Isolation and characterization of hydrophobic polypeptides in human bile


M. Stark, Department of Medical Biochemistry and Biophysics, Karolinska Institutet, S-171 77 Stockholm, Sweden. Fax: + 46 8 33 7462, Tel.: + 46 8 728 7699, E-mail:


Polypeptides were isolated from human bile by extraction with chloroform/methanol, followed by reversed-phase chromatography in methanol/ethylene chloride and gel filtration in chloroform/methanol. Peptides were characterized by SDS/PAGE, sequence analysis and matrix-assisted laser desorption ionization/time-of-flight mass spectrometry. This identified haemoglobin α chain, ATP synthase lipid-binding protein subunit 9, an N-terminal fragment of mac25/insulin-like growth factor-binding protein 7 and an internal fragment of monocyte differentiation antigen CD14, all not described previously in bile. In addition, α1-antitrypsin, known in bile from previous work, was also identified. The hydrophobic character of haemoglobin α chain is not apparent from its amino acid sequence, but the other polypeptides all have major hydrophobic segments. These results show that several proteins are removed upon organic solvent extraction used for delipidation during the preparation of samples for proteome analysis. Several of the polypeptides found are unexpectedly present in bile, suggesting that specific excretion mechanisms may be involved.




matrix-assisted laser desorption ionization/time-of-flight

The protein content in human gallbladder bile is ≈ 5% by dry weight [1]. Many of the proteins identified in bile are derived from serum, but they may also originate from hepatocytes or the bile duct epithelium [2]. Two-dimensional PAGE of rat or human bile shows a population of at least 30 major proteins, and up to several hundred protein spots can be visualized by silver staining. Examples of bile proteins already identified after two-dimensional electrophoresis are albumin, transferrin, haptoglobin, α2-macroglobulin, immunoglobulins, α-fibrinogen, β-fibrinogen and α1-antitrypsin [3,4]. However, they are all common serum proteins and are probably of little special interest in the present context. Anionic polypeptide fraction and calcium-binding protein [5,6], apolipoproteins [7] and complement factors [8] have been identified previously by SDS/PAGE and immunoassays while fragments of C1 inhibitor and α1-antitrypsin [9] have been identified by sequence analysis.

In the pathogenesis of human cholesterol gallstones, gallbladder bile supersaturated with cholesterol is necessary. This cannot account solely for gallstone formation, however, because a large fraction of the human population has biliary cholesterol supersaturation but does not form gallstones [10,11]. Apparently the presence of a nucleating factor [12] or nucleation inhibiting factor is essential. Some proteins in bile (e.g. glycoproteins as immunoglobulins) act as pronucleating agents [11,13] and others (e.g. apolipoprotein A-I and A-II) act as antinucleating agents [14,15]. It has been proposed that the more hydrophobic the protein, the higher the antinucleating potency, and vice versa [16].

The aim of the present study was to identify hydrophobic polypeptides in human gallbladder bile by analysis of the unpolar phase formed from the two-phase system chloroform/methanol/bile (8 : 4 : 3, v/v/v).

Materials and methods


Sephadex LH-60 was from Pharmacia Biotech and Lipidex-5000 from Packard Instrument. Precast Tricine gels and Mark 12 molecular mass standard were purchased from Novex. All solvents and chemicals were of the highest available quality. Glass columns equipped with a glass filter and Teflon tap were used for chromatographies in organic solvents.

Isolation of hydrophobic polypeptides

Human gallbladder bile was collected (with permission from the ethical committee of Karolinska Institutet) during cholecystectomy and stored at −20 °C until processed. It was pooled in volumes of 20–30 mL (from 1–4 patients). Cells and solids were removed by centrifugation at 3000 g for 15 min at 15 °C.

The supernatant was mixed with 4 vol. of chloroform/methanol (2 : 1, v/v), to yield a two-phase system of chloroform/methanol/bile (8 : 4 : 3, v/v/v). The phases were allowed to separate overnight at ambient temperature. The lower, unpolar phase containing lipids and lipid-associated proteins, was evaporated to dryness, redissolved in methanol/ethylene chloride, 4 : 1 (v/v), and subjected to reversed-phase chromatography on Lipidex-5000 in the same solvent system. Between 0.6 and 0.8 g was applied (not more than 9 mg·mL−1 gel), and fractions were collected at a flow rate of 75 mL·h−1 (1 drop·s−1). In this solvent system the proteins elute before the majority of the phospholipids in the first bed volume [17], while neutral lipids elute significantly later [18]. The fractions containing proteins were pooled, evaporated to dryness, redissolved and subjected to size-exclusion chromatography on Sephadex LH-60 in chloroform/methanol/0.1 m HCl (19 : 19 : 2, v/v/v). The column size was 48 × 1.1 cm, 20–40 mg was applied and fractions were collected at a flow rate of 6 mL·h−1 (1 drop per 10 s).

Reverse-phase HPLC

Separation was performed on a 4.6 mm × 25 cm C18 column with 5-µm particle size (Vydac) using a Waters instrument.

Water containing 0.1% trifluoroacetic acid was added slowly to the sample dissolved in chloroform/methanol/0.1 m HCl (3 : 48 : 47, v/v/v), in order to obtain a methanol concentration of 15%. Before injection the sample was centrifuged for 10 min at 19 000 g. Fifteen per cent methanol in water with 0.1% trifluoroacetic acid was used as the initial mobile phase, and elution was performed with a linear gradient of methanol containing 0.1% trifluoroacetic acid. The flow rate was 1 mL·min−1, the absorbance of the eluate was monitored continuously at 214 nm and peaks were collected manually.

Protein and phospholipid analysis

Acid hydrolysis was performed in 6 m HCl/0.5% phenol in evacuated glass tubes for 20–24 h at 110 °C. For Lipidex 5000 and Sephadex LH60 eluates, protein, phosphatidylethanolamine and phosphatidylserine were determined by phenylthiocarbamyl amino acid analysis [17]. HPLC eluates were analysed with a ninhydrin-based LKB 4151 Alpha Plus Instrument. SDS/PAGE was performed on precast 10–20% Tricine gels [19] under reducing conditions and the gels were stained with silver [20].

Sequence analysis

Fractions collected after LH-60 chromatography were subjected to N-terminal amino acid sequence analysis by Edman degradation in protein sequencers 470A, 477A, 494 or 494 cLC (Applied Biosystems).

Mass spectrometry

For matrix-assisted laser desorption ionization/time-of-flight (MALDI-TOF) mass spectrometry, samples were dried on a stainless steel grid together with a matrix of α-cyano-4-hydroxycinnamic acid (≈ 5 µg). Molecular masses were determined with a Finnigan MAT Lasermat 2000 instrument operated in the positive ion mode. Calibration was performed using a matrix component with m/z 190.2 [M + H]+.


Polypeptides present in the chloroform/methanol extracts of human bile were separated from the majority of the phospholipids by Lipidex-5000 chromatography (Fig. 1A). The protein-containing fractions that eluted at 30–45% of the Lipidex-5000 bed volume were further separated by gel filtration on Sephadex LH-60. The chromatogram shows two peaks containing amino acids (Fig. 1B). The early eluting proteins after Sephadex LH-60 chromatography (elution position at 15–40% of one column volume) were well separated from phospholipids and pigments, while the late eluting proteins co-eluted with phospholipids and coloured compounds (elution position at 40–60% of one column volume; Fig. 1B). Analysis of the LH-60 fractions by SDS/PAGE showed that the early eluting proteins contained a 17 kDa polypeptide and several polypeptides in the range 3–6 kDa, while the lipids that start to elute at 45% of one column volume disturbed the electrophoresis of the late eluting proteins (Fig. 2).

Figure 1.

Figure 1.

Isolation of hydrophobic bile proteins. (A) Gallbladder bile was extracted with chloroform/methanol and the unpolar phase was subjected to chromatography on Lipidex-5000 in the solvent system methanol/ethylene chloride 4 : 1. Fractions which were separated further are marked with a bar. (B) The pooled polypeptide fraction after Lipidex-5000 chromatography was subjected to gel filtration on Sephadex LH-60 in the solvent system methanol/chloroform/0.1 m HCl (19 : 19 : 2). The numbered fractions were further analysed by SDS/PAGE (Fig. 2), sequence analysis (Table 1), MALDI/TOF (Fig. 3) and RP-HPLC (Fig. 4).

Figure 2.

Figure 2.

SDS/PAGE of fractions after Sephadex LH-60 chromatography. The lanes are labelled with the fraction numbers from Fig. 1B. The lane labelled St contains molecular mass standard. The figures on the right of the figure indicate the mass in kDa. The bands around 60 kDa present in all lanes are artefacts obtained at silver staining.

The lipid-free LH-60 fractions were further analysed by N-terminal sequence analysis (Table 1) and MALDI-TOF mass spectrometry (Fig. 3). This showed that the 17 kDa polypeptide was haemoglobin α chain and that the 3–6 kDa polypeptides contained ATP synthase lipid-binding protein subunit 9, a fragment of monocyte differentiation antigen CD14 (positions 75–103), an N-terminal fragment of mac25/insulin-like growth factor-binding protein 7 (positions 1–22) and a C-terminal fragment of α1-antitrypsin (position 359–). An α1-antitrypsin fragment (359–394) has been identified previously in bile [9]. The C-terminal ends of the CD14 and mac25 fragments were determined from the masses obtained by MALDI-TOF mass spectrometry (Fig. 3). The theoretical masses of the fragments are 3052 for CD14 (75–103) and 2365 for mac25 (1–22), indicating that the ions revealed by MALDI-TOF mass spectrometry for the mac25 fragment were sodium adducts. The amount of each polypeptide in bile was estimated from the sequence analysis (Table 1).

Table 1. Results from N-terminal sequence analysis of polypeptides in individual Sephadex LH-60 fractions. Comparison between amino acid sequences determined (upper) and sequences obtained from Swissprot (lower). In cases where several proteins or fragments were obtained in the same fraction, it was possible to unambiguously follow the sequences because the polypeptides were present at different amounts. X denotes unidentified residue. Preparation I, two patients; II, one patient; III, four patients. The amounts in bile were estimated from the amounts found during sequence analysis, expressed as µg polypeptide per ml of starting material.
LH-60 fraction
Found in
Haemoglobin α chain
ATP synthase lipid-binding protein
CD14 (75-)
α1-antitrypsin (359–)
Figure 3.

Figure 3.

MALDI/TOF mass spectrum offraction 9 after Sephadex LH-60chromatography.m/z 3053 corresponds to CD1475–103[M + H]+ and m/z 2387 corresponds to mac251–22[M + Na]+.

Several attempts were made to identify the late-eluting polypeptides after LH-60 chromatography. However, because of contaminating lipids further purification was necessary. To this end, fractions 11–13 were pooled and subjected to acidic organic solvent extraction, using chloroform/methanol/0.1 m HCl. Analysis of the polar methanol/water phase and the unpolar chloroform phase showed that ≈ 30% of the protein was present in the methanol/water phase while the majority of lipids remained in the chloroform phase. The contents of the methanol/water phase were subjected to reverse-phase HPLC (Fig. 4). Two peaks, eluting at 67 and 97% methanol, respectively, contained amino acids; predominantly Asp, Ser, Glu and Gly, but Ala, Val, Met, Ile, Leu, Tyr, Phe, Lys and Arg were also detected. Further analysis of these peaks by sequence analysis revealed a high background of amino acids, but no detectable peptide sequence.

Figure 4.

Figure 4.

Reversed-phase HPLC of lipidsand peptides. Fractions 11–13 after Sephadex LH-60 chromatography (containing ≈ 70 nmol amino acids) were subjected to extraction with chloroform/methanol (2 : 1) with 0.1 m HCl. The contents of the polar methanol/water phase were separated by reversed-phase HPLC using a C18 column. The arrows indicate peaks containing amino acids.


Bile acids, phospholipids, cholesterol and bilirubin constitute ≈ 90% of the solutes in gallbladder bile. These components interact with bile proteins, disturb analytical methods and cause problems in protein purification. Two-dimensional electrophoresis is a powerful method for separating complex mixtures, but for bile and other lipid-rich samples, a delipidation procedure is indispensable. Also, in some protocols [6] for the purification of bile proteins, lipids have to be removed prior to chromatography. The delipidation procedure generally involves precipitation of the proteins with organic solvents, while lipids remain soluble. However, very hydrophobic or lipid-associated polypeptides may not precipitate and will therefore be inadvertently removed.

In the purification procedure described here, the focus was on proteins that remain in the unpolar phase after organic solvent extraction of gallbladder bile. After reverse-phase chromatography in methanol/ethylene chloride and gel-filtration in chloroform/methanol/0.1 m HCl peptides geqslant R: gt-or-equal, slanted 3 kDa were pure enough for analysis by SDS/PAGE, MALDI-TOF mass spectrometry and Edman degradation. It was therefore disclosed that organic solvent extracts of human bile contain several unexpected polypeptides (Table 1).

ATP synthase lipid-binding protein subunit 9 contains 75 amino acids, is an extremely hydrophobic inner mitochondrial membrane component [21] and accumulates in lysosomes in ceroid lipofuscinosis [22,23]. In agreement with our results, this protein has an apparent Mr of 3500 when analysed by SDS/PAGE. Monocyte differentiation antigen CD14 is a 356 amino acid leucine-rich glycoprotein expressed primarily by monocytes and macrophages, attached to the cell membrane by a glycosylphosphatidylinositol anchor [24]. A soluble form of the protein has been identified in human plasma with a mass of 48 kDa [25]. CD14 has been identified as an lipopolysaccharide (LPS) receptor [26] with the LPS receptor activity located in the N-terminal half of the molecule. The fragment covering positions 75–103 now found in bile is cleaved after a dibasic site and is located in the N-terminal part of the molecule. Mac25/insulin-like growth factor-binding protein-7 is a 277-residue polypeptide first described in meningiomas [27] and later found in normal human urine, cerebrospinal fluid, amniotic fluid and also at low levels in normal sera [28].

Hydrophobicity profiles for the proteins and peptide fragments found (Fig. 5), show that the ATP synthase lipid-binding protein, the CD14 fragment and the mac25 fragment exhibit extensive hydrophobic parts, while the α1-antitrypsin fragment contains two shorter hydrophobic internal segments. The fragments of CD14 and mac25 found here are most likely products of general protein degradation, while the α1-antitrypsin fragment is generated as a result of its inhibitory action. Bile may thus be a suitable way of disposing of pronouncedly hydrophobic peptides.

Figure 5.

Figure 5.

Hydropathy profiles of the 3–6 kDa polypeptides found in bile. The average hydropathy indices calculated for consecutive seven-residue fragments are plotted against the number of the central residue according to the method by Kyte and Doolittle [30]. (A) ATP synthase lipid-binding protein1–75, (B) CD1475–103, (C) mac251–22, (D) α1-antitrypsin359–394.

The amount of the α1-antitrypsin fragment (359–394) was consistent with an earlier report where polypeptides were isolated from the phospholipid fraction of human bile [9]. However, the amounts of specific polypeptides have been found to vary widely, e.g. for apolipoprotein A1 (2–320 µg·mL−1) [14,29] and for complement factor B (0.3–8 µg·mL−1) [8]. The amounts of the hydrophobic polypeptides found here, i.e. CD14 fragment (0.4 µg·mL−1), ATP synthase lipid-binding protein (0.2 µg·mL−1) and the mac25 fragment (0.2 µg·mL−1) are in the same range as those found for other proteins in gallbladder bile.

In cholesterol gallstone formation, it is likely that the balance between antinucleators and nucleators is important. The composition of hydrophobic polypeptides may vary depending on several factors and it is likely that a complex balance between hydrophobic and hydrophilic polypeptides, rather than individual polypeptides, plays a role in gallstone formation.

The presence of the haemoglobin α chain, but not the β chain, in the unpolar phase after organic solvent extraction is of interest. In the polar phase the α chain and β chain are found in approximately equal amounts, but the α chain, or a derivative thereof, apparently partitions exclusively into the unpolar phase. The hydropathy profiles of the haemoglobin α chain and β chain are very similar, which may suggest that a fraction of the haemoglobin α chain is modified in a manner to make it more hydrophobic. Notably, when 250 mL human plasma was subjected to the same extraction and purification procedure as described here, no haemoglobin chain was found in the unpolar phase.

In summary, by extraction with organic solvents and subsequent chromatography, we have purified and identified several polypeptides not described previously in human gallbladder bile. Because hydrophobic polypeptides are excluded using conventional separation protocols, procedures like those used here may be useful for the complete characterization of expressed proteins, in particular in lipid-rich tissues.


Human bile was kindly collected by Dr Sol-Britt Curstedt, Department of Anaesthesiology, Danderyd Hospital. This study was supported by the Swedish Medical Research Council.