The tauopathy model used in the current studies is based on expression of human tau in Drosophila
. In our model the bipartite UAS/GAL4 expression system 
is used to express human tau in a panneural pattern with the elav-GAL4
driver. The model has been previously described in detail 
. To determine if the UPR is activated in our tauopathy model we utilized a UAS-Xbp1-EGFP
reporter system developed by Ryoo et al. 
. In the reporter construct, EGFP is located after the IRE-1 splice site in Xbp1 such that EGFP is only in frame after Xbp1 has been spliced by IRE-1. Thus the UAS-Xbp1-EGFP transgene acts as a reporter for UPR activation as expression of EGFP indicates that Xbp1 has been cleaved by IRE-1. The UAS-Xbp1-EFGP
transgene was first co-expressed with human tau in Drosophila
neurons using the elav-GAL4
driver. The FTDP-17 associated mutant version of tau, tauR406W
, was used in these experiments because expression of mutant tauR406W
induces a level of toxicity well suited for experimental analyses 
. We observed activation of the UPR reporter in adult transgenic tau animals (). Activation was apparent throughout the brain, consistent with the panneural pattern of the elav-GAL4
driver and the widespread neurodegeneration present in tau transgenic flies 
. To determine the cell type with reporter activation, we double stained for elav, a marker of neurons, or repo, a glial marker. We observed activation of the reporter in neurons (, upper panels), but not glia (, lower panels). To confirm activation of the UPR in tau transgenic flies we used an enhancer trap P element inserted into the Hsc70-3
locus. We found significant activation of Hsc70-3 in tau transgenic flies (27.6±2.6 ß-galactosidase immunoreactive cells per brain; genotype: elav-GAL4/Hsc70-3G0407;UAS-tauR406W/+
) compared to controls (1.2±0.5 ß-galactosidase immunoreactive cells per brain; genotype: elav-GAL4/Hsc70-3G0407
) at 10 days of age.
Upregulation of unfolded protein response.
To further the explore the relationship between tau expression and UPR activation we quantified the degree of UPR activation flies expressing tau transgenes that cause varying levels of neurotoxicity. At 10 days of age, expression of human wild-type tau (tauWT
) in Drosophila
neurons is mildly toxic, expression of the FTDP-17 linked mutant tau (tauR406W
) is moderately toxic, and expression of a phosphomimetic mutant tau in which 14 disease-associated SP/TP phosphorylation sites are mutated to glutamate (tauE14
) is substantially more toxic 
. Western blot analysis confirmed that differences in neurotoxicity were not due to different levels of tau expression (Figure S1
), consistent with prior characterization of these mutant lines 
. For quantitative analysis, the number of EGFP positive nuclei was counted in 4 µm serial sections of entire fly brains. UPR activation was absent in control flies and minimal in flies expressing only the UAS-Xbp1-EGFP
reporter at 10 days of age (), or at one day or 30 days of age (data not shown). In contrast, activation of the reporter was robust in flies expressing all forms of tau (). The degree of UPR activation correlated very well with the degree of toxicity caused by the various forms of tau (tauWT
43.0±2.9). Activation of the UPR in tau transgenic flies was age dependent (tauR406W
6.7±1.1 at one day of age), consistent with our prior characterization of age-dependent neurodegeneration in tau transgenic flies 
. Our findings suggest a plausible link between tau neurotoxicity and UPR activation. Further, the substantially increased UPR activation seen with expression of pseudophosphorylated tau supports a role for tau phosphorylation upstream of UPR activation.
Quantitative analysis of Xbp1-EGFP reporter activity.
To determine whether activation of the UPR influences neurotoxicity, or is merely a marker of cellular toxicity in our model, we determined if manipulation of the UPR modifies tau toxicity. To decrease function of the UPR we utilized a loss of function Xbp1 allele (Xbp1k13803
homozygous for the Xbp1k13808
allele have severely reduced Xbp1 transcript levels, exhibit growth retardation and die before reaching the pupal stage 
. To assess apoptosis in adult fly brains a less severe manipulation of Xbp1 levels was required so we examined Drosophila Xbp1k13808
heterozygotes. Although heterozygosity for Xbp1k13808
itself does not cause apoptosis, quantification of the number of TUNEL-positive cells in 10-day-old fly brains revealed that reducing Xbp1 function with one copy of Xbp1k13808
significantly increased apoptosis mediated by expression of tau (tauR406W
44.9±4.4, p<0.05; ). Thus Xbp1 protects against the toxicity caused by expression of tau. We confirmed these results using an RNAi against Xbp1 (Xbp1-IR). Although expression of Xbp1-IR alone resulted in few TUNEL positive cells (0.8±0.7), its co-expression with tau more than doubled the number of TUNEL positive cells compared to expression of tau alone (tauR406W
52.1±6.9, p<0.001; ). Enhancement of tau toxicity by reducing Xbp1 levels did not result from increased tau expression because Western blot analysis showed similar levels of tau in control flies (genotype: elav-GAL4/+;UAS- tauR406W/+)
and tau transgenic flies with reduced levels of Xbp1 (genotypes: elav-GAL4/+;UAS- tauR406W/UAS-Xbp1-IR
and elav-GAL4/+;Xbp1k13808/+;UAS- tauR406W/+
; Figure S2
Quantitative analysis of cell death using TUNEL staining.
Since cell death is apoptotic in our model and can be decreased by inhibiting caspase activity 
, the increased TUNEL seen in tau transgenic fly brains when an Xbp1 loss of function allele or Xbp1-IR is present suggests that effector caspase activity is increased by reducing Xpb1. To further probe activation of caspase in our system we employed a genetically encoded reporter construct. Transgenic flies have been created that express a caspase substrate, human poly-ADP-ribose polymerase-1 (PARP). Human PARP is cleaved by mammalian caspase-3 and by homologous Drosophila
effector caspases. Processed PARP can then be recognized by an antibody specific to cleaved human PARP 
. Human PARP was co-expressed with tau alone and with tau in combination with Xbp1k13808
or Xbp1-IR. An antibody to cleaved PARP was used to detect cleavage and the number of cell bodies with cleaved PARP was quantified. Although heterozygosity for Xbp1k13808
or expression of Xbp1-IR alone resulted in few cells showing evidence of PARP cleavage (Xbp1k13803
1.0±0.5 and Xbp1-IR/PARP
4.8±1.8), decreasing Xbp1 significantly increased the amount of PARP cleavage observed with tau transgenic flies (tauR406W
27.6±1.9, p<0.05, and Xbp1-IR/tauR406W
35.8±3.1, p<0.001; ). Of note, the number of neurons showing UPR activation as measured by the Xbp1-EGFP reporter is similar to the number of TUNEL- and cleaved PARP-positive neurons, consistent with a role for the UPR in tau neurotoxicity.
Increased cleaved PARP in flies with reduced levels of Xbp1.
We have previously demonstrated that cell cycle activation causes neurodegeneration in our Drosophila
tauopathy model 
. To determine if reducing Xbp1 promotes tau-mediated neurodegeneration via cell cycle activation, fly brains were immunostained with an antibody against the proliferating cell nuclear antigen (PCNA), a marker of cell cycle activation 
, and the number of PCNA positive foci were counted. Although aberrant PCNA expression is minimal with Xbp1-IR alone (0.2±0.2% of tauR406W
), Xbp1-IR greatly increases the aberrant PCNA expression seen in tau brains (160.2±17.7% of tauR406W
, p<0.01; ).
Increased cell cycle activation in flies with reduced levels of Xbp1.