..The pairing of different combinations of Ig heavy and light chains.

..Recombination of V, D and J segments (VJ for light chains)

Together these potentially generate some 5x10⁶ different antibodies.

Added Value

For the heavy chain the functional coding frame is fixed for V segments and for J segments (ie each have only a single useful frame. D segments however can be read in more than 1 frame. As the human locus has 25 functional D segments, the total possible number of D products would be 75 if all D segments could be read in all frames. In fact only 53 possible reading frames exist (that is there are in-frame stop codons within the other 22 D-frames). That is why the number of possible heavy chain V regions is 16218 (51 x 53 x 6) and not 7650. It also means that instead of 1/3 rearrangements being in-frame the true number is 1/4.25 (23.5%).

..Variability in the joins of the recombined DNA segments.

This arises from

a) Imprecise joining by the recombinational machinary

b) the addition of extra random nucleotides by Terminal deoxynucleotide transferase (N region addition, not seen in light chains)

..Somatic hypermutation, a poorly understood mechanism for introducing mutations into V regions of activated B cells (antigen driven).

Antibody molecules show the greatest sequence diversity at the V -> C junction. This forms the CDR3 and is encoded largely by D and J gene segments and thus benefits from both recombinational and N region diversity.

Added Value
One of the students doing the 1997/8 course asked me a couple of questions about Ig rearrangement and gene expression, You might find them interesting.
The questions were:

Q.

After VDJ gene rearrangement at the IgH locus is completed there seem to remain (dependent on splicing) several copies of (different) J segments at the end of the now complete VDJ heavy chain. - ie it is more like a VDJJJJ chain where the number of Js depends on the D->J rearrangement. - For example in the picture the rearranged sequence is shown to be V(H1) D(H26) J(H5) J(H6) C(H1).
So my question is what happens after the VDJ rearrangement. - Does transcription of the VDJJC (using above example) proceed as one piece and is then maybe spliced to VDJ(H5)C before expression or does the final protein actually contain two J segments?

Q.

A second question which follows logically from this is what happens when there are several V segments at the beginning: eg. If V -> DJ rearrangement occurs so that the result is V(H1) V(H2) D(H26) ...
Is a transcript of VVD... made and V(H1) is spliced out or how is this regulated?

You can see my answers here.

Information provided by: http://www-immuno.path.cam.ac.uk