Abstract |
In animals, mtDNA is maternally transmitted, therefore all progeny carries one single
mitochondrial haplotype, or otherwise mitotype. As a result, all progeny is expected to carry only
one mitotype, a condition described as homoplasmy. However, more and more individuals are
found to carry more than one mitotypes, namely they are heteroplasmic. Heteroplasmy can
emerge in two different ways. The first one is due to mutations. The second one is due to
circumstancial transmission of paternal mtDNA, along with the inheritance of maternal mtDNA,
which is described by the term ‘paternal leakage’. Studying paternal leakage is interesting, since
it can have great impact on mtDNA fitness and on its evolution, in general. Specifically, mtDNA
accumulates deleterious mutations faster compared to nDNA, because of its asexual mode of
transmission, however mtDNA remains functional. Leakage may contribute to mtDNA’s
preservation, since when two mitotypes are co-existing in an individual, recombination between
the different mtDNA molecules may occur. In turn, the detrimental accumulation of mutations
can be mitigated. Consequently, studying heteroplasmy and its pattern is of great significance.
In the first set of experiments, we used chimeric Drosophila lines that possess nDNA from D.
melanogaster and mtDNA from D. simulans, to study heteroplasmy. The lines used carry a specific
nuclear background (DGRP-820) and have been found to be heteroplasmic after replacing their
original mtDNA with mtDNA from D. simulans. On the contrary, other lines that did not carry this
particular nuclear background were homoplasmic for the inserted mitotype. After the first series
of experiments, we deduced that the emergence of the mel mitotype from the DGRP-820 line
was associated with the X chromosome, and specifically with the region between the phenotypic
markers f and mal. However, the fact that only the mel mitotype leaked to the next generation
from the heteroplasmic lines, and not the si mitotype, led to the hypothesis the observed
patterns were not due to true heteroplasmy, but due to a mtDNA fragment embedded in the
nDNA (Nuclear Mitochondrial, NUMT). In order to resolve this issue, we designed another set of
experiments and found that there is a large NUMT embedded in a 3.15Mb region of the X
chromosome, between f and mal markers. We estimated that the NUMT should be at least
16225bp in size, which is the largest NUMT found in the D. melanogaster species. Given the
abundance of NUMTs among genomes, heteroplasmy data should be handled with caution, so
they are not interpreted as heteroplasmy, while they are truly NUMTs.
In the next set of experiments, we tried to resolve the problem that arose with heteroplasmy
in chimeric lines, using a bioinformatic approach. Specifically, whole-genome sequencing data
were used in order to detect the presence of chimeril reads that align in both the mtDNA and the
X chromosome. For the analysis, we used reads that were flagged as unmapped after mapping
with a reference genome and we performed a BLAST search in order to find reads with similarities
with the X chromosome and the mtDNA. Several filters were applied, so that we could detect
pairs of reads that correspond to the junctions of the NUMT. However, BLAST hits were hard to
reduce even after filtering and, furthermore, positive control showed that reads that correspond
to the junctions are probably eliminated during the filtering process. We assumed that the major
cause for the inability to detect chimeric reads is the relatively short reads that were used,
compared to the large size of the NUMT that we were trying to locate.
In the last set of experiments, we used Drosophila collected from a wild population to assess
if paternal leakage happens with a different frequency, dependent on the nuclear genome. For
this purpose, we constructed isofemale lines of D. simulans and we crossed females from these
lines with D. mauritiana males. In total, we analysed 2292 individuals for the presence of the
paternal mtDNA, which came from 15 different isofemale lines. Statistical analysis of the results
showed that paternal leakage is more frequently detected in males from three specific lines
compared to the other lines. A similar difference was detected when analysing all progeny,
regardless of sex. Interestingly, we found two lines where there was no difference in leakage
detected between males and females, opposed to what was expected from previous studies. We
deduced that paternal leakage is associated with the nuclear background of the D. simulans lines,
suggesting that leakage does not happen randomly due to failure of mechanisms that eliminate
paternal mtDNA from transmitting to the next generation, but it is a process cotrolled by the
nDNA.
|