Substrate specificity and kinetic properties of enzymes belonging to the hormone-sensitive lipase family: Comparison with non-lipolytic and lipolytic carboxylesterases

Biochimica et Biophysica Acta 1738 (2005) 29 – 36 http://www.elsevier.com/locate/bba

Substrate specificity and kinetic properties of enzymes belonging to the hormone-sensitive lipase family: Comparison with non-lipolytic and lipolytic carboxylesterases Henri Chahinian a,*, Yassine Ben Ali a, Abdelkarim Abousalham a, Stefan Petry b, Luigi Mandrich c, Guiseppe Manco c, Stephane Canaan a, Louis Sarda d a

Laboratoire d’Enzymologie Interfaciale et de Physiologie de la lipolyse, UPR9025-CNRS, 31 Chemin J. Aiguier, 13402 Marseille Cedex 20, France b Sanofi-Aventis Germany, 65926 Frankfurt am Main, Germany c Istituto di Biochimica delle Proteine, CNR, via P. Castellino 111, 80131 Napoli, Italy d Laboratoire de Biochimie, Universite´ de Provence, Centre St-Charles, 3 Place V. Hugo, 13331, Marseille Cedex 3, France Received 21 June 2005; received in revised form 16 September 2005; accepted 4 November 2005 Available online 17 November 2005

Abstract We have studied the kinetics of hydrolysis of triacylglycerols, vinyl esters and p-nitrophenyl butyrate by four carboxylesterases of the HSL family, namely recombinant human hormone-sensitive lipase (HSL), EST2 from Alicyclobacillus acidocaldarius, AFEST from Archeoglobus fulgidus, and protein RV1399C from Mycobacterium tuberculosis. The kinetic properties of enzymes of the HSL family have been compared to those of a series of lipolytic and non-lipolytic carboxylesterases including human pancreatic lipase, guinea pig pancreatic lipase related protein 2, lipases from Mucor miehei and Thermomyces lanuginosus, cutinase from Fusarium solani, LipA from Bacillus subtilis, porcine liver esterase and Esterase A from Aspergilus niger. Results indicate that human HSL, together with other lipolytic carboxylesterases, are active on short chain esters and hydrolyze water insoluble trioctanoin, vinyl laurate and olive oil, whereas the action of EST2, AFEST, protein RV1399C and non-lipolytic carboxylesterases is restricted to solutions of short chain substrates. Lipolytic and non-lipolytic carboxylesterases can be differentiated by their respective value of K 0.5 (apparent K m) for the hydrolysis of short chain esters. Among lipolytic enzymes, those possessing a lid domain display higher activity on tributyrin, trioctanoin and olive oil suggesting, then, that the lid structure contributes to enzyme binding to triacylglycerols. Progress reaction curves of the hydrolysis of p-nitrophenyl butyrate by lipolytic carboxylesterases with lid domain show a latency phase which is not observed with human HSL, non-lipolytic carboxylesterases, and lipolytic enzymes devoid of a lid structure as cutinase. D 2005 Elsevier B.V. All rights reserved. Keywords: Carboxylesterase; EST2; AFEST; Hormone-sensitive lipase; HSL family; Kinetic study

1. Introduction Carboxylesterases (carboxylesterases) include two classes of enzymes widely represented in all living organisms, commonly designated as esterases and lipases. Esterases and lipases were first differentiated on the basis of their substrate specificity. It was early established that esterase activity was restricted to aqueous solutions of short acyl chain esters whereas lipases specifically hydrolyzed emulsions of water insoluble triacylglycerols [1]. Later, an alternative classification of carbox* Corresponding author. Fax: +33 4 91 16 45 36. E-mail address: [email protected] (H. Chahinian). 1388-1981/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.bbalip.2005.11.003

ylesterases based on their amino acid sequence homology has been proposed. According to Hemila et al. [2], enzymes of the esterase/lipase superfamily have been divided into three families, namely the LPL family, which includes lipoprotein lipase, hepatic and pancreatic lipases, the EST family (cholinesterase and lipases from Geotrichum candidum and Candida rugosa) and the hormone-sensitive lipase (HSL) family. The HSL family includes mammalian HSL and several proteins of bacterial origin among which are brefeldin A esterase (BFAE) from Bacillus subtilis [3], heroin esterase from Rhodococcus sp. strain H1 [4], acetyl hydrolase from Streptomyces viridochromogenes [5], lipase 2 from Moraxella TA 114 [6,7] and two thermophilic carboxylesterases, namely

30

H. Chahinian et al. / Biochimica et Biophysica Acta 1738 (2005) 29 – 36

EST2 from Alicyclobacillus acidocaldarius [8] and AFEST from Archeoglobus fulgidus [9]. More recently, several other proteins of the HSL family, as the protein Rv1399c protein from Mycobacterium tuberculosis (protein Rv1399c) have been biochemically characterized [10]. Enzymes of the HSL family contain the consensus sequence Gly-X-Ser-X-Gly containing the active residue of serine which, together with an histidine and an aspartic, or glutamic acid, form the charge relay network of the catalytic triad common to all carboxylesterases and serine proteases. Apart from this sequence, they contain the His – Gly – Gly – Gly sequence pattern located at about 70– 100 amino acids upstream of the catalytic serine. This sequence contributes to the formation of the so-called oxyanion hole. Members of the HSL family have an average molecular mass of 34– 40 kDa whereas that of HSL is in the range of 85 –120 kDa. A two-domain structure has been proposed for HSL [11]. The N-terminal domain, which contains around 300 amino acids, shows no sequence homology to any other protein. It is generally assumed that this domain mediates the interaction of HSL with adipocyte lipid-binding protein [12]. The C-terminal domain of HSL (around 450 residues), known as the catalytic domain, contains the residues of the catalytic triad [13,14]. It shows significant identity with carboxylesterases of the HSL family of lower molecular mass. In HSL, this region contains a large sequence stretch of about 150 amino acids known as the regulatory module where are located the phosphorylation sites of the enzyme. Recent studies have suggested that specific sequences within the Cterminal region of the regulatory module of HSL is involved in a lipid pocket and is crucial for determining lipolytic activity [15]. The three-dimensional (3D) structure of several bacterial enzymes of the HSL family including BFAE [3], EST2 [16] and AFEST [9], have been determined in the recent years. Crystallographic investigations have identified a common protein fold shared with other hydrolytic enzymes, known as the a/h hydrolase fold. This particular structure is characterized by a central parallel h-sheet core domain, surrounded on both sides by a-helices, which contain the residues of the catalytic triad. In addition, the 3D-structure of BFAE, EST2 and AFEST shows the presence of two separate helical regions, external to the a/h core, that cover the active site in such a way that free access of a long chain triacylglycerol molecule is impaired. This structure is reminiscent of the lid domain found in well-characterized long chain triacylglycerol hydrolyzing enzymes as fungal lipases from Thermomyces lanuginosus (TLL) or Mucor miehei (MML), and mammalian lipases of the LPL family [17 – 19]. As yet, no crystal structure of HSL has been reported and no putative lid domain has been localized in the protein. However, recent studies on the inhibition of HSL by specific serine enzyme inhibitors, as diethyl-p-nitrophenyl phosphate (E600), or phenylmethylsulfonyl fluoride (PMSF), however, have shown that HSL, in sharp contrast to classical lipases possessing a lid preventing free access to active serine, reacts with these inhibitors in the absence of micelles of detergent, which

suggests that HSL lacks this domain [20]. On the other hand, superimposition of the crystallograhic structures of a mutant of EST2 and of the lipase from Candida rugosa (CRL), in its closed conformation, shows the concomitant occurrence of the CRL lid and the N-terminal region of EST2. In both proteins, these structures cover the active site but in opposite relative orientation and it is generally agreed that the two structures have different functions [21]. For EST2 from A. Acidocaldarius, it was recently reported that a deletion of 35 amino acids at the N-terminus has a dramatic effect on the kinetic properties of the enzymes [22]. To date, only few comparative kinetic studies of the carboxylesterases of the HSL family have been reported. In this study, we report the kinetic properties of four enzymes of the HSL family, namely recombinant human HSL, EST2, AFEST and protein Rv1399c. Results obtained with triacylglycerols and vinyl esters substrates have been compared to those collected with various lipolytic and non-lipolytic carboxylesterases including human pancreatic lipase (HPL) and guinea pig pancreatic lipase related protein 2 (GPL-RP2), lipases from Mucor miehei and Thermomyces lanuginosus, cutinase from Fusarium solani, LipA from Bacillus subtilis, porcine liver esterase and Esterase A from Aspergilus niger, using triacylglycerols, vinyl esters and p-nitrophenyl butyrate ( p-NPB) as substrates. Our results show that EST2 and AFEST display substrate specificity and kinetic properties similar to those of nonlipolytic carboxylesterases, in contrast to human HSL whose kinetic behaviour is similar to that of lipolytic enzymes. They further confirm that non-lipolytic and lipolytic carboxylesterases can be differentiated by their apparent K m (K 0.5) for the hydrolysis of short acyl chain ester substrates [23]. 2. Materials and methods 2.1. Chemicals Triacylglycerols (triacetin, tripropionin, tributyrin and trioctanoin) were supplied by Acros-Organics (F-9316 Noisy-le-Grand, France). Olive oil was from local origin. Vinyl esters (vinyl acetate, vinyl propionate, vinyl butyrate and vinyl laurate) and p-NPB were purchased from Sigma-Aldrich-Fluka (F38297 St-Quentin-Fallavier, France). The solubility limit of triacetin, tripropionin and tributyrin, in 2.5 mM Tris – HCl buffer, pH 7.5, with 0.1 M NaCl, is 330 mM, 10 mM and 0.4 mM, respectively, and that of vinyl acetate, vinyl propionate and vinyl butyrate is 315 mM, 86 mM and 22 mM, respectively [23]. The solubility limit of p-NPB is 1 mM [24].

2.2. Enzymes Recombinant human HSL was expressed and purified from baculovirus infected insect cells as previously described [20]. Protein Rv1399c [10] was prepared by Dr. S. Canaan at the laboratory. Pure EST2 and AFEST were produced at the Institute of Protein Biochemistry, Napoli, Italy [8,9]. Protein Esterase A (EstA) from Aspergillus niger [25] was a gift of Dr. Y. Bourne (CNRS, Marseille, France). Porcine liver esterase (PLEst) was from SigmaAldrich-Fluka. HPL, GPL-RP2 and porcine colipase were provided by Dr F. Carriere (CNRS, Marseille, France). Recombinant Fusarium solani cutinase was a gift of Dr. M. Egmond (Utrecht University, The Netherlands). MML and TLL were provided by Dr. S. Patkar (Novo Nordisk A/S, Bagsvaerd, Denmark). Bacillus subtilis LipA was a gift from Dr. W. Quax (Groningen, The Netherlands).

H. Chahinian et al. / Biochimica et Biophysica Acta 1738 (2005) 29 – 36

2.3. Determination of enzyme activity Enzymatic hydrolytic activity against triacylglycerols, vinyl esters and pNPB were determined potentiometrically at 25 -C under the standard conditions described in a previous report [26]. Hydrolysis of p-NPB was measured at pH 7.5. All assays were performed in 15 mL of 2.5 mM Tris – HCl buffer containing 0.1 M NaCl. At substrate concentration exceeding the solubility limit, the esters were emulsified by mechanical stirring. No detergent was added except in the case of olive oil which was previously emulsified with gum arabic [27]. The pH dependence curves of carboxylesterase activity against p-NPB were studied from pH 4.5 to pH 9. Correction was made for the partial dissociation of butanoic acid and p-nitrophenol assuming values of pK a of 4.55 and 7.15, respectively. The ester concentration corresponding to half maximal activity (apparent K m or K 0.5) was determined by plotting the rate of hydrolysis (v) against the substrate concentration (S). Values of S of solutions and emulsions of esters were arbitrarily expressed as mmoles per liter (mM). Enzyme activity was expressed as units. One enzyme unit corresponds to the release of one micromole of acid per minute. Specific activity of the carboxylesterases were determined at optimal concentration of substrate and expressed as units per milligram of protein. For each enzyme, the value of the catalytic rate constant, k cat (s 1), was calculated from the specific activity and molecular mass. Values in the figures and tables correspond to mean values of two independent assays with a deviation less than 5%. The protein

31

Table 1 Kinetic parameters k cat and K 0.5 of the hydrolysis of short chain triacylglycerols by members of the HSL family and other lipolytic and non-lipolytic carboxyesterases Enzymes

PLEst EstA EST2 AFEST Rv1399c HSL HPL MML TLL GPL-RP2 Cutinase LipA

Substrates Triacetin (C2)

Tripropionin (C3)

Tributyrin (C4)

k cat

K 0.5

k cat

K 0.5

k cat

K 0.5

60 265 945 40 100a 0 8 10 0 555 115 120

3 15 12 10 3a – 190 210 – 120 70 110

50 60 835 110 715a 7 3320 200 225 750 210 150

0.30 0.35 0.75 0.30 0.30 15 8 18 14 10.5 10 12

70 20 1290 95 245a 18 7885 1080 2250 1500 130 100

0.05 nd 0.07 0.05 0.05 0.40 0.40 0.45 0.50 0.45 0.35 0.40

k cat and K 0.5 are expressed as s a Data from reference [10].

1

and mM, respectively. nd: not determined.

concentration of enzyme samples was estimated by the colorimetric method of Bradford with bovine serum albumin as reference protein [28].

3. Results 3.1. Kinetics of hydrolysis by enzymes of the HSL family. Comparison with other carboxylesterases 3.1.1. Hydrolysis of triacylglycerols The variation of the initial rate (v) of hydrolysis of tripropionin by EST2 as a function of ester concentration (S) is shown in Fig. 1A. It can be noticed from the v/S curve that EST2 is fully active at low substrate concentration, well below the solubility limit of the ester. In contrast, as reported previously [20], HSL is only partially active on soluble tripropionin and displays maximal activity against emulsion. The values of K 0.5 (apparent K m), estimated for the v/S curves, are 0.75  10 3 M and 15  10 3 M for EST2 and HSL, respectively. Values of k cat for EST2 and HSL are 835 s 1 and 7 s 1, respectively. k cat and K 0.5 were also determined from the v/S curves of the hydrolysis of triacetin and tributyrin by EST2 (data not shown) and compared to those obtained with AFEST, protein Rv1399c, PLEst and EstA. Data are reported in Table 1. Table 1 also includes the values of k cat and K 0.5 determined for HPL, GPL-RP2, MML, TLL, cutinase and LipA. The latter enzymes, together with HSL, in contrast to EST2, AFEST, Rv1399c, PLEst and EstA, hydrolyze emulsions of trioctanoin and olive oil (Table 2).

Fig. 1. Effect of substrate concentration on the rate of hydrolysis of (A) tripropionin and (B) vinyl acetate by EST2. Enzyme activity was measured at 25 -C and pH 7.5 in 2.5 mM Tris – HCl buffer with 0.1 M NaCl under conditions described in Materials and methods. Activity is expressed as percentage of maximal activity (V max). The solubility limit of tripropionin (10 mM) is indicated by the dotted line whereas that of vinyl acetate (315 mM) is not indicated.

3.1.2. Hydrolysis of vinyl esters Kinetic studies were performed with carboxylesterases against vinyl acetate, vinyl propionate, vinyl butyrate and vinyl laurate. The v/S curve of the hydrolysis of vinyl acetate by EST2 is shown in Fig. 1B. This curve, as v/S curve of Fig. 1A, do not represent proper hyperbolic curves pointing then to complexed kinetic systems as previously reported [32]. Curves obtained with vinyl acetate, vinyl

32

H. Chahinian et al. / Biochimica et Biophysica Acta 1738 (2005) 29 – 36

Table 2 Hydrolysis of trioctanoin and olive oil by lipolytic carboxyesterases Enzymes

respectively. Fig. 2B shows the v/S curves found with HSL, TLL and GPL-RP2. Values of K 0.5 for these enzymes and for MML and cutinase (data not shown) are 1.0  10 3 M, 2.5  10 3 M, 1.0  10 3 M, 2.1 10 3 M and 0.4  10 3 M, respectively. Kinetic parameters for the enzymes of the HSL family and other carboxylesterases are listed in Table 4. Table 4 does not include values for EstA and HPL which show no detectable activity against p-NPB under the same assay conditions.

Substrates

PLEst EstA EST2 AFEST Rv1399ca HSL HPL MML TLL GPL-RP2 Cutinase LipA

Trioctanoin (C8)

Olive oil (C18)

k cat

k cat

0 0 0 0 0 6 5400 3450 2900 130 305 240

0 0 0 0 0 6 3360 1620 1885 375 245 165

3.2. Analysis of the hydrolysis progress curves by carboxylesterases Here, we have compared the reaction progress curves for the hydrolysis of p-NPB, vinyl butyrate and tripropionin by carboxylesterases. Fig. 3A shows the time course of the reaction progress of the hydrolysis of p-NPB by TLL, determined at optimal ester concentration (V max) and at ester concentration corresponding to V max/2. In both cases, the kinetic curves show a progressive increase in hydrolytic rate until a steady state reaction is reached, after 5 to 6 min reaction. The lag period, s, calculated from the intercept between the asymptote of the progress curve and the x axis, is around 1.5 min. The value of s depends neither on the substrate concentration nor on the amount of enzyme used in the assays (data not shown). The same latency phase was observed in the hydrolysis of vinyl butyrate and tripropionin by TLL (data not shown). MML showed the same kinetic behaviour as TLL (data not shown). The reaction progress curves for the hydrolysis of p-NPB by HSL are reported in Fig. 3B. The kinetics of hydrolysis of pNPB, tripropionin and vinyl butyrate by HSL show no latency period. This observation indicates that the latency phase observed with TLL and MML during titrimetric determination of activity is not artefactual but reflects a difference in the behaviour of the carboxylesterases during the pre-steady state phase of enzymatic hydrolysis. Kinetics of hydrolysis by EST2, AFEST, protein Rv1399c and other non-lipolytic

k cat is expressed as s 1. a Data from reference [10].

propionate and vinyl butyrate for EST2, AFEST and protein Rv1399c are of the same type (data not shown). The curves representative of the effect of ester concentration on the rate of hydrolysis of short acyl chain vinyl esters by HSL have been previously reported [20]. The kinetic parameters, k cat and K 0.5, for the hydrolysis of the vinyl esters by HSL and other members of the HSL family are reported in Table 3, together with the values of the kinetic parameters of other non-lipolytic and lipolytic carboxylesterases. 3.1.3. Hydrolysis of p-nitrophenyl butyrate Kinetic studies of the hydrolysis of p-NPB were performed with HSL, EST2, AFEST, protein Rv1399c in comparison with PLEst, TLL, MML, GPL-RP2, cutinase and LipA. The v/S curves for EST2 and PLEst are shown in Fig. 2A. Values of K 0.5 for the two enzymes are 0.05  10 3 M and 0.08  10 3 M, respectively. Values of K 0.5, calculated from the v/S curves of the hydrolysis of p-NPB by AFEST and protein Rv1399c (data not shown), are 0.045  10 3 M and 0.05  10 3 M,

Table 3 Kinetic parameters k cat and K 0.5 of the hydrolysis of vinyl esters by members of the HSL family and other lipolytic and non-lipolytic carboxyesterases Enzymes

PLEst EstA EST2 AFEST Rv1399c HSL HPL MML TLL GPL-RP2 Cutinase LipA

Substrates Vinyl acetate (C2)

Vinyl propionate (C3)

Vinyl butyrate (C4)

Vinyl laurate (C12)

k cat

K 0.5

k cat

K 0.5

k cat

K 0.5

k cat

K 0.5

315 400 110 50 320a 140 170 950 455 70 240 100

4 3 2.5 4 8a 170 315 120 280 295 140 280

295 160 980 200 855a 105 800 3675 1950 270 255 1935

5 3 6 7 15 70 75 35 100 50 40 75

460 25 1930 210 590a 210 625 5390 2600 200 50 635

2 2 3 2.5 2 11 25 15 20 20 12 40

0 0 0 0 0a 31 375 5400 12,000 125 285 145

– – – – – 3 4 2 4 1.25 5 5

k cat and K 0.5 are expressed as s a Data from reference [10].

1

and mM, respectively.

H. Chahinian et al. / Biochimica et Biophysica Acta 1738 (2005) 29 – 36

33

latency phase (data not shown) although, as mentioned above, these lipolytic enzymes, as TLL and MML, have the capacity to hydrolyze water-insoluble middle and long acyl chain esters, such as trioctanoin, olive oil and vinyl laurate. In this respect, it may be observed that crystallographic studies have clearly established that the lid domain that covers the active site in classical lipases is absent in cutinase and LipA and that the lid domain in the guinea pig enzyme is restricted to 5 amino acid residues instead of 23, 15 and 12 amino acids in HPL, TLL and MML, respectively. 3.3. pH-dependence of the hydrolysis of p-nitrophenyl butyrate and vinyl butyrate by carboxylesterases The pH-activity profiles of HSL,TLL, EST2, AFEST and protein Rv1399c, studied with p-NPB as substrate, is shown in Fig. 4. Curves of Fig. 4 show that comparable pH-activity profiles are obtained for all carboxylester hydrolyses. In all cases, activity is maximal at slightly basic pH with peak activity between pH 7.0 and pH 9. The curves of Fig. 4 further indicate that the enzymatic hydrolysis is dependent on the

Fig. 2. Effect of substrate concentration on the rate of hydrolysis of p-NPB by (A) EST2 (black circles) and PLEst (open circles) and by (B) human HSL (open squares), TLL (black triangles) and GPL-RP2 (black squares). Enzyme activity was determined at 25 -C and pH 7.5 in 2.5 mM Tris – HCl buffer with 0.1 M NaCl under conditions described in Materials and methods. The solubility limit of p-NPB (1 mM) is indicated by the dotted line.

carboxylesterases, such as PLEst and EstA, are identical to those found with HSL (data not shown). Interestingly, the kinetics of hydrolysis of p-NPB, tripropionin and vinyl butyrate by cutinase, GPL-RP2 and LipA do not show a Table 4 Kinetic parameters k cat and K 0.5 of the hydrolysis of p-NPB by members of the HSL family Enzymes

PLEst EST2 AFEST Rv1399c HSL MMP TLL GPL-RP2 Cutinase LipA

p-NPB (C4) k cat

K 0.5

18 480 140 185 190 420 580 50 525 50

0.08 0.05 0.04 0.05 1 2.1 2.5 1 0.4 0.75

Comparison with other lipolytic and non-lipolytic carboxylesterases. k cat and K 0.5 are expressed as s 1 and mM, respectively.

Fig. 3. Progress curves of the hydrolysis of p-NPB by (A) TLL and (B) HSL. Curves (a) represent the kinetics of hydrolysis of p-NPB at optimal ester concentration corresponding to ten-fold K 0.5, namely 25 mM for TLL and 10 mM for HSL. Curves (b) show kinetics of hydrolysis at p-NPB concentration corresponding to K 0.5 (see Table 3).

34

H. Chahinian et al. / Biochimica et Biophysica Acta 1738 (2005) 29 – 36

lower pK a value (pK a: 5.30) is found for protein Rv1399c. Similar values of pK a were estimated from the pH-activity profiles of the carboxylesterases studied with vinyl butyrate as substrate (data not shown). 4. Discussion

Fig. 4. Effect of pH on the rate of hydrolysis of p-NPB by HSL (black circles), TLL (black squares), EST2 (open squares), AFEST (open triangles) and Rv1399c (black lozenges). Activity was determined titrimetrically in 15 mL of 2.5 mM Tris buffer, at pH ranging from 4.5 to 8.5, on a solution of p-NPB at concentration of 0.9 mM in the case of EST2, AFEST and Rv1399c and on emulsion at the concentration of 2 mM with HSL and TLL. Activity is expressed relatively to maximal activity. Correction was made for partial dissociation of butanoic acid and p-nitrophenol assuming pK a values of 4.55 and 7.15, respectively.

unprotonated form of an essential amino acid residue, likely histidine, with an apparent pK a around 6. The pK a values for HSL, TLL, EST2 and AFEST, estimated graphically from the curves of Fig. 4, are 6.0, 6.15, 5.85 and 5.70, respectively. A

Carboxylesterases differ by their substrate specificity and kinetic properties. Among these enzymes, a number of proteins of microbial origin, which are structurally related to HSL, constitute the HSL family. The alignment of the amino acid sequence from the active serine to the histidine residue of the catalytic triad of EST2, AFEST, and protein Rv1399c with that of the catalytic domain of rat and human HSL shows high sequence identity (Fig. 5). It is noteworthy that, in all these enzymes, a constant number of thirty-three amino acids are found between the aspartic acid and histine of the catalytic triad. So far, few comparative studies of the biochemical properties of carboxylesterases of the HSL family have been reported. In the present work, we have studied the kinetic properties of EST2, AFEST, protein Rv1399c and recombinant human HSL. Hydrolytic activity was determined against short, middle and long acyl chain triacylglycerols and vinyl esters and compared to that of various carboxylesterases including PLEst, EstA, HPL, TLL, MML, GPL-RP2, cutinase and LipA, under conditions previously used in kinetic studies carried out to differentiate non-lipolytic from lipolytic carboxylesterases [23]. We have also investigated the kinetic behaviour of the above enzymes against p-NPB. Table 1 clearly shows that short chain triacylglycerols are hydrolyzed by all carboxylesterases but at

Fig. 5. Amino acid sequence alignment of the partial sequence of five homologous proteins belonging to the HSL family which include EST2, AFEST, protein Rv1399c, rat HSL and human HSL. For each protein, the aligned sequence regions include the residues of serine, aspartic acid and histidine of the catalytic triad (indicated with arrowheads). Identical amino acids are indicated in bold letters. The phosphorylation sites in rat and human HSL catalytic domain sequence are indicated by asterisks.

H. Chahinian et al. / Biochimica et Biophysica Acta 1738 (2005) 29 – 36

varied rates. It can be noticed that the values of K 0.5 for EST2, AFEST and protein Rv1399c are very similar to those found for the non-lipolytic carboxylesterases PLEst and EstA and differ markedly from those of the lipolytic carboxylesterases HPL, TLL, MML, GPL-RP2, cutinase and LipA. In contrast to lipolytic carboxylesterases, EST2, AFEST and Rv1399c are inactive against water-insoluble trioctanoin and olive oil (Table 2). Although HSL displays low activity against triacylglycerols (Tables 1 and 2), it must be noticed that the values of K 0.5 for the hydrolysis of tripropionin and tributyrin by HSL are comparable to those of classical lipolytic enzymes (lipases). Parallel studies of the kinetic properties of carboxylesterases against vinyl esters are reported in Table 3. Again, it appears that the K 0.5 values for the hydrolysis of short chain vinyl esters by EST2, AFEST and protein Rv1399c are similar to those found for PLEst and EstA. The values of K 0.5 for all lipolytic carboxylesterases, including HSL, are significantly higher than those found for the nonlipolytic carboxylesterases. Taken together, these results fully confirm that values of K 0.5 for the hydrolysis of short chain triacylglycerols and vinyl esters can be taken into account to differentiate non-lipolytic from lipolytic carboxylesterases. Among the members of the HSL family, HSL can be classified unambiguously as a lipolytic enzyme whereas EST2, AFEST and Rv1399c are non-lipolytic carboxylesterases. It can be further observed from data shown in Tables 1 and 2 that the values of k cat of the hydrolysis of tributyrin, trioctanoin and olive oil by HSL, cutinase and LipA, which are devoid of the lid structure, are markedly lower than those of HPL, TLL and MML which possess a lid. Then, it appears that, in lipolytic enzymes, the presence of the lid domain is correlated to the capacity of the enzymes to catalyze hydrolysis of emulsions of triacylglycerols at high rate. This conclusion is in accordance with the previous observations suggesting that, in lipases in their active conformation, the open lid regulates substrate specificity [29 – 31]. Table 4 shows the results of kinetic studies of the hydrolysis of p-NPB by carboxylesterases. Although the solubility limit of p-NPB is as low as 1 mM, this ester is suitable to differentiate non-lipolytic from lipolytic carboxylesterases. As shown above, in the case of short acyl chain triacylglycerols and vinyl esters, the values of K 0.5 for the hydrolysis of p-NPB by non-lipolytic carboxylesterases are far lower than those found for lipolytic enzymes. Kinetic studies with carboxylesterases further show the existence of a lag period s of around 1.5 min in the time course of hydrolysis of short chain triacylglycerols and vinyl esters and p-NPB by TLL (Fig. 3A) and MML (data not shown). The same s value is observed in kinetics carried out at substrate concentration corresponding to maximal and half-maximal rate of hydrolysis. The lag period is independent of the amount of enzyme used in assays. Progress reaction curves with human HSL (Fig. 3B), cutinase, LipA and GPL-RP2 (data not shown) do not show a lag period. Likewise, no lag period is found with EST2, AFEST, protein Rv1399c, PLEST and EstA. Taken together, these observations suggest that the lag period observed with classical lipases possessing a lid likely corresponds to the pre-steady state of the hydrolysis reaction and

35

reflect sequential events such as the binding of the lipolytic enzyme at interface and/or transition from an inactive closed conformation to an open active form upon rearrangement of the lid domain [32]. It is worth noticing that cutinase and LipA are devoid of the lid and that, in GPL-RP2, the lid consists only of a short peptide loop (mini-lid). As reported recently, human HSL reacts with specific inhibitors of the serine hydrolases as diethyl-p-nitrophenyl phosphate (E600) in the absence of interface which is an evidence that the enzyme lacks the lid domain [20]. It further appears, then, that the helical regions surrounding the active site of the 3-D structure of EST2 and AFEST and the lid found of classical lipases have different functions. Finally, we have compared the pH-dependence activity of the carboxylesterases of the HSL family to that of TLL (Fig. 4). It appears that the activity of both non-lipolytic and lipolytic carboxylesterases depends on the deprotonation of an ionizable group, likely the histidine residue of the catalytic triad, with a pK a around 6.0, except for protein Rv1399c which shows a more acidic pK a value (5.30). A low pK a value has been previously reported in the case of protein Rv1399c [10]. Following the proposed model of the 3D-structure of protein Rv1399c, the presence of the side-chain of the arginine residue at position 213 at close vicinity of the histidine residue of the catalytic triad (His290) might account for the low pK a of this residue. Actually, it can been observed, from the comparison of the sequence of members of the HSL family (Fig. 5), that this arginine residue is changed for a glycine, a glutamine or a leucine in EST2, AFEST and HSL, respectively. In conclusion, studies of the substrate specificity and kinetic properties of carboxylesterases structurally related to HSL, in comparison with other enzymes of the esterase/lipase superfamily, have shown that HSL displays biochemical characteristics similar to those of lipolytic enzymes whereas HSL homologs have the same kinetic behaviour as typical nonlipolytic carboxylesterases. In these studies, we have observed that the kinetics of hydrolysis of short acyl chain esters by lipolytic enzymes possessing a lid show a latency phase which is observed neither with lipolytic enzymes devoid of the lid domain or human HSL nor with non-lipolytic carboxylesterases. Further studies on the mode of action of lipolytic carboxylesterases will help to understand whether this lag period is related to the opening of the lid upon interaction of enzymes with organized molecules of lipid substrate. Acknowledgements The authors are indebted to Dr. R. Verger and F. Carrie`re from the Laboratoire d’Enzymologie Interfaciale et de Physiologie de la Lipolyse, CNRS, Marseille, France, for fruitful discussions and help during the preparation of the manuscript. References [1] T. Tsujita, K. Shirai, Y. Saito, H. Okuda, Isozymes: Structures, Function and Use in Biology and Medicine, Wiley-Liss, New York, 1990, pp. 915 – 933.

36

H. Chahinian et al. / Biochimica et Biophysica Acta 1738 (2005) 29 – 36

[2] H. Hemila, T.T. Koivula, I. Palva, Hormone-sensitive lipase is closely related to several bacterial proteins, and distantly related to acetylcholinesterase and lipoprotein lipase: identification of a superfamily of esterases and lipases, Biochim. Biophys. Acta 1210 (1994) 249 – 253. [3] Y. Wei, J.A. Contreras, P. Sheffield, T. Osterlund, U. Derewenda, R.E. Kneusel, U. Matern, C. Holm, Z.S. Derewenda, Crystal structure of brefeldin a esterase, a bacterial homolog of the mammalian hormonesensitive lipase, Nat. Struct. Biol. 6 (1999) 340 – 345. [4] X. Zhu, N.A. Larsen, A. Basran, N.C. Bruce, I.A. Wilson, Observation of an arsenic adduct in an acetyl esterase crystal structure, J. Biol. Chem. 278 (2003) 2008 – 2014. [5] W. Wohlleben, W. Arnold, I. Broer, D. Hillemann, E. Strauch, A. Puhler, Nucleotide sequence of the phosphinothricin N-acetyltransferase gene from Streptomyces viridochromogenes Tu494 and its expression in Nicotiana tabacum, Gene 70 (1988) 25 – 37. [6] D. Langin, H. Laurell, L.S. Holst, P. Belfrage, C. Holm, Gene organization and primary structure of human hormone-sensitive lipase: possible significance of a sequence homology with a lipase of Moraxella TA144, an antarctic bacterium, Proc. Natl. Acad. Sci. U. S. A. 90 (1993) 4897 – 4901. [7] G. Feller, M. Thiry, C. Gerday, Nucleotide sequence of the lipase gene lip2 from the antarctic psychrotroph Moraxella TA144 and site-specific mutagenesis of the conserved serine and histidine residues, DNA Cell. Biol. 10 (1991) 381 – 388. [8] G. Manco, E. Adinolfi, F.M. Pisani, G. Ottolina, G. Carrea, M. Rossi, Overexpression and properties of a new thermophilic and thermostable esterase from Bacillus acidocaldarius with sequence similarity to hormonesensitive lipase subfamily, Biochem. J. 332 (Pt. 1) (1998) 203 – 212. [9] G. De Simone, V. Menchise, G. Manco, L. Mandrich, N. Sorrentino, D. Lang, M. Rossi, C. Pedone, The crystal structure of a hyper-thermophilic carboxylesterase from the archaeon Archaeoglobus fulgidus, J. Mol. Biol. 314 (2001) 507 – 518. [10] S. Canaan, D. Maurin, H. Chahinian, B. Pouilly, C. Durousseau, F. Frassinetti, L. Scappuccini-Calvo, C. Cambillau, Y. Bourne, Expression and characterization of the protein Rv1399c from Mycobacterium tuberculosis. A novel carboxyl esterase structurally related to the HSL family, Eur. J. Biochem. 271 (2004) 3953 – 3961. [11] T. Osterlund, B. Danielsson, E. Degerman, J.A. Contreras, G. Edgren, R.C. Davis, M.C. Schotz, C. Holm, Domain – structure analysis of recombinant rat hormone-sensitive lipase, Biochem. J. 319 (1996) 411 – 420. [12] W.J. Shen, Y. Liang, R. Hong, S. Patel, V. Natu, K. Sridhar, A. Jenkins, D.A. Bernlohr, F.B. Kraemer, Characterization of the functional interaction of adipocyte lipid-binding protein with hormone-sensitive lipase, J. Biol. Chem. 276 (2001) 49443 – 49448. [13] C. Holm, R.C. Davis, T. Osterlund, M.C. Schotz, G. Fredrikson, Identification of the active site serine of hormone-sensitive lipase by site-directed mutagenesis, FEBS Lett. 344 (1994) 234 – 238. [14] T. Osterlund, J.A. Contreras, C. Holm, Identification of essential aspartic acid and histidine residues of hormone-sensitive lipase: apparent residues of the catalytic triad, FEBS Lett. 403 (1997) 259 – 262. [15] J. Wang, W.J. Shen, S. Patel, K. Harada, F.B. Kraemer, Mutational analysis of the ‘‘regulatory module’’ of hormone-sensitive lipase, Biochemistry 44 (2005) 1953 – 1959. [16] G. De Simone, S. Galdiero, G. Manco, D. Lang, M. Rossi, C. Pedone, A snapshot of a transition state analogue of a novel thermophilic esterase

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

belonging to the subfamily of mammalian hormone-sensitive lipase, J. Mol. Biol. 303 (2000) 571 – 761. D.M. Lawson, A.M. Brzozowski, S. Rety, C. Verma, G.G. Dodson, Probing the nature of substrate binding in Humicola lanuginosa lipase through X-ray crystallography and intuitive modelling, Protein Eng. 7 (1994) 543 – 550. L. Brady, A.M. Brzozowski, Z.S. Derewenda, E. Dodson, G. Dodson, S. Tolley, J.P. Turkenburg, L. Christinasen, B. Huge-Jensen, L. Norskov, L. Thim, U. Menge, A serine protease triad forms the catalytic centre of a triacylglycerol lipase, Nature 343 (1990) 767 – 770. C.S. Wang, J. Hartsuck, W.J. McConathy, Structure and functional properties of lipoprotein lipase, Biochim. Biophys. Acta 1123 (1992) 1 – 17. Y. Ben Ali, H. Chahinian, S. Petry, G. Muller, F. Carriere, R. Verger, A. Abousalham, Might the kinetic behavior of hormone-sensitive lipase reflect the absence of the lid domain? Biochemistry 43 (2004) 9298 – 9306. G. De Simone, V. Menchise, V. Alterio, L. Mandrich, M. Rossi, G. Manco, C. Pedone, The crystal structure of an EST2 mutant unveils structural insights on the H group of the carboxylesterase/lipase family, J. Mol. Biol. 343 (2004) 137 – 146. L. Mandrich, L. Merone, M. Pezzullo, L. Cipolla, F. Nicotra, M. Rossi, G. Manco, Role of the N terminus in enzyme activity, stability and specificity in thermophilic esterases belonging to the HSL family, J. Mol. Biol. 345 (2005) 501 – 512. H. Chahinian, L. Nini, E. Boitard, J.P. Dubes, L.C. Comeau, L. Sarda, Distinction between esterases and lipases: a kinetic study with vinyl esters and TAG, Lipids 37 (2002) 653 – 662. K. Shirai, R.L. Jackson, Lipoprotein lipase-catalyzed hydrolysis of pnitrophenyl butyrate. Interfacial activation by phospholipid vesicles, J. Biol. Chem. 257 (1982) 1253 – 1258. Y. Bourne, A.A. Hasper, H. Chahinian, M. Juin, L.H. De Graaff, P. Marchot, Aspergillus niger protein EstA defines a new class of fungal esterases within the alpha/beta hydrolase fold superfamily of proteins, Structure (Camb.) 12 (2004) 677 – 687. H. Chahinian, L. Nini, E. Boitard, J.P. Dubes, L. Sarda, L.C. Comeau, Kinetic properties of Penicillium cyclopium lipases studied with vinyl esters, Lipids 35 (2000) 919 – 925. P. Canioni, R. Julien, J. Rathelot, L. Sarda, Pancreatic and microbial lipases: a comparison of the interaction of pancreatic colipase with lipases of various origins, Lipids 12 (1977) 393 – 397. M.M. Bradford, A rapid and sensitive method for quantitation of microgram quantities of protein utilizing the principle of protein-dye binding, Anal. Biochem. 72 (1976) 248 – 254. H. van Tilbeurgh, M.-P. Egloff, C. Martinez, N. Rugani, R. Verger, C. Cambillau, Interfacial activation of the lipase-procolipase complex by mixed micelles revealed by X-ray crystallography, Nature 362 (1993) 814 – 820. J. Schmitt, S. Brocca, R.D. Schmid, J. Pleiss, Blocking the tunnel: engineering of Candida rugosa lipase mutants with short chain length specificity, Protein Eng. 15 (2002) 595 – 601. S. Brocca, F. Secundo, M. Ossola, L. Alberghina, G. Carrea, M. Lotti, Sequence of the lid affects activity and specificity of Candida rugosa lipase isoenzymes, Protein Sci. 12 (2003) 2312 – 2319. R. Verger, in: B.B.A.H.L. Brockman (Ed.), Lipases, Elsevier, Amsterdam, 1984, pp. 83 – 149.