アーリック™(ERLIC™)ミックス・クロマトグラフィ分離モードにおける
揮発性塩(volatile salts)でのリン酸化ペプチド(phosphopeptides)分離

新しい分離モードであるアーリック(ERLIC: Electrostatic Repulsion／Hydrophilic Interaction Chromatography)イオン的反発作用／親水性相互作用ミックス・クロマトグラフィのユニークなメカニズムは、以前に紹介しました(2007年7月掲載)。その後多くの研究者から質量分析装置などの使用に際して、脱塩を必要としない揮発性塩の要望があり、ここではその可能性について検討しました。そして今回は揮発性塩として、ぎ酸アンモニウムを使用した幾つかのオプションが試され、複数個のリン酸基を持つトリプシン酵素で消化されたペプチド断片の分離に関しても、検討していますのでご紹介いたします。


Isolation of Tryptic Phosphopeptides by ERLIC
(Electrostatic Repulsion-Hydrophilic Interaction Chromatography)

Andrew Alpert¹, Goran Mitulović², and Karl Mechtler²

¹PolyLC Inc./ 9151 Rumsey Road, ste. 180/ Columbia, MD 21045 USA/ aalpert@polylc.com
²IMP/IMBA Research Institutes/ Dr. Bohr-Gasse 3/ A-1030 Vienna, Austria

ABSTRACT

Peptides with one phosphate group cannot be cleanly separated from peptides with no phosphate by anionexchange chromatography, since the electrostatic repulsion of the positively charged termini outweighs the electrostatic attraction of the phosphate group to the stationary phase. When the column is operated in the HILIC mode, though, then the hydrophilicity of the phosphate group plus its electrostatic attraction accomplishes this separation despite the electrostatic repulsion of the termini. This combination is called ERLIC: Electrostatic Repulsion-Hydrophilic Interaction Chromatography. Selectivity for phosphate groups is ensured by operation at pH 2. A gradient to 0.2 M triethylamine phosphate elutes peptides with 1-4 phosphate groups. Retention is much greater using volatile salts such as ammonium formate, to the point that it is practical to isolate phosphopeptides via solid-phase extraction in the ERLIC mode. Unlike high-affinity methods involving titania or IMAC, ERLIC is sensitive to aspects of peptide composition besides the phosphate group. This makes it suitable as a high-resolution mode for samples containing thousands of phosphopeptides.

In general, peptides running in this mode seem to be oriented with the C-terminus facing the stationary phase. The basic sidechain of the C-terminal residue may be forming a zwitterion with the carboxyl- group, making this end less basic and so less repelled in the ERLIC mode than the N-terminus. This has implications for selectivity in HILIC and cation- and anion-exchange as well.

INTRODUCTION

Anion-exchange chromatography (AEX) has been examined as a method for enrichment of phosphopeptides. However, at a pH low enough to uncharge Asp- and Glu- residues (so as to distinguish them from phosphate groups), the electrostatic repulsion of the positively-charged termini causes the elution of tryptic fragments prior to the void volume [1], and attachment of a single phosphate group does not suffice to overcome this repulsion. Such phosphopeptides aren't well-resolved from nonphosphopeptides in AEX [2]. When the column is operated in the HILIC (Hydrophilic Interaction Chromatography) mode, though, then the considerable hydrophilicity of the phosphate group plus its electrostatic attraction accomplishes this separation despite the repulsion from the termini. This combination is called ERLIC (Electrostatic Repulsion-Hydrophilic Interaction Chromatography) [3]. A gradient from 20 mM sodium methylphosphonate to 0.2 M triethylamine phosphate elutes peptides with 1-4 phosphate groups [2]. Volatile salts such as ammonium formate are much weaker displacing agents than phosphate or methylphosphonate salts. Retention of phosphopeptides is much greater with them, to the point that it is practical to isolate phosphopeptides via solid-phase extraction (SPE). This poster explores the use of ERLIC with ammonium formate mobile phases for isolation of phosphopeptides, either via gradient elution of a column or via SPE.

MATERIALS AND METHODS

The HPLC system was a model Essence® from Scientific Systems Inc. (State College, PA). Control and data collection was via EZStart® (Scientific Software). The HPLC column used for AEX and ERLIC was PolyWAX LP™, 5-µm, 300-Å (PolyLC Inc., Columbia, MD); 100x4.6-mm unless indicated otherwise. The SPE cartridges were LooseTips (item# LT200WAX) packed with 20-µm PolyWAX LP (PolyLC).

Reagents: Phosphoric acid was HPLC-grade from Fisher Scientific while all other reagents were from Fluka. Methylphosphonic acid was purum-grade. Formic acid was puriss./HPLC grade. Triethylamine was ultra grade. All other reagents were HPLC-grade. A 1 M stock solution of ammonium formate was prepared by weighing formic acid into a beaker, adding a magnetic stir bar and water, adjusting the pH to 2.2 with ammonium hydroxide, then diluting to the mark in a volumetric flask. The pH of the final mobile phases was not measured after addition of acetonitrile (ACN). Sodium methylphosphonate (Na-MePO4) stock solution, pH 2.0, was prepared by addition of NaOH to a solution of methylphosphonic acid in water.Triethylammonium phosphate (TEAP) stock solution, pH 2.0, was prepared by addition of triethylamine to a solution of phosphoric acid in water.

Peptides: The following peptides were synthesized as described [4] using standard Fmoc solid-phase chemistry on an ABI 433A peptide synthesizer (Applied BioSystems, Foster City, CA): a) The peptide GGAAGLGY(p)LGK; b) The set of peptides with the sequence WWGSGPSGSGGSGGGK, with phosphate groups on 0-4 Ser- residues; c) The sequence variant peptides NAAAAAAWK, AAANAAAWK, AAAAAAWNK and their amidated analogs. Peptides SLYSSSPGGAYVTR (Vimentin(51-64)), SLYSSS(p)PGGAYVTR, SVNFSLTPNEIK (MAP 1B(1271-1282)), and SVNFSLT(p)PNEIK were a gift of Ken Jackson (Molec. Biol.-Proteomics Facility, Univ. of Oklahoma Health Sciences Center). Peptide TRDIYETDYYRK (Insulin Receptor (1142-1153)) and its phosphorylated analogue were purchased from Quality Controlled Biochemicals (Hopkinton, MA).

Elution with a gradient from Na-MePO4 to TEAP: Mobile phase A was 20 mM Na-MePO4, pH 2.0, with 70% ACN. Mobile phase B was 0.2 M TEAP, pH 2.0, with 60% ACN. After 5' at 100% A, there was a 20' linear gradient to 100% B followed by a 10' hold at 100% B. Flow rate: 1 ml/min.

Phosphate vs. Formate in Gradient Elution

Column: PolyWAX LP, 100x4.6-mm; 5-µm, 300-Å
Detection: A220 (blue), or A280 (red) 1 ml/min
Gradient: See Methods.
P0: WWGSGPSGSGGSGGGK
P1: WWGSGPSGSGGS(p)GGGK
P2: WWGSGPSGS(p)GGS(p)GGGK
P3: WWGSGPS(p)GS(p)GGS(p)GGGK
P4: WWGS(p)GPS(p)GS(p)GGS(p)GGGK

Fig. 1. Selective retention of phosphopeptides. At pH 2.2, the formate is 〜 97% in the form of the unbuffered acid. Unbuffered formic and acetic acids are extremely weak eluting agents; far weaker here than is TEAP.
RESULT: Far stronger retention of phosphopeptides. Presumably less ammonium formate would be needed for elution at a higher pH, but then Asp- and Glu- residues would become charged and the selectivity for phosphopeptides would be lost. Under the present conditions, the electrostatic repulsion still causes the nonphosphorylated peptide to elute in the void volume.

Fig. 2. ERLIC vs. AEX with formate buffer; variants of WWGSGPSGSGGSGGGK
Column: PolyWAX LP. A 280 Gradient: 0-5': 0% B; 5-25': 0-100% B; 25'+: 100% B
Mobile Phase: A) 20 mM NH4-Formate, pH 2.2; B) 1 M NH4-Formate, pH 2.2; with starting and final % ACN as indicated
[TOP] As expected, singly phosphorylated peptides elute near the void volume in AEX.
[MIDDLE] With ERLIC, retention of the singly phosphorylated peptide is good but retention of the triphosphopeptide is inconveniently long.
[BOTTOM] A switch from ERLIC conditions to AEX conditions during the gradient affords both adequate retention of the singly phosphorylated peptide plus more convenient elution of the more highly phosphorylated peptides.

Fig. 3. ERLIC of Phosphopeptides: Effect of Salt Mismatch
Sample: WWGSGPSGSGGSGGGK with 0-4 phosphates, dissolved in 20 mM NH4-Formate, pH 2.2, w. 70% ACN HPLC analysis: 104WX0503 column with Na-MePO4 - TEAP gradient per standard ERLIC method
RED: Sample freeze-dried and redissolved in Na-MePO4 mobile phase (MP A) before analysis (no mismatch)
BLUE: 5 µl of sample in NH4-Formate solvent mixed with 15 µl of Na-MePO4 (MP A) for injection (salt mismatch)
∴ Mismatch between counterions severely affects the peak shape of multiphosphorylated peptides

Elution of Ideal Tryptic Monophosphopeptide Standards: ↑ [Salt] vs. ↓ [ACN]

Column: PolyWAX LP Flow rate: 1 ml/min. Detection: As noted
Gradient: 0-5': 0%B; 5-45': 0-100% B; 45-50': 100%B
MP A: 20 mM NH4-Formate, pH 2.2, with 70% ACN
MP B: [TOP] 100 mM NH4-Formate, pH 2.2, w. 64% ACN;
[BOTTOM] 20 mM NH4-Formate, pH 2.2, w. 10% ACN

Fig. 4. Elution with formate buffer; alternative gradients. Increasing the salt content of the mobile phase is a standard way to elute solutes in ion-exchange, and works well here [TOP]. Selectivity and peak shapes are quite good. An alternative gradient of decreasing [ACN] was also tried, switching the mode from ERLIC to AEX, since tryptic monophosphopeptides are not wellretained in the AEX mode [BOTTOM]. The selectivity was retained although peak shapes deteriorated to some extent. The ACN gradient has two significant advantages: 1) Since the salt concentration doesn't vary, the absorbance baseline is steady. This suggests the possibility of monitoring 〜 220 nm for peptides that lack aromatic residues; the baseline would be 〜 0.4 AU but steady. 2) The use of only 20 mM ammonium formate would be quite convenient for direct flow to a mass spectrometer. Of course, a salt gradient would be necessary to elute peptides with more than one phosphate.

Fig. 5. Elution of Nonideal Tryptic Phosphopeptide Standards: ↑ [Salt] vs. ↓ [ACN]
CONDITIONS: Per Fig. 4. SAMPLE: Variants of Insulin Receptor (1142-1153)
These peptides have an extra basic residue at the C-terminus, a "ragged end", as well as one near the N-terminus. Basic residues are quite hydrophilic. This additional hydrophilic interaction led to the failure of the monophosphopeptide to be eluted by the gradient to 100 mM salt that eluted the standards in Fig. 4. The alternative gradient of decreasing [ACN] eliminated the hydrophilic interaction. The electrostatic repulsion of the basic residues now led to elution in the same time frame as the standards in Fig. 4.
Peptides with ragged ends and missed cleavages are commonly encountered in complex tryptic digests. CONCLUSION:
A decreasing [ACN] gradient is preferable to increasing [salt] for elution of singly phosphorylated tryptic fragments.

Fig. 6. ERLIC-SPE Fractionation of a Mixture of Phosphopeptides
SPE cartridge: Item# LT200WAX Sample: WWGSGPSGSGGSGGGK with 0-4 phosphates
Binding solvent: 20 mM NH4-Formate, pH 2.2, w. 70% ACN. Eluting solvents (all 10% ACN): 1) 20 mM NH4-Formate, pH 2.2; 2) 1 M NH4-Formate, pH 2.2; 3) 300 mM TEAP, pH 2.0 (desalted for HPLC)
HPLC analysis: PolyWAX LP column with Na-MePO4 - TEAP gradient per standard ERLIC method
∴ Singly phosphorylated peptides will elute with a step to 10% ACN in 20 mM NH4-Formate. Some doubly phosphorylated peptides might elute with a step to 1 M NH4-Formate. However, elution of peptides with 3 or 4 phosphates and some with 2 phosphates requires a stronger eluting salt like TEAP (or possibly KH2PO4).

Fig. 7. OBSERVATION: The elution order of the phosphopeptide standards in Fig. 4 seems to correlate with how close the phosphate group is to the C-terminus. That suggests that it's the preferred contact region. This speculation was tested with the above sequence variants.
CONCEPT: Asparagine is polar and contributes to retention in HILIC (and ERLIC is a form of HILIC). Here it serves as a reporter group on the orientation of the peptide during chromatography; the peptide that's best-retained will be the one that's oriented with the Asn- closest to the stationary phase, where it can interact most effectively and promote retention.
RESULTS: See Fig. 8

Fig. 8. Orientation Study of Nonphosphorylated Peptide Sequence Variants.
Column: PolyWAX LP, 200x4.6-mm; 5-µm, 300-Å 1 ml/min A280 Mobile Phase: 10, 20 or 40 mM ammonium formate, pH 2.2, with ACN (% as indicated) In nearly every case, the variant with Asn- at the C-terminal end elutes last. This indicates that these peptides are oriented with the C-terminal end facing the stationary phase.
NOTE: Salt shields electrostatic effects, repulsive in this case, so retention increases with [salt].

SPECULATION:

At low salt concentration, the sidechain of the Arg- or Lys- group of a tryptic peptide could form a zwitterion with the C-terminus. This zwitterion would be less basic than the N-terminus and would be less repelled in ERLIC, hence the orientation with the C-terminus down. At higher salt concentration, the concentration of counterions would be high enough to outcompete the internal charged groups for interaction, and the zwitterionic behavior would be reduced.

This speculation was tested by removing the ability to form a zwitterion; the peptide orientation standards were resynthesized with the C-terminus amidated.

 

Fig. 9. Observations: 1) Amidation of the C-terminus leads to earlier elution in every case.
2) When the C-terminus is free, the variant with Asn- at the C-terminal end elutes last in nearly every case. Amidation of the C-terminus abolishes the C-terminal orientation; all variants practically coelute except for the control missing the Asn- residue.
These data support the hypothesis that the C-terminus of a tryptic peptide is in a zwitterionic bond to some extent with the basic C-terminal residue's side chain. The side chain then repels the stationary phase less, so that end is the preferred contact region and retention is increased.

OBSERVATIONS AND PREDICTIONS:

1) C-terminal fragments from a tryptic digest should elute later than blocked N-terminal fragments as a class in either anion- or cationexchange chromatography or HILIC. With blocked N-terminal peptides the two charged groups are both at the C-terminus and might form a zwitterion that doesn't interact well with any ion-exchange material. With a C-terminal fragment the two charged groups are on opposite ends of the molecule. The peptide can assume the orientation necessary to bring a group into proximity with a stationary phase of the opposite charge. This would also permit the amino- terminus, a polar group, to approach a HILIC surface in isolation rather than in a less polar zwitterion.

2) ZIC-HILIC material (Sequant AB): This material has the following structure, and is ostensibly a neutral zwitterionic stationary phase for HILIC. It's actually a good analogue of a tryptic C-terminal end. It acts like a zwitterion < 6 mM salt, a cationexchanger between 6-20 mM salt, and a neutral material > 20 mM salt (high enough to shield the electrostatic effects). This is evident in chromatography of sialylated glycans [5] and charged pharmaceuticals [6,7]. Also, in general, C-terminal fragments do elute later from this material than do blocked N-terminal fragments [8].

DISCUSSION

Using volatile solvents in ERLIC, separation of phosphopeptides from nonphosphopeptides is easy. Phosphopeptides can also be separated from each other with high resolution, an important consideration in analysis of complex mixtures via mass spectrometry. Elution of peptides with more than one phosphate is more difficult and may require nonvolatile salts. We are presently testing the following alternative protocols with the tryptic digest of a HeLa cell lysate:
1) Fractionation with a 4-solvent gradient (20 mM NH4-Formate,70-10% ACN; 20-1000 mM NH4-Formate; 1000 mM NH4-Formate-0.3 M TEAP). Fractions will be collected and freeze-dried or desalted, then analyzed via RPC-MS.
2) Fractionation via SPE with a PolyWAX LP LooseTip with elution with the same 3 solvents; the three eluates plus the filtrate to be analyzed via RPC-MS. This is an easy protocol affording a lower degree of fractionation.
3) The eluate obtained from protocol 2 with 20 mM NH4-Formate/10% ACN will be applied to a regular PolyWAX LP column and resolved with a gradient from 70-10% ACN in 20 mM NH4-Formate, with numerous fractions to be collected and analyzed further via RPC-MS. Every component of the sample is guaranteed to elute during this gradient. Since singly phosphorylated peptides comprise 〜 82% of a complex tryptic digest [9], then it is worthwhile applying this high-resolution step to this particular SPE eluate in order to maximize the number of phosphopeptides identified.
4) The same sample as in 3) will be run on a capillary of PolyWAX LP with elution directly into a mass spectrometer, skipping the fraction collection and RPC steps.

Observations of note:
1) The fractions eluted with 0.3 M TEAP are likely to contain numerous peptides with > 3 phosphates. The limiting factor in their identification at present is not the chromatography but the capabilities of mass spectrometry.
2) Tryptic peptides (ideal ones, anyway) appear to be oriented with the C-terminus facing the stationary phase. The closer the phosphate group is to the C-terminus, the more it promotes retention. This appears to be a major factor governing selectivity in this mode.

REFERENCES

1) A. Motoyama, T. Xu, C.I. Ruse, J.A. Wohlschlegel, and J.R. Yates, III, Anal. Chem. 79 (2007) 3623-34 [Fig. 4H].

2) A.J. Alpert, S.P. Gygi, A.K. Shukla. Desalting Phosphopeptides by Solid-Phase Extraction. Poster# MP438, 55th ASMS Conference, June 2007.

3) A.J. Alpert. Electrostatic Repulsion Hydrophilic Interaction Chromatography for Isocratic Separation of Charged Solutes and Selective Isolation of Phosphopeptides. Anal. Chem. 80 (2008) 62-76.

4) M. Mazanek, G. Mitulović, F. Herzog, C. Stingl, J.R.A. Hutchins, J.-M. Peters, and K. Mechtler, Nat. Procotols 1 (2006) 1977-87.

5) Y. Takegawa, K. Deguchi, H. Ito, T. Keira, H. Nakagawa, and S.-I. Nishimura, J. Sep. Sci. 29 (2006) 2533-40.

6) Y. Guo and S.J. Gaiki, J. Chromatogr. A, 1074 (2005) 71-80.

7) Y. Guo, S. Srinivasan, and S.J. Gaiki, Chromatographia 66 (2007) 223-29.

8) P.J. Boersema, N. Divecha, A.J.R. Heck, and A. Mohammed, J. Proteome Res. 6 (2007) 937-46.

9) J.V. Olsen, B. Blagoev, F. Gnad, B. Macek, C. Kumar, P. Mortensen, and M. Mann, Cell 127 (2006) 635-48.

 

Topics>>