Solving the mystery of the internal structure of casein micelles

Abstract

The interpretation of milk X-ray and neutron scattering data in relation to the internal structure of the casein micelle is an ongoing debate. We performed resonant X-ray scattering measurements on liquid milk and conclusively identified key scattering features, namely those corresponding to the size of and the distance between colloidal calcium phosphate particles. An X-ray scattering feature commonly assigned to the particle size is instead due to protein inhomogeneities.

---

Despite the widespread consumption of milk and dairy products as a calcium-rich food source, controversy still remains over the internal structure of the casein micelle.^1,2 The casein micelle is a network of α- and β-casein protein molecules, with κ-casein on the surface. These caseins interact through hydrophobic bonding. Calcium is present within casein micelles in the form of colloidal calcium phosphate (CCP) nanoclusters around 2 nm in size.^3,4 Recent electron microscopy studies show that the internal micelle structure is quite uniform,^4,5 as opposed to the earlier hypothesis that the casein micelle was composed of sub-micelles^6–8 – an idea that is inconsistent with the chemistry as it is now understood. Earlier electron microscopy studies may have been prone to artefacts.⁹ In recent times various models have been proposed, including the nanocluster,^2,10,11 dual-binding^12,13 and interlocking lattice models.⁵ These models all recognise CCP as being a key structural component, binding to phosphoserine groups on the α- and β-casein proteins.

Small-angle scattering techniques are generally non-destructive¹⁴ and can be used to obtain structural information on the 1–1000 nm length scale of casein micelles in solution, thus being less prone to artefacts than electron microscopy. For small-angle X-ray scattering (SAXS), intensity arises due to X-rays scattering from areas of electron density contrast. In the case of small-angle neutron scattering (SANS), neutrons scatter in a similar way but according to nuclear cross-sections.

A number of features are observed in SAXS and SANS data for native micelles¹⁵ at various values of the scattering vector, q = (4π/λ)sin θ, where λ is the wavelength and θ is half of the scattering angle. A feature centred at q ∼ 0.005 Å⁻¹, observed in both SAXS and SANS, arises from the casein micelle itself, roughly 100 nm in diameter.^4,15–22 In this case the contrast is between the micelle and the liquid serum phase. In many SAXS experiments a feature centred at q ∼ 0.01 Å⁻¹ is observed, which is more prominent in concentrated or dried systems and has been attributed to separate ‘mini-micelles’,^17,21 a second oscillation of a core–shell form factor¹² or other sub-micellar structures.¹⁸ In SANS experiments and SAXS of dried powders, a feature centred at q ∼ 0.035 Å⁻¹ is observed,^19,23,24 but is usually absent in SAXS of micelles in solution.¹⁹ There is lingering controversy over the identity of this feature. It has been interpreted as the submicelle size¹⁶ or interaction distance,^24,25 or the CCP spacing.^3,19,23 Finally, a feature centred at q ∼ 0.08–0.1 Å⁻¹ is generally attributed to the CCP form factor.^{4,11,18,19,21,22} although a recent report considering the electron density contrast and known CCP concentration within the micelle, suggested that the scattering may instead arise from protein inhomogeneities – local fluctuations in density due to hydrophobic interactions of the caseins – on the same length scale (1–3 nm).³ In a recent study where individual casein micelles were trapped inside pores of a filtration membrane, the q ∼ 0.08 Å⁻¹ feature was absent.²⁶ Its disappearance was attributed to internal structural changes of the casein proteins as a result of the applied forces.

Several different approaches have been taken in the literature to fitting small-angle scattering data from casein micelles. The micelle itself is usually modelled as a smooth sphere;^17,23 despite the known presence of the κ-casein ‘hairy layer’, a soft diffuse surface layer is not required to adequately fit the data.^3,19 Under filtration forces the micelles deform and become ellipsoidal in shape.^21,22,26,27 Under osmotic pressure the micelle is compressed and can be fitted using a sponge model consisting of hard and soft regions.¹⁸ The internal structure is often fitted using hierarchical analytic functions^{4,15,16,20,24,25} or form factor models, where CCP may be modelled as spheres,^3,23 ellipsoids,^11,19 core–shell spheres (where the shell is a surrounding protein region)¹⁹ or cylinders,²¹ often with a hard sphere structure factor included to describe the inter-particle interaction.^3,11,21 The parameters used to describe the different levels can be constrained based on prior knowledge of the relative concentrations of calcium and protein.^3,15 The protein structures, if they are included in the model, may be described as chains or random coils.^15,26,28

Conventional SAXS measurements lack elemental information about the scattering objects. SANS experiments are often performed using different H₂O/D₂O ratios to vary the contrast between the serum, protein, and nanoclusters.^16,23,24 We have performed resonant soft X-ray scattering (RSoXS)^29,30 measurements on cow skim milk to conclusively identify the scattering features corresponding to CCP. This technique utilises the energy-dependence of the complex index of refraction near X-ray absorption edges to highlight contributions to the scattering profile from a particular element or chemical of interest.

Skim cow milk powders were obtained from Fonterra Co-operative Group Limited, New Zealand and reconstituted in water at a 1 : 10 w/w ratio. For RSoXS measurements, liquid samples were loaded into a sandwiched silicon nitride cell with a 1 μm spacer between two membranes. Measurements were performed at various X-ray energies around the oxygen K edge (∼540 eV) and Ca L₂/L₃ absorption edges (∼350 eV) on beamline 11.0.1.2 at the Advanced Light Source, Lawrence Berkeley National Laboratory.³¹ SAXS measurements were recorded at an X-ray energy of 8200 eV on the SAXS/WAXS beamline at the Australian Synchrotron, Melbourne.³²

RSoXS data collected on cow skim milk at various energies around the Ca L₂/L₃ absorption edges (∼350 eV) are shown in Fig. 1a, together with conventional SAXS collected with 8200 eV X-rays. The range of the RSoXS data is limited due to the long wavelength and does not capture the turnover at low q corresponding to the micelle size. At the absorption edge, a prominent peak centred at q = 0.035 Å⁻¹ emerges, with a shoulder at q = 0.08 Å⁻¹. The two features track together and show strong resonant dependence when the energy is scanned across the Ca L edges (Fig. 1b), indicating that they correspond to scattering objects containing calcium, i.e. CCP. At 350 eV they are no longer visible, which suggests contrast matching between the CCP and the matrix. The intensity at low q increases slightly at the same energies, consistent with the average contrast of the casein micelle increasing relative to the serum due to the presence of Ca within the micelle. Although we cannot extract a value for the radius of gyration R_g of the casein micelle from the RSoXS data (e.g. from a Guinier plot), the curves do not significantly change shape with energy in the low q region, indicating that R_g is unchanged (as also observed in SANS contrast studies³). However, unlike neutron scattering studies where the deuteration level of the water changes the contrast between the serum and micelle, in RSoXS both the casein micelle and the CCP clusters will contribute to the scattering in the low-q region differently at different energies, making an extraction of R_g non-trivial.

Fig. 1 (a) Resonant soft X-ray scattering (RSoXS) of liquid milk at various energies near the Ca L₂-edge as labelled, shown together with typical SAXS of liquid milk collected at 8200 eV (for which the intensity scale has been translated). (b) Intensity variation at fixed q (0.033 Å⁻¹) corresponding to the structure factor peak, as a function of incident X-ray energy. (c) Fit to a difference plot of I(349.2 eV) − I(350 eV) using a single population of interacting spheres to describe the CCP particle scattering (curve i) and two populations of non-interacting spheres to describe the casein micelle scattering (curve ii).

The difference data (I(349.2 eV) − I(350 eV)) were fitted to a simple model of separate populations of spheres to describe (a) the micelle scattering (two non-interacting populations) and (b) the CCP (one population with hard sphere structure factor³³). The data and fit are shown in Fig. 1c. The peak arises from the structure factor component (interacting hard spheres 15 ± 2 nm apart), while the shoulder arises from the form factor (sphere 4 ± 1 nm in diameter). These findings are in agreement with SANS studies by Holt et al.²³ who attributed the q = 0.035 Å⁻¹ feature to inter-particle scattering.

We note that the peak at q = 0.035 Å⁻¹ is not observed in the conventional hard X-ray small-angle scattering for cow skim milk (Fig. 1a). In contrast, the prominent SAXS feature at q = 0.08 Å⁻¹, previously attributed to the CCP size, is only weakly present in the RSoXS. The absence of a feature at q = 0.035 Å⁻¹ in conventional SAXS is compelling evidence that the q = 0.08 Å⁻¹ feature observed in SAXS is not primarily due to the form factor of the CCP nanoparticles.

Conclusions

These results provide conclusive evidence as to the nature of the principal scattering features observed in cow milk. The feature observed in SANS and some, but not all, SAXS experiments at q = 0.035 Å⁻¹ is a structure factor peak corresponding to the inter-particle separation of CCP nanoclusters present within casein micelles. The feature observed in SAXS experiments at q = 0.08–0.1 Å⁻¹ is not primarily due to the form factor from CCP nanoclusters. Instead, it is more likely to be due to protein inhomogeneities on a 1–3 nm length scale as suggested by de Kruif et al.^3,15 The overall picture of the casein micelle obtained from these results is consistent with a model where the casein micelle structure is stabilized primarily through both calcium phosphate nanoclusters and protein hydrophobic bonding.^2,12,23

Acknowledgements

This work was funded by the New Zealand Ministry of Business, Innovation and Employment (MBIE) under contract C08X1003. Portions of this research were undertaken on the SAXS/WAXS beamline at the Australian Synchrotron, Victoria, Australia, and beamline 11.0.1 at the Advanced Light Source, Berkeley. The Advanced Light Source is supported by the Director, Office of Science, Office of Basic Energy Sciences, of the U.S. Department of Energy under Contract no. DE-AC02-05CH11231.

Notes and references

1. P. F. Fox and A. Brodkorb, Int. Dairy J., 2008, 18, 677.
2. D. G. Dalgleish, Soft Matter, 2011, 7, 2265.
3. C. G. de Kruif, T. Huppertz, V. S. Urban and A. V. Petukhov, Adv. Colloid Interface Sci., 2012, 171, 36.
4. S. Marchin, J.-L. Putaux, F. Pignon and J. Léonil, J. Chem. Phys., 2007, 126, 045101.
5. D. J. McMahon and B. S. Oommen, J. Dairy Sci., 2008, 91, 1709.
6. C. W. Slattery and R. Evard, Biochim. Biophys. Acta, 1973, 317, 529.
7. P. Walstra, Int. Dairy J., 1999, 9, 189.
8. H. M. Farrell Jr, E. L. Malin, E. M. Brown and P. X. Qi, Curr. Opin. Colloid Interface Sci., 2006, 11, 135.
9. D. J. McMahon and W. R. McManus, J. Dairy Sci., 1998, 81, 2985.
10. C. Holt, Adv. Protein Chem., 1992, 43, 63.
11. A. Shukla, T. Narayanan and D. Zanchi, Soft Matter, 2009, 5, 2884.
12. D. S. Horne, Int. Dairy J., 1998, 8, 171.
13. D. S. Horne, Colloids Surf., A, 2003, 213, 255.
14. In the case of small-angle X-ray scattering, if the X-ray dose is too high, radiation damage can occur.
15. C. G. de Kruif, J. Appl. Crystallogr., 2014, 47, 1479.
16. S. Hansen, R. Bauer, S. B. Lomholt, K. B. Quist, J. S. Pedersen and K. Mortensen, Eur. Biophys. J., 1996, 24, 143.
17. P. Müller-Buschbaum, R. Gebhardt, S. V. Roth, E. Metwalli and W. Doster, Biophys. J., 2007, 93, 960.
18. A. Bouchoux, G. Gésan-Guisiou, J. Pérez and B. Cabane, Biophys. J., 2010, 99, 3754.
19. J. P. Mata, P. Udabage and E. P. Gilbert, Soft Matter, 2011, 7, 3837.
20. F. Pignon, G. Belina, T. Narayanan, X. Paubel, A. Magnin and G. Gésan-Guiziou, J. Chem. Phys., 2004, 121, 8138.
21. R. Gebhardt, M. Burghammer, C. Riekel, S. V. Roth and P. Müller-Buschbaum, Macromol. Biosci., 2008, 8, 347.
22. R. Gebhardt, C. Vendrely and U. Kulozik, J. Phys.: Condens. Matter, 2011, 23, 444201.
23. C. Holt, C. G. de Kruif, R. Tuinier and P. A. Timmins, Colloids Surf., A, 2003, 213, 275.
24. P. H. Stothart, J. Mol. Biol., 1989, 208, 635.
25. P. H. Stothart and D. J. Cebula, J. Mol. Biol., 1982, 160, 391.
26. R. Gebhardt, J. Appl. Crystallogr., 2014, 47, 29.
27. R. Gebhardt, T. Steinhauer, P. Meyer, J. Sterr, J. Perlich and U. Kulozik, Faraday Discuss., 2012, 158, 77.
28. H. Sørensen, J. S. Pedersen, K. Mortensen and R. Ipsen, Int. Dairy J., 2013, 33, 1.
29. S. Swaraj, C. Wang, H. Yan, B. Watts, J. Luning, C. R. McNeill and H. Ade, Nano Lett., 2010, 10, 2863.
30. C. Wang, D. H. Lee, A. Hexemer, M. I. Kim, W. Zhao, H. Hasegawa, H. Ade and T. P. Russell, Nano Lett., 2011, 11, 3906.
31. E. Gann, A. T. Young, B. A. Collins, H. Yan, J. Nasiatka, H. A. Padmore, H. Ade, A. Hexemer and C. Wang, Rev. Sci. Instrum., 2012, 83, 045110.
32. N. M. Kirby, S. T. Mudie, A. M. Hawley, D. J. Cookson, H. D. T. Mertens, N. Cowieson and V. Samardzic-Boban, J. Appl. Crystallogr., 2013, 46, 1670.
33. J. S. Pedersen, J. Appl. Crystallogr., 1994, 27, 595.

---