Shift in the isoelectric-point of milk proteins as a consequence of adaptive divergence between the milks of mammalian species
© Khaldi and Shields; licensee BioMed Central Ltd. 2011
Received: 10 September 2010
Accepted: 29 July 2011
Published: 29 July 2011
Milk proteins are required to proceed through a variety of conditions of radically varying pH, which are not identical across mammalian digestive systems. We wished to investigate if the shifts in these requirements have resulted in marked changes in the isoelectric point and charge of milk proteins during evolution.
We investigated nine major milk proteins in 13 mammals. In comparison with a group of orthologous non-milk proteins, we found that 3 proteins κ-casein, lactadherin, and muc1 have undergone the highest change in isoelectric point during evolution. The pattern of non-synonymous substitutions indicate that selection has played a role in the isoelectric point shift, since residues that show significant evidence of positive selection are much more likely to be charged (p = 0.03 for κ-casein; p < 10-8 for muc1). However, this selection does not appear to be solely due to adaptation to the diversity of mammalian digestive systems, since striking changes are seen among species that resemble each other in terms of their digestion.
The changes in charge are most likely due to changes of other protein functions, rather than an adaptation to the different mammalian digestive systems. These functions may include differences in bioactive peptide releases in the gut between different mammals, which are known to be a major contributing factor in the functional and nutritional value of mammalian milk. This raises the question of whether bovine milk is optimal in terms of particular protein functions, for human nutrition and possibly disease resistance.
This article was reviewed by Fyodor Kondrashov, David Liberles (nominated by David Ardell), and Christophe Lefevre (nominated by Mark Ragan).
The isoelectric point (pI) and charge of a protein is important for solubility, subcellular localization, and interaction. There is a correlation between subcellular location and protein pI [1, 2]. Proteins in the cytoplasm possess an acidic pI (pI < 7.4), while those in the nucleus have a more neutral pI (7.4 < pI < 8.1) [1, 2]. It has also been shown that the pI can vary greatly, depending on both insertion and deletions between orthologs, and the ecology of the organism . Kirga et al  have shown that the pI of membrane proteins of bacteria correlates with their ecological niche, and changes dramatically from acidic to basic. For example, some prokaryotes that infect human have a pI that reflects their localization in the human body, compensating for the pH change. E. coli that resides in the intestines has more acidic proteins, and H. pylori that infects the acidic stomach has more negatively charged proteins .
For highly abundant proteins, shifts in their pI can impact on the function of organs that interact with them. Purtell et al  examined the effects of change in isoelectric point (pI) on renal handling of albumin molecules. The authors showed that the increase of the pI caused an increase in heterologous albumin secretion and increased nephron permeability.
Milk proteins travel through the various mammalian digestive systems with their different compartments and pH levels. For example, carnivorous species possess very acidic stomachs compared to herbivores, and orthologous milk proteins need to travel and perform their function in all these systems. Because of these differences we might expect to observe adaptation of the milk proteins in order to perform orthologous functions, or an adaptation of the pI to a new acquired functionality. Large differences in the pI of milk proteins might have important consequences on the structure, properties, functionality and interaction of these proteins.
Function of milk proteins
α-S1-casein, β-caseinm and κ-casein
~80% of bovine and 20-45% of human milk protein. Phosphoprotein carriers of minerals and trace elements.
Calcium and other carrier, lactose synthesis 
Also known as Milk Fat globule factor 8 (Mfge8); bactericidal and apoptotic properties .
Milk fat globule; digestion resistant
Modulates bacterial adhesion 
Milk fat globule; digestion resistant
Milk fat globule
~40% of protein in Milk Fat Globule Membrane; fat globule secretion 
Milk fat globule; rapidly degraded
We show that the shifts do not simply reflect differences in sequence lengths between the milk orthologous proteins, and are likely driven by selection. Both sequence length and selection have been recently shown to explain the observed differences in pI between mammalian orthologs . We argue that the differences in the digestive systems due to pH and compartmentalization of the different mammals is not the sole driver of major changes in pI, and that these selective changes might be due to functional divergence of the protein.
Results and discussion
Calculation of pI
It has been shown that platypus contains two extra copies of κ-casein . These two copies have very different pI values ranging from acidic to basic, with pI = 5.9 for FJ548612, to pI = 8.8 for FJ548626 (Figure 1). Contrary to the other observed shifts in pI represented in Figure 1, the great shift in pI between the κ-casein copies cannot be explained by interspecies differences. It is noticeable that the pI of the current κ-casein orthologs is much higher than that of the ancestor values (Figure 2). However mouse and guinea pig seem to be an exception to this observation. It is unclear how much this is due to real pI shifts or to and artifact of the method of pI calculation.
Lactadherin has shifted at least twice on the tree. It is basic in the two outgroup species opossum (pI = 7.83) and platypus (pI = 7.75), in primates (pI = 7.96 human, pI = 8.17 chimp, pI = 7.42 macaque), and in guinea pig (pI = 7.76), but seems to have shifted independently twice to acidic/neutral, once in rodents (pI = 6.36 in mouse, and pI = 6.9 in rat), and another time in dog (pI = 6.45 in dog).
The pattern of pI change of muc1 protein shows a number of potential changes, shifting in two independent lineages to a lower pI in both rodents (mouse pI = 5.45, and rat pI = 5.09) and horse (pI = 5.83).
Are the shifts in the pIof some milk proteins important compared to whole proteome comparison?
What appear as dramatic changes between the pIs of κ-casein, lactadherin, and muc1 orthologs, might not seem so dramatic compared to the changes across the entire proteome for non-milk protein orthologs.
To investigate this, we considered all the orthologous proteins in the 13 mammals (human, chimp, monkey, mouse, rat, guinea pig, rabbit, cow, horse, dog, cat, opossum, platypus). We considered a shift in pI between human and mouse to be high if it was greater than 0.92, and an identical value between human and cow (Additional File 1; see Methods for the rationale behind the choice of these cut-offs). We further tested for the significance of this threshold by randomly assigning pI values to proteins, and found that our set thresholds are in all cases significant (p < = 0.01).
Differences in length between orthologs due to insertions or deletions are associated with the pI shift in certain proteins
Sequence lengths for eight milk proteins in human, mouse, and cow
butyrophilin subfamily 1 member A1
These results show that for Lactadherin the change in pI is mainly due to the mouse insertion. However this scenario does not account for the change in pI for muc1 and κ-casein where both shifts are accounted for by amino acid replacements between human and mouse.
Selection causing pI change
Given that so many residues are experiencing adaptation in the human κ-casein and have a direct impact on the pI argues for adaptive changes in the pI of κ-casein. To further examine if positive selection has played a role in the evolution of κ-casein, we calculate the ratio of the rate of non-synonymous over synonymous substitutions (dN/dS). Figure 2 shows that the mouse κ-casein has undergone the greatest ratio indicating the action of positive selection on this protein in the mouse lineage. This also happens to correspond to the lineage undergoing the highest shift in pI (Figure 1, Figure 2). This positive selection seems to have consequently shifted the pI of mouse κ-casein. Two other orthologs seem to have also undergone some sort of fast evolutionary divergence (Figure 2 shows that horse and cow have dN/dS > 1) even though the dN/dS value might be too week to speak about positive selection, the cow ortholog happens to have also diverged in its pI (Figure 1 and Figure 2, horse however seems to have diverged in sequence but retained a closer basic pI to the other mammals studied in this work.).
For muc1, we detected 25 sites under positive selection (p < = 4.7.10-2), 15 of these have changed the overall pI of the protein and also changed its net charge. Here again, we find that there are significantly more sites undergoing positive selection and that these have an impact on the net charge of the protein compared to all other neutral sites (p = 2.4 × 10-9; 28% for charged residues versus 1.9% for neutral residues). Under a random distribution of the positively selected sites detected in the human muc1 protein sequence, we will expect an average of 4.4% sites that undergo positive selection whether these are charged or neutral, which is less than the observed 28% charged sites that have undergone positive selection.
Put together these results show that selection has played a part in the change of pI and consequently on the overall net charge of the protein.
Selection pressures for changes in pI: the roles of dietary, morphological, and intrinsic milk protein factors
What is driving this selection on the pI? Can it be the important differences in pH and compartmentalization between the digestive systems of different mammals?  Milk proteins travel down the digestive system. Some, such as the caseins, get broken down in the highly acidic conditions of the stomach, whereas others such as lactadherin and lactoferrin [9, 10] travel intact or partially intact to be broken down further down in the digestive tract. Given the very large shifts in pI, we would anticipate that the processing and breakdown of milk proteins are likely to differ substantially. Thus, if we were to replace the human κ-casein with that of mouse, it seems unlikely that they will interact with their environment and function in an identical way, given that the mouse and human κ-casein pI is 4.75 in mouse, but 8.59 in human.
We might imagine that the greatest shifts during evolution might occur when animals shift between largely carnivorous or omnivore diets and herbivore diets, since the more complex stomachs of some herbivores, and the more acid stomach pHs of some carnivores might alter functional constraints. However, inspection of Figure 1 indicates that many large shifts occur between species that have largely similar overall dietary strategies (dog and cat; mouse and rat). This suggests that the shifts in functional constraints may be associated with factors that are not linked with the gross morphology or diet of major clades. Similarly, the values of the posterior stomach pH in the different mammals represented in Figure 2 do not clearly argue for a stomach-pH change that is driving the shift in pI for κ-casein, including the significant pI shift observed in mouse (Fig1, Fig2, and Fig3). Besides, the great difference observed between the pI values of the two extra copies of κ-casein in platypus (Figure 1; pI = 5.9 for FJ548612, to pI = 8.8 for FJ548626) does not argue for a stomach pH driven selection on milk proteins' pI.
It is interesting to speculate on how extrinsic factors, such as commensal and pathogenic bacteria, may exert selection pressures on milk protein function, but also of interest to consider how alterations in intrinsic milk protein functions may relate to adaptive changes. Milk proteins are known to yield many bioactive peptides that modulate and participate in various regulatory processes in the body . These peptides are usually cleaved by digestive enzymes such as trypsin, pepsin, and chymotrypsin. Some proteases cleave near positively charged residues, such as trpysin, while others avoid positive charge in their substrate region (pepsin), and the adaptive requirements for the gain and loss of proteolytic cleavage sites in certain regions of the gut (e.g. the duodenum versus the stomach) may have some an impact on pI. In particular, when we consider the casoxins , known bioactive peptides released from bovine κ-casein that have opioid antagonist and anti-opioid activities- we note that although casoxin A, and C are released in cow, this is not the case in human and mouse, since the cleavage sites are not the same between the species (Figure 4). It is interesting to note that 3 residues of the 14 residues that we found to be positively selected on in κ-casein are found on the borders of the three peptides casoxin A, B, and C (Figure 4), indicating possible selection on the cleavage sites. Also, Figure 4 shows that 3 other positively selected sites are located within the peptides casoxin A, and B sequence, indicating adaptation of the individual peptides at least to cow. Thus, the shift in pI may be associated with divergence in functional requirements for either rates of digestion, or for functional components of the milk.
Phosphorylation and glycosylation
We observe that all the proteins that have shifted dramatically are ones that also happen to be highly glycosylated and phosphorylated. Indeed the three proteins κ-casein, muc1, and lactadherin have more glycosylation sites than the other milk proteins with an average of 7 glycosylations in human (9 glycosylations in muc1, 7 in κ-casein, and 5 in lactadherin; these include referenced, probable, and potential sites), and in cow, as opposed to an average of 1.3 in the remaining 6 milk proteins in human, and in cow. Besides, we do also observed differences in phosphorylation sites, for example we have 3 referenced phosphorylations in cow κ-casein and none in human and mouse. Also, there are 9 referenced phosphorylations in human muc1, while there are 6 and 7 by similarity in cow and mouse respectively.
Our analyses of pI did not take into account these post-translational modifications. To examine if post-translational modifications can reduce the difference in the isoelectric point, we used experimentally validated phosphorylation and glycosylation sites, which are defined in cow, human and to a weaker extent in mouse. For κ-casein (Figure 4), the cow pI shifts from 5.93 to 5.34 when the two experimentally verified phosphorylations are added. Human remains the same pI = 8.68 (no experimentally validated phosphorylation so far), and mouse shifts from 4.67 to 4.52 (1 potential phosphorylation site; Figure 4.). The phosphorylation sites for muc1 in both cow and mouse are potential sites found with similarity rather than experimentally validated sites. These results show that despite shifting the pI of κ-casein and muc1 towards a more acidic pH as a result of phosphorylation in the three different species, the difference in pI remains very important between these two proteins.
The differences in glycosylation between human and cow for κ-casein might somewhat further reduce the pI shift between both these species. Indeed, in κ-casein (Figure 4) we have 7 glycosylations in human as opposed to 6 in cow (none have been experimentally validated so far in mouse). For muc1, experimental validation is only available for human that has 4 O-linked, and 5 N-linked glycosylations. These might also narrow down the gap in the muc1 pI between the different species. Nonetheless, both cases where the pI difference is reduced or not are interesting. Indeed if the pI difference is reduced and becomes very close between both species, this reflects that the protein has adapted its pI so that the final product with the different number of glycosylations and phosphorylations becomes the same. Indeed, if the pI was initially not different, the addition of glycosylation will then further the gap between the pIs.
Although the production of milk is conserved between mammals for over 190 MA, our results argue that common proteins that have been shared by mammals are functionally diverging. Many humans consume cow's milk on a daily basis, and yet the pI of κ-casein in cow is very different from our κ-casein. We have shown that selection has acted on the residues that affect the protein's pI. The simplest explanation was the adaptation of the protein to the different digestive systems to accommodate reactions to changes in pH of the different compartments. However, we found the pattern of change did not correlate strongly with the greatest shifts in compartmentalization and pH during evolution, suggesting that other factors, potentially including milk proteins' functional features, may be associated with the adaptive changes.
Differences in the function of κ-casein between various species, raises the question of whether κ-casein of cow can functionally replace that of human. κ-casein is known to yield many bioactive peptides [12, 13] which, as we have discussed, might have different affinities and functionalities between human and cow. Such functional changes may relate to regional positive selection seen within κ-casein in the family bovidae .
It is of interest to note that two of the proteins showing the most striking shifts in pI are also glycosylated extensively (κ-casein and muc1). It is not clear if this is merely coincidental, or whether glycosylated proteins play a particular role in the gut that is subjected to shifting selection pressures over evolutionary time. An obvious candidate function would be bacterial interactions, which are heavily influenced by glycosylated proteins, and κ-casein is known to play a role in altering Helicobacter pylori adhesion  (review ). Exactly how shifting the pI of these milk proteins might benefit the neonate is not entirely clear. However, given the ability of pathogens such as H. pylori to modify the host stomach pH , the ability of milk proteins to coat particular compartments or infected regions of altered pH is an obvious candidate factor to investigate. In this context, a specific question raised by our study is whether the muc1 and κ-casein in cow's milk provide optimal protection against bacterial infections of the stomach and intestine for human neonates.
The human, chimp, monkey macaque, mouse, rat, guinea pig, rabbit, cat, dog, horse, cow, opossum, and platypus protein sequences were downloaded by FTP from the ENSEMBL database at: ftp://ftp.ensembl.org/pub/release-63/fasta/
Out of seventeen identified major milk proteins  we picked a subset for analysis on the basis of their belonging to at least 8 mammalian species out of the 13 (Table 2). In addition the 8 species needed to include human, chimp, cow, and mouse. These proteins represent the three parts of milk (Table 1): whey, casein, and milk fat globule. We used the 9 major milk proteins defined in human and cow to detect their orthologs in the 13 other genomes, defined by reciprocal hits.
Orthologs and sequence evolution
To find orthologous non-milk proteins, we identified 13-way mutual best BLASTP hits among human, chimp, monkey macaque, mouse, rat, guinea pig, rabbit, cat, dog, horse, cow, opossum, and platypus. This method resulted in 1412 sets of putative orthologs that were present among all 13 species. Each set of 13 proteins was aligned using ClustalW .
Calculating the isoelectric point
We first cleaved off the signal peptide from each protein using a HMM search with SignalP-HMM . The rest of the sequence was incorporated into an in-house perl script for the calculation of the pI that uses the Henderson-Hasselbach equation. The script searched for the number of R, K, Y, C, H, E, and D that are implicated in the pI of a protein. Each of the previous amino acids was assigned a pKa value, 12.48, 10.54, 10.46, 8.18, 6.04, 4.07, and 3.9 respectively, 8.0 for the N-terminus, and 3.1 for the C-terminus. The charge due to arginine for example is the product of the corresponding pKa with the number of instances or R in the sequence. We can then calculate an estimated charge for the protein at any particular pH. To determine the pI that is the pH value at which the estimated charge is zero, we estimated an initial pH at which the overall charge of the protein is positive and one where the charge is negative. We then used a bisection method to estimate to a 10-2 precision the value that renders the overall charge null.
Defining significant pI shifting proteins
A protein is considered as significantly shifting in, for example mouse, if the distance between its pI and that of its ortholog in human is higher than a threshold that is determined from the differences in pI of all orthologs between human and mouse (Additional File 1). Setting a threshold of pI between two species is somewhat arbitrary because the data does not follow a known distribution, for this reason we used a non-parametric formula to define the threshold of significance. This threshold is calculated using the median, and third quartile of the absolute shift in pI between orthologous proteins this is: threshold = 2 × (3rd quartile - median).
Ancestral Reconstruction and amino acid substitution rate
To reconstruct the ancestral sequences of the current κ-casein protein, we aligned the κ-casein orthologs in the 12 species represented in Figure 2 using T-coffee ; this step was followed by a maximum likelihood reconstruction using codeml from the paml package .
To calculate the amino acid substitution we gathered the DNA coding sequences of κ-casein proteins from the ENSEMBL database. We could not locate good quality sequences for guinea pig, cat, and dog. We aligned the other 9 κ-casein protein orthologs using T-coffee . The DNA sequences were aligned based on the protein alignment. We implemented codeml  on the DNA alignment to calculate the synonymous dS and non-synonymous dN substitutions.
Detecting selection in the charged residues
To examine if the significant variation between human, chimp, mouse, and cow, in amino acid composition is due to selection, we gathered the DNA coding sequences of all milk-specific proteins from the ENSEMBL database. We aligned the proteins using T-coffee  and implemented a script that aligns to DNA based on the protein's alignment. We removed poorly aligned positions and divergent regions of a DNA alignment using Gblocks . We used the SLR method with the default parameters to detect positions that are likely to be under positive selection . These positions are indicated on Figure 4.
Sitewise Likehood Ratio.
This work was funded by the Irish Research Council for Science, Engineering & Technology (IRCSET), Marie Curie Actions under FP7, and Food for Health Ireland (FHI).
Referee 1, Fyodor Kondrashov
This is a very interesting study of proteins present in milk. My only suggestion is to provide more background in the introduction section on the function of the proteins that were studied to give a reader that may not be well oriented in protein function a better understanding of the implications of their evolution.
We have added a table (Table 1) summarizing the functions of the milk proteins used in this study. We refer to this table in different parts of the text including the introduction.
Referee 2, David Liberles
"Shift in the isoelectric-point of milk proteins as a consequence of adaptive divergence between the milks of mammalian species" by Khaldi and Shields is an interesting paper examining functional shifts in mammalian milk proteins. An additional table and additional analyses are suggested to maximize the ability to interpret the data presented in the study.
1) First, a table with more detail on the known functions of milk proteins would be desirable and help the reader assess other adaptive processes beyond adaptation to digestive system pH.
We have added a table (Table 1) summarizing the functions of the milk proteins used in this study. We refer to this table in different parts of the text including the introduction.
2) Additionally, a species tree showing known mammalian digestive system pH values at the tips coupled to reconstruction as continuous data over the tree would be informative. This couples to two additional pieces of information from the sequences. The first is a free ratios model of dN/dS (where supported) over each gene tree. The other is ancestral sequence reconstruction of sequences at nodes in trees and calculation of pI values for those ancestral states. These pieces of information can be used to ask if positive selection correlates with lineages where pI is changing and if this correlates with changes in digestive system pH values. Positive selection that is not explained by pI and pH changes would be strong candidates for alternative sources of adaptation.
With this, the authors will have performed a nice study of the evolution of milk proteins in mammals.
We have added a figure (Figure 2) representing the ancestral reconstruction of κ-casein and the summary of the pI values and the dN/dS ratios (when possible) for each of the studied species (Figure 2). Figure 2 also shows most of the pH interval values of the anterior and posterior stomachs for the different mammals of this study. The results of Figure 2 are discussed in the results and discussion section.
This reviewer provided no comments for publication.
- Andrade MA, O'Donoghue SI, Rost B: Adaptation of protein surfaces to subcellular location. J Mol Biol. 1998, 276 (2): 517-25. 10.1006/jmbi.1997.1498.PubMedView ArticleGoogle Scholar
- Nandi S, et al: Comparison of theoretical proteomes: identification of COGs with conserved and variable pI within the multimodal pI distribution. BMC Genomics. 2005, 6: 116-10.1186/1471-2164-6-116.PubMedPubMed CentralView ArticleGoogle Scholar
- Kiraga J, et al: The relationships between the isoelectric point and: length of proteins, taxonomy and ecology of organisms. BMC Genomics. 2007, 8: 163-10.1186/1471-2164-8-163.PubMedPubMed CentralView ArticleGoogle Scholar
- Purtell JN, et al: Isoelectric point of albumin: effect on renal handling of albumin. Kidney Int. 1979, 16 (3): 366-76. 10.1038/ki.1979.139.PubMedView ArticleGoogle Scholar
- Alendé N, et al: Evolution of the isoelectric point of mammalian proteins as a consequence of indels and adaptive evolution. Proteins. 2011, 79 (5): 1635-48. 10.1002/prot.22990.PubMedView ArticleGoogle Scholar
- Kraft C, Reggiori F, Peter M: Selective types of autophagy in yeast. Biochim Biophys Acta. 2009, 1793 (9): 1404-12. 10.1016/j.bbamcr.2009.02.006.PubMedView ArticleGoogle Scholar
- Massingham T, Goldman N: Detecting amino acid sites under positive selection and purifying selection. Genetics. 2005, 169 (3): 1753-62.PubMedPubMed CentralView ArticleGoogle Scholar
- Martinez M, et al: Applying the biopharmaceutics classification system to veterinary pharmaceutical products. Part II. Physiological considerations. Adv Drug Deliv Rev. 2002, 54 (6): 825-50. 10.1016/S0169-409X(02)00071-6.PubMedView ArticleGoogle Scholar
- Troost FJ, et al: Gastric digestion of bovine lactoferrin in vivo in adults. J Nutr. 2001, 2101-2104. 131
- Peterson JA, et al: Human milk fat globule (HMFG) glycoproteins: their association with lipid micelles in skim milk and survival in the stomach of milk-fed and pre-term infants. Pediatric Research. 1997, 41 (4, Part 2): 87-View ArticleGoogle Scholar
- Meisel H: Overview on Milk Protein-derived Peptides. Int Dairy J. 1998, 8 (5-6): 363-373. 10.1016/S0958-6946(98)00059-4.View ArticleGoogle Scholar
- Chiba H, Tani F, Yoshikawa M: Opioid antagonist peptides derived from kappa-casein. J Dairy Res. 1989, 56 (3): 363-6. 10.1017/S0022029900028818.PubMedView ArticleGoogle Scholar
- Murray BA, FitzGerald RJ: Angiotensin converting enzyme inhibitory peptides derived from food proteins: biochemistry, bioactivity and production. Curr Pharm Des. 2007, 13 (8): 773-91. 10.2174/138161207780363068.PubMedView ArticleGoogle Scholar
- Ward TJ, Honeycutt RL, Derr JN: Nucleotide sequence evolution at the kappa-casein locus: evidence for positive selection within the family Bovidae. Genetics. 1997, 147 (4): 1863-72.PubMedPubMed CentralGoogle Scholar
- Stromqvist M, et al: Human milk kappa-casein and inhibition of Helicobacter pylori adhesion to human gastric mucosa. J Pediatr Gastroenterol Nutr. 1995, 21 (3): 288-96. 10.1097/00005176-199510000-00006.PubMedView ArticleGoogle Scholar
- Lonnerdal B: Nutritional and physiologic significance of human milk proteins. Am Clin J Nutr. 2003, 77 (6): 1537S-1543S.Google Scholar
- Argent RH, et al: Toxigenic Helicobacter pylori infection precedes gastric hypochlorhydria in cancer relatives, and H pylori virulence evolves in these families. Clin Cancer Res. 2008, 14 (7): 2227-35. 10.1158/1078-0432.CCR-07-2022.PubMedView ArticleGoogle Scholar
- Rudolph MC, McManaman JL, Hunter L, Phang T, Neville MC: Functional development of the mammary gland: use of expression profiling and trajectory clustering to reveal changes in gene expression during pregnancy, lactation, and involution. J Mammary Gland Biol Neoplasia. 2003 Apr, 8 (3): 287-307.PubMedView ArticleGoogle Scholar
- Thompson JD, Higgins DG, Gibson TJ: CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 1994, 22 (22): 4673-80. 10.1093/nar/22.22.4673.PubMedPubMed CentralView ArticleGoogle Scholar
- Nielsen H, Krogh A: Prediction of signal peptides and signal anchors by a hidden Markov model. Proc Int Conf Intell Syst Mol Biol. 1998, 6: 122-30.PubMedGoogle Scholar
- Notredame C, Higgins DG, Heringa J: T-Coffee: A novel method for fast and accurate multiple sequence alignment. J Mol Biol. 2000, 302 (1): 205-17. 10.1006/jmbi.2000.4042.PubMedView ArticleGoogle Scholar
- Yang Z: PAML 4: phylogenetic analysis by maximum likelihood. Mol Biol Evol. 2007, 24 (8): 1586-91. 10.1093/molbev/msm088.PubMedView ArticleGoogle Scholar
- Castresana J: Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Mol Biol Evol. 2000, 17 (4): 540-52.PubMedView ArticleGoogle Scholar
- Benton MJ, Donoghue PC: Paleontological evidence to date the tree of life. Mol Biol Evol. 2007, 24 (1): 26-53.PubMedView ArticleGoogle Scholar
- DeSesso JM, Williams AL: Annual Reports in Medicinal Chemistry. Contrasting the Gastrointestinal Tracts of Mammals. Edited by: JE. Macor. 2008, Wallingford CT, United States, 43: 353-371.Google Scholar
- Jollès P, et al: Analogy between fibrinogen and casein. Effect of an undecapeptide isolated from kappa-casein on platelet function. Eur Biochem J. 1986, 158 (2): 379-82. 10.1111/j.1432-1033.1986.tb09764.x.View ArticleGoogle Scholar
- Permyakov EA, Berliner LJ: alpha-Lactalbumin: structure and function. FEBS Lett. 2000, 473 (2): 269-74.PubMedView ArticleGoogle Scholar
- Baker EN, Baker HM: Molecular structure, binding properties and dynamics of lactoferrin. Cell Mol Life Sci. 2005, 62 (22): 2531-9. 10.1007/s00018-005-5368-9.PubMedView ArticleGoogle Scholar
- Harmsen MC, et al: Antiviral effects of plasma and milk proteins: lactoferrin shows potent activity against both human immunodeficiency virus and human cytomegalovirus replication in vitro. J Infect Dis. 1995, 172 (2): 380-8. 10.1093/infdis/172.2.380.PubMedView ArticleGoogle Scholar
- Peterson JA, et al: Structural and functional aspects of three major glycoproteins of the human milk fat globule membrane. Adv Exp Med Biol. 2001, 501: 179-87.PubMedView ArticleGoogle Scholar
- Vorbach C, Scriven A, Capecchi MR: The housekeeping gene xanthine oxidoreductase is necessary for milk fat droplet enveloping and secretion: gene sharing in the lactating mammary gland. Genes Dev. 2002, 16 (24): 3223-35. 10.1101/gad.1032702.PubMedPubMed CentralView ArticleGoogle Scholar
- Silanikove N, et al: Role of xanthine oxidase, lactoperoxidase, and NO in the innate immune system of mammary secretion during active involution in dairy cows: manipulation with casein hydrolyzates. Free Radic Biol Med. 2005, 38 (9): 1139-51. 10.1016/j.freeradbiomed.2004.12.011.PubMedView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.