Functional analysis of Clostridium difficile sortase B reveals key residues for catalytic activity and substrate specificity

Most of Gram-positive bacteria anchor surface proteins to the peptidoglycan cell wall by sortase, a cysteine transpeptidase that targets proteins displaying a cell wall sorting signal. Unlike other bacteria, Clostridium difficile, the major human pathogen responsible for antibiotic-associated diarrhea, has only a single functional sortase (SrtB). Sortase's vital importance in bacterial virulence has been long recognized, and C. difficile sortase B (Cd-SrtB) has become an attractive therapeutic target for managing C. difficile infection. A better understanding of the molecular activity of Cd-SrtB may help spur the development of effective agents against C. difficile infection. In this study, using site-directed mutagenesis, biochemical and biophysical tools, LC-MS/MS, and crystallographic analyses, we identified key residues essential for Cd-SrtB catalysis and substrate recognition. To the best of our knowledge, we report the first evidence that a conserved serine residue near the active site participates in the catalytic activity of Cd-SrtB and also SrtB from Staphylococcus aureus. The serine residue indispensable for SrtB activity may be involved in stabilizing a thioacyl-enzyme intermediate because it is neither a nucleophilic residue nor a substrate-interacting residue, based on the LC-MS/MS data and available structural models of SrtB–substrate complexes. Furthermore, we also demonstrated that residues 163–168 located on the β6/β7 loop of Cd-SrtB dominate specific recognition of the peptide substrate PPKTG. The results of this work reveal key residues with roles in catalysis and substrate specificity of Cd-SrtB.

Most of Gram-positive bacteria anchor surface proteins to the peptidoglycan cell wall by sortase, a cysteine transpeptidase that targets proteins displaying a cell wall sorting signal. Unlike other bacteria, Clostridium difficile, the major human pathogen responsible for antibiotic-associated diarrhea, has only a single functional sortase (SrtB). Sortase's vital importance in bacterial virulence has been long recognized, and C. difficile sortase B (Cd-SrtB) has become an attractive therapeutic target for managing C. difficile infection. A better understanding of the molecular activity of Cd-SrtB may help spur the development of effective agents against C. difficile infection. In this study, using site-directed mutagenesis, biochemical and biophysical tools, LC-MS/MS, and crystallographic analyses, we identified key residues essential for Cd-SrtB catalysis and substrate recognition. To the best of our knowledge, we report the first evidence that a conserved serine residue near the active site participates in the catalytic activity of Cd-SrtB and also SrtB from Staphylococcus aureus. The serine residue indispensable for SrtB activity may be involved in stabilizing a thioacyl-enzyme intermediate because it is neither a nucleophilic residue nor a substrateinteracting residue, based on the LC-MS/MS data and available structural models of SrtB-substrate complexes. Furthermore, we also demonstrated that residues 163-168 located on the ␤6/␤7 loop of Cd-SrtB dominate specific recognition of the peptide substrate PPKTG. The results of this work reveal key residues with roles in catalysis and substrate specificity of Cd-SrtB.
Clostridium difficile infection (CDI) 3 is a global healthcare problem associated with morbidity and mortality for hospitalized patients (1,2). CDI frequently occurs in older and severely ill patients who are in long-term care facilities (2). The mortality rate of CDI patients ranges from 5 to 40% (3,4), and the rate of recurrence occurring within 30 days after treatment is to to 30% (5). C. difficile is a Gram-positive, spore-forming, and anaerobic bacterium (6), which is the causing agent of a multitude of intestinal diseases ranging from mild diarrhea to severe inflammatory bowel perforations or pseudomembranous colitis (7,8). CDI is transmitted through bacterial spores or from person to person by the fecal-oral route. Patients infected with C. difficile spores are mostly by direct contact with contaminated surfaces and symptomatic patients in the hospital setting (9). Current treatment of CDI mainly relies on the administration of antibiotics such as metronidazole, vancomycin, and fidaxomicin to alleviate immediate symptoms for patients (2,10,11). Furthermore, alternative treatment options are considered, such as fecal microbiota transplantation as a means of re-establishing a normal microbiota profile for patients with recurrent CDI (12). Nevertheless, fecal microbiota transplantation is still not widely applied, and broad-spectrum antibiotic therapy remains the first choice in managing CDI (6). However, antibiotic use is a major risk factor for recurrent CDI and C. difficile superinfection (5), because of the disruption of the normal gut microbiota (6). The incidence rate of multiple recurrent CDI and frequency of antimicrobial treatment failures have significantly increased (13). Moreover, C. difficile 630 is a multidrug-resistant strain whose genome was isolated from a patient with pseudomembranous colitis (14). Given that the prevalence of antibiotic-resistant bacteria is rising, nonconventional antimicrobial therapies are in demand, and efforts searching for developing alternative anti-infective drugs for the treatment of CDI patients are growing (15).
For Gram-positive bacteria, the attachment of virulence-associated surface proteins to the peptidoglycan cell wall is mediated by sortase enzymes (16). Sortases are cysteine transpeptidases that function in covalently anchoring of surface proteins to the cell wall envelopment (17) and in constructing pili (18,19). Based on the primary sequences and biological roles, sortases are classified into six classes (A-F) (20 -22). All the characterized sortases possess a signal sequence that enables their translocation across the membrane via the Sec apparatus and target proteins consisting of a cell wall sorting signal in the C-terminal region (23, 24). The characteristic five-residue sortase-recognition sequence motif located within cell wall sorting signal of substrate proteins is class-and/or bacteria-specific. Class A sortase enzymes anchor many surface proteins in cell wall and play a housekeeping role. Sortase A from Staphylococcus aureus (Sa-SrtA), the best studied sortase, recognizes the LPXTG motif of its substrates and initiates catalysis by employing the thiolate of the active site cysteine residue to cleave the peptide bond between Thr and Gly residues (25). This process results in generation of an acyl-enzyme intermediate by a thioester linkage between the cysteine residue of sortase and its substrate (26,27). Subsequently, a secondary substrate (lipid II or pilin) is recognized by sortase that catalyzes a reaction in which the amine group from lipid II or a lysine residue within a pilin subunit nucleophilic attacks the thioacyl bond and relieves the sortase-protein thioacyl intermediate (23). Class B sortase can perform distinct tasks including heme uptake and pilus polymerization. In contrast to many other bacteria that typically have multiple sortases, C. difficile possesses only sortase B (Cd-SrtB), which attaches seven proteins to cell wall and appears to play a general role (28). Cd-SrtB recognizes a sorting signal containing a (S/P)PXTG motif (29) that differs from the conserved class A sortase-recognition LPXTG motif and the S. aureus sortase B (Sa-SrtB) NP(Q/K)TN sorting motif (30,31). The molecular origin of how Cd-SrtB discriminates (S/P)PXTG from LPXTG or NP(Q/K)TN is not well-understood.
Structural and computational studies have provided in-depth insights into the molecular mechanism of sortase-mediated catalysis and have advanced our understanding on the complicated process of substrate recognition (32)(33)(34). Structures of the catalytic domains of sortases share a conserved eight-stranded ␤-barrel core harboring a His-Cys-Arg triad essential for catalysis (23, 33). Characteristic structural features and variations among different classes of sortases are observed, and the structural differences dictating the class-specific function and substrate specificity are described (23, 24). The key catalytic residues His, Cys, and Arg are structurally equivalent in the family of sortase and are in proximity to one another within the active site located at the edge of ␤-barrel. The His residue functions as a general acid/base during acyl and deacyl process (35,36); the Arg residue is believed to play an important role in stabilizing the acyl-enzyme intermediate by forming an oxyanion hole (21,33,36,37). Although the residues that constitute the active site are believed to be His-Cys-Arg, studies also reported that other residues also participate in catalysis (23, 38, 39). The crystal structure of Sa-SrtB revealed that Asp-196 also constitutes the catalytic site, similar to the catalytic His-Cys-Asp triad of Cys/Ser protease (39). It remains to be explored whether other residues located near the active site also contribute to catalytic activity.
Crystallographic structures of Cd-SrtB determined by our group (40) and Chambers et al. (41) reveal that the overall structure of Cd-SrtB conforms the canonical sortase fold. In addition, our previous study also constructed and validated an in silico model of a Cd-SrtB-PPKTG complex and elucidated the molecular interaction governing the PPKTG recognition (40). It was suggested that all sortases form similar sorting signalbinding grooves. The direct evidences came from the currently available structures of sortase-substrate analog complexes: Sa-SrtA-LPAT* (32), Sa-SrtB-NPQT* (33), and Bacillus anthracis sortase A complexed with LPAT* (34). The binding grooves are primarily formed by strands ␤4 and ␤7 and loops connecting ␤2/␤3, ␤3/␤4, ␤6/␤7, and ␤7/␤8. The ␤6/␤7 loop plays a substantial role in interacting and discriminating sorting motif because studies showed that the replacement of the ␤6/␤7 loop in Sa-SrtA with the corresponding site from Sa-SrtB results in converting the specificity profile of Sa-SrtA to Sa-SrtB (33). Our previous work has defined two residues located within ␤6/␤7 loop of Cd-SrtB, Ser163 and Tyr167, that are to be in the direct contact with the substrate peptide PPKTG (40). The structural variation in ␤6/␤7 loop is significant between class A and class B sortase. It draws our attention whether the less dissimilar ␤6/␤7 loop in class B sortase also acts as a determinant for bacteria-specific sorting signal recognition.
Significant efforts have been made to seek the novel therapeutics for CDI, and sortase is one of the most considered targets (28,38,41,42). Therefore, a better understanding of the molecular basis of Cd-SrtB could provide insightful information to facilitate the development of Cd-SrtB-based agents against CDI. Cd-SrtB is not as extensively studied as Sa-SrtA. This work was initiated by a sequence alignment and structural superimposition of class B sortases showing a conserved serine residue near the active site that drew our attention to investigate whether this serine residue participates in catalytic activity. Moreover, we also performed the ␤6/␤7 loop swap between Cd-SrtB and Sa-SrtB to study whether the ␤6/␤7 loop also dictates substrate specificity of class B sortases among different bacteria. Our results demonstrated that the conserved serine residue in proximity to the active site is indispensable for the catalytic activity of Cd-SrtB and Sa-SrtB and that the ␤6/␤7 loop dominates the molecular interactions governing the specific motif recognition.

Ser-207 is essential for the catalytic activity of Cd-SrtB
A sequence alignment of SrtB from Gram-positive bacteria including B. anthracis, S. aureus, Streptococcus pyogenes, and other Gram-positive bacteria showed that a highly conserved serine residue is located near the catalytic cysteine residue (Fig.  S1, A and B). Moreover, superimposition of structures of SrtB from C. difficile (PDB code 5GYJ) (40), S. aureus (PDB code 1NG5) (39), B. anthracis (PDB code 1RZ2) (39), and S. pyogenes (PDB code 3PSQ) (43) also revealed the structural conservation of the serine residue (Fig. S2). These suggest that the conserved Functional residues of C. difficile sortase B serine residue in proximity to the active site may have a functional role in SrtB catalysis.
To investigate whether the equivalent conserved serine residue Ser-207 in C. difficile is involved in the catalytic activity of Cd-SrtB, a mutant replacing Ser-207 with Ala was generated by site-directed mutagenesis. Purified recombinant WT Cd-SrtB and S207A mutant with a deletion of 26 residues at the N-terminal transmembrane region are designated as Cd-SrtB ⌬N26,WT and Cd-SrtB ⌬N26,S207A . The catalytic activity of Cd-SrtB ⌬N26,S207A was assessed by FRET-based assay using PPKTG-containing peptide (Dabcyl-PVPPKTGDSTTIIGE-Edans) as a substrate described previously (40). The results showed that Cd-SrtB ⌬N26,S207A exhibited reduced cleavage activity compared with Cd-SrtB ⌬N26,WT (Fig. 1, A and B), indicating the participation of Ser-207 in catalysis. To study whether two single-site mutations on Ser-207 and Cys-209 produce additive effect, the double mutant Cd-SrtB ⌬N26,S207A/C209A was generated. The results from FRET-based assay showed a nonadditive effect of the double mutations on catalytic activity of Cd-SrtB ( Fig. 1, C and D). To confirm the substrate specificity of recombinant Cd-SrtB ⌬N26,WT , LC-MS/MS was performed to identify the substrate cleavage site. The PPKTG-containing peptide (Dabcyl-PVPPKTGDSTTIIGE-Edans) and scrambled peptide (H-PVGSSTPDSTTIIGE-OH) were incubated with Cd-SrtB ⌬N26,WT for LC-MS/MS studies. The MS/MS spectrum of reaction products revealed a precursor ion at m/z 570.7421 with z ϭ 2 that corresponds to the predicted peptide mass of GDSSTTIIGE-Edans ( Fig. 2, A, upper panel, and B), confirming that Cd-SrtB ⌬N26,WT cleaved PPKTG-containing peptide between Thr and Gly residues (Fig. 2, A and B). In addition, the MS/MS spectrum of reaction mixture of Cd-SrtB ⌬N26,WT with scrambled peptide showed a predominant ion at m/z 730.8409 with z ϭ 2, corresponding to the mass of the intact scrambled peptide (Fig. 2C). Furthermore, the LC-MS/MS results showed that the mutant Cd-SrtB ⌬N26,C209A did not cleave the substrate peptide between Thr and Gly ( Fig. 2A, lower panel), verifying that Cys-209 is the active nucleophilic residue of Cd-SrtB.
To further understand whether the mutation at Ser-207 influences the structure of Cd-SrtB that leads to the reduced enzymatic activity, the crystal structure of Cd-SrtB ⌬N26,S207A was determined at 2.6 Å resolution (Fig. 3A). Structural superposition of Cd-SrtB ⌬N26,S207A with Cd-SrtB ⌬N26,WT showed that Cd-SrtB ⌬N26, S207A mutant has a nearly identical overall structure of Cd-SrtB ⌬N26,WT with a root-mean-square deviation (RMSD) of 0.174 Å for 176 C ␣ coordinates (Fig. 3B). These data revealed that the mutation at Ser-207 did not alter the Functional residues of C. difficile sortase B main-chain conformation on the Cd-SrtB structure but had impact on the catalytic function of Cd-SrtB.

The conserved serine residue in Sa-SrtB is correspondingly indispensable for activity
To explore whether the catalytic contribution of the conserved serine residue is specific to C. difficile, we extended our studies to Sa-SrtB. The analogous experiments were performed with Sa-SrtB to study whether the equivalent serine residue Ser-192 in Sa-SrtB also participates in catalytic activity. The Sa-SrtB ⌬N29,WT , Sa-SrtB ⌬N29,S192A , Sa-SrtB ⌬N29,C194A , and Sa-SrtB ⌬ N29,S192A C194A were generated, purified, and subjected to FRET-based assay using NPQTN-containing peptide as a substrate. Similar to the results from Cd-SrtB, the data showed that the Sa-SrtB ⌬N29,S192A displayed reduced enzymatic activity comparable with the catalytic mutant Sa-SrtB ⌬N29,C194A , and no additive effect was observed on double mutations (Fig. 4). These studies demonstrated that a conserved serine residue plays a role in catalysis in both Cd-SrtB and Sa-SrtB.

Characterize effects of mutations at His-Cys-Arg triad on Cd-SrtB activity
The conserved serine residue was found to be essential for the catalytic activity of Cd-SrtB and Sa-SrtB in this study that led us to be keen on reconfirming the functional contribution of His-116 and Arg-217 in the active site of Cd-SrtB. The mutational impact of His-Cys-Arg catalytic triad on sortase activity has been studied in S. aureus (19,33), B. anthracis (39), and S. pyogenes (44) but has never been assessed in C. difficile. Therefore, we created mutants Cd-SrtB ⌬N26,H116A and Cd-SrtB ⌬N26,R217A , in addition to Cd-SrtB ⌬N26,C209A . The recombinant purified Cd-SrtB ⌬N26,H116A and Cd-SrtB ⌬N26,R217A were subjected to FRET-based assay. To our surprise, the Cd-SrtB ⌬N26,H116A mutant exhibited a comparable level of enzymatic activity as Cd-SrtB ⌬N26,WT , indicating that the mutation of His-116 to alanine has no detectable effect on the catalytic activity. The Cd-SrtB ⌬N26,R217A mutant exhibited significantly decreased enzymatic activity as expected (Fig. 5). It is believed that the

Functional residues of C. difficile sortase B
conserved active-site arginine residue plays a critical role in stabilizing the oxyanion transition state of the enzyme via electrostatic interactions (33). To confirm that the mutation of Arg to alanine has no effect on the global structure, we determined the crystal structure of Cd-SrtB ⌬N26,R217A to 3.1 Å resolution. Superimposition of structure of Cd-SrtB ⌬N26,WT with that of Cd-SrtB ⌬N26,R217A revealed very small structural variation (RMSD ϭ 0.239 for 176 C ␣ atoms), and the structures differ mostly in the replacement of the Arg-217 side chain with Ala methyl group (Fig. 6). The results suggest that the absence of the positively charged Arg-217 guanidine group results in the disruption of the electrostatic interactions, which is essential for stabilization of the tetrahedral oxyanion intermediate.

The ␤6/␤7 loop is the specificity determinant of Cd-SrtB
Our previous study provided a computational model of the Cd-SrtB ⌬N26 -PPKTG complex in which Ser-163 is hydrogenbonded with the P2 lysine residue of PPKTG motif and Tyr-167 interacts with P4 proline noncovalently (40). The mutagenic study also confirmed that the abolishment of these specific interactions affected the cleavage activity of Cd-SrtB, confirming that Ser-163 and Tyr-167 play important roles in specific substrate-binding (40). The Ser-163 and Tyr-167 residues are located in the ␤6/␤7 loop, which has been demonstrated to play a crucial role in substrate specificity between class A and B sortases (45)(46)(47). Larger structural variation in ␤6/␤7 loop between SrtA and SrtB is observed than the differences among SrtB from different bacteria (33,48). Whether the ␤6/␤7 loop of bacteria-specific SrtB can also discriminates its specific motif has not yet been studied.

Discussion
To the best of our knowledge, this work is the first report that reveals an essential serine residue located near the active site of Cd-SrtB contributing to the catalytic activity of Cd-SrtB. In addition, we demonstrated that the corresponding serine residue in Sa-SrtB also participates in the Sa-SrtB-catalyzed cleavage activity, suggesting that the role of the conserved serine residue in SrtB among Gram-positive bacteria is indispensable. Furthermore, we also demonstrated the ␤6/␤7 loop of Cd-SrtB is the specificity determinant for substrate recognition of PPKTG motif, revealing that the ␤6/␤7 loop governs the molecular interactions for classspecific and bacteria-specific motif recognition.

Functional residues of C. difficile sortase B
Our LC-MS/MS analysis confirmed the substrate specificity of Cd-SrtB ⌬N26,WT , and showed that Cd-SrtB ⌬N26,C209A is catalytically inactive (Fig. 2), indicating that Ser-207 is not a nucleophilic residue. In addition, the location of Ser-207 is distant from substrate-binding pocket based on the structural models of Sa-SrtB-NPQT* (33) and Cd-SrtB-PPTKG (40) (Fig.  S3), suggesting that Ser-207 does not directly interact with substrate peptide. Furthermore, crystallographic studies of the Cd-SrtB ⌬N26,S207A mutant suggest that Ser-207 does not play a structural role in Cd-SrtB (Fig. 3). Taken together, Ser-207 may be involved in stabilizing a thioacyl-enzyme intermediate during catalytic process.
Compared with SrtA structure, SrtB contains additional helices at the N terminus and an additional ␣-helix in the ␤6/␤7 loop (23, 33). The ␤6/␤7 loop in Sa-SrtA undergoes significant conformational change that the disordered ␤6/␤7 loop transits to an ordered state upon substrate binding (47). In contrast, the ␤6/␤7 loop in Ba-SrtA and Sa-SrtB forms a well-defined binding pocket for substrates (33). In addition, the replacement of the ␤6/␤7 loop in Sa-SrtA with that of Sa-SrtB shifts the specificity profile of Sa-SrtA to Sa-SrtB, demonstrating that the ␤6/␤7 loop plays a major role in distinguishing the sorting signal between class A and class B sortases (45). Nevertheless, whether the ␤6/␤7 loop is also the specificity determinant for the same class of sortase enzymes has not been investigated. Our results established that the ␤6/␤7 loop also plays a dominate role in recognizing the specific sorting signal of bacteriaspecific SrtB enzymes by loop-swapping mutagenesis. The loop-swapped mutant Cd-SrtB ⌬N26,LS is able to recognize the Sa-SrtB-specific NPQTN motif instead of the cognate PPKTG motif (Fig. 8). It is concluded here that the ␤6/␤7 loop in sortase enzymes is the specificity determinant for the class-specific and bacteria-specific sorting motif.
The mutagenesis studies on His-Cys-Arg triad of SrtB from B. anthracis, S. aureus, and S. pyogenes have demonstrated the essential roles of the catalytic residues (19,33,39,44). Studies have also reported that other residues are involved in catalysis of sortases (19,23,38,39). Trp-194 in Sa-SrtA was shown to assist the thiolate-imidazolium ion-pair formation in active site (19). Asp-196 and Asp-234 were also shown to participate in catalytic site of Sa-SrtB and Ba-SrtB, respectively (39). The discovery of the catalytically essential serine residue near the active site led us to revisit the His-Cys-Arg triad in Cd-SrtB. Our results showed that mutation of Arg-217 significantly decreased the enzymatic activity of Cd-SrtB. The arginine residue in catalytic triad is proposed to be essential for stabilizing the oxyanion intermediate by hydrogen bonding and to facilitate catalysis (33). Structural studies and computational modeling of the Sa-SrtB-NPQT* and Ba-SrtA-LPAT* complexes showed that the active site arginine is hydrogenbonded with P1 threonine (33). Our crystallographic study on Cd-SrtB ⌬N26,R217A showed that the major structural difference between Cd-SrtB ⌬N26,WT and Cd-SrtB ⌬N26,R217A is the side chain at position 217. This may explain how the absence of the arginine guanidino group abolishes the hydrogen-bonding interaction essential for intermediate stabilization. Surprisingly, point mutation on His-116 of Cd-SrtB did not affect the catalytic activity in our study. The corresponding histidine residue in Sa-SrtB, His-130, has been demonstrated to be catalytically essential, because H130A mutant enzyme exhibited no detectable activity (35). Various roles of the catalytically essential histidine residue have been proposed. It was originally thought that the histidine residue activates the nucleophilic cysteine to form a histidine-cysteine ion pair (49). It is now believed that the histidine functions as a general acid/base (36,37). Our results suggest the possibility that Cd-SrtB employs a different functional residue in the active site.
In summary, our work defined key residues essential for Cd-SrtB catalysis and substrate recognition. Our studies provide information that may be useful for developing therapeutic strategies against CDI by manipulating the actions of Cd-SrtB without disrupting the beneficial bacteria in intestinal flora.

Site-directed mutagenesis
The primers listed in Table S1 were used to introduce a desired mutation in dsDNA by an overlapping and a back-toback orientation. KOD FX DNA polymerase (Toyobo), the high-fidelity DNA polymerase enzyme, was used in PCR to prevent polymerase errors during PCR. Subsequently, the reactions were performed in the PCR machine (GeneAmp PCR system 2400, PerkinElmer), and the PCR products were treated by DpnI restriction enzyme (20 units/l; New England Biolabs) to Functional residues of C. difficile sortase B digest methylation DNA from original DNA template and incubated at 37°C for 2 h. Finally, all reaction products were directly transformed into Escherichia coli DH5␣ and confirmed by DNA sequencing.

Protein overexpression and purification
The gene encoding Cd-SrtB was cloned into pMCSG7 vector with an N-terminal His 6 tag described in our previous work (50). Residues 2-27 within the transmembrane domain of Cd-SrtB were deleted to improve solubility of the recombinant protein, designated as Cd-SrtB ⌬N26,WT . The gene encoding Sa-SrtB amplifying from chromosomal DNA of S. aureus was cloned into pET21b ϩ vector with a C-terminal His 6 tag. Residues 2-29 within the transmembrane domain of Sa-SrtB were also deleted to improve solubility of the recombinant protein, designated as Sa-SrtB ⌬N29,WT . The genes encoding sortase mutants generated by site-directed mutagenesis were cloned to pMCSG7 and pET21b ϩ for Cd-SrtB mutants and Sa-SrtB mutants, respectively. All plasmids were transformed to E. coli BL21(DE3). The overexpression of sortases were induced by adding 0.5 mM isopropyl-␤-D-thiogalactopyranoside when cells density reached an A 600 of 0.5-0.6, at the temperatures of 37°C (for Cd-SrtB ⌬N26,WT and mutants) and 25°C (for Sa-SrtB Fractions containing SrtB proteins of mutants from C. difficile and S. aureus were further purified through HiLoad TM 26/60 Superdex TM 75 prep grade column in the ÄKTA prime plus system (GE Healthcare). Size-exclusion chromatography was performed at the flow rate at 1 ml/min with FPLC running buffer (Cd-SrtB ⌬N26,WT and mutants: 10 mM HEPES, pH 7.4, and 150 mM NaCl; Sa-SrtB ⌬N29,WT and mutants: 10 mM HEPES, pH 8.0, and 150 mM NaCl).

FRET assay
The PPKTG-and NPQTN-containing substrate peptides for Cd-SrtB and Sa-SrtB were conjugated with Edans as a fluorophore and Dabcyl as a quencher. FRET assay was carried out as described in our previous work (40). Briefly, when substrate peptide was cleaved by Cd-SrtB ⌬N26,WT /Sa-SrtB ⌬N29,WT , the fluorescence signal could be detected. The reaction of FRET assay was in a total volume of 100 l containing 120 M of Cd-SrtB ⌬N26,WT /Sa-SrtB ⌬N29,WT and 20 M of peptide in black polystyrene 96-well plate (Nunc) with FRET buffer. The plate was incubated at 37°C and was monitored at an excitation/ emission wavelength of 340/490 nm) for every hour during the first 8 h and then at 24 h by using FlexStation 3. All experiments were performed in triplicate, and the data were calculated for the means and the standard error by using GraphPad Prism software (GraphPad Software). The statistical significance of differences between groups was evaluated by two-tailed unpaired Student's t test with GraphPad Prism. A p value of Ϯ0.05 was considered statistically significant: *, p Յ 0.05; **, p Յ 0.01; ***, p Յ 0.001; and ****, p Ͻ 0.0001.

nanoLC-MS/MS analysis and MS/MS database searching
The substrate peptide (Dabcyl-PVPPKTGDSTTIIGE-Edans) and scrambled peptide (PVGSSTPDSTTIIGE) were incubated with Cd-SrtB ⌬N26,WT and Cd-SrtB ⌬N26,C209A , and the reaction mixtures were subjected to LC-MS/MS studies. The peptide mixtures were desalted by C18 Zip-tip (Millipore, Bedford, MA) and evaporated to dryness using a SpeedVac. Dried peptides were dissolved in 5% acetonitrile and 0.1% formic acid, and 5 l of the solution was loaded onto a manually packed precolumn (150-m inner diameter ϫ 30 mm, 5 m, 200 Å) at a 10 l/min flow rate. The peptides were analyzed with a 7-Tesla LTQ-FT Ultra mass spectrometer (linear quadrupole ion trap Fourier transform ion cyclotron resonance; Thermo Scientific, San Jose, CA) coupled to an Agilent 1100 Series binary HPLC pump (Agilent Technologies, Palo Alto, CA), and a FAMOS autosampler (LC Packing, San Francisco, CA). Chromatographic separation was performed over 60 min on a manually packed reversed phase C18 nanocolumn (75-m inner diameter ϫ 200 mm, 3 m, 200 Å) using 0.1% formic acid in water as mobile phase A, 0.1% formic acid in 80% acetonitrile as mobile phase B, and a split flow rate of 300 nl/min. The full-scan mass rage was set from m/z 320 to 2000 with 1000,000 resolution at m/z ϭ 400. The top five most intense ions were sequentially isolated for CID MS/MS fragmentation and detection in the linear ion trap (AGC target at 10,000) with previously selected ions dynamically excluded for 15 s. Ions with Functional residues of C. difficile sortase B singly and unrecognized charge state were also excluded. The electrospray voltage was maintained at 1.7 kV, and the capillary temperature was set to 200°C.
All MS and MS/MS raw data were processed with Proteome Discoverer version 2.3 (Thermo Scientific), and the peptides were identified from the MS/MS data searched against the target substrate peptide sequence (PVPPKTGDSTTIIGE) database using the Mascot search engine 2.6.2 (Matrix Science). Searches were limited to peptide mass tolerance of Ϯ1.0 Da and MS/MS ion mass tolerance of Ϯ1.0 Da. The variable modifications considered were N-terminal proline Dabcyl modification (peptides molecular ϩ 252.1 Da, C 15 H 14 N 3 O) and C-terminal glutamic acid Edans modification (peptides molecular ϩ 250.05 Da, C 12 H 12 NO 3 S). The significant peptide hits defined as peptide score must be higher than Mascot significance threshold (p Ͻ 0.05) and therefore considered reliable, and that manual interpretation confirmed agreement between spectra and peptide sequence. The false discovery rate of the peptides and protein groups was set to 1% for the MS/MS spectra automatically processed by Proteome Discoverer for statistical validation and quantification.

X-ray data collection and structure determination
The crystallographic data of Cd-SrtB ⌬N26,S207A and Cd-SrtB ⌬N26,R217A were collected at Beamline BL13B1 equipped with CCD detector (Q315, ADSC) at the National Synchrotron Radiation Research Center in Hsinchu, Taiwan. Diffraction results were processed by HKL2000 (51). The native data were collected at the wavelength of 1.000 Å. Both Cd-SrtB ⌬N26,S207A and Cd-SrtB ⌬N26,R217A were crystallized in the same space group I23. The unit cell parameters were a ϭ b ϭ c ϭ 120.969 Å and ␣ ϭ ␤ ϭ ␥ ϭ 90°for Cd-SrtB ⌬N26,S207A and a ϭ b ϭ c ϭ 120.895 Å and ␣ ϭ ␤ ϭ ␥ ϭ 90°f or Cd-SrtB ⌬N26,R217A . All diffraction statistics of SrtB mutants were listed in Table 1. The structures of Cd-SrtB ⌬N26,S207A and Cd-SrtB ⌬N26,R217A were solved by molecular replacement by using the structure of Cd-SrtB ⌬N26 (PDB code 5GYJ) (40) as a search model. The manual model rebuilding was performed by using COOT (52) with the guidance of 2F o Ϫ F c and F o Ϫ F c density maps. The iterative refinement was performed by using program CCP4 (53) and PHENIX (54). The structures of Cd-SrtB mutants S207A and R217A were solved at resolutions of 2.6 and 3.1 Å, respectively. The refinement statistics are shown in Table 1. Coordinates and structure factors with the identifier (PDB code 6KYD-R217A) and (PDB code 6KYC-S207A) have been deposited in the PDB.  Functional residues of C. difficile sortase B