Chemical cross-linking and mass spectrometric identification of sites of interaction for UreD, UreF, and urease.

Synthesis of active Klebsiella aerogenes urease requires four accessory proteins to generate, in a GTP-dependent process, a dinuclear nickel active site with the metal ions bridged by a carbamylated lysine residue. The UreD and UreF accessory proteins form stable complexes with urease apoprotein, comprised of UreA, UreB, and UreC. The sites of protein-protein interactions were explored by using homobifunctional amino group-specific chemical cross-linkers with reactive residues being identified by matrix-assisted laser desorption-ionization time-of-flight mass spectrometry (MALDI-TOF MS) of tryptic peptides. On the basis of studies of the UreABCD complex, UreD is capable of cross-linking with UreB Lys(9), UreB Lys(76), and UreC Lys(401). Furthermore UreD appears to be positioned over UreC Lys(515) according to decreased reactivity of this residue compared with its reactivity in UreD-free apoprotein. Several UreB-UreC and UreC-UreC cross-links also were observed within this complex; e.g. UreB Lys(76) with the UreC amino terminus, UreB Lys(9) with UreC Lys(20), and UreC Lys(515) with UreC Lys(89). These interactions are consistent with the proximate surface locations of these residues observed in the UreABC crystal structure. MALDI-TOF MS analyses of UreABCDF are consistent with a cross-link between the UreF amino terminus and UreB Lys(76). On the basis of an unexpected cross-link between UreB Lys(76) and UreC Lys(382) (distant from each other in the UreABC structure) along with increased side chain reactivities for UreC Lys(515) and Lys(522), UreF is proposed to induce a conformational change within urease that repositions UreB and potentially could increase the accessibility of nickel ions and CO(2) to residues that form the active site.

Synthesis of active Klebsiella aerogenes urease requires four accessory proteins to generate, in a GTP-dependent process, a dinuclear nickel active site with the metal ions bridged by a carbamylated lysine residue. The UreD and UreF accessory proteins form stable complexes with urease apoprotein, comprised of UreA, UreB, and UreC. The sites of protein-protein interactions were explored by using homobifunctional amino group-specific chemical cross-linkers with reactive residues being identified by matrix-assisted laser desorption-ionization time-of-flight mass spectrometry (MALDI-TOF MS) of tryptic peptides. On the basis of studies of the UreABCD complex, UreD is capable of cross-linking with UreB Lys 9 , UreB Lys 76 , and UreC Lys 401 . Furthermore UreD appears to be positioned over UreC Lys 515 according to decreased reactivity of this residue compared with its reactivity in UreD-free apoprotein. Several UreB-UreC and UreC-UreC cross-links also were observed within this complex; e.g. UreB Lys 76 with the UreC amino terminus, UreB Lys 9 with UreC Lys 20 , and UreC Lys 515 with UreC Lys 89 . These interactions are consistent with the proximate surface locations of these residues observed in the UreABC crystal structure. MALDI-TOF MS analyses of UreABCDF are consistent with a cross-link between the UreF amino terminus and UreB Lys 76 . On the basis of an unexpected cross-link between UreB Lys 76 and UreC Lys 382 (distant from each other in the UreABC structure) along with increased side chain reactivities for UreC Lys 515 and Lys 522 , UreF is proposed to induce a conformational change within urease that repositions UreB and potentially could increase the accessibility of nickel ions and CO 2 to residues that form the active site.
The hydrolysis of urea to form carbonic acid and two molecules of ammonia is catalyzed by urease, an enzyme found in all plants and many species of algae, fungi, and bacteria (1). The primary role of plant urease is to recycle arginase-derived urea nitrogen during germination (2), whereas microbial ureases decompose many environmental sources of urea to supply ammonia as a nitrogen source (3). In addition, the bacterial enzyme is a virulence factor that participates in gastroduodenal infection (4), urinary stone formation (5), ammonia encepha-lopathy, and other human and animal disease states (for reviews, see Refs. 3 and 6). Urease sequences from diverse sources are more than 60% identical despite having differences in the numbers of subunits. Plant and fungal ureases contain a single type of subunit, such as the 840-residue polypeptide from jack bean (7). In contrast, Helicobacter pylori urease is comprised of two subunits containing 238 and 569 residues that match distinct portions of the eukaryote enzyme (8). Finally bacteria such as Klebsiella aerogenes possess a threesubunit urease (UreA, UreB, and UreC of 100, 106, and 567 residues, respectively) (9) again matching segments of the plant enzyme sequence. The homology is reflected in the similarity of three-dimensional structures determined for the enzymes from K. aerogenes (10,11), Bacillus pasteurii (12), H. pylori (13), and jack bean (14). In all examples the active site is deeply buried in the protein and includes a dinuclear nickel metallocenter with the two metals (separated by ϳ3.6 Å) bridged by a carbamylated lysine residue. As described below, synthesis of the unique urease metallocenter is a highly complex process requiring nickel ions, carbon dioxide (needed for lysine carbamylation), several accessory proteins, and GTP (for a review, see Ref. 15).
The current model for urease activation initiates with ribosomal synthesis and chaperonin-facilitated folding (16) of the urease apoprotein UreABC. 1 Structural characterization of UreABC reveals an identical fold to holoenzyme with no metal or carbamylation present (17). Studies with the recombinant K. aerogenes system have shown that UreABC forms complexes with the UreD, UreF, and UreG accessory proteins (encoded adjacent to the structural genes in the ureDABCEFG cluster). In particular, UreABCD (18,19), UreABCDF (20), and UreAB-CDFG complexes have been characterized (21,22). Deletion of any of the K. aerogenes ureD, ureF, or ureG genes results in undetectable urease activity, thus highlighting their essential nature (23). Fungal and plant homologues to ureD, ureF, and ureG also are essential for generation of urease activity (24). UreABC and the larger urease apoprotein complexes of K. aerogenes are partially activated by incubation with nickel ions and bicarbonate (as a source of carbon dioxide) with full activity achieved by additional inclusion of GTP and UreE (25). GTP hydrolysis occurs only within the UreABCDFG complex in a UreG-dependent reaction (22). UreE serves a metallochaperone role by delivering the requisite metal ion to the UreABC-DFG complex. Structures of UreE from K. aerogenes (26) and B.   Fig. 1 and the UreAB, UreD, and UreF samples shown in Supplemental Figs. S1, S2, and S3.
b Numbers 1 and 2 in parentheses refer to the number of predicted missed cleavage sites. Met-ox denotes oxidized methionine residues. C-cam is used to designate carboxamidomethylated cysteine residues. pyro-Glu indicates cyclization of an amino terminal Gln residue.
c Two possible calculated options that agree with the observed mass are indicated.
pasteurii (27) along with results from mutagenesis studies (28,29) indicate that the critical metal-binding site is located at the dimer interface generated by one domain of the two-domain protein. The second UreE domain resembles the structure of a domain found in an Hsp40 chaperone, leading to speculation that this UreE domain might participate in protein target recognition.
Very little is known about the sites of binding between the accessory components and urease subunits. Yeast two-hybrid analyses coupled with immunoprecipitation studies (using anti-UreC or anti-UreD antibodies) identified UreA-UreC, UreA-UreA, UreE-UreE, UreC-UreD, and UreD-UreF interactions in the Proteus mirabilis system (30). Yeast two-hybrid studies of the H. pylori system identified interactions between UreD (alternatively named UreH in that microorganism) and both UreA (equivalent to the fusion of UreA plus UreB in most bacteria) and UreF (31,32). Additional evidence was consistent with the following interactions in H. pylori: UreA-UreA, UreA-UreB (where UreB is equivalent to UreC in other bacteria), and UreE-UreG.
In this study we examined accessory protein-urease interactions for two stable complexes, UreABCD and UreABCDF, of K. aerogenes by using a chemical cross-linking approach. UreAB- CDFG was not examined because it is less stable (due to UreG dissociation) and because of its increased complexity. Analysis of the tryptic peptides derived from the cross-linked samples made use of matrix-assisted laser desorption-ionization timeof-flight (MALDI-TOF) mass spectrometry (MS). We tentatively identify specific sites of cross-linking between UreD and both UreB and UreC. In addition, we provide evidence for a previously undescribed interaction between UreF and UreB and identify the putative sites of cross-linking between these proteins. Furthermore we establish several UreF-dependent changes in reactivity of the urease side chains with cross-linking reagents consistent with a significant conformational change occurring in the presence of UreF that may be relevant to metallocenter assembly.

EXPERIMENTAL PROCEDURES
Materials-Dimethyladipimidate (DMA), dimethylsuberimidate (DMS), and bis(sulfosuccinimidyl)suberate (BS 3 ) were purchased from Pierce. Adipic acid, anhydrous 1,4-dioxane, 1,3-dicyclohexylcarbodiimide (1.0 M solution in dichloromethane), glycine, dithiothreitol, iodoacetamide, trifluoroacetic acid, and ␣-cyano-4-hydroxycinnamic acid were purchased from Sigma. Acetonitrile was from EM Science (Gibbstown, NJ). Sequencing grade modified trypsin was from Promega (Madison, denotes cross-linking between two peptides by the reagent. C-pam represents reaction of a Cys with acrylamide to form Cys-S-␤-propionamide. Numbers 1 and 2 in parentheses refer to the number of predicted missed cleavage sites. Met-ox denotes oxidized methionine residues. C-cam is used to designate carboxamidomethylated cysteine residues. FIG. 4. Three-dimensional structure of UreABC. Three views (A, B, and C) are shown for the (UreABC) 3 structure (Protein Data Bank accession number 1FWJ) whose surface is identical for the apoprotein and holoprotein (10,17). UreA is shown in yellow, UreB is shown in blue, and UreC is shown in green. The amino termini are depicted in orange, and all Lys residues are highlighted in red. The residue numbering is provided (as subunit letter-residue number or subunit letter-N for amino terminus) in one-third of the trimeric structure. The buried active site of the holoprotein is encircled.  a Numbers 1 and 2 in parentheses refer to the number of predicted missed cleavage sites. Met-ox denotes oxidized methionine residues. C-cam is used to designate carboxamidomethylated cysteine residues. pyro-Glu indicates cyclization of an amino terminal Gln residue. WI). Disuccinimidyladipate (DSA) was synthesized on a Schlenk line by following a published procedure that involved carbodiimide coupling of adipic acid to N-hydroxysuccinimide in 1,4-dioxane solvent (33).
Culture Conditions and Disruption-Escherichia coli DH5 or DH5␣ carrying pKAU22⌬ureD-1 (23), pKAUD2 (18), or pKAUD2Fϩ⌬ureG (20) were grown in 4 liters of Luria-Bertani medium with 100 mg/liter ampicillin. All cultures were harvested by centrifugation and resuspended in 40 ml of HEDG buffer (25 mM Hepes, 1 mM EDTA, 1 mM dithiothreitol, 1% glycerol, pH 7.4). The resuspended cells were dis-rupted by one to two passages through a French pressure cell (American Instrument Co.) at 18,000 pounds/square inch. Cell extracts were obtained after removal of cell debris and membranes by centrifugation at 100,000 ϫ g at 4°C for 60 min.
Protein Purification-UreABC, UreABCD, and UreABCDF forms of urease apoprotein were purified according to published procedures (18, 20, 34) except that HEDG buffer was used in all purification steps and NaCl was used for linear salt gradient elution of the proteins. The proteins were greater than 95% homogeneous on the basis of denatur- a C-pam represents reaction of a Cys with acrylamide to form Cys-S-␤-propionamide. Numbers 1 and 2 in parentheses refer to the number of predicted missed cleavage sites. Met-ox denotes oxidized methionine residues. C-cam is used to designate carboxamidomethylated cysteine residues. pyro-Glu indicates cyclization of an amino terminal Gln residue. denotes cross-linking between two peptides by the reagent. Met-ox denotes oxidized methionine residues. C-cam is used to designate carboxamidomethylated cysteine residues. pyro-Glu indicates cyclization of an amino terminal Gln residue.

MALDI-MS analysis of tryptic peptides derived from BS 3 -cross-linked bands of UreABCDF
ing SDS-PAGE analysis (35). Molecular weight markers were obtained from Bio-Rad.
Cross-linking Experiments-Selected urease complexes were treated with the following cross-linking reagents: BS 3 , DMA, DMS, and DSA. Samples of UreABC, UreABCD, and UreABCDF (25 l of 50 M "Ure-ABC unit") were incubated at room temperature with BS 3 (3200-, 1600-, 600-, and 400-fold molar excess) in 15 mM sodium phosphate buffer, pH 7.5, containing 150 mM NaCl, and after 60 min the reactions were quenched by adding 15 l of 1 M glycine, pH 6.5. UreABCD (300 l of 5 M UreABC unit) was mixed with DMA or DMS (20 l of 125 mM) in HEDG buffer, pH 7.4, and incubated for 50 min at room temperature before quenching with 30 l of 1 M glycine, pH 6.5. UreABCD (200 l of 5 M UreABC unit) was incubated with DSA (0.5 l of 2 mM in dimethyl sulfoxide) in HEDG buffer, pH 7.4, for 18 h at room temperature and then quenched by adding 5 l of 100 mM ammonium bicarbonate, pH 8.5.
Gel Electrophoresis and In-gel Digestion-The cross-linked samples and controls were denatured in sample buffer containing 60 mM Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, and 0.01% bromphenol blue by boiling for 2 min. Electrophoresis was carried out at room temperature using 12 or 13.5% acrylamide running gels. Selected bands were excised with a razor blade, chopped into 1-mm 3 pieces, and collected in siliconized microcentrifuge tubes. A similar size gel piece from the blank region was also cut out and treated in parallel. The gel pieces were washed with 100 l of 100 mM ammonium bicarbonate buffer (pH 8.5) for 5 min, dehydrated in 50 l of acetonitrile at room temperature for 15 min, and dried in a Speed Vac for 15 min. The gel particles were rehydrated with 50 l of 10 mM dithiothreitol in 100 mM ammonium bicarbonate buffer and heated at 56°C for 30 min to reduce the disulfides. The gel pieces were dehydrated twice in 50 l of acetonitrile at room temperature for 5 min and dried in a Speed Vac for 15 min. 50 l of 55 mM iodoacetamide in 100 mM ammonium bicarbonate buffer was added to each sample and incubated in the dark at room temperature for 20 min to alkylate cysteine residues. The gel pieces were washed briefly with the ammonium bicarbonate buffer and incubated with 100 l of the same solution at room temperature for 15 min. The gel particles were dehydrated with 50 l of acetonitrile for 15 min and dried completely in the Speed Vac. 20 l of 13 ng/l sequencing grade modified trypsin (in 50 mM ammonium bicarbonate buffer) was added to each of the samples and the controls to cover the gel pieces completely. After digestion (37°C overnight), the supernatants were collected by washing the gel pieces twice with 15 l of 60% acetonitrile, 1% trifluoroacetic acid. The pooled supernatants were dried in a Speed Vac, and the peptides were dissolved in 8 l of 3% trifluoroacetic acid in water.
MALDI-TOF MS-MALDI-TOF MS typically was performed on a Voyager-DE STR time-of-flight instrument (Applied Biosystems) equipped with a nitrogen laser operating at 337 nm. In addition, one sample was examined by MS/MS methods using a ThermoFinnigan Atmospheric Pressure MALDI ion trap mass spectrometer. All MALDI-MS results were obtained in the linear positive mode using ␣-cyano-4-hydroxycinnamic acid (saturated solution in 50% acetonitrile with 0.1% trifluoroacetic acid) as the UV-absorbing matrix. Analytes were prepared by mixing 1 l of peptide digest with 1 l of matrix solution on a MALDI plate and allowed to air dry at room temperature in a hood before inserting into the MALDI-TOF mass spectrometer. Mass spectra were externally calibrated with des-Arg 1 -bradykinin (905.05 Da), angiotensin (1297.51 Da), Glu 1 -fibrinopeptide B (1571.61 Da), and insulin (5733.54 Da). In addition, peptides derived from trypsin (843.02 and 2212.44 Da) were used for validation when clearly observed. All masses are reported as average values.
MALDI-TOF mass spectra were analyzed to identify tryptic peptides of the urease proteins, tryptic peptides that reacted with one end of the reagents with the other end being hydrolyzed or reacted with quenching agent, and cross-linked tryptic peptides. Peptides were identified by using the ProteinProspector program 2 to perform theoretical trypsin digests of K. aerogenes urease subunits (UreA, UreB, and UreC) and accessory proteins (UreD and UreF) and searching for potential unmodified tryptic peptides (with up to two missed cleavages) or suspected modified species. Methionine residues were considered as either normal Met or their oxidized form (Met-ox), and cysteine residues were considered to be carbamidomethylated (C-cam) or reacted with acrylamide to produce Cys-S-␤-propionamide (C-pam) (36).

MALDI-TOF MS Analysis of Untreated Urease-related Proteins-
The SDS-PAGE gel-isolated UreC, UreD, and UreF peptides, as well as a mixture of the UreA and UreB peptides, were subjected to trypsin cleavage and analyzed by MALDI-TOF MS. A representative spectrum for the UreC sample is shown in Fig. 1, and spectra of the remaining samples are provided as supplementary data (see Supplemental Figs. S1-S3). Peaks consistent with the expected peptides are summarized in Table  I. For UreA, four peptides (10,12,19, and 25 residues) were identified that cover 66 of the total 100 amino acids, accounting for 66% of the sequence. Nine UreB peptides (6,7,10,11,12,19,23,23, and 27 residues) were represented, covering 91 of the 106 amino acids and accounting for 86% of the sequence. In 2 See prospector.ucsf.edu.  Fig. 4 with the proposed positions of UreD and UreF indicated. A, UreD (light blue) partially protects UreC Lys 515 from reactivity with the reagents and is positioned so as to be able to cross-link to UreB Lys 9 , UreB Lys 76 , and UreC Lys 401 . UreC Lys 522 is non-reactive, whereas a fraction of UreB Lys 9 , UreB Lys 50 , and UreC Lys 20 exhibits half-reactivity with various reagents. Cross-links also occur between UreB Lys 9 and UreC Lys 20 , UreB Lys 76 and the UreC amino terminus, and UreC Lys 89 and UreC Lys 515 . Other surface residues could potentially react with the reagents, but the expected MALDI-TOF MS features were not detected. B, UreF (orange) forms a partial cross-link to UreB Lys 76 while preventing UreD from forming cross-links to the residues indicated above. In addition, UreF induces a urease conformational change (indicated by a shift of the UreB subunit on the left to the rear and of the UreB subunit on the right to the front) that enhances reactivity of UreC Lys 515 and UreC Lys 522 toward the reagent and allows a portion of the sample to form a cross-link between UreB Lys 76 and UreC Lys 382 . the case of UreC, the spectrum was consistent with the presence of 14 peptides (7,7,10,14,14,16,17,19,22,22,26,27,28, and 35 residues) covering 237 of 567 amino acids and accounting for 42% of the sequence. The MALDI-TOF spectrum for peptides derived from UreD was less complete with only four clearly identified peptides (9,12,13, and 17 residues) covering 51 of 270 amino acids and accounting for 19% of the sequence. Better coverage was observed with UreF where nine peptides (7,8,9,11,12,14,15,16, and 22 residues) covered 114 of 224 amino acids and accounted for 50% of the sequence. BS 3 -treated UreABC and UreABCD-As a first attempt to define the sites of interaction between UreD and urease, Ure-ABC and UreABCD were treated with varied concentrations of BS 3 and analyzed by SDS-PAGE with the results of one such experiment depicted in Fig. 2. The treated samples exhibited decreases in intensity of bands corresponding to the ureaserelated peptides, the appearance of new bands above UreC, and extensively cross-linked material that barely entered the gel. Whereas BS 3 -treated UreABC gave rise to two new bands (labeled a and b, faster and slower bands, respectively), the UreABCD sample yielded the same bands as well as a third faint band (labeled c). Gel band a and bands b plus c (because of their proximity) were isolated from BS 3 -treated UreABCD, digested with trypsin, and analyzed by MALDI-TOF MS ( Fig. 3 and Table II). Analogous studies were carried out with a mixture of the new bands from the BS 3 -treated UreABC sample (see Supplemental Table SI). In addition, the remaining UreC bands in the BS 3 -treated samples (including the UreABCDF sample) were analyzed by protease digestion and MALDI-TOF MS (see Supplemental Table SII).
The MALDI-TOF spectrum of the cross-linked UreABCD b/c sample closely resembled that of band a in containing UreC and UreB peptides but also contained three features consistent with the presence of UreD. In particular, a MALDI peak putatively assigned to UreD- (8 -23) was subjected to MS/MS analysis and confirmed to represent this peptide (e.g. the LDLRF internal ion at m/z 645.78, the DLRFHQAGG internal ion less NH 3 at m/z 966.05, the WQATLDLR internal ion at m/z 985.14, the TLDLRFGQAGG internal ion at m/z 1197.35, and the WQATLDLRFHQAG internal ion minus NH 3 at m/z 1508.69 were observed). We conclude that the sample contains UreB, UreC, and UreD cross-linked together or, more likely, a mixture of bands that together contain these peptides. As in the band a sample, features observed in the band b/c sample were consistent with cross-links involving UreB Lys 76 plus the UreC amino terminus and UreC Lys 89 plus UreC Lys 515 along with half-reactivity of UreB Lys 50 and Lys 76 . Of greater interest, two new peaks were compatible with cross-linking of UreB Lys 76 with UreD (UreD-(1-6) and UreD- (7)(8)(9)(10)(11)(12)(13)(14)(15)(16)).
Cross-linking of UreABCD with Other Reagents-To further characterize the sites of interaction of UreD with UreABC, the complex was cross-linked with several other homobifunctional, amino group-specific reagents and analyzed by the SDS-PAGE, trypsin digestion, MALDI-TOF approach described above. Specifically samples were examined after treating with the positively charged reagents DMA and DMS along with the uncharged reagent DSA to complement the studies described above using the negatively charged BS 3 .
Treatment of UreABCD with DMA or DMS, differing only by the number of methylene units (four and six, respectively) joining their imidoester reactive groups, yielded single crosslinked (or a mixture of two closely spaced cross-linked) bands migrating more slowly than UreC during SDS-PAGE (data not shown). MALDI-TOF analysis of tryptic peptides derived from these cross-linked samples (Tables III and IV and see Supplemental Fig. S4) yielded important similarities as well as surprising differences. The spectra provided compelling evidence for half-reactivity involving UreB Lys 50 and a cross-link between UreB Lys 9 plus UreC Lys 20 on the basis of the 28-Da mass shift due to the increased size of the cross-linker. Whereas no potential cross-link involving UreB Lys 76 plus the UreC amino terminus was observed using the smaller reagent (DMA), a feature consistent with this cross-link was detected using DMS. Use of the latter reagent also led to features consistent with half-reactivity of UreC Lys 20 and UreB Lys 9 . Most notably, no features suggesting the presence of UreD were found in the DMA-cross-linked sample, while the band from the DMS-treated sample yielded features consistent with the presence of two UreD peptides (UreD-(234 -244) and UreD-(8 -23)) and a cross-link involving UreC Lys 401 plus the UreD amino terminus. As shown in Fig. 4, all of the UreB and UreC reactivities are reasonable and involve surface groups. Differences in reactivity between DMA and DMS are attributed, at least in part, to the different lengths of their spacer arms (8.6 and 11 Å, respectively). We conclude that DMA-treated Ure-ABCD sample contains UreB and UreC cross-linked together, while DMS-treated UreABCD sample contains both UreB-UreC and UreD-UreC cross-links.
DSA-treated UreABCD yielded a single cross-linked band (data not shown) that migrated more slowly than UreC when analyzed by SDS-PAGE. MALDI-TOF analyses of tryptic peptides of the sample (Table V and see Supplemental Fig. S5) provided evidence for two UreD peptides (UreD-(126 -137) and UreD-(178 -211)) and a putative cross-link between the amino terminus of UreD plus UreB Lys 9 . In addition, MS features were consistent with a cross-link involving UreC Lys 20 plus UreB Lys 9 along with half-reactivity of UreB Lys 9 and UreB Lys 50 . BS 3 -Treated UreABCDF-When UreABCDF was treated with BS 3 , two new cross-linked species were generated (Fig. 5) and shown to migrate either more slowly (band d) or more rapidly (band e) than UreC. MALDI-TOF MS analysis of the tryptic digest for band d indicates the presence of UreB and UreC peptides in the sample (Table VI). No UreD peptides or cross-links involving UreD were detected, suggesting that UreF overlaps UreD and protects it from the reagent. Significantly a putative cross-link was observed between UreC Lys 382 and UreB Lys 76 in this sample. Because these residues are quite distant in the three-dimensional structure of UreABC (Fig. 4), this result is consistent with a conformational change that alters the relative locations of these residues in the Ure-ABCDF complex. 3 The cross-linking/MS data also are consistent with half-reactivity of BS 3 involving three surface residues: UreB Lys 76 , UreC Lys 515 , and UreC Lys 522 . The latter residue is partially buried in the UreABC structure (Fig. 4) and was non-reactive in other samples, suggesting that a conformational change leads to exposure of this residue in the UreAB-CDF complex. Similarly UreC Lys 515 was not observed to react with the various reagents when using UreABCD, but it was accessible in the UreABCDF sample. We assign band d to a UreC-UreB cross-linked species.
MALDI-TOF analysis of band e indicated the presence of both UreB and UreF peptides (Table VI and  Conclusions-On the basis of MALDI-TOF MS studies that identify putative cross-linking sites and sites of protection from reagents, we propose that UreD (M r 29,840) and UreF (M r 25,222) interact with urease as illustrated in Fig. 6. We observed two MS features consistent with UreD cross-linking to UreB Lys 76 (using BS 3 ), one compatible with UreD attaching to UreC Lys 401 (using DMS) and another suggesting that UreD can link to UreB Lys 9 (using DSA). Significantly, interaction between UreD and UreB has not been previously defined. Furthermore UreD appears to be positioned over UreC Lys 515 according to the decreased reactivity of this residue toward the reagents in UreABCD compared with UreABC. For UreAB-CDF, the MALDI-TOF results indicate that the UreF amino terminus can be cross-linked to UreB Lys 76 ; interactions between these proteins had not been described previously. This UreB Lys 76 residue alternatively appears capable of crosslinking to UreC Lys 382 , a situation that suggests a UreF-induced conformational change of the urease apoprotein. 3 For example, Fig. 6 depicts a hingelike shift in position of the major domain of the UreB subunits in urease so as to bring UreB Lys 76 closer to UreC Lys 382 . Such a shift could reasonably increase the dynamic flexibility of the neighboring portion of UreC, perhaps accounting for the increased reactivities of UreC Lys 515 and UreC Lys 522 with BS 3 . More significant to the mechanism of metallocenter assembly is that this conformational change could reasonably lead to increased accessibility of nickel ions and CO 2 to the developing active site. No evidence was observed for cross-links involving UreD peptides in BS 3treated UreABCDF. We interpret this result in terms of UreF protecting UreD from reactivity with the reagent. This interpretation is consistent with findings derived from previously reported immunological cross-reactivity studies where antibodies developed to UreD were able to recognize UreABDC in native gels but did not detect this protein in the UreABCDF sample (20). No UreA-associated cross-links were observed nor were UreD-UreF cross-links detected. Several expected UreB-UreC and UreC-UreC cross-links were observed; e.g. UreB Lys 76 with the UreC amino terminus (via BS 3 and DMS), UreB Lys 9 with UreC Lys 20 (via DMA, DMS, and DSA), and UreC Lys 515 with UreC Lys 89 (in BS 3 -treated UreABCD sample). All urease residues proposed to react with the cross-linkers are located on the protein surface according to the apoprotein and holoprotein structures determined by crystallography (10,17). These studies demonstrate the feasibility of using a chemical cross-linking/SDS-PAGE/proteolysis/MALDI-TOF MS analysis protocol to obtain new insights into the structures of stable protein complexes.