Problem with IP-RP HPLC column

[kenna_s]

Hi all,

I'm using silica column (Varian Helix). Some time ago 'something wrong' started to happen - peaks from the shortest fragments of DNA standard (pUC18 digested with HaeIII) became small and doubled. I washed the column with 0.5M disodium EDTA according to Varian's instruction and after that the column completely broke down. Washing with ACN-THF-ACN composition didn't help.
Do you have any idea:
1) what happened?

I wonder if it could be due to any contamination in the water used to analysed PCR and EDTA solution making (detergents from dirty glass bottle?) Or maybe I destroyed the column with too high pH (that's the suggestion of my colleagues, who work on Waters silica columns)? My buffers are quite old, with measured pH=7.4 (whereas Varian specifies its buffor pH 7.0), EDTA solution pH=8.0.

2) what else I can do?

Thanks.

[rhaefe]

I don't think that the pH was too high unless you flushed it with Disodium EDTA solution for days at elevated temperature. You should definitely use fresh buffers.
Did you try to reverse the column? Partially blocked and contaminated inlet frits can cause peak broadening and splitting. Surfactants from polymerase cocktails (Tween 20 is common) can be adsorbed onto the stationary phase and cause peak splitting. If this is the case usually a flush with 75% acetonitrile for 30 minutes solves that problem.

Since you are using the Varian Helix column I assume that you use Varians Helix system as well and this one is (if memory serves me well) a bio-compatible system. If not and the system is a regular stainless steel system, the column can be poisoned by very small amounts of metals (mainly Fe and Cr) which can also cause peak splitting. In such a case the system should be passivated with nitric acid and the column replaced.

[kenna_s]

Did you try to reverse the column? Partially blocked and contaminated inlet frits can cause peak broadening and splitting.

No, not yet. I'll do it, but I'm affraid it's too late now, after the fatal EDTA washing. 'Peak broadening and splitting' - that's OK, but dividing one peak on two??? I've never seen that before. I checked the plasmid, it is OK for sure.

If this is the case usually a flush with 75% acetonitrile for 30 minutes solves that problem.

I flushed with 75% acetonitrile at the very begining, nothing changed so I tried pure acetonitrile.

Since you are using the Varian Helix column I assume that you use Varians Helix system as well and this one is (if memory serves me well) a bio-compatible system. If not and the system is a regular stainless steel system, the column can be poisoned by very small amounts of metals (mainly Fe and Cr) which can also cause peak splitting. In such a case the system should be passivated with nitric acid and the column replaced.

Yes, I'm using Varian Helix system. Nevertheless, I suspected metal ions contamination, that's why I used EDTA.

I'll change the column eventually, but first I want to find the answer "what's wrong with my system or my reagents" in order not to broke the new one.

Thanks a lot for your help.

[tom jupille]

A few random points:

1. How old was your column (and how many injections had it seen)? Columns *are* a consumable; yours may have died a natural death

2. A partially-plugged inlet frit can cause split peaks (or shoulders, or tailing, or . . . ); it all depends on what the obstruction does to the flow profile at the column inlet. In general, if it's an obstruction, you'll see the same pattern on all peaks in the chromatogram.

3. Trying to save money by stretching the lifetime of buffers is a false economy (like not changing the oil in your car!).

[kenna_s]

1. How old was your column (and how many injections had it seen)? Columns *are* a consumable; yours may have died a natural death

I made 480 injections before any problems started, what means about 140 hours of column work. Not many, in my opinion. Varian specifies 1000 injections as an expected lifetime of that column. BUT, I used it for dHPLC analysis, at the temperatures 50 - 64 C degree.

2. A partially-plugged inlet frit can cause split peaks (or shoulders, or tailing, or . . . ); it all depends on what the obstruction does to the flow profile at the column inlet. In general, if it's an obstruction, you'll see the same pattern on all peaks in the chromatogram.

So, that must have been somenthing else.
Backflushing didn't change anything, except for small pressure decrease. I'll change the column. Buffers too.

Thanks a lot again.

[rhaefe]

What fragments were affected? I still would run the column in reverse and see if this solves or minimizes the problem. What Tom said is correct, usually all peaks are affected by a partial obstruction, however these signs can sometimes be hard to detect on later eluting peaks, which spend more time on the column. I have seen cases were the first two or three (18, 80 and 102bp) fragments of pUC 18 Haee III were split but not the larger ones (102 too 587bp).