(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



A major endogenous glycoside hydrolase mediating quercetin uptake in Bombyx mori [1]

['Ryusei Waizumi', 'Institute Of Agrobiological Sciences', 'National Agriculture', 'Food Research Organization', 'Naro', 'Tsukuba', 'Ibaraki', 'Chikara Hirayama', 'Shuichiro Tomita', 'Tetsuya Iizuka']

Date: 2024-02

Quercetin is a common plant flavonoid which is involved in herbivore–plant interactions. Mulberry silkworms (domestic silkworm, Bombyx mori, and wild silkworm, Bombyx mandarina) take up quercetin from mulberry leaves and accumulate the metabolites in the cocoon, thereby improving its protective properties. Here we identified a glycoside hydrolase, named glycoside hydrolase family 1 group G 5 (GH1G5), which is expressed in the midgut and is involved in quercetin metabolism in the domestic silkworm. Our results suggest that this enzyme mediates quercetin uptake by deglycosylating the three primary quercetin glycosides present in mulberry leaf: rutin, quercetin-3-O-malonylglucoside, and quercetin-3-O-glucoside. Despite being located in an unstable genomic region that has undergone frequent structural changes in the evolution of Lepidoptera, GH1G5 has retained its hydrolytic activity, suggesting quercetin uptake has adaptive significance for mulberry silkworms. GH1G5 is also important in breeding: defective mutations which result in discoloration of the cocoon and increased silk yield are homozygously conserved in 27 of the 32 Japanese white-cocoon domestic silkworm strains and 12 of the 30 Chinese ones we investigated.

Quercetin is one of the most abundant flavonoids present in plants. This flavonoid is involved in herbivorous insect–plant interactions. Insects utilize it for host–plant recognition, coloration, and protection from ultraviolet and oxidative stress. However, the molecular mechanism of quercetin metabolism in insects remains unclear. Mulberry silkworms (domestic silkworm, Bombyx mori, and wild silkworm, Bombyx mandarina) take up quercetin from mulberry leaves and sequester it into their cocoon to improve its protective properties. In this study, we identified an endogenous glycoside hydrolase in the domestic silkworm, named glycoside hydrolase family 1 group G 5 (GH1G5). This enzyme mediates quercetin uptake into the midgut cells by deglycosylating mulberry leaf-derived quercetin glycosides. This is the first discovery of a rutin glycoside hydrolase in an animal. Furthermore, we found that defective mutations of GH1G5 have been broadly disseminated within the domestic silkworm population due to the improved cocoon color (i.e., discoloration to white) and increased silk yield. This study illuminates the unique mechanism of quercetin uptake in the domestic silkworm and uncovers an important event in the history of silkworm breeding.

Here, we performed a quantitative trait locus (QTL) analysis focused on cocoon flavonoid content in the domestic silkworm and identified a novel locus, Green d (Gd), which is associated with this trait. Within the locus, we identified a glycoside hydrolase gene, glycoside hydrolase family 1 group G 5 (GH1G5), which mediates quercetin uptake into the midgut cells by deglycosylating mulberry leaf-derived quercetin glycosides. Genetic dissection of the novel gene revealed the contribution of the gene to improvement of the cocoon in a commercial context through breeding.

The color of the cocoon of the domestic silkworm has been diversified through breeding, which indicates that the kinds and amounts of flavonoids that accumulate in the cocoons differ between strains [ 11 , 12 ]. Particularly, cocoons containing high flavonoid concentrations express a yellow-green color and are known as “green cocoons”. Because accumulation in the cocoon is the end step of flavonoid metabolism, forward-genetic analysis focused on cocoon flavonoid content can reveal the genes involved in individual steps of flavonoid metabolism. Indeed, several loci associated with flavonoid metabolism have already been identified through this approach: the Green b locus, which encodes a uridine 5´-diphospho-glucosyltransferase (UGT) with a rare enzymatic activity glycosylating the 5-O position of quercetin [ 14 ], and the New Green Cocoon (Gn) locus, which encodes clustered glucose transporter (GLUT)-like sugar transporters presumed to import quercetin glucosides from the hemolymph to the silk gland [ 18 ]. However, these findings explain only a part of the process from quercetin uptake to the final accumulation of its metabolites in the cocoon. Elucidating the steps of flavonoid metabolism in the silkworm can provide valuable insights into the understanding of herbivore–plant interactions.

Mulberry silkworms (domestic silkworm, Bombyx mori, and wild silkworm, Bombyx mandarina) accumulate various compounds derived from plants, including quercetin glucosides, kaempferol glucosides, and carotenoids, in their silk glands and colored cocoons ( S1 Fig ) [ 11 , 12 , 13 ]. Quercetin glucosides stored in the body and in the cocoon are reported to have antioxidant, ultraviolet-protective, and antibacterial properties [ 14 , 15 , 16 ]. Quercetin is found in the leaves of the mulberry tree (Morus alba), the sole food source of mulberry silkworms, as a series of glycosides formed by glycosylation at the 3-O position. The three most common quercetin glycosides in mulberry are quercetin-3-O-rutinoside (rutin), quercetin-3-O-malonylglucoside (Q3MG) and Q3G, which account for 71%–80% of the total flavonol content in the leaves [ 17 ].

Quercetin (3,3´,4´,5,7-pentahydroxyflavone) is a flavonoid abundantly found in a wide variety of plants [ 1 ]. Previous studies have found quercetin glycosides possess oviposition and feeding stimulant activity in lepidopteran, orthopteran and coleopteran insects [ 2 , 3 , 4 , 5 , 6 ]. These observations suggest that quercetin is widely ingested by insects. Quercetin ingestion by insects is likely not merely a consequence of the identification and feeding on host plants; it may have adaptive significance. In the common blue butterfly (Polyommatus icarus), which sequesters quercetin-3-O-galactoside in its wings, flavonoid content is higher in the female than in the male and positively correlated with female sex attraction [ 7 , 8 , 9 ]. The yellow pigment in the wings of a grasshopper (Dissosteira carolina), which may contribute to their camouflage in plants, results from sequestration of quercetin-3-O-glucoside (isoquercitrin, Q3G) [ 10 ]. Although these studies strongly emphasize the broad significance of quercetin in herbivore–plant interactions, the underlying molecular mechanisms of quercetin metabolism in insects are poorly understood.

Results

QTL analysis identified a novel locus associated with flavonoid content in cocoons To identify genes involved in quercetin metabolism by means of a forward-genetics approach, we prepared a green-cocoon strain (p50; alias: Daizo) and a white-cocoon strain (J01; alias: Nichi01). The two strains exhibited a distinct difference in cocoon color and flavonoid content (Fig 1A and 1B). A gradated range of cocoon colors in their F 2 intercross offspring implied that the genetic differences between the two strains associated with flavonoid content were composite (Fig 1C). Therefore, we conducted a QTL analysis, which allows for simultaneous identification of multiple genetic loci involved in a phenotype of interest. The flavonoid content of the cocoons of the F 2 population were scored by an absorbance-based method according to a previous report [19]. The QTL analysis was performed by using the phenotypic data of 102 individuals and 1038 genetic markers obtained from double-digest restriction-associated DNA sequencing data [20] (S1 Table). From the composite interval mapping, we identified three significant QTLs for cocoon flavonoid content on chromosomes 15, 20, and 27 (Fig 1C). Previous linkage studies suggested a locus, named Green c (Gc), associated with the yellow-green color of cocoons, is located at an unknown position on chromosome 15 [21,22]. The QTL on chromosome 27 was presumed to correspond to Gn [18]. Since no green cocoon-associated locus has yet been reported on chromosome 20, we named that locus Green d (Gd). The contribution of the Gd locus to cocoon flavonoid content was the second largest of the three QTLs, with a percentage of phenotypic variation explained by each QTL (PVE) value of 24.57% (Fig 1C and S2 Table). The 95% Bayes credible interval of the Gd locus was 7,980,189–10,504,065 bp. The nearest marker to Gd was located at 10,265,033 bp, with a logarithm of odds (LOD) score of 19.99. The PVE values of the QTLs on chromosomes 15 and 27 were 7.04% and 56.05%. Significant additive effects were detected between the QTLs on chromosomes 15 and 27 and between those on chromosomes 20 and 27, with corresponding PVEs of 1.93% and 5.47% (S2 Table). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Quantitative trait loci (QTLs) associated with cocoon flavonoid content. (A) Photographs of representative cocoons of the p50 and J01 silkworm strains. Bar = 30 mm. (B) Flavonoid content of p50 and J01 cocoons. Data are means of n independent biological replicates ± SD. (C) QTL analysis for cocoon flavonoid content. The horizontal dotted line indicates the threshold of the permutation test (trials = 1000). Phenotype scores on the frequency distribution are relative to a maximum measurement of 10. LOD, logarithm of odds; PVE, percentage of phenotypic variation explained by each QTL. According to a previous study [21], the Gc locus is located on chromosome 15, but its detailed genetic or physical position is unknown. Therefore, it cannot be concluded that the QTL peak on chromosome 15 identified here corresponds to the Gc locus. (D) Gene models present within the Gd locus of the p50T genome assembly. The genes annotated as encoding glycoside hydrolase family 1 proteins are highlighted in black; their IDs are KWMTBOMO12222, KWMTBOMO12223, KWMTBOMO12224, KWMTBOMO12225, KWMTBOMO12227, KWMTBOMO12229, KWMTBOMO12230, KWMTBOMO12233, and KWMTBOMO12236 from the upstream side. (E) Expression of the candidate Gd genes in the midgut of third-day final instar male larvae. Data are means of n independent biological replicates ± SD. TPM, transcripts per million. https://doi.org/10.1371/journal.pgen.1011118.g001

Gd locus contains a glycoside hydrolase family 1 gene cluster Using the genome assembly and gene models of p50T [23], we found a cluster of nine genes encoding glycoside hydrolases (KWMTBOMO12222–25, KWMTBOMO12227, KWMTBOMO12229, KWMTBOMO12230, KWMTBOMO12233, and KWMTBOMO12236) on the Gd locus, which were annotated as glycoside hydrolase family 1 (GH1), according to the nomenclature of carbohydrate-active enzymes provided by CAZy [24] (Fig 1D). In mammals, a critical step in quercetin metabolism is deglycosylation of quercetin glucosides by lactase/phlorizin hydrolase (LPH), a member of GH1. LPH is expressed in intestinal epithelial cells and is anchored on the brush border membrane where it hydrolyzes flavonoid glycosides [25,26,27,28]. The resulting free quercetin aglycon is then passively absorbed into the intestinal cells due to its increased lipophilicity. We hypothesized that the GH1 genes clustered within the Gd locus are involved in quercetin metabolism in the midgut lumen of the domestic silkworm, playing a role similar to that of LPH in mammals. In an inferred phylogenetic tree of all 21 GH1 proteins in B. mori and their homologous proteins in representative Holometabola insects (the fruit fly, Drosophila melanogaster; the honeybee, Apis mellifera; the red flour beetle, Tribolium castaneum), the glycoside hydrolases encoded in the Gd locus formed a distinct clade (group G), which was supported by a high bootstrap value (100%), suggesting that divergence of the group G-glycoside hydrolases had occurred after insect order divergence (S2 Fig). According to the phylogeny and their genomic positions, we named them as glycoside hydrolase family 1 group G 1–9 (GH1G1–9). The sequence identity among the group G glycoside hydrolase proteins was highest between GH1G2 and GH1G4 at 79.13%, and the sequence similarity was highest between GH1G1 and GH1G9 at 95.56% (S3A and S3B Fig). In addition, signal peptides were predicted in the N-terminal region of GH1G1, GH1G2, GH1G3, GH1G5, GH1G7 and GH1G9, suggesting they are secreted enzymes (S3A Fig). To identify which of them act in the midgut, we performed RNA-seq-based expression analysis. Three of the candidate genes, GH1G1, GH1G5, and GH1G9, were found to be strongly expressed in the midgut of p50T final instar larvae; the expression levels of these genes were significantly lower in the midgut of strain J01 (Fig 1E). Further examination of the expression profiles of the three genes with high expression in the midgut revealed that GH1G5 and GH1G9 were expressed specifically in the midgut, whereas GH1G1 was expressed mostly in Malpighian tubules (S4 Fig). Taken together, we identified GH1G1, GH1G5, and GH1G9 as candidate Gd genes which are involved in quercetin metabolism in the domestic silkworm.

Functional analysis of the candidate Gd genes by CRISPR-Cas9 To investigate the involvement of these candidate genes in quercetin metabolism, we attempted to use a microinjection-mediated CRISPR-Cas9 system to establish p50T lineages in which they had been knocked out. Although we failed to establish knockout lineages for GH1G1 and GH1G9 due to the lethality of homozygous frame-shift mutations, we did manage to obtain two knockout lineages of GH1G5 with different types of frameshift mutations in exon 5. We designated the one with a 5-bp deletion as ΔGd1 and the other with a 2-bp deletion as ΔGd2 (Fig 2A). The GH1G5 mutations resulted in a premature stop codon at exon 5 along with shortened amino acid sequence lengths from 492 to 215 in ΔGd1, and from 492 to 216 in ΔGd2. The mutants produced discolored cocoons compared to p50T (Fig 2B). In addition, we observed a reduction of fluorescence under ultraviolet irradiation in the midgut, hemolymph, and silk glands of the mutants (Fig 2C–2E). Such fluorescence is characteristic of the accumulation of the two major quercetin metabolites in silkworm, quercetin-5-O-glucoside and quercetin-5,4´-di-O-glucoside [14,15,29]. Although knockout of GH1G5 reduced the total flavonoid content in the cocoon to less than half that in the original p50T strain, it was still much larger than the effect predicted for the Gd locus in the QTL analysis (Fig 2F and S2 Table). We found similar reductions in the midgut, hemolymph, and middle and posterior silk gland. Because the flavonoid content in the cocoon differed largely between the insects reared with an artificial diet and those reared with fresh mulberry leaves (Figs 1B and 2F), we confirmed the flavonoid content reductions in the mutants in an experiment using fresh mulberry leaves (S5 Fig). Together, these results indicated that knockout of GH1G5 resulted in malfunction of quercetin uptake into the midgut. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Reduction of flavonoid content in GH1G5-knockout mutants. (A) Disrupted target sequences in GH1G5. (B)–(E) Cocoons (B), midguts (C), hemolymph (D), and silk glands (E) of the p50T strain and the GH1G5 mutants. Organs and tissues were taken from sixth-day final instar female larvae. Bars = 10 mm. (F) Total flavonoid content of cocoons, organs, and tissues of the p50T strain and the KWMTBOMO12227 mutant lineages. Data are means of n independent biological replicates ± SD. All insects were reared on a commercial artificial diet. SG, silk gland. https://doi.org/10.1371/journal.pgen.1011118.g002

Sequence and isoform determination for GH1G5 The predicted gene model for GH1G5, KWMTBOMO12227, consists of 11 exons, encoding a total of 492 amino acids. The accuracy of the predicted sequence and its dominance among isoforms were confirmed using Sanger sequencing and the RNA-seq data obtained from final instar larvae of strain p50T. We cloned the putative longest open reading frame in the transcripts of GH1G5, and sequenced it by Sanger sequencing (S6A Fig). The determined sequence was completely identical to the predicted sequence of GH1G5. Previously, our research group reported RNA-seq data for final instar larva tissues of the p50T strain [30]. By mapping the reads of the data derived from the midgut to the genomic sequence of the p50T strain reported by Kawamoto et al. [23], we determined the transcript isoforms of GH1G5. The gene model accurately represented the dominant isoform; another isoform extending 15-bp downstream of the 10th exon was also detected (S6B Fig). The termination codon in the extended region shortened the predicted protein sequence length of this isoform to 456 amino acids. Since the mean coverage of each base of the 15-bp extended region was only about 15% of that of the original 10th exon of GH1G5 (S6C Fig), we concluded this to be a minor isoform.

GH1G5 mediates quercetin uptake by hydrolysis of quercetin glycosides Together, our knockout analysis and sequence characterization suggested that GH1G5 is secreted into the midgut lumen and mediates the uptake of mulberry-derived quercetin by deglycosylation of quercetin glycosides. However, the knockout mutants still accumulated some flavonoids in the midgut cells (Fig 2F). This might be due to the diversity of quercetin glycosides in mulberry leaves and the substrate specificity of GH1G5. To confirm whether GH1G5 is involved in deglycosylation of quercetin glycosides, we investigated the hydrolytic activity of midgut tissue of the knockout mutants on the three major quercetin glycosides in mulberry leaf: rutin, Q3MG, and Q3G (Fig 3A). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Hydrolytic activity of GH1G5 on the three major quercetin glycosides in mulberry leaf. (A) Structural formulae of quercetin and the three major quercetin glycosides present in mulberry leaf. (B) Hydrolytic activity of the midgut on the quercetin glycosides. Data are means of n independent biological replicates ± SD. The values above the graph are p-values calculated by a two-tailed Student’s t-test. All samples were from fifth-day final instar female larvae reared on a commercial artificial diet. (C) Photograph of representative cocoons of the p50T strain and the mutants reared on semi-synthetic diets containing quercetin or rutin. Bar = 20 mm. (D) Total flavonoid content in cocoons of the p50T strain and the mutants reared on semi-synthetic diets containing quercetin or rutin. Data are means of n independent biological replicates ± SD. https://doi.org/10.1371/journal.pgen.1011118.g003 The homogenate of the p50T midgut exhibited hydrolytic activity against all three types of quercetin glycoside (Fig 3B). However, the activities were decreased in the knockout mutants, indicating that GH1G5 is involved in the hydrolysis of all three quercetin glycosides. Notably, GH1G5 knockout completely abolished the hydrolytic activity against rutin. Furthermore, partial reductions were observed in the hydrolytic activity of the other two glycosides, Q3MG and Q3G, suggesting that the silkworm has other glycoside hydrolases with hydrolytic activities against those molecules (Fig 3B). Knocking out GH1G5 reduced the hydrolytic activity of midgut homogenates against Q3MG and Q3G by 0.6 to 0.8-fold and 0.7 to 0.8-fold, respectively. These results suggested that GH1G5 is an important protein for the hydrolysis of quercetin glycosides in the silkworm lumen. In 1972, Fujimoto and Hayashiya reported that the domestic silkworm accumulates flavonoids in its cocoon when reared on artificial diets containing isolated quercetin or rutin [31]. To investigate whether the uptake of rutin-derived quercetin is dependent on deglycosylation by GH1G5, we reared insects on semi-synthetic diets supplemented with rutin or quercetin but without mulberry leaf powder and measured the flavonoid content in the cocoons. The p50T strain accumulated flavonoids in its cocoon irrespective of diet (Fig 3C and 3D and S3 Table). Although the GH1G5-knockout mutants accumulated the same amount of flavonoids as did p50T when reared on the quercetin diet, they accumulated only 6% of that accumulated by p50T when reared on the rutin diet (Fig 3C and 3D). These results were consistent with the report by Fujimoto and Hayashiya, and suggested that the uptake of quercetin from rutin is strongly dependent on deglycosylation by GH1G5.

Evolution of GH1G5 The phylogeny of the Holometabola GH1 proteins suggested that the divergence of the group G glycoside hydrolases had occurred after insect order divergence (S2 Fig). To estimate when GH1G5 arose during the evolution of Lepidoptera, we constructed a maximum likelihood-inferred phylogenetic tree of lepidopteran-wide orthologous proteins of the group G glycoside hydrolases in B. mori (BmorGH1G1–9) (Fig 4A). The tree classified BmorGH1G1–9 into five clades: one clade including BmorGH1G5, one clade including BmorGH1G1 and BmorGH1G9, one clade including BmorGH1G2, BmorGH1G3 and BmorGH1G4, one clade including BmorGH1G6 and BmorGH1G7, and one clade including BmorGH1G8. The clade including BmorGH1G5 consisted of only proteins from species belonging to Macroheterocera (including Noctuoidea, Lasiocampoidea, and Bombycoidea), suggesting that GH1G5 had evolved through gene duplication after the divergence of Pyraloidea and Macroheterocera (Fig 4A and 4B). Sequence identities and similarities of the GH1G5-class proteins, excluding the one from B. mandarina, to BmorGH1G5 were at a maximum of 57.92% and 89.98% (Fig 4C). OrthoFinder [32], the tool we used for the collection of orthologous proteins of BmorGH1G1–9, detected gene duplication events that had occurred within each orthologous group while simultaneously estimating orthologous relationships. Interestingly, it suggested that the group G glycoside hydrolases had undergone notably frequent duplication events; the estimated number of duplications which occurred within the ortholog group including GH1G5 (also including GH1G2, GH1G4, and GH1G6–9) was 46, which ranked 197th out of 19436 total groups and 1st out of 54 groups including B. mori proteins annotated as glycoside hydrolases (InterPro entry: IPR001360 or GO annotation: 0016798) (S4 and S5 Tables). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Phylogenetic tree of lepidopteran-wide orthologous proteins of the group G glycoside hydrolases in B. mori. (A) We constructed the phylogenetic tree by the maximum likelihood method. Values on the nodes represent the bootstrap scores (trials = 100). Unreliable nodes with bootstrap values under 50 are shown as multi-branching nodes. The tree was rooted using Rat LPH as an outgroup. Red branches represent clades including the group G glycoside hydrolases in B. mori. The clade containing BmorGH1G5 is highlighted with a box. (B) Species tree of Lepidoptera drawn/constructed with reference to the report by Kawahara et al. [33]. Red circles indicate species harboring a GH1G5 ortholog. (C) Sequence identities and similarities of the GH1G5-class proteins to BmorGH1G5. Similarity of amino acid residue property is determined according to the groups of strongly similar properties described at Clustal Omega FAQ (https://www.ebi.ac.uk/seqdb/confluence/display/THD/Help+-+Clustal+Omega+FAQ). https://doi.org/10.1371/journal.pgen.1011118.g004

[END]
---
[1] Url: https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1011118

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.