Structural Analysis of Escherichia Coli ThiF

https://doi.org/10.1016/j.jmb.2005.04.011Get rights and content

Escherichia coli ThiF is an enzyme in the biosynthetic cascade for generating the essential cofactor thiamin pyrophosphate. In this cascade, ThiF catalyzes adenylation of the C terminus of ThiS. We report here the crystal structures of ThiF, alone and in complex with ATP. The structures provide insight into a preference for ATP during adenylation of the protein ThiS. Additionally, the structures reveal an ordered crossover loop predicted to clamp the flexible tail of ThiS into the ThiF active site during the adenylation reaction. The importance of the crossover loop for ThiF activity is highlighted by mutational analysis. Comparison of ThiF with the structural homologues MoeB, APPBP1-UBA3, and SAE1-SAE2 reveals that the ATP-binding site, including an arginine-finger, is maintained throughout evolution, and shows divergence occurring in protein substrate-binding sites and regions devoted to unique steps in the specific function of each enzyme.

Introduction

Thiamin pyrophosphate is an essential cofactor in carbohydrate and branched amino acid metabolism. The biosynthesis of thiamin involves a complex cascade of enzymes with distinct paths for synthesizing the thiazole and pyrimidine moieties, which are then coupled and phosphorylated to yield thiamin pyrophosphate.1, 2, 3, 4 In Escherichia coli, formation of the thiazole moiety involves the actions of six distinct gene products: ThiS, ThiF, ThiG, ThiH, IscS and ThiI.5, 6, 7 The thiazole moiety is synthesized from deoxy-d-xylulose 5-phosphate (DXP), tyrosine, and a sulfur atom carried by the C terminus of ThiS, a 66 residue protein (Figure 1(a)).

ThiF is an essential enzyme for thiamin biosynthesis.5, 8 Current evidence suggests that ThiF is involved in two distinct steps in the generation of the thiazole moiety in E. coli. First, ThiF catalyzes formation of a critical ThiS-acyl-adenylate (ThiS-COAMP) intermediate. Evidence for ThiF-catalyzed adenylation at the ThiS C terminus is as follows: (1) a purified ThiF–ThiS complex was shown to catalyze formation of pyrophosphate in a MgATP-dependent manner; and (2) mass spectrometric characterization of MgATP-treated ThiF–ThiS purified from a ThiI strain of E. coli revealed the presence of the ThiS-COAMP product.8 Following formation of ThiS-COAMP, IscS catalyzes transfer of sulfur from cysteine to yield ThiS-thiocarboxylate. ThiS-thiocarboxylate formation has been achieved in vitro in the presence of ThiF and IscS.6, 9 ThiI has been implicated in formation of the ThiS-thiocarboxylate intermediate, although ThiI is not essential, and the exact catalytic function of ThiI remains poorly understood.8 Subsequent to formation of the ThiS-thiocarboxylate, E. coli ThiF has been implicated in a second role in thiazole biosynthesis: a covalently linked ThiF–ThiS complex was identified from a partially purified mixture of ThiF and ThiS.10 Mass spectrometry revealed this complex contains an acyldisulfide linkage between the sulfur atoms of the ThiS-thiocarboxylate and Cys184 of ThiF.10 Mutation of ThiF Cys184 to serine did not affect generation of the ThiS-thiocarboxylate intermediate, indicating that formation of the acyldisulfide-linked ThiF–ThiS complex follows ThiS-thiocarboxylate formation.10 Although a covalently linked ThiF–ThiS complex has not been detected from Bacillus subtilis,11 the physiological importance of the ThiF–ThiS covalent complex in E. coli is underscored by the finding that the ThiF Cys184Ser mutant did not complement the growth defect of ThiF E. coli.10 Exogenous addition of thiazole restores normal E. coli growth, attributing the defect to thiazole biosynthesis.10 Thus, previous results suggest two critical functions for ThiF in thiazole biosynthesis in E. coli: adenylation of the ThiS C terminus, and formation of a covalent acyldisulfide complex.10

Ultimately, the thiazole moiety is generated from DXP, tyrosine, and the sulfur atom carried at the C terminus of ThiS, in a reaction catalyzed by ThiG, or thiazole synthase.4, 11, 12 Interestingly, ThiG has been shown to compete with ThiF for binding to ThiS, indicating a requirement for the ThiF-mediated and ThiG-mediated reactions to occur sequentially.13

The structures of E. coli and B. subtilis ThiS have been determined recently, both alone in solution (from E. coli), and in complex with ThiG (from B. subtilis).13, 14 The ThiS structure consists of a globular domain containing a four-stranded β-sheet and an α-helix, and a flexible C-terminal tail that terminates in the sequence Gly–Gly. Interestingly, the ThiS structure revealed striking similarity to the E. coli protein MoaD, which also carries a sulfur atom at its C terminus for the biosynthesis of molybdopterin.15 In addition, ThiS resembles the eukaryotic protein ubiquitin, and ubiquitin-like proteins (ubls) such as NEDD8, SUMO and ISG15, raising the possibility of common ancestry for ThiS and ubiquitin and ubls.14 Ubiquitin and ubls are small proteins with C termini that become covalently attached to other proteins as post-translational modifications in eukaryotes.16 Ubiquitin and ubls are involved in essential eukaryotic mechanisms for regulating processes such as cell division, development, the immune system and cellular trafficking.16, 17 These post-translational modifications serve as tags that alter the activities of their targets, such as changing a protein's half-life, subcellular localization, or catalytic activity.18, 19, 20, 21, 22

Further evidence for an evolutionary relationship between the thiamin biosynthetic pathway in prokaryotes and ubiquitin and ubl post-translational modification pathways in eukaryotes comes from sequence similarity between ThiF and portions of the E1 enzymes that initiate ubiquitin and ubl conjugation.17 Ubiquitin and ubls are conjugated to their protein targets through parallel, multienzyme cascades.17, 23 For each ubl, a dedicated E1 catalyzes multiple reactions. First, the E1 binds ATP, Mg2+, and its specific ubl, and catalyzes adenylation of the ubl's C terminus. The E1 subsequently forms a covalent thioester intermediate complex between its catalytic cysteine residue and the C terminus of the ubl. The sequential E1-mediated adenylation and covalent thioester complex formation at the C terminus of ubiquitin and ubls is reminiscent of the ThiF-mediated adenylation and covalent acyldisulfide complex formation with ThiS (Figure 1(b)).8, 10 The ThiF-catalyzed adenylation of the ThiS C terminus also resembles the adenylation reaction catalyzed at the C terminus of E. coli MoaD by MoeB, which is essential for molybdopterin biosynthesis.24, 25 ThiF and MoeB share 44% sequence identity, and are likely evolutionary ancestors of the eukaryotic E1s, with both including a Gly-X-Gly-X-X-Gly nucleotide-binding motif.26

In order to better understand the function of ThiF, we examined the nucleotide specificity of ThiF-catalyzed modification of the ThiS C terminus, and present here data supporting a strong preference for ATP over GTP, CTP or TTP. In order to gain insight into the mechanism of ThiF, we determined the crystal structures of ThiF, alone and in complex with ATP. The structures reveal similarities and differences between ThiF, MoeB, and APPBP1-UBA3 and SAE1-SAE2, the E1s for the ubls NEDD8 and SUMO, respectively.25, 27, 28 The structures shed light on the function of ThiF, and on the evolution of the common ThiF/MoeB/E1 fold.

Section snippets

Nucleotide-specificity of ThiF-mediated modification of ThiS

Previous studies have revealed that purified ThiF–ThiS complexes incubated in the presence of Mg2+ and ATP generate ThiS-COAMP and inorganic pyrophosphate.8, 11 Indeed, we monitored ThiS-COAMP formation by incubating purified glutathione-S-transferase (GST)-tagged E. coli ThiF, purified E. coli ThiS with Mg2+ and [α-32P]ATP, and observing generation of the ThiS-COAM32P product (Figure 2, lane 5). Control reactions performed lacking ThiF, ThiS, Mg2+, or [α-32P]ATP, or performed with ThiF, ThiS, Mg

Comparison to structures of E. coli MoeB, and the eukaryotic E1s for ubiquitin and ubiquitin-like proteins

In order to gain insight into the function of ThiF, we compared the structure of ThiF with the sequences, and where available, structures of E. coli MoeB, and human E1s for ubiquitin (UBE1), and the ubls NEDD8 (the APPBP1-UBA3 heterodimer), SUMO (the SAE1-SAE2 heterodimer) and ISG15 (UBE1L). ThiF and MoeB have a monomer molecular mass of ∼27 kDa and exist as homodimers, while the E1s for ubiquitin, NEDD8, ISG15, and SUMO are ∼110 kDa, and are either a single polypeptide chain, or a heterodimeric

Cloning, protein expression and protein purification

The cDNAs for ThiF and ThiS were obtained by PCR from E. coli strain DH10B cells obtained from Incyte Genomics, and subcloned into the BamHI and NotI sites of pGEX4T3 (GE Healthcare) and the BamHI and EcoRI sites of pGEX2TK (GE Healthcare), respectively. Mutations were introduced by PCR. The entire coding sequence for each construct was verified by sequencing.

GST–ThiF and GST–ThiS were expressed at 16 °C overnight in E. coli strain BL21(DE3) grown in LB induced with 0.6 mM isopropyl-β,d

Acknowledgements

We are grateful to members of the Schulman laboratory for many insightful discussions, M. Suria and C. Ross for computer support, M. Becker at the X25 beamline at NSLS, J. Holton at the 8.3.1 beamline at ALS, and J. Chrzas and S. Foundling at the SER-CAT beamline at APS for synchrotron support. This work was supported by ALSAC, the NIH (P30CA21765, R01GM69530), the Philip and Elizabeth Gross Foundation, the Beckman Foundation, the Pew Scholar's Program, and a postdoctoral fellowship from the

References (47)

  • R.K. Wierenga et al.

    Prediction of the occurrence of the ADP-binding beta alpha beta-fold in proteins, using an amino acid sequence fingerprint

    J. Mol. Biol.

    (1986)
  • C. Abad-Zapatero et al.

    Refined crystal structure of dogfish M4 apo-lactate dehydrogenase

    J. Mol. Biol.

    (1987)
  • C.U. Lenzen et al.

    Crystal structure of the hexamerization domain of N-ethylmaleimide-sensitive fusion protein

    Cell

    (1998)
  • K. Karata et al.

    Dissecting the role of a conserved motif (the second region of homology) in the AAA family of ATPases. Site-directed mutagenesis of the ATP-dependent protease FtsH

    J. Biol. Chem.

    (1999)
  • M.R. Singleton et al.

    Crystal structure of T7 gene 4 ring helicase indicates a mechanism for sequential hydrolysis of nucleotides

    Cell

    (2000)
  • H. Walden et al.

    The structure of the APPBP1-UBA3-NEDD8-ATP complex reveals the basis for selective ubiquitin-like protein activation by an E1

    Mol. Cell

    (2003)
  • Z. Otwinowski et al.

    Processing of X-ray diffraction data collected in oscillation mode

    Methods Enzymol.

    (1997)
  • I.K. McDonald et al.

    Satisfying hydrogen bonding potential in proteins

    J. Mol. Biol.

    (1994)
  • F. Jordan

    Current mechanistic understanding of thiamin diphosphate-dependent enzymatic reactions

    Nature Prod. Rep.

    (2003)
  • T.P. Begley et al.

    Thiamin biosynthesis in prokaryotes

    Arch. Microbiol.

    (1999)
  • P.B. Vander Horn et al.

    Structural genes for thiamine biosynthetic enzymes (thiCEFGH) in Escherichia coli K-12

    J. Bacteriol.

    (1993)
  • R. Kambampati et al.

    IscS is a sulfurtransferase for the in vitro biosynthesis of 4-thiouridine in Escherichia coli tRNA

    Biochemistry

    (1999)
  • J. Xi et al.

    Biosynthesis of the thiazole moiety of thiamin in Escherichia coli: identification of an acyldisulfide-linked protein–protein conjugate that is functionally analogous to the ubiquitin/E1 complex

    Proc. Natl Acad. Sci. USA

    (2001)
  • Cited by (50)

    • The ubiquitin-like modification by ThiS and ThiF in Escherichia coli

      2019, International Journal of Biological Macromolecules
      Citation Excerpt :

      volcanii, the ThiF homolog UbaA is modified at K87 by SAMP2, K113 by SAMP1/2, and K157 by SAMP1 [16,17]. ThiF C169 is essential to coordinate Zn2+, which is important for the activity of ThiF [18]. Furthermore, three candidate substrates (GroEL, FtsA, PriC) were identified that can be thisylated by ThiS/ThiF in vitro.

    • The Structure of Escherichia coli TcdA (Also Known As CsdL) Reveals a Novel Topology and Provides Insight into the tRNA Binding Surface Required for N<sup>6</sup>-Threonylcarbamoyladenosine Dehydratase Activity

      2015, Journal of Molecular Biology
      Citation Excerpt :

      AMP (or ATP) adopts an identical orientation as seen in ThiF and MoeB, which was expected from conservations of the ATP-binding residues. AMP (or ATP) is buried in a pocket with numerous networks of contact as described in previous studies [10,12]. In brief, the AMP/ATP-binding pocket is formed by a cleft pillared by β1, α3, β2 and 310A, and further interactions to AMP/ATP are mediated by multiple loop residues of β1–α3, β2–310A, β3–α5 and β4–α6 (Fig. 6).

    • Structural basis of Atg8 activation by a homodimeric E1, Atg7

      2011, Molecular Cell
      Citation Excerpt :

      Specific recognition of the appropriate Ubl and cognate E2 by E1 is critical for proper modification of target molecules. Crystallographic and biochemical studies on E1 antecedents MoeB (Lake et al., 2001) and ThiF (Duda et al., 2005), as well as on E1 enzymes for Ub (Lee and Schindelin, 2008), NEDD8 (Huang et al., 2005, 2007; Walden et al., 2003a, 2003b), and SUMO (Lois and Lima, 2005; Olsen et al., 2010) revealed the common architecture of canonical E1s and established the molecular mechanisms of the specific recognition and reactions catalyzed by E1s; these canonical E1s have a heterodimeric architecture, consisting of one active adenylation domain (AD) and one inactive AD. In addition, they possess a C-terminal ubiquitin-fold domain (UFD) and a catalytic cysteine domain composed of two half domains.

    View all citing articles on Scopus
    View full text