Interfacial Reaction Dynamics and Acyl-enzyme Mechanism for Lipoprotein Lipase-catalyzed Hydrolysis of Lipid p-Nitrophenyl Esters*

The fatty acyl (lipid) p-nitrophenyl esters p-nitro-phenyl caprylate, p-nitrophenyl laurate and p-nitro- phenyl palmitate that are incorporated at a few mol '3, into mixed micelles with Triton X-100 are substrates for bovine milk lipoprotein lipase. When the concentration of components of the mixed micelles is approx- imately equal to or greater than the critical micelle Concentration, time courses for lipoprotein lipase-cat- alyzed hydrolysis of the esters are described by the integrated form of the Michaelis-Menten equation. Least square fitting to the integrated equation there- fore allows calculation of the interfacial kinetic parameters K;t, and v*,, from single runs. The computa- tional methodology used to determine the interfacial kinetic parameters is described in this paper and is used to determine the intrinsic substrate fatty acyl specificity of lipoprotein lipase catalysis, which is re- flected in the magnitude of kzat/K% and kZat. The results for interfacial lipoprotein lipase catalysis, along with previously determined kinetic parameters for the water-soluble estersp-nitrophenyl acetate andp-nitro-phenyl butyrate, indicate that lipoprotein lipase has highest specificity for the substrates that have fatty acyl chains of intermediate length (Le. p-nitrophenyl butyrate and p-nitrophenyl caprylate).

alyzed by soluble enzymes at lipid-water interfaces, increases in a manner functionally described by the Michaelis-Menten equation as the analytical concentration of lipid substrate increases (1, 2). However, the kinetic parameters Pzp and VZg do not bear the same relationship to the various microscopic steps of the catalytic mechanism as they do for singlesubstrate enzyme reactions of soluble enzymes and monomolecularly dispersed ( i e . water-soluble) substrates. This difference is illustrated by considering the Verger-deHaas model for interfacial enzyme catalysis (2).

SCHEME I
The definitions of terms in Scheme I are E = free lipoprotein lipase; E* = lipoprotein lipase bound to the substrate-containing particle interface, but with unoccupied active sites (the penetration complex); ES* = Michaelis complex of lipoprotein lipase and substrate formed at the interface; xs = concentration of substrate in the interface in mole fraction units (N. B., Verger and deHaas use concentration per surface area to define concentration); M is the micellar or lipoprotein complex that contains substrate monomers; P = reaction products. The velocity equation for the mechanism of Scheme I is as follows.
In Equation 1, [ M I is the analytical concentration of all the components that comprise the substrate-containing particle (N.B., Verger and deHaas use interface area per volume).

Equation 1 shows that velocity increases as [ M I increases,
because the total interface area of substrate-containing particles concomitantly increases and eventually the enzyme is saturated by interface (i.e. all enzyme is in the E* form). Clearly, a method for directly determining the interfacial Michaelis-Menten parameters P , and Vm., would be of great value in determining how such factors as pH, temperature, substrate structure, added nucleophiles, and the physiological activator apolipoprotein C-I1 affect the molecular dynamics of lipoprotein lipase catalysis. This paper describes the development of micellar substrates for lipoprotein lipase that consist of Triton X-100 and the p-nitrophenyl esters of caprylic, lauric, or palmitic acid. Because lipoprotein lipase-catalyzed hydrolysis of these esters releases p-nitrophenoxide, which absorbs light at 400 nm, reaction of the micellar substrates can be followed continuously by spectrophotometry. Time courses for lipoprotein lipase-catalyzed hydrolysis of the lipid p-nitrophenyl esters are first-order at concentrations below the critical micelle concentration of the mixed micelles. However, at concentrations near and above the critical micelle concentration, the kinetics become mixed-order and are well described by the integrated form of the Michaelis-Menten equation. Hence, least square analysis of the mixed-order time courses allows calculation of V-, and P , . This method is used herein to characterize the mechanism of interfacial lipoprotein lipase-catalyzed hydrolysis of lipid p-nitrophenyl esters.

DISCUSSION
Physiological lipoprotein lipase catalysis involves the hydrolysis of triacylglycerols and phospholipids of the triacylglycerol-rich lipoproteins, very low density lipoproteins and chylomicrons (1, 20). This process is perforce an interfacial reaction, which makes the molecular dynamics of lipoprotein lipase catalysis difficult to characterize for the reasons discussed in the Introduction. In this manuscript we describe the development and characterization of a novel lipoprotein lipase reaction, the hydrolysis of lipid p-nitrophenyl esters that are contained in mixed micelles with Triton X-100. This system offers advantages for the study of interfacial lipoprotein lipase catalysis. (a) The reaction is biomimetic. Both lipid p-nitrophenyl ester and lipoprotein lipase are bound to Triton X-100 micelles (cf. Fig. 1 alogs of phosphatidylcholine and phosphatidylethanolamine that are contained in Triton X-100 micelles. Hence, our use of Triton X-100 mixed micelles as model substrates for lipoprotein lipase is not without precedent in the literature of lipolytic enzyme mechanisms. A hallmark of lipoprotein lipase-catalyzed hydrolysis of triacylglycerols that are contained in emulsions or lipoproteins is product inhibition by fatty acids (13-17) in the absence of a fatty acid acceptor such as bovine serum albumin.
The results reported herein are unusual in that fatty acid products do not inhibit the lipoprotein lipase reaction. Plots such as the inset of Fig. 2B show that lipoprotein lipasecatalyzed hydrolysis of lipid p-nitrophenyl esters is well described by the integrated Michaelis-Menten equation. Progressive product inhibition would be accompanied by systematic upward curvature of the plot. Lack of product inhibition is not peculiar to the composition of the micelle, since the interfacial kinetic parameters do not depend on the mole fraction of substrate in the micelle. The most surprising finding, however, is that addition of lauric acid to p-nitrophenyl laurate micelles actiuates the reaction. Hence, product inhibition in lipoprotein or emulsion substrates likely does not occur by mechanisms that involve competitive binding at the lipoprotein lipase active site or binding on an effector site that is distal from the active site (See Ref. 1 for a presentation of possible fatty acid inhibition mechanisms). Bengtsson and Olivecrona (14) showed that, in the absence of albumin, lipoprotein lipase synthesizes acylglycerols from [3H]oleic acid in the presence of a trioleylglycerol/gum arabic emulsion. This is probably not the product inhibition mechanism for lipid p-nitrophenyl ester substrates because the p-nitrophenoxide product is neither lipophilic nor nucleophilic and does not accumulate in the micelle or participate in the reverse reaction. It is also possible that the reverse reaction does not occur because the fatty acid products rapidly diffuse away from the micelles. The linearity of the plot in the Fig. 2B inset is consistent with the lack of a reverse reaction. These considerations suggest that mixed micelles of Triton X-100 and lipid p-nitrophenyl esters are good systems for further probing the mechanism of product inhibition of lipoprotein lipase.
As mentioned earlier, lipoprotein lipase-catalyzed hydrolysis of mixed micellar lipidp-nitrophenyl esters is a biomimetic reaction. If this is so, the reaction ought to share some of the features of lipoprotein lipase-catalyzed hydrolysis of physiological substrates. Shinomiya et al. (23) studied the lipoprotein lipase-catalyzed hydrolysis of an homologous series of phosphatidylcholines in mixed micelles with Triton X-100. They found that in the absence of the activator apolipoprotein C-I1 the reaction velocity decreased with increasing fatty acyl chain length. Bengtsson and Olivecrona (24) found that bovine milk lipoprotein lipase more readily hydrolyzes trioctanoin than triolein when the lipids are contained in Triton X-100-stabilized emulsions. We have observed decreasing lipoprotein lipase activity with increasing chain length for the lipid p-nitrophenyl esters p-nitrophenyl caprylate, p-nitrophenyl laurate, and p-nitrophenyl palmitate, as reflected in both k& and kZat/K+, values (cf. Table I). For the water-soluble substrates p-nitrophenyl acetate and p-nitrophenyl butyrate the data of Table I show that lipoprotein lipase prefers the longer acyl chain. These trends are consistent with highest intrinsic lipoprotein lipase specificity for fatty acyl chains of intermediate length. The fact that the trend in fatty acyl specificity for lipoprotein lipase-catalyzed hydrolysis of lipid p-nitrophenyl esters parallels that for hydrolysis of triacylglycerols and phosphatidylcholines suggests that the reaction Interfacial Lipoprotein Lipase Reaction Dynamics 3 system described in this paper is indeed biomimetic. The plots of Fig. 4 show that the nucleophile hydroxylamine increases K*, and Xax by equal amounts, and hence the Lineweaver-Burk plots are parallel. Scheme I1 depicts a mechanism that is consistent with these kinetics. Nucleophilic attack by the active site serine (1,19,25) and consequent loss of p-nitrophenoxide produces the acyl-enzyme intermediate. Hydroxylamine activates the reaction by providing an alternate route for decomposition of the acyl-lipoprotein lipase intermediate. This mechanism is an interfacial catalysis analog of the mechanism of nucleophilic activation of serine proteases (26,27). For such a mechanism, a steady-state derivation of the kinetic parameters gives: In these equations Nu is hydroxylamine. K: in Equation 9 is the Michaelis constant for the interfacial lipoprotein lipasesubstrate complex. K$ contains k, and k2 (not shown in Scheme II), which are, respectively, the rate constants for association and dissociation of substrate monomers at the lipoprotein lipase active site. Equations 8 and 9 are predicated on the concept that decomposition of the acyl-lipoprotein lipase intermediate is the rate-determining step. These equations predict linear and matching increases in Vm,, and Km as [Nu] increases. The plots of Fig. 4B show that this is the case. Chemical analysis of reaction products for hydroxylamineactivated lipoprotein lipase-catalyzed hydrolysis of p-nitrophenyl laurate (cf. "Results") also supports the mechanism of Scheme 11. This is the first example of a kinetic signature of an acyl-enzyme mechanism for any lipoprotein lipase-catalyzed reaction. Therefore, micellar lipid p-nitrophenyl esters offer a system for which the molecular details of interfacial lipoprotein lipase catalysis are reasonably well defined.
The ability to measure the interfacial kinetic parameters KZ and V,,, opens additional aspects of lipoprotein lipase reaction dynamics to scrutiny. For example, solvent isotope effects can determine whether interfacial lipoprotein lipase catalysis involves transition states that are stabilized by proton transfer (19). The effect of mechanism-based inhibitors on interfacial lipoprotein lipase reaction dynamics can be defined. Lipid p-nitrophenyl esters can be incorporated into different micelles, such as phosphatidylcholine micelles, which can provide a yet more biomimetic substrate system. These and other investigations of interfacial lipoprotein lipase catalysis are being pursued in our laboratory. 14.

MATERIALS AbD llBTBODS
Il&rinls -LpL was purified from skimmed bovine milk by using previously described procedures (4,5). The purified enzyme showed a slngle dextran. the sodium salt of heparin from porcine intestinal mucosa, TX100, band when assayed for homogeneity by SDS-PAGE. PNPC, PNPL, PNPP, blue Sepharose CL-68 and rhodamine 6G were purchased from Sigma Chemical Company and were used a s received. All other materials were commercially available reagent-grade products.
substrates were prepared-g on an analytical balance the requi- Control experiments showed that p-nitropheno!jdePthat was produced by hydrolysis of PNPL in the absence of LpL was quantitatively extracted into the methanolic aqueous phase. while by e~aporating~the solvent, adding 1 mL of 1 N NaOH to hydrolyze the ester,

( 3 1
A A and A a r e absorbances at tlme t, at t = 0 and at t = 11, respectively; k'is'the frcst-order rate constant and always e q u a l s Vmax/Km foc single substrate enzyme reactions.
For reactions conducted n e a r or above the CHC, tlmecourses were described by the integrated form of the Michaelis-Menten equation, which is expressed in terms of absorbance in eq 4:

phase.
A laurylhydroxamate TLC standard was synthesized from lauryi chloride and NH20H'HC1 in pyridine and purified by recrystallization from hot EtOH. LpL catalysis of laurylhydroxamate formation was establlshed by extracting Selected reaction mixtures a s described above, followed by quantitation of laurylhydroxamate by measuring the absorbance at 520 nm of the dye. Therefore, the dye is likely "at inducing micellization of the TXlOO lipid p-nitrophenyl ester mixed micelles. nixed micelles of TXlOO and PNPi were further characterized by a e l filtration chromatooraohv on a colnrnn o f micelles coelute. which demonstrates that TX100 and PNPL a r e contained in Sepharose CL-68. As Fig 1ii ghhows, TXlOO and PNPL cbn;tit;e"~s~ormix;a the same micelles. Moreover, the mixed micelles elute at the same volume ( 5 6 k 2 mL, 3 determinations) a s that of pure TXlOO micelles ( 5 6 2 2 mL 5 determinetionsl. A broad shoulder that elutes at the column volume contalns TXlOO but n o PNPL monomers. Figure lB shows that LpL elutes with micelles when the enzyme and TXlOO are co-chrornatographed. When the chromatographic run is repeated Without TXlOO micelles, LpL is retained by the column and cannot be eluted by the buffer. Therefore, the coelution of LpL and TXlOO shown in Fig 18 demonstrates that LpL associates with the micelles to form a complex that is sufficiently tight-to overcome the association of the enzyme with Sepharose CL-68. Binding of LpL to TXlOO micelles does not radically alter the micelles. sincetheCUCisthesame in the ~r e s e n r m a n d a h a r n r r n f The kinetics of LpL-catalyzed hydrolysis of micellar lipid p-nitrophenyl ester-. change a s the analytical concentration of the micellar components is increased. Below the CUC the kinetics a r e first-order a s the fit to eq 3 of timec0uK.e data for LpL-catalyzed hydrolysis of PN;L shows (cf. Fig ZA).
At the CHC and above. the kinetics are described by the integrated form of the Michaelis-Menten equation, a 6 illustrated by the linear-least squares fit in the Fig 28 inset of data for LpL-catalyzed hydrolysis of PNPL.
Similar linear fits w e~e also obtained for LpL catalyzed hydrolysis of PNPC and PNPP. LpL molecular weight of 55,000 11) was used to calc~late kcat.
Sodium phosphate buffer that contains 0.1 N NaCl and 1 0 pq/mL of heparin.
bRate constants for PNPA and PNPB are kcat and kcat/Km. Rate constants for PNPA were calculated from data of Q u l n n et a l .
( 1 8 1 ; rate constants for PNPB were calculated from data of Q u i n n 119). Initial velocities were also measured at constant analytical concentration be described mathematically by eq 7, wh,?tpYs the interfaclal catalysis The resultinq plot should a n a l o g of the differentla1 Hichaells-Menten equatlon: The plots Of Figure 3 show that thls prediction 1 s Correct. F l y u r e 3 R is a Lineweaver-Burt plot of the reciprocal transform of the data of F l g u r e 3A. A nonlinear-least squares fit to eq 7 yields the following interfacial kinetic parameters: K * = 0.06 t 0 . pattern of l i n e s , which arises from equal increases ~n V:>" and Kf, is consistent with nucleophilic trapping of an acylenzyme intermedlate whose hydrolysis is the rate-determining step in the absence of nucleophiles. . .