How sequence alterations enhance the stability and delay expansion of DNA triplet repeat domains

Abstract

DNA sequence alterations within DNA repeat domains inexplicably enhance the stability and delay the expansion of interrupted repeat domains. Here we propose mechanisms that rationalise such unanticipated outcomes. Specifically, we describe how interruption of a DNA repeat domain restricts the ensemble space available to dynamic, slip out, repeat bulge loops by introducing energetic barriers to loop migration. We explain how such barriers arise because some possible loop isomers result in energetically costly mismatches in the duplex portion of the repeat domain. We propose that the reduced ensemble space is the causative feature for the observed delay in repeat DNA expansion. We further posit that the observed loss of the interrupting repeat in some expanded DNAs reflects the transient occupation of loop isomer positions that result in a mismatch in the duplex stem due to 'leakiness' in the energy barrier. We propose that if the lifetime of such a low probability event allows for recognition by the mismatch repair system, then 'repair' of the repeat interruption can occur; thereby rationalising the absence of the interruption in the final expanded DNA 'product.' Our proposed mechanistic pathways provide reasoned explanations for what have been described as 'puzzling' observations, while also yielding insights into a biomedically important set of coupled genotypic phenomena that map the linkage between DNA origami thermodynamics and phenotypic disease states.

Keywords: DNA conformational space; DNA expansion/contraction; repeat DNA origami; repeat mismatch repair; stability-functional correlations.

Figure 1.

(a) Cartoon version of a repeat bulge loop within a larger repeat domain. Such a static representation does not highlight the reality that such repeat bulge loops really are dynamic ensembles of loop isomers as indicated in (b) (Völker et al., 2012). Note that in these cartoon representations, we present the repeat bulge loop as a ‘unstructured ring’ to account for the (largely unknown) fluctuating microstructures that likely make up the repeat bulge loop ensemble within a given rollamer isomer (Völker et al., 2008).

Figure 2.

Schematic representation of the impact of repeat interruptions on the repeat bulge loop ensemble as shown for the [CAG]₈ˑ[CTG]₄ system containing a CAA interruption in place of the 6^th CAG repeat (Mutated base pair indicated in red letters). The [CAG]₈ˑ[CTG]₄ complex results in a 4-repeat bulge loop that can be positioned in 5 possible loop positions, identified by roman numerals I-V, in the 5′ to 3′ directions. The colour coding of the CAG repeat segments into yellow (repeats 1–4) and red (repeats 5–8) is meant as a visual aid to help identify which repeat is partitioned into which domain in each of the 5 loop isomers. Loop position determines whether the CAA repeat is part of the duplex or the loop domain with the blue ball highlighting the position of the mutated A in each loop isomer. When the CAA triplet is partitioned into the loop domain, the complementary TTC triplet in the opposing strand forms base pairs with a CAG repeat that is part of the upstream duplex region, resulting in a GˑT mismatch (green ellipsoid). Note that the different loop isomers can be classified into 3 general groups; as defined by the energetic impact of the repeat interruptions on the repeat bulge loop isomer; with the differential energetic impacts dictating the differential loop populations. Group 1: isomers I and II contain a conventionally base paired CAA/GTT triplet in the downstream duplex domain, (potential loop Isomers with the CAA/GTT triplet in the upstream duplex domain are not shown in this example but can be considered essentially equivalent); Group 2: Loop isomers IV and V contain a mismatched triplet CAG/GTT in the upstream duplex domain. Analogous to group 1 above, potential loop Isomers with the mismatch in the downstream duplex domain are not shown, but also can be considered essentially equivalent; and Group 3: Loop isomer III contains the GˑT mismatch at the 5′ junction and we propose that it likely is part of an expanded loop domain. Note that a mismatch at the 3′ junction is possible, depending on the nature of the repeat interruption, as defined by the triplet sequence. An altered base in 1^st, 2^nd, or 3^rd positions has unique impacts at the 3′ and 5′ junctions, but only impacts group 1 and 2 insofar as it alters nearest neighbours in the duplex. Fig. 2 is intended as an illustrative example. Longer repeats, larger slip outs, and repeat interruptions at different positions will produce different loop isomer arrangements, but conceptually they are represented by the three group classifications shown in Fig. 2.