Structure of the Nitrosomonas europaea Rh protein

Li et al. 10.1073/pnas.0709710104.

Supporting Information

Files in this Data Supplement:

SI Table 1
SI Figure 3
SI Text
SI Figure 4
SI Figure 5
SI Figure 6
SI Figure 7
SI Table 2
SI Figure 8

SI Figure 3

Fig. 3. Comparison of the N. europaea Rh and E. coli AmtB proteins, in stereo. (A) N. europaea Rh protein colored in rainbow from blue to red. (B) Overlap of N. europaea Rh protein (in cyan) and E. coli AmtB protein (in magenta). The N. europaea Rh protein and E. coli AmtB are shown as Ca traces. The sequence number of the Rh protein is labeled every 50 residues.

SI Figure 4

Fig. 4. Transmembrane a-helices of the N. europaea Rh protein (Upper) and E. coli AmtB (Lower). To highlight the extent of proline kinking, the orientations of the a-helices have been adjusted so that the cytoplasmic side of the a-helix is vertical with the page and the kink is perpendicular.

SI Figure 5

Fig. 5. Comparison of the extracellular vestibules of the N. europaea Rh (Left) and E. coli AmtB (Right). (Upper) View from periplasm. (Lower) View from membrane bilayer. The proteins are depicted as a ribbons diagram with the side chains shown in stick form. The ethylene glycol and methylammonium ligands bound to N. europaea Rh and E. coli AmtB, respectively, are shown in ball-and-stick form. The ribbons trace and carbon atoms are colored in cyan for the Rh protein and green for AmtB. The remaining atoms are colored in CPK.

SI Figure 6

Fig. 6. Fit of different models to the electron density in the extracellular vestibule. Weighted density maps using 2mF_obs-DF_calc (in blue, contoured at 2.0s) and mF_obs-DF_calc (in green, contoured at 3.5s; and in red, contoured at -3.5s) with terms as determined by REFMAC5 (1). Proteins are shown as stick models, and waters, CO₂, and formate are shown as ball-and-stick models. (A-D) Electron density maps obtained from refinement of native Rh protein crystals with ligand density modeled as no ligands (A), ethylene glycol (B), two waters (C), and two waters in dual conformations (D). (E-H) Electron density maps obtained from refinement of CO₂-pressurized Rh protein crystals with ligand density modeled as no ligands (E), ethylene glycol (F), two waters (G), and two waters in dual conformations (H). All of the maps were calculated after 30 cycles of restrained refinement with REFMAC5 (1).

1. Murshudov G, Vagin A, Dodson E (1997) Acta Crystallogr D 53:240-255.

SI Figure 7

Fig. 7. Fit of different models to the electron density in the CO₂ binding pocket. Weighted density maps using 2mF_obs-DF_calc (in blue, contoured at 1.0s) and mF_obs-DF_calc (in green, contoured at 3.0s; and in red, contoured at -3.0s) with terms as determined by REFMAC5 (1). Proteins are shown in stick form, and waters, CO₂, and formate are shown as ball-and-stick models. (A-E) Electron density maps obtained from refinement of native Rh protein crystals with ligand density modeled as no ligands (A), one water (B), two waters (C), CO₂ (D), and formate (E). (F-J) Electron density maps obtained from refinement of CO₂-pressurized Rh protein crystals with ligand density modeled as no ligands (F), one water (G), two waters (H), CO₂ (I), and formate (J). All of the maps were calculated after 30 cycles of restrained refinement with REFMAC5 (1).

1. Murshudov G, Vagin A, Dodson E (1997) Acta Crystallogr D 53:240-255.

SI Figure 8

Fig. 8. C termini length distribution of Rh protein subfamilies. Shown is length distribution of Rh protein C termini for the different subfamilies based on the number of residues after the conserved proline (Pro-390 in N. europaea Rh protein).

SI Text

Bacterial Host Strains and Membrane Protein Expression Vectors. The full-length N. europea Rh gene (GenBank accession no. AY377923), a gift from C.-H. Huang, was first cloned into a membrane protein expression cassette that contains the pmo1 promoter derived from the pmo1 operon of Methylococcus capsulatus (Bath), the membrane targeting sequence (30 aa) derived from the leader sequence of the M. capsulatus (Bath) pmoC1 protein, and the transcription terminator sequence of the M. capsulatus (Bath) pmo1 operon. The Rh protein expression cassette was subsequently cloned into a broad host range vector for expression in an engineered M. capsulatus (Bath) host strain in which the pmo1 locus was deleted. The vectors containing this cassette could be used to overexpress recombinant membrane-bound Rh protein in E. coli.

The broad host range vector containing the Rh membrane protein expression cassette was introduced into M. capsulatus (Bath) by conjugation using the E. coli S17-1 strain as donor. For expression in E. coli, the vectors were introduced into the cells by heat-shock method. The E. coli host strains used for recombinant membrane protein production are mutants that have been screened for their ability to tolerate elevated expression levels of toxic membrane proteins. These E. coli mutants were then subsequently engineered to contain extra copies of genes involved in membrane protein biosynthesis. The extra membrane protein biosynthesis genes were placed under the control of the lac promoter so their expression could be induced by the addition of IPTG.

Expression of Recombinant Rh Protein and Membranes Isolation. The host strains (M. capsulatus and E. coli) harboring the Rh membrane protein expression vector were cultivated in appropriate mediums (NMS for M. capsulatus and LB for E. coli) that also contain additional hydrocarbons and other nutrients to relieve the growth arrest when recombinant membrane proteins are overexpressed. The overexpression of recombinant Rh protein was induced by addition of Cu(II) ions in the case of M. capsulatus host or by IPTG in the case of the E. coli host. The cells were grown to an OD of at least 2.0 and then harvested. The cells were washed, resuspended with buffer A (50 mM phosphate/200 mM NaCl/10% wt/vol glycerol, pH 7.2) and lysed by using a French press (15,000 psi, two passes). Cell debris and unlysed cells were removed by centrifuging the lysate at 6,000 rpm for 10 min. The membrane fractions were then pelleted from the cell-free extract by ultracentrifugation (45,000 rpm for 2 h). The membranes were collected and resuspended in buffer A by using a Dounce homogenizer.

Purification of Recombinant Rh Protein. The washed membranes were treated with n-octylglucopyranoside (5% wt/vol), incubated on ice for 30 min to ensure complete solubilization. The solubilized membranes were then centrifuged at 15,000 rpm to remove unsolubilized materials. The clear supernatant was the applied onto a Talon cobalt metal affinity resin column that had been preequilibrated with buffer B (50 mM phosphate/200 mM NaCl/10% wt/vol glycerol/0.04% wt/vol n-octyl-glucoside, pH 7.2). The column was washed extensively (30-40 times column volume) with buffer B and buffer B containing 10 mM imidazole. The recombinant Rh protein was then eluted out of the column by using imidazole step gradient with buffer B containing 50, 75, 100, 150, and 250 mM imidazole. Fractions containing pure Rh protein as judged by SDS/PAGE were pooled, concentrated to at least 5 mg/ml, and frozen in liquid nitrogen.

Preparation of Se-Met Rh Derivative. The Se-Met Rh derivative was prepared by using the E. coli expression method. The E. coli host strain harboring the Rh expression vector was cultivated in a minimal medium containing essential amino acids, other nutrient supplements (vitamins, trace elements), and glucose as carbon source. At an OD of 0.6, the temperature was dropped to 24°C, and, once the temperature was equilibrated, L-selenomethionine (25 mg/liter) was added to the medium and IPTG was added 15 min later. The cells were harvested after 18-24 h. The Se-Met Rh derivative was purified by using the same procedures as the wild-type protein.

Native and CO₂-Bound Rh Protein Crystals Preparation. Crystals of native Rh protein were grown at room temperature in 14.5% PEG 2000 MME, 1 mM EDTA, and 0.1 M Mes buffer (pH 6.45) via the vapor diffusion method. Crystals of the Se-Met derivative were grown in 25.8% PEG 550 MME and 0.1 M Mes buffer (pH 6.20) at 16°C. To reduce ice formation, fomblin oil was added onto the crystallization drop before mounting and flash-cooling in liquid nitrogen.

CO₂-bound Rh protein crystals were derived from native Rh protein crystals. Fresh crystals lifted onto Hampton loops in fomblin were placed under 200 psi CO₂ for 2 min using the gas pressure cell developed at SSRL (1). Upon release of the gas pressure, these CO₂-treated crystals were immediately flash-cooled in liquid nitrogen. Longer pressurization times were explored but were found to damage the crystal.

Rh protein crystals were also prepared in the presence of ammonium or methyl ammonium. However, the resolution (3.4 Å) of even the best of these crystals, grown from (MeNH₄)Cl, was insufficient to identify novel methyl ammonium binding sites. No conformational changes of the protein were observed.

Data Collection and Structure Determination. All data sets were collected at the Stanford Synchrotron Radiation Laboratory, using the BLu-ICE interface (2). Native Rh protein and Se-Met derivative crystals were collected on beamline 9-2 using a MarMosaic-325 CCD detector, and CO₂-treated crystals were collected on beamline 9-1 using a Quantum-315 CCD detector.

The crystallographic results are summarized in SI Table 1. All crystals belong to the space group H3 with almost identical cells. The data were integrated and scaled with MOSFLM (3) and SCALA (4). The phases of the Se-Met Rh-protein were obtained by the multiwavelength anomalous diffraction (MAD) method using SHELX (5, 6) and SOLVE (7) implemented by the automated MAD phasing script developed at SSRL (http://smb.slac.stanford.edu/research/developments/MAD-scripts). The initial model was built by using the program COOT (8) and refined to 2.9-Å resolution by using the program REFMAC5 (3) within the CCP4 package using the CCP4i interface (4). The C-terminal helix H12 was visible in the experimental Se-Met phased map with unambiguous electron density protruding from the cytoplasmic side of the protein.

The first 30 residues at the N terminus of the Se-Met-labeled Rh protein were not observed. The program SignalP 3.0 (9) predicts a probable signal peptide cleavage site between residues Ala-26 and Val-27. Cleavage of the N terminus was confirmed by MALDI mass spectrometry. The Rh protein was mixed with 5 volumes of the matrix sinapinic acid and then used for mass determination on a Bruker Reflex III MALDI-TOF. The protein molecular mass was determined to be 42,859 ± 504 Da, which is smaller than the mass calculated from the coding sequence and consistent with the protein mass predicted by SignalP 3.0.

The phases for the native structure and CO₂-bound structure were obtained by molecular replacement with MOLREP (3) and refined to 1.85 Å by using Refmac5 with 5% set aside for the free-R calculation. Solvent molecules were identified by the automated package ARP/wARP. One site with 2F_O-F_C density consistent with two water molecules in the native Rh protein structure had a much longer strand of 5.0 s density in the F_O-F_C map of the CO₂-treated crystal. A CO₂ ligand was built and refined. The occupancy of CO₂ ligand was refined with Phenix CCI Apps (13) as 0.73 with B factor 24.7 Å². The CO₂-bound Rh protein structure was nearly identical to the native with an rmsd of 0.1 Å for 388 Ca atoms.

The quality of the final structures was evaluated by using the program PROCHECK (14). Additionally, composite omit maps were calculated by using CNS (15) and compared as a check of potential model bias. The secondary structures are assigned as follows: (helix) residues, (M1) 34-58, (M2) 66-89, (M3) 102-122, (M4) 128-148, (M5) 170-184, (M6) 186-190, (M7) 198-220, (M8) 228-253, (M9) 260-274, (M10) 283-311, (M11) 320-338, (short helix 1) 343-369, (short helix 2) 380-384, (H12) 395-411.

1. Soltis SM, Stowell MHB, Wiener MC, Phillips GN, Rees DC (1997) J Appl Crystallogr 30:190-194.

2. McPhillips T, McPhillips S, Chiu H, Cohen A, Deacon A, Ellis P, Garman E, Gonzalez A, Sauter N, Phizackerley R, et al. (2002) J Synchrotron Radiat 9:401-406.

3. Murshudov G, Vagin A, Dodson E (1997) Acta Crystallogr D 53:240-255.

4. Collaborative Computational Project No. 4 (1994) Acta Crystallogr D 50:760-763.

5. Schneider TR, Sheldrick GM (2002) Acta Crystallogr D 58:1772-1779.

6. Terwilliger T (2000) Acta Crystallogr D 56:965-972.

7. Terwilliger TC, Berendzen J (1999) Acta Crystallogr D 55:849-861.

8. Emsley P, Cowtan K (2004) Acta Crystallogr D 60:2126-2132.

9. Bendtsen J, Nielsen H, von Heijne G, Brunak S (2004) J Mol Biol 340:783-795.

10. Avent ND, Butcher SK, Liu W, Mawby WJ, Mallinson G, Parsons SF, Anstee DJ, Tanner MJ (1992) J Biol Chem 267:15134-15139.

11. Eyers SA, Ridgwell K, Mawby WJ, Tanner MJ (1994) J Biol Chem 269:6417-6423.

12. Avent ND, Liu W, Warner KM, Mawby WJ, Jones JW, Ridgwell K, Tanner MJ (1996) J Biol Chem 271:14233-14239.

13. Adams PD, Grosse-Kunstleve RW, Hung LW, Ioerger TR, McCoy AJ, Moriarty NW, Read RJ, Sacchettini JC, Sauter NK, Terwilliger TC (2002) Acta Crystallogr D 58:1948-1954.

14. Laskowski RA, Macarthur MW, Moss DS, Thornton JM (1993) J Appl Crystallogr 26:283-291.

15. Brunger AT, Adams PD, Clore GM, DeLano WL, Gros P, Grosse-Kunstleve RW, Jiang JS, Kuszewski J, Nilges M, Pannu NS, et al. (1998) Acta Crystallogr D 54:905-921.