Development of High Affinity Calcitonin Analog Fragments Targeting Extracellular Domains of Calcitonin Family Receptors

The calcitonin and amylin receptors (CTR and AMY receptors) are the drug targets for osteoporosis and diabetes treatment, respectively. Salmon calcitonin (sCT) and pramlintide were developed as peptide drugs that activate these receptors. However, next-generation drugs with improved receptor binding profiles are desirable for more effective pharmacotherapy. The extracellular domain (ECD) of CTR was reported as the critical binding site for the C-terminal half of sCT. For the screening of high-affinity sCT analog fragments, purified CTR ECD was used for fluorescence polarization/anisotropy peptide binding assay. When three mutations (N26D, S29P, and P32HYP) were introduced to the sCT(22–32) fragment, sCT(22–32) affinity for the CTR ECD was increased by 21-fold. CTR was reported to form a complex with receptor activity-modifying protein (RAMP), and the CTR:RAMP complexes function as amylin receptors with increased binding for the peptide hormone amylin. All three types of functional AMY receptor ECDs were prepared and tested for the binding of the mutated sCT(22–32). Interestingly, the mutated sCT(22–32) also retained its high affinity for all three types of the AMY receptor ECDs. In summary, the mutated sCT(22–32) showing high affinity for CTR and AMY receptor ECDs could be considered for developing the next-generation peptide agonists.


Introduction
Calcitonin (CT) a 32 amino acid peptide hormone is secreted from thyroid glands and activates the calcitonin receptor (CTR) to control calcium homeostasis. CTR is a drug target for osteoporosis treatment. Salmon CT (sCT) was developed as a peptide drug targeting human CTR due to its higher affinity and potency than human CT [1][2][3]. Interestingly, CTR can form a complex with an accessory protein called receptor activity-modifying protein (RAMP). The CTR:RAMP complexes gain affinity for the peptide hormone amylin and are known as the amylin receptor (AMY receptor). AMY receptor activation regulates blood glucose levels by reducing food intake, inhibiting glucagon secretion, and slowing gastric emptying [4]. Apparently, the AMY receptor is a drug target for diabetes treatment and also its activation holds a potential for treating other metabolic diseases including obesity [5][6][7]. A rat amylin analog pramlintide was developed and is available in clinics to treat diabetes as co-therapy with insulin [8]. In addition, lots of effort has been focused on developing next-generation peptide drugs targeting AMY receptors as exemplified with dual amylin calcitonin receptor agonists (DACRA) and long-acting amylin/calcitonin receptor agonists [9][10][11][12][13].
CTR belongs to class B G protein-coupled receptors (GPCR) that are characterized by a transmembrane domain (TM) and a large extracellular domain (ECD). The N-terminal half of CT binds CTR TM triggering G protein association and initiating cell signaling [14]. The CTR ECD is an important peptide binding site for the C-terminal half of sCT as reported in crystal structures [15,16]. Several studies have suggested the mechanisms of peptide interactions with CTR [14,15,[17][18][19]. The cryo-EM structure of CTR indicated Dulbecco's Modified Eagle Medium (DMEM) with 4.5 g/L glucose, L-glutamine, and sodium pyruvate was obtained from Corning (Mediatech, Inc., Manassas, VA, USA) for HEK293T and HEK239S GnTI − mammalian cell culture. The mixture of non-essential amino acids (NEAA, 100X) was purchased from Lonza (Basel, Switzerland). Fetal bovine serum (Cat.# F2442) was purchased from Sigma-Aldrich (St. Louis, MO, USA). DNA assembly master mix and restriction enzymes used for DNA cloning were purchased from New England Biolabs (Ipswich, MA, USA). All other reagents were purchased from Sigma-Aldrich, unless otherwise noted.

Cell Lines Used
HEK293T cells were purchased from ATCC (Manassas, VA, USA) for the expression of CTR ECD, RAMP1-CTR ECD fusion, and RAMP2-CTR ECD fusion proteins. HEK293S GnTI − cells were also purchased from ATCC (Manassas, VA, USA) for RAMP3-CTR ECD fusion protein expression.

Expression and Purification of CTR ECD, RAMP1-CTR ECD Fusion, and RAMP2-CTR ECD Fusion Proteins
The general procedures of the receptor ECD expression from HEK293T cells were previously described [18]. Briefly, HEK293T cells transiently transfected with DNA expression vectors were incubated for 4 days at 37 • C. Cell culture media were collected and they proceeded to purification steps. All protein purification was performed at 4 • C unless otherwise noted. The cell culture media were initially dialyzed to the dialysis buffer and were followed by immobilized metal affinity column (IMAC) chromatography and size exclusion column (SEC) chromatography. The procedures of these column chromatography used in this study were previously described [18]. The final fractions from SEC chromatography containing purified receptor ECDs were dialyzed to storage buffer and stored at −80 • C until their use.

Expression and Purification of the RAMP3-CTR ECD Fusion Protein
HEK293S GnTI − cells were used to express the RAMP3-CTR ECD fusion protein. The general procedures of receptor ECD expression and purification from HEK293S GnTI − cells were previously described [22]. HEK293S GnTI − cells were transiently transfected with the DNA expression vector (H-pSL001) using polyethylenimine (PEI) at 1:1.5 ratio (DNA:PEI, w/w). Transfected cells were incubated for 4 days at 37 • C. All protein purification was performed at 4 • C unless otherwise noted. Cell culture media were collected and dialyzed to dialysis buffer overnight. The next day, dialyzed cell culture media were loaded to IMAC chromatography. Peak fractions from IMAC chromatography were spin-concentrated (MWCO 10 kDa) and injected into SEC. Peak fractions from SEC chromatography were collected ( Figure S1A) and the purified RAMP3-CTR ECD fusion protein was confirmed with SDS-PAGE ( Figure S1B). When N-glycans of the RAMP3-CTR ECD fusion protein were removed by PNGase F treatment, the band of the RAMP3-CTR ECD fusion protein located closely to 25 kDa, its expected MW without any N-glycans ( Figure S1C). The purified RAMP3-CTR ECD fusion protein showed a selective binding profile with an antagonistic amylin analog AC413 compared to CTR ECD alone suggesting that the purified RAMP3-CTR ECD fusion protein showed amylin receptor phenotype ( Figure S1D). The purified RAMP3-CTR ECD fractions were dialyzed to storage buffer and stored at −80 • C until their use for peptide binding assay.

Fluorescence Polarization/Anisotropy (FP) Peptide Binding Assay
The overall procedures of FP peptide binding assay with receptor ECDs and FITClabeled peptide probes were previously described [22]. FITC-labeled sCT(22-32) (10 nM) was used to evaluate peptide ligand affinity for the CTR ECD. AC413 has been reported as an antagonistic amylin peptide analog [23] and FITC-labeled AC413(6-25) with Y25P mutation (10 nM) was used for FP assay with AMY receptor ECDs since the Y25P mutation dramatically increased the AC413(6-25) affinity for AMY receptor ECDs [18]. A SpectraiD5 (Molecular Devices, San Jose, CA, USA) was used to measure fluorescence polarization/anisotropy. Background (reaction buffer only) was subtracted for anisotropy calculation. For the SpectraiD5, G factor (0.38 for FITC-sCT (22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32) and 0.41 for FITC-AC413(6-25) and FITC-AC413(6-25) Y25P) was used to correct the instrumental bias for anisotropy calculation. The polarization (mP) of the FITC-peptide probes only (No receptor ECD) was set close to 50 mP. For the saturation binding assay, the anisotropy values were re-calculated when total fluorescence intensity was changed by more than 10% as previously described [22]. However, effects of fluorescence intensity changes on the affinity values (K I ) obtained from the competition binding assay was reported to be minimum [24] and re-calculation of anisotropy was not applied. For the competition binding assay, the receptor ECD concentrations (CTR ECD 500 nM, RAMP1-CTR ECD fusion protein 77 nM, RAMP2-CTR ECD fusion protein 67 nM, and RAMP3-CTR ECD fusion protein 252 nM) that produced a half of the maximal anisotropy values with the respective peptide probes were used. Anisotropy values were used to produce non-linear regression curves with PRISM 5.0 (GraphPad software, San Diego, CA, USA) as previously described [22]. K I indicates affinity values obtained from competition binding assay and it was calculated from the non-linear regression curves of anisotropy by using equations previously described [22,24]. Concentrations of the FITC-labeled peptide and the total receptor protein and the affinity of the FITC-labeled peptide for the receptor protein were incorporated into the equations for the K I calculation of the competitive peptide ligand. Two technical replicates were used for each receptor concentration in the peptide-binding curve. SEM of the anisotropy values of the two replicates at each receptor concentration were presented as error bars in the representative peptide-binding curves. When the error bars were shorter than the height of the symbol, they were omitted in the representative curves. At least three independent peptide binding experiments were performed to obtain three independent peptide-binding curves. Mean and standard deviation of peptide ligand affinity were calculated from them.

Building Hypothetical Structures
Hypothetical structures were represented by using Pymol (Schrodinger, New York, NY, USA). The crystal structure of N-glycosylated CTR ECD with sCT(16-32) peptide (PDB 6PFO, Mol A) was used for figure representation. Selective mutations were introduced to the sCT(22-32) structure by using the mutagenesis function in Pymol. Hypothetical structures of the CTR ECD complexes with RAMP1/2/3 ECD were generated by placing RAMP1/2/3 ECD structures reported in the cryo-EM structures of CGRP and AM1/2 receptors as follows. CLR ECD structures of the CGRP receptor (PDB 6E3Y) and AM1 (PDB 6UUN) and AM2 receptors (PDB 6UVA) were aligned with the CTR ECD (PDB 6PFO, Mol A). Then, CLR ECD and other domains were removed and only RAMP ECD structures were shown with the CTR ECD to build the hypothetical structures of the RAMP-CTR ECD fusion proteins. Likewise, crystal structures of RAMP1-CLR ECD fusion (PDB 4RWG) and RAMP2-CLR ECD fusion (PDB 4RWF) proteins were used to build hypothetical AMY receptor 1/2 ECD structures by placing RAMP1/2 ECD next to the CTR ECD. These hypothetical structures were represented in Figure S2A,B. For the hypothetical structure of the CTR ECD complex with RAMP3 ECD, the additional cryo-EM structure of the AM2 receptor where AM peptide bound (PDB 6UUS) was used and the structure was shown in Figure S2C. For hydroxyproline mutations, a plugin PyTMs was installed in Pymol [25] and proline hydroxylation was used to make sCT S29HYP and P32HYP mutations.

Statistical Analysis
PRISM 5.0 (GraphPad software, San Diego, CA, USA) was used for one-way ANOVA and Tukey's post hoc test (unless otherwise noted), when more than two groups were compared. PRISM 5.0 was also used for Student's t-test (two-tailed) for the statistical analysis of two groups. p < 0.05 was considered as a statistical significance.

Effects of sCT S29 to Hydroxyproline Mutation on sCT(22-32) Affinity
Hydroxyproline (HYP) is an uncommon amino acid produced from proline post-translational protein modification through hydroxylation [27]. To confirm tha and hydrophobic side chains at the sCT S29 position are more suitable for CTR ECD ing, an HYP mutation was introduced to the sCT S29 position. When sCT S29 was m to HYP, the hypothetical structure suggested that its hydroxyl group located towa E123 of the CTR ECD with a 1.0 Å distance producing a visual overlap between HY the CTR E123 residue (Figure 3a). Consistent with the hypothetical structure, sCT(   13-fold compared to the affinity of sCT (22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32) with N26D/S29P mutations (Figure 3b and Table 2). These results are consistent with the idea that small and hydrophobic residues at the sCT S29 position are more suitable for CTR ECD binding.

Effects of sCT S29 to Hydroxyproline Mutation on sCT(22-32) Affinity
Hydroxyproline (HYP) is an uncommon amino acid produced from proline during post-translational protein modification through hydroxylation [27]. To confirm that small and hydrophobic side chains at the sCT S29 position are more suitable for CTR ECD binding, an HYP mutation was introduced to the sCT S29 position. When sCT S29 was mutated to HYP, the hypothetical structure suggested that its hydroxyl group located toward the E123 of the CTR ECD with a 1.0 Å distance producing a visual overlap between HYP and the CTR E123 residue (Figure 3a). Consistent with the hypothetical structure, sCT(22-32) with N26D/S29HYP mutations showed a marked affinity decrease for the CTR ECD by 13-fold compared to the affinity of sCT (22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32) with N26D/S29P mutations (Figure 3b and Table 2). These results are consistent with the idea that small and hydrophobic residues at the sCT S29 position are more suitable for CTR ECD binding.

Mutational Effects of sCT P32HYP on sCT(22-32) Affinity for AMY Receptor ECDs
sCT is also known as a dual agonist for CTR and AMY receptors. sCT has been used to activate AMY receptors in pre-clinical research areas [28][29][30] since it displayed high binding affinity and potency for AMY receptors and showed no discrimination between CTR and AMY receptors [1][2][3]. I investigated whether the mutational effects of P32HYP, N26D/S29P, and combined N26D/S29P/P32HYP mutation(s) are conserved for AMY receptor ECDs. AMY receptor 1 and 2 ECDs were previously reported as an ECD fusion protein of RAMP1/2 ECD and CTR ECD [18]. These ECD fusion proteins were functional AMY receptor ECDs that showed selective peptide ligand binding [18]. The current study reported for the first time to my knowledge, the successful purification of functional AMY receptor 3 ECD. The RAMP3 ECD-CTR ECD fusion protein was expressed from mammalian cells and purified ( Figure S1). The fusion protein showed selective peptide ligand binding suggesting that it is a functional AMY receptor 3 ECD ( Figure S1).
First, the hypothetical structures of all three AMY receptor ECDs were constructed by using the CTR ECD crystal structure (PDB 6PFO, Mol A) and recently reported cryo-EM structures of CGRP (PDB 6E3Y) and AM1/2 receptors (PDB 6UUN and 6UVA). The overview of the hypothetical AMY receptor ECD structures was shown in Figure 5a. These hypothetical structures predicted that the sCT P32HYP mutation did not make any visual clash with nearby RAMP1/2/3 ECD residues. The main chain of RAMP1 ECD W84 is located at a 4.5Å distance from the hydroxyl group of sCT HYP32 (Figure 5b). The main chain of RAMP2 ECD F111 is located at a 5.9Å distance from the hydroxyl group of sCT HYP32 (Figure 5c) and the side chain of RAMP3 ECD W84 located at a 5.2Å distance from the hydroxyl group of sCT HYP32 (Figure 5d). In addition, there are more RAMP1/2/3 ECD structures reported. When they were used to build hypothetical AMY receptor ECD structures ( Figure S2), the distance from the hydroxyl group of sCT HYP32 to the RAMP ECD residues ranged from 3.9Å to 5.2Å. In these hypothetical structures, RAMP ECD residues at AMY receptor ECDs would not inhibit CTR ECD interactions with sCT HYP32. These predictions are in line with the idea that the mutational effects of P32HYP on sCT (22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32) affinity for CTR ECD will be conserved for AMY receptor ECDs.
CTR forms a complex with RAMP and the complexes gain affinity for the peptide hormone amylin [1,2,34]. RAMP ECDs are known to provide limited access to the peptide hormones at CGRP and AM receptors and only C-terminal residues of CGRP and AM peptides were shown to contact RAMP ECDs [33]. Accordingly, the amylin C-terminal residue Y37 has been suggested as a potential site for RAMP interaction [21,35]. Amylin Y37 was reported to enhance amylin potency for AMY receptor 1/3 activation [35] and it was also reported to interact with RAMP2 residue E101 [21]. While amylin Y37 holds a hydroxyl group in its bulky side chain, sCT C-terminal residue P32 has a relatively small side chain. When the sCT P32HYP mutation introduced a hydroxyl group to P32, the additional hydroxyl group did not appear to interact with the RAMP ECD residues (Figure 5a-d). Consistently, the affinity enhancement mediated by the sCT P32HYP mutation was also conserved with the AMY receptor ECDs (Figure 5e-g). Structural information from either crystal structures of the AMY receptor ECDs or cryo-EM structures of full length AMY receptors will be essential to clearly elucidate how RAMPs interact with peptide ligands at AMY receptors.
One of the assumptions that initiated this study was that the CGRPmut and sCT fragments have a similar binding mode at their respective receptor ECDs. Booe et al. have reported extensive peptide interaction studies with CGRP and AM receptors to develop peptide analogs with improved affinity/potency profiles [36,37]. Their high-affinity CGRP and AM peptide analogs might be applicable to CTR and AMY receptors. The binding profiles of those peptide ligands for CTR and in particular for all three AMY receptor ECDs remain to be tested, although the receptor binding of those peptides would not be selective for CTR and AMY receptors.
The next important question is how to make peptide agonists with enhanced affinity and potency for CTR and AMY receptors. This study covers the sCT (22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32) fragment that binds the ECD portion of CTR and AMY receptors. These receptors are class B GPCRs and their activation model has been suggested [38,39]. The dynamic ECD of these receptors facilitates the initial binding of the peptide C-terminal part. The following interaction of the N-terminal part of the peptides with the receptor TM activates the receptors for G protein association triggering cell signaling. Booe et al. developed CGRP and AM peptide fragments with nanomolar affinity for their respective receptor ECDs [36,37]. Using these high-affinity peptide fragments, they further developed picomolar affinity antagonists targeting full-length CGRP and AM receptors [37]. Unexpectedly, when they made the agonist versions of these peptides with affinity-enhancing mutations, the potency enhancement of those agonists was not apparent as the affinity enhancement shown for the receptor ECDs [37]. However, the affinity enhancement for receptor ECDs was shown to increase peptide residence time at the CGRP and AM receptors and the agonists with affinity-enhancing mutations turned out to be long-acting agonists [37]. Whether the agonist version of sCT (22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32) with affinity-enhancing mutations increases the residence time at the CTR and whether it shows the long-acting property at the CTR are of great interest and remain to be investigated in future studies.
The efficacy of the dual agonists of CTR and AMY receptors for metabolic diseases has been largely attributed to their activity on the AMY receptor [40]. However, recent reports showed that the dual agonist KBP-088 and the combined use of amylin and CT were superior to activating either the AMY receptor or CTR alone [41,42]. These results suggest that the activation of CTR itself is involved in those metabolic processes and that CTR may be a valid player for the efficacy of the dual agonists. In addition, a long-acting amylin analog (LAAMA) was reported to decrease body weight gain in RAMP1 or RAMP3 knockout mice given a high fat diet [9]. These results indicate that CTR alone activated by LAAMA was enough to prevent body weight gain in mice. The current study provided the peptide analog that showed a higher affinity for CTR and AMY receptor ECDs than wild-type sCT (22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32). The mutations that enhanced sCT (22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32) affinity could be exploited to develop the peptide agonists with improved affinity and potency for these two important receptors.

Data Availability Statement:
The data underlying this article will be shared on reasonable request.