Surface loops of extracellular phospholipase A(1) determine both substrate specificity and preference for lysophospholipids

Members of the pancreatic lipase family exhibit both lipase activity toward triacylglycerol and/or phospholipase A(1) (PLA(1)) activity toward certain phospholipids. Some members of the pancreatic lipase family exhibit lysophospholipase activity in addition to their lipase and PLA(1) activities. Two such enzymes, phosphatidylserine (PS)-specific PLA(1) (PS-PLA(1)) and phosphatidic acid (PA)-selective PLA(1)α (PA-PLA(1)α, also known as LIPH) specifically hydrolyze PS and PA, respectively. However, little is known about the mechanisms that determine their substrate specificities. Crystal structures of lipases and mutagenesis studies have suggested that three surface loops, namely, β5, β9, and lid, have roles in determining substrate specificity. To determine roles of these loop structures in the substrate recognition of these PLA(1) enzymes, we constructed a number of PS-PLA(1) mutants in which the three surface loops are replaced with those of PA-PLA(1)α. The results indicate that the surface loops, especially the β5 loop, of PA-PLA(1)α play important roles in the recognition of PA, whereas other structure(s) in PS-PLA(1) is responsible for PS preference. In addition, β5 loop of PS-PLA(1) has a crucial role in lysophospholipase activity toward lysophosphatidylserine. The present study revealed the critical role of lipase surface loops, especially the β5 loop, in determining substrate specificities of PLA(1) enzymes.     

Biochemical and molecular characterization of two phosphatidic acid-selective phospholipase A1s,mPA-PLA1alpha and mPA-PLA1beta

We have identified a novel phospholipase A1, named mPA-PLA1beta, which is specifically expressed in human testis and characterized it biochemically together with previously identified mPA-PLA1alpha. The sequence of mPAPLA1beta encodes a 460-amino acid protein containing a lipase domain with significant homology to the previously identified phosphatidic acid (PA)-selective PLA1, mPA-PLA1alpha. mPA-PLA1beta contains a short lid and deleted beta9 loop, which are characteristics of PLA1 molecules in the lipase family, and is a member of a subfamily in the lipase family that includes mPA-PLA1alpha and phosphatidylserine-specific PLA1. Both mPA-PLA1beta and mPA-PLA1alpha recombinant proteins exhibited PA-specific PLA1 activity and were vanadate-sensitive. When mPAPLA1beta-expressing cells were treated with bacterial phospholipase D, the cells produced lysophosphatidic acid (LPA). In both mPA-PLA1alpha and beta-expressing cells, most of the PA generated by the phospholipase D (PLD) treatment was converted to LPA, whereas in control cells it was converted to diacylglycerol. When expressed in HeLa cells most mPA-PLA1alpha protein was recovered from the cell supernatant. By contrast, mPA-PLA1beta was recovered almost exclusively from cells. Consistent with this observation, we found that mPA-PLA1beta has higher affinity to heparin than mPA-PLA1alpha. We also found that the membrane-associated mPA-PLA1s were insoluble in solubilization by 1% Triton X-100 and were detected in Triton X-100-insoluble buoyant fractions of sucrose gradients. The present study raises the possibility that production of LPA by mPA-PLA1alpha and -beta occurs on detergent-resistant membrane domains of the cells where they compete with lipid phosphate phosphatase for PA.     

Structure and function of phosphatidylserine-specific phospholipase A1

Phospholipase A1 (PLA1) is an enzyme that hydrolyzes the sn-1 fatty acids from phospholipids and produces 2-acyl-lysophospholipids. Although PLA1 activities are detected in many tissues and cell lines, a limited number of PLA1s have been purified and cloned so far. These include phosphatidylserine (PS)-specific PLA1 (PS-PLA1) from rat platelets, PLA1 from vespid venom, and phosphatidic acid (PA)-preferential PLA1 (PA-PLA1). Structurally, the former two PLA1s belong to the lipase family, where they form a subfamily among the lipase family. An alignment of the PLA1s with other members of the lipase family revealed two molecular characteristics of PLA1: the presence of extremely short lids and deleted beta9 loops. The two surface loops have been implicated in the ligand recognition in human pancreatic lipase (PL) and guinea pig PL-related protein 2. Under physiological conditions, accessibility of PS-PLA1 to its substrate is limited as it is a secreted enzyme and PS is normally located in the inner leaflet of the lipid bilayer. However, PS-PLA1 efficiently hydrolyzes PS exposed on the surface of cells such as apoptotic cells and activated platelets, and produces 2-acyl-lysophosphatidylserine (lysoPS), which is a lipid mediator for mast cells, T cells and neural cells. Identification of PS-PLA1 reveals the presence of PLA1 subfamily within the lipase family and suggests that PLA1 has a role in the production of lysophospholipid mediators.     

The triglyceride lipases of the pancreas

Pancreatic triglyceride lipase (PTL) and its protein cofactor, colipase, are required for efficient dietary triglyceride digestion. In addition to PTL, pancreatic acinar cells synthesize two pancreatic lipase related proteins (PLRP1 and PLRP2), which have a high degree of sequence and structural homology with PTL. PLRP1 has no known activity. PTL and PLRP2 differ in substrate specificity, behavior in bile salts and dependence on colipase. Each protein has a globular amino-terminal (N-terminal) domain, which contains the catalytic site for PTL and PLRP2, and a beta-sandwich carboxyl-terminal (C-terminal) domain, which includes the predominant colipase-binding site for PTL. Inactive and active conformations of PTL have been described. They differ in the position of a surface loop, the lid domain, and of the beta5-loop. In the inactive conformation, the lid covers the active site and, upon activation by bile salt micelles and colipase or by lipid-water interfaces, the lid moves dramatically to open and configure the active site. After the lid movement, PTL and colipase create a large hydrophobic plateau that can interact with the lipid-water interface. A hydrophobic surface loop in the C-terminal domain, the beta5' loop, may also contribute to the interfacial-binding domain of the PTL-colipase complex.     

Substrate specificity of lipoprotein lipase and endothelial lipase: studies of lid chimeras

The triglyceride (TG) lipase gene subfamily, consisting of LPL, HL, and endothelial lipase (EL), plays a central role in plasma lipoprotein metabolism. Compared with LPL and HL, EL is relatively more active as a phospholipase than as a TG lipase. The amino acid loop or "lid" covering the catalytic site has been implicated as the basis for the difference in substrate specificity between HL and LPL. To determine the role of the lid in the substrate specificity of EL, we studied EL in comparison with LPL by mutating specific residues of the EL lid and exchanging their lids. Mutation studies showed that amphipathic properties of the lid contribute to substrate specificity. Exchanging lids between LPL and EL only partially shifted the substrate specificity of the enzymes. Studies of a double chimera possessing both the lid and the C-terminal domain (C-domain) of EL in the LPL backbone showed that the role of the lid in determining substrate specificity does not depend on the nature of the C-domain of the lipase. Using a kinetic assay, we showed an additive effect of the EL lid on the apparent affinity for HDL(3) in the presence of the EL C-domain.     

Reactivation of the totally inactive pancreatic lipase RP1 by structure-predicted point mutations

Both classical pancreatic lipase (DPL) and pancreatic lipase-related protein 1 (DPLRP1) have been found to be secreted by dog exocrine pancreas. These two proteins were purified to homogeneity from canine pancreatic juice and no significant catalytic activity was observed with dog PLRP1 on any of the substrates tested: di- and tri-glycerides, phospholipids, etc. DPLRP1 was crystallized and its structure solved by molecular replacement and refined at a resolution of 2.10 Å. Its structure is similar to that of the classical PL structures in the absence of any inhibitors or micelles. The lid domain that controls the access to the active site was found to have a closed conformation. An amino-acid substitution (Ala 178 Val) in the DPLRP1 may result in a steric clash with one of the acyl chains observed in the structures of a C11 alkyl phosphonate inhibitor, a transition state analogue, bound to the classical PL. This substitution was suspected of being responsible for the absence of DPLRP1 activity. The presence of Val and Ala residues in positions 178 and 180, respectively, are characteristic of all the known PLRP1, whereas Ala and Pro residues are always present in the same positions in all the other members of the PL gene family. Introducing the double mutation Val 178 Ala and Ala 180 Pro into the human pancreatic RP1 (HPLRP1) gene yielded a well expressed and folded enzyme in insect cells. This enzyme is kinetically active on triglycerides. Our findings on DPLRP1 and HPLRP1 are therefore likely to apply to all the RP1 lipases. Proteins 32:523–531, 1998. © 1998 Wiley-Liss, Inc.     

The open lid mediates pancreatic lipase function

Pancreatic triglyceride lipase (PTL) and the homologous pancreatic lipase related protein 2 (PLRP2) provide a unique opportunity to understand the molecular mechanism of lipolysis. They differ in substrate specificity, sensitivity to bile salts, and colipase dependence despite their close amino acid and tertiary structure identity. One important structure, present in both lipases, is the lid which occupies different positions in the inactive and active forms of PTL. We investigated the role of the lid in lipase function by site-specific mutagenesis. By exchanging the lids between PTL and PLRP2, we created two chimeric lipases. Additionally, we made multiple substitution mutations in the PTL lid. PLRP2 with the PTL lid had kinetic properties similar to PLRP2. PTL with the PLRP2 lid was greatly impaired and had no activity at micellar bile salt concentrations even in the presence of colipase. Both chimeras showed interfacial activation suggesting that the closed lid position was maintained. A series of substitution mutations were made in positions Arg257 and Asp258. These mutations demonstrated the importance of these two residues to maintaining the normal activity, triglyceride acyl chain specificity, and colipase interaction of PTL. The preserved interfacial activation in the chimeras, the similar crystal structure of the two lids in the closed position, and the importance of Arg257 and Asp258 in mediating the open conformation of the lid argue that the position of the open lid influences the differences in activity against triglycerides, in sensitivity to bile salts, and in colipase dependence between PTL and PLRP2.     

Porcine pancreatic lipase related protein 2 has high triglyceride lipase activity in the absence of colipase

Efficient dietary fat digestion is essential for newborns who consume more dietary fat per body weight than at any other time of life. In many mammalian newborns, pancreatic lipase related protein 2 (PLRP2) is the predominant duodenal lipase. Pigs may be an exception since PLRP2 expression has been documented in the intestine but not in the pancreas. Because of the differences in tissue-specific expression, we hypothesized that the kinetic properties of porcine PLRP2 would differ from those of other mammals. To characterize its properties, recombinant porcine PLRP2 was expressed in HEK293T cells and purified to homogeneity. Porcine PLRP2 had activity against tributyrin, trioctanoin and triolein. The activity was not inhibited by bile salts and colipase, which is required for the activity of pancreatic triglyceride lipase (PTL), minimally stimulated PLRP2 activity. Similar to PLRP2 from other species, PLRP2 from pigs had activity against galactolipids and phospholipids. Importantly, porcine PLRP2 hydrolyzed a variety of dietary substrates including pasteurized human mother's milk and infant formula and its activity was comparable to that of PTL. In conclusion, porcine PLRP2 has broad substrate specificity and has high triglyceride lipase activity even in the absence of colipase. The data suggest that porcine PLRP2 would be a suitable lipase for inclusion in recombinant preparations for pancreatic enzyme replacement therapy.     

A novel fluorogenic substrate for the measurement of endothelial lipase activity

Endothelial lipase (EL) is a phospholipase A1 (PLA1) enzyme that hydrolyzes phospholipids at the sn-1 position to produce lysophospholipids and free fatty acids. Measurement of the PLA1 activity of EL is usually accomplished by the use of substrates that are also hydrolyzed by lipases in other subfamilies such as PLA2 enzymes. In order to distinguish PLA1 activity of EL from PLA2 enzymatic activity in cell-based assays, cell supernatants, and other nonhomogeneous systems, a novel fluorogenic substrate with selectivity toward PLA1 hydrolysis was conceived and characterized. This substrate was preferred by PLA1 enzymes, such as EL and hepatic lipase, and was cleaved with much lower efficiency by lipases that exhibit primarily triglyceride lipase activity, such as LPL or a lipase with PLA2 activity. The phospholipase activity detected by the PLA1 substrate could be inhibited with the small molecule esterase inhibitor ebelactone B. Furthermore, the PLA1 substrate was able to detect EL activity in human umbilical vein endothelial cells in a cell-based assay. This substrate is a useful reagent for identifying modulators of PLA1 enzymes, such as EL, and aiding in characterizing their mechanisms of action.     

Characterization of the lipolytic activity of endothelial lipase

Endothelial lipase (EL) is a new member of the triglyceride lipase gene family previously reported to have phospholipase activity. Using radiolabeled lipid substrates, we characterized the lipolytic activity of this enzyme in comparison to lipoprotein lipase (LPL) and hepatic lipase (HL) using conditioned medium from cells infected with recombinant adenoviruses encoding each of the enzymes. In the absence of serum, EL had clearly detectable triglyceride lipase activity. Both the triglyceride lipase and phospholipase activities of EL were inhibited in a dose-dependent fashion by the addition of serum. The ratio of triglyceride lipase to phospholipase activity of EL was 0.65, compared with ratios of 24.1 for HL and 139.9 for LPL, placing EL at the opposite end of the lipolytic spectrum from LPL. Neither lipase activity of EL was influenced by the addition of apolipoprotein C-II (apoC-II), indicating that EL, like HL, does not require apoC-II for activation. Like LPL but not HL, both lipase activities of EL were inhibited by 1 M NaCl. The relative ability of EL, versus HL and LPL, to hydrolyze lipids in isolated lipoprotein fractions was also examined using generation of FFAs as an end point. As expected, based on the relative triglyceride lipase activities of the three enzymes, the triglyceride-rich lipoproteins, chylomicrons, VLDL, and IDL, were efficiently hydrolyzed by LPL and HL. EL hydrolyzed HDL more efficiently than the other lipoprotein fractions, and LDL was a poor substrate for all of the enzymes.     

Kinetic and structural characterization of triacylglycerol lipases possessing phospholipase A1 activity

The pancreatic lipase gene family displays various substrate selectivities for triglycerides and phospholipids. The structural basis for this difference in substrate specificity has not been definitively established. Based on a kinetic comparative study between various pancreatic lipase family members, we showed here that porcine pancreatic lipase (PPL), which was so far classified as “classical lipase”, was able to hydrolyze phosphatidylcholine (PC). Amino acid sequence alignments revealed that Val260 residue in PPL lid could be critical for the interaction with lipid substrate. Molecular dynamics was applied to investigate PC binding modes within the catalytic cavity of PPL and human pancreatic lipase (HPL), aiming to explain the difference of specificity of these enzymes towards phospholipids. Results showed that with HPL, the oxyanion hole was not able to accommodate the PC molecule, suggesting that no activity could be obtained. With PPL, the formation of a large pocket involving Val260 allowed the PC molecule to come near the catalytic residues, suggesting that it could be hydrolyzed. One more interesting finding is that human pancreatic lipase related protein 2 could hydrolyze phospholipids through its PLA₁ and PLA₂ activities. Overall, our study shed the light on new structural features of the phospholipase activity of pancreatic lipase family members.     

Increased fat mass and insulin resistance in mice lacking pancreatic lipase-related protein 1

Pancreatic triglyceride lipase (PTL) and its cofactor, colipase, are required for efficient dietary triglyceride digestion. In addition to PTL, pancreatic acinar cells synthesize two pancreatic lipase-related proteins (PLRP1 and PLRP2), which have a high degree of sequence and structural homology with PTL. The lipase activity of PLRP2 has been confirmed, whereas no known triglyceride lipase activity has been detected with PLRP1 up to now. To explore the biological functions of PLRP1 in vivo, we generated Plrp1 knockout (KO) mice in our laboratory. Here we show that the Plrp1 KO mice displayed mature-onset obesity with increased fat mass, impaired glucose clearance and the resultant insulin resistance. When fed on high-fat (HF) diet, the Plrp1 KO mice exhibited an increased weight gain, fat mass and severe insulin resistance compared with wild-type mice. Pancreatic juice extracted from Plrp1 KO mice had greater ability to hydrolyze triglyceride than that from the wild-type littermates. We propose that PLRP1 may function as a metabolic inhibitor in vivo of PLT-colipase-mediated dietary triglyceride digestion and provides potential anti-obesity targets for developing new drugs.     

High expression in adult horse of PLRP2 displaying a low phospholipase activity

The physiological role of the two lipase-related proteins, PLRP1 and PLRP2, still remains obscure although some propositions have been made concerning PLRP2. In this paper, we report the presence of high amounts of PLRP2 in adult horse pancreas whereas no PLRP1 could be detected. As well, a non-parallel expression of PLRP2 and PLRP1 is observed in adult cat and dog, since no PLRP2 could be detected in these two species. In adult ox, neither PLRP2 nor PLRP1 could be found. These findings are in favor of a different regulation of the expression of the genes encoding pancreatic lipase and the related proteins according to the species. The cDNA encoding horse PLRP2 has been cloned and the protein expressed in insect cells. Both native and recombinant PLRP2 display the same catalytic properties. They possess a moderate lipase activity, inhibited by bile salts and not restored by colipase. Interestingly, they differ from PLRP2 from other species by their very low phospholipase activity indicating that PLRP2 could not be considered as a general phospholipase as previously postulated. This work highlights the variability of the properties of PLRP2 and rises the question of the physiological function of this protein in adult according to the species.     

Lipase maturation factor 1 is required for endothelial lipase activity

Lipase maturation factor 1 (Lmf1) is an endoplasmic reticulum (ER) membrane protein involved in the posttranslational folding and/or assembly of lipoprotein lipase (LPL) and hepatic lipase (HL) into active enzymes. Mutations in Lmf1 are associated with diminished LPL and HL activities ("combined lipase deficiency") and result in severe hypertriglyceridemia in mice as well as in human subjects. Here, we investigate whether endothelial lipase (EL) also requires Lmf1 to attain enzymatic activity. We demonstrate that cells harboring a (cld) loss-of-function mutation in the Lmf1 gene are unable to generate active EL, but they regain this capacity after reconstitution with the Lmf1 wild type. Furthermore, we show that cellular EL copurifies with Lmf1, indicating their physical interaction in the ER. Finally, we determined that post-heparin phospholipase activity in a patient with the LMF1(W464X) mutation is reduced by more than 95% compared with that in controls. Thus, our study indicates that EL is critically dependent on Lmf1 for its maturation in the ER and demonstrates that Lmf1 is a required factor for all three vascular lipases, LPL, HL, and EL.     

Endothelial lipase: a new lipase on the block

Endothelial lipase (EL) is a newly described member of the triglyceride lipase gene family. It has a considerable molecular homology with lipoprotein lipase (LPL) (44%) and hepatic lipase (HL) (41%). Unlike LPL and HL, this enzyme is synthesized by endothelial cells and functions at the site where it is synthesized. Furthermore, its tissue distribution is different from that of LPL and HL. As a lipase, EL has primarily phospholipase A1 activity. Animals that overexpress EL showed reduced HDL cholesterol levels. Conversely, animals that are deficient in EL showed a marked elevation in HDL cholesterol levels, suggesting that it plays a physiologic role in HDL metabolism. Unlike LPL and HL, EL is located in the vascular endothelial cells and its expression is highly regulated by cytokines and physical forces, suggesting that it may play a role in the development of atherosclerosis. However, there is only a limited amount of information available about this enzyme. Some of our unpublished data in addition to previously published data support the possibility that the enzyme plays a role in the formation of atherosclerotic lesion.     

Pancreatic lipase-related protein-2 (PLRP2) can contribute to dietary fat digestion in human newborns

In newborn mice, PLRP2 is essential for fat digestion. In human infants, the role of PLRP2 in fat digestion is unclear, as it has poor activity against long-chain triglycerides in vitro. Also, many infants carry a genetic polymorphism resulting in a truncated protein, PLRP2 W340X, which may impact function significantly. We re-examined the properties of recombinant human PLRP2 and studied the impact of W340X mutation on its function. In the presence of bile salt micelles and colipase, human PLRP2 hydrolyzed long-chain tri-, di-, and monoglycerides. It hydrolyzed triolein at a level much lower than that of pancreatic triglyceride lipase, but close to that of carboxyl ester lipase, after a long lag phase, which could be eliminated by the addition of oleic acids. Human PLRP2 W340X was poorly secreted and largely retained inside the cell. The retention of the mutant protein triggered endoplasmic reticulum stress and unfolded protein responses. Our results show that earlier studies underestimated human PLRP2 activity against triolein by employing suboptimal assay conditions. In vivo, dietary fat emulsions contain fatty acids as a result of the action of gastric lipase. Consequently, PLRP2 can contribute to fat digestion during early infancy. Furthermore, infants with homozygous W340X alleles will not secrete functional PLRP2 and may have inefficient dietary fat digestion, particularly when breastfeeding is unavailable. Additionally, the aberrant folding of W340X mutant may cause chronic cellular stress and increase susceptibility of pancreatic exocrine cells to other metabolic stressors.     

Structure of human pancreatic lipase-related protein 2 with the lid in an open conformation

Access to the active site of pancreatic lipase (PL) is controlled by a surface loop, the lid, which normally undergoes conformational changes only upon addition of lipids or amphiphiles. Structures of PL with their lids in the open and functional conformation have required cocrystallization with amphiphiles. Here we report two crystal structures of wild-type and unglycosylated human pancreatic lipase-related protein 2 (HPLRP2) with the lid in an open conformation in the absence of amphiphiles. These structures solved independently are strikingly similar, with some residues of the lid being poorly defined in the electron-density map. The open conformation of the lid is however different from that previously observed in classical liganded PL, suggesting different kinetic properties for HPLRP2. Here we show that the HPLRP2 is directly inhibited by E600, does not present interfacial activation, and acts preferentially on substrates forming monomers or small aggregates (micelles) dispersed in solution like monoglycerides, phospholipids and galactolipids, whereas classical PL displays reverse properties and a high specificity for unsoluble substrates like triglycerides and diglycerides forming oil-in-water interfaces. These biochemical properties imply that the lid of HPLRP2 is likely to spontaneously adopt in solution the open conformation observed in the crystal structure. This open conformation generates a large cavity capable of accommodating the digalactose polar head of galactolipids, similar to that previously observed in the active site of the guinea pig PLRP2, but absent from the classical PL. Most of the structural and kinetic properties of HPLRP2 were found to be different from those of rat PLRP2, the structure of which was previously obtained with the lid in a closed conformation. Our findings illustrate the essential role of the lid in determining the substrate specificity and the mechanism of action of lipases.     

Comparative study on digestive lipase activities on the self emulsifying excipient Labrasol, medium chain glycerides and PEG esters

Labrasol is a lipid-based self-emulsifying excipient used in the preparation of lipophilic drugs intended for oral delivery. It is mainly composed of PEG esters and glycerides with medium acyl chains, which are potential substrates for digestive lipases. The hydrolysis of Labrasol by porcine pancreatic extracts, human pancreatic juice and several purified digestive lipases was investigated in the present study. Classical human pancreatic lipase (HPL) and porcine pancreatic lipase, which are the main lipases involved in the digestion of dietary triglycerides, showed very low levels of activity on the entire Labrasol excipient as well as on separated fractions of glycerides and PEG esters. On the other hand, gastric lipase, pancreatic lipase-related protein 2 (PLRP2) and carboxyl ester hydrolase (CEH) showed high specific activities on Labrasol. These lipases were found to hydrolyze the main components of Labrasol (PEG esters and monoglycerides) used as individual substrates, whereas these esters were found to be poor substrates for HPL. The lipolytic activity of pancreatic extracts and human pancreatic juice on Labrasol(R) is therefore mainly due to the combined action of CEH and PLRP2. These two pancreatic enzymes, together with gastric lipase, are probably the main enzymes involved in the in vivo lipolysis of Labrasol taken orally.     

Higher order lipase gene association with plasma triglycerides

Lipoprotein lipase, HL, and endothelial lipase (EL) are proteoglycan-bound enzymes that regulate plasma lipoprotein levels through coordinated triglyceride (TG) lipase and phospholipase activity. We hypothesized that single nucleotide polymorphisms (SNPs) in lipase genes would have higher order impact on plasma lipoproteins beyond the influence of individual SNPs. In a sample of asymptomatic Caucasian subjects (n = 738), we used a two-stage approach, first identifying groups of subjects with similar multilocus lipase genotypes and then characterizing the relationships between genotype groups and plasma lipids. Using complementary methods, including a permutation test procedure and a mixed-effects modeling approach, we found a higher order interaction between four SNPs in three lipase genes (EL 2,237 3' untranslated region, EL Thr111Ile, HL -514C/T, and LPL HindIII) and plasma TG levels. Subjects who were heterozygous for all four lipase SNPs had significantly higher plasma TG levels beyond the effect of individual lipase SNPs and environmental factors, even after correcting for multiple comparisons. In conclusion, lipase genes had synergistic association with plasma TG beyond individual gene effects. Higher order multilocus genotype contributions to dyslipidemia and atherosclerotic cardiovascular disease need to be considered a priori because they may have an important effect even in the absence of significant main effects of the individual genes.