Utilizing a chromosomal-length genome assembly to annotate the Wnt signaling pathway in the Asian citrus psyllid, Diaphorina citri

The Asian citrus psyllid, Diaphorina citri, is an insect vector that transmits Candidatus Liberibacter asiaticus, the causal agent of the Huanglongbing (HLB), or citrus greening disease. This disease has devastated Florida’s citrus industry, and threatens California’s industry as well as other citrus producing regions around the world. To find novel solutions to the disease, a better understanding of the vector is needed. The D. citri genome has been used to identify and characterize genes involved in Wnt signaling pathways. Wnt signaling is utilized for many important biological processes in metazoans, such as patterning and tissue generation. Curation based on RNA sequencing data and sequence homology confirms 24 Wnt signaling genes within the D. citri genome, including homologs for beta-catenin, Frizzled receptors, and seven Wnt-ligands. Through phylogenetic analysis, we classify D. citri Wnt ligands as Wg/Wnt1, Wnt5, Wnt6, Wnt7, Wnt10, Wnt11, and WntA. The D. citri version 3.0 genome with chromosomal length scaffolds reveals a conserved Wnt1-Wnt6-Wnt10 gene cluster with a gene configuration like that in Drosophila melanogaster. These findings provide greater insight into the evolutionary history of D. citri and Wnt signaling in this important hemipteran vector. Manual annotation was essential for identifying high quality gene models. These gene models can be used to develop molecular systems, such as CRISPR and RNAi, which target and control psyllid populations to manage the spread of HLB. Manual annotation of Wnt signaling pathways was done as part of a collaborative community annotation project.

biology, the D. citri genome has been manually annotated to curate accurate gene model predictions. Accurate gene models can be used to develop novel insect control systems that utilize molecular therapeutics such as CRISPR (clustered regularly interspaced short palindromic repeats) and RNA interference (RNAi) to control the spread of D. citri [3,4]. These molecular therapeutics would be gene-specific, thus would reduce reliance on broad-spectrum insecticides that have given rise to resistant D. citri populations [5][6][7].

Context
Here, we report D. citri genes involved in both canonical and noncanonical Wnt signaling.
Wnt signaling is important for many biological processes in metazoans, such as patterning, cell polarity, tissue generation, and stem cell maintenance [8][9][10]]. In the model insects Drosophila melanogaster and Tribolium castaneum, knockout and knockdown of Wnt ligands and other Wnt signaling components have detrimental effects on embryo development and adult homeostasis [11][12][13][14][15][16]. Wnt signaling components could therefore be effective knockout targets to limit the spread of D. citri, thus reducing HLB incidence. We curated a comprehensive repertoire of Wnt signaling genes in D. citri. Twenty-four gene models corresponding to canonical and noncanonical Wnt signaling genes have been annotated, including seven Wnt ligands, three frizzled homologs, arrow, armadillo/beta-catenin, and receptor tyrosine kinases ROR and doughnut. We were unable to find Wnt8/D, Wnt9, and Wnt16 as well as Wnt2-4, which have been lost in insects. The mechanisms of Wnt signaling appear to be mostly conserved and comparable to those found in D. melanogaster (Table 1). A model for canonical Wnt signaling in D. citri based on curated genes is shown (Figure 1). This is an important first step towards understanding critical biological processes that might be targeted to control the spread of D. citri, and may provide broader insights into the mechanisms of Wnt signaling in this important hemipteran vector.

METHODS
We used the psyllid genome curation workflow used for community annotation ( Figure 2) [17].
High-scoring MCOT models (accessions available in Table 2) were then searched on the NCBI protein database using NCBI BLAST to confirm the viability of the predicted MCOT models. The high-scoring MCOT models that had promising NCBI search results were used to search the D. citri genome. Genome regions containing computationally predicted gene models with high sequence identity to the query sequence from the MCOT transcriptome were investigated within JBrowse (RRID:SCR_001004). Gene models were modified using the Apollo (RRID:SCR_001936) gene annotation platform, based on mapped DNA-Seq, RNA-Seq, Iso-Seq, orthologous proteins, and other lines of evidence to edit and confirm manual annotations and gene structure. The gene models were analyzed with NCBI BLAST to assess  (RRID:SCR_000667) [23]. Neighbor-joining trees were constructed using MEGA7 with p-distance for determining branch length and 1000 bootstrapping replications to measure the precision of branch placement. In special cases, phylogenetic analysis in conjunction with NCBI BLAST scores was used to properly name and characterize the manually annotated gene models.

Data validation and quality control
The loss of Wnt ligand genes is more common in insects than in other metazoans [20], which leads to a highly variable array of Wnt genes and Wnt signaling components from species to species [15,21,22,42]. We performed a phylogenetic analysis to characterize the D. citri Wnt repertoire ( Figure 3). The ortholog sequences used for this analysis were collected from the NCBI protein database [18]; see the 'Availability of Data and Materials' section for accession numbers. Seven D. citri Wnts were identified and classified as Wnt1 (also known as wingless), Wnt5, Wnt6, Wnt7, Wnt10, Wnt11, and WntA (Figures 3 and 4). In comparison, seven Wnt genes have been identified in D. melanogaster, nine in T. castaneum, and six in Acyrthosiphon pisum [22,42]. The collection of Wnt genes found in D. citri is like that found in other insects, and no Wnt subfamilies have been identified as being unique to D. citri. Contrary to previous reports [43], D. citri does appear to possess a Wnt6 gene.
Wnt1, Wnt6, and Wnt10 typically occur close together in a highly conserved gene cluster [44,45]. The chromosomal length genome assembly in v3.0 suggests that this cluster is also Ortholog sequences were collected from NCBI protein database [18]. Analysis was performed using MEGA7 [23]. divergence of cnidarians and bilaterians [45]. The orientation of these clustered D. citri Wnt genes is like that found in D. melanogaster and differs from what may be a basal arthropodal organization of Wnts found in species of Coleoptera, Hymenoptera, and Cladocera ( Figure 5). Wnt9 is also associated with this gene cluster when present in the genome. However, as with A. pisum, Wnt9 was not found in the D. citri genome and appears to have been lost during evolution. A second Wnt cluster, Wnt5 and Wnt7, is also common among non-insect metazoans. This cluster is not seen in D. citri; however, D. citri Wnt5 and Orthologs for Wnt2, Wnt3, Wnt4, Wnt8/D, Wnt9, and Wnt16 were not located in the D.
citri genome. The close identity of certain Wnt subfamilies makes it difficult to distinguish between them; however, the loss of Wnt2-4 is expected because they are absent in all insects [20]. Apis mellifera and the hemipteran A. pisum have been reported to lack Wnt8/D.
Perhaps this Wnt subfamily has been lost in the divergence from other insect groups [22]. and 272 TPM, respectively. This is considerably higher than all other Wnt genes in these tissues, which only average between 0. 26 (Figure 8). The hemipteran clade suggests that these genes might belong to a different subfamily of Frizzled, maybe one specific to Hemiptera. However, this ortholog has not been reported in the A. pisum genome [22].
Orthologs for both ROR1 and ROR2 have been identified. Interestingly, ROR1 has two isoforms, the first of which contains an immunoglobulin (IG) domain that is lacking from isoform 2 (Figure 9). ROR1 isoform 2 (Dcitr05g14430. 1.2) appears to average higher transcript levels in D. citri egg, nymph, and adult tissues than ROR1 isoform 1 (Dcitr05g14430.1.1) based on CGEN data ( Figure 10). Many transcripts for isoform 2 were detected in the psyllid egg ( Figure 10). This suggests that expression of isoform 2 may be important in the early developmental stages of D. citri.

Conclusion
Controlling the spread of D. citri is an important strategy for reducing the spread of HLB.
With this study, we hope to provide a greater insight into D. citri biology, as well as accurate gene models that can be used in future research and applications. We have curated a comprehensive repertoire of Wnt signaling genes in D. citri. In total, 24 gene models  There are 24 gene models in total. Each gene model has been assigned an identifier, and the evidence used to validate or modify the structure of the gene model has been listed. MCOT transcriptome identifiers that best support the manual annotation are also listed. The table is marked with an 'X' when supporting evidence of de novo transcriptome, Iso-Seq, RNA-Seq and ortholog support is present. MCOT: comprehensive transcriptome based on genome MAKER, Cufflinks, Oases, and Trinity transcript predictions; MAKER: gene predictions; De novo transcriptome: an independent transcriptome using Iso-Seq long-reads and RNA-Seq data; Iso-Seq transcripts: full-length transcripts generated with Pacific Biosciences technology; RNA-Seq: reads mapped to genome are also used as supporting evidence for splice junctions; Ortholog evidence: proteins from related hemipteran species and Drosophila melanogaster.
Dcitr01g03830. corresponding to canonical and noncanonical Wnt signaling have been annotated. The mechanisms of Wnt signaling appear to be mostly conserved and comparable to those found in D. melanogaster and other insects. These findings provide a greater insight into the evolutionary history of D. citri and Wnt signaling in this important hemipteran vector. Manual annotation and an improved genome assembly with chromosomal length scaffold were essential for identifying high quality gene models.

REUSE POTENTIAL
The manually curated genes will be included in the Citrus Greening Expression Network (CGEN) [48] as a part of the Official Gene Set version 3. This visualization tool is useful for understanding psyllid biology and comparative analysis because it contains public transcriptomics data for Diaphorina citri from various tissues, life stages, CLas infection levels and citrus hosts. Future work could utilize these gene models in developing CRISPR and RNAi systems that target and disrupt critical biological processes in D. citri, thus controlling the spread of HLB. This work was done as part of a collaborative community annotation project [49]. Ortholog sequences were collected from the NCBI protein database [18]. Some NCBI sequences (such as XP_006568530.1, XP_008188372.2, and XP_022194032.1) may have numeric labels derived from computational predictions that do not reflect sequence or functional similarity. Analysis was performed using MEGA7 [23].

DATA AVAILABILITY
Our annotation and gene curation workflow is described by Shippy et al. [17]. The Diaphorina citri genome assembly, official gene sets, and transcriptome data are accessible on the Citrus Greening website [50]. All accessions for genes used for phylogentic analysis are provided within this report (Tables 2, 3, Figure 8). We have included the Newick and Multiple Sequence Alignment files used to construct the Wnt neighbor-joining phylogenetic tree and other data is available in the GigaScience GigaDB repository [51].

EDITOR'S NOTE
This article is one of a series of Data Releases crediting the outputs of a student-focused and community-driven manual annotation project, curating gene models and -if requiredcorrecting assembly anomalies, for the Diaphorina citri genome project [46].