Endosymbionts of citrus leafminer Phyllocnistis citrella Stainton among different citrus orchards in China

Endosymbionts regulate the behavior of pest species, which could provide insights into their control. The citrus leafminer (Phyllocnistis citrella Stainton) is a widely distributed pest associated with diseases of citrus, especially of young trees. Here, we determined the endosymbiont composition of P. citrella in citrus orchards across China. The resulting dataset comprised average 50,430 high-quality reads for bacterial 16S rRNA V3-V4 regions of endosymbionts from 36 P. citrella larvae sampled from 12 citrus orchards across China. The sequencing depth and sampling size of this dataset were sufficient to reveal most of the endosymbionts of P. citrella. In total, 2,875 bacterial amplicon sequence variants were obtained; taxonomic analysis revealed a total of 372 bacterial genera, most of which were Proteobacteria phylum with Undibacterium being the most abundant genus. This dataset provides the first evidence of P. citrella endosymbionts that could support the development of pest management approaches in citrus orchards.


Background & Summary
The citrus leafminer (Phyllocnistis citrella Stainton) is a moth that poses a threat to citrus and related ornamental plants 1 .Adult P. citrella are small, silvery moths with a wingspan of 4 mm 2 .Although P. citrella is considered to have originated in India and southern Asia 3 , it has now spread to most citrus-growing regions in the world, including China 4 .The damage caused in P.citrella can stunt the growth of young trees, increasing their susceptibility to other stressors; by contrast, mature trees are better able to tolerate such damage 5 .Larvae of P. citrella create serpentine mines on the leaves, which reduce the photosynthetic capacity of, and cause deformities in, young leaves 6 .The endosymbionts of insects have been reported to play a role in various aspects of insect biology, including pesticide degradation and interactions with plant pathogens 7 .In citrus pests, such as the P. citrella, understanding the endosymbionts may have implications for pest control efforts and the management of citrus diseases 8 .Research has focused on the endosymbionts of citrus psyllids (Diaphorina citri Kuwayama), which are the important carriers of Candidatus Liberibacter asiaticus, the causal agent of huanglongbing in citrus [9][10][11] .However, there is a lack of information about the endosymbionts specific to P. citrella Here, we used high-throughput sequencing based on the bacterial 16S rRNA gene to characterize the endosymbionts of 36 larval P. citrella from 12 citrus orchards across China (Fig. 1).Basic information of sequencing and taxonomic annotation is presented in Table 1.In total, 2,208,938 raw reads were obtained ranging from 43,113 to 111,529 per sample.After quality control, the average number of clean reads among all samples were 50,430, which was clustered into 2,875 bacterial amplicon sequence variants (ASVs).Taxonomic annotation revealed a total of 22 phyla, 52 classes, 135 orders, 206 families, and 372 genera.Most of these ASVs belonged to three bacterial phyla: Proteobacteria, Firmicutes, and Bacteroidota (Fig. 2).At the genera level, Acinetobacter, Pseudomonas, Sphingomonas, and Staphylococcus were most represented by ASVs (Fig. 2).Only a few ASVs were detected in each larvae, suggesting that there are significant differences in endosymbiont composition among P. citrella individuals.In terms of the endosymbionts, Proteobacteria was the most dominant bacterial phylum, accounting for 93.36% of the total bacterial community (Fig. 3a).Undibacterium was the most abundant bacterial genus in the gut endosymbiont of P. citrella (43.79%), followed by Achromobacter (20.31%), and the relative abundance of all other genera was <5% (Fig. 3b).
In summary, our sequencing data offer the first snapshot of the diversity and composition of endosymbionts from P. citrella.This information provides insights into the co-evolution of endosymbionts and their insect hosts, and further supports the development of potential approaches for management of P. citrella in citrus orchards.

Methods
Insect sampling and DNA isolation.At each citrus orchard in this study (Fig. 1), several citrus leaves infested with P. citrella were randomly collected.The P. citrella larvae were removed from leaves, added to a plastic tube, and transported to the laboratory.The larvae from each citrus orchard were then washed several times with sterile normal saline, air dried in a bioclean room, randomly divided into three groups (dozens individuals per group), and stored at −80 °C for further analyses.In total, 36 samples were obtained from 12 citrus orchards (three per orchard).The total genomic DNA of each sample was extracted using a QIAamp Power Fecal DNA Kit (QIAGEN, Dusseldorf, Germany) according to the manufacturer's instructions.Agarose gel electrophoresis (1.5% concentration) was applied to evaluate whether DNA was extracted successfully.Then, the purity and concentrations of the extracted DNA were measured using a NanoPhotometer Classic (IMPLEN, Munich, Germany).All acceptable DNA samples of acceptable purity and concentration levels were stored at −20 °C until further use.
High-throughput sequencing.The primers 341 F (CCTACGGGNGGCWGCAG) and 806 R (GGACTACHVGGGTATCTAAT) with adapter sequence and barcodes combined at the end of the reverse primer were applied to amplify the V3-V4 regions of the bacterial 16S rRNA gene from each sample of DNA 12 .A 20-µL mixture (0.4 μL of FastPfu Polymerase, 4 μL of 5 × FastPfu Buffer, 0.8 μL of each primer (5 μM), 2 μL of dNTPs (2.5 mM), and 11.8 μL ddH 2 O) was used to perform the PCR reaction with the template DNA (10 ng).Thermal cycling comprised initial denaturation at 95 °C for 2 min, followed by 30 cycles of denaturation at 95 °C for 5 s, annealing at 55 °C for 30 s, elongation at 72 °C for 30 s, and a final extension at 72 °C for 6 min.All PCR products were extracted and purified using an AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA) with 2% agarose gels.Then, Qubit ® 3.0 (Life Invitrogen) was used to quantify the concentra- tions of each purified PCR product, which were then mixed in equal volumes to generate the amplicon libraries using a TruSeq Nano DNA LT Library Prep Kit (Illumina, USA).Library quality was assessed using the Agilent Bioanalyzer 2100 system (Agilent, USA) and QuantiFluor dsDNA system (Promega, USA).Finally, an Illumina Novaseq 6000 platform was applied to sequence these libraries with a 250 bp paired-end strategy.

Data processing and taxonomic annotation.
Based on the unique barcode combined in the primer, sequencing reads were appointed to each sample, and each barcode sequence was then truncated by a self-written script.The quality of raw reads was controlled according to the follow standards: average Phred scores >20, no ambiguous bases, no mismatches in the primers, homopolymer runs <8, and read length >250 bp 13 .Quality control, paired reads assembly, chimeras elimination, and ASV clustering were performed using the DADA2 plugin unit in the QIIME2 program 14 with the default parameters.Based on these results, all ASVs were assigned to a taxonomy using the SILVA database (Release 138) 15 and singletons (an ASV represented by only one count) were abandoned.Finally, the sequencing dataset was rarefied using a standard number of reads according to the sample with the lowest read number (34,051), and the then ASV abundance table was transformed to the relative abundance (%) for further analyses.
Visualization and statistics analysis.All visualization and statistics analysis were performed using R v4.2.2 16 .Species rarefaction and cumulative curves of sequenced data sets were extrapolated by the "iNEXT" package 17 .A phylogenetic tree of bacterial ASVs was visualized by the "ggtree" package 18 and all other plots were drawn using the "ggplot2" package 19 .

Data Records
The raw sequencing data (fastq format) were deposited in NCBI's Sequence Read Archive (PRJNA1072277) 20 .

Technical Validation
Low-quality reads, singletons, and chimera were removed from the sequencing data; the remaining clean reads accounted for >80% of the data (Table 1).To ensure unbiased data production, randomization principles were used for sample collection, DNA extraction, and sequencing processes.Using the software and database described above, we confirmed the technical validation of the ASV clustering, taxonomic assignments, and abundance account.Species rarefaction and cumulative curves were established (Fig. 4), and a straight horizontal strait line was observed, suggesting sufficient sequencing depth and sample size of our dataset.

Fig. 1
Fig. 1 Map of sample collection sites (The bottom left photo shows the leaves used for collecting Phyllocnistis citrella Stainton larvae, and the black lines on the leaves represent the Phyllocnistis citrella Stainton larvae).

Fig. 2
Fig. 2 Phylogenetic tree of bacterial ASVs detected from the endosymbionts of P. citrella.Only the top 500 most abundant ASVs are shown.The color of the branches represents the ASV taxonomy at the phylum level.Color for the blocks at the middle layer represents the ASV taxonomy at the genus level.The barplot on the outer ring represents the detected ratio of ASVs in the collected samples.

Fig. 3 Fig. 4
Fig. 3 Average relative abundance (%) of dominant bacterial phyla (a) and genera (b) among the endosymbionts of all collected samples of P. citrella.

Table 1 .
Statistics of high-throughput sequencing and taxonomic annotation.