Enhancing plasma peptide MALDI-TOF-MS profiling by mesoporous silica assisted crystallization

Talanta 80 (2010) 1532–1538

Contents lists available at ScienceDirect

Talanta journal homepage: www.elsevier.com/locate/talanta

Enhancing plasma peptide MALDI-TOF-MS profiling by mesoporous silica assisted crystallization Rosa Terracciano a,∗ , Francesca Casadonte a , Luigi Pasqua b , Patrizio Candeloro a , Enzo Di Fabrizio a , Andrea Urbani c , Rocco Savino a a b c

Department of Experimental and Clinical Medicine, “Magna Graecia” University of Catanzaro, 88100 Catanzaro, Italy Department of Chemical Engineering and Materials, University of Calabria, 87036 Cosenza, Italy University of Rome “Tor Vergata”, Department of Internal Medicine and S. Lucia Foundation - IRCCS, Rome, Italy

a r t i c l e

i n f o

Article history: Available online 5 April 2009 Keywords: Clinical proteomics Biomarkers discovery MALDI-TOF-MS Mesoporous materials Human plasma

a b s t r a c t Promising profiling techniques based on new material/solid phase extraction for capturing “molecular signatures” from body fluids are being coupled to MALDI-TOF-MS. Sample preparation significantly influences spectrum quality in this ionization method. Mesoporous silica beads (MSB), by the means of nano-sized porous channels with high surface area, enable harvesting of peptides from plasma and serum excluding large size proteins. We have investigated the morphology of a sample slurry, developed as a new tool for plasma peptides enrichment based on mesoporous materials. Our study highlights a correlation between crystals morphology and enhanced performances in MALDI-TOF-MS analysis. This is the first report which correlates the increase in signal intensity with crystal formation in samples preparations which make use of various kinds of slurries for the analysis of samples clinically relevant like human plasma. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Discovery of biomarkers for disease diagnosis is the major task for modern predictive medicine. Besides traditional and wellestablished immunoassays (ELISA, RIA) and other technologies used for the detection of a single over-expressed molecule, proteome profiling of human body fluids based on mass spectrometry (MS) is now becoming an important method for investigating and detecting novel disease-associated markers [1–4]. The ability to screen and discover multiple biomarkers simultaneously has been advanced by the recent success of MS, especially surfaceenhanced laser desorption/ionization time-of-flight (SELDI-TOF) MS and matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) MS [5]. The demand for reducing sample complexity in order to increase the detection of low-abundance proteins has generated the development and implementation of proteomic technologies in the field of protein biomarker discovery. Most of these proteomic technology platforms are centered on the implementation of MS in conjunction with several other analytical techniques, such as solid phase extraction, chromatography and

∗ Corresponding author at: Department of Experimental and Clinical Medicine, “Magna Graecia”, Laboratory of Mass Spectrometry and Proteomics, University of Catanzaro, University Campus, Europa Avenue, 88100, Germaneto, Catanzaro, Italy. Tel.: +39 09613694080; fax: +39 09613694090. E-mail address: [email protected] (R. Terracciano). 0039-9140/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2009.03.060

electrophoresis [6]. A new emerging field of study is addressed to development, improvement and standardization of pre-MS separations approaches. Among well-established methods for sample preparation such as magnetic beads (MB) [7], or surface-derivatized chips (SELDI) [8], new separation/fractionation platforms are based on innovative application of both traditional and novel materials for selective capture of low molecular weight peptides and polypeptides from biofluid samples [9–12]. Material science is a mine of richness not yet exploited by proteomic community which could contain premises and promises for future generations of “biomarker discovery” pioneers. That is why the design and generation of material-based platforms for capturing “molecular signatures” from body fluids has gained increasing interest in recent years. Sample preparation is crucial step to generate reproducible and good quality mass spectra which are essential prerequisites for successful proteomic pattern analysis. Therefore optimization and more comprehensive investigations of sample preparation methods could be of high impact for MALDI-TOF-MS analysis of biofluids proteome. We report herein further investigations and optimization of our recently developed procedure based on mesoporous silica particles for selective binding and enrichment of low molecular weight plasma/serum proteins by MALDI-TOF-MS analysis [11]. Mesoporous silica beads (MSB) are incubated with the sample, washed, then suspension in ␣-cyano-4-hydroxycinnamic acid (CHCA) is loaded on the MALDI target plate. In comparison to the previous standard protocol [11], the results emphasize a significant increase in signal intensity and decrease in background by new modified

R. Terracciano et al. / Talanta 80 (2010) 1532–1538

procedure. We have additionally investigated by scanning electron microscopy (SEM) on MALDI samples spots obtained from two different preparation procedures based on MSB. A different behavior in matrix crystallization is observed when sample is spotted on MALDI target plate as matrix-particles suspension, which changes dramatically plasma analysis performances.

1533

2. Experimental

Insuline (bovine). Stock solutions of standard peptides were all prepared in 0.1% of TFA. 1 ␮L of standard solution was directly spotted onto a well of MALDI plate and air-dried. Then 1 ␮L of CHCA matrix solution (prepared as described above) was overlaid on the sample and air-dried. In parallel, 1 mg of MSB was added to100 ␮L of standard solution, then 1 ␮L of this suspension was spotted onto a well of MALDI plate and air-dried. Then 1 ␮L of CHCA matrix solution was overlaid on the sample and air-dried.

2.1. Material preparation and characterization

2.4. MALDI-TOF-MS analysis

The synthesis of MSB was performed under hydrothermal condition in presence of cationic surfactant dodecyltrimethylammonium bromide [C12 H25 N(CH3 )3 Br] (CTABr). The molar composition of the starting mixture was: SiO2 :0.2 CTABr, 0.2 NaOH, 0.04 Al(OH)3 , 40 H2 O. 10 g of fumed silica were added to a solution of 0.52 g of Al(OH)3 , 12.1 g of CTABr, and 1.35 g of NaOH in 120 g of H2 O. The gel was aged for 2 h at room temperature and then heated in oven for 24 h at 140 ◦ C. The product obtained was filtered, washed with deionized water, and dried at 80 ◦ C for 12 h. The activated sample was obtained by solvent extraction in a Soxhlet apparatus. The nitrogen adsorption–desorption volumetric isotherms at 77 K were measured on a Micrometritics Asap 2010 apparatus. Surface area of the samples was obtained by Brunauer–Emmet and Teller (BET) linearization.

MALDI-TOF-MS analysis was performed on a mass spectrometer (Voyager-DE STR, Applied Biosystems, Foster City, CA, USA) equipped with a 337 nm nitrogen laser. Analyses were performed in linear positive-ion mode, using delayed extraction. The acceleration voltage was 20 kV, guide wire was 0.05% of the accelerating voltage, grid voltage was 91.5%, and the delay time was 220 ns. Same TOF-MS settings and laser shot acquisition of the sample spots were applied to carefully compare spectral read-outs from different MALDI preparations. In particular five 100-laser shots were averaged for each mass spectrum to obtain a good statistical analysis. All spectra were normalized to the base peak and a baseline correction was performed using Data Explorer 4.0 (Applied Biosystems, Framingham, MA, USA). All experiments were carried out at least in triplicate to ensure the reproducibility of the spectra. Same instrument settings were used for both elution and suspension procedures.

2.2. Plasma collection

2.5. Scanning electron microscopy

Blood intended for plasma preparation was collected into the following tubes: BD Plus Plastic K2EDTA, 10 mL, # 367525. The samples were centrifuged at 1300 × g for 10 min under refrigerated conditions (2–6 ◦ C), within 15 min of the draw. The resultant plasma was pooled into one secondary conical bottom BD (TM) Falcon tube. The secondary tube was centrifuged at 2400 × g for 15 min to remove potentially remaining cellular material and to prepare platelet poor plasma. Samples were then aliquoted and immediately frozen at −80 ◦ C. All experiments were performed on aliquots of first-thawed poor platelet plasma.

The samples were first prepared on MALDI stainless steel target plate following the same procedures of the samples for MALDITOF-MS analysis. The samples topography was characterized by scanning electron microscopy, using the electron column of a dual beam system (FEI Nova 600 Nanolab) equipped with a thermal field emitter as electron source. Low beam currents in combination with

2.3. Experimental procedures Substance P was spiked in human plasma at a concentration of 12.5 ng/ml. Plasma-Substance P was diluted 1:5 (40 ␮L of plasma in 160 ␮L of deionized water), MSB (2 mg) were mixed with diluted human plasma-Substance P sample and shaken at room temperature for 1 h. The suspension was centrifuged at 2000 × g for 2 min, and then MSB were separated from the supernatant and washed with deionized water (4× 30 ␮L). After the last wash, bound species were directly extracted in CHCA (30 ␮L of a 4 mg/mL solution prepared dissolving the matrix in a 1:1 mixture of acetonitrile and 0.1% trifluoroacetic acid). A volume of the above slurry (1 ␮L) was loaded on MALDI target plate in the “suspension” procedure, otherwise, after centrifugation at 2000 × g for 2 min, 1 ␮L of supernatant solution (30 ␮L) with extracted peptides, was deposited on MALDI probe (“elution” procedure). Both in suspension and in elution procedure, equal volume of matrix solution (30 ␮L) was used for extracting bound species from beads, and in both cases 1 ␮L was deposited on MALDI probe. Protein concentrations were determined using the Bio-Rad Laboratories (Hercules, CA, USA) protein assay kit (Bradford method) and bovine gamma-globulin as standard following the instructions furnished by the manufacturer. Standard solution was prepared by mixing 100 ␮L of a 1 mg/mL stock solution of Substance P, with 100 ␮L of a 1 mg/mL stock solution of Dynorfine and with 100 ␮L of a 1 mg/mL stock solution of

Fig. 1. (A) Nitrogen adsorption–desorption isotherms and (B) Barret–Joyner– Halenda (BJH) pore size distribution of mesoporous silica beads used in proteomic experiments.

1534

R. Terracciano et al. / Talanta 80 (2010) 1532–1538

a small accelerating voltage of 5 kV were employed in order to prevent the sample from any possible damage. Furthermore, the low beam currents limited the charging effects and consequently no previous metal coating of the sample was required for the analysis. 3. Results and discussion Mesoporous materials constitute a new generation of materials that show ordered arrangements of channels and cavities of different geometry built up from SiO2 unities [13]. The pore size is variable (2 nm < ˚p < 50 nm) and can be controlled and modified, in a reasonable range, using in situ [14] and ex situ [15] synthetic strategies. These materials find many uses in catalysis, metal ion extraction, optical applications. Moreover silica particles with highly ordered mesostructures, for example, M41s [16] and SBA-15 [17], have been widely applied in the fields of separation and adsorption [18–20]. MSB used in the experiments hereby described belong to highly ordered mesoporous silica particles synthesized according to reported procedure [21]. Representative nitrogen adsorption–desorption isotherms (Fig. 1, panel A) and the pore size distribution diagram for MSB (Fig. 1, panel B) demonstrate the existence of highly ordered mesopores running along the same direction. In particular nitrogen adsorption–desorption isotherms of sample (SBET = 848 m2 /g, pore volume = 1.21 cm3 /g, show two mesopore filling steps at P/P0 range 0.2–0.4 and 0.9–1.0, indicating a bimodal pore system. The Barret–Joyner–Halenda (BJH) pore

size distribution is centered at 25 and 390 Å. The first peak is associated with primary mesopores generated by surfactant micelles, the second one is associated with secondary mesopores resulting from secondary silica condensation. Here we extended our previous studies in which we developed a strategy based on MSB for plasma peptides profiling [11]. Given the high surface area mesoporous silicates offer the desired adsorptive capacity for binding and enrichment of low molecular weight peptides present in body fluids. Tuning the pore size distribution and surface adsorptive properties, MSB selectively capture peptides and small size molecules excluding large size proteins such as Human Serum Albumin. Plasma samples are exposed to mesopores, then captured molecular species are extracted and profiled by MALDI-TOF-MS. In order to make the analysis more rapid and to circumvents the risk of sample loss, we refined the previous protocol by direct spotting of MSB-loaded suspension on the MALDI target plate. As illustrated in Fig. 2 we called “elution” procedure the previous protocol in order to distinguish from the new modified procedure called “suspension” procedure. A detailed description of procedure is summarized in experimental section. We prepared our samples with CHCA given its optimal chemical–physical properties as a matrix for the MALDI mass spectrometric analysis of peptides. A comparison of the performance in MALDI-TOF-MS analysis of human plasma, between the “suspension” procedure and the “elution” procedure is shown in Fig. 3 where the MALDI-TOF-MS spectra are evaluated in absolute ion counts units for the same m/z range. A significant enhancement in signal intensity as well as the

Fig. 2. Extraction of plasma peptides by MSB and MALDI sample preparation on-target plate by “elution” and “suspension” procedure. MSB were mixed with diluted human plasma and shaken at room temperature for 1 h (adsorption step). The suspension was centrifuged, the MSB were separated from the supernatant (separation step) and washed with deionized water (washing step). After the last wash, bound species were directly extracted in CHCA (extraction step). 1 ␮L of the above slurry was loaded on MALDI target plate in the “suspension” procedure, otherwise, after centrifugation 1 ␮L of supernatant solution, was deposited on MALDI probe (“elution” procedure). SEM images of two different preparations with magnification inserts are provided.

R. Terracciano et al. / Talanta 80 (2010) 1532–1538

Fig. 3. Performance of the “elution” and “suspension” procedures in MALDI-TOF-MS analysis. (A) MSB were incubated with the sample, washed, captured peptides were eluted and analyzed (“elution” procedure). (B) MSB were incubated with the sample, washed, matrix solution was added and finally 1 ␮L of slurry was loaded on MALDI target plate. In both cases Substance P was spiked at 12.5 ng/ml (green asterisk). High, medium, and low intensity peaks within the mass range of m/z 1200–3500 were randomly selected (asterisks) for the variability assessment. Signal intensity is plotted in the same absolute units for both procedures.

number of peaks detected is clearly visible in the new “suspension procedure” (Fig. 3, panel B). A typical effect observed in the MALDI technique is that the spectra do not show good stability of the ion current intensity, therefore to obtain a good statistical analysis we used five hundreds laser pulses to average the mass spectra. Moreover, to carefully compare spectral read-outs from different MALDI preparations, same TOF-MS conditions and laser shot acquisition of the sample spots were applied both in “suspension” and “‘elution” procedure.

1535

In order to evaluate the increased detection sensitivity achieved by suspension procedure, Substance P was added to human plasma sample at the concentration of 12.5 ng/ml in both elution and suspension procedures. Substance P was detected with a signal over 10-fold higher when the sample was deposited on MALDI probe as slurry instead of as simple solution (Table 1 and Fig. 3). In order to better discriminate whether the enhancement in signal intensity we observed was due to different analyte concentration, total protein concentration of the samples spotted on MALDI plate was measured by Bradford assay in three different experiments. Interestingly we found that the supernatant solution by the “elution procedure” contained a total protein concentration of 3.9 ± 0.1 ␮g/␮L, while, in the case of the slurry, the measured protein concentration was 3.68 ± 0.03 ␮g/␮L. Therefore, we consider highly unlikely that the poorer mass spectrum observed in the “elution” procedure may be due to lower analyte concentration. To better analyze the gain in ion yield, we randomly selected a list of ten peaks (indicated with asterisk in Fig. 3, panel B) of low, medium and high intensity in a set of three replicate experiments. The results are illustrated in Table 1, which shows average absolute areas and area ratios of peaks obtained with suspension and elution procedure. We observed a substantial enhancement in peak areas which for some peaks resulted more than hundred fold, whereas a minority of peaks were detected with roughly the same intensity (Table 1). Although mass accuracy was slightly lower in “suspension procedure”, we found a higher reproducibility in peaks area in “suspension procedure” (see percentage coefficient of variance in Table 1). In the higher m/z range between 3500 and 7500, the MALDI mass spectrum from suspension procedure (Fig. 4, panel B) shows more spectral features with higher S/N when compared to the MALDITOF-MS spectrum from elution preparation (Fig. 4, panel A), even though we notice the presence of some isolated little peaks, more evident in elution preparation, marked with red arrows in Fig. 4, panel A. The superior performance of suspension preparation is evident as the highest signal intensities were gained with almost no background noise (Fig. 4, panel B). Signal-to-noise ratio was increased ninefold for the m/z 3519 peak (from 4.7 to 43.4), and 16-fold for the m/z 3830 peak (from 5 to 83) when slurry was used

Fig. 4. MALDI-TOF mass spectra of human plasma treated with MSB and spotted on MALDI probe with “elution” (A) and “suspension” (B) procedures in the mass range of m/z 3500-7500. Same amount of plasma and MSB were used in both procedures as well as the volume of matrix solution used for extraction and for MALDI sample preparation on the probe. Instrument settings were the same in both experiments. Red arrows in panel A show little isolated peaks more evident in “elution” vs. “suspension” procedure. Signal intensity is plotted in the same absolute units for both spectra.

1536

R. Terracciano et al. / Talanta 80 (2010) 1532–1538

Fig. 5. Comparison of S/N ratio for “elution” (cyan) and “suspension” (magenta) procedures. The values were averaged on three independent experiments and reported, with error bars, on logarithmic scale.

in sample preparation. Enhancement of S/N was also estimated for peaks 4301 (from 8 to 22) and 4965 (from 6 to 25) in suspension preparation as well as the peaks 5707, 6179, 6435 and 6633 were lost in elution preparation (Fig. 4, panel A and B). To complete MS spectra analysis we also evaluated signal to noise ratios for the same set of peaks of Table 1. Fig. 5 illustrates the stringent increase of performance in suspension procedure with

respect to elution procedure with S/N values which increases several fold (over 100-fold for peaks 1896 and 2021). To gain deeper insight in the mechanisms determining the increased performance of the “suspension procedure”, inspection of sample morphology of the two different preparations on MALDI plate was investigated by SEM. This study revealed a higher density of CHCA crystals over the entire well area of sample in the new

Fig. 6. SEM analysis. (A) SEM image of on-target sample preparation by “elution” procedure; (B) SEM image of on-target sample preparation by “suspension” procedure; (C) A superb flower-shaped crystal obtained from analytes in ␣-cyano-4-hydroxycinnamic acid matrix, grown among mesoporous silica beads on a typical stainless steel MALDI probe. The crystals radially grow around a bud and a flower-shaped crystal consists of approximately 20 crystalline parts looking like “petals” of the flowers. Promotion of these micron-sized crystals significantly improves MALDI mass spectral read-outs of human plasma peptides extracted by mesoporous silica beads and deposited as slurry on MALDI probe.

R. Terracciano et al. / Talanta 80 (2010) 1532–1538

1537

Table 1 Comparison of mass accuracy and absolute peak areas in suspension and elution procedure. m/za

%CV on mass accuracyb e

%CV on mass accuracyc e

Average peak areab (%CV)

Average peak areac (%CV)

Peak area ratiod 14.40 24.51 10.61 1.26 231.09 102.87 21.71 1.15 2.22 1.70

1349 1499 1741 1848 1896 2021 2107 2485 2861 3065

0.00 (i.s.) 0.033 0.0094 0.043 0.044 0.048 0.052 0.025 0.060 0.047

0.00 (i.s.) 0.041 0.016 0.030 0.029 0.027 0.043 0.029 0.012 0.016

10347 (12.39) 16859 (24.55) 6462 (3.17) 18667 (5.21) 84094 (15.45) 44434 (11.45) 19874 (23.39) 78137 (7.82) 8140(9.17) 32487 (3.65)

718 (19.77) 688 (36.30) 609 (1.65) 14853 (6.88) 364 (15.33) 432 (20.50) 915 (25.72) 68071 (23.86) 3658 (36.85) 19099 (29.92)

Mean

0.036

0.024

%CV = 11.62

%CV = 21.68

a b c d e

41.15

Analysis and variability assessment for peak areas are performed on MALDI-TOF spectra acquired from replicate analyses (n = 3). Suspension procedure. Elution procedure. %Coefficient of variance is in parentheses. Suspension vs. elution peak area ratios. Internal standard.

“suspension procedure”, whereas typically “sweet spots” are observed at the edge of the spot in the standard procedure (see Supplementary material Fig. 1). We speculate that when silica particles are present in sample preparation they represent centers of nucleation for crystals growth after the deposition on MALDI probe. When the particles are absent, as it happens in elution procedure, the crystallization takes place on the smooth surface of the stainless-steel probe thus determining a lower density of crystals per unit area. We observed with SEM that matrix-analytes solution (elution procedure) and matrix-analytes-particles suspension (suspension procedure) crystallize in different shapes and dimensions (Fig. 6). In the case of elution procedure crystallites with roughly spherical shapes are predominant with an average diameter of 7 ␮m (Supplementary material Fig. 2), whereas larger flower-like crystals are predominant in the suspension procedure (Fig. 6, panel B and C). Careful inspection by SEM revealed that these flowers have different sizes and shapes with petals from 10 to 20 ␮m (Fig. 6, panel C and Supplementary material Fig. 3).

To further support the hypothesis that the increased performance observed in the case of the “suspension procedure” is due mainly to the crystal shape and not to other reasons (for instance, different analyte concentration) we performed an additional experiment in which MSB were exclusively used for assisting crystallization and not for peptide harvesting. A standard solution containing Substance P, Dynorfine and Insuline was deposited on MALDI probe with and without mesoporous silica, air-dried and then covered with a CHCA solution. As shown in Fig. 7, MALDI and SEM data confirmed the results obtained with human plasma peptides. An enhancement in signal intensity (Fig. 7, panels A and C) was observed in the presence of MSB, which correlated with the appearance of flower-like crystals (Fig. 7, panel B); considering that in this case the peptide concentration is identical in the two samples, the most likely explanation for the observed difference is the different crystal shapes obtained from two sample preparations (Fig. 7, panels B and D and supplementary material Fig. 4). It is important to underline that the crystals obtained in this

Fig. 7. MALDI-TOF MS spectra and SEM images of a standard solution from two different preparations on MALDI probe. The standard solution was prepared with Substance P (m/z 1348), Dynorfine (m/z 2148) and Insuline (m/z 5734); each peptide at a concentration of 0.33 ␮g/␮L. Panel A illustrates the mass spectrum obtained when 1 ␮L of standard solution containing mesoporous silica particles was deposited on MALDI plate, air-dried and then the solution of CHCA was added. Panel B shows a high magnification SEM image of floral-shaped crystals obtained from above described preparation. Panel C illustrates the mass spectrum obtained when 1 ␮L of standard solution without mesoporous silica particles was deposited on MALDI plate, air-dried and then the solution of CHCA was added. MALDI spectra are compared at the same absolute counts. Panel D shows a high magnification SEM image of spherulitic crystals obtained from preparation without silica particles.

1538

R. Terracciano et al. / Talanta 80 (2010) 1532–1538

experiment, with the standard solution in absence of MSB, retain the spherical shape already observed in the “elution” procedure. Flower-like shapes are still observed when MSB are used, but in comparison to previously described suspension procedure, these flowers are obviously slightly different, due to the different crystallization conditions. Sample morphology is considered critical in determining the ion yield in MALDI-MS of proteins [22]. It is tempting to speculate that the shape of the crystals obtained with the “suspension procedure”, with a more favorable surface to volume ratio, offers a larger surface to the laser beam, allowing a more efficient desorption/ionization process compared to the crystals obtained with the “elution procedure”. Although MALDI is a very popular mass analysis technique, it is still a poorly understood process and one has to be cautious in extrapolating the results. Our experimental observations regarding the possible correlation between the crystals/analyte distribution and resultant MALDI signals have been reported in other systems [23]. Even if other studies investigated on the dependence of matrix crystal morphology on analyte ion yield [24], this is the first report which correlates the increase in signal intensity with crystal formation in samples preparations which make use of various kinds of slurries for the analysis of samples clinically relevant like human plasma. Materials with micrometer-scale regularity are of great technological applications in microelectronics, information storage, optics, biomedical implants, catalysis and separation technologies [25]. It is hoped that this study would ultimately lead to rational design of improved methods for sample preparation by controlling crystallization conditions in the wide range of biofluid proteomics applications. Detailed studies on influence of particles suspension on matrix-analytes crystallization as well as inspection of the crystallites morphology might help elucidating the mechanisms and facilitating MS-based proteomics research. Acknowledgment This work was supported by an AIRC 2005 grant and a Ministero della Salute, Dipartimento dell’Innovazione (Direzione Generale della Ricerca Scientifica e Tecnologica) 2006 grant.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.talanta.2009.03.060. References [1] J. Villanueva, J. Philip, D. Entenberg, C.A. Chaparro, M.K. Tanwar, E.C. Holland, P. Tempst, Anal. Chem. 76 (2004) 1560–1570. [2] X. Zhang, S.M. Leung, C.R. Morris, M.K. Shigenaga, J. Biomol. Technol. 15 (2004) 167–175. [3] G.L. Hortin, Clin. Chem. 52 (2006) 1223–1237. [4] G.L. Hortin, S.A. Jortani, J.C. Ritchie, R. Valdes, D.W. Chan, Clin. Chem. 52 (2006) 1218–1222. [5] R.E. Banks, M.J. Dunn, D.F. Horchstrasser, J.C. Sanchez, W. Blackstock, D.J. Pappin, P.J. Selby, Lancet 356 (2000) 1749–1756. [6] J.L. Luque-Garcia, T.A. Neubert, J. Chromatogr. A 1153 (2007) 259–276. [7] J. Villanueva, J. Philip, C.A. Chaparro, Y. Li, R. Toledo-Crow, L. De Noyer, M. Fleisher, R.J. Robbins, P. Tempst, J. Proteome Res. 4 (2005) 1060–1072. [8] T.W. Hutchens, T.T. Yip, Rapid Commun. Mass. Spectrom. 7 (1993) 576–580. [9] I. Feuerstein, M. Rainer, K. Bernardo, G. Stecher, C.W. Huck, K. Kofler, A. Pelzer, W. Horninger, H. Klocker, G. Bartsch, G.K. Bonn, J. Proteome Res. 4 (2005) 2320–2326. [10] X.L. Kong, L.C.L. Huang, C.M. Hsu, W.H. Chen, C.C. Han, H.C. Chang, Anal. Chem. 77 (2005) 259–265. [11] R. Terracciano, M. Gaspari, F. Testa, L. Pasqua, G. Cuda, P. Tagliaferri, M.C. Cheng, A.J. Nijdam, E.F. Petricoin, L.A. Liotta, M. Ferrari, S. Venuta, Proteomics 6 (2006) 3243–3250. [12] H. Chen, X. Xu, N. Yao, C. Deng, P. Yang, X. Zhang, Proteomics 8 (2008) 2778–2784. [13] J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge, K.D. Schmitt, C.T.W. Chu, D.H. Olson, E.W. Sheppard, S.B. Higgins, J.L. Schlenker, J. Am. Chem. Soc. 114 (1992) 10834–10843. [14] X.S. Zhao, G.Q. Lu, X. Hu, Chem. Commun. (1999) 1391–1392. [15] A. Firouzi, F. Atef, A.G. Oertly, G.D. Stucky, B. Chmelka, J. Am. Chem. Soc. 119 (1997) 3596–3610. [16] C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli, J.S. Beck, Nature 359 (1992) 710–712. [17] D.Y. Zhao, J.L. Feng, Q.S. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D. Stucky, Science 279 (1998) 548–552. [18] M. Hartmann, Chem. Mater. 17 (2005) 4577–4593. [19] G. Buchel, M. Grun, K.K. Unger, A. Matsumoto, K. Tsutsumi, Supramol. Sci. 5 (1998) 253–259. [20] M. Grun, C. Buchel, D. Kumar, K. Schumacher, B. Bidlingmaier, K.K. Unger, Stud. Surf. Sci. Catal. 128 (2000) 155. [21] L. Pasqua, F. Testa, R. Aiello, Stud. Surf. Sci. Catal. 146 (2003) 497–500. [22] M. Sadeghi, A. Vertes, Appl. Surf. Sci. 127–129 (1998) 226–234. [23] Y. Day, R.M. Whittal, L. Li, Anal. Chem. 68 (1996) 2494–2500. [24] I.D. Figueroa, O. Torres, D.H. Russell, Anal. Chem. 70 (1998) 4527–4533. [25] V. Gupta, P. Scharff, N. Miura, Mater. Lett. 60 (2006) 2278–2281.