Search tips
Search criteria 


Logo of actafjournal home pagethis articleInternational Union of Crystallographysearchsubscribearticle submission
Acta Crystallogr Sect F Struct Biol Cryst Commun. 2012 December 1; 68(Pt 12): 1460–1464.
Published online 2012 November 19. doi:  10.1107/S1744309112045241
PMCID: PMC3509965

Structure of a periplasmic glucose-binding protein from Thermotoga maritima


ABC transport systems have been characterized in organisms ranging from bacteria to humans. In most bacterial systems, the periplasmic component is the primary determinant of specificity of the transport complex as a whole. Here, the X-ray crystal structure of a periplasmic glucose-binding protein (GBP) from Thermotoga maritima determined at 2.4 Å resolution is reported. The molecule consists of two similar α/β domains connected by a three-stranded hinge region. In the current structure, a ligand (β-d-glucose) is buried between the two domains, which have adopted a closed conformation. Details of the substrate-binding sites revealed features that determine substrate specificity. In toto, ten residues from both domains form eight hydrogen bonds to the bound sugar and four aromatic residues (two from each domain) stabilize the substrate through stacking interactions.

Keywords: glucose-binding proteins, ABC transporters, hinge motion, closed conformation

1. Introduction  

Transport of molecules across membrane lipid bilayers is central and critical to cell physiology, including uptake of nutrients, elimination of waste products, energy generation and cell signaling. ABC transporter proteins for the import or export of essential nutrients and other substances are widely distributed in nature (Higgins, 1992 [triangle]; Dean, Hamon et al., 2001 [triangle]; Dean, Rzhetsky et al., 2001 [triangle]) and form approximately 20–30% of the genes in a whole genome. The range of substrates that they transport varies from ions and small organic molecules to lipids, carbohydrates, amino acids, peptides and even whole proteins. Clinically relevant ABC exporters have been implicated in various pathologic conditions such as multi-drug resistance of cancer cells (Gottesman & Ambudkar, 2001 [triangle]; Gadsby et al., 2006 [triangle]), cystic fibrosis and Stargardt disease (Dean, Hamon et al., 2001 [triangle]; Dean, Rzhetsky et al., 2001 [triangle]). Glucose-binding proteins (GBPs) belong to a large family of soluble periplasmic proteins that support distinct biological functions in Gram-negative bacteria (Mowbray et al., 1990 [triangle]). In bacteria, this periplasmic component is the primary determinant of the specificity of the transport complex as a whole. Progression from the resting state to the transition state is driven by the engagement of substrate-loaded periplasmic binding protein rather than the binding of ATP to the Walker A and B motifs in ABC transporter proteins. Here, we report the X-ray crystal structure of a periplasmic glucose-binding protein from Thermotoga maritima bound to glucose at 2.4 Å resolution. T. maritima GBP consists of two similar structural domains connected by a three-stranded hinge region and is believed to exist in a dynamic equilibrium between closed (ligand-bound) and open (apo) conformations in solution. In our crystal structure, β-­d-glucose is bound at the domain interface, where it makes extensive hydrogen-bonding and hydrophobic interactions with both domains.

2. Materials and methods  

2.1. Protein production  

The target gene (tm0114) for NYSGXRC target 11013q was cloned using polymerase chain reaction from T. maritima genomic DNA using a forward (ACCATAGGTGTTATCGGAAA) and a reverse (CTTTTATTGGAATTCCGAGTTCTT) primer. The predicted signal peptide (1–31) was not included in cloning. The amplified gene was gel-purified and cloned into BC-pSGX3 vector designed to express the protein of interest with a C-terminal hexahistidine affinity tag. Protein expression/purification utilized previously published protocols described in detail in the PSI Knowledgebase (

2.2. Crystallization and data collection  

Initial crystallization screening of GBP was carried out with the high-throughput Index and Crystal Screens (Hampton Research) using a TECAN crystallization robot. The optimized sitting-drop vapor-diffusion crystallization involved mixing 2 µl protein solution (21 mg ml−1) with 2 µl reservoir solution consisting of 0.2 M ammonium sulfate, 30% PEG 4000 and equilibrating against 650 µl of the same reservoir solution. Good-quality crystals were flash-cooled in liquid nitrogen after transferring them into mother liquor containing 15–20%(v/v) glycerol. X-ray diffraction data were collected under standard cryogenic conditions on beamline X29A at the National Synchrotron Light Source (NSLS), Brookhaven National Laboratory. Data were integrated and scaled using HKL-2000 (Otwinowski & Minor, 1997 [triangle]). The cystals belonged to space group R3, with unit-cell parameters a = b = 123.3, c = 280.6 Å, γ = 120.0° (hexagonal setting). The calculated Matthews coefficient was 2.98 Å3 Da−1, assuming the presence of four molecules per asymmetric unit, which corresponds to an estimated solvent content of ~60%. Data-collection and refinement statistics are provided in Table 1 [triangle].

Table 1
Data-collection and refinement statistics

2.3. Crystal structure determination  

The crystal structure was determined at 2.4 Å resolution via the single-wavelength anomalous dispersion (SAD) technique using selenomethionine-substituted protein. The selenium subsubstructure was determined using SHELXD (Sheldrick, 2008 [triangle]) and revealed all 11 possible Se positions per molecule. SHARP (de La Fortelle & Bricogne, 1997 [triangle]) was used for phase refinement, and density modification was performed with SOLOMON (Abrahams & Leslie, 1996 [triangle]). The resulting phases were used for automated model building with ARP/wARP (Perrakis et al., 1999 [triangle]), which yielded ~90% of the polypeptide chain. The model-building tool O (Jones et al., 1991 [triangle]) was used to build the remaining polypeptide chain. All four molecules (A, B, C and D) in the asymmetric unit were then generated using noncrystallographic symmetry (NCS) operators derived from the Se positions. Well defined residual electron density was observed in all four molecules and could be modeled as β-d-glucose even though no ligand was added during crystallization. The atomic model was refined to convergence with the CNS (Brünger et al., 1998 [triangle]) slow-cool annealing method alternating with manual model building (final R factor and R free values of 0.21 and 0.24, respectively). The final atomic model consisted of 1204 residues, four β-d-glucose molecules and 287 water molecules. Model validation with PROCHECK (Laskowski et al., 1993 [triangle]) revealed that 89.8% of the residues were in the most favorable region and 9.4% were in the generously allowed region of the Ramachandran diagram. All figures were drawn using PyMOL (DeLano, 2002 [triangle]). Refined atomic coordinates and experimental structure factors have been deposited in the Protein Data Bank (; PDB entry 2qvc). While the refinement of this crystal structure was in progress, the crystal structure of an identical protein with bound β-d-glucose was published in a different space group (Tian et al., 2007 [triangle]). These two structures agree with an r.m.s.d. of 0.34 Å for 297 equivalent pairs of Cα atoms.

3. Results and discussion  

3.1. The overall structure  

The asymmetric unit of our T. maritima GBP crystals contained four structurally similar protomers (pairwise Cα root-mean-square deviations of 0.15–0.26 Å), which occur as monomers in solution as judged from size-exclusion chromatography (data not shown). Each protomer has approximate dimensions of 55 × 35 × 35 Å and is composed of two similar Rossmann-fold domains (pairwise r.m.s.d. of ~1.6 Å for 74 aligned Cα-atom pairs; Fig. 1 [triangle]). Residues 32–132 and 267–308 form domain 1 and residues 133–266 and 309–333 form domain 2. Each domain consists of an α/β-fold with a core of β-sheet flanked by layers of α-helices on both sides (Fig. 1 [triangle]). Domain 1 consists of a seven-stranded β-sheet (β1–5 and β11–12) with three α-helices (α1 and α9–10) on one side of the sheet and two α-helices (α2–3) on the other side. Domain 2 includes six β-strands (β6–10 and β13) with two α-helices (α4–5) lying on one side of the sheet and four α-helices (α6–8 and α11) on the opposite face.

Figure 1
Cartoon representation of the overall structure of GBP, with the bound β-d-glucose shown as a stick figure. The hinge region is indicated by an arrow and secondary-structural elements are labeled.

The domains are connected by a hinge made up of three segments (132–134, 265–267 and 308–312) of the polypeptide chain, with the ligand-binding site formed by the cleft between the domains. Sugar binding proceeds from an encounter complex with the open form of the protein, which is thought to predominate in the absence of ligand (Zou et al., 1993 [triangle]), followed by hinge bending to give the closed form observed in our cocrystal structure.

3.2. The binding cleft  

The sugar-binding site is buried within the interface between the two domains. β-d-Glucose (chair conformation) makes extensive hydrogen-bonding interactions with both domains. The location of this sugar-binding site is similar among structurally characterized periplasmic sugar-binding proteins (Zou et al., 1993 [triangle]; Borrok et al., 2007 [triangle]). The side chains of residues Lys39, Trp45, Asn167, Arg171, Asp246 and Gln266 and the main-chain N atom of Ala221 form an intricate network of hydrogen bonds with the sugar-ring hydroxyl groups and the sugar-ring O atom (Fig. 2 [triangle] a). In the binding cleft, most of the polar groups are hydrogen-bonded to the sugar molecule. In addition to the interacting polar residues, four aromatic groups (Tyr44 and Trp45 from domain 1 and Tyr220 and Tyr222 from domain 2) make hydrophobic contacts with the bound ligand (Fig. 2 [triangle] b). β-­d-­Glucose is stacked between residues Tyr44 and Tyr220, while Trp45 lies perpendicular to the Tyr44 ring. The precise configuration of the aromatic groups in the binding cleft is thought to be an important determinant of binding specificity (Chaudhuri et al., 1999 [triangle]).

Figure 2
(a) Hydrogen-bonding interactions between sugar and GBP. H atoms were positioned stereochemically. However, the hydrogen-bond distances are between the heavy atoms. (b) Sugar–aromatic interactions.

3.3. The hinge region  

Three short polypeptide-chain segments (linking the domains) together provide a functionally important hinge that support interdomain conformational changes. The major differences between the unliganded and glucose-bound structures of GPBs lie within the three hinge segments (Fig. 1 [triangle]). The first hinge segment (residues Asp308 and Asp312) connects β12 from domain 1 and β13 from domain 2. The second hinge segment (Gly265 and Arg267) links β9 from domain 2 with α9 from domain 1. Gly132 and Asp134 of the third hinge segment connect β5 from domain 1 and α4 from domain 2. Movement in the hinge region must be coordinated among these three segments and the relative domain motion is vital for correct recognition of the membrane component for transport (Björkman & Mowbray, 1998 [triangle]). HINGEprot (Emekli et al., 2008 [triangle]) analysis identified three hinge residues (Asp134, Gln266 and Gly310).

3.4. Comparison of the binding sites with similar structures  

The residues that are found in the binding cleft of GBP from T. maritima (PDB entry 2qvc) are similar to those found in GBP from Salmonella typhimurium (PDB entry 3gbp; Mowbray et al., 1990 [triangle]) and in E. coli ribose-binding protein (RBP; PDB entry 2dri; Björkman et al., 1994 [triangle]). Of the seven residues that make contact with β-d-glucose in 2qvc, six equivalent residues interact with the ribose in 2dri, while seven equivalent residues are observed to bind glucose in 3gbp (Fig. 3 [triangle]). In 2qvc, the number of hydrophobic residues forming the binding surface exceeds those observed in the other two proteins (Table 2 [triangle]). Although none of these hydrophobic interactions are conserved, the general characteristics of the binding site are very similar. A superposition of these three cocrystal structures is depicted in Fig. 4 [triangle].

Figure 3
Multiple sequence alignment of similar structures. Residues involved in hydrogen-bond interactions are marked in red in the present structure and in light blue in the comparative structures.
Figure 4
Cα superposition of 2qvc (green), 2dri (cyan) and 3gbp (magenta) showing a common closed sugar-bound conformation.
Table 2
Residues forming the binding cleft in β-D-glucose-binding protein (2qvc) and the corresponding residues in E. coli RBP (2dri) and S. typhimurium GBP (3gbp)

4. Conclusions  

The overall crystal structure of GBP bound to β-d-glucose is comparable with those of other periplasmic sugar-binding proteins from Gram-negative bacteria, all of which exhibit a closed ligand-bound conformation. The characteristic features of this protein subfamily, such as domain structure and organization, the tripartite hinge region and the location of the binding cleft within the domain interface, are all similar. Both domains participate in ligand binding in the closed liganded structures, burying the bound ligand. β-­d-­Glucose binds to T. maritima GBP in its lowest energy chair form, which has also been observed in other members of this subfamily. Hydrogen-bonding and hydrophobic interactions of the sugar with both domains of the protein stabilize the closed form. In summary, these interactions both substitute for water binding to this highly soluble class of ligands and exploit aromatic sugar stacking, which is thought to contribute to binding.

Supplementary Material

PDB reference: glucose-binding protein, 2qvc


This research was supported by a U54 award from the National Institute of General Medical Sciences to the NYSGXRC (GM074945; PI SKB) under DOE Prime Contract No. DEAC02-98CH10886 with Brookhaven National Laboratory. We gratefully acknowledge data-collection support from beamline X29 at the National Synchrotron Light Source.


  • Abrahams, J. P. & Leslie, A. G. W. (1996). Acta Cryst. D52, 30–42. [PubMed]
  • Björkman, A. J., Binnie, R. A., Zhang, H., Cole, L. B., Hermodson, M. A. & Mowbray, S. L. (1994). J. Biol. Chem. 269, 30206–30211. [PubMed]
  • Björkman, A. J. & Mowbray, S. L. (1998). J. Mol. Biol. 279, 651–664. [PubMed]
  • Borrok, M. J., Kiessling, L. L. & Forest, K. T. (2007). Protein Sci. 16, 1032–1041. [PubMed]
  • Brünger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W., Jiang, J.-S., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T. & Warren, G. L. (1998). Acta Cryst. D54, 905–921. [PubMed]
  • Chaudhuri, B. N., Ko, J., Park, C., Jones, T. A. & Mowbray, S. L. (1999). J. Mol. Biol. 286, 1519–1531. [PubMed]
  • Dean, M., Hamon, Y. & Chimini, G. (2001). J. Lipid Res. 42, 1007–1017. [PubMed]
  • Dean, M., Rzhetsky, A. & Allikmets, R. (2001). Genome Res. 11, 1156–1166. [PubMed]
  • DeLano, W. L. (2002). PyMOL
  • Emekli, U., Schneidman-Duhovny, D., Wolfson, H. J., Nussinov, R. & Haliloglu, T. (2008). Proteins, 70, 1219–1227. [PubMed]
  • Gadsby, D. C., Vergani, P. & Csanády, L. (2006). Nature (London), 440, 477–483. [PMC free article] [PubMed]
  • Gottesman, M. M. & Ambudkar, S. V. (2001). J. Bioenerg. Biomembr. 33, 453–458. [PubMed]
  • Higgins, C. F. (1992). Annu. Rev. Cell Biol. 8, 67–113. [PubMed]
  • Jones, T. A., Zou, J.-Y., Cowan, S. W. & Kjeldgaard, M. (1991). Acta Cryst. A47, 110–119. [PubMed]
  • La Fortelle, E. de & Bricogne, G. (1997). Methods Enzymol. 276, 472–493.
  • Laskowski, R. A., MacArthur, M. W., Moss, D. S. & Thornton, J. M. (1993). J. Appl. Cryst. 26, 283–291.
  • Mowbray, S. L., Smith, R. D. & Cole, L. B. (1990). Receptor, 1, 41–53. [PubMed]
  • Otwinowski, Z. & Minor, W. (1997). Methods Enzymol. 276, 307–326.
  • Perrakis, A., Morris, R. & Lamzin, V. S. (1999). Nature Struct. Biol. 6, 458–463. [PubMed]
  • Sheldrick, G. M. (2008). Acta Cryst. A64, 112–122. [PubMed]
  • Tian, Y., Cuneo, M. J., Changela, A., Höcker, B., Beese, L. S. & Hellinga, H. W. (2007). Protein Sci. 16, 2240–2250. [PubMed]
  • Zou, J., Flocco, M. M. & Mowbray, S. L. (1993). J. Mol. Biol. 233, 739–752. [PubMed]

Articles from Acta Crystallographica Section F: Structural Biology and Crystallization Communications are provided here courtesy of International Union of Crystallography