Das et al. 10.1073/pnas.0609033103. |
Fig. 5. Structural assessment of RIa (residues 119-244) based on the correlation between measured and computed Ca (a) and Cb (b) chemical shifts.
Fig. 6. Differences between the cAMP- and C-bound structures of RIa (residues 119-244) (1, 2) in terms of local rmsds mapped into the structure of cAMP-bound RIa (residues 119-244) (1). Residues in dark blue are characterized by local rmsds greater than the average + 1 SD, and residues in light blue correspond to local rmsds between the average and the average + 1 SD. All remaining residues are colored in light gray. Residues that interact directly with the C-subunit (2) are marked by a sphere and define three major regions of C-interaction (i.e., a-B', C-helix and the a-XN/A-helix loop) indicated by black dashed lines. Sites that do not interact with the C-subunit directly but are subject to significant conformational changes upon C binding are highlighted with red lines. The W188 region is not highlighted because conformational changes at this site may, at least in part, be accounted for by C-independent intrinsic variability as explained in SI Materials and Methods. Selected secondary structure elements and residues are labeled. cAMP is shown as an atom-type color-coded CPK model. All figures were created by using the program MOLMOL (3) and atomic coordinates are from the Protein Data Bank entry 1RGS (1).
1. Su Y, Dostmann WR, Herberg FW, Durick K, Xuong NH, Ten Eyck L, Taylor SS, Varughese KI (1995) Science 269:807-813.
2. Kim K, Xuong NH, Taylor SS (2005) Science 307:690-696.
3. Koradi R, Berman HM, Wüthrich K (1996) J Mol Graphics 14:51-55.
| Intramolecular hydrogen bonds | Intermolecular contacts‡ | |
Residue | Tight* | Loose† |
|
G199 | -- | CO: R209 Guanidinium | NH: cAMP Ribose O2' |
|
|
| CO: cAMP Exocyclic O |
E200 | NH: E200 OE1 | OE1: L201 NH | OE1: cAMP Ribose O2' |
CO: I204 NH | CO: L203 NH | ||
| OE2: R241 Guanidinium | CO: I204 NH |
|
L201 | CO: Y205 NH | NH: E200 OE1 | -- |
A202 | -- | CO: G206 NH | NH: cAMP Exocyclic O |
|
| CO: T207 NH |
|
L203 | CO: R230 Guanidinium | NH: E200 CO | -- |
I204 | NH: E200 CO | -- | -- |
Y205 | NH: L201 CO | -- | -- |
G206 | -- | NH: A202 CO | -- |
T207 | -- | NH: A202 CO | OG1: Water 701‡ |
P208 | -- | -- | CO: Water 701‡ |
R209 | CO: G166 NH | Guanidinium: D170 NH | Guanidinium: cAMP Exocycl. O |
NH: D167 CO | Guanidinium: G199 CO | ||
| Guanidinium: N171 CO |
|
|
A210 | -- | -- | NH: cAMP Exocyclic O NH: Water 701‡ |
* With H---O distance < 2.40 Å and H-N....O angle < 35o in the 1RGS structure (1). These criteria may not be fulfilled by all salt bridges.
† With 2.40 Å ≤ H---O distance < 2.85 Å and 35o ≤ H-N....O angle < 47o in the 1RGS structure (1). These criteria may not be fulfilled by all salt bridges.
‡
From ref. 2. The water molecule W701 is also hydrogen-bonded to one of the cAMP Exocyclic oxygen atoms.1. Su Y, Dostmann WR, Herberg FW, Durick K, Xuong NH, Ten Eyck L, Taylor SS, Varughese KI (1995) Science 269:807-813.
2. Wu J, Jones JM, Nguyen-Huu X, Ten Eyck LF, Taylor SS (2004) Biochemistry 43:6620-6629.
Supporting Materials and Methods
Expression, Purification, and NMR Sample Preparation of RIa (Residues 119-244).
The cells were grown at 37°C until an OD600 of 0.8 was reached, at which point bacteria were induced by adding 0.1 mM isopropyl b-D-thiogalactoside for 20 h at 22°C (1). After harvesting the cells were resuspended in the lysis buffer (20 mM Mes, pH 6.5/100 mM NaCl/2 mM EGTA/2 mM EDTA/5 mM DTT) with protease inhibitors and lysed by passing three times through a French press (1). The cell debris was cleared by centrifugation at 20,000 × g, and the protein was then fractionated by 40% ammonium sulfate (1). The precipitated protein was collected by centrifugation and resuspended with lysis buffer. The protein solution was then incubated with a cAMP-conjugated resin overnight (1). The resin was washed with high salt buffer (lysis buffer containing 700 mM NaCl) before eluting the protein with 25, 35, and 40 mM cAMP (Sigma) (1). The protein was further purified and buffer-exchanged to 20 mM KH2PO4 pH 6.5, 100 mM NaCl, and 1 mM cAMP with a FPLC gel filtration column (HiLoad 16/26 Superdex-S75 column from GE Healthcare) (1). Then the purified protein was dialyzed against 20 mM KH2PO4 (pH 6.5), 100 mM NaCl without cAMP for 36 h with three buffer changes and finally dialyzed against the cryoprobe compatible buffer used for the NMR experiments (NMR buffer: 50 mM Mes, pH 6.5/100 mM NaCl /0.02% NaN3). A first NMR sample (defined as sample A) was prepared by concentrating the protein to 0.1 mM and adding D2O to a final concentration of 5% (vol/vol). Potential high molecular weight oligomers were removed by passing the sample through a 100-kDa spin filter (Millipore). Sample B containing excess cAMP was prepared by adding 1 mM cAMP (final concentration) to an aliquot of sample A. Both samples with and without excess cAMP (i.e., B and A) were stable.The sample of RIa (residues 119-244) used for the Nz experiment, which requires detectable amounts of free protein, was prepared through an unfolding/refolding protocol: 8 M urea was added to sample A and it was dialyzed against 500 ml of a 6 M urea solution for 18 h with three changes of the dialysis buffer at 6-h intervals. The urea concentration was brought down to 2 M by stepwise addition of the NMR buffer solution, while incubating on ice and stirring occasionally. The partially refolded protein was then dialyzed against 1 and 0.5 M urea solutions, respectively, for 4 h each time. The protein was finally dialyzed against 50 mM Mes (pH 6.5), 100 mM NaCl, and 0.02% NaN3 to remove the urea before the start of the NMR experiments. This unfolding/refolding protocol resulted in a sample (C) with detectable amounts of free RIa (residues 119-244). After adding excess cAMP, the HSQC of this sample becomes superimposable to that of sample B above (data not shown), indicating successful refolding of the protein.
NMR Spectroscopy.
The probe temperature was calibrated by using a thermocouple and an ethylene glycol sample. For the 15N and 1H dimensions 128 and 512 complex points and spectral widths of 31.8 and 14.2 ppm were used, respectively, unless otherwise specified. In all spectra, the 1H and 15N carrier frequencies were centered on the water resonance and in the middle of the amide 15N region, respectively. In all experiments, 15N decoupling during 1H acquisition was obtained through a GARP pulse train with a 1.32-kHz RF pulse strength. A 100-point sine-bell shape was used for all pulse-field gradients. The Xwinnmr (Bruker Inc.) or NMRPipe (2) programs were used for spectral processing. Phase-shifted squared sine bell window functions and zero filling were applied in both dimensions, unless otherwise indicated. The Gaussian line fitting protocol implemented in Sparky 3.111 (3) was used for the measurement of the cross-peak fit heights, unless otherwise specified. Estimations of the fit heights errors were calculated based on the SD of the cross-peak intensity difference distribution obtained from replicate spectra, as explained (4). 1H ppm values were calibrated by using DSS, whereas the 15N chemical shifts were indirectly referenced through the gN/gH ratio (5). 2D-HSQC experiments were carried out with 128 dummy scans, eight scans, and an interscan delay of 1 s using sensitivity, gradient, and water-flip back enhanced pulse sequences, unless otherwise specified. Standard triple-resonance experiments were used to assign the backbone resonances of cAMP-bound RIa (residues 119-244) as indicated (6).Nz Exchange.
When the exchange between the free and bound states occurs in the ms time scale, the binding-induced chemical shift changes can be probed either indirectly by variations in the relaxation dispersion resulting from slight perturbations of the binding equilibrium (7) or more directly by Nz-exchange spectroscopy. The latter approach requires higher concentrations of the poorly soluble cAMP-free protein as compared with the relaxation dispersion method and therefore it relies on less stable samples; however, when the Nz-spectrum can be acquired with sufficient sensitivity and resolution it provides a more direct determination of the binding-induced chemical-shift changes. Here, the HSQC cross-peaks for the free state of RIa (residues 119-244) were assigned through NZ exchange spectra (8, 9) based on the assignment of cAMP-bound RIa (residues 119-244). For the acquisition of the NZ exchange data sample C was used. The relaxation delay between subsequent scans was 2 s and 128 scans were accumulated per t1 transient after 128 dummy scans leading to a total acquisition time of ~21.5 h. Some aggregation was noticed during the acquisition of the Nz spectrum; however, a control HSQC spectrum acquired on a 10-fold diluted sample indicates that the partial precipitation does not significantly affect the measured chemical-shift changes. The resolution along the indirectly detected dimension was enhanced through semiconstant time frequency labeling (10) and linear prediction. The compounded 1H,15N free-bound chemical-shift variation was computed as ((Dd 1HN)2 + (Dd15NH/6.5)2)1/2, as explained (11).H/D Exchange.
The sample for H/D exchange was prepared by passing the concentrated (1 mM) sample A through a Sephadex G10 column preequilibrated with 50 mM Mes (pH 6.5), 100 mM NaCl in 100% D2O at 4°C. The dead time for the H/D exchange was reduced by preoptimizing the HSQC acquisition parameters on a sample in the same buffer and with similar height as that in 100% D2O. The measurement of the H/D exchange rates was repeated after adding 1 mM excess cAMP to both sample A and the preequilibration buffer. A 3 ´ 3 matrix centered at peak maxima was used for the quantification of the HSQC intensities with NMRPipe (2) and the spectral noise standard deviation was used to derive the uncertainties in the cross-peak intensities (4). The H/D exchange rates were obtained through the program Curvefit (12) by implementing a Levenberg-Marquardt nonlinear least-squares exponential fitting. Estimates of the uncertainties in the fitted decay rates were obtained both from the covariance matrix and Monte Carlo simulations (12). PFs were computed as described (13) assuming an EX2 Linderstrom-Lang mechanism because the pH is 6.5. The intrinsic exchange rates required for the computation of the PFs were computed by using the program SPHERE implemented with a rate basis from Alanine oligopeptides (14, 15). If a rate basis from poly-DL-Ala is used, the intrinsic exchange rates decrease by ~50%, resulting in a constant offset of ~0.3 in the logPF values reported in Fig. 2 (14, 15). The errors in the logPF values were determined by error propagation based on the experimental uncertainty for the measured H/D exchange rates.H/H Exchange.
A total of 32 scans were accumulated per serial file with an interscan delay of 2 s and a RF strength of 6.9 kHz for the CLEANEX-PM (16, 17) mixing block, which minimizes offset effects. Replicate spectra were acquired for the short mixing times (5 and 10 ms). Despite the coaddition of the duplicate CLEANEX-PM-HSQC data sets at short mixing times, for selected slow exchanging residues, detectable CLEANEX cross-peaks appeared only at longer mixing periods. Build-up curves were analyzed as described using an FHSQC spectrum to determine reference peak heights (16, 17). Protection factors from H/H exchange rate were computed as kintrinsic.H/H/kobservedH/H. The intrinsic exchange rates were computed as for the case of H/D exchange but using H2O as solvent rather than D2O in the implementation of the SPHERE computation (14, 15).Structural and Chemical Shift Analyses.
For all structure-based analyses the coordinates of the (residues 119-244) fragment of the RIa (residues 91-376) crystal structure (Protein Data Bank ID Code 1RGS) were used because the 1RGS-based computed Ca and Cb chemical shifts correlate very well with the corresponding values measured for RIa (residues 119-244) in solution (Fig. 5). The Ca and Cb chemical shifts were computed through the program ShiftX (18, 19) based on the (residues 119-244) fragment of the 1RGS structure (20). The correlation coefficients are 0.969 and 0.994 for the Ca and Cb chemical shift plots (Fig. 5), respectively, which are in the range expected based on the accuracy of the chemical shift prediction and on the resolution of the 1RGS structure (2.8 Å) (18, 19). Furthermore, the theoretical vs. experimental rmsds are 0.773 and 0.883 ppm for the Ca and Cb chemical shifts correlations, respectively, which are well within the error of the chemical shift computation (i.e., 0.98 and 1.10 ppm for Ca and Cb atoms, respectively) (18, 19).Hydrogen bonds and interresidue contacts were analyzed by using Molmol (21) and the 1RGS (119-244) structure (20) with computed hydrogen atoms. SASAs were computed by using the program Getarea (22) applied with a probe radius of 1.4 Å. The shielding effect of tightly bound water molecules and cAMP was not considered. The secondary structure elements were identified based on the hydrogen bonding patterns according to the Kabsch/Sander algorithm (23). The local rmsds between the cAMP-bound (20) and the C-bound structures of RIa (residues 119-240) (24) were computed for each residue after superimposing the backbone heavy atoms for the amino acid triplet formed by the selected and the two adjacent residues. Residues 240-244 were excluded from this analysis because of their intrinsic flexibility, which would bias the local rmsd computation. The pairwise rmsd for the backbone heavy atoms of all three residues was then calculated. Considering that the local rmsd is calculated between residue triplets, an error margin of ± one residue should be considered in the local rmsd interpretation. To eliminate local rmsd biases caused by intrinsic variability we repeated the local rmsd computation after replacing the cAMP-bound structure of the R-subunit solved at 2.8-Å resolution with the structure of the same R-construct but bound to the cAMP agonist Sp-cAMPS and solved at higher resolution (2.3 Å) (25). The Sp-cAMPS-bound analysis confirmed the pattern for the major C-induced conformational variations identified above using the cAMP-bound structure with the exception of the W188 site, which as suspected may be affected by intrinsic variability.
1. Hamuro Y, Anand GS, Kim JS, Juliano C, Stranz DD, Taylor SS, Woods VL, Jr (2004) J Mol Biol 340:1185-1196.
2. Delaglio F, Grzesiek S, Vuister GW, Zhu G, Pfeifer J, Bax A (1995) J Biomol NMR 6: 277-293.
3. Goddard TD, Kneller DG (2006) SPARKY 3 (University of California, San Francisco).
4. Farrow NA, Muhandiram R, Singer AU, Pascal SM, Kay CM, Gish G, Shoelson SE, Pawson T, Forman-Kay JD, Kay LE (1994) Biochemistry 33:5984-6003.
5. Cavanagh J, Fairbrother WJ, Palmer AG I, Skelton NJ (1996) Protein NMR: Principles and Practice (Academic, London).
6. Esposito V, Sjoberg T, Das R, Brown S, Taylor SS, Melacini G (2006) J. Biomol. NMR, in press.
7. Das, R, Abu-Abed M, Melacini G (2006) J Am Chem Soc 128:8406-8407.
8. Farrow NA, Zhang O, Forman-Kay JD, Kay LE (1994) J Biomol NMR 4:727-734.
9. Vialle-Printems C, van Heijenoort C, Guittet E (2000) J Magn Reson 142:276-279.
10. Rodriguez JC, Jennings PA, Melacini G (2004) J Biomol NMR 30:155-161.
11. Mulder FAA, Schipper D, Bott R, Boelens R (1999) J Mol Biol 292:111-123.
12. Palmer AGI (1998) Curvefit (Columbia University, New York).
13. Bai Y, Englander JJ, Mayne L, Milne JS, Englander SW (1995) Methods Enzymol 259:344-356.
14. Bai Y, Milne JS, Mayne L, Englander SW (1993) Proteins Struct Funct Genet 17:75-86.
15. Zhang YZ (1995) Ph.D. Thesis (University of Pennsylvania, Philadelphia, PA).
16. Hwang TL, van Zijl PCM, Mori S (1998) J Biomol NMR 11:221-226.
17. Huang H, Melacini G (2006) Anal Chim Acta 564:1-9.
18. Neal S, Nip AM, Zhang HY, Wishart DS (2003) J Biomol NMR 26:215-240.
19. Zhang HY, Neal S, Wishart DS (2003) J Biomol NMR 25:173-195.
20. Su Y, Dostmann WR, Herberg FW, Durick K, Xuong NH, Ten Eyck L, Taylor SS, Varughese KI (1995) Science 269:807-813.
21. Koradi R, Berman HM, Wüthrich K (1996) J Mol Graphics 14:51-55.
22. Fraczkiewicz R, Braun W (1998) J Computl Chem 19:319-333.
23. Kabsch W, Sander C (1983) Biopolymers 22:2577-2637.
24. Kim C, Xuong NH, Taylor SS (2005) Science 307:690-696.
25. Wu, J, Jones JM, Nguyen-Huu X, Ten Eyck LF, Taylor SS (2004) Biochemistry 43:6620-6629.