The ring size of monocyclic ET‐1 controls selectivity and signaling efficiency at both endothelin receptor subtypes

Cardiovascular diseases (CVDs) like hypertension are a major cause for death worldwide. In the cardiovascular tissue, the endothelin system—consisting of the receptor subtypes A (ETAR) and B (ETBR) and the mixed agonist endothelin 1 (ET‐1)—is a major key player in the regulation of vascular tone and blood pressure. Tight control of this system is required to maintain homeostasis; otherwise, the endothelin system can cause severe CVDs like pulmonary artery hypertension. The high sequence homology between both receptor subtypes limits the development of novel and selective ligands. Identification of small differences in receptor–ligand interactions and determination of selectivity constraints are crucial to fine‐tune ligand properties and subsequent signaling events. Here, we report on novel ET‐1 analogs and their detailed pharmacological characterization. We generated simplified ET‐1‐derived monocyclic peptides to provide an accessible synthesis route. By detailed in vitro characterization, we demonstrated that both G protein signaling and the subsequent arrestin recruitment of activated ETBR remain intact, whereas activation of the ETAR depends on the intramolecular ring size. Increasing of the intramolecular ring structure reduces activity at the ETAR and shifts the peptide toward ETBR selectivity. All ET‐1 analogs displayed efficient ETBR‐mediated signaling by G protein activation and arrestin 3 recruitment. Our study provides in‐depth characterization of the ET‐1/ETAR and ET‐1/ETBR interactions, which has the potential for future development of endothelin‐based drugs for CVD treatment. By identification of Lys9 for selective labeling, novel analogs for peptide‐mediated shuttling by ET‐1 are proposed.

ET-1 is a crucial regulator of cardiovascular homeostasis. ET-1-mediated receptor activation leads to long-lasting vasoconstriction (ET A R activation) or vasodilatation (ET B R activation). 9,10 Dysregulation of the endothelin signaling axis can cause severe CVDs like congenital heart defects, pulmonary fibrosis, and pulmonary artery hypertension. [11][12][13] Thus, the endothelin system is an interesting target for the development of novel treatment options due to its limited distribution in the human body. Because the endothelin receptor subtypes share high sequence homology but convert vastly different physiological effects, the development of selective ligands is of great pharmacological interest. Precise pinpointing of the structural requirements for both ET A R and ET B R activation can reveal novel differences between these closely related GPCRs. This is vital to the development of novel and potentially selective agonists for each receptor. Previous structure-activity relationship (SAR) studies were able to reveal the importance of different ET-1 elements like the compact secondary structure due to intramolecular disulfide bonds. 14 Here, we report on the synthesis of different monocyclic ET-1 derivatives and ET B R-selective peptide agonists. By substitution of amino acid residues, one intramolecular disulfide bridge was selectively removed and the ring size of the remaining disulfide bridge was stepwise adjusted, keeping the characteristic structure of the wild-type (wt) endothelin peptides. The pharmacology of both ET A R and ET B R was investigated in response to the ET-1 derivatives. We applied detailed analysis of GPCR signaling by studying the G protein activation profiles and the arrestin recruitment and internalization of activated receptors. We demonstrate that the size of the Cys 1 -Cys 15 intramolecular disulfide acts as crucial selectivity determinant between the endothelin receptor subtypes. Increasing the size of this bond reduced ET A R activity, pushing the peptide selectivity toward the ET B R. Furthermore, by introducing a fluorescent label, the receptor-mediated internalization of the peptides into ET A R-and ET B R-expressing cells was monitored, identifying Lys 9 as potential attachment site for different cargos.

| Solid phase peptide synthesis
Peptides were synthesized by solid-phase peptide synthesis (SPPS), using the 9-fluorenylmethyloxycarbonyl (Fmoc)/tert-butyl (tBu) strategy. For automated SPPS, performed on a SYRO I peptide synthesizer, pre-loaded NovaSyn W-TGA resins (15-μmol scale), using an eightfold molar excess of N α -Fmoc-protected amino acids (and N α -Boc-protected Boc-Cys(Trt)-OH and Boc-Ala-OH at Position 1), Oxyma, and DIC, dissolved in DMF. All coupling steps were carried out twice with a reaction time set to 40

| Generation of the nanoluciferase-arrestin 3 construct
Overlap extension (OE) polymerase chain reaction (PCR) was used to fuse the Nluc to the N-terminus of arr3. The arr3 cDNA was amplified and modified with from the pcDNA-Rluc3-Arr3 plasmid. 16    Image processing was performed with Zeiss ZEN 2 software.  (Table 1) were synthesized by automated and manual SPPS.

| Synthesis of monocyclic ET-1 derivatives
First, the four cysteines were substituted by alanine, removing the bicyclic structure of the peptide. To further simplify the synthesis and increase the stability toward oxidation of peptide 1, we replaced Met 7 with isosteric norleucine. Next, we included one disulfide bridge (Cys 1 -Cys 15 ; peptide 2). For derivative 3, we increased the size of the intramolecular bridge from Cys 1 -Cys 15 by thioacetal formation ( Figure 1). We incorporated an additional methylene moiety between the sulfur atoms, forming the disulfide bond. To test whether the distance of the sulfur atoms is important for peptide-receptor interaction, we replaced Cys 1 with hCys (peptide 4), as well as Cys 1 and Cys 15 (peptide 5) to further increase the size of the disulfide bridge.
To visualize receptor-mediated peptide uptake into cells by fluorescence microscopy, peptides 1-5 were additionally labeled with 6-carboxytetramethylrhodamine (TAMRA) moieties (peptides 1F-5F) at Lys 9 to identify a potential modification site for cargo loading. The homogeneity of all synthesized peptides was assessed by RP-HPLC, and the molecular identity was verified by MALDI-ToF-and ESI-MS (Table S1 and Figures S1-S10). All peptides were biologically characterized and compared to wt ET-1 and the linearized ET-1 derivative (peptide 1).

| Identical folding properties of ET-1 derivatives in comparison to the wt peptide
Because all members of the endothelin peptides share high sequence identity, we used CD spectroscopy to estimate the structural similarities of the synthesized ET-1 analogs in comparison to ET-1 ( Figure 2).   Varying ring sizes were investigated by incorporation of different cysteine-derived residues. A ring size of four atoms was accomplished by disulfide formation between two cysteine residues, which was increased to five atoms by formation of a thioacetal. A five-atom linkage was also achieved by incorporation of a homocysteine (position 1) and was elongated to six atoms by homocysteines at positions 1 and 15. At position 7, a norleucine was introduced. K 9 was exploited for fluorophore attachment. Abbreviation: X = norleucine

| Lys 9 -A promising modification site for cargo attachment
To investigate the trafficking of the ET-1-derived peptides, fluorescently labeled analogs were used. Lys 9 , which is present in wt ET-1, was chosen for side-chain modification. We attached TAMRA at the Lys 9 N ε by SPPS proceedings, generating peptides 1F-5F. To validate the compatibility of this modification with receptor-ligand interaction, we determined the G protein activation mediated by the TAMRAlabeled analogs in inositol phosphate accumulation assays ( Figure 5).
Similar to the G protein activation of the nonfluorescent ET-1-analogs, no nanomolar ET A R activation was detectable for linear Abbreviations: n.d., not detected; SEM, standard error of the mean. Note: G protein activation was investigated in inositol phosphate accumulation assays (n ≥ 3), and arrestin 3 engagement was analyzed by BRET assay (n = 2). Data were normalized to ET-1 (wt) and are represented as mean value over all assay replicates (performed in triplicates) ± standard error of the mean (SEM).
peptide 1F, the thioacetal bridged peptide 3F, and the hCyscontaining peptides 4F and 5F. Derivative 2F exhibited an EC 50 of 1.7 nM, which was 2.5-fold reduced compared to the endogenous ligand ET-1 (0.7 nM; see Table 3) and approximately eightfold reduced compared to the nonfluorescent peptide 2 (0.2 nM) but fully activated the receptor. All TAMRA analogs activated the ET B R at low nanomolar concentrations. Surprisingly, the fluorescently labeled peptides were less potent to induce intracellular G protein signaling for both receptor species compared to the unlabeled peptides.
Next, we studied the cellular uptake of the TAMRA-labeled analogs 1F-5F into transiently transfected HEK293 cells ( Figure 6).
Because ET-1 was not available with a fluorescent label as external control peptide, we used peptide 1F as control for the internalization studies. In agreement with the G protein activation studies, this peptide did not induce internalization of the ET A R (as shown in Figure 6A), because of the low receptor activation at the applied con-   Note: Data were normalized to ET-1 (wt) and are represented as mean value over all assay replicates (performed in triplicates) ± standard error of the mean (SEM). n.d.   39,40 In the absence of agonist, membrane-embedded ET A R subpopulations remain at the cell surface, whereas ET B R subpopulations undergo constitutive internalization. 41,42 To discriminate between ligand-independent and ligand-dependent ET B R internalization and to verify cellular uptake of the ET-1 analogs, the TAMRA fluorophore was introduced into the peptides as orthogonal fluorescent reporter. The organic dye was linked to the side chain of the endogenous Lys 9 of ET-1 (peptides 1F-5F), which circumvents the need for additional amino acid substitution. The fluorescent ET-1 analogs 1F-5F displayed a minor loss of ET A R and ET B R activation potency compared to the respective nonfluorescent peptides (5-to 12-fold decrease concerning ET B R activation), but EC 50  CVDs is limited due to linear peptides being prone to enzymatic degradation in the blood stream. [64][65][66] Metabolic stabilization can be achieved by incorporation of unnatural amino acids and artificial intramolecular bridges, resistant toward degradation. 67 We propose that monocyclic peptides, displaying ET B R-selectivity, are promising vectors to combat CVDs due to the favorable physiological responses of the ET B R (e.g., ligand clearing). Furthermore, Lys 9 is a suitable position for drug attachment as demonstrated by fluorescence labeling.

| CONCLUSION
Here, we report on the development of new ET B R-selective peptides with minimal modifications compared to the wt agonist ET-1 and their detailed characterization with regard to endothelin receptor signaling.
Monocyclization of ET-1 by removal of the Cys 3 -Cys 11 disulfide bridge and increase of the size of the monocycle from Cys 1 -Cys 15 to hCys 1 -hCys 15 led to a loss of potency at the ET A R but kept activity at the ET B R. We established a methylene thioacetal formation of monocyclic ET-1 analogs to probe the acceptance of artificial bonds. Additionally, by introduction of a fluorescent label, we identified Lys 9 as a potential modification side for the attachment of fluorophores and other cargos in the future. Receptor-mediated peptide internalization in ET A R-and ET B R-expressing cells was studied by TAMRA-labeled compounds.