Crystallization and preliminary X-ray studies of Escherichia coli glycerol kinase

J. Mot Biol. (1989) 207, 637-639

Crystallization and Preliminary X-ray Studies of Escherichia coli Glycerol Kinase Escherichia coli glycerol kinase, a major regulatory enzyme which catalyzes the reversible MgATP-dependent phosphorylation of glycerol has been crystallized by the hanging drop vapor diffusion method at room temperature. Three different crystal forms have been obtained in the presence of glycerol and appear to be suitable for X-ray crystallographic studies. Vapor diffusion against 55% ammonium sulfate and 1 y0 /3-octyl glucpside (pH 7.0) yields rhombohedral crystals with space group R32, a = b = 277.1 A, c = 78.7 A (hexagonal indexing) containing a dimer of 2M, 112,000 in the asymmetric unit (V,,, = 2.64 As/dalton). Vapor diffusion against sodium chloride in the presence of 10% (w/v) polyethylene glycol (pH 65 to 7.0) yields two different crystal forms, both with space group P21. The first form has a = 881 A, b = 993 A, c = 1146 A, fi = 119”, the second form has a = 92.5 A, b = 117.6 A, c = 108.3 A, fi = 9364”. Addition of ADP enhances growth of the monoclinic forms. These forms appear to contain an entire tetramer of M, 224,000 in the asymmetric unit and have V, values of 2.28 and 2.65 a3/dalton, respectively. All forms diffract to better than 3.0 A resolution while the second monoclinic form diffracts to approximately 1.8 A.

reduction in reaction velocity. Saier and co-workers (Novotony rt al., 1985) have demonstrated that factor IIIg” of the phosphotransferase system is also an allosteric inhibitor of glycerol kinase and by mutational analysis showed that the binding site for this regulator is distinct from that of FBP. Factor IIIg’” regulates other pathways as well, so glycerol kinase may contain a common regulatory “receiver module” such as that recently suggested to be involved in the regulation of chemotactic activity in E. coli (Kofoid & Parkinson, 1988). We have further charact,erized GKase including cloning and sequencing the gene of the wild-type of the enzyme (Pettigrew, 1987; Pettigrew et al., 1988). The molecule has 501 amino acid residues for a tetramer M, of 224,000. The evidence gathered to date suggests that in the presence of glycerol, there are two non-equivalent pairs of nucleotide binding sites per enzyme tetramer (D. W. Pettigrew & Y. Liu, unpublished results). Interestingly, all of the crystals characterized appear to have at least a dimer in the asymmetric unit. Therefore it is possible that the enzyme has only S-fold symmetry and pairs of subunits that are chemically identical but not structurally so.

Glycerol kinase (GKaset, EC 2.7.1.30), since its isolation from Escherichia coli (Hayashi & Lin, 1967) and characterization, has been the subject, of considerable interest. The enzyme is inducible by glycerol and is required for growth on glycerol (Lin et al.. 1962). lt catalyzes the MgATP-dependent phosphorylation of glycerol and thus supplies two major metabolic pathways, glycolysis and phospholipid biosynthesis. Glycerol kinase is the rate-determining regulatory enzyme in E. coli glycerol metabolism. Subsequent characterization by Paulus and coworkers has revealed the enzyme to have very unusual kinetic properties. Thorner & Paulus (1973) discovered that GKase is a velocity modulated chnzyme (making it one of only about a half dozen discovered to date and the only one to our knowledge that, has been crystallized) which displays a complicated response to ATT’, ADP, AMP and fructose 1.6.bisphosphate (FBI’). Double reciprocal plots of reaction velocity vf?rfiu.s lMgATP concentration are curved, providing evidence for two apparent K, values, one in the range of 80 to 100 PM, the other in the range of 400 to 500 PM, indicative of possiblr negative co-operativity. MgADP has characteristics of both a competitive and non-competitive inhibitor depending on relative concentration. but, is clearly a non-competitive inhibitor with respect t,o glycerol. de Riel & Paulus (1978a,b,c) have shown that there is a dimer-tetramer equilibrium at physiological concentrations with an apparent dissociation c*onstant of about 5 x IV8 M (subunit). FRP binds selectively to the tetramer, stabilizing this state by two orders of magnitude and can cause up to 947;

(a) Overproduction

and purijkation

glycerol

qf

kinase

An overexpression system has been constructed for the glycerol kinase (gZpK) gene. This was accomplished by inserting the 2.8 kb HindIII restriction fragment from a pCJlO2 plasmid (Pettigrew et aE., 1988) into the Hind111 site of pTTQl8 (Amersham). The new plasmid (pRFQl8) was used to transform JJ161, a strain of E. co& that is Hfr, RelAl, glpkl, and metB1. The purification procedure of Thorner & Paulus (1971), as modified by Pettigrew (1986), has been further modified for use with this overproducing strain.

t Abbreviations used: GKase, glycerol kinase; FBP, fructose 1,6-bisphosphate; kb, 10S bases; IPTG, isopropyl-B-n-thiogalactoside. 637 0(322%2836/89/l

10637-43

$03.00/O

0

1989 Academics

Press Limited

638

H. R. Faber et al.

a a Figure

2. Packing diagram of t,he YZ projection for ITT of glycerol kinase (axes not to scale). Each H is intended to represent a dimer which may not

crystal

form

have t,he Z-fold

Z is to the right,

Figure 1. 12” OkE precession photograph of crystal TIT of glycerol kinase. 1 is up. k is to the right.

and preliminary

implied

by this representation.

Y is up.

form

A cell culture of 4 liters is grown in Luria broth until the end of logarithmic phase. At this point the cells are induced with 1 m&r-IPTG and allowed to continue growth for two to three hours before harvesting. The standard buffer used throughout the purification consists of 20 mmTris (pH 8.0 unless noted), 10 mlrr-glycerol, and 1 mlcl-/3-mercaptoethanol. After ammonium sulfate precipitation (Thorner & Paulus, 1971) the protein is loaded on a 5 cm x 20 cm DEAE-cellulose column equilibrated with standard buffer containing 91 M-KCl, and eluted with a 91 M to 93 M-KC1 linear gradient (400 x 400 ml). The second column is 5 cm x 20 cm DEAF-Sephadex run as above but at pH 7.0. The purified protein is then dialyzed into 50 mM-Na/Kphosphate (pH 6*5), 10 mM-&CerOl, and 1 mM$mercaptoethanol. The protein is stored as described previously. Typically, a 4-liter culture will yield 150 to 300 mg of purified enzyme. (b) Crystallization

symmetry

X-ray studies

Crystallization of glycerol kinase was achieved by the hanging drop method of vapor diffusion. Two different sets of conditions have led to three crystal forms, which have been characterized by X-ray diffraction. Drops of protein solution (5 ~1, 10 to 30 mg/ml in 50 mM-Na/K-phosphate pH 65 to 7.0, 10 mM-&CerOl, 2 mM-j?-mercaptoethanol) were mixed with an equal volume of precipitant, placed on silanized coverslips and suspended over wells (Linbro tissue culture plates) containing 1 ml pre cipitant. In one series of experiments, the well solution was 55% saturated ammonium sulfate prepared by mixing saturated ammonium sulfate with the appropriate amount of the buffer in which the protein was dissolved. To each 5 ~1 protein drop was added 5 ~1 of well solution and 1 ~1 10% (w/w)

/I-octyl glucoside. In several weeks, rhombohedral crystals with maximum dimensions of about 62 mm x 62 mm x O-2 mm formed amid a cloudy precipitate. In a second series of experiments, the well solution consisted of 300 to 500 m&f-sodium chloride. To each 5 ~1 protein drop was added 5 ~1 of 20% (w/w) polyethylene glycol, &l, 1550 to 20,000 in 50 m&i-Na/K-phosphate, pH 65 to 7.0. Use of higher molecular weight polyethylene glycol requires less sodium chloride in the well. Addition of 95 ~1 100 mM-ADP to the protein drop enhances the growth of these crystals. Crystals appeared in a few days and reached maximum sizes of about 0.5 mm x 0.5 mm x 1.0 mm after several weeks. Precipitate did not always appear in these experiments. We have found that in order to obtain the largest crystals, it is important to use the freshest possible protein. Preliminary characterization by precession photography of crystals obtained under these conditions on a Rigaku RU-100 rotating anode X-ray generator operated at 40 kV, 130 mA yielded the following crystal forms. (I) From vapor diffusion against ammonium sulfate, the space group is R32 with a = 6 = 277.1 A, c = 78.7 A (hexagonal indexing). Assuming a dimer of molecular weight 112,000 in the asymmetric unit. the crystal packing parameter (Matthews, 1968) If,,,= 2.64 A3/dalton, which is quite typical of protein crystals (1 A = 0.1 nm). (II) From vapor diffusion against polyethylene glycol ilf, 1550 to 8000, space group P2, with a = 88*1 A, b=99.3 A, c=1146 A and /3=119” V,= 228 A3/dalton for a tetramer in the asymmetric unit. (III) Under the same conditions as in (II) above, also P2, with a = 925 A, b = 117.6 A, c = 108.3 A for a tetramer and /I = 93*64”, V, = 2.65 A3/dalton in the asymmetric unit. Crystals of form III are currently the most pro mising for structural work as they diffract to

639

Letters to the Editor

approximately 1.8 A resolution. It has not yet been possible to determine t,he difference in crystallization conditions leading to crystal forms II and III. Form II crystals tend to be elongated in the c direction compared to form III but it is not always possible to distinguish the two without a diffraction pattern. It may be that forms II and III can grow in the same drop. Inspection of the diffraction patterns of crystal form III reveals the presence of an approximate superlattice (note the alternating light and dark rows in Fig. 1 and the curious pattern of pseudosystematic absences). These pseudosystematic absences can be shown to be due to a molecular S-fold axis exactly parallel to the crystallographic b axis, located at 2 = l/8. A packing diagram (Fig. 2, axes not drawn to scale) illustrates our reasoning: if the rows with 1 odd were systematically absent, and the extinctions for 1 even complete, this would correspond to a unit cell centered and halved in the YZ projection (dotted molecules). However, the dotted molecules would only contribute to systematic absences with 1 odd because of the If2 cell translation in 2. Therefore, the pseudosystematic absences for 1 even are due to the special arrangement of the molecules shown with continuous lines and this requires the molecule to have a 2-fold axis. Part of the native data set for form III has been collected by rotation photography. With this partial dataset, a nat,ive Patterson map was calculated, and shows a peak approximately 80% of the origin height at V = l/2, II: = l/5 and W = l/4. This confirms our reasoning about the molecular 2-fold; it is located at X = l/IO and Z = l/S, parallel to Y. We intend to solve these structures by multiple isomorphous replacement and the search for heavy-atom derivatives is underway. We gratefully acknowledge the advice of Dr Masazumi Matsumura and the technical assistance of Karen Kallio

and David Worthylake. This work was supported in part by a grant from the Kational Science Foundation and from the Texas Agricultural Experiment Station. H. R. Faber’

D. W. Pettigrew’ S. J. Remington’ ‘Institute of Molecular Biology and Department of Physics University of Oregon Eugene, OR 97403. LT.&A. ‘Department of Biochemistry and Biophysics Texas A&M University College Station. TX 77843, U.S.A. Received 17 November

1988

References De Riel, J. K. & Paulus, H. (197%). Riochemistry, 17, 5134-5140. 17, De Riel, J. K. & Paulus, H. (19786). Bioch,emistry, 514lL5146. De Riel. J. K. & Paulus. H. (1978c). Hiochemistry, 17, 5146-5150. Hayashi, S. & Lin, E. (1. C. (1967). .I. Riol. Ohem. 242, 1030-1035. Kofoid, E. C. & Parkinson, J. 6. (1988). Proc. Nat. Acad. Sci., U.S.A. 85, 4981-4985. Lin. E. C. C.. Koch. J. P.. Chused. T. ,M. & Jorgensen, S. E. (1962). Proc. Nat. Acad. Sci., r1.S.A. 48. 2145 2149. Matthews, B. W. (1968). J. Mol. Hiol. 33. 491497. Novotony. M. ,J., Frederickson, E. B. W. & Saier. M. H., (1985). .J. Bacterial. 162, 810-816. Pettigrew, Il. W. (1986). Biochemistry, 25, 471 l-4718. Pettigrew. D. W. (1987). Biochemistry, 26, 1723 ~1727. Pettigrew, D. W., Ma, D.-P., Conrad. C. A. & *Johnson, .J. R. ( 1988). J. Biol. C’hem. 263, 135139. Thorner. J. W. & Paulus, H. (1971). J. Biol. (‘hwn. 296, 38853894. Thorner, J. W. & Paulus. H. (1973) J. Hiol. Chsm. 248, 3922-3932. t*

Edited by R. Huber