KW, KS and LHS declare that they have no competing interest. SAJ holds the patent to the immunsignature and related technologies and has equity in a company related to these technologies. This does not alter the authors’ adherence to PLOS ONE policies on sharing data and materials.
Conceived and designed the experiments: LHS SAJ KS. Performed the experiments: KW LHS. Analyzed the data: KW LHS KS. Contributed reagents/materials/analysis tools: KW LHS. Wrote the paper: KW SAJ KS LHS.
Antigen-antibody complexes are central players in an effective immune response. However, finding those interactions relevant to a particular disease state can be arduous. Nonetheless many paths to discovery have been explored since deciphering these interactions can greatly facilitate the development of new diagnostics, therapeutics, and vaccines.
Characterizing the interactions between disease-specific antibodies and their cognate antigens has proven highly informative in the study of host-pathogen relationships and critical in the development of effective biomedical products. The discovery of modified antigens or autoantigens that are specifically recognized by patient antibodies is of growing importance in disease research and target development for diagnostics, vaccines, and therapeutics. These complexes are typically found by querying immune sera against possible ligands in lysates, or in libraries of proteins or peptides made recombinantly or synthetically. Myriad of binding assays such as immunoblots [
In its original description, phage display was used to survey a library of peptides for binding to a given antibody[
In vitro combinatorial synthesis of peptides on beads and microarrays of either proteins or peptides have been explored as library formats for surveying target binding [
However most of the applications of the above methods focused on known biological targets. This limited their applications. Each high throughput library can only cover limited proteome. On the other side, the antigens or epitopes within the antigen may not be in the known proteome. Studies can be confounded by the fact that immune sera often carry antibodies to mutant, unknown, or even exogenously-derived antigens of a host or pathogen proteome.
Random peptide libraries, whether in vivo or in vitro, provide a means for identifying mimotopes of unknown antigens and are far more economical since one library can be used for all screens. For example, an antibody panned against phages displaying random sequence peptide fusions might select recombinants that mimic or hold similarity to a previously unknown ligand [
Recently, the immunosignatures technology has been developed as a universal platform to profile complex antibody mixtures [
Here we employ the IMS as a universal platform to design a method for identifying mimics of disease-relevant epitopes and then show that these mimotopes can capture the disease-relevant antibodies for the further investigation, such as identification of the unknown biological antigens.
Unique from other platforms that focused on analyzing the purified targets [
A number of free peptides were synthesized by Sigma, Inc. Beads for antibody affinity purification were prepared in the lab. Peptides were directly synthesized on the TentaGel Beads. The sequences of these peptides and their assigned names are presented in
Assigned Name | Sequence | Description |
---|---|---|
SMCfs | CCGIYCHEEPQREDSSI | human SMC1A frameshift 17mer peptide |
SMCfs-27 | TAIIGPNGSGCCGIYCHEEPQREDSSI | human SMC1A frameshift 27mer peptide |
RP1 | TISKYVMVEPMRQHEEWGSC | random peptide SMCfs mimotope |
RP2 | AVSHQEMNEGEQGPMREGSC | random peptide SMCfs mimotope |
RP3 | RVGEMPMREYDISGGSGGSC | random peptide SMCfs mimotope |
RP4 | TAFYRTLTKHEVDPGIAGSC | random peptide SMCfs mimotope |
CP1 | AVLLMCQLYQPWMCKEYYRLL | negative control peptide which is not a mimotope of SMCfs |
SMCfs-B | CCGIYCHEEPQREDSSI | human SMC1A frameshift 17mer peptide conjugated to Tentagel beads |
RP1-B | TISKYVMVEPMRQHEEWGSC | random peptide SMCfs mimotope conjugated to Tentagel beads |
RP2-B | AVSHQEMNEGEQGPMREGSC | random peptide SMCfs mimotope conjugated to Tentagel beads |
RP3-B | RVGEMPMREYDISGGSGGSC | random peptide SMCfs mimotope conjugated to Tentagel beads |
RP4-B | TAFYRTLTKHEVDPGIAGSC | random peptide SMCfs mimotope conjugated to Tentagel beads |
CP2-B | ATKAAIPGPNTVPRAP | negative control peptide which is not a mimotope of SMCfs conjugated to Tentagel beads |
Anti-human SMC1Afs serum was generated by Global Peptide Service LLC (Fort Collins, CO). The 17 amino acid (a.a.) SMC1A frame shift mutant sequence (SMCfs), identified in human tumor cDNA (CCGIYCHEEPQREDSSI), was synthesized by the peptide synthesis lab at Arizona State University and then conjugated to keyhole limpet hemocyanin (KLH) by Global Peptide Service LLC.A New Zealand white rabbit was immunized with the SMCfs-KLH conjugate. Blood for the experiment was collected at exsanguinations after two immunizations.
ELISA plates were coated with 50 μL of 10 μg/mL of peptide or protein in carbonate coating buffer and incubated at 4°C overnight. The coated plates were washed 3X with PBST and blocked with 200 μL of 3% BSA in PBST at 37°C for 30 minutes. The blocked plate was washed 3X with PBST and 50 μL of primary anti-serum or purified antibody diluted in 3% BSA in PBST was applied. The plate was then incubated at 37°C for 1 hr. After the incubation, the plate was washed 3X with PBST. The antibody was detected with 50 μL HRP-goat anti-rabbit IgG diluted 1:2000 in 3% BSA in PBST. After the plate was incubated at 37°C for 1 hr, the plate was washed 3X and developed with 50 μL TMB for 10 minutes at room temperature. The development was stopped by adding 50 μL of 0.5 N HCl, and the plate was read with a SpectraMax 190 Molecular Devices instrument at OD 450 nm.
Specific antibodies were absorbed from the rabbit anti-SMCfs sera by applying the sera to the SMCfs coated plate. The rabbit sera was diluted 1:250 with 3% BSA in PBST and incubated with the SMCfs peptide coated plate at 37°C for 1 hr. The unbound antibody in the supernatant was then removed and applied to another peptide coated well to remove more antibody specific for the peptide. This process was repeated up to 20 times, and this serum was then later applied to the peptide microarray at a dilution of 1:500. This same method was used to produce negative control antibody absorbed sera using the negative control CP1 peptide.
The 10,000 non-natural randomly generated peptide sequences were custom synthesized by Sigma, Inc. These 20-mers were designed with 17 non-natural sequence a.a. residues (excluding cysteine) and a 3 amino acid (GSC) linker sequence on the C terminus. The C terminal cysteine binds to a sulfo-SMCC coated aminosilane glass slide. The solutions of different non-natural sequence peptides were printed onto the glass slide using a Nanoprint 60 instrument (Array It Technologies).
Rabbit sera samples were applied to the non-natural sequence peptide microarray using a Tecan HS 4800 Pro microarray hybridization station. Slides were first washed for 30s with TBST, and then blocked with a blocking buffer consisting of BSA, mercaptohexanol, Tween 20, and PBS for 1 hr at 23°C. Duplicate samples of sera were diluted 1:500 in an incubation buffer consisting of BSA, Tween 20, and PBS and incubated with the slide for 1 hr at 37°C. The slide was then washed, and 5 nM of goat anti-rabbit IgG conjugated with Alexa Fluor 647 dye was applied for 1 hour at 37°C. The slide was then washed and dried for 5 min.
The slides were scanned with an Agilent Technologies DNA Microarray Scanner with Surescan High-Resolution Technology instrument and analyzed with GenePix Pro 6.0 software to determine the fluorescence intensity of each spot. GeneSpring GX 7.3, Microsoft Excel, simple custom Java code, and GraphPad Prism 4 were then used to perform further analysis of this data.
Total IgG of the rabbit anti-SMCfs sera was purified using Pierce Protein A/G Agarose beads according to the protocol of the manufacturer. 1 ml of the TentaGel beads with specific peptides were embedded in a column. 3 ml of the purified total IgG was added to the column and incubated with the beads for 1 hr at room temperature. The column was then washed with 10 ml PBST. The specific IgG was eluted with 4 fractions of 1 ml IgG Elution Buffer (Pierce, Inc) and each fraction was neutralized with 100 ul 1M TRIS (pH8.0). All of the eluted fractions were measured at an absorbance of 280nm. The two fractions with the highest absorbance were combined and used for further analysis at a 1 to 40 dilution in 3% BSA in PBST.
GLAM2 software [
To identify non-natural sequences recognized by the immune serum, both naïve and SMCfs-immune rabbit sera were applied to the array. We identified a total of 108 non-natural sequence peptides which exhibited high specificity to the immune serum relative to the naïve rabbit control serum based on the t-test (p<0.0001)(
Polyclonal rabbit serum generated against the SMCfs peptide conjugated to KLH and naive rabbit serum were applied to an array of 10,000 synthetic peptides of randomly generated sequence. Student’s t-test analyses comparing the probing results identified 108 non-natural sequence 20-mers with differential binding to the immune versus naïve sera (p-values <0.0001). These differences in peptide-binding intensities are visually represented as a heatmap.
Since the KLH is a large and highly immunogenic antigen, the rabbit anti-serum contained high titer antibodies to the KLH protein. Mimotopes to KLH among the 108 SMCfs peptides were investigated (
Compared to KLH, the SMCfs peptide is much smaller and lower immunogenic. To distinguish the mimotopes of the SMCfs peptide from that of KLH, we specifically depleted the anti-SMCfs antibodies from the immune serum through iterative antibody absorption steps. ELISA analysis confirmed that anti-SMCfs antibodies were partially depleted from the original serum after 11 rounds of absorption and further depleted after 20 rounds (
The anti-SMCfs antibody-depleted samples, as well as the negative control serum, were applied to the CIM10K array. Reactivity profiles of these samples against the 108 peptides selected in the original screen were compared (
A high score indicates that the given random peptide binds to antibodies with high specificity for the SMCfs sequence, and a low score indicates the opposite. Smax is the maximum signal intensity of the peptide with unabsorbed SMCfs sera. S20X is the intensity of the peptide with sera after SMCfs binding antibodies have been absorbed twenty times. SCP is the intensity of the peptide with sera after CP1 binding antibodies have been absorbed twenty times. The CP1 peptide is a negative control peptide with no sequence similarity to SMCfs.
To assess the specificity of anti-serum binding to the four putative mimotopes, we first analyzed the reactivity of the selected peptides to the immune serum by ELISA (
To test the possibility that RP1-RP4 are mimotopes of a KLH protein epitope rather than SMCfs, we measured the reactivity of the four selected random peptides, the original SMCfs peptide, and CP1 to the polyclonal KLH immune serum (
To determine whether RP1 to RP4 can be used to selectively purify antibodies specific to the original immunogen, we prepared affinity purification columns with each of four random peptides (RP1 to RP4), SMCfs, and an irrelevant peptide as negative control (CP2-B). The binding specificities of the affinity-purified antibodies were measured by determining the activity of the eluted antibodies against SMCfs-27 or KLH antigen (
The SMCfs peptide, four highly array-reactive peptides, and a control peptide (CP2-B) were synthesized on Tenta-gel beads. These were used to prepare individual affinity columns. Purified total IgG of the anti-SMCfs serum was applied; bound antibodies were eluted. These peptide-purified antibody samples were analyzed by ELISA against SMCfs-27 coated or KLH coated plates. Error bar indicates SD of the duplication.
The antibody samples that were affinity-purified from the SMCfs immune serum with the SMCfs peptide or with the array-selected mimotopes display more than one cross-reactivity group as measured by ELISA (
The antibodies affinity purified by SMCfs-B or four library-selected peptide bound to beads (RP1-B through RP4-B) were measured for reactivity against the SMCfs and mimotope peptides by ELISA and displayed in a three-dimensional bar graph.
From the mimotpe identification to the validation, we did not use the sequence information of the original SMCfs peptide. To test the possibility of using the sequence analysis of the mimotopes to identify epitope of the unknown antigen, we did the sequence analysis of both the SMCfs and RP1 to RP4 peptides. The epitope search software Bepipred was used to predict strong B cell epitopes within the unique SMC1A frameshifted sequence. The highest amino acid score was assigned to the proline within HEEPQRE. The GLAM2 software was used to identify sequences in RP1 through RP4 with similarity to sequence stretches found in SMCfs. The RP1 peptide contains two motifs: HEE and YXXXXPMRQ, although they are in reverse order relative to those sequences of SMCfs. Note that previous experiments have also identified high binding peptides with inverted sequences relative to the original epitope, but these inversions typically involved just two amino acids [
Sequence ID | Sequence | Motif in Peptide Matching with SMCfs | Matching SMCfs sequence |
---|---|---|---|
SMCfs | CCGIYCHEEPQREDSSI | - | - |
RP1 | TISKYVMVEPMRQHEEWGSC | YXXXXPMRQ | YXXXXPQRE |
PMRQ | PQRE | ||
HEE | HEE | ||
RP2 | AVSHQEMNEGEQGPMREGSC | PMREGS | PQREDS |
PMRE | PQRE | ||
RP3 | RVGEMPMREYDISGGSGGSC | EMPMRE | EEPQRE |
PMRE | PQRE | ||
RP4 | TAFYRTLTKHEVDPGIAGSC | HE | HE |
Note that the RP1-4 peptides which have been the focus of this paper were not the only peptides among the 108 selected peptides which exhibited good mimotope matches of the SMCfs peptide (
There was a good correspondence between GLAM2 scores for RP1-4 aligned with the SMCfs peptide and scores from the peptide microarray. The GLAM2-assigned similarity scores of the mimotopes to SMCfs led to a ranking of the top 4 peptides in the same order that they were ranked by the array absorption scoring described in section 3.2. For example, the RP1 peptide holds two motifs in common with the original antigen, exhibits the highest intensity on the peptide microarray, and scored the highest by the GLAM2 software. By contrast, the absorbance intensities obtained in the ELISA experiment did not yield rankings identical with the other readouts. This suggests that the array format may enable higher resolution detection of antibody-peptide interactions. Taking together, the sequence analysis of the specific array peptides showed the advantage of our high density random peptide array for identifying the epitopes of unknown antigen.
The techniques demonstrated here could be inexpensively and generally applied for both mimotope and true-epitope discovery. We have demonstrated that non-natural sequence peptide arrays can be used to identify mimotopes of specific antibodies without knowing the sequence information of the primary epitope. This high density pepetide array platform is sensitive enough to identify at least two different specific antibodies to a simple peptide from the highly reactive antibodies which recognize KLH, a highly immunogenic antigen. The sequence analysis of the identified mimotopes can help identify the true B cell epitopes of an immunogen. Furthermore, the selected mimotopes can be used to purify this specific polyclonal antibody from the complex milieu of a serum sample for further analysis. In addition to be being a universal platform for the epitope exploration of any antibody, the non-natural sequence peptide array requires no assumptions or information relative to the true epitope of an immunogen, and therefore even mimotopes of newly mutated antigens can be discovered. As the SMCfs peptide was not an annotated, but frame-shift antigen, this method may allow discovery of the immunogenic epitopes of non-annotated antigens, such as novel mutations in tumor cells.
We suggest that mimotopes can be used to identify medically important antibodies and their natural targets. Additionally, mimotopes can be used as alternatives to a true epitope if they are found to carry advantageous properties. These may include greater stability, simpler production, or differential affinity relative to an autoantigen. These are all important attributes for successful drug or therapeutic development.
The intensities of all 10,000 peptides are available upon request.
(DOCX)
(DOCX)
The intensity of all 10,000 peptides are available upon request.
(DOCX)
(DOCX)
We appreciate Dr. Bart Legutki for tips and advice on applying sera to the peptide array and for providing us mouse anti-KLH serum for an experiment; Dr. Phillip Stafford for suggestions on array analysis; Dr. Zhan-gong Zhao for advice and help with peptide synthesis and Tentagel beads.
Bovine serum albumin
Horse radish peroxidase
Keyhole limpet hemocyanin
Phosphate buffered saline
Phosphate buffered saline Tween-20
Structural maintenance chromosome 1A
Structural maintenance chromosome 1A frame-shift
A 17 amino acid SMC1A frameshift mutant
Tris Buffered Saline Tween-20
3,3’,5,5’-Tetramethylbenzidine
tris (hydroxymethyl) aminomethane