Presenilin controls CBP levels in the adult Drosophila central nervous system

Abstract

Background: Dominant mutations in both human Presenilin (Psn) genes have been correlated with the formation of amyloid plaques and development of familial early-onset Alzheimer's disease (AD). However, a definitive mechanism whereby plaque formation causes the pathology of familial and sporadic forms of AD has remained elusive. Recent discoveries of several substrates for Psn protease activity have sparked alternative hypotheses for the pathophysiology underlying AD. CBP (CREB-binding protein) is a haplo-insufficient transcriptional co-activator with histone acetly-transferase (HAT) activity that has been proposed to be a downstream target of Psn signaling. Individuals with altered CBP have cognitive deficits that have been linked to several neurological disorders.

Methodology/principal findings: Using a transgenic RNA-interference strategy to selectively silence CBP, Psn, and Notch in adult Drosophila, we provide evidence for the first time that Psn is required for normal CBP levels and for maintaining specific global acetylations at lysine 8 of histone 4 (H4K8ac) in the central nervous system (CNS). In addition, flies conditionally compromised for the adult-expression of CBP display an altered geotaxis behavior that may reflect a neurological defect.

Conclusions/significance: Our data support a model in which Psn regulates CBP levels in the adult fly brain in a manner that is independent of Notch signaling. Although we do not understand the molecular mechanism underlying the association between Psn and CBP, our results underscore the need to learn more about the basic relationship between Psn-regulated substrates and essential functions of the nervous system.

Figure 1. Developmental phenotypes of flies compromised for Notch, Psn or CBP using transgenic RNAi.

(A) represents a wildtype wing and (B–D) show phenotypes of the UAS-RNAi lines crossed to c96-Gal4. Silencing either gene with this driver results in irregular or missing wing margins and wrinkling and notching of the wing blade. (E) is a wildtype thorax and (F–H) are representative samples of the UAS-RNAi lines crossed to pnr-Gal4. Silencing the genes with this driver leads to improper sensory organ development that gives rise to balding and/or supernumerary machrochete on the thorax and notum. (I) is a wildtype eye and (J–L) represent UAS-RNAi lines crossed to GMR-Gal4, which causes blistered and fused ommatidia, and with the Psn knockdowns in particular (K), sporadic elongated sensory hairs.

Figure 2. Notch is not required for CBP expression.

(A) shows the location of four putative binding sites for Su(H) near the transcription unit of CBP. Boxes indicate exons and filled regions represent sequences that are translated. The position of each element is indicated relative to the start of transcription (+1, arrow). (B) is a western analysis of Notch and CBP protein levels under conditions in which Notch is silenced with RNAi by applying multiple heat shocks (see Materials and Methods). Two blots were made from each protein sample. One blot was incubated with Notch antibody (top row) and the other was incubated with CBP antibody (3^rd row). Both blots were cut and incubated separately with tubulin antibody to serve as a control for protein loading and transfer (2^nd and 4^th rows). (C–D) are confocal images of dissected brains from control (no Gal4, UAS-Ni) and Notch-silenced experimental animals (hsGal4; UAS-Ni). Both groups were heat-shocked using the same regimen as B, and both were stained with CBP antibody and imaged using the same conditions and microscope settings.

Figure 3. Psn is necessary for maintaining proper CBP protein levels in the adult Drosophila CNS.

(A) shows a western analysis from fly heads in which CBP protein levels are reduced >90% (hsGal4; UAS-Psni; + heat shocks) compared to the controls (heat shocked animals without a driver or an RNAi responder) using tubulin as an internal standard for protein loading and transfer. (B) is a table of qRT-PCR results assaying CBP transcript levels (normalized relative to control  =  no Gal4; UAS-Psni) under conditions where Psn is silenced (hsGal4; UAS-Psni). An antibody stain on dissected heads indicates that CBP levels are drastically reduced when Psn is silenced (D) compared to a control without a driver (C). Arrowheads in (D) indicate non-specific autofluorescence of red pigment remaining after dissection.

Figure 4. The HAT domain of CBP is required for H4K8ac in the adult Drosophila CNS.

(A) depicts a typical roughened eye phenotype produced when UAS-CBP-FLAD is expressed with the GMR-Gal4 driver during development. (B) displays a wing phenotype with a prominent notch (arrowhead) that is typically observed when UAS-CBP-FLAD is expressed with the c96-Gal4 driver. Both phenotypes are similar to those detected when UAS-CBPi is expressed with the same drivers (Fig. 1). (C) is a control brain (no Gal4; UAS-CBP-FLAD; + heat shocks) stained for H4K8ac, and (D) is an experimental brain (hsGal4; UAS-CBP-FLAD; + heat shocks) stained with the same antibody under the same conditions. Note the dramatic reduction in staining intensity when the HAT-defective form of CBP is expressed.

Figure 5. Psn affects H4K8ac levels in the adult brain.

The above images are whole-mount adult fly brains stained for H4K8ac as an indicator of global acetylation levels. (A,D,G) represent one control group (hsGal4; UAS-RNAi; no heat shocks), (B,E,H) depict a second control group (no Gal4; UAS-RNAi; + heat shocks), and (C,F,I) represent the experimental group (hsGal4; UAS-RNAi; + heat shocks). Note that when CBP is silenced there is a dramatic effect on H4K8ac levels (compare C with the A and B controls); when Psn is silenced there is a significant but less dramatic effect (F); and when Notch is silenced (I) there is no noticeable effect. All brains were dissected, stained, and photographed using the same conditions and microscope settings.

Figure 6. Assaying geotaxis using a countercurrent device.

Pictured is the countercurrent device we built for assaying geotaxis behavior. This apparatus was modeled directly from a similar design created by Seymour Benzer . Each experiment starts as flies are placed in tube set 1 and given 3 sharp taps on a rubber mat to knock them to the bottom. Flies are then given 10 seconds to climb into the upper tube, after which time it is moved over to the bottom of set 2. They are again tapped to the bottom, allowed to climb for 10 seconds, and moved to the bottom of set 3. The procedure continues until set 6 is reached. At the end of the experiment, the percentage of the fly cohort in each tube is calculated. The experiment is calibrated in such a way that most of the wildtype flies (exhibiting normal negative geotaxis) accumulate in tube number six.

Figure 7. Flies compromised for CBP exhibit poor geotaxis.

(A) represents the mean distribution of flies within each tube after 6 successive trials. Tube distribution for each group is presented by a best-fit line, which is based on a linear regression. The slope of the experimental group (hs-Gal4; UAS-CBPi) indicates that the majority of these cohorts remained in the first sets of tubes, whereas the control groups accumulated in the last sets of tubes. (B) is the mean score for each group with a significant difference between the experimental group (hsGal4; UAS-CBPi; +heat shocks) and all three control groups (F _(3,53) = 11.8; P<0.0001). There were no significant differences between genders or control groups. A (+) indicates a heat shock treatment and (-) indicates no heat shock.