Cytosolic O-GlcNAcylation and PNG1 maintain Drosophila gut homeostasis by regulating proliferation and apoptosis

Tissue homeostasis requires a delicate balance between stem cell self-renewal, proliferation, and differentiation. Essential to this process is glycosylation, with both intra-and extra-cellular glycosylation being required for stem cell homeostasis. However, it remains unknown how intracellular glycosylation, O-GlcNAcylation, interfaces with cellular components of the extracellular glycosylation machinery, like the cytosolic N-glycanase NGLY1. In this study, we utilize the Drosophila gut and uncover a pathway in which O-GlcNAcylation cooperates with the NGLY1 homologue PNG1 to regulate proliferation in intestinal stem cells (ISCs) and apoptosis in differentiated enterocytes. Further, the CncC antioxidant signaling pathway and ENGase, an enzyme involved in the processing of free oligosaccharides in the cytosol, interact with O-GlcNAc and PNG1 through regulation of protein aggregates to contribute to gut maintenance. These findings reveal a complex coordinated regulation between O-GlcNAcylation and the cytosolic glycanase PNG1 critical to balancing proliferation and apoptosis to maintain gut homeostasis.

express Piezo, a cation channel that senses mechanical tension [24,25]. Previous studies have identified the HBP as a key player regulating ISC response to nutrition and midgut adaptation [26,27]. We have previously investigated the role of O-GlcNAcylation in ISCs/EBs through genetic mutation of the enzymes involved in O-GlcNAc cycling, OGT and OGA. We reported that O-GlcNAcylation controls Drosophila intestinal stem cell/progenitor cell homeostasis [27]. These data indicated that O-GlcNAc is important for stem cell maintenance, function, and tissue development through regulation of the DNA damage response and stress-induced proliferation [27]. However, it remains unknown if intracellular glycosylation interfaces with cytosolic glycanases like NGLY1 in adult stem cells, and further how this interaction is utilized in response to stress to maintain normal gut homeostasis.
In this study we uncover a pathway by which loss of PNG1 interacts with the enzymes of O-GlcNAc cycling in ISC/EBs and further differentiated EC cells to maintain gut homeostasis. Genetic analysis suggests that O-GlcNAc and NGLY1 act within the same pathway to regulate proliferation in ISC/EBs and apoptosis in ECs. Surprisingly, we find coordinate regulation of PNG1 and OGT, with PNG1 acting to stabilize OGT. Further investigation revealed that modulation of Nrf2 antioxidant signaling acts downstream of PNG1 and OGT to regulate the proliferation/apoptosis balance required in ISC/EBs and ECs for tissue maintenance. Similarly, Endo-beta-N-acetylglucosaminidase (ENGase) which is a key enzyme involved in the processing of free oligosaccharides in the cytosol, also participates in maintaining gut homeostasis through the PNG1/OGT pathway. Thus, abnormal phenotypes of OGT/PNG1 knockdown can be rescued by their downstream effectors, CncC and ENGase through regulation of protein aggregates and proteasome activity. The results highlight the importance of glycosylation in nutrient-sensitive stem cell diseases like cancer. We also present an underlying molecular mechanism and unexpected pathway that can be targeted for treating NGLY1-deficient patients.

PNG1 regulates stem cell proliferation and is elevated in ROS-induced and OGA knockdown midgut ISCs/EBs
Previous reports indicate that Both PNG1 (NGLY1) and OGT have critical roles within the gut that contribute to normal gut homeostasis and importantly the overall health of the organism. PNG1 null larvae have specific developmental abnormalities in their midgut that contributes to their lethality [28]. Further, intestinal inflammation in Crohn's disease is associated with increased O-GlcNAc modification [29]. Our previous study also showed that increased O-GlcNAc promotes gut dysplasia through regulation of DNA damage [27]. Thus, Png1 or O-GlcNAc might be associated with gut dysfunction in a disease context. To define the interaction between intracellular glycosylation and the cytosolic glycosidase NGLY1 we first investigated if the Drosophila homologue of NGLY1, PNG1, contributes to proliferation in adult stem cells. We used flies in which the enzyme PNG1 was knocked down (UAS-Png1 RNAi #103607; UAS-Png1 RNAi #BL54853), mutated (UAS-Png1 C303A and Png1 ex18 ), or Wild-Type Myo1A ts >Png1 C303A , Myo1A ts >Png1 RNAi (BL54853), Myo1A ts >Ogt RNAi , and Myo1A ts >Oga RNAi flies. (F) The number of PH3-positive cells in midguts from flies of the indicated genotype. (G) Immunofluorescence staining of NGLY1 (red) in esg-GFP-positive cells (green) in midgut of 7-day-old esg ts >+, PQ-treated esg ts >+, esg ts >Oga RNAi , and esg ts ;Oga del.1 flies. (H) Quantification of NGLY1 mean fluorescence per esg-GFP-positive cell. (I) Immunofluorescence staining of NGLY1 (red) and O-GlcNAc (white) in esg-GFP-positive cells (green) from the midgut of esg ts ;Oga del.1 flies. Outline indicates esg-positive cells. Data are represented as mean ± SD. � p< 0.05. �� p< 0.01. ��� p< 0.001. ���� p< 0.0001. n.s., not significant., see S1 (UAS-Png1 WT ) specifically in ISCs/EBs (esg ts , ISC/EB cell-specific inducible GAL4) or ECs (Myo1A ts ,UAS-GFP,Gal4; EC-cell-specific inducible GAL4). After incubating at 29˚C, we observed that GFP-positive cells (indicating ISC/EB) and PH3-positive cells (phospho-histone H3; a marker of cell proliferation) of esg ts >Png1 WT did not significantly change compared to control (Fig 1B-1D). However, there was a significant decrease in the number of GFP-positive and PH3-positive cells in the esg ts >Png1 RNAi , esg ts >Png1 C303A and esg ts >Png1 ex18 midguts compared to control (Fig 1B-1D). Interestingly, in the EC-specific PNG1 knockdown (Myo1A ts >Png1 RNAi ) and PNG1 mutant midguts (Myo1A ts >Png1 C303A ), the number of PH3-positive cells were significantly increased (Fig 1E and 1F). As shown in S1 Fig, ISC proliferation was induced while mature EC cells were reduced in EC-specific Png1 knockdown midgut compared to control. The percentage of Delta positive cells (ISCs) vs. Myo1A-GFP positive cells (mature cells) increased in Png1 knockdown midgut compared to control. These phenotypes are similar to the ISC/EB-and EC-specific OGT knockdown phenotypes described previously [27], and confirmed here (Fig 1B-1F). Thus, our data indicated that loss of PNG1 in ISCs/EBs suppressed ISC proliferation and loss of PNG1 in ECs promoted ISC hyperproliferation, similar to what was observed with loss of OGT. Because the PNG1 and OGT mutant phenotypes looked similar, we next investigated the association between PNG1 levels and O-GlcNAcylation. We assessed PNG1 levels in Paraquat (PQ)-treated (an inducer of extrinsic oxidative stress), OGA knockdown, and OGA del.1 mutant midgut ISCs/EBs, all of which have been shown to elevate O-GlcNAc levels [27]. We found that PNG1 levels were significantly increased, concomitant with O-GlcNAc in the PQ-treated, OGA knockdown, and OGA del.1 mutant midgut ISCs/EBs (Fig 1G-1I). We next assessed ISC proliferation in PNG1 mutants (Png1 ex18 ; FRT82B) and OGT mutants (sxc 7 ; FRT82B) using the MARCM (mosaic analysis with a repressible cell marker) system. We found that the number of clones and clone size decreased in Png1 ex18 and sxc 7 mutant midguts (S2A- S2C Fig). O-GlcNAc and PNG1 were reduced in Png1 ex18 and sxc 7 mutant midguts (S2D and S2E Fig). Additionally, our previous data showed that ISC proliferation and O-GlcNAc levels increased in OGA del.1 , FRTB2B mutant ISCs compared to control [27]. Therefore, our data revealed that increased O-GlcNAcylation is correlated with an increase in PNG1 levels.

PNG1 knockdown rescues dysplasia induced by hyper-O-GlcNAcylation in ISCs/EBs
Our data suggested that PNG1 levels increased in OGA knockdown and OGA del.1 mutant midgut ISCs/EBs. Additionally, the increased PNG1 levels significantly correlated with O-GlcNAc in ISCs/EBs (Fig 1), suggesting a link between intracellular O-GlcNAc and PNG1. Thus, we hypothesized an interaction between O-GlcNAc and PNG1 in stem cells and progenitor cells. To test this hypothesis, we asked if the mutant Png1 C303A could rescue ISC hyperproliferation, increased O-GlcNAcylation and elevated ROS levels in our ISC/EB-specific OGA knockdown. Our data confirmed ISC hyperproliferation in OGA knockdown ISCs/EBs [27] (Fig 2A-2C). Next, we assessed ISC proliferation rate through quantification of GFP-positive and PH3-positive cells and found that the double knockdowns rescued the hyperproliferation observed in the OGA single-knockdown midguts (Fig 2A-2C). These findings suggested that PNG1 was required for ISC hyperproliferation in the OGA knockdown midguts. Next, we assessed O-GlcNAc and PNG1 levels in the double-mutant fly midguts. There were high levels of O-GlcNAc and PNG1 staining in the single OGA knockdown and low levels of staining in the Png1 C303A single mutant midgut ISCs/EBs compared to control (Fig 2D-2G). We found a significant decrease in both O-GlcNAc and PNG1 levels in esg ts >OGA RNAi +Png1 C303A midgut ISCs/EBs compared to esg ts >OGA RNAi ISCs/EBs (Fig 2D-2G). Additionally, the elevated DHE (dihydroethidium) signal (ROS detection marker) detected in the single OGA-RNAi mutant was rescued when combined with the Png1 C303A mutant in ISCs/EBs (Fig 2H and 2I). Therefore, OGA knockdown-induced ISC proliferation, increased PNG1, and elevated ROS levels were rescued by loss of PNG1 function in ISCs/EBs.

Regulation of PNG1 and OGT levels in ISCs/EBs
To understand how O-GlcNAc might be regulated by PNG1, we assessed the association between PNG1 and OGT protein levels. We utilized a fly in which a Myc tagged OGT is overexpressed. This allows detection of the overexpressed OGT protein using an anti-Myc antibody. We examined the levels of Myc-OGT in esg ts >Myc-Ogt, esg ts >Png1 RNAi , and esg ts >Myc-Ogt+Png1 RNAi . We found that Myc-OGT levels significantly decreased in esg ts >Myc-Ogt+-Png1 RNAi midgut compared to esg ts >Myc-Ogt midgut ISCs/EBs (Fig 3A and 3B). This indicates that PNG1 was required to maintain Myc-OGT stability. Next, we tested if OGT was required for expression of PNG1. Analysis of NGLY1 levels in OGT knockdown ISCs/EBs indicated a significant decrease as compared to control (Fig 3C and 3D). Therefore, both OGT and PNG1 are required for each other's normal expression patterns, suggesting a mechanism by which PNG1 and OGT interaction contributes to ISC/EB proliferation.

Induction of O-GlcNAcylation in ECs rescues gut dysfunction induced by PNG1 knockdown
Interestingly, our data indicated that ISC hyperproliferation was induced by the loss of PNG1 in ECs (Fig 1). We also observed that hyperproliferation, increased levels of PNG1, O-GlcNAc and ROS induced by ISC/EB-knockdown of OGA were rescued by loss of PNG1 activity ( Fig  2). Thus, we wanted to explore whether PNG1-related hyperproliferation is rescued by elevating O-GlcNAcylation in ECs. To do this, we combined the Png1 C303A mutant with knockdown of OGA in ECs using an EC-specific GAL4 system. We have previously reported that ISC proliferation within the midguts of EC-specific OGA knockdown was similar to control, but ISC proliferation and cell death increased in EC-specific OGT knockdown [27]. After incubation at 29˚C for 5 days, we observed a significant increase in PH3-positive cells and cCaspase-positive cells (apoptosis marker) in Myo1A ts >Png1 C303A midgut compared to controls (Fig 4A-4D), consistent with the results in Fig 1. Interestingly, there were significant decreases in both PH3-positive cell and caspase-positive cell staining in midguts from the Myo1A ts >Oga RNAi +-Png1 C303A double-mutant flies compared to esg ts >Png1 C303A midguts (Fig 4A-4D). Thus, these data suggest that EC-specific PNG1 deficiency-induced ISC hyperproliferation and cell death can be rescued by OGA knockdown, indicating that increased O-GlcNAcylation is responsible for this rescue. Next, we assessed PNG1 and O-GlcNAc levels in these doublemutant fly midguts. PNG1 and O-GlcNAc levels were undetectable in Myo1A ts >Png1 C303A midgut GFP-positive cells (EC cells) (Fig 4E-4H). Conversely, NGLY1 and O-GlcNAc levels were significantly elevated in Myo1A ts >Oga RNAi midgut GFP-positive cells (EC cells) (Fig 4E  and 4F). Interestingly, NGLY1 and O-GlcNAc levels were significantly decreased in Myo1A ts >Oga RNAi +Png1 C303A double-mutant fly midgut GFP-positive cells as compared to Myo1A ts >Oga RNAi (Fig 4E-4H). These findings indicate that the gut dysfunction associated with loss of PNG1 activity could be rescued by loss of OGA in ECs. Thus, PNGase and O-staining (red) in midgut of flies. (I) Quantification of DHE mean fluorescence per esg-GFP-positive cell from the indicated genotype. Outline indicates esg-positive cells. Data are represented as mean ± SD. � p< 0.05. ���� p< 0.0001., see S1 GlcNAc appear to act within the same pathway in ISC/EBs and ECs to maintain intestinal homeostasis.

Consequences of OGT or PNG1 loss in ISCs/EBs can be rescued by modulating CncC activities
We have shown that both OGT and PNG1 are required for ISC proliferation and that both O-GlcNAc and PNG1 levels increase under stress conditions [27] (Fig 1). Cap'n'collar (CncC, the Drosophila Nrf2 homologue) is a transcription factor that regulates cellular redox homeostasis which has been shown to crosstalk with both O-GlcNAc and PNG1 [30]. Therefore, we investigated if we could rescue the proliferative in OGT knockdown and PNG1 mutation through modulation of CncC activity. To increase CncC activity, we used the chemical activator Oltipraz and examined ISC proliferation in esg ts >Png1 RNAi , esg ts >Png1 C303A and esg ts >Ogt RNAi flies treated with Oltipraz or left untreated. The number of GFP-positive cells and PH3-positive cells were not significantly changed in the Oltipraz-treated control midguts (Fig 5A-5C). Interestingly, GFP-positive and PH3-positive cells increased with Oltipraz-treatment in esg ts >Png1 RNAi , esg ts >Png1 C303A and esg ts >Ogt RNAi midguts compared to the non-treatment groups (Fig 5A-5C). Because proliferation in progenitor cells is often associated with apoptosis of differentiated cells, we examined apoptotic markers in these mutant midguts. We found that PNG1 or OGT knockdown alone did not impact apoptosis (S3 Fig). However, there was a significant increase in cCaspase-3 staining in differentiated cells after Oltipraz treatment in

PLOS GENETICS
Cytosolic O-GlcNAcylation and PNG1 maintain Drosophila gut homeostasis

PLOS GENETICS
both PNG1 and OGT knockdown midguts compared to non-treated groups (S3 Fig). Thus, CncC activation rescued ISC proliferation and promoted apoptosis of differentiated cells. Therefore, activation of CncC can rescue the depressed ISC proliferation of the OGT and PNG1 knockdown ISCs/EBs. We further investigated whether there was poly-ubiquitination (Poly-UB) accumulation (protein aggregation marker) and changes in proteasome activity in PNG1 knockdown associated with CncC activation. CncC is critical for activity of the proteasome and a substrate of the 26S proteasome in Drosophila S2 cells. In addition, the 26S proteasome is important for maintaining low levels of CncC under normal conditions but under stressed conditions CncC promotes proteasomal gene expression [31]. NGLY1 removes N-glycans from misfolded proteins before they can be degraded by the proteasome [17]. Our data showed that poly-UB accumulation increased in in PNG1 knockdown ISCs/EBs (Fig 5D and 5E). This increase could be rescued by CncC overexpression and chemical activation (Fig 5D and 5E). The increase in Poly-UB phenotype was associated with changes in 26S proteasome expression in the PNG1 knockdown, that was also rescued with CncC overexpression or chemical activation (Fig 5F and 5G). Therefore, the rescue of PNG1 by CncC activation was related to the regulation of protein aggregation and proteasomal degradation.

Loss of OGT or PNG1 in ISCs/EBs can be rescued by modulating ENGase activities
Endo-β-N-acetylglucosaminidase (Engase), is a cytoplasmic glycosidase that cleaves glycans on misfolded proteins. Importantly, lethality of NGLY1 deficient mice was partially rescued by additional deletion of Engase [17]. Thus, we wanted to define the interaction between PNG1, OGT and ENGase and test whether PNG1 and OGT mutant phenotypes could be rescued using an ENGase inhibitor (Rabeprazole; proton pump inhibitor) [32]. First, we assessed ENGase expression in ISCs/EBs-specific PNGase and OGT knockdown midguts using an anti-ENGase antibody. Interestingly, we found that ENGase levels were upregulated in esg ts >Png1 R-NAi and esg ts >Ogt RNAi midgut ISCs/EBs compared to control (Fig 6A and 6B). However, ENGase levels decreased upon Rabeprazole treatment (Fig 6A and 6B). To test whether ENGase inhibition could rescue the proliferation phenotypes of the PNG1 or OGT knockdown, we assessed GFP-positive and PH3-positive cells in esg ts >Png1 RNAi and esg ts >Ogt RNAi treated with Rabeprazole and found significant increases in proliferation as compared to the non-treated controls (Fig 6C-6E). Consistently, cCaspase-3 staining was elevated in differentiated cells from the Rabeprazole treated PNG1 and OGT knockdown midguts compared to non-treated groups (S3 Fig). To look more specifically at ERAD and ER stress and tease out if extracellular N-Glycosylation might be playing a role, we assessed ISC proliferation and levels of ENGase using double combination PNG1 mutant and ENGase knockdown flies. Our data showed that the number of GFP-positive and PH3-positive cells of esg ts >ENGase RNAi were similar to control, and ENGase was undetectable in ENGase RNAi ISCs/EBs (Fig 6F, 6G and  6H). Interestingly, PNG1 levels did not change in ENGase RNAi ISCs/EBs compared to control (Fig 6J). We found that ISC proliferation and PNG1 levels in PNG1 mutant flies were rescued by ENGase knockdown (Fig 6F, 6G, 6J and 6K). The increased ENGase levels were rescued by ENGase knockdown in ISCs/EBs (Fig 6H and 6I). We next investigated protein aggregation and assessed poly-ubiquitination accumulation in PNG1 mutant ISCs/EBs. PNG1 knockdown alone increased poly-ubiquitination accumulation, however, this was rescued by ENGase knockdown in ISCs/EBs (Fig 6L and 6M). Thus, our data indicated that protein aggregation and ENGase activity was likely associated with stem cell proliferation in loss of PNG1. Moreover, PNG1-related gut homeostasis is controlled through ENGase activity. Therefore, our

EC-specific OGT or PNG1 knockdown-induced ISC proliferation was rescued by CncC or ENGase activities
Our data indicated ISC-specific OGT or PNG1 knockdown induced ENGase expression and decreased ISC proliferation, which could be rescued by ENGase inhibition or CncC activation (Figs 5 and 6). Next, we assessed whether ENGase inhibition or CncC activation could also rescue the phenotypes of PNG1 or OGT knockdown in differentiated cells. As shown Fig 7A and  7B, ISC hyperproliferation of Myo1A ts >Png1 RNAi and Myo1A ts >Ogt RNAi was rescued by Rabeprazole treatment. Interestingly, ENGase expression was increased in Myo1A ts >Png1 RNAi and Myo1A ts >Ogt RNAi midgut and was decreased in the Rabeprazole treated compared to nontreated groups (Fig 7C). Further, EC-specific PNG1 knockdown-induced cell death was decreased in Myo1A ts >Png1 RNAi with Rabeprazole ( Fig 7D). Thus, in ECs, loss of OGT or PNG1 induced ISC proliferation, ENGase, and cell death was rescued by inhibition of ENGase. Next, we tested whether CncC activation could rescue hyperproliferation of EC-specific PNG1 or OGT knockdown. As shown Fig 7E and 7F, ISC hyperproliferation of Myo1A ts >Png1 RNAi and Myo1A ts >Ogt RNAi was rescued by Oltipraz treatment. Moreover, EC-specific PNG1 knockdown-induced cell death was decreased in Myo1A ts >Png1 RNAi with Oltipraz ( Fig 7G). Thus, loss of OGT or PNG1 induced ISC proliferation and cell death was rescued by CncC activation. Therefore, ENGase and CncC regulate OGT and PNG1 to maintain tissue homeostasis. A model depicting the roles of PNG1 and OGT in ISCs/EBs and ECs resulting in altered proliferation and apoptosis is shown in Fig 7H. Thus, O-GlcNAcylation and PNG1 have key roles in both progenitor (ISCs/EBs) and differentiated cells (ECs) contributing to tissue homeostasis. PNG1 regulates stem cell proliferation and is elevated in OGA knockdown midgut ISCs/EBs. PNG1 knockdown was rescued by O-GlcNacylation-induced dysplasia. PNG1 and OGT regulate each other's levels in ISCs/EBs. PNG1 knockdown-induced gut dysfunction was rescued by induction of O-GlcNacylation in ECs. OGT and PNG1 knockdown in ISCs/ EBs induced ENGase, resulting in reduced proliferation. In ECs, OGT and PNG1 knockdown induced ENGase and cell death further promoting hyperproliferation of ISC/EBs. Phenotypes of OGT and PNG1 knockdown in ISCs/EBs or ECs were rescued by modulation of ENGase or CncC level (Fig 7H). Taken together, OGT and PNG1 are important factors regulating ISC proliferation, and cell death in Drosophila whose effects can be mitigated by modulating CncC and ENGase.

Discussion
Intestinal stem cells regulate tissue homeostasis by balancing self-renewal, proliferation, and differentiation all of which are supported by elevated flux through the HBP [26]. Both N-linked glycosylation and intracellular O-GlcNAc modifications are regulated by the HBP pathway in a nutrient-sensing manner [2,3]. However, how NGLY1 is utilized to control stem cell homeostasis and differentiation in cells remains largely unknown. This is a critical question as   [14,15]. Elevation of NGLY1 was observed in patients' tumor samples, suggesting a function in oncogenic signaling [20]. In Drosophila, PNG1 mutants had severe developmental defects and reduced viability, with the surviving adults frequently sterile [16]. Here, we identified a pathway by which PNG1 regulates ISC homeostasis in vivo. In this study, we show that PNG1 levels increased in ISC/EBs concomitant with O-GlcNAc. This interaction between PNG1 and O-GlcNAcylation is critical for maintaining normal ISC proliferation and differentiation. Thus, through their mutual regulation, OGT and PNG1 have key roles in both progenitor (ISCs/EBs) and differentiated cells (ECs) contributing to tissue homeostasis. Previous reports indicated that PNG1 null larvae have specific developmental abnormalities in their midgut that contributes to their lethality [28]. Further, intestinal inflammation in Crohn's disease is associated with increased O-GlcNAc modification [29]. Our previous study also showed that increased O-GlcNAc promotes gut dysplasia through regulation of DNA damage [27]. Thus, PNG1 or O-GlcNAc might still be associated with gut dysfunction in a disease context.
The regulation of O-GlcNAc by PNG1 and the interaction between PNG1 and O-GlcNAc has been implicated previously. In fact, GlcNAc supplementation partially rescued lethality associated with PNG1 knockdown [33]. Although the mechanism by which GlcNAc supplementation rescued these mutant flies has not been fully worked out, Gfat1 transcript levels were downregulated in PNG1 knockdown flies [33]. Gfat1 is the enzyme that controls the rate limiting step in the HBP to produce UDP-GlcNAc. Thus, PNG1 through regulation of Gfat1 could impact levels of UDP-GlcNAc and ultimately O-GlcNAc [26,33]. Additionally, it has been hypothesized that the loss of PNG1 could increase the presence of intracellular N-GlcNAc modification, potentially interfering with O-GlcNAc mediated signaling [34]. Therefore, alterations in UDP-GlcNAc levels or presence of intracellular N-GlcNAc upon PNG1-deficiency can interact with O-GlcNAc to regulate stem cell homeostasis.
Previous reports have shown that Nrf1 undergoes NGLY1-mediated deglycosylation, followed by proteolytic cleavage and translocation into the nucleus as an active transcription factor [13]. Loss of NGLY1 caused Nrf1 dysfunction, as evidenced by an enrichment of deregulated genes encoding proteasome components and proteins involved in oxidation reduction [33]. Proteasome activity can induce an apoptotic cascade that leads to growth arrest and, subsequently, cell death. Our data indicated that PNG1 or OGT knockdown suppressed ISC proliferation, which was rescued by Oltipraz (CncC activation) treatment in ISCs/EBs ( Fig  5). Furthermore, there was increased apoptosis in PNG1 or OGT knockdown with treatment compared to non-treated groups (S3 Fig). Interestingly, we found O-GlcNAc-induced intestinal dysplasia was rescued by knockdown of PNG1 in ISCs/EBs through regulation of ROS levels (Fig 2G). Similarly, increases in global O-GlcNAcylation in embryos of diabetic mice caused an overproduction of ROS and subsequent oxidative and ER stress. It is known that activation of SKN-1A/Nrf1 also requires deglycosylation by PNG-1/NGLY1 in C. elegans [18]. Further, SKN-1 is O-GlcNAc modified [35] and translocates to the nucleus in ogt-1(ok430)null worms [36]. Together, these studies all suggest conserved functional connections between O-GlcNAc and Nrf family transcription factors [37]. Here, we also showed EC-specific OGT or PNG1 knockdown-induced hyperproliferation and cell death was decreased by CncC activation (Fig 7). This data indicated OGT or PNG1 can be regulated by CncC activity in ISCs/ EBs and ECs. CncC has high activity within ISCs/EBs of unstressed as well young ISCs and positive cells (green) in midgut of flies from the indicated genotype with or without Oltipraz. (H) O-GlcNAc and PNG1 cooperate to influence intestinal homeostasis in Drosophila which can be modulated by Nrf2 and ENGase activity. Data are represented as mean ± SD. � p< 0.05. ��� p< 0.001. ���� p< 0.0001. n.s., not significant., see S1 Table for N values. https://doi.org/10.1371/journal.pgen.1010128.g007 quiescent ISCs but decreases with age and damage [38,39]. These data indicated that CncC acts to properly balance between signaling and damage responses necessary for tissue homeostasis [38]. We found that CncC activation increased ISC proliferation in ISCs/EBs and decreased ISC proliferation in ECs of OGT or PNG1 knockdown contributing towards tissue homeostasis. Another study showed that inhibition of NGLY1 resulted in Nrf1 being misprocessed, mislocated, and inactive, thus indicating that functional NGLY1 is essential for Nrf1 processing, nuclear translocation, and transcription factor activity [40]. Therefore, our data suggests that PNG1 and OGT modulated by CncC activation contribute to ISC proliferation and ultimately regulating tissue homeostasis. Nrf2 activation was able to rescue the developmental growth of NGLY1 deficiency in worm and fly models [41]. In cancer-initiating cells, ER stress-dependent (ROS-independent) CncC induction is an event necessary to maintain stemness [42]. Our data showed that PNG1 knockdown-induced Poly-UB accumulation and 26S proteasome expression that was rescued by CncC overexpression and chemical activation (Fig 5). Through functioning as a sensor of cytosolic proteasome activity and an activator of aggresomal formation, Nrf2 alleviates cell damages caused by proteasomal stress [43]. Expression of proteasome subunit genes and mitophagy-related genes were broadly enhanced after sulforaphane (Keap1 inhibitor) treatment and pharmacologically induction of Nrf2 promotes mitophagy and ameliorates mitochondrial defect in Ngly1 −/− cells [42]. Thus, we believe that the sensitized background of the OGT or PNG1 mutant provides an environment where CncC activation promotes proliferation to the normal level through regulation of proteasome activity and protein aggregation.
In our previously published paper [27], we showed OGT overexpression and OGA knockdown in ISCs/EBs both increased O-GlcNAc levels and induced hyperproliferation of the stem cells, whereas OGT knockdown decreased proliferation. However, in differentiated ECs, OGT overexpression and OGA knockdown phenotypes were similar to the normal gut, whereas OGT knockdown elevated proliferation and cell death. In general, EC death promotes proliferation in order to maintain gut homeostasis [24]. Here, NGLY1 knockdown in ISCs/EBs decreased proliferation and clone size (Figs 1 and S2) but NGLY1 knockdown in ECs induced hyperproliferation and cell death and importantly decreased O-GlcNAc levels. Thus, the phenotypes of OGT and NGLY1 were similar, demonstrating that maintenance of OGT and NGLY1 protein expression is highly interdependent for the maintenance of tissue homeostasis. It is interesting that the progenitor and differentiated cell types within the gut respond differently to changes in O-GlcNAc. It is possible that a certain level of O-GlcNAcylation is needed to maintain stem cells and promote proliferation and self-renewal, however, differentiated cells that do not have the same energy and growth requirements are not as reliant on high levels of O-GlcNAc. On the other hand, both ISC/EBs and ECs require some level of O-GlcNAc and without OGT there is decreased proliferation in progenitor cells and increased cell death of ECs. There are a few possibilities how NGLY1 and OGT can collaboratively work, however, it is unlikely that they share protein targets. First, a previous publication showed that additional deletion of ENGase, another N-deglycosylating enzyme that leaves a single GlcNAc residue, alleviates some of the lethality of Ngly1-deficient mice [17]. Thus, it is possible with the accumulation of aggregation prone intracellular N-GlcNacylated proteins, there is disruption of normal O-GlcNac signaling. Our data also showed increased protein aggregation in OGT or NLGY1 knockdown that was rescued by ENGase knockdown (Fig 6). In addition, MYC-OGT protein levels in OGT overexpression fly guts were decreased by PNG1 knockdown (Fig 3). It is possible that loss of NGLY1 disrupts normal OGT degradation and thus impacts levels global of O-GlcNAcylation.
In this study we have shown that ENGase levels increased in PNG1 or OGT knockdown ISCs/EBs and ECs (Figs 6 and 7). PNGase is involved in the process of endoplasmic reticulum associated degradation (ERAD), acting as a deglycosylating enzyme that cleaves N-glycans attached to ERAD substrates [12]. The small molecule ENGase inhibitors have potential to treat pathogenesis associated with NGLY1 deficiency [44]. Rabeprazole, a proton pump inhibitor, was identified as a potential ENGase inhibitor [44]. We demonstrated that the consequences of knockdown of OGT or PNG1 on ISC proliferation and ENGase activity was rescued by Rabeprazole treatment in ISCs/EBs or ECs (Figs 6 and 7). Our data showed that cell death was elevated in ISCs/EBs-specific PNG1/OGT knockdown with Rabeprazole treatment compared to non-treated groups concomitant with an increase in ISC proliferation (S4 Fig). On the other hand, cell death decreased in EC-specific PNG1 knockdown treated with Rabeprazole resulting in a decrease in ISC proliferation (Fig 7). It is known that loss of PNG1 function in cells can cause the accumulation of aberrant proteins in the cytosol and the interruption of ERAD. Further, downregulation of ER stress-related genes has been reported in B-cell-specific OGT mutant mice [45]. The protective effects of O-GlcNAc are not limited to mitochondrial function but also rescue injury caused by ER stress [46]. Therefore, NGLY1/ OGT seems to be functionally associated with the ERAD machinery [17]. More recently, using a model ERAD substrate, it was reported that the ablation of Ngly1 causes a disruption in the ERAD process in mouse embryonic fibroblast (MEF) cells [47]. Moreover, lethality of mice bearing a knockout of the Ngly1-gene was partially rescued by the additional deletion of the Engase gene [17]. Interestingly, we showed that OGA knockdown rescued ENGase levels of PNG1 knockdown ISCs/EBs (S4 Fig). Hence, these findings suggest that there is a correlation between OGT/PNG1 and ENGase contributing to tissue maintenance.
Taken together, our findings implicate O-GlcNAc and PNG1 as key regulators of tissue maintenance. PNG1 can impact stem cell homeostasis through regulation of O-GlcNAc both in ISCs/EBs or ECs. Of significance is the finding that PNG1 and OGT phenotypes are rescued by modulating CncC and ENGase activity in ISCs/EBs or ECs. Thus, our findings reveal that nutrient-driven glycosylation contribute towards control of ISC and progenitor cell proliferation and EC cell death via regulation of CncC and ENGase. Our study provides a platform for future designs of interventions in which changes in O-GlcNAc can be utilized as a therapeutic for stem-cell-derived diseases like cancer. We also present a molecular mechanism and unexpected pathway that can be targeted for treating NGLY1-deificient patients.

Materials availability
This study did not generate new unique reagents.

Drosophila stocks, culture, and husbandry
Fly stocks were maintained at 25˚C on standard food under a *12 h/12 h light/dark cycle. Food consisted of 15.8g yeast, 9g soy flour, 5.2g agar, 67g cornmeal, and 0.5% propionic acid. To avoid larval overpopulation, <30 adult flies per vial were transferred to new food vials every 2-3 days.
For transgene expression at specific developmental stages, the Gal80ts technique was used. The flies were set up and maintained at 22˚C until adulthood. After maintaining the flies at 29˚C, the midguts were dissected and analyzed.
Paraquat feeding assay. Flies were treated with 10 mM paraquat for 20-22 h at 29˚C. After feeding, the midguts were dissected and analyzed.
Oltipraz feeding assay. Two-day-old flies were fed 100 uM Oltipraz (Selleckchem, #S7864) or DMSO only mixed in standard food for 7 days or 5 days at 29˚C. Flies were transferred to new Oltipraz-containing food vials every 2 days.
Rabeprazole feeding assay. Two-day-old flies were fed 1 mM Rabeprazole (Selleckchem, #S4845) or DMSO only mixed in standard food for 7 days or 5 days at 29˚C. Flies were transferred to new Rabeprazole -containing food vials every 2 days. DHE assay. Fly midguts were dissected in Schneider's medium (HyClone). After incubation in 30 μM DHE (Invitrogen) for 7 min in the dark at room temperature, midguts were washed three times and mounted. Images were captured immediately with a Zeiss LSM700 confocal microscope.
MARCM. For MARCM experiments, flies were maintained at 23˚C until 3-5 days after eclosion, heat-shocked at 37˚C for 60 min, and then maintained back at 23˚C before dissection and markers were analyzed at 7 day after induction.
Quantitative of PH3 positive cell. To quantitatively analyze PH3-positive cells, the number of PH3-positive cells in the whole gut was counted. N represents the number of guts.
Quantitative of Delta-or GFP-positive cell. To quantitatively analyze Delta-or GFPpositive cells, the number of Delta-or GFP-positive cells in the field was counted. N represents the number of guts.
Measurement of O-GlcNAc, NGLY1, Myc-OGT, DHE, 26S proteasome, poly-UB, and ENGase fluorescence in ISC. The fluorescence images of O-GlcNAc, NGLY1, Myc-OGT, DHE, 26S proteasome, poly-UB, and ENGase staining were captured at the same exposure time in each experiment and was measured by quantifying the level of fluorescence (IHC staining) in individual ISCs normalized to nearby background in FIJI (ImageJ) software. The mean fluorescence was analyzed after exclusion of the mean of the background region (from two spots, excluding the nuclear portion in the posterior midgut), with background fluorescence set to 0. At least 10 ISCs were quantified in each image, and >10 images (>1 image per fly) were used to calculate the average intensity of fluorescence in ISCs of each fly. N represents the number of guts. n represents the number of cells.

Statistical analysis
Data representation and statistical analysis were performed using GraphPad Prism software. Statistical analysis was performed using a t-test and multiple comparisons were performed with a One-Way ANOVA. All experiments were replicated independently 2-3 times. N represents the number of guts. n represents the number of cells.