Protien separation retention time shifts

[KarenJ]

Hi,

I have a method for measuring a 98KDa protein in buffer solutions. I was using a C8, 300A, 5um, 2.1mm x 25 cm column. Column temp = 40C. Mobile phase is a gradient from 30% to 80% B. MP A is 0.1% TFA in water, MP B is 0.1% TFA in ACN. When I initially developed this method a couple of months ago, my retention time was very solid at 16.2 min. for every injection every day. However, now, my retention time is variable and on average is getting later.

Two things had happened that I thought might have caused this problem. First, I had to switch the instrument to normal phase for a week. After I switched back and replaced the pump seals, the retention time was 17.3. Second, I started assaying some filtered whole broth samples which had a lot more proteins and salts in them. The peak of interest began shifting later until it was 20.3 min.

At this point, I tried cleaning the column according the manufacturer's recommendations and the retention times. The retention times briefly improved back to 17 min. but quickly began shifting later again. I tried a new column C8 and still saw the retention times shifting later, even with the cleaner samples of just protein in buffer.

Since I felt that the method performance had changed so much that I needed to repeat my validation, I decided I might as well try a new column. It was recommended that a C4 column might be better for a protein this large. I now am using a C4, 300A, 5um, 2.1mm x 25 cm column with all other conditions being the same. My results are the same as with the old C8 column. The peak of interest began coming out around 19 min. In two weeks it has drifted to 22 min. Any changes that I make to the gradient cause the peak to be even later - I can't seem to make it come out earlier. When I have very long runs, I can see the peak move as much as 2 minutes (e.g. from 20 to 22 minutes) during the run. This has really messed up the integration method because the peak moves out of the window.

Does any one have any ideas about why this peak keeps shifting to later times and what I can do to stabilize it?

Sorry for the very long message.

KarenJ

[tom jupille]

First question: was the original C8 column "end-capped" by any chance?

Short-chain bonded phases are notoriously unstable at low pH (the short alkyl chain is not hydrophobic enough to be kept in place if the silyl ether linkage holding it to the silica matrix is hydrolyzed). End capping is usually done with trimethylsilyl groups (C1). As that comes off, your protein sticks to the newly-exposed silanols.

In general, stability at low pH follows the order C18 > C8 > C4 > C1.

If this were my problem, I would go back to a C8, but check with various vendors (I'm sure some of them will chime in here ) to get a recommedation for a column stable at low pH.

[Uwe Neue]

When you used the new C8 column, was the retention back to 16.2 minutes or thereabouts?

I agree with Tom. My first suspicion is also that the column is aging, and I would not switch to the C4 column. This could be especially true, if you stored the column in TFA during the time when you did other stuff.

[KarenJ]

Hi,

Thanks for the responses. My C8 columns are Vydac 208TPC8 columns that are endcapped. The C4 column is a Vydac 214TPC4 column that is also endcapped. When I injected my protein on the new C8 column the retention time was 18.3 min. and the area was about 2/3 of the area that I was getting with the old C8 column. The area with the new C4 column was just slightly lower than the area with the old C8 column.

The columns end up sitting with 70% water, 30% ACN, 0.1% TFA after assays are completed because I haven't added a separate wash to my sequence.

One other thing to mention. The old C8 column was not new when I first developed this assay. This was a used column that the customer gave me. I believe that it had been used for other proteins before this but I don't know. However, retention times and peak areas were very stable and I had a good validation package started before this trouble began.

I am also using a small C8 guard cartridge in the front of the analytical column. I have replaced this guard cartridge since I started having the problems. I did not have a C4 cartridge when I started this, so I kept the C8 in place hoping that there was not enough packing material in that little cartridge to significantly affect the separation and that the cartridge would at least help protect the analytical column.

Thanks so much for your help.

KarenJ

[Mark Tracy]

Under the mobile phase conditions you describe, you will lose some of the endcapping when the column is new, then it will stabilize. The long alkyl chains will come off more slowly. For best column life, don't store it in TFA-containing mobile phase.

[Uwe Neue]

With your next column, wash out the TFA if the column is not used. If you do not use it for a week, I would put it in 100% acetonitrile.

[KarenJ]

I'm sending an overlay of the chromatograms collected using my new C8 column (first image). Is this variability really normal for a new column even if it has been exposed to TFA for a week or two? I didn't see this type of variability with the old column (second image) until after the events that I described even though the old column had been exposed to TFA for a couple of months.

Thanks,

Karen

[KarenJ]

Oops - here is the chromatogram from the old column. This is the way my method was performing before the events that I described earlier. The old chromatograms (shown below) are from standards of three different concentrations. Precision was good and linearity was good. The chromatograms from the new column (previous post) are out of the same vial. Precision is terrible and linearity is terrible. Any suggestions would be greatly appreciated.

Thanks,

KarenJ

[tom jupille]

If your library keeps LC-GC, look up the following article:
Marchand, Zhu, and Dolan; LC-GC 14(12) 1028 (1996)
The chromatograms are almost a dead-on match to the problem you're seeing. The underlying causes were pump malfunction along with the narrow-range gradient problem I alluded to in another thread (http://www.sepsci.com/chromforum/viewtopic.php?t=4110).

If you can't find the article (LC-GC's on-line archives only go back to 2000), e-mail me and I'll send you a pdf version.

[KarenJ]

Hi Tom,

Just wanted to say Thank You for your help. We have found that we did have a small leak in our pump. We've changed the seals which has helped. I also have some new outlet ball valves ordered. Hopefully this should take care of it. The article that you recommended was really good.

Thanks,

KarenJ

[HW Mueller]

Karen, in your original statement you mention that cleaning the column resulted in a short reprieve. It wouldn´t do this if you ruined it via TFA or if you had a permanent leak. When I did HPLC with proteins in the matrix (not protein chromatogr.) I had to clean the column at least once a day (runs during day time only, cleaning before shutdown in the evening, keeping methanol in the column over night could increase backpressure, as, apparently, not all the proteins had always come off). It seems that part of the problem is that you don´t clean the column often enough. (Also, can´t you do this without TFA?)