OUP user menu

Involvement of thiaminase II encoded by the THI20 gene in thiamin salvage of Saccharomyces cerevisiae

Mari Onozuka, Hiroyuki Konno, Yuko Kawasaki, Kenichi Akaji, Kazuto Nosaka
DOI: http://dx.doi.org/10.1111/j.1567-1364.2007.00333.x 266-275 First published online: 1 March 2008

Abstract

The physiological significance of thiaminase II, which catalyzes the hydrolysis of thiamin, has remained elusive for several decades. The C-terminal domains of THI20 family proteins (THI20/21/22) and the whole region of PET18 gene product of Saccharomyces cerevisiae are homologous to bacterial thiaminase II. On the other hand, the N-terminal domains of THI20 and THI21 encode 2-methyl-4-amino-5-hydroxymethylpyrimidine kinase and 2-methyl-4-amino-5-hydroxymethylpyrimidine phosphate kinase involved in the thiamin synthetic pathway. In this study, it was first indicated that the C-terminal domains of the THI20 family and PET18 are not required for de novo thiamin synthesis in S. cerevisiae, using a quadruple deletion strain expressing the N-terminal domain of THI20. Biochemical analysis using cell-free extracts and recombinant proteins demonstrated that yeast thiaminase II activity is exclusively encoded by THI20. It appeared that Thi20p has an affinity for the pyrimidine moiety of thiamin, and HMP produced by the thiaminase II activity is immediately phosphorylated. Thi20p was found to participate in the formation of thiamin from two synthetic antagonists, pyrithiamin and oxythiamin, by hydrolyzing both antagonists and phosphorylating HMP to give HMP pyrophosphate. Furthermore, 2-methyl-4-amino-5-aminomethylpyrimidine, a presumed naturally occurring thiamin precursor, was effectively converted to HMP by incubation with Thi20p. It is proposed that the thiaminase II activity of Thi20p is involved in the thiamin salvage pathway by catalyzing the hydrolysis of HMP precursors in S. cerevisiae.

Keywords
  • thiamin
  • thiamin salvage
  • multifunctional enzyme
  • THI20
  • yeast

Introduction

Yeast Saccharomyces cerevisiae is able to synthesize thiamin pyrophosphate de novo which involves the independent formation of pyrimidine and thiazole units, their phosphorylation and subsequent condensation to form thiamin phosphate, its hydrolysis to free thiamin, and finally its pyrophosphorylation (Fig. 1) (Hohmann & Meacock, 1998; Nosaka, 2006). In addition, this organism can efficiently utilize thiamin from the extracellular environment to produce thiamin pyrophosphate (Iwashima et al., 1973). The expression of these genes involved in the synthesis of thiamin pyrophosphate and the utilization of thiamin (THI genes) is induced when the supply of thiamin is limited, a mechanism called the yeast THI regulatory system (Hohmann & Meacock, 1998; Nosaka, 2006). On the other hand, the existence of thiamin-degrading enzyme, thiaminase II (EC 3.5.99.2), has been reported from S. cerevisiae (Kimura & Iwashima, 1987). Thiaminase II catalyzes the simple hydrolysis of thiamin to 2-methyl-4-amino-5-hydroxymethylpyrimidine (HMP) and 4-methyl-5-β-hydroxyethylthiazole (HET). Thiaminase II activity has been detected in various bacteria and fungi (Fujita, 1972; Murata, 1982), and the protein has been purified from Bacillus aneurinolyticus (Ikehata, 1960; Nakatsuka et al., 1987). In addition, the structure of thiaminase II from Bacillus subtilis (TenA) has been determined, and the catalytic mechanism was elucidated (Toms et al., 2005). However, the physiological significance of thiaminase II has remained unclear for several decades. Kimura (1995) reported that thiaminase II activity is low in yeast cells grown with exogenous thiamin and hardly detectable in the extract of the thi2 mutant, lacking a positive regulatory factor in the THI regulatory system. It appears to be paradoxical that the expression of an enzyme degrading thiamin is upregulated in response to thiamin deprivation. It is, therefore, suggested that yeast thiaminase II is involved in thiamin production rather than thiamin degradation and thiamin itself is not the physiological substrate. Very recently, Jenkins (2007) clarified a new thiamin salvage pathway in Bacillus halodurans. In this pathway, a thiazole-degraded thiamin such as an N-formyl-5-aminomethyl derivative of 2-methyl-4-aminopyrimidine is transported into the cell and deformylated, and the resultant 2-methyl-4-amino-5-aminomethylpyrimidine (aminomethylpyrimidine) is hydrolyzed to HMP by TenA. Thus, they clearly demonstrated the physiological significance of bacterial thiaminase II (Bettendorff, 2007).

1

The later steps in the biosynthesis of thiamin pyrophosphate in Saccharomyces cerevisiae. Yeast genes are indicated in italics. Abbreviations are as defined in the text.

The phosphorylation of HMP in the thiamin synthesis pathway proceeds in two-steps catalyzed by HMP kinase (EC 2.7.1.49) and HMP phosphate kinase (HMP-P kinase; EC 2.7.4.7), both of which are encoded by THI20 gene family members (THI20/21) in S. cerevisiae (Llorente et al., 1999; Kawasaki et al., 2005) (Fig. 1). THI20 and THI21 encode proteins containing 551 amino acids (aa) with 86% identity. There is a third member of the family, THI22, whose product is nearly 80% identical to those of the other two members. However, Thi22p seems not to be capable of synthesizing HMP pyrophosphate (HMP-PP) from HMP-P or HMP (Llorente et al., 1999; Kawasaki et al., 2005). These three genes are each composed of two distinct domains and appear to be the result of gene fusions during evolution: the N- and C-terminal parts are homologous to bacterial thiD and tenA genes, respectively. The HMP kinase and HMP-P kinase activities of Thi20p and Thi21p reside in the N-terminal parts (1–300 aa). The single null strains for THI20 and THI21 grow at the same rate as the parental strain in medium lacking thiamin, but the simultaneous disruption of both genes leads to thiamin auxotrophy. Although THI20 and THI21 are functionally redundant in terms of de novo thiamin synthesis, Thi20p has higher ability to synthesize HMP-PP from HMP than Thi21p (Kawasaki et al., 2005). The sequences of the C-terminal parts of THI20/21/22 are similar to the sequence of the S. cerevisiae PET18 gene over its full length (about 20% aa identity). The expression of PET18 is also induced in the absence of thiamin (Llorente et al., 1999; Nosaka et al., 2005), but the function of the PET18 product is unknown. Recently, Haas (2005) reported the ability of Thi20p to catalyze the hydrolysis of thiamin. It is notable and apparently curious that both thiamin synthetic and degradative enzymes exist in one fusion protein, Thi20p. This fact also suggests the physiological function of yeast thiaminase II. In this study, the thiaminase II activities of the THI20 family members and PET18 were compared and their physiological significance is discussed. It is proposed that the thiaminase II activity of Thi20p participates in salvage thiamin synthesis by catalyzing the hydrolysis of precursors of HMP such as aminomethylpyrimidine in S. cerevisiae.

Materials and methods

Strains and media

Escherichia coli strain DH5α was used to amplify plasmids and express the recombinant proteins (Thi6p, Thi20p, Thi21p), and BL21(DE3)pLys was used to express recombinant Pet18p. Yeast strains used in this study are listed in Table 1. Yeast quadruple deletion strain of thi20/21/22 and pet18 (NKC24) was constructed as follows: deletion of THI22 was introduced into BY4741 by a PCR-directed integration method (Baudin et al., 1993) using LEU2 as a selectable marker to give strain NKC22; a 3.5-kb DNA fragment containing thi22::LEU2 was PCR-amplified from NKC22 genomic DNA and transformed into NKC11 (Δthi20Δthi21) (Kawasaki et al., 2005); the resultant triple deletion strain, NKC23, was then transformed with a 2.1-kb DNA fragment containing pet18::kanr, which was PCR-amplified from the genomic DNA of a strain BY4741/pet18::kanr (Open Biosystems). The primers used in these procedures are indicated in supplementary Table S1. The newly constructed strains were confirmed to be correctly disrupted by restriction mapping of PCR-isolated fragments from the genomic DNA. Standard media and growth conditions for yeast cells were as described previously (Nosaka et al., 1994, 2005).

View this table:
1

Yeast strains used in this study

StrainGenotypeSources or references
BY4741MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0Open Biosystems
BY4741/thi20thi20::kan r in BY4741Open Biosystems
BY4741/thi21thi21::kan r in BY4741Open Biosystems
BY4741/thi22thi22::kan r in BY4741Open Biosystems
BY4741/pet18pet18::kan r in BY4741Open Biosystems
YPH500MATαade2-101 his3-Δ200 leu2-Δ1 lys2-801 trp1-Δ63 ura3-52Our stock
NKC11thi20::HIS3 thi21::TRP1 in YPH50013
NKC22thi22::LEU2 in BY4741This study
NKC23thi20::HIS3 thi21::TRP1 thi22::LEU2 in YPH500This study
NKC24thi20::HIS3 thi21::TRP1 thi22::LEU2 pet18::kan r in YPH500This study

Plasmids

Yeast and bacterial expression vectors used in this study are listed in Table 2. The newly constructed vectors were pYES2/NT-THI20(1–302), pYES2/NT-THI20(300–551), and pRSET-PET18. The N- and C-terminal coding regions of THI20 were PCR-amplified from the vector pYES2/NT-THI20 using specific primers into which BamHI and XhoI sites were designed, and then the fragment obtained was digested with the restriction enzymes and subcloned into pYES2/NT B. To construct pRSET-PET18, the coding sequence of PET18 was isolated from yeast genomic DNA using PCR with specific primers containing BamHI and XhoI sites at their respective termini, and cloned into pRSET B at the BamHI/XhoI sites. All the plasmids were verified to be constructed correctly by sequencing using an ABI Prism 377 DNA sequencer (Applied Biosystems).

View this table:
2

Plasmids used in this study

PlasmidDescriptionSources
pYES2/NT BYeast expression vector, GAL1 promoter, 2μ origin, URA3Invitrogen
pYES2/NT-THI20THI20 with a six histidine-tag at N-terminus in pYES2/NT B13
pYES2/NT-THI20 (1–302)1–302 aa region of THI20 with a six histidine-tag at N-terminus in pYES2/NT BThis study
pYES2/NT-THI20 (300–551)300–551 aa region of THI20 with a six histidine-tag at N-terminus in pYES2/NT BThis study
pTrcHis2-TOPOBacterial expression vector, trc promoter, ampr, pBR322 originInvitrogen
pTrc-THI6THI6 with a six histidine-tag at C-terminus in pTrcHis2-TOPO13
pTrc-THI20THI20 with a six histidine-tag at C-terminus in pTrcHis2-TOPO13
pTrc-THI21THI21 with a six histidine-tag at C-terminus in pTrcHis2-TOPO13
pRSET BBacterial expression vector, T7 promoter, ampr, pUC originInvitrogen
pRSET-PET18PET18 with a six histidine-tag at N-terminus in pRSET BThis study

Chemicals

HMP, HMP-P, and aminomethylpyrimidine were kind gifts from the late S. Yurugi (Takeda Pharmaceutical Industries, Osaka, Japan), and thiamin acetic acid was synthesized previously by K. Sempuku in the authors' laboratory according to the method of Neal (1969). The identity of aminomethylpyrimidine and thiamin acetic acid was confirmed by their nuclear magnetic resonance (NMR) spectra using a Bruker AM-300 instrument. 1H NMR (D2O) of aminomethylpyrimidine: δ 2.53 (s, 3H), δ 4.14 (s, 2H), δ 8.18 (s, 1H). 1H NMR (D2O) of thiamin acetic acid: δ 2.52 (s, 6H), δ 3.85 (s, 2H), δ 4.83 (s, 1H), δ 5.47 (s, 2H), δ 8.04 (s, 1H). Fast atom bombardment (FAB) MS of aminomethylpyrimidine using a JEOL JMS-SX102A gave a single ion with (m/z): [M+H]+calcd. for C6H11N4, 139.0984; found, 139.0987. HET was purchased from Sigma, and pyrithiamin and oxythiamin were from Aldrich. All other chemicals were of analytical grade.

Preparation of yeast cell-free extract

Yeast strains, BY4741 (wild-type) or the derivatives, were first grown in 25 mL of thiamin-lacking medium with 2% glucose to OD600 nm of around 0.8, pelleted, and grown for 16 h in 200 mL of the same medium. When thi20Δ strain harboring the plasmid was harvested, the cells were first grown with 2% raffinose and subsequently grown with 2% galactose. All subsequent steps were done at 4 °C. The cells were pelleted and resuspended in 1.5 mL of 50 mM Na2HPO4 pH 7.5 containing 300 mM NaCl, 1 mM phenylmethylsulfonyl fluoride (PMSF), and 10 μL mL−1 of Protease Inhibitor Cocktail for Yeast Cells (Sigma). Glass beads were added, and the cells were lysed in a bead beater (BioSpec Products, Bartlesville, OK) by beating for five 60 s pulses with 5 min intervals on ice. After incubating for a further 20 min on ice, the lysate was aspirated from the beads. Cell debris was removed by two rounds of centrifugation at 14 000 g for 20 min, and the supernatant (about 1 mL) was dialyzed twice against 500 mL of 50 mM Tris-HCl pH 7.5 and 1 mM PMSF. The prepared cell-free extract was stored at −20 °C after adding an equal volume of glycerol.

Purification of recombinant proteins

The pTrcHis2-based plasmids and pRSET-PET18 were transformed into E. coli strains DH5α and BL21(DE3)pLys, respectively. Bacterial cells were grown overnight at 37 °C in 10 mL of Luria–Bertani medium containing 50 μg mL−1 ampicillin, pelleted, and grown for 3 h in 100 mL of fresh medium. The cells were further shaken at 28 °C for 2 h with 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG) to induce expression of the recombinant protein. All subsequent steps except the elution were done at 4 °C. The cells were then pelleted and resuspended in 10 mL of solution L (50 mM Na2HPO4 pH 7.5, 300 mM NaCl) containing 20 mM imidazole, 1 mM PMSF and 10 μL mL−1 Protease Inhibitor Cocktail for Histidine-Tagged Proteins (Sigma), and sonicated four times in ice-cold water using a Bioruptor (Cosmo Bio, Tokyo, Japan) at 200 W for 30 s each with a 120 s interval. Cell debris was removed by centrifugation at 14 000 g for 20 min, and the supernatant served as a crude extract. The prepared crude extract was combined with 1 mL of a 50% slurry of Ni-NTA agarose resin (Qiagen) and mixed on a rotating wheel for 60 min. The resin was washed with 10 mL of solution L containing 20 mM imidazole. The purified protein was eluted with 2 mL of solution L containing 125 mM imidazole and desalted by Sephadex G-25 (Amersham) equilibrated with 50 mM Tris-HCl pH 7.5. Combined fractions were concentrated to about 0.5 mL by ultrafiltration using Centricon YM-10 (10 kDa cutoff, Millipore).

Enzyme assay

The thiaminase II activity was assessed by determining the decrease in the amount of substrates or the increase in the amount of products. In the standard assay, 0.5 mL of a mixture containing 50 mM Tris-HCl pH 7.5, 100 μM thiamin or other substrate, 1 mM dithiothreitol, and enzyme source was incubated at 37 °C. The reaction was stopped with 0.1 mL of 30% trichloroacetic acid (TCA). After TCA was removed by two rounds of extraction each with two volumes of ethyl ether, the amount of substrate (thiamin, thiamin phosphate, thiamin pyrophosphate, thiamin acetic acid) in the aqueous layer was determined by HPLC after conversion to thiochrome by alkaline oxidation with cyanogen bromide as described previously (Nosaka et al., 1989). When the hydrolysis of thiamin by the purified protein was analyzed, the amount of HET in the reaction mixture was determined, without removing TCA, by HPLC using UV absorbance. When the hydrolysis of aminomethylpyrimidine by the purified protein was monitored, HMP and aminomethylpyrimidine in the reaction mixture were detected, without adding TCA, by HPLC using UV absorbance. The synthesis of thiamin from HMP- and HET-precursors was assessed by a defined biochemical system. The reaction mixture contained 50 mM Tris-HCl pH 7.5, 10 mM ATP, 20 mM MgCl2, 1 mM dithiothreitol, Thi6p (2 μg), Thi20p or Thi21p (2 μg), and a combination of 10 μM precursors in a final volume of 1.0 mL. After incubation at 37 °C, the reaction was stopped with 0.2 mL of 30% TCA. The amount of thiamin phosphate in the mixture was determined fluorometrically by HPLC after removing trichloroacetic acid as above.

The HPLC system consisted of a model LC-10AD liquid delivery module (Shimadzu, Kyoto), a model 9725 sample injection valve (Rheodyne), a Capcell Pak C18 UG120 column (150 mm × 4.6 mm i.d., Shiseido, Tokyo), a model RF-10AXL fluorescence detector (excitation 370 nm, emission 430 nm, Shimadzu), and a model SPD-10AV UV detector (254 nm, Shimazdu). The solvent to determine the fluorescence of thiochromes from thiamin and thiamin derivatives was 25 mM potassium phosphate buffer pH 8.4 containing 8% acetonitrile. The flow rate was 1.4 mL min−1 and the corresponding thiochromes of thiamin pyrophosphate, thiamin phosphate, thiamin acetic acid, and thiamin eluted at 1.7, 1.9, 3.0, and 9.3 min, respectively. The solvent to determine the UV absorbance was 20 mM potassium phosphate buffer pH 6.6 containing 10% methanol. The flow rate was 1.0 mL min−1, and the retention times were as follows: HMP-P, 2.0 min; aminomethylpyrimidine, 2.7 min; thiamin, 3.9 min; HMP, 4.8 min; HET, 15.1 min. Under these conditions, large amounts of TCA were eluted overlapping with aminomethylpyrimidine, thiamin, and HMP.

Results

Effect of simultaneous disruption of the THI20 gene family members and PET18

A yeast deletion strain of both THI20 and THI21 genes displays thiamin auxotrophy (Llorente et al., 1999; Kawasaki et al., 2005). A deficiency in HMP-P kinase activity involved in de novo thiamin synthesis is responsible for this thiamin auxotrophy, because the forced expression of the N-terminal part of Thi20p in the mutant cells led to thiamin prototrophy (Fig. 2) (Llorente et al., 1999). To investigate the involvement of the C-terminal part of the THI20 gene family and PET18 in de novo thiamin synthesis, a quadruple deletion strain of thi20/21/22 and pet18 was constructed, and the N-terminal part of Thi20p was expressed in the cells. As shown in Fig. 2, similar to the double mutant cells, the quadruple mutant cells with pYES2-THI20(1–302) could grow in medium without thiamin. This result clearly showed that the C-terminal parts of the THI20 gene family and PET18 are not required for de novo thiamin synthesis in S. cerevisiae.

2

Growth of double and quadruple mutant strains in medium lacking thiamin. Yeast strains harboring pYES2/NT B or pYES2-THI20 (1–302) were grown in 1 mL of medium containing 10 nM thiamin with raffinose and harvested at c. 3 × 107 cells mL−1. Aliquots (4 μL) of 10−1 serial dilutions beginning with 1.5 × 106 cells mL−1 were spotted onto a thiamin-lacking medium with galactose. Agar plates were incubated at 30°C for 3 days.

Thiaminase II activity in yeast cell-free extract

Thiaminase II activity was determined, using thiamin as the substrate, in cell-free extracts of S. cerevisiae and the deletion strains for THI20 gene family members and PET18. Optimal pH for the reaction was around 7.5 and dithiothreitol was necessary for expression of the activity (data not shown). Predictably, marked activity was detected in the extract of the wild-type strain BY4741 when thiamin was withdrawn from the medium (Table 3). It was found that thiaminase II activity was almost completely abolished in the thi20 deletion strain, and the activity in the thi21 strain decreased to some extent. On the other hand, both thi22 and pet18 deletion strains expressed the same activity as the parental strain. Then an attempt was made to produce the entire Thi20p peptide, and also the N- and C-terminal parts of Thi20p, in the thi20 strain from the GAL1 promoter of the expression vector pYES2/NT B. The expression of the transformed gene was verified by reverse transcriptase (RT)-PCR analysis (data not shown) and subsequently the thiaminase II activity was investigated using cell-free extracts. As shown in Table 3, the activity in the thi20 deletion strain was elevated by expressing recombinant Thi20p or Thi20p(300–551), whereas the cells with pYES2-THI20(1–302) had slight activity as low as that with pYES2-NT B. These results indicated that thiaminase II activity in S. cerevisiae is mostly derived from the THI20 gene and the catalytic domain of the activity is located in the C-terminal part of Thi20p.

View this table:
3

Thiaminase II activity of yeast cell-free extract

StrainPlasmidActivity (nmol thiamin decreased mg−1 h−1)
WildNone4.8
WildNone0.52
thi20ΔNone0.46
thi21ΔNone3.0
thi22ΔNone4.4
pet18ΔNone4.7
thi20ΔpYES2/NT B0.61
thi20ΔpYES2-THI202.6
thi20ΔpYES2-THI20 (1–302)0.55
thi20ΔpYES2-THI20 (300–551)1.6
  • * The cells were grown in medium containing 1 μM thiamin.

  • Yeast strains were grown in thiamin-lacking medium unless otherwise noted, and the cell-free extracts were prepared as described in ‘Materials and methods’. The thiaminase II activity was assessed using 100 μg of cell-free extract. After incubation for 3 h, the amount of thiamin was determined. Each value is the mean from two experiments.

Catalytic properties of thiaminase II activity of Thi20p

To compare the catalytic activity of thiaminase II, Thi20p, Thi21p, and Pet18p were separately overexpressed in E. coli and partially purified (see supplementary Fig. 1S). Recombinant Thi22p could not be obtained by affinity chromatography using Ni-NTA resin for an unexplained reason, as observed previously (Kawasaki et al., 2005). The activity was assessed by determination of the amount of HET released from thiamin. In accordance with the observation using cell-free extract, the purified fraction of Thi20p showed high thiaminase II activity of 78.3 nmol HET formed mg−1 min−1. The activity of the purified Thi21p fraction was only 0.1-fold of Thi20p and activity was not detected in the Pet18p fraction. The catalytic properties of Thi20p regarding its thiaminase II activity was further characterized. Although thiamin was hydrolyzed time-dependently by Thi20p, the reaction proceeded linearly for at most the initial 1 min. This phenomenon suggested that the activity was strongly inhibited by the hydrolysis product, and it was difficult to determine the catalytic parameters of thiaminase II activity of Thi20p. Then, the inhibitory effects of HMP and HET on thiaminase II activity of Thi20p were investigated. The influence of the presence of the HET, which was assessed by determination of the decrease in thiamin, was barely observed (data not shown). On the other hand, the hydrolysis of thiamin was inhibited by 98% when HMP was added to the reaction mixture at a ratio of HMP to thiamin of 1 : 1 (Fig. 3). Even at a ratio of 1 : 100, the activity was inhibited by as much as 44%. This finding indicated that thiaminase II activity of Thi20p exhibited a strong product inhibition. However, HMP-P showed no inhibitory effect on the hydrolysis of thiamin. Given the fact that HMP-P is synthesized from HMP by the HMP kinase activity of Thi20p (Haas et al., 2005; Kawasaki et al., 2005), the effect of the addition of ATP with Mg to the thiaminase II reaction mixture was investigated. As shown in Fig. 3, the hydrolysis of thiamin increased about 135% following the addition of ATP, suggesting that HMP produced from thiamin by Thi20p is immediately phosphorylated by the N-terminal domain of the same protein.

3

Effect of HMP, HMP-P, aminomethylpyrimidine, and ATP on hydrolysis of thiamin by Thi20p. The hydrolysis of thiamin was assessed using purified Thi20p (1 μg) and 100 μM thiamin as described in ‘Materials and methods’, with the indicated concentrations of compounds. Effect of MgATP was carried out by addition of 1 mM ATP with 5 mM MgCl2. After incubation for 5 min, the amount of HET was determined. Each value is the mean from two experiments. Note that the hydrolysis might be inhibited to some extent by HMP released from thiamin during the 5 min incubation. amino-MP, aminomethylpyrimidine.

Effect of pyrithiamin and oxythiamin on the growth of yeast cells

Saccharomyces cerevisiae can grow in the presence of two synthetic thiamin antagonists, pyrithiamin and oxythiamin, although its growth is inhibited by each antagonist alone (Iwashima et al., 1984). Pyrithiamin is modified at the thiazole portion of thiamin, while the 4′-amino group of the pyrimidine portion of thiamin is substituted by a hydroxy group in oxythiamin. Kimura & Iwashima (1987) proposed that pyrithiamin and oxythiamin are hydrolyzed by thiaminase II present in the yeast cytosol, and their respective degradation products, HMP and HET, are utilized as precursors for thiamin synthesis. The effect of both antagonists on the growth of the deletion strains for THI20 gene family members and PET18 was examined. As shown in Fig. 4, the wild-type cells grew in the presence of both pyrithiamin and oxythiamin. It was confirmed that the simultaneous addition of both antagonists led to three times higher contents of intracellular thiamin in comparison with normal growth condition (2.4 vs. 0.83 pmol total thiamin per 106 cells), suggesting that the degradation products, HMP and HET, were utilized as thiamin precursors. However, among the deletion strains tested, only thi20 mutant cells could not grow in the presence of both antagonists (Fig. 4), indicating that the THI20 gene is necessary for the utilization of pyrithiamin and oxythiamin as precursors for thiamin. This result seems to be in accordance with a significant loss in thiaminase II activity in the cell-free extract of the thi20 deletion strain.

4

Growth of yeast strains in the presence of thiamin antagonists. Yeast strains were grown in 1 mL of medium without thiamin and harvested at c. 3 × 107 cells mL−1. Aliquots (4 μL) of 10−1 serial dilutions beginning with 1.5 × 106 cells mL−1 were spotted onto a thiamin-lacking medium supplemented with 1 μM pyrithiamin and/or 10 μM oxythiamin. Agar plates were incubated at 30°C for 4 days.

To confirm the involvement of Thi20p in thiamin synthesis from pyrithiamin and oxythiamin, the biosynthesis of thiamin from both antagonists was examined in a defined biochemical system containing Thi20p and Thi6p, a bifunctional thiamin phosphate synthase (EC 2.5.1.3) and HET kinase (EC 2.7.1.50) (Fig. 1) (Nosaka et al., 1994). The reaction mixture contained pyrithiamin, oxythiamin, ATP, MgCl2, dithiothreitol, Thi6p, and Thi20p, and the level of synthesized thiamin phosphate was determined. As shown in Fig. 5a, 3.0% conversion based on 10 μM antagonist was detected after 2 h. On the other hand, the conversion was not observed from the same concentration of antagonists when Thi21p was supplied instead of Thi20p. Thiamin synthesis was not detected when Thi20p or Thi6p was omitted from the reaction mixture. These results indicated that both synthetic antagonists were cleaved at the C-N bond by Thi20p, and the respective products, HMP and HET, were incorporated into thiamin structure after being phosphorylated (Fig. 5b).

5

Thiamin formation from pyrithiamin and oxythiamin. (a) Reconstitution of thiamin synthesis from pyrithiamin and oxythiamin by purified proteins. The reaction mixture and the assay procedure are described in ‘Materials and methods’. The starting concentrations of pyrithiamin and oxythiamin were 10 μM. The amounts of thiamin phosphate in the reaction mixture (1 mL) are shown. Each point is the mean from two experiments. Symbols indicate the following combination of proteins: ●, Thi20p and Thi6p; ○, Thi21p and Thi6p; ◆, Thi20p alone; ◊, Thi6p alone. (b) Thiamin synthetic pathway from pyrithiamin and oxythiamin catalyzed by Thi6p and Thi20p.

Examination of aminomethylpyrimidine as a candidate for the physiological substrate of Thi20p

Because Thi20p has a high affinity for the pyrimidine part of thiamin and subsequently phosphorylates HMP via the activity of its N-terminal domain, it seems likely that the physiological role of the thiaminase II activity of Thi20p is the salvage of HMP from HMP precursors. Three thiamin metabolites, thiamin phosphate, thiamin pyrophosphate, and thiamin acetic acids as substrates for Thi20p, were tested. Thiamin acetic acid has been identified as the initial metabolite of thiamin in a soil microorganism (Neal, 1969). However, the hydrolysis of these three compounds was below the lower limit of detection.

On the other hand, Diorio & Lewin (1968) reported that certain thiamin-requiring strains of Neurospora crassa excrete presumed thiamin precursors into the medium. They include 5-formyl, 5-methoxymethyl, and 5-aminomethyl derivatives of 2-methyl-4-aminopyrimidine in addition to the 5-hydroxymethyl derivative (HMP). Of these compounds, 5-aminomethyl-2-methyl-4-aminopyrimidine (aminomethylpyrimidine) has the C–N bond at C-5′ of the pyrimidine. Kimura (1995) had already observed that thiamin production from aminomethylpyrimidine by resting yeast cells is almost the same as that from HMP, and yeast cell-free extracts have been shown to contain an enzyme activity catalyzing the conversion of aminomethylpyrimidine into HMP by bioautography. Moreover, in the course of partial purification of thiaminase II from S. cerevisiae, it was found that thiaminase II is eluted from a DEAE-cellulose column with the activity for converting aminomethylpyrimidine into HMP. Therefore, the ability of Thi20p to catalyze the hydrolysis of aminomethylpyrimidine was investigated. As shown in Fig. 6a, HMP was released from aminomethylpyrimidine by incubation with purified Thi20p. Then, the synthesis of thiamin from aminomethylpyrimidine and HET was examined using the defined biochemical system containing Thi20p and Thi6p as above. To compare the catalytic efficiency with respect to the substrate for thiaminase II activity, thiamin or pyrithiamin were supplied instead of aminomethylpyrimidine in the defined system. After 2 h incubation, 25% of aminomethylpyrimidine at a concentration of 10 μM was converted into thiamin phosphate, whereas 9.6% and 3.9% conversion was detected from the same concentrations of thiamin and pyrithiamin, respectively (Fig. 6b). These results suggested that aminomethylpyrimidine was hydrolyzed more than two times faster than thiamin by Thi20p. It was also found that the hydrolysis of thiamin by Thi20p was inhibited in the presence of aminomethylpyrimidine to the same extent as in the presence of HMP (Fig. 3), strongly suggesting that Thi20p has much higher affinity for aminomethylpyrimidine than thiamin.

6

Hydrolysis of aminomethylpyrimidine to give HMP. (a) HPLC chromatogram showing the conversion of aminomethylpyrimidine to HMP. Aminomethylpyrimidine (10 μM) was incubated with or without purified Thi20p (1 μg) for 30 min as described in ‘Materials and methods’, and the reaction mixture was analyzed. Peaks indicate the following compounds: 1, aminomethylpyrimidine; 2, HMP; 3, dithiothreitol. (b) Reconstitution of thiamin synthesis from aminomethylpyrimidine and HET by Thi6p and Thi20p. The procedures were performed as described in the legend of Fig. 5a, except for the starting materials. The starting concentrations of all precursors were 10 μM. Each point is the mean from two experiments. Note that HET should be produced by the hydrolysis of thiamin. Symbols indicate the following combinations of precursors: ●, aminomethylpyrimidine and HET; ○, thiamin and HET; ◆, pyrithiamin and HET.

Discussion

By biochemical analysis using cell-free extracts and recombinant proteins, the fact that thiaminase II activity is exclusively encoded by the THI20 gene in S. cerevisiae was demonstrated. This implies that Thi20p is a trifunctional protein with HMP kinase, HMP-P kinase, and thiaminase II activities, as Haas (2005) had already noted. It is proposed that the physiological function of the thiaminase II activity of Thi20p is the salvage of HMP for thiamin synthesis from the following facts: the expression of the THI20 gene is induced in response to thiamin deprivation under the control of the yeast THI regulatory system (Llorente et al., 1999; Nosaka et al., 2005); Thi20p is a fusion protein that consists of thiaminase II and the enzyme involved in thiamin synthesis; the thi20 deletion strain is deficient in the utilization of pyrithiamin and oxythiamin as precursor of thiamin synthesis; the thiaminase II domain of Thi20p is not required for de novo thiamin synthesis. It is highly likely that HMP is not an obligatory intermediate in the pathway of de novo thiamin synthesis, and that HMP kinase is involved in the salvage synthesis of HMP-P from HMP in S. cerevisiae (Nosaka, 2006). Because not only thiaminase II activity but also HMP kinase activity of Thi21p is slight (Kawasaki et al., 2005), the utilization of HMP precursors for HMP-P synthesis is primarily catalyzed by Thi20p in S. cerevisiae. In the catalytic reaction of Thi20p, a product of the hydrolysis can be effectively utilized as the substrate for the subsequent phosphorylations. There is another example of gene fusion in thiamin synthesis in S. cerevisiae: THI6 consists of a thiamin phosphate synthase domain and a HET kinase domain (Nosaka et al., 1994). Interestingly, both of the C-terminal domains of the THI6 and THI20 genes code for the salvage enzyme, HET kinase and thiaminase II, respectively. It is known that S. cerevisiae developed mechanisms to utilize extracellular HMP and HET to form thiamin phosphate. HMP is taken up by a common transport system with thiamin (Iwashima et al., 1990), whereas HET is taken up by diffusion, followed by metabolic trapping by HET kinase-catalyzed phosphorylation (Iwashima et al., 1986). An assumption was made that the synthesis of both the pyrimidine and thiazole parts of thiamin via the salvage pathways play a significant role in retaining intracellular levels of this vitamin in S. cerevisiae.

From observations of thiamin-requiring strains of Neurospora crassa, Diorio & Lewin (1968) postulated that aminomethylpyrimidine is deaminated to the 5-formyl compound, which is then reduced to HMP. However, whether such a two-step processing pathway to yield HMP from aminomethylpyrimidine exists in yeast and other organism remains to be clarified. This study demonstrated that aminomethylpyrimidine is a good substrate for the hydrolysis catalyzed by Thi20p. It was therefore suggested that aminomethylpyrimidine is a physiological substrate for Thi20p and utilized as a direct precursor of HMP in S. cerevisiae. On the other hand, the hydrolysis of aminomethylpyrimidine was scarcely detected by incubation with Thi21p or Pet18p (data not shown). The possibility that unidentified HMP analogs other than aminomethylpyrimidine serve as the physiological substrate of Thi20p and its family members cannot be ruled out.

There is another type of thiamin-degrading enzyme, thiaminase I (EC 2.5.1.2), which catalyses the cleavage of thiamin by an exchange of the thiazole moiety with a variety of nucleophiles (Fujita, 1972; Murata, 1982; Bos & Kozik, 2000). One might assume that thiaminase I is also concerned with the salvage of pyrimidine or thiazole moieties. It has been noted that thiaminase II has been discovered exclusively in organisms able to synthesize thiamin de novo such as bacteria, fungi, and yeast. In these microorganisms, HMP produced by thiaminase II can be phosphorylated and coupled with HET phosphate to generate thiamin phosphate. On the other hand, thiaminase I has been found not only in bacteria and plants but also shellfish, freshwater fish, and some sea fish (Fujita, 1972; Murata, 1982; Bos & Kozik, 2000). In the latter organisms, thiamin is a nutrient and the thiaminase I-catalyzed products cannot be utilized as thiamin precursors. Thiaminase I may have an undefined physiological role other than salvage thiamin synthesis.

Supplementary material

The following supplementary material is available for this article online:

Fig. S1. SDS-PAGE analysis of the purified proteins.

Table S1. Oligonucleotides used in this study.

Acknowledgements

The authors are grateful to Dr K. Oda (Kyoto Pharmaceutical University) for FAB mass spectrometry measurement. This work was supported in part by Grants-in-Aid for Scientific Research 18580095 from the Japan Society for the Promotion of Science and by Individual Research Grants of Doshisha Women's College of Liberal Arts.

Footnotes

  • Editor: Terrance Cooper

References

View Abstract