Aptamers (protein binding oligonucleotides) have potential as a new class of targeted therapeutics. For applications requiring chronic systemic administration, aptamers must achieve high-affinity target binding while simultaneously retaining high in vivo stability, tolerability, and ease of chemical synthesis. To this end, we describe a method for generating aptamers composed entirely of 2'-O-methyl nucleotides (mRmY). We present conditions under which 2'-O-methyl transcripts can be generated directly and use these conditions to select a fully 2'-O-methyl aptamer from a library of 3 x 10(15) unique 2'-O-methyl transcripts. This aptamer, ARC245, is 23 nucleotides in length, binds to vascular endothelial growth factor (VEGF) with a Kd of 2 nM, and inhibits VEGF activity in cellular assays. Notably, ARC245 is so stable that degradation cannot be detected after 96 hr in plasma at 37 degrees C or after autoclaving at 125 degrees C. We believe ARC245 has considerable potential as an antiangiogenesis therapeutic.

The structures of beta class carbonic anhydrases (beta-CAs) determined so far fall into two distinct subclasses based on the observed coordination of the catalytic zinc (Zn2+) ion. The subclass of beta-CAs that coordinate Zn2+ tetrahedrally with four protein-derived ligands is represented by the structures of orthologues from Porphyridium purpureum, Escherichia coli, and Mycobacterium tuberculosis. Here we present the structure of an additional member of that subclass, that from Haemophilus influenzae, as well as detailed kinetic analysis, revealing the correspondence between structural classification and kinetic profile for this subclass. In addition, we identify a unique, noncatalytic binding mode for the substrate bicarbonate that occurs in both the H. influenzae and E. coli enzymes. The kinetic and structural analysis indicates that binding of bicarbonate in this site of the enzyme may modulate its activity by influencing a pH-dependent, cooperative transition between active and inactive forms. We hypothesize that the two structural subclasses of beta-CAs may provide models for the proposed active and inactive forms of the H. influenzae and E. coli enzymes.