An RBPJ- Drosophila Model Reveals Dependence of RBPJ Protein Stability on the Formation of Transcription-Regulator Complexes

Abstract

Notch signaling activity governs widespread cellular differentiation in higher animals, including humans, and is involved in several congenital diseases and different forms of cancer. Notch signals are mediated by the transcriptional regulator RBPJ in a complex with activated Notch (NICD). Analysis of Notch pathway regulation in humans is hampered by a partial redundancy of the four Notch receptor copies, yet RBPJ is solitary, allowing its study in model systems. In Drosophila melanogaster, the RBPJ orthologue is encoded by Suppressor of Hairless [Su(H)]. Using genome engineering, we replaced Su(H) by murine RBPJ in order to study its function in the fly. In fact, RBPJ largely substitutes for Su(H)'s function, yet subtle phenotypes reflect increased Notch signaling activity. Accordingly, the binding of RBPJ to Hairless (H) protein, the general Notch antagonist in Drosophila, was considerably reduced compared to that of Su(H). An H-binding defective RBPJ^LLL mutant matched the respective Su(H)^LLL allele: homozygotes were lethal due to extensive Notch hyperactivity. Moreover, RBPJ^LLL protein accumulated at lower levels than wild type RBPJ, except in the presence of NICD. Apparently, RBPJ protein stability depends on protein complex formation with either H or NICD, similar to Su(H), demonstrating that the murine homologue underlies the same regulatory mechanisms as Su(H) in Drosophila. These results underscore the importance of regulating the availability of RBPJ protein to correctly mediate Notch signaling activity in the fly.

Keywords: Drosophila; Notch signaling pathway; RBPJ; Su(H); functional conservation; model system; regulation.

Figure 1

Genome engineering at the Su(H) locus to integrate murine RBPJ. (A) Comparison of Drosophila Su(H) protein with RBPJ protein from Mus musculus. N-terminal domain (NTD, light blue); ß-trefoil domain (BTD, green); C-terminal domain (CTD, yellow) and the N-terminally located alpha1-helix, which makes contact to CTD (yellow), are well conserved with identity between 60% and 90%. The flanking parts of the proteins, however, show little conservation. Numbers above the proteins depict codons, framing the respective domains. Structure of RBPJ bound to DNA (PDB ID: 3BRG) [38]. DNA is colored in gray; domains in RBPJ are colored as above. (B) Flow chart of strategy used to exchange Su(H) with murine RBPJ by genome engineering. The founder line Su(H)^attP was used to integrate RBPJ^wt and RBPJ^LLL cloned in pGE-attB^GMR via PhiC31-integrase mediated recombination at the attP landing site. Subsequently, vector sequences and the white⁺ marker, flanked by loxP sites, were excised with the help of the Cre-recombinase to yield the final fly strains RBPJ^wt and RBPJ^LLL. (C) Splicing of RBPJ^wt mRNA occurred as expected in the RBPJ^wt strain, leading to a PCR product of about 410 bp (open arrow). RT-PCR was performed on cDNA from the given strains, using y¹ w^67c23 flies as controls; (+) with reverse transcriptase and (–) no-RT control. Primer pairs S.up and R.lo are depicted schematically in (B). Tubulin primers served as controls for intact mRNA (arrows). As size standard (M), a 100 bp ladder was used. ^* Label unspecific bands.

Figure 2

Adult phenotypes in RBPJ adults. (A,B) Scanning electron micrographs of fly heads (A) and thoraces (B) of the given genotype. In comparison to the control Su(H)^gwt or the heterozygous RBPJ^wt / +, the homozygous RBPJ^wt flies have fewer macro- and microchaetae. This phenotype is enhanced in a RBPJ^wt/Su(H)^attP background or in the RBPJ^wt/RBPJ^LLL combination (arrows point to examples of missing bristles). RBPJ^LLL/+ heterozygotes match the control. Scale bars: 200 µm. (C) Viability of heterozygous RBPJ^wt and RBPJ^LLL flies of the homozygotes and the RBPJ^wt/Su(H)^attP combination, respectively, was determined relative to their balanced siblings (hatching ratio). Bars depict the fraction of the expected offspring (RBPJ^wt black, RBPJ^LLL grey, and CyO light grey); numbers show total animals analyzed. The heterozygotes balanced over CyO were crossed to the flies with the genotypes given above. Note that RBPJ^wt flies always hatch at the expected numbers, whereas RBPJ^LLL/RBPJ^wt heterozygotes are slightly underrepresented. (D) Average number of macrochaetae in adult females of the given genotype (n = 20). Note significantly reduced numbers in the homozygous RBPJ^wt and RBPJ^wt/Su(H)^attP flies, as well as in the RBPJ^wt/RBPJ^LLL combination. (E) Average number of microchaetae determined from scanning electron micrographs; the evaluated sector is highlighted in the control in (B). Number of animals analyzed is given in each bar. Note significant reduction in RBPJ^wt/RBPJ^wt and RBPJ^wt/Su(H)^attP flies, as well as in the heterozygous RBPJ^wt/RBPJ^LLL animals. (D,E) Statistical analyses were performed with ANOVA Tukey–Kramer approach relative to wild type control (*** p < 0.001; * p < 0.05). (F) Typical examples of wings from female flies of the given genotype are depicted. Sixty-eight percent of the RBPJ^wt homozygotes are characterized by a shortened L5 vein (arrow, n = 28), an effect which is enhanced in RBPJ^wt/Su(H)^attP flies, where additionally 100% of L4 and 50% of L2 veins are shortened (open arrows, n = 23). A likewise enhancement is seen in the RBPJ^wt/RBPJ^LLL combination (n = 29).

Figure 3

Assembly of repressor and activator complexes by RBPJ. (A) Yeast two-hybrid assay for the interaction of H with Su(H) and RBPJ variants, respectively. H fused to the lexA-DNA binding domain (DBD) provided in pEG vector; CSL variants fused to the trans-activator domain (TAD) in pJG vector. Interaction results in transcription of the lacZ reporter, as shown in the scheme. No binding is seen between the H-binding deficient Su(H)^LLL and RBPJ^LLL isoforms, whereas both Su(H)^wt and RBPJ^wt show binding, however the latter is much weaker than Su(H)^wt. Interaction assays were done with the corresponding full-length proteins. (B) Yeast protein three-hybrid assay for formation of a ternary complex with the D. melanogaster components, co-activator Mam (aa 118–194) provided in pEG vector, RICN (intracellular Notch including RAM domain; aa 1762–2176) provided in pESC vector, and CSL variants provided in pJG vector. Ternary complex formation results in transcription of the lacZ reporter, as shown in the scheme. (C) Yeast protein three-hybrid assay as in (B) with M. musculus components, co-activator MamL (aa 12–74) provided in pEG vector, NICD (aa 1751–2293) provided in pESC vector, and CSL variants provided in pJG vector. (A’–C’) Quantification of the interactions shown in (A–C) is given in Miller Units. At least six different clones from two independent experiments were quantified and statistically analyzed with ANOVA and two-tailed Tukey–Kramer test relative to RBPJ. (*** p < 0.001; ns: not significant).

Figure 4

RBPJ^LLL homozygotes display strong gain of Notch activity. Wing imaginal discs derived from homozygous larvae, as indicated, were assayed for (A) Wingless or (B) Pebbled (also named Hindsight) protein expression. (A) Compared to the control Su(H)^gwt, the wing blade area is slightly enlarged in the RBPJ^wt homozygote and strongly hypertrophied in the RBPJ^LLL mutant discs (double headed arrow). (B) Sensory organ precursors express Pebbled (arrows point to examples). Their number is strongly reduced in RBPJ^LLL mutant discs. Note complete absence of the presumptive triple row in the presumptive wing field, marked by an arrowhead in the controls. Size bar: 100 µm.

Figure 5

The Notch target Dpn is de-repressed in RBPJ^LLL homozygous cells. (A–D) Clonal analysis to monitor expression of Deadpan (Dpn; magenta); wild type cells are labeled with GFP (green), whereas homozygous cells of the indicated genotype are unlabeled. Mutant cell clones are outlined for clarity. Dpn expression is undisturbed in Su(H)^gwt control clones (A), but appears mildly upregulated in RBPJ^wt clones (arrow) (B). In contrast, cell clones homozygous mutant for Su(H)^LLL (C) or RBP^LLL (D) display a robust upregulation of Dpn expression (arrows). Size bars: 50 µm.

Figure 6

Quantification of altered transcription resulting from the failure of repressor complex formation. Expression of dpn (A), E(spl)mß (B), and peb (C) transcripts, respectively, was quantified by qRT-PCR relative to Su(H)^gwt; cyp33 and Tbp were used as reference genes. mRNA was prepared from larval wing discs isolated from 25 homozygous larvae, each of the indicated genotype. Data were gained from four biological and two technical replicates. An increase in dpn and E(spl)mß transcription levels was observed in Su(H)^LLL, RBPJ^LLL, and H^attp mutants. In contrast, peb transcripts were reduced. Median corresponds to expression ratio; mini-max depicts 95% confidence. The p-values are given above each bar; significance was tested using PFRR from REST (p < 0.05).

Figure 7

RBPJ^LLL protein abundance is lowered compared to RBPJ^wt. (A,B) Clonal analysis: cells with a wild type Su(H) gene copy are labeled by GFP (green); RBPJ protein expression is shown (magenta). Homozygous RBPJ* cell clones are outlined for clarity. RBPJ protein accumulates to a higher level in homozygous RBPJ^wt cells (arrow) than in heterozygous cells (RBPJ^wt/+). (B) RBPJ protein is barely visible in the RBPJ^LLL heterozygous cells, and is likewise lowered in homozygous cells (outlined). Note, however, that nuclear accumulation of RBPJ protein in a stripe of cells along the dorso–ventral boundary. Insets show enlargements of framed region. Size bars represent 50 µm. (C) Western blots on protein extracts from homozygous larvae: Note reduced level of RBPJ protein (double headed arrows) in the RBPJ^LLL mutants compared to RBPJ^wt control. M, pre-stained protein marker (in kDa). Tubulin was used as a loading control (Tub). Blot was cut for parallel detection of RBPJ and Tubulin. Quantification of signals from four independent Western blots with Image J gel analysis program in relation to the beta-Tubulin signals. Error bars denote standard deviation; Student’s t test was applied (*** p < 0.001). Uncropped blots used for quantification are shown in supplemental Figure S4.