Materials and Methods

Chemicals.

ε-Caprolactone and ε-caprolactam were purchased from Acros Organics USA (Fair Lawn, NJ). Phenylacetate, paraoxon, and all other lactones, thiolactones, lactams, and reagents were purchased from Sigma-Aldridge (Milwaukee, WI). Mevastatin (Compactin) and spironolactone were obtained from Sigma (St. Louis, MO); simvastatin (Zocor) and lovastatin (Mevacor) were obtained from Merck and Co. (West Point, PA) and extracted from the tablet formulations. The cyclic carbonate ester compound KB-R4899 (sodium 4-[(5-methyl-2-oxo-1,3-dioxol-4-yl) methylthio] benzenesulfonate ½ hydrate) was obtained from Kanebo, Ltd (Osaka, Japan).

Phenotyping and Purification of Human PON1 Types Q and R.

Individual human plasma/serum was phenotyped for the PON1 Q192R polymorphism by dividing the paraoxonase activity in the presence of 1 M NaCl by the arylesterase activity (with phenyl acetate as the substrate) as described (Eckerson et al., 1983). Units of activity are defined as micromoles (arylesterase) or nanomoles (paraoxonase) of substrate hydrolyzed per minute. Human PON1 types Q and R were purified from outdated citrate plasma (obtained from the University of Michigan blood bank) as previously described (Gan et al., 1991) and modified (Kuo and La Du, 1995). The average purified PON1 preparation contains the pooled sera from three to five previously phenotyped individuals.

UV Spectrophotometric Assay.

UV spectrophotometry was used to quantify the hydrolysis of lactones exhibiting reasonable absorbance differences between substrate and product, particularly for aromatic lactones. In a typical experiment the cuvette contained 1.0 mM substrate in 50 mM Tris/HCl, pH 8.0, in a total volume of 2.0 ml. The reaction was initiated by the addition of enzyme, and the increases in absorbance at 270 nm (for dihydrocoumarin), 274 nm (for 2-coumaranone), and 290 nm (for homogentisic acid lactone) were recorded. Differences in molar extinction coefficients of substrate and product were used to calculate the rate of hydrolysis. These are 876, 1295, and 816 M⁻¹ cm⁻¹ for dihydrocoumarin, 2-coumaranone, and homogentisic acid lactone, respectively.

Phenol Red Assay.

Another general assay procedure was used for substrates with inadequate spectral differences between substrate and their corresponding hydrolysis products, but with sufficient solubility to achieve millimolar concentrations. This assay used the pH indicator dye phenol red to follow hydrogen ion release from carboxylic acid formation, the product of lactone and carbonate ester hydrolysis. The method is a modification of one developed by Sharp and Rosenberry (1982) as outlined previously (Billecke et al., 1999). Briefly, a substrate solution containing 0.004% (106 μM) phenol red, 0.005% bovine serum albumin, 2.0 mM HEPES, pH 8.0, and 1.0 mM CaCl₂ with substrate was prepared. Upon addition of enzyme, serum, or recombinant expression medium, the increase in absorbance at 422 nm was monitored and recorded. Slope values (in units of change in absorbance per minute) were multiplied by a rate factor (1900 units/ml) derived from a standard acid titration curve and divided by the sample volume (in microliters). Units were defined as the number of micromoles of acid produced (i.e., substrate hydrolyzed) per unit time.

Thiolactone Hydrolysis Assay.

Thiolactone hydrolysis was measured using Ellman's procedure for monitoring the accumulation of free sulfhydryl groups via coupling with 5,5′-dithio-bis-2-nitrobenzoic acid (Ellman et al., 1961).

A dual-beam Cary 3E spectrometer (Varian, Mulgrave, Australia) with a temperature controller adjusted to 25°C was used for the above assays. All assay buffers contained 1.0 mM CaCl₂unless otherwise specified. Reference cuvettes containing the appropriate buffer and substrate were included in each assay to compensate for the effects of any spontaneous hydrolysis.

HPLC Analysis of Hydrolysis Products.

The hydrolyses of spironolactone and the statin lactones (mevastatin, lovastatin, and simvastatin) were analyzed by HPLC using a Beckman System Gold high-performance liquid chromatograph with a model 126 programmable solvent module, a model 168 diode array detector set at 238 nm, a model 7125 rheodyne manual injector valve with a 20-μl loop, and a Beckman ODS Ultrasphere column (C₁₈, 250 × 4.6 mm, 5 μm). In a final volume of 1 ml, 50 μl of enzyme solution and 10 μl of substrate solution in methanol (0.5 mg/ml) were incubated at 25°C in 25 mM Tris/HCl (pH 7.6) and 1 mM CaCl₂. Aliquots (100 μl) were removed at specified times and added to acetonitrile (100 μl), mixed with a vortex, and centrifuged for 1 min at maximum speed (Beckman Microfuge). The supernatants were poured into new tubes, capped, and stored on ice until HPLC analysis. Samples were eluted isocratically at a flow rate of 1.0 ml/min with a mobile phase consisting of the following: A = acetic acid/acetonitrile/water (2:249:249) and B = acetonitrile, in A/B ratios of 70:30, 50:50, 45:55, and 40:60 for spironolactone, mevastatin, lovastatin, and simvastatin, respectively. Under the above conditions the retention times for the carboxylic acid formed and the lactone substrate were as follows: spironolactone (3.1 and 5.5 min), mevastatin (4.5 and 6.4 min), lovastatin (4.4 and 5.6 min), and simvastatin (4.8 and 6.6 min). Response factors for the acid products were calculated from the peak heights after complete alkaline hydrolysis of the lactones in 0.02 M NaOH for 2 to 24 h at 25°C.

Recombinant PON1 Cysteine Mutant.

The cysteine at residue 284 of human PON1 type Q was replaced with alanine (C284A) using site-directed mutagenesis techniques as described previously (Sorenson et al., 1995). Serum-free Ultraculture media (Bio-Whittaker, Walkersville, MD) from Chinese hamster ovary cells stably expressing the C284A and wild-type (PON1 type Q) recombinants and a control cell line stably transfected with the pGS vector were concentrated 20-fold using Centricon-30 concentrators (Amicon, Beverly, MA) and assayed with phenyl acetate, paraoxon, the aromatic substrate homogentisic acid lactone, and the aliphatic substrate undecanoic-δ-lactone.

Removal of Calcium from PON1 and Replacement with Metal Ions.

A chelating resin (Chelex 100, Bio-Rad, Hercules, CA) was used to remove calcium from PON1 type Q enzyme as described (Kuo and La Du, 1998). Equal volumes of buffer containing either 2 mM calcium, magnesium, or zinc chloride were added to aliquots of enzyme preparation immediately after elution from the Chelex column. Care was taken to not irreversibly inactivate the enzyme after removal of the high-affinity bound calcium. Assays were performed with the appropriate metal present at 1 mM in the reaction cuvette.

Stimulation of PON1 Enzymatic Activities by Dilauroyl Phosphatidylcholine (PC).

Two milligrams of dilauroyl PC were dissolved in 1.0 ml of 50 mM Tris/HCl (pH 8.0), 1.0 mM CaCl₂ buffer and dispersed by sonication for 45 sec. One volume of purified PON1 type Q or type R was mixed with an equal volume of the phospholipid suspension. Activities of stimulated and nonstimulated PON1 samples were obtained using phenyl acetate and the lactone substrates dihydrocoumarin, 2-coumaranone, homogentisic acid lactone, γ-butyrolactone, ε-caprolactone, and undecanoic-δ-lactone. The percent stimulation or inhibition of PON1 with these substrates was estimated from changes in substrate hydrolysis after treatment.

PON1 Inhibition with Lactams.

Inhibition of PON1 arylesterase activity by lactams was achieved through either addition of the lactam to the reaction cuvette (initial inhibition) or preincubation with the enzyme and reaction buffer for 15 min, followed by addition of the substrate (phenyl acetate) to initiate the reaction. IC₅₀ values were approximated by titrating arylesterase inhibition with the lactams. Reversibility was determined by incubating the enzyme with lactam for 1 h and then removing the lactam by washing with 4 volumes of enzyme buffer using a Centricon-30.

Determination of Kinetic Constants.

Kinetics values were derived both from double reciprocal plots and by fitting experimental data to the Michaelis-Menten equation using GraphPad Prism software (version 3.00).

Results and Discussion

Previous investigations on PON1's hydrolytic activities with aromatic esters and organophosphates have identified a number of important properties that should be studied for a clear understanding of the enzyme's hydrolytic characteristics. Here we analyze PON1's lactonase activity using several of the same enzymatic characteristics—including substrate structural requirements, basic kinetics, important amino acids, and calcium requirements—previously studied for its arylesterase and organophosphatase activities.

Substrate Specificity.

Lactone Substrates

Lactones with varying ring sizes and substituents were tested with purified human PON1 types Q and R to gain information about the enzyme's structure/activity profile. To obtain values appropriate for comparison of the PON1 isozymes, we chose to relate all lactonase activities to the arylesterase activity with phenyl acetate. This reference compound was selected because the Q and R isoforms have approximately the same specific activity with this substrate under the above assay conditions. Purified PON1 Q and R preparations were diluted to have 100 units/ml arylesterase activity (with 1 mM substrate) for all subsequent assays unless otherwise specified. Current studies at this laboratory suggest that a suitable lactone may need to be selected for similar activity comparisons with other PON family members (i.e., PON2 and PON3) because of their limited or absent arylesterase activity.

The results with the lactone substrate assays are compiled in Table1. All assays were conducted at 1 mM and at 5 or 10 mM substrate concentrations, depending on the limitations in the solubility of the substrate. Note also that these restrictions limit our ability to predict the rate of hydrolysis of these substrates at saturating conditions. These concentrations were chosen arbitrarily to display the range of substrates hydrolyzed and the isozyme- and stereoisomer-specific variability of hydrolysis at specific assay conditions. Furthermore, different purified PON1 preparations (both type Q and R) were assayed with phenyl acetate and representative lactones to show the relative consistency of the values obtained from the preparations used to compile Table 1 compared with other preparations. These data are presented in Table2.

PON1 has a broad range of lactone substrates. Lactone rings containing from 4 to 7 atoms are hydrolyzed. Of the aliphatic lactone substrates, δ-valerolactone (6 member ring) is hydrolyzed at a faster rate (75.4 units/ml at 1 mM for PON1 type Q) than γ-butyrolactone (2.46 units/ml, 5 member ring) and ε-caprolactone (14.8 units/ml, 7 member ring). Introduction of hydroxyl group(s) reduces (R-α-hydroxy-γ-butyrolactone,S-β-hydroxy-γ-butyrolactone, S- andR-dihydro-5-(hydroxymethyl)-2(3H)-furanone; Table 1) or eliminates lactonase activity, as observed with some sugar lactones (i.e., l- andd-gulono-γ-lactone). The introduction of a bromide to the ring (α-bromo-γ-butyrolactone) enhances the hydrolysis compared with the unsubstituted compound (47.2 units/ml at 1 mM for PON1 type Q versus 2.46 units/ml, respectively). This suggests that PON1 has a greater affinity for the bromide form, which saturates the enzyme under these reaction conditions. Aromatic lactones are hydrolyzed at much higher rates than most aliphatic ones at 1 mM concentrations, suggesting a higher affinity and enzymatic efficiency for aromatic lactones. This is confirmed by kinetic analysis (Table3). Interestingly, coumarin, which contains an α,β double bond in the lactone ring, is not hydrolyzed by PON1.

Of special interest are the differences in rates of hydrolysis (up to 4-fold for lactones and over 6-fold for the carbonate ester KB-R4899) between the human PON1 Q and R isozymes. The latter substrate was designed by Tanaka et al. as a convenient, soluble substrate to measure arylesterase activity in human serum as an index of liver damage (Tanaka et al., 1987; Kawai et al., 1990). It is one of the few substrates we have found, in addition to paraoxon, that is hydrolyzed much more rapidly by the R than the Q isozyme.

At 1 mM concentration, some lactones (2-coumaranone, angelicolactone, and δ-valerolactone) are hydrolyzed appreciably faster by the Q form, whereas a few others (thiolactones, γ-butyrolactone, and KB-R4899) are hydrolyzed faster by the R form. We also found 2- to 9-fold differences in the rate of hydrolysis with stereoisomers [R- andS-α-hydroxy-γ-butyrolactone; R- andS-dihydro-5-(hydroxymethyl)-2(3H)-furanone; (1S,5R)- and (1R,5S)-oxabicyclooctenone]. Although thorough kinetic analyses are required for complete characterization of these isozymic differences, these data demonstrate that there is appreciable isozymic variation in the hydrolysis of lactone and carbonate ester substrates.

Augustinsson and Ekedahl (1962) proposed that only esters with an aromatic ring or a double bond adjacent to the ester group would be substrates for PON1. We confirmed their observation that ethyl acetate and cyclopentane acetate are not substrates, whereas vinyl acetate and Δ²-cyclopentenyl acetate are both hydrolyzed. Cyclic esters hydrolyzed by PON1 that fit the hypothesis of Augustinsson and Ekedahl are angelicolactone, KB-R4899, and the aromatic lactones. However, we noted that the carbon-carbon double bond requirement predicted by Augustinsson and Ekedahl does not hold for many of the lactone substrates and cyclic carbonate esters (Table 1). It is now clear that spatial restrictions imposed by the lactone ring structure allow some substrates to interact with the active center of the enzyme in a manner that leads to their hydrolysis. This interaction is not possible with the corresponding open ester forms. For example, ethyl acetate is not hydrolyzed, but γ-butyrolactone, containing the same number of atoms and ester moiety, is very well hydrolyzed. These new findings about the substrate requirements greatly extend the range of potential substrates for PON1 and represent an entirely new substrate class for the enzyme.

Fishbein and Bessman (1966a,b) reported lactonization of γ-hydroxybutyric acid catalyzed by a partially purified serum lactonase, which we now believe is PON1. However, we have not observed PON1 lactonizing activity with the open hydroxy acid forms of the aromatic lactones, when followed by direct UV absorption. Because we were unable to study lactonization of aliphatic hydroxy acids due to limitations in our phenol red assay, PON1's possible lactonizing activity with aliphatic hydroxy acids cannot be excluded.

Kinetic data for select lactone substrates show that the affinity or the turnover for some exceed that of phenyl acetate, which has long been considered the optimal PON1 substrate (Table 3). Values for the hydrolysis of the organophosphates paraoxon, sarin, and soman are included in Table 3 for comparison. Estimates of enzymatic efficiency, presented as the product ofV_max/K_m, show that PON1 is most active with the aromatic lactones. This is due, at least in part, to the higher affinity of the enzyme for substrates with an aromatic component.

Some Lactone Drugs Are Substrates for PON1.

Based on the structural requirements for PON1's lactonase activity, it is reasonable to expect that some lactone- and carbonate ester-containing drugs or prodrugs should be hydrolyzed by this enzyme. We found that the diuretic spironolactone and three hydroxymethylglutaryl-CoA reductase inhibitors, i.e., mevastatin, lovastatin, and simvastatin, are hydrolyzed by PON1 (Table4). Both Q and R isozymes of PON1 appear to hydrolyze these compounds with approximately the same efficiency under the experimental conditions used.

Earlier studies reported a “statinase” activity of human and animal plasma with some of the above hydroxymethylglutaryl-CoA reductase inhibitors (Vickers et al., 1990; Tang and Kalow, 1995), and we can now attribute this hydrolytic activity, at least in part, to PON1. Other examples of drugs hydrolyzed by PON1 and the pharmacological benefit of these activities have recently been reported. These include the systemic metabolism of glucocorticoid δ-lactones (Biggadike et al., 2000) and the activation of the carbonate ester prodrug prulifloxacin [(±)-6-fluoro-1-methyl-7-[4-(5-methyl-2-oxo-1,3-dioxolen-4-yl)methyl-1-piperazinyl]-4-oxo-4H-[1,3]thiazeto[3,2-a]quinoline-3-carboxylic acid or NM441] (Tougou et al., 1998). Thus, PON1 lactonase activity may be utilized for its ability to metabolize topically administered agents and thereby limit their unwanted systemic effects (Biggadike et al., 2000). In addition, this serum esterase/lactonase can be used to activate prodrugs (Tougou et al., 1998). Interestingly, PON1 type R serum was shown to hydrolyze prulifloxacin at a higher rate than type Q serum, an observation that is also true for paraoxon, as well as KB-R4899.

Isozymic differences in affinity and turnover, coupled with quantitative differences in the amount of the enzyme (which may vary up to 10-fold or more), can influence both the tissue concentration and toxicity of some lactone-containing drugs in different individuals. Recently, the importance of both the quantity and quality of serum PON1 in regard to organophosphate hydrolysis and toxicity has been demonstrated (Richter and Furlong, 1999). The level and quality of PON1 may also be of clinical significance by affecting the degree of individual variations in the metabolism of some lactone-containing drugs.

Thiolactones are less well hydrolyzed by PON1 than lactones.

Thiolactones are hydrolyzed by PON1, but at much lower rate than the structurally homologous lactones (Table 1). Although HTL is a relatively poor substrate for PON1, it appears to be hydrolyzed only by PON1 in human serum (Jakubowski, 2000).

Elevated levels of homocysteine are an independent risk factor for cardiovascular disease in humans (Langman and Cole, 1999), and there is evidence suggesting that homocysteine can be harmful to human cells because of its metabolic conversion to HTL (Jakubowski, 2000). HTL reacts with proteins by a mechanism involving homocysteinylation of protein lysine residues leading to protein damage, increased LDL aggregation and enhanced macrophage scavenging (Jakubowski, 2000;Langman and Cole, 1999; McCully, 1996).

It has been demonstrated that PON1 protects against homocysteinylation by hydrolyzing HTL (Jakubowski, 2000). We have previously shown that PON1 protects LDL from oxidation induced by either copper ion or amidinopropane hydrochloride (Aviram et al., 1998a,b). Thus, serum PON1 may have two independent antiatherogenic roles: one by preventing the accumulation and promoting the removal of oxidized lipids from LDL and the second by detoxifying HTL. Interestingly, the Q type isozyme is more efficient in protecting against LDL oxidation than the R type PON1, but conversely the R type hydrolyzes thiolactones more readily than the Q isoform. This may help explain the ambiguity observed in some studies focusing on PON1's role in atherosclerosis.

Lactams Inhibit PON1 Activity.

Lactams are isosteric forms of lactones in which the ring oxygen is replaced by a nitrogen. It appears that lactams are not hydrolyzed by PON1 but rather inhibit the enzyme. The in vitro inhibition studies thus far performed have identified seven lactams as potent PON1 inhibitors: oxindole, isatin, δ-valerolactam, ε-caprolactam, 2-hydroxyquinoline, 3,4-dihydro-2(1H)-quinoline andN-bromo-ε-caprolactam (Table5). For some of these compounds (i.e., 2-hydroxyquinoline and 3,4-dihydro-2(1H)-quinoline), the IC₅₀ values are in the micromolar range. Isatin and N-bromo-ε-caprolactam appear to inhibit PON1 arylesterase activity irreversibly; these lactams are being further investigated to identify residues that are presumably components of the enzyme's active center.

Enzyme Requirements for Lactonase Activity.

PON1's Free Cysteine (Residue 284) Is Important for Lactonase Activity

Augustinsson and Ekedahl (1962) hypothesized that the mechanism of arylester hydrolysis involves the formation of thioester intermediates from the enzyme sulfhydryl and the aryl moiety of the substrate. We have previously demonstrated that PON1's free cysteine (residue 284) is not required for the arylesterase and paraoxonase activities (Sorenson et al., 1995), but is essential for the enzyme's ability to protect LDL against oxidation (Aviram et al., 1998). To determine whether the conserved free cysteine is required for PON1 lactonase activity, we compared wild-type recombinant PON1 type Q with a C284A mutant (Table 6).

Activities for phenyl acetate and paraoxon are approximately 44- and 20-fold lower for the C284A mutant, respectively, than for wild-type PON1, but are consistently more than 5 times greater than background (not shown), in agreement with earlier reported values (Sorenson et al., 1995). However, no lactonase activity was observed with either homogentisic acid lactone or undecanoic-δ-lactone, but both lactones were hydrolyzed very efficiently by the expressed wild-type PON1. These data indicate that the free cysteine at residue 284 is necessary for lactonase activity with the representative lactones and may be a general requirement for lactone hydrolysis. Furthermore, the lactonase activity observed with the wild-type PON1 recombinant proves that lactone hydrolysis is attributed to PON1 and not to some contaminant within PON1 preparations purified from human serum.

Metal Ion Requirements.

Both arylesterase and paraoxonase activities of PON1 require calcium for activity (Kuo and La Du, 1998). Activities of Ca²⁺ ion-depleted preparations in which Ca²⁺ ions were replaced by either Zn²⁺ or Mg²⁺ are compared with the calcium preparations in Table 7.

PON1 has two calcium binding sites, a higher affinity site (K_d = 3.6 ± 0.9 × 10^-7 M), which maintains both the hydrolytic activity and stability, and one with a lower affinity (K_d = 6.6 ± 1.2 × 10^-6 M), which is required for catalytic activity (Kuo and La Du, 1998). The enzyme activities with all substrates were greatly reduced in the presence of zinc and magnesium ions (Table 7). It was also found that percent lactonase activity compared favorably with percent arylesterase activity, with the greatest recovered activity occurring with magnesium-PON1 type Q hydrolysis of ε-caprolactone (9.0%). This residual activity is probably due to incomplete removal of calcium ions. The results indicate that calcium was required for PON1 lactonase activity in the same way as for the other types of hydrolytic activity. A recent report confirms these observations for thiolactonase activity (Jakubowski, 2000).

Modification of Lactonase Activity by Phospholipids: Stimulation by Dilauroyl PC.

The paraoxonase and arylesterase activities of purified human serum PON1 are stimulated by a number of phospholipids with the greatest stimulation obtained in the presence of dilauroyl PC (Kuo and La Du, 1995). Stimulation is achieved through improvement of the enzyme'sV_max but not K_mvalues. An interaction between serum PON1 and phospholipids is probably important due to the enzyme's strong association with lipids in HDL (Sorenson et al., 1999). Because of this association, we have tested stimulation of PON1 types Q and R with dilauroyl PC to determine whether it has an effect on lactonase activity as well.

The effect of phospholipids on lactonase activity is variable (Table8). Percent stimulation between Q and R types differs markedly among the substrates with the most obvious variation appearing with dihydrocoumarin, where R type activity is inhibited 17% whereas Q type is stimulated 41%. This is the only incidence where inhibition in the presence of dilauroyl PC is apparent, although the hydrolysis of other aromatic lactones, 2-coumaranone and homogentisic acid lactone, also shows less stimulation with R type PON1. Otherwise, percent stimulations for both PON1 isoforms seem to peak at the 5-membered ring γ-butyrolactone with stimulation decreasing as the ring size increases. The observed inhibition of PON1 hydrolytic activity with dihydrocoumarin in the presence of dilauroyl PC was unexpected. To our knowledge, this is the first description of differences in phospholipid stimulation for Q and R type PON1 isozymes.

Lactonase Evolution: Evolutionary Aspects of the PONs.

There is appreciable structural homology between a fungal lactonase ofFusarium oxysporum and human PON1 (Kobayashi et al., 1998). Kobayashi et al. logically suggest an evolutionary relationship between these fungal and human enzymes, and recent work in our laboratory supports this idea. Surprisingly, compounds such as homogentisic acid lactone and dihydrocoumarin were substrates for the fungal lactonase (Kobayashi et al., 1998) and for both the human (this study) and rabbit (Watson et al., 2000) PON1 enzymes. The appreciable degree of sequence identity and homology between the other human PON-like proteins make it reasonable to expect that they also may have lactonase activity. Although this conservation of enzymatic activity does not appear true for arylesterase and organophosphatase activities, these two activities may reflect recently acquired characteristics as opposed to a more basic lactonase function. In fact, lactonase activity has recently been observed with rabbit PON3 (Draganov et al., 2000) and lactone substrates may be helpful in identifying other members of the mammalian PON family.