Sequence analysis of enzymes of the shikimate pathway: Development of a novel multiple sequence alignment algorithm

Bell, Lachlan Hamilton (1996) Sequence analysis of enzymes of the shikimate pathway: Development of a novel multiple sequence alignment algorithm. PhD thesis, University of Glasgow.

Full text available as:
Download (12MB) | Preview


The possibility of homology modelling the shikimate pathway enzymes, 3-dehydroquinate synthase (el), 3-dehydroquinase (e2), shikimate dehydrogenase (e3), shikimate kinase (e4) and 5-enolpyruvylshikimate 3 -phosphate (EPSP) synthase (e5) is investigated. The sequences of these enzymes are analysed and the results found indicate that for four of these proteins, el, e2, e3, and e5, no structural homologues exist. Developing a model structure by homology modelling is therefore not possible. For shikimate kinase, statistically significant alignments are found to two proteins with known structures, adenylate kinase and H-ras p21 protein. These are also judged to be biologically significant alignments. However, the alignments obtained show too little sequence identity to permit homology modelling based on primary sequence data alone. An ab initio based methodology is next applied, with the initial step being careful evaluation of multiple sequence alignments of the shikimate pathway enzymes. Altering the parameters of the available multiple sequence alignment algorithms, produces a large range of differing alignments, with no objective way to choose a single alignment or construct a composite from the many produced for each shikimate pathway enzyme. This problem with obtaining a reliable alignment for the shikimate pathway enzyme will occur in other low sequence identity protein families, and is addressed by the development of a novel multiple sequence alignment method, Mix'n'Match. Mix'n'Match is based on finding alternating Strongly Conserved Regions (SCRs) and Loosely Conserved Regions (LCRs) in the protein sequences. The SCRs are used as 'anchors' in the alignment and are calculated from analysis of several different multiple alignments, made using varying criteria. After divided the sequences into Strongly Conserved Regions (SCRs) and Loosely Conserved Regions (LCRs), the 'best' alignment for each LCR is chosen, independently of the other LCRs, from a selection of possibilities in the multiple alignments. To help make this choice for each LCR, the secondary structure is predicted and sliown alongside each different possible alignment. One advantage of this method over automatic, non-interactive, methods, is that the final alignment is not dependent on the choice of a single set of scoring parameters. Another is that, by allowing interactive choice and by taking account of secondary structural information, the final alignment is based more on biological, rather than mathematical factors. This method can produce better alignments than any of the initial automatic multiple alignment methods used. The SCRs identified by Mix'n'Match, are found to show good correlation with the actual secondary structural elements present in the enzyme families used to test the method. Analysis of the Mix'n'Match alignment and consensus secondary structure predictions for shikimate kinase, suggest a closer match with the actual secondary structure of adenylate kinase, than is found between their amino acid sequences. These proteins appear to share functional, sequence and secondary structural homology. The proposal is made that a model structure of shikimate kinase, based on the structure of adenylate kinase, could be constructed using homology modelling techniques.

Item Type: Thesis (PhD)
Qualification Level: Doctoral
Additional Information: Adviser: James E Milner
Keywords: Biochemistry
Date of Award: 1996
Depositing User: Enlighten Team
Unique ID: glathesis:1996-71741
Copyright: Copyright of this thesis is held by the author.
Date Deposited: 17 May 2019 09:31
Last Modified: 17 May 2019 09:31

Actions (login required)

View Item View Item