Reductio Ad Amino Acid

Date: February 1, 1999Fusion/Tag Proteases TableProteolytic Enzymes TableTable 3Table 4 A proteome analysis aims to characterize all proteins expressed by an organism or tissue. The next step will be to correlate a protein profile with the appropriate genome, and beyond that researchers will want to understand the correlations between levels of proteins, co- and post-translational modifications, and cell or tissue activity. Many of the technologies that are necessary to realize this goal are developing simultaneously: large-scale, high-throughput DNA sequencing efforts are decoding genomes at an unprecedented rate; numerous expression monitoring approaches, expressed sequence tags, and two-dimensional protein gels are cataloging expression patterns; and a variety of staining methodologies are beginning to reveal co- and post-translational protein modifications. However, the basic identification of the proteins present in spots on 2-D gels remains a time-consuming (and sometimes unfathomable) problem. Flow path for proteomics, from M. Lopez, "Automating Proteome Analysis, " in American Biotechnology Laboratory, in press. Figure provided by Genomic Solutions. Proteome-level analysis involves physical separation (of often a subfraction) of the proteins from the cell or tissue of interest followed by characterization of the separated species. Wilkins and Gooley1 describe several methods for characterizing proteins at the primary (whole protein) and secondary (fragmented protein) levels. Primary methods include isoelectric point and apparent and actual mass, as well as N- and C-terminal sequence tags and extensive sequence analysis. Secondary methods include peptide mass fingerprinting (using MALDI, matrix assisted laser desorption/ionization, or ESI-MS, electrospray ionization mass spectroscopy, analysis of peptides following digestion of gel-resolved proteins), peptide fragmentation data (fragmentation of peptides following mass fingerprinting), and amino acid composition studies. This mini-review will examine the proteases that are commonly used to reduce whole proteins to collections of peptides in the production of peptide mass fingerprint data. Any technique that is slow, labor intensive, or expensive is not a candidate for proteome-level protein analysis, but peptide mass fingerprinting is a relatively straightforward technique that appears amenable to automated and high-throughput approaches. Peptide mass fingerprinting generally involves the following seven steps: (1) preparation of the protein sample, (2) separation of the protein sample using 2-D gel electrophoresis, (3) blotting of the gel onto a membrane (optional but usually performed), (4) excision of spots and destaining, (5) digestion with a protease, (6) use of MALDI-TOF MS or ESI-MS to measure the masses of the peptides produced, and (7) matching the sizes of the peptides produced with a protein database. Several programs to assist with protease digestion analysis are available over the web. MS-Digest, available at http://prospector.ucsf.edu/, allows for the "in silico" digestion of a protein sequence with a variety of proteolytic agents including trypsin, chymotrypsin, V8 protease, Lys-C, Arg-C, Asp-N, and CNBr. The program calculates the expected mass of fragments from these virtual digestions and allows the effects of protein modifications such as N-terminal acetylation, oxidation, and phosphorylation to be considered (a Java-enabled updated version is available from falson.ludwig.ucl.ac.uk/mshome.htm). A similar program is available from the OWL resource at prowl.rockefeller.edu/PROWL/ as well as from other locations on the web. (A list of proteome-related software resources is maintained at www.expasy.ch/tools.html.) In order to increase the confidence with which proteins can be identified by peptide mass fingerprinting, researchers have added a variety of "second layers" of analysis to the initial set of peptides. These analyses might involve partial Edman sequencing, amino acid composition, or a second round of fragmentation (often using a physical rather than enzymatic approach). Duplicate experiments where the protein is initially cleaved with a protease of different specificity also lead to high confidence in the protein's identity. Trypsin, because it is inexpensive, stable, and widely available, has been the enzyme of choice in most peptide mass fingerprinting and other diagnostic protein digestion experiments. However, of the several hundred proteases that have been characterized, a handful of others either have been used in peptide mass fingerprinting experiments or have been incorporated into the software that is used for experimental analysis. The most common of these are chymotrypsin, V8 protease, endoproteinase Lys-C, endopeptidase Arg-C, and endopeptidase Asp-N. We'll look at these in more detail presently. One of the earliest proteomics-relevant reports of protein analysis by the proteolytic digestion of proteins separated by electrophoresis came from Aebersold et al.2 The authors ran shark myelin proteins on 1-D or 2-D gels and electroblotted the bands or spots onto nitrocellulose. Regions of interest were excised from the membrane and treated with polyvinylpyrollidone-40 (PVP-40) to prevent the protease from binding to the membrane. Proteins were digested in situ using trypsin or V8 protease and the resulting peptides analyzed by HPLC. Kawasaki and Suzuki3 investigated methods to digest proteins directly following SDS PAGE and to analyze the results using HPLC. By using a DEAE precolumn, they were able to eliminate the adverse effects of SDS on the HPLC reverse phase column peptides that resulted from target proteins digested using endoproteinase Lys-C or V8 protease in the presense of SDS. James et al.4 ran 2-D gels, excised the spots of interest directly from the gel (no blotting), digested the proteins in situ using trypsin or endoproteinase Asp-N, and analyzed the resulting peptides using capillary electrophoresis. In the same year, Yates et al.5 published a theoretical paper that showed the feasibility of identifying a protein by screening a series of sized peptides against a database. They created a collection of microcapillary HPLC ESI-MS masses for a series of cytochromes-C that they screened against theoretical digests (trypsin, V8 protease, endoproteinase Lys-C, and CNBr) of The Protein Identification Resource, (PIR), a publicly accessible protein sequence database). Wheeler et al.6 brought many of these technologies together. They ran 2-D gels of human heart proteins, blotted them onto PVDF, and stained with coomassie brilliant blue. Stained protein spots were excised from the blots and destained in acetonitrile. They placed small squares of the excised membrane regions in tubes, digested them overnight using modified trypsin or endopeptidase Lys-C, and finally subjected the digested proteins to MALDI-MS. The resulting peptide masses were screened against several databases. Seven of the twelve proteins studied were identified in agreement with assignments using other methods, three proteins that had been previously assigned identities failed to be confirmed, and two previously unidentified proteins were tentatively assigned. Several of these techniques have evolved from protein microsequencing efforts. Many eukaryotic proteins have blocked amino termini, rendering them impossible to sequence directly using Edman degradation, and several laboratories overcame this problem by digestion with CNBr, which cleaves proteins at methionine residues (see, for example, Vandekerckhove et al.7). Other groups (see, for example, Stone et al.8) developed techniques for complete digestion of the proteins using trypsin or endoproteinase Lys-C followed by HPLC purification of the resulting peptides. It was found that, in order to obtain complete digestion, it was important to purify the starting material away from residual SDS. Depending on the prevailing salt concentrations, this was achieved either by dialysis and vacuum concentration or by TCA precipitation. Many of these extra steps can be eliminated by blotting the proteins onto PVDF and digesting the protein on the membrane. However, when the proteins are blotted onto PVDF it becomes necessary to treat the membrane to prevent the protease from simply sticking. Previously, this was generally done with PVP-40, but PVP-40 resulted in high backgrounds and increased processing time after digestion, and a new approach using hydrogenated Triton X-1009 seems to be gaining in popularity. Even this approach has problems with MALDI though, and Pappin et al10 developed a trypsin digestion method using octyl glucoside that overcomes many of the difficulties. The serine, threonine, and cysteine proteases form a group of enzymes that is distinct from the aspartic and metallo proteinases. Trypsin is perhaps the archetypal protease in that it is the one we all learned about in our biochemistry classes. Trypsin is one of the serine proteases. Barrett et al.11 describe about 40 families of serine proteases (based on amino acid sequence) grouped into six clans (based on tertiary structure). The serine proteases are so named because of a highly reactive serine residue in the active site (threonine or cysteine in the threonine and cysteine proteinases). This amino acid forms part of the catalytic triad, generally in an almost linear arrangement with a histidine and an aspartic acid residue. The residues of the catalytic triad form a charge transfer relay: the Asp residue strongly polarizes the His residue, which in turn accepts a hydrogen ion from the Ser residue. In the related aspartic and metallo proteinases, a water molecule initiates the nucleophilic attack on the substrate, whereas in the serine proteases, it is this ionized serine residue that initiates the attack. One consequence of this is that there is an acyl intermediate formed between the protease and its substrate in the serine, threonine, and cysteine proteinases but not in the aspartic or metalloproteinases. 1. Recombinant Enterokinase Titration. A pET fusion protein (3 µg) was digested with increasing amounts of Recombinant Enterokinase (rEK) in separate reactions. Samples were analyzed by SDS-PAGE (4-20% gradient gel) followed by staining with Coomassie blue. The 0.06 units lane corresponds to 1 rEK unit per 50 µg target protein, which exhibits >95% cleavage. One unit (Novagen) cleaves 50 µg Cleavage Control Protein (a 48 kDa pET fusion protein) in sixteen hours at room temperature. Trypsin (EC 3.4.21.4) is classified in clan SA, family S1, in the Handbook of Proteolytic Enzymes. Most trypsin that is used for protein sequencing or fingerprinting is of porcine origin. Trypsin hydrolyzes peptide bonds at the carboxyl side of lysine and arginine residues. The trypsin substrate binding pocket contains an aspartate residue (Asp 189) at the bottom. This Asp residue forms a bond (salt bridge) with the positively charged group at the end of its substrate-specifying amino acids, lysine and arginine. Most trypsins have a molecular weight of approximately 24 kD. Inactive trypsinogen is cleaved between Lys 6 and Ile 7 to form trypsin. This cleavage is carried out primarily by another proteolytic enzyme, enterokinase, but trypsin itself can autolytically carry out the cleavage as well. Biologically, trypsin's main role is as the activator of many pancreatic enzymes. Trypsin also exhibits autolysis and corresponding loss of activity. This autolysis can be prevented by the addition of calcium ions. However, it is often not desirable to have excess calcium in the reaction. Many companies (for example, Promega) sell an enzyme that has been subjected to reductive methylation. First published in the late seventies, methylation of trypsin results in an enzyme that retains high catalytic activity but is not susceptible to autolysis. Chymotrypsin (EC 3.4.21.1), like trypsin, is also classified in clan SA, family S1. Two major forms, chymotrypsin A and chymotrypsin B share over 80 percent sequence similarity. While the precise kinetic characters may differ, the overall specificity of the different forms remains the same: chymotrypsin generally cleaves on the C-terminal side of Tyr, Phe, and Trp residues (occasional cleavage at Lys has been reported). This specificity is largely determined by the structure of chymotrypsin's substrate binding pocket, which is considerably larger than trypsin's and can thus accommodate the large hydrophobic side chains of tyrosine, phenylalanine, and tryptophan. Chymotrypsin undergoes a similar but more complex activation process than trypsin: an initial activation event involving cleavage between Arg 15 and Ile 16 (in chymotrypsinogen A or B) is followed by at least two other digestive events. The most active form of chymotrypsin A or B comprises peptides 1-13, 16-147, and 149-246 held together by disulphide bonds. This protease is listed as an option in MS-digest, but you won't find it in the Handbook of Proteolytic Enzymes, in a Medline search, or in a search of the web. It doesn't have an EC number, and you cannot (intentionally) purchase it from any reputable supplier! MS-Digest helpfully lists slymotrypsin's cleavage sites though--F (Phe) W (Trp) Y (Tyr) K (Lys) R (Arg), which is the combined active site of trypsin and chymotrypsin. This activity is not derived, however; it is simply a mixture of the two enzymes. While slymotrypsin-like activity has been attributed to "bad" or poorly controlled batches of trypsin, this is probably only part of the story. Trypsin's autolytic reactions can lead to an enzyme that has more commonly been referred to as pseudotrypsin, which in addition to trypsin-like cleavage, exhibits chymotrypsin-like activity as well. Perhaps there are times when this mixed activity is desirable, but for the most part it would seem to be something to be avoided. Fortunately, it is an easy problem to eliminate if you stick with methylated rather than native trypsin. V8 protease (EC 3.4.21.19) is a clan SA, family S2 proteinase that is most commonly commercially produced from Staphylococcus aureus . The family S2 group of proteases comprises mostly bacterial endoproteinases that are secreted or membrane bound. V8 protease, a glutamyl endopeptidase I, is part of one of the four subfamilies, S2B that is known only from gram-positive bacteria. Confusion can arise when discussing this enzyme because, although the NC-IUBMB recommends the term glutamyl endoproteinase I, it is more commonly called endoproteinase Glu-C. Molecular surfaces of trypsin and chymotrypsin. Image provided by Molecular Simulations (San Diego, Calif.). References are made to two different V8 protease activities: V8E and V8DE. This is due to the altered specificity exhibited by the enzyme in different buffer systems. In ammonium bicarbonate or ammonium acetate buffers, glutamyl endoproteinase I cleaves only at the carboxylic side of glutamic acid residues; in phosphate and other buffers, glutamyl endoproteinase I cleaves at the carboxyl side of both glutamic and aspartic acid residues. (Glutamyl endoproteinase I has been extensively used in ammonium bicarbonate buffers for protein fragmentation prior to sequencing. In phosphate buffer, the activity is up to 10-fold higher.) V8 protease contains 268 amino acids, and similar to many other proteolytic enzymes, it is initially synthesized as a pre-pro-enzyme; the 69 amino acid leader segment contains a signal peptide and a propeptide. Activation appears to be by another peptidase. V8 protease is unique in its group, having a long (50 amino acid) tail that is composed of a tripeptide repeat that is of unknown function. Endopeptidase Lys-C (EC3.4.21.50) is also a clan SA, family S2 proteolytic enzyme that is now commercially produced from Pseudomonas aeruginosa. Originally isolated from Achromobacter lyticus and called Achromobacter protease I (API), the NC-IUBMB recommends the name lysyl endopeptidase. Lysyl endopeptidase cleaves proteins at the carboxyl side of lysine residues. Most of the enzymology of this protein is understood from studies of the Achromobacter protein, but the Pseudomonas protein is believed to be very similar. While the catalytic center is a classic serine protease catalytic triad (composed of His 57, Asp 113, and Ser 194), the substrate binding subsites (His 210, Gly 211, Gly 212, Asp 225) determine the lysine specificity. Lysyl endopeptidase is a single-chain molecule comprising 268 amino acids and 3 disulphide bonds. The tertiary structure resembles that of trypsin despite only low sequence similarity. Lysyl endopeptidase is synthesized as an inactive preprotein (653 amino acids) that comprises a signal peptide, propeptide, and C-terminal sequences. Activation is believed to be autocatalytic. Probably due to the extensively disulphide-bonded single-chain structure, lysyl endopeptidase is stable through a wide range of pH and temperature, but more interestingly is stable to low concentrations of SDS. Endopeptidase Arg-C (EC3.4.22.8) is a clan CX, family C11 proteolytic enzyme, and as such is a very different beast from those already discussed. Clostripain is produced by the bacterium Clostridium histolyticum and (to further exemplify the confusion that can be found in the proteolytic enzyme literature) has also been referred to as clostridiopeptidase B or gamma-protease. Clostripain cleaves at the carboxyl side of arginine residues, although under some conditions a minor lysyl bond cleavage can be observed. Clostripain is a heterodimeric protein with noncovalently bonded chains of 43 kD and 15.5 kD. Clostripain is produced from a precursor protein that comprises a 27 amino acid signal peptide, a 23 amino acid propeptide, the 131 amino acid light chain subunit, a 9 amino acid linker peptide, and the 336 amino acid heavy chain subunit. The precise details of the activation process are unknown. Endopeptidase Asp-N (EC3.4.24.33) is a clan MX, family M99 secreted endoproteinase that is commercially produced from a mutant strain of Pseudomonas fragi. This endopeptidase exhibits the unusual cleavage specificity of hydrolysing bonds that are N-terminal (not, as with the other proteolytic enzymes in this review, C-terminal) to aspartic acid residues (cleavage N-terminal to cysteic acid residues has also been observed). Endopeptidase Asp-N has also been called X-Asp metalloendopeptidase but should more correctly be called peptidyl-Asp metallopeptidase. Very little is known about this protein, which appears to be a single-chain protein of approximately 25 kD. The partial sequence contains a potential zinc binding site, although if the enzyme is inactivated by the chelator 1,10-phenanthroline, activity cannot be recovered by adding Zn2+. For peptide mass fingerprinting experiments, it is advisable to avoid slymotrypsin-like enzymes and other poorly controlled activities and to use "sequencing-grade" proteolytic enzymes. These grades probably cost more but have been purified more rigorously, often by affinity processes. Similar to restriction endonuclease quality control, sequencing-grade proteolytic enzymes can also be evaluated by their activity. HPLC analysis of the peptides produced from overdigestion of a test protein or peptide is a good indication that there are no additional activities (assuming the target material used in the test contains sites for expected additional cleavage reactions). The bottom line is to be sure you know exactly what you're working with and, for proteome-level studies that may be eventually compared with many other studies, to recharacterize the enzyme yourself using appropriate control targets. By Laura De Francesco Leading the list of hot proteases in molecular biology circles are those used to cleave fusion proteins. Fusion proteins--genetically engineered proteins containing one or more heterologous elements that improve the stability, solubility, or purification prospects of the protein-are being used increasingly for the production of proteins in expression systems. While the presence of a fusion element imparts a benefit, it also can be detrimental to some kinds of studies. For experiments requiring active enzyme, for example, native protein is often needed, which requires the removal of some fusion elements. In such instances, fusions can be removed either enzymatically or chemically, though often enzymatic means are preferred (because of milder conditions, greater specificity). Proteases with restricted specificities are available for liberating the protein from certain carrier or fusion elements. Coagulation factor Xa and the mammalian serine proteases thrombin and enterokinase have been used in this context. What these enzymes have in common--not found in those described previously--is large substrate recognition sequences, a property that reduces the likelihood of unwanted cleavage here in the protein. Having said that, it is important to note that unlike restriction endonucleases, which also have long recognition sequences, these proteases do not have nearly the specificity of the nucleases, and commonly cut at secondary sites. Nonetheless, it is becoming commonplace to see this class of enzyme referred to as restriction proteases. Key issues for customers choosing a vendor of fusion proteases can be boiled down to the following: Performance (activity, fidelity, and so on) Purity Price Completeness of kit Options for inactivation or removal after use A vexing aspect of price comparisons (and activity measure, for that matter) is the old and impractical unit definitions that are still in use by some. While they are somewhat "standard", measuring thrombin based on blood clotting times, for example, is a little silly when it has little to do with measuring enzyme purity or effectiveness on a purified protein. Increasingly, companies are beginning to incorporate fusion proteins into their quality control assays and unit definitions. Enterokinase (more correctly called enteropeptidase, but it has been known for so long as enterokinase, or EK, the name remains acceptable) has specificity for the sequence Asp-Asp-Asp-Asp-Lys, cleaving on the carboxy-terminal side of the lysine. This feature gives EK the attractive property of creating a protein with a normal N-terminus. Other properties adding to EK's usefulness are the wide range of conditions under which it is active and the availability of cloned enzyme. This last property is important because, when isolated from bovine or porcine intestine, preparations can be contaminated with other serine proteases (trypsin and chymotrypsin, for example). A number of suppliers (Stratagene, CalBiochem, Invitrogen, New England BioLabs, Novagen) now provide cloned enzyme, which eliminates this problem. A number of expression systems (ThioFusion Expression System, His-Patch ThioFusion System, Xpress Invitrogen; some pET vectors from Novagen; ESP Yeast pCAL-EK, Stratagene, pMAL vectors from New England BioLabs, and Kodak's FLAG expression system® produce proteins with the protein thioredoxin (or other EK recognition sequence) juxtaposed to the polylinker. These fusion proteins can be cleaved with EK to produce native protein. Factor Xa, derived from plasma, cleaves after the arginine residue in the sequence Ile-(Glu or Asp)-Gly-Arg, although it has been known to cleave at other basic residues. Several commercially available cloning vectors contain fusions with factor Xa cleavage sites: some of the pMAL vectors (New England BioLabs), GST vectors (Amersham Pharmacia), and PinPoint (Promega). According to Promega literature, factor Xa cleavage sites can also be cleaved with serine proteases (trypsin and endopeptidase Arg-C, aka clostripain). While these enzymes have less specificity than factor Xa, they can be useful for cleaving fusion proteins that are refractory to cleavage with factor Xa. Thrombin cleaves after arginine and lysine residues but has optimum cleavage at sites containing certain surrounding sequences: to wit, P4-P3-Pro-Arg/Lys-P1'-P2', where P3 and P4 are hydrophobic amino acids, and P1' and P2' are nonacidic amino acids (with cleavage occurring s at the Arg/Lys-P1' bond), or P2-Arg/Lys-P1', where P2 and P1' are Gly (with cleavage occurring once again at the Arg/Lys-P1' bond). Proteins containing glutathione S-transferase (GST) (such as Pharmacia's GST Vectors and Stratagene's ESP Yeast System) are susceptible to cleavage by thrombin, as well as some of Stratagene's and Novagen's expression vectors. A couple of unique, in some cases proprietary enzymes are also a part of the fray. Life Technologies incorporates into its expression vectors a cleavage site for TEV protease, an enzyme isolated from the tobacco etch virus. This protease, has a seven amino acid recognition site, (Glu-Asn-Leu-Tyr-Phe-Gln-Gly) with cleavage occurring between the Gln and Gly, resulting in a Gly at the amino terminus of the protein of interest. A recombinant form of this enzyme, containing a histidine tag, which facilitates removal of the protease after cleavage, is available from Life Technologies. Amersham Pharmacia Biotech offers PreScission Protease, a genetically engineered fusion protein consisting of a 3C protease of human rhino virus type 14 and GST. This enzyme specifically cleaves between the Gln and Gly residues of the recognition sequence of Leu-Glu-Val-Leu-Phe-Gln/Gly-Pro. AP Biotech's pGEX-6P vectors have the recognition sequence for PreScission Protease positioned between an affinity tag and the multiple cloning site. According to AP Biotech, the recognition sequence for PreScission Protease occurs at a frequency similar to that for thrombin and factor Xa. And finally, New England BioLabs offers Genenase™ I, a genetically modified subtilisin that shows less secondary site cleavage than EK and thrombin while cleaving a similar site (His-Tyr). NEB also has a system (IMPACT-Intein Mediated Purification with an Affinity Chitin-binding Tag) with a self-cleaving tag added to an affinity tag, such that the fusion protein is released from an affinity column and from the fusion peptides simultaneously. A list of the suppliers of these enzymes can be found in the Fusion/Tag Proteases Table. A more complete table of information on the enzymes can be found on our web site at www.the-scientist.lib.upenn.edu/yr1999/feb/profile1_990201.html The author, Laura De Francesco, can be reached at ldefrancesco@the-scientist.com. Fusion/Tag Proteases TableProteolytic Enzymes TableTable 3Table 4 M. R. Wilkins and A. A. Gooley, "Protein identification on proteome projects" in Proteome Research: New frontiers in Functional Genomics, M.R. Wilkins et al., eds., Springer, Berlin, 1997. R. Aebersold et al., "Internal amino acid sequence analysis by in situ protease digestion on nitrocellulose of proteins separated by one- or two-dimensional gel electrophoresis", Proceedings of the National Academy of Sciences USA, 84:6970-4, 1987 H. Kawasaki and K. Suzuki, "Separation of peptides in sodium dodecyl sulphate solution by reverse phase liquid chromatography: removal of sodium dodecyl sulphate from peptides using and ion exchange pre-column", Analytical Biochemistry , 186: 264-8, 1990. P. James, "Of genomes and proteomes", Biochemical and Biophysical Research Communications, 195: 58-64, 1993. J. R. Yates III et al., "Peptide mass maps, a highly informative approach to protein identification", Analytical Biochemistry, 214: 397-408, 1993. C.H. Wheeler et al., "Characterization of proteins from 2-D gels by matrix assisted laser desorption mass spectrometry and amino acid compositional analysis", Electrophoresis, 17: 580-7, 1996. Vandekerckhove et al., in Methods in Protein Sequence Analysis, KA Walsh, ed., Humana Press, 1986. K.L. Stone et al, "Enzymatic digestion of proteins and HPLC peptide isolation", in A Practical Guide To Protein And Peptide Purification For Microsequencing, T. Matsudaira, ed., Academic Press, San Diego, 1989. J. Fernandez, et al., "An improved procedure for enzymatic digestion of polyvinylidene difluoride bound proteins for internal sequence analysis", Analytical Biochemistry, 218:112-7. 1994. D. J. C. Pappin et al., "Peptide-mass fingerprinting as a tool for the rapid identification and mapping of cellular proteins" in Methods in Protein Structure Analysis, M.Z. Atassi, and E. Appella, Plenum Press, New York, 1995. A. J. Barrett et al., in The Handbook of Proteolytic Enzymes, A. J. Barrett et al., eds., Academic Press, San Diego, 1998. The author, Bob Sinclair, can be contacted at bobsinclair@tech-write-edit.com.

Reductio Ad Amino Acid

Feb 1, 1999

Interested in reading more?

Become a Member of