How Did They Figure It Out

[apricimo]

There is kind of a catch 22 going on. In order to figure out what a gene is people use databases and find homologies in their sequence of interest. How did it start?

 

For example, how did they go about figuring out that haemoglobin is haemoglobin? So, there's mRNA floating about in our erythrocytes, ok great, but how can someone have the slightest idea that this mRNA codes for a protein that binds oxygen. I just don't get how one can go about figuring out for what a piece of mRNA codes.

 

Let's say I isolate mRNA from some bacterium and then I say ok what now. Without using any databases how do I go about figuring out what this mRNA could be.

 

There's all these proteimics and bla bla but they draw from a starting point.

 

Can someone tell a step by step way of taking an organism islolating a cell and its mRNA and then identifying what those genes code for. That would be awesome!

[Mr Skeptic]

We understand transcription, where codons (sets of 3 RNA bases) code for certain amino acids. We can also find what amino acids a protein is made of. From there we can work backward and see what RNA sequences would make that protein, and match them to the DNA.

 

You are right though, that knowing what sequence of amino acids a protein is made of does not yet tell us their function. The Folding@Home project is trying to rectify that. Once we get that understanding, we could start to design proteins for specific functions.

[CharonY]

apricimo, the information came from various sides, not just from one point. Without the use of DNA sequence traditionally proteins were sequenced with Edman. So even without a database to begin with, you could decipher the sequence of a protein, and give it a name. Large scale sequencing projects have changed that somewhat.

The scenario described above does not make sense in the absence of sequence data as the question would be, how did you isolate that particular mRNA to begin with? And why? What is the hypothesis? What you describe is more a less a blind screening procedure and these require databases.

 

Realistically, however there are other, more likely scenarios in the absence of sequence data. The most common one is reverse genetics. It goes like this: you want to identify e.g. the gene for a given function. You make e.g. a transposon mutagenesis and screen for cells that lose this function, due to disruption with that transposon. As the transposon has a known sequence, you can proceed to identify the DNA region in which it jumped into (e.g. by direct sequencing or Southern). Once you got the region, you can sequence it, based on the sequence you can deduce the protein sequence and as you already know its function (due to the method of identifying it) you can give it a nice name and put it in a database for everyone to see.

[Psycho]

A lot of the research was originally done in bacteria so you can work out where the promoters and and terminators are and knock out the gene and see what happens in the mutate strain, of course this only works if it isn't a essential genes. However then you can make point mutations within the gene and see what is down regulated.

 

Once this is found out you can try the crystallise the protein and use x-ray crystallography to see its structure and eventually also do this with its substrate bound, however this normally requires a mutant strain of the gene so you can get a locked conformation.

[CharonY]

And again, this was an approach adopted after sequence databases were common. Before that it was quite a challenge to find the the position of a gene (in minutes of conjugation, for instance), not mentioning the trouble in the identification of the actual promoter...

[apricimo]

apricimo, the information came from various sides, not just from one point. Without the use of DNA sequence traditionally proteins were sequenced with Edman. So even without a database to begin with, you could decipher the sequence of a protein, and give it a name. Large scale sequencing projects have changed that somewhat.

The scenario described above does not make sense in the absence of sequence data as the question would be, how did you isolate that particular mRNA to begin with? And why? What is the hypothesis? What you describe is more a less a blind screening procedure and these require databases.

 

Realistically, however there are other, more likely scenarios in the absence of sequence data. The most common one is reverse genetics. It goes like this: you want to identify e.g. the gene for a given function. You make e.g. a transposon mutagenesis and screen for cells that lose this function, due to disruption with that transposon. As the transposon has a known sequence, you can proceed to identify the DNA region in which it jumped into (e.g. by direct sequencing or Southern). Once you got the region, you can sequence it, based on the sequence you can deduce the protein sequence and as you already know its function (due to the method of identifying it) you can give it a nice name and put it in a database for everyone to see.

 

I guess my question was not that clear. So you figure out a sequence of a protein but what makes you think it binds oxygen, or some other molecule and makes it into something else. How do you start doing assays to say what this protein does. Without databases to relate homology how would you begin to test what a protein/enzyme does?

[CharonY]

Again, you are starting from the wrong assumption. Sequence based functional analysis was made possible by sequence databases. Before they were present you had to do e.g. reversed genetics, as mentioned above. In other words, you figure function first and then proceed to identify the sequence in question.