Subsystem Info

Last Modified:7-3-2015:22:37:47
DescriptionSTATUS: Prediction invalidated, function still unknown.
Predicted to be a thiamine-related transporter, but E. coli protein did not mediate uptake of thiazole or hydroxymethylpyrminidine
Notes1. COG2363 is absent from archaea, present sporadically in diverse bacteria (a-, b-, g-, d-proteobacteria, Firmicutes, Cyanobacteria, Chlamydiae, Bacteroidetes, Acidobacteria) and in plants, animals, and Schizosaccharomyces.

2. COG2363 is a small protein with 3-4 predicted TM helices. There is no obvious pattern of distribution with respect to the thiamin pathway related transporters annotated by Rodionov.

3. COG2363 is clustered with enzymes of thiamin synthesis and metabolism. As thiamin synthesis pathways in bacteria are fairly well known, and as COG2363 is absent from many (thiamine producing) prokaryote genomes, but present in (thiamine non-producing) animals it is not a missing thiamin metabolism enzyme. It is not likely to be a thiamine transporter since it is present in E. coli in which the thiamine transporter is known (sfuABC).

4. COG2363 could be a thiazole transporter; thiamine synthesis mutants of E. coli can be rescued with thiazole, and E. coli, S. typhimurium, B. subtilis, S. pombe (all of which have COG2363), and S. cereviseae (which lacks COG2363) can take up thiazole, but the transporter is not known. Since thiazole and (more so) its oxidation product 4-methylthiazole-5-acetate are urinary catabolites of thiamine in animals, transport of the thiazole moiety in animals must also occur. Since S. cereviseae can transport thiazole but (in common with most ascomycetes) lacks COG2363, it may have another thiazole transporter.

5. Thiazole transport in E. coli & Salmonella shown to be by facilitated diffusion not active transport (unlike thiamine, which is ABC-mediated active transport). Facilitated diffusion is consistent with COG2363 structure i.e. not an ABC component. Thiazole transport NEM-sensitive (Yamasaki Y et al 1973) in E. coli. E. coli COG2363 (YgdD) contains 1 Cys (residue 89), predicted to be outside by TMHMM.

6. Consistent with a function in thiazole transport in bacteria, COG2363 usually (but not invariably) occurs in genomes that have ThiE (the 'marker' enzyme for thiamine synthesis that joins the thiazole and pyrimidine moieties). Genomes where COG2363 is present but ThiE is absent include Francisella (obligate intracellular pathogen) and Parachlamydia (obligate intracellular bacterium). To test for a thiazole uptake function: a COG2363 KO in an E. coli mutant in one of the thiazole synthesis genes (thiFGHIE) should stop rescue by thiazole but not thiamine.

7. E. coli essentiality data non-essential, but YgdD KO shows reduced growth in medium without aromatic amino acids (Smith LK et al 2007), i.e. deleting YgdD creates a partial requirement for aromatic amino acids. As the WT strain had no such requirement, YgdD cannot be required for aromatic amino acid uptake, but must be involved in aromatic amino acid production. Note that thiazole has an aromatic ring.

8. COG2363 not likely to be a hydroxymethylpyridine (HMP) transporter. A very good candidate ThiYXZ (an ABC system) is identified in the Rodionov thiamine biosynthesis subsystem: have upstream Thi elements; binding protein ThiY is similar to HMP synthesis enzymes of fungi). ThiYXZ is not in E. coli. ThiYXZ often co-occurs with COG2363, i.e. there is not an inverse distribution with ThiYXZ (which if present might suggest the same role). Nor is there any inverse relationship between absence of ThiC (required for HMP synthesis) presence of COG2363 (which if present might suggest that HMP is supplied via COG2363. A function in HMP uptake could be ruled out by testing whether an E. coli thiC COG2363 double mutant can be rescued by thiamine but not HMP.

9. Arabidopsis expression data: there is a strong negative correlation (< -0.7) with the two thiamine synthesis genes Thi4 and ThiC (which are highly positively correlated with each other). This could be consistent with recyling of thiamine fragments (involving intercompartmental transport) predominating when synthesis is minimal.

10. Proteomics: COG2363 apparently not localized yet by proteomics studies in Arabidopsis. Generally predicted to be mitochondrial. Has 15-20 residue N-terminal extension relative to most prokaryote COG2363 proteins.

11. Channels often have a small number of TM helices. Therefore COG2363 structure is consistent with a solute channel.
Plant-Prokaryote comparative genomics