Abstract
Abstract
Disulfide bridges are side-chain-mediated covalent bonds between cysteines that stabilize many protein structures. Disulfide mapping experiments to resolve these linkages typically involve proteolytic cleavage of the protein of interest followed by mass spectroscopy to identify fragments corresponding to linked peptides. Here we report the sequence-based “DIMPL” web tool to facilitate the planning and analysis steps of experimental mapping studies. The software tests permutations of user-selected proteases to determine an optimal peptic digest that produces cleavage between cysteine residues, thus separating each to an individual peptide fragment. The webserver returns fragment sequence and mass data that can be dynamically ordered to enable straightforward comparative analysis with mass spectroscopy results, facilitating dipeptide identification.
1. Introduction
A
The experimental steps of a disulfide mapping study using the “bottom-up” approach (Chait, 2006) involve protein digestion using proteolytic enzymes or chemical reagents, followed by mass determination of the fragments (Gorman et al., 2002; Tsai et al., 2013). Disulfide-linked peptides can be identified using the profile comparison approach, where differences in the mass spectra of reduced and unreduced aliquots of the protein digest correspond to linked peptides. If disulfide reduction using reagents such as dithiothreitol (DTT) is inefficient or leads to sample loss due to the aggregation of unfolded protein, it is still possible to identify disulfide-linked peptides by comparing the mass spectrum of the unreduced peptic digest with a theoretical complete digest of the fully reduced protein calculated using the same protease set.
A key decision in the planning of disulfide mapping studies is the choice of proteolytic enzymes or chemical reagents and the order of their application. An effective digestion for mapping studies produces cleavage of the substrate between each cysteine residue as this eliminates the possibility of fragments containing intrapeptide disulfides. However, testing permutations of proteases to achieve optimal cleavage can be arduous, particularly when there are a large number of cysteines in the target protein sequence. Here, we report the DIsulfide Mapping PLanner (DIMPL) web tool for in silico protein digestion, which determines the minimal set of user-selected proteases that will produce segregation of cysteines in a user-provided protein sequence. In addition to reporting the optimal protease set, the software returns the protein fragments and the corresponding masses and provides a range of options to order and analyze these data.
2. Methods
The core of DIMPL was developed using the Python programming language (v.2.7) and the software is hosted on a Linux platform. The web interface (www.lampert-lab.com/dimpl) is written in server- and user-sided programming languages, including PHP, MySQL, HTML5, CSS, and the JavaScript-based jQuery framework.
2.1. Input
The amino acid sequence of the protein of interest in FASTA format or as raw text is entered via the web interface. Alternatively, a UniProt accession number can be entered for dynamic sequence retrieval from the UniProt server following web-form submission. Next, the endopeptidase enzymes (or chemical reagents: cyanogen bromide, iodosobenzoic acid, and hydroxylamine) are selected for the digestion. For convenience, the proteases are grouped into “Common” (e.g., trypsin, chymotrypsin, and pepsin), “Less Common,” and “Uncommon” sets, which can be selected via a drop-down menu. Each protease can also be selected or deselected individually.
To further extend the number of digestion options available, custom proteases can be created with a user-provided substrate specificity. The custom digestion pattern consists of four amino acid positions before and two after the cutting site and the user can provide amino acids in single letter code that are to be matched or avoided at each position. Unset amino acid positions are kept as nonspecific.
Given that digestion reactions are not always 100% efficient, the user is provided with the option (via a drop-down menu) to include up to three “miscleavages” in their results. These are intact peptides that contain an uncleaved substrate site.
A final input option is a text box for an e-mail address to allow the user to receive results via e-mail.
2.2. Enzyme scoring scheme
Following web-form submission, each of the selected proteases is tested for its ability to digest the input protein sequence using the specificity rules for endopeptidase activity listed in the EXPASY PeptideCutter library (Gasteiger et al., 2005). Proteases unable to cleave the sequence at least once are rejected. The fragmentation patterns formed by the remaining proteases are then analyzed and scored (S) according to the following function:
where “F” represents the total number of fragments produced and “C” is the number of separated cysteines. The “C” term is squared to increase the score of proteases that separate a greater number of cysteines. Similarly, when different digestions result in an equal number of separated cysteines, the proteases that produce fewer fragments are scored higher.
If individual proteases are unable to produce full cysteine segregation, then the activity of multiple enzymes is assessed. However, determining the fragmentation pattern resulting from the simultaneous action of a combination of proteases is problematic as the activity of one protease may eliminate substrate sites for a second protease and vice versa. Therefore, DIMPL adopts the more experimentally relevant approach of digesting the protein substrate sequentially with proteases to produce a series of nested fragments. To determine whether full cysteine segregation is achievable, the software performs iterative testing of substrate digestion with protease permutations beginning with two enzymes and including additional proteases up to a maximum of five. If a satisfactory solution cannot be found with user-selected proteases, the substrate is then tested with other proteases to suggest an alternative digestion strategy to the user.
2.3. Output
The results page from the DIMPL server comprises two sections. The first section gives the list of proteases and the chronological order for their application to produce the requisite intercysteine cleavage (or when full cysteine separation is not possible, the selection of proteases that gives the submaximal cysteine segregation).
The second section details the fragment data produced by the in silico digestion. Fragment sequences are listed with corresponding masses, the order with which they occur in the protein primary sequence, and whether a cysteine is present in the fragment; the user can dynamically sort these results by their preferred field. The selection (by double clicking) of two fragments produces the calculated sum of their masses.
Ambiguous digestion sites in the sequence are designated with multiple arrows “>>.” These sites occur when two or more cleavage sites are present concurrently in the primary sequence. For example, trypsin cleaves after lysine (K) and will generate fragments of varying lengths when multiple consecutive “K's” are encountered. To account for these ambiguous sites, a feature of the results section is that the mass for any sequence of contiguous amino acids can be calculated by highlighting the corresponding region in the full-length protein sequence. Indeed, this allows the mass for a fragment of any length to be calculated, thus further enabling comparative analysis with mass spectroscopy results.
The results are available for download in XML or CSV files and are also posted by e-mail if requested via the input page. A list with all possible fragments bound together by a disulfide bridge is available for download as CSV file. In addition, the results page is accessible for 30 days via a uniquely generated URL.
3. Discussion
The DIMPL web tool provides an in silico digest of a protein sequence using user-selected proteases (and chemical cleavage reagents) and reports the order of digestion steps that produces the maximum number of cysteines separated onto the fewest number of peptide fragments. The molecular mass, sequence, and the order that each fragment occurs in the submitted protein are also reported.
A great variety of software tools exist for both the prediction and in silico study of disulfide connectivity in proteins (O'Connor and Yeates, 2004; Craig and Dombkowski, 2013; Yachdav et al., 2014). For example, the sequence-based software tools DISULFIND (Ceroni et al., 2006) and DiANNA (Ferrè and Clote, 2005) can accurately predict the disulfide connectivity of a correctly folded protein. DIMPL provides a complementary software tool that can support experimental efforts to validate these predictions. The aim of DIMPL is to provide researchers with an optimized workflow for disulfide mapping that is both time effective—involving the fewest number of digestion steps—and resource efficient in terms of protein sample and available digestion enzymes and reagents.
DIMPL is targeted to researchers adopting the bottom-up mass spectroscopy approach for protein characterization (Chait, 2006), which involves an initial sample digestion step. In general, trypsin is the sole protease used for digestion before fractionation of peptides by liquid chromatography and application to a mass spectrometer. Unfortunately, the slightly alkaline conditions required for optimal trypsin activity can produce the phenomenon of disulfide shuffling (Ryle and Sanger, 1955), which can confound the mapping process. Other proteases (e.g., pepsin) can be used to circumvent this problem and so DIMPL provides the user with a wide range of digestion options, including the ability for users to define their own unique protease specificity. This feature is advantageous as it also provides a measure of future proofing for the software.
Depending on the mass spectroscopy technique and ionization source, a multitude of different ions can be generated per peptide. Before a mass spectroscopy run, software such as Skyline (MacLean et al., 2010) can be used to calculate all the masses of a complex ionic mixture and plot a predicted mass spectrum. In comparison, DIMPL software adopts the comparatively straightforward approach of reporting just the mass of each full peptide fragment with charged N- and C-termini. This is consistent with the focus of DIMPL as an experimental planning software aid, as it aims to inform the users of the mass range of digestion products, thus allowing them to adjust their digestion strategy so that a suitable fragment range for their mass spectrometer can be generated. Finally, for researchers using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) or other techniques instead of mass spectroscopy to size their proteolytic fragments, the results returned by DIMPL should prove similarly useful.
DIMPL is a user-friendly web tool that facilitates the planning and data analysis steps of disulfide mapping studies. The use of DIMPL can save time and labor and can make efficient use of protein samples as it can determine the minimum number of stepwise proteolytic reactions required for intercysteine cleavage. Its application can therefore support the many diverse experimental studies that assess the contribution of disulfide bridges to protein structural integrity.
Footnotes
Acknowledgments
This research was supported by the MINT Excellence stipend of the Manfred Lautenschläger Stiftung to A.M.K., a German Research Foundation grant to A.L. (DFG LA2740/2-1), and a European Commission Marie Curie Research Fellowship to A.O.R. (FP7-PEOPLE-2010-IEF, No. 275768).
Author Disclosure Statement
No competing financial interests exist.