Figure 6.  Structural basis for the selective excision of AZTMP.  AZTMP is excised much more efficiently than other thymidine analogs (d4TMP, ddTMP).  We proposed that the long azido of ATTMP group interferes with the formation of a stable closed complex with an incoming dNTP.  This would give an AZTMP-terminated primer better access to the N site, thus enhancing excision.  The available biochemical data supports this model; the figure presents the structural data that support the model.  (A) Comparison of the structure of the ternary complex (Huang et al., Science 282:1669-1675, 1998) with an AZTMP-terminated primer bound at the N site.  The end of the primer and the incoming dNTP are shown in gold for the ternary complex.  The AZTMP-terminated primer is shown in purple.  The active site aspartic acid side chains (110, 185, 186) are shown, as in the side chain of the glutamine, at position 151.  Overall, the structures are similar and AZTMP is, as expected, well accommodated at the N site.  Formation of the phosphate bond brings the 5'OH of the incoming dNTP closer to the nucleotide in the P site; incorporation appears to cause inversion of the alpha phosphate.  (B) Comparison of the structures of RTs in complexes with a DNA primer with a normal nucleotide at the 3' ends (shown in green) or AZTMP (shown in white).  The 3' ends of both primers are in the P site.  The side chains of two of the active site aspartic acids (110, 185) are shown.  The long azido group of AZTMP lies close to these amino acids.  The AZTMP is rocked up, presumably to avoid a steric clash, which moves the pseudo ribose ring of AZTMP out of the normal position, and puts the thymidine out of position as well.  (C) Model showing potential steric conflict of an incoming dNTP at the N site of a complex with an AZTMP-terminated primer at the P site.  The incoming dNTP is shown in blue, with the three phosphates in orange.  The side chains of 110 and 185 are shown.  The AZTMP is shown both as backbone (white) and space filling (yellow).  Because the AZTMP is not in the normal position for a nucleotide at the P site (see panel B), docking an incoming dNTP causes steric conflicts (shown in purple) involving the azido group and the alpha phosphate of the incoming dNTP and the bases of the incoming dNTP and AZTMP.  These steric conflicts would explain why it is difficult to form a stable closed complex if the primer strand is terminated with AZTMP.