The Applied Biosystems (ABI) series of protein sequencers use reverse-phase chromatography to separate the derivatized amino acids. The typical separation consists of gradient elution using two solvents; solvent A is tetrahydrofuran (THF) in water with sodium acetate added to adjust the pH and ionic strength, and solvent B is acetonitrile (MeCN). Careful adjustment of the solvent A ionic strength, pH, and gradient parameters are necessary for successful separation of most of the naturally occurring amino acids. Recently, ABI introduced PreMix (1), which includes an ion-pairing reagent in addition to sodium acetate, designed to improve the peak shape and enhance the reproducibility of the retention time of PTH-His, PTH-Arg, and PTH-pyridylethyl-Cys (PE-Cys). ABI also suggested changing solvent B from 100% MeCN to 88% MeCN/12% isopropanol to separate PTH-Trp from diphenylurea (2). In our laboratory, we implemented both changes and, after minor adjustments to the standard conditions, were satisfied with the resolution of the PTH-amino acids (Figure 1).
Most columns lasted six months, or almost 2,000 injections, before resolution deteriorated. The amount of PreMix added to solvent A for these columns at the end of their lifetime was 20 ml. We recently began using a new lot of ABI PTH C-18 column for sequence analysis using a 477A protein sequencer. With this column, after only one month (450 injections), we had increased the amount of PreMix from 15 ml to 30 ml per liter of solvent A but still were not achieving adequate resolution of the protonated amino acids PTH-His, PTH-Arg, and PTH-PE-Cys from PTH-Ala, PTH-Tyr, and PTH-Pro, respectively (Figure 2). We received a new column with the same lot number as the previous column, but after two months (550 injections) we were in the same situation: 35 ml PreMix/liter solvent A was needed to separate all PTH-amino acids.
It was at this point that I turned to the ABRF electronic mailing list to ask if anyone else in the group had experienced this particular technical problem, and how it was solved. Len Packman (Cambridge University) suggested that we dilute the Premix buffer levels and return to 15 ml/l. This will have the effect of placing Arg beyond Tyr and His beyond Ala and will result in a stable separation pattern for quite some time. He speculated that there are sites on the column that require saturation with a component of the Premix before the chromatography becomes stable, and this can only be achieved at high Premix levels. Once saturated, the sites remain unreactive for quite some time, even at reduced Premix levels.
I responded that reducing the PreMix concentration would cause the PE-Cys derivative to elute in the middle of the Pro/Met/Val trio. Our laboratory was not using PE-Cys, but it is included in the ABI calibration standards.
Several people suggested alternatives to pyridylethylation of cysteine. Carol Beach (University of Kentucky) suggested alkylating cysteine with 3-bromopropylamine (3, 4), noting that PTH-aminopropyl-Cys elutes after PTH-leucine. John Hempel (University of Pittsburgh) suggested I try N-isopropyliodo-acetamide (NIPIA) to derivatize cysteine to carboxamidopropyl-cysteine (CAP-Cys) in place of iodoacetamide or vinylpyridine. He noted that the PTH-derivative elutes in a nice open area of the chromatogram between PTH-Tyr and the PTH-P/V/M cluster.
This derivative had been used in our laboratory, and I had noted that PTH-CAP-Cys elutes with a retention time nearly identical to PTH-PE-Cys. Due to the lack of a PTH-CAP-Cys standard, I did not know if itwould also elute in the P/M/V cluster. To test these suggestions, I reduced the amount of Premix in solvent A until the His and Arg were eluting after Ala and Tyr, respectively, and then tried sequencing a peptide with the sequence KCTCCA that had been alkylated with NIPIA using the method of Krutzsch et al. (5). I was pleasantly surprised to find that while PE-Cys is sensitive to Premix concentration, the CAP-Cys derivative is not. A chromatogram from one cycle using the reduced-PreMix gradient is shown in Figure 3. This chromatogram was generated when the PTH-CAP-Cys from KCTCCA and calibration standard mixture were co-injected to show that PTH-CAP-Cys elutes between PTH-Arg and PTH-Pro. It also shows the reversal in elution order for H/A and R/Y.
John Hempel noted that the PTH-CAP-Cys would not be expected to be sensitive to ionic strength because it does not protonate, as the PTH derivatives of Arg, His and PE-Cys do. It was his understanding that the variability in elution of these derivatives is due to gradual erosion of alkyl groups from the HPLC column, leaving SiO3 groups to ion pair with basic side-chains. By increasing ionic strength as the column ages, this interaction is suppressed.
There are many variables to optimize for successful sequence analysis of proteins and peptides. The use of Premix buffer in solvent A, isopropanol in solvent B, and alkylation of cysteine residues with NIPIA provides a method of identifying all common amino acids. The ABRF electronic mailing list provided a means by which easy and rapid exchange of ideas helped solve this particular problem. The problem might have been solved by trial and error, but in this case, I was able to draw on the experience of others to save many hours in front of the HPLC. Special thanks to Len Packman, John Hempel, and Carol Beach for their assistance in this matter and to all others not mentioned here who participated in this electronic dialogue.
References
--------------------------------------------------------------------- Date: Mon, 12 Dec 1994 From: William Chestnut (Burroughs Wellcome Co.) [much text deleted] My question is if anyone else had the same experience with this C-18 column? The column Lot# = R8-716-004 Buffer A = 3.5% THF inH2O with Premix added Buffer B = 12% IPA in Acetonitrile with DMPTU (noacetone) Column Temp = 53 degrees C. FAST and NORMAL gradients both haveHis/Ala, Arg/Tyr resolution problems. --------------------------------------------------------------------- Date: Tue, 13 Dec 94 From: Len Packman (Cambridge University) Regarding high Premix volumes for ABI PTH columns........ I saw a similareffect on batches of columns over a year ago but not on recent batches(currently using lot 49036). We operate a 477 and use standard sol B (noiPrOH). The way round this is a little counter-intuitive. I recommend thatyou now DILUTE OUT the Premix buffer levels and return to 15 ml/l. Thiswill have the effect of placing Arg beyond Tyr and His beyond Ala, butthat's a workable situation and in my experience will result in a STABLEseparation pattern for quite some time. On the worst column we had, it wasnecessary to get to 25 ml/l then switch to 12.5 ml/l. Over a period oftime, drift in His and Arg demanded a rise in Premix again, but once at 25ml/l, the volume was dropped back to 12.5 ml/l and there was no furtherdrift after that. I think that there are sites on the column that requiresaturation with a component of the Premix before the chromatography becomesstable and this can only be achieved at high Premix levels. Once saturated,the sites remain unreactive for quite some time, even at reduced Premixlevels. --------------------------------------------------------------------- Date: Tue, 13 Dec 1994 From: William Chestnut (Burroughs Wellcome Co.) My records show that you are using an old lot of ABI column (I used minefrom that lot in Feb 1994). We have switched to iPrOH to separate Trp fromDPU.[The way round this is a little counter- intuitive. I recommend that you nowDILUTE OUT the Premix buffer levels and return to 15 ml/l. This will havethe effect of placing Arg beyond Tyr and His beyond Ala, but that's aworkable situation and in my experience will result in a STABLE separationpattern for quite some time.]That may be true but unacceptable to those of us who identify Cys bypyridylethylation (PE-Cys) or other derivatives that elute just before Pro.When adjusted using the method you suggest, the Cys derivatives tend tocome out in the middle of the Pro/Met/Val trio. --------------------------------------------------------------------- Date: Fri, 16 Dec 1994 From: John Hempel (University of Pittsburgh) Re: the message above, if you will try N-isopropyliodoacetamide in place ofiodoacetamide or vinylpyridine, as described by Krutzsch & Inman, AnalBiochem 209:109-116, 1993, I think you will be pleasantly surprised thatyour problems with Cys-derivative elution positions become moot. The PTHruns in a nice open area between Tyr and the P/V/M cluster. --------------------------------------------------------------------- Date: Fri, 16 Dec 94 From: Carol Beach (University of Kentucky) Bill Chestnut might try alkylating cysteine with 3-bromopropyl- amine(R.A.Jue & J.E.Hale, Anal. Biochem. 210 (1993) 39-44 and for "on sequencer"alkylation see Jue & Hale in "Techniques in Protein Chemistry V", JohnCrabb, ed., 1994, Academic Press, pp. 179-188). The PTH-aminopropyl-Cyselutes after PTH-leucine and interferes with only PTH-norleucine. If youuse PTH-norleucine as in internal standard (as I do) then PTH-norvalineseems to be a decent substitute. It elutes between valine and DPTU. Iassume these relative elution positions hold for systems using Premix. --------------------------------------------------------------------- Date: Fri, 16 Dec 1994 From: William Chestnut (Burroughs Wellcome Co.) As a matter of fact, CAP-Cys (carboxamidopropyl-Cys) runs right before Pro,but I am going to try the 10 ml Premix w/CAP-Cys to see if it ends up underPro the same way PE-Cys does. If not, then that's the ticket. I will letthe list know as soon as I find out. --------------------------------------------------------------------- Date: Thu, 22 Dec 1994 From: Bill Chestnut (Burroughs Wellcome Co.) Hello all; I would like to thank all the people who responded with theirthoughts and experiences concerning separation of PTH aa's on the ABIsequencers. Although skeptical, I tried the scheme that many out there havebeen using. I reduced the amount of Premix in solvent A until the His andArg were eluting after Ala and Tyr, respectively, and then tried sequencinga peptide that had been alkylated with N-isopropyliodoacetamide (to giveCAP-Cys). I was pleasantly surprised to find that while PE-Cys is sensitiveto Premix concentration, it appears that the CAP-Cys derivative is not. Itelutes between PTH-Arg and PTH-Pro and is even separated from one of theThr by-products. Alkylating with NIPIA is as straightforward asiodoacetamide, so this will be our method of choice. --------------------------------------------------------------------- Date: Thu, 22 Dec 1994 From: John Hempel (University of Pittsburgh) Glad you are pleased with NIPIA. The PTH would not be expected to besensitive to ionic strength since it does not protonate, as PTH Arg, Hisand PECys do. It has been my understanding that the variability in elutionof these derivatives is due to gradual erosion of alkyl groups from the LCcolumn, leaving SiO3 groups to ion pair with basic side-chains. Byincreasing ionic strength as the column ages, you suppress thisinteraction. ---------------------------------------------------------------------
Return to the The ABRF Home Page