Photodynamic Activation of Cholecystokinin 1 Receptor with Different Genetically Encoded Protein Photosensitizers and from Varied Subcellular Sites

Cholecystokinin 1 receptor (CCK1R) is activated by singlet oxygen (1O2) generated in photodynamic action with sulphonated aluminum phthalocyanine (SALPC) or genetically encoded protein photosensitizer (GEPP) KillerRed or mini singlet oxygen generator (miniSOG). A large number of GEPP with varied 1O2 quantum yields have appeared recently; therefore, in the present work, the efficacy of different GEPP to photodynamically activate CCK1R was examined, as monitored by Fura-2 calcium imaging. KillerRed, miniSOG, miniSOG2, singlet oxygen protein photosensitizer (SOPP), flavin-binding fluorescent protein from Methylobacterium radiotolerans with point mutation C71G (Mr4511C71G), and flavin-binding fluorescent protein from Dinoroseobacter shibae (DsFbFP) were expressed at the plasma membrane (PM) in AR4-2J cells, which express endogenous CCK1R. Light irradiation (KillerRed: white light 85.3 mW‧cm−2, 4’ and all others: LED 450 nm, 85 mW·cm−2, 1.5′) of GEPPPM-expressing AR4-2J was found to all trigger persistent calcium oscillations, a hallmark of permanent photodynamic CCK1R activation; DsFbFP was the least effective, due to poor expression. miniSOG was targeted to PM, mitochondria (MT) or lysosomes (LS) in AR4-2J in parallel experiments; LED light irradiation was found to all induce persistent calcium oscillations. In miniSOGPM-AR4-2J cells, light emitting diode (LED) light irradiation-induced calcium oscillations were readily inhibited by CCK1R antagonist devazepide 2 nM; miniSOGMT-AR4-2J cells were less susceptible, but miniSOGLS-AR4-2J cells were not inhibited. In conclusion, different GEPPPM could all photodynamically activate CCK1R. Intracellular GEPP photodynamic action may prove particularly suited to study intracellular GPCR.

CCK1R is unique among A class G protein-coupled receptors (GPCR) in that it is permanently activated ligand-independently by the lowest lying excited state molecular oxygen, the delta singlet oxygen (∆ 1 O 2 or 1 O 2 ), usually generated in type II photodynamic action with chemical photosensitizer sulphonated aluminum phthalocyanine (SALPC) after a brief cellular incubation [29,30] in freshly isolated rat pancreatic acini or with genetically encoded protein photosensitizer (GEPP) KillerRed or mini singlet oxygen generator (miniSOG) target-expressed at the plasma membrane (PM) in rat pancreatic acinar tumor cell AR4-2J [31,32].
A typical photodynamic action involves three elements: light, a light-absorbing molecule (photosensitizer), and molecular oxygen. After absorption of a photon of a certain wavelength by a photosensitizer, the excited state photosensitizer molecule eventually undergoes either electron transfer or the energy transfer process. Electron transfer leads to the production of oxygen radicals (type I), such as superoxide anion (O 2 −. ). Energy transfer to ground state molecular oxygen results in the production of 1 O 2 (type II) [29][30][31]. The absorption spectrum of the photosensitizer molecule determines the most effective wavelength of the irradiation light. The photophysicochemical property of the photosensitizer determines the 1 O 2 quantum yield (φ 1 O 2 ) in type II photodynamic action. In contrast to chemical photosensitizers such as porphyrins and phthalocyanines, genetic manipulations will ensure that GEPP could be targeted to specified cell types or subcellular organelles with high precision. KillerRed is the first major GEPP to emerge, initially thought to generate solely superoxide anion (O 2 −. ) [33][34][35] but has since been reported to probably generate 1 O 2 also [31,36], and later measured to have a 1 O 2 quantum yield (φ∆ 1 O 2 0.008) of more than eight-fold higher than O 2 −. [37].
In view of the above developments, it has become pertinent to examine whether the newly emerged GEPP could also be used for photodynamic CCK1R activation. Further, numerous works have reported that G protein-coupled receptors (GPCR) function not only from PM but, also, from intracellular membranes [60,61]. The role of intracellular miniSOG photodynamic action on CCK1R therefore also needs to be examined. Such works would potentially provide novel platforms upon which to elucidate the in vivo CCK1R function with high temporal and spatial precision.
In the present work, it was found that plasma membrane (PM)-expressed KillerRed, miniSOG, miniSOG2, SOPP, Mr4511 C71G , and DsFbFP in AR4-2J cells after light irradiation all photodynamically activated endogenous CCK1R in the rat pancreatic acinar tumor cell AR4-2J, triggering persistent calcium oscillations, with DsFbFP being the least effective. Interestingly, light irradiation of AR4-2J cells with miniSOG expressed at the plasma membrane (PM), mitochondria (MT), or lysosomes (LS) was found to trigger similarly persistent calcium oscillations, which were inhibited by CCK1R antagonist devazepide 2 nM with a graded sensitivity of PM > MT > LS. Therefore, GEPP expressed either at PM or intracellularly are both effective to photodynamically activate CCK1R, suggesting that photodynamic action may be particularly suited for the study of intracellular GPCR without the need for extracellularly added agonists to overcome multiple diffusion barriers.
Solid E. coli medium LB/kana and LB/amp were sterilized and culture plates made. Liquid E. coli medium LB/kana and LB/amp had the same composition but without agar.

Vector Constructs
Plasmid pKillerRed PM was bought from Evrogen (Moscow, Russia), proliferated in and harvested from competent E. coli. A mammalian codon-optimized miniSOG gene (GenBank accession number JX999997) was synthesized de novo from nucleotides at Genscript (Nanjing, China) with the following full sequence: ATGGAAAAGAGCTTTGTGATTACCGATCCGCGCCTGCCAGACAACC CGATCATTTTCGCGAGCGATGGCTTTCTGGAGTTAACCGAATATTCTCGTGAGGAAATTCTGGGT CGCAATGGCCGTTTCTTGCAGGGTCCGGAAACGGATCAAGCCACCGTGCAGAAAATCCGCGAT GCGATTCGTGACCAACGCGAAATCACCGTTCAGCTGATTAACTATACGAAAAGCGGCAAGAAA TTTTGGAACTTACTGCATCTGCAACCGATGCGCGATCAGAAAGGCGAATTGCAATATTTCATTG GTGTGCAGCTGGATGGCTAG. This synthesized full miniSOG gene sequence was inserted into plasmid pKillerRed PM (Evrogen, Moscow, Russia) to replace the KillerRed sequence. Competent E. coli were infected with the recombinant plasmid, cultured on solid LB/kana. Bacteria colonies were picked and further cultured in liquid LB/kana with shaking overnight. Proliferated plasmid was extracted with sequence verification. The plasmid so obtained was named pminiSOG PM due to the presence of the PM-localization sequence in the original Evrogen plasmid. After transfection with plasmid pKillerRed PM or pminiSOG PM , positive expressing AR4-2J cells were named KillerRed PM -or miniSOG PM -AR4-2J cells, as reported before [31,32].
For the construction of plasmid pminiSOG2 PM , the miniSOG2 [48] gene was synthesized de novo after rat codon optimization.
The miniSOG sequence in plasmid pminiSOG PM was replaced with the synthesized miniSOG2 sequence to obtain plasmid pminiSOG2 PM (Genscript, Nanjing, China).
For the construction of plasmid pSOPP PM , the SOPP amino acid sequence [50] was used. The SOPP gene was synthesized de novo after rat codon optimization. The miniSOG sequence in plasmid pminiSOG PM was replaced to obtain plasmid pSOPP PM (Genscript, Nanjing, China), where the SOPP gene sequence was: ATGGAGAAGAGCTTCGTGATCACCGACCCCAGGCTGCCTGACAACCCA ATCATCTTCGCCAGCGACGGCTTCCTGGAGCTGACCGAGTACTCCAGGGAGGAGATCCTGG GAAGGAACGGCCGGTTCCTGCAGGGACCCGAGACCGACCAGGCCACCGTGCAGAAGATC AGAGACGCCATCAGAGACCAGCGCGAGATCACCGTGCAGCTGATCAACTACACCAAGTCC GGCAAGAAGTTCTGGAACCTGCTGCACCTGCAGCCCATGCGGGACCAGAAGGGCGAGCTG CAGTACTTCATCGGCGTGCTGCTGGACGGCTAA. AR4-2J cells transfected with plasmid pSOPP PM as verified were named SOPP PM -AR4-2J cells.
The plasmid pminiSOG MT was prepared by replacing the PM-localizing sequence ATGCTGTGCTGTATGAGAAGAACCAAACAGGTTGAAAAGAATGATGAGGACCAAAAGATC in pminiSOG PM with the mitochondrial (MT)-targeting sequence (MTS: ATGTCCGTCCTGACGCC GCTGCTGCTGCGGGGCTTGACAGGCTCGGCCCGGCGGCTCCCAGTGCCGCGCGCCAAGAT CCATTCGTTGGGGGATCCACCGGTCGCCACC) (Genscript, Nanjing, China). pminiSOG LS was prepared by replacing the PM-localizing ATGCTGTGCTGTATGAGAAG AACCAAACAGGTTGAAAAGAATGATGAGGACCAAAAGATC sequence in pminiSOG PM with the lysosomal (LS) sequence (LTS: ATGAAGGGACAGGGAAGCATGGACGAGG GAACCGCCGACGAGAGGGCCCCCCTGATCCGGACC). Competent E. coli were infected with the plasmid, further cultured on solid LB/kana. Bacteria colonies were picked and cultured in liquid LB/kana with shaking overnight. Propagated plasmids were extracted for sequence verification. The plasmid constructs were designated pminiSOG MT and pminiSOG LS ; transformed AR4-2J cells were named miniSOG MT -AR4-2J and miniSOG LS -AR4-2J cells, accordingly.

Data Presentation and Statistical Analysis
All calcium tracings and other graphs were plotted with SigmaPlot. For calculation and comparison of the strength of induced calcium responses, the calcium peak area above the baseline was integrated (usually per 10 min, unless stated otherwise). Statistical data from N (as indicated) independent experiments were presented in bar graphs as mean ± SEM, unless specifically stated otherwise. Student's t-test was used for statistical analysis against controls, and p < 0.05 was taken as significant and indicated with an asterisk (*).

PM-Delimited
Photodynamic CCK1R Activation in AR4-2J Cells with GEPP KillerRed, miniSOG, miniSOG2, SOPP, Mr4511, and DsFbFP Plasmid pKillerRed PM , pminiSOG PM , pminiSOG2 PM , pSOPP PM , pMr4511 C71G PM , and pDsFbFP (for the full protein sequence of each of these GEPP, see Table A2) were conducted into AR4-2J cells, and the protein expression was confirmed by GEPP fluorescence imaging (Figure 1a-f). In comparison with the other GEPP (KillerRed, miniSOG, miniSOG2, SOPP and Mr4511 C71G ), DsFbFP demonstrated markedly dimmer fluorescence, suggesting poorer protein expression, as shown in the confocal images ( Figure 1f). RT-PCR experiments revealed that DsFbFP was abundantly expressed at the mRNA level after transfection with plasmid pDsFbFP PM (Figure 1g). The reason for the rather dim DsFbFP fluorescence will be further elaborated below and in Discussion.
Biomolecules 2020, 10, x 6 of 21 was integrated (usually per 10 min, unless stated otherwise). Statistical data from N (as indicated) independent experiments were presented in bar graphs as mean ± SEM, unless specifically stated otherwise. Student's t-test was used for statistical analysis against controls, and p < 0.05 was taken as significant and indicated with an asterisk (*). Plasmid pKillerRedPM, pminiSOGPM, pminiSOG2PM, pSOPPPM, pMr4511 C71G PM, and pDsFbFP (for the full protein sequence of each of these GEPP, see Table A2) were conducted into AR4-2J cells, and the protein expression was confirmed by GEPP fluorescence imaging (Figure 1a (Figure 2g). This weak calcium response combined with the dim fluorescent imaging shown in Figure 1 would suggest that DsFbFP indeed was conducted into AR4-2J cells, as verified by RT-PCR experiments (Figure 1g). A quantitative analysis of photodynamically-induced calcium spiking in each case found no statistically significant difference in photodynamic CCK1R activation with KillerRed, miniSOG, miniSOG2, SOPP, or Mr4511 C71G as the photosensitizer, but the effect was markedly diminished or weaker with DsFbFP PM as the photosensitizer (p < 0.05, Figure 2h).

PM-Delimited Photodynamic CCK1R Activation in AR4-2J Cells with GEPP KillerRed, miniSOG, miniSOG2, SOPP, Mr4511, and DsFbFP
Biomolecules 2020, 10, x 7 of 21 images (Figure 1f). RT-PCR experiments revealed that DsFbFP was abundantly expressed at the mRNA level after transfection with plasmid pDsFbFPPM (Figure 1g). The reason for the rather dim DsFbFP fluorescence will be further elaborated below and in Discussion.
The colored calcium tracings in (a-g) were each from individual cells from one of N (as indicated) identical experiments, ordinate being the fluorescence ratios in F340 / F380 and abscissa being time in min, as indicated. Integrated calcium peaks (per 10 min) in calcium tracings as represented in (a-g) from N The colored calcium tracings in (a-g) were each from individual cells from one of N (as indicated) identical experiments, ordinate being the fluorescence ratios in F 340 /F 380 and abscissa being time in min, as indicated. Integrated calcium peaks (per 10 min) in calcium tracings as represented in (a-g) from N experiments were calculated and plotted as a bar graph (h). The thin, pink, dashed horizontal line in (h) indicates the level of miniSOG photodynamic effect. No statistical difference was found between miniSOG and KillerRed (KR) or miniSOG2, SOPP, Mr4511 C71G , but the DsFbFP photodynamic effect was significantly less compared with miniSOG (p < 0.05), as indicated by an asterisk ( * ). CT: control-parental AR4-2J cells not transfected with any genetically encoded protein photosensitizer (GEPP).

miniSOG Photodynamic Action in AR4-2J Cells at PM, MT, or LS with Graded Sensivitivity towards Extracellular CCK1R Antagonist
Other than targeting miniSOG to the plasma membrane (PM) like previously reported [31,32], we now examined the miniSOG photodynamic action from the mitochondria (MT) or lysosomes (LS). The miniSOG was targeted to the mitochondria or lysosomes by transfection with vectors pminiSOG MT and pminiSOG LS (Figure 3a), with a mitochondrial localization sequence from human cytochrome C oxidase subunit VIII (MSVLTPLLLRGLTGSARRLPVPRAKIHSLGDPPVAT) [67,68] or with a lysosomal localization sequence from the C-terminal tail sequence of lysosomal-associated membrane protein 1 (LIMP II: KGQGSMDEGTADERAPLIRT) (LTS), respectively [69]. Twenty-four (24) hours after the transfection of AR4-2J cells with plasmid pminiSOG MT or pminiSOG LS , confocal imaging confirmed miniSOG expression in the mitochondria or lysosomes, as judged by their respective co-localization with MitoTracker or LysoTracker (Figure 3b,c).
Quantitative analysis of calcium responses in original tracings, as represented in Figure 3d-h, from N (3-7) identical experiments was calculated and plotted as bar graphs (i). Note the sharp difference in the calcium response after LED light irradiation in parental AR4-2J cells, and dark or light responses in MT or LS miniSOG-transfected AR4-2J cells (i).
In parallel experiments, it was found that calcium oscillations induced by LED light irradiation in miniSOG PM -AR4-2J cells (i.e., miniSOG PM photodynamic action at the plasma membrane) were inhibited nearly completely by the CCK1R antagonist devazepide 2 nM (Figure 4a). LED light irradiation-triggered calcium oscillations were inhibited in miniSOG MT -AR4-2J cells by the CCK1R antagonist devazepide 2 nM but to a lesser extent; a complete blockade was seen with increased devazepide at 10 nM (Figure 4b). LED light irradiation-induced calcium oscillations in miniSOG LS -AR4-2J cells, however, were not inhibited by devazepide 2 nM (Figure 4c).   identical experiments, the ordinate being fluorescence ratios in F340 / F380 and the abscissa being time in min, as indicated. Integrated calcium peaks (per 10 min) from N experiments were plotted, and statistically significant differences between before (Bef.) and during (Dur.) or after (Aft.) devazepide application were indicated by an asterisk (*) at p < 0.05 (d), for the devazepide inhibition of miniSOG photodynamically induced calcium oscillations with miniSOGPM or miniSOGLS. Note the thin pink (PM), cyan (MT), and red (LS) dotted horizontal lines (3) at the level before the devazepide application in miniSOGPM-AR4-2J, miniSOGMT-AR4-2J, and miniSOGLS-AR4-2J cells, respectively (d). In miniSOGMT-AR4-2J cells before devazepide 2 nM, the calcium response was 100%, during devazepide 2 nM, 65%, during devazepide 10 nM, 29%, and after devazepide, 106%. For better consistence, for miniSOGMT-AR4-2J cells, the statistical data in (d) were from all individual cells in one out of three identical experiments shown in (b).

Discussion
In the present work, it was found that PM-expressed KillerRed, miniSOG, miniSOG2, SOPP, Mr4511 C71G , and DsFbFP after light irradiation all photodynamically activated CCK1R to induce persistent cytosolic calcium oscillations in AR4-2J cells, but the photodynamic effect of DsFbFP was much reduced in comparison, likely due to poor protein expression. Permanent photodynamic CCK1R activation was achieved in AR4-2J cells by miniSOG expression not only at the plasma membrane (PM) but, also, in mitochondria (MT) and lysosomes (LS). Calcium oscillations induced by miniSOG photodynamic action at intracellular sites showed reduced sensitivity to inhibition by CCK1R antagonist devazepide 2 nM with the order of PM > MT > LS.
Although the DsFbFP protein level was low, as shown by the dim DsFbFP fluorescence, DsFbFP mRNA were expressed at sufficient levels in AR4-2J cells, as verified by RT-PCR experiments (Figure 1g). Although the DsFbFP fluorescence quantum yield (Φ Fluo = 0.35) [56] is the highest of all GEPP examined in the present work, DsFbFP is also known to be easily photobleached, photobleaching eight times faster than miniSOG (t bl50% miniSOG: 2.85 min and DsFbFP: 0.35 min) [65]. This was borne out by the fact that, although we were able to observe with the naked eye in transfected AR4-2J cells moderate DsFbFp fluorescence under the confocal microscope, photobleaching by the scanning laser light was so fast that it was difficult to capture the fluorescence image. However, the photobleaching of DsFbFP by the excitation light (340, 380 nm) might not be as significant during Fura-2 calcium imaging; therefore, we were able to observe distinct calcium spikes in DsFbFP PM -AR4-2J cells after blue LED irradiation (Figure 2g). DsFbFP (from Dinoroseobacter shibae) has been known to be expressed in CHO-K1 cells as a dimer, which might also affect its fluorescence, unlike the monomeric miniSOG, miniSOG2, or SOPP, all sourced from Arabidopsis thaliana [65,70]. The weak DsFbFP fluorescence in transfected AR4-2J cells is also likely a reflection of poor protein translation from mRNA, since the DsFbFP photodynamic activation of CCK1R was much diminished in comparison with miniSOG ( Figure 2). Additional code optimization for mammalian cell expression might improve DsFbFP protein expression. Alternatively, a heterodimeric construct similar to phiSOG [53] might facilitate DsFbFP fluorescence imaging or stabilize its photodynamic efficacy. miniSOG produces 1 O 2 after light irradiation in a type II photodynamic action [49,52,71], with an φ 1 O 2 of 0.03 (as reviewed in [43]). KillerRed was originally thought to be primarily a type I photosensitizer to generate mainly superoxide anion (O 2 −. ) [43], but more recently, it has been suggested that KillerRed in the cellular milieu almost certainly also generates 1 O 2 [31,36,43]. Photochemical measurements indeed revealed that KillerRed actually generates eight times more 1 O 2 than O 2 −. , with an φ 1 O 2 of 0.008, whilst monomeric KillerRed SuperNova has an φ 1 O 2 of 0.02 [37]. miniSOG2 is reported to be about eight times as efficient as miniSOG to produce reactive oxygen species [48]. The singlet oxygen protein photosensitizer (SOPP or miniSOG Q103L ) has a φ 1 O 2 of 0.25 [50]. Mr4511 C71G sourced from Methylobacterium radiotolerans has a φ 1 O 2 of 0.19 [57]. DsFbFP has the highest φ 1 O 2 among the lot at 0.33 [56]. Although these protein photosensitizers have varied values of φ 1 O 2 from 0.008 to 0.33 (see Table A1), no significant difference in their photodynamic activation of CCK1R (i.e., integrated calcium oscillations) was found except DsFbFP, which has the highest φ 1 O 2 value of 0.33 but the weakest photodynamic efficacy (Figure 2). This indicates that photodynamic CCK1R activation could be achieved with GEPP with wide-ranging values of φ 1 O 2 . However, care must be taken with protein photosensitizers with higher values of φ 1 O 2 to avoid photobleaching or possible activation by the ambient light during imaging or calcium measurements. It has been found by others that blue LED irradiation (450 nm, 0.8 W·cm −2 , 1 min) induced phototoxicity in a small percentage of cells two hours after light irradiation of miniSOG-HEK293, miniSOG2-HEK293, and SOPP-HEK293 cells, but no marked difference was found in HEK293 cells expressing SOPP (9%) with a φ 1 O 2 of 0.25 or miniSOG (11%) with a φ 1 O 2 of 0.03 [48]. The reason for the satisfactory performance by miniSOG with a φ 1 O 2 of only 0.03 to permanently activated photodynamically CCK1R in miniSOG PM -AR4-2J cells in the present work might be related to the progressive photochemical transformation of the fluorophore flavin mononucleotide (FMN) to lumichrome and the photo-oxidization of internal residues in miniSOG to significantly increase its φ 1 O 2 up to 10-fold [72]. In this regard, it is interesting to note that with AsLOV2 (not studied in the present work, and its φ 1 O 2 is not determined; see Table A1), the LOV2 domain of Avena sativa phototropin 1, light irradiation was found to induce progressive photochemical dissociation or the release of FMN from the AsLOV2 protein moiety, leading to significantly increased φ 1 O 2 [59].
Of the GEPP KillerRed, miniSOG, miniSOG2, SOPP, Mr4511 C71G , and DsFbFP examined in the present work, KillerRed is excited by the red component (585 nm) of the visible spectrum (full-spectrum white light was used in the present work), whereas all others by blue light (450 nm) ( Figure 2) (for the photophysical parameters of GEPP, see Table A2). Although the KillerRed structure is completely different from miniSOG and other flavin-binding protein photosensitizers, the most satisfactory performance by KillerRed reported in the present work ( Figure 2) might warrant further photochemical studies to examine possible increases in φ 1 O 2 during light irradiation.
For possible future in vivo applications, KillerRed might be better-suited due to its excitation by red light, instead of by blue light for miniSOG, miniSOG2, SOPP, Mr4511 C71G , and DsFbFP. It would be ideal if KillerRed could be subjected to further annotations, to shift its maximal excitation peak toward even longer wavelengths (a red shift), possibly by genetic code expansion [73], instead of the blue shifts observed in KillerOrange [46,47] and GreenSuperNova [45]. Although KillerRed is twice the size of miniSOG and larger than MR4511 C71G , DsFbFP (see Table A2), KillerRed fusion to CCK1R (CCK1R-KillerRed) were found to still result in, after light irradiation, the effective photodynamic activation of the in-frame CCK1R [31]. In the future, efficient red light-powered KillerRed photodynamic action might be applied to modulate other vital protein machines, such as the membrane molecular architecture of zymogen granules [74].
Other than the plasma membrane expression of miniSOG, mitochondrial (MT) or lysosomal (LS) miniSOG were found to photodynamically trigger persistent calcium oscillations similarly ( Figure 3). LED irradiation (450 nm, 85 mW·cm −2 , 1.5 min)-induced calcium oscillations were inhibited by CCK1R antagonist devazepide 2 nM significantly in miniSOG PM -CHO-K1 cells, slightly in miniSOG MT -AR4-2J cells, but not at all in miniSOG LS -AR4-2J cells (Figure 4), although LED light irradiation induced-calcium oscillations were significantly inhibited by devazepide 10 nM in miniSOG MT -AR4-2J cells (Figure 4). The calcium oscillations elicited by miniSOG photodynamic action at the plasma membrane and mitochondria were all likely due to CCK1R activation; at lysosomes, it could be due to CCK1R activation, but in lysosomes, activated CCK1R may not be inhibited by an antagonist, possibly due to the partial proteolysis of the receptor protein ( Figure 5). We believe that miniSOG MT photodynamic CCK1R activation might be due to 1 O 2 diffusion from the mitochondria to the plasma membrane at mitochondrion-plasma membrane contact sites, which have been found to be widely present [75][76][77]. Such mitochondrion-plasma membrane contact sites may be tethered by proteins such as Num-1 [75] and are nanometers across in size [78]; therefore, miniSOG MT photodynamically generated 1 O 2 with a diffusion distance of several tens of nanometers [43] could easily diffuse to the plasma membrane to oxidatively activate CCK1R in miniSOG MT -AR4-2J cells.
Biomolecules 2020, 10, x 13 of 21 may be tethered by proteins such as Num-1 [75] and are nanometers across in size [78]; therefore, miniSOGMT photodynamically generated 1 O2 with a diffusion distance of several tens of nanometers [43] could easily diffuse to the plasma membrane to oxidatively activate CCK1R in miniSOGMT-AR4-2J cells.
Although the emphasis in the present work was the subcellular localization of miniSOG expression and therefore of subcellular 1 O2 generation, due to its limited lifetime of 1 µs [83][84][85] the effective diffusion distance of 1 O2 in the cellular milieu has been suggested to be in the tens of nanometers or more (20-150 nm) [43,52,83,[86][87][88]. Therefore, photodynamically generated 1 O2 could diffuse from the plasma membrane (PM), mitochondrial (MT), or lysosomal (LS) membranes as the origin to within a circle with a radius of 20-150 nm. The lysosomal accumulation of endocytosized CCK1R and partial CCK1R degradation [62,[79][80][81] could possibly account for the miniSOG LS photodynamic CCK1R activation and the little inhibition afforded by CCK1R antagonist devazepide in miniSOG LS -AR4-2J cells. The insensitivity of lysosomal CCK1R to the antagonist devazepide 2 nM might also be due to limited accessibility of lysosomal CCK1R to extracellularly added devazepide 2 nM. Such a reduced sensitivity of intracellular GPCR to ligands was noted before for the nuclear membrane GPCR in cardiomyocytes, for example [82].
Although the emphasis in the present work was the subcellular localization of miniSOG expression and therefore of subcellular 1 O 2 generation, due to its limited lifetime of 1 µs [83][84][85] the effective diffusion distance of 1 O 2 in the cellular milieu has been suggested to be in the tens of nanometers or more (20-150 nm) [43,52,83,[86][87][88]. Therefore, photodynamically generated 1 O 2 could diffuse from the plasma membrane (PM), mitochondrial (MT), or lysosomal (LS) membranes as the origin to within a circle with a radius of 20-150 nm.
It may be noted that there is abundant evidence for GPCR expression, localization, and function at nuclear [89], mitochondrial [90,91], or other intracellular membranes, such as melanosomal membranes [92], in addition to GPCR transport from the endoplasmic reticulum (ER) to the Golgi apparatus, transport vesicles, PM, and then from the PM by endocytosis to endosomes/lysosomes [61,93,94]. Cardiomyocytes, for example, are endowed with multiple nuclear-membrane GPCR of the A class, such as α1-, β1-, β3-adrenergic receptors, and AT1R and AT2R angiotensin receptors [95][96][97]. The limited diffusion distance of 1 O 2 could well help to further investigate GPCR functions at these intracellular sites. Photodynamic GPCR activation/modulation might offer distinct advantages over conventional receptor pharmacology in that no ligand is needed for photodynamic activation after GEPP (KillerRed, miniSOG, miniSOG2, SOPP, Mr4511 C71G , and DsFbFP) expression at defined intracellular sites. Only light irradiation is required to permanently activate the intracellular GPCR. The limited diffusion distance of photodynamically generated 1 O 2 (20-150 nm) may ensure spatial precision and specificity.
In the present work, all GEPP examined were found to elicit persistent calcium oscillations (i.e., permanent CCK1R activation) photodynamically, either from the PM, MT, or LS. We have previously found that the SALPC photodynamic activation of CCK1R in rat pancreatic acini involved the near quantitative transformation of the CCK1R protein dimer to the monomer [98]. In addition, CCK-induced CCK1R monomerization in purified membrane proteins under sub-threshold SALPC photodynamic action closely followed the CCK dose response curve for amylase secretion in intact rat pancreatic acini, especially in the low, physiological range of CCK concentrations [98]. It would be interesting to see whether GEPP photodynamic CCK1R activation would also involve receptor monomerization. Future works should help to identify the essential structural motif(s) for the permanent 1 O 2 activation of CCK1R.

Conclusions
In conclusion, representative GEPP (KillerRed, miniSOG, miniSOG2, SOPP, Mr4511 C71G , and DsFbFP) reported in the literature were found to photodynamically activate the endogenous CCK1R in AR4-2J cells after plasma membrane expression. The miniSOG expression at intracellular sites was also found to induce persistent calcium oscillations or CCK1R activation ( Figure 5). The present work provides an effective means to activate CCK1R photodynamically, with the potential for in vivo applications in peripheral physiology and in central nervous system functional studies. Photodynamic activation might also prove suitable for the study of intracellular GPCR, which will involve no ligand additions in the extracellular fluid, therefore overcoming the diffusion barrier imposed by the plasma membrane and without the need for long-distance ligand diffusion through the cytosol. The present work might imply that it is now possible to examine the functional status of the CCK1R protein at different maturation and proteolysis stages in the protein trafficking (ER → Golgi apparatus → transport vesicles → PM) and degradation (endosomes, lysosomes, and proteosomes) pathways.

Conflicts of Interest:
The authors declare no conflict of interest.  The color code corresponds to the translated sequences of: the plasma membrane localization sequence, KillerRed, miniSOG, miniSOG2, SOPP/miniSOG Q103L , Mr4511 C71G , DsFbFP, and the mitochondrial/lysosomal localization sequence. Note that the Evrogen pKillerRed plasmid does not encode the initial two residues MG.