Structure-activity studies of Streptococcus pyogenes enzyme SpyCEP reveal high affinity for CXCL8 in the SpyCEP C-terminal

The Streptococcus pyogenes cell envelope protease (SpyCEP) is vital to streptococcal pathogenesis and disease progression. Despite its strong association with invasive disease, little is known about enzymatic function beyond the ELR+ CXC chemokine substrate range. As a serine protease, SpyCEP has a catalytic triad consisting of aspartate (D151), histidine (H279), and serine (S617) residues which are all thought to be mandatory for full activity. We utilised a range of SpyCEP constructs to investigate the protein domains and catalytic residues necessary for enzyme function. We designed a high-throughput mass spectrometry assay to measure CXCL8 cleavage and applied this for the first time to study the enzyme kinetics of SpyCEP. Results revealed a remarkably low Michaelis-Menton constant (KM) of 82 nM and a turnover of 1.65 molecules per second. We found that an N-terminally-truncated SpyCEP C-terminal construct containing just the catalytic dyad of H279 and S617 was capable of cleaving CXCL8 with a similar KM of 55 nM, albeit with a reduced substrate turnover of 2.7 molecules per hour, representing a 2,200- fold reduction in activity. We conclude that the SpyCEP C-terminus plays a key role in high affinity substrate recognition and binding, but that the N-terminus is required for full catalytic activity.


60
Group A Streptococcus (GAS) or Streptococcus pyogenes is a leading human pathogen that manifests 61 clinically as a broad spectrum of diseases, ranging from less severe, usually self-limiting infections to 62 life-threatening invasive diseases such as necrotizing fasciitis and toxic shock syndrome. While much 63 of the global health burden of S. pyogenes can be attributed to auto-immune sequelae such as 64 SpyCEP is expressed by S. pyogenes as a 1647 amino acid, 180 kDa subtilisin-like serine protease, the 86 crystal structure of which was recently solved to 2.8 Å resolution [10] and further refined to 2.2 Å 87 resolution [11]. It is a member of the S8 subtilase family, members of which are characterised by a 88 catalytic triad consisting of an aspartate, histidine, and serine residue, each surrounded by a region 89 of highly conserved amino acids [12]. SpyCEP is unique among streptococcal proteases in that, 90 during maturation, it is autocatalytically cleaved between residues Q244 and S245 into 2 distinct 91 polypeptides, a 30 kDa N-terminal polypeptide and a 150 kDa C-terminal polypeptide. The two 92 polypeptides harbour the separate residues required for the formation of the catalytic site [6]; the 93 N-terminal fragment contains the catalytic D151 residue, and the C-terminal fragment contains the 94 catalytic H279 and S617 residues. Upon cleavage, the two polypeptides re-associate non-covalently 95 to reconstitute the active enzyme [6]. The recent crystal structures have shed further light upon the 96 structure of SpyCEP, describing 9 separate domains, the first 5 of which contain the catalytic triad 97 necessary for enzymatic activity [10]. 98 99 SpyCEP has been included as a target antigen in several recent S. pyogenes vaccine designs due to its 100 cell surface expression, highly conserved nature, and central role in S. pyogenes pathogenesis. 101 Immunisation with SpyCEP successfully elicits a SpyCEP-specific neutralising antibody response, 102 providing protection against systemic bacterial dissemination and reducing disease severity in S. 103 pyogenes intramuscular, skin infection models and non-human primate infection models [13][14][15][16][17][18]. 104 Data that demonstrate vaccine dependence on enzyme inhibition highlight the importance of 105 understanding the enzymatic activity of SpyCEP and the potential to improve upon vaccine or 106 inhibitor design [13]. 107 108 Despite this, little is known about the catalytic properties of the enzyme, except preliminary 109 structure-function studies [6,10,11,19]. Published data largely focus on the impact of SpyCEP on 110 streptococcal pathogenesis and disease progression. In this study we generated multiple SpyCEP 111 constructs to confirm domains necessary for catalytic activity. We then used mass spectrometry to 112 determine the enzyme kinetics of two SpyCEP constructs for the natural substrate CXCL8. 113 114 116 Codon-optimised SpyCEP gene constructs were expressed in Escherichia coli using synthetic gene 117 sequences (GenScript) from Spy_0416 in the SF370 S. pyogenes M1 genome [6,20] representing full-118 length enzyme, SpyCEP 34-1613 , the N-terminal polypeptide, SpyCEP 34-244 , and C-terminal polypeptide 119

Cloning and purification of recombinant SpyCEP constructs in Escherichia coli
SpyCEP 245-1613 . Constructs were also generated with alanine substitutions to replace catalytic 120 residues in the N-terminal fragment (SpyCEP 34-244 D151A), and the C-terminal fragment (SpyCEP 245-121 1613 S617A). To enable downstream protein purification, both the full-length enzyme and C-terminal 122 polypeptides were expressed with a C-terminal 6-histidine tag and the N-terminal polypeptides were 123 expressed with an N-terminal FLAG tag and TEV linker. 124 All SpyCEP constructs were cloned into the vector pET-24B and expressed in BL21 (DE3) competent 125 E. coli, cultured in Terrific Broth medium supplemented with 50 µg/mL kanamycin at 37°C and 126 shaken at 200 rpm, for 3 hours. The cultures were induced with 0.5 mM Isopropyl β-D-1-127 thiogalactopyranoside (IPTG), cooled to 15 o C and shaken at 200 rpm for 16 hours before lysis by 128 sonication on ice. Full-length and C-terminal constructs were purified by Ni-IDA affinity 129 chromatography (GenScript) as per the manufacturer's instructions. N-terminal constructs were 130 purified by anti-flag M2 agarose resin chromatography (Sigma-Aldrich) as per manufacturer's 131 instruction, and further concentrated and purified by SP ion-exchange (Sigma-Aldrich) and Q FF ion-132 exchange chromatography (GE Healthcare) as per manufacturer's instruction. All SpyCEP constructs 133 were subsequently concentrated and purified by size exclusion chromatography on a HiLoad 134 to encode a hexa-histidine tag by inverse PCR using pUCMUTCEP as the template and the primers 5'-157 TATCCTAGGTAGTGTTGTGAATTCGTAATCATGGTCATAG-3' and 5'-158 TATCCTAGGATGATGATGATGATGATGGGCTTTTGCTGAGGTCGTTG-3'. The amplification was instructions. AvrII restriction sites, incorporated at the terminal ends of the primer sequences, 161 facilitated re-circulation of the amplified plasmid. The presence of a 6-His sequence in the modified 162 pUCMUTCEP construct (denoted pUCMUTCEP-HIS) was confirmed by Sanger sequencing using the 163 pUCMUT sequencing primers 5'-GACAGCAACATCTTTGTGAAAGATGG-3' and 5'-164 CATTAATGCAGCTGGCACGAC-3'. The pUCMUTCEP-HIS construct was introduced into H292 by 165 electroporation and crossed into the chromosome by homologous recombination as previously 166 described [21] to generate strain H1317. Secretion of 6-His-tagged SpyCEP into the culture 167 supernatant of H1317 was confirmed by western blotting (data not shown). To purify SpyCEP, S. To visualise generation of both intact CXCL8 and the larger (N-terminal) CXCL8 cleavage product, a 186 rabbit antiserum was raised against the CXCL8 neo-epitope (anti-ENWVQ) that is exposed following 187 SpyCEP cleavage of CXCL8 [22]. SpyCEP constructs were incubated at 37 o C with 18.75 pmol of 188 recombinant human CXCL8 in a final volume of 20 µl, at a 1:50 molar ratio in favour of CXCL8. 189 Digests were halted and separated by SDS-PAGE gel electrophoresis as described above. terminal SpyCEP, 5 pmol CXCL8 were incubated with C-terminal constructs at a range of enzyme: 210 chemokine molar ratios, 1:5 -1:250, in a final volume of 100 µl at room temperature for 60 minutes. 211 Full-length SpyCEP and C-terminal S617A were included as controls at a molar ratio of 1:1000 and 1:5, 212 respectively. To compare SpyCEP cleavage rates of CXCL8 and CXCL1, 5 fmol or 10 fmol SpyCEP were 213 incubated with 2 pmol human CXCL8 and CXCL1 respectively (R&D Systems), in a final volume of 100 214 µl and incubated at room temperature for 30 minutes. All reactions were halted at defined 215 timepoints with the addition of concentration of Pefabloc (Sigma-Aldrich) to a final concentration of 216 2 mg/ml (8.34 mM). 217 Linear regression analyses of the initial five time points of CXCL1 and CXCL8 cleavage (0, 1, 2, 3 and 4 218 minutes) were utilised to determine the maximal rate of SpyCEP activity. 219 Mass spectrometry analysis of CXCL8 cleavage 220 Analysis of CXCL8 cleavage was assayed on a SCIEX API6500 triple quadrupole electrospray mass 221 spectrometer coupled to a high-throughput robotic sample preparation and injection system, 222

259
Screening of CXCL8 cleavage activity using fluorescent western blotting 260 261 To initially assess the activity of recombinant SpyCEP constructs we screened constructs for specific 262 CXCL8 cleavage activity using two-colour multiplex western blotting. Human CXCL8 was incubated 263 for 2 hours at 37 o C with the different SpyCEP constructs, and the reaction products were separated 264 by SDS-PAGE, then immunoblotted using separate antibodies that detect either intact or cleaved 265 CXCL8. Detection of full-length CXCL8 (8 kDa, green bands) or the CXCL8 neo epitope (ENWVQ), 266 exposed after SpyCEP cleavage, (6 kDa, red bands) was evident with this system (Figure 1). 267 Recombinant full-length SpyCEP expressed in E. coli successfully cleaved CXCL8 to completion, with 268 no difference observed between CXCL8 cleavage with S. pyogenes SpyCEP and E. coli SpyCEP ( Figure  269 1). As expected, no CXCL8 cleavage was observed when using the catalytically inactive mutant, 270 SpyCEP D151A, S617A . As has been previously reported [6], the SpyCEP N-and C-termini, when 271 independently expressed and purified, can be re-associated to form an active enzyme R.E SpyCEP, 272 which successfully cleaved CXCL8. The N-terminal fragment of SpyCEP alone could not cleave CXCL8. 273 Unexpectedly, however, the C-terminal fragment of SpyCEP was observed to cleave CXCL8, albeit 274 not to completion. This independent catalytic activity was negated by mutation of the catalytic S617 275 to alanine (C-terminal S617A construct, Figure 1). Cleavage of CXCL8 by the SpyCEP C-terminal fragment 276 was enhanced when re-associated with the catalytically inert N-terminal mutant (N-terminal D151A 277 construct); indeed, activity was equivalent to that observed with the re-associated enzyme under 278 these conditions, with cleavage of CXCL8 to near completion. However, the N-terminal D151A construct 279 was unable to restore catalytic activity to the C-terminal S617A construct when the two were 280

combined. 281
The results demonstrated for the first time that the SpyCEP C-terminal fragment alone is sufficient 282 for catalytic cleavage of CXCL8 in this assay system. The data also established that, although the D151 was dispensable for enzymatic activity. We further sought to examine whether the C-terminal 285 displayed activity against the newly described substrate, LL-37 [9]. Western blot analysis confirmed 286 cleavage of LL-37 by full-length recombinant SpyCEP, demonstrated by a reduction in band size, 287 however the C-terminal fragment was unable to cleave LL-37 despite a high molar ratio of enzyme to 288 LL-37 (1: 10) and a prolonged 16-hour incubation at 37 °C ( Figure S1).

300
To determine whether the catalytic activity of the C-terminal SpyCEP fragment was reproducible 301 over a shorter incubation period, we assessed CXCL8 cleavage by ELISA, where CXCL8 cleavage is 302 detected through a reduction in substrate concentration. SpyCEP C-terminal constructs were 303 incubated with CXCL8 at molar ratios ranging from 1:5 -1:250 (enzyme: chemokine) over a 60-304 minute time course at room temperature. Full-length recombinant SpyCEP and the inactive C-terminal fragment, C-terminal S617A , were included as controls at a molar ratio of 1:1000 and 1:5, 306 respectively. Under these conditions, near complete CXCL8 cleavage was observed for full-length 307 SpyCEP and a dose-dependent increase in catalytic activity was observed for the C-terminal fragment 308 ( Figure 2). After 5 minutes incubation full-length SpyCEP cleaved over 50% of the starting CXCL8 309 input, with only 6.25% of CXCL8 remaining after 1 hour. At the highest concentration of C-terminal 310 SpyCEP tested, a 1:5 molar ratio, the C-terminal fragment alone cleaved 9% of the starting CXCL8 by 311 5 minutes, 25% by 30 minutes and 42% by 60 minutes. Indeed, at the lowest concentration tested, a 312 1:250 molar ratio, the C-terminal of SpyCEP cleaved 9% of CXCL8 input by 1 hour. As demonstrated 313 by immunofluorescent western blotting, the serine residue at position 617 was vital for SpyCEP 314 catalytic function, as no CXCL8 cleavage was observed using the C-terminal S617A construct, even when 315 employed at the highest enzyme: chemokine ratio. After a 1-hour incubation, full-length SpyCEP and 316 C-terminal constructs (assayed using a molar enzyme: chemokine ratio of 1:5, 1:25 and 1:50) cleaved 317 significantly more CXCL8 compared to the C-terminal S617A construct. SpyCEP constructs were co-incubated with 5 pmol CXCL8. Graphs show residual CXCL8 after a 60-321 minute room temperature incubation, using full-length SpyCEP at a 1:1000 ratio to CXCL8; the C-322 terminal SpyCEP construct at a 1:5 -1:250 molar ratio to CXCL8; and the C-terminal S617A mutant at a 323 1:5 molar ratio to CXCL8. Reactions were halted at specified timepoints by the addition of Pefabloc 324 to a final concentration of 2 mg/ml. N=6 experimental replicates for each construct, data points 325 show means, error bars represent SD. ns p > 0.05, * p ≤ 0.05, **** p ≤ 0.0001, at 60 minutes as 326 determined by ordinary one-way ANOVA. 327 SDS-PAGE analysis of the catalytic activity of the C-terminal SpyCEP construct additionally showed 329 that C-terminal activity was not restricted to the CXCL8 substrate. Over 2 hours at 37 o C using a 1:5 330 molar ratio of enzyme: substrate, the SpyCEP C-terminal construct was capable of cleaving human 331 CXCL1 to near completion ( Figure S2). 332 To further assess the activity of SpyCEP and to interrogate reaction rates against separate 333 chemokines, full-length SpyCEP was incubated with CXCL1 and CXCL8 and the remaining chemokine 334 levels determined by ELISA. Human CXCL1 or human CXCL8 were incubated with SpyCEP at room 335 temperature over a 30-minute time course, at 1:200 or 1:400 molar ratios respectively (enzyme: 336 chemokine). 5 fmol of SpyCEP rapidly and efficiently cleaved 2 pmol CXCL8, with ~ 15% of the 337 chemokine input remaining after 10 minutes of incubation ( Figure 3A). This contrasted with CXCL1 338 cleavage, that required 10 fmol SpyCEP to cleave just 25% of the chemokine input over the same 10-339 minute period. Indeed, by 30 minutes SpyCEP had yet to cleave half of the starting amount of CXCL1 340 ( Figure 3B). Utilising a linear regression of the initial 5 timepoints, where SpyCEP activity is maximal, 341 we found that 5 fmol of SpyCEP was able to cleave 284 fmol of CXCL8 per minute, and 10 fmol of 342 SpyCEP was capable of cleaving 62 fmol of CXCL1 per minute. These data confirmed the activity of 343 recombinant SpyCEP and highlighted differential cleavage efficiency across the CXC substrate range 344 -a feature which has been previously recognised but not quantified [8].

365
To directly compare the active full-length and C-terminal SpyCEP constructs and to understand 366 relative catalytic efficiencies, a mass spectrometry approach to continuously assay the generation of 367 the 13 amino acid peptide cleaved from the native substrate CXCL8, following incubation with 368 enzyme, was employed. A range of CXCL8 concentrations (25-2000 nM) were incubated with a fixed 369 concentration of enzyme, either 250 pM of full length SpyCEP, or 40 nM of the C-terminal SpyCEP 245-370 1613 and the production of the 13 amino acid peptide was monitored over time. 250 pM full-length 371 SpyCEP cleaved CXCL8 to near completion over 30 minutes when substrate concentrations were less 372 than 250 nM; incomplete CXCL8 cleavage was observed when substrate concentration was in excess 373 of 250 nM ( Figure S3A). In contrast, 40 nM of the C-terminal SpyCEP 245-1613 cleaved CXCL8 to near 374 completion over 4 hours when the CXCL8 concentration was 250 nM or less; incomplete CXCL8 375 cleavage was again observed when substrate concentrations were over 250 nM ( Figure S3B). Under 376 these conditions the C-terminal SpyCEP fragment maintained measurable catalytic activity 377 throughout, though with reduced efficacy compared to the full-length construct. A 160-fold increase 378 in enzyme concentration and additional 3.5 hours reaction time were required to cleave a 379 comparable amount of CXCL8. 380 Linear regression analyses of 5 time points, where the rate of cleaved CXCL8 production was linear, 381 were used to derive Michaelis-Menton plots (Figure 4) and KM and Kcat values for each construct 382 (Table 2) SpyCEP is a serine protease and a leading virulence factor of S. pyogenes, with a narrow range of 407 substrate specificity, restricted to the family of ELR + CXC chemokines which modulate neutrophil 408 mediated immune responses [3,4,8] and . Autocatalytic processing of SpyCEP results in the 409 generation of two individual fragments that re-assemble to form an active enzyme [6]. Here, we 410 describe the enzyme kinetics of full-length SpyCEP and report the KM of the enzyme for its natural 411 substrate to be remarkably low, just 81.76 nM, consistent with high efficiency. Furthermore, we 412 demonstrate that the C-terminal SpyCEP fragment can catalyse the cleavage of CXCL1 and CXCL8 413 independent of the N-terminal fragment. Indeed, when KM values were compared, they were found 414 to be similar, suggesting that substrate binding may be confined to the C-terminal domain of 415 SpyCEP. The enzymatic activity of the C-terminal SpyCEP fragment was, however, markedly reduced 416 compared to full-length SpyCEP. The N-terminal and N-terminal D151A constructs were equally able to 417 restore full catalytic activity of the C-terminal SpyCEP fragment. Collectively, this suggests that 418 although the aspartate 151 of the N-terminal fragment may be dispensable for catalysis, the domain 419 itself is important for optimal enzyme activity. 420 Serine proteases are ubiquitous and comprise up to one third of all proteolytic enzymes currently 421 described. They are currently categorised by the MEROPS database [23] into 13 distinct clans, being 422 differentiated into groups of proteins based on their evolution from the same common ancestor, 423 with SpyCEP belonging to the S8 family of the SB clan. S8 serine proteases are typified by a classical 424 catalytic triad composed of serine, histidine and aspartic acid residues that together contribute to 425 the hydrolysis of a peptide bond within the substrate. It is recognised that a number of serine 426 protease clans employ a variation on the S8 catalytic triad, utilising instead a triad of serine, 427 histidine, and glutamic acid, or serine, glutamic acid, and aspartic acid residues. Other clans utilise 428 catalytic dyads of lysine and histidine or histidine and serine for proteolytic activity. Our data suggest 429 that SpyCEP activity can reside in a catalytic dyad of histidine and serine, albeit at a greatly reduced efficacy. It is likely this large gulf in efficiency explains the failure of previous studies to detect 431 catalytic activity within the isolated C-terminal SpyCEP fragment [6]. 432 Within serine proteases there are additional features, beyond the catalytic triad residues, which can 433 contribute to activity. The oxyanion hole for example, a pocket in the active site composed of 434 backbone amide NH groups, may provide substrate stabilisation and help drive catalysis [24]. 435 Additional residues located in close proximity to the catalytic pocket can also mitigate a loss of 436 activity resulting from a missing residue, and water also has the potential to moonlight as a missing 437  [27,28]. Enzyme specificity, a constant which measures the cleavage 451 efficiency of enzymes, (K cat /K M ), for full-length SpyCEP was estimated to be 2.02 x10 7 M -1 s -1 , a 452 value in the order of magnitude typical for serine proteases [24]. The specificity constant of the C-453 terminal fragment was ~1500-fold less, 1.36 x10 4 M -1 s -1 , and Kcat 2200 fold less, a reduction that is in 454 line with previously reported aspartic acid mutants from a systematic mutational study of the Bacillus amyloliquefaciens subtilisin catalytic triad [26]. The development of an MS-based assay of 456 SpyCEP activity, that is not dependent on western blotting or ELISA, provides potential for future 457 high-throughput analysis of SpyCEP activity and detection of SpyCEP inhibitors. 458 The kinetic assays attributed a marked increase in CXCL8 turnover to the additional presence of the 459 N-terminal fragment. Although we did not specifically evaluate the role of the aspartate residue at 460 position 151 on the kinetics of SpyCEP activity, this residue did not contribute appreciably to 461 cleavage of CXCL8 when evaluated using immunoblotting. This raises a question as to whether the N-462 terminal fragment confers some additional structural contribution to enzyme activity. Our data 463 strongly suggest that substrate binding is likely to be attributed to the C-terminal fragment, a finding 464 consistent with related cell envelope proteinases of Lactococci [12] and the closely related 465 streptococcal protein, C5a peptidase [29]. 466 The implications of our findings relating to activity of the C-terminal SpyCEP fragment in S. pyogenes 467 pathogenesis are currently unclear; without the N-terminus, the enzymatic activity detected may be 468 too low to be of consequence for chemokine cleavage in vivo, and in nature both N-and C-terminal 469 fragments are likely to co-exist. SpyCEP has been a focus of S. pyogenes vaccine development in 470 recent years [30], used either in isolation or combination with other antigenic targets [11, 14, 16, 17, 471 31]. Vaccine-induced SpyCEP specific antibodies appear not to act through traditional opsonic means 472 [13] and so may act through inhibition of SpyCEP activity. Many vaccine preparations evaluated have 473 been based on 'CEP5'; a polypeptide spanning residues 35 -587 of SpyCEP which contains the N-474 terminal fragment and only part of the C terminal fragment [18]. Our findings relating to enzyme 475 function suggest that antibodies targeting the C-terminal fragment of SpyCEP are more likely to 476 provide greater neutralizing activity, and potentially improve vaccine efficacy. 477