当前位置:首页 >> >>

High-throughput expression, purification, and characterization of recombinant Caenorhabditis elegans


BBRC
Biochemical and Biophysical Research Communications 307 (2003) 928–934 www.elsevier.com/locate/ybbrc

High-throughput expression, purication, and characterization of recombinant Caenorhabditis elegans proteins
Raymond Y. Huang,a Simon J. Boulton,b,1 Marc Vidal,b Steve C. Almo,c Anne R. Bresnick,c and Mark R. Chancea,c,*
a

Center for Synchrotron Biosciences and Department of Physiology and Biophysics, Albert Einstein College of Medicine, Bronx, NY 10461, USA b Dana-Farber Cancer Institute and Department of Genetics, Harvard Medical School, Boston, MA 02115, USA c Department of Biochemistry, Albert Einstein College of Medicine, Bronx, New York 10461, USA Received 23 June 2003

Abstract Modern proteomics approaches include techniques to examine the expression, localization, modications, and complex formation of proteins in cells. In order to address issues of protein function in vitro using classical biochemical and biophysical approaches, high-throughput methods of cloning the appropriate reading frames, and expressing and purifying proteins eciently are an important goal of modern proteomics approaches. This process becomes more dicult as functional proteomics eorts focus on the proteins from higher organisms, since issues of correctly identifying intron–exon boundaries and eciently expressing and solubilizing the (often) multi-domain proteins from higher eukaryotes are challenging. Recently, 12,000 open-reading-frame (ORF) sequences from Caenorhabditis elegans have become available for functional proteomics studies [Nat. Gen. 34 (2003) 35]. We have implemented a high-throughput screening procedure to express, purify, and analyze by mass spectrometry hexa-histidine-tagged C. elegans ORFs in Escherichia coli using metal anity ZipTips. We nd that over 65% of the expressed proteins are of the correct mass as analyzed by matrix-assisted laser desorption MS. Many of the remaining proteins indicated to be “incorrect” can be explained by high-throughput cloning or genome database annotation errors. This provides a general understanding of the expected error rates in such high-throughput cloning projects. The ZipTip puried proteins can be further analyzed under both native and denaturing conditions for functional proteomics eorts. 2003 Elsevier Inc. All rights reserved.

Gene function is determined by the expression, localization, modications, and activities of the encoded protein, including its interactions with other gene products. Prior to the genomic era, the characterization of gene function has been limited to the study of one or a few genes at a time. The availability of genome sequences for an increasing number of organisms allows scientists to map gene functions using parallel, largescale approaches [1–6]. Classically, biological chemists have characterized proteins by in vitro methods to study their biochemical, biophysical, and structural properties. This approach has been carried out one protein at a
* Corresponding author. Fax: 1-718-430-8587. E-mail address: mrc@aecom.yu.edu (M.R. Chance). 1 Present address: DNA Damage Response Lab, Cancer Research UK, London Research Institute, Clare Hall, UK.

time. With the expanding sequence information currently available, thousands of genes can be identied and cloned with the correct reading frame and, in the case of higher organisms, with the exons correctly joined. A number of gene cloning projects are currently underway in multiple organisms including prokaryotes [7], yeast [8], plant (Arabidopsis thaliana, Institut National de la Recherche Agronomique, France), worm [6], y (Berkeley Drosophila Genome Project, CA, USA), mouse (Riken, Japan), and human (FLEXGene, MA). To keep pace, in vitro methods of biophysical and biochemical characterization must also adopt highthroughput approaches. The Caenorhabditis elegans ORFeome project has reported a successful initial cloning of over 12,000 open reading frame (ORF) sequences [6]. These cloned ORFs are available as Gateway (Invitrogen) “Entry” vectors

0006-291X/03/$ - see front matter 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0006-291X(03)01265-8

R.Y. Huang et al. / Biochemical and Biophysical Research Communications 307 (2003) 928–934

929

Fig. 1. Schematic diagram of ZipTip/MALDI-MS method. The cloning, gene transformation, cell harvesting, ZipTip protein purication, and MALDI-MS data collection were all performed in a 96-well format (details described in Materials and methods).

that can be readily transferred via recombination to various expression systems [9]. These ORFs and those that will arise from the myriad of cloning projects require functional and structural characterization. Cloning and expression testing of open reading frame sequences is typically conrmed by gel analyses, which are often inaccurate and time-consuming when a large number of genes need to be screened [10]. Highthroughput techniques for screening protein expression and solubility have been developed, such as the ELISAbased assays using anti-His-tag antibodies to determine the presence of proteins in the soluble fraction [11]. Although more rapid than gel-analysis, this method does not measure protein mass so incorrect expressors and signals resulting from non-specic antibody binding are false positives. Mass spectrometry is increasingly used in high-throughput identication of proteins and protein complexes [1,4,12,13]. For example, matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) can detect picomole to femtomole quantities of protein within a few seconds [14]. In addition to high sensitivity, the accuracy of MALDI-MS for protein mass identication in the 10–100 kDa range is <0.5%, allowing the detection of deviations from the predicted mass due to post-translational modication or sequence variations [14]. To explore the rapid characterization of the C. elegans ORFs, we developed a high-throughput protein expression and solubility screening protocol that incorporates a column purication step in 96-well format prior to MALDI-MS analysis. A set of C. elegans genes from the ORFeome project [6] are expressed as His-tag

Fig. 2. Result of ZipTip/MALDI-MS expression and solubility testing for the 86 C. elegans ORFS. The 96 square grid represents the conguration of the actual 96-well plate of ORFs. The color indicates the various screening outcomes. Ten of the wells (H3–H12) contain no plasmid and are designated as “Empty.” The two mass spectra at the bottom are obtained from two of the 86 clones demonstrating a high signal-to-noise ratio. In these examples, the singly charged species are detected at $48 and 72 kDa, respectively, and the doubly charged species were also observed.

930

R.Y. Huang et al. / Biochemical and Biophysical Research Communications 307 (2003) 928–934

fusion proteins in Escherichia coli, puried using metal anity tips, and examined by MALDI-MS. We show that these high-throughput MALDI methods can be used to verify the identity of expressed genes and correct database annotation errors. We further demonstrate that the puried proteins bound to the tips can be eciently digested on-tip with proteases and the peptide products can be analyzed.
Table 1 Comparisons of SDS–PAGE and MALDI measurements Gene T03F1.3 T22E5.5 K12C11.2 F54B11.5 C47E8.5 F25H5.4 F13G3.4 Y15E3A.1 K04C2.4
a b

Materials and methods
Expression constructs, cell culture, and lysis. The open reading frame sequences tested in this study were identied as DNA damage response (DDR) pathway interacting proteins in C. elegans that were mostly hypothetical and uncharacterized [2]. The 86 selected ORFs, which had sizes of the encoded proteins ranging from 10 to 110 kDa, were subcloned into the GATEWAY (Invitrogen) Entry vectors and subsequently transferred to the Destination vector (pDEST17) con-

Predicteda 44,113 47,041 10,222 23,804 80,283 94,796 11,913 73,666 73,312

SDS–PAGE (kDa) 48 61 12 25 80 15 52 58 79

MALDI-MS 48,206 50,872 14,032 28,018 12,261 18,973 50,666 57,233 17,143

Sequencingb 48,063 50,586 14,145 27,876 84,206 18,904 49,109 57,207 17,081

Comment Correct massc Correct mass Correct mass Correct mass Non-specic binding Actual gene B0336.10 Actual gene F45F2.11 Alternative splicing Inversion, early stop

Mass calculated based on the ORF sequence from www.wormbase.org N- and C-terminus sequence ($4 kb) were not included in the calculate. Sequenced mass is calculated from actual nucleotide sequence as conrmed by gene sequencing. c Correct mass is designated when the error of MALDI-MS measurement is within 0.5% of the mass calculated from the DNA sequence or within 1% of the expected mass plus the 4 kb N- and C-terminus sequence.

Fig. 3. MALDI-MS analysis of on-tip trypsin digested peptides. (A) Schematic diagram of tryptic digestion while proteins are bound to the ZipTips. The procedure consists of equilibrization, binding, washing, proteolytic digestion, and elution. The eluted peptides were analyzed by MALDI-MS as described in Materials and methods. (B) A MALDI-MS spectrum of expressed and digested TO3F1.3 (48 kDa). (C) Digestion coverage of TO3F1.3: more than 95% of the protein was covered by peptides from the tryptic digest. The ladder of purple lines indicates peptides detected by MALDI-MS that match with the predicted sites with trypsin digestion. The red segment of the bottom line indicates regions of TO3F1.3 protein covered by the peptides, which represent approximately 95% of the full protein.

R.Y. Huang et al. / Biochemical and Biophysical Research Communications 307 (2003) 928–934 taining N-terminal 6 Histidine-tag [9]. The nucleotide gene sequences of the ORFs were veried. The 96-well protocol for cell transformation, culture, and harvesting were developed previously [10]. Briey, expression target vectors were transformed into a 96-well block of E. coli competent cells (BL21-Star-pLyss strain, Invitrogen, Carlsbad, California) prepared using standard CaCl2 protocol. A liquid handling system (Biomek FX, Beckman) was used to automate most steps involving 96-well processing (Fig. 1). The transformed cells were cultured in 2 ml well containing 1.6 ml Luria–Bertani (LB) media in a 96-well block and allowed to grow at 37 °C and 250 rpm to OD 0.6 before 0.5 mM IPTG was added to each well. After the IPTG induction, the cells were shaken at 225 rpm for an additional 3 h followed by centrifugation at 2500g. The resulting cell pellets were lysed in 80 ll/well of lysis buer containing 75 ll BugBuster (Novagen, Madison, Wisconsin), 0.2% Benzonase (Novagen), 2% Protease Inhibitor Cocktail III (Novagen), 1 mM b-mercaptoethanol, and 500 mM NaCl. After cell lysis, the 96-well block was again centrifuged at 2500g and the resulting supernatant was isolated. The pellet remaining in each well was resuspended and solubilized in 8 M urea (pH 7.0, 20 mM Tris–HCl, and 20 mM NaH2 PO4 ). The supernatant and the urea-solubilized pellet were separately analyzed by SDS– PAGE using molecular weight standards. We dened proteins as soluble when they were observed in the supernatant fraction and as insoluble when they were not detected in the supernatant but in the pellet fraction. Anity tip protein purication. The cell lysates (both the soluble and the insoluble fractions) prepared in 96-well plates were aspirated into pre-equilibrated (10 mM Tris, pH 7.0) ZipTips (Millipore) containing a metal anity resin, washed twice with 10 mM Tris, pH 7, and twice with 30% methanol, and eluted with the manufacturer supplied buer (Fig. 1). As much as 0.5 ll of the eluted solution was spotted onto a MALDI-MS target plate, mixed with 0.5 ll of saturated sinapinic acid in 50% acetonitrile and 0.1% triuoroacetic acid. MALDI mass spectrometry. The MALDI-MS experiment was performed on a Voyager-DE STR Biospectrometry Workstation (Applied Biosystems, Foster City, CA). The mass/charge scale was externally calibrated using cytochrome-c and albumin for proteins and bradykinin for peptides. Fifty spectra were collected and averaged for each sample. Automated data collection was set up using Voyager’s data collection sequence option. On-tip proteolysis and MALDI-MS peptide analyses. After the protein binding and washing steps of the ZipTip purication protocol, 5 ll of 4 lg/ml trypsin (10 mM Tris, pH7) was aspirated into the protein bound ZipTips and incubated at 37 °C for at least 4 h. The evaporation of digestion solution is compensated by immersing the ZipTip into an additional 5 ll of equal concentration of trypsin solution. Following the digestion period, 0.5 ll of the solution from the ZipTip was spotted onto a MALDI target plate and mixed with the matrix solution for peptide analysis. The peptides on the MALDI-MS spectrum were compared to the predicted proteolytic products using PAWS (Proteometrics, New York).

931

minal extra sequences) by less than 1%. The accuracy of the MALDI-MS measurements was determined by comparing to the mass calculated from the actual DNA sequences (Table 1). To unambiguously conrm that the correct proteins were expressed, we digested selected ZipTip-bound proteins with trypsin and analyzed the resulting peptides by MALDI-MS (Fig. 3). In the example demonstrated here, the observed peptides covered >95% of the protein sequence. The peptide mass list can be used to identify the expressed protein using peptide mapping search engines (ProFound: http://prowl.rockefeller.edu/ PROWL/prowl.html). Incorrect size expressions The number of genes with “incorrect size” detected by MALDI-MS is 19%. To determine the reason for the observed size discrepancy in these “incorrect” clones, we conrmed the expression using SDS–PAGE and examined the DNA sequence for a subset of these genes in order to verify the identity of the clones (Table 1). A number of potential explanations for the false positive results must be considered. First, endogenous E. coli proteins can remain bound to the anity ZipTips, resulting in false-positive signals. In these cases, the DNA sequences were correct and the proteins showed the correct size using SDS–PAGE, but were incorrectly identied in the MALDI-MS spectra. This problem was more commonly observed early in the development phase when the washing condition was not optimized; the protein masses of these false-positives are seen to be present in the spectra from multiple genes and thus are clear examples of non-specic binding. A second potential explanation for the observation of incorrect size may be due to cloning or annotation error. For example, the sequences of two genes (F25H5.4 and F13G3.4) were found to be inadvertently dierent C. elegans genes (B0336.10 and F45F2.11, respectively), which could be due to either errors in the initial annotation or non-specic priming in the PCRs from cDNA. In another case, gene K04C2.4 was found to have an inversion in part of the sequence causing early termination in translation and the resulting MALDI-MS measurement reected this change (Table 1). These examples demonstrate that the MALDI-MS method can be used to assist in verifying the delity of the cloning step for high-throughput cloning projects. A third explanation for incorrect size may be due to the reported 29% discrepancy existing between the exon–intron structure of cloned ORFs (using ORF sequence tag analysis (OST) and that based on the GeneFinder predictions [6]. For example, OST analysis found that gene Y15E3A.1 has a dierent structure compared to that predicted (Table 1). MALDI-MS showed a peak at 57,233 for this gene, dierent from the predicted mass (73,666 Da) and we assigned it initially as

Results Protein expression and solubility screening The results of expressing 86 putative C. elegans DDR genes using the ZipTip/MALDI-MS method show that $62% of the genes were expressed and 5.4% of these were also soluble in the solvent tested (Fig. 2). The expressed genes were assigned as “correct mass” when the MALDIMS measurements dier from the estimated mass based on the sequence prediction (gene plus 4 kb N- and C-ter-

932

R.Y. Huang et al. / Biochemical and Biophysical Research Communications 307 (2003) 928–934

“incorrect mass.” After the OST correction, the mass predicted from the corrected sequence was 57,207 mass units, agreeing with the measured mass. Analysis of non-expressors The existence of false negatives using the ZipTip/ MALDI was considered and is as follows: (1) Candidate proteins may not bind to the ZipTips due to the N-terminal sequence not being available (see below). This could be addressed by examining both N- and C-terminal fusions. (2) Proteins may be expressed at too low quantities to be detected, based on the MALDI detection limits and the amounts of culture examined this denes the minimum protein required for detection. (3) Proteins were not soluble in the denaturing or elution buer(s). For example, one of the tested genes, K12D12.5 (well position A2), showed a band of correct size on the SDS–PAGE gel of the whole cell lysate, but no signal was detected by MALDI-MS. We studied this protein by scaling up the purication using Ni–NTA column (Invitrogen) and found that the protein did not bind to the column even in the denaturing condition (data not shown). The presence of the His-tag in the Nterminus was conrmed by DNA sequencing and there were no mutations causing an early stop or frame-shift. It is possible that the N-terminal segment of the protein was cleaved in the cell before binding to the column, or that it is buried within a protein segment resistant to the denaturing condition. In the latter case, we may be able to increase the anity of the protein to the ZipTip by placing the His-tag at the C-terminus. In another example, we found that the expressed protein C01G5.6 (well position D2) could not be puried by the ZipTip because it is not soluble in 8 M urea and was perhaps trapped inside the ZipTip and not eluted.

Discussion When eukaryotic genes are expressed in prokaryotic organisms such as E. coli, they may not fold properly and may form aggregates (in the form of inclusion bodies) due to the absence of appropriate post-translational chaperones or processing [15]. However, the manipulability, short growth time, and low cost render E. coli the most widely used expression system for recombinant proteins. Therefore, many of the structural genomics projects have focused on searching for genes that express and fold correctly in E. coli so that they can be rapidly puried and crystallized [10]. In this study we screened the expression and solubility of 86 His-tagged C. elegans proteins in E. coli using metal anity ZipTips and MALDI-MS. The procedures are fully automated and were completed within 2 h, much less than is required for gel analyses. In theory, the intrinsic high

sensitivity of MALDI-MS, that is, its ability to detect proteins at picomole quantities should enable the ZipTip/MALDI-MS method to identify low expressors better than other conventional methods. In practice, however, there are several technical issues that need to be improved, particularly proteins not binding to the anity column. Furthermore, the results of screening using the ZipTip/MALDI-MS method can conrm the correctness of the initial cloning. For example, we estimate that about 80% of the genes identied as “incorrect size” could be attributed to errors in cloning or annotation, and only the remaining were perhaps due to non-specic binding or other technical limitations associated with ZipTip purication. Thus, the accuracy of the ZipTip MALDI method in detecting the actual protein expressed is from 90% to 95%. The expressed protein bound to the ZipTips can be characterized using dierent biochemical approaches. First, the ability to perform proteolytic digestion on-tip allows for the unambiguous identication of the expressed proteins. The trypsin digested peptide mass can be compared to the list expected from the protein sequences. In particular, proteolytic peptide analyses allow detection of proteins with large masses that cannot be easily measured by MALDI-MS. In addition to conrming protein expression, it is possible to detect alternatively spliced forms of the expressed gene by examining the observed tryptic digest. Second, the ZipTip/MALDI-MS method can rapidly test the solubility of proteins in dierent buers. Although only 3.4% of the genes examined were shown soluble by this method (Fig. 2), these selected genes were not predicted to be soluble before screening, and many of the targets could be membrane associated or hydrophobic. The rapidity with which the ZipTip/MALDI-MS method can be carried out allows the application of screening methods to enhance expression, lysis eciency, and solubility of expressed proteins. Such screening methods have indicated that the ideal conditions for protein solubility and expression are very much dependent on the individual protein [16]. Also, for the proteins puried under denaturing conditions of high urea, on-tip refolding could be attempted through systematic variation of urea concentrations. Since the refolding of denatured proteins involves kinetic competition between refolding and aggregation, having the proteins bound to the tip during refolding would decrease the aggregation rate enhancing the possibility of refolding [17]. The ZipTip technique can be further developed to capture other fusion proteins that are readily constructed by the Gateway system. It has been shown previously that both glutathione-S-transferase (GST) [18] and maltose binding protein (MBP) [19,20] enhance the solubility of the fused proteins. For example, among the MBP fusions of the C. elegans ORFs studied, 24%

R.Y. Huang et al. / Biochemical and Biophysical Research Communications 307 (2003) 928–934

933

are soluble as compared to 5% of their His-tagged counterparts [10]. Therefore, the development of GSTand MBP-ZipTip methods should increase the rate of success in obtaining soluble proteins. Third, we can obtain structural information about the puried proteins using recently developed hydroxyl radical based surface mapping (footprinting) techniques. Surface accessible residues are modied by hydroxyl radicals produced by radiolysis [21,22] or Fenton reagents [23]. The oxidized side chains can by analyzed by quantitative LC-MS or MALDI to determine the extent of oxidation and by tandem methods to determine the oxidation site. This method has been applied to analyze protein structure and dynamics [21,22,24] protein unfolding [25], and protein–ligand interactions [26–28]. Since expressed proteins can be presented on ZipTips, protein structure changes at the protein surface induced by ligand binding can be explored in a high-throughput fashion by a combination of these approaches. This will allow predicted protein-protein interactions to be directly tested as well as provide information on their interacting surfaces [1–4]. In conclusion, high-throughput structural and functional proteomics approaches to accompany highthroughout cloning projects for higher organisms are essential steps to validating such proteome resources such that biochemical and biophysical experiments on the proteins can move ahead. Many protein chip experiments that rely on antibody identication of the fusion construct (e.g., anti-GST) [29] may unduly suer from false negatives due to incorrectly identifying proteins as expressed but non-functional when the incorrect protein has been expressed. Using MALDI approaches, we can tentatively validate over 65% of the constructs that exhibit expressed proteins from the C. elegans clones we have examined. This provides a general understanding of the expected error rates in such highthroughput cloning projects. Acknowledgments
The authors thank Tim Blankenship and Elena Chernokalskaya from Millipore for their technical assistance in developing the ZipTip protein purication and Edward Nieves for his helpful suggestions on MALDI-MS methods. This research is supported in part by The Protein Structure Initiative P50-GM-62529, P41-EB-01979, and R33CA-83179. The MALDI-MS experiments were performed in The Laboratory for Macromolecular Analysis and Proteomics (LMAP) at the Albert Einstein College of Medicine, which is supported in part by the Albert Einstein Comprehensive Cancer Center (CA13330) and the Diabetes Research and Training Center (DK20541).

References
[1] Y. Ho, A. Gruhler, A. Heilbut, et al., Systematic identication of protein complexes in Saccharomyces cerevisiae by mass spectrometry, Nature 415 (2002) 180–183.

[2] S.J. Boulton, A. Gartner, J. Reboul, et al., Combined functional genomic maps of the C. elegans DNA damage response, Science 295 (2002) 127–131. [3] J.F. Gera, T.R. Hazbun, S. Fields, Array-based methods for identifying protein–protein and protein–nucleic acid interactions, Methods Enzymol. 350 (2002) 499–512. [4] Y. Zhao, T.W. Muir, S.B. Kent, et al., Mapping protein–protein interactions by anity-directed mass spectrometry, Proc. Natl. Acad. Sci. USA 93 (1996) 4020–4024. [5] A.C. Gavin, M. Bosche, R. Krause, et al., Functional organization of the yeast proteome by systematic analysis of protein complexes, Nature 415 (2002) 141–147. [6] J. Reboul, P. Vaglio, J.F. Rual, et al., C. elegans ORFeome version 1.1: experimental verication of the genome annotation and resource for proteome-scale protein expression, Nat. Genet. 34 (2003) 35–41. [7] S.A. Lesley, P. Kuhn, A. Godzik, et al., Structural genomics of the Thermotoga maritima proteome implemented in a high-throughput structure determination pipeline, Proc. Natl. Acad. Sci. USA 99 (2002) 11664–11669. [8] J.R. Hudson Jr., E.P. Dawson, K.L. Rushing, et al., The complete set of predicted genes from Saccharomyces cerevisiae in a readily usable form, Genome Res. 7 (1997) 1169–1173. [9] A.J. Walhout, G.F. Temple, M.A. Brasch, et al., GATEWAY recombinational cloning: application to the cloning of large numbers of open reading frames or ORFeomes, Methods Enzymol. 328 (2000) 575–592. [10] M.R. Chance, A.R. Bresnick, S.K. Burley, et al., Structural genomics: a pipeline for providing structures for the biologist, Protein Sci. 11 (2002) 723–738. [11] S.A. Lesley, J. Graziano, C.Y. Cho, et al., Gene expression response to misfolded protein as a screen for soluble recombinant protein, Protein Eng. 15 (2002) 153–160. [12] J.S. Andersen, B. Svensson, P. Roepstor, Electrospray ionization and matrix assisted laser desorption/ionization mass spectrometry: powerful analytical tools in recombinant protein chemistry, Nat. Biotechnol. 14 (1996) 449–457. [13] A.J. Link, J. Eng, D.M. Schieltz, et al., Direct analysis of protein complexes using mass spectrometry, Nat. Biotechnol. 17 (1999) 676–682. [14] M. Mann, R.C. Hendrickson, A. Pandry, Analysis of proteins and proteomes by mass spectrometry, Annu. Rev. Biochem. 70 (2001) 437–473. [15] F. Baneyx, Recombinant protein expression in Escherichia coli, Curr. Opin. Biotechnol. 10 (1999) 411–421. [16] N. Armstrong, A. de Lencastre, E. Gouaux, A new protein folding screen: application to the ligand binding domains of a glutamate and kainate receptor and to lysozyme and carbonic anhydrase, Protein Sci. 8 (1999) 1475–1483. [17] M.E. Goldberg, R. Rudolph, R. Jaenicke, A kinetic study of the competition between renaturation and aggregation during the refolding of denatured-reduced egg white lysozyme, Biochemistry 30 (1991) 2790–2797. [18] D.B. Smith, K.S. Johnson, Single-step purication of polypeptides expressed in Escherichia coli as fusions with glutathione Stransferase, Gene 67 (1988) 31–40. [19] H. Bedouelle, P. Duplay, Production in Escherichia coli and onestep purication of bifunctional hybrid proteins which bind maltose. Export of the Klenow polymerase into the periplasmic space, Eur. J. Biochem. 171 (1988) 541–549. [20] C.V. Maina, P.D. Riggs, A.G. Grandea III, et al., An Escherichia coli vector to express and purify foreign proteins by fusion to and separation from maltose-binding protein, Gene 74 (1988) 365–373. [21] S.D. Maleknia, C.Y. Ralston, M.D. Brenowitz, et al., Determination of macromolecular folding and structure by synchrotron X-ray radiolysis techniques, Anal. Biochem. 289 (2001) 103–115.

934

R.Y. Huang et al. / Biochemical and Biophysical Research Communications 307 (2003) 928–934 [26] S.C. Goldsmith, J.Q. Guan, S. Almo, et al., Synchrotron protein footprinting: a technique to investigate protein–protein interactions, J. Biomol. Struct. Dyn. 19 (2001) 405–418. [27] J.Q. Guan, S. Vorobiev, S.C. Almo, et al., Mapping the G-actin binding surface of colin using synchrotron protein footprinting, Biochemistry 41 (2002) 5765–5775. [28] H. Rashidzadeh, S. Khrapunov, M.R. Chance, et al., Solution structure and interdomain interactions of the Saccharomyces cerevisiae “TATA binding protein” (TBP) probed by radiolytic protein footprinting, Biochemistry 42 (2003) 3655–3665. [29] H. Zhu, M. Bilgrin, R. Bangham, et al., Global analysis of protein activities using proteome chips, Science 293 (2001) 2101–2105.

[22] J.G. Kiselar, S.D. Maleknia, M. Sullivan, et al., Hydroxyl radical probe of protein surfaces using synchrotron X-ray radiolysis and mass spectrometry, Int. J. Radiat. Biol. 78 (2002) 101–114. [23] J.S. Sharp, J.M. Becker, R.L. Hettich, Protein surface mapping by chemical oxidation: structural analysis by mass spectrometry, Anal. Biochem. 313 (2003) 216–225. [24] J.G. Kiselar, P.A. Janmey, S.C. Almo, et al., Visualizing the Ca2 dependent activation of gelsolin by using synchrotron footprinting, Proc. Natl. Acad. Sci. USA 100 (2003) 3942–3947. [25] M.R. Chance, Unfolding of apomyoglobin examined by synchrotron footprinting, Biochem. Biophys. Res. Commun. 287 (2001) 614–621.


相关文章:
更多相关标签: