Column lifetime

[DaveM]

Hello,

I work in a non-GLP facility, carrying out method development for a range of analytes, from small molecules, through oligonucleotides, to whole proteins. As such I use reverse phase, size exclusion, and ion exchange and will potentially bring in HILIC columns in the future.

I've been using these columns fairly extensively for a while now, and the chromatography has stayed constant (I've been following the instructions!). Is the number of injections the usual way of monitoring column lifetime, or is it the volume or flow of solvent, or a combination of both? Does it change for each type of column?

Finally, is it predictable? Most of the answers I've found so far are on the lines of 'it depends', mostly on the kind of sample I'm using. Is it the most common practice (in GLP and non-GLP) to run a certain number of injections then change the column, or to run it until the chromatography changes?

With thanks,

DM

[danko]

Yes it depends! Some column degradation is caused by the mobile phase chemistry, for instance pH (too low/too high). In such a case it’s time and volume of mobile phase flown through that is important. In the more physical part of the scale, the temperature could be mentioned as an important factor.
Finally and especially if some of the sample deposits on the column/stationary phase and stays there, then the number of injections would be a dominating factor.

So, you’ll need to evaluate all the relevant factors and make a decision based on that evaluation. Regarding the way of monitoring the column’s performance I am a SST enthusiast. If you define relevant SST parameters and limits you have all the good reasons to keep using one given column, until the SST tells you otherwise.

Best Regards

[DaveM]

Thankyou Danko, that figures. Good to know that both ways are useful too. I'm using a size exclusion column beyond the parameters stated in the data sheet, after checking with the vendors and effectively getting a response that said 'it should be OK but we've not tested it, so it's up to you'. Which is fair enough.

It looks like I do a form of SST without knowing that it has formal name. Almost certainly not sufficient for regulatory guidelines (so perhaps it's not SST after all), but it acts as a check on the column performance.

DM.

[ScottHorn]

I noticed you said that you used a size exclusion column. Which column are you talking about? Also which do you use for large proteins, and how many injections have you gotten out of it? The reason I ask is because I'm using a Tosoh SuperSW3000 to analyze antibodies (mostly polyclonal) and conjugated antibodies. I've been through two guard columns w/ roughly 250 injections each. This most recent guard column change didn't help things nearly as well as the first, so I'm wondering whether I'm premature in suspecting the column iteself.

[danko]

Hi ScottHorn,

You’re right in thinking of the column (and the guard for that matter) as a possible factor. The compounds you’re working with can bind non- specifically (secondary/hydrophobic interactions) which can cause quite a rapid deterioration of the stationary phase and thereby the chromatography. How do you experience the deterioration?
A revision of the mobile phase could be helpful in some situations, but finding more hydrophilic stationary phase could be one of the best solutions.

Best Regards

[DaveM]

ScottHorn,

I'm using a Waters Biosuite 250 UHR SEC column, 4um, 300x4.6mm. I've only put 50 or so injections through it so far, and of those, 30 or so have been at these extended conditions (1.5M NaCl and 8% organic to try and disrupt protein-protein interactions; the protein recovery is 95% so I don't think I'm promoting too many interactions with the column itself). Which is a long-winded way of saying that I don't yet know how long it's going to last.

DM.

[ScottHorn]

Danko:
From what I've seen since I made that post I believe the column was on it's last legs. We replaced it today and there was a dramatic improvement. Based on the symptoms, my hunch is that there was a dead space at the top of the column. My retention times were increasing gradually, split peaks or shoulders, increasingly broad peaks, and whereas previously the uv signal dropped to zero within 15 minutes, it'd started taking almost thirty to settle down. I'm thinking my mobile phase had gradually eaten the coating off the silica and allowed it to start dissolving. All this was after about 500 injections, and I'd just put on my third guard column.
As far as mobile phase changes, I recently made a post about my initial "scouting" attempts. I had decided to use a 100mM phosphate buffer w/ 0.4M NaCl in order to disrupt ionic interactions between my sample and the column. With this most recent column I decided to try replacing the NaCl with sodium perchlorate as per this article:

http://www.biocompare.com/Articles/Appl ... olumn.html

Though they state that sodium perchlorate is aggressive towards the bonded phase, I've seen other sources that claim that sodium chloride is bad for the hydrophilic bonded phase in question. Any ideas as to which is more correct? Am I killing my column with this perchlorate? Finally, you mention using a more hydrophilic stationary phase. Can you think of an example? I don't want to go back to using superdex and lose the speed and resolution of this column.

DaveM:

Could you show me a chromatogram obtained using your setup? I want to compare it to the results from my column. Also, 1.5M NaCl seems quite high. Did you see an improvement in your chromatography by using that? As I said above, I had to replace my column after about 500 injections and two guard columns, but I probably should've done so earlier. I talked to a Phenomenex rep today that claimed I could get over 1000 injections using their equivalent of this column. That sounds a little fishy to me, but they also offered 50% off the price of the column, so maybe I'll give it a shot.

[tom jupille]

To put in perspective, check out Ron Majors' article on column usage in the January LC-GC North America:
http://chromatographyonline.findanalyti ... =&pageID=5

[danko]

ScottHorn,

The trick is not to have a bonded phase. My feeling is as follows: You’re killing the column by washing the bonded phase away which exposes your antibodies to the bare silica which is somewhat hydrophobic and the rest is obvious……….
Regarding the superdex column you’ve tried; which one was it? Because speed is tightly related to the column format (length etc.). As for the resolution; there are different particle size superdex media. Which one did you use? If you choose smaller particle size media, you’ll experience quite a boost of the resolution.

Finally, maybe ionic interactions (according to your suspicions) are not the greatest problem with your setup, but hydrophobic ditto, perhaps.

Best regard

[ScottHorn]

My concern with damage to the bonded phase was also related to the dissolution of the silica in the aqueous buffers I use. I agree that a polymeric stationary phase has a lot of advantages (ruggedness, inertness, etc).

The superdex column I've used, for analytical purposes, was the superdex 10/30 GL (superdex 200, 10mmx30cm, 13um average). At a flow rate of .4 ml/min one injection takes about an hour. Raising that flow rate leads to unacceptable resolution. The only other particle size of superdex 200 that I'm aware of is the superdex 200 prep grade, which I use for large prep columns. It's got a 34um average particle size. I'm unaware of any polymeric media smaller than superdex 200 which will separate in the size ranges I work with. That's why I went with the Tosoh column. Better resolution, 4x the speed, but with all the problems mentioned above.

[danko]

Hi ScottHorn,

For analytical purposes I’d recommend the Superdex 200, 5/150 (Tricorn), which is shorter and thus enables shorter analysis time. This column is also of smaller diameter (5 mm) which is nearly the perfect column thickness for analytical purposes in SEC context. I’m practically sure that you’ll achieve both higher resolution and faster runs if you choose this column over the Superdex 10/30, which I would choose for more preparative (in small scale) tasks.
You might also like to look at the Superose material, which provides a wider fractionation range if the separation is inadequate due to exclusion or total inclusion of some of the analytes.
Finally, one should keep in mind that the resolution is not only a function of the stationary phase/column but of the mobile phase as well (along with other more hardware related factors). So, looking into the mobile phase chemistry etc. would be a natural optimization step – in my opinion anyway.

Best Regards

[Gerhard Kratz]

Hello,

I work in a non-GLP facility, carrying out method development for a range of analytes, from small molecules, through oligonucleotides, to whole proteins. As such I use reverse phase, size exclusion, and ion exchange and will potentially bring in HILIC columns in the future.

I've been using these columns fairly extensively for a while now, and the chromatography has stayed constant (I've been following the instructions!). Is the number of injections the usual way of monitoring column lifetime, or is it the volume or flow of solvent, or a combination of both? Does it change for each type of column?

Finally, is it predictable? Most of the answers I've found so far are on the lines of 'it depends', mostly on the kind of sample I'm using. Is it the most common practice (in GLP and non-GLP) to run a certain number of injections then change the column, or to run it until the chromatography changes?

With thanks,

DM

Hi Dave,
Do you have for example Tween 80 in your sample? Such detergent will bond to the sationary phase. Try SPE to get rid of Tween 80. Otherwise number of injections is low, with or without guard column. Specifications of both mentioned column brands are the same, so the problems are the same. BR Gerhard

[HW Mueller]

Welcome back, Gerhard!

Some of you who have been around for some time might remember that I had a permanent deterioration of resolution in a Tosoh Super SW after it was exposed to TFA. I never found out what caused it. These columns seem to be very touchy, but their performance is really super. I almost completely separated a monoclonal antibody from its Fab fragment (after the TFA there was a bit more overlap), while a German manufacturer of polymer SEC columns got a, as they called it, "separation" of the Mab and its Fab in the form of widening of the peak (the mixture of Mab and Fab gave a peak slightly broader than the Mab alone, no sign of two peaks, whatever).

[DaveM]

ScottHorn,

I can't put a chromatogram up (or rather I'm not allowed to, sorry), but the high [NaCl] did not improve the look of it anyway. What it did achieve was a separation of two components that stuck together in lower salt. So the protein sizes were such that the complex ran essentially to the same place as the larger component (it's a wide peak, eluting near the void volume, separating from something late on). Analysis of this larger peak post-column showed that the smaller component was removed at high salt but not a low salt.

Gerhard,

I have DDM (dodecyl maltoside) in the sample. Maybe that will also bind the stationary phase. It's in there because it's a hepatmeric membrane protein I'm trying to isolate and it's not happy without some sort of detergent. It will be a problem if that is damaging the column as I also have one those bosses who thinks that one column will do everything, and will also do it for ever.

DM.

[Gerhard Kratz]

Dave, it`s a long time ago but I know that problem . SEC columns based on silica were developed to separate proteins. HPLC column manufacturers are producing packing material and pack it into columns. They are not producing MABs. Beside detergences there are a lot of other compounds used to stabilize MABs, PEG proteins etc., and the column manufacturer don`t know these compounds. If your detergent is absolutely necessary for the isolation of your membran protein, I only can recommend to clean the column very carefully, a wash step after 5 injections. Please use this column, once you injected your sample with the detergent, ONLY for this application. Or try an SEC column based on polymethacrylate. IF it is for isolation only and the resolution is still good it may serve you. Ask the manufacturer for a cleaning procedure to get rid of the detergence bonded to the surface of the silica packing.
Do you use a guard column? if yes, good, if not, also good, because it will not help in that case. Gerhard