Mapping polyclonal antibody responses to bacterial infection using next generation phage display

Abstract

Mapping polyclonal antibody responses to infectious diseases to identify individual epitopes has the potential to underpin the development of novel serological assays and vaccines. Here, phage-peptide library panning coupled with screening using next generation sequencing was used to map antibody responses to bacterial infections. In the first instance, pigs experimentally infected with Salmonella enterica serovar Typhimurium was investigated. IgG samples from twelve infected pigs were probed in parallel and phage binding compared to that with equivalent IgG from pre-infected animals. Seventy-seven peptide mimotopes were enriched specifically against sera from multiple infected animals. Twenty-seven of these peptides were tested in ELISA and twenty-two were highly discriminatory for sera taken from pigs post-infection (P < 0.05) indicating that these peptides are mimicking epitopes from the bacteria. In order to further test this methodology, it was applied to differentiate antibody responses in poultry to infections with distinct serovars of Salmonella enterica. Twenty-seven peptides were identified as being enriched specifically against IgY from multiple animals infected with S. Enteritidis compared to those infected with S. Hadar. Nine of fifteen peptides tested in ELISA were highly discriminatory for IgY following S. Enteritidis infection (p < 0.05) compared to infections with S. Hadar or S. Typhimurium.

Figure 1. Screening of peptides specific for S. Typhimurium infection in pigs.

Peptides identified in silico as being enriched against IgG from pigs experimentally infected with S. Typhimurium were assessed for binding to sera antibodies from 24 animals taken before challenge (pre-infection) and after challenge (post-infection). Synthetic peptide was immobilised on maxisorb plates and probed with the sera, bound antibody was then detected with an anti-porcine-IgG-AP conjugate. All samples were analysed in duplicate, ELISA signals after 2 hours incubation with substrate were taken. Binding was defined by applying a cut-off value for each peptide that was determined as the mean +3SD of the signals for the 24 non-infected samples (dashed line in C,E). ROC analysis for each peptide revealed 22 peptides (out of 27) with associated p values < 0.05 indicating that they are highly discriminatory for sera from infected pigs. Binding of sera samples to these 22 peptides is shown (A) as being positive (grey boxes) or negative (white boxes). Animal numbers are shown and those in italics provided sera samples for the phage panning. Analysis of the 22 peptides showed binding to all 24 sera samples from infected pigs but not to any sera from non-infected pigs. Examples of binding of individual peptides are also shown for AEGEFVQATDTNS (B,C) and AEGEFPLHNGNERL (D,E) and both bound to 13 of the 24 samples from infected pigs. ROC analysis is shown (B,D) along with the ELISA signals for each of the 24 infected and non-infected animals binding to peptides (C,E). These peptides had 100% specificity and 54% sensitivity, the positive and negative predicted values are shown for the samples tested.

Figure 2. Screening of peptides specific for S. Enteritidis infection in chickens.

Peptides identified in silico as being enriched against IgY from chickens experimentally infected with S. Enteritidis compared to birds infected with S. Hadar were assessed for binding to IgY from 19 birds infected with S. Enteritidis, 9 infected with S. Hadar and 16 infected with S. Typhimurium. Synthetic peptide was immobilised on maxisorb plates and probed with the IgY, bound antibody was then detected with an anti-IgY-AP conjugate. All samples were analysed in duplicate, ELISA signals after 2 hours incubation with substrate were taken. Binding was defined by applying a cut-off value for each peptide that was determined as the mean +3SD of the signals for the 24 non-infected samples (dashed line in C). ROC analysis for each peptide revealed 9 peptides (out of 15) with associated p values < 0.05 indicating that they are highly discriminatory for sera from S. Enteritidis infected chickens. Binding of IgY samples to these 9 peptides is shown (A) as being positive (grey boxes) or negative (white boxes). Animal numbers are shown and those in italics provided IgY for the phage panning. The most discriminatory individual peptides was AEGEFEPQQSARPS that bound to 7 of the 19 samples from S. Enteritidis infected chickens. ROC analysis is shown (B) along with the ELISA signals (C). This peptide had 100% specificity and 37% sensitivity, the positive and negative predicted values are shown for the samples tested.

Figure 3

Panning of peptide libraries against purified polyclonal IgG (A) or IgY (B). Panel A: Phage libraries were bound in parallel to 12 IgG samples from pigs infected with S. Typhimurium and after elution these 12 sub-libraries of phage were propagated and groups of 3 were pooled (round1, dashed arrows) in equal phage numbers. Each sub-library pool was then bound to IgG from the same 3 animals used in round 1 this time using IgG taken post-challenge (grey boxes) and IgG taken pre-challenge (white boxes) with the pathogen (round 2, dashed arrows). Bioinformatic analysis to assess for enrichment of individual peptides specifically against IgG from infected animals was carried out by comparing the binding of peptides to IgG taken pre- and post-infection with the pathogen (solid arrows). Panel B: Phage libraries were bound in parallel to 9 IgY samples from chickens infected with S. Enteritidis and after elution these 9 sub-libraries of phage were propagated. Each sub-library was then bound to IgY from the same S. Enteritidis-infected chicken used in round 1 (grey boxes) and also against IgY from one of nine chickens infected with S. Hadar (white boxes) as indicated by dashed arrows. Bioinformatic analysis to assess for enrichment of individual peptides specifically against IgY from S. Enteritidis-infected animals was carried out by comparing the binding of peptides to IgY from S. Enteritidis-infected birds with those infected with S. Hadar (solid arrows).