Deuterium nuclear magnetic resonance studies of bile salt/phosphatidylcholine mixed micelles

Mixed micelles of deoxycholate (DOC) and 1-palmitoyl-2-oleoylphosphatidylcholine (POPC) have been prepared in which the POPC was specifically deuterated in the 2-, 6-, 10-, or 16-position of the palmitoyl chain or in the N-methyl position of the choline head group. The deuterium nuclear magnetic resonance (2H NMR) spectrum of each of these specifically deuterated mixed micelles consists of a singlet whose line width depends upon the position of deuteration. Spin-spin relaxation times indicate a gradient of mobility along the POPC palmitoyl chain in the mixed micelle, with a large increase in mobility on going from the 10- to the 16-position. Spin-lattice relaxation times (T1's) demonstrate a similar gradient of mobility. Both trends in NMR relaxation behavior are consistent with a bilayer arrangement for the solubilized POPC. 2H T1 times for DOC/POPC micelles are significantly shorter than those measured in other bilayer systems, indicating unusually tight phospholipid acyl chain packing in the mixed micelle.     

Structure of human milk bile salt activated lipase

The structure and some functional sites of human milk bile salt activated lipase (BAL) were studied by cDNA cloning and chemical analysis of the enzyme. Eighteen cDNA clones of human BAL were identified from lactating human breast cDNA libraries in lambda gt11 and lambda gt10 with antibody and synthetic oligonucleotides as probes. The sequence of four clones was sufficient to construct a 3018-bp BAL cDNA structure. This sequence codes for an open reading frame of 742 amino acid residues. There is a putative signal sequence of 20 residues which is followed by the amino-terminal sequence of BAL, and the mature BAL contains 722 amino acid residues. The cDNA sequence also contains a 678-base 5'-untranslated sequence, a 97-base 3'-untranslated region, and a 14-base poly(A) tail. The sequence of a 1.8-kbp insert of clone G10-4A differs from that of the other cDNA in that it contains a deletion of 198 bases (1966-2163) corresponding to 66 amino acid residues. By use of BAL cDNA as probe, it was found that the major molecular species of BAL mRNA in human mammary gland HBL-100 cells had a size of 2.9 kb and two minor species had sizes of 3.8 and 5.1 kb by Northern blot analyses. The deduced BAL protein structure contains in the carboxyl-terminal region 16 repeating units of 11 amino acids each. The repeating units have the basic structure Pro-Val-Pro-Pro-Thr-Gly-Asp-Ser-Gly-Ala-Pro with only minor substitutions. The amino acid sequence of human BAL is related to that of pancreatic lysophospholipase, cholesterol esterase, cholinesterase, acetylcholinesterase, and thyroglobulin. Ten of the 14 cyanogen bromide fragments of diisopropyl fluorophosphate inhibited human milk BAL were isolated, determined for N-terminal sequences, analyzed for amino sugars, and tested for some functional properties. These chemical studies established that the active site of human milk BAL is located at serine-194, the N-glycosylation site is present at asparagine-187, the O-glycosylation region is in the 16 repeating units near the C-terminus, and the heparin binding domain is in the N-terminal region. We have also determined the location of disulfide bridges as Cys64-Cys80 and Cys246-Cys257. The cyanogen bromide cleavage and the partial sequencing of CNBr peptides also confirmed the location of methionines in the polypeptide chain as well as the deduced cDNA sequence of BAL.     

Studies in bile salt solutions. Deoxycholate stimulation of human milk lipase

Stimulation of human milk lipase by deoxycholate and its taurine and glycine conjugates was demonstrated by measuring the esterolysis reaction of 4-nitrophenylacetate. The steroidal surfactants did not bind strongly to the polar substrate but they did bind effectively to a hydrophobic site on the enzyme and these bile salt-enzyme complexes were effective catalysts. These results are compared with those for stimulation of the enzyme by cholate surfactants and it has been demonstrated that the absence of a 7 alpha-OH substituent on the steroid nucleus does not prevent stimulation of either the esterolytic or lipolytic activity of the enzyme.     

500 MHz H-NMR studies of bile salt-phosphatidylcholine mixed micelles and vesicles Evidence for differential motional restraint on bile salt and phosphatidylcholine resonances

¹H nuclear magnetic resonance (NMR) spectra at 500 MHz have been obtained for taurocholate/egg phosphatidylcholine mixtures of varying composition. The excellent chemical shift dispersion permits identification of most resonances for each component. This high-resolution character of the NMR spectra is retained until the phosphatidylcholine (PC) mole fraction exceeds 60–70% (the exact limit depends on ionic strength). ¹H linewidths have been monitored as a function of solute composition in order to evaluate trends in local molecular mobility of each component as the distribution of aggregate particles is varied, and to examine the effects of added NaCl in altering micellar size and shape. Although prior light scattering studies (Mazer, N.A., Benedek, G.B. and Carey, M.C. (1980) Biochemistry 19, 601–615) and our own work indicate a 6-fold increase in particle hydrodynamic radius from pure taurocholate micelles to 1 : 1 taurocholate/PC mixtures containing 150 mM NaCl, both lipid components retain substantial motional freedom and exhibit narrow NMR signals in this compositional region. As the solubilization limit for PC is approached (approx. 2:1 PC:taurocholate), differential behavior is observed for the two components: the motion of taurocholate becomes preferentially restricted, while polar portions of the PC remain mobile until large multilayers predominate.     

Pancreatic bile salt dependent lipase from cod (Gadus morhua): purification and properties.

The enzymatic basis for cod digestive lipolysis has been investigated. Lipase activity was found in aqueous extracts from pyloric caeca as well as in pancreatic tissue surrounding the caeca and the bile duct. A bile salt-dependent lipase (BSDL) was purified from either defatted powder of cod pyloric caeca or aqueous pancreatic extracts by combined affinity chromatography on cholate-Sepharose and gel filtration on Sephacryl S-200 HR. By SDS-PAGE analysis the molecular weight of purified cod BSDL was estimated to 60 kDa. The enzyme was totally dependent on bile salts for hydrolysis of insoluble fatty acid esters. Antiserum raised against purified cod BSDL reacted specifically with selected mammalian pancreatic BSDLs by Western blot analysis. Results presented in this paper strongly suggest that the bile salt-dependent lipase is the only pancreatic enzyme involved in lipid digestion in cod. The enzyme has been characterized and compared to human pancreatic BSDL with respect to substrate specificity, temperature- and pH-dependence and inhibitors. Both soluble and insoluble fatty acid esters were hydrolysed and the enzyme was 1,3-specific in hydrolysis of triolein. The enzyme was inhibited by di-isopropyl fluorophosphate and phenyl boronic acid, but not significantly by phenyl methyl sulfonyl fluoride. The cod BSDL is probably homologous to mammalian pancreatic BSDLs.     

500 MHz 1H-NMR studies of bile salt-phosphatidylcholine mixed micelles and vesicles Evidence for differential motional restraint on bile salt and phosphatidylcholine resonances

¹H nuclear magnetic resonance (NMR) spectra at 500 MHz have been obtained for taurocholate/egg phosphatidylcholine mixtures of varying composition. The excellent chemical shift dispersion permits identification of most resonances for each component. This high-resolution character of the NMR spectra is retained until the phosphatidylcholine (PC) mole fraction exceeds 60–70% (the exact limit depends on ionic strength). ¹H linewidths have been monitored as a function of solute composition in order to evaluate trends in local molecular mobility of each component as the distribution of aggregate particles is varied, and to examine the effects of added NaCl in altering micellar size and shape. Although prior light scattering studies (Mazer, N.A., Benedek, G.B. and Carey, M.C. (1980) Biochemistry 19, 601–615) and our own work indicate a 6-fold increase in particle hydrodynamic radius from pure taurocholate micelles to 1 : 1 taurocholate/PC mixtures containing 150 mM NaCl, both lipid components retain substantial motional freedom and exhibit narrow NMR signals in this compositional region. As the solubilization limit for PC is approached (approx. 2:1 PC:taurocholate), differential behavior is observed for the two components: the motion of taurocholate becomes preferentially restricted, while polar portions of the PC remain mobile until large multilayers predominate.     

Protons @ interfaces: implications for biological energy conversion

The review focuses on the anisotropy of proton transfer at the surface of biological membranes. We consider (i) the data from "pulsed" experiments, where light-triggered enzymes capture or eject protons at the membrane surface, (ii) the electrostatic properties of water at charged interfaces, and (iii) the specific structural attributes of proton-translocating enzymes. The pulsed experiments revealed that proton exchange between the membrane surface and the bulk aqueous phase takes as much as about 1 ms, but could be accelerated by added mobile pH-buffers. Since the accelerating capacity of the latter decreased with the increase in their electric charge, it was concluded that the membrane surface is separated from the bulk aqueous phase by a barrier of electrostatic nature. The barrier could arise owing to the water polarization at the negatively charged membrane surface. The barrier height depends linearly on the charge of penetrating ions; for protons, it has been estimated as about 0.12 eV. While the proton exchange between the surface and the bulk aqueous phase is retarded by the interfacial barrier, the proton diffusion along the membrane, between neighboring enzymes, takes only microseconds. The proton spreading over the membrane is facilitated by the hydrogen-bonded networks at the surface. The membrane-buried layers of these networks can eventually serve as a storage/buffer for protons (proton sponges). As the proton equilibration between the surface and the bulk aqueous phase is slower than the lateral proton diffusion between the "sources" and "sinks", the proton activity at the membrane surface, as sensed by the energy transducing enzymes at steady state, might deviate from that measured in the adjoining water phase. This trait should increase the driving force for ATP synthesis, especially in the case of alkaliphilic bacteria.     

ProFace: a server for the analysis of the physicochemical features of protein-protein interfaces

Background  

Molecular recognition is all pervasive in biology. Protein molecules are involved in enzyme regulation, immune response, signal transduction, oligomer assembly, etc. Delineation of physical and chemical features of the interface formed by protein-protein association would allow us to better understand protein interaction networks on one hand, and to design molecules that can engage a given interface and thereby control protein function on the other hand.     

Structural basis for bile salt inhibition of pancreatic phospholipase A2

Bile salt interactions with phospholipid monolayers of fat emulsions are known to regulate the actions of gastrointestinal lipolytic enzymes in order to control the uptake of dietary fat. Specifically, on the lipid/aqueous interface of fat emulsions, the anionic portions of amphipathic bile salts have been thought to interact with and activate the enzyme group-IB phospholipase A2 (PLA2) derived from the pancreas. To explore this regulatory process, we have determined the crystal structures of the complexes of pancreatic PLA2 with the naturally occurring bile salts: cholate, glycocholate, taurocholate, glycochenodeoxycholate, and taurochenodeoxycholate. The five PLA2-bile salt complexes each result in a partly occluded active site, and the resulting ligand binding displays specific hydrogen bonding interactions and extensive hydrophobic packing. The amphipathic bile salts are bound to PLA2 with their polar hydroxyl and sulfate/carboxy groups oriented away from the enzyme's hydrophobic core. The impaired catalytic and interface binding functions implied by these structures provide a basis for the previous numerous observations of a biphasic dependence of the rate of PLA2 catalyzed hydrolysis of zwitterionic glycerophospholipids in the presence of bile salts. The rising or activation phase is consistent with enhanced binding and activation of the bound PLA2 by the bile salt induced anionic charge in a zwitterionic interface. The falling or inhibitory phase can be explained by the formation of a catalytically inert stoichiometric complex between PLA2 and any bile salts in which it forms a stable complex. The model provides new insight into the regulatory role that specific PLA2-bile salt interactions are likely to play in fat metabolism.     

Activation of bile salt dependent lipase by (lyso)phosphatidic acid and platelet activating factor

The activity of breast milk BSDL was assayed with or without phospholipids as extra-intestinal effector candidates. Phosphatidic acid, lysophosphatidic acid and platelet activating factor but not phosphatidylcholine and lysophosphatidylcholine stimulated BSDL activity at least as efficiently as taurocholate. The apparent dissociation constants of PA and LPA at saturating concentrations of three different substrates were between 0.1 and 13.4 μM and that of PAF was below or equal to 200 pM. Kinetic data suggested the existence of at least one binding site for each of these effectors. PA, LPA and PAF are likely extra-intestinal modulators of BSDL activity.     

Preparation of monodisperse vesicles with variable size by dilution of mixed micellar solutions of bile salt and phosphatidylcholine

Uniformly 15N-labelled triostin A and echinomycin have been prepared by growing the producing organisms on enriched media and their 15N nuclear magnetic resonance spectra partially assigned by a combination of nuclear Overhauser effect and scalar coupling constant measurements. Selective feeding experiments using unlabelled L-tryptophan-supplemented media have shown that N-1 and N-4 of the quinoxaline rings have their origins in the indole and amino groups of tryptophan, respectively.     

Asymmetric distribution of phosphatidylcholine and sphingomyelin between micellar and vesicular phases. Potential implications for canalicular bile formation.

Both phosphatidylcholine (PC) and sphingomyelin (SM) are the major phospholipids in the outer leaflet of the hepatocyte canalicular membrane. Yet, the phospholipids secreted into bile consist principally (>95%) of PC. In order to understand the physical;-chemical basis for preferential biliary PC secretion, we compared interactions with bile salts (taurocholate) and cholesterol of egg yolk (EY)SM (mainly 16:0 acyl chains, similar to trace SM in bile), buttermilk (BM)SM (mainly saturated long (>20 C-atoms) acyl chains, similar to canalicular membrane SM) and egg yolk (EY)PC (mainly unsaturated acyl chains at sn-2 position, similar to bile PC). Main gel to liquid-crystalline transition temperatures were 33. 6 degrees C for BMSM and 36.6 degrees C for EYSM. There were no significant effects of varying phospholipid species on micellar sizes or intermixed-micellar/vesicular bile salt concentrations in taurocholate-phospholipid mixtures (3 g/dL, 37 degrees C, PL/BS + PL = 0.2 or 0.4). Various phases were separated from model systems containing both EYPC and (EY or BM)SM, taurocholate, and variable amounts of cholesterol, by ultracentrifugation with ultrafiltration and dialysis of the supernatant. At increasing cholesterol content, there was preferential distribution of lipids and enrichment with SM containing long saturated acyl chains in the detergent-insoluble pelletable fraction consisting of aggregated vesicles. In contrast, both micelles and small unilamellar vesicles in the supernatant were progressively enriched in PC. Although SM containing vesicles without cholesterol were very sensitive to micellar solubilization upon taurocholate addition, incorporation of the sterol rendered SM-containing vesicles highly resistant against the detergent effects of the bile salt. These findings may have important implications for canalicular bile formation.     

Motility of bile canaliculi in the living animal: implications for bile flow

Modern fluorescence microscopic techniques were used to image the bile canalicular system in the intact rat liver, in vivo. By combining the use of sodium fluorescein secretion into bile, with digitally enhanced fluorescence microscopy and time-lapse video, it was possible to capture and record the canalicular motility events that accompany the secretion of bile in life. Active bile canalicular contractions were found predominantly in zone 1 (periportal) hepatocytes of the liver. The contractile movements were repetitive, forceful, and appeared unidirectional moving bile in a direction towards the portal bile ducts. Contractions were not seen in the network of canaliculi on the surface of the liver. Cytochalasin B administration resulted in reduced canalicular motility, progressive dilation of zone 1 canaliculi, and impairment of bile flow. Canalicular dilations invariably involved the branch points of the canalicular network. The findings add substantively to previous in vitro studies using couplets, and suggest that canalicular contractions contribute physiologically to bile flow in the liver.     

Genetic analysis of two bile salt hydrolase activities in Lactobacillus acidophilus NCFM

Two genes, bshA and bshB, encoding bile salt hydrolase enzymes (EC 3.5.1.24) were identified in the genome sequence of Lactobacillus acidophilus NCFM. Targeted inactivation of these genes via chromosomal insertion of an integration vector demonstrated different substrate specificities for these two enzymes.     

Binding of bis-ANS to Bacillus subtilis lipase: A combined computational and experimental investigation

8-Anilino-1-naphthalene sulfonate (ANS) and its covalent dimer bis-ANS are widely used for titrating hydrophobic surfaces of proteins. Interest to understand the nature of interaction of these dyes with proteins was seriously pursued. However as the techniques used in these studies varied, they often provided varied information regarding stoichiometry, binding affinity, actual binding sites etc. In the present study, we used combination of computation methods (docking and MD simulation) and experimental methods (mutations, steady-state and time-resolved fluorescence) to investigate bis-ANS interaction with Bacillus subtilis lipase. We identified seven binding sites for bis-ANS on lipase using computational docking and MD simulation and verified these data using a set of single amino acid substituted mutants. Docking and MD simulation studies indicated that the binding sites were various indentations and grooves on protein surface with hydrophobic characteristics. Both hydrophobic and ionic interactions were involved in each of these binding events. We further examine the fluorescence properties of bis-ANS bound to mutant lipases that either gained or lost a binding site. Our results indicated that neither gain nor loss of single binding site caused any change in fluorescence lifetimes (and their relative amplitudes) of mutant lipase-bound bis-ANS in comparison to that bound to wild type; hence, it suggested that nature of bis-ANS binding to each of the sites in lipase was very similar.     

A model of pancreatic lipase and the orientation of enzymes at interfaces

A model for the interfacial orientation and the mode of action of lipase is proposed. Lipase is oriented so that its active site is near the oil-water boundary. This orientation is achieved by oil-enzyme bonding at the “hydrophobic head” of the enzyme, a region free of electric charges and relatively resistant to unfolding. The measured K_M is a complex constant including the dissociation constant of this oil-enzyme “complex”. The interfacial orientation of lipase is further aided by hydrophilic negative charges on the “back” of the enzyme and by a hydrophilic carbohydrate “tail”.It is suggested that similar hydrophobic heads and hydrophilic tails and asymmetric charge distributions establish the orientation of many enzymes which act at interfaces. Many phospholipases, for instance, appear to be charge-oriented, and the carbohydrate residues of ribonucleases and many other glycoproteins may be hydrophilic tails.Lipase is probably a serine enzyme with a catalytic center similar to that of chymotrypsin, but more hindered, perhaps owing to the presence of a leucine residue, and there is no binding of substrate lipid chains in the “active complex”. The substrate molecule is fixated on the enzyme in a two-dimensional orientation, because its leaving alkoxy group must be received by the serine hydroxyl hydrogen which is directed towards the imidazol ring of the reactive histidine through a hydrogen bond. The high turnover rate of lipolysis, 5 × 10⁵/min, exceptional even for an enzyme, results from the extremely high substrate concentration near the active site, and from an almost complete extrusion of water because of the hydrophobicity of both the active site and the substrate. In addition, both substrate and enzyme, because of their polarity, are already so favorably positioned at the interface that the formation of the “active complex” requires only a proper two-dimensional alignment, perhaps with partial extraction of the substrate molecule from the lipid phase.     

Limitations of phosphatidylcholine/deoxycholate mixtures for the analysis of phospholipase A2 inhibition and activation: illustration with annexins.

The effects of human recombinant lipocortin I (annexin I) and bovine lung calpactin I (annexin II) on porcine pancreatic phospholipase A2 (PLA2) activity in phosphatidylcholine (PC)/deoxycholate (DOC) mixtures were investigated. Annexin-associated decreases in PLA2 activity were observed under some conditions, for example, at high DOC/PC molar ratios; however, activation was observed under other conditions. NaCl, which lowers the non-critical micellar concentration (NCMC) of deoxycholate, caused significant decreases in control PLA2 activity in the absence of annexins, and greater decreases in PLA2 activity in annexin-containing samples, resulting in an apparent increase in inhibition. The PC/DOC substrate mixtures themselves appeared unstable. Despite a large excess of detergent, precipitates were, at times, observed upon incubation of some PC/DOC mixtures at 37 degrees C. Such behavior is of interest in view of the numerous reports of PLA2 inhibition by annexins and annexin-derived peptides in the PC/DOC system. The influence of the annexins on activity in this system is consistent with effects on the phase behavior of the PC/DOC mixture and/or competition with the enzyme for available Ca2+. These results caution against use of the PC/DOC system for analysis of potential PLA2 inhibitors unless the phase behavior of the system is more fully delineated.     

Medium-chain versus long-chain triacylglycerol emulsion hydrolysis by lipoprotein lipase and hepatic lipase: implications for the mechanisms of lipase action

To explore how enzyme affinities and enzyme activities regulate hydrolysis of water-insoluble substrates, we compared hydrolysis of phospholipid-stabilized emulsions of medium-chain (MCT) versus long-chain triacylglycerols (LCT). Because substrate solubility at the emulsion surface might modulate rates of hydrolysis, the ability of egg yolk phosphatidylcholine to solubilize MCT was examined by NMR spectroscopy. Chemical shift measurements showed that 11 mol % of [13C]carbonyl enriched trioctanoin was incorporated into phospholipid vesicles as a surface component. Similar methods with [13C]triolein showed a maximum solubility in phospholipid bilayers of 3 mol % (Hamilton & Small, 1981). Line widths of trioctanoin surface peaks were half that of LCT, and relaxation times, T1, were also shorter for trioctanoin, showing greater mobility for MCT in phospholipid. In assessing the effects of these differences in solubility on lipolysis, we found that both purified bovine milk lipoprotein lipase and human hepatic lipase hydrolyzed MCT at rates at least 2-fold higher than for LCT. With increasing concentrations of MCT, saturation was not reached, indicating low affinities of lipase for MCT emulsions, but with LCT emulsion incubated with lipoprotein lipase, saturation was reached at relatively low concentration, demonstrating higher affinity of lipase for LCT emulsions. Differences in affinity were also demonstrated in mixed incubations where increasing amounts of LCT emulsion resulted in decreased hydrolysis of MCT emulsions. Increasing MCT emulsion amounts had little or no effect on LCT emulsion hydrolysis.(ABSTRACT TRUNCATED AT 250 WORDS)     

Theaflavin-3-gallate specifically interacts with phosphatidylcholine,forming a precipitate resistant against the detergent action of bile salt

Black tea is a highly popular beverage, and its pigments, polymerized catechins such as theaflavins (TFs), are attracting attention due to their beneficial health effects. In this study, to test the inhibitory activities of TFs on the intestinal absorption of cholesterol, we investigated their effects on phosphatidylcholine (PC) vesicles in the absence or presence of a bile salt. (?)-Epicatechin gallate, (?)-epigallocatechin gallate, and TFs formed insoluble complexes with PC vesicles. Galloylated TFs such as TF2A, TF2B, and TF3 precipitated far more than other polyphenols. The subsequent addition of taurocholate redispersed the polyphenol-PC complexes, except that a large amount of TF2A remained insoluble. After incubation with taurocholate-PC micelles, TF2A elevated the turbidity of the micelle solution, providing red sediments. The TF2A-specific effect was dependent on the PC concentration. These results suggest that TF2A interacts with PC and aggregates in a specific manner different from catechins and other TFs.     

Influence of total lipid concentration, bile salt:lecithin ratio, and cholesterol content on inter-mixed micellar/vesicular (non-lecithin-associated) bile salt concentrations in model bile.

We modified classic equilibrium dialysis methodology to correct for dialysant dilution and Donnan effects, and have systematically studied how variations in total lipid concentration, bile salt (taurocholate):lecithin (egg yolk) ratio, and cholesterol content influence inter-mixed micellar/vesicular (non-lecithin-associated) concentrations (IMC) of bile salts (BS) in model bile. To simulate large volumes of dialysant, the total volume (1 ml) of model bile was exchanged nine times during dialysis. When equilibrium was reached, dialysate BS concentrations plateaued, and initial and final BS concentrations in the dialysant were identical. After corrections for Donnan effects, IMC values were appreciably lower than final dialysate BS concentrations. Quasielastic light scattering was used to validate these IMC values by demonstrating that lipid particle sizes and mean scattered light intensities did not vary when model biles were diluted with aqueous BS solutions of the appropriate IMC. Micelles and vesicles were separated from cholesterol-supersaturated model bile, utilizing high performance gel chromatography with an eluant containing the IMC. Upon rechromatography of micelles and vesicles using an identical IMC, there was no net transfer of lipid between micelles and vesicles. To simulate dilution during gel filtration, model biles were diluted with 10 mM Na cholate, the prevailing literature eluant, resulting in net transfer of lipid between micelles and vesicles, the direction of which depended upon total lipid concentration and BS/lecithin ratio. Using the present methodology, we demonstrated that inter-mixed micellar/vesicular concentrations (IMC) values increased strongly (5 to 40 mM) with increases in both bile salt (BS):lecithin ratio and total lipid concentration, whereas variations in cholesterol content had no appreciable effects. For model biles with typical physiological biliary lipid compositions, IMC values exceeded the critical micellar concentration of the pure BS, implying that in cholesterol-supersaturated biles, simple BS micelles coexist with mixed BS/lecithin/cholesterol micelles and cholesterol/lecithin vesicles. We believe that this methodology allows the systematic evaluation of IMC values, with the ultimate aim of accurately separating micellar, vesicular, and potential other cholesterol-carrying particles from native bile.