You can clone your gene for expression of cellulase downstream to your TEVp cleavage sequence, i.e., Strep-tag-TEVp cleavage site-Cellulase.

The first point that I would note for you is that pET-52b doesn't contain a TEV site! It has a 3C protease site (which is a very good protease - the advantage over TEV is that it is very active at 4 C, and so you can perform the digestion in the fridge and protect your protein). If you have to have TEV, then you will need to engineer in the TEV site onto the N-terminus of your protein. 3C is a good though!

For your application, you will really want the minimum cloning artefact on your final, digested protein. I would therefore recommend that you try to engineer the final construct so that your protein of interest (POI) start codon follows the Gly just after the 3C protease site. I would therefore incline to use the XmaI site (don't use SmaI for restriction cloning - it leaves a blunt end), to give ...GGACCCGGGATG... (underlined bit is your POI). For the 3' end of the gene, probably the BamHI or SacI sites would be fine. You need to include the stop codon from your POI so that there is no artefact at the C-terminus.

If you are able to - my lab usually now uses the Gibson assembly method for pretty much all cloning. Usually we use NEB's "NEBuilder" kits (e.g. https://www.neb.com/products/e2621-nebuilder-hifi-dna-assembly-master-mix). It is a bit more expensive than restriction cloning, but is fantastically robust and reduces the numbers of failed experiments a lot. You also have the flexibility to engineer into the vector wherever will work best for your experiment, rather than wherever the restriction site is.

Hope this helps!