S9 protease WprP₂ catalyzes uniform cleavage on the precursor peptide in RiPP biosynthesis

Abstract

Serine proteases in ribosomally synthesized and post-translationally modified peptides (RiPPs) catalyze the cleavage on the precursor peptides in the biosynthesis of RiPP natural products. Here, we identified an uncharacterized serine protease WprP₂ from Streptomyces venezuelae NPDC049867, encoded next to the radical S-adenosyl-L-methionine (SAM) enzyme WprB₂ involved in the biosynthesis of cyclophane natural products. In vitro characterization of S9 protease WprP₂ revealed that the precursor peptide WprA₂ is uniformly cleaved. The cleavage activity of WprP₂ has not been seen in any serine proteases and expands the S9 protease in RiPP biosynthesis.

Ribosomally synthesized and post-translationally modified peptides (RiPPs) require precise proteolytic cleavage to generate bioactive natural products, yet the diversity of serine proteases involved remains underexplored. Here, the authors identify and characterize the serine protease WprP₂ from Streptomyces venezuelae NPDC049867, revealing its unique cleavage activity on precursor peptides WprA₂ involved in the biosynthesis of cyclophane RiPP natural products.

Introduction

Ribosomally synthesized and post-translationally modified peptides (RiPPs) are a rapidly growing class of peptide natural products defined by their post-translational modifications^1,2. The backbone structure of RiPPs is encoded in the core peptide region of the precursor peptide, which is modified by post-translational modification enzymes (maturases) and then removed from the leader peptide by proteases and released as mature natural products (Fig. 1A)^1,2. In general, the removal of the leader peptide in RiPPs is crucial for the generation of bioactive natural products³. The class I lanthipeptide nisin A is the oldest and most extensively studied RiPP natural product; isolated in 1928, with its structure determined in 1971, assigned to RiPP in 1988, it possess antimicrobial activities against Gram-positive bacteria^4–7. Cleavage of the leader peptide is necessary for the production of biologically active nisin⁸.

Fig. 1. Overview of serine proteases in RiPP biosynthesis.

A General biosynthesis pathway in RiPPs. B Phylogenetic tree of known and newly characterized serine proteases in RiPPs. C Cleavage sites of known and newly characterized serine proteases in RiPP biosynthesis. The known and newly characterized serine proteases in RiPPs are shown as grey and bold black colored letters, respectively. Protease cleavage sites are shown as dashed lines. Core peptides are shown as blue colored bold letters. Residues within the precursor sequence are numbered ‘+1’ from the start of the core peptide.

Proteases involved in RiPP biosynthesis include: (1) cysteine family proteases that usually cleave after a typical double glycine-like motif, (2) serine family proteases that have serine as the nucleophilic residue in the active site, (3) metalloproteases that use metal ions to catalyze cleavage, and (4) other uncategorized proteases⁹. Over the past decade, the number of characterized serine proteases in RiPP biosynthesis has increased (Fig. 1B). To date, only two groups of serine proteases involved in the maturation of RiPP natural products (Fig. 1C): (1) the S8 family which includes AmyP¹⁰, PenP¹⁰, BthP¹⁰, LicP¹¹, NisP¹², ElxP¹³, CylA¹⁴, AprE¹⁵, and CerP¹⁶, involved in the biosynthesis of class I to III lanthipeptides, and PatA/G involved in the biosynthesis of cyanobactin^17,18; and (2) the S9 family, which contains only FlaP¹⁹, OphP²⁰, and MpgP²¹ involved in the biosynthesis of class III lathipeptides, omphalotins, and clavusporins, respectively. The catalytic triad order of the S8 proteases is Asp/His/Ser, while the catalytic triad of the S9 proteases appear in the order of Ser/Asp/His²². The protein structure of the S8 proteases feature a substilisin-like fold, while the S9 proteases feature a α/β hydrolase fold²².

Cross-linking of peptides is an important chemical feature that improves the stability and biological activity of peptides^23,24. An important example is the natural product, darobactin A, a heptapeptide composed of two three-residue motif cyclophanes that targets the essential outer membrane protein BamA of Gram-negative bacteria^24,25. Recently, radical S-adenosyl-L-methionine (SAM) proteins were discovered as a group of RiPP enzymes that catalyze the formation of a three-residue motif cyclophane between one aromatic amino acid and one aliphatic amino acid on the precursor peptide^26–28. However, in many cases, these radical SAM-RiPP biosynthetic gene clusters (BGCs) do not contain proteases^24,29–34, and thus their mature natural products remain elusive, with the exception of the radical SAM-RiPP natural products, darobactin A and dynobactin A isolated from native strains^24,34,35. To date, only one protease annotated as peptidase C39, a cysteine protease has been characterized in radical SAM-RiPP biosynthesis, which cleaves after a canonical double glycine (GG) motif to generate a mature natural product, xenorceptide (Fig. S1)²⁹.

In our previous study, we discovered the radical SAM cyclophane synthase WprB₁, which catalyzes a cross-link between Trp-C5 and Arg-Cγ at all three-residue WPR motifs on the precursor peptide WprA₁ (Fig. 2A)³³. Unlike other radical SAM-RiPP precursor peptides, WprA contains three repeats of the WPR motif. However, no proteases were found in the flanking region of the BGC. We are interested in searching the proteases that can cleave the precursor peptide WprA after installing the cyclophanes at WPR motifs.

Fig. 2. Overview of the WPR motif cross-linking formation by radical SAM enzyme WprB₂ and cleavage activity of S9 protease WprP₂.

A E. coli coexpression of WprA₁B₁C₁ from Xenorhabdus littoralis psl showed cross-linking at all WPR motifs³³. B E. coli coexpression of WprA₂B₂C₂ from Streptomyces venezuelae NPDC049867 followed by Ni-affinity purification and GluC digestion yielded peptide fragments 1-3. C In vitro characterization of native or boiled His₆-WprP₂ + modified His₆-WprA₂ yielded peptide fragments 4-8. D Cleaved unmodified peptide fragments 9-13 generated by WprP₂. Cross-link formation on the peptide sequences is shown as red connectors.

Results and discussion

To this end, we used Position-Specific Iterative Basic Local Alignment Search Tool (PSI-BLAST)³⁶ to expand for radical SAM sequences homologous to WprB₁³³. We identified 16 homologous WprB whose cognate precursor peptides contain one or three repeated WPR motifs and then searched for the presence of a protease in the flanking region (Fig. S2 and Table S1). We detected a S9 protease in two distinct BGCs: one from Streptomyces venezuelae NPDC049867, which encoded a precursor peptide containing three WPR motifs; and the other from Streptomyces jietaisiensis NBC_00023, where the precursor peptide contained only one WPR motif. Based on the presence of three WPR repeats in the precursor peptide, as in our previous study, we aimed to elucidate the cleavage activity by the S9 protease WprP₂ from S. venezuelae NPDC049867 (Fig. 2B). WprP₂ shares amino acid identity with FlaP (28.7%)¹⁷, OphP (24.3%)¹⁸, MpgP (31.4%)²¹, and no significant similarity with all characterized S8 proteases found in RiPPs^10–18.

We first validated the cross-linking activity at WPR motif(s) by coexpression of WprA₂B₂C₂. The N-terminal His₆ tag precursor peptide WprA₂ expressed in Escherichia coli NiCo21(DE3) alone or coexpressed with radical SAM enzyme WprB₂ and/or chaperone WprC₂, purified by Ni-affinity chromatography and digested with GluC. We used GluC instead of trypsin to digest the precursor peptide because trypsin cannot cleave after Arg residue involved in cross-linking, as observed in our previous study³³. Comparative LC-MS analysis of the digests led to the identification of peaks 1-3 with -2 Da mass loss relative to the unmodified peptide fragment (Figs. 2B and S3–S5). The results showed that WprB₂ similar to WprB₁³³, requires the chaperone WprC₂ to catalyze the formation of a cross-link at all three WPR motifs on the precursor peptide. The nature of the cross-links at all three WPR motifs on WprA₂ was assumed to be identical to those generated by WprB₁ (43.8% amino acid identity with WprB₂)³³.

Next, we investigated the S9 protease WprP₂ activity through in vitro experiments (Fig. 2C). As a substrate for biochemical characterization of WprP₂, we used the Ni-affinity purified His₆ tag precursor peptide WprA₂, expressed alone or coexpressed with WprB₂C₂ in E. coli NiCo21(DE3). Separately, N-terminal His₆ tag WprP₂ was expressed in E. coli NiCo21(DE3), purified by Ni-affinity chromatography and analyzed by SDS-PAGE (Fig. S6). When modified full-length His₆-WprA₂ was incubated with native or boiled enzyme His₆-WprP₂, we observed that modified full-length His₆-WprA₂ was present only in the boiled enzyme but not in the native enzyme (Figs. 2D and S7). The native His₆-WprP₂ cleaved after all three WPR motifs and generated three peptide fragments 4-6. Interestingly, further analysis showed that WprP₂ can catalyze the second cleavage on peptide fragments 4-6 before the Pro amino acid at the 12^th residue preceding the WPR motif, generating shorter peptide fragments 7-8, where the cleavage products of 5 and 6 have identical amino acid sequences (Figs. 2C and S8–S10). By detecting peptide fragments of varying lengths, we concluded that WprP₂ catalyzes sequential cleavage. Surprisingly, the native His₆-WprP₂ also cleaves the unmodified full-length precursor peptide in a similar manner as to the modified full-length precursor peptide (Figs. 2D and S11–S13). This indicates that cyclophanes are not necessary for the cleavage activity of WprP₂.

To understand the recognition sequence of WprP₂, we designed two precursor variants His₆-WprA₂-eng1/2 with multiple mutations located away from the cleavage sites D51A/T52A/D53 A/T54A and D42A/F43A/P47A/E48A, and another four precursor variants with a single mutation located at the cleavage sites P45A, Q44A, H38A and R59A. The results showed that the D51A/T52A/D53A/T54A and D42A/F43A/P47A/E48A mutations at the proximal WPR motif do not affect the cleavage activity by WprP₂, but it impaired the cross-link formation by WprB₂ (Figs. S14 and S15). The WprA₂ H38A results showed that the cleavage activity of WprP₂ does not require the His residue after the WPR motif, and thus all expected peptide fragments were detected (Fig. S16). Surprisingly, a single mutation at R59A in WprA₂ prevented cleavage at this position and generated a longer peptide fragment 19, indicating that the first cleavage recognition sequence is the WPR-Xxx site (Figure S17). We next found that the second cleavage recognition sequence is the Gln-Pro site, as single mutation at P45A or Q44A prevented cleavage at this position and generated peptide fragments 21 (22) and 23 (Figs. 3A and S18–S19).

Fig. 3. Overview of mutational study and computational modeling to identify recognition sequence.

A, B Recognition sequence for the cleavage activity of S9 protease WprP₂. C Cleavage sites and catalytic triad of different S9 subfamily proteases. D AlphaFold3 predicted the complex structure of WprP₂ with a truncated unmodified substrate of WprA₂ showing the first cleavage site in the catalytic triad. Conserved and mutated residues on the precursor sequences are shown as asterisk symbols and red colored bold letters, respectively.

We then asked whether WprP₂ could cleave a related substrate WprA₁, which has different amino acid composition and length (Fig. 3B). To test this, we used the modified full-length WprA₁ in our hand for in vitro assay of WprP₂. The LC-MS data detected cleavage after all three WPR motifs by WprP₂ generating peptide fragments 24-26, but we did not observe the second cleavage before the Pro residue preceding the WPR motif (Figs. 3B and S20–S21). Notably, all three Pro residues in WprA₂ are preceded by a Gln, whereas the Pro preceding amino acids in WprA₁ are Asn, Val and Asn, respectively (Fig. S22). This prompted us to design the precursor variant WprA₁ N36Q, and as expected, a second cleavage activity was detected at the Gln-Pro site and generated peptide fragments 28 (Figs. 3B and S23). To cross-validate the recognition sequence of the first cleavage, we used two precursor variants WprA₁ with a single mutation at H52A or R33G in our hand. The results showed that WprA₁ H52A could detect all expected peptide fragments (Fig. S24). While WprA₁ R33G prevented cleavage at this position as expected and generated a longer peptide fragment 29 containing two WPR motifs (Figs. 3B and S25).

Several members of S9 family proteases are pharmacologically relevant, for example prolyl oligopeptidase is a drug target for celiac sprue and diabetes^37,38, while acyl aminoacyl peptidase has been reported to be associated with Alzheimer’s disease, cataract formation and cancer^39,40. The S9 family proteases can be classified into four subfamilies: (1) S9A prolyl oligopeptidase, which cleave peptide bonds after the Pro residue^19,20, while oligopeptidase B cleaved after the Arg residue⁴¹, (2) S9B dipeptidyl peptidase, which cleave peptide bonds at the penultimate Pro residue at the N-terminus⁴², (3) S9C acyl aminoacyl peptidase, which cleave peptide bonds after a N-acetylated Pro residue⁴³, and (4) S9D carboxypeptidase, which sequentially cleave peptide bonds at the C-terminal residue⁴¹. All S9 subfamily proteases have a signature catalytic triad consisting of Ser, Asp, and His residues (Figs. 3C and S26–S27), and a single mutation at S507A, D590A or H621A completely abolished the cleavage activity of WprP₂ (Fig. S28 and S29).

We used AlphaFold3⁴⁴ to predict the complex structure of WprP₂ with truncated unmodified substrates of WprA₂. The results showed when substrate sequence C-terminus contains “WPRHTE”, the WPR-His cleavage site is placed in spatial proximity to catalytic triad (Figs. 3D and S30–S31). As expected, when the C-terminus of the substrate sequence is truncated, the Gln-Pro cleavage site is located near the catalytic triad (Fig. S30), supporting the first and second cleavage sites. In addition, complex structure with truncated modified substrates were constructed and optimized, and the results were consistent with the truncated unmodified substrates (Fig. S32). To provide concrete evidence of sequential cleavages by WprP₂, we performed time-course experiments ranging from 1 to 120 min. Although we did not obtain satisfactory results, we were able to observe that the longer peptide fragments 4-6 have higher intensity than the shorter peptide fragments 7-8 in the first 1 min time-course experiment (Fig. S33). To date, only three S9 proteases FlaP¹⁹, OphP²⁰ and MpgP²¹ have been characterized in RiPP biosynthesis, and WprP₂ differs from them in three ways: (1) our phylogenetic analysis showed WprP₂ does not form a clade with FlaP and OphP, and (2) unlike known S9 proteases, WprP₂ cleaves the peptide bond at the N-terminus of the Pro residue and (3) catalyzes uniform cleavage on the precursor peptide. From a biotechnological perspective, commercial trypsin cannot cleave after cross-linked Arg residues of the modified full-length WprA₁, while WprP₂ can (Fig. S34), highlighting the potential applicability of WprP₂ in cleaving peptide bond.

Conclusion

We identified an uncharacterized serine protease WprP₂ from Streptomyces venezuelae NPDC049867, encoded next to the radical SAM enzyme WprB₂ involved in the biosynthesis of cyclophane natural products. WprP₂ catalyzes the uniform cleavage on the precursor peptide WprA₂, with the first cleavage occurring after WPR motif and second cleavage occurring before the Pro amino acid at the 12^th residue preceding the WPR motif. Such cleavage has not been seen in any serine proteases from RiPP biosynthesis. Furthermore, WprP₂ recognized both Gln-Pro and WPR-Xxx cleavage sites regardless of the amino acid composition and length in between two cleavage sites. This work adds to the growing list of RiPP proteases for peptide bond cleavage and highlights WprP₂ promiscuity in recognizing a wide range of substrates, providing immense potential for generating bioactive peptides.

Methods

Protein expression and purification of precursor peptides

A colony from the transformation above was picked up by a toothpick and added to 3 mL TB medium supplemented with appropriate antibiotics in a 15 mL falcon tube. The 50 mL culture was grown overnight at 37 °C and shaken at 250 rpm. The overnight culture was used to inoculate either 250 mL of antibiotic-supplemented TB media in a 1 L Erlenmeyer flask in a 1:100 (v:v) ratio. The cells were then grown at 37 °C, 250 rpm until OD_{600 nm} reached 1.8-2.4. The culture was then placed on ice water for 30 min and protein expression was induced by addition of IPTG at a 0.1 mM final concentration. After induction, the culture was shaken at 16 °C, 250 rpm for 18 h. The cells were collected by centrifugation at 4000 rpm for 10 min. The denaturing lysis buffer (100 mM NaH₂PO₄, 10 mM Tris, 8 M Urea, 10 mM imidazole, pH 8) was added to cell pellets in a ratio of 3:1 (v:w). The cell pellets were reconstituted and lysed by sonication with a Titanium 7 mm solid probe (20 sec on and 15 sec off for 25 cycles at 50% amplitude). After sonication, the cell debris was removed by centrifugation at 15,000 rpm for 15 min. HisPur Ni-NTA resin (0.7 mL) was added to ~15-20 mL of supernatant in a 50 mL falcon tube and gently shaken for 1 h to allow binding of the precursor peptide to the Ni-NTA resin. Peptide-bound Ni-NTA resin was then washed with denaturing lysis buffer (2 ×1 mL for 0.7 mL resin if this buffer was used to resuspend the cell pellet), NPI-20 (50 mM NaH₂PO_4, 300 mM NaCl, 20 mM imidazole, pH 8, 5 ×1 mL for 0.7 mL resin) and eluted with NPI-250 (50 mM NaH₂PO_4, 300 mM NaCl, 250 mM imidazole, pH 8, 2.5 mL for 0.7 mL resin). Elution fractions were desalted into 50 mM Tris buffer (pH 8.0) using PD Minitrap G-10 columns. The full-length His₆-precursor peptide obtained by coexpression of WprA₂B₂C₂ was digested with GluC (10 µg per 1 mL eluate) at 37 °C for 16 h and analyzed by LC-MS to detect activity from the radical SAM enzyme WprB₂, while the undigested full-length His₆-precursor peptide was used as a substrate for the biochemical characterization of S9 protease WprP₂.

Protein expression and purification of S9 protease WprP₂

A colony from the transformation above was picked up by a toothpick and added to 5 mL TB medium supplemented with appropriate antibiotics in a 15 mL falcon tube. The 50 mL culture was grown overnight at 37 °C and shaken at 250 rpm. The overnight culture was used to inoculate either 250 mL of antibiotic-supplemented TB media in a 1 L Erlenmeyer flask in a 1:100 (v:v) ratio. The cells were then grown at 37 °C, 250 rpm until OD_{600 nm} reached 1.8-2.4. The culture was then placed on ice water for 30 min and protein expression was induced by addition of IPTG at a 0.1 mM final concentration. After induction, the culture was shaken at 16 °C, 250 rpm for 18 h. The cells were collected by centrifugation at 4000 rpm for 10 min. The non-denaturing buffer (50 mM NaH₂PO₄, 10 mM Tris, 10 mM imidazole, 300 mM NaCl, 10% glycerol, pH 8) was added to cell pellets in a ratio of 3:1 (v:w). The cell pellets were reconstituted and lysed by sonication with a Titanium 7 mm solid probe (20 sec on and 15 sec off for 25 cycles at 50% amplitude). After sonication, the cell debris was removed by centrifugation at 15,000 rpm for 15 min. HisPur Ni-NTA resin (0.7 mL) was added to ~15-20 mL of supernatant in a 50 mL falcon tube and gently shaken for 1 h to allow binding of the precursor peptide to the Ni-NTA resin. Peptide-bound Ni-NTA resin was then washed with non-denaturing buffer (2 ×1 mL for 0.7 mL resin if this buffer was used to resuspend the cell pellet), NPI-20 (50 mM NaH₂PO_4, 300 mM NaCl, 20 mM imidazole, pH 8, 5 ×1 mL for 0.7 mL resin) and eluted with NPI-250 (50 mM NaH₂PO_4, 300 mM NaCl, 250 mM imidazole, pH 8, 2.5 mL for 0.7 mL resin). Elution fractions were desalted into 50 mM Tris buffer (pH 8.0) using PD Minitrap G-10 columns. Purified His₆-WprP₂ was determined by SDS-PAGE and subjected to in vitro assay.

In vitro assay of S9 protease WprP₂

His₆-WprP₂ biochemical characterization was carried out in 50 mM Tris-HCl buffer (pH 8.0) with purified full-length His₆-precursor peptide (0.1 - 0.2 mM), 3 mM of dithiothreitol (DTT) and 5 uM of enzyme His₆-WprP₂ in a reaction volume of 100 uL. For the negative control, enzyme His₆-WprP₂ was boiled by heating at 80 °C for 10 min. The reaction solution was incubated at 25 °C for 2 h and quenched with absolute methanol. After centrifugation at 15,000 rpm for 10 min, the products were then subjected to LC-MS analysis as mentioned in the supplementary information.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

J.R.H. and C.-S.P. designed the study; J.R.H., A.H.K. and C.-S.P. designed and carried out experiments for characterization of enzymes; S.M. performed computational optimization; J.R.H. and C.-S.P. drafted the initial version of the manuscript; S.D. and C.-S.P. edited the manuscript before submission.

Peer review

Peer review information

Communications Chemistry thanks Huan Wang and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Data availability

The experimental details; figures for coexpression of WprA₂B₂C₂ and in vitro assay of WprP₂; list of 16 homologous BGCs of wpr; gene sequences and primers used in this study can be found in the Supplementary Information.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s42004-026-01915-w.