Nanopore-Based Proteolytic Reactor for Sensitive and

In minutes, many more sample peptides from the in-nanopore digestion of protein ..... A “teardown” method to create large mesotunnels on the pore ...
1 downloads 0 Views 318KB Size
Download PDF
Anal. Chem. 2006, 78, 4811-4819

Nanopore-Based Proteolytic Reactor for Sensitive and Comprehensive Proteomic Analyses Wenqing Shui,†,‡ Jie Fan,†,‡ Pengyuan Yang,† Chunli Liu,† Jianjun Zhai,† Jie Lei,† Yan Yan,† Dongyuan Zhao,*,† and Xian Chen*,†,§

Department of Chemistry, Fudan University, Shanghai 200433, P. R. China, and BN-2, Biosciences Division, Los Alamos National Laboratory, Los Alamos, New Mexico 87545

Various silica-based microreactors have been designed that use enzyme immobilization to address technical concerns in proteolysis including inefficient and incomplete protein digestion. Most of current designs for proteolytic reactors can improve either protease stability or proteolysis efficiency of individual protein(s). However, the desired features such as rapid digestion, larger sequence coverage, and high sensitivity have not been achieved by a single microreactor design for broad range proteins with diverse physical properties. Here, unlike conventional enzyme immobilization strategies, we describe a novel proteolytic nanoreactor based on the unique three-dimensional nanopore structure of our newly synthesized mesoporous silica (MPS), FDU-12, which integrates substrate enrichment, “reagent-free” protein denaturation, and efficient proteolytic digestion. In our design, protein substrates were first captured by MPS nanopore structure and were concentrated from the solution. Following the pH change and applying trypsin, the denaturation and concurrent proteolysis of broadrange proteins were efficiently achieved. In minutes, many more sample peptides from the in-nanopore digestion of protein mixtures were detected by mass spectrometry, resulting in the identifications of a broad range of diverse proteins with high sequence coverage. The unique features of FDU-12 nanostructure that allow rapid, complete proteolysis and resulting enhanced sequence coverage of individual proteins were investigated by using Raman spectroscopy and comparative studies with respect to other MPSs. Proteolysis by sequence-specific proteases that convert proteins to characteristic sets of peptides is an essential step for the mass spectrometric (MS)-based characterization of complex protein mixtures on a genomic scale.1 Current available techniques of proteolysis, either in-gel or in-solution, are time-consuming. Further, peptide recovery from in-gel or in-solution digestions is highly dependent upon the structural properties of various * To whom correspondence may be addressed. E-mail: [email protected] or [email protected]. † Fudan University. ‡ These authors contributed equally to this work. § Los Alamos National Laboratory. (1) Aebersold, R.; Mann, M. Nature 2003, 422, 198-207. 10.1021/ac060116z CCC: $33.50 Published on Web 06/08/2006

© 2006 American Chemical Society

proteins; e.g., proteins with rigid structures tend to be resistant to a complete digestion. Insufficient sequence coverage could compromise the accuracy of proteome characterization. In fact, high sequence coverage of peptide sets derived from individual proteins is critical (i) to reduce ambiguity in the mass-to-charge (m/z)-based genome-scale searches for protein identities and (ii) to detect functionally related posttranslational modifications, protein variants including sequence point mutations, truncations, and isoforms. Enzyme/protease immobilization on microfluidic devices via covalent bonding with silica supports,2-4 physical adsorption,5,6 and encapsulation in gel matrixes2,3,7,8 has been used to facilitate proteolysis with enhanced protease stability and reusability. In practice, high-concentration trypsin is immobilized on different silica supports such as confined zones of capillary,9,10 packed beads,11 porous silicon,12 or porous polymer monoliths3 resulting in the hours-long procedure of conventional digestion being shortened to minutes. However, enzymatic activity, completion of proteolysis, detection sensitivity, and diversity of protein substrates may be compromised to achieve the rapid digestion. Additionally, most available methods for protease immobilization require multistep reactions. As a consequence, few of these microfluidic devices are useful for real-case proteomic analysis of protein mixtures in cellular extracts. Recently, mesoporous silicas (MPSs) have emerged as a new type of solid host for enzyme immobilization due to their open pore structure, uniform pore size distribution, and large pore volume.13-18 Importantly, many of these MPS nanostructures are (2) Peterson, D. S.; Rohr, T.; Svec, F.; Frechet, J. M. J. Anal. Chem. 2002, 74, 4081-4088. (3) Peterson, D. S.; Rohr, T.; Svec, F.; Frechet, J. M. J. J. Proteome Res. 2002, 1, 563-568. (4) Ekstrom, S.; Onnerfjord, P.; Nilsson, J.; Bengtsson, M.; Laurell, T.; MarkoVarga, G. Anal. Chem. 2000, 72, 286-293. (5) Craft, D.; Doucette, A.; Li, L. J. Proteome Res. 2002, 1, 537-547. (6) Doucette, A.; Craft, D.; Li, L. Anal. Chem. 2000, 72, 3355-3362. (7) Ye, M. L.; Hu, S.; Schoenherr, R. M.; Dovichi, N. J. Electrophoresis 2004, 25, 1319-1326. (8) Kato, M.; Sakai-Kato, K.; Jin, H. M.; Kubota, K.; Miyano, H.; Toyo′oka, T.; Dulay, M. T.; Zare, R. N. Anal. Chem. 2004, 76, 1896-1902. (9) Amankwa, L. N.; Kuhr, W. G. Anal. Chem. 1992, 64, 1610-1613. (10) Licklider, L.; Kuhr, W. G.; Lacey, M. P.; Keough, T.; Purdon, M. P.; Takigiku, R. Anal. Chem. 1995, 67, 4170-4177. (11) Wang, C.; Oleschuk, R.; Ouchen, F.; Li, J. J.; Thibault, P.; Harrison, D. J. Rapid Commun. Mass Spectrom. 2000, 14, 1377-1383. (12) Bengtsson, M.; Ekstrom, S.; Marko-Varga, G.; Laurell, T. Talanta 2002, 56, 341-353. (13) Diaz, J. F.; Balkus, K. J. J. Mol. Catal., B 1996, 2, 115-126.

Analytical Chemistry, Vol. 78, No. 14, July 15, 2006 4811

Figure 1. (a) TEM image viewed along [110] direction of the ordered nanoporous silica FDU-12. The image was taken by the JEOL 2011 TEM operating at 200 kV. This FDU-12 MPS was synthesized as previous described.22 (b) Schematic representation of in-nanopore substrate entrapment, unfolding, in situ proteolytic digestion, and subsequent MS identification.

compatible with the physical properties of proteins. However, their potential use as a matrix for biocatalysis has only been demonstrated through the enzymatic hydrolysis of small organic molecules, and the reaction efficiency was generally lower than that of free enzymes in solution,13,15,19-21 We demonstrate here a novel design of a proteolytic nanoreactor using our newly synthesized three-dimensional (3D) nanopore-based MPS, FDU-12,22,23 that has the versatile functions of sample enrichment, purification, and highly efficient proteolysis to meet the practical demands of rapid and comprehensive proteomic analysis of protein mixtures. Their structure-to-function (14) Takahashi, H.; Li, B.; Sasaki, T.; Miyazaki, C.; Kajino, T.; Inagaki, S. Chem. Mater. 2000, 12, 3301-3305. (15) Han, Y. J.; Watson, J. T.; Stucky, G. D.; Butler, A. J. Mol. Catal., B 2002, 17, 1-8. (16) Deere, J.; Magner, E.; Wall, J. G.; Hodnett, B. K. J. Phys. Chem. B 2002, 106, 7340-7347. (17) Lei, C. H.; Shin, Y. S.; Liu, J.; Ackerman, E. J. J. Am. Chem. Soc. 2002, 124, 11242-11243. (18) Thomas, J. M. J. Mol. Catal., A. 1999, 146, 77-85. (19) Yiu, H. H. P.; Wright, P. A.; Botting, N. P. J. Mol. Catal., B 2001, 15, 8192. (20) Mody, H. N.; Mody, K. H.; Jasra, R. V.; Shin, H. J.; Ryong, R. Indian J. Chem. Sect. A-Inorg. 2002, 41, 1795-1803. (21) Macario, A.; Calabro, V.; Curcio, S.; De Paola, M.; Giordano, G.; Iorio, G.; Katovic, A. In Impact of Zeolites and Other Porous Materials on the New Technologies at the Beginning of the New Millennium, Parts A and B; Aiello, R., Giordano, G., Testa, F., Eds. Proc. 2nd Int. FEZA Conf., Taormina, Italy, Elsevier: Amsterdam, 2002; Vol. 142, pp 1561-1568. (22) Fan, J.; Yu, C. Z.; Gao, T.; Lei, J.; Tian, B. Z.; Wang, L. M.; Luo, Q.; Tu, B.; Zhou, W. Z.; Zhao, D. Y. Angew. Chem., Int. Ed. 2003, 42, 3146-3150. (23) Fan, J.; Yu, C. Z.; Lei, J.; Zhang, Q.; Li, T. C.; Tu, B.; Zhou, W. Z.; Zhao, D. Y. J. Am. Chem. Soc. 2005, 127, 10794-10795.

4812 Analytical Chemistry, Vol. 78, No. 14, July 15, 2006

relationship in the proteolytic process was also investigated by using Raman spectroscopy and comparative studies with respect to other MPSs. EXPERIMENTAL SECTION Synthesis of MPS FDU-12.22,23 In a typical synthesis, 0.5 g of triblock copolymers EO106PO70EO106 (F127, BASF), 0.6 g of TMB, and 2.5 g of KCl were dissolved in 30 mL of 2 M HCl and stirred at 15 °C for 24 h. To this solution, 2.08 g of TEOS was added under stirring. After stirring for 24 h at 15 °C, the mixture was autoclaved at 100 °C for 24 h and for another 72 h at 140 °C. The solid product was collected by filtration and dried at room temperature in air. The organic templates were completely removed by microwave-assisted digestion.24 In-Nanopore Digestion (Figure 1b). For the proteolysis in MPSs, ∼2.5 mg of MPS solid materials in a 2.5-mL vial was first incubated with 0.4 mL of protein solution containing a total of 50 µg of individual protein (myoglobin or convalbumin) in sodium phosphate buffer (pH 6.0). The mixture was then agitated on an Eppendorf Thermomixer at 8 °C for 10 min for complete protein adsorption in MPSs. The MPS beads were pelleted by centrifugation and washed with Millipore water twice. Then, 10 µg of trypsin in 0.4 mL of sodium phosphate buffer (pH 6.0) was added into the vial containing the MPS beads and the same buffer for 10 min of vortexing at 8 °C. The beads containing protein/trypsin/ FDU-12 composite were then quickly washed and pelleted as (24) Tian, B. Z.; Liu, X. Y.; Yu, C. Z.; Gao, F.; Luo, Q.; Xie, S. H.; Tu, B.; Zhao, D. Y. Chem. Commun. 2002, 1186-1187.

described previously before removing the supernatant as minimum tryptic digestion occurs at both acidic pH and low temperature. There was no absorption at λ ) 280 nm for any supernatant or wash-offs measured by UV-visible spectrometry. Finally, the MPS beads were incubated with 0.2 mL of 25 mM NH4HCO3 buffer and shaken for 15 min at 37 °C. The digestion products in the supernatant were obtained by spinning down MPS beads, and the supernatant was diluted by 10-fold in 0.1% TFA for subsequent MS analysis. For in-nanopore proteolysis of the protein mixture, the procedure is very similar except that we loaded 2 mL of the total sample at pH 6.0 to ensure complete capture of each protein, and finally, the nanopore beads were incubated with 0.25 mL of NH4HCO3 for rapid digestion. After centrifugation, the supernatant was taken directly for subsequent MS analysis. In-Solution Digestion without MPSs. A 50-µg aliquot of predenatured individual protein or 61 µg of total protein mixture (thermodenaturation was performed at 100 °C for 5 min) in NH4HCO3 buffer was mixed with 10 µL of trypsin solution (2.5 µg/µL) to reach a total volume of 0.25 or 2 mL, respectively, and the mixture was incubated at 37 °C for up to 12 h. Different timecourse aliquots of in-solution digestion were taken for MS analysis. We fixed the protease/substrate ratio as 1:2 in all cases, rather than entrapping a large amount of enzymes to digest trace proteins as is often done for sol-gels and silica supports. Reusability Test of the In-Nanopore Digestion. The process of the proteolysis between myoglobin and trypsin in MPS is similar to the above description. After the reaction was completed, the MPS beads were recycled by centrifugation and washed with Millipore water twice for next-day digestion. The supernatants were collected each day for final Matrix-assisted laser desorption/ ionization time-of-flight (MALDI-TOF) measurements. This analytical process was carried out over a time period of 16 days. Preparation of Liver Nuclear Extracts. Mouse liver nuclear extracts were prepared using the protocol for nuclear protein extraction by Kawakami et al.25 with the following modifications. All procedures were carried out at 4 °C. The nuclear pellet was washed in the buffer (10 mM HEPES at pH 7.9, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT, and 0.5 mM PMSF) and centrifuged at 400g for 10 min. Following washing twice, the supernatant was discarded and the pellet was resuspended in the lysis buffer (8 M urea, 2% CHAPS, 18 mM DTT, and 1 mM PMSF). The nuclei were homogenized and then stirred for 30 min. After centrifugation at 21000g for 30 min, the supernatant was gathered and stored in aliquots at -80 °C for future use. In-Solution and In-Nanopore Digestion of Liver Nucleoprotein Fraction. For in-solution digestion, 10 µg of nuclear fraction of proteins dissolved in 100 µL of lysis buffer was mixed with 250 ng of trypsin (E/S ) 1:40 w/w) and incubated for 16 h. Similar to the in-nanopore digestion protocol as described above, ∼250 µg of FDU-12 beads was incubated with 10 µg of nucleoprotein extract diluted in 0.4 mL of sodium phosphate buffer (50 mM, pH 6.0). A total of 1.3 µg of trypsin (E/S ) 1:8 w/w) was then loaded into the same bead pellet. After washing and pelleting the beads, the buffer was changed to 100 µL of 20 mM NH4HCO3 for a 30-min digestion. In both cases, half of the final

products were taken to be concentrated into 5 µL by a N2 thermal concentrator. Capillary(µ)LC Separation of Both In-Solution and InNanopore Digests of Liver Nucleoprotein Fraction. Capillary LC separations were carried out as previously described.26 Briefly, the separation of peptide digests was performed on a Famos/ Switchos/Ultimate chromatography system (Dionex/LC Packings) equipped with a Probot device for MALDI plate spotting. A 5-µL samples of concentrated nucleoprotein digests from either the nanoreactor or in-solution control were injected and captured onto a trap column (PepMap C18, 5 µm, 100 Å, 300 µm i.d. × 5 mm) at 30 µL/min. After switching the trap column in line with the nanoflow solvent delivery system, peptides were eluted and separated on an analytical nano-LC column (Vydac C18, 3 µm, 100 A, 75 µm i.d. × 15 cm). Mobile phase A is 0.1% formic acid and 5% acetonitrile. Mobile phase B is 0.1% formic acid and 95% acetonitrile. The gradient was kept at 5% B for 5 min, then ramped linearly from 5 to 50% B in 60 min, and then jumped to 80% B and kept for 10 min. Then the gradient was jumped back to the start point, and the column was equilibrated for 20 min. The flow rate was 200 nL/min. The nano-LC eluant was supplemented with 5 mg/mL R-cyano-4-hydroxycinnamic acid (in 50/50 ACN/water containing 0.1% TFA) from a syringe pump at a flow rate of 1 L/min and spotted directly onto the ABI 4700 576-well target plates using the Probot. MALDI Sample Preparation and Instrumentation. The protein digests of individual and a mixture of proteins were mixed with an equal volume of the matrix solution (saturated R-cyano4-hydroxycinnamic acid in 50% ACN, 0.1% TFA). A 0.5-µL sample of the mixture was spotted on the MALDI sample plate. All tryptic digests were analyzed by both MS precursor ions and MS/MS peptide sequencing on an Applied Biosystems 4700 proteomics analyzer. The acquired MS or MS/MS data were then submitted to Swissprot database for protein identification using GPS Explorer Software with mass errors less than 50 ppm. Raman Characterization of Myoglobin/FDU-12 Composite. A 10-mg sample of FDU-12 beads was mixed with 0.20 mg of myoglobin diluted in 1.6 mL of NH4HCO3 buffer (pH 6.5). The mixture was then agitated on an Eppendorf Thermomixer at 8 °C for 10 min. The MPS beads were pelleted by centrifugation (18 000 rpm) and washed with Millipore water twice. The final solid sample was air-dried at room temperature. A micro-Raman instrument (LaRam-1B, Dilor) was employed to record Raman spectra. The resonance Raman experiment was carried out at room temperature with laser excitation at 632.8 nm from a He-Ne laser. The laser power used for the measurement was ∼6 mW.

(25) Kawakami, K.; Yanagisawa, K.; Watanabe, Y.; Tominaga, S.; Nagano, K. FEBS Lett. 1993, 335, 251-254.

(26) Gu, S.; Chen, J.; Dobos, K. M.; Bradbury, E. M.; Belisle, J. T.; Chen, X. Mol. Cell. Proteomics 2003, 2, 1284-1296.

RESULTS AND DISCUSSIONS Efficiency and Reproducibility of In-Nanopore Digestion of Individual Proteins. Here MPS FDU-12 is employed as a nanoreactor for protein capture and digestion. It has a highly ordered (fcc) three-dimensional interconnected nanopore networks in which ultra-large nanopores (∼27 nm) are connected by expanded entrances (∼17 nm) (Supporting Information Figure 1). Figure 1a shows the TEM image of MPS FDU-12 along the [110] direction and more details about the structure features of

Analytical Chemistry, Vol. 78, No. 14, July 15, 2006

4813

Figure 2. MALDI MS spectra obtained from (a) in FDU-12 nanopore and (b) in solution digestion of myoglobin. Identified peptide peaks (S/N >80) and tryptic autolysis peaks are labeled with asterisks in red and black, respectively. All the identified peptide sequences are provided in Supporting Information Table 1.

FDU-12 are reported elsewhere.22,27 The proteolysis of myoglobin (MW 16 900, pI ) 7.36) by the protease (trypsin, MW 25 400, pI ) 8.23) was first carried out with and without MPS to evaluate the reaction efficiency. At pH 6.0, trypsin and myoglobin are all positively charged and were rapidly (in less than 1 min) captured into the inner surface of FDU-12 nanopores that are negatively charged. After the entrapment, no protein was detected in the supernatant with UV-visible spectrometry. It is widely accepted that when the mesopore diameter is sufficiently large for “comfortable” entrapment of biomolecules, proteins penetrate deep into mesoporous networks rather than adsorb onto the external surface.13,16,28 In our case, the pore entrances (∼17 nm) of FDU-12 are much larger than the globular diameter of proteins/enzymes (∼4 nm). Therefore, we believe that the majority of the proteases and protein substrates were completely entrapped within the mesopores. Moreover, the rapid entrapment of 10 wt % myoglobin or trypsin within FDU-12 was also observed. This rapid and high-capacity absorption is caused by in-pore entrapment instead of external site binding.22 (27) Fan, J.; Shui, W. Q.; Yang, P. Y.; Wang, X. Y.; Xu, Y. M.; Wang, H. H.; Chen, X.; Zhao, D. Y. Chem.-Eur. J. 2005, 11, 5391-5396. (28) Balkus, K. J.; Diaz, J. F. Abstr. Pap. Am. Chem. Soc. 1996, 212, 231-INOR.

4814

Analytical Chemistry, Vol. 78, No. 14, July 15, 2006

The subsequent proteolysis of the entrapped myoglobin was triggered by adjusting the buffer pH from acidic (pH 6.0) to basic (pH 8.2) condition. During the rapid buffer pH alteration, no protein was released from the mesochannels to solution as monitored by UV-visible spectrometry. MALDI-TOF mass spectrometry was employed to analyze digestion products generated from different MPS nanoreactors. Proteolysis efficiency was evaluated based on protein identification results using peptide mass fingerprinting (PMF) and the peptide peak intensity in the PMF spectra (Figure 2).29,30 Surprisingly, the mass spectrum of the product obtained from the 15-min proteolysis of myoglobin inside the nanopores of FDU-12 yielded intense peaks (S/N >80). This mass spectrum allowed for confident identification of 12 peptides (Table 1, Figure 2), with a MOWSE score of 170; e.g., a score of 61 is the identification threshold value for peptide peaks of S/N >3. Thus, highly efficient proteolysis was achieved in FDU12 nanopores. In contrast, even after a 12-h incubation of predenatured myoglobin in solution in the presence of trypsin, myoglobin was not identified by the PMF (i.e., MOWSE score of 57, below the threshold for identification). In fact, in-solution digestion generated only two peptides at a similar S/N level after (29) Park, Z. Y.; Russell, D. H. Anal. Chem. 2000, 72, 2667-2670. (30) Park, Z. Y.; Russell, D. H. Anal. Chem. 2001, 73, 2558-2564.

Table 1. Amino Acid Sequence Coveragea and MOWSE Scores for Protein Identification via Tryptic Digests digestion condition

myoglobin (%)

convalbumin (%)

in solutionc

no denaturation thermal denaturation

0b 27 (57)

0b 13 (63)

in MPS nanospaces

FDU-12 SBA-15 MCF

84 (170) 31 (65) 22 (49)

29 (169) 10 (66) 6 (48)

a The sequence coverage and score (in parentheses) are commonly used to evaluate digestion efficiency. Here we use a high S/N cutoff (>80) to identify strong peptide signals. b There is no peptide identified. c The ratio of enzyme/substrate is 1:2 (w/w).

5-h reaction and three peptides after overnight incubation. The difficulty in digesting this rigid-structure protein by in-solution enzymes has been reported in other studies.29 It is noteworthy that in our case, we applied the same enzyme/substrate ratio (E:S ) 1:2) for both in-solution and in-nanoreactor digestion. In MS measurements of in-solution digests (Figure 2b as the peptide sequences are given by Supporting Information Table 1), the autolysis signals from in-solution trypsin significantly suppressed sample peptide signals and complicated mass spectra. In contrast, as shown in Figure 2a, enzyme entrapment in nanoreactors minimized undesired trypsin autodigestion while maintaining a remarkably high proteolytic efficiencysone advantage of enzyme immobilization. Furthermore, a close comparison of the spectral intensity of a single peptide at m/z ) 748.4, generated from both types of digestions using a conventional protocol, revealed that the S/N ratio was increased 15-fold by in-nanopore digestion under the same instrumental settings (Figure 2). Although the signal intensity of MALDI is generally semiquantitative, the reproducible PMF from five pairs of in-solution and in-nanopore digestion samples, and the large difference in the observed peak intensity between in-nanoreactor and in-solution samples, strongly indicated that proteolysis within nanoreactors generated far more abundant peptides in much less time than those derived from in-solution reactions. We have examined the proteolytic activity and stability of the entrapped trypsin in FDU-12 nanopore structure. Figure 3 illustrates the reusability of the trypsin-immobilized FDU-12 nanoreactor, by doing repetitive digestion and examining the peptide recovery over 16 days using this nanoreactor (See the Experimental Section for details). The MALDI spectrum acquired over the time, as well as the inset plot of protein identification scores, indicating for proteins in a particular pI range the trypsin deposited in FDU-12 nanopores could maintain required activity and stability over a long period of time. Interestingly, we found that the digestion efficiency of the trypsin-immobilized FDU-12 reactor varied according to the pI of the protein substrates (Supporting Information Figure 2). For example, those proteins with relatively acidic pIs tended to be digested incompletely. We then found that the proteolysis of a broad range of proteins/mixtures could only be optimized by our novel design shown in Figure 1b. These results suggest that the proteins trapped inside FDU-12 cavities may undergo a significant pH-dependent unfolding, making their cleavage sites fully accessible to trypsin. Addition of fresh trypsin on the inner surface of the FDU-12 cavities could then activate

Figure 3. MALDI spectrum obtained from the digestion of myoglobin in FDU-12 nanopore. The asterisks indicate identified peptide peaks in the 16th test. Inset plot is MOWSE score of myoglobin identification as a function of time.

this nanopore-based microreactor for complete proteolysis in a pH-dependent manner. Note that actual pIs of proteins may significantly change due to posttranslational modifications. However, in those cases, the actual pIs of ovalbumin and conalbumin would be even lower after modification with phosphate or sialic acid with respect to those of theoritic calculating values. In fact, these shifts of pI values do not change the conclusion that acidic pIs generally lead to lower digestion efficiencies in the nanopores. In-Nanopore Digestion of Mixtures of Proteins with Diverse Physical Properties. Our nanopore-based proteolytic reactor is effective not only for individual protein but also for protein mixtures. At pH 6.0, a three-protein mixture with a broadrange diversity in MW, pI, and abundance (Table 2) was incubated with FDU-12 beads. After the entrapment, no protein was detected in the supernatant with UV-visible spectrometry. Trypsin was then added into the substrate-enriched nanopores and in situ proteolysis was immediately triggered by changing the buffer pH from 6.0 to 8.5. and the reaction was conducted for 15 min at 37 °C. For comparison, the in-solution digestion for the mixture with the same protein composition was processed up to 12 h to reach the reaction equilibrium and to give the best yield of peptides. Comparing the MALDI spectra obtained from in-nanopore and in-solution digestion, much more peptides at a 10-fold increase in peak intensity were present in the PMF from the 15-min innanopore digestion (Figure 4a as the peptide sequences are given in Supporting Information Table 2) than the peptide peaks from the 5-h in-solution digestion (Figure 4b). It should be noted that in order to minimize trypsin autolysis signals, a smaller amount of enzyme was added into the solution at E:S ) 1:40. Thermal denaturation was also performed in the control to unfold proteins for exposing their cleavage sites to trypsin.29 However, neither of these optimizations for currently available protocols of in-solution digestion showed the comparable efficiency as that displayed by in-nanopore digestion. Efficiency and Sequence Coverage of In-Nanopore Digestion of Nuclear Fraction Proteins. To examine the proteolytic efficiency of our MPS-based naoreactors for real-case samples of biological complexity, we compared the proteolytic profile of innanopore versus in-solution digestion on a nuclear protein fraction Analytical Chemistry, Vol. 78, No. 14, July 15, 2006

4815

Table 2. MS Identification Results of a Protein Mixture in Different Digestion Systems in solutione mixture componenta

component absolute quantity (µg)

component concnb (nM)

MALDI spottingc (fmol)

in FDU-12 nanospaces d

thermal denaturation

without denaturation

ovalbumin conalbumin cytochrome c

45 15 1

527 97 43

262 48 21

++ ++ ++

++ + -

-

a Ovalbumin (42.7 kDa, pI 5.2); conalbumin (77.7 kDa, pI 6.85); cytochrome c (11.7 kDa, pI 10.4). b FDU-12 solids were incubated with 2 mL of protein solution for adsorption. c The amount of protein digest spotted on MALDI plate was calculated on the assumption that all peptides were released from nanospaces after digestion. d S:E ) 2:1 (w/w) “++” indicates unambiguous identification through PMF and database search; “+” indicates discovery of a few poor peptide signals yet no confident identification; “-” indicates no peptide signals. e S:E ) 40:1 (w/w), as commonly used for in-solution digestion to minimize trypsin autolysis.

isolated from the mouse liver cells. To evaluate the multiple strengths such as protein denaturation of the nanoreactor for “crude” protein mixtures, in-solution digestion was performed on the native protein extract the same way as does in the nanopore reactor, i.e., without predenaturation. The proteolytic digests resulting from both types of digestions of 5 µg of nuclear proteins were subjected to microcapillary (µ)LC separation followed by MALDI TOF-TOF analyses. In Figure 5, the base peak chromatograph showed far more peptide signals resulted from in-nanopore digestion (Figure 5b) than from in-solution digestion of the same sample (Figure 5a). Using MS/MS peptide sequencing with stringent criteria for protein identification,31 a total of 98 proteins were unambiguously identified from the in-nanopore tryptic profile compared to only 6 proteins confidently identified from the insolution digestion of the native isolate (Supporting Information Tables 3 and 4). For the proteins identified in both cases, the sequence coverage of the detected peptides for the corresponding

Figure 4. MALDI spectra obtained from the digestion of a protein mixture, (a) In FDU-12 nanopores (15 min) and (b) in solution (5 h). The identified peptide peaks are annotated to the corresponding proteins. Their sequences and database search results for the spectrum in (a) are provided in Supporting Information Table 2. 4816

Analytical Chemistry, Vol. 78, No. 14, July 15, 2006

individual proteins was significantly enhanced via in-nanopore digestion. The failure of conventional digestion of the protein extract at low concentration (0.1 µg/µL in total) is partially attributed to the inherent kinetic limitation of most proteases in solution5 (Km values on the order of 5-50 mM), while in our case, most protein components are at submicromolar or nanomolar level. Other important factors that reduce the activity of trypsin for the in-solution digestion here might be the leftover detergents and trypsin inhibitor in the lysis buffer (e.g., urea, CHAPS, and PMSF), which are used to maintain protein stability and solubility during the sample preparation.32 Importantly, the new protocol designed for our MPS-based nanoporereactor involves a buffer exchange step before enzyme entrapment and interaction with substrates. Therefore, in addition to its enriching and unfolding effects, the nanoreactor allows for removal of any impurities to maximize tryptic activity on the fully accessible proteins, which further improves the sensitivity and resolution of MS analysis. Exploration of Mechanistic Aspects of In-Nanopore Digestion. For all the digestions of either an individual protein, a protein mixture, or a real biological complex sample, the reaction in the nanopores of FDU-12 is much more efficient than that occurring in solution according to signals of peptide peaks in PMF, protein identification results, and digestion time variation. It is known that enzymatic reaction rate is dependent on various factors, including concentrations of the enzyme and substrate, pH, temperature, and presence of inhibitors or activators. In our cases, all these factors remained constant for both in-solution and innanopore proteolysis. It allows us focus on distinct features of MPS as the proteolytic nanoreactors compared with in-solution reaction or other microreactor systems.27 These effects, such as substrate enrichment (resulting in a 2 order of magnitude enhancement of reaction rate), MPS surface chemistry, and MPS morphology, have been reported elsewhere.27 Here we mainly discuss our interesting finding of substrate unfolding, which can substantially facilitate proteolysis in MPS-based nanoreactors, as well as other factors that play an important role in proteolysis reaction. (1) “Reagent-Free” Protein Unfolding within MPS. Myoglobin and conalbumin are two proteins known to be resistant to proteolysis, given that most of their hydrophobic domains are (31) Gu, S.; Liu, Z. H.; Pan, S. Q.; Jiang, Z. Y.; Lu, H. M.; Amit, O.; Bradbury, E. M.; Hu, C. A. A.; Chen, X. Mol. Cell. Proteomics 2004, 3, 998-1008. (32) Yu, Y. Q.; Gilar, M.; Lee, P. J.; Bouvier, E. S. P.; Gebler, J. C. Anal. Chem. 2003, 75, 6023-6028.

Figure 5. Base peak chromograph of (a) in-solution digestion and (b) in-nanopore digestion of the nuclear fraction of mouse liver tissue. The protein identification data are included in Supporting Information Tables 3 and 4.

deeply buried and difficult to be accessed by proteases.29 Therefore, thermal or chemical denaturation is usually applied before in-solution digestion to unfold their rigid globular structures for exposing their cleavage site. Whereas, successful proteolysis within MPS nanopores without predenaturation suggests that the proteins trapped inside FDU-12 nanopores may undergo a significant structure transition, making their cleavage sites fully accessible to trypsin. To test this hypothesis, resonance Raman spectroscopy was used to monitor possible conformational change of the protein substrates accommodated in FDU-12 nanopores as a comparison with that of native or denatured proteins in solution (Figure 6). The high-frequency region of the Raman spectrum of myoglobin is sensitive to the spin, coordination, and oxidation state of the heme iron.33 In our study, the marker bands at 1120, 1209, 1541, and 1607 cm-1 changed following the myoglobin entrapment within FDU-12 nanopores, which were upshifted by ∼5-15 cm-1 compared with those of native myoglobin in solution. This peak shifting is also observed in other unfolding processes.34 Further(33) Hu, S. Z.; Smith, K. M.; Spiro, T. G. J. Am. Chem. Soc. 1996, 118, 1263812646. (34) Tang, Q.; Kalsbeck, W. A.; Olson, J. S.; Bocian, D. F. Biochemistry 1998, 37, 7047-7056.

more, Raman spectrum of the myoglobin entrapped within MPS is very similar to that after a thermal-induced unfolding, suggesting that myoglobin in MPS has structural features very close to that after thermal denaturation. Although the information from this Raman study was not conclusive to picture the exact way myoglobin reorganizes its structure within MPS cavities, it indicated that myoglobin changes their conformation after the entrapment. In fact, our observation is in line with a previous report that the native myoglobin conformation was strongly disturbed when it was adsorbed onto silica particles, leading to a more heterogeneous structure with partial unfoldings.35 A mechanism was then postulated to explain this reagent-free protein unfolding process in the nanopores of MPS. Protein substrates were confined in the 3D nanopores containing abundant negatively charged silanol groups. The interaction between silanols and those basic amino acid residues at the protein surface can initiate an unfolding of the protein structures. This process may be further facilitated by the formation of hydrogen bonds between the silanol walls of FDU-12 and interior sites of the proteins, which (35) Buijs, J.; Ramstrom, M.; Danfelter, M.; Larsericsdotter, H.; Hakansson, P.; Oscarsson, S. J. Colloid Interface Sci. 2003, 263, 441-448.

Analytical Chemistry, Vol. 78, No. 14, July 15, 2006

4817

due to the advantages in the mass diffusion and transportation of 3D interconnected nanopores of FDU-12 compared with the SBA15 structure. However, with disordered mesocellar foams (MCFs)37 that have large cavities (27 nm) and a wide entrance (15 nm) similar to those of FDU-12 nanopores, only a low sequence coverage at 22% was obtained. This result implies that FDU-12 contributed to the enhanced proteolytic efficiency through its wellorganized cavities that provided a more suitable microenvironment for protein denaturation and concurrent proteolysis.

Figure 6. Resonance Raman spectra of myoglobin under different conditions: (a) in FDU-12 nanopore arrays, (b) after thermal denaturation, (c) in buffer, and (d) solid sample.

could disrupt the forces maintaining original conformation and lead to in situ protein unfolding. (2) Entrapment Effect. Since reagent-free structural unfolding is associated with in-nanopore digestion, the efficient entrapment of protein substrates within MPS is the key for a relatively complete digestion. For example, a poor in-nanopore digestion was observed for conalbumin when the protein was loaded at the high pH of 8.5. UV spectrum and gel electrophoresis (data not shown) revealed that most of the conalbumin remained in the supernatant. The low-level adsorption of conalbumin in the nanopore is due to its relatively low pI (6.85). In the buffer at pH 8.5, both the protein and silicate nanopore surface carry negative net charges. This electric repulsion prevents protein molecules from entering nanopores of MPS and undergoing subsequent structural unfolding and proteolysis. This phenomenon becomes even more pronounced when a protein mixture is digested. Thus, a slightly acidic buffer (e.g., pH 6.0) is applied in our experiments to ensure all proteins in the mixture are completely entrapped within the nanopores. (3) MPS Structural Effect. We also tested two other typical MPSs with different large-pore mesoporous structures for the efficiency of myogloin digestion. In the case of MPS SBA-15 (2D hexagonal packing ∼8-nm mesochannels) 36 as the host to protein substrates, the peptide peaks detected at high S/N (>80) only covered 31% of the protein sequence as compared with the 84% coverage at the same S/N threshold using the FDU-12 nanopores (Table 1). This improvement of proteolysis efficiency might be (36) Zhao, D. Y.; Feng, J. L.; Huo, Q. S.; Melosh, N.; Fredrickson, G. H.; Chmelka, B. F.; Stucky, G. D. Science 1998, 279, 548-552.

4818

Analytical Chemistry, Vol. 78, No. 14, July 15, 2006

CONCLUSION There are several unique features distinguishing our nanoporebased proteolytic reactors from other currently available microdevices. The distinctive nanoscale confinement of FDU-12 offers multiple functions that could overcome some of the weakness of the conventional digestion system. For instance, most of microdevices use the E/S (w/w) generally in the range of 5:1-100: 1,5,38,39 and even higher than 4000:1 for on-line digestion in a capillary,7 while in-solution digestion usually uses E/S in the range of 1:20-1:40.30 The rapid digestion and reasonable protein identification 7,39 based on these microreactors is largely compromised by complicated fabrication. Besides, rapid proteolysis by a few microreactors tended to be incomplete as many miscleaved large peptides were observed.2 In our case, E/S was at 1:2 for the digestion of standard proteins and at 1:8 for the digestion of the nuclear proteome. The comparisons between in-nanopore and in-solution digestion of standard proteins were made under the same E/S. Thus, our rapid digestion, high sequence recovery with fewer miscleavages, as well as low consumption of peptide samples in MS analysis, suggested that the significant enhancement of proteolytic efficiency was due to the intrinsic properties of the 3D well-ordered nanoporous structure. Furthermore, enzyme immobilizations via most available methods are based on time-consuming multistep reactions. Our one-step adsorption of protein substrates, denaturation, and concurrent trypsin flashing takes less than 15 min due to the ultrahigh surface area and favorable interactions of FDU-12 nanopores with proteins. In particular, our nanopores of uniformly biocompatible sizes provide a highly permeable support for both enzyme and proteins, whereas sol-gel matrixes commonly used for hydrolysis of small peptides have widely varied pore sizes and they are rather difficult for protein to penetrate, thus hindering digestion of larger substrates.7,8 Therefore, most of these designs have been tested only on a limited number of selected substrates rather than a wide range of proteins present in biological samples. Our findings also suggest this novel nanoporous materials provide both an enzyme-friendly and a mass spectrometrycompatible environment for real-case proteomic practices. Because of substrate enrichment effect,27 the high efficiency of both proteolysis in nanopores and peptide release afterward, our design showed high sensitivity by detecting protein digests at low femtomoles using MALDI-MS/MS. In comparison, for most of the current microreactors, peptide samples loaded onto MALDIMS or injected into ESI-MS were in the range of high femtomoles (37) Schmidt-Winkel, P.; Lukens, W. W.; Zhao, D. Y.; Yang, P. D.; Chmelka, B. F.; Stucky, G. D. J. Am. Chem. Soc. 1999, 121, 254-255. (38) Sakai-Kato, K.; Kato, M.; Toyo’oka, T. Anal. Chem. 2003, 75, 388-393. (39) Qu, H. Y.; Wang, H. T.; Huang, Y.; Zhong, W.; Lu, H. J.; Kong, J. L.; Yang, P. Y.; Liu, B. H. Anal. Chem. 2004, 76, 6426-6433.

to picomoles.2,3,39 Conventional in-solution digestion efficiency is intrinsically hindered as the concentration of protein is reduced to below micromolar levels.6 Hence the detection sensitivity is much enhanced in our new nanoreactors. In short, our nanoreactor design represents a new step toward exploring the “multifunctional” capabilities of unique nanoporous materials for sensitive and comprehensive analyses of cellular protein mixtures. Our study has provided an interesting clue to understand the largely unknown interactions of biomolecules and nanoporous confinement. In a broad sense, this design of nanoproteolytic reactor lays a foundation for developing efficient, biocompatible, and multifunctional microdevices for large-scale protein characterization.

Natural Science Foundation of China (Grants 20328508, 20233030), National HighTec Research Developing Programme (02BAC11A11), and Shanghai Sci.&Tech. Research Program (03DZ14024) for financial support. We thank Dr. Sheng Gu of Los Alamos for his assistance in running the LC profile of the digest of nuclear protein mixture. We also thank Dr. X. Y. Wang and Dr. H. L. Wu for helpful discussion in enzyme immobilization and thank Dr. W. Zhou for helpful discussion of TEM measurements.

ACKNOWLEDGMENT We thank Dr. E. Morton Bradbury for his critical comments on the manuscript. We thank the National Basic Research Priorities Programme (01CB510202 and 2000077500), National

Received for review January 16, 2006. Accepted April 28, 2006.

SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

AC060116Z

Analytical Chemistry, Vol. 78, No. 14, July 15, 2006

4819