Conversion factor: 3 bp per amino acid
Convert exons 1 through 4 into amino acid residue lengths: 8 - 3 - 2 - 6.
Now, proteins can exist as different isoforms via alternative splicing. In alternative splicing, you can imagine it as taking some exons out. They tell you that the isoforms exist as 16 or 17 amino acid residues long. There are only two possible combinations of exons here that give you 16 or 17 and they are:
8 + 3 + 6
8 + 2 + 6
These correspond to exons:
1 + 2 + 4
1 + 3 + 4
Now, the idea is to take the transcripts (after splicing has occurred) and make the cDNA so that you get two different pieces of cDNA - one corresponding to exons 1-2-4 and one to exons 1-3-4. Then you just do electrophoresis to separate the two.
Now for the reasons why B-D are wrong:
B) If you're using a probe for exon 3, you're not going to detect the first piece of mRNA above (1-2-4).
C) Genomic DNA is not spliced so you're just going to amplify the region 1 through 3 uniformly.
D) Again with the genomic DNA.