Cardiac Contractility Structure-Activity Relationship and Ligand-Receptor Interactions; the Discovery Of Unique and Novel Molecular Switches in Myosuppressin Signaling

Peptidergic signaling regulates cardiac contractility; thus, identifying molecular switches, ligand-receptor contacts, and antagonists aids in exploring the underlying mechanisms to influence health. Myosuppressin (MS), a decapeptide, diminishes cardiac contractility and gut motility. Myosuppressin binds to G protein-coupled receptor (GPCR) proteins. Two Drosophila melanogaster myosuppressin receptors (DrmMS-Rs) exist; however, no mechanism underlying MS-R activation is reported. We predicted DrmMS-Rs contained molecular switches that resembled those of Rhodopsin. Additionally, we believed DrmMS-DrmMS-R1 and DrmMS-DrmMS-R2 interactions would reflect our structure-activity relationship (SAR) data. We hypothesized agonist- and antagonist-receptor contacts would differ from one another depending on activity. Lastly, we expected our study to apply to other species; we tested this hypothesis in Rhodnius prolixus, the Chagas disease vector. Searching DrmMS-Rs for molecular switches led to the discovery of a unique ionic lock and a novel 3–6 lock, as well as transmission and tyrosine toggle switches. The DrmMS-DrmMS-R1 and DrmMS-DrmMS-R2 contacts suggested tissue-specific signaling existed, which was in line with our SAR data. We identified R. prolixus (Rhp)MS-R and discovered it, too, contained the unique myosuppressin ionic lock and novel 3–6 lock found in DrmMS-Rs as well as transmission and tyrosine toggle switches. Further, these motifs were present in red flour beetle, common water flea, honey bee, domestic silkworm, and termite MS-Rs. RhpMS and DrmMS decreased R. prolixus cardiac contractility dose dependently with EC50 values of 140 nM and 50 nM. Based on ligand-receptor contacts, we designed RhpMS analogs believed to be an active core and antagonist; testing on heart confirmed these predictions. The active core docking mimicked RhpMS, however, the antagonist did not. Together, these data were consistent with the unique ionic lock, novel 3–6 lock, transmission switch, and tyrosine toggle switch being involved in mechanisms underlying TM movement and MS-R activation, and the ability of MS agonists and antagonists to influence physiology.


Introduction
Peptidergic signaling plays numerous critical roles in transmitting and regulating physiological processes. Therefore, delineating the mechanisms that underlie these events is a powerful approach to identifying target molecules to influence health. An important first step in signaling is when a ligand binds to a G protein-coupled receptor protein (GPCR). Ligand-receptor binding disrupts molecular switches which cause TM movement and ultimately results in receptor activation reviewed in [1].
Myosuppressin dramatically decreases cardiac contractility and gut motility [2,3]. First isolated as a cockroach brain peptide that affects spontaneous contractions of the gut, myosuppressin was subsequently found to be distributed throughout the invertebrates. The conservation of its structure and activities, and its widespread distribution are consistent with myosuppressin playing an important role in physiology.
Myosuppressins are members of a family of peptides with an identical C-terminal RF-NH 2 , however, the N-terminal extension is unique and differs in length and sequence [4]. The identical C terminus and variant N terminus are both important in binding and activating signaling pathways [5,6]. FMRF-NH 2 , the first RF-NH 2 -containing peptide isolated, was identified from a neural extract applied to a clam heart preparation [7]. This superfamily of peptides is grouped based on XRF-NH 2 , where myosuppressins often contain X = L. Myosuppressins are typically represented by X1DVX2HX3FLRF-NH 2 , where X1 = pQ, P, T, A; X2 = D, G, V; X3 = V, S [8].
Drosophila melanogaster myosuppressin (DrmMS; TDVDHVFLRF-NH 2 ) is representative of its peptide family [9]. DrmMS is pleotropic decreasing the frequency of both cardiac contractility and gut motility. The DrmMS structure-activity relationship (SAR) for its effect on cardiac contractility and gut motility is reported; the data are consistent with DrmMS having distinct signaling pathways in heart and gut [8]. In addition, DrmMS binds to two putative GPCR proteins, DrmMS-R1 and DrmMS-R2 [10]. Apart from these facts, little is known about MS signaling. No mechanism which underlies MS receptor activation, a crucial step in signal transduction, is described in literature. And, the design and characterization of MS antagonists in a disease vector are molecularly and physiologically limited in scope in the literature.
The kissing bug, Rhodnius prolixus, is a vector of the Chagas disease, an important health problem [11]. RhpMS, pQDIDHVFMRF-NH 2 , contains two substitutions compared to the MS consensus structure; V3 ! I3 and L8 ! M8 [12]. The physicochemical characteristics of myosuppressins are conserved with the residue replacements [12,13]; RhpMS affects R. prolixus heart rate [14]. The unique MS provides an opportunity to further explore a pathway that affects a crucial physiological function in a disease vector. Previously, little was known about RhpMS signaling; its receptor sequence and structure was unidentified and its SAR uncharacterized.
This study tested our prediction that MS signaling would mimic mechanisms involved in Rhodopsin activation. We searched for molecular switch motifs across DrmMS-Rs and discovered the unique ionic lock and novel 3-6 lock, in addition to the transmission and tyrosine toggle switches. We also tested our belief that DrmMS-DrmMS-R1 and DrmMS-DmrMS-R2 interactions would reflect our SAR data, which it did. When DrmMS and its N-terminal truncation and alanyl-substituted analogs were docked to the DrmMS-Rs, the ligand contacts were distinct between receptors consistent with our SAR data. DrmMS interactions with DrmMS-R2 mirrored the cardiac contractility SAR data and DrmMS-DrmMS-R1 interactions reflected the gut motility data. Additionally, we hypothesized agonist-and antagonist-receptor contacts would differ from one another reflecting activity and inactivity. The docking data confirmed this prediction; agonists mirrored DrmMS interactions, yet an inactive analog failed to mimic parent peptide contacts.
Lastly, we expected our study to apply to other species; we tested this hypothesis in R. prolixus. We investigated RhpMS-R structure, binding pockets, ligand contacts, and SAR. We identified the R. prolixus receptor and found it shared substantial sequence identity to DrmMS-R1 and DrmMS-R2, 56% and 51%, respectively. The predicted protein was modeled to find the receptor contained typical GPCR features. RhpMS-R contained the unique myosuppressin ionic lock and novel 3-6 lock, and the transmission and tyrosine toggle switches, which, upon ligand binding, promoted TM movement and receptor activation. We obtained further support for the role of the unique and novel locks by identifying and modeling the red flour beetle, common water flea, honey bee, domestic hornworm, and termite MS-Rs; their structures, too, contained the myosuppressin motifs, which were likely involved in TM movement and receptor activation.
Due to the conservation of physicochemistry of the amino acids, which differed between the peptides and the high receptor sequence identity, we predicted RhpMS and DrmMS would be alike in activity and ligand contacts. RhpMS and DrmMS decreased R. prolixus cardiac contractility dose dependently with EC 50 values of 140 nM and 50 nM, respectively. Based on ligand-receptor contacts, analogs were predicted to be an RhpMS agonist or inactive and act as an antagonist. [5][6][7][8][9][10] RhpMS mimicked the full-length peptide; [6][7][8][9][10]RhpMS applied to R. prolixus heart was inactive and blocked the effect of the parent peptide. Together, data from these studies confirmed tissue specificity in MS signaling, and supported the roles of a unique and a novel lock in MS-R activation.

Ethics statement
This research utilized an invertebrate, R. prolixus, reared according to a protocol in complete agreement with the recommendation established the by Directive 2010/63/EU pf the European Parliament and the Resolution 1047/2005 Annex II of the National Council of Scientific and Technical Researches (COINCET; Buenos Aires, Argentina) related to the protection of animals used for scientific purposes.

Animals
R. prolixus were maintained at 28°C under high humidity in a cycle of 12h:12h (L:D). The insects used were 5 th instar 2-3 weeks post-ecdysis fed on chicken blood.

Chemicals
All peptides were synthesized on a 433A Applied Biosystems peptide synthesizer using standard Fmoc procedures and purified by reversed-phase high performance liquid chromatography. Each synthesis was obtained with 95% purity and identified by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Peptides were diluted in series with physiological saline to obtain the working solutions.

Bioassays
Fifth instar R. prolixus were placed ventral surface down on a dish with paraffin wax and dissected under physiological saline. Segments four to seven from the dorsal abdominal cuticle were removed using minuten pins leaving the dorsal vessel exposed. The semi-intact preparation was equilibrated in 20 μl saline for 20 minutes at room temperature (25°± 2°C), replacing the fluid with fresh saline every five minutes. Heart rate was measured under saline after the final bath, counting the number of contractions in one minute. This counting was replicated three times for each preparation, and the results of the three countings were averaged. After this, the saline was removed and replaced with an equal volume of solution containing a peptide. The heart rate was measured and expressed as % of the control (saline). Ten animals were analyzed for control and for each experiment. The investigator who performed the bioassay was unaware of the peptide structures being tested.

Data analysis
Dose response curves were generated and the effective concentrations at half-maximal response (EC 50 ) values were calculated using a non-linear regression with GraphPad Prism version 6.03. Activities of RhpMS, DrmMS, and analogs were compared to saline, set at 100%, using Single Factor ANOVA to calculate p values; significance was set at p 0.05. tyrosine toggle switch [1]. No MS-R molecular switch was previously reported in the literature. The DrmMS-R structures contained structural motifs reminiscent of those present in Rhodopsin. The ionic lock switch was represented by the WRY motif on TM3 [ Fig. 1]. It was present in the DrmMS-Rs in the same location as ERY in Rhodopsin (PDB ID: 1F88) [20]. The WRY motif of DrmMS-R interacted with multiple T residues on TM6. Even so, DrmMS-R lost an electrostatic interaction between TM3 and TM6 due to the absence of a negatively-charged residue on TM6 within the range of a salt bridge. Also, because WRY did not contain E it could not form an intra-motif interaction with R130/125 (notation used to indicate R130 in DrmMS-R1 and R125 in DrmMS-R2).
A novel 3-6 lock was also discovered. It sat between TM3 and TM6 limiting the movement of TM6 and maintaining the inactive state of the receptor. The 3-6 lock mirrored the 3-7 lock of Rhodopsin. When the lock was disrupted, TM6 likely moved and the receptor was activated [21]. TM7 in DrmMS-Rs did not contain a P residue, thus, it was more rigid than Rhodopsin and less likely to disrupt the resting state of the receptor. This was one reason stabilization of transmembrane 7 was unnecessary. TM6 was stabilized by the 3-6 lock with a salt bridge interaction between H116/114 on TM3 and E265/369 on TM6.
The DrmMS-Rs contained the TEFP transmission switch on TM6 in the same location as the CWLP motif in Rhodopsin [22]. Further, F361/365 located one helix turn away from E365/ 369 on TM6, H116/114 on TM3, and P284/292 located nearby on TM5 in DrmMS-Rs, were in the same position as in Rhodopsin. A notable difference was E365/369 replaced W265, which was stabilized through ionic interactions with H116/114 on TM3, K281/289 on TM5, and Q368/372 on TM7 in the DrmMS-Rs. Lastly, in the DrmMS-Rs, NFILY was similar to the Rhodopsin tyrosine toggle switch motif, NPVIY [23]. The loss of P on DrmMS-R TM7 suggested the helix would be less likely to move. TM7 movement in Rhodopsin positions Y306 to extend a hydrogen bond network from the binding pocket to the intracellular side of the receptor. A loss of DrmMS-R TM7 movement suggested Y in the toggle switch may be distanced from the hydrogen bond network. Further, in the DrmMS-Rs, S292/300 replaced Y223 on Rhodopsin TM5, which also participated in this network. Due to the distance apart, S292/300 and the ionic lock did not interact, although water-mediated bonds might exist. Either configuration would result in a less extensive hydrogen bond network than observed for Rhodopsin. Together, the presence of DrmMS-R molecular switches resembled those present in Rhodopsin, with predictably similar roles in receptor activation.
DrmMS-Rs were compared to additional Rhodopsin-like receptors to determine the uniqueness of the ionic and 3-6 locks. The inactive states of the A 2A adenosine ( [24,25,26,27,28]. The ionic lock in DrmMS-Rs was similar to that in H 1 R, which formed a hydrogen bond between R125 and Q416; there was lack of evidence for an ionic lock in A 2A R, β 2 AR, and CKCR4, and the lock in D 3 R resembled that of Rhodopsin. DrmMS-Rs were the only receptors to show the presence of a 3-6 lock, all other receptors, with the exception of A 2A R, contained a 3-7 lock. The transmission and tyrosine toggle switches between MS-Rs and additional Rhodopsin-like receptors are highly similar. Next, we believed DrmMS-R1 and DrmMS-R2 interactions would reflect our SAR data [8]. This hypothesis required identifying and characterizing the DrmMS-R1 and DrmMS-R2 binding pockets. The DrmMS-R binding pockets (Fig. 2, Fig. 3) were similar to each other in shape, size, and physicochemical characteristics, but each was characterized by some unique features ( Table 1, Table 2). In DrmMS-R2, W165 was between TM4 and TM5 excluding Y284 on TM5; DrmMS-R1 model. The exposed atoms at the surface of the ligand-binding pocket are highlighted as carbon (gray), oxygen (red), and nitrogen (blue). The pocket contained a polar region between TM6 and TM7, and separate hydrophobic and aromatic regions near TM1 and TM5. DrmMS-R1 was undefined between TM1 and TM7, and TM4 and TM5.
doi:10.1371/journal.pone.0120492.g002 due to its size, S167 in DrmMS-R1 did not mimic W165. Q269 in DrmMS-R1 and R277 in DrmMS-R2 differed in charge. In DrmMS-R1, K25 rotated out, yet, in DrmMS-R2, K23 was exposed resulting in a charge on TM1. In DrmMS-R1, H116 rotated toward K281 on TM6 to block access to a hydrophobic region to the center of the pocket in DrmMS-R1. Contrary, in DrmMS-R2, H114 rotated toward TM7 increasing access to this region. Y276 on TM5 in DrmMS-R1 increased the hydrophobic character, yet, the rotation of Y284 in DrmMS-R2 moved F281 further into the pocket. Together with the H114 rotation, the absence of Y284 in the pocket led to accessibility of the hydrophobic region around TM4. Also, ECL2 in DrmMS-R2 was structured with helices and beta sheets, but, in DrmMS-R1, it was predicted to have no secondary structure, making it flexible enough to be modeled between TM1 and TM7. The unique binding pockets were consistent with tissue-specific signaling as inferred by our SAR data. [6][7][8][9][10]DrmMS was the heart rate active core, yet [4][5][6][7][8][9][10]DrmMS was the gut motility active core. The alanine scans performed in heart and gut showed the residues crucial to DrmMS activity and binding are different between the two tissues; F7 and L8 were essential in gut, and F7  and F10 in heart. We believed DrmMS-DrmMS-R1 and DrmMS-DrmMS-R2 interactions would reflect our SAR data. N-terminal truncated and alanyl-substituted analogs (Table 3) were docked to the receptors; additional data are available in Supporting Information. In DrmMS docked to DrmMS-R1 ( Fig. 4, Table 4), F7 made multiple, strong hydrophobic and aromatic contacts, which L8 extended by interacting with hydrophobic residues on TM1, TM2 and TM7. H5, V6, and F10 formed additional hydrophobic contacts between TM4, TM5, and TM6; H5 made extensive interactions within this region. The N-terminal residues formed weaker hydrophobic and polar interactions to TM2 and TM3. In DrmMS-R2 (Fig. 5, Table 5), F7 formed contacts between TM1 and TM2; the position of F10 on TM7 allowed for hydrophobic and aromatic interactions with ECL3. D2 and R9 interacted and formed an ionic network near TM6. D4 made extensive intramolecular interactions with the backbone to stabilize the peptide. The N terminus spanned the pocket to make hydrophobic and polar interactions with TM3, TM4, and TM5. The residues which formed strong interactions differed between the DrmMS-Rs, in particular, L8, F7, and F10 showed receptor-specific ligand binding in line with our tissue-specific SAR data.
In [4][5][6][7][8][9][10]DrmMS, active in heart and the active core in gut, docked to DrmMS-R1 ( Fig. 6, Table 6), F7 initiated a hydrophobic network that was extended by L8 to the TM1, TM2, and TM7 interface. V6 and F10 contributed to this network by making contacts near TM2 and TM3, respectively. In DrmMS-R2, the analog mimicked DrmMS, though the contacts were not made by the same residues ( Fig. 7, Table 7). F10 spanned the pocket to pi-stack to TM5. The C-terminal amide docked between TM6 and TM7 leading R9 to form hydrogen bonds with TM3, D4, and H5, which created a large hydrogen bond network extending to the bottom of the pocket. V6 and F7 maintained DrmMS-like contacts and interacted with each other, creating a small hydrophobic network between TM1 and TM2. The absence of three N-terminal residues did not change the backbone conformation or contacts relative to DrmMS. Thus, the contact sites between [4][5][6][7][8][9][10] DrmMS and the DrmMS-Rs were consistent with the analog being active in heart and gut.
When docked to DrmMS-R2, DrmMS, [4][5][6][7][8][9][10]DrmMS, and [A8]DrmMS contacted H114, K289, and Q372, receptor residues comprising and interacting with the 3-6 lock. H114 rotated to the extracellular side of the receptor to make strong contacts with the ligand, distancing it Discovery of Novel Myosuppressin Signaling Switches Table 5. DrmMS contact sites on DrmMS-R2 a . RhpMS: receptor, molecular switches, ligand contacts, and SAR We identified RhpMS-R in the R. prolixus genome database (www.vectorbase.org) using DrmMS-Rs as queries; it was contained in a single exon gene, Supercontig GL56309. RhpMS-R shared 56% sequence identity with DrmMS-R1 and 52% with DrmMS-R2. The hydrophobic and aromatic residues on TM2 and TM3 created a region analogous to that seen in the DrmMS-Rs (Fig. 13). K218 on TM5 was rotated out compared to DrmMS-R1 and DrmMS-R2. F210 and Y213 on TM5 were available to make strong aromatic interactions. Finally, residues on TM6 and TM7, including E281, Q284, D308, and D311, created a polar region that included a nearby residue, H114. RhpMS-R was searched to identify molecular switches; it contained a unique ionic lock represented by a WRY motif on TM3 (Fig. 14). R128 of RhpMS-R formed a hydrogen-bond network with multiple T residues on TM6. The ionic lock did not form an electrostatic interaction due to the lack of a negatively charged residue on TM6 or within the WRY motif. The novel 3-6 lock was also present in RhpMS-R. Similar to DrmMS-Rs, RhpMS-R did not contain a P residue on TM7, indicating that stabilization of TM7 through a 3-7 lock would not be necessary. Rather, the 3-6 lock of RhpMS-R involved the formation of a salt bridge between H114 on TM3 and E281 on TM6, limiting the movement of TM6 and stabilizing the inactive state of  the receptor. Receptor activation would involve the disruption of the 3-6 lock to allow for the movement of TM6. RhpMS-R contained a TEFP motif on TM6 identical to that of the DrmMS-Rs (Fig. 14). The polar character of the RhpMS-R pocket favored the presence of E281 in comparison to W265 in Rhodopsin. Further, E281 was stabilized through the 3-6 lock with an ionic interaction to H114. The conservation of the motif and presence of F227 on TM6, H114 on TM3, and P221 on TM5 indicated that the function of the transmission switch, inducing receptor activation through TM3-TM5-TM6 structural changes upon agonist binding [22], was maintained. The NFMIY motif on TM7 in RhpMS-R was structurally similar to NFILY in DrmMS-Rs, its tyrosine toggle switch. The lack of a P residue in TM7 suggested that Y325 was moved from a hydrogen bond network to the intracellular side of the receptor upon activation. Further, RhpMS-R contained S229 on TM5 which was unable to reach the hydrogen bond network. Additional work to model and search B. mori (BomMS-R), A. mellifera (ApmMS-R1 and ApmMS-R2), D. pulex (DapMS-R), T. castaneum (TrcMS-R), and Z. nevadensis (ZonMS-R) showed a conservation of the 3-6 lock within the MS-Rs (Fig. 15).
In RhpMS docked to RhpMS-R (Fig. 16, Table 13) F7 and F10 interacted with hydrophobic and aromatic regions, and R9 formed a salt bridge between TM6 and TM7. D4 made strong contacts, forming salt bridge interactions and extending an ionic network initiated by R9. H5 interacted extensively with the backbone and C-terminal amide to stabilize the ligand. I3, H5, V6, M8 and R9 contacts in RhpMS resembled those of DrmMS docked to DrmMS-R2. However, RhpMS was slightly shifted; Y391 on TM7 that pi-stacked with DrmMS F10 was modeled as part of ECL3 in RhpMS-R, which caused F10 to pi-stack with F210 on TM5. K23 on TM1 of DrmMS-R2 was R25 in RhpMS-R, which altered the orientations of nearby residues and resulted in F7 docking between TM2 and TM3 instead of TM1 and TM7 as in DrmMS. K218 on TM5, analogous to K289 on DrmMS-R2 that contacted D2 in DrmMS, was outside of the pocket in RhpMS-R and, thus, was unable to dock favorably. However, D4 was rotated and interacted with R9 maintaining an ionic network to TM7 as in DrmMS. Due to its size, pQ1  could not dock as deeply between TM3, TM4, and TM5, as T1 did in DrmMS, thus, it docked upwards between TM4 and TM5.

Discussion
In this paper we described molecular switches involved in TM movement that underlies MS-R activation; no prior publication reports these motifs and mechanisms. We found a unique ionic lock and novel 3-6 lock held MS-Rs in an inactive state that, upon ligand binding, may lead to TM6 movement, a crucial step for receptor activation [20,21,29]. A conserved TEFP present in MS-Rs mimicked the Rhodopsin transmission switch, CWLP [22,30]. Lastly, NF (M/I)(I/L)Y in MS-Rs resembled the Rhodopsin tyrosine toggle switch motif, NPVIY.

Discovery of Novel Myosuppressin Signaling Switches
The motifs were present and made contacts consistent with MS-R activation. Even so, the motifs of the MS-Rs were unique compared to Rhodopsin, in line with the MS peptide, receptor structures, and ligand contacts, and consistent with the DrmMS and RhpMS SAR data. The loss of contacts and networks, and weakened interactions observed in MS-Rs compared to Rhodopsin may indicate their transition from the inactive to active state occurs on a different time scale or energy level.
Myosuppressins are likely to play crucial roles in physiology; however, ligand-receptor contact data remained unpublished. DrmMS-R1 and DrmMS-R2 share high sequence identity and bind the same ligand, yet, their binding pockets differed physicochemically. The unique DrmMS-DrmMS-R1 and DrmMS-DrmMS-R2 contact data that we report and our previous SAR data [8] were consistent with tissue-specific signaling in heart and gut. Next, we explored MS ligand binding and receptor activation in R. prolixus. Although RhpMS-R shared high sequence identity with the DrmMS-Rs, its binding pocket was different in shape and size, and the residues which projected into it. These data were in line with the unique RhpMS SAR compared to DrmMS data, in particular, the differences in the RhpMS active core and antagonist structures for R. prolixus cardiac contractility.            Together, the data from this study describe molecular switches involved in receptor activation and ligand contacts which provide insight into how the motifs are involved in MS signaling. Additionally, a bioassay, and binding pockets, size and physicochemistry of the residues available to make contacts supported published data demonstrating tissue-specificity of myosuppressin signaling and yielded information on MS SAR and its receptor in a disease vector.    S8 Table. DrmMS ligand-receptor contact sites on RhpMS-R a . a Residues numbered 1-10 are in DrmMS or RhpMS. (NH) and (CO) indicate that the residue backbone group was contacted. In the case in which a residue was contacted twice by the backbone or side chain of the same ligand residue, O and H (backbone atoms), OH (hydroxyl of Y), and CO (carbonyl of Bpa) are used to distinguish the contacts. (DOCX)