Evolutionary biology

Mitochondrial genes on the move

Article metrics

In flowering plants, genes have frequently been transferred from mitochondria to the cell nucleus by way of a remarkable evolutionary rapid-transit system.

Mitochondria, the cell's energy-generating organelles, have their own genome, known as mitochondrial (mt) DNA. This genome is, however, minuscule compared with that of the free-living bacterium from which mitochondria originated1 in eukaryotes — organisms whose cells have a defined nucleus. Gene transfer from the mitochondrial to the nuclear genome has played a central role in this process of reductive genome evolution2. The protein product of such a transferred gene still functions in the mitochondrion, but the gene instead operates from the nucleus.

On page 354 of this issue, Adams et al.3 pose the question, “During the evolution of a particular eukaryotic lineage, how many times and how often has a given mitochondrial gene been relocated from organelle to nucleus?”. In the case of angiosperms (flowering plants), the answer is, “Very many times and very often, indeed”.

Even the most gene-rich mtDNAs encode at least ten times fewer genes than the smallest bacterial genomes. So most of the reduction in mitochondrial genome size and gene content is assumed to have occurred early in mtDNA evolution. However, the extremely variable gene content of contemporary mitochondrial genomes implies that mitochondrial gene loss has been an on-going process. By mapping gene losses onto phylogenetic trees, we can infer that the same mitochondrial gene has been transferred to the nucleus in separate events in different eukaryotic lineages4. Individual instances of mitochondrion-to-nucleus gene transfer have been well documented5. Up to now, however, we have had little appreciation of the overall frequency and timing of such gene relocations during mtDNA evolution.

Adams et al.3 report the remarkably frequent loss from angiosperm mtDNA of the gene encoding ribosomal protein S10 (a component of the mitochondrial ribosome) and the transfer of this gene (rps10) to the nucleus. Their results conjure up a picture of an evolutionary rapid-transit system, in which rps10 is being continually and frequently transported from the mitochondrial to the nuclear genome in different groups of angiosperms — and at a dizzying pace in evolutionary terms.

The authors used part of the rps10 gene as a hybridization probe to score the presence or absence of this gene in DNA extracted from 277 angiosperms representing 169 families. The lab concerned has creatively exploited this valuable resource in other studies6,7. The approach works well in plants for two reasons. First, there are many copies of the mitochondrial genome. Second, angiosperm mitochondrial genes evolve very slowly compared with single-copy nuclear genes — not to mention their counterparts in the mitochondria of animals, fungi and protists.

When plotted onto a phylogenetic tree of the angiosperms examined, the hybridization results indicated a loss of rps10 on 26 separate occasions. In 16 of these cases, the authors amplified a nuclear rps10 counterpart (not detectable by hybridization) using the polymerase chain reaction. Because a single mitochondrion-to-nucleus transfer of rps10 could conceivably be followed by many independent losses of this gene from the mitochondrial genome, several of the nuclear rps10 genes were characterized in more detail to see how many separate rps10 transfers these 16 losses may represent.

To become active in the nucleus, a gene acquired from mtDNA must be inserted into the nuclear genome in such a way that a mature, translatable messenger RNA can be produced. Moreover, the resulting protein (made outside the nucleus, in the cell's cytosol) must be targeted to and imported into mitochondria. What emerges from the analysis of such transferred mitochondrial rps10 genes is an intriguing diversity in the mechanism of functional activation in different angiosperms.

In some cases, pre-existing copies of other nuclear genes have been parasitized, with the rps10 coding sequence being inserted into the host gene. In several instances, a mitochondrial targeting sequence (presequence) has been effectively co-opted to provide entry for the RPS10 protein back into the mitochondrion; however, different nuclear genes are used as the source of the presequence in different plants. In other cases, the nucleus-encoded RPS10 protein appears to be imported into mitochondria despite the absence of an obvious presequence. From these results and other published data, the authors infer that there have been at least seven separate transfers of rps10 to the nucleus in angiosperms, and probably many more.

Is this just the tip of the iceberg? Apparently it is. Adams et al.3 point out that their study encompasses only about 2% of living angiosperm genera, so that rps10 may have been “independently transferred hundreds of times during angiosperm evolution”. Other genes that encode ribosomal proteins also seem to have been lost frequently from angiosperm mitochondrial genomes. Individual examples of such transfers have been characterized, again illuminating the myriad mechanisms that seem to be employed for functional activation. In grasses, for instance, a relocated rps14 gene has ended up within a nuclear gene (sdhB) that encodes a subunit of the mitochondrial enzyme succinate dehydrogenase. Co-expression of the sdhB and rps14 coding sequences, followed by alternative splicing of the resulting mRNA, gives rise to two different proteins having the same mitochondrial targeting signal8,9.

Further studies have yielded evidence of comparably high and variable rates of loss of a majority of 13 other mtDNA-encoded ribosomal protein genes during angiosperm evolution (K. Adams, Y.-L. Qiu and J. D. Palmer, personal communication). In contrast, only two of a sample of 13 genes involved in energy generation showed a similarly high rate of loss from angiosperm mtDNA. A particularly striking point is that in many plants the relative rate of ribosomal-protein gene loss actually exceeds the rate of nucleotide substitution at a single silent site (a site at which a mutation does not result in an amino-acid change): this is a conclusion that Adams et al. (and I) find astonishing.

In other eukaryotes (animals, for instance), further transfer of genes from mitochondria to the nucleus has effectively been blocked by changes in the mitochondrial genetic code: essentially, mitochondrial gene content has been frozen. In plants, mitochondrion-to-nucleus gene relocation is undoubtedly facilitated by the fact that there is no genetic code barrier to such transfer: the systems for translating RNA into protein use the same genetic code in both mitochondria and the cytosol. Even so, we could hardly have imagined the relative ease with which angiosperms seem to be able to shuttle mitochondrial genes into the nucleus and activate them there. Unravelling the genetic and biochemical workings of this highly efficient rapid-transit system will undoubtedly provide further surprises down the road.


  1. 1

    Gray, M. W., Burger, G. & Lang, B. F. Science 283, 1476– 1481 (1999).

  2. 2

    Berg, O. G. & Kurland, C. G. Mol. Biol. Evol. 17, 951–961 (2000).

  3. 3

    Adams, K. L., Daley, D. O., Qiu, Y.-L, Whelan, J. & Palmer, J. D. Nature 408, 354– 357 (2000).

  4. 4

    Gray, M. W. Curr. Opin. Genet. Dev. 9, 678–687 (1999).

  5. 5

    Schuster, W. & Brennicke, A. Annu. Rev. Plant Physiol. Plant Mol. Biol. 45, 61–78 ( 1994).

  6. 6

    Cho, Y., Qiu, Y.-L, Kuhlman, P. & Palmer, J. D. Proc. Natl Acad. Sci. USA 95, 14244–14249 (1998).

  7. 7

    Qiu, Y.-L, Cho, Y., Cox, J. C. & Palmer, J. D. Nature 394, 671–674 (1998).

  8. 8

    Figuerosa, P., Gómez, I., Holuigue, L., Araya, A. & Jordana, X. Plant J. 18, 601–609 (1999).

  9. 9

    Kubo, N., Harada, K., Hirai, A. & Kadowaki, K.-i Proc. Natl Acad. Sci. USA 96, 9207–9211 (1999).

Download references

Author information

Correspondence to Michael W. Gray.

Rights and permissions

Reprints and Permissions

About this article

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.