The Active N-terminal Region of p67 phox STRUCTURE AT 1.8 Å RESOLUTION AND BIOCHEMICAL CHARACTERIZATIONS OF THE A128V MUTANT IMPLICATED IN CHRONIC GRANULOMATOUS DISEASE*

Upon activation, the NADPH oxidase from neutro-phils produces superoxide anions in response to micro-bial infection. This enzymatic complex is activated by association of its cytosolic factors p67 phox , p47 phox , and the small G protein Rac with a membrane-associated flavocytochrome b 558 . Here we report the crystal struc- ture of the active N-terminal fragment of p67 phox at 1.8 Å resolution, as well as functional studies of p67 phox mutants. This N-terminal region (residues 1–213) consists mainly of four TPR (tetratricopeptide repeat) motifs in which the C terminus folds back into a hydrophobic groove formed by the TPR domain. The structure is very similar to that of the inactive truncated form of p67 phox bound to the small G protein Rac previously reported, but differs by the presence of a short C-terminal helix (residues 187–193) that might be part of the activation domain. All p67 phox mutants responsible for Chronic Granulomatous Disease (CGD), a severe defect of 1 resolution electrospray ionization mass spectrometry consistent with 100% selenium incorporation. CGD Mutants of p67 Nter — Mutants G78E and A128V of p67 phox -(1– constructed using site-directed mutagenesis (Stratagene kit) on the native cDNA cloned in the pET-15b vector. The mutant A128V expressed in E. coli BL21 (DE3) at 15 for 16 h instead of the 3 h at 37 °C employed for the native protein. The first steps of the purification were the same as for the native protein, but an additional purification step was carried out by gel filtration on a Superdex Hiload 16/60 column (Amersham Pharmacia Biotech) equlibrated in 20 m M Hepes, pH 7.5 and 200 m M NaCl. The protein was concentrated to 5 mg/ml with Centricon10. Proteolysis— Native p67 phox -(1–213) and the A128V mutant at concentration of 0.6 mg/ml submitted to limited proteolysis for one hour at 25 °C by trypsin Biochemicals) at a protease/protein ratio of 1:200 (w/w). Dichroism (CD)— CD were recorded on a Jobin Yvon dichrograph from 190 to 260 nm with a 1-mm path length cell and a thermostated cell holder. The protein concentrations were 0.5

all of which are prone to protein-protein interactions. Through these various domains, the cytosolic factors are able, upon stimulation, to associate, translocate to the membrane, and finally activate the electron transfer from NADPH to O 2 through the flavocytochrome b 558 . Over the past few years, serious efforts have been made to delineate the nature and the succession of intra-and intermolecular interactions between the cytosolic factors leading to the activated complex (3,4). Most, if not all, pairwise interacting protein domains have been identified, but the sequence of the molecular rearrangements as well as the structural modifications governing this cascade are still a matter of debate. It is known that Rac can interact in a GTP-dependent manner with the N-terminal region of p67 phox comprising amino acids 1-199 (5). Moreover, all known missense mutations in p67 phox responsible for CGD are localized in this N-terminal region (6). This region of p67 phox has been shown to contain four TPR motifs (2) that were first identified as degenerate 34 amino acid sequences present in a variety of proteins from bacteria to eukaryotes (7,8). One TPR motif consisting of two antiparallel ␣ helices termed A and B and a TPR domain (a succession of TPR motifs) adopts an overall superhelical fold (9). Using various binding assays, a first study suggested that p67 phox interacts with Rac only through 30 residues downstream of the TPR domain (10). A second study showed that Rac-p67 phox interaction involves the TPR domain itself (11). To unravel the properties and Rac binding sites on p67 phox , structures of the non-complexed p67 phox and of the Rac-p67 phox complex had to be deciphered. Recently, a first piece of the puzzle was achieved with the structure at 2.4 Å of the complex between the N-terminal region of p67 phox (residues 1-203) and Rac-GTP, showing an unusual interaction of a TPR domain with a partner protein (12).
In this study, we report the structure at 1.8 Å resolution of the catalytically active N-terminal region (residues 1-213) of p67 phox that contains four TPR motifs. We compared it to the Rac-p67 phox structure as well as to other TPR domain structures. We have also investigated the molecular and structural basis of the defects due to some of the CGD mutations. In particular, we report biochemical and biophysical evidence for the instability of the A128V mutant at physiological temperature, in agreement with our structural analyses.

EXPERIMENTAL PROCEDURES
Native and Selenomethionyl Protein Production-The cDNA for the N-terminal part of p67 phox corresponding to amino acids 1-213 was obtained by polymerase chain reaction and cloned into pET-15b (Novagen). The protein was expressed in Escherichia coli BL21 (DE3) and purified by affinity chromatography in two steps: first on a Ni 2ϩ -column equilibrated in 20 mM Hepes, pH 7.5, 250 mM NaCl and eluted with a linear gradient of imidazole and second on an SP-Sepharose column equilibrated in 20 mM Hepes, pH 7.5 and eluted with a linear gradient * This work was supported in part by the Association pour la Recherche sur le Cancer. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
CGD Mutants of p67 Nter -Mutants G78E and A128V of p67 phox -(1-213) were constructed using site-directed mutagenesis (Stratagene kit) on the native cDNA cloned in the pET-15b vector. The mutant A128V was expressed in E. coli BL21 (DE3) at 15°C for 16 h instead of the 3 h at 37°C employed for the native protein. The first steps of the purification were the same as for the native protein, but an additional purification step was carried out by gel filtration on a Superdex 200 Hiload 16/60 column (Amersham Pharmacia Biotech) equlibrated in 20 mM Hepes, pH 7.5 and 200 mM NaCl. The protein was concentrated to 5 mg/ml with Centricon10.
Limited Proteolysis-Native p67 phox -(1-213) and the A128V mutant at a concentration of 0.6 mg/ml were submitted to limited proteolysis for one hour at 25°C by trypsin (Roche Molecular Biochemicals) at a protease/protein ratio of 1:200 (w/w).
Circular Dichroism (CD)-CD spectra were recorded on a Jobin Yvon CD6 dichrograph from 190 to 260 nm with a 1-mm path length cell and a thermostated cell holder. The protein concentrations were 0.5 mg/ml. The spectra were recorded on one sample per protein from 15 to 35°C by increments of 5°C. At each temperature, the sample was incubated for 15 min.
NADPH oxidase activity-The NADPH oxidase activating potency was assessed in a semi-recombinant cell-free system (14) containing 3.5 g of membrane protein, 20 pmol of recombinant p47 phox , 20 -400 pmol of an N-terminal fragment of p67 phox , 2 mM MgCl 2 and 10 pmol of Rac in a final volume of 200 l. Rac was loaded with GTP-␥-S in the presence of 4 mM EDTA, followed by addition of MgCl 2 to 20 mM. An optimal amount of arachidonic acid (5-20 nmol) was added with strong agitation. After a 10-min activation, the elicited oxidase activity was assessed using the superoxide dismutase inhibitable cytochrome c reduction in the presence of 250 M NADPH and 100 M cytochrome c, followed at 550 nm using a Labsystem IEMS microplate reader.
Assay of Rac Binding-Rac fused to GST or GST alone were immobilized on glutathione-Sepharose beads (Amersham Pharmacia Biotech). Rac was loaded with GTP-␥-S on the beads. The beads were washed and incubated for 2 h at 4°C with a stoichiometric amount of p67 phox -(1-213) or the A128V mutant. After washing, proteins bound to the beads were analyzed by SDS-polyacrylamide gel electrophoresis.
Crystallization-Crystals of the native protein or the Se-Met protein were grown at 20°C by vapor diffusion in hanging drops by mixing equal volumes of protein (5-6 mg/ml) and reservoir solutions (17% polyethylene glycol monomethyl ether 2000, 100 mM sodium citrate, pH 4.5, 10% glycerol, and 10 mM dithiothreitol in the case of the Se-Met protein). The crystals grew as thin needles (20 ϫ 20 ϫ 200 m 3 ) and belong to the trigonal space group P3 1 (a ϭ b ϭ 67.7 Å, c ϭ 50.2 Å) with one molecule in the asymmetric unit. Crystals exhibit various amounts of merohedral twinning; however, non-twinned crystals could be selected after analyzing the diffraction data. 2 Data Collection and Processing-A native data set was collected (beamline ID14-EH1, ESRF-Grenoble) to 1.8 Å resolution (Table I), integrated with DENZO (15) and scaled with SCALA (16). A SAD data set was collected on a selenomethionyl-substituted crystal at the K absorption edge of selenium (beamline ID29, ESRF-Grenoble). Data were integrated and scaled with the DENZO/SCALEPACK programs. The program Shake and Bake (17) located 5 of 7 selenium atoms expected in the asymmetric unit.
Structure Determination and Refinement-Initial phases calculated with MLPHARE (16) using the 5 selenium sites allowed the location of a sixth selenium atom detected in an anomalous Fourier difference map. Final phases from MLPHARE resulted in a figure of merit of 0.37-2.8 Å resolution. Phases were extended to 1.8 Å on the native data set using the program DM (16) assuming a solvent content of 45% of the unit cell volume (Table I). 85% of the model was built automatically using the program wARP (18). Loop residues 151-158, residues 2-3 at the N terminus and 191-193 at the C terminus were added manually using the program O (19). The structure was refined with CNS (20) to a final crystallographic R factor of 18.2% and an R free of 20.5% at 1.8 Å resolution (Table II). The model consists of residues 2-193 of p67 phox , 160 water molecules, and one citrate anion, the citrate being essential for crystallization. All non-glycine residues are in the most favored or additionally allowed regions of the Ramachandran plot according to PROCHECK (21). The figures were prepared with MOLSCRIPT (22) and Raster3D (23).

RESULTS AND DISCUSSION
Overview of the Structure and Physiological Relevance-The N-terminal region of p67 phox used in this study was shown to be fully competent in NADPH oxidase activation both in vitro and in vivo (24,25). The electron density map obtained from SAD phasing was clearly interpretable (Fig. 1). p67 phox -(1-213) consists of four TPR motifs followed by an extended loop, which inserts into the hydrophobic groove formed by the helical organization of the TPRs (Fig. 2). The first three TPRs are contiguous and 16 residues (105 to 120) forming two antiparallel ␤-strands are inserted between TPR3 and TPR4. A comparison with other recently reported TPR domain structures (9, 26) highlights a remarkable conservation of the overall structure. For example, the TPR1 domain of Hop (PDB accession code 1ELW), composed of three TPR motifs, could be superimposed on three consecutive TPR motifs of p67 phox with rmsd values of 0.95 Å and 1.40 Å with the first three and the last three p67 phox TPRs, respectively. This indicates that the fold of TPR domains is highly conserved despite a rather small sequence identity; in addition, the insertion of amino acids 105-120 in p67 phox does not disrupt the superhelical structure of the TPR domain. Following TPR4, a helix (residues 156 -166) terminates the TPR domain and an extended structure (residues 168 -186) folds back into the internal hydrophobic groove of the superhelix. The structure ends with a short helix (residues 187-193),  mainly composed of polar residues. Residues 170 -185 interact extensively with residues belonging to the A helices of TPR motifs or to the inserted ␤-strands (Fig. 2b), probably stabilizing the overall structure. In particular, Arg-181 interacts with Gln-115. In the Rac-p67 phox complex, Arg-181 is replaced by Lys-181 and the same type of interaction is observed, consistent with the existence of a polymorphism (6). The sequence of p67 phox (1-213) is extremely rich in basic amino acids with 22 lysine and 6 arginine residues located mainly on the external surface of the super-helix. These residues are uniformly distributed at the surface without forming a highly basic patch. Residues downstream from Leu-193 are not visible in the electron density map. The recent structure of Rac-p67 phox -(1-203) (12) showed that all the residues of p67 phox involved in the complex formation belong to the inserted ␤-strands (Arg-102, Asn-104, Leu-106, Asp-108) or to the loops connecting TPR1 to TPR2 (S37) and TPR2 to TPR3 (D67, H69). The structures of both the non-complexed and complexed forms of p67 phox can be remarkably well superimposed with a rmsd of 0.57 Å on main chain atoms, showing that the interaction of p67 phox (1-213) with Rac does not require structural rearrangements of this N-terminal region. In particular, the residues involved in the interaction with Rac are totally accessible in the non-complexed structure, their side chains pointing toward the solvent. Although the two constructs of p67 phox (1-203 in the complex and 1-213 in this study) differ only by ten amino acids in length, the first was reported to be inactive whereas the second fully activates NADPH oxidase in vitro (24). This drastic difference could be related to the presence of the C-terminal ␣-helix (residues 187-193) present in our model that is not seen in the Rac-p67 phox complex structure, although amino acids 1-203 of p67 phox are present in the crystal. The folding of this helix is probably facilitated by the presence of amino acids up to 213. From functional studies of various truncated forms of p67 phox , amino acids 199 -210 of p67 phox were defined as an NADPH oxidase activation domain (24), and this region was reported to be involved in the regulation of electron transfer (27). The comparison of the structure of p67 phox -(1-213) to that of p67 phox -(1-203) in the Rac-p67 phox complex (12) suggests that the activation domain includes helix 187-193.
Mutations in p67 phox That Cause Chronic Granulomatous Disease-CGD is an inherited disorder of neutrophil function characterized by an increased susceptibility to infection because of a defect in the NADPH oxidase components. Mutations involving p67 phox are single cases and are located in the Nterminal region of the protein. An in-frame deletion of K58 was found to prevent Rac binding to p67 phox (28). Deletion of amino acids 19 -21 renders the protein inefficient for oxidase activation (29). These amino acids are located between helices A and B of TPR1 and are exposed to solvent. Various point mutations encountered in p67 phox of CGD patients (R77Q, G78E, A128V, D160V/K161E) were studied and shown to be associated with the absence of the protein in vivo (6,30). This observation accounts either for mRNA or protein instability or for an increased sensitivity to proteases. Amino acids Gly-78 and Ala-128 are located in similar positions in the ␣-helices of the TPR motifs (position 8 of helix A). Their mutations are likely to destabilize TPR packing. A sequence alignment of multiple TPRs shows that position 8 in helix A is restricted to glycine, alanine, or serine residues.
We produced G78E and A128V CGD mutants of p67 phox -(1-213). The G78E mutant was insoluble in bacteria. Because the A128V mutant showed lower solubility than the native protein, it was expressed at 15°C instead of 37°C. Surprisingly, this mutant was still able to bind Rac and to activate the NADPH oxidase in a cell-free system at 25°C (Fig. 3), suggesting correct folding. Moreover, the CD spectra at 25°C of the mutant and of the native proteins were identical and characteristic of ␣-helices, indicating no modification in the secondary structure (Fig.  4, inset). Altogether, these experiments do not explain the CGD phenotype of this mutant. To assess slight differences in the tertiary structure, the native protein and the A128V mutant were subjected to limited proteolysis by trypsin at 25°C. As shown in Fig. 5, the native protein is poorly degraded up to 60 min, whereas the A128V mutant shows notable degradation starting at 15 min. Additionally, the CD spectrum as a function of the temperature shows that the native protein is still stable at 40°C whereas the A128V mutant begins to loose its helical folding at 30°C (Fig. 4) and precipitates at 40°C.
The behavior of the two CGD mutants can be interpreted with respect to the structure of the native protein p67 phox -(1-213). The mutation G78E leads to a misfolding of the protein by steric hindrance within the TPR2 motif. In the less drastic A ribbon representation of the structure shows the four TPR motifs labeled in a, the A and B type helices are shown in b. The TPRs adopt an overall super-helix fold (colored from blue to green). The C terminus in red is in an extended conformation locked in the hydrophobic groove formed by the TPRs and ends with a short helix. The inserted ␤-strands between TPR 3 and 4 are shown in yellow. a and b are two perpendicular views of the molecule. mutant A128V, Val-128 would be located at 2.2 Å of the carbonyl of Ala-140 or of the hydroxyl of Tyr-172 (Fig. 6). This hydroxyl in the native structure is stabilized by two hydrogen bonds. Although the environment should be able to accommodate this moderate steric change, the hydrogen-bond network is perturbed and the structure of the A128V mutant is weakened. Within the cellular environment at 37°C, the decrease in stability because of this mutation probably leads to misfolded p67 phox protein. This mutation is akin to other point mutations introduced at position 8 of the A helix of a TPR motif in p62 cdc23 (Ala3 Thr or Gly3 Asp), which also confer thermolability to the protein (31).
Thus, the structure of the active N-terminal domain of p67 phox -(1-213) highlights a short ␣-helix that participates to the NADPH oxidase activation domain. Interestingly, most of the CGD mutants of p67 phox are located in the N-terminal region of p67 phox . The structural and functional studies reported here permit an elucidation of the instability of the G78E and A128V mutants and may be extended to mutations that affect TPR folding. The coordinates and the structure factors have been deposited in the Protein Data Bank with ID code 1hh8.