Maternal mitochondrial function affects paternal mitochondrial inheritance in Drosophila

Abstract The maternal inheritance of mitochondria is a widely accepted paradigm, and mechanisms that prevent paternal mitochondria transmission to offspring during spermatogenesis and postfertilization have been described. Although certain species do retain paternal mitochondria, the factors affecting paternal mitochondria inheritance in these cases are unclear. More importantly, the evolutionary benefit of retaining paternal mitochondria and their ultimate fate are unknown. Here we show that transplanted exogenous paternal D. yakuba mitochondria can be transmitted to offspring when maternal mitochondria are dysfunctional in D. melanogaster. Furthermore, we show that the preserved paternal mitochondria are functional, and can be stably inherited, such that the proportion of paternal mitochondria increases gradually in subsequent generations. Our work has important implications that paternal mitochondria inheritance should not be overlooked as a genetic phenomenon in evolution, especially when paternal mitochondria are of significant differences from the maternal mitochondria or the maternal mitochondria are functionally abnormal. Our results improve the understanding of mitochondrial inheritance and provide a new model system for its study.


Introduction
It is widely accepted that mitochondria are inherited maternally.Indeed, mechanisms preventing paternal mitochondrial transmission to offspring in spermatogenesis and post-fertilization have been reported (Wallace 2007;Levine and Elazar 2011;Sato and Sato 2017).In C. elegans, paternal mitochondria are identified post-fertilization, degenerated in a CPS-6 (endonuclease G)-mediated pathway (Zhou et al. 2016), and then encapsulated by autophagosomes and delivered to lysosomes (Al Rawi et al. 2011;Sato and Sato 2012).In D. melanogaster, most of the paternal mitochondrial DNA (mtDNA) is degraded by the endonuclease G-mediated pathway during spermatogenesis, and the remaining paternal mtDNA is removed via cellular remodeling during spermatid tail formation (DeLuca and O'Farrell 2012).After fertilization, the paternal mitochondria are poly ubiquitinated, surrounded by MVB (multivesicular body)-like vesicles and finally degraded through autophagy (Politi et al. 2014).
Compared with D. melanogaster and C. elegans, the mechanism of paternal mitochondria degradation pathway is more complex in mammals.In mice, paternal mitochondria are poly ubiquitinated during sperm formation before entering the egg (Sutovsky et al. 1999).Shortly after fertilization, paternal mitochondria are recognized and surrounded by autophagy-related factors (Al Rawi et al. 2011).It remains unclear whether paternal mitochondria are eliminated during embryogenesis.Some studies suggest that paternal mitochondria are not eliminated, but rather unevenly distributed to daughter cells during embryonic mitoses (Luo et al. 2013), while others report that paternal mitochondria are cleared by mitophagy and/or the proteasome system (Rojansky et al. 2016;Song et al. 2016).Although detailed mechanisms remain to be elucidated, it is clear that various pathways work to ensure that maternal inheritance of mitochondria occurs in animals.
Despite the numerous pathways that either eliminate paternal mitochondria or prevent their replication, paternal mitochondrial leakage is still detected in many species such as fish (Peng et al. 2018), insects (Kondo et al. 1990;Ma et al. 2014;Polovina et al. 2020), mice (Gyllensten et al. 1991;Kidgotko et al. 2013), sheep (Zhao et al. 2004), and even human beings (Luo et al. 2018).One study found that paternal mitochondria in mice are eliminated in intraspecific crosses but not in interspecific crosses (Kaneda et al. 1995).Another group hypothesized that the elimination of paternal mitochondria postfertilization is mainly due to the combined effects of a nuclear coding factor that recognizes sperm mitochondria and another factor in eggs (Rokas et al. 2003).Given the same genetic background during intraspecific hybridization, the two factors are thought to coordinate to eliminate paternal mitochondria, whereas the genetic background variation during interspecific hybridization weakens the elimination process and leads to paternal mitochondrial inheritance.However, paternal mitochondrial leakage during intraspecies hybridization has been detected in Drosophila (Matsuura et al. 1991;Nunes et al. 2013), which indicates that the genetic background difference cannot entirely explain all cases of paternal mitochondrial leakage.
Paternal mitochondrial inheritance in humans is highly controversial.The first report of paternal mitochondrial leakage in humans was published in 2002 (Schwartz and Vissing 2002).In 2018, a study was published reporting a high degree of mitochondrial heteroplasmic variants in 17 individuals from 3 unrelated families, and the authors proposed that the inheritance of paternal mitochondria is similar to autosomal dominant inheritance (Luo et al. 2018).This work created significant controversy and has been intensely debated in the field.Indeed, it has since been argued that there is insufficient evidence to confirm paternal mitochondrial inheritance in humans from that study, mainly because the mitochondrial heteroplasmic types in the offspring could not be established in their father or grandfather (Lutz-Bonengel and Parson 2019).Yet another study sequenced the mitochondrial DNA from 41 patients with mitochondria diseases, as well as their parents, and found no case supporting paternal mitochondrial inheritance (Rius et al. 2019).Moreover, the analysis of 33,105 whole genome sequences revealed that the rare inherited nuclear-encoded mitochondrial segments (NUMTs) can create the impression of heteroplasmy, similar to paternal mitochondrial inheritance (Wei et al. 2020).Consequently, there remains little supporting evidence for paternal mitochondrial inheritance in humans.
Although paternal mitochondrial inheritance has been reported in several species, the question of whether paternal mitochondria can be systematically inherited has been debated for decades.The controversy remains ongoing due to the lack of effective ways to detect paternal mitochondria and measure the efficiency of paternal mitochondrial inheritance.We thus do not know under which circumstances paternal mitochondria are preserved, or if there is a mechanism regulating paternal mitochondria retention.It is also important to discern whether the preserved paternal mitochondria are functional, as there would be little apparent evolutionary benefit from retaining nonfunctional mitochondria.Previously, there have been a few studies showing paternal mitochondrial leakage in fruit flies and mice (Kidgotko et al. 2013;Wolff et al. 2013;Polovina et al. 2020), but no studies showing the functional state of the retained paternal mitochondria.
The ability to select offspring with paternal mitochondria, track them, and reveal the genetic mechanism of paternal mitochondria transmission is a challenging task.In our study, as a first step toward this goal, we developed a research model to efficiently select offspring with paternal mitochondria.In this model, we utilized two D. melanogaster strains.One is a temperature-sensitive (29°C) lethal mtDNA mutant strain, designated mt: CoI ts , which contains a point mutation in the mitochondrial gene CoI (Hill et al. 2014).The other strain is a heteroplasmic fly (Het) strain with two kinds of mitochondria, the proportionally minor D. melanogaster mitochondrion (mt: CoI ts , mt: nd2 del1 ) and D. yakuba mitochondrion (Lieber et al. 2019).We crossed the mt: CoI ts (female) with Het (male) and recovered F1 survivors at restrictive temperature.Subsequent analyses confirmed successful paternal transmission of functional mtDNA to offspring.Interestingly, F1 survivors were similarly recovered from crosses of mt: CoI ts females with different wild type melanogaster strains, although definite evidence for the presence of paternal mtDNA in the offspring remains elusive.Our results nevertheless indicate that paternal mitochondrial inheritance might be a regulated process, and this study established a genetically tractable model for its study.

Fly genetics
The fly strains used in this study are CS, w 1118 , mt: CoI ts (gifted by Dr. Xu Hong), Het fly strain (gifted by Dr. Thomas R. Hurd).Flies were raised in an approximate 12 h light (700 lux)/12 h dark cycle, at temperatures of 18°C or 29°C.

Genetic crosses
The crosses of F0 were conducted at 18°C, then the eggs were collected within 12 hours, and transferred to 18°C (control group in Fig. 3a) or 29°C for culturing until pupal eclosion (Fig. 1b and c, Fig. 2, Fig. 3b, Fig. 4, Supplementary Fig. 1).The F1 pupal eclosion rate of each vial was counted and the average rate of each group and condition was calculated.The mean values are presented.The cross of F1 females with mt: CoI ts males was also conducted at 18°C, then the eggs were collected within 12 hours and were transferred to 29°C until pupal eclosion.The subsequent generations after F1 were kept at 29°C for culturing for their life-time.

Mitochondrial respiratory chain complex IV activity assay and statistical analysis
For each experiment, 50 4-day-old flies from every group were collected and homogenized in 1 ml mitochondrial extracting buffer and analyzed using the Mitochondrial Respiratory Chain Complex IV Activity Assay Kit (D799473, Sangon Biotech).The enzyme activity for each group was calculated following the standard formula recommended in the kit.Each data point was the average of at least 3 independent experiments.The mean values and SE of each group are presented.The P values were determined with t-tests, as implemented in Prism 8 for Windows 10 GraphPad Software.

DNA extraction, PCR, and restriction enzyme treatment
Individual flies were homogenized in 100 μL solution A (0.1M Tris, 0.1M EDTA, 0.1%SDS, pH 9.0) and incubated at 70°C for 25 min for cell lysis and DNA denaturation.14 μL KAc (8M) was added for renaturation on ice for 30 min.DNA extracts were separated at 4°C by centrifugation for 15 min at 14,000 rpm.After that, total DNA was precipitated with 60 μL isopropyl alcohol and washed with 70% ethanol, then finally diluted in 10 μl ultra-pure water. 2 μL DNA extract was used for each PCR reaction (Taq DNA polymerase, M0273L, NEB).
For enzyme treatment in Supplementary Fig. 1b and c, 25 μL PCR product was purified and treated with XhoI restriction enzyme (XhoI, R0146S, NEB) at 37°C overnight.
The separation of mitochondria and nucleus in Fig. 2c was conducted following the protocol from the Qproteome Mitochondria Isolation Kit (QIAGEN,37612).About 30 1-day-old flies were homogenized and subjected to the mitochondria isolation process.The pellets containing mitochondria and nuclei were later subjected to the DNA extraction process.Resultant pellets containing mtDNA or nuclear DNA were suspended in 30 μL water.PCR was carried out using 2 μL mtDNA and nuclear DNA extraction.

qPCR quantification of mtDNA
One-day-old flies were homogenized and subjected to DNA extraction.The resultant pellet was suspended in water (10 μL ultra-pure water/1 single fly).qPCR was carried out using 1 μL of DNA extraction and 400 nM of each primer pair with a SLAN-96P Real-time PCR machine and GoTaq qPCR Master Mix (2X) (Promega, A6001).The PCR program was: 2 min at 50°C, 5 min at 95 °C, 40 cycles of 95 °C for 30 s and 60 °C for 1 min.Dissociation curves were generated through a thermal denaturation step and used to verify amplification specificity.The data were calculated by the ΔΔCt method and rp49 was used as the reference gene.The P values were determined with t-tests.The data were normalized to wild-type samples and plotted as fold changes.

PCR primers
The primers binding specifically to mtDNA of either D. melanogaster or D. yakuba are: The primers for amplification of mtDNA fragments including the XhoI restriction site in D. melanogaster are: The primers used in qPCR are:

The F1 progeny of mt: CoI ts females × Het males develop into adults, and the pupal eclosion rate correlates to the age of female flies
To study the transmission and functionality of paternal mitochondria in D. melanogaster and to construct an operational model to investigate the pathways regulating paternal transmission, we utilized 2 D. melanogaster lines, the mt: CoI ts and the Het fly (Fig. 1a).The mt: CoI ts is a temperature sensitive mutant line containing a point mutation in the mitochondrial gene CoI (Hill et al. 2014).The mt: CoI ts flies develop normally at a permissive temperature (18°C), but fail to eclosion from pupae at a restrictive temperature (29°C) (Fig. 1b).If eclosion is allowed to occur at the permissive temperature and then is followed by a shift to 29°C, the life span of the mt: CoI ts flies is no more than 5 days (Hill et al. 2014).The Het fly contains 2 kinds of mitochondria, D. melanogaster mitochondrion (mt:CoI ts , mt: nd2 del1 ) and D. yakuba mitochondrion (Lieber et al. 2019).
We crossed mt: CoI ts females with Het males and, surprisingly, found that the pupal eclosion rate of the F1 flies was greater than zero (ranging from 4.26 to 27.82%) at the restrictive temperature (Fig. 1c).Since there was a high pupal eclosion rate variance, we wondered if this correlated to the age of the female flies.By counting and comparing of the F1 eclosion rate of females of different ages, we found that the F1 eclosion rate decreased with the age of female flies.The finding that F1 flies were recovered strongly supports that paternal D. yakuba mitochondria are retained and functional.We also found that the majority of surviving F1 flies were females (91.65%).We do not have an explanation for the observation.

Detection of paternal D. yakuba mtDNA in F1 flies from the cross mt: CoI ts females × Het males
There have been several previous attempts to detect paternal mitochondrial leakage, including by analyzing interspecific crosses in Drosophila (Kondo et al. 1990(Kondo et al. , 1992;;Dokianakis and Ladoukakis 2014;Polovina et al. 2020) and by examination of naturally occurring heteroplasmic populations (Nunes et al. 2013;Wolff et al. 2013).In our system, we detect the inheritance of paternal mitochondria in a simple way, by performing PCR to specifically detect paternal mtDNA in F1 flies.Primers are chosen to bind specifically to unique regions of mtDNA from either D. melanogaster or D. yakuba (Fig. 1a).The D. yakuba primers are specific and we did not detect any D. yakuba mtDNA in extracts from mt: CoI ts flies (Fig. 2a).
We used this detection method to analyze YF1 flies (from the crosses of mt: CoI ts × Het).PCR analysis revealed the presence of D. yakuba mtDNA fragments in YF1 flies, implying paternal D. yakuba mitochondria transmission (Fig. 2a).The PCR bands differ in intensity, implying the transmission efficiency among individual flies is variable, which may also contribute to the eclosion rate variability observed before.Sequencing results confirmed the identification of D. yakuba mtDNA sequences in YF1 flies (Fig. 2b).
Previous reports have shown that mtDNA can integrate into chromosomes to form NUMTs, which are mainly caused by chromosome breakage (Gray et al. 1999;Wei et al. 2020;Tigano et al. 2021).To exclude any effect of NUMTs, we prepared mitochondrial extracts and conducted PCR with the specific D. yakuba primers.As expected, D. yakuba fragments could be amplified from mitochondrial extracts, but not from nuclear extracts (Fig. 2c).Thus, the effect of NUMTs can be excluded.Moreover, qPCR analysis showed YF1 flies acquired about 5.85% D. yakuba mtDNA of Het flies (Fig. 2d).These results strongly suggest that YF1 flies inherited paternal D. yakuba mitochondria.
To further support that paternal D. yakuba mitochondria are inherited and functional, we conducted a Mitochondrial Respiratory Chain Complex IV Activity Assay on YF1 flies.We found that YF1 flies acquired about 45.67% complex IV activity of Het flies (Fig. 2e).

Dysfunctional female mitochondria facilitate paternal D. yakuba mitochondrial transmission
Paternal mitochondrial leakage has been previously reported in Drosophila and Mus musculus.For example, 32-48% of paternal mitochondrial inheritance was reported in interspecific crosses of Drosophila (Sherengul et al. 2006).Other studies reported lower numbers, in one case 0.66% (Wolff et al. 2013) and in another 6% (Nunes et al. 2013).So far, there is no evidence that maternal mitochondria affect the inheritance of paternal mitochondria.In order to study the effect of maternal mitochondria, we conducted crosses and looked for evidence of paternal inheritance in the context of functional or dysfunctional maternal mitochondria.
First, we crossed mt: CoI ts (female) with the Het fly (male) and cultured them at the permissive temperature of 18°C.The maternal mitochondria functioned normally and we could not detect paternal D. yakuba mitochondria by PCR analyses, indicating no or negligible paternal mitochondrial inheritance (Fig. 3a and c).Next, we crossed w 1118 (female) with Het fly (male) and cultured them at 29°C.The maternal mitochondria in this cross also functioned normally and we could not detect paternal D. yakuba mtDNA in F1 flies either (Fig. 3b and c).When we crossed mt: CoI ts (female) × Het flies (male) at 29°C, where the maternal mitochondria in mt: CoI ts were dysfunctional, the pupal eclosion rate ranged from 4.26 to 27.82% (Fig. 1c), suggesting successful paternal inheritance.These results suggest that the functionality of maternal mitochondria can affect the efficiency of paternal D. yakuba mitochondria elimination in fertilized eggs.

mtDNA of paternal origin can be transmitted to subsequent generations
Our results show that paternal D. yakuba mitochondrial transmission can confer viability to offspring and rescue the mitochondrial respiratory chain complex IV activity in YF1 flies (Figs. 1c and 2e), indicating the rescue of mitochondrial function in YF1 flies.We then asked if paternal D. yakuba mitochondria could be prorogated in subsequent generations, although via the normal mode of maternal inheritance.
We collected YF1 females and backcrossed them with mt: CoI ts flies (male), due to insufficient YF1 males.We found the pupal eclosion rate of YF2 (YF1 females × mt: CoI ts male) flies at 29°C was about 12%, and not higher than that of YF1.This result could be explained by a dearth of paternal D. yakuba mitochondria in the YF1 fly reproductive system due to uneven distribution of paternal D. yakuba mitochondria in the first mitoses of YF1 embryos.Succeeding YF2 male and female flies were then collected, crossed, and cultured at 29°C for their life-time.Paternal D. yakuba mtDNA could be detected in YF3 and YF4 flies (Fig. 4a and b).The intensity of the PCR bands from YF1 flies was much lower compared to the Het flies (Fig. 2a), while the intensity of the PCR bands from YF3-4 flies was nearly identical to that of Het flies (Fig. 4a and  b), suggesting the content of paternal mtDNA in YF3 and YF4 flies was proportionally higher than that in YF1 flies.Paternal mtDNA was also present in somatic cells and reproductive systems in YF3 flies (Fig. 4c).
A previous study showed that deleterious mitochondria are gradually replaced by healthy mitochondria (Hill et al. 2014).We therefore asked if the content of D. yakuba mtDNA changes after many generations.In order to answer this, we collected flies over 30 generations and then quantified the D. yakuba mtDNA content by qPCR.After 30 generations, the content of D. yakuba mtDNA increased identical to that of Het flies (Fig. 2d).Our results strongly demonstrate that paternally inherited mitochondria can be transmitted to subsequent generations just like normal mitochondria.

The F1 progeny of mt: CoI ts females with D. melanogaster strains can develop into adults
Since the D. yakuba mitochondria in Het flies are interspecific, it is possible that our result simply reflects the failure of elimination mitochondria in an interspecific cross.To investigate this possibility, we crossed mt: CoI ts females with w 1118 males that contain wild-type D. melanogaster mitochondria (Supplementary Fig. 1a).The pupal eclosion rate of these WF1 flies was again greater than zero (ranging from 3.55 to 13.32%), though lower than the eclosion rate of YF1 flies from the cross of mt: CoI ts (female) × Het fly (male) (Fig. 1c).To help rule out the effect of nuclear background, we crossed mt: CoI ts females with another D. melanogaster strain Canton S (CS), and obtained a pupal eclosion rate of 19.91% ±11.32% (Supplementary Fig. 1f).
We then looked for the presence of paternal mitochondrial DNA in the WF1 (from the cross of mt: CoI ts ×w 1118 ) flies.The mitochondrial CoI gene of w 1118 contains a XhoI restriction site, CTCGAG, whereas the restriction site is mutated to TTCGAG in the mt: CoI ts mutants.DNA fragments amplified from w 1118 and mt: CoI ts flies were treated with XhoI restriction enzyme.While XhoI treatment of the amplified DNA fragments from w 1118 yielded 2 small bands (230bp + 440 bp), the fragments from the mt: CoI ts mutant fly did not (Supplementary Fig. 1b).XhoI treatment of the fragments of WF1 flies also failed to yield 2 small bands, which would have confirmed the presence of paternal mitochondrial DNA.
We considered that the amount of paternal mitochondria (with the XhoI restriction site) in WF1 flies may be below the detection limits.To gather support for this possibility, we mixed mt extracts from mt: CoI ts and w 1118 flies (Supplementary Fig. 1c), PCR amplified the target fragments, and treated them with XhoI.At a ratio of 2:1 (mt: CoI ts : w 1118 ) the digested bands can be detected, but they are not detectable upon further dilution.This result supports the above possibility.
We used an alternative approach to investigate whether the WF1 generation contains paternal mtDNA, by measuring the Mitochondrial Respiratory Chain Complex IV activity in WF1 flies.Previous research has shown the cytochrome oxidase activity in mt: CoI ts flies was unaffected at 18°C, but significantly reduced at the restrictive temperature of 29°C (Hill et al. 2014).We found that WF1 flies acquired about 40% complex IV activity compared to wild-type flies (Supplementary Fig. 1d).
Based on these results, we speculate that a similar phenomenon of paternal leakage exists in the F1 flies from crosses between mt: CoI ts females and D. melanogaster strains containing wild type mitochondria, when maternal mitochondria are dysfunctional.

Discussion
It is widely accepted that mitochondria are inherited maternally.Several pathways mediating paternal mitochondria elimination during spermatogenesis and zygote formation have been elucidated.However, paternal mitochondrial leakage has been reported in species such as fruit flies (Sherengul et al. 2006;Nunes et al. 2013), mice (Shitara et al. 1998), andsheep (Zhao et al. 2004).Heteroplasmic variants in patients with mitochondrial diseases in three independent families were reported in 2018, suggesting paternal mitochondrial inheritance in humans (Luo et al. 2018).Subsequently, paternal mitochondrial inheritance has become controversial.Some researchers believe that paternal mitochondrial leakage in interspecific hybridization is due to different genetic backgrounds, while others have proposed that a complex regulatory mechanism mediates paternal mitochondrial leakage.There are also those who believe that paternal mitochondrial leakage is a rare accident (Pagnamenta et al. 2021).Our results showed that exogenous paternal mitochondria can be transmitted to YF1 flies especially when maternal mitochondria are dysfunctional.It's exciting to discover that the inherited mitochondria are functional and their levels increase in subsequent generations.Interestingly, we observed a similar phenomenon when only melanogaster mtDNA is present in both parents.Until we can definitively identify wild type D. melanogaster mtDNA in WF1 progenies, it remains highly speculative that paternal leakage occurs under the condition when maternal mitochondria are dysfunctional.
Nevertheless, we established a genetically tractable system for the study of paternal mitochondria leakage under any conditions.We envision that induced mutations under the genetic background of mt: CoI ts would uncover novel pathways regulating paternal mitochondrial transmission.

Fig. 1 .
Fig. 1.The pupal eclosion rate from the cross mt: CoI ts females with Het males correlates with the age of female flies.a) Schematics of the mitochondrial genome and the D loop regions of D. yakuba and D. melanogaster.PCR primers specifically targeting the unique regions of D. yakuba and D. melanogaster mtDNA sequences are shown.b) The pupal eclosion rates of w 1118 , mt: CoI ts , and the Het flies.The results shown are mean ± SE. c) The pupal eclosion rate of F1 flies from the cross of mt: CoI ts females (ages, 1-15 days) with Het males.The results shown are mean ± SE.

Fig. 2 .
Fig. 2. Detection of paternal mtDNA in F1 flies from the cross mt: CoI ts females with Het males.a) Paternal D. yakuba mtDNA can be detected in F1 flies from the cross mt: CoI ts females with Het males.Amplified DNA fragments from mtDNA of mt: CoI ts and Het are used as negative and positive controls.b) Amplified DNA fragments of F1 flies from the cross mt: CoI ts females with Het males are shown aligned with D. yakuba mtDNA sequence.c) Mitochondrial and nuclear extracts from mt: CoI ts , Het, and F1 flies were separated and amplified with D. yakuba specific primers.Amplified fragments can be detected in the mitochondrial extracts from Het and F1 flies, but not in the nuclear extracts.d) Real-time RT-PCR analysis of mtDNA content.d) yakuba mtDNA in F1 flies is about 5.85% and in Fn (n > 30) is about 103.76% of Het flies.Four data sets were averaged.The content of D. yakuba mtDNA in F1 (*P < 0.05) and Fn (****P < 0.0001) flies is significantly different from that in mt: CoI ts flies.e) Mitochondrial Respiratory Chain Complex IV Activity of Het, mt: CoI ts and F1 flies at 29°C.F1 flies acquire about 45.67% complex IV activity of wild-type flies.The results shown are mean ± SE.Six data sets were averaged.The complex IV Activity in F1 flies is significantly different from that in mt: CoI ts flies (***P < 0.001).

Fig. 4 .
Fig. 4. Detection of paternal mtDNA in the F3, F4 generations and the bodies/ovaries of F3 flies.a) Detection of paternal D. yakuba mtDNA in F3 flies.Amplified D. yakuba DNA fragments from mt:CoI ts and Het flies were used as negative and positive controls.b) Detection of paternal D. yakuba mtDNA in F4 flies.Amplified DNA fragments from mt:CoI ts and Het flies were used as negative and positive controls.c) Detection of D. yakuba mtDNA in the bodies and ovaries of F3 flies.Dissected ovaries and bodies from F3 flies were separately subjected to DNA extraction.Amplified D. yakuba DNA fragments were detected in F3 ovaries and bodies.Fig. 3. Detection of paternal mtDNA in F1 flies from control crosses.a) Detection of paternal D. yakuba mtDNA in F1 flies from the cross mt: CoI ts females with Het males at 18°C.Amplified DNA fragments from mtDNA of mt: CoI ts and Het flies are used as negative and positive controls.No amplified bands could be detected from mt: CoI ts × Het F1 flies cultured at 18°C.b) Detection of paternal D. yakuba mtDNA in F1 flies from the cross w 1118 females with Het males at 29°C.Amplified DNA fragments from mtDNA of mt: CoI ts and Het flies are used as negative and positive controls.c) Real-time RT-PCR analysis of mtDNA content.The content of D. yakuba mtDNA in F1 flies from mt: Col ts females ×Het males at 18°C and w 1118 females ×Het males at 29°C has no significant difference to that in mt: Col ts flies.Three data sets were averaged.