Activating an invertebrate bistable opsin with the all-trans 6.11 retinal analogue

1 Animal vision depends on opsins, a category of G protein-coupled receptor (GPCR) that achieves light 2 sensitivity by covalent attachment to retinal. Typically binding as an inverse agonist in the 11-cis form, 3 retinal photoisomerizes to the all-trans isomer and activates the receptor, initiating downstream 4 signaling cascades. Retinal bound to bistable opsins isomerizes back to the 11-cis state after 5 absorption of a second photon, inactivating the receptor. Bistable opsins are essential for invertebrate 6 vision and non-visual light perception across the animal kingdom. While crystal structures are available 7 for bistable opsins in the inactive state, it has proven difficult to form homogeneous populations of 8 activated bistable opsins either via illumination or reconstitution with all-trans retinal. Here we show 9 that a non-natural retinal analogue, all-trans retinal 6.11 (ATR6.11), can be reconstituted with the 10 invertebrate bistable opsin, Jumping Spider Rhodopsin-1 (JSR1). Biochemical activity assays 11 demonstrate that ATR6.11 functions as an agonist of JSR1. ATR6.11 binding also enables complex 12 formation between JSR1 and downstream signaling partners. Our findings demonstrate the utility of 13 retinal analogues for biophysical characterization of bistable opsins, which will deepen our 14 understanding of light perception in animals. 15 16 Main Text 17 18

19 20 Jumping Spider Rhodopsin-1 is a light-sensitive GPCR that the jumping spider requires for depth 21 perception (1). JSR1 achieves light sensitivity by covalent binding of 11-cis retinal chromophore to a 22 lysine side chain via a protonated Schiff base (PSB) (2). Photoisomerization of retinal to the all-trans 23 isomer, triggers conformational rearrangements in the receptor, resulting in adoption of an active 24 conformation capable of catalyzing nucleotide exchange in intracellular G proteins (2, 3). The receptor 25 thereby transduces an optical signal to initiate cellular signaling cascades. Unlike vertebrate visual 26 opsins, which are bleached and lose the retinal after photo-activation, all-trans retinal bound to 27 invertebrate opsins and non-visual vertebrate opsins reverts to the 11-cis state after a second photo-28 isomerisation event (4). They are therefore classified as ’bistable’ due to the thermal stability of the 29 PSB in both the inactive and active states. Bistable opsins are potential optogenetic switches to control 30 G protein signaling pathways, and in vivo studies have demonstrated the capability of JSR1 as an 31 optogenetic tool to manipulate neuronal signaling in animals (5). 32 33 High-resolution crystal structures of JSR1 and squid opsin have provided insights into the architecture 34 of the retinal binding site in bistable opsins (3, 6). These structures identify several structural 35 differences between monostable (bleachable) and bistable (non-bleachable) opsins, particularly in the 36 retinal binding site (3, 6). In particular, the two categories of receptor have evolved distinct counterion 37 systems that stabilize the positive charge of the PSB within the hydrophobic core of the receptor. 38 There are currently no active state structures of bistable opsins interacting with downstream signaling 39 partners and it is therefore unclear how these receptors change conformation after retinal 40 isomerization. Structural studies depend on obtaining a homogeneous population of the receptor in the 41 active state. In the case of JSR1, it has not been possible to reconstitute the receptor with the native 42 agonist, all-trans retinal. In addition, due to the overlapped absorption maxima (λmax) of the inactive 43 and active states, illumination of the receptor generates a mixture of states not easily amenable to 44 structural characterization. It is not uncommon for the λmax of bistable opsins to be similar in the active 45 and inactive states, as shown for JSR1, melanopsin and squid rhodopsin (2, 7, 8). The present study 46 focuses on the use of locked retinal analogues to activate JSR1 without light. 47 48

49 50 JSR1 was expressed recombinantly in the apo state, and addition of 9-cis retinal (Figure 1a) to the cell 51 lysate enabled reconstitution with the inverse agonist, as evidenced by the 505 nm absorption 52 maximum (Figure 1b). Adding the endogenous agonist, all-trans retinal, to cell lysate containing JSR1 53 failed to reconstitute the protein with the chromophore and to form the protein-chromophore covalent 54 bond (Figure 1b). We therefore sought for non-natural agonists of JSR1 that could activate the 55 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.08.588240doi: bioRxiv preprint 3 receptor. We attempted reconstitution of JSR1 with 9-cis retinal 6.11 (9CR6.11) and ATR6.11, which 56 have a six-membered ring cyclized around the 11-double bond of retinal (Figure 1a). JSR1 bound both 57 ATR6.11 and 9CR6.11, as demonstrated by the absorption spectrum of the protein after purification 58 (Figure 1b). The 513 nm and 509 nm λmax of JSR1 bound to ATR6.11 and 9CR6.11 respectively 59 suggests that they are covalently bound to JSR1 via a PSB (Figure 1b). 60 61 We performed an acid denaturation experiment to confirm that ATR6.11 and 9CR6.11 were covalently 62 bound to JSR1 via a protonated Schiff base. The λmax shifts from ≈510 nm to 445 nm (Figure 1c & 1d) 63 for both retinal isomers after addition of HCl, as expected if the protein unfolds and the PSB becomes 64 solvent exposed (8). The molar extinction coefficients of JSR1/ATR6.11 and JSR1/9CR6.11 were 65 determined experimentally using a hydroxylamine assay. Addition of hydroxylamine to the opsin 66 cleaves the PSB and releases the retinal oxime (λmax 365 nm). The extinction coefficient is calculated 67 using the extinction coefficients determined for ATR6.11 and 9CR6.11, and the ratio between the 68 decrease in absorbance at 509 nm and increase at 365 nm. We determined the extinction coefficient 69 of JSR1/ATR6.11 to be 49,500 M−1 cm−1 at 509 nm (Figure 1e), while JSR1 in complex with 9CR6.11, 70 has an extinction coefficient of 37,430 M−1cm−1 at 509 nm (Figure 1f). The higher extinction coefficient 71 of ATR6.11 compared to 9CR6.11 corresponds well to the higher extinction coefficient of JSR1 bound 72 to all-trans retinal compared to 9-cis, 37,560 M−1 cm−1 and 32,660 M−1 cm-1 respectively (2). 73 74 75 While the C11=C12 bond of the retinal analogues is constrained and cannot isomerize, other bonds may 76 still be susceptible to photo-isomerization. We measured UV-Vis spectra of JSR1 bound to the retinal 77 analogues before and after illumination at 519 nm for defined time periods. In the case of 78 JSR1/ATR6.11, we observe a marked increase in absorption within the first 10 seconds of illumination 79 (Figure 1g), with the greatest increase in absorption at 491 nm (Figure 1h). We also observe increased 80 absorption by JSR1/9CR6.11 within the first 10 seconds of illumination (Figure 1i), with the formation 81 of a photoproduct with a higher extinction coefficient at 488 nm (Figure 1j). As the 9-cis bond is liable 82 to isomerize, it is possible that an all-trans isomer is formed, although the formation of a di-cis species 83 cannot be excluded (9). While the molecular bases of these changes have not been resolved, it is 84 evident that both retinal 6.11 analogues remain light sensitive after binding to JSR1. 85 86 After confirming covalent binding of both retinal analogues, we sought to determine whether they 87 function as agonists. Based on phylogenetic analysis, JSR1 is expected to catalyze exchange of GDP 88 for GTP in the Gq heterotrimer when in the active state (10), although it also couples to human Gi in 89 vitro (3). A GTPase-Glo assay was used to measure the amount of GTP remaining after incubating 90 GTP with JSR1 and either Gq or Gi heterotrimer, and hence the ability of JSR1 to catalyze nucleotide 91 exchange in the G proteins. In the absence of JSR1, Gi demonstrates a markedly higher intrinsic rate 92 of nucleotide turnover than Gq, with almost 60% of nucleotides depleted compared to 15% for Gq 93 (Figure 2). Addition of JSR1/9-cis retinal causes no increase in the rate of nucleotide exchange in the 94 Gi or Gq heterotrimers (Figure 2a), as expected, since 9-cis retinal stabilizes JSR1 in an inactive 95 conformation (3). Steady state illumination of 9-cis retinal bound JSR1 generates a dynamic 96 equilibrium of JSR1 conformational states, due to the overlapped absorption profiles of the 9-cis, 11-97 cis and all-trans isomers, with up to 73% of the molecules adopting the active all-trans conformation 98 (2). In keeping with this, we observe an increase in GTP depletion by Gi from 57% in the presence of 99 JSR1/9-cis retinal to 92% after illumination (Figure 2a). Similarly, nucleotide depletion by Gq increases 100 from 7% to 44%, a clear marker of light-induced receptor activation (Figure 2a). 101 102 In contrast, JSR1/ATR6.11 catalyzes nucleotide exchange without illumination, increasing GTP 103 depletion in the presence of Gi from 58% to 80% and from 13% to 28% in the presence of Gq (Figure 104 2b). Illumination of JSR1/ATR6.11 causes a further increase in activity in Gi and Gq to 91% and 47% 105 respectively (Figure 2b). While the C11=C12 double bond of the ligand is constrained in a trans 106 configuration, absorption of light may induce structural changes in the chromophore that increase its 107 potency as an agonist of JSR1. In the dark state, JSR1/9CR6.11 demonstrates no catalytic effect on 108 nucleotide depletion by Gi or Gq (Figure 2c), similar to the effect of 9-cis retinal. Illumination of the 109 sample allows the receptor to catalyze nucleotide exchange in Gi and Gq to similar levels observed for 110 illuminated JSR1/9-cis retinal and JSR1/ATR6.11, with 91% depletion in the presence of Gi and 51% 111 with Gq (Figure 2c). The light induced increase in nucleotide depletion may be caused by 112 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.08.588240doi: bioRxiv preprint 4 isomerization around the C9=C10 double bond of 9CR6.11, yielding ATR6.11 chromophore. However, 113 light induced formation of di-cis retinal analogues has also been shown to activate monostable opsins 114 previously (9). 115 116 Finally, we sought to determine whether JSR1/ATR6.11 would form a stable complex with the human 117 Gi and Gq heterotrimers. Purified receptor and G protein was incubated together in the presence of 118 apyrase. The receptor and the G protein heterotrimers were subjected to size exclusion 119 chromatography (SEC) after incubation together. Both samples show a pronounced peak at an elution 120 volume of ~8.8 ml corresponding to the receptor alone, with a relatively high ratio of absorbance at 505 121 nm compared to the 280 nm (Figure 2d & 2e). An additional peak is visible at 7.8 ml at both 122 wavelengths when JSR1/ATR6.11 is incubated with the Gq heterotrimer (Figure 2e). The shorter 123 retention volume indicates formation of a higher molecular weight complex. Similarly, when 124 JSR1/ATR6.11 was incubated with the Gi heterotrimer, a species with a shorter retention volume (8.3 125 ml) formed with absorbance at both 280 nm and 505 nm (Figure 2d). SDS-PAGE analysis confirmed 126 formation of the full complex of JSR1/ATR6.11 with the Gi and Gq heterotrimers. This demonstrates 127 that ATR6.11 is a sufficiently potent agonist to induce complex formation with downstream signaling 128 partners. We note that significant populations of the receptor and G protein remain unbound to each 129 other in the Gi sample and this complex may therefore be less stable than the Gq complex (Figure 2d). 130 131

132 133 Locked retinal analogues have been used to investigate the photochemistry underlying signaling by 134 retinal proteins, both with microbial opsins (11, 12) and monostable vertebrate opsins (9, 13–15). We 135 show that their use can be expanded to biochemical and biophysical studies of bistable opsins. 136 Despite their physiological importance, our understanding of the mechanisms by which these proteins 137 achieve bistability remains limited, largely due to difficulties with recombinant expression and 138 successful reconstitution of the receptors with their native chromophores. We demonstrate that retinal 139 analogues function as tool molecules to inform our understanding of bistable opsins. In particular, we 140 show that reconstitution of ATR6.11 with JSR1 forms a population of activated receptor suitable for 141 structural and biophysical studies. The underlying reason for the selective binding of ATR6.11 over all-142 trans retinal remains unclear. ATR6.11 may simply bind with higher affinity, or the aldehyde group may 143 be better positioned to form a Schiff base bond with the binding site lysine residue. It should be noted 144 that most bistable opsins have not evolved to bind all-trans retinal and light-independent binding of a 145 natural agonist would contribute to dark noise, with light independent receptor activation. We envisage 146 that these compounds will be of use to the broader opsin community, as their ability to activate a given 147 protein can be easily tested using biochemical and cellular signaling assays. 148 149

150 151 Full materials and methods may be found in the Supplementary Information. 152 153 154 Acknowledgments 155 156 We would like to thank Mara Wieser, Dr. Elena Lesca and Dr. Niranjan Varma for their 157 contributions in JSR1 biochemistry. This project has received funding from the European 158 Research Council (ERC) under the European Union's Horizon 2020 research and innovation 159 programme (Grant agreement No. 951644 to GFXS, and Marie Skłodowska-Curie grant 160 agreement No. 701647 to MJR). M.S. thanks the Kimmelman Center for Biomolecular Structure 161 and Assembly for partial support. M.S. holds the Katzir-Makineni Chair in Chemistry. 162 163 164 165 166 167 168 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.08.588240doi: bioRxiv preprint 5

169 170 1. T. Nagata, et al., Depth perception from image defocus in a jumping spider. Science 335, 469–471 171 (2012). 172 2. D. Ehrenberg, et al., The two-photon reversible reaction of the bistable jumping spider rhodopsin-1. 173 Biophysical Journal 116, 1248–1258 (2019). 174 3. N. Varma, et al., Crystal structure of jumping spider rhodopsin-1 as a light sensitive GPCR. Proc. 175 Natl. Acad. Sci. U.S.A. 116, 14547–14556 (2019). 176 4. A. Terakita, E. Kawano-Yamashita, M. Koyanagi, Evolution and diversity of opsins. WIREs Membr 177 Transp Signal 1, 104–111 (2012). 178 5. H. Hagio, et al., Optogenetic manipulation of Gq- and Gi/o-coupled receptor signaling in neurons 179 and heart muscle cells. eLife 12, e83974 (2023). 180 6. M. Murakami, T. Kouyama, Crystal structure of squid rhodopsin. Nature 453, 363–367 (2008). 181 7. R. Hubbard, R. C. C. St. George, The rhodopsin system of the squid. Journal of General Physiology 182 41, 501–528 (1958). 183 8. T. Matsuyama, T. Yamashita, Y. Imamoto, Y. Shichida, Photochemical properties of mammalian 184 melanopsin. Biochemistry 51, 5454–5462 (2012). 185 9. S. Gulati, et al., Photocyclic behavior of rhodopsin induced by an atypical isomerization mechanism. 186 Proc. Natl. Acad. Sci. U.S.A. 114, E2608–E2615 (2017). 187 10. M. Koyanagi, T. Nagata, K. Katoh, S. Yamashita, F. Tokunaga, Molecular evolution of arthropod color 188 vision deduced from multiple opsin genes of jumping spiders. Journal of Molecular Evolution 66, 189 130–137 (2008). 190 11. A. Albeck, N. Friedman, M. Sheves, M. Ottolenghi, Role of retinal isomerizations and rotations in the 191 photocycle of bacteriorhodopsin. J. Am. Chem. Soc. 108, 4614–4618 (1986). 192 12. T. L. Brack, et al., Picosecond time-resolved absorption and fluorescence dynamics in the artificial 193 bacteriorhodopsin pigment BR6.11. Biophys J 65, 964–972 (1993). 194 13. H. Akita, S. P. Tanis, M. Adams, V. Balogh-Nair, K. Nakanishi, Nonbleachable rhodopsins retaining 195 the full natural chromophore. J. Am. Chem. Soc. 102, 6370–6372 (1980). 196 14. W. J. de Grip, et al., 10,20-Methanorhodopsins: (7E,9E,13E)-10,20-methanorhodopsin and 197 (7E,9Z,13Z)-10,20-methanorhodopsin. 11-cis-locked rhodopsin analog pigments with unusual 198 thermal and photo-stability. Eur J Biochem 191, 211–220 (1990). 199 15. S. Bhattacharya, K. D. Ridge, B. E. Knox, H. G. Khorana, Light-stable rhodopsin. I. A rhodopsin analog 200 reconstituted with a nonisomerizable 11-cis retinal derivative. J Biol Chem 267, 6763–6769 (1992). 201 202 203 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.08.588240doi: bioRxiv preprint 6 Figures 204 205 206 207 Figure 1. (a) Chemical structures of all-trans retinal, 9-cis retinal, all-trans retinal 6.11 and 9-cis 208 retinal 6.11. (b) UV-Vis spectra of purified JSR1 after reconstitution with 9-cis retinal, all-trans 209 retinal, 9-cis retinal 6.11 and all-trans retinal 6.11. (c) UV-Vis spectra of JSR1/ATR6.11 before 210 and after acid denaturation. (d) UV-Vis spectra of JSR1/9CR6.11 before and after acid 211 denaturation. (e) UV-Vis spectra of JSR1/ATR6.11 after addition of hydroxylamine. (f) UV-Vis 212 spectra of JSR1/9CR6.11 after addition of hydroxylamine. (g, h) UV-Vis spectra and difference 213 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.08.588240doi: bioRxiv preprint 7 spectra of JSR1/ATR6.11 after repeated illumination at 519 nm. (i, j) UV-Vis spectra and 214 difference spectra of JSR1/9CR6.11 after repeated illumination at 519 nm. 215 216 217 218 219 220 Figure 2. GTP depletion by the human Gi and Gq heterotrimers in the presence and absence of 221 JSR1 bound to (a) 9-cis retinal, (b) all-trans retinal 6.11, (c) 9-cis retinal 6.11, with and without 222 illumination. Size exclusion chromatography (SEC) profiles for dark state JSR1/ATR6.11 223 incubated with the (d) human Gi heterotrimer and (e) human Gq heterotrimer. 224 225 226 227 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted April 9, 2024. ; https://doi.org/10.1101/2024.04.08.588240doi: bioRxiv preprint