Reads falling in between exon and intro may have to be automatically trimmed.

Reads in that region will fail to map if you are only aligning against CDS sequences in your reference. Aligners will generally soft clips reads (if the alignment ends partway in the read).

What exactly are you wanting to do?

It is doable with BWA or Bowtie2. The main issue is that you'll need to tweak the gap parameters to assign a very low penalty to gaps, since I guess that some of your CDS are entire mRNA without the introns.

In BWA you might be interested in the following parameters:

• -O INT Gap open penalty. [6]
• -E INT Gap extension penalty. A gap of length k costs O + k*E (i.e. -O is for opening a zero-length gap). [1]
• -L INT Clipping penalty. When performing SW extension, BWA-MEM keeps track of the best score reaching the end of query. If this score is larger than the best SW score minus the clipping penalty, clipping will not be applied. Note that in this case, the SAM AS tag reports the best SW score; clipping penalty is not deducted. [5]

Source: http://bio-bwa.sourceforge.net/bwa.shtml

Starting from there, I think it is just trial and error to find the correct parameters