Tb927.10.6900 encodes the glucosyltransferase involved in synthesis of base J in Trypanosoma brucei

Abstract

Base J is a DNA modification found in the genome of Trypanosoma brucei and all other kinetoplastids analyzed, where it replaces a small fraction of Ts, mainly in telomeric and chromosome-internal transcription initiation and termination regions. The synthesis of base J is a two-step process whereby a specific T is converted to HOMedU (hydroxymethyldeoxyuridine) and subsequently glucosylated to generate J. The thymidine hydroxylases (JPB1 and JBP2) that catalyze the first step have been characterized, but the identity of the glucosyltransferase catalyzing the second step has proven elusive. Recent bioinformatic analysis by Iyer et al. [Nucleic Acids Res. 2013; 41, 7635] suggested that Tb927.10.6900 encodes the glucosyltransferase (HmdUGT) responsible for converting HOMedU to J in T. brucei. We now present experimental evidence to validate this hypothesis; null mutants of Tb927.10.6900 are unable to synthesize base J. Orthologues from related kinetoplastids show only modest conservation, with several insertion sequences found in those from Leishmania and related genera.

The only modified DNA base identified to-date in the nuclear genome of kinetoplastid parasites is β-d-glucopyranosyloxymethyluracil (base J), which replaces ∼1% of T residues and is predominantly localized within repetitive DNA sequences at telomeres [1-5]. However, a minor fraction of J is also located at chromosome-internal regions that coincide with RNA polymerase (RNAP) II transcriptional initiation and termination sites [6,7], and is required to prevent transcription read-through, at least in Leishmania [7]. Interestingly, J is present only in the bloodstream form of T. brucei [1,2], where it is found in transcriptionally silent variant surface glycoprotein (VSG) expression sites (ES), but not in the single active ES, suggesting a role in regulating antigenic variation [8,9]. Base J synthesis occurs in two steps: hydroxylation of a specific thymine to form the intermediate, hydroxymethyldeoxyuridine (HOMedU), and glucosylation of HOMedU to form J [10,11]. Two thymidine hydroxylases, JBP1 and JBP2, which belong to the TET/JBP superfamily of Fe²⁺/αKG enzymes [12], carry out the first step, but the glucosyltransferase (HmdUGT) that catalyzes the second step has proven elusive. A recent bioinformatic analysis by the Aravind laboratory [13] identified a putative HmdUGT gene immediately downstream of a TET/JBP gene in several phage genomes, and they proposed that the most closely related gene in T. brucei (Tb927.10.6900) encodes the long sought-after glucosyltransferase involved in J synthesis. Here, we provide genetic evidence that validates this hypothesis.

In anticipation that the HmdUGT would not be essential in T. brucei (since JBP1 and/or JBP2 null mutants are viable) [11], we generated Tb927.10.6900 double knockout (dKO) transfectants of bloodstream forms, which normally contain J. We employed an efficient and rapid PCR fusion method [14] to construct DNA fragments containing selectable markers and ∼500 nt from the 5′ and 3′ regions of the Tb927.10.6900 coding sequence. The selectable markers were amplified from vectors that included 5′ and 3′ untranslated regions flanking the EP/PARP (procyclin) and ribosomal protein L4 genes, respectively, to ensure high expression. Transfections were carried out using T. brucei SM427 cells [15]. Single knock-out (sKO) and dKO cell lines were generated by sequential selection for hygromycin (HYG) and/or blasticidin (BSD) resistance. Successful replacement of Tb927.10.6900 was confirmed by PCR of genomic DNA using primers flanking the gene. The Tb927.10.6900 null mutants grew normally and did not display significant phenotypic differences from wild-type (WT). Genomic DNA was isolated from Tb927.10.6900 sKO and dKO cell lines obtained from two independent transfections, digested with DdeI, and J levels were quantified using dot-blot analysis with anti-J antiserum [16]. A slight reduction (∼10-30%, based on fluorescence intensity) in J levels in the sKO versus WT was observed, although this was only marginally statistically significant for one clone (Figure 1). However, all dKO clones appeared to lack J, since they showed only background levels of fluorescence (<20% of WT) at all DNA concentrations tested (Figure 1). An additional blot that included procyclic (J null) T. brucei gDNA as a negative control, and that used a different blocking agent (BLOTTO/TBS + 0.2% Tween 20 rather than 5% powdered skim milk in TBS + 0.2% Tween 20), showed no residual signal in the HmdUGT dKO or procyclic cell gDNA (Supplementary Figure S1). These results indicate that Tb927.10.6900 encodes the HmdUGT responsible for the second step in base J synthesis in T. brucei, consistent with previous analyses showing that Tb927.10.6900 mRNA is more abundant in bloodstream form T. brucei (which have J) than in procyclic forms (which lack J) [17-19].

Fig. 1. Tb927.10.6900-null bloodstream forms lack base J.

(A) Genomic DNA from T. brucei SM427 wild-type (WT), Tb927.10.6900/Δ Tb927.10.6900∷HYG (sKO1), Tb927.10.6900/ΔTb927.10.6900∷BSD (sKO2), Δ Tb927.10.6900∷BSD/ΔTb927.10.6900∷HYG (dKO1A and dKO1B), and ΔTb927.10.6900∷HYG/ΔTb927.10.6900∷BSD (dKO2) bloodstream forms was digested with DdeI, spotted onto a nitrocellulose membrane, and probed with anti-J antiserum and secondary antibody (IRDye® 680RD goat anti-rabbit IgG (H+L)). (B) Fluorescence intensities from three independent dot-blots were measured using LI-COR Image Studio software, normalized as a percentage of the signal obtained for 100 ng WT DNA, and the results plotted as mean plus one standard deviation. Samples that showed statistically significant reduction from the corresponding WT sample for each DNA concentration are indicated by asterisks (* p<0.05, ** p<0.01, *** p<0.001 by two-sample, one-tailed t-test, assuming equal sample variance). dKO1B was present on only one dot-blot and was omitted from this analysis.

Tb927.10.6900 has an orthologue (in a syntenic location) in all other trypanosomatid genomes analyzed, so the corresponding amino acid sequences were retrieved from GenBank and aligned using ClustalW [20] (Supplemental Figure S2). The HmdUGT orthologues showed only modest sequence conservation, ranging from ∼42% identity between T. brucei and T. cruzi, ∼20% between Trypanosoma and Leishmania or Crithidia, and 14-20% between Trypanosoma and Phytomonas, Angomonas or Strigomonas (Supplemental Table S1). Indeed, there was only 80-90% identity between the various Leishmania species and ∼54% identity between Leishmania and Crithidia. In addition, the Leishmania proteins (as well as Crithidia, Phytomonas, Angomonas or Strigomonas) contain several insertion sequences (of ∼25-100 amino acids) that are absent in Trypanosoma. It remains to be seen whether these differences have any functional significance in terms of enzyme activity and/or specificity. We (and our collaborators) are currently expressing HmdUGT protein from several species for x-ray crystallography, as well as attempting to generate conditional knockouts in Leishmania (for which J is essential) to further elucidate the role of HmdUGT in these parasites.

After submission of this manuscript, a study published elsewhere [21] confirmed the complete loss of J after knockout of Tb927.10.6900 in T. brucei and that reintroduction of the gene into Tb927.10.6900-null T. brucei restored J synthesis. That study also showed that reduction of HmdUGT mRNA by RNAi causes reduced J and increased HOMedU levels and that the glucosyltransferase uses uridine diphosphoglucose to transfer glucose to HOMedU.

Highlights.

• In trypanosomatid genomic DNA base J replaces ∼1% of T nucleotides

• Bloodstream forms of Tb927.10.6900 null mutants lack base J

• Tb927.10.6900 encodes the hydroxymethyldeoxyuridine glucosyltransferase (HmdUGT)

• HmdUGT orthologues in other trypanosomatids show only modest sequence conservation