Mass Spectrometric Identification of Serum Peptides Employing

used to search against the Swiss-Prot database, using the Mascot search engine for the identification ... Keywords: Identification • MALDI−MS/...
0 downloads 0 Views 189KB Size
Mass Spectrometric Identification of Serum Peptides Employing Derivatized Poly(glycidyl methacrylate/divinyl benzene) Particles and µ-HPLC Matthias Rainer,† Muhammad Najam-ul-Haq,† Rania Bakry,* Christian W. Huck, and Gu1 nther K. Bonn Institute of Analytical Chemistry and Radiochemistry, Leopold-Franzens University, Innrain 52a, 6020 Innsbruck, Austria Received August 21, 2006

Abstract: Biomarkers play a key role in preclinical screening and diagnosis of a disease. Various support materials are utilized for this task, in combination with MALDI-TOFMS. The way to effectively bind serum contents and their profiling is well-elaborated by the material-enhanced laser desorption ionization (MELDI) approach. In this particular work, focus is placed on the development of a strategy to identify low molecular weight serum peptides. Poly(GMA/DVB) is derivatized in a way to achieve an affinity termed as immobilized metal ion affinity chromatography (IMAC). Iminodiacetic acid (IDA) is used as a chelating ligand, whereas copper (Cu2+) acts as a metal ion for complexing peptides and proteins out of blood serum. Polymer binds the serum compounds over a broad mass range, which includes low mass peptides and high mass albumin (66 kDa). Bound contents are eluted from material by an acetonitrile/trifluoroacetic acid mixture, which proves the reversible nature of metal and amino acid linkage. Polystyrene/divinyl benzene (PS/DVB) monolithic capillary column is used for fractionation through RPHPLC, prior to the target spotting. The tandem TOF fragment ion mass spectra of each fraction is acquired and used to search against the Swiss-Prot database, using the Mascot search engine for the identification of peptides. Keywords: Identification • MALDI-MS/MS • MELDI-MS • Poly(GMA/DVB)-IDA-Cu2+ • µ-RP-HPLC • Serum

Introduction The field of proteomics provides opportunities to explain disease mechanisms and to discover diagnostic markers for early detection of diseases, but the complexity of proteomic samples makes their comprehensive analysis and characterization very difficult. Several sophisticated approaches, such as two-dimensional (2D) gel electrophoresis,1 have been devel* Address correspondence to Dr. Rania Bakry, Institute of Analytical Chemistry and Radiochemistry, Leopold-Franzens University of Innsbruck, Innrain 52a, 6020 Innsbruck, Austria. Tel.: +43/(0)512/507-5125. Fax: +43/ (0)512/507-2965. E-mail: [email protected]. † Authors contributed equally to this work.

382

Journal of Proteome Research 2007, 6, 382-386

Published on Web 11/18/2006

oped to resolve and visualize complex peptide and protein mixtures and their subsequent analysis by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDITOF-MS)2,3,4 or liquid chromatography-electrospray ionization (LC-ESI) MS.5,6 In comparison to ESI, MALDI-TOF-MS systems were primarily not suitable for generating MS/MS data, but since MALDI mass spectrometers were equipped with MS/ MS capabilities,7,8 it became possible to identify many more peptides and proteins. Moreover, the implementation of MS/ MS capable mass analyzers has boosted LC-MALDI in proteomic studies and has been widely used for the identification of peptides and proteins.9,10 As another approach to biomarker discovery, proteomic pattern analysis has been developed for the detection of novel diagnostic markers from body fluids when comparing massfingerprints from patients with those from control. The application of surface-enhanced laser desorption/ionization (SELDI)11,12 based on the ProteinChip technology (Ciphergen Biosystems, Inc., Fremton, CA) represents such an approach. Peptides and proteins are selectively immobilized to chemical or biological sites on the ProteinChip surface and subjected to on-chip analysis by MALDI-MS. Differences in protein profiles of disease and control related samples can be caused by overexpression, abnormally shed proteins or protein fragments, modified proteins, proteolytically cleaved proteins, or degradation due to the proteosome pathway.13 Peaks (m/z) of interest are statistically evaluated, but the identification of markers is still a problem solely using pattern analysis. Recently, a material-based approach referred as materialenhanced laser desorption/ionization (MELDI)14-17 has been introduced, utilizing various functionalized carrier materials for the selective binding of biomolecules. In comparison to the conventional SELDI method, MELDI does not exclusively focus on chemically or biologically treated docking-sites of the carrier, but also emphasizes its morphology, i.e., porosity, particle size, or effects of hydrophobicity and hydrophilicity, thus, posing a great impact on the resulting mass pattern as shown in case of silica supports.18 The ongoing development of the MELDI technique is offering high sensitivity, which is crucial for disease markers to be identified with MALDI-TOF-MS. In combination with liquid chromatography and MALDI-MS/MS, the described MELDI method can be successfully employed for the identification of peptides.19 Some peptides are already identi10.1021/pr060426y CCC: $37.00

 2007 American Chemical Society

technical notes

Figure 1. Strategy for identification of low molecular weight peptides present in serum samples by poly(GMA/DVB)-IDACu2+ particles.

fied as biomarkers for specific diseases like the amyloid beta peptide for Alzheimer’s disease20 and resistin for diabetes and obesity.21 In this work, we describe the use of a polymer material synthesized by glycidyl methacrylate (GMA) and divinyl benzene (DVB) derivatized by iminodiacetic acid (IDA) and loaded with copper (Cu2+), as an IMAC support material,22,23 for the selective enrichment of peptides and proteins from human serum samples for the application in peptide identifications (peptidomics). Peptidic species mostly lie out of the range of conventional 2D gel electrophoresis. Therefore, the MELDI technique is applied to reduce the complexity of the serum samples, while reversed-phase HPLC is used for fractionation, followed by target spotting and MALDI-MS/MS including database searching analysis.

Rainer et al.

Figure 2. Comparison of serum mass fingerprinting on poly(GMA/DVB)-IDA-Cu2+, from 2 to 10 kDa with HCCA (A) and sinapinic acid (B) as matrix, recorded by MALDI-TOF-MS by averaging 400 laser shots of 337 nm nitrogen laser in linear mode.

Experimental Procedures Chemicals and Reagents. Iminodiacetic acid (IDA, 98%), acetonitrile (ACN) (for HPLC, g99.9%), methanol (for HPLC, g99.9%), sinapinic acid (SA), divinyl benzene (technical grade, 80.0%), and glycidyl methacrylate (for GC, g97.0%) were purchased from Sigma Aldrich (St. Louis, MO). Trifluoroacetic acid (TFA, for protein sequence analysis), copper sulfate (anhydrous, g99.9%), R,R′-azoisobutyronitrile (AIBN), and R-cyano-4-hydroxycinnamic acid (HCCA, matrix substance for MALDI-MS, g99.0%) were obtained from Fluka (Buchs, Switzerland). Serum samples were provided by the Department of Urology at Medical University of Innsbruck, Austria. Preparation and Derivatization of Poly(GMA/DVB). Glycidyl methacrylate was copolymerized with divinyl benzene by thermal polymerization to yield GMA/DVB polymer beads. The obtained beads were derivatized with iminodiacetic acid (IDA) under alkaline conditions and additionally complexed with copper(II) ions.23 MELDI Sample Preparation. Thirty milligrams of poly(GMA/ DVB)-IDA-Cu2+ particles were filled into 1.5 mL centrifuge tubes and equilibrated twice with 400 µL of phosphate-buffered saline (PBS; pH ) 7.4, created by 0.01 M Na2HPO4/NaH2PO4 and 0.15 M NaCl). One hundred microliters of human serum,

Figure 3. Comparison of MELDI (A) versus MALDI after elution (B) mass spectra using poly(GMA/DVB)-IDA-Cu+2, from 2 to 10 kDa with HCCA as matrix, recorded by MALDI-TOF-MS by averaging 400 laser shots of 337 nm nitrogen laser in linear mode.

diluted in 300 µL of PBS was incubated with equilibrated polymer for 5 min at 30 °C. Nonbound peptides and proteins as well as salts and other contaminants were washed away three times with 400 µL of PBS buffer and once with an equal volume of deionized water. For MELDI-MS analysis, 1 µL of the washed polymer slurry was placed on a stainless steel target, and 1 µL of saturated R-cyano-4-hydroxycinnamic acid (HCCA) or sinapinic acid (SA) in 50% (v/v) ACN/0.1% TFA was added. Peptideand protein-loaded polymers were directly analyzed using Journal of Proteome Research • Vol. 6, No. 1, 2007 383

MS Identification of Serum Peptides Using Poly(GMA/DVB)/µHPLC

technical notes

Figure 4. Profiling by MELDI (A), MALDI spectra of LC-fraction after RP-HPLC using PS/DVB monolithic column (B), and tandem TOF/ TOF fragment ion mass spectra of m/z 2028.058 (C).

MALDI-TOF-MS (Ultraflex MALDI-TOF/TOF, Bruker Daltonics, Bremen, Germany) equipped with a 337 nm nitrogen laser. Measurements were recorded in positive linear mode for crude profiling and in reflector mode for MS/MS analysis. Preparative HPLC. Preparative HPLC was performed on an A¨ ktapurifier system (GE Healthcare Life Sciences, Germany), connected with fraction collector (frac 950) and multiple wavelengths UV detector. Poly(GMA/DVB)-IDA-Cu2+ material was packed in a polyetheretherketone (PEEK) column 3 × 1.6 cm using a slurry packing technique. The column was equilibrated with PBS buffer (3 column volumes) under a flow rate of 1 mL/min. Serum sample (0.5 mL) was diluted with an equal volume of PBS buffer and loaded into a poly(GMA/DVB)-IDACu2+-packed cartridge. The column was washed with PBS buffer to remove nonbound peptides and proteins followed by washing with water. The bound constituents were eluted with 0.1% TFA in 50% ACN (5 column volumes). Fractions (0.5 mL) 384

Journal of Proteome Research • Vol. 6, No. 1, 2007

were collected in the fraction collector throughout the whole elution program. Both sample loading and elution steps were monitored online with a UV detector at 214 and 280 nm. The fractions were concentrated to 50 µL volume using a speed vacuum, and the samples were analyzed with µLC-MALDI/ TOF-MS. Reversed-Phase (RP) µ-HPLC Separation. The eluted samples were separated using a monolithic reversed-phase capillary column, equipped with a 1100 series Nanoflow LC (Agilent Technologies, Waldbronn, Germany). The microfraction collector/spotter module was equipped with the following accessories to use its spotting capabilities: KDS 200 syringe pump, 1 mL SGE syringe, Valco micro T-piece, 125 mm capillary (PEEK-coated fused silica). For system control, Agilent ChemStation A 10.02 Software was used. The mobile phase for RP separations were solution A, 0.1% (v/v) TFA in water, and solution B, 0.1% TFA (v/v) in ACN. The samples (8 µL) were

technical notes

Rainer et al.

Table 1. List of Proteins Identified by LC-MALDI-MS/MS of Human Serum after Preconcentration with Poly(GMA/DVB)-IDA-Cu+2 Particles

protein name

Complement C3 precursor Complement C3 precursor Complement C3 precursor Complement C3 precursor Fibrinogen alpha chain precursor Fibrinogen alpha chain precursor Fibrinogen alpha chain precursor Clusterin precursor Apoliprotein A-IV precursor Apoliprotein A-IV precursor Inter-alpha-trypsin inhibitor heavy chain H4 precursor Inter-alpha-trypsin inhibitor heavy chain H4 precursor Ig gamma-1 chain C region

matched sequence

sequence position

peptide mass (Da)

score

P01024 P01024 P01024 P01024 P02671 P02671 P02671 P10909 P06727 P06727 Q14624

SSKITHRIHWESASLLR SSKITHRIHWESASLL SKITHRIHWESASLL THRIHWESASLL SSSYSKQFTSSTSYNRGDSTFES SSSYSKQFTSSTSYNRGDSTFESKSY HRHPDEAAFF DTASTGKTFPG RPHFFFPKSRIV GNTEGLQKSLAELGGHLDQQVEEF SLAELGGHLDQQVEEF QLGLPGPPDVPDHAAYHPF

1304-1320 1304-1319 1305-1319 1308-1319 575-597 575-601 501-521 215-226 280-303 288-303 669-687

2020.145 1863.97 1777389 1449.594 2553.057 2931.109 2288.571 1530.751 2599.2949 1771.779 2028.073

55 50 56 52 63 61 50 50 67 87 69

Q14624

RSQLGLPGPPDVPDHAAYHPF

667-687

2271.13

55

P01857

PEVKFNWYVD

154-163

1296.535

70

access key

first loaded onto a trapping column (PS/DVB monolith, 0.2 mm × 20 mm) at 8 µL/min with 2% solution B. After 5 min, the 10-port valve was switched, and the sample was eluted onto the analytical separation column (PS/DVB, 0.2 mm × 50 mm) (LC packings, Amsterdam, Netherlands) in back-flush mode using a micropump operated at 2 µL/min. The separation was achieved by gradient starting with 2% B at 0 min to 60% B in 45 min. The matrix HCCA (3 mg/mL in ethanol/acetone 2:1) was added to the LC effluent 1 µL/min using a syringe pump connected to a T-connector. The fractions were spotted every 20 s on the MALDI target. The target was then analyzed using a MALDI-TOF/TOF mass spectrometer (Bruker Daltonics, Germany). We further identified peptides by acquiring MS/MS spectra using MALDI-TOF/TOF-MS, followed by database searching with Mascot against the Swiss-Prot database. Peptide mass tolerance was kept at 150 ppm, and MS/MS tolerance was adjusted to 1.0 Da.

Results and Discussion The entire focus of this work is placed on the best developed strategy (Figure 1) to identify low mass peptides and proteins, which can be a potential way to approach biomarkers relating to diseases like cancers. Blood serum is chosen for our investigations, as it is understood that biomarkers can stably and rapidly travel from the site of disease into biofluids. Direct profiling of the serum contents bound to poly(GMA/DVB)IDA-Cu2+ is not sufficient for identification through MS/MS data acquisition. The complexity of peptides and proteins in the profiling pattern can be a simple reason for this failure. This complexity is, however, reduced by eluting the contents from the MELDI-material and passing over an RP-HPLC fractionating capillary column. As an alternative, the serum sample can be passed through a preparative separating column packed with poly(GMA/DVB)-IDA-Cu2+, followed by fractionation on RP-HPLC. The fractions thus obtained are handier to identify by MS/MS. The above-mentioned strategy for identification of peptides and proteins leading toward biomarkers identification is based on the reproducibility of material and methodology employed. Poly(GMA/DVB)-IDA-Cu2+ serves as an effective carrier to bind analytes from complex biological samples and analyze directly with MALDI-TOF-MS. Reproducibility was confirmed by recording three spectra from standard serum sample with

different batches. High qualitative reproducibility (number and location of detected mass signals) is proved by comparing the recorded spectra. Minor divergences in signal intensity are attributed to desorption and ionization process of the mass spectrometer. The profiling pattern shows an enormous capacity of this polymeric material in terms of binding peptides and proteins in the mass range from 2 to 10 kDa. The studies show that the poly(GMA/DVB)-IDA-Cu2+ material has also the potential to bind proteins in the higher mass range, such as human serum albumin at the mass of 66 kDa. This is of great importance in biomarker serum analysis as the crude serum sample contains hundreds of peptides and protein associated with human serum albumin.24 The choice of matrix is very crucial not only for the sensitivity of protein profiles, but also in terms of quality, capacity, and resolution of the spectra. Normally, sinapinic acid is considered to be the ultimate choice of matrix for protein mass range.25 The comparison was made with HCCA in this regard. The mass fingerprinting with the HCCA (Figure 2A) matrix is quite enriched in quality parameters when compared with sinapinic acid (Figure 2B). Significantly more signals are observed in the lower mass range (