Identification of an allatostatin C signaling system in mollusc Aplysia

Neuropeptides, as pervasive intercellular signaling molecules in the CNS, modulate a variety of behavioral systems in both protostomes and deuterostomes. Allatostatins are neuropeptides in arthropods that inhibit the biosynthesis of juvenile hormones. Based on amino acid sequences, they are divided into three different types in arthropods: allatostatin A, allatostatin B, allatostatin C. Allatostatin C (AstC) was first isolated from Manduca sexta, and it has an important conserved feature of a disulfide bridge formed by two cysteine residues. Moreover, AstC appears to be the ortholog of mammalian somatostatin, and it has functions in common with somatostatin, such as modulating feeding behaviors. The AstC signaling system has been widely studied in arthropods, but minimally studied in molluscs. In this study, we seek to identify the AstC signaling system in the marine mollusc Aplysia californica. We cloned the AstC precursor from the cDNA of Aplysia. We predicted a 15-amino acid peptide with a disulfide bridge, i.e., AstC, using NeuroPred. We then cloned two putative allatostatin C-like receptors and through NCBI Conserved Domain Search we found that they belonged to the G protein-coupled receptor (GPCR) family. In addition, using an inositol monophosphate 1 (IP1) accumulation assay, we showed that Aplysia AstC could activate one of the putative receptors, i.e., the AstC-R, at the lowest EC50, and AstC without the disulfide bridge (AstC') activated AstC-R with the highest EC50. Moreover, four molluscan AstCs with variations of sequences from Aplysia AstC but with the disulfide bridge activated AstC-R at intermediate EC50. In summary, our successful identification of the Aplysia AstC precursor and its receptor (AstC-R) represents the first example in molluscs, and provides an important basis for further studies of the AstC signaling system in Aplysia and other molluscs.


Results
Identifying a precursor for AstC and predicting peptides in Aplysia. To identify putative precursor and receptors for Aplysia AstC, we began with a bioinformatic analysis. For the precursor, searching "Aplysia allatostatin C" in NCBI returned two entries: a predicted sequence (accession number: XM_005112737.3, which corresponds to a genome sequence: NW-004798659.1) (Fig. 1A), and an AstC precursor deposited in 2010 (mRNA accession number: GU973882) (Fig. 1B), which is likely based on an early large-scale sequencing project 55 . Using the RNA sequence from NCBI (XM_005112737.3), we also found an DNA sequence (DNA:  Figure 1. Gene expression mapping of the Aplysia Allatostatin C precursor. (A) A genome sequence from NCBI (NW-004798659.1) expresses 5 genes: LOC101859220; LOC101859459 (product: enolase); LOC101860306 (product: phosphatidylinositol); LOC101859837 (product: glyoxylate reductase/ hydroxypyruvate reductase); LOC101860069 (product: glyoxylate reductase/hydroxypyruvate reductase). The first gene (LOC101859220) corresponds to a mRNA (XM_005112737.3, with an intron between 8955 to 20,579 bp), which produces an uncharacterized protein (XP-005112794.1). This protein is similar to Allatostatin C from other molluscs. (B) A mRNA from NCBI (GU973882), named Aplysia californica allatostatin C mRNA, complete CDS, was submitted by Moroz et al. 55 . It produces a protein, Allatostatin C (ADX20596.1), which is identical to XP-005112794.1 in (A), but its untranslated regions at 5' and 3' ends are different from those of (XM_005112737.3) in (A). (C) DNA contig-1403 (Note that the nucleotide number on top starts from the right) from Aplysia gene nucleotide databases (the AplysiaTools) expresses the same mRNA as in (A) (XM_005112737.3) and the protein generated from this mRNA is the same as those in (A,B). Note that the numbers for the proteins refer to base pairs of corresponding mRNAs, not amino acids. contig_1403) in AplysiaTools (see Materials and Methods) that produces an identical mRNA sequence including the 5' and 3' ends of the noncoding regions as in mRNA (accession number: XM_005112737.3) (Fig. 1C). Note that the mRNA sequence (GU973882) produces an identical protein, but its noncoding regions are very different from the RNA sequence from NCBI (XM_005112737.3) and the AplysiaTools sequence, suggesting that the noncoding regions of the former sequence (GU973882) did not originate from the AstC precursor gene. Given that there are two precursor genes for AstC in arthropods 4 , with the second gene next to the first one, we examined whether there is a second mRNA following the AstC precursor that may encode an AstC peptide. In the genome scaffold (including NW-004798659.1 from NCBI and DNA contig-1403 from AplysiaTools that encode the AstC mRNA sequence in Fig. 1A,C), we found only one gene for the AstC related sequence, suggesting that it is likely that there is only a single AstC precursor in Aplysia.
After using bioinformatics to find a potential allatostatin C gene in Aplysia, it was important to identify the peptides that are generated by the precursor gene, and to find receptors that are responsive to the peptides. Here, we first designed primers (Supplementary Table 1) using the sequence we found, and performed PCR on cDNA of Aplysia californica, and obtained an mRNA of 351 bp in length ( Fig. 2A, see Supplementary Fig. 1 for the complete gels). The gene sequence has been deposited into the NCBI database (GenBank accession number: OL546292). This mRNA sequence is similar to sequences from NCBI (i.e., XM_005112737.3 and GU973882) and AplysiaTools, and these sequences differ from each other in only one or two nucleotides ( Supplementary  Fig. 2). Moreover, these single nucleotide polymorphisms (SNPs) did not affect protein sequences, and the proteins from all four genes are identical. Thus, combined with earlier work, these data support the presence of an AstC precursor in Aplysia.
Next, we determined the identities and similarities between the Aplysia AstC precursor with AstC precursors from some selected invertebrate species and somatostatin precursors in some vertebrate species (Table 1, see  Supplementary Table 2 for information on these sequences). Of the precursors examined, the Aplysia precursor is most closely related to molluscan AstC precursors, e.g., gastropods Theba pisana with a similarity of 53.1%, and Deroceras reticulatum with a similarity of 50.4%. Furthermore, we compared the 7 AstC precursors in molluscs and 4 in annelids (Fig. 3A, see Supplementary Table 2).
Finally, we used NeuroPred 56 to predict possible peptides that might be generated from the AstC precursor (Fig. 3B). They include at least two peptides: AstC: SHYSSMCMFNVVACY, and AstC gene-derived peptide (AstCG-DP): AVIPVSSPEPMEEASALQLLPGKIamide. Similar to AstC in other invertebrates, Aplysia AstC has two Cysteines with 6 amino acids in between. We also compared Aplysia AstC with AstCs in other invertebrate species and with somatostatin in mammals (Fig. 4, see Supplementary Table 3 for information on these sequences). Given that AstCs with the disulfide bridge in arthropods are bioactive, we hypothesized that Aplysia AstC with the disulfide bridge could activate an Aplysia AstC receptor (see Fig. 7).

Identifying putative receptors for AstC in Aplysia.
To identify putative receptors, we searched "Aplysia allatostatin C receptor" in NCBI, but this search did not return any sequences. Because of the homology of allatostatin C and somatostatin 4,8,9,11 , we then tried to search "Aplysia somatostatin receptor" in NCBI, which did return seven sequences (Supplementary Table 4). In Supplementary Table 4, we also indicated whether these seven sequences are present in AplysiaTools. Next, we used NCBI Conserved Domain Search and TMHMM sever 2.0 to predict whether these seven sequences are GPCRs. Among them, three are predicted to have 7 transmembrane domains (Fig. 5A), which presumably are complete GPCR sequences. These three sequences are also present in AplysiaTools databases (see Supplementary Table 4).
To determine whether the three putative GPCRs might be related to AstC receptors, we blasted each sequence in NCBI in species where more protein sequences have been studied, i.e., Caenorhabditis elegans, Drosophila melanogaster, Danio rerio and Mus musculus. For the protein with accession number: XP_005095139.1 (mRNA:   Table 5), suggesting that this protein might be related to AstC receptors. We therefore tentatively named it as AstC-R. For the protein with accession number: XP_005096107.1 (mRNA: XM_005096050.3), named somatostatin receptor type 3-like receptor in NCBI, and the protein with accession number: XP_012945941.1 (mRNA: XM_013090487.2), also named somatostatin receptor type 2-like receptor in NCBI, these searches did not return useful known proteins with low E-values (< 1E-4) and high query coverage (> 50%) (Supplementary Table 5). Therefore, we used Pfam (http:// pfam. xfam. org/ search# tabvi ew= tab1) to blast both proteins, and both are classified as Class A GPCRs (rhodopsin family). Thus, we tentatively named the first protein as Class-A_GPCR1, and the second one as Class-A_GPCR2 (Supplementary Table 4 and 5).
To obtain a better view of phylogenetic relationship between the three proteins, we decided to construct a phylogenetic tree with a number of Class A GPCRs in Lottia giagantea and Crassostrea gigas from the Supplementary File, sd02.rtf, of Jekely 2013 8 and two other sequences of Crassostrea gigas (Cg_XP_011450519.2, Cg_XP_011428314.2). We selected protein sequences with more than 340 amino acids, and eliminated duplicated sequences from file "sd02.rtf ". As described in the last paragraph, we blasted these sequences in Caenorhabditis elegans, Drosophila melanogaster, Danio rerio and Mus musculus and annotated them (Supplementary Table 5, except Cg_XP_011450519.2, which did not return a proper name, so we named it "Class-A_GPCR C gigas" based on Pfam). Then, we added the three Aplysia sequences, together with AstC-R sequence of Crassostrea gigas (see also Supplementary Table 5), and re-ran the phylogenetic tree (Fig. 5B). The tree showed that Aplysia AstC-R clustered together with Crassostrea gigas AstC-R, supporting the hypothesis that Aplysia AstC-R might be an AstC receptor, whereas the two other Aplysia proteins, Class-A_GPCR1 and Class-A_GPCR2, are not closely related to AstC receptors. Finally, we generated a phylogenetic tree of Aplysia AstC-R with AstC-Rs from selected species in arthropods, lophotrochozoa and some somatostatin receptors in vertebrates (Fig. 6, see Supplementary Table 6 for information on these sequences). We also compared most of the putative molluscan AstC-Rs with Aplysia AstC-R using BioEdit ( Supplementary Fig. 3, see Supplementary Table 6). The data indicate that of the sequences examined, Aplysia AstC-R is most closely related to the sequence from pulmonate gastropod, Table 1. Identities and similarities of AstC precursors from invertebrate species, and with somatostatins in vertebrates using BioEdit (pairwise alignment-calculate identity/similarity for two sequences).  in Supplementary Table 6). We chose to pursue the research by first cloning Aplysia AstC-R. We also cloned Aplysia Class-A_GPCR1, which was to be used as a control. We designed primers (Supplementary Table 1)   www.nature.com/scientificreports/ number: OL546293; Class-A_GPCR1 GenBank accession number: OL546294). To search for other sequences that might be related to AstC-R, we used the cloned AstC-R sequence to blast both the transcriptome and the genome of the AplysiaTools databases, but we did not find any additional related sequences.

Activation of putative receptors by AstC peptides.
We cloned AstC-R and Class-A_GPCR1 into pcDNA3.1 plasmids, and expressed them in CHO cells. We then used the IP1 accumulation assay that detects IP1 generated in the Gαq pathway (see Methods) to determine whether our predicted Aplysia peptides (i.e., AstC or AstCG-DP) and other related peptides which we synthesized (see Fig. 7G) could activate the receptors. In preliminary experiments, we only transfected plasmids for a putative receptor in CHO cells. However, none of the receptors responded to AstC or AstCG-DP, suggesting that it is possible that these receptors did not associate with the native Gαq in CHO cells. Thus, for all the experiments shown in Fig. 7, we co-transfected plasmids for the putative receptor and Gα16 plasmids in CHO cells. Because Gα16 is a promiscuous Gαq protein that will bind to most GPCRs 10,53 , this procedure would ensure an IP1 response when a potential ligand binds to its GPCR.
We first screened IP1 responses of the two peptides, AstC and AstCG-DP, at two concentrations (10 -10 M and 10 -5 M) on the two receptors: AstC-R and Class-A_GPCR1 (Fig. 7A). At 10 -10 M, a peptide activated a receptor minimally, if at all, so it is used as a control. We also tested the effects of several other peptides, including AstC without the disulfide bridge (named AstC'), Aplysia SPTR (apSPTR-GP-DP2), another Aplysia peptide also with a disulfide bridge 52 . In addition, we selected several AstCs predicted from four other molluscan species (Fig. 4A), i.e., AstC-T (Theba pisana), AstC-D (Deroceras reticulatum), AstC-C (Crassostrea gigas) and AstC-L (Lottia giagantea) (Fig. 7G). We found that AstC-R was responsive to AstC and AstC' , but not to AstCG-DP or SPTR (Fig. 7A). The data suggest that AstC and AstC' were the active peptides, and AstC-R is their receptor.     www.nature.com/scientificreports/ Moreover, Aplysia AstC-R responded to all AstCs from the four other molluscan species (Fig. 7B). On the other hand, Class-A_GPCR1 did not respond significantly to any of the peptides tested (Fig. 7A,B), supporting the bioinformatic analysis showing that Class-A_GPCR1 is not an AstC receptor. Furthermore, for the six peptides that had a significant effect on AstC-R in the initial screening (Fig. 7A,B), we used multiple concentrations of the peptides, ranging from 10 -12 M to 10 -4 M to determine the dose response curve of peptide activation on the AstC-R (Fig. 7C-E). We found that log[EC 50 ] of AstC is the lowest, −8.2 ± 0.3 M (n = 3), whereas that of AstC' is the highest, −6.4 ± 0.5 M, which was significantly higher than that of AstC (Fig. 7F,G), indicating that the disulfide bridge plays an important role in the activity of the receptor. The peptides with the disulfide bridge but having different variations of amino acid sequences from four other molluscan species had log[EC 50 ] at intermediate values (Fig. 7F,G), indicating that both the disulfide bridge and amino acid sequences play some roles in the receptor activity.

Discussion
Growing databases of the Aplysia genome and transcriptome are becoming available, but relatively few studies have taken advantage of this information (see 53,54 ). Here, we have used bioinformatics, molecular biology, and a cell-based assay to demonstrate for the first time that an allatostatin C signaling system is present in the mollusc Aplysia. This appears to be the first example of a clear demonstration of an AstC signaling system in molluscs. Previous bioinformatic studies in several molluscs predicted the presence of an allatostatin C precursor (Fig. 3, e.g., 57 ) and of AstC with the disulfide bridge (Fig. 4) as well as the receptors (Fig. 6, e.g., 58,59 ). However, these studies did not report activation of any of the putative AstC receptors. In contrast, there is a recent work that identified allatostatin C and its putative receptor in an annelid, Platynereis dumerilii 10 . However, the EC 50 of this Platynereis receptor was quite high (1000 or 1200 nM) compared with the EC 50 of Aplysia AstC-R (7.8 nM). Thus, it is possible that there is another AstC receptor in Platynereis that could respond to AstC with a lower EC 50 . Annelids are related to molluscs in that they are both within the superphylum: lophotrochozoa.
In this age of genomics, a large amount of sequence information that is available has facilitated the analysis of evolutionary relationships of neuropeptide signaling systems across phyla 4,[8][9][10][11] , which has in turn predicted the presence of diverse peptide signaling systems in specific species. However, the ultimate evidence for the   Table 6 for information on the sequences). This phylogenetic tree indicates that Aplysia AstC-R is more closely related to a sequence in mollusc Biomphalaria. * indicates that the receptor has been studied/verified. "Parathyroid hormone receptor 1 C. gigas" is a Class B GPCR used as an outgroup (see Supplementary Table 5 www.nature.com/scientificreports/ presence of any particular signaling system in a species needs to be obtained using molecular biology, cell-based assays and other approaches that conclusively prove such a prediction to be correct. In the present work, we have successfully identified one precursor and one receptor for AstC ( Figs. 2A,B, 7) in Aplysia, although we cannot exclude the possibility as yet that there may be one more precursor or one more receptor, as in some arthropods. More work is needed to resolve these issues. Notably, we demonstrated the important role of the disulfide bridge in AstC because compared with the AstC with the disulfide bridge, the AstC without the disulfide bridge had a significantly higher log[EC 50 ] on AstC-R (Fig. 7F). In addition, the amino acid sequence of Aplysia AstC is also important for receptor activity because four molluscan AstCs with the disulfide bridge had intermediate log [EC 50 ] and the four sequences vary somewhat from Aplysia AstC, whereas Aplysia SPTR with a disulfide bridge is inactive apparently due to the fact that its sequence is very different from Aplysia AstC (Fig. 7). www.nature.com/scientificreports/ Our analyses indicate that the protein sequence encoded by the AstC precursor we identified is identical to ones available in current databases ( Fig. 1 and Supplementary Fig. 2), suggesting that sequences in the available public databases for Aplysia are credible and useful in predicting the precursor gene and putative peptides. In contrast, the receptors we identified required much more work to determine their sequences and whether they actually function as a receptor for AstC. Indeed, the significance of the present work can be appreciated by the fact that searching Aplysia somatostatin receptors have returned up to seven predicted sequences in NCBI, but only three of them are complete sequences, and among them, only one is significantly activated by Aplysia AstC. Thus, bioinformatics and experimental work are both necessary for identification of a peptide signaling system in a particular species. This is also a cautionary tale for evolutionary analysis of peptide precursors and especially receptors based only on bioinformatics.
Overall, our study is an important initial advance toward studying allatostatin C signaling system in Aplysia. Future work will use mass spectrometry (i.e., to demonstrate the specific form of AstC expressed in the Aplysia CNS), in situ hybridization, physiological techniques to provide further proof of a functional AstC signaling system in Aplysia, and to demonstrate its specific physiological roles. Indeed, our work provides an important basis for the study of functional roles of AstC signaling system invertebrates, other than arthropods. Although allatostatin C was initially found based on its bioassay to inhibit the biosynthesis of juvenile hormones in retrocerebral corpora allata in insects 3 , growing evidence suggests that it has multiple functions (see 60 ) as do many peptides 41,61,62 . For example, both allatostatin C 60 and somatostatin 14,15 may play a modulatory role in feeding. In addition, a recent work has shown that the AstC signaling system plays an important role in modulating circadian activity in Drosophila 63 , supporting diverse roles of AstC. Invertebrates other than arthropods, such as Aplysia, do not have corpora allata or juvenile hormones. Thus, it would be of great interest to determine whether Aplysia AstC plays an important role in modulating motivated behaviors such as feeding and locomotion. Because feeding and locomotor networks are well studied in Aplysia [17][18][19]21,[23][24][25][29][30][31][32][33]51,52,[64][65][66][67][68] , we expect that future studies will provide novel understanding of synaptic and circuit roles of the AstC signaling system in various behavioral networks. Finally, studying the AstC signaling system in molluscs will inform us how the genes and functions of the AstC and somatostatin signaling systems may have evolved across phyla.

Material and methods
Subjects and reagents. Experiments were performed on Aplysia californica (100-350 g) obtained from Marinus, California, USA. Aplysia are hermaphroditic (i.e., each animal has reproductive organs normally associated with both male and female sexes). Animals were maintained in circulating artificial seawater at 14-16 °C and the animal room was equipped with a 24 h light cycle with light period from 6:00 am to 6:00 pm. All chemicals were purchased from Sigma-Aldrich unless otherwise stated.

Bioinformatic analysis of peptide precursors and receptors.
We first used NCBI to search specific sequences of interests. In addition, we also searched AplysiaTools databases (Dr. Thomas Abrams, University of Maryland, USA) to obtain additional sequences for comparison. These latter databases (http:// aplys iatoo ls. org) include databases for Aplysia transcriptome and Aplysia genome.
The open reading frames (ORFs) of the AstC precursor, putative receptor full-length cDNA sequences were obtained using ORF Finder (https:// www. ncbi. nlm. nih. gov/ orffi nder/). For the AstC precursor, the putative signal peptide was predicted using SignalP-5.0 (http:// www. cbs. dtu. dk/ servi ces/ Signa lP/) and the putative peptides encoded by the AstC precursor were predicted using NeuroPred (http:// stagb eetle. animal. uiuc. edu/ cgi-bin/ neuro pred. py). We also compared the AstC precursor and neuropeptide sequences with those of other species using BioEdit software and generated a frequency plot of each amino acid (aligned from c-terminus) using a Weblogo software (http:// weblo go. berke ley. edu/ logo. cgi). For the putative AstC receptors, transmembrane domains were predicted using TMHMM Server v. 2.0 (http:// www. cbs. dtu. dk/ servi ces/ TMHMM/). For proteins that were difficult to annotate using blast, we also used Pfam database (http:// pfam. xfam. org/ search# tabvi ew= tab1) to determine what type of a protein it is. The phylogenetic trees of sequences from different species were constructed by MEGA X software (https:// www. megas oftwa re. net/) using alignment by MUSCLE and the maximum likelihood method with 1000 replicates. For Fig. 5B, we used LG + G + F model to generate our final tree; for Fig. 6, JTT model was performed which was different from Fig. 5B. The selection of the models was based on the results of MEGA analysis.
Cloning of mRNA in Aplysia. RNA extraction. After anesthesia with 30-50% of the body weight with 333 mM MgCl 2 , Aplysia cerebral, pleural-pedal, buccal and abdominal ganglia were dissected out and maintained in artificial seawater containing the following (in mM): 460 NaCl, 10 KCl, 55 MgCl 2 , 11 CaCl 2 , and 10 HEPES buffer, pH 7.6, in a dish lined with Sylgard (Dow Corning). RNA was prepared from the Aplysia ganglia using the TRIzol reagent method. Specifically, the dissected ganglia were placed into 200 μl TRIzol (Sigma, T9424) and stored at -80 °C until use. The frozen ganglia in TRIzol were thawed and homogenized with a plastic pestle, then TRIzol was added to a total volume of 1 ml, which were incubated at room temperature for 10 min. Then, 200 μl chloroform was added, and the solution was mixed thoroughly by shaking, and let stand on ice for 15 min. The solution was centrifuged (12,000 × g, 4 °C, 15 min), and the supernatant was added to an equal volume of isopropanol. The tube was shaken gently by hand and let stand at -20 °C for 2 h. After 2 h, it was centrifuged (12,000 × g, 4 °C, 15 min) again, the supernatant was discarded, 1 ml of 75% ethanol/water was added, and the centrifuge tube was shaken gently by hand to suspend the pellet. It was centrifuged (12,000 × g, 4 °C, 10 min), the supernatant discarded and the precipitant was dried at room temperature for 5-10 min. Finally, 30 μl of nuclease-free water was added to dissolve the RNA pellet, and the RNA concentration was determined with a Nanodrop ND-1000 spectrophotometer (Thermo Fisher Scientific). IP1 accumulation assay. Inositol monophosphate 1 (IP1) accumulation assay measures concentration of IP1, that is hydrolyzed from the second messenger, inositol monophosphate 3 (IP3), generated by Gαq pathway when a G-protein coupled receptor (GPCR) expressed in CHO-K1 cells is activated by an appropriate ligand. In order to express the Aplysia putative receptors transiently in CHO-K1, the cDNA was cloned into the mammalian expression vector pcDNA3.1( +). CHO-K1 cells (Procell, CL-0062) were cultured in F-12K medium (Gibco, 21127-022) with 10% fetal bovine serum (Genial, G11-70500) at 37 °C in 5% CO 2 . Transfection experiments were performed when the cells were grown to 70-90% confluence. In preliminary experiments, for each dish (60-mm diameter), 4 μg of the putative receptor plasmids [in pcDNA3.1( +)] were mixed with 400 μl of Opti-MEM (Gibco, 11058021), followed by the addition of 15 μl of Turbofect (Thermo Fisher Scientific, R0531). However, none of the receptors responded to AstC or AstC gene-derived peptide (AstCG-DP, see Fig. 3B and Results), suggesting that these receptors may not associate with the native Gαq in CHO cells. Thus, for all experiments shown in Fig. 7, 3 μg of the putative receptor plasmids [in pcDNA3.1( +)] and 3 μg of Gα16 plasmids [in pcDNA3.1( +)] were co-transfected in the above procedure to ensure that the ligand-receptor binding would generate a IP1 response because Gα16 is a promiscuous Gαq protein that will bind to most GPCRs 10,53 . The CHO cells with the reagents added above were mixed gently, and incubated at room temperature for 15 min. The DNA/Turbofect mixture dropwise was then added to the dish, and the cells were incubated at 37 °C in 5% CO 2 overnight. The next day, the cells were trypsinized and reseeded in opaque white 96-well half-area (Corning, 3688) or 384-well tissue culture-treated plates (Corning, 3570) at a density of 20,000 cells/well in F-12K and 10% FBS and incubated at 37 °C in 5% CO 2 overnight. On the third day, the activation of the putative receptor was detected by monitoring IP1 accumulation using IP1 detection kit (Cisbio, 62IPAPEB) in Tecan Spark. Except for using 0.5 × reagent, all other procedures were performed in accordance with the IP1 detection kit manufacturer's instructions. Peptides are synthesized by Sangon Biotech, Guoping Pharmaceutical or ChinaPeptides (Supplementary Fig. 4), and are aliquoted in 50 nmol EP tubes, stored at −20 °C until use.