Background
Chemotaxis inhibitory protein of
Staphylococcus aureus (CHIPS) is a potent inhibitor of neutrophil chemotaxis and activation by specifically binding and blocking the G-protein coupled Complement fragment C5a receptor (C5aR) and formylated peptide receptor (FPR) [
1,
2]. C5a is a complement polypeptide with many functions. It exerts pro-inflammatory effects through the C5aR and is involved in host defense against microorganisms. The C5a/C5aR interaction mediates immunomodulatory and inflammatory activities such as chemotaxis, degranulation, vascular permeabilisation and cytokine regulation [
3‐
6]. C5a plays a role in a wide variety of inflammatory disorders like rheumatoid arthritis, inflammatory bowel disease, immune complex disease, ischemia-reperfusion injury and sepsis [
7‐
13]. Since C5a is generated early in the inflammatory cascade it is a promising target for anti-inflammatory therapy. Several studies have demonstrated the beneficial effects of targeting the C5aR in inflammatory diseases [
14‐
19].
The potent ability of CHIPS to inhibit activation of the C5aR is an important asset which might be useful for development of a potential new anti-inflammatory drug. However, since the CHIPS gene is present in over 60% of
Staphylococcus aureus strains and
S. aureus is a common bacterium, a majority of the population encounters CHIPS early in life. Human serum contains anti-CHIPS antibodies that have been shown to interfere with CHIPS function
in vitro [
20]. These antibodies may therefore neutralize the
in vivo effect of CHIPS or give rise to antibody mediated immune reactions. As such, the potential of CHIPS to function as an anti-inflammatory molecule is hampered. An improved CHIPS molecule will be characterized by decreased reactivity with pre-existing antibodies, but preserved C5aR blocking activity. Mapping the epitopes of human IgG on the CHIPS protein is an important step in the understanding of how antibodies interfere with CHIPS.
Antibody screening of phage-displayed peptide libraries is a useful approach to identify antibody epitopes [
21,
22]. Previous studies showed the potential of using random peptide phage libraries in identifying linear epitopes and conformational epitopes for monoclonal and polyclonal antibodies. For example, Luzzago
et al. identified discontinuous epitopes in human H-subunit ferritin by the use of phage display and verified the potential epitopes by design of variants with point mutations [
23]. Rowley
et al. studied autoantibodies in primary biliary cirrhosis and could predict the major conformational antibody epitope using phage display and the known NMR structure of the target protein [
24]. These studies show that peptides expressed by phage display are capable of adopting a conformation that mimics the epitopes.
In this paper we focus on mapping epitopes on CHIPS by panning of phages displaying 7-mer peptides against polyclonal human IgG affinity-purified on N- and C-terminally truncated CHIPS. Peptides selected by phage display could be grouped into two distinct groups based on their sequence. These were then verified in ELISA and by the use of a synthetic peptide. By mapping the selected peptide sequences on the CHIPS surface using the known NMR structure of CHIPS
31–121 [
25], potential conformational epitopes for human IgG could be identified. The epitopes were subsequently verified by mutation of key amino acids in the CHIPS molecule and the activity of the mutants was studied by C5aR inhibition. The mutational analysis of the epitopes revealed amino acid residues that are important for the interaction of CHIPS with human antibodies. In addition, amino acids involved in the interaction with the C5aR were identified.
Discussion
Antibodies often react with tertiary structures formed through protein folding. In contrast to linear epitopes, these epitopes are formed when remote amino acids in the primary sequence of a protein are brought into close proximity in the folded protein, hence they are referred to as discontinuous or conformational epitopes. Antibody screening of phage-displayed peptide libraries is an efficient method to identify both linear and conformational epitopes [
21,
22]. Epitopes have been revealed both for monoclonal [
28,
29] and polyclonal antibodies [
24,
30,
31]. In this study, we used two phage-displayed random peptide libraries to map the epitopes for polyclonal human anti-CHIPS ΔN/C IgG on CHIPS. Phage selection depends on many factors, for instance, arginines in the displayed peptide sequence interfere with secretion of the coat protein pIII and consequently clones with peptides containing arginines are strongly selected against [
32]. Furthermore, the stringency and nature of the wash steps can favor certain phages [
33]. To increase the specificity of selected phages for binding to CHIPS IgG, we used competition elution with a high concentration of CHIPS in the last round of selection.
Peptides found to be involved in human anti-CHIPS ΔN/C IgG binding all originated from the phage library which expressed linear peptides, and not from the library with cyclic peptides. We speculate that linear peptides might have been favorable in these selections since they may be able to adopt more conformations than the more rigid cyclic peptides. Mapping of the linear peptides obtained from phage selections with anti-CHIPS ΔN/C IgG identified seven potential epitope surfaces, out of which six represent conformational epitopes. This is in accordance with the result from the pepscan ELISA, where mainly peptides spanning the N-terminal part of CHIPS were recognized by anti-CHIPS IgG. Antibody binding to CHIPS is thus likely to be a combination of interaction with conformational and linear epitopes.
The interaction between an antibody and its epitope is dependent on several amino acid residues [
34]. The nature of the amino acid substituted in a mutagenesis study can affect the ability to detect its involvement in the investigated epitope [
35]. Here, we have substituted residues of the suggested epitopes by amino acids with different characteristics in an effort to detect the epitopes recognized by polyclonal anti-CHIPS ΔN/C IgG. Furthermore, single mutations are likely to affect the binding of only a fraction of all antibodies in a polyclonal preparation. Therefore, the effect of site-directed mutagenesis of a single residue may be difficult to detect by ELISA, as was found in this study. If the mutations mediating an effect on antibody binding were combined in one sequence they may have a synergistic effect to further decrease the antibody interaction with CHIPS.
Nevertheless, we were able to identify amino acid residues important for antibody interaction. By mutagenesis of residues in the different proposed surfaces, we retrieved information on which of the suggested surfaces were likely to be true epitopes. Surface 1.2 and 1.6 are the surfaces affected most by the mutagenesis. For surface 1.2, mutagenesis in three out of three positions resulted in lower interaction with the IgG. N55 in the loop between the α-helix and β1, K100 in the β3 sheet and S107 located in the short loop between β3 and β4 were all shown to be involved in antibody interaction. Surface 1.6 contains two amino acids (K95 and Y97 in the β3 sheet) that upon mutagenesis reduced the antibody binding to less than 50%, respectively. N111, located in a well defined long loop between the β2 and β3 strands, was also shown to contribute to antibody binding. Surface 1.1–1.4 are overlapping structures. Surface 1.1, 1.3 and 1.4 were considered less likely to be true epitopes since mutagenesis of amino acids T53 in surface 1.1 and 1.4, Q58 in 1.3 as well as T53 and K54 in 1.4 did not affect the binding to the IgG. Surface 1.5 was another alternative epitope originating from the group 1 consensus sequence. This was the only linear epitope suggested, but the relevance of this surface was difficult to verify, since two of the five proposed residues (G70 and Y72) were not mutated. G70 and Y72 were considered important for maintaining the structure and were not mutated for stability reasons. Surface 2.1 was the only mapping possibility found for the group 2 consensus sequence. We consider this to be a less likely epitope, since mutagenesis of its amino acids did not reduce the antibody binding to a similar extent as that observed for 1.2 or 1.6.
During the mutational analysis for verification of antibody epitopes, amino acids involved in the interaction with the C5aR were also identified. The β3 sheet is comprised of amino acids Y94-K100 and is most likely critical for binding to the C5aR since we show that substitutions K95S, Y97K and K100A as well as S107N affect the C5aR inhibiting activity of CHIPS, while leaving the CHIPS structure intact. These data confirm previous findings where the substitution K95A was implicated in C5aR blocking activity [
25]. Some of the amino acid residues found to be important for the antibody interaction are therefore also most likely important for the receptor binding properties of CHIPS. This corroborates data from Wright
et al. who recently showed that CHIPS-specific antibodies from human serum could inhibit the binding and activity of CHIPS to the C5aR [
20].
The data about which amino acids are involved in C5aR binding is a finding that may be utilized in the design of an optimal CHIPS molecule that requires retained C5aR binding and blocking in addition to minimal binding of human antibodies. Our study suggests residues that must be left untouched or substituted in a more conservative manner. For certain applications, there may also be other aspects to consider, such as analysis of the presence of T-cell epitopes. Such a process can be added in future studies when designing new CHIPS variants.
Methods
Cloning, expression and purification of recombinant proteins
Full-length CHIPS
1–121 and CHIPS with N-terminal truncation, composed of amino acids 31–121 (CHIPS ΔN) and CHIPS with N- and C-terminal truncations, composed of amino acids 31–113 (CHIPS ΔN/C) were cloned, expressed and purified as described previously [
27]. CHIPS with C-terminal truncation (CHIPS ΔC) and variants thereof were cloned in a modified pRSET B vector (Invitrogen, Carlsbad, CA). As a result of cloning, this protein was 112 amino acids long with two additional non-relevant amino acids included in the C-terminal end of the expressed protein. Cultivation was performed as described by Haas
et al [
27], but the protein was purified from inclusion bodies. Briefly, the CHIPS variants were purified from
E. coli culture by lysis of the bacteria in 1/10 culture volume binding buffer (8 M urea, 20 mM Sodium Phosphate, 500 mM NaCl, pH 7.8). After sonication and filtration, the lysate was loaded onto a HisTrap Ni-column (GE Healthcare, Uppsala, Sweden). The column was washed with binding buffer, pH 7.8, pH 6.0 and pH 5.6, respectively. CHIPS was eluted with binding buffer, pH 5.6 with 50 mM EDTA and fractions containing protein as measured by A
280 were pooled and refolded by dialysis against 50 mM Tris-HCl, pH 8.0. The His-tag was removed as described [
27].
Site-directed mutagenesis and expression of CHIPS variants in plate format
To be able to compare many mutants of truncated CHIPS, a tag is required to capture the proteins and determine relative concentrations. However, CHIPS activity has proven to be sensitive to the addition of His-tags in both the N- and C-terminus. These have to be cleaved off before functional assays can be performed. Instead, we used a mAb (2H7) recognition site located between residue 15 and 30 within the native CHIPS N-terminus that does not interfere with C5aR blocking activity [
27]. Therefore the mutations were generated in CHIPS with C-terminal truncation, leaving the N-terminal 30 amino acids untouched for recognition by the 2H7 mAb in ELISA. Site-directed mutagenesis was performed in order to introduce point mutations into the CHIPS sequence with C-terminal truncation. This was performed using the QuikChange II mutagenesis kit (Stratagene, La Jolla, CA) according to the manufacturer's recommendations. The new CHIPS variants were sequence verified and transformed into
E. coli BL-21 Star(DE3)pLysS (Invitrogen, Carlsbad, CA) for protein expression. Variants of CHIPS with C-terminal truncation were expressed in LB medium containing 50 μg/ml ampicillin and 34 μg/ml chloramphenicol in deep-well plates. Overnight cultures were diluted 1/50 in fresh LB containing 50 μg/ml ampicillin and cultured at 37°C for 3 hours. IPTG (Isopropyl β-D-Thiogalactoside) (final concentration 0.5 mM) (BDH, Poole, UK) was added to log-phase cultures to induce protein expression followed by cultivation for three hours. Protein was prepared from
E. coli lysates by freeze-thawing the
E. coli pellet in a buffer consisting of PBS-0.05% Tween-20, Complete EDTA-free protease inhibitor (Roche, Basel, Switzerland), 25 U/ml Benzonase (Sigma-Aldrich, St Louis, MO) and 1 KU/ml rLysozyme (EMD Chemicals, Darmstadt, Germany). The lysates were incubated for 30 min at room temperature with shaking and centrifuged at 13,000 rpm for 15 min to remove debris and were stored at -20°C.
Affinity purification of anti-peptide IgG and anti-CHIPS IgG
A peptide, comprising the phage derived sequence MNKSYTI, was synthesized with an additional C-terminal spacer of three glycines and a cysteine for efficient coupling. (Bio-synthesis; Lewisville TX). A 38-mer peptide comprising the N-terminal part of CHIPS (first 37 amino acids plus an additional cysteine (Pepscan, Lelystad, The Netherlands) served as control. In order to couple the peptides to a solid matrix, peptides were first reduced using agarose linked Tris(2-Carboxyethyl) Phosphine (TCEP) (Pierce, Rockford, IL) and subsequently mixed with Sulfo-Link agarose beads (Pierce, Rockford, IL) in 50 mM Tris/HCl, pH 8,3 with 5 mM EDTA for 2 hours at room temperature. Unreacted groups were blocked with L-cysteine and beads were extensively washed with coupling buffer and PBS. Small 1 ml columns were used for affinity purification of IgG from a pooled human immunoglobulin preparation for intravenous use (IV-IgG) (Sanquin, Amsterdam, The Netherlands). IgG was eluted with 0.1 M glycine-HCl buffer pH 2.9 and neutralized with 1 M Tris pH 8.0. Pooled fractions were dialyzed overnight against PBS and the purified anti-peptide IgG was stored at 4°C.
Purified CHIPS variants (full-length or CHIPS ΔN/C) were coupled to CNBr activated sepharose 4B (Amersham Biosciences, Uppsala, Sweden) and packed on a column according to manufacturer's instructions. Affinity purification was performed on an ΔKTA Prime system (Amersham Biosciences) according to the manufacturer's protocol. A total of 1 g IV-IgG (20 mg/ml) over the column. Bound human IgG was eluted with 0.1 M Glycine pH 3.0 and the pH neutralized with 1 M Tris, pH 8.0. Eluted fractions containing protein were pooled and buffer was changed to PBS on PD-10 columns (Amersham Biosciences). The affinity-purified human anti-CHIPS IgG was stored at 4°C.
ELISA
ELISA was used for many applications throughout the study. Nunc Maxisorb clear or white 96 well plates were coated with the specific protein in PBS. Incubations were carried out in a volume of 50 μl for 1 hour at RT of not described differently, always followed by washing trice with PBS-0.05%Tween-20. Dilution buffer for all reagents was PBS-0.05%Tween-20 with 1% BSA or milk powder added.
Anti-CHIPS and Pepscan ELISA
Microtiter plates were coated with CHIPS (1 μg/ml) overnight at 4°C. Plates were blocked with PBS-0.05%Tween-20 4% BSA and incubated with different anti-CHIPS antibodies in dilution buffer with 1% BSA. Bound antibodies were detected with goat-α-human-IgG conjugated with peroxidase (Southern, Birmingham, USA) and TMB as substrate. For peptide experiments, synthetic 25-mer CHIPS derived peptides spanning the complete CHIPS
1–121 sequence were synthesized as described earlier [
27]. Plates were coated with peptide (10 μM) overnight at 4°C and ELISA was performed as described above with IgG purified on full-length CHIPS
1–121.
Binding of human anti-CHIPS ΔN/C IgG to CHIPS variants with C-terminal truncation
For quantification of expressed proteins, plates were coated overnight at 4°C with 3 μg/ml monoclonal anti-CHIPS Ab 2H7 directed against a peptide of CHIPS amino acids 24–30 [
27]. Plates were blocked in PBS-0.05% Tween-20 with 3% milk powder, washed and incubated with serial dilutions of lysates from CHIPS variants in dilution buffer with 1% milk powder. Binding was detected with 3 μg/ml polyclonal rabbit anti-CHIPS N-terminal IgG (IgG produced by immunization of a rabbit with a KLH-coupled synthetic peptide corresponding to CHIPS N-terminal amino acids 1–14) and goat anti-rabbit IgG-HRP (Southern Biotech, Birmingham, AL)). Super Signal ELISA Pico Chemiluminescent Substrate (Pierce, Rockford, IL) was added and luminescence was measured.
For detection of binding of human anti-CHIPS IgG to the CHIPS variants, plates were coated, blocked and incubated with E. coli lysates as described above. Affinity-purified human anti-CHIPS ΔN/C IgG was added and binding was detected with goat-anti-human IgG HRP (Jackson ImmunoResearch, West Grove, PA) and substrate as described above.
Random peptide phage libraries and phage selection
The Ph.D.-7™ and Ph.D.-C7C™ libraries from New England Biolabs (Ipswich, MA) were used to map the epitopes for human IgG on the surface of the CHIPS protein.
The Ph.D.-7™ Phage Display Peptide library consists of 7-mer random linear peptides fused with a linker sequence (GGGS) to the N-terminus of the major coat protein pIII of bacteriophage M13. The randomized segment of the Ph.D.-C7C™ library is flanked by a pair of cysteine residues, which are oxidized during phage assembly to a disulfide linkage, resulting in the displayed peptides being presented to the target as loops [
36]. Protein G coated beads (100 μl) (Dynal, Norway) were washed three times with 1 ml PBS-0.05% Tween-20. The washed beads were blocked with 1 ml PBS-0.05% Tween-20, 5% BSA for 1 h at 22°C. Beads were washed four times and resuspended in 1 ml PBS-0.05% Tween-20. One half of the blocked beads was used for preclearing the phage stock. Therefore, 10 μl Ph.D.-7™ and 10 μl Ph.D.-C7C™ were incubated with blocked beads in PBS-0.05% Tween-20 and were incubated for 30 min at 22°C under continuous agitation. Affinity-purified human anti-CHIPS ΔN/C IgG (final concentration ~10 nM) was added to the precleared phages and incubated at 22°C for 30 min. The phage/IgG suspension was added to the remaining blocked beads and incubated at 22°C for 30 min. The beads were washed 10 times with PBS-0.05% Tween-20 to wash away unbound phages. The bound phages were eluted with 0.2 M Glycine, pH 2.2, 0.1% BSA for 8 min followed by neutralization of the pH with 1 M Tris-HCl, pH 8. The eluate was amplified and 10 μl of amplified phages was used as input for the next selection round. To increase the specificity of the phage selection, the bound phages in the fourth round were eluted using competition elution with the CHIPS protein. Bound phages were eluted by overnight incubation with 1.8 mg/ml CHIPS.
Phage titration, amplification and characterization
LB medium was inoculated with a single colony of ER2738 E. coli and incubated at 37°C with vigorous shaking until mid-log phase (OD600 ~0.5). Top agar (50% LB agar, 50% LB medium) was melted and cooled to approximately 45°C. Melted top agar (3 ml) was added to 200 μl ER2738 E. coli and poured on top of a LB/0.5 mM IPTG/80 μg/ml X-gal plate. 10 μl of 10-fold dilutions in LB medium was spotted on the prepared culture plates and incubated overnight at 37°C. The next day, plaques were counted in order to calculate phage titers. The remaining phage eluate was added to ER2738 E. coli culture at early log phase (OD600 0.4–0.5) and incubated with vigorous shaking at 37°C for 4.5 h. Cultured cells were pelleted at 4°C and phages were precipitated overnight at 4°C in 25% PEG6000 (Sigma-Aldrich, St Louis, MO) 3 M NaCl. The precipitated phages were centrifuged for 15 min at 10,000 rpm, 4°C. The pellet containing the amplified phages was resuspended in 200 μl PBS and titrated as described above. After the fourth selection round no phage amplification was performed and phages were directly characterized by DNA sequencing.
48 different plaques from the titration plates were stabbed with a pipette tip and used to infect 1 ml 1/100 diluted overnight culture of ER2738 E. coli and incubated for 4.5~5 h at 37°C. Cultures were centrifuged and 500 μl of the supernatant was precipitated with PEG6000, 3 M NaCl for 10 min at 22°C. The samples were centrifuged for 10 min at 13,600 rpm and the pellet was resuspended in 100 μl 4 M NaI, 10 mM EDTA, pH 8. 250 μl 95% EtOH was added and the samples were incubated for 10 min at 22°C to preferentially precipitate the single-stranded phage DNA. Samples were centrifuged for 10 min at 13,600 rpm and the pellet was washed with 70% EtOH, dried and sent for sequencing to MWG Biotech (Martinsried, Germany) using the PIII-96seq primer (New England Biolabs, Ipswich, MA).
A phage ELISA was performed to test the binding specificity of the selected phages for affinity-purified human anti-CHIPS ΔN/C IgG. Affinity-purified human anti-CHIPS ΔN/C IgG was coated overnight with at 4°C. Coated and non-coated wells were blocked with PBS-0.05% Tween-20, 5% BSA. Serial dilutions of the purified phage stocks in dilution buffer with 1% BSA were added and incubated for 1 h at 37°C. Detection was performed with mouse anti-M13-mAb (1 μg/ml) (Amersham Biosciences, Uppsala, Sweden) and rabbit anti-mouse IgG-HRP (Dako A/S, Glostrup, Denmark) and OPD substrate (O-phenylenediamine) (Sigma Aldrich, St Louis, MO).
Analysis of peptide antibody binding to CHIPS
Binding of affinity-purified antibodies and pooled human IgG (IV-IgG) to CHIPS was studied by ELISA as described for anti-CHIPS ΔN/C antibodies and by SPR on a Biacore 1000 instrument. 20 μl 1 mg/ml CHIPS was directly coupled to an activated N-ethyl-N'(di-methylaminopropyl)carbodiimide (EDC) and N-hydroxysuccinimide (NHS) carboxymethyl dextran sensor chip (CM5). Unreacted groups were blocked by injection of 50 μl 1 M ethanolamine-HCl pH 8.5. Binding assays were performed at a constant flow rate of 5 μl/min at 25°C. Affinity-purified antibodies and IV-IgG were diluted in HBS-EP buffer (10 mM HEPES (pH 7.4) containing 150 mM NaCl, 3 mM EDTA and 0.005% surfactant P20). Antibodies were allowed to interact with CHIPS for 210 s followed by a 120 s dissociation phase. Additionally, the antibodies were preincubated with 1 mg/ml CHIPS protein to study competition. Affinity-purified anti-peptide antibodies were tested at a concentration of 10 μg/ml. Residual bound antibody was removed from the sensor chip surface by washing the chip for three minutes with 10 mM glycine-HCl (pH 1.5).
Epitope mapping
The amino acid sequences of the selected phages were aligned using the Clustal-W alignment tool
http://www.ebi.ac.uk/clustalw/. Consensus sequences were manually mapped onto the surface of the CHIPS protein using the CHIPS
31–121 NMR structure (PDB code: 1XEE) [
25] and the PyMOL molecular graphics system [
37].
Biological activity assay
The C5aR inhibiting capacity of the purified CHIPS variants with C-terminal truncation was tested by measuring calcium release from C5a stimulated U937 cells stably transfected with the C5aR [
1]. U937 cells (human promonocytic cell line) transfected with the C5aR (U937/C5aR) were a generous gift from Dr. E. Prossnitz (University of New Mexico, Albuquerque, NM). Cells were grown in 75 cm
2 cell culture flasks in a 5% CO
2 incubator at 37°C and were maintained in RPMI 1640 medium with L-glutamine (Cambrex, Verviers, Belgium) and 10% FBS (Cambrex, Verviers, Belgium). Briefly, 5 × 10
6/ml U937/C5aR cells were incubated with 2 μM Fluo-3AM (Sigma Aldrich, St Louis, MO) in RPMI 1640 medium with 0.05% BSA for 30 min at room temperature. After washing, the cells were preincubated with buffer or a 3-fold dilution series of CHIPS variants (1 to 0.003 μg/ml) at room temperature for 30 min. Every sample was measured on a FACSsort flow cytometer (BD Biosciences, San José, CA) and C5a (Sigma Aldrich, St Louis, MO) (final concentration 1 nM) was added. The cell population was gated on forward and side scatter and results are expressed as percentage inhibition of calcium release as compared to cells without addition of CHIPS.
CD spectroscopy
The structures of the purified C-terminally truncated CHIPS variants were compared by the use of CD spectroscopy. The experiment was carried out on a Jasco J-720 spectropolarimeter (Jasco Inc., Easton, MD) in a 2 mm cuvette at a protein concentration of ~20 μM. Spectra were recorded from 195 to 250 nm at 20°C, at a scan speed 50 nm/min, a time constant of 4 s, a bandwidth of 1 nm, resolution of 1 nm and sensitivity of 50 mdeg.
Data analysis
Data from ELISA and cellular experiments were plotted by the use of the Prism 4 software package (GraphPad Software Inc.). The cellular data were fitted in a nonlinear regression model (sigmoidal dose-response curve with variable slope).
Authors' contributions
EG carried out the mutational analysis, functional assays and structural comparisons, participated in the design of the study and purification of antibodies and drafted the manuscript. PJH carried out the phage selections and pepscan experiments and participated in the design of the study and the epitope mapping and drafted the manuscript. BW carried out the epitope mapping and participated in the design of the study and helped to draft the manuscript. MH carried out the SPR measurements. CF participated in the design of the study and helped to draft the manuscript. MO participated in the design of the study and helped to draft the manuscript. JvS participated in the design of the study. KvK participated in the design of the study, purification of antibodies and draft of the manuscript. All authors read and approved the final manuscript.