I’m not sure how you’re trying to do the experiment. Are the enzyme sites really in the sequence as you amplify it, or are you designing primers so that the sites will be there after pcr? And then how or what kind of site-directed protocol are you trying to use? The easiest way to introduce point mutations—if that’s what you’re doing—is to use something like a QuikChange kit from Stratagene (soon to be part of Agilent). The process involves pcr with at least 2 mutagenic primers, but you shouldn’t have to digest out any fragments unless you’re doing some kind of cassette-style transfers to other vectors. It does rely on methylated/unmethylated DNA to discriminate against native, non-mutated DNA, so you have to be careful what strains you use to prepare the original template. It’s possible, I guess, that you’re shuttling between an E coli vector for propagation and a mammalian expression vector when you’re going into B-cells.

If you are excising a SalI-HindIII fragment from a pcr product with designed restriction sites at the end, then your primer design needs to include, if it doesn’t already, some extra junk bases at the ends to ensure that the enzymes cut efficiently. If the sites are actually internal to a longer pcr product, that should not be an issue. If you intend to clone/sub-clone from one vector to another using SalI-HindIII, then yes, both vectors must have the sites, or at least sticky-end compatible sites. SalI and XhoI, for example, have compatible sticky ends and you can ligate a SalI-cut piece of DNA into an XhoI site of a vector, but be warned: you cannot re-cut the product with either enzyme once you’ve done it. That may or may not be a problem, but you should be aware of things like that when you decide on your strategy.

A possible “way around” restriction site dependence is to use something like a Gateway vector from Invitrogen. There, once you clone your fragment of interest into donor vector—and you could also do your mutagenesis in the donor as well—you can transfer to any number of pre-designed acceptor vectors for expression in bacterial or eukaryotic or insect cells. Invitrogen’s system uses lambda recombinase to effect the transfer. I think Clontech’s system uses Cre/lox recombinase and a different series of donor/acceptors, but it’s the same sort of idea.

That was a bit of a ramble. I can’t say anything that specific unless I understand what exactly it is you’re trying to do.

Well, that’s clearer. But, you might want to look at the plasmids supplied with the kit. My recollection is that there is pUC18 or pUC19, which is a transfection control, though, if you grow some up you may be able to use it as a cloning vehicle. The other plasmid is a mutagenesis control plasmid that contains a mutation in the lacZ gene that produces "white" colonies on X-gal/IPTG plates. The manufacturer supplies you with two mutagenic oligonucleotides that, if everything works, will convert a cryptic stop codon to the correct amino acid and produce blue colonies on X-gal/IPTG plates. Neither were intended to be your cloning vehicle, though, as I say, you could grow up some pUC plasmid and use that if you like. As long as the propagation vector has SalI-HindIII, you should be OK shuttling that way. If the eukaryotic expression vector has XhoI-HindIII in the correct order given your regulatory sequences you should be OK there as well. Just remember that you’ll have to go back to your propagation clone every time you want to use the SalI-HindIII fragment.