Sidues (Ile-208, Leu-303, Ala-321, Leu-232 and Val-352). The C-domain also contains one isopeptide bond, formed 1113-59-3 site between Lys-363 and Asp-487 (Fig. 4c). In this domain the bond links the first and last b-strands of the domain, b20 and b30, that run parallel to each other in the S5 sheet. The formation of the bond is assisted by the presence of Glu-452, previously described as the E-box motif. The BIBS39 custom synthesis glutamic acid forms only one hydrogen bond to the carbonyl oxygen of the isopeptide crosslink. Similar to the M-domain, the bond is surrounded by aromatic and hydrophobic residues, such as Tyr-375, Phe-380, Pro-456, Tyr459 and Val-485. There is no isopeptide bond observed in the N-domain but when superimposing the C-domain with the N-domain it 18334597 is obvious that the N-domain also has all the necessary amino acids, Lys-52, Asn-183 and Glu-145, to form an isopeptide linkage. However, in our crystal structure, Asn-183 is located on the hinge between the N- and M-domains, and its CG atom is too far away from the lysine NZ atom for a bond to form, although the two side chains are connected via a water molecule (Fig. 4a). Instead the residue following the lysine, Cys-53, forms a disulfide bridge with Cys-74, thereby contributing to the stability of the S1 sheet by other means. If formed, the isopeptide bond in the N-domain would link the two parallel strands of the three-stranded sheet, similar to the bond in the C-domain. Interestingly, the Asn-183 residue expected to form the isopeptide bond in the N-domain, is positioned next to Lys-182, the lysine that is crucial for formation of the intermolecular bond with Thr-499 of another FimP subunit upon pili polymerization. The presence of the two isopeptide bonds was evident from the continuous electron density between the lysines and the asparagine (in the M-domain) or the aspartic acid (in the C-domain). Their existence was also confirmed by ESI-TOF-MS. Site directed mutagenesis was applied to generate two mutants, D230A and E452A, each replacing the catalytic residue of the isopeptide triads with alanine. Mass analyses of the mutant and wild type proteins confirmed the approximate mass losses of NH3 (M-domain), H2OFigure 2. Overall structure of FimP. Ribbon representation of FimP31?91. The N-, M- and C-domains are in coral, light blue, and blue respectively. The residues forming isopeptide bonds in the M- and Cdomains are shown as coral stick models and the disulfides in the Nand C-domains as green sticks. The lysine in the pilin motif is shown as a stick 24786787 model in yellow and blue. The Ca2+ ion bound to the C domain is depicted as a sphere in magenta. Residues in the N-domain putatively involved in isopeptide bond formation are shown in blue. N- and Ctermini are indicated. doi:10.1371/journal.pone.0048364.gFimP Structure and Sequence AnalysesFigure 3. Domains and topology diagram. Each domain (N-, M- and C-) with labeled secondary structure. A: N-domain in coral, B: M-domain in light blue, C: C-domain in blue. The Ca2+ ion is depicted as a magenta sphere. D: Topology diagram of FimP with domains colored as in A . The isopeptides are depicted as red, and disulfides as green, bars. The Ca2+ ion coordinated by the loop is shown as a grey sphere. doi:10.1371/journal.pone.0048364.g(C-domain) or both (wild type) upon isopeptide bond formation compared to the masses calculated from the primary FimP sequence (Table 2). Thus the mutants, where the respective catalytic residue has been removed, form one isopeptide bo.Sidues (Ile-208, Leu-303, Ala-321, Leu-232 and Val-352). The C-domain also contains one isopeptide bond, formed between Lys-363 and Asp-487 (Fig. 4c). In this domain the bond links the first and last b-strands of the domain, b20 and b30, that run parallel to each other in the S5 sheet. The formation of the bond is assisted by the presence of Glu-452, previously described as the E-box motif. The glutamic acid forms only one hydrogen bond to the carbonyl oxygen of the isopeptide crosslink. Similar to the M-domain, the bond is surrounded by aromatic and hydrophobic residues, such as Tyr-375, Phe-380, Pro-456, Tyr459 and Val-485. There is no isopeptide bond observed in the N-domain but when superimposing the C-domain with the N-domain it 18334597 is obvious that the N-domain also has all the necessary amino acids, Lys-52, Asn-183 and Glu-145, to form an isopeptide linkage. However, in our crystal structure, Asn-183 is located on the hinge between the N- and M-domains, and its CG atom is too far away from the lysine NZ atom for a bond to form, although the two side chains are connected via a water molecule (Fig. 4a). Instead the residue following the lysine, Cys-53, forms a disulfide bridge with Cys-74, thereby contributing to the stability of the S1 sheet by other means. If formed, the isopeptide bond in the N-domain would link the two parallel strands of the three-stranded sheet, similar to the bond in the C-domain. Interestingly, the Asn-183 residue expected to form the isopeptide bond in the N-domain, is positioned next to Lys-182, the lysine that is crucial for formation of the intermolecular bond with Thr-499 of another FimP subunit upon pili polymerization. The presence of the two isopeptide bonds was evident from the continuous electron density between the lysines and the asparagine (in the M-domain) or the aspartic acid (in the C-domain). Their existence was also confirmed by ESI-TOF-MS. Site directed mutagenesis was applied to generate two mutants, D230A and E452A, each replacing the catalytic residue of the isopeptide triads with alanine. Mass analyses of the mutant and wild type proteins confirmed the approximate mass losses of NH3 (M-domain), H2OFigure 2. Overall structure of FimP. Ribbon representation of FimP31?91. The N-, M- and C-domains are in coral, light blue, and blue respectively. The residues forming isopeptide bonds in the M- and Cdomains are shown as coral stick models and the disulfides in the Nand C-domains as green sticks. The lysine in the pilin motif is shown as a stick 24786787 model in yellow and blue. The Ca2+ ion bound to the C domain is depicted as a sphere in magenta. Residues in the N-domain putatively involved in isopeptide bond formation are shown in blue. N- and Ctermini are indicated. doi:10.1371/journal.pone.0048364.gFimP Structure and Sequence AnalysesFigure 3. Domains and topology diagram. Each domain (N-, M- and C-) with labeled secondary structure. A: N-domain in coral, B: M-domain in light blue, C: C-domain in blue. The Ca2+ ion is depicted as a magenta sphere. D: Topology diagram of FimP with domains colored as in A . The isopeptides are depicted as red, and disulfides as green, bars. The Ca2+ ion coordinated by the loop is shown as a grey sphere. doi:10.1371/journal.pone.0048364.g(C-domain) or both (wild type) upon isopeptide bond formation compared to the masses calculated from the primary FimP sequence (Table 2). Thus the mutants, where the respective catalytic residue has been removed, form one isopeptide bo.
