Ubiquitin system inside the mitochondria!

Mitochondria are the cellular organelles responsible for converting food into the available energy that fuels our cells metabolic activity (ATP). A decline in the mitochondrial functional capacity triggers the aging proccess (my opinion) and therefore its maintenance is of uttermost importance. Indeed, human maximal endurence and heart rate capacities decline ~10% and ~4.5% per decade respectively.

How are the mitochondria maintained? The current view is that the mitochondria are "offsprings" of bacteria, and are maintained by a repertoire of bacterial-like proteins. I compared the maintenance mechanisms of mitochondria with those of bacteria and found that the mitochondria lack several universal bacterial systems. Specifically, these systems are the transfer-messenger RNA (tmRNA) that recycles stalled ribosomes and tags unfinished proteins for degradation with SsrA peptide, the N-end rule ClpS adaptor protein its downstream protease complex ClpAP, and the "proteasome ancestor" ATP-dependent protease complex HslUV. *In some low eukaryotes, the mitochondria still contain the tmRNA (jakobid protists) and HslUV (Plasmodium species).

The ClpAP and HslUV protease complexes resemble the eukaryotic proteasome. Their functions to degrade N-end rule substrates (ClpAP) and to degrade unfolded proteins following heat shock (HslUV) is, in eukaryotes, carried out by the ubiquitin proteasome system (UPS). Moreover, the role of the bacterial tmRNA and ClpS is preformed by ubiquitin ligases in eukaryots (LTN1 and HEL2, and UBR1 and UBR2 respectively). These evidence suggest that the eukaryotic UPS replaced the ancient mitochondrial bacterial-like protein quality control/assurence systems. Further support for this suggestion comes from a different prespective. The UPS is known to be essential for DNA repair in the eukaryotic nucleus (e.g., DNA double strand break repair by Ku70/Ku80). How is this DNA repair supposedly done in the mitochondria if the UPS system is not active there? Time has come to re-examine the received view that the UPS is absent inside the mitochondria.

Figure 1. Mitochondria convert food into usable enrgy (ATP) and produce harmful radical oxygen species (ROS) as a byproduct. In endotherms, the mitochondria also generate body heat through the proccess of uncoupling. At the same time, mitochondria have to maintain a functional proteome and repair their mitochondrial DNA (mtDNA). Do they really perform these tasks like bacteria do or did they evolve to recruit the powerful UPS for their needs?

Mitochondrial purification workflow

The mitochondrial purification procedure that I followed is described Here. Shortly, I treated yeast cells with zymolyase and lyzed them with homogenizer. I added protease inhibitors and did a series of centrifugations of varying speeds to pellet different parts of the cells. After the ordinary centrifugations, I did discontinous sucrose gradient and used ultracentrifuge to further isolate "pure" mitochondria (figure 2).

Figure 2. After differential centrifugation, I used discontinous sucrose density gradient to further purify yeast mitochondria (Link to the protocol). The Pure mitochondria appear as brown membrane at the interface of the 60% and 32% sucrose. To extract the pure mitochondria without contaminating it, I used a syringe and cut through the plastic tube.

The standard "pure" mitochondria were not enough for me. I wanted to have only the mitochodnrial matrix proteins and the inner membrane proteins that face the matrix. The mitochondrial outer membrane and the inter-membrane space (IMS) proteins were contaminations for my purpose of identifying mitochondrial matrix ubiquitin conjugates and UPS machinery. Therefore, I put the mitochondria in hypotonic buffer (EM buffer) that causes them to swell and generates mitoplasts (mitochondria without the outer membrane and IMS proteins). Then, I gently used dounce homogenizer to dislodge the outer membranes from the mitoplasts. By centrifugation at 12,000g for 10 min I pelleted the mitoplasts. From the supernatant of the mitoplasts, I pelleted and collected the outer membrane by using ultracentrifuge. To cut the mitoplasts inner membrane proteins that have outward orientation, I put the mitoplasts in SEM buffer containing trypsin (0.05 mg/ml) on ice for 30 minuts. I stopped the reaction with trypsin inhibitors and washed the mitoplasts several times. Finally, I separated the mitoplasts into inner membrane and matrix fractions by rupting them in pure water combined with insulin syringe (29 gauge needle).

* Notes *

Working with mitochondria is difficult. The mitochondrial volume is ~10% of the cells and during the purification process a lot of mitochondria get damaged or are lost. Therefore, very large quantities of cells are need to start with. Then, the classical purification process of serial differential centrifugations is several hours long meaning that until the mitochondria are purified, much has happened in the samples, even if on ice and with protease inhibitors. Furthermore, there is always a question mark over the purity of the samples, especially if the downstream analysis is a sensitive mass spectrometry. From my experience, mass spectrometry detected in my mitochondrial samples also proteins that are not reported to be mitochondrial. After considering the pros and cons, I decided to work with the yeast model organism. The major advantage of using yeast compared to mammalian cells is that its samples complexity is much lower. Yeast has only ~6000 protein-coding genes and usually these genes dont encode isoforms. This fact, I reasoned, will ease the detection of low abundance proteins by the mass spectrmetry instrument. More advantages are that yeast is cheaper to work with and that I could easily grow overnight as many cells as I wanted for starting material. The disadvantage is of course that I am much more interested in humans.

Biochemical evidence for ubiquitin-conjugates in the mitochondria

I Used ultracentrifuge to pellet the mitochondrial outer membrane from the soluble fraction after generating the mitoplasts and did a combined analysis of western blot and mass-spectrometry which resulted in the absolute identification of ubiquitinated mitochondrial matrix protein SSC1 (the yeast homolog of the human protein HSPA9 also known as mortalin; figure 3).

Figure 3. Analyses of several mitochondrial fractions: mitoplasts, soluble fraction (matrix), outer membranes, and outer membrane + soluble (before their separation). The fractions were size separated as duplicates by a 10 lanes SDS-PAGE. The SDS-PAGE was then cut into two halves, with one half used for western blot with anti-ubiquitin antibodies (gel on the top right corner led to the membrane), and the second half was stained and later used for mass-spectrometry (gel at the top left corner). Because the two halves were identical, I could overlay the stained gel on top of the blotted membrane and cut precisely the place with the strongest ubiquitin bands. The mass-spectrometry results detected the ubiquitin remnant motif (K-ε-GG) on multiple peptides from the mitochondrial matrix protein SSC1 (in my soluble frction) and on multiple peptides from ATP2 (in my outer membrane fraction). As control, the mass-spectrometry did not identify ubiquitinated SSC1 peptides in the corresponding gel slice from the outer membrane fraction (and also didnt find the ubiquitinated ATP2 in the soluble fraction).

The identification of SSC1 and ATP2 as the ubiquitinated proteins is certain because several of their peptides were identified with ubiquitin remnant motif (K-ε-GG; Here is the mass-spec identification of the GlyGly-SSC1 peptides). Added to that, SSC1 is known to reside in the mitochondrial matrix (ATP2 is part of the mitochondrial ATP synthase complex with its precise localization unknown). It is interesting that the matrix protein SSC1 appears to form one strong ubiquitinated band rather than a polyubiquitin ladder. This suggests that the protein is only monubiquitinated, or perhaps to some extent monoubiquitinated at multiple sites. However it needs to be in mind that by the time I purified the mitochondria and reached western blot (over 6 hours after the yeast cells lysis), polyubiquitinated proteins might have been degraded or lost.

I did several more experiments and found that more mitochondrial proteins are only monoubiquitinated. This time, I did not purify the mitochodnria but immediately pulled down the mitochondrial matrix proteins which I knew were ubiquitinated - ADH3 and HSP78. For the pulldown I transfected my yeast strains with His-tagged ubiquitin and HBT-tagged ubiquitin. The HBT tag is large and HBT-ubiquitin conjugation should shift proteins molecular weight to larger extent which is desirable for HSP78.

Figure 4. Yeast expressing TAP-tagged ADH3 and TAP tagged HSP78 were transfected with His-ubiquitin or HBT-ubiquitin. The yeast cells were lysed with glass beads in the presence of 8M urea (denaturating buffer) and from the lysate I pulled down the tagged-ubiquitin. To the left, pull down of his-ubiquitin and then blot with anti TAP-ADH3 antibody revealed a ubiquitinated form of ADH3. Similarly, pull down of HBT-ubiquitin and then blot with anti TAP-HSP78 antibody revealed ubiquitinated forms of HSP78. To the right, control for the success of the tagged ubiquitin pulldown. The his-ubiquitin pulldown for HSP78 failed. This explains why no ubiquitinated HSP78 were identified in the "HSP78: his-ub" lane (they are the same sample, just run in different SDS-pages and blotted with different antibodies).

In summary: the mitochondria have changed since their days as bacteria. They have lost several bacterial quality control and assurence mechanisms and gained access to the ubiquitin proteasome system!

* Further information: more biochemical information about the ubiquitin system in the mitochondria, including the identification of the ubiquitin ligase DMA1 in the mitochondria can be found in my paper Here. I also did comprehensive bioinformatics analysis of the link between the UPS and the mitochondria, including UPS components found in the mitochondria or predicted to contain mitochondrial transit signal Here. A PowerPoint .pptx file that I presented in 2017 at the Max Planck institute in Munich is available here.