Footprinting Proteins

Protein footprinting is a methodology to gain information about protein conformation and interactions by probing the structure of polypeptide chain that is labeled at one end. The protein of interest is end-labeled and reacted with a probe in such a way that susceptibility of a site on the protein to the probe leads to breakage of the polypeptide backbone nearby. The sites of backbone scission are deduced from the lengths of peptides containing the end label, which can be estimated readily by SDS-PAGE. Alteration of the backbone cleavage pattern indicates that the end-labeled protein has undergone a conformational change and/or interaction with other macromolecules. The principle is the same as that for footprinting nucleic acids, which has been used widely. Protein footprinting has only recently become feasible with the advent of techniques to end-label proteins.

 Information on scission sites or modification sites can be obtained by peptide mapping as well, but footprinting has three advantages:

1. Facility Peptide mapping requires purification of individual cleavage products and amino acid analysis and/or the N-terminal sequence of each. Footprinting determines the lengths of many peptides, and hence many cleavage sites, in a single operation without isolating individual fragments.

2. Sensitivity Determination of the N-terminal amino acid sequence of a peptide typically requires 10 pmol of peptide. Footprinting is generally more sensitive; for instance, radiolabeling has virtually no detection limit.

3. Selectivity In large multisubunit complexes, peptide mapping will be complicated by the plethora of peptides to be purified and analyzed. In footprinting, information is obtained only on the polypeptide chain that was end-labeled.

1. Methods to End-Label Proteins

Currently there are three ways to end-label proteins: 

1. Chemical modification of the N-terminus (1, 2). The protein is first subjected to one round of Edman Degradation. All amino groups become phenylthiocarbamylated, except for the one at the very N-terminus, which was originally the imino group of the second amino acid residue before the Edman reaction. This unique amino group can be reacted with either a radioactive or a fluorescent reagent, resulting in labeling of the N-terminus. This method inevitably destroys the folded structure of the labeled protein prior to probing, and information can be obtained only at the level of the primary structure.

2. Specific antibody binding to either terminus (3, 4). In this method, an antibody raised against a peptide comprising the amino acid sequence of either terminus is used to detect peptides containing the terminus. A more convenient alternative is to place an epitope, against which an antibody is already available, at one end of a protein and to express the epitope-tagged protein through recombinant DNA techniques.

3. Phosphorylation at either terminus (5, 6). This is based on the sequence specificity of protein kinases, such as heart muscle kinase, which is less stringent than antigen-antibody recognition. An amino acid sequence that can be phosphorylated by a kinase is attached to a protein at the gene level. The site is labeled using the kinase with [g-³²P] or [g-³⁵S] ATP.

2. Methods to Probe Protein Structure

 As is apparent from the principle of nucleic acid footprinting, structure probing has to result in scission of the polypeptide backbone, either directly or indirectly. There are common advantages and disadvantages for each type of method. In direct probing methods, probing immediately leads to backbone scission. An advantage of direct methods is their simplicity in comparison with indirect ones, which involve additional reactions. One major drawback is that scission of the peptide backbone often renders the fragments generated more susceptible to ensuing scissions, particularly when the scission has occurred within a folded structural unit. This causes a problem, especially in quantitative analyses, when the first cleavage sites have to be identified. In indirect methods, chemical modification of a side chain is followed by scission of an adjacent peptide bond (or prevention thereof) after one or more subsequent steps. Because modification of a solvent-accessible side chain in one entire protein molecule is unlikely to perturb the folded structure extensively, or to enhance the reactivity of other residues of the same molecule, indirect methods should not suffer from the problems that direct methods do.

Hitherto, five probing methods, two direct and the other three indirect, have been devised and employed: 

1.One direct method is proteolysis (4-6). Endoproteinases have been extensively used to delineate the domain structures of proteins, because well-folded structural units are more resistant to proteolysis than regions linking them. Interaction with other macromolecules may block some cleavage sites. In addition, conformational changes are often reflected in changes in susceptibility of sites cleavable by proteinases. Although proteolysis has been useful, the biophysical basis of the interaction between the probing proteinase and the probed protein molecule is not entirely understood.

2. The other direct method uses oxygen radical, typically generated by the Fenton reaction, to cleave the backbone (7). It is likely that the reactivity of the oxygen radical with the peptide backbone is more directly related to solvent accessibility than is that of proteinases. The extent of side-chain damage by the oxygen radical per backbone breakage is not known.

3. One indirect method utilizes a combination of reversible and irreversible modifications of lysine residues (8). The protein is first subjected to limited citraconylation. Removal of the citraconyl groups after irreversible acetylation of the remaining lysines recovers unmodified lysyl residues only at the sites where the first modification was introduced. Complete digestion of these polypeptide chains by a lysine-specific endoproteinase generates fragments that end at a first modification site.

4.A second indirect method utilizes oxidation of methionine residues, which blocks cleavage at such residues by cyanogen bromide (9). Oxidation of individual methionines is observed as a decrease in the amount of peptide fragments generated by cyanogen bromide. In such cases where modification prevents cleavage, the extent of cleavage must be carefully controlled to ensure that all observable peptides will not be reduced to the shortest peptide containing the end label.

5. The last method utilizes cyanylation of cysteine residues by 2-nitro-5-thiocyanobenzoic acid (10). Raising the pH induces slow cleavage of the peptide bond at the S-cyanocysteine, thus turning the modified side chain into a cleaver. Because the cyano group can transfer to a free sulfhydral, ie, an unmodified cysteine within the same peptide, during the slow cleavage, unreacted cysteine residues have to be blocked beforehand, eg, with N-ethylmaleimide.

The latter three methods were devised on the basis of side-chain modification schemes developed in the 1960s and 1970s. More of them may be revived to create new footprinting methods.

3. Variations and Future Prospects

 Various tagging technologies have been used widely in molecular and cellular biology. This indicates that use of a label is not limited to detection of peptides. For instance, some end labels can also serve to isolate terminal fragments. DNA-binding sites of a protein were delimited by immunoprecipitating peptide fragments with a terminal epitope tag, following transfer of radioactivity from DNA and partial proteolysis of the radiolabeled protein (11). Although SDS-PAGE has a fairly good resolution, other methods of determining molecular weights of proteins, such as mass spectrometry, can also be employed after peptides are separated with the use of a tag. In another example of use of a tag other than labeling, a protein kinase site was used to identify protein–protein interactions that blocked phosphorylation of the site (12).

 A complex of iron and EDTA, Fe-EDTA, that generates oxygen radical can be crosslinked to specific sites on macromolecules (13). This attachment of a cleavage center onto a macromolecule rather than remaining free in solution allows mapping in the vicinity of the cleavage center (13), when combined with end labeling of its interaction partner, be it a protein molecule or a DNA molecule.

 Because the principle of footprinting is the same for proteins and for nucleic acids, much parallel can be drawn from the longer experience of nucleic acids footprinting. The interference experiment scheme (14), in which macromolecules are first subjected to modification and later separated according to their remaining functionality, has yet to be applied to proteins.

References

1. D. G. Jay (1984) J. Biol. Chem. 259, 15572–15578. 

2. R. A. Jue and R. F. Doolittle (1985) Biochemistry 24, 162–170. 

3. P. Matsudaira, R. Jakes, L. Cameron, and E. Atherton (1985) Proc. Natl. Acad. Sci. USA 82, 6788-6792.

4. J. E. Lindsley and J. C. Wang (1991) Proc. Natl. Acad. Sci. USA 88, 10485–10489. 

5. R. Hori, S. Pyo, and M. Carey (1995) Proc. Natl. Acad. Sci. USA 92, 6047–6051. 

6. V. Nktinis, J. Turner, and M. O''Donnell (1996) Cell 84, 137–145. 

7. T. Heyduk and N. Baichoo, and F. Heyduk (2001) Metal Ions Biol. Syst. 38, 255–287. 

8. R. Hanai and J. C. Wang (1994) Proc. Natl. Acad. Sci. USA 91, 11904–11908. 

9. M. V. de Arruda, H. Bazari, M. Wallet, and P. Matsudaira (1992) J. Biol. Chem. 267, 13079–13085.

10. B. P. Tu and J. C. Wang (1999) Proc. Natl. Acad. Sci. USA 96, 4862–4867. 

11. N. F. Lue, A. Sharma, A. Mondragon, and J. C. Wang (1995) Structure 3, 1315–1322. 

12. P. T. Stukenberg, J. Turner, and M. O''Donnell (1994) Cell 78, 877–887. 

13. S. A. Datwyler and C. F. Meares, (2001) Metal Ions Biol. Syst. 38, 213–254. 

14. U. Seibenlist and R. B. Simpson, and W. Gilbert (1980) Cell 20, 269–281. 

15. T. M. Rana and C. F. Meares (1991) Proc. Natl. Acad. Sci. USA 88, 10578–10582. 

16. U. Siebenlist, R. B. Simpson, and W. Gilbert (1980) Cell 20, 269–281.