This article argues that hydrogenosomes are derived from mitochondria. I don’t see anaerobic organisms as oxygen-’shunning’ and also not that without a doubt htey are mitochondria in the endosymbiotic sense. In the eukaryotic sense, yes. He seems to see the presence of DNA per se as good evidence for a derived mitochondrion, whereas in my opinion it is an intermediate or protomitochondrion.
Abstract: Mitochondria typically respire oxygen and possess a small DNA genome. But among various groups of oxygen-shunning eukaryotes, typical mitochondria are often lacking, organelles called hydrogenosomes being found instead. Like mitochondria, hydrogenosomes are surrounded by a double-membrane, produce ATP and sometimes even have cristae. In contrast to mitochondria, hydrogenosomes produce molecular hydrogen through fermentations, lack cytochromes and usually lack DNA. Hydrogenosomes do not fit into the conceptual mold cast by the classical endosymbiont hypothesis about the nature of mitochondria. […] A new report has uncovered DNA in the hydrogenosomes of anaerobic ciliates. The sequences show that these hydrogenosomes are, without a doubt, mitochondria in the evolutionary sense, even though they differ from typical mitochondria in various biochemical properties.
However, to date, in studies of the ciliates, as with other eukaryotic groups , either mitochondria or hydrogenosomes have been found, never any intermediates. Therefore, one might ask, where are the missing links that might connect the aerobic and anaerobic forms? New findings by Brigitte Boxma et al.  reveal that the first bona fide intermediate organelles that unite the hallmark features of mitochondria (a DNA genome) and hydrogenosomes (H2-production) are found in the anaerobic ciliate Nyctotherus ovalis, which inhabits the cockroach hindgut. The mitochondria of N. ovalis that Boxma et al.  describe have rather well-defined cristae.
Various hydrogenosome-bearing ciliates harbour methanogenic endosymbionts, which conduct interspecies hydrogen transfer with the organelles and are sometimes also physically associated with them [2,3]. In the case of Nyctotherus, this is a sensitive bioassay indicating that the organelles do indeed produce H2 , upon which the methanogenic endosymbionts depend for their growth and survival.
It seems that hydrogenosomes can be fit in as an intermediate, having already some parts of the repiratory chain, but without oxygen as terminal acceptor, which can be later in evolution added as an extension to existing functionality.
This DNA appears to be a rather typical ciliate mitochondrial genome like those found among aerobic ciliates: it encodes ribosomal proteins and components of the mitochondrial respiratory chain , which is a significant finding.
Furthermore, the Nyctotherus hydrogenosomes share attributes not only with O2-respiring mitochondria (the genome) but also with mitochondria that produce ATP without the need for oxygen but that do not produce H2 – organelles designated as anaerobic mitochondria . Similar to the anaerobic mitochondria of some metazoans, the mitochondrion of Nyctotherus possesses an incomplete citric acid cycle . […]. substantial amounts of the 14C label were excreted as acetate and succinate, end products that are typical of anaerobic mitochondria because they occur in protists and various metazoans that, for part or all of their life cycle, are exposed to anoxic environments . The detection of radioactively labelled succinate is significant because it indicates that endogenously produced fumarate is serving as a terminal electron acceptor in Nyctotherus. Fumarate reduction in anaerobic mitochondria usually requires the presence of rhodoquinone with its stronger reducing potential  rather than ubiquinone (typical of aerobic mitochondria). In line with that view, Boxma et al.  detected rhodoquinone in N. ovails, albeit at relatively low amounts, leaving open an interesting issue as to its functional significance.
Finally, Boxma et al.  also report several expressed sequence tags from Nyctotherus that indicate the presence of typically mitochondrial functions in these hydrogenosomes. Inhibitor studies suggest the presence of functional complex I and complex II in the mitochondrion along with the lack of a terminal oxidase, while also confirming that N. ovalis is indeed an anaerobe because prolonged exposure to O2 kills the organism. Like most anaerobes, it can survive short-term oxygen exposure because it possesses biochemical mechanisms to remove O2 that do not involve a respiratory chain.