During the past year the Ellington Lab has focused on the design of self-evolving nucleic acid enzymes. We have designed a cross-catalytic amplification system based on the fast and efficient 10-23 deoxyribozyme (Ellington and Levy, 2003). In this system complementary deoxyribozyme cleavases are inactivated by circularization (Figure 1). Linearization results in activation of the deoxyribozyme and leads to the initiation of a cascade of cleavage reactions that displays exponential growth kinetics (Figure 2). This system represents the first in vitro system capable of exponential growth in the absence of protein enzymes.

The conformational change from circular substrate to linear catalyst resulting in exponential growth suggested the potential for the evolution of both sequence and function. To test this hypothesis, we randomized opposing portions of the binding arms generating two pools of deoxyribozymes (Figure 3; CE.1 and CE.2). When incubated together, members of each population that were capable of activating their complements were preferentially amplified. A portion of this reaction was then transferred to a new tube containing a population of un-reacted circular pool. After four rounds of serial transfer, sequence analysis indicated the sequences expected to dominate in the selection had indeed been selected (Figure 4, GTG and TCC). In addition to these two species other variants were also isolated. Examination of these sequences suggested that they might form 2 out of 3 base pairs with their complement. If selection had occurred, then it is expected that these minor variants would also be active catalysts, albeit to a lesser extent. To test this we assayed the alternate variant GGG in exponential growth assay. As expected, the GGG variant is not as good a catalyst as the dominant GTG variant (Figure 5). The successful design and implementation of this scheme for selection via serial transfer represents an in vitro selection experiment conducted in the absence of any protein enzymes. To our knowledge this is the first time this has been accomplished.

Currently we are continuing our work with the cross catalytic clevase cascade. To this end we have designed two new circular deoxyribonucleic acid (DNA) pools (Figure 6). Unlike the previous pool design in which the random region opposed a constant region, the new pools have been designed such that random regions oppose random regions. In vitro selection experiments using these larger and more diverse pools should allow us to begin to elucidate some of the parameters surrounding the evolution of sequence and function in this exponential replicating system.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

As part of our efforts to develop unnatural organisms (currently unfunded, but under the general aegis of the Astrobiology Institute), we have interacted with Dr Chris McKay at NASA-Ames. One of my graduate students, Tim Riedel, spent a year at Ames working with Dr. McKay.