CylA is a Sequence-Specific Protease Involved in Toxin Biosynthesis

Abstract

CylA is a subtilisin-like protein belonging to a recently expanded serine protease family related to class II lanthipeptide biosynthesis. As a leader peptidase, CylA is responsible for maturation of the enterococcal cytolysin, a lantibiotic important for Enterococcus faecalis virulence. In vitro reconstitution of CylA reveals that it accepts both linear and modified cytolysin peptides with a preference for cyclized peptides. Further characterization indicates that CylA activates itself by removing its N-terminal 95 amino acids. CylA achieves sequence-specific traceless cleavage of non-cognate peptides even if they are post-translationally modified, which makes the peptidase a powerful tool for mining novel lanthipeptides by providing a general strategy for leader peptide removal. Knowledge about the substrate specificity of CylA may also facilitate the development of protease inhibitors targeting cytolysin biosynthesis as a potential therapeutic approach for enterococcal infections.

INTRODUCTION

Lanthipeptides are members of the family of ribosomally synthesized and post-translationally modified peptides (RiPPs) [1,17,34,45]. Like most other RiPPs, lanthipeptides are generated from a linear precursor peptide with a leader peptide at the N-terminus that is important for recognition by the post-translational modification machinery [32,48]. The modifications take place in the core peptide that is located at the C-terminus of the precursor peptide (Fig. 1a) [2,34]. The dehydro-amino acids dehydroalanine (Dha) and dehydrobutyrine (Dhb) are first generated by a dehydratase from serine and threonine residues, respectively (Fig. 1a and 1b). Subsequently, these electrophilic residues are attacked by cysteine thiols via a Michael-type addition catalyzed by a cyclase. The resulting thioether cross-linked amino acids are termed lanthionine (Lan) and methyllanthionine (MeLan) (Fig. 1a and 1b). After installation of the post-translational modifications, the leader peptide is removed by one or more proteases and the mature lanthipeptide is exported from the producer cells (Fig. 1a) [32]. Lanthipeptides with antimicrobial activities are termed lantibiotics and some of these exhibit potent growth inhibition activities against pathogens [6,28]. One unique lanthipeptide produced by clinical isolates of Enterococcus faecalis, the enterococcal cytolysin, displays lytic activities against eukaryotic cells including human immune cells, in addition to its antimicrobial activities. Cytolysin has been linked to virulence enhancement in animal models infected by E. faecalis and acute patient mortality in the clinic [8,43].

Fig. 1 |. Biosynthesis of class II lanthipeptides.

(a) Biosynthetic pathway of the class II lanthipeptide cytolysin S (b) Dehydration and cyclization reactions that take place post-translationally. (c) The biosynthetic gene cluster of enterococcal cytolysin and the sequential cleavage event employed during cytolysin maturation. Abu, α-aminobutyric acid.

Lanthipeptides have been classified into four classes, differentiated by the biosynthetic enzymes responsible for installation of the thioether rings [18,34]. For class I lanthipeptides, the dehydration and cyclization reactions are catalyzed by separate enzymes named LanB and LanC [22,30], whereas for class II lanthipeptides, the two reactions are carried out by a single bifunctional enzyme LanM [10]. Class III and IV lanthionine synthetases are trifunctional enzymes with separate Ser/Thr kinase, phosphoThr/phosphoSer lyase, and cyclase domains [16,27]. Compared to these intensively-studied lanthionine synthetases, the proteases responsible for leader peptide removal of lanthipeptides are much less characterized, presumably because these proteins are hard to access through heterologous expression.

A bifunctional LanT transporter protein with an N-terminal papain-like cysteine protease domain is responsible for leader peptide removal of most class II lanthipeptides [11,29]. A subset of class II lanthipeptides require one more cleavage step for maturation performed by LanP proteases [4,5,24,46]. These subtilisin-like proteases are homologous to class I lanthipeptide leader peptidases but are located in a separate clade in a Markov chain Monte Carlo (MCMC) phylogenetic tree [31]. They appear to remove a short oligopeptide after a LanT protein removes the majority of the leader peptide at a double Gly-type cleavage site (GG/GA/GS). For example, CylA is an extracellular serine protease required for the biosynthesis of the enterococcal cytolysin. After installation of thioether rings in the precursor peptides CylL_L and CylL_S by CylM, CylB (a LanT protein) removes the majority of the leader peptide to generate CylL_L’ and CylL_S’ (Fig. 1c) [7]. Extracellularly, CylA further trims these peptides by removing six amino acids at the N-terminus of the modified core peptides to form CylL_L” and CylL_S” (cytolysin L and S), the two peptides that make up mature cytolysin (Fig. 1c) [4]. A recent study on LicP, a peptidase involved in the biosynthesis of the class II lanthipeptide lichenicidin, suggested that the protein was post-translationally cleaved into two polypeptides in an autocatalytic process. The resulting N and C-terminal fragments formed a complex mediated by extensive hydrophobic interactions, which is different from most subtilisin-related proteases. LicP acts in a sequence-specific manner with potential for future utility in bioengineering [36]. Importantly, these LanP peptidases encoded in class II lanthipeptide gene clusters are more widely-occurring than previously anticipated [36], and therefore may serve as a tool for sequence-specific tag or leader peptide removal purposes.

All currently annotated class II LanP proteins contain N-terminal secretion signal peptides and are therefore believed to be extracellularly located [36]. It has been confirmed for LicP that the removal of such a secretion signal peptide is achieved by a self-catalyzed mechanism, in which the first 100 amino acids are tailored during maturation [36]. Similar observations were reported for two class I LanPs, NisP and EpiP, which lack the first 195 and 99 amino acids in their mature forms, respectively [13,41]. The removal of the pro-sequences was suggested to activate the proteases for cleaving their LanA substrates [41,47]. However, to date the full length proteins have never been accessed for any of these LanPs. As a result, no activity comparison has been performed between the mature, processed form of LanP and its full length version to confirm the activation upon removal of the pro-sequence.

In order to better understand the maturation process of the enterococcal cytolysin and to mine more lanthipeptidases with novel sequence-specificities, we focused on the characterization of CylA. In this study, we successfully reconstitute the activity of CylA in vitro, provide evidence that it self-activates, and demonstrate that it recognizes a specific cleavage sequence but is rather tolerant of sequence context. Similar to other sequence-specific proteases, CylA may have a broad range of applications in biotechnology [23,49].

RESULTS AND DISCUSSION

Expression and purification of CylA

The gene encoding CylA was synthesized codon-optimized for E. coli expression (Table S1). Residues 1–26 that are predicted to constitute a secretion signal peptide were omitted. The protein was expressed in E. coli with an N-terminal His₆-tag and purified by immobilized metal affinity chromatography. Purified CylA showed three bands when analyzed by SDS-PAGE, with one band corresponding to the full length His₆-CylA-27–412 and two other bands appearing at molecular weights of about 35 kDa and 10 kDa (Fig. 2a). Matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) analysis indicated masses of 9,591 Da and 34,815 Da (Fig. 2b), consistent with an N-terminal fragment (His₆-CylA-27–95; calculated 9,591 Da) and a C-terminal fragment (CylA-96–412; calculated 34,812 Da). This observation suggested a cleavage event after Glu95, in accordance with a previous report that CylA cleaves itself during maturation [4]. Moreover, His₆-CylA-27–412-E95A expressed and purified only as the full-length protein, supporting the proposed cleavage site (Fig. S1).

Fig. 2 |. Analysis of His₆-CylA-27–412.

(a) SDS-PAGE analysis of heterologously expressed and purified His₆-CylA-27–412. MALDI-TOF mass spectra of the N-terminal fragment (b), and the C-terminal fragment and full length form of His₆-CylA-27–412 (c).

The relative abundance of the full-length protein and proteolytic fragments was determined by gel electrophoresis and subsequent densitometry revealing a molar ratio of 1:1:6 for His₆-CylA-27–412, CylA-96–412, and His₆-CylA-27–95 (Fig. S2). Both His₆-CylA-27–412 and His₆-CylA-27–95 have a hexahistidine tag resulting in binding to the nickel-charged affinity resin. CylA-96–412, however, is not expected to have strong affinity for the resin. The presence of CylA-96–412 in the protein elution fragments may be explained by either post-elution cleavage of the full-length protein His₆-CylA-27–412 or co-purification with the N-terminal fragment His₆-CylA-27–95 through non-covalent interactions. We currently cannot exclude either possibility.

CylA is a subtilisin-like serine protease with a conserved catalytic triad consisting of aspartate, histidine and serine. To test whether the observed cleavage was autocatalytic, another CylA mutant was constructed with the catalytic Ser359 substituted by Ala. His₆-CylA-S359A was purified in its full length form (Fig. S3), and hence we conclude that CylA itself catalyzes the cleavage at position 95. These findings mirror two recent reports, one on the class I lanthipeptidase EpiP and one on the class II peptidase LicP, both of which employ a self-catalyzed mechanism to cleave a pro-sequence from the mature protease [19,36]. We also considered an intermolecular process, however, His₆-CylA-27–412 was fully processed into His₆-CylA-27–95 and CylA-96–412 within 4 h, whereas incubation of His₆-CylA-27–412-S359A with His₆-CylA-27–412 resulted in the removal of its pro-sequence much more slowly (>8 h) under the same conditions (Fig. S4). Although we cannot rule out the possibility of intermolecular cleavage in the cell, our observations favor the hypothesis that CylA maturates through intramolecular removal of a pro-sequence.

CylA removes the leader peptides of CylL_L and CylL_S

We next tested the activity of CylA against the CylL_L and CylL_S peptides. Dehydrated and cyclized CylL_L and CylL_S were obtained by coexpression with their lanthionine synthetase CylM in E. coli [37,39]. Instead of using the membrane protein CylB [15], we employed the commercial protease AspN, which specifically cleaved N-terminal to Asp–5, leaving five amino acids (DVQAE) on the core peptides. These peptides were incubated with CylA and the five amino acids were successfully removed (Fig. 3a). We also incubated CylA with full-length modified CylL_L and CylL_S, and MS analysis demonstrated clean removal of the entire leader peptides (Fig. S5). Cytolysin obtained in this way exhibited the anticipated antimicrobial activity against Lactococcus lactis HP (Fig. S5) and hemolytic activity against rabbit red blood cells (Fig. 3b). Similar results were observed for lichenicidin, where the pre-treatment with LicT is not absolutely necessary for LicP recognition [36]. Importantly, CylA did not require post-translational modifications of the precursor peptides as it also removed the leader peptides from linear CylL_L and CylL_S (Fig. 3c).

Fig. 3 |. In vitro activity of CylA.

(a) MALDI-TOF mass spectra of CylL_L” (left) and CylL_S” (right), both carrying an N-terminal 5-amino acid oligopeptide extension (DVQAE) remaining from the leader, incubated with (magenta) or without (blue) CylA. (DVQAE-CylL_L”, calculated [M + H]⁺ = 3981.6, average mass; observed [M + H]⁺ = 3981.1, average mass. CylL_L”, calculated [M + H]⁺ = 3439.0, average mass; observed [M + H]⁺ = 3438.5, average mass. DVQAE-CylL_S”, calculated [M + H]⁺ = 2576.0, average mass; observed [M + H]⁺ =2575.4, average mass. CylL_S”, calculated [M + H]⁺ = 2033.4, average mass; observed [M + H]⁺ = 2034.0, average mass.) (b) Hemolytic activity of mature cytolysin obtained by CylA digestion of CylM processed precursor peptides. (c) MALDI-TOF mass spectra of linear CylL_L (upper) and CylL_S (lower) incubated with CylA. (CylL_L leader, calculated [M + H]⁺ = 4807, average mass; observed [M + H]⁺ = 4804, average mass. CylLS leader, calculated [M + H]⁺ = 6797, average mass; observed [M + H]⁺ = 6796, average mass. The linear CylL_L core peptide was not observed, presumably because of less efficient ionization of the hydrophobic peptide.)

Post-translationally modified CylL peptides are preferred substrates

To further understand the substrate preference of CylA with respect to linear or modified substrates, we performed a MALDI-TOF MS-based competitive assay for semi-quantitative analysis, where CylA was supplied to a mixture of equimolar amounts of modified and linear CylL_S and the consumption of precursor peptides as well as the production of leader peptides was monitored over time. A Pro to Gly mutation was introduced between the hexa-histidine tag and linear CylL_S peptide (G-CylL_S) to differentiate the otherwise isobaric leader peptides (ΔΜ = 40 Da for the resulting leader peptides). A 60-fold excess of modified and linear CylL_S were incubated with CylA (Fig. 4). The peak area of normalized mass spectra were integrated and the consumption of precursor peptides and the formation of leader peptides were fitted (Fig. S6a). The half-life of modified CylL_S precursor peptide was 6.8 min, whereas the half-life of linear CylL_S was 87 min, 12 times longer than that of modified CylL_S. In addition, a delay of cleavage was observed for linear CylLS in the initial 20 min of incubation, during which modified CylL_S was robustly consumed, suggesting that CylA has a clear preference for the modified precursor peptide. Similar data were also observed when comparing the relative efficiency of CylA cleavage of linear and modified CylL_L, but ion suppression of other ions by the ion of the modified core peptide did not allow a competition experiment. When the peptides were incubated with CylA separately, modified CylL_L was completely cleaved after 80 min whereas under the same conditions linear CylL_L required 16 h (Fig. S6b,c). Thus, our observations suggest that, although both can be accepted as substrates, CylA prefers modified substrates over the linear versions.

Fig. 4 |. Time-dependent MALDI-TOF MS analysis of modified CylL_S and linear G-CylL_S treated with CylA.

Peptides were each supplied with a final concentration of 12 μM in the presence of 0.2 μM His₆-CylA-27–412. The intensity of modified CylL_S precursor peptide in starting material was set to 100% in the region of 8,280–8,420 Da, whereas the intensity of the product CylL_S P-1G leader peptide at the completion of the assay was set to 100% in the region of 6,220–6,380 Da. For a quantitative plot of time-dependent consumption of modified CylL_S and linear G-CylL_S, see Fig. S6.

CylA cleaves non-cognate peptides

It has been suggested that sequence-specific cleavage could be generally achieved by class II lanthipeptidases with LicP as one example [36]. To test whether this hypothesis holds for CylA, we engineered the proposed recognition sequence GDVQAE into two linear lanthipeptide precursor peptides, ProcA1.7 and NisA (Fig. S7). The putative recognition sequence was installed between the leader and core peptides by substituting the native residues at positions −6 to −1 (P1 through P6). Upon incubation with CylA, the leader peptides of both GDVQAE-ProcA1.7 and GDVQAE-NisA were successfully removed (Fig. S7b,c). The tolerance of CylA for the amino acid at the P1’ position was evaluated by altering this position in GDVQAE-ProcA1.7 from threonine to a variety of amino acids with different side chains (Table 1). CylA completely cleaved mutant peptides in which the P1’ position was Ala, Ile, Leu, Asp, Glu, Trp, and Phe (Fig. S8). Cleavage of the T1G and T1K mutants was less efficient as demonstrated by MALDI-TOF MS, and ProcA1.7-T1P was not accepted as substrate (Fig. S8). Interestingly, the T1E mutant of ProcA1.7 was cleaved after Glu1 as well as after Glu–1, producing a mixture of products. Peptides with N-terminal Cys residues have great utilization in native chemical ligation or expressed protein ligation [9,26,40]. However, such peptides are often not accessible by proteolytic cleavage as quite a few commonly used proteases do not accept a Cys at the P1’ position of substrates. In contrast, CylA tolerates Cys in P1’ quite well as demonstrated by clean removal of the leader peptides of both GDVQAE-ProcA1.7-T1C and GDVQAE-NisA-I1C (Fig. S9). Collectively, these results show that the activity of CylA is highly portable but does have limits (Table 1).

Table 1 |.

Peptides with different P’ sequences that are accepted by CylA.

Peptide[a]	P’ positions (P1’–P5’)[a]	Relative cleavage efficiency[b]
CylLL[c]	TTPVC (or MeLan[d])	+++
CylLS[c]	TTPAC (or MeLan[d])	+++
NisA	ITSIS	+
NisA-I1C	CTSIS	+
ProcA1.7	TIGGT	+++
ProcA1.7-T1A	AIGGT	+++
ProcA1.7-T1I	IIGGT	+++
ProcA1.7-T1L	LIGGT	+++
ProcA1.7-T1D	DIGGT	+++
ProcA1.7-T1W	WIGGT	+++
ProcA1.7-T1F	FIGGT	+++
ProcA1.7-T1E	EIGGT	++[e]
ProcA1.7-T1K	KIGGT	+
ProcA1.7-T1C	CIGGT	+
ProcA1.7-T1G	GIGGT	+
ProcA1.7-T1P	PIGGT	−
HalA1	CAWYN	+++
HalA2[f]	TTWPC (MeLan)	+++
His6-ProcA1.7[g]	TDPNS	+

^[a]All peptides have the GDVQAE sequence before the P1’ position.

^[b]+++ refers to the wild type activity (see text). ++ indicates slower but still complete cleavage + indicates slower and incomplete cleavage; - indicates that no cleavages was observed.

^[c]These peptides are substrates with the linear sequences shown and also

^[d]after formation of a MeLan between the Cys in the 5^th position and the Thr at position 1 (see text).

^[e]The ProcA1.7-T1E mutant was cleaved substrate was cleaved at both Glu1 and after Glu–1.

^[f]The first T residue is involved in the formation of MeLan.

^[g]the GDVQAE sequence was inserted between the His-tag sequence and ProcA1.7.

To expand the possible utility of CylA for mining novel RiPPs, we tested CylA’s tolerance with post-translationally modified peptides. The GDVQAE sequence was engineered into HalA1 and HalA2, the precursor peptides for haloduracin α and β (Halα and Haβ), which constitute a two-component lantibiotic [20,25]. GDVQAE-HalA1 and GDVQAE-HalA2 peptides were coexpressed with their cognate lanthionine synthetases HalM1 or HalM2 in E. coli, resulting in the anticipated post-translational modifications including three thioether crosslinks for HalA1 and four such rings in addition to three dehydrations in HalA2 (Fig. S10 and S11). These modified peptides were then incubated with CylA and their leader peptides were successfully removed (Fig. S10 and S11), suggesting that CylA accepts peptides with non-proteinogenic amino acids located in close proximity to the cleavage site as substrates. Encouraged by these results, we further engineered the cleavage site between the hexa-histidine tag and modified ProcA1.7 peptide [38]. Incubation with CylA indeed resulted in the removal of the hexa-histidine sequence (Fig. S12), confirming that CylA could be applied for expression tag removal purposes.

Self-cleavage activates CylA

We next returned to the importance of the self-catalyzed processing step for activity. To assess the effect of self-cleavage on the rate of substrate cleavage, CylA-96–412 and His₆-CylA-27–412-E95A were incubated with modified CylL_S, and the formation of CylL_S” was monitored by liquid chromatography MS. CylA-96–412 catalyzed substrate proteolysis at a rate of 20 min⁻¹ under the conditions used, whereas His₆-CylA-27–412-E95A exhibited an approximate 10-fold lower rate (Fig. S13). Thus, the self-cleavage event leads to activation of CylA, although it is not absolutely required for protease activity under our conditions. The recent X-ray structure of a homolog of the class I protease EpiP illustrates that upon cleavage, the pro-domain interacts non-covalently with the catalytic domain through complementary electrostatic surfaces [19]. This interaction leaves the active site of the processed enzyme more exposed, which may explain the increased activity observed in this work for processed CylA.

CONCLUSION

In summary, we have successfully reconstituted the proteolytic activity of CylA in vitro and demonstrated that it accepts both modified and linear cytolysin precursor peptides. Consistent with what has been reported for NisP, LicP and a prolyloligopeptidase-type protease FlaP responsible for the biosynthesis of class III lanthipeptide flavipeptin [36,41,44], CylA cleaves modified CylL_S and CylL_L more efficiently than the corresponding linear peptides. Moreover, we have provided evidence that CylA can serve as a sequence-specific protease that exhibits considerable tolerance for the P1’ site and cleaves before alanine, glycine, (iso)leucine, cysteine, aspartic acid, and aromatic residues (phenylalanine, tryptophan) (Table 1); Glu, Pro and Lys at the P1’ position are not processed or poorly processed. CylA can also remove small expression tags when the recognition sequence is installed. The removal of leader peptides of lanthipeptides is often difficult as most of the currently available commercial proteases do not tolerate non-proteinogenic amino acids such as dehydro amino acids, lanthionine or methyllanthionine located in close proximity to cleavage sites [3,12]. CylA does tolerate these structures and provides a significant advantage when acting on such peptides, making it a valuable tool for synthetic biology efforts of RiPPs. A general strategy for the leader peptide removal step provided by CylA will also greatly aid genome mining for novel RiPPs.

Although multiple groups have reported that mature LanPs purified from the supernatant of their producing strains lack the first 95–195 residues from their N terminus [4,13,41,47], our results serve as the first evidence in the lanthipeptide family that a LanP protease cleaves itself for activation [33]. A recent study on LicP revealed a novel calcium-independent mechanism for stabilization of a subtilase, which involves the removal of a pro-sequence followed by the insertion of a tryptophan into a hydrophobic pocket in the protease domain [36]. This model was supported by the observation that removal of the pro-sequence was detrimental for the production of soluble LicP. A different maturation/stabilization mechanism may be adopted by CylA, given that it can be purified as an active full length protein with the pro-sequence covalently attached. Although maturation of CylA is likely the result of intramolecular cleavage given its sequence homology with LicP and the robustness of intramolecular hydrolysis, we cannot fully exclude the possibly that removal of the pro-sequence may take place (in part) by intermolecular catalysis. Structural information for CylA is needed to unveil the answers to these questions.

E. faecalis and E. faecium infections account for more than 90% of all enterococcal infections, which constitute a growing health issue worldwide [42,43]. The secretion of cytolysin has been linked to the virulence of E. faecalis, supported by both infection model studies and clinical data. The biosynthetic machinery that installs cytolysin could serve as a potential drug target for treating E. faecalis infections using an anti-virulence strategy [43]. As the final step of cytolysin biosynthesis, the removal of the GDVQAE oligopeptide from CylL_L’ and CylL_S’ by CylA is essential for the production of mature toxin [4]. Drug targeting could be further assisted by the notion that CylA is secreted, unlike the other enzymes involved in cytolysin biosynthesis that require inhibitors to be membrane permeable. With heterologously expressed and purified CylA in hand, high-throughput screening of compound libraries can be performed to identify possible inhibitors and drug candidates.

Materials and Methods

General methods

All polymerase chain reactions (PCRs) were carried out on a C1000™ thermal cycler (Bio-Rad). DNA sequencing was performed by ACGT, Inc. Preparative HPLC was performed using a Waters Delta 600 instrument equipped with appropriate columns. Solid phase extraction was performed with Strata X polymeric reverse phase columns (Phenomenex). MALDI-TOF MS was carried out on a Bruker Daltonics UltrafleXtreme MALDI TOF/TOF instrument (Bruker) or a Voyager-DE-STR instrument (Applied Biosystems). The detection of peptides with low molecular weights (700–3,500 Da), peptides with medium molecular weights (5,000–20,000 Da) and proteins with high molecular weights (20,000–50,000 Da) was achieved by using different instrument settings optimized for these mass ranges. LC-ESI-Q/TOF MS analyses were conducted using a Micromass Q-Tof Ultima instrument (Waters) equipped with a Vydac C18 column (5 μm; 100 A; 250 × 1.0 mm). Absorbance of rabbit hemoglobin solution was measured in 96-well plates with a Synergy™ H4 Microplate Reader (BioTek). Negative numbers are used for amino acids in the leader peptide counting backwards from the leader peptide cleavage site.

Materials

The gene encoding CylA was synthesized by GeneArt (Invitrogen) with codon usage optimized for E. coli expression. The DNA sequence of cylA is listed in Table S1. All other oligonucleotides were synthesized by Integrated DNA Technologies and used as received. Restriction endonucleases, DNA polymerases, and T4 DNA ligase were obtained from New England Biolabs. Media components were purchased from Difco Laboratories. Trypsin was purchased from Worthington Biochemical Corporation; Factor Xa was obtained from New England Biolabs and other endoproteinases were ordered from Roche Biosciences. Defibrinated rabbit blood was purchased from Hemostat Laboratories and used within 10 days of receipt. Chemicals were ordered from Sigma Aldrich or Fisher Scientific. Miniprep, gel extraction and PCR purification kits were purchased from Qiagen.

Strains and plasmids

The indicator strain, Lactococcus lactis HP, was obtained from American Type Culture Collection. E. coli DH5α and E. coli BL21 (DE3) cells were used as host for cloning and plasmid propagation, and host for protein expression, respectively. The co-expression vector pRSFDuet-1 was obtained from Novagen.

Construction of pRSFDuet-1 derivatives for expression of CylA, CylL_L and CylL_S

The cylA, cyL_L and cylL_S genes were amplified using appropriate primers and cloned into the MCS1 of a pRSFDuet-1 vector using restriction sites EcoRI and NotI to generate the plasmids pRSFDuet-1/CylA-27–412, pRSFDuet-1/CylL_L and pRSFDuet-1/CylL_S. Primer sequences are listed in Table S2.

Construction of pRSFDuet-1 derivatives for expression of GDVQAE-ProcA1.7 and GDVQAE-NisA peptides

Genes encoding the mutant peptides were amplified by multi-step overlap extension PCR. First, the amplification of the 5’ leader part was carried out by 30 cycles of denaturing (95 °C for 10 s), annealing (55 °C for 30 s), and extending (72 °C for 15 s) using forward primers forprocA1.7 and nisA and appropriate reverse primers (Table S2) to generate a forward megaprimer (FMP). In parallel, PCR reactions using appropriate forward primers and reverse primers for procA1.7 and nisA (Table S2) were performed to produce 3’ fragments (reverse megaprimer, RMP). The 5’ FMP fragment and the 3’ RMP fragment were purified by 2% agarose gel followed by use of a Qiagen gel extraction kit. The two fragments were combined in equimolar amounts (approximately 20 ng each for a 50 μL PCR) and amplified using the same PCR conditions as above with procA1.7 and nisA primers. The resulting PCR products were purified, digested and then cloned into the MCS1 of pRSFDuet-1 to generate pRSFDuet-1/GDVQAE-ProcA1.7 and pRSFDuet-1/GDVQAE-NisA.

Construction of pRSFDuet-1 derivatives for expression of GDVQAE-ProcA1.7-T1X mutants, CylL_L-T27A, GDVQAE-NisA-I1C, CylA-27–412-E95A, and CylA-27–412-S359A

Plasmids pRSFDuet-1/GDVQAE-ProcA1.7-T1G, pRSFDuet-1/GDVQAE-ProcA1.7-T1F, pRSFDuet-1/GDVQAE-ProcA1.7-T1W, pRSFDuet-1/GDVQAE-ProcA1.7-T1C, pRSFDuet-1/GDVQAE-NisA-T1C, pRSFDuet-1/CylA-27–412-E95A and pRSFDuet-1/CylA-27–412-S359A were generated using Quikchange methodology based on pRSFDuet-1/GDVQAE-ProcA1.7, pRSFDuet-1/GDVQAE-NisA and pRSFDuet-1/CylA-27–412, respectively. Primer sequences are listed in Table S2. Plasmids pRSFDuet-1/DGVQAE-ProcA1.7 T1A, pRSFDuet-1/GDVQAE-ProcA1.7 T1K, pRSFDuet-1/GDVQAE-ProcA1.7 T1D, pRSFDuet-1/GDVQAE-ProcA1.7 T1E, pRSFDuet-1/GDVQAE-ProcA1.7 T1I, pRSFDuet-1/GDVQAE-ProcA1.7 T1L and pRSFDuet-1/DGVQAE-ProcA1.7 T1P, were generated by linearizing pRSFDuet-1 with EcoRI and HindIII, and subsequent purification using a Qiaquick Gel Extraction kit (Qiagen). The genes encoding the mutations were amplified by PCR using primers Proc1.7fp_EcoR1and Proc1.7rp_HindIII (Table S2, encoding homology for overlap with backbone, shown in lowercase letters), genes ProcA1.7-T1X as template (Table S2) and touchdown PCR with the annealing temperature decreasing from 70 °C to 54 °C over 80 cycles (−0.2 °C/cycle). An example PCR amplification cycle consisted of denaturing (98 °C for 10 s), annealing (70 °C to 55 °C, 0.2 °C lower every cycle for 30 s), and extension (72 °C for 30 s). The PCR product (containing homology for Gibson assembly into an EcoRI-HindIII pRSFDuet digest) was purified from agarose gel electrophoresis and extracted using a Qiaquick Gel Extraction kit. The insert was then Gibson assembled into the EcoRI/HindIII-linearized pRSFDuet backbone using a ratio of 10:1 (insert : backbone). The gene encoding CylL_L-T27A was amplified by PCR as described above, using primers CylLL-T27Afp_YF_EcoR1 and CylLL-T27Arp_HindIII for insertion into pRSFDuet-1 that was linearized with EcoRI and HindIII. The plasmid was assembled in a ratio of 10:1 (insert : backbone) using Gibson assembly. The pRSFDuet-1/CylM (MCSII) was linearized with EcoR1 and SacI, followed by purification by gel extraction using a Qiaquick Gel Extraction kit (Qiagen). The gene encoding CylLL-T27A was amplified by PCR as described above, using primers CylLL-T27Afp_EcoR1 and CylLL-T27Arp_HindIII for insertion into lthe inearized pRSFDuet-1/CylM(MCSII) and assembled in a ratio of 10:1 (insert : backbone) using Gibson assembly.

Construction of pRSFDuet-1 derivatives for co-expression of HalM1 and HalM2 with GDVQAE-HalA1 and GDVQAE-HalA2, respectively

Mutant peptide genes were generated by a similar multi-step overlap extension PCR procedure as described above and cloned in the MCS1 of pRSFDuet-1/HalM1–2 and pRSFDuet-1/HalM2–2, respectively, to generate pRSFDuet-1/GDVQAE-HalA1/HalM1–2 and pRSFDuet-1/GDVQAE-HalA2/HalM2–2 plasmids. Primer sequences are listed in Table S2.

Construction of pRSFDuet-1 derivatives for co-expression of ProcM with His₆-GDVQAE-ProcA1.7

pRSFDuet-1/His₆-GDVQAET-ProcA1.7G-1E/ProcM-2 was employed to produce modified His₆-GDVQAE-ProcA1.7 and was constructed using a 2-step Quikchange methodology based on a previously reported plasmid pRSFDuet-1/ProcA1.7G-1E/ProcM-2 [35,38]. In the first quick change reaction, the gene encoding GDVQAE oligopeptide was inserted between the His₆-tag and ProcA sequences, resulting in pRSFDuet-1/His₆-GDVQAE-ProcA1.7G-1E/ProcM-2. The second step reaction provided the pRSFDuet-1/His₆-GDVQAET-ProcA1.7G-1E/ProcM-2 plasmid by switching the Gln residue at the P1’ position to Thr. Primer sequences are listed in Table S2.

Expression and purification of CylA and CylA mutants

E. coli BL21 (DE3) cells were transformed with pRSFDuet-1/CylA-27–412, pRSFDuet-1/CylA-27–412-E95A or pRSFDuet-1/CylA-27–412-S359A and plated on an LB plate containing 50 mg/L kanamycin. A single colony was picked and grown in 20 mL of LB with kanamycin at 37 °C for 12 h and the resulting culture was inoculated into 2 L of LB. Cells were cultured at 37 °C until the OD at 600 nm reached 0.5, cooled to 18 °C, and IPTG was added to a final concentration of 0.1 mM. Cells were cultured at 18 °C for another 10 h before harvesting. The cell pellet was resuspended on ice in LanP buffer (20 mM HEPES, 1 M NaCl, pH 7.5 at 25 °C) and lysed by homogenization. The lysed sample was centrifuged at 23,700×g for 30 min and the pellet was discarded. The supernatant was passed through 0.45μm syringe filters and the protein was purified by immobilized metal affinity chromatography (IMAC) as previously described [21]. The proteins were generally eluted from the column at an imidazole concentration between 150 mM and 300 mM and the buffer was exchanged using GE PD-10 desalting columns pre-equilibrated with LanP buffer. Protein concentration was quantified by its absorbance at 280 nm. The extinction coefficient for His₆-CylA-27–412, His₆-CylA-27–412-E95A and His₆-CylA-27–412-S359A was calculated as 30,830 M⁻¹ cm⁻¹. Aliquoted protein solutions were flash-frozen and kept at −80 °C until further usage.

Expression and purification of modified His₆-CylL_L, His₆-CylL_S, His₆-GDVQAE-HalA1, His₆-GDVQAE-HalA2 and His₆-GDVQAE-ProcA1.7

Modified peptides were obtained using a similar procedure described previously using the corresponding co-expression plasmids [35,39].

Expression and purification of linear His₆-CylL_L, His₆-CylL_S, His₆-GDVQAE-ProcA1.7, His₆-GDVQAE-NisA, and mutant peptides

E. coli BL21 (DE3) cells were transformed with pRSFDuet-1/CylL_L, pRSFDuet-1/CylL_S, pRSFDuet-1/GDVQAE-ProcA1.7, pRSFDuet-1/GDVQAE-NisA, pRSFDuet-1/GDVQAE-ProcA1.7-T1G, pRSFDuet-1/GDVQAE-ProcA1.7-T1F, pRSFDuet-1/GDVQAE-ProcA1.7-T1W, pRSFDuet-1/GDVQAE-ProcA1.7-T1C or pRSFDuet-1/GDVQAE-NisA-T1C and plated on an LB agar plate containing 50 mg/L kanamycin. A single colony was picked and grown in 10 mL of LB with kanamycin at 37 °C for 12 h and the resulting culture was inoculated into 1 L of LB-kan. Cells were cultured at 37 °C until the OD at 600 nm reached 0.5 and IPTG was added to a final concentration of 0.2 mM. Cells continued to be cultured at 37 °C for another 3 h before harvesting. The cell pellet was resuspended at room temperature in LanA start buffer (20 mM NaH₂PO₄, pH 7.5 at 25 °C, 500 mM NaCl, 0.5 mM imidazole, 20% glycerol) and lysed by sonication. The sample was centrifuged at 23,700×g for 30 min and the supernatant was discarded. The pellet was then resuspended in LanA buffer 1 (6 M guanidine hydrochloride, 20 mM NaH₂PO₄, pH 7.5 at 25 °C, 500 mM NaCl, 0.5 mM imidazole) and sonicated again. The insoluble portion was removed by centrifugation at 23,700×g for 30 min and the soluble portion was passed through 0.45-μm syringe filters. His-tagged peptides were purified by immobilized metal affinity chromatography (IMAC) as previously described [21]. The eluted fractions were desalted using reversed phase HPLC using a Waters Delta-pak C4 column (15 μm; 300 Å; 25 × 100 mm) or Phenomenex Luna C5 (5 μm 100 Å, 250 × 10 mm) for the ProcA and NisA peptides or a Strata XL polymeric reversed phase SPE column for the cytolysin peptides. Peptides were lyophilized and stored at −20 °C for future usage.

Intermolecular cleavage of His₆-CylA-27–412-S359A by His₆-CylA-27–412

Parallel reactions were set up with His₆-CylA-27–412 only (final protein concentration of 0.8 mg/mL in LanP buffer), His₆-CylA-27–412-S359A only (final protein concentration of 0.2 mg/mL in LanP buffer), and His₆-CylA-27–412 and His₆-CylA-27–412-S359A together (final protein concentrations of 0.8 mg/mL and 0.2 mg/mL, respectively). The three reactions were allowed to proceed at room temperature for 0, 1, 2, 4, 8 and 22 h before being stopped by addition of SDS loading buffer and boiling at 95 °C for 10 min.

Proteolytic cleavage of the leader peptides

Peptides were dissolved in H₂O to a final concentration of 3 mg/mL. To a 85 μL solution of peptides, 10 μL of 500 mM HEPES buffer (pH 7.5) was added followed by 5 μL of 0.5 mg/mL AspN protease (for modified CylL_L and CylL_S peptides), 0.1 mg/mL CylA protease (for modified and unmodified CylL_L and CylL_S peptides). In cleavage tests for the engineered GDVQAE-containing peptides, CylA was added to a final concentration of 5 μg/mL (for GDVQAE-ProcA1.7-T1C and GDVQAE-NisA-T1C peptides), 25 μg/mL (for His₆-GDVQAE-ProcA1.7) or 10 μg/mL (for all other peptides), whereas the substrate peptide was added to a concentration of 0.3 mg/mL in 50 mM HEPES buffer (pH 7.5). TCEP (1 mM) was employed for modified GDVQAE-HalA1, GDVQAE-ProcA1.7-T1C and GDVQAE-NisA-T1C. The cleavage reactions were kept at 25 °C for 1 to 48 h or left at 37 °C for less than 24 h. Osmotic pressure was adjusted with 150 mM NaCl if desired. The digested peptide mixture was directly used in antimicrobial and hemolytic assays without further purification.

Competition assay of CylA activity with modified and linear CylL_S

To a reaction vessel with 70 μL deionized H₂O, 5 μL each of 2 mg/mL modified CylL_S and linear G-CylL_S peptides were added (final peptide concentration of 12 μM each) followed by 10 μL of 500 mM HEPES buffer (pH 7.5). Then, 10 μL of 0.1 mg/mL CylA was supplied (final protein concentration of 0.2 μM) and the reaction was incubated at room temperature before being stopped by addition of formic acid to a final concentration of 1% at different time points for mass spectrometry analysis.

Data analysis of competition cleavage assay

Mass spectrometry data was processed using the MALDIquant package [14]. Raw data was smoothed using Savitzky-Golay filtering. Following base line removal, peaks were detected and aligned. Signal intensities were quantified by integrating the peak area spanning +/− 10 Da of peak centers. The following normalization is based on the hypothesis that the ionization efficiencies of modified and linear CylL_S are slightly different and the difference is proportionally stable, whereas both leader peptides resulting from CylA-cleaved modified and linear CylL_S have the same ionization efficiency. We first calculated the difference in ionization efficiency for modified CylL_S and linear G-CylL_S peptides based on the mass spectrum of the starting material in which CylL_S and linear G-CylL_S were supplied in a 1:1 ratio. The resulting constant was applied to all other mass spectra to correct for the difference in ionization efficiency for precursor peptides. Next, we calculated the normalization coefficients for precursor peptide peaks detected in the region of 8250–8450 Da and leader peptide peaks identified in the region of 6200–6300 Da by solving a system of linear equations such that at each time point, both the sum of proportions of G-CylL_S precursor and leader peptide, and the sum of proportions of modified CylL_S precursor and leader peptides, equal to 100%. Finally, these coefficients were applied to calculate the relative quantities of remaining starting materials (modified CylL_S and linear G-CylL_S) and formed products (leader peptides) and the reaction kinetics were fit using a sigmoid function.

Cleavage rate comparison of modified and linear CylL_L

Modified CylL_L-T27A and linear YF-CylL_L-T27A were added in 50 μΜ final concentration to separate 50 mM HEPES (pH 7.5) buffered reactions containing 2 μΜ CylA. An aliquot of 10 μL was taken at various time points and quenched with 1% final concentration formic acid for mass spectrometric analysis.

Cleavage rate analysis of full length CylA and CylA-96–412 with modified CylL_S

As full length CylA is not available due to the self-cleavage, His-CylA-27–412-E95A was chosen to serve as a substituent of full length CylA because the self-cleavage in His-CylA-27–412-E95A was abolished, whereas the conserved catalytic C-terminal region of CylA remained unchanged. To obtain the mature protease CylA-96–412, His-CylA-27–412 was kept at 4 °C for 12 h to allow the self-cleavage to proceed until CylA-96–412 was the dominant peak monitored by MALDI-TOF MS (Fig. S14). CylA-96–412 obtained using this method was directly evaluated for its cleavage activity without further purification. To test their activities, CylA-96–412 and His-CylA-27–412-E95A were used at a final concentration of 22 nM and 110 nM, respectively, with modified CylL_S provided at a concentration of 36 μΜ. The reactions were stopped at 3, 6, 12 and 24 min with 1% TFA and the formation of mature CylL_S” was monitored by liquid chromatography MS (LC/MS). A 5 μL volume of sample obtained from the cleavage reaction was applied onto the column that was pre-equilibrated in aqueous solvent A. The solvents used for LC were: solvent A = 0.1% formic acid in 95% water / 5% acetonitrile and solvent B = 0.1% formic in 95% acetonitrile / 5% water. A solvent gradient of 0%-80% B over 30 min was employed and the fractionated sample was directly subjected to ESI-Q/TOF MS analysis. The production of core peptide was analyzed by extracted ion chromatography monitoring the desired product mass 1017 (M+2H⁺).

Antimicrobial assay

L. lactis HP cells were grown in GM17 media under anaerobic conditions at 25 °C for 16 h. Agar plates were prepared by combining 15 mL of molten GM17 agar (cooled to 42 °C) with 150 μL of dense cell culture. The seeded agar was poured into a sterile 100 mm round dish (VWR) to solidify. Peptide samples were directly spotted on the seeded and solidified agar. Plates were incubated at 30 °C for 16 h and the antimicrobial activity was determined by the size of the zone of growth inhibition. Halα and Halβ were obtained by factor Xa cleavage of modified HalA1Xa and HalA2Xa peptides following a reported procedure [35]. CylL_L” and CylL_S” were prepared using modified CylL_L-E-1K and CylL_S-E-1K peptides described previously [39].

Hemolytic assay for cytolysin

A sample of 1 mL of defibrinated rabbit blood was diluted by 20 mL of PBS in a 50 mL conical tube and mixed gently. The PBS-diluted blood sample was centrifuged at 800×g for 5 min at 4 °C and the supernatant containing lysed blood cells and released hemoglobin was discarded. The process was repeated 2 to 4 times until the supernatant was clear. The blood cells were then diluted with PBS to make a 5% solution, which was immediately used to test the hemolytic activity of the peptides. To an 1.5 mL eppendorf tube, 50 μL of 5% washed red blood cell sample was added followed by the addition of the desired peptide samples or controls. PBS was used to adjust the final volume to 85 μL All tubes were kept in a 37 °C incubator to allow the lytic reaction to proceed. At each time point, 8 or 10 μL of reaction mixture was taken out, diluted with 190 μL of fresh PBS and centrifuged at 800×g for 5 min. The supernatant (170 μL) was transferred to a new well and the absorbance was measured at 415 nm. The absorbance of prepared blood sample at each time point was analyzed in triplicate and the maximum absorbance was determined by adding 35 μL of 0.1% Triton in PBS to 50 μL of 5% blood sample and using the same analysis procedure.