Production and evaluation of the utility of novel phage display-derived peptide ligands to Salmonella spp. for magnetic separation

Authors


Correspondence

Irene R. Grant, Institute for Global Food Security, School of Biological Sciences, Queen's University Belfast, 97 Lisburn Road, Belfast BT9 7BL, UK. E-mail i.grant@qub.ac.uk

Abstract

Aims

The objectives of this study were to produce Salmonella-specific peptide ligands by phage display biopanning and evaluate their use for magnetic separation (MS).

Methods and Results

Four-phage display biopanning rounds were performed, and the peptides expressed by the two most Salmonella-specific (on the basis of phage-binding ELISA results) phage clones, MSal020401 and MSal020417, were chemically synthesized and coupled to MyOne™ tosylactivated Dynabeads®. Peptide capture capability for whole Salmonella cells from nonenriched broth cultures was quantified by MS + plate counts and MS + Greenlight™ detection and compared to capture capability of anti-Salmonella (antibody-coated) Dynabeads®. MS + Greenlight™ gave a more comprehensive picture of capture capability than MS + plate counts and showed that Peptide MSal020417-coated beads exhibited at least similar, if not better, capture capability to anti-Salmonella Dynabeads® (mean capture values of 36·0 ± 18·2 and 31·2 ± 20·1%, respectively, over Salmonella spp. concentration range 3 × 101–3 × 106 CFU ml−1) with cross-reactivity of ≤1·9% to three other foodborne pathogens: Escherichia coli, Listeria monocytogenes and Campylobacter jejuni.

Conclusions

One of the phage display-derived peptide ligands was demonstrated by MS + Greenlight™ to be a viable antibody alternative for MS of Salmonella spp.

Significance and Impact of the Study

This study demonstrates an antibody-free approach to Salmonella detection and opens substantial possibilities for more rapid tests for this bacterium.

Introduction

Salmonella, a genus belonging to the Enterobacteriaceae family, is comprised of two species, bongori and enterica (Jacobsen and Bech 2012). Salmonella enterica is further classified into six subspecies with over 2500 serovars currently known, of which 1531 are part of the subspecies enterica and are the causative agents of more than 99% of human Salmonellosis cases (Jacobsen and Bech 2012). Traditional methods for detection and isolation of Salmonella spp. in food are culture-based with their multi-step format rendering them time-consuming and laborious (Walker et al. 2001; Alakomi and Saarela 2009). A variety of more rapid methods have been developed that utilize different technologies for example, immuno-based methods such as biosensor (Chai et al. 2012), enzyme-linked immunosorbent assay (ELISA; Bang et al. 2012), antibody array (Karoonuthaisiri et al. 2009; Charlermroj et al. 2011) and lateral flow device (Torlak et al. 2012) and molecular techniques such as real-time polymerase chain reaction (Tatavarthy and Cannons 2010), amongst others.

Immuno-based methods require animal-derived antibodies that are difficult and expensive to produce. In addition, the scientific research community is making concerted efforts to reduce, refine and replace the use of animals in scientific research in accordance with EU legislation, Directive 2010/63/EU (European Union 2010). One way to generate antibody-alternative binders is by phage display biopanning, a technique extending from the invention of phage display in Smith (1985). Previous research in relation to phage display and Salmonella has predominantly been therapeutic interventional strategies aimed for the medical industry where the phage library was used to obtain peptide ligands to Salmonella spp. indirectly by using lipopolysaccharide (LPS), whole LPS or LPS moiety, from Salmonella spp. or other Gram-negative bacteria as the target (Noda et al. 2001; Thomas et al. 2003; Zhu et al. 2003; Kim et al. 2005, 2006; Guo and Chen 2006). Diagnostic applications aimed towards the environmental and food industries used phage display biopanning to obtain antibody-alternative binders for use in Salmonella detection systems (Sorokulova et al. 2005; Petrenko 2008).

This study employs phage display biopanning against gamma-irradiated whole Salmonella cells to produce high-affinity peptide ligands for coupling to magnetic beads and subsequent use for magnetic separation (MS). The cocktail contained the eight Salmonella serovars most commonly associated with foodborne outbreaks in Europe (European Food Safety Authority and European Centre for Disease Prevention and Control 2012). Capture capability of peptide-coated beads was quantified by MS + plate counts and MS + Greenlight™ detection and compared to that of commercially available antibody-coated beads, with the end goal to determine the utility of phage display-derived peptide ligands as antibody-alternative binders for MS of Salmonella spp.

Material and methods

Preparation of target antigen and biopanning plates

Target antigen for phage display biopanning consisted of a cocktail of Salm. enterica subspecies enterica serovars (Table 1). Each serovar culture was standardized to a cell concentration of 2 × 109 CFU ml−1, equal volumes of each suspension mixed, centrifuged (3000 g, 15 min, 20°C) and the pellet resuspended to the original volume with phosphate buffered saline pH 7·4 (PBS). This Salmonella cocktail was subjected to a 10 kGy dose of gamma radiation (using a Gammabeam 650 irradiator located at Agri-Food and Biosciences Institute for Northern Ireland, Belfast). Residual individual preparations of each serovar were also centrifuged, resuspended and irradiated as described above. Prior to the coating of four 60-mm Petri dishes (Sarstedt, Leicester, UK), the irradiated Salmonella spp. cocktail underwent 10-fold concentration [centrifugation at 3000 g, 15 min, 20°C, with resuspension of the pellet in coating buffer (0·1 mol l−1 sodium hydrogen carbonate pH 8·6) to 1/10 original volume] to 2 × 1010 CFU ml−1. The Petri dishes were incubated overnight in a humidified container with agitation at 4°C.

Table 1. Bacterial strains employed in this study
BacteriumSerogroupSource
  1. a

    National Collection of Type Cultures, Colindale, London, UK.

  2. b

    Originally isolated and serotyped by Salmonella Reference Laboratory, Agri-Food and Biosciences Institute for Northern Ireland, Belfast, UK and kindly provided by Dr Robert Madden.

  3. c

    Laboratory of the Government Chemist, Middlesex, UK.

Salmonella Enteriditis9,12:g,m:7NCTCa 6676
Salmonella Typhimurium4,12:i:2Pig carcass swabb
Salmonella Dublin9,12:g,pPorkb
Salmonella Infantis6,7:r:5Raw chickenb
Salmonella Senftenberg3,19:g,s,tAnimal feedb
Salmonella Hadar6,8:z10:e,n,xQA sample- LGCc
Salmonella Mbandaka6,7:z10:e,n,z15Hygiene swabb
Salmonella Virchow6,7:r:2NCTC 5742
Escherichia coli K12NCTC 10538
Listeria monocytogenes NCTC 4885
Campylobacter jejuni NCTC 11351

Production and ELISA screening of phage clones to Salmonella

Four biopanning rounds were performed using a phage display peptide library [PhD-12; New England Biolabs (NEB), Herts, UK] according to the manufacturer's instructions in PhDTM Phage Display Libraries Instruction Manual (www.neb.com). After both third and fourth-round biopanning, 18 phage clones were randomly selected for individual amplification, DNA extraction and sequencing (using −96gIII sequencing primer) by DNA Sequencing and Services, University of Dundee, Dundee, Scotland. Corresponding 12-mer peptide sequences were deciphered using FinchTV software (http://www.geospiza.com/Products/finchtv.shtml) and ExPASy translate tool (http://web.expasy.org/translate/), respectively. Consensus between peptide sequences was assessed using online ClustalW software (http://www.ebi.ac.uk/Tools/msa/clustalw2/). The SAROTUP tool (Huang et al. 2010) was used to ensure the deciphered peptide sequences were unique and not previously recognized target-unrelated peptide sequences (Vodnik et al. 2011). Unique peptide sequences were screened for their ability to bind to irradiated Salmonella spp. and three heat-killed nontarget bacteria (Table 1), at bacterial cell concentration of 1 × 108 CFU ml−1. The phage-binding ELISA protocol followed the manufacturer's instructions (http://www.neb.com/nebecomm/manualfiles/manuale8101.pdf).

Chemical synthesis of 12-mer peptides expressed by Salmonella-binding phages, their coupling to magnetic beads and subsequent use in MS

The 12-mer peptides expressed by two phage clones from fourth-round biopanning, Peptide MSal020401 and Peptide MSal020417, were chemically synthesized (GL Biochem, Shanghai, China) at >85% purity with a glycine-glycine-glycine-serine-cysteine spacer at the C-terminus as suggested by the manufacturer (http://www.neb.com/nebecomm/manualfiles/manuale8101.pdf). The synthetic peptides were reconstituted with molecular-grade water to 10 mg ml−1, and 100 μl of peptide coupled to 250 μl of MyOne™ tosylactivated Dynabeads® (bead size of 1 μm, bead concentration of 1 × 1012 beads ml−1; Life Technologies Limited, Paisley, UK) as per the manufacturer's instructions and the coupled beads resuspended with a final volume of 500 μl of PBS with 0·1% bovine serum albumin (bead concentration of 5 × 109 beads 10 μl−1).

A Dynal® BeadRetriever (Life Technologies Limited) was used for automated MS. The test sample was nonirradiated Salmonella spp. cocktail (Table 1). Serial dilutions (10-fold) of standardized cocktail were prepared (10−1–10−7) in nutrient broth. While the bead concentration for the peptide-coated beads was known, it was not known for the commercial antibody-coated beads (proprietary information), therefore the comparison study simply used the same bead volume for each bead type. The MS protocol used 10 μl of antibody- or peptide-coated beads on duplicate aliquots of six Salmonella cell concentrations (3 × 101–3 × 10CFU ml−1) and was as follows: capture of beads for 30 min at very slow speed (1 ml bacterial dilution and 10-μl beads), 10 s release of beads at very fast speed, with two washes for 1 min in 1 ml PBS with 0·05% tween 20 at very slow speed followed by 10-s release of beads at fast speed with final 2-min elution at fast speed into 100 μl of nutrient broth (a 10-fold concentration was achieved during MS). Alongside, the peptide-coated beads negative and positive control beads (commercial anti-Listeria Dynabeads® and anti-Salmonella Dynabeads®, respectively; Life Technologies Limited) were included in every run, which was performed in triplicate. The number of Salmonellae in MS products was quantified by MS + plate counts and MS + Greenlight™ detection.

Evaluation of coated beads by MS + plate counts

Plate counts were performed pre- and post-MS for each bead type by spreading 100 μl sample, or appropriate dilution, onto a nutrient agar plate and incubating overnight at 37°C. The chosen concentration was 3 × 103 CFU ml−1 because a countable number of colonies would result. The protocol was performed in triplicate and mean % capture values for each bead type were subsequently calculated by dividing the number of colonies post-MS by the number pre-MS for the same dilution and expressing this as a percentage.

Evaluation of coated beads by MS + Greenlight™ detection

Luxcel Biosciences Limited (Cork, Republic of Ireland) has developed an oxygen-sensitive, fluorescence probe, GreenLight™ probe, whose signal increases when oxygen levels decrease. This signal, termed Threshold Time (h), denotes the time taken for the first measureable change in oxygen levels within the sample that is compared against a predetermined standard curve for the target bacterium and a bacterial concentration (CFU ml−1) attributed to unknown samples. The Greenlight™ 960 system is a 96-well format allowing 96 samples (including controls) to be measured simultaneously in real-time via a fluorescence-based plate reader, FLUOstar Omega reader (BMG LABTECH, Bucks, UK) over the required incubation period for the target bacterium. Data processing and calculations are automatically performed by FLUOstar Omega Analysis software.

GreenLight™ probe was reconstituted in nutrient broth and mixed 1 : 1 (100 μl : 100 μl) with test samples in the wells of a sterile 96-well plate (Sarstedt). Each test sample was overlaid with 100 μl of mineral oil to prevent ambient oxygen influencing the closed environment of the probe/sample mix. Standard curve determination for Salmonella spp. was performed in duplicate using the dilutions prepared for MS (containing 101–10CFU ml−1). The MS products from evaluating the negative control bead, positive control bead and peptide-coated beads with 10-fold dilutions of standardized Salmonella spp. (102–10CFU ml−1) were also analysed in duplicate. The protocol ran for approx. 13 h at 37°C taking measurements every 10 min. When complete, threshold time values were exported to Microsoft Excel 2007 and the Salmonella spp. standard curve for pre-MS dilutions generated along with its linear regression equation and correlation coefficient (R2). In order to determine the CFU ml−1 for post-MS samples, the threshold time values were related back to numbers of Salmonellae using this standard curve equation. The protocol was performed in triplicate and mean % capture values for each bead type were subsequently calculated by dividing the number of salmonellae post-MS by the number pre-MS for the and expressing this as a percentage.

Cross-reactivity of coated beads by MS + plate counts and MS + Greenlight™ detection

Cross-reactivity of Peptide MSal020417-coated bead and anti-Salmonella Dynabeads® was assessed either by MS + plate counts or by MS + Greenlight™ detection depending on the nontarget bacterium of interest. The test samples for MS were cultures of nonirradiated Listeria monocytogenes, Escherichia coli K12 and Campylobacter jejuni (Table 1). Each culture was standardized to a cell concentration of 3 × 108 CFU ml−1 and 10-fold serial dilutions were prepared (10−1–10−7) in nutrient broth. Nutrient broth was used as a negative control. Automated MS was performed using 10 μl of antibody- or peptide-coated beads on duplicate aliquots of two dilutions, 10−5 and 10−6 (should contain 3 × 104–3 × 10bacterial cells ml−1 due to the 10-fold concentration during MS). The number of L. monocytogenes and E. coli K12 captured was quantified by two separate Greenlight™ protocols, one at 30°C for L. monocytogenes (approx. 23 h) and the other at 37°C for E. coli K12 (approx. 13 h). Standard curves were generated for L. monocytogenes and E. coli K12 in duplicate using the dilutions prepared for MS (containing 102–10CFU ml−1). The protocol was performed in triplicate and mean % capture values for each bead type and each bacterium were subsequently calculated. Bead capture capability for Camp. jejuni was quantified by plate counts only because microaerophilic conditions cannot be achieved with Greenlight™ detection. Post-MS 100 μl beads were spread onto the surface of plates of Columbia blood agar plates (Oxoid Limited, Hampshire, UK) supplemented with 5% sterile defibrinated blood and incubated for 48 h at 42°C under microaerophilic conditions. The protocol was performed in triplicate and mean % capture values for each bead type were subsequently calculated.

Statistical analysis of results

Statistical analysis of results was carried out using GraphPad Instat version 3.10 (GraphPad, San Diego, CA, USA). Analysis of variance (anova) was used to compare % capture of target Salmonella spp. by the four different bead types, and a paired t-test was used to compare % capture of nontarget bacteria (E. coli, L. monocytogenes and Camp. jejuni) by the peptide MSal020417-coated and anti-Salmonella antibody-coated beads.

Results

Phage binders from third-round biopanning

DNA sequences were obtained for 11 of the 18 phage clones; however, no two sequences were the same (Table 2). The 11 phage clones were screened via phage-binding ELISA for their ability to bind irradiated Salmonella spp. (data not shown), and the four phage clones exhibiting best binding capability were further characterized via phage-binding ELISA for their ability to bind eight individual irradiated Salmonella serovars and three heat-killed nontarget bacteria (Fig. 1a). OD measurements for specific binding ranged from 0·009 to 0·317 and for cross-reactivity with nontarget bacteria from 0·045 to 0·243. Only one of the four phage clones, MSal020317 (deduced peptide HPLLTYSASQKG), exhibited significantly higher binding to the eight Salmonella serovars than to the three nontarget bacteria, the other three did not.

Table 2. Outcomes from third-round biopanning of irradiated Salmonella spp. cocktail
Clone identifier(s)Amino acid sequenceFrequencyaAbility to bind Salmonella spp. cocktail
  1. a

    Number of clones with sequence/total number of clones tested after 3rd-round biopanning.

  2. b

    This amino acid sequence is obtained when a peptide sequence is not expressed on outside of phage coat.

  3. c

    Phage clones not tested by phage-binding ELISA.

MSal020301M Q G H – E L G F M K I1/18No
MSal020302 A E T V E S C L A K S Hb6/18c
MSal020303M H Q G P T S Y N L Q M1/18No
MSal020304H P N G H I L L E L R Q1/18Yes
MSal020305L S N T L Q T – T N K K1/18No
MSal020306E L Q M R A L Q R L Q P1/18No
MSal020307C L R N A S P H L G C L1/18Yes
MSal020308Unsuccessful sequencing1/18
MSal020309H T S D G L Q L S N V Q1/18No
MSal020310W P Q G P D – R A V L G1/18No
MSal020314S V H L Y S Q M P T K K1/18No
MSal020316S R M P T M I M E K G W1/18Yes
MSal020317H P L L T Y S A S Q K G1/18Yes
Figure 1.

Phage-binding ELISA results for phage clones were obtained after (a) third-round biopanning and (b) fourth-round biopanning demonstrating the degree of binding to eight individual Salmonella serovars (white bars from left to right: Salm. Infantis, Salm. Virchow, Salm. Senftenberg, Salm. Typhimurium, Salm. Mandaka, Salm. Dublin, Salm. Hadar and Salm. Enteriditis). For both graphs, the normalized OD data were obtained by subtracting the OD measurement for negative control (coating buffer well) from OD measurement for bacteria coated well. (image_n/jam12207-gra-0001.png) Salmonella spp.; (image_n/jam12207-gra-0002.png) L. monocytogenes; (image_n/jam12207-gra-0003.png) E. coli K12 and (image_n/jam12207-gra-0004.png) Camp. jejuni

Phage binders from fourth-round biopanning

DNA sequences were obtained for 14 of the 18 phage clones and three clones had the same sequence, HIRWDVNHNSMS (Table 3). The 14 phage clones were screened via phage-binding ELISA for their ability to bind irradiated Salmonella spp. (data not shown), and the four phage clones exhibiting best binding capability were further characterized via phage-binding ELISA for their ability to bind eight individual irradiated Salmonella serovars and three heat-killed nontarget bacteria (Fig. 1b). OD measurements for specific binding ranged from 0·098 to 0·793 and for cross-reactivity with nontarget bacteria from 0·001 to 0·265. Two of the four phage clones, MSal020401 (deduced peptide SEAYKHRQMHMS) and MSal020417 (deduced peptide NRPDSAQFWLHH), exhibited higher binding to the eight Salmonella serovars than to the three nontarget bacteria, with the possible exception of Salmonella Dublin, the other two phage clones did not.

Table 3. Outcomes from fourth-round biopanning of irradiated Salmonella spp. cocktail
Clone identifier(s)Amino acid sequenceFrequencyaAbility to bind Salmonella spp. cocktail
  1. a

    Number of clones with sequence/total number of clones tested after 4th-round biopanning.

  2. b

    Recognized or suspected plastic binder (Huang et al. 2010).

  3. c

    Phage clones not tested by phage-binding ELISA.

MSal020401S E A Y K H R Q M H M S1/18Yes
MSal020402Y Y P H P P W S P Q H1/18No
MSal020403 H I R W D V N H N S M S3/18Yes
MSal020404V P W V T T Y E P W G M1/18Yes
MSal020406V S A A R A D F Y A A M1/18No
MSal020407S S Y Y P Q L T A H R F1/18No
MSal020408S A K T H P W S I W A Yb1/18c
MSal020412W H N A W E S W H Y A Nb1/18
MSal020414K L P P S F D L T G A N1/18Yes
MSal020415G P A D N T S K H V I R1/18Yes
MSal020417N R P D S A Q F W L H H1/18Yes
MSal020418S W M P H P R W S P Q H1/18Yes
Four clonesUnsuccessful sequencing4/18

Evaluation of coated beads by MS + plate counts

Figure 2 shows mean % capture results obtained using MS + plate counts. Both anti-Salmonella peptide-coated beads were able to capture Salmonella spp. from a broth suspension containing 3 × 103 CFU ml−1 pre-MS. However, Peptide MSal020417-coated beads were able to capture a significantly higher % of Salmonellae than Peptide MSal020401-coated beads (mean 85·6 vs 3·2%, anova P < 0·05). Mean % capture of Salmonellae by Peptide MSal020417-coated beads and anti-Salmonella antibody-coated beads was not significantly different (mean 85·6 vs 33·6%, anova P > 0·05), due to the large variability in % capture observed for Peptide MSal020417-coated beads. No colonies were obtained after MS for the negative control bead (anti-Listeria Dynabeads®).

Figure 2.

Capability of antibody- and peptide-coated magnetic beads to capture Salmonella spp. from a nutrient broth suspension containing 3 × 103 CFU ml−1 as assessed by MS + plate counts. Experiment performed in triplicate (n = 3). Error bars represent ± standard error of mean. MS, magnetic separation.

Evaluation of coated beads by MS + Greenlight™ Model 960

Figure 3 shows capture results obtained using MS + Greenlight™ detection. The standard curve for Salmonella spp. cocktail, over concentration range 3 × 101–3 × 107 CFU ml−1 pre-MS, had R2 value of 0·9958 and linear regression equation y = −1·2699x + 9·0145. This equation was used to calculate the number of Salmonella captured during MS and then to calculate % capture relative to Salmonella concentration pre-MS. Both types of peptide-coated beads were able to capture Salmonella spp., but to differing degrees. Peptide MSal020417-coated beads were consistently able to capture a higher % of Salmonella spp. than Peptide MSal020401-coated beads (mean 36·0 vs mean 4·9%, anova P < 0·001) over concentration range 3 × 101–3 × 106 CFU ml−1 pre-MS. There was no significant difference between % capture by the Peptide MSal020417-coated beads and anti-Salmonella antibody-coated beads when 3 × 10, 3 × 102, 3 × 103 or 3 × 106 Salmonellae per millilitre were present pre-MS (P > 0·05, Fig. 3). However, when 3 × 104 and 3 × 105 Salmonellae per millilitre were present pre-MS the peptide-coated beads achieved significantly greater % capture than the anti-Salmonella antibody-coated beads (P < 0·05 and P < 0·01, respectively, Fig. 3). The mean % capture value for the negative control bead (anti-Listeria Dynabeads®) was 0·4%.

Figure 3.

Capability of antibody- and peptide-coated magnetic beads to capture Salmonella spp. from nutrient broth suspensions containing 3 × 101–3 × 106 Salmonellae ml−1 as assessed by magnetic separation (MS) + Greenlight™ detection. Experiment performed in triplicate (n = 3) and error bars represents ± standard error of mean. Statistical significance of difference between % capture by the commercial anti-Salmonella antibody-coated Dynabeads and the Peptide MSal020417-coated beads is indicated above bars, nsP > 0·05, *P < 0·05 and **P < 0·01. (□) Anti-Listeria antibody-coated beads; (image_n/jam12207-gra-0001.png) Anti-Salmonella antibody-coated beads; (image_n/jam12207-gra-0002.png) Peptide MSal020417-coated beads and (image_n/jam12207-gra-0004.png) Peptide MSal020401-coated beads.

Cross-reactivity of coated beads by MS + plate counts and MS + Greenlight™ detection

Table 4 shows the cross-reactivity of Peptide MSal020417-coated bead and anti-Salmonella Dynabeads® with viable nontarget bacteria, L. monocytogenes, E. coli K12 and Camp. jejuni, as assessed by MS + Greenlight™ detection or MS + plate counts. MS + Greenlight™ detection results showed that the L. monocytogenes standard curve over concentration range 3 × 101–3 × 107 CFU ml−1 pre-MS had R2 value of 0·9846 with linear regression equation y = −3·1080x + 24·5900, and the E. coli K12 standard curve over the same concentration range had R2 value of 0·9863 with linear regression equation y = −1·3176x + 11·2440. No standard curve was generated for Camp. jejuni because Greenlight™ detection could not be used for microaerophilic incubation. Cross-reactivity for both bead types with any of the nontarget bacteria tested was not significantly different (≤1·9% for Peptide MSal020417-coated beads and ≤0·8% for anti-Salmonella antibody-coated beads, paired t-test P > 0·05).

Table 4. Capability of antibody- and peptide-coated magnetic beads to capture nontarget bacteria from nutrient broth suspensions containing 3 × 104 and 3 × 103 CFU ml−1 as evaluated by either magnetic separation (MS) + plate counts or MS + Greenlight™ detection. Experiment performed in triplicate, and data represent mean % capture ± standard deviation
Nontarget bacteriumNo. bacteria pre-MS (CFU ml−1) Mean % capture by bead type
Anti-Salmonella antibody-coated Dynabeads®)Peptide MSal020417-coated bead
  1. a

    Evaluated by MS + Greenlight™ detection at 37°C.

  2. b

    Evaluated by MS + Greenlight™ detection at 30°C.

  3. c

    Evaluated by MS + plate counts.

Escherichia coli a 3 × 1040·5 ± 0·51·9 ± 1·5
3 × 1031·2 ± 2·01·2 ± 1·6
Listeria monocytogenes b 3 × 1040·0 ± 0·01·1 ± 0·7
3 × 1030·0 ± 0·01·2 ± 0·7
Campylobacter jejuni c 3 × 1040·0 ± 0·00·2 ± 0·1
3 × 1030·0 ± 0·00·1 ± 0·1

Discussion

Phage display biopanning is an in vitro selection process, whereby high-affinity phage binders to a target are selected for and amplified in number through a series of biopanning rounds. Each successive biopanning round should produce phage binders with increased specificity for the target and have increased representation within the phage library thereby reducing library diversity and highlighting the amino acid sequences important for phage-target binding. The phage-binding ELISA results obtained in this study demonstrate that with successive biopannings specificity to the target did increase along with consensus; the OD measurements for specific binding increased from third to fourth-round biopanning while the cross-reactivity with nontarget bacteria remained relatively stable (Fig. 1), and no two peptides from third-round biopanning had the same sequence (Table 2), while three peptides from fourth-round biopanning had the same sequence (Table 3).

This study employed gamma-irradiated Salmonella spp. as the target for phage display biopanning. Gamma-irradiated cells are nonviable, but their cell morphology is retained along with their antigen characteristics (Stewart et al. 2012) and was a technique successfully employed by Grant et al. (1998) and Stewart et al. (2012) where gamma-irradiated mycobacteria were used as the target/immunogen in polyclonal and monoclonal antibody production. The use of irradiated bacteria rendered the bacteria safe for transfer between collaborators, and as the antigenic properties are maintained, the peptide ligands produced should be viable binders for use in detecting Salmonella. It should be noted that the nontarget bacteria employed in the cross-reactivity experiments with phage clones were heat-killed cells and may not be an accurate representation of viable cells as heat treatment has a damaging effect on cells especially heat-sensitive surface antigens. The nontarget bacteria employed in the cross-reactivity study with peptide-coated beads were viable cells, therefore, the results obtained are probably a more accurate representation of cross-reactivity to nontarget bacteria.

The two best binding phage clones from fourth-round biopanning, MSal020401 and MSal020417, had their 12-mer peptides chemically synthesized as 17-mer peptides (12-mer sequence plus 5-mer sequence spacer) and their binding properties were evaluated. The synthetic peptides, Peptide MSal020401 and Peptide MSal020417, were coupled to MyOne™ tosylactivated Dynabeads® and used for MS of Salmonella spp. The number of Salmonellae captured by each bead type was quantified by MS + plate counts and MS + Greenlight™ detection. Under the same MS conditions for each bead type (10-μl bead volume), both methods demonstrated that the two anti-Salmonella peptide-coated beads, MSal020401 peptide-coated beads and MSal020417 peptide-coated beads, were able to capture Salmonella spp. but to varying degrees. The general trend was that Peptide MSal020417-coated beads had much higher capture capability for Salmonella spp. than Peptide MSal020401-coated beads, which also exceeded that of the anti-Salmonella antibody-coated beads at some cell concentrations. Information regarding the commercial antibody-coated Dynabeads®, such as bead concentration, bead size, coating concentration of antibody and number of binding sites per bead, is proprietary. Hence, we were unable to directly compare the antibody-coated beads and the peptide-coated beads (the latter have bead size of 1 μm, bead concentration of 5 × 1011 beads ml−1). In the absence of this information, MS evaluations were made using the same bead volume for each bead type. Bead size and numbers will influence surface area available for binder coating and hence bacterial capture (Tu et al. 2009), so these factors may have influenced the capture capability of the antibody-coated beads.

The larger set of results via MS + Greenlight™ detection provided a more complete picture of capture capability. MS + Greenlight™ detection results showed that both Peptide MSal020417-coated beads and the anti-Salmonella antibody-coated Dynabeads® were capable of capturing Salmonella spp. at as low a concentration as 3 × 101 CFU ml−1 pre-MS demonstrating the detection sensitivity achievable with MS + Greenlight™ detection. The results confirm the successful coupling of the synthetic peptides to the magnetic beads and the successful binding of each peptide-coated bead to Salmonella spp., suggesting the synthetic peptides retained the binding properties demonstrated by the original phage clones. It would appear Peptide MSal020417 retained more of the binding properties of its original phage clone than Peptide MSal020401, given that phage-binding ELISA results indicated that binding to Salmonella spp. was greater for phage clone MSal020401 than for phage clone MSal020417 (Fig. 1b). The binding properties of the synthetic peptide may differ from those of the original phage clone; a synthetic peptide is a lone entity with possibly a different conformation to that expressed on the outside of the phage. The peptides were synthesized with the N-terminal free and the C-terminal fused to a glycine-glycine-glycine-serine-cysteine spacer to mimic the phage clone structure. Differences in binding properties between the synthetic peptide and the original phage clone, as observed for Peptide MSal020401 and Peptide MSal020417, may be attributable to conformational differences.

The variation within the triplicate data set derived from MS + plate counts and MS + Greenlight™ detection was deemed acceptable for microbiological analysis except for one result with Peptide MSal020417-coated beads via MS-plate counts (Fig. 2). Plate spreading was more difficult for the peptide-coated beads post-MS than for the antibody-coated beads and may explain the large variation obtained with Peptide MSal020417-coated beads. The hydrophobic nature of MyOne™ tosylactivated Dynabeads® may cause plate spreading difficulties when the beads are suspended in a hydrophilic environment like nutrient broth. The same high level of variation was not found with Peptide MSal020401-coated beads or the antibody-coated beads perhaps due to the reduced capture capability of Peptide MSal020401-coated beads and the use of a different bead type (proprietary information) for the commercial antibody-coated beads, respectively. Agglutination issues with magnetic beads, including Dynabeads®, have been previously reported (Kretzer et al. 2007). In addition to agglutination issues, publications exist to say plate counts are not an optimal quantification method (Augustin and Carlier 2006; Corry et al. 2007; Jewell et al. 2009). Much research has been carried out on optimal spreading technique, and new ways to enumerate bacteria, to minimize CFU inconsistencies and to maximize viable bacterial counts (Herigstad et al. 2001; Robinson et al. 2004; Thomas et al. 2012) therefore, we speculate that the hydrophobic nature of Peptide MSal020417-coated bead contributed to the MS + plate counts variation observed in this study.

In addition to the generation of a more complete picture of capture capability, the Greenlight™ system is automated, hands-on time is minimal, a simple ‘mix and measure’ protocol is used, it is high-throughput, producing 96 simultaneous results and in real-time with results available within 12–24 h. These attributes have also been reported by Hempel et al. (2011) and Borchert et al. (2012).

Peptide-MSal020417 is a 17-mer peptide with formula weight 1869·01 Da and while information surrounding the commercial bead is proprietary, an antibody has typical formula weight approx. 150 kDa (Janeway et al. 2001). It is a significant finding that a peptide ligand has a similar level of specific binding, if not better, than that of an antibody ligand of which it is approx. 100 times smaller than. Another significant finding is the low % capture under nonselective culture conditions to two other Gram-negative bacteria, Camp. jejuni and E. coli K12, and to L. monocytogenes, a Gram-positive bacterium, as cross-reactivity issues between magnetic beads and nontarget bacteria have been reported previously (Jenikova et al. 2000; Uyttendaele et al. 2000; Vytrasova et al. 2005). This study evaluated capture capability with a cocktail of eight Salmonella spp. covering the majority of important serovars associated with foodborne illness, along with the capture capability with three other foodborne pathogens. Future studies will evaluate capture capability with a range of other Salmonella serovars and other nonpathogenic foodborne bacteria.

Not only were nonselective culture conditions employed in this study, but no selective or nonselective enrichment step preceded MS. As described by Taskila et al. (2012), at least one enrichment cultivation step is generally required for other rapid methods currently available for Salmonella detection. However, it would be interesting to evaluate the use of pre-enrichment and selective media, and their affect on capture capability and sensitivity.

In conclusion, a phage display library was screened against gamma-irradiated whole Salmonella spp. cells to produce Salmonella-specific peptide ligands of which one was demonstrated by MS + Greenlight™ to be a viable antibody alternative for MS of Salmonella spp. In addition, results of the study suggested that the detection sensitivity of peptide-mediated MS (at least 10 salmonellae ml−1) may be sufficient to eliminate the need for pre-enrichment, or at least limit length of enrichment period, thereby speeding up acquisition of results. Phage display-derived peptide ligands can be produced within 3 weeks. Whereas animal-derived antibody production is labour intensive and can take months to complete. This proof of principle study demonstrates the feasibility of using phage display-derived peptide ligands as antibody alternatives. Further studies will help to validate the utility of Peptide MSal020417 for MS and will include capture capability with other Salmonella serovars and other foodborne bacteria, food matrix effect, the use of Salmonella-selective media and testing of artificially spiked foods, with the aim of fully developing a new rapid method for Salmonella detection in food.

Acknowledgements

We would like to thank the Department of Employment and Learning (DEL) for the award of a PhD studentship to Mary Josephine Morton as part of the ASSET project, part of the Strengthening All-island Programme. Thanks are due to Conn Carey, James Hynes and Breda Kearney at Luxcel Biosciences Limited, Cork for training and support in use of the Greenlight™ Model 960. This work was supported by the EU FP7 Incoming Marie Curie Fellowship awarded to Dr Karoonuthaisiri.

Ancillary