Dm-myb mutant lethality in Drosophila is dependent upon mip130: positive and negative regulation of DNA replication

mip130 mutants are viable and show decreased DmMyb stability

We used “ends in” gene targeting to generate mutant alleles in the largest subunit of the DmMyb complex, mip130 (Rong and Golic 2000, 2001; Rong et al. 2002). To do this, we engineered an I-SceI restriction endonuclease site and point mutations in the coding region of mip130 on either side of the I-SceI site to produce truncated forms of Mip130 upon targeted integration. The mutagenesis required step-wise strain construction: Transgenic animals carrying the mutant mip130 construct within a P element (that also harbored binding sites for FLP recombinase flanking the insert) were first generated by standard methods. The mip130 mutant was then obtained after crossing this P element-harboring line to animals that express both the I-SceI restriction enzyme and FLP recombinase. Expression of both FLP and I-SceI in an individual resulted in excision of the donor sequences from the P element and generated a double-stranded DNA break within mip130. The double-stranded DNA break stimulated homologous recombination between the donor DNA and the corresponding chromosomal target locus. The recombination resulted in two mutant copies of mip130 at the normal genomic locus that were separated by intervening sequences as outlined in Figure 1 (mip130^{1-723 + 1-36}). We engineered point mutations in the donor construct such that the left copy of mip130 contained a new EcoRI site 2136 nucleotides from the start codon that could produce a 723-amino acid, C-terminal truncated protein. The right copy contained a new BamHI site 45 nucleotides from the start codon that could produce a 36-amino acid protein (of which the last 21 amino acids were not Mip130 sequence). Of two potential candidates, we found that one gave the correct restriction endonuclease pattern for a targeted insertion at the proper location on the X chromosome for mip130 (Fig. 1B).

Because one gene copy at the mip130^{1-723 + 1-36} locus can encode a truncated Mip130 protein of substantial size, we reduced the tandem duplication to a single copy in order to generate a true null mutant. We expressed I-CreI endonuclease in the mip130^{1-723 + 1-36} heterozygous mutant animals to generate a double-stranded DNA break at the specific I-CreI-site in the locus (Fig. 1A). Repair of the break by recombination between the homologous regions flanking the double-stranded DNA break can eliminate the intervening sequences, including the marker gene, thereby providing a simple eye-color selection to detect reduction events. Reduction can lead to recovery of either mip130 allele, in addition to the wild-type configuration at the locus. We found the reduction to be very efficient, and we recovered several lines for each allele. Southern blot analysis confirmed that we had successfully isolated each allele (Fig. 1C).

Although we found that all alleles were viable (Table 1A,B), they were generally not healthy, appeared smaller than nonmutant siblings, and the mip130^{1-723 + 1-36} and mip130^1-723 alleles displayed a mild rough-eye phenotype (data not shown). Interestingly, we found that there were significant differences in the alleles. The mip130^1-723 allele was more severely affected than the mip130^{1-723 + 1-36} and mip130^1-36 alleles, and we were unable to generate mip130^1-723 mutant females for fertility tests and staining assays due to a very short adult lifespan. Based on these data and data to be presented below, we believe that we have generated an allelic series ranging from a null mutant (mip130^1-36) to a weak neomorph (mip130^{1-723 + 1-36}) to a strong neomorph (mip130^1-723).

Immunoblot analysis of the three mip130 mutants showed that none produced a full-length Mip130 protein (Figs. 2A [left panel], 6D [below]). In fact, the mip130^{1-723 + 1-36} and mip130^1-723 alleles produced a truncated Mip130 protein of a molecular weight as anticipated to be encoded by the mip130^1-723 allele. Interestingly, we found that DmMyb protein levels were severely reduced in all mip130 mutants (Figs. 2A [left panel, lanes 2,4], 6D [below, both panels, lane 4]), whereas the levels of the other complex members were only modestly reduced (Fig. 2A, middle panel). Because low levels of DmMyb were detected in the mip130^1-36 mutant (Fig. 6D [below], upper panel, cf. lanes 3 and 4), we speculate that in vivo, the truncated Mip130 protein present in the DmMyb complex in either the mip130^{1-723 + 1-36} or the mip130^1-723 mutant affects DmMyb degradation most effectively. As we will show below, the phenotypic differences in the allelic series are not due to differences in residual DmMyb protein levels. We suspect that the allelic differences may be due to the expression levels of the truncated Mip130 protein and that the viability differences between alleles may in fact be the result of different levels of the truncated Mip130 protein in a particular cell type. In extracts derived from mutant adult males, ∼20% more of the truncated Mip130 protein was measured in mip130^1-723 animals versus mip130^{1-723 + 1-36} animals (Supplementary Fig. 1), consistent with the most severely affected allele expressing the highest levels of the truncated Mip130 protein.

Immunofluorescence analysis of mutant ovaries from mip130^{1-723 + 1-36} (data not shown) and mip130^1-36 females also showed an absence of DmMyb protein staining in the follicle cell nuclei surrounding the developing oocyte (see Fig. 5A, below). RT-PCR analysis of RNA isolated from mip130^{1-723 + 1-36} mutant ovaries showed that the Dm-myb transcript levels were comparable to those from wild-type ovaries (Fig. 2B), suggesting that the absence of DmMyb protein in mip130^{1-723 + 1-36} mutants is due to protein instability and not the result of reduced transcription.

Because we found that DmMyb protein levels were reduced in mip130 mutants, we were interested in determining at what stage in mip130 mutant development DmMyb protein levels decline. We created a mip130^{1-723 + 1-36} mutant strain in which the nonmutated X chromosome harbored a green fluorescent protein (GFP)-producing transgene so that mutant animals at any stage could be sorted from nonmutant animals. When heterozygous mutant females were crossed to mutant males, the absence of the GFP signal using a fluorescence microscope allowed for such manual sorting of homozygous mutant animals. We found that Mip130 levels severely dropped after the maternal stores were depleted following embryogenesis, and that there was a corresponding drop in DmMyb levels shortly thereafter (Fig. 2C). When we performed the reciprocal experiment, we found that with Dm-myb mutants, Mip130 levels were reduced but not eliminated (Fig. 2D). Together, these data suggest that DmMyb requires the presence of full-length Mip130, possibly in association with the other DmMyb complex members, for normal protein levels. The DmMyb protein itself may be unstable without Mip130, perhaps due to misfolding or perhaps due to targeted degradation when not in complex with the other members. The finding that Mip130 levels were not eliminated in Dm-myb mutant animals is also consistent with the presence of a Mip130-containing complex lacking DmMyb in such mutant strains.

Because we saw such a dramatic decrease in DmMyb protein levels in mip130^{1-723 + 1-36} mutants, we were curious to examine the level of each DmMyb complex member when a specified member was removed. To this end, we used RNA-interference (RNAi) on Drosophila Schneider S2 cells and analyzed cell pellets after a 3-d incubation with the double-stranded RNA for the presence of each DmMyb complex member, using immunoblot analysis. During the RNAi analysis, no effects upon the growth rates of the RNAi-treated cells were detected (data not shown). We found that DmMyb was the most sensitive to the loss of another complex member, as depletion of Mip120, Mip130, or Caf1 p55 had dramatic effects upon DmMyb protein levels (Fig. 3A). Similar results were obtained using another Drosophila cell line (Kc), and separate experiments using two distinct, nonoverlapping double-stranded RNAs for Mip120 and Mip130 showed identical results (data not shown). The reduced protein levels in the RNAi experiments were caused by posttranscriptional mechanisms, because the only DmMyb complex member transcript affected in each RNAi experiment was that to which the double-stranded RNA was targeted (Fig. 3B). Depletion of Caf1 p55 affected all complex members except Mip120, consistent with its role as a platform for protein complex assembly in many other complexes. Conversely, Caf1 p55 appeared to be the least sensitive to loss of any DmMyb complex member, presumably due to its presence in a number of complexes in Drosophila. To summarize, we schematically illustrated the interactions between the DmMyb complex members in Figure 3C. The bold arrows indicate that loss of a given target protein had a profound effect upon the levels of the subunit to which the arrow is directed, and dotted lines depict very reproducible but smaller effects. We are uncertain as to the true contacts made by individual complex members; however, it seems that the levels of DmMyb and Mip40 are the most sensitive to another subunit loss. Furthermore, loss of Caf1 p55 and Mip130 affected the levels of several other complex members most significantly and therefore may be part of the “core” of the DmMyb complex. Previous biochemical fractionation experiments showed that DmMyb was always found in complex with the other four subunits, as measured by immunoblot analysis, through the final purification step (data not shown). Further fractionation experiments following the other complex members will be needed to explore the nature of other subcomplex assemblies.

mip130 mutant females display defects in chorion gene amplification

One of the hallmark features of some viable replication mutants is female sterility resulting from reduced chorion gene amplification that leads to thin eggshell production (Calvi and Spradling 1999). As with the viability results, we found a phenotypic range for the different mip130 mutants ranging from partial fertility for the null allele (mip130^1-36) to complete infertility for the mip130^{1-723 + 1-36} allele (Table 2). Moreover, the “fertility” displayed by the mip130^1-36 mutant females was severely impaired: Most individual mip130^1-36 females had only one or two larvae crawling in vials after 7 d of egg laying, whereas the control heterozygous sisters had upwards of 100 larvae crawling in a vial from a single female. Eggs that were laid by mip130 mutant females were flaccid and contained thin eggshells, as exemplified by mip130^{1-723 + 1-36} in Figure 4. Often, the dorsal appendages were thin and fragile, another phenotype associated with defective chorion gene amplification. When we measured the amplification levels in mip130^{1-723 + 1-36} mutants, we found that the levels were reduced twofold compared to heterozygous mutant sisters (data not shown), consistent with the notion that the mip130^{1-723 + 1-36} mutant female sterility is in part due to problems with chorion gene amplification.

In order to ensure that the mutant phenotypes that we were studying were due to loss of Mip130 only, we performed rescue-complementation experiments using the null, mip130^1-36 allele, and transgenic animals carrying the mip130 cDNA under the control of the Gal4 transcription factor. We used the e22c “driver” that expressed Gal4 (and therefore, Mip130) at several developmental stages, including in the ovary. We found that homozygous mip130^1-36 mutant females that contained both the e22c driver and mip130 cDNA were significantly more fertile than their sisters that contained the e22c driver only (Table 2). The complementation restored fertility to 80%-90% of wild-type levels. Importantly, expression of Mip130 resulted in a restoration of the expression of DmMyb (see Fig. 6E [below], cf. lanes 1 and 2).

In the follicle cells surrounding the Drosophila oocyte, genome-wide replication normally ceases prior to stage 10. Subsequently, DNA replication starts again, but at this time, DNA replication initiation occurs at only four loci. These amplification foci can be visualized as four spots of BrdU incorporation or as intense spots of DmOrc2 staining in follicle cell nuclei (Fig. 5B, mip130^{1-723 + 1-36}/FM7; Calvi et al. 1998; Royzman et al. 1999). For both mip130^{1-723 + 1-36} and mip130^1-36 mutants, the normal shut-off of DNA replication prior to stage 10 occurred (Supplementary Fig. 2). However, beginning at stage 10, when we looked at BrdU incorporation in mip130 mutant ovaries, we found that the patterns were very different from wild type: In most nuclei, BrdU incorporation occurred throughout the entire nucleus (Fig. 5A, mip130^1-36 and Fig. 5B, mip130^{1-723 + 1-36}) and was mirrored by a punctate but dispersed DmOrc2 staining pattern (Fig. 5B, mip130^{1-723 + 1-36}; data not shown). Superimposed upon the general BrdU and DmOrc2 staining, many mip130 mutant nuclei displayed rather large and dense staining centers that are suspected to be the chorion amplification foci. This staining pattern was mosaic: In a small fraction of mip130 mutant nuclei, the BrdU and DmOrc2 staining patterns were near normal (data not shown). Even though the abnormal BrdU incorporation persisted into the latest stages in egg chamber development (data not shown), an extra complete endocycle was not manifest in mip130 mutants (Supplementary Fig. 2). As discussed below, such phenotypes are strikingly similar to those reported for e2f2 mutants (Cayirlioglu et al. 2001). These data provide evidence that Mip130 has a role as a negative regulator to keep genome-wide replication off in regions of the chromosome not normally targeted for replication.

Dm-myb mip130 double mutants are viable

Mutations in Dm-myb are lethal, and mutants die as third instar larvae/prepupae (Katzen and Bishop 1996; Manak et al. 2002). We were therefore surprised to find DmMyb protein levels nearly absent in some mip130 mutants, because mip130 mutants were viable. In order to determine whether low levels of DmMyb were responsible for mip130 mutant viability, we made double mutant animals that contained both a lethal mutation in Dm-myb and a mutation in mip130. Because both genes reside on the X chromosome, we generated recombinants that contained both mutations on a single X chromosome. Different lines were generated for each of the alleles of mip130 in combination with the lethal allele of Dm-myb. The presence of both the mutant mip130 and Dm-myb alleles was confirmed by Southern and immunoblot analyses, and representative examples are provided (Fig. 6A,B; data not shown). To our surprise, we found that all of the Dm-myb mip130 double mutant combinations were viable (Table 1A,B). Moreover, when combined with the Dm-myb mutation, the phenotypic differences in the mip130 allelic series were even more apparent. Given that the Dm-myb lethal mutation was constant in these strains and that the null double mutant was the most robust, the allelic differences likely result from the truncated Mip130 protein levels in the different alleles, as discussed (Supplementary Fig. 1). The double mutant animals were not healthy and displayed the same phenotypes as those observed for mip130 single mutants: Males were fertile, females were generally sterile (Table 2), and BrdU incorporation and anti-DmOrc2 staining in most follicle cell nuclei occurred throughout the nucleus (Fig. 5B, right panel, Dm-myb mip130^{1-723 + 1-36} double mutant). Summarizing the behavior of the various double mutants, it appears that mip130 is epistatic to Dm-myb. Based on the above data, we conclude that residual DmMyb activity is not responsible for the observed viability of mip130 mutants. Furthermore, as the Dm-myb mip130 double mutants were viable, it appears that loss of Mip130 works to suppress the normal requirement for DmMyb in critical but unknown cell types.

Plasmid construction

For the mip130 donor element, 4.3 kb of DNA containing all of exon 2 and flanking genomic DNA (containing 485 bp upstream and 1007 bp downstream of the coding sequence) was amplified from genomic DNA in three pieces and subcloned into pHSX. Nucleotide changes were introduced in the PCR fragments as follows to generate the mip130^1-723 and mip130^1-36 alleles: The mip130^1-723 allele genomic sequence was changed from 5′-CAGAATCTCGGCGGCGACTGCG-3′ to 5′-CAGAATTCTCGGCGGCGACTGC-3′ by introduction of a new EcoRI site as indicated, and the mip130^1-36 allele genomic sequence was changed from 5′-GCAGGAGCTTCCGTATCCGACTCCGC-3′ to 5′-GCAGGATCCTTCCGTATCCGACTCCGC-3′ by introduction of a new BamHI site as indicated. An I-SceI site was introduced into the unique Xmn site between the new EcoRI and BamHI sites using PCR. The 4.3-kb DNA was flanked by NotI sites, the modified fragment was cloned into the pT2 transformation vector (Rong et al. 2002), and Drosophila transformants were made using standard methods.

Genetics

Targeting cross: We screened for targeting events by the loss of white+/white mosaicism induced by FLP recombinase as described (Rong and Golic 2001; Rong et al. 2002). Details of the crosses used can be found in the above references. Southern blotting was used to identify and verify the targeted mutant using standard methods.

I-CreI reduction to single alleles: Reduction was carried out using the following cross: mip130^{1-723 + 1-36}/FM7 virgins were crossed to FM7w; 70I-CreI, Sb/TM6 males (Rong et al. 2002). FM7w/mip130^{1-723 + 1-36}; 70I-CreI, Sb females were crossed to FM7w males. After 3 d, the parents were removed and the vials heat-shocked (36°C, 1h). Heterozygous female heat-shocked progeny were crossed to FM7w males, and the white-eyed, Sb+ progeny were selected as reduction allele candidates. Southern blotting was used to identify and verify that both alleles were recovered, using standard methods.

Generation of myb mip130 double mutants: The original mip130 mutant (mip130^{1-723 + 1-36}), mip130^1-36 allele line 1A2 and mip130^1-723 allele line 1A7 were used to make double mutant lines as follows: Heterozygous mip130 mutant/FM7a females were crossed to MH107/Dp(1;Y)shi[+]1,y[+] males (Dm-myb null mutant in which the region on the X chromosome spanning the Dm-myb locus is also carried on the Y chromosome as a duplication; Katzen et al. 1998) and non-FM7 female progeny were crossed to FM7cKrGFP males. Single heterozygous FM7cKrGFP virgin progeny were used to generate stocks. Double mutants were identified using PCR with the following primer pairs: to detect the Dm-myb mutation (MH107), 5′-GC TAAGGACCAGAGAAAATGCG-3′ and 5′-CTTAATAGCA CACTTCGGCACG-3′; to detect the mip130^1-36 allele, 5′-GAG CAAGTAATGAAGCGAAGG-3′ and 5′-CTCATCACGGCAT TCGAATAG-3′; to detect the mip130^1-723 allele, 5′-CTAC GACTCACTAAACGCCAT-3′ and 5′-GTCAGTCAGAAAGC TATGCGG-3′. Double mutant lines were confirmed by Southern analysis.

Developmental time course: mip130^{1-723 + 1-36}/FM7cKrGFP virgins were crossed to FM7cKrGFP males, and progeny were sorted at each stage using a Leica MZFLIII microscope. Samples were homogenized in sample buffer and analyzed using 9% SDS-PAGE followed by immunoblotting with affinity-purified anti-Myb, anti-Mip130, or anti-Tubulin (Sigma) antibodies.

Complementation: For the mip130 complementation, transgenic lines containing the Mip130 cDNA under the control of Gal4 with the use of the pUAST transformation vector were constructed. The “driver” strain (the source of Gal4) was y[1] w[^*]; Pw[+mW.hs] = en2.4-GAL4e22c/SM5 (Bloomington). The cross for complementation was as follows: mip130^1-36/FM7; P e22c/CyO; +/+ virgin females were crossed to mip130^1-36; +/+; pUASTmip130/TM3 males. Two independent UAST-Mip130 lines were tested for complementation. Ten independent experiments in which 10 single-pair matings using females of the following genotype (for a total of 100 females) were scored for fertility: mip130^1-36/mip130^1-36; Pw[+mW.hs] = en2.4-GAL4e22c/+; pUASTMip130/+. As a control, mip130^1-36/mip130^1-36; Pw[+mW.hs] = en2.4-GAL4e22c/+; TM3/+ females were scored for fertility.

RNAi and RT-PCR analysis

Double-stranded RNA production: Gene-specific PCR products (∼700 bp) flanked by T7 polymerase-binding sites (5′-CGGCCAGTGAATTGTTTAATACGACTC) were used as templates in in vitro transcription reactions to produce dsRNA with the RiboMAX large-scale RNA production system (Promega). The dsRNAs were cleaned up using the RNeasy Mini Kit (QIAGEN), eluted in RNA annealing buffer (100 mM NaCl, 20 mM Tris at pH 8.0, 1 mM EDTA), and annealed by incubation at 68°C for 10 min followed by 30 min at 37°C. An aliquot of each dsRNA was analyzed on a 1% agarose gel to check the size and integrity of the dsRNA. Primers used to generate the specific PCR products were as follows: mip40, FlyBase accession CG15119, sense-primer 5′-14355765-14355747, antisense-primer 5′-14355044-14355065-3′; caf1 p55, FlyBase accession CG4236, sense-primer 5′-11051773-11051792-3′, antisense-primer 5′-11052588-11052569-3′; Myb, FlyBase accession CG9045, sense-primer 5′-15588042-15588023-3′, antisense-primer 5′-15587607-15587628-3′; mip120, FlyBase accession CG6061, sense-primer 5′-8575300-8575318-3′, antisense-primer 5′-8576076-8576058-3′; mip130, FlyBase accession CG3480, sense-primer 5′-1675099-1675117-3′, antisense-primer 5′-1675824-1675806-3′.

RNAi: Drosophila L2 cells were diluted to a final concentration of 1 × 10⁶ cells/mL. For each RNAi experiment, 3 mL of cells were washed with serum-free medium, dsRNA (60 μg) added directly to the medium, and the cells were rocked for 2 h at 25°C. Three milliliters of medium supplemented with 10% FCS was added, and the cells were rocked at 25°C for 3 d.

RT-PCR and immunoblot analysis: For RT-PCR, ∼5 × 10⁶ cells were collected and total RNA was isolated using the RNeasy Mini Kit (QIAGEN). RNA samples were treated with RQ1 Rnase-free DNAse (Promega) to eliminate contaminating genomic DNA, and subsequently cleaned up with the RNeasy mini Kit (QIAGEN). Total RNA (2 μg) was used for cDNA synthesis using Superscript One-Step RT-PCR with Platinum Taq (Invitrogen). RT-PCR was performed with gene-specific primers using standard methods. For immunoblot analysis, total protein (corresponding to ∼0.15 × 10⁶ cells) were separated using 12% SDS-PAGE and transferred to nitrocellulose; proteins were detected using affinity-purified polyclonal rabbit antibodies against individual complex members as described (Beall et al. 2002). Anti-tubulin monoclonal antibodies were from Sigma.

Antibody staining and BrdU labeling of ovaries

Ovaries were fixed and processed as described (Beall et al. 2002). Staining was carried out with anti-BrdU monoclonal (Becton-Dickinson), affinity-purified anti-Myb (Beall et al. 2002), and/or anti-Orc2 (Austin et al. 1999) polyclonal antibodies followed by fluorophore-conjugated secondary antibodies (Molecular Probes). Ovaries were visualized with a Zeiss LSM510 confocal microscope.