Three-dimensional structure of the alkaline protease of Pseudomonas aeruginosa: a two-domain protein with a calcium binding parallel beta roll motif

A receptor-binding region in Escherichia coli alpha-haemolysin

Escherichia coli alpha-hemolysin (HlyA) is a 107-kDa protein toxin with a wide range of mammalian target cells. Previous work has shown that glycophorin is a specific receptor for HlyA in red blood cells (Cortajarena, A. L., Goñi, F. M., and Ostolaza, H. (2001) J. Biol. Chem. 276, 12513-12519). The present study was aimed at identifying the glycophorin-binding region in the toxin. Data in the literature pointed to a short amino acid sequence near the C terminus as a putative receptor-binding domain. Previous sequence analyses of several homologous toxins that belong, like HlyA, to the so-called RTX toxin family revealed a conserved region that corresponded to residues 914-936 of HlyA. We therefore prepared a deletion mutant lacking these residues (HlyA Delta 914-936) and found that its hemolytic activity was decreased by 10,000-fold with respect to the wild type. This deletion mutant was virtually unable to bind human and horse red blood cells or to bind pure glycophorin in an affinity column. The peptide Trp914-Arg936 had no lytic activity of its own, but it could bind glycophorin reconstituted in lipid vesicles. Moreover, the peptide Trp914-Arg936 protected red blood cells from hemolysis induced by wild type HlyA. It was concluded that amino acid residues 914-936 constitute a major receptor-binding region in alpha-hemolysin.

An extracellular calcium-binding domain in bacteria with a distant relationship to EF-hands

Extracellular Ca(2+)-dependent nuclease YokF from Bacillus subtilis and several other surface-exposed proteins from diverse bacteria are encoded in the genomes in two paralogous forms that differ by a approximately 45 amino acid fragment, which comprises a novel conserved domain. Sequence analysis of this domain revealed a conserved DxDxDGxxCE motif, which is strikingly similar to the Ca(2+)-binding loop of the calmodulin-like EF-hand domains, suggesting an evolutionary relationship between them. Functions of many of the other proteins in which the novel domain, named Excalibur (extracellular calcium-binding region), is found, as well as a structural model of its conserved motif are consistent with the notion that the Excalibur domain binds calcium. This domain is but one more example of the diversity of structural contexts surrounding the EF-hand-like calcium-binding loop in bacteria. This loop is thus more widespread than hitherto recognized and the evolution of EF-hand-like domains is probably more complex than previously appreciated.

Crystal structures of a psychrophilic metalloprotease reveal new insights into catalysis by cold-adapted proteases

Enzymes from psychrophilic organisms differ from their mesophilic counterparts in having a lower thermostability and a higher specific activity at low and moderate temperatures. It is in general accepted that psychrophilic enzymes are more flexible to allow easy accommodation and transformation of the substrates at low energy costs. Here, we report the structures of two crystal forms of the alkaline protease from an Antarctic Pseudomonas species (PAP), solved to 2.1- and 1.96-A resolution, respectively. Comparative studies of PAP structures with mesophilic counterparts show that the overall structures are similar but that the conformation of the substrate-free active site in PAP resembles that of the substrate-bound region of the mesophilic homolog, with both an active-site tyrosine and a substrate-binding loop displaying a conformation as in the substrate-bound form of the mesophilic proteases. Further, a region in the catalytic domain of PAP undergoes a conformational change with a loop movement as large as 13 A, induced by the binding of an extra calcium ion. Finally, the active site is more accessible due to deletions occurring in surrounding loop regions.

The structure of bacterial outer membrane proteins

Integral membrane proteins come in two types, alpha-helical and beta-barrel proteins. In both types, all hydrogen bonding donors and acceptors of the polypeptide backbone are completely compensated and buried while nonpolar side chains point to the membrane. The alpha-helical type is more abundant and occurs in cytoplasmic (or inner) membranes, whereas the beta-barrels are known from outer membranes of bacteria. The beta-barrel construction is described by the number of strands and the shear number, which is a measure for the inclination angle of the beta-strands against the barrel axis. The common right-handed beta-twist requires shear numbers slightly larger than the number of strands. Membrane protein beta-barrels contain between 8 and 22 beta-strands and have a simple topology that is probably enforced by the folding process. The smallest barrels form inverse micelles and work as enzymes or they bind to other macromolecules. The medium-range barrels form more or less specific pores for nutrient uptake, whereas the largest barrels occur in active Fe(2+) transporters. The beta-barrels are suitable objects for channel engineering, because the structures are simple and because many of these proteins can be produced into inclusion bodies and recovered therefrom in the exact native conformation.

Purification of secreted leukotoxin (LtxA) from Actinobacillus actinomycetemcomitans

The RTX (repeats in toxin) family of toxins is important in the pathogenesis of many Gram-negative bacteria. The oral and systemic human pathogen Actinobacillus actinomycetemcomitans produces a member of this family known as leukotoxin (LtxA). Previously, we found that LtxA is secreted into culture supernatants of A. actinomycetemcomitans and that this protein is abundant and relatively pure. Here, we report a large-scale method for the isolation and purification of LtxA from culture supernatants of A. actinomycetemcomitans strain JP2. The purification scheme involves ammonium sulfate precipitation of culture supernatants, dialysis, and ultrafiltration to concentrate LtxA to approximately 10mg/ml. We found that LtxA remained soluble in buffer that contained at least 250mM NaCl. Purified LtxA was >98% pure and the final preparations were active against HL-60 cells. The entire purification protocol can be completed within 2 days. The ability to readily obtain a large amount of purified leukotoxin should accelerate investigations into the structure and biology of this important virulence factor.

His-859 is an essential residue for the activity and pH dependence of Escherichia coli RTX toxin alpha-hemolysin

Escherichia coli alpha-hemolysin (HlyA) is a toxin protein that, in common with other members of the RTX family, contains a calcium-binding domain consisting of a number of Gly- and Asp-rich nonapeptides (17 in this case) repeated in tandem. Amino acid number 6 in these nonapeptides is almost invariably Asp, and occasionally Asn, but HlyA contains a His residue (number 859 in the chain) in position 6 of the last-but-one nonapeptide. HlyA mutants have been prepared, by site-directed mutagenesis, in which His-859 has been replaced by an Asn (H859N) or by Asp (H859D). HlyA exists in aqueous media in an aggregate-monomer equilibrium, but only the monomer containing bound Ca(2+) (HlyA.Ca) appears to be competent to achieve target membrane insertion and subsequent lysis. In mutant H859N, equilibrium appears to be shifted toward the aggregate, therefore the protein does not exchange Ca(2+) with the aqueous environment, no HlyA.Ca monomers are detected, and the protein lacks any membrane lytic activity. Mutant H859D in turn is almost indistinguishable from the wild-type regarding its calcium binding and membrane lytic activity, however, it differs significantly in its pH dependence. Wild-type HlyA activity decreases sigmoidally with pH, following rather closely the protonation curve of a His residue (apparent pK(a) approximately 6.5). With mutant H859D activity decreases almost linearly with pH and to a smaller extent. It can be concluded that His-859 plays a critical role in several aspects of HlyA activity, namely self-aggregation properties, calcium binding, hemolysis, and pH dependence.

Caulobacter crescentus synthesizes an S-layer-editing metalloprotease possessing a domain sharing sequence similarity with its paracrystalline S-layer protein

Strains of Caulobacter crescentus elaborate an S-layer, a two-dimensional protein latticework which covers the cell surface. The S-layer protein (RsaA) is secreted by a type I mechanism (relying on a C-terminal signal) and is unusual among type I secreted proteins because high levels of protein are produced continuously. In efforts to adapt the S-layer for display of foreign peptides and proteins, we noted a proteolytic activity that affected S-layer monomers with foreign inserts. The cleavage was precise, resulting in fragments with an unambiguous N-terminal sequence. We developed an assay to screen for loss of this activity (i.e., presentation of foreign peptides without degradation), using transposon and traditional mutagenesis. A metalloprotease gene designated sap (S-layer-associated protease) was identified which could complement the protease-negative mutants. The N-terminal half of Sap possessed significant similarity to other type I secreted proteases (e.g., alkaline protease of Pseudomonas aeruginosa), including the characteristic RTX repeat sequences, but the C-terminal half which normally includes the type I secretion signal exhibited no such similarity. Instead, there was a region of significant similarity to the N-terminal region of RsaA. We hypothesize that Sap evolved by combining the catalytic portion of a type I secreted protease with an S-layer-like protein, perhaps to associate with nascent S-layer monomers to "scan" for modifications.

Disease-causing mutations in cartilage oligomeric matrix protein cause an unstructured Ca2+ binding domain

Chondrocytes from pseudoachondroplasia (PSACH) and multiple epiphyseal dysplasia (EDM1) patients display an enlarged rough endoplasmic reticulum that accumulates extracellular matrix proteins, including cartilage oligomeric matrix protein (COMP). Mutations that cause PSACH and EDM1 are restricted to a 27-kDa Ca(2+) binding domain (type 3 repeat). This domain has 13 Ca(2+)-binding loops with a consensus sequence that conforms to Ca(2+)-binding loops found in EF hands. Most disease-causing mutations are found in the 11-kDa C-terminal region of this domain. We expressed recombinant native and mutant forms of the type 3 repeat domain (T3) and its 11-kDa C-terminal region (T3-Cterm). T3 and T3-Cterm bind approximately 13 and 8 mol of Ca(2+)/mol of protein, respectively. CD, one-dimensional proton, and two-dimensional (1)H-(15)N HSQC spectra of Ca(2+)-bound T3-Cterm indicate a distinct conformation that has little helical secondary structure, despite the presence of 13 EF hand Ca(2+)-binding loops. This conformation is also formed within the context of the intact T3. 19 cross-peaks found between 9.0 and 11.4 ppm are consistent with the presence of strong hydrogen bonding patterns, such as those in beta-sheets. Removal of Ca(2+) leads to an apparent loss of structure as evidenced by decreased dispersion and loss of all down field resonances. Deletion of Asp-470 (a mutation found in 22% of all PSACH and EDM1 patients) decreased the Ca(2+)-binding capacity of both T3 and T3-Cterm by about 3 mol of Ca(2+)/mol of protein. Two-dimensional (1)H-(15)N HSQC spectra of mutated T3-Cterm showed little evidence of defined structure in the presence or absence of Ca(2+). The data demonstrate that Ca(2+) is required to nucleate folding and to maintain defined structure. Mutation results in a partial loss of Ca(2+)-binding capacity and prevents Ca(2+)-dependent folding. Persistence of an unstructured state of the mutated Ca(2+) binding domain in COMP is the structural basis for retention of COMP in the rough endoplasmic reticulum of differentiated PSACH and EDM1 chondrocytes.

The N terminus of the HasA protein and the SecB chaperone cooperate in the efficient targeting and secretion of HasA via the ATP-binding cassette transporter

Secretion of the HasA hemophore is mediated by a C-terminal secretion signal as part of an ATP-binding cassette (ABC) pathway in the Gram-negative bacterium Serratia marcescens. We reconstituted the HasA secretion pathway in Escherichia coli. In E. coli, this pathway required three specific secretion functions and SecB, the general chaperone of the Sec pathway that recognizes HasA. The secretion of the isolated C-terminal secretion signal was not SecB-dependent. We have previously shown that intracellular folded HasA can no longer be secreted, and we proposed a step in the secretion process before the recognition of the secretion signal. Here we show that the secretion of a fully functional HasA variant, lacking the first 10 N-terminal amino acids, was less efficient than that of HasA and was SecB-independent. The N terminus of HasA was required, along with SecB, for the efficient secretion of the whole protein. We have also previously shown that HasA inhibits the secretion of metalloproteases from Erwinia chrysanthemi by their specific ABC transporter. Here we show that this abortive interaction between HasA and the E. chrysanthemi metalloprotease ABC transporter required both SecB and the N terminus of HasA. N-terminal fragments of HasA displayed this abortive interaction in vivo and also interacted specifically in vitro with the ABC protein of the Prt system. SecB also interacted specifically in vitro with the ABC protein of the Prt system. Finally, the HasA variant, lacking the first 10 N-terminal amino acids did not display this abortive interaction with the Prt system. We suggest that the N-terminal domain of HasA specifically recognizes the ABC protein in a SecB-dependent fashion, facilitating functional interaction with the C-terminal secretion signal leading to efficient secretion.

The functions of Ca(2+) in bacteria: a role for EF-hand proteins?

In bacteria, Ca(2+) is implicated in a wide variety of cellular processes, including the cell cycle and cell division. Dedicated influx and efflux systems tightly control the low cytoplasmic Ca(2+) levels in prokaryotes. Additionally, the growing number of proteins containing various Ca(2+)-binding motifs supports the importance of Ca(2+), which controls various protein functions by affecting protein stability, enzymatic activity or signal transduction. The existence of calmodulin-like proteins (containing EF-hand motifs) in bacteria is a long-standing hypothesis. Analysis of the prokaryotic protein sequences available in the databases has revealed the presence of several calmodulin-like proteins containing two or more authentic EF-hand motifs, suggesting that calmodulin-like proteins could be involved in Ca(2+) regulation in bacteria.

The conformations of polypeptide chains where the main-chain parts of successive residues are enantiomeric. Their occurrence in cation and anion-binding regions of proteins

We have investigated the shapes of polypeptides where successive residues have main-chain phi,psi conformations of opposite hand. A graph not unlike a Ramachandran plot is presented illustrating the various possible conformations. All are ring-shaped or extended. Some of these conformations occur in native proteins, the commonest approximating to a feature we propose calling a nest, described in the accompanying paper, where the main-chain NH groups point inwards relative to the ring and give rise to an anion-binding site. Another conformation is related but more extended and is found uniquely in the four stretches of polypeptide that line the tetrameric K(+) channel; their CO groups bind the K ions in the channel. In a different ring-shaped conformation that we propose calling a catgrip, the main-chain CO groups point into the ring; this is employed for specific Ca ion binding in the annexin, phospholipase A2 and subtilisin loops, and the regularly arranged beta-roll loops of the serralysin protease family.

A theoretical model of a plant antifreeze protein from Lolium perenne

Antifreeze proteins (AFPs), found in certain organisms enduring freezing environments, have the ability to inhibit damaging ice crystal growth. Recently, the repetitive primary sequence of the AFP of perennial ryegrass, Lolium perenne, was reported. This macromolecular antifreeze has high ice recrystallization inhibition activity but relatively low thermal hysteresis activity. We present here a theoretical three-dimensional model of this 118-residue plant protein based on a beta-roll domain with eight loops of 14-15 amino acids. The fold is supported by a conserved valine hydrophobic core and internal asparagine ladders at either end of the roll. Our model, which is the first proposed for a plant AFP, displays two putative, opposite-facing, ice-binding sites with surface complementarity to the prism face of ice. The juxtaposition of the two imperfect ice-binding surfaces suggests an explanation for the protein's inferior thermal hysteresis but superior ice recrystallization inhibition activity and activity when compared with fish and insect AFPs.

Ca(2+)-induced folding of a family I.3 lipase with repetitive Ca(2+) binding motifs at the C-terminus

In order to understand a role of the Ca(2+) ion on the structure and function of a Ca(2+)-dependent family I.3 lipase from Pseudomonas sp. MIS38, apo-PML, holo-PML, holo-PML*, and the N-terminal domain alone (N-fragment) were prepared and biochemically characterized. Apo-PML and holo-PML represent refolded proteins in the absence and presence of the Ca(2+) ion, respectively. Holo-PML* represents a holo-PML dialyzed against 20 mM Tris-HCl (pH 7.5). The results suggest that the C-terminal domain of PML is almost fully unfolded in the apo-form and its folding is induced by Ca(2+) binding. The folding of this C-terminal domain may be required to make a conformation of the N-terminal catalytic domain functional.

Protease C of Erwinia chrysanthemi: the crystal structure and role of amino acids Y228 and E189

PrtC, a metallo-protease secreted by Erwinia chrysanthemi, is a member of the serralysin family and hence belongs to the metzincin superfamily. While the crystal structures of representatives of all metzincin subfamilies have been elucidated in the past, there is still some controversy about the reaction mechanism and the role of certain characteristic amino acids in the active centre. In this study, we probed the role of Tyr228 and Glu189 by site-directed mutagenesis and X-ray crystallography. There is evidence that these residues participate in catalysis, although conflicting hypotheses have been proposed. The crystal structures of wild-type and mutants have been refined to an R(free) of about 0.20 at resolutions of 2.0 A or better. Exchange of Glu189 versus alanine reduces the catalytic efficiency to less than 0.5 % using resorufin casein as substrate and to about 3 % using an assay utilising the thiol ester Ac-Pro-Leu-Gly-[(S)Leu]-Leu-Gly-OEt. The drop in activity is caused by a reduction in k(cat) while the K(M) values are virtually the same. In the resorufin casein assay, the mutant Y228F shows about 3 % of the wild-type activity and in the thiol ester assay this increases to about 56 %. In the latter case, the K(M) value of the mutant is increased from 5.3 mM to 9.0 mM with only little reduction in k(cat). The different behaviour of this mutant with respect to the two substrates can be explained by a switch in the rate-determining step during catalysis. The study presented here provides clear evidence that Glu189 of the HEXXHXXGXXH motif is the catalytic base, while Tyr228 is more likely involved in substrate binding and the stabilisation of the tetrahedral transition state.

Structural consequences of divalent metal binding by the adenylyl cyclase toxin of Bordetella pertussis

Adenylyl cyclase toxin of Bordetella pertussis has been shown by several investigators to require Ca(2+) for its actions on target cells, but little is known about the nature and specificity of divalent metal binding to this novel toxin. Calcium is the preferred divalent metal since toxic actions are markedly reduced in the presence of divalent species other than calcium. Mn(2+) EPR was used to quantitate and characterize divalent metal binding and revealed that the toxin contains approximately 40 divalent metal sites, consisting of at least one class of high-affinity sites that bind Mn(2+) with a K(D) of 0.05 to 0.35 microM and one or more classes of lower affinity sites. Water proton relaxation data indicate that approximately 30 of these sites are completely inaccessible to bulk solvent. Our observations, together with the sequence homology between adenylyl cyclase toxin and the alkaline protease of Pseudomonas aeruginosa, indicate that the formation of five beta-sheet helices within the repeat domain of the toxin upon binding Ca(2+) is required for cell intoxication.

The architecture of parallel beta-helices and related folds

Three-dimensional structures have been determined of a large number of proteins characterized by a repetitive fold where each of the repeats (coils) supplies a strand to one or more parallel beta-sheets. Some of these proteins form superfamilies of proteins, which have probably arisen by divergent evolution from a common ancestor. The classical example is the family including four families of pectinases without obviously related primary sequences, the phage P22 tailspike endorhamnosidase, chrondroitinase B and possibly pertactin from Bordetella pertusis. These show extensive stacking of similar residues to give aliphatic, aromatic and polar stacks such as the asparagine ladder. This suggests that coils can be added or removed by duplication or deletion of the DNA corresponding to one or more coils and explains how homologous proteins can have different numbers of coils. This process can also account for the evolution of other families of proteins such as the beta-rolls, the leucine-rich repeat proteins, the hexapeptide repeat family, two separate families of beta-helical antifreeze proteins and the spiral folds. These families need not be related to each other but will share features such as relative untwisted beta-sheets, stacking of similar residues and turns between beta-strands of approximately 90 degrees often stabilized by hydrogen bonding along the direction of the parallel beta-helix.Repetitive folds present special problems in the comparison of structures but offer attractive targets for structure prediction. The stacking of similar residues on a flat parallel beta-sheet may account for the formation of amyloid with beta-strands at right-angles to the fibril axis from many unrelated peptides.

Crystal structure of a complex between Pseudomonas aeruginosa alkaline protease and its cognate inhibitor: inhibition by a zinc-NH2 coordinative bond

Serralysins are a family of metalloproteases secreted by Gram-negative bacteria into the medium in the form of inactive zymogens. Usually, all serralysin secretors have on the same operon a gene coding for a periplasmic 10-kDa protein, which is an inhibitor of the secreted protease. The recent characterization of the inhibitor of the alkaline protease from Pseudomonas aeruginosa revealed a surprisingly low dissociation constant of 4 pm, contrary to earlier studies on homologous systems, where inhibition constants in the microm range were reported. To approach a more accurate understanding, the crystal structure of the complex between inhibitor and protease from P. aeruginosa was determined at 1.74 A resolution and refined to R(free) = 0.204. The structure reported here shows clearly that the N terminus of the inhibitor forms a coordinative bond to the catalytic Zn(2+) ion with a nitrogen-zinc distance of 2.17 A. We conclude that this interaction adds substantially to the complex stability and show also that similar interactions are found in other metzincin-inhibitor complexes.

Membrane interaction of Escherichia coli hemolysin: flotation and insertion-dependent labeling by phospholipid vesicles

The 1,024-amino-acid acylated hemolysin of Escherichia coli subverts host cell functions and causes cell lysis. Both activities require insertion of the toxin into target mammalian cell membranes. To identify directly the principal toxin sequences dictating membrane binding and insertion, we assayed the lipid bilayer interaction of native protoxin, stably active toxin, and recombinant peptides. Binding was assessed by flotation of protein-liposome mixtures through density gradients, and insertion was assessed by labeling with a photoactivatable probe incorporated into the target lipid bilayer. Both the active acylated hemolysin and the inactive unacylated protoxin were able to bind and also insert. Ca(2+) binding, which is required for toxin activity, did not influence the in vitro interaction with liposomes. Three overlapping large peptides were expressed separately. A C-terminal peptide including residues 601 to 1024 did not interact in either assay. An internal peptide spanning residues 496 to 831, including the two acylation sites, bound to phospholipid vesicles and showed a low level of insertion-dependent labeling. In vitro acylation had no effect on the bilayer interaction of either this peptide or the full-length protoxin. An N-terminal peptide comprising residues 1 to 520 also bound to phospholipid vesicles and showed strong insertion-dependent labeling, ca. 5- to 25-fold that of the internal peptide. Generation of five smaller peptides from the N-terminal region identified the principal determinant of lipid insertion as the hydrophobic sequence encompassing residues 177 to 411, which is conserved among hemolysin-related toxins.

The catalytic activities of the bifunctional Azotobacter vinelandii mannuronan C-5-epimerase and alginate lyase AlgE7 probably originate from the same active site in the enzyme

The Azotobacter vinelandii genome encodes a family of seven secreted Ca(2+)-dependent epimerases (AlgE1--7) catalyzing the polymer level epimerization of beta-D-mannuronic acid (M) to alpha-L-guluronic acid (G) in the commercially important polysaccharide alginate. AlgE1--7 are composed of two types of protein modules, A and R, and the A-modules have previously been found to be sufficient for epimerization. AlgE7 is both an epimerase and an alginase, and here we show that the lyase activity is Ca(2+)-dependent and also responds similarly to the epimerases in the presence of other divalent cations. The AlgE7 lyase degraded M-rich alginates and a relatively G-rich alginate from the brown algae Macrocystis pyrifera most effectively, producing oligomers of 4 (mannuronan) to 7 units. The sequences cleaved were mainly G/MM and/or G/GM. Since G-moieties dominated at the reducing ends even when mannuronan was used as substrate, the AlgE7 epimerase probably stimulates the lyase pathway, indicating a complex interplay between the two activities. A truncated form of AlgE1 (AlgE1-1) was converted to a combined epimerase and lyase by replacing the 5'-798 base pairs in the algE1-1 gene with the corresponding A-module-encoding DNA sequence from algE7. Furthermore, substitution of an aspartic acid residue at position 152 with glycine in AlgE7A eliminated almost all of both the lyase and epimerase activities. Epimerization and lyase activity are believed to be mechanistically related, and the results reported here strongly support this hypothesis by suggesting that the same enzymatic site can catalyze both reactions.

Serratia ATP-binding cassette protein exporter, Lip, recognizes a protein region upstream of the C terminus for specific secretion

Serratia marcescens ATP-binding cassette (ABC) exporter, the Lip system, secretes lipase (LipA(SM)), metalloproteases, and a cell surface layer protein homologue but not a heme acquisition protein, HasA (HasA(SM)). Secretion of HasA(SM) is limited to the Has(SM) system. However, HasA proteins from Pseudomonas fluorescens (HasA(PF)) and Pseudomonas aeruginosa were exported through the Lip and Has(SM) systems. To investigate the specificity in Lip exporter-mediated secretion, secretion analysis was performed using chimeras containing the HasA(PF) and HasA(SM) sequences. The segment Val-Ala-Leu (designated R1 to R3 sites), which is present close to the C terminus of HasA(PF) but not HasA(SM), was revealed to be involved in the substrate specificity of the Lip exporter. Introduction of amino acid substitutions into the R1-R5 region demonstrated that R1, R3, R4, and R5 sites require some specific amino acid residues for Lip-mediated secretion. The amino acid sequence of the region was conserved considerably among the proteins secreted by the Lip exporter. On the contrary, the region was not related to HasA secretion through the Has(SM) system. Interestingly, a typical C-terminal motif, so far regarded as a secretion signal, was not necessary for secretion through either the Lip or the Has(SM) exporter. In LipA(SM) secretion via the Lip system, the typical C-terminal motif was not essential either, but the presence of a sequence similar to Val-Ala-Leu and its location from the C terminus greatly affect the secretion level. Secretion analyses using hybrid exporters and competitors exhibited that the R1-R5 region was recognized by an ABC protein of the Lip exporter, LipB, and that the mutations aborting Lip-mediated secretion in the region resulted in a loss of the affinity to LipB. Thus, a determinant within the secretory protein for Lip-mediated secretion was fully defined.

RTX toxin structure and function: a story of numerous anomalies and few analogies in toxin biology

It can be agreed that RTX toxins contribute to the pathogenesis of different diseases by causing dysfunction of the general cellular reactions of the immune response. The suggestion that RTX toxins induce cytokine production in nonimmune cells that would ultimately cause tissue damage is an expansion of their role in disease pathogenesis (Uhlen et al. 2000). Investigators in the RTX toxin field may not agree with me, but precise and satisfactory answers to the following questions are not yet available. How do RTX toxins mechanistically damage a cell? Do RTX toxins have receptors in the classic sense, in which there is a reversible ligand and receptor complex? What is responsible for the common Ca2+ ion influx in affected cells? The recent observation that an RTX toxin stimulates host-cell-mediated Ca2+ ion oscillation in part challenges the long held concept that these toxins damage cells by the direct formation of pores. Are the Ca2+ ion fluxes truly the noxious cellular insult? What is the final molecular structure of RTX toxins at the time they cause cellular death? How does the common requirement for acyl modification among RTX toxins fit into the toxin structure and mechanism of cellular killing, particularly when mixtures of unusual fatty acids are used by some toxins? There are a number of outstanding laboratories throughout the world that are seeking answers to these questions. We can reasonably expect that during the next decade research on the structure and function of RTX toxins will lead to new chemotherapeutic targets and reagents for basic cell biology and biotechnology.

Calcium-dependent conformation of a heme and fingerprint peptide of the diheme cytochrome c peroxidase from Paracoccus pantotrophus

The structural changes in the heme macrocycle and substituents caused by binding of Ca(2+) to the diheme cytochrome c peroxidase from Paracoccus pantotrophus were clarified by resonance Raman spectroscopy of the inactive fully oxidized form of the enzyme. The changes in the macrocycle vibrational modes are consistent with a Ca(2+)-dependent increase in the out-of-plane distortion of the low-potential heme, the proposed peroxidatic heme. Most of the increase in out-of-plane distortion occurs when the high-affinity site I is occupied, but a small further increase in distortion occurs when site II is also occupied by Ca(2+) or Mg(2+). This increase in the heme distortion explains the red shift in the Soret absorption band that occurs upon Ca(2+) binding. Changes also occur in the low-frequency substituent modes of the heme, indicating that a structural change in the covalently attached fingerprint pentapeptide of the LP heme occurs upon Ca(2+) binding to site I. These structural changes may lead to loss of the sixth ligand at the peroxidatic heme in the semireduced form of the enzyme and activation.

Crystal structure of E. coli beta-carbonic anhydrase, an enzyme with an unusual pH-dependent activity

Carbonic anhydrases fall into three distinct evolutionary and structural classes: alpha, beta, and gamma. The beta-class carbonic anhydrases (beta-CAs) are widely distributed among higher plants, simple eukaryotes, eubacteria, and archaea. We have determined the crystal structure of ECCA, a beta-CA from Escherichia coli, to a resolution of 2.0 A. In agreement with the structure of the beta-CA from the chloroplast of the red alga Porphyridium purpureum, the active-site zinc in ECCA is tetrahedrally coordinated by the side chains of four conserved residues. These results confirm the observation of a unique pattern of zinc ligation in at least some beta-CAS: The absence of a water molecule in the inner coordination sphere is inconsistent with known mechanisms of CA activity. ECCA activity is highly pH-dependent in the physiological range, and its expression in yeast complements an oxygen-sensitive phenotype displayed by a beta-CA-deletion strain. The structural and biochemical characterizations of ECCA presented here and the comparisons with other beta-CA structures suggest that ECCA can adopt two distinct conformations displaying widely divergent catalytic rates.

Review: proteins with repeated sequence--structural prediction and modeling

The relationship between the amino acid sequence and the three-dimensional structure of proteins with internal repeats is discussed. In particular, correlations between the amino acid composition and the ability to fold in a unique structure, as well as classification of the structures based on their repeat length, are described. This analysis suggests rules that can be used for the structural prediction of repeat-containing proteins. The paper is focused on prediction and modeling of solenoid-like proteins with the repeat length ranging between 5 and 40 residues. The models of leucine-rich repeat proteins and bacterial proteins with pentapeptide repeats are examined in light of the recently solved structures of the related molecules.

Interaction of Escherichia coli hemolysin with biological membranes. A study using cysteine scanning mutagenesis

Escherichia coli hemolysin (HlyA) is a membrane-permeabilizing protein belonging to the family of RTX-toxins. Lytic activity depends on binding of Ca2(+) to the C-terminus of the molecule. The N-terminus of HlyA harbors hydrophobic sequences that are believed to constitute the membrane-inserting domain. In this study, 13 HlyA cysteine-replacement mutants were constructed and labeled with the polarity-sensitive fluorescent probe 6-bromoacetyl-2-dimethylaminonaphthalene (badan). The fluorescence emission of the label was examined in soluble and membrane-bound toxin. Binding effected a major blue shift in the emission of six residues within the N-terminal hydrophobic domain, indicating insertion of this domain into the lipid bilayer. The emission shifts occurred both in the presence and absence of Ca2(+), suggesting that Ca2(+) is not required for the toxin to enter membranes. However, binding of Ca2(+) to HlyA in solution effected conformational changes in both the C-terminal and N-terminal domain that paralleled activation. Our data indicate that binding of Ca2(+) to the toxin in solution effects a conformational change that is relayed to the N-terminal domain, rendering it capable of adopting the structure of a functional pore upon membrane binding.

Influence of environmental conditions on the activity of the recombinant mannuronan C-5-epimerase AlgE2

The mannuronan C-5-epimerase AlgE2 is one of a family of Ca(2+)-dependent epimerases secreted by Azotobacter vinelandii. These enzymes catalyze the conversion of beta-D-mannuronic acid residues (M) to alpha-L-guluronic acid residues (G) in alginate. AlgE2 had a pH optimum between 6.5 and 7 and a temperature optimum around 55 degrees C. Addition of low molecular weight organic compounds, including buffers, amino acids and osmoprotective compounds, affected the activity of the enzyme. The charge, size and stereochemistry of the added compounds were important. The activity of AlgE2, dissolved in various buffers (same pH), decreased with increasing fraction of positively charged buffer ions. Mono- and divalent metal ions also influenced the activity. When Ca(2+) was omitted only Sr(2+), of the metal ions tested, supported some activity of AlgE2. At high concentration of Ca(2+) (3.3 mM) these ions had a negative effect on the activity, whereas at low Ca(2+) concentration (0.58 mM) the activity was enhanced by addition of Sr(2+), and to some degree also by addition of Mg(2+) and Mn(2+). During epimerization AlgE2 occasionally causes cleavage of the alginate chain. These chain breaks could not be prevented by changes in the conditions during the epimerization. The composition and sequential structure of epimerized alginate was not altered by changes in the epimerization conditions.

Structure of the zinc-binding site in the crystal structure of a zinc endoprotease from Streptomyces caespitosus at 1 A resolution

A zinc endoprotease produced by Streptomyces caespitosus (ScNP) contains a H83E84TGH87VLG90LPD93-Met103 sequence. Except for D93, this amino acid sequence is the same as a characteristic consensus HEXXHXXGXXH-M motif found in one class of zinc endoprotease called 'metzincins'. We analyzed the structural and functional role of the consensus sequence located around a catalytically essential zinc ion based on the crystal structure of ScNP. The structure was determined at the highest level on resolution (1 A resolution) and accuracy among crystal structures of zinc endoproteases ever determined. The zinc ion of ScNP is tetrahedrally coordinated by three amino acid side-chains (H83, H87 and D93) and a water molecule. The distances between the zinc ion and the coordinating atoms are 2.01, 2.01 and 1.95 A for H83N epsilon, H87N epsilon and D93O delta, respectively. These distances agree very well with those normally found in crystal structures of zinc-containing small molecules in the Cambridge Structural Database. On the other hand, the distance between the zinc ion and the coordinating water molecule (1.93 A) is slightly shorter than the typical value (2.01 A) found in the database. In addition, E84O epsilon makes a short hydrogen bond to this water molecule with the distance of 2.54 A. Two hydrogen bonds (H83N delta-L102O, H87N delta-L91O) and van der Waals interactions between the side-chain of M103 and the two imidazole rings of H83 and H87 are also observed. These interactions are probably important for the imidazole rings to construct the tetrahedral coordination arrangement toward the zinc ion.

Identification of the histidine and aspartic acid residues essential for enzymatic activity of a family I.3 lipase by site-directed mutagenesis

A lipase from Pseudomonas sp. MIS38 (PML) is a member of the lipase family I.3. We analyzed the roles of the five histidine residues (His(30), His(274), His(291), His(313), and His(365)) and five acidic amino acid residues (Glu(253), Asp(255), Asp(262), Asp(275), and Asp(290)), which are fully conserved in the amino acid sequences of family I.3 lipases, by site-directed mutagenesis. We showed that the mutation of His(313) or Asp(255) to Ala almost fully inactivated the enzyme, whereas the mutations of other residues to Ala did not seriously affect the enzymatic activity. Measurement of the far- and near-UV circular dichroism spectra suggests that inactivation by the mutation of His(313) or Asp(255) is not due to marked changes in the tertiary structure. We propose that His(313) and Asp(255), together with Ser(207), form a catalytic triad in PML.

Protein secretion mechanisms in Gram-negative bacteria

Gram-negative bacteria have developed a variety of secretion pathways to secrete toxins and enzymes into the extracellular medium. These pathways are very different with respect to their functional mechanism and complexity, and each system has its own advantages and limitations, regarding the number, size, folding state and fate of their substrates. Pseudomonas aeruginosa secretes many different proteins into the extracellular medium, using at least four secretion pathways. Most of the exoproteins are secreted via the type II system, composed of the 12 Xcp proteins. The only outer membrane protein of the system, XcpQ, belongs to a large family of proteins, designated secretins, which participate in a variety of different transport processes. Other Xcp proteins, XcpT-X, show homology to the subunits of the retractile type IV pili. Further analogies between the type II system and the assembly of retractile pili suggest a mechanism for type II secretion, in which a pilus-like structure, composed of XcpT-X, facilitates the transport of exoproteins through the channel formed by the secretin XcpQ.

When protein folding is simplified to protein coiling: the continuum of solenoid protein structures

Solenoid proteins contain repeating structural units that form a continuous superhelix. This category of proteins conveys the least complicated relationship between a sequence and the corresponding three-dimensional structure. Although solenoid proteins are divided into different classes according to commonly used classification schemes, they share many structural and functional properties.

The Sinorhizobium meliloti ExpE1 protein secreted by a type I secretion system involving ExpD1 and ExpD2 is required for biosynthesis or secretion of the exopolysaccharide galactoglucan

In Sinorhizobium meliloti the biosynthesis of the exopolysaccharide galactoglucan (EPS II) is directed by the exp genes. The expD1 and expD2 gene products are homologous to components of type I secretion systems. ExpE1, the gene of which is located adjacent to expD1 and expD2, was detected in S. meliloti cells and culture supernatants. ExpD1 and ExpD2 were required for the secretion of ExpE1, indicating that ExpE1 is secreted by a type I secretion system involving ExpD1 and ExpD2. ExpE1 contains 15 aspartate- and glycine-rich nonapeptide repeats that were suggested to bind Ca(2+). The ability to bind Ca(2+) was demonstrated for a recombinant ExpE1 protein. Extracellular EPS II was not detected in cultures of non-polar expD1, expD2 and expE1 deletion mutants implying that these three genes are required for biosynthesis or secretion of galactoglucan.

Alkaline proteinase inhibitor of Pseudomonas aeruginosa. Interaction of native and N-terminally truncated inhibitor proteins with Pseudomonas metalloproteinases

The apr locus of Pseudomonas aeruginosa encodes alkaline proteinase (APR), a member of the metzincin metalloendopeptidase superfamily, and an 11.4-kDa alkaline proteinase inhibitor (APRin). We describe here the expression in Escherichia coli and characterization of full-length and N-terminally truncated APRin proteins. Fluorescence and circular dichroism spectra indicated that the recombinant proteins were folded into native-like structures. Analytical ultracentrifugation showed that APRin was monomeric and formed a 1:1 complex with APR. Binding of wild-type APRin to APR occurred with association (k(on)) and dissociation (k(off)) rate constants of 0.29 +/- 0.06 x 10(6) m(-1) s(-1) and 1.15 +/- 0.08 x 10(-6) s(-1) to give an equilibrium dissociation constant (K(D)) of approximately 4 x 10(-12) m (25 degrees C, pH 7.0, ionic strength 2.4 m). The association rate decreased by approximately 2-fold in 20% glycerol and increased by approximately 3-fold in 0.1 m NaCl. The glycerol effect suggests a diffusion-limited reaction, and the small salt effect indicates that electrostatic interactions contribute little to binding. Deletion of residues 1-10, 1-6, or 6-10 abolished inhibition, and deletion of residues 1-2, 1-3, 1-4, and 1-5 resulted in a progressively decreased affinity of APRin for APR (K(D) = 0.12 micrometer the Delta(1-5) mutant). Substitution of APRin residues 6-10 with a (Gly)(5) or (Pro)(5) linker restored inhibitory activity of the Delta(6-10) mutant but with a 100- and 50-fold reduction in K(D). Log k(on) for the full-length and truncated inhibitors correlated with the solvent-accessible surface area of their N-terminal regions, suggesting that increased interactions and/or desolvation of these residues in the transition state for binding contribute to the enhanced association rate. Treatment of APRin with pseudolysin, also secreted by P. aeruginosa, resulted in removal of residues 1-5. APRin was neither an inhibitor nor a substrate of other metzincins, including collagenase or gelatinases A or B.

Cloning and sequencing of a gene of organic solvent-stable protease secreted from Pseudomonas aeruginosa PST-01 and its expression in Escherichia coli

A gene of organic solvent-stable protease (PST-01 protease) secreted by Pseudomonas aeruginosa PST-01 was cloned and its nucleotide was sequenced. The nucleotide sequence analysis revealed that the PST-01 protease was a pseudolysin, which was an elastase produced by P. aeruginosa and was well characterized by the previous investigators. The PST-01 protease produced in recombinant Escherichia coli was not secreted into the extracellular medium, but its proenzyme was released by the lysis of the cells and became a 33.1kDa mature enzyme autoproteolytically. Its characteristics including organic solvent stability were as same as those of the PST-01 protease secreted by P. aeruginosa PST-01.

Purification, physico-chemical characterization and sequence of a heat labile alkaline metalloprotease isolated from a psychrophilic Pseudomonas species

The psychrophilic alkaline metalloprotease (PAP) produced by a Pseudomonas bacterium isolated from Antarctica has been purified and characterized. The gene encoding PAP has been cloned and sequenced and the derived amino acid sequence shows 66% identity with the mesophilic alkaline metalloprotease from Pseudomonas aeruginosa IFO 3455 (AP). Compared to the purified AP, PAP is three times more active at 20 degrees C, is very sensitive to chelating agents and is rapidly inactivated at 45 degrees C. The lower thermostability of PAP can tentatively be explained by a loss of a stabilizing Ca(2+), a decrease in the content of hydrophobic residues and a smaller aliphatic index.

Secretion of the Caulobacter crescentus S-layer protein: further localization of the C-terminal secretion signal and its use for secretion of recombinant proteins

The secretion signal of the Caulobacter crescentus S-layer protein (RsaA) was localized to the C-terminal 82 amino acids of the molecule. Protein yield studies showed that 336 or 242 C-terminal residues of RsaA mediated secretion of >50 mg of a cellulase passenger protein per liter to the culture fluids.

Overproduction in Escherichia coli, purification and characterization of a family I.3 lipase from Pseudomonas sp. MIS38

Determination of the nucleotide sequence of the gene encoding a lipase from Pseudomonas sp. MIS38 (PML) revealed that PML is a member of the lipase family I.3 and is composed of 617 amino acid residues with a calculated molecular weight of 64510. Recombinant PML (rPML) was overproduced in Escherichia coli in an insoluble form, solubilized in the presence of 8 M urea, purified in a urea-denatured form and refolded by removing urea in the presence of the Ca(2+) ion. Gel filtration chromatography suggests that this refolded protein is monomeric. rPML showed relatively broad substrate specificities and hydrolyzed glyceryl tributyrate and olive oil with comparable efficiencies. rPML was active only in the form of a holo-enzyme, in which at least 12 Ca(2+) ions bound. These Ca(2+) ions bound too tightly to be removed from the protein upon dialysis, but were removed from it upon EDTA treatment. The resultant apo-enzyme was fully active in the presence of 10 mM CaCl(2), but was inactive in the absence of the Ca(2+) ion. PML has a GXSXG motif, which is conserved in lipases/esterases and generally contains the active-site serine. The mutation of Ser(207) within this motif to Ala completely inactivated PML, suggesting that Ser(207) is the active-site serine of PML.

alpha-Helical solenoid model for the human involucrin

Involucrin is a key component of the cross-linked envelope of terminally differentiated keratinocytes. The human molecule largely consists of 10 residue repeats and forms a thin 460 A long rod. Summarized experimental data and a detailed stereochemical analysis made with computer modeling resulted in a structural model for the involucrin molecule. The suggested structure is a left-handed alpha-helical solenoid built of a tandem array of helix-turn-helix folds. The structure enables us to explain the whole set of experimental data and residue conservations within the repeats. It is ideally suited to serve as a scaffold for cell envelope assembly and proposes a possible mode of the intermolecular interactions of involucrin during cell cornification.

Folding of a synthetic parallel beta-roll protein

Recently, the design of beta-sheet proteins and concomitant folding studies have attracted increasing attention. A unique natural all-beta domain occurs in a family of cytolytic bacterial toxins, the so-called RTX toxins. This domain consists of a variable number (about 6-45) of tandem repeats of a glycine-rich nine-residue motif with the consensus sequence GGXGXDX(L/I/F)X. The analysis of the three-dimensional structure of alkaline protease from Pseudomonas aeruginosa which possesses six of these repeats revealed that they fold into a novel 'parallel beta-roll' where calcium is bound within the turns connecting the beta-strands. A 75-mer peptide of the sequence NH(2)-WLS-[GGSGNDNLS](8)-COOH was chemically synthesised. Circular dichroism spectroscopy showed that this polypeptide folds in the presence of Ca(2+) and polyethylene glycol into a beta-structure which is presumably identical with the parallel beta-roll. This synthetic beta-roll behaves similarly to the isolated beta-roll domains from Escherichia coli haemolysin or Bordetella pertussis cyclolysin in terms of calcium binding and polymerisation behaviour.

Do parallel beta-helix proteins have a unique fourier transform infrared spectrum?

Several polypeptides have been found to adopt an unusual domain structure known as the parallel beta-helix. These domains are characterized by parallel beta-strands, three of which form a single parallel beta-helix coil, and lead to long, extended beta-sheets. We have used ATR-FTIR (attenuated total reflectance-fourier transform infrared spectroscopy) to analyze the secondary structure of representative examples of this class of protein. Because the three-dimensional structures of parallel beta-helix proteins are unique, we initiated this study to determine if there was a corresponding unique FTIR signal associated with the parallel beta-helix conformation. Analysis of the amide I region, emanating from the carbonyl stretch vibration, reveals a strong absorbance band at 1638 cm(-1) in each of the parallel beta-helix proteins. This band is assigned to the parallel beta-sheet structure. However, components at this frequency are also commonly observed for beta-sheets in many classes of globular proteins. Thus we conclude that there is no unique infrared signature for parallel beta-helix structure. Additional contributions in the 1638 cm(-1) region, and at lower frequencies, were ascribed to hydrogen bonding between the coils in the loop/turn regions and amide side-chain interactions, respectively. A 13-residue peptide that forms fibrils and has been proposed to form beta-helical structure was also examined, and its FTIR spectrum was compared to that of the parallel beta-helix proteins.

Mannuronan C-5 epimerases and cellular differentiation of Azotobacter vinelandii

Differentiation in Azotobacter vinelandii involves the encystment of the vegetative cell under adverse environmental circumstances and the germination of the resting cell into the vegetative state when growth conditions are satisfactory again. Morphologically, the encystment process involves the development of a protective coat around the resting cell. This coat partly consists of multiple layers of alginate, which is a copolymer of beta-D-mannuronic acid (M) and alpha-L-guluronic acid (G). Alginate contributes to coat rigidity by virtue of a high content of GG blocks. Such block structures are generated through a family of mannuronan C-5 epimerases that convert M to G after polymerization. Results from immunodetection and light microscopy, using stains that distinguish between different cyst components and types, indicate a correlation between cyst coat organization and the amount and appearance of mannuronan C-5 epimerases in the extracellular medium and attached to the cells. Specific roles of individual members of the epimerase family are indicated. Calcium and magnesium ions appear to have different roles in the structural organization of the cyst coat. Also reported is a new gene sharing strong sequence homology with parts of the epimerase-encoded R-modules. This gene is located within the epimerase gene cluster of Azotobacter vinelandii.

Effects of calcium and protons on the secondary structure of the nodulation protein NodO from Rhizobium leguminosarum biovar viciae

NodO, a 30-kDa nodulation protein secreted by Rhizobium leguminosarum biovar viciae, belongs to a family of proteins produced by Gram-negative bacteria containing a variable number of glycine/aspartates nonapeptides. In some instances, these are organized into a parallel beta-roll structure and bind Ca(2+) (one ion per repeat). To gain insight into NodO's secondary and tertiary structures, and their dependence upon Ca(2+) binding, we performed fluorescence experiments and FTIR spectroscopy. We found that calcium binds to the protein, promoting about a 10% increase in beta-structure mainly to the expense of random-coil. Protons can also induce a reversible change in NodO structure, as indicated by quenching of intrinsic tryptophan fluorescence and binding of ANS, albeit probably via a different mechanism. Tb(3+), a trivalent lanthanide, can compete with Ca(2+) for the same binding sites, but with higher affinity. The number of Ca(2+) binding sites, estimated by FTIR spectroscopy, was found to be consistent with the number of predicted repeats.

High resolution crystal structures of the Escherichia coli lytic transglycosylase Slt70 and its complex with a peptidoglycan fragment

The 70 kDa soluble lytic transglycosylase (Slt70) from Escherichia coli is an exo-muramidase, that catalyses the cleavage of the glycosidic bonds between N -acetylmuramic acid and N -acetylglucosamine residues in peptidoglycan, the main structural component of the bacterial cell wall. This cleavage is accompanied by the formation of a 1,6-anhydro bond between the C1 and O6 atoms in the N -acetylmuramic acid residue (anhMurNAc). Crystallographic studies at medium resolution revealed that Slt70 is a multi-domain protein consisting of a large ring-shaped alpha-superhelix with on top a catalytic domain, which resembles the fold of goose-type lysozyme. Here we report the crystal structures of native Slt70 and of its complex with a 1,6-anhydromuropeptide solved at nominal resolutions of 1.65 A and 1.90 A, respectively. The high resolution native structure reveals the details on the hydrogen bonds, electrostatic and hydrophobic interactions that stabilise the catalytic domain and the alpha-superhelix. The building-block of the alpha-superhelix is an "up-down-up-down" four-alpha-helix bundle involving both parallel and antiparallel helix pairs. Stabilisation of the fold is provided through an extensive packing of apolar atoms, mostly from leucine and alanine residues. It lacks, however, an internal consensus sequence that characterises other super-secondary helical folds like the beta-helix in pectate lyase or the (beta-alpha)-helix in the ribonuclease inhibitor. The 1, 6-anhydromuropeptide product binds in a shallow groove adjacent to the peptidoglycan-binding groove of the catalytic domain. The groove is formed by conserved residues at the interface of the catalytic domain and the alpha-superhelix. The structure of the Slt70-1, 6-anhydromuropeptide complex confirms the presence of a specific binding-site for the peptide moieties of the peptidoglycan and it substantiates the notion that Slt70 starts the cleavage reaction at the anhMurNAc end of the peptidoglycan.

Structural predictions of AgfA, the insoluble fimbrial subunit of Salmonella thin aggregative fimbriae

The unusually stable and multifunctional, thin aggregative fimbriae common to all Salmonella spp. are principally polymers of the fimbrin subunit, AgfA. AgfA of Salmonella enteritidis consists of two domains: a protease-sensitive, 22 amino acid residue N-terminal region and a protease-resistant, 109 residue C-terminal core. The unusual amino acid sequence of the AgfA core region comprises two-, five- and tenfold internal sequence homology patterns reflected in five conserved, 18-residue tandem repeats. These repeats have the consensus sequence, Sx5QxGx2NxAx3Q and are linked together by four or five residues, (x)xAx2. The predicted secondary structure for this unusual arrangement of tandem repeats in AgfA indicates mainly extended conformation with the beta strands linked by four to six residues. Candidate proteins of known structure with motifs of alternating beta strands and short loops were selected from folds described in SCOP as a source of coordinates for AgfA model construction. Three all-beta class motifs selected from the Serratia marcescens metalloprotease, myelin P2 protein or vitelline membrane outer protein I were used for initial AgfA homology build-up procedures ultimately resulting in three structural models; beta barrel, beta prism and parallel beta helix. The beta barrel model is a compact, albeit irregular structure, with the beta strands arranged in two antiparallel beta sheet faces. The beta prism model does not reflect the 5 or 10-fold symmetry of the AgfA primary sequence. However, the favored, parallel beta helix model is a compact coil of ten helically arranged beta strands forming two parallel beta sheet faces. This arrangement predicts a regular, potentially stable, C-terminal core region consistent with the observed tandem repeat sequences, protease-resistance and strong tendency of this fimbrin to oligomerize and aggregate. Positional conservation of amino acid residues in AgfA and the Escherichia coli AgfA homologue, CsgA, provides strong support for this model. The parallel beta helix model of AgfA offers an interesting solution to a multifunctional fimbrin molecular surface having solvent exposed areas, regions for major and minor subunit interactions as well as fiber-fiber interactions common to many bacterial fimbriae.