Overexpression of Different Types of Microbial Rhodopsins with a Highly Expressible Bacteriorhodopsin from Haloarcula marismortui as a Single Protein in E. coli

Microbial rhodopsins (M-Rho) are found in Archaea, Bacteria and some species of Eukarya and serve as light-driven ion pumps or mediate phototaxis responses in various biological systems. We previously reported an expression system using a highly expressible mutant, D94N-HmBRI (HEBR) from Haloarcula marismortui, as a leading tag to assist in the expression of membrane proteins that were otherwise difficult to express in E. coli. In this study, we show a universal strategy for the expression of two M-Rho proteins, either the same or different types, as one fusion protein with the HEBR system. One extra transmembrane domain was engineered to the C-terminal of HEBR to express another target M-Rho. The average expression yield in this new system reached a minimum of 2 mg/L culture, and the maximum absorbance of the target M-Rho remained unaltered in the fusion forms. The fusion protein showed a combined absorbance spectrum of a lone HEBR and target M-Rho. The function of the target M-Rho was not affected after examination with functional tests, including the photocycle and proton pumping activity of fusion proteins. In addition, an otherwise unstable sensory rhodopsin, HmSRM, showed the same or even improved stability under various temperatures, salt concentrations, and a wide range of pH conditions. This HEBR platform provides the possibility to construct multi-functional, stoichiometric and color-tuning fusion proteins using M-Rho from haloarchaea.

To create even more versatile applications, expressing different types of M-Rho as one fusion protein is indeed a logical approach, since it will enable the design of new, multi-capability proteins, which can be fixed at a desired stoichiometric ratio of component moieties. In fact, a study 23 showed stoichiometric and co-localized expression of a channelrhodopsin and halorhodopsin as one fusion protein and further demonstrated that this new protein can excite or inactivate neurons upon illumination with different wavelengths of light.
Previously, we reported 24 a system that adopted a highly expressible bacteriorhodopsin (HEBR) system in E. coli that served as an expression tag preceding target membrane proteins to enhance target protein functional expression in E. coli and achieved a yield of more than 2 mg/mL. In this study, we successfully extended one transmembrane segment from HEBR to fuse with various target M-Rho proteins, including Halobacterium salinarum bacteriorhodopsin (HsBR), Natronomonas pharaonis halorhodopsin (NpHR), Haloarcula marismortui sensory rhodopsin II (HmSRII), and Haloarcula marismortui sensory rhodopsin M (HmSRM). In each construct, the optical property, function and stability of the target M-Rho protein were examined and found to be intact or unaltered.

Results
Construction of HEBR-assisted M-Rho plasmids for E. coli expression. The D94N mutant of H. marismortui bacteriorhodopsin I, designated HEBR for short, was constructed as described previously 25 . The yield of HEBR alone reached 60-70 mg/L in culture without any reconstitution or refolding 24 , and the maximum absorbance (Abs-max) of HEBR 26 was 552 nm, almost identical to the wild type protein, HmBRI. Such an overexpression yield in E. coli system inspired the adoption of HEBR as a co-expression tag for other kinds or types of M-Rho proteins to create fusion proteins featuring multi-capability yet with ample expression yield for further creative applications.
To fuse and co-express another M-Rho protein with HEBR ( Fig. 1a), the C-terminus of HEBR was first extended using the first transmembrane region (TM1), a total of 37-a.a. in length, from HmHtrI. The TM1 was then connected with a flexible loop comprised of Ala-Ser-Ala-Ser-Asn-Gly-Ala-Ser-Ala followed by a designated target M-Rho containing a six-histidine tag appended to its C-terminus. The expression cassettes and restriction enzyme cutting sites for all the constructs in this study were summarized in Fig. 1b. Among them, a) HEBR-HsBR contains a bacteriorhodopsin from H. salinarum, which is known to have an extremely low expression efficiency in E. coli; b) HEBR-NpHR has a halorhodopsin from N. pharaonic that is a light-driven inward chloride transporter; c) HEBR-HmSRII attaches a sensory rhodopsin from H. marismortui, which is known to sense wavelengths approximately 498 nm and mediates a photorepellent response; and d) HEBR-HmSRM has a new type of functionally unknown sensory rhodopsin 25 from H. marismortui, which was reported to be extremely unstable after purification.
Visual colors of purified target proteins and UV-vis absorbance spectra. The expression of these four fusion proteins is described in the Methods. Briefly, plasmids of fusion proteins were first transformed into E. coli C43, and an overnight single colony was inoculated in LB broth until OD 600 = 0.4-0.6 before being induced with IPTG to a final concentration of 1 mM and 5-10 μM of all-trans retinal (Sigma, R-2500, USA). After another 4-6 hours of culture, the cells were harvested using centrifugation, and the pellets were observed with unaided eyes for their visual colors (Fig. 2) after removal of the supernatant. The cell pellets expressing different fusion proteins were visually distinguishable by the color difference compared with the target protein alone. Such visual color differences can serve as early indications for successful fusion protein expression. On the practical side, our system paves the way for the potential to develop a screening system to evaluate the expression efficiency of target proteins achieved by the visual color of ITPG-induced E. coli cell pellets and can also be applied to search particular color tuning mutants. All the proteins expressed were first purified as described in the Methods. A final minimum yield of 2 mg/L culture was observed in HEBR-HsBR, while all the other target proteins had yields of more than 4 mg/L culture. Different E. coli strains were used (Supplementary Fig. 1) to prove that this fusion strategy was not due to effects of the C43 strain.
The two primary optical properties are the maximum absorbance (Abs-max) and the kinetics of the photocycle. To test whether the HEBR fusion tag disturbs the Abs-max of the individual target M-Rho in fusion proteins, the purified (Fig. 3, inlets) HEBR-HsBR, HEBR-NpHR, HEBR-HmSRII, and HEBR-HmSRM were subjected to UV-Vis spectrophotometric scanning ( Fig. 3a-d). When compared with the Abs-max of singly expressed target M-Rho proteins, the UV-vis scanning profiles of the fusion proteins each showed a pure combination of the scanning profile of a singly expressed M-Rho merged with that of HEBR, thus indicating that the designed fusion tag, HEBR, TM1 of HmHtrI and the fusion loop did not disturb the optical property of either the individual target M-Rho or HEBR itself, in each case.
Photocycle and light-driven proton assays showed the un-altered function of the target M-Rho. The second optical property, the kinetics of the photocycle and the light-induced photocurrent were examined for the HEBR-NpHR fusion protein. In the laser-induced photolysis measurements, the ground state photocycle profile of the HEBR-NpHR fusion protein (Fig. 4a, bottom, 567 nm) showed a pattern resembling that of a merged HEBR (Fig. 4a, top, 550 nm) and NpHR alone (Fig. 4a, middle, 580 nm). The 410 nm M-intermediate that only appeared in HEBR alone and the 650 nm O-intermediate that was only detectable in NpHR were both recorded in the HEBR-NpHR fusion protein, indicating that both HEBR and NpHR were still active.
The functionality of HEBR-NpHR was probed with the whole cell E. coli pH and protein-based photocurrent assays that we previously adopted and developed 27 . After expressing HEBR-NpHR in E. coli, the cells were centrifuged and resuspended in a non-buffer solution. Upon illumination of cells with a 532-nm laser (Fig. 4b, top-bottom, indicated with green bar), the pH changes in the solution were monitored. In E. coli cells expressed with HEBR alone (Fig. 4b top), the illumination induced a pH decrease in each repeat (black line), while those pH changes were eliminated in the presence of an ionophore CCCP (red line). In the cells expressing solely NpHR (Fig. 4b middle), the whole-cell measurement showed the pattern established to indicate light-driven inward chloride pumping action, which is known to create a passive proton influx that consequently leads to an increase in the environmental pH. In the cells expressing the HEBR-NpHR fusion protein (Fig. 4b bottom), it showed a pattern resembling that of NpHR cells alone but is surmised to be a combined pattern of NpHR and HEBR with the NpHR component being dominant. The rationale is that NpHR executes light-driven inward chloride pumping 10-times faster than the light-driven proton outward pumping of HEBR.
In addition to the whole-cell measurements, we also conducted purified protein-based in vitro proton pumping measurements using pH sensitive ITO glass to detect the proton flow. HEBR (Fig. 4c, top) showed an outward proton flow, since it is classified as a proton pump. The same phenomenon was observed in NpHR (Fig. 4c, middle), which was reported to undergo a proton circulation on the cytoplasmic side 27 . The same measurement was conducted on the HEBR-NpHR fusion protein (Fig. 4c, bottom) to examine its biological functionalities. In the fused form, a light-triggered proton outward flux signal was observed, indicating that both HEBR and NpHR were functionally intact.
Due to the lack of a valid in vitro functional assay for sensory rhodopsins, the status of HmSRII and HmSRM was examined with their optical properties. In a flash laser-induced photolysis experiment, the decay and regeneration of the ground state upon light-activation is a solid implication of the stability, integrity, and functionality of sensory rhodopsins 28 . In the case of the fusion rhodopsins, HEBR-HmSRII was observed to complete its photocycle at the time-scale of ~2 seconds when monitored with the Abs max of HEBR (Fig. 5a, upper) and that of HmSRII (Fig. 5a, lower), both patterns being comparable to both HmSRII alone (Fig. 5c) and HEBR alone 26 .
A faster recovery kinetic of 4 seconds in the HEBR-HmSRM protein was observed at both 505 nm and 550 nm (Fig. 5b), which are the Abs-max for HmSRM and HEBR, respectively, in comparison to the ~10 seconds measured in HmSRM alone (Fig. 5d). These results indicated that the target M-Rho retained its original photocycle properties in the fused form. The otherwise unstable HmSRM even showed an accelerated photocycle, from 10 s to 4 s in HEBR-HmSRM form (Fig. 5b,d), in addition to only marginally shifted spectra under extreme conditions of temperatures, salt concentrations, and pH conditions ( Supplementary Fig. 2), indicating that the HmSRM was stabilized under such a fusion system. We therefore conclude that sensory type M-Rho can be fused with HEBR to obtain functionally intact proteins.

Discussion
Microbial rhodopsins have been widely adopted for biotechnological and biophysical applications during the previous decades. They are primarily utilized in optical appliances, therapeutic/medical applications and certain computational purposes 21 , including the most recent optogenetic studies 29 . In this study, we successfully developed an expression strategy using the HEBR system that we previously reported to create a single fusion protein with a combination of two M-Rho proteins. All the fusion proteins were functional and showed reasonable yields.
Our study demonstrated that when two M-Rho proteins were fused together in our HEBR system, both proteins can still conserve their individual functionality and biophysical properties. As described, a study 23 fused the NpHR from Natronomonas pharaonis with a light-gated inward cation channel ChR2 from Chlamydomonas reinhardtii, and when expressed in nerve cells, this fusion protein could deactivate and activate nerve fibers upon illumination of ~580 nm or ~480 nm of light, respectively. It should be noted that even though this study showed a Western blot to confirm the expression of the proteins in cells, it lacked further confirmation of direct protein purification, Abs-max determination, photocycle measurements, and direct in vitro functionality examinations. However, it clearly showed wavelength-dependent excitation or inhibition effect on nerve activity when expressed

Second, M-Rho proteins fused with HEBR retain their biophysical properties.
Our data concludes that HEBR shows no alteration in both the photocycle and Abs-max in both ion-type and sensory-type target M-Rho proteins. In the four HEBR fusion proteins studied, all individual target M-Rho conserved their Abs-max and photocycle. Therefore, it can be claimed that HEBR shows no negative effects toward the target fused proteins, at least in this study.
Third, the M-Rho proteins fused to HEBR are all still functionally intact. In addition, the light-driven ion transportation functions of BR and HR are conserved in the fused form. At the same time, HEBR does not disturb the otherwise unstable lone sensory type M-Rho. Interactions with the transmembrane region of their corresponding cognate transducers was known to stabilize sensory M-Rho proteins as reported in HsSRI 30,31 , NpSRII 32 and HmSRM 25 . These results suggest that the HEBR fusion M-Rho might increase or at least maintain the stability of the target proteins.
In summary, based on the success of our previous co-expression and fusion protein study, we combined the previous conclusions and applied them to our HEBR system. We show a universal strategy to overexpress two M-Rho proteins as a single fusion protein, which reaches ample expression efficiency for in vitro purification and investigation, while at the same time conserving the individual biochemical and biophysical properties of the two proteins. Therefore, these advantages increase the confidence to design multi-functional M-Rho using our HEBR system for further creative applications. coli C43(DE3) and E. coli C41(DE3) were used to express protein. The rhodopsin and transducer genes were amplified using PCR from the genomic DNA of H. marismortui as described previously 24 . Primers for the first transmembrane domain of HmHtrI (forward, 5′-ATATAAGCTTGCGTCGGCGTCGA ACGGCGCGTCGGCGATGACTATTTCTAGTGTC-3′, reverse, 5′-ATATCTCGAGTGCGGCCGCCATGGT GAGCTCGCGTCCCTCAATCGATGCAACATCTTGTGT-3′) were designed by introducing a HindIII restriction enzyme cutting site into the N-terminal and a NcoI, SacI, and XhoI restriction enzyme cutting site into its C-terminal. The DNA fragments were digested using HindIII-XhoI and ligated into the pET-21b(+) containing HEBR. The genes of HsBR and NpHR were previously constructed and digested using NcoI and XhoI. Primers for sensory rhodopsins, HmSRII (forward, 5′-ATATGAGCTCATGGCAACGATAACAACC-3′, reverse, 5′-ATATCTCGAGGTCCCCTGCAACCGCTGT-3′) and HmSRM (forward, 5′-ATATGAGCTCAT GGCACAAGAGATCGTT-3′, reverse, 5′-ATATCTCGAGCTTGGCGGGAGCTACGGA-3′) were designed by introducing a SacI restriction enzyme cutting site into the start codon and a XhoI restriction enzyme cutting site in opposition to the stop codon. The SacI-XhoI DNA fragments were digested and ligated into HEBR-HmHtrI contained in pET-21b(+). Consequently, the vector encoding six histidines at the C terminus, which resulted in the following N-and C-terminal peptide sequences: M-LEHHHHHH, in which the hyphens represent the protein sequence for the gene of interest, and the underlined amino acids represent the introduced XhoI site. The accuracy of all the constructs was confirmed using nucleotide sequencing.
Protein expression and purification. A single colony of transformed E. coli C43 (DE3) or E. coli C41(DE3) cells was inoculated in LB (Luria-Bertani) medium supplemented with 50 μg/mL of ampicillin and incubated at 37 °C overnight. For large-scale protein expression, a 1:100 (v/v) dilution of the overnight culture was added to fresh LB/ampicillin media and incubated at 37 °C. When the OD 600 of the culture reached 0.4-0.6, IPTG (isopropyl β -D-1-thiogalactopyranoside; final concentration 1 mM) and all-trans retinal (final concentration 5-50 μM) were added for induction. Following subsequent incubation for 4-6 h in the dark, the cells were collected using centrifugation at 6750 × g for 10 min at 4 °C (Hitachi CR-21, R10A3). The cells collected were resuspended in buffer A (50 mM Tris/HCl, 4 M NaCl, 14.7 mM 2-mercaptoethanol and 0.2 mM PMSF, pH 7.8) and broken by ultrasonic processing (S-3000; MISONIC). To separate the membrane fraction, total cell-extract centrifugation was performed at 6750 × g for 10 min at 4 °C (Hitachi CR-21, R20A2). The supernatant was centrifuged at 169,538 × g for 1 h at 4 °C (Hitachi CP80WX, P70AT). The sediment was dissolved in buffer B (buffer A supplemented with 1% DDM (n-dodecyl-β -D-maltoside) for at least 12 h at 4 °C, followed by centrifugation at 32816 × g for 45 min at 4 °C (Hitachi CR-21, R20A2) to separate the detergent-soluble fraction.  proteins were purified by affinity purification using the Ni-NTA (Ni2+-nitrilotriacetate) method. The detergent-soluble solution containing 20 mM imidazole was incubated with Ni-NTA agarose at 4 °C for 6-8 h on an orbital shaker. It was transferred to a chromatography column and washed with buffer C (buffer A with 0.05% DDM and 50 mM imidazole). The target proteins were eluted with buffer D (buffer A with 0.05% DDM and 250 mM imidazole). The purified proteins were concentrated and exchanged into buffer E (50 mM MES, 4 M NaCl and 0.05% DDM, pH 5.8) with a protein concentrator (Millipore, Amicon, cut-off size of 30 kDa).

UV/Vis spectrum analysis and stability analysis.
The absorption spectroscopy and stability analysis of proteins were measured using UV/Vis spectroscopy as described previously 26 . Briefly, the purified and concentrated proteins were diluted 100-fold into different pH value (KCl, citric acid, phosphate, Tris or NaHCO3, each at 100 mM concentration), salt concentration (7.81 mM to 4 M NaCl) and temperatures (4 °C to 64 °C) buffers under red dim light before absorption spectra measurements were determined (U1900 UV/Vis, Hitachi, Japan). The results were used to evaluate the stability of the proteins in different environments. All the measurements were performed at 25 °C.
Flash-laser-induced photocycle measurements. Flash-induced absorption transients were monitored using a flash-photolysis system designed by our group and described previously 25 . The flash laser was a Nd-YAG laser (532 nm, 6 ns pulse, 40 mJ). The purified proteins were suspended in buffer E (50 mM MES, 4 M NaCl and 0.05% DDM, pH 5.8) to reach 0.3 at OD λmax , and the transient absorbance changes were recorded at selected wavelengths. The curves represent the loss and recovery of the absorbance at the indicated wavelength upon green laser (532 nm) excitation. Transients were collected and averaged for each measurement. All the measurements were performed at 25 °C.

Light-driven photocurrent measurements.
A slight modification of the electrochemical cell designed by Chu et al. 33 was used for light-driven photocurrent measurements, and a detailed description can be found in our previous study 27 . Briefly, a 0.5-W 532-nm continuous laser was used to stimulate M-Rho located in the photochemical cell, which was composed in the following order: ITO-coated glass slide, sample chamber, dialysis membrane, blank solution chamber, and then another ITO-coated slide. The two ITO-coated slides were connected by a wire and a signal amplifier (SR570, Stanford Research Systems, Sunnyvale, CA) for photocurrent measurements. For each measurement, 64 trials were averaged. The purified proteins were dialyzed against a non-buffer solution (10 mM NaCl and 0.02% DDM).