I am looking into trying to characterise a protein secreted from a gram positive bacteria. I have started by aligning the sequence to a number of strains of the particular species to see which residues are conserved. What kind of things do you look for in a protein sequence (i.e.residues or repeated codons), that might suggest its toxicity?
Interesting .... never heard this question asked before but it is a good one.
"How can you tell a protein is toxic from its sequence?"
The problem is there are many modes of toxicity:
1) Toxicity by catalytic action
2) Toxicity by affecting membrane permeability
3) Toxicity by mimicry of host proteins or other compounds
4) Immunotoxicty
5) Allergo-toxicity
6) Modulation of gene expression by mimicry/competition of gene regulators
7) Mimicry of DNA or RNA precursors
There are probably a thousand or more variations on 1) alone .....
So your question is actually complicated but the process of taking your best guess is simple.....
1) BLASTP your protein and see what hits come up. Overall conservation is not as important as regions of high conservation or motifs. This may give hints on where to start.
2) Send your sequence to the BTXpred server
http://www.imtech.res.in/raghava/btxpred/
All prediction servers are risky to have much faith in, but give it a try as it may lead you in the right direction.
3) Look for hydrophobic regions with repeats indicative of a protein that can form pores in membranes. If your protein is pure try adding it to some blood from your host (metal cofactors may be required)
4) Purify your protein, de-endotoxin it and add it to a cell line from your predicted host. (Do another protein purified from the same organism you think is not toxic for a negative control). A negative result here is not definitive as your expressed protein may be misfolded or require cofactors for activity
Interesting .... never heard this question asked before but it is a good one.
"How can you tell a protein is toxic from its sequence?"
The problem is there are many modes of toxicity:
1) Toxicity by catalytic action
2) Toxicity by affecting membrane permeability
3) Toxicity by mimicry of host proteins or other compounds
4) Immunotoxicty
5) Allergo-toxicity
6) Modulation of gene expression by mimicry/competition of gene regulators
7) Mimicry of DNA or RNA precursors
There are probably a thousand or more variations on 1) alone .....
So your question is actually complicated but the process of taking your best guess is simple.....
1) BLASTP your protein and see what hits come up. Overall conservation is not as important as regions of high conservation or motifs. This may give hints on where to start.
2) Send your sequence to the BTXpred server
http://www.imtech.res.in/raghava/btxpred/
All prediction servers are risky to have much faith in, but give it a try as it may lead you in the right direction.
3) Look for hydrophobic regions with repeats indicative of a protein that can form pores in membranes. If your protein is pure try adding it to some blood from your host (metal cofactors may be required)
4) Purify your protein, de-endotoxin it and add it to a cell line from your predicted host. (Do another protein purified from the same organism you think is not toxic for a negative control). A negative result here is not definitive as your expressed protein may be misfolded or require cofactors for activity
You could take a closer look at some of the tools created by The Center for Genomic Epidemiology. They have tools for detecting pathogenicity and virulence. Hope this helps. If you have a phenotype for different genotypes of the protein, then you I recommend you take a closer look at SigniSite: http://www.cbs.dtu.dk/services/SigniSite/
http://www.genomicepidemiology.org/
So you want to "characterise a protein" by "aligning the sequence to a number of strains of the particular species to see which residues are conserved"? You mentioned your protein is gram positive - how about other strains you tried to align to? If your sequence data have gram positive and gram negative groups, there are many things you can test.
by Multiple sequence alignment of different strains of particular species using clustalW or Muscle alignment, you may find some conserved pattern in group of sequences. for predicting toxicity, you can use ToxinPred tool. on this conserved sequence and moreover can compare it with wet lab data(preliminary) to design new experiments for lowering toxicity, if the tool is able to predict the wet lab activity of sequences..
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics