Introduction
Human milk oligosaccharides (HMOs) are a diverse group of carbohydrate-based bioactive compounds primarily composed of glucose, galactose, fucose, N-acetylhexosamine (HexNAc), and sialic acids (N-acetylneuraminic acid and N-glycotylneuraminic acid). HMOs represent the third most predominant solid component of human milk after lactose and fat (Andreas et al. Reference Andreas, Kampmann and Le-Doare2015). Over the past decade, research has demonstrated that HMOs contribute to infant brain development and promote the growth of probiotics, particularly bifidobacteria, a phenomenon referred to as the bifidogenic effect (Katayama et al. Reference Katayama, Sakuma and Kimura2004). As a result, the addition of HMOs to infant formulas has garnered significant interest in efforts to reduce the health gap between formula-fed and breastfed infants.
Currently, the European Union has approved seven HMOs as novel additives for infant formula and food supplements, including 2’-fucosyllactose (2′-FL), 3’-fucosyllactose (3′-FL), 3’-sialyllactose (3′-SL), 6’-sialyllactose (6′-SL), lacto-N-tetraose (LNT), lacto-N-neotetraose (LNnT), and 2’,3-difucosyllatcose (2’,3-FL or DFL). Based on the substitution at the lactose core structure, HMOs are generally classified into neutral, fucosylated, sialylated, and both fucosylated- and sialylated oligosaccharides, accounting for 42–55%, 35–50%, and 12–14% of total colostrum HMOs (Totten et al. Reference Totten, Zivkovic and Wu2012). Due to their high abundance in human milk and relatively simple structures, 2’-FL and 3’-FL have become targets for synthetic production using synthetic biology strategies. To achieve high yields of 2’-FL and 3’-FL, several α-1,2-fucosyltransferases (EC2.4.1.69) and α-1,3-fucosyltransferases (EC2.4.1.152) from diverse microorganisms like Helicobacter pylori (Li et al. Reference Li, Zhang and Li2024), Escherichia coli (Engels and Elling et al. Reference Engels and Elling2014), and Bacteroides gallinaceum (Chen et al. Reference Chen, Wu and Zhu2022) have been integrated into engineered cell factories.
Due to their effectiveness in growth and development, there has been significant interest in identifying novel HMOs. To date, more than 200 oligosaccharides have been discovered in human and animal milk (Huang et al. Reference Huang, Li and Luo2021). Oligosaccharides from animal milk have been shown to exhibit physiological functions similar to HMOs. Therefore, in addition to human and cow milk, other dairy animals, such as goats, buffalos, camels, horses, and donkeys, could serve as potential oligosaccharide sources. Given the trace amounts of most milk oligosaccharides, various extraction and analytical methods have been employed over the years to facilitate their identification. Extraction methods such as solid-phase extraction, enzymatic digestion, and anion-exchange chromatography, combined with advanced analytical techniques like Nano-LC-Chip/Quadropole time-of-flight (QToF) liquid chromatography/mass spectroscopy and ultrahigh-performance liquid chromatography-electrospray ionization-tandem mass spectrometry (UHPLC-ESI-MS/MS), have been used to characterize novel oligosaccharides (Karav et al. Reference Karav, Salcedo and Frese2018; Manabat et al. Reference Manabat, Bhattacharya and Completo2023; Remoroza et al. Reference Remoroza, Liang and Mak2020; Shi et al. Reference Shi, Han and Zhang2021; Wang et al. Reference Wang, Zhang and Huo2023; Yan et al. Reference Yan, Ding and Jin2018).
Buffaloes (Bubalus bubalis) are a multipurpose livestock species primarily found in South Asia, South America, and the Mediterranean region of Europe. Buffalo milk accounts for 15–20% of total global milk production. Due to its mild flavor and high nutritional value, characterized by high levels of protein, fat, and other essential nutrients, buffalo milk is gaining popularity worldwide. The crossbred (Nili-Ravi × Murrah × local) buffalo is an important breed combination that includes Nili-Ravi (originating from Pakistan), Murrah (originating from India), and domestic buffaloes in Guangxi, China. Given this breed’s excellent ability to digest low-quality roughage (Zou et al. Reference Zou, Liang and Yang2007), investigating the milk oligosaccharide compositions of these ternary hybrid (Nili-Ravi × Murrah × local) buffaloes is of particular scientific interest. In this study, buffalo milk samples were collected at both early- and late-lactation stages, and their composition and quantification were analyzed using ultra-performance liquid chromatography coupled with quadrupole time-of-flight mass spectrometry (UPLC-QTOF-MS).
Materials and methods
Milk samples were collected from crossbred (Nili-Ravi × Murrah × local) buffaloes at the Buffalo Research Institute Farm from the Guangxi Academy of Agricultural Sciences (Nanning, China). Standards of 3’-SL, 6’-SL, 3’-FL, and LNnT were purchased from Sigma-Aldrich (St. Louis, MO, USA). All reagents were of analytical or chromatographic grade. Sixteen healthy buffaloes (∼550 kg, within the second parity) were selected for the study, representing early (19, 29, 71, 96, 97, 137, 140, and 146 d) and late (292-1, 292-2, 295, 304, 312, 465, 472, and 647 d) lactation stages. All buffaloes were fed the same diet (Table S1). On May 31, 2024, at 4 a.m., representative milk samples (∼500 mL) were collected from the 16 buffaloes. Immediately after collection, the samples were frozen and shipped on dry ice to Zhejiang University (Hangzhou, China) for further analysis. Buffalo milk oligosaccharides were extracted according to a previously described method (Zhang et al. Reference Zhang, Wang and Chen2019) with minor modifications. Briefly, the samples were centrifuged at 10,000 × g for 30 min at 4°C to remove the upper fat layer. Whey protein was precipitated by adding two volumes of ethanol. After incubating at −20°C for 4 h, the mixture was centrifuged at 4,000 × g for 30 min at 4°C to ensure complete phase separation. The supernatant containing the oligosaccharide fraction was carefully transferred to a new tube and dried by evaporation to obtain enriched oligosaccharides.
Chromatographic separation of oligosaccharides was performed using an ACQUITY UPLC BEH amide column (2.1 mm × 100 mm, 1.7 μm; Waters, Milford, MA) on an ACQUITY UPLC system (Waters) equipped with an evaporative light-scattering detector. The column temperature was maintained at 30℃, with a flow rate of 0.25 mL/min. The mobile phase consisted of acetonitrile (A) and 0.1% formic acid (B). The optimized gradient elution conditions were as follows: 0−1 min, 2% B; 1−10 min, 2%−50% B; 10−12 min, 50% B; 12−13 min, 50−2% B; and 13−18 min, 2% B. The injection volume was 5 μL for both standards and milk samples. For TOF-MS analysis, the analytes were detected using electrospray ionization mass spectrometry (ESI-MS) on a SYNAPT G2-Si mass spectrometer (Waters). The ESI interface was operated in negative-ion mode (ESI–) at 2.2 kV. Desolvation and blocking were performed at 350°C and 120°C, respectively, with a nebulizer and auxiliary gas nitrogen flow rate of 900 L/h. The UPLC system was coupled to a Waters Xevo triple quadrupole detector (TQD), in a heated interface. ESI voltage was 2.2 kV (negative ions), and the mass spectrometer was operated in negative mode. The MS parameters were as follows: source temperature of 120°C, desolvation gas flow rate of 900 L/h, collision gas flow rate of 0.30 mL/min, and capillary voltage of 3 kV.
A series of oligosaccharide standards (3’-SL, 6’-SL, 3’-FL, and LNnT) at concentrations ranging from 25 to 200 µg/mL were prepared using Milli-Q water. All standard samples were analyzed under the optimized conditions described above. The sample concentration (mg/L) was set as the independent variable (X), while the peak area of extracted-ion chromatograms (EIC) in negative ion mode [M-H+] was set as the dependent variable (Y; Table S2). Experimental results were obtained by matching the sugar chain structures to the mass-to-charge ratio of the primary spectra using GlycoWorkbench 2.1 (v2.1 2011, http://www.eurocarbdb.org/applications/ms-tools) software, with reference to the GlycomeDB database (parent ion tolerance: 1 Da). The peak area of EIC was used for oligosaccharide quantification. Statistical analyses were carried out using Student’s t-test with GraphPad Prism v.8.0 (San Diego, CA, USA) (*P < 0.05; **P < 0.01; ***P < 0.001).
Results and discussion
Due to their crucial roles in multiple biological processes, significant efforts have been made over the past decade to identify novel oligosaccharides in human and dairy animal milk (Remoroza et al. Reference Remoroza, Mak and Leoz2018; Yan et al. Reference Yan, Ding and Jin2018). In this study, a total of 97 oligosaccharides with distinct retention times and monosaccharide compositions were identified in the milk of crossbred (Nili-Ravi × Murrah × local) buffaloes (Table S3). The representative UPLC profiles are shown in Fig. S1. Even though buffalo milk contains fewer oligosaccharides than human milk (Huang et al. Reference Huang, Li and Luo2021), it exhibits greater diversity compared to the milk of other non-human mammals, such as cow (Remoroza et al. Reference Remoroza, Mak and Leoz2018; Shi et al. Reference Shi, Han and Zhang2021), goat (Remoroza et al. Reference Remoroza, Liang and Mak2020; Shi et al. Reference Shi, Han and Zhang2021), donkey (Garhwal et al. Reference Garhwal, Sangwan and Mehra2022), elephant (Osthoff et al. Reference Osthoff, Wiese and Deacon2023), and camel (Albrecht et al. Reference Albrecht, Lane and Mariño2014; Shi et al. Reference Shi, Han and Zhang2021).
Based on their glycosylation modification of the lactose core, HMOs are generally classified into three main categories: neutral, fucosylated, and sialylated (acidic) oligosaccharides. The 97 oligosaccharides in buffalo milk were categorized into four groups: 19 neutral (19.59%), 26 fucosylated (26.80%), 32 sialylated (32.99%), and 20 both fucosylated and sialylated (20.62%). Compared to the oligosaccharide composition in other dairy animals (Albrecht et al. Reference Albrecht, Lane and Mariño2014; Karav et al. Reference Karav, Salcedo and Frese2018; Remoroza et al. Reference Remoroza, Mak and Leoz2018; Zhang et al. Reference Zhang, Vervoort and Pan2022), buffalo milk exhibits a distribution pattern more similar to human milk. Specifically, human milk contains approximately 10.39% neutral, 48.05% fucosylated, 15.58% sialylated, and 25.97% both fucosylated and sialylated oligosaccharides. This similarity suggests that buffalo milk may have potential applications in the development of infant formula (Shi et al. Reference Shi, Han and Zhang2021; Totten et al. Reference Totten, Zivkovic and Wu2012).
A recent study on Philippine buffalo milk identified 66 oligosaccharides using Nano-LC-Chip/Quadropole TOF MS (Manabat et al. Reference Manabat, Bhattacharya and Completo2023). Notably, the 97 oligosaccharides identified in crossbred (Nili-Ravi × Murrah × local) buffalo milk in this study included all 66 previously reported oligosaccharides, along with an additional 31 oligosaccharides. As expected, low-molecular weight (<1,000 Da) oligosaccharides commonly found in mammal milk, such as 3’-FL, 3’-SL, 6’-SL, and LNnT, were detected in buffalo milk (Table S3). It is well documented that 2’-FL, the most abundant HMO, is synthesized via the transfer of a fucosyl residue from 5′-diphosphate- L-fucose (GDP-L-fucose) to lactose by α-1,2-fucosyltransferase (α-1,2-FucT, EC2.4.1.69; Lee et al. Reference Lee, Shin and Han2022). Subsequently, 2′-FL and GDP-L-fucose can be converted into DFL by α-1,3-fucosyltransferase (α-1,3-FucT, EC2.4.1.152; Liang et al. Reference Liang, He and Liu2024). Interestingly, 2’-FL was absent in buffalo milk, whereas both 3’-FL and DFL were present. This suggests that α-1,3-FucT may play a dominant role in catalyzing the conversion of 2′-FL and GDP-L-fucose. Further investigation into the α-1,3-FucT gene(s) in buffalo could provide insights. In addition, 17 novel high-molecular weight (> 1,272 Da) oligosaccharides (2141, 6120, 2112, 3122, 7120, 2160, 4131, 8020, 5031, 7030, 2151, 3051, 5022, 6122, 2142, 5041, and 4222), were identified for the first time in this study (Fig. S2). However, further analysis is required to elucidate their detailed structural characteristics.
Human milk contains the highest concentration of oligosaccharides among all mammalian species, ranging from 12−16 g/L in colostrum to 5−11 g/L in mature milk (Borewicz et al. Reference Borewicz, Gu and Saccenti2019; Zhang et al. Reference Zhang, Vervoort and Pan2022). In contrast, oligosaccharide concentrations in dairy animals, including cows, goats, sheep, camels, donkeys, and elephants, vary significantly from 0.02 to 4.6 g/L (Albrecht et al. Reference Albrecht, Lane and Mariño2014; Licitra et al. Reference Licitra, Li and Liang2019; Shi et al. Reference Shi, Han and Zhang2021; Wang et al. Reference Wang, Zhang and Huo2023; Zhang et al. Reference Zhang, Vervoort and Pan2022; Zhong et al. Reference Zhong, Yang and Han2024). In this study, oligosaccharide concentrations were determined using calibration curves (Table S2), yielding 416.6 ± 25.86 in early-lactation and 368.3 ± 10.29 mg/L in late-lactation buffalo milk. The total oligosaccharide concentration in buffalo milk, particularly in mature milk, was comparable to that of camels (0.40 − 0.48 g/L; Shi et al. Reference Shi, Han and Zhang2021; Zhang et al. Reference Zhang, Vervoort and Pan2022) and donkeys (∼0.42 g/L; Garhwal et al. Reference Garhwal, Sangwan and Mehra2022; Licitra et al. Reference Licitra, Li and Liang2019). However, it was considerably higher than in other dairy animals such as cows (0.08 − 0.10 g/L; Zhong et al. Reference Zhong, Yang and Han2024), goats (0.06 − 0.35 g/L; Shi et al. Reference Shi, Han and Zhang2021; van Leeuwen et al. Reference van Leeuwen, Te Poele and Chatziioannou2020; Zhang et al. Reference Zhang, Vervoort and Pan2022), sheep (0.02 − 0.04 g/L; Albrecht et al. Reference Albrecht, Lane and Mariño2014; Zhang et al. Reference Zhang, Vervoort and Pan2022), and elephants (Osthoff et al. Reference Osthoff, Wiese and Deacon2023). Moreover, buffalo milk offers a longer lactation period (> 10 months) compares to donkeys (∼7 months), goats (∼9.5 months), and sheep (∼7 months), making it a potentially valuable milk source for both adult and infant formula production. Notably, the relative abundance of the four oligosaccharide categories remained stable across early- and late-lactation stages in crossbred (Nili-Ravi × Murrah × local) buffalo milk: neutral (8.44%–9.40%), fucosylated (31.69%–33.99%), sialylated (42.20%–45.18%), and both fucosylated and sialylated (14.30%–14.69%; Fig. 1). The most abundant oligosaccharides in mature buffalo milk were 3’-SL, DFL, 6’-SL, and 2142 at 108.80 ± 2.72, 104.60 ± 3.69, 44.25 ± 5.46 and 27.09 ± 1.07 mg/L, respectively (Table 1).
Figure 1. Percentage distribution of neutral-, fucosylated-, sialylated- and both fucosylated and sialylated-oligosaccharides in buffaloes. (A) early-lactation stage; (B) late-lactation stage.
Table 1. Concentrations of top four milk oligosaccharides in early- and late-lactation stages of crossbred (Nili-Ravi × Murrah × local) buffaloes (mg/L)
3’-SL, 3’-sialyllactose; DFL, difucosyllactose; 6’-SL, 6’-sialyllactose. Data represent the mean ± SD (n = 8). Statistical significance between oligosaccharides from early- and late-lactation stages was assessed using Student’s t-test.
* P < 0.05; *** P < 0.001.
Sialylated oligosaccharides, such as 3’SL and 6’SL, have been shown to exhibit prebiotic effects, prevent bacterial adhesion, inhibit pathogens, promote infant brain development, and regulate both local and systemic immunity (Zhu et al. Reference Zhu, Zhang and Zhang2023). Notably, necrotizing enterocolitis, a leading cause of morbidity and mortality in premature or very low-birth-weight infants, has been linked to the activation of the toll-like receptor 4 or nuclear factor kappa-B pathways. Studies have demonstrated that sialylated oligosaccharides, particularly 3’SL and 6’SL, can protect against necrotizing enterocolitis in mice by inhibiting these pathways (Huang et al. Reference Huang, Li and Luo2021; Sodhi et al. Reference Sodhi, Ahmad and Fulton2023, Reference Sodhi, Wipf and Yamaguchi2021; Wang et al. Reference Wang, Zhang and Guo2019). In human milk, the concentration of 3’SL ranges from 130 to 190 mg/L in colostrum and mature milk (Soyyilmaz et al. Reference Soyyilmaz, Miks and Röhrig2021). In this study, 3’-SL was the most abundant oligosaccharide in buffalo milk, with concentrations ranging from 108.80 ± 2.72 to 124.55 ± 5.05 mg/L across early- and late-lactation stages (Table 1). A previous study found that Holstein cow colostrum contained a significantly higher 3’-SL concentration (717 ± 27 mg/L; Nakamura et al. Reference Nakamura, Kawase and Kimura2003). However, in mature milk, 3’-SL levels decline to 47–128 mg/L (Fong et al. Reference Fong, Ma and McJarrow2011; Wang et al. Reference Wang, Zhou and Gong2020), which is comparable to the levels found in buffalo milk in this study. Interestingly, 3’-SL is also the predominant oligosaccharide in the milk of cows (Albrecht et al. Reference Albrecht, Lane and Mariño2014; Nakamura et al. Reference Nakamura, Kawase and Kimura2003; Zhong et al. Reference Zhong, Yang and Han2024), and camels (Jiang et al. Reference Jiang, Sun and Lin2024; Shi et al. Reference Shi, Han and Zhang2021; Zhang et al. Reference Zhang, Vervoort and Pan2022), suggesting that its high abundance may be a unique feature of ruminant milk.
Fucosylated oligosaccharides, such as 2’FL, 3’FL and DFL, selectively promote the proliferation of probiotics, particularly Bifidobacterium species, in the gastrointestinal tracts of both humans and animals. Specifically, 2’FL, 3’FL, and DFL support the growth of Bifidobacterium bifidum, Bifidobacterium breve, Bifidobacterium infantis, and Bifidobacterium longum, as well as dominant commensal fibrolytic bacteria such as Bacteroides thetaiotaomicron, Bacteroides fragilis, and Bacteroides vulgatus (Dedon et al. Reference Dedon, Hilliard and Rani2023; Salli et al. Reference Salli, Hirvonen and Siitonen2021). Both 2’FL and 3’FL serve as carbon sources for certain potentially pathogenic bacteria, including Clostridium perfringens, Cronobacter sakazakii, and Escherichia coli. Interestingly, DFL did not promote the growth of any of the 16 potentially pathogenic bacterial species evaluated by Salli et al. (Reference Salli, Hirvonen and Siitonen2021). Due to its exclusive role in supporting probiotics and its ability to convert fucose-derived pyruvate into acetyl phosphate and formate, DFL has garnered considerable attention. Unlike the widely observed fucosylated-oligosaccharides, such as 2’-FL and 3’-FL, which are commonly found in human and animal milk, DFL is rarely detected in non-human mammals. In this study, the concentration of DFL in buffalo milk during late lactation was 103.88 ± 2.27 mg/L, comparable to that in mature human milk (110–120 mg/L; Gu et al. Reference Gu, Wang and Beijers2021; Olivares et al. Reference Olivares, Albrecht and De Palma2015). Considering its critical roles in various biological functions (Dedon et al. Reference Dedon, Hilliard and Rani2023; Salli et al. Reference Salli, Hirvonen and Siitonen2021), buffalo milk is a promising candidate for infant formula development. In addition, the functional properties of buffalo milk oligosaccharides, such as 2142, remain unknown, highlighting the need for further in-depth biological characterization. In summary, a total of 97 oligosaccharides were identified in crossbred (Nili-Ravi × Murrah × local) buffaloes from Guangxi, China, demonstrating greater diversity than oligosaccharides present in milk of other non-human mammals. The concentrations of oligosaccharides ranged from 368.3 to 416.6 mg/L across the early- and late-lactation stages. Notably, 3’-SL, DFL, and 6’-SL, which have multiple biological functionalities, were the most abundant oligosaccharides in buffalo milk. Additionally, 17 novel oligosaccharides including an abundant one (2142) were identified. These findings suggest that buffalo milk has the potential to be developed as a high-quality source for adults and infants.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/anr.2025.10011.
Acknowledgements
This work was supported by the National Key R&D Program of China (2023YFD1300902), the Fundamental Research Funds for the Central Universities (226-2024-0Z0139), and the Guangxi Science and Technology Major Project (guike AA22068099-5). The authors thank Mr. Wei-Dong Zeng of The Experimental Teaching Center, College of Animal Sciences, Zhejiang University, for facility support.
Author contributions
J.H.: Investigation, Methodology, Data curation; H.W.: Investigation, Software; Y.L.: Investigation; D.W.: Investigation; X.X.: Investigation; H.Z.: Methodology; S.Y.: Supervision; J.W.: Supervision, Funding acquisition; B.L.: Supervision, Funding acquisition; Q.W.: Conceptualization, Supervision, Funding acquisition, Writing.
Conflicts of interest
The authors report no declarations of interest.
Open access
