Analysis of the LAGLIDADG interface of the monomeric homing endonuclease I-DmoI

Abstract

The general structural fold of the LAGLIDADG endonuclease family consists of two similar alpha/beta domains (alphabetabetaalphabetabetaalpha) that assemble either as homodimers or monomers with the domains related by pseudo-two-fold symmetry. At the center of this symmetry is the closely packed LAGLIDADG two-helix bundle that forms the main inter- or intra-molecular contact region between the domains of single- or double-motif proteins, respectively. In this work, we further examine the role of the LAGLIDADG residues involved in the helix-helix interaction. The interchangeability of the LAGLIDADG helix interaction was explored by grafting interfacial residues from the homodimeric I-CreI into the corresponding positions in the monomeric I-DmoI. The resulting LAGLIDADG exchange mutant is partially active, preferring to nick dsDNA rather than making the customary double-strand break. A series of partial revertants within the mutated LAGLIDADG region are shown to restore cleavage activity to varying degrees resulting in one I-DmoI mutant that is more active than wild-type I-DmoI. The phenotype of some of these mutants was reconciled on the basis of similarity to the GxxxG helix interaction found in transmembrane proteins. Additionally, a split variant of I-DmoI was created, demonstrating that the LAGLIDADG helices of I-DmoI are capable of forming and maintaining the protein-protein interface in trans to create an active heterodimer.

Figure 1

Schematic alignment of the LAGLIDADG residues of I-DmoI. Pertinent residues in P1 and P2 are identified below the consensus LAGLIDADG sequence. The alignment of the GxxxG motif is shown on the bottom. Numbers in the consensus sequences represent generic positions within the LAGLIDADG motif. Numbers in P1/P2 represent positions of residues in I-DmoI. Gray barrels indicate helical structure. A and B designate the two α/β domains that form the bulk of the protein. Dashed line represents the short (six to eight amino acids) linker between the subdomains A and B in I-DmoI.

Figure 2

Structure of I-DmoI and the LAGLIDADG motif. (A) Overall fold of I-DmoI. Left, top view down the two-helix bundle axis. Right, side view perpendicular to bundle axis. White spheres represent the van der Waals radii of the interfacial LAGLIDADG residues highlighting the majority of subdomain contacts between helices α1 and α4. Linker residues 98–105 are shown in red. (B) The α1/α4 helices (P1 and P2 regions) of I-DmoI with the XLAGLIDADG residues labeled. Black type, residues above the plane of the page; gray type, residues below the plane of the page. A dashed horizontal line demarcates the conserved sequence. At the right is the superposition of the LAGLIDADG helices of four known structures (I-DmoI, I-CreI, PI-SceI and PI-PfuI) demonstrating the extent of structural similarity in naturally occurring proteins. Although E-DreI, I-MsoI and I-AniI display a similar overlap, for clarity they were omitted from the superposition. (C) Creating I-DmoI_CL. The consensus sequence and generic numbering are shown on the left. The interfacial P1/P2 regions of I-DmoI were replaced by the P1/P1′ of I-CreI to generate I-DmoI_CL; bold type, residues changed in the LAGLIDADG of I-DmoI; gray type, residues facing the protein core and thus not changed to their I-CreI counterparts; italics, the GxxxG motif residues within the LAGLIDADG helices. Revertants are indicated to the left and right of the helices. See Figure 4 and the text for details.

Figure 3

Cleavage activity of I-DmoI and I-DmoI_CL. (A) Typical time-course activity assay. The 7 min assay was performed with pUD718 plasmid dsDNA and I-DmoI (top) or I-DmoI_CL (bottom); 20 µl samples were removed at the times indicated. Closed-circular plasmid DNA (C) was cleaved forming a linear product (L); nicked, open circles (O) appeared as intermediates of the reaction. Residual nuclease activity was ruled out by incubating each construct with target plasmid lacking only the 14 bp recognition sequence. No nicking or cleavage activity was observed (G. H. Silva, unpublished results). (B) Plots of data in (A). Percent reaction products are shown over time. Diamonds, squares and triangles correspond to O, L and C, respectively, in (A).

Figure 4

Relative cleavage activity of the I-DmoI constructs. (A) Time-course assays. Activity assays were performed in triplicate (average deviation <5%) as noted in Figure 3 for each construct listed below the plots. Percent cleavage reflects the ratio of linear to closed-circular dsDNA. (B) Relative cleavage activity. ¹Cleavage activity on circular target, resulting in linear and nicked dsDNA; ²relative cleavage to wild-type I-DmoI based on the slope of the best fit line to each curve in (A); ³CL is I-DmoI_CL base construct with residue changes as indicated (see text for details); ⁴CL_{G116A, D115V} is termed I-DmoI_H in later experiments; ⁵I-DmoI_S corresponds to split construct I-DmoI_102/101

Figure 5

Creating a functional heterodimer. The wild-type I-DmoI plasmid was used as the template in efforts to create I-DmoI_S via two pathways as shown. T7, promoter region; SD, Shine–Dalgarno sequence; ATG and TAA are the start and stop codons, respectively. Coding regions are indicated by arrows. (1) Domain constructs: individual A and B domain expression vectors were created. Protein expression was carried out separately for each domain and then lysates were combined and tested for heterodimer formation based on DNA cleavage activity. No active constructs were generated by this method. (2) Tandem constructs: individual A and B domain coding sequences were cloned into a single plasmid and co-expressed under the control of a single T7 promoter. Constructs were tested for heterodimer formation based on DNA cleavage activity as described in Table 4, Figure 6 and the text.

Figure 6

Cleavage-site mapping of I-DmoI_CL, I-DmoI_H and I-DmoI_S. The cleavage positions (arrows) on the top (left) and bottom (right) strand were mapped for I-DmoI (lane 1), I-DmoI_CL (lane 2), I-DmoI_H (lane 3) and I-DmoI_S (lane 4). Lane 5, no-protein control. A, C, G and T indicate the base in the sequencing ladder; left and right, the sequence readout for the area of interest as marked; bottom, the I-DmoI homing site with the cleavage position indicated.

Figure 7

The extent of heterodimer formation in I-DmoI_S. (A) Schematic of the I-DmoI constructs used to probe heterodimer formation. I-DmoI constructs were created with the N-terminal FLAG (black square) and/or C-terminal c-myc (black circle) epitopes. Abbreviations for construct names appear in parenthesis to the left of each diagram. (B) Western blots quantifying I-DmoI and I-DmoI_S. For all panels, sizes indicate expected position of wild-type I-DmoI (22 kDa) and split constructs (11 kDa; I-DmoI_S and I-DmoI_SS). Panel 1 shows that I-DmoI_S does not contain detectable full-length protein using anti-I-DmoI polyclonal antibody. Shaded triangles indicate decreasing protein concentration. Panel 2 shows that no cross-reactivity is evident for untagged or c-myc-tagged I-DmoI and I-DmoI_S using the anti-FLAG antibody (the shadow band between lanes _fW and W_c is an exposure artifact). Panel 3 shows the dramatically reduced amount of domain B in protein lysates (total protein amounts loaded are as in panel 2) when c-myc-tagged I-DmoI, I-DmoI_S and I-DmoI_SS proteins were probed with anti-c-Myc antibody. S_c1 and S_c2 represent crude and cleared lysates of S_c protein, respectively. (C) Schematic of IP/western blot protocol. This method was used to determine relative dimerization. See text for details. (D) Western blot of dimerization. Results from the protocol outlined in (C). Panels 1a and 1b show different exposures of the same experiment; Panels 2a and 2b show the results of two separate experiments. Size markers are as in (B). IP indicates antibody used for immunoprecipitation: L, lysate-only control (no IP); F, anti-FLAG antibody; and C, anti-c-Myc antibody. Western indicates antibody used for western blot. By-products of the IP step: LCc, light chain of the anti-c-Myc antibody; LCf, light chain of the anti-FLAG antibody.