Sequential Autolytic Processing Activates the Zymogen of Arg-gingipain

Abstract

Most proteases are synthesized as inactive precursors to protect the synthetic machinery of the cell and allow timing of activation. The mechanisms used to render latency are varied but tend to be conserved within protease families. Proteases belonging to the caspase family have a unique mechanism mediated by transitions of two surface loops, and on the basis of conservation of mechanism one would expect this to be preserved by caspase relatives. We have been able to express the full-length precursor of the Arg-specific caspase relative from the bacterium Porphyromonas gingivalis, Arg-gingipain-B, and we show that it contains N- and C-terminal extensions that render a low amount of latency, meaning that the zymogen is substantially active. Three sequential autolytic processing steps at the N and C terminus are required for full activity, and the N-propeptide may serve as an intramolecular chaperone rather than an inhibitory peptide. Each step in activation requires the previous step, and an affinity probe reveals that incremental activity enhancements are achieved in a stepwise manner. Most proteases are synthesized as inactive precursors to protect the synthetic machinery of the cell and allow timing of activation. The mechanisms used to render latency are varied but tend to be conserved within protease families. Proteases belonging to the caspase family have a unique mechanism mediated by transitions of two surface loops, and on the basis of conservation of mechanism one would expect this to be preserved by caspase relatives. We have been able to express the full-length precursor of the Arg-specific caspase relative from the bacterium Porphyromonas gingivalis, Arg-gingipain-B, and we show that it contains N- and C-terminal extensions that render a low amount of latency, meaning that the zymogen is substantially active. Three sequential autolytic processing steps at the N and C terminus are required for full activity, and the N-propeptide may serve as an intramolecular chaperone rather than an inhibitory peptide. Each step in activation requires the previous step, and an affinity probe reveals that incremental activity enhancements are achieved in a stepwise manner. full-length wild-type 2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol matrix-assisted laser desorption ionization time-of-flight Ac-biotinyl-Lys-Tyr-6-aminohexanoic-Arg-acyloxymethyl ketone t-butoxycarbonyl 7-amino-4-methylcoumarin synthetic complete Proteases of the gingipain family are virulence factors of the periodontal pathogenic bacterium Porphyromonas gingivalis(1Nakayama K. Kadowaki T. Okamoto K. Yamamoto K. J. Biol. Chem. 1995; 270: 23619-23626Abstract Full Text Full Text PDF PubMed Scopus (217) Google Scholar, 2Curtis M.A. Aduse-Opoku J. Rangarajan M. Crit. Rev. Oral Biol. Med. 2001; 12: 192-216Crossref PubMed Scopus (121) Google Scholar, 3Genco C.A. Potempa J. Mikolajczyk-Pawlinska J. Travis J. Clin. Infect. Dis. 1999; 28: 456-465Crossref PubMed Scopus (86) Google Scholar). This group contains two genes that encode Arg-specific proteases (rgpA and rgpB) and one gene encoding a Lys-specific protease (kgp). Gene ablation studies have shown that RgpA and RgpB are required for the activation of Kgp, placing the Arg-specific proteases at the top of a proteolytic pathway required for bacterial growth (4Veith P.D. Talbo G.H. Slakeski N. Dashper S.G. Moore C. Paolini R.A. Reynolds E.C. Biochem. J. 2002; 363: 105-115Crossref PubMed Google Scholar, 5Kadowaki T. Nakayama K. Yoshimura F. Okamoto K. Abe N. Yamamoto K. J. Biol. Chem. 1998; 273: 29072-29076Abstract Full Text Full Text PDF PubMed Scopus (145) Google Scholar). The protein encoded byrgpB is predicted to consist of three distinct segments (see Fig. 1), but only two are found in the mature product isolated from bacterial cultures, the catalytic unit and an Ig domain. Consequently the 205-amino acid N-terminal segment may constitute an activation peptide that restrains the activity of the protease until it reaches its site of action. Structural analysis of the catalytic unit of RgpB (6Eichinger A. Beisel H.G. Jacob U. Huber R. Medrano F.J. Banbula A. Potempa J. Travis J. Bode W. EMBO J. 1999; 18: 5453-5462Crossref PubMed Scopus (155) Google Scholar) demonstrates that it shares its evolutionary origin with a common ancestor of caspases, proteases involved in apoptosis and cytokine activation (7Thornberry N.A. Lazebnik Y. Science. 1998; 281: 1312-1316Crossref PubMed Scopus (6159) Google Scholar). Moreover, homology mapping suggests that the clan encompassing gingipains and caspases also may contain bacterial clostripain, plant and animal legumains (processing proteases) (8Chen J.M. Rawlings N.D. Stevens R.A. Barrett A.J. FEBS Lett. 1998; 441: 361-365Crossref PubMed Scopus (196) Google Scholar), and separase (required for sister chromatid separation during anaphase) (9Uhlmann F. Wernic D. Poupart M.A. Koonin E.V. Nasmyth M. Cell. 2000; 103: 375-386Abstract Full Text Full Text PDF PubMed Scopus (654) Google Scholar). This clan is known as protease clan CD (10Barrett A.J. Rawlings N.D. Biol. Chem. 2001; 382: 727-733Crossref PubMed Google Scholar) or the caspase-hemoglobinase fold (11Aravind L. Koonin E.V. Proteins. 2002; 46: 355-367Crossref PubMed Scopus (137) Google Scholar). The majority of proteases are synthesized as zymogens that await activation at a suitable time to protect the biosynthetic machinery of the cell against activation and to act as a timing event in biological function (12Neurath H. Trends Biochem. Sci. 1989; 14: 268-271Abstract Full Text PDF PubMed Scopus (156) Google Scholar). Thus, one of the key events in any proteolytic pathway is the conversion of the zymogen to the active enzyme. Different protease clans utilize distinct strategies for zymogen maintenance and activation, but within clans there seems to be conservation of a particular strategy. On the basis of conservation of mechanism one would imagine that protease clan CD would embrace a similar activation pathway, meaning that the zymogens of caspases and RgpB should be stabilized by homologous molecular interactions. The molecular determinants of caspase activation have been elucidated (13Renatus M. Stennicke H.R. Scott F.L. Liddington R.C. Salvesen G.S. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 14250-14255Crossref PubMed Scopus (370) Google Scholar, 14Chai J. Wu Q. Shiozaki E. Srinivasula S.M. Alnemri E.S. Shi Y. Cell. 2001; 107: 399-407Abstract Full Text Full Text PDF PubMed Scopus (214) Google Scholar, 15Riedl S.J. Fuentes-Prior P. Renatus M. Kairies N. Krapp R. Huber R. Salvesen G.S. Bode W. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 14790-14795Crossref PubMed Scopus (194) Google Scholar), and this seems not to involve the removal of N-terminal segments common for other protease clans (12Neurath H. Trends Biochem. Sci. 1989; 14: 268-271Abstract Full Text PDF PubMed Scopus (156) Google Scholar). Consequently the understanding that the precursor of RgpB (pro-RgpB) may require truncation at its N terminus (or even C terminus) for its activation serves as a good model to test the conservation hypothesis for protease zymogen activation since these observations would seem to contrast with the caspase activation mechanism. This study presents a detailed investigation of the autocatalytic processing of recombinant pro-RgpB, including the characterization of intermediates on the activation pathway, to clarify the mechanism of pro-RgpB maturation. Saccharomyces cerevisiaestrain YG227 (Mata,Δalg6::HIS3,ade2–101, his3Δ200,ura3–52, lys2–801) was kindly provided by Markus Aebi (Institute of Microbiology, ETH Zentrum, Zurich).Escherichia coli DH5α was used as the host for construction and propagation of all plasmids. Standard and synthetic media were prepared and supplemented with nutrients appropriate for selection and maintenance of plasmids as described previously (16Johnston J.R. Molecular Genetics of Yeast: A Practical Approach. 141. IRL Press at Oxford University Press, Oxford1994: 123-134Google Scholar). Yeast cells were grown in 2% glucose as carbon source and 2% galactose and 1% raffinose to induce protein expression fromGAL1 promoter. The pRS316Gal(ΔAcc65I) expression vector was produced for internal laboratory purposes as follows, although the deletion of Acc65I was not specifically necessary for the cloning strategy used in this report. The pRS316(ΔAcc65I) was first generated by digestion of pRS316 (17Sikorski R.S. Hieter P. Genetics. 1989; 122: 19-27Crossref PubMed Google Scholar) with Acc65I and blunt ending of the overhang by T4 polymerase and religation in the presence ofAcc65I. Clones lacking the KpnI/Acc65I site were identified by restriction enzyme digest and used to generate the expression vector. The sequence encoding the Gal1promoter was amplified from genomic yeast DNA using primers Gal1f (5-ttggagctcacatggcattaccaccatatacatatc-3) and Gal1r (5-agagcggccgccggtaccgttttttctccttgacgttaaagt-3). This fragment was introduced intro pRS316(ΔAcc65I) as aSacI-NotI insert resulting in pRS316Gal plasmid. The full-length pro form of RgpB was amplified from genomic DNA isolated from P. gingivalis strain HG66. The 24-residue signal peptide of gingipain was replaced by the signal peptide of yeast carboxypeptidase Y amplified from pRA21 plasmid (18Nielsen T.L. Holmberg S. Petersen J.G. Appl. Microbiol. Biotechnol. 1990; 33: 307-312Crossref PubMed Scopus (36) Google Scholar). The insert was then ligated into the yeast expression plasmid pRS316Gal, and insertion of a FLAG epitope sequence (DYKDDDDK) at the C terminus resulted in the pro-RgpB FL1-WT construct (Fig. 1). The XhoI restriction site encoding two amino acids (Leu and Glu) separates the FLAG tag sequence from the authentic C-terminal sequence of pro-RgpB. Specific mutants (Fig.1) were constructed using an overlap polymerase chain reaction technique and the pro-RgpB FL-WT constructs as a template. The double mutant 2The minus sign before the number indicates residues in the RgpB prodomain counting backward using the first residue of the mature protein as the origin. Arg−1→ Ala/Arg−103 → Ala was generated in the same way but with pro-RgpB FL Arg−103 → Ala as a template. All constructs were sequenced completely to confirm that no undesired mutations were present. Yeast cells were transformed by the lithium and were on for selection D. A. R.A. PubMed Scopus Google Scholar). expression was as were used to of and were grown at for by to of and the was to for cells were in and into 2% 1% were by in and in the protease acid and H. A Practical Approach. IRL Press, Scholar). A of was and the cells were in a were by at for by and the was used for Yeast was to a with and with of the same protein was with a and were for activity, and analysis was using an with recombinant protein were and by by to the to at the was on the with protein was with were and for active enzyme or by with the recombinant protein were and at of the recombinant protein was achieved by using N-terminal and processing was in two pro-RgpB was to by the recombinant protein in for at of the reaction was used for affinity and activity the other of the reaction was by and were by with or to the mature RgpB J. R. Travis J. Infect. 1995; PubMed Google Scholar). pro-RgpB was for at to test for intramolecular activation. The RgpB was against a peptide to the N-terminal residues of mature RgpB and not require a N terminus for as All used in the of the ketone were from and The was synthesized by a from an ketone to the described in A. L. J. Chem. 1999; Scopus Google with The ketone was synthesized as described using is and is A.J. R.A. A. Chem. Scopus Google Scholar). from the and was on a and by The of was by against RgpB of known activity from P. gingivalis J. Mikolajczyk-Pawlinska J. D. D. Travis J. J. Biol. Chem. 1998; 273: Full Text Full Text PDF PubMed Scopus Google Scholar). was used at a of activity was by of as described of RgpB with was by analysis as was in for at and then to a reaction and in to a of were by a of the to the The of the of was used to the site was with a and a affinity probe The time and required to and probe were to the is the of required for enzyme to by A of recombinant to was in for at was to a of for at probe and for at probe site was by for by in the activity of recombinant of gingipain was by the of 7-amino-4-methylcoumarin generated by of at by the in at using an in the were in of Specific activity was as the amount of of recombinant protein used for The catalytic and were by a to the using at a from to enzyme of recombinant wild-type RgpB was on active site with of the low of double → Ala/Arg−103 → Ala pro-RgpB mutant and its activity, we were not able to the enzyme by active site we the enzyme on protein by a with for the active pro-RgpB using of expression of and active full-length or mutant of RgpB coli was we the constructs to a Saccharomyces expression that is to and We a FLAG epitope tag at the C terminus of all constructs for of and were by a using by and by protein (Fig. and using a FLAG (Fig. were low and from to of of yeast of the full-length pro-RgpB resulted in the protein but also in processing by of molecular and The N-terminal sequence not be for this to a N but on the molecular in and of by sequence the form to full-length zymogen with signal peptide This was since the expression was for with signal peptide The may the recombinant product was not from the the of the protein from the cell N-terminal sequence analysis that the protein at the protein at (see Fig. for a of the of these residues in the it is that are generated by a mechanism. clarify the conversion or it is to a protease we a catalytic → of Ala for the catalytic completely the an mechanism A and the of we constructed → Ala at the and a double mutant pro-RgpB Arg−103 → Ala and the double mutant were found to be an form to pro-RgpB at Arg−103 was in the → Ala mutant (Fig. This indicates that pro-RgpB sequential processing in at Arg−103 is required for processing at mature RgpB from P. gingivalis is at the C terminus (6Eichinger A. Beisel H.G. Jacob U. Huber R. Medrano F.J. Banbula A. Potempa J. Travis J. Bode W. EMBO J. 1999; 18: 5453-5462Crossref PubMed Scopus (155) Google Scholar), and we were to this protein the C-terminal FLAG tag would be this we generated the construct that the pro form lacking the C-terminal sequence (Fig.1) and of this protein a similar to that of a full-length but as was (Fig. pro-RgpB were at for in (see and to for autolytic this time the full-length wild-type was completely to the mature form A and The mature recombinant enzyme to be in to that of the enzyme at the same The that a yeast was for processing is the protease acid and the and the protease no on the the gingipain at completely Moreover, the catalytic mutant → Ala in the same The double mutant → Ala/Arg−103 → to not generate any by (Fig. A affinity that resulted from an (Fig. C The product not during and we not an in activity as by of or affinity (Fig. and of pro-RgpB mutants were to and by using and RgpB full-length catalytic and of pro-RgpB double recombinant → Ala/Arg−103 → Ala mutant was for at and the reaction were by or affinity with with with RgpB with a of with a amount of and are by an of pro-RgpB by The recombinant pro-RgpB were to by for at was by The were on for the same of time as the and for at in the to → Ala/Arg−103 → We the of processing by using using against the N-terminal and C-terminal with the N-terminal but lacking the C-terminal FLAG was not This reveals that removal of the C-terminal requires removal of the N-terminal pro conversion to the mature form was in the mutant zymogen to the of the C-terminal of the protein in The be inactive it not in although in yeast have been to the that mature gingipain be in by sequential with the product lacking the N-terminal and the C-terminal as in the active enzyme isolated from P. it is not the product the active or the relative of the zymogen and this and to test the of the we used active site of recombinant zymogen and its This technique on the of a active affinity probe to with that the enzyme catalytic J.M. J. Biol. Chem. Full Text Full Text PDF PubMed Scopus Google Scholar). All intermediates in the processing of pro-RgpB have a active site and an catalytic as by (Fig. C and but this not that have two of an enzyme with are in an be only is in the is in is the relative should the activity of the the active form be other low activity, This to enzyme activity as a function of The probe we was for with Arg-specific proteases but was of was using a low probe and time of (see and not to the active (Fig. this we also the intermediates with probe to completely all active (Fig. The of the at probe was to the of with the protein of form (Fig. and 1). low probe a completely with the even it is in the (Fig. and 1). the the the low probe that the enzyme activity with of the three processing The same were also for with the a in was in wild-type and no in activity was for enzyme and double mutant (Fig. the that pro-RgpB activation with the enzyme processing and is by sequential of N-terminal by of the C processing be intramolecular the catalytic site within the or a catalytic site the in an be by the is or H. J. K. Salvesen 1990; PubMed Scopus Google Scholar). The of was on precursor (Fig. a this is with a we that processing is although we intramolecular is with an in activity during and we to confirm this by the of full-length wild-type protein to the mature form by and full-length double mutant → Ala/Arg−103 → the zymogen with the full-length zymogen and an product of (Fig. C and The to a C-terminal to since it is not by the to the first residues of mature The to be a of Arg−103 and and be into for activity to the full-length analysis that the full-length zymogen to of the RgpB as we the protein to be of that by a (see and The full-length form to of activity as (Fig. we that of the against is to the full-length This is on the that probe catalytic these be into for an of enzyme and on the we an in catalytic during conversion of zymogen to the mature enzyme. the of wild-type pro-RgpB are to that of the enzyme isolated of recombinant mature RgpB and zymogen mutant form → Ala/Arg−103 → → Ala/Arg−103 → in a The product from the is a precursor protein with a 24-residue signal N-terminal pro and a C-terminal that are not found in the mature enzyme. previously J. T. N. Travis J. Potempa J. Biol. Chem. 1998; PubMed Scopus Google Scholar, K. Microbiol. PubMed Scopus Google Scholar), the presence of an RgpB precursor not previously been in P. gingivalis, and of recombinant active gingipains in E. coli with activity to the been J. J. J.M. Rangarajan M. M.A. Infect. 1995; PubMed Google 2000; 18: PubMed Scopus Google Scholar). The activity of all proteases is by in a precursor that be to generate the active biosynthetic machinery and for activation (12Neurath H. Trends Biochem. Sci. 1989; 14: 268-271Abstract Full Text PDF PubMed Scopus (156) Google Scholar). This study to the precursor form of RgpB with a to understanding the for the was it to the zymogen or was it required for of the catalytic RgpB is a product we that a would the and processing of a yeast expression We that recombinant pro-RgpB be to generate a mature form of the enzyme to that isolated from P. gingivalis of pro-RgpB the sequential of intermediates to the product The first at Arg−103 is for processing since at this site of authentic The processing step the of the N-terminal and the step the mature enzyme by removal of a C-terminal peptide. we were not able to the site at the C terminus of the in the peptide. conversion of pro-RgpB into the mature enzyme be to the of the enzyme The conversion during expression but also in with the other bacterial factors may processing in but that an processing pathway This that the precursor have at a of proteolytic activity, to the relative activity of the for to full-length wild-type pro-RgpB of processing in and we to at the and processing we as the This full-length was active with an than that of the mature is not for protease zymogens to have proteolytic activity, but this activity is of than that of the active enzyme (12Neurath H. Trends Biochem. Sci. 1989; 14: 268-271Abstract Full Text PDF PubMed Scopus (156) Google Scholar). are a of protease precursors with activity with its including A. M. Science. PubMed Scopus Google Scholar), W. Trends Med. 1998; PubMed Scopus Google Scholar), and (13Renatus M. Stennicke H.R. Scott F.L. Liddington R.C. Salvesen G.S. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 14250-14255Crossref PubMed Scopus (370) Google Scholar, H.R. Salvesen G.S. J. Biol. Chem. 1999; Full Text Full Text PDF PubMed Scopus Google Scholar). in of these activity of the form is low and requires by This not to be the with pro-RgpB since the zymogen form catalytic activity with an of than proteases on synthetic the activity of the enzyme to with processing step as by an with the affinity probe (Fig. it is to the probe the in probe is with an in the activity of the zymogen may the to activity during protein and gingivalis, but there is for the of the We were not able to protein from a construct encoding the mature This indicates that the may serve an intramolecular chaperone function to the of on demonstrates that the is not required for activity and is by the of the protein a M. PubMed Scopus Google Scholar). This intramolecular chaperone function for protease N-terminal is not to but also other D. Biol. PubMed Scopus Google Scholar, J.R. P. Proc. Natl. Acad. Sci. U. S. A. PubMed Scopus Google Scholar). the intramolecular chaperone also act as of the enzyme activity of 1998; PubMed Scopus Google Scholar). the RgpB seems to activity it may function as an but it is that its is to serve in the of the catalytic of of this is the of group and other to express the catalytic in an active or form J. J. J.M. Rangarajan M. M.A. Infect. 1995; PubMed Google Scholar, 2000; 18: PubMed Scopus Google Scholar). the is to the of the RgpB to the activation mechanism of its the and On the basis of conservation of mechanism one would expect the activation mechanisms to be yeast and separase autolytic this not seem to be for activation of the F. Biol. 2002; 12: Full Text Full Text PDF PubMed Scopus Google Scholar, J. Wernic D. J. Biol. 2002; 12: Full Text Full Text PDF PubMed Scopus Google Scholar). may require a known as to allow and of a active been to an C-terminal J.M. M. Barrett A.J. Biochem. J. 2000; PubMed Google Scholar) similar to the for On the other of zymogen of J. Wu Q. Shiozaki E. Srinivasula S.M. Alnemri E.S. Shi Y. Cell. 2001; 107: 399-407Abstract Full Text Full Text PDF PubMed Scopus (214) Google S.J. Fuentes-Prior P. Renatus M. Kairies N. Krapp R. Huber R. Salvesen G.S. Bode W. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 14790-14795Crossref PubMed Scopus (194) Google Scholar) and (13Renatus M. Stennicke H.R. Scott F.L. Liddington R.C. Salvesen G.S. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 14250-14255Crossref PubMed Scopus (370) Google Scholar) a mechanism. the zymogens show at of activity than that of the active H.R. Salvesen G.S. J. Biol. Chem. 1999; Full Text Full Text PDF PubMed Scopus Google Scholar, Q. Salvesen G.S. Biochem. J. PubMed Scopus Google Scholar), and this is by a of two that contain the activity and determinants of the are distinct for and for activation from of these two the precursors of caspases a to affinity J.M. J. Biol. Chem. Full Text Full Text PDF PubMed Scopus Google Scholar, M. Stennicke H.R. Salvesen G.S. J. Biol. Chem. 1998; 273: Full Text Full Text PDF PubMed Scopus Google Scholar, S. Y. S. K. N.A. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: PubMed Scopus Google Scholar), is not of protease and we this as that pro-RgpB is in a similar (Fig. This would that the may not of to a active site as in protease families. the with the the catalytic and of the activation (Fig. removal would then allow activation by a but should await We Scott for