Cabbage looper (Trichoplusia ni Hübner) labial glands contain unique bacterial flora in contrast with their alimentary canal, mandibular glands, and Malpighian tubules

Abstract In recent years, several studies have examined the gut microbiome of lepidopteran larvae and how factors such as host plant affect it, and in turn, how gut bacteria affect host plant responses to herbivory. In addition, other studies have detailed how secretions of the labial (salivary) glands can alter host plant defense responses. We examined the gut microbiome of the cabbage looper (Trichoplusia ni) feeding on collards (Brassica oleracea) and separately analyzed the microbiomes of various organs that open directly into the alimentary canal, including the labial glands, mandibular glands, and the Malpighian tubules. In this study, the gut microbiome of T. ni was found to be generally consistent with those of other lepidopteran larvae in prior studies. The greatest diversity of bacteria appeared in the Firmicutes, Actinobacteria, Proteobacteria, and Bacteriodetes. Well‐represented genera included Staphylococcus, Streptococcus, Corynebacterium, Pseudomonas, Diaphorobacter, Methylobacterium, Flavobacterium, and Cloacibacterium. Across all organs, two amplicon sequence variants (ASVs) associated with the genera Diaphorobacter and Cloacibacterium appeared to be most abundant. In terms of the most prevalent ASVs, the alimentary canal, Malpighian tubules, and mandibular glands appeared to have similar complements of bacteria, with relatively few significant differences evident. However, aside from the Diaphorobacter and Cloacibacterium ASVs common to all the organs, the labial glands appeared to possess a distinctive complement of bacteria which was absent or poorly represented in the other organs. Among these were representatives of the Pseudomonas, Flavobacterium, Caulobacterium, Anaerococcus, and Methylobacterium. These results suggest that the labial glands present bacteria with different selective pressures than those occurring in the mandibular gland, Malpighian tubules and the alimentary canal. Given the documented effects that labial gland secretions and the gut microbiome can exert on host plant defenses, the effects exerted by the bacteria inhabiting the labial glands themselves deserve further study.


| INTRODUC TI ON
Cabbage looper (Trichoplusia ni) is a voracious insect pest, which can eat ~160 different plant hosts. It prefers cruciferous species such as Brassica oleracea, hence the name cabbage looper. The plant response to insect infestation depends on the mode of infestation.
Piercing/sucking insects with haustellate mouthparts do not create much damage upon infestation but generally induce the salicylic acid pathway and the defense genes involved in slowing plant pathogens.
Chewing insects with mandibulate mouthparts typically cause the induction of the jasmonic acid pathway, which produces a cascade of defense genes that can deter insects via toxicity or slowing down digestive processes (Stahl, Hilfiker, & Reymond, 2018). However, plant responses to an insect pest are driven by more than just the mechanical damage caused by insect mouthparts. Plant responses to herbivory are altered by exposure to components of oral secretions from the labial (salivary) and mandibular glands, regurgitant from the alimentary canal, and frass. Additionally, microbes associated with the insect have been shown to alter how plants perceive herbivory.
Generally, both chemical components within the oral secretions, as well as specific microbes isolated from the insect pest, can change the plant response (Stahl et al., 2018).
In recent years, the mechanisms that plants use to defend themselves against lepidopteran larvae, and the mechanisms the larvae employ to defeat those defenses, have been the subject of numerous investigations. Increasingly, these studies suggest that the insect-plant interaction can be substantially influenced or mediated by bacteria inhabiting the alimentary canal of the insect. In some instances, the resident bacteria favor the insect, while in other cases the plant benefits. Studies indicate that bacterial populations within two insect species inhabiting the same plant, or a single insect species on different plants can be quite variable, and appear to be highly influenced by the bacteria inhabiting the host plant (Jones, Mason, Felton, & Hoover, 2019).
One of the more straightforward ways gut bacteria can influence the interaction between an insect and its host plant is to augment the digestive processes of the insect to overcome factors that impair digestion. Multiple lines of evidence suggest that proteases produced by the gut bacteria of the velvetbean caterpillar allow the larvae to overcome the protease inhibitors produced by soybean (Visôtto, Oliveira, Guedes, Ribon, & Good-God, 2009;Visôtto, Oliveira, Ribon, Mares-Guia, & Guedes, 2009). In other insect-plant combinations, such as fall armyworm feeding on maize, it appears that gut bacteria may augment certain plant defenses to the detriment of the insect . Acevedo et al. (2017) demonstrated that a number of bacteria of the Enterobacteriaceae isolated from fall armyworm modulated jasmonic acid-mediated plant defenses in a plant dependent way. Pantoea ananatis and an isolate termed Enterobacteriaceae-1 downregulated polyphenol oxidase and a trypsin protease inhibitor in tomato, but upregulated a maize proteinase inhibitor in maize.
In addition to the advances in understanding the role of the gut microbiome in plant-caterpillar interactions, the role of the larval labial gland has also been examined using proteomic and transcriptomic methods (de la Paz Celorio-Mancera et al., 2011;Rivera-Vega, Acevedo, & Felton, 2017;Rivera-Vega, Galbraith, Grozinger, & Felton, 2017;Rivera-Vega, Stanley, Stanley, & Felton, 2018) Labial gland secretions from the cotton bollworm, Helicoverpa armigera, included a variety of enzymes with apparent digestive functions, such as proteases, lipases, and amylases. A variety of antimicrobial peptides and lysozymes were also detected, as was glucose oxidase (GOX), which decreases wound-inducible nicotine production in tobacco (Musser et al., 2005;de la Paz Celorio-Mancera et al., 2011). Transcriptome and proteome profiles of Trichoplusia ni labial glands were significantly different when the larvae were fed either tomato, considered a more challenging host, or cabbage. A variety of proteins involved in digestion and response to host defenses were upregulated when larvae were fed tomato relative to cabbage. In particular, the levels of catalase, which inhibits foliar peroxidase by reducing levels of H 2 O 2 , were found to be increased in the labial glands of larvae fed tomato (Rivera-Vega, Galbraith, et al., 2017;Rivera-Vega et al., 2018).
The labial glands, as well as the mandibular glands, also open directly into the oral cavity of the alimentary canal (Eaton, 1988). The alimentary canal of lepidopteran larvae is a tube of epithelium which runs the entire length of the larvae, from the oral cavity to the anus.
It can be subdivided into three major functional regions, the foregut, midgut, and the hindgut. The foregut is a relatively narrow, muscular tube that conveys chewed leaf material from the oral cavity through the head and thorax to the midgut, which begins near the junction of the thorax and abdomen, occupying most of the latter. The midgut In the current study, we examined the alimentary canal microbiome of T. ni, including accessory structures that communicate directly with the gut lumen: the labial glands, mandibular glands, and the Malpighian tubules. We were particularly interested in determining whether the insect labial glands contained populations of bacteria, and if so, whether these populations were similar to those inhabiting the alimentary canal. To initiate the study K E Y W O R D S cabbage looper, gut, labial glands, microbiome, plant-herbivore interactions, Trichoplusia ni of interactions between T. ni and Brassica oleracea, the bacteria associated with T. ni labial glands, mandibular labial glands, the alimentary canal, and the Malpighian tubules were characterized using next-generation sequencing of 16S ribosomal RNA gene amplicons.

| Plant/insect preparation
Two flats of Champion collard seeds (Johnny's Selected Seeds) were sown in sunshine mix #1 (Griffin) and grown for 3 weeks. One flat of collards were moved to a greenhouse where the temperature mimicked the outdoor temperatures.
Four to six squares of wax paper with approximately 100 eggs each of recently oviposited T. ni eggs (Benzon Research) were pinned to the leaves throughout the flat that had been placed inside a mesh cage. The second flat of collards was added to the cage when larvae consumed ¾ of the first flat. Larvae were monitored until they reached the 5th instar and were collected for dissection.

| Larval dissection
Larvae were sedated by placing at −20°C for 1-2 min to facilitate immobilization before dissection of labial glands, mandibular glands, Malpighian tubules, and midgut tissue ( Figure 1). Five to ten organs (or sets of organs) were pooled for each biological replicate and frozen. Midgut tissue was rinsed in PBS (phosphate-buffered saline) before freezing. Tissue was stored at −80°C until genomic DNA extraction.

| Bacterial genomic DNA extraction and purification
Both GeneJET Genomic DNA purification (Thermo Scientific) and Gentra Puregene (Qiagen) kit reagents were used to isolate microbial genomic DNA to ensure liberation of all bacterial cells from insect organ tissue and efficient lysis of both gram-negative and gram-positive bacterial cells. Insect tissue samples stored at −80°C in microcentrifuge tubes were thawed on ice and ground with sterile mini pestles until a homogeneous mixture was achieved.

| GeneJET protocol
Immediately following tissue disruption, 9 μl GeneJET digestion solution per mg of tissue was added and samples were incubated at 56°C for 3 hr on a thermomixer (150 rpm every 10 min) until all particulates disappeared. The solution was centrifuged for 5 min at 16,000 × g to pellet any unlysed gram-positive or gram-negative bacterial cells (pellets were placed on ice for gram-positive bacterial genomic DNA isolation using Gentra Puregene). The supernatant was transferred to new microcentrifuge tubes and further processed for gram-negative bacterial DNA purification with GeneJET.
To continue with gram-negative bacterial DNA purification, 20 μl of GeneJET RNase A solution was added to the supernatant, mixed by inversion and incubated for 10 min at room temperature. 200 μl of GeneJET lysis solution was added and mixed by inversion until a homogenous mixture was obtained. 400 μl of 50% ethanol was added and mixed by inversion followed by GeneJET column purification as per manufacturer instructions.

| Gentra Puregene protocol
The pellet from the initial centrifugation step in the GeneJET protocol was processed for gram-positive bacterial cells as follows.
300 μl of Gentra Puregene cell suspension solution was added to the pellet and heated to 95°C for 10 min, then cooled to 37°C.
3 μl Gentra Puregene lytic enzyme solution was added, mixed by inversion 25 times, incubated for 30 min at 37°C, and finally centrifuged 1 min 16,000 × g to pellet cells. The supernatant was discarded, 300 μl Gentra Puregene cell lysis solution was added, and the remaining pellet was gently resuspended by flicking the tube.
Resuspended pellet solutions were incubated at 80°C for 5 min to complete gram-positive cell lysis. 1.5 μl of Gentra Puregene RNase A solution was added, and tubes were mixed by inversion 25 times then incubated for 1 hr at 37°C followed immediate cooling on ice for 1 min. 100 μl Gentra Puregene protein precipitation solution was added, vortexed for 20 s then centrifuged 16,000 × g for 3-8 min until a tight protein pellet formed. The supernatant was transferred to a new 1.5 ml microcentrifuge tube containing 300 μl isopropanol and inverted gently 50 times. The DNA was pelleted by centrifugation at 16,000 × g for 1 min, washed with 70% ethanol and air-dried for 5 min at room temperature. 100 μl Gentra Puregene DNA Hydration Solution was added, and the DNA was incubated at 65°C for 1 hr to aid in dissolving the DNA.
Purified genomic DNA samples from both GeneJET and Gentra workflows for each replicate were combined prior to DNA quality assessment.

| DNA quality assessment and concentration measurements
Before library preparation, DNA quality was assessed. First, concentrations of the DNA samples were determined with a plate fluorometer. To begin library preparation, 30 µl of 25 ng/µl DNA was necessary to build Illumina compatible 16S libraries (Illumina, 2013). The concentrations of the samples that met these criteria, and a random selection of samples were analyzed using a Fragment Analyzer to ensure reliable quality control prior to next-generation sequencing (NGS).

| 16S Library preparation and assessment
Samples were shipped to the Georgia Genomics and Bioinformatics Core for library preparation and sequencing. The library preparation process began with normalizing DNA samples to 5 ng/µl.
A 5 µl of each sample was used to proceed with the first PCR.

| RE SULTS AND D ISCUSS I ON
Consequently, 9,000 reads were used for the sampling depth. The The normalized frequencies of the ten most abundant ASVs for each organ, and their levels in the other organs, are shown in Table 2.
Among the organs examined, the labial glands appeared to have the most distinctive microbiome, possessing several abundant ASVs that were either absent or of low numbers in the other organs, and conversely, lacking ASVs that appeared in the other organs (Table 2). Among those, ASVs that were predominant in the labial glands and not elsewhere are three apparent pseudomonads and a Flavobacterium (Figure 3).   (2012) isolated Pseudomonas nitroreducens from the gut of the scarab, Holotrichia parallela, and found it to be cellulolytic, suggesting that the ASV found in the current study could contribute to early digestive processes in T. ni.
The mandibular and labial glands both open into the oral cavity of the larva, which would seemingly be exposed to the same inoculating population of bacteria. Yet their microbiomes appear to be quite different. This suggests that the conditions within the two organs exert different selective pressures on bacteria attempting to colonize them. The functions of the two glands, and their internal chemistries, have been shown to be significantly different.
In addition to silk production, the labial glands of lepidopteran insects have been shown to produce many enzymes and peptides with both digestive and defensive roles (Rivera-Vega, Galbraith, et al., 2017). While the mandibular glands are less well studied, several species have been shown to produce 2-acyl-1,3-cyclohexanediones, which can serve as larval trail marking pheromones or cuticular hydrocarbons (Fitzgerald, Kelly, Potter, Carpenter, & Rossi, 2015;Howard & Baker, 2004). In addition, many chemosensory and odorant-binding proteins have been found in mandibular gland secretions, suggesting that the gland is involved in chemical communication (de la Paz Celorio-Mancera et al., 2012). Whether this is due to a difference in pH, differential production of antimicrobial factors by the two glands, or altogether different elements, remains to be elucidated.

ACK N OWLED G M ENTS
No funding is declared.

CO N FLI C T O F I NTE R E S T
None declared.

E TH I C S S TATEM ENT
None required.