Our previous studies revealed that not only mammalian Opn5 but also non-mammalian Opn5m forms a UV-sensitive bistable opsin [8, 9]. Moreover, in these studies, X. tropicalis Opn5m showed the highest expression yield in cultured cells among various Opn5m recombinant proteins that we analyzed. Wild-type X. tropicalis Opn5m directly binds all-trans retinal to produce a visible light-absorbing form (Fig. 2A). Yellow light irradiation induces the formation of a UV light-absorbing form, and subsequent UV and yellow light irradiations result in inter-conversion between the UV light- and visible light-absorbing forms. These spectral changes are triggered by the photoisomerization of the retinal between 11-cis and all-trans forms [8], which enables calculation of the absorption spectra of the 11-cis retinal and all-trans retinal bound forms of Opn5m (Fig. 2B). In this study, we performed mutational analysis of X. tropicalis Opn5m to search for the amino acid residue(s) responsible for the UV light sensitivity. The spectral tuning of UV-sensitive opsins has been extensively analyzed in vertebrate cone visual pigments and insect visual opsins. Previous studies revealed the importance of the amino acid residues in the extracellular side of Helix II for the spectral tuning of UV-sensitive opsins [22,23,24,25,26,27]. S90C mutation of chicken violet-sensitive cone pigment resulted in UV light sensitivity of the pigment, whereas C90S mutation of zebra finch UV-sensitive cone pigment resulted in violet light sensitivity of the pigment [27]. In addition, F86Y mutation of mouse UV-sensitive cone pigment produced a violet light-sensitive pigment, whereas Y86F mutation of bovine violet-sensitive cone pigment produced a UV light-sensitive pigment [22]. These results suggested that the introduction of amino acids whose side chain contained a hydroxyl group at positions 86 and 90 led to the acquisition of visible light sensitivity. The analysis of Drosophila opsins showed that K90 mutants of UV-sensitive opsins formed violet-sensitive pigments, which suggested the importance of the lysine residue for the UV light sensitivity [24, 25]. Comparison of the sequences in the extracellular side of Helix II among vertebrate Opn5m highlighted the conservation of Lys91 in this region (Fig. 1A). Based on these previous studies, we prepared five mutants (I86S, I86T, G90S, G90T and K91T) of Opn5m (Fig. 1B) and analyzed their spectral property.
Fig. 2Spectral property of wild-type and mutants of Xenopus Opn5m. A Absorption spectra of wild-type Opn5m purified after incubation with all-trans retinal. Spectra were recorded in the dark (curve 1), after yellow light (> 500 nm) irradiation (curve 2), after subsequent UV light (360 nm) irradiation (curve 3) and after yellow light re-irradiation (curve 4). (inset) Spectral changes of wild-type Opn5m caused by yellow light irradiation (curve 1), subsequent UV light irradiation (curve 2) and yellow light re-irradiation (curve 3). B Calculated absorption spectra of wild-type Opn5m. The method of calculating the spectra of 11-cis (black curve) and all-trans retinal (red curve) bound forms of Opn5m is described in previous papers [8, 9]. The spectrum of 11-cis retinal or all-trans retinal bound forms has a peak in the UV region (360 nm) or in the blue region (474 nm), respectively. C-H Spectral changes of detergent-solubilized cell membranes expressing Opn5m. The cell membranes containing wild-type (C), I86S (D), I86T (E), G90S (F), G90T (G) and K91T (H) after the addition of all-trans retinal were solubilized with 1% DDM and their absorption spectra were recorded in the dark and after light irradiation. Spectral changes caused by yellow light irradiation (curve 1), subsequent UV light irradiation (curve 2), yellow light re-irradiation (curve 3) and UV light re-irradiation (curve 4) are shown
We obtained cell membranes containing the wild-type and five mutants of Opn5m after the addition of all-trans retinal and solubilized them with 1% DDM. Yellow light irradiation of the wild-type and four of the mutants (I86S, I86T, G90S and G90T) resulted in highly similar spectral changes, namely, a decrease of absorbance in the blue region and a concomitant increase of absorbance in the UV region (Figs. 2C-G). Spectral changes induced by subsequent UV and yellow light irradiations were mirror images of each other in these samples. Detailed comparison of the spectral change showed that the negative maximum, which corresponds to the peak of the all-trans retinal bound form, was slightly red-shifted in G90 mutants and slightly blue-shifted in I86 mutants (Fig. S1). These results indicated that I86S, I86T, G90S and G90T mutants are UV-sensitive bistable opsins, like wild-type, and these mutations had a slight effect on the spectral peak of the all-trans retinal bound form. By contrast, yellow light irradiation of K91T mutant resulted in a quite small increase of absorbance in the UV region accompanied by a large decrease of absorbance in the visible region (Fig. 2H). The spectral changes induced by subsequent UV and yellow light irradiations in K91T mutant were mirror images of each other. This result suggested that K91T mutation substantially affected the spectral property of Opn5m without hindering the bistable photoreaction.
Molecular property of K91T mutantTo analyze the detailed spectral property of K91T mutant, we purified K91T mutant after the addition of 11-cis retinal (Fig. 3A). The absorption spectrum had a peak at around 470 nm and UV light irradiation decreased the absorbance in the blue region. Subsequent yellow light irradiation induced a decrease of absorbance in the blue region and a quite small increase of absorbance in the UV region. Analysis of the retinal configuration revealed that the dark state predominantly contained all-trans retinal, which is formed by thermal isomerization of 11-cis retinal in culture medium, and yellow and UV light irradiations triggered the photoisomerization of the retinal into the 11-cis form (Fig. S2A). We also purified K91T mutant after the addition of all-trans retinal and analyzed the spectral and retinal configuration changes by light irradiations (Figs. 3B and S2B). In the dark, the K91T mutant sample had a spectral peak at around 470 nm and predominantly contained all-trans retinal. Yellow light irradiation led to a decrease of absorbance in the blue region by the conversion from all-trans retinal to 11-cis retinal, and subsequent UV light irradiation resulted in some recovery of absorbance in the blue region by the conversion of the retinal into all-trans form. Yellow and UV light re-irradiations induced repetitive spectral changes, which indicated that K91T mutant retained the bistable photoreaction. Based on the results shown in Figs. 3B and S2B, we calculated the absorption spectra of the 11-cis and all-trans retinal bound forms of K91T mutant (Figs. 3C and S2C). K91T mutation induced an ~ 75 nm red shift of the absorption maximum (λmax) in the 11-cis retinal bound form (435 nm) to form a visible light-sensitive pigment. By contrast, K91T mutation had only a slight effect on λmax of the all-trans retinal bound form (472 nm). Moreover, K91T mutant exhibited preferential binding to all-trans retinal.
Fig. 3Spectral property of K91T mutant of Opn5m. A Absorption spectra of K91T mutant of Opn5m purified after the incubation with 11-cis retinal. Spectra were recorded in the dark (curve 1), after UV light (360 nm) irradiation (curve 2), after subsequent yellow light (> 500 nm) irradiation (curve 3), after UV light re-irradiation (curve 4), after yellow light re-irradiation (curve 5) and after UV light re-irradiation (curve 6). (inset) Spectral changes of K91T mutant caused by UV light irradiation (curve 1), subsequent yellow light irradiation (curve 2), UV light re-irradiation (curve 3), yellow light re-irradiation (curve 4) and UV light re-irradiation (curve 5). B Absorption spectra of K91T mutant of Opn5m purified after the incubation with all-trans retinal. Spectra were recorded in the dark (curve 1), after yellow light irradiation (curve 2), after subsequent UV light irradiation (curve 3), after yellow light re-irradiation (curve 4), after UV light re-irradiation (curve 5) and after yellow light re-irradiation (curve 6). (inset) Spectral changes of K91T mutant caused by yellow light irradiation (curve 1), subsequent UV light irradiation (curve 2), yellow light re-irradiation (curve 3), UV light re-irradiation (curve 4) and yellow light re-irradiation (curve 5). C Calculated absorption spectra of K91T mutant of Opn5m. λmax of the 11-cis retinal and all-trans retinal bound forms are 435 nm and 472 nm, respectively. The method of calculating the spectra of 11-cis (black curve) and all-trans retinal (red curve) bound forms of K91T mutant is provided in Fig. S2
Spectral property of other K91 mutantsThe mutational analysis on the extracellular side of Helix II suggested the possibility that the well-conserved lysine residue at position 91 is a spectral tuning site in UV-sensitive Opn5m. Therefore, next, we analyzed the effects of other mutations at Lys91 on the spectral property of Opn5m. We replaced Lys91 with the other 18 naturally occurring amino acid residues and prepared the cell membranes containing these mutants after the addition of all-trans retinal (Fig. S3). After solubilizing the cell membranes with 1% DDM, we observed substantial spectral changes caused by light irradiations. Thus, all of these mutants formed photo-pigments upon the reconstitution with retinal. Comparison of the spectral changes showed that, in all of the mutants except K91R mutant, yellow light irradiation induced a large decrease of absorbance in the visible region, as in the wild-type (Fig. 2C). This suggested that K91R mutant has lost the ability to directly bind all-trans retinal. In addition, K91Y mutant exhibited an increase of absorbance in the UV region after yellow light irradiation, as did the wild-type (Fig. 2C), whereas the other mutants exhibited quite a small increase of absorbance in the UV region after yellow light irradiation, as did K91T mutant (Fig. 2H). We also prepared the cell membranes containing these mutants after the addition of 11-cis retinal (Fig. S4). After the cell membranes were solubilized with 1% DDM, K91R exhibited a decrease of absorbance in the UV region and a concomitant increase of absorbance in the blue region after UV light irradiation. This suggests that K91R mutant maintains the ability to directly bind 11-cis retinal. Moreover, in many of these mutants (K91A, K91C, K91D K91F, K91G, K91I, K91N, K91P, K91S, K91V and K91W), as in K91T mutant (inset of Fig. 3A), UV light irradiation did not induce a substantial increase of absorbance in the blue region. This suggested that these mutants have decreased affinity for 11-cis retinal. Based on these spectral analyses, we speculated that K91R and K91Y mutants form UV light-sensitive pigments, like wild-type, whereas the other mutants form visible light-sensitive pigments, like K91T mutant.
Molecular property of K91A and K91Q mutantsNext, we analyzed the detailed spectral property of several mutants. Among the 18 mutants shown in Figs. S3 and S4, we purified K91A and K91Q mutants, which are considered to form visible light-sensitive pigments, after the addition of 11-cis retinal (Fig. 4A and C). These mutant samples had spectral peaks in the blue region and contained not only 11-cis retinal but also a substantial amount of all-trans retinal (Figs. S5A and S5C). In K91A mutant, UV light irradiation and subsequent yellow light irradiation caused the isomerization of all-trans retinal to 11-cis retinal and decreased the absorbance in the blue region (Figs. 4A and S5A). In K91Q mutant, UV light irradiation increased the absorbance in the blue region by the isomerization of 11-cis retinal to all-trans retinal, and subsequent yellow light irradiation decreased the absorbance in the blue region by the isomerization of all-trans retinal to 11-cis retinal (Figs. 4C and S5C). Moreover, we prepared purified samples of K91A and K91Q mutants after the addition of all-trans retinal (Fig. 4B and D). These mutant samples had spectral peaks in the blue region and predominantly contained all-trans retinal (Figs. S5B and S5D). Yellow light irradiation of these samples decreased the absorbance in the blue region, and subsequent UV light irradiation resulted in some recovery of absorbance in the blue region. These spectral changes occurred as a result of photo-conversion between 11-cis retinal and all-trans retinal. Based on these results, we calculated the absorption spectra of the 11-cis retinal and all-trans retinal bound forms of K91A and K91Q mutants (Fig. 4E and F). As speculated based on the results shown in Figs. S3 and S4, the 11-cis retinal bound forms of K91A and K91Q mutants were maximally sensitive to violet light (λmax at 428 nm and 424 nm, respectively). By contrast, K91A and K91Q mutations had a quite small effect on the spectral sensitivity of the all-trans retinal bound form. Moreover, we confirmed that K91A and K91Q mutations resulted in preferential binding to all-trans retinal. The spectral property and binding preference of retinal isomers of these mutants are similar to those of K91T mutant (Fig. 3).
Fig. 4Spectral property of K91A and K91Q mutants of Opn5m. A, C Absorption spectra of K91A (A) and K91Q (C) mutants of Opn5m purified after the incubation with 11-cis retinal. Spectra were recorded in the dark (curve 1), after UV light (360 nm) irradiation (curve 2), after subsequent yellow light (> 500 nm) irradiation (curve 3), after UV light re-irradiation (curve 4) and after yellow light re-irradiation (curve 5). (inset) Spectral changes caused by UV light irradiation (curve 1), subsequent yellow light irradiation (curve 2), UV light re-irradiation (curve 3) and yellow light re-irradiation (curve 4). B, D Absorption spectra of K91A (B) and K91Q (D) mutants of Opn5m purified after the incubation with all-trans retinal. Spectra were recorded in the dark (curve 1), after yellow light irradiation (curve 2), after subsequent UV light irradiation (curve 3), after yellow light re-irradiation (curve 4) and after UV light re-irradiation (curve 5). (inset) Spectral changes caused by yellow light irradiation (curve 1), subsequent UV light irradiation (curve 2), yellow light re-irradiation (curve 3) and UV light re-irradiation (curve 4). E, F Calculated absorption spectra of K91A (E) and K91Q (F) mutants of Opn5m. λmax of the 11-cis retinal and all-trans retinal bound forms of K91A are 428 nm and 466 nm, respectively. λmax of the 11-cis retinal and all-trans retinal bound forms of K91Q are 424 nm and 468 nm, respectively. The method of calculating the spectra of 11-cis (black curve) and all-trans retinal (red curve) bound forms is provided in Fig. S5
Molecular property of K91R and K91Y mutantsWe also analyzed the molecular property of K91R and K91Y mutants, because these mutants are considered to form UV light-sensitive pigments. We purified these mutants after the addition of 11-cis retinal (Fig. 5A and C). K91R mutant sample had a spectral peak in the UV region (Fig. 5A) and predominantly contained 11-cis retinal (Fig. S6A). UV light irradiation shifted the spectrum into the blue region and subsequent yellow light irradiation recovered the spectrum in the UV region (Fig. 5A). These spectral changes were triggered by the photo-conversion between 11-cis retinal and all-trans retinal (Fig. S6A). K91Y mutant sample had substantial absorbance in the blue region (Fig. 5C). This was probably because this sample contained a substantial amount of all-trans retinal in addition to 11-cis retinal (Fig. S6B). UV light irradiation and subsequent yellow light irradiation induced an increase and a decrease of absorbance in the blue region, respectively (Fig. 5C). These spectral changes were mainly caused by the inter-conversion between 11-cis retinal and all-trans retinal (Fig. S6B). We also purified these mutants after the addition of all-trans retinal (Fig. 5B and D). K91R mutant sample exhibited no clear spectral peak. Yellow light irradiation induced no spectral change and subsequent UV light irradiation induced only a quite small increase in the blue region (Fig. 5B), which is consistent with the result shown in Fig. S3. This spectral property indicated that K91R mutant lost the ability to directly bind all-trans retinal. K91Y mutant had a spectral peak in the blue region (Fig. 5D) and predominantly contained all-trans retinal (Fig. S6C). Yellow light irradiation shifted the spectrum into the UV region and subsequent UV light irradiation resulted in some recovery of absorbance in the blue region (Fig. 5D). During these spectral changes, inter-conversion between 11-cis retinal and all-trans retinal was observed (Fig. S6C). Based on these results, we calculated the absorption spectra of the 11-cis retinal and all-trans retinal bound forms of K91R and K91Y mutants (Fig. 5E and F). As speculated based on the results shown in Figs. S3 and S4, the 11-cis retinal bound form of K91R mutant was maximally sensitive to UV light (λmax at 362 nm), like wild-type (Fig. 5E). The all-trans retinal bound form of K91R mutant had a spectral peak at 454 nm, ~ 20 nm blue-shifted from that of wild-type. Thus, K91R mutation resulted in maintenance of UV light sensitivity and preferential binding to 11-cis retinal. By contrast, the 11-cis retinal bound form of K91Y mutant had a spectral peak in the UV region and a spectral shoulder at around 430 nm (Fig. 5F). K91Y mutation had a small effect (~ 9 nm blue-shift) on the spectral sensitivity of the all-trans retina bound form. This spectral property showed that K91Y mutant contained UV light-sensitive and visible light-sensitive components in the 11-cis retinal bound form.
Fig. 5Spectral property of K91R and K91Y mutants of Opn5m. A, C Absorption spectra of K91R (A) and K91Y (C) mutants of Opn5m purified after the incubation with 11-cis retinal. Spectra were recorded in the dark (curve 1), after UV light (360 nm) irradiation (curve 2), after subsequent yellow light (> 500 nm) irradiation (curve 3), after UV light re-irradiation (curve 4) and after yellow light re-irradiation (curve 5). (inset) Spectral changes caused by UV light irradiation (curve 1), subsequent yellow light irradiation (curve 2), UV light re-irradiation (curve 3) and yellow light re-irradiation (curve 4). B, D Absorption spectra of K91R (B) and K91Y (D) mutants of Opn5m purified after the incubation with all-trans retinal. Spectra were recorded in the dark (curve 1), after yellow light irradiation (curve 2), after subsequent UV light irradiation (curve 3), after yellow light re-irradiation (curve 4) and after UV light re-irradiation (curve 5). (inset) Spectral changes caused by yellow light irradiation (curve 1), subsequent UV light irradiation (curve 2), yellow light re-irradiation (curve 3) and UV light re-irradiation (curve 4). E, F Calculated absorption spectra of K91R (E) and K91Y (F) mutants of Opn5m. λmax of the 11-cis retinal and all-trans retinal bound forms of K91R are 362 nm and 454 nm, respectively. λmax of the 11-cis retinal and all-trans retinal bound forms of K91Y are 360 nm and 465 nm, respectively. The method of calculating the spectra of 11-cis (black curve) and all-trans retinal (red curve) bound forms is provided in Fig. S6
Our spectral analysis was performed at pH 6.5. In this pH condition, a unique mixture of UV light-sensitive and visible light-sensitive components was observed in K91Y mutant. These two spectrally distinguishable components may result from the protonation and deprotonation of the Schiff base within the protein. Thus, we analyzed the spectral property of K91Y mutant under different pH conditions. K91Y mutant purified after the addition of all-trans retinal had a spectral peak in the blue region at pH 6 (Fig. 6A) and pH 7 (Fig. 6B). These two samples predominantly contained all-trans retinal (Fig. S7A and S7B). Yellow light irradiation of these two samples decreased the absorbance in the blue region and increased the absorbance in the UV region (Fig. 6A and B). In this process, the spectral change in the UV region at pH 7 was larger than that at pH 6 (insets of Fig. 6A and B). Moreover, the profiles of the retinal configuration changes caused by this yellow light irradiation were quite similar to each other (Figs. S7A and S7B). Based on these results, we calculated the absorption spectra of 11-cis retinal and all-trans retinal bound forms of K91Y mutant at pH 6 and pH 7 (Fig. 6C and D). Comparison of these spectra of the 11-cis retinal bound form showed that the ratio of the UV light-sensitive component at pH 7 was larger than that at pH 6. This can be explained by the increase of the ratio of the component with the deprotonated Schiff base at pH 7 compared to that at pH 6. Thus, these results suggested that K91Y mutant contained two spectrally distinguishable components which have protonated and deprotonated Schiff base under neutral conditions.
Fig. 6pH-dependent spectral changes of K91Y mutant of Opn5m. A, B Absorption spectra of K91Y mutant of Opn5m purified after the incubation with all-trans retinal. Spectra were recorded in the dark (curve 1), after yellow light irradiation (> 500 nm) (curve 2), after subsequent UV light (360 nm) irradiation (curve 3), after yellow light re-irradiation (curve 4) and after UV light re-irradiation (curve 5) at pH 6 (A) or pH 7 (B). (inset) Spectral changes caused by yellow light irradiation (curve 1), subsequent UV light irradiation (curve 2), yellow light re-irradiation (curve 3) and UV light re-irradiation (curve 4). C, D Calculated absorption spectra of K91Y mutant of Opn5m at pH 6 (C) or pH 7 (D). The method of calculating the spectra of 11-cis (black curve) and all-trans retinal (red curve) bound forms is provided in Fig. S7
Comments (0)