|Home | About | Journals | Submit | Contact Us | Français|
Covalent modifications of histones play a role in regulating telomere attrition and cellular senescence. Biotinylation of lysine (K) residues in histones, mediated by holocarboxylase synthetase (HCS), is a novel diet-dependent mechanism to regulate chromatin structure and gene expression. We have previously shown that biotinylation of K12 in histone H4 (H4K12bio) is a marker for heterochromatin and is enriched in pericentromeric alpha satellite repeats. Here we hypothesized that H4K12bio is also enriched in telomeres. We used human IMR-90 lung fibroblasts and immortalized IMR-90 cells overexpressing human telomerase (hTERT) in order to examine histone biotinylation in young and senescent cells. Our studies suggest that one out of three histone H4 molecules in telomeres is biotinylated at K12 in hTERT cells. The abundance of H4K12bio in telomeres decreased by 42% during telomere attrition in senescent IMR-90 cells; overexpression of telomerase prevented the loss of H4K12bio. Possible confounders such as decreased expression of HCS and biotin transporters were formally excluded in this study. Collectively, these data suggest that H4K12bio is enriched in telomeric repeats and represents a novel epigenetic mark for cell senescence.
Telomeres, the heterochromatic end caps of linear chromosomes, are comprised of thousands of TTAGGG repeats in human chromatin, and bind proteins such as TRF1 (Telomeric Repeat Binding Factor 1), TRF2, and tankyrase [1-3]. Telomeres play an important role in chromosome stability by preventing the loss or mutations of genes and the fusion of chromosomes [1,4,5]. Due to the inability of DNA polymerases to replicate the end of chromosomes, telomere length is shortened by approximately 50-150 bp with each cell division . Division ceases when telomeres reach a critical threshold length in senescent cells [6-8]. In germline cells, immortalized cell lines, and cancer cells, telomere shortening is prevented by telomerase, which is composed of two major components, the telomerase reverse transcriptase and telomerase RNA template. Telomerase helps to maintain telomere length by adding TTAGGG repeats to the chromosome ends during replication [9-13].
Chromatin structure and gene expression is regulated by covalent modifications of histones H1, H2A, H2B, H3, and H4 [14,15]. DNA and histones form the nucleosomal core particles, which consist of 146 base pairs of DNA wrapped around an octamer of core histones (one H3-H3-H4-H4 tetramer and two H2A-H2B dimers) . Histone H1 binds to the DNA in-between two nucleosomal core particles, completing the nucleosomal assembly . The amino terminal tails of histones that protrude from the nucleosomal surface are subjected to various covalent modifications [17,18]. Acetylation, phosphorylation, and methylation of histone tails play a major role in regulating telomere attrition by controlling telomerase activity [19-21].
Recently, it was shown that biotinylation of lysine (K) residues in histones, a novel diet-dependent histone modification, plays a role in gene regulation [22-24]. Biotinylation is mediated by holocarboxylase synthetase (HCS) [23,25]. Human histone biotinylation sites include K9, K13, K125, K127, and K129 in histone H2A , K4, K9, K18, and perhaps K23 in histone H3 [27,28], and K8 and K12 in histone H4 .
We have shown that K12-biotinylated histone H4 (H4K12bio) is enriched in repeat regions such as pericentromeric alpha satellite repeats, and retrotransposons, and also participates in gene repression [24,30]. Here, we tested the hypothesis that H4K12bio is a general marker for repeat regions and therefore is enriched in telomeric repeats. We propose that decreased abundance of H4K12bio is an epigenetic mark for cell senescence.
In this study we used primary human IMR-90 lung fibroblasts as a model to investigate histone biotinylation in aging cells. IMR-90 cells were chosen because they undergo telomere attrition and enter a state of irreversible growth arrest, designated cellular senescence, after about 50-55 population doublings (PD). IMR-90 cells overexpressing human telomerase reverse transcriptase (hTERT cells) do not undergo senescence  and were used as immortal controls. Our goals were to determine (1) whether H4K12bio is enriched in telomeric repeats in human cells, (2) whether the abundance of H4K12bio decreases with aging, and (3) whether overexpression of telomerase prevents loss of H4K12bio.
Human lung fibroblasts IMR-90 cells were purchased from American Type Culture Collection (Manassas, VA). These cells were passaged 2-3 times per week and replated at 60,000 to 80,000 cells in T75 cell culture flasks with 15 mL medium. IMR-90 cells were harvested at timed intervals. Young cells at PD 35 were used as baseline. IMR-90 cells were considered to be senescent when they did not need to be passaged for 2 weeks, and at this stage they were at a PD of 50-55. hTERT cells were kindly provided by Dr. Judith Campisi (Lawrence Berkeley National Laboratory, CA) at PD 55. hTERT cells at PD 60 were used as baseline and harvested at 5 or 10 additional PD. Cells were cultured in Minimum Essential Medium (MEM) supplemented with 10% (v/v) fetal bovine serum, 0.1% nonessential amino acids, 100,000 units/L penicillin and 100 mg/L streptomycin.
Senescent cells are known to express β-gal, which produces a blue color in the β-gal assay; the β-gal assay was performed as described .
Telomerase activity was detected by telomerase repeat amplification protocol (TRAP) assay  with slight modifications. In this assay, telomerase from cell extracts ligates synthetic TTAGGG repeats to the 3’ end of the telomerase substrate oligonuceotide “TS” (5’-AATCCGTCGAGCAGAGTT-3’). The extended product is then amplified by PCR using TS and the reverse primer CX (5’-CCTTACCCTTACCCTTACCCTAA-3’), which generates a ladder of PCR products with size increasing in 6-base increments starting at 50 nucleotides. Cell lysate  was incubated on ice for 30 min and centrifuged at 12,000 g for 20 min at 4°C. The protein concentrations in lysates were adjusted to 1 g/L by dilution with lysis buffer; 2 μL of lysate was used in TRAP reaction as follows. PCR amplification of telomerase extension products was performed in 50 μL TRAP reaction mixes containing 20 mM Tris-HCl (pH 8.3), 1.5 mM MgCl2, 68 mM KCl, 0.005% Tween 20, 1 mM EGTA, 50 μM dNTP, TS and CX primers at 1.8 mg/L, and 2 U of Taq polymerase. Samples were denatured for 2 min at 94 °C, followed by 30 cycles of PCR amplification at 94 °C for 10 sec, 50 °C for 25 sec and 72 °C for 30 sec in a thermal cycler. Samples were analyzed using 10% TBE gels (Invitrogen) with 0.5 mg/L ethidium bromide staining.
Relative telomere lengths were determined as described previously  with slight modifications using the iCycler iQ™ multicolour real-time detection system (Bio-Rad, Palo Alto, CA). The relative telomere lengths were determined as the factor by which a genomic DNA sample differs from a reference sample in its ratio of telomere (T) repeat copy number to a single (S) copy gene number. The cycle threshold (Ct) values generated were used to calculate T/S values for each sample: T/S =2-ΔCt (where ΔCt = Ctsinglecopygene - Cttelomere). Long telomeres produce low Ct values because more PCR product is generated due to many PCR initiation sites; this results in higher T/S ratios. Genomic DNA was extracted from IMR-90 (PD 35 and 45) and hTERT cells (PD n and PD n+10). Standard curves were generated for telomeres and the single-gene copy gene; DNA concentrations in standards ranged from 10 to 100 ng/reaction. Triplicate PCR reactions were carried out in a 25 μL total reaction mix containing ABsolute QPCR SYBR Green fluorescein mix (ABgene, Rochester, NY). Primers for telomeres (T1a and T1b) and the single copy gene 36B4 (acidic ribosomal phosphoprotein PO) were added to obtain final concentrations of 0.2 μM and 0.3 μM, respectively .
IMR-90 and hTERT cells were used to quantify the relative enrichment of proteins in telomeric repeats. ChIP assays were conducted as described . The following antibodies were used for chromatin immunoprecipitations; polyclonal antisera against H4K12bio , and antibodies against telomere protein TRF2 and the C-terminus in histone H3 (Abcam). Efficiency of anti-H4K12bio in ChIP assays was tested by comparing nuclear extracts from ChIP assays before and after treatment with anti-H4K12bio; in these assays supernatants from before and after immunoprecipitation were probed with anti-H4K12bio using Western blot analysis . Chromatin immunoprecipitated with the preimmune serum of H4K12bio served as the negative control. Each immunoprecipitation was repeated three times, using 47 million cells per precipitation. Cross-links were reversed, and DNA was recovered for ligation-mediated PCR. For total DNA samples (denoted input DNA), aliquots corresponding to 1/6th of the lysate volume used for immunoprecipitation were processed along with the rest of immunoprecipitated samples during the crosslink reversal step. DNA was amplified by ligation-mediated PCR (LMPCR) as described [35,36] with the following modifications. An asymmetric double-stranded T7 linker composed of a 25-mer (5'-GGCTAATACGACTCACTATAGGGAG-3’) annealed to a 13-mer (5’-CTCCCTATAGTGA-3’) was prepared by dissolving 670 pmol of each oligonucleotide in 100 μl of water, heating at 95°C for 5 min, and cooling to room temperature. Blunt-end DNA was produced by mixing 25 ng of input or antibody-precipitated DNA and 5 U T4 DNA polymerase (Invitrogen, Carlsbad, CA). Blunt-end DNA was purified by using QIAprep Spin Miniprep Kit (Qiagen, Inc., Valencia, CA) and resuspended in 20 μl of 10 mM Tris-Cl, pH 8.5. DNA was ligated to the T7 linker by incubating with 5 U T4 DNA Ligase (Ambion, Austin, TX) at 16°C overnight. Ligation products were purified by eluting with 20 μl of 10 mM Tris-HCl, pH 8.5, using QIAprep Spin Miniprep Kits. Ligation products were PCR-amplified twice. First, 18 μl ligation products were amplified for 20 cycles by using 50 pmol 25-mer linker primer and 5 U Taq DNA polymerase (New England Biolabs, Ipswich, MA) as follows: 94°C for 30 s, 55°C for 30 s, 72°C for 1 min and 15 s, followed by a final extension at 72°C for 4 min. Ligation products were purified by eluting with 20 μl of 10 mM Tris-HCl, pH 8.5 using Miniprep Kits, and subjected to another 20 cycles of PCR under the same conditions as above. Ligation products were purified by eluting with 20 μl of ddH2O using Spin Miniprep Kits.
Equal amounts of input and immunoprecipitated DNA was dot blotted onto a nylon membrane and fixed with UV-crosslinking. Membranes were dried, and hybridized using 32P-labeled (CCCTAA)4 as a probe for telomeric repeats. Intensities of dots were determined using Image J 1.38x software . Dot blot analysis data (relative intensities of dots) was obtained for three independent blots. Input DNA represents the total telomeric DNA in cell extracts. The amount of telomeric DNA immunoprecipitated in each ChIP was calculated based on the signal relative to the corresponding total telomeric (input) DNA. ChIP data are reported as ratios of immunoprecipitated DNA to an equal amount of input DNA.
Co-immunoprecipitation experiments were conducted to confirm physical interaction between TRF2 and H4K12bio. In these experiments nuclear extracts from ChIP assays were precipitated with anti-H4K12bio, followed by Western blot analysis with anti-TRF2 [29,38].
Nuclear histones were extracted from IMR-90 and hTERT cells using 1 M HCl . Histones were resolved on 18% Tris-Glycine gels (Invitrogen, Carlsbad, CA). Transblots were probed with streptavidin peroxidase  and the following primary antibodies: rabbit anti-human H4K12bio , rabbit anti-human H3K9bio, and rabbit anti-human H3K18bio . Rabbit anti-human H4K20me3 (Abcam, Cambridge, MA) is a known marker for aging  and was used as a control. Antibody and avidin blots were quantified using chemiluminescence and gel densitometry . Equal loading of lanes and sample integrity was confirmed by using bicinchoninic assay (Pierce, Rockford, IL), densitometric quantitation of gels stained with coomassie blue, and by western-blot analysis using an antibody to the C-terminus in histone H3 (Santa Cruz Biotechnology; Santa Cruz, CA).
Abundance of HCS in whole cell homogenates was quantified by western-blot analysis as described previously . Histone H3 (Santa Cruz Biotechnology) was used as loading control . Abundance of mRNA coding for HCS was quantified as described . HCS is also responsible for catalyzing the attachment of biotin to acetyl-CoA carboxylase (AC) 1 and 2, pyruvate carboxylase (PC), propionyl-CoA carboxylase (PCC), and 3-methylcrotonyl carboxylase (MCC) . Abundance of these holocarboxylases was quantified by streptavidin blotting as described previously . Equal loading of protein was confirmed by gel densitometry after probing with anti-human histone H3.
The abundance of biotin transporter (SMVT) protein and mRNA were quantified as described . Rates of biotin transport into IMR-90 cells were determined using [3H]biotin at a physiological concentration of 475 pmol/L as described previously .
Homogeneity of variances among groups was confirmed using Bartlett's test . Significance of differences among groups was tested by one-way ANOVA. Fisher's protected least significant difference procedure was used for post hoc testing . StatView 5.0.1 (SAS Institute; Cary, NC) was used to perform all calculations. Differences were considered significant if P ≤ 0.05. Data are expressed as mean ± SD.
IMR-90 cell cultures were harvested at PD 35, 40 and 45. hTERT cells were received at PD 55 and harvested at PD 60, 65, and 70, which are referred to as n, n+5, and n+10 in the text. Telomere length, telomerase activity, and β-galactosidase (β-gal) activity were measured to validate our cell models.
Relative telomere length was quantified as the ratio of telomere (T) repeat copy number to a single (S) copy gene number (T/S ratio) in a genomic DNA sample. The T/S ratio is directly proportional to the average telomere length [34,45]. The T/S ratios for IMR-90 cells at PD 35, 40 and 45 were 1.0 ± 0.1, 0.9 ± 0.01, and 0.7 ± 0.03, respectively, consistent with telomere attrition in aging cells. The T/S ratios of hTERT cells (control) were 1.0 ± 0.1 and 1.1 ± 0.1 at PD n and n+10, respectively, indicating no change in telomere length. Telomerase activity was detectable in hTERT cells (positive control) but absent in both IMR-90 cells and heat-treated hTERT cells (negative control, Figure 1). β-gal is a known marker for cell senescence ; β-gal activity was detected in IMR-90 (PD 45) but not in hTERT (n+10) cells (Figure 2). IMR-90 cells ceased to proliferate at about PD 55, whereas hTERT cells continued to proliferate during the entire observation period (data not shown). These results confirm that IMR-90 cells undergo senescence and show decreasing telomere lengths with aging, whereas telomerase-expressing hTERT cells escape senescence and maintain their telomeres during cell division.
Previous studies suggested that H4K12bio is enriched in repeat regions such as pericentromeric alpha satellite repeats, retrotransposons [24,30]. Here we used chromatin immunoprecipitation (ChIP) assays in combination with dot blotting to demonstrate that H4K12bio also is enriched in telomeric repeats.
Chromatin from hTERT and IMR-90 cells was precipitated with antibodies to H4K12bio, the C-terminus in histone H3, and TRF2, spotted onto a nylon membrane, and probed with a telomere-specific oligonucleotide (Figure 3). Relative enrichment of telomeric repeats in each ChIP was quantified as the signal intensity relative to the corresponding telomeric signal in total genomic DNA (input DNA). In this assay, input DNA represents the total amount of telomeric repeats, whereas precipitation with anti-H4K12bio, anti-histone H3, and anti-TRF2 is a marker for the relative enrichment of these proteins in telomeric repeats. All three proteins were associated with telomeric repeats (Figure 3A). Dot blot analyses of hTERT ChIP suggested that 31 ± 4% of the histone H4 in telomeric repeats is biotinylated at K12 (n=3). This calculation was based on the fact that equal amounts of H3 and H4 exist in nucleosomes [16,46], and that the relative enrichments of telomeric repeats in pull-down DNA with anti-H4K12bio, anti-histone H3, and anti-TRF2 were 0.3, 1.0, and 1.5, respectively, in hTERT cells. Note that the C-terminus in histone H3 is not known to be covalently modified and that chromatin immunoprecipitation with anti-H3 is a well-established marker for quantitation of nucleosomal abundance [47-49]. The relative enrichments depended on the PD (0.8, 1.0, and 1.1 at PD 35 vs. 0.6, 1.0, and 1.1 at PD 45) in IMR-90 cells (Figure 3B). Furthermore, ChIP data from IMR-90 showed that the relative enrichment of telomeric repeats in total telomeric DNA (input DNA) was reduced by 23%, whereas the pull-down DNA with anti-H4K12bio, anti-histone H3, and anti-TRF2 decreased by 42, 25, and 24 %, respectively, at PD 45 compared to those of PD 35, confirming the association of H4K12bio with telomeres. Co-immunoprecipitation experiments confirmed that TRF2 physically interacts with H4K12bio, consistent with enrichment of H4K12bio in telomeres (Figure 3C). Note that the abundance of H4K12bio in telomeres is an estimate that should be interpreted carefully, as it somewhat depends on the affinity and quality of antibodies used in chromatin immunoprecipitation assays. Importantly, our studies suggest that our anti-H4K12bio antibody in combination with our protocol for ChIP assays produced a near-complete precipitation of H4K12bio in nuclear extracts (Figure 4); incubation with pre-immune serum did not produce a meaningful depletion of H4K12bio in nuclear extracts (negative control).
Based on our observation that H4K12bio is enriched in telomeres we expected to see an age-dependent decrease in biotinylation of histones in bulk histone extracts from IMR-90 nuclei but not hTERT cell nuclei. When IMR-90 nuclear histones were probed with streptavidin, which is a probe for biotin regardless of the biotinylation sites in histones, the abundance of biotinylated histones was 20 ± 4.3% lower in PD 45 cells compared with PD 35 cells (Figure 5A) as judged by gel densitometry. Next we looked at distinct biotinylation sites. The abundance of H4K12bio significantly decreased by 85 ± 8.4% at PD 45 compared to PD 35 in IMR-90 cells (Figure 5B). In contrast, biotinylation of K9 and K18 in histone H3 showed no change (Figure 5C and 5D). The known aging marker K20 trimethylated in histone H4 (H4K20me3)  increased with aging in IMR-90 cells (positive control, Figure 5E). There was no change in hTERT cells if histones were probed with antibodies to H4K12bio, H3K9bio, H3K18bio, and H4K20me3. Staining with coomassie blue and probing with anti-histone H3 was used to confirm equal loading of histones from within each group (Figures 5F and 5G).
Theoretically, the expression of HCS [23,25] and the biotin transporter SMVT  might affect histone biotinylation. Therefore we investigated the expression of these proteins in IMR-90 and hTERT cells. The expression of HCS (as judged by mRNA and protein abundance) and SMVT (abundance of mRNA and protein, and biotin transport rates) did not change in hTERT cells during the observation period (data not shown). Similarly, the expression of HCS did not change in IMR-90 cells (data not shown). In contrast, the abundance of SMVT mRNA and protein decreased by 25 ± 2.1% and 65 ± 5.2%, respectively, in PD 45 compared to PD 35 in IMR-90 cells (Figures 6A and 6B). Equal loading of proteins were confirmed by probing with anti-histone H3 (Figure 6B). The decreased expression of SMVT in aging IMR-90 cells was paralleled by a decrease in biotin transport rates. Biotin transport rates decreased from 49 ± 1.2 to 18 ± 1.1 amol/(μg protein*30 min) at PD 35 and PD 45, respectively, in IMR-90 cells (Figure 6C).
Holocarboxylases are a well established marker of biotin status [42,51]. Therefore we quantified holocarboxylases to determine whether the decreased expression of biotin transporters in aging IMR-90 cells decreased the intra-cellular levels of biotin. The abundance of biotinylated pyruvate carboxylase (PC), propionyl-CoA carboxylase (PCC), and 3-methylcrotonyl-CoA carboxylase (MCC) did not change in hTERT cells (PD n+10 vs. PD n). In contrast, biotinylation of PC decreased by 40 ± 4.2% at PD 45 compared with PD 35 in IMR-90 cells (Figure 7A, P < 0.05, PD 45 vs. PD 35; n=3). Biotinylation of PCC and MCC decreased to a similar extent in aging IMR-90 cells. Equal loading of carboxylases were confirmed by probing with anti-histone H3 (Figure 7B).
Here we report a novel epigenetic mark for telomeric repeats in humans, H4K12bio. Our studies suggest that about one-third of the histones H4 in telomeres are biotinylated in position K12. Consistent with this observation, telomere attrition in aging cells was associated with a loss of H4K12bio. This loss of H4K12bio was prevented by overexpression of hTERT. Aging had no effect on the biotinylation of histones other than H4. These observations are consistent with previous studies in our laboratory in which we demonstrated that H4K12bio is a mark for repeat regions and repressed genes [24,41].
In this study we have formally excluded a number of potential confounders that might affect biotinylation of histones. First we have shown that the expression of HCS does not depend on aging. Therefore, the low abundance of H4K12bio in aging cells is not caused by a decrease in the levels of this essential histone biotinyl ligase . In previous studies we have demonstrated that another putative histone biotinyl ligase, biotinidase, is less important than HCS for histone biotinylation . Second, we have shown that the cellular availability of biotin in aging IMR-90 cells was sufficient to maintain biotinylation of carboxylases despite a decreased expression of the determined biotin transporter SMVT. We demonstrated that biotin transport rates decreased by about 60% in aging cells compared with young cells but that effects on biotinylation of carboxylases, PC, PCC, and MCC, were moderate (<40%) compared with decreased abundance of H4K12bio (>85%) in these cells. Previous studies in human cell lines suggest that biotinylation of carboxylases is an early and sensitive marker for cellular biotin levels [42,52,53].
Theoretically, the difference between the depletion of holocarboxylases and H4K12bio could have been caused by distinct turnover rates of these proteins. The half-lives of carboxylases are only a few days [54-58], whereas half-lives of histones may extend to several weeks depending on the type of tissue and class of histone investigated [59,60]. Therefore, it is unlikely that the decrease in H4K12bio seen in this study was caused by a rapid turnover of this protein, although we cannot formally exclude this possibility. We propose that telomere attrition is responsible for the depletion of H4K12bio in aging cells.
In this study the abundance of H3K9bio and H3K18bio in bulk histones was not affected by aging. Functions of K9- and K18- biotinylated histone H3 are unknown and are currently being investigated in our laboratory along with other known histone biotinylation marks [26-29]. Preliminary studies suggest that some of these biotinylation marks are enriched in transposable elements, such as long terminal repeats in humans and mice .
The structural function of H4K12bio in telomeres is unknown, but it appears to be related to gene repression and heterochromatin formation . In addition, please note that telomeres bind to the nuclear envelope, mediated by proteins such as Membrane Telomere Binding Protein (MTBP), SUN1, and Sun2 [61-63]. HCS also localizes primarily to the nuclear lamina , but any mechanistic links between HCS, nuclear compartmentalization, and telomeres are unknown. We are now in the process of identifying proteins that interact with HCS and our ongoing studies tentatively revealed a number of proteins that may play critical roles in chromatin biology.
Grants, Sponsors and Funding sources: A contribution of the University of Nebraska Agricultural Research Division, supported in part by funds provided through the Hatch Act. Additional support was provided by NIH grants DK063945, DK077816 and ES015206, USDA CSREES grant 2006-35200-17138, and by NSF EPSCoR grant EPS-0701892.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.