Functional analysis of the nuage, a unique germline organelle, in drosophila melanogaster 4

71 3.1.3 KRIMP interacts genetically with other nuage components Previous work has suggested that the nuage components interact genetically: localisation of AUB to the nuage depends on

70

Figure 3.1.9 UASp-krimp-venus transgene fully rescues krimp mutant defects By crossing flies harbouring the venus-krimp transgene into krimp mutant background, (a)

the oocyte nucleus compacts into a karyosome (blue) and C(3)G (red) becomes

chromosomal Bar is 5 µm (b) MAEL and AGO3, whose localisation depends on

KRIMP, localises to the perinuclear regions of the germline cells Bar is 10 µm (c) osk

mRNA is repressed normally Bar is 20 µm (d) GRK expression is comparable to the

wild-type ovariole and the protein localises to the anterior-dorsal region of the stage 8 egg chamber Bar is 20 µm

c

d

71

3.1.3 KRIMP interacts genetically with other nuage components

Previous work has suggested that the nuage components interact genetically: localisation

of AUB to the nuage depends on VAS, and MAEL localisation depends on VAS and AUB (Findley et al., 2003) To determine if KRIMP exhibits genetic interaction with other nuage components, KRIMP localisation to the perinuclear nuage was examined in

squ, spn-E, vas, aub, cuff, and mael mutants In all of the examined mutants except mael,

KRIMP perinuclear localisation was affected, while the localisation of AGO3 and MAEL appeared to depend on KRIMP, as well as on SQU, SPN-E, VAS, AUB, and CUFF (Figure 3.1.10; Findley et al., 2003) Moreover, AGO3 and MAEL localisation to the perinuclear nuage were rescued in 100% (n = 30) and % (n = 29) of the ovarioles in the

presence of a UASp-krimp-venus transgene that was driven by nosgal4VP16 in krimp

mutant (Figure 3.1.9b), thereby confirming the dependency of both proteins on KRIMP expression It was also noticeable that VAS localisation depends partially on proper AUB

localisation Although VAS foci were apparent in aub mutant, cytoplasmic VAS was

visibly more abundant than in the wild-type (Figure 3.1.10) Hence, a feedback mechanism may exist between VAS and AUB

The genetic interaction among the different nuage components suggests that there may exist a “platform” that lies in the perinuclear vicinity to facilitate the recruitment of multiple nuage components It is possible that these examined components have comparable activities in a common molecular machine; or alternatively, they have distinct functions that are carried out using a common platform, the nuage

72

Figure 3.1.10 Nuage components interact genetically with one another Ovaries from

different mutant flies are immunostained for the nuage components Homozygous mutant alleles or allelic combinations are used for all the mutants Localisation of the nuage components at the perinuclear regions of the germline cells reflects a hierarchical assembly All the nuage components VAS, AUB, KRIMP, AGO3, and MAEL, depend

on SQU to localise normally to the perinuclear regions AUB, KRIMP, AGO3 and MAEL depend on SPN-E and VAS for proper localisation; KRIMP, AGO3 and MAEL depend on AUB and CUFF to localise to the nuage; AGO3 and MAEL depend on VAS, AUB, and KRIMP to localise normally Bar is 10 µm

3.1.4 KRIMP interacts physically with other nuage components

The genetic dependency of the nuage components on one another, as well the co-localisation of different proteins, is suggestive of the presence of a macromolecular complex or sub-complexes at the perinuclear regions of the germline cells To determine

if the nuage components are functioning as complexes, the physical interactions between

in vitro translated KRIMP and some nuage components including AUB, CUFF, AGO3,

and MAEL were examined When KRIMP-MYC was pulled down, AUB-HA,

CUFF-73

HA, and AGO3-HA were co-immunoprecipitated, indicating that KRIMP interacts directly with these three nuage components (Figure 3.1.11) On the other hand,

MAEL-HA was not, or weakly, pulled down with KRIMP-MYC, suggesting that the nuage is made up of sub-complexes

Figure 3.1.11 KRIMP interacts directly with AUB, CUFF, and AGO3 Fusion

proteins are synthesized using TNT® Coupled Rabbit Reticulocyte Lysate System (Promega) MYC-tagged KRIMP is immunoprecipitated with IgG-coupled Protein A/G beads, in the presence of either HA-tagged AUB, CUFF, AGO3 or MAEL KRIMP-MYC interacts directly with AUB-HA, CUFF-HA, and AGO3-HA The asterisk indicates the IgG band

To confirm KRIMP association with the nuage components in vivo, KRIMP and one of

the interacting nuage components AGO3 were, respectively, expressed as a fusion protein

*

74

of Venus and HA in the wild-type ovary using the UAS/GAL4 expression system Co-immunoprecipitation was then performed with lysates prepared from the ovaries Venus-KRIMP co-immunoprecipitated with AGO3-HA (Figure 3.1.12), implying that Venus-KRIMP interacts physically with at least one or more nuage components in the germline

Figure 3.1.12 KRIMP interacts with AGO3 in vivo Venus-tagged KRIMP and

HA-tagged AGO3 are overexpressed in wild-type ovary using the UAS/GAL4 system and lysates are prepared for co-immunoprecipitation KRIMP-Venus co-immunoprecipiates with AGO3-HA in ovary lysates

3.1.5 KRIMP participates in retroelement repression

Past work has shown that some nuage components including SPN-E, AUB, and VAS repress the expression of retroelements in the germline (Savitsky et al., 2006; Vagin et al., 2004; Vagin et al., 2006) To investigate if KRIMP has a similar function,

semi-quantitative RT-PCR was performed on total RNA isolated from krimp control and

mutant ovaries Retroelements that were examined include the repetitive long

interspersed nuclear elements (LINEs) such as HeT-A, TAHRE, and I-element (Aravin et al., 2003), and a tandem-repeat that lies near the β-heterochromatin mst40 (Steven and

75

Russel, 1994) In krimp mutant ovary, all of the examined retroelements were

de-repressed (Figure 3.1.13), suggesting that KRIMP shares a common function with

SPN-E, AUB, and VAS

Figure 3.1.13 Retroelements are de-repressed in krimp mutant Semi-quantitative

RT-PCR of the retroelements HeT-A, TAHRE, I-element, and mst40 in krimp mutant ovary All examined retroelements are de-repressed in krimp mutant

3.1.6 KRIMP’s domains display distinct functions

To further characterise KRIMP functions, as well as its genetic and physical interactions with other nuage components during oogenesis, truncated N-terminus (NT) and C-terminus (CT) KRIMP transgene variants were generated KRIMP-NT harboured both the coiled-coil domain and CCCH-type zinc finger motif, while KRIMP-CT only

contained the tudor domain (Figure 3.1.14) These transgenes were introduced into krimp

76

mutant and expressed using UAS/GAL4 expression system Ovaries were then dissected

and examined for krimp mutant phenotypes that were rescued

Figure 3.1.14 Schematic drawing depicting KRIMP transgene variants KRIMP-NT

harbours the coiled-coil domain and CCCH-type zinc finger motif KRIMP-CT only contains the tudor domain ZnF, zinc finger

NT, when expressed, localised predominantly as perinuclear foci, while

KRIMP-CT appeared diffuse in the cytoplasm of wild-type ovary (Figure 3.1.15) In krimp mutant ovary that expressed KRIMP-NT, krimp mutant phenotypes including AGO3 and MAEL

perinuclear localisation, GRK expression level and anterior-dorsal localisation, and karyosome formation, were rescued On the other hand, the introduction of the transgene

variant for KRIMP-CT into krimp mutant background only rescued the precocious translation of osk mRNA (Figure 3.1.15) Although the anterior-dorsal localisation of GRK appeared to be partially rescued in krimp mutant ovary expressing KRIMP-CT, ectopic GRK spheroid aggregates were detected This phenocopies bicaudal-C mutant,

which is known to affect the secretory pathway and cause the accumulation of