Please note! This essay has been submitted by a student.
The c323a driver is expressed strongly in follicle cells of D. melanogaster ovaries, particularly in the later stages of 9 to 11. Its expression is weak in germline cells, but is relatively active in the gut during larval development (Beaucher et al., 2007). Micro-cytoophidia in follicle cells of female c323a-GAL4>Oregon-R were truncated and compacted. This phenotype was obvious as early as in stage 7 egg chambers. Shortening was unexpectedly even more clearly displayed by nurse cells; truncation of macro-cytoophidia was very much apparent in a majority of egg chambers, and by stage 11 would have been almost completely reduced. Greater numbers of oocyte micro-cytoophidia were observed instead.
T-155 is an enhancer sequence specifically active in ovarian follicle cells, with strong signatures in accessory glands and columnar cells of the male reproductive tract (Hrdlicka et al., 2002). It is most active between stage 6 to 12. No changes were visible to follicle cell cytoophidia in T155-GAL4>Oregon-R flies; as was normal, one micro-cytoophidium of average length was seen per cell. Instead, nurse cell macro-cytoophidium was noticeably shorter than average. This phenotype was consistently seen in egg chambers of late stage 7 to late stage 10. Truncation was nonetheless moderate, unaccompanied by an increase in micro-cytoophidia numbers, and overall lacking the severity observed in c323a-GAL4>Oregon-R flies.
We utilized two nos-GAL4 lines for this screening: they were 32563 and 107748, each respectively obtained from Bloomington and Kyoto resource centres. They both carry constructs tethering the GAL4 gene to the nanos or nos-NGT driver. Expression is therefore restricted to germline cells; 32563 and 107748 flies should in theory be crossed only to UASp-carrying fly lines. This variant of the nanos gene driver in particular is a driver of early oogenesis, most active in the germarium (Staller et al., 2013). Nos-GAL4>Oregon-R F1 flies did not display obviously altered cytoophidia lengths, shape, nor compaction within their nurse cells, when compared to Oregon-R. Follicle cell micro-cytoophidia remained duly unaffected. These phenotypes were consistently maintained up to early stage 11, where cytoophidia within both nurse and follicle cell would become reduced naturally.
Act5c is a ubiquitous driver of D. melanogaster. It is expected to assert strong expression signatures in egg chambers at all stages of development, and in both soma and germ-cell types. Act5c-GAL4>Oregon-R females display shorter micro-cytoophidia within the follicle cells of their egg-chambers. Slight shortening of nurse-cell macro-cytoophidia was also observed, often along with compaction and increased numbers of micro-cytoophidia. These phenotypes were seen in egg chambers as young as stage 5, to those of late stage 10.
Crosses were made between a virgin driver-GAL4 female and a UAS-miRX male; X represents any miRNA. Each cross was repeated at least thrice. Reciprocals were made at random, to assess whether outcomes were sex-influenced. A total of 123 miRNAs borne by 234 fly lines were screened. The characteristics under scrutiny were primarily cytoophidia length, numbers, and density or compaction. Affected polarity and the discombobulation of cell shape and size are secondarily considered traits. In the first screening phase, a candidate miRNA is shortlisted if its overexpression was found to cause any of these changes; to be retained in the next list, they would need to show replicability, in that the same phenotype was observable in the next two bioreplicates of the cross. Some miRNAs were represented by more than one UAS-miRX lines i.e. ‘cousin-constructs’. They differ by point of insertion and chromosomal location within the D. melanogaster genome. In such cases, the miRNA’s overexpression by the cousin-construct should also produce a similar outcome, although leniency was given in the way of phenotypic strength.
Both follicle cell drivers (FCD) were utilized throughout the primary stages of screening. These were considered lowest-tier as compared to well-known drivers such as nos or Act5c, they were likely to be less active and have weaker expression signatures. This very trait nonetheless makes them the best candidates for early screening purposes: a positive hit would mean that a candidate miRNA is of consequence enough that even its mild overexpression is sufficient to induce changes to cytoophidia. We have established that cytoophidia-truncating side-effects of GAL4 were possible for either driver line. By simple logic, it would therefore be much easier to identify cytoophidia-elongating candidates (CytEl-miRNAs) than those which shorten them (CytSh-miRNAs). As expected, we were able to isolate more elongating miRNAs, at nine and four CytEl-miRNAs and CytSh-miRNAs, respectively.
CytEl-miRNAs were those which display much lengthened cytoophidia, either in follicle or nurse cells. Given that the structure would already be considerably shortened from ectopic GAL4 expression under C323a, those which confer a sufficient diminishing to cytoophidia length and/or a surge in the number of nurse-cell micro-cytoophidia under T155-GAL4 driver expression is shortlisted as CytSh-miRNAs.
Apart from length, size and compaction-based identification, we identified an additional three candidates which may distort the polarity of follicle-cell micro-cytoophidium. This structure as it appears in egg chambers overexpressing CytPol-miRNAs tend to be misaligned from the micro-cytoophidia of neighbouring follicle cells.
As described previously, nos-GAL4 is a germline cell driver. The variant carried by functionally-redundant lines of 32563 and 107748 is NGT-nos, which is expressed within the germarium (Staller et al., 2013). If miRNA overexpression does indeed affect cytoophidia, phenotypic alterations are therefore expected to be retained through to the later-stages of oogenesis. To fully enable this driver, the corresponding miRNA gene must be tethered to a UASp construct, rather than UASt, as the latter is not expressible in the female germline (Rørth, 1998). We crossed all 48 available UASp lines to ¬nos-GAL4. These included two each of those already identified as either CytEl or CytSh-miRNAs. No CytPol-miRNAs were UASp-enabled. Immunostaining showed that whilst CytSh-miRNAs maintained their ability to shorten cytoophidia in nurse cells, neither CytEl-miRNA induced prominent changes. Another two CytSh-miRNA candidates were also discovered from the remaining UASp lines.
As a final screening step, all lines were crossed to Act5c-GAL4 flies. This was to both further strengthen the efficacy of shortlisted candidates, and to determinedly eliminate thus-far inconsequential miRNAs. Not all crosses were expected to produce adult F1 flies. A total of 63 of the 126 miRNAs involved were found to be lethal under ubiquitous expression by Act5c-GAL4 (Schertel et al., 2012). We proceeded with indiscriminate crossing regardless, and observed that despite their reported lethality, of-genotype F1 emergent flies were retrieved for 7 of those 63 miRNAs. All had construct alternatives, i.e. redundant UAS-lines. Act5c-GAL4 was found to exacerbate the elongating and shortening effects of most of the shortlisted candidates, but did not appear to affect polarity.
Extremity of phenotypic manifestations and reproducibility under this varied array of drivers allowed us to rank CytEl, CytSh, and CytPol-miRNAs, for the purpose of downstream prioritization.
Nine CytEl-miRNAs were identified altogether. Two were considered prime candidates due to their consistency at inducing cytoophidia elongation i.e. miR-975 and miR-1014. Both miRNAs are insect exclusive, and neither have reported human paralogues. The miRNA-975 is carried by 9 lines, in various combinations. On the Drosophila genome, its gene is a part of a tri-miRNA cluster, alongside miRNAs 976 and 977 (Ryazansky, Gvozdev, & Berezikov, 2011). This cluster is regionally associated with miR-4966 upstream, as well as miR-978 and miR-979 genes downstream, although their expressions are dependent on separate transcription start sites.
Throughout the overexpression screen, egg chambers overexpressing miR-975 alone consistently showed obvious lengthening of cytoophidia. Under FCD-driven expression, cytoophidia elongation is more obvious within nurse cells, whereas the elongated phenotype is obvious in both follicle and nurse cells with Act5c-GAL4. Out-of-the-ordinary numbers of micro-cytoophidia did not accompany macro-cytoophidia elongation in either case. These observations were repeatable as miR-975 was expressed in a cluster with miR-976 and miR-977, or as a pair with miR-978, though to a lesser degree. The only instance whereby miR-975 did not assert its cytoophidia-lengthening effects was when it shared a construct with miR-4966, alluding to the possibility that these genes may co-regulate each other’s activity. MiR-975 overexpression nonetheless did not seem to bear cell-death inducing effects. In fact, the opposite might be the case; where its expression was driven by Act5c-GAL4, we have observed instances whereby hyperactive cytoophidia formation coincided with the doubling of nurse cell numbers within an egg chamber. The frequencies in which this phenotype was found was nonetheless low i.e. in only ~2.06% of ovarioles (n=348).
MiR-1014 is the second CytEl-miRNA identified as a prime candidate. Its gene is carried by three fly lines. In vivo, miR-1014 occupies a locus on chromosome 3, within the intronic regions of the salt gene. Its overexpression consistently induced the lengthening of cytoophidia regardless of driver and UAS-construct, though one bore greater effects above the rest. Though elongation patterns are akin to those described for miR-975, excessive miR-1014 levels did appear to impact cell survival negatively. Nearly a quarter of egg chambers of stage 8 and beyond (i.e. ~22.35%; n=464) could discernibly be seen to be undergoing apoptosis at any given time. This is a significant increase from natural cell death seen in Act5ct-GAL4>Oregon-R controls, as well as amongst miR-975 overexpressing flies. Accordingly nurse cell nuclei doubling phenotypes were not recovered under miR-1014 overexpression.
Six miRNAs were determined to shorten cytoophidia. Based on their effectiveness and reproducibility, the two prime CytSh-miRNAs are miR-190 and miR-932. Both miRNAs are carried exclusively within UASp constructs, enabling germline-cell based overexpression analysis. Four UAS-lines encode singularly for miR-190. Two carry the gene on chromosome 2. On the other two, the construct has been inserted into chromosome 3, where the endogenous dme-miR-190 should also be found. Its locus is intronic to the rhea gene. Flies overexpressing the miRNA exhibit cytoophidia shortening consistently. The truncation of these filaments is often accompanied by its compaction, as well as greater numbers of micro-cytoophidia. Confidence in a phenotypic outcome is dependent upon pre-established, driver-specific reference phenotypes. The effects of its overexpression is therefore clearest under nos-GAL4, followed by Act5c-GAL4. As truncation was already anticipated in any crosses involving FCDs, shortening could only be ascertained by the near disappearance of cytoophidia. Chromosomal positioning of the overexpression construct itself did not appear to significantly affect outcomes. Two points should be noted. The first is that although the ubiquitous overexpression of miR-190 was expected to be lethal (Schertel et al., 2012), our own observations proved that this was not guaranteed. Act5c-GAL4>UAS-miR190 individuals were easily recovered, with flies which presented neither compromised fitness nor fertility. The second and more importantly, is that whilst FCD-driven truncation of follicle-cell cytoophidia affects nurse cell macro-cytoophidia greatly, neither nos-GAL4 nor Act5c-GAL4-driven truncation of macro-cytoophidia within germline nurse cells bear any affect upon follicle cell cytoophidia.
The second prime CtySh-miRNA is identified as miR-932. Like miR-975 and miR-1014, this miRNA is also exclusive to arthropods. Its overexpression construct is carried by three UASp-lines. In vivo, the dme-miR-932 gene is found in chromosome 2, intronic to the gene Nlg2. Its effects on cytoophidia typically culminate in the severe shortening of the filaments, as well as an increase in micro-cytoophidia numbers in both nurse cells and the developing oocyte.
Truncation of follicle-cell cytoophidia in flies overexpressing miR-932 was comparatively more discernible than that of flies overexpressing miR-190. However, phenotypes conferred when the transgene is carried by fly-line IMCB#3038 are inconsistent at times, where misshapen and highly-dense macro-cytoophidia appeared alongside truncated ones. Akin to dme-miR-190, Schertel et al. has also reported the lethality of overtly-excessive levels of dme-miR-932. We found that eventhough the recovery of Act-GAL4>UAS-miR932 flies was possible, their numbers were somewhat lower than Act-GAL4 variants of the other three prime candidates. This adverse effect on population size was especially true when the driver was paired to a chromosome 2 located UASp-miR932 insert, with recovered F1 numbers falling to half of the other two overexpression constructs (#3038 vs #3036 vs #116480: 21±4.74 vs 138±11.32 vs 108±9.86).
Apart from cytoophidia-length, we initially considered the distortion of polarity of these structures as they appear in follicle-cells to be indicative of a response to miRNA-overexpression. Those which conferred such an effect are classified as CytPol-miRNAs. ‘Normal’ polarity is conformity; in follicle-cell cross sections within controls, a majority of micro-cytoophidia may appear to be perpendicular to cell surface (dorsal to ventral). CytPol-miRNAs are therefore those which confer an extensive degree of directional disorganization upon these filamentous structures. Follicle-cell drivers were the most effective in identifying such miRNAs. Act5c-GAL4 exacerbates other effects of overexpression as well, and therefore was less effective for CytPol-miRNA-screening. However, we ultimately opted to not investigate any of the three CytPol-miRNAs as prime candidates. The reasoning behind this decision will be explained further in the following Discussion section.