Simultaneous determination of choline, carnitine and betaine in premixes by non-suppressed ion chromatography

Abstract

A simple, highly sensitive ion chromatographic method for the simultaneous determination of betaine, choline and carnitine in feed premixes is described. Premixes were extracted via ultrasonication in water for 20 min and extracts were analyzed by ion chromatography using an aqueous 3.0 mmol L⁻¹ methane sulfonic acid solution containing 15% (v/v) acetonitrile as the eluent and an IonPac SCS1 column. The composition of the mobile phase was optimized for the efficient separation of the three main analytes from each other as well as from common inorganic cations and trimethylamine, which is an ingredient for the synthesis of the analytes. The recoveries of choline, carnitine and betaine spikes added to premixes were all greater than 90% with replicate relative standard deviations of less than 10%. The limits of detection of this method for these analytes in premixes were calculated to be 10, 6 and 20 mg kg⁻¹, respectively.

---

1 Introduction

Choline is a common dietary supplement for both animals and humans, owing to its role in several metabolic pathways, including carotene metabolism and the formation of cell structures. It is also a major source of methyl groups within metabolic processes and a precursor of acetylcholine.^1–3 Carnitine is a quaternary ammonium compound that is biosynthesized from the amino acids lysine and methionine. The primary metabolic role of carnitine is the transport of long-chain fatty acids across the mitochondrial membrane, as a prelude to the β-oxidation of the acids to produce energy.^4,5 Betaine is formed by the oxidation of choline¹ and functions as a dietary source of methyl groups and has a role in cell volume regulation in animals under osmotic stress conditions.^6,7 The molecular structures of these compounds are provided in Fig. 1. Several studies have reported that dietary supplementation with these compounds significantly improves both the growth performance and carcass quality of animals raised for food purposes,^8–13 and thus the hydrochloride salts of these chemicals are commonly used as feed or premix additives to increase the feed efficiency.

Fig. 1 The molecular structures of (1) trimethylamine, (2) carnitine, (3) choline and (4) betaine.

Existing analytical techniques for the determination of choline, carnitine and betaine in various matrices include enzymatic reactions,^14–16 liquid chromatography^17–20 and mass spectrometry,^21–23 with the enzymatic methods the most commonly used. Enzymatic techniques, however, are sensitive to matrix interference. The influence of proteins and inorganic salts in the catalase assay of carnitine in serum has already been reported¹⁵ and similar interference may occur in complex feed matrices. Several enzymatic methods for choline have been developed that rely on reactions catalyzed by choline oxidase (ChO_x) to convert choline into betaine,¹⁶ but the use of such methods does not allow the simultaneous analysis of choline and betaine within the same sample. As an alternative, high pressure liquid chromatography (HPLC) may be used to quantify these analytes in a wide variety of products, although the majority of HPLC methods for feed compounds require derivatization or the use of complex purification protocols prior to analysis.^17–20 As an example, the pre-column derivatization of carnitine to yield a fluorescent derivative is one of the strategies used to address the lack of a chromophore in this molecule.¹⁹ Assays used to determine betaine concentrations in feed ingredients also require pre-purification using an SCX column.¹⁷ Mass spectrometry is another analytical technique commonly used to quantify analytes in plant tissues and plasma. The required instrumentation, however, is quite expensive and for this reason is not commonly available in nutrition laboratories. Another option is ion chromatography (IC), which has been applied to the determination of choline in infant formula and betaine in feed,^24,25 but has not been reported for the simultaneous determination of choline, carnitine and betaine in premixes. To address the deficiencies noted above, we have developed a simple and fast ion chromatographic method for the simultaneous detection of these analytes. As part of the method development, the mobile phase composition was optimized for the efficient separation of the analytes from common inorganic cations as well as from trimethylamine (TMA), which is a raw ingredient for the synthesis of the analytes and therefore may be expected to be present in premixes.

2 Materials and methods

2.1 Reagents and solutions

All reagents were of analytical grade unless otherwise stated. Methane sulfonic acid (MSA), TMA hydrochloride and choline chloride were obtained from Sigma Aldrich Co., LLC (St. Louis, MO, USA). L-Carnitine and betaine hydrochloride were obtained from Aladdin Reagent Co. (Shanghai, China). Water was purified by a Milli-Q Plus water system (Millipore, Bedford, MA, USA). Standard solutions of inorganic cations were obtained from the National Institute of Metrology (Beijing, China). All solutions used for IC were first filtered through a 0.45 μm filter and degassed. Blank premixes and premixes containing the desired analytes were obtained from a local market in China.

Stock solutions were prepared in methanol at a concentration of 1 mg mL⁻¹ and kept at 0 to 4 °C for no longer than six months. Working solutions used for method validation were prepared in water, sealed and kept at 0 to 4 °C for no longer than four weeks.

2.2 Instrumental conditions

Chromatographic analysis was performed on a Dionex ICS 2500 IC system (Dionex, Sunnyvale, CA, USA) equipped with a Dionex GP50 Pump, a Dionex ED50A conductivity detector with non-suppressed conductivity detection, a column oven and a Dionex AS 3000 auto-injector. Chromeleon chromatography workstation software was used for instrument control, data collection, data processing, and analysis.

2.3 Sample preparation

Each sample, ranging from 1 to 5 g in mass, was accurately weighed and transferred to a 50 mL test tube, to which 20 mL of water was added. The sample was vortexed for 20 s and ultrasonicated for 20 min to extract the analytes, followed by centrifugation at 8000 rpm for 5 min. The supernatant was diluted with sufficient water so as to obtain a final solution with analyte concentrations in the range of 2 to 200 mg mL⁻¹. The resulting solutions were filtered through a 0.45 μm nylon filter before IC analysis.

2.4 IC conditions

The chromatographic column was a Dionex IonPac SCS 1 analytical column (250 mm × 4 mm i.d.) with a Dionex IonPac SCG 1 guard column (50 mm × 4 mm i.d.). The optimized eluent was an aqueous solution of 3.0 mmol L⁻¹ MSA with 15% (v/v) acetonitrile. Separation was carried out at a flow rate of 1 mL min⁻¹ and a constant column temperature of 30 °C.

3 Results and discussion

3.1 Optimization of chromatographic performance

All of the analytes, as well as TMA, have the same amine functional group and similar structures, and therefore their chromatographic separation with baseline resolution is not trivial. To determine the optimum MSA concentration in the mobile phase, different MSA concentrations ranging from 1.5 mmol L⁻¹ to 4.5 mmol L⁻¹ were employed. Each of these MSA concentrations produced good chromatographic separation of betaine, carnitine and choline. However, there were significant effects of the MSA concentration on the resolution of TMA and carnitine, as illustrated in Fig. 2, such that the resolution of these two analytes decreased with increasing MSA concentration. A concentration of 3 mmol L⁻¹ MSA was found to satisfactorily separate TMA and carnitine (R = 1.60). Although lower MSA concentrations resulted in better separation, they also produced longer retention times and unsatisfactory peak shapes and hence 3 mmol L⁻¹ MSA was considered to be optimal for the separation of carnitine and TMA.

Fig. 2 The resolution of TMA and carnitine peaks at mobile phase MSA concentrations from 2.0 to 4.0 mmol L⁻¹.

A further challenge was presented by unsatisfactory resolution between choline and Mg²⁺ and betaine and Na⁺ when using the 3 mmol L⁻¹ MSA mobile phase. Since many feed premixes contain significant quantities of Mg and Na, different mobile phase acetonitrile percentages ranging from 0% to 20% (v/v) were tested while keeping the concentration of MSA constant, to determine the optimum acetonitrile concentration. The significant effects of mobile phase acetonitrile concentration on the analyte retention times are summarized in Fig. 3. As a result of their interactions with acetonitrile, the three amine analytes displayed more rapid decreases in retention times with increasing acetonitrile concentration as compared to the inorganic cations, leading to improved resolution. Taking peak resolution, peak shape and detector sensitivity into account, a mobile phase acetonitrile concentration of 15% (v/v) was selected. The effects of column temperature were also analyzed and increased temperatures were found to shorten retention times but also to marginally reduce peak heights and area counts. As a compromise, a column temperature of 30 °C was employed. Fig. 4 shows the chromatogram of a standard solution obtained under these optimized experimental conditions. All three target analytes, as well as the common inorganic cations and TMA, are well separated within a 15 min run time.

Fig. 3 The effects of mobile phase acetonitrile content on the retention times of betaine, choline, TMA, carnitine and common inorganic cations.

Fig. 4 Chromatographic separation of a standard solution showing peaks for (1) betaine, (2) sodium, (3) potassium, (4) ammonium, (5) carnitine, (6) TMA, (7) choline, (8) magnesium and (9) calcium.

3.2 Optimization of the extraction method

A number of extraction parameters were evaluated, including the extraction solvent composition (water, water–methanol (50/50, v/v), methanol, 0.1 mol L⁻¹ HCl), type of extraction (oscillatory or ultrasonic) and extraction time (10, 20 or 30 min). These extraction experiments were designed to evaluate the influence of each parameter on the recovery of analytes from premix samples. The results from these tests indicated that ultrasonic extraction was superior to oscillatory extraction, however there were no significant differences in extraction efficiency observed between the four different solvents. Based on data obtained from these tests, it was concluded that the optimal conditions consisted of ultrasonic extraction in water for 20 min, and these conditions were subsequently applied to sample extraction. Fig. 5 shows the chromatogram of a typical premix sample extraction.

Fig. 5 Representative chromatogram of a premix sample showing peaks for (1) betaine, (2) carnitine and (3) choline.

3.3 Linearity, LOD and LOQ

To determine the linearity of this method with regard to choline, carnitine and betaine, standard solutions of each analyte at concentrations of 1, 2, 5, 10, 20, 50, and 100 mg mL⁻¹ were analyzed. The resulting data were used to generate calibration curves to which straight line fits were applied, as summarized in Table 1. All three analytes generated highly linear calibrations with correlation coefficients greater than 0.999.

Table 1 Summary of linearity, LOD and LOQ data

Analyte	Concentration range (mg mL^−1)	Straight line equation	R	LOD (mg kg^−1)	LOQ (mg kg^−1)
Choline	1–100	Y = 0.0586X − 0.0333	0.9999	3.3	10
Carnitine	1–100	Y = 0.0686X − 0.0231	0.9997	2.2	6
Betaine	1–100	Y = 0.0186X − 0.0015	0.9995	5.2	15

---

The limit of detection (LOD) for each analyte was calculated as the concentration that produced a signal-to-noise ratio of 3, and was determined for all compounds by spiking a blank premix with decreasing concentrations until this ratio was observed. The limit of quantification (LOQ) was similarly set at a signal-to-noise ratio of 10. Replicate data exhibited acceptable precision (relative standard deviation, RSD ≤ 10%). A summary of results is presented in Table 1.

3.4 Recovery and repeatability

The recoveries of analyte spikes of different levels from a variety of blank premixes and from four commercial premix products (two solid and two liquid) with known analyte concentrations were evaluated (Table 2). The majority of the commercial premixes contained only one or two of the three analytes, and these products were chosen so as to cover a representative concentration range of the analytes, from low (<200 mg kg⁻¹) to medium (200–5000 mg kg⁻¹) to high (>5000 mg kg⁻¹). Three replicates of each sample with different analyte concentrations were performed and the results are summarized in Table 2. Spike recoveries ranged from 92 to 98%, 91 to 99% and 94 to 103% for betaine, carnitine and choline, respectively, with RSDs less than 10%. The results of the recovery assays demonstrated that no significant loss of analytes occurred during the extraction process.

Table 2 Analyte recovery from premixes (n = 5)^a

Sample Type	Choline spike (mg kg^−1)	Choline found (mg kg^−1)	Recovery (%)	Carnitine spike (mg kg^−1)	Carnitine found (mg kg^−1)	Recovery (%)	Betaine spike (mg kg^−1)	Betaine found (mg kg^−1)	Recovery (%)
a ND = none detected.
Blank premix	50	46.4	92.8	50	45.9	91.8	50	47.9	95.8
	250	238	95.2	250	231	92.4	250	241	96.4
	500	480	96	500	477	95.4	500	485	97.0
Premix	0	523	—	0	1019	—	0	ND	—
	250	755	92.8	250	1265	98.4	250	247	98.8
	1000	1491	96.8	1000	2002	98.1	1000	986	98.6
Premix	0	23.9	—	0	ND	—	0	550.6	—
	100	118	94.1	100	95.7	95.7	100	650.3	99.7
	500	495	94.2	500	479	95.8	500	1043.8	98.6
Liquid premix	0	578	—	0	54.9	—	0	ND	—
	100	676	98	100	155	100	100	97.4	97.4
	1000	1498	92	1000	1004	94.9	1000	967.8	96.8
Liquid premix	0	4.31 × 10^4	—	0	3.97 × 10^4	—	0	ND	—
	10 000	5.23 × 10^4	92	10 000	4.92 × 10^4	95	10 000	0.99 × 10^4	99
	50 000	9.32 × 10^4	100.2	50 000	9.00 × 10^4	100.6	50 000	4.92 × 10^4	98.4

---

The RSD values obtained from repeatability tests using nine premixes were 2.5%, 5.4% and 2.1% for betaine, carnitine and choline, respectively, while the intermediate reproducibility RSD values for these same analytes were 4.7%, 3.5% and 2.4% (Table 3).

Table 3 Repeatability and intermediate reproducibility data obtained from premix analyses (n = 6)

Sample type	Choline			Carnitine			Betaine
	Conc. range (mg kg^−1)	RSDr, (%)	RSDir, (%)	Conc. range (mg kg^−1)	RSDr (%)	RSDir (%)	Conc. range (mg kg^−1)	RSDr (%)	RSDir (%)
Liquid premix	523–22 700	3.9–5.0	4.4–5.7	0–1500	3.4–4.9	4.6–9.3	0–12 500	4.5–7.8	5.9–8.5
Solid premix	230–500 000	3.9–7.7	5.5–8.1	0–100 000	5.4–6.8	6.6–7.9	0–300 000	3.1–9.4	5.2–8.8

---

4 Conclusions

A simple and efficient IC method using non-suppressed detection for the simultaneous determination of carnitine, choline and betaine was developed. After extraction, the total run time of this new method is only 15 min, making it significantly faster than existing assay techniques for the same analytes. This new method also offers a high degree of specificity and is therefore well suited for the routine determination of these analytes in premixes.

Acknowledgements

This project was supported by the Special Fund for Agro-scientific Research in the Public Interest of China (no. 200903006).

References

1. S. H. Zeisel and J. K. Blusztajn, Annu. Rev. Nutr., 1994, 14, 269–296.
2. S. H. Zeisel, Nutrition, 2000, 16, 669–671.
3. S. H. Zeisel, Encyclopedia of Food Sciences and Nutrition. 2003, p. 1251.
4. A. C. Feller and D. Rudman, J. Nutr., 1988, 188, 541–547.
5. A. Steiber, J. Kerner and C. Hoppel, Mol. Aspects Med., 2004, 25(5–6), 455–460.
6. I. A. Brouwer, P. Verhoef and R. Urgert, Arch. Intern. Med., 2000, 160, 2546–2555.
7. J. D. Finkelstein, J. J. Martin and B. J. Harris, J. Nutr., 1983, 113, 519–521.
8. D. J. Bindel, E. C. Titgemeyer, J. S. Drouillard and S. E. Ives, J. Anim. Sci., 2005, 83, 1625–1630.
9. R. G. Twibell and P. B. Brown, J. Nutr., 2000, 130, 95–99.
10. K. Q. Owen, J. L. Nelssen, R. D. Goodband, M. D. Tokach and K. G. Friesen, J. Anim. Sci., 2001, 79, 1509–1515.
11. R. H. Greenwood, E. C. Titgemeyer, G. L. Stokka, J. S. Drouillard and C. A. Loest, J. Anim. Sci., 2001, 79, 254–260.
12. B. V. Lawrence, A. P. Schinckel, O. Adeola and K. Cera, J. Anim. Sci., 2002, 80, 475–482.
13. Y. Z. Wang, Z. R. Xu and J. Feng, Anim. Feed Sci. Technol., 2004, 116, 151–159.
14. D. C. Woollard and H. E. Indyk, J. AOAC Int., 2000, 83, 131–138.
15. J. Galbán, O. Sánchez-Monreal, Y. Andreu, S. de Marcos and J. R. Castillo, Anal. Biochem., 2004, 334, 207–215.
16. G. Panfili, P. Manzi, D. Compagnone and L. Scarciglia, J. Agric. Food Chem., 2000, 48, 3403.
17. T. P. Chendrimada, M. G. Neto, G. M. Pesti, A. J. Davis and R. I. Bakalli, J. Sci. Food Agric., 2002, 82, 1556–1563.
18. R. Maeba and N. Ueta, Anal. Biochem., 2004, 331, 169–176.
19. T. Yoshida, A. Aetake, H. Yamaguchi, N. Nimura and T. Kinoshita, J. Chromatogr., A, 1988, 445, 175–182.
20. K. Li, W. Li and Y. Huang, Clin. Chim. Acta, 2007, 378, 159–163.
21. P. Andrieux, T. KIlinc, C. Perrin and E. Campos-GimÉnez, J. AOAC Int., 2008, 91, 777–785.
22. P. Holm, P. M. Ueland, G. Kvalheim and E. A. Lien, Clin. Chem., 2003, 49, 286–296.
23. L. Vernez, G. Hopfgartner, M. Wenk and S. Krähenbühl, J. Chromatogr., A, 2003, 984, 203–213.
24. M. Laikhtman and J. S. Rohrer, J. AOAC Int., 1999, 82, 1156–1162.
25. The people's Republic of China agriculture industry standard, NY/T 1619-20004, Determination of betaine in feeds ion chromatography.

---