Each target is going to be different, and in addition the design constraints for the different engineering platforms will differ. CRISPRs have a much smaller recognition sequence than TALENs or ZFNs, so may offer more opportunities, at the cost of greater off target. Sigma Aldrich (declaration of interest: my employers) offers ZFNs (declaration of interest: I use them daily for cell genome engineering!) with a design density averaging about every 50bp, and they are validated, so that might be a good option. We also make CRISPRs, which are cheaper and effective, but the customer needs to validate them.
Efficiencies of the nuclease platforms will give workable efficiencies, typically in excess 10% cleavage, resulting in ~1% of clones having two alleles disrupted by insertion or deletion (and in your case a frameshift is not necessary!). Critically important here is having an assay that allows to detect the desired genetic change. (To paraphrase: If a gene gets targeted in the forest and nobody hears it, did it ever get disrupted!?)
Finally, an important overriding consideration for many is the licensing of these technologies. Beware: down stream commercial applications may be restricted by using some platform technologies.
CRISPR is much easier to conduct than TALEN and there is also study indicated that the former had a higher knockout efficiency than the latter. There are a bunch of articles talking about the design and efficiency.
use CRISPR- check out Invitrogen product options. also look into use of miRNA mirVana inhibitors- alternative solution (not complete inhibition of course, + temporary - but fast and easy approach)