Zinc is among the most common transition metals bound to proteins. It plays important role in various biological functions, such as DNA synthesis, brain development, cell growth and structure stabilization. The accurate prediction of zinc-binding proteins and zinc-binding sites from sequences are of interest to researchers in various disciplines. In proteins, zinc binds almost only to four types of amino acids, namely Cys, His, Asp and Glu (CHDE). These four types of amino acids account for ~98% of all residues that bind to zinc. Among these four amino acids, Cys and His (CH) are dominant, which account for ~84% of all zinc-binding residues. One zinc atom binds to three or four amino acid residues. Zinc bound by three amino acid residues are often catalytic zinc and those bound by four are usually structural zinc. Figure 1 shows a protein (alcohol dehydrogenases, PDBID 2ohx) with one catalytic zinc-binding site and one structural zinc-binding site. The structural zinc (Zn402) is bound by four cystines and the presence of this zinc stabilizes the protein structure. The catalytic zinc (Zn401) is bound by three amino acids (two cysteines and one histidine) while the fourth position is bound by the DMS ligand which further interacts with the NAD ligand so that the catalytic function can be carried out. Despite this simplicity, zinc-binding CHDEs take up only a small proportion of all CHDEs. The proportion of each amino acid that binds to zinc is Cys: 5.8%, His: 2.4%, Asp: 0.28% and Glu: 0.13%. Moreover, even given the 3D structure, the detection of binding sites solely from geometric criteria in proteins without bound metal (e.g. apoproteins) is non-trivial, since the residues that bind to a metal often undergo conformational changes upon binding. Due to the fact that metal-binding sites are often highly conserved, sequence profiles derived from multiple sequence alignments are usually applied in predicting functional metal-binding sites.

2. Summary

The accurate prediction of zinc-binding proteins and zinc-binding sites from sequences are of interest to researchers in various disciplines. We have developed a web server for predicting zinc-binding proteins and zinc-binding sites from sequences. In order to cover most residues of interest while enabling a reasonable prediction accuracy, we focused on predicting four types of amino acids, i.e. cysteins, histidines, asparates and glutamates (CHDE) which compose ~98% of all residues that bind to zinc. A five-fold cross-validation indicates that our server predicts zinc-binding Cys and His with 76% precision at 60% recall and on the chain level 60% precision at 60% recall.

3. Usage

Input to the server is one or several amino acid sequences (up to ) in FASTA format. The user can either paste one or more sequences in the text-area provided, or, alternatively, upload a file in ASCII format. Example input: >2az4_A mol:protein length:429 hypothetical protein EF2904 MESKAKTTVTFHSGILTIGGTVIEVAYKDAHIFFDFGTEFRPELDLPDDHIETLINNRLVPELKD LYDPRLGYEYHGAEDKDYQHTAVFLSHAHLDHSRMINYLDPAVPLYTLKETKMILNSLNRKGDFL IPSPFEEKNFTREMIGLNKNDVIKVGEISVEIVPVDHDAYGASALLIRTPDHFITYTGDLRLHGH NREETLAFCEKAKHTELLMMEGVSISFPEREPDPAQIAVVSEEDLVQHLVRLELENPNRQITFNG YPANVERFAKIIEKSPRTVVLEANMAALLLEVFGIEVRYYYAESGKIPELNPALEIPYDTLLKDK TDYLWQVVNQFDNLQEGSLYIHSDAQPLGDFDPQYRVFLDLLAKKDITFVRLACSGHAIPEDLDK IIALIEPQVLVPIHTLKPEKLENPYGERILPERGEQIVL

4. Output

The output for the example sequence can be found here

5. References

PredZinc: [Please cite this paper if you find PredZinc useful in your research] Nanjiang Shu, Tuping Zhou and Sven Hovmöller. Prediction of Zinc-Binding Sites in Proteins from Sequence. Bioinformatics, 2008;24(6):775-82 [PubMed]

5. Contact