September 15, 2010 (Vol. 30, No. 16)
Advanced Platform Finds Applications in Characterizing Antibodies in Complex Samples
Peptides are valuable affinity reagents in proteomics research, useful in generating and characterizing antibodies and in serving as robust probes for protein binding analyses. Array formats are popular because they increase throughput and require only small quantities of reagents.
Label-free analysis of peptide-protein interactions eliminates the risk of introducing experimental artifacts by chemically modifying protein targets with a label, and reduces the expense and difficulty of reproducibly labeling multiple targets for multiplex analysis. This article summarizes the use of label-free peptide arrays to characterize antibodies in egg yolk and serum samples.
The surface plasmon resonance imaging (SPRi) array analysis described in this article requires a gold substrate (chip). A dependable method for making peptide arrays on gold is to synthesize peptides with a terminal cysteine residue, plus a Ser-Gly-Ser-Gly spacer. The thiol group on the cysteine residue is then covalently linked in a specific orientation to the thiol-reactive surface on the chip, while the spacer serves to improve availability of the peptide probe for binding to target proteins.
For peptides lacking cysteines, an amine-reactive surface such as N-Hydroxysuccinimide has been used for array fabrication. Such surfaces are designed to link probes covalently via available primary amines.
Egg Extracts
An array with replicates of four cys-terminated peptides was fabricated on thiol-reactive SpotReady® gold substrates; egg yolk extracts from chickens immunized with peptides were then exposed to the array and binding was monitored in real time using GWC Technologies’ SPRimager®II label-free array reader (Figure 1).
The array was first exposed to egg extracts made just seven days after immunization with peptide A (a segment of the PLA4 protein). As no antibodies had yet been generated, no specific binding was observed, but exposing the array to the egg extract served to block most nonspecific sites.
Next, the array was exposed to egg extract from chickens hyperimmunized with peptide B, a different segment of the PLA4 polypeptide. Significant binding of extracts was observed only to peptide B, confirming the presence of antipeptide B antibodies. The low level of binding to the other three peptides is considered nonspecific and was subtracted from the peptide B binding signals to generate corrected curves (Figure 1A inset).
The peptide array was regenerated, then exposed to egg extract made seven days after a third boost with peptide A. Strong antibody binding was observed to both peptide A probes (Figure 1B, A1 and A2 have the Cys-spacer sequence on opposite ends). The observed association rate for binding to peptide A1 was faster than for A2 (3.1 x 10-2 sec-1 vs 2.2 x 10-2 sec-1) as judged by fitting the data to standard models for simple bimolecular interactions.
Figure 1. Binding of egg extracts to peptide arrays, plotted as reflectivity changes (change%R) over time as monitored by SPRi: Peptide arrays were exposed to egg extract from a chicken immunized with peptide B (A) and peptide A (B). Arrows show time of addition of egg extract or PBS. Curves show average binding signals to replicate peptide spots for A1 and A2 (red and pink), D (blue), and B (green). Insets show net binding to immunizing peptides after subtracting average of signals for controls. Methods: Thiol-reactive surfaces were prepared on SpotReady®16 gold chips by overnight incubation in 1 mM amino octane thiol in absolute ethanol followed by activation in 1 mM succinimidyl 4-[N-maleimidomethyl]cyclohexane-1-carboxylate in PBS pH 7.4. Cysteine-terminated peptides were spotted at 1–4 mg/mL for 60 min. Peptide A1 is from PLA4 (Cys on N-terminus); Peptide A2, same as A1 except Cys on C-terminus; Peptide B is from PLA4; Peptide D is a control from TLR4. For exposure to the array, egg yolks were extracted in acidified PBS (pH 3) and then diluted 1:4,000 with PBS pH 7.4. After exposure to antipeptide B extract, the array was regenerated by washing in 0.1 M glycine pH 2.6, then rinsing with PBS.
Mouse Serum
In this set of experiments, peptides (derived from shiga-toxin), conjugated peptides, and proteins were all spotted on the same array. The array was fabricated on SpotReady substrates with amine-reactive surfaces; binding of sera to the array was monitored in real time by SPRi.
To confirm the integrity of the method, an array of conjugated and unmodified peptides was exposed to purified mAbs in PBS without blocking (Figure 2A). Each of the three mAbs specifically recognized the peptide used as immunogen, as expected. Peptide 149 was arrayed in both pure and ovalbumin-conjugated form, and stronger binding was observed to the conjugated version. This may reflect superior presentation of the antigen to the analyte when the peptide is conjugated to ovalbumin rather than bound directly to the array surface.
An array was prepared with three peptides and a protein, ovalbumin. For serum analysis, the array was first blocked by exposing it to 5% v/v preimmune serum in PBS for 10 minutes on the instrument. A large nonspecific binding signal was observed on all array elements.
In many SPR instruments, strong serum nonspecific binding signals can occupy most of the linear dynamic range, precluding quantitative analysis of subsequent binding events. GWC’s SPRimager II instrument has a broad 40–70° angle adjustment range that enables the detector to be adjusted back into linear range following such strong binding signals.
Following blocking and readjustment of the angle, serum from a mouse immunized with peptide 148 conjugated to ovalbumin showed specific binding to peptide 148, confirming the success of the immunization scheme (Figure 2). The serum also bound to ovalbumin (Figure 2B). Despite the complexity of the serum analyte, the faster association rate for binding to the pure peptide was readily distinguished from the slower rate of binding to the ovalbumin.
Figure 2. Binding of mAbs and peptide-immunized sera to peptide arrays, plotted as reflectivity changes (change%R) over time as monitored by SPRi. (A) Purified mAbs binding to a peptide array: Arrows show time of addition of anti-148 mAb (blue), PBS wash (gray, open arrowheads), anti-149 mAb (red), and anti-162 mAb (green). “C” suffix indicates immobilized peptides conjugated to ovalbumin. Curves show binding of mAbs to peptides 148 (blue), 149 (red), conjugated 149 (pink), and conjugated 162 (green). For the anti-148 and -149 mAbs, antibody was added first at 10 nM then at 100 nM; association rates increased as they should for the higher mAb concentrations. For anti-162 mAb, only 100 nM antibody was added. (B) Sera binding to a peptide array: Arrows show time of addition of nonimmune sera (gray) and serum from a mouse immunized with peptide 148 conjugated to ovalbumin (green). Curves show average binding to replicate negative control peptide 149 spots (red), immunizing peptide 148 (blue), and ovalbumin (pink) on the array. Inset shows array image at the end of the experiment. Bright spots indicate binding has occurred. Methods: Amine-reactive surfaces were prepared by soaking SpotReady 16 substrates overnight in dithiobis[succinimidyl propionate] then rinsing with ethanol and drying before spotting. Peptides and peptide-ovalbumin conjugates were spotted at ~1 mg/mL and incubated for 60 min.
Conclusion
Valuable information on the characteristics of antibodies in complex analytes can be obtained directly using label-free SPRi analysis of peptide arrays. Distinct rates of association of antibodies in serum are readily distinguished, despite the complexity of the analyte.
Elucidation of the distinct binding properties observed for the same peptide immobilized in opposite orientations, and for peptides arrayed in pure versus conjugated form, underscores the value of the SPRi platform for optimizing the fabrication and analysis of arrays. Moreover, both small peptide probes and much larger protein probes may be spotted and analyzed on the same array, a convenience that greatly extends the versatility of SPRi.
Timothy G. Burland, Ph.D. ([email protected]), is president and CEO, and Voula Kodoyianni, Ph.D., is CSO at GWC Technologies.
