Elsevier

New Biotechnology

Volume 60, 25 January 2021, Pages 173-182
New Biotechnology

Full length Article
Combinatorial mutagenesis with alternative CDR-L1 and -H2 loop lengths contributes to affinity maturation of antibodies

https://doi.org/10.1016/j.nbt.2020.09.002Get rights and content

Highlights

  • Libraries constructed at CDR-L1, -L3 and -H2 sites with varied loop lengths.

  • Stable scFvs with non-canonical loop lengths were identified by phage display.

  • Affinity improvement was mostly attributable to a change in CDR-L1 or -H2.

  • Over 20-fold improvement in binding affinity was obtained for hapten targets.

  • Length variation in CDR-L1 and -H2 is a potent strategy for affinity maturation of Abs.

Abstract

Loop length variation in the complementary determining regions (CDRs) 1 and 2 encoded in germline variable antibody genes provides structural diversity in naïve antibody libraries. In synthetic single framework libraries the parental CDR-1 and CDR-2 length is typically unchanged and alternative lengths are provided only at CDR-3 sites. Based on an analysis of the germline repertoire and structure-solved anti-hapten and anti–peptide antibodies, we introduced combinatorial diversity with alternative loop lengths into the CDR-L1, CDR-L3 and CDR-H2 loops of anti-digoxigenin and anti-microcystin-LR single chain Fv fragments (scFvs) sharing human IGKV3−20/IGHV3−23 frameworks. The libraries were phage display selected for folding and affinity, and analysed by single clone screening and deep sequencing. Among microcystin-LR binders the most frequently encountered alternative loop lengths were one amino acid shorter (6 aa) and four amino acids longer (11 aa) CDR-L1 loops leading up to 17- and 28-fold improved affinity, respectively. Among digoxigenin binders, 2 amino acids longer (10 aa) CDR-H2 loops were strongly enriched, but affinity improved anti-digoxigenin scFvs were also encountered with 7 aa CDR-H2 and 11 aa CDR-L1 loops. Despite the fact that CDR-L3 loop length variants were not specifically enriched in selections, one clone with 22-fold improved digoxigenin binding affinity was identified containing a 2 residues longer (10 aa) CDR-L3 loop. Based on our results the IGKV3−20/IGHV3−23 scaffold tolerates loop length variation, particularly in CDR-L1 and CDR-H2 loops, without compromising antibody stability, laying the foundation for developing novel synthetic antibody libraries with loop length combinations not existing in the natural human Ig gene repertoire.

Introduction

The adaptive immune system has a remarkable ability to generate antibodies against a variety of antigens with a limited number of genetic elements as source material. The variable (V) genes, diversity (D) and joining (J) segments rearrange in B cells into complete immunoglobulin variable domains which are responsible for the antigen recognition. The majority of the length variation in the matured variable Ig domains is located at the third complementarity determining region (CDR-3) forming a loop structure at the C-terminal end of the domain as a result of V-, D- and J-segment combinations and junctional diversity at the V-J (VL genes) and V-D-J (VH genes) segment interfaces [1]. In contrast to CDR-3 regions, CDR-1 and -2 benefit neither from the combinatorial nor junctional diversity as they are encoded within the center of the germline V gene regions distant from the junction sites. Consequently, the lengths of CDR-1 and CDR-2 loops in the mature antibodies are limited to the available germline V gene sequences. Despite active somatic hypermutation of the Ig genes during B cell affinity maturation, insertions and deletions in the V gene region covering CDR-1 and -2 are rare events, being present at 1.9 % and 2.6 % frequencies in the memory IgG cell population, respectively [2]. Against this background, it is not surprising that the natural length variation at the CDR-H3 site is mimicked in several synthetic antibody library designs while the lengths of the other CDR loops are kept constant adhering to the corresponding V gene equivalents [[3], [4], [5]].

Only a handful of studies on combinatorial antibody libraries have explored alternative loop lengths in CDRs other than H3. In the n-CoDeR library, in vivo-formed CDR sequences were amplified from cDNA by PCR and incorporated into a specific single framework showing functionality of the idea of exploiting CDRs and alternative loop lengths in a universal antibody library context [6]. However, due to the human-derived source material, the incorporated CDRs were very close to germline sequences, which does not allow full exploration of the boundaries of feasible antibody structures. In another example, a library of heavy (H) chains was paired with a modified version of the IGKV3−20*01 (A27) gene containing a 5 amino acid (aa) longer invariant CDR-L1 loop copied from IGKV4−1*01 (B3) that is often encountered in anti-peptide antibodies [7]. Furthermore, alternative CDR-L3 loop libraries with different lengths were incorporated into a universal synthetic Fab library, demonstrating that CDR-H3 contacts are not mandatory for antigen recognition as they could be replaced by L3 [8]. From a wider perspective, different loop lengths have been the most comprehensively explored with an alternative binder scaffold, the fibronectin 3 domain (Fn3) [9], in which, loops BC, DE and FG (resembling CDRs) were diversified to contain combinatorial libraries in four different loop lengths [10]. It was observed that the Fn3 molecules with the highest affinity to their targets contained novel altered loop lengths which were non-existent in nature, exemplifying the importance of loop length variation as a diversity element.

In single framework antibody libraries, analogous to the alternative scaffold binders, the surface topology of the antigen-binding site is globally constrained to a fixed configuration of CDR main chain conformations dictated by the selected pair of V domains. As is well known, CDR loop length is the major determining factor of the paratope topography [11,12] and, therefore, loop length variation could be especially useful as an extension to combinatorial single framework Ab libraries in order to sample a larger structural sequence-space.

Targeted randomization of residues in CDR regions is widely used in affinity maturation [[13], [14], [15], [16], [17]], but alternative loop lengths have been less investigated as a source of structural diversity. Previously, random amino acid insertions in CDR loops H2 and L3 were explored to obtain higher affinity anti-estradiol variants [18] and to change the specificity profile of an anti-digoxigenin antibody [19]. In the fully synthetic HuCAL Gold library a limited natural loop length variation within a germline family was introduced to a consensus framework of the respective family [20]. Use of the CDR cassettes in affinity maturation of an antibody isolated from the library resulted in an affinity improved clone with an altered CDR-H2 length [21,22].

Deep sequencing allows population-wide sequence analysis of natural and synthetic antibody repertoires [5,[23], [24], [25], [26], [27], [28], [29]]. In phage display, it has helped to follow enrichment and isolate new binders [23,[30], [31], [32], [33], [34]] and, with the increased sequencing read length, even all six CDRs in single-framework scFv antibodies have been retrieved [31]. Furthermore, NGS has been used to obtain sequence information about the overall fitness landscape of the antibodies [35] and to identify affinity-improving mutations in antibodies [36,37]. Here, we have studied the possibility to affinity mature hapten- and peptide-binding antibodies by altering the lengths and sequences of CDR-L1, -L3 and -H2 loops in scFv fragments sharing a fixed KV3−20/HV3−23 framework. The first target of detection, microcystin-LR, is a cyclic heptapeptide toxin produced by blue-green algae and has been a focus for immunoassay development at our department aiming at water quality monitoring [38], while the second target, digoxigenin, is a plant steroid found in foxgloves (Digitalis sp.), often used as a model hapten in protein engineering and a widely used tool for molecular labeling in biotechnology [[39], [40], [41], [42]]. The libraries consisting of the variable domains Vκ3 and VH3 in scFv format were selected with phage display and studied for folding (by protein L), function (by low stringency selection), stability and affinity (by high stringency selection) to digoxigenin and microcystin-LR (McLR) by deep sequencing and single clone analysis.

Section snippets

Reagents and buffers

Rabbit anti-mouse (RAM) IgG (DAKO, Glostrup, Denmark) was labelled with Eu (N1)-chelate using procedure described earlier [43]. The europium (Eu)-labelling of digoxigenin (DIG), microcystin-LR (McLR), and streptavidin, and the synthesis of biotinylated digoxigenin (bio-DIG) and microcystin-LR (bio-McLR) was as described previously [3]. Biotinylated protein L was from Pierce Technology (Rockford, IL, USA). Digoxigenin (DIG) was from Sigma (St. Louis, MO, USA) and McLR was kindly provided by Dr

Experimental setup

We studied the possibility to affinity mature hapten and peptide binding antibodies by altering the length and sequence of CDR-L1, -L3 and -H2 loops in scFv fragments sharing a fixed KV3−20/HV3−23 framework. The antibodies targeted for this study were 10 anti-microcystin-LR (anti-McLR) and 16 anti-digoxigenin (anti-DIG) clones (termed parental) previously isolated by phage display from synthetic human single framework antibody libraries scFvM and scFvP, designed especially for hapten and

Discussion

In this study, it was demonstrated that a synthetic single framework antibody library does not need to be confined to the original CDR loop configuration at the minor CDR-1 and CDR-2 loop sites. Alternative loop lengths, typical of other germline gene families, were tolerated by the scFv structure consisting of the IGKV3−20/IGHV3−23 frameworks and combinatorial libraries derived from it resulted in the discovery of novel structural solutions for molecular recognition. Indeed, the highest

Conclusions

Loop length variation with concomitant codon randomization proved to be a valid strategy for affinity maturation of antibodies isolated from a synthetic single framework library. Significant structural plasticity was observed in the IGKV3−20/IGHV3−23 antibody framework as new loop lengths were encountered in scFvs selected with protein L, as well as in stringent and non-stringent antigen selection conditions against the cognate antigens. Most prominently, 6 aa CDR-L1, 11 aa CDR-L1 and 10 aa

Declaration of Competing Interest

None.

Acknowledgements

We thank Maria Wallenius, Riikka Lammi and Pekka Puolasmaa for technical assistance in performing the experiments. This work was supported by a grant (proChassis, 272621) from the FINSynBio program of Academy of Finland to E.B.

References (53)

  • W.P. Yang et al.

    CDR walking mutagenesis for the affinity maturation of a potent human anti-HIV-1 antibody into the picomolar range

    J Mol Biol

    (1995)
  • U. Lamminmaki et al.

    Expanding the conformational diversity by random insertions to CDRH2 results in improved anti-estradiol antibodies

    J Mol Biol

    (1999)
  • C. Rothe et al.

    The human combinatorial antibody library HuCAL GOLD combines diversification of all six CDRs according to the natural immune system with a novel display method for efficient selection of high-affinity antibodies

    J Mol Biol

    (2008)
  • S. Steidl et al.

    In vitro affinity maturation of human GM-CSF antibodies by targeted CDR-diversification

    Mol Immunol

    (2008)
  • H. Peng et al.

    Mining naïve rabbit antibody repertoires by phage display for monoclonal antibodies of therapeutic utility

    J Mol Biol

    (2017)
  • J. Lövgren et al.

    Next generation sequencing of all variable loops of synthetic single framework scFv-Application in anti-HDL antibody selections

    N Biotechnol

    (2016)
  • S. Schlehuber et al.

    A novel type of receptor protein, based on the lipocalin scaffold, with specificity for digoxigenin

    J Mol Biol

    (2000)
  • A. Krebber et al.

    Reliable cloning of functional antibody variable domains from hybridomas and spleen cell repertoires employing a reengineered phage display system

    J Immunol Methods

    (1997)
  • B.H. Nilson et al.

    Protein L from Peptostreptococcus magnus binds to the kappa light chain variable domain

    J Biol Chem

    (1992)
  • T. Arakawa et al.

    Suppression of protein interactions by arginine: a proposed mechanism of the arginine effects

    Biophys Chem

    (2007)
  • M. Gani et al.

    Monoclonal antibodies against progesterone: effect of steroid-carrier coupling position on antibody specificity

    J Steroid Biochem Mol Biol

    (1994)
  • A.R. Rees

    Understanding the human antibody repertoire

    MAbs

    (2020)
  • B.S. Briney et al.

    Location and length distribution of somatic hypermutation-associated DNA insertions and deletions reveals regions of antibody structural plasticity

    Genes Immun

    (2012)
  • T. Huovinen et al.

    Two ScFv antibody libraries derived from identical VL-VH framework with different binding site designs display distinct binding profiles

    Protein Eng Des Sel

    (2013)
  • T. Tiller et al.

    A fully synthetic human Fab antibody library based on fixed VH/VL framework pairings with favorable biophysical properties

    MAbs

    (2013)
  • E. Söderlind et al.

    Recombining germline-derived CDR sequences for creating diverse single-framework antibody libraries

    Nat Biotechnol

    (2000)
  • Cited by (5)

    View full text