Molecular Properties of New Enzyme Rhodopsins with Phosphodiesterase Activity.

The choanoflagellate Salpingoeca rosetta contains a chimeric rhodopsin protein composed of an N-terminal rhodopsin (Rh) domain and a C-terminal cyclic nucleotide phosphodiesterase (PDE) domain. The Rh-PDE enzyme (SrRh-PDE), which decreases the concentrations of cyclic nucleotides such as cGMP and cAMP in light, is a useful tool in optogenetics. Recently, eight additional Rh-PDE enzymes were found in choanoflagellate species, four from Choanoeca flexa and the other four from other species. In this paper, we studied the molecular properties of these new Rh-PDEs, which were compared with SrRh-PDE. Upon expression in HEK293 cells, four Rh-PDE proteins, including CfRh-PDE2 and CfRh-PDE3, exhibited no PDE activity when assessed by in-cell measurements and in vitro HPLC analysis. On the other hand, CfRh-PDE1 showed light-dependent PDE activity toward cGMP, which absorbed maximally at 491 nm. Therefore, CfRh-PDE1 is presumably responsible for colony inversion in C. flexa by absorbing blue-green light. The molecular properties of MrRh-PDE were similar to those of SrRh-PDE, although the λmax of MrRh-PDE (516 nm) was considerably red-shifted from that of SrRh-PDE (492 nm). One Rh-PDE, AsRh-PDE, did not contain the retinal-binding Lys at TM7 and showed cAMP-specific PDE activity both in the dark and light. These results provide mechanistic insight into rhodopsin-mediated, light-dependent regulation of second-messenger levels in eukaryotic microbes.

Introduction

Rhodopsins are heptahelical membrane proteins containing retinal as a chromophore. They are subdivided into animal and microbial rhodopsins.[1] While animal rhodopsins serve almost exclusively as G protein-coupled receptors,[2−5] the functions of microbial rhodopsins are diverse, including light-driven cation and anion pumps, light-gated cation and anion channels, positive and negative phototaxis sensors, photochromic sensors, and light-activated enzymes.[1,6−9] The first discovered light-activated enzyme was an algal histidine kinase (HK).[10] In addition, the discovery of enzyme rhodopsins functioning as a light-activated guanylyl cyclase (GC)[11−20] and phosphodiesterase (PDE)[21−24] attracted much attention from researchers working in optogenetics.

Optogenetics has revolutionized life sciences where light-gated channels and light-driven ion pumps have been used to excite and silence neurons.[25,26] This is highly advantageous for applications in neuroscience when fast temporal resolution is required to change transmembrane potential. In contrast, for the optogenetic control of a wider spectrum of biological functions, light-induced enzyme activation is more attractive because intracellular signaling processes can be manipulated by light.[27] In particular, the optogenetic control of secondary messengers such as cAMP and cGMP is in high demand. Therefore, the emergence of Rh-GC and Rh-PDE was welcomed by the field of optogenetics. Their molecular properties have been investigated.[11−24]

In the case of Rh-GC, several homologous proteins were reported in addition to the first reported Rh-GC from Blastocladiella emersonil (BeRh-GC), such as Rh-GC from Catenaria anguillulae (CaRh-GC).[11−20] Therefore, optimization of these proteins can be tested using site-directed mutagenesis. This is in stark contrast with discovery for Rh-PDEs. After the first report in March 2017,[24] homologous proteins have thus far not been found. The Rh-PDE from the choanoflagellate Salpingoeca rosetta (SrRh-PDE) was the only protein whose molecular properties were investigated.[21−24] Nevertheless, SrRh-PDE exhibits considerable dark activity under heterologous expression conditions, which makes it difficult to improve for optogenetic applications.

In June 2019, the situation changed dramatically after the discovery of additional Rh-PDEs from a multicellular choanoflagellate. Brunet et al. reported that a newly described choanoflagellate species, Choanoeca flexa (C. flexa), forms cup-shaped colonies that invert their curvature in response to light and that these new Rh-PDEs are responsible for light-signal transduction through the cGMP signaling.[28] Those authors showed that C. flexa contains four Rh-PDE homologs and an additional four Rh-PDEs from other species. This was an exciting finding for both fundamental and applied research. Although it is not clear why C. flexa contains four Rh-PDEs, these new Rh-PDEs may be more suitable for optogenetics than the currently known Rh-PDE.

In this paper, we characterized the molecular properties of these eight new Rh-PDEs. We first synthesized the full-length DNA, which was expressed in HEK293 cells. The PDE activity in cells was examined by the GloSensor assay. We also investigated the PDE activity biochemically using a crude membrane of HEK cells. Consequently, we observed the PDE activity from four new Rh-PDEs. While CfRh-PDE1, CfRh-PDE4, and MrRh-PDE exhibited light-dependent PDE activity, AsRh-PDE without the retinal-binding Lys at TM7 showed cAMP-specific PDE activity both in the dark and light. The remaining four Rh-PDEs showed no enzyme activity upon expression in HEK293 cells. Using the hydroxylamine bleach method, we obtained the absorption spectra of six new Rh-PDEs, whose maxima were distributed from 491 to 527 nm. The molecular mechanism of these new Rh-PDEs is discussed based on the present observations.

Materials and Methods

Molecular Biology

Amino acid sequences of eight Rh-PDEs were provided by Dr. N. King of UC Berkeley. Full-length genes encoding Rh-PDEs were synthesized after human codon optimization (Gen Script) and cloned into the pCS2+ vector between BamHI and XhoI sites using the In-Fusion HD Cloning Kit (Takara Bio). All constructs were verified by DNA sequencing.

Assay of the Enzymatic Activity of Rh-PDEs in Mammalian Cells

Enzymatic activity was evaluated by the GloSensor assay as described previously.[21] HEK293 cells were purchased from the JCRB Cell Bank and cultured in an E-MEM medium with -glutamine and phenol red (Wako) containing 10% (v/v) FBS and penicillin–streptomycin. The cells were co-transfected with the pCS2+ vector carrying the Rh-PDE genes and the pGloSensor-42F cGMP or pGloSensor-22F cAMP vector (Promega) by using Lipofectamine 2000 (Invitrogen). Transfected cells were incubated for 66–24 h at 37 °C in a medium containing 0.5 μM all-trans-retinal (Toronto Research Chemicals). Before measurements, the culture medium was replaced with a CO2-independent medium containing 10% (vol/vol) FBS and 2% (vol/vol) GloSensor cAMP or cGMP stock solution (Promega). Cells were then incubated for 2 h at room temperature in the dark. The intracellular cAMP or cGMP level was observed by monitoring luminescence with a microplate reader (Corona Electric) at 27 °C. The cells were treated with 3.5 μM forskolin (Wako), a direct activator of the endogenous adenylyl cyclase, to elevate the intracellular cAMP level. Alternatively, the cells were treated with 100 μM sodium nitroprusside (SNP) (Wako), a direct activator of the endogenous guanylyl cyclase, to elevate intracellular cGMP level. The cells were illuminated with 520 nm light at 0.35 μW/mm2 from a xenon lamp (LAX-103, Asahi Spectra Co., Ltd., Japan) through an interference filter. Light intensity was measured using a power meter LP1 (Sanwa Electric Instruments Co., Ltd., Japan). Six independent experiments were averaged.

HPLC Analysis for the in Vitro Assay of the Enzymatic Activity of Rh-PDEs

HEK293T cells were transfected with plasmid pCS2+_Rh-PDE by the calcium phosphate method.[21] The DMEM/F-12 medium contained 0.5 μM all-trans-retinal as well as penicillin and streptomycin. The cells were harvested after 24 h and washed with buffer A (140 mM NaCl, 3 mM MgCl2, 50 mM HEPES-NaOH, pH 7.5). The cells were resuspended in buffer A and homogenized using a Potter-Elehjem Grinder (Wheaton) and a syringe with a 27G needle. The syringe was filled and drained five times while stirring the homogenate. The amount of protein was determined by the BCA protein assay (Thermo Fisher Scientific). Samples were kept in the dark for at least 2 h before measurement. Catalytic activity was measured at room temperature in 100 μL of buffer A with 16–18 μg (in the case of cAMP) or 1.6–1.8 μg (in the case of cGMP) of total protein in a 1.5 mL sample tube. The sample was illuminated with a xenon lamp (MAX-303, Asahi Spectra Co., Ltd.) through a Y52 filter (7 mW/mm2). The reaction was initiated by adding cyclic nucleotides (final concentration, 100 μM). Aliquots were removed at different time points, and the reactions were immediately terminated by adding 100 μL of 0.1 N HCl and frozen in liquid nitrogen. After thawing, samples were centrifuged to remove the membranes and denatured proteins. Nucleotides (20 μL of aliquot) were separated by HPLC (Shimadzu systems) with a C18 reversed-phase column (Waters) and 100 mM potassium phosphate (pH 5.9), 4 mM tetrabutylammonium iodide, and 10% (vol/vol) methanol as the eluent.[29] Nucleotides were monitored at 254 nm. Data were evaluated with LabSolutions (Shimadzu). Peak areas were integrated and assigned to the educt cyclic nucleotide based on retention times of a corresponding standard compound.

Spectroscopy

Absorption maxima of the new Rh-PDEs were determined without purification.[24,30] Each rhodopsin molecule expressed in HEK293T cells was suspended in 1.5% DDM, 0.3% cholesteryl hemisuccinate, 50 mM hydroxylamine, 100 mM NaCl, and Na2HPO4 (pH 7.0) and illuminated at room temperature for 1 min with a 1 kW tungsten halogen projector lamp (Master HILUX-HR, Rikagaku, Japan) through a glass filter (Y-52, AGC Techno Glass, Japan) at wavelengths of >500 nm.

Phylogenetic Analysis of Rhodopsin Genes

The amino acid sequences of either full-length proteins, transmembrane domains, or enzyme domains of selected rhodopsins were aligned by using MEGA10 software.

Results

New Rh-PDEs from the Choanoflagellate

The genome of the choanoflagellate C. flexa (sequenced by the Broad Institute, NCBI accession PRJNA37927) contains four Rh-PDE genes (NCBI Gene ID: QDH43408.1 for CfRh-PDE1, QDH43410.1 for CfRh-PDE2, QDH43407.1 for CfRh-PDE3, and QDH43409.1 for CfRh-PDE4).[28] In addition, Brunet et al. reported four Rh-PDE genes in a recent paper (so far, undeposited in NCBI).[28]Figure a is the schematic drawing of the Rh-PDE architecture, in which the dimer is a functional unit. An N-terminal with eight transmembrane helical rhodopsin domains[22,23] and a cytoplasmic PDE domain constitutes Rh-PDEs (see Figure S1 for the complete amino acid sequence). Rh-PDEs are a microbial rhodopsin that binds all-trans-retinal as a chromophore and where light absorption triggers protein structural changes to induce (or modulate) their enzymatic activity.[21−24] A decrease in intracellular cyclic nucleotide leads to each function through intracellular signaling pathways, such as colony inversion in a multicellular choanoflagellate.[28]

Figure 1

(a) Rh-PDE composed of eight transmembrane helical rhodopsin domains and the cytoplasmic PDE domain, which forms a dimer. (b) Phylogenetic tree of the eight new Rh-PDEs, SrRh-PDE, and bacteriorhodopsin (BR). Full-length sequences were aligned.

Figure b shows the phylogenetic tree of Rh-PDEs based on their full-length amino acid sequences (Figure S1). Amino acid identities among the nine Rh-PDEs are shown in Figure S2. Phylogenetic trees of Rh-PDEs based on the amino acid sequences of the transmembrane (TM) and PDE domains are shown in Figure S3a,b, respectively. Many common residues of microbial rhodopsins are conserved in Rh-PDEs, but considerable uniqueness can be observed in the sequences (Figure S1). First, AsRh-PDE does not contain retinal-binding Lys, which is replaced by Asn. Therefore, AsRh-PDE probably lacks the ability to absorb light. L93 in TM3,[31] P186 in TM6,[32] and A215 in TM7[33] in bacteriorhodopsin (BR), known as color determinants in microbial rhodopsin, are also denoted as the L/Q, G/P, and A/T switches, respectively. In each switch, Leu, Gly, and Ala exhibit red-shifted absorption, while Gln, Pro, and Thr (Ser) exhibit blue-shifted absorption.

Enzymatic Activity of Rh-PDEs in Mammalian Cells

To measure the enzymatic activity of Rh-PDEs in cells, we used the GloSensor assay, which is based on a cyclic nucleotide-dependent luciferase.[21] In this assay, the Rh-PDE-expressing cells were first incubated in a serum-containing culture medium after the addition of all-trans-retinal in the dark at room temperature (∼27 °C). Luminescence intensity was increased by forskolin or sodium nitroprusside (SNP), which activates AC or GC, resulting in an elevation of the cytoplasmic cAMP or cGMP level, respectively. Then, photoactivated PDE activity was measured by observing the decrease in luminescence intensity in a light-dependent manner.

Figure S4 shows the time course after the addition of forskolin into HEK293 cells expressing new Rh-PDEs (red line) and SrRh-PDE (black line) together with the luciferase. The observed luminescence of red and black lines was smaller than that of the gray line, whose cells expressed only a luciferase. This suggests that the dark activity of Rh-PDEs, or that co-expression of the Rh-PDE and luciferase may affect the level of luminescence from the luciferase. Only AsRh-PDE exhibited no increase after the addition of forskolin, presumably originating from its strong dark activity. It should be noted that AsRh-PDE does not contain retinal-bonding Lys at TM7, which may be related to the present observation.

Then, we illuminated the sample and observed changes in PDE activity. Figure a shows the light-induced PDE activity of new Rh-PDEs (red line) and SrRh-PDE (black line). Clear light-dependent PDE activity was observed for CfRh-PDE1 and MrRh-PDE, as well as SrRh-PDE, while a small signal was observed for CfRh-PDE4. In contrast, light-dependent PDE activity was not observed for CfRh-PDE2, CfRh-PDE3, CpRh-PDE1, CpRh-PDE2, and AsRh-PDE. Thus, the Glosensor assay showed that only three Rh-PDEs exhibit light-dependent PDE activity among the eight proteins (Figure b). After the light was switched off, the decreased luminescence by light gradually recovered, and τ1/e values were 2.8, 7.6, and 3.4 min for CfRh-PDE1, CfRh-PDE4, and MrRh-PDE, respectively (Figure a). These values are close to that of SrRh-PDE (3.5 min).

Figure 2

In-cell measurements of the light-dependent PDE activity toward cAMP. (a) Changes of luminescence signals upon 2 min irradiation of 510 nm light (light-blue dot) of HEK293 cells with the empty vector (mock) and Rh-PDEs. Red lines represent the data for the new Rh-PDEs. Time constants (τ1/e) of the recovery kinetics after the light is switched off are 2.8 ± 0.3, 7.6 ± 2.8, 3.5 ± 0.3, and 3.4 ± 0.6 min for CfRh-PDE1, CfRh-PDE4, SrRh-PDE, and MrRh-PDE, respectively. (b) Averaged light-dependent PDE activity.

Figure S5a shows the results on PDE activity toward cGMP in the dark. Compared to the mock data (gray line in Figure S5a), no increase upon addition of SNP was observed for CfRh-PDE1, CfRh-PDE4, and MrRh-PDE (blue lines in Figure S5a) as well as for SrRh-PDE (black line in Figure S5a).[21] This fact suggests that the light-dependent PDE activity toward cGMP can be tested only for the remaining five proteins, among which AsRh-PDE does not contain the retinal-binding Lys. When the cells were illuminated, none of the Rh-PDEs showed light-dependent PDE activity toward cGMP (Figure S5). This may be reasonable as CfRh-PDE1, CfRh-PDE4, and MrRh-PDE, whose light-dependent activity toward cAMP was detectable (Figure b), showed no increase in luminescence (Figure S5a), presumably because of high dark activity.

HPLC Analysis of the Enzymatic Activity of Rh-PDEs

We next characterized the enzymatic activity of Rh-PDEs by HPLC analysis using crude membranes of HEK293 cells. In these experiments, cAMP or cGMP was mixed with the crude cell membranes expressing Rh-PDEs, and time- and light-dependent hydrolysis of cyclic nucleotides into 5′AMP or 5′GMP was monitored by HPLC. Figures S6 and S7 show the time course of the HPLC profile for cAMP and cGMP, respectively, whose concentrations were monitored. A clear decrease was only observed for MrRh-PDE, AsRh-PDE, and SrRh-PDE in the case of cAMP (Figure S6), while the decrease was observed for CfRh-PDE1, MrRh-PDE, and SrRh-PDE in the case of cGMP (Figure S7).

Figure summarizes the PDE activity studied by in vitro HPLC analysis. Among the eight new Rh-PDEs, four proteins, CfRh-PDE2, CfRh-PDE3, CpRh-PDE1, and CpRh-PDE2, exhibited no PDE activity from within the cells (Figure ) or in vitro (Figure ) measurements. The lack of a decrease in cyclic nucleotide concentration in the HPLC analysis (Figure ) strongly suggests that the lowered luminescence after expressing these proteins (Figure S3) did not originate from the dark PDE activity. This is an entirely unexpected result as these four Rh-PDEs are homologous proteins to a known Rh-PDE, SrRh-PDE. The reason is discussed below.

Figure 3

In vitro HPLC analysis of PDE activity toward (a) cAMP and (b) cGMP. HEK293 cell membranes expressing Rh-PDEs were incubated with (a) cAMP or (b) cGMP, and the decrease in each substrate was measured by HPLC in the dark (D) and light (L). Error bars represent SD.

While CfRh-PDE1 and CfRh-PDE4 showed in-cell PDE activity in response to cAMP (Figure b), no clear activity, as assessed by HPLC, was observed for CfRh-PDE4, both in the dark and light. In contrast, small but clear PDE activity was observed by HPLC for CfRh-PDE1 in response to cGMP, but the response to cAMP was unclear. One possibility is that the preparation of the crude membrane sample caused loss of enzyme activity. Another possibility is that both CfRh-PDE1 and CfRh-PDE4 possess PDE activity and the in-cell measurement was more sensitive than the HPLC measurement.

MrRh-PDE showed the largest light-dependent PDE activity in response to cAMP in the in-cell measurement (Figure b), which was larger than that of SrRh-PDE. In the case of cGMP, dark activity was so strong in the in-cell measurement that detectable luminescence was not observed before illumination. PDE activities in the dark and light were more quantitatively obtained by HPLC analysis, where the activities of MrRh-PDE in response to cAMP were larger than those of SrRh-PDE in both the dark and light (Figure a). Nevertheless, the light–dark activity ratio of MrRh-PDE (1.73 times) was lower than that of SrRh-PDE (3.11 times) (Figure a). We previously reported that the light–dark activity ratio of SrRh-PDE was 1.64 times.[21] A change in the construct of SrRh-PDE, mainly the removal of the 1D4 tag, resulted in an improved light–dark activity ratio for SrRh-PDE (3.11 times). Although we used the same construct for MrRh-PDE without the 1D4 tag, its light–dark activity ratio was not high. The tendency was more significant in response to cGMP. Figure b shows that the light–dark activity ratio of MrRh-PDE was only 1.12 times while that of SrRh-PDE was 1.90 times.

Finally, AsRh-PDE had a unique feature. The PDE activity in response to cAMP was significant, but it was the same in the dark and in the light. This is reasonable since AsRh-PDE does not contain retinal-binding Lys at TM7. On the other hand, the PDE activity in response to cGMP was negligible, indicating that AsRh-PDE is a cAMP-specific PDE.

Molecular Properties of Purified Full-Length Rh-PDEs

We next measured the absorption spectra of Rh-PDEs. To determine the absorption properties of rhodopsins, proteins are normally solubilized with a detergent and purified by column chromatography. In contrast, here we determined the absorption maxima of the new Rh-PDEs without purification by applying the hydroxylamine bleach method to gently solubilized proteins.[24,30]Figure shows the difference absorption spectra of Rh-PDEs before illumination minus after illumination in the presence of 50 mM hydroxylamine. The positive spectra correspond to those of Rh-PDEs. Among the eight Rh-PDEs, six absorption maxima were determined. The lack of an absorption spectrum for AsRh-PDE is reasonable since the retinal-binding Lys is replaced by Asn. Even though CfRh-PDE2 possesses retinal-binding Lys, we were unable to obtain a spectrum. It should be noted that only CpRh-PDE1 reacted with 50 mM hydroxylamine in the dark, forming a retinal oxime.

Figure 4

Light-induced difference absorption spectra of Rh-PDEs in the presence of 50 mM hydroxylamine. Positive and negative signals show the spectra before and after illumination, corresponding to those of the rhodopsin and retinal oxime, respectively. Red and black lines represent the spectra of new Rh-PDEs (λmax is shown in each panel) and SrRh-PDE (λmax: 492 nm), respectively. Only CpRh-PDE1 reacted with hydroxylamine in the dark.

The obtained absorption maxima, which were distributed between 491 and 527 nm, were roughly classified into three groups. In the first group, CfRh-PDE1 (491 nm) and CfRh-PDE4 (496 nm) possessed blue-shifted absorption, similar to SrRh-PDE (492 nm). The second group included MrRh-PDE (516 nm) and CpRh-PDE2 (519 nm), and the most red-shifted third group is composed of CfRh-PDE3 (524 nm) and CpRh-PDE1 (527 nm). There is currently no structural information about these Rh-PDEs, and Figure shows the 25 residues surrounding the retinal chromophore in a light-driven proton pump BR.[34] Among the nine Rh-PDEs, only the sequence of AsRh-PDE is different, including the lack of retinal-binding Lys and the counterion Glu. The remaining eight Rh-PDEs are highly homologous. In fact, among the 25 residues in Figure , 15 residues are identical for eight Rh-PDEs (gray background in Figure ). This fact strongly suggests that the remaining 10 residues are responsible for color variation over 36 nm, from 491 (CfRh-PDE1) to 527 nm (CpRh-PDE1).

Figure 5

Twenty-five amino acid residues surrounding the retinal chromophore in BR (PDB: 1C3W).[34]

The unique residue of MrRh-PDE and CpRh-PDE2 is Gly at TM6, whereas other Rh-PDEs contain Pro. The corresponding position was recently identified as the color determinant of microbial rhodopsins (G/P switch at TM6).[32] Pro is highly conserved at TM6 in microbial rhodopsins (P186 in BR) (1), while the replacement of Pro by Thr or Gly led to a spectral red-shift in the light-driven sodium pump KR2. In addition, a red-shifted sodium pump protein is found in nature.[32] The red-shifted spectra in MrRh-PDE and CpRh-PDE2 with Gly in the G/P switch are fully consistent, although this needs to be experimentally validated in the future. On the other hand, the origin of the spectral red-shift in CfRh-PDE3 and CpRh-PDE1 is not obvious from known color determinants. From Figure , their sequential uniqueness can be seen (i) in Q or E at TM1 and (ii) in N at TM4.

Discussion

Enzymatic Activity of New Rh-PDEs

Here, we report the enzymatic and molecular properties of novel Rh-PDEs from choanoflagellates by heterologously expressing them in HEK293 cells and comparing them with those of SrRh-PDE. Among the eight Rh-PDEs, AsRh-PDE did not contain the retinal-binding Lys at TM7, so we expected light-dependent PDE activity for the remaining seven proteins. However, the HPLC analysis in Figure showed no PDE activity for half of the new Rh-PDEs, CfRh-PDE2, CfRh-PDE3, CpRh-PDE1, and CpRh-PDE2, which is consistent with the results of the GloSensor assay (Figure ). The lack of enzyme activity may originate from different membrane and cellular environments between native choanoflagellate cells and HEK293 cells, such that the PDE domain, or other regulatory protein(s) associated with it, is post-translationally modified (phosphorylated) in a native environment. Another possibility is that these proteins do not possess PDE activity in a native choanoflagellate. C. flexa contains four Rh-PDEs, among which CfRh-PDE2 and CfRh-PDE3 may be pseudogenes or have different functional roles other than serving simply as light-activated enzymes.

Brunet et al. reported that Rh-PDEs are responsible through the cGMP signaling for colony inversion of C. flexa.[28] The present study revealed the PDE activity of CfRh-PDE1 and CfRh-PDE4 for cAMP but only CfRh-PDE1 for cGMP (Figures and 3). Thus, it is likely that CfRh-PDE1 contributes to colony inversion, though CfRh-PDE4 may have the PDE activity toward cGMP under physiological conditions. Then, the action spectrum of colony inversion should have a peak at around 490–500 nm since CfRh-PDE1 and CfRh-PDE4 absorb maximally at 491 and 496 nm, respectively. If CfRh-PDE3 contributes to that function, the action spectrum is more red-shifted since CfRh-PDE3 absorbs maximally at 524 nm. Since each protein was not purified in this study, information on their molecular mechanism is very limited. Nevertheless, recovery kinetics of luminescence after illumination (Figure a) provides the reaction cycle of PDE activity. The time constant of CfRh-PDE1 (τ1/e = 2.8 min) or CfRh-PDE4 (τ1/e = 7.6 min) was faster and slower, respectively, than that of SrRh-PDE (τ1/e = 3.5 min). Note that the photocycle of the Rh domain is much faster in SrRh-PDE.[21,24]

The PDE activity of MrRh-PDE is very similar to that of a known Rh-PDE, SrRh-PDE. The enzymatic reaction turnover between MrRh-PDE (τ1/e = 3.4 min) and SrRh-PDE (τ1/e = 3.5 min) was very similar. Higher PDE activity toward cGMP than toward cAMP was also observed for MrRh-PDE and SrRh-PDE. These similarities are reasonable as the amino acid sequence of MrRh-PDE is closest to that of SrRh-PDE (identity: 79% in Figure S2) among the eight new Rh-PDEs. Nevertheless, MrRh-PDE shows more enhanced dark activity.

Color Tuning Mechanism of Rh-PDEs

No absorption spectra were obtained for CfRh-PDE2 and AsRh-PDE, and the latter result is presumably because the retinal-binding Lys at TM7 is replaced by Asn (Figure ). The obtained absorption maxima for the remaining seven Rh-PDEs were distributed from 491 to 527 nm (Figure ). Figure also shows that 15 residues among 25 residues surrounding the retinal chromophore in these Rh-PDEs are identical. This fact strongly suggests that the remaining 10 residues determine color variation over 36 nm, from 491 (CfRh-PDE1) to 527 nm (CpRh-PDE1). Color-determining residues, which have thus far been reported for microbial rhodopsins, are discussed next.

The L/Q switch is a famous color determinant, where Leu and Gln at position 105 of TM3 discriminate red-shifted (green-absorbing) and blue-shifted (blue-absorbing) proteorhodopsins, respectively.[31] In the case of Rh-PDEs, Leu is highly conserved, while MrRh-PDE contains Ile at the corresponding position. A previous mutation study of green-absorbing proteorhodopsin (GPR) reported that L105I GPR shows an about 2 nm red-shift for the deprotonated form and an about 10 nm red-shift at pH 8.0.[35] A larger red-shift for the latter originates from the higher pKa value in the L105I mutant. Thus, the L/Q switch partly affects the spectral red-shift in MrRh-PDE.

The A/T switch (A/TS switch) is located at TM7, one residue before the retinal-binding Lys. It is well known that the introduction of an O–H-bearing residue such as Ser and Thr causes a spectral blue-shift.[1,30,33,36−38] In the case of Rh-PDEs, Ala is fully conserved except for AsRh-PDE (Figure ). The presence of Ala at the corresponding position suggests that Rh-PDEs have an intrinsically red-shifted nature from the A/S switch. This further suggests that absorption maxima of CfRh-PDE (491 nm), SrRh-PDE (492 nm), and CfRh-PDE4 (496 nm) are attained by a specific spectral blue-shift mechanism. A previous FTIR study of SrRh-PDE suggested a strong electrostatic interaction of the protonated Schiff base with D292 at TM7.[24]

We recently reported the presence of another color determinant, a G/P switch at TM6.[32] Pro is highly conserved at TM6 in microbial rhodopsins (P186 in BR), while replacement of Pro into Thr or Gly led to a spectral red-shift in a naturally found light-driven sodium pump KR2. MrRh-PDE and CpRh-PDE2 contain Gly at the corresponding position, while other Rh-PDEs contain Pro. The red-shifted spectra in MrRh-PDE (516 nm) and CpRh-PDE2 (519 nm) with Gly in the G/P switch is fully consistent with established knowledge. MrRh-PDE and CpRh-PDE2 contain Ile and Leu in the L/Q switch, respectively, and a similar λmax suggests that the G/P switch is more influential for the color tuning of MrRh-PDE and CpRh-PDE2.

Interestingly, CfRh-PDE3 (524 nm) and CpRh-PDE1 (527 nm) exhibit the most red-shifted spectra, although the mechanism cannot be explained by known color switches. Unique residues can be seen at TM1 and TM4. CfRh-PDE3 and CpRh-PDE1 contain Gln and Glu, respectively, at the position of M20 in BR (TM1), while other Rh-PDEs possess Ala or Gly. CfRh-PDE3 and CpRh-PDE1 contain Asn at the position of I119 in BR (TM4), while other Rh-PDEs possess Leu or Val. These residues may cause a spectral red-shift, or other residues not listed in Figure might contribute to the spectral red-shift. Identification of the color determinant is our future focus.

Applicability of Rh-PDEs for Optogenetics

The strong motivation of the present study was to find a better optogenetic tool than SrRh-PDE. However, the functional experiments provided rather negative results regarding application to optogenetics. Among the eight new Rh-PDEs, AsRh-PDE did not contain retinal-binding Lys. In addition, four proteins showed no enzymatic activity. Even though they are functional under native cell conditions, these proteins cannot be used in mammalian cells. Light-dependent PDE activity was observed for CfRh-PDE1, CfRh-PDE4, and MrRh-PDE, but the activity of the former two proteins was not significant. The activity of MrRh-PDE is comparable to that of SrRh-PDE, but the dark activity is higher for MrRh-PDE. Thus, the best Rh-PDE for optogenetics is SrRh-PDE at present, which is the first discovered Rh-PDE.

Although new Rh-PDEs are not promptly usable for optogenetics, there are several useful aspects. The PDE activity of AsRh-PDE is specific to cAMP, a property that is in contrast to SrRh-PDE, which has PDE activity toward both cAMP and cGMP, the latter being stronger. Thus, for example, the recovery of the retinal-binding Lys may lead to a substrate-specific Rh-PDE. The present data will stimulate further experimental and theoretical efforts to understand the molecular mechanism underlying the function of enzyme rhodopsins, which is also important for application to optogenetics.