Point-Substitution of Phenylalanine Residues of 26RFa Neuropeptide: A Structure-Activity Relationship Study

26RFa is a neuropeptide that activates the rhodopsin-like G protein-coupled receptor QRFPR/GPR103. This peptidergic system is involved in the regulation of a wide array of physiological processes including feeding behavior and glucose homeostasis. Herein, the pharmacological profile of a homogenous library of QRFPR-targeting peptide derivatives was investigated in vitro on human QRFPR-transfected cells with the aim to provide possible insights into the structural determinants of the Phe residues to govern receptor activation. Our work advocates to include in next generations of 26RFa(20–26)-based QRFPR agonists effective substitutions for each Phe unit, i.e., replacement of the Phe22 residue by a constrained 1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid moiety, and substitution of both Phe24 and Phe26 by their para-chloro counterpart. Taken as a whole, this study emphasizes that optimized modifications in the C-terminal part of 26RFa are mandatory to design selective and potent peptide agonists for human QRFPR.


Introduction
The neuropeptides 26RFa (QRFP26) and its N-terminal extended form 43RFa (QRFP, Figure 1) are the endogenous ligands of the G i/s/q coupled QRFP receptor (QRFPR), formerly known as the orphan receptor GPR103 [1 -3]. Since the isolation of 26RFa from a European green frog Rana esculenta brain extract [4], prepro-26RFa cDNA has been characterized in numerous genomes of diverse vertebrates from fish to mammals [5]. However, very few mature peptides have been identified so far, leaving the 26RFa/QRFP precursor post-translational processing uncertain. Indeed, in avians, 26RFa orthologs have been isolated and sequenced from the Japanese quail [6] and zebra finch [7] while, in rodents, only the 43RFa counterpart has been characterized from rat [2,8]. In humans, both 26RFa-and 43RFa-immunoreactive forms have been detected in hypothalamic and spinal cord extracts [9]. Interestingly, the presence of a tribasic cleavage site (Arg-Lys-Arg/Lys) within the 26RFa sequence suggests that the C-terminal heptapeptide GGFSFRF-NH 2 , named 26RFa (20)(21)(22)(23)(24)(25)(26) , could also be produced from the precursor. However, to date, this fragment, whose sequence is highly conserved in tetrapods, has not been identified as a mature product of the 26RFa/QRFP gene transcription. In rats, the 26RFa/QRFP gene is primarily expressed in hypothalamic neurons which project in various brain areas [1, 2,8,10]. Consistent with the widespread distribution of 26RFa/QRFP-containing fibers, 26RFa and/or QRFP are involved in the regulation of multiple physiological activities [5]. In particular, it is now firmly established that 26RFa and QRFP stimulate food intake in rodents [1,8,[10][11][12][13][14][15][16]. There is also evidence that 26RFa and/or QRFP regulate several other neuroendocrine and cognitive functions including reproduction [17,18], anxiety [19], memory [20], cardiovascular activity [8] and nociceptive transmission [21][22][23], some of which depend, at least in part, on off-target interaction with QRFPR-related receptors [24,25]. In addition, recent data reveal that the 26RFa/QRFPR system controls glucose homeostasis at the periphery by increasing insulin sensitivity and inhibiting hepatic glucose production [26][27][28][29]. Thus, QRFPR positions as an attractive target for the development of innovative drugs. Indeed, there is now clinical evidence for possible indications of QRFPR ligands for the treatment of metabolic disorders, obesity and diabetes [5].

Impact of Modifications of Each Phenylalanine Residue
It is well-accepted that the (hetero)aromatic units of peptides and proteins are master residues for DNA recognition [41], folding [42] and receptor-ligand interaction [43]. In particular, His, Phe, Trp and Tyr contribute to three types of non-covalent interactions including π-hydrogen bonds, electrostatic cation-π interactions and van der Waals π-π interactions [44] for governing molecular recognition of a ligand into the binding site of the receptor and transduction processes.
Thus, the aim of the present study was to investigate molecular diversity at the unexplored Phe positions to go deeper into the structure-activity relationships of 26RFa for the design of potent hQRFPR agonists.

Impact of Modifications of Each Phenylalanine Residue
It is well-accepted that the (hetero)aromatic units of peptides and proteins are master residues for DNA recognition [41], folding [42] and receptor-ligand interaction [43]. In particular, His, Phe, Trp and Tyr contribute to three types of non-covalent interactions including π-hydrogen bonds, electrostatic cation-π interactions and van der Waals π-π interactions [44] for governing molecular recognition of a ligand into the binding site of the receptor and transduction processes.
Compounds inactive as agonists were evaluated as possible antagonists of 26RFaevoked calcium increase in cultured G α16 -hQRFPR-transfected CHO cells. None of the members of this series were able to reverse the stimulatory effect of 10 −7 M 26RFa on [Ca 2+ ] i , except LV-2068 (11) and LV-2188 (39) which antagonized 45-46% of the agonistic response with, respectively, modest and very low IC 50 (Table 2, Figure 4).
Compounds inactive as agonists were evaluated as possible antagonists of 26RFaevoked calcium increase in cultured Gα16-hQRFPR-transfected CHO cells. None of the members of this series were able to reverse the stimulatory effect of 10 −7 M 26RFa on [Ca 2+ ]i, except LV-2068 (11) and LV-2188 (39) which antagonized 45-46% of the agonistic response with, respectively, modest and very low IC50 (Table 2, Figure 4). To summarize, the three phenylalanine units exhibit differential contributions to the biological activity of 26RFa (20)(21)(22)(23)(24)(25)(26). The Phe 22 residue appears to be the most permissive as it tolerated N-substitution like in [Tic 22 ]26RFa(20-26) (LV-2211, 28) which was 5-folds more potent than the lead peptide. Conversely, most modifications of Phe 24 and Phe 26 did not improve the activity with the exception of para substituents of lesser size than that of a tertbutyl group.  (20)(21)(22)(23)(24)(25)(26) (LV-2194, 19) were 2-, 3.5-and 3.3-fold more potent than 26RFa (20)(21)(22)(23)(24)(25)(26) (LV-2021, 1). Although the amide NH of Phe 24 residue was probably not involved in H-bond, our results suggest that some flexibility of the peptide backbone is required at this point. Noteworthy, not only the H-bond capacity, but other features were also impacted by these modifications such as the side chain presentation geometry, amide cis/trans isomerization equilibrium, and/or β-sheet potential of the analog with a wide range of steric, electronic and hydrophobic characteristics.

Calcium Mobilization Assays
Changes in intracellular Ca 2+ concentrations induced by 26RFa (20)(21)(22)(23)(24)(25)(26) analogs in CHO-G α16 -hQRFPR-transfected cells were measured on a benchtop scanning fluorometer Flexstation III (Molecular Devices, Sunnyvale, CA, USA) as previously described [36][37][38]55,56]. Briefly, 96-well assay black plates with clear bottoms (Corning international, Avon, France) were seeded at a density of 40,000 cells/well 24 h prior to assay. For profiling agonistic experiments, cells were loaded with 2 µM Fluo-4 acetoxymethyl ester (AM) (Invitrogen) for 1 h in the presence of 0.01% pluronic acid, washed thrice, and incubated for 30 min with standard HBSS containing 2.5 mM probenecid and 5 mM HEPES. Compounds to be tested were added at final concentrations ranging from 10 −11 to 10 −5 M in HBSS, and the fluorescence intensity was measured during 3 min. To evaluate the antagonistic potency of the test compounds, cells were incubated with each compound over 15 min after Fluo-4 AM loading. Then, during fluorescence recording, a pulse of 26RFa was administered at a final concentration of 10 −7 M. After subtraction of the mean fluorescence background, the baseline was normalized to 100%. Fluorescence peak values were determined for each concentration of compound.

Statistical Analysis
Calcium experiments were performed in triplicate, and data, expressed as mean ± SEM of at least three distinct experiments, were analyzed with the Prism 8.0 software (Graphpad Software, San Diego, CA, USA). EC 50 and the IC 50 values were determined from concentration-response curves using a sigmoidal dose-response fit with variable slope from at least three independent determinations. Differences between 26RFa (20)(21)(22)(23)(24)(25)(26) and analog activities were analyzed by the Mann-Whitney test. p values < 0.05 were considered significant.

Nomenclature of Targets and Ligands
All targets and ligands used throughout this manuscript conform with the guidelines outlined by the International Union of Basic and Clinical Pharmacology and British Pharmacological Society (IUPHAR/BPS) Guide to Pharmacology [5,57].

Conclusions
The C-terminal extremity of 26RFa was previously identified as a pivotal region to modulate its signaling at hQRFPR [36]. In this work, the Phe 22 , Phe 24 and Phe 26 residues of 26RFa (20)(21)(22)(23)(24)(25)(26) were modified using series of point substitutions with natural, side chainconstrained, side chain-modified residues and peptidomimetic building blocks. Subtle chemical modifications in the sequence led to significant improvement in compound potency to activate hQRFPR. As such, a single modification of Phe 22 with the steric restricted Tic residue decreased the EC 50 value from 1640 ± 259 to 327 ± 170 nM, providing the most efficient modification at this sequence position. Anchored substituents in the para position of each Phe unit were the most versatile investigated modification. Thereby, the introduction of a para-chloro-phenylalanine in place of the native Phe 24 moiety emerged as a favorable replacement. This finding follows the same trend observed in [Pcp 26 ]26RFa (20)(21)(22)(23)(24)(25)(26) , highlighting a limited room to accommodate aromatic residues around the Phe aryl group. Although combination of multiple point-effective modifications does not necessarily translate into an additive or synergic effect, we have explored each position of 26RFa (20)(21)(22)(23)(24)(25)(26) for complete optimization of its sequence. Future challenges will be to convert several of these point modifications to low molecular weight 26RFa analogs for metabolic disorder, obesity or diabetes therapies. We are confident that, by utilizing subtle amendments, we can design 26RFa (20)(21)(22)(23)(24)(25)(26) -based compounds with nanomolar potency, functional selectivity and in vivo bioavailability.