Distinctive Properties of Dark Reversion Kinetics between Two Red/Green-Type Cyanobacteriochromes and their Application in the Photoregulation of cAMP Synthesis†
Abstract
Cyanobacteriochromes (CBCRs) are photoreceptors that bind to a linear tetrapyrrole within a conserved cGMP-phosphodiesterase/adenylate cyclase/FhlA (GAF) domain and exhibit reversible photoconversion. Red/green-type CBCR GAF domains that photoconvert between red- (Pr) and green-absorbing (Pg) forms occur widely in various cyanobacteria. A putative phototaxis regulator, AnPixJ, contains multiple red/green-type CBCR GAF domains. We previously reported that AnPixJ's second domain (AnPixJg2) but not its fourth domain (AnPixJg4) shows red/green reversible photoconversion. Herein, we found that AnPixJg4 showed Pr-to-Pg photoconversion and rapid Pg-to-Pr dark reversion, whereas AnPixJg2 showed a barely detectable dark reversion. Site-directed mutagenesis revealed the involvement of six residues in Pg stability. Replacement at the Leu294/Ile660 positions of AnPixJg2/AnPixJg4 showed the highest influence on dark reversion kinetics. AnPixJg2_DR6, wherein the six residues of AnPixJg2 were entirely replaced with those of AnPixJg4, showed a 300-fold faster dark reversion than that of the wild type. We constructed chimeric proteins by fusing the GAF domains with adenylate cyclase catalytic regions, such as AnPixJg2-AC, AnPixJg4-AC and AnPixJg2_DR6-AC. We detected successful enzymatic activation under red light for both AnPixJg2-AC and AnPixJg2_DR6-AC, and repression under green light for AnPixJg2-AC and under dark incubation for AnPixJg2_DR6-AC. These results provide platforms to develop cAMP synthetic optogenetic tools.
Introduction
Phytochromes (Phys) and cyanobacteriochromes (CBCRs) belong to a large superfamily of photoreceptors that possess a conserved cGMP-phosphodiesterase/adenylate cyclase/FhlA (GAF) domain 1-5. The GAF domain covalently binds a linear tetrapyrrole (bilin pigment), such as phycocyanobilin (PCB) and biliverdin IXα (BV) (see Figure S1), and commonly exhibits reversible photoconversion, which is provoked by Z/E isomerization of a double bond between C15 and C16 positions in the chromophore 6, 7. Only the GAF domain of the CBCRs is necessary for chromophorylation and proper photoconversion, whereas most Phys require a large unit consisting of Per/Arnt/Sim (PAS), GAF and PHY (phytochrome-specific) domains 2. Typically, the Phys reversibly respond to red and far-red light, whereas the CBCRs sense an extremely wide wavelength range from ultraviolet to far-red light 8-32. Among them, PCB-binding red/green-type CBCR GAF domains that reversibly photoconvert between the red light-absorbing form with 15Z-isomer (Pr) and the green light-absorbing form with 15E-isomer (Pg) are found widely in various cyanobacteria 10-12, 19, 23, 29 and have been extensively characterized using various spectroscopic techniques 33-39.
We previously reported that AnPixJ contains four tandemly arranged GAF domains with a histidine kinase/adenylate cyclase/methyl-binding protein/phosphodiesterase (HAMP) and a methyl-accepting chemotaxis protein (MA) domains at the C-terminus, which are considered to be involved in phototaxis regulation in the heterocyst-forming cyanobacterium Anabaena sp. PCC 7120 (see Fig. 1c, 29). The second, third and fourth GAF domains are red/green-type CBCR GAF domains, whereas the first GAF domain is not. The second and fourth GAF domains (AnPixJg2 and AnPixJg4) but not the third GAF domain efficiently binds to PCB. From that study, we concluded that AnPixJg2, but not AnPixJg4, showed red/green reversible photoconversion. Furthermore, the structure of AnPixJg2 in its Pr form was confirmed using X-ray crystallography, which is the first known report of the crystal structure of CBCR 40. Based on this structural information, nuclear magnetic resonance imaging and Raman spectroscopic studies have illustrated detailed photoconversion mechanisms 35, 36, 38. In addition, Rockwell et al. 19 reported that many red/green-type CBCR GAF domains are present in Nostoc punctiforme, and some of them show photoconversion from Pr to Pg and rapid dark reversion from Pg to Pr. Notably, there are no other CBCR subfamily GAF domains reported to show fast dark reversion except for DXCIP-type CBCR GAF domains 41.
Optogenetics, a biological technique that combines optics and genetics to regulate and monitor various biological functions in living organisms with the help of light, was developed in 2005 and has since been adopted rapidly and widely 42. Light is a useful tool that can be manipulated with a high spatiotemporal resolution, and is controllable in its quality and intensity. In optogenetics, light-switching systems that comprise a light-sensing input module and an enzymatic or transcriptional output module are designed to regulate various biological activities in response to light stimulation. Recently, it was reported that synthetic and catabolic enzymes for cyclic adenosine monophosphate (cAMP)—a second messenger involved in various responses in diverse organisms—were regulated by light stimulation 43, 44. Although these studies focused on Phys as a light-input module, Phys have a drawback in requiring a large unit for proper photoconversion, which may cause the unexpected localization of functional molecules. In this context, CBCRs might be preferable owing to their small molecular size and their considerable wide range of wavelengths that can be detected.
Here, we report a dark reversion mechanism of the red/green-type CBCR GAF domains and the creation of red/green and red/dark reversible cAMP synthetic systems. Although AnPixJg2 and AnPixJg4 share a high sequence identity, photoconversion was previously observed for AnPixJg2, but not for AnPixJg4. However, we found that AnPixJg4 exhibited photoconversion from Pr to Pg, and rapid dark reversion from Pg to Pr, akin to certain red/green-type CBCRs in N. punctiforme. Comparison between sequences of AnPixJg2 and AnPixJg4 (Fig. 1a,b) led to the identification of amino acid residues involved in dark reversion of the red/green-type CBCRs.
Materials and Methods
Bacterial strains and growth media
The Escherichia coli (E. coli) strain JM109 was used for cloning plasmid DNA, and the E. coli strain C41 harboring PCB synthetic system, pKT271 45, was used for protein expression. Bacterial cells were grown in Luria–Bertani (LB) medium containing kanamycin with or without chloramphenicol at 20 μg ml−1. For protein expression, the cells were grown in LB medium at 37°C until the absorbance at 600 nm was 0.4–0.8, and then Isopropyl β-D-1-thiogalactopyranoside (IPTG) was added at a final concentration of 0.01–2 mm. Subsequently, the cells were cultured at 18°C overnight.
Bioinformatics
Motif analyses were performed by SMART searches run on the Internet 46. Multiple sequence alignment was constructed using CLUSTAL_X 47. The crystal structure of AnPixJg2 Pr (PDB ID: 3W2Z) 40 was used to explore the key amino acid residues involved in dark reversion.
Plasmid construction
Plasmids expressing the His-tagged AnPixJg2 (amino acid residues 221–397) and AnPixJg4 (amino acid residues 592–763) constructed in the previous study were used 29. Site-directed mutagenesis of AnPixJg2 and AnPixJg4 was performed using the PrimeSTAR Max Basal Mutagenesis kit (TaKaRa) with appropriate nucleotide primers (Table S1). The adenylate cyclase (AC) catalytic region (amino acid positions 386–859) of the full-length CyaB1 was amplified from the Anabaena sp. PCC 7120 genomic DNA using PrimeSTAR Max DNA polymerase (TaKaRa) and the appropriate nucleotide primers (Table S1). The catalytic region was fused with each GAF domain using In-fusion HD Cloning System (TaKaRa) and Gibson assembly system (New England BioLabs, Japan). All expression constructs were verified by nucleotide sequencing.
Protein expression and purification
AnPixJg2, AnPixJg4 and their modified proteins were expressed in E. coli C41 pKT271. The cells were disrupted in a buffer (20 mm HEPES-NaOH, pH 7.5, 0.1 m NaCl and 10% (wt/vol) glycerol) by three passages through an Emulsiflex C5 high-pressure homogenizer at 83 MPa (Avestin). The mixture was centrifuged at 165 000 g for 30 min, and then, the supernatants were loaded onto a nickel affinity His-trap column (GE Healthcare) using the ÄKTAprime plus chromatography system (GE Healthcare) after filtration with a 0.2-μm cellulose ether membrane. The column was washed using a buffer containing 100 mm imidazole, and then, the His-tagged proteins were eluted with a linear gradient of the buffer containing 100–400 mm imidazole. Following incubation with 1 mm EDTA for 1 h, the proteins were dialyzed against the buffer without EDTA and imidazole.
Electrophoresis and zinc-induced fluorescence assay
Purified proteins in 2% (wt/vol) sodium dodecyl sulfate (SDS), 60 mm dithiothreitol and 60 mm Tris-HCl (pH 8.0) were separated by SDS polyacrylamide gel electrophoresis (PAGE) using a 10–12% (wt/vol) acrylamide gel, followed by staining with Coomassie Brilliant Blue R-250 (CBB). For the zinc-induced fluorescence assay after SDS-PAGE, the gel was soaked in 20 mm zinc acetate at room temperature for 30 min 48. Fluorescence was then visualized through a 600-nm-long path filter upon excitation with blue light (λmax = 470 nm), and green light (λmax = 527 nm) through a 562-nm short-path filter using a WSE-6100 LuminoGraph (ATTO) and a WSE-5500 VariRays (ATTO).
Spectroscopy and dark reversion kinetics
Ultraviolet and visible (UV–vis) absorption spectra of the proteins were recorded with a UV-2600 spectrophotometer (SHIMADZU) at 15, 20, 25 and 30°C using a temperature controller. Monochromic light of various wavelengths for photoconversion was generated using an Opto-Spectrum Generator (Hamamatsu Photonics, Inc.).
To monitor the photoconversion and dark reversion processes, absorbance at 648 nm of those proteins against various intensities of red light (650 nm; 1700, 1400, 1000, 700, 300, 130, 75 and 40 μmol m−2 s−1) was measured for 2 min with dark intervals of 5 min. The half-lives and the Arrhenius parameters were estimated from the dark reversion kinetics at the different temperatures.
AC assay
Reaction buffer contained 50 mm Tris-HCl (pH 8.0), 50 mm NaCl, 10 mm MgCl2 and 100 μm adenosine triphosphate (ATP). Concentrations of the chimeric proteins were measured using the Bradford method (BIO-RAD). The reaction was initiated by adding the chimeric proteins (final concentration: 1 μm for AnPixJg2-AC and AnPixJg4-AC; 2 μm for AnPixJg2_DR6-AC) to the reaction buffer at 25°C. The mixtures were illuminated by red (λmax = 650 nm) or green light (λmax = 540 nm) for AnPixJg2-AC, and illuminated by red light or kept in the dark for AnPixJg4-AC and AnPixJg2_DR6-AC before and during the reaction to keep Pg or Pr form. At 0, 1, 5, 10, 20, 30 and 60 min after addition of the proteins, the mixtures were immediately heated at 95°C for 3 min to stop the reaction, followed by the addition of nicotine adenosine dinucleotide (NAD; final concentration, 1 mm) for the internal standard (IS). Following centrifugation to remove large aggregates, the samples were filtered through a 0.2-μm PTFE membrane.
Elution of nucleotides was performed with an HPLC system (SHIMADZU) using a reverse-phase HPLC column, Kinetex C18 (2.1 i.d. × 100 mm, 1.7 μm; Phenomenex). Each 10-μL sample was injected and monitored at 260 nm. The molecules NAD, cAMP and ATP were separated using a gradient mode buffer (100 mm potassium phosphate, 4 mm tetrabutylammonium hydrogen sulfate, pH 6.0)/MeOH 95:5 to 80:20 (0–4 min) and 60:40 (4–5 min) at 0.4 mL min−1 and at 50°C. The molecules cAMP and ATP were assigned on the basis of the retention time (tR) of the standard compounds, and were quantified using standard curves.
Results
Re-analysis of AnPixJg2 and AnPixJg4 considering dark reversion
Although we had previously analyzed AnPixJg4 and observed it to not show any photoconversion 29, we speculate that its photoconversion was undetectable at the time because of rapid dark reversion, as seen in the case of certain red/green-type CBCRs in N. punctiforme 19. In order to verify this, His-tagged AnPixJg2 and AnPixJg4 were expressed in PCB-producing E. coli C41 pKT271, and were purified to near homogeneity by nickel affinity chromatography (Figure S2). Purified AnPixJg2 showed red/green reversible photoconversion as shown previously (Fig. 2a and Table 1) 29. Purified AnPixJg4 showed peak absorbance at 648 nm, which corresponds well to the Pr form of AnPixJg2, as shown previously (Fig. 2b and Table 1) 29. To detect dark reversion, the absorbance change at 648 nm was monitored by toggling the red light on and off at 25°C. Dark reversion from Pr to Pg for AnPixJg2 was barely detectable, whose half-life was up to 29.1 h (Figs. 2a and 3a–c, Table 2). By contrast, the red light-dependent absorbance decrease followed by recovery under dark incubation was observed for AnPixJg4 (Fig. 3a,b). This clearly showed that AnPixJg4 underwent both photoconversion and dark reversion. To detect a red light-activated spectrum, the absorption spectrum during red light illumination was monitored. As a result, a Pg form peaking at 563 nm was detected (Fig. 2b and Table 1). The Pg form of AnPixJg4 was about 20-nm red-shifted compared with that of AnPixJg2. Pg absorbance relative to Pr absorbance of AnPixJg4 was lower than that of AnPixJg2 (Fig. 2a,b).
| λ max, Pg | λ max, Pr | SAR | |
|---|---|---|---|
| AnPixJg2 | |||
| WT | 544 | 648 | 1.4 |
| H293Y | 544 | 648 | 1.3 |
| L294I | 549 | 648 | 1.0 |
| F308L | 544 | 650 | 1.5 |
| F319S | 547 | 650 | 0.6 |
| S320P | 544 | 649 | 1.1 |
| I331V | 541 | 648 | 0.9 |
| DR6 | 549 | 648 | 0.2˜0.6a |
| AnPixJg4 | |||
| WT | 563 | 648 | 1.6 |
| Y659H | 559 | 648 | 0.4 |
| I660L | 546 | 650 | 1.4 |
| L674F | 557 | 647 | 0.9 |
| S685F | n/a | n/a | n/a |
| P686S | 557 | 647 | 0.4 |
| V697I | ˜556 | ˜648 | 0.1 |
| Chimeric protein | |||
| AnPixJg2-AC | 543 | 649 | 0.5 |
| AnPixJg4-AC | 562 | 648 | 0.7 |
| AnPixJg2 DR6-AC | 550 | 648 | 0.1 |
- n/a, not applicable.
- SAR, specific absorbance ratio, which was calculated as peak red light absorption maximum/peak 280 nm absorption, providing a relative measure of chromophore-binding efficiency.
- a Different culture conditions resulted in various SAR values ranging from 0.2 to 0.6.
| Half-life at each temperature (°C) | ||||
|---|---|---|---|---|
| 15°C | 20°C | 25°C | 30°C | |
| AnPixJg2 | ||||
| WT | n.t. | n.t. | 29.1 h | n.t. |
| H293Y | 31.7 h | 33.6 h | 23.8 h | 14.9 h |
| L294I | 2.59 h | 1.20 h | 43.5 min | 24.4 min |
| F308L | 34.0 h | 26.9 h | 15.9 h | 7.81 h |
| F319S | 16.1 h | 9.22 h | 4.49 h | 2.71 h |
| S320P | 28.5 h | 20.1 h | 13.3 h | 5.47 h |
| I331V | 20.1 h | 20.1 h | 11.6 h | 6.01 h |
| DR6 | 31.8 min | 13.2 min | 5.60 min | 2.39 min |
| AnPixJg4 | ||||
| WT | 47.2 s | 27.2 s | 17.2 s | 14.4 s |
| Y659H | 1.62 min | 1.06 min | 55.7 s | n.d. |
| I660L | 11.8 min | 5.38 min | 2.60 min | 1.32 min |
| L674F | 44.3 s | 27.3 s | 20.2 s | n.d. |
| S685F | n.t. | n.t. | n.t. | n.t. |
| P686S | 1.26 min | 40.9 s | 25.1 s | 20.7 s |
| V697I | n.d. | n.d. | n.d. | n.d. |
| Chimeric protein | ||||
| AnPixJg2-AC | n.t. | n.t. | 19.9 h | n.t. |
| AnPixJg4-AC | 1.28 min | 42.3 s | 23.4 s | 14.6 s |
| AnPixJg2 DR6-AC | 14.2 min | 6.93 min | 3.26 min | 1.56 min |
- n.t., not tested, owing to no chromophorylation; n.d., not detectable, owing to very low chromophorylation or inactivation of photoactive components under high temperature.
To measure the detailed kinetics of AnPixJg4 photoconversion and dark reversion, we monitored absorbance changes at 648 nm during the toggling of the red light with different light intensities at different temperatures (Fig. 3a,b). Based on this analysis, the half-lives of AnPixJg4 dark reversion were estimated to be at 47.2, 27.2, 17.2 and 14.4 s under 15, 20, 25 and 30°C, respectively (Fig. 3c and Table 2). The half-life of AnPixJg4 at 25°C was about 6000-fold shorter than that of AnPixJg2. Activation energy of AnPixJg4 was ascertained to be 58.5 kJ mol−1, using the Arrhenius plot (Fig. 3d). This value is similar to those belonging to other CBCR GAF domains showing rapid dark reversion 41. One of them, NpF2164g7, has been previously shown to function as a light intensity sensor 19. In this context, AnPixJg4 is also likely to function as a power sensor. At all the tested temperatures, the light-dependent absorbance changes decreased with lower light intensities (Fig. 3a). Plotting the absorbance change at 648 nm (∆A648) versus the light intensity demonstrated that ∆A648 increased with lower temperature, and was nearly saturated only at 15°C (Fig. 3b). These results, together with the dark reversion kinetics, implied that suppression of dark reversion at lower temperatures results in a shift of equilibrium in favor of the Pg form. Such a temperature-dependent manner may mean that AnPixJg4 integrates light power signal and temperature signal.
Site-directed mutagenesis for identifying key residues in dark reversion
Although AnPixJg2 and AnPixJg4 possess distinct dark reversion properties (barely detectable reversion and rapid dark reversion, respectively), they both share sequence homology at a 66% identity. This pushed us to explore key amino acid residue(s) playing a role in dark reversion. We regarded residues within 6 Å of the chromophore as candidates crucial for the Pg stability. Based on the AnPixJg2 Pr structure, we extracted 30 residues within 6 Å of the chromophore. Among them, six residues of AnPixJg2 were different from those of AnPixJg4—His293, Leu294, Phe308, Phe319, Ser320 and Ile331 for AnPixJg2, and Tyr659, Ile660, Leu674, Ser685, Pro686 and Val697 for AnPixJg4 (Fig. 1a,b). Therefore, we performed g2-to-g4 and g4-to-g2 site-directed mutagenesis individually—H293Y, L294I, F308L, F319S, S320P and I331V for AnPixJg2, and Y659H, I660L, L674F, S685F, P686S and V697I for AnPixJg4.
The variant proteins were isolated using nickel affinity chromatography (Figure S2). Considering the zinc-induced fluorescence and absorption spectroscopy, it was evident that all AnPixJg2 variant proteins covalently bound to PCB. The specific absorbance ratio (SAR) values showed that the F319S variant exhibited low albeit covalent chromophore binding, whereas the other variants showed moderately efficient chromophorylation (Figure S3a and Table 1). All AnPixJg2 variants, apart from L294I, showed red/green reversible photoconversion, which was similar to that of the wild-type AnPixJg2. The absorption maximum of the L294I Pg form was observed at 549 nm, which was red-shifted by 5 nm as compared with that of the wild type; however, its Pr form showed a peak at 648 nm, the same as that of the wild type.
All AnPixJg4 variant proteins, except for S685F, bound covalently to PCB. Based on the SAR values, Y659H, P686S and V697I variants showed a reduced albeit covalent chromophore binding, whereas the other variants showed moderately efficient chromophorylation (Figure S3b and Table 1). All AnPixJg4 variants, except for S685F, showed photoconversion from Pr to Pg, and dark reversion from Pg to Pr. Absorption maxima of the Pg forms of all variants were blue-shifted, but those of their Pr forms were almost the same as those of the wild-type AnPixJg4. In particular, the Pg form of I660L variant was largely blue-shifted by 17 nm relative to that of the wild type.
Next, we measured dark reversion kinetics of all the variant proteins from Pg to Pr at different temperatures (15, 20, 25 and 30°C) (Table 2). We could not measure dark reversion kinetics of the S685F and V697I variants of AnPixJg4 because the former showed no chromophorylation and that of the latter was too low. Furthermore, we were unable to measure dark reversion kinetics of Y659H and L674F variants of AnPixJg4 under 30°C owing to the inactivation of their photoactive components. As a result, Pg forms of all AnPixJg2 variants converted to Pr faster than that of AnPixJg2 wild type, whereas Pg forms of all AnPixJg4 variants, except for L674F, converted to Pr slower than that of AnPixJg4 wild type. Amino acid replacements at the same positions of AnPixJg2 and AnPixJg4, L294I and I660L, showed the most considerable effects on dark reversion kinetics (Fig. 4b and Table 2). In particular, L294I replacement in AnPixJg2 resulted in a 40-fold faster dark reversion than the wild type at 25°C, whereas I660L replacement in AnPixJg4 resulted in a nine-fold slower dark reversion than the wild type. Interestingly, these replacements also had the largest effects on the absorption maxima of Pg forms as described earlier (Figure S3a,b and Table 1). Each replacement resulted in a shift of the absorption maximum on approaching to the other Pg form. In summary, Leu/Ile residues at 294/660 positions of AnPixJg2 and AnPixJg4 are likely to be most crucial residues for spectral tuning and stability of the Pg form among the residues studied presently (Fig. 1b, inset).
Because all single replacements of AnPixJg2 resulted in a faster dark reversion than the wild type, we constructed a mutant that possessed all the mutations (AnPixJg2_DR6) in order to obtain a protein with rapid dark reversion. Judging from its low SAR value (~0.2), purified AnPixJg2_DR6 covalently bound to PCB with reduced efficiency when 0.1 mm IPTG was added to the 1-L LB medium (Table 1). Because a very high quantity of AnPixJg2_DR6 was isolated, low chromophorylation may be due to a high-yield production of apo-AnPixJg2_DR6, and an accompanying shortage of PCB supply. Thus, we expressed AnPixJg2_DR6 with different IPTG concentrations and different incubation volumes. All these expression conditions resulted in cell pellets with a deep blue color (representative cell pellets are shown in Figure S4). These purified proteins showed SAR values ranging from 0.2 to 0.6 (Table 1). This strongly indicated that PCB levels largely affect chromophorylation, and apo-AnPixJg2_DR6 itself can efficiently bind to PCB. The Pr form showed a peak absorbance at 648 nm, which was the same as that of both AnPixJg2 and AnPixJg4, whereas its Pg form showed a peak absorbance at 549 nm, which was 5-nm red-shifted relative to AnPixJg2 Pg (Fig. 4a). As for the dark reversion kinetics from Pg to Pr, half-lives were calculated at 31.8, 13.2, 5.60 and 2.39 min under 15, 20, 25 and 30°C, respectively (Fig. 5c and Table 2). The dark reversion of AnPixJg2_DR6 at 25°C was about 300-fold and eight-fold faster than that of AnPixJg2 and AnPixJg2 L294I, respectively (Fig. 4b). It was, however, still slower than that of AnPixJg4 and AnPixJg4 I660L (Fig. 4b). These results suggested the presence of other key amino acid residues that remotely affect spectral tuning and stability of the Pg form. The activation energy of AnPixJg2_DR6 was shown to be 125 kJ mol−1 by Arrhenius plot, which was about two-fold higher than that of AnPixJg4 (Fig. 5d). AnPixJg2_DR6 also exhibited a reduced light-dependent absorbance change with lower light intensities under the different temperatures, as well as AnPixJg4 (Fig. 5a,b). Full Pg formation of AnPixJg2_DR6 under high light illumination was achieved even at high temperatures (20 and 25°C), while full Pg formation under high light irradiation was not achieved over 20°C, but only at 15°C, for AnPixJg4 (Figs. 3b and 5b). This is consistent with the fact that the dark reversion of AnPixJg2_DR6 is slower than that of AnPixJg4.
AnPixJg2_DR6 showed a moderately slow dark reversion as compared with AnPixJg4 (Table 2). Thus, we could generate the Pr form with and without green light illumination—photoactivated (PA) Pr form and dark reverted (DR) Pr form (Figure S5a), which were spectrally identical to each other. Furthermore, in response to red light irradiation (650 nm; 40 μmol m−1 s−1), both the PA and DR-Pr forms photoconverted to identical Pg forms, with kinetics almost identical to those of the wild-type protein (Figure S5a,b). These results indicate that dark reversion and light conversion processes from Pg provide identical Pr products.
Creation of red/green and red/dark reversible cAMP synthetic systems
We attempted to create cAMP synthetic light switches using AnPixJg2 and AnPixJg4 for optogenetics. We chose the CyaB1 protein from Anabaena spp. as a catalytic protein of adenylate cyclase (AC) whose enzymatic properties are characterized in detail 49. AnPixJg2-AC and AnPixJg4-AC were constructed by replacing the second GAF domain of CyaB1 with AnPixJg2 and AnPixJg4, and by removing its N-terminal GAF domain (Fig. 1c). It should be noted that the purified chimeric proteins showed red/green and red/dark reversible conversion almost similar to that of the GAF domains alone (AnPixJg2-AC, λmax, Pg 543 nm, λmax, Pr 649 nm; AnPixJg4-AC, λmax, Pg 562 nm, λmax, Pr 648 nm) (Figures S2 and S6a,b, Table 1). On the other hand, the dark reversion kinetics of each chimeric protein were slightly different from that of the GAF domain alone (Table 2). Half-life of AnPixJg2-AC was estimated to be 19.9 h at 25°C, which was 1.5-fold shorter than that of the GAF domain alone. Half-lives of AnPixJg4-AC were 1.28 min and 42.3, 23.4 and 14.6 s under 15 and 20, 25 and 30°C, respectively. The durations were about 1.5-fold longer than that of the GAF domain alone.
Next, we measured AC activities of the chimeric proteins (representative results are shown in Figure S7). AnPixJg2-AC showed significant light quality-dependent activity, wherein the AC activity under red light irradiation was two to three times higher than that under green light (Fig. 6a). As for AnPixJg4-AC, we could not detect a significant difference between AC activity under red light and under dark incubation (Fig. 6b). Instead, to develop a red/dark reversible cAMP synthetic system, we introduced the “DR6″ residues into AnPixJg2-AC, namely AnPixJg2_DR6-AC. Although its chromophore-binding efficiency was lower than that of AnPixJg2-AC and AnPixJg4-AC, purified AnPixJg2_DR6-AC covalently bound to PCB (Figures S2 and S6c, Table 1). AnPixJg2_DR6-AC showed a peak absorbance at 550 nm for Pg and 648 nm for Pr. Dark reversion kinetics of AnPixJg2_DR6-AC were calculated at 14.2, 6.93, 3.26 and 1.56 min under 15, 20, 25 and 30°C, respectively (Table 2). The dark reversion was found to be 1.5-fold to two-fold faster than that of the GAF domain alone. Using this chimeric protein, we could detect a significant difference between the AC activity under red light irradiation and under dark incubation. In particular, the AC activity under red light was about 1.5 times higher than that under dark incubation (Fig. 6c).
Discussion
In this report, we revealed that AnPixJg4 did show photoconversion from Pr to Pg and rapid dark reversion from Pg to Pr, whereas AnPixJg2 showed red/green reversible photoconversion and a barely detectable dark reversion. Half-life of the dark reversion of AnPixJg4 at 25°C was about 17 s, which is comparable to that of NpF2164g7 (about 4 s at 25°C), a red/green-type CBCR GAF domain in N. punctiforme. This GAF domain likely serves as a power sensor, but not as a color sensor, because of its rapid dark reversion 19. Similarly, AnPixJg4 may also serve as a power sensor. In fact, red light irradiation with low light intensity did not achieve full Pg formation (Fig. 3a). Because AnPixJg2 and AnPixJg4 are tandemly arranged in one protein, AnPixJ may integrate signals of both light quality and light intensity.
Site-directed mutagenesis identified six key amino acid residues for dark reversion. Among them, replacements at the same positions of AnPixJg2 and AnPixJg4, L294I and I660L showed the most significant effect on dark reversion kinetics (Table 2). In the AnPixJg2 Pr structure, Leu294 of AnPixJg2 is located 4–5 Å from the C- and D-rings of PCB, and serves to form a hydrophobic cavity for these rings (Fig. 1b, inset) 40. Considering the absorption peaks and shapes, the Pr forms of AnPixJg2 and AnPixJg4 are very similar to each other. In this context, Ile660 of AnPixJg4 likely serves to form the hydrophobic cavity for the C- and D-rings in the Pr form, as well as Leu294 of AnPixJg2. Based on the structural information of TePixJ, whose structures have been revealed in both 15Z- and 15E-isomers 40, 50, 51, the “flip-and-rotate” model originally proposed for the Phys was also applicable to the CBCRs 52. The initial D-ring flip triggers a subsequent rotation of the chromophore in the plane of the B- and C-rings within the pocket. It is further suggested that the Asp492 residue of TePixJ (corresponding to Asp291/Asp657 of AnPixJg2/AnPixJg4) then switches interaction partners upon this structural arrangement—Asp492 interacts with the A-, B- and C-rings in the 15Z-isomer, and with the D-ring in the 15E-isomer. Because Asp291/Asp657 of AnPixJg2/AnPixJg4 are located three residues upstream of Leu294/Ile660, spatial positioning of Leu294/Ile660 is also likely to be largely affected by this structural arrangement, which may lead to close contact between Leu294/Ile660 and the D-ring in the Pg forms (Figure S8). Consequently, subtle differences between Leu and Ile may result in distinctive effects on the Pg stability, in which Leu294, but not Ile660, stabilizes the Pg form. The other residues may also be structurally affected by photoconversion to facilitate dark reversion in AnPixJg4 directly or indirectly. To discuss further the detailed mechanisms, the Pg structures of AnPixJg2 and AnPixJg4 need to be determined.
Rockwell et al. 19 have reported that certain red/green-type CBCR GAF domains in N. punctiforme show photoconversion from Pr to Pg and rapid dark reversion from Pg to Pr. Among them, NpF2164g7 and NpR4776g2 show rapid dark reversions, whose half-lives are 4 and 25 s, respectively, at 25°C. Although we compared the “DR6” residues of AnPixJg4 with the corresponding residues of NpF2164g7 and NpR4776g2, we could not identify any common residues. This may reflect convergent evolution in which mutations are independently accumulated to facilitate dark reversion.
We also developed the red/green and red/dark reversible cAMP synthetic systems based on the PCB-binding red/green-type CBCR GAF domains, which have been reported for the first time, according to our knowledge. Because the molecular size of the CBCR GAF domains is about 25 kDa and is much smaller than the chromophore-binding region of the Phys, the CBCR-based systems developed in this study are beneficial for heterologous expression, although there is still plenty of room for improvement. Induction ranges for AnPixJg2-AC and AnPixJg2_DR6-AC are about three-fold and 1.5-fold, respectively (Fig. 6), in which repressed Pg forms of both AnPixJg2-AC and AnPixJg2_DR6-AC show significant enzymatic activities. Reduction of these activities should largely contribute to improved induction ranges. Previous studies have shown that alteration of linker length and mutations within photosensory and AC regions largely affect the induction range 43, 44; similar trials may be effective for our chimeric proteins. In the case of AnPixJg2_DR6-AC, its binding efficiency to PCB is not high, judging from fluorescence detection and SAR value (Figures S6c and S2, Table 1). If apo-AnPixJg2_DR6-AC holds some AC activity, apparent induction range should be smaller than the actual induction range displayed by holo-AnPixJg2_DR6-AC. Improvement of the binding efficiency may result in a larger induction range.
Till date, no CBCRs were shown to bind to BV, and we have presently identified that some red/green-type CBCR GAF domains from the chlorophyll d-bearing cyanobacterium Acaryochloris marina covalently bind not only to PCB but also to BV with high efficiency 10-12. Because the conjugated π system of BV is longer than that of PCB (Figure S1), these BV-binding CBCRs can sense light quality for longer than the PCB-binding ones do. The BV-binding ones showed far-red/orange reversible photoconversion, whereas PCB-binding ones showed red/green reversible photoconversion. Light quality of far-red and near-infrared regions were the primary focus of recent studies on optogenetic light sources because of their efficient penetration into deep animal tissues, wherein heme and water absorb light of shorter visible and far-infrared wavelengths 53. Moreover, BV, but not PCB, is an intrinsic chromophore that is synthesized in mammalian cells, and therefore, neither the addition of BV nor the introduction of BV biosynthetic pathway is needed for optogenetic control. In this context, if we can succeed in constructing far-red/dark reversible cAMP synthetic systems based on the BV-binding CBCR GAF domains, such systems would be advantageous for application to animal tissues. We are currently attempting to create BV-based chimeric proteins.
Acknowledgements
We thank Prof. K. Awai for the experimental support and helpful discussion (Shizuoka University). This work was supported by JST, PRESTO and JSPS KAKENHI Grant Number 26702036 (R.N.). The authors would like to thank Enago (www.enago.jp) for the English language review. We really appreciate Prof. Wolfgang Gärtner's considerable contribution to the photobiological research area.
