Expression, limited proteolysis and preliminary crystallographic analysis of IpaD, a component of the Shigella flexneri type III secretion system

IpaD, the putative needle-tip protein of the S. flexneri type III secretion system, has been crystallized in a variety of crystal forms using in-drop proteolysis. Native and selenomethionine-labelled data collection and preliminary analyses are reported.


Introduction
Shigella flexneri, the causative agent of shigellosis, is a Gram-negative bacterial pathogen that initiates infection by invading cells of the colonic epithelium in a process which is dependent on a type III secretion system (T3SS; Phalipon & Sansonetti, 2003). The T3SS is composed of a basal body, which traverses both bacterial membranes, and an external needle through which effector proteins are secreted (Blocker et al., 2001). During S. flexneri infection, secretion is activated by contact of the tip of the needle complex with host cells, resulting in formation of a pore in the host-cell membrane that is contiguous with the needle. Other effector proteins are injected via this apparatus directly into the host-cell cytoplasm (for a review, see Johnson et al., 2005).
Of the T3SS effectors, IpaB, IpaC and IpaD are essential for invasion (Menard et al., 1993). IpaB and IpaC have been demonstrated to insert into the host-cell membrane and are therefore suggested to form the translocation pore (Blocker et al., 1999). Cells lacking IpaD are not only incapable of pore insertion and invasion, but also demonstrate impaired effector secretion control (Picking et al., 2005). IpaD was originally proposed to form a plug in the T3SS with IpaB (Menard et al., 1994), but more recent data has demonstrated that its role is more complex (Picking et al., 2005). Deletions within the N-terminal third of the molecule do not affect invasion of the bacteria into host cells, although the insertion of IpaB/IpaC into membranes is slightly impaired. However, even short deletions (five residues) in the C-terminus of IpaD completely abolish the invasive phenotype and pore insertion. These observations, combined with the functional homology between IpaD and LcrV, a protein recently shown to be assembled at the tip of the Yersinia pestis T3SS needle (Mueller et al., 2005), suggest that IpaD may play a role in regulating insertion of the IpaB/IpaC translocon pore from the tip of the S. flexneri needle.
Our laboratory has recently determined the crystal structure of MxiH, the needle subunit, and assembled a molecular model of the T3SS needle by docking it into a 16 Å EM reconstruction of the S. flexneri needle (Cordes et al., 2003;. As part of an ongoing study into the interactions of the T3SS with host cells, we have now expressed, purified and crystallized IpaD. The three-dimensional structure of IpaD, in conjunction with the structure of the MxiH needle and other tip proteins such as LcrV (Derewenda et al., 2004), will hopefully allow us to delineate the mechanism of translocon insertion.

Expression and purification
The DNA fragment of the ipaD gene encoding residues 15-332 was produced using the methods described previously (Harrington et al., 2003;Picking et al., 2005) and subcloned into the pET-15b vector. This construct placed the gene in frame with an N-terminal hexahistidine tag. IpaD NÁ14 was expressed in Escherichia coli BL21 (DE3) cells grown in LB media containing 100 mg ml À1 ampicillin. Cells were grown at 310 K to an A 600 nm of 0.6 and protein overexpression was induced by the addition of 1.0 mM IPTG. After 3 h, cells were harvested by centrifugation (4000g, 15 min, 277 K) and pellets were frozen at 193 K. Cell pellets were resuspended in lysis buffer [20 mM Tris-HCl pH 7.5, 500 mM NaCl and Complete EDTA-free protease inhibitor cocktail (Roche)] and lysed using an Emulsiflex-C5 Homogeniser (Glen Creston, UK). The resultant cell suspension was centrifuged (20 000g, 30 min, 277 K) and the soluble fraction was applied onto a pre-charged 5 ml His-Trap HP nickel-affinity column (Amersham Biosciences). Protein was eluted using a gradient of 0-1 M imidazole in lysis buffer. Fractions containing IpaD NÁ14 were further purified by gel-filtration chromatography using a HiLoad 26/60 Superdex 75 column (Amersham Biosciences) equilibrated in 20 mM Tris-HCl pH 7.5, 100 mM NaCl, 10 mM DTT. IpaD NÁ14 eluted as a single monomeric peak and was demonstrated to be pure by SDS-PAGE; mass-spectrometric analysis (data not shown) confirmed the molecular weight of the protein minus the N-terminal methionine (37 374 Da). Fractions containing purified IpaD NÁ14 were pooled, concentrated using 5 kDa molecular-weight VivaSpin centrifugal concentrators (VivaScience, Sartorius Group) to approximately 70 mg ml À1 and stored at 193 K in 50 ml aliquots.
SeMet-labelled IpaD NÁ14 was produced by expression in the E. coli met À auxotrophic strain B834 (DE3). Cultures were grown in LB media to an A 600 nm of 0.2, pelleted (4000g, 15 min, 277 K) and washed in phosphate-buffered saline three times. Washed cells were used to inoculate SelenoMet Medium Base containing SelenoMet Nutrient Mix, l-selenomethionine (40 mg ml À1 ; Molecular Dimensions Ltd, UK) and ampicillin (100 mg ml À1 ). Cultures were grown at 310 K until an A 600 nm of 0.6 was reached, whereupon the solutions were cooled to 293 K and protein overexpression was induced by the addition of 1.0 mM IPTG. After 16 h, cells were harvested and SeMet-labelled protein was purified as described above. Mass-spectrometric analysis confirmed full incorporation of SeMet, including the N-terminal methionine (data not shown), in contrast to the native protein. We assume this is a consequence of the incorporation of the selenium making the N-terminal peptide bond less likely to be cleaved.

Crystallization and limited proteolysis
Initial crystallization conditions were obtained by sparse-matrix screening (Jancarik & Kim, 1991) using the sitting-drop vapourdiffusion technique. Drops were prepared by mixing 0.2 ml protein solution (70 mg ml À1 , 20 mM Tris-HCl pH 7.5, 100 mM NaCl, 10 mM DTT) with 0.2 ml reservoir solution using a Tecan crystallization robot (Tecan UK, Theale, UK) and were equilibrated against 100 ml reservoir solution at 293 K. Initial crystals of IpaD NÁ14 (crystal form 1) grew in 3 d in condition No. 33 of Molecular Dimensions Screen 1 [30%(w/v) PEG 4000, 0.1 M Tris-HCl pH 8.5, 0.2 M MgCl 2 ]. Subsequent screening using handmade conditions yielded diffractionquality crystals in PEG 4000 concentrations ranging from 25 to 28%(w/v). Mercury derivatives were prepared by adding 1 ml of 10 mM HgCl 2 in reservoir solution to the 0.4 ml drop. Crystals were soaked for 3 h at 293 K.
One IpaD NÁ14 crystal form (crystal form 2) grew over a period of 12 months in 10%(v/v) PEG 400, 0.1 M NaCl, 0.1 M Tris-HCl pH 7.0 (Fig. 1a) and was discovered to be an N-terminal truncation of IpaD NÁ14 . In an attempt to reproduce this and find new crystal forms, in-drop limited proteolysis was carried out (as in Gaur et al., 2004). The broad-spectrum serine protease subtilisin Carlsberg from Bacillus licheniformis (Sigma-Aldrich) was added to IpaD NÁ14 in a 1:5000 enzyme:protein ratio by weight before sparse-matrix screening was carried out as described above. Diffraction-quality native crystals were obtained in

Results and discussion
Intact IpaD NÁ14 yielded plate-shaped crystals (CF-1) belonging to space group P2 1 2 1 2 1 ( Table 1). Calculation of a native Patterson map revealed a peak at $50% of the origin on Harker section v = 0.5, consistent with two molecules per asymmetric unit related by a pure translation (Fig. 2). The value of the Matthews coefficient is 2.1 Å 3 Da À1 for two molecules, corresponding to a solvent content of 42% (Matthews, 1968). Attempts to grow diffraction-quality SeMet crystals in this crystal form were unsuccessful. Soaking the crystals with mercury increased the quality of diffraction of the crystals but did not yield a reliable set of Hg sites, despite using the available Patterson-solving software.
During optimization of CF-1 a second crystal form was observed (CF-2) which took 12 months to grow and belongs to space group C2 (Fig. 1). One crystal was used for data collection and the remainder were re-solubilized, run on SDS-PAGE and N-terminal sequenced. This revealed a major band corresponding to an N-terminal truncation beginning at residue 144 (data not shown). The value of the Matthews coefficient is 2.4 Å 3 Da À1 for two molecules, corresponding to 49% solvent content.
The contents of CF-2 revealed that the N-terminal third of IpaD is susceptible to degradation and suggested that limited proteolysis could trim these regions. In an attempt to reproduce CF-2 and search for new crystal forms, the broad-spectrum protease subtilisin Carlsberg was added in trace amounts to IpaD NÁ14 before screening. Plate-shaped crystals were grown in three new crystal forms (CF-3, CF-4 and CF-5; Table 1) and analysis by SDS-PAGE revealed various degrees of N-terminal truncation (Fig. 1b). The value of the Matthews coefficient is 2.1 and 2.5 Å 3 Da À1 for CF-3 and CF-4, respectively, for two molecules per asymmetric unit and is 2.6 Å 3 Da À1 in CF-5 for four molecules.
Anomalous difference Patterson maps for SeMet derivatives of CF-4 were calculated within autoSHARP (Vonrhein et al., 2005) using (E 2 À 1) coefficients and strict outlier rejection. The positions of ten Se atoms were obtained using the Patterson-solving software within  Table 1 IpaD X-ray diffraction data-collection statistics.
Values in parentheses are for the highest resolution shells.

Figure 2
Harker section v = 0.5 of the native Patterson map of IpaD crystal form 1 calculated using PATTERSON (Collaborative Computational Project, Number 4, 1994) at 3.5 Å . The map is drawn with a minimum contour level of 1.5 with 0.5 increments.
autoSHARP and were refined using SHARP. The ten sites clustered into two sets of five related by the twofold peak in the self-rotation function (Fig. 3). Structure determination and refinement in the various crystal forms is under way.
SJ is funded by a Medical Research Council of the UK grant (G0400389) to SML. PR is funded by a Wellcome Trust Grant (No. 077082) to SML and PR. JED is funded by an Australian National Health and Medical Research Council CJ Martin Postdoctoral Fellowship (ID-358785). WDP's laboratory was supported by PHS grants AI034428 and RR017708 and the University of Kansas Research Development Fund. AB is funded by a Guy G. F. Newton Senior Research Fellowship. We are grateful to Ed Lowe for assistance with data collection and to Tony Willis for N-terminal sequencing. Marc Morgan and Jenny Gibson assisted with the crystallization robot and we are very grateful to Clemens Vonrhein for masterful processing of the orthorhombic native data.