To further understand the regulation of mtDNA by other proteins, we analyzed publicly available ChIP-seq datasets from ENCODE, modENCODE and mouseENCODE for evidence of nuclear transcription factor binding to the mitochondrial genome. ChIP-seq is a powerful tool for understanding the interactions between proteins and the mitochondrial genome, and future studies promise to further our understanding of how mtDNA is regulated in the nucleoid.
Background
First, mutations at specific TFAM binding sites lead to inactivation of promoter activity (Dairaghi, Shadel et al. 1995). Further analytical ultracentrifugation experiments indicate that the C-terminal tail is required for dimerization (Wong, Rajagopalan et al. 2009).
Overview of Thesis
34;Import of mitochondrial transcription factor A (TFAM) into rat liver mitochondria stimulates transcription of mitochondrial DNA.". 34;Overexpression of mitochondrial transcription factor a ameliorates mitochondrial deficiencies and heart failure after myocardial infarction." Circulation.
Introduction
Importantly, mice deficient in any of the fusion proteins die mid-digestion (Chen, Detmer et al. 2003, Davies, Hollins et al. 2007), indicating that mitochondrial fusion is essential for viability. Loss of fusion leads to defects in respiratory capacity and potential morphological and membrane heterogeneity (Chen, Chomyn et al. 2005). Furthermore, maintenance of proper mitochondrial function appears to be essential for development, as evidenced by the embryonic lethality of splicing-deficient mice (Chen, Detmer et al. 2003).
Depletion of mtDNA copy number is known to cause cellular bioenergetics defects (Baron, Kudin et al. 2007). Interestingly, many neurodegenerative diseases, including ALS and FRDA (Baron, Kudin et al. 2007), as well as many mitochondrial myopathies (Alberio, Mineri et al. 2007), have also been associated with mtDNA copy number deficiencies.
Results
The average ratio of mtDNA to nuclear genome (Shen-Li, O'Hagan et al.) in MEFs was found to be 440 mtDNA:nDNA. Strikingly, while absence of either Mfn1 or Mfn2 alone did not affect mtDNA copy number, absence of both Mfns, leading to loss of all outer membrane fusion, or absence of Opa1, leading to loss of all inner membrane fusion, dramatically results in depleted mtDNA (Figure 2.2A). To verify that this finding was due to the loss of the mitofusins in Mfn-null cells, and not due to other conditions that would cause different mtDNA:nDNA copy numbers in these cell lines, ds-Red (mock), Mfn1-myc, or Mfn2-myc was reintroduced into Mfn-null cells by retroviral infection and protein levels were monitored via Western at 2 and 4 weeks post-infection; mtDNA:nDNA was also tracked simultaneously.
This observation is confirmed by data from mouse studies by Hsiuchen Chen in our laboratory, which show that mutant mice homozygous for both Mfn1 and Mfn2 in skeletal muscle (MLC-Cre/dm) also exhibit a severe decrease in mtDNA: nDNA, a phenotype not seen. in Mfn1-/-,Mfn2+/- or Mfn1+/-,Mfn2-/- mice.
Discussion
The implications of dysfunction are serious, as mutations in the proteins involved in fusion and fission lead to neurodegenerative diseases (Alexander, Votruba et al. 2000, Delettre, Lenaers et al. 2000, Zuchner and Vance 2006). In addition to mutations in the machinery that directly affect the machinery, phenotypic perturbations in mitochondrial dynamics are correlated with many of the most common neurodegenerative diseases and myopathies in humans, such as AD and PD (Knott and Bossy-Wetzel 2008). Furthermore, depletion of mtDNA, which causes further mitochondrial dysfunction, has been implicated in several of these neurodegenerative diseases (Baron, Kudin et al. 2007), suggesting that depletion could contribute to the severity of the phenotype.
It is certainly possible that mtDNA depletion due to fusion ablation is the result of loss of exchange of mitochondrial content, thus leading to heterogeneous distribution of proteins required for genome propagation and a persistent depletion phenotype. Further studies addressing the interaction between proteins directly involved with mtDNA copy number, such as POLG, Twinkle, mtSSB, TFAM, POLRMT, and TFB2M, and fusion proteins will shed light on this intriguing effect.
Methods
Figure Legends
34;Mitofusins Mfn1 and Mfn2 coordinately regulate mitochondrial fusion and are essential for embryonic development." Hum Mol Genet.
Abstract
Introduction
Damage or depletion of mtDNA causes numerous inherited disorders, including Alpers' disease, ataxia neuropathy spectrum, and progressive external ophthalmoplegia (Suomalainen and Isohanni 2010, Stumpf, Saneto et al. 2013). Furthermore, homozygous knockout mice have no detectable levels of mtDNA and die during embryogenesis (Larsson, Wang et al. 1998), highlighting the importance of TFAM in maintaining mtDNA levels and in cellular and organismal viability. Although it is not well understood how TFAM packages mtDNA, it is known to bind non-specifically to DNA (Fisher, Parisi et al. 1989) and is estimated to be sufficiently abundant to completely cover the genome (Alam , Kanki et al. 2003, Ekstrand , Falkenberg et al. 2004, Kaufman, Durisic et al. 2007).
A TFAM variant that is deficient in transcriptional activation but competent in DNA binding is able to prevent mtDNA depletion (Kanki, Ohgaki et al. 2004). Canonical ChIP-seq was developed (Johnson, Mortazavi et al. 2007) in order to capture the in vivo interactions of transcription factors associated with the nuclear genome.
Results
Regarding the mitochondrial proteins, neither Mfn2 nor Opa1 was seen in the RIPA sample, although this could be attributed to the low relative amounts of these proteins in cells. Because partial copies of the mitochondrial genome are also present in the nuclear genome, not all reads derived from mtDNA can be uniquely mapped. We attribute this to the largely uniform coverage of the genome by TFAM, resulting in ubiquitous signaling.
Thus, there is no evidence for sequence-dependent DNA bending in the vicinity of the tip. On the other hand, TFAM is localized only upstream of OL in mouse mitochondria.
Discussion
Inspection of the ATP2A2 gene revealed no TFAM enrichment either in the promoter region or anywhere else in the vicinity of the gene (Figure 3.10E). We found no correlation between irregularities in TFAM signal distribution and characteristics of the mitochondrial genome such as GC content. The peak patterns mirrored that of the input to these regions, and no peaks were observed in seq-ChIP exhibiting canonical strand asymmetry in read distribution.
The localized nature of the ChIP signal at this site suggests a higher volume of TFAM. ChIP-chip on the yeast mitochondrial genome has shown that metabolic changes can lead to differential binding of the yeast TFAM homolog, Abf2p (Kucej, Kucejova et al. 2008).
Methods
Cells were harvested by 1000 rpm centrifugation for 10 min, then washed with ice-cold PBS and pelleted at 1000 rpm again. Crosslinking was performed by overnight 65C incubation and resulting DNA was purified via phenol-chloroform extraction followed by PCR purification (Qiagen #28104), substituting buffer PM for buffer PB in the classical protocol. Immunocytochemistry to visualize colocalization of mitochondrial nucleoids and TFAM was performed sequentially because both antibodies were raised in mice.
ChIP experiments and preparation of DNA for sequencing were performed according to standard procedures (Johnson, Mortazavi et al. 2007) with some modifications. Cells were fixed in 1% formaldehyde for 10 min at room temperature, harvested using a cell scraper, washed once in ice-cold PBS, and resuspended in RIPA buffer with protease inhibitor.
Figure Legends
ChIP-seq analysis of DNA submitted for sequencing shows that the majority was cut at ~222 bp. After removing multireads and alignments to the mitochondrial genome, peaks in the nuclear genome were called using MACS2. All ChIP-seq replicates resulted in at least 30% of reads mapping to the mitochondrial genome, much higher than the 0.4–1.9%. of reads mapped to mtDNA in the input datasets.
Schematic of TFAM ChIP-seq plus and minus strand and input read density signal over chrM. A, E) Annotation of protein-coding transcripts (green in forward/heavy strand, red in reverse/light strand), ribosomal RNA (blue) and tRNA (blue in forward/heavy strand, gray in reverse/light string). Note that the input signal is amplified 60-fold compared to the ChIP-seq signal to visualize coverage irregularities. from TFAM ChIP-seq largely follows that of the input, showing generalized connectivity across the mitochondrial genome.
Figures
The mammalian mitochondrial genome encodes 13 proteins, all of which are components of the electron transport chain, as well as 22 tRNAs and two rRNAs (Anderson, Bankier et al. 1981; Bibb, Van Etten et al. 1981). The mitochondrial localization of the estrogen receptor is also well established, for both the ERα and ERβ isoforms, and has also been suggested to bind to the D-loop (Monje and Boland 2001, Chen, Delannoy et al. 2004). Mitochondrial localization has also been reported for STAT1 and STAT5 (Boengler, Hilfiker-Kleiner et al. 2010, Chueh, Leong et al. 2010).
During our study of TFAM occupancy in mitochondrial and nuclear genomes (Wang, Marinov et al. 2013), we noticed that a number of nuclear transcription factors exhibit local enrichment in certain regions of the mitochondrial genome in ChIP-seq data (Figure 4.1). It is well known that the nuclear genome contains partial copies of the mitochondrial genome (NUMTs) (du Buy and Riley 1967, Hazkani-Covo, Zeller et al. 2010).
Loop ChIP-seq signal is likely artifactual
An experimental argument against unknown NUMTs comes from the strength of the ChIP-seq signal we see in the mitochondrial genome. Human transcription factors with canonical ChIP-seq peaks (representing typical strand asymmetry in the read distribution around the putative binding site) outside the D-loop. The unique mappability trace for the mitochondrial genome is shown in red in the outer trace (see Methods for details).
Mouse transcription factors with canonical ChIP-seq peaks (representing typical strand asymmetry in read distribution around the putative binding site) outside the D-loop. The inner circle shows the motif occurrences in the mitochondrial genome for each factor as black.
