Chiral CE of aromatic amino acids by ligand-exchange with zinc(II)–L-lysine complex

Li QiYanli HanMin ZuoYi Chen

Beijing National Laboratory ofMolecular Science,Laboratory of AnalyticalChemistry for Life Science,Institute of Chemistry,Chinese Academy of Sciences,Beijing, P. R. China

Received November 14, 2006Revised February 15, 2007Accepted February 21, 2007

Research Article

Chiral CE of aromatic amino acids by ligand-exchange with zinc(II)–L-lysine complex

A novel method of chiral ligand-exchange CE was developed with either L- or D-lysine (Lys)as a chiral ligand and zinc(II) as a central ion. This type of chiral complexes was explored forthe first time to efficiently separate either individual pairs of or mixed aromatic amino acidenantiomers. Using a running buffer of 5 mM ammonium acetate, 100 mM boric acid,3 mM ZnSO4?7H2O and 6 mM L-Lys at pH 7.6, unlabeled D,L-tryptophan, D,L-phenylala-nine, and D,L-tyrosine were well separated, giving a chiral resolution of up to 7.09. The bestseparation was obtained at a Lys-to-zinc ratio of 2:1, zinc concentration of 2–4 mM andrunning buffer pH 7.6. The buffer pH was determined to have a strong influence on reso-lution, while buffer composition and concentration impacted on both the resolution andpeak shape. Boric acid with some ammonium acetate was an adoptable buffer system, andsome additives like ethylene diamine tetraacetic acid capable of destroying the complexshould be avoided. Fine-tuning of the chiral resolution and elution order was achieved byregulating the ratio of L-Lys to D-Lys; i.e. the resolution increased from zero to its highestvalue as the ratio ascended from 1:0 to 1:infinitive, and L-isomers eluted before or afterD-isomers in excessive D- or L-Lys, respectively.

Keywords:

Aromatic amino acid / Enantioseparation / Ligand-exchange CE / Zinc–L-Lys com-plex DOI 10.1002/elps.200600740

Electrophoresis 2007, 28, 2629–2634 2629

1 Introduction

The principle of ligand-exchange chromatography wasintroduced by Davankov and Rogozhin in 1971 [1]. It hassince become one of the key principles in chiral separations,applicable to the chiral separation of amino acids (AA) andother chelate complex forming compounds [2]. Zare and co-workers [3] found that this principle was also adoptable inconducting chiral CE of dansyl AA with Cu(II)–histidine or–aspartame complex as a selector [4]. This is actually a type offlexible chiral ligand-exchange CE (CLE-CE) since the ligandis, in theory, variable. In practice, it has been shown that theligand histidine or aspartame could indeed be replaced by,e.g. N,N-didecyl-L-alanine, arginine, prolinamide, L-proline,and L-4-hyproline [5–8]. Andrighetto et al. [9] found that theselectivity of CLE-CE was strongly affected by CD and itsderivatives due to the presence of a guest–host interaction.

Cucinotta et al. [10, 11] and Wu et al. [12] revealed that thisinteraction could directly be introduced into CLE-CE usingCD-derivatives such as 3-amino-b-CD or 6A-(2-aminoethyl-amino)-6A-deoxy-b-CD as a ligand. In addition, Schmid et al.[13] developed a CEC method using chiral ligand exchangestationary phase, while Karbaum and Jira [14] explored non-aqueous CLE-CE. The application of CLE-CE has been wellreviewed by Gübitz and Schmid [15].

Although ligands are important in performing CLE-CEdue to the exchange mechanism [16], the frequently usedcentral ion Cu(II) [17] is also replaceable, namely by Mn(II),Zn(II), Ni(II), and even B(III) or borate [18–26]. Kodama et al.[23] have demonstrated that borate was effective in theseparation of D,L-pantothenic acid because it is capable offorming complex with the cis-1,2- or 1,3-hydroxy. Zheng et al.[24] have studied the use of Mn(II), Zn(II), or Ni(II)–alaninecomplex (in combination with b-CD) in CLE-CE for theseparation of dansyl AA. Linder et al. [25] employed chiraltriamine–Zn(II) complexes for the separation of dansylatedAA. Horimai et al. [26] discovered that a Zn(II)–D-phenylala-nine complex together with g-CD generated good resolutionof six new chiral quinolone drugs. However, Zn(II) and othercentral ions remain unexplored as compared with Cu(II),possibly due to their role in causing the evident loss of chiralrecognition [6]. In order to inspect the reason(s), a systematicstudy has been conducted with Zn(II) as a center ion and

Correspondence: Professor Yi Chen, Institute of Chemistry, Chi-nese Academy of Sciences, P. O. Box 2709, Beijing 100080, Peo-ple’s Republic of ChinaE-mail: [email protected]: 186-10-62559373

Abbreviations: AA, amino acid(s); CLE-CE, chiral ligand-exchange

© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

2630 L. Qi et al. Electrophoresis 2007, 28, 2629–2634

various AA as ligands. The results strongly suggested thatthe effectiveness of a central ion largely depended on theselection of running conditions.

In this study, the use of Zn(II)–lysine (Lys) complex isdiscussed in CLE-CE of aromatic D,L-tryptophan (D,L-Trp),D,L-phenylalanine (D,L-Phe), and D,L-tyrosine (D,L-Tyr). It wasfound that the key to achieve a successful chiral separationwas to carefully search the pH. The composition of the run-ning buffer and its concentration also impacted greatly onthe enantioresolution (Rs) and peak shape. After optimiza-tion, all the tested amino acidic enantiomers were well sepa-rated. To our knowledge, this is the first study to use aZn(II)–Lys complex in CLE-CE of aromatic AA.

2 Materials and methods

2.1 Chemicals

D,L-Trp, D,L-Phe, and D,L-Tyr were purchased from SigmaChemical (St. Louis, USA). NiCl2 was obtained from BritishDrug House (London, UK). Boric acid, Tris, PEG 10 000,EDTA, and the other reagents of analytical reagent gradewere all obtained from Beijing Chemical Factory (Beijing,China).

2.2 apparatus CE

CE was carried out using P/ACE model 2050 (BeckmanCoulter, USA). A bare fused-silica capillary (YongnianOptical Fiber Factory, Hebei, China) of 75 mm id657 cmlenght (50 cm effective) was mounted. Prior to injection,the capillary was sequentially rinsed with 0.1 M HNO3,water, 0.1 M NaOH, water, and running buffer for 2 mineach. Samples were then injected at 0.5 psi for 3 s andseparated at 120 kV and 257C. The analytes were detectedby UV absorption at 214 nm and acquired at 4 Hz. Thepeaks were identified by spiking standard AA in samplesolutions. The peaks with increased height were con-sidered to be the targets.

2.3 Buffer and sample preparation

CE running buffer, unless stated otherwise, was com-posed of 5 mM ammonium acetate, 100 mM boric acid,0–6 mM ZnSO4?7H2O, and 0–12 mM L-Lys adjusted topH 7.60 with Tris. Before use, all running buffer solu-tions were filtered through a membrane with 0.45 mmpores and degassed by sonication for 2 min. The waterused was doubly distilled using a distillation apparatusmodel SZ-93 (Yarong Biochemical Instrument, Shanghai,China).

Stock solutions of samples were prepared by dissolving2.0 mg of AA in 1 mL of double-distilled water and stored at47C. Injected samples were prepared by diluting the stocksolutions with water by a factor of 10–80.

3 Results and discussion

3.1 Effectiveness of the Zn(II)–Lys complex

With Lys as the ligand, Zn(II) produced the best Rs overthe other tested ions including Cu(II), Co(II), Mg(II), andMn(II) as shown in Table 1. Co(II) was the second effectivecentral ion but it yielded poor resolution of D,L-Tyr and noresolution of D,L-Phe. Other ions did not show obviouschiral selectivity for the tested three pairs of AA. This isparallel to the migration time (t) offered by the ions, that isZn(II) .Co(II) .Mn(II) .Mg(II) .Cu(II), implying that along separation duration is preferred for chiral resolutionas in common CE. However, it should be noted that Rs isalso parallel to the radius-to-charge ratio of the central ions[24]:

Zn(II) .Co(II) < Cu(II) .Mn(II) .Mg(II)

with the exception of Cu(II). This means that the size andcharge of a central ion should be considered in conductingCLE-CE of D,L-AA. It is better to select an ion with a radius-to-charge ratio similar to or a little greater than Cu(II).

Table 1. Influence of central ion on the performance of CLE-CEa)

Ion Trp Phe Tyr

tDa) (min) tL

a) (min) Rsa) aa) tD

a) (min) tLa) (min) Rs

a) aa) tDa) (min) tL

a) (min) Rsa) aa)

Zn 37.23 40.69 7.09 1.09 39.52 42.22 3.69 1.07 43.17 45.06 2.47 1.04Co 23.14 23.70 1.43 1.02 23.70 23.70 0 1.00 25.09 25.58 0.76 1.00Mn 19.52 19.52 0 1.00 17.75 17.75 0 1.00 21.46 21.46 0 1.00Mg 9.44 9.44 0 1.00 9.92 9.92 0 1.00 9.92 9.92 0 1.00Cu 7.20 7.20 0 1.00 7.53 7.53 0 1.00 7.60 7.60 0 1.00

a) Running buffer: 5 mM ammonium acetate, 100 mM boric acid, 3 mM central ions, and 6 mM L-Lys at pH 7.6; for other conditions refer toSection 2; Rs = 2 (tL 2 tD)/(WL 1 WD); a = tL/tD; t: migration time.


Electrophoresis 2007, 28, 2629–2634 CE and CEC 2631

Lys was selected because it has a higher isoelectric point(pI = 9.4) [27] than the sample AA (pI , 6.0). Its amino butylside chain could interact with or arrest the negative chargedAA at pH .pI and borate anion as well. A better resolutioncould thus be achieved.

3.2 Complex concentration and Lys-to-Zn(II) ratio

The concentrations of Lys and Zn(II) and their molar ratioare theoretically important to regulate Rs and are confirmedby experiments. Figure 1A shows that the best ratio of Zn(II)to Lys is 1:2, when their total concentrations are fixed,implying that Zn(II)–Lys has the same structure as Cu(II)-based complexes. The chiral selection mechanism is thusconsidered also to be the same as that involved in Cu(II)complex systems:

Zn(L-Lys)2 1 (D-AA)) !KD (L-Lys)Zn(D-AA) 1 (L-Lys)

Zn(L-Lys)2 1 (L-AA)) !KL (L-Lys)Zn(L-AA) 1 (L-Lys)

where KD and KL are the equilibrium constants of the ligandexchanging reaction for D-and L-AA, respectively. The chiralresolution power should root in the difference between KD

and KL.Figure 1A also reveals that the Rs of D,L-Trp responds to

the Zn(II)-to-Lys ratio more sharply than that of the otherAA. In addition, all the AA tend to overlap far beyond theoptimum ratio, especially at a ratio lower than 1:1. Thismeans that the ligand exchange will become impossiblewhen Lys is not sufficient. In other words, ligand exchange isthe prerequisite to conduct chiral separation in this study.

Somewhat unexpectedly, the concentration of the com-plex was found to have a strong impact on the chiral resolu-tion. Figure 1B shows clearly that Rs increases sharply at,2 mM Zn(II) and decreases also fast at .4 mM Zn(II).This phenomenon should be anticipated by equilibriumshifting: for a given amount of solute, the ligand-exchangereaction will move toward left when the complex is not suf-ficient, also causing insufficient exchange and consequentlyreducing the resolution; in contrast, when the complex is tooconcentrated to make solutes incapable of leaving the com-plex, the ligand exchange among the solutes will be equal-ized, causing also a loss of chiral resolution. For higher Rs,the complex concentration should be confined in the middle,namely in between 2 and 4 mM Zn(II) in this study. A betteror compromised one was 3 mM Zn(II).

3.3 Running buffer

The CE running buffer may contain various compositionsdepending on the solutes and the separation model adopted.In common, buffering reagents, their concentration, pH,solvent and additive are considered. In this study, becausewater was found to be sufficient as the solvent, the influencesof solvents were not studied in detail. A few organic solventswill, however, be mentioned as additives in Section 3.3.3.

3.3.1 Buffer reagent and concentration

Buffer reagent and concentration had a great influence onseparation and even on the peak shape. To find as good abuffer system as possible, several reagents including ammo-nium acetate, carbonate, phosphate, and borate were

Figure 1. Dependence of enantioresolution (Rs) on complex concentration. (A) Variation of Rs against Zn (II)-to-Lys ratio. Running buffer:5 mM ammonium acetate, 100 mM boric acid, 3 mM Zn(II), and cLys mM L-Lys at pH 7.6; capillary: 75 mm id657 cm lenght (50 cm effective);injection: 0.5 psi for 3 s; voltage: 20 kV; temperature: 257C; UV detection: 214 nm. (B) Plot of Rs against zinc concentration at a constant ratioof Zn(II) to L-Lys. Running buffer: 5 mM ammonium acetate, 100 mM boric acid, cZn mM Zn (II) and 2cZn mM L-Lys. Other conditions as in (A).



checked at the concentrations of 5–200 mM at pH 7.6. Gen-erally, borate yielded the best chiral resolutions, ammoniumacetate the second, while carbonate or phosphate the worst.Further studies showed that mixed buffers of borate andammonium acetate produced even better resolution than theborate alone. On testing from 5 to 20 mM ammonium ace-tate and from 50 to 200 mM boric acid, a buffer of 5 mMammonium acetate and 100 mM boric acid at pH 7.6 wasfound to be the best for the separation of mixed samples. Thepositive effect of borate on the performance was ascribed toits coordination with hydroxyl and amino groups [23, 28].This happens in the case of Tyr with a hydroxyl group on itsphenyl ring and for Trp with an amino group.

Figure 2 shows that boric acid concentration (pH 7.6) hasstriking influences on peak shape, resolution, eluting timeand elution order. Symmetrical peaks obtained at concentra-tions above 50 mM boric acid and Rs increased with the con-centration for any single pair of AA. Nevertheless, the bestchiral separation for mixed AA happened at around 100 mM

boric acid (Fig. 2B), or D-Phe/D-Trp, and L-Phe/L-Trp over-lapped easily once the boric acid content was varied. Itshould specially be noted that boric acid is better kept below200 mM to speed up the separation and reduce the heatingeffect.

3.3.2 pH

As is well known that, the stability of complexes stronglydepends on pH which controls the dissociation of not onlyligands but also the analytes. Tested with a buffer of 5 mMammonium acetate and 100 mM boric acid, the resolutionwas determined to sharply depend on acidity at pH 6.0–9.0,reaching the maximum at pH 7.6 (Fig. 3). Clearly, the pHwindow was quite narrow (between pH 7.4 and 7.8) usablefor mixed AA, which may explain why zinc complexes easilyproduce poor chiral resolution before this investigation. Itmust be noted that the working pH should be kept below 9.0to prevent the precipitation of the Zn(II) form.

Figure 2. Electropherogram of mixed aromatic AA measured in a running buffer of 5 mM ammonium acetate, 3 mM Zn (II), 6 mM L-Lys and(A) 50, (B) 100, (C) 150, or (D) 200 mM boric acid adjusted to pH 7.6 with Tris. Other conditions as in Fig. 1. Peak identity: 1D = D-Trp, 1L = L-Trp, 2D = D-Phe, 2L = L-Phe, 3D = D-Tyr, 3L = L-Tyr.


Electrophoresis 2007, 28, 2629–2634 CE and CEC 2633

Figure 3. Plots of Rs versus pH measured from a buffer of 5 mMammonium acetate, 100 mM boric acid, 3 mM Zn (II), and 6 mM L-Lys adjusted to the required pH with crystal Tris. Other conditionsas in Fig. 1.

3.3.3 Non-chiral additive and interference

Non-chiral additives are often used to improve CE perfor-mance but considering that the metal ion was used in CLE-CE, the negative effect or interference of the additives wasspecially studied in addition to the positive effect. For thispurpose, several representative additives were inspectedincluding EDTA, PEG, SDS, and other organic modifiers.PEG 10 000 (5.0 mg/mL) was able to reduce electroosmosisor to prolong elution but had no effect on enhancing thechiral resolution. SDS and other organic modifiers such asmethanol and tert-butyl alcohol produced only negligibleeffect on varying the enantioselectivity.

However, the composition that could compete with AA toform complexes impacted largely on the chiral separation.Tartaric acid, citric acid, EDTA, and so forth should beexcluded from the running buffers. These strong ligandsstart disturbing the enantioresolution even at a very low

concentration (EDTA, for example, at 0.4 mM as shown inTable 2) and also ruin the separation at low concentrationitself (EDTA at 3.0 mM).

3.4 Tuning the elution and resolution

It has been demonstrated that both the resolution and theelution order of a pair of enantiomers could be adjusted orcontrolled by regulating not only the buffer features (Figs. 2and 3) but also the complex-related factors, namely its con-centration and Zn-to-Lys ratio (Fig. 1). Further study showedthat fine-tuning of the resolution or elution order of a pair ofD,L-AA was also possible. Figure 4 displays a sequential elec-tropherogram corresponding to the variation of D-Lys-to-L-Lys ratio. Clearly, D-Trp migrated ahead of L-Trp in excessiveL-Lys or conversely in excessive D-Lys; the resolutionincreased with the excessiveness of either L-Lys or D-Lys.Thus, the best resolution happened only when either L-Lys orD-Lys was present and non separation was obtained whenD-Lys was equal to L-Lys.

Figure 4. Tuning the elution order and enantioresolution of D,L-Trp by varying the D-to-L ratio of ligand Lys at (a) 6:0, (b) 4:2, (c)3:3, (d) 2:4, and (e) 0:6 while keeping the molar ratio of Zn(II) to allLys at 1:2. Other conditions as in Fig. 1.

Table 2. Impact of EDTA on the enantioseparation of CLE-CEa)

C (mM) Trp Phe Tyr

tD (min) tL (min) Rs a tD (min) tL (min) Rs a tD (min) tL (min) Rs a

0 37.23 40.69 7.09 1.09 39.52 42.22 3.69 1.07 43.17 45.06 2.47 1.040.4 30.88 33.78 6.59 1.09 32.85 35.10 2.61 1.07 35.85 37.24 1.58 1.040.6 28.58 31.25 2.96 1.09 30.51 32.58 2.21 1.07 33.03 34.26 1.83 1.040.8 26.26 28.34 2.17 1.08 27.86 29.79 2.11 1.07 29.79 30.72 1.58 1.031.0 12.29 12.67 2.43 1.03 12.54 12.86 1.64 1.03 13.22 13.44 1.08 1.023.0 8.86 8.97 0.74 1.03 8.97 9.25 1.44 1.03 9.25 9.42 1.28 1.024.0 6.14 6.14 0 1.00 6.14 6.14 0 1.00 6.26 6.26 0 1.00

a) The conditions and symbols are the same as in Table 1 except that the central ion is Zn(II).



4 Concluding remarks

A Zn(II)–Lys complex system was shown for the first time tobe suitable for the enantioseparation of native aromatic AAby CE with common UV absorption detection. The tried zinccomplex was convenient in use but the running buffershould be selected in a way a bit different from that for cop-per complexes. To achieve baseline enantioseparation, boratebuffer containing ammonium acetate was the best choice.Except for strong chelating reagents such as EDTA, which isharmful and should be avoided, other buffer additives likeSDS, PEG, and organic solvents had only negligible influ-ences on Rs. This new method is also extendable or applica-ble to the separation of other analogs of AA.

However, it should be mentioned that, although thismethod is in principle applicable to the enantioseparation ofaliphatic AA, we have not yet obtained confidential data dueto the detection problem. Either indirect detection or labelingtechnique has been tried to detect the peaks but new prob-lems have been encountered because these modificationslargely disturbed the complexation. Further study should beconducted to extend this method to the analysis of non-aro-matic AA.

We gratefully acknowledge the financial support from NSFC(Nos. 20375042, 20420130137 and 20435030), Chinese Acad-emy of Sciences (no. KJCX2-SW-H06), and Ministry of Scienceand Technology of China (No. 2002CB713803).

5 References

[1] Davankov, V. A., Rogozhin, S. V., J. Chromarogr. 1971, 60,284–312.

[2] Davankov, V. A., Navratil, J. D., Walton, H. F., LigandExchange Chromatography 1988, CRC Press, Boca Raton, FL1988.

[3] Gassmann, E., Kuo, J. E., Zare, R. N., Science 1985, 230, 813–814.

[4] Gozel, P., Gassmann, E., Michelsen, H., Zare, R. N., Anal.Chem. 1987, 59, 44–49.

[5] Cohen, A. S., Paulus, A., Kager, B. L., Chromatographia 1987,24, 15–24.

[6] Yuan, Z. B., Yang, L., Zhang, S. S., Electrophoresis 1999, 20,1842–1845.

[7] Chen, Z. L., Nilitsuma, M., Nakagama, T., Uchiyama, K.,Hobo, T., J. Sep. Sci. 2002, 25, 1197–1201.

[8] Schmid, M. G., Gübitz, G., Entiomer 1996, 1, 23–27.

[9] Andrighetto, P., Carofiglio, T., Fornasier, R., Tonellato, U.,Electrophoresis 2000, 21, 619–626.

[10] Cucinotta, V., Giuffrida, A., Mendola, D. L., Maccarrone, G. etal., Chromatogr. B 2004, 800, 127–133.

[11] Cucinotta, V., Giuffrida, A., Maccarrone, G., Messina, M.,Vecchio, G., Electrophoresis 2006, 27, 1471–1480.

[12] Wu, B. D., Wang, Q. Q., Liu, Q., Xie, J. W., Yun, L. H., Elec-trophoresis 2005, 26, 1013–1017.

[13] Schmid, M. G., Grobuschek, N., Tuscher, C., Gübitz, G. et al.,Electrophoresis 2000, 21, 3141–3144.

[14] Karbaum, A., Jira, T., J. Chromarogr. A 2000, 874, 285–292.

[15] Gübitz, G., Schmid, M. G., Electrophoresis 2000, 21, 4112–4135.

[16] Chen, Y., Lu, X. N., Han, Z. Q., Qi, L. et al., Electrophoresis2005, 26, 833–840.

[17] Zheng, Z. X., Wei, Y. L., Lin, J. M., Electrophoresis 2004, 25,1007–1012.

[18] Vallee, B. L., Falchuk, K. H., Physiol. Rev. 1993, 73, 79–118.

[19] Berg, J. M., Shi, Y., Science 1996, 271, 1081–1085.

[20] O’Halloran, T. V., Science 1993, 261, 715–725.

[21] Fraker, P. J., Telford, W. G., Proc. Soc. Exp. Biol. Med. 1997,215, 229–236.

[22] Costello, L. C., Franklin, R. B., Prostate 1998, 35, 285–296.

[23] Kodama, S., Yamamoto, A., Lio, R., Sakamoto, K. et al., An-alyst 2004, 129, 1238–1242.

[24] Zheng, Z. X., Qu, F., Lin, J. M., Chin. J. Chem. 2003, 21, 1478–1484.

[25] Linder, W., Lepage, J. N., Davies, G., Seitz, D. E., Karger, B. L.,J. Chromatogr. 1979, 186, 323–344.

[26] Horimai, T., Ohara, M., Ichinose, M., J. Chromatogr. A 1997,760, 235–244.

[27] Lu, X. N., Chen, Y., Guo, L., Yang, Y. F., J. Chromatogr. A2002, 945, 249–255.

[28] Wang, X. Y., Chen, Y., Li, Z., Wang, Z. X., J. Liq. Chromatogr.Technol. 2002, 25, 589–600.


Chiral CE of aromatic amino acids by ligand-exchange with zinc(II)–L-lysine complex

Documents

Transcript of Chiral CE of aromatic amino acids by ligand-exchange with zinc(II)–L-lysine complex