Drosophila SAGA harbors one of the two fly ADA2-type adaptors, dADA2b, while dADA2a is present in the ATAC HAT complex [1, 7]. Within the yeast (y)SAGA ADA2 and ADA3 are believed to play a role in extending the lysine specificity of GCN5 HAT . In accord with this function, yADA2 physically interacts with yGCN5 . The SANT domain of yADA2 was mapped as a region critical for the interaction and the N-terminal half of this domain was found to be required for effective histone acetylation by yGCN5 in the context of SAGA . In Drosophila too, the absence of dADA2b makes dSAGA an incompetent nucleosomal histone H3 acetyltransferase [11, 12].
We present here biochemical and genetic evidences that the two isoforms of dADA2b are expressed in different quantities during Drosophila development. It might be surprising that despite the detection of two different forms of dAda2b messages in earlier studies [1, 11] the functions of the two protein isoforms have not been studied. Since dADA2bS has only a short unique peptide segment distinguishing it from dADA2bL, discriminating between the two isoforms might be difficult by mass spectrometric determination of dSAGA constituents. The antibody we used in this study for dADA2b detection is highly specific and the complete absence of both dADA2b isoforms from extracts of dAda2b null animals (Figure 2A and 3A) proves that both proteins are bona fide products of dAda2b.
Our analysis of dAda2bS transgene-carriers provides information on the in vivo function of the short dADA2b isoform. The partial phenotypic rescue of dAda2b null mutants by dAda2bS transgenes proves that the shorter isoform alone can contribute to dSAGA functions required for the completion of development. The decreased viability of adults possessing only dADA2bS suggests functional deficiencies, which might be related to dSAGA function in neuronal development . The dSAGA-specific histone H3 modification pattern of animals expressing only dADA2bS indicates a reduction in the level of H3K14ac and particularly in that of H3K9ac (Figure 4). The high level of dADA2bS expression in the transgene-carriers makes it unlikely that the amount of dSAGA complexes would be limiting. Rather, the HAT activity of dSAGA complexes containing only dADA2bS seems to be generally decreased. A change in activity of a C-terminally truncated dADA2b form is in accord with our recent observation that the C-terminal regions of dADA2 proteins play important roles in targeting them to GCN5-containing complexes .
Microarray analysis of total RNA profiles revealed that the mRNA levels corresponding to several genes affected by dAda2b mutation were partially restored in dAda2bS transgene-carriers. Significantly, the levels of those mRNAs, which were either up- or down-regulated in dAda2b null mutants, were changed towards the wild type control in dAda2bS transgene-carriers (Figure 6). Furthermore, the short dADA2b isoform alone seems to be more potent to restore mRNA levels of genes up-regulated in null mutants.
For a large number of genes we observed that in dADA2bS expressing animals the mRNA levels were close to those detected in dAda2b null mutants, so dADA2bS could not restore the normal expression levels of these genes. In the case of a smaller number of genes the mRNA levels in dAda2bS transgene-carriers were between those detected in wild type and dAda2b samples. The intermediate mRNA levels of these genes are indicative of a partial dSAGA function. Perhaps the most interesting are those few genes for which the presence of dADA2bS resulted in mRNA levels close to the wild type (Figure 6B). All these genes are represented with increased RNA level in null mutants, but in the presence of the short dADA2b isoform their expression is kept at a lower level. The expression of dADA2bS alone results in a decreased level of mRNA for more genes than the expression of none of the dADA2b isoforms (Figure 5A). One explanation for these observations could be that the negative function of dADA2b-containing complex(es) is restored better if at least the short isoform is present, than if none.
The microarray data unfortunately does not provide information on which step(s) of transcription is affected when only dADA2bS is expressed. SAGA complexes have several structural modules, to which specific functions can be assigned . The ADA2 type proteins belong to the GCN5-HAT module, but this does not exclude the possibility that they modify SAGA function through mechanisms other than histone acetylation. For example mutations affecting WDA and dADA2b proteins cause similar phenotypes though the former is part of the deubiquitinating module . Furthermore, the acetyltransferase function of SAGA could affect gene expression by several mechanisms including promoter specific modification, global histone acetylation, factor acetylation, acting at PIC formation or at elongation . The various SAGA functions explain the observations that up- and down-regulated genes are detected in comparable numbers in dAda2b mutants. The phenotype and histone acetylation pattern of animals expressing only dADA2bS indicate that with the short dADA2b isoform dSAGA complexes with reduced activity are formed. Whether the activity changes of these complexes affect the expression of genes selectively and differently at specific stages and/or conditions remains to be revealed. Our data on the alterations of the transcriptome in the presence of dADA2bS alone suggest that the use of distinct ADA2b adaptor isoforms could be a way to achieve gene-selective effects.