Many HJ resolvases, including RuvC, cleave HJs by two consecutive, but uncoupled, strand cleavages 
. To ensure that bilateral strand cleavage is achieved within the lifetime of a single resolvase-HJ complex, incision of the second strand is accelerated compared to that of the first strand. One possible explanation for this is that cleavage of the first strand is slowed by the need to distort the junction in order to position the scissile bond within the active site 
. However, once the first strand is cleaved, the junction becomes more flexible, and therefore the second scissile bond can be more readily located into the resolvase's active site. One of the supporting pieces of evidence for this is that the rate of strand cleavage in a nicked junction, containing a consensus RuvC cleavage site, is 8-fold higher than in the same junction without a nick 
. Our own data shows that the rate of cleavage of M1n is ~2.4 fold higher than M1, which is consistent with the idea that increased junction flexibility aids second strand cleavage. However, the rate of cleavage is increased by a further ~3.1 fold if the nick contains a 5′ phosphate. In the case of X0, which contains a very poor RuvC cleavage site, the effect is even more dramatic with the 5′ phosphate stimulating cleavage by more than 50 fold compared to the same junction without a 5′ phosphate. These data suggest that the acceleration of second strand cleavage during the resolution of a HJ by RuvC depends to a large extent on the exposure of a 5′ phosphate.
Why is the 5′ phosphate critical for accelerating second strand cleavage by RuvC? One possibility is that a nicked HJ without a 5′ phosphate is not as flexible as one with a 5′ phosphate. We are unaware of any study that has directly addressed this possibility, however it has been reported that the phosphates at the centre of an intact HJ influence junction conformation 
. However, we think that the presence of a 5′ phosphate is unlikely to have any major effect on the flexibility of a nicked HJ. Certainly it does not improve the efficiency of X0n cleavage by Mus81, which is thought to require considerable junction flexibility for proper complex formation 
A second possibility is that the 5′ phosphate provides a molecular “handle” for RuvC to interact with thereby enabling it to influence junction conformation in a way that enhances second strand cleavage. This would be analogous to another member of the RNase H/Integrase superfamily, Tn5 transposase, which interacts with the 5′ phosphate exposed by hairpin cleavage during the transposition reaction 
. The coordination of the 5′ phosphate involves residues of the (R)YREK motif that is common to the IS4
transposase family, and stabilizes a DNA conformation that dramatically enhances strand transfer of the donor DNA into the target by promoting target DNA capture and/or the strand transfer reaction itself 
A third possibility is that the putative interaction between RuvC and the 5′ phosphate generates a conformational change in RuvC itself that, together with the additional flexibility of the nicked HJ, stimulates second strand cleavage. Here we imagine that charge repulsion or attraction between the exposed phosphate and residue(s) in the active site of the first monomer might help to promote a conformational change, which could in some way be relayed to the active site of the second monomer aiding its interaction with the scissile bond. Indeed the idea that a conformational change in one subunit can be relayed to a second subunit has been mooted to explain the enhancement of second strand cleavage by the HJ resolvase Ydc2 
. Structural studies of RuvC and its interaction with nicked HJs with and without a 5′ phosphate will be needed to determine whether or not the phosphate promotes protein and/or DNA conformational changes that can account for the dramatic stimulation of second strand cleavage during junction resolution.
In contrast to RuvC, Mus81 does not need a 5′ phosphate at the nick site to stimulate its ability to cleave nicked HJs. The presence of the nick itself regardless of its terminal chemistry seems to be sufficient for optimal cleavage efficiency. Recently we showed that nicked HJs are bound with higher affinity than intact HJs in the presence of a relatively low concentration of divalent metal ion 
. Similar to cleavage efficiency, the binding affinity of Mus81 for nicked HJs is unaffected by the presence of a 5′ phosphate at the nick site. This correlation between binding affinity and cleavage efficiency contrasts with RuvC, which binds equally well to intact and nicked HJs (with and without a 5′ phosphate at the nick site) even though optimal cleavage of a nicked HJ depends on the presence of a 5′ phosphate at the nick site. We suspect that optimal binding and cleavage by Mus81 simply requires a junction with the level of flexibility that is achieved by the presence of a strand nick at or close to the junction crossover point.
Although the presence of a 5′ phosphate at the nick site of a nicked HJ has no effect on the activation of Mus81 cleavage, it does influence cleavage site selection. A recent model of the Mus81-Eme1-nicked HJ complex shows how the exposed 5′ DNA end may be close to residues in and around helix 5 of Mus81 
. These residues include a conserved aspartate, and therefore it is possible that charge repulsion could cause movement of the 5′ side of the nick away from helix 5, which in turn would “drag” the cleavage site further from the junction crossover point.
In this study we have shown that the presence of a 5′ phosphate at the strand discontinuity in a nicked HJ plays an important role in stimulating junction cleavage by RuvC. From this we conclude that the acceleration of second strand cleavage during HJ resolution by RuvC is not solely promoted by increased junction flexibility caused by incision of the first strand as previously proposed 
. Whether a 5′ phosphate is similarly important for efficient bilateral strand cleavage by other HJ resolvases is yet to be determined. However, our observation that Mus81 cleaves nicked HJs with and without a 5′ phosphate with equal efficiency suggests that at least in some cases a nick may only be needed to impart junction flexibility.