Strategy to Identify and Test Putative Light-Sensitive Non-Opsin G-Protein-Coupled Receptors: A Case Study.
An implicit assumption when studying eyes and other photoreceptive structures is that the initial step of light perception is mediated by a single protein subfamily, opsin. This is a reasonable starting point because, despite the morphological diversity and nonhomology of eyes, it is well documented that the majority of eumetazoans use an opsin-based phototransduction system (Feuda et al., 2012; Porter et al., 2012; Ramirez et al., 2016). However, this assumption may limit our understanding of the biological reality, as other proteins can play a role in photoreception. There are multiple examples of photoreception that employ non-opsin-based mechanisms to detect light, including a gustatory receptor (Gr) homolog in nematodes (Liu et al., 2010), cryptochrome in fly and mammals (Cashmore et al., 1999), cytochrome c oxidase in sponge larvae (Bjorn and Rasmusson, 2009), and photoactivated adenylyl cyclase in euglenoids (Ntefidou et al., 2003). Thus, when photoreception is being studied, an agnostic and unbiased approach should be applied to identify and characterize the initiator of the phototransduction cascade in non-model organisms. One such approach is expression profiling, or transcriptomics, which can capture all messenger RNA (mRNA) transcripts of a specific tissue under particular conditions or time points via RNA sequencing.
Whole-transcriptome shotgun sequencing, or RNA sequencing (herein RNA-seq), has revolutionized the way we study biology and the way we understand biological systems, including photoreception (Wang et al., 2009). One of the strengths of RNA-seq is that it allows the capture of all gene transcripts in a given sample. The transcripts are sequenced in small fragmerits, and these short reads (25-100 bp long) are assembled to the full-length, or near-full-length, reads of the mRNAs. Since the development of RNA-seq technologies, many sophisticated software programs have been developed to assemble a transcriptome de novo. The ability of de novo assembly is critical when working with non-model organisms that lack a reference genome or are distantly related to organisms that could be used as a reference (Grabherr et al., 2011; Schulz et al., 2012). After transcriptome assembly, the transcripts can be annotated automatically (e.g., BLAST; Altschul et al., 1990), and assigned a molecular function and a biological process using gene ontology (GO) (Conesa et al., 2005) or Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis (KEGG, 2017). These software packages and database resources create robust pipelines that have revolutionized the output of transcriptome analysis and the study of gene expression. In recent years, many studies have taken advantage of these pipelines and characterized in great detail transcriptomes associated with light perception or vision in non-model organisms (Pairett and Serb, 2013; Bok et al., 2014; Delroisse et al., 2014; Hering and Mayer, 2014; Henze and Oakley, 2015; Sander and Hall, 2015; Porath-Krause et al., 2016). In turn, this growing data set has made it possible to place annotated genes of the phototransduction cascade into specific functional classes with an evolutionary context (e.g., Speiser et al., 2014; Ramirez et al., 2016).
However, the same strength provided by the annotation of newly assembled genes into defined gene ontology classes may constrain the discovery of new genes and their functions. Because of its nature, transcriptome annotation relies on a biased approach, stemming from (1) the narrow taxonomic representation of different reference databases (e.g., National Center for Biotechnology Information [NCBI], UniProt, Swiss-Prot) and (2) experimental evidence for functional annotation based on only a few model organisms (e.g., GO and KEGG). As an example, the annotation of a transcript as an "opsin" is often based on its sequence similarity to class A/rhodopsin-like members of the G-protein-coupled receptor (GPCR) superfamily and the presence of a lysine residue "invariantly" located in the seventh transmembrane helix (TM7) (i.e., "active-site" lysine). This combination of traits is highly suggestive that the transcript produces an apoprotein able to bind to a retinal-based chromophore to form a photopigment capable of absorbing light. However, this limited view precludes the discovery of "new" genes with the function of interest. In the opsin example, if a transcript does not meet specific criteria described above, it excludes any possibility of discovering alternative modes of chromophore binding (i.e., binding sites outside of TM7) as well as the possibility that other non-opsin class A/rhodopsin-like receptors might be able to absorb light. Interestingly, recent work from Devine et al. (2013) on bovine rhodopsin (Bta-RHO) demonstrates that the active-site lysine can be moved to 4 other locations (position 90 and 94 in TM2, position 186 in the 2-stranded (3-sheet connecting TM4 and TM5 and position 293 in TM7) while still retaining the ability to form a photopigment with the retinal chromophore. These results indicate that the conserved location of active-site lysine in TM7 is not required for photosensitivity in GPCR proteins, and the current approaches to annotate transcriptomic data from eyes and other photoreceptive structures may miss novel light-sensitive proteins.
In this study, we propose a four-part strategy to discover novel light-sensitive proteins from transcriptomic data (Fig. 1). We use this pipeline to test the hypothesis that naturally occurring genes belonging to the class A/rhodopsin-like GPCR class possess an active-site lysine outside of the seventh alpha helix and are capable of binding to a retinal-based chromophore in photoreceptive structures. To test this hypothesis, we use the scallop as our model system for three reasons. First, they possess multiple light-sensing structures: vision-capable eyes and dermal photoreception along the edge of the mantle lobes, the membranous organ that lines the shells regulating the "shadow response" in molluscs (Kennedy, 1960; Wilkens, 2008). Second, the molecular components used in scallop photoreception are in the preliminary stages of identification through gene-targeted and transcriptomic sequencing (Kojima et al., 1997; Gomez et al., 2011; Pairett and Serb, 2013; Serb et al., 2013; Porath-Krause et al., 2016; Wang et al., 2017). Third, we leveraged the large number of unidentified transmembrane proteins (including GPCRs) from eye-specific transcriptomes in scallop (Pairett and Serb, 2013), suggesting a rich system for gene discovery.
In our case study, we followed the four-part strategy outlined in Figure 1. We first identified a set of 165 transcripts in the class A/rhodopsin-like GPCRs from transcriptomes of mantle, eye, and adductor muscle from the common bay scallop Argopecten irradians (Lamarck, 1819). We then made an alignment of the candidate GPCRs with Bta-RHO to confirm the presence of lysine in any of the six positions (the five proposed by Devine et al , plus the canonical lysine at position 296), reducing our candidate list to three transcripts. Using three-dimensional (3D) homology modeling, we demonstrated that only two of three candidates have a lysine directed toward the inner protein core, a requirement for chromophore attachment. We then expressed the two final candidates in our heterologous expression system. We showed for the first time that non-opsin photosensitive GPCRs may potentially exist in the animal kingdom.
Materials and Methods
Nomenclature and numbering system
Gene and protein nomenclature follow the rules described in Porath-Krause et al. (2016). Briefly, the first three letters of a gene name are determined by the first capital letter of the genus and the first two letters in lowercase of the species name. A hyphen separates the abbreviated Latin binomial from the alphanumeric code identifying the class (homolog) of the gene. The alphanumeric code is three to six characters long, italicized, and in lowercase for a gene or its transcript, while proteins use the same name but without italics and the alphanumeric code in all uppercase. For example, gene and protein nomenclature for bovine rhodopsin (GenBank NP_001014890.1) are indicated as Bta-rho and Bta-RHO, respectively. We use a relative numbering system for proteins, where the first amino acid position corresponds to the first methionine (start codon) of the given protein.
Bioinformatics and modeling
To identify class A/rhodopsin-like GPCR homologs, we screened tissue-specific transcriptomes (eyes, mantle, and adductor muscle) generated under dark and light conditions from the common bay scallop Argopecten irradians (for details on the transcriptomic data, see Porath-Krause et al., 2016). First, we conducted a BLAST search (Altschul et al., 1990) using full-length sequences of A. irradians [G.sub.q]-protein-coupled opsins (Air-OPNGqs) (Porath-Krause et al., 2016) to identify putative GPCR sequences. Those proteins with a lysine residue in TM7 were treated as an opsin and were removed from the list. Then, the remaining non-opsin candidates were blasted against the whole transcriptomic data set to generate the final list of putative class A/rhodopsin-like GPCRs. The resulting 165 sequences were aligned against Bta-RHO and Tpa-OPNGq1 (Todarodes pacificus, rhodopsin; GenBank CAA49906.1) using ClustalW in MEGA7 (Kumar et al., 2016). The final alignments were formatted using JalView (Waterhouse et al., 2009).
Based on Devine et al. (2013), we were interested in candidates that had a lysine residue at 1 of 5 sites homologous to the positions 90, 94, 117, 186, or 293 plus the canonical 296 position in Bta-RHO. Of the 165 sequences, only 3 candidates remained after this screen. We then bioinformatically characterized these candidate transcripts by identifying putative homologous genes and possible functions using the information generated by the Trinotate annotation suite (Grabherr et al., 2011): Air-CCAPR1-Like (similar to cardioacceleratory peptide receptor), Air-AAR1-Like (similar to allatostatin-A receptor), and Air-OX2R1-Like (similar to orexin receptor type 2). GenBank accession numbers for these sequences are Air-CCAPR1-Like, MF360796; Air-AAR1-Like, MF360797; Air-OX2R1-Like, MF360798. Next, we developed 3D models of the candidate proteins using the GPCR-I-TASSER server (Zhang et al., 2015). These models were superimposed on the 3D template of Bta-RHO (Okada et al., 2002), Protein Data Bank (PDB) entry 1L9H. For each candidate protein, a model was selected with the best C-score, a confidence score in the range of -5 to 2 for estimating the quality of the modeling process, where the higher the value, the higher the confidence in the model. All manipulations and image generation of the 3D models were performed using UCSF Chimera (Pettersen et al., 2004). Image files were created in Photoshop CC 2015.5.0 (Adobe, San Jose, CA). The flowchart template used to create Figure 1 is from Harryarts - Freepik (Graphic Resources, S. L., Malaga, Spain). Finally, we determined whether the relative expression levels for the two candidates differed among tissue types or across light treatments. Using the transcriptomic data from Porath-Krause et al. (2016), we calculated the expression levels of the candidate genes in fragments per kilobase of exon per million fragments mapped (FPKM) using the "built in" option of the Trinity assembly package (Grabherr et al., 2011).
Generation of expression constructs
As demonstrated by others (Terakita et al., 2008; Koyanagi et al., 2013), truncation of the C-terminus of long-tailed opsins improves expression without affecting their spectroscopic properties. Since the two candidates selected for in vitro study, Air-AAR1 -Like and Air-CCAPR1 -Like, have relatively long C-terminus tails, we followed the same strategy. Synthetic genes for Air-AAR1-Like[DELTA]58 and Air-CCAPR1-Like[DELTA]105 were produced from GenScript (Piscataway, NJ) with the last 58 and 105 amino acids, respectively, removed. The two synthetic constructs were inserted in plasmid p1D4-hrGFP II (Morrow and Chang, 2010) to generate pJS81 [p1D4-hrGFP II-PCMV: :hrAir-ccapr1-Like[DELTA]105; PCMV: :gfp] and pJS83 [p1D4-hrGFP II-PCMV: :hrAir-aar1-Like[DELTA]58; PCMW: :gfp]. The "hr" before the gene name means that codons of the synthetic gene were altered to match the most common sequence for that codon in human (humanization), which optimizes in vitro expression in HEK293T cells.
Plasmid purification for transient transfection was performed with QIAGEN (Hilden, Germany) HiSpeed Plasmid Maxi Kit following the manufacturer's instructions.
Site-directed mutagenesis was used to generate specific mutants of Air-aar1-Like[DELTA]58 and Air-ccapr1-Like[DELTA]105. Primers PR278 (5'-CTGCAGTTTCTGGGGATCGTGGTGCTG) and PR279 (CAGCACCACGATCCCCAGAAACTGCAG-3') were used to generate Air-ccapr1 -Like[DELTA]105K132G. We initially tried to generate Air-aar1-Like[DELTA]58K113G using primers PR280 (5'-ATCTGCGTGCCAGTGGGGGGCGTGACCTT C-3') and PR281 (5'-GAAGGTCACGCCCCCCACTGGCA CGCAGAT-3'), but we were unsuccessful, likely because the long stretch of 6 guanines in the forward primer that includes the "GGG" codon for glycine may be prone to secondary structures that interfere with the polymerase chain reaction (PCR) (Burge et al., 2006). We developed a new strategy and decided to generate Air-aar1 -Like [DELTA]58K113A instead. We designed 2 primers, PR286 (5'-ATCTGCGTGCCAGTGGAGGGCGTGAC CTTC-3') and PR287 (5'-GAAGGTCACGCCCTCCACTG GCACGCAGAT-3'), to generate an intermediate Air-aar1-Like[DELTA]58K113E mutant to facilitate the development of Air-aar1-Like[DELTA]58K113A. We used Air-aar1-Like[DELTA]58K113E as a template for a second round of mutagenesis to generate Air-aar1-Like[DELTA]58K113A with primers PR290 (5'-ATCTGCGTG CCAGTGGCGGGCGTGACCTTC-3') and PR291 (5'-GAA GGTCACGCCCGCCACTGGCACGCAGAT-3'). Air-aar1-Like[DELTA]58K205A was generated with a first round of PCR with primers PR292 (5'-GGCACCACAAGGAGGGCCTTCTGG TGC-3') and PR293 (5'-GCACCAGAAGGCCCTCCTTGT GGTGCC-3') to generate Air-aar1 -Like[DELTA]58K205R. This product was used as template with primers PR300 (5'-GTGGGC ACCACAAGGGGGGCCTTCTGGTGCCGC-3') and PR301 (5'-GCGGCACCAGAAGGCCCCCCTTGTGGTGCCCAC-3') to generate Air-aar1-Like [DELTA]58K205G. DNA Polymerase PfuTurbo (Agilent, Santa Clara, CA) was used for all mutagenesis experiments following the manufacturer's instructions with the following thermocycling profile: 95 [degrees]C for 2 min; 12 cycles of 95 [degrees]C for 30 s, 55 [degrees]C for 1 min, and 68 [degrees]C for 12 min; and 72 [degrees]C for 10 min, followed by a 2.5-h digestion with DpnI (New England Biolabs, Ipswich, MA) at 37 [degrees]C. Five microliters of the reaction was then used to transform TOP 10 chemically competent Escherichia coli cells (Thermo Fisher Scientific, Waltham, MA). Positive colonies were identified by Sanger sequencing. The following plasmids were generated by mutagenesis: pJS82 [p1D4-hrGFP II-PCMV: :hrAir-ccapr1-Like[DELTA]105K132G; PCMV: :gfp], pJS86 [p1D4-hrGFP II-PCMV: :hrAir-aarI-Like[DELTA]58K113A; PCMV: :gfp], and pJS88 [p1D4-hrGFP II-PCMV: :hrAir-aar1-Like[DELTA]58K205G; PCMV: :gfp].
Transient transfection, pull-down purification, and spectroscopy
To express the two GPCR proteins in vitro, we transiently transfected HEK293T cells (ATCC, Manassas, VA). Because protein expression levels differed between candidates, we transfected either 15 or 30 plates (Corning Falcon Standard Tissue Culture Dishes, 10 cm, ref. 353003; Tewksbury, MA) of confluent HEK293T cells per candidate protein with 8 [micro]g DNA and 20 [micro]l 293fectin Transfecting Reagent (Thermo Fisher Scientific) per plate following the manufacturer's instructions. Twenty-four hours after transfection, the medium was exchanged with new medium enriched with 5 [micro]mol [l.sup.-1] 11-cis retinal or 5 [micro]mol [l.sup.-1] all-trans retinal. Thus, all subsequent manipulations were performed under dim red light. Forty-eight hours after transfection, the cells were scraped from the plates 2 times with 5 ml buffer A (3 mmol [l.sup.-1] MgC12, 140 mmol [l.sup.-1] NaCl, 50 mmol [l.sup.-1] HEPES pH 6.6, aprotinin [10 mg m[l.sup.-1]], leupeptin [10 mg m[l.sup.-1]]). Cells were collected by centrifugation (10 min at 1620 relative centrifugal force [RCF]) and resuspended in 10 ml buffer A. This step was repeated twice.
After the second wash, the cells were resuspended in 2 ml per plate of buffer A with 5 [micro]mol [l.sup.-1] 11-cis retinal or 5 [micro]mol [l.sup.-1] all-trans to regenerate the photopigment. Cells were nutated for 1 h at 4 [degrees]C. The regenerated cells were collected by centrifugation at 38,360 RCF for 20 min and resuspended in solubilization buffer (buffer A plus 1 % n-dodecyl [beta]-D-maltoside and glycerol [20% w/v]) using 1 ml solubilization buffer per plate. The resuspended material was nutated for 1 h at 4 [degrees]C, and any remaining debris was separated by centrifugation for 20 min at 42,740 RCF. The supernatant was incubated with 100 [micro]l slurry resin (1:1 v/v resin/resin buffer) composed of 1D4 antibody (University of British Columbia, Canada) conjugated to sepharose beads. After 3 washes with 5 ml washing buffer (buffer A with 1% n-dodecyl [beta]-D-maltoside and glycerol [20% w/v] without aprotinin and leupeptin), the protein was eluted with 2 ml elution buffer (washing buffer with 40 [micro]mol [l.sup.-1] Rho1D4 peptide [TETSQVAPA]), adapted from Oprian et al. (1987). The eluate was concentrated to ~300 [micro]l using 4 Amicon Ultra 0.5-ml 10-kDa centrifugal filters (Millipore, Billerica, MA).
Ultraviolet-visible absorption spectra (250-750 nm) of purified proteins were measured at 15 [degrees]C using a Hitachi (Chiyoda, Tokyo, Japan) U-3900 spectrophotometer. Data analysis was performed on the mean value of five spectral measurements with the software UV Solutions, version 4.2 (Hitachi), and with R scripts (R Core Team, 2017). Interpolation of the data points was performed with the R function smooth.spline.
To test the proteins for photoreactivity, we first measured the "dark" absorbance of the protein, e.g., the naive protein that has been incubated and regenerated with the chromophore (11-cis retinal or all-trans retinal). We tested each protein independently with 11-cis and all-trans retinal because (1) based on opsin literature these 2 retinals are the preferential chromophore for most opsins and (2) different classes of opsins form stable pigments only in the presence of a particular retinal. For example, ciliary opsins generally favor 11-cis retinal, whereas retinochromes prefer all-trans retinal.
The maximum absorbance of the 2 retinals is very similar when the apoprotein is not present: 380 nm for all-trans retinal and 379 nm for 11-cis retinal in ethanol. Thus, any light-dependent isomerization converting free 11-cis retinal to all-trans retinal or vice versa will be undetectable in our experimental system. Therefore, we consider that the most plausible explanation for any observed change in spectral absorbance is due to a conformational change of the retinal covalently bonded to the apoprotein.
For the "light" spectra, we bleached the extracted proteins at different wavelengths and then measured the absorbance. Extracted proteins were first exposed to light at ~592 nm using 2 amber LEDs (MR16-YLX1-DI; superbrightleds.com, St. Louis, MO) simultaneously irradiating both transparent sides of the cuvette (Hellma Analytics 104002B-10-40; Mullheim, Germany) for 3 min. The absorbance (light spectrum) was recorded again. After this, the protein was exposed to light at ~474 nm using 2 blue LEDs (MR16-B24-15-DI; superbrightleds.com) and the absorbance recorded. A final exposure to light at ~405 nm using 2 violet LEDs (T1.5-UVHP; superbrightleds.com) was performed, followed by an absorbance measurement. The rationale behind the selection of these three wavelengths in this order was to start bleaching from the longest wavelength and move sequentially toward shortest wavelengths. This systematic approach makes it possible to "dissect" the absorbance of unknown putative lightsensitive proteins in discrete units (i.e., a light-sensitive protein may not respond to long wavelengths but show some response at medium-length wavelengths and even have an increased absorbance at shorter wavelengths). The differential spectra presented were mainly calculated from two adjacent light treatments (i.e., the absorbance spectrum after exposure to blue light was subtracted from the exposure to blue light) to minimize the possibility that any possible observed difference was in fact unrelated to the treatment but rather due to such external factors as handling of the probe or degeneration of the protein over time. Last, 2 [micro]l 12 N HC1 was added to the cuvette and the absorption measured of the acid-trapped form of the photosensitive GPCR.
Identification of two putative light-sensitive non-opsin GPCRs
We screened six transcriptomes generated in our lab (Porath-Krause et al., 2016) to identify putative GPCR genes that may have a chromophore-binding site in an alternative position to canonical opsins. This first screening (Fig. 1, point 1) provided a list of 165 putative GPCRs. We subsequently aligned all the candidates against Bta-RHO and Tpa-OPNGq 1 to test whether any of the candidates had a lysine at any of the five positions identified by Devine et al. (2013, 2016) (Fig. 1, point 2). Of the 165 candidates, 3 passed this screen and were considered for further analysis. The 3 candidates possessed lysine at sites located in TM2 and extracellular loop (EL) 2 (Air-AAR1-Like), TM3 (Air-CCAPR1-Like), or TM6 (Air-OX2R1-Like) (Fig. 2). Based on the results generated by the Trinotate annotation suite (Grabherr et al., 2011), we named the three candidates Air-CCAPR1-Like (similar to cardioacceleratory peptide receptor), Air-AAR1-Like (similar to allatostatin-A receptor), and Air-OX2R1-Like (similar to orexin receptor type 2).
To further refine our selection, we predicted the tertiary structure of the three candidates (Fig. 1, point 3). The model selected was based on the best C-score: Air-AAR1-Like, C-score = -2.46; Air-CCAPR1-Like, C-score = -3.93; and Air-OX2R1-Like, C-score = 0.45. In order to verify topologically whether the lysine was in a favorable position in the folded protein, we superimposed the predicted 3D structure of each of the 3 candidates on the Bta-RHO pdb entry 1L9H (Okada et al., 2002). The isomers of lysine 113 in Air-AAR1-Like and lysine 132 in Air-CCAPR1-Like were adjusted using Dunbrack's list of rotamers (Dunbrack, 2002). For lysine 113 in Air-AAR1-Like, we selected a rotamer with a probability of 0.220876, whereas for lysine 132 in Air-CCAPR1-Like a rotamer with a probability of 0.169323 was selected. As depicted in Figure 3, after rotamer adjustment, lysine 113 in Air-AAR1-Like and lysine 132 in Air-CCAPR1-Like were favorably positioned toward the interior of the chromophore-binding pocket. In contrast, lysine 302 of Air-OX2R1-Like was not in the proximity of the predicted chromophore pocket but rather on a very distant region of the sixth helix. Gene Air-ox2r1-like was therefore removed from further analysis.
Expression analysis of Air-ccapr1-Like and Air-aar1-Like
We then characterized the expression patterns of Air-ccapr1-Like and Air-aar1-Like among the tissue-specific (adductor, eye, and mantle) transcriptomes generated under two different light conditions (light or dark) (Fig. 1, point 1). Because these data are from a single biological replicate, differences in expression levels suggest only general trends rather than quantitative expression. We found that Air-ccapr1-Like and Air-aar1-Like have two different expression patterns. Air-ccapr1-Like is ubiquitously expressed in all three tissues. It has the lowest level of expression in the adductor muscle and the highest expression level in the photosensitive mantle (Fig. 4). Interestingly, Air-ccapr1-Like expression in the eye is relatively less than its expression in the mantle, and we could identify its transcript in the eyes only after light exposure (Fig. 4). Air-aar1-Like appears to be expressed exclusively in the mantle (Fig. 4).
Air-CCAPR1-Like and Air-AAR1-Like are photoreactive and respond differently to light after incubation with 11-cis retinal and all-trans retinal
To verify whether the two GPCRs were indeed able to form photopigments in a manner similar to opsin proteins, the two proteins were expressed in HEK293T cells and incubated with either 11-cis retinal or all-trans retinal. Previous work has shown that opsins with a long C-terminus express more efficiently in vitro if part of the C-terminus is removed (Terakita et al., 2008; Koyanagi et al., 2013). We therefore generated chimeric truncated constructs expressing Air-CCAPR1-Like[DELTA]105 and Air-AAR1-Like[DELTA]58, optimized for transcription and translation in HEK293T cells.
Both candidates seem to serve as apoproteins in photopigments. Incubation of Air-AAR1-Like[DELTA]58 with 11-cis retinal generated a photopigment that apparently reacted after exposure to blue light (~474 nm) but not amber light (~592 nm; not shown) (Fig. 5A). Upon incubation with all-trans retinal, Air-AAR1-like[DELTA]58 produced a dark pigment that clearly reacted to amber light (~592 nm) (Fig. 5B inset, spectrum 1) and had a pattern indicating bistability by first responding to blue light (~474 nm) (Fig. 5B inset, spectrum 2) and then again responding to violet light (~405 nm) irradiation (Fig. 5B inset, spectrum 3), which appears very similar to the initial darklight plot (Fig. 5B inset, spectrum 1).
The sinusoidal shape of the differential spectra of Air-CCAPR1-Like[DELTA]105 incubated with either 11-cis or all-trans retinal clearly indicates a change after exposure to light (Fig. 5C, D insets, spectra 1). This may be due to a change in the properties of the apoprotein, changes in the conformation of the chromophore, and/or changes in the interaction between the apoprotein and the chromophore. In addition, Air-CCAPR1-Like[DELTA]105 incubated with 11-cis retinal appears to possess a bistable nature, as demonstrated by the sinusoidal shape of spectrum 1, complementary to spectrum 2 (Fig. 5C, inset). On the contrary, we could not provide evidence of bistability for Air-CCAPR1-Like[DELTA]105 when incubated with all-trans retinal. Even after exposure to blue or violet light, the differential spectra after the first light bleaching remained flat (Fig. 5D inset, spectra 2 and 3).
Lysine 132 in Air-CCAPR1-Like and lysine 113 in Air-AAR1-Like are required for chromophore-mediated photoreaction
To determine whether the photoreactivty of Air-CCAPR1-Like and Air-AAR1-Like requires the presence of lysine 132 in Air-CCAPR1-Like[DELTA]105 and lysine 113 or lysine 205 in Air-AAR1 -Like[DELTA]58 (Fig. 3), we tested mutants where we substituted lysine for either glycine or alanine. We chose glycine or alanine because they are the same two amino acids chosen by Devine et al. (2013) for similar functional experiments in Bta-RHO. Initially, we wanted to mutate all lysines to glycine, but all attempts to convert the "AAG" triplet of lysine 113 in Air-AAR1-Like[DELTA]58 to "GGG" failed, and we decided to mutate it to alanine, "GCG." The mutants were tested with the chromophore that gave the best result for each of the two GCPRs: 11-cis retinal for Air-CCAPR1-Like[DELTA]105 and all-trans retinal for Air-AAR1-Like[DELTA]58. Air-CCAPR1-Like[DELTA]105K132G incubated with 11-cis retinal no longer reacted to light when compared to wild-type Air-CCAPR1-Like[DELTA]105 (Fig. 5C vs. Fig. 6A). Similarly, Air-AAR1-Like[DELTA]58K113A failed to react to light when compared to wild-type Air-AAR1-Like[DELTA]58 (Fig. 5B vs. Fig. 6B). In contrast, when lysine was replaced at position 205, Air-AAR1-Like[DELTA]58K205G maintained similar function to wild-type Air-AAR1-Like[DELTA]58 (Fig. 5B vs. Fig. 6C).
The diversity and sheer number of proteins belonging to the superfamily of GPCRs are a testament to their evolutionary success and biological importance. Thanks to advances in protein crystallography methods, significant progress has been made in recent years to catalog and describe the variety of known GPCRs (Fredriksson et al., 2003; Fredriksson and Schioth, 2005; Venkatakrishnan et al., 2013). "Deorphanization," the process of identifying the signaling molecule that "triggers" GPCR activation, has been demonstrated to be one of the most powerful in vitro tools to investigate the biological function of "orphan" GPCRs (Tang et al., 2012; Bauknecht and Jekely, 2015). Nevertheless, hundreds of GPCRs are still poorly characterized or not characterized at all. The lack of information is even more pronounced in non-model organisms, where often the only available information is derived by manual or automatic annotation of newly discovered GPCRs.
In this study, we asked whether concealed among the uncharacterized GPCRs of any given organism there are proteins that cannot be categorized as canonical opsin but can conjugate with a retinal chromophore to become photosensitive. Taking advantage of two recent studies demonstrating that in Bta-RHO the lysine can be relocated to four non-canonical positions (G90K, T94K, S186K, and F293K) and retain light absorbance (Devine et al., 2013, 2016), we performed a set of experiments to verify whether we could use this information to discover putatively light-sensitive nonopsin GPCRs. Initially, we did not have any particular expectations about the number and diversity of the possible GPCR candidates. Interestingly, of the three candidates that we identified, all showed a different placement of the putative chromophore-binding lysine: Air-CCAPR1-Like K132 (aligned to Bta-RHO A117); Air-OX2R1-Like K302 (aligned to Bta-RHO K296); and two positions for Air-AAR1-Like, K113 and K205 (aligned to Bta-RHO T94 and S186) (Fig. 2).
Three-dimensional protein modeling played an important role in visualizing and determining the possibility of an apoprotein-chromophore interaction. Notably, while zones of a multiprotein two-dimensional (2D) alignment may indicate homology between certain domains, the same alignment may not necessarily be useful in conveying significant information on the function of a given protein or the clear position of allosteric components. For example, we were able to exclude Air-OX2R1-Like from the candidate list because its lysine at position 302 was modeled to occur outside the inner core of the GPCR where the chromophore pocket would be located. In contrast, when adjusted for the rotamer (see Results), we have some confidence that lysine 132 in Air-CCAPR1-Like and lysine 113 in Air-AAR1 -Like were indeed oriented toward the inner core of the protein (Fig. 3), and it appears that the internal volume of the two proteins could harbor a retinal-based chromophore. Of course, the 3D modeling alone does not provide a detailed mechanical modeling of the interior of Air-CCAPR1-Like and Air-AAR1 -Like, so we cannot exclude that steric clashes with other amino acids protruding internally may prevent a retinal chromophore from docking. Last, it is worthwhile to note that Air-AAR1-Like possesses a second lysine at position 205 that may be suitable to form a Schiff base with the chromophore (Figs. 2, A1). The 3D model of Air-AAR1-Like has this lysine 205 depicted as being part of a protruding loop. This conformation could either reflect a true folding of the protein or be an artifact of the model and not representative of the residue's true location. These examples prove the usefulness of 3D modeling, in particular to remove false positives.
To capture a better insight at the in vivo distribution of Air-aar1-Like and Air-ccapr1 -Like, we examined their pattern of relative expression among the six transcriptomes (Fig. 4). Air-ccapr1-Like is expressed at different levels and at different conditions in all three tissues, including the adductor muscle, whereas Air-aar1-Like is absent in all tissues except the mantle. We interpret the small decrease in expression of Air-ccapr1 Like in the adductor muscle after exposure to light as an experimental fluctuation. Light exposure appears to be "turning on" transcription of this gene in eyes. In contrast, Air-aar1-Like maintains a relatively high expression in the mantle under both light treatments. Although these expression levels have been determined by a single RNA-seq experiment for each treatment and cannot provide any quantitative interpretation, these data suggest that the two GPCRs have a very different tissue compartmentalization that may be indicative of tissue-specific roles for each of them. Additional replicates using more sensitive quantitative methods are needed to determine whether the patterns reported here hold.
As validation that the two proteins could conjugate with a retinal chromophore and become light sensitive, we tested the remaining GPCR candidates in vitro (Fig. 1, point 4). Surprisingly, both proteins became photosensitive when incubated with 11-cis retinal and all-trans retinal (Fig. 5). We could not define the maximal absorbance of the two GPCRs; but we demonstrated that Air-AAR1-Like[DELTA]58 is bistable if incubated with all-trans retinal (Fig. 5B, inset), and Air-CCAPR1-Like[DELTA]105 is bistable if incubated with 11-cis retinal (Fig. 5C, inset). In the inset in Figure 5D, curves 2 and 3 demonstrate what we would expect if the proteins were not photosensitive, e.g., the differential spectra are flat.
The acid-trapping experiments we conducted on both GPCRs were inconclusive because of a very low signal: noise ratio; therefore, we could not exclude that the absorbance we observed was generated by the chromophore residing inside the chromophore pocket of Air-AAR1-Like[DELTA]58 and Air-CCAPR1-Like[DELTA]105 rather than an actual covalent bond to a lysine and a consequential formation of a Schiff base.
To explicitly test whether the lysines we had identified are necessary to make the two GPCRs light sensitive, we generated and tested mutants where we substituted the candidate lysine with a glycine or an alanine (Fig. 1, point 4). Remarkably, Air-CCAPR1-Like[DELTA]105K132G and Air-AAR1-Like [DELTA]58K113A failed to react to light compared to their wild-type counterpart (Fig. 5C vs. Fig. 6A and Fig. 5B vs. Fig. 6B). Conversely, Air-AAR1-Like[DELTA]58K205G kept its ability to photo-react (Fig. 6C). These results provide clear evidence that the photosensitivity of Air-CCAPR1 -Like and Air-AAR1 -Like requires a lysine at position 132 in Air-CCAPR1-Like and position 113 in Air-AAR1-Like[DELTA]58. However, both proteins are missing the glutamate counterion at the two homologous canonical positions in a true "opsin" protein (stars in Figs. 2, A1). This may indicate an inability to stabilize the chromophore after binding. It is therefore possible that what we observe in our spectra experiment is an artifact, independent of the fact that a covalent bond is effectively established between the apoprotein and the chromophore. Nevertheless, the circumstantial evidence that we provide for Air-AAR1-Like and Air-CCAPR1-Like (i.e., expression in eyes and/or light-sensitive mantle), combined with the light sensitivity after incubation with 11-cis retinal and all-trans retinal and loss of photosensitivity after removal of key lysines, points toward a putative light sensitivity of these two GPCRs.
If what we observe has a biologically relevant role, two different scenarios are possible. One scenario is where Air-AAR1-Like and Air-CCAPR1 -Like act as light-sensitive proteins in a similar manner to opsins. This suggests that nonopsin light-sensitive GPCRs may have evolved multiple times during the history of the GPCR superfamily. Under a second scenario, Air-AAR1-Like and Air-CCAPR1-Like may primarily or secondarily serve other GPCR-related roles, and the two proteins may have been evolutionarily co-opted to serve both functions. Further experimentation is required to confirm or falsify either of these hypotheses, including demonstrating that these GPCR proteins can transduce a light signal via a photo-transduction pathway. Future work could address this short-coming by identifying homologs and generating in vivo mutants in model organisms to test for light-dependent phenotypes or by identifying G protein partners and testing for their light-dependent activation. These data are important to provide the biological and molecular context of the two candidate GPCRs. Taken together, these results provide evidence that the workflow described in Figure 1 is a viable option for the discovery of putative light-sensitive non-opsin GPCRs and a starting point to formulate functional hypotheses.
We thank Autum Pairett for tissue preparation and transcriptome analyses, Rosalie Crouch (Storm Eye Institute, Medical University of South Carolina) and the National Eye Institute for supplying 11-cis-retinal, and Belinda Chang for providing the expression vector p1D4-hrGFP II. We are grateful to Don Sakaguchi for providing guidance, bench space, and the use of equipment; Brent Danielson for his ingenuity and encouragement; and Glene Mynhardt and Dalton Smedley for comments on an earlier draft. Finally, we thank Dan Speiser and Bill Kier for the invitation to participate in the virtual symposium "New Insights from Genetic Data Sets on the Function and Evolution of Visual Systems." This work was supported by the Carl A. and Grace A. Bailey Research Career Development Award and the College of Agriculture and Life Sciences at Iowa State University to JMS.
Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990. Basic local alignment search tool. J. Mol. Biol. 215: 403-410.
Bauknecht, P., and G. Jekely. 2015. Large-scale combinatorial deorphanization of Platynereis neuropeptide GPCRs. Cell Rep. 12: 684-693.
Bjorn, L., and A. Rasmusson. 2009. Photosensitivity in sponge due to cytochrome c oxidase? Photochem. Photobiol. Sci. 8: 755-757.
Bok, M. J., M. L. Porter, A. R. Place, and T. W. Cronin. 2014. Biological sunscreens tune polychromatic ultraviolet vision in mantis shrimp. Curr. Biol. 24: 1636-1642.
Burge, S., G. Parkinson, P. Hazel, A. Todd, and S. Neidle. 2006. Quadruplex DNA: sequence, topology and structure. Nucleic Acids Res. 34: 5402-5415.
Cashmore, A., J. Jarillo, Y. Wu, and D. Liu. 1999. Cryptochromes: blue light receptors for plants and animals. Science 284: 760-765.
Conesa, A., S. Gotz, J. M. Garcia-Gomez, J. Terol, M. Talon, and M. Robles. 2005. Blast2GO: a universal tool for annotation, visualization and analysis in functional genomics research. Bioinformatics 21: 3674-3676.
Delroisse, J., E. Ullrich-Liiter, O. Ortega-Martinez, S. Dupont, and M. Arnone. 2014. High opsin diversity in a non-visual infaunal brittle star. BMC Genomics 15: 1035.
Devine, E. L., D. D. Oprian, and D. L. Theobald. 2013. Relocating the active-site lysine in rhodopsin and implications for evolution of retinylidene proteins. Proc. Natl. Acad. Sci. U.S.A. 110: 13351-13355.
Devine, E. L., D. L. Theobald, and D. D. Oprian. 2016. Relocating the active-site lysine in rhodopsin. 2. Evolutionary intermediates. Biochemistry 55: 4864-4870.
Dunbrack, R. L. 2002. Rotamer libraries in the 21st century. Curr. Opin. Struct. Biol. 12:431-440.
Feuda, R., S. C. Hamilton, J. O. McInerney, and D. Pisani. 2012. Metazoan opsin evolution reveals a simple route to animal vision. Proc. Natl. Acad. Sci. U.S.A. 109: 18868-18872.
Fredriksson, R., and H. Schioth. 2005. The repertoire of G-proteincoupled receptors in fully sequenced genomes. Mol. Pharmacol 67: 1414-1425.
Fredriksson, R., M. C. Lagerstrom, L.-G. Lundin, and H. B. Schioth. 2003. The G-protein-coupled receptors in the human genome form five main families: phylogenetic analysis, paralogon groups, and fingerprints. Mol. Pharmacol 63: 1256-1272.
Gomez, M. D. P., L. Espinosa, N. Ramirez, and E. Nasi. 2011. Arrestin in ciliary invertebrate photoreceptors: molecular identification and functional analysis in vivo. J. Neurosci. 31: 1811-1819.
Grabherr, M. G., B. J. Haas, M. Yassour, J. Z. Levin, D. A. Thompson, I. Amit, X. Adiconis, L. Fan, R. Raychowdhury, Q. Zeng et al. 2011. Full-length transcriptome assembly from RNA-Seq data without a reference genome. Nat. Biotechnol. 29: 644-652.
Henze, M. J., and T. H. Oakley. 2015. The dynamic evolutionary history of pancrustacean eyes and opsins. Inregr. Comp. Biol. 55: 830-842.
Hering, L., and G. Mayer. 2014. Analysis of the opsin repertoire in the tardigrade Hypsibius dujardini provides insights into the evolution of opsin genes in Panarthropoda. Genome Biol. Evol. 6: 2380-2391.
KEGG (Kyoto Encyclopedia of Genes and Genomes). 2017. KEGG pathway analysis. [Online]. Available: http://www.genome.jp/kegg/ [2017, August 25].
Kennedy, D. 1960. Neural photoreception in a lamellibranch mollusc. J. Gen. Physiol. 44: 277-299.
Kojima, D., A. Terakita, T. Ishikawa, Y. Tsukahara, A. Maeda, and Y. Shichida. 1997. A novel [G.sub.o]-mediated phototransduction cascade in scallop visual cells. J. Biol. Chem. 272: 22979-22982.
Koyanagi, M., E. Takada, T. Nagata, H. Tsukamoto, and A. Terakita. 2013. Homologs of vertebrate Opn3 potentially serve as a light sensor in nonphotoreceptive tissue. Proc. Natl. Acad. Sci. U.S.A. 110: 4998-5003.
Kumar, S., G. Stecher, and K. Tamura. 2016. MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol. Biol. Evol. 33: 1870-1874.
Liu, J., A. Ward, J. Gao, Y. Dong, N. Nishio, H. Inada, L. Rang, Y. Yu, D. Ma, T. Xu et al. 2010. C. elegans phototransduction requires a G protein-dependent cGMP pathway and a taste receptor homolog. Nat. Neurosci. 13: 715-722.
Morrow, J. M., and B. S. W. Chang. 2010. The plD4-hrGFP II expression vector: a tool for expressing and purifying visual pigments and other G protein-coupled receptors. Plasmid 64: 162-169.
Ntefidou, M., M. Iseki, M. Watanabe, M. Lebert, and D. Hader. 2003. Photoactivated adenylyl cyclase controls phototaxis in the flagellate Euglena gracilis. Plant Physiol. 133: 1517-1521.
Okada, T., Y. Fujiyoshi, M. Silow, J. Navarro, E. M. Landau, and Y. Shichida. 2002. Functional role of internal water molecules in rhodopsin revealed by X-ray crystallography. Proc. Natl. Acad. Sci. U.S.A. 99: 5982-5987.
Oprian, D. D., R. S. Molday, R. J. Kaufman, and H. G. Khorana. 1987. Expression of a synthetic bovine rhodopsin gene in monkey kidney cells. Proc. Natl. Acad. Sci. U.S.A. 84: 8874-8878.
Pairett, A. N., and J. M. Serb. 2013. De novo assembly and characterization of two transcriptomes reveal multiple light-mediated functions in the scallop eye (Bivalvia: Pectinidae). PLoS One 8: e69852.
Petersen, E. F., T. D. Goddard, C. C. Huang, G. S. Couch, D. M. Greenblatt, E. C. Meng, and T. E. Ferrin. 2004. UCSF Chimera--a visualization system for exploratory research and analysis. J. Comput. Chem. 25: 1605-1612.
Porath-Krause, A. J., A. N. Pairett, D. Faggionato, B. S. Birla, K. Sankar, and J. M. Serb. 2016. Structural differences and differential expression among rhabdomeric opsins reveal functional change after gene duplication in the bay scallop, Argopecten irradians (Pectinidae). BMC Evol. Biol. 16: 250.
Porter, M. L., J. R. Blasic, M. J. Bok, E. G. Cameron, T. Pringle, T. W. Cronin, and P. R. Robinson. 2012. Shedding new light on opsin evolution. Proc. R. Soc. Biol. Sci. B 279: 3-14.
R Core Team. 2017. R: a language and environment for statistical computing. [Online]. R Foundation for Statistical Computing, Vienna. Available: http://www.R-project.org/ [2017, September 28].
Ramirez, M. D., A. N. Pairett, M. S. Pankey, J. M. Serb, D. I. Speiser, A. J. Swafford, and T. H. Oakley. 2016. The last common ancestor of bilaterian animals possessed at least 7 opsins. Genome Biol. Evol. 8: 3640-3652.
Sander, S. E., and D. W. Hall. 2015. Variation in opsin genes correlates with signaling ecology in North American fireflies. Mol. Ecol. 24: 4679-4696.
Schulz, M. H., D. R. Zerbino, M. Vingron, and E. Birney. 2012. Oases: robust de novo RNA-seq assembly across the dynamic range of expression levels. Bioinformatics 28: 1086-1092.
Serb, J. M., A. J. Porath-Krause, and A. N. Pairett 2013. Uncovering a gene duplication of the photoreceptive protein, opsin, in scallops (Bivalvia: Pectinidae). Integr. Comp. Biol. 53: 68-77.
Speiser, D. I., M. Pankey, A. K. Zaharoff, B. A. Battelle, H. D. Bracken-Grissom, J. W. Breinholt, S. M. Bybee, T. W. Cronin, A. Garm, A. R. Lindgren et al. 2014. Using phylogenetically-informed annotation (PIA) to search for light-interacting genes in transcriptomes from non-model organisms. BMC Bioinformatics 15: 350.
Tang, X., Y. Wang, D. Li, J. Luo, and M. Liu. 2012. Orphan G protein-coupled receptors (GPCRs): biological functions and potential drug targets. Acta Pharmacol. Sin. 33: 363-371.
Terakita, A., H. Tsukamoto, M. Koyanagi, M. Sugahara, T. Yamashita, and Y. Shichida. 2008. Expression and comparative characterization of Gq-coupled invertebrate visual pigments and melanopsin. J. Neurochem. 105: 883-890.
Venkatakrishnan, A. J., X. Deupi, G. Lebon, C. G. Tate, G. F. Schertler, and M. M. Babu. 2013. Molecular signatures of G-protcin-coupled receptors. Nature 494: 185-194.
Vogel, R., M. Mahalingam, S. Liideke, T. Huber, F. Siebert, and T. P. Sakmar. 2008. Functional role of the "ionic lock"--an intcrhclical hydrogen-bond network in family A heptahelical receptors. J. Mol. Biol. 380: 648-655.
Wang, Z., M. Gerstein, and M. Snyder. 2009. RNA-Seq: a revolutionary tool for transcriptomics. Nat. Rev. Genet. 10: 57-63.
Wang, S., J. Zhang, W. Jiao, J. Li, X. Xun, Y. Sun, X. Guo, P. Huan, B. Dong, L. Zhang et al. 2017. Scallop genome provides insights into evolution of bilaterian karyotype and development. Nat. Ecol. Evol. 1: doi:10.1038/s41559-017-0120.
Waterhouse, A., J. Procter, D. Martin, M. Clamp, and G. J. Barton. 2009. Jalview version 2--a multiple sequence alignment editor and analysis workbench. Bioinformatics 25: 1189-1191.
Wilkens, L. A. 2008. Primary inhibition by light: a unique property of bivalve photoreceptors. Am. Malacol. Bull. 26: 101-109.
Zhang, J., J. Yang, R. Jang, and Y. Zhang. 2015. GPCR-I-TASSER: a hybrid approach to G protein-coupled receptor structure modeling and the application to the human genome. Structure 23: 1538-1549.
DAVIDE FAGGIONATO AND JEANNE M. SERB (*)
Department of Ecology, Evolution, and Organismal Biology, Iowa State University, Ames, Iowa 50011
Received 4 April 2017; Accepted 12 September 2017; Published online 13 November 2017.
(*) To whom correspondence should be addressed. E-mail: email@example.com.
Abbreviations: Bta-RHO, bovine rhodopsin; EL, extracellular loop; FPKM, fragments per kilobase of exon per million fragments mapped; GO, gene ontology; GPCR, G-protein-coupled receptor; Gr, gustatory receptor; KEGG, Kyoto Encyclopedia of Genes and Genomes; mRNA, messenger RNA; TM, transmembrane helix; Tpa-OPNGq1, Todarodes pacificus, rhodopsin.
|Printer friendly Cite/link Email Feedback|
|Author:||Faggionato, Davide; Serb, Jeanne M.|
|Publication:||The Biological Bulletin|
|Date:||Aug 1, 2017|
|Previous Article:||Opsin Expression in the Central Nervous System of the Mantis Shrimp Neogonodactylus oerstedii.|
|Next Article:||Expression of G Proteins in the Eyes and Parietovisceral Ganglion of the Bay Scallop Argopecten irradians.|