Adaptation of cone pigments found in green rods forscotopic vision through a single amino acid mutationKeiichi Kojimaa,1, Yuki Matsutania,1, Takahiro Yamashitaa, Masataka Yanagawab, Yasushi Imamotoa, Yumiko Yamanoc,Akimori Wadac, Osamu Hisatomid, Kanto Nishikawae, Keisuke Sakuraif, and Yoshinori Shichidaa,g,2

aDepartment of Biophysics, Graduate School of Science, Kyoto University, Kyoto 606-8502, Japan; bCellular Informatics Laboratory, RIKEN, Wako 351-0198,Japan; cDepartment of Organic Chemistry for Life Science, Kobe Pharmaceutical University, Kobe 658-8558, Japan; dDepartment of Earth and Space Science,Graduate School of Science, Osaka University, Osaka 560-0043, Japan; eGraduate School of Human and Environmental Studies, Kyoto University, Kyoto606-8501, Japan; fFaculty of Life and Environmental Sciences, University of Tsukuba, Tsukuba 305-8577, Japan; and gResearch Organization for Science andTechnology, Ritsumeikan University, Kusatsu 525-8577, Japan

Edited by King-Wai Yau, Johns Hopkins University School of Medicine, Baltimore, MD, and approved April 17, 2017 (received for review December 5, 2016)

Most vertebrate retinas contain a single type of rod for scotopicvision and multiple types of cones for photopic and color vision.The retinas of certain amphibian species uniquely contain twotypes of rods: red rods, which express rhodopsin, and green rods,which express a blue-sensitive cone pigment (M1/SWS2 group).Spontaneous activation of rhodopsin induced by thermal isomer-ization of the retinal chromophore has been suggested to con-tribute to the rod’s background noise, which limits the visualthreshold for scotopic vision. Therefore, rhodopsin must exhibitlow thermal isomerization rate compared with cone visual pig-ments to adapt to scotopic condition. In this study, we determinedwhether amphibian blue-sensitive cone pigments in green rodsexhibit low thermal isomerization rates to act as rhodopsin-likepigments for scotopic vision. Anura blue-sensitive cone pigmentsexhibit low thermal isomerization rates similar to rhodopsin,whereas Urodela pigments exhibit high rates like other vertebratecone pigments present in cones. Furthermore, by mutational anal-ysis, we identified a key amino acid residue, Thr47, that is respon-sible for the low thermal isomerization rates of Anura blue-sensitivecone pigments. These results strongly suggest that, through thismutation, anurans acquired special blue-sensitive cone pigmentsin their green rods, which could form the molecular basis for sco-topic color vision with normal red rods containing green-sensitiverhodopsin.

photoreceptor cell | visual pigment | amphibian | molecular evolution |color discrimination

Vertebrate vision consists of scotopic and photopic vision,triggered, respectively, by light absorption by the rod and

cone photoreceptor cells of the retina (1). Scotopic vision re-quires high sensitivity and low threshold to be able to detect afew photons (2–4). Thus, electrical signals generated by single-photon absorptions in rods need to be reliably transmitted tohigher-order retinal neurons in the presence of overwhelmingintrinsic noise in rods, which is composed of two components:discrete and continuous noise (5). The former originates fromthermal activation of the visual pigment, rhodopsin, and thelatter results from spontaneous activation of the phosphodies-terase in the phototransduction cascade. Although the continuousnoise can be separated from the background noise, the discretenoise events are indistinguishable from true single-photon re-sponses (4, 6). Thus, the discrete noise sets a limit for the ab-solute visual threshold in scotopic vision (4, 7, 8). Thermalactivation of rhodopsin originates from the thermal isomeriza-tion of the retinal chromophore (9, 10). Taken together, sup-pressing the thermal isomerization rates of the retinal inrhodopsin improves the threshold of light detection.Most vertebrates have one type of rod containing rhodopsin

and multiple types of cones containing different cone visualpigments. Thus, they have the ability to discriminate color onlyunder photopic conditions. By contrast, amphibians have diver-sified photoreceptor systems in their retinas. Most species in the

order Gymnophiona are fossorial and have small eyes coveredwith skin, which express only rhodopsin without cone pigments(11). Moreover, amphibian species belonging to Anura and sev-eral genera in Urodela, including Hynobius and Ambystoma, butneither Cynops nor Salamandra, are unique in possessing twotypes of rods, red and green rods. Red rods are the normal rodsthat express rhodopsin, whereas green rods express a blue-sensitive cone pigment (12–16). Thus, our goal was to deter-mine whether the blue-sensitive cone pigments in green rods ex-hibit molecular properties similar to that of rhodopsin or to thatof a typical cone pigment in cones. In this study, we compared thethermal isomerization rates of blue-sensitive cone pigments fromseveral amphibian species with those of rhodopsin and other conepigments. Our biochemical analysis revealed that Anura blue-sensitive cone pigments (Anura blues) acquired low thermalisomerization rates similar to rhodopsin through a single aminoacid mutation. We discuss the specialization process of blue-sensitive cone pigments in anuran green rods for scotopic vision.

Results and DiscussionComparison of Thermal Isomerization Rates of Amphibian Blue-Sensitive Cone Pigments. The thermal isomerization rates of theretinal chromophore of visual pigments have been extensivelymeasured by electrophysiological analyses of photoreceptor

Significance

Anurans are unique in possessing two types of rod photore-ceptor cells, red and green rods. Red rods express rhodopsin,whereas green rods express blue-sensitive cone visual pig-ment. Rhodopsin exhibits a low rate of thermal isomerizationof the retinal chromophore, which enables rods to detectphotons with extremely high signal-to-noise for scotopic vi-sion. Here, we show that anuran blue-sensitive cone pigmentsacquired a rhodopsin-like property through a single amino acidmutation at position 47 in the evolutionary process from othercone pigments. Thus, anurans have special blue-sensitive conepigments for the contribution of green rods to the low thresholdof light detection, which could form the molecular basis in tan-dem with red rods containing rhodopsin in scotopic color vision.

Author contributions: K.K., Y.M., T.Y., and Y.S. designed research; K.K. and Y.M. per-formed research; M.Y., Y.I., Y.Y., A.W., O.H., K.N., and K.S. contributed new reagents/analytic tools; K.K., Y.M., T.Y., and Y.S. analyzed data; and K.K., T.Y., and Y.S. wrotethe paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Data deposition: The sequences reported in this paper have been deposited in the GenBankdatabase (accession nos. LC180360, LC180361, and LC180362).1K.K. and Y.M. contributed equally to this work.2To whom correspondence should be addressed. Email: shichida@rh.biophys.kyoto-u.ac.jp.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1620010114/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1620010114 PNAS | May 23, 2017 | vol. 114 | no. 21 | 5437–5442

BIOCH

EMISTR

Y

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

020

cells from a series of genetically modified animals (17–19).Recently, to better understand the molecular mechanism thatregulates the retinal thermal isomerization, we developed abiochemical method to estimate thermal isomerization rates byusing recombinant visual pigment proteins (10). Thermal isom-erization rates (kth) are calculated using three experimentallydetermined values in the following manner. First, we assumed asimplified two-step reaction as shown in Fig. 1, where R is thepigment in the inactive state and R* is the activated pigment bythe thermal isomerization of the retinal chromophore. Giventhat the thermal isomerization occurs much slower than thedecay of R* (kth << kd), a steady-state approximation can beapplied to the concentration of R*. (Here, kd is the spontaneousdecay rate of the activated pigment.) Therefore, we obtain thefollowing:

�R*

�=kthkd

½R�0, [1]

where [R]0 is the initial concentration of the visual pigment.Namely, it can be considered that a small but constant concen-tration of R* exists in a solution of the purified pigments in thedark. Because R* is essentially the same as the light-inducedMeta II state (20), the initial rate of G protein activation by apigment in the dark (vdark) can be approximated by

vdark =kthkd

vlight, [2]

where vlight is the initial rate of G protein activation by an acti-vated pigment. Therefore, we can estimate the kth from threeexperimentally determined values (vdark, vlight, and kd) as

kth =vdarkvlight

kd. [3]

By using this biochemical method, we have previously ana-lyzed the thermal isomerization rate of Xenopus tropicalis blue-sensitive cone pigment (Xenopus blue) (10). The measuredthermal isomerization rate was about 100-fold lower than ratesof normal cone pigments in mouse and chicken green-sensitivecones. Moreover, the measured rate was closer to that of rho-dopsin. This finding is consistent with electrophysiological evi-dence that green rods in a toad, Rhinella marina, formerly Bufomarinus, exhibit low discrete noise like normal rods (21, 22). Toexamine whether Xenopus blue is a special case exhibiting lowthermal isomerization rate, we determined the thermal isomer-ization rates of the blue-sensitive cone pigments of other anuranspecies: the American bullfrog, Lithobates catesbeianus, formerlyRana catesbeiana, and the mantelline frog, Mantella baroni (Figs.S1−S3 and S4A). Our results show that these pigments also exhibitrhodopsin-like thermal isomerization rates (Fig. 2A). Therefore, weconcluded that Anura blues changed their properties to mimic

that of rhodopsin in the course of molecular evolution afteranurans had acquired green rods.Some urodelans also have green rods in their retina that

contain blue-sensitive cone pigments. Therefore, we determinedthe thermal isomerization rates of blue-sensitive cone pigmentsfrom three urodelan species (Figs. S1−S3 and S4B). The tigersalamander (Ambystoma tigrinum) was reported to have greenrods that express blue-sensitive cone pigments (14). The Mexicansalamander (Ambystoma mexicanum) is a closely related speciesto the tiger salamander (23). The Japanese fire-bellied newt(Cynops pyrrhogaster), however, has no green rods, and its blue-sensitive cone pigments are expressed only in blue cones (12).Our results show that the tiger salamander and Mexican sala-mander blue-sensitive cone pigments (tiger salamander blue andMexican salamander blue, respectively) exhibit thermal isomer-ization rates about 50-fold higher than that of bovine rhodopsin,whereas Japanese fire-bellied newt blue-sensitive cone pigmentexhibits a rate about 300-fold higher than that of bovine rho-dopsin (Fig. 2B). Although the rates of the former two speciesare sixfold lower than that of the latter, these rates are similar tothose of typical cone pigments but considerably different from

Fig. 1. Two-step reaction scheme of thermal activation and deactivation of visual pigments. R and R* indicate visual pigments in the inactive and activestates, respectively. An opsin regenerated with normal 11-cis-retinal spontaneously converts to R* by thermal isomerization of retinal in the dark. After thefirst reaction, R* is degraded into opsin and retinal.

Fig. 2. Comparison of the thermal isomerization rates (kth) of wild-typevisual pigments. (A) Comparison of the thermal isomerization rates (kth) ofbovine and Xenopus rhodopsin and Xenopus, American bullfrog, man-telline frog, and zebrafish blue estimated from the data presented in Fig.S4A. (B) Comparison of the thermal isomerization rates (kth) of bovine andXenopus rhodopsin, tiger salamander, Mexican salamander, newt, andzebrafish blue estimated from the data in Fig. S4B. All error bars in Fig. 2represent the SEM of more than three independent measurements. Anasterisk (*) indicates a significant difference in relative rate constants be-tween visual pigments and Xenopus rhodopsin (P < 0.05; Student’s t test,two-tailed).

5438 | www.pnas.org/cgi/doi/10.1073/pnas.1620010114 Kojima et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

020

that of rhodopsin. This finding is consistent with electrophysio-logical evidence that blue cones in tiger salamander exhibit higherdiscrete noise than rods although a rod contains ∼30 times asmany pigment molecules as a cone (24). Therefore, unlike an-urans, urodelans express typical blue-sensitive cone pigmentirrespective of its presence or absence in green rods.

Identification of a Key Amino Acid Residue Responsible for the LowThermal Isomerization Rates of Anura Blue-Sensitive Cone Pigments.To investigate the mechanism underlying the low thermal isom-erization rates of Anura blues, we sought to identify the aminoacid residue(s) responsible for the low thermal isomerizationrates. We previously identified two amino acid residues, at po-sitions 122 and 189, which are responsible for the low thermalisomerization rate of rhodopsin (10). However, Anura blues con-tain amino acid residues common in cone pigments at positions

122 and 189 (Fig. S5). Thus, our working hypothesis is thatamino acid residue(s) at position(s) other than 122 or 189 may beresponsible for the low thermal isomerization rates of Anurablues. We compared amino acid residues located within 10 Å ofthe retinal chromophore in the crystal structure of bovine rho-dopsin between Anura and other animals’ blue-sensitive conepigments and found that threonine at position 47 is conservedonly in Anura blues (25) (Fig. 3 A and B and Fig. S5). Mutationof this threonine in Xenopus blue increased the thermal isom-erization rate about 60-fold (Fig. 3C), which is comparable tothat of zebrafish blue. Therefore, the threonine at position47 appears to be a determinant of the low thermal isomerizationrates of Anura blues.To study whether or not introduction of threonine at position

47 causes a decrease of thermal isomerization rate of tiger sal-amander blue, we prepared an M47T mutant of tiger salamander

Fig. 3. A key amino acid residue regulates the thermal isomerization rates (kth) of amphibian blue-sensitive cone pigments. (A) Comparison of amino acidresidue at position 47. The residues of Xenopus, American bullfrog, tiger salamander, newt, and zebrafish blue and bovine rhodopsin are shown. (B) Ho-mology model of Xenopus blue that was constructed based on the crystal structure of the dark state of bovine rhodopsin (PDB 1U19). Thr47, Lys296, and 11-cis-retinal are colored blue, black, and red, respectively. (C) Comparison of the thermal isomerization rates (kth) of bovine rhodopsin, wild-type, and T47Lmutant of Xenopus blue and zebrafish blue estimated from the data in Fig. S4C. (D) Comparison of the thermal isomerization rates (kth) of bovine rhodopsin,wild-type, and M47T mutant of tiger salamander blue estimated from the data in Fig. S4C. All error bars in Fig. 3 represent the SEM of more than threeindependent measurements. An asterisk (*) indicates a significant difference in relative rate constants between visual pigments and Xenopus blue in C or tigersalamander blue in D (P < 0.05; Student’s t test, two-tailed).

Kojima et al. PNAS | May 23, 2017 | vol. 114 | no. 21 | 5439

BIOCH

EMISTR

Y

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

020

blue and determined its thermal isomerization rate. This muta-tion decreases the thermal isomerization rate by only a quarter(Fig. 3D). This finding suggests that the introduction of threo-nine at position 47 together with other mutations in Urodelablue-sensitive cone pigments can suppress the thermal isomeri-zation rates to the same level as Anura blues.

Mechanistic Implication of the Low Thermal Isomerization Rates.Previous electrophysiological and theoretical studies showed thatthe thermal isomerization rates can be expressed as

kth  =A× e−EaRT

Xm1

1ðm− 1Þ!

�Ea

RT

�m−1

based on the Hinshelwood distribution (R is the gas constant,T is absolute temperature, and m is the number of molecularvibrational modes contributing thermal energy to pigment acti-vation), where Ea is inversely proportional to the pigment’s ab-sorption maximum (λmax), and the preexponential factor, A, isindependent of the pigment’s λmax but dependent on the chro-mophore binding site structure (open vs. closed) (22, 26). Toassess the molecular mechanism about the regulation of kth, wepreviously determined that there is a strong relationship betweenthermal isomerization rates of the retinal chromophore in thedark state and decay rates of the active state (kd) in visual pig-ments and their mutants (10). These results indicated that thetwo rates are regulated by a common molecular feature suchas flexibility of the protein moiety. Therefore, we proposed amolecular model in which the flexibility of the chromophorebinding site is responsible for the thermal isomerization ratesof the retinal chromophore and decay rates of the active state(10, 27, 28). Interestingly, we found that the results we obtainedin this study can be plotted on a correlation line as before (Fig.4). In particular, the mutations at position 47 caused little spec-tral shift in Xenopus and tiger salamander blues (Fig. S6) butsignificantly changed the thermal isomerization rates, probablyby interacting with Lys296 (Fig. 3B). This suggests that the ac-quisition of threonine at position 47 in Anura blues decreased Ato suppress their thermal isomerization rates. This effect is com-monly observed between rhodopsin and green-sensitive cone pig-ments (i.e., A of rhodopsin is 2 orders of magnitude lower thanthose of green-sensitive cone pigments) (10). Taken together,this structural flexibility could also be an important factor inregulating the difference in A.A recent study suggests that the ability to exchange of 11-cis-

retinal in a pigment with 9-cis-retinal in solution in the dark canbe explained by the differences between the structures of thechromophore binding site (open or closed) (29). We assessed thedegree of retinal exchange in rhodopsin, zebrafish blue, andXenopus blue (Fig. S7). As shown previously (29), we observedthe spectral shift induced by the chromophore exchange inzebrafish blue, but not in rhodopsin. However, we could alsodetect the spectral shift in wild-type and T47L mutant of Xen-opus blue. This finding suggests that the low thermal isomeri-zation rate of the retinal in Xenopus blue cannot be explainedsimply by a more open or closed chromophore binding pocketthat would be predicted by the ability to exchange retinal.

Diversity of the Molecular Properties of Blue-Sensitive Cone Pigmentsin Amphibians. Green rods are uniquely observed in amphibianspecies belonging to Anura and several genera in Urodela, in-cluding Hynobius and Ambystoma, but not Cynops nor Salaman-dra. Considering the phylogenetic relationship of amphibians (23,30), it has been proposed that green rods emerged before thedivergence of Anura and Urodela and were lost from severalgenera of Urodela, including Cynops and Salamandra (12). Ouranalysis of the thermal isomerization rates of blue-sensitive cone

pigments suggests that, after the divergence of Anura and Uro-dela, Anura blues specifically acquired threonine at position47 to suppress thermal isomerization rates (Fig. 5). Electro-physiological evidence from frogs shows that green rods trans-duce signals to most retinal ganglion cells, some of which canalso receive inputs from red rods (31–34). Thus, the decrease ofthe thermal isomerization rates of anuran blue-sensitive conepigments in green rods enables them to function in tandem withgreen-sensitive red (normal) rods for scotopic vision, whichcould provide a molecular basis for color discrimination underthe dark conditions. This theory is consistent with behavioralevidence that indicates that frogs can distinguish the colors blueand gray at low illumination levels, probably using green rods, inaddition to color vision in photopic condition (35–37). Scotopiccolor vision may be advantageous for nocturnality that is ob-served in many anuran species (such as X. tropicalis and theAmerican bullfrog). On the other hand, a few anuran species(such as the mantelline frog) are exceptionally diurnal. They arethought to have changed their behavioral rhythms in severalphylogenetic branches after the divergence of nocturnal species(23). Our analysis of mantelline frog blue-sensitive cone pigmentshowed that the pigment maintains the low thermal isomeriza-tion rate. Therefore, diurnal frogs may have adapted to photopicconditions using a cone photoreceptor system containing red-and violet-sensitive cone pigments without changing the molec-ular properties of the blue-sensitive cone pigment.

ConclusionAnuran species acquired the low thermal isomerization ratesof the retinal chromophore of blue-sensitive cone pigmentsexpressed in green rods by a single amino acid replacement atposition 47 in the evolutionary process. We previously deter-mined that the low thermal isomerization rate of rhodopsin wasachieved by two amino acid replacements at positions 122 and189 (10). Therefore, we propose that the contribution of normal

Fig. 4. Correlation between the thermal isomerization rates (kth) and kd ofthe visual pigments and their mutants. We plotted lnkth and lnkd of thevisual pigments. Red circles indicate data from this study, and black circlesindicate data from our previous study (10). Error bars represent the SEM ofmore than three independent measurements. The regression line derivedfrom the data in this study (red circles) is shown by a red dashed line; r = 0.96(P < 0.05).

5440 | www.pnas.org/cgi/doi/10.1073/pnas.1620010114 Kojima et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

020

rods and anuran green rods for scotopic vision results from theconvergent evolution of the visual pigments to suppress thethermal isomerization rates of retinal chromophore.

Materials and MethodsAnimals. X. tropicalis were obtained from The Institute for Amphibian Biol-ogy (Hiroshima University). Mantelline frog (M. baroni), tiger salamander(A. tigrinum) and Mexican salamander (A. mexicanum) were obtainedfrom local pet shops. The use of animals was in accordance with guidelinesestablished by the Ministry of Education, Culture, Sports, Science, andTechnology, Japan. The Amphibians of the World database of amphibianspecies (38) was used for the most current taxonomy of amphibians.

Heterologous Expression and Purification of the Visual Pigments. The cDNAs ofX. tropicalis rhodopsin (NM_001097334) and blue-sensitive cone pigment(XM_002937226) were isolated by PCR from the first strand cDNA from eyes.The cDNA of tiger salamander (A. tigrinum) blue-sensitive cone pigment wasalso isolated from the first strand cDNA from eyes based on the sequenceinformation (AF038946) from the National Center for Biotechnology In-formation (NCBI) database. The ORF sequence of our clone had severaldifferent amino acid residues, including position 47, reported previously(AF038946). Thus, we deposited the sequence of our clone in the NCBI da-tabase under accession number LC180360. Mexican salamander (A. mexicanum)blue-sensitive cone pigment (LC180361) was cloned by PCR from the firststrand cDNA from eyes based on the homology to the sequences of tigersalamander and Japanese fire-bellied newt (C. pyrrhogaster) blue-sensitivecone pigments. Primer sequences are as follows: 5′-GCTTTGGAGTCAGACC-GGAG-3′ (forward) and 5′-GGTTTGTGTGAGGGTGCCCTA-3′ (reverse). Theisolated clone contained the full-length ORF sequence. To isolate the cloneof mantelline frog (M. baroni) blue-sensitive cone pigment, we firstobtained the sequences of exons 1 and 5 (including start and stop codons ofORF) by PCR from genomic DNA based on the homology to the sequences ofother frog blue-sensitive cone pigments. Primer sequences are as follows: 5′-GGGCCCAGGCCAGACCTTTTC-3′ (forward) and 5′-ACCAGGATGTAGTTAAG-GTGGG-3′ (reverse) for exon 1 and 5′-ATGATGAAGCTGATCTTCTGTGG-3′(forward) and 5′-GCGTGGCAGAACAGGAGCTG-3′ (reverse) for exon 5. Afteridentification of start and stop codons, the full-length ORF sequence ofmantelline frog blue-sensitive cone pigment (LC180362) was obtained byPCR from the first strand cDNA from eyes. Primer sequences are as follows:5′-ATGAGCAAAGGTCGACCAGA-3′ (forward) and 5′-TTATGCAGGAGTCACC-TGGCTGC-3′ (reverse). American bullfrog (L. catesbeianus) blue-sensitivecone pigment (AB010085) and Japanese fire-bellied newt (C. pyrrhogaster)blue-sensitive cone pigment (AB040148) were isolated as described pre-viously (12, 13). The cDNAs containing a point mutation were constructedusing an In-Fusion cloning kit (Clontech) according to the manufacturer’sinstructions. The cDNA of bovine rhodopsin (K00506) was inserted into themammalian expression vector, pUSRα. The cDNAs of other visual pigmentswere tagged with the epitope sequence of the anti-bovine rhodopsinmonoclonal antibody Rho1D4 at the C terminus and inserted into themammalian expression vector pMT4 (39). Expression of the visual pigments

in HEK293 cells and sample preparation of the visual pigments for themeasurements of vdark, vlight, and kd were performed as previously de-scribed (10). The cell membranes expressing the visual pigments were di-vided into two aliquots. One was regenerated with 11-cis-retinal and7-membered-ring 11-cis-retinal (7mr) (10, 40), and the other was regen-erated by only 7mr. After regeneration, they were solubilized with Buffer A(50 mM Hepes, 140 mM NaCl, pH 6.5) containing 1% dodecyl maltoside(DDM) and purified using Rho1D4-conjugated agarose. The purified visualpigments were eluted with 0.02% DDM in Buffer A containing the syntheticC-terminal peptide of bovine rhodopsin. All of the experiments after re-constitution of the visual pigments with 11-cis-retinal were performed underinfrared light. We refer to the purified samples regenerated by both 11-cis-retinal and 7mr, or only 7mr, as “pigment name-n” or “pigment name-7mr,”respectively. We verified that the concentrations of the visual pigmentscontained in the two samples were similar by Western blot analysis (Fig. S1).The samples of the visual pigments for the analysis of the chromophore-exchange reaction were regenerated with 11-cis-retinal or 9-cis-retinal andpurified as described above.

Measurement of vdark, vlight, and kd. The vdark was measured by a [35S]GTPγSbinding assay in the dark under infrared light as previously described (10).The assay mixture consisted of 300 nM pigments, 1 μM transducin (Gt), 5 μMGTPγS, 25 nM [35S]GTPγS, 0.015% DDM, 50 mM Hepes (pH 6.5), 140 mMNaCl, 5.8 mM MgCl2, and 1mM DTT. Experimental data were fitted by asingle exponential function, and the vdark was estimated by the differencebetween the initial rates between two samples (‘pigment name-n’ and ‘pigmentname-7mr’). The vlight was measured by fluorescence assay as previouslydescribed (41, 42). The assay mixture consisted of 20 nM pigments, 0 or 1 μMGt, 5 μM GTPγS, 0.015% DDM, 50 mM Hepes (pH 6.5), 140 mM NaCl, 5.8 mMMgCl2, and 1 mM DTT. Experimental data were fitted by a single exponentialfunction, and the vlight was estimated as previously described (41, 42). The kdwas measured by a fluorescence assay as previously described (42, 43).The assay mixture consisted of 20 nM pigments (60 nM for zebrafish, tigersalamander, Mexican salamander and newt blue, Xenopus blue T47L, andtiger salamander blue M47T to increase signal to noise ratio), 5 μM GTPγS,0.015% DDM, 50 mM Hepes (pH 6.5), 140 mM NaCl, 5.8 mM MgCl2, and1 mM DTT. Experimental data were fitted by a single exponential functionto estimate kd.

Western Blotting. Extracts from visual pigment-transfected HEK293 cells weresubjected to SDS/PAGE, transferred onto a polyvinylidene difluoride mem-brane, and probed with Rho1D4. Immunoreactive proteins were detected byECL (GE Healthcare) and visualized by a luminescent image analyzer (LAS4000mini; GE. Healthcare).

Spectroscopic Measurements.Absorption spectra of the samples were recordedwith a UV–visible spectrophotometer (Shimadzu UV-2450, UV-2400). Sampleswere kept at 0 °C using a cell holder equipped with a temperature-controlledcirculating water bath.

Fig. 5. Evolutionary relationship of the molecular properties of blue-sensitive cone pigments.

Kojima et al. PNAS | May 23, 2017 | vol. 114 | no. 21 | 5441

BIOCH

EMISTR

Y

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

020

Retinal Exchange Reaction. The samples of visual pigments purified afterreconstitution with 11-cis-retinal were incubated with a 10-fold molar excessof 9-cis-retinal in the dark for 15 h at 20 °C. Then, after the addition ofneutralized hydroxylamine (final concentration: 5 mM in blue-sensitive conepigments or 50 mM in mouse rhodopsin), the samples were incubated in thedark for 5 h. Absorption spectra were recorded before and after yellow light(>410-nm light in blue-sensitive cone pigments or >500-nm light in rhodopsin)irradiation.

ACKNOWLEDGMENTS. We thank Dr. M. Kono for a critical reading of themanuscript. We also thank Prof. R. S. Molday for the generous gift of a Rho1D4-producing hybridoma and The Institute for Amphibian Biology (HiroshimaUniversity) for X. tropicalis through the National Bio-Resource Project of theMinistry of Education, Culture, Sports, Science and Technology of Japan (MEXT).This work was supported in part by MEXT Grants-in Aid for Scientific Research26650119 and 16H02515 (to Y.S.) and 15H00812 (to T.Y.), and a grant from theTakeda Science Foundation (to T.Y.). K.K. was supported by a Japan Society forthe Promotion of Science Research Fellowship for Young Scientists (15J02054).

1. Fu Y, Yau KW (2007) Phototransduction in mouse rods and cones. Pflugers Arch 454:805–819.

2. Field GD, Sampath AP, Rieke F (2005) Retinal processing near absolute threshold:From behavior to mechanism. Annu Rev Physiol 67:491–514.

3. Hecht S, Shlaer S, Pirenne MH (1942) Energy, quanta, and vision. J Gen Physiol 25:819–840.

4. Pahlberg J, Sampath AP (2011) Visual threshold is set by linear and nonlinear mech-anisms in the retina that mitigate noise: How neural circuits in the retina improve thesignal-to-noise ratio of the single-photon response. BioEssays 33:438–447.

5. Baylor DA, Matthews G, Yau KW (1980) Two components of electrical dark noise intoad retinal rod outer segments. J Physiol 309:591–621.

6. Field GD, Rieke F (2002) Nonlinear signal transfer frommouse rods to bipolar cells andimplications for visual sensitivity. Neuron 34:773–785.

7. Aho AC, Donner K, Hydén C, Larsen LO, Reuter T (1988) Low retinal noise in animalswith low body temperature allows high visual sensitivity. Nature 334:348–350.

8. Naarendorp F, et al. (2010) Dark light, rod saturation, and the absolute and in-cremental sensitivity of mouse cone vision. J Neurosci 30:12495–12507.

9. Ala-Laurila P, Donner K, Crouch RK, Cornwall MC (2007) Chromophore switch from11-cis-dehydroretinal (A2) to 11-cis-retinal (A1) decreases dark noise in salamanderred rods. J Physiol 585:57–74.

10. Yanagawa M, et al. (2015) Origin of the low thermal isomerization rate of rhodopsinchromophore. Sci Rep 5:11081.

11. Mohun SM, et al. (2010) Identification and characterization of visual pigments incaecilians (Amphibia: Gymnophiona), an order of limbless vertebrates with rudi-mentary eyes. J Exp Biol 213:3586–3592.

12. Takahashi Y, Hisatomi O, Sakakibara S, Tokunaga F, Tsukahara Y (2001) Distributionof blue-sensitive photoreceptors in amphibian retinas. FEBS Lett 501:151–155.

13. Hisatomi O, Takahashi Y, Taniguchi Y, Tsukahara Y, Tokunaga F (1999) Primary struc-ture of a visual pigment in bullfrog green rods. FEBS Lett 447:44–48.

14. Ma J, et al. (2001) A visual pigment expressed in both rod and cone photoreceptors.Neuron 32:451–461.

15. Nilsson SE (1964) Receptor cell outer segment development and ultrastructure of thedisk membranes in the retina of the tadpole (Rana Pipiens). J Ultrastruct Res 11:581–602.

16. Sherry DM, Bui DD, Degrip WJ (1998) Identification and distribution of photoreceptorsubtypes in the neotenic tiger salamander retina. Vis Neurosci 15:1175–1187.

17. Sakurai K, et al. (2007) Physiological properties of rod photoreceptor cells in green-sensitive cone pigment knock-in mice. J Gen Physiol 130:21–40.

18. Fu Y, Kefalov V, Luo DG, Xue T, Yau KW (2008) Quantal noise from human red conepigment. Nat Neurosci 11:565–571.

19. Kefalov V, Fu Y, Marsh-Armstrong N, Yau KW (2003) Role of visual pigment prop-erties in rod and cone phototransduction. Nature 425:526–531.

20. Yau KW, Matthews G, Baylor DA (1979) Thermal activation of the visual transductionmechanism in retinal rods. Nature 279:806–807.

21. Matthews G (1984) Dark noise in the outer segment membrane current of green rodphotoreceptors from toad retina. J Physiol 349:607–618.

22. Luo DG, Yue WW, Ala-Laurila P, Yau KW (2011) Activation of visual pigments by lightand heat. Science 332:1307–1312.

23. Pyron RA, Wiens JJ (2011) A large-scale phylogeny of Amphibia including over2800 species, and a revised classification of extant frogs, salamanders, and caecilians.Mol Phylogenet Evol 61:543–583.

24. Rieke F, Baylor DA (2000) Origin and functional impact of dark noise in retinal cones.Neuron 26:181–186.

25. Okada T, et al. (2004) The retinal conformation and its environment in rhodopsin inlight of a new 2.2 A crystal structure. J Mol Biol 342:571–583.

26. Ala-Laurila P, Donner K, Koskelainen A (2004) Thermal activation and photo-activation of visual pigments. Biophys J 86:3653–3662.

27. Piechnick R, et al. (2012) Effect of channel mutations on the uptake and release of theretinal ligand in opsin. Proc Natl Acad Sci USA 109:5247–5252.

28. Lórenz-Fonfría VA, Furutani Y, Ota T, Ido K, Kandori H (2010) Protein fluctuations asthe possible origin of the thermal activation of rod photoreceptors in the dark. J AmChem Soc 132:5693–5703.

29. Yue WW, et al. (2017) Spontaneous activation of visual pigments in relation toopenness/closedness of chromophore-binding pocket. eLife 6:6.

30. Anderson JS, Reisz RR, Scott D, Fröbisch NB, Sumida SS (2008) A stem batrachian fromthe Early Permian of Texas and the origin of frogs and salamanders. Nature 453:515–518.

31. Reuter T, Virtanen K (1972) Border and colour coding in the retina of the frog. Nature239:260–263.

32. Reuter T, Virtanen K (1976) Color discrimination mechanisms in retina of toad (Bufo-Bufo). J Comp Physiol 109:337–343.

33. Kicliter E, Kay CJ, Chino YM (1981) Spectral opponency of on-type ganglion cells andthe blue preference of Rana pipiens. Brain Res 210:103–113.

34. Bäckström AC, Reuter T (1975) Receptive field organization of ganglion cells in thefrog retina: Contributions from cones, green rods and red rods. J Physiol 246:79–107.

35. Gniubkin VF, Kondrashev SL, Orlov OIu (1975) [Constancy of color perception in thegrey toad]. Biofizika 20:725–730. Russian.

36. Govardovskii VI, Reuter T (2014) Why do green rods of frog and toad retinas lookgreen? J Comp Physiol A Neuroethol Sens Neural Behav Physiol 200:823–835.

37. Maximov VV, Orlov OY, Reuter T (1985) Chromatic properties of the retinal afferentsin the thalamus and the tectum of the frog (Rana temporaria). Vision Res 25:1037–1049.

38. Frost DR (2016) Amphibian species of the world: An online reference, version 6.0.Available at research.amnh.org/vz/herpetology/amphibia/index.php. Accessed Sep-tember 9, 2016.

39. Oprian DD, Molday RS, Kaufman RJ, Khorana HG (1987) Expression of a syntheticbovine rhodopsin gene in monkey kidney cells. Proc Natl Acad Sci USA 84:8874–8878.

40. Akita H, Tanis SP, Adams M, Baloghnair V, Nakanishi K (1980) Non-bleachable rho-dopsins retaining the full natural chromophore. J Am Chem Soc 102:6370–6372.

41. Kojima K, Imamoto Y, Maeda R, Yamashita T, Shichida Y (2014) Rod visual pigmentoptimizes active state to achieve efficient G protein activation as compared with conevisual pigments. J Biol Chem 289:5061–5073.

42. Imamoto Y, Seki I, Yamashita T, Shichida Y (2013) Efficiencies of activation of trans-ducin by cone and rod visual pigments. Biochemistry 52:3010–3018.

43. Matsuyama T, Yamashita T, Imai H, Shichida Y (2010) Covalent bond between ligandand receptor required for efficient activation in rhodopsin. J Biol Chem 285:8114–8121.

5442 | www.pnas.org/cgi/doi/10.1073/pnas.1620010114 Kojima et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

020