Exploring the Secretomes of Microbes and Microbial Communities Using Filamentous Phage Display

Abstract

Microbial surface and secreted proteins (the secretome) contain a large number of proteins that interact with other microbes, host and/or environment. These proteins are exported by the coordinated activities of the protein secretion machinery present in the cell. A group of bacteriophage, called filamentous phage, have the ability to hijack bacterial protein secretion machinery in order to amplify and assemble via a secretion-like process. This ability has been harnessed in the use of filamentous phage of Escherichia coli in biotechnology applications, including screening large libraries of variants for binding to "bait" of interest, from tissues in vivo to pure proteins or even inorganic substrates. In this review we discuss the roles of secretome proteins in pathogenic and non-pathogenic bacteria and corresponding secretion pathways. We describe the basics of phage display technology and its variants applied to discovery of bacterial proteins that are implicated in colonization of host tissues and pathogenesis, as well as vaccine candidates through filamentous phage display library screening. Secretome selection aided by next-generation sequence analysis was successfully applied for selective display of the secretome at a microbial community scale, the latter revealing the richness of secretome functions of interest and surprising versatility in filamentous phage display of secretome proteins from large number of Gram-negative as well as Gram-positive bacteria and archaea.

Keywords: adhesins; bacteriophage; metagenomics; next generation sequencing; phage display; secretome.

FIGURE 1

The Ff bacteriophage structure and virion proteins used most commonly in phage display. (A) Ff virion visualized by atomic force microscope (M. Russel and P. Model, sample prepared by J. Rakonjac). (B) Schematic diagram of Ff bacteriophage. (C) Ribbon representations (top and side view) of the pVIII coat protein (RCSB PDB database accession number 2cOw; (Marvin et al., 2006) arranged around bacteriophage single stranded DNA (not shown). (D) Ribbon representation of N1 and N2 domains of pIII (RCSB PDB database accession number1g3p; Lubkowski et al., 1998).

FIGURE 2

Ff virion protein targeting, virion assembly, phagemid display system, and display cassette in phagemid vectors. (A) In Escherichia coli phagemids can be replicated as plasmids or alternatively, in the presence of a helper phage, packaged as “transducing” or “phagemid” particles (PPs). Phagemid encodes phage protein pIII as a fusion partner for display. The resulting phagemid particles may incorporate either pIII derived from the helper phage (red lollipop-like structure) or the polypeptide-pIII fusion protein (red lollipop-like structure decorated with a yellow star), encoded by the phagemid. Ab^R, antibiotic resistance marker, Ff ori; PS, filamentous phage origin of replication and packaging signal; Plasmid ori, plasmid origin of replication; gIII, gene III; pIII, pV, pVI, pVII, pVIII, and pIX, filamentous phage proteins. Phagemid particles are produced at a 10–100-fold excess over the helper phage (denoted by solid vs. faded lines for phagemid particle vs. helper phage. (B) Typical phage display cloning cassette in a phagemid vector: promoter, a ribosome binding site (RBS), signal sequence (commonly used PelB signal sequence from Erwinia carotovora (Lei et al., 1987); multiple cloning site (MCS), affinity tag (tag), and sequence encoding the mature portion (C-terminal domain) of pIII (gIIIC; required for assembly of the fusion into the virion). Two types of inserts derived from fragmented bacterial or archaeal genomic DNA are shown above the cloning site. Insert without signal sequence, in order to result in displayed peptide, has to correspond to a CDS that is in frame with the upstream signal sequence and downstream gIII (encoding the C-terminal domain). It is typically truncated at the 5′ and 3′ ends to avoid stop codons that terminate translation. A second type of insert, that contains signal sequence, can be displayed if the CDS is truncated at the 3′ end, and is in frame with gIII. The latter type of inserts typically carries its own promoter and RBS. E. coli host strains that contain suppressor mutations (such as supE44) can read through amber stop codons at 50% efficiency and result in display of fusions that include this codon.

FIGURE 3

Biopanning – a basic selection for binding peptides. (1) Display: Filamentous phage displaying variants of proteins/peptides/antibodies is created, cloned into phage, or phagemid vectors as fusions to a coat protein gene(s) and displayed on the surface of the virions. (2) Panning: The phage library displaying variant peptides or proteins (different-color stars) is exposed to immobilized ligand (yellow pentagon) and phage with appropriate binding specificity is captured (yellow star). (3) Enrichment: Non-binding phages are washed off and bound phage(s) is (are) eluted by conditions that disrupt the peptide-ligand interactions, leading to enrichment for a specific binder. (4) Amplification: Eluted phage is then amplified by infection of a suitable E. coli strain. This amplified phage population is greatly enriched in recombinant phage clones displaying peptides that bind to the target. The biopanning steps (two to four) are repeated for several (three to five) rounds, ultimately resulting in a clonal population of recombinant phages that bind to the target used for affinity panning of the library. Captured putative binder can then be identified by sequencing (5) and functionally analyzed.

FIGURE 4

The construction of shotgun metagenomic and metasecretome filamentous phage display libraries. Metagenomic DNA is randomly sheared and cloned into phagemid that contains signal sequence (+ss) or as in case for metasecretome into phagemid without signal sequence (-ss). In both cases constructed metagenomic inserts contain endogenous signal sequences, represented by red ovals. Depending on helper phage used [wild-type (wt) or gIII-deleted helper phage] for aid in replication and assembly of recombinant virions, the library will contain virions displaying the whole metaproteome (metagenome phage display) or it will consist of virions capped by insert-pIII fusion proteins (signal sequence-positive clones) that are resistant to sarkosyl (Sarkosyl^R, virions inside the dotted line) and uncapped virions (signal sequence-negative clones) that are sensitive to sarkosyl (Sarkosyl^S). Sarkosyl resistance is used as a basis for selection in metasecretome phage display. Single stranded DNA (ssDNA) purified from Sarkosyl^R virions after the selection can be used to obtain the amplified metasecretome plasmid library for preliminary assessment of metasecretome diversity by next-generation sequencing [Taken from (Ciric et al., 2014) with permission].

FIGURE 5

Taxonomic distribution, signal sequence types, and cellulosome components enrichment in metasecretome phage display of fiber-adherent rumen microbial community. (A) Distribution of signal sequence types and transmembrane helices in selected metasecretome recombinants; Abbreviations: Type I ss, classical signal sequence; Type II ss, lipoprotein signal sequence; Type IV ss, prepilin-like signal sequence; TMH, N-terminal or internal transmembrane α- helix/helices; background, ORFs encoding putative proteins without a predicted membrane-targeting signal/non-classical secretion, or ORFs encoding putative proteins and peptides ≤24 amino acid residues. (B) Taxonomic distribution of the phage-display-selected metasecretome; The taxonomic assignments, at the phylum level, were based on the distribution of the best BLASTP hits at a 30% amino acid sequence identity threshold for protein-coding genes predicted in metagenome and metasecretome datasets. Each section of the stacked columns represents the percentage of total protein-coding genes assigned to the corresponding phylum. The section labeled ‘Other’ contains putative protein-coding genes assigned to a phylogenetic group with low abundance in the dataset (<0.1%), while the section labeled ‘Unassigned’ corresponds to putative protein-coding genes with best BLASTP hit below 30% identity cut-off. (C) Enrichment for the components of cellulosome; Frequency of three putative distinct ‘signature’ cellulosome modules: cohesins (blue); dockerins (red) and surface S-layer homology (SLH) domains (green) in three datasets: MS, metasecretome dataset; MG, metagenome dataset and published deep-sequenced metagenome (DMG). The latter dataset is from (Hess et al., 2011). (A,C) are taken from (Ciric et al., 2014) and (B) from (Ciric, 2014) with permission.