| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Proteomics and Protein Markers |
1 Department of Clinical Medicine, Division of Clinical Chemistry, Biomedicum, University of Helsinki, Helsinki, Finland.
2 Department of Surgery, Clinic of Transplantation and Liver Surgery, and3
Second Department of Surgery, Helsinki University Central Hospital, Helsinki, Finland.
aAddress correspondence to this author at: Department of Clinical Medicine, Division of Clinical Chemistry, Biomedicum Helsinki, University of Helsinki, PO Box 63 (Haartmaninkatu 8), Helsinki FIN-00014, Finland. Fax 358-9-47171731; e-mail leena.valmu{at}helsinki.fi.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: We used a quadrupole/time-of-flight hybrid mass spectrometer equipped with an electrospray ionization source to analyze the molecular identity of PSTI purified from the urine of 12 patients with pancreatitis and from 3 controls. We also developed a rapid small-scale capture procedure to isolate and analyze PSTI from small volumes of urine.
Results: The mutations responsible for mass shifts of different PSTI variants could be verified. We observed differences in the expression of different variants as well as a novel proteolytic fragment of PSTI. Small-scale magnetic bead–mediated immunoaffinity chromatography PSTI enabled easy and rapid purification from small urine volumes, facilitating mass spectrometric analysis with adequate sensitivity.
Conclusions: Pancreatitis-related PSTI variants occurring at nanomolar concentrations in urine can be detected and quantified by immunoaffinity purification and mass spectrometry. In addition, the N34S variant occurs at higher concentrations than the wild type. This finding casts new light on the possible role of PSTI as a cause of hereditary pancreatitis.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
PSTI is encoded by a gene comprising 4 exons located on chromosome 5 (6). The gene encodes a single polypeptide chain consisting of 79 amino acids, including a 23-amino acid signal sequence (6). Mature PSTI is thus composed of 56 amino acids, contains 3 disulfide bridges (7), and has a calculated molecular weight (Mr) of 6241.0 based on the amino acid composition (2).
A fairly common mutation in the PSTI gene has been associated with chronic pancreatitis, which usually first occurs during childhood or adolescence (8). An association between exonic PSTI mutations, particularly the asparagine-to-serine mutation at amino acid 34 (N34S mutation), and idiopathic chronic pancreatitis has been demonstrated (9)(10). Because the PSTI N34S mutation is also quite common in the general population (11), it is unlikely that this mutation alone initiates the development of chronic pancreatitis, but it may rather act as a disease modifier (12). Recently, rare intronic mutations leading to complete functional loss of the protein and carrying a clear disease-causing effect have been found (13).
The mechanism by which mutations in the PSTI gene affect disease development remains to be elucidated. The N34S mutation is thought to lower the threshold of development of pancreatitis by reducing the capacity of PSTI to inhibit trypsin and prevent inadvertent trypsin activation (8). However, because decreased trypsin-inhibitory activity is not shown by recombinant PSTI with the N34S mutation (14), other changes in the trypsin-inhibitory capacity of the N34S mutant have been hypothesized to underlie the predisposition to pancreatitis in patients carrying this mutation. Such changes include the effect of 2 intronic mutations that cosegregate with the N34S mutant, which might cause abnormal splicing of PSTI as well as enhanced digestion of the N34S mutant by enzymes other than trypsin (15).
The association of 2 exonic PSTI mutants, N34S and a proline-to-serine mutation at amino acid 55 (P55S), with chronic (11) as well as with acute (16) pancreatitis has recently been analyzed in the Finnish population. The N34S mutation was found to be associated with these diseases, being present in 12% of patients with chronic and 7.8% of those with acute pancreatitis compared with 2.6% in controls (P <0.0001), but no significant difference in frequency of pancreatitis in patients and controls was found to be associated with the P55S mutation.
Traditionally, the analysis of genetic variability has been performed at the DNA level. The development and automation of proteomic technology have raised the possibility of analysis of genetic variants at the protein level (17)(18), facilitating studies of differences in the expression of mutated variants and identification of protein posttranslational modifications that are not evident at the DNA level.
Pancreatic acinar cells secrete PSTI into the pancreatic juice, where it constitutes 0.1%–0.8% of the total protein (2). The mean concentration of TATI in serum of healthy individuals is 1.8 nmol/L (reference interval, 0.5–3.5 nmol/L) (4). Because of its small size, it is rapidly cleared from the circulation by renal excretion with a half-life of 6 min (19). The concentration of PSTI in urine of healthy individuals is therefore somewhat higher than that in serum, the reference interval being 1–8.5 nmol/L. In pancreatitis, PSTI concentrations in urine may increase more than 1000-fold (20)(21)(22). The aim of the present study was to explore the capability of mass spectrometry (MS) in identifying different PSTI variants occurring at nanomolar concentrations in body fluids in a specific and sensitive way. Because of the easy access of urine obtainable by noninvasive procedures, we performed proteomic characterization of PSTI in urine of pancreatitis patients and healthy controls. We also developed a rapid method for capture of PSTI from small volumes of urine.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
patients and control individuals
We obtained urine from 12 patients with pancreatitis and from 3 controls without pancreatic disease (Table 1
). Of the 12 patients with pancreatitis, 3 (25%) had severe disease and 9 (75%) mild disease according to the Atlanta classification (24). None of the pancreatitis episodes was fatal. One of the patients (individual 9) suffered from pancreatic cancer in addition to pancreatitis. Patients were selected for the study on the basis of molecular genetic analysis of PSTI variants performed earlier (11)(16), in which the diagnosis of pancreatitis was made on the basis of a typical clinical picture, a serum amylase more than 3-fold higher than the upper reference limit, and/or typical findings on computed tomography. Two patients (individuals 1 and 2) and 2 controls (individuals 13 and 14) expressed wild-type (WT) PSTI protein, 1 patient (individual 3) was homozygous for the N34S mutation, 6 patients (individuals 4–9) and 1 control (individual 15) were heterozygous for the N34S mutation, and 3 patients (individuals 10–12) were heterozygous for the P55S mutation. The patient group included 4 women and 8 men (median age, 50 years; range, 14–69 years), and the control group included 2 women and 1 man (median age, 36 years; range, 19–38 years).
|
samples
Urine samples were either stored frozen at –80 °C or immediately subjected to PSTI purification. The amount of urine used for chromatographic purification was 10–1500 mL. The concentration of PSTI in urine and serum was determined by a previously described time-resolved immunofluorometric assay (23) using MAb 6E8 as the capture antibody and europium-labeled MAb 11B3 as the detection antibody. The concentrations of PSTI in the urine samples used for purification varied from 1.5 nmol/L to 2.6 µmol/L.
Serum samples from 2 pancreatitis patients were collected and stored frozen at –80 °C. The volume of serum used for purification was 15 mL, and the amount of PSTI in these samples was 0.3 nmol according to the time-resolved immunofluorometric assay.
chromatographic purification of psti
Urine containing 1.6–80 nmol of PSTI was clarified by filtration through a 0.45 µm pore size Millex low protein binding Durapore membrane (Millipore). The filtrate was applied to an anti-PSTI affinity column (4 mL) containing MAb 6E8 coupled to CNBr-activated Sepharose 4B (Pharmacia) according to the instructions of the manufacturer. The column was equilibrated with 50 mmol/L Tris-HCl (pH 7.2). After application of the urine samples, the column was first washed with equilibration buffer containing 1 mol/L NaCl, 10 mmol/L benzamidine, and 10 mL/L isopropanol and then with 10 mmol/L ammonium acetate (pH 4.5). Bound protein was eluted with 1 mL/L trifluoroacetic acid (TFA), and 1-mL fractions were collected and immediately neutralized with 1 mol/L Tris-HCl (pH 9.0). Fractions containing PSTI were pooled and subjected to reversed-phase (RP) chromatography on a C18 column [Symmetry C18; 20 x 3.9 mm (i.d.); pore size, 300 Å; bead size, 5 µm; Waters]. PSTI was eluted with a 30-min linear gradient of acetonitrile (0%–80%) in 1 mL/L TFA with a flow rate of 0.5 mL/min. Fractions containing PSTI were pooled for further analyses.
To purify PSTI from serum, we subjected the sample to gel filtration on a Sephacryl S-100 HR (Amersham Biosciences) column [700 x 10 mm (i.d.)] in 0.1 mol/L ammonium bicarbonate (pH 8) before immunoaffinity chromatography. Fractions containing proteins smaller than 20 kDa were collected and subjected to immunoaffinity purification as described above.
in-liquid trypsin digestion of psti
For in-liquid digestion, the PSTI protein was first reduced with dithiothreitol and alkylated with 4-vinylpyridine (Aldrich). The alkylated protein was desalted by RP-HPLC on a C18 column (Symmetry C18; column, bead, and pore sizes are the same as above) by elution with a 30-min linear acetonitrile gradient (0%–80%) in 1 mL/L TFA. PSTI-containing fractions were pooled and dried, 50 mg/g sequencing grade trypsin (Promega Ltd.) in 10 mmol/L ammonium bicarbonate was added, and digestion was carried out at 37 °C overnight.
small-scale immunoaffinity capture of psti from urine
For small-scale immunoaffinity capture of PSTI from urine, MAb 11B3 was covalently coupled to protein G–coated magnetic beads (25 µg of MAb/100 µL of beads; Dynal Biotech) according to the manufacturer’s instructions with dimethyl pimelimidate (Pierce) used as a cross-linker. Immunoaffinity capture with the MAb coupled to CNBr-activated Sepharose beads or magnetic beads gave essentially similar results, but because the magnetic beads procedure was more rapid, it was chosen for the study. An aliquot of urine containing 10–100 pmol of PSTI was incubated with 20 µL of MAb-coated beads for 2 h at room temperature with constant agitation. The reaction mixture was diluted with phosphate-buffered saline (10 mmol/L phosphate, 0.15 mol/L NaCl) containing 0.1 mL/L Tween 20 to prevent aggregation of the beads. The beads were washed 3 times with the same buffer, 3 times with phosphate-buffered saline without Tween, and once with water to remove detergent, which may interfere with the MS analysis. Protein bound to the magnetic beads was eluted with 10 µL of 1 mL/L TFA immediately before liquid chromatography (LC)-MS analysis. The magnetic beads were removed with a MagneSphere 12-position magnetic separation stand (Promega).
ms
Electrospray ionization (ESI)-MS analysis of chromatographically purified PSTI as such or as trypsin-digested peptides was performed with a Micromass Q-TOF Micro quadrupole/time-of-flight hybrid mass spectrometer (Waters). The peptides were injected into the mass spectrometer via a nanoflow interface with either a Hamilton syringe pump with a flow rate of 0.3 µL/min or after fractionation by nanoscale RP-HPLC on the CapLC (Waters) with a 150 x 0.075 mm (i.d.) C18 column (Symmetry C18; pore size, 300 Å; bead size, 3.5 µm; Waters) that was eluted with a 30-min linear gradient of acetonitrile (5%–50%) in 1 mL/L formic acid. The flow rate was 0.25 µL/min, and the eluate was injected directly into the mass spectrometer. For analysis of intact PSTI, the Q-TOF was calibrated by use of 400 fmol/µL myoglobin (Sigma). For analysis of tryptic peptides, the mass spectrometer was calibrated with 2 pmol/µL glufibrinogenic peptide B (Sigma) fragments. Tandem MS (MS/MS) fragmentation spectra of the peptides were acquired by colliding the doubly or triply charged precursor ions with argon collision gas at acceleration voltages of 30–45 V. Data analysis was carried out with MassLynx software (Waters) and PAWS proteomic analysis software (ProteoMetrics).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). PSTI appeared in the mass spectra mainly as [M + 4 H]4+, [M + 5 H]5+, and [M + 6 H]6+ (Fig. 1
, A, C, and E). The mass of PSTI was deconvoluted from these mass spectra with MassEnt1 software (Waters; Fig. 1
, B, D, and F). The deconvoluted mass of WT PSTI was 6241.5 Da (Fig. 1B
), and PSTI isolated from the urine of a heterozygous N34S patient showed an additional component with a mass of 6214.6 Da (Fig. 1D
). PSTI isolated from the urine of a patient heterozygous for the P55S mutation displayed masses of 6231.3 and 6241.7 Da (Fig. 1F
). Each of the masses corresponded to the theoretical masses calculated for WT PSTI (6241.0 Da), N34S mutated PSTI (6214.0 Da), and P55S mutated PSTI (6231.0 Da) with all putative disulfide bridges present. A smaller proteolytic fragment corresponding to the mass of PSTI lacking 3 N-terminal amino acids (Table 2A
) was also detected (Fig. 1
, B, D, and F).
|
|
WT and mutated PSTI variants (N34S and P55S) were further digested with trypsin to verify the origin of the mass difference in the intact protein masses. ESI-MS analysis of the tryptic digest of WT PSTI revealed all of the potential tryptic peptides of PSTI within 20 ppm mass accuracy (Table 2B
). The tryptic digest from the urine of an N34S heterozygous patient contained an additional tryptic peptide of m/z 664.30 (Fig. 2A
), which corresponded by mass to the [M + 2 H]2+ peptide comprising amino acids 9–18 with an asparagine-to-serine mutation. The WT and mutated peptides with m/z 677.80 and 664.30, which likely represent the 2 variants of peptide 9–18, were selected for fragmentation, which verified the identity of the peptides (Fig. 2C
). In the tryptic digest of PSTI isolated from a P55S heterozygous patient, a peptide of m/z 1009.77 was observed, corresponding by mass to [M + 3 H]3+ of amino acids 19–42 with the proline-to-serine mutation (Fig. 2D
). Fragmentation of peptides m/z 1013.12 and 1009.77 showed that these represent the 2 variants of amino acid sequence 19–42 carrying either proline (Fig. 2E
) or serine (Fig. 2F
) at position 55, respectively.
|
analyses of intact psti variants from urine
PSTI was purified from 15 urine samples by serial immunoaffinity chromatography and C18 RP chromatography (Table 1
). The amount of purified PSTI measured by time-resolved immunofluorometric assay (23) varied from 1.6 to 80 nmol. Samples were diluted to 2 pmol/µL and subjected to ESI-MS.
Intact PSTI revealed the same deconvoluted mass patterns of 6241.0, 6214.0, or 6231.0 Da as seen in Fig. 1
, but there were differences in the proportions of the protein variants (Table 1
). The WT samples and the single N34S homozygous sample contained only the expected PSTI variant, but the urine of heterozygous patients did not contain the 2 variants at the expected 1:1 proportions. The 6 patients and the control individual heterozygous for the N34S mutation had more of the variant than the WT protein. In 3 samples, the proportion of the variant exceeded 94%. Because the intensity values were derived only from the deconvoluted ESI-MS spectra, we further subjected the purified PSTI samples to LC-MS analyses. When integrated mass chromatograms [M + 5 H]5+ of intact PSTI were generated and the peak area was calculated, the proportion of PSTI variants measured differed by <5% from those calculated on the basis of the deconvoluted mass intensities. The proportion of WT PSTI was significantly lower in heterozygous individuals with the N34S mutation than in those with the P55S mutation (P = 0.017). The intensities of PSTI variants isolated from P55S heterozygotes showed only minor differences in the proportions of the variants with a slight excess of the WT variant.
To determine whether the increased proportion of the N34S variant was reproducible over time, we purified PSTI from multiple samples from a healthy individual heterozygous for the N34S mutation. The proportions of WT and the N34S variant remained constant.
We also purified PSTI from sera of 2 patients heterozygous for the N34S mutation (individuals 6 and 7). ESI-MS analysis showed that the N34S mutant also dominated in serum, representing 81% and 61% of intact PSTI.
All urine and serum samples analyzed contained a proteolytic fragment corresponding to PSTI lacking 3 N-terminal amino acids (Table 2A
). The proportion of this fragment varied from 1% to 50% (data not shown). No other fragments of PSTI were visualized.
rapid immunoaffinity capture of psti from small volumes of urine
We performed the initial ESI-MS analyses on PSTI purified from the large volumes of urine (10–1500 mL) needed for thorough molecular characterization of PSTI variants. To facilitate analyses of larger number of samples, we developed a rapid method to analyze PSTI from small volumes of urine. For small-scale immunoaffinity capture, magnetic beads with covalently bound MAb 11B3 were used to capture 10–100 pmol of PSTI from <1 mL of urine. PSTI was eluted with 1 mL/L TFA and injected directly into the LC-MS (Fig. 3B
). The amount analyzed equaled 10 pmol of PSTI if all PSTI from the urine was bound to the magnetic beads. According to MS measurements, the actual amount of PSTI bound was comparable to 10 pmol of purified PSTI (Fig. 3A
). As expected, PSTI was the main peptide/protein component detected by LC-MS, and detergents were the only minor interfering components remaining in the sample after magnetic bead capture. PSTI eluted from the C18 RP column as 2 peaks (Fig. 3
, A and B), one indicating the proteolytic fragment lacking 3 N-terminal amino acids (Fig. 3C
) and the other the intact protein (Fig. 3D
). The mutated PSTI variants were also captured by small-scale immunoaffinity and detected in the same proportions as by ESI-MS analyses of purified PSTI (data not shown). The retention times of the different PSTI variants varied slightly, with the P55S variant eluting first, the WT protein second, and N34S last. The 3 PSTI variants can therefore be separated by C18 RP chromatography, but the separation is not complete.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
PSTI is a good candidate for proteomic analysis because of its small size, lack of posttranslational modifications other than 3 disulfide bridges, and relatively high concentrations in urine of patients with pancreatitis. Mutations in the PSTI gene are associated with pancreatitis, and PSTI is also used as a tumor marker (TATI) (4). The N34S mutation is associated with a predisposition to chronic pancreatitis (8)(25)(26)(27), but controversy exists regarding its role in the development of this disease (12)(15). Analysis by MS facilitates quantification of expression differences of these variants, which could contribute to disease development.
Our results demonstrate the capability of an ESI-connected Q-TOF instrument to distinguish between WT and variant forms of PSTI. The ±1 Da mass accuracy for analysis of intact proteins enabled us to demonstrate the mass differences between different variants. We also performed de novo sequencing of the tryptic peptides from different variants to unambiguously assign the amino acid mutation, as shown for both the N34S and P55S variants.
A surprising finding was the enrichment of the N34S variant in urine of heterozygous carriers. Although the number of patients studied was small, the fact that all heterozygous individuals with the N34S variant had an excess of this variant suggests that the N34S variant is more stable than WT PSTI. This finding contradicts the hypothesis that susceptibility of the N34S variant to enzyme digestion could be a potential cause of pancreatitis in patients carrying this mutation (15). Although the ESI method does not provide a quantitative comparison of peptides in separate or even adjacent fractions, we measured the intensities in variants coinjected directly into the mass spectrometer, where the difference in ion suppression is minimal because the mutated amino acids in the PSTI variants are neutral. We used 2 MAbs with different epitopes and obtained very similar variant ratios, thus excluding the possibility that different variant proportions were attributable to the binding difference of the antibody used in the purification. We cannot rule out the possibility that some WT PSTI was present in complex with trypsin in urine and thus was not analyzable with our method. We therefore hypothesize that the N34S variant is more stable than WT PSTI. The proportion of the N34S variant was also increased in PSTI purified from sera of 2 N34S heterozygotes, a finding that favors this hypothesis.
Abnormal splicing of the PSTI RNA has also been suggested to produce protein splice variants that might cause a predisposition to pancreatitis (15). In this study we observed no new splice variants at the protein level. Most of the PSTI proteins detected were intact, and the only new variant detected was a proteolytic fragment lacking the first 3 N-terminal amino acids. Detection of such a fragment has not been reported, but a variant lacking 5 N-terminal amino acids has been observed in PSTI purified from pancreatic juice (28). The possibility that PSTI forms with differentially cleaved NH2 termini occur in different body fluids can not be ruled out, but because earlier methods purified PSTI from pancreatic juice by trypsin affinity chromatography, it is possible that the first 5 amino acids were cleaved by trypsin at the specific trypsin cleavage site after the fifth amino acid. Proteolytic cleavage of PSTI at its NH2 terminus comes as no surprise, because nuclear magnetic resonance analysis has shown that the NH2 terminus is relatively flexible (29), and the first 3 N-terminal amino acids have been found to be disordered in the crystal structure of PSTI (30). The third amino acid being leucine raises the possibility that, before renal excretion, PSTI is cleaved by either chymotrypsin or pepsin. This hypothesis is supported by the fact that the proteolytic fragment lacking 3 N-terminal amino acids was also found in the 2 serum samples analyzed. The ability of this proteolytic fragment to inhibit trypsin remains to be investigated, and the role of the fragmentation needs to be clarified.
This study demonstrates that the 3 known molecular variants of PSTI can be identified and characterized by use of immunoaffinity chromatography combined with MS. Extensive characterization of PSTI by protease digestion and de novo sequencing of the peptide fragments required large volumes of urine, but a small-scale immunoaffinity technique was developed to facilitate rapid analysis of PSTI variants. This method enables rapid PSTI detection in <1 mL of urine containing a PSTI concentration as low as 1.6 nmol/L, which is within the reference interval. Small-scale immunoaffinity capture with magnetic beads is a fast technique that lends itself to automation, and the immunoaffinity beads are reusable. To date, detection of genetic PSTI variants has been based on time-consuming DNA analyses. A typical PCR-based solid-phase minisequencing method (11)(16) takes a few days to detect PSTI variants, whereas the small-scale immunoaffinity capture described here can be performed within hours. In addition to the small-scale technique presented here, a previously described on-line immunoaffinity LC-MS method (31) could also be adapted to further facilitate analysis of PSTI variants.
Although the methods used in this study are not intended for quantitative analysis of PSTI, they demonstrate the possibility for further development of quantitative MS methods incorporating enrichment of urine with known amounts of either labeled or otherwise mass-mutated recombinant PSTI protein. These methods could be applicable to other protein variants and to analysis of various diseases.
In conclusion, MS can be used to characterize variants of small proteins occurring at nanomolar concentrations in urine and serum. Interestingly, the N34S variant of PSTI occurred at higher concentrations than the WT. The pathophysiologic implications remain to be clarified, but this finding demonstrates the power of MS to reveal expression of genetic variants at the protein level.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1-antitrypsin variants in plasma with the use of proteomic technology. Clin Chem 2001;47:2012-2022.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
