Mutational analysis of a sequence-specific ssDNA binding lupus autoantibody

11F8 is a murine anti-ssDNA monoclonal autoantibody isolated from a lupus prone autoimmune mouse. This mAb binds sequence specifically, and prior studies have defined the thermodynamic and kinetic basis for sequence-specific recognition of ssDNA (Ackroyd, P. C., et al. (2001) Biochemistry 40, 2911-2922; Beckingham, J. A. and Glick, G. D. (2001) Bioorg. Med. Chem. 9, 2243-2252). Here we present experiments designed to identify the residues on 11F8 that mediate sequence-specific, noncognate, and nonspecific recognition of ssDNA and their contribution to the overall binding thermodynamics. Site-directed mutagenesis of an 11F8 single-chain construct reveals that six residues within the complementarity determining regions of 11F8 account for ca. 80% of the binding free energy and that there is little cooperativity between these residues. Germline-encoded aromatic and hydrophobic side chains provides the basis for nonspecific recognition of single-stranded thymine nucleobases. Sequence-specific recognition is controlled by a tyrosine in the heavy chain along with a somatically mutated arginine residue. Our data show that the manner in which 11F8 achieves sequence-specific recognition more closely resembles RNA-binding proteins such as U1A than other types of nucleic acid binding proteins. In addition, comparing the primary sequence of 11F8 with clonally related antibodies that differ by less than five amino acids suggests that somatic mutations which confer sequence specificity may be a feature that distinguishes glomerulotrophic pathogenic anti-DNA from those that are benign.

Dissection of binding interactions in the complex between the anti-lysozyme antibody HyHEL-63 and its antigen

Alanine-scanning mutagenesis, X-ray crystallography, and double mutant cycles were used to characterize the interface between the anti-hen egg white lysozyme (HEL) antibody HyHEL-63 and HEL. Eleven HEL residues in contact with HyHEL-63 in the crystal structure of the antigen-antibody complex, and 10 HyHEL-63 residues in contact with HEL, were individually truncated to alanine in order to determine their relative contributions to complex stabilization. The residues of HEL (Tyr20, Lys96, and Lys97) most important for binding HyHEL-63 (Delta G(mutant) - Delta G(wild type) > 3.0 kcal/mol) form a contiguous patch at the center of the surface contacted by the antibody. Hot spot residues of the antibody (Delta Delta G > 2.0 kcal/mol) are organized in two clusters that juxtapose hot spot residues of HEL, resulting in energetic complementarity across the interface. All energetically critical residues are centrally located, shielded from solvent by peripheral residues that contribute significantly less to the binding free energy. Although HEL hot spot residues Lys96 and Lys97 make similar interactions with antibody in the HyHEL-63/HEL complex, alanine substitution of Lys96 results in a nearly 100-fold greater reduction in affinity than the corresponding mutation in Lys97. To understand the basis for this marked difference, we determined the crystal structures of the HyHEL-63/HEL Lys96Ala and HyHEL-63/HEL Lys97Ala complexes to 1.80 and 1.85 A resolution, respectively. Whereas conformational changes in the proteins and differences in the solvent networks at the mutation sites appear too small to explain the observed affinity difference, superposition of free HEL in different crystal forms onto bound HEL in the wild type and mutant HyHEL-63/HEL complexes reveals that the side-chain conformation of Lys96 is very similar in the various structures, but that the Lys97 side chain displays considerable flexibility. Accordingly, a greater entropic penalty may be associated with quenching the mobility of the Lys97 than the Lys96 side chain upon complex formation, reducing binding. To further dissect the energetics of specific interactions in the HyHEL-63/HEL interface, double mutant cycles were constructed to measure the coupling of 13 amino acid pairs, 11 of which are in direct contact in the crystal structure. A large coupling energy, 3.0 kcal/mol, was found between HEL residue Lys97 and HyHEL-63 residue V(H)Asp32, which form a buried salt bridge surrounded by polar residues of the antigen. Thus, in contrast to protein folding where buried salt bridges are generally destabilizing, salt bridges in protein-protein interfaces, whose residual composition is more hydrophilic than that of protein interiors, may contribute significantly to complex stabilization.

The role of somatic mutation in determining the affinity of anti-DNA antibodies

Combinatorial antibody libraries were constructed from the spleen of a patient with concomitant systemic lupus erythematosus and idiopathic thrombocytopenia. Following selection of the libraries with DNA, a panel of 15 anti-DNA Fabs was isolated. Sequence analysis of these antibodies coupled with measurements of their affinities for ss- and dsDNA were used to investigate the role of somatic mutation in affinity maturation of the anti-DNA response. Examination of the germline genes used by these Fabs supports previous studies that suggest there is no restriction of the gene usage in the anti-DNA response. However, data are presented indicating that VH3 genes and the A27 V(kappa) paired with the J(kappa)1 may be over-expressed in the anti-DNA repertoire. Analysis of the role of somatic mutation in increasing affinity for DNA indicates that affinity maturation has occurred and suggests that the CDR1 and CDR2 of the heavy chain are of importance in this process.

Elucidation of the paratope of scFv-8H9D4, a PAI-1 neutralizing antibody derivative

Interfering with increased levels of plasminogen activator inhibitor-1 (PAI-1) might offer new therapeutic strategies for a variety of cardiovascular diseases. Inactivation of PAI-1 can be accomplished by a number of monoclonal antibodies (MA), including MA-8H9D4. In a previous study, a single-chain variable fragment (scFv-8H9D4) was cloned and found to have the same properties as the parental MA-8H9D4. In the present study, we identified the residues of scFv-8H9D4 that contribute significantly to the paratope. The complementarity determining region 3 from the heavy (H3) and the light (L3) chain were analysed through site-directed mutagenesis. Out of twelve mutations, only four residues appeared to contribute to the paratope. The affinity of scFv-8H9D4-H3-L97D for PAI-1 was 38-fold decreased (K(A) = 4.8 +/- 0.2 x 10(7) M(-1) vs. 1.8 +/- 0.7 x 10(9) M(-1) for scFv-8H9D4) whereas scFv-8H9D4-H3-R98Y did not bind to PAI-1. The affinities of scFv-8H9D4-L3-Y91S and scFv-8H9D4-L3-F94D for PAI-1 were 9- and 5-fold reduced, respectively, whereas the combined mutation resulted in an 86-fold decreased affinity (K(A) = 2.1 +/- 0.2 x 10(7) M(-1)). In accordance with the affinity data, these mutants had no, or a reduced, PAI-1 inhibitory capacity, confirming that these four particular residues form the major interaction site of scFv-8H9D4 with PAI-1. In combination with the three-dimensional structure, these data contribute to the rational design of PAI-1 neutralizing compounds.

B cell selection and affinity maturation during an antibody response in the mouse with limited B cell diversity

The quasi-monoclonal mouse has limited B cell diversity, whose major (approximately 80%) B cell Ag receptors are comprised of the knockin V(H) 17.2.25 (V(H)T)-encoded H chain and the lambda1 or lambda2 L chain, thereby being specific for 4-hydroxy-3-nitrophenylacetyl. The p-nitrophenylacetyl (pNP) was found to be a low affinity analog of nitrophenylacetyl. We examined affinity maturation of anti-pNP IgG by analyzing mAbs obtained from quasi-monoclonal mice that were immunized with this low affinity Ag. The results are: 1) Although V(H)T/lambda1 and V(H)T/lambda2 IgM were equally produced, V(H)T/lambda2 IgG almost exclusively underwent affinity maturation toward pNP. 2) A common mutation in complementarity-determining region 3 of V(H)T (T313A) mainly contributed to generating the specificity for pNP. 3) Because mutated V(H)T-encoded gamma-chains could form lambda1-bearing IgG in Chinese hamster ovary cells, apparent absence of V(H)T/lambda1 anti-pNP IgG may not be due to the incompatibility between the gamma-chains and the lambda1-chain, but may be explained by the fact that V(H)T/lambda1 B cells showed 50- to 100-fold lower affinity for pNP than V(H)T/lambda2 B cells. 4) Interestingly, a pNP-specific IgM mAb that shared common mutations including T313A with high affinity anti-pNP IgG was isolated, suggesting that a part of hypermutation coupled with positive selection can occur before isotype switching. Thus, even weak B cell receptor engagement can elicit an IgM response, whereas only B cells that received signals stronger than a threshold may be committed to an affinity maturation process.

Generation and iterative affinity maturation of antibodies in vitro using hypermutating B-cell lines

We show that iterative antigen-mediated selection of B-cell lines that constitutively hypermutate their immunoglobulin V genes during culture can be exploited to generate antibodies in vitro. From Ramos, a hypermutating human B-cell line expressing IgM of unknown specificity, we derived descendants that exhibit stepwise improved binding to streptavidin. Binding is initially conferred by mutations in complementarity-determining regions (CDRs), but maturation is due to strategic framework mutations. A more powerful system is provided by a hypermutating chicken B-lymphoma line, owing to its rapid proliferation, high rate of mutation accumulation, and genetic tractability. Starting from a single cell, we selected parallel lineages of derivatives, making mutated antibodies of increasing affinity to independent test antigens. Selection is initiated at an exceedingly low affinity threshold, but antibodies can be delivered with nanomolar affinities. The strategy could prove useful for in vitro generation of antigen-specific monoclonal antibodies and may be extendable to the maturation of other protein-ligand interactions.

Induction of somatic hypermutation in immunoglobulin genes is dependent on DNA polymerase iota

Somatic hypermutation of immunoglobulin genes is a unique, targeted, adaptive process. While B cells are engaged in germinal centres in T-dependent responses, single base substitutions are introduced in the expressed Vh/Vl genes to allow the selection of mutants with a higher affinity for the immunizing antigen. Almost every possible DNA transaction has been proposed to explain this process, but each of these models includes an error-prone DNA synthesis step that introduces the mutations. The Y family of DNA polymerases--pol eta, pol iota, pol kappa and rev1--are specialized for copying DNA lesions and have high rates of error when copying a normal DNA template. By performing gene inactivation in a Burkitt's lymphoma cell line inducible for hypermutation, we show here that somatic hypermutation is dependent on DNA polymerase iota.

How do two unrelated antibodies, HyHEL-10 and F9.13.7, recognize the same epitope of hen egg-white lysozyme?

The anti-hen egg-white lysozyme (HEWL) antibodies HyHEL-10 and F9.13.7 recognize a common epitope. The structures of the complexes differ, however, in the numbers of electrostatic and hydrogen-bond interactions and in the distributions of contacts between the light and heavy chains. The equilibria and kinetics characterizing the F9.13.7 complex formation were evaluated for both wild-type and mutant derivatives of HEWL to help to understand how the different contacts are effectively used in the complexes with the two antibodies. Three epitope hot spots, Y20, K96, and R73 (destabilization > 4 kcal/mole), were found by alanine scanning mutagenesis. The first two constitute two of the three hot spots in the HyHEL-10 complex. The hot spots of the HyHEL-10 paratope are centered on the HEWL epitope; whereas R73 (HEWL), the only important light-chain-contacting residue, is clearly separated from the other hot spots of the F9.13.7 complex. The larger number of epitope warm plus hot spots found in the F9.13.7 complex compared with that of HyHEL-10 shows that the specificity of the former is greater even though the K(D) value is 20-fold larger. Conservative mutations showed that the specificity enhancement is related to the greater number of functional polar and hydrogen bond interactions in the F9.13.7 complex. Alanine scanning mutagenesis would not have illuminated these distinctions. It is shown that the concept of antigen specificity, as defined by cross-reactivity with natural variant antigens, is flawed by phylogenetic bias, and that specificity can only be defined by the use of unbiased epitopes, which are conveniently accessed by site-directed mutagenesis.

In human IgA nephropathy uteroglobin does not play the role inferred from transgenic mice

BACKGROUND

Uteroglobin (UG)-knockout and UG-antisense transgenic mice develop clinical and pathological features of immunoglobulin A (IgA) nephropathy with heavy proteinuria. These models suggested that UG, an anti-inflammatory protein with high affinity for fibronectin (Fn), prevents the formation of IgA-Fn complexes and mesangial deposits in mice. We aim to elucidate whether similar mechanisms underlie the development and severity of human IgA nephropathy.

METHODS

Specific enzyme-linked immunosorbent assays were devised to detect serum levels of UG binding to Fn or incorporated into IgA-Fn complexes and IgA binding to Fn or collagen IV. Sera from 75 patients with IgA nephropathy with normal renal function and various degrees of proteinuria (0.2 to 5 g/d of protein) stable over the previous 3 months without therapy were investigated and compared with healthy controls.

RESULTS

Levels of UG binding to Fn were similar in patients with IgA nephropathy and healthy controls. UG incorporated into circulating IgA-Fn complexes, as well as levels of IgA-Fn complexes and IgA binding Fn and collagen IV, were significantly greater in patients than healthy controls. Greater amounts of UG incorporated into IgA-Fn complexes reduced the risk for proteinuria with protein greater than 1 g/d (odds ratio = 0.67; P < 0.001). Logistic regression analysis assigned a predictive value for proteinuria persistently greater than 1 g/d of protein to lower amounts of UG incorporated into IgA-Fn complexes (R = -0.267; P = 0.008) and increased binding of IgA to collagen IV (R = 0.214; P = 0.0003).

CONCLUSION

This first report of human IgA nephropathy after the publication of the mouse model shows that UG is not reduced in circulation and is even increased in IgA-Fn complexes. Because aberrant IgA1 glycosylation is the event initiating IgA nephropathy in humans, we speculate that the enhanced incorporation of UG into IgA-Fn complexes might represent feedback to reduce the formation of macromolecular aggregates.

TCR-like human antibodies expressed on human CTLs mediate antibody affinity-dependent cytolytic activity

The permanent genetic programming via gene transfer of autologous T cells with cell surface receptors directed toward tumor-related Ags holds great promise for the development of more-specific tumor therapies. In this study we have explored the use of Abs directed to MHC-peptide complexes (or TCR-like Abs) to engraft CTLs with exquisite specificity for cancer cells. First, we affinity matured in vitro a previously selected TCR-like Ab, Fab-G8, which is highly specific for the peptide melanoma-associated Ag-A1 presented by the HLA-A1 molecule. A combination of L chain shuffling, H chain-targeted mutagenesis, and in vitro selection of phage display libraries yielded a Fab-G8 Ab derivative, Fab-Hyb3, with an 18-fold improved affinity yet identical peptide fine specificity. Fab-G8 and Fab-Hyb3 were expressed on primary human T lymphocytes as cell surface-anchored Fab, demonstrating that T cells expressing the high-affinity Fab-Hyb3 molecule eradicate tumor cells much more effectively. Furthermore, the gain in ligand-binding affinity resulted in a 2-log improvement in the detection of peptide/MHC complexes on melanoma-associated Ag-A1 peptide-loaded cells. In summary, an affinity-matured Ab specifically recognizing a cancer-related peptide/MHC complex was generated and used to improve the tumor cell killing capacity of human T cells. This strategy, based on engraftment of T cells with in vitro engineered Abs, is an attractive alternative to the laborious, and in many cases unsuccessful, generation of highly potent tumor-specific T lymphocytes.

Contribution of light chain rearrangement in peripheral B cells to the generation of high-affinity antibodies

Recently, peripheral B cells have been shown to undergo secondary V(D)J rearrangement of immunoglobulin genes, but the physiological role of this event has not been fully elucidated. To investigate whether rearrangement of L chain genes in the periphery is involved in the generation of high-affinity antibodies (Ab), we used the 17.2.25 rearranged VHDJH gene (VHT)-knockin mouse whose B cell diversity is limited due to the expression of the site-directed transgene. Immunization of the mouse with p-nitrophenylacetyl (pNP)-conjugated chicken gamma-globulin preferentially led to the production of anti-pNP IgG Ab comprised of non-VHT-encoded H chains and lambda chains. lambda(+) IgG constituted a majority of high-affinity Ab to this hapten. RAG-2 mRNA and the recombination signal sequence break of the lambda1 gene increased in the draining lymph node of immunized mice, but not of nonimmunized animals. There was a close correlation between the levels of these parameters implicating lambda gene rearrangement and the production of lambda(+ )high-affinity anti-pNP IgG. These observations were reproduced in RAG-1-deficient mice that were reconstituted with the spleen cells ofthe knockin mouse. Thus, our findings suggest that L chain rearrangement that occurs in the periphery can contribute to affinity maturation of Ab.

Quantitative screening of yeast surface-displayed polypeptide libraries by magnetic bead capture

Magnetic bead capture is demonstrated here to be a feasible alternative for quantitative screening of favorable mutants from a cell-displayed polypeptide library. Flow cytometric sorting with fluorescent probes has been employed previously for high throughput screening for either novel binders or improved mutants. However, many laboratories do not have ready access to this technology as a result of the limited availability and high cost of cytometers, restricting the use of cell-displayed libraries. Using streptavidin-coated magnetic beads and biotinylated ligands, an alternative approach to cell-based library screening for improved mutants was developed. Magnetic bead capture probability of labeled cells is shown to be closely correlated with the surface ligand density. A single-pass enrichment ratio of 9400 +/- 1800-fold, at the expense of 85 +/- 6% binder losses, is achieved from screening a library that contains one antibody-displaying cell (binder) in 1.1 x 10(5) nondisplaying cells. Additionally, kinetic screening for an initial high affinity to low affinity (7.7-fold lower) mutant ratio of 1:95,000, the magnetic bead capture method attains a single-pass enrichment ratio of 600 +/- 200-fold with a 75 +/- 24% probability of loss for the higher affinity mutant. The observed high loss probabilities can be straightforwardly compensated for by library oversampling, given the inherently parallel nature of the screen. Overall, these results demonstrate that magnetic beads are capable of quantitatively screening for novel binders and improved mutants. The described methods are directly analogous to procedures in common use for phage display and should lower the barriers to entry for use of cell surface display libraries.

Selection of high affinity p-azophenyarsonate Fabs from heavy-chain CDR2 insertion libraries

The length of the heavy chain complementarity-determining region two (HCDR2) of the unmutated anti-p-azophenylarsonate (Ars) monoclonal antibody (36-65 mAb) was extended by three residues in order to test whether this insertion can provide additional contacts between the Ab and the antigen. Two libraries were generated using 36-65 heavy and light chain genes which were cloned as Fab in the phage-display vector pComb3. In the first library, three randomized amino acids were inserted between residues Gly 54 and Asn 55, which are the most solvent exposed residues in the HCDR2 loop. In the second library, in addition to the 3-mer randomized insertion, the flanking residues at positions 54 and 55 were also randomized to allow additional loop flexibility for binding to Ars. Solid-phase and solution phase affinity panning were used to select for clones that bind to Ars. Results indicate that diverse 3-mer HCDR2 insertions can be tolerated, and affinities 10-fold higher than germline encoded 36-65 Ab can be obtained. The sequence diversity of the insertion among the selected clones from both libraries suggests that the insertion increases contact between the Ab and the protein carrier rather than the hapten alone.

The impact of H2-DM on humoral immune responses

H2-DM (DM, previously H2-M) facilitates the exchange of peptides bound to MHC class II molecules. In this study, we have used H2-DM-deficient (DM(-/-)) mice to analyze the influence of DM in the priming of B cell responses in vivo and for Ag presentation by B cells in vitro. After immunization, IgG Abs could be raised to a T-dependent Ag, 4-hydroxy-5-nitrophenylacetyl-OVA, in DM(-/-) mice, but closer analysis revealed the IgG response to be slower, diminished in titer, and composed of low-affinity Abs. The Ab response correlated with a vast reduction in the number of germinal centers in the spleen. The presentation of multiple epitopes by H2-A(b) from distinct Ags was found to be almost exclusively DM-dependent whether B cells internalized Ags via fluid phase uptake or using membrane Ig receptors. The poor B cell response in vivo could be largely, but not completely restored by expression of a H2-Ea(d) transgene, despite the fact that Ag presentation by H2-E(d/b) molecules was found to be highly DM dependent. Hence, while substantial Ab responses can be raised in the absence of DM, this molecule is a crucial factor both for Ag processing and for the normal maturation of T-dependent humoral immune responses in vivo.

Analysis of antibodies of known structure suggests a lack of correspondence between the residues in contact with the antigen and those modified by somatic hypermutation

Forty unique murine antibody-antigen complexes determined at 2.5 A or less resolution are analyzed to determine whether the residues in direct contact with the antigen are modified by somatic hypermutation. This was done by taking advantage of the recent characterization of the pool of Vkappa germline genes of the mouse. The average number of residues in contact with the antigen in the V(L) gene, which contains the CDRL-1, CDRL-2, and all but one residue of CDRL-3, was six. The average number of somatic mutations was similar (around five). However, as many as 53% of the antibodies did not show somatic replacements of residues in contact with the antigen. Another 28% had only one. Overall, the frequency of antibodies with increasing number of somatic replacements in residues in contact with the antigen decreased exponentially. A possible explanation of this finding is that mutations in the contacting residues have an adverse effect on the antigen-antibody interaction. This implies that most of the observed mutations are those remaining after negative (purifying) selection. Therefore, efficient strategies of site-directed mutagenesis to improve the affinity of antibodies should be focused on residues other than those directly interacting with the antigen.

Copper chaperone for superoxide dismutase co-aggregates with superoxide dismutase 1 (SOD1) in neuronal Lewy body-like hyaline inclusions: an immunohistochemical study on familial amyotrophic lateral sclerosis with SOD1 gene mutation

The copper chaperone for superoxide dismutase (CCS) interacts with Cu/Zn-binding superoxide dismutase 1 (SOD1) specifically and delivers copper to SOD1. To determine the role of the CCS-SOD1 interaction in the pathogenesis of SOD1-mutated familial amyotrophic lateral sclerosis (FALS) patients, we produced an affinity-purified rabbit antibody against CCS and investigated the immunohistochemical localization of both CCS and SOD1 in neuronal Lewy body-like hyaline inclusions (LBHIs) in the spinal cords of two FALS patients with a two-base pair deletion at codon 126 in the SOD1 gene and three FALS patients with an Ala to Val substitution at codon 4. The LBHIs in anterior horn cells from the five FALS patients showed identical immunoreactivities for CCS: the reaction product deposits with the antibody against CCS were generally restricted to the periphery of the core and halo-type LBHIs. The localizations of the immunoreactivities for CCS and SOD1 were similar in the inclusions: both CCS and SOD1 colocalized in neuronal LBHIs in the five mutant SOD1-linked FALS patients. Our results suggest that the specific interaction and aggregation of CCS-SOD1 (probably CCS-mutant SOD1) in SOD1-mutated FALS patients may amplify the formation of inclusions and emphasize a more marked mutant SOD1-mediated toxicity.

Somatic diversity of the immunoglobulin repertoire is controlled in an isotype-specific manner

We have studied two aspects of the IgE immune response. First, we have compared the kinetics of the IgE response to the T cell-dependent antigen ph-Ox coupled to ovalbumin with that of the IgG1 response and we have assessed the quality of the IgE response. Second, we have studied the generation of somatic diversity, understood as the combined effect of somatic mutation and the selection of D(iversity) and J(oining) elements, in germinal center B cells at the molecular level, using the germ-line sequence of the prototype anti-ph-Ox heavy chain variable element V(H)Ox1 as reference. We evaluated sequences derived from mu-, gamma 1- and epsilon-variable elements and showed that somatic diversification was different for all isotypes studied. We further compared the IgE responses of wild-type mice with those of mice expressing a truncated cytoplasmic IgE tail (IgE(KVK Delta tail)). IgE(KVK Delta tail) mice showed a more diverse sequence pattern. We corroborated previous results suggesting that short CDR3 regions are indicative for high-affinity antibodies by measuring relative affinities of phage-expressed Fab fragments with prototype long and short CDR3 regions. Therefore, the composition of the antigen-receptor is responsible for the selection process and the expansion of antigen-specific cells, leading to an isotype-specific antibody repertoire.

Decreased frequency of somatic hypermutation and impaired affinity maturation but intact germinal center formation in mice expressing antisense RNA to DNA polymerase zeta

To examine a role of DNA polymerase zeta in somatic hypermutation, we generated transgenic mice that express antisense RNA to a portion of mouse REV3, the gene encoding this polymerase. These mice express high levels of antisense RNA, significantly reducing the levels of endogenous mouse REV3 transcript. Following immunization to a hapten-protein complex, transgenic mice mounted vigorous Ab responses, accomplished the switch to IgG, and formed numerous germinal centers. However, in most transgenic animals, the generation of high affinity Abs was delayed. In addition, accumulation of somatic mutations in the V(H) genes of memory B cells from transgenic mice was decreased, particularly among those that generate amino acid replacements that enhance affinity of the B cell receptor to the hapten. These data implicate DNA polymerase zeta, a nonreplicative polymerase, in the process of affinity maturation, possibly through a role in somatic hypermutation, clonal selection, or both.

Mutations of an epitope hot-spot residue alter rate limiting steps of antigen-antibody protein-protein associations

The antibodies, HyHEL-10 and HyHEL-26 (H10 and H26, respectively), share over 90% sequence homology and recognize with high affinity the same epitope on hen egg white lysozyme (HEL) but differ in degree of cross-reactivity with mutant lysozymes. The binding kinetics, as measured by BIAcore surface plasmon resonance, of monovalent Fab from both Abs (Fab10 and Fab26) to HEL and mutant lysozymes are best described by a two-step association model consistent with an encounter followed by docking that may include conformational changes. In their complexes with HEL, both Abs make the transition to the docked phase rapidly. For H10, the encounter step is rate limiting, whereas docking is also partially rate limiting for H26. The forward rate constants of H10 are higher than those of H26. The docking equilibrium as well as the overall equilibrium constant are also higher for H10 than for H26. Most of the free energy change of association (Delta G degrees) occurs during the encounter phase (Delta G1) of both Abs. H10 derives a greater amount and proportion of free energy change from the docking phase (Delta G2) than does H26. In the H10--HEL(R21Q) complex, a significant slowing of docking results in lowered affinity, a loss of most of Delta G2, and apparently faster dissociation. Slower encounter and docking cause lowered affinity and a loss of free energy change primarily in the encounter step (Delta G1) of H26 with mutant HEL(R21Q). Overall, in the process of complex formation with lysozyme, the mutations HEL(R21X) affect primarily the docking phase of H10 association and both phases of H26. Our results are consistent with the interpretation that the free energy barriers to conformational rearrangement are highest in H26, especially with mutant antigen.

The integrity of the ball-and-socket joint between V and C domains is essential for complete activity of a humanized antibody

AF2 is a high affinity murine Ab possessing potent neutralizing activity against human IFN-gamma. In carrying out the modifications to humanize this Ab, we discovered that an initial version displayed affinity for IFN-gamma that was slightly less than that of AF2, but exhibited IFN-gamma-neutralizing activity that was severely diminished. Characterization via site-directed mutagenesis revealed that the majority of this loss in IFN-gamma-neutralizing activity was due to altering the V(H) framework residue at position 11. V(H) position 11 is distal to the binding surface of the Ab; however, it, along with residues 110 and 112, have been identified as forming the socket of a molecular ball-and-socket joint between the V and C domains of the Ig Fab, which influences the elbow angle between these domains. To determine whether disrupting the structure of this joint was the basis for reduced IFN-gamma-neutralizing capacity, we altered residue 148 of C(H1), which with residue 149 comprises the corresponding ball portion of the joint. Changing this single C(H1) domain residue diminished the ability of the Ab to neutralize IFN-gamma to a level similar to that observed with the V(H) alteration. Thus, an intact ball-and-socket joint between the V and C domains in AF2 is required for potent neutralization of IFN-gamma. These results suggest the importance of the elbow angle between Ig V and C domains in Ab activity, and support the hypothesis that this joint can be an important functional element of Ab structure.

Tailoring in vitro evolution for protein affinity or stability

We describe a rapid and general technology working entirely in vitro to evolve either the affinity or the stability of ligand-binding proteins, depending on the chosen selection pressure. Tailored in vitro selection strategies based on ribosome display were combined with in vitro diversification by DNA shuffling to evolve either the off-rate or thermodynamic stability of single-chain Fv antibody fragments (scFvs). To demonstrate the potential of this method, we chose to optimize two proteins already possessing favorable properties. A scFv with an initial affinity of 1.1 nM (k(off) at 4 degrees C of 10(-4) s(-1)) was improved 30-fold by the use of off-rate selections over a period of several days. As a second example, a generic selection strategy for improved stability exploited the property of ribosome display that the conditions can be altered under which the folding of the displayed protein occurs. We used decreasing redox potentials in the selection step to select for molecules stable in the absence of disulfide bonds. They could be functionally expressed in the reducing cytoplasm, and, when allowed to form disulfides again, their stability had increased to 54 kJ/mol from an initial value of 24 kJ/mol. Sequencing revealed that the evolved mutant proteins had used different strategies of residue changes to adapt to the selection pressure. Therefore, by a combination of randomization and appropriate selection strategies, an in vitro evolution of protein properties in a predictable direction is possible.

N-terminal mutations in the anti-estradiol Fab 57-2 modify its hapten binding properties

Recombinant antibodies often contain N-terminal mutations arising from the use of degenerate cloning primer sets and/or the introduction of restriction sites in the framework 1 regions. We studied the effects of such mutations in a recombinant anti-estradiol Fab fragment derived from the hybridoma cell line 57-2. The 5' ends of the heavy and light chain genes were originally modified to introduce the restriction sites XhoI and SacI, respectively, for cloning purposes. However, the affinity and specificity of the recombinant Fab were lowered compared to the proteolytic Fab' fragment of the parental hybridoma IgG. Replacing the mutated sites with authentic amino acid coding sequences restored the binding properties as well as increased the bacterial production levels fivefold and 10-fold at 30 and 37 degrees C, respectively. Local changes in the antigen binding site were probed by determining the affinity constants (Kd) for estradiol and four related steroids. It was found that the mutated heavy chain amino terminus specifically increased the Kd for testosterone whereas the mutated light chain amino terminus decreased the Kd for all of the steroids to the same extent; the heavy and light chain effects were additive. Analysis of a newly determined crystal structure of the authentic Fab 57-2 in complex with estradiol suggests that mutations in the residue 2 in V(H), and 2 and 4 in the V(L) domain were those responsible for the observed effects. Their general roles as structure-determining residues for the CDR3 loops imply that similar effects can occur with other recombinant antibodies as well.