As a mastering engineer here is how I prefer to have mixes prepared before receiving them:
No limiting, clipping, or EQ on the master whatsoever. Compression is fine, and encouraged if it's integral to the overall tone of your track but only if it's done properly and not smashed or pumping all over the place.
Short answer: compression can be ok, nothing else. If you're unsure about your compression then take it off.
Format-wise: .wav or .aiff uncompressed. 24-bit or 32-bit float is preferred (especially if intending to do a 'Mastered for iTunes' release) but 16-bit isn't the end of the world. Sample-rate should be whatever rate it was mixed at, never upsampled.
Other often overlooked things:
"Tips & Tails" - aka the lead in and fade out of your track. If you've started your track with a big transient (i.e. a kick or a crash cymbal) right on beat 1 of bar 1, some DAWs will cut it off slightly when it renders, so check that it hasn't been cut off, or provide a bit of silence at the beginning of the track to be safe. Similarly at the very end of the track make sure you haven't cut off the reverb tail or anything like that, give it a little extra room to breathe because it can easily be edited and faded back down in mastering.
Snap, Crackle, Pop - not necessarily distortion or clipping, but watch out for the occasional click or pop from a buffer error, or from a piece of audio/sample that has not been cut at a zero-crossing. A good mastering engineer will notice these and have tools to remove/repair them as transparently as possible on the master, but fixing it in the mix will ALWAYS be have better results for these things.
