Directing Sequence-specific Proteolysis to New Targets
1998; Elsevier BV; Volume: 273; Issue: 8 Linguagem: Inglês
10.1074/jbc.273.8.4323
ISSN1083-351X
AutoresGary S. Coombs, Robert C. Bergstrom, Edwin L. Madison, David R. Corey,
Tópico(s)Peptidase Inhibition and Analysis
ResumoWe have previously used substrate phage display to identify peptide sequences that are efficiently and selectively cleaved by tissue-type plasminogen activator (t-PA) or urokinase-type plasminogen activator (u-PA). We demonstrate that this information can be used to direct selective proteolysis to new protein targets. Sequences that were labile to selective cleavage by t-PA or u-PA when in the context of a peptide were introduced into the 43–52 (or Ω) loop of staphylococcal nuclease. Both t-PA and u-PA hydrolyze the engineered proteins at the inserted target sequences, and Km values for protein cleavage were reduced up to 200-fold relative to values for cleavage of analogous sequences within 15 residue peptides. Variation of loop size surrounding a target sequence affects the efficiency of t-PA approximately 5-fold more strongly than that of trypsin, suggesting that cleavage by t-PA is more dependent on target site mobility. Cleavage of proteins by t-PA and u-PA is sequence selective. u-PA is 47-fold more active than t-PA for cleavage of a sequence known to be u-PA selective within small peptide substrates, whereas t-PA is 230-fold more active toward a t-PA-selective sequence. We have previously used substrate phage display to identify peptide sequences that are efficiently and selectively cleaved by tissue-type plasminogen activator (t-PA) or urokinase-type plasminogen activator (u-PA). We demonstrate that this information can be used to direct selective proteolysis to new protein targets. Sequences that were labile to selective cleavage by t-PA or u-PA when in the context of a peptide were introduced into the 43–52 (or Ω) loop of staphylococcal nuclease. Both t-PA and u-PA hydrolyze the engineered proteins at the inserted target sequences, and Km values for protein cleavage were reduced up to 200-fold relative to values for cleavage of analogous sequences within 15 residue peptides. Variation of loop size surrounding a target sequence affects the efficiency of t-PA approximately 5-fold more strongly than that of trypsin, suggesting that cleavage by t-PA is more dependent on target site mobility. Cleavage of proteins by t-PA and u-PA is sequence selective. u-PA is 47-fold more active than t-PA for cleavage of a sequence known to be u-PA selective within small peptide substrates, whereas t-PA is 230-fold more active toward a t-PA-selective sequence. To regulate biological processes proteases must efficiently hydrolyze selected peptide bonds in target proteins while leaving other proteins intact. Many members of the chymotrypsin family of serine proteases, including those involved in blood clotting (1Mann K.G. Jenny R.J. Krishnaswamy S. Annu. Rev. Biochem. 1988; 57: 915-956Crossref PubMed Scopus (484) Google Scholar), fibrinolysis (2Madison E.L. Fibrinolysis. 1994; 8: 221-236Crossref Scopus (48) Google Scholar), complement activation (3Sim R.B. Kolble K. McAleer M.A. Dominguez O. Dee V.M. Int. Rev. Immunol. 1993; 10: 65-86Crossref PubMed Scopus (36) Google Scholar), and growth and development (4Matrisian L.M. Hogan B.L.M. Curr. Top. Dev. Biol. 1990; 24: 219-259Crossref PubMed Scopus (205) Google Scholar, 5Werb Z. Alexander C.M. Adler R.R. Matrix. 1992; 1: 337-343PubMed Google Scholar, 6Hamilton R.T. Bruns K.A. Delgado M.A. Shim J.K. Fang Y. Denhardt D.T. Nilsen-Hamilton M. Mol. Reprod. Dev. 1991; 30: 285-292Crossref PubMed Scopus (35) Google Scholar, 7Niedbala M.J. Agents Actions. 1993; 42: 179-193PubMed Google Scholar), possess such specificity. Understanding mechanisms by which such proteases restrict their specificities in vivo may aid identification of physiologically relevant substrates and facilitate design of proteases with novel, highly restricted specificities. Such engineered proteases would be useful additions to the repertoire of biological research tools and might have wide-ranging therapeutic applications. Tissue-type plasminogen activator (t-PA) 1The abbreviations used are: t-PA, tissue-type plasminogen activator; SNase, staphylococcal nuclease; u-PA, urokinase-type plasminogen activator; Plg, plasminogen. is an attractive model for the study of mechanisms that restrict proteolysis by highly specific serine proteases (2Madison E.L. Fibrinolysis. 1994; 8: 221-236Crossref Scopus (48) Google Scholar). t-PA is stringently selective for its physiologic substrate plasminogen (Plg), even though its protease domain has high homology to the nonselective protease trypsin (43% overall and 87% within residues conserved in all known trypsins) (8Rypniewski W.R. Perrakis A. Vorgias C.E. Wilson K.S. Protein Eng. 1994; 7: 57-64Crossref PubMed Scopus (105) Google Scholar). In previous reports we have shown that much of this specificity is inherent in the protease domain (9Madison E.L. Coombs G.S. Corey D.R. J. Biol. Chem. 1995; 270: 7558-7562Abstract Full Text Full Text PDF PubMed Scopus (74) Google Scholar) and have used substrate phage display (10Smith M.M. Shi L. Navre M. J. Biol. Chem. 1995; 270: 6440-6449Abstract Full Text Full Text PDF PubMed Scopus (138) Google Scholar, 11Matthews D.J. Wells J.A. Science. 1993; 260: 1113-1117Crossref PubMed Scopus (323) Google Scholar) to define consensus sequences for optimal cleavage by t-PA (12Ding L. Coombs G.S. Strandberg L. Navre M. Corey D.R. Madison E.L. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7627-7631Crossref PubMed Scopus (89) Google Scholar, 13Ke S.H. Coombs G.S. Tachias K. Corey D.R. Madison E.L. J. Biol. Chem. 1997; 272: 20456-20462Abstract Full Text Full Text PDF PubMed Scopus (102) Google Scholar) and urokinase-type plasminogen activator (u-PA), a related protease that also targets Plg (14Ke S.H. Coombs G.S. Tachias K. Navre M. Corey D.R. Madison E.L. J. Biol. Chem. 1997; 272: 16603-16609Abstract Full Text Full Text PDF PubMed Scopus (55) Google Scholar). The consensus sequence for t-PA cleavage was X(Y/F/R)GR↓(X′)A, whereX is a large hydrophobic residue and X′ can be several different residues but is most often arginine, whereas the optimal sequence for u-PA was GSGR↓SA. We then obtained 6–14 amino acid peptides containing these sequences and demonstrated that hydrolysis by either t-PA or u-PA occurred with the selectivity predicted by the consensus sequences derived from the substrate phage display. We now examine whether highly selective proteolysis can be targeted to engineered protein substrates. Proteolytic cleavage within internal surface loops is likely to be more challenging than hydrolysis of peptides because target sequences are constrained within a structured scaffold at both their amino and carboxyl termini. Little information has been reported regarding the interaction of plasminogen activators with any protein substrate except their physiologic target, Plg. However, data that does exist suggests significant differences in the behavior of t-PA toward peptides and proteins. First, peptides containing the amino acid sequence of the physiological cleavage site within Plg are poor substrates, but Plg is efficiently hydrolyzed (9Madison E.L. Coombs G.S. Corey D.R. J. Biol. Chem. 1995; 270: 7558-7562Abstract Full Text Full Text PDF PubMed Scopus (74) Google Scholar). Second, introduction of a peptide sequence identified by phage display as being highly t-PA labile into an amino-terminal extension of a protein resulted in a 950-fold reduction in Km relative to the same target in a peptide (15Coombs G.S. Dang A.T. Madison E.L. Corey D.R. J. Biol. Chem. 1996; 271: 4461-4467Abstract Full Text Full Text PDF PubMed Scopus (43) Google Scholar). Finally, Thornton and co-workers (16Hubbard S.J. Campbell S.F. Thornton J.M. J. Mol. Biol. 1991; 220: 507-530Crossref PubMed Scopus (346) Google Scholar, 17Hubbard S.J. Eisenmenger F. Thornton J.M. Protein Sci. 1994; 3: 757-768Crossref PubMed Scopus (213) Google Scholar) have shown that the crystallographically determined conformations of 9 known sites of proteolysis within unrelated proteins are not similar to the crystallographically determined conformations of the reactive site loops of proteinaceous protease inhibitors in protease-inhibitor complexes. These authors conclude that up to 12 residues surrounding the scissile bonds would have to be deformed for these target sites to form interactions with proteases that are similar to those observed in trypsin-inhibitor complexes (interactions presumed to be similar to those formed within protease-substrate complexes). These observations suggest that the context of a target sequence can determine its ability to be a substrate and that interactions separate from those of the primary subsites in the substrate binding cleft may make significant contributions to proteolysis. In this report, we investigate the importance of target site mobility and primary sequence for selective cleavage of engineered protein substrates by t-PA and u-PA. We replaced residues 44–51 within the 43–52 (or Ω) loop of staphylococcal nuclease (SNase) with optimal t-PA or u-PA cleavage sites and measured kinetic parameters for proteolysis of these engineered protein substrates. t-PA and u-PA efficiently cleave the engineered SNase variants despite substrate loop constraint due to amino- and carboxyl-terminal attachment to core-forming structural elements of the protein, and both enzymes exhibit the same relative sequence selectivity observed for cleavage of peptide substrates, confirming that proteins can be engineered to be selectively labile to t-PA or u-PA. These studies indicate that highly selective proteases can hydrolyze introduced sequences within proteins. Purified t-PA (ActivaseTM) was provided by Genentech (San Francisco, CA). Bovine trypsin was purchased from Sigma. Purified u-PA isolated from human urine was purchased from Calbiochem. Enzyme concentrations were determined by titration with 4-methylumbelliferyl p-guanidinobenzoate (Sigma) using a Perkin-Elmer LS 50B luminescence fluorometer (9Madison E.L. Coombs G.S. Corey D.R. J. Biol. Chem. 1995; 270: 7558-7562Abstract Full Text Full Text PDF PubMed Scopus (74) Google Scholar). Titrations of trypsin were performed in 100 mm NaCl, 20 mm CaCl2, 50 mm Tris-HCl (pH 8.0). Titrations of t-PA and u-PA were performed in 150 mm NaCl, 10 mm Tris-HCl (pH 7.5). The plasmid pONF1 (18Takahara M. Hibler D.W. Barr P.J. Gerlt J.A. Inouye M. J. Biol. Chem. 1985; 260: 2670-2674Abstract Full Text PDF PubMed Google Scholar), which expresses wild-type SNase, under control of the lac promoter was the expression vector for all of the mutants described. The wild-type sequence was initially altered by deleting the coding region for residues 45–50, and altering the codons for residues 44 and 51 to incorporate a SmaI site. This mutagenesis was accomplished by polymerase chain reaction (19Higuchi R. Krummel B. Saiki R.K. Nucleic Acids Res. 1988; 16: 7351-7367Crossref PubMed Scopus (2149) Google Scholar) using universal primers 5′-AGGCCTCTAGATAACGAGGCG-3′ and 5′-ACTCAAGCTTCGTTTACCATT-3′ situated at unique XbaI and HinDIII sites respectively, and the oligonucleotides 5′-GTTGATACACCTGAACCCGGGGAGAAATATGGTCCTG-3′ and 5′-AGGACCATATTTCTCCCCGGGTTCAGGTGTATCAACC-3′ as mutagenic primers. All mutations, with one exception, were introduced by synthesizing oligonucleotides coding for a desired loop sequence, annealing and ligating them into the introduced SmaI site. The exception is a mutant in which the codon for residue 51 is deleted by the polymerase chain reaction method used to create theSmaI site. Escherichia coli strains HB101 and DH5α were used for expression of the mutants. SNase was prepared as described (20Shortle D. J. Cell. Biochem. 1986; 30: 281-289Crossref PubMed Scopus (80) Google Scholar). Cultures were grown toA600 = 0.9 and induced by addition of 2 mm lactose. Cultures were harvested 3–4 h after induction, and purification of SNase mutants was performed as described previously using BioRex 70 cation exchange resin (Bio-Rad). Purified nuclease mutants were dialyzed against 50 mm NaCl, 2 mmHEPES, pH 6.8, at 4 °C. Where necessary, nuclease solutions were concentrated by centrifugation at 1,000 rpm in 3-ml microconcentrators (Filtron; Northborough, MA). A280 of each purified SNase mutant was measured in a 1-ml quartz cuvette using a Hewlett-Packard 8452 diode array spectrophotometer. Concentrations were determined using the relation [SNase] =A280/ɛ, where the molar absorption coefficient ɛ is 18280 m−1 cm−1(corresponds to 0.93 mg/A280) (19Higuchi R. Krummel B. Saiki R.K. Nucleic Acids Res. 1988; 16: 7351-7367Crossref PubMed Scopus (2149) Google Scholar). Purity was confirmed by polyacrylamide gel electrophoresis. Calf thymus DNA was purchased from Sigma and dissolved in 10 mmCaCl2, 40 mm Tris-Cl, pH 7.4, to a stock concentration of 1.86 mg/ml as determined by measuringA260 and using the molar absorption coefficient ɛ = 0.025 ml/μg−1 cm−1. This stock was denatured by heating to 100 °C for approximately 40 min after which remaining solid was removed by filtration through Whatman #4 filter paper (Whatman International, Maidstone, United Kingdom). Dilutions were then made in 10 mm CaCl2, 40 mm Tris-Cl, pH 7.4, in concentrations ranging from 5–60 μg/ml. DNA solutions of varying concentrations were then added to 1-ml quartz cuvettes and rapidly mixed with nuclease to final concentrations between 6 and 350 nm. Hydrolysis of the DNA was continuously monitored by following the increase in absorbance at 260 nm in a Hewlett-Packard 8452 diode array spectrophotometer over 40 min to 1 h for each digestion. Data was interpreted by Eadie-Hofstee analysis to obtain Km and Vmax for DNA hydrolysis. Errors were calculated as described (21Taylor J.R. An Introduction to Error Analysis. The Study of Uncertainties in Physical Measurements. University Science Books, Mill Valley, CA1982Google Scholar). Vmax is usually reported rather than kcat because of the heterogeneous nature of the chromosomal DNA substrate, but for purposes of comparison with the uncatalyzed pseudo-first order rate we assumed an average molecular weight for tetranucleotide substrate of 1400 and a change in absorbance of 0.3 OD for complete hydrolysis of 50 mg/ml DNA (22Cuatrecasas P. Fuchs S. Anfinsen C. J. Biol. Chem. 1967; 242: 1541-1547Abstract Full Text PDF PubMed Google Scholar,23Serpersu E.H. Shortle D. Mildvan A.S. Biochemistry. 1987; 26: 1289-1300Crossref PubMed Scopus (159) Google Scholar). Potential protein substrates were incubated for various time periods at 37 °C with 10–50 nm trypsin or at concentrations of t-PA and u-PA ranging from 10 to 320 nm. Substrate protein concentrations varied from 1 to 20 μm. For kinetic assays, t-PA and u-PA concentrations were 50 nm, and 5–9 substrate concentrations were used within the range listed above. Proteolytic digests were terminated between 10 and 20% completion by addition of loading buffer and heating to 100 °C for 10–20 min. Samples of each digest were separated on 15% SDS-polyacrylamide gels and stained with Coomassie Brilliant Blue. After destaining, substrate to proteolytic product ratios were determined by densitometric scanning on a model 300A scanning densitometer (Molecular Dynamics, Sunnyvale, CA) operating with ImageQuant 3.0 software. The ratios obtained were used to determine initial velocities of cleavage. The site of cleavage was confirmed by amino-terminal amino acid sequencing of the carboxyl-terminal proteolytic product. Values for kcat and Km were derived from Eadie-Hofstee analysis. For some protein substrates, it was not possible to achieve substrate concentrations that approached Km. For these proteins, individual parameters kcat and Km could not be determined. However, for [S] ≪Km, the Michaelis-Menten equation, V0=Vmax[S]Km+[S]Equation 1 simplifies to the following. V0=VmaxKm[S]Equation 2 Thus Vmax/Km can be obtained from the slope of linear plots of V0(initial velocity) versus [S] (substrate concentration). Data were fit to Equation 2 by linear regression and in all casesR 2 > 0.9. Under these conditionsKm was estimated to be at least 5-fold higher than the highest substrate concentration used in the assay, because if [S] was any larger than 20% of Km, it would significantly affect the denominator of the above Michaelis-Menten equation, and the resultant systematic deviation from linearity would be about 2-fold greater than the errors we observed. This assumption allowed estimation of Vmax and Km as described (24Chapman K.T. Kopka I.E. Durette P.L. Esser C.K. Lanza T.J. Izquierdo-Martin M. Niedzwiecki L. Chang B. Harrison R.K. Kuo D.W. Lin T.-Y. Stein R.L. Hagmann W.K. J. Med. Chem. 1993; 36: 4293-4301Crossref PubMed Scopus (72) Google Scholar, 25Hedstrom L. Szilagyi L. Rutter W.J. Science. 1992; 255: 1249-1253Crossref PubMed Scopus (457) Google Scholar). We used SNase to study the activity of t-PA and u-PA toward protein substrates. SNase contains one of the trypsin cleavage sites examined by Thornton and co-workers (16Hubbard S.J. Campbell S.F. Thornton J.M. J. Mol. Biol. 1991; 220: 507-530Crossref PubMed Scopus (346) Google Scholar,17Hubbard S.J. Eisenmenger F. Thornton J.M. Protein Sci. 1994; 3: 757-768Crossref PubMed Scopus (213) Google Scholar), carboxyl-terminal to lysine 48 within a surface loop formed by residues 43–52, and NMR studies (26Torchia D.A. Sparks S.W. Bax A. Biochemistry. 1989; 28: 5509-5524Crossref PubMed Scopus (142) Google Scholar, 27Kay L.E. Torchia D.A. Bax A. Biochemistry. 1989; 28: 8972-8979Crossref PubMed Scopus (1829) Google Scholar) and x-ray crystallography (28Hynes T.R. Fox R.O. Proteins. 1991; 10: 92-95Crossref PubMed Scopus (251) Google Scholar) conclude that this loop is highly flexible. These observations suggested that this site would be accessible to u-PA or t-PA. We prepared SNase variants in which residues 44–51 within the 43–52 loop were replaced by sequences that are selective for cleavage by t-PA or u-PA. NMR and kinetic studies have demonstrated that variants containing deletions within the 43–52 loop have essentially the same structure as wild type (29Baldisseri D.M. Torchia D.A. Biochemistry. 1991; 30: 3628-3633Crossref PubMed Scopus (35) Google Scholar) and are catalytically active (30Poole L.B. Loveys D.A. Hale S.P. Gerlt J.A. Stanczyk S.M. Bolton P.H. Biochemistry. 1991; 30: 3621-3627Crossref PubMed Scopus (34) Google Scholar), but to provide experimental evidence that this was also true for our variants we obtained kinetic parameters for their hydrolysis of calf thymus DNA. Vmax/Km for the variants was decreased from 2000–20,000-fold relative to wild type (Table I). These substantial reductions were expected since residues 44–51 play a significant role in the rate-limiting processes of substrate binding and product release (31Hale S.P. Poole L.B. Gerlt J.A. Biochemistry. 1993; 32: 7479-7487Crossref PubMed Scopus (37) Google Scholar) and are adjacent to glutamate 43, a residue critical for efficient catalysis (32Judice K.J. Gamble T.R. Murphy E.C. de Vos A.M. Schultz P.G. Science. 1993; 261: 1578-1581Crossref PubMed Scopus (75) Google Scholar).Table IKinetic constants for hydrolysis of denatured calf thymus DNA by wild type and nine SNase variants containing amino acid substitutions for the sequences TKHPKKGV at residues 44–51Inserted sequenceVmaxKmVmax/KmΔA260/(μg nuclease)(min)μg/ml DNAwild type (TKHPKKGV)0.7 ± 0.092 ± 0.70.4 ± 0.1 ( I)PFGRSA0.008 ± 0.0017180 ± 464e−5 ± 9e−6 (II)PFGRSAG0.001 ± 0.000310 ± 81e−4 ± 6e−5 (III)PPFGRSAG0.005 ± 0.000821 ± 82e−4 ± 5e−5 (IV)PGPFGRSAG0.001 ± 0.000315 ± 87e−5 ± 3e−5 (V)PGPFGRSAGG0.003 ± 0.000415 ± 52e−4 ± 5e−5 (VI)PHYGRSGG0.0008 ± 0.0000422 ± 24e−5 ± 3e−6 (VII)PQRGRSAG0.003 ± 0.000551 ± 166e−5 ± 1e−5 (VIII)PGSGRSAG0.004 ± 0.000918 ± 9.92e−4 ± 8e−5 (IX)PGSGRSASGTTGTG0.008 ± 0.0006365 ± 302e−5 ± 2e−6 Open table in a new tab Vmax is used to characterize the kinetics of SNase catalysis because of the heterogeneous nature of the DNA substrate, but it can be converted into an approximatekcat value assuming the molecular weight of an average substrate tetranucleotide to be 1400 and a change in absorbance of 0.3 A260 for complete hydrolysis of 50 μg/μl DNA (22Cuatrecasas P. Fuchs S. Anfinsen C. J. Biol. Chem. 1967; 242: 1541-1547Abstract Full Text PDF PubMed Google Scholar, 23Serpersu E.H. Shortle D. Mildvan A.S. Biochemistry. 1987; 26: 1289-1300Crossref PubMed Scopus (159) Google Scholar). Conversion of Vmax intokcat allows direct comparison with the background rate of phosphodiester hydrolysis. The lowestkcat we calculate for any variant, 0.1 s−1, was over 12 orders of magnitude greater than the pseudo-first order rate constant determined for the uncatalyzed reaction, 5.7 (10Smith M.M. Shi L. Navre M. J. Biol. Chem. 1995; 270: 6440-6449Abstract Full Text Full Text PDF PubMed Scopus (138) Google Scholar)−14 s−1 (23Serpersu E.H. Shortle D. Mildvan A.S. Biochemistry. 1987; 26: 1289-1300Crossref PubMed Scopus (159) Google Scholar, 33Kumamoto J. Cox J.R. Westheimer F.H. J. Am. Chem. Soc. 1956; 77: 4858-4860Crossref Scopus (132) Google Scholar, 34Bunton C.A. Mhala M.M. Oldham K.G. Vernon C.A. J. Chem. Soc. (Lond.). 1960; : 3293-3301Crossref Scopus (75) Google Scholar, 35Guthrie J.P. J. Am. Chem. Soc. 1977; 99: 3391-4001Crossref Scopus (146) Google Scholar). Retention of substantial catalytic activity suggests that most of the structural features necessary for catalysis are present in the variants and that their overall structure is similar to the wild-type enzyme. Structural similarity between variants is likely becauseVmax/Km values varied by only 10-fold between the most and the least active variants and because 7 of the 9 mutants exhibited differences in Km of less than 5-fold. We separately incubated t-PA, u-PA, or trypsin with the engineered SNase variants. Product separation by polyacrylamide gel electrophoresis demonstrated site-specific proteolysis by t-PA and u-PA of every variant (results not shown), and amino-terminal sequencing confirmed that this cleavage occurred within the inserted sequences at the predicted P1 arginine. Trypsin readily cleaved the variants at other sequences after the initial cleavage within the target loop, whereas u-PA and t-PA did not cleave other sequences even when digestions were allowed to run to completion. To confirm that kinetic data could be obtained using SNase as a substrate we determined the range of concentrations of protease for which cleavage of SNase was linearly dependent on amount of the protease present and used protease concentrations within this range for all subsequent assays. We then determined the range of times in which appearance of cleavage products increased linearly and in which the only measurable cleavage occurred at the engineered target site (Fig. 1, A-D). For trypsin, appearance of product correlates linearly with time for digestions that were stopped prior to cleavage of 8% or less of total substrate (Fig. 2, A and C). For t-PA and u-PA, linear initial rates were maintained until cleavage of 20% of total substrate had occurred (Fig. 1,B and D for t-PA; u-PA data not shown). The linear range for trypsin-mediated hydrolysis is less than that for t-PA or u-PA because, as noted, trypsin has a much greater propensity to cut at other sites after initial hydrolysis within the 44–51 insertion.Figure 2Selective cleavage of SNase variants by t-PA and u-PA at engineered target sequences. Each lane was loaded with 20-μl volumes containing 15 μm wild-type SNase, variant (VII), or variant (VIII). Concentrations of t-PA or u-PA were 75 nm. Each reaction was incubated at 37 °C for 4 h.View Large Image Figure ViewerDownload (PPT) As noted above, Thornton and co-workers (16Hubbard S.J. Campbell S.F. Thornton J.M. J. Mol. Biol. 1991; 220: 507-530Crossref PubMed Scopus (346) Google Scholar, 17Hubbard S.J. Eisenmenger F. Thornton J.M. Protein Sci. 1994; 3: 757-768Crossref PubMed Scopus (213) Google Scholar) have shown that sequences within proteins must deform from their crystallographically determined conformations to engage in productive, “canonical” interactions (i.e. similar to those observed between trypsin and bovine pancreatic trypsin inhibitor) with proteases. The need to deform a sequence prior to binding, and/or cleavage may become rate-limiting for hydrolysis of protein substrates, suggesting that the efficiency of proteolysis of a target sequence within a protein surface loop may be sensitive to small changes in the size of the loop, and that this dependence may differ among related proteases. To test this hypothesis we synthesized 5 substrates with sequences of increasing length replacing residues 44–51. Each substrate contained a 6-residue P4 to P2′ sequence previously shown to be highly labile for cleavage by t-PA (12Ding L. Coombs G.S. Strandberg L. Navre M. Corey D.R. Madison E.L. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7627-7631Crossref PubMed Scopus (89) Google Scholar) and flanking regions that varied from no amino acids, SNase (I), to 4 amino acids, SNase (V) (Table II). For both trypsin and t-PA, SNase (I) was the least efficiently cleaved substrate. Trypsin and t-PA differed, however, in that catalytic efficiency of trypsin-mediated cleavage increased 7.5-fold as the length of the introduced sequences increased from 6 to 10 residues, while efficiency of t-PA-mediated cleavage increased 38-fold (Table II) as the length increased from 6 to 8 residues. Increased efficiency for t-PA-mediated cleavage was largely due to a greater than 15-fold decrease inKm, from ≥140 μm for SNase (I) to 9.3 μm for SNase (V). By contrast, Km for trypsin-mediated cleavage of the same variants decreased less than 3-fold.Table IIKinetic constants for hydrolysis by trypsin and t-PA of a series of SNase variants (I–V) that contain an optimal P4–P2′ t-PA target sequence (PFGRSA) and that vary insert length from 6 to 10 residues by addition of Gly or Pro residues at subsites flanking the P4–P2′ target sequenceInserted sequence (P4,P3,P2,P1↓P1′,P2′)t-PATrypsinkcatKmkcat/KmkcatKmkcat/Kms−1μmm−1s−1s−1μmm−1s−1(I) PFGR↓SA≥0.0041≥14029 ± 1.70.51 ± 0.1431 ± 1016,000 ± 4800(II) PFGR↓SAG0.018 ± 0.00979 ± 47230 ± 1300.72 ± 0.0912 ± 2.460,000 ± 9700(III) PPFGR↓SAG0.011 ± 0.00129.9 ± 3.11100 ± 2300.60 ± 0.0812 ± 3.250,000 ± 9900(IV) PGPFGR↓SAG0.0075 ± 0.00049.9 ± 1.1760 ± 601.0 ± 0.3511 ± 3.090,000 ± 31,000(V) PGPFGR↓SAGG0.01 ± 0.0019.3 ± 2.21100 ± 1801.7 ± 0.4314 ± 5.4120,000 ± 38,000 Open table in a new tab SNase variants (VI) and (VII) were obtained by replacing the wild-type sequence of residues 44–51 with sequences HYGRSG and QRGRSA, which had been identified as t-PA selective within a peptide context (12Ding L. Coombs G.S. Strandberg L. Navre M. Corey D.R. Madison E.L. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7627-7631Crossref PubMed Scopus (89) Google Scholar, 14Ke S.H. Coombs G.S. Tachias K. Navre M. Corey D.R. Madison E.L. J. Biol. Chem. 1997; 272: 16603-16609Abstract Full Text Full Text PDF PubMed Scopus (55) Google Scholar). Proteolysis of the variant nucleases was highly selective with hydrolysis observable only at the inserted sequences (Fig. 2). Both variants (VI) and (VII) were hydrolyzed by t-PA with akcat/Km value of 670m−1 s−1 (Table III), similar tokcat/Km values of 322 and 850m−1 s−1 for hydrolysis of the respective analogous peptide substrates (12Ding L. Coombs G.S. Strandberg L. Navre M. Corey D.R. Madison E.L. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7627-7631Crossref PubMed Scopus (89) Google Scholar, 14Ke S.H. Coombs G.S. Tachias K. Navre M. Corey D.R. Madison E.L. J. Biol. Chem. 1997; 272: 16603-16609Abstract Full Text Full Text PDF PubMed Scopus (55) Google Scholar). There are striking differences for the individual constants kcatand Km for protein versus peptide cleavage, and these will be discussed below. u-PA proteolyzed the variants less efficiently than t-PA withkcat/Km values 230-fold less than t-PA for (VI) and 92-fold less than t-PA for (VII). For (VI) this decrease was largely due to a decreased kcat (136-fold), while the decreased value for (VII) was more evenly influenced by bothkcat and Km. The 230- and 92-fold t-PA/u-PA selectivities for cleavage of HYGRSG and QRGRSA in a protein context compare favorably with the 21- and 19-fold selectivities observed in the peptide context. To further the comparison between t-PA and u-PA we examined cleavage of variants (I) and (V) by u-PA. These were characterized bykcat/Km values of 1.5 and 25m−1 s−1 respectively, affording t-PA/u-PA selectivities of 19-fold for (I) and 44-fold for (V).Table IIIKinetic parameters for hydrolysis by t-PA and u-PA of SNase variants containing t-PA or u-PA specific target sequences identified in peptidesInserted sequence (P4,P3,P2,P1↓P1′,P2′)t-PAu-PAkcatKmkcat/KmkcatKmkcat/Kms−1μmm−1s−1s−1μmm−1s−1t-PA selective sequences (VI)PHYGR↓SGG0.012 ± 0.00318 ± 7.0670 ± 2208.8(10)−5 ± 1.7(10)−530 ± 9.32.9 ± 0.73 (VII) PQRGR↓SAG0.006 ± 0.0038.9 ± 0.6670 ± 190≥0.0009≥1307.3 ± 1.1u-PA selective sequences (VIII) PGSGR↓SAG3.7(10)−4 ± 5.6(10)−517 ± 4.722 ± 4.70.013 ± .00353 ± 21250 ± 75 (IX) PGSGR↓SASGTTGTG≥0.0015≥8019 ± 2.5≥0.063≥70900 ± 120 Open table in a new tab We also assayed SNase variants (VIII) and (IX) containing the u-PA specific target sequence GSGRSA in loops of varying size. u-PA cleaved both substrates more efficiently than did t-PA (Table III, Fig. 2). The variant containing GSGRSA within an 8-residue insert (VIII) displayed selectivity for u-PA cleavage of 12-fold, whereas the variant containing a 14-residue insert was cleaved by u-PA with a higher efficiency (kcat/Km = 900m−1 s−1) and with a u-PA/t-PA selectivity of 47-fold. The analogous sequence in a peptide context was cleaved with a u-PA/t-PA specificity of 63 so that, by contrast to cleavage of t-PA selective
Referência(s)