The first Drosophila melanogaster UGT gene to be identified, Dorothy (currently Ugt36A1), was named after a character of The Wizard of Oz (Rodriguez et al., 1996). A little later, five other D. melanogaster UGT genes, Ugt35a, Ugt35b, Ugt37a1, Ugt37b1, and Ugt37c1 (lowercase letters were initially used to indicate subfamily membership), were among the first UGT genes to be named in consultation with the UGT Nomenclature Committee (Wang et al., 1999). Subsequently, several other D. melanogaster UGTs were directly named in FlyBase according to their cytogenetic locations (e.g., Ugt36Ba – Ugt36Bc, Ugt58Fa, and Ugt86Da – Ugt86Dj) (Table 2), which is evidently confusing given the superficial resemblance between this notation and the UGT Committee nomenclature. Ahn et al. (2012) revised and curated the D. melanogaster UGTs, employing the systematic names to maintain consistency with the universal nomenclature and the five previously assigned official names. In the current study, we have completed the list of D. melanogaster UGT genes and have updated the gene symbols and names within FlyBase to adopt the systematic nomenclature. Furthermore, we have added a UGT “gene group” page to FlyBase that conveniently lists all these genes in a single report to facilitate further analysis and download of associated data².

Wang et al. (1999) identified 9–10 putative UGT gene sequences, including the five named ones (see above), from cDNA libraries and the incomplete genome databases available at the time. Upon completion of the D. melanogaster genome (Adams et al., 2000), the first genome-wide annotation of multiple UGT genes was conducted and a total of 33 putative UGT genes were reported together with a phylogenetic and genomic analysis (Luque and O’Reilly, 2002). Ahn et al. (2012) revised the sequences in detail and identified an additional gene (Ugt50B3). The current study has added one further gene (Ugt305A1), resulting in a complete repertoire of 35 UGT genes in D. melanogaster (Table 2). They are grouped into 13 families according to the nomenclature system: UGT35 (6 genes), UGT36 (4 genes), UGT37 (8 genes), UGT49 (3 genes), UGT50 (1 gene), UGT301 (1 gene), UGT302 (3 genes), UGT303 (4 genes), and 1 gene in each of UGT304, UGT305, UGT307, UGT316, and UGT317 (Table 2 and Figure 1).

All 35 UGT genes are found on the two major autosomes (chromosome 2 with 16 genes and chromosome 3 with 19 genes); none are located on the minor autosome (chromosome 4) or the sex chromosomes (Table 2 and Supplementary Figure 1). Among different chromosomal arms, about half (17 UGT genes) lie on 3R (the right arm of chromosome 3), followed by 2L (11 UGT genes), 2R (5 genes) and 3L (2 genes). A large cluster of UGT genes is found on 3R at the cytogenetic location of 86D4 – 86D6, where ten closely related UGT genes are positioned in tandem. The other multiplied gene families are found in one or two genomic locations in close proximity, whereas the members of another large family, UGT37, are spread across three different chromosomal arms (five in 2L, one in 2R, and two in 3R) (Table 2 and Supplementary Figure 1). It is noteworthy that 3L harbors only two UGT genes (Ugt305A1 and Ugt316A1), both of which seem to be unique in their sequences, and are unusually long (Table 2).

All 35 UGT genes are interrupted by intron(s) except for Ugt37C1 and Ugt37C2 (Table 2). These two intron-less genes do not seem to originate from bacterial UGT genes due to their sequence similarity to animal UGTs (see Ahn et al., 2014). D. melanogaster UGT genes are composed of one to six exons: a majority of genes (19 genes; 54%) comprise 2 exons and the rest of genes have 1, 3, 4 or 5 exons, except one gene (Ugt50B3) has 6 exons in its coding sequence (Table 2 and Supplementary Figure 2). The lengths of intron sequences are mostly within the range of 48–85 bp (41 introns) or 108–584 bp (14 introns). Exceptionally, Ugt50B3 is interrupted by three long introns (1,389, 1,0432, and 8,198 bp) followed by two short ones (63 and 52 bp) (Supplementary Table 1 and Supplementary Figure 3). This, together with the fact it is phylogenetically distinguished from the others (Figure 1) and highly conserved in insects in general (Ahn et al., 2012), suggests Ugt50B3 is one of the oldest UGT genes.

Splicing variants are found in two UGT genes, Ugt50B3 and Ugt303B1, where two alternative transcripts have been reported (Table 2). The Ugt50B3 variant is annotated to have an alternative start codon in the middle of what is otherwise the third exon, producing a protein that is 89 amino acids (aa) shorter than the normal one. The Ugt303B1 variants seem to be derived from alternative splicing sites at the 3’-end of the first exon, resulting in a difference of only 9 nucleotides (3 aa) (Table 2).

The average length of D. melanogaster UGT proteins is 532 aa with two outliers, Ugt305A1 (583 aa) and Ugt316A1 (636 aa), which, as noted above, are phylogenetically unique and located in different genomic positions from the other UGT genes. All the UGTs contain an N-terminal signal peptide and a C-terminal transmembrane (TM) domain (Table 2 and Supplementary Figure 4), indicating that the D. melanogaster UGTs are located in the endoplasmic reticulum (ER) with their catalytic domains facing the ER lumen, as shown in other animals (Meech et al., 2012). The UGT-defining 44-aa signature sequence in the C-terminal domain, which is predicted to be intimately involved in the binding of UDP-sugar (Meech et al., 2019), is well conserved across the 35 UGTs (Supplementary Figure 5). However, variations shown in some residues in the signature sequence imply different specificity to different sugar donors other than UDP-glucose.

A consensus Maximum-likelihood tree constructed with deduced amino acid sequences revealed lineage-specific gene amplifications in several families such as UGT35, UGT36, UGT37, UGT49, UGT302, and UGT303 (Figure 1). For example, upon divergence from a common ancestor with Ugt307A1, UGT37 seems to have diversified into the largest gene family in D. melanogaster UGTs. It is noteworthy that the UGT37 members are spread across five different genomic locations. On the other hand, other multiplied UGTs are most likely diversified by tandem gene duplications, as they are found in the same genomic scaffolds in close proximity (Supplementary Figure 1).

Tissue-specific expression patterns of D. melanogaster UGT genes were analyzed previously by Ahn et al. (2012) using microarray data present in FlyAtlas (Chintapalli et al., 2007). Here, we have revisited this analysis using the higher quality RNAseq data available from the FlyAtlas2 database (Leader et al., 2018) – full data for adult males, adult females and larvae are included in Supplementary Table 2; representative data for adult males and larvae are in Figure 2. UGTs from each family are expressed in every adult and larval tissue at some level. Some UGT genes belonging to multi-gene families (Ugt35D1 and Ugt37E1) are undetectable in any tissue, while several others are expressed only in restricted patterns. In contrast, many UGT genes appear to be expressed ubiquitously, with high expression levels often seen within the digestive and excretory systems, particularly for members of the UGT35 and UGT37 families. Across all UGTs, the highest expression is seen within the adult midgut and larval Malpighian tubules. Of note, Ugt50B3, the sole representative of the UGT50 family, shows unusually high expression within the male accessory gland and the female spermatheca, whereas Ugt305A1 is only expressed at appreciable levels in the testis. Such restricted expression patterns suggest particularly important roles of Ugt50B3 and Ugt305A1 within these tissues.

Given the documented role of some UGTs in detoxification, we also examined whether D. melanogaster UGT gene expression is induced after exposure to various environmental and chemical treatments by examining RNAseq data generated by the modENCODE project (Brown et al., 2014) – the full dataset is in Supplementary Table 3; representative subsets are in Figure 3. The expression of most UGT genes is not upregulated in response to the majority of treatments. However, six genes from four different UGT families (Ugt35A1, Ugt37A2, Ugt37A3, Ugt37D1, Ugt49B1, and Ugt302C1) clearly show upregulated expression in response to the addition of caffeine, rotenone or ethanol to the diet, or exposure to Sindbis virus. On the other hand, certain treatments, including cold exposure and increased dietary copper or zinc, have no/little effect on the expression of any UGT gene.

We identified UGT genes in 18 additional Drosophila species and deduced their orthologous relationships to the D. melanogaster genes (Figure 4; see section “Materials and Methods”). The total number of UGT genes per genome varies from 29 in D. elegans, D. pseudoobscura, and D. mojavensis, to 50 in D. takahashii. Some UGT families have been preserved, whereas others have been multiplied or lost through evolution (Figure 4 and Supplementary Table 4). The conserved UGT families are mostly single-member families, such as UGT50, UGT301, UGT304, UGT305, UGT307, UGT316, and UGT317, and show little or no gene additions/losses. The other UGT families comprising multiple genes show variable gene additions or losses in the different species (Supplementary Table 4). One of the most fluctuating families is UGT37: there are 8 gene members in D. melanogaster, but the number increases up to double (16 genes) in D. rhopaloa followed by D. willistoni (15 genes), and decreases down to half (4 genes) in D. erecta and D. grimshawi. The UGT49 family also shows a high degree of species difference: there are 3 gene members in D. melanogaster, but the number increases up to 11 in D. bipectinata followed by 8 in D. ananassae.

Two UGTs that are not orthologous with any D. melanogaster UGTs were detected in both D. virilis and D. mojavensis. One pair is an additional member of the UGT50 family, named as the UGT50F subfamily in this study. The other pair defines a new UGT family, named here as Ugt401A. By BLAST search in NCBI, additional UGT50F members were found in three other species not included in this study (D. arizonae, D. navojoa, and D. hydei), whereas orthologs of UGT401A were present in seven other species (D. arizonae, D. navojoa, D. hydei, D. novamexicana, D. albomicans, D. innubila, and D. busckii). As all of these species form a distant group (“repleta-virilis” group) from D. melanogaster, the UGT401A genes might have been lost after divergence of two sub-genera, Sophophora and Drosophila, or newly emerged in this group, probably playing a unique role.

Further comparative analyses amongst Drosophila and related species will become possible as additional genomes are sequenced and annotation pipelines are improved. This will likely reveal other interesting evolutionary patterns. For example, our preliminary analysis of the genome (Gloss et al., 2019) and transcriptome (Whiteman et al., 2012) of Scaptomyza flava, a herbivorous leaf-mining species belonging to the Drosophilidae family (Whiteman et al., 2011), reveals that this species has only 23 UGT genes (data not shown), the smallest number among the species surveyed in this study.

Genomic data for D. melanogaster UGTs were obtained from FlyBase (flybase.org; Thurmond et al., 2019) using release FB2020_05, which includes D. melanogaster genome annotation R6.36. Genomic data for other Drosophila species were obtained from NCBI – sequence assemblies and annotation versions are given in Supplementary Table 4. Supplementary Data File 1 contains all Drosophila UGT protein sequences in fasta format. The signal peptides and transmembrane domains shown in Supplementary Table 1 were predicted by SignalP-5.0 Server³ and TMHMM Server v. 2.0⁴, respectively.

Deduced amino acid sequences of 35 D. melanogaster UGT sequences were aligned by ClustalW and a consensus phylogenetic tree was constructed using the Maximum Likelihood method and JTT matrix-based model with 1,000 bootstrappings. As an outgroup, fringe (CG10580), an N-acetylglucosaminyltransferase, was used. Evolutionary analyses were conducted in MEGA X (Kumar et al., 2018). The species phylogenetic tree of Drosophila used in Figure 4 was adapted from that in (Seetharam and Stuart, 2013).

Tissue expression (RNAseq) data were downloaded from FlyAtlas2 (⁵ Leader et al., 2018). Gene FPKM (Fragments Per Kilobase of transcript per Million mapped reads) and Enrichment (measuring the abundance of a gene in a particular tissue relative to that in the whole fly) data for adult males, adult females and larvae were downloaded as TSV files and processed in Excel (Supplementary Table 2). FPKM data for adult males and larvae are presented in Figure 2.

modENCODE treatment expression (RNAseq) data (Brown et al., 2014) for were obtained from FlyBase (⁶ Thurmond et al., 2019) using the Batch Download tool operated on the gene_rpkm_report precomputed file. Data were processed in Excel (Supplementary Table 3) and a subset of representative data are presented in Figure 3.

UDP-glycosyltransferases genes in 18 non-melanogaster species were additionally identified, which are D. ananassae (taxID: 7217), D. biarmipes (taxID: 125945), D. bipectinata (taxID: 42026), D. elegans (taxID: 30023), D. erecta (taxID: 7220), D. eugracilis (taxID: 29029), D. ficusphila (taxID: 30025), D. grimshawi (taxID: 7222), D. mojavensis (taxID: 7230), D. persimilis (taxID: 7234), D. pseudoobscura (taxID: 7237), D. rhopaloa (taxID: 1041015), D. sechellia (taxID: 7238), D. simulans (taxID: 7240), D. takahashii (taxID: 29030), D. virilis (taxID: 7244), D. willistoni (taxID: 7260), and D. yakuba (taxID: 7245), in alphabetic order. All the UGTs were classified into families/subfamilies using three complementary approaches. First, D. melanogaster UGT gene/protein sequences were used as queries of other Drosophila genomes available at NCBI using NCBI BLAST. In case of multiple genes in a same gene family, genomic locations were further compared with those of D. melanogaster to confirm the orthologous families/subfamilies they belong. Second, the InterPro database (release 82.0;⁷ Mitchell et al., 2019) was queried using the InterPro signature “UDP-glucuronosyl/UDP-glucosyltransferase” (IPR002213), which is diagnostic of UGT proteins, within the Drosophila genus (taxon ID 7215). Third, the OrthoDB v10.1 database (⁸ Kriventseva et al., 2019) was also queried using the IPR002213 signature within the Drosophila genus (taxon ID 7215) to identify orthologous groups comprising UGT genes. In addition, OrthoDB v9.1 data were obtained via D. melanogaster orthology data present in FlyBase (FB2020_05), primarily to obtain OrthoDB groupings for genes in Drosophila species absent from v10.1 (D. simulans, D. sechellia, D. persimilis). Data were cross-referenced using the NCBI gene IDs, FlyBase gene IDs and/or UniProt accessions present in each database, and the integrated data are shown in Supplementary Table 4. There is a large (mainly 1:1) agreement between the UGT subfamilies defined by the UGT Nomenclature Committee and the orthologous groups defined by OrthoDB (see Supplementary Table 5 for details). Note that several UGT gene models are incorrectly annotated at FlyBase/NCBI, e.g., some gene models need to be split, others need to be merged, others require extending (see Supplementary Table 4 for details). Also note that all non-melanogaster gene models and IDs have been retired from FlyBase and are now annotated and maintained by the NCBI (see the FB2018_06 and FB2020_03 release notes⁹). However, since archived non-melanogaster data are still present in FlyBase, and FlyBase IDs/symbols are still present in many databases, FlyBase gene IDs for the non-melanogaster species are included in Supplementary Table 4.