The age of reverse biochemistry

In any field of science, new information comes in cycles of feast and famine. Knowledge is harvested steadily, until an impasse is encountered for lack of technological capabilities. Upon development of new methodologies, a flood of new information emerges. It appears that we are witnessing a revolution in biochemistry with new approaches in structural biology. Traditional biochemical approaches entail identification of a cellular function of interest, purification of the protein(s) responsible for the function, and characterization of its activity in vitro, culminating with solution of the atomic structure by X-ray crystallography or nuclear magnetic resonance spectroscopy. With a proliferation of structural biologists and the exponential growth in numbers of structures now available in databases, a “reverse biochemistry” approach to identification of protein function is now possible. Computational tools are now available for scanning a three-dimensional structure against databases to identify similarly structured proteins. A beautiful example of this approach is reported by Robert Liddington and colleagues in this issue of Protein Science (Aoyagi et al. 2007). Liddington's group crystallized the vaccinia virus N1L protein, which was known to be important for viral virulence (Kotwal et al. 1989), but its function was unknown. It has no significant amino acid sequence similarity to any other protein. Comparison of the N1L structure with that of proteins in the DALI server database (Holm and Sander 1993) revealed a striking similarity to the bcl-2 family of proteins that function in modulating cellular apoptosis, or programmed cell death. A bcl-2-like function was confirmed by binding assays with peptides corresponding to interaction domains on the pro-apoptotic BH3 proteins. When a virus invades a cell, a battle ensues. A successful virus infection will result in virus multiplication, allowing the virus to spread to other cells. Higher eukaryotic cells mount several different defenses aimed at slowing or stopping the replication of the virus by the auto-destruct process of apoptosis. The cell attempts to block virus replication by dismantling its own structure (Holm and Sander 1993). The triggering of apoptosis results in inhibition of protein synthesis through a pathway regulated by the double-stranded RNA activated protein kinase, PKR. Mitochondrial function is destroyed in a process requiring the Bak and Bax proteins. Inner mitochondrial membrane potential is lost and the SMAC, TtrA2, apoptosis inducing factor, endonuclease G, and cytochrome c proteins are released from the organelle (Benedict et al. 2002), presumably by opening of the mitochondrial transition pores. The bcl-2 proteins are believed to act as signal transduction checkpoint regulators of mitochondrial destruction by modulating the activity of Bak and Bax proteins by heterodimerization (Wang 2001). The release of cytochrome c leads to the activation of caspase proteases that proceed to cleave the nuclear matrix proteins to dismantle the nucleus. Not to be outdone, viruses have the ability to counter all of these defenses. For maximal production of eukaryotic viruses, the infection generally must proceed for several days. Therefore, it is not surprising that the virus must develop some means to block apoptosis if high yields of progeny are to be produced. Another vaccinia virus protein, the F1L gene product, was previously reported to have bcl-2-like activity (Wasilenko et al. 2003). This activity was discovered through a more traditional approach of screening viruses that have genome deletions that allow increased apoptosis during infection. The vaccinia N1L protein bears six of the seven signature α-helices characteristic of bcl-2 proteins in a tight bundle. The N1L helices are marginally smaller than those of other bcl-2 proteins; however, minimized size relative to cellular counterparts is a feature common to many vaccinia virus proteins. The amino acid sequence of N1L has barely any detectable similarity to that of other family bcl-2 family members. A forced fit reveals only 11% amino acid identity with other members. Such extreme divergence of vaccinia proteins from cellular counterparts is not without precedent. The vaccinia DNA topoisomerase is structurally similar to other DNA recombinases, yet only the active site residues of the enzyme are conserved (Cheng et al. 1998). It would seem that the lesson here is that there is more than one way to build a protein.