Anal. Chem. 2000, 72, 5431-5436
Cleavage and Identification of Proteins: A Modified Aspartyl-Prolyl Cleavage Adrianne Kishiyama, Zemin Zhang,† and William J. Henzel*
Department of Protein Chemistry, Genentech, Inc., 1 DNA Way, South San Francisco, California 94080
We have developed a method for rapidly cleaving and identifying proteins electroblotted onto poly(vinylidene difluoride) membranes. Cleavage is performed with 10% acetic acid in 7 M guanidine chloride at pH 2.5 for 1 h at 90 °C, resulting in fragmentation primarily at aspartylprolyl bonds. Peptides resulting from non-Asp-Pro cleavage are N-terminally blocked by reaction with orthophthalaldehyde (OPA) prior to automated Edman degradation. Reaction with OPA after cleavage blocks all amino acids containing primary amino groups. Only peptides containing an N-terminal amino acid with a secondary amino group (proline) will be available for reaction with the Edman reagent. The sequences obtained are used for protein database searching. Using this approach, proteins that are found to be N-terminally blocked can be removed from the sequencer, cleaved with acetic acid, blocked with OPA, and reapplied to the sequencer. The protein can then be identified from a database search using the sequence mixture obtained. With the recent completion of over 30 genomes, understanding protein function has challenged biochemist and biologists to develop new tools. Peptide mass fingerprinting has become a frequently utilized tool for the identification of proteins that have sequences in a protein sequence database.1-5 This method has increasingly replaced automated Edman degradation for protein identification, especially for proteins that contain a blocked N-terminus. A limitation of peptide mass fingerprinting is the requirement of peptide fragments to have sequence identity with database matches. Since many genomes have not been completely sequenced, peptide mass fingerprinting can fail to identify a protein when the sequence is not in a protein sequence database. Direct sequencing by Edman degradation can identify proteins by homology when the species of interest is not contained in a sequence database. However, many proteins are N-terminally blocked, preventing direct N-terminal analysis. Cleavage of the * To whom correspondence should be addressed: (fax) (650) 225-5945; (email)
[email protected]. † Department of Bioinformatics. (1) Henzel, W. J.; Billeci, T. M.; Stults, J. T.; Wong, S. C.; Grimley, C.; Watanabe, C. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 5011-5015. (2) James, P.; Quadroni, M.; Carafoli, E.; Gonnet, G. Biochem. Biophys. Res. Commun. 1993, 195, 58-64. (3) Pappin, D. J. C.; Hojrup, P.; Bleasby, A. J. Curr. Biol. 1993, 3, 327-332. (4) Yates, J. R.. III; Speicher, S.; Griffin, P. R.; Hunkapiller, T. Anal. Biochem. 1993, 214, 397-408. (5) Mann, M.; Hojrup, P.; Roepstorff, P. Biol. Mass Spectrom. 1993, 22, 338345. 10.1021/ac000547w CCC: $19.00 Published on Web 10/05/2000
© 2000 American Chemical Society
blocked protein with cyanogen bromide followed by mixture sequence analysis of the resulting peptides can allow rapid protein identification of blocked proteins. This method is only useful when the protein of interest contains only a few methionine residues. An alternative approach is the use of chemical cleavage between aspartic acid and proline residues. Most proteins contain less copies of this dipeptide sequence than methionine residues. We have developed a method that utilizes Edman degradation of peptide fragments produced by dilute acid cleavage to identify proteins. Dilute acid at elevated temperatures has been utilized as a method to cleave proteins at aspartic acid residues.6,7 The reaction can be shifted toward specific cleavage between aspartic acid and proline residues by using lower temperatures and longer reaction times. A mechanism for the aspartyl-prolyl cleavage reaction by Piszkiewicz et al.8 proposed that the reaction occurs via intramolecular catalysis by carboxylate anion displacement of the protonated nitrogen of the peptide. The aspartyl-prolyl cleavage rate is enhanced due to the greater basicity of the imino ring of the proline nitrogen. Reaction times previously reported for aspartyl-prolyl cleavage vary from 24 to 120 h. Various acids including 10% acetic acid adjusted to pH 2.5, 70-90% formic acid, or 0.015-0.03 N HCl9 have been used. Addition of a denaturing agent to the acid solution, longer reaction times,9 or higher temperature10 was found to increase cleavage yields. Higher temperature and increased reaction time can result in the hydrolysis of other labile peptide bonds as well as amide groups on asparaginyl and glutaminyl residues. We have utilized a primary amine-blocking reagent to prevent detection of non-Asp-Pro cleavage peptides. The reagent, orthophthalaldehyde (OPA), has been used as a blocking agent, reacting with primary but not secondary amines such as proline.11 When the peptides obtained from the dilute acid cleavage are reacted with OPA, only peptides with an N-terminal proline residue will remain unblocked. On the basis of studies with known and unknown proteins, we describe a protocol for cleavage of aspartyl(6) Inglis, A. S. In Methods in Enzymology; Hirs, C., Timasheff, S. N., Eds.; Academic Press: San Diego, 1983; Vol. 91, pp 324-332 (7) Schultz, J. In Methods in Enzymology; Hirs, C., Ed.; Academic Press: New York, 1967; Vol. 11, pp 255-263. (8) Piszkiewicz, D.; Landon, M.; Smith, E. L. Biochem. Biophys. Res. Commun. 1970, 40, 1173-1178. (9) Landon, M. In Methods in Enzymology; Hirs, C., Timasheff, S. N., Eds.; Academic Press: .New York, 1977; Vol. 47, pp 145-149. (10) Marcus, F. Int. J. Protein Res. 1985, 25, 542-546. (11) Brauer, A.; Oman, C. L.; Margloies, N. Anal. Biochem. 1984, 137, 134142.
Analytical Chemistry, Vol. 72, No. 21, November 1, 2000 5431
prolyl bonds at 90 °C followed by reaction with OPA as a method for the identification of proteins using database searching.
Table 1. Asp-Pro Cleavage of BSA at 37 °C % cleavage yield
EXPERIMENTAL SECTION Materials. Methanol was purchased from Burdick & Jackson (Muskegon, MI). Dithiothreitol was obtained from Diagnostic Chemicals (Oxford, CT). n-Isopropyliodoacetamide was from Molecular Probes (Eugene, OR). Thioglycolic acid (mercaptoacetic acid), 3-(cyclohexylamino)-1-propanesulfonic acid (CAPS), ammonium hydroxide, bovine carbonic anhydrase, bovine albumin, rabbit muscle phosphorylase, and sodium borate were purchased from Sigma (St. Louis, MO). Acetic acid was obtained from Fluka (Rokonkoma, NY), and guanidine hydrochloride (8.0 M sequenal grade) was from Pierce (Rockford, IL). Chromatographic grade o-phthalaldehyde was obtained from Pickering Laboratories (Mt. View, CA). Electrophoresis purity grade 2-mercaptoethanol was purchased from BioRad (Hercules, CA). All water used was from an ultrapure Milli-Q system by Millipore (Milford, MA). Alkylation and Electroblotting. Proteins were reduced in 20 µL of BioRad Laemmli sample buffer adjusted to pH 8.3 containing 10 mM DTT at 85 °C for 5 min. Alkylation was performed by the addition of 2 µL of 200 mM (0.08 mg) n-isopropyliodoacetamide12 in methanol at 25 °C for 20 min. Proteins were separated on BioRad precast gels and electroblotted on Problott membranes (PE-Applied Biosystems) in a BioRad Trans-Blot transfer cell using 10 mM CAPS, pH 11.0, 10 mM thioglycolic acid, 10% methanol as the transfer buffer for 1 h at 250 mA constant current13. The poly(vinylidene difluoride) (PVDF) membrane was stained with 0.1% Coomassie Blue R-250 in 50% methanol for 0.5 min and destained with 10% acetic acid in 50% methanol for 2-3 min. The membrane was thoroughly washed with water and allowed to dry before storage at -20 °C. Asp-Pro Cleavage. The bands containing the protein of interest were excised from the PVDF and wetted with 1 µL of methanol. Proteins were cleaved with 30 µL of 10% acetic acid in 7 M guanidine hydrochloride, pH 2.5, for 1 h at 90 °C. The solution was then removed, and the membrane was washed three times with Milli-Q water. The PVDF bands were treated at 25 °C for 5 min, with 30 µL of 30 mM OPA, 57 mM 2-mercaptoethanol in 0.1 M sodium borate (50 mg of OPA was dissolved in 1250 µL of methanol and then added to 11.2 mL of 0.1 M sodium borate containing 50 µL of 2-mercaptoethanol). Samples were washed six times with 200 µL of water, dried in a SpeedVac, and stored at 5 °C prior to analysis. Cyanogen Bromide Cleavage. Reduced and alkylated protein blots were treated with 20 µL of 0.1 N HCl at 45 °C for 3 h using a single crystal of cyanogen bromide and then dried in a SpeedVac.14 Automated Edman Degradation. Automated protein sequencing was performed on PE-Applied Biosystems Procise 494A protein sequencers. The Procise sequencers were equipped with 6-mm-diameter microcartridges and on-line PTH analyzers. The coupling buffer used was n-methylpiperidine (PE-Applied Biosystems or distilled in-house) in 1-propanol and water (25:60:15). (12) Krutzsch, H. C.; Inman, J. K. Anal. Biochem. 1993, 209, 109-116. (13) Matsudaira, P. J. Biol. Chem. 1987, 262, 10035-10038. (14) Zalut, C.; Henzel, W. J.; Harris, W. H. J. Biochem. Biophys. Methods 1980, 3, 11-30.
5432 Analytical Chemistry, Vol. 72, No. 21, November 1, 2000
cleavage timea (h)
peptide 1b (PNTLCDEFXA)
peptide 2 (PHACYSTVFD)
24 48 72
11c 19 21
11 18 28
a BSA was cleaved with 10% acetic acid in 7 M guanidine chloride (pH 2.5), and then blocked with OPA. b The residue labeled X, is a lysine residue that was not detected by Edman sequencing. The sidechain amino group of lysine was modified with OPA, which coeluted with the DPTU peak. c All values represents the average of triplicate analysis.
Twenty-minute Edman cycles were used as described.15 PTH columns (2.0 × 150 mm) containing Haisil resin were supplied by Higgins Analytical (Mt. View, CA). Acetone was routinely added to solvent A to balance the baseline. Peaks were integrated with Justice Innovation software using Nelson Analytical 760 interfaces (PE-Applied Biosystems). Sequence interpretation was performed on a DEC Alpha.16 SEQSORT. Mixture sequences were sorted using the SEQSORT algorithm.15,17 The algorithm finds patterns specified as regular-expression syntax. Sequence mixtures can be sorted by using a known sequence or by comparing the sequence mixture with all proteins in a protein sequence database. RESULTS AND DISCUSSION Optimization of the Asp-Pro Cleavage Conditions. To determine the optimal cleavage time and temperature for AspPro cleavage, several time course experiments were performed. Table 1 shows the cleavage yields for the two expected aspartylprolyl peptides derived from a cleavage of BSA at 37 °C. The peptide yields shown in the table were determined by quantitative Edman degradation of the mixture sequence present on the PVDF membrane. The percent cleavage was calculated by dividing the initial yields of the Asp-Pro cleavage peptides by the initial yield of the protein N-terminal sequence obtained from an identical blot of BSA that was not subjected to cleavage. The reaction of OPA with peptides containing lysine results in derivatization of the -amino group of lysine. This phenylthiohydantoin (PTH) lysine derivative coelutes with diphenylthiourea (DPTU) on our chromatography system, preventing its quantitation. After 72 h of cleavage, the yields for the two peptides derived from Asp-Pro cleavage of BSA were only 21% for peptide 1 and 28% for peptide 2. By increasing the reaction temperature to 75 °C, the reaction time was significantly decreased. Table 2 shows the cleavage yields for the same two aspartyl-prolyl peptides derived from cleavage of BSA at 75 °C. The yield for peptide 1 obtained after 3 h of cleavage at 75 °C was 31% as compared to 21% obtained after 72 h of cleavage at 37 °C. The yields obtained for peptide 2 were similar for both the 3 h 75 °C and the 72 h 37 °C time points as shown in Tables 1 and 2. The cleavage reaction was also (15) Henzel, W. J.; Tropea, J.; Dupont, D. Anal. Biochem. 1999, 267, 148-160. (16) Henzel, W. J.; Rodriquez, H.; Watanabe, C. J. Chromatogr. 1987, 404, 4152. (17) Henzel, W. J.; Stults, J. T.; Wong, S. C.; Namenuk, A.; Yashio, J.; Watanabe, C. In Techniques in Protein Chemistry; Marshak, D. R., Ed.; Academic Press: San Diego, 1995; Vol. VII, pp 341-346.
Table 2. Asp-Pro Cleavage of BSA at 75 °C % cleavage yield cleavage timea (h)
peptide 1b (PNTLCDEFXA)
peptide 2 (PHACYSTVFD)
1 3 6
15c 31 41
21 31 32
a BSA was cleaved with 10% acetic acid in 7 M guanidine hydrochloride (pH 2.5) and then blocked with OPA. b The residue labeled X, is a lysine residue that was not detected by Edman sequencing. cThe 1-h time point values represents the average of duplicate analysis. The 3- and 6-h time point values were obtained from four replicate experiments. The side chain amino group of lysine was modified with OPA, which coeluted with the diphenylthiourea peak (DPTU).
Table 3. Asp-Pro Cleavage of BSA at 90 °C % cleavage yield 1b
cleavage timea (h)
peptide (D:PNTLCDEFXA)c
peptide 2 (D:PHACYSTVFD)
0.5 1 2 3 3c 6 7c
21d 51 48 50 24 41 24
20 56 39 45 27 33 21
a BSA was cleaved with 10% acetic acid in 7 M guanidine hydrochloride pH 2.5, and then blocked with OPA. b Guanidine hydrochloride omitted from cleavage solution. These experiments were performed in duplicate. c The residue labeled X, is a lysine residue that was not detected by Edman sequencing. The side chain amino group of lysine was modified with OPA, which coeluted with the DPTU peak. d The 0.5 and the 1 h time point values represents the average of duplicate analysis. The 2, 3 and 6 h time point values were obtained from four replicate experiments.
Table 4. BSA Peptides Present after Asp-Pro Cleavage peptidea DTHKSEIAHR (N-terminus) D/PNTLCDEFKA D/PHACYSTVF
starting residue no.
pmol
% yield
1
6.6
100%
119 365
3.3 2.8
51% 42%
0.8 1.6 1.0 1.2 0.9 1.4 0.2 1.5
11% 24% 15% 19% 11% 21% 3% 23%
Nonspecific Peptides Found D/LGEEHFKGLV 14 D/SPDLPKLKPD 109 D/EKKFWGKYLY 130 D/KPLLEKSHCI 280 D/FAEDKDVCKN 308 D/AFLGSFLYEY 324 D/EPQNLIKQNC 382 D/TEQIKKQTA 518
a BSA was cleaved for 1 h at 90 °C with 10% HOAc, 7 M guanidine chloride, pH 2.5.
performed at 90 °C as shown in Table 3. The optimal cleavage time at this temperature is 1 h. Cleavage at 90 °C resulted in significantly higher yields for peptide 1 (50%) and peptide 2 (45%) than were obtained with cleavage at 37 °C or at 75 °C. The yields for both peptides 1 and 2 decreased after 3 h at 90 °C, which may be a result of additional peptide bond cleavage resulting from longer exposure to the acetic acid at the higher temperature. As shown with the 3- and 6-h samples, the inclusion of 7 M guanidine
Figure 1. (A) PTH chromatogram showing the first cycle obtained from BSA electroblotted onto PVDF and cleaved with 10% acetic acid in 7 M guanidine hydrochloride, pH 2.5, and then blocked with OPA. (B) The same condition as shown in (A) except the protein was not reacted with OPA after cleavage. The peak labeled J represents a peak associated with the OPA reagents.
hydrochloride can increase the amount of Asp-Pro cleavage significantly. This is somewhat surprising since the protein was reduced and alkylated. Effect of Reaction with OPA. In another experiment, BSA was cleaved at 90 °C and not treated with OPA, in an effort to identify and quantitate other cleavage sites. Non-Asp-Pro cleavage increased beyond acceptable levels after 1 h (Table 4). In addition to the two peptides sequences derived from cleavage between aspartic acid and proline, eight other sequences that are a result of cleavage after aspartic acid residues were found. This is expected, since acetic acid at elevated temperature has been utilized to cleave at aspartic acid residues.6,7 However, the yields of the two peptides obtained from cleavage of the Asp-Pro peptide bond are more then twice the yield of the other peptides reported in Table 4. Analytical Chemistry, Vol. 72, No. 21, November 1, 2000
5433
Table 5. N-Terminally Blocked Proteins Identified by Database Searching from Asp-Pro Cleavagea Sequence Data accession no.
protein name
SY16_HUMANb
SY_16 human small inducible cytokine a16 precursor lithostathine 1-β precursor human bovine carbonic anhydrase
LITB_HUMANb,c CAH2_BOVINb JW0047d RAGE_MOUSEb PHS2_RABITb
class I cytokinase receptor precusor human advanced glycosylation end product specific receptor precursor mouse glycogen phosphorylase rabbit
molecular weight
no. of sequences found/expected
starting residue no.
11 184
1/1
91
16 414
2/2
28 508
2/2
40 431
2/2e
40 448
1/1
54 105 41 179 144 361 128
D/PASELTASVPN
97 158
3/3
80 322 635
D/PXRIYYLSLEF D/PVRTNFDAFPD D/PVVGDRLRVIF
sequence found D/PNLPLLPTRN D/PETWVDADLYCQN D/PKKNRRWHWSSGS D/PALXPLALVYG D/PGSLLPNVLDY D/PLEATVHWAPD/PLEXLNWVRL
a All proteins were cleaved with 10% acetic acid in 7 M guanidine hydrochloride, pH 2.5, at 90 °C for 1 h and then blocked with OPA b Swissprot database. c The protein was subjected to 10 cycles of automated Edman degradation, removed from the sequencer cleaved as described above and resequenced. d Genebank translated database. e The protein was expressed recombinantly without the transmembrane domain, which contains a third Asp-Pro site.
As the number of peptides increase, the sequences become difficult to interpret and sort. However, when the mixture of peptides on the PVDF membrane is reacted with OPA after the cleavage reaction, all primary amino groups are blocked. Since proline is a secondary amino acid and does not react with OPA, only sequences derived from Asp-Pro cleavage will be detected after OPA treatment. The improvement in sequence interpretation is shown in Figure 1. Two duplicate blots of BSA on PVDF were cleaved with 10% acetic acid for 1 h at 90 °C. The PTH chromatograms comparing the first cycle of the PVDF membrane treated with OPA contained only proline. In contrast, the BSA that was not reacted with OPA contains proline, the N-terminal residue aspartic acid, and other amino acids (Figure 1B), indicating the presence of peptides resulting from non-Asp-Pro cleavages (Table 4). Protein Identification. The SEQSORT algorithm was used to search the database when multiple sequences were observed. All combinations of adjacent residues in the sequence mixture were searched against a protein sequence database. Identifying proteins using a mixture of two to five sequences with the use of the SEQSORT algorithm is dependent upon identification of the majority of the amino acid residues in the peptide mixture. The SEQSORT algorithm only matches identical amino acid residues. Error tolerance is utilized in the algorithm by allowing mismatches. Low-yielding amino acids such as serine, threonine, and amino acids that can be easily oxidized (tryptophan and methionine) are often difficult to detect. The use of mismatches by the SEQSORT algorithm allows sequence database matches to be found without identifying every amino acid on every cycle. When mismatches are utilized, each peptide sequence found can contain from one to the maximum number of mismatches that were specified. More complex mixtures require longer sequences and more mismatches. Mismatches also enable protein identification by homology when the sequence of the protein species of interest is not contained in a database. Table 5 contains a list of N-terminally blocked proteins that were identified by database searching using sequence data obtained from automated Edman degradation of peptides generated by aspartyl-prolyl cleavage. Using 10-13 amino acid 5434 Analytical Chemistry, Vol. 72, No. 21, November 1, 2000
Figure 2. Percentage of proteins containing Asp-Pro sites versus molecular weight in a human protein sequence database. The average values of Asp-Pro sites were determined for all proteins between 20 000 and 300 000. The number of Asp-Pro sites were averaged in intervals of 10 000 for protein up to 100 000 and in intervals of 50 000 for proteins above 100 000.
residues, proteins that ranged in molecular weight from 11 000 to 97 000 and containing from 1 to 3 Asp-Pro cleavage sites were identified. After cleavage, all of the predicted peptides were found in the automated Edman sequence data. Carbonic anhydrase, Sy16, lithostathine, and glycogen phosphorylase all contain one or more Asn-Pro sites. The high temperature utilized for cleavage could theoretically lead to deamidation of asparagine residues resulting in new Asp-Pro cleavage sites. However, we have not found any cleavage at Asn-Pro sites using the 1-h cleavage protocol. Figure 2 is a plot of the percentage of proteins in a human protein sequence database that have an Asp-Pro sequence as a function of molecular weight. The database contained over 29 000 nonredundant human sequences. The percentage of proteins containing an Asp-Pro cleavage site increases as the size of the protein increases. Fifty percent of the 30 000 and 80% of the proteins of >70 000 contain an Asp-Pro sequence. Figure 3A shows a plot of the average number of Asp-Pro cleavage sites
Figure 3. (A) Average number of Asp-Pro sites in a human protein sequence database per protein versus molecular weight. (B) Average number of methionine residues per protein versus molecular weight in a human protein sequence database. The average number of AspPro sites or methionine residues was determined for all proteins between 20 000 and 300 000. The number of Asp-Pro sites or methionine residues were averaged in intervals of 10 000 for protein up to 100 000 and in intervals of 50 000 for proteins above 100 000.
per protein contained within a human protein sequence database for proteins between 20 000 and 300 000. The number of AspPro sequences on average in large proteins is relatively low. A protein with a molecular mass of 100 kDa has on average only three Asp-Pro sites. Asp-Pro Cleavage and Resequencing of Blocked Proteins. One of the proteins shown in Table 5 is lithospertathine 1-β precursor. Automated Edman degradation of this protein showed that the protein was N-terminally blocked. The PVDF blot was removed from the sequencer after 10 cycles, acid cleaved at 90 °C for 1 h, blocked with OPA, and then reapplied to the sequencer. The PTH chromatograms of this protein are shown in Figure 4. Only residues derived from the Asp-Pro cleavage were observed. Since the protein was initially sequenced, all lysine residues had been reacted with phenylisothiocyanate resulting in the conversion of the side-chain -amino groups into phenylthiocarbamyl derivatives. This blocks the lysine side chain from reaction with OPA, allowing lysine to be detected as PTH lysine. Lithospertathine 1-β was identified from a database search using this mixture sequence data (Table 5).
Figure 4. PTH chromatograms of the automated Edman degradation of cycles 1-7 of lithospertathine 1-β precursor. The PVDF blot was initially placed in the protein sequencer reaction cartridge and sequenced for 10 cycles. No sequence was detected. The PVDF band was then removed from the sequencer, cleaved at 90 °C with 10% acetic acid in 7 M guanidine hydrochloride, pH 2.5, and then blocked with OPA. The protein was again subjected to automated Edman degradation.
Use of Cyanogen Bromide Cleavage after Attempted AspPro Cleavage. Cyanogen bromide cleavage has been the most frequently utilized technique for the cleavage of blocked proteins. The use of cyanogen bromide, which results in peptide bond cleavage at methionine residues, results in the formation of many more peptides on average when compared with the use of AspPro cleavage. Figure 3B shows a plot of the average number of peptides per protein resulting from cleavage at methionine plotted with their corresponding molecular weight. Proteins 50T000 Da or larger have on average 10 or more methionine residues. Cyanogen bromide cleavage of proteins larger than 50T000 Da Analytical Chemistry, Vol. 72, No. 21, November 1, 2000
5435
Table 6. N-Terminally Blocked Protein Identified by Database Searching from CNBr Cleavage Sequence Data after Attempted Asp-Pro Cleavage accession no.
protein name
molecular weight
no. of sequences found/ expected
starting residue no.
sequence found
AF092137_1a,b
FK506-binding protein
22 856
3/2
85 104 197
W/FVLGVGQVIX M/CPGEXRXVVI M/DNDRQLSXAE
a Genebank translated database. b The protein was subjected to 10 cycles of automated Edman degradation and found to be N-terminally blocked. The PVDF band was then removed from the sequencer, cleaved with 10% acetic acid in 7 M guanidine hydrochloride, pH 2.5, at 90 °C for 1 h, blocked with OPA, and placed back in the sequencer for 10 cycles. No sequence was obtained since the protein does not contain Asp-Pro residues. The sample was removed from the sequencer a second time, cleaved with CNBr, and again placed in the sequencer (third time), which resulted in the sequence shown above. The residues labeled as X in the mixture sequence are lysine residues that were not observed.
are likely to result in very complex sequence mixtures. As the number of sequences in the mixture increases, more residues are required for protein identification to prevent false positives. Another problem that can occur with the use of cyanogen bromide cleavage is when methionine residues are oxidize to methionine sulfoxide. These residues are not cleaved with cyanogen bromide. Oxidation can occur during protein purification, electrophoresis, and electroblotting. The use of Asp-Pro cleavage avoids these problems. The Asp-Pro cleavage procedure can be utilized on proteins that have been subjected to automated Edman degradation and shown to be N-terminally blocked, as was illustrated above for lithospertathine 1-β precursor. However, if this strategy fails, a cyanogen bromide cleavage can then be performed and a third attempt at sequencing can be made. Table 6 shows a blocked protein (FK506-binding protein) that was initially sequenced, removed from the sequencer, cleaved with acetic acid, blocked with OPA, and resequenced. No sequence was observed. The protein was then cleaved with cyanogen bromide and again placed in the sequencer for 10 cycles of Edman degradation. A mixture of three sequences was observed which were used in a database search to identify the 20-kDa band as FK506-binding protein. One (18) Savinge, W. E.; Fontana, A. In Methods in Enzymology; Hirs, C., Timasheff, S. N., Eds., Academic Press: New York, 1977; Vol. 47, pp 459-469.
5436 Analytical Chemistry, Vol. 72, No. 21, November 1, 2000
of the peptides was cleaved after a tryptophan residue, which can occur if tryptophan is oxidized.18 Another advantage of the use of aspartyl-prolyl cleavage for protein identification is the generation of a sequence-specific tag. When Asp-Pro cleavage is utilized with OPA blocking, the first two amino acid residues used for database searching would be Asp-Pro. The use of two amino acids allows a much higher search specificity compared to single amino acids from CNBr (Met) or Lys-C (Lys) digestion. Ten-fifteen Edman cycles are usually sufficient to uniquely identify a protein containing two to three Asp-Pro residues. The 1-h Asp-Pro cleavage combined with OPA blocking provides a rapid but highly specific chemical cleavage method. The low number of Asp-Pro bonds in most proteins makes this an ideal approach for the identification of blocked proteins using automated Edman degradation. ACKNOWLEDGMENT We thank Alison Bruce for graphics.
Received for review May 15, 2000. Accepted September 11, 2000. AC000547W
