Accurate Characterisation of Weak Protein- Protein Interactions by Titration of NMR Residual Dipolar Couplings

Transcription

1 Accurate Characterisation of Weak Protein- Protein Interactions by Titration of NMR Residual Dipolar Couplings Jose Luis Ortega-Roldan, Malene Ringkjøbing Jensen*, Bernhard Brutscher, Ana I. Azuaga, Martin Blackledge* and Nico A.J. van Nuland Supporting Information S1

2 METHODS Protein expression and purification For clarity, we use the residue numbering of CD2AP SH3-C of PDB entry 2JTE throughout the text. Unlabelled, 15 N-labelled and 15 N, 13 C-labelled CD2AP SH3-C was obtained as described. 1 Unlabelled and 15 N-labelled ubiquitin was purchased from both Cortecnet and Spectra Stable Isotopes. 15 N/ 13 C-labelled ubiquitin was kindly provided by Varian Inc. Protein concentrations were determined by absorption measurements at 280 nm using 1 an extinction coefficient of and 1450 cm 1 M for SH3-C and ubiquitin, respectively, determined using the ProtParam algorithm ( NMR chemical shift perturbation All NMR titration experiments were performed at 25ºC on a Varian NMR Direct- Drive Systems 600 MHz spectrometer ( 1 H frequency of MHz) equipped with a triple-resonance PFG-XYZ probe. CD2AP SH3-C and ubiquitin samples were prepared for NMR experiments in 93% H 2 O/7% D 2 O, 50 mm NaPi, 1 mm DTT at ph 6.0. The backbone amide and 15 N frequencies of CD2AP SH3-C under the above conditions, previously assigned at ph were obtained first by comparing 2D 15 N- HSQC spectra at ph 2.0, 3.0, 6.0 and 7.0 and confirmed by a single HNCACB triple resonance experiment acquired on 15 N, 13 C-labelled CD2AP SH3-C at ph 6.0. A HNCACB triple resonance spectrum was also recorded on a 13 C/ 15 N-labelled ubiquitin to confirm backbone assignment at ph 6.0. The SH3-binding site on ubiquitin was obtained by titrating with increasing amounts of unlabeled CD2AP SH3-C domain into a 0.25 mm 15 N-ubiquitin sample at ph 6.0, 25 ºC. Similarly, the ubiquitin-binding site on CD2AP SH3-C was obtained by titrating with increasing amounts of unlabeled ubiquitin into a 0.25 mm 15 N-SH3-C sample under the same conditions. The progress of the titrations was monitored by recording one-dimensional 1 H and two-dimensional 1 H- 15 N HSQC spectra. The magnitude of the chemical shift deviations (Δδ) were calculated using the equation: "# ppm = "# HN $ ' ) % 6.51( ( ) 2 + "# N & 2 S2

3 where the difference in chemical shift is that between the bound and free forms of the different proteins. All NMR data were processed using NMRPipe 2 and analyzed by NMRView. 3 NMR Spectroscopy Assignment of SH3-C in the complex was confirmed by a single HNCACB triple resonance experiment acquired on 15 N, 13 C-labelled CD2AP SH3-C in complex with unlabelled ubiquitin (molar ratio SH3-C:Ubiquitin of 1:3) at ph 6.0. In order to obtain atom-specific ambiguous interaction restraints (AIRs, see below), we recorded the following experiments on both free SH3-C as well as in complex with ubiquitin: HNCO, HNCACB, HBHA(CO)NH and aromatic (HB)CB(CGCD)HD and (HB)CB(CGCDCE)HE. These experiments thus provided us with chemical shift perturbations upon ubiquitin binding for all 1 H, 15 N, 1 Hα, 13 Cα, 1 Hβ, 13 Cβ, 13 C atoms and the 1 Hδ and 1 Hε atoms of aromatic and Asn and Gln residues. Measurement of Residual Dipolar Couplings RDCs were measured in samples partially aligned in a liquid-crystalline medium consisting of a mixture of 5% penta-ethyleneglycol monododecyl ether (C 12 E 5 ) and hexanol. 4 For the 13 C/ 15 N-labelled protein (SH3) in the free form or in diverse mixtures of free and complex, a set of 4 different RDCs ( 1 D NH, 1 D CaC, 2 D HN C and 1 DCaHa ) was measured per sample using 3D BEST-type HNCO or HNCOCA experiments. 5, 6 Coupling constants were obtained from line splittings in the 13 C dimension using the nmrpipe nlinls fitting routine or using Sparky. 7 For the 15 N- labelled protein (ubiquitin) in free or in the complex, 1 D NH were measured from a pair of spin-state-selected 1 H- 15 N correlation spectra recorded using the pulse sequence shown in figure S4. The pulse sequence uses a DIPSAP filter 8 for J-mismatch compensated spin-state selection, and the BEST concept for longitudinal relaxation and sensitivity enhancement. In addition, signals from the doubly ( 13 C/ 15 N) labeled binding partner are removed by additional transfer steps from 15 N to 13 CO. After this transfer step the presence of orthogonal coherences for the spin systems from 15 N- only and 13 C/ 15 N labeled proteins is exploited to suppress the unwanted signals by S3

4 means of pulses field gradients and phase cycling. Total measurement time for one titration point is approximately 1 day. Sweep widths and data points used in different experiments: 1H sw 13Csw 15Nsw complex points(1hx13cx15n) nt 3D HNCO_jnh: x120x D HNCOCA_jcaha x120x D HNCO_jcoh x120x D HNCO_jcc x320x CfiltDIPSAP x Structural model of the SH3-C domain of CD2AP in complex with ubiquitin using HADDOCK Docking of the CD2AP SH3-C/Ubiquitin complex was performed using the software HADDOCK2.0 9 in combination with CNS 10 based on the chemical shift perturbation data obtained in this work. The starting structures for the docking were the 10 lowestenergy RDC-refined NMR structures of CD2AP SH3-C obtained in this work and the X-ray structure of ubiquitin (PDB entry 1UBQ). 11 Active and passive residues defined for HADDOCK were chosen based on the chemical shift perturbation data and solvent accessibility. We first selected all the residues in ubiquitin having a combined backbone 1 H- 15 N chemical perturbation upon complex formation higher than the mean chemical shift perturbation (0.05 ppm). We then selected 1 H, 15 N, 1 Hα, 13 Cα, 1 Hβ, 13 Cβ, 13 C atoms and the 1 Hδ and 1 Hε atoms of aromatic and Asn and Gln residues in SH3-C showing chemical shift differences between the free and bound form bigger than the mean plus one standard deviation. We calculated the solvent accessibility using the program NACCESS, 12 which is based upon the Lee & Richards algorithm 13, over the ensemble of SH3-C structures and the single ubiquitin X-Ray structure and selected as active residues all the amino acids showing an average relative solvent accessibility for the backbone and/or side-chains higher than 40%. We then selected all surface neighbors amino acids having a high solvent accessibility (>40%) as passive residues. These selection criteria resulted in active residues 6, 8, 42, 44, 46, 47, 48, 68, 71, 72, 73, 75 and 76 and passive residue 70 and 74 in ubiquitin. A 2 Å distance was used to define the ambiguous interaction restraints (AIR) for ubiquitin. Atom-specific AIRs for SH3-C involving active residues 15, 18, S4

5 19, 21, 36-38, 40, 42, 43 and are collected in Table S1 using a 6 Å distance for all protons and a 7 Å distance for 15 N and 13 C atoms. Residues 20 and 39 were included as passive residues. Semi-flexible regions for the two proteins were defined as residues 12-24, and for SH3-C and residues 6-10, and for ubiquitin. During the rigid body energy minimization, 1000 structures were calculated. The 200 best solutions based on the intermolecular energy structure were used for the semi-flexible simulated annealing followed by refinement in explicit water. Finally, the solutions were clustered using a 3 Å rmsd after superposition on the backbone of the residues in the above defined semi-flexible regions. From each of the 10 resulting clusters, one structure was then used as input for the combined chemical shift perturbation - RDC protocol. RDC Refinement of the SH3-C domain of CD2AP in complex with ubiquitin The program SCULPTOR 14 has recently been developed as an addition to the program CNS. The refinement protocol involves a restrained MD calculation using the standard CNS force field. Starting structures were taken from a selection of 10 structures determined using HADDOCK 9 based on ambiguous intermolecular restraints (AIRs). Initial sampling was further increased by including a sampling period of 10ps at 700K that allows for the SH3 domain to reorient freely without RDCs or AIR restraints. Following this both ubiquitin and SH3 conformations are fixed and the alignment tensor is allowed to evolve freely to determine initial estimates of the alignment tensor on the SH3 structure (there are four RDC types available for this structure). Both molecules and tensors are then freed with the initial sampling period at 1000K for 5ps during which time both AIRS and RDCs are scaled from 0.1% to 100% of their final values. The backbone conformation of the segment (1-70) of ubiquitin is restrained to its initial coordinates using a harmonic potential, while the experimentally measured noes from the free form of SH3 are used as restraints. AIR restraints are used as described in the supporting information. Sampling of restraints is followed by a 5ps sampling stage and slow cooling over 5ps to 100K and energy minimization. The protocol is repeated 30 times for each starting structure. S5

6 Table S1: Atom-specific ambiguous interaction restraints for SH3-C in complex with ubiquitin. Residue 15 N 1 HN 13 Cα 1 Hα 13 CO 13 Cβ 1 Hβ 1 Hδ 1 Hε in SH3 Tyr15 X X X X X X Thr18 X X Asn19 X X X X Asp21 X X X Lys36 X X Glu37 X X Thr38 X X Glu40 X X Gly42 X Trp43 X X Pro56 X X Asp57 X X Asn58 X Phe59 X X X X X S6

7 Table S2. Alignment tensors of different mixtures between two different simulated tensors. p 0.0 a A a (10-4 ) A r (10-4 ) α ( ) β ( ) γ ( ) a : RDC data from the two limiting tensors (p=0.0 and 1.0) were simulated and co-added to mimic the case of free and bound form mixtures. The data were then fitted to the same structure, and the effective alignment tensors are shown here. The fit is perfect in each case. S7

8 Table S3. Structural statistics for the 10 models of the CD2AP SH3-C:Ubiquitin complex 1 Restraints statistics NOE violations > 0.5 Å 26 ± 9 NOE violations > 0.3 Å 42 ± 14 NOE violations > 0.1 Å 66 ± 23 AIR violations > 0.2 Å 0.3 ± 0.3 φ/ψ restraint violations > ± 0.20 Reduced Chi2 (χ 2 /N) from experimental residual dipolar coupling restraints (Q-factor given in square parentheses): 2) SH3 C α -H α (42) 1.5 ± 0.2 [0.10] N-H N (56) 0.8 ± 0.1 [0.09] C -H N (55) 1.1 ± 0.1 [0.13] C α -C (53) 1.2 ± 0.1 [0.12] Ubiquitin N-H N (56) 0.9 ± 0.1 [0.09] Alignment tensor eigenvalues : Deviation from the idealized covalent geometry A a A r 12.9± ± CNS Interaction Energies (Kcal/mol) Bonds [Å] Angles [ ] Impropers [ ] ± ± ± 0.02 Coordinate precision (Å) 3 E elec E vdw -445 ± 104 kcal.mol ± 18 kcal.mol -1 Backbone N, C α, C 1.0±0.1 S8

9 Ramachandran plot 4 Most favored regions (%) 87.0 Additional allowed regions (%) 10.4 Generously allowed regions (%) 2.6 Disallowed regions (%) The statistics is obtained from an ensemble of ten models for the SH3-C:Ubiquitin complex with the lowest target function. 2 Restraint statistics reported for the refined complex. 3 Coordinate precision is given as the pair-wise Cartesian coordinate Root Mean Square Deviations from the average structure of the entire complex over the ensemble (SH3 and Ubiquitin). 4 Values obtained from the PROCHECK analysis over all residues. S9

10 Supplementary Figure Legends Figure S1. Representative cartoon presentation of starting structures from HADDOCK. Stereo representation of 5 out of the 10 HADDOCK structures that were used as starting structures in the combined CS-RDC protocol. Structures were superimposed on the backbone atoms of residues 4-73 of the ubiquitin molecule. For clarity only one ubiquitin molecule is presented in blue, while the 5 CD2AP SH3-C molecules are in different colours. S10

11 Figure S2. Fits of experimental RDC data sets to the 1d3z conformation of ubiquitin. N-H N RDCs are reproduced by the structure for all four data sets using the best fitting alignment tensors (table 1). (A) free form, (B) mixture m 1, (C) mixture m 2, (D) mixture m 3. S11

12 Figure S3. Fits of experimental RDC data sets to the RDC refined free SH3 conformation. (A, C, E) Comparison of N-H N RDCs to mixtures m 1, m 2 and m 3 respectively for the best fitting alignment tensors (table 1). (B, D, F) Comparison of C α H α RDCs to mixtures m 1, m 2 and m 3 respectively for the best fitting alignment tensors (table 1). While inferior agreement of experimental data with the SH3 freeform structure may indicate that more structural rearrangements are taking place in this protein upon formation of the complex, it may also reflect the lower effective resolution of the free-form structure than is available for ubiquitin. S12

13 Figure S4. BEST-type pulse sequence for the measurement of ( 1 J NH + 1 D NH ) coupling constants in a mixture of 15 N- and 13 C/ 15 N-labelled proteins. Only signals from the 15 N-labelled protein are detected in this experiment, while all signals from the doubly labeled protein are suppressed by a 15 N- 13 CO transfer filter. Filled and open pulse symbols indicate 90 and 180 rf pulses. Unless indicated, all pulses are applied with phase x. All selective 1 H pulses are centered at 8.2 ppm, covering a bandwidth of 4.0 ppm, with the following shapes: EBURP2 6 for excitation, REBURP 6 for refocusing, and time-reversed EBURP2 6 for flip back purposes. The additional H pulses applied during the DIPSAP filter (open squares) have the shape of BIP pulses 7. CO pulses have the shape of the center lobe of a sinx/x function. The transfer delays are set to: τ 1 =1/4J HN 2.4 ms-δ/2, and T= 1/8J NCO 8.5 ms, with δ the length of the REBURP 180 pulse. Pulsed field gradients, G 1 G 8 are applied along the z-axis (PFG z ) with durations of 200 µs to 2 ms and field strengths ranging from 5 40 G/cm. A 2-step phase cycle is applied by simultaneously changing the sign of ϕ 1 and the receiver phase ϕ rec. Quadrature detection in t 1 is obtained by time-proportional phase incrementation of ϕ 2 according to TPPI-States. For DIPSAP spin-state selection 8 3 repetitions of the experiment are performed with the following settings: (A) ε=0, ϕ 1 = x, (B) ε =1/8J HN 1.3 ms, ϕ 1 = y, and (C) ε =1/4J HN 2.6 ms, ϕ 1 = x. The spectra containing the upfield and downfield components of the 15 N doublets are then calculated from the 3 recorded data sets: 0.73 A 0.27 C ± B. S13

14 Figure S5. Correlation plot of 1 H- 15 N scalar couplings in free and complexed form. A. Correlation between scalar couplings measured in free SH3 versus scalar couplings measured in mixture m 1. To create two of the mixtures (m 2 and m 3 ), ubiquitin was added to the aligned sample, so that scalar couplings were not remeasured. For mixture m 1 scalar couplings were measured before adding the alignment medium. The high correlation between measured scalar couplings in the free and free/complex mixture m 1 justifies the approximation used here. S14

15 Figure S6. The ability of the RDC titration procedure to determine free data sets, compared to a simple extrapolation from a single mixture with the relative scaling determined from the D 2 O splitting measured in the free and bound samples. All 15 N- 1 H N RDCs were removed from mixture m 3 and : (A) predicted on the basis of values measured at m 0 and m 1 using a standard m extrapolation ( D 3 m ij = (" 1 p 3 D 1 ij # " 0,sh 3 (p 3 # p 1 )D free ij ) /" 3 p 1 ). The populations p are assumed to be accurately known in both cases and the scaling terms λ represent the measured D 2 O splitting (table 2). (B) the RDC titration analysis described in the article was repeated and experimental and 15 N- 1 H N predicted values from mixture m 3 were compared. S15

16 S16

17 1 Ortega Roldan, J.L.; Ab, E.; Romero Romero, M.L.; Ora, A.; Lopez Mayorga, O.; Azuaga, A.I.; van Nuland, N.A.J. J. Biomol. NMR , Delaglio, F.; Grzesiek, S.; Vuister, G.W.; Zhu, G.; Pfeifer, J.; Bax, A J. Biomol. NMR , Johnson, B.A.; Blevins, R.A. J. Biomol. NMR , Ruckert, M.; Otting, G. J.Am.Chem.Soc , Schanda, P.; Van Melckebeke, H.; Brutscher, B. J. Am. Chem. Soc , Lescop, E.; Schanda, P.; Brutscher, B. J. Magn. Reson , Goddard, T.D.; Kneller, D.G University of California 8 Brutscher, B. J. Magn. Reson , Dominguez C, Boelens R, Bonvin AM (2003) J. Am. Chem. Soc. 125: Brunger AT, Adams PD, Clore GM, DeLano WL, Gros P, Grosse-Kunstleve RW, Jiang J, Kuszewski J, Nilges M, Pannu NS, Read RJ, Rice LM, Simonson T, Warren GL (1998) Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Cryst D54: Vijay-Kumar S, Bugg CE, Cook WJ Structure of ubiquitin refined at 1.8 A resolution. J. Mol. Biol. 194, (1987) 12 Hubbard SJ, Thornton JM NACCESS. Department of Biochemistry and Molecular Biology, University College, London (1993) 13 Lee B, Richards FM The interpretation of protein structures: estimation of static accessibility J. Mol. Biol. 55, (1971) 14 Sibille, N.; Pardi, A.; Simorre, J.P.; Blackledge, M. J.Am.Chem.Soc , S17