GC separation of amino acid enantiomers via derivatization with heptafluorobutyl chloroformate and Chirasil-L-Val column

Abstract

Heptafluorobutyl chloroformate (HFBCF), a recently introduced derivatization reagent, was examined in enantioseparation of amino acids (AAs) by GC. Twenty proteinogenic AAs, plus ornithine, cystine and 4-fluorophenylalanine (internal standard) were treated with the reagent and separation properties of the derivatives were assessed on a Chirasil-Val capillary column. Nineteen AA enantiomers were efficiently separated in 43 min except proline, arginine and cystine. The HFBCF derivatives of the studied DL-AAs show improved separation over other chloroformate-based derivatives hitherto reported. A combination of the improved and faster separation with a simple derivatization protocol, involving an immediate one-step reaction–extraction in two-phase aqueous-organic medium, and low elution temperatures extend application of HFBCF to chiral AA analysis.

1 Introduction

Chirality is an inherent property of many natural substances including proteinogenic amino acids (AAs) and separation of the enantiomeric pairs represents an important field of their chemical analysis. A number of methodologies have been developed to fulfill this task including the applications of capillary GC 1–3. Among the current GC methods, separation of AA enantiomers on a Chirasil-(L or D)-Val column has been most widely used 4–12. A valuable alternative is based on the preparation of diastereomeric derivatives after treating the particular analytes with a chiral reagent 13–19 and subsequent separation of the derivatives on a traditional organosiloxane (nonchiral) phase. Owing to the lability of the silylated primary amino groups, silylation is not a suitable reaction for this purpose. Traditional two-step acylation–esterification procedure implemented by Frank et al. 20 for the chiral separations is rather laborious and lengthy, but reliable and robust, and has been widely used in numerous applications 5, 6, 9–12, 21–23.

Derivatization of AAs with an alkyl chloroformate and analogous alcohol enables immediate formation of N,O,S-alkoxycarbonyl-alkyl esters in water-containing media 24. Soon after publishing the first report on the treating AAs with ethyl chloroformate and ethanol 25, 26, 16 AA enantiomeric AA pairs were separated on a mixed cyclodextrin phase and on a Chirasil-L-Val column 27. Further, the simultaneous treatment with trifluoroethyl chloroformate and trifluoroethanol resulted in the formation of volatile species, allowing separation of 16 AA enantiomeric pairs on Chirasil-Val 28. Others examined ethyl chloroformate with 2,2,3,3,4,4,-heptafluoro-1-butanol (HFBOH) that enabled separation of 11 29 or 15 AA chiral pairs simultaneously, but without DL-serine, DL-threonine (Thr), asparagine (Asn), DL-glutamine (Gln) and DL-proline (Pro) 30, 31. Combined action of methyl chloroformate with HFBOH was also described for the separation of 14 proteinogenic AA pairs on a Chirasil-Val column but without DL-Asn, Gln, Pro and DL-histidine 32.

We recently reported pentafluoropropyl (PFPCF) and heptafluorobutyl chloroformates (HFBCF) as new efficient derivatization reagents 33. The HFBCF reagent is more reactive than traditional alkyl chloroformates and PFPCF. With HFBCF most protic functional groups in protein AAs, except the guanidino group in arginine (Arg), are smoothly derivatized without any need for the presence of HFBOH 33, 34. The arising derivatives of AAs are volatile and exhibit excellent GC properties. The advantageous features of the novel fluoroalkyl chloroformate reagents led us to examine the reagents for the separation of AA enantiomers. We first prepared and tested PFPCF for chiral AA analysis in cyanobacteria extracts 35 and for determination of chirality in Carbetocin peptide 36 with satisfactory results. The HFBCF reagent was successfully used for GC-MS determination of more than 30 biomarkers related to folate and cobalamin status in human serum 37. In this paper, we extended the HFBCF derivatization strategy for the first time to chiral GC AA analysis. Twenty proteinogenic AAs plus ornithine, cystine and 4-fluorophenylalanine (internal standard) were treated with HFBCF in an aqueous medium and simultaneously extracted into isooctane layer amenable directly to GC analysis. The separation properties of the formed enantiomeric derivatives were assessed on a Chirasil-Val capillary column.

2 Materials and methods

2.1 Chemicals and laboratory equipment

DL-Alanine, DL-Arg, DL-Asn, DL-aspartic acid, DL-cysteine (Cys), DL-cystine, 4-fluoro-DL-phenylalaine (4FPhe), DL-glutamic acid, Gln, glycine, DL-histidine, DL-isoleucine, DL-leucine (Leu), DL-lysine (Lys), DL-methionine, DL-ornithine, DL-phenylalanine, Pro, DL-serine, Thr, DL-tryptophan (Trp), DL-tyrosine (Tyr), DL-valine (Val) were obtained from Fluka Chemie (Buchs, Switzerland, all of purity >99%). Asn, Gln and Trp were dissolved in deionized water; other 20 AA standards in 0.1 mol/L HCl. Isooctane (for residue analysis) and pyridine were purchased from Sigma-Aldrich (Prague, Czech Republic, all p.a. grades). Sodium carbonate was obtained from Lachema (Brno, Czech Republic, p.a. grade). HFBCF reagent was prepared in the author's laboratory 37 followed by the Abbe method 28 and can be purchased at the Biology Centre (České Budějovice, Czech Republic) upon request.

2.2 Instrumentation

A Shimadzu gas chromatograph GC-2014 equipped with an AOC-20s autosampler and flame ionization detector, all controlled by a GC Solution software from Shimadzu (Japan), were used. Prior to chiral measurement, AAs were converted to the corresponding N(O,S)-heptafluorobutoxycarbonyl O-heptafluorobutyl esters. Separation of derivatized AA enantiomers was accomplished on a Chirasil-L-Val fused-silica capillary column (25 m×0.25 mm id, 0.12 μm film thickness; Varian, Palo Alto, CA, USA). Hydrogen carrier gas flow rate was 1.4 mL/min. The injector and detector temperatures were 220 and 210°C, respectively. The oven temperature program was the following: initial temperature 95°C, the rise of 3°C/min to 200°C, held for 10 min. A 1-μL sample was injected in split mode, split ratio 1:8.

2.3 Derivatization procedure

Reactions were carried out in 1.1-mL tapered polypropylene tubes (Continental Lab Products, San Diego, CA, USA) without any closure. Adjustable 50 and 100-μL Brand Transferpettor pipettes with glass capillaries were supplied by Merck (Prague, Czech Republic), and engaged for manipulation with the reactive reagents and their mixtures in the organic solvent. A vortex mixer (50–2400 rpm) was supplied by P-Lab (Prague, Czech Republic).

The reaction conditions were identical to those described earlier 33. Briefly, the reactive organic medium for the phase-transfer derivatization was a mixture of isooctane with HFBCF (5:3, v/v). First, to 150 μL of 50 mmol/L aqueous sodium carbonate, 50 μL of the particular reactive organic medium were admixed and the two-phase system was vigorously shaken for 10–15 s, rather intermittently, using the vortex mixer. After about 1 min, the second medium, i.e. 50 μL of 100 mmol/L sodium carbonate with admixed pyridine (5:1, v/v) was added, followed by vortexing for about 5–7 s to clear the milky organic phase. After addition of 70–100 μL of isooctane and mixing for about 3 s, an aliquot of the organic phase was withdrawn for the GC-FID analysis.

HFBCF is harmful in contact with skin and if swallowed, toxic by inhalation, may cause burns. The toxicological properties of this substance have not been fully investigated.

Manipulation of the reagent should be performed in a well-ventilated area (fume-hood). After application, autosampler syringes should be rinsed with propan-2-ol to prevent corrosion of their plungers.

2.4 Resolution, precision and LOD

All parameters were calculated on the basis of five repeated GC-FID measurements of 10 nanomol of each AA derivatized by HFBCF. The resolution (R) of the enantiomeric AA pairs was determined by both the temperature gradient 29–32, 38, 39 and the isothermal elution 27, 28, 34, 40; the latter measurement was carried out at a temperature nearest to that obtained from the gradient elution 27, 28, 34, 40. The resolution of substance i and previously eluted i−1 was calculated according to the formula:

• 

(RT, retention time, W, width of peak).

Precision of the method was measured by means of relative standard deviation of peak areas of AA derivatives divided by an average of peak area of D-4FPhe and 4FPhe (RSD in %, n=5). LOD were determined for the HFBCF derivatives with signal-to-noise ratio 3:1 (n=5).

3 Results and discussion

The resolution data of the particular enantiomeric HFBCF-treated DL-AA pairs obtained on the Chirasil-L-Val capillary column under isothermal or temperature gradient temperature conditions are summarized in Table 1; the latter values appeared to be always higher. The AA retention times under the temperature gradient conditions, precision of the enantiomeric peak area measurements (RSD, %) and LOD (expressed in nmol per injection) are shown in Table 2.

Table 1. Values of resolution at isothermal and programmed temperature operation

	Resolution
	Isothermal operation/	Temperature gradienta)
Name	Temperature (°C)	95–200°C, 3°C/min
^a Ala, alanine; Ile, isoleucine; Glu, glutamic acid; Met, methionine; Phe, phenylalanine; Orn, ornithine. a) a) See Section 2.2.
Ala	4.00/110	5.68
Val	3.58/110	4.18
Ile	2.72/120	3.85
Leu	5.45/120	6.72
ThrOR	3.32/130	4.51
Asp	1.68/130	2.06
SerOR	2.17/140	2.79
Glu	2.66/150	4.47
Met	2.98/150	5.56
Phe	2.31/150	3.70
Asn	0.42/150	0.66
Cys	1.77/150	1.69
4FPhe	2.43/150	4.09
Gln	3.03/160	4.90
His	1.94/170	2.36
Orn	2.14/180	3.07
Tyr	1.20/190	2.43
Lys	1.34/190	2.29
Trp	0.95/200	1.36

Table 2. Retention data of the (N,O,S)-HFBOC-HFB ester derivatives on a Chirasil-L-Val column obtained by the gradient elution

Name	RT	RSD (%)	LOD (nmol per injection into a GC column)
^a Ala, alanine; Orn, ornithine; Phe, phenylalanine; Glu, glutamic acid. Precision of the retention time and LOD (S/N=3).
D-Ala	5.20	1.83	0.21
Pro	5.47	2.13	0.52
L-Ala	5.52	2.64	0.21
D-Val	5.94	2.09	0.08
L-Val	6.24	3.34	0.08
Gly	6.93	2.82	1.30
D-Ile	7.80	3.20	0.07
L-Ile	8.09	3.91	0.07
D-Leu	9.29	2.93	0.09
L-Leu	9.84	2.97	0.09
D-ThrOR	12.15	4.27	0.58
L-ThrOR	12.57	5.52	0.59
D-Asp	13.70	3.37	0.14
L-Asp	13.89	3.45	0.14
D-SerOR	15.88	5.75	0.90
L-SerOR	16.06	5.82	0.90
D-Met	17.52	1.59	0.26
L-Met	18.01	3.38	0.26
D-Glu	18.28	5.02	0.26
D-Phe	18.51	4.01	0.15
L-Glu	18.70	3.86	0.23
L-Phe	18.74	3.43	0.15
D-Asn	19.50	4.94	0.31
L-Asn	19.56	3.66	0.31
D-4FPhe	19.87	5.02	0.15
D-Cys	20.10	7.60	1.80
L-4FPhe	20.26	4.98	0.15
L-Cys	20.33	7.61	1.80
D-Gln	22.62	3.91	0.65
L-Gln	23.08	6.85	0.72
D-His	26.79	8.76	0.24
L-His	27.00	8.36	0.23
D-Orn	30.90	5.62	0.90
L-Orn	31.19	6.62	0.90
D-Lys	32.26	7.41	0.16
D-Tyr	32.39	8.50	0.26
L-Lys	32.47	5.99	0.16
L-Tyr	32.50	8.76	0.26
D-Trp	41.64	5.04	0.80
L-Trp	41.89	5.09	0.80

GC-FID chromatograms of the particular AA enantiomers of proteinogenic AAs, 4FPhe, Cys and DL-ornithine after HFBCF derivatization are in Fig. 1A–C. Twenty-one AAs were eluted in 43 min from the capillary column, except Arg, the guanidino group of which resisted an effective derivatization 33, 34. The temperature limit of the chiral phase disabled elution of cystine. Nineteen enantiomeric pairs were separated with favorable resolution ranging from 0.95 (Trp) to 5.45 (Leu) and from 1.36 (Trp) and 6.72 (Leu) for isothermal and temperature gradiental elution, respectively. The separation data of DL-Asn were lower, with respective R values of 0.42 and 0.66. It is clear from Fig. 1 that several peaks of AA HFBCF derivatives are partly overlapping (Pro and L-Ala, DL-methionine and unknown impurity from Thr, L-4FPhe and L-Cys and four enantiomers, D-Lys, D-Tyr, L-Lys and L-Tyr forming a quadruplet, are not baseline resolved).

Figure 1. Temperature gradient GC separation of equimolar mixture (10 nmol each) of AA standards as N(O,S)-heptafluorobutoxycarbonyl heptafluorobutyl esters on a Chirasil-L-Val column. Experimental conditions refer to 2.2.

In accordance with other reports using chloroformate derivatives 28, 32, 33, 35, the DL-Pro HFBCF derivatives were not resolved on the Chirasil-Val phase. Derivatization with isopropyl chloroformate and methanol followed by separation on a chiral cyclodextrin phase was the exception 40; nevertheless, only eight enantiomeric AA pairs were resolved simultaneously. The same was true when the N-alkoxycarbonyl AA trifluoroethyl esters were subjected to nucleophilic substitution of the ester group with amines, e.g. isobutyl amine 41, where separation of six AA chiral pairs was demonstrated. The DL-Pro enantiomers were resolved on a Chirasil-Val phase in this case (α=1.00–1.15); however, the methods were useful only for several low molecular weight AAs.

HFBCF-treated Thr was alkylated on the side chain hydroxyl in a higher yield than the corresponding PFPCF derivative previously described 35 and the alkylated ThrOR enantiomers were more efficiently resolved (R=3.32 at 130°C against 2.55 at 150°C for HFBCF and PFPCF derivatives, respectively).

Separation of DL-Cys was efficient (R=1.77 with temperature gradient). Till now, using chloroformates, DL-Cys has been separated only in two cases, 32, 38, but with a lower gradiental resolution 0.69 and 1.35, respectively, and a longer separation time (over 60 or 70 min, respectively).

The capability to measure a larger number of enantiomers simultaneously is favorable for HFBCF derivatives. In the earlier papers, using chloroformates, separation of six 34, 15 30, 38, 16 27, 28 or 18 35 AA pairs was shown. The information about the separation efficiency summarized in Table 3 is not easy to evaluate because of incomplete data published and different experimental conditions used, even on the identical GC columns. In addition, some works reported resolution 28–32, 38, 39 and the other separation factors 27, 34, 41. The most efficient GC separation using the alkyl chloroformate derivatization methodology was achieved for AAs of the diastereomeric N-ethoxycarbonyl-(S)-1-phenylamide products of 12 AA enantiomeric pairs 39. However, the 3-stage derivatization procedure involved rather complicated and time-consuming steps with separation times mostly between 48 and 70 min.

Table 3. Comparison of retention characteristic under different condition found in literature

Derivatization procedure	Separation efficiencies		Separation system	Temperature program	Enantiomeric pairs	Reference
	R	α
^a ECF, ethyl chloroformate; PEA, phenylethylamine; 2-ClPrOH, 2-chloro-1-propanol; iPrCF, isopropyl chloroformate; TFECF, trifluoroethyl chloroformate; TEA, triethylamine; TFE, 2,2,2-trifluoroethanol; PFPOH, 2,2,3,3,3-pentafluoropropyl-1-propanol; EtOH, ethanol; MCF, methyl chloroformate.
ECF/EtOH		1.000–1.057	Chirasil Val	Isothermal	16	27
ECF/HFBOH	1.03–2.42		Chirasil Val	Gradient	11	29,31
ECF/HFBOH	1.13–5.21		Chirasil Val	Gradient	15	30
ECF/2-ClPrOH	0.45–3.45		Chirasil Val	Gradient	15	38
iPrCF or iBuCF /TFE/alkylamines		1.000–1.225	Chirasil Val	Isothermal	6	41
MCF/HFBOH	0.69–6.24		Chirasil Val	Gradient	14	32
PFPCF/PFPOH	0.44–4.62		Chirasil Val	Isothermal	17	35
TFECF/TFE		0.53–4.52	Chirasil Val	Isothermal	16	28
iPrCF/MeOH		1.047 Asp	Modified cyclodextrin	Isothermal	8	40
ECF/2-ClPrOH	0.48–1.67		CP-Sil 19 CB	Gradient	9	38
ACF/TEA/PEA	2.3–10.2		DB-5	Gradient	12	39
ACF/TEA/PEA	3.1–107		DB-17	Gradient	12	39

The chiral separation data summarized in the Tables 2 and 3 indicate that the HBFCF-treated AA derivatives exhibit advanced features, particularly in terms of resolution of the particular enantiomers and in a number of AA to be analyzed simultaneously. Moreover, fast sample preparation in aqueous medium and the analysis speed are shorter than previously published 27, 29–32, 35, 38–41.

For practical applications, a Chirasil-L-Val column is preferred to the D-Val counterpart as the D-enantiomers are eluted in front of the L-forms for all described AAs. This might be advantageous in samples of natural origin where the D-AAs are minority by a rule.

4 Concluding remarks

The HFBCF derivatization proved to be a useful tool for the chiral analysis of proteinogenic AAs, except Arg, Pro and cystine. The HFBCF treatment enabled 19 enantiomeric pairs to be separated in a single GC analysis. The derivatives are readily prepared in aqueous media while simultaneously transferred into an isooctane phase, directly amenable to GC analysis. Most enantiomers are separated with improved resolution over the existing chloroformate-based derivatization methods using a simple temperature gradient and in a short analysis time.

Acknowledgements

The study was supported by the Grant Agency of the Czech Republic, research project No. 203/09/2014, which is highly acknowledged.

The authors have declared no conflict of interest.