UGPase (UDP-glucose pyrophosphorylase) is highly conserved among eukaryotes. UGPase reversibly catalyses the formation of UDP-glucose and is critical in carbohydrate metabolism. Previous studies have mainly focused on the UGPases from plants, fungi and parasites, and indicate that the regulatory mechanisms responsible for the enzyme activity vary among different organisms. In the present study, the crystal structure of hUGPase (human UGPase) was determined and shown to form octamers through end-to-end and side-by-side interactions. The observed latch loop in hUGPase differs distinctly from yUGPase (yeast UGPase), which could explain why hUGPase and yUGPase possess different enzymatic activities. Mutagenesis studies showed that both dissociation of octamers and mutations of the latch loop can significantly affect the UGPase activity. Moreover, this latch effect is also evolutionarily meaningful in UGPase from different species.

INTRODUCTION

UGPase (UDP-glucose pyrophosphorylase) (EC 2.7.7.9) which is ubiquitously distributed in micro-organisms, plants and animals, plays very important roles in carbohydrate metabolism. UGPase can reversibly catalyse the formation of UDP-glucose and pyrophosphate from UTP and glucose 1-phosphate in the presence of Mg2+ [1]. UDP-glucose, the active form of glucose, plays very important roles in saccharide metabolism. Polysaccharides such as sucrose, cellulose and starch in plants, or glycogen in animals, are all synthesized from UDP-glucose [25]. UDP-glucose is also involved in the synthesis of glycoproteins, glycolipids and proteoglycans [68]. In addition, UDP-glucose is the precursor of UDP-galactose and UDP-glucuronic acid which also function as glycosyl donors [911]. As UGPase lies at the crossroads of saccharide metabolism, its regulation at different levels has been well characterized in plants [12,13] and fungi [1416].

At present, the crystal structures of apo and substrate-bound UGPase in some species have been solved, which provide insight into their oligomeric status and catalytic mechanism. Plant UGPases have been well investigated because UGPases play critical roles in cell wall synthesis [1720] and in the regulation of glucose and hypoxia states [11], which are very important for plant survival. UGPases from Arabidopsis thaliana and Hordeum vulgare were found to be only enzymatically active in the monomeric form, and polymer formation shows negligible activity because the catalytic pocket is blocked [21,22]. Therefore polymerization can be recognized as a regulatory mechanism for UGPase.

Similarly to plants, UGPase from protozoan parasites including Leishmania major [23] and Trypanosoma brucei [24] also function as monomers. The conformations of the NB loop (nucleotide-binding loop) and the SB loop (substrate-binding loop) from the UGPase of L. major change during catalysis [23]. These conformational changes are also observed in A. thaliana UGPase [21]. Since glycoproteins play very important roles in the survival and infection of parasites, structural information of UGPase from parasites would provide the basis for the design of species-specific UGPase inhibitors.

In contrast with parasite and plant UGPases, yUGPase (yeast UGPase) from Saccharomyces cerevisiae is active as a homo-octamer [25]. The C-terminal left-handed β-helices were demonstrated to be important for the formation of the yUGPase octamer. The octameric complex is considered important for effective UGPase activity regulation and recruitment of UGPase to where increased activity is required [26]. The UGPase of S. cerevisiae is essential for survival [27].

Alignment of UGPase amino acid sequences from different species shows that this enzyme is highly conserved among eukaryotes, especially in the C-terminal region. The ~50% sequence identity with yUGPase indicates that UGPase in higher organisms may have the same quaternary structure as yUGPase. Currently, there is no structural report on any mammalian UGPase. Therefore obtaining the three-dimensional structure of an animal UGPase is extremely important.

In order to provide more evidence for interpreting the molecular mechanism of animal UGPases activity, the structure of hUGPase (human UGPase) was determined in the present study. We show that hUGPase exists as octamers, which is similar to its yeast homologue. However, comparison of the crystal structures of hUGPase and yUGPase reveals obvious differences between the latch loops, which results in different enzymatic activities. Mutagenesis, enzymatic assays, AUC (analytical ultracentrifugation) and transmission electron microscope analyses confirmed further the latch effect.

EXPERIMENTAL

Plasmid construction and site-directed mutagenesis

The hUGPase isoform II gene (residues Met1–Ala11 of this isoform are lacking compared with the isoform I canonical sequence, Gene ID 48255967) was amplified using PCR primers described in Supplementary Table S1 at http://www.BiochemJ.org/bj/442/bj4420283add.htm and subcloned into pET28a (Novagen) with an N-terminal hexahistidine tag. zUGPase [zebrafish (Danio rerio) UGPase], dUGPase (Drosophila melanogaster UGPase), cUGPase (Caenorhabditis elegans UGPase), yUGPase (a gift from Dr Andreas Bracher, Max Planck Institute of Biochemistry, Martinsried, Germany) and hPGM1 (human phosphoglucomutase-1) genes were also cloned into the pET28a vector. All mutation constructs of hUGPase except N491P/L492E were made using a reverse complementary method. Mutants of N491P/L492E were constructed by directly introducing the mutation in the reverse primers. All constructed plasmids were confirmed by DNA sequencing.

Protein expression and purification

The hUGPase plasmid was transformed into Escherichia coli BL21(DE3) strain, overexpressed and purified by Ni2+-chelating chromatography followed by gel-filtration chromatography. In detail, cells were harvested and resuspended in buffer A (20 mM Tris/HCl, pH 8.0, and 200 mM NaCl). After sonication of the harvested cells, ultracentrifugation at 17800 rev./min for 60 min at 4°C using a R20A2 rotor (Hitachi, CR21GII) rotor was performed to obtain the supernatant. The protein in the supernatant was purified using an ÄKTA FPLC (GE Healthcare) with a 5 ml HiTrap™ chelating HP (GE Healthcare) column using a stepwise elution method. The supernatant was loaded on to a Ni2+-chelating column equilibrated with buffer A. The column was washed with seven column volumes of buffer B (20 mM Tris/HCl, pH 8.0, 200 mM NaCl and 100 mM imidazole). The UGPase protein was eluted further using buffer C (20 mM Tris/HCl, pH 8.0, 200 mM NaCl and 500 mM imidazole). Gel filtration with a Superdex-200 size-exclusion column (GE Healthcare) that was equilibrated with buffer A was used to further purify the protein. The purified protein was concentrated to ~10 mg/ml for subsequent experiments. All of the hUGPase mutants, zUGPase, dUGPase, cUGPase, yUGPase and hPGM1 were expressed and purified using the same method.

The polymeric states of hUGPase, yUGPase, hUGPase mutants E412D and N491P/L492E were also analysed by analytical gel-filtration chromatography using Superdex 200 10/300 GL (GE Healthcare). Standard proteins with the molecular masses of 699, 158, 43, 13.7 and 1.35 kDa were used as calibrating markers.

Crystallization and data collection

Crystals of hUGPase were grown at 20°C using the hanging-drop vapour-diffusion method. The reservoir solution contained 100 mM Hepes (pH 6.5), 5 mM MgSO4, 15% (w/v) PEG [poly(ethylene glycol)] 3350 and 20% (v/v) glycerol. Each drop contained 2 μl of the protein solution mixed with 1 μl of the reservoir solution equilibrated against 400 μl of the reservoir solution. Crystals were flash-frozen in liquid nitrogen, and X-ray diffraction data were collected on a MX225CCD detector (Rayonix) on the beamline BL17U1 at the SSRF (Shanghai Synchrotron Radiation Facility) with a wavelength of 1.0 Å (1 Å=0.1 nm). The data were then processed with HKL-2000 [28]. The crystallographic parameters and data collection statistics are presented in Table 1.

Table 1
Crystallographic data collection and refinement statistics of hUGPase

Values in parentheses refer to the outer resolution shell. Rmergehkl|I−<I>|/ΣhklI. Rwork=Σ‖Fobs|−|Fcalc‖/Σ|Fobs|. Rfree was calculated as Rwork using 5.0% of the data that were omitted from structure refinement. SSRF, Shanghai Synchrotron Radiation Facility.

Dataset hUGPase 
Data collection  
 Beamline SSRF, BL17U1 
 Wavelength (Å) 1.000 
 Resolution range (Å) 120–3.60 (3.77–3.60) 
 Space group P3121 
 Cell dimensions  
  a, b, c (Å) 140.45, 140.45, 311.72 
  α, β, γ (°) 90.00, 90.00, 120.00 
Rmerge 0.068 (0.276) 
 Completeness 93.9 (62.1) 
 Monomers/asymmetry unit 
Refinement  
 Resolution range (Å) 20–3.60 
 Reflections (test set) 38652 (1929) 
Rwork 0.248 
Rfree 0.306 
 Number of atoms 13983 
 RMSD bond (Å) 0.013 
 RMSD angles (°) 1.534 
 B-factor (Å273.7 
Ramachandran plot (%)  
 Most favoured region 84.8 
 Additionally allowed 12.6 
PDB code 3R2W 
Dataset hUGPase 
Data collection  
 Beamline SSRF, BL17U1 
 Wavelength (Å) 1.000 
 Resolution range (Å) 120–3.60 (3.77–3.60) 
 Space group P3121 
 Cell dimensions  
  a, b, c (Å) 140.45, 140.45, 311.72 
  α, β, γ (°) 90.00, 90.00, 120.00 
Rmerge 0.068 (0.276) 
 Completeness 93.9 (62.1) 
 Monomers/asymmetry unit 
Refinement  
 Resolution range (Å) 20–3.60 
 Reflections (test set) 38652 (1929) 
Rwork 0.248 
Rfree 0.306 
 Number of atoms 13983 
 RMSD bond (Å) 0.013 
 RMSD angles (°) 1.534 
 B-factor (Å273.7 
Ramachandran plot (%)  
 Most favoured region 84.8 
 Additionally allowed 12.6 
PDB code 3R2W 

Phase determination and structure refinement

Primary co-ordinates of hUGPase were obtained by the molecular replacement method. Initially, the backbone of the yUGPase end-to-end dimer structure (PDB code 2I5K) was used as the model; however, this approach revealed no possible rotation function or translation function for molecular replacement. Subsequently, the backbone of one subunit of the yUGPase structure was processed as the model. The PHASER program in the CCP4i package [29] found three, but not four, subunits in one asymmetry unit. More subunits were subsequently generated with crystal symmetry elements to fulfil the crystal lattice using PyMOL (http://www.pymol.org). This resulted in an end-to-end dimer model as reported previously for yUGPase [26] with a slight twist. This end-to-end interacting dimer was then used as a model for the Molrep-auto MR program in CCP4i [30] to obtain the four subunits' co-ordinates. After rigid body refinement, the structure was reconstructed using WinCoot 6.0 [31]. The final structure was refined with Refmac5.5 [32] using TLS (Translation/Liberation/Screw) refinement [33]. The TLS file was obtained online (http://skuld.bmsc.washington.edu/~tlsmd) [34]. Data check and validation were performed using the CCP4i package. Figures were prepared using PyMOL.

In vitro enzymatic activity assay

The enzymatic activity of hUGPase was examined in the reverse direction by a coupled enzymatic assay as reported previously [35]. A 0.1–0.6 μg amount of hUGPase, yUGPase and other mutant proteins were added into a 200 μl reaction system containing 100 mM Hepes (pH 7.5), 1 mM MgCl2, 2 units of G6PDH (glucose-6-phosphate dehydrogenase) (Sigma), 2 units of purified recombinant hPGM1, 0.5 mM NADP+, 0.75 mM pyrophosphate and UDP-glucose. Pyrophosphate was added last to start the reaction at 37°C. Concentrations of UDP-glucose were varied from 0.05 to 0.5 mM to draw double-reciprocal plots for the determination of Km and Vmax. Purified hPGM1 was confirmed to be active using glucose 1-phosphate as the substrate. The whole reaction was monitored every 0.5 s with TriStar LB941 (Berthold). NADPH fluorescence was measured to indirectly monitor the reaction process with an excitation wavelength of 355 nm and an emission wavelength at 460 nm. The data recorded were compared with the NADPH standard curve to calculate the formation velocity of NADPH. Finally, the formation velocity of glucose 1-phosphate, which is equal to that of NADPH, could be obtained. Each measurement was performed five times, and a Lineweaver–Burk curve was drawn to calculate the Michaelis–Menten constant Km and Vmax. Eadie–Hofstee and Hanes plots were also constructed using the same data. The activities of zUGPase, dUGPase, cUGPase, yUGPase and all other mutants were measured using the same method. All double-reciprocal plots are shown in Supplementary Figure S1 at http://www.BiochemJ.org/bj/442/bj4420283add.htm. All kinetic parameters are listed in Table 2.

Table 2
Kinetic parameters of hUGPase, zUGPase, dUGPase, cUGPase, yUGPase and hUGPase mutants
Protein Km (μM) Vmax (μM/min per μg) 
hUGPase 253±34 385±26 
zUGPase 183±32 368±32 
dUGPase 316±35 575±44 
cUGPase 715±39 632±52 
yUGPase 830±173 849±172 
S309N/S311R 235±38 322±23 
T406K/M407L 219±42 502±47 
K410S 151±17 988±51 
E412Q 155±16 460±25 
E412D 205±10 676±13 
E412K 67±7 85±3 
P414G/T415P 202±15 333±11 
V416N 217±27 497±30 
N491P/L492E 198±20 673±38 
hUGPase Y-latch* 338±44 570±34 
yUGPase H-latch† 396±16 237±6 
Protein Km (μM) Vmax (μM/min per μg) 
hUGPase 253±34 385±26 
zUGPase 183±32 368±32 
dUGPase 316±35 575±44 
cUGPase 715±39 632±52 
yUGPase 830±173 849±172 
S309N/S311R 235±38 322±23 
T406K/M407L 219±42 502±47 
K410S 151±17 988±51 
E412Q 155±16 460±25 
E412D 205±10 676±13 
E412K 67±7 85±3 
P414G/T415P 202±15 333±11 
V416N 217±27 497±30 
N491P/L492E 198±20 673±38 
hUGPase Y-latch* 338±44 570±34 
yUGPase H-latch† 396±16 237±6 
*

hUGPase with the latch loop of yUGPase.

yUGPase with the latch loop of hUGPase.

Transmission electron microscopy

The hUGPase, yUGPase and hUGPase mutant proteins were negatively stained using uranyl acetate on glow-discharged carbon-coated transmission electron microscope grids. Specimens were imaged at room temperature (20°C) using a Tecnai G2 20 transmission electron microscope (FEI) equipped with an LaB6 filament and operated at an acceleration voltage of 120 kV.

Analytical ultracentrifugation

The sedimentation coefficients of recombinant hUGPase, yUGPase, and hUGPase mutants E412D and N491P/L492E were analysed using a Beckman Optima XL-1 analytical ultracentrifuge with an AN50-Ti rotor. Absorbance data (72 scans at 280 nm) were collected and analysed using the SEDFIT program [36].

Multi-sequence alignment

Multi-sequence alignment was performed by using ClustalW [37] and the print picture generated at ESPript (http://espript.ibcp.fr/ESPript/ESPript) [38]. The UGPase sequences of Homo sapiens (Q16851), Bos taurus (Q07130), Mus musculus (Q91ZJ5), Gallus gallus (Q5ZKW4), Xenopus tropicalis (Q68F95), D. rerio (B8JMZ1), D. melanogaster (A5XCL5), Lepeophtheirus salmonis (C1BTD6), Brugia malayi (A8Q2J5), C. elegans (Q9XUS5), Schistosoma mansoni (C4Q2V7), Monosiga brevicollis (A9VAS8), S. cerevisiae (C7GP37), L. major (Q1PQK3), A. thaliana (Q9M9P3), Solanum tuberosum (P19595) and H. vulgare (Q43772) were all obtained from ExPASy Proteomics Server (UniProtKB) (http://www.uniprot.org).

RESULTS

Protein expression and structure determination

hUGPase, zUGPase, dUGPase, cUGPase, yUGPase, hPGM1 and other hUGPase mutants were expressed and purified in E. coli. SDS/PAGE analyses showed that the purified recombinant proteins have more than 95% purity.

The space group of the hUGPase crystal was P3121 with lattice dimensions of 140.45, 140.45 and 311.72 Å. The 3.6 Å resolution structure of hUGPase (PDB code 3R2W, Figure 1) was solved using a molecular replacement method using the structure of yUGPase (PDB code 2I5K) as the model, which has 50% sequence identity with hUGPase. Four hUGPase subunits were contained in an asymmetric unit. Most of the backbone atoms can be traced, except for a few disordered loops (residues Lys58–Trp64 and Ala194–Ala206 for subunit A; residues Glu57–Trp64 and Lys195–Ala217 for subunit B; residues Lys58–Trp64, Leu190–Trp207 and Met257–Arg264 for subunit C; and residues Lys58–Trp64 and Ala194–Glu205 for subunit D) that have poor electron density and could not be modelled. The N-terminal residues (residues Met1–Val12 for subunit A; Met1–Phe9 for subunit B; Met1–Arg14 for subunit C; and Met1–Gln10 for subunit D) and the hexahistidine tag were highly flexible and were not traced because of insufficient electron density. Subunits A, B and D had well-defined electron density, whereas subunit C had poorer electron density when compared with the other three subunits. Most of the residues were in favoured or allowed regions of the Ramachandran plot. The detailed refinement results are shown in Table 1.

Overall structure of hUGPase

Figure 1
Overall structure of hUGPase

(A) The hUGPase monomer is shown in the ribbon diagram (left) and surface diagram (right). The N-terminal domain, the catalytic domain and the C-terminal domain are indicated in blue, green and yellow respectively. The catalytic pocket is illustrated by the arrow. (B) Two kinds of interactions observed between hUGPase subunits, the end-to-end interaction (left) and the side-by-side interaction (right). The interacting parts of the two subunits are tinted in blue and magenta. (C) Side view of the octameric complex of hUGPase (left) and its surface model (right). The monomers within a building block are coloured the same. (D) Top view of the octameric complex of hUGPase (left) and its surface model (right). The monomers within a building unit are tinted in the same colour.

Figure 1
Overall structure of hUGPase

(A) The hUGPase monomer is shown in the ribbon diagram (left) and surface diagram (right). The N-terminal domain, the catalytic domain and the C-terminal domain are indicated in blue, green and yellow respectively. The catalytic pocket is illustrated by the arrow. (B) Two kinds of interactions observed between hUGPase subunits, the end-to-end interaction (left) and the side-by-side interaction (right). The interacting parts of the two subunits are tinted in blue and magenta. (C) Side view of the octameric complex of hUGPase (left) and its surface model (right). The monomers within a building block are coloured the same. (D) Top view of the octameric complex of hUGPase (left) and its surface model (right). The monomers within a building unit are tinted in the same colour.

Overall structure

Similarly to the structures of homologues of UGPases reported previously [2124,26], each subunit of hUGPase contains three domains: an N-terminal domain, a catalytic domain and a C-terminal domain (Figure 1A). The N-terminal domain is largely flexible. It contains two long α-helices (residues Ile13–Glu57) and two loops (residues Ser180–Gly211 and Met339–Ala360). This domain is thought to correlate with the regulatory function of UGPase [14]. The central catalytic domain belongs to the SGC domain family consisting of a mixed eight-stranded β-sheet flanked by α-helices that resembles a Rossmann fold, which is a characteristic structural fold observed in nucleotidyltransferases and many nucleotide-binding proteins [39]. This is the most conserved domain that shows high structural similarity to other UGPases, with an RMSD (root mean square deviation) from yUGPase of 0.867 Å for 234 Cα atoms. The C-terminal domain starts at Thr386 and contains a left handed β-helix structure, in which the last β-strand is extended and forms an end-to-end contact to the opposing hUGPase molecule in the asymmetric unit (Figure 1B). This β-helix is interrupted by a short α-helix (residues Val428–Asp445) and a small loop (residues Asn470–Arg474).

Octamer assembly of hUGPase

Multi-sequence alignment revealed that the residues in the C-terminal domain participating in the formation of the enzyme octameric complex are highly conserved (Supplementary Figure S2A at http://www.BiochemJ.org/bj/442/bj4420283add.htm). It is likely that all homologues of UGPases from higher organisms form similar octamers. As we expected, hUGPase subunits assemble into an octamer that is similar to yUGPase. Two types of interactions between subunits contribute to the formation of the octamer. The end-to-end interaction formed by β-interactions of Ile487–His497 in the C-terminus mainly contributes to the protomer formation (Figure 1B). The protomer that consists of two subunits is termed the building block. Four building blocks constitute an octamer. The side-by-side interaction is mainly because of hydrophobic interactions between two nearby subunits from different building blocks (Figure 1B). Residues on one face of the C-terminal β-helix, Gly463-Thr-Val-Ile-Ile-Ile-Ala469 and Asn485-Ile-Val-Ser-Gly490 (hydrophobic residues are highlighted in bold), interact with residues in the same face of another subunit. These two kinds of interactions have been observed previously in yUGPase [26]. Residues involved in these interactions are also highly conserved.

Previous studies have shown that UGPase from yeast exists in an active form as octamers [25,26,40]. To confirm further the oligomeric state of the recombinant hUGPase, transmission electron microscopy, gel-filtration and AUC analyses were performed. Similarly to yUGPase, negative-staining electron microscopy of hUGPase showed a four-leaf-clover-shaped structure, which is compatible with a flat tetramer or an octamer (Figure 2A). Consistently, gel-filtration analysis of the samples of hUGPase and yUGPase gave a similar elution volume (Figure 2B). Calculation of the molecular mass of the proteins revealed that the approximate molecular masses are 439 and 489 kDa for hUGPase and yUGPase respectively. Both hUGPase and yUGPase are octamers because the molecular mass of one hUGPase or yUGPase subunit is 55 kDa. Furthermore, AUC analyses revealed the same sedimentation coefficient between hUGPase and yUGPase (Table 3). These results indicate further that hUGPase forms a homo-octamer, which is consistent with the crystal structure data. In addition, gel-filtration analyses of zUGPase, dUGPase and cUGPase revealed the same elution volume (results not shown), indicating that these proteins are all octamers.

Transmission electron microscope micrograph and analytical gel-filtration chromatograph analyses of hUGPase, yUGPase, and hUGPase mutants E412D and N491P/L492E

Figure 2
Transmission electron microscope micrograph and analytical gel-filtration chromatograph analyses of hUGPase, yUGPase, and hUGPase mutants E412D and N491P/L492E

(A) All particles were negatively stained with uranyl acetate. Arrows indicate the octamer particles in hUGPase, yUGPase and mutant E412D, or the monomer in mutant N491P/L492E. (B) Elution profiles of hUGPase, yUGPase, mutants E412D and N491P/L492E compared with standard proteins that are marked with a broken line. Standard proteins used for molecular mass calibrations were 699, 158, 43, 13.7 and 1.35 kDa for peaks 1–5 respectively. The approximate molecular masses of hUGPase, yUGPase, mutant E412D and mutant N491P/L492E are calculated accordingly, which are 439, 489, 436 and 79 kDa respectively. OD280=A280. mAU, milli-absorbance units.

Figure 2
Transmission electron microscope micrograph and analytical gel-filtration chromatograph analyses of hUGPase, yUGPase, and hUGPase mutants E412D and N491P/L492E

(A) All particles were negatively stained with uranyl acetate. Arrows indicate the octamer particles in hUGPase, yUGPase and mutant E412D, or the monomer in mutant N491P/L492E. (B) Elution profiles of hUGPase, yUGPase, mutants E412D and N491P/L492E compared with standard proteins that are marked with a broken line. Standard proteins used for molecular mass calibrations were 699, 158, 43, 13.7 and 1.35 kDa for peaks 1–5 respectively. The approximate molecular masses of hUGPase, yUGPase, mutant E412D and mutant N491P/L492E are calculated accordingly, which are 439, 489, 436 and 79 kDa respectively. OD280=A280. mAU, milli-absorbance units.

Table 3
AUC results of hUGPase, yUGPase, E412D and N491P/L492E
Protein Sedimentation coefficient (S) 
hUGPase 14.3 
yUGPase 14.3 
E412D 14.3 
N491P/L492E 5.88 
Protein Sedimentation coefficient (S) 
hUGPase 14.3 
yUGPase 14.3 
E412D 14.3 
N491P/L492E 5.88 

We termed the eight subunits in the octamer A, A', B, B', C, C', D and D' (Supplementary Figure S3 at http://www.BiochemJ.org/bj/442/bj4420283add.htm). A–A' represents the end-to-end interaction which forms a building block, as does B–B', C–C' and D–D'. A–B, B'–C', C–D and D'–A' represent the side-by-side interactions. These two types of interactions organized eight hUGPase monomers to form a large particle with the C-terminal domains binding each other, and the catalytic pocket of each subunit showed no interference (Figures 1C and 1D).

Structural differences between hUGPase and yUGPase

Although the three-dimensional structures of hUGPase and yUGPase are very similar to each other (RMSD 0.937 Å for 390 Cα atoms), the enzymatic activity assays showed that hUGPase and yUGPase possess different activities. The Vmax of hUGPase and yUGPase were determined to be 385 and 849 μM/min per μg respectively (Table 2). This indicates that particular structural differences may exist between these two proteins, and such differences may influence enzyme activity. Superimposition of the two structures revealed that the building block of hUGPase is slightly twisted compared with yUGPase (Figure 3A), which could explain why the standard molecular replacement method did not work when using the yUGPase building block as the model. The RMSD between the building block of hUGPase and yUGPase is 1.544 Å over 799Cα atoms, which is higher than the RMSD determined between the human and yeast monomers.

Superimposition of the structures of hUGPase and yUGPase reveals obvious conformational differences

Figure 3
Superimposition of the structures of hUGPase and yUGPase reveals obvious conformational differences

(A) Superimposition of the building blocks of hUGPase (yellow) and yUGPase (cyan) by fixing a subunit. The arrow represents the building block twist direction of hUGPase when compared with yUGPase. (B) Superimposition of monomers of hUGPase (yellow) and yUGPase (cyan). The conformational difference between the loops including the latch loop and the 309 loop are indicated by arrows.

Figure 3
Superimposition of the structures of hUGPase and yUGPase reveals obvious conformational differences

(A) Superimposition of the building blocks of hUGPase (yellow) and yUGPase (cyan) by fixing a subunit. The arrow represents the building block twist direction of hUGPase when compared with yUGPase. (B) Superimposition of monomers of hUGPase (yellow) and yUGPase (cyan). The conformational difference between the loops including the latch loop and the 309 loop are indicated by arrows.

hUGPase showed high similarity with yUGPase in the subunit structure except for some loops such as the 309 loop and latch loop (Figure 3B). Sequence alignment also revealed diversities among these loops (Supplementary Figure S4 at http://www.BiochemJ.org/bj/442/bj4420283add.htm). Interestingly, the loop from Thr406 to Val416 showed a meaningful conformational change that positioned this region closer to the nearby subunit of its side-by-side interacting partner. This loop positions between the SB loop and the 309 loop (from Ser309 to Ser311) and forms a latch that prevents the conformational change of the 309 loop when the substrate binds. This is therefore named the ‘latch’ loop. This latch effect is not observed in yUGPase, and we postulate that this is the reason for the lower activity of hUGPase when compared with yUGPase.

The latch effect

In the present study, we tried to co-crystallize and determine the substrate-bound structure of hUGPase, but the experiments were not successful. Superimposition of hUGPase with UGPase from L. major revealed very similar structures. The RMSD between hUGPase and the apo form of L. major UGPase (PDB code 2OEF) is 1.988 Å for 364Cα atoms, and 2.834 Å over 365Cα atoms between hUGPase and the substrate-bound form of L. major UGPase (PDB code 2OEG) (Supplementary Figure S5A at http://www.BiochemJ.org/bj/442/bj4420283add.htm). Therefore the conformational changes of hUGPase occurring upon substrate binding could be inferred by comparison of the apo and substrate-bound L. major UGPase structures. Superimposition of the apo and substrate-bound UGPase structures of L. major revealed that a conformational change occurs in the SB loop upon substrate binding [21,23]. In addition, we noticed a second conformational change involving the 309 loop that is linked to the SB loop by a β-hairpin (Figure 4A). The 309 loop leans towards the substrate in the same direction as the SB loop; however, the 309 loop does not bind directly to the substrate. Similar conformational changes were observed in A. thaliana as well (PDB codes 2ICX and 2ICY, Figure S5B). Thus we deduced that the 309 loop of hUGPase would undergo similar conformational changes upon ligand binding as observed for UGPases of L. major and A. thaliana. The conformational change of the 309 loop will further influence the SB loop, which will subsequently affect enzymatic activity.

The latch effect

Figure 4
The latch effect

(A) Superimposition of apo and substrate-bound structures of L. major UGPase (PDB codes 2OEG and 2OEF). The open conformation is coloured yellow and the closed conformation is coloured cyan. The SB loop and the 309 loop are highlighted in red and blue for the closed conformation (represented as ‘C’), and deep salmon and light blue for the open conformation (represented as ‘O’) respectively. (B) Superimposition of hUGPase (green) and L. major UDP-glucose-bound UGPase (cyan, PDB code 2OEG). (C) Superimposition of hUGPase (green) and yUGPase (cyan). The latch loop and the 309 loop are highlighted. ‘H’ and ‘Y’ represent human and yeast respectively. (D) Amino acid sequence alignment of the latch loop of selected eukaryotic UGPases. Strictly conserved residues are shown in bold on a red background. The latch loop is shown in the black frame and the latch residue Glu412 is highlighted in green. These species are divided into two groups. Group 1 represents the vertebrate group and group 2 indicates the lower organisms. (E) Amino acid sequence alignment of the SB loop of selected eukaryotic UGPases. The SB loop is shown in the black frame. Strictly conserved residues are shown in bold on a red background. (F) Amino acid sequence alignment of the 309 loop of selected eukaryotic UGPases. The 309 loop is shown in the black frame.

Figure 4
The latch effect

(A) Superimposition of apo and substrate-bound structures of L. major UGPase (PDB codes 2OEG and 2OEF). The open conformation is coloured yellow and the closed conformation is coloured cyan. The SB loop and the 309 loop are highlighted in red and blue for the closed conformation (represented as ‘C’), and deep salmon and light blue for the open conformation (represented as ‘O’) respectively. (B) Superimposition of hUGPase (green) and L. major UDP-glucose-bound UGPase (cyan, PDB code 2OEG). (C) Superimposition of hUGPase (green) and yUGPase (cyan). The latch loop and the 309 loop are highlighted. ‘H’ and ‘Y’ represent human and yeast respectively. (D) Amino acid sequence alignment of the latch loop of selected eukaryotic UGPases. Strictly conserved residues are shown in bold on a red background. The latch loop is shown in the black frame and the latch residue Glu412 is highlighted in green. These species are divided into two groups. Group 1 represents the vertebrate group and group 2 indicates the lower organisms. (E) Amino acid sequence alignment of the SB loop of selected eukaryotic UGPases. The SB loop is shown in the black frame. Strictly conserved residues are shown in bold on a red background. (F) Amino acid sequence alignment of the 309 loop of selected eukaryotic UGPases. The 309 loop is shown in the black frame.

To examine whether mutations in the 309 loop affected the enzyme activity, we constructed a mutant S309N/S311R of hUGPase in which Ser309 was mutated to asparagine and Ser311 to arginine in the 309 loop to mimic the 309 loop of yUGPase. This mutant revealed ~84% activity when compared with the activity of the wild-type protein (Table 2), indicating that the 309 loop does not play a critical function in influencing UGPase activity.

In monomeric plant and protozoan UGPases, the 309 loop movement does not interfere with or appear to have a role in regulating substrate binding. In octameric yUGPase, residues of the 309 loop show no interaction to the latch loop (Figure 4C). Whereas, in hUGPase, the latch loop locates between the SB loop and the 309 loop of its nearby subunit from different building blocks and will therefore further prevent the 309 loop approaching the substrate. This latch effect synchronously interferes with the movement of the 309 loop and the SB loop (Figures 4B and 4C), and subsequently interferes with enzymatic activity. The plug residue Glu412 of hUGPase on the latch loop seems to function as the ‘latch’.

To determine further whether the latch loop will affect enzyme activity, we exchanged the latch loops of yUGPase and hUGPase. The latch loop of hUGPase (Thr406-Met-Ser-Glu-Lys-Arg-Glu-Phe-Pro-Thr-Val416) was mutated to the latch loop of yUGPase (Lys409-Leu-Asp-Pro-Ser-Arg-Phe-Gly-Pro-Asn418), and the latch loop of yUGPase was replaced by that of hUGPase, and then the enzyme activities were examined. The results showed that, compared with the wild-type hUGPase, replacement of the latch loop of hUGPase with that of yUGPase resulted in a higher enzymatic activity. This is due to the elimination of the latch effect. In contrast, substitution of the latch loop of yUGPase with hUGPase's latch loop led to a significant decrease in the enzyme activity in comparison with the wild-type yUGPase (Table 2). These results suggest that the latch effect does affect UGPase enzymatic activity.

Identification of the residues important for the latch effect

To identify the residues that are important for the latch effect of hUGPase, a detailed comparison of the sequence around the latch loop was performed. The result showed that Glu412 is highly conserved in vertebrates, and is a different amino acid in lower organisms and is absent from S. cerevisiae (Figure 4D). Using the structural analysis presented above, a series of Glu412 mutants were constructed and examined. Although the oligomerization status of these mutants did not change (Figure 2 and Table 3), their enzymatic activities showed differences (Table 2). Mutant E412D, with a shorter side chain, showed weaker inhibition on the 309 loop, and therefore significantly promoted hUGPase activity to 176% when normalized against the wild-type hUGPase activity (Table 2). This mutant may not eliminate the latch effect completely, because the E412D mutation does not significantly change the side-chain structure. As such the activity of this mutant is not as high as yUGPase. Substitution of glycine for Glu412 (E412Q) changed the charge property, but not the length of side chain and showed only a marginal increase in activity (19%) when compared with the wild-type protein. In contrast, mutant E412K that has a longer side chain with a reverse in charge showed obvious inhibitory effects which gave only 22% activity when compared with the wild-type hUGPase. Thus we can deduce that this latch effect is mainly due to steric hindrance, but not electrostatic interactions or hydrogen-bonding. The result that mutation on the 309 loop did not obviously affect hUGPase activity also indicates that the latch effect does not rely on electrostatic interactions or hydrogen-bonding.

Several factors may stabilize the conformation of the latch loop and assure its latch effect in hUGPase. One is within the latch loop where an N–O hydrogen bond is formed between Arg411 and Thr415 (Supplementary Figure S6 at http://www.BiochemJ.org/bj/442/bj4420283add.htm), This results in the three residues between positions 411 and 415 (Glu412, Phe413 and Pro414) to be fixed to a relatively stable conformation so that they can function as a latch. In addition, the negatively charged side chain of Glu412 just lies in a cavity harbouring positive charge. Also, Phe413 is located in the hydrophobic region between the SB loop and the 309 loop. These factors affect the conformation of the latch loop moderately, therefore the latch effect only reduces, but does not eliminate, catalytic activity.

Mutations were also introduced for residues near to Glu412 because these residues on the latch loop vary among different organisms and may influence the loop direction and further affect enzymatic activity. Mutant K410S significantly increased the activity to 257% when compared with the wild-type hUGPase, making the activity of this mutant similar to yUGPase (Table 2). The reason for this increase in activity is still not clear. Because mutant K410S can strongly change the residue charge property, we suspected that this mutation may influence the conformation of Arg411 and affect further the hydrogen bond between Arg411 and Thr415 (Supplementary Figure S6). This will affect the orientation of the latch loop and eliminate the blocking of the latch loop towards the 309 loop. The other mutants, including T406K/M407L, P414G/T415P and V416N, showed activity similar to that of the wild-type hUGPase (Table 2), indicating that these mutations do not dramatically affect the loop direction and leave the latch loop unchanged. Arg411 forms an N–O hydrogen bound with the main-chain oxygen atom of Thr415 (Supplementary Figure S6), which could explain why mutant P414G/T415P does not affect catalytic activity very much.

Gel-filtration analysis revealed that all of the mutants are octamers, suggesting that the mutations do not change the oligomeric state.

The latch effect is evolutionarily meaningful

The UGPases from different species can be divided into two groups according to the multi-sequence alignment of their latch loops (Figure 4D). One is the vertebrate group including UGPases from H. sapiens, M. musculus, X. tropicalis and D. rerio, in which the latch residue at the 412 position is a glutamic acid residue that is conserved and critical to the latch effect, as shown in hUGPase. The other is the lower organisms group that includes UGPases from D. melanogaster, C. elegans and S. cerevisiae. In this group, the sequences of the latch loop are more diverse and different from that of the vertebrates, especially at position 412 where the residue is replaced by other residues such as methionine or serine. As shown above, different latch loops in hUGPase and yUGPase resulted in different catalytic activities. We speculated that the enzymatic activity of UGPase from the two groups is different. To prove this speculation, a species of UGPases from different evolutionary levels such as S. cerevisiae (fungi), C. elegans (nematode), D. melanogaster (insect), D. rerio (vertebrate) and H. sapiens (vertebrate) were selected, and their UGPase activities were examined and compared. The Vmax values were 385, 368, 575, 632 and 849 μM/min per μg for hUGPase, zUGPase, dUGPase, cUGPase and yUGPase respectively (Table 2). As expected, the vertebrate group possesses lower catalytic activity because of the latch effect. In contrast, the UGPases of the lower organisms group show higher activities owing to the absence of the latch effect. This result indicates that the latch loop is evolutionarily meaningful.

The amino acid sequences of the SB loop and the 309 loop from different organisms were also analysed, and the results revealed that the SB loop is conserved in all of the species (Figure 4E). Similarly, the 309 loop showed less divergence among species in comparison with the latch loop (Figure 4F). These results indicate that the SB loop and the 309 loop are highly conserved among species and may not correlate with the differences in UGPase activity between the lower organisms group and the vertebrate group. This was confirmed by the 309 loop mutation experiment as shown in Table 2.

Disruption of the octamerization of hUGPase increases the pyrophosphorylase activity

Previous studies have shown that deoligomerization of plant UGPases increase enzyme activity significantly [21,22]. Structural analysis revealed that Asn491 and Leu492 are involved in the end-to-end interaction (Supplementary Figure S2B). Multi-sequence alignment of the C-terminal sequences of UGPases from plants to animals also indicated that these two residues are different between the monomeric group and octameric group, and therefore these two residues are thought to be critical for octamer formation (Supplementary Figure S2A). To investigate whether deoligomerization of hUGPases has a similar effect on enzyme activity, mutant N491P/L492E was constructed to depolymerize hUGPase octamers. Enzymatic activity assays showed that, compared with the wild-type UGPase (Vmax of 385 μM/min per μg), mutant N491P/L492E showed a higher activity with a Vmax of 673 μM/min per μg (Table 2). To determine whether the increase in UGPase activity resulted from deoligomerization of the protein, the oligomeric state of mutant N491P/L492E was examined by gel-filtration chromatography, AUC analysis and negative-staining electron microscopy. As expected, gel-filtration analysis revealed that the molecular mass of the mutant protein is much lower than the wild-type hUGPase (Figure 2B). The calculated molecular mass of mutant N491P/L492E is approximately 79 kDa which is close to the molecular mass of the UGPase subunit. Disruption of the octamerization of hUGPase was confirmed further by AUC analysis (Table 3). The result showed that the sedimentation coefficient of mutant N491P/L492E was 5.88, which is much smaller than the value for the wild-type hUGPase (14.3). Consistently, electron microscopy analyses showed that the four-leaf-clover-shaped structure disappeared, and, instead, monomeric particles were present in the N491P/L492E sample (Figure 2A). These experiments indicate that substitution of Asn491 and Leu492 in the C-terminal left-handed β-helix changed the oligomerization state of hUGPase and affected the enzyme activity.

DISCUSSION

In order to interpret the catalytic mechanism of mammalian UGPases, we have solved the crystal structure of human UGPase. In contrast with the monomeric plant and protozoan UGPases, hUGPase forms an octameric complex, which is similar to yUGPase. However, enzymatic analysis showed that hUGPase possesses only ~45% activity when compared with the activity of yUGPase, suggesting a different molecular basis for the UGPase from human and yeast even though they share 50% sequence identity, similar structural fold and identical oligomeric states.

The most significant discovery of the present study is the identification of the unique latch effect in hUGPase. Besides sequence differences (Figure 4D), comparison of the structures reveals an obvious conformational difference between the latch loop of hUGPase and yUGPase (Figures 4B and 4C). Glu412 in the latch loop of hUGPase latches directly into the interspaces of the SB loop and the 309 loop of its side-by-side interacting partner subunit. Further inspection of the structures of apo and substrate-bound UGPase from A. thaliana and L. major reveals that the SB loop undergoes a conformational change by moving towards the glucose moiety when UDP-glucose binds to the catalytic pocket. In addition to the observed conformational change of the SB loop, we also found that the 309 loop, which is linked to the SB loop by a β-hairpin, changed its position as the SB loop binds to the substrate. This has not been mentioned in previous papers [21,23].

A series of mutants for residues located in the latch loop were constructed to verify whether changes of the latch loop really affected enzymatic activity. We found that although the polymerization states of these mutants did not differ from the wild-type hUGPase, mutants that either changed the length of the side chain of the latch (E412D and E412K) or strongly changed the charge property of the residue on the loop close to the latch (K410S) affected the hUGPase activity significantly. These mutations influence the blocking effect of the latch loop on the 309 loop, and therefore affect enzymatic activity (Table 2). Our results indicate that the latch effect is mainly due to steric hindrance, but not electrostatic interactions or hydrogen-bonding. Moreover, depolymerization of hUGPase (N491P/L492E) resulted in monomers and higher enzymatic activity (Figure 2 and Tables 2 and 3). This observation also suggests that the blocking effect of the latch loop on the 309 loop had been removed because the neighbouring subunit to the latch loop did not exist any more. These studies provided clear evidence for the effect of the latch loop on the enzymatic activity in hUGPase, which has not been observed for the octameric yUGPase [26]. In the case of yUGPase, no inhibition of the 309 loop occurs and therefore higher activity is shown. Replacement of the latch loop of hUGPase with that of yUGPase results in an increase in the UGPase activity, whereas substitution of the latch loop of hUGPase for the yUGPase latch loop leads to a decrease in enzymatic activity. This confirms further that the latch effect of hUGPase functions importantly in UGPase activity.

Multi-sequence alignment revealed that all vertebrates, including mammals, birds, amphibians and fish, share almost the same sequence in the latch loop (Figure 4D), indicating that they may all possess the same latch effect as that of hUGPase and may have similar activity to hUGPase. In invertebrates and simpler organisms, the amino acid sequences of the latch loop are diverse, suggesting that the UGPase from lower organisms may have no latch effect and therefore possesses higher enzymatic activity. As expected, our enzymatic analyses showed that the UGPase from simpler organisms such as yUGPase, cUGPase and dUGPase has higher enzymatic activity than hUGPase and zUGPase (Table 2). Therefore the difference in the latch loops can explain the different activities between lower and higher organisms. According to our results and previous reports, although further studies are needed, we hypothesize that, in lower organisms, UGPases function in the monomeric or oligomeric states without the latch effect because of the large amount of UDP-glucose required for survival. In contrast, the octameric vertebrate UGPases use the latch effect to control their activities and satisfy the lower requirement for UDP-glucose.

Concluding remarks

In the present study, we have revealed a 3.6 Å resolution crystal structure of human UGPase. An unprecedented latch effect is observed and confirmed, which could explain the activity differences among hUGPase, yUGPase and a hUGPase depolymerization mutant. This latch effect was also proven to be evolutionarily meaningful.

Abbreviations

     
  • AUC

    analytical ultracentrifugation

  •  
  • hPGM1

    human phosphoglucomutase-1

  •  
  • RMSD

    root mean square deviation

  •  
  • SB loop

    substrate-binding loop

  •  
  • UGPase

    UDP-glucose pyrophosphorylase

  •  
  • cUGPase

    Caenorhabditis elegans UGPase

  •  
  • dUGPase

    Drosophila melanogaster UGPase

  •  
  • hUGPase

    human UGPase

  •  
  • yUGPase

    yeast UGPase

  •  
  • zUGPase

    zebrafish UGPase

AUTHOR CONTRIBUTION

Quan Yu and Xiaofeng Zheng designed experiments and wrote the paper. Quan Yu performed experiments.

We thank Professor Ming Luo and Professor Yicheng Dong for helpful discussions. Data were collected at the Shanghai Synchrotron Radiation Facility. We give sincere thanks for the help of Dr Ping Zhu for transmission electron microscopy experiments. We thank Dr Andreas Bracher for providing us with the yeast UGPase plasmid.

FUNDING

This work was supported by the National High Technology and Development Program of China 973 Program [grant numbers 2010CB911800 and 2007CB914303] and the 863 Program [grant number 2006AA02A314] and the International Centre for Genetic Engineering and Biotechnology (ICGEB) [project number CRP/CHN09-01].

References

References
1
Aksamit
R. R.
Ebner
K. E.
Purification, properties and kinetic analysis of UDP-glucose pyrophosphorylase from bovine mammary tissue
Biochim. Biophys. Acta
1972
, vol. 
268
 (pg. 
102
-
112
)
2
Alonso
M.
Lomako
J.
Lomako
W.
Whelan
W.
A new look at the biogenesis of glycogen
FASEB J.
1995
, vol. 
9
 (pg. 
1126
-
1137
)
3
Gibeaut
D. M.
Nucleotide sugars and glycosyltransferases for synthesis of cell wall matrix polysaccharides
Plant Physiol. Biochem.
2000
, vol. 
38
 (pg. 
69
-
80
)
4
Kleczkowski
L. A.
Glucose activation and metabolism through UDP-glucose pyrophosphorylase in plants
Phytochemistry
1994
, vol. 
37
 (pg. 
1507
-
1515
)
5
Hellerstein
M. K.
Letscher
A.
Schwarz
J.
Cesar
D.
Shackleton
C.
Turner
S.
Neese
R.
Wu
K.
Bock
S.
Kaempfer
S.
Measurement of hepatic Ra UDP-glucose in vivo in rats: relation to glycogen deposition and labeling patterns
Am. J. Physiol. Endocrinol. Metab.
1997
, vol. 
272
 (pg. 
E155
-
E162
)
6
Flores-Díaz
M.
Alape-Giron
A.
Persson
B.
Pollesello
P.
Moos
M.
von Eichel-Streiber
C.
Thelestam
M.
Florin
I.
Cellular UDP-glucose deficiency caused by a single point mutation in the UDP-glucose pyrophosphorylase gene
J. Biol. Chem.
1997
, vol. 
272
 (pg. 
23784
-
23791
)
7
Sandhoff
K.
Van Echten
G.
Schroder
M.
Schnabel
D.
Suzuki
K.
Metabolism of glycolipids: the role of glycolipid-binding proteins in the function and pathobiochemistry of lysosomes
Biochem. Soc. Trans.
1992
, vol. 
20
 (pg. 
695
-
699
)
8
Silbert
J. E.
Sugumaran
G.
Intracellular membranes in the synthesis, transport, and metabolism of proteoglycans
Biochim. Biophys. Acta
1995
, vol. 
1241
 (pg. 
371
-
384
)
9
Frey
P. A.
The Leloir pathway: a mechanistic imperative for three enzymes to change the stereochemical configuration of a single carbon in galactose
FASEB J.
1996
, vol. 
10
 (pg. 
461
-
470
)
10
Aw
T. Y.
Jones
D. P.
Direct determination of UDP-glucuronic acid in cell extracts by high-performance liquid chromatography
Anal. Biochem.
1982
, vol. 
127
 (pg. 
32
-
36
)
11
Aw
T. Y.
Jones
D. P.
Control of glucuronidation during hypoxia: limitation by UDP-glucose pyrophosphorylase
Biochem. J.
1984
, vol. 
219
 (pg. 
707
-
712
)
12
Ciereszko
I.
Johansson
H.
Hurry
V.
Kleczkowski
L. A.
Phosphate status affects the gene expression, protein content and enzymatic activity of UDP-glucose pyrophosphorylase in wild-type and pho mutants of Arabidopsis
Planta
2001
, vol. 
212
 (pg. 
598
-
605
)
13
Ciereszko
I.
Johansson
H.
Kleczkowski
L. A.
Sucrose and light regulation of a cold-inducible UDP-glucose pyrophosphorylase gene via a hexokinase-independent and abscisic acid-insensitive pathway in Arabidopsis
Biochem. J.
2001
, vol. 
354
 (pg. 
67
-
72
)
14
Rutter
J.
Probst
B. L.
McKnight
S. L.
Coordinate regulation of sugar flux and translation by PAS kinase
Cell
2002
, vol. 
111
 (pg. 
17
-
28
)
15
Nishizawa
M.
Tanabe
M.
Yabuki
N.
Kitada
K.
Toh-e
A.
Pho85 kinase, a yeast cyclin-dependent kinase, regulates the expression of UGP1 encoding UDP-glucose pyrophosphorylase
Yeast
2001
, vol. 
18
 (pg. 
239
-
249
)
16
Hedegaard
J.
Horn
P.
Lametsch
R.
Søndergaard Møller
H.
Roepstorff
P.
Bendixen
C.
Bendixen
E.
UDP-glucose pyrophosphorylase is upregulated in carriers of the porcine RN− mutation in the AMP-activated protein kinase
Proteomics
2004
, vol. 
4
 (pg. 
2448
-
2454
)
17
Worrell
A. C.
Bruneau
J. M.
Summerfelt
K.
Boersig
M.
Voelker
T. A.
Expression of a maize sucrose phosphate synthase in tomato alters leaf carbohydrate partitioning
Plant Cell
1991
, vol. 
3
 (pg. 
1121
-
1130
)
18
Huber
S. C.
Huber
J. L.
Role of sucrose-phosphate synthase in sucrose metabolism in leaves
Plant Physiol.
1992
, vol. 
99
 (pg. 
1275
-
1278
)
19
Aloni
Y.
Delmer
D. P.
Benziman
M.
Achievement of high rates of in vitro synthesis of 1,4-β-D-glucan: activation by cooperative interaction of the Acetobacter xylinum enzyme system with GTP, polyethylene glycol, and a protein factor
Proc. Natl. Acad. Sci. U.S.A.
1982
, vol. 
79
 (pg. 
6448
-
6452
)
20
Ghangas
G. S.
Steffens
J. C.
UDPglucose: fatty acid transglucosylation and transacylation in triacylglucose biosynthesis
Proc. Natl. Acad. Sci. U.S.A.
1993
, vol. 
90
 (pg. 
9911
-
9915
)
21
McCoy
J. G.
Bitto
E.
Bingman
C. A.
Wesenberg
G. E.
Bannen
R. M.
Kondrashov
D. A.
Phillips
G. N.
Jr
Structure and dynamics of UDP-glucose pyrophosphorylase from Arabidopsis thaliana with bound UDP-glucose and UTP
J. Mol. Biol.
2007
, vol. 
366
 (pg. 
830
-
841
)
22
Meng
M.
Fitzek
E.
Gajowniczek
A.
Wilczynska
M.
Kleczkowski
L. A.
Domain-specific determinants of catalysis/substrate binding and the oligomerization status of barley UDP-glucose pyrophosphorylase
Biochim. Biophys. Acta
2009
, vol. 
1794
 (pg. 
1734
-
1742
)
23
Steiner
T.
Lamerz
A. C.
Hess
P.
Breithaupt
C.
Krapp
S.
Bourenkov
G.
Huber
R.
Gerardy-Schahn
R.
Jacob
U.
Open and closed structures of the UDP-glucose pyrophosphorylase from Leishmania major
J. Biol. Chem.
2007
, vol. 
282
 (pg. 
13003
-
13010
)
24
Marino
K.
Guther
M. L.
Wernimont
A. K.
Amani
M.
Hui
R.
Ferguson
M. A.
Identification, subcellular localization, biochemical properties, and high-resolution crystal structure of Trypanosoma brucei UDP-glucose pyrophosphorylase
Glycobiology
2010
, vol. 
20
 (pg. 
1619
-
1630
)
25
Turnquist
R. L.
Gillett
T. A.
Hansen
R.
Uridine diphosphate glucose pyrophosphorylase: crystallization and properties of the enzyme from rabbit liver and species comparisons
J. Biol. Chem.
1974
, vol. 
249
 (pg. 
7695
-
7700
)
26
Roeben
A.
Plitzko
J. M.
Korner
R.
Bottcher
U. M.
Siegers
K.
Hayer-Hartl
M.
Bracher
A.
Structural basis for subunit assembly in UDP-glucose pyrophosphorylase from Saccharomyces cerevisiae
J. Mol. Biol.
2006
, vol. 
364
 (pg. 
551
-
560
)
27
Daran
J. M.
Dallies
N.
Thines-Sempoux
D.
Paquet
V.
Francois
J.
Genetic and biochemical characterization of the UGP1 gene encoding the UDP-glucose pyrophosphorylase from Saccharomyces cerevisiae
Eur. J. Biochem.
1995
, vol. 
233
 (pg. 
520
-
530
)
28
Otwinowski
Z.
Minor
W.
Processing of X-ray diffraction data collected in oscillation mode
Methods Enzymol.
1997
, vol. 
276
 (pg. 
307
-
326
)
29
McCoy
A. J.
Grosse-Kunstleve
R. W.
Adams
P. D.
Winn
M. D.
Storoni
L. C.
Read
R. J.
Phaser crystallographic software
J. Appl. Crystallogr.
2007
, vol. 
40
 (pg. 
658
-
674
)
30
Vagin
A.
Teplyakov
A.
MOLREP: an automated program for molecular replacement
J. Appl. Crystallogr.
1997
, vol. 
30
 (pg. 
1022
-
1025
)
31
Emsley
P.
Lohkamp
B.
Scott
W.
Cowtan
K.
Features and development of Coot
Acta Crystallogr. Sect. D Biol. Crystallogr.
2010
, vol. 
66
 (pg. 
486
-
501
)
32
Murshudov
G.
Vagin
A. A.
Dodson
E. J.
Refinement of macromolecular structures by the maximum-likelihood method
Acta Crystallogr. Sect. D Biol. Crystallogr.
1997
, vol. 
53
 (pg. 
240
-
255
)
33
Painter
J.
Merritt
E. A.
Optimal description of a protein structure in terms of multiple groups undergoing TLS motion
Acta Crystallogr. Sect. D Biol. Crystallogr.
2006
, vol. 
62
 (pg. 
439
-
450
)
34
Painter
J.
Merritt
E. A.
TLSMD web server for the generation of multi-group TLS models
J. Appl. Crystallogr.
2006
, vol. 
39
 (pg. 
109
-
111
)
35
Tsuboi
K.
Fukunaga
K.
Petricciani
J.
Purification and specific kinetic properties of erythrocyte uridine diphosphate glucose pyrophosphorylase
J. Biol. Chem.
1969
, vol. 
244
 (pg. 
1008
-
1015
)
36
Schuck
P.
A model for sedimentation in inhomogeneous media. I. Dynamic density gradients from sedimenting co-solutes
Biophys. Chem.
2004
, vol. 
108
 (pg. 
187
-
200
)
37
Higgins
D. G.
Thompson
J. D.
Gibson
T. J.
Using CLUSTAL for multiple sequence alignments
Methods Enzymol.
1996
, vol. 
266
 (pg. 
383
-
402
)
38
Gouet
P.
Courcelle
E.
Stuart
D. I.
ESPript: analysis of multiple sequence alignments in PostScript
Bioinformatics
1999
, vol. 
15
 (pg. 
305
-
308
)
39
Unligil
U. M.
Rini
J. M.
Glycosyltransferase structure and mechanism
Curr. Opin. Struct. Biol.
2000
, vol. 
10
 (pg. 
510
-
517
)
40
Keppler
D.
Decker
K.
Studies on the mechanism of galactosamine-1-phosphate and its inhibition of UDP-glucose pyrophosphorylase
Eur. J. Biochem.
1969
, vol. 
10
 (pg. 
219
-
225
)

Author notes

The atomic co-ordinates and structural factors of human UDP-glucose pyrophosphorylase have been deposited in the RCSB PDB under accession code 3R2W.

Supplementary data