How paternal mitochondria are destroyed in an embryo

Molecular cytogenetics is hardly my field, so this paper was a bit hard for me, but the results were so interesting that I’ll do my best to present it. The paper is by Quingua Zhou et al. and was just published in the early, non-print edition of Science. (reference and free download below). It’s about what happens to mitochondria in embryos. There’s another paper in the same issue that shows the same thing, by S. Al Rawi et al, (reference below), but I didn’t read that one. There’s also a perspective on both papers by Beth Levine and Zvulun Elazar, which I also haven’t read (I’m busy!)

As you may remember from your biology courses, mitochondria are “organelles” in the cell that function in respiration and metabolism (the generation of energy) and in “apoptosis,” or programmed cell death. They contain a circular DNA molecule that produces gene products (proteins) as well as transfer RNAs—the small molecules that lock onto amino acids and help assemble them into proteins.  One of the big findings of my lifetime, suggested by Lynn Margulis among others, is that mitochondria are actually the evolutionary remnants of bacteria-like prokaryotes that were ingested (probably by another prokaryote), forming a mutualistic relationship that led to the first “true” cell: a eukaryote.

The interesting thing about mitochondria is that in nearly all sexually reproducing species they’re inherited purely maternally:  while both mother and father have them, only the mother’s mitochondria are passed to its adult offspring. This is why, for instance, we’re able to identify a “mitochondrial Eve”: the ancestral woman from whom all the mitochondrial DNA of living humans is descended. (Of course the rest of our DNA, which is far greater in extent, comes from a diversity of other “Eves,” so this doesn’t support the Biblical narrative in any way.)

Although a late-stage developing embryo contains only the mitochondria from the mother, sperm that penetrate the egg do have paternal mitochondria, and, at least in the “model organism” studied (the tiny worm Caenorhabditis elegans), they are present in early embryos.  But then something happens—something that is amazing.  You can see it happening in the photograph below.

The first photo, “A,” shows normal mitochondria in C. elegans sperm. B-C are the paternal mitochondria, derived from sperm, in embryos. Shortly after the embryo is formed, abnormal aggregates begin to form in those mitochondria (blue arrows; scale bars are 300 nanometers, or 0.3 microns, long). The cristae, or internal mitochondrial membranes, begin to break down (B & C). Then an “autophagosome” (yellow arrow) begins to form around the paternal mitochondrion (C); this is a membrane that the cell puts around structures that it intends to destroy. The contents of the autophagosome are subsequently degraded by special organelles called lysosomes. In photo “D”, the mitochondrion is on its way out; its membrane has broken down and it will soon degenerate. The paternal mitochondria are all destroyed in this way. Only the maternal ones remain as the embryo develops into an adult worm.

Screen Shot 2016-06-24 at 9.45.13 AM

There are four questions here:

  • Why does the embryo destroy the paternal but not maternal DNA?  There are evolutionary theories for this based on preventing the spread of selfish DNA in organelles (see here for one example), but these are only untested speculations. But there is almost certainly some evolutionary reason for it, since uniparental inheritance of mitochondria is so common.
  • How does the embryo distinguish between paternal and maternal mitochondria, and destroy only the former? Short answer: we don’t know. But I bet we will within a few years.
  • What are the genes involved in the destruction of paternal mitochondria? The authors of the Zhou et al. paper answered that one using clever techniques. By genetically removing RNA of genes involved in mitochondrial membrane production, they showed that one gene, cps-6, is probably involved. When the RNA product of that gene is deleted from the embryo, the paternal mitochondria persist until the late embryo stage, something that doesn’t happen with the product of any other gene. It turns out that the cps-6 gene is a nuclease (it destroys chains of nucleotides, like RNA and DNA) and is imported into paternal mitochondria by the cell, where it breaks them down.
  • What happens to the cell if paternal mitochondria aren’t destroyed? If you inactivate the cps-6 gene so that paternal mitochondria remain in the embryo intact, there’s elevated mortality of those embryos. It’s not total, but increases by 5.9%—a serious loss of offspring in evolutionary terms. This suggests that it’s to the cell’s advantage to destroy paternal mitochondria, and directs us again to an evolutionary explanation, one still not fully understood.


Zhou, Q., H. Li, H. Li, A. Nakagawa, J. L. J. Lin, E.-S. Lee, B. L. Harry, R. R. Skeen-Gaar, Y. Suehiro, D. William, S. Mitani, H. S. Yuan, B.-H. Kang, and D. Xue. 2016. Mitochondrial endonuclease G mediates breakdown of paternal mitochondria upon fertilization. Science, published online: DOI: 10.1126/science.aaf4777

Al Rawi, S., Louvet-Vallee, S., Djeddi, A., Sachse, M., Culetto, E., Hajjar, C., Boyd, L., Legouis, R., & Galy, V. (2011). Postfertilization Autophagy of Sperm Organelles Prevents Paternal Mitochondrial DNA Transmission Science, 334 (6059), 1144-1147 DOI:10.1126/science.1211878


  1. Posted June 24, 2016 at 11:13 am | Permalink

    Has it been definitively shown that male mDNA is, in fact, destroyed? An alternative theory is that the male contribution is so small that it it very difficult or impossible to identify. Females (humans) contribute upwards to 100k mitochondria with the male only adding a hundred or so. Still it seems that with current technology this should be far and away enough to recognize.

    I’m more partial to the idea that upon fertilization the ovum goes on a mad hunt for anything foreign. This reproduction thing, after all, is what life and evolution is all about. Mess it up here and the egg is in trouble.

    Thanks for posting – it’s a fascinating area of research.

  2. Posted June 24, 2016 at 11:15 am | Permalink

    Reblogged this on My Selfish Gene and commented:
    Go to the link below for Professor Ceiling Cat’s take on this fascinating bit of research. Go here to see my previous post on the topic.


  3. GBJames
    Posted June 24, 2016 at 11:23 am | Permalink

    Really interesting.

  4. Dee
    Posted June 24, 2016 at 11:30 am | Permalink

    I had no idea this happened. Damn cool stuff.

  5. Posted June 24, 2016 at 11:38 am | Permalink

    Wow – I think if ID and creationist Christians made it a habit of reading this sort of science, as I did before I deconverted, it would lead many more people to reason.

    Although this did make me think of the concept of Jewish status being passed down through the mother – Mitochondrial in nature perhaps? 😉

    • Posted June 24, 2016 at 3:45 pm | Permalink

      Mama’s baby, papa’s maybe. I think that’s the basis for the rule of maternal descent.

      • Posted June 24, 2016 at 3:48 pm | Permalink

        Oh, it is for sure, but for it still came to mind reading this 😉

  6. Reggie Cormack
    Posted June 24, 2016 at 11:43 am | Permalink

    Thank you for that, P.C.C.(E). Great analysis and well explained to the non-geneticist.

  7. ThyroidPlanet
    Posted June 24, 2016 at 11:44 am | Permalink

    I never knew! Interesting!

  8. Posted June 24, 2016 at 11:45 am | Permalink

    I would like to say I have wondered about that, but frankly it never occurred to me. But now that it comes up, it is extremely interesting. Thanks for posting a comprehensible digest of the paper.

  9. aldoleopold
    Posted June 24, 2016 at 11:56 am | Permalink

    Thank you for calling our attention to this new literature!! All questions that I had for my biology/evolutionary professors in undergrad – was always a primary interest. Great summary of a dense paper!

  10. Posted June 24, 2016 at 11:59 am | Permalink

    Given that Murphy’s law applies to organisms, what would happen if the *paternal* mitochondria remained *without* the maternal? (The last question in the post addresses the both, AFAIK.)

  11. Kevin
    Posted June 24, 2016 at 1:00 pm | Permalink

    Wow. Excellent article and very international piece: Boulder, Japan, China, Florida.

    There are probably simple biochemical markers that cause the cell to terminate one of the mitochondria. Some maternal DNA supplied information somewhere to basically get rid of the paternal mitochondria. It probably happened this way ‘once’ and then happened again zillion times.

    Can cells function with more than two mitochondria?

    • Posted June 24, 2016 at 1:25 pm | Permalink

      The number of mitochondria in a cell is a function of how much metabolic activity is going on – but cells can have a very large number of these organelles. And they are not necessarily even clonal. They are just not paternal in origin. (As I note below there are also multiple mechanisms of destroying the paternal ones – so it’s obviously important – but no idea why, not my field!)

      • Posted June 25, 2016 at 12:13 am | Permalink

        I just read the number of mitochondria — I suppose in muscles — can be increased with exercise. How does that work? I know how cells split by mitosis, but how does a mitochondrian get constructed? It’s more than just a protein.

        • John Harshman
          Posted June 25, 2016 at 4:17 pm | Permalink

          Think of a mitochondrion as a little cell within the cell. It assembles by splitting, just like any other cell. The mitochondrial genome replicates just like a bacterial genome does, and the various proteins and other molecules are produced by the usual means. The only weird thing about it is that most of the mitochondrial proteins are coded for by the nuclear genome, translated by cytoplasmic ribosomes, and transported to the mitochondrion.

  12. Posted June 24, 2016 at 1:12 pm | Permalink

    Interesting article.

  13. Simon Hayward
    Posted June 24, 2016 at 1:22 pm | Permalink

    I wonder how many mechanisms exist. I read a paper a couple of years ago on paternal mitochondrial loss in Drosophila – and that involved destruction of the mitochondrial DNA by an endonuclease. So they just wear out and don’t replace themselves.

    DeLuca and Farrell Developmental Cell 2012, 22, 660–668

    This suggests that there are multiple possible options to destroy paternal mitochondria. Another question, that you note we can’t answer yet is, why bother? Since multiple mechanisms exist it seems reasonable to assume there is a reason for this sort of selection.

  14. mikeyc
    Posted June 24, 2016 at 1:46 pm | Permalink

    Inheritance of maternal mitochondria is almost but not quite universal. Paternal mitochondrial inheritance is very rare, but they have been seen.

    There are reports of paternal mitochondria being passed down in mice (see Gyllensten et al Nature 1991;352:255-257). It has even happened in humans; a report about a man suffering from mitochondrial myopathy (poor fellow as so afflicted that he suffers from severe exercise intolerance and could not run even a few steps). Because of the affliction (due to a 2-bp mtDNA deletion in the ND2 gene) they looked at his mitochondrial genome and discovered the mutation was of paternal origin. See; Schwartz and Vissing, N Engl J Med 2002; 347:576-580

    • John Harshman
      Posted June 25, 2016 at 4:20 pm | Permalink

      In mussels, females always inherit their mother’s mitochondria, but males always inherit their father’s. How weird is that?

      There is occasional leakage between sexes, so that male and female mitochondria within species tend to be more closely related to each other than female mitochondria between species or male mitochondria between species.

  15. Billy Bl.
    Posted June 24, 2016 at 2:26 pm | Permalink

    Nick Lane has suggested that uniparental inheritance would optimise the the association between mitochondrial and nuclear genes for optimising respiration. Maternal and paternal mitochrondria have different genomes, but the nuclear genomes in all cells of a body are clonal.

    • Brad
      Posted June 25, 2016 at 9:02 am | Permalink

      But the mitochondria aren’t necessarily genetically distinct. For example, consider female A who has daughters B and C. B has daughter D who has daughter E, while daughter C has daughter F who has daughter G who has son H. The offspring of E and H will inherit genetically identical mitochondria from both maternal and paternal gametes, discounting mutations. That’s not even particularly close inbreeding in humans, and probably quite common among herd animals and the like.
      I am reluctant to use the E word, but my money is on something epigenetic.
      Also, Jerry, thanks for another fascinating science post. I do read them even if I seldom comment.

  16. Torbjörn Larsson
    Posted June 24, 2016 at 2:34 pm | Permalink

    Biology strikes again! Since sex is ubiquitous in eukaryotes, I would think this trait evolved in the stem lineage. “Given its ubiquity and shared core features, sex is thought to have arisen once in the last common ancestor to all eukaryotes.” [ ]

    One of the big findings of my lifetime, suggested by Lynn Margulis among others, is that mitochondria are actually the evolutionary remnants of bacteria-like prokaryotes that were ingested (probably by another prokaryote), forming a mutualistic relationship that led to the first “true” cell: a eukaryote.

    One of the big findings of *my* lifetime, suggested by James Lake in 1984 apparently, is that eukaryotes clades with archaea [ ], first tested by Anja Spang et al at my alma mater with gene phylogenies and then by Laura Hug et al by concatenating a set of ribosomal proteins into the latest massive tree that Jerry has described [ ].

    I have very vague notions of these things of course, but the pathway to multicellular complex life interests me. (As part of astrobiology.)

    It is therefore interesting that these advances has allowed a mechanistic proposal of a “long, slow dance” wherein the archaeon retooled both the bacterial motor protein dynamin used for cell division and its own complex shape regulation eukaryote like tool set to allow for membrane compartmentalization and fusion. “The putative archaeal host existed in a stable symbiotic relationship with one or more bacterial species, with the capacity for both gene and lipid exchange between species. The archaeal host, with a large complement of cytoskeletal genes and regulatory GTPases, was probably capable of complex shape regulation.” One can even – I think – understand why the archaeon changed its outer membrane in order to achieve detailed membrane regulation. [ ; the image elaborate on membrane regulation.]

    The specific proposal is in turn testable. “The researchers predict that, when Loki [now closer Thor, see Hug et al] is finally isolated or cultured, “it will look more like an archaeon than a proto-eukaryote and will not have internal compartments or a vesicle-trafficking network.” But its morphology and/or cell cycle might have complexities more often associated with eukaryotes.”

    And after dancing, sex.

  17. Eduardo
    Posted June 24, 2016 at 2:35 pm | Permalink

    Very nice!

  18. Merilee
    Posted June 24, 2016 at 3:12 pm | Permalink


  19. Posted June 24, 2016 at 4:02 pm | Permalink

    Liked this.


  20. Lars
    Posted June 24, 2016 at 4:15 pm | Permalink

    Thanks for posting this useful summary – handy for when I teach human biology again.

  21. John Frum
    Posted June 24, 2016 at 6:44 pm | Permalink

    This begs the question then that why would sperm have mitochondria?
    If it is always destroyed there is no selection pressure so you would think that mutations would long ago have changed the production of sperm to not include it.

    • Posted June 24, 2016 at 6:58 pm | Permalink

      To power them, like all other cells.

    • Peter Lund
      Posted June 25, 2016 at 12:45 pm | Permalink

      Sperm cells move a lot. It takes energy to flail that tail. “The early sperm gets the cell” or something like that.

  22. Ken Pidcock
    Posted June 24, 2016 at 6:51 pm | Permalink

    Thanks for this. It’s always inspiring to see some mechanistic insight into what we still can’t comprehend.

    A bit pedantic, but please don’t use “prokaryote” as a noun. It is phylogenetically meaningless.

    • Posted June 24, 2016 at 9:43 pm | Permalink

      I can’t read the article without paying for it. Prokaryote is defined as “a microscopic single-celled organism that has neither a distinct nucleus with a membrane nor other specialized organelles.” Those types of cells exist, so why not have a name for them?

      • Ken Pidcock
        Posted June 25, 2016 at 9:23 am | Permalink

        This is a nice extract from that essay. (Although “…as shown in the figure overleaf.” will make no sense!)

        The campaign is really Norman Pace’s tribute to Carl Woese, whom he believes should be better known.

        • Gregory Kusnick
          Posted June 25, 2016 at 11:13 am | Permalink

          If we’re not allowed to have names for paraphyletic groups, how are we going to talk sensibly about fish, reptiles, and a host of others?

          And so what if “prokaryote” is negatively defined? So is “atheist”. That doesn’t make it meaningless.

  23. Posted June 24, 2016 at 7:00 pm | Permalink

    Note that the same applies to the other organelle type descended from free-living bacteria: plants’ chloroplasts. They too are nearly always inherited from the mother, although again there are a few exceptions. At any rate this independent occurrence of uniparental inheritance of an organelle with its own genome adds weight to the idea that there is some evolutionary reason behind it.

    • Gregory Kusnick
      Posted June 24, 2016 at 7:25 pm | Permalink

      OK, I’ll bite. Why do pollen grains need chloroplasts?

      • Posted June 24, 2016 at 10:06 pm | Permalink

        I don’t know if they need them, but my hunch is that having the organelles tag along is just standard procedure for eukaryote cells. When not in use for photosynthesis, chloroplasts lose colour and turn into different shapes with their own names (see the nice diagram close to the top). I assume pollen would contain proplastids or etioplasts…?

        An example of a group where the chloroplast is usually inherited from the father are the conifers, by the way, so there the organelles must definitely travel down the pollen tube. I assume the zygote or embryo would have to get rid of the ones from the mother, perhaps in a similar way as described for mitochondria in this post.

  24. EvolvedDutchie
    Posted June 25, 2016 at 10:27 am | Permalink

    I’m in the middle of exam weeks and essay deadlines, so I don’t have a lot of time to comment. 😦 But I do want to say I appreciate the science posts very, very much.

  25. Posted June 26, 2016 at 9:11 am | Permalink

    When mitochondria work intensively, their electron transport chain produces reactive oxygen species in dangerous proximity to the mitochondrial DNA. It is like having a firework factory next to a library. The DNA gets damaged and mutated. This is observed in tissues with high oxygen metabolism, such as muscles… and sperm. By the time the sperm cell reaches the oocyte, its mitochonidrial DNA must be damaged beyond rescue. It must be destroyed so that not to “pollute” the embryo’s mitochondrial gene pool. This is what I think.
    However, I didn’t expect elevated mortality caused by persisting paternal mitochondrial DNA. Possibly the damaged mitochondria trigger apoptosis in cells that shouldn’t undergo it. (And C. elegans with its rigidly pre-planned development is notoriously unable to make up for unexpected events during embryogenesis.)

    Posted June 27, 2016 at 5:03 am | Permalink

    For a person with only knowledge of high school biology, this is phenomenal. This is the stuff that Dawkins talks about, about the awe of discovery and science and why it is so wonderful to be part of this discovery. Inspires you to read and learn.

  27. cruzrad
    Posted June 30, 2016 at 6:05 pm | Permalink

    Very cool science. Thanks for posting.

  28. Posted July 12, 2016 at 2:29 pm | Permalink

    Late to the party — very interesting.

    I feel micro-aggressed …

%d bloggers like this: