Nutritional Factors Modulating Alu Methylation in an Italian Sample from The Mark-Age Study Including Offspring of Healthy Nonagenarians

Alu hypomethylation promotes genomic instability and is associated with aging and age-related diseases. Dietary factors affect global DNA methylation, leading to changes in genomic stability and gene expression with an impact on longevity and the risk of disease. This preliminary study aims to investigate the relationship between nutritional factors, such as circulating trace elements, lipids and antioxidants, and Alu methylation in elderly subjects and offspring of healthy nonagenarians. Alu DNA methylation was analyzed in sixty RASIG (randomly recruited age-stratified individuals from the general population) and thirty-two GO (GeHA offspring) enrolled in Italy in the framework of the MARK-AGE project. Factor analysis revealed a different clustering between Alu CpG1 and the other CpG sites. RASIG over 65 years showed lower Alu CpG1 methylation than those of GO subjects in the same age class. Moreover, Alu CpG1 methylation was associated with fruit and whole-grain bread consumption, LDL2-Cholesterol and plasma copper. The preserved Alu methylation status in GO, suggests Alu epigenetic changes as a potential marker of aging. Our preliminary investigation shows that Alu methylation may be affected by food rich in fibers and antioxidants, or circulating LDL subfractions and plasma copper.


Supplementary
. Schematic representation of an Alu element, composed by a left arm and a right arm, divided by an A-rich region (A n in the center of the graph). The bended arrow on the left represents the transcription start site. A typical Alu element is followed by a poly-A tail (A n on the right of the graph) and flanked two direct repeats (target site duplications, TSD in the graph). The other letters in the graph indicate the main transcription factor binding sites, known as A box, B box and A' box. The five small "lollipops" above the central part of the graph indicate the position of the five CpG methylation sites which have been assayed by bisulfite pyrosequencing in this work. The small arrow on the right of the "lollipops" indicates the region corresponding to the pyrosequencing primer, and the sequencing direction. The scale, and the position of the sequence features in the scheme, are calculated based on a consensus sequence of Alu Sx subfamily, accessed through the Repbase Update database (Jurka et. al., 2005) on the Genetic Information Research Institute web site ( www.girinst.org). The exact position of the internal features in the Alu Sx sequence is based on previous publications (Cui et al., 2011;Luo et al., 2014).

Supplementary Figure 2S. Position of the assayed CpGs in the Alu Sx consensus sequence (before and after bisulfite conversion)
Upper box: the consensus sequence of Alu Sx subfamily, obtained through the Repbase Update database (Jurka et. al., 2005) available at the Genetic Information Research Institute web site ( www.girinst.org). The region analysed in the bisulfite pyrosequencing assay is evidenced in blue. The letters in bold represent the 5 assayed CpGs. The sequencing direction is antisense with respect to the consensus, hence the CpG n.1 in the assay corresponds to the CpG located more near to the 3'-end in the above reported sequence. The underlined traits (starting from the 5' end) represent the following features of the Alu sequence: the A box; the B box; the A-rich region; the A' box (Cuo et al., 2011;Luo et al., 2014). Lower box: the region of the bisulfite-converted Alu Sx consensus corresponding to the assay target is reported. Letters " Y" indicates CpG cytosines, which have been converted into thymines only when not methylated. The PCR primers and the sequencing primer are also shown. The box indicates the region assayed by pyrosequencing, the arrow represents the sequencing direction, and the letters evidenced in yellow represent the five CpG cytosines (partially converted in thymines) for which the percentage of methylation is measured.

Supplementary Table 1S
Details of the pyrosequencing assay for Alu elements Sequence to analyze RYRYRCCACYAYRCCYRACTAA Dispensation order CGACTGACTGACTACTCACTGACTAGACT Details of the pyroseqeuncing assay for PyroMark Control Oligo (Qiagen) TAYGGTTTGC dispensation order CTGACTGTG Supplementary Figure 3S. Calibration curves obtained by pyrosequencing analysis of a control, bisulfite-converted genomic DNA sample. To generate each point of the curve, a sample with known methylation has been obtained by mixing different proportions of a totally methylated and a totally unmethylated converted genomic DNA sample (control samples obtained from Qiagen). Note: "True methylation": known methylation of the control sample; "Estimated methylation": methylation value obtained by the Alu pyrosequencing assay.

Supplementary Figure 4S. Alu bisulfite pyrosequencing: in silico determination of the assay targets
The complementarity of the primers used for the bisulfite pyrosequencing assay has been checked against the consensus sequences of all Alu subfamilies present in the RepBase Update database (www.girinst.org, data downloaded on 31-10-2017). Only part of the Alu S subfamilies have a complete complementarity (after bisulfite conversion) with all the primers (the two PCR primers and the pyrosequencing primer) used in the assay: Alu Sx, Alu Sg, Alu Sz, Alu Sz6, Alu Sg1, Alu Sg4, Alu Sq, Alu Sq4, Alu Sx1, Alu Sg7, Alu Sq2, Alu Sq10 and Alu Sp. As shown in the Supplemental figure below, only part of the Alu subfamilies targeted by the assay contain in their consensus sequences all the 5 CpG sites analysed.

Supplementary Table 2S. Alu bisulfite pyrosequencing: genomic distribution of target Alu subfamilies.
By using the CLC Biomedical Genomics Workbench software (Qiagen), the genomic positions of the Alu elements of the targeted subfamilies have been annotated with respect to known genes and regulatory regions (see table below). All annotations are referred to the human hg19 genome.

Alu elements of target subfamilies 1
Overlapping with genes 2

Overlapping with exons 3
Overlapping with regulatory regions 4 Position with respect to genomic annotations

Intragenic
Intronic 329 Table 6S. Automatic linear regression analysis for variables independently associated with mean levels of Alu methylation at CpG2, CpG3, CpG4 and CpG5 in RASIG and GO donors a For meat consumption: =1 serv/day was compared to consumption of meat 2-7 serv/week (used as reference). b For brown bread consumption: = 1-6 serv/week was automatically combined with ≥ 1 serv/day, used as reference and compared to < 1 serv/day.