Supplementary Figure 1

Monitoring the thermal stability of WT Gαi1 by differential scanning fluorimetry (DSF) upon addition of increasing concentrations of nucleotides.

a, b, Melting curves of WT Gαi1 in the presence of GDP (a) and stabilization effect concentration-response curve (b). Stability increase [(Tm (w/ GDP) - Tm (wo/ GDP)] is plotted against the concentration of GDP. Tm of WT Gαi1 in addition of 1mM GDP reflects the thermal stability of WT Gαi1 in the GDP-bound state [Gαi1(WT)-GDP]. c, d, Melting profiles of WT Gαi1 in presence of GTPγS (c) and stability increase concentration-response curve (d). Tm of WT Gαi1 in addition of 100 µM GTPγS was finally chosen to reflect the thermal stability of WT Gαi1 in the GTPγS-bound state. The thermal stability measurement of all Gαi1 alanine mutants in the nucleotide-bound state was performed upon addition of 1mM GDP or 100 µM GTPγS. e, Thermal profile of F336A

G.H5.8 in the nucleotide-bond state. The melting curve shows that alanine

replacement of F336 G.H5.8

in Gαi1 completely impairs the protein stability and activity of coupling with nucleotides. The details of DSF experiments are described in the methods section.

Nature Structural & Molecular Biology: doi:10.1038/nsmb.3070

Supplementary Figure 2

High-throughput (HTP) assay for monitoring the effect of Gαi1 alanine mutants on the Rho*–Gi complex.

The recombinant Gαi1 alanine mutants were prepared by HTP purification as shown in the left panel, and the endogenous rhodopsin and βγ subunit were prepared from bovine retinas. For the formation of Rho*-Gi complex, Gαi1 and βγ subunit were reconstituted to form Gi heterotrimer, followed by mixing with rhodopsin and light activation. The formed Rho*-Gi

complex was visualized by native gel electrophoresis and the gel bands of complex were quantified by ImageJ software. In each round, WT Gαi1 was always included as the reference control. The experimental details of HTP assay are described in methods section.

Nature Structural & Molecular Biology: doi:10.1038/nsmb.3070

Supplementary Figure 3

Characterization of Gαi1 alanine mutants by native gel electrophoresis.

a, Visualization of Rho*-Gi (WT) complex. The Rho*-Gi (WT) complex (+/- GTPγS) was visualized by the native gel electrophoresis, as well as with WT Gαi1, βγ subunit, reconstituted Gi heterotrimer, and the activated rhodopsin as reference markers. The data clearly show that the Rho*-Gi (WT) complex is stable in the absence of nucleotides and that the addition of GTPγS dissociates the complex. b, Gαi1 alanine mutants impairing the formation of Rho*-Gi complex. The Rho*-Gi complexes were formed with WT, Y320A

G.S6.2, L348A

G.H5.20, G352A

G.H5.24 and L353A

G.H5.25 of Gαi1 at 4 ºC and

heated at 36.3 ºC, followed by the native gel electrophoresis. The results clearly show that these four Gαi1 alanine mutants are severely impaired in their ability to form the Rho*-Gi complex. c, Monitoring the thermal dissociation (Td) of Rho*-Gi complex. The Rho*-Gi complex formed with WT Gαi1 was heated at the indicated temperature and visualized by the native gel electrophoresis. The gel bands of Rho*-Gi complex were integrated and quantified by ImageJ software. The determined Td50 value of Rho*-Gi(WT) is 36.0 ± 0.1 ºC. The same experiment was also performed with Rho*-Gi(R32A

G.hns1.3), Rho*-Gi(I55A

G.H1.10) and Rho*-Gi(N331A

G.H5.3) and the determined Td50 values are 36.7 ± 0.1 ºC, 37.4 ±

0.1 ºC and 37.8 ± 0.1 ºC, respectively, indicating these three alanine mutants stabilize the receptor-bound state. Data points represent mean ± s.d. from three individual experiments.

Nature Structural & Molecular Biology: doi:10.1038/nsmb.3070

Supplementary Figure 4

Characterization of Gαi1 alanine mutants in stabilizing the Rho*–Gi complex.

a, Visualization of Rho*-Gi complex by the native gel electrophoresis. The Rho*-Gi complexes were formed with WT, R32A

G.hns1.3, K51A

G.H1.6, I56A

G.H1.11, K54A

G.H1.9, I55A

G.H1.10, H57A

G.H1.12, R176A

H.HF.6, N331A

G.H5.3 and V332A

G.H5.4 of Gαi1 at

4 ºC and heated at 36.3 ºC, followed by the native gel electrophoresis as described in methods section. b, Gαi1 alanine mutants stabilize the Rho*-Gi complex. The gel bands of Rho*-Gi complexes were integrated and quantified using the ImageJ software. The complex formation efficiency (%) at 4 ºC and complex stability (%) at 36.3 ºC were defined and determined as described in methods section. The results show that these Gαi1 alanine mutants obviously enhance the thermal stability of Rho*-Gi complex. c, Tm of Gαi1 alanine mutants which stabilize the receptor-bound state. Tm was measured upon addition of 1mM GDP or 0.1mM GTPγS using the DSF assay as described in methods section. Data points represent mean ± s.d. from three individual experiments.

Nature Structural & Molecular Biology: doi:10.1038/nsmb.3070

Supplementary Figure 5

Effect of alanine substitution of residues involved in the activation and stabilization clusters on formation of the Rho*–Gi

complex.

a-d, Effect on Rho*-Gi complex formation of mutation of residues involved in the activation cluster I (a) and stabilization cluster II (b) of the GTPase domain, stabilization cluster III of helical domain (c) and the inter-domain interface (d). The increases in Δcomplex formation efficiency are colored in red and the reductions are colored in blue. The definition of Δcomplex formation efficiency is described in methods section, and the derived numbers are shown in Supplementary Table1.

Nature Structural & Molecular Biology: doi:10.1038/nsmb.3070

Supplementary Figure 6

Characterization of heterotrimer (Gi) formation by analytical size-exclusion chromatography (FSEC).

a-f, Characterization of heterotrimer reconstitution of last 48 Gαi1 alanine mutants which are inefficient in formation of Rho*-Gi complex. The reconstitution of Gi and analysis by FSEC were described in methods section. The retention time of WT Gαi1, βγ subunit and reconstituted Gi are 11.15 min, 11.45 min and 10.26 min, respectively. a, b, Retention time of alanine mutants which are efficient in Gi reconstitution. c, d, Retention time of inefficient alanine mutants in formation of heterotrimer. e, f, Retention time of three alanine mutants forming oligomers upon reconstitution.

Nature Structural & Molecular Biology: doi:10.1038/nsmb.3070

Supplementary Figure 7

Structural overlay of Gαi1–GDP with G protein α subunit from the plant Arabidopsis thaliana (AtGPA1).

The crystal structure of Gαi1-GDP (PDB 1GDD11

) was mapped with the measured ΔTm (in addition of GDP) as spectrum ranging from blue over white to red, as on Fig 1. The structure of AtGPA1-GTPγS (PDB 2XTZ

32) is shown in light pink.

The enlarged opening of the inter-domain interface suggests that the inter-domain interactions in AtGPA1 are much weaker than those in the Gαi1.

Nature Structural & Molecular Biology: doi:10.1038/nsmb.3070

Supplementary Figure 8

Mutations that affect the GTP-bound state cluster around the γ-phosphate of the GTPγS.

a, Correlation of destabilization upon mutation between GDP- and GTP-bound states. The red line shows the linear correlation, blue lines indicate the 95% confidence interval. The most destabilizing residue is N269A

G.S5.7, involved in

nucleotide binding. b, Alanine substitutions that affect GTP bound state locate near the gamma phosphate of GTPγS. ΔTm values for each single alanine mutant are mapped onto the crystal structure of GTPγS-bound Gαi1 (PDB 1GIA, ref. 11), as a spectrum ranging from blue over white to red. Alanine substitutions that specifically affect the GTP-bound state are deviated from a and the correlated residues are shown as spheres. In addition, alanine mutations in helix α2 also specifically affect the GTP-bound state, consistent with the conformational changes leading to the dissociation of the Gβγ subunit.

Nature Structural & Molecular Biology: doi:10.1038/nsmb.3070

SupplementarynoteProbingGi1ProteinActivationatSingleAminoAcidResolution

Dawei Sun1,2, Tilman Flock3,7, Xavier Deupi1,4,7, Shoji Maeda1,6, Milos Matkovic1,2, Sandro

Mendieta1, Daniel Mayer1,2, Roger Dawson5, Gebhard F.X. Schertler1,2, M. Madan Babu3 and

Dmitry B. Veprintsev1,2

The human Gai construct used in our study contained N terminal MKK sequence for

increase expression, 10xHis-tag followed by TEV proteiase cleavage site. The protein

sequence of the construct used in our experiments was:

MKKHHHHHHHHHHENLYFQGGSMGCTLSAEDKAAVERSKMIDRNLREDGE

KAAREVKLLLLGAGESGKSTIVKQMKIIHEAGYSEEECKQYKAVVYSNTI

QSIIAIIRAMGRLKIDFGDSARADDARQLFVLAGAAEEGFMTAELAGVIK

RLWKDSGVQACFNRSREYQLNDSAAYYLNDLDRIAQPNYIPTQQDVLRTR

VKTTGIVETHFTFKDLHFKMFDVGGQRSERKKWIHCFEGVTAIIFCVALS

DYDLVLAEDEEMNRMHESMKLFDSICNNKWFTDTSIILFLNKKDLFEEKI

KKSPLTICYPEYAGSNTYEEAAAYIQCQFEDLNKRKDTKEIYTHFTCATD

TKNVQFVFDAVTDVIIKNNLKDCGLF*

The amino acid numbers refer to the positions in the WT protein (starting MGC…).

Usabilityofstabilizingmutationsforstructuralstudies

We identified twelve mutations that stabilize the Rho*-Gi complex (R32(G).HNS1.3, K51(G).H1.6,

K54(G).H1.9, I55(G).H1.10, I56(G).H1.11, H57(G).H1.12, R176(H).HF.6, E245(G).H3.4, T327(G).S6H5.4 ,

N331(G).H5.3, V332(G).H5.4, D350(G).H5.22). These mutations are involved in the activation

process, stabilization or destabilization of the inter-domain interface, or interactions with the

receptor or the βγ subunit, and have been discussed in detail earlier. We also identified eight

mutations (G42(G).S1H1.3, A59(G).H1HA.2, T177(G).HFS2.1, D200(G).S3.7, A226(G).S4.7, E297(G).H4.2 ,

A300(G).H4.5, F334(G).H5.6) that stabilized the GDP-bound state, and four (V50(G).H1.5,

A59(G).H1HA.2, R178(G).HFS2.2, K180(G).HFS2.4)) that stabilized the GTP-bound state (Table I).

The most stabilizing mutation in the in the GDP-bound state was D200(G).S3.7A, which

increased stability by 5 °C. The stability of this mutant was almost identical in GDP- and

Nature Structural & Molecular Biology: doi:10.1038/nsmb.3070

GTPγS-bound states, in contrast to the wild type protein, which was more stable in the

GTPγS-bound state. D200(G).S3.7 is located at the C-terminal part of the strand β3, which is

the starting point of switch III, a region that is disordered in the GDP-bound state but is

folded in the GTP-bound state 11. It is possible that D200(G).S3.7A stabilizes this region in a

state similar to that observed for the GTP bound state.

The next mutation with the largest effect, A59G.h1ha.2G, stabilized both the GDT and GTP

bound states. This residue is located in loop L1 connecting the GTPase and the helical

domains, so a Gly at this position may be thermodynamically preferable due to the increased

flexibility of this segment. The mechanism of stabilization by the G42A mutation is not

obvious, and it may be due to a better local packing. The rest of the mutations that

moderately stabilized one or the other state are located in proximity to the nucleotide binding

pocket and probably provide better packing, or in case of the T177, R178 and K180, are

directly involved in the nucleotide interactions.

Due to the high sequence and structural conservation among G proteins, these mutations

are very likely to be transferrable to other G protein subtypes. These mutations can be used

to facilitate structural and biophysical studies of receptor – G protein complexes, as well as

of the Gα complexes with other interacting proteins.

Nature Structural & Molecular Biology: doi:10.1038/nsmb.3070