Simple Liquid Chromatography–Mass Spectrometry Method for

Article pubs.acs.org/JAFC

Simple Liquid Chromatography−Mass Spectrometry Method for Quantification of Major Free Oligosaccharides in Bovine Milk Zhiqian Liu,† Peter Moate,‡ Ben Cocks,†,§ and Simone Rochfort*,†,§ †

Biosciences Research Division, Department of Environment and Primary Industries, AgriBio, 5 Ring Road, Bundoora, Victoria 3083, Australia ‡ Farming Systems Research Division, Department of Environment and Primary Industries, Ellinbank Centre, 1301 Hazeldean Road, Ellinbank, Victoria 3821, Australia § School of Applied Systems Biology, La Trobe University, Bundoora, Victoria 3083, Australia S Supporting Information *

ABSTRACT: Free oligosaccharides (OS) are a significant functional component of milk that are difficult to quantitate. A simple method for quantitative analysis of the major free OS in bovine milk is described. Following a defatting step, protein elimination was performed by ultrafiltration. OS were separated by hydrophilic interaction liquid chromatography (HILIC) and detected by an Orbitrap mass analyzer in negative mode. The method is sensitive [with a limit of detection (LOD) for all representative OS of 0.99). Precision and Accuracy of the Method. This method also shows excellent reproducibility in the peak area for all representative OS and all injection levels. The peak area variation between repeated analyses of the five representative OS standards is below 9% for the low injection level (0.5 ng) and below 3% for intermediate (5 ng) and high (100 ng) injection levels (Table 4). The overall measurement accuracy of the method ranges from 91 to 104% for the three concentrations of OS standards (Table 4). Recovery of Spiked Standards. The overall recovery of the four representative OS (stachyose, verbascose, 6′-SL, and 6′-SLN) spiked at two different levels (0.1 and 1 μg in 0.4 mL of defatted milk) is between 91 and 105% (Table 5). This suggests that no significant loss of the OS occurred during the ultrafiltration step nor was there a significant matrix effect on the detection of these compounds. Method Application. The OS concentrations in 32 samples of milk were determined using the validated method. The major OS detected in these samples were the same as those found in the reference sample. The relative abundances (averaged from the 32 samples) of the 13 OS are shown in Table 6. Again, six OS (Trisa, OS-A, 3′-SL, 6′-SL, DSL, and GNL) were found in a larger amount, with 3′-SL and 6′-SL being the most dominant species, followed by Trisa, whereas OS-B, OS-C, OS-D, OS-E, OS-F, OS-G, and 6′-SLN were the lower abundance OS species in these samples. Large variation was also observed between samples in the concentration of all major OS. For example, the 3′-SL content varied from 116 to 199 μg/mL across the 32 samples; the 6′-SL content varied from 6 to 32 μg/mL across the 32 samples; and the 6′-SLN content varied from 0.27 to 1.2 μg/mL across the 32 samples (Figure 2). The magnitude of variation appeared to be similar for milks from three different diets. It is also interesting to note that the contents (as measured by peak areas) of some of the major OS are correlated with each other, with the most significant correlation (R2 > 0.6) being found between the pairs 6′-SL/6′-SLN, OS-C/GNL, and OS-D/OS-G (Figure 3). When the whole data matrix (13 OS contents of 32 samples) was subjected to PCA analysis, no clear separation was observed between the three feed groups, indicating that the intake of almond hulls or citrus pulp had little effect on the OS

quantifiable and are thus considered as major OS (Table 1 and Figure 1). It is worth noting that all 13 major OS are of lower degree of polymerization (DP) type and the dominant species (Trisa, OS-A, 3′-SL, 6′-SL, DSL, and GNL) contain three or four monosaccharide units. The composition of the major OS identified using accurate mass data was further confirmed by the standards (in the case of 3′-SL, 6′-SL, and 6′-SLN) and/or MS/MS data (see Figure S-1 of the Supporting Information). It is interesting to note that, of the pure hexose-based OS category, species containing more than three sugar units were not found in the sample nor were any fucose-containing OS species. Lactose, lactose phosphate, and N-acetyllactosamine were not considered as OS in this work. Comparison of Sample Pretreatment Methods. Stachyose was used as an internal standard in our preliminary analysis, and comparable results were obtained with and without the internal standard. Therefore, no internal standards were used in subsequent experiments. The concentrations (as judged by peak areas) of the 13 major OS in the reference milk sample following different protein elimination procedures were measured in parallel. The experiment was designed in such a way that a same dilution factor was applicable to all three methods. The peak areas of the 13 OS generated with method 2 (acetonitrile precipitation) were highly correlated with those of method 1 (standard method), but the overall response was about 20% lower (y = 0.7992x; R2 = 0.9996). This suggests that using acetonitrile to precipitate milk protein could lead to a significant underestimation of OS. In contrast, the peak areas of the major OS measured using protein elimination method 3 (ultrafiltration) were not only highly correlated with those obtained with method 1 but also showed slightly higher overall ion abundance (y = 1.0733x; R2 = 0.9998), implying a higher OS recovery using this new sample pretreatment method. The ultrafiltration method (method 3) also displayed better measurement reproducibility than the standard method (method 1), with a RSD of the peak area of replicate extractions below 4% for 12 of the 13 major OS (Table 2). Consequently, the ultrafiltration was chosen as the protein elimination technique for further method validation. LOD, LOQ, and Linear Range of Representative OS. This LC−MS method is highly sensitive for the detection of both acidic and neutral OS with a LOD for all species below 0.1 C

dx.doi.org/10.1021/jf5037849 | J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry

Article

Figure 1. Extracted ion chromatogram of the 13 major OS detected in the reference milk sample. The deprotonated ion of each compound was extracted using a window of calculated accurate mass of m/z ±0.005.

determining the concentration of total fructo-oligosaccharide in plant tissue.26 As a result, measuring the concentration of the major OS species is important to evaluate the potential beneficial value of OS in a milk sample. Irrespective of analytical methods, absolute quantification of individual OS is currently possible only for a few major species, for which the standards are available.16,18 Consequently, relative quantification (using ion intensity or peak area) is the only feasible way for estimating the content of most OS in milk. Indeed, ion intensity was used in a number of studies to estimate the relative content of OS in various milk samples,13,14,20 but a robust and validated method for relative quantification of major OS is still lacking. Besides, all of the

accumulation in milk (see Figure S-2 of the Supporting Information).

■

DISCUSSION The beneficial biological functions of OS are increasingly being recognized especially for neonates. It is generally accepted that the OS found in bovine milk are structurally similar to those in human milk.5 Consequently, the use of bovine milk OS in infant formula and pharmaceutical and food industries is expected to intensify. Because OS in bovine milk are present as a mixture of a large number of linear and branched molecules with 3−10 monosaccharide units, measuring the total OS concentration in bovine milk is not as straightforward as D

dx.doi.org/10.1021/jf5037849 | J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry

Article

Table 2. Measurement Reproducibility of Methods 3 and 1

Table 5. Recovery (%) of Spiked Standards (n = 3) spike I (0.1 μg)

peak area RSD (%) (n = 3) OS name

method 3

Trisa 3′-SL 6′-SL 6′-SLN DSL GNL OS-A OS-B OS-C OS-D OS-E OS-F OS-G

method 1

1.68 1.55 2.25 2.80 0.47 0.77 2.02 3.13 2.46 8.18 0.37 2.45 3.71

4.50 1.97 1.00 1.09 3.50 6.49 1.07 4.02 4.94 7.81 4.86 7.16 8.08

OS

SD (%)

RSD (%)

recovery (%)

SD (%)

RSD (%)

6′-SL 6′-SLN stachyose verbascose

94.93 103.70 95.28 93.09

3.48 3.43 2.34 2.75

3.66 3.31 2.45 2.95

93.71 105.29 94.23 91.07

2.75 3.61 4.23 1.53

2.93 3.43 4.49 1.68

Table 6. Relative Abundance (Peak Area) of the 13 Major OS (n = 32)

3′-SL 6′-SL 6′-SLN stachyose verbascose

ion quantified [M [M [M [M [M

− − − − −

−

H] H]− H]− H]− H]−

mean ± SD

OS name OS-A OS-B OS-C OS-D 6′-SL 3′-SL OS-E DSL GNL Trisa 6′-SLN OS-F OS-G

Table 3. LOD, LOQ, and Linearity Range of Five Representative OS OS

spike II (1 μg)

recovery (%)

LOD (ng)

LOQ (ng)

linear range (ng)

R2

0.05 0.05 0.05 0.05 0.05

0.25 0.25 0.25 0.25 0.25

0.25−500 0.25−250 0.25−250 0.25−100 0.25−250

0.9945 0.9970 0.9974 0.9982 0.9937

3716387 751707 528108 45560 15374375 87002494 703177 2730046 3592667 9060101 653564 183795 113112

± ± ± ± ± ± ± ± ± ± ± ± ±

1179608 349185 227577 24728 6650642 12984817 221299 652066 3376552 2021898 286207 60414 55621

Moreover, the Luna HILIC stationary phase that we used separated all major OS from lactose, thus avoiding the potential ion suppression effect of this most abundant sugar on the detection of other minor analogues. The minimal matrix effect observed in the recovery test may well have resulted from this. It is worth indicating that, except for 3′-SL and 6′-SL, no attempt was made to separate and quantify individual isomers of other major OS species. As a result, most major OS elute as a single peak, facilitating relative quantification. Because of the limited availability of OS standards, we have used two galacto-oligosaccharides, stachyose (DP of 4) and verbascose (DP of 5), in the method validation to represent neutral OS. Given the similarity in molecular size and structure between these two compounds and the major neutral OS in milk, it is appropriate to assume that they have comparable ionization efficiency for MS detection. The fact that all of the tested OS standards had an on-column LOD of 0.05 ng suggests that, without any enrichment step, low-abundance OS (from 0.01 μg/mL) can be detected by this method. It should be pointed out that all OS species showed much higher response in negative ESI mode than in positive mode in our system and deprotonated molecular ions are dominant in all cases. All OS detected in the reference sample have been identified previously in different studies.9,14 Among the 13 major OS (of

previous analyses adopted a standard and complex protocol for protein elimination in milk samples, which consists of chloroform/methanol extraction, residual protein precipitation by ethanol, and OS fraction drying/reconstitution. Such a multi-step sample pretreatment procedure constitutes a bottleneck for measurement throughput. To simplify the sample preparation step, we have evaluated the performance of two new methods that are simpler than the standard method. Acetonitrile precipitation appeared to be effective for protein elimination from milk, but this solvent also caused partial loss of OS in the supernatant, leading to a significant reduction in the OS yield. This may simply result from co-precipitation of OS in the presence of acetonitrile as reported by Ku et al. in the case of fructo-oligosaccharides.27 In contrast, ultrafiltration is not only simple to perform but also afforded a higher OS yield with better reproducibility compared to the standard method. Consequently, we believe ultrafiltration has clear advantages over the standard method for protein elimination in milk samples prior to OS analysis. Nano-LC−chip/TOF MS has been a technique of choice for OS profiling in milk and also for glycan profiling in human serum.19,28,29 We have shown that HILIC separation coupled to an Orbitrap MS is also highly suitable for OS profiling, owing to its superior sensitivity and high mass accuracy capability.

Table 4. Method Precision (% RSD; n = 5) and Accuracy (% ± SD; n = 5) 0.5 ng OS

precision (% RSD)

3′-SL 6′-SL 6′-SLN stachyose verbascose

4.53 8.09 2.78 3.71 7.65

5 ng

accuracy (% ± SD)

precision (% RSD)

± ± ± ± ±

1.76 2.49 1.56 1.19 1.98

91.1 93.4 96.9 93.5 97.8

4.1 7.6 2.7 3.5 7.5

100 ng accuracy (% ± SD) 95.5 99.5 100.1 101.4 99.4

E

± ± ± ± ±

1.7 2.5 1.6 1.2 2.0

precision (% RSD) 1.64 1.65 1.02 2.82 1.38

accuracy (% ± SD) 96.5 99.9 102.2 103.6 99.5

± ± ± ± ±

1.6 1.6 1.0 2.9 1.4

dx.doi.org/10.1021/jf5037849 | J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry

Article

Figure 3. Significant correlations between the contents of some OS species (n = 32).

Figure 2. Variation in 3′-SL, 6′-SL, and 6′-SLN concentrations across the 32 samples (numbers 1−10, ALH; numbers 11−20, CIT; and numbers 21−32, CON.

contents produced by individual cows of the same breed and under the same feeding regime. More samples over different seasons need to be tested to ascertain whether this is genetically based. Moreover, to increase the concentration of OS in cow milk requires a thorough understanding of OS metabolism in the mammary gland and in milk itself.21 It has already been demonstrated that key genes for oligosaccharide synthesis undergo temporal changes in expression, and this expression differs between species.30 In this regard, our finding regarding the correlation between some OS, which has not been reported previously to the best of our knowledge, is valuable for elucidating the biosynthesis pathways of OS. The correlation found between some OS species may be an indication of common biochemical steps used in their biosynthesis. Our work also demonstrates the variability in the concentration of OS in milk from different cows in the same herd that is not associated with the feeding regime used in this study. This suggests that cows may be selected for their ability to produce larger quantities of certain OS that may be of value in human health products.

which two are isomers) quantifiable, the structure of 12 have been reported previously.9 Our observation of 3′-SL and 6′-SL being the most abundant OS across all samples has confirmed the results of other researchers.12,16,18 The fact that 3′-SLN was not detected in any sample is also in agreement with the observation by Fong et al.18 However, our finding of 6′-SLN being a low-abundance OS is significantly different to the results of Martin-Sosa et al.15 and Tao et al.,9 who found this same compound to be one of the most abundant OS. Such a discrepancy may be attributable to samples being derived from different cow breeds and/or lactation stages. An exhaustive search for more minor OS through sample enrichment was not the aim of this work. Most of the previous work focused on the comprehensive identification of OS in human milk and animal milk and on the variation of the most abundant OS species, as influenced by various factors, e.g., lactation stages, mammal species, and bovine breeds. Our study has revealed the large variation of OS F

dx.doi.org/10.1021/jf5037849 | J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry

Article

(13) Tao, N.; DePeters, E. J.; German, J. B.; Grimm, R.; Lebrilla, C. B. Variation in bovine milk oligosaccharides during early and middle lactation stages analysed by high-performance liquid chromatography− chip/mass spectrometry. J. Dairy Sci. 2009, 92, 2991−3001. (14) Sundekilde, U. K.; Barile, D.; Meyrand, M.; Poulsen, N. A.; Larsen, L. B.; Lebrilla, C. B.; German, J. B.; Bertram, H. C. Natural variability in bovine milk oligosaccharides from Danish Jersey and Holstein-Friesian breeds. J. Agric. Food Chem. 2012, 60, 6188−6196. (15) Martin-Sosa, S.; Martin, M. J.; Garcia-Pardo, L.; Hueso, P. Sialyloligosaccharides in human and bovine milk and in infant formulas: Variations with the progression of lactation. J. Dairy Sci. 2003, 86, 52−59. (16) Nakamura, T.; Kawase, H.; Kimura, K.; Watanabe, Y.; Ohtani, M.; Arai, I.; Urashima, T. Concentrations of sialyloligosaccharides in bovine colostrum and milk during the prepartum and early lactation. J. Dairy Sci. 2003, 86, 1315−1320. (17) Fong, B.; Ma, K.; McJarrow, P. Quantification of bovine milk oligosaccharides using liquid chromatography−selected reaction monitoring−mass spectrometry. J. Agric. Food Chem. 2011, 59, 9788−9795. (18) Marino, K.; Lane, J. A.; Abrahams, J. L.; Struwe, W. B.; Harvey, D. J.; Marotta, M.; Hickey, R. M.; Rudd, P. M. Method for milk oligosaccharide profiling by 2-aminobenzamide labelling and hydrophilic interaction chromatography. Glycobiology 2011, 21, 1317−1330. (19) Ninonuevo, M. R.; Lebrilla, C. B. Mass spectrometric method for analysis of oligosaccharides in human milk. Nutr. Rev. 2009, 67 (Supplement 2), S216−S226. (20) Barile, D.; Marotta, M.; Chu, C.; Mehra, R.; Grimm, R.; Lebrilla, C. B.; German, J. B. Neutral and acidic oligosaccharides in HolsteinFriesian colostrum during the first 3 days of lactation measured by high-performance liquid chromatography on a microfluidic chip and time-of-flight spectrometry. J. Dairy Sci. 2010, 93, 3940−3949. (21) Wickramasinghe, S.; Hua, S.; Rincon, G.; Islas-Trejo, A.; German, J. B.; Lebrilla, C. B.; Medrano, J. F. Transcriptome profiling of bovine milk oligosaccharide metabolism genes using RNAsequencing. PLoS One 2011, 6 (4), No. e18895. (22) Aldredge, D. L.; Geronimo, M. R.; Hua, S.; Nwosu, C. C.; Lebrilla, C. B.; Barile, D. Annotation and structural elucidation of bovine milk oligosaccharides and determination of novel fucosylated structures. Glycobiology 2013, 23, 664−676. (23) National Health and Medical Research Council (NHMRC), Australian Government. Australian Code of Practice for the Care and Use of Animals for Scientific Purposes, 7th ed.; NHMRC, Australian Government: Canberra, Australian Capital Territory (ACT), Australia, 2004; http://www.nhmrc.gov.au/publications/synopses/ea16syn.htm. (24) Coppa, G. V.; Pierani, P.; Zampini, L.; Carloni, I.; Carlucci, A.; Gabrielli, O. Oligosaccharides in human milk during different phases of lactation. Acta Paediatr. Suppl. 1999, 430, 89−94. (25) Liu, Z.; Rochfort, S. A fast chromatography−mass spectrometry (LC−MS) method for quantification of major polar metabolites in plants. J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2013, 912, 8− 15. (26) Liu, Z.; Mouradov, A.; Smith, K.; Spangenberg, G. An improved method for quantitative analysis of total fructans in plant tissues. Anal. Biochem. 2011, 418, 252−259. (27) Ku, Y.; Jansen, O.; Oles, C. J.; Lazar, E. Z.; Rader, J. I. Precipitation of inulins and oligoglucoses by ethanol and other solvents. Food Chem. 2003, 81, 125−132. (28) Hua, S.; An, H. J.; Ozcan, S.; Ro, G. S.; Soares, S.; DeVereWhite, R.; Lebrilla, C. B. Comprehensive native glycan profiling with isomer separation and quantitation for the discovery of cancer biomarkers. Analyst 2011, 136, 3663−3671. (29) Ninonuevo, M. R.; Park, Y.; Yin, H.; Zhang, J.; Ward, R. E.; Clowers, B. H.; German, J. B.; Freeman, S. L.; Killeen, K.; Grimm, R.; Lebrilla, C. B. A strategy for annotating the human milkglycome. J. Agric. Food Chem. 2006, 54, 7471−7480. (30) Maksimovic, J.; Sharp, J. A.; Nicholas, K. R.; Cocks, B. G.; Savin, K. Conservation of the ST6Gal I gene and its expression in the mammary gland. Glycobiology 2011, 21, 467−481.

In conclusion, a novel LC−MS method has been developed, which, using a routine HILIC−LC and an Orbitrap mass analyzer, is able to provide relative quantification of 13 major OS in bovine milk with minimum sample preparation. The method is sensitive, accurate, and reproducible and can be used as an alternative to the widely used nano-LC−chip/TOF-based method. The applicability of this method to a large number of mature milk samples collected from cows after different feed treatments allowed us to confirm that almond hulls and citrus pulp when used as a feed supplement have no adverse effect on the accumulation of OS in milk.

■

ASSOCIATED CONTENT

S Supporting Information *

MS/MS pattern of 13 OS (Figure S-1) and PCA plot of PC1 and PC2 showing unsupervised classification of 32 milk samples based on the content of 13 major OS (Figure S-2). This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Stelwagen, K.; Carpenter, E.; Haigh, B.; Hodgkinson, A.; Wheeler, T. T. Immune components of bovine colostrum and milk. J. Anim. Sci. 2009, 87, 3−9. (2) Bentley, P. A.; Shao, Y.; Misra, Y.; Morielli, A. D.; Zhao, F. Q. Characterization of bovine glucose transporter 1 kinetics and substrate specificities in Xenopus oocytes. J. Dairy Sci. 2012, 95, 1188−1197. (3) Boehm, G.; Stahl, B. Oligosaccharides from milk. J. Nutr. 2007, 137 (Supplement 2), 847S−849S. (4) Gopal, P. K.; Gill, H. S. Oligosaccharides and glycoconjugates in bovine milk and colostrum. Br. J. Nutr. 2000, 84 (Supplement 1), S69−S74. (5) Zivkovic, A. M.; Barile, D. Bovine milk as a source of functional oligosaccharides for improving human health. Adv. Nutr. 2011, 2, 284−289. (6) Ninonuevo, M. R.; Park, Y.; Yin, H.; Zhang, J.; Ward, R. E.; Clowers, B. H.; German, J. B.; Freeman, S. L.; Killeen, K.; Grimm, R.; Lebrilla, C. B. A strategy for annotating the human milkglycome. J. Agric. Food Chem. 2006, 54, 7471−7480. (7) Wang, B. Sialic acid is an essential nutrient for brain development and cognition. Annu. Rev. Nutr. 2009, 29, 177−222. (8) Tao, N.; Wu, S.; Kim, J.; An, H. J.; Hinde, K.; Power, M. L.; Gagneux, P.; German, J. B.; Lebrilla, C. B. Evolutionary glycomics: Characterization of milk oligosaccharides in primates. J. Proteome Res. 2011, 10, 1548−1557. (9) Tao, N.; DePeters, E. J.; Freeman, S.; German, J. B.; Grimm, R.; Lebrilla, C. B. Bovine milk glycome. J. Dairy Sci. 2008, 91, 3768−3778. (10) Urashima, T.; Taufik, E.; Fukuda, K.; Asakuma, S. Recent advances in studies on milk oligosaccharides of cows and other domestic farm animals. Biosci., Biotechnol., Biochem. 2013, 77, 455− 466. (11) Veh, R. W.; Michalski, J. C.; Corfield, A. P.; Sander-Wewer, M.; Gies, D.; Schauer, R. New chromatographic system for the rapid analysis and preparation of colostrum sialyloligosaccharides. J. Chromatogr. A 1981, 212, 312−322. (12) MaJarrow, P.; van Amelsfort-Schoonbeek, J. Bovine sialyl oligosaccharides: Seasonal variations in their concentrations in milk, and a comparison of the colostrums of Jersey and Friesian cows. Int. Dairy J. 2004, 14, 571−579. G

dx.doi.org/10.1021/jf5037849 | J. Agric. Food Chem. XXXX, XXX, XXX−XXX