The focus of this study, on the importance of loop 1 on pol λ’s substrate specificity, largely derives from previous studies indicating that loop 1 in the palm subdomain of TdT and pol µ is an important determinant of DNA substrate specificity during NHEJ of double-strand breaks in DNA. Loop 1 in TdT is long (20 residues) and has been suggested to preclude binding of the template strand and thereby contribute to its template-independent polymerization activity (19
). Pol µ also contains a long loop 1, deletion of which abrogates its ability to catalyze template-independent synthesis as well as template-dependent extension of a primer terminus lacking its template partner during a reconstituted NHEJ reaction (13
). Nonetheless, pol µ lacking loop 1 still retains the ability to perform more standard template-dependent gap-filling synthesis (e.g. when the primer is paired with its complementary template base), either alone or with its partners in NHEJ (13
). Thus, loop 1 is not required for all template-dependent synthesis activity by pol µ, but it is an important determinant of the substrate specificity of pol µ. The present study leads to a similar general interpretation on the role of loop 1 in human pol λ, namely that loop 1 is not required for template-dependent synthesis, but it does have an important role in defining substrate specificity. Thus, wt pol λ and its loop 1 deletion derivative have similar kinetic constants for correct dNTP insertion (), similar specific activity with activated DNA, similar ability to extend a simple primer-template (), and similar abilities to participate in NHEJ (). If anything, the loop 1 deletion derivative performs slightly better than wt pol λ in the latter three assays. This may partly be a consequence of its slightly higher processivity (B), a property suggesting that loop 1 may modulate movement of the template strand during catalytic cycling (see below), thus affecting translocation. However, in a manner generally akin to loop 1 in pol µ, loop 1 in pol λ also appears to have an important role in determining substrate specificity. This is because the loop 1 deletion enzyme is considerably less accurate than wt pol λ, as evidenced by the increased diversity of DNA products generated during NHEJ ( and ), by the increased rate of misinsertion of dGTP opposite template T (), and by the increased error rates for a variety of errors (), especially all 12-single base–base mismatches (). The rate at which pol λ DL stably incorporates certain of these single base–base mismatches into duplex DNA approaches 1%. This rate is similar to the rate at which wt pol λ generates certain single-base deletions (39
), an observation that motivated successful efforts to obtain crystal structures of pol λ bound in a catalytically competent conformation to misaligned template-primers (42
). Thus, the high error rates in suggest that it might be possible to obtain crystal structures of ternary complexes of pol λ DL protein containing mismatches in the nascent base pair binding pocket and/or at the primer terminus.
A previous study (43
) showed that, when bound to Bacillus stearothermophilus
DNA polymerase I large fragment (family A), the 12 single base–base mismatches are structurally diverse. It is therefore somewhat surprising that pol λ DL has increased error rates for all 12-single base–base mismatches (), which indicates that pol λ DL can accommodate, and relatively efficiently extend from, all 12 mispairs. Furthermore, the fact that modulation of base substitution fidelity by pol λ DL is not mispair specific is consistent with the idea that truncation of loop 1 likely does not alter the shape of the nascent base pair binding pocket [as, for example, is the case with mutator pol β R283K (44
) or R517K pol λ (Bebenek,K., Garcia-Diaz,M. and Kunkel,T.A., unpublished data)]. Instead removal of loop 1 may allow the enzyme to more readily achieve the active conformation under less than ideal conditions, thus increasing a chance of misincorporation.
Clues to how the loop 1 deletion results in such a general loss of discrimination come from considering loop 1 interactions during the pol λ catalytic cycle. In a binary complex of wt pol λ bound to a gapped primer-template (1XSL), prior to binding of a correct dNTP (22
), β3, β4 and the hairpin are stabilized by a network of hydrogen bonds. These include interactions between two residues of the hairpin: N467 and Q470 and H530 (2.6 Å) and E498 (2.9 Å), respectively. In addition β3, which is positioned parallel to the template strand, makes van der Waals contacts with the backbone of the template upstream of the primer terminal base pair. Upon dNTP binding, however, the switch from an inactive to active conformation involves partial unraveling of the β strands and displacement of the loop to permit the template strand to adopt its catalytic conformation. It is tempting to speculate that this loop is part of a system of checks and balances that controls insertion fidelity. The energetic penalty of relocating the loop would serve to preserve an inactive conformation, and would be generally only overcome upon binding of a correct nucleotide. Thus, the loop deletion, by lowering the energetic penalty of adopting an active conformation, eliminates one of the checks and facilitates the adoption of an active conformation, thus reducing the discrimination between correct/incorrect insertions. This would explain the general increase in base substitution mutagenesis, and would constitute a clear example of how actions ‘at a distance’ (i.e. modifications distant from the active site that influence active site function) can affect the fidelity of a polymerase.
The differences observed for K472 between the binary complexes of wt pol λ and pol λ DL described in ‘Results’ section are consistent with the possibility that K472 may be involved in modulating the activity and the fidelity of DNA synthesis by pol λ. K472 is conserved in family X polymerases (), as a lysine in TdT and pol β, and as an arginine in pol µ (Arg387). Recently Andrade et al.
) reported that substituting Arg387 with Ala resulted in a loss of template-independent synthesis, whereas a change to Lys increased this activity. Thus, Arg387 plays a key role in modulating template-independent synthesis by pol µ. Substituting the homologous lysine in TdT with arginine or alanine (20
) also results in loss of template-independent activity, although the properties of the two TdT mutants are not identical. These results and the structures of TdT and pol µ led Andrade et al.
) to propose that Arg387 stabilizes the position of the primer terminus and, through its interaction with the primer strand, controls the repositioning of the primer terminus upon dNTP incorporation to allow binding and incorporation of the subsequent dNTP during template-independent synthesis. According to their model, the repositioning of the primer is rate limiting for template-independent synthesis. Substituting Arg387 with Lys may allow the primer terminus to more readily adopt the catalytically active conformation. Our results suggest that K472 may help to modulate template-dependent synthesis. In the wt pol λ binary complex (1XSL), K472 is within H-bonding distance of the 3′-O of the primer terminal nucleotide. A hydrogen bond between K472 and the primer terminus that could stabilize the inactive conformation would need to be disrupted in order for the 3′-O to assume its catalytically competent position. A weakened interaction between K472 and the primer terminus, as might be the case in pol λ DL, would allow the 3′-O to more easily adopt a conformation that would support catalysis with an incorrect nucleotide bound, reducing the discrimination between correct and incorrect incorporation, as observed. Finally, the possible role of K472 is consistent with a quantum mechanical/molecular mechanical study of the catalytical mechanism of pol λ, indicating this lysine as one of the residues important for catalysis (46