The evolution of immunological molecular recognition
Background
The immune system provides an unparalleled opportunity to understand the potential contribution of dynamics to molecular recognition. Within days of encountering an antigen, antibodies are produced that selectively bind and help eradicate the foreign molecule. Sequence diversity prior to encountering antigen (i.e. the repertoire of germline antibodies) is introduced during the combinatorial recombination of the different gene segments that encode the heavy (VH) and light (VL) chains, the N-terminal ‘variable regions’ that pack on one another to form the binding or ‘combining’ site of the antibody. However, the number of germline antibodies that may be displayed is limited by the number of available B cells, and is insufficient to include antibodies that bind any of the infinite number of foreign antigens. Thus, the immune system introduces additional diversity in a process called affinity maturation.
B cell(s) displaying germline Abs that recognize an offending target with at least low affinity proliferate in the spleen and lymph nodes, forming germinal centers, and compete for a limiting amount of antigen. Those that compete less effectively for antigen undergo apoptosis, while those that compete more effectively are stimulated to divide and multiply in a process called clonal expansion. During clonal expansion, mutations are introduced into the antibody genes, and the process continues among the mutant progeny of the initially selected B cells. Through iterative cycles of somatic hypermutation and clonal expansion, the process of affinity maturation typically results in the introduction of 1 to 15 amino acid mutations per antibody, increasing affinity for the antigen up to 105-fold. After these mature Abs help to fight off the infection, the B cell-displayed mature Abs may be committed to the memory repertoire, thereby granting the host lasting immunity.
The observation that highly somatically mutated, mature antibodies are typically exquisitely specific for binding their target antigen, while germline antibodies must be broadly polyspecific, suggests that affinity maturation evolves polyspecific antibodies into highly specific antibodies. One biophysical property that distinguishes polyspecificity from specificity is protein flexibility. A flexible combining site is able to adopt different conformations that recognize different antigens, while a rigid combining site is locked into a conformation that is specific for a given antigen. This suggests that nature solves the problem of recognizing an infinite number of potential foreign molecules, without recognizing any self molecules (which would cause autoimmunity), by evolving flexible, polyspecific germline antibodies, which are kept at low concentration (to prevent autoimmunity), into more rigid and specific mature Abs. Only the mature and exquisitely specific antibodies are produced in sufficient quantities to fight off an infection. Importantly, this suggests that Ab dynamics are tailored during affinity maturation to mediate immunological molecular recognition; and this provides the key to determining which, if any, of the observed dynamics are biologically relevant – biologically relevant dynamics will change as a function of evolution. Furthermore, detailed characterization of changes of biologically relevant motions as a function of evolution should provide unprecedented insight into both immunology and evolution.
Somatic mutations in the anti-fluorescein antibody 4-4-20 restrict conformational dynamics
To directly test the hypothesis that affinity maturation tailors Ab dynamics and conformational heterogeneity, we have employed antibodies that were evolved to bind chromophoric antigens, such as fluorescein (Fl). Chromophoric antigens are ideal, because they allow for the use of spectroscopic methods, such as three-pulse photon echo peak shift (3PEPS) spectroscopy, but otherwise are expected to be recognized just like any other antigen. Much of our work has employed the monoclonal anti-Fl Ab 4-4-20. We identified the germline precursor of Ab 4-4-20 via DNA sequencing and comparison of the 5’-untranslated region (5’UTR) of the 4-4-20 hybridoma and BALB/c genomic DNA. For both the VL and VH genes, comparison of 500 to 600 nucleotides of 5’UTR DNA was sufficient to unambiguously assign the germline gene. Sequence comparison of the germline Ab and 4-4-20 coding regions revealed that affinity maturation introduced 10 and 2 mutations into the VH and VL, respectively, as shown below.
To characterize how affinity changed during somatic evolution, we first determined the Fl binding constants for germline and mature Ab 4-4-20 and the VLgl chimera using surface plasmon resonance (See Table, below). The total increase in affinity is 170-fold, with a 14-fold increase resulting from VH evolution (germline Ab vs. VLgl chimera) and a 12-fold increase from VL evolution (VLgl chimera vs. mature Ab). In addition, we constructed mutants of both the mature Ab and the VLgl chimera, where, one at a time, each somatic mutation was varied from its mature to its germline residue. Interestingly, two of the ten VH mutations, SH32Y and CH38R (corresponding to mutants VHY32S and VHR38C, respectively, in the Table, below), appear to account for most of the increased binding affinity associated with VH evolution - despite being 10 and 20 Å removed from the binding site, respectively. These changes resulted from changes in both kon and koff, with kon increasing 4-fold with VH evolution and 2-fold with VL evolution, while koff decreased 4-fold with VH evolution and 6-fold with VL evolution. Thus, the 170-fold increase in affinity results predominantly, although not exclusively, from a decrease in the rate of Fl dissociation.
Affinity maturation and the evolution of fast Ab dynamics
We used 3PEPS and DSS to characterize the fs to ns dynamics of the Ab-Fl complexes as a function of affinity maturation. As is apparent in the data shown below, the 3PEPS and DSS signals differ significantly among the Ab-Fl complexes. Strikingly, we observed a large long-time signal offset in the 3PEPS decay for the germline Ab, which is significantly reduced in the VLgl chimera, and virtually absent in the mature Ab. This demonstrates significant inhomogeneous broadening (i.e. conformational heterogeneity) in the germline Ab that is dramatically reduced with affinity maturation. Time-resolved fluorescence data further support this conclusion. The fluorescence lifetime distribution of the germline Ab shows multiple maxima, indicating the population of different conformations that do not interconvert during the fluorescence lifetime. The lifetime distribution narrows for the VLgl chimera, and shows only a single peak for the mature Ab, consistent with the conclusion that affinity maturation restricted the conformational heterogeneity of the Ab.
By combining the 3PEPS and DSS data, we determined λinh and the ρB(ω) between 0.1 and 500 cm-1 for each Ab, and found that on the fs to ns timescale, affinity maturation consistently and significantly rigidified the Ab-Fl complex. Conformational heterogeneity (λinh) is systematically reduced by both VH and VL maturation. VH maturation had the biggest effect on the low frequency part of ρB(ω), corresponding to motions on the ns timescale described by λK2 and τK2. Such motions are prominent in the germline Ab, less prominent in the VLgl chimera, and virtually absent in the mature Ab. VL maturation reduced the amplitude (λK1) of the ps timescale motions and increased the corresponding time constant (τK1), i.e. reflecting conformational substates that are separated by higher barriers in the mature Ab, and less configurational space is sampled through these motions. No such ps dynamics is observed in the germline Ab, possibly because τK2 is too fast to be distinguished from the fs dynamics (i.e. interconversion between conformational substates is essentially barrierless in the germline protein). Interestingly, the data reveal a correlation between conformational heterogeneity (λinh) and the amplitude of low frequency motions. Both are significant in the germline Ab, reduced in the VLgl chimera, and virtually absent in the mature Ab. The correlation between conformational heterogeneity and low-frequency motions suggests that affinity maturation not only rigidified the Ab combining site, but also localized it to a single conformation, likely the one best suited for binding Fl.
In addition to 3PEPS, we are also using NMR spectroscopy to characterize the evolution of conformational heterogeneity in antibodies. Using the single-chain variant (scFV) of the 4-4-20 Fab used in our previous work, NMR is making it possible for us to observe dynamics over a broad range of timescales (sub-ns to ms, or longer) and further extend our model.
To address the dynamics underlying the divergent evolution of molecular recognition, we have also raised Abs to the chromophore, 8-methoxypyrene-1,3,6,trisulfonic acid (MPTS). The initial characterization of one mature anti-MPTS Ab, 6C8, demonstrated that it is an excellent chromophore for the characterization of Ab dynamics via ultrafast laser spectroscopy. Subsequently, the VH and VL gene sequences of 6C8, and several other anti-MPTS Abs revealed that 6C8 is homologous to another mature Ab, 8B10 (differentiated by 5 VH and 2 VL residues), and we identified a second pair of homologous mature Abs, 10A6 and 5D11 (differentiated by 4 VH and 3 VL residues). The V-J and V-D-J recombination junctions, which are unique to each germline Ab, are identical in each pair, demonstrating that the Abs of each pair evolved from common germline antecedent. Germline sequences for Abs 10A6, 5D11, 6C8, and 8B10 were identified. Because the homologous Abs have identical recombination junctions as well as both common and unique somatic mutations, we can conclude that 10A6 and 5D11, and 6C8 and 8B10 do not represent different points along one B cell lineage, but rather they are ‘siblings’ that diverged during evolution from a common germline Ab. More than half of the mutations introduced are outside the CDR loops, suggesting that they do not directly contact MPTS. The ongoing characterization of these anti-MPTS Abs is providing a unique opportunity to characterize the divergent evolution of molecular recognition.
A general model for the evolution of novel proteins. We are also working to generalize the concepts developed in our work focused on understanding evolution within the immune system to the problem of how all proteins are evolved. This project is new and expected to draw not only on biophysical techniques and approaches, but also on sequencing, phylogenetics, and expression analysis.
Jump to a publication for more detail: