Diaphorina citri Genome Possesses a Complete Melatonin Biosynthesis Pathway Differentially Expressed under the Influence of the Phytopathogenic Bacterium, Candidatus Liberibacter asiaticus

Simple Summary The indole-like compound, melatonin, is a tryptophan-derivative that is secreted by the pineal gland. Melatonin is ubiquitously distributed in both prokaryotes and eukaryotes including animals and plants. In animals, melatonin plays pleiotropic regulatory roles in several biological and physiological functions including sleep, circadian rhythm, oxidative stress, immune response, aging, apoptosis, and autophagy. Moreover, it might have anti-inflammatory, anti-tumor, and anti-cancer activities. Although most, if not all, of these genes were cloned and characterized previously in numerous animal species, none of them have been identified from the Asian citrus psyllid, Diaphorina citri, in the vector of Huanglongbing yet. In the current study, we performed a genome-wide analysis and introduces a shortlist of six putative melatonin biosynthesis-related genes included two putative tryptophan 5-hydroxylase (DcT5H-1 and DcT5H-2), a putative aromatic amino acid decarboxylase (DcAADC), two putative arylalkylamine N-acetyltransferase (DcAANAT-1 and DcAANAT-2), and putative N-acetylserotonin O-methyltransferase (DcASMT), which could indicate sites of functional or structural constraint. All these genes were differentially expressed under the influence of the phytopathogenic bacterium, Candidatus Liberibacter asiaticus, and after melatonin supplementation. Our findings could be a further step for optimization and cloning of melatonin biosynthesis genes of Diaphorina citri. Abstract Melatonin is synthesized from the amino acid L-tryptophan via the shikimic acid pathway and ubiquitously distributed in both prokaryotes and eukaryotes. Although most of melatonin biosynthesis genes were characterized in several plants and animal species including the insect model, Drosophila melanogaster, none of these enzymes have been identified from the Asian citrus psyllid, Diaphorina citri. We used comprehensive in silico analysis and gene expression techniques to identify the melatonin biosynthesis-related genes of D. citri and to evaluate the expression patterns of these genes within the adults of D. citri with gradient infection rates (0, 28, 34, 50, 58, and 70%) of the phytopathogenic bacterium Candidatus Liberibacter asiaticus and after the treatment with exogenous melatonin. We showed that the D. citri genome possesses six putative melatonin biosynthesis-related genes including two putative tryptophan 5-hydroxylase (DcT5H-1 and DcT5H-2), a putative aromatic amino acid decarboxylase (DcAADC), two putative arylalkylamine N-acetyltransferase (DcAANAT-1 and DcAANAT-2), and putative N-acetylserotonin O-methyltransferase (DcASMT). The infection with Ca. L. asiaticus decreased the transcript levels of all predicted genes in the adults of D. citri. Moreover, melatonin supplementation induced their expression levels in both healthy and Ca. L. asiaticus-infected psyllids. These findings confirm the association of these genes with the melatonin biosynthesis pathway.

The listed putative gene candidates were identified based on the top-matches of recent available data in GenBank, the national center for biotechnology information website (NCBI, http://www.ncbi.nlm.nih.gov/ gene/, 12 February 2021) with the query sequences. The shortlist of top-matches was generated based on the phylogenetic trees, identity more than 50% (except for DcAANATs), and excluding all the hypothetical and low-quality proteins that have these characteristics. b The listed genes are the top-matched sequences of the "Diaphorina citri OGS v2.0 CDS" and "Diaphorina citri OGS v2.0 proteins" BLAST datasets available on the Citrus Greening Solutions website (https://citrusgreening.org/organism/Diaphorina_citri/genome, 12 February 2021) using the Nucleotide-Nucleotide BLAST (BLASTn) and Protein-Protein BLAST (BLASTp) tools [56,57] using the NCBI query sequences. c Top-matched sequences from D. citri with tryptophan hydroxylase (GenBank Accession no. NP_612080.1) from fruit fly (Drosophila melanogaster) [49] using the Protein-Protein BLAST (BLASTp), based on recent available data in GenBank, the national center for biotechnology information website (NCBI, http://www.ncbi.nlm.nih.gov/gene/, 12 February 2021), using the compositionally adjusted substitution matrices [57]. d Top-matched sequences from D. citri with dopa decarboxylase, isoform B (GenBank Accession no. NP_724164.1) from fruit fly (D. melanogaster) [51,52] using the Protein-Protein BLAST (BLASTp), based on recently available data in GenBank, the national center for biotechnology information website (NCBI, http://www.ncbi.nlm.nih.gov/gene/, 12 February 2021), using the compositionally adjusted substitution matrices [57]. e Top-matched sequences from D. citri with arylalkylamine N-acetyltransferase 1, isoform A (GenBank Accession no. NP_523839.2) from the fruit fly (D. melanogaster) [53,54,59] using the Protein-Protein BLAST (BLASTp), based on recently available data in GenBank, the national center for biotechnology information website (NCBI, http://www.ncbi.nlm.nih.gov/gene/, 12 February 2021), using the compositionally adjusted substitution matrices [57]. f Top-matched sequences from D. citri with N-acetylserotonin O-methyltransferase-like protein, isoform X1 (GenBank Accession no. XP_014251646.1) from bed bug (Cimex lectularius) using the Protein-Protein BLAST (BLASTp), based on recently available data in GenBank, the national center for biotechnology information website (NCBI, http://www.ncbi.nlm.nih.gov/gene/, 12 February 2021), using the compositionally adjusted substitution matrices [57].

Evolutionary Analysis by Maximum Likelihood Method
The evolutionary history of all matched sequences for each gene was inferred using the maximum likelihood method and JTT matrix-based model [60]. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the JTT model, and then selecting the topology with a superior log-likelihood value. Evolutionary analyses were conducted in MEGA X [61].

Multiple Sequence Alignment Analysis
Amino acid sequences from D. citri that produced significant alignments with known melatonin biosynthesis-related genes were simultaneously aligned using the Constraint-Based Alignment tool (COBALT; https://www.ncbi.nlm.nih.gov/tools/cobalt/re_cobalt. cgi, 12 February 2021) for multiple protein sequences [62]. Moreover, the top-matched sequences (amino acid and nucleotide sequences) of D. citri producing significant alignments with known melatonin-biosynthetic genes were used to generate the multiple sequence alignment by ClustalW (http://www.genome.jp/tools-bin/clustalw, 12 February 2021) [63], and the version 3.21 of BOXSHADE (https://embnet.vital-it.ch/software/BOX_form.html, 12 February 2021) was used to visualize conserved regions in the alignment.

Rearing of Healthy and Ca. L. asiaticus-Infected D. citri Colonies
Healthy colonies of D. citri were continuously reared in a secured growth room (27 ± 2 • C, 60 ± 5% relative humidity, and 16:8 h L/D photocycle) at Citrus Research and Education Center (CREC), University of Florida (28 • 10 N, 81 • 71 E), Lake Alfred, FL, USA. Insects were maintained on HLB-free alemow (Citrus macrophylla) trees. Random samples of D. citri adults and alemow leaves were collected monthly and tested for the presence of Ca. L. asiaticus using polymerase chain reaction (PCR) assay as described by Tatineni et al. [70] to ensure that colonies remained HLB-free.
To obtain the Ca. L. asiaticus-infected colonies, D. citri from the healthy colonies were reared on HLB-symptomatic and PCR-positive Ca. L. asiaticus-infected alemow trees and maintained in a separate secured growth room, under the same conditions as described above, to avoid cross-contamination. To obtain gradient infection rates of D. citri, the infection rates were tested monthly and prior to each experiment using two different methods: (I) Testing the presence of Ca. L. asiaticus visually in 50 individual psyllids using binocular laboratory compound microscope; and (II) the infection rates with Ca. L. asiaticus were further confirmed by PCR as described by Tatineni et al. [70]. Based on these examinations, D. citri colonies were categorized into five infection rates (24, 34, 50, 58, and 70%), in addition to the healthy colony (0%), which were all tested in this study. The infection rates were consistent without significant differences between them throughout the whole study. Briefly, newly emerged adults (~2-days old) from each infection rate were collected using an aspirator without any sex-based discrimination.

Treatment with Exogenous Melatonin
Newly emerged D. citri adults (healthy versus Ca. L. asiaticus-infected (50% infection rate)) were caged in the feeding system described in our previous study [15] and were fed on 100 µL of 20% sucrose suspension as an artificial diet (mock control versus 500 µg mL −1 of melatonin in 20% sucrose solution). The artificial diet was placed between double-layered parafilm stretched over an acrylic feeding chamber that contained 50 adults (10 replicates/treatment). Insects were maintained at the same conditions as described above, and 72 h post-treatment (hpt), psyllids were collected using an aspirator for gene expression analysis.

Gene Expression Analysis Using Quantitative Real-Time PCR (RT-PCR)
Total RNA was extracted from five individual insects per replicate (10 replicates/treatment) using TriZol ® reagent (Ambion ® , Life Technologies, New York, NY, USA), and the gene expression analysis was carried out as described in our previous study [15]. Quantification of transcript levels was used as a measure of the gene expression. Samples were analyzed in triplicate for each biological replicate. Primers for 6 melatonin biosynthesis-related genes (Table S1) were used to measure the gene expression. The relative expression of the consensus sequence among PCR products was determined according to the 2 −∆∆C T method [71]. For all gene expression experiments, data were normalized using two reference genes, α-Tubulin and actin, which previously showed high stability for transcript normalization in D. citri under biotic stress [14].

Statistical Analysis
All experiments were designed in a completely randomized design using 10 biological replicates per treatment. The analysis of variance technique (ANOVA) was used for statistical comparison between more than two treatments, followed by post-hoc pairwise comparisons using the Tukey-Kramer honestly significant difference test (Tukey HSD; p < 0.05). Additionally, a two-tailed t-test was used for statistical comparison between only two treatments (healthy versus infected) or (mock-treated versus melatonin-treated), and statistical significance was established as p < 0.05. Moreover, simple linear regression (SLR) analysis was performed to model the relationship between Ca. L. asiaticus infection rates (as an independent variable) and gene expression (as a dependent variable). The fitted regression line is expressed as a significant equation, as determined by the F test (p < 0.05). Both coefficients of determination (R 2 ) and adjusted coefficient of determination (R 2 adj ) were also obtained. Further, due to the observed nonlinear phenomena, data were fitted with a second-degree polynomial regression model (quadratic model) to understand the curvilinear relationship between Ca. L. asiaticus infection rates (as an independent variable) and gene expression (as a dependent variable). Polynomial regression models, the 95% confident curves for the estimated regression, quadratic equation, R 2 , R 2 adj , and p-value based on the F test (p < 0.05) were also obtained. JMP Statistical Software (SAS Institute, Cary, NC, USA) was used for all statistical analyses listed above.

D. citri Genome Possesses a Putative Melatonin Biosynthetic Pathway
The predicted melatonin biosynthesis pathway in D. citri was dissected using a comparative in silico analysis. Putative candidate genes involved in the melatonin biosynthesis pathway were presented as the top-matched sequences producing significant alignments of melatonin-biosynthetic genes from model insects and were selected based on sequence similarity, the phylogenetic relationships with the query sequences, and based on the sequence identity between query sequences and predicted ones, after excluding all the hypothetical and low-quality proteins that have these characteristics (Table 1). 3.1.1. D. citri Genome Encodes for two Putative Tryptophan 5-hydroxylase (DcT5H) Using the Protein-Protein BLAST (BLASTp) tool, our findings showed that the D. citri genome possesses about six predicted amino acid sequences (based on NCBI database, Table S2) and nine sequences (based on Diaphorina citri OGS v2.0 proteins dataset, Table S3) that produce significant similarities to tryptophan hydroxylase (DmT5H, GenBank Accession no. NP_612080.1) from fruit fly (D. melanogaster). Although the multiple protein sequence alignment using COBALT analysis showed that all predicted sequences have relatively high homology with DmT5H protein, the phylogenetic analysis showed that only two proteins from D. citri were phylogenetically closer to the query sequence (approximately 55%) ( Figure 1A). Those two proteins included putative tryptophan 5-hydroxylase 1-like (henceforth DcT5H-1) encoded by the D. citri locus LOC113470334 (GenBank Accession no. XP_026684504.1) and protein henna-like (henceforth DcT5H-2) by the D. citri locus LOC103524631 (GenBank Accession no. XP_017305180.1) ( Figure 1A).

D. citri Genome Possesses a Putative Melatonin Biosynthetic Pathway
The predicted melatonin biosynthesis pathway in D. citri was dissected using a comparative in silico analysis. Putative candidate genes involved in the melatonin biosynthesis pathway were presented as the top-matched sequences producing significant alignments of melatonin-biosynthetic genes from model insects and were selected based on sequence similarity, the phylogenetic relationships with the query sequences, and based on the sequence identity between query sequences and predicted ones, after excluding all the hypothetical and low-quality proteins that have these characteristics (Table 1). 3.1.1. D. citri Genome Encodes for two Putative Tryptophan 5-hydroxylase (DcT5H) Using the Protein-Protein BLAST (BLASTp) tool, our findings showed that the D. citri genome possesses about six predicted amino acid sequences (based on NCBI database, Table S2) and nine sequences (based on Diaphorina citri OGS v2.0 proteins dataset, Table S3) that produce significant similarities to tryptophan hydroxylase (DmT5H, Gen-Bank Accession no. NP_612080.1) from fruit fly (D. melanogaster). Although the multiple protein sequence alignment using COBALT analysis showed that all predicted sequences have relatively high homology with DmT5H protein, the phylogenetic analysis showed that only two proteins from D. citri were phylogenetically closer to the query sequence (approximately 55%) ( Figure 1A). Those two proteins included putative tryptophan 5-hydroxylase 1-like (henceforth DcT5H-1) encoded by the D. citri locus LOC113470334 (Gen-Bank Accession no. XP_026684504.1) and protein henna-like (henceforth DcT5H-2) by the D. citri locus LOC103524631 (GenBank Accession no. XP_017305180.1) ( Figure 1A). gure 1. In silico analysis of tryptophan 5-hydroxylase (DcT5H) of Diaphorina citri. (A) Evolutionary analysis using aximum likelihood method and its associated multiple protein sequences alignments using Constraint-Based Alignment ol (COBALT) analysis. The AA sequences were identified using Protein-Protein BLAST (BLASTp) using tryptophan droxylase (DmT5H; GenBank Accession no. NP_612080.1) from Drosophila melanogaster, as a query sequence, against iaphorina citri genome available in GenBank, the national center for biotechnology information website (NCBI, tp://www.ncbi.nlm.nih.gov/gene/, 12 February 2021). The tree with the highest log likelihood (−5340.89) is shown. The ee is drawn to scale, with branch lengths measured in the number of substitutions per site (next to the branches). The oportion of sites where at least one unambiguous base is present in at least 1 sequence for each descendent clade is own next to each internal node in the tree. Evolutionary analyses and the joint tree were conducted in MEGA-X software. (A) Evolutionary analysis using maximum likelihood method and its associated multiple protein sequences alignments using Constraint-Based Alignment tool (COBALT) analysis. The AA sequences were identified using Protein-Protein BLAST (BLASTp) using tryptophan hydroxylase (DmT5H; GenBank Accession no. NP_612080.1) from Drosophila melanogaster, as a query sequence, against Diaphorina citri genome available in GenBank, the national center for biotechnology information website (NCBI, http: //www.ncbi.nlm.nih.gov/gene/, 12 February 2021). The tree with the highest log likelihood (−5340.89) is shown. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site (next to the branches). The proportion of sites where at least one unambiguous base is present in at least 1 sequence for each descendent clade is shown next to each internal node in the tree. Evolutionary analyses and the joint tree were conducted in MEGA-X software. In the COBALT analysis, residues were colored using a column-based method according to their relative entropy threshold. Aligned columns with no gaps are colored blue and red, where the red color indicates highly conserved columns and blue indicates less conserved ones. (B) The protein functional and conserved domains analysis of DmT5H (NP_612080.1), DcT5H-1 (XP_026684504.1), and DcT5H-2 (XP_017305180.1) using the InterPro Scan tool (https://www.ebi.ac.uk/interpro/, 12 February 2021). FYWHYDRXLASE: Biopterin-dependent aromatic amino acid hydroxylase signature; ArAA_hydroxylase_Fe/CU: Aromatic amino acid hydroxylase, iron/copper-binding site; and BH4_AAA_HYDROXYL_1: Non-heme iron and tetrahydrobiopterin (BH4)-dependent enzymes.
The NCBI protein sequences of DcT5H-1 and DcT5H-2 were aligned with the sequence of the top-matched proteins from the D. citri database (DcitrP076520.1.1 and Dc-itrP012845.1.1, respectively). The AA alignment showed high similarity and conserved sequences in both proteins ( Figures S1 and S2). Furthermore, the nucleotide sequence of DcT5H-1 (GenBank Accession no. XM_026828703.1) and DcT5H-2 (GenBank Accession no. XM_017449691.2) had high similarity and conserved sequences when aligned with the mRNA sequences from the D. citri database (DcitrC076520.1.1 and DcitrC012845.1.1, respectively) ( Figures S3 and S4). Collectively, these findings suggest sequences retrieved from the NCBI database were highly similar and homology with those of the D. citri database. Therefore, we focused on these proteins for further in silico analysis.

D. citri Genome Encodes for a Putative Aromatic Amino Acid Decarboxylase (DcAADC)
In silico analysis using the BLASTp tool showed that the D. citri genome possesses about 15 sequences (based on NCBI database, Table S4) and 16 sequences (based on Diaphorina citri OGS v2.0 proteins dataset, Table S5) that produce significant similarities to dopa decarboxylase, isoform B (DmDDC, also known as aromatic L-amino acid decarboxylase (DmAADC); GenBank Accession no. NP_724164.1) from D. melanogaster. The multiple protein sequence alignment using COBALT analysis using the NCBI sequences showed that all predicted sequences have relatively high homology with DmDDC protein ( Figure 4A). The phylogenetic analysis showed that only three proteins from D. citri were phylogenetically closer to the query sequence ( Figure 4A). These proteins included aromatic L-amino acid decarboxylase, isoform X1 (henceforth DcAADC-1) encoded by the D. citri locus LOC103520978 (481aa; GenBank Accession no. XP_008484302.1), aromatic L-amino acid decarboxylase (henceforth DcAADC-2) encoded by the D. citri locus LOC103510318 (484; GenBank Accession no. XP_017300015.1), and aromatic L-amino acid decarboxylase-like encoded by the D. citri locus LOC103510317 (93 aa; GenBank Accession no. XP_026680193.1) ( Figure 4A). However, the latest sequence was excluded from our further analysis because it was very short compared with the query sequence, and it had low query cover (14% ; Table S4).
The prediction of the conserved domains using the InterPro Scan tool suggests a high topological similarity among DmAADC, DcAADC-1, and DcAADC-2 ( Figure 4B). All sequences had two families included aromatic L-amino acid decarboxylase (IPR010977) and pyridoxal phosphate-dependent decarboxylase (IPR002129); three homologous pyridoxal phosphate-dependent transferase superfamilies (IPR015424, IPR015421, and IPR015422); and two binding sites including the pyridoxal-phosphate binding site (IPR021115) and DDC/GAD/HDC/TyrDC pyridoxal-phosphate attachment site (PS00392) ( Figure 4B).  shown. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site (next to the branches). The proportion of sites where at least one unambiguous base is present in at least one sequence for each descendent clade is shown next to each internal node in the tree. Evolutionary analyses and the joint tree were conducted in MEGA-X software. In the COBALT analysis, residues were colored using a column-based method according to their relative entropy threshold. Aligned columns with no gaps are colored blue and red, where the red color indicates highly conserved columns and blue indicates less conserved ones.  The proportion of sites where at least one unambiguous base is present in at least one sequence for each descendent clade is shown next to each internal node in the tree. Evolutionary analyses and the joint tree were conducted in MEGA-X software.
Likewise, DcAADC-1 and DcAADC-2 were also predicted as homodimers that combined two units (model A and model B) with slight differences between both units in each model ( Figure 5). Briefly, the predicted model of DcAADC-1 covered 98% with the target protein (residues Asp 3 to Glu 476) with notable statistics (GMQE = 0.87, QMEAN = −0.90, QSQE = 0.98, seq identity = 73.78%, seq similarity = 54%, and confidence = 100%) and were composed of 21 α-helix ribbons for model A or 22 α-helices for model B and 17 stranded β-sheets for each model (six of them were short) ( Figure 5D,E, respectively) with considerable predicted local similarity to target ( Figure 5F). The extra α-helix of model B (α18) was represented by four residues including Leu 334, Lys 335, His 336, and Asp 337 ( Figure 5E).
Furthermore, the structure and/or function of the three protein sequences (DmAADC, DcAADC-1, and DcAADC-2) were deeply analyzed using the Phyre2 protein fold recognition server ( Figure 6). The crystallographic 3D structures of the three proteins were predicted using the same protein template (PDB ID 3k40.1.A). All 3D structures were predicted as monomers and they were almost identical ( Figure 6A-C) except for only one αhelix ribbon that was present in DmAADC, DcAADC-1 ( Figure 6D,E, respectively), but it was absent in DcAADC-2 ( Figure 6F). Additionally, Phyre2-based predicted topology suggested that the three predicted proteins (DmAADC, DcAADC-1, and DcAADC-2) might act as transporters. The predicted topology of DmAADC showed that it contains four transmembrane domains (S1-S4) (three connecting loops, and internal N-and C-termini ( Figure 6G)) DcAADC-1 contains five transmembrane domains (S1-S5) (four connecting loops, internal N-terminal, and external C-terminal ( Figure 6H)), and DcAADC-2 contains six transmembrane domains (S1-S6) (five connecting loops, and internal N-and C-terminal extremities ( Figure 6I)).   Figure 7B) and centroid (normalized MFE = −0.2374; Figure 7E) secondary structures. Moreover, the mountain plot representation of the MFE structure, the centroid structure, the thermodynamic ensemble of RNA structures, and the positional entropy of DmAADC ( Figure 7G), DcAADC-1 ( Figure 7H), and DcAADC-2 ( Figure 7I) suggested that all predicted secondary RNA structures were thermodynamically stable. Further, no significant disparities were observed among all predicted mRNA hairpins which can be considered as proof of the stability of the secondary structures. BLASTp analysis showed that D. citri genome encodes three AA sequences (based on NCBI database, Table S6) and only two protein sequences (based on Diaphorina citri OGS v2.0 proteins dataset, Table S7) with significant similarity to arylalkylamine Nacetyltransferase 1, isoform A (DmAANAT1, also known as DmSNAT; GenBank Accession no. NP_523839.2, 240 aa) from the fruit fly (D. melanogaster). The COBALT-based multiple protein sequences alignment showed that the predicted proteins had relatively high homology with DmAANAT1 protein; however, Dopamine N-acetyltransferase-like, isoform X1 (henceforth DcAANAT-1) encoded by the D. citri locus LOC103507708 (GenBank Accession no. XP_026678312.1; 217 aa) and dopamine N-acetyltransferase-like (henceforth DcAANAT-2) encoded by the D. citri locus LOC103507696 (GenBank Accession no. XP_017298946.1; 220 aa) were phylogenetically closer to DmAANAT1 ( Figure 8A).
The protein tertiary structures of DmAANAT1, DcAANAT-1, and DcAANAT-2 were predicted using the crystal structure of D. melanogaster dopamine N-acetyltransferase in The tree is drawn to scale, with branch lengths measured in the number of substitutions per site (next to the branches). The proportion of sites where at least one unambiguous base is present in at least 1 sequence for each descendent clade is shown next to each internal node in the tree. Evolutionary analyses and the joint tree were conducted in MEGA-X software. In the COBALT analysis, residues were colored using a column-based method according to their relative entropy threshold. Aligned columns with no gaps are colored blue and red, where the red color indicates highly conserved columns and blue indicates less conserved ones. InterPro-based analysis of conserved domains suggests a comparable topological similarity between DmAANAT1, DcAANAT-1, and DcAANAT-2 ( Figure 8B). All sequences had an acyl-CoA N-acyltransferase (IPR016181) homologous superfamily, and three unintegrated domains including dopamine N-acetyltransferase (PTHR20905:SF31), NAT_SF (cd04301), and N-acetyltransferase-related (PTHR20905) ( Figure 8B). However, a Gcn5related N-acetyltransferases (GNAT) domain (IPR000182) was predicted only in DmAANAT1 from D. melanogaster, but not in any DcAANAT genes from D. citri ( Figure 8B).

D. citri Genome Encodes for a Putative N-acetylserotonin O-methyltransferase (DcASMT)
Digging the D. citri genome using the BLASTp tool retrieved two sequences from th NCBI database (Table S8) and only one sequence from the Diaphorina citri OGS v2.0 pr teins dataset (Table S9) that produced significant similarities to N-acetylserotonin O-m thyltransferase-like protein, isoform X1 (ClASMT; 260 aa; GenBank Accession n XP_014251646.1) from bed bug (Cimex lectularius). The multiple protein sequence alig ment using COBALT analysis showed that all predicted sequences have relatively hig homology with ClASMT protein ( Figure 11A). One of these sequences (103 aa; GenBan Accession no. XP_026688590.1) was excluded because it was a partial sequence. Howeve septum formation protein Maf-like encoded by the D. citri locus LOC113468045 (162 a GenBank Accession no. XP_026680472.1) was used for further analysis. Interestingly, th Digging the D. citri genome using the BLASTp tool retrieved two sequences from the NCBI database (Table S8) and only one sequence from the Diaphorina citri OGS v2.0 proteins dataset (Table S9) that produced significant similarities to N-acetylserotonin O-methyltransferase-like protein, isoform X1 (ClASMT; 260 aa; GenBank Accession no. XP_014251646.1) from bed bug (Cimex lectularius). The multiple protein sequence alignment using COBALT analysis showed that all predicted sequences have relatively high homology with ClASMT protein ( Figure 11A). One of these sequences (103 aa; GenBank Accession no. XP_026688590.1) was excluded because it was a partial sequence. However, septum formation protein Maf-like encoded by the D. citri locus LOC113468045 (162 aa; GenBank Accession no. XP_026680472.1) was used for further analysis. Interestingly, this AA sequence was highly similar and showed conserved sequences to N-acetylserotonin O-methyltransferase-like protein from the D. citri database (henceforth DcASMT; 274 aa; DcitrP032285.1.1) ( Figure S11). Likewise, the nucleotide sequence of DcASMT (746 bp, Gen-Bank Accession no. XM_026824671.1) had high similarity and conserved sequences when aligned with the mRNA sequences from the D. citri database (825 bp, DcitrC032285.1.1) ( Figure S12). Together, these findings suggest that the DcASMT sequence retrieved from the NCBI database presented the same protein sequence from the D. citri database.  Figure S11). Likewise, the nucleotide sequence of DcASMT (746 b GenBank Accession no. XM_026824671.1) had high similarity and conserved sequen when aligned with the mRNA sequences from the D. citri database (825 b DcitrC032285.1.1) ( Figure S12). Together, these findings suggest that the DcASMT quence retrieved from the NCBI database presented the same protein sequence from t D. citri database. The prediction of the conserved domains using the InterPro Scan tool suggests a hi topological similarity among ClASMT and DcASMT ( Figure 11B). Both sequences had t domains included nucleoside triphosphate pyrophosphatase Maf-like prot (IPR003697) and bifunctional DTTP/UTP pyrophosphatase/methyltransfera (PTHR43213); a homologous inosine triphosphate pyrophosphatase-like superfam (IPR029001); and an unintegrated bifunctional DTTP/UTP pyrophosphatase/methyltra ferase (PTHR43213:SF5) domain ( Figure 11B).
The crystallographic 3D structures of ClASMT and DcASMT were predicted usi the crystal structure of the Maf domain of human N-acetylserotonin O-methyltransfera like protein (PDB ID: 2p5x.2.A) and refined to 2.00 Å resolution with good statistics (F ure 12). ClASMT was predicted as a monomer that represented approximately 76% co erage (residues Leu 9 to Asp 203) with remarkable statistics (GMQE = 0.58, QMEAN −0.06, QSQE = 0.00, seq identity = 46.46%, seq similarity = 43%, and confidence = 100 ( Figure 12A). The predicted model of ClASMT composed of nine α-helices and 11 β-she ( Figure 12A,B) with good local quality estimate ( Figure 12C). Likewise, about 93% Figure 11. In silico analysis of N-acetylserotonin O-methyltransferase (DcASMT) of Diaphorina citri. (A) Evolutionary analysis using maximum likelihood method and its associated multiple protein sequences alignments using Constraint-Based Alignment tool (COBALT) analysis. The AA sequences were identified using the Protein-Protein BLAST (BLASTp) using N-acetylserotonin O-methyltransferase-like protein, isoform X1 (ClASMT; GenBank Accession no. XP_014251646.1) from bed bug (Cimex lectularius) as a query sequence, against Diaphorina citri genome available in GenBank, the national center for biotechnology information website (NCBI, http://www.ncbi.nlm.nih.gov/gene/, 12 February 2021). The tree is drawn to scale, with branch lengths measured in the number of substitutions per site (next to the branches). The proportion of sites where at least one unambiguous base is present in at least 1 sequence for each descendent clade is shown next to each internal node in the tree. Evolutionary analyses and the joint tree were conducted in MEGA-X software. In the COBALT analysis, residues were colored using a column-based method according to their relative entropy threshold. Aligned columns with no gaps are colored blue and red, where the red color indicates highly conserved columns and blue indicates less conserved ones. (B) The protein functional and conserved domains analysis of ClASMT (XP_014251646.1) and DcASMT (XP_026680472.1) using the InterPro Scan tool (https://www.ebi.ac.uk/interpro/, 12 February 2021).
Additionally, the mRNA hairpins of ClASMT (1091 bp; XM_014396160.1) and DcASMT (746 bp; GenBank Accession no. XM_026824671.1) were predicted using the nucleotide sequence with strengths of base pairing probabilities of minimum free energy (MFE; Figure 13A,B, Insects 2021, 12, 317 23 of 34 respectively) and centroid secondary structures ( Figure 13C,D, respectively). The RNAfold analysis demonstrated that the mRNA hairpins of DmASMT and DcASMT could produce stable MFE secondary structures (MFE = −255.50 and −190.80 Kcal/mol, respectively) and also stable centroid secondary structures (MFE = −197.20 and −152.04 Kcal/mol, respectively). Moreover, the mountain plot representation of the MFE structure, the centroid structure, the thermodynamic ensemble of RNA structures, and the positional entropy of DmASMT ( Figure 13E) and DcASMT ( Figure 13F) suggested that both predicted secondary RNA structures were thermodynamically stable.  Figure 13E) and DcASMT ( Figure 13F) suggested that both predicted secondary RNA structures were thermodynamically stable. We investigated the transcript levels of six melatonin biosynthesis-related genes in D. citri (Figure 14). These genes included two DcT5Hs (DcT5H-1 and DcT5H-2), DcAADC,

Ca. L. asiaticus Infection Downregulated the Expression of Melatonin Biosynthesis-Related Genes of D. citri
We investigated the transcript levels of six melatonin biosynthesis-related genes in D. citri (Figure 14). These genes included two DcT5Hs (DcT5H-1 and DcT5H-2), DcAADC, two DcAANATs (DcAANAT-1 and DcAANAT-2), and DcASMT ( Figure 14A-F, respectively). , respectively. Gene expressions were normalized using two housekeeping genes (α-tubulin and actin), and the changes were analyzed using the 2 −ΔΔC T method. Open dots indicate the raw data (n = 10), the black ones indicate potential outliers, whereas horizontal thick lines indicate the medians. Whiskers indicate the minimum, and the maximum values of the data and boxes show the interquartile ranges (twenty-fifth to the seventy-fifth percentile of the data). Different letters indicate statistically significant differences among different infection rates, while the same letter signifies no significant differences between them using Tukey-Kramer honestly significant difference test (Tukey HSD; p < 0.05). The full list of expressed genes, names, accession numbers, and primers are available in Table 1.   , aromatic amino acid decarboxylase (DcAADC), arylalkylamine N-acetyltransferase genes (DcAANAT-1 and DcAANAT-2), and N-acetylserotonin O-methyltransferase (DcASMT), respectively. Gene expressions were normalized using two housekeeping genes (α-tubulin and actin), and the changes were analyzed using the 2 −∆∆C T method. Open dots indicate the raw data (n = 10), the black ones indicate potential outliers, whereas horizontal thick lines indicate the medians. Whiskers indicate the minimum, and the maximum values of the data and boxes show the interquartile ranges (twenty-fifth to the seventy-fifth percentile of the data). Different letters indicate statistically significant differences among different infection rates, while the same letter signifies no significant differences between them using Tukey-Kramer honestly significant difference test (Tukey HSD; p < 0.05). The full list of expressed genes, names, accession numbers, and primers are available in Table 1. (G-L) Simple linear regression and quadratic polynomial regression analysis between Ca. L. asiaticus infection rates and the expression levels of DcT5H-1, DcT5H-2, DcAADC, DcAANAT-1, DcAANAT-2, and DcASMT, respectively. Open dots present the row data (n = 10). The fitted regression line is presented as a black solid-line, while the polynomial regression model is presented as a blue dashed line. The 95% confidence intervals (95% CI) for the estimated polynomial regression are blue-shaded and edged by doted-lines. Regression equations, R 2 , R 2 adj , and p-value based on the F test (p < 0.05) were also obtained and presented within the graph.
Our findings showed that the Ca. L. asiaticus infection significantly reduced the transcript levels of all studied melatonin biosynthesis-related genes. The downregulation of melatonin biosynthesis-related genes was proportionally over the Ca. L. asiaticus infection rates. Although the transcript levels of DcT5H-1 were gradually decreased with Ca. L. asiaticus infection rates, no significant differences were observed between the high Ca. L. asiaticus infection rates (58 and 70% Ca. L. asiaticus) ( Figure 14A). Likewise, Ca. L. asiaticus infection decreased the expression levels of DcT5H-2; however, no significant differences were observed between low infection rates (24 and 35%), and also no significant differences were observed between the high infection rates (50, 58 and 70 % Ca. L. asiaticus) ( Figure 14B). Moreover, DcAADC was downregulated in Ca. L. asiaticus-infected D. citri compared with healthy insects, without significant differences among the high infection rates (50,58, and 70%) ( Figure 14C). On the other hand, low infection rates (less than 50%) did not affect the expression levels of DcAANAT-1 and DcAANAT-2 ( Figure 14D,E, respectively). The transcript levels DcASMT were also gradually reduced under different Ca. L. asiaticus infection rates without significant differences between the high infection rates (50,58, and 70%) ( Figure 14F).

Melatonin Supplementation Induced the Expression Levels of Melatonin Biosynthesis-Related Genes of D. citri
We investigated the transcript levels of the melatonin biosynthesis-related genes in healthy and Ca. L. asiaticus-infected D. citri adults after treatment with exogenous melatonin ( Figure 15). Generally, the pretreatment with 500 µg mL −1 melatonin upregulated the gene expression of all studied genes in both healthy and infected insects. Although the expression levels of melatonin biosynthesis-related genes were lower in Ca. L. asiaticus-infected psyllids compared with the control (healthy adults), exogenous melatonin supplementation enhanced the transcript levels of all studied genes to reach the control without significant differences between them (treated-control versus treated-infected psyllids). Accordingly, these findings suggest that melatonin supplementation enhanced the gene expression of melatonin biosynthetic genes of D. citri. . Gene expressions were normalized using two housekeeping genes (α-tubulin and actin), and the changes were analyzed using the 2 −ΔΔC T method. Open dots indicate the raw data (n = 10), the black ones indicate potential outliers, whereas horizontal thick lines indicate the medians. Whiskers indicate the minimum, and the maximum values of the data and boxes show the interquartile ranges (twenty-fifth to the seventy-fifth percentile of the data). Treatments (healthy versus infected and untreated versus treated) were compared using a two-tailed t-test, and statistical significance was established as p < 0.05. The full list of expressed genes, names, accession numbers, and primers are available in Table 1.

Discussion
Previously, we reported that Ca. L. asiaticus infection significantly reduced the e dogenous melatonin content of D. citri [15]. However, it is not clear whether this ma happen due to the utilization of insect melatonin directly by Ca. L. asiaticus, or if it was common cause due to the Ca. L. asiaticus infection, which might affect the physiologic and transcriptional capacities of D. citri. The genome sequencing of Ca. L. asiaticus r vealed that it cannot synthesize the amino acid L-tryptophan, the precursor of melaton [72], from metabolic intermediates [73], and it should acquire it from its host. Furthe there is no evidence for melatonin biosynthesis by Ca. L. asiaticus which supports the id that Ca. L. asiaticus depends on its host (psyllid vectors or citrus plant) for its melaton needs. Nevertheless, recently we showed that melatonin might play an antibacterial ro  (DcT5H-1 and DcT5H-2, respectively), (C) relative gene expression of aromatic amino acid decarboxylase (DcAADC), (D,E) relative gene expression of arylalkylamine N-acetyltransferase genes (DcAANAT-1 and DcAANAT-2, respectively), and (F) relative gene expression of N-acetylserotonin O-methyltransferase (DcASMT). Gene expressions were normalized using two housekeeping genes (α-tubulin and actin), and the changes were analyzed using the 2 −∆∆C T method. Open dots indicate the raw data (n = 10), the black ones indicate potential outliers, whereas horizontal thick lines indicate the medians. Whiskers indicate the minimum, and the maximum values of the data and boxes show the interquartile ranges (twenty-fifth to the seventy-fifth percentile of the data). Treatments (healthy versus infected and untreated versus treated) were compared using a two-tailed t-test, and statistical significance was established as p < 0.05. The full list of expressed genes, names, accession numbers, and primers are available in Table 1.

Discussion
Previously, we reported that Ca. L. asiaticus infection significantly reduced the endogenous melatonin content of D. citri [15]. However, it is not clear whether this may happen due to the utilization of insect melatonin directly by Ca. L. asiaticus, or if it was a common cause due to the Ca. L. asiaticus infection, which might affect the physiological and transcriptional capacities of D. citri. The genome sequencing of Ca. L. asiaticus revealed that it cannot synthesize the amino acid L -tryptophan, the precursor of melatonin [72], from metabolic intermediates [73], and it should acquire it from its host. Further, there is no evidence for melatonin biosynthesis by Ca. L. asiaticus which supports the idea that Ca. L. asiaticus depends on its host (psyllid vectors or citrus plant) for its melatonin needs.
Nevertheless, recently we showed that melatonin might play an antibacterial role against Ca. L. asiaticus [18], which suggests that utilization of melatonin by Ca. L. asiaticus is not the main reason for melatonin reduction within infected psyllid.
Furthermore, another potential reason for the reduction of melatonin is that it could be a common cause due to the Ca. L. asiaticus infection, which disrupts the physiological and transcriptional capacities of D. citri [13][14][15], particularly melatonin biosynthesis-related genes. However, to the best of our knowledge, melatonin biosynthetic genes are not wellannotated yet from the Asian citrus psyllid, D. citri. Previously, we roughly identified four genes to be associated with the melatonin biosynthesis in D. citri, which were downregulated in Ca. L. asiaticus-infected psyllids [15]. However, some of these genes have been updated based on the most recent available data in GenBank, the national center for biotechnology information website (NCBI, http://www.ncbi.nlm.nih.gov/gene/, 12 February 2021) including PREDICTED: Diaphorina citri protein henna-like (LOC103524631; GenBank Accession no. XM_017449691.2), PREDICTED: Diaphorina citri tyrosine 3-monooxygenase (LOC103505706; GenBank Accession no. XM_017442547.2), and PREDICTED: Diaphorina citri aromatic L-amino acid decarboxylase-like (LOC103510318; GenBank Accession no. XM_017444526.2), and, as a result of standard genome annotation processing (see www.ncbi.nlm.nih.gov/genome/annotation_euk/process/, 12 February 2021 for more information), even some sequences were removed including PREDICTED: Diaphorina citri dopamine N-acetyltransferase-like (LOC103507708; GenBank Accession no. XM_008472208.2) and PREDICTED: Diaphorina citri arf-GAP domain and FG repeatcontaining protein 1 (LOC103510708; GenBank Accession no. XM_017444656.1).
Therefore, herein, we carried out a comprehensive in silico and bioinformatics analysis to deeply identify the melatonin biosynthesis-related genes of D. citri using two major databases including the D. citri-specific dataset of "Diaphorina citri OGS v2.0 proteins" available on Citrus Greening Solutions website (https://citrusgreening.org/organism/ Diaphorina_citri/genome, 12 February 2021) [58] and the most popular database of Gen-Bank, the national center for biotechnology information website (NCBI, http://www.ncbi. nlm.nih.gov/gene/, 12 February 2021). A proposed melatonin biosynthesis pathway and its associated genes in D. citri is presented in Figure 16.
Briefly, we suggest that melatonin in D. citri is synthesized from the L -tryptophan via four enzymatic steps. The first step is the oxidation of L -tryptophan to 5-hydroxytryptophan [74] via the activity of DcT5H. Our findings showed that the D. citri genome possesses at least two sequences that were relatively homologous and were phylogenetically closer to DmT5H from D. melanogaster [48,49] including putative tryptophan 5-hydroxylase 1-like (DcT5H-1; DcitrP076520.1.1) and protein henna (DcT5H-2; DcitrP012845.1.1). T5H is the rate-limiting enzyme in serotonin biosynthesis [28,29] that works in the presence of molecular oxygen and requires tetrahydrobiopterin as a cofactor.
Our InterPro-based analysis showed that both DcT5Hs have a tetrahydrobiopterindependent aromatic amino acid (ArAA hydroxylase) family [75] that catalyzes ring hydroxylation of aromatic amino acids, using tetrahydrobiopterin (BH4) as a substrate. Previous studies showed that all eukaryotic T5H are homotetramers and include a regulatory N-terminal domain, a catalytic domain, and a C-terminal oligomerization motif [76,77]. In agreement with these studies, the crystallographic 3D structures of both DcT5Hs modeled with the crystal structure of human tryptophan hydroxylase 2 (TPH2), catalytic domain (4v06.1.A) [78], encompassed two close conserved histidines that are involved in the binding to iron [79] and one more distant acidic residue, usually glutamic acid in our predicted models. This arrangement of ligands suggests that DcT5Hs are metalloproteins. Interestingly, the expression levels of both DcT5Hs were decreased in Ca. L. asiaticus-infected psyllids and showed an almost identical profile with endogenous melatonin content from our previous study [15]. Collectively, these findings suggest that DcT5Hs (DcT5H-1 and DcT5H-2) could play a key role in melatonin biosynthesis. However, further studies are required to clarify the functional and/or regulatory roles of DcT5Hs in D. citri. databases including the D. citri-specific dataset of "Diaphorina citri OGS v2.0 proteins available on Citrus Greening Solutions website (https://citrusgreening.org/organism/D aphorina_citri/genome, 12 February 2021) [58] and the most popular database of Gen Bank, the national center for biotechnology information website (NCB http://www.ncbi.nlm.nih.gov/gene/, 12 February 2021). A proposed melatonin biosynthe sis pathway and its associated genes in D. citri is presented in Figure 16. The second step in the melatonin biosynthesis pathway is the decarboxylation of 5-hydroxytryptophan to the indoleamine serotonin using the pyridoxal 5-phosphate (PLP)-dependent enzyme aromatic L-amino acid decarboxylase (AADC) [28,40]. Our findings showed that the D. citri genome could encode for a putative Dopa decarboxylase (DcDDC; DcitrP031955.1.1) that produced significant similarities to DmDDC (also known as DmAADC) from D. melanogaster [50][51][52]. This protein was described in the current study as DcAADC, indicating its aromatic L-amino acid decarboxylase activity. Based on the InterPro analysis, this enzyme is a pyridoxal phosphate-dependent decarboxylase that belongs to the group II decarboxylases [80,81] and shares a region of a conserved lysine residue, which provides the binding site for the PLP group [80,82]. Further, InterPro analysis showed that DcAADC seems to share regions of sequence similarity with aromatic-L-amino acid decarboxylase (DDC; also known as L-dopa decarboxylase or tryptophan decarboxylase), glutamate decarboxylase (GAD), histidine decarboxylase (HDC), and tyrosine decarboxylase (TyrDC) [80][81][82][83] since it had a DDC/GAD/HDC/TyrDC pyridoxal-phosphate attachment site. These enzymes are collectively known as group II decarboxylases [81].
Moreover, our findings suggest that DcAADC might act as an amino acid-transporter and could be essential for primary carbon metabolism. Although the homo-oligomerization state of AADC proteins has not been extensively investigated, our findings suggest that DcAADC might form a homodimer in the membrane; nevertheless, previous studies suggest that this is dispensable for transport [84]. Our findings showed that Ca. L. asiaticus infection diminished the transcript levels of DcAAAD of D. citri compared to uninfected psyllids. The reduction in the DcAAAD expression was consistent with the profile of endogenous melatonin upon different Ca. L. asiaticus infection rates as reported in our previous study [15]. Moreover, melatonin supplementation enhanced the DcAADC expression and reverses the negative effects of Ca. L. asiaticus. Taken together, these findings suggest that DcAADC might play a dual role in melatonin biosynthesis and amino acid transportation. However, further studies are required to investigate the functional and/or regulatory roles of DcAADC in D. citri.
The third step in the melatonin biosynthesis pathway is the N-acetylation of serotonin to form N-acetylserotonin using serotonin N-acetyltransferase (SNAT; also known as arylalkylamine N-acetyltransferase [AANAT]) [29]. Our findings showed that the D. citri genome possesses two putative dopamine N-acetyltransferase proteins (DcitrP025630.1.1 and DcitrP085745.1.1) with significant similarity to arylalkylamine N-acetyltransferase 1, isoform A (DmAANAT1) from D. melanogaster [53][54][55]. Both proteins were described in the current study as arylalkylamine N-acetyltransferase (DcAANAT-1 and DcAANAT-2) indicating their N-acetylation activity. These findings are in agreement with studies on D. melanogaster where two AANAT variants (AANATA and AANATB) were identified [59,85]. Although both variants were physiologically relevant in D. melanogaster, they were differentially expressed with respect to tissue distribution and developmental stages [59,85]. In contrast, the expression patterns of DcAANAT-1 and DcAANAT-2 in the current study were almost identical with slight differences which suggest that both enzymes serve the same metabolic role in D. citri.
AANAT is an acetyl-CoA-dependent enzyme that belongs to GCN5 N-acetyltransferases (GNATs) family [86]. It is the penultimate enzyme in melatonin biosynthesis that catalyzes the transfer of the acetyl group of acetyl-CoA to the primary amine of serotonin to form N-acetylserotonin and CoA [86]. In the current study, InterPro analysis showed that both DcAANAT-1 and DcAANAT-2 have a structural domain of Acyl-CoA N-acyltransferase homologous superfamily. This domain has a triple-layer α/β/α structure that contains mixed β-sheets. The crystallographic 3D structures showed that DcAANAT-1 and DcAANAT-2 have three antiparallel β-strands like DmAANAT. Although coenzyme A binding pockets were observed via the InterPro analysis, no acyl-CoA binding sites were observed in the tertiary structures of DcAANAT-1, and DcAANAT-2. Our findings in the current study showed that low infection rates of Ca. L. asiaticus (less than 50%) did not affect the expression levels of DcAANAT-1 and DcAANAT-2; however, our previous study showed that low infection rates significantly reduced the endogenous melatonin content [15]. On the other hand, the higher infection rates reduced the expression levels of both genes in agreement with the endogenous melatonin profile of our previous study [15]. Together, these findings suggest that DcAANAT may not be a rate-determining enzyme in melatonin biosynthesis in D. citri. However, further studies are required to investigate the functional and/or regulatory roles of DcAANAT in D. citri.
Finally, N-acetylserotonin is subsequently methylated by N-acetylserotonin O-methyltransferase (ASMT) as a final step in the melatonin production [28,[30][31][32]. Our findings showed that the D. citri genome encodes for a putative N-acetylserotonin O-methyltransferase-like (DcASMT; DcitrP032285.1.1) that is producing significant similarities to N-acetylserotonin O-methyltransferase-like protein, isoform X1 (ClASMT) from bed bug (C. lectularius). Herein, we used ClASMT from bed bug as a query sequence because, to the best of our knowledge, no ASMT genes have been identified from D. melanogaster yet. Like other melatonin biosynthesis-related genes in this study, DcASMT was significantly downregulated in Ca. L. asiaticus-infected psyllids and agreed with the melatonin profile in our previous study [15].
Moreover, we proposed an alternative route for melatonin biosynthesis from serotonin ( Figure 16). Briefly, we suggest that serotonin could be methylated first, rather than N-acetylated as discussed above, to form 5-methoxytryptamine using ASMT. Subsequently, 5-methoxytryptamine is N-acetylated to form melatonin using SNAT. However, a previous study showed that the enzymatic activity of ASMT was about 14-times greater when it reacted with N-acetylserotonin than when it reacted with serotonin [87] which suggests methylation, but not N-acetylation, of N-acetylserotonin, as the last step of melatonin. Nevertheless, this alternative route might occur under specific circumstances [31,33]. As summarized above, melatonin biosynthesis is controlled by four successive enzymes; however, SNAT was proposed to be the key rate-limiting enzyme in this pathway in vertebrates [28], but not at night [88]. Instead, ASMT may be a rate-limiting enzyme during the nocturnal production of melatonin [88].

Conclusions
In conclusion, in the present study, we computationally identified six melatonin biosynthesis-related gene candidates (two DcT5Hs, one DcAADC, two DcAANATs, and one DcASMT) in D. citri. These genes were definitionally expressed within the adults of D. citri after challenging with gradient infection rates of the phytopathogenic bacterium Ca. L. asiaticus. Moreover, the expression patterns of these genes demonstrated a piece of indirect evidence for the return of melatonin to its normal levels in Ca. L. asiaticusinfected psyllids after melatonin supplementation and confirmed the association of these genes with the melatonin biosynthesis pathway. However, further investigations are required to explore the functional and/or regulatory roles of these genes in melatonin biosynthesis. Our findings of this study are a further step for optimization and cloning of melatonin biosynthesis genes of D. citri. They could rapidly be identified via in silico analysis and subsequently subjected to in vitro and in vivo confirmatory studies, since our previous study showed that inhibition of melatonin biosynthesis was associated with reduced longevity of D. citri [15]. Therefore, the identified genes in this study could be good candidates that serve as potential targets for RNA interference (RNAi)-based control or other sustainable control strategies of D. citri.
Author Contributions: N.K., together with Y.N., conceptualized the study and contributed to the experimental design. Y.N. and N.K. contributed with analytic tools, data analysis, and finalized the figures. Finally, N.K. and Y.N. wrote and finalized the manuscript. All authors have read and agreed to the published version of the manuscript. Data Availability Statement: Data will be shared upon request to the corresponding author.