In addition to being energy generators, mitochondria control many cellular processes including apoptosis. They are dynamic organelles, and the machinery of membrane fusion and fission is emerging as a key regulator of mitochondrial biology. A novel and conserved mitochondrial rhomboid intramembrane protease has been identified that controls membrane fusion in Saccharomyces cerevisiae by processing the dynamin-like GTPase, Mgm1, thereby releasing it from the membrane (McQuibban, 2003). The genetics of mitochondrial membrane dynamics has until now focused primarily on yeast. In Drosophila, the mitochondrial rhomboid (Rhomboid-7) is required for mitochondrial fusion during fly spermatogenesis and muscle maturation, both tissues with unusual mitochondrial dynamics. Mutations in Drosophila optic atrophy 1-like (Opa1-like or CG8479), the ortholog of yeast mgm1, display similar phenotypes, suggesting a shared role for Rhomboid-7 and Opa1-like, as with their yeast orthologs. Loss of human OPA1 leads to dominant optic atrophy, a mitochondrial disease leading to childhood onset blindness. rhomboid-7 mutant flies have severe neurological defects, evidenced by compromised signaling across the first visual synapse, as well as light-induced neurodegeneration of photoreceptors that resembles the human disease. rhomboid-7 mutant flies also have a greatly reduced lifespan (McQuibban, 2006).

Mitochondria form a tubular network in most eukaryotic cells, the result of a balanced set of opposing membrane fusion and fission reactions. Disturbance of this balance disrupts mitochondrial architecture and function. Mitochondria are surrounded by double membranes and must maintain effective separation from the surrounding cytoplasm; there is considerable complexity in the mechanism underlying the membrane dynamics required for fusion and fission. In Saccharomyces cerevisiae, two dynamin-related large GTPases, Mgm1 (Guan, 1993; Sesaki, 2003a) and Fzo1 (Rapaport, 1998; Hermann, 1998), regulate the fusion reaction, although their precise functions remain unknown. It has recently become clear that Mgm1 is regulated by a two-stage proteolytic release from the inner mitochondrial membrane: first, the mitochondrial targeting sequence is removed by MPP (mitochondrial processing peptidase), a matrix-localized protease; second, the intramembrane serine protease activity of the mitochondrial rhomboid family member Rbd1/Pcp1 releases a soluble form of Mgm1 into the intermembrane space (McQuibban, 2003; Herlan, 2004; Herlan, 2003; Sesaki, 2003a). Loss of Rbd1/Pcp1 leads to a phenotype similar to loss of Mgm1, implying that this intramembrane cleavage is necessary for full Mgm1 function (McQuibban, 2006).

Rbd1/Pcp1 and Mgm1 are conserved throughout eukaryotes, but much less is known about the genetic networks or mechanisms that control mitochondrial membrane dynamics in higher organisms. In humans, disruption of mitochondrial function is a major cause of inherited disease, including many neurological defects. Of particular relevance to thee studies, mutation of OPA1, the human ortholog of Mgm1, leads to autosomal dominant optic atrophy, the most common familial, childhood-onset cause of blindness (Delettre, 2002). This implies that normal membrane dynamics are necessary for human mitochondrial function and provides a strong incentive for understanding the underlying biological control of this process (McQuibban, 2006).

The only gene in Drosophila known to be needed specifically for mitochondrial fusion is fuzzy onions (fzo), which regulates mitochondrial fusion only in the male germline of flies (Hales, 1997). In an attempt to extend the genetic analysis of mitochondrial membrane fusion in Drosophila, focus was placed on Rhomboid-7, which is predicted to be mitochondrial by virtue of an N-terminal targeting sequence. When Rhomboid-7 was expressed in mammalian cells, the protein exclusively colocalized with Mitotracker and the F₁F₀ ATP synthase, respectively, both well-characterized markers for the mitochondrial network. As an initial approach to revealing the function of Rhomboid-7, RNAi was used to disrupt gene expression in cultured Drosophila S2 cells. Treatment with double-stranded (ds) RNA that targets rhomboid-7 resulted in a majority of cells with a highly fragmented mitochondrial network and a quantitative reduction in rhomboid-7 expression. By homology searching, the apparent Drosophila ortholog (CG8479) of mgm1 and OPA1 was identified, that is designate as optic atrophy 1-like (opa1-like) after the mammalian ortholog. Cells treated with dsRNA that targets opa1-like resulted in a fragmented mitochondrial network, a phenotype indistinguishable from rhomboid-7 disruption, and a quantitative reduction in Opa1-like expression. These results suggest that Rhomboid-7 and Opa1-like participate in a similar process of maintaining mitochondrial morphology and that, as in other organisms, disruption of mitochondrial fusion leads to fragmentation. They also suggest that the machinery of membrane fusion control may be conserved between yeast and metazoans (McQuibban, 2006).

To complement the Drosophila RNAi analysis and to investigate whether a similar function was conserved in mammalian cells, mouse mitochondrial rhomboid (called PSARL, presenilins-associated rhomboid-like; Pellegrini, 2001) and mouse OPA1 were expressed in COS-7 cells. Cells that expressed high levels of either PSARL or OPA1 but not a catalytic mutant of PSARL displayed complete aggregation of the mitochondrial network, in a phenotypically identical fashion. This mitochondrial aggregation phenotype resembles the excess fusion caused by overexpression of the mitofusins (Santel, 2001), the mammalian homologs of Fzo. Cells that expressed either PSARL or OPA1 at low levels did not generally disrupt the normal tubular network seen in COS-7 cells. In parallel, the effects were examined of overexpressing Rhomboid-7 in Drosophila S2 cells. As with mammalian cells, this led to mitochondrial aggregation. These data are consistent with the Drosophila RNAi experiments: together, they strongly suggest that, as in yeast, mitochondrial rhomboids and Opa1-like GTPases are involved in the control of mitochondrial membrane dynamics in both flies and mammals (McQuibban, 2006).

Since all other known phenotypes associated with loss of rhomboid-7 or its yeast or mammalian homologs can be attributed to defects in mitochondrial fusion, it is suspected that this is the cause of the abnormal synaptic transmission. There are many studies implicating mitochondria in synaptic function, although their precise role remains to be determined. The current data strongly suggest that fully efficient mitochondrial membrane fusion is necessary for normal photoreceptor activity. Consistent with this, loss of synaptic mitochondria prevents mobilization of the reserve pool of synaptic vesicles. It will be interesting to determine whether impairment of mitochondrial fusion in synapses leads to the same specific defect. In fact, the basic need for mitochondrial fusion and fission in any cells is not fully understood; various theories have been proposed, but the precise reasons for mitochondrial networks being so dynamic remain mysterious (McQuibban, 2006).

Mutations in the human OPA1 gene cause dominant optic atrophy, a common form of childhood-onset blindness. Therefore light-induced photoreceptor degeneration was measured in rhomboid-7 and control flies. Control flies show no defects after 3 days of continuous light. In contrast, over the same period (the maximum length the mutants survive), rhomboid-7^pΔ1 mutant photoreceptors degenerated severely, as shown by the almost complete loss of distinct rhabdomeres. This sensitivity to light implies that, in the absence of efficient mitochondrial membrane dynamics, photoreceptor neurons acquire activity-dependent damage at a much higher rate than normal, an interesting possible parallel with human dominant optic atrophy. Importantly, the rhomboid-7^pΔ1 mutant eyes showed no detectable degeneration under normal light or dark conditions, and it was these flies that showed the clear defect in signal transfer across the first visual synapse. This suggests that the synaptic defects are primary consequences of Rhomboid-7 loss, and not secondary to photoreceptor loss. Consistent with this conclusion that apoptosis may not be a primary phenotype of Rhomboid-7 loss, Drosophila S2 cells in which rhomboid-7 was reduced by RNAi showed no extra sensitivity to cycloheximide-induced apoptosis (McQuibban, 2006).

The results imply that not all mitochondrial membrane dynamics require functional Rhomboid-7 (otherwise the mutant would be fully lethal, as is the case for the mouse mitofusin genes MFN1 and 2). Instead, Rhomboid-7 is apparently required in places with a high requirement for mitochondrial fusion and/or energy demands. Loss of Opa1-like causes similar phenotypes to loss of Rhomboid-7 which, coupled with what has previously been shown in yeast, strongly suggests that they work in the same processes. It is worth noting, however, that opa1-like mutants are more severe than rhomboid-7 mutants, indicating that Opa1-like has additional functions that are Rhomboid-7 independent. A similar conclusion has been reached in yeast, where loss of Mgm1 causes more severe failure of mitochondrial fusion than loss of Rbd1/Pcp1 (Sesaki, 2003b). One goal of this work was to investigate the consequences of disrupting mitochondrial fusion in metazoans, where the potential for complex phenotypes is greater than in yeast. Although the full mechanisms that underlie the synaptic transmission defects are not know, the results lead to the idea that they are caused by loss of full mitochondrial function at the synapse. The mechanism underlying the striking longevity defect is less obvious but could be related to neural malfunction in regions other than the retina (McQuibban, 2006).

GENE STRUCTURE

cDNA clone length - 1301 bp

Bases in 5' UTR - 87

Exons - 6

Bases in 3' UTR - 158

PROTEIN STRUCTURE

Amino Acids - 351

Structural Domains

For information about Rhomboid proteins see Pfam

date revised: 26 January 2007

Home page: The Interactive Fly © 2006 Thomas Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.