Diet affects glycosylation of serum proteins in women at risk for cardiometabolic disease

Abstract

Background: Glycoproteomics deals with glycoproteins that are formed by post-translational modification when sugars (like fucose and sialic acid) are attached to protein. Glycosylation of proteins influences function, but whether glycosylation is altered by diet is unknown.

Objective: To evaluate the effect of consuming a diet based on the Dietary Guidelines for Americans on circulating glycoproteins that have previously been associated with cardiometabolic diseases.

Design: Forty-four women, with one or more metabolic syndrome characteristics, completed an 8-week randomized controlled feeding intervention (n = 22) consuming a diet based on the Dietary Guidelines for Americans (DGA 2010); the remaining consumed a 'typical American diet' (TAD, n = 22). Fasting serum samples were obtained at week0 (baseline) and week8 (post-intervention); 17 serum proteins were chosen for targeted analyses. Protein standards and serum samples were analyzed in a UHPLC-MS protocol to determine peptide concentration and their glycan (fucosylation or sialylation) profiles. Data at baseline were used in correlational analyses; change in proteins and glycans following intervention were used in non-parametric analyses.

Results: At baseline, women with more metabolic syndrome characteristics had more fucosylation (total di-fucosylated proteins: p = 0.045) compared to women with a lesser number of metabolic syndrome characteristics. Dietary refined grain intake was associated with increased total fucosylation (ρ = - 0.530, p < 0.001) and reduced total sialylation (ρ = 0.311, p = 0.042). After the 8-week intervention, there was higher sialylation following the DGA diet (Total di-sialylated protein p = 0.018, poly-sialylated orosomucoid p = 0.012) compared to the TAD diet.

Conclusions: Based on this study, glycosylation of proteins is likely affected by dietary patterns; higher sialylation was associated with a healthier diet pattern. Altered glycosylation is associated with several diseases, particularly cancer and type 2 diabetes, and this study raises the possibility that diet may influence disease state by altering glycosylation.

Clinical trial registration: NCT02298725 at clinicaltrials.gov; https://ichgcp.net/clinical-trials-registry/NCT02298725 .

Keywords: Dietary Guidelines for Americans; Fucosylation; Glycan; Glycoproteomics; Glycosylation; Post-translational modification; Sialylation.

Conflict of interest statement

S. Adams is founder and principal of XenoMed LLC, a company with no competing interests with the studies presented herein.

© 2021. The Author(s).

Figures

Fig. 1

All significantly different glycovariant mol% of individual or total peptides arranged within each panel from left to right in increasing order of glycosylation (none-mono-di-poly). Box and whisker plots showing the median ± interquartile range values, with p-values inset. Panel A shows differences based on screening characteristics—DL = dyslipidemic (n = 18), DL + GIT (n = 26) = dyslipidemic + glucose intolerant. Panel B presents differences between pre (n = 23- and post-menopausal women (n = 21). Panel C shows differences between overweight (OW—BMI between 25 and 30 kg/m2, n = 15) and obese (OB, BMI between 30 and 40 kg/m2, n = 29) individuals. VTNC Vitronectin, CERU ceruloplasmin, TOTAL all peptides together, KLKB1 Kallikrein, sialyl—sialylated, fucosyl fucosylated

Fig. 2

Correlation based significant associations between glycovariant mol% (y-axis) and HEI sub-category scores (x-axis) with inset Spearman’s rho (ρ) and p values. For total vegetables, greens and beans, seafood and plant proteins, total dairy and total score higher score reflects both higher intake of these food groups and a ‘healthy’ diet. For refined grain a higher score indicates lower intake and a ‘healthy’ diet

Fig. 3

Mol% glycoprotein changes (wk8–wk0) comparing DGA and TAD groups. Box plot represents median + IQR, and points show data (there were no statistical outliers) with p values inset. Only analytes with significant diet differences (p < 0.05) are depicted here. A2MG alpha-2-macroglobulin, KNG-1 Kininogen, CERU ceruloplasmin, AGP-1 alpha-1-acid glycoprotein, sialyl sialylated, fucosyl fucosylated, TAD typical American diet, DGA Dietary guidelines for American diet

Fig. 4

Loadings and scores plot of a PLS-DA model generated to predict ‘Group’ using difference in wk8–wk0 in anthropometric, clinical and glycovariant data. The scores plot (a) shows the participant distribution across the n-dimensions is inset within the loadings plot (b) which shows the variables (dimensions). In the scores plot (a) the black dots represent scores from subjects fed the TAD and orange dots represent scores of participants fed the DGA. In both scores and loadings plot (b), the orange highlight ellipses represent DGA and dark grey ellipses highlight TAD group. c Displays the VIP variables with VIP score > 1, which significantly contribute to the model discrimination of DGA and TAD groups, coded with orange for variables that are associated with change in DGA and black for TAD