Structural studies on Helicobacter pyloriATP-dependent protease, FtsH.

The ATP-dependent protease, FtsH, degrades misassembled membrane proteins for quality control like SecY, subunit a of FoF1-ATPase, and YccA, and digests short-lived soluble proteins in order to control their cellular regulation, including sigma32, LpxC and lambdacII. The FtsH protein has an N-terminal transmembrane segment and a large cytosolic region that consists of two domains, an ATPase and a protease domain. To provide a structural basis for the nucleotide-dependent domain motions and a better understanding of substrate translocation, the crystal structures of the Helicobacter pylori (Hp) FtsH ATPase domain in the nucleotide-free state and complexed with ADP, were determined. Two different structures of HpFtsH ATPase were observed, with the nucleotide-free state in an asymmetric unit, and these structures reveal the new forms and show other conformational differences between the nucleotide-free and ADP-bound state compared with previous structures. In particular, one HpFtsH Apo structure has a considerable rotation difference compared with the HpFtsH ADP complex, and this large conformational change reveals that FtsH may have the mechanical force needed for substrate translocation.

Introduction

ATP-dependent proteases play vital roles in protein quality control and in regulating the levels of certain cellular proteins (Sauer et al., 2004 ▶). Escherichia coli and other bacteria, including Bacillus subtilis, Staphylococcus aureus and Helicobacter pylori, contain at least five types of ATP-dependent proteases, including ClpXP, ClpAP, HslUV, FtsH and Lon. The first three (ClpXP, ClpAP and HslUV) have a composite structure of the ATPase complex and protease complex. In the case of the last two (FtsH and Lon), ATPase and protease exist in different domains in the same subunit (Gottesman, 1996 ▶).

FtsH belongs to the AAA (ATPases Associated with different cellular Activities) protein superfamily, which is characterized by a highly conserved domain comprised of 230–250 amino acid residues. FtsH is divided into an N-terminal transmembrane segment and a large cytosolic region that contains the ATPase and protease domains. The ATPase domain includes the conserved Walker A, Walker B and second region of homology (SRH) motifs. The protease domain contains a zinc-binding motif, which is consistent with the finding that its proteolytic activity is stimulated by a zinc ion (Ito & &Akiyama, 2005 ▶).

The crystal structures of the FtsH ATPase domain from Escherichia coli (Ec) and Thermus thermophilus (Tt), and the whole cytosolic region from Thermotoga maritima (Tm) and Thermus thermophilus (Tt), have been reported previously (Krzywda et al., 2002 ▶; Suno et al., 2006 ▶; Bieniossek et al., 2006 ▶). These structures suggest that substrate polypeptides bind to the lateral portion of a hexagonal plate and are then translocated to the central pore. Moreover, FtsH forms a threefold symmetric hexamer with open and closed subunits, which suggests that the polypeptide substrate is then delivered from the pore region to the protease catalytic site. However, the mechanism by which FtsH translocates substrate polypeptides is still not fully understood, in part because there is no structural information that suggests how they move from the lateral region of the FtsH hexagonal plate to the central pore region.

Crystal structures of the HpFtsH ATPase domain

Overall structure

The crystal structures of the ATPase domain of H. pylori FtsH in the nucleotide-free state and in complex with ADP were determined by molecular replacement at resolutions of 3.2 and 3.3 Å, respectively. The HpFtsH ATPase domain shows the general structural architecture characteristic of the AAA superfamily and is comprised of two structural subdomains: a large N-subdomain (residues 160–339) and a smaller C-subdomain (residues 344–418). The centre of the N-subdomain contains a 350 sheet made up of five parallel β strands in the order β2–β3–β4–β1–β5 that are surrounded by eight α-helices. The C-subdomain is composed of four helices: α9, α10, α11 and α12 (Fig. 1 ▶ a).

Figure 1

Crystal structure of the HpFtsH ATPase domain. (a) Stereoview of a ribbon diagram showing the overall structure of the HpFtsH ATPase. The α-helices are labelled α1–α12, and the β-strands are labelled β1–β5. (b) Stereoview of the ADP binding site. The bound ADP molecule is shown as the ball-and-stick model.

We observed that within the asymmetric unit under nucleotide-free conditions the HpFtsH ATPase domain assumes two different conformations, which hereafter we will refer to as HpFtsH Apo_A and Apo_B. Although the overall structures of HpFtsH Apo_A and Apo_B are quite similar, a dramatic conformational difference is apparent when their N-subdomains are superimposed: there is a rotational difference of 126.6° in the C-subdomain relative to the N-subdomain, and the root mean square deviation (RMSD) is 9.01 Å.

The ADP binding sites

The ATPase catalytic site is situated at the interface between the N- and C-subdomains (Fig. 1 ▶ b). This binding site and the conformation of ADP within the structure of the HpFtsH-ADP complex are similar to those seen within the structures of the TtFtsH-ADP and TmFtsH-ADP complexes, which reflects the strict conservation of the active-site residues. The Walker A motif provides the binding site for the ribose and diphosphate parts of the ADP molecule. The diphosphate group interacts with the strictly conserved Gly213, Thr214, Lys216 and Thr217 residues via hydrogen bonds, confirming that the TtFtsH-ADP and TmFtsH-ADP complexes have highly conserved fingerprint sequences in this region (Suno et al., 2006 ▶). On the other hand, the structure of the adenine and ribose-binding pocket shows greater variability. Gly164 and His343 in the TmFtsH-ADP complex and Ala159 in the TtFtsH-ADP complex each form hydrogen bonds with the adenine ring of ADP. By contrast, the adenine ring of the HpFtsH-ADP complex is stabilized primarily by a hydrogen bond between the N and O atoms of Ala173 and the N∊2 atom of His354, as well as by hydrophobic interactions with Leu218, Ile350 and Ala379.

Comparison of the FtsH ATPase domain structures

Compared with the TtFtsH-ADP complex [Protein Data Bank (PDB) ID: 2DHR], the differences in the rotation angles of the C-subdomain in the two Apo structures of TtFtsH (PDB IDs: 1IY2, 1IXZ) were 18.8° and 36.4°, respectively, which is similar to the differences seen in HslU. The crystal structures of HslU indicate that it assumes four different conformational states during the catalytic cycle, and that it rotates about 21.5° in going from the nucleotide-free state to the ADP-bound state (Wang et al., 2001 ▶). These large motions are often associated with the binding and release of nucleotides and, presumably, generate the force mediating translocation of protein substrates. It is noteworthy that the rotational difference relative to two subdomains within the structures of the HpFtsH Apo_A and other structures is considerably larger than that seen with HslU. Compared with the HpFtsH Apo_A, the rotational differences relative to the N-sub­domain between the C-subdomains in the HpFtsH Apo_B, HpFtsH-ADP and TtFtsH-ADP complexes were 126.6°, 134.5° and 126.7°, respectively.

By comparing the Apo structures (1IXZ, 1IY2 and HpFtsH Apo_A) with the structure of the TtFtsH-ADP complex, it is possible to predict the trajectories of the movement during the course of ADP binding. The trajectory of the movement of the HpFtsH ATPase domain was easily constructed by overlapping the Apo structures onto the N-subdomain of the TtFtsH-ADP complex ATPase domain. This analysis showed sequential rotations of 18.8°, 29.9° and 92.1° from theTtFtsH ADP complex to 1IY2, from 1IY2 to 1IXZ, and from 1IXZ to HpFtsH Apo_A, respectively (Fig. 2 ▶). This is the first description of the significant movement that occurs when a nucleotide binds to the active site of HpFtsH. It suggests that movement of this subdomain may generate the mechanical force needed for substrate translocation.

Figure 2

Conformational changes in the Apo structures of HpFtsH (PDB IDs: 1IXZ, 1IY2 and HpFtsH Apo_A) relative to the ATPase domain of the TtFtsH-ADP complex (PDB ID: 2DHR). The TtFtsH-ADP complex and Apo structures (1IY2, 1IXZ and HpFtsH Apo_A) are coloured grey, cyan, orange and pink, respectively. This illustrates the sequential rotations of 18.8°, 29.9° and 92.1° from the TtFtsH-ADP complex to Apo structure 1IY2, from Apo structure 1IY2 to Apo structure 1IXZ, and from Apo structure 1IXZ to HpFtsH Apo_A, respectively.