Structures of topoisomerase V in complex with DNA reveal unusual DNA-binding mode and novel relaxation mechanism

Topoisomerase V is a unique topoisomerase that combines DNA repair and topoisomerase activities. The enzyme has an unusual arrangement, with a small topoisomerase domain followed by 12 tandem (HhH)2 domains, which include 3 AP lyase repair domains. The uncommon architecture of this enzyme bears no resemblance to any other known topoisomerase. Here, we present structures of topoisomerase V in complex with DNA. The structures show that the (HhH)2 domains wrap around the DNA and in this manner appear to act as a processivity factor. There is a conformational change in the protein to expose the topoisomerase active site. The DNA bends sharply to enter the active site, which melts the DNA and probably facilitates relaxation. The structures show a DNA-binding mode not observed before and provide information on the way this atypical topoisomerase relaxes DNA. In common with type IB enzymes, topoisomerase V relaxes DNA using a controlled rotation mechanism, but the structures show that topoisomerase V accomplishes this in different manner. Overall, the structures firmly establish that type IC topoisomerases form a distinct type of topoisomerases, with no similarities to other types at the sequence, structural, or mechanistic level. They represent a completely different solution to DNA relaxation.


32
The topological state of DNA in cells is regulated by the action of DNA topoisomerases (Bush,  are not only involved in DNA repair but may also serve as a processivity factor by encircling the DNA.

94
Overall, the structures demonstrate a highly unusual DNA/protein complex that helps explain many of the     Figure 1A). The crystals of the complex formed by Topo-97(∆RS2) with a 38 129 bp asymmetric DNA oligonucleotide containing one abasic site diffract to a 3.24 Å in the best direction but 130 only to ~3.9 Å in the other directions (Table II). This anisotropy resulted in an electron density map of 131 uneven quality and although diffraction extends to higher resolution, the map mostly corresponds to a 132 medium resolution structure. Nevertheless, the placement of the protein molecules was unambiguous and 133 a previously unseen region, repeat 7 and Linker helix II, linking repeats 7 and 8, could be seen clearly 134 (Figure 1). The region was built with the aid of the structure from the symmetric complexes, which diffract  or fully disordered. In the complex structure this region is ordered in one of the monomers and shows that 148 the seventh repeat forms an (HhH)2 domain followed by a long helix (Linker helix II) that connects it to the 149 eighth (HhH)2 domain. This long helix sits in the major groove of one DNA molecule, making the (HhH)2 150 domains wrap around the DNA ( Figure 1C). Aside from the ordering of this region and the swiveling of the 151 last three (HhH)2 domains, no other major changes are seen in the protein with respect to the structure of 152 the same fragment in the absence of DNA (Rajan et al., 2016). Thus, in the closed conformation the main 153 change in the protein structure is the movement of (HhH)2 repeats 8-10 to enclose the DNA and the ordering 154 of the seventh domain and the linker region.

155
The protein binds the two DNA molecules through the (HhH)2 domains. One of the DNA molecules

161
but mostly having canonical DNA parameters. Interestingly, the same DNA sequence without the abasic 162 site also crystallized under the same conditions, suggesting that the presence of the DNA abasic site had 163 no effect on the binding of the protein or the conformation of the DNA. Not surprisingly, the DNA abasic site 164 was not apparent in the structure and does not appear to cause any deviations from canonical B-DNA.

165
Overall structure of topoisomerase V with symmetric DNA

166
The use of a symmetric DNA with two abasic sites resulted in a new crystal form with one DNA 167 molecule bound by two Topo-97(∆RS2) monomers (Figure 2 -Figure Supplement 1). Different lengths 168 of DNA were tried to improve the crystal quality, but all of them suffered from anisotropic diffraction. The 169 best diffraction was from crystals with a 40 bp DNA oligonucleotide with 2 base overhangs at each end 170 (Materials and Methods). These crystals diffract to 2.92 Å in the best direction but only to ~3.5 Å in the 171 other directions (Table III). Crystals with 39 or 40 bp DNA and including 2 base overhangs were also 172 anisotropic, but served to provide information on the path of the DNA.

173
In the structure, one DNA molecule is surrounded by two protein monomers through the (HhH)2 174 domains (Figure 2A, 2C). The two proteins in the asymmetric unit are very similar in conformation and are 175 related by an almost perfect non-crystallographic two-fold axis (rmsd between Cα ~ 1.0 Å). For this reason, 176 only one monomer is described hereafter. The (HhH)2 domains can be divided into two subsets, (HhH)2 177 domains 1 to 7 and 8 -10. Each (HhH)2 domains subset has the same conformation as in the DNA-free

190
The structure with a 38 bp oligonucleotide revealed that the non-crystallographic two-fold axis

208
The bending around the abasic site allows the DNA to approach the active site (Figure 2). In

218
The (HhH)2 domains surround most of the DNA coming in close contact with the phosphate 219 backbone at many points. They form a loop with a positively charged interior that accommodated the DNA, 220 but appears to have few close contacts (Figure 2). There are no contacts of the (HhH)2 domains with the 221 bases, only with the backbone, which is not unusual for a non-sequence specific DNA binding protein.

222
Repeats 3-6 and repeat 9 make contacts with the phosphate backbone while the rest of the repeats only 223 surround it, but do not come close to it. Interestingly, even though many of the repeats do not contact the 224 DNA directly, they have positively charged residues facing the DNA, creating an overall positively charged 225 environment around the DNA (Figure 2).

226
The abasic sites are not in direct contact with the protein even though the abasic sites were 227 introduced as a possible target for the single intact repair (HhH)2 domain. Repeat 6 contains an AP/dRP 228 lyase active site that includes lysines 566, 570, and 571; mutations of any of these three residues is 229 deleterious for activity (Rajan et al., 2013). In the complex structure, lysines 570 and 571 face the DNA

235
The active site is exposed in the structure and reveals a highly positively charge region where the 236 DNA enters. As mentioned above, the DNA in this region melts due to a sharp bend. This groove is narrow

257
Arg131 and Arg144 both contact the P-1 phosphate, the phosphate group immediately 5' of the scissile one 258 (Figure 4). Their orientation is such that they could also contact the P0 phosphate if it was present as they 259 are in the region between the two phosphates. His200 and Glu215 are hydrogen bonded to each other. In 260 the complex structure, they are too far to contact the DNA directly, but Glu215 could contact the P0 261 phosphate with minimal side chain rearrangements. Finally, Lys218 is not making contacts with the DNA 262 but is close enough to Tyr226 that it could contact the P0 phosphate. Thus, all residues that have been

265
It is not surprising that the only major change around the active site between the free protein and 266 the protein/DNA complex is the movement of the HhH2 domains. At this resolution, it is not possible to 267 discern large changes in conformation in any of the side chains around the active site tyrosine or in the 268 nearby regions contacting the DNA. It appears that the active site is largely preassembled, even though in 269 the closed conformation it is buried. It is likely that some changes may occur during cleavage and religation, 270 but these changes do not have to be major.

272
The general outlines of the DNA binding mode and relaxation mechanism is understood for many 273 topoisomerases (Bush et al., 2015;Corbett & Berger, 2004), but this is not the case for type IC enzymes.

274
In the latter case, the absence of information on the way they recognize and bind DNA has limited our 275 understanding of the mechanism of this topoisomerase subtype. In addition, the lack of sequence or 276 structural similarities between topoisomerase V and other topoisomerases meant that it was not possible 277 to deduce the DNA binding mechanism based only on similarities. The presence of multiple (HhH)2 domains 278 as well as biochemical information suggested that these domains were involved in DNA binding (Belova et

299
The overall DNA binding mode by topoisomerase V is unusual. The (HhH)2 domains surround the 300 DNA loosely covering almost four helical turns. Topoisomerase V clamps around DNA by having two sets 301 of tandem repeats of (HhH)2 domains that follow the path of DNA in opposite directions. The region between 302 repeats 7 and 8 serves as the turning point to permit the repeats to change direction and encircle the DNA.

303
The Topo-97(∆RS2) fragment is missing the last two (HhH)2 domains plus a few amino acids of unknown 304 structure at the C-terminus. Based on the structures described here, it is likely that the last two (HhH)2 305 domains continue the same path as the previous three and interact with the topoisomerase domain and

325
The active site of topoisomerase V had not been observed in an accessible conformation before.       Redinbo et al., 1998). Single stranded DNA binding and recognition is part of the type IA mechanism, 381 but these enzymes work by an enzyme bridged strand passage mechanism, which is fundamentally 382 different from the swiveling mechanism employed by type IB and IC enzymes. These observations suggest 383 that despite the apparent similarities, type IC enzymes employ a different relaxation strategy. Unlike type 384 IB enzymes, type IC molecules bend the DNA to create a single stranded region that is likely to facilitate 385 swiveling by freeing the two DNA strands around the cleavage site. It is not clear whether the rotation of 386 the strands involves only rotation of the strands or also movement of the topoisomerase domain; it is 387 possible that the topoisomerase domain moves as the strands rotate (Figure 5). Similar to type IB enzymes, 388 the broken strand is captured after swiveling around the other strand. Finally, it is interesting to note that 389 whereas type IB enzymes surround the DNA during the reaction, type IC enzymes do not appear to do so.

390
The (HhH)2 domains surround the DNA, but their role seems to be to act as a processivity factor as these

426
The abasic sites are denoted by a B. They correspond to a tetrahydrofuran abasic site to mimic an 427 apurinic/apyrimidinic (AP) site.

429
Individual oligonucleotides were resuspended to 1mM in water. For annealing, complementary 430 oligonucleotides were mixed at an equimolar ratio, heated to 85°C for 2.5 minutes, cooled down to 5°C 431 below their calculated melting temperature for 5 minutes, and then transferred to ice for at least 10 minutes.

432
Annealed oligonucleotides were used directly in crystallization experiments.

451
For data collection, crystals were first transferred to cryoprotectant (see Table I Table II) (Table IV).