Dragonfly red opsins share a common tuning mechanism with mammalian red opsins and further enhancement of near-infrared sensitivity

Abstract

Some animals, such as primates and insects have color vision including sensitivity to red light (red vision). Red vision is basically achieved through opsins sensitive to the red region (red opsins), which independently evolved in different lineages. In dragonfly red vision, which is known to sense longer-wavelength light compared with humans, however, the underlying opsins and the spectral tuning mechanism are largely unknown. Here we investigated dragonfly opsins and found that RhLWA2s are the longest-wavelength-sensitive opsins, so-called red opsins in dragonflies. Spectroscopic analysis of the recombinant pigment of RhLWA2 from Asiagomphus melaenops (Am_RhLWA2) revealed that it has an absorption maximum at 580 nm and exhibits bistability, indicating that Am_RhLWA2 is the longest-wavelength-sensitive bistable opsin to date. Mutational analysis of Am_RhLWA2 revealed that position 292 is responsible for the red shift. The spectral tuning site as well as the mechanism for the red shift (S292A) is shared with that of mammalian red opsins, showing parallel evolution between mammalian and insect green/red opsins, and the substitution from Ala to Val (A292V) in a dragonfly lineage further enhanced the red sensitivity to near-infrared region. Furthermore, we succeeded in engineering red-shifted Am_RhLWA2 mutant having an absorption maximum at 590 nm by introducing V211C mutation. Cultured cells expressing the red-shifted Am_RhLWA2 mutant exhibited significant Ca²⁺ responses to 738 nm light, showing the potential of near-infrared sensitive optogenetic tools to control GPCR-signaling. Based on the analysis of body coloration of a related dragonfly species, the longer-wavelength sensitivity of Am_RhLWA2 could confer an advantage in sex recognition.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-025-06017-9.

Introduction

Opsins, members of G protein coupled receptors (GPCRs) bind to a chromophore retinal to form photosensitive pigments (opsin-based pigments) for vision and non-visual function of animals. Absorption spectra of opsin-based pigments are tuned from UV to red regions based on amino acid residues basically surrounding the chromophore, which underlies spectral sensitivities of various photoreception [1]. Diversity in spectral sensitivity is particularly crucial for color vision. Some animals have color vision involving red sensing (red vision), reaching 700 nm light, which had evolved multiple times in the animal kingdom [2–4]. The red-vision is basically achieved by long-wavelength-sensitive (LW) opsins evolved independently in different lineages, and therefore the spectral tuning mechanisms could vary [5–8].

Spectroscopic analyses of recombinant LW opsin-based pigments expressed in cultured cells and their site-specific and/or chimeric mutants have uncovered the amino acid residues responsible for the red-shift of absorption spectra. In many vertebrate LW opsins, whose wavelength of maximum absorption (λ_max) is ~ 560 nm, chloride ion binding to the site involving histidine residue at position 181 (bovine rhodopsin numbering system) results in a red shift of λ_max by 15–45 nm [9, 10], and amino acid residues at positions 289 and 292 are also responsible for the chloride effect in some mammalian LW opsins [11]. In primate LW opsins (red and green opsins), which diverged around 30 million years ago, the ~ 30 nm difference in λ_max between red and green opsins is explained by substitutions at positions 164 (5 nm), 261 (10 nm) and 269 (17 nm) [12, 13]. In mammalian LW opsins, the ~ 50 nm difference in λ_max between human red opsin and mouse green opsin is primarily explained by the substitution at position 292 (18 nm), along with that at position 181 (28 nm) [12, 14].

Red vision has also been reported in some arthropods, such as butterflies, dragonflies, and crustacean mantis shrimps [2, 15]. In the case of these arthropods, the red sensitivities in vision, having peak wavelengths of up to 700 nm, are achieved through combinations of the absorption spectra of LW opsins and filtering pigments present in ommatidia in compound eyes [16]. Therefore, elucidating the absorption spectra of LW opsins is particularly important for understanding the molecular basis of arthropod red vision. However, due to difficulty in expression of invertebrate opsins in cultured cells, the molecular basis for arthropod red vision including the red-shift mechanism, has been uncovered only to a limited extent. We have developed the cultured-cell expression system for invertebrate opsins [8, 17–21], and using the system, we succeeded in the spectroscopic analysis of the recombinant LW opsin (PxRh3)-based pigment of the swallowtail butterfly Papilio xuthus, revealing that the λ_max is ~ 570 nm [7]. Furthermore, we experimentally revealed, using heterologous action spectroscopy (HAS) [22], that helix III, particularly the amino acid residues at positions 116 (+ 6 nm) and 120 (+ 3 nm), are crucial for the red shift in PxRh3, demonstrating a different spectral tuning mechanism from that of vertebrate LW opsins. Absorption spectra of LW opsin in some other arthropods have also been determined, with their λ_max ranging from ~ 518 nm (Callophrys sheridanii LW) to ~ 578 nm (Arhopala. japonica LW) in lepidopterans, 542 nm (RhLWD1) and 557 nm (RhLWA2) in a dragonfly Sympetrum frequens, and 539 nm (NoL6) and 611 nm (NoL14) in a mantis shrimp Neogonodactylus oerstedii, although spectral tuning sites as well as mechanisms underlying their red shift have yet to be revealed [23, 24]. Interestingly, in dragonflies, LW opsins are extraordinarily diversified [25] and only two LW opsins have been investigated, with the longer one exhibiting a λ_max of 557 nm [24], whereas photoresponses of photoreceptor cells in some dragonfly species showed a sensitivity peak at ~ 600 nm [2]. These facts raise the possibility that longer-wavelength-sensitive LW opsins may exist in dragonflies. In addition, exploring LW opsins sensitive to longer-wavelength light would be valuable for optogenetic applications because such light penetrates biological tissues effectively, enabling efficient photoactivation of tools in deeper regions of the body [26–28].

Here we systematically investigated spectral sensitivities of several kinds of LW opsins of dragonflies to understand the red shift mechanism in dragonfly LW opsins that evolved independently from those of other animals, and to develop powerful optogenetic tools. We succeeded in spectroscopic analysis of the recombinant pigment of Asiagomphus melaenops RhLWA2 (Am_RhLWA2) and revealed that it has an absorption maximum at 580 nm and exhibits bistability, indicating that Am_RhLWA2 is the longest-wavelength-sensitive bistable red opsin to date. Furthermore, we determined that the position 292 is responsible for the red shift in dragonfly LW opsins, showing parallel evolution between the mammalian and insect LW opsins, and found that further enhancement of red sensitivity occurred by the substitution to the other amino acid in a dragonfly lineage. We also successfully engineered the red-shifted Am_RhLWA2 and demonstrated significant cellular responses of cultured cells expressing the red-shifted Am_RhLWA2 against near-infrared (NIR) light (738 nm), showing a potential for NIR-sensitive optogenetic tool for controlling GPCR-signaling.

Materials and methods

Plasmid construction

The cDNAs of dragonfly RhLWs and their mutants were tagged with the epitope sequence of the anti-bovine rhodopsin antibody Rho1D4 (ETSQVAPA) at their C-termini. The tagged cDNA was inserted into a pMT plasmid vector [29]. For heterologous action spectroscopy (HAS), Gs-coupled dragonfly RhLWs were constructed by replacing the third cytoplasmic loop of dragonfly RhLWs with that of Gs-coupled jellyfish opsin (Fig. S1), as described [19], according to the previous report [22]. Site-directed mutants were produced by overlap extension PCR using PrimeSTAR Max DNA Polymerase (TAKARA, Shiga, Japan) with site-specific primers. The amino acid sequences used in this study are available from the INSDC databases under the accession numbers as follows: Sympetrum frequens (RhLWA1, LC009060; RhLWA2, LC009061; RhLWF1, LC009066; RhLWF2, LC009067; RhLWF3, LC009068; RhLWF4, LC009069); Orthetrum albistylum RhLWA2, LC009081; Somatochlora uchidai RhLWA2, LC009101; Macromia amphigena RhLWA2, LC009125; Anotogaster sieboldii RhLWA2, LC009150; Tanypteryx pryeri RhLWA2, LC009170; Asiagomphus melaenops RhLWA2, LC009188; Anax parthenope RhLWA2, LC009205; Epiophlebia superstes RhLWA2, LC009239; Ischnura asiatica RhLWA2, LC009262; Mnais costalis RhLWA2, LC009278; Indolestes peregrinus RhLWA2, LC009293.

Expression of opsin-based pigments and spectroscopy

Pigment expression in HEK293S cells and purification were carried out as previously described [30]. Briefly, the expression constructs were transfected into HEK293S cells by the calcium-phosphate method. The cells were cultured for 48 h for protein expression. For pigment reconstitution, the expressed proteins were incubated with an excess amount of 11-cis retinal overnight at 4 °C in the dark. The pigments were then extracted with 1% (w/v) n-dodecyl-β-D-maltoside (DM) in 50 mM HEPES buffer (pH 6.5) containing 140 mM NaCl (Buffer A). The lipid-containing extracts were used for measurement of photoreactions as previously described [31]. Pigments in the extract were bound to 1D4-agarose, washed with 0.02% (w/v) DM in Buffer A, and eluted with 0.02% (w/v) DM in Buffer A containing the 1D4 peptide. The absorption spectra of purified pigments were measured using a V-750 UV-Vis Spectrophotometer (JASCO International). The absorption spectra in the longer wavelength region were fitted with the rhodopsin nomogram [32] to estimate the λ_max. Red and green lights were supplied by a 1-kW halogen lamp (Philips, Eindhoven, Netherlands) with a R64 glass cutoff filter and a 550-nm interference filter (Toshiba, Tokyo, Japan), respectively.

HPLC analysis of chromophore configuration

High-performance liquid chromatography (HPLC) was carried out using a Shimadzu LC-7 A interfaced with CR-5 A to analyze the conformations of retinal present in the purified pigments as described previously [17]. Briefly, 100 µL of purified pigments were mixed with 210 µL of cold 90% methanol which was stored in − 20 °C and 30 µL of 1 M hydroxylamine to convert retinal chromophore in a sample into retinal oxime. The retinal oxime was extracted with 700 µL of n-hexane. 200 µL of the extract were injected into a YMC- Pack SIL column (particle size 3 μm, 150 × 6.0 mm2) and eluted with n-hexane containing 15% (vol/vol) ethyl acetate and 0.15% (vol/vol) ethanol at a flow rate of 1 mL/min while being monitored by the absorbance at 360 nm.

Bioluminescent reporter assays for cAMP and Ca²⁺

The intracellular cAMP and Ca²⁺ levels in opsin-expressing HEK293S cells were measured using the GloSensor cAMP assay and the aequorin assay, as previously described [21, 33, 34]. In detail, 1.5 µg of each opsin plasmid was transfected into HEK293S cells (~ 80% confluent) in 35 mm culture dishes (IWAKI) with 1.5 µg of either the pGloSensor-22F cAMP plasmid or the aequorin plasmid. The pGloSensor-22F cAMP plasmid (Promega) was used for the GloSensor cAMP assay. The aequorin mutant obtained by introducing a reverse mutation, A119D, into the plasmid [pcDNA3.1+/mit-2mutAEQ] (Addgene #45539) [35] was used for the aequorin assay. The expression constructs for the dragonfly opsins and their mutants were cotransfected with the plasmids containing pGloSensor-22F cAMP or the aequorin mutant by using the polyethyleneimine (PEI) transfection method. After transfection, the cells were incubated for 24 h at 37 °C under 5% CO₂, followed by treatment with 11-cis retinal and overnight incubation at 25 °C. Before the luminescence measurements, the culture medium was replaced with a medium containing either GloSensor cAMP Reagent stock solution (Promega) or Coelenterazine h (Wako), for cAMP and Ca²⁺ measurements, respectively, and the cells were incubated to equilibrate with the medium at 25 °C for at least 2 hours. Luminescence was measured using a GloMax 20/20n Luminometer (Promega). LEDs with spectral emission peaks of 510 nm and 630 nm arrayed on a board (SPL-25-CC; REVOX Inc., Kanagawa, Japan) were used as light sources for HAS [22]. The quantal intensities of these two lights were adjusted to 1.9 *10¹⁵ photons/cm²/sec, a level that does not induce saturated responses, using interference filters (MZ0510 and MZ0630; Asahi Spectra Co., Ltd.) and neutral-density filters (SIGMAKOKI Co., Ltd., Saitama, Japan and Shibuya Optical Co., Ltd., Saitama, Japan), and applied for 5 s in the GloSensor cAMP assay as light stimuli, as described [7]. A broadband green LED light (536 nm; 1.9 *10¹⁵ photons/cm²/sec) with a band-pass filter (540 nm, Toshiba, Tokyo, Japan) and a narrow-band infrared LED light (738 nm; 1.6 *10¹⁶ photons/cm²/sec, 7.9 *10¹⁵ photons/cm²/sec, 5.2 *10¹⁵ photons/cm²/sec, 3.0 *10¹⁵ photons/cm²/sec, and 1.6 *10¹⁵ photons/cm²/sec) with a long-pass filter (LV0730, Asahi Spectra Co., Ltd., Osaka, Japan) were applied for 1 s in the aequorin assay as light stimuli.

Protein structure prediction

The protein structure of Am_RhLWA2 was predicted using ColabFold [36] with the default parameters. Five structural models were generated, and the model with the highest average pLDDT (predicted local distance difference test) score was selected for further analysis. The protein structure of Am_RhLWA2 was aligned with spider Rh1 (PDB ID: 6I9K) [37] and bovine rhodopsin (PDB ID: 1U19) [38]. Figures displaying three-dimensional structures were prepared with PyMOL.

Reflectance measurement of body coloration

Gomphidae dragonflies, Sieboldius albardae were collected in Katano city, Mino city, Izumisano city and Botanical Gardens of Osaka Metropolitan University, Osaka, Japan in June–July 2025. Reflectance spectra of the yellow coloration of the thorax in the dragonflies were measured using a Flame-S-UV-VIS spectrometer and OceanView software (OceanOptics) from 250 to 750 nm. Reflectance spectra were expressed relative to a white diffuse reflectance standard (WS1, Ocean Optics).

cDNA cloning

Total RNA extracted from the compound eye of Sieboldius albardae were reverse transcribed to cDNAs by using an oligo (dT) primer, and the cDNAs were used as templates for PCR amplification. cDNA fragments of RhLWA2 from this species were obtained by PCR cloning using the following primers designed based on the 5’- and 3’-untranslated regions from Am_RhLWA2: 5′-ccacaagaaaactacagccaaaatg-3′ as the sense primer and 5′-gtgaaataaacaaaagaaaagcggtattta-3′ as the antisense primer. PCR amplifications were carried out with annealing temperatures of 50 °C. The nucleotide sequence of Sa_RhLWA2 cDNA, encoding RhLWA2 of Sieboldius albardae, has been deposited in the DNA Data Bank of Japan (DDBJ) and is available in the INSDC databases under accession number LC899447.

Results

Heterologous action spectroscopy-based initial screening of red opsins in dragonflies

Insect visual opsins are phylogenetically classified into three groups, members of which are sensitive to UV, short-wavelength and long-wavelength light, respectively [6]. A comprehensive study on opsins of 12 dragonfly species uncovered 200 visual opsins, and 136 opsins fall into the insect LW group, which is further divided into six subgroups (LWA, LWB, LWC, LWD, LWE and LWF) [25]. Among the dragonfly LW opsins, RhLWA2 in the LWA subgroup and RhLWF1-4 in the LWF subgroup are expressed in the ventral region of adult eyes of Sympetrum frequens, where electrophysiological responses to longer-wavelength light have been recorded [25]. Therefore, we focused on the five LW opsins as candidates for longest-wavelength-sensitive LW opsins, so-called red opsins. Since expression and purification of recombinant photopigments of arthropod opsins in cultured cells often fails, we applied heterologous action spectroscopy (HAS), which is based on light-induced increases of cAMP levels in cultured cells expressing Gs-coupled opsins under each color of light with an equal photon number [22], to roughly estimate spectral sensitivities of the dragonfly LW opsins. For HAS, we generated Gs-coupled mutants of Sympetrum frequens RhLWA2 (Sf_RhLWA2) and RhLWF1-4 (Sf_RhLWF1-4), in which the third cytoplasmic loops were replaced with that of a Gs-coupled jellyfish opsin [19] (Fig. 1A). We performed a simplified version of HAS using two wavelengths of light in this initial screening. We selected 510 nm and 630 nm, which are expected to cause a similar level of light-induced cAMP increases, assuming a λ_max of ~ 575 nm as a reference for LW opsins (Fig. 1B). As results, the Gs-coupled Sf_RhLWA2 exhibited a similar level of cAMP increases for both wavelengths of light, like the Gs-coupled butterfly LW opsin (PxRh3), whereas the Gs-coupled Sf_RhLWF1-4 exhibited much higher cAMP increases for 510 nm light, strongly suggesting that the λ_max of RhLWA2 is the longest among those of dragonfly LW opsins (Fig. 1C). We also investigated spectral sensitivities of RhLWA1 because they are closely related to RhLWA2 in the molecular phylogenetic tree, whereas they are highly expressed during larval stages [25]. HAS revealed that cultured cells expressing the Gs-coupled Sf_RhLWA1 exhibited a significant response to 630 nm light; however, the amplitude was smaller than that to 510 nm light (Fig. 1C). This suggests that the λ_max of RhLWA1 is longer than those of RhLWF1-4 but is shorter than that of RhLWA2, allowing for focusing on RhLWA2 as red opsins in dragonflies. Then, we surveyed the RhLWA2 that showed the highest cAMP increases, which suggests high expression level in cultured cells among RhLWA2s of 12 dragonfly species. As results, we found that cultured cells expressing the Gs-coupled version of Asiagomphus melaenops RhLWA2 (Am_RhLWA2) exhibited much larger light responses compared to the Gs-coupled RhLWA2s of other dragonflies (Fig. 1D), enabling further studies with Am_RhLWA2.

Fig. 1.

Cellular response-based screening of longer-wavelength-sensitive LW opsins in dragonflies (A) Dragonfly RhLWs (Dragonfly RhLW) were engineered to activate Gs type G protein (Dragonfly Gs-RhLW) by replacing their third cytoplasmic loops (iL3) with that of the Gs-coupled jellyfish opsin (Jellyfish Gs-Op) to enable light-induced increases in intracellular cAMP levels. (B) Experimental design for rough estimation of the absorption maximum (λ_max) by irradiation with 510 nm (cyan vertical line) and 630 nm (yellow vertical line) light. The two wavelengths were selected based on the assumption that opsin-based pigments with a λ_max at ~ 575 nm would exhibit similar levels of light-induced cAMP increases upon irradiation with 510 nm and 630 nm. (C and D) Peak values of cAMP increases in HEK293 cells expressing various dragonfly Gs-RhLWs by irradiation with 510 nm (cyan bars) and 630 nm (yellow bars) light. A butterfly LW opsin (PxRh3) having the λ_max at ~ 570 nm was included as a control. Error bars indicate the means ± SE (n = 3) of the measured peak values of cAMP increases. (C) Comparison of the relationships between responses to 510 nm and 630 nm light among RhLWs of S. frequens, suggesting that RhLWA2 has the longest-wavelength sensitivity. (D) Comparison of the relationships between responses to 510 nm and 630 nm light among RhLWA2s of 12 dragonfly species. Am, Asiagomphus melaenops; Ma, Macromia amphigena; Sf, Sympetrum frequens; Ip, Indolestes peregrinus; Ap, Anax parthenope; Es, Epiophlebia superstes; As, Anotogaster sieboldii; Tp, Tanypteryx pryeri; Mc, Mnais costalis; Ia, Ischnura asiatica; Su, Somatochlora uchidai; Oa, Orthetrum albistylum

Absorption spectrum of Am_RhLWA2 and spectral tuning sites for the red shift

We expressed Am_RhLWA2 in cultured cells and succeeded in obtaining the recombinant pigment of Am_RhLWA2. The spectroscopic analysis revealed that Am_RhLWA2 exhibited an absorption maximum at 580 nm in the dark, which is longer than that of butterfly PxRh3 [7] (Fig. 2A). It should be noted that the fitting curve does not closely match the actual absorption spectrum in the shorter-wavelength region because scattering affects absorption spectra in that region, especially in low-concentration samples [7]. We then analyzed the spectroscopic property of Am_RhLWA2 in detail. Red-light irradiation of Am_RhLWA2 resulted in decrease and increase of the absorbance around 640 nm and around 550 nm, respectively (Fig. 2B). The spectral change by the light irradiation is explained by the conversion of the dark state to the photoproduct having blue-shifted absorption spectrum (Fig. S4). Subsequent green-light irradiation caused the opposite reaction, which indicates the reverse conversion of the photoproduct to the dark state. A second red-light irradiation resulted in a spectral change that was the mirror image of those caused by the first and second green-light irradiations (Fig. S4). The photoreaction clearly fits the criteria of bistable opsins, which convert to stable photoproducts upon light absorption, and the photoproducts revert to the original dark state by subsequent light absorption [30, 31]. Furthermore, we also carried out HPLC analysis to examine the retinal isomerization in Am_RhLWA2 by light irradiations. Consistent with the mirror-image difference spectra, isomerization of 11-cis to all-trans and all-trans to 11-cis were observed by red light preferentially absorbed by the dark state and green-light preferentially absorbed by the photoproduct, respectively (Fig. 2C). These results clearly demonstrated that Am_RhLWA2 is a bistable red opsin. Next, we investigated amino acid residues responsible for the red shift of RhLWA2 compared with RhLWF1–4, which have shorter-wavelength sensitivity (Fig. 1C), to understand the spectral tuning mechanism in dragonfly LW opsins. Cultured cells expressing other kinds of LW opsins, RhLWF1-4 exhibited a significant response to 540 nm light, whereas they responded slightly to 630 nm light (Fig. 1C). The HAS profiles of RhLWF1-4 are similar to that of PxRh1 with λ_max at ~ 540 nm in our previous report [7], suggesting that their λ_max values are around ~ 540 nm in the green region. Then, we compared the amino acid resides within 5Å of the chromophore retinal, which could play significant roles in spectral tuning [39, 40] between RhLWA2 and RhLWF1-4 to find the residues responsible for the red shift of RhLWA2. For this purpose, we deduced these residues from the crystal structure of jumping spider rhodopsin (6I9K) [37], which is closely related to dragonfly opsins. The comparison revealed that positions 114, 186 and 292 are occupied by mutually exclusive amino acids in RhLWA2s and RhLWF1-4 (Fig. 2D). We conducted spectroscopic analysis of Am_RhLWA2 mutants at these positions having amino acid residues in RhLWF1-4. As results, although the G114A and S186A mutants did not form functional photopigments, we observed an obvious absorption spectrum for the V292 mutant. Remarkably, the λ_max of the V292S mutant is 546 nm, similar to the estimated λ_max of RhLWF1-4, demonstrating that the substitution at position 292 accounts for most of the spectral shift between Am_RhLWA2 and RhLWF1-4 (Fig. 2E). Interestingly, among dragonfly RhLWA2, only Am_RhLWA2 has Val at position 292, whereas RhLWA2 of other dragonfly species have Ala (Fig. 2D). The Am_RhLWA2 mutant having Ala at position 292 (V292A) exhibited an absorption spectrum having λ_max at 562 nm, which is 16 nm longer than λ_max of the V292S mutant and 18 nm shorter than λ_max of the wild type (Fig. 2E, F). These results revealed that position 292 plays a key role in spectral tuning of RhLW, and Am_RhLWA2 is particularly long-wavelength-sensitive RhLWA2.

Fig. 2.

Spectroscopic characteristics and the spectral tuning site of A. melaenops RhLWA2 (Am_RhLWA2) (A) The absorption spectrum of purified Am_RhLWA2 in the dark (black) fitted with rhodopsin nomogram (red) to estimate that Am_RhLWA2 has an absorption maximum at ~ 580 nm. (B) Difference spectra of lipid-containing extracts of Am_RhLWA2-expressing cells showing spectral changes induced by irradiation with red light (1: orange), subsequent green light (2: green), a second red light (3: grey) and a second green light (4: blue). The mirror-image relationship of the green and blue curves and the grey curve indicates that Am_RhLWA2 is a bistable opsin (See also Fig. S4, which illustrates the relationship between difference spectra and absolute spectra). (C) The configuration of retinal in Am_RhLWA2 in the dark (black), after irradiation with red light (red trace), subsequent green light (green trace), a second red light (yellow trace), and a second green light (blue trace) analyzed by HPLC. Retinal was extracted in its oxime form. AT, all-trans-retinal oxime; 11, 11-cis-retinal oxime. (D) Comparison of residues near the retinal between dragonfly RhLWA2s and RhLWFs. Residues near the retinal were selected with reference to the crystal structure of jumping spider rhodopsin (PDB ID: 6I9K). Amino acid residues at positions 114, 186 and 292 that differ between LWA2 group (yellow) and LWF group (cyan) are highlighted. Putative counterion residue E181 [66, 67] and the retinal-binding K296 are also shown (grey). Residues are numbered according to bovine rhodopsin. (E) Absorption spectra of the V292S mutant (cyan) fitted with rhodopsin nomogram (red) and the wild type (WT, grey) of the Am_RhLWA2. (F) Absorption spectra of the V292A mutant (yellow) fitted with rhodopsin nomogram (red) and the wild type (WT, grey) of the Am_RhLWA2. Each absorption spectrum was normalized to the maximum absorbance in the visible light region

Engineering further red-shifted red opsins and NIR-sensitive GPCR tools

In addition to their biological importance, rhodopsins have become powerful tools for optogenetics, a technique that genetically introduces light-sensitive proteins into target cells and manipulates the cellular activity with light, particularly in neural cells where rhodopsin genes are genetically expressed [41–44]. It is well known that longer wavelength light can penetrate deeper into biological tissues [26]. Therefore, long wavelength-sensitive tools have a big advantage for non-invasive as well as efficient optical controls of physiological responses at deeper tissue [27, 28]. Given these advantages, we aimed to develop further red-shifted red opsins compared to the native ones for near-infrared (NIR)-sensitive optogenetic applications, and to understand the potential of red sensitivity of animal opsins. Then we introduced amino acid substitutions that cause red shift in red opsins of other animals into Am_RhLWA2 and carried out spectroscopic analysis (Fig. 3A). Unfortunately, however, Am_RhLWA2 having mutations at red-shift sites of human red opsin (OPN1LW) (181, 261, 269) and butterfly PxRh3 (116, 120) did not exhibit any absorption spectra derived from recombinant pigments, probably due to their low expression level in cultured cells and/or low stability in the detergent by mutations. In mantis shrimp LW opsins, the red-shift sites have not been determined but the comparison of amino acids within 5Å of the chromophore between green-sensitive NoL6 and red-sensitive NoL14 suggested positions 90, 118, 208, 211 and 212 might be candidates for the red-shift sites [24] (Fig. 3A). Since amino acids at positions 90, 118, and 212 were shared between NoL14 and Am_RhLWA2, we replaced amino acids at positions 208 and 211 in Am_RhLWA2 with those in NoL14. Although the S208T mutant exhibited an absorption spectrum not significantly different from Am_RhLWA2 wild type (Fig. 3B), the V211C mutant exhibited a significantly red shifted spectrum (Fig. 3C). Based on the rhodopsin nomogram, the λ_max of the V211C mutant is 590 nm, demonstrating the successful engineering of red opsins for a further red shift by introducing the “heterologous” mutation.

Fig. 3.

Absorption spectrum of the red-shifted Am_RhLWA2 mutant (A) Spectral tuning sites of LW opsins in other animals. Amino acid residues in red opsins and green opsins are highlighted in yellow and blue, respectively. Hs and Mm indicates human and mouse, respectively. In the case of NoL from a mantis shrimp, candidate spectral tuning sites were indicated. (B) Absorption spectra of the S208T mutant (magenta) and the wild type (grey) of the Am_RhLWA2, showing no spectral shift. (C) Absorption spectra of the V211C mutant (magenta) and the wild type (grey) of the Am_RhLWA2. The estimated spectrum using rhodopsin nomogram (yellow) indicates that λ_max of the V211C mutant is 590 nm. The rhodopsin nomogram for the wild type (cyan) is also shown. Each absorption spectrum was normalized to the maximum absorbance in the visible light region. Note that the S208T/V211C double mutant did not form functional photopigments

We then examined optogenetic potentials of dragonfly red opsins for NIR-sensitive GPCR optogenetic tools. Because insect visual opsins are known to be primarily Gq-coupled opsins [45], we measured 738-nm-light-induced intracellular Ca²⁺ elevation in cultured cells expressing dragonfly red opsins in different light intensities. Aequorin assay, which detects intracellular Ca²⁺ levels as luminescence levels, revealed that cultured cells expressing Am_RhLWA2 wild type or the V211C mutant exhibited significant Ca²⁺ responses to 738-nm light, whereas those expressing jumping spider Rh1 (SpiRh1), a model of Gq-coupled bistable opsin [6, 20, 37, 46], did not (Fig. 4A). It is remarkable that the response amplitude of cultured cells expressing the Am_RhLWA2 V211C mutant was up to ten times larger than that of the wild type (Fig. 4B), clearly demonstrating the high performance of the red-shifted mutant of Am_RhLWA2 for longer-wavelength light. In contrast, 536-nm light stimulation caused a larger elevation of intracellular Ca²⁺ levels in cultured cells expressing Am_RhLWA2 than in those expressing Am_RhLWA2 V211C mutant, showing biased sensitivity of the V211C mutant toward longer-wavelength light, consistent with their absorption spectra (Fig. 4C and D and Fig. S5). These results provide promising NIR-sensitive tools for controlling intracellular Ca²⁺ levels.

Fig. 4.

High performance of the red-shifted Am_RhLWA2 mutant in NIR-light-induced intracellular Ca²⁺ increases in cultured cells Aequorin Ca²⁺ assay in HEK293 cells expressing Am_RhLWA2 wild type (cyan) and the V211C mutant (yellow) and jumping spider rhodopsin (SpiRh1) (grey) in response to light. (A) Sensitivity curves for light-induced Ca²⁺ responses based on the peak responses to NIR light (738 nm) in different light intensities.(B) Statistical analysis of the Ca²⁺ responses to 738 nm light at 7.9 *10¹⁵ photons/cm²/sec (A), indicating that the V211C mutant exhibits 10-hold higher responsiveness to NIR light compared with the wild type, whereas SpiRh1 exhibits negligible response. (C) Light-induced Ca²⁺ responses to green light (536 nm). Arrowheads at time 0 indicate the timing of light stimulation. The colors correspond to those in (A). (D) Statistical analysis of the Ca²⁺ responses to 536 nm light (C), based on peak values, confirming the functionality of SpiRh1 in cultured cells. Peak values were normalized to those for Am_RhLWA2 wild type (B, D). Luminescence values were normalized to the baseline and then subtracting 1 from all values (“Rel. Luminescence”), and error bars indicate the means ± SE (n = 3) of the measured relative luminescence. Welch’s t-test was used to compare results between each opsin (*P < 0.05, **P < 0.01, ***P < 0.001) with Bonferroni adjustment

Discussions

Visible light varies depending on animals, and the long-wavelength limit of visible-light sensitivity is determined by red vision based on red opsins. Red vision is particularly important for animals because extension of the long-wavelength limit could give rise to unique visual ecologies, such as in foraging and mating. Here we succeeded in clarifying the molecular basis of red vision of dragonflies, which have been known to have one of the longest-wavelength sensitivities in the animal kingdom [2]. Our HAS and spectroscopic analysis of recombinant photopigments uncovered that RhLWA2 is red opsins (λ_max > 560 nm) of dragonflies, and particularly the RhLWA2 from A. melaenops (Am_RhLWA2) has the λ_max at 580 nm (Figs. 1 and 2). Considering that the absorption spectrum of the mantis shrimp NoL14 was reported to have λ_max at 611 nm [24], Am_RhLWA2 is the second longest-wavelength-sensitive animal opsin identified to date. On the other hand, Am_RhLWA2 is the only red opsin demonstrated to exhibit bistability (Fig. 2B), indicating that Am_RhLWA2 is the bistable red opsin having the longest-wavelength-sensitivity. Our mutational analysis of Am_RhLWA2 revealed that position 292 is responsible for the red shift, following the rule that Ser as in RhLWFs, Ala as in RhLWAs and Val as in Am_RhLWA2 at this position result in λ_max values of 546 nm, 562 nm and 580 nm, respectively (Fig. 2). Estimated λ_max values of RhLWA2s other than Am_RhLWA2 are ~ 560 nm, which is consistent with the previous report that λ_max of RhLWA2 of Sympetrum frequens is 557 nm [24], validating our conclusion based on mutagenesis. The spectral tuning through Ser/Ala substitutions at position 292 has been known in many vertebrate opsins including rhodopsins in many vertebrates as well as mammalian red opsins [14, 47–50]. The spectral tuning mechanism employing Ser/Ala at position 292 is well understood. Position 292 is located close to the retinylidene Schiff base, which is formed between retinal and Lys296. A hydrogen bond between Ser at position 292 and the Schiff base stabilizes the positive charge on the Schiff base nitrogen, and the substitution with Ala reduces the stabilization of the positive charge, contributing to the spectral tuning [14]. Ser/Ala substitutions at position 292 also account for 10–17-nm spectral shift in green sensitive Rh1 (λ_max = 480 nm) and Rh6 (λ_max = 514 nm) of Drosophila [51], demonstrating that the mechanism is conserved in invertebrate opsins as well. In the case of invertebrate red opsins, however, position 292 is not involved in the spectral shift in the butterfly red opsin PxRh3; instead, helix III, particularly positions 116 and 120, plays the crucial role [7]. In addition, in both the green-sensitive NoL6 and red-sensitive NoL14 of the mantis shrimp, position 292 is occupied by the same amino acid (Ser), suggesting that amino acid resides at other positions are involved in the red shift. Therefore, it is remarkable that Ser/Ala substitution at position 292 is predominantly responsible for the spectral shift in dragonfly LW opsins, as in mammalian red opsins, showing parallel evolution between the red/green opsins of insects and mammals. In addition, the Ala/Val substitution at the same position occurred in the lineage to A. melaenops leaded to further red shift of RhLWA2 (Fig. 2F). Taken together with the physicochemical properties of Ala and Val, the red shift is suggested to result from the substitution with a larger and more hydrophobic residue (Val), which could further reduce stabilization of the positive charge on the Schiff base nitrogen to destabilize the ground state of the retinal chromophore. Since almost all opsins have Ser or Ala at position 292, the effect of Val at position 292 had remained unknown. Therefore, current findings extend the potential of position 292 to induce a red shift, probably depending on hydrophobicity and/or volume of the amino acid residue. Furthermore, we attempted to artificially generate red shifted Am_RhLWA2 variants by assembling so-called red-shift residues from red opsins of other animals. Fortunately, we engineered Am_RhLWA2 to have 10-nm red-shifted absorption spectrum through V211C mutation derived from a candidate residue for the red shift of mantis shrimp NoL14 (Fig. 3C). In the spectral tuning of mammalian red and green opsins, substitutions from nonpolar to polar amino acid residues near the ß ionone ring of retinal, like positions 261 and 269 suggested to cause red shift [52]. Interestingly, although the amino acid residue at position 211 is not close to the ß ionone ring in bovine rhodopsin (1U19) [38], it is close to the ß ionone ring in jumping spider rhodopsin (6I9K) [37] (Fig. 5A and B). Therefore, the red shift through V211C mutation in Am_RhLWA2 could be explained by the proximity of position 211 in arthropod rhodopsins, as in mammalian red and green opsins. At the same time, the results strongly suggest the spectral tuning site of mantis shrimp red opsin.

Fig. 5.

Structural model for the spectral tuning in Am_RhLWA2 (A) Aligned 3D structure of bovine rhodopsin (BovRh) (PDB ID: 1U19) (magenta) and jumping spider rhodopsin (SpiRh1) (PDB ID: 6I9K) (cyan), focusing on retinal and amino acid residues at position 211. (B) Aligned 3D structure of SpiRh1 (cyan) and Am_RhLWA2 (red), focusing on retinal and amino acid residues at position 211. The protein structure of Am_RhLWA2 was predicted using ColabFold [36] with the default parameters

The red-shifted mutant of Am_RhLWA2 also has a significant impact on optogenetics. Optogenetic tools sensitive to longer-wavelength light, such as NIR, have been required because such light penetrates biological tissues more effectively, enabling efficient photoactivation of tools in deeper regions of the body [26–28]. In the case of microbial rhodopsins, Chrimson, the red-shifted channel rhodopsin having the λ_max at 590 nm from Chlamydomonas noctigama is used for optogenetic stimulation of neural activities with 660–735 nm light [53]. On the other hand, in the case of animal opsins, human red (OPN1LW) opsin-based tools were applied to induce intracellular cAMP changes in cultured cells with 544-nm-light stimulation; however, they did not produce statistically significant changes in cAMP levels [54]. We previously demonstrated that butterfly PxRh3-based tools can induce intracellular cAMP increases in cultured cells with 630-nm-light stimulation [7, 28]. In this paper, we successfully induced intracellular Ca²⁺ increases using the red-shifted mutant of Am_RhLWA2 with 738-nm-light stimuli, providing a NIR-sensitive GPCR optogenetic tool for Ca²⁺ regulation (Fig. 4). Recently, we and others demonstrated the utility and advantages of bistable opsins as optogenetic tools compared with bleaching opsins, such as vertebrate rhodopsins and human red opsin [33, 55–62]. Our data also showed that a bistable animal opsin mosquito Opn3 (MosOpn3) exhibited ~ 7,000 times higher sensitivity than channelrhodopsin-2 (ChR2), a light-gated cation channel in the light-induced behavior of C. elegans. Therefore, although a direct comparison was not performed, the red-shifted mutant of Am RhLWA2 could be more sensitive to NIR light than Crimson.

In this study, we used retinal (A1 retinal) for reconstituting photopigments, whereas both A1 retinal and 3-hydroxyretinal (A3 retinal) are used in compound eyes of dragonflies [63]. In general, photopigments containing A3-retinal exhibit slightly blue-shifted spectra compared with those containing A1-retinal [64], suggesting that red vision of dragonflies might have slightly shorter wavelength sensitivity than the wavelengths deduced from our analysis. Notably, however, most dragonfly species predominantly use A3 retinal, but members in the family Gomphidae, including A. melaenops, exceptionally prefer A1 retinal as a chromophore [63], enhancing the red sensitivity in Gomphidae dragonflies compared with other dragonfly species. Taken together with this evidence, our findings that among 12 dragonfly species tested, only A. melaenops possesses red-shifted RhLWA having λ_max at 580 nm (Fig. 2), prompted consideration of its physiological and ecological relevance. To address this issue, we focused on body coloration because dragonflies are particularly colorful, with body colors varying by species, ranging from blue to red, and in fact, the red wing pigment in the damselfly Mnesarete pudica plays a key role in sexual recognition, particularly courtship and territorial behaviors [65]. Interestingly, Gomphidae dragonflies including A. melaenops typically exhibit black body coloration with stripes in yellow, which is strongly absorbed by RhLWA2 (Fig. 6A). Then, we measured the reflectance spectra of the yellow area in the body coloration of another species of Gomphidae dragonfly Sieboldius albardae to investigate variations within the yellow coloration. As results, there is a marked difference in body coloration between females and males in the longer wavelength region (> ~ 530 nm) (Fig. 6B). In agreement with this, the difference in the longer wavelength region could be detected more efficiently by longer-wavelength-sensitive opsins. Therefore, the data theoretically suggest that the differences between females and males would be more clearly distinguished with 580-nm-red opsin in Gomphidae dragonflies compared to relatively shorter-wavelength-sensitive red opsins in other dragonflies (Fig. 1D), including 557-nm-red opsin in Sympetrum frequens [24], although to our knowledge, no research on the specific visual ecology regarding sex recognition of Gomphidae dragonflies has been reported to date. We confirmed that Sieboldius albardae also has RhLWA2 with Val at position 292 (Fig. 6C) and it is actually a 580-nm-red opsin (Fig. 6D), which supports the idea that red-shifted RhLWA2 is at least advantageous for distinguishing between females and males based on differences in body coloration in Gomphidae dragonflies. At this point, it is difficult to exclude other potential advantages of red-shifted RhLWA2 including color discrimination ability. Further investigations of visual ecology of Gomphidae dragonflies will be necessary to clarify the biological meaning of the red-shifted RhLWA2 in Gomphidae dragonflies.

Fig. 6.

Difference in thorax coloration between males and females, and absorption spectrum of RhLWA2 of the Gomphidae dragonfly, Sieboldius albardae (A) Images of representative thoraces of female (Upper) and male (Lower) Sieboldius albardae. For each sex, two individuals were shown to demonstrate the variation. Regions used for reflectance measurements were marked with white dotted circles. (B) Reflectance spectra of female (orange, n = 3 individuals) and male (green, n = 21 individuals) Sieboldius albardae ± SEs of the mean. Normalized absorption spectra with peaks at 560 nm (P560) and 580 nm (P580) are also shown as gray curves. (C) Comparison of residues around position 292 in RhLWA2 between Gomphidae dragonflies (red) and a species in another family (black). Amino acid residues at position 292 and the retinal-binding K296 are highlighted. Residues are numbered according to bovine rhodopsin. TM7, seventh transmembrane region. (D) Absorption spectra of the Sa_RhLWA2 (cyan) and the Am_RhLWA2 (grey), showing that both are 580-nm-red opsins

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

GPCRs

G protein-coupled receptors

LW

long-wavelength-sensitive

NIR

near-infrared

HAS

heterologous action spectroscopy

Authors’ contributions

M.K. conceived and designed the study. R.S. and M.K. carried out the experiments. R.S., A.T. and M.K. analyzed the data. R.S. and M.K. wrote and edited the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by Ministry of Education, Culture, Sports, Science and Technology (MEXT) KAKENHI Grant Numbers JP22H02663 (to M.K.), JP23H02516 (to A.T.) and JP23KJ1845 (to R.S.), Ohsumi Frontier Science Foundation (to M.K.), Japan Science and Technology Agency (JST) Precursory Research for Embryonic Science and Technology (PRESTO) grant JPMJPR13A2 (to M.K.). and Core Research for Evolutional Science and Technology (CREST) grant JPMJCR25B5 (to M.K.) R.S. was supported by Grant-in-Aid for JSPS Fellows.

Data availability

All data are available in the main text or the supplementary materials.

Declarations

Competing Interests

Authors declare that they have no competing interests.