Bioinformatics Advance Access originally published online on December 16, 2004
Bioinformatics 2005 21(8):1308-1310; doi:10.1093/bioinformatics/bti227
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Structural drift: a possible path to protein fold change
1Howard Hughes Medical Institute, University of Texas Southwestern Medical Center 5323, Harry Hines Blvd, Dallas, TX 75390-9050, USA
2Department of Biochemistry, University of Texas Southwestern Medical Center 5323, Harry Hines Blvd, Dallas, TX 75390-9050, USA
*To whom correspondence should be addressed.
 |
Abstract |
|---|
 TOP Abstract REFERENCES |
|---|
Summary: Along with their mutating sequences, protein structures change in time. Analyzing a formate dehydrogenase domain that is evolutionarily related to ferredoxin, but simultaneously contains all the structural elements of a ß-Grasp fold, we illustrate here a mechanism termed as structural drift, by which changes to a protein fold can occur.
Contact: grishin{at}chop.swmed.edu
It is commonly accepted that spatial structures of proteins are more conserved than their sequences. Many evolutionarily related proteins with pronounced structural resemblance and undetectable sequence similarity are known (Andreeva et al., 2004). Often, structural features remain preserved even after sequence signal is lost to mutations, insertions and deletions. With more protein structures being solved and compared with each other, we are beginning to see many exceptions to this general rule (Andreeva et al., 2004; Grishin, 2001). Moreover we are starting to catalogue possible evolutionary scenarios that lead to significant structural changes between homologues (Grishin, 2001). Here we discuss an example of how such changes can occur.
The modular and repetitive nature of protein structures has been well characterized by Thornton and colleagues as the Russian doll effect, namely when smaller substructures are contained within larger structures (Swindells et al., 1998). In proteins that display the Russian doll effect, there are multiple levels of increased structural complexity. The protein at every new level is formed by the addition of secondary structural elements to the core of the previous level (Swindells et al., 1998). However, the common minimum core (smallest doll) always remains the same by definition, and thus the fold does not change.
As an alternative to this description, one can imagine an evolutionary scenario in which addition of a few elements to a protein structure may partially or completely disrupt the old core and form a distinct subdomain that includes the newly added elements along with parts of the old core, while not the entire old core like in the Russian doll effect. This newly formed subdomain may share structural similarity to other evolutionarily unrelated proteins. We call such a phenomenon structural drift. A protein experiencing structural drift can be described as a hybrid of two overlapping subdomains, which also have overlapping elements from their respective cores. The first subdomain constitutes the old protein core and can share sequence, structural and functional similarities with its homologues. The second subdomain covers several newly added elements and a part of the old core, and thus overlaps with the first subdomain. This subdomain may show analogous structural resemblance to other folds. Structural drift is a special case of gregariousness, a property described by Harrison et al. (2002) that measures how many other folds have a significant structural overlap with a given fold irrespective of evolutionary reasons. We illustrate the structural drift concept by analyzing the ferredoxin domains of formate dehydrogenase (FDH) small subunit (Raaijmakers et al., 2002) and speculate about its evolutionary implications.
Ferredoxins are small Fe–S proteins functioning as electron carriers in a variety of redox reactions. Ferredoxins are present in all organisms and adopt several different structural folds. The bacterial ferredoxin fold (Andreeva et al., 2004) has a two-layer 
+ ß architecture and is made of four ß-strands and two -helices. This – ß plait structure (Orengo et al., 2003) is one of the most common folds in nature. In addition to 4[Fe–S]-ferredoxins, this fold is seen in a functionally diverse set of proteins encompassing 48 superfamilies from the Structural Classification of Proteins (SCOP) database 1.65 (Andreeva et al., 2004).
The 214-residue small subunit of FDH binds three 4[Fe–S] clusters and consists of two homologous ferredoxin domains (Raaijmakers et al., 2002) originated by a duplication event (George et al., 1985) (PDB Entry 1h0h
[PDB]
, chain B, Fig. 1a, b and e). In addition to the ferredoxin core structural elements (ß-strand a, -helix A, ß-hairpin bc, -helix B and ß-strand d, Fig. 1a), the first domain contains a C-terminal extension. This extension is comprised of a long -helix followed by a ß-hairpin (elements C, e and f). The second ferredoxin domain (Fig. 1b) is inserted inside the first domain between ß-strand c and -helix B (Fig. 1a, c and e), circularly permuted with respect to the first domain, and consists of just the ferredoxin core secondary structural elements (BdaAbc). Sequence, structural and functional arguments (Raaijmakers et al., 2002) leave little doubt that these two ferredoxin domains of FDH are evolutionarily related (i.e. homologous) and the elements aAbcBd form their evolutionarily conserved functional cores that bind 4[Fe–S] clusters. However, the C-terminal extension of the first domain provokes a few thoughts.
|
This C-terminal extension (Fig. 1a, elements C, e and f shown in white; Fig. 1e traced in black thick lines) together with two ß-strands from the ferredoxin core (Fig. 1a, ß-strands a and d; Fig. 1e, red thick lines) forms a substructure (Fig. 1c, colored yellow and blue) that has a pronounced similarity to proteins from the ß-Grasp fold. The structural core of ß-Grasp (ubiquitin-like) fold is made up of two ß-hairpins arranged to form a mixed ß-sheet and connected by an
-helix: ß2 ß2 (Fig. 1d). A structure similarity search with the first domain of the FDH small subunit (all elements including the C-terminal extension, but not the second ferredoxin domain, Fig. 1a and c) against a representative set of SCOP domains using the program MAMMOTH (Ortiz et al., 2002) finds hits to itself (Z = 20.6), other ferredoxin domains (Z = 5.67.8) and the ß-Grasp fold domains (Z = 5.7). MAMMOTH Z-scores for the hits to homologous ferredoxins and structurally similar ß-Grasp are comparable. The ß-Grasp-like core of the FDH small subunit (Fig. 1c, colored yellow and blue) can superimpose with the immunoglobulin light chain-binding domain of protein L (1hz6
[PDB]
, Fig. 1d) with an RMSD of 1.27 Å over 52 C atoms from all four ß-strands (2.34 Å over 84 C atoms, including residues from the -helix). This superposition is comparable to that of the second ferredoxin domain onto the first one, which can be superimposed over 62 C atoms with an RMSD of 1.19 Å (1.90 Å over 86 C atoms when residues from strands a and d are included). As demonstrated by structural searches and superpositions, the first domain of the FDH small subunit is gregarious (Harrison et al., 2002) and shows almost equal structural similarity to the ferredoxin fold (aAbcBd) and to the ß-Grasp fold (adCef). The cores of both folds overlap within the FDH structure, namely, ß-strands a and d belong to both cores simultaneously. Since clear homology exists between the ferredoxin cores (Raaijmakers et al., 2002), the structural similarity to ß-Grasp fold is analogous and probably fortuitous due to the overlap with the ferredoxin core over the ß-hairpin ad, i.e. aAbcBd is the evolutionary unit with Cef being an extension that disrupts the aAbcBd ferredoxin unit by recruiting the ad ß-hairpin to form a ß-Grasp core adCef.
It is also noteworthy that the domain delineation program PUU (Holm and Sander, 1994) defines the first domain of the FDH small subunit to cover the elements of a ß-Grasp fold rather than a ferredoxin fold. This definition is mostly due to the well-defined hydrogen bonds between the ß-strands a, d, e and f that constitute the ß-sheet in ß-Grasp (Fig. 1e, ß-strands shown as thick lines and ball-and-stick representation with hydrogen bonds highlighted as orange dashed lines). In contrast, the ß-sheet of the ferredoxin fold is irregular (Fig. 1a, colored yellow and blue) and ß-strands a and c are not hydrogen-bonded in both ferredoxin domains (Fig. 1a and b). Such absence of hydrogen bonds is characteristic of the entire superfamily of 4[Fe–S] ferredoxins, although in other superfamilies of the ferredoxin-like fold (e.g. ribosomal protein S6) ß-strands a and c are hydrogen-bonded to form a regular four-stranded ß-sheet. The second ferredoxin domain of the FDH small subunit is a typical representative of the 4[Fe–S]-ferredoxin superfamily. It lacks the hydrogen bonds between the ß-strands a and c, but all four ß-strands (a, b, c and d) are positioned roughly in the same plane (150°) (Fig. 1b). The mutual orientation of the ß-hairpins ad and bc is different in the first FDH small subunit domain. The presence of the C-terminal ß-hairpin ef twists the ad ß-hairpin by 
45° and causes it to move further away (3 Å) from the bc ß-hairpin (Fig. 1a) as compared to the second ferredoxin domain (Fig. 1b), and the two ß-hairpins of the ferredoxin core (ad and bc) are now positioned at an angle (105°) rather than being in the same plane.
Thus, the presence of the C-terminal extension structurally disrupts the evolutionary ferredoxin core of the first domain making it less regular than that of the second ferredoxin domain. Such a disruption of the ferredoxin core is compensated by the formation of a regular ß-Grasp core (Fig. 1c and e). We call this change a structural drift from a classic ferredoxin domain to a ß-Grasp domain by insertion (Cef), incorporation of the insertion into the core, and formation of a new core resulting in the distortion of the former core. One could imagine a further evolutionary development, in which other elements of the ferredoxin fold (bc ß-hairpin, -helices A and B) may deteriorate after functional diversification of the domain. The resulting domain will then display only the elements of the ß-Grasp core. Such a domain will be structurally analogous and evolutionarily unrelated to the ß-Grasp fold proteins, since it contains parts of its relative, a ferredoxin domain.
|
Acknowledgments |
|---|
We are grateful to Lisa Kinch and Sara Cheek for the helpful comments on our manuscript. This work was supported by NIH grant GM67165 to N.V.G.
Received on October 21, 2004; revised on November 22, 2004; accepted on December 13, 2004
|
REFERENCES |
|---|
|
TOPAbstractREFERENCES |
|---|
Andreeva, A., Howorth, D., Brenner, S.E., Hubbard, T.J., Chothia, C., Murzin, A.G. (2004) SCOP database in 2004: refinements integrate structure and sequence family data. Nucleic Acids Res., 32, D226–D229
Esnouf, R.M. (1999) Further additions to MolScript version 1.4, including reading and contouring of electron-density maps. Acta Crystallogr. D Biol. Crystallogr., 55, 938–940[CrossRef][Medline].
George, D.G., Hunt, L.T., Yeh, L.S., Barker, W.C. (1985) New perspectives on bacterial ferredoxin evolution. J. Mol. Evol., 22, 20–31[CrossRef][Medline].
Grishin, N.V. (2001) Fold change in evolution of protein structures. J. Struct. Biol., 134, 167–185[CrossRef][ISI][Medline].
Harrison, A., Pearl, F., Mott, R., Thornton, J., Orengo, C. (2002) Quantifying the similarities within fold space. J. Mol. Biol., 323, 909–926[CrossRef][ISI][Medline].
Holm, L. and Sander, C. (1994) Parser for protein folding units. Proteins, 19, 256–268[CrossRef][ISI][Medline].
Orengo, C.A., Pearl, F.M., Thornton, J.M. (2003) The CATH domain structure database. Methods Biochem. Anal., 44, 249–271[Medline].
Ortiz, A.R., Strauss, C.E., Olmea, O. (2002) MAMMOTH (matching molecular models obtained from theory): an automated method for model comparison. Protein Sci., 11, 2606–2621
Raaijmakers, H., Macieira, S., Dias, J.M., Teixeira, S., Bursakov, S., Huber, R., Moura, J.J., Moura, I., Romao, M.J. (2002) Gene sequence and the 1.8 A crystal structure of the tungsten-containing formate dehydrogenase from Desulfovibrio gigas. Structure (Camb.), 10, 1261–1272.
Swindells, M.B., Orengo, C.A., Jones, D.T., Hutchinson, E.G., Thornton, J.M. (1998) Contemporary approaches to protein structure classification. Bioessays, 20, 884–891[CrossRef][ISI][Medline].
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  L. Xie and P. E. Bourne Detecting evolutionary relationships across existing fold space, using sequence order-independent profile-profile alignments PNAS, April 8, 2008; 105(14): 5441 - 5446. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Cheng, B.-H. Kim, and N. V. Grishin MALISAM: a database of structurally analogous motifs in proteins Nucleic Acids Res., January 11, 2008; 36(suppl_1): D211 - D217. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Shi, Y. Zhong, I. Majumdar, S. Sri Krishna, and N. V. Grishin Searching for three-dimensional secondary structural patterns in proteins with ProSMoS Bioinformatics, June 1, 2007; 23(11): 1331 - 1338. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Y. Niv, D. R. Ripoll, J. A. Vila, A. Liwo, E. S. Vanamee, A. K. Aggarwal, H. Weinstein, and H. A. Scheraga Topology of Type II REases revisited; structural classes and the common conserved core Nucleic Acids Res., April 1, 2007; 35(7): 2227 - 2237. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||