I have always had good luck with BamHI, and the HF version I suspect should be even better. Never had a problem with XhoI either that I can recall, just NdeI -- it may be due to something else but it had made me paranoid about trying it in the future, as of now I wouldn't knowingly use it to design a cloning strategy, but good luck anyways, can you use something other than NdeI? If it were me I and doing something different was not a real option I might even try blunting the NdeI site(s) and have BamHI at one end and a blunt end at the other, which I have done before and worked very well (the blunt end was via SmaI though).
I am curious about the NdeI -- awhile back in the middle of doing some more difficult cloning, I went ahead and made a deletion clone of a construct I had using NdeI, which should be the absolute easiest cloning you could possibly do, I was cutting out a chunk of DNA using just NdeI and religating (there was two sites in the insert). The digest looked perfect on a gel, and gave the expected two bands, and I gel purified the larger fragment and religated it -- literally about as simple as you can get as far as cloning -- and the first surprise was very few transformed colonies (should get TONS doing this) and second surprise was of the colonies I did get, none were correct, some strange plasmids (not the original clone either). I tried this again, and was extra careful and the same result -- since this was not anything critical I never spent time trying to figure out what was wrong, but I have done alot of cloning and never saw something like that with something seemingly so "easy". It makes me wonder about 1) using NdeI for cloning or 2) maybe some bizarre problem with the 2 bp AT overhangs and gel purification. Obviously I did try using lowered ligation temps (and using PEG which is already in the ligation buffer I use). But there is something weird there, and like someone already mentioned I am paranoid about using NdeI for cloning (although I use it for restriction digestions for sizing and it works fine).
Hmm... looks like you aren't the only one being paranoid about Nde1.
I have BamH1(HF) in my lab and it cuts pretty well. I thought about switching BamH1(HF) to Xho1 and just use Nde1 and Xho1 for my digestion. But I am wondering if I should stick to BamH1(HF) as other people in my lab have told me that BamH1(HF) cuts pretty well and change Nde1 to another enzyme.
I do want to use BamH1(HF) too, for it works really well. But looking at the vector map, I will have to stick to using Nde1 as it will affect my gene orientation if I were to switch Nde1 to something else.
Actually, if you are using PCR to amplify the fragment you are cloning, you can still use BamHI and just leave the other end blunted. So, you can just phosphorylate the opposite PCR primer (near your NdeI site) and digest the PCR product with BamHI and clone it that way (you'll have to blunt the vector too, for example baqsed on your vector, cutting with XbaI and blunting it, or you can PCR amplify the vector as well). Having one sticky end and one blunt end is actually still pretty easy to clone, because the correct clone is still heavily favored in the reaction. Done it many times and you'll get fewer colonies than you would with two different sticky ends but you still should get plenty and most of them should be what you want (very low background). Sometimes you have to get creative like this when you run into these problems (many times its easier to just order new primers, all things considered!) Warren..