How do we make a billion different antibodies?

Why so many? Well if there are half a million different species of bugs, and they each have a number of different molecules on the surface, and our immune system needs to be able to recognise them to fight them.....

Do we have a billion different antibody genes?  Well no actually.  For a lot of proteins one gene is transcribed into one messenger RNA molecule which makes one protein. Occasionally a messenger RNA molecule can be spliced differently and so would make more than one kind of protein.  But even so we still wouldn’t have enough DNA in our genome to make a billion different antibodies (Let alone all the other proteins we need to build our cells).

So how do we do it?  Pick and mix!  In a process called gene rearrangement we mix multiple different gene segments to make the final gene. 
One antibody is made of two identical copies of heavy chain and two identical copies of light chain.
The heavy chain variable region is made of three gene segments, the light chain variable region is made of two gene segments. See the picture below for how it works:


So in the genome there are a finite number of antibody gene segments.  The picture represents a cut down version of the heavy chain gene locus.  On the top line there are 17 genes.  However, because the antibody gene is made up of one red (IGHV), one green (IGHD) and one blue (IGHJ) gene mixed together randomly, then the real number of different antibody genes you can make with the 6 red, 5 green and 6 blue genes is 6x5x6 = 180.

In reality there are approximately 46 red, 27 green, 6 blue, this produces  7452 different possible heavy chains. In a similar way, but without the green (D segment) 421 different light chains can be made.
So when a heavy and a light chain are recombined randomly we get 7452x421 = 3,137,292 different antibodies.
Not only that, but when the gene segment join together they don’t do it precisely, and there is another enzyme which randomly inserts nucleotides into the joins. The enzyme is called TdT for short (Terminal deoxynucleotidyl transferase) and adds “N nucleotides”.  Plus when a B cell is activated it changes its antibody further by mutation in order to improve the fit of the antibody against the antigen it is reacting to.  (A process called the germinal centre reaction).  So now you can see how we get our estimates of a billion!  Of course we haven’t actually counted them all yet, but all indications so far say this estimate is in the right ball park.

Detailed gene tables can be found on Marie Paul Lefranc’s IMGT website.

Dunn-Walters’ Lab
Faculty of Health and Medical Sciences, Duke of Kent Building, University of Surrey,
Guildford, GU2 7XH
d.dunn-walters[at sign]