A novel approach identified the FOLR1 gene, a putative regulator of milk protein synthesis
- First Online:
- Cite this article as:
- Menzies, K.K., Lefèvre, C., Sharp, J.A. et al. Mamm Genome (2009) 20: 498. doi:10.1007/s00335-009-9207-4
This study has utilised comparative functional genomics to exploit animal models with extreme adaptation to lactation to identify candidate genes that specifically regulate protein synthesis in the cow mammary gland. Increasing milk protein production is valuable to the dairy industry. The lactation strategies of both the Cape fur seal (Artocephalus pusillus pusillus) and the tammar wallaby (Macropus eugenii) include periods of high rates of milk protein synthesis during an established lactation and therefore offer unique models to target genes that specifically regulate milk protein synthesis. Global changes in mammary gene expression in the Cape fur seal, tammar wallaby, and the cow (Bos taurus) were assessed using microarray analysis. The folate receptor α (FOLR1) showed the greatest change in gene expression in all three species [cow 12.7-fold (n = 3), fur seal 15.4-fold (n = 1), tammar 2.4-fold (n = 4)] at periods of increased milk protein production. This compliments previous reports that folate is important for milk protein synthesis and suggests FOLR1 may be a key regulatory point of folate metabolism for milk protein synthesis within mammary epithelial cells (lactocytes). These data may have important implications for the dairy industry to develop strategies to increase milk protein production in cows. This study illustrates the potential of comparative genomics to target genes of interest to the scientific community.
Increasing milk protein production is an ongoing challenge for the dairy industry. The use of comparative functional genomics to exploit animal models with extreme adaptation to lactation, in particular, a high rate of milk protein production, is an alternative approach to identify candidate genes that specifically regulate protein synthesis in the mammary gland. Generally, eutherians such as cows (Bos taurus) secrete milk that does not change dramatically in composition after production of colostrum (Akers 2002; Jenness 1986), and the concentrations of protein and lipid remain relatively constant for individual cows. Previous studies in the dairy cow have shown that the mammary gland is not operating at maximum capacity to produce milk protein (Girard and Matte 2005; Mackle et al. 2000). Therefore, these candidate genes may provide either new opportunities for improved breeding programs or potential nutritional intervention to improve efficiency of milk protein production.
Milk protein content of otariid species (fur seals and sea lions) is more than double that of cow’s milk and is the major component of milk that is synthesized de novo in the mammary gland (Akers 2002; Georges et al. 2001). Some species of fur seals, including the Cape fur seal (Artocephalus pusillus pusillus), feed their young on shore for approximately 3 days and then forage continuously at sea for up to 3 weeks (Bonner 1984; Oftedal et al. 1987). During time at sea the mammary gland downregulates milk production to less than 20%, but maximum milk production resumes each time the mother returns to shore to feed the pup (Arnould and Boyd 1995). The protein concentration of Cape fur seals is among the highest of any mammal and ranges from 10 to 20% (Georges et al. 2001; Goldsworthy and Crowley 1999; Trillmich and Lechner 1986) compared to that of cow milk at 2.5–4.5% (Akers 2002). Fur seal milk is also rich in lipid, but very little de novo lipid synthesis occurs in the mammary gland as lipids are synthesised from circulating triacylglycerols absorbed from the animal’s marine diet (Georges et al. 2001; Iverson et al. 1997). Lactose synthesis by the mammary gland is absent and a low concentration of glucose is present in milk (Urashima et al. 2001). The increase in milk production during established lactation on-shore correlates with an upregulation of milk protein gene expression and milk protein synthesis (Sharp et al. 2005).
During lactation in the tammar wallaby (Macropus eugenii), milk protein content and production increase dramatically in the second half of lactation (Nicholas 1988) to assist the physiological and nutritional development of the altricial young (Green 1984; Nicholas 1988). Milk production during the first 200 days of lactation is low in lipid and protein but rich in oligosaccharides. At around day 200 of lactation when the young becomes endothermic and begins periodically to exit the pouch, milk production increases significantly and milk is low in complex carbohydrates and high in lipid and protein content (Messer and Green 1979; Nicholas 1988). In addition, the tammar practices asynchronous concurrent lactation (Nicholas 1988) which suggests that this substantial increase in milk protein production is regulated locally within the mammary gland.
Both the Cape fur seal and the tammar show periods of increased milk protein production during an established lactation. Furthermore, lipid and lactose syntheses are minimal in the mammary gland of the lactating Cape fur seal. Collectively, these two species provide mammary gland models to directly target genes and molecular processes regulating milk protein synthesis. This study has exploited these models with microarray technology to identify differentially expressed mammary genes common to the lactating cow mammary gland and the mammary glands of the Cape fur seal and tammar at periods of increased milk protein production during an established lactation. This gene list was used to subsequently identify genes that specifically regulate protein synthesis in the mammary gland of the dairy cow.
Materials and methods
Microarray analysis of RNA from four early-lactating tammars (two animals at day 30 of lactation [d30L], two animals at d105L) and four late-lactating (two animals at d193L, two animals at d238L) tammars was performed on a custom tammar spotted cDNA microarray as previously described (Lefevre et al. 2007). This tammar array contains 14,837 ESTs clustered into 1929 genes from the mammary transcriptome (Lefevre et al. 2007). RNA from mammary tissue of two on-shore and one off-shore lactating Cape fur seal was extracted using an RNeasy Lipid Kit (Qiagen, Valencia, CA). This RNA was hybridized to canine Affymetrix GeneChips as previously described (Sharp et al. 2006). High sequence conservation between the Cape fur seal and dog, 95% similarity at the DNA level (Sharp et al. 2006), permits a significant detection rate of measurable hybridization signals between seal cDNA and the Affymetrix canine microarray (Affymetrix, Santa Clara, CA). Mammary tissue was obtained by biopsy (Sheehy et al. 2004) from three multiparous late-pregnant cows (20 days prior to expected calving date) and again 30 days after parturition (d30L). Daily milk production of these cows 1 week before biopsy was 30–35 L, and 24–48 h post biopsy the cows were producing 29–40 L/day. RNA was extracted from biopsied samples using Tri-Reagent (Sigma Aldrich, St. Louis, MO), purified by application to an RNeasy column (Qiagen), and hybridized to bovine Affymetrix GeneChips under contract to the Australian Genome and Research Facility.
Species-specific signal intensities from Affymetrix microarrays for the cow or the Cape fur seal were normalised using the robust multiarray average function in Bioconductor (http://www.bioconductor.org) (Gentleman et al. 2004; Irizarry et al. 2003). Tammar cDNA microarray signal intensities were single-channel normalized using functions of the limma package (Smyth and Speed 2003).
In order to retrieve expression data for orthologous genes in the three species, cDNA sequence (tammar) or Unigene sequences (canine and bovine) were mapped across species using reciprocal top blast hit (filtering e value < 10−8). The number of genes mapped across arrays is approximate due to redundancy on the array (number of probes and number of genes are different), nonsymmetric mapping between species, dependence upon threshold for mapping, sequence quality, and reliance on Unigene gene assemblies.
Genes with a minimum intensity above background and twofold change in expression in both cow and Cape fur seal were retrieved from the Affymetrix microarray data using the Unigene sequence mapping. When available, orthologous cDNAs from the tammar were identified and expression profiles were retrieved from a comprehensive tammar expression microarray database and visually inspected to assess increased gene expression during lactation. Fold changes of the six commonly regulated genes (seven gene probes) in the cow, tammar wallaby, and Cape fur seal are expressed as averages of the gene probe intensity in each species. Significant changes in FOLR1 gene expression in the cow and tammar wallaby were assessed by t test. FOLR1 gene expression data are presented as the mean ± SEM, except for off-shore lactating Cape fur seal where n = 1.
Microarray technology was used to transcript profile the mammary glands of four tammars in early lactation and four tammars in late lactation, the mammary glands of two on-shore and one off-shore lactating Cape fur seals, and the mammary glands of three cows during late-pregnancy and at peak lactation. Commonly regulated genes between the Cape fur seal and the cow were identified and compared to differentially regulated genes in the tammar. Finally, a gene list of differentially regulated genes common to all three species was generated.
This study used comparative functional genomics to exploit animal models with high milk protein production to identify candidate genes that specifically regulate protein synthesis in the mammary gland. Changes in global gene expression in mammary tissues at the transition from low to increased milk protein production in the tammar during late lactation, from off-shore and on-shore lactating Cape fur seal, and pregnant and lactating cows highlighted a substantial increase in expression of folate receptor α (FOLR1) gene in all three species. This indicated that folate metabolism, and particularly FOLR1 gene expression, may be an important regulatory point of milk protein synthesis in mammary epithelial cells (lactocytes). Although the analysis was limited due to redundancy in gene probes on the arrays, nonsymmetric mapping of genes between species, dependence upon gene threshold for mapping and sequence quality, and reliance on Unigene gene assemblies, the upregulation of FOLR1 was consistent with previous studies demonstrating a central role for folate in the one-carbon pool, generation of methionine, methylation reactions, and protection from genome damage (Bailey and Gregory 1999; Choi and Mason 2000). Circulating folate is present in the monoglutamate form which is required for active transport across cell membranes by the membrane-bound receptor (Birn 2006). The importance of the FOLR1 for cellular uptake of folate was established by the analysis of renal folate handling in mice with targeted gene knockouts of folate binding proteins 1 and 2 (folbp1 and folbp2, equivalent to human and cow FOLR1 and FOLR2) (Birn et al. 2005).
Although milk is a nutritional source of folate for the suckling young, it is unlikely that the increase in FOLR1 gene expression is related to the presence of folate in milk. Folate occurs in milk predominantly in the polyglutamate form and is bound to a folate-binding protein (Jones and Nixon 2002). Given that the FOLR1 protein has a high affinity for folate in the monoglutamate form (Birn 2006) and that milk folate is present in milk in a polyglutamate state, the folate-binding protein in milk is likely be different than the FOLR1 protein.
The data presented in the current study are consistent with an increase in FOLR1 gene expression upon stimulation of milk protein synthesis in hormone-induced mammary explants from pregnant cows (Menzies et al. 2009). The expression of the gene for FOLR1 increased with the induction of milk protein gene synthesis in response to insulin in cultured mammary explants (Menzies et al. 2009). Furthermore, genes encoding enzymes involved in the metabolism of folate to its active cellular form of tetrahydrofolates (Shane 1989) were upregulated in bovine cultured mammary explants when milk protein gene expression was induced (Menzies et al. 2009). Although the bifunctional methylenetetrahydrofolate dehydrogenase (MTHFD2) gene was upregulated in both lactating cow and on-shore lactating seal mammary tissues in the current study (data not shown), an assessment in tammar mammary tissue was not possible because the expressed sequence tags (ESTs) for this gene and other genes involved in folate metabolism were not present on the EST array for the tammar.
Folate supplementation experiments in multiparous, but not primiparous, dairy cows have shown a positive milk protein production response (Girard and Matte 1998, 2005; Graulet et al. 2007). Lactation increases the demands for both methylated compounds and methionine to support milk protein synthesis (Girard and Matte 2005; Xue and Snoswell 1985). Folate may potentially reduce the competition for precursors between gluconeogenesis and methylneogenesis and improve metabolic efficiency of the mammary gland or regulate gene expression by methylation reactions to influence milk protein production.
It remains to be established whether folate has a direct effect on the mammary gland to enhance milk protein synthesis. The current data suggest that the folate receptor population may play a crucial role in the capacity of lactocytes to respond and utilize circulating serum folates. Molecular studies in human, monkey, and mouse cell lines have shown that the regulation of FOLR1 protein population is complex and occurs at multiple levels. For example, FOLR1 protein population is influenced by extracellular folate concentrations and intercellular homocysteine concentrations (Antony et al. 2004; Kane et al. 1988). The FOLR1 may be recycled between the plasma membrane and an internal compartment, and regulation of FOLR1 protein synthesis is at the translation level as well as transcription of the gene (Antony et al. 2004; Birn 2006; Kamen and Smith 2004; Kamen et al. 1989; Sabharanjak and Mayor 2004). Functional studies exploiting bovine mammary culture models will be important to understand the regulation of the folate receptor in lactocytes and establish whether folate has a direct effect on the mammary gland. Understanding the molecular mechanisms of FOLR1 regulation at the level of the mammary gland more thoroughly may provide new opportunities for intervention to achieve more consistent milk protein responses to folate supplements in primiparous and multiparous cows (Girard and Matte 1998, 2005; Graulet et al. 2007).
This study demonstrated the potential of exploiting animals with extreme adaptation to lactation, combined with biotechnology platforms, to identify candidate regulatory genes of interest to the dairy industry. Previous studies in both cows and mice have examined global changes in gene expression at the transition from pregnancy to lactation both in vitro and using mammary culture models to identify important genes involved in lactation (Finucane et al. 2008; Menzies 2008; Menzies et al. 2009; Naylor et al. 2005; Rudolph et al. 2003; Stiening et al. 2008). However, both these experimental approaches include lactogenesis and are limited in their capacity to specifically study key regulatory mechanisms of milk protein synthesis. The pool of data extrapolated from such studies may pertain not only to synthesis of other milk constituents such as lipid and lactose, but also to the development of the extracellular matrix signalling and the induction of the cellular processes involved in milk synthesis, including the induction of milk protein synthesis. Recent comparative genomic studies utilised two lactating mouse models, one with a normal milk production phenotype (CBA mice) and one with an increased total milk production (QSi super-lactating mice) (Ramanathan et al. 2007). This experimental approach removed the interference from genes specifically involved in induction and maintenance of lactation and allowed identification of key genes for milk production, with potential use in the dairy industry (Ramanathan et al. 2007). In the current study, the use of comparative genomics was further exploited using species with extreme adaptation to lactation to focus on key genes regulating one aspect of milk production, milk protein synthesis.
Comparative genomics utilising the Cape fur seal and the tammar wallaby was undertaken in the current study because the lactation strategies of these mammals are characterised by periods of high milk protein production during an established lactation. Furthermore, de novo synthesis of lipids and lactose is minimal in the Cape fur seal, allowing the study to aggressively target key regulatory genes involved in milk protein synthesis. As genome coverage of other lactating species becomes better understood at the DNA level [e.g., the genome of the platypus, a monotreme, has recently been sequenced (Warren et al. 2008)], more opportunities will be created to study key aspects of lactation. This not only will be applicable to the dairy industry but will have broader relevance to other areas of the biological sciences such as human nutrition and development. In addition, the current study was limited to genes that had probes in common with the microarray Affymetrix GeneChips used to analyse the cow and fur seal mammary RNA and the EST array used to analyse the tammar wallaby mammary tissue. Therefore, it is likely that there are additional genes commonly expressed in the mammary glands of these species and not just the six observed in the current study. In the future, extensive sequencing technologies will potentially provide better transcriptome coverage in the tammar (and other models), and this will help overcome the limitation of redundancy in available gene probes between species and expand the current data. Despite the limitations of this study, the results were useful and this novel approach successfully highlighted a potential key role for FOLR1 in regulating milk protein synthesis. This illustrates the concept that comparative genomics combined with bioinformatics is a useful method to target genes of interest to the scientific community.
This work was supported by the CRC for Co-operative Research of Innovative Dairy Products, Dairy Australia, and the Geoffrey Gardiner Foundation. The assistance of Herman Oosthuizen, John Arnould, and colleagues in the collection of Cape fur seal tissue samples is gratefully acknowledged. We also thank the captain and crew of the MV Sardinops for logistical support without which this work could not have been conducted.