Intrabodies are recombinant antibodies that are generated using mRNA display (1), an in vitro selection technique developed by our collaborator, Richard Roberts, in combination with a novel in vivo selection procedure developed in my laboratory. mRNA display can be used to select proteins that bind to targets at picomolar affinity and with high specificity from large random libraries (2). Furthermore, selections produce not only protein binders, but the genes that encode them.
However, although molecules selected by mRNA display have been shown to bind targets in vitro or in an extracellular context (3), they tend not to fold properly and/or aggregate in an intracellular environment. To overcome this limitation we developed a novel in vivo test, based on protein localization, that can be used to screen binders selected by mRNA display to identify those that also work well intracellularly. We take DNA encoding protein binders selected by mRNA display, engineer fusions with GFP and express the GFP-tagged binders in COS cells. To assess binding capability we co-express the target protein from the mRNA selection fused to a Golgi-targeting signal (4). Approximately 1 out of 10 binders colocalize with target protein at the Golgi indicating that they (1) fold correctly and remain structurally stable (2) bind to target at high affinity (3) do not bind appreciably to non-target proteins, all in an intracellular environment. Using mRNA display in combination with our in vivo screen we have successfully generated four different intrabodies that bind to endogenous targets (PSD95, Gephyrin, CAM Kinase IIα, or Kv4.2) with very high affinity and specificity when expressed in native cells. Because PSD95 and Gephyrin are markers of synaptic strength, intrabodies against these proteins allow the distribution of synaptic inputs to individual neurons to be mapped in real time, in individual neurons in vivo.
We are currently developing and optimizing ablating intrabodies- tools for manipulating the ubiquitin/proteasome pathway to induce degradation of specific proteins in a fast and efficient manner. Ablating intrabodies are fusions between intrabodies, which are recombinant antibody-like proteins, and E3 ligases, which mediate the transfer of ubiquitin onto target proteins causing them to be degraded (Fig. 1). Ablating intrabodies mediate the direct degradation of proteins, unlike RNAi or gene deletion strategies, which work at the level of nucleic acids. Thus, degradation mediated by ablating intrabodies has a number of unique properties that could be very useful. For instance, it is not necessary to wait for protein turnover, as with siRNA or gene deletion, and thus protein ablation can potentially be very fast. Also, because ablating intrabodies are bifunctional molecules, they can be made inducible by splitting them into their component parts and using a rapamycin-based system to inducibly join them together (Fig. 2). Such inducible ablating intrabodies combined with light-activatable rapamycin could be used to exert precise spatial and temporal control over protein degradation. We are on developing ablating intrabodies against PSD95 and Gephyrin using intrabodies that we currently have in hand that identify these proteins with very high affinity and specificity.
Such tools will enable us to manipulate individual excitatory and inhibitory synapses in neurons in vivo using photoactivation. Furthermore, the methodology that we are testing can theoretically be applied to virtually any protein, and thus has the potential to make a major impact on the neurosciences and beyond.