8-HQA adjusts the number and diversity of bacteria in the gut microbiome of Spodoptera littoralis

Abstract

Quinolinic carboxylic acids are known for their metal ion chelating properties in insects, plants and bacteria. The larval stages of the lepidopteran pest, Spodoptera littoralis, produce 8-hydroxyquinoline-2-carboxylic acid (8-HQA) in high concentrations from tryptophan in the diet. At the same time, the larval midgut is known to harbor a bacterial population. The motivation behind the work was to investigate whether 8-HQA is controlling the bacterial community in the gut by regulating the concentration of metal ions. Knocking out the gene for kynurenine 3-monooxygenase (KMO) in the insect using CRISPR/Cas9 eliminated production of 8-HQA and significantly increased bacterial numbers and diversity in the larval midgut. Adding 8-HQA to the diet of knockout larvae caused a dose-dependent reduction of bacterial numbers with minimal effects on diversity. Enterococcus mundtii dominates the community in all treatments, probably due to its highly efficient iron uptake system and production of the colicin, mundticin. Thus host factors and bacterial properties interact to determine patterns of diversity and abundance in the insect midgut.

Keywords: 8-hydroxyquinoline-2-carboxylic acid; Enterococcus mundtii; Spodoptera littoralis; kynurenine 3-monooxygenase; microbiome.

Figure 1

Initial steps in the biosynthesis of 8-HQA. Tryptophan is converted to formylkynurenine by tryptophan dioxygenase, which is then converted to kyurenine by kynurenine formamidase. Kynurenine in turn is converted to 3-hydroxykynurenine by kynurenine monooxygenase (KMO). The toxic 3-hydroxykynurenine is converted to xanthurenic acid by transamination (here shown via alanine glyoxalate aminotransferase, represented by two enzymes AGT1 and AGT2 encoded by separate genes in Spodoptera littoralis) followed by cyclization. Labelling experiments have shown that 3-hydroxykynurenine is a precursor of 8-HQA (Pesek et al., 2015) therefore enzymes involved in its synthesis and conversion were targeted by CRISPR/Cas9.

Figure 2

Mutated target sequences in AGT2 and KMO genes. (A) 3 bp insert (above) and 2 bp insert (below) occurred in the AGT2 gene compared to the WT allele. The target sequence is highlighted in green in the WT. (B) Examples of KMO-ko alleles compared to the WT allele. The two target sequences are highlighted in green in the WT.

Figure 3

Detection and quantification of 8-HQA. (A) Derivatization of 8-HQA with PFBB and mass spectrum of the derivatized compound. (B) Concentration of 8-HQA in μg ml⁻¹ of regurgitant of the larvae in 5th instar wildtype (WT) and KMO gene knockout larvae. On the average, a WT larva has 142 μg ml⁻¹ of 8-HQA in the regurgitant (Wilcoxon test, p < 0.05).

Figure 4

Bacterial abundance and composition in wild-type and KMO knockout larvae. (A) 16S rRNA gene copy numbers signifying the bacterial abundance as determined by quantitative PCR during different stages of the life cycle of S. littoralis (2nd, 3rd, 4th and 5th instars) and the genotypes of KMO-ko and WT. The KMO-ko harbor a higher bacterial number than their corresponding WTs. (B) Comparison of relative microbial composition at the phylum level in KMO-ko vs. WT S. littoralis through the larval instars. (C) Box plot for comparison of alpha diversity of species in KMO-ko and WT guts of S. littoralis, through the larval instars, based on Shannon (species richness) index of alpha diversity (Wilcoxon test p < 0.05 for 3rd and 4th instars). (D) Beta diversity analysis between the KMO-ko and WT lines of S. littoralis guts by Canberra distances on PCoA plot (Adonis test p = 0.001). Each point represents an individual sample and clustering of points indicates similarity of bacterial composition in the individuals.

Figure 5

Results of feeding 8-HQA to KMO knockout larvae. (A) Normalized 16S copy numbers of the same groups of larvae, a measure of the bacterial abundance in each condition: 5th instar larvae-KMO-ko fed with 0, 1, 3, 5, 10 mm 8-HQA, and WT. The bacterial numbers of KMO-ko approach that of WT upon administration of the compound (ANOVA, p < 0.05, Tukey’s test between 0 mm 8-HQA and WT < 0.05). (B) Bacterial composition of KMO-ko larvae fed with the same increasing concentration of 8-HQA, KMO-ko control, and WT. The taxa distribution of KMO-ko do not resemble that of WTs, and Bacilli is the most abundant class in KMO-ko even after administration of 8-HQA.