374

Biochimica et BiophysicaAct~ 1076(1991) 374-378 © 1991 ElsevierSciencePub~lishersB.V.0167-4838/91/$03.50 ADONIS 0167483891001087

BBAPRO ~ 8 ~

The importance of glycine-30 for enzymatic activity of phospholipase A2 August C_A.P.A. Bekkers, Peet A. Franken, Ellis Toxopeus, Hubertus M. Verheij and Gerard H. de Haas Department of Enzymology and Protein Engineering Universityof Utrecht, CBLE, University Centre "De Uithof" Utrecht (The Nellu~rlands)

(Received23 July 1','90) Key words: PhospholipaseA2; Calciumicn binding; Catalyticmechanism The nearly conserved glycine-30 in porcine pancreatic phosphalipase A z I n s been replaced by serine. The resulting mutant G30S was expressed in Escherlchla coil, purified and characterized. T h e mutation caused a significant drop in enzymatic activity towards monomeric and aggregated substrates, bat had a limited effect on substrate binding. In contrast the alTmity for calcium ions, the essential cofactor, was reduced 10-fold. The reduced enzymatic activity is attn'buted to a reduced stabilization of the transition state. The results are discussed in view of naturally occurring inactive phospholipase A 2 homologues from snake venom. Introduction Many tissues contain phospholipase A 2 (EC 3.1.1.4., phosphatidylcholine 2-acylhydrolase, PLA2) catalyzing the hydrolysis of fatty acid ester bonds at the sn-2 position of 3-sn-phospholipids in a calcium dependent reaction. The extracellular enzymes from snake venom, mammalian pancreas and the cellular enzymes from mammalian cells have significant sequence homology (for a recent review see Ref. 1). The three-dimensional structore of several extracellular PLA2s from both mammalian pancreas and snake venom has been elucidated. As discussed in detail [2], these structures share significant spatial homology as well. Based on the structural data and on the results of chemical modification reactions a mechanism for the catalytic action of PLA2was postulated [3]. In this model the catalytic unit comprises His-48, Asp-99 and a water molecule. Their function is thought to be analogous to the catalytic residues in serine proteinases. The catalytically important calcium ion is figanded by the carbonyl oxygens

Abbreviations: PLA2, phospholipaseA2; diCsPC , 1,2-dioctanoyl-snglycero-3-phosphocholine; diC6dithioPC, Rac-l,2-dihexanoyldithinpropyl-3-phosphocholine;CI2PN, n-dodecylphosphocholine;C16PN, n-hexadecylphosphocholine. Correspondence:H.M. Verhcij,Department of Enzymologyand Prorein Engineering,Universityof Utrecht, CBLE, P.O. Box 80.054,3508 TB Utrecht, The Netherlands.

of Tyr*28, Gly-30, Gly-32 and by the side chain oxygens of Asp-49 [2]. The role of the calcium ion was proposed [3] to be 2-fold: to fix the phosphate of the substrate and to stabilize together with the backbone amide of Gly-30 the formed tetrahedral intermediate. Many snake venoms contain PLA2s , some of which exert (neuro)toxic properties. The most thoroughly characterized venoms contain several isozymes with significant sequence variations and thus display different enzymatic and pharmacological activities [4]. The observed variation of activities can be moderate, but several PLA 2 variants are known which lack all enzymatic and toxic activities tested for [5]. In some sequences the lack of activity can be explained quite easily, since residues that are thought to be essential for catalysis have been replaced. Examples are the sequence of a PLA 2 homologue in the venom of Vipera ammodytes ammodytes where the essential His-48 has been replaced by Gin-48 [6] and the PLA 2 variant in the venom of Agkistrodon piscivorus piscivorus where the calcium ligand Asp-49 has been replaced by Lys-49 [5,7]. For the non-neurotoxic, non-enzymatic PLA 2 homologue notechis II-1 from the venom of the Australian tiger snake Notechis scutatus scutams [8,9] it is more difficult to define the structural basis for its lack of activity. This inactive homologue differs from two active and toxic isozymes t~otexin and notechis I1-5 from the same venom by not less than 41 and 42 substitutions, respectively, in a sequence of 119 amino acids. Because the inactive homologue notechis II-1 contains

375 all of the residues that appear to be invariant in active PLA2, except for position 30 in the sequence where a serine is substituted for an otherwise invariant glycine, it was suggested [9] that this substitution is the main cause of the loss in activity. In the present paper we describe the substitution of serine for glycine at position 30 of the cloned porcine pancreatic PLA2 by site-directed mutagenesis. Thus we were able to study the effect of a single-point mutation on the catalytic function and on the binding properties for substrate (analogues) and calcium ions of this mutant PLA 2. Materlals and Methods Construction and purification of mutant PLA 2 Escherichia coil K-12 strain PC 2494 (/t(lac-pro), supE, t h i / F ' , traD36, proA+B +, lacl q, IacZAM15, Phabagen collection, Utrecht) was used for plasmid constructions and as a host for M13-derived vectors. HB 2154 (ara, A(lac-pro), t h i / F ' , proA+B +, lacI q, lacZAM15, mutL::Tnl0) [10] was used as a recipient strain in the mutagenesis experiments. Substitutions in the proPLA2-cDNA [11] were introduced by the gapped duplex procedure, using amber selection [12]. The mutageulc oli8onuclentide 5" CACCTAGGCTGCAGTAGCAG 3' was used for site-directed mutagenesis at position 30 and to introduce a Pstl-site to be used for screening procedures. It was synthesized on a Biosearch 6800 DNA synthesizer. The sites of mutation in the sequence of the oligonucleotide are underlined. The resulting mutant proPLA2-cDNA was sequenced by the didcoxy chain ~ermination method after p!aque purifying [13] and a BamHI-BstXl fragment of the mutant proPLA2-cDNA was cloned into the expression vector described before [14]. After tranaformation and expression in E. coli K-12 strain MC 4100 [15], containing plasmid pCI857, the mutant PLA 2 was obtained by tryptic cleavage of reoxidized fusion protein [11] and purified by ion exchange chromatography at pH 5 and 6 using CM-cellulose. Final purification to homogeneity was achieved on Mono-Q (Pharmacia) at pH 8.3. Protein concentrations were determined from the absorbance at 280 nm, using an Atl~,~ = 13.0 for the wildtype and the mutant enzyme. Amino acid analysis Amino acid analyses samples were hydrolyzed in vacuo for 24 h at 110°C in 6 M HCL Analyses were carried out on an LKB 4151 alpha plus analyzer. The values for Thr and Ser were corrected for 5 and 10% loss, respectively. Phosphoh'pid~ The 1,2-dioctanoyl-sn-glycero-3-phosphocholine (diCsPC) used in this study was synthesized according to Cubero Robles and van den Berg [16]. Rac-l,2-di-

hexanoyldithio-propyl-3-phosphocholine (diC6dithioPC) was obtained as described by Volwerk and co-workers [17], n-dodecylphosphocholine (Ct2PN) and n-hexadecylphosphocholine (Ci6PN) were prepared as described before [18]. Phospholipase A 2 assays Quantitative measurements with diCsPC as a substrate were carried out with a titrimetric assay at pH 6 in the presence of 1 mM acetate, varying concentrations of CaCI 2 and 100 mM NaC! at 25°C [19]. The burette of the Radiometer ABU was filled with 10 mM NaOH. Activities on monomeric diC6dithioPC were determined at pH 6 in the presence of 200 mM acetate, 10O mM NaCI and 50 mM CaCI2, as described previously [17]. PLA 2-activities on egg-yolk lipoproteins as a substrate were determined as described before [20]. Direct binding studies Direct binding of calcium and substrate analogues was measured with ultraviolet absorbance spectroscopy. Difference spectra were recorded on an Aminco Model DW-2a speetrophotometer equipped with a Midan data analyser, coupled to an Apple !I desk-top computer. Measurements were carried out as described earfier [18]. The enzymes were dissolved up to a concentration of 20 p M in a buffer containing 100 mM NaAc (pH 6.0) and 0.1 M NaCI. Calcium chloride was titrated up to 100 mM. The affinity of PLA2 for monomers and micelles was determined by following the increase of tyrosine and tryptophan fluorescence upon addition of increasing concentrations of the non-hydrolyzable substrate analogues Ct2PN (CMC 1.3 mM) for monomer binding and Cz6PN (CIVIC 10 pM) for miceUe binding. Fluorescence studies were performed in a buffer containing 100 n~M NaAc, 50 mM CaCI 2 and 100 mM NaCI at pH 6.0 on a Perkin-Elmer LS-5 luminescence spectrometer, at 25°C. Enzyme concentrations were 5-7 FM. Excitation wavelength was at 280 nm and the emission spectra were recorded from 310-370 nm; slit-widths were 5 nm. From saturation curves, obtained with fipid monomers, a Kd-ValUe can directly be derived, The data concerning micelle binding were analyzed in terms of the binding of the enzyme to a theoretical fipid particle consisting of N monomers with a dissociation constant K d. The N - K d value is the experimental concentration at which 50% of the enzyme is saturated with micelles [21]. Circular dichroism Spectra for G30S and wildtype PLA 2 were obtained on a Jasco J-600 spectropolarimeter at a concentration of 7.2/tM in 10 mM Tris (pH 8.0), in the absence and presence of 25 mM CaCI 2. All recordings were made at room temperature with a cell path length of 0.1 cm.

376 p-Bromophenacyl bromide inactivation Rates of modification of PLA 2 by p-bromophenacyl bromide were determined at 30°C in a 0.1 M cacodylate buffer (pH 6.0). The inactivation rates were determined both in the absence and in presence of varying coneentrations CaC12. Protein concentrations were 2 0 / z M and the reaction was started by the addition of 0.2 rag reagent in 10 pl acetonitrile to a final vol. of I ml. The rate of inactivation was determined by taking samples and analyzing them for residual activity. Half-time values for the reaction were determined from the slopes of semi-logarithmlc plots of residual activity vs. time as described before [22].

strates and we conclude that this mutation does not induce large conformational changes in the protein. This notion was supported by the observation that circular dichroism spectra obtained with the mutant and native PLA2 were similar, either in the presence or in the absence of 25 m M CaCI2 (results not shown). The fact that the k ~ value is significantly lower for the mutant than for the native PEA 2 might be explained by a small local effect in the active site affecting the positioning of the catalytically important residues Asp99, His-48, Asp-49 or of the essential cofactor calcium. This proposed local change is apparently too small to be detectable by means of circular dichrolsm.

Resdls

Calcium binding of mutant and native PLA2 In all PLA2s calcium is figanded by the side-chain of Asp-49 and the backbone carbonyls of Tyr-28, Gly-30 and Gly-32 [2]. Thus substitution of serine for glycine at position 30 might change the affinity for calcium ions a n d / o r its exact positioning. The affinity for calcium ions was determined with three independent techniques: (1) kinetically in the presence of substrate; (2) by ultraviolet difference spectroscopy; and (3) from the protective effect of calcium ions on alkylatiun of His-48. Based on the kinetic measurements with varying calcium concentrations, it was found that in the presence of substrate the mutant PLA 2 binds calcium ions about 20-fold less efficiently than native PLA 2 does (20 m M vs. 1 m M at p H 6.0). This difference in aff'mity was confh'med by ultraviolet difference spectroscopy. Addition of calcium ions up to 10 m M induced a differenee spectrum at 242 u m similar to that observed with native PLA 2. At higher calcium concentrations a second spectral change, characterised by a trough at 279 nm, was superimposed on the spectra of the mutant PLA2, whereas the spectra of the native enzyme did not show this behaviunr (Fig. 1A). In Fi~ 1B the signal at 242 n m is plotted a~ainst the calcium concentration. For native PLA 2 a saturation curve was obtained from which a dissociation constant of 2 m M was calculated. The curve obtained with the mutant protein first increases, but decreases at higher calcium concentrations as the trough develops. Although the shape of the latter curve does not allow for exact calculations, a dissociation constant of 10-20 m M was estimated for the mutant PEA 2 based on the fwst part of the binding curve. It is well established that calcium ions protect porcine pancreatic PLA 2 from being inactivated by pbromophenacyl bromide [22]. Analysis of the rate of inhibition in the absence and presence of calcium ions yields information about the degree of protection as well as the dissociation constant of the calcium-PLA 2 complex. Inhibition of native and mutant PLA 2 with p-bromophenacyl bromide in the absence of calcium ions at p H 6.0 and 30°C occurred with half-times of

Purification and enzymatic properties The mutant PLA 2 G30S was readily purified from E. coil cells essentially as has been described for the native PLA 2 (see Materials and Methods). Three consecutive inn-exchange columns yielded pure and homogeneous material. The purified material was stored frozen in distilled water and lyophiliTation was avoided since the mutant, in contrast to native PLA2, did not yield clear solutions after lyophiliTatlon. The enzymatic activity was measured with three assay systems, i.e., in mixed mieelles of egg-yolk fipoproteins with dcoxycholate, with micelles of a short-chain phosphatidylcboline (diCsPC) and with monomers of diC~dithioPC. The results are summarized in Table I. In the three assay systems the turnover number of the G30S mutant is about 2 ~ relative to native PLA2, whereas the affinity of the mutant for substrates is largely unaffected. The unchanged affinity for substrates was confirmed by direct binding studies to non-hydrolyzable substrate analogues Ct2PN and ClsPN. The affmity for the monomeric analogue C12PN was hardly affected: 1.05 m M for G30S and 0.58 m M for native PLA2. Also the affinity for C~6PN micelles was similar for both proteins since the N - K d value was 180/tM for G30S and 59 p M for native PEA 2. These results suggest that the substitution of serine for glycine at position 30 does not destroy the binding sites for either monomeric or aggregated sub-

TABLE I Kinetic properties of native and mutant porcine pancreatic phospholipases A2 on different substrates The aetiviti~ on the syntheticlecithinswere assayed at 25°C (pH 6) in the presenceof 50 mM CaC12.For further experimentaldetails see Materials and Method. ~zyme

E88-yolk specificactivity 0tmol-min.mg)

diCsPC kca~ (rain-l)

diC6dithioPC Km kcat~Kin (raM) (s-LM-I)

wt G30S

1400 30

370 6

2.4 3.0

580 14

377 A

B !

40

I

!

;

I

I

'

-~

40

t

-

,

t

/0......0-----------0

o

0

/°

..

o

~ 20 0

~43m - e - O ' { 3 4 ~

\-

IIK>IO -IOll4:l- -

a"

~4

0/rl