Derivatized Mesoporous Silica Beads for MALDI-TOF MS Profiling of

Apr 2, 2009 - Matrix-assisted laser desorption/ionization-time-of-flight mass spectrometry (MALDI-TOF MS) is a promising tool for large-scale screenin...
12 downloads 4 Views 2MB Size
Download PDF
Bioconjugate Chem. 2009, 20, 913–923

913

Derivatized Mesoporous Silica Beads for MALDI-TOF MS Profiling of Human Plasma and Urine Rosa Terracciano,*,† Luigi Pasqua,‡ Francesca Casadonte,† Stella Frasca`,† Mariaimmacolata Preiano`,† Daniela Falcone,† and Rocco Savino† Department of Experimental and Clinical Medicine, University of Catanzaro “Magna Graecia”, School of Medicine, 88100 Catanzaro, Italy, and Department of Chemical Engineering and Materials, University of Calabria, 87036 Cosenza, Italy. Received November 30, 2008; Revised Manuscript Received February 24, 2009

Matrix-assisted laser desorption/ionization-time-of-flight mass spectrometry (MALDI-TOF MS) is a promising tool for large-scale screening of body fluids for the early detection of human diseases. Proteins, peptides, and metabolites present in cells, tissues, or in body fluids constitute the molecular signatures of individuals. The design and generation of material-based platforms for capturing molecular signatures from body fluids has gained increasing interest in recent years. Highly selective materials are attractive candidates for a wide range of applications in biofluid proteomics. We have therefore developed a procedure based on mesoporous silica particles for the selective binding and enrichment of low molecular weight plasma/serum proteins by MALDI MS analysis (Terracciano, R., Gaspari, M., Testa, F., Pasqua, L., Cuda G., Tagliaferri, P., Cheng, M. C., Nijdam, A. J., Petricoin, E. F., Liotta, L. A., Ferrari, M., and Venuta, S. (2006) Selective binding and enrichment for low-molecular weight biomarker molecules in human plasma after exposure to nanoporous silica particles. Proteomics 6, 3243–3250). Mesoporous silica beads (MSB) are able to harvest peptides from plasma and serum by means of nanosized porous channels with high surface area, while excluding large size proteins. Moreover, the absorption properties can be modified since the pore walls can be functionalized with different chemical species due to the high concentration of silanol groups at the surface. In this study, we performed derivatization of MSB with different functionalities, and we evaluated the derivatized materials for plasma and urine peptidomic profiling. Aminopropyl, N-(2-aminoethyl)-3-aminopropyl, and N,N,N′ tris-carboxymethyl ethylene diamine, have been introduced onto the mesoporous silica surfaces in order to modulate selective peptide enrichment. We also explored various experimental conditions in order to optimize the performance of chemically modified MSB in the peptide profiling of human plasma and urine. These new derivatized mesoporous surfaces, in addition to the previous nonderivatized MSB, constitute an extended and reliable platform of five distinct chromatographic phases with defined surface functionality and porosity. Several plasma and urine peptides were extracted from derivatized MSB and then profiled by MALDI-TOF MS. The reproducibility of sample preparation by different functionalized beads was evaluated via three replicate analyses of plasma and urine samples. Lower coefficients of variation in the mass values and peak intensities resulted for plasma in comparison to those of urine samples; nevertheless, these where satisfactory for diagnostic purposes. For human urine, a linear correlation was found between spiked peptide concentrations and their peak areas (R2 > 0.98) with a limit of detection in the low-nanogram per milliliter range, thus confirming the high sensitivity of the methodology, previously demonstrated for human plasma. Different panels of peptide repertoires have thus been collected from highly porous substrates chemically conjugated with different functional groups, and these may be used in biomarker discovery for disease diagnosis.

INTRODUCTION The comparison of protein expression between healthy individuals and patients is one of the promising tools for biomarker identification. A fast and sensitive protein profiling method for biomarker discovery is an essential prerequisite for high-throughput and large-scale clinical studies. The establishment of proteomic approaches to biomarker discovery and disease profiling in recent years has tested the limits of existing tools for the capture and detection of low-abundance proteins in biological samples. The ability to isolate target proteins from complex protein mixtures, as well as improving the sensitivity toward peptides of low abundance, is a challenge for mass spectrometric analysis. * Corresponding author. Tel: ++39/09613694080. Fax: ++39/ 09613694090. E-mail: [email protected]. † University of Catanzaro. ‡ University of Calabria.

Proteomic pattern analysis by mass spectrometry (MS) is emerging as the preferred tool for robust and rapid expression profiling studies (1-9) and represents an alternative to the classical and laborious two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) (10). For biofluids, surface-enhanced laser desorption/ionizationtime-of-flight (SELDI-TOF) or matrix-assisted laser desorption/ ionization-time-of-flight (MALDI-TOF) MS are rapidly growing in popularity because of their ability to screen and discover multiple biomarkers simultaneously in clinical proteomics. In addition, SELDI-TOF MS and MALDI-TOF MS focus especially on the low molecular mass region analytes (1000-20000 Da), which include the small proteins, protein/peptide fragments, and peptides that constitute the peptidome or fragmentome, which are often overlooked by traditional techniques (11, 12). Affinity binding reagents have played a crucial role in the translation of proteomic discoveries to clinical diagnostics due to their ability to isolate target proteins from complex protein mixtures (13). The design and generation of material-based

10.1021/bc800510f CCC: $40.75  2009 American Chemical Society Published on Web 04/02/2009

914 Bioconjugate Chem., Vol. 20, No. 5, 2009

platforms for capturing molecular signatures written in body fluids have also gained increasing interest in recent years (14-17). Additionally, novel bioconjugate materials, designed specifically for microscale genomic and proteomic analyses, are facilitating new separation techniques (18, 19). Mesoporous silica has captured considerable interest for a diverse range of applications, such as catalysis, filtration and separation, molecular collection and storage, nanofluidics, medical imaging, drug delivery, and sensors (20). Recently, our group has explored a new application of mesoporous silica. We have previously developed a strategy based on mesoporous silica beads (MSB) for profiling low molecular weight plasma peptides (21). In later correlated studies, nanoporous silicon and glass beads were used to harvest distinct subsets of the proteome from serum/plasma samples (22, 23). According to IUPAC nomenclature, porous materials are divided into 3 classes: microporous material with a pore diameter below 2 nm, mesoporous material with a pore diameter between 2 and 50 nm, and macroporous material with a pore diameter above 50 nm. We speculated that mesopores would have the right dimensions for tunable peptide entrapment, while micropores might be too small and macropores too large for such a purpose. Given the high surface area, we demonstrated that mesoporous silica offered the desired adsorptive capacity for binding and enrichment of low molecular weight peptides present in human plasma (21). These results encouraged us to further investigate and explore the use of MSB as harvesting materials not only for human plasma but also for other clinically relevant body fluids such as human urine. The present study starts from the rationale that the absorption properties of mesoporous silica can be modified since the pore walls exhibit a high concentration of silanol groups on the surface, which can be functionalized with different chemical species. Aminopropyl, N-(2-aminoethyl)-3-aminopropyl, and N,N,N′-tris-carboxymethyl ethylene diamine (TED) have been introduced onto mesoporous silica surfaces in order to selectively modulate peptide enrichment. Plasma and urine peptides were extracted from derivatized mesoporous silica beads and then profiled by MALDI-TOF MS. The reproducibility of the method was also evaluated for the beads with different groups, both for plasma and urine samples. Depending on the chemical group and the chromatographic features, different subproteomes in plasma and urine were extracted and converted into reproducible MALDI profiles.

EXPERIMENTAL SECTION Reagents and Materials. Reagents for sol-gel synthesis, sodium silicate solution, fumed silica, tetraethylorthosilicate (TEOS), 3-aminopropyltriethoxysilane (APTES), and hexadecyltrimethylammonium bromide [C16H25N(CH3)3Br] (CTABr), were purchased from Sigma Aldrich, (St. Louis, MO, USA). Polyoxyethylene(10)isononylphenylether (Nonfix10), was obtained from Condea (Houston, TX, USA), n-(Trimethoxysilylpropyl)ethylenediaminetriacetic acid and trisodium salt (EDATAS, 50 wt.% in water) were purchased from Gelest Inc. (Tullytown, PA, USA). Pluronic P123 (poly(ethylene oxide)poly(propylene oxide)-poly(ethylene oxide) triblock copolymer, PEO20PPO70PEO20) was obtained from BASF (Mount Olive, NJ, USA). Acetonitrile (ACN, HPLC grade) and trifluoroacetic acid (TFA, ACS grade) were purchased from Merck (Darmstadt, Germany). The MALDI matrix R-cyano-4-hydroxycinnamic acid (CHCA) was purchased from Fluka (St. Louis, MO, USA). Peptide and protein standards, for mass spectrometer external calibration, were prepared from the Sequazime peptide mass standard kit (Applied Biosystems, Framingham, MA, USA). Angiotensin I

Terracciano et al.

and adrenocorticotropic hormone, ACTH (clip 1-17), for the spiking experiment were purchased from Sigma (St. Louis, MO, USA). The water used was obtained from a Milli-Q water purification system (Millipore, Bedford, MA, USA). Preparation of Materials. MSB-A and MSB-B Preparation. Mesoporous silica beads A and B were obtained according to two different synthetic procedures described elsewhere (24, 25). Preparation of Material SBA-15 and Large Pore Mesoporous Silica Materials. SBA-15 Mesoporous Material. SBA15 mesoporous material has been synthesized using a triblockcopolymer neutral surfactant PEO20-PPO70-PEO20 (Pluronic P123) in accordance with the procedure described by Zhao et al. (26). In a typical preparation, 4.0 g of Pluronic P123 was dissolved in 30 g of water and 120 g of 2 M HCl solution with stirring at 35 °C. Then 8.50 g of TEOS was added to the solution under vigorous stirring at 35 °C and maintained like this for 20 h. The mixture was aged at 80 °C overnight without stirring. The solid product was recovered, washed, and air-dried at room temperature. Large Pore Mesoporous Silica Material. Large pore mesoporous silica spheres were synthesized by using TEOS as the silica source, P123 as the template, CTABr as the cosurfactant, and ethanol as the cosolvent. Typically, 0.3 g of triblock copolymer P123 and 0.05 g of CTABr were dissolved in a solution formed by mixing 6 mL of 2 M HCl, 3 mL of H2O, and 2.5 mL of ethanol. Then, 1 mL of TEOS was added to the aqueous solution at room temperature under magnetic stirring. The reactant molar ratio TEOS/P123/CTAB/HCl/EtOH/H2O in the final reaction mixture was 1:0.011:0.031:2.67:9.18:116. After 30 min of stirring, the solution was transferred to a Teflonlined steel Parr autoclave and heated at 80 °C for 5 h and then kept at 120 °C for 12 h. The white precipitate was recovered by filtration, dried at 90 °C for 24 h, and calcined in air at 550 °C for 5 h to remove the templates. Preparation of Functionalized MSB-C, MSB-D, and MSB-E. In the preparation of 3-aminopropyl mesoporous silica (MSB-C) 0.6 g of calcined mesoporous material SBA-15 were suspended and stirred for 40 h in a solution obtained by dissolving 2.5 mL of 3-aminopropyltriethoxysilane in 10 mL of ethanol at room temperature. The suspension had been filtered, washed, and dried. In the preparation of N-(2-aminoethyl)-3-aminopropyl mesoporous silica (MSB-D), 0.6 g of calcined mesoporous material SBA-15 was suspended and stirred for 8 h in a solution obtained by dissolving 2.3 mL of N-(2-aminoethyl)-3-aminopropyltrimethoxysilane in 10 mL of ethanol under reflux. The suspension was filtered, washed, and dried. In the preparation of N-(3-propyl) ethylenediaminetriacetic acid, trisodium salt mesoporous silica (MSB-E) and 0.4 g of calcined large pore mesoporous silica were suspended in a solution obtained by dissolving 1.0 mL of N-(trimethoxy silylpropyl)ethylenediaminetriacetic acid and trisodium salts 50% in water. The suspension was stirred under reflux for 6 h and then filtered, washed, and dried. Material Characterization. All of the samples were characterized by X-ray diffraction on powder (XRD), thermogravimetrical analysis (TG), energy dispersive spectroscopy (EDS), X-ray microanalysis, and nitrogen adsorption-desorption isotherms. The XRD spectra were acquired on a Philips PW1710 diffractometer using Cu-KR radiation (40 kV, 20 mA) over the range 1° < 2θ < 15° with a rate scan of 0.005 2θ/s. TG analyses were carried out with a Netzsch STA 409 apparatus between 20 and 850 °C at a ramp of 10 °C/min in air with a flow rate of 5 cc/min. EDS chemical analyses were performed using a scanning electron microscope FEI Quanta 200.

Functionalized Mesoporous Silica Beads

Surface area of the samples was obtained by Brunauer-Emmet and Teller (BET) linearization in the pressure range 0.05-0.2 P/P0. The nitrogen adsorption-desorption volumetric isotherms at 77 K were measured on a Micrometritics Asap 2010 apparatus. Average pore diameter was calculated according to the BJH (Barret-Joyner-Halenda) model applied to the desorption branch of the isotherm. Plasma Collection. Venous blood samples were obtained from a total of 10 healthy adult males (age range 25-45 years) after they provided written informed consent. Blood intended for plasma preparation was collected into the following tubes: BD Plus Plastic K2EDTA, 10 mL, # 367525 (BD Bioscience, Franklin Lake, NJ, USA). The samples were centrifuged at 1300g for 10 min under refrigerated conditions (2-6 °C). The resultant plasma was immediately transferred into 2 mL Eppendorf vials and frozen in aliquots at -80 °C. Urine Collection. Midstream, first-void morning urine samples were collected from healthy consenting male volunteers after they provided written informed consent. For each participant, the urine sample (150 mL) collected in standard urine container flasks without additives or preservatives, was centrifuged at 1000g for 10 min at 4 °C. The supernatant was aliquoted and immediately stored at -80 °C. Preparation of Samples. MSB were incubated with the plasma or urine sample for 1 h at room temperature. The beads were subsequently separated from the supernatant by centrifugation, and the washing step was performed according to the protocol optimized for each bead functionality (see the next subsections). The bound peptides were eluted using the appropriate CHCA solution, and 1 µL of this eluate was then deposited on the MALDI probe and analyzed by MALDI-TOF. MSB Protocols for Human Plasma Peptidome Harvesting. For MSB-A and MSB-B, which have no functionalities, we used the following protocol. Briefly, aliquots (5 mg) of MSB were mixed with 500 µL of diluted human plasma sample (100 µL of plasma in 400 µL of deionized water) and shaken at room temperature for 1 h. The suspension was centrifuged at 2000g for 2 min, and then the MSB were separated from the supernatant and washed with deionized water (4 × 100 µL). After the last wash, bound species were directly extracted into 100 µL of a matrix solution, made up of 4 mg/mL CHCA in a 1:1 (v/v) mixture of acetonitrile and 0.1% TFA. For MSB-C and MSB-D, with, respectively, the aminopropyl and N-(2-aminoethyl)-3-aminopropyl groups, two sets of experimental conditions for peptide harvesting were tested. In a first set of experimental conditions, aliquots (5 mg) of beads were preconditioned with 250 µL of 25 mM Tris HCl at pH 8.8 for 5 min. MSB were then mixed with 500 µL of diluted human plasma sample (100 µL of plasma in 400 µL of deionized water) and shaken at room temperature for 1 h. The suspension was centrifuged at 2000g for 2 min, and then the MSB were separated from the supernatant and washed with 25 mM Tris HCl at pH 8.8 (2 × 100 µL). In a second set of experimental conditions, aliquots (5 mg) of beads were directly mixed with 500 µL of diluted human plasma sample (100 µL of plasma in 400 µL of deionized water) and shaken at room temperature for 1 h. The suspension was centrifuged at 2000g for 2 min, and then the MSB were separated from the supernatant and washed with 25 mM Tris HCl at pH 8.8 (2 × 100 µL). In both cases, after the last wash, the bound species were directly extracted into 100 µL of a matrix solution, made up of 4 mg/ mL CHCA in a 1:1(v/v) mixture of acetonitrile and ammonium formate (100 mM). To determine the pH value for optimal biomarker elution, the pH of the elution solution was adjusted to the values of 3, 4, and 5. For MSB-E, a surface-derivatized with the TED group, we adopted the following protocol. Aliquots (5 mg) of beads were

Bioconjugate Chem., Vol. 20, No. 5, 2009 915

preconditioned with 60 µL of HCl 10 mM, followed by 60 µL of ammonium acetate buffer at pH 6.6; this process was repeated twice. The MSB were then mixed with 500 µL of diluted human plasma sample (100 µL of plasma in 400 µL of deionized water) and shaken at room temperature for 1 h. The suspension was centrifuged at 2000g for 2 min, and then the MSB were separated from the supernatant and washed twice with ammonium acetate buffer at pH 6.6, followed by a third washing step with 100 µL of deionized water. After the last wash, the bound species were directly extracted into 100 µL of a matrix solution, made up of 4 mg/mL CHCA in a 1:1 (v/v) mixture of acetonitrile and 0.5% TFA. Extraction conditions were also tested in a solution of CHCA in ACN and formic acid (70:30 v/v). In all preparations, after centrifugation at 2000g for 2 min, 1 µL of supernatant CHCA solution with the extracted peptides was analyzed by MALDI-TOF MS. The samples were manually spotted by a pipettor onto the MALDI probe. MSB Protocols for Human Urine Peptidome Harvesting. For nonderivatized MSB-A and MSB-B, the following protocol was adopted. Aliquots (2 mg) of MSB were mixed with 125 µL of 4-fold concentrated human urine sample and shaken at room temperature for 1 h. The suspension was centrifuged at 2000g for 2 min, and then the MSB were separated from the supernatant and washed with 0.1% TFA (4 × 100 µL). After the last wash, the bound species were directly extracted into 15 µL of a matrix solution, made up of 4 mg/mL CHCA in a 1:1 (v/v) mixture of acetonitrile and 0.1% TFA. For MSB-C and MSB-D, derivatized, respectively, with the aminopropyl and N-(2-aminoethyl)-3-aminopropyl groups, two sets of experimental conditions for peptide harvesting were tested. In a first set of experimental conditions, aliquots (2 mg) of beads were preconditioned with 50 µL of 25 mM Tris HCl at pH 8.8 for 5 min. The MSB were then mixed with 125 µL of 4-fold concentrated human urine sample and shaken at room temperature for 1 h. The suspension was centrifuged at 2000g for 2 min, and then MSB were separated from the supernatant and washed with 25 mM Tris HCl at pH 8.8 (2 × 100 µL) and subsequently with deionized water (2 × 100 µL). In a second set of experimental conditions, aliquots (2 mg) of beads were directly mixed with 125 µL of 4-fold concentrated human urine sample and shaken at room temperature for 1 h. The suspension was centrifuged at 2000g for 2 min, and then the MSB were separated from the supernatant and washed with 25 mM Tris HCl at pH 8.8 (2 × 100 µL). In both cases, after the last wash, bound species were directly extracted into 40 µL of a matrix solution, made up of 4 mg/mL CHCA in a 1:1 (v/v) mixture of acetonitrile and ammonium formate (100 mM, pH 3). For MSB-E, surface-derivatized TED, we adopted the following protocol. Aliquots (2 mg) of beads were preconditioned with 60 µL of 10 mM HCl, followed by 60 µL of ammonium acetate buffer at pH 6.6; this process was repeated twice. The MSB were then mixed with 120 µL of a 4-fold concentrated human urine sample, and the resulting slurry was shaken at room temperature for 1 h. To determine the pH value for optimal biomarker harvesting, the pH of the resulting slurry was also adjusted to pH 4.5. The suspension was centrifuged at 2000g for 2 min, and then the MSB were separated from the supernatant and washed twice with 100 µL of ammonium acetate buffer at pH 6.6, followed by a third washing step with 100 µL of deionized water. After the last wash, the bound species were directly extracted into 60 µL of a matrix solution, made up of 4 mg/mL CHCA in a 1:1(v/v) mixture of acetonitrile and 0.5% TFA. In all preparations, after centrifugation at 2000g for 2 min, 1 µL of supernatant CHCA solution with the extracted peptides

916 Bioconjugate Chem., Vol. 20, No. 5, 2009

Terracciano et al.

Scheme 1. Functionalization of Mesoporous Silica

was analyzed by MALDI-TOF MS. The samples were manually spotted by a pipettor onto the MALDI probe. Urine Sample Processing with Standard Peptides. Urine samples were spiked with angiotensin I and adrenocorticotropic hormone, ACTH (clip 1-17), at eight different concentrations: 1000, 500, 250, 125, 62.5, 26, 6.5, and 3.7 ng/mL for angiotensin I, and 1614, 807, 404, 202, 101, 42, 10.5, and 6 ng/mL for ACTH (clip 1-17). The samples were processed with MSB-C as described in the previous section in three independent assays. MALDI MS Analysis. MALDI MS analysis was performed on a MALDI-TOF mass spectrometer (Voyager DE-STR, Applied Biosystems, Foster City, CA) equipped with a 337 nm nitrogen laser. Analyses were performed in linear positive-ion mode using delayed extraction. The acceleration voltage was 20 kV, the guide wire was 0.05% of the accelerating voltage, the grid voltage was 91.5%, and the delay time was 220 ns. Five 100-laser shots were averaged for each mass spectrum. All spectra were processed using Data Explorer Software (Applied Biosystems, Framingham, MA). All experiments were carried out at least in triplicate to ensure the reproducibility of the spectra. All signals with a signal-to-noise (S/N) ratio > 10 in a mass range of 1100-10000 for plasma and 1100-6000 for urine were collected in the Data Explorer Software. Peptide sequencing was carried out on a 4800 MALDI-TOF/TOF mass spectrometer (Applied Biosystems, Foster City, CA) equipped with a neodymium:yttrium-aluminum-garnet laser emitting at λ ) 355 nm with a 200-Hz firing rate.

RESULTS AND DISCUSSION The possibility of synthesizing mesoporous materials with regularly sized pores by means of the supramolecolar templating of a sol-gel process represents a starting point for the design of functional nanostructured materials. The MSB used in our approach belong to the class of ordered mesoporous silica synthesized for the first time in 1992 by the hydrolysis and condensation of inorganic precursors (the sol-gel process) in the presence of surfactant micelles (templates) (27). The term ordered in this context means that the pores are ordered. It is widely accepted that when the mesopore diameter is sufficiently large for the comfortable entrapment of biomolecules, analytes penetrate deep into mesoporous networks rather

than become adsorbed onto the external surface (28). The analytes entrapped by the mesoporous channels and consequently the resulting peptidomic profile of a given body fluid or complex protein mixture depend greatly on the size of the mesopores. Another way to modulate the selectivity of substrates may be by introducing functional groups. Indeed, it has been demonstrated that 17 and 70 nm pore sized aminopropyl-coated glass beads were able to adsorb serum peptides, but in a number ranging from one to a few dozens (23). On the contrary, hundreds of peptides were retained when mesoporous silica particles (MSB-A and MSB-B) containing pores with a size in the range of 2-3 nm were used (21). Therefore, we prepared a series of hybrid mesoporous silica in order to differentially modulate the selective binding and enrichment of low molecular weight proteins (LMWP) from plasma and urine, without significantly changing the physical properties (pore size, pore volume, and surface area) of the previously used MSB-A and MSB-B (21). Briefly, hybrid inorganic-organic materials are produced when chemically active groups are covalently linked to the inorganic framework of mesoporous materials either by postsynthesis grafting or by the simultaneous condensation of siloxane and organic siloxane precursors, in which the latter contain a nonhydrolyzable Si-C bond (29). In our case, three hybrid materials with different functional groups, namely, aminopropyl-, N-(2-aminoethyl)-3-aminopropyl-, and N-(3-propyl) ethylenediaminetriacetic acid, were prepared by postsynthesis modification of two different mesoporous silica obtained using Pluronic P-123, a triblockcopolymer neutral surfactant. The derivatization procedure is shown in Scheme 1. The derivatized particles MSB-C, MSBD, and MSB-E were characterized according to analyses described in the Experimental Section to obtain information on the size, pore size, shape, surface area, and pore volume (more details on the materials are available in Supporting Information). The beads show an irregular shape, with the average pore size being approximately 4 nm for MSB-C and MSB-D, and 7 nm for MSB-E, and surface area in the order of 200 to 300 m2/g. A more detailed description of textural properties of these new hybrid materials along with the previously used silica A and B is reported in Table 1. As briefly reported in the Experimental

Functionalized Mesoporous Silica Beads

Bioconjugate Chem., Vol. 20, No. 5, 2009 917

Table 1. Textural Properties of Mesoporous Silica Bead Platforms

a

Pore size is expressed in nm. b Pore volume is expressed in cm3/g. c Surface area is expressed in m2/g.

Section under the paragraph on material preparation, MSB can be readily synthesized by a simple, inexpensive procedure. Additionally, given the good mechanical properties, (rigidity and stability), MSB can act as an excellent substrate for the pretreatment of body fluids. Plasma and urine represent convenient diagnostic specimens for clinical analysis. Metabolites and peptides from these body fluids have the potential to serve as reliable indicators of the progression from a normal to a diseased state. Human urine contains the proteins filtered from blood within the glomerulus as well as the proteins secreted from the kidneys and the urogenital tract. Changes in kidney and urogenital tract function may be detected in the urine proteome. In addition, as a filtrate of blood, urine contains protein components that are similar to those found in the blood. Therefore, pathological changes in human organs, which could be found in blood, may also be reflected and detected in urine. In this sense, urine proteome analysis could be used in disease diagnostics (30). In spite of multidimensional liquid chromatography-MS, which is usually unsuitable for clinical high-throughput screening, MALDI-TOF MS represents a key tool for rapidly analyzing clinical samples. Unfortunately, salts, lipids, and high concentrations of proteins in plasma as well as high concentrations of salts and metabolic wastes in urine make it difficult to detect and analyze peptides by MS. MSB represent a rapid procedure for reducing plasma and urine complexity prior to mass spectrometric analysis. We have tested the new derivatized mesoporous silica beads for the generation of molecular signatures from plasma and urine samples. In MSB-A and MSB-B, previously tested on plasma only, peptide harvesting is largely regulated by the mesoporous properties and the electrostatic interactions of the silica (MSBA) and the silanol groups (MSB-B) with the peptides (21). In the new hybrids, MSB-C, MSB-D, and MSB-E, weak basic and acid functionalities present on the surface differently modulate the electrostatic interactions toward a particular subset of negatively or positively charged peptides. In particular, in MSBC, the aminopropyl group has one ion-exchange site with a pKa of 9.8, while MSB-D, which contains both a primary and secondary amine (PSA), exhibits two ion-exchange sites (pKa 10.1 and 10.9). MSB-E, derivatized with the chelating N,N,N′triscarboxymethyl ethylene diamine group (TED), offers the possibility of forming pentadentate complexes when used in conjunction with specific metal ions used for the separation of proteins (31). Moreover, TED contains weak acid groups whose conjugate bases can act as weak cation exchangers; therefore, MSB-E also possesses cation exchange properties in addition to its chelating character. In this exploratory study, we tested this substrate as a weak cation exchanger. More detailed and

additional investigations are planned by our group to study the feasibility of using this substrate as an immobilized metal affinity chromatography (IMAC) material (32). In Scheme 2, we illustrate the workflow of the derivatized MSB for MALDI-TOF analysis of plasma and urine peptidome. It is divided into four main steps: adsorption, separation, washing, and extraction. In the first step, a small amount of the particles are simply suspended in plasma or urine and left to act as sponges. During this step, only peptides and proteins with sizes smaller than or comparable to the pore size are adsorbed into the mesochannels. Moreover, given the specific chemical functionality present on the mesoporous surface, the adsorption process will also be driven by the electrostatic interactions of the positively (arginine, lysine, histidine) or negatively (glutamic and aspartic acids) charged residues of the peptides and proteins present in the plasma or urine, with the chemical groups conjugated to the mesoporous silica surface. Typically, the adsorption of body fluid peptides into MSB can often change with pH. Thus, in a neutral solution, MSB-C and MSB-D (pKa around 10) retain a positively charged surface, which can adsorb negatively charged peptides. On the contrary, MSB-E possesses a negatively charged surface, which can adsorb positively charged peptides. From this point of view, our strategy in choosing the appropriate experimental conditions was driven by the chemical group conjugated to silica pore walls. In the Experimental Section, we report the detailed conditions used for each derivatized surface. Different pH values were tested in order to ensure that the maximum number of peptides were captured from plasma and urine. We found that for MSB-C, a primary amine phase, and MSB-D, a primary and secondary amine phase (PSA), better working conditions (in terms of plasma and urine peptide adsorption) were obtained in a pH range between 8 and 9, rather than under neutral conditions (data not shown). For MSB-E, containing the TED group, we also observed a strong pH dependence on the adsorption efficiency. In Figure 1, we report the comparison between the MALDI-TOF mass spectra of urinary peptides captured by MSB-E at two different pH values. When the (concentrated) urine sample was incubated with MSB-E at pH 6.6, we observed an increase in the number of peptides detected by MALDI-TOF mass spectrometry (Figure 1a) in comparison to that at pH 4.5 (Figure 1b). In our experience with MSB, we found that the separation and washing steps were critical in order to ensure MALDI spectra of good quality. Separation of the beads from the supernatant requires gentle centrifugation and preferably a single pipettor aspiration, in order to avoid diffusion of adsorbed peptides into the mother liquor. At the same time, washing needs to be as fast as possible. The washing step is crucial in order to

918 Bioconjugate Chem., Vol. 20, No. 5, 2009

Terracciano et al.

Scheme 2. Workflow Diagram of Plasma and Urine Sample Preparation with Mesoporous Silica Beads (MSB)

eliminate salts and other contaminants, which can interfere with MS analysis. Too many washes should be avoided due to the risk of sample loss. To overcome these problems, different washing conditions were tested, and the best compromise was found when the number of washes did not exceed three (data not shown). Our derivatized beads have ionizable groups; therefore, the elution parameters are affected by a number of factors, and the ionic strength, temperature, and pH of the eluent may play a crucial role. In order to prevent the degradation of the human plasma and urine peptides harvested on the beads and to make

Figure 1. MALDI-TOF spectra of peptides extracted from a urine sample by weak cation mesoporous silica beads using two different pH values during the adsorption step. The resulting spectrum at pH 6.6 (a) reveals an increased number of peptide peaks compared to that in the spectrum at pH 4.5 (b).

Figure 2. Effect of pH during the extraction step, on the mass spectral profiles of plasma processed with mesoporous silica beads derivatized with an aminopropyl group (MSB-C). The figure shows the change in the plasma profile when plasmatic peptides are extracted from the beads at pH 5 (a), pH 4 (b), and pH 3 (c). The optimal MS profile is obtained using a CHCA solution in ACN and 100 mM ammonium formate at pH 3 (c).

Functionalized Mesoporous Silica Beads

Figure 3. MALDI-TOF spectra of peptides extracted from human plasma using different mesoporous silica beads (MSB). Human plasma samples were incubated with mesoporous beads followed by washing and elution with a matrix CHCA solution. Eluates from MSB-A (a), MSB-B (b), MSB-C (c), MSB-D (d), and MSB-E (e) were analyzed by MALDI-TOF MS. The top panel (control) shows the MALDI spectrum of an unprocessed plasma sample. Asterisks mark peaks, sequenced by MS/MS experiments, which are retained by a single MSB (indicated in red), by two different MSB (indicated in black), or by all three MSB analyzed (indicated in blue).

the analysis more rapid, one elution step at room temperature was preferred to fractionation, which requires more time. Moreover, the extraction was performed directly with the matrix solution for MALDI-TOF analysis. CHCA was used as the matrix because it is more suitable for peptide analysis compared to other matrices used for MALDI-MS (sinapinic acid or DHB). Different elution parameters, in particular the eluent composition and pH, were assessed in order to maximize the elution of all of the adsorbed peptides in one step. We observed different MALDI MS spectra when plasma peptides were extracted from MSB-C at three different pH values (Figure 2). Comparing elution at pH 5 (Figure 2, panel a) and elution at pH 4 (Figure 2, panel b), the most significant differences were observed with increased mass signals in the m/z range between 2600 and 5000 at pH 4. The most efficient peptide extraction was accomplished at pH 3. We observed no substantial differences in peptide patterns when comparing the MALDI spectra at pH 4 and 3 (Figure 2, panels b and c) apart from a substantial variation in the relative signal intensities. Therefore, for anionic mesoporous beads, optimal elution conditions were obtained with a CHCA solution in ACN and 100 mM ammonium formate at pH 3. Plasmatic and urinary peptides bound to siloxane (MSB-A) and silanol (MSB-B) beads were eluted satisfactorily with a CHCA solution in ACN and 0.1% TFA 1:1 (v/v). For cationic beads, a solution of CHCA in ACN and 0.5% TFA (50:50, v/v) was more efficient compared to a solution of ACN and formic acid (70:30, v/v, data not shown).

Bioconjugate Chem., Vol. 20, No. 5, 2009 919

Figure 4. Processed MALDI spectra of peptides extracted from human urine samples using five different beads. The mass range from 1100 to 6000 Da is shown. Human urine was analyzed using MSB-A (a), MSB-B (b), MSB-C (c), MSB-D (d), and MSB-E (e). The top panel (control) shows the MALDI spectrum of an unprocessed urine sample. Table 2. Mean Number of Peaks Detected by Each Bead Type from Human Plasma human plasma bead type

total peaks detecteda

peaks detected range 1100-5000 Da

peaks detected range 5000-10000 Da

MSB-A MSB-B MSB-C MSB-D MSB-E

122 138 160 114 119

117 130 147 110 119

5 8 13 4 0

a The number of peaks detected in the range between 1100 and 10000 Da with a S/N > 10.

The working conditions used for capturing, washing, and eluting the plasma and urine peptides, in order to ensure the best MALDI-TOF profiling in terms of the number of peaks detected, reproducibility, signal-to-noise ratio, and signal intensity are summarized in Tables S1 and S2 of Supporting Information. Representative examples of the profiles of human plasma and urine obtained with different bead types, using the experimentally determined optimal conditions, are shown in Figures 3 and 4. The use of MSB appears to provide sufficiently rich fingerprints of plasma and urine, as revealed by comparing the mass spectra of unprocessed plasma (Figure 3, control) and urine

920 Bioconjugate Chem., Vol. 20, No. 5, 2009

Terracciano et al.

1100-5000 for plasma and 1100-6000 for urine, exhibiting some sporadic high intensity peaks outside this range. The numbers of peaks detected over the entire range from 1100 to 10000, in the case of plasma samples processed with the various kinds of silica, are reported in Table 2. Among all bead types, aminopropyl-derivatized MSB-C provided the best performance in resolving the plasma peptidome (160 peaks detected), thus indicating that mesoporous silica materials with a pore size of roughly 4 nm and with a surface area of around 270 m2/g could be the starting point for future chemical tailoring aimed at increasing the affinity for plasma peptides. Our results extend the findings of Geho and co-workers, who used aminopropylcoated 17 and 70 nm pore-sized glass beads from Sigma for serum fractionation, obtaining, however, only a few dozen peptides (23). The data presented here (Table 2) strongly suggest that, besides functionalization, tuning the pore size, pore volume, and the surface area of mesoporous silica may also be a viable approach to biomarker harvesting. Future development of our technology along this line will include surface functionalization with alkyl and aryl groups, which could increase the hydro-

Table 3. Mean Number of Peaks Detected by Each Bead Type from Human Urine human urine bead type

peaks detecteda range 1100-6000 Da

MSB-A MSB-B MSB-C MSB-D MSB-E

92 40 82 44 199

a The number of peaks detected in the range between 1100 and 6000 Da with a S/N > 10.

(Figure 4, control) to mass spectra generated after MSB processing (Figure 3 and 4). All MALDI spectra were generated using the same settings in the range from 800 to 10000 Da, imposing the stringent condition of a signal-to-noise ratio of higher than 10 for peak detection (see Experimental Section). Apart from some peak clusters related to the matrix CHCA used in MALDI analysis in the range between 800 and 1100 Da, most of the detected signals for all of the bead functionalities were in the range from

Table 4. Peptides Identified by MALDI-MS/MS Fragmentation from Human Plasma after Processing on MSB-A, MSB-B, and MSB-C observed mass

theoretical mass

mass error

MS/MS sequence

m/z

m/z

Da

peptide

1136.67 1348.72 1441.84 1501.80 1629.89 1846.92 2209.65 2310.21 2578.28

1136.61 1348.66 1441.79 1501.74 1629.83 1846.87 2209.59 2310.16 2578.23

0.06 0.06 0.05 0.06 0.06 0.05 0.06 0.05 0.05

MRDVVLFEK RHDWGHEKQR MKPVPDLVPGNFK MELERPGGNEITR DSHSLTTNIMEILR ADSGEGDFLAEGGGVRGPR KHNLGHGHKHERDQGHGHQ HTFMGVVSLGSPSGEVSHPRKT DSHSLTTNIMEILRGDFSSANNR

signal-to-noise ratio MSB-A

MSB-B

MSB-C

68.30 13.47

11.66 165.44

12.82 50.83 224.25

155.46

115.68 101.69 39.48 10.1 115.44 63.05

40.33 257.95

79.46 231.00

Table 5. Reproducibility Assessment for Peak m/z and Signal Intensity in Acquired MALDI TOF Spectra from Replicate Analyses (n ) 3) of Human Plasma Samples Processed by Derivatized Mesoporous Silica Beads Human Plasma MSB-C

MSB-D

MSB-E

m/z

CV (%) on m/z

CV (%) on height

m/z

CV (%) on m/z

CV (%) on height

m/z

CV (%) on m/z

CV (%) on height

1350.70 1451.77 1501.14 1896.31 2106.99 2289.36 2484.54 3215.56 3315.67 4576.47 mean

0.0087 0.019 0.0094 0.012 0.014 0.014 0.013 0.017 0.014 0.017 0.014

14.11 1.93 17.30 9.20 31.28 5.31 33.08 8.92 9.54 12.94 14.36

1515.63 1532.29 1880.90 1896.49 1936.06 2210.97 2256.66 2349.95 2366.77 3066.33

0.049 0.022 0.017 0.017 0.015 0.017 0.014 0.013 0.012 0.0091 0.018

8.42 4.63 10.64 1.96 6.06 10.57 9.80 9.45 10.29 7.59 7.94

1084.54 1098.55 1184.94 1921.70 2022.81 2046.83 2193.15 2314.22 2866.20 3113.44

0.0023 0.0018 0.0073 0.015 0.0023 0.0064 0.0023 0.0022 0.00092 0.00023 0.0041

6.41 21.77 5.68 11.09 4.11 9.84 5.37 4.10 3.31 7.69 7.94

Table 6. Reproducibility Assessment for Peak m/z and Signal Intensity in Acquired MALDI TOF Spectra from Replicate Analyses (n ) 3) of Human Urine Samples Processed by Derivatized Mesoporous Silica Beads Human Urine MSB-C

MSB-D

MSB-E

m/z

CV (%) on m/z

CV (%) on height

m/z

CV (%) on m/z

CV (%) on height

m/z

CV (%) on m/z

CV (%) on height

1181.33 1770.73 1897.32 1914.97 2194.99 2439.33 2619.08 2793.13 2978.65 3391.03 mean

0.028 0.020 0.0021 0.031 0.041 0.046 0.0066 0.043 0.028 0.026 0.027

16.83 32.05 4.14 33.71 5.85 7.55 8.62 20.31 32.23 10.61 17.17

1369.69 1656.70 1770.27 1914.54 2067.57 2194.74 2439.99 2774.59 2924.64 3415.09

0.0054 0.0057 0.0066 0.0067 0.0044 0.0034 0.018 0.0079 0.011 0.022 0.0091

17.49 8.29 1.72 0.00 7.56 13.58 9.65 13.76 12.56 11.68 9.63

1219.95 2238.03 1520.92 1914.19 2080.66 2193.91 2438.26 2791.65 4293.66 4752.78

0.0094 0.0026 0.011 0.015 0.031 0.017 0.022 0.020 0.017 0.021 0.017

23.69 10.68 27.32 2.68 18.59 6.79 5.68 0.00 1.09 28.56 12.51

Functionalized Mesoporous Silica Beads

Figure 5. Peak areas of the tested peptides spiked in the urine sample, plotted against their concentrations. The assayed peptides are (a) angiotensin I (1297.51 m/z) and (b) adrenocorticotropic hormone (ACTH; clip 1-17; 2094.46 m/z). A linear equation is used in curve fitting. The error bars are the standard deviations of the mean peak areas.

phobicity of the mesoporous silica and allow the selective capture of hydrophobic peptides for mass spectrometric assays. MSB-A and MSB-B, previously tested on plasma samples (21), have also been tested for urine in this investigation exhibiting very good mass coverage in the m/z range 1100-6000 (Figure 4, panels a and b). Urine analysis using MSB-A in fact produced 92 signals, while a lower number of peaks was generated in this range by MSB-B (40 peaks) (Table 3). MALDI spectra from anionic beads MSB-C and MSB-D show 82 and 44 peaks, respectively. Very high mass coverage is observed in the case of MSB-E with 199 peaks detected (Figure 4, panel e and Table 3). Mass profiles obtained from urine processed by MSB-C and MSB-D share many features in common as shown in Figure 4c and d, but closer examination and comparison of the m/z values and peak intensities reveal a higher efficiency for MSB-D at low m/z, ranging from 1100 to 1880 in comparison to that of MSB-C. In the case of plasma, a few common peaks are more abundant when MSB-D is used instead of MSB-C (Figure 3, panels c and d). One possible explanation for the lower efficiency of MSB-D in comparison to that of MSB-C is the different chemical functionalities present on the mesoporous surface. It may be speculated that the N-(2aminoethyl)-3-aminopropyl groups of MSB-D possess higher selectivity toward plasma and urine components in comparison to that of the aminopropyl group of MSB-C. For plasma samples, there are very few signals in the mass spectrum generated by MSB-E (Figure 3, panel e). Although valuable for selective binding studies, MSB-E does not provide the best choice for plasma peptidome profiling. On the contrary, this kind of beads significantly amplifies the peak detection of urine peptides as clearly shown in Figure 4e. These results indicate that different mesoporous silica with unique surface characteristics produce different MALDI mass spectra for an identical plasma or urine sample; hence, these MSB could be used to obtain complementary proteomic information with high sample throughput. To assess the selectivity of sample processing with MSB, as proof-of-principle we identified by MALDI-TOF/TOF some peaks, both from nonfunctionalized (MSB-A and MSB-B) and from functionalized (MSB-C) silicas. Therefore, in a preliminary

Bioconjugate Chem., Vol. 20, No. 5, 2009 921

analysis, we randomly selected for identification the following peaks, retained from plasma by MSB-A, MSB-B, and MSB-C: peak at m/z 1442, retained by MSB-A only; peak at m/z 1847, retained by MSB-B only; peak at m/z 2210, retained by MSB-C only; peaks at m/z 1630, and 2578, retained by MSB-A and MSB-B but not MSB-C; peak at m/z 1137, retained by MSB-B and MSB-C but not MSB-A; finally, peaks at m/z 1349, 1502, and 2310, retained by all three MSB examined. The selected peaks are marked with an asterisk in the mass spectra shown in Figure 3 panels a, b, and c. The amino acid sequences of each of the selected peptide are shown in Table 4, together with the corresponding mass errors and signal-to-noise ratios. Interestingly, peaks extracted selectively by a single MSB corresponded to specific peptides of unique sequence (Table 4), strongly suggesting that sample processing with MSB is compatible with peptidomic analysis. Moreover, even when the same peptide is extracted from plasma by more that one MSB, there is often a significant difference in the signal-to-noise ratio. For instance, the peptide at m/z 2578, retained by both MSB-A and MSB-B, is characterized by a signal-to-noise ratio of 63.05 in the mass spectrum of the MSB-A eluate and of 257.95 in the mass spectrum of MSB-B eluate, further suggesting that each MSB determines a significant increase in specific peptide components (Table 4). Considering the stringent conditions used for peak detection, a signal-to-noise ratio of greater than 10, the MALDI spectra generated from plasma and urine, after MSB processing, resulted in a remarkable improvement in peak detection compared to that in the SELDI spectra reported in the literature (23, 33-35). In contrast to the SELDI chip arrays, MSB, with their greater surface area, can provide higher binding capacity, and moreover, the possibility of thoroughly washing the beads allows the effective removal of salts and other contaminants, which could negatively affect MALDI spectra quality. This particular limitation of SELDI, which resides in the small surface area of its flat chips directly used as MALDI targets, is also reported elsewhere (36). In a recent study, Tempst and co-workers specifically report that the use of porous particles for sample pretreatment is more sensitive than surface capture on chips because particles have a larger combined surface area than smalldiameter spots (14). Before peptidome analysis based on mass spectrometric screening can proceed from the bench to the bedside, a large number of important obstacles need to be overcome, among which are the valid and reproducible analyses of peptide patterns (37, 38). For this reason, we have investigated the reproducibility of the peptidome profiling of human plasma and urine by derivatized MSB in combination with MALDI-TOF MS. Plasma and urine samples were run in triplicate with all of the sample preparation methods, then the spectra obtained were used for data analysis. Triplicate MALDI spectral readouts of plasma and urine samples processed by the derivatized MSB are shown in Supporting Information (Figures S7 and S8). The reproducibility of the mass spectra was evaluated with respect to the relative peak intensities and mass accuracy; variability was measured by coefficient of variance (CV). Tables 5 and 6 list CVs of the m/z values and the intensities of peaks randomly selected over a wide mass range for plasma and urine, respectively. In the plasma experiments, average CVs of 0.014%, 0.018%, and 0.0041% of the mean m/z values were obtained for MSB-C, MSB-D, and MSB-E, respectively. The average CV of the mean intensity values was 14.36% for MSB-C and 7.94% both for MSB-D and MSB-E (Table 5), ensuring good reproducibility for MALDI MS-based profiling studies. In the urine experiments, the average % CV corresponding to the mean m/z values was 0.027% for MSB-C, 0.0091% for MSB-D, and 0.017% for MSB-E. The mean CVs of the signal intensities for

922 Bioconjugate Chem., Vol. 20, No. 5, 2009

Terracciano et al.

Table 7. Accuracy and Precision Assessment of m/z Values for the Standard Peptides Spiked into Urine Sample and Processed by MSB-C MALDI-TOF MS standard peptide

standard peptide +

ACTH (clip 1-17) (theoretical mass, MH+ 2094.46)

angiotensin I (theoretical mass, MH 1297.51) concentration (ng/mL)

mean m/z

% error

CV (%) on m/z

concentration (ng/mL)

meana m/z

% error

CV (%) on m/z

1000 500 250 125 62.5 26 6.5 3.7

1297.83 1297.82 1297.09 1297.38 1297.67 1297.34 1297.33 1297.33

-0.025 -0.024 0.032 0.010 -0.012 0.013 0.014 0.014

0.06 0.02 0.02 0.02 0.02 0.05 0.04 0.07

1614 807 404 202 101 42 10.5 6.0

2094.83 2094.90 2094.17 2094.69 2096.95 2097.18 2096.06 2097.11

-0.018 -0.021 0.014 -0.011 -0.023 0.013 0.019 0.017

0.06 0.02 0.02 0.01 0.02 0.03 0.04 0.04

a

a

The m/z values are the means from three independent experiments.

10 urinary signals were 17.17% (ranging from 4% to 33%) for MSB-C, 9.63% (ranging from 0 to 17%) for MSB-D, and 12.51% (ranging from 0 to 28%) for MSB-E (Table 6). In order to provide more information about the robustness, reproducibility, and linearity of the methodology, we added internal standards to urine before processing with MSB-C with a series of dilutions of angiotensin I (m/z 1297.51) and adrenocorticotropic hormone ACTH (clip 1-17) (m/z 2094.46). The analysis of 3 replicates of urine aliquots spiked with these two standard peptides at decreasing concentrations showed that there was a linear correlation between peak area and peptide concentration (R2 > 0.98) in the range of 1 to 2000 ng/mL (Figure 5). We were able to detect these standard peptides in human urine with concentrations as low as 3.7 ng/mL for angiotensin I and 6.0 ng/mL for the fragment 1-17 of ACTH (Figure 5). Therefore, also for human urine, we confirmed the low-nanogram per milliliter sensitivity of the methodology, previously demonstrated for human plasma (21). The precision and the accuracy of the MSB-MALDI-TOF MS methodology was evaluated as percentage error on the theoretical mass values, resulting for the tested standards with vaues