Multiple forms of equine α-lactalbumin: evidence for N-glycosylated and deamidated forms

https://doi.org/10.1016/j.idairyj.2003.08.001Get rights and content

Abstract

Equine raw milk whey proteins were separated by anion-exchange fast protein liquid chromatography and characterised by alkaline polyacrylamide gel electrophoresis (PAGE), sodium dodecyl sulphate PAGE, and bi-dimensional PAGE. Approximately, 1% of α-lactalbumin (α-LA) was N-glycosylated. A minor N-glycosylated form of lysozyme was also found in equine milk. On the other hand, two non-glycosylated α-LA isoforms with similar molecular masses (14,215±4 Da) were shown to be present. Their respective apparent isoelectric points were 5.25 and 4.94. These isoforms did not correspond to different genetic variants and were not the result of α-LA modulation by calcium ions. They corresponded rather to a non-enzymatic deamidation process of a single asparagine side-chain, the most acidic isoform being spontaneously generated from the less acidic isoform by simple incubation of α-LA at 37°C. The initial rate of this chemical degradation was 4.5 μm ammonia liberated per hour, in 150 mm sodium phosphate buffer pH 7.4 at 37°C. Deamidation induced a slight variation in secondary structure content, but no significant change in the tertiary structure of the equine α-LA was studied by circular dichroism in the near- and far-UV regions.

Introduction

The physical, chemical and functional properties of α-lactalbumin (α-LA) and lysozyme (LZ) have been extensively compared. The two proteins have striking structural homology in spite of marked differences in their functions (Acharya, Stuart, Phillips, McKenzie, & Teahan, 1994; Qasba & Kumar, 1997). α-LA is a part of the lactose synthetase complex that promotes the synthesis of lactose in the lactating mammary gland in modifying the substrate specificity of the UDP-galactosyltransferase (Kuhn, Carrick, & Wilde, 1980). Human α-LA has a strong Ca2+ binding site and a weaker secondary binding site (Chandra, Brew, & Acharya, 1998). Calcium binding strongly influences the molecular stability of α-LA and is required for refolding and disulphide bond formation in the reduced, denatured protein. α-LA can also bind different metal ions in addition to calcium (Permyakov & Berliner, 2000). LZ in equine milk has the essential features of a chicken-type (c-type) LZ (McKenzie & Shaw, 1985) that catalyses the hydrolysis of glycosidic bonds of mucopolysaccharides in bacterial cell walls. Equine milk LZ is a calcium binding protein (Lyster, 1992), although bovine and hen egg white LZs have no calcium. Therefore, on the basis of metal binding properties, the equine LZ is more similar to bovine α-LA than to bovine and hen LZs (Sugai, Ikeguchi, & Shimizu, 1999).

The refolding of bovine α-LA (e.g., Dolgikh et al., 1981) and equine LZ (Sugai et al., 1999; Morozova-Roche, Jones, Noppe, & Dobson, 1999) has been extensively studied in order to elucidate the mechanisms by which proteins fold. The two proteins show an intermediate, partially unfolded state termed “molten globule” state that does not bind calcium any more.

In the mare, three genetic variants of α-LA have been identified (Kaminogawa, McKenzie, & Shaw, 1984; Godovac-Zimmermann, Shaw, Conti, & McKenzie, 1987). Donkey's α-LA presents two isoforms A and B which show an isoelectric point (IEP) difference of 0.23 pH units (Giuffrida, Cantisani, Napolitano, Conti, & Godovac-Zimmermann, 1992). Their aminoacid sequences reveal, however, no differences. This apparent heterogeneity of α-LA would be due to binding of Ca2+ (Giuffrida et al., 1992).

Among the different species investigated, α-LA has been reported to be glycosylated either totally (Prasad, Hudson, Butkowsky, Hamilton, & Ebner, 1979; Cantisani, Napolitano, Giuffrida, & Conti, 1990), or partially (Brew, Castellino, Vanaman, & Hill, 1970; MacGillivray, Brew, & Barnes, 1979; Hopp & Woods, 1979; Halliday, Bell, McKenzie, & Shaw, 1990; Giuffrida et al., 1997). No glycosylation has been reported up to now for horse's and donkey's α-LAs. Moreover, no glycoform of LZ has been found, except the hen egg white LZ, for which only 0.3% of the total protein is glycosylated (Trudel & Asselin, 1995).

The aims of this study were (i) to evidence the existence of minor glycosylated forms of equine α-LA and (ii) to show that the existence of two major forms of equine α-LA was not the result of a modulation by Ca2+ as suggested in the case of the donkey's α-LAs, but was the consequence of a protein degradation process identified as the deamidation of one amidated side-chain. This degradation process was identified by in vitro study as non-enzymatic deamidation and its effects on the Ca2+ binding and on the secondary and tertiary structures of the equine α-LA were discussed.

Section snippets

Preparation of equine whey soluble proteins

Milk samples were collected from a herd of Haflinger mares and immediately stored at −20°C until used. The milk was skimmed by centrifugation (2100×g at 32°C for 30 min) and the whole casein was precipitated at pH 4.2 with 1 m HCl (the minimal solubility of the equine caseins was reached for pH 4.2 instead of pH 4.6 for bovine caseins). Milk whey was neutralised with 1 m NaOH, dialysed against distilled water in the presence of thymol at 4°C for 72 h, and the whey soluble proteins were lyophilised.

Anion-exchange FPLC

Chromatography of equine whey proteins

The whey soluble proteins were separated in four fractions by anion-exchange FPLC (Fig. 1). Each of the four fractions was checked for purity by anion-exchange FPLC and was submitted to amino-terminal microsequencing (Fig. 1). LZ was not retained on the Mono Q column and was entirely recovered in F1. Its amino-terminal region was identical to that reported for equine LZ (accession number P11376; McKenzie & Shaw, 1985). α-LA displayed two isoforms recovered in F2 and F3, respectively. The two

Acknowledgments

We thank Dr. G. Humbert and F. Saulnier (SCSP, UHP-Nancy 1, France) for the peptide sequencing, Dr. D. Mollé (Laboratoire de Recherches de Technologie Laitière, INRA, Rennes, France) for the mass spectrometry, Dr. L. Miclo (Laboratoire des BioSciences de l’Aliment, UHP-Nancy 1, France) and Dr. C. Barbey (Laboratoire de Chimie Structurale, Université Paris 13, France) for good advices and Fanny Bassora for technical assistance.

References (43)

  • R.T. MacGillivray et al.

    The amino acid sequence of goat α-lactalbumin

    Archives of Biochemistry and Biophysics

    (1979)
  • L.A. Morozova-Roche et al.

    Independent nucleation and heterogeneous assembly of structure during folding of equine lysozyme

    Journal of Molecular Biology

    (1999)
  • D. Pâquet et al.

    Study of a hydrophobic protein fraction isolated from milk proteose peptone

    Journal of Dairy Science

    (1988)
  • E.A. Permyakov et al.

    α-LactalbuminStructure and function

    FEBS Letters

    (2000)
  • R. Prasad et al.

    Resolution of the charge forms and amino acid sequence and location of a tryptic glycopeptide in rat α-lactalbumin

    The Journal of Biological Chemistry

    (1979)
  • M.P. Thompson et al.

    The calcium dependent electrophoretic shift of α-lactalbumin, the modifier protein of galactosyl transferase

    Biochemical and Biophysical Research Communications

    (1988)
  • M. Xie et al.

    Secondary structure and protein deamidation

    Journal of Pharmaceutical Sciences

    (1999)
  • K.R. Acharya et al.

    Models of the three dimensional structures of echidna, horse, and pigeon lysozymesCalcium-binding lysozymes and their relationship with α-lactalbumins

    Journal of Protein Chemistry

    (1994)
  • B.P. Alston-Mills et al.

    A theoritical approach to possible biological functions of the milk-whey proteins, α-lactalbumin and β-lactoglobulin

    Comments in Agricultural and Food Chemistry

    (1996)
  • M.A. Andrade et al.

    Evaluation of secondary structure of proteins from UV circular dichroism spectra using an unsupervised learning neutral network

    Protein Engineering

    (1993)
  • D.W. Aswald et al.

    Isoaspartate in peptides and proteinsFormation, significance, and analysis

    Journal of Pharmaceutical and Biomedical Analysis

    (2000)
  • Cited by (24)

    • Antibacterial potential of donkey's milk disclosed by untargeted proteomics

      2021, Journal of Proteomics
      Citation Excerpt :

      Nearly 10–15% of α-La in the milk of ruminants is N-glycosylated at the level of canonical consensus triplets N-X-S/T (where X is any amino acid except P). The primary structure of donkey’s α-La lacks consensus sequences, but, like the human homologue [47], it could be 1% N-glycosylated at the N71 belonging to the N-X-C triplet [48]. However, in absolute terms glycosylated α-La could contribute to the overall compute of milk glycoproteins to a non-negligible extent.

    • Effect of nonenzymatic deamidation on the structure stability of Camelus dromedarius α-lactalbumin

      2019, Food Chemistry
      Citation Excerpt :

      The secondary and tridimensional structures of A1, A2, and A3 were investigated by CD in the far- and near-UV regions (Fig. 2). The near-UV CD spectra had the typical features of the spectrum of holo-α-La with a minimum in the region 270–275 nm (corresponding to Tyr) and a maximum at 292 nm for Trp (Girardet et al., 2004; N'Negue et al., 2006; Saricay, Wierenga, & de Vries, 2014). According to former studies, the spectrum of apo-α-La is different but nevertheless retains the characteristic shape of the spectrum of the holo-α-La with much lower intensity (Saricay et al., 2014), or is featureless (N'Negue et al., 2006).

    • Analytical methods for measuring or detecting whey proteins

      2018, Whey Proteins: From Milk to Medicine
    View all citing articles on Scopus
    1

    Present address: EMBRAPA-CAPRINOS, Cx. Postal D10, CEP 62011-970, Sobral, Ceara, Brazil.

    View full text