Protein cleavage strategies for an improved analysis of the membrane proteome
© Fischer and Poetsch; licensee BioMed Central Ltd. 2006
Received: 28 September 2005
Accepted: 02 March 2006
Published: 02 March 2006
Membrane proteins still remain elusive in proteomic studies. This is in part due to the distribution of the amino acids lysine and arginine, which are less frequent in integral membrane proteins and almost absent in transmembrane helices. As these amino acids are cleavage targets for the commonly used protease trypsin, alternative cleavage conditions, which should improve membrane protein analysis, were tested by in silico digestion for the three organisms Saccharomyces cerevisiae, Halobacterium sp. NRC-1, and Corynebacterium glutamicum as hallmarks for eukaryotes, archea and eubacteria.
For the membrane proteomes from all three analyzed organisms, we identified cleavage conditions that achieve better sequence and proteome coverage than trypsin. Greater improvement was obtained for bacteria than for yeast, which was attributed to differences in protein size and GRAVY. It was demonstrated for bacteriorhodopsin that the in silico predictions agree well with the experimental observations.
For all three examined organisms, it was found that a combination of chymotrypsin and staphylococcal peptidase I gave significantly better results than trypsin. As some of the improved cleavage conditions are not more elaborate than trypsin digestion and have been proven useful in practice, we suppose that the cleavage at both hydrophilic and hydrophobic amino acids should facilitate in general the analysis of membrane proteins for all organisms.
Major achievements have been accomplished in the field of mass spectrometry, namely the mild ionization techniques of MALDI  and ESI , which have been honoured in 2002 by the Nobel Prize in chemistry . These breakthroughs have raised protein analysis to a level, where in one experiment the study of the complete cellular set of all proteins, the proteome, has become possible. For the task of protein identification, nowadays MALDI and ESI ionisation sources are employed that are commonly coupled to quadrupols, TOFs or ion traps for ion separation. The proteins are cleaved with chemicals or proteases to obtain smaller peptide fragments. In usual high-throughput MALDI experiments, peptides are measured in the range of 600–4000 Da. For ESI, either a quadrupole with a common mass range of 0 to ~4000 Da or an ion trap with a common mass range of ~600 to ~4000 Da is used for molecule separation and analysis. The layout of ESI instruments is highly suitable for molecule fragmentation by collision with an inert gas, which can be used to determine the amino acid sequence of the peptide from the mass differences of the fragments. Nonetheless, the technology to study the proteome, proteomics, is far from mature. Certain protein classes, especially integral membrane proteins, still remain elusive. Different, complementary separation methods exist for integral membrane proteins. One possibility is to separate solubilized, intact membrane proteins, which can be carried out with blue native electrophoresis, SDS-PAGE, chromatography, and combinations of the techniques . Alternatively, the membrane fraction is treated with proteases to yield peptide fragments that are separated by multidimensional chromatography, e.g. MudPIT  and their sequences determined automatically, for instance with the SEQUEST algorithm . Several factors may hamper the detection of integral membrane proteins, such as solubilization problems, and low protein abundance. Another problem is the inability to obtain peptides from protein cleavage that are suitable for protein identification. The amino acids lysine and arginine are less frequent in membrane proteins compared to cytosolic proteins . Furthermore, these positively charged amino acids are not uniformly distributed along the protein sequence, but are mainly present in hydrophilic domains and almost absent in transmembrane helices. Hence the commonly used protease trypsin is unfavorable for the cleavage of integral membrane proteins. Alternative cleavage conditions, such as the consecutive use of cyanogen bromide and trypsin , as well as other proteases like proteinase K have been described  and proven superior to trypsin in practice. However, the latter is only suitable if MS/MS technologies can be employed for protein identification and it generally produces complex peptide mixtures. Furthermore, no transmembrane (TM) domains were detected by proteinase K treatment due to experimental limitations. As an efficient solubilization of membrane proteins is a necessary prerequisite for their identification by mass spectrometry and the detection of TM domains, new detergents  or solubilization conditions employing organic solvents  were developed. Yet far more possible cleavage conditions exist than experimentally tested for membrane proteins, but it is often not feasible to test them all in practice. A faster alternative for the screening for new, superior cleavage conditions is to perform in silico digestion (i.e. computational simulation) of entire membrane proteomes. In silico digestion has been carried out before to determine the suitability of single cleavage reagents , and to perform statistical analysis of proteomes . However, it has not been performed on the whole proteome scale to discover better cleavage conditions for integral membrane proteins, nor have the membrane proteomes from eukaryotes, archae, and prokaryotes been compared.
Taking the mass range of common instruments, and the MS/MS capability of ESI into account, in silico proteome analysis was carried out to determine the best experimental procedures for the identification of integral membrane proteins from Saccharomyces cerevisiae, Halobacterium sp. NRC-1, and Corynebacterium glutamicum. We show that the physiochemical properties of the membrane proteins significantly differ between the three kinds of organisms with consequences for protein identification by mass spectrometry. Finally, the practical applicability of new and potentially superior cleavage conditions were tested for bacteriorhodopsin as model membrane protein.
Physicochemical features of the membrane proteomes
Protein sequence coverage of peptide fragments
Abbreviations used for the cleavage conditions
Trypsin/ iodosobenzoic acid
Trypsin / chymotrypsin
Chymotrypsin/ cyanogen bromide
Chymotrypsin / staphylococcal peptidase I
Trypsin / iodosobenzoic acid / cyanogen bromide
Chymotrypsin / endopeptidase Lys-C
Trypsin / staphylococcal peptiase I / endopeptidase Asp-N
Trypsin / staphylococcal peptiase I
Pentafluoropropionic acid / cyanogen bromide
In a more comprehensive analysis a variety of cleavage conditions was tested and the percentage of peptides in the mass detection window of 600–4000 Da was calculated. This mass window is typical for common MALDI-TOF instruments and ESI ion traps, though for some peptides and instruments the window may be larger or smaller in reality.
In the first step, analysis of the complete C. glutamicum proteome (integral membrane and other proteins) was carried out, which shows that trypsin (KR) is the best single "cutter" with coverage of 69.9% – slightly better than chymotrypsin (FYW) with 67.5% (data not shown). Combinations of cleavage reagents can further improve the sequence coverage: FYWE and KRMW are the best combinations of cleavage sites in this mass window and result in coverage of 78.7% and 77.8%, respectively. It is possible to obtain a better sequence coverage using all cutters, if the mass window is extended up to 5000 Da. In this case, the cleavage sites FYWK give the best result with coverage of 82.5%. If only proteins are considered that are not predicted to be membrane integral (e.g. cytosolic), the best single cutter for proteins is trypsin (KR) with a sequence coverage of 74.5% in the 600–4000 Da window. Only a minor increase of sequence coverage (78.5%) is possible for non membrane integral proteins if the best combination of cleavage agents (FYWE) is used.
The Halobacterium sp. membrane proteome is covered poorly in the 600–4000 Da mass range if trypsin is used as protease (52.0% sequence coverage), while much better results are obtained for cleavages at FYW (71.3%) with chymotrypsin. Yet, if perfect cleavage occurs at FYWL, the sequence coverage drops again to a value of 60.1%. The reason is that additional cleavages at leucine produces peptides too small to be detected – the same effect can be observed for C. glutamicum and S. cerevisiae. Higher sequence coverage was obtained with combinations of "cutters"; cleavages at KRE, KRDE, and KRM already outperform trypsin but not chymotrypsin, while KRMW, FYWE, and FYWK permit even higher sequence coverage than chymotrypsin. The cleavage conditions KRE and KRDE achieve almost the same coverage, yet the most losses were found in the high mass range for KRE, while they were in the low mass range for KRDE with more cleavage sites. The best combination is a combination of chymotrypsin and staphylococcal peptidase I (FYWE) with 78.7% coverage. As for C. glutamicum, the best results are obtained if cleavage at hydrophilic and hydrophobic residues occurs.
The S. cerevisiae membrane proteome is covered well if only tryspin is used, i.e. 69.7% coverage is obtained. Cleavages at FYW permit a marginally better coverage of 72.7%. Combination of cleavage reagents were found that yield higher coverage than trypsin or chymotrypsin alone, but the increase in coverage is not as high as for C. glutamicum and Halobacterium. The best cleavage in the 600–4000 Da mass window would be with chymotrypsin and staphylococcal peptidase I at FYWE (78.7% coverage). KRM is also a good alternative for MALDI using yeast, since the coverage of 76.0% is not much lower.
In practice, one would prefer conditions that require the least experimental effort and are most reproducible. This could be the combination trypsin / chymotrypsin (FYWKR) since the digestion can be carried out simultaneously with both proteases and a good coverage of 75.7% can be attained.
From the results for all the organisms a hypothetical mean membrane proteome was calculated to examine whether one cleavage condition exists that is significantly better suited for membrane proteins than trypsin. Indeed, cleavages at KRMW, FYWK and FYWE yield a significantly better coverage. Furthermore it can be stated for all three organisms that if the number of cleavable amino acids with the same physicochemical properties (hydrophilic or hydrophobic) is increased, the coverage does not significantly increase and indeed, sometimes decreases. This observation is based on the fact that the amino acid sequences of proteins are not random; instead they reflect evolutional processes that lead to a better adaptation of function and stability. If it were random, sequence coverage of membrane proteins from C. glutamicum should be almost the same for the combinations KRM and KRE, since under this assumption the average peptide length would be 9.8 amino acids for KRM and 8.8 amino acids for KRE.
Fraction of unique peptides in the mass window
For C. glutamicum, the variation in the total number of peptides and of unique peptides in the assumed 600–4000 Da mass window is not very high. Here, the range for the number of peptides is from 42,919 total / 41,955 unique for cleavage at FYW to 63,563 total / 61,677 unique for cleavage at FYWKR. The ratio of unique / total peptides is almost identical for all tested cleavage conditions. The lowest number was calculated for KRDE (96.3% unique) and the highest for FYW (97.8% unique). In summary, the total number of peptides varies about 50% between the tested conditions, and about 97% of all peptides in the mass window are unique; thus, one peptide is almost always sufficient to identify the corresponding protein. Due to this, cleavage conditions can be used that achieve the best sequence coverage.
For Halobacterium sp., the number of peptides varies more between the different cleavage conditions than it is the case for C. glutamicum. The range is from 29,230 total / 28,532 unique for cleavage at FYW to 45,109 total / 43,684 unique for cleavage at FYWL. The fraction of unique peptides is almost similar for all cleavage conditions and ranges from 96.1% for KRDE to 97.6% for FYW. These calculations show that the cleavage conditions for highest sequence coverage (FYWE, FYWKR) are also suitable in respect to the number and uniqueness of the peptides.
The S. cerevisiae proteome consists of more proteins than C. glutamicum and Halobacterium sp. While the number of proteins in the yeast proteome is about 50% higher than in C. glutamicum, the number of peptides in the ESI mass window is more than doubled for the tested cleavage conditions. The range extends from 145,900 total / 131,707 unique for cleavage at FYW to 194,537 total / 170,538 unique for cleavage at FYWK. Among the three organisms, the percentage of unique peptides is lowest for yeast and rises from 86.2% for KRDE to 90.3% for FYW. In the case of yeast, cleavage at KR with trypsin already was suitable to obtain high sequence coverage, and only a slight improvement was obtained with other conditions, e.g. KRM. Both cleavages yield about the same percentage of unique peptides, which is 89.7% for KR and 89.1% for KRM.
Dependence of the peptide number from protein size and cleavage condition
For Halobacterium sp., the cleavage at KR is most unfavorable among all three organisms. On average, 8 peptides are obtained for membrane proteins larger than ~27 kDa. The relative suitability of all tested cleavage conditions is similar to C. glutamicum. MST and FYWKR are the best cleavage conditions and produce on average 8 peptides for proteins above ~13 kDa. Under these conditions, about twice as much peptides are obtained in the MALDI-TOF mass range (600–4000 Da) in comparison to trypsin, which should greatly increase the chance of identifying integral membrane proteins from this organism by the PMF technique.
For C. cerevisiae, the number of produced peptides does not vary much between the different cleavage conditions in contrast to the other two organisms. Proteolysis at KR with trypsin already produces 8 peptides per protein at ~17 kDa. Once again, MST and FYWKR are most suitable to increase the peptide number; both yield 8 peptides at ~14 kDa. Interestingly, the cleavages at MST and FYWKR give almost the same results in all three organisms. Furthermore, in respect to KR, KRM and FYW yield more peptides per protein in yeast compared to Halobacterium sp. and C. glutamcium. FYW is a largely complementary cleavage condition to KR, which was found by calculating for each membrane protein the number of obtained peptides in the MALDI range (data not shown). The number of peptides differed largest between these two conditions – up to 19 more peptides could be generated for some proteins below 40 kDa by FYW cleavage, yet up to ten less peptides were observed, too.
Bacteriorhodopsin – a practical approach
Result summary for the bacteriorhodopsin MALDI experiment
Mascot scores (MOWSE scoring)
Sequence coverage [%]
No. of identified peptides
Matched peaks [%]
Identification rank (Mascot list)
The other spectra showed a much higher number of peaks; especially the use of chymotrypsin leads to much more "noise" in the resulting spectra. In these cases the peak match rate decreases to 35% (6 of 17 peaks) for the single chymotrypsin digest and even to 26% (4 of 19 peaks) for the trypsin / chymotrypsin digest (Tab. 2). 50% (7 of 14 peaks) of all peaks could be related to bacteriorhodopsin peptides in the trypsin / CNBr digested sample. The trypsin digested sample showed the clearest MALDI-TOF spectrum with a total protein coverage of 8%. Nevertheless, this coverage is quite low when compared to the trypsin / chymotrypsin digest (20%) or to the chymotrypsin and trypsin / CNBr digest (27%). All the described data had an influence on the probability based scoring system i.e. the MOWSE score system Pappin, 1993 #14]. We noted that when searching the archaea database, Mascot scores above 61 were significant to identify a protein. Only cleavage with chymotrypsin (Mascot score 70) or trypsin / CNBr (Mascot score 92) resulted in a significant Mascot score for bacteriorhodopsin (see Tab. 2).
Predictions and limitations of the digest model
In this study in silico digestion was used to find cleavage conditions for an improved analysis of integral membrane proteins for the three model organisms S. cerevisiae, Halobacterium sp. NRC-1, and Corynebacterium glutamicum. The TMHMM algorithm  was used to identify integral membrane proteins from these organisms, since no complete experimentally determined dataset exists. While one has to be aware that certain integral membrane proteins (e.g. single transmembrane proteins and integral porins) are missed in the bioinformatics prediction, this and other state of the art algorithms are overall very specific and sensitive . To ascertain whether this in silico approach is pragmatic, we verified our predictions experimentally using the model membrane protein bacteriorhodopsin. We considered this experimental verification necessary, because the in silico model contains several limitations. For instance the used cleavage model assumes that with the exception of blocking, no missed cleavages occur. This is unrealistic since often in MALDI analysis, even with a peptide occurring at high frequency, one missed cleavage site is observed. This is mainly caused by inaccessibility of the cleavage site for the protease due to incomplete unfolding. In practice unfolding is improved by the use of chaotropes like urea , methanol , or an acid-labile surfactant . Furthermore, KK and RR are often not digested completely, since trypsin is an endoprotease and has weak exoprotease activity. Others have approached these problems by calculating best case (no missed cleavage) and worst case scenarios (one or two misses) for comparison [19, 13]. Different values exist in the literature for the minimum number of peptides that are necessary to identify a protein with MALDI-TOF PMF. They range from 2.2 and 2.3  to 5 . Further on, this number depends on the peptide mass, since fewer proteins match to large peptides . As often small membrane proteins do not yield enough peptides for identification, our in silico study recommends alternative cleavage procedures, e.g. a combination of trypsin and chymotrypsin for all three organisms, S. cerevisiae, Halobacterium sp., and C. glutamicum. For the identification of membrane proteins larger than 20 kDa by MALDI-TOF PMF trypsin alone appears sufficient.
When the mean number of peptides per membrane protein generated by different cleaving agents was calculated, trypsin was clearly inferior to a combination of trypsin/CNBr, or chymotrypsin. An inspection of the peptide sequences from membrane protein cleavages above 4000 Da revealed that indeed a high number of these are either uncleaved transmembrane segments, if trypsin (KR) is used as a protease, or long hydrophilic stretches of soluble protein domains, if chymotrypsin (FYW) is used. Therefore all three organisms exhibit the common characteristic that the best combination of "cutters" for integral membrane proteins is a cleavage at both hydrophilic and hydrophobic amino acids. Among these combinations, one was superior to trypsin for all three organisms, the combination of chymotrypsin and staphylococcal peptidase I (FYWE). In Shotgun proteomics, the peptide mass and sequence is generally the only information to identify the original protein. The analysis of the C. glutamicum proteome showed that in the 600–4000 Da mass range more than 96% of the peptides contain unique sequences and thus almost always lead to protein identification. This agrees with an in silico analysis of E. coli, S. pombe, and S. cerevisiae proteome – around 1%, 1.5%, and 2% of the respective ensemble of peptides with at least seven residues (i.e. ~800 Da), were redundant . However, it can be seen from the analysis for C. glutamicum, that also in the mass range below 600 Da at least more than 60% of the peptides are unique for trypsin cleavage, which leaves this range still attractive for analysis with instruments that allow low mass detection (e.g. Q-TOF).
Improvements in membrane protein analysis – theory and practice
We believe the best criterion to evaluate the suitability of the predictions would be to carry out a practical test using a MALDI-TOF instrument. For this purpose bacteriorhodopsin was digested with trypsin (KR), chymotrypsin (FYWL), trypsin / chymotrypsin (FYWLKR) and trypsin / CNBr (KRM). From this experiment, it can be concluded that the predictions agree well with the practical results. As predicted, the highest sequence coverage was obtained for digestion with chymotrypsin or trypsin / CNBr (27%). Additionally, the trypsin / CNBr digest yielded the most theoretical peptides and also for this sample, the highest number of peptides could be matched. It was observed that fewer peaks could be matched in the chymotryptic digests than in the tryptic digests. This is a consequence of the lower specificity of chymotrypsin and a higher autoproteolysis rate in comparison to trypsin. Thus the predictive power of the in silico digests is better for enzymes that are highly specific. Nonetheless, the predictions for chymotrypsin were good enough to point out the applicability of this protease for membrane protein identification. Comparison between the bacteriorhodopsin sequence coverage of chymotrypsin and trypsin /CNBr further revealed that although under both conditions equal values are obtained, the matched sequence parts are largely different. Two different digestions, one with chymotrypsin and one with trypsin / CNBr appear attractive if high sequence coverage is desired.
So far no in silico study has been published that analyses cytosolic and membrane proteins separately on the complete proteome level. It became evident that it is necessary to use different cleavage conditions for membrane and soluble proteins if high sequence coverage is desired. To better simulate the experimental conditions, it was chosen to limit the considered peptides for the in silico analyses to mass windows that are typical for the used mass spectrometry technology. Nevertheless, in practice the window can be even smaller. For instance, peptide analysis by MALDI-TOF is often unsuccessful for masses less than 780 Da, since the matrix produces a high number of peaks in this range, too.
We set the mass range to 600–4000 Da for the analysis of membrane proteome coverage and the number of unique peptides. This in our opinion, a reasonable range for MALDI and ESI, but we are aware that in reality the size of the useful mass window can vary. To allow for this variability, the membrane proteome coverage in relation to the peptide mass was depicted. It can be said in general that optimal cleavage conditions minimize the number of peptides and proteome coverage outside the accessible mass range. For the 600–4000 Da window, cleavage conditions superior to trypsin reduced the coverage of the membrane proteome beyond 4000 Da peptide size, while not too many small peptides were generated which either escape detection or are not informative for protein identification. It has been observed in practice that the peptide sequence affects its ionisation in MALDI and ESI. Although it was not considered in our study, trypsin usually delivers peptides that tend to ionise well, and this can be used to calculate the confidence for an identified peptide . While for the in silico analysis, a mass range of 600–4000 Da was assumed, doubly charged peptides with less than 800 Da are rarely used for database searching in ESI-MS/MS, since the two terminal amino acids fragment together, leaving only a very short stretch of peptide for sequence deduction. For the small proteome of C. glutamicum a sequence fragment of 4 amino acids is often sufficient to identify a protein, but this is not the case for organisms with a higher number of proteins.
The reason for the high number of unique peptides for the three organisms in the 600–4000 Da mass range is their relatively small number of coding genes. In reality, the number of peptides will be larger, due to effects like protein splicing, posttranslational modifications, missed cleavages, etc. For eukaryotic organisms it can indeed be important to verify if protein identification by a peptide eliminates the possibility of existing splice variants as demonstrated by Kuster et al. . In addition, it is estimated that by using MudPIT, about 23,000 peptides can be separated in one experiment ; this number is smaller than the number of predicted peptides from in silico digestion that fall into the defined mass range. These observations raise the question, whether the proposed strategy from in silico, i.e. multiple cleavage combinations, is applicable in reality. Our own findings for the membrane fraction of C. glutamicum  proof that the combination of trypsin and chymotrypsin or trypsin and cyanogen bromide  can be superior to trypsin alone. Others have found even for higher eukaryotes (brain homogenisate) that cleavage with unspecific proteinase K was extremely useful for membrane protein identification. However, it must be stressed that with current instrument limitations, prefractionation steps for membrane enrichment are mandatory.
By carrying out in silico proteome analysis, integral membrane protein cleavage conditions were identified that are superior to trypsin for all three examined organisms, Halobacterium sp. NRC-1, Saccharomyces cerevisiae, and Corynebacterium glutamicum. While the best result was obtained with a combination of chymotrypsin and staphylococcal peptidase I, in general all superior conditions are combined cleavages at hydrophobic and hydrophilic amino acids. The combined cleavages improved the sequence coverage for ESI-MS/MS and the identification of small integral membrane proteins by MALDI-TOF PMF. The practicability of cleavage conditions was assessed by digestion of bacteriorhodopsin and MALDI-TOF PMF. It can be concluded that the in silico predictions agree well with the practical results, particularly if highly specific proteases are used. Since some of the improved cleavage conditions (e.g. chymytropsin + trypsin) are not more elaborate than the classical trypsin digestion, they should useful alternatives to trypsin for a better analysis of integral membrane proteins.
Datasets and bioinformatics
A file containing the translated predicted ORFs for Halobacterium sp. NRC-1 (NC_002607) was downloaded from the NCBI FTP website. Predicted protein sequences for Saccharomyces cerevisae were obtained from the European Bioinformatics Institute (EBI) website . The translated predicted ORFs for Corynebacterium glutamicum were supplied by J. Kalinowski (Bielefeld University) and are part of the publicly available dataset at EBI. The in silico digestion software was written in PERL and is taking advantage of BioPerl 1.4 modules. Cleavage rules (simple model for trypsin and chymotrypsin) for the in silico digestions were employed according to the Expasy PeptideCutter website , aside from these rules perfect cleavage was assumed. Membrane proteins were predicted for all three organisms using the software TMHMM 2.0 .
Experimental validation of cleavage conditions
In each case 20 μg (protein amount) of isolated purple membranes were dissolved in sample buffer (10% v/v glycerol, 5% v/v mercaptoethanol, 3% w/v SDS, 62 mM Tris-HCl, pH 6.8, 0.01% bromophenol blue) and incubated for 30 min at 60°C. SDS-PAGE was performed according to Laemmli . After electrophoresis the proteins were stained by colloidal Coomassie blue using the protocol of Neuhoff .
In-gel digestion of bacteriorhodopsin
The Coomassie stained protein bands were excised from the SDS-polyacrylamide-gel and completely destained with 100 μL 25 mM ammonium hydrogencarbonate and 50% v/v acetonitrile (three times for 20 min by 37°C). The gel pieces were dried in a Speed Vac and subsequently covered with a 25 mM ammonium hydrogencarbonate buffer containing either 12.5 μg trypsin (three samples) or 12.5 μg chymotrypsin (one sample). Proteolysis was carried out overnight at 37°C. The next day, peptide samples of one tryptic digest and of the chymotryptic digest were eluted by adding 10 μL elution buffer (50% v/v acetonitrile, 0.5% TFA) with subsequent sonication for 20 min. The gel pieces of the second and third tryptic digests were completely dried in a vacuum concentrator and used for a second cleaving procedure. The dried gel piece of the second tryptic digest was covered with chymotrypsin in ammonium hydrogencarbonate buffer and incubated for 7 h at 37°C. Elution of the generated peptides was carried out as described above. With the third sample a CNBr cleavage was carried out in the dark at room temperature for 4 h, while a small crystal of CNBr was dissolved in 70 % TFA and added to the dried gel piece, followed by several washing steps according to the protocol of Van Monfort .
MALDI-TOF-MS and protein identification
Aliquotes of 0.6 μL peptide digest were applied onto a target plate and immediately mixed with an equal volume of α-cyano-hydroxycinnamic acid (10 mg/ml in 60% v/v acetonitrile and 1% v/v TFA). MALDI-TOF mass spectra were recorded using an Applied Biosystems Voyager DE-Pro instrument. Each spectrum was first externally calibrated by an Applied Biosystems external peptide standard and in a second step internally calibrated using trypsin autoproteolytic products. Monoisotopic peptide masses were used to search the ncbi database using the Mascot algorithm (version 2.1) . Peptides with missing cleavage sides of 2 were accepted and the mass tolerance of the peptide mass ions was set to < 100 ppm.
Purple membranes were a generous gift from Norbert Dencher (Physical Biochemistry, TU Darmstadt). This work was supported by grants from the Bundesministerium für Bildung und Forschung (BMBF 0312843E), and by the Fonds der Chemischen Industrie.
- Karas M, Hillenkamp F: Laser desorption ionization of proteins with molecular masses exceeding 10,000 daltons. Anal Chem 1988, 60(20):2299–2301. 10.1021/ac00171a028View ArticlePubMedGoogle Scholar
- Yamashita M, Fenn JB: J Phys Chem. 1984, 88: 4451. 10.1021/j150664a002View ArticleGoogle Scholar
- Cho A, Normile D: Nobel Prize in Chemistry. Mastering macromolecules. Science 2002, 298(5593):527–528. 10.1126/science.298.5593.527bView ArticlePubMedGoogle Scholar
- Kashino Y: Separation methods in the analysis of protein membrane complexes. J Chromatogr B Analyt Technol Biomed Life Sci 2003, 797(1–2):191–216.View ArticlePubMedGoogle Scholar
- Wolters DA, Washburn MP, Yates JR 3rd: An automated multidimensional protein identification technology for shotgun proteomics. Anal Chem 2001, 73(23):5683–5690. 10.1021/ac010617eView ArticlePubMedGoogle Scholar
- MacCoss MJ, Wu CC, Yates JR 3rd: Probability-based validation of protein identifications using a modified SEQUEST algorithm. Anal Chem 2002, 74(21):5593–5599. 10.1021/ac025826tView ArticlePubMedGoogle Scholar
- Kyte J, Doolittle RF: A simple method for displaying the hydropathic character of a protein. J Mol Biol 1982, 157(1):105–132. 10.1016/0022-2836(82)90515-0View ArticlePubMedGoogle Scholar
- van Montfort BA, Doeven MK, Canas B, Veenhoff LM, Poolman B, Robillard GT: Combined in-gel tryptic digestion and CNBr cleavage for the generation of peptide maps of an integral membrane protein with MALDI-TOF mass spectrometry. Biochim Biophys Acta 2002, 1555(1–3):111–115.View ArticlePubMedGoogle Scholar
- Wu CC, MacCoss MJ, Howell KE, Yates JR: A method for the comprehensive proteomic analysis of membrane proteins. Nat Biotechnol 2003, 21(5):532–538. 10.1038/nbt819View ArticlePubMedGoogle Scholar
- Luche S, Santoni V, Rabilloud T: Evaluation of nonionic and zwitterionic detergents as membrane protein solubilizers in two-dimensional electrophoresis. Proteomics 2003, 3(3):249–253. 10.1002/pmic.200390037View ArticlePubMedGoogle Scholar
- Blonder J, Goshe MB, Moore RJ, Pasa-Tolic L, Masselon CD, Lipton MS, Smith RD: Enrichment of integral membrane proteins for proteomic analysis using liquid chromatography-tandem mass spectrometry. J Proteome Res 2002, 1(4):351–360. 10.1021/pr0255248View ArticlePubMedGoogle Scholar
- Wise MJ, Littlejohn TG, Humphery-Smith I: Peptide-mass fingerprinting and the ideal covering set for protein characterisation. Electrophoresis 1997, 18(8):1399–1409. 10.1002/elps.1150180815View ArticlePubMedGoogle Scholar
- Cagney G, Amiri S, Premawaradena T, Lindo M, Emili A: In silico proteome analysis to facilitate proteomics experiments using mass spectrometry. Proteome Sci 2003, 1(1):5. 10.1186/1477-5956-1-5PubMed CentralView ArticlePubMedGoogle Scholar
- Kamo M, Tsugita A: Specific cleavage of amino side chains of serine and threonine in peptides and proteins with S-ethyltrifluorothioacetate vapor. Eur J Biochem 1998, 255(1):162–171. 10.1046/j.1432-1327.1998.2550162.xView ArticlePubMedGoogle Scholar
- Krogh A, Larsson B, von Heijne G, Sonnhammer EL: Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. J Mol Biol 2001, 305(3):567–580. 10.1006/jmbi.2000.4315View ArticlePubMedGoogle Scholar
- Chen CP, Kernytsky A, Rost B: Transmembrane helix predictions revisited. Protein Sci 2002, 11(12):2774–2791. 10.1110/ps.0214502PubMed CentralView ArticlePubMedGoogle Scholar
- Rabilloud T, ed: Proteome Research: Two-Dimensional Gel Electrophoresis and Identification Methods. Berlin Heidelberg: Springer; 2000.Google Scholar
- Suder P, Bierczynska A, Konig S, Silberring J: Acid-labile surfactant assists in-solution digestion of proteins resistant to enzymatic attack. Rapid Commun Mass Spectrom 2004, 18(7):822–824. 10.1002/rcm.1411View ArticlePubMedGoogle Scholar
- Fenyo D, Qin J, Chait BT: Protein identification using mass spectrometric information. Electrophoresis 1998, 19(6):998–1005. 10.1002/elps.1150190615View ArticlePubMedGoogle Scholar
- Perkins DN, Pappin DJ, Creasy DM, Cottrell JS: Probability-based protein identification by searching sequence databases using mass spectrometry data. Electrophoresis 1999, 20(18):3551–3567. 10.1002/(SICI)1522-2683(19991201)20:18<3551::AID-ELPS3551>3.0.CO;2-2View ArticlePubMedGoogle Scholar
- Parker KC: Scoring methods in MALDI peptide mass fingerprinting: ChemScore, and the ChemApplex program. J Am Soc Mass Spectrom 2002, 13(1):22–39. 10.1016/S1044-0305(01)00320-8View ArticlePubMedGoogle Scholar
- Kuster B, Mortensen P, Andersen JS, Mann M: Mass spectrometry allows direct identification of proteins in large genomes. Proteomics 2001, 1(5):641–650. Publisher Full Text 10.1002/1615-9861(200104)1:5<641::AID-PROT641>3.0.CO;2-RView ArticlePubMedGoogle Scholar
- Fischer F, Wolters D, Rogner M, Poetsch A: Towards the complete membrane proteome: High coverage of integral membrane proteins through transmembrane peptide detection. Mol Cell Proteomics 2005.Google Scholar
- Schluesener D, Fischer F, Kruip J, Rogner M, Poetsch A: Mapping the membrane proteome of Corynebacterium glutamicum. Proteomics 2005, 5(5):1317–1330. 10.1002/pmic.200400993View ArticlePubMedGoogle Scholar
- Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970, 227(259):680–685. 10.1038/227680a0View ArticlePubMedGoogle Scholar
- Neuhoff V, Arold N, Taube D, Ehrhardt W: Improved staining of proteins in polyacrylamide gels including isoelectric focusing gels with clear background at nanogram sensitivity using Coomassie Brilliant Blue G-250 and R-250. Electrophoresis 1988, 9(6):255–262. 10.1002/elps.1150090603View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.