-Many microorganisms have polysaccharide deacetylases such as
Streptococcus pneumoniae PgdA (6, 7) and B. subtilis PdaA (10,11,26). Fig. 1 shows identified or predicted polysaccharide deacetylases in B. subtilis and some other Gram-positive bacteria. It is known that PdaA and PdaB play roles in sporulation (10,28,29) but the other gene products, YjeA (renamed PdaC), YlxY, YxkH, and YheN, are not associated with sporulation and germination (28). Thus, these disruptants were created, and the sensitivity of these strains toward lysozyme (muramidase that digests a linkage of MurNAc-GlcNAc in peptidoglycan) was determined. As shown in Fig. 2, only the pdaC mutant was sensitive toward lysozyme because the growth rate was decreased (closed circles in Fig. 2). To confirm the pdaC mutant phenotype, a new strain, PdaCp (P spac -pdaC) was constructed.
When the strain was incubated without IPTG (weak PdaC expression), it was sensitive toward lysozyme (closed squares in supplemental Fig. S1). The strain with 1 mM IPTG (PdaC expression) was resistant toward lysozyme (closed diamonds in supplemental Fig. S1), suggesting that PdaC is necessary for protection of peptidoglycan toward lysozyme because lysozyme cannot digest deacetylated peptidoglycan (30).
PdaC Deacetylates B. subtilis Peptidoglycan-Because the pdaC deletion mutant was sensitive toward lysozyme (Fig. 2), it was predicted that PdaC deacetylates B. subtilis peptidoglycan. From the SOSUI algorithm that can be used to predict the membrane regions of a target protein, PdaC seems to have a transmembrane region (from amino acids 7 to 29). Moreover, ϳ50 kDa PdaC is detected from the membrane fraction by proteome analysis described by Eymann et al. (31). Therefore, the truncated PdaC (from amino acids 30 to 467) lacking the transmembrane region was overexpressed in E. coli and then purified by affinity chromatography. As shown in lane 3 in supplemental Fig. S2, the truncated PdaC (53.0 kDa) could be purified (one band on SDS-PAGE).
The deacetylase activity of PdaC toward peptidoglycan was measured by released acetic acid. As shown in Table 1 and   supplemental Fig. S3, PdaC dialyzed against EDTA weakly deacetylated the peptidoglycan without addition of a divalent cation ("EDTA-treated PdaC"). Moreover, PdaC had the maximum activity with Mn 2ϩ . Therefore, these results suggest that PdaC in the presence of Mn 2ϩ efficiently deacetylates B. subtilis peptidoglycan.

PdaC Works as a MurNAc Deacetylase toward Glycan
Strands Containing L-Ala-D-Glu-As shown in Table 2, PdaC deacetylates glycan strands consisting of (-GlcNAc-MurNAc[-L-Ala-D-Glu]-) n . The purified glycan strands were deacetylated with or without PdaC, and then the samples were treated with a muramidase, followed by reduction and separation of the sample by RP-HPLC. In Fig. 3A, the without PdaC sample contained two major peaks (1 and 3). Both materials were collected and analyzed by ESI-MS. The peak 1 material in the without PdaC sample showed fragment ions at m/z 721.5 and 697.3 in the positive and negative modes, respectively (supplemental Fig. S4, A and B, respectively) Table S3).
The with PdaC sample contained not only two materials (peaks 1 and 3), but also additional material (peak 2) (Fig. 3A).  Fig. S6, A and B, respectively). The peak 2 material was further analyzed by ESI-MS/MS in the positive and negative modes. As shown in Fig. 4, A and B PdaC Also Deacetylates Reduced 4S2P-After glycan strands containing L-Ala-D-Glu had been treated with PdaC followed by muramidase digestion, the main product was deacetylated
Thus, PdaC was able to deacetylate 4S2P derived from B. subtilis peptidoglycan. Fig. S7, the amino sugar deacetylated by PdaC seems to be MurNAc. To confirm this result, reduced 4S2P was deacetylated by PdaC (this sample is exactly the same as the with PdaC sample in Fig. 3B) and digested with mutanolysin, which can digest glycosidic linkages such as -GlcN-MurNAc2GlcNAc-MurNAc-, -GlcNAc-MurNAc2GlcN-MurNAc-and -GlcN-MurNAc2GlcN-MurNAc-, but not -GlcNAc-Mur-GlcNAc-MurNAc- (26,27). After the sample had been reduced, it was separated by RP-HPLC. As shown in Fig. 3C, the peak 3 material (reduced 4S2P) was completely digested to 2S1P and reduced 2S1P, followed by NaBH 4 reduction to yield only reduced 2S1P (the peak 1 material), indicating the mutanolysin works very well in this reaction. However, the peak 2 material (deacetylated compound of reduced 4S2P) was not digested. Therefore, this suggested that the deacetylated product is was subjected to mutanolysin, it was not digested (data not shown). Therefore, these results strongly suggest that PdaC is a MurNAc deacetylase.
The deacetylase activities of PdaC toward GlcNAc, several chitin oligomers (GlcNAc 2-5 ), chitin (GlcNAc polymer), and chitosan (partially deacetylated GlcNAc polymer) were measured by quantitating the released acetic acid. As shown in Fig.  5, PdaC displayed deacetylase activities toward GlcNAc 4 and GlcNAc 5 , whereas it did not deacetylate GlcNAc and had reduced activities toward GlcNAc 2 and GlcNAc 3 . PdaC was essentially not active toward chitosan and had very low activity toward chitin (Fig. 5). These results suggest that PdaC deacetylates chitin oligomers (GlcNAc 4,5 ) as an N-acetylglucosamine deacetylase.
The deacetylase activity of PdaC toward GlcNAc 4 (a suitable substrate for PdaC) with divalent cations was measured by released acetic acid. As shown in Table 3 and supplemental Fig.  S8, PdaC showed the maximum activity with Mn 2ϩ . Therefore, the result suggests that PdaC in the presence of Mn 2ϩ efficiently deacetylates not only B. subtilis peptidoglycan but also the GlcNAc oligomer, GlcNAc 4 .
To identify the position of deacetylated GlcNAc by PdaC toward the chitin oligomer, GlcNAc 4 was deacetylated. Then the samples were separated by normal-phase HPLC with a TSKgel Amide-80 column to remove salts and buffer, followed by collection of the peaks containing the tetrasaccharide materials (data not shown). These peaks were further separated by RP-HPLC. As shown in Fig. 6, A and B, the non-deacetylated sample contained two major peaks (peaks 2 and 3 in Fig. 6B) and the deacetylated sample contained an additional peak (peak 1 in  Fig. S6, A and B) and ESI-MS/MS analyses (see Fig. 4, A and B). B, the reduced substrate, tetrasaccharide dipeptides (peak 3 material) was deacetylated with PdaC or without PdaC, followed by separation by RP-HPLC. Control sample contains PdaC only. The peak 3 materials in the with PdaC and without PdaC are the same although their retention times are slightly different. The produced peak 2 material in the with PdaC sample was identified as GlcNAc-Mur(-L-Ala-D-Glu)-GlcNAc-MurNAcr(-L-Ala-D-Glu) by ESI-MS and ESI-MS/MS analyses in the negative mode (supplemental Fig. S7, A and B). C, RP-HPLC of the reduced fragments of mutanolysin digests. Deacetylated or non-deacetylated reduced-4S2P material was digested with a muramidase, mutanolysin, followed by reduction with NaBH 4 . Peak 1, reduced 2S1P; peak 2, GlcNAc-Mur(-L-Ala-D-Glu)-GlcNAc-MurNAcr(-L-Ala-D-Glu). The arrow indicates the elution position of reduced 4S2P. Fig. 6A), in addition to peaks 2 and 3. Peaks 2 and 3 in Fig. 6, A and B, had a fragment ion at m/z 832.1 by ESI-MS in the positive mode, corresponding to [M ϩ H] ϩ of GlcNAc 4 (M r 830.5) (data not shown), and ESI-MS/MS analysis indicated that the sample was actually GlcNAc 4 (data not shown). The reason for the different elution times (peaks 2 and 3) is unknown, but it may reflect the difference of ␣ and ␤ anomers at the reducing end as described previously (32). Peak 1 in Fig. 6A had a fragment ion at m/z 790.0 by ESI-MS in the positive mode, corresponding to [M ϩ H] ϩ of deacetylated GlcNAc 4 (M r 788.5) (Fig. 6C). Moreover, as shown in Fig. 6D, the substrate was identified by ESI-MS/MS as GlcNAc-GlcNAc-GlcN-GlcNAc (Fig. 6E). To confirm the structure, the peak 1 material was reduced, and then analyzed by ESI-MS and ESI-MS/MS. The reduced material had a fragment ion at m/z 792.1 by ESI-MS in the positive mode,

. ESI-MS and ESI-MS/MS analyses of GlcNAc-Mur(-L-Ala-D-Glu)-GlcNAc-MurNAcr(-L-Ala-D-Glu).
A and B, peak 2 material was analyzed by ESI-MS (supplemental Fig. S6, A and B) and ESI-MS/MS in positive and negative modes (panels A and B, respectively). The fragment peaks b2, b3, and y2 correspond to the structure of the deacetylated compound of reduced 4S2P (panel C). C, identified structure and molecular weight of each fragment in the peak 2 material in Fig. 3A.  Fig. S9A), and the fragments (b2Ј and b3Ј in supplemental Fig. S9B) by ESI-MS/MS strongly support that the structure of the reduced compound of deacetylated GlcNAc 4 is GlcNAc-GlcNAc-GlcN-GlcNAcr (Fig. 6E).
MurNAc and GlcNAc Deacetylation by PdaC-PdaC has both MurNAc and GlcNAc deacetylase activities; however, Fig.  5 shows that the activity toward MurNAc seems to be higher than toward GlcNAc, because the released acetic acid from peptidoglycan and GlcNAc oligomers are different (released acetic acid, 83.4 M from peptidoglycan and 23.1 M from GlcNAc 4 ). To determine the kinetics for GlcNAc and MurNAc deacetylation, GlcNAc 4 and peptidoglycan were utilized as substrates, and the released acetic acid was measured. For the purpose of the kinetics for peptidoglycan, the molarity was calculated using the assumption that peptidoglycan consists of a repeating unit, GlcNAc-MurNAc-L-Ala-D-Glu-A 2 pm (M r 868.8), because this is the main unit contained in B. subtilis peptidoglycan (24). Initial velocities for GlcNAc 4 and peptidoglycan were measured after a 30-min incubation at different substrate concentrations because the rates of released acetic acid from both substrates were constant within at least 30 min (Fig. 7, A and B), and the measured initial velocities fit Michaelis-Menten kinetics. The K m for GlcNAc 4 and peptidoglycan were 12.3 Ϯ 1.84 and 4.8 Ϯ 0.30 mM, respectively, and the values of k cat for GlcNAc 4 and peptidoglycan were 0.24 Ϯ 0.031 and 0.32 Ϯ 0.010 s Ϫ1 , respectively (Fig. 7C). The value of K m /k cat for peptidoglycan (0.067 mM Ϫ1 s Ϫ1 ) is much larger than the value for GlcNAc 4 (0.020 mM Ϫ1 s Ϫ1 ), and non-treated peptidoglycan was utilized instead of peptidoglycan treated with DL-endopeptidase, although the treated peptidoglycan (-GlcNAc-MurNAc[-L-Ala-D-Glu]-) is a more suitable substrate for PdaC (Table 2). Thus, PdaC seems to prefer to deacetylate MurNAc from peptidoglycan. Moreover, because GlcNAc 3 deacetylation by S. pneumoniae PgdA had values of K m ϭ 3.8 mM, k cat ϭ 0.55 s Ϫ1 , and k cat /K m ϭ 0.15 mM Ϫ1 s Ϫ1 (7), the deacetylation toward GlcNAc by PdaC was much weaker than that by PgdA.
No Nuclease Activity of PdaC-Ng et al. (33) described that PdaC (YjeA) has DNase activity. Our results indicate that PdaC is both a GlcNAc and MurNAc deacetylase. Thus, DNase activity by PdaC was determined under the same conditions as described by Ng et al. (33). Results indicated that PdaC had no DNase activity at all (supplemental Fig. S10A). This may be because Ng et al. (33) overexpressed the enzyme in E. coli and purified the protein from the culture (not cytoplasm). As well, the region of PdaC expressed by Ng et al. (33) (from amino acid 50 to the end) was slightly different from the region in this study (from amino acid 30 to the end). Another reason may be that the longer incubation of DNA with the crude enzyme solution described by Ng et al. (33) may easily lead to DNA digestion.
Interestingly, divalent cations strongly affected MurNAc deacetylase (Table 1 and supplemental Fig. S3). PdaC dialyzed against EDTA showed weak deacetylase activity (15.1 M). On the other hand, PdaC with 5 mM Mn 2ϩ displayed strong activity (92.4 M). Because divalent cations are not required for GlcNAc 4 deacetylation, the deacetylation by PdaC toward peptidoglycan seems to be different from that toward the GlcNAc oligomer.
PdaC had no deacetylase activity toward the -GlcNAc-MurNAc-polymer ( Table 2). The difference between GlcNAc and MurNAc is the C3 position (MurNAc has a propionic acid group instead of a hydroxyl group). Moreover, when the substrate had peptide side chains (L-Ala-D-Glu) covalently bound to the glycan strand, the deacetylase activity by PdaC was increased compared with the substrate with no peptide side chains (Table 2). Thus, the C3 position of the amino sugar is very important for the deacetylation reaction and/or substrate recognition by PdaC.
It has been reported that two of peptidoglycan modifications, O-acetylation on the C-6 hydroxyl moiety of MurNAc residues and N-deacetylation of GlcNAc and MurNAc residues, are involved in lysozyme resistance of peptidoglycan (2,6,39). Because the pdaC mutant is sensitive to lysozyme treatment (Fig. 2) and PdaC is a deacetylase (Figs. 3-7), PdaC deacetylates MurNAc and/or GlcNAc in vivo. Atrih et al. (24) reported that B. subtilis 168 peptidoglycan contained deacetylated residues of GlcNAc at 17.3% (per muropeptides prepared from peptidoglycan digested with a muramidase). In contrast, Zipperle et al. (40) reported that B. subtilis peptidoglycan contained both deacetylated muramic acid residues (33% per total muramic acid) and deacetylated glucosamine residues (19% per total glucosamine). Muramic acid residues have also been identified in peptidoglycans from Bacillus cereus, Bacillus anthracis, and Bacillus thuringiensis (40), and S. pneumoniae (6). Because PdaC is found in the B. subtilis membrane by proteome analysis (31) and is not detected in the cell surface and extracellular fractions (41,42), at least PdaC seems to work on the membrane as a MurNAc deacetylase.
Structural Comparison among PdaC, PdaA, and PgdA-The polysaccharide deacetylase domains of PdaC and PdaA in B. subtilis and PgdA in S. pneumoniae R6 are similar (Fig. 1). Moreover, the other domains of PdaC are conserved in PgdA (Fig. 1) although these domains have unknown function(s). The predicted secondary and tertiary structures of PdaC were created using the ESyPred3D algorithm. The results indicated that the secondary and tertiary structures of PdaC were very similar to PdaA and PgdA (supplemental Fig. S11 and data not shown). The catalytic residues in PgdA (7) are conserved in PdaC and PdaA (in supplemental Fig. S11). Blair et al. (7) described that the Asp-276, His-326, and His-330 in PgdA interact with a zinc ligand, and a zinc ion is necessary for the deacetylase activity of PgdA. These amino acid residues are also conserved in B. subtilis PdaC and PdaA (closed triangles in supplemental Fig. S11). Very recently S. mutants PgdA (homologue of S. pneumoniae) was reported to only digest a hexamer of N-acetylglucosamine, but not peptidoglycan, and required zinc ions for its activity (37). Therefore, the various deacetylase activities are regulated by ion type and substrate although their amino acid similarities are high.
It has also been elucidated that several pathogenic bacteria contain deacetylases, although the enzymes do not contain the three-domain structure. Severin et al. (46) demonstrated that a pathogenic bacterium, B. cereus has the deacetylated peptidoglycan. Psylinakis et al. (38) described that the BC_1960 and BC_3618 proteins in B. cereus are peptidoglycan GlcNAc deacetylases, and these enzymes contribute to increasing resistance to lysozyme digestion. Interestingly, the Gram-negative pathogenic bacterium, Helicobacter pylori, induces a peptidoglycan deacetylase, HP310 (PgdA) under oxidative stress conditions (47) and the deacetylation contributes to the bacterial survival by mitigating host immune responses (48). However, no MurNAc deacetylase activities have been characterized in these bacteria. This paper describes for the first time a unique enzyme that has both GlcNAc and MurNAc deacetylase activities. Moreover, this is only the second MurNAc deacetylase to be reported aside from B. subtilis PdaA, which exhibits very narrow substrate specificity.