AlgK Is a TPR-Containing Protein and the Periplasmic Component of a Novel Exopolysaccharide Secretin

Structure

Article AlgK Is a TPR-Containing Protein and the Periplasmic Component of a Novel Exopolysaccharide Secretin Carrie-Lynn Keiski,1,2 Michael Harwich,3 Sumita Jain,3,6 Ana Mirela Neculai,1 Patrick Yip,1 Howard Robinson,4 John C. Whitney,1,2 Laura Riley,1 Lori L. Burrows,1,5 Dennis E. Ohman,3,* and P. Lynne Howell1,2,* 1Molecular

Structure and Function, Hospital for Sick Children, Toronto, ON M5G 1X8, Canada of Biochemistry, University of Toronto, Toronto, ON M5S 1A8, Canada 3Department of Microbiology and Immunology, Virginia Commonwealth University Medical Center, Richmond, VA 23298-0678, USA 4Biology Department, Brookhaven National Laboratory, Upton, NY 11973-5000, USA 5Department of Biochemistry and Biomedical Sciences and The Michael G. DeGroote Institute for Infectious Disease Research, McMaster University, Hamilton, ON L8N 3Z5, Canada 6Present address: Department of Periodontics, University of Washington School of Dentistry, Seattle, WA 98195, USA *Correspondence: [email protected] (P.L.H.), [email protected] (D.E.O.) DOI 10.1016/j.str.2009.11.015 2Department

SUMMARY

The opportunistic pathogen Pseudomonas aeruginosa causes chronic biofilm infections in cystic fibrosis patients. During colonization of the lung, P. aeruginosa converts to a mucoid phenotype characterized by overproduction of the exopolysaccharide alginate. Here we show that AlgK, a protein essential for production of high molecular weight alginate, is an outer membrane lipoprotein that contributes to the correct localization of the porin AlgE. Our 2.5 A˚ structure shows AlgK is composed of 9.5 tetratricopeptide-like repeats, and three putative sites of protein-protein interaction have been identified. Bioinformatics analysis suggests that BcsA, PgaA, and PelB, involved in the production and export of cellulose, poly-b-1,6-N-Acetyl-D-glucosamine, and Pel exopolysaccharide, respectively, share the same topology as AlgK/E. Together, our data suggest that AlgK plays a role in the assembly of the alginate biosynthetic complex and represents the periplasmic component of a new type of outer membrane secretin that differs from canonical bacterial capsular polysaccharide secretion systems.

INTRODUCTION Biofilms, defined as surface-attached bacteria encased in a hydrated polymeric matrix, are the cause of many persistent and chronic bacterial infections. Biofilms are notoriously difficult to eradicate because the embedded bacteria are less susceptible to antibiotics and the host immune response (Hall-Stoodley et al., 2004). The typical polymeric matrix is composed primarily of exopolysaccharides that are secreted by the bacteria and remain loosely associated with their surfaces (Branda et al., 2005). Exopolysaccharides have also been implicated in adhesion, development, and maintenance of the structure of mature

biofilms (Danese et al., 2000; O’Toole et al., 2000; Ryder et al., 2007). Examples of biofilm exopolysaccharides include colanic acid, poly-b-1,6-N-acetyl-D-glucosamine (b-1-6-GlcNAc, also known as polysaccharide intercellular adhesin or PIA), cellulose, alginate, and the polysaccharides encoded by the pel and psl operons of Pseudomonas aeruginosa. The genetic loci responsible for the production of each of these polysaccharides have been identified, but many questions regarding how the exopolysaccharides are polymerized, exported to the periplasm, and secreted remain to be answered. A second type of polysaccharide secreted by bacteria are capsular polysaccharides, which unlike exopolysaccharides remain tightly associated with the cell surface and are important virulence factors (Whitfield, 2006). Two different mechanisms have been proposed for capsular polysaccharide production. The difference between the two mechanisms occurs at the inner membrane. Group I and IV capsular polysaccharides in Escherichia coli are produced by the Wzy-dependent pathway, which uses the integral inner membrane protein Wzx to ‘‘flip’’ the polysaccharide repeat units across the inner membrane from the cytoplasm to the periplasm for polymerization by Wzy, while group II and III capsular polysaccharides require an ABC-transporter to transport the polymerized polysaccharide across the inner membrane (Whitfield, 2006). Export across the outer membrane occurs via the novel a-helical barrel transporter Wza and its ortholog KpsD in group I/IV and II/III capsular polysaccharide biosynthesis, respectively. Colanic acid is produced by the Wzy-dependent pathway (Whitfield, 2006) and the Psl exopolysaccharide of P. aeruginosa may be synthesized by a similar mechanism given the presence of genes encoding putative homologs of this system in the psl biosynthetic locus (Friedman and Kolter, 2004). In contrast, the gene clusters responsible for the production of poly-b-1,6-GlcNAc, Pel-polysaccharide, cellulose, and alginate lack signature homologs from either of the known capsular polysaccharide biosynthetic pathways, suggesting that these exopolysaccharides may be synthesized and secreted by a novel mechanism. Alginate production by P. aeruginosa is one of the best-characterized exopolysaccharide biosynthetic systems. P. aeruginosa establishes chronic biofilm infections in the lungs of cystic

Structure 18, 265–273, February 10, 2010 ª2010 Elsevier Ltd All rights reserved 265

Structure Structure of AlgK

fibrosis patients, where specific environmental cues induce the bacteria to secrete copious amounts of alginate (Govan and Deretic, 1996). Alginate is a negatively charged polymer of b1-4linked D-mannuronate and its C5 epimer, L-guluronate. Thirteen genes have been implicated in the biosynthesis and secretion of alginate, and with the exception of algC, all are located in the algD operon (Jain and Ohman, 2004). AlgA, AlgC, and AlgD are cytoplasmic enzymes that synthesize the activated sugar nucleotide precursor of alginate, GDP-mannuronate (Jain and Ohman, 2004). Alginate polymerization and transport across the inner membrane appears to require the action of two inner membrane proteins: Alg8, which has sequence homology to family 2 glycosyltransferases (Remminghorst and Rehm, 2006b), and the PilZ domain-containing protein Alg44 (Merighi et al., 2007; Remminghorst and Rehm, 2006a). Once in the periplasm, D-mannuronate residues in the nascent polymer can be epimerized to L-guluronate by AlgG, selectively acetylated at the 20 and/or 30 hydroxyls by the concerted action of AlgI, AlgJ, and AlgF, or remain unaltered (Jain and Ohman, 2004). AlgK and AlgX are periplasmic proteins essential for the secretion of high molecular weight alginate; however, their exact role in alginate biosynthesis remains unclear (Jain and Ohman, 1998; Robles-Price et al., 2004). The algD operon also encodes a lyase, AlgL, that degrades alginate polymers that accumulate in the periplasmic space (Jain and Ohman, 2005). Alginate is secreted from the cell via the outer membrane b-barrel porin AlgE (Rehm et al., 1994). Together these proteins are believed to assemble into a large heterooligomeric complex, which like other bacterial secretion machineries, is thought to span the entire cell envelope, a hypothesis that is supported by the finding that both inner and outer membrane components are needed for alginate polymerization in vitro (Remminghorst and Rehm, 2006b). In the current study, we examined the role of AlgK in alginate biosynthesis. We show that AlgK is an outer membrane lipoprotein that contributes to the localization of the AlgE porin to the outer membrane. The structure of AlgK was determined to 2.5 A˚ resolution and reveals 9.5 tetratricopeptide repeat (TPR)-like motifs. The presence of these protein-protein interaction motifs suggests that AlgK plays an important role in the assembly of the alginate biosynthetic complex. Bioinformatics analyses and literature searches revealed that the BcsC, PgaA, and PelB proteins, implicated in the production and export of cellulose, poly b-1-6-GlcNAc, and Pel polysaccharide, respectively, are each predicted to contain both porin and TPR domains. These findings suggest AlgK/E may assemble into a novel exopolysaccharide secretin, where the TPR domain acts as a scaffold for the assembly of the biosynthetic complex and the porin domain facilitates the transport of alginate across the outer membrane. RESULTS AlgK Is an Outer Membrane Lipoprotein AlgK is expressed in the cytoplasm as a 52.5 kDa preprotein that is processed within cells to a 50 kDa mature protein localized in the periplasm (Aarons et al., 1997; Jain and Ohman, 1998). Sequence analysis suggests that AlgK is a lipoprotein, with a canonical lipobox containing a lipidation site at Cys28. Mutating Cys28 to serine has been shown to interfere with alginate production (see Tables S1 and S2 and Figures S1 and S2

Figure 1. Lipidation and Subcellular Localization of AlgK (A) Autoradiogram of E. coli cells grown in the presence of C14-palmitic acid and transformed with either the empty vector or a vector expressing AlgK. (B) Membrane fractions of FRD1 cells (top panels) blotted for AlgK and AlgE and of FRDDalgK cells (Table S1) (bottom panel) probed for AlgE. The membrane fractions correspond to the inner membrane (IM), outer membrane 1 (OM1), outer membrane 2 (OM2), and a mixture of inner and outer membranes (M). Succinate dehydrogease (SDH) activity (nanomoles of DCPIP reduced per minute per milligram of total protein) is presented numerically beneath the corresponding membrane fraction. (See also Figures S1 and S2.)

available online). To confirm that AlgK is lipidated in vivo, we supplemented P. aeruginosa FRD1 cultures with C14-labeled palmitic acid. However, this approach resulted in most of the expressed proteins being labeled because the bacteria catabolized the palmitic acid (data not shown). When the experiment was repeated in Escherichia coli JM109(DE3) carrying a plasmid expressing AlgK, the autoradiogram revealed a band of a mass consistent with that of AlgK that was C14-labeled by palmitic acid and absent in cells transformed with the vector alone (Figure 1A). As lipoproteins can be tethered to either the inner or outer membrane, we used sucrose gradient fractionation of FRD1 cells to separate inner and outer membrane components, which were then probed for AlgK using specific antisera. The activity of the inner membrane enzyme succinate dehydrogenase was used as a fractionation control. We observed AlgK only in fractions containing the outer membrane; AlgK was not observed in the inner membrane fraction (Figure 1B). These results, coupled with previous data (Aarons et al., 1997; Jain and Ohman, 1998), show that AlgK is a periplasmic lipoprotein that is sorted to the outer membrane. AlgK Contributes to the Proper Localization of AlgE Given the role that lipoproteins play in the insertion of proteins into the outer membrane (Bayan et al., 2006), we investigated whether the absence of AlgK affected the localization of AlgE, the b-barrel porin responsible for transporting alginate across the outer membrane (Rehm et al., 1994). Sucrose density fractionation of FRD1 cells revealed that AlgE is present only in outer membrane fractions (Figure 1B). This result is in sharp contrast to its localization in FRDDalgK cells, where AlgE is found in both inner and outer membrane fractions. In FRDalgK-C28S cells, an FRD1-derivative expressing a C28S mutant of AlgK lacking its lipidation site, AlgE is also found in both the inner and outer

266 Structure 18, 265–273, February 10, 2010 ª2010 Elsevier Ltd All rights reserved

Structure Structure of AlgK

Figure 2. AlgK Structure (A) Ribbon representation of molecule C of AlgK shown in two orthogonal orientations. The TPRlike motifs R1–R10 are labeled. Helix A and B of each repeat are found on the concave and convex surface of the protein, respectively. Cys168 and Cys178 involved in the disulphide bond are colored in red and labeled. Residues 35–45 are highlighted in hot pink. (B) Schematic diagram showing the organization of a helices (labeled H1, H22, and H23) and TPRlike motifs (labeled R1–R10) in AlgK. Discontinuities in the schematic diagram represent regions of molecule C that could not be modeled in the electron density. The disulphide bond between TPRs 4 and 5 is shown in red. The numbers above and below each schematic helix represent the amino acid residues found in the helix. a and b refer to the two helices that make up each TPR motif. (See also Figures S3–S5.)

membrane (Figure S2B), further supporting our conclusion that AlgK plays an important role in the proper localization of AlgE.

dues 35–45 may also fold into a helix that corresponds to R1A and that AlgK likely carries ten TPR-like repeats (Figure S4).

Structure of AlgK As no good quality crystals of P. aeruginosa AlgK could be obtained, we determined the structure of P. fluorescens WCS 374r AlgK at 2.5 A˚ resolution (Figure 2 and Table 1). AlgK from P. fluorescens WCS374r has 58.4% sequence identity and 76% similarity with its P. aeruginosa homolog. Although AlgK was found by analytical ultracentrifugation to be predominantly a monomer in solution (data not shown), four molecules are present in the asymmetric unit. These molecules exhibit significant conformational heterogeneity, suggesting that AlgK is a flexible molecule (Figure S3). The most well ordered and complete molecule of AlgK in the asymmetric unit, molecule C, contains 22 a helices that pack into a right-handed superhelix that is 70 A˚ long and 50 A˚ wide with an internal diameter of 20–30 A˚

Structural Homologs of AlgK Structural homologs of AlgK were identified using the DALI server (Holm and Sander, 1993). Although the order and score of the proteins identified depended on which AlgK molecule was used as the query, not surprisingly given the arrangement of pairs of antiparallel a helices found in AlgK, the list of homologs is dominated by proteins belonging to the TPR superfamily including: O-linked N-acetylglucosamine transferase from Homo sapiens, rmsd of 8.9 A˚ for 284 residues (Jinek et al., 2004); PilF from P. aeruginosa, rmsd of 2.5 A˚ for 155 residues (Kim et al., 2006; Koo et al., 2008) (Figure S5); and HcpC from Helicobacter pylori, rmsd of 2.9 A˚ for 152 residues (Luthy et al., 2004).

(Figure 2; Table S3). At the N terminus, a helix H1 (residues 17– 32) and residues 33–45 (Figure 2; Figure S4) pack against the concave surface of the superhelix (Figure 2). Please note that the numbering used to describe the structure refers to the mature protein lacking its signal sequence. The body of the superhelix is composed of 21 antiparallel a helices with residues 55–384 forming 9.5 TPR-like motifs (Figure 2; Figure S4). These repeats vary in length from 34 to 40 residues. The inter-repeat packing angle between repeats R3–R4, R4–R5, and R5–R6 (17 , 51 , and 16 , respectively) are atypical for TPRs and are most likely the result of the highly conserved disulfide bond between helices R4B and R5A (Cys 168 and 178). In addition, while secondary structure prediction algorithms suggest that residues 35–45 fold into a helix, in the structure they are predominantly in an extended conformation. Since TPRs are normally formed by pairs of helices, we examined the amino acid sequence of this putative helix and found that it carries most of the helix A TPR consensus sequence found for AlgK. This suggests that under more biologically relevant conditions, resi-

Potential Sites of Protein-Protein Interaction Given the role TPR motifs play in mediating protein-protein interactions, potential binding sites on AlgK were identified. A multiple sequence alignment of closely and distantly related AlgK homologs shows that most of the highly conserved residues map to the N- and C-terminal regions of the protein (Figure S4). The highly conserved residues, which are not part of the TPR consensus sequence, were mapped onto the surface of the AlgK structure (Figure 3) and, interestingly, we found that few conserved residues are found on the concave surface of the protein; these are located mostly on the convex surface or in loop regions. Several clusters of highly conserved residues are visible on the surface of the protein, which may correspond to potential sites of protein-protein interaction. The first potential binding site is on helix H1, where residues Asp18, Ala22, Asn25, Leu29, and Ala30 all lie on the same face of the helix. These residues are buried when helix H1 interacts with the concave surface of the superhelix (Figure 3A). The second potential site is composed of several charged residues, Arg70, Arg77, and Glu105, on repeats 2 and 3, which cluster on the concave

Structure 18, 265–273, February 10, 2010 ª2010 Elsevier Ltd All rights reserved 267

Structure Structure of AlgK

Table 1. X-Ray Data Collection and Refinement Statistics Native

SeMet-SAD

P21

P21

Data Collection Space group Cell dimensions a, b, c (A˚)

79.1, 107.8, 119.0

78.2, 107.0 120.1

a, b, g ( )

90.0, 97.0, 90.0

90.0, 96.1, 90.0

Resolution (A˚)

50.0–2.50 (2.60–2.50)

30.00–2.80 (2.90–2.80)

Rmerge

0.102 (0.546)

0.109 (0.694)

I/sI

14.96 (4.19)

43.9 (3.15)

Completeness (%)

99.9 (100)

99.2 (98.6)

Redundancy

7.49 (7.53)

14.4 (10.9)

Refinement Resolution (A˚)

49.05–2.50

No. reflections

68663

Rwork / Rfreea

0.22 / 0.28

No. atoms Proteins

11461

Glycerol

24

Chloride

3

Figure 3. Putative Sites of Protein-Protein Interaction

Water

853

(A) Schematic ribbon representation of helix H1, in isolation and in the context of the rest of the protein. The conserved residues of the first putative binding site are shown in orange. (B) Surface representation of AlgK in three orthogonal orientations. Helix H1 and residues 33–45 and 428–438 have been omitted from the structure prior to rendering the surface. Conserved residues from putative binding sites 2 and 3 are colored in purple and green, respectively. (See also Figure S4.)

Average B factors Proteins

48.7

Glycerol

62.4

Chloride

66.6

Water

49.4

Rmsd Bond lengths (A˚)

0.010

Bond angles ( )

1.5

Values in parentheses are for the highest resolution shell. a Rfree was calculated using 5% of the reflections.

surface of the superhelix (Figure 3B). Arg70 and Arg77 overlap with the ‘‘hypervariable’’ positions identified in TPRs as being involved in substrate binding (Magliery and Regan, 2005). Arg74 also clusters with these residues in the structure and is conserved in most homologs, suggesting that it could also be part of this binding site. If residues 35–45 do form helix R1A, then residues Glu36 and Asp43 have the potential to contribute to this binding site. The third putative binding site consists of residues in the loops that connect helices A and B of repeat 9 (Gly331, Tyr332, Leu333, Gly334, and Lys335) and repeat 10 (Lys368, Gly369, and Asp373). These loops are located next to each other, on the same surface of the superhelix (Figure 3B). Outer Membrane Proteins Containing a TPR and Porin Domain Because alginate secretion involves both a TPR protein tethered to the outer membrane and an outer membrane porin, we investigated whether similar pairs of proteins are implicated in other exopolysaccharide secretion systems. Interestingly, PgaA, a protein required for the secretion of poly-b-1,6-GlcNAc, was recently predicted to have a large N-terminal TPR domain similar to O-linked N-acetylglucosamine transferase and a C-terminal

porin domain similar to FadL (Itoh et al., 2008). The results of our bioinformatics analysis support the presence of both TPR and porin domains in PgaA and predict that the PgaA homologs BpsA in Bordetella bronchiseptica and HmsH in Yersinia pestis would have a similar topology (Figure 4). Furthermore, we found that the bcs operons in E. coli K12 and Salmonella typhimurium and the pel operon in P. aeruginosa, which are responsible for the polymerization and export of cellulose and the Pel polysaccharide, respectively, each encode an outer membrane protein that we predict has an N-terminal TPR domain followed by a C-terminal porin domain (Figure 4).

DISCUSSION In this study, we showed that AlgK is an outer membrane lipoprotein that contributes to the localization of the porin AlgE. This finding correlates well with the role of other outer membrane lipoproteins in outer membrane biogenesis, e.g., transport and/or insertion of proteins, capsular polysaccharides, and lipopolysaccharides into the outer membrane (Nesper et al., 2003; Tanaka et al., 2001; Wu et al., 2005). How AlgK contributes to the localization of AlgE in the outer membrane is currently unknown. Similarly to other porins, AlgE is thought to be secreted across the inner membrane in an unfolded form by the SecYEG translocon, cleaved by signal peptidase I to remove its signal peptide, and transported to the outer membrane by chaperones such as SurA, Skp, and DegP. At the outer membrane, the BAM complex is thought to be responsible for folding and inserting AlgE into the

268 Structure 18, 265–273, February 10, 2010 ª2010 Elsevier Ltd All rights reserved

Structure Structure of AlgK

Figure 5. Bacterial Polysaccharide Biosynthetic Complexes

Figure 4. Exopolysaccharide Secretins Schematic representation of predicted TPR and porin domains in AlgK (P. aeruginosa PAO1), AlgE (P. aeruginosa PAO1), PgaA (E. coli K12), HmsH (Y. pestis), BpsA (B. bronchiseptica), BcsC E. coli (E. coli K12), BcsC S. typh (S. typhimurium), and PelB (P. aeruginosa PAO1). The porin domain is colored in yellow with the number of b strands predicted listed on the right. The predicted TPR domain is dark blue and the regions predicted to fold into helices are light blue. AlgK and AlgE are shown as they are organized in the algD operon in P. aeruginosa PAO1. The stop and start codons of the two genes overlap. The TPR and porin domain boundaries are listed beneath the proteins. Predicted signal sequences are in red.

membrane (Knowles et al., 2009). It is difficult to pinpoint where AlgK could be acting within this pathway: it could be involved in the translocation of AlgE across the periplasm or could act in conjunction with the BAM complex in promoting the folding and/or insertion of AlgE into the outer membrane. In both scenarios, the absence of AlgK would be expected to cause a ‘‘back-up’’ in the AlgE biogenesis pathway, with accumulation of AlgE at the inner membrane either in a misfolded/aggregated form or tethered to the membrane by an uncleaved signal peptide. Given that in the absence of AlgK, some AlgE is detected in the outer membrane, it would appear that while AlgK may facilitate AlgE localization, it is not essential. However, it is not clear whether the portion of AlgE found in the outer membrane fractions of the algK mutant is correctly folded and/ or functional. In general, bacterial polysaccharide secretion appears to be mediated by a trans-envelope complex, consisting of a minimum of an inner membrane transporter, an outer membrane transporter, and an inner membrane component or components that link the two transporters in the periplasm (Whitfield and Naismith, 2008). The best described polysaccharide secretion systems in bacteria are the Wzy-dependent and ABC transporter-dependent pathways responsible for capsular polysaccharide production in E. coli, which are summarized in Figure 5. The alginate operon also appears to encode all three components of a trans-envelope secretion complex, each of which is distinct from their counterparts in the capsular biosynthetic systems. The integral inner membrane glycosyltransferase

Cartoon representation of the multiprotein complexes thought to be responsible for synthesis and secretion of alginate, poly b-1-6-GlcNAc, cellulose exopolysaccharide, and group 1/IV and II/III capsular polysaccharide production. The systems only include those proteins thought to form the basic components of the trans-envelope complex and the enzymes responsible for polymerization. The glycosyltransferases and Wzy polymerase are orange and yellow, respectively. Proteins responsible for the transport of capsular polysaccharides across the inner membrane are shown in green. Inner membrane proteins needed for polymerization that bridge the inner and outer membrane biosynthetic machinery are shown in blue. The outer membrane secretin is pink for all systems, except alginate, where AlgE is pink and AlgK is light pink. Capsular polysaccharide models were adapted from Figures 4 and 7 in Whitfield (2006). IM and OM represent the inner and outer membrane, respectively.

Alg8, along with Alg44, is believed to be responsible for synthesizing and transporting the polymer across the inner membrane (Ogelsby et al., 2008; Remminghorst and Rehm, 2006b). The integral outer membrane b-barrel protein AlgE is involved in alginate transport across the outer membrane into the extracellular milieu. Given the structure of AlgK, we propose that this protein acts as a scaffold for the assembly of a multiprotein complex that links the inner and outer membrane components of the alginate secretion complex (Figure 5). This hypothesis is in keeping with previous studies that have suggested that AlgK is a member of a periplasmic channel responsible for alginate secretion, as the FRDDalgK mutant was non-mucoid and secreted low molecular weight uronic acids rather than high molecular weight alginate (Jain and Ohman, 1998). The uronic acids were identified to be AlgL degradation products. A similar non-mucoid, uronic acidproducing phenotype has also been observed for the FRDDalgG and FRDDalgX mutants (Jain et al., 2003; Robles-Price et al., 2004). AlgL is a periplasmic lyase that appears to play a dual role in alginate production, in the degradation of polymer, and as a member of the secretion complex itself: in the absence of AlgL, mucoid cells rapidly accumulate alginate in their periplasm and burst (Jain and Ohman, 2005). In light of these findings, AlgK, along with AlgX, AlgG, and AlgL, has been proposed to assemble into a periplasmic channel for alginate secretion (Jain et al., 2003; Jain and Ohman, 1998, 2005; Robles-Price et al., 2004). AlgE is a b-barrel protein predicted to have 18 transmembrane strands interspersed by short periplasmic and longer extracellular loops (Rehm et al., 1994; Whitney et al., 2009). This topological prediction suggests that AlgE lacks a domain equivalent to the large 100 A˚ long periplasmic domain of Wza. This deficit would require that other periplasmic proteins or domains form the bulk of the periplasmic conduit, a role that we propose is fulfilled, at least in part, by the outer membrane lipoprotein AlgK. The large number of TPR-like motifs in AlgK makes it an excellent

Structure 18, 265–273, February 10, 2010 ª2010 Elsevier Ltd All rights reserved 269

Structure Structure of AlgK

candidate to bind to multiple proteins and to act as a scaffold for the assembly of the periplasmic conduit linking AlgE to Alg8 and Alg44 in the inner membrane. This idea is not without precedent, as multiple proteins have been shown to bind the TPR-containing proteins Hop, PscG, and YscG (Quinaud et al., 2007; Scheufler et al., 2000; Sun et al., 2008). Three clusters of highly conserved residues were identified on the surface of our model, corresponding to potential sites of protein-protein interaction. Two of these putative binding sites are close to the N terminus, which is attached to the outer membrane via its lipid moiety and thus positioned to interact directly with the periplasmic loops of AlgE. While the conservation of residues on one face of helix H1 at the N terminus of the protein may suggest that this region is involved in protein-protein interactions, we hypothesize that the association of helix H1 and residues 35–45 to the concave surface of the superhelix, observed in the crystal structure, is not biologically relevant. We believe that this conformation is the consequence, in part, of the absence of AlgK’s binding partner(s) and the lack of lipidation at the N terminus of this helix. As detailed above, we predict that residues 35–45 form helix R1A of the first TPR-like motif and as such would be expected to bind to the hydrophobic face of helix R1B. This interaction would in turn alter the location and packing of helix H1. The third putative binding site in AlgK is located at the C terminus of the protein and therefore could bind to periplasmic and/or inner membrane proteins. Periplasmic proteins that could bind to the C terminus of AlgK include AlgG, AlgX, and AlgL, all of which, as described above, have been implicated in the formation of the periplasmic conduit (Jain et al., 2003; Jain and Ohman, 2005; Robles-Price et al., 2004). AlgK could also interact with the inner membrane protein Alg44. Alg44 has a cytoplasmic N-terminal PilZ domain (Merighi et al., 2007), a predicted single transmembrane region and a C-terminal periplasmic domain that is thought to be structurally similar to HlyD, AcrA, and MexA (Merighi et al., 2007; Remminghorst and Rehm, 2006a). HlyD, AcrA, and MexA are members of the membrane fusion protein family that couple an inner membrane transporter to an outer membrane factor, such as TolC, to form a channel to translocate substrates across the outer membrane (Eswaran et al., 2004). The predicted structural similarity of Alg44 to HlyD, AcrA, and MexA and the observation that the in vivo stability of AlgE appears to be dependent on the presence of Alg44 (Ogelsby et al., 2008) suggests that Alg44’s periplasmic domain may interact with AlgE and/or AlgK in the outer membrane. Closer examination of the proteins involved in the biosynthesis of the homopolymeric alginate, cellulose, and poly b-1-6GlcNAc polysaccharides reveals remarkable similarities between the systems. The bcs and pga operons responsible for cellulose and poly-b-1,6-GlcNAc polymerization and export in E. coli each encode four proteins, BcsABZC and PgaABCD, respectively (Romling, 2002; Wang et al., 2004). While these systems have fewer than the seven proteins implicated in alginate polymerization and export that are encoded by the algD operon, each appears to have four elements in common and to contain the necessary building blocks for a trans-envelope secretion complex. First, the alg, bcs, and pga operons each encode an integral inner membrane protein that is required for the polymerization of the homopolymer. Each of these proteins, namely Alg8, PgaC, and BcsA, is predicted to have multiple

transmembrane helices and a large soluble domain that exhibits sequence similarity to family 2 glycosyltransferases (Itoh et al., 2008; Ogelsby et al., 2008; Romling, 2002; Wang et al., 2004; Wong et al., 1990). Moreover, each of these polymerases appears to be associated with a second inner membrane protein that has been shown to be essential for polymerization, namely Alg44, BcsB, and PgaD (Itoh et al., 2008; Ogelsby et al., 2008; Romling, 2002; Wang et al., 2004; Wong et al., 1990). Perhaps of greater significance, given the structure of AlgK presented here, is that the bcs and pga operons each encode a protein, BcsC and PgaA, respectively, that is predicted to contain a large periplasmic N-terminal TPR-containing domain and a C-terminal b-barrel porin (Figure 4). In alginate production, the TPR and porin domains are encoded by algK and algE proteins, respectively (Aarons et al., 1997). The evolutionary fusion of two genes to form a single gene is a well established phenomenon (Marcotte et al., 1999) and further strengthens our proposal that AlgK and AlgE interact directly and suggests that AlgK/AlgE, PgaA, and BcsC define a new family of secretins where the porin domain mediates secretion of polysaccharide across the outer membrane and the TPR domain acts as scaffold for the assembly of the periplasmic and inner membrane components of the biosynthetic complex. We propose that in cellulose and poly-b-1,6-GlcNAc production, like alginate production, the outer membrane TPR/porin protein forms a stable trans-envelope complex with the inner membrane glycosyltransferase and/or its associated protein and that the large TPR-containing protein is responsible for binding or recruiting various proteins for the assembly of the complex (Figure 5). It is interesting to note that while the inner membrane proteins encoded by the pel operon appear to be different from those encoded by the algD, bcs, and pga operons, our bioinformatics analysis did reveal that PelB has a similar predicted topology to AlgK/E, PgaA, and BcsC, i.e., a large TPR-containing N-terminal domain fused to a C-terminal porin domain. Query of the superfamily server (Gough et al., 2001) also shows this domain combination to be present in a number of other uncharacterized bacterial proteins, suggesting that this secretion mechanism may be more widespread than currently appreciated. In summary, we have shown that AlgK is an outer membrane lipoprotein with at least 9.5 TPR-like repeats that contributes to the proper localization of AlgE in the outer membrane. In addition to its potential role in AlgE biogenesis, we propose that AlgK acts as a scaffold that binds to multiple Alg proteins, such as AlgE and Alg44, to form a trans-periplasmic channel for the secretion of alginate. The biosynthetic machinery responsible for alginate production appears to have a different architecture compared to well established capsular polysaccharide systems and may represent the prototype of a novel biosynthetic/secretion system. The secretion of other exopolysaccharides such as cellulose and poly-b-1,6-GlcNac appears to be similar to alginate, with AlgK/AlgE, PgaA, and BcsC defining a new family of polysaccharide secretins. Improving our understanding of the mechanism underlying alginate, cellulose and poly b-1-6-GlcNac synthesis and secretion is of great biomedical interest given that such biofilm exopolysaccharides are key virulence factors of several important human pathogens (Darby, 2008; Govan and Deretic, 1996; O’Gara, 2007; Zogaj et al., 2001).

270 Structure 18, 265–273, February 10, 2010 ª2010 Elsevier Ltd All rights reserved

Structure Structure of AlgK

EXPERIMENTAL PROCEDURES Palmitic Acid Labeling An overnight culture of FRD1 (Table S1) was subcultured in 10 ml of modified alginate-producing medium (see Supplemental Information) to an OD600 of 0.5, at which point 10 mCi of 14C- palmitic acid was added to the medium. Cells were cultured for an additional 90 min, and then harvested and resuspended in 150 ml of Laemmli sample buffer. Samples were separated using SDS-PAGE, and then transferred to a PVDF membrane and exposed to film (Kodak). A similar protocol was used to radiolabel and detect AlgK expressed in E. coli JM109(DE3) cells transformed with pSJ172. Membrane Fractionation Overnight cultures of FRD1 and the FRDDalgK mutant (Table S1) were subcultured in 4 liters of broth (10 g tryptone, 5 g yeast extract, and 5 g NaCl, per liter) at 37 C with vigorous shaking to an OD600 of 1.0. Cells were harvested and fractionated using sucrose gradient density fractionation following the protocol described by Hancock and Nikaido (1978) with the following modifications: (1) Tris-HCl (pH 8.3) instead of Tris-HCl (pH 8.0) was used to wash cells and resuspend cells prior to lysis; (2) cells were passed through a French press twice at 40,000 lbin2 instead of 15,000 lbin2; (3) for the second sucrose gradient, cells were spun for 18 hr instead of 14. The four resulting fractions correspond to the inner membrane, outer membrane, and a mixture of the inner and outer membranes. Structure Determination of AlgK The expression, purification, crystallization, and preliminary X-ray analysis of the native crystals of P. fluorescens WCS 374r AlgK were described previously (Keiski et al., 2007). Recombinant AlgK from P. fluorescens WCS 374r contains 4 methionines in 452 residues, not including the N-terminal methionine, and therefore to increase the phasing power obtainable in a selenomethionine (SeMet) single anomalous dispersion (SAD) experiment, a methionine triple mutant was generated. A multiple sequence alignment of AlgK homologs from P. fluorescens WCS 374r, P. aeruginosa PAO1, P. putida KT2440, P. syringae pv. tomato, and Azotobacter vinelandii was generated using T-Coffee (Notredame et al., 2000) and used to determine which residues would be the best candidates for mutagenesis. Residues Leu32, Ile44, and Leu399 were chosen as these positions are methionines in at least one of the other homologs and because leucine to methionine mutants have been shown to have the smallest isomorphous difference with the wild-type protein (Gassner and Matthews, 1999). The three mutations, L32M, I44M, and L399M, were introduced sequentially into pAlgK26ctag using the QuickChange SiteDirected Mutagenesis strategy (Stratagene) and verified by DNA sequencing. BL21 CodonPlus(DE3) cells (Stratagene) transformed with the triple mutant vector were grown in 1 liter of Luria broth at 37 C. When the OD600 of the culture reached 0.9, 140 ml of the culture was centrifuged at approximately 3000 3 g. The pellet was rinsed with M9 minimal media and resuspended in 1 liter of M9 supplemented with SeMet. The cells were grown to an OD600 of 0.6, induced with 0.5 mM IPTG, and grown overnight at 25 C. The triple mutant was purified as described previously for the native protein (Keiski et al., 2007) with the following modifications: (1) 2 mM DTT was used instead of 15 mM bmercaptoethanol in buffer C (0.1 mM MgCl2, 1 mM phenylmethylsulfonyl fluoride, and 1 tablet of EDTA-free Roche Complete protease inhibitor cocktail); (2) the protein-bound NiNTA resin was washed with 75 ml instead of 140 ml of buffer E [20 mM Tris-HCl (pH 7.5), 300 mM NaCl, 20 mM imidazole, 1 mM phenylmethylsulfonyl fluoride, 5 mM b-mercaptoethanol, and 1/2 tablet of EDTAfree Roche Complete protease inhibitor cocktail]; and (3) the SeMet protein eluted from the nickel column was concentrated to 20 mg/ml and exchanged into buffer G [20 mM Tris-HCl (pH 8.5), 20 mM NaCl, 0.2 mM EDTA, and 5 mM b-mercaptoethanol] using a concentrator with a 10 kDa MW cut-off (Amicon) prior to crystallization. Crystals of the triple mutant were grown using the same conditions as the native protein (Keiski et al., 2007) but with a protein concentration of 20 mg ml1 and without detergent. Prior to data collection at 172 C the crystals were cryo-protected as described previously (Keiski et al., 2007). Two SAD data sets were collected at the selenium peak on the same crystal at beam lines X25 and X29 (National Synchrotron Light Source) and the data were processed and merged using HKL-2000 (Otwinowski and Minor, 1997) (Table 1). Twenty-two selenium sites were located using Hkl2map

(Pape and Schneider, 2004) and density modified phases were calculated using Solve/Resolve (Terwilliger and Berendzen, 1999). The initial model of AlgK built into the experimental map consisted of residues 19–34 and 69– 314 and the NCS operators were used to generate the entire contents of the asymmetric unit. In the initial rounds of model refinement, NCS restraints were applied to all residues and in subsequent rounds of refinement to residues 108–301. After the R factor attained 30%, the NCS restraints were released and waters, glycerols, and chloride ions were added. The parameter and topology files for glycerol were obtained from the HIC-Up server (Kleywegt, 2007). Following numerous rounds of automated and manual model building using Resolve (Terwilliger, 2003) and Coot (Emsley and Cowtan, 2004), refinement using CNS (Brunger et al., 1998) converged to an R/Rfree of 0.22/0.28 (Table 1). Procheck v3.5.4 (Laskowski et al., 1993), ADIT (Kleywegt and Jones, 1996; Westbrook et al., 2003), amide_check v1.5 (G.D. Smith, Hospital for Sick Children), and Molprobity (Davis et al., 2007) were used during the model building process and to validate the final structure. The quality of the electron density prevented several residues in each molecule from being built (Table S3). Identification of Putative AlgK Binding Sites T-Coffee (Notredame et al., 2000) was used to generate a multiple sequence alignment of AlgK homologs from Azotobacter vinelandii, P. aeruginosa PAO1, P. fluorescens WCS 374r, Marinobacter algicola DG893, and Alcanivorax borkumensis SK2/ATCC700651/DSM11573. CONTACTS (G.D. Smith) and the protein interfaces, surfaces, and assemblies server PISA were used to identify and study molecular interfaces in the asymmetric unit (Krissinel and Henrick, 2007). Bioinformatic Characterization of Proteins The protein sequences for PgaA, HmsH, and BcsC from E. coli K12 and S. typhimurium were obtained from the UniProt Knowledgebase (SwissProt/ TrEMBL) and the sequences for BpsA and PelB were obtained from the National Center for Biotechnology Information protein database. PSORTb V2.0 (Gardy et al., 2005) and the Signal P 3.0 server (Bendtsen et al., 2004) were used to predict the subcellular localization and detect potential signal peptides, respectively. PSIPRED v2.6 was used to predict the secondary structure of the proteins (Jones, 1999). InterProScan was used to identify protein motifs (Zdobnov and Apweiler, 2001). The PRED-TMBB server was used to predict the porin domain (Bagos et al., 2004). The full-length sequences of BcsC E. coli and BpsA were submitted to the PRED-TMBB server; however, truncated versions of PgaA (residues 421–807), HmsH (residues 500–822), BcsC S. typh (776–1143), and PelB (880–1167) were needed to obtain successful b-barrel predictions. For E. coli BcsC, the porin domain predicted by PRED-TMBB spanned residues 703 to 1140, which overlapped with TPR domain predicted by InterProScan (40–771). Secondary structure predictions suggest that region 703–771 folds into helices and is therefore more likely to be part of the TPR domain than the porin domain. ACCESSION NUMBERS The coordinates and structure factors have been deposited in the Protein Data Bank under accession number 3E4B. SUPPLEMENTAL INFORMATION Supplemental Information includes three tables, five figures, Supplemental Results, and Supplemental Experimental Procedures and can be found with this article online at doi:10.1016/j.str.2009.11.015. ACKNOWLEDGMENTS The authors thank G.D. Smith, S.-Y. Ku, and J. Marsh for access to various computer programs; the Advanced Protein Technology Centre at the Hospital for Sick Children for assistance with DNA sequencing; and R. Zhoa and W. Houry at the University of Toronto for the analytical ultracentrifiguation analysis. This work is supported by grants from the Canadian Institutes of Health Research (CIHR; MT13337) to P.L.H. and from the Public Health Service

Structure 18, 265–273, February 10, 2010 ª2010 Elsevier Ltd All rights reserved 271

Structure Structure of AlgK

(AI-19146), the National Institute of Allergy and Infectious Disease, the Cystic Fibrosis (CF) Foundation, and Veterans Administrations Medical Research Funds to D.E.O. P.L.H., L.L.B., and M.N. are recipients of a Canada Research Chair, a CIHR New Investigator award, and a CIHR postdoctoral fellowship, respectively. C.-L.K. was funded, in part, by graduate scholarships from the Natural Sciences and Engineering Research Council, the Canadian CF Foundation, the Ontario Graduate Scholarship Program, Ontario Student Opportunities Trust Fund, and the Hospital for Sick Children Foundation Student Scholarship Program. Beam lines X25 and X29 at the National Synchrotron Light Source are supported by the US Department of Energy. Received: October 21, 2009 Revised: November 19, 2009 Accepted: November 24, 2009 Published: February 9, 2010 REFERENCES Aarons, S.J., Sutherland, I.W., Chakrabarty, A.M., and Gallagher, M.P. (1997). A novel gene, algK, from the alginate biosynthesis cluster of Pseudomonas aeruginosa. Microbiology 143, 641–652. Bagos, P.G., Liakopoulos, T.D., Spyropoulos, I.C., and Hamodrakas, S.J. (2004). PRED-TMBB: a web server for predicting the topology of beta-barrel outer membrane proteins. Nucleic Acids Res. 32, W400–W404. Bayan, N., Guilvout, I., and Pugsley, A.P. (2006). Secretins take shape. Mol. Microbiol. 60, 1–4. Bendtsen, J.D., Nielsen, H., von Heijne, G., and Brunak, S. (2004). Improved prediction of signal peptides: SignalP 3.0. J. Mol. Biol. 340, 783–795. Branda, S.S., Vik, S., Friedman, L., and Kolter, R. (2005). Biofilms: the matrix revisited. Trends Microbiol. 13, 20–26. Brunger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., GrosseKunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M., Pannu, N.S., et al. (1998). Crystallography & NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr. D Biol. Crystallogr. 54, 905–921. Danese, P.N., Pratt, L.A., and Kolter, R. (2000). Exopolysaccharide production is required for development of Escherichia coli K-12 biofilm architecture. J. Bacteriol. 182, 3593–3596. Darby, C. (2008). Uniquely insidious: Yersinia pestis biofilms. Trends Microbiol. 16, 158–164. Davis, I.W., Leaver-Fay, A., Chen, V.B., Block, J.N., Kapral, G.J., Wang, X., Murray, L.W., Arendall, W.B., 3rd, Snoeyink, J., Richardson, J.S., and Richardson, D.C. (2007). MolProbity: all-atom contacts and structure validation for proteins and nucleic acids. Nucleic Acids Res. 35, W375–W383. Emsley, P., and Cowtan, K. (2004). Coot: model-building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132. Eswaran, J., Koronakis, E., Higgins, M.K., Hughes, C., and Koronakis, V. (2004). Three’s company: component structures bring a closer view of tripartite drug efflux pumps. Curr. Opin. Struct. Biol. 14, 741–747. Friedman, L., and Kolter, R. (2004). Two genetic loci produce distinct carbohydrate-rich structural components of the Pseudomonas aeruginosa biofilm matrix. J. Bacteriol. 186, 4457–4465. Gardy, J.L., Laird, M.R., Chen, F., Rey, S., Walsh, C.J., Ester, M., and Brinkman, F.S. (2005). PSORTb v.2.0: expanded prediction of bacterial protein subcellular localization and insights gained from comparative proteome analysis. Bioinformatics 21, 617–623. Gassner, N.C., and Matthews, B.W. (1999). Use of differentially substituted selenomethionine proteins in X-ray structure determination. Acta Crystallogr. D Biol. Crystallogr. 55, 1967–1970. Gough, J., Karplus, K., Hughey, R., and Chothia, C. (2001). Assignment of homology to genome sequences using a library of hidden Markov models that represent all proteins of known structure. J. Mol. Biol. 313, 903–919. Govan, J.R., and Deretic, V. (1996). Microbial pathogenesis in cystic fibrosis: mucoid Pseudomonas aeruginosa and Burkholderia cepacia. Microbiol. Rev. 60, 539–574.

Hall-Stoodley, L., Costerton, J.W., and Stoodley, P. (2004). Bacterial biofilms: from the natural environment to infectious diseases. Nat. Rev. Microbiol. 2, 95–108. Hancock, R.E., and Nikaido, H. (1978). Outer membranes of gram-negative bacteria. XIX. Isolation from Pseudomonas aeruginosa PAO1 and use in reconstitution and definition of the permeability barrier. J. Bacteriol. 136, 381–390. Holm, L., and Sander, C. (1993). Protein structure comparison by alignment of distance matrices. J. Mol. Biol. 233, 123–138. Itoh, Y., Rice, J.D., Goller, C., Pannuri, A., Taylor, J., Meisner, J., Beveridge, T.J., Preston, J.F., 3rd, and Romeo, T. (2008). Roles of pgaABCD genes in synthesis, modification, and export of the Escherichia coli biofilm adhesin, poly-{beta}-1,6-N-acetyl-D-glucosamine (PGA). J. Bacteriol. 190, 3670–3680. Jain, S., Franklin, M.J., Ertesvag, H., Valla, S., and Ohman, D.E. (2003). The dual roles of AlgG in C-5-epimerization and secretion of alginate polymers in Pseudomonas aeruginosa. Mol. Microbiol. 47, 1123–1133. Jain, S., and Ohman, D.E. (1998). Deletion of algK in mucoid Pseudomonas aeruginosa blocks alginate polymer formation and results in uronic acid secretion. J. Bacteriol. 180, 634–641. Jain, S., and Ohman, D.E. (2004). Alginate biosynthesis. In Pseudomonas, J.-L. Ramos, ed. (New York: Kluwer Academic/Plenum Publishers), pp. 53–81. Jain, S., and Ohman, D.E. (2005). Role of an alginate lyase for alginate transport in mucoid Pseudomonas aeruginosa. Infect. Immun. 73, 6429–6436. Jinek, M., Rehwinkel, J., Lazarus, B.D., Izaurralde, E., Hanover, J.A., and Conti, E. (2004). The superhelical TPR-repeat domain of O-linked GlcNAc transferase exhibits structural similarities to importin alpha. Nat. Struct. Mol. Biol. 11, 1001–1007. Jones, D.T. (1999). Protein secondary structure prediction based on positionspecific scoring matrices. J. Mol. Biol. 292, 195–202. Keiski, C.L., Yip, P., Robinson, H., Burrows, L.L., and Howell, P.L. (2007). Expression, purification, crystallization and preliminary X-ray analysis of Pseudomonas fluorescens AlgK. Acta Crystallogr. Sect. F Struct. Biol. Cryst. Commun. 63, 415–418. Kim, K., Oh, J., Han, D., Kim, E.E., Lee, B., and Kim, Y. (2006). Crystal structure of PilF: functional implication in the type 4 pilus biogenesis in Pseudomonas aeruginosa. Biochem. Biophys. Res. Commun. 340, 1028–1038. Kleywegt, G.J. (2007). Crystallographic refinement of ligand complexes. Acta Crystallogr. D Biol. Crystallogr. 63, 94–100. Kleywegt, G.J., and Jones, T.A. (1996). Phi/psi-chology: Ramachandran revisited. Structure 4, 1395–1400. Knowles, T.J., Scott-Tucker, A., Overduin, M., and Henderson, I.R. (2009). Membrane protein architects: the role of the BAM complex in outer membrane protein assembly. Nat. Rev. Microbiol. 7, 206–214. Koo, J., Tammam, S., Ku, S.Y., Sampaleanu, L.M., Burrows, L.L., and Howell, P.L. (2008). PilF is an outer membrane lipoprotein required for multimerization and localization of the Pseudomonas aeruginosa Type IV pilus secretin. J. Bacteriol. 190, 6961–6969. Krissinel, E., and Henrick, K. (2007). Inference of macromolecular assemblies from crystalline state. J. Mol. Biol. 372, 774–797. Laskowski, R.A., MacArthur, M.W., Moss, D.S., and Thornton, J.M. (1993). PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Crystallogr. 26, 283–291. Luthy, L., Grutter, M.G., and Mittl, P.R. (2004). The crystal structure of Helicobacter cysteine-rich protein C at 2.0 A resolution: similar peptide-binding sites in TPR and SEL1-like repeat proteins. J. Mol. Biol. 340, 829–841. Magliery, T.J., and Regan, L. (2005). Sequence variation in ligand binding sites in proteins. BMC Bioinformatics 6, 240. Marcotte, E.M., Pellegrini, M., Thompson, M.J., Yeates, T.O., and Eisenberg, D. (1999). A combined algorithm for genome-wide prediction of protein function. Nature 402, 83–86. Merighi, M., Lee, V.T., Hyodo, M., Hayakawa, Y., and Lory, S. (2007). The second messenger bis-(30 -50 )-cyclic-GMP and its PilZ domain-containing receptor Alg44 are required for alginate biosynthesis in Pseudomonas aeruginosa. Mol. Microbiol. 65, 876–895.

272 Structure 18, 265–273, February 10, 2010 ª2010 Elsevier Ltd All rights reserved

Structure Structure of AlgK

Nesper, J., Hill, C.M., Paiment, A., Harauz, G., Beis, K., Naismith, J.H., and Whitfield, C. (2003). Translocation of group 1 capsular polysaccharide in Escherichia coli serotype K30. Structural and functional analysis of the outer membrane lipoprotein Wza. J. Biol. Chem. 278, 49763–49772. Notredame, C., Higgins, D.G., and Heringa, J. (2000). T-Coffee: a novel method for fast and accurate multiple sequence alignment. J. Mol. Biol. 302, 205–217. O’Gara, J.P. (2007). ica and beyond: biofilm mechanisms and regulation in Staphylococcus epidermidis and Staphylococcus aureus. FEMS Microbiol. Lett. 270, 179–188. O’Toole, G., Kaplan, H.B., and Kolter, R. (2000). Biofilm formation as microbial development. Annu. Rev. Microbiol. 54, 49–79. Ogelsby, L.L., Jain, S., and Ohman, D.E. (2008). Membrane topology and roles of Pseudomonas aeruginosa Alg8 and Alg44 in alginate polymerization. Microbiology 154, 1605–1615. Otwinowski, Z., and Minor, W. (1997). Processing of X-ray diffraction data collection in oscillation mode. Methods Enzymol. 276, 307–326. Pape, T., and Schneider, T.R. (2004). HKL2MAP: a graphical user interface for phasing with SHELX programs. J. Appl. Crystallogr. 37, 843–844. Quinaud, M., Ple, S., Job, V., Contreras-Martel, C., Simorre, J.P., Attree, I., and Dessen, A. (2007). Structure of the heterotrimeric complex that regulates type III secretion needle formation. Proc. Natl. Acad. Sci. USA 104, 7803–7808. Rehm, B.H., Boheim, G., Tommassen, J., and Winkler, U.K. (1994). Overexpression of algE in Escherichia coli: subcellular localization, purification, and ion channel properties. J. Bacteriol. 176, 5639–5647. Remminghorst, U., and Rehm, B.H. (2006a). Alg44, a unique protein required for alginate biosynthesis in Pseudomonas aeruginosa. FEBS Lett. 580, 3883– 3888. Remminghorst, U., and Rehm, B.H. (2006b). In vitro alginate polymerization and the functional role of Alg8 in alginate production by Pseudomonas aeruginosa. Appl. Environ. Microbiol. 72, 298–305. Robles-Price, A., Wong, T.Y., Sletta, H., Valla, S., and Schiller, N.L. (2004). AlgX is a periplasmic protein required for alginate biosynthesis in Pseudomonas aeruginosa. J. Bacteriol. 186, 7369–7377. Romling, U. (2002). Molecular biology of cellulose production in bacteria. Res. Microbiol. 153, 205–212. Ryder, C., Byrd, M., and Wozniak, D.J. (2007). Role of polysaccharides in Pseudomonas aeruginosa biofilm development. Curr. Opin. Microbiol. 10, 644–648. Scheufler, C., Brinker, A., Bourenkov, G., Pegoraro, S., Moroder, L., Bartunik, H., Hartl, F.U., and Moarefi, I. (2000). Structure of TPR domain-peptide

complexes: critical elements in the assembly of the Hsp70-Hsp90 multichaperone machine. Cell 101, 199–210. Sun, P., Tropea, J.E., Austin, B.P., Cherry, S., and Waugh, D.S. (2008). Structural characterization of the Yersinia pestis type III secretion system needle protein YscF in complex with its heterodimeric chaperone YscE/YscG. J. Mol. Biol. 377, 819–830. Tanaka, K., Matsuyama, S.I., and Tokuda, H. (2001). Deletion of lolB, encoding an outer membrane lipoprotein, is lethal for Escherichia coli and causes accumulation of lipoprotein localization intermediates in the periplasm. J. Bacteriol. 183, 6538–6542. Terwilliger, T.C. (2003). Automated main-chain model building by template matching and iterative fragment extension. Acta Crystallogr. D Biol. Crystallogr. 59, 38–44. Terwilliger, T.C., and Berendzen, J. (1999). Automated MAD and MIR structure solution. Acta Crystallogr. D Biol. Crystallogr. 55, 849–861. Wang, X., Preston, J.F., 3rd, and Romeo, T. (2004). The pgaABCD locus of Escherichia coli promotes the synthesis of a polysaccharide adhesin required for biofilm formation. J. Bacteriol. 186, 2724–2734. Westbrook, J., Feng, Z., Burkhardt, K., and Berman, H.M. (2003). Validation of protein structures for protein data bank. Methods Enzymol. 374, 370–385. Whitfield, C. (2006). Biosynthesis and assembly of capsular polysaccharides in Escherichia coli. Annu. Rev. Biochem. 75, 39–68. Whitfield, C., and Naismith, J.H. (2008). Periplasmic export machines for outer membrane assembly. Curr. Opin. Struct. Biol. 18, 466–474. Whitney, J.C., Neculai, A.M., Ohman, D.E., and Howell, P.L. (2009). Expression, refolding, crystallization and preliminary X-ray analysis of Pseudomonas aeruginosa AlgE. Acta Crystallogr. Sect. F Struct. Biol. Cryst. Commun. 65, 463–466. Wong, H.C., Fear, A.L., Calhoon, R.D., Eichinger, G.H., Mayer, R., Amikam, D., Benziman, M., Gelfand, D.H., Meade, J.H., Emerick, A.W., et al. (1990). Genetic organization of the cellulose synthase operon in Acetobacter xylinum. Proc. Natl. Acad. Sci. USA 87, 8130–8134. Wu, T., Malinverni, J., Ruiz, N., Kim, S., Silhavy, T.J., and Kahne, D. (2005). Identification of a multicomponent complex required for outer membrane biogenesis in Escherichia coli. Cell 121, 235–245. Zdobnov, E.M., and Apweiler, R. (2001). InterProScan—an integration platform for the signature-recognition methods in InterPro. Bioinformatics 17, 847–848. Zogaj, X., Nimtz, M., Rohde, M., Bokranz, W., and Romling, U. (2001). The multicellular morphotypes of Salmonella typhimurium and Escherichia coli produce cellulose as the second component of the extracellular matrix. Mol. Microbiol. 39, 1452–1463.

Structure 18, 265–273, February 10, 2010 ª2010 Elsevier Ltd All rights reserved 273