Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2019 Jan 2;116(1):303-312.
doi: 10.1073/pnas.1816656115. Epub 2018 Dec 12.

MITO-Tag Mice Enable Rapid Isolation and Multimodal Profiling of Mitochondria From Specific Cell Types in Vivo

Affiliations
Free PMC article

MITO-Tag Mice Enable Rapid Isolation and Multimodal Profiling of Mitochondria From Specific Cell Types in Vivo

Erol C Bayraktar et al. Proc Natl Acad Sci U S A. .
Free PMC article

Abstract

Mitochondria are metabolic organelles that are essential for mammalian life, but the dynamics of mitochondrial metabolism within mammalian tissues in vivo remains incompletely understood. While whole-tissue metabolite profiling has been useful for studying metabolism in vivo, such an approach lacks resolution at the cellular and subcellular level. In vivo methods for interrogating organellar metabolites in specific cell types within mammalian tissues have been limited. To address this, we built on prior work in which we exploited a mitochondrially localized 3XHA epitope tag (MITO-Tag) for the fast isolation of mitochondria from cultured cells to generate MITO-Tag Mice. Affording spatiotemporal control over MITO-Tag expression, these transgenic animals enable the rapid, cell-type-specific immunoisolation of mitochondria from tissues, which we verified using a combination of proteomic and metabolomic approaches. Using MITO-Tag Mice and targeted and untargeted metabolite profiling, we identified changes during fasted and refed conditions in a diverse array of mitochondrial metabolites in hepatocytes and found metabolites that behaved differently at the mitochondrial versus whole-tissue level. MITO-Tag Mice should have utility for studying mitochondrial physiology, and our strategy should be generally applicable for studying other mammalian organelles in specific cell types in vivo.

Keywords: MITO-Tag Mice; lipidomics; metabolomics; mitochondria; proteomics.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.
Fig. 1.
Characterization of MITO-Tag Mice and associated in vivo workflow. (A) Schematic demonstrating the design of the MITO-Tag Mice and how they can be utilized for rapid, cell-type–specific isolation and multimodal analysis of mitochondria in vivo. MITO-Tag Mice contain the 3XHA-EGFP-OMP25 (HA-MITO) gene knocked into the Rosa26 locus. The expression of this gene is dependent on the presence of Cre recombinase due to an upstream LSL cassette. Mating MITO-Tag Mice with mice expressing Cre under the control of certain promoters, such as the Albumin promoter as in this study, allows for cell-type-specific expression of the HA-MITO gene and, consequently, allows for the rapid, cell-type-specific isolation of mitochondria in vivo (mitochondrial isolation takes ∼10 min once liver tissue has been homogenized). Because KPBS, the mitochondrial isolation buffer utilized in our workflow, is compatible with mass spectrometric analyses, the isolated mitochondria can subsequently be interrogated with either proteomics or metabolomics. The actual immunopurification workflow used here overall is similar to that used in our prior work (2, 4); however, it is important to note that the material inputs in this work are liver tissue from Rosa26LSL-3XHA-EGFP-OMP25/+ mice (Control-MITO mice) and Alb-Cre+/−, Rosa26LSL-3XHA-EGFP-OMP25/+ mice (HA-MITO mice) and that no measurements of mitochondrial volume were done in this study because we performed relative rather than absolute quantification of metabolites. Arrowheads indicate loxP sites. (B) Genomic PCR analysis of liver tissue taken from mice with the indicated genotypes. Upper (2,675 bp) and Lower (169 bp) bands indicate the presence of the desired knock-in at the Rosa26 locus and successful Cre-mediated excision of the floxed region, respectively. The presence of nonexcised versions of the knock-in cassette in the HA-MITO liver tissue could be secondary to heterogeneity of Albumin promoter activity across hepatocytes and there being a significant number of cells in liver that are not hepatocytes and thus not expected to express Cre recombinase (35). (C) Immunoblot analysis of liver tissue taken from mice with the indicated genotypes. ACTB was used as a loading control. Control-MITO mice (n = 3); HA-MITO mice (n = 3). (D) Immunofluorescence of liver tissue taken from mice with the indicated genotypes. HA (green), COX4l1 (red), and Hoescht (blue) signals are shown. (Scale bars in the large and small images represent 100 and 50 μm, respectively.) (E) Immunoblot analysis of liver tissue and the anti-HA immunoprecipitates (anti-HA IP) from mice with the indicated genotypes. The line indicates that two parts of the same membrane have been brought together. The names of the protein markers used and the corresponding subcellular compartments appear to the left and right of the immunoblots, respectively. ER, endoplasmic reticulum; Golgi, Golgi complex; IMS, mitochondrial intermembrane space; Matrix, mitochondrial matrix; OMM, outer mitochondrial membrane. The enrichment in the HA-MITO IP, relative to the HA-MITO liver tissue, of the markers VDAC, PEX13, and CALR was quantified and is shown in the table below. All mice used for these experiments were fed ad libitum. For AE, the following abbreviations are used: Control-MITO, Control-MITO mice; HA-MITO, HA-MITO mice.
Fig. 2.
Fig. 2.
Mitochondria isolated with MITO-Tag Mice can be used for multiple downstream analyses and possess both protein and metabolite features characteristic of mitochondria. (A) Flowchart of how proteomics analysis was done. In brief, the 1,204 proteins in our primary proteomics data were filtered based on a stringent enrichment criterion of having a mean pseudonumbered protein signal in the HA-MITO IPs that was at least five times greater than that in the Control-MITO IPs. Of these 511 proteins with signal considerably above that of the Control-MITO IPs (i.e., the background), 75.9% of them were found in MitoCarta2.0 by cross-referencing the NCBI gene IDs, with the respective numerical breakdown indicated in the pie chart wedges (20). Control-MITO mice (n = 3); HA-MITO mice (n = 3). Mice used for these experiments were fed ad libitum (Dataset S1). (B) Lipidomic analysis of IPs for cardiolipin species. Negative lipid ion peak area signals for various cardiolipins (CLs) were used to generate this graph. Control-MITO IPs (n = 4 mice); HA-MITO IPs (n = 4 mice). Mice were fasted for 12 h and then refed for 4 h. Data are shown as means with SDs. All CLs presented here were categorized as grade A and satisfy the criteria of being considered mitochondrial, as defined in Materials and Methods (see also Dataset S2). (C) Targeted metabolomic analysis of IPs for various polar metabolites. Shown are metabolite peak area ratios, which are calculated by dividing the metabolite peak areas by internal standard peak areas, for the Control-MITO and HA-MITO IPs. NAD, NADH, NADP, NADPH, CoA, alpha-ketoglutarate, ATP, 3-hydroxybutyrate, and saccharopine are known mitochondrial metabolites, whereas cystine (lysosome) and sedoheptulose 7-phosphate (cytosol) are markers for extramitochondrial compartments. Control-MITO IPs (n = 4 mice); HA-MITO IPs (n = 4 mice). Mice were fasted for 12 h and then refed for 4 h. Data are shown as means with SDs. With the exception of cystine and sedoheptulose 7-phosphate, all metabolites shown here meet the criteria of being considered mitochondrial, as defined in Materials and Methods (see also Dataset S3).
Fig. 3.
Fig. 3.
In vivo dynamics of hepatocyte mitochondrial metabolites during fasting and refeeding. (A) Schematic demonstrating general design of fasting and refeeding experiment. In brief, Control-MITO and HA-MITO mice were fasted overnight for 12 h and then either fasted or refed for another 4 h before liver tissue was harvested for mitochondrial or whole-tissue polar metabolomics. (B) Immunoblot analysis of liver tissue taken from Control-MITO and HA-MITO mice during fasted and refed conditions. Protein names appear to the left of the membrane. P-S6, phospho-S6 [Ser240/244]. GAPDH was used as a loading control. (C) Heat map of changes in metabolites as assessed at the mitochondrial and whole-tissue level during fasted and refed conditions. The data are presented as the log2-transformed mean fold difference (fast/refed). For a metabolite to be included in this heat map, it had to be considered mitochondrial in at least the fasted or refed state. See Materials and Methods for criteria for being considered mitochondrial. See also Dataset S3 for more detail. (D) Comparison of select metabolites during fasted and refed conditions. Metabolite peak area ratios, which are calculated by dividing the metabolite peak areas by internal standard peak areas, are used to generate these graphs. See also Dataset S3. Data are presented as means with SDs. *P < 0.05 and **P < 0.01 as determined by unpaired, parametric, two-tailed Student’s t test. For AD, fasted Control-MITO mice (n = 4), fasted HA-MITO mice (n = 5), refed Control-MITO mice (n = 4), and refed HA-MITO mice (n = 4).
Fig. 4.
Fig. 4.
Interrogation of mitochondrial metabolites during fasting and refeeding using untargeted metabolomics. (A) Scatter plot of compounds in the HA-MITO IPs (i.e., isolated mitochondria) as generated from the untargeted analysis. The data presented here are the log2-transformed fold differences (median fasted HA-MITO IP signal/median refed HA-MITO IP signal). Ad hoc adjusted P values (Benjamini–Hochberg method) are presented here, and the line indicates where the adjusted P value = 0.05. All points above the line are considered significant. Compounds that were subsequently validated using authentic chemical standards are highlighted in red and indicated by their metabolite name. Note that many compounds appear along the x axis as a result of their adjusted P values. See also Dataset S4 for more detail. (B) Metabolomic analysis of IPs for O-adipoylcarnitine. Shown are metabolite peak area ratios, which are calculated by dividing the metabolite peak areas by internal standard peak areas, for the Control-MITO and HA-MITO IPs during fasted and refed conditions. O-adipoylcarnitine meets the criteria of being mitochondrial in both fasted and refed conditions, as described in Materials and Methods. Data are shown as means with SDs. See also Dataset S4. (C) Metabolomic interrogation of O-succinylcarnitine and 4-guanidinobutyric acid during fasted and refed conditions. Metabolite peak area ratios, which are calculated by dividing the metabolite peak areas by internal standard peak areas, were used to generate these graphs. See also Dataset S4. Data are shown as means with SDs. ***P < 0.001 as determined by unpaired, parametric, two-tailed Student’s t test. (D) Scatter plot of compounds from untargeted analysis when assessed on the mitochondrial versus whole-tissue level. The data presented here are the log2-transformed fold differences for the HA-MITO IP samples (median fasted HA-MITO IP signal/median refed HA-MITO IP signal) and the HA-MITO whole-tissue samples (median fasted HA-MITO whole-tissue signal/median refed HA-MITO whole-tissue signal). Methionine and alanine are highlighted in red as the behavior of these two metabolites was confirmed in our targeted analysis. See Datasets S3 and S4 for more detail. (E) Metabolomic interrogation of methionine and alanine during fasted and refed conditions. Metabolite peak area ratios, which are calculated by dividing the metabolite peak areas by internal standard peak areas, are used to generate these graphs. See also Dataset S3. Data are shown as means with SDs. *P < 0.05 as determined by unpaired, parametric, two-tailed Student’s t test. n.s., not statistically significant. For AE, fasted Control-MITO mice (n = 4), fasted HA-MITO mice (n = 5), refed Control-MITO mice (n = 4), and refed HA-MITO mice (n = 4).

Comment in

  • Probing mitochondrial metabolism in vivo.
    McElroy GS, Chandel NS. McElroy GS, et al. Proc Natl Acad Sci U S A. 2019 Jan 2;116(1):20-22. doi: 10.1073/pnas.1819614116. Epub 2018 Dec 18. Proc Natl Acad Sci U S A. 2019. PMID: 30563855 Free PMC article. No abstract available.

Similar articles

See all similar articles

Cited by 12 articles

See all "Cited by" articles

Publication types

LinkOut - more resources

Feedback