Substrate spectrum of l -rhamnulose kinase related to models derived from two ternary complex structures

Substrate spectrum of l -rhamnulose kinase related to models derived from two ternary complex structures

FEBS Letters 581 (2007) 3127–3130 Substrate spectrum of L -rhamnulose kinase related to models derived from two ternary complex structures Dirk Gruen...

821KB Sizes 0 Downloads 16 Views

FEBS Letters 581 (2007) 3127–3130

Substrate spectrum of L -rhamnulose kinase related to models derived from two ternary complex structures Dirk Grueninger, Georg E. Schulz* Institut fu¨r Organische Chemie und Biochemie, Albert-Ludwigs-Universita¨t, Albertstrasse 21, 79104 Freiburg im Breisgau, Germany Received 18 April 2007; revised 29 May 2007; accepted 29 May 2007 Available online 6 June 2007 Edited by Hans Eklund

Abstract The enzyme L -rhamnulose kinase from Escherichia coli participates in the degradation pathway of L -rhamnose, a common natural deoxy-hexose. The structure of the enzyme in a ternary complex with its substrates ADP and L -rhamnulose ˚ resolution and refined to has been determined at 1.55 A Rcryst/Rfree values of 0.179/0.209. The result was compared with the lower resolution structure of a corresponding complex containing L -fructose instead of L -rhamnulose. In light of the two established sugar positions and conformations, a number of rare sugars have been modeled into the active center of L -rhamnulose kinase and the model structures have been compared with the known enzymatic phosphorylation rates. Rare sugars are of rising interest for the synthesis of bioactive compounds.  2007 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. Keywords: Epimer tolerance; Ligand fitting; Phosphoryl transfer; Rare 2-ketoses; X-ray diffraction

1. Introduction In numerous plants the sugar L -rhamnose is a frequent cell wall constituent [1] that is metabolized by various bacteria, among them Escherichia coli [2,3]. In E. coli the respective operon is composed of L -rhamnose permease, L -rhamnose isomerase, L rhamnulose kinase (RhuK, EC 2.7.1.5), L -rhamnulose-1-phosphate aldolase and two regulatory proteins [4]. Here, we are concerned with the kinase that uses ATP to phosphorylate the sugar. The reaction results in L -rhamnulose-1-phosphate, which is then split by the aldolase into L -lactaldehyde and dihydroxyacetone phosphate [5]. RhuK belongs to the hexokinase-actin-hsp70 family exhibiting the characteristic b-sheet in an N-terminal as well as a C-terminal domain and the ATP binding site in a cleft between the two domains [6]. The similarity is significant indicating a common evolutionary origin [7]. A comparison between holo and apo structures of RhuK showed a large induced-fit during substrate binding involving a relative rotation of about 10 between the N- and C-terminal domains [6]. Kinases require such a conformational change to exclude water, which would otherwise compete for the phosphoryl group. An induced-fit was initially observed for hexokinase [8]. For the ATP hydrolases actin and hsp70, a corresponding rotation on phosphoryl transfer serves as a conformational signal [9] or as a power stroke [10]. * Corresponding author. Fax: +49 761 203 6161. E-mail address: [email protected] (G.E. Schulz).

Abbreviations: RhuK, L -rhamnulose kinase; E. coli, Escherichia coli

2. Materials and methods The substrate L -rhamnulose-1-phosphate was produced by an aldol condensation of dihydroxyacetone phosphate and lactaldehyde using L -rhamnulose-1-phosphate aldolase [11,12]. The reaction mixture contained 2.4 mmol dihydroxyacetone phosphate and 6 mmol D /L -lactaldehyde in 180 ml buffer of 20 mM Tris–HCl, pH 7.2. The reaction was started by adding 1.7 lmol (50 mg) of the aldolase and ran at 37 C for 2 h. The protein was removed by denaturation (95 C, 5 min) and centrifugation. The solution was concentrated and run through a silicagel column (Silicagel 60, Merck Darmstadt) using a 2:1 (v/v) mixture of iso-propanol and saturated NH3 in water. The product L -rhamnulose-1-phosphate was identified by thin-layer chromatography, confirmed by 1H and 31P NMR, separated and dried using a rotary evaporator. Since the charge removal at three surface residues in RhuK mutant E69A-E70A-R73A gave rise to well-ordered crystals [6], we kept these mutations in the reported analysis. The triple mutant was overexpressed in E. coli JM105 and purified as described [6]. The protein was dialyzed against 10 mM b-mercaptoethanol and concentrated to 20 mg/ ml. Co-crystallization was carried out at 20 C using the hanging drop method. The drops consisted of 2 ll of a solution of 12 mg/ml protein, 20 mM ADP and 20 mM L -rhamnulose-1-phosphate (unbuffered) and 2 ll of the reservoir containing 17% (w/v) PEG 8000 and 120 mM LiCl (unbuffered). Crystals grew to maximum sizes of about 300 · 150 · 100 lm3 within three days. For X-ray analysis, the crystals were transferred to 34% (w/v) PEG 8000 with 120 mM LiCl and shockfrozen in a nitrogen gas stream at 100 K. X-ray diffraction data were collected at the Swiss Light Source (SLS, Villigen, Switzerland) and processed with XDS [13]. Using REFMAC5 [14] an initial model was obtained by rigid body refinement of the established RhuK structure (PDB code 2CGL). The substrate L rhamnulose was generated with PRODRG2 [15] and introduced into the model. Water molecules were added with ARP/wARP [16]. The model was improved with COOT [17] and refined with REFMAC5. Models of the rare sugars were produced with PRODRG2 [15] starting from the furanose conformation observed for L -fructose [6] and L rhamnulose bound to RhuK. They were manually placed using COOT [17]. The figures were prepared using POVscript [18] and POVRAY (http://www.povray.org). The coordinates and structure factors are deposited in the Protein Data Bank accession code 2UYT.

3. Results and discussion RhuK is a monomer with a mass of 54 110 Da comprising two approximately equal sized domains, each containing a characteristic five-stranded b-sheet covered on both sides by helices [6]. The known structure of residues 2–480 and ADP was placed into the crystal unit cell. After adding b-L -rhamnulose and water mole˚ resolution yielding good cules, the model was refined to 1.55 A quality indices (Table 1). Given the high resolution, 14 residues were observed in double conformations. None of them was in the active center. ADP and the natural substrate L -rhamnulose were well outlined by the electron density. Since the average B-factors of both substrates matched that of the surrounding protein,

0014-5793/$32.00  2007 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2007.05.075

3128

D. Grueninger, G.E. Schulz / FEBS Letters 581 (2007) 3127–3130

Table 1 Structure determinationa Data collection ˚) Resolution range (A Unique reflections Multiplicity Completeness RsymI (%) Average I/r

79–1.55 (1.62–1.55) 60 401 (6646) 6.9 (5.5) 98.6 (89.4) 5.1 (37) 23.2 (4.7)

Refinement Protein atoms Ligand atoms Solvent atoms Rcryst/Rfree (5% test set) ˚ 2) B-factor average total/water (A ˚ )/bond angles () Rmsd bond lengths (A Ramachandran: favored/allowed/disallowed (%)

3783 38 322 0.179/0.209 14.5/25.6 0.010/1.34 93.9/5.9/0.2

a

Values in parentheses are for the highest shell. The data were collected ˚ at beamline PX-I of the Swiss Light at a wavelength of 0.91841 A Source. The crystal belongs to spacegroup P212121 with 1 molecule per ˚, asymmetric unit, 36% solvent, and unit cell parameters a = 51.0 A ˚ , c = 158.9 A ˚. b = 51.1 A

the sites were considered fully occupied. The substrate binding structure is shown in Fig. 1. The co-crystallization experiment with ADP and L -rhamnulose-1-phosphate but without the divalent cation required for catalysis [12,19] were designed to show RhuK in complex with the reactants. Unfortunately, the transferred phosphoryl group was missing in the crystal. A disordered phosphate was ruled out because the corresponding space was filled by well-ordered water molecules. We conclude that the phosphoryl group was hydrolyzed during crystallization.

The adenine moiety of ADP binds in a non-polar crevice made up of Leu315, Pro316 and Ile319 on one side, and Gly402, Gly403 and Gln405 on the other. The ribose binds with its 3 0 -hydroxyl group to Gln304-Ne. The conformation of the diphosphate moiety is stabilized by hydrogen bonds to backbone amides and side-chains from the two b-hairpins around positions 14 and 259, as well as from Gly402-N. The bound ADP corresponds closely to that of the L -fructose complex [6] except for the b-phosphoryl, which is rotated about 25 around the Pb– ˚ Oab bond. As a consequence the bc-bridge oxygen is at a 4.5 A distance from the 1-hydroxyl nucleophile of the bound L ˚ shorter than observed in the L -frucrhamnulose, which is 0.2 A tose complex, corroborating the associative type of phosphoryl transfer [20]. As shown in Fig. 1, the c-phosphoryl group can be modeled in a staggered conformation between nucleophile and leaving group in an ideal geometry for a direct in-line transfer. The sugar L -rhamnulose was bound in the cleft at the same position as L -fructose (Fig. 1, Table 2). Its 6-methyl group fitted well into a non-polar pocket formed by Gly83, Val84, Tyr102 and Phe134 from the N-domain (Fig. 2). The hydroxyl groups form hydrogen bonds to His236, Asp237, Thr238, Asn296, and via two water molecules to Glu284, all from the C-domain. The 1-hydroxyl group of L -rhamnulose is hydrogen-bonded to Asp237 functioning as the catalytic base. The binding site is virtually the same as observed with the L -fructose complex [6]. Whereas its natural phosphoryl acceptor is L -rhamnulose, activity data suggested that RhuK plays a role in the metabolism of quite a number of ‘rare sugars’ [21], phosphorylating a broad spectrum of 2-ketoses [12,19,22]. In Table 2, we list the 16 ketoses tested for RhuK activity and state their steric rela-

Fig. 1. Stereoview of the active center displayed in a ribbon plot of RhuK. The two domains have different colors. The (Fo  Fc)-difference electron density map of ADP and b-L -rhamnulose is shown at the 3.5 r contour level. Some residues and their interactions are shown. Hydrogen bonds are indicated as broken lines. The c-phosphoryl group before the transfer (pink) is modeled into the trigonal bipyramid describing the transfer reaction. The presence of the modeled Mg2+ ion (pink) at the expected position between the b- and c-phosphoryl groups most likely causes a rearrangement of the water molecules.

D. Grueninger, G.E. Schulz / FEBS Letters 581 (2007) 3127–3130

3129

Table 2 Activity of RhuK from E. coli for 2-keto-furanoses Substrate

Model Fig. 2

L -rhamnulose

A A – B A – – B B – – B A – – –

6-deoxy-D -sorbose L -xylulose L -fructose L -fuculose D -ribulose D -psicose L -tagatose D -sorbose D -xylulose D -fructose L -ribulose 6-deoxy-L -psicose D -tagatose L -sorbose D -rhamnulose

C atom configurationa 3

4

5

R R R R R R R R R Se S Se Se Se Se Se

S S S S Rd Rd R Rd S Rd R S S S Rd Rd

S Rb Sc S – Rb,c Sc Rb – Rb – S Rb Sc Rb

Residue at atom C5

Activity (%)

Reference

CH3 CH3 H CH2OH CH3 H CH2OH CH2OH CH2OH H CH2OH H CH3 CH2OH CH2OH CH3

100 87 11, 87 77 30, 56 24 0, 23 19 16 2 1 1 0 0 0 0

– [22] [12,19] [12] [12,19] [12] [12,19] [12] [12] [19] [19] [19] [22] [19] [19] [22]

a

We observed the 2S-configuration in both structurally established complexes with RhuK and used it throughout for modeling. The 5-residue makes tight contacts in the non-polar pocket formed by Gly83, Val84, Tyr102 and Phe134. c The 5-hydroxyl forms a hydrogen bond to Gly83-O. d The 4-hydroxyl loses its hydrogen bonds to Asn296 but finds a reasonably large low-polarity pocket. e The 3-hydroxyl protrudes into a narrow pocket and loses its hydrogen bonds to His236, Thr238 and water molecule II. b

tion to the prime substrate L -rhamnulose. Since L -rhamnulose and L -fructose assume the S-configuration at the C2 atom (banomer) when bound to RhuK and since RhuK obviously requires the R-configuration at the C3 atom (Table 2), we suggest that all 2-ketoses bind as 2S-anomers, which are the aand b-anomers in the D - and L -series, respectively. The 2Ranomer would not fit the binding site (Fig. 2). Accordingly, we modeled all sugars as 2S-anomers manually into the active center so that the catalytic base Asp237 forms hydrogen bonds to the 1- as well as to the 2-hydroxyls. Given this fit, the remainder of the sugars is spatially rather restricted so that we refrained from using docking programs, but positioned the sugars by mere rotation/translation. None of the modeled ˚ or a van-dersugars forms a hydrogen bond shorter than 2.5 A ˚ , indicating that the models are Waals contact closer than 3.5 A consistent with binding to a slightly relaxed protein. The fits explain the observed activities [12,19,22] rather well. In particular, we show in Fig. 2 multiple hydrogen bonds around the 3R-hydroxyl, which are lost in the 3S-epimer. The surrounding water molecule I is suspended from Tyr102 and polarizes the nucleophile. Water molecules II and III are held by Glu284 (forming a buried saltbridge to Arg292) and promote binding of the 3and 4-hydroxyls. All sugar models are compatible with the three bound water molecules although in some cases their 3- and 4hydroxyls make close contacts to them. However, a contact with bound water is rather easily relaxed, in particular when it is (like molecule III) backed up by other water. The residue at the C5 atom of the sugar seems to be of lesser importance as it faces a non-polar pocket formed by Gly83, Val84, Tyr102 and Phe134. In summary, the data of Table 2 show rather clearly that the configuration at the C3 atom is most important for activity. Actually the configuration at C2 is equally decisive (Fig. 2) but for a 2-ketose the 2S-epimer is just one of two available anomers. The 4S-epimer is of much lesser importance because the 4R-sugar still finds enough space. The 5S-epimer faces a rather spacious non-polar pocket so that a replacement is not a severe problem. Accordingly, the observed activity values of rare 2-ketoses [12,19,22] are rather well explained by the active center structure. The 2-ketoses are a subgroup of the large

family of ‘rare sugars’ that have been found in a number of organisms [23–25]. Recently, the interest in rare sugar production increased appreciably because they are useful ingredients of various bioactive compounds [11,12,23–25]. Acknowledgements: We thank C. Beierlein for the gift of lactaldehyde and for discussions as well as the team of the Swiss Light Source (Villigen, Switzerland) for their help in data collection.

References [1] McNeil, M., Darvill, A.G., Fry, S.C. and Albersheim, P. (1984) Structure and function of the primary cell walls of plants. Annu. Rev. Biochem. 53, 625–663. [2] Chiu, T.-H. and Feingold, D.S. (1969) L -rhamnulose-1-phosphate aldolase from Escherichia coli. Crystallization and properties. Biochemistry 8, 98–108. [3] Graninger, M., Kneidinger, B., Bruno, K., Scheberl, A. and Messner, P. (2002) Homologs of the Rml enzymes from Salmonella enterica are responsible for dTDP-b-L -rhamnose biosynthesis in the Gram-positive thermophile Aneurinibacillus thermoaerophilus DSM 10155. Appl. Environ. Microbiol. 68, 3708–3715. [4] Moralejo, P., Egan, S.M., Hidalgo, E. and Aguilar, J. (1993) Sequencing and characterization of a gene cluster encoding the enzymes for L -rhamnose metabolism in Escherichia coli. J. Bacteriol. 175, 5585–5594. [5] Kroemer, M., Merkel, I. and Schulz, G.E. (2003) Structure and catalytic mechanism of L -rhamnulose-1-phosphate aldolase. Biochemistry 42, 10560–10568. [6] Grueninger, D. and Schulz, G.E. (2006) Structure and reaction mechanism of L -rhamnulose kinase from Escherichia coli. J. Mol. Biol. 359, 787–797. [7] Buss, K.A., Cooper, D.R., Ingram-Smith, C., Ferry, J.G., Sanders, D.A. and Hasson, M.S. (2001) Urkinase: structure of acetate kinase, a member of the ASKHA superfamily of phosphotransferases. J. Bacteriol. 183, 680–686. [8] Bennett Jr., W.S. and Steitz, T.A. (1978) Glucose-induced conformational change in yeast hexokinase. Proc. Natl. Acad. Sci. USA 75, 4848–4852. [9] Page, R., Lindberg, U. and Schutt, C.E. (1998) Domain motions in actin. J. Mol. Biol. 280, 463–474. [10] Jiang, J., Prasad, K., Lafer, E.M. and Sousa, R. (2005) Structural basis of interdomain communication in the hsc70 chaperone. Mol. Cell 20, 513–524. [11] Liu, K.K.-C., Kajimoto, T., Chen, L., Zhong, Z., Ichikawa, Y. and Wong, C.-H. (1991) Use of dihydroxyacetone

3130

D. Grueninger, G.E. Schulz / FEBS Letters 581 (2007) 3127–3130

Fig. 2. Stereoview of the sugar binding site of RhuK showing the established bound substrates (black) and models of half of the enzymatically tested 2ketoses (Table 2). Hydrogen bonds are black dashed lines. Chain cuts are marked by halos. The residues are given in their domain colors (Fig. 1). (A) ˚ Bound b-L -rhamnulose (black) with hydrogen bonds. The red dotted line runs to the c-phosphoryl-donating oxygen of the b-phosphoryl group at a 4.5 A distance from the 1-hydroxyl nucleophile (Fig. 1). The modeled rare sugars 6-deoxy-a-D -sorbose (5-epimer, blue), b-L -fuculose (4-epimer, green) and 6deoxy-b-L -psicose (3-epimer, magenta) are also shown. (B) Same as panel (A) but bound b-L -fructose (black) with hydrogen bonds [6]. The modeled rare sugars are a-D -sorbose (5-epimer, blue), b-L -tagatose (4-epimer, green) and b-L -ribulose (3-epimer without the 6-CH2OH residue, magenta).

[12] [13] [14] [15] [16]

[17] [18]

phosphate-dependent aldolases in the synthesis of deoxyazasugars. J. Org. Chem. 56, 6280–6289. Fessner, W.-D., Badia, J., Eyrisch, O., Schneider, A. and Sinerius, G. (1992) Enzymatic syntheses of rare ketose 1-phosphates. Tetrahedron Lett. 33, 5231–5234. Kabsch, W. (1993) Automatic processing of rotation diffraction data from crystals of initially unknown symmetry and cell constants. J. Appl. Crystallogr. 26, 795–800. Murshudov, G.N., Vagin, A.A. and Dodson, E.J. (1997) Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr. D 53, 240–255. Schu¨ttelkopf, A.W. and van Aalten, D.M.F. (2004) PRODRG: a tool for high throughput crystallography of protein-ligand complexes. Acta Crystallogr. D 60, 1355–1363. Perrakis, A., Sixma, T.K., Wilson, K.S. and Lamzin, V.S. (1997) wARP: improvement and extension of crystallographic phases by weighted averaging of multiple refined dummy atomic models. Acta Crystallogr. D 53, 448–455. Emsley, P. and Cowtan, K. (2004) Coot: model-building tools for molecular graphics. Acta Crystallogr. D 60, 2126– 2132. Fenn, T.D., Ringe, D. and Petsko, G.A. (2003) POVscript+: a program for model and data visualization using persistence of vision ray-tracing. J. Appl. Crystallog. 36, 944–947.

[19] Chiu, T.H. and Feingold, D.S. (1964) The purification and properties of L -rhamnulokinase. Biochim. Biophys. Acta 92, 489– 497. [20] Abele, U. and Schulz, G.E. (1995) High resolution structures of adenylate kinase from yeast ligated with inhibitor Ap5A, showing the pathway of phosphoryl transfer. Protein Sci. 4, 1262– 1271. [21] Badia, J., Gimenez, R., Baldoma, L., Barnes, E., Fessner, W.-D. and Aguilar, J. (1991) L -Lyxose metabolism employs the L rhamnose pathway in mutant cells of Escherichia coli adapted to grow on L -lyxose. J. Bacteriol. 173, 5144–5150. [22] Schneider, A. (1994) Isolierung und Charakterisierung der LFuculose-1-phosphat-Aldolase und der L-Rhamnulose-Kinase, Enzymatische Synthese seltener Ketose-phosphate. Dissertation. Albert-Ludwigs-Universita¨t Freiburg im Breisgau. [23] Khoo, K.-H., Suzuki, R., Dell, A., Morris, H.R., McNeil, M.R., Brennan, P.J. and Besra, G.S. (1996) Chemistry of the lyxosecontaining Mycobacteriophage receptors of Mycobacterium phlei/ Mycobacterium smegmatis. Biochemistry 35, 11812–11819. [24] Muniruzzaman, S., Pan, Y.T., Zeng, Y., Atkins, B., Izumori, K. and Elbein, A.D. (1996) Inhibition of glycoprotein processing by L -fructose and L -xylulose. Glycobiology 6, 795–803. [25] Ahmed, Z. (2001) Production of rare pentoses using microorganisms and their enzymes. Electron. J. Biotechnol. 4, 103–111.