The Waddlia Genome: A Window into Chlamydial Biology

Growing evidence suggests that a novel member of the Chlamydiales order, Waddlia chondrophila, is a potential agent of miscarriage in humans and abortion in ruminants. Due to the lack of genetic tools to manipulate chlamydia, genomic analysis is proving to be the most incisive tool in stimulating investigations into the biology of these obligate intracellular bacteria. 454/Roche and Solexa/Illumina technologies were thus used to sequence and assemble de novo the full genome of the first representative of the Waddliaceae family, W. chondrophila. The bacteria possesses a 2′116′312bp chromosome and a 15′593 bp low-copy number plasmid that might integrate into the bacterial chromosome. The Waddlia genome displays numerous repeated sequences indicating different genome dynamics from classical chlamydia which almost completely lack repetitive elements. Moreover, W. chondrophila exhibits many virulence factors also present in classical chlamydia, including a functional type III secretion system, but also a large complement of specific factors for resistance to host or environmental stresses. Large families of outer membrane proteins were identified indicating that these highly immunogenic proteins are not Chlamydiaceae specific and might have been present in their last common ancestor. Enhanced metabolic capability for the synthesis of nucleotides, amino acids, lipids and other co-factors suggests that the common ancestor of the modern Chlamydiales may have been less dependent on their eukaryotic host. The fine-detailed analysis of biosynthetic pathways brings us closer to possibly developing a synthetic medium to grow W. chondrophila, a critical step in the development of genetic tools. As a whole, the availability of the W. chondrophila genome opens new possibilities in Chlamydiales research, providing new insights into the evolution of members of the order Chlamydiales and the biology of the Waddliaceae.


Introduction
The Chlamydiaceae were long considered a phylogenetically isolated group of closely-related bacteria. However, during the past decades, the order Chlamydiales has been enriched by the discovery of five additional families: Criblamydiaceae, Parachlamydiaceae, Rhabdochlamydiaceae, Simkaniaceae, and Waddliaceae [1,2,3,4]. Members of these families are commonly called Chlamydia-related bacteria due to their phylogenetic relationship to Chlamydiaceae and their Chlamydia-like cycle of replication with two developmental stages; the elementary body (EB) or infectious particle, and the reticulate body (RB) or replicative form.
Globally, members of the Chlamydiaceae family are recognized as a widespread and clinically significant cause of disease in humans and animals. It comes as no surprise that evidence is emerging to support a pathogenic role for Chlamydia-related bacteria as well [5].
Simkania negevensis and Parachlamydia acanthamoebae are suspected to cause respiratory tract infections [6,7]. Moreover, P. acanthamoebae was recently shown to be associated with abortion in cows [8].
Waddliaceae might also cause bovine abortion since two strains of W. chondrophila have been isolated independently from aborted bovine fetuses [9,10] and a serological study in cows supported an abortigenic role [11]. More recently, a prospective study demonstrated an association between miscarriage and the presence of anti-Waddlia antibodies in humans [12]. Moreover, DNA of W. chondrophila was detected in one respiratory sample of a patient with community-acquired pneumonia [13] and in samples taken from children with bronchiolitis [14], suggesting that, as previously described for Coxiella burnetii, this emerging agent of miscarriage may also cause respiratory tract infections. The pathogenic potential of W. chondrophila is further indicated by its rapid growth within human macrophages [15].
The advent of genomics was a fundamental step in the characterization of obligate intracellular bacteria such as the Chlamydiaceae, which are widely recognized pathogens. Since 1998, the release of fourteen complete genome sequences from members of the Chlamydiaceae, ranging in size from 1Mb to 1.2Mb, provided major advances in the study of their biology and the identification of virulence factors [16]. The genome analysis revealed a high overall similarity in gene content and gene order between the various Chlamydiaceae, although regions of gene rearrangement, referred to as the ''plasticity zone'', can be found near the terminus of replication [17]. Incomplete pathways for tricarboxylic acid cycle or biosynthesis of key amino acids, nucleotides or cofactors revealed a strong dependency on host-derived metabolites [16]. Insights into bacterial virulence were provided by the discovery of a complete type three secretion system (T3SS) [18] and conserved virulence factors such as CPAF [19]. Furthermore, periplasmic and outer membrane proteins unique to the Chlamydiaceae were discovered and compose a highly disulphide crosslinked matrix that supply the structural resilience usually provided by the peptidoglycan layer in most Gram-negative bacteria [20]. These include the abundant cysteine-rich proteins OmcA and OmcB as well as the major components of the chlamydial outer membrane complex, the beta-barrel porins OmpA and PorB [16]. Moreover, a chlamydial specific family of autotransporters, the highly diverse polymorphic outer membrane proteins (pmps), has been implicated in adhesion and in the host immune response [21,22]. Most of these cell wall proteins are highly immunogenic and are used for serological diagnosis or vaccine development [23,24].
Among the Chamydia-related bacteria, sequences from only two strains belonging both to the Parachlamydiaceae family have been released to date. P. acanthamoebae Hall's coccus was published as an unfinished genome of 3Mb in a combined proteomics and genomics approach [25] whereas the environmental Protochlamydia amoebophila UWE25 was fully sequenced and exhibited a 2.4Mb genome [26], i.e. approximately twice the size of classical chlamydia. P. amoebophila showed limited conservation of genome structure together with the presence of several repetitive elements [26]. Moreover, its chromosome contains a 100 kb-long genomic island encoding a potentially functional F-like DNA conjugative system [27] and more than 70 leucine-rich repeat proteins [28]. Although the bacterium displayed improved biosynthetic abilities as compared to Chlamydia species, a similar dependency on host derived metabolites was observed [26]. Intriguingly, no homologs to major outer membrane proteins and polymorphic membrane proteins have been identified, suggesting these highly immunogenic proteins might be Chlamydiaceae family specific.
The availability of genome sequences within new families of the Chlamydiales order is crucial to better understand the biology of Chlamydia-related bacteria. Therefore, we sequenced the full genome of W. chondrophila using both 454/Roche and Solexa/ Illumina technologies, also uncovering the presence of a low copy number plasmid. The genome annotation revealed numerous intriguing features presented here which we anticipate will help stimulate and drive further research into this fascinating and medically important bacterial order.

General genome features
The genome of Waddlia chondrophila WSU 86-1044 consists of a circular chromosome of 291169312 bp with a G+C content of 43.8% and a 159593 bp circular plasmid with a G+C content of 37.6% ( Figure 1 and Table 1). The chromosome sequence displays a typical ''V''-inverted shape on a cumulative GC skew plot ( Figure S1), allowing the origin and terminus of replication to be located and the assembly accuracy to be confirmed [29,30]. Another commonly used marker for the origin of replication is dnaA [31]. However, like other Chlamydiales, W. chondrophila encodes two copies of dnaA, none of which is linked to the minimum of the cumulative GC skew ( Figure S1). Using bioinformatics analyses, a large number of repetitive sequences (.200 bp) were identified encompassing 4.9% of the chromosome, a significantly higher proportion than other sequenced Chlamydiales ( Figure 1 and Table 1).
Two sets of rRNAs and 37 tRNA genes were identified as well as 1934 protein coding genes, which represent 92% of the whole genome. A putative function or family membership could be inferred for 1243 (65%) of them, whereas 253 (13%) are conserved hypothetical proteins and the remaining 438 (23%) show no similarity to known proteins (Table S1 and Figure S2). Of the conserved hypothetical proteins, the major group of 156 proteins is most similar to hypothetical proteins from the Parachlamydiaceae family. The remainder show best BLAST hits against Eukaryotes (2), Archaea (4), Chlamydiaceae (4) and various other bacterial phyla (87). As expected, all essential components for DNA replication, transcription and RNA translation were successfully identified and mostly belong to the core set of Chlamydiales genes. The comparison of W. chondrophila encoded proteins with those of C. trachomatis and P. amoebophila showed the large proportion of family-specific proteins and proteins poorly conserved at the amino acid level within this highly diverse bacterial order ( Figure  S3). The X-plot representation of conserved genes highlighted small collinear regions between Waddlia and Protochlamydia, even though numerous rearrangements occurred, changing gene orientation and position. In contrast, conserved gene order is hardly distinguishable between Waddlia and Chlamydia.

Waddlia chondrophila plasmid
The comparison between the average read depth of the W. chondrophila chromosome (40x) and its plasmid (440x) indicates that the plasmid is present in about 11 copies per cell. It encodes 22 proteins that mostly show no homology to other chlamydial plasmid proteins, with the exception of an integrase that exhibits 54% identity to the plasmid integrase pCpA1_003 of C. psittaci. The Waddlia chromosome contains numerous small regions (16-24 bp) identical to sequences in the plasmid. In addition, 7 chromosomal regions ranging from 57 bp to 849bp that encode entire or partial integrated transposases share between 99% and 100% identity with the two plasmid transposases ( Figure 1). Each of these two transposases is strongly similar to several proteins encoded on the P. amoebophila genome. Finally, two adjacent genes were found to be integrated into the Waddlia chromosome sharing 88% nucleic acid identity to their plasmid counterparts. One of the encoded proteins is homologous to MazF, an endoribonuclease of the MazEF module, one of the most thoroughly studied toxin-antitoxin systems [32], that might be involved in the stable maintenance of the plasmid during cell division. Normally, mazE encodes a labile antitoxin that prevents the lethal effect of the stable toxin encoded by mazF. Although the neighbouring genes of mazF do not show any sequence similarity to mazE, one of them may represent the necessary antitoxin.
Virulence factors and resistance to environmental stresses As frequently described for intracellular bacteria, W. chondrophila encodes a T3SS devoted to inoculation of bacterial effectors into the host cytoplasm. Clusters of genes encoding the T3SS are spread over the bacterial chromosome as previously shown for other Chlamydiales ( Figure S4) [18,33]. The functionality and the requirement of the T3SS for Waddlia survival and replication in human macrophages were demonstrated by the effective inhibition of bacterial growth with T3SS specific inhibitors [34] (Figure 2). Interestingly, several genes encoding homologs to SycE and SycD chaperones are located adjacent to genes encoding hypothetical proteins that could represent putative T3SS effectors ( Figure S4).
Several systems specific to Waddlia are likely involved in resistance to professional phagocytes such as amoebae and/or macrophages where the bacterium is able to escape lysosomal degradation and to grow rapidly [15,35]. The mrp and trk systems as well as a putative carbonate permease and a carbonic anhydrase are probably involved in pH homeostasis by importing protons. Furthermore, homologs to superoxide dismutase and catalase as well as a nitric oxide reductase are likely involved in defense against radical oxygen species and nitrous oxides.
To resist environmental stresses, W. chondrophila possesses at least four heavy metal exporters for copper, zinc, cadmium and a possible tellurite reductase/permease, that could be involved in defense against toxic metals. Moreover, the bacterium encodes several multidrug efflux pumps of the RND family that may confer resistance to detergents, lipophilic drugs or bile salt derivatives.   A putative class-C b-lactamase might be responsible for the previously described in vitro resistance of W. chondrophila to ampicillin and ceftriaxone [36]. A putative peptidase S66 family protein and a putative undecaprenyl-diphosphatase 1 also indicate a possible resistance to microcin and bacitracin. Finally, several antibiotic resistance mechanisms were identified with the presence of proteins related to multidrug resistance MarC and MATE families suggesting that the bacteria could present a large pattern of resistance.

Host parasitism and bacterial metabolism
Waddlia genome analysis revealed a degree of host independence, compared with other members of the Chlamydiales order, with the ability to produce energy independently from its host through oxidative phosphorylation. Reduced cofactors issued from complete TCA cycle and glycolysis are funneled along the electron transport chain to produce ATP. As in P. amoebophila, the presence of a F O F 1 ATP synthase complex, in addition to a V 1 V 0 ATPase complex conserved in the Chlamydiaceae, enhances its energy production capacity and improves its adaptability in energydepleted environments. Furthermore, W. chondrophila contains the enzymatic components of the glyoxylate bypass enabling the utilization of fatty acids or acetate, in the form of acetyl-CoA, as a carbon source.
As summarized in Figure 3, the bacterium displays enhanced anabolic capabilities for key molecules such as cofactors, nucleotides or amino acids. No homolog to P. amoebophila NAD+/ADP transporter (ntt_4) [37] was found, but enhanced biosynthetic abilities indicate that W. chondrophila likely synthesizes NAD from an intermediary metabolite such as quinolinate or nicotinamide imported through another system. Five nucleotide transporters similar to ntt_1, 2 and 3 of P. amoebophila potentially enable the import of all nucleotides [37,38]. Despite the presence of genes for nucleotide parasitism and unlike other Chlamydiales, W. chondrophila possesses all enzymes to convert L-glutamine in UMP and all pyrimidine derivatives necessary for replication and transcription ( Figure S5). In contrast, a complete purine biosynthesis pathway could not be reconstructed but an active purine conversion, that is not present in other members of the Chlamydiales order, was identified ( Figure S5).
W. chondrophila harbors the genetic material to produce at least ten of the twenty classical amino acids (Table S2). This bacterium completely lacks genes for the biosynthesis of tryptophan that are at least partially present in other Chlamydia and seems unable to produce tyrosine and phenylalanine but, instead, encodes five transporters devoted to general or specific aromatic amino acid import. Furthermore, many oligo-peptides and amino acid transporters or permeases have been identified and can likely import a variety of amino acids from the environmental medium.
Lipid metabolism also exhibits interesting features with the presence of additional enzymes for glycerophospholipid, glycerolipid and sphingolipid metabolism compared to other Chlamy- diales. More interestingly, unlike P. amoebophila and C. trachomatis, W. chondrophila possesses a complete operon encoding the mevalonate pathway in the biosynthesis of isoprenoids precursors, whereas only one gene could be identified in the non-mevalonate pathway used by both P. amoebophila and C. trachomatis.

Bacterial cell wall
The chlamydial cell wall differs from that of the majority of extracellular Gram-negative bacteria, the classical protective peptidoglycan being replaced by a highly disulphide-linked proteinaceous layer in the infectious EB [39]. Upon entering the cell, the EB is released from the constraints imposed by its protective corsett by reducing the disulphide-linked network of proteins, allowing it to swell in size as the replicating body forms. This process must be tightly regulated and W. chondrophila possesses several conserved periplasmic chlamydial redox enzymes. Major components of the proteinaceous network are OmcA and OmcB, a highly diverse family of polymorphic outer membrane proteins (pmp) and the outer membrane protein (omp) beta-barrel porins OmpA and PorB. A striking feature of these porins are conserved cysteine rich clusters of CxCxC or CxxC or CC or CxxCxxC signature sequences, essential in the covalent cross-linking of the periplasmic Omc proteins and the outer membrane proteins [40,41].
Most impressive in W. chondrophila, is a novel OMP family of 11 putative beta-barrel proteins or porins with C-rich signatures, partially shared with the Chlamydiaceae (Figure 4). In addition, conserved motives and structural analyses revealed the presence of a putative autotransporter protein that shows similarity to a gene in the P. amoebophila genome ( Figure S6). These proteins might belong to the chlamydial pmps, a highly diverse family of autotransporters unique to the Chlamydia and ranging from 8 members in C. trachomatis to 21 members in C. pneumoniae. In addition to the expected omcA and omcB genes, we detected five adjacent genes sharing a similar N-terminus and conserved cysteine residues, which may form an extended omc family, both in W. chondrophila and P. amoebophila.
Finally, elements of the cell division and peptidoglycan pathways appear to be conserved with Chlamydiaceae suggesting that, like its close relatives, W. chondrophila uses a FtsZ-independent cell division mechanism (divisome) proposed to involve remnants of peptidoglycan and lipid II pathways [20,42] (Figure S7). Although the gene organization is different, after a detailed structural comparison, we propose that a gene previously annotated in the chlamydial genomes to date as a hypothetical protein might be the missing cell division gene ftsL. Moreover, W. chondrophila displays additional genes for osmoregulated periplasmic glycan synthesis and cell wall biosynthesis indicating that the cell wall probably differ from other Chlamydiales, which will affect host recognition, membrane structure and function.

Discussion
The full genome sequence of the first representative within the Waddliaceae family, Waddlia chondrophila, revealed numerous features that provide major insights not only into genome evolution of the order Chlamydiales but also into the pathogenic potential of this strict intracellular bacterium. The Waddlia genome exhibits numerous repeated sequences and transposases indicating a different genome dynamics from classical Chlamydia which almost completely lack repetitive elements. If these repetitions might partially account for Waddlia's larger genome size compared to classical Chlamydia, it is then even more interesting that P. amoebophila, which has an even larger genome, exhibits fewer repeated elements and reduced metabolic capacities compared to Waddlia. Among other features that could explain the increased size of P. amoebophila is the F-like DNA conjugative system encoded on a 100 kb-long genomic island [27]. The latter F-like operon is not present in W. chondrophila whereas it seems to be at least partially present in another member of the Parachlamydiaceae family, Parachlamydia acanthamoebae [25], suggesting that it might have been acquired after divergence of Waddliacaeae and Parachlamydiaceae families.
The W. chondrophila plasmid is significantly larger than those of classical Chlamydia and encodes few genes with known function. Thanks to its multiple copies, this genetic element might represent an interesting target for diagnostic PCRs of increased sensitivity. Nevertheless, the recent difficulties that appeared with a C. trachomatis Swedish variant, where the introduction of a diagnostic PCR targeting the plasmid was rapidly followed by a deletion in the targeted region [43], implies that a multi-target approach should be used. Interestingly, the putative toxin/anti-toxin system indicates that the plasmid might be stably present in the bacteria. Another crucial step for the development of a genetic system is the ability to grow the bacteria in a cell-free medium. W. chondrophila may represent the best candidate among Chlamydiales for axenization, i.e. growth in absence of eukaryotic cells. Indeed, the bacterium possesses all components necessary for the generation of energy and unexpectedly encodes extensive capabilities for de novo biosynthesis of essential components such as nucleotides, amino acids, lipids and cofactors. This also suggests that the common ancestor of Chlamydiales might have been less dependent upon its eukaryotic host. Different pressures have selected for maintenance or elimination of different metabolic pathways within the different families, compensated by the acquisition of transporters to import the metabolites that cannot be synthesized.
The predicted increased host independence of W. chondrophila may partially explain its ability to grow efficiently within a broad host range which includes at least amoebae [44], McCoy cells [10], bovine turbinate cells, P388D1 mouse macrophages [45] and monocyte-derived human macrophages [15]. Moreover, the presence of various mechanisms for pH regulation and defense against radical oxygen species represent key systems for the initial resistance to destruction by professional phagocytes. In macrophages, survival is then achieved by evading the early phagolysosome into a replicative vacuole that forms an intimate association with mitochondria and the endoplasmic reticulum [35]. To modulate and interfere with host functions, the secretion of effectors through a functional type III secretion system is likely essential since a specific inhibition of the T3SS efficiently prevents the growth of Waddlia in PBMC-derived human macrophages. The availability of the full genome sequence facilitates the identification of new effectors and virulence factors that will enable us to better understand the mechanisms used by this bacterium to divert the host cell.
The chlamydial outer membrane is unique amongst bacteria and here we demonstrate that not only does Waddlia possess OmcA and OmcB, but there is an extended family of OmpArelated proteins which are rich in conserved cysteine clusters and predicted to form outer membrane beta-barrels or porins. Added to this is a predicted outer membrane autotransporter, a putative member of the chlamydial pmp family, that present a homolog in P. amoebophila genome. The pmp family members were originally identified as highly immunogenic proteins of the chlamydial cell wall [24] and later shown to be chlamydial adhesins [21,22]. Further investigations are now needed to confirm the potential roles as adhesins for the putative autotransporters identified. The OmpA family, on the other hand, appears to be confined to the Chlamydiaceae and Waddliaceae, indicating that its potential role as an adhesin is selective for a common host and tissue range. In this context, it is tempting to speculate that the unique diversity of the OmpA family in the Waddliaceae could reflect a wider host specificity as well as a mechanism to avoid immune surveillance, assuming here the role played by the similarly diverse pmps in the Chlamydiaceae.
The uniqueness of the Waddlia cell wall, combined with what appears to be a family specific complement of proteins well established within the Chlamydiaceae to be both highly immunogenic and surface exposed in both the EB and RB forms, raises the exciting prospect that we have identified here prime candidates for serological testing and vaccine development. We can expect that the discovery and isolation of more species within the Waddliaceae family will add additional layers of diversity, emphasizing the importance of intensifying investigations into these pathogens.
As a whole, the complete genome sequence of W. chondrophila provides new insights into the evolution of the order Chlamydiales. The release of further new sequences is now needed to better understand the genetic composition and the genome dynamics of Chlamydia-related bacteria. We highlighted the unexpected high biosynthetic capabilities for essential compounds, the presence of several functional virulence factors and discovered large families of outer membrane proteins that might represent good candidates for vaccine development. The availability of the W. chondrophila genome will stimulate further research into this medically important bacterial phylum and provides the basis to develop new diagnostic tools which will assist in understanding the pathogenic potential of this bacterium considered as an emerging agent of miscarriage and respiratory tract infections.

Culture and purification of W. chondrophila
Waddlia chondrophila WSU 86-1044, ATCC number VR-1470, was grown at 32uC within Acanthamoeba castellanii ATCC 30010 in 75 cm 2 cell culture flasks (Becton Dickinson, Franklin Lakes, USA) with 30 ml of peptone-yeast extract glucose broth. To purify W. chondrophila, amoebae were removed from culture media using a first centrifugation step at 1206g for 10 min. Amoebal debris were next removed from the resuspended bacterial pellet by centrifugation at 6500 x g for 30 min onto 25% sucrose (Sigma Aldrich, St Louis, USA) and then at 32000 x g for 70 min onto a discontinuous Gastrographin (Bayer Schering Pharma, Zurich, Switzerland) gradient (48%/36%/28%). The bacteria clustering in the Gastrographin gradient at a large lower band were collected, centrifuged at 5800 x g and resuspended in PBS twice, and finally stocked at 280uC. The absence of contaminants was confirmed by plating frozen material on Chocolate agar. Since no growth was observed on agar after 72 h of incubation, frozen material was inoculated onto A. castellanii and immunofluorescence was performed using specific anti-Waddlia antibodies as well as DAPI-staining. We observed no DAPI-positive particles that were not stained with the anti-Waddlia antibodies. In addition, a PCR targeting Eubacteria 16S rRNA followed by sequencing was performed with primers FD1 (59agagtttgatcctggctcag39) and RP2 (59acggctaccttgttac-gactt39).

Genome sequencing, assembly and gap closure
Isolation of genomic DNA from W. chondrophila was performed with the Wizard Genomic DNA purification kit (Promega Corporation, Madison, USA). DNA was sequenced using both 454/Roche [46] and Solexa/Illumina [47] technologies, respectively on the Genome Sequencer FLX by Roche Applied Science (Penzberg, Germany) and the Genome Analyzer GAII by Fasteris (Plan les Ouates, Switzerland). GS FLX reads were assembled using Newbler V1.1.02.15 and the 90 large contigs obtained with 40x coverage served as the basis for the gap closure. To scaffold the contigs, a fosmid library with 40 kb DNA inserts was build in the vector pEpiFOS (Epicentre Biotechnologies, Madison, USA) by IIT-Biotech (Bielefeld, Germany). Fosmid walking as well as PCR-based techniques were used to close the gaps. Solexa reads were then mapped to the final assembly with Phrap and visualized with Consed [48]. Thus, 22 homopolymer errors were corrected after manual inspection of discrepancies.

Genome annotation
Curation and annotation of the genome was performed using the genome annotation system GenDB 2.4 [49]. Prediction of coding sequences (CDS) was accomplished using Critica [50], Glimmer [51] and Reganor [52]. All predicted ORFs were automatically submitted to similarity searches against nr, Swissprot, KEGG, InterPro, Pfam and TIGRfam databases. Putative signal peptides, transmembrane helices and nucleic acid binding domains were predicted using SignalP [53], TMHMM [54] and Helix-Turn-Helix [55], respectively. The automatic annotation of each CDS was manually checked and corrected according to the most congruent tool results. The complete annotated genome sequences have been deposited at GenBank under the accession numbers CP001928 and CP001929.

Genome analysis
The circular genome plot was created with DNA plotter [56]. Repeats were identified using REPuter [57]. W. chondrophila gene content was compared to P. amoebophila UWE25 (NC_005861) and C. trachomatis D/UW-3/CX (NC_000117) with EDGAR [58], which defines orthologous proteins based on bidirectional best blast hit and then calculates BLASTP score ratio values (SRV). Paralogous genes might be discarded during the analysis. For each comparison, SRV distribution was fitted with binormal or bibeta distribution with a self written R script, and a cutoff was determined at the point where the probability to belong to one or the other peak is equal. Accordingly, a general cutoff of 0.21 was used to retrieve the core genes and singletons.

Outer membrane proteins analysis
Genes identified as coding for putative outer membrane proteins (omp) and polymorphic membrane proteins (pmp) were aligned using Tcoffee [59] on EMBL-EBI interface (http://www. ebi.ac.uk/Tools/t-coffee/index.html). For omps, a corresponding neighbor-joining tree was calculated using MEGA4 [60] with the following parameters: 1000 bootstrap, pairwise comparison, poisson distribution, gamma parameter equal to 1. The presence of membrane spanning beta-barrel and beta-helical domains were predicted using the programs Partifold [61] and Beta-Wrap Pro [62], respectively.

Supporting Information
Methods S1 Supplementary methods.   Author Contributions