Probing the urea dependence of residual structure in denatured human alpha-lactalbumin

Abstract

Backbone (15)N relaxation parameters and (15)N-(1)H(N) residual dipolar couplings (RDCs) have been measured for a variant of human alpha-lactalbumin (alpha-LA) in 4, 6, 8 and 10 M urea. In the alpha-LA variant, the eight cysteine residues in the protein have been replaced by alanines (all-Ala alpha-LA). This protein is a partially folded molten globule at pH 2 and has been shown previously to unfold in a stepwise non-cooperative manner on the addition of urea. (15)N R(2) values in some regions of all-Ala alpha-LA show significant exchange broadening which is reduced as the urea concentration is increased. Experimental RDC data are compared with RDCs predicted from a statistical coil model and with bulkiness, average area buried upon folding and hydrophobicity profiles in order to identify regions of non-random structure. Residues in the regions corresponding to the B, D and C-terminal 3(10) helices in native alpha-LA show R(2) values and RDC data consistent with some non-random structural propensities even at high urea concentrations. Indeed, for residues 101-106 the residual structure persists in 10 M urea and the RDC data suggest that this might include the formation of a turn-like structure. The data presented here allow a detailed characterization of the non-cooperative unfolding of all-Ala alpha-LA at higher concentrations of denaturant and complement previous studies which focused on structural features of the molten globule which is populated at lower concentrations of denaturant.

Fig. 1

Schematic representation of the native structure of human α-lactalbumin. The α- and β-domains and the five α-domain helices are labelled. The diagram was generated using MOLSCRIPT (Kraulis 1991)

Fig. 2

¹⁵N relaxation data for all-Ala α-LA plotted as a function of the amino acid sequence. a {¹H}–¹⁵N NOE as I_sat/I_nonsat, b R₁ and c R₂ values for all-Ala α-LA in 4 M (black), 6 M (red), 8 M (cyan) and 10 M (gold) urea are shown. All data were collected at 600 MHz and 20°C. The secondary structure and domain organization found in native α-lactalbumin is summarized above

Fig. 3

Reduced spectral density mapping for all-Ala α-LA at a 0 MHz (J(0)), b 60 MHz (J(ω_N)) and c 522 MHz (J(0.87ω_H)) in 4 M (black), 6 M (red), 8 M (cyan) and 10 M (gold) urea. The secondary structure and domain organization found in native α-lactalbumin is summarized above

Fig. 4

a Difference in ¹⁵N chemical shifts for all-Ala α-LA at 6 and 10 M urea at 20°C [δ_15N(10 M)–δ_15N(6 M)]. b–e Experimental RDCs recorded for all-Ala α-LA in compressed polyacrylamide gels at 20°C (black) are compared with predicted RDCs for all-Ala α-LA based on a coil-model (red). Measurements were made in b 4 M, c 6 M, d 8 M and e 10 M urea. The secondary structure and domain organization found in native α-lactalbumin is summarized above

Fig. 5

a Experimental RDCs recorded for all-Ala α-LA in 10 M urea (black) are compared with the bulkiness profile (red) calculated as described by Cho et al. (2007). b Normalised values for all Ala α-LA of the average area buried upon unfolding (AABUF) (black) (Rose et al. 1985) are compared with the hydrophobicity using the Abraham & Leo scale (red) (Abraham and Leo 1987). All values were averages over 7-residue windows. The dashed grey line shows the mean value of AABUF and hydrophobicity