Zinc-finger BED domains drive the formation of the active Hermes transpososome by asymmetric DNA binding

Lannes, Laurie; Furman, Christopher M.; Hickman, Alison B.; Dyda, Fred

doi:10.1038/s41467-023-40210-3

Download PDF

Article
Open access
Published: 25 July 2023

Zinc-finger BED domains drive the formation of the active Hermes transpososome by asymmetric DNA binding

Nature Communications volume 14, Article number: 4470 (2023) Cite this article

1520 Accesses
Metrics details

Subjects

Abstract

The Hermes DNA transposon is a member of the eukaryotic hAT superfamily, and its transposase forms a ring-shaped tetramer of dimers. Our investigation, combining biochemical, crystallography and cryo-electron microscopy, and in-cell assays, shows that the full-length Hermes octamer extensively interacts with its transposon left-end through multiple BED domains of three Hermes protomers contributed by three dimers explaining the role of the unusual higher-order assembly. By contrast, the right-end is bound to no BED domains at all. Thus, this work supports a model in which Hermes multimerizes to gather enough BED domains to find its left-end among the abundant genomic DNA, facilitating the subsequent interaction with the right-end.

Structural insight into Tn3 family transposition mechanism

Article Open access 18 October 2022

Selective TnsC recruitment enhances the fidelity of RNA-guided transposition

Article 24 August 2022

IS21 family transposase cleaved donor complex traps two right-handed superhelical crossings

Article Open access 22 April 2023

Introduction

Transposable elements (TE) are discrete genomic regions that can move from one position to another within genomes. TEs have been found in all eukaryotic kingdoms as both active and inactive forms and can make up a large portion of eukaryotic genomes¹. TEs are a major force in the shaping of genomes and the evolution of species by establishing novel cellular functions and pathways². Furthermore, the origin of some human diseases is linked to TEs^3,4.

Most of the eukaryotic class II or DNA transposons move by cut-and-paste transposition that consists of the excision of the transposon by generating double-stranded breaks (DSB) at the ends of the element and its reintegration elsewhere, typically without specificity (Fig. 1a insert). An autonomous transposon is delimited by two ends featuring terminal inverted repeats (TIRs), with one or more genes between them (Fig. 1b exemplified with Hermes transposon). One of these genes encodes the transposase, an enzyme that mobilizes the transposon by carrying out the necessary nuclease and trans-esterification activities (Fig. 1a exemplified with Hermes)⁵. The transposase typically assembles into a multimer that recognizes and brings together the TIRs by relying on site-specific DNA-binding domain(s) such as helix-turn-helix or zinc-finger domains, which can be either upstream or downstream of the catalytic domain⁶. After the formation of the transposition complex or transpososome, DSBs are created at the ends of the TIRs, liberating the transposon, and the assembly finds its target DNA to integrate the TE.

**Fig. 1: The eukaryotic *Hermes* transposon system.**

The Hermes DNA transposon, belonging to the largest cut-and-paste superfamily called hAT, was isolated from the genome of the housefly Musca domestica⁷. It has been characterized biochemically^8,9, and the three-dimensional structure of its N-truncated form with its catalytic core (hereafter referred to as “∆-BED” Hermes) has been solved (Fig. 1c)^9,10,11. The Hermes transposon is ~3 kb and is delimited by two dissimilar ends of ~450 bp referred to as the left-end (LE) and the right-end (RE)⁷. The ends are capped by two imperfect 17 bp TIRs that differ from each other by two swapped base pairs (Fig. 1b)⁷. The Hermes transposase shows a preference for 8 bp target sites with a 5’-nTnnnnAn-3’ pattern such that staggered integration of the transposon across the target site generates 8 bp target site duplications (TSD) (Fig. 1a)¹².

The Hermes transposase is composed of 612 amino acids and is organized in four domains (Fig. 1c)^7,10,13. In principle, two transposase protomers, each carrying one catalytic domain with its active site, assembled into a dimer should be sufficient to liberate two transposon ends and integrate them into a target site. In fact, a number of DNA transposases from both the prokaryotic and eukaryotic domains appear to work as dimers^{14,15,16,17,18}, suggesting that a dimer is a minimal oligomerization unit. However, there are also a few examples of higher-order oligomers. The bacteriophage MuA transposase forms a tetramer upon synapsing its ends^19,20, whereas the related retroviral integrases can form even higher-order oligomers, at least in vitro^21,22.

Architecturally, the Hermes transposase is unusual as it spontaneously forms a ring-shaped tetramer of dimers (Fig. 1c). The dimers tightly associate using an “intertwined” dimerization domain (DD; Fig. 1c). The quaternary structure is held together by the swapping over of an α-helix from the insertion domain between adjacent dimers^10,11. Deletion of the swapped helix prevents octamer formation, converting Hermes into dimers. Remarkably, these dimers are fully competent for all the enzymatic activities of Hermes in vitro; in fact, they are more active than wild-type Hermes. Yet they are inactive for transposition in cell-based assays¹¹. The available biochemical and structural data provide no explanation for this substantial contradiction between in vitro and in-cell experiments.

All hAT transposases are predicted to contain an N-terminal BED zinc-finger domain organized around a conserved CCHH or CCHC motif that coordinates Zn²⁺¹³, but no structural information is available for Hermes as the constructs used in previous structural work lacked the first 78 amino acids. Nevertheless, this domain is indispensable for the excision step in vitro¹¹. We have previously suggested that the BED domain of the Hermes transposase recognizes short subterminal repeats (STRs) interspersed in its transposon ends¹¹. Even though the chemical steps of transposition occur at the very tips of TEs, several hAT transposons have DNA end requirements that span far beyond the TIRs. For instance, Tol2 needs the terminal 200 and 150 bp of its LE and RE, respectively, for excision and transposition²³; Tag1 requires about 100 bp from both ends to generate comparable transposition rates as the full ends²⁴; and Ac at least 200 bp from both ends²⁵. In the case of Tol2, it has been shown that mutation of STRs reduced the excision activity²³. Given the number of STRs in the Hermes ends, a dimer with only two BED domains does not appear sufficient if the binding of more than two is needed. Therefore, it is plausible to assume that interaction with multiple STRs is needed for in-cell activity, and that a higher-order assembly, such as the octamer with its eight BED domains, would be one way to supply a sufficient number of BED domains. However, the architecture of such an assembly is not known.

In this work, we aimed to decipher the role of the BED domain in transposon end binding and to understand how the transposase interacts with its ends. To explain the striking discrepancy between in vitro and in-cell results, we undertook a mechanistic study using biochemistry, X-ray crystallography, cryo-electron microscopy (cryo-EM), and in-cell transposition assays.

Results

The N-terminal domain of the Hermes transposase is a DNA-binding domain

Although the first 78 amino acids of Hermes that include the BED domain are indispensable for transposition in vivo and cleavage in vitro, they do not directly participate in recognition of the TIR or in the catalytic activity of the transposase^9,11. Despite the critical role of the BED domain, its DNA-binding site has not yet been established. We, therefore, first sought to confirm the DNA-binding activity of Hermes’ BED domain and determine if such activity was specific. We expressed the region 1-78 of the Hermes transposase (hereafter “BED”) in E. coli and purified it (Supplementary Fig. 1a) to perform interaction assays. While the LE and RE of the Hermes transposon are 449 bp and 464 bp, respectively⁷, the proximity in sequence space of the BED domain to the part of Hermes with known three-dimensional structure suggests that DNA binding would likely occur towards the interior portion of the TIR. This notion is also supported by previous activity data using mutated Hermes LE¹¹.

We used a double-stranded DNA (dsDNA) that spanned bp 11-27 of the LE (LE11-27, Fig. 2a) to probe the DNA binding of the BED domain. As a control, a randomized 17-mer (ran17) was used that had no common features with LE11-27. The DNAs were titrated with an increasing amount of protein from one to three equivalents and analyzed by size exclusion chromatography (SEC). The resulting chromatograms are presented in Fig. 2a. Binding was assessed by the shift of the DNA peak, the shape of any resulting complex peak, and whether we observed free DNA or protein as a function of titration. Although ran17 showed evidence for only weak binding by BED, LE11-27 formed a tight and stable complex that saturated at a 1:2 ratio of DNA to protein.

**Fig. 2: The N-terminal BED domain of the Hermes transposase.**

The N-terminal domain of Hermes folds in a CCHC zinc-finger BED motif that specifically interacts with a subterminal repeat

Crystals were obtained of the 2:1 BED/LE11-27 complex that were ultimately optimized to diffract to 2.5 Å, and the structure was solved by zinc multiwavelength anomalous diffraction (Zn-MAD; Supplementary Table 1). The complex crystallized in space group P6₅22 with a crystallographic twofold axis perpendicular to the oligonucleotide that intersected it at its central base pair. Thus, the asymmetric unit contained one BED domain and one DNA strand of the duplex. The blunt-ended double-stranded LE11-27 DNA used in crystallization was not perfectly palindromic and packed end-to-end in the crystal lattice. The crystallographic twofold axis, therefore, related the top half of the oligonucleotide to the bottom half, resulting in electron density that was the superposition of the two crystallographically related orientations. Iterative model building and refinement led to the atomic model that is composed of two crystallographically identical BED domains bound to LE11-27. We modeled LE11-27 in the two overlapping and opposite orientations at 50% relative occupancy. We observed alternate conformations for some of the base-interacting amino acid side chains of the BED domain, consistent with the two orientations of LE11-27. This indicates that the BED domain has the ability to recognize the two halves of the imperfect palindrome by relying on these alternate side-chain conformations. The atomic model for the whole assembly is shown in Fig. 2b.

Each BED domain contains at its C-terminal a CCHC zinc-finger involving residues C51, C54, H71, and C73 that coordinate a bound Zn²⁺ ion. The two BED domains bound to the imperfect palindrome do not contact each other, but each C-terminal α-helix between T65 and R70 (α3; see Fig. 2b) is deeply inserted into the major groove. According to a DALI search²⁶, the only identifiable structural homolog that binds Zn²⁺ is the BED domain of the human transcription factor ZBED2 protein, whose structure was determined by NMR (ZBED2; PDB 2DJR, RIKEN Structural Genomics/Proteomics Initiative, 2006). This protein was recently implicated in the development of lineage plasticity in pancreatic cancer cells²⁷. The summary of the protein–DNA interactions is presented in Fig. 2c.

Key to binding the imperfect palindrome is the ability of the N67/R70 pair of the BED domain to assume alternate conformations. These, together with S66, Q64, and the main chain carbonyl of R63 and T65 at the apical loop (between β2 and α3) of the zinc-finger, form all the base-specific interactions. In one half of the palindrome, N67 of BED1 (shown in yellow, Fig. 2d and Supplementary Fig. 2a with the electron density) interacts with the Hoogsteen face of A18, while R70 interacts with the stacked G17 and T18. In the other half (BED2, shown in green, Fig. 2d and Supplementary Fig. 2a), the position that was occupied by A18 is now replaced by G20, which is less competent to interact with N67. We observe that N67 is completely displaced, no longer facing the DNA but instead interacting with K59 and the phosphate of G19, with R70 now in its place to interact with G19. For both BED domains, other contributions to LE11-27 binding include several nonspecific contacts between protein side chains of Lys, Arg, Ser, Thr, and Trp residues and the DNA phosphate backbone. Thus, the structure shows that the single BED domain of the Hermes monomer is able to form specific interactions with two related but distinct DNA motifs.

Analysis of the DNA parameters by Curves+²⁸ indicated several distortions relative to B-form DNA. The LE11-27 DNA experiences a slight bending of the double helix that is reflected in the positive roll and negative slide of base pair dyads (Supplementary Fig. 2b). The bending is directed towards the edges of the major groove and the C-termini of the BED domains (Fig. 2e). The minor groove shows widening over the base pairs flanking the central base pair (i.e., A16, G17, T18 and G20, C21, T22, respectively) of ~2 Å compared to B-form, whereas its major groove is narrowed around the central base pair before widening ~2 Å at the termini (Supplementary Fig. 2b). On the other hand, the major groove is deeper by ~4 Å over the region A16-T22, while the minor groove is shallower over this region (Fig. 2e). This suggests that BED binding to the imperfect palindrome may be allosteric, in which the insertion of one α-helix into the major groove and the resulting DNA distortion makes the binding of the second BED easier.

The mapping of the STRs

The interactions observed in the crystal structure of the BED/LE11-27 complex suggests that the BED domain binding site is 5’-AAG(T/C)−3’. There are 17 and 14 occurrences of this motif in the LE and RE of Hermes, respectively, concentrated in particular towards the tips of the transposon, with five copies in the first 55 base pairs of both ends (Fig. 2f and Supplementary Fig. 3a). We designate these putative BED binding sites as subterminal repeats (STRs), and number them from “STR1” onward moving in from the LE transposon tip but from STR0 on the RE as the first AAGT motif is within the TIR. With the exception of LE-STR1/RE-STR1 and the pair LE-STR3/RE-STR2, the rest of the STRs of the two ends do not align with each other: their arrangement is asymmetric at the two ends. Interestingly, on the LE, except for STR1, STR2 to STR7 are all oriented antisense, while RE-STR1 to RE-STR5 are all oriented in the sense direction. Beyond LE-STR1/LE-STR2, only one other pair of STRs is organized as a palindrome, LE-STR14/LE-STR15, far interior in the LE beginning at bp 276.

Reconstitution of the Hermes transpososome in vitro

As our attempts to crystallize either the full-length Hermes transposase or its complexes with transposon ends had failed, we turned to cryo-EM to determine the architecture of the transpososome, the assembly containing two bound transposon ends. Although the LE and RE TIRs only differ from each other by two base pairs which do not interact with the protein, yet the transposase has a much higher affinity for the LE compared to the RE^8,11. Therefore, to reconstitute the transpososome in vitro, we evaluated the interaction of various LE-DNA oligonucleotides complexed with the purified transposase to identify the best substrate for structural studies. We tested a range of oligonucleotides that extended the LE-TIR into the transposon, such as LE-TIR + 13 and LE-TIR + 30 as well as those that were nicked and gapped at the non-transferred strand at the cleavage site but also extended into the flanking DNA, i.e., 8 + LE-TIR + 7 and 8 + LE-TIR + 30 (Supplementary Fig. 4a and Fig. 3b). Samples were prepared by mixing purified Hermes transposase with sub-stoichiometric or stoichiometric amounts of DNA and dialyzed to lower salt concentration prior to analysis by SEC and mass photometry (MP).

**Fig. 3: Stoichiometry of Hermes/LE-DNA transpososomes.**

The SEC elution profiles corresponding to the titration of Hermes with the various LE-DNAs are presented in Supplementary Fig. 4b–g. The results were consistent with the crystal structure of the ∆-BED-transpososome¹¹, in that close to one LE-TIR bound per protomer (Supplementary Fig. 4e). However, for longer DNAs, the stoichiometry changed, with a maximum of 0.5 DNA bound per Hermes protomer (Supplementary Fig. 4c, d, f, g). LE-TIR + 13, LE-TIR + 30, and 8 + LE-TIR + 30 led to precipitation of the complex when added to the protein in a stoichiometric 1:1 mix (Supplementary Fig. 4c, d, g), in contrast to 8 + LE-TIR + 7 whose complex remained stable but did not bind more DNA in going from a 1:0.5 to 1:1 mix (Supplementary Fig. 4f).

The peaks corresponding to the protein–DNA complexes were not monodisperse. This was also evident during the purification of the transposase, where we similarly observed heterogeneity and unusual behavior when the samples were analyzed by SEC. As shown in Supplementary Fig. 4h, the purified transposase itself eluted as two populations, P1 and P2, where the relative ratio between the two elution peaks varied as a function of the salt concentration in the buffer, with the higher concentration favoring P1 (a species consistent with an octamer) and lower concentrations promoting P2 (a species consistent with a hexamer). Through a combination of SEC (Supplementary Fig. 4h) and MP analysis (Fig. 3a), we established that the Hermes transposase exists in solution as two oligomeric states composed of six or eight monomers.

We used MP (Fig. 3b) to measure the composition of the DNA-bound complexes observed by SEC. These results were consistent with the SEC data showing that the Hermes transposase assemblies bound fewer molecules as the DNA was lengthened. For example, the MP experiments revealed that LE-TIR was mainly bound to the hexameric form of Hermes and that approximately five DNAs were bound. The longer oligonucleotides also generated more heterogeneous samples with the appearance of several masses under 400 kDa and broad distributions that led to poor fitting (standard error >30 kDa). Nevertheless, under our experimental condition, the Hermes hexamer could bind up to two of either LE-TIR + 13 and 8 + LE-TIR + 30 DNAs; the Hermes octamer bound up to four LE-TIR + 13, and three LE-TIR + 30 and 8 + LE-TIR + 30 DNAs.

Three LE-STRs bind to BED domains within the transpososome

On the basis of the SEC and MP experiments, for cryo-EM studies we selected the 8 + LE-TIR + 30 oligonucleotide as it contained both flanking region and the longest transposon DNA. It formed both hexameric and octameric complexes. As we assumed that in a biologically relevant assembly, two transposon ends are bound, we minimized the loading of the transposase and used a ratio of 1:0.25 protein-to-DNA. We also anticipated that the TIRs would be synapsed inside the same Hermes dimer given that the transposase integrates its transposon across an 8-bp target site to generate 8 bp TSDs, consistent with the distance between the two active sites in one Hermes dimer. The resulting complex was purified by SEC and the fraction with the highest absorbance was used for structure determination.

As shown in Fig. 4a, the 2D classes of the top view of the complex resembled an octamer, as observed in the ∆-BED-transpososome crystal structure¹¹, with resolved features for two transposase dimers facing each other and linked by two equatorial dimers to form a closed ring. While the top views did not give clear indication for the presence of DNA, in the side views, bound DNA was clearly visible. The particles were further cleaned by 3D classification, and the refined maps of their best classes are shown in Fig. 4b, c, respectively (see details in Supplementary Fig. 5). The reconstruction (from two 3D classes representing 37% of the total particles) was composed of a closed octameric ring (Fig. 4c). On the other hand, the shape of the reconstruction presented in Fig. 4b from one 3D class that represented 19% of the particles was consistent with an octameric transpososome that appeared to have lost one of its equatorial dimers as no potential density was observed for it. Importantly, both maps indicated that two DNAs bridged two Hermes dimers, the top dimer A and the bottom dimer C. The maps aligned well and displayed the same features for these two dimers. In contrast, the equatorial transposase dimers (dimers B and D) were poorly resolved and featureless in both reconstructions, likely due to the sparsity of interactions with the DNAs (see local resolution mapping in Supplementary Fig. 5).

**Fig. 4: The cryo-EM structure of the LE-LE *Hermes* transpososome.**

In order to resolve the interactions between the BED domains and DNA, we performed multi-body 3D refinement on the three best 3D classes (regardless of the oligomerization state of the transposase) using masks to divide the transpososome into three bodies. Body 1 was composed of the DNAs, the Hermes dimers A and C and the BED domains originating from the equatorial Hermes dimers bound to the DNAs, and the other two corresponded to the N-truncated Hermes dimers B and D, respectively. The resulting composite map is presented in Fig. 4d. The map for Body 1 was refined at 4.64 Å resolution, while the other two maps were of much lower resolution (Body 2, 10.21 Å; Body 3, 10.94 Å, Supplementary Fig. 6). The resolution of the Body 1 map was sufficient to perform model building (Fig. 4f, Supplementary Fig. 8, maps-model FSC curves are reported in Supplementary Fig. 7). In conjunction with rigid-body fitting of several copies of the crystal structure of the apo Δ-BED Hermes monomer (PDB: 2BW3)¹⁰, we obtained the structure presented in Fig. 4e.

The analysis of the final composite map in Fig. 4d revealed that two Hermes dimers A and C (with each monomer A1/A2 and C1/C2 in pink and orange, respectively) are bridged by two DNA duplexes. There are also two DNA-bound BED domains (in green) that have been contributed by the B dimer. Thus, three BED domains bind to each DNA. It was possible to unambiguously assign the orientation of the DNA molecules (hereafter 5’ and 3’ always refer to the top strand) since the 5’-flanking region, as well as the two-nucleotide gap on the top strand of the 8 + LE-TIR + 30 DNA, were clearly visible. The two LE-DNAs are antiparallel, with each TIR interacting with a different Hermes dimer next to the 3’-end of the other oligonucleotide (Fig. 4f, g). The configuration of the transpososome with two antiparallel DNAs was the only configuration we saw, yet such an arrangement with the TIRs bound by opposite dimers cannot support the integration of the Hermes transposon across an 8 bp target site.

The 8 + TIR-LE + 30 DNA includes the TIR and STR1 through STR4. STR1 is clearly bound by the BED domain that belongs to the same Hermes protomer (A1 or C1) that also binds to the TIR as indicated by distinct linker density between the BED domain and the dimerization domain (DD in Fig. 1c). The equatorial BED domains bound to STR2 are closest to dimer B whereas dimer D is further away on the other side of the DNAs. Furthermore, the STR2 motifs are also bound in the hexameric Hermes transpososome where only one equatorial dimer is present at the position of dimer B. Therefore, we conclude that the equatorial Hermes dimer B engages both its BED domains with the STR2 motifs. The final map also showed weak density that linked the STR3-bound BED domains to the core of protomers A2 or C2. STR4 was not bound by any BED domain; rather, the 3’-end of each oligonucleotide (bp 36–47) was positioned similarly as the TIR entering the catalytic domain of the A2 or C2 protomer, with bp 47 sitting at the place of the TIR bp 3. The insertion of the 3’-ends inside a Hermes dimer in a TIR-like manner was unexpected as it appears to place a full-length transposon end at risk if it were to be inadvertently cleaved. However, the DNA sequence at the oligonucleotide 3’-end has no similarity to that of the TIRs. Of note, no protein-protein interactions between the BED domains or between the BED domains and the rest of their transposase were observed.

The STRs of the RE do not interact with the BED domains inside the transpososome

As we were unable to generate interpretable cryo-EM data for a single Hermes assembly that bound one LE and RE transposon end at the same time, we determined the structure of the transpososome bound to two REs. Using a similar approach as described for the LE-LE version based on partial signal subtraction to focus the reconstruction on the DNAs and the Hermes dimers A and C (Supplementary Fig. 9), a reconstruction at 5.1 Å resolution was obtained and a model could be generated (Fig. 5e and Supplementary Fig. 10). The 3D class that refined best clearly showed 5’-flanks synapsed in the top A1/A2 dimer revealing two parallel DNAs. In marked contrast to the LE-LE transpososome, none of the refined 3D classes showed density corresponding to either free or bound BED domains (Supplementary Fig. 9). The 3’-ends of 8 + RE-TIR + 30 DNAs were both inserted in the C1/C2 dimer in a TIR-like fashion, suggesting that this interaction was not a function of STR/BED interactions, but a possible consequence of the particular length of the DNA used.

**Fig. 5: The cryo-EM structure of the RE-RE *Hermes* transpososome.**

The minimal BED binding site is the AAGT motif, but it must be followed by an AT-rich region

It was very surprising that in the RE-RE complex, no BED domain was bound to RE-STR1 as we had anticipated it would be bound as seen for LE-STR1. This forced us to reevaluate the notion that the AAG(T/C) motif indeed represents the BED domain recognition site. We first asked whether RE11-27, that contains RE-STR1 (5’-AAGC), binds the BED domain (Supplementary Fig. 11). SEC analysis showed that RE11-27 DNA binds similarly as the randomized control, ran17, and did not form a stable complex. Likewise, a mutant RE11-27 (RE11-27T) with an AAGT motif replacing AAGC did not form a stable complex either. However, on the LE, LE11-27mut, where STR2 (5’-AAGC) was replaced by 5’-AAAA, formed a stable 1:1 protein-to-DNA complex. These results suggested that an isolated AAGT motif was necessary but not sufficient for BED binding.

We noted that the AAGT motif in LE11-27mut was flanked by two AT-rich regions, whereas RE11-27T presented the AT-rich region only at its 5’-end. We then tested if a BED domain interacts with an isolated AAGT motif only if it is followed by an AT-rich region (Supplementary Fig. 11, LE11-27mut-5’G vs. LE11-27mut-5'3’G), and we found this was indeed the case. We verified whether it was also the case for an isolated AAGC motif, as the RE-STR1 did not show binding. The SEC data obtained with LE11-27mutC showed only weak binding comparable to ran17, suggesting that an isolated AAGC motif was not sufficient. Collectively, these experiments revealed that the AAGT motif is the effective BED binding site, but it must be accompanied by an AT-rich region to achieve high-affinity DNA binding.

The interaction of the BED domains with the quasi-palindromic LE-STR1-STR2 is cooperative

The characterization of the BED binding site suggested that the AAGC motif might need the presence of the AAGT motif adjacent to it in a palindromic configuration for optimal binding. Furthermore, the crystal structure suggested the possibility of cooperative binding between the two BED domains as the two α3 helices from two protomers are inserted into a deepened major groove adjacent to each other. To test this hypothesis, we used electrophoretic mobility shift assays (EMSA) with fluorescently labeled LE11-27 and LE11-27mut DNAs and the BED domain. The resulting polyacrylamide gels are shown in Fig. 6b. As expected, the DNAs generated two different delayed bands resulting from the interaction of two and one BED with LE11-27 and LE11-27mut, respectively. No band corresponding to a single binding event, even in the early stage of the titration, was detected for LE11-27. Therefore, only simultaneous double interaction events were observed consistent with cooperative binding. As the crystal structure of the BED/DNA complex showed no evidence of protein-protein contacts, it appears therefore, that cooperativity is the result of DNA deformation.

**Fig. 6: Cooperative interaction of the BED domain with *Hermes* left-end (LE) STR1-STR2 motif.**

Restoring asymmetric BED binding sites in the LE/LE transposon does not rescue transposition activity in cells

The cryo-EM reconstructions indicated that the LE/LE pair, in contrast with the RE/RE pair, is in antiparallel orientation, a configuration that cannot support the integration of the transposon. We suspected that the BED/STR interactions within the LE/LE pair were responsible for the antiparallel orientation. Previously, we showed that a modified Hermes transposon with symmetrized LE/LE ends were inactive in insect cells¹¹. We asked if we could rescue the transposition of a LE/LE transposon carrying a puromycin marker, by mutating the first three STRs of one end to make it resemble the RE, contained on a LEmut donor plasmid, p-donor (Fig. 7b). The transposase was expressed from a helper plasmid, p-helper. Both plasmids were transfected into HEK293T cells. After several days under puromycin selection, colonies were counted as a read-out (Fig. 7a). As Hermes transposition in HEK293T cells has not been previously reported, we first established it was active on its wild-type ends. As a comparison, Hermes showed a higher transposition rate than that of wild-type piggyBac under identical transfection conditions. (Hermes was also reported as highly active in S. pombe, so the T317A mutation had to be introduced to tame it)¹². Symmetrized LE/LE p-donor led to a drastic loss of mobilization in Drosophila S2 DEV8 cells¹¹, and similarly, in HEK293T cells, LE/LE p-donor retained little mobilization, while the RE/RE transposon showed no activity. Finally, the LE/LEmut p-donor showed the same activity as the LE/LE p-donor. Thus, restoring the STR asymmetry in the first 50 bp in one of the LEs was not sufficient to rescue the mobilization of the LE/LE transposon. Like previous experiments in insect cells, transposition activity in HEK293T cells also required substantially longer transposon ends. Complexes assembled with such ends were insoluble, so we have thus far been unable to study them with structural tools. However, it appears clear that the asymmetry of the transposon ends is an important requirement for activity.

**Fig. 7: *Hermes* transposition assay in HEK293T cells.**

The minimal end requirement for Hermes transposition in HEK293T cells

The transposition assay showed that the divergence of sequence between the ends must span beyond the first 50 bp to have a mobile system. Hence, we sought to establish how much of the LE is needed to observe transposition while keeping the RE unchanged. We tested nine different LE/RE p-donors wherein the LE was gradually truncated (LE50 to LE306) as well as one in which the region spanning over the first three STRs was deleted (LEΔ), and the results are presented in Fig. 7c. The respective lack of significant integration of the transposons LE50/RE and LEΔ/RE confirmed that neither the 50 bp downstream of the LE-TIR nor the rest of the LE are sufficient for mobilization in cells when combined with the full RE.

The truncated LE constructs showed that mobility equivalent to the full-length LE was retained up to 223 bp. A slight decline in integration started at 190 bp to finally reached a drastic downturn at 140 bp and shorter. The loss of transposition activity was steady, allowing us to affirm that the critical truncation point on the LE lies between 140 and 190 bp when associated with the full-length RE. Experiments with partially scrambled LE sequences suggest that there is no influence on transposition by not keeping constant the distance between the LE-TIR and the CMV enhancer (Fig. 7e). Similarly, we also asked what the minimal end requirement was for the RE when combined with LE223 (Fig. 7d). We observed that the RE truncated after 327 bp resulted in similar integration as the full-length RE. The transposition activity was slightly reduced with 215 bp of RE, while it is severely impaired when only 144 bp of RE were retained. We conclude that the Hermes transposon needs at least ~220 terminal bp of both its LE and RE to be effectively mobilized in cells.

The binding of auxiliary Hermes transposases inside the transposon ends to support transposition in cells is unlikely

In the subterminal regions of the LE and RE that are critical for activity, there are several STRs that are putative BED binding sites. On the LE, there are two putative BED binding sites between bp 140 and 223 at positions 183 and 190 (LE-STR5 and LE-STR6). There are also two STRs on the RE between bp 144 and 327 at positions 172 and 228 (RE-STR4 and RE-STR5). To determine if these STRs are important for Hermes transposition in cells, we tested the mobility of a transposon donor in which the LE putative BED binding sites were all mutated (LEmut2/RE, Fig. 7e). We also tested donors in which the LE sequence was scrambled after bp 50 (LE50Scr/RE) or after bp 140 (LE140Scr/RE). Whereas LEmut2/RE retained ~70% of wild-type activity, the scrambled LE sequences in LE50Scr/RE and LE140Scr/RE p-donors showed no or little mobility (Fig. 7e). It, therefore, appears that the putative LE BED binding motifs downstream of the first 50 bp are not needed for transposition, ruling out a model in which auxiliary transposases besides the octameric transpososome are required. Nonetheless, it appears that some as-yet-unidentified feature inside the Hermes transposon ends is needed to support transposition in cells.

Discussion

The cut-and-paste Hermes transposon stands out among the transposases that have been biochemically or structurally characterized so far. While most DNA transposases function as dimers or tetramers, Hermes is unique that it forms closed octameric rings as a tetramer of dimers. While a mutant version of Hermes, incapable of forming octameric rings, that forms only dimers, is hyperactive for transposition in vitro at low salt conditions, it is inactive at physiological ionic strengths or in cells¹¹. We sought to establish a mechanistic explanation of the higher-order organization necessary for in-cell activity.

The cryo-EM structures of the cores of the LE-LE and RE-RE transpososomes determined here provide a new framework for understanding the role of Hermes’ unique ring-shaped organization. In the LE-LE complex, although the two ends were antiparallel (an arrangement that clearly cannot support transposition), each of the first three STRs within the first 35 bp of the LE interacted with a BED domain originating from three different Hermes dimers (Fig. 8a, left). This included the high-affinity LE-STR1-STR2 site that is present only on the LE of Hermes. The RE adopted a similar but parallel positioning within the transpososome, but it did not interact with BED domains (Fig. 8a, right). Taken collectively, we propose that transposition activity in cells relies on a transpososome assembly that can supply a sufficient number of BED domains and, in the correct configuration, first recognize and tightly bind to its transposon LE. Seemingly, a ring-shaped octamer that assembles a stable closed-form oligomer has evolved to serve this purpose. Furthermore, the binding of an LE, as it interacts with three dimers out of four, could shape the three-dimensional organization of the Hermes assembly such that it might not be able to accommodate three STR/BED interactions with a second LE bound parallel to the first. Thus, the antiparallel LE binding we observed might be the most energetically stable solution for the Hermes assembly in the presence of multiple LEs. We propose that the cooperative binding of two BED domains to the LE-STR1-STR2 palindrome is the key to the recognition of the LE and this is the first binding event (Fig. 8b). The cryo-EM structure of the core of the RE-RE transpososome showed that the RE does not interact with any BED domain, suggesting that RE binding relies on the interaction of its TIR with one dimer in the octamer, perhaps aided by nonspecific contributions from dimer C across the octamer. Assisting the process is the fact that once the LE is bound, the RE is only a few kb away at the other end of the transposon (Fig. 8b).

**Fig. 8: Model of the formation of the *Hermes* transpososome in cells.**

The LE-STR/BED interaction relies on a rigid pattern with the STR motifs precisely positioned relative to one another and to the TIR, compatible with the three-dimensional arrangement of the BED domains inside the transposase octamer. For example, the cooperative binding of two BED domains with the LE-STR1 and LE-STR2 is only possible due to their palindromic arrangement, and we have previously shown that even a single bp shift in the position of the LE-STR1 and LE-STR2 away from the LE-TIR severely impairs the transposon cleavage activity¹¹. These observations suggest that the exact phasing of the LE-STRs is crucial. Certainly, the closed architecture of a ring-shaped assembly is more rigid when compared to the alternative of linear arrangement, and we suggest that this is an important aspect of supporting the proper spatial organization of the BED domains and their interactions with the LE-STRs. It is possible that the role of the equatorial dimers is to provide BED domains to the complex and hold the ring together so that additional nonspecific interactions can form with the dimer that is diametrically opposite to the catalytically active dimer.

Conceptually, the accumulation of zinc-finger (ZF) motifs for high-affinity binding is similar to that of transcription factors (TFs) that are composed of ZF arrays. For example, the essential insulator protein CTCF has eleven ZFs²⁹, and KRAB-ZF proteins feature twelve ZFs on average³⁰. The ZFs of these polydactyl proteins recognize a wide variety of DNA motifs of typically three or four bases avoiding redundancy within the same protein. The organization in arrays enables the combination of ZFs to interact with longer target sequences and increases the affinity of the protein toward DNA sites. As the Hermes transposase did not evolve the luxury of multiple BED domains arranged as beads on a string as did ZF-TFs, it apparently solved its affinity problem by high-order multimerization to make an array out of a single domain on the polypeptide. The use of a short recognition motif has a problem, though, as it is present throughout the genome. Thus, while the 5’-AAGT motif is the essence of the minimal BED binding site, high-affinity interaction occurs only if an AT-rich region follows. AT-rich sequences tend to be intrinsically bent, and we observed that in our crystal structure of the BED/DNA complex, three lysines interact with the phosphate group of three nucleobases positioned at the 3’-end of the specific binding sites. This interaction seems to be facilitated by the bending of the DNA, and in the case of an AAGT motif followed by an AT-rich sequence, it is plausible that the intrinsic bending of the 3’-end of the motif favors the interaction of these three lysines, thereby stabilizing BED binding.

It was previously suggested that the Hermes STR was 5’-GTGGC as it was repeated several times close to the TIRs, and also found in the closely related hobo transposon^11,31. Our binding data suggests that this should be reevaluated, as 5’-AAGT represents the BED binding site, although its binding mode is complex as it also depends on the sequence following the tetranucleotide. The Hermes BED domain has some nonspecific DNA-binding affinity but reaches high affinity specific binding only if the AAGT motif is followed by an AT-rich sequence. Under these constraints, Hermes has at most eight other putative BED binding sites on the LE after bp 50, and eight on the RE (Supplementary Fig. 3b). The Hermes ~200 bp minimal ends requirement for in-cell transposition strongly suggests that the sequence beyond the terminal 50 bp of both ends is involved in the formation of a competent transpososome yet mutation of the putative BED binding sites inside the LE did not dramatically reduce transposition activity. This suggests that the binding of additional Hermes transposases is not required in cells, and instead, we hypothesize that a cellular factor may be involved in directing the formation of the synaptic complex in HEK293T cells. For example, DNA-bending proteins have been identified as transposition cofactors both in prokaryotes and eukaryotes. The bacterial HU protein was reported to associate with bacteriophage Mu^20,32, and the IHF protein with Tn10³³, Tn4652³⁴, and the Tn7-like transposon from the type I-F Vibrio cholerae CRISPR-associated transposase system³⁵. In mammalian cells, the HMGB1 protein interacts with Sleeping Beauty³⁶, as well as the recombination signal sequence (23RSS) of the RAG1/RAG2 V(D)J recombination complex closely related to transposition^37,38. By bending the DNA, these proteins are proposed to facilitate the formation of the synaptic complex. We have not been able to identify a candidate motif inside the Hermes transposon ends: HMGB1, for example, does not have a well-defined binding motif; rather, it interacts with non-canonical and pre-bent DNA^39,40. However, the Hermes LE and RE have an AT-content of 64 and 70%, respectively, with a high potential for intrinsic bending. A better understanding of the influence of host cell factors on the mobility of the Hermes transposon would be of great interest, especially in the scope of the development of genome editing tools based on Hermes and other hAT superfamily transposons.

Our results have implications beyond Hermes to other members of the hAT superfamily. Kunze and coworkers have reported that the N-terminal region of the hAT Activator (Ac) transposase recognizes several STRs scattered in both ends of Ac in a cooperative manner^41,42,43. We suggest that it is the cooperative binding of Ac’s C₂H₂-BED domains that might be responsible for this. Most of the hAT transposons have sets of STRs close to the TIRs⁴⁴. Interestingly, the hAT Tol2 transposon presents five 5’-AAGT motifs in the first 50 bp of its LE and RE, and the LE somewhat resembles Hermes LE, featuring a perfect palindrome (5’-AAGTACTT) directly after the TIR followed by a third antisense STR 10 bp downstream⁴⁵. Furthermore, all the hAT transposases have a BED domain at their N-terminus⁴⁴. However, to our knowledge Hermes is the only hAT transposase that was expressed and purified to obtain soluble transposase; therefore, biochemical and biophysical data involving the crucial question of multimerization by other hAT transposases are unfortunately not currently available. It has been suggested that Tol2, Tgf2, and the domesticated Kat1 can form oligomers (bigger than dimers) spontaneously or upon DNA binding^46,47,48. Therefore, it is possible that the ability to deploy a multitude of BED domains and to cooperatively bind to STRs might be generalizable to other hAT transposons as well.

The hAT transposons have been co-opted several times as a source of the coding sequence for the emergence of new genes and functions⁴⁴. Even though domesticated proteins have evolved from intact transposases, it seems that their enzymatic activity is rarely retained, contrary to their DNA-binding capacity^49,50. The hAT BED domain has been conserved in many cases and even sometimes replicated, as in the vertebrate ZBED4 and ZBED6 proteins that have four and two BED domains, respectively⁵¹. The ZBED transcription factors are expressed in various vertebrate tissues and have been found to be involved in the regulation of many functions, such as the expression of ribosomal protein genes⁵², embryogenesis and carcinogenesis⁵¹, retinal morphogenesis⁵³, or muscle development⁵⁴. Our structural results might help in better understanding how these proteins function.

The simplicity of cut-and-paste transposon systems makes them appealing to re-purpose as genomic tools as transposons are naturally occurring genome editing systems. Importantly, cut-and-paste DNA transposition does not require extensive DNA repair and, therefore can work in cell lines where repair mechanisms are not efficient. Currently, the two most widely used transposon systems to modify mammalian genomes are piggyBac and Sleeping Beauty^55,56. However, as wild-type transposases typically have suboptimal activity since they are not under positive selection, this gives us the opportunity to generate hyperactive transposons. A spectacular example of this possibility is Sleeping Beauty, whose transposase was first reactivated and then at least two orders of magnitude of increased transposition activity was achieved using educated guesses from sequence alignment and randomized genetic screens^57,58. The Hermes transposon is mobile in a variety of organisms from yeasts^59,60,61, to various insects^62,63,64,65, and here we showed also in human tissue culture and at activity levels comparable to piggyBac. The understanding of the principles of transpososome assembly and organization at the three-dimensional level should open up possibilities for the rational redesign of Hermes to join the array of transposon-based tools available for the modification of mammalian genomes.

Methods

Isolated BED domain expression and purification

The region coding for the N-terminal BED domain of Hermes (Uniprot Q25442) spanning from residues 1 to 78 (BED) was cloned into a pET15b expression vector in such a way that the recombinant protein was not tagged. The protein BED was expressed in Rosetta2(DE3) E. coli cells (Novagen) by growth at 37 °C in LB medium supplemented with 50 µg/mL carbenicillin until reaching OD_600nm ~ 0.6, followed by cooling at 16 °C and induction by addition of IPTG to the final concentration of 0.25 mM. After 16 h, 1 L of culture was harvested by centrifugation and resuspended in 50 mM Tris-HCl pH 7.5, 100 mM NaCl, 0.3 mM TCEP, 2.5 mM MgCl₂, DNaseI, and protease inhibitors (Roche). The cells were disrupted by sonication and the soluble part of the lysate was purified on a 5 mL-HiTrap heparin column (GE Healthcare) at 4 °C. The column was equilibrated with 25 mM HEPES.Na pH 7.5, 100 mM NaCl, 0.3 mM TCEP, and protease inhibitors. BED was eluted with a linear salt gradient (100 mM to 1 M NaCl). The fractions that contained BED were concentrated and further purified at 4 °C on a preparative Superdex 75 16/600 size exclusion column (GE Healthcare) equilibrated with 25 mM HEPES.Na pH 7.5, 500 mM NaCl, and 0.3 mM TCEP (see chromatogram in Supplementary Fig. 1a). The protein and its purity were checked at each step of the purification by SDS-PAGE electrophoresis (4–12% Bis-Tris NuPAGE, MES as running buffer, Invitrogen) stained with SimplyBlue SafeStain (Novex) (Supplementary Fig. 1a).

Hermes transposase expression and purification

The region coding for the full-length Hermes transposase (Uniprot Q25442), with the mutations Q2E and K128G) was cloned into pBAD/Myc-His (Invitrogen) in such a way that the recombinant protein was not tagged. Hermes was expressed in Top10 (Invitrogen) E. coli cells by growth at 37 ˚C in LB medium supplemented with 50 µg/mL carbenicillin until reaching OD_600nm ~ 0.6, then followed by cooling at 16 ˚C and induction by addition of arabinose to the final concentration of 0.01%. After 16 h, 4 L of culture were harvested by centrifugation and resuspended in 50 mM Tris-HCl pH 7.5, 500 mM NaCl, 0.3 mM TCEP, 40 µM MgCl₂, DNaseI, and protease inhibitors (Roche). The cells were disrupted by sonication and the soluble part of the lysate was purified on two 5 mL-HiTrap heparin columns (GE Healthcare) at 4 ˚C. The column was equilibrated with 25 mM HEPES.Na pH 7.5, 100 mM NaCl, 0.3 mM TCEP, and protease inhibitors. Hermes was eluted with a linear salt gradient (640 mM to 1 M NaCl). The fractions that contained the protein were concentrated and further purified at 4 ˚C on a preparative Superose 6 XK 16/70 size exclusion column (GE Healthcare) equilibrated with 25 mM HEPES.Na pH 7.5, 750 mM NaCl, 0.3 mM TCEP, and protease inhibitors (see chromatogram in Supplementary Fig. 1b). The fractions containing Hermes were then concentrated to be directly used or frozen in liquid nitrogen to be stored at −80 ˚C after the addition of glycerol to a final concentration of 15%. Hermes eluted as a peak with a low molecular weight shoulder; we excluded the fractions corresponding to the shoulder (Supplementary Fig. 1b).

Preparation of the double-stranded DNA samples

All the oligonucleotides were purchased from Integrated DNA Technologies (IDT). The lyophilized DNAs were dissolved in 10 mM Tris-HCl pH 8.0 to a concentration of 1 mM or 1.5 mM for the unlabeled DNAs and to a concentration of 100 µM for the 6FAM-labeled DNAs. The double-stranded DNA samples were prepared by mixing stoichiometrically the complementary strands to a final concentration of 500 µM (unlabeled) or 10 µM (6FAM-labeled). The mixes were heated to 95 ˚C for 10 min and slowly cooled to room temperature, and stored at −20 ˚C. All the DNAs used in this study are reported in Supplementary Table 2.

BED/DNA interaction assay monitored by size exclusion chromatography (SEC)

Purified BED was dialyzed overnight at 4 ˚C against 25 mM HEPES.Na pH 7.5, 150 mM NaCl, 0.3 mM TCEP, and protease inhibitors (Roche) and subsequently concentrated (Vivaspin 20 3 kDa MWCO, GE Healthcare) to ~7 mg/mL. The samples (60 µL) were prepared by mixing the protein and the oligonucleotide from the 500 µM stock solution in various ratios (1:0, 0:1, 1:1, 2:1, and 3:1 DNA-to-protein ratio with 1 equivalent corresponding to 100 µM) and equilibrated on ice for at least 15 min. About 10 µL of the sample (10 µl injection loop loaded with 50 µL sample) were injected on a Superdex 75 PC 3.2/30 (GE Healthcare) analytical size exclusion column preequilibrated with the protein buffer. The UV absorbance at 280 and 260 nm were monitored to identify the different species.

Hermes transposase stoichiometry determined by SEC

Purified Hermes at 1 mg/mL in B750 buffer (750 mM NaCl, 25 mM HEPES.Na pH 7.5, 0.3 mM TCEP, and protease inhibitors) was either dialyzed overnight or 3 h against buffer B500 (500 mM NaCl, 25 mM HEPES.Na pH 7.5 and 0.3 mM TCEP). The latter sample was then transferred in B250 (250 mM NaCl, 25 mM HEPES.Na pH 7.5, and 0.3 mM TCEP) either overnight or for 3 h. The latter sample was finally transferred to B150 (150 mM NaCl, 25 mM HEPES.Na pH 7.5, and 0.3 mM TCEP) for 16 h. The three samples were analyzed by SEC (Superose 6 30/100, GE Healthcare) in their respective buffer. The absorbance at 280 nm was monitored. Calibration curves (MW as a function of the elution volume) in B500, B250, and B150 were obtained, using three proteins from the Sigma-Aldrich Gel Filtration Markers kit: thyroglobulin (669 kDa), apoferritin (443 kDa), and β-amylase (200 kDa).

Hermes/DNA interaction assay monitored by SEC

Purified Hermes was concentrated (Vivaspin 20 50 kDa MWCO, GE Healthcare) to ~3 mg/mL. The samples were prepared by mixing the protein and the double-stranded DNAs from the 500 µM stock solution in various ratios (1:0, 0:0.25, 1:0.25, and 1:0.5 protein-to-DNA ratio with 1 equivalent corresponding to 34 µM). The samples were successively dialyzed at 4 ˚C against 500, 250, and 150 mM NaCl containing buffers (25 mM HEPES.Na pH 7.5, 0.3 mM TCEP, and protease inhibitors) for 2, 4 h, and overnight, respectively. About 100 µL of the sample were injected on a Superose 6 30/100 (GE Healthcare) analytical SEC column preequilibrated with the last dialysis buffer. The UV absorbance at 280 and 260 nm were monitored to identify the different species.

Hermes transposase stoichiometry determined by mass photometry (MP)⁶⁶

The day before the experiment, three samples of frozen Hermes transposase (15% glycerol, 25 mM HEPES.Na pH 7.5, 750 mM NaCl, and 0.3 mM TCEP) were dialyzed for 2 h against buffer B500 (500 mM NaCl, 25 mM HEPES.Na pH 7.5, and 0.3 mM TCEP). One sample was left in B500 to equilibrate overnight, while the other two were transferred in B250 (250 mM NaCl, 25 mM HEPES.Na pH 7.5, and 0.3 mM TCEP) for 3 h. Finally, one sample was left overnight in B250, while the last sample was transferred to B150 (150 mM NaCl, 25 mM HEPES.Na pH 7.5, and 0.3 mM TCEP) for 16 h. A volume increase, as well as light precipitation, was observed as the NaCl concentration of the dialysis buffer decreased. The concentrations of the dialyzed samples were 10.0 µM (in B500), 6.5 µM (in B250), and 6.0 µM (in B150). Just before performing the MP experiments, the samples were diluted 10, 50, and 100 times in their respective buffers, and equilibrated for at least 10 min at room temperature. The detailed protocol was published by the NHLBI biophysics facility with the silicon gasket applied on a glass coverslip as a sample holder⁶⁷. Data collection was performed on a G10 RefeynOne mass photometer. About 10 µL of buffer was used to optimize the focus, then 10 µL of the sample was mixed into the buffer drop for data collection (1 min acquisition). The concentrations that gave the best signal-to-noise were 50, 65, and 30 nM in B500, B250, and B150, respectively. The movie frames were processed with the built-in software DiscoverMP. The contrast values were converted to masses using a calibration curve obtained with BSA (monomer and dimer), alcohol dehydrogenase (monomer and dimer), ovalbumin, and thyroglobulin in a PBS buffer. The mass distributions were fitted using a Gaussian distribution model implemented into DiscoverMP. The theoretical mass of a Hermes monomer, dimer, hexamer, and octamer are 70.1, 140.2, 420.7, and 561.0 kDa, respectively.

Transposase/DNA complexes stoichiometry determined by MP

A typical sample (100 µL) was composed of 10 µM Hermes mixed with 2.5, 5, or 10 µM DNA in 25 mM HEPES.Na pH 7.5, 750 mM NaCl, and 0.3 mM TCEP. The samples were successively dialyzed as described for the transposase alone to reach the equilibrium in B150. Just before the experiments, the samples were diluted 50 and 100 times and equilibrated for at least 10 min at room temperature. The same experimental procedure was used as described for the apoprotein. The composition of the protein/DNA complexes (P1 and P2 masses) were determined by subtracting the mass of either hexameric Hermes or octameric Hermes, and the remaining mass was then divided by the mass of the given DNAs. The theoretical MW of the Hermes hexamer and octamer are 420.7 and 561.0 kDa, respectively. The theoretical MWs of the LE-TIR, LE-TIR + 13, LE-TIR + 30, and 8 + LE-TIR + 30 DNAs are 10.1, 18.1, 32.2, and 33.2 kDa, respectively.

Crystallization, X-ray diffraction data collection, structure determination, and model refinement of the BED/LE11-27 complex

After purification, BED was dialyzed overnight at 4 ˚C against 25 mM HEPES.Na pH 7.5, 100 mM NaCl, 0.3 mM TCEP and subsequently concentrated (Vivaspin 20 3 kDa MWCO, GE Healthcare) to ~10 mg/mL. The LE11-27 DNA was dialyzed against the same buffer and concentrated to ~2.5 mM (Vivaspin 500, 3 kDa MWCO, Merck). The complex was formed by mixing the protein and the DNA in a 2:1 protein-to-DNA ratio (final concentration of 800 and 400 µM, respectively). Crystals were grown at 20 ˚C by the hanging drop method. About 1.7 µL of the sample were mixed with 2.3 µL of crystallization solution composed of 100 mM Bis-Tris pH 6.5 (Hampton Research) and 25% PEG 4000 (Hampton Research). Crystals grew over 16 days to a size of ~0.3 × 0.3 × 0.3 mm and were cryoprotected by a quick transfer to a stabilizing solution at 20% ethylene glycol, 12.5 mM HEPES.Na pH 7.5, 50 mM NaCl, 25 mM Bis-Tris pH 6.5, 12.5% PEG 4000 prior to freezing in liquid nitrogen. The X-ray diffraction data were collected at the Advanced Photon Source beamline 22-ID, operated by SER-CAT on an Eiger X16M detector. Three anomalous diffraction data sets were collected on the same crystal around the zinc absorption edge, peak, and remote wavelengths (9661 eV and 1.28335 Å; 9665 eV and 1.28282 Å; 10000 eV and 1.23984 Å, respectively). The diffraction data were integrated and scaled with XDS and XSCALE⁶⁸. Initial experimental electron density maps were computed with Autosol from the Phenix package⁶⁹. The maps were improved by incorporating the anomalous signal of the DNA backbone’s 8 P atoms and the 4 S atoms from the protein using an additional dataset collected at 8.03 keV using a rotating anode source equipped with an Eiger 4 M detector. Phase calculations were performed by using Sharp⁷⁰. The density-modified map enabled to build an initial model in COOT⁷¹. The “edge” diffraction data processed as a non-anomalous dataset (Friedel law considered as true) and treated for anisotropy (Ian J. Tickle, The STARANISO Server, http://staraniso.globalphasing.org/cgi-bin/staraniso.cgi) was used to further improve and refine the model in Phenix and Buster (Global Phasing Limited) at 2.5 Å resolution^69,72. Detailed crystallographic statistics are in Supplementary Table 1.

Cryo-EM specimen preparation

Frozen or freshly purified Hermes transposase was mixed in an 8:2 protein-to-DNA ratio with either 8 + LE-TIR + 30 or 8 + RE-TIR + 30 in a 250 µL sample (1 equivalent was equal to 5 µM). The samples were dialyzed against a series of buffers as described for the SEC interaction assay with the final buffer composed of 25 mM HEPES.Na pH 7.5, 150 mM NaCl, 0.3 mM TCEP, and protease inhibitors (Roche). The samples were concentrated to 100 µL and purified by SEC (Superose 6 30/100, GE Healthcare) at 4 °C. The fraction with the highest absorbance was concentrated to reach the absorbance of ~0.60 at 260 nm and ~0.5 at 280 nm. The protein concentration was estimated to be ~0.2 mg/mL by comparing the SDS-PAGE band intensity of the cryo-EM sample with that of a dilution series of apo-Hermes.

The samples were frozen on copper or gold grids with holey carbon film coated with a 2 nm continuous carbon film (Quantifoil R1.2/1.3 ultrathin carbon, 300 mesh) freshly glow discharged for 20 s at 15 mA (PELCO easiGlow). A Vitrobot Mark IV (FEI) rapid plunging device was used for the specimen preparation. The Vitrobot chamber was at room temperature and the relative humidity was set at 100%. Three µL of samples were applied on the grid and after 10 s wait, the excess sample was blotted for 3 or 4 s (force 1) and flash frozen in liquid ethane cooled by liquid nitrogen.

Cryo-EM data collection

For the LE-LE transpososome, we collected ~9600 movies from two grids on a 200 kV Glacios TEM (FEI) equipped with a K3 direct electron detector camera (Gatan). The movies were recorded in super-resolution counting mode at a nominal magnification of 130,000, corresponding to a calibrated super-resolution pixel size of 0.58 Å per pixel with a defocus range from −1 to −2.5 µm. The acquisition was supervised by the semi-automated program SerialEM⁷³. The dose rate on the camera was set at 15 electrons per physical pixel per second. The total exposure time for each movie was 2 s with a total exposure dose of 22.3 e⁻/Å² (1.39 e⁻/Å² per frame). Each movie was composed of 16 frames, with 125 ms per frame.

For the RE-RE transpososome, ~9500 movies were recorded on a 300 kV Titan Krios TEM (FEI) equipped with a K3 camera. The movies were recorded in super-resolution counting mode at a nominal magnification of 105,000, corresponding to a calibrated super-resolution pixel size of 0.43 Å per pixel with a defocus range from −1 to −2.5 µm. The acquisition was supervised by the semi-automated program SerialEM⁷³. The dose rate on the camera was set at 22 electron per physical pixel per second. The total exposure time for each movie was 1.66 s with a total exposure dose of 48.7 e⁻/Å² (2.21 e⁻/Å² per frame). Each movie was composed of 22 frames, with 75 ms per frame.

Cryo-EM single particle analysis

The single particle analyses were performed with RELION 3.1 ran on the NIH HPC Biowulf cluster (http://hpc.nih.gov)^74,75,76. The processing of each dataset followed the same protocol unless stated differently. The processing workflows are presented in Supplementary Figs. 5, 6,9. UCSF Chimera was used to visualize the maps⁷⁷. The figures of cryo-EM map and models were prepared with Chimera or ChimeraX^77,78, and PyMol (http://www.pymol.org).

The movies were motion corrected in RELION 3.1 and binned by a factor 2, resulting in a pixel size of 1.16 Å (Glacios data) and 0.86 Å (Krios data), respectively, for further processing. The contrast transfer function (CTF) parameters were estimated with gctf1.06⁷⁹. The program crYOLO was used for particle picking with a general model on motion-corrected movies (~3.85 million and ~2.92 million initial particles for LE-LE and RE-RE transpososome, respectively)⁸⁰. The best particles were selected by several rounds of 2D classifications. The particle stacks of the LE-LE transpososome were joined and ran through a last 2D classification. The resulting particle stack (~641,300 particles and ~186,493 particles for LE-LE and RE-RE transpososome, respectively) was used to generate a 3D initial model that was used as a reference for a 3D classification job (six classes). The best classes were used for gold-standard refinement (class #4 and #6 of the LE-LE transpososome and class #2 of the RE-RE transpososome in Supplementary Figs. 5, 9, respectively). For both samples, the equatorial Hermes dimers (B and D) were poorly resolved, and we opted for a multi-body refinement for the LE-LE transpososome and a partial signal subtraction strategy to improve the alignment of the core of the complex.

For the LE-LE transpososome a “consensus” gold-standard refinement was performed with the joined particles from 3D classes #4, #5, and #6 (~359,800 particles), followed by per-particle CTF refinement, Bayesian polishing, and a new 3D refinement with the shiny particles (4.89 Å resolution). The map was divided into three body masks. Mask Body 1 covered the DNAs, the Hermes dimers A and C, and the BED domains from Hermes dimer B bound to the DNAs. The masks Body 2 and 3 covered the N-truncated Hermes dimers B and D, respectively. These masks were used to run a multi-body refinement to obtain a composite map of the LE-LE transpososome. The Body 1 map was gold-standard refined at 4.64 Å resolution. The Body 2 and 3 maps could not be gold-standard refined but were estimated at 10.21 and 10.94 Å resolution, respectively.

For the RE-RE transpososome, the 3D class #2 was gold-standard refined to generate a mask that only covered the core of the transpososome, i.e., the DNAs and the Hermes dimers A and C, leaving out the equatorial Hermes dimers B and D. Partial signal subtraction was performed on the particle stack of class #2 (53,656 particles). The resulting “edited” particles were subjected to focused refinement with local orientational search. The resolution of the final map was 5.1 Å (0.143 FSC threshold).

Model building - LE-LE transpososome

Two models of the N-truncated Hermes dimer bound to two nicked TIRs (PDB:6DX0)⁹ and four BED/LE11-27 models (PDB: 8EB5, from this publication) were rigid-body fitted inside the Body 1 map (UCSF Chimera)⁷⁷. The DNA base pairs and the BED domains that lay out of the density were removed. The contiguous pieces of DNAs were then merged as one dsDNA, and the sequence was corrected in COOT 0.9.6.2-pre EL⁸¹ to generate two LE-DNAs. The BED domains of protomers A1, A2, C1, and C2 were linked to the N-truncated Hermes structures in COOT as well. B-form restraints were applied to the DNA base pairs and the all-atom real-space refinement against the Body 1 map was performed. The resulting model was further real-space refined against the cryo-EM map and validated in Phenix 1.19.2 (cryo-EM module). The refinement statistics are summarized in Supplementary Table 3. Two models of the apo N-truncated Hermes monomer (PDB: 2BW3)¹⁰ were rigid-body fitted inside the Body 2 and Body 3 maps.

Model building—core of the RE-RE transpososome

The model of the core of the RE-RE transpososome was generated in a similar fashion as the core of the LE-LE transpososome (Body 1), but the crystal structure of the N-truncated Hermes dimer bound to two cleaved TIRs (PDB: 4D1Q)¹¹ was also used as a starting point. The refinement statistics are summarized in Supplementary Table 3.

BED/DNA electrophoretic mobility shift assay (EMSA)

The samples were prepared by mixing 0.9 µM of cold DNA in B150 (25 mM HEPES.Na pH 7.5, 150 mM NaCl, and 0.3 mM TCEP) with 0.1 µM of 6FAM-labeled DNA in B150 (5’−6FAM top strand), 0.5 µM of ran17 DNA, various amount (0 to 10 µM) of BED protein in B150 and loading dye (0.5X TBE, 10% glycerol and bromophenol blue). The samples were equilibrated for 1 h at 4 ˚C, while the PAGE gels were pre-run at 80 V (15 % polyacrylamide, 0.5X TBE, 1 mm gels). The samples were equilibrated for 15 min at room temperature prior being loaded (10 µL) onto the gels. The PAGE was run for 4 h at 80 V in an ice box. The EMSAs were performed as technical triplicates. The gels were scanned with a Typhoon FLA7000 (GE Healthcare) fluorescence imager avoiding the saturation of the detector. We used ImageJ to extract the intensity of the delayed bands⁸², and transformed as percentage of DNA bound. The means of the triplicate points were plotted against the concentration of BED and the standard deviation was reported as bars. The resulting binding curves were fitted in GraphPad Prism 9.3.1 with the equation: Y = Bmax × X^h/(Kd^h + X^h).

In-cell transposition assay

The helper plasmid pFV4a-Hermes and the donor plasmids pHermesWT-CMVpuro, pHermes2LE-CMVpuro, pHermes2RE-CMVpuro, pHermes2LEmut-CMVpuro, pHermesLE50Scr-CMVpuro, pHermesLE140Scr-CMVpuro, and pHermes2LEmut2-CMVpuro were ordered from GenScript using the plasmids pFHelR and pHelR-CMV-puro from Grabundzija et al. (2018) as backbones⁸³. The rest of the donor plasmids were obtained by deletion mutagenesis of pHermesWT-CMVpuro⁸⁴. The sequences of the transposon ends are reported in Supplementary Table 4. The helper plasmid contains the gene of the Hermes transposase, and the donor plasmid carries the puromycin resistance gene flanked by the Hermes LE and RE and variants to form the puro^R transposon.

HEK293T cells (0.5 × 10⁶) were seeded onto six-well plates 1 day before transfection. The cells were transfected with 1 µg of helper plasmid and 0.5 µg of the donor plasmid. All the transfections were performed with Lipofectamine 3000 (Thermo Fisher Scientific) according to the manufacturer’s protocol. Two days post-transfection, the cells was replated onto 100 mm dishes at a 50-fold dilution (i.e., 2% of the cells) and selected for transposon integration with 2 µg/mL puromycin. The selection medium was changed every 3 days. After 8–12 days, the cell colonies were fixed directly into their dish for 20 min with 4% formaldehyde in PBS and subsequently stained overnight with 1% methylene blue in PBS for counting. All experiments were independently replicated at least four times. Negative control experiments were performed by replacing the helper plasmid with pUC19. The piggyBac system was used as a procedure control.

All statistical analyses were performed using GraphPad Prism 9.3.1. Two-tailed Student’s unpaired t-tests were used to compare means between experimental samples and their corresponding pUC19 controls and means between some experimental samples and pHermesWT-CMVpuro. The bars correspond to the standard deviations.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The structure factors and the crystal structure of the Hermes transposase BED domain bound to the Hermes transposon LE quasi-palindrome LE-STR1-STR2 have been deposited in the Protein Data Bank (PDB) under the accession code 8EB5. The cryo-EM maps of the core of the LE-LE and of the RE-RE Hermes transpososomes were deposited in the Electron Microscopy Data Bank (EMDB) under the accession codes 28034 and 40553, respectively. Their related structure models were deposited in the PDB with the codes 8EDG and 8SJD, respectively.

References

Biémont, C. & Vieira, C. Junk DNA as an evolutionary force. Nature 443, 521–524 (2006).
Article ADS PubMed Google Scholar
Wells, J. N. & Feschotte, C. A field guide to eukaryotic transposable elements. Annu. Rev. Genet. 54, 539–561 (2020).
Article CAS PubMed PubMed Central Google Scholar
Hancks, D. C. & Kazazian, H. H. Roles for retrotransposon insertions in human disease. Mob. DNA 7, 9 (2016).
Article PubMed PubMed Central Google Scholar
Payer, L. M. & Burns, K. H. Transposable elements in human genetic disease. Nat. Rev. Genet. 20, 760–772 (2019).
Article CAS PubMed Google Scholar
Yuan, Y.-W. & Wessler, S. R. The catalytic domain of all eukaryotic cut-and-paste transposase superfamilies. Proc. Natl Acad. Sci. USA 108, 7884–7889 (2011).
Article ADS CAS PubMed PubMed Central Google Scholar
Feschotte, C. & Pritham, E. J. DNA transposons and the evolution of eukaryotic genomes. Annu. Rev. Genet. 41, 331 (2007).
Article CAS PubMed PubMed Central Google Scholar
Warren, W. D., Atkinson, P. W. & O’brochta, D. A. The Hermes transposable element from the house fly, Musca domestica, is a short inverted repeat-type element of the hobo, Ac, and Tam3 (hAT) element family. Genet. Res. 64, 87–97 (1994).
Article CAS PubMed Google Scholar
Zhou, L. et al. Transposition of hAT elements links transposable elements and V(D)J recombination. Nature 432, 995–1001 (2004).
Article ADS CAS PubMed Google Scholar
Hickman, A. B. et al. Structural insights into the mechanism of double strand break formation by Hermes, a hAT family eukaryotic DNA transposase. Nucleic Acids Res. 46, 10286–10301 (2018).
CAS PubMed PubMed Central Google Scholar
Hickman, A. B. et al. Molecular architecture of a eukaryotic DNA transposase. Nat. Struct. Mol. Biol. 12, 715–721 (2005).
Article CAS PubMed Google Scholar
Hickman, A. B. et al. Structural basis of hAT transposon end recognition by Hermes, an octameric DNA transposase from Musca domestica. Cell 158, 353–367 (2014).
Article CAS PubMed PubMed Central Google Scholar
Evertts, A. G., Plymire, C., Craig, N. L. & Levin, H. L. The Hermes transposon of Musca domestica is an efficient tool for the mutagenesis of Schizosaccharomyces pombe. Genetics 177, 2519–2523 (2007).
Article CAS PubMed PubMed Central Google Scholar
Aravind, L. The BED finger, a novel DNA-binding domain in chromatin-boundary-element-binding proteins and transposases. Trends Biochem. Sci. 25, 421–423 (2000).
Article CAS PubMed Google Scholar
Davies, D. R., Goryshin, I. Y., Reznikoff, W. S. & Rayment, I. Three-dimensional structure of the Tn5 synaptic complex transposition intermediate. Science 289, 77–85 (2000).
Article ADS CAS PubMed Google Scholar
Kosek, D., Hickman, A. B., Ghirlando, R., He, S. & Dyda, F. Structures of ISCth4 transpososomes reveal the role of asymmetry in copy-out/paste-in DNA transposition. EMBO J. 40, e105666 (2021).
Article CAS PubMed Google Scholar
Liu, C., Yang, Y. & Schatz, D. G. Structures of a RAG-like transposase during cut-and-paste transposition. Nature 575, 540–544 (2019).
Article ADS CAS PubMed PubMed Central Google Scholar
Richardson, J. M., Colloms, S. D., Finnegan, D. J. & Walkinshaw, M. D. Molecular architecture of the Mos1 paired-end complex: the structural basis of DNA transposition in a eukaryote. Cell 138, 1096–1108 (2009).
Article CAS PubMed PubMed Central Google Scholar
Ghanim, G. E., Kellogg, E. H., Nogales, E. & Rio, D. C. Structure of a P element transposase–DNA complex reveals unusual DNA structures and GTP-DNA contacts. Nat. Struct. Mol. Biol. 26, 1013–1022 (2019).
Article CAS PubMed PubMed Central Google Scholar
Lavoie, B. D., Chan, B. S., Allison, R. G. & Chaconas, G. Structural aspects of a higher order nucleoprotein complex: induction of an altered DNA structure at the Mu-host junction of the Mu type 1 transpososome. EMBO J. 10, 3051 (1991).
Article CAS PubMed PubMed Central Google Scholar
Montaño, S. P., Pigli, Y. Z. & Rice, P. A. The Mu transpososome structure sheds light on DDE recombinase evolution. Nature 491, 413–417 (2012).
Article ADS PubMed PubMed Central Google Scholar
Li, M., Mizuuchi, M., Burke, T. R. & Craigie, R. Retroviral DNA integration: reaction pathway and critical intermediates. EMBO J. 25, 1295–1304 (2006).
Article CAS PubMed PubMed Central Google Scholar
Passos, D. O. et al. Cryo-EM structures and atomic model of the HIV-1 strand transfer complex intasome. Science 355, 89–92 (2017).
Article ADS CAS PubMed PubMed Central Google Scholar
Urasaki, A., Morvan, G. & Kawakami, K. Functional dissection of the Tol2 transposable element identified the minimal cis-sequence and a highly repetitive sequence in the subterminal region essential for transposition. Genetics 174, 639–649 (2006).
Article CAS PubMed PubMed Central Google Scholar
Liu, D. et al. Functional dissection of the cis-acting sequences of the Arabidopsis transposable element Tag1 reveals dissimilar subterminal sequence and minimal spacing requirements for transposition. Genetics 157, 817–830 (2001).
Article CAS PubMed PubMed Central Google Scholar
Coupland, G., Plum, C., Chatterjee, S., Post, A. & Starlinger, P. Sequences near the termini are required for transposition of the maize transposon Ac in transgenic tobacco plants. Proc. Natl Acad. Sci. USA 86, 9385–9388 (1989).
Article ADS CAS PubMed PubMed Central Google Scholar
Holm, L. Dali server: structural unification of protein families. Nucleic Acids Res. 50, W210–W215 (2022).
Article CAS PubMed PubMed Central Google Scholar
Somerville, T. D. D. et al. ZBED2 is an antagonist of interferon regulatory factor 1 and modifies cell identity in pancreatic cancer. Proc. Natl Acad. Sci. USA 117, 11471–11482 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Lavery, R., Moakher, M., Maddocks, J. H., Petkeviciute, D. & Zakrzewska, K. Conformational analysis of nucleic acids revisited: Curves. Nucleic Acids Res. 37, 5917–5929 (2009).
Article CAS PubMed PubMed Central Google Scholar
Yin, M. et al. Molecular mechanism of directional CTCF recognition of a diverse range of genomic sites. Cell Res. 27, 1365–1377 (2017).
Article CAS PubMed PubMed Central Google Scholar
Urrutia, R. KRAB-containing zinc-finger repressor proteins. Genome Biol. 4, 231 (2003).
Article PubMed PubMed Central Google Scholar
Kim, Y. J., Hice, R. H., O’Brochta, D. A. & Atkinson, P. W. DNA sequence requirements for hobo transposable element transposition in Drosophila melanogaster. Genetica 139, 985–997 (2011).
Article CAS PubMed Google Scholar
Surette, M. G. & Chaconas, G. A protein factor which reduces the negative supercoiling requirement in the Mu DNA strand transfer reaction is Escherichia coli integration host factor. J. Biol. Chem. 264, 3028–3034 (1989).
Article CAS PubMed Google Scholar
Ilves, H., Hõrak, R., Teras, R. & Kivisaar, M. IHF is the limiting host factor in transposition of Pseudomonas putida transposon Tn4652 in stationary phase. Mol. Microbiol. 51, 1773–1785 (2004).
Article CAS PubMed Google Scholar
Sewitz, S., Crellin, P. & Chalmers, R. The positive and negative regulation of Tn10 transposition by IHF is mediated by structurally asymmetric transposon arms. Nucleic Acids Res. 31, 5868–5876 (2003).
Article CAS PubMed PubMed Central Google Scholar
Walker, M. W., Klompe, S. E., Zhang, D. J. & Sternberg, S. H. Novel molecular requirements for CRISPR RNA-guided transposition. Nucleic Acids Res. 51, 4519–4535 (2023).
Article PubMed PubMed Central Google Scholar
Zayed, H., Izsvák, Z., Khare, D., Heinemann, U. & Ivics, Z. The DNA-bending protein HMGB1 is a cellular cofactor of Sleeping Beauty transposition. Nucleic Acids Res. 31, 2313 (2003).
Article CAS PubMed PubMed Central Google Scholar
Van Gent, D. C., Hiom, K., Paull, T. T. & Gellert, M. Stimulation of V(D)J cleavage by high mobility group proteins. EMBO J. 16, 2665–2670 (1997).
Article PubMed PubMed Central Google Scholar
Kim, M. S. et al. Cracking the DNA code for V(D)J recombination. Mol. Cell 70, 358–370.e4 (2018).
Article CAS PubMed PubMed Central Google Scholar
Bianchi, M. E., Beltrame, M. & Paonessa, G. Specific recognition of cruciform DNA by nuclear protein HMG1. Science 243, 1056–1059 (1989).
Article ADS CAS PubMed Google Scholar
Sánchez-Giraldo, R. et al. Two high-mobility group box domains act together to underwind and kink DNA. Acta Crystallogr. Sect. D. Biol. Crystallogr. 71, 1423–1432 (2015).
Article Google Scholar
Becker, H. A. & Kunze, R. Maize activator transposase has a bipartite DNA binding domain that recognizes subterminal sequences and the terminal inverted repeats. Mol. Gen. Genet. 254, 219–230 (1997).
Article CAS PubMed Google Scholar
Feldmar, S. & Kunze, R. The ORFa protein, the putative transposase of maize transposable element Ac, has a basic DNA binding domain. EMBO J. 10, 4003–4010 (1991).
Article CAS PubMed PubMed Central Google Scholar
Kunze, R. & Starlinger, P. The putative transposase of transposable element Ac from Zea mays L. interacts with subterminal sequences of Ac. EMBO J. 8, 3177–3185 (1989).
Article CAS PubMed PubMed Central Google Scholar
Atkinson, P. W. hAT transposable elements. Microbiol. Spectr. https://doi.org/10.1128/microbiolspec.mdna3-0054-2014 (2015).
Koga, A., Suzuki, M., Maruyama, Y., Tsutsumi, M. & Hori, H. Amino acid sequence of a putative transposase protein of the medaka fish transposable element Tol2 deduced from mRNA nucleotide sequences. FEBS Lett. 461, 295–298 (1999).
Article CAS PubMed Google Scholar
Shibano, T. et al. Recombinant Tol2 transposase with activity in Xenopus embryos. FEBS Lett. 581, 4333–4336 (2007).
Article CAS PubMed Google Scholar
Jiang, X. Y. et al. The N-terminal zinc finger domain of Tgf2 transposase contributes to DNA binding and to transposition activity. Sci. Rep. 6, 27101 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Chiruvella, K. K., Rajaei, N., Jonna, V. R., Hofer, A. & Åstrom, S. U. Biochemical characterization of Kat1: a domesticated hAT-transposase that induces DNA hairpin formation and MAT-switching. Sci. Rep. 6, 21671 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Sinzelle, L., Izsvák, Z. & Ivics, Z. Molecular domestication of transposable elements: from detrimental parasites to useful host genes. Cell. Mol. Life Sci. 66, 1073–1093 (2009).
Article CAS PubMed Google Scholar
Etchegaray, E., Naville, M., Volff, J. N. & Haftek-Terreau, Z. Transposable element-derived sequences in vertebrate development. Mob. DNA 12, 1–24 (2021).
Article CAS PubMed PubMed Central Google Scholar
Hayward, A., Ghazal, A., Andersson, G., Andersson, L. & Jern, P. ZBED evolution: repeated utilization of DNA transposons as regulators of diverse host functions. PLoS ONE 8, e59940 (2013).
Article ADS CAS PubMed PubMed Central Google Scholar
Yamashita, D. et al. hDREF regulates cell proliferation and expression of ribosomal protein genes. Mol. Cell. Biol. 27, 2003–2013 (2007).
Article CAS PubMed PubMed Central Google Scholar
Mokhonov, V. V., Theendakara, V. P., Gribanova, Y. E., Ahmedli, N. B. & Farber, D. B. Sequence-specific binding of recombinant Zbed4 to DNA: insights into Zbed4 participation in gene transcription and its association with other proteins. PLoS ONE 7, e35317 (2012).
Article ADS CAS PubMed PubMed Central Google Scholar
Chen, T. et al. Identification of zinc-finger BED domain-containing 3 (Zbed3) as a novel Axin-interacting protein that activates Wnt/beta-catenin signaling. J. Biol. Chem. 284, 6683–6689 (2009).
Article CAS PubMed PubMed Central Google Scholar
Kim, A. & Pyykko, I. Size matters: versatile use of PiggyBac transposons as a genetic manipulation tool. Mol. Cell. Biochem. 354, 301–309 (2011).
Article CAS PubMed Google Scholar
Kebriaei, P., Izsvák, Z., Narayanavari, S. A., Singh, H. & Ivics, Z. Gene therapy with the Sleeping Beauty transposon system. Trends Genet. 33, 852–870 (2017).
Article CAS PubMed Google Scholar
Ivics, Z., Hackett, P. B., Plasterk, R. H. & Izsvák, Z. Molecular reconstruction of Sleeping Beauty, a Tc1-like transposon from fish, and its transposition in human cells. Cell 91, 501–510 (1997).
Article CAS PubMed Google Scholar
Mátés, L. et al. Molecular evolution of a novel hyperactive Sleeping Beauty transposase enables robust stable gene transfer in vertebrates. Nat. Genet. 41, 753–761 (2009).
Article PubMed Google Scholar
Gangadharan, S., Mularoni, L., Fain-Thornton, J., Wheelan, S. J. & Craig, N. L. DNA transposon Hermes inserts into DNA in nucleosome-free regions in vivo. Proc. Natl Acad. Sci. USA 107, 21966–21972 (2010).
Article ADS CAS PubMed PubMed Central Google Scholar
Park, J. M., Evertts, A. G. & Levin, H. L. The Hermes transposon of Musca domestica and its use as a mutagen of Schizosaccharomyces pombe. Methods 49, 243–247 (2009).
Article CAS PubMed PubMed Central Google Scholar
Patterson, K. et al. Functional genomics for the oleaginous yeast Yarrowia lipolytica. Metab. Eng. 48, 184–196 (2018).
Article CAS PubMed Google Scholar
Guimond, N., Bideshi, D. K., Pinkerton, A. C., Atkinson, P. W. & O’Brochta, D. A. Patterns of Hermes transposition in Drosophila melanogaster. Mol. Genet. Genom. 268, 779–790 (2003).
Article CAS Google Scholar
Sarkar, A., Yardley, K., Atkinson, P. W., James, A. A. & O’Brochta, D. A. Transposition of the Hermes element in embryos of the vector mosquito, Aedes aegypti. Insect Biochem. Mol. Biol. 27, 359–363 (1997).
Article CAS PubMed Google Scholar
Sarkar, A. et al. The Hermes element from Musca domestica can transpose in four families of cyclorrhaphan flies. Genetica 99, 15–29 (1997).
Article CAS PubMed Google Scholar
Pinkerton, A. C., O’Brochta, D. A. & Atkinson, P. W. Mobility of hAT transposable elements in the Old World bollworm, Helicoverpa armigera. Insect Mol. Biol. 5, 223–227 (1996).
Article CAS PubMed Google Scholar
Young, G. et al. Quantitative mass imaging of single biological macromolecules. Science 360, 423–427 (2018).
Article ADS CAS PubMed PubMed Central Google Scholar
Wu, D. & Piszczek, G. Standard protocol for mass photometry experiments. Eur. Biophys. J. 50, 403–409 (2021).
Article CAS PubMed PubMed Central Google Scholar
Kabsch, W. Integration, scaling, space-group assignment and post-refinement. Acta Crystallogr. Sect. D. Biol. Crystallogr. 66, 133–144 (2010).
Article CAS Google Scholar
Adams, P. D. et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr. Sect. D. Biol. Crystallogr. 66, 213–221 (2010).
Article CAS Google Scholar
De La Fortelle, E. & Bricogne, G. [27] Maximum-likelihood heavy-atom parameter refinement for multiple isomorphous replacement and multiwavelength anomalous diffraction methods. Methods Enzymol. 276, 472–494 (1997).
Article PubMed Google Scholar
Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr. Sect. D. Biol. Crystallogr. 66, 486–501 (2010).
Article CAS Google Scholar
Bricogne, G. Direct phase determination by entropy maximization and likelihood ranking: status report and perspectives. Acta Crystallogr. D Biol. Crystallogr. 49, 37–60 (1993).
Article CAS PubMed Google Scholar
Mastronarde, D. N. Automated electron microscope tomography using robust prediction of specimen movements. J. Struct. Biol. 152, 36–51 (2005).
Article PubMed Google Scholar
Zivanov, J. et al. New tools for automated high-resolution cryo-EM structure determination in RELION-3. Elife 7, e42166 (2018).
Article PubMed PubMed Central Google Scholar
Scheres, S. H. W. A Bayesian view on cryo-EM structure determination. J. Mol. Biol. 415, 406–418 (2012).
Article CAS PubMed PubMed Central Google Scholar
Scheres, S. H. W. RELION: Implementation of a Bayesian approach to cryo-EM structure determination. J. Struct. Biol. 180, 519–530 (2012).
Article CAS PubMed PubMed Central Google Scholar
Pettersen, E. F. et al. UCSF Chimera–a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612 (2004).
Article CAS PubMed Google Scholar
Goddard, T. D. et al. UCSF ChimeraX: meeting modern challenges in visualization and analysis. Protein Sci. 27, 14–25 (2018).
Article CAS PubMed Google Scholar
Zhang, K. Gctf: Real-time CTF determination and correction. J. Struct. Biol. 193, 1–12 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Wagner, T. et al. SPHIRE-crYOLO is a fast and accurate fully automated particle picker for cryo-EM. Commun. Biol. 2, 1–13 (2019).
Article Google Scholar
Casañal, A., Lohkamp, B. & Emsley, P. Current developments in Coot for macromolecular model building of electron cryo-microscopy and crystallographic data. Protein Sci. 29, 1069–1078 (2020).
Article PubMed PubMed Central Google Scholar
Schneider, C. A., Rasband, W. S. & Eliceiri, K. W. NIH Image to ImageJ: 25 years of image analysis. Nat. Methods 9, 671–675 (2012).
Article CAS PubMed PubMed Central Google Scholar
Grabundzija, I., Hickman, A. B. & Dyda, F. Helraiser intermediates provide insight into the mechanism of eukaryotic replicative transposition. Nat. Commun. 9, 1–14 (2018).
Article CAS Google Scholar
Makarova, O., Kamberov, E. & Margolis, B. Generation of deletion and point mutations with one primer in a single cloning step. Biotechniques 29, 970–972 (2000).
Article CAS PubMed Google Scholar
Sanchez-Garcia, R. et al. DeepEMhancer: a deep learning solution for cryo-EM volume post-processing. Commun. Biol. 4, 1–8 (2021).
Article Google Scholar

Download references

Acknowledgements

This work was supported by the Intramural Program of the National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health (NIDDK/NIH) under Grant DK036153-16. X-ray diffraction data were collected at the Southeast Regional Collaborative Access Team (SER-CAT) 22-ID beamline at the Advanced Photon Source (APS), Argonne National Laboratory. SER-CAT is supported by its member institutions (https://www2.ser.aps.anl.gov/about_us.html#TITLE_Memberships), and equipment grants (S10_RR25528 and S10_RR028976 and S10_OD027000) from the NIH. The use of the Advanced Photon Source was supported by the US Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract No. W-31-109-Eng-38. The computations for the cryo-EM single particle analysis were carried out using the High-Performance Computing Systems at the NIH. The authors thank Greg Piszczek and Di Wu of the Biophysics Core Facility at the National Heart, Lung, and Blood Institute (NHLBI/NIH) for the access to the mass photometer. The authors would like to also thank Huaibin Wang and Yanxiang Cui for their help with cryo-EM data collection at the Multi Institute Cryo-EM Facility (MICEF), NIDDK/NIH. The authors are also thankful to Istvan Botos (NIDDK/NIH) for on-site computational support.

Funding

Open Access funding provided by the National Institutes of Health (NIH).

Author information

Authors and Affiliations

Laboratory of Molecular Biology, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD, 20892, USA
Laurie Lannes, Christopher M. Furman, Alison B. Hickman & Fred Dyda

Authors

Laurie Lannes
View author publications
You can also search for this author in PubMed Google Scholar
Christopher M. Furman
View author publications
You can also search for this author in PubMed Google Scholar
Alison B. Hickman
View author publications
You can also search for this author in PubMed Google Scholar
Fred Dyda
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

L.L. performed most of the experiments in the study except for the in-cell transposition assays on the truncated LE that were carried out by C.M.F. L.L. analyzed all the data generated in the study. L.L. and F.D. analyzed the X-ray data. L.L. prepared all the samples used in this study, including the crystals and the cryo-EM specimens. A.B.H. initiated the project. F.D. supervised the study. All authors participated in manuscript preparation.

Corresponding author

Correspondence to Fred Dyda.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Reporting Summary

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Lannes, L., Furman, C.M., Hickman, A.B. et al. Zinc-finger BED domains drive the formation of the active Hermes transpososome by asymmetric DNA binding. Nat Commun 14, 4470 (2023). https://doi.org/10.1038/s41467-023-40210-3

Download citation

Received: 17 October 2022
Accepted: 18 July 2023
Published: 25 July 2023
DOI: https://doi.org/10.1038/s41467-023-40210-3

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.