A bioinformatic strategy for the detection, classification and analysis of bacterial autotransporters

Abstract

Autotransporters are secreted proteins that are assembled into the outer membrane of bacterial cells. The passenger domains of autotransporters are crucial for bacterial pathogenesis, with some remaining attached to the bacterial surface while others are released by proteolysis. An enigma remains as to whether autotransporters should be considered a class of secretion system, or simply a class of substrate with peculiar requirements for their secretion. We sought to establish a sensitive search protocol that could identify and characterize diverse autotransporters from bacterial genome sequence data. The new sequence analysis pipeline identified more than 1500 autotransporter sequences from diverse bacteria, including numerous species of Chlamydiales and Fusobacteria as well as all classes of Proteobacteria. Interrogation of the proteins revealed that there are numerous classes of passenger domains beyond the known proteases, adhesins and esterases. In addition the barrel-domain-a characteristic feature of autotransporters-was found to be composed from seven conserved sequence segments that can be arranged in multiple ways in the tertiary structure of the assembled autotransporter. One of these conserved motifs overlays the targeting information required for autotransporters to reach the outer membrane. Another conserved and diagnostic motif maps to the linker region between the passenger domain and barrel-domain, indicating it as an important feature in the assembly of autotransporters.

Figure 1. Size distribution of predicted autotransporters.

A HMM built to describe the known 47 autotransporters was used to screen bacterial genomes and identified a total of 373 putative autotransporters (using an E-value of 10⁻⁵ as a cut-off score). These stage 1 results were plotted to the size of the protein sequence (blue bars). A second HMM based on features of the barrel-domains was then used to screen sequences that fell in the range (10⁻⁵ to 10⁻²) that might be false-negatives in the primary search, and these sequences represented with pink bars. Characterized autotransporters are indicated according to their size either in blue (stringent search criteria) or pink (relaxed search criteria). Details of the largest autotransporter, BigE, are shown in Figure S1. INSET: Representation of the small autotransporters from three species of Bordetella: NP_883896.1 [Bordetella parapertussis 12822, TynE, 491 residues], NP_889647.1 [Bordetella bronchiseptica RB50, 528 residues] and NP_879974.1 [Bordetella pertussis Tohama I, TcfA-tracheal colonization factor precursor, 647 residues]. Representative sections of multiple sequence alignment from the passenger domain (colored) and barrel-domain (grey) are shown. The numbering refers to the residues from TcfA, and proline residues (which are prevalent in the first two autotransporters) are colored yellow. The grey histogram plots represent sequence identity (as determined by PsiPred) across all three sequences.

Figure 2. The twin HMM search strategy for autotransporter detection.

Panels (A) and (B) depict the building of HMMs: both AT47-HMM and AT47-bb-HMM were initiated from the manually curated list of 47 autotransporters reported in the literature. (A) AT47-HMM was built from full-length autotransporter sequences. (B) AT47-bb-HMM was built from barrel-domains predicted using the secondary structure prediction tool DomPred. (C) Workflow of HMM analysis is shown wherein 1,210 bacterial genomes, comprising a total of 4,264,032 protein sequences, were filtered with AT47-HMM to produce two lists at E-value cutoffs of 10⁻² and 10⁻⁵. In the second step, 38,786 sequences collected with the E-value cutoff of 10⁻² were filtered with AT47-bb-HMM to produce 1,523 AT candidate sequences at the E-value cutoff of 10⁻⁴. Sequence lists (or sequence databases) are shown as simple rectangles, HMMs are depicted as ovals, simple arrows represent filtering of sequence lists, while thick arrows represent the process of building a HMM from a sequence list.

Figure 3. Phylogenetic analysis identifies major families of functional domains.

Sequences corresponding to the passenger domain plus α−linker-domain of autotransporters were defined and subject to phylogenetic analysis (Methods). Sub-domain signatures were identified using Pfam analysis of all sequences, and these are represented as radiating coloured symbols. The length of this line is proportional to the number of residues in the passenger-α-linker-domains. Major functional categories are shown (light blue arcs) based on the conservation of Pfam signatures and phylogenetic clustering. Non-proteobacterial sequences are coloured green (Chlamydiales) and yellow (Fusobacteria). Figure S2 provides the tree in a form where the accession numbers of every sequence can be viewed.

Figure 4. Conserved sequence motifs are not spatially conserved in tertiary structure.

(A) MEME was used to identify six common motifs in the barrel-domains of the 47 autotransporter sequences in the starting set (Methods). The six motifs were mapped onto the barrel-domain region of the five autotransporters for which crystal structures have been solved: BrkA, EspP, EstA, Hbp and NalP and are indicated by the respective colours. (B) In 1038 sequences, motif number 5 (the β-motif) was confined to the last β-strand and the final residue in the motif was always a bulky hydrophobic residue. In some of these sequences, a second occurrence, or even a third occurrence of the β-motif was detected in strands other than the terminal strand (paler yellow). In 473 sequences the β-motif is found internally within the barrel-domain. In 12 sequences a short “α-tail” segment succeeds the β-motif of the barrel. Given the topology of bacterial β-barrel proteins, the α-helical tail would presumably be positioned within the periplasmic space. (C) Sequences corresponding to the passenger-and α−linker-domains of autotransporters were defined (Methods) and subject to motif analysis using MEME. A single motif corresponding to the α−linker-domains of 31 of the 47 autotransporter sequences in the starting set (Methods), including EspP, EstA and Hbp was identified and is indicated by blue shading on the crystal structure representations of these proteins. Note, that in the crystallized form of BrkA the corresponding segment of the protein was not visualized.

Figure 5. Classification of autotransporter barrel-domains.

Sequences corresponding to the barrel-domain of autotransporters were defined and subject to phylogenetic analysis (Methods). Motif analysis of all sequences is represented as radiating coloured stripes, where the colours correspond to those used in Figure 4. The length of the line coloured by these stripes is proportional to the number of residues in the predicted barrel-domain. Fourteen types of barrel-domains were categorized based on conservation of motif placement and phylogenetic clustering. Each type is named for a recognized autotransporter found in that group, for example: Group 1 includes PspA and PspB, Group 4 includes EstA and BigE, Group 6 includes NalP. Group 11 sequences have in common that they have no obvious β-motif, and Group 13 has proteins with longer β-barrel sequences including AIDA-I and Ag43. Each sequence has also been categorized as having a “small”, “medium” or “large” passenger domain, with the purple shading indicating total protein sizes. Figure S3 summarizes the sequence features of the C-terminal (yellow) motifs for each of the 14 groups.