Structural determinants of procryptdin recognition and cleavage by matrix metalloproteinase-7.

The bactericidal activity of mouse Paneth cell alphadefensins, or cryptdins, is dependent on processing of cryptdin precursors (pro-Crps) by matrix metalloproteinase-7 (MMP-7) (Wilson, C. L., Ouellette, A. J., Satchell, D. P., Ayabe, T., Lopez-Boado, Y. S., Stratman, J. L., Hultgren, S. J., Matrisian, L. M., and Parks, W. C. (1999) Science 286, 113-117). To investigate the mechanisms of pro-Crp processing by this enzyme, recombinant pro-Crp4, a His-tagged chimeric pro-Crp (pro-CC), and site-directed mutant precursors of each were digested with MMP-7, and the cleavage products were analyzed by NH(2)-terminal peptide sequencing. Proteolysis of pro-Crp4 with MMP-7 activated in vitro bactericidal activity to the level of the mature Crp4 peptide by cleaving pro-Crp4 at Ser(43) downward arrow Ile(44) and Ala(53) downward arrow Leu(54) in the proregion and near the Crp4 peptide NH(2) terminus between Ser(58) downward arrow Leu(59). Because the Crp4 NH(2) terminus occurs at Gly(61), not Leu(59), MMP-7 is necessary but insufficient to complete the processing of Crp4. Crp activating proteolysis at S58 downward arrow L59 was unaffected by I44S/I44D or L54S/L54D loss-of-function mutations in pro-Crp4, and a (L59S)-pro-CC mutant was cleaved normally at Ser(43) downward arrow Val(44) and Ser(53) downward arrow Leu(54) sites but not at the peptide NH(2) terminus. C57BL/6 mice contain an abundant (L59S)-Crp4 mutant peptide with Leu(54) at its NH(2) terminus resulting from Ala(53) downward arrow Leu(54) cleavage and loss-of-function at the Ser(58) downward arrow Ser(59) cleavage site. Thus, alpha-defensins resulting from mutations at MMP-7 cleavage sites exist in mouse populations. A pro-CC substrate containing both L54S and L59S mutations resisted cleavage at Ser(43) downward arrow Val(44) completely, showing that cleavage at one or both downstream sites must precede proteolysis at Ser(43) downward arrow Val(44). These findings show that MMP-7 activation of pro-Crps can occur without proteolysis of the proregion, and prosegment fragmentation depends, at least in part, on the release of the Crp peptide from the precursor.

The release of endogenous antimicrobial peptides by mammalian epithelial cells contributes to innate mucosal immunity (1,2). In the small intestine of most mammals, Paneth cells that reside at the base of the crypts synthesize and secrete microbicidal ␣-defensins, termed cryptdins in mice, as components of apical secretory granules (3)(4)(5)(6). The granules are released by Paneth cells in response to cholinergic agonists or when exposed to bacterial stimuli (7)(8)(9)(10). Cryptdin peptides constitute ϳ70% of the bactericidal peptide activity released by mouse Paneth cells, and cryptdin concentration at the point of secretion is at least 1000 times greater than the minimal bactericidal concentration of the peptides (9). The production of functional ␣-defensins involves proteolytic processing of inactive precursor forms by mechanisms that differ between mice and humans (5,11).
In mouse small intestine, MMP-7 is expressed only by Paneth cells (17), where the enzyme activates all ␣-defensins from 8.4-kDa proforms (18). Previously, both procryptdin-1 (pro-Crp1) and a COOH-terminal His 6 tagged pro-Crp chimera (pro-* This work was supported by National Institutes of Health Grants DK10184 (to D. P. S), DE14040 (to C. L. W.), and DK44632 (to A. J. O.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. CC) containing sequence from pro-Crp1, pro-Crp4, and pro-Crp15 were shown to be activated to 3.5 kDa ␣-defensins in vitro by MMP-7-catalyzed cleavage at conserved sites in the proregion and at the junction of the propeptide and the NH 2 terminus of the mature cryptdin peptide (5,18). In those studies, MMP-7 was found to cleave between Ser 43 2Val 44 in the prosegment and at Ser 58 2Leu 59 , where Leu 59 is the NH 2terminal residue for all known mouse cryptdins except Crp4 and Crp5 (5). Additional preliminary evidence showed that MMP-7 cleaved pro-Crp4 at Ala 53 2Leu 54 (5), a site that corresponds to processing intermediates isolated from mouse small intestine (19). Thus, mouse pro-Crps contain conserved sites within the precursor proregion that MMP-7 recognizes and cleaves, but their role in pro-Crp activation is uncharacterized.
In this study, cryptdin biosynthesis was investigated by analyzing the processing of recombinant pro-Crps by MMP-7 in vitro. We focused on Crp4, because it is the most bactericidal mouse ␣-defensin, and comparisons of pro-Crp4 cleavage with cleavage of the pro-Crp chimera pro-CC allowed the generality of MMP-7 site usage to be assessed. The products of in vitro pro-Crp cleavage by MMP-7 and the effects of eliminating the proregion processing sites by site-directed mutagenesis have been determined. The results show that cryptdin activation by cleavage of the peptide bond at the peptide NH 2 terminus is independent of proteolysis within the prosegment; however, cleavage upstream at Ser 43 2Val 44 is dependent on cleavage downstream at the Ser 58 2Leu 59 or Ala/Ser 53 2Leu 54 sites. Also, C57BL/6 mice accumulate Crp4-related processing intermediates that result from loss-of-function mutations at MMP-7 processing sites.

EXPERIMENTAL PROCEDURES
Animals and Tissue Preparation-All procedures on mice were performed with approval and in compliance with the policies of the Institutional Animal Care and Use Committees of the University of California, Irvine (UCI), and Washington University School of Medicine. Sixweek-old adult male C57BL/6 mice were purchased from Charles River Breeding Laboratories, Inc. (North Wilmington, MA). MMP-7 null mice were 6 -8-week-old males backcrossed for 10 generations onto the C57BL/6 background. Mice were housed in a specific pathogen-free facility under 12 h cycles of light and dark and had free access to standard rat chow and water.
Purification of Recombinant Mouse Pro-Crp4 -Recombinant pro-Crp4 and pro-Crp4 variants with mutated MMP-7 recognition sites were prepared using the pET-28a expression system to produce NH 2terminal His 6 -tagged fusion proteins (Novagen, Madison, WI). By PCR amplification, a Met-coding trideoxynucleotide was incorporated 5Ј of codon 20 in the Crp4 precursor cDNA and cloned in-frame with the amino-terminal His 6 in the EcoRI/SalI sites of pET-28a. For cloning pro-Crp4, forward primer pETPCr4-F (5Ј-GCGCGAATTCATGGATCC-TATCCAAAACACA) was paired with reverse primer SLpMALCrp4R (5Ј-ATATATGTCGACTGTTCAGCGGCGGGGGCAGCAGTACAA), corresponding to nucleotides 104 -119 and 301-327 in prepro-Crp4 cDNA (20). Reactions were performed using the GeneAmp PCR Core Reagents (Applied Biosystems, Foster City, CA) by incubating the reaction mixture at 95°C for 5 min, followed by successive cycles at 60°C for 1 min, 72°C for 1 min, and 94°C for 1 min for 40 cycles. The underlined codon in the pETPCr4-F primer denotes a Met codon that was introduced immediately upstream of the pro-Crp4 NH 2 -terminal Asp residue to incorporate a CNBr cleavage site.
Following PCR amplification, samples (25 l) of individual reactions were gel purified using 2% agarose gels, and extracted using QIAEX II (Qiagen Inc., Valencia CA). In most cases, amplification products were cloned in pCR2.1-TOPO, sequenced, digested with EcoRI and SalI, and gel-purified EcoRI/SalI inserts were ligated into EcoRI and SalI-digested pET-28a plasmid DNA, and transformed into both Escherichia coli XL-2 Blue and BL21(DE3) Codon Plus cells (Stratagene Cloning Systems, Inc., La Jolla, CA).
Recombinant proteins were expressed for 6 h at 37°C in E. coli BL21(DE3) Codon Plus cells growing exponentially in Terrific Broth medium by induction with 0.2 mM isopropyl-1-thio-␤-D-galactopyranoside under kanamycin selection. Bacterial cells were harvested by centrifugation and stored at Ϫ20°C. Cells were lysed in 6 M guanidine HCl, 100 mM Tris-Cl (pH 8), and clarified by centrifugation in a Sorvall SA-600 rotor at 30,000 ϫ g for 30 min at 4°C. Fusion proteins were purified immediately after lysate clarification.
Recombinant precursor fusion proteins were purified by nickel-nitrilotriacetic acid (Ni-NTA, Qiagen) resin affinity chromatography and recovered from fusions after CNBr cleavage. His-tagged fusion proteins were eluted from Ni-NTA resin with 2 column volumes of buffer consisting of 6 M guanidine HCl, 1 M imidazole, and 100 mM Tris-HCl (pH 6.0). Fusion proteins, containing an NH 2 -terminal His 6 tag, a 26-amino acid spacer, and pro-Crp4, were dialyzed against 5% acetic acid and lyophilized. Lyophilized proteins dissolved in 80% formic acid were adjusted to 10 mg/ml CNBr in 80% formic acid and incubated under N 2 overnight at room temperature. Cleavage was terminated by addition of 10 volumes of H 2 O, proteins were lyophilized, dissolved in 5% acetic acid, and purified by C-18 RP-HPLC by eluting peptides over 120 min with an 20 -40% acetonitrile gradient. The identities of recombinant pro-Crp4 molecules were verified by MALDI-TOF MS at the UCI Biomedical Protein and Mass Spectrometry Resource Facility and by acidurea PAGE (21).
Site-directed Mutagenesis of Pro-Crp4 -Mutations were introduced into recombinant pro-Crp4 molecules by PCR. In reaction 1, the mutant forward primer, e.g. pc4I44Dfor, containing the mutant codon flanked by three natural codons was paired with SLpMALCrp4R, the normal reverse primer at the 3Ј-end of the desired sequence. In reaction 2, the mutant reverse primer, e.g. pc4I44Drev, the exact complement of the mutant forward primer, was paired with the normal forward primer pETPCr4-F at the 5Ј-end of the desired sequence, and sequences were amplified from the pET-28a pro-Crp4 construct as described above: 95°C for 5 min, followed by successive cycles at 60°C for 1 min, 72°C for 1 min, and 94°C for 1 min for 40 cycles. Products from reactions 1 and 2 were purified electrophoretically, and 0.5-l samples of gelpurified DNA were combined as templates in PCR reaction 3, using normal external primers, SLpMALCrp4R and pETPCr4-F, as amplimers. The full-length, mutated pro-Crp4 product of reaction 3 was cloned sequentially into the vectors pCR2.1-TOPO and pET-28a as noted above, and all mutations were verified by DNA sequencing prior to expression.
Purification of Recombinant Mouse Chimeric Pro-CC-The pro-CC construct for production in the baculovirus expression system has been described (18). Briefly, prepro-Crp15 cDNA was amplified using a forward primer derived from sequence encoding the signal peptide and a reverse primer that changed a Met residue in the COOH terminus of the mature peptide to a Thr (characteristic of Crp1). In addition, the COOH-terminal Arg residue of the precursor was converted to Pro-Arg-Arg-His-His-His-His-His-His; Pro-Arg-Arg is the Crp4 COOH-terminal sequence. The amplified sequence was cloned into the transfer vector pVL1393 and transfected into Sf9 insect cells along with BaculoGold DNA (BD Pharmingen, La Jolla, CA) to produce recombinant baculovirus. Following a 4 -5-day infection of HighFive insect cells (Stratagene) with recombinant baculovirus, cells were harvested by centrifugation and stored at Ϫ20°C. Pellets were thawed and lysed in 6 M guanidine HCl, 100 mM sodium phosphate (pH 8), and 10 mM Tris-HCl (pH 8).
Lysates were passed several times through an 18-gauge needle and centrifuged at 4°C in a Sorvall SS-34 rotor at 12,000 ϫ g. Supernatants were incubated batchwise at room temperature with Ni-NTA resin (Qiagen) for 1-3 h with mixing. Bound His-tagged precursor was eluted with 8 M urea, 100 mM sodium phosphate (pH 4.5), 10 mM Tris-HCl (pH 4.5), and 1% Triton X-100 in 0.5-1.0-ml fractions. Peak fractions were pooled and dialyzed against 50 mM NaCl, 20 mM Tris-HCl (pH 7.5). Protein concentration and purity were assessed by reducing Tris-Tricine SDS-PAGE and GelCode Blue (Pierce) staining. If further purification was required, proteins were concentrated to 0.6 ml using Centricon YM-3 centrifugal filter devices and subjected to gel filtration chromatography on a Superdex 75 HR/10/30 column connected to an Ä KTAFPLC system (Amersham Pharmacia Biotech).
Site-directed Mutagenesis of Pro-CC-As with pro-Crp4, mutations were introduced into recombinant pro-CC by PCR. To generate the L59S mutation, the forward primer used originally to construct pro-CC, CRPBam-F (5Ј-GCGGATCCTCCTGCTCACCAATCCTCCA-3Ј) was paired with the mutant reverse primer CRP-L59S (5Ј-ACCAGATCTC-TCGACGATTCCTCT-3Ј), where the underlined sequence corresponds to BamHI and BglII restriction sites, respectively. Conditions for amplification from the pro-CC template were as follows: 94°C for 5 min, 30 cycles of 94°C for 1 min, 53°C for 1 min, 72°C for 2 min, and a final extension at 72°C for 15 min. The 220-bp product was cloned into the pCR 2.1-TOPO vector (Invitrogen); the BamHI-BglII fragment from this plasmid construct was used to replace the BamHI-BglII fragment in the pro-CC cDNA in pGEM-7Zf(ϩ) (Promega). The full-length (L59S)-pro-CC cDNA, flanked by BamHI and EcoRI sites, was then transferred to the pVL1393 baculovirus expression vector using these two restriction enzymes. To add the L54S mutation, pGEM7-(L59S)pro-CC was used as a template for forward primer CRPBam-F and mutant reverse primer CRPL54/59SBglII-R containing a BglII restriction site (underlined) (5Ј-AGATCTCTCGACGATTCCTCTTGACTAGA-AGAGCC-3Ј). Amplification parameters were similar to those used for (L59S)-pro-CC except that the annealing temperature was increased to 65°C and the number of cycles was reduced to 25. The 220-bp product was subcloned into pVL1393 as described above for pro-CCL59S. To eliminate only the Ser 43 2Val 44 and Ser 53 2Leu 54 MMP-7 cleavage sites in pro-CC, a two-step process was used. First, the forward primer CRPBam-F was paired with the mutant reverse primer CRPL54SBg-lII-R (5Ј-AGATCTCTCAACGATTCCTCTTGACTAGAAG AGCC-3Ј), where the BglII restriction is underlined. The PCR product was subcloned and transferred to pGEM7-pro-CC as described above. In the second step, this new construct, pGEM7-(L54S)-pro-CC, was used as a template for the mutant forward primer CRPV44SbbsI-F (5Ј-GAAGA-CGACCAGGCTGTGTCTTCCTCTTTTGGAGAC-3Ј), where the BbsI restriction site is underlined, and a normal downstream primer containing a PstI restriction site (underlined), CRPPstI-R (5Ј-CTGCAGG-TCCCATTTATGTGT-3Ј). The 130-bp product was cloned into pCR2.1-TOPO; the BbsI-PstI fragment from this construct was used to replace the BbsI-PstI fragment in pGEM7-(L54S)-pro-CC. Full-length cDNA from the resulting plasmid, pGEM7-(V44SL54S)-pro-CC, was subcloned into the BamHI and EcoRI sites of pVL1393. All mutations were verified by DNA sequencing prior to expression.
Purification of Crp4 and Pro-Crp4 Variants from C57BL/6 and MMP-7-null Mice-A naturally existing ␣-defensin that has a L59S mutation was purified from C57BL/6 mice, and the corresponding precursor to that Crp4(B6a) variant was isolated from MMP-7 null mouse small intestine by extraction with 30% acetic acid as described (5,18). Protein samples were applied to analytical C-18 RP-HPLC columns (Vydac 218TP54) in aqueous 0.1% trifluoroacetic acid and eluted at ϳ35 min using a 10 -45% acetonitrile gradient developed over 55 min. Protein fractions containing apparent pro-Crp4 were analyzed by acid urea-(AU)-PAGE as described (18,21), and their identities were deduced from a combination of NH 2 -terminal peptide sequencing, MALDI-TOF-MS, and comparisons with the corresponding cDNA sequence in RIKEN (accession number AK008107). Purification of the Crp4 mutant and corresponding pro-Crp was completed subsequently by C-18 reverse phase-HPLC using a 120-min, 20 -40% acetonitrile gradient, from which cryptdin precursors eluted between 23 and 30% acetonitrile. Peptide concentrations were determined using bicinchoninic acid assay (Pierce, Rockford, IL).
Acid Urea-Polyacrylamide Gel Electrophoresis-Peptide samples were lyophilized, dissolved in 20 l of 5% acetic acid containing 3.0 M urea, and electrophoresed on 12.5% AU-PAGE for 1 h at 100 V and for 3.5 h at 250 V (21). Resolved proteins were visualized by staining with Coomassie R-250 after fixation in formalin-containing acetic acid/methanol. Crp4 and pro-Crp4 were identified by co-migration with authentic mouse Crps and pro-Crps in AU-PAGE (Ͼ0.6 ϫ R F of methyl green dye) as described (22) and confirmed by MALDI-TOF-MS and NH 2terminal sequencing.
Protein Analysis by Mass Spectrometry-For reduction and alkyla-tion, recombinant peptides dissolved at 500 g/ml in 6 M guanidine HCl, 100 mM Tris-HCl (pH 8.0) were reduced with dithiothreitol at 50°C for 3-4 h using 5 mol of dithiothreitol per mol of polypeptide cysteine. After cooling, a 3-fold mass excess of iodoacetic acid to dithiothreitol was added to the reduced peptide solution, incubated for 10 min, and residual iodoacetic acid was reacted with excess dithiothreitol. The native and alkylated peptides were purified on C-18 RP-HPLC, and the molecular masses were determined by MALDI-TOF MS, followed by NH 2terminal sequencing in the UCI Biomedical Protein and Mass Spectrometry Resource Facility. MMP-7 Cleavage of Mouse Pro-Crps in Vitro-Recombinant pro-Crp4 and pro-CC molecules were digested with MMP-7, analyzed by AU-PAGE and SDS-PAGE, and samples of the proteolytic digests were analyzed by NH 2 -terminal sequencing. Samples (1 g) of pro-Crp4 and all variants, as well as pro-CC and corresponding variants, were incubated with activated recombinant human MMP-7 (0.3-1.0 g) catalytic domain (Calbiochem, La Jolla, CA, or Chemicon International, Inc., Temecula, CA) in buffer containing 10 mM HEPES (pH 7.4), 150 mM NaCl, 5 mM CaCl 2 for 18 -24 h at 37°C. Samples of pro-Crp4 digests were analyzed by AU-PAGE, and ϳ200 ng quantities of complete digests were subjected to 5 or more cycles of NH 2 -terminal peptide sequencing at the UCI Biomedical Protein and Mass Spectrometry Resource Facility. Pro-CC digests were analyzed by Tris-Tricine SDS-PAGE (15% polyacrylamide), and bands were visualized using GelCode TM Blue staining reagent (Pierce). For NH 2 -terminal sequencing by Edman degradation, 3 g of precursor incubated with MMP-7 were separated by Tris-Tricine SDS-PAGE and transferred to and visualized on mini-ProBlott polyvinylidene difluoride membranes (Applied Biosystems) according to the manufacturer's instructions. Also, samples of the Crp1 prosegment corresponding to residues 19 -58 in prepro-Crp1 (23,24) were digested with MMP-7 and sequenced. The Crp1 proregion consisting of the primary structure, DPIQNTDEETKT-EEQPGEDDQAVSVSFGDPEGTSLQEES, was synthesized by Quality Controlled Biochemicals, Inc., (Hopkinton, MA). The composition and properties of the synthetic prosegment has been reported previously (5).
Bactericidal Peptide Assays of Pro-Crp4 Activated by MMP-7-The activation of pro-Crp4 was assayed by conducting bactericidal peptide assays with pro-Crp4 following digestion with MMP-7 under conditions of quantitative cleavage at Leu 59 (above). Samples consisting of exponentially growing bacterial cells (ϳ1 ϫ 10 6 colony forming units (CFU)/ ml) were incubated with 0 -20 g/ml Crp4 or pro-Crp4 with or without MMP-7-mediated proteolysis (above). After 60 min at 37°C, 20 l of each incubation mixture was diluted 1:1000 with 10 mM PIPES (pH 7.4), and 50 l of the diluted samples were plated on trypticase soy agar using a Spiral Biotech Autoplate 4000 (Spiral Biotech Inc., Bethesda, MD). Surviving bacteria were quantitated as colony forming units on plates after incubation at 37°C for 12 h.

MMP-7-mediated Cleavage of Pro-Crp4 -Pro-Crp4
is cleaved by MMP-7 at sites corresponding to those identified previously in natural pro-Crps (5). Recombinant pro-Crp4 molecules expressed in E. coli were purified to homogeneity by RP-HPLC (Fig. 1). Because the pro-Crp4 protein lacks methionine, CNBr provided a means for quantitative chemical cleavage of affinity purified fusion proteins from which the pro-Crp4 component could be separated from the His-tagged fusion partner by sequential RP-HPLC fractionation. The mass of the pro-Crp4 molecule was verified by MALDI-TOF-MS to be 8231 atomic mass units, and its homogeneity was judged by analytical RP-HPLC, AU-PAGE, and NH 2 -terminal sequencing. The sequence DPIQNT . . . , the consensus for mouse pro-␣-defensins, was the only NH 2 terminus detected. Also, cleavage of recombinant pro-Crp4 with MMP-7 in vitro produced a single evident product that migrated only slightly slower than Crp4 in AU-PAGE gels and was comparable with Crp4 in bactericidal activity (Fig. 1).
Specificity of in Vitro Pro-Crp-4 Cleavage by MMP-7-Previously, Paneth cell ␣-defensin precursor substrates isolated from MMP-7 null mouse small intestine were cleaved in vitro with MMP-7, identifying cleavage sites at Ser 43 2Val 44 and Ser 58 2Leu 59 , the latter site corresponding to the NH 2 terminus of the known mouse cryptdins except for Crp4 and Crp5 (5,24). In addition, MALDI-TOF-MS analyses of partially purified mouse intestinal proteins had identified an apparent pro-Crp processing intermediate with Leu 54 as its amino terminus (19). Because Crp4 is the most potent of the mouse ␣-defensins, and the Crp4 proregion differs from other pro-Crps at several positions at or proximal to the predicted MMP-7 cleavage sites (Fig.  2), we tested whether MMP-7 processed pro-Crp4 in vitro and whether the products would correspond to those produced by hydrolysis of pro-Crp1. NH 2 -terminal peptide sequence analysis of MMP-7 digests of pro-Crp4 showed that pro-Crp4 contains cleavage sites common to natural pro-Crps (5,19). Consistently, four NH 2 termini were detected in the digests: DPIQ . . . , ISFG . . . , LHEKS, and LRGLL_Y_ . . . , where the underscores denote empty sequencing cycles characteristic of Cys residues. These NH 2 termini correspond to Asp 20 at the pro-Crp4 NH 2 terminus, Ile 44 , Leu 54 , and Leu 59 in the polypeptide chain (Fig. 2). These NH 2 termini result from MMP-7 cleavage of pro-Crp4 peptide bonds at Ser 43 2Ile 44 and Ala 53 2Leu 54 in the proregion and at Ser 58 2Leu 59 near the Crp4 peptide NH 2 terminus. When natural pro-Crp6 purified from MMP-7-null mouse small intestine was analyzed similarly after MMP-7 digestion, the NH 2 termini detected were DPIQNT . . . , VSFGDP . . . LQEES, and LRDLV . . . , the same positions identified in pro-Crp4 and in previous reports (19) (Fig. 2) (5). No other MMP-7 cleaved sites were apparent in these studies. Also, MMP-7 digested the 39-residue, synthetic Crp1 proregion at Ser 43 2Val/Ile 44 and Ser/Ala 53 2Leu 54 , the same positions cleaved in pro-Crp4 and pro-Crp6 (Fig. 2). This finding shows that the specificity of MMP-7 cleavage of proregion sites is independent of the COOH-terminal presence of an ␣-defensin. Because the NH 2terminal amino acid of natural Crp4 is Gly 61 , not Leu 59 , MMP-7 proteolysis is not sufficient to catalyze complete activation of pro-Crp4 to the fully mature Crp4 peptide. The Arg 60 2Gly 61 cleavage step does, however, require prior hydrolysis at Ser 58 2Leu 59 by MMP-7, because MMP-7 null mice lack mature Crp4 with Leu 59 or Gly 61 NH 2 termini (5,18). The enzyme that removes the Leu 59 -Arg 60 dipeptide at Arg 60 2Gly 61 is unknown but may be a trypsin-like aminodipeptidase.
MMP-7 Cleavage of Pro-Crp4 Activates Crp4 Bactericidal Activity-To test whether MMP-7 mediated pro-Crp4 proteolysis results in the production of functional Crp4 peptide, bactericidal peptide activity assays were performed on MMP-7-digested pro-Crp4. Salmonella typhimurium PhoP(Ϫ) cells were combined with 0 -3 M Crp4 or pro-Crp4 that had been digested with a 0.2 mol eq of MMP-7 or incubated without enzyme ("Experimental Procedures"). No bacterial cell killing was detected when cells were exposed to MMP-7 alone (not shown) or to intact pro-Crp4 (Fig. 3). Conversely, quantitative Ser 58 2Leu 59 hydrolysis near the Crp4 peptide NH 2 terminus activated pro-Crp4 to a bactericidal peptide activity level that was equivalent to equimolar quantities of mature Crp4 with the Gly 61 NH 2 terminus (Fig. 3). Consistent with the biochemical evidence of pro-Crp4 proteolysis by MMP-7 ( Figs. 1 and 2), MMP-7 cleavage produces a functional Crp4 peptide from its inactive precursor. These findings provided rationale for testing whether structural determinants of those cleavage events exist in the precursor molecule.
Disulfide Bonds Protect the Crp4 Peptide from MMP-7 Proteolysis during Pro-Crp4 Activation-The formation of disulfide bonds within the pro-Crp4 molecule confers protection of the Crp4 peptide from MMP-7-mediated proteolysis. To determine whether the specificity of MMP-7 recognition and hydrolysis would be modified by disrupting the disulfide array of the Crp4 peptide region in pro-Crp4, the MMP-7 digestion products of native, reduced, and reduced and alkylated pro-Crp4 were characterized ("Experimental Procedures"). The extent of Cys alkylation in reduced and alkylated pro-Crp4 was confirmed by MALDI-TOF-MS, which showed that the mass of reduced and alkylated pro-Crp4 was 350.22 atomic mass units greater than that of native pro-Crp4 (theoretical increase ϭ 348 atomic mass units), an indication that the 6 Cys residues in pro-Crp4 had been acetylated. The specificity of proteolysis at MMP-7 sites within the prosegment was the same for native pro-Crp4 and reduced and alkylated pro-Crp4 (Fig 2). However, MMP-7 also cut reduced pro-Crp4 and reduced and alkylated (L54D)-Crp4 substrates at Leu 62 2Leu 63 , Cys 64 2Tyr 65 , and Phe 87 2Leu 88 , all sites within the polypeptide backbone of the Crp4 ␣-defensin moiety (Fig. 2). From the relative recovery of individual NH 2 termini, the Phe 87 -Leu 88 peptide bond in reduced and alkylated pro-Crp4 appears to be more susceptible to MMP-7 proteolysis than the Leu 62 2Leu 63 and Cys 64 2Tyr 65 cleavage sites that were cleaved to similar extents (data not shown). In more than 20 MMP-7 digests of native pro-Crp4, pro-Crp4 variants, and other pro-Crps isolated without reduction, MMP-7 cleavage events in the Crp4 peptide had not been detected (Figs. 2, 4-6). These findings are evidence that the Crp4 tridisulfide array protects the peptide from MMP-7-mediated proteolysis during pro-Crp4 processing.
The activating pro-Crp4 cleavage step at Ser 58 2Leu 59 does not require cleavage to occur at either MMP-7 recognition site in the proregion. Because MMP-7 processed the Ser 58 2Leu 59 site when the Ser 43 2Ile 44 or Ala 53 2Leu 54 sites were mutagenized individually (Fig. 4B), pro-Crp4 lacking both prosegment cleavage sites was prepared by introducing an L54D mutation in (DEDED)-pro-Crp4. The resulting (DEDED/ L54D)-pro-Crp4 molecule was digested quantitatively by MMP-7 at Ser 58 2Leu 59 (Fig. 4B). Sequencing MMP-7 digests of (DEDED/L54D)-pro-Crp4 failed to detect new NH 2 termini that might have been produced by alternative cleavage events associated with the mutagenesis. Thus, even though MMP-7 cleaves the pro-Crp4 proregion with specificity, activation of Crp4 in vitro is independent of the processing steps in the prosegment.
The effects of Ser mutations at Leu 54 and Leu 59 on MMP-7 recognition of the upstream Val 44 site (Ile 44 in pro-Crp4) were tested further by analyzing the MMP-7 digestion products of additional chimeric recombinant pro-Crp molecules. For example, pro-CC is a chimeric pro-Crp derived from pro-Crp15, which differs from pro-Crp1 at one amino acid position in the proregion and at 3 residue positions in the mature ␣-defensin peptide (Fig. 5A) (18,24). The molecule also contains a COOHterminal His 6 tag linked to a PRR tripeptide sequence that was introduced immediately following the COOH-terminal cysteines in the Crp15 sequence ("Experimental Procedures," Fig.  5A). In previous studies, the patterns of pro-CC and pro-Crp1 cleavage by MMP-7 were found to be similar, in that cleavage at Ser 58 2Leu 59 yielded mature peptides for both precursors (18). Recombinant pro-CC molecules containing an L59S mutation, (L59S)-pro-CC, or both L54S and L59S mutations, (L54S/L59S)-pro-CC (Fig. 5A), were produced in the baculovirus expression system and affinity purified from insect cell lysates ("Experimental Procedures"). As shown by SDS-PAGE and NH 2 -terminal sequencing, MMP-7 cleaved (L59S)-pro-CC at Ser 53 2Leu 54 but not at Ser 58 2Ser 59 (Fig. 5B). Furthermore, MALDI-TOF analyses of the digests confirmed cleavage in the proregion at Ser 43 2Val 44 in both the pro-CC and (L59S)pro-CC substrates. However, when both Leu 54 and Leu 59 were converted to Ser in the (L54S/L59S)-pro-CC molecule, the Ser 43 2Val 44 site no longer was digested by MMP-7 (Fig. 5C). The (L54S/L59S)-pro-CC double mutant resisted cleavage by MMP-7 completely, and no alternate sites to Ser 43 2Val 44 proteolysis could be detected. Consistent with analysis of MMP-7digested (DEDED/L54D)-pro-Crp4 (Fig. 3B), V44S and L54S mutations in the upstream sites of the pro-CC proregion did not affect the activating cleavage step at Leu 59 in pro-CC as determined by NH 2 -terminal sequencing of MMP-7 digests (data not shown). Thus, MMP-7 hydrolysis of the Ser 43 2Val 44 site requires (a) that cleavage events at Leu 54 , Leu 59 , or both, release a nearly full-length prosegment from pro-CC and (b) that those cleavage steps precede hydrolysis of the Ser 43 -Val 44 peptide bond.
Mutation of the Ser 58 2Leu 59 Processing Site in C57BL/6 Mouse Pro-Crp4 Variants-Paneth cells in C57BL/6 mice contain a Crp4 variant peptide with an L59S mutation that abrogates cleavage at that position by MMP-7. Comparative AU-PAGE analyses of intestinal protein extracts showed that the abundant apparent ␣-defensins of C57BL/6 mice do not comigrate with Crps from inbred or outbred strains of mice (Fig.  6A). To investigate the structural basis for the differences, the distinctive C57BL/6 Crp peptides were isolated by RP-HPLC and identified as potential Crps by co-migration with Crp markers in AU-PAGE and by MALDI-TOF MS (Fig. 6B). Comparisons of selected peptides before and after reduction and alkylation with iodoacetamide identified molecules with 6 Cys residues as peptides with masses that increased by 344.9 (theoretical value ϭ 348 atomic mass units when acylated).
Two novel Crp4-related peptides purified from C57BL/6 mouse small intestine were characterized by NH 2 -terminal peptide sequencing. The NH 2 terminus of the abundant peptide, Crp4(B6a), was LHEKSSRDLI_Y_RKGG_N-RGEQVYGT_ . . . , where the underscore characters denote deduced Cys residues (Fig. 6C). Analyses of Paneth cell secretory granules partially purified from mouse crypts showed that Crp4(B6a) was present at the same relative concentration as in extracts of intact small bowel, consistent with the localization of all known ␣-defensins in mouse small intestine. In contrast to the abundance of Crp4(B6a), the Crp4(B6b) peptide was recovered only at very low levels from C57BL/6 mouse small bowel, and the NH 2 -terminal sequence of the peptide was LSRDLI_L_RNR . . . (Fig. 6B). A BLASTP search revealed that both sequences corresponded to Crp4-related small intestinal cDNAs in the RIKEN data base (accession numbers AK008107 and AK008266, respectively). Alignment of the peptide sequences, their deduced pro-Crps and pro-Crp4 identified the peptides as Crp4 variants. The Crp4(B6a) and Crp4(B6b) proregions resemble the Crp4 prosegment more closely than other Crp prosegments (Fig. 6C), and both peptides lack 3 amino acid residues between the fourth and fifth cysteine residue positions. That 3-codon deletion displaces amino acids that are present in all other ␣-defensins, and the deletion, found only in Crp4, would shorten the loop formed by the Cys 3 -Cys 5 disulfide bond (23)(24)(25). Previously, extensive sequencing of intestinal cDNAs and genomic clones from outbred Swiss mice, and C3H/ HeJ, BALB/cJ, and 129/SvJ inbred mouse strains had not de- FIG. 4. MMP-7 proteolysis of pro-Crp4 variants with site-directed mutations in MMP-7 processing sites. Panel A, recombinant pro-Crp4 peptides were mutagenized, expressed, and purified as described ("Experimental Procedures") and judged to be homogeneous by detection of only one terminus by NH 2 -terminal peptide sequencing (data not shown) and by nonreducing AU-PAGE analysis. Samples containing 2 g of protein/lane were electrophoresed in AU-PAGE and stained with Coomassie Blue. Lanes: 1, pro-Crp4; 2, (I44D)-pro-Crp4; 3, (I44S)-pro-Crp4; 4, (L54D)-pro-Crp4; 5, (L54S)-pro-Crp4; 6, (DEDED)-pro-Crp4; 7, (DEDED/L54D)-pro-Crp4; 8, Crp4. Upper arrow at the right indicates pro-Crp4 and pro-Crp4 variants; lower arrow points to a mature Crp4 peptide standard. Panel B, samples consisting of 2 g each of pro-Crp4 and pro-Crp4 variants were incubated overnight with 0.5 mol eq of MMP-7 (Fig. 1), and digests were analyzed by 5 cycles of NH 2 -terminal peptide sequencing. Cleavage sites disclosed by protein sequencing are noted by downwards arrows (2) that interrupt the individual sequences. Numerals above the (I44D)-Crp4 primary structure refer to residue positions at the beginning of each detected NH 2 terminus, numbered with the initiating Met residue in prepro-Crp4 as residue position 1. Filled, inverted triangles are positioned above positions with site-directed mutations, and substituted residues are denoted in bold underlined typeface. The mature Crp4 peptide sequence is shown in bold italic type. Asterisks above the reduced (L54D)-pro-Crp4 sequence denote sites of proteolysis in the Crp4 peptide that are not cleaved when the Crp4 tridisulfide array is intact. tected variant Crp4 coding sequences (Refs. 20 and 23-25 and data not shown).
Crp4(B6a) and Crp4(B6b) peptides both arise from Crp4related genes with loss-of-function mutations at the MMP-7 cleavage site near the Crp4 peptide NH 2 terminus. In the case of Crp4(B6a), an L59S substitution, similar to that introduced into pro-CC (Fig. 5), eliminates in vitro MMP-7 cleavage of pro-Crp4(B6a) near the peptide NH 2 terminus (Fig. 6D). The consequence of the mutation in vivo in C57BL/6 small bowel is the accumulation of the abundant Crp4(B6a) peptide, which terminates at Leu 54 instead of Leu 59 as a product of MMP-7mediated pro-Crp4(B6a) cleavage at Ala 53 2Leu 54 . Furthermore, both Crp4(B6a) and Crp4(B6b) contain a Asp 61 substitution for Gly 61 , indicating that other processing sites are subject to mutation.
The Crp4(B6b) peptide also harbors an L59S mutation that would abrogate MMP-7-mediated proteolysis at Ser 58 2Leu 59 . In Crp4(B6b), however, an additional S57L mutation appears to have provided an alternative MMP-7 cleavage site at Lys 56 2Leu 57 that leads to the appearance of Crp4(B6b), although at very low levels. The order in which the S58L and L59S mutations appeared cannot be inferred from these data, but the prospect that the S58L change may have rescued an earlier L59S loss-of-function substitution is an attractive but speculative notion. Whether Crp4(B6b) levels are low because MMP-7 does not cleave pro-Crp4(B6b) efficiently at Lys 56 2Leu 57 is not known. Prepro-Crp4(B6b) mRNA levels are very low relative to those of other ␣-defensin mRNAs (data not shown).
Loss-of-function Mutation at Ser 58 2Leu 59 Is Associated with Attenuated Bactericidal Activity-To test whether L59S inactivation of the Ser 58 2Leu 59 MMP-7 cleavage site has potential effects on innate immunity, the bactericidal activities of Crp4(B6a) and Crp4 were compared against four species of bacteria. Relative to Crp4, Crp4(B6a) was less active against all bacterial species tested (Fig. 7). Against Staphylococcus aureus and Listeria monocytogenes, both peptides had equivalent bacterial cell killing activities at concentrations of 5 g/ml or greater (Fig. 7B), but Crp4(B6a) had markedly lower activity against Gram-negative E. coli and the defensinsensitive S. typhimurium PhoP(Ϫ) strain (Fig. 7A). Although these experiments show that Crp4(B6a) is less potent than FIG. 5. Cleavage of pro-CC serine mutants by MMP-7. A, the amino acid sequence of the pro-Crp chimera, pro-CC, is shown compared with mutants (L59S)-pro-CC and (L54/59S)-pro-CC, which contain one and two Leu to Ser substitutions, respectively. Numbering of the residues below the sequence follows the same scheme as in Fig. 3. The residues underlined in the pro-CC sequence are those that differ between pro-CC and pro-Crp1. The arrows indicate the sites of cleavage by MMP-7, with cleavage at the third site in pro-CC liberating the mature peptide (as defined by NH 2 termini of small intestinal peptides isolated from outbred Swiss mice (24) Crp4, the differences in activity cannot be attributed solely to the 5 additional amino acids at the peptide NH 2 terminus. Crp4 and Crp4(B6a) are different at 7 residue positions and the composition and length of their COOH termini also differ greatly. Nevertheless, the findings show that mutations at MMP-7 cleavage sites exist in mouse populations and that the mutated processing intermediates can accumulate as abundant peptide variants. The boxed region denotes the cryptdin-containing portion of the gel, the large arrow at the right indicates the position of the Crp4(B6a) peptide, and lower arrow shows the position of Crp3. Note that the C57BL/6 cryptdins do not co-migrate with those of the other strains. Panel B, samples consisting of 2 g of the four purified peptides shown and 500 g of total protein extracted from C57BL/6 mouse small intestine were analyzed by AU-PAGE under standard conditions ("Experimental Procedures"). Lanes: 1, pro-Crp4; 2, Crp4; 3, pro-Crp4(B6a); 4, Crp4(B6a), extract from C57BL/6 mouse small bowel. Upper arrow shows the position of Crp4 and Crp4(B6a) precursors, the large middle arrow indicates the position of Crp4(B6a) peptide, and the lowest arrow shows the position of Crp4. Panel C, NH 2 -terminal sequences obtained for Crp4(B6a) and Crp4(B6b) peptides purified from C57BL/6 small intestine are shown aligned with their predicted precursors deduced from their corresponding cDNAs. Filled, inverted triangles indicate sites of loss-of-function mutagenesis in Crp4(B6a) and apparent introduction of an alternative MMP-7 recognition site in Crp4(B6b). Peptide sequences are shown in bold type, and peptide sequences determined by NH 2 -terminal protein sequencing are shown underlined. Panel D, samples consisting of 2 g each of pro-Crp4 and pro-Crp4(B6a) purified from MMP-7 null mice were incubated overnight with 0.5 mol eq of MMP-7 (Fig. 1), and digests were analyzed by 5 cycles of NH 2 -terminal peptide sequencing. Cleavage sites disclosed by protein sequencing are noted by downwards arrows (2) that interrupt the individual sequences.

DISCUSSION
In mouse small intestinal Paneth cells, pro-Crp activation is mediated intracellularly by MMP-7, which processes 60 -70% of pro-Crps prior to secretion (5). The studies reported here show that the processing reactions include cleavage steps at two conserved sites in the proregion and a site at or near the Crp NH 2 terminus. Despite the specificity of MMP-7 proteolysis at these positions, loss-of-function mutations in the two proregion cleavage sites had no effect on the activating cleavage step at the NH 2 terminus of the mature Crp4 or Crp15 chimeric peptide, pro-CC. Cleavage of the Ser 43 2Ile 44 and Ser 53 2Leu 54 sites in the synthetic cryptdin-1 prosegment alone by MMP-7 occurred without the ␣-defensin moiety being present, and thus the Crp peptide does not play a critical role in making the Ser 43 2Val 44 and Ser 53 2Leu 54 sites accessible to the processing enzyme. However, the forced retention of CC in the proform by mutagenizing both Ser 53 2Leu 54 and Ser 58 2Leu 59 sites (Fig. 5) eliminates cleavage at the Ser 43 2Val 44 position.
One objective of these studies was to distinguish (a) whether the activation cleavage at the peptide NH 2 terminus depends on partial degradation of the proregion, (b) whether prosegment degradation depends on the activation cleavage step, or (c) whether the different cleavage steps are independent events. From the data in Fig. 5, we infer that prosegment proteolysis does not regulate Crp peptide activation. Instead, it appears that the proregion remains intact until the activating cleavage event at the Crp peptide NH 2 terminus or near that site at Ser 53 2Leu 54 is completed. Possibly, MMP-7 catalyzed proteolysis in the proregion at Ser 43 2Val 44 eliminates the ability of the proregion to inhibit Crp bactericidal activity (5) by fragmenting the 39-and 34-amino acid prosegment fragments that are released from pro-Crps by proteolysis at Ser 58 2Leu 59 or Ser 53 2Leu 54 , respectively. To summarize, proregion fragmentation depends on Crp peptide activation rather than the converse.
In mature human and rabbit phagocytes, ␣-defensins predominantly exist in the fully processed state, which is mediated by analogous cleavage events (12,13,15). Over a 4 -24-h period, three primary cleavage events generate HNP-1 major intermediates of 75 and 56 amino acids, as well as the 30residue mature peptide, in cells of myeloid origin (12,13,15). Extensive amino acid deletions from the COOH-terminal region of the pro-HNP-1 propeptide, but not from the NH 2 -terminal region of the prosegment, impaired pro-HNP-1 processing (13). Both sets of observations are consistent with MMP-7 processing of pro-Crps in that mutations at the COOH terminus affect cleavage upstream. Our results suggest that the Crp prosegment is removed nearly intact then proteolyzed subsequently, whereas pro-HNPs are sequentially truncated at the NH 2 terminus; nevertheless, the end result is the same. Although the enzymes that mediate pro-␣-defensin processing differ depending on the system, the overall scheme of ␣-defensin processing is similar in myeloid and epithelial cells. The anionic prosegments of both cryptdins and HNPs inhibit ␣-defensin bactericidal activity in vitro (5,15), and thus may confer cytoprotection during granulogenesis until processing is complete (14).
The molecular details of human Paneth cell ␣-defensin processing provide interesting contrasts and comparisons with the biology of pro-Crp activation in the mouse. Mouse Paneth cells secrete mature, 3.5-kDa cryptdins as components of secretory granules, because MMP-7-mediated pro-Crp processing takes place intracellularly before secretion (5). On the other hand, human Paneth cells release the ␣-defensin HD5 precursor, pro-HD5-(20 -94), into the small intestinal lumen (11), and pro-HD5(20 -94) is processed rapidly after secretion by anionic and meso isoforms of trypsin and not by MMP-7, which human Paneth cells lack (10, 11). 2 A trypsin cleavage site in pro-HD5-(20 -94) at Arg 62 2Ala 63 gives rise to the major form of the mature peptide found in washes of the small intestinal lumen (11). However, other processing intermediates, resulting from hydrolysis within the prosegment, also have been identified, suggesting that alternative sites may be used. For example, an HD5 peptide with the NH 2 terminus Gly 37 has been isolated from supernatants of human small intestinal crypts stimulated to release Paneth cell granules with carbamyl choline (10). The extensive differences between the processing of ␣-defensins by mice and humans suggest that it is the capacity for releasing mature, microbicidal ␣-defensins that is conserved, and that Paneth cells from mice and humans evolved differing mechanisms to ensure the delivery of functional peptides into the lumen.
The identification of the Crp4(B6a) variant peptide with a loss-of-function L59S mutation (Fig. 6) shows that mouse populations can accumulate defective pro-Crps. These two variants of Crp4 represent the only Crp4 variants identified in inbred populations and revealed that the Crps from the C57BL/6 strain differ markedly from those of 129/SvJ, C3H/HeJ, BALB/ cJ, and outbred Swiss mice (Fig. 6A, and not shown). This observation should be considered before extrapolating from C57BL/6 genomic DNA and cDNA sequences to other strains of 2 C. L. Wilson and C. L. Bevins, unpublished data.

FIG. 7. Relative bactericidal activities of Crp4 and Crp4(B6a).
Exponentially growing bacterial cells were exposed to the indicated concentrations of Crp4 or Crp4(B6a) for 1 h, and surviving bacteria were quantitated as described in the legend to Fig. 3 mice, at least with respect to Paneth cell gene products. One consequence of the L59S mutation in vivo is an abundance of the Crp4(B6a) peptide in C57BL/6 small bowel (Fig. 6), a molecule that has less bactericidal peptide activity and is especially attenuated against the two Gram-negative bacterial species tested (Fig. 7). Thus, one possibility is that the production of alternative cryptdin peptides in different strains of mice may contribute to variation in susceptibility to enteric pathogens. If mutations at MMP-7 cleavage sites can persist in mouse populations and if the mutated processing intermediates can accumulate as abundant, attenuated peptide variants, it seems likely that comparable defects could disrupt HD5 or HD6 processing in human Paneth cells. Such mutations could predispose certain individuals to increased susceptibility to enteric bacterial infections, a notion that is speculative but testable.