We are trying to separate a group of unknowns. Their structures (4-5 compounds) include a small peptide (up to 8 mers) bonded through a cysteine thio-ester linkage to a fatty acid. So far we have looked at reverse phase C18 (4.6x250mm, 5um)using water + 0.1% TFA (A) and 1:1 CH3CN:IPA + 0.1% TFA. The gradient is 85% B to 100% over 20 mins then another 20 mins at 100%B. Compounds are eluting in a poorly resolved group between 25 and 35 mins. We have a number of ideas for a next step that might include swapping B for THF+0.1% TFA, swapping A for a buffer, or even going to a new system (normal phase). On the TLC, we get reasonable separation on silica with CHCl3/MeOH/H2O (65/25/4). Can anyone suggest an alternate approach for this separation ? we dont have much experience with peptides.

Acc to my experiences with peptides, phosphate buffers separate much better than TFA buffer. ACN seems to be the best choice for 215 nm detection because of stable gradient baseline. Try the real flat gradient - 5% change in 60 min. It works! I like Kromasil (Eka Nobel), Luna CN (Phenomenex) or Discovery RPAmideC16 (Supelco) columns. The peptide separation is the challenge. Good luck.

Peptides normally need a very shallow gradient to be adequately resolved. Try to use no more than 0.1% change/min in your conditions. I agree with another message posted here about phosphate buffers, they seem to separate better than TFA.

You may also want to try a different source of columns. We have seen drastic differences between suppliers. Zorbax and Diazem columns have been good for us.

Another approach is to change the TFA for a higher acid, such as Pentafluoropropionic acid, however, this could retain your compounds too long.

Good luck

The good news is you have a lot of variables to play with. The bad news is . . . you have a lot of variables to play with. The *really* bad news is that there's no "magic bullet" and no guarantee that what worked for someone else's peptides will work for yours! (Having said that, the advice given in both of the preceding posts is sound!)

Sticking with the reversed-phase column, the variables (in rough order of ease-of-use) are:
- gradient steepness / shape
- temperature
- organic solvent
- buffer type
- pH
- column type
(yes, I realize that other people may rank these differently).

A good rule of thumb for method development is:
1. Pick a variable
2. Change it (a lot!)
3. If the selectivity (spacing of the peaks) changed, then spend some time exploring intermediate values.
4. If the selectivity didn't change, then set that variable at a reasonable value and go on to the next one.

At LC Resources, we have software (DryLab) that lets you model the effect of these variables, given appropriate calibration experiments. If you're interested, e-mail me privately and I'll use it to take a look at your separation.

-- Tom Jupille / LC Resources