Evolutionary biology

The hydrogenosome's murky past

The evolution of specialized cellular powerhouses called hydrogenosomes has long confounded biologists. The discovery that in some cases they have their own genome sheds some much-needed light on the issue.

Hydrogenosomes are double-membraned subcellular structures that generate hydrogen while making the energy-storage compound ATP. They are found in certain eukaryotic (nucleus-containing) microbes that inhabit oxygen-deficient environments1. The evolution of the hydrogenosome has remained obscure, mainly because these organelles seemed not to have a genome2,3 — until now. On page 74 of this issue, Boxma et al.4 report the characterization of what seems to be an authentic hydrogenosomal genome in the anaerobic microbe Nyctotherus ovalis, an inhabitant of the termite hindgut.

In eukaryotes that live in oxygen-rich (aerobic) environments, organelles called mitochondria are responsible for making ATP. Although an evolutionary relationship between hydrogenosomes and mitochondria has been postulated, this hypothesis remains contentious2,3. Mitochondria contain a small genome (mtDNA) that retains traces of their evolutionary origin from a bacterial symbiont5,6. Interestingly, the hydrogenosomal DNA isolated by Boxma et al. exhibits hallmarks of a bona fide mitochondrial genome.

Adding to this story are two recent papers7,8 that probe the evolutionary history of the hydrogenosome from another anaerobic microbe, the parasite Trichomonas vaginalis. The absence of a hydrogenosomal genome in this organism9 makes it a challenging task to infer the origin of its hydrogenosome. Indeed, on this point, the two groups7,8 come to rather different conclusions, even though they analyse the same Trichomonas hydrogenosomal proteins.

In animals and fungi, the mitochondrial genome encodes a small number of essential inner-membrane proteins (components of respiratory complexes I–IV and complex V, a specialized type of ATP-synthesizing enzyme) that function in electron transport and ATP production5. In addition, mtDNA specifies the RNA components of the mitochondrial protein-synthesis system and, in plants and many algae and protozoa, some of the proteins of this system too5.

The report by Boxma et al.4 extends their earlier observations10 — which were provocative but not compelling — that the hydrogenosome of Nyctotherus might contain DNA. Having purified Nyctotherus hydrogenosomes, the authors isolated a 14-kilobase stretch of DNA and sequenced it4. They identified genes that encode homologues of mitochondrial proteins; that is, the mitochondrial and hydrogenosomal counterparts are close relatives with similar sequences. These genes encode four subunits of complex I (Fig. 1), and two proteins and two RNAs from the protein-synthesis system. The properties of these sequences — for instance, characteristic codon-usage patterns and a similarity to mitochondrial genes from aerobic microbes of the same group as Nyctotherus (the ciliate protozoa) — make a convincing case that this DNA is part of an mtDNA-like hydrogenosomal genome.

Figure 1: Subunits of mitochondrial respiratory complex I.

The membrane-integrated (green rectangle) and peripheral (purple rectangle) regions include numbered subunits that are encoded in one or more mitochondrial genomes; for example, animal mtDNA specifies seven subunits, 1–6 and 4L. Several subunits (yellow) have been identified in Nyctotherus ovalis hydrogenosomal DNA6. Genes encoding the 51- and 24-kDa subunits (dark green) have only been found in nuclear genomes. A nucleus-encoded 75-kDa subunit has also been reported in N. ovalis6. In this step of respiration, the oxidation of nicotinamide adenine dinucleotide (NADH) and reduction of ubiquinone (Q) provide protons and electrons to be passed along the respiratory chain, eventually producing ATP and water.

Additionally, Boxma et al.4 identify several proteins in Nyctotherus that are encoded by genes in the nucleus but are typically transported to and function in mitochondria; these include three additional subunits of complex I (of molecular mass 24 kilodaltons (kDa), 51 kDa and 75 kDa; Fig. 1) and components of complex II. Phylogenetic reconstructions aimed at inferring the evolutionary history of these proteins show an affiliation with mitochondrial (specifically ciliate) homologues. Not unexpectedly, biochemical analyses suggest that Nyctotherus hydrogenosomes do not have complexes III and IV, which are responsible for the final stages of aerobic respiration. Nor is there any evidence of a mitochondrial-type ATP synthase (complex V) in this organism.

These and other observations imply that the Nyctotherus hydrogenosome represents an intermediate form between mitochondria, which possess a membrane-bound electron-transport chain, and previously characterized hydrogenosomes, which do not — a “true missing link”, in the words of the authors. In parallel, the results suggest that the Nyctotherus hydrogenosomal genome, whose total size, shape and gene content have yet to be determined, is probably a reduced ciliate-type mtDNA, lacking those mtDNA-encoded genes that normally specify components required to construct a complete mitochondrial respiratory chain.

The genome-less Trichomonas hydro-genosome has been much less forthcoming about its evolution, with sequence-based analysis necessarily limited to nuclear genes that specify the constituent proteins of this organelle. The simultaneous discovery by two groups7,8 of Trichomonas homologues of the 51- and 24-kDa components of mitochondrial complex I (Fig. 1) is a notable development. These proteins (termed Ndh51 and Ndh24, respectively) are the first Trichomonas counterparts of components of the mitochondrial respiratory chain to be identified. However, the two groups differ sharply in their conclusions about the evolutionary origin of these proteins, and hence of the hydrogenosome itself.

Both groups used a standard, rigorous approach for reconstructing evolutionary relationships by comparing protein sequences. However, Hrdy et al.8 conclude that Trichomonas Ndh51 shares a specific common ancestry with its mitochondrial counterpart, whereas Dyall et al.7 argue that it does not (Fig. 2). So, why the difference, and who is right?

Figure 2: The conflicting evolutionary positions of the Trichomonas vaginalis hydrogenosome.

In phylogenetic reconstructions based on an alignment of the Ndh51 (51-kDa subunit) protein sequence, Dyall et al.7 place the T. vaginalis hydrogenosome at the base of the α-proteobacterial lineage, not specifically related to mitochondria, whereas Hrdy et al.8 position the hydrogenosome as a specific relative of mitochondria, to the exclusion of α-proteobacteria. Numbers are statistical probabilities that strongly support the associated branches. (Figure courtesy of R. Watkins.)

These conflicting conclusions illustrate a common conundrum in using molecular-sequence data to infer ancient evolutionary events. In parasites such as Trichomonas, whose position in the eukaryotic lineage is uncertain to begin with, protein sequences tend to change relatively rapidly in the course of evolution. This can confound their accurate placement in phylogenetic trees, causing so-called long branches. Moreover, Trichomonas Ndh51 proved to have a very different amino-acid composition from its counterpart in other organisms, another phenomenon that can severely compromise phylogenetic analysis.

Hrdy et al.8 tried to offset the bias caused by the divergent amino-acid composition by assigning each of the 20 possible amino acids to one of six groups of amino acids that have similar chemistries and commonly replace one another in protein sequences. They then reconstructed the alignment of the Ndh51 and comparison sequences using just the six groups of amino acids and reanalysed the data. This technique has the effect of shortening long branches and homogenizing the amino-acid composition among compared sequences. Using this additional approach, Hrdy et al. deduced a common origin for the Trichomonas and mitochondrial 51-kDa proteins (Fig. 2).

Several points emerge from these three reports. First, Boxma et al.4 are the first to show that a putative evolutionary relative of the mitochondrion contains (and indeed encodes) homologues of proteins specified by mtDNA. By contrast, although Dyall et al.7 and Hrdy et al.8 also identified and studied two complex I homologues in hydrogenosomes, the genes encoding these two proteins (24 and 51 kDa) have not been found in any mtDNA to date but reside exclusively in the nuclear genome. Admittedly, there is strong evidence that the mitochondrial 24- and 51-kDa subunits of complex I originate from the proto-mitochondrial genome via gene transfer to the nucleus. Nevertheless, their connection (and that of their hydrogenosomal counterparts, Ndh24 and Ndh51) to the proto-mitochondrion is less direct than in the case of proteins whose genes have been retained in at least some extant mitochondrial genomes.

Second, the mitochondrial affiliation demonstrated with Ndh51 by Hrdy et al.8 is consistent with other data — particularly characteristics of the protein import system in hydrogenosomes — that unite these organelles with mitochondria2,3. By contrast, there is no solid evidence that specifically affiliates the hydrogenosome of any anaerobic eukaryote with a different eukaryotic bacterial group, in particular an anaerobic hydrogen-producing lineage.

Finally, the sporadic phylogenetic distribution of hydrogenosomes and the intimate phylogenetic intermingling of their anaerobic ‘hosts’ with aerobic, mitochondrion-containing relatives imply that hydrogenosomes are derived secondarily from mitochondria. Indeed, it seems that nature can evolve a hydrogenosome from a mitochondrion with relative ease.

This story is far from complete, because mitosomes — putative remnant mitochondria that lack the ability to make ATP — have recently been discovered in several microbial lineages that do not have conventional mitochondria11. The evolutionary and biochemical connections among mitochondria, hydrogenosomes and mitosomes must be elucidated if we are to truly understand the pathways and mechanisms of eukaryotic cell evolution.


  1. 1

    Müller, M. J. Gen. Microbiol. 139, 2879–2889 (1993).

    Article  Google Scholar 

  2. 2

    Embley, T. M. et al. IUBMB Life 55, 387–395 (2003).

    CAS  Article  Google Scholar 

  3. 3

    Dyall, S. D., Brown, M. T. & Johnson, P. J. Science 304, 253–257 (2004).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Boxma, B. et al. Nature 434, 74–79 (2005).

    ADS  CAS  Article  Google Scholar 

  5. 5

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

    ADS  CAS  Article  Google Scholar 

  6. 6

    Andersson, S. G. E., Karlberg, O., Canbäck, B. & Kurland, C. G. Phil. Trans. R. Soc. Lond. B 358, 165–179 (2003).

    CAS  Article  Google Scholar 

  7. 7

    Dyall, S. D. et al. Nature 431, 1103–1107 (2004).

    ADS  CAS  Article  Google Scholar 

  8. 8

    Hrdy, I. et al. Nature 432, 618–622 (2004).

    ADS  CAS  Article  Google Scholar 

  9. 9

    Clemens, D. L. & Johnson, P. J. Mol. Biochem. Parasitol. 106, 307–313 (2000).

    CAS  Article  Google Scholar 

  10. 10

    Akhmanova, A. et al. Nature 396, 527–528 (1998).

    ADS  CAS  Article  Google Scholar 

  11. 11

    Roger, A. J. & Silberman, J. D. Nature 418, 827–829 (2002).

    ADS  CAS  Article  Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Gray, M. The hydrogenosome's murky past. Nature 434, 29–31 (2005). https://doi.org/10.1038/434029a

Download citation

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.