|Home | About | Journals | Submit | Contact Us | Français|
DNA transposition is the movement of a defined segment of DNA from one location to another. Although the enzymes that catalyze transposition in bacterial systems have been well characterized, much less is known about the families of transposase enzymes that function in higher organisms. Active transposons have been identified in many insect species, providing tools for gene identification and offering the possibility of altering the genotypes of natural insect populations. One of these active transposons is Hermes, a 2749-base-pair element from Musca domestica that encodes its own transposase. An N-terminally deleted version of the Hermes transposase (residues 79–612) has been overexpressed and purified, and crystals that diffract to 2.1 Å resolution have been obtained at 277 K by the hanging-drop method.
Hermes, a transposable element isolated from the genome of the housefly Musca domestica (Warren et al., 1994 ), belongs to the hAT family (Kempken & Windhofer, 2001 ; Rubin et al., 2001 ), one of ten superfamilies into which eukaryotic transposons can be classified (Kapitonov & Jurka, 2004 ). In general, hAT elements have short terminal inverted repeats (5–27 bp), generate 8 bp target-site duplications upon transposition and encode a single transposase protein that catalyzes the DNA breakage and rejoining reactions required for transposition (Kempken & Windhofer, 2001 ). All hAT transposases display significant amino-acid sequence similarity, with the highest primary structure conservation at their C-termini (Calvi et al., 1991 ; Feldmar & Kunze, 1991 ).
Various functional and structural subdomains of the 612-residue Hermes transposase (Hermes; 70.1 kDa) have been determined or predicted (Fig. 1 ). For example, the N-terminus of Hermes contains residues important for nuclear localization (Michel & Atkinson, 2003 ) and has been proposed to contain a DNA-binding BED domain (residues 25–78; Aravind, 2000 ), while the C-terminus contains a sequence (residues 551–569) demonstrated to be important for multimerization (Michel et al., 2003 ). It has been proposed that hAT transposases may carry a DSE catalytic triad (Bigot et al., 1996 ), a variant of the DDE motif found in the active site of other transposases such as HIV-1 integrase and Mu transposase (Dyda et al., 1994 ; Rice & Mizuuchi, 1995 ), and that two acidic residues in particular, Asp402 and Glu572, are necessary for transposition (Michel et al., 2002 ). However, more recent biochemical characterization of Hermes in vitro, combined with secondary-structure prediction, indicates that Hermes may be a member of the retroviral integrase superfamily and that residues Asp180, Asp248 and Glu572 are likely to form a metal-ion-binding site (Zhou et al., 2004 ).
Our current understanding of the mechanisms of DNA transposition is derived from well studied prokaryotic and eukaryotic systems (Steiniger-White et al., 2004 ; Chaconas & Harshey, 2002 ; Rio, 2002 ), although high-resolution structural data are currently limited to prokaryotic transposases. To date, only one full-length eukaryotic transposase, Mos1 mariner from Drosophila mauritiana, has been crystallized (Richardson et al., 2004 ). However, the structure of this mariner/Tc1 superfamily member has not yet been reported. Structural data are also available for the bipartite DNA-binding domain (residues 1–135) of the Caenorhabditis elegans Tc3 transposase in complex with transposon DNA (van Pouderoyen et al., 1997 ; Watkins et al., 2004 ). Although the structure provides insight into how the Tc3 transposase recognizes DNA, no structural information has been garnered about the remaining portion of this enzyme.
Given the limited structural information available on eukaryotic transposases, we have cloned, overexpressed and crystallized a portion of Hermes with the objective of obtaining a high-resolution crystal structure. We anticipate the structural analysis of this transposase will provide insight into the mechanism by which hAT transposases catalyze DNA transposition.
The full-length Hermes transposase (Hermes; residues 1–612) was subcloned into pET-15b (Novagen) for expression in Escherichia coli as an N-terminal His-tag fusion protein. Although we were able to obtain copious amounts of soluble protein, when it was subjected to analytical gel-filtration chromatography on a calibrated Superose 6 column (Pharmacia) full-length Hermes migrated at a position consistent with a large aggregate (>700 kDa). Consequently, in efforts to identify a version that was more suitable for structural studies, Hermes was subjected to limited proteolysis by papain and trypsin. The digested protein was analyzed by SDS–PAGE and a prominent stable cleavage product common to both proteases (MW 45 kDa) was characterized by N-terminal sequencing and MALDI–TOF mass spectrometry. The results were consistent with the removal of residues 1–78 from the N-terminus and ~130 residues from the C-terminus (Fig. 1 ). Given the known importance of the C-terminal region, we elected to investigate the properties of a Hermes construct missing only the N-terminal 78 residues. Thus, primers were designed to remove codons 1–78 from full-length Hermes and DNA encoding residues 79–612 in pET-15b was generated using a QuikChange kit (Stratagene). Soluble protein was obtained using the same expression procedure as for full-length Hermes.
Soluble Hermes 79–612 (both wild-type and a single point mutant, S163G; see below) was obtained by expression in Escherichia coli BL21(DE3) cells which were grown at 310 K until OD600 = 0.6. Cells were then rapidly cooled on ice to 292 K and protein expression was induced by addition of IPTG to a final concentration of 0.5 mM.
Cells collected from 8 l culture were harvested 16–20 h post-induction. The pellet was resuspended in 300 mM NaCl, 12 mM phosphate pH 7.4, flash-frozen in liquid nitrogen and then stored at 193 K. Unless noted otherwise, all purification steps were performed at 277 K. After thawing, cells were lysed by sonication in the presence of 500 mM NaCl, 5 mM imidazole (Im), 25 mM Tris pH 7.5 and 2 mM β-mercaptoethanol (BME). Following centrifugation of the cell lysate at 100000g for 45 min, the supernatant was loaded onto a Hi-Trap metal-chelation column (Amersham Biosciences) previously equilibrated with NiSO4. The column was washed extensively with 20 mM Tris pH 7.5, 2 mM Im and 500 mM NaCl followed by the same buffer containing 22 mM Im. Hermes 79–612 was eluted from the column using a gradient of 22–400 mM Im. After visualization on an SDS–PAGE gel, fractions containing Hermes 79–612 were combined and dialyzed against 20 mM Tris pH 7.5, 1 mM EDTA, 500 mM NaCl, 4 mM BME and 10%(w/v) glycerol. This was followed by dialysis against a single change of the same buffer containing 5 mM DTT in place of BME (TSK buffer). To remove the polyhistidine tag, 10 units of thrombin (Sigma) were added per milligram of protein and incubated overnight. Thrombin was removed by passage over a 1 ml benzamidine Sepharose 4B (Pharmacia) column. The recovered protein was concentrated to at least 10 mg ml−1 prior to size-exclusion chromatography on a TSK-Gel G3000SW column equilibrated in TSK buffer.
The elution profile from the TSK column (Fig. 2 a) revealed three forms of Hermes 79–612, which has a molecular weight of 61 425 Da. Subsequent analysis of the three species on a calibrated analytical size-exclusion Superdex 200 column demonstrated that peaks 1, 2 and 3 migrated with positions consistent with a high-molecular-weight aggregate, a species that was approximately hexameric and one that was approximately dimeric, respectively. The material in peak 2, but not that in peak 3, is active in in vitro assays of strand transfer and hairpin formation (Zhou & Craig, 2004 ).
When fractions corresponding to the three protein species were analyzed by SDS–PAGE (Fig. 2 b), we noted that while the protein in peak 2 was quite homogenous, the dimeric protein consistently eluted with at least three other smaller proteins of ~9–12 kDa. N-terminal sequencing indicated that the co-migrating contaminants all had the same N-terminal sequence, GSHMQSRE, corresponding to the four residues contributed by the remnants of the polyhistidine tag (GSHM) followed by the Hermes sequence beginning at residue 79. The most prominent contaminant, indicated by the arrow in Fig. 2 (b), was analyzed by MALDI–TOF mass spectrometry using a PerSeptive Biosystems Voyager-DE instrument. The measured molecular weight was 9750 Da, consistent with a Hermes 79–162 fragment after removal of the polyhistidine tag. All attempts to remove the Hermes 79–162 fragment under conditions that maintained the native folded tertiary structure of the protein were unsuccessful. Efforts were subsequently made to eliminate the cleavage site by conservative mutation of Ile162 and Ser163. Although mutations failed to eliminate cleavage, a S163G mutant yielded considerably more soluble protein than the wild type and was therefore used for subsequent crystallization studies. This change in soluble protein recovery was the only reproducible difference between wild-type Hermes 79–612 and the S163G point mutant, which are otherwise indistinguishable in their purification properties, in vitro activities and crystallizability.
The proteins corresponding to material in peaks 2 and 3 were separately pooled, concentrated and used for crystallization trials. Although crystals were obtained of both species, only those corresponding to the approximately dimeric species were of diffraction quality. Initial crystals of the approximately dimeric species of Hermes 79–612 S163G were grown at 277 K by the hanging-drop method by mixing protein at 10 mg ml−1 in TSK buffer in a 1:1 ratio with 25%(v/v) polyethylene glycol 550 monomethyl ether (PEG 550 MME; Fluka), 0.1 M MES pH 6.5. Streak-seeding into a 1:1 mixture of protein and 23–25%(v/v) PEG 550 MME, 0.1 M MES pH 6.7 containing 0.25 M NaCl (total volume of 10 µl) yielded diffraction-quality crystals (Fig. 3 a). Crystals were extracted from the hanging drop, washed and analyzed by SDS–PAGE (Fig. 3 b); in addition to the expected Hermes 79–612 species, the persistent contaminants were also present. Crystals were cryoprotected by brief transfer into 10 mM Tris pH 7.5, 50 mM MES pH 6.5, 0.375 M NaCl, 2.5 mM DTT, 25%(v/v) PEG 550 MME, 10%(w/v) glycerol and frozen by rapid immersion in liquid propane.
Crystals of the approximately dimeric form of Hermes 79–612 S163G were initially obtained under numerous conditions from ‘The PEGS’ screening suite from Nextal (Montreal, Canada) both at room temperature and 277 K. Optimization by the hanging-drop method at 277 K, combined with several rounds of streak-seeding, yielded crystals that nucleated overnight and continued to grow for up to a week to final maximum dimensions of 0.5 × 0.4 × 0.1 mm. The crystals belong to space group C2, with unit-cell parameters a = 116.3, b = 84.9, c = 73.8 Å, β = 93.8°. Given the molecular weight of 79–612 S163G after thrombin cleavage, this corresponds to 49.2% solvent and a Matthews coefficient of 2.97 Å3 Da−1. Statistics for data collection are shown in Table 1.
Subsequent experiments have shown that the asymmetric unit contains one 79–612 monomer bound to a 79–162 fragment, consistent with the SDS–PAGE analysis of the dissolved crystals.
We thank C. Bradley for helpful comments on the manuscript.