bioRxiv preprint doi: https://doi.org/10.1101/2020.08.05.238105; this version posted November 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Epitomic analysis of the specificity of conformation-dependent, anti-Aß amyloid monoclonal antibodies. Jorge Mauricio Reyes-Ruiz1, Rie Nakajima2, Ibtisam Baghallab3, Luki Goldschmidt4, Justyna Sosna1, Phuong Nguyen Mai Ho1, Taha Kumosani3, Philip L. Felgner2, and Charles Glabe1* 1Department of Molecular Biology and Biochemistry, University of California, Irvine, Irvine, CA 29697- 3900 3900 2Vaccine Research and Development Center, Department of Physiology and Biophysics, University of California Irvine, Irvine, California, USA. 3Biochemistry Department, Faculty of Science, King Abdulaziz University, Jeddah - Saudi Arabia 4Institute for Protein Design, Department of Biochemistry, University of Washington, NanoES 396, Seattle, WA 98195-7370 *Corresponding author: Charles Glabe Email: [email protected] Running title: Epitomic analysis of anti-Aß antibodies Keywords: Bioinformatics, Alzheimer Disease, Amyloid antibodies, Discontinuous epitopes, Epitope mapping, Phage display, Immunotherapy 1 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.05.238105; this version posted November 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Abstract arrangement, the segments of the sequence that form intermolecularly hydrogen bonded ß-sheets Antibodies against Aß amyloid are indispensable and the locations where the sheets fold and how the research tools and potential therapeutics for sheets stack together. Many proteins are able to Alzheimer’s Disease, but display several unusual form amyloids upon unfolding or misfolding and properties, such as specificity for aggregated forms are frequently associated with neurodegenerative of the peptide, ability to distinguish polymorphic diseases such as Alzheimer’s disease (AD), aggregate structures and ability to recognize Parkinson’s disease, amyotrophic lateral sclerosis generic aggregation-related epitopes formed by and prion diseases (3). Alzheimer’s disease unrelated amyloid sequences. Understanding the contains two canonical amyloids: Aß amyloid, mechanisms underlying these unusual properties of derived from APP and tau amyloid and in anti-amyloid antibodies and the structures of their approximately 30% of AD, cortical Lewy bodies corresponding epitopes is crucial for the are also present containing α-synuclein amyloid. understanding why antibodies display different Monoclonal antibodies against Aß are a leading therapeutic activities and for the development of class of therapeutic for AD. Many of these more effective therapeutic agents. Here we antibodies have been evaluated in clinical trials, but employed a novel “epitomic” approach to map the so far none have demonstrated consistent fine structure of the epitopes of 28 monoclonal therapeutic activity in slowing the progression of antibodies against amyloid-beta using AD and none have been approved by the FDA immunoselection of random sequences from a (reviewed in (4,5)). phage display library, deep sequencing and pattern Conformation-dependent monoclonal antibodies analysis to define the critical sequence elements against Aβ are also an invaluable tool for research recognized by the antibodies. Although most of the in the role of amyloids in Alzheimer’s disease antibodies map to major linear epitopes in the because they recognize epitopes that are amino terminal 1-14 residues of Aß, the antibodies differentially displayed on distinct structural display differences in the target sequence residues polymorphs or folded states of the peptide, that are critical for binding and in their individual providing insight into the role of polymorphisms in preferences for non-target residues, indicating that the pathogenic spectrum of the disease (6). Many of the antibodies bind to alternative conformations of the antibodies raised against Aß amyloid the sequence by different mechanisms. Epitomic specifically bind to aggregated oligomeric or analysis also identifies more discontinuous, non- fibrillar forms of the peptide and do not bind to overlapping sequence Aß segments than peptide monomer or the amyloid precursor protein (7-10). array approaches that may constitute the Moreover, many of these aggregation state-specific conformational epitopes that underlie the antibodies recognize aggregates formed from aggregation specificity of antibodies. Aggregation unrelated sequences, indicating that amyloids specific antibodies recognize sequences that display generic epitopes as a consequence of their display a significantly higher predicted propensity aggregated structure. Understanding the for forming amyloid than antibodies that recognize mechanisms underlying the unusual specificities of monomer, indicating that the ability of random anti-amyloid antibodies and the structures of their sequences to aggregate into amyloid is a critical corresponding epitopes is crucial for the element of their binding mechanism. understanding of the immune response to amyloid and for the development of effective therapeutic Introduction agents. We have previously reported polyclonal rabbit sera Amyloids are intermolecularly hydrogen bonded ß- that distinguish two classes of amyloid aggregates. sheet aggregates that have a regularly repeating A11 sera was raised against oligomeric Aß mimics lattice structure (1,2). Unlike natively folded made of Aß42 coupled via the carboxyl terminus to proteins that adopt a single or limited range of colloidal gold that recognizes prefibrillar structures or states, amyloids can adopt a large oligomers, but not fibrils or monomer (7). number of different ß-sheet aggregate structures Subsequently we reported OC serum that was that vary in the parallel vs antiparallel strand raised against Aß42 fibrils that is specific for fibrils 2 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.05.238105; this version posted November 19, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. and fibrillar oligomers but not prefibrillar Library randomness and amino acid bias are oligomers or monomer (11). A11 and OC serum important for distinguishing specific and non- recognize prefibrillar oligomers and fibrils specific binding. In order to assess the randomness respectively from several other amyloid forming and amino acid bias in the library as purchased from proteins, including α-synuclein and islet amyloid the manufacturer, DNA from an aliquot of the polypeptide (IAPP), indicating that some of the library was extracted, the random sequences antibodies in the sera recognize generic amyloid amplified by PCR and deep sequenced, yielding aggregates that do not depend on the precise peptide 19,434 sequences of which 97.4% were unique sequence. We also raised rabbit antiserum against sequences (Supplemental Information file 1, Aß42 annular protofibrils, which are pore-like LibrarySeqOnly, column A, Original unamplified structures, but this serum labels the same type of library). We also amplified the library 2 successive prefibrillar oligomers as A11, indicating that these rounds and examined the effect on the number of sera have related specificities (12). We cloned 6 unique sequences observed. (Supplemental monoclonals from A11 serum-producing rabbits Information file 1, LibrarySeqOnly, columns B, C, (9), 2 monoclonals from annular protofibril Amplified round 1 and round 2). In the first round vaccinated rabbits (Glabe, unpublished) and 23 of amplification, 93.4% of the sequences were monoclonals from OC serum-producing rabbits unique while in the second round of amplification (10). For the OC monoclonals (mOC), the epitope 93.9% were unique. To further compare the was mapped using the PepSpots method of randomness of the library, we compared the unique overlapping 10mers that vary by a single amino sequences between the original library and the first acid from position -5 to 45. Nineteen of the 23 mOC amplification and found that 2.1% of the unique monoclonals gave a pattern of reactivity that sequences were observed in both libraries. This is mapped to either a linear or discontinuous epitope the same as the number of unique sequences that in the amino terminal 2/3 of the molecule (10) while were observed twice in the same original 16 map to the amino terminal residues from 1-11 unamplified library (2.1%). These results indicate (10). Although these antibodies bind to the same that while some sequences may have amplified regions of Aß, they bind selectively to alternative preferentially, the vast majority of sequences in the fibrils structures of Aß, indicating that this region is library remain random after amplification. conformationally polymorphic (13). Here we report that the antibodies have very The library has an inherent bias in terms of the different epitopes and binding modes in terms of the frequency of nucleotides observed as shown in 1) target residues that participate in binding. 2) Supplemental Information Table I. Thymidine is target residues that are not important for binding 3) over-represented, while adenine is the lowest non-target residues that are preferred over target abundance. We also calculated the frequency of residues for binding.