Mar 102009
 
ResearchBlogging.orgIn a post last week I mentioned a technique for obtaining the high-resolution structure of a protein inside a living cell, but I also pointed out that this technique was difficult and expensive, and might not be applicable to large proteins. Techniques improve and become more powerful, of course, but you might not want to wait for NMR to catch up to your question. Fortunately, high-resolution in vivo structures may not be necessary if you already have relevant dilute-solution structures of your protein and merely want to distinguish between different known conformational states. In a recent paper in PNAS, researchers from San Francisco used conformation-specific antibodies to locate activated caspase-1 in cultured cells.

Caspase-1 is a cysteine protease that plays a role in the immune response, as well as being released during apoptosis. From crystal structures we know that this protein can adopt two different structures, of which only one represents a catalytically competent state of the enzyme (the on-form). Caspase-1 also possesses an allosteric site where an inhibitor can bind, locking the enzyme in an inactive conformation (the off-form). When it’s not bound to anything (the apo-form) caspase-1 is presumed to have a conformation similar to the off-form. Like many proteases, caspase-1 has a large, inactivating tail when it is made (the pro-form) that must be cleaved off before activation is possible. The structure of the caspase-1 proenzyme is not known. Current models of inflammatory response propose that after processing, the on-form binds to scaffolding proteins in an “inflammasome”. In order to confirm this proposition, the authors decided to generate antibodies that would bind specifically to the on-form or the off-form of caspase-1.

The key to this experiment was combining irreversible inhibitors that could essentially lock the caspase into one conformation with the phage-display technique for optimizing antibody recognition. The authors had the advantage that both the active site and the allosteric site have cysteines in them. In an oxidizing environment, small molecules can covalently bind to the protein via disulfide bonds, thus locking the enzyme into the on-form or off-form. The authors immobilized these “locked” forms of caspase-1 and used them to screen antibody fragments (Fabs) using phage display. In addition to the typical selection approach, the authors performed anti-selection at one point using the “wrong” conformation to increase the specificity. After several rounds of selection, and some controls to ensure that the antibodies were binding to caspase and not the inhibitors, Gao et al. had several candidates for further optimization and screening. After they completed that process, they had two antibodies, Fabon and Faboff, specific for the two conformations. Each antibody bound to its intended target with a KD of less than 5 nM. The authors also made full antibodies (IgGon and IgGoff) from these Fabs for expression in mammalian cells.

The authors took these new antibodies for a spin with the apo-form of caspase-1. One might naively expect that only Faboff would bind to this protein, but in fact Fabon bound as well, albeit with substantially reduced affinity relative to the on-form. One possible interpretation of this finding is that the apo-form is equivalent to the off-form, but that Fabon can convert it to the on-form via an induced-fit mechanism. If this is the case, then we would expect Faboff to have the same affinity for the apo-form as it has for the off-form. However, the authors find that Faboff has reduced affinity for the apo-form relative to the off-form. This indicates that the apo-form is an ensemble of conformational states, most of which more closely resemble the off-form than the on-form. Consistent with this view, the authors found that they can activate or inhibit apo-form activity by adding Fabon or Faboff, respectively.

By contrast, IgGon did not bind detectably to a model of the pro-form, suggesting that this form’s conformational ensemble contains no members that are close in structure to the active form. The weak affinity of IgGoff for the pro-form suggests that there are substantial differences between this conformation and the off-form as well.

At this point we know that IgGon will bind tightly to the on-form of caspase-1, weakly to the apo-form, but not to the pro-caspase. This means it will likely be an effective probe of active caspase-1 in cells. The authors performed this experiment in THP-1 cells that they differentiated into macrophages. While IgGoff produced diffuse fluorescence in these cells, IgGon stained small, concentrated bodies in a fraction of the cells. This suggests that active caspase-1 is localized to supramolecular structures in these cells, which the authors argue are identical to a structure previously identified as the “pyroptosome”.

Although this particular experiment took advantage of binding-site cysteines that are particular to caspase-1, it should be possible to extend this approach to other proteins. Even non-covalent inhibitors or activators should be useful in this approach as long as the concentration is held high enough to saturate the target site during the selection step. Of course, the conformational change must alter the structure enough that the antibodies have something to grasp — it may not be possible to get specific antibodies if the shift is too subtle. If this requirement is met, however, it should be possible to determine the distribution of specific conformational states in cells, or even (as the authors suggest) to use antibodies as activators or inhibitors in vivo.

J. Gao, S. S. Sidhu, J. A. Wells (2009). Two-state selection of conformation-specific antibodies Proceedings of the National Academy of Sciences, 106 (9), 3071-3076 DOI: 10.1073/pnas.0812952106