Transmutation of human glutathione transferase A2-2 with peroxidase activity into an efficient steroid isomerase

A major goal in protein engineering is the tailor-making of enzymes for specified chemical reactions. Successful attempts have frequently been based on directed molecular evolution involving libraries of random mutants in which variants with desired properties were identified. For the engineering of enzymes with novel functions, it would be of great value if the necessary changes of the active site could be predicted and implemented. Such attempts based on the comparison of similar structures with different substrate selectivities have previously met with limited success. However, the present work shows that the knowledge-based redesign restricted to substrate-binding residues in human glutathione transferase A2-2 can introduce high steroid double-bond isomerase activity into the enzyme originally characterized by glutathione peroxidase activity. Both the catalytic center activity (k(cat)) and catalytic efficiency (k(cat)/K(m)) match the values of the naturally evolved glutathione transferase A3-3, the most active steroid isomerase known in human tissues. The substrate selectivity of the mutated glutathione transferase was changed 7000-fold by five point mutations. This example demonstrates the functional plasticity of the glutathione transferase scaffold as well as the potential of rational active-site directed mutagenesis as a complement to DNA shuffling and other stochastic methods for the redesign of proteins with novel functions.     

Crystal structure of human glutathione S-transferase A3-3 and mechanistic implications for its high steroid isomerase activity

The crystal structure of human class alpha glutathione (GSH) S-transferase A3-3 (hGSTA3-3) in complex with GSH was determined at 2.4 A. Despite considerable amino acid sequence identity with other human class alpha GSTs (e.g., hGSTA1-1), hGSTA3-3 is unique due to its exceptionally high steroid double bond isomerase activity for the transformation of Delta(5)-androstene-3,17-dione (Delta(5)-AD) to Delta(4)-androstene-3,17-dione. A comparative analysis of the active centers of hGSTA1-1 and hGSTA3-3 reveals that residues in positions 12 and 208 may contribute to their disparate isomerase activity toward Delta(5)-AD. Substitution of these two residues of hGSTA3-3 with the corresponding residues in hGSTA1-1 followed by kinetic characterization of the wild-type and the mutant enzymes supported this prediction. On the basis of our model of the hGSTA3-3.GSH.Delta(5)-AD ternary complex and available biochemical data, we propose that the thiolate group of deprotonated GSH (GS(-)) serves as a base to initiate the reaction by accepting a proton from the steroid and the nonionized hydroxyl group of catalytic residue Y9 (HO-Y9) functions as part of a proton-conducting wire to transfer a proton back to the steroid. Residue R15 may function to stabilize the deprotonated thiolate group of GSH (GS(-)), and a GSH-bound water molecule may donate a hydrogen bond to the 3-keto group of Delta(5)-AD and thus help the thiolate of GS(-) to initiate the proton transfer and the subsequent stabilization of the reaction intermediate.     

Targeting human glutathione transferase A3-3 attenuates progesterone production in human steroidogenic cells

hGSTA3-3 (human Alpha-class glutathione transferase 3-3) efficiently catalyses steroid Delta(5)-Delta(4) double-bond isomerization in vitro, using glutathione as a cofactor. This chemical transformation is an obligatory reaction in the biosynthesis of steroid hormones and follows the oxidation of 3beta-hydroxysteroids catalysed by 3beta-HSD (3beta-hydroxysteroid dehydrogenase). The isomerization has commonly been ascribed to a supplementary function of 3beta-HSD. The present study is the first to provide evidence that hGSTA3-3 contributes to this step in steroid hormone biosynthesis in complex cellular systems. First, we find glutathione-dependent Delta(5)-Delta(4) isomerase activity in whole-cell extracts prepared from human steroidogenic cells. Secondly, effective inhibitors of hGSTA3-3 dramatically decrease the conversion of Delta(5)-androstene-3,17-dione into Delta(4)-androstene-3,17-dione in cell lysates. Thirdly, we show that RNAi (RNA interference) targeting hGSTA3-3 expression decreases by 30% the forskolin-stimulated production of the steroid hormone progesterone in a human placental cell line. This effect is achieved at low concentrations of two small interfering RNAs directed against distinct regions of hGSTA3-3 mRNA, and is weaker in unstimulated cells, in which hGSTA3-3 expression is low. The results concordantly show that hGSTA3-3 makes a significant contribution to the double-bond isomerization necessary for steroid hormone biosynthesis and thereby complements the indispensable 3beta-hydroxysteroid oxidoreductase activity of 3beta-HSD. The results indicate that the lower isomerase activity of 3beta-HSD is insufficient for maximal rate of cellular sex hormone production and identify hGSTA3-3 as a possible target for pharmaceutical intervention in steroid hormone-dependent diseases.     

Structure-guided activity restoration of the silkworm glutathione transferase Omega GSTO3-3

Glutathione transferases (GSTs) are ubiquitous detoxification enzymes that conjugate hydrophobic xenobiotics with reduced glutathione. The silkworm Bombyx mori encodes four isoforms of GST Omega (GSTO), featured with a catalytic cysteine, except that bmGSTO3-3 has an asparagine substitution of this catalytic residue. Here, we determined the 2.20-Å crystal structure of bmGSTO3-3, which shares a typical GST overall structure. However, the extended C-terminal segment that exists in all the four bmGSTOs occupies the G-site of bmGSTO3-3 and makes it unworkable, as shown by the activity assays. Upon mutation of Asn29 to Cys and truncation of the C-terminal segment, the in vitro GST activity of bmGSTO3-3 could be restored. These findings provided structural insights into the activity regulation of GSTOs.     

The hereditary transmission of high glutathione transferase activity towards trans-stilbene oxide in human mononuclear leukocytes

Summary High glutathione transferase activity towards trans-stilbene oxide has been observed in resting mononuclear leukocytes only in a portion of the individuals examined. Approximately 46% of a population of 248 individuals demonstrated this high activity. In addition, eight families have been investigated in order to elucidate the hereditary transmission of this activity. The results are consistent with a dominant expression of a single gene located on an autosomal chromosome for this high glutathione transferase activity.     

Human glutathione transferase A3-3, a highly efficient catalyst of double-bond isomerization in the biosynthetic pathway of steroid hormones.

The cDNA of a novel human glutathione transferase (GST) of the Alpha class was cloned, and the corresponding protein, denoted GST A3-3, was heterologously expressed and characterized. GST A3-3 was found to efficiently catalyze obligatory double-bond isomerizations of Delta(5)-androstene-3,17-dione and Delta(5)-pregnene-3,20-dione, precursors to testosterone and progesterone, respectively, in steroid hormone biosynthesis. The catalytic efficiency (k(cat)/K(m)) with Delta(5)-androstene-3,17-dione was determined as 5 x 10(6) m(-1) s(-1), which is considerably higher than with any other GST substrate tested. The rate of acceleration afforded by GST A3-3 is 6 x 10(8) based on the ratio between k(cat) and the rate constant for the nonenzymatic isomerization of Delta(5)-androstene-3,17-dione. Besides being high in absolute numbers, the k(cat)/K(m) value of GST A3-3 exceeds by a factor of approximately 230 that of 3beta-hydroxysteroid dehydrogenase/isomerase, the enzyme generally considered to catalyze the Delta(5)-Delta(4) double-bond isomerization. Furthermore, GSTA3-specific polymerase chain reaction analysis of cDNA libraries from various tissues showed a message only in those characterized by active steroid hormone biosynthesis, indicating a selective expression of GST A3-3 in these tissues. Based on this finding and the high activity with steroid substrates, we propose that GST A3-3 has evolved to catalyze isomerization reactions that contribute to the biosynthesis of steroid hormones.     

Folding and assembly of dimeric human glutathione transferase A1-1

Glutathione transferases function as detoxification enzymes and ligand-binding proteins for many hydrophobic endogenous and xenobiotic compounds. The molecular mechanism of folding of urea-denatured homodimeric human glutathione transferase A1-1 (hGSTA1-1) was investigated. The kinetics of change were investigated using far-UV CD, Trp20 fluorescence, fluorescence-detected ANS binding, acrylamide quenching of Trp20 fluorescence, and catalytic reactivation. The very early stages of refolding (millisecond time range) involve the formation of structured monomers with native-like secondary structure and exposed hydrophobic surfaces that have a high binding capacity for the amphipathic dye ANS. Dimerization of the monomeric intermediates was detected using Trp fluorescence and occurs as fast and intermediate events. The intermediate event was distinguished from the fast event because it is limited by a preceding slow trans-to-cis isomerization reaction (optically silent in this study). At high concentrations of hFKBP, dimerization is not limited by the isomerization reaction, and only the fast event was detected. The fast (tau = 200 ms) and intermediate (tau = 2.5 s) events show similar urea-, temperature-, and ionic strength-dependent properties. The dimeric intermediate has a partially functional active site ( approximately 20%). Final reorganization to form the native tertiary and quaternary structures occurs during a slow, unimolecular, urea- and ionic strength-independent event. During this slow event (tau = 250 s), structural rearrangements at the domain interface occur at/near Trp20 and result in burial of Trp20. The slow event results in the regain of the fully functional dimer. The role of the C-terminus helix 9 (residues 210-221) as a structural determinant for this final event is proposed.     

Functional studies of single-nucleotide polymorphic variants of human glutathione transferase T1-1 involving residues in the dimer interface

Glutathione transferase T1-1 catalyses detoxication and bioactivation processes in which glutathione conjugates are formed from endogenous and xenobiotic substrates, including alkylating agents and halogenated alkanes. Although the common null polymorphism of the human GSTT1 gene has been studied extensively, little is known about the consequences of GSTT1 single-nucleotide polymorphisms (SNPs). Here, we have examined the effects of two SNPs that alter amino acid residues in the dimer interface of the GST T1-1 protein and one that causes a conservative substitution in the core of the subunit. Variant proteins were expressed in an Escherichia coli strain in which the metabolism of ethylene dibromide to a glutathione conjugate leads to lacZ reversion mutations. We measured the kinetic properties of the enzymes with the characteristic substrate 1,2-epoxy-3-(p-nitrophenoxy)propane (EPNP) and determined the specific activities with several other substrates. Circular dichroism spectroscopy was used to measure protein thermal denaturation profiles. Variant T104P, which has been reported as inactive, showed weak but detectable activity with each substrate. Variant R76S was expressed at lower levels and showed much-reduced thermal stability. The results are interpreted in the context of the three-dimensional structure of human GST T1-1.     

Identification of a novel glutathione transferase in human skin homologous with class alpha glutathione transferase 2-2 in the rat.

Six forms of glutathione transferase with pI values of 4.6, 5.9, 6.8, 7.1, 8.5 and 9.9 have been isolated from the cytosol fraction of normal skin from three human subjects. The three most abundant enzymes were an acidic Class Pi transferase (pI 4.6; apparent subunit Mr 23,000), a basic Class Alpha transferase (pI 8.5; apparent subunit Mr 24,000) and an even more basic glutathione transferase of Class Alpha (pI 9.9; apparent subunit Mr 26,500). The last enzyme, which was previously unknown, accounts for 10-20% of the glutathione transferase in human skin. The novel transferase showed greater similarities with rat glutathione transferase 2-2, another Class Alpha enzyme, than with any other known transferase irrespective of species. The most striking similarities included reactions with antibodies, amino acid compositions and identical N-terminal amino acid sequences (16 residues). The close relationship between the human most basic and the rat glutathione transferase 2-2 supports the classification of the transferases previously proposed and indicates that the similarities between enzymes isolated from different species are more extensive than had been assumed previously.     

Cysteine residues are not essential for the catalytic activity of human class Mu glutathione transferase M1a-1a

To investigate the possible involvement of a Cys thiol in the catalysis of the human glutathione transferase M1a-1a, we constructed mutants of this enzyme wherein the four Cys residues present in the native enzyme were replaced by Ala residues. Three mutants, one where all four Cys residues had been replaced and two mutants where three out of four Cys residues were changed into Ala, were characterized regarding their catalytic activities with three different substrates as well as by their binding of three different inhibitors. All three Cys-deficient mutant forms of glutathione transferase M1a-1a were catalytically active with the tested substrates and their binding of inhibitors, measured by I50, were not significantly different from the values previously obtained for the wild-type enzyme. We therefore conclude that none of the Cys residues in this class Mu glutathione transferase are directly involved in the catalysis performed by this enzyme.     

Aromatic residues in the C-terminal region of glutathione transferase A1-1 influence rate-determining steps in the catalytic mechanism

Human glutathione transferase A1-1 (GST A1-1) has a flexible C-terminal segment that forms a helix (alpha 9) closing the active site upon binding of glutathione and a small electrophilic substrate such as 1-chloro-2,4-dinitrobenzene (CDNB). In the absence of active-site ligands, the C-terminal segment is not fixed in one position and is not detectable in the crystal structure. A key residue in the alpha 9-helix is Phe 220, which can interact with both the enzyme-bound glutathione and the second substrate, and possibly guide the reactants into the transition state. Mutation of Phe 220 into Ala and Thr was shown to reduce the catalytic efficiency of GST A1-1. The mutation of an additional residue, Phe 222, caused further decrease in activity. The presence of a viscosogen in the reaction medium decreased the kinetic parameters K(cat) and K(cat)/K(m) for the conjugation of CDNB catalyzed by wild-type GST A1-1, in agreement with the view that product release is rate limiting for the substrate-saturated enzyme. The mutations cause a decrease of the viscosity dependence of both kinetic parameters, indicating that the motion of the alpha 9-helix is linked to catalysis in wild-type GST A1-1. The isomerization reaction with the alternative substrate Delta(5)-androstene-3,17-dione (AD) is affected in a similar manner by the viscosogens. The transition state energy of the isomerization reaction, like that of the CDNB conjugation, is lowered by Phe 220 as indicated by the effects of the mutations on K(cat)/K(m). The results demonstrate that Phe 220 and Phe 222, in the dynamic C-terminal segment, influence rate-determining steps in the catalytic mechanism of both the substitution and the isomerization reactions.     

Glutathione induces helical formation in the carboxy terminus of human glutathione transferase A1-1

The structure and dynamic properties of the C-terminal region of the human class alpha glutathione transferase A1-1 have been investigated with high-resolution NMR methods. On the basis of crystallographic and fluorescence measurements, this 13-residue segment of the enzyme is presumed to be disordered in the unliganded enzyme. When the product or product analogue is bound, a C-terminal alpha-helix is observed in crystal structures. Conflicting data exists regarding the structure of this region when one of the substrates, glutathione (GSH), is bound. The NMR studies presented here show that in the unliganded protein, this region of the protein samples different conformations, most likely an ensemble of helix-like structures. Addition of either GSH or the conjugate between GSH and ethacrynic acid (EASG) causes this segment to become a stable alpha-helix. In the GSH complex, the ends of this helix exhibit dynamic behavior on both the millisecond and nanosecond time scales. In contrast, there is no evidence of millisecond dynamics in the EASG complex. The ligand-induced ordering of the enzyme reduces the intrinsic affinity of the enzyme for its product, facilitating enzymatic turnover.     

Aromatic residues in the C-terminal region of glutathione transferase A1-1 influence rate-determining steps in the catalytic mechanism

Human glutathione transferase A1-1 (GST A1-1) has a flexible C-terminal segment that forms a helix (alpha9) closing the active site upon binding of glutathione and a small electrophilic substrate such as 1-chloro-2,4-dinitrobenzene (CDNB). In the absence of active-site ligands, the C-terminal segment is not fixed in one position and is not detectable in the crystal structure. A key residue in the alpha9-helix is Phe 220, which can interact with both the enzyme-bound glutathione and the second substrate, and possibly guide the reactants into the transition state. Mutation of Phe 220 into Ala and Thr was shown to reduce the catalytic efficiency of GST A1-1. The mutation of an additional residue, Phe 222, caused further decrease in activity. The presence of a viscosogen in the reaction medium decreased the kinetic parameters k(cat) and k(cat)/K(m) for the conjugation of CDNB catalyzed by wild-type GST A1-1, in agreement with the view that product release is rate limiting for the substrate-saturated enzyme. The mutations cause a decrease of the viscosity dependence of both kinetic parameters, indicating that the motion of the alpha9-helix is linked to catalysis in wild-type GST A1-1. The isomerization reaction with the alternative substrate Delta(5)-androstene-3,17-dione (AD) is affected in a similar manner by the viscosogens. The transition state energy of the isomerization reaction, like that of the CDNB conjugation, is lowered by Phe 220 as indicated by the effects of the mutations on k(cat)/K(m). The results demonstrate that Phe 220 and Phe 222, in the dynamic C-terminal segment, influence rate-determining steps in the catalytic mechanism of both the substitution and the isomerization reactions.     

Structures of thermolabile mutants of human glutathione transferase P1-1

An N-capping box motif (Ser/Thr-Xaa-Xaa-Asp) is strictly conserved at the beginning of helix alpha6 in the core of virtually all glutathione transferases (GST) and GST-related proteins. It has been demonstrated that this local motif is important in determining the alpha-helical propensity of the isolated alpha6-peptide and plays a crucial role in the folding and stability of GSTs. Its removal by site-directed mutagenesis generated temperature-sensitive folding mutants unable to refold at physiological temperature (37 degrees C). In the present work, variants of human GSTP1-1 (S150A and D153A), in which the capping residues have been substituted by alanine, have been generated and purified for structural analysis. Thus, for the first time, temperature-sensitive folding mutants of an enzyme, expressed at a permissive temperature, have been crystallized and their three-dimensional structures determined by X-ray crystallography. The crystal structures of human pi class GST temperature-sensitive mutants provide a basis for understanding the structural origin of the dramatic effects observed on the overall stability of the enzyme at higher temperatures upon single substitution of a capping residue.     

Characterization of porcine alpha-class glutathione transferase A1-1

An Alpha-class glutathione transferase (GST) has been cloned from pig gonads. In addition to two conservative point mutations our nucleotide sequence presents a frame shift resulting from a missing A as compared to a previously published porcine GST A1-1 sequence. The deduced C-terminal amino-acid segment of the protein differs between the two variants. Repeated sequencing of cDNA isolated from different tissues and animals ruled out the possibility of a cloning artifact, and the deduced amino acid sequence of our clone showed higher similarity to related mammalian GST sequences. Hereafter, we refer to our cloned enzyme as GST A1-1 and to the previously published enzyme as GST A1-1^∗. The study of the tissue distribution of the GSTA1 mRNA revealed high expression levels in many organs, in particular adipose tissue, liver, and pituitary gland. Porcine GST A1-1 was expressed in Escherichia coli and its kinetic properties were determined using alternative substrates. The catalytic activity in steroid isomerization reactions was at least 10-fold lower than the corresponding values for porcine GST A2-2, whereas the activity with 1-chloro-2,4-dinitrobenzene was approximately 8-fold higher. Differences in the H-site residues of mammalian Alpha-class GSTs may explain the catalytic divergence.     

Differences between bovine and human steroid double-bond isomerase activities of Alpha-class glutathione transferases selectively expressed in steroidogenic tissues

Bovine glutathione transferase A1-1 (bGST A1-1) and human GST A3-3 (hGST A3-3) share both high amino acid sequence similarity and selective expression in steroidogenic organs. hGST A3-3 is the most efficient steroid isomerase known in mammals, and is thought to catalyze isomerization reactions in the biosynthesis of steroid hormones. We observed that four out of five residues essential to the high steroid isomerase activity of hGST A3-3 are conserved in bGST A1-1. The bovine GST was cloned, heterologously expressed, and purified to homogeneity. Its specific activity towards classical GST substrates and two steroids, Delta(5)-androstene-3,17-dione and Delta(5)-pregnene-3,20-dione, was studied, and the steady-state kinetic parameters with the steroids were determined. We find that bGST A1-1 exhibits enzymatic activities comparable to those of hGST A3-3 towards non-steroid substrates. However, the bovine enzyme had 100 times lower catalytic efficiency in steroid isomerization reactions than the human GST. Nevertheless, bGST A1-1 was found as efficient as bovine 3beta-hydroxysteroid dehydrogenase as a steroid isomerase. We discuss likely reasons for the contrasting steroid isomerase activities of bGST A1-1 and hGST A3-3, and alternative roles of bGST A1-1.     

Role of cysteine residues in the activity of rat glutathione transferase P (7-7): elucidation by oligonucleotide site-directed mutagenesis

To clarify the role(s) of thiol (sulfhydryl) groups of cysteine (Cys) residues in the activity of the rat glutathione transferase P (7-7) form (GST-P), a cDNA clone, pGP5, containing the entire coding sequence of GST-P (Y. Sugioka et al., (1985) Nucleic Acids Res. 13, 6044-6057) was inserted into the expression vector pKK233-2 and the recombinant GST-P (rGST-P) expressed in E. coli JM109. All four Cys residues in rGST-P were independently substituted with alanine (Ala) by site-directed mutagenesis, the resultant mutants as well as the rGST-P being identical to GST-P purified from liver preneoplastic nodules with regard to molecular weight and immunochemical staining. Since all mutants proved as enzymatically active towards 1-chloro-2,4-dinitrobenzene as liver GST-P, it was indicated that none of the four Cys residues is essential for GST-P activity. However, the mutant with Ala at the 47th position from the N-terminus (Ala47) became resistant to irreversible inactivation by 0.1 mM N-ethylmaleimide (NEM), whereas the other three mutants remained as sensitive as the nonmutant type (rGST-P). Ala47 was also resistant to inactivation by the physiological disulfides, cystamine or cystine, which cause mixed disulfide and/or intra- or inter-subunit disulfide bond formation. These results suggest that the 47-Cys residue of GST-P may be located near the glutathione binding site, and modulation of this residue by thiol/disulfide exchange may play an important role in regulation of activity.     

The role of glutathione in the isomerization of delta 5-androstene-3,17-dione catalyzed by human glutathione transferase A1-1

Human glutathione transferase (GST) A1-1 efficiently catalyzes the isomerization of Delta(5)-androstene-3,17-dione (AD) into Delta(4)-androstene-3,17-dione. High activity requires glutathione, but enzymatic catalysis occurs also in the absence of this cofactor. Glutathione alone shows a limited catalytic effect. S-Alkylglutathione derivatives do not promote the reaction, and the pH dependence of the isomerization indicates that the glutathione thiolate serves as a base in the catalytic mechanism. Mutation of the active-site Tyr(9) into Phe significantly decreases the steady-state kinetic parameters, alters their pH dependence, and increases the pK(a) value of the enzyme-bound glutathione thiol. Thus, Tyr(9) promotes the reaction via its phenolic hydroxyl group in protonated form. GST A2-2 has a catalytic efficiency with AD 100-fold lower than the homologous GST A1-1. Another Alpha class enzyme, GST A4-4, is 1000-fold less active than GST A1-1. The Y9F mutant of GST A1-1 is more efficient than GST A2-2 and GST A4-4, both having a glutathione cofactor and an active-site Tyr(9) residue. The active sites of GST A2-2 and GST A1-1 differ by only four amino acid residues, suggesting that proper orientation of AD in relation to the thiolate of glutathione is crucial for high catalytic efficiency in the isomerization reaction. The GST A1-1-catalyzed steroid isomerization provides a complement to the previously described isomerase activity of 3beta-hydroxysteroid dehydrogenase.     

Structural characterization of acid-induced intermediates of human glutathione transferase P1-1

The acid denaturation of human glutathione transferase P1-1 (hGSTP1-1) has been performed to investigate the unfolding intermediates of the protein and their possible involvement in the refolding mechanism. The acid-induced structures of GSTP1-1 have been characterized by activity, gel filtration, intrinsic fluorescence and far-u.v. circular dichroism (CD) techniques. Because of the non-identity of the different transitions monitored, the acid denaturation of hGSTP1-1 appears to be a multistep process during which several intermediates coexist in equilibrium. The dependence of inactivation on the protein concentration, as well as gel-filtration experiments, indicate that the inactivation transition, centred at about pH 4.0, corresponds to the monomerization of the protein. At pH 2.0, when the enzyme is completely inactive, the protein retains a small, but significant, amount of secondary structure. This means that the dimeric arrangement of the molecule is important for maintaining the native-like secondary structure of the monomer. The results show that, at low pH, the compact state of the GST monomer, even upon the addition of salts, does not possess native-like secondary structure as described for many monomeric proteins (molten globule). In the presence of physiological concentrations of salts, the protein solution at pH 2.0 leads to a dead-end aggregation process, suggesting that this compact state cannot represent a productive intermediate of the refolding pathway.     

The trans-stilbene oxide-active glutathione transferase in human mononuclear leucocytes is identical with the hepatic glutathione transferase mu.

A glutathione transferase from human mononuclear leucocytes with high activity towards trans-stilbene oxide (GT-tSBO) was purified. GT-tSBO is expressed in only about 50% of the individuals studied. As judged from activity measurements, immunological studies and the fact that only those individuals who express glutathione transferase mu have high activity towards trans-stilbene oxide, it is concluded that the hepatic transferase mu is identical with the glutathione transferase (GT-tSBO) in mononuclear leucocytes.