1 Introduction

The heterogeneous nuclear ribonucleoproteins (hnRNPs) are conserved in eukaryotes and are involved in a variety of RNA processing events, including splicing, transport and translation (Daneholt 2001; Guisbert et al. 2005; Martinez-Contreras et al. 2007; Olson et al. 2007; Blanchette et al. 2009; He and Smith 2009; Chaudhury et al. 2010; Han et al. 2010; Nilsen and Graveley 2010). In addition, some of them are known to be important in chromatin organization and regulation of gene expression (Piacentini et al. 2009) and to affect neurodegeneration (Sengupta and Lakhotia 2006; Sofola et al. 2007; Mallik and Lakhotia 2010). The different hnRNPs are grouped into 13 families in mammals, of which 11 are represented in Drosophila (Busch and Hertel 2012). The hnRNP A family in Drosophila is represented by Hrb87F/Hrp36, Hrb98DE/Hrp38 and Rb97D proteins (Busch and Hertel 2012). The Hrb27C/Hrp48 and Squid/Hrp40 of Drosophila, which share sequence and functional similarities with the Hrp36 and Hrp38 proteins, were considered as members of hnRNP A/B family by Blanchette et al. (2009), while Busch and Hertel (2012) grouped Squid/Hrp40 in the hnRNP D family. Hrp36 and Hrp38 seem to have overlapping functions so that a genetic null condition for any one of them is not lethal (Haynes et al. 1991; Zu et al. 1996). Hrp40 and Hrp48 share many common pre-mRNA targets with Hrp36 and Hrp38 during splicing (Blanchette et al. 2009).

It has been shown earlier that several hnRNPs including Hrp36, Hrp38, Hrp40, Bancal (hnRNP K/Hrb57A) and other related RNA-processing proteins including NonA (DBHS or Drosophila Behavior, Human Splicing protein family member; Kozlova et al. 2006), Sxl, etc., associate, in addition to their presence at many chromosome regions, with the nucleus limited non-coding hsrω-n transcripts of the hsrω gene in the nucleoplasmic omega speckles (Lakhotia et al. 1999; Prasanth et al. 2000; Jolly and Lakhotia 2006; Onorati et al. 2011). A disruption of omega speckles either in the absence of hsrω-n RNA in hsrω nulls or following down-regulation of these transcripts through RNAi results in a diffused distribution of the hnRNPs in nucleoplasm, and therefore, the hsrω-n RNA is considered to be essential for organization of omega speckles and association of the hnRNPs in the speckled nuclear compartment (Prasanth et al. 2000; Mallik and Lakhotia 2011; Lakhotia et al. 2012). Omega speckles are believed to function as storage sites for the unengaged hnRNPs and other related RNA-processing proteins and thus regulate their availability according to cellular needs (Lakhotia 2003; Jolly and Lakhotia 2006; Lakhotia 2011). Heat shock causes a dynamic redistribution of RNA Pol II and the omega-speckle-associated proteins inside the nucleus (Saumweber et al. 1980; Bonner and Kerby 1982; Zeng et al. 1997; Lakhotia et al. 1999; Prasanth et al. 2000; Boehm et al. 2003; Gerber et al. 2005; Lakhotia 2011). A recent study from our laboratory reported that altered levels of the hsrω transcripts affect the remobilization of hnRNPs, RNA Pol II and HP1 during recovery from heat shock, resulting in severely compromised thermo-tolerance (Lakhotia et al. 2012).

Earlier studies have shown that Hrp36, along with Hrp38, Hrp40 and Hrp48, affect alternative splicing of several transcripts in a sequence-specific manner (Zu et al. 1996; Borah et al. 2009; Blanchette et al. 2009; Nilsen and Graveley 2010). Piacentini et al. (2009) reported that Hrp36, PEP, HP1 and DDP1 dominantly suppress position effect variegation (PEV) and thus seem to be involved in heterochromatin formation. Hrp36 is also reported to be involved in neuronal development and its enhanced expression suppresses neurodegeneration in Fragile X-associated tremor/ataxia syndrome through its direct interaction with riboCGG repeats (Sofola et al. 2007). Depletion of hrp36 leads to eye roughening with extra/fragmented photoreceptors and also enhances polyQ-induced neurodegeneration in Drosophila (Sengupta and Lakhotia 2006; Mallik and Lakhotia 2010). It is interesting that, in spite of the abundance of Hrp36 and its above noted roles in RNA processing, etc., its complete absence is reported to be without a phenotype and thus this protein has been suggested to be dispensable (Zu et al. 1996; Haynes et al. 1997). It is believed that the closely related Hrp38 provides functional redundancy so that a complete absence of Hrp36 is tolerated (Zu et al. 1996; Blanchette et al. 2009). Since the Hrp36 is a major constituent of omega speckles, we wanted to know the status of omega speckles in hrp36 null individuals. Our present studies reveal that, like the hsrω-n transcripts, Hrp36 too is essential for formation of omega speckles since in its absence the hsrω-n RNA and the other hnRNPs do not assemble into speckled structures. Contrary to the earlier reports (Zu et al. 1996; Haynes et al. 1997), we found that a complete absence of this protein actually results in pleiotropic effects like developmental delay, reduced female fecundity, severely compromised stress tolerance, etc. Therefore, this protein is not completely dispensable.

2 Materials and methods

2.1 Fly stocks and rearing condition

Fly stocks were routinely reared on standard cornmeal-agar food medium at 24°C ± 1°C. In one set of experiments, larvae, pupae and adults were reared at 18°C or 24°C or 30°C (±1°C) on standard food. Oregon R + was used as the wild type. The P{w + Tsr + }/P{w + Tsr + }; ry Df(3R)Hrb87F/ry Df(3R)Hrb87F stock (Zu et al. 1996; Haynes et al. 1997), obtained from Dr S Haynes, was used as the hrp36 null line. The entire hrp36 coding region along with approximately 0.6 kb of Tsr gene (CG9654, http://flybase.org ), encoding a Testis-specific RRM protein, is deleted in the Df(3R)Hrb87F chromosome, and therefore, ry Df(3R)Hrb87F/ry Df(3R)Hrb87F male flies are completely sterile, but P{w + Tsr + }/P{w + Tsr + }; ry Df(3R)Hrb87F/ry Df(3R)Hrb87F males and females carrying the P{w + Tsr + } transgene on chromosome 2 are fertile (Haynes et al. 1997). The P{w + Tsr + }/P{w + Tsr + }; ry Df(3R)Hrb87F/ry Df(3R)Hrb87F flies will be referred to as hrp36 null in the following.

2.2 Developmental assay

For developmental assay, eggs were collected at 2 h intervals from wild type or hrp36 null flies and those hatching within a 2 h interval were collected and transferred to food vials. The vials were monitored at 24 h intervals until fly emergence for progression of development. To examine the effect of temperature, developmental assay was carried out at 18°C ± 1°C, 24°C ± 1°C and 30°C ± 1°C separately.

2.3 Mouth hook preparation

For mouth hook preparation, wild type and hrp36 null larvae were taken out of the food vials at 24 h intervals after hatching of embryos. After washing with water, they were dried on blotting paper and fixed in glycerol : acetic acid (1:4) fixative at 60°C for 48 h, following which the mouth hooks and cuticle were separated and mounted in Hoyer’s solution and kept at 60°C for 48 h. Mouth hooks of different age larvae were examined under light microscope. Mouth hooks of 25 larvae of each genotype were prepared at 24 h intervals till 96 h after hatching.

2.4 Life span assay

Flies of desired genotype were collected within 4 h of eclosion and transferred to fresh food vials at 24°C ± 1°C. Twenty-five flies were transferred in each vial. The number of dead flies was counted at 24 h intervals and the live flies were transferred to fresh food vials, without anesthetization, at 48 h intervals. Ten replicates of each set of experiment were examined.

2.5 Fecundity assay

To examine the role of Hrp36 in fecundity, following crosses of three days old 20 male and 20 female flies were set up: (a) wild type ♀ × wild type ♂, (b) hrp36 null ♀ × wild type ♂, (c) wild type ♀ × hrp36 null ♂ and (d) hrp36 null ♀ × hrp36 null ♂. Flies were allowed to lay eggs on cornmeal-agar food plates. Numbers of eggs laid during 24 h intervals in each plate were counted for the next seven days. Hatching of eggs was examined after 24 and 48 h from egg laying. Three replicates were kept for each set of experiment.

2.6 Number of ovarioles in ovaries

Ovaries from three- and ten-day-old wild type and hrp36 null females, which were allowed to mate normally, were dissected out in Poels’ salt solution (PSS; Lakhotia and Tapadia 1998), followed by staining in 2% aceto-carmine for 5 min, washing and mounting in 45% acetic acid with a lightly pressing coverslip. They were examined under a Nikon E800 microscope for the number of ovarioles in each of the two ovaries of a female.

2.7 Starvation resistance assay

Three-day-old wild type and hrp36 null flies were transferred to empty bottles containing only water-soaked filter paper strips. The bottles were kept at 24°C ± 1°C and water-soaked filter paper strips were replaced at 8 h intervals. The number of dead flies was monitored at 8 h intervals after the starvation was started. Two replicates were carried out with each genotype.

2.8 Thermo-tolerance assay

Wild type and hrp36 null actively wandering third instar larvae were heat-shocked at 37°C for 1 or 2 h. Larvae of the desired genotypes were collected in 1.5 mL microfuge tubes lined with moist filter paper and incubated in water bath maintained at 37°C ± 1°C for 1 or 2 h, following which the larvae were transferred to fresh food vials at 24°C ± 1°C and monitored until fly emergence. In another set, three-day-old healthy wild type and hrp36 null flies were transferred in separate empty food vials and transferred to water bath maintained at 37°C ± 1°C for 1 h. The numbers of flies of each of the two genotypes that were immediately knocked down after heat shock and the numbers of flies that died during the next 5 days were counted. Each set of experiments was carried out at least in triplicate.

2.9 Size and polyteny levels of wild type and hrp36 null larval Malpighian tubules (MT) and salivary glands (SG)

To compare the width of MT and levels of polyteny in principal cells of MT, actively wandering wild type and hrp36 null late third instar larvae were dissected in PSS and the MT were immediately fixed in freshly prepared 3.7% paraformaldehyde in PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4, pH 7.4) for 20 min. After washing with PBS, tissues were stained with 4, 6-diamidino-2-phenylindole (DAPI, 1 μg/mL) and mounted on bridged slides in 1,4-diazabicyclo[2.2.2]octane (DABCO) antifade mountant. These were examined for measurement of width of MT under a Nikon E800 phase contrast microscope while nuclear diameter and total nuclear DAPI fluorescence intensity were measured from the confocal projection images.

To examine polytene chromosomes in principal cell nuclei, MT of actively wandering wild type and hrp36 null late third instar larvae were dissected out in PSS, fixed in 3.7% formaldehyde in PBS followed by 3.7% formaldehyde in 45% acetic acid for 1 min each. They were washed briefly in 45% acetic acid and finally squashed under a coverslip in 45% acetic acid by tapping with rubber-ended pencil, followed by firm thumb pressure. Squash preparations of SG from actively wandering wild type and hrp36 null late third instar larvae were also prepared similarly. Slides were briefly dipped in liquid nitrogen and the coverslips flipped off with a sharp blade. The slides were dipped immediately in 90% alcohol and after air-drying; the preparations were rehydrated in PBS, stained with DAPI and mounted in DABCO for examination under confocal microscope.

2.10 Whole organ immunostaining

MT from actively wandering wild type or hrp36 null third instar larvae were dissected out in PSS and immediately fixed in freshly prepared 3.7% paraformaldehyde in PBS for 20 min and processed further for immunostaining with the desired antibodies (see below) as described earlier (Prasanth et al. 2000). Ovaries from wild type or hrp36 null flies were stained with Tetramethyl Rhodamine Isothiocyanate (TRITC)-conjugated Phalloidin at 1:200 dilution to localize the F-actin. Chromatin was counterstained with DAPI. Tissues were mounted in DABCO antifade mountant for confocal microscopy.

2.11 Immunostaining of polytene chromosome squashes

Actively wandering wild type and hrp36 null late third instar larvae were heat shocked at 37°C for 40 min and their SG were dissected out in PSS either immediately or after 1 or 2 h recovery at 24°C. Parallel controls were maintained at 24°C. SG were dissected out from the control and treated larvae in PSS and transferred to 1% Triton X-100 in PBS for 30 s, following which they were treated consecutively with 3.7% formaldehyde in PBS and 3.7% formaldehyde in 45% acetic acid for 1 min each. Finally, the SG were incubated in 45% acetic acid for 1 min and squashed under coverslip in the same solution by first gentle tapping with rubber-ended pencil, followed by hard tapping and thumb pressure. Polytene chromosome spreads were examined under phase contrast microscope for appropriate spreading. Slides were briefly dipped in liquid nitrogen and the coverslips were flipped off with a sharp blade and the slides immediately dipped in 90% alcohol. These were processed further for immunostaining with the desired antibodies (see below) as described earlier (Lakhotia et al. 2012).

To examine the mode of association of proteins present at 93D locus after heat shock, one set of heat-shocked SG was treated with DNAse-free RNase A (1 mg/ml, Sigma-Aldrich, India) for 15 min at 24°C prior to fixation as described earlier (Lakhotia et al. 2012). The glands were processed further for polytene chromosome squash preparation and immunostaining as above.

2.12 Western blotting

Western blotting was done to estimate levels of HSP70, Hrb87F, NonA or β-tubulin. Following the desired treatment, protein samples were prepared either from whole larvae or a particular tissue (see Results) in SDS-PAGE sample buffer (Prasanth et al. 2000). After the electrophoretic separation on 10% SDS-PAGE, the polypeptides were blotted on PVDF membrane and processed for Western detection of the desired protein (see below) as described earlier (Mallik and Lakhotia 2009).

2.13 Antibodies

Primary antibodies used for immunostaining were mouse monoclonal anti-Hrp36 (P11; Saumweber et al. 1980) at 1:20 dilution, mouse monoclonal anti-NonA (Bj6, Saumweber et al. 1980) at 1:10 dilution, mouse monoclonal anti-Hrb57A (Q18; Saumweber et al. 1980) at 1:10 dilution, mouse monoclonal anti-Megator (BX34; Zimowska and Paddy 2002) at 1:10 dilution, mouse monoclonal anti-HP1 (C1A9; Developmental Studies Hybridoma Bank, USA) at 1:50 dilution, and rabbit anti-ISWI (Corona et al. 2007) at 1:2000 dilution. Appropriate secondary antibodies conjugated either with Cy3 (1:200, Sigma-Aldrich, India) or Alexa Fluor 488 (1:200; Molecular Probes, USA) or Alexa Fluor 546 (1:200; Molecular Probes, USA) were used to detect the given primary antibody.

The primary antibodies used for detection of HSP70, Hrb87F, NonA and β-tubulin in Western blots were 7Fb (Velazquez and Lindquist 1984; 1:1000 dilution), P11 (Saumweber et al. 1980; 1:200 dilution), Bj6 (Saumweber et al. 1980; 1:100 dilution) and E7 (Sigma-Aldrich, India; 1:200 dilution), respectively. The Western blots were developed with Horseradish peroxidase (HRP)-conjugated respective goat anti-rat HRP or goat anti-mouse HRP (Bangalore Genei, India) at 1:1000 dilution, using the Supersignal West Pico Chemiluminescent Substrate kit (Pierce, USA).

2.14 RNA isolation and RT-PCR

For semi-quantitative estimation of levels of the hsrω-n transcripts in wild type and hrp36 null larval SG, RT-PCR was carried out with total RNA isolated from 20 pairs of SG, using the TRI Reagent as per the manufacturer’s recommended protocol (Sigma-Aldrich, India). RNA pellets were resuspended in nuclease free water and incubated with 2U of RNase-free DNaseI (MBI Fermentas, USA) for 40 min at 37°C. First strand cDNA was synthesized using 2 μg of total RNA, 200U of Revert AidTM reverse transcriptase (MBI Fermentas, USA) and 80 pmol of oligo d(T)18 primer (New England Biolabs, USA). The samples were stored at −70°C until further use.

PCR was carried out for G3PDH transcripts, as loading control, using the forward 5′-CCACTGCCGAGGAGGTCAACTA-3′ and reverse 5′-GCTCAGGGTGATTGCGTATGCA-3′ primer pairs. The hsrω-n1 and hsrω-n2 (Mallik and Lakhotia 2011) cDNAs were co-amplified using forward primer P1 (5′-GGAAACAATGAAACCATACGC-3′) and reverse primer R1 (5′-TTGCGCTCACAGGAGATCAA-3′) pair. Another set of forward primer LP (5′-GGCAGACATACGTACACGTGGCAGCAT-3′) in combination with the R1 reverse primer (see above) was used to amplify the total hsrω-n cDNA (Mallik and Lakhotia 2011). The thermal cycling parameters included an initial 4 min denaturation at 94°C followed by 30 cycles of 30 s at 94°C, 30 s at 60°C and 45 s at 72°C. Final extension was carried out at 72°C for 10 min. The PCR products were electrophoresed on 1.5% agarose gel with a 100 bp ladder marker (Bangalore Genei, India).

2.15 RNA-RNA in situ hybridization

MT from wild type and hrp36 null late third instar larvae were dissected out in PSS, fixed in 4% paraformaldehyde for 2 min followed by incubation in 45% acetic acid for 4 min and partially squashed in same solution on poly-L-lysine-coated slide. After freezing in liquid nitrogen, the coverslip was flipped off and slide was dehydrated in 90% ethanol. FRISH was done with DIG-labelled antisense riboprobe from pDRM30 clone that specifically detects hsrω-n transcripts following Prasanth et al. (2000). Chromatin was counterstained with DAPI and fluorescence was examined under confocal microscope.

2.16 Confocal microscopy

Zeiss LSM510 Meta laser scanning confocal microscope was used with Plan-Apo 40X (1.3-NA) or 63X (1.4-NA) oil immersion objectives. Quantitative estimates of the proteins on different regions of polytene chromosome or in intact nuclei were obtained with the help of profile and histo options of the Zeiss LSM Meta 510 software. All images were assembled using the Adobe Photoshop 7.0 software.

3 Results

In this study we compared several phenotypes of wild type and hrp36 null (P{w + Tsr + }/P{w + Tsr + }; ry Df(3R)Hrb87F/ry Df(3R)Hrb87F) individuals. Since the hrp36 null genotype also carries the P{w + Tsr + } transgene, we examined possible effects of this transgene insertion on the phenotypes displayed by hrp36 null individuals in pilot experiments. For this purpose, the phenotypes of P{w + Tsr + } /P{w + Tsr + }; +/+ (only transgene insertion on chromosome 2), +/+; ry Df(3R)Hrb87F/ry Df(3R)Hrb87F, P{w + Tsr + }/P{w + Tsr + }; ry Df(3R)Hrb87F/ry Df(3R)Hrb87F and P{w + Tsr + }/P{w + Tsr + }; ry Df(3R)Hrb87F/+ individuals were compared. The +/+; ry Df(3R)Hrb87F/ry Df(3R)Hrb87F (without the Tsr transgene insertion on chromosome 2 but hrp36 null), (P{w + Tsr + }/P{w + Tsr + }; ry Df(3R)Hrb87F/ry Df(3R)Hrb87F (with the Tsr transgene insertion on chromosome 2 but hrp36 null) displayed comparable phenotypic changes, except that the former males were also sterile as reported earlier (Haynes et al. 1997). Therefore, in all subsequent studies, the (P{w + Tsr + }/P{w + Tsr + }; ry Df(3R)Hrb87F/ry Df(3R)Hrb87F) genotype was used as hrp36 null.

3.1 Absence of Hrp36 protein delays development and affects viability

We examined the time taken by hrp36 null and wild type from hatching of embryos to pupation when grown at 18°C, 24°C or at 30°C (±1°C). Eggs of the two genotypes that hatched within a period of 2 h were selected for monitoring their subsequent development at 24 h intervals. As the data in figure 1 show, there is some variation in the time required by different larvae of a given genotype to pupate. However, the hrp36 null larvae always started pupating much later than the wild type larvae. Interestingly, the time gap between first and last pupation was significantly greater in hrp36 null than in wild type and increased with lower rearing temperature (figure 1). A comparable delay in the emergence of adults was also noted for the hrp36 null pupae (not shown).

Figure 1
figure 1

Absence of Hrp36 protein substantially delays development. Graphical presentation of the time (in days, X-axis) of pupation after hatching of wild type and hrp36 null first instar larvae reared at 18°C (N = 289 and 212, respectively), 24°C (N = 826 and 651, respectively) or at 30°C (N = 578 and 165, respectively). The Y-axis shows mean percentages (±S.E., based on three replicates) of larvae that pupated till a given day.

Full size image

To examine if certain larval stages were specifically affected, we examined the number of teeth on mouth hooks of 24°C reared larvae at 24 h intervals after hatching. After 48 h of hatching, 100% of the wild type larvae (N = 25) reached second instar stage while only 20% of the hrp36 null larvae (N = 25) were at the second instar stage. After 72 h of hatching, 100% wild type larvae reached third instar stage compared to only 15% hrp36 null larvae. Only after 96 h, all the hrp36 null larvae were at the third instar stage. Thus, each instar’s duration is extended in hrp36 nulls.

Another interesting difference between the wild type and hrp36 null larvae was the site of pupation. 57% of the hrp36 null larvae (N = 212) pupated on the surface of food, while in the wild type only 3% larvae (N = 289) pupated on food and the rest pupated on drier surfaces away from the food.

As reported earlier by Haynes et al. (1997), no significant larval (figure 1) or pupal (not shown) lethality was observed in hrp36 nulls that were reared at 24°C or at 18°C. However, the hrp36 nulls grown at 24°C appeared a little weak and lethargic compared to the wild type, and with age, their weakness increased substantially. Life span assay revealed that the median life span of hrp36 null flies at 24°C was significantly less (23.9 ± 1.6, N = 250) than in the wild type (30.7 ± 1.6, N = 250) flies reared in parallel.

The hrp36 null pupae and adults were very sensitive to rearing at 30°C. Nearly 70% of 30°C reared hrp36 null larvae died as late pupae or pharates or soon after emergence (figure 2A and B). The surviving flies were very weak, unable to fly, laid very few eggs and died within a few days. Wild type larvae reared at 30°C did not show any significant pupal or adult death.

Figure 2
figure 2

The hrp36 nulls are sensitive to rearing at 30°C. (A) Graphical presentation showing mean (±S.E., 3 replicates) percentages of wild type (N = 578) and hrp36 null (N = 165) larvae (X-axis) that pupated and subsequently emerged as adult flies (Y-axis) when reared at 30°C ± 1°C. (B) Shows mean (±S.E., 20 replicates of 25 pupae each) percentages of hrp36 null pupae which died as late pupae (a), pharates (b), freshly emerged flies (c) or within 3 days after emergence (d) when reared from egg hatching onwards at 30°C ± 1°C.

Full size image

3.2 hrp36 null females have reduced fecundity and show abnormal ovarian follicles

Since the number of progeny in the hrp36 null stock appeared much less than in wild type stock grown in parallel and since the above data indicated little larval or pupal death of hrp36 null progeny, we examined to check if the fecundity of hrp36 null females was compromised. Examination of ovaries in wild type and hrp36 null females revealed that unlike in wild type females, one of the two ovaries in hrp36 null females was often much smaller than the other (figure 3A–C). The numbers of ovarioles (sum of the two ovaries) was also significantly less in hrp36 null females compared to that in same age (3 or 10 days old) wild type females (table 1).

Figure 3
figure 3

Absence of Hrp36 affects fecundity and oogenesis. (A) Graphical presentation of the numbers of ovarioles in each of the two ovaries in 3-day-old Oregon R + (left, N = 31) and hrp36 null (right, N = 31) females; the pairs of bars (blue and magenta) represent the numbers of ovarioles in each of the two ovaries of an individual fly; these are arranged in decreasing order keeping the ovary with lesser number of ovarioles (magenta bars) as reference. DAPI stained (white) ovaries of 7-day-old wild type (B) and hrp36 null (C) female flies. Note the fewer ovarioles (also see Table 1) and mature eggs and consequently much smaller size of one of the two ovaries in hrp36 null females. (D) Graphical presentation showing the mean numbers of eggs laid/female (fecundity) on different days (X axis) by wild type ♀ × wild type ♂ (stippled bars) and by hrp36 null ♀ × wild type ♂ (hatched bars). (E) Graphical presentation showing the mean percentage hatching of eggs laid by (a) WT ♂ × WT ♀ (93.9%), (b) WT ♂ × hrp36 null ♀ (83.4%), (c) hrp36 null ♂ × WT ♀ (91.5%) and (d) hrp36 null ♂ × hrp36 null ♀ (82.3%) parents. (F and G) Phalloidin stained (green) single ovarioles of wild type (F) and hrp36 null (G) flies showing higher numbers of circular muscle fibres in the epithelial sheath in the hrp36 null chambers at different stages (G= germarium, S1-S8, stage 1 to stage 8, respectively); on the basis of DAPI staining (not shown), the chamber behind the S7 in (G) appears apoptotic; insets at the lower right corners in (F) and (G) are higher magnifications projection images of top three optical sections of Phalloidin stained stage 6 chambers, while the insets on upper left corners show higher magnification image of single optical sections of the follicle cell layer and parts of the underlying nurse cells of Phalloidin stained stage 6 egg chambers of respective genotype; the red arrows point to the F-actin layer at the nurse cell face of follicle cell layer, while the white arrows point to the sectional view of circular muscle fibres in the external epithelial sheath. (H and I) Single confocal sections of ovarioles showing distribution of F-actin (green) and chromatin (DAPI stained, white) in different stages of egg chambers from wild type (H) and hrp36 null (I) ovaries: red arrow in (I) points to an apoptotic chamber in hrp36 null ovariole. The images in insets in (H) and (I) are high-magnification projection images of DAPI stained stage 7 nurse cell nuclei showing properly dispersed chromosomes in wild type but blob-like arrangement of chromatin in hrp36 null nurse cells.

Full size image
Table 1 Ovaries of hrp36 null females have fewer ovarioles than in wild type flies of comparable age
Full size table

We counted the numbers of eggs laid by wild type or hrp36 null females mated with wild type males. It was seen that between day 3 to day 9 of adult life, the hrp36 null females always laid fewer eggs than corresponding age wild type females (figure 3D). In another set, we compared the hatchability of eggs laid by wild type females that were mated with either wild type or hrp36 null males and by hrp36 null females that were mated with wild type or hrp36 null males. It was seen that eggs laid by hrp36 null mothers suffered a greater (about 17–18%) embryonic death irrespective of their fathers being wild type or hrp36 null males (figure 3E). It is interesting that the Df(3R)Hrb87F/+ zygotes suffered embryonic death only when the mother’s genotype was Df(3R)Hrb87F/Df(3R)Hrb87F since the Df(3R)Hrb87F/+ eggs laid by wild type females that were mated with Df(3R)Hrb87F/Df(3R)Hrb87F males did not show significant embryonic lethality (figure 3E). This clearly indicates a maternal effect of hrp36 null genotype.

Further examination of ovarioles revealed additional defects. Organization of F-actin in the thin circular muscles of the epithelial sheath surrounding each of the ovarioles in hrp36 null ovaries was altered. Unlike the thin and rather discontinuous actin filaments in the epithelial sheath covering the wild type ovarioles (figure 3F), continuous and a higher number actin filaments wrapped the entire ovariole in hrp36 null ovaries (figure 3G). The F-actin layer at the interface of follicle and nurse cells in hrp36 null chambers was thinner than in the wild type egg chambers (see insets in figure 3 F and G). Further, many of the hrp36 null ovarioles did not contain egg chambers beyond stage 7–8 (figure 3G), possibly due to a high incidence of apoptosis in hrp36 null flies (figure 3G and I). Ring canals appeared normal in the mutant egg chambers.

Organization of chromatin in the hrp36 null ovarioles was also affected. As reported by King (1970) the highly endoreplicated chromosomes of nurse cell nuclei in stage 6 and later chambers appear as fully dispersed chromatin (inset in figure 3H). However, in hrp36 null egg chambers, the highly endoreplicated chromatin remained more condensed in blob-like structures even in later stages (inset in figure 3I), similar to that reported earlier for Squid, hrb27C or otu mutants (King et al. 1981; King and Storto 1988; Goodrich et al. 2004).

3.3 Malpighian tubules are enlarged in width with higher levels of polyteny while salivary gland chromosomes show distinct telomere elongation in hrp36 null larvae

The MT of hrp36 null larvae were nearly two times thicker (figure 4C, D, E) than in the wild type (figure 4A, B, E). Correspondingly, the nuclei in the polytenized principal cells of MT of hrp36 null larvae were also nearly two times larger than in Oregon R + larvae of comparable stage (figure 4F–H). Squash preparations of late third instar hrp36 null larval MT revealed well-spread polytene chromosomes with distinct band/interband organization (figure 4J) compared to the poorly polytenized chromosomes seen in wild type MT (figure 4I). Total DAPI fluorescence of the principal cell nuclei in hrp36 null larval MT was also greater than in the corresponding wild type MT (figure 4K). The size of the chromocentre, which is known to be not involved in endoreplication (Lakhotia 1974, 1984), was comparable in wild type and hrp36 null larvae. The sizes of polytene chromosomes in SG were comparable between wild type and hrp36 null (figure 4L-N). Their total DAPI fluorescence values were also comparable (data not presented). Interestingly, however, all the polytene chromosome arms in all hrp36 null larval SG nuclei showed distinct elongation of telomeres when compared with wild type chromosome arms tips and the extended telomeres in hrp36 nuclei displayed occasional association of different telomeric regions (figure 4L–N).

Figure 4
figure 4

Malpighian tubules (MTs) are enlarged because of increased polyteny of the principal cells in hrp36 null larvae. Phase contrast images of MT from wild type (A and B) and hrp36 null (C and D) at lower (A and C) and higher (B and D) magnifications. Graphical presentation of mean (±S.E, N = 25 for each value) width of MTs from Oregon R + and hrp36 null (E) larvae. Confocal projection images of DAPI-stained MT principal cell nuclei in intact MTs from Oregon R + (F) and hrp36 null (G) late third instar larvae. Graphical presentation of mean (±S.E, N = 50 for each value) nuclear diameter of polytene nuclei from Oregon R + and hrp36 null larval MT (H). Confocal images of DAPI-stained squash preparations of polytene chromosomes in MT principal cells from Oregon R + (I) and hrp36 null (J) late third instar larvae. Graphical presentation of the mean (±S.E, N = 8 for each value) total nuclear DAPI fluorescence of polytene nuclei from wild type and hrp36 null larval MT (K). Salivary gland polytene chromosomes of wild type (L) and hrp36 null (M) larvae showing telomere association (arrow) and telomere elongation in hrp36 null; tips of different chromosome arms (X, 2L, 2R, 3L and 3R) are marked. Magnified images of pairs of telomeric regions of the different long arms (X, 2L, 2R, 3L and 3R) from wild type (left) and hrp36 null (right) chromosome spreads in (L) and (M), respectively, are presented in (N) to show the substantially elongated telomeres in hrp36 null chromosomes.

Full size image

3.4 Hrp36 is essential for organization of the omega speckles

The Hrp36 is one of the major hnRNPs which is physically associated with the non-coding hsrω-n transcripts in the nucleoplasmic omega speckles (Prasanth et al. 2000; Lakhotia et al. 2012). In order to find out the role of Hrp36 in organization of omega speckles, we examined nuclear distribution of hsrω-n transcripts and the omega speckle interacting proteins like NonA (Onorati et al. 2011) and Bancal/Hrb57A (Prasanth et al. 2000) in the principal cells of MT in wild type and hrp36 null late third instar larvae.

In situ hybridization of hsrω-n specific riboprobe to MT cells revealed that unlike the specific localization of the hsrω-n transcripts in the nucleoplasmic omega speckles and the site of transcription in wild type MT nuclei (figure 5A and B), the hrp36 null MT cells displayed few omega speckles. Instead, the hsrω-n transcripts in these cells were distributed more or less in a diffuse manner in the nucleoplasm with a prominent cluster at the site of transcription (figure 5C and D).

Figure 5
figure 5

Typical omega speckles do not form in the absence of Hrp36. Confocal projection images to show in situ localization of hsrɷ-n RNA (red, A–D), NonA (red, E–H ) and Bancal (red I–L) in Malpighian tubule principal cell nuclei of different genotypes (noted above columns). Chromatin, counterstained with DAPI (blue), is seen in the merged images in B, D, F, H, J and L.

Full size image

Immunostaining of NonA and Bancal proteins also revealed almost complete disruption of the omega speckles in MT nuclei of hrp36 null larvae. Both these proteins were, like the hsrω-n transcripts, diffused through the nucleoplasm (figure 5G, H, K and L) rather than as individual small nucleoplasmic speckles seen in wild type (figure 5E, F, I and J). In addition, these proteins were also present in small clusters on chromatin regions. Similar to MT principal cell nuclei, speckled pattern of these hnRNPs was also disrupted in MT stellate and other polytene and diploid cell nuclei in gut (not shown).

3.5 Absence of Hrp36 affects localization of nuclear matrix protein Megator and chromatin remodelling protein ISWI in MT nuclei

Earlier studies have reported a role of hnRNP A in organization of the nuclear matrix (He et al. 1991; Zimowska et al. 1997) in higher eukaryotes. Therefore, we compared the localization of a nuclear matrix protein Megator in MT principal cell nuclei of hrp36 null and wild type. Interestingly, while in wild type nuclei, the Megator was seen in small clusters of speckles in the nucleoplasm (figure 6A), in the hrp36 null cells (figure 6B), this protein was present in larger clusters in the interchromatin space with less distinct punctate distribution (figure 6B). However, the presence of Megator as a rim at the nuclear periphery was similar in the hrp36 null and wild type nuclei (figure 6A and B).

Figure 6
figure 6

Absence of Hrp36 affects the localization and abundance of nuclear matrix and chromatin remodelling proteins. Confocal projection images showing immunolocalization of Megator (red, A and B) and ISWI (red, D and E) in unstressed MT principal cell nuclei of wild type (A and D) and hrp36 null (B and E) larvae. DNA is stained with DAPI (blue). (C) Graphical representation of the relative levels of Megator (excluding the peripheral nuclear rim) in MT nuclei of wild type and hrp36 null larvae, expressed as ratios of Megator and DAPI fluorescence in a nucleus (N = 60 optical medial sections from 20 nuclei for each genotype). (F) Graphical representation of the relative levels of ISWI in MT nuclei of wild type and hrp36 null larvae, expressed as ratios of ISWI and DAPI fluorescence in a nucleus (N = 20 for each genotype).

Full size image

The nucleoplasmic distribution of ISWI protein, a component of chromatin remodelling complexes, was also affected in hrp36 null cells since its punctated distribution in nucleoplasm in wild type nuclei (figure 6D) was not seen in hrp36 null nuclei (figure 6E).

The immunofluorescence for Megator as well as ISWI proteins in the hrp36 null MT nuclei appeared to be much greater than in the wild type nuclei (figure 6A, B, D and E). In order to know if the enhanced immunostaining for these proteins in hrp36 null principal cell nuclei was correlated with their increased DNA content, we compared the ratios of the nuclear Megator (excluding the nuclear rim associated fluorescence) or ISWI and nuclear DNA (DAPI fluorescence) in wild type and hrp36 null MT nuclei (figure 6C and F). This comparison clearly revealed that compared to the wild type, the increase in levels of Megator and ISWI in principal cells of MT of hrp36 null larvae is much more than the increase in DNA.

3.6 Absence of Hrp36 affects starvation and thermal stress tolerance

3.6.1 Starvation stress

Three-day-old Oregon R + and hrp36 null flies were starved and their survival monitored at 8 h intervals. As seen from survival curves in figure 7A, death of starved hrp36 null flies started as early as 24 h of starvation, while the Oregon R + flies began to die only after 48 h of starvation. After 50 h of starvation, ~90% hrp36 null flies were dead while in the case of Oregon R + flies, a comparable percentage of death was observed after ~90 h of starvation (figure 7A).

Figure 7
figure 7

The hrp36 null flies are highly sensitive to starvation and heat stresses. (A) Survival curves of flies of Oregon R + (N = 172) and hrp36 null (N = 297) flies measured at 8 h intervals after beginning of the starvation treatment. The mean (±S.E.) % surviving flies (Y-axis) for each time point (X-axis) is based on two replicates. (B) Graphical presentation of mean (±S.E., based on 3 replicates) % of Oregon R + (N = 200) and hrp36 null (N = 200) larvae that emerged as young flies after heat shock for 1 or 2 h (X-axis) at late third instar stage. (C) Graphical presentation showing mean (±S.E., based on 2 replicates) percentage of 3-day-old Oregon R + (N = 220) and hrp36 null (N = 322) flies that were knocked down just after heat shock (left) and of those surviving for the next 5 days (right). The mean percentage values (Y-axis) are based on two replicates.

Full size image

3.6.2 Thermo-tolerance

To examine a role of Hrp36 in thermo-tolerance, actively wandering third instar wild type and hrp36 null larvae were heat-shocked at 37°C for 1 or 2 h and their survival was monitored until the emergence of flies. Data from three replicates revealed that hrp36 nulls are more thermo-sensitive since following 1 h exposure to 37°C, only 65% of the larvae survived to adulthood compared to emergence of 97% of similarly treated wild type larvae. Likewise, only 25% of hrp36 null larvae exposed to 37°C for 2 h emerged as flies in contrast to emergence of 88% wild type larvae (figure 7B). In the absence of Hrp36 protein, adult flies also showed greater thermo-sensitivity since after 1 h of heat shock at 37°C all the hrp36 null flies were knocked down compared to the knock down of only 26% of the wild type flies (figure 7C). Five days after the heat shock, compared to 95% survival of wild type flies, only 53% hrp36 null flies survived (figure 7C).

3.7 Absence of Hrp36 protein does not affect the heat shock induced synthesis of Hsp70

In order to see if the heat shock response was compromised because of the complete absence of Hrp36 protein, we examined the inducibility of Hsp70 in response to heat shock. As shown in figure 8, the status of Hsp70 in control, heat shocked and 2 h recovered wild type and hrp36 null larvae was comparable. As noted below, the different heat shock puffs were also comparably induced in wild type and hrp36 null larval SG (figure 9).

Figure 8
figure 8

Levels of the stress inducible HSP70 (upper panel) are comparable in wild type and hrp36 null. Western blot of total larval protein challenged with the 7Fb antibody to detect stress-inducible Hsp70 in control (C), heat shocked larvae (HS) and in those that recovered from heat shock for 2 h (R). The lower panel shows the levels of β-tubulin in the corresponding samples.

Full size image
Figure 9
figure 9

Immunolocalization of NonA protein (red) on salivary gland polytene chromosome (white) in wild type and hrp36 null (mentioned above each column) during control (A and B), 40 min heat shock at 37°C (C and D) and 2 h recovery from 40 min heat shock at 37°C (E and F) conditions. The white arrow indicates 93D locus, while red and yellow arrows indicate the 87A and 87C loci, respectively. Scale bar represents 20 μm. (G) Graphical presentation of the amount of NonA on polytene chromosome 3R segment, 93D, 87A and 87C puffs immediately after heat shock and on the 3R segment and the 93D puff after 2 h recovery from heat shock in wild type and hrp36 null chromosome spreads (N = 12 for each genotype). (H) Western blot showing the levels of NonA (top row) and Hrp36 (middle row) proteins in wild type and hrp36 null salivary glands in control (C), after heat shock (HS), and after 2 h recovery from heat shock (R). Levels of β tubulin, used as loading control, are seen in the lower row. As expected, Hrp36 is completely absent in hrp36 null in all given condition.

Full size image

3.8 Absence of Hrp36 protein affects the heat-shock-induced redistribution of omega-speckle-interacting proteins

Heat shock results in disappearance of the nucleoplasmic omega speckles and at the same time, most of the hnRNPs and other proteins associated with omega speckles and chromosome regions get relocalized to the 93D or hsrω site (Prasanth et al. 2000). To investigate the role of Hrp36 in this mobilization, we examined localization of NonA, which is one of the omega-speckle-interacting proteins (Onorati et al. 2011), on salivary gland polytene chromosome spreads from actively wandering third instar wild type and hrp36 null larvae. Under unstressed condition, the amount of NonA protein at the 93D locus in hrp36 null polytene chromosomes was distinctly less than that in wild type (figure 9A and B). Following heat shock, the NonA protein moved away from most of the chromosomal sites in wild type cells and strongly accumulated at the 93D puff; after 2 h recovery from heat shock the distribution of NonA resembled the pre-stress condition (figure 9C and E). It is notable that the level of NonA protein at the 93D site in heat-shocked hrp36 null cells was always much less than in wild type cells, while that on the other chromosomal regions was greater than in heat shocked wild-type chromosomes (figure 9D and F). We quantitated the amount of NonA protein at different cytological regions including the 93D, 87C and 87A puffs and a segment of right arm of chromosome 3 (100F to 93E region) in polytene chromosome spreads of wild type and hrp36 null SG following heat shock and after 2 h recovery from heat shock. As shown in figure 9G, while the level of NonA at 93D puff in heat-shocked hrp36 null cells was approximately 50% of that in wild type cells, it was much more abundant at the 87A and 87C puffs and the 100F to 93E chromosome region in hrp36 null than in the wild type. Interestingly, after a 2 h recovery from heat shock, the levels of NonA on the 3R segment were comparable in hrp36 null and wild type cells but the 93D site showed much less fluorescence. Compromised clustering of RNA-processing proteins during heat shock was also observed in other polytene as well as diploid cell nuclei of hrp36 null (not shown).

In order to see if the varying levels of NonA at the 93D puff site and other chromosome regions was related to any difference in the total cellular levels of this protein, we estimated NonA and Hrp36 in wild type and hrp36 null control and heat-shocked larvae and in those that had recovered for 2 h from the heat shock through Western blotting. The levels of NonA remained constant and equal in all conditions in wild type and hrp36 null larvae (figure 9H). The levels of Hrp36 in wild type larvae also did not change under the different experimental condition and, as expected, the hrp36 null larval samples did not show any trace of signal in any of the lanes.

We also examined the localization of two chromatin remodelling proteins, viz heterochromatin protein HP1 and ISWI, on heat-shock-induced 87A, 87C and 93D puffs in wild type and hrp36 null polytene chromosome spreads. As reported earlier (Lakhotia et al. 2012), HP1 protein is present at the 93D site in control cells (not shown). Heat shock causes a strong accumulation of HP1 at the 93D puff in wild type (figure 10A). Significantly, while the level of HP1 in control hrp36 null polytene cells was comparable to that in wild type control cells, in the heat-shocked hrp36 null cells, much less HP1 protein was present at the 93D puff than in wild type (figure 10B). As also reported earlier (Lakhotia et al. 2012), HP1 was found to be present in very small amounts at the 87A and 87C puffs in wild type as well as the hrp36 null cells (figure 10A and B).

Figure 10
figure 10

Confocal images showing the localization of HP1 (green) and ISWI (red) on salivary gland polytene chromosome (white) after 40 min of heat shock in wild type (A and C) and hrp36 null (B and D). Higher magnification images are shown in insets. White, yellow and red arrows indicate 93D, 87C and 87A site respectively. Immunolocalization of NonA (E), HP1 (F) and ISWI (G) after heat shock (upper panel) and heat shock followed by RNase treatment (lower panel) at 93D locus in wild type. Accumulation of HP1 and NonA is dramatically disappeared while ISWI equally accumulated even after RNase treatment.

Full size image

In agreement with a previous report (Sala et al. 2008), ISWI was detectably present at the 87A and 87C sites in unstressed wild type and control cells (not shown) but was completely absent in both genotypes at the heat-shock-induced 87A and 87C puffs (figure 10C and D). ISWI protein was distinctly seen at the 93D site in control cells of both genotypes (not shown), but unlike its disappearance at the 87A and 87C puffs, its levels at the 93D puff increased significantly after heat shock in wild type as well as hrp36 null polytene cells. Interestingly, unlike the hnRNPs and HP1, the strong accumulation of ISWI at the 93D puff following heat shock was comparable in wild type and hrp36 null larvae. ISWI protein was present at low levels at the other heat shock gene sites in control cells of both genotypes with no appreciable change detectable in any of the genotypes following the heat shock.

In view of the comparable levels of ISWI at the 93D puff site in hrp36 null and wild type polytene chromosomes but much less accumulation of NonA and HP1 after heat shock in the former, we examined if the association of these proteins at 93D locus is sensitive to RNase. Interestingly, RNase treatment of live heat-shocked glands removed most of the NonA and HP1 proteins from the 93D puff site while immunostaining for ISWI at the 93D puff site was not affected by the RNAse treatment (figure 10E–G). This shows that while NonA and HP1 accumulate through their association with RNA at the 93D site after heat shock, ISWI directly binds with DNA at 93D locus.

3.9 Levels of hsrω-n transcripts are enhanced by heat shock equally in wild type and hrp36 null larval SG but the splicing of the hsrω-n1 transcripts is affected in absence of Hrp36

Semi-quantitative RT-PCR to assess the levels of hsrω-n in control and heat-shocked wild type and hrp36 null larval SG showed almost equal enhancement following heat shock in both genotypes (figure 11). Interestingly, however, while in wild type heat-shocked glands, the spliced hsrω-n2 transcripts were more abundant than the unspliced hsrω-n1 species, in heat-shocked hrp36 null SG, the unspliced form was more abundant than the spliced hsrω-n2 form, suggesting that in the absence of Hrp36, splicing of hsrω-n1 transcripts is compromised in heat shocked SG cells (figure 11).

Figure 11
figure 11

Absence of Hrp36 affects splicing of hsrω-n transcripts. RT-PCR amplicons showing the levels of total hsrω-n and its unspliced hsrω-n1 and spliced hsrω-n2 transcripts in control (Con) and after heat shock (HS) in wild type and hrp36 null salivary glands. G3PDH amplicons (lowest row) of corresponding samples were used as loading control to normalize the amount of hsrω-n transcripts.

Full size image

4 Discussion

The Df(3R)Hrb87F is a small deletion that removes most of the Hrp36-encoding gene and an adjacent gene, Tsr, which is essential for male fertility (Zu et al. 1996; Haynes et al. 1997). Haynes et al. (1997) reported that Df(3R)Hrb87F homozygous flies are viable, and females fertile but males sterile. The sterility of Df(3R)Hrb87F homozygous males is rescued by a P[Tsr] transgene inserted on chromosome 2 (Haynes et al. 1997) so that P[Tsr]/P[Tsr]; Df(3R)Hrb87F/Df(3R)Hrb87F (hrp36 null) stock can be maintained. Survival of organisms without the abundantly and universally expressed Hrp36 protein was suggested to be due to the functional redundancy with Hrp38, which shows 76% identity to RNP (RRM1) domain and 67% identity to the glycine-rich domain (GRD) of Hrp36 (Haynes et al. 1990, 1991; Zu et al. 1996). Although the initial studies by Haynes et al. (1997) indicated that the hrp36 null individuals are without any apparent phenotype and viable, an earlier study in our laboratory (Sengupta and Lakhotia 2006) reported that eyes of hrp36 null individuals show roughening due to loss of photoreceptors and that the Df(3R)Hrb87F acts as a dominant enhancer of neurodegeneration caused by expanded polyQ proteins in fly models. Our present study reveals that hrp36 null individuals suffer from several other sub-lethal phenotypic consequences like developmental delay, altered site selection for pupation, significantly compromised female fecundity, reduced starvation and thermo-tolerance, etc. None of these phenotypes are due to the absence of the endogenous Tsr gene or to the presence of the P[Tsr] transgene insertion on chromosome 2 since we found that individuals homozygous for Df(3R)Hrb87F but not carrying the P[Tsr] transgene on chromosome 2 also displayed comparable defects (data not presented), although such males were, as expected, sterile (Haynes et al. 1997). It may be mentioned here that Manita et al. (2011) did not report any delay in development of Hrb87F KG 02089 and Hrb87F BG02743 mutants because they compared the number of wild type and the Hrb87F mutant flies that had eclosed by day 15 without considering earlier time points. When grown at 24°C, all hrp36 null pupae emerge by day 15, while all the Oregon R + flies emerge by day 12 after embryo hatching. Therefore, Manita et al. (2011) would have missed the developmental delay seen in hrp36 nulls. Thus, we suggest that notwithstanding the high degree of homology between Hrp36 and Hrp38 (Haynes et al. 1991, 1997), Hrp36 is not fully dispensable but is essential for normal development and female fecundity and, more importantly, for surviving stresses.

The nucleoplasmic omega speckles, organized by the nucleus limited long noncoding hsrω-n transcripts, help regulate the functional availability of a variety of hnRNPs and some other RNA-processing proteins (Lakhotia et al. 1999; Prasanth et al. 2000; Jolly and Lakhotia 2006; Mallik and Lakhotia 2011; Lakhotia 2011). Since we found the speckled pattern of nucleoplasmic hsrω-n RNA and the omega-speckle-associated proteins like NonA and Bancal to be nearly completely disrupted, it appears that, like the hsrω-n transcripts (Prasanth et al. 2000; Mallik and Lakhotia 2011), Hrp36 is also a core constituent of omega speckles. A physical interaction of Hrp36 with hsrω transcripts in omega speckles was shown through the presence of these transcripts in Hrp36-immunoprecipitate (Prasanth et al. 2000; Mallik and Lakhotia 2011). The RBD-2 and GRD of Hrp36 are reported to show high affinity binding with UAGGUUAGG, which is very similar to the conserved nonamer AUAGGUAGG present at ~100 b intervals in the repeats present in the hsrω-n transcripts (Mayeda et al. 1998; Zu et al. 1998). Thus, it appears that Hrp36 directly binds with the hsrω-n transcripts and the two together provide the core platform for other proteins to assemble at omega speckles. However, proteins like NonA can still associate with the hsrω transcripts since the heat-shock-induced accumulation at the 93D puff in hrp36 null cells, although less than in cells that have Hrp36, was found to be sensitive to RNAse.

The observed effect on the distribution and abundance of the nuclear-matrix-associated protein like Megator may be a consequence of the well-known association of hnRNP A with the nuclear matrix (He et al. 1991; Zimowska et al. 1997). The enhanced levels of Megator in nuclear matrix in the endoreplicated MT nuclei does not appear to be simply due to the increased DNA content in hrp36 null MT cells since the ratio of fluorescence intensities of nucleoplasmic Megator and DNA (DAPI) in the hrp36 null principal cells was several folds greater than in corresponding wild type MT. The disruption of omega speckles in hrp36 null cells may also be related to the altered organization of nuclear matrix since these speckles are associated with the nuclear matrix (our unpublished observations). Additionally, the observed increase in the ISWI content in hrp36 nulls may also affect omega speckles since this chromatin remodeller is known to be involved in organizing omega speckles (Onorati et al. 2011). These possibilities need further studies.

In the absence of organized omega speckles, the regulated availability of a variety of hnRNPs and other RNA-processing proteins for different cellular events in all body cells is expected to be compromised (Mallik and Lakhotia 2011; Lakhotia et al. 2012). Such a global effect of the absence of organized omega speckles on the orderly availability of nuclear RNA-processing proteins and the above-noted involvement of Hrp36 in processing of a large number of nuclear RNA species is expected to affect timely progression of several cellular processes, which cumulatively result in varyingly delayed development of hrp36 null organisms. Ji and Tulin (2012) recently reported that depletion of Hrp38 causes substantial pre-pupal lethality. On the other hand, we found that a complete absence of Hrp36 slows down the development but does not affect larval or pupal survival, although the adult life span is reduced. Therefore, it appears that because of the buffering of Hrp36 functions by the other hnRNP A1 homologs, its complete absence does not cause larval or pupal death under normal developmental conditions but still has a sub-lethal effect.

The Hrp36, together with hsrω-n RNA and omega-speckle-associated proteins like the Hrp40/Squid, Hrp48/Hrb27C, etc., is abundantly expressed during oogenesis (Mutsuddi and Lakhotia 1995; Lakhotia et al. 2001; Norvell et al. 2005). Although a number of hnRNPs have been shown to perform specific functions during oogenesis in Drosophila, including localization and translation of antero-posterior axis determining transcripts and proteins (Goodrich et al. 2004; Steinhauer and Kalderon 2005; Norvell et al. 2005; Geng and MacDonald 2006; Clouse et al. 2008; Ji and Tulin 2012), a specific role of Hrp36 in ovary development and oogenesis is not yet known. Our results show that in the absence of Hrp36, the overall development of ovaries is affected so that fewer ovarioles are formed and among them also, many of the follicles fail to complete differentiation resulting in the overall reduced fecundity. Involvement of Hrp36 in early differentiation of ovary is also indicated by the unequal size of two ovaries in hrp36 null females. It appears that the other related hnRNPs like Hrp38, Hrp40 and Hrp48 cannot completely compensate for the complete absence of Hrp36 so that under the limiting conditions, only one of the developing ovaries is able to attain a threshold state for differentiation of ovarioles to some extent. A global overexpression of hsrω transcripts because of the Act5C-GAL4-driven expression of the EP3037 allele of hsrω, which sequesters a greater amount of hnRNPs in omega speckle clusters resulting in their functional depletion (Mallik and Lakhotia 2009, 2011), also results in unequal size of the two ovaries with greatly reduced numbers of follicles and mature eggs (Soundarya Iyer, Moushami Mallik and SC Lakhotia, unpublished). Significantly, depletion of Hrp38 is not reported to result in development of unequal sized ovaries although the stem cell maintenance and oocyte location are affected (Ji and Tulin 2012). Hrp36, therefore, appears to be involved in other events of early steps of development and differentiation of ovarioles. The observed alterations in the F-actin in the hrp36 null ovarioles may also contribute to mis-organized egg chambers. The failure of chromatin dispersal in hrp36 null nurse cells older than stage 5 (King 1970; Dej and Spradling 1999) is comparable to that seen in Hrb27C, Squid or otu mutants (King et al. 1981; King and Storto 1988; Goodrich et al. 2004). Therefore, it appears that these proteins may act in concert to bring about the dispersal of nurse cell chromatin before the stage 6. All these factors together contribute to poor growth and consequent apoptosis of egg chambers at stage 7–8. Although the hrp36 null egg chambers that progress beyond stage 7–8, and develop to apparently normal-looking eggs, the observed maternal effect leading to embryonic death of ~17–18% of eggs laid by hrp36 null mothers clearly indicates a continuing role of Hrp36 protein in later stages of oogenesis, which can only be partially compensated by the other hnRNPs. Further studies are required to understand the specific roles of Hrp36 in differentiation of ovaries and egg chambers.

It is unlikely that the additional endoreplication cycles in MT cells is a direct consequence of the longer period of development of the hrp36 null larvae since several other mutations that prolong the larval period are not known to show enhanced levels of polyteny in MT cells. Since a unique feature of larval MT is that it continues as such in adults without histolysis during pupal metamorphosis, the regulation of endoreplication cycles in MT is different from that in the other endoreplicating larval tissues which histolyse following the pupal moult. Enhanced polyteny and consequent larger nuclear size in principal cells of MT nuclei and persistence of blob-like forms of endoreplicated advanced stage nurse cell nuclei suggest a possible role of Hrp36 in higher-order chromatin organization in at least some of the endoreplicating nuclei. In this context, it is significant that Manita et al. (2011) have shown a direct interaction between Hrp36 and DNA topoisomerase I. The cytologically analysable polytene chromosomes in MT of hrp36 null individuals provide, besides the classically used salivary gland polytene chromosomes, an additional tissue for studies on gene expression at chromosomal level in Drosophila.

The substantial extension of telomeres of most chromosomes in hrp36 null larval SG is remarkable. Hrp36 and other hnRNPs are known to be present at polytene telomeres (Lakhotia et al. 2012) and hnRNPs are known to be involved in telomere organization in mammalian cells (He and Smith 2009). Further studies on the role of Hrp36 in telomere maintenance in Drosophila are currently in progress and would be reported elsewhere.

It is significant that as observed earlier in the case of hsrω-deficient or hsrω-overexpressing individuals (Lakhotia et al. 2012), the hrp36 null larvae and adults also display high sensitivity to elevated temperature in spite of the characteristic inducion of the heat shock puffs and the induced synthesis of Hsp70. Besides the induction of heat shock genes, the well-documented dynamic relocalization of omega speckles and their associated hnRNPs following heat shock and during subsequent recovery from the stress (Lakhotia et al. 1999; Prasanth et al. 2000) is equally essential for surviving the thermal stress (Lakhotia et al. 2012). Although the relocalization of omega-speckle-associated NonA or the HP1 protein on general chromosome regions following recovery from heat shock was not as severely affected in the hrp36 null cells as when the omega speckles were disrupted because of down-regulation or absence of the hsrω-n transcripts (Lakhotia et al. 2012), the association of these proteins at the hsrω gene locus and the reformation of omega speckles were affected in hrp36 null cells. Therefore, Hrp36, like the hsrω-n transcripts (Lakhotia et al. 2012), has an essential role in the dynamic redistribution of proteins like NonA and HP1 following stress and recovery from stress, and therefore, its absence results in high sensitivity to stress.

Although ISWI, an important member of chromatin remodelling complexes, is required for assembly of typical omega speckles (Onorati et al. 2011), its accumulation at the hsrω gene site (93D puff) following heat shock was not found to be affected by the absence of Hrp36. Since RNAse treatment of live SG did not affect its presence at the 93D site, ISWI seems to be associated with chromatin rather than with the heat-shock-induced hsrω transcripts with which HP1 and NonA proteins are bound. Sala et al. (2008) reported that ISWI, which normally binds with the uninduced hsp70 DNA at 87A and 87C sites, gets displaced as chromatin contents of these puffs get increasingly poly-ADP-ribosylated (PARylated) following heat shock. In contrast, the ISWI protein’s level at the 93D site was more abundant at the 93D puff in heat-shocked cells (figure 10C, D and G). Interestingly, we noticed that ISWI is also present at heat shock puffs other than the 87A and 87C (details not shown, but see figure 10C and D). This difference in the association of ISWI with heat-shock-induced 87A/87C puffs on one hand and the 93D and other heat shock puffs on the other, requires further study.

Ji and Tulin (2009) suggested that heat-shock-induced PARylation of chromatin associated hnRNPs causes these hnRNPs to come to nucleoplasm where they get de-PARylated and thus finally get sequestered at the 93D puff site because the de-PARylated hnRNPs associate with hsrω-n transcripts. In this context, it remains to be examined if the reduced levels of proteins like NonA at the 93D puff after heat shock in the hrp36 null cells is due to inefficient de-PARylation of these proteins in the absence of Hrp36 or due to a direct role of Hrp36 in the movement of other hnRNPs and NonA, etc., to the 93D puff after heat shock.

In agreement with the report of Ji and Tulin (2009) that in poly(ADP-ribose) glycohydrolase (PARG)-deficient cells, which accumulate much less hnRNPs at the 93D puff after heat shock, the splicing of hsrω nuclear transcripts to remove the intron is affected in hrp36 null cells as well. It may be noted here that while the RT-PCR primers used in Ji and Tulin’s (2009) study could not discriminate between splicing of the hsrω-n (hsrω-RB) and hsrω-pre-c (hsrω-RA) transcripts, the primers designed by Mallik and Lakhotia (2011) and used in the present study unambiguously identify that splicing of hsrω-n (hsrω-RB) transcript is affected in the hrp36 null cells. The effect on splicing on the hsrω-n transcripts in hrp36 null cells agrees with the known role of Hrp36 in alternative splicing (Zu et al. 1996, 1998; Olson et al. 2007; Borah et al. 2009; Blanchette et al. 2009; Nilsen and Graveley 2010). The functional consequences of retention of intron in the hsrω-n1 transcript on recovery from heat shock remain to be analysed.

A role of Hrp36 in starvation tolerance is indicated by our studies, although the underlying mechanism is not clear. It is to be expected that starvation stress would affect cell physiology, gene expression and RNA processing in many ways, some or all of which may involve the hnRNPs. For example Griffith et al. (2006) have shown a role of hnRNP A and some other hnRNPs in regulating splicing of G6PD during starvation in mice. In another study in our lab (Dwivedi et al. 2012), an association has been found between enhanced levels of hnRNPs (including Hrp36), faster development and improved tolerance against starvation and thermal stresses. Present results are in agreement with these findings since absence of Hrp36 is associated with delayed development and poor stress tolerance.

In view of the very wide variety of functions mediated by hnRNP A1 family members in eukaryotic cells (Martinez-Contreras et al. 2007; He and Smith 2009; Chaudhury et al. 2010; Han et al. 2010; Nilsen and Graveley 2010), it is not surprising that a complete absence of the Hrp36 has such pleiotropic, although sub-lethal, phenotypic consequences. Interestingly, while its complete absence does not have lethal consequences under normal laboratory developmental conditions presumably because of the overlapping functions provided by related hnRNPs like Hrp38, Hrp40 and Hrp48, the hrp36 null condition is highly detrimental under unfavourable conditions like starvation and thermal stresses. We believe that this essential requirement of Hrp36 under conditions of stress is primarily because of its interaction with the hsrω-n transcripts to organize the omega speckles and facilitate the recruitment of other proteins to omega speckles. Since environmental stresses exert strong selection force under natural conditions, and since the roles of Hrp36 under stress cannot be complemented by the other related hnRNPs, the hrp36 gene is retained in the fly genome. Persistence of multiple members of a given family of hnRNP in higher eukaryotes indicates that while they share common structural and functional properties, each one also performs certain unique sets of functions that are essential for the fine tuning of cell’s activities in relation to requirements of differentiation and environmental conditions.