Intellectual disability-associated disruption of O-GlcNAc cycling impairs habituation learning in Drosophila.

O-GlcNAcylation is a reversible co-/post-translational modification involved in a multitude of cellular processes. The addition and removal of the O-GlcNAc modification is controlled by two conserved enzymes, O-GlcNAc transferase (OGT) and O-GlcNAc hydrolase (OGA). Mutations in OGT have recently been discovered to cause a novel Congenital Disorder of Glycosylation (OGT-CDG) that is characterized by intellectual disability. The mechanisms by which OGT-CDG mutations affect cognition remain unclear. We manipulated O-GlcNAc transferase and O-GlcNAc hydrolase activity in Drosophila and demonstrate an important role of O-GlcNAcylation in habituation learning and synaptic development at the larval neuromuscular junction. Introduction of patient-specific missense mutations into Drosophila O-GlcNAc transferase using CRISPR/Cas9 gene editing leads to deficits in locomotor function and habituation learning. The habituation deficit can be corrected by blocking O-GlcNAc hydrolysis, indicating that OGT-CDG mutations affect cognition-relevant habituation via reduced protein O-GlcNAcylation. This study establishes a critical role for O-GlcNAc cycling and disrupted O-GlcNAc transferase activity in cognitive dysfunction, and suggests that blocking O-GlcNAc hydrolysis is a potential strategy to treat OGT-CDG.

Introduction

O-GlcNAcylation is an essential and dynamic co-/posttranslational modification that is characterized by the attachment of an N-acetylglucosamine (GlcNAc) molecule to serine or threonine residues of intracellular proteins. O-GlcNAcylation is implicated in a wide range of cellular processes, such as: chromatin remodeling [1-3], transcription [4,5] and translation [6], Ras-MAPK and insulin signaling [7-9], glucose homeostasis [10], mitochondrial trafficking [11], and control of the circadian clock [12]. The addition and removal of the O-GlcNAc modification, termed O-GlcNAc cycling, is controlled by two evolutionarily conserved enzymes, O-GlcNAc transferase (OGT) and O-GlcNAc hydrolase (OGA).

OGT, responsible for the addition of O-GlcNAc, is abundantly expressed in neurons and is enriched in the postsynaptic density (PSD) [13], a protein-dense structure that organizes the postsynaptic signaling machinery. O-GlcNAcylation is altered in brains of patients with Alzheimer’s disease [14] and animal and cellular models of major neurodegenerative diseases [15-19]. In C. elegans models of neurodegeneration, O-GlcNAcylation protects against neurotoxicity [20]. Furthermore, it plays an important role in neuronal regeneration through the synchronization of insulin signaling-dependent regenerative processes [8]. Recent studies also point to important functions in neuronal development, such as neuronal differentiation [21,22], assembly and axonal transport of neurofilaments [23], and synapse maturation [24]. Missense mutations in human OGT gene, located on the X chromosome, are associated with intellectual disability (ID) [25-31], a severe neurodevelopmental disorder that is characterized by impaired cognition. Patients with OGT mutations suffer from a wide array of clinical features, including intrauterine growth retardation, developmental delay, delayed or restricted language skills and severe learning difficulties. The syndrome has been termed OGT-associated Congenital Disorder of Glycosylation (OGT-CDG) [32]. These findings and animal studies suggest that O-GlcNAcylation plays an important function in cognitive processes, such as learning [33,34].

The OGT protein consists of an N-terminal tetratricopeptide (TPR) domain, which contributes to substrate recognition and binding [35] and a C-terminal catalytic domain that is responsible for glycosylation of the TPR-bound proteins. OGT-CDG mutations have been found in both domains. Unlike the mutations in the catalytic domain, the mutations in the TPR domain have not been shown to significantly affect global protein O-GlcNAc levels but they do affect the OGT-TPR domain substrate binding and glycosylation kinetics, as derived from in vitro assays and crystal structure analysis [26,36,37]. However, it remains to be known whether impaired glycosylation is the mechanism that leads to developmental and cognitive defects caused by identified TPR domain mutations. This question is pertinent as the TPR domain has been shown to be essential for cellular functions other than glycosylation [38].

Drosophila as a model organism has contributed to understanding the disease pathogenesis of numerous (ID) syndromes. It offers a combination of well-established gene targeting approaches and a plethora of morphological and functional disease-relevant phenotypes. The Drosophila OGT orthologue is encoded by the polycomb group gene super sex combs (sxc) [39]. It is highly similar to human OGT [40]. Complete loss of sxc results in severe homeotic transformations of adult body structures and pupal lethality [41]. This lethality can be restored by ubiquitous overexpression of human OGT, demonstrating functional conservation [39]. sxc has also been associated with neuronal function in circadian rhythm regulation [42]. Therefore, Drosophila is highly suited to investigate the disrupted mechanisms underlying OGT-CDG.

Here we genetically manipulated both O-GlcNAc transferase and O-GlcNAc hydrolase activity using established sxc and Oga mutants and investigated the effect of O-GlcNAc cycling on cognitive function and neuronal development. We turned to habituation, an evolutionary conserved form of non-associative learning that is characterized by response decrement towards a repeated, non-meaningful stimulus. At the neuronal level, habituation is mediated by adaptive changes in the excitatory activity of the stimulus response pathway, causing attenuated downstream neuronal responses to repeated, familiar stimuli. Habituation thus serves as a filter mechanism that prevents information overload and allows cognitive resources to focus on relevant stimuli [43]. It represents a prerequisite for higher cognitive functions [44-46]. Deficits in habituation have been reported in a number of neurodevelopmental disorders [47] and in more than a hundred Drosophila ID models [48,49]. Synaptic morphology at the Drosophila neuromuscular junction, an established model synapse [50], was assessed as a measure of neuronal development. To independently validate our findings we targeted the endogenous sxc locus by CRISPR/Cas9 editing [51] and generated a strong hypomorph mutation in the O-GlcNAc transferase domain (sxc). With the same technique we introduced three ID-associated missense mutations found in the conserved TPR domain of OGT: R284P [28], A319T [25] and L254F [29] and generated equivalent sxc, sxc and sxc alleles, respectively. We evaluated their effect on protein O-GlcNAcylation, developmental viability, adult lifespan and locomotor activity, and assessed their effect on cognitive function and neuronal development.

We find that appropriate O-GlcNAc cycling is required for habituation, a fundamental form of learning that is widely disrupted in Drosophila models of ID, and for synaptic development. We show that OGT missense mutations, implicated in ID, outside the catalytic O-GlcNAc transferase domain lead to deficits in habituation learning and that these deficits are caused by disrupted O-GlcNAc transferase activity. We thus unambiguously demonstrate the role of O-GlcNAcylation in brain development and function.

Results

Alteration of O-GlcNAc transferase activity leads to a deficit in habituation

To investigate whether the catalytic activity of Drosophila OGT is required for cognitive function, we tested the effect of an sxc mutation with diminished catalytic activity, sxc, on habituation. The homozygous hypomorphic flies are viable and except mild wing vein and scutellar bristle phenotypes do not present with any morphological abnormalities [51]. In habituation, an initial strong response towards a repeated but harmless stimulus gradually wanes based on prior experience. To assess this phenotype we used light-off jump habituation, an established non-associative learning assay that meets the strict habituation criteria, including spontaneous recovery and dishabituation with novel stimulus and excluding sensory adaptation and motor fatigue [46,52]. We subjected sxc homozygous (sxc), heterozygous (sxc), and genetic background control flies (+/+) to 100 light-off pulses in the light-off jump habituation assay. While sxc and control flies exhibited good initial jump responses to the light-off stimuli (61% and 67% initial jumpers out of N = 96 tested flies per genotype; above a required threshold of 50% [49]), sxc flies were impaired (36% initial jumpers, N = 64), identifying broader defects that preclude assessment of habituation. Compared to control flies that habituated quickly to the repeated light-off stimulus, sxc flies displayed significantly slower habituation and needed significantly more light-off pulse trials to suppress their jump response (Trials To no-jump Criterion, TTC). Some mutant flies were not able to suppress their jump response during the entire course of the experiment, as reflected by the high baseline of the average jump response curve (). These results suggest that partial loss of O-GlcNAc transferase activity or altered O-GlcNAcylation kinetics in sxc mutants impairs habituation.

Catalytic activity of O-GlcNAc transferase in neurons is required for habituation learning.

Jump responses were induced by 100 light-off pulses with 1 s intervals between pulses. The jump response represents the % of jumping flies in each light-off trial. The mean number of trials that flies needed to reach the no-jump criterion (Trials To Criterion, TTC) ± SEM is also shown. (A) Defective habituation of sxc flies (N = 59, mean TTC ± SD: 12.8 ± 6.7 p = 0.001, in red) compared to their respective genetic background control flies (+/+, mean TTC ± SD: 4.4 ± 0.9, N = 61, in blue). ** p<0.01, based on lm analysis. (B) Habituation defect of sxc flies (N = 72, mean TTC ± SD: 15.5 ± 7.8 padj = 0.045, in red) is corrected by removing one Oga allele in sxc; Oga flies (N = 70, mean TTC ± SD: 10.1 ± 5.9, padj = 0.024, in cyan) to the level of control flies (N = 72, mean TTC ± SD: 9.4 ± 3.5, padj = 0.677, in blue). Habituation of Oga flies (N = 76, mean TTC ± SD: 17.1 ± 10.4, in purple) is slower but not significantly different from the control flies (padj = 0.467). (C) Defective habituation of elav-Gal4>UAS-sxc flies (N = 55, mean TTC ± SD: 12.8 ± 4, padj = 3.89x10-12, in dark blue) compared to control elav-Gal4/+ flies (N = 38, mean TTC ± SD: 2.6 ± 1.3 in light blue). (D) Habituation defect of sxc; elav-Gal4/+ flies (N = 40, mean TTC ± SD: 14.6 ± 6, padj = 4.92x10-12, in red) is corrected by selective expression of UAS-sxc in neurons (sxc; elav-Gal4>UAS-sxc, N = 43, mean TTC ± SD: 4.2 ± 1.3, padj = 8.94x10-7, in green), to the level of the genetic background control flies (+/+, N = 65, mean TTC ± SD: 4.4 ± 0.9, padj = 0.68, in blue). * padj<0.1, *** padj<0.001, n.s. not significant, based on lm analysis with Bonferroni-Holm correction for multiple comparisons. A complete list of p-values and summary statistics is provided in .

We validated our conclusion by employing a knockout-out allele of Drosophila Oga [53]. We asked whether partial inhibition of O-GlcNAc hydrolysis, by removing one copy of Oga (Oga) could improve habituation of the sxc flies. Habituation of Oga flies was also slower but not significantly different from the control flies. The transheterozygous sxc; Oga flies showed good initial jump responses, and their habituation was not significantly different to that of control flies, identifying a significant improvement compared to sxc flies (). The fatigue assay (see Materials and Methods) confirmed that the lower TTC values were not a result of increased fatigue (). These results show that Drosophila is a suitable model to study the role of OGT in cognitive functioning and demonstrate that tight control of protein O-GlcNAcylation is required for proper habituation learning in Drosophila.

O-GlcNAc transferase is required for habituation in neurons

We next sought to determine whether the sxc habituation deficit originates from reduced OGT function in neurons. We therefore induced neuronal knockdown of sxc by crossing the pan-neuronal elav-Gal4 driver line (see Materials and Methods) to an inducible RNAi line against sxc obtained from Vienna Drosophila Resource Center (#18610, zero predicted off-targets). Progeny from crossing the driver line to the isogenic genetic background of the RNAi line (#60000) were used as controls. While control flies (elav-Gal4/+) showed good initial jump response (56%, N = 64), elav-Gal4>UAS-sxc flies—similar to sxc flies—exhibited very low initial jump response to light-off stimulus (19% initial jumpers, N = 96). While this detrimental effect prevented assessment of sxc neuron-specific knockdown in habituation learning, it does argue that i) the failed jump response of sxc flies is likely due to loss of OGT activity in neurons, and ii) OGT activity is indispensable for basic neuronal function or neuronal development.

We used an alternative strategy to test whether habituation deficits of sxc flies are of neuronal origin. We asked whether restoration of OGT activity in neurons can correct habituation deficits of sxc flies, by inducing pan-neuronal overexpression of functional wild-type sxc in the heterozygous sxc as well as control background. Neuronal overexpression of functional sxc in control flies (elav-Gal4>UAS-sxc) resulted in a habituation deficit (). This is consistent with our previous findings of habituation deficits in Oga flies, which also show increased protein O-GlcNAcylation [53]. In contrast, re-expression of functional sxc in the sxc flies completely corrected their habituation deficits (). The lower TTCs were not a result of increased fatigue (). Therefore, appropriate levels of OGT activity and O-GlcNAcylation, specifically in neurons, are required for habituation learning.

Neuronal O-GlcNAc transferase activity controls synaptic development

Synapse biology is important for brain development and cognition, and abnormalities in synaptic architecture are characteristic of multiple Drosophila models of neurodevelopmental disorders [54-59]. For these reasons, we asked whether sxc mutants show any defects in the morphology of the third instar larval neuromuscular junction (NMJ), a well-established model synapse. We labeled NMJs by immunostaining for the postsynaptic membrane marker anti-discs large (Dlg1) to visualize the overall morphology of the NMJ terminal, and for synaptotagmin (Syt), a synaptic vesicle marker that visualizes synaptic boutons. We did not observe a significant change in synaptic length, area, or perimeter (), nor in the number of branches and branching points () in NMJs of sxc mutant larvae but we observed a significant increase in bouton number compared to the genetic background control (). Increasing OGT activity by presynaptic overexpression of wild-type sxc (elav-Gal4>UAS-sxc) resulted in an opposite phenotype: a decreased number of boutons compared to both controls, UAS transgene and driver alone. NMJ length was also reduced (). Both parameters were normalized when we neuron-specifically expressed functional sxc in the sxc mutant background (sxc, elav-Gal4>UAS-sxc) (). These results demonstrate a role of sxc in the synaptic bouton number and NMJ morphology and indicate that tight control of O-GlcNAcylation is important for normal synaptic development.

Both reduced and increased O-GlcNAc transferase activity in neurons cause defects in synaptic morphology.

Data presented as individual data points with mean ± SD. (A) NMJs on muscle 4 of sxc larvae have a significantly higher number of synaptic boutons (N = 29, in red) as compared to their genetic background control (+/+, N = 24, p = 0.009, in blue,) but not significantly different NMJ length (p = 0.872), area (p = 0.314) and perimeter (p = 0.935). ** p<0.01, based on one-way ANOVA. (B) Elav-Gal4>UAS-sxc larvae have a significantly lower number of boutons (N = 29, in dark blue) compared to their respective background controls elav-Gal4/+ (N = 30, padj = 2.1x10-4, in light blue) and UAS-sxc/+ (N = 26, padj = 0.009, in grey), significantly reduced NMJ length (padj/ = 0.013, padj/ = 0.02) and a smaller NMJ perimeter (padj/ = 0.008). (C) Neuron-selective expression of sxc in sxc larvae shows no change in bouton numbers (sxc; elav-Gal4>UAS-sxc (N = 29, in green)), compared to sxc, UAS-sxc/+ larvae (N = 28, padj = 0.102, in red) and compared to control (N = 27, padj = 0.977, in blue) and no change in NMJ length compared to control (+/+, N = 28, padj = 0.974) and sxc, UAS-sxc/+ larvae (padj = 0.935). Also NMJ area (padj = 0.849, padj = 0.691) and NMJ perimeter were not changed (padj = 0.066), padj/ = 0.995). (D) Number of synaptic boutons (padj = 0.003), NMJ length (padj = 0.01), area (padj = 0.028) and perimeter (padj = 4.5x10-4) are significantly increased in sxc larvae (N = 31, in brown) compared to the genetic background control larvae (+/+, N = 28, in blue) and partially normalized in the sxc; Oga larvae (N = 30, in cyan, boutons: padj/ = 0.011, padj/ = 0.923; length: padj/ = 0.896, padj/ = 0.001; area: padj/ = 0.501, padj/ = 0.431; perimeter: padj/ = 0.08, padj/ = 0.26). None of the parameters is significantly affected in the Oga larvae (N = 30, in purple; Oga experiments were performed simultaneously and first published here [53] with significantly increased bouton counts (p <0.05) without multiple testing correction). * padj <0.05, ** padj <0.01, *** padj <0.001, based on one-way ANOVA with Tukey’s multiple comparisons test. A complete list of p-values and summary statistics is provided in . (A’- D’) Representative NMJs of wandering third instar larvae labeled with anti-discs large 1 (Dlg, magenta) and anti-synaptotagmin (Syt, green). When appropriate, type 1b synapses are distinguished from other synapses with white arrow. Scale bar, 20μm. The quantitative parameter values of the representative images: (A’) (+/+ | sxc): #Boutons (31 | 39), Length (103.7 | 137.8), Area (374.4 | 369.4), Perimeter (245.5 | 306.7) (B’) (elav-Gal4/+ | UAS-sxc/+ | Elav-Gal4>UAS-sxc): #Boutons (45 | 37 | 28), Length (160.0 | 137.5 | 93.3), Area (480.4 | 478.4 | 363.9), Perimeter (407.6 | 290.8 | 243.9) (C’) (+/+ | sxc, UAS-sxc/+ | sxc; elav-Gal4>UAS-sxc): #Boutons (27 | 34 | 31), Length (107.6 | 144.9 | 114.9), Area (430.6 | 464.7 | 361.7), Perimeter (256.0 | 354.8 | 299.4) (D’) (+/+ | sxc | Oga | sxc; Oga): #Boutons (23 | 43 | 35 | 30), Length (111.5 | 205.6 | 122.9 | 142.5), Area (351.7 | 691.7 | 417.8 | 377.6), Perimeter (245.6 | 544.0 | 274.1 | 318.9).

The decrease of synaptic length caused by neuronal overexpression of sxc in the control but not sxc background indicates that strong dysregulation of sxc might be required to uncover its function in synaptic growth. Accordingly, we found a significant increase in NMJ length and perimeter in homozygous sxc larvae. The increase in the number of synaptic boutons did not reach significance (), potentially due to greater variability. This effect also did not withstand multiple testing correction in the sxc; UAS-sxc/+ larvae that were used as a control in the rescue experiment (). However, there is a quantitatively very consistent increase in bouton number across the three tested H537A mutant conditions (fold changes 1.19, 1.18 and 1.15). We therefore conclude that the bouton phenotype associated with H537A mutation is mild. To validate the synaptic bouton and length/growth phenotypes, we used CRISPR/Cas9 gene editing to generate a stronger catalytic hypomorph, sxc (). The in vitro catalytic activity of sxc was reported to be 3% relative to wildtype OGT activity, less than the reported catalytic activity of sxc (5.6% activity relative to wildtype) [40]. Indeed, we found total O-GlcNAc levels in sxc homozygous embryos to be reduced (). sxc homozygous flies are viable, confirming minimal requirement of endogenous OGT activity for completion of development in Drosophila (). We found that NMJs of sxc larvae display a significant increase in synaptic bouton number, NMJ length, area and perimeter (), reflecting a more severe NMJ phenotype and indicating the effect of sxc on synaptic morphology is mild and/or not fully penetrant. We also subjected the sxc flies to light-off jump habituation assay but homozygous as well as heterozygous sxc flies showed impaired jump response (38% and 42% initial jumpers), similar to homozygous sxc and pan-neuronal sxc knockdown flies. This precluded the assessment of habituation.

A knockout of Oga normalized bouton number and partially also the area and perimeter of the sxc NMJs (). These data show that O-GlcNAcylation controls bouton number and partially also NMJ size.

Characterization of development and locomotor function of sxc mutations associated with Intellectual Disability

Recent studies have reported three hemizygous missense mutations (R248P, A319T, L254F) in human OGT in male individuals with ID. The de novo R248P mutation was identified by trio whole exome sequencing in an affected individual with ID and developmental delay [28]. A319T and L254F mutations were identified by X chromosome exome sequencing. The A319T mutation, present in three individuals with severe ID, was inherited from the mother but segregated with an uncharacterized missense mutation in MED12, a gene already implicated in ID [25]. The L254F mutation was present in three related individuals with moderate to mild ID [27,29]. These mutations reside in the conserved TPR domain, outside of the catalytic O-GlcNAc transferase domain [25,27-29]. To investigate the functional consequences of these mutations, we introduced the equivalent missense mutations (R313P, A348T, L283F) into the sxc gene using CRISPR/Cas9 editing () and generated three novel sxc ID alleles sxc, sxc, and sxc.

We first characterized the development of the patient-related mutant sxc alleles. We transferred embryos at stage 11–16 to vials with fresh food and counted the number of resulting pupae and adult flies. Homozygous sxc, sxc, and sxc embryos developed normally to adulthood without apparent delay and the percentage of pupae and adults did not statistically differ from the genetic background controls ().

We next investigated locomotor phenotypes in adult sxc, sxc, and sxc flies using the island and negative geotaxis assays. In the island assay, flies were thrown onto a white platform surrounded by water, and the number of individuals remaining on the platform was quantified over time. Homozygous sxc, sxc and sxc flies escaped from the platform with similar efficiency as the genetic background control (), indicating that their startle response is not affected.

In the negative geotaxis assay, climbing performance of homozygous and heterozygous sxc flies was significantly slower while homozygous sxc and sxc flies showed an average climbing speed similar to the control (). We also tested Drosophila lines with homozygous H537A (sxc) or H596F (sxc) catalytic hypomorph mutations to investigate whether impaired O-GlcNAc transferase activity affects climbing speed in the negative geotaxis assay. The catalytic hypomorphs exhibited similar climbing speed as the control group (). This suggests that the deficit in coordinated locomotor behavior in the negative geotaxis assay of sxc flies is independent of O-GlcNAc transferase activity.

Locomotor and biochemical characterization of sxc, sxc and sxc flies.

(A) Climbing locomotor behaviour was assessed based on the climbing speed (mm/s) in an automated negative geotaxis assay. The sxc and sxc (N = 9) flies showed reduced climbing speed compared to background control (N = 9) indicating locomotor dysfunction. sxc, sxc, sxc and sxc flies (N = 9 for all genotypes) did not show significantly reduced climbing speed. Data presented as mean ± SD. * padj < 0.05 based on one-way ANOVA with Tukey’s multiple comparisons of mean climbing speed. A complete list of p-values and summary statistics is provided in . (B) Western blot on head samples from 1–4 days old male adult Drosophila indicate no significant alteration in the level of protein O-GlcNAcylation in sxc, sxc and sxc samples compared to the genetic background controls, while the homozygous Oga allele caused an increase of O-GlcNAcylation. Western blot was probed with a monoclonal anti-O-GlcNAc antibody (RL2). (C) Quantification of O-GlcNAcylated proteins revealed that protein O-GlcNAcylation in sxc, sxc and sxc flies remain at a similar level as in the control samples. Data presented as mean ± SD. * padj = 0.035, based on one-way ANOVA with Tukey’s multiple comparisons test, n = 3 for all lines. A complete list of p-values and summary statistics is provided in .

Taken together, similar to catalytic hypomorphs ([51] and ) the sxc, sxc, and sxc mutants are fully viable and develop normally to adulthood. Neither the reduction of protein O-GlcNAcylation induced by catalytic hypomorph alleles nor the sxc and sxc ID alleles cause severe locomotor defects in adult flies. Only the sxc allele negatively affects climbing performance. This effect appears to be independent of O-GlcNAc transferase activity. However, a contribution of a potential second site mutation affecting another gene that was not eliminated by six generations of backcrossing cannot be formally excluded.

Patient-related sxc mutant alleles do not affect global protein O-GlcNAcylation

We investigated whether the ID-associated alleles in sxc affect protein O-GlcNAcylation by subjecting lysates from adult heads of sxc, sxc, and sxc flies to Western blotting. Labeling with anti-O-GlcNAc antibody (RL2) that is able to capture O-GlcNAcylation changes in the catalytic hypomorphs ( and [40]) indicated that the levels of protein O-GlcNAcylation are not significantly altered in each of the three mutants (). This is in line with normal levels of O-GlcNAcylation in patient-derived fibroblasts and human embryonic stem cell models of the R313P and L283F equivalent mutations [26,28]. O-GlcNAcylation of the human A348T equivalent has not been investigated. The cell models of OGT-CDG mutations show downregulation of Oga, which may compensate for decreased O-GlcNAcylation. Although existence of such regulatory mechanism has not been shown in Drosophila, we analyzed the O-GlcNAcylation levels in Oga knockout background. Blocking O-GlcNAc hydrolysis in patient-related sxc mutant alleles with Oga increased O-GlcNAc levels to the same degree as in Oga samples (). We thus conclude that flies carrying ID-associated sxc mutations do not have a grossly affected protein O-GlcNAcylation.

sxc and sxc display defective habituation learning

Because of their role in ID in humans, we also investigated the effect of the novel sxc, sxc, and sxc alleles on habituation learning. We first subjected the sxc flies to 100 light-off pulses in the habituation assay. Despite sufficient locomotor abilities to perform in the island test and negative geotaxis assays, the initial jump response of the sxc homozygous and heterozygous flies was below the required threshold of 50% therefore deemed non-performers (sxc: 36% initial jumpers, N = 96; sxc: 47% initial jumpers, N = 96). Insufficient performance at the beginning of the assay thus precluded the assessment of habituation in these flies. We observed the same phenotype also for the sxc homozygous flies (49% initial jumpers, N = 64). The initial response in sxc heterozygous flies was sufficient (67%) and they were not able to suppress their jump response to the repeated light-off stimuli as efficiently as the genetic background control flies (), revealing a learning deficit. Flies heterozygous for the sxc allele showed a good initial jump response and habituated similar to the control, while sxc homozygous flies showed a habituation deficit (). In summary, deficits in habituation learning were observed for the R313P (heterozygous) and A348T (homozygous) mutations, while evaluation of the L254F homo- and heterozygous as well as R313P homozygous conditions was precluded by a poor initial jump response.

Assessment of sxc and sxc flies in habituation learning.

Jump responses were induced by 100 light-off pulses with 1 s interval between pulses. The jump response represents the % of jumping flies in each light-off trial. The mean number of trials that flies needed to reach the no-jump criterion (Trials To Criterion, TTC) ± SEM is also shown. (A) Deficient habituation of sxc flies (N = 73, padj = 6.18x10-6, in red) compared to their respective genetic background controls (control, N = 65, in blue). (B) Deficient habituation of sxc flies (N = 76, padj = 2.1x10-14, in brown) and no significant habituation deficit of sxc flies (N = 72, padj = 0.095, in red) compared to the genetic background control (control, N = 65, in blue). (C) Deficient habituation of sxc flies (N = 81, padj = 2.63x10-6, in red) is restored in sxc; Oga flies (N = 53, padj = 4.89x10-5, in cyan). (D) Deficient habituation of sxc flies (N = 79, padj = 8.84x10-10, in brown) is restored in sxc; Oga flies (N = 62, padj = 1.07x10-4, in cyan). (E) Deficient habituation of sxc flies (N = 81, padj = 2.63x10-6, in red) is restored in sxc; Oga flies (N = 64, padj = 9.09x10-6, in cyan). (F) Deficient habituation of sxc flies (N = 79, padj = 8.84x10-10, in brown) is restored in sxc; Oga flies (N = 56, padj = 3.74x10-7, in cyan). *** padj<0.001, n.s. not significant, based on lm analysis with Bonferroni-Holm correction for multiple comparisons. A complete list of p-values and summary statistics is provided in .

Blocking O-GlcNAc hydrolysis corrects habituation deficits of sxc and sxc

The apparently unaltered levels of O-GlcNAcylation in ID-associated sxc mutants () may suggest that O-GlcNAc-independent mechanisms underlie their cognitive phenotypes. To test this experimentally, we performed the habituation assay in flies carrying sxc allele and either the Oga allele or an Oga mutation that specifically blocks its O-GlcNAc hydrolase activity (Oga). Notably, heterozygous Oga and Oga flies habituated to a similar degree as the genetic background control flies. When introduced into an sxc background, Oga and Oga alleles fully rescued defects seen in the sxc mutants alone (). Similarly, we attempted a rescue of habituation deficient homozygous sxc with homozygous Oga and Oga alleles. We have previously shown that these homozygous Oga mutants also exhibit habituation deficits [53]. Strikingly, blocking O-GlcNAc hydrolysis by Oga or Oga in sxc flies was sufficient to completely rescue habituation deficits of either single mutant condition (). All tested flies show a good initial jump response and lower TTCs in the rescue experiments were not caused by fatigue (). In summary, despite seemingly normal gross O-GlcNAc levels in the sxc and sxc ID alleles, these genetic experiments provide evidence that their deficits in habituation learning depend on defective OGT enzymatic activities.

sxc and sxc and sxc show deficits in synaptic morphology

To determine whether synaptic phenotypes seen in larvae with hypomorphic catalytic domain mutations are recapitulated in larvae with patient mutations we assessed synaptic morphology at the NMJs of homozygous sxc and sxc and sxc larvae. We found that the NMJs of larvae carrying any of the three patient-related sxc mutant alleles display a significant increase in synaptic bouton number and sxc larvae also display significantly increased length and perimeter of the NMJ. NMJ length is also increased in the sxc and sxc larvae albeit not significantly (). In addition, sxc and sxc larvae show an increased number of NMJ branches (). Overall, the NMJ morphology phenotypes of the patient-related sxc mutant alleles resemble those of the catalytic hypomorphs sxc () and sxc (). This is in line with the observation that the patient-related sxc and sxc alleles affect O-GlcNAc transferase activity, which is indispensable for habituation (). The shared NMJ phenotype signature between the catalytic and patient-related mutants is the increase of synaptic bouton number. Because R313P and L283F mutations also significantly affect other NMJ parameters, we conclude that they are stronger/more detrimental than A348T mutation. This is in line with the observed effect on the jump performance in the light-off jump habituation assay (sxc, sxc and sxc non-performers).

Synaptic morphology of sxc, sxc and sxc.

Data presented as individual data points with mean ± SD. (A) NMJs on muscle 4 of sxc, sxc and sxc larvae have a significantly higher number of synaptic boutons (sxc: N = 21, p = 6.8x10-5, in red; sxc: N = 21, p = 0.011, in blue; sxc: N = 20, p = 0.0225) compared to the control (+/+, N = 42, in grey). sxc larvae have also significantly higher NMJ length (p = 0.0125) and perimeter (p = 0.001).). * padj <0.05, ** padj <0.01, *** padj <0.001, based on one-way ANOVA with Tukey’s multiple comparisons test. A complete list of p-values and summary statistics is provided in . (A’) Representative NMJs of wandering third instar larvae labeled with anti-discs large 1 (Dlg, magenta) and anti-synaptotagmin (Syt, green). When appropriate, type 1b synapses are distinguished from other synapses with white arrow. Scale bar, 20μm. The quantitative parameter values of the representative images (+/+ | sxc | sxc | sxc): #Boutons (20 | 36 | 29 | 37), Length (84.5 | 113.8 | 107.4 | 128), Area (399 | 447.9 | 360.9 | 349.9), Perimeter (254.2 | 329.5 | 313.8 | 302.4).

The Drosophila O-GlcNAc proteome is enriched in genes with function in neuronal development, learning & memory, and human ID gene orthologs

ID-associated mutations in sxc do not globally reduce protein O-GlcNAcylation yet blocking O-GlcNAc hydrolysis can completely restore the learning deficits in light-off jump habituation. We hypothesized that altered O-GlcNAcylation of specific sxc substrates is responsible for the habituation deficits. We therefore attempted to predict candidate substrates by exploration of the Drosophila O-GlcNAc proteome, as previously determined through enrichment with a catalytically inactive bacterial O-GlcNAcase [60]. We first performed an enrichment analysis of neuronal and cognitive phenotypes (as annotated in Flybase, see Materials and Methods) among the Drosophila O-GlcNAc substrates (encoded by in total 2293 genes) and found that they are significantly enriched in phenotype categories learning defective (Enrichment = 1.5, padj = 0.032), memory defective (Enrichment = 1.5, padj = 0.016), neurophysiology defective (Enrichment, 1.5, padj = 3.2x10-5) and neuroanatomy defective (Enrichment = 1.9, p = 5.27x10-35) (genes listed in ), supporting the importance of O-GlcNAcylation for neuronal development and cognitive function. We also found that human orthologs of 269 genes from the O-GlcNAc proteome are proven or candidate monogenic causes of ID (Enrichment = 1.7, padj = 1.01x10-16). When restricting this analysis to proteins with high-confidence mapped O-GlcNAc sites (in total 43) [60], we found orthologs of nine O-GlcNAcylated proteins to be implicated in Intellectual Disability, again representing a significant enrichment (Enrichment = 3.4, padj = 0.01). These orthologs are: Atpalpha (human ATP1A2), Gug (human ATN1), Hcf (human HCF1), LanA (human LAMA2), mop (human PTPN23), NAChRalpha6 (human CHRNA7), Ndg (human NID1), Nup62 (human NUP62) and Sas-4 (human CENPJ). They represent potential downstream effectors of sxc that may control habituation learning in the wild-type condition and may contribute to habituation deficits in catalytic and ID-associated sxc mutant conditions. Further analysis will be required to answer the question whether impaired O-GlcNAc transferase activity towards one or more of these targets is responsible for habituation deficits that are associated with OGT-CDG mutations.

Discussion

O-GlcNAcylation is important for habituation and for neuronal development in Drosophila

Habituation, the brain’s response to repetition, is a core element of higher cognitive functions [44-46]. Filtering out irrelevant familiar stimuli as a result of habituation allows to focus the cognitive resources on relevant sensory input. Abnormal habituation was observed in a number of neurodevelopmental disorders, including ID and Autism [47] and characterizes > 100 Drosophila models of ID [48,49]. To address the role of OGT and its O-GlcNAc transferase activity in this cognition-relevant process, we investigated heterozygous sxc flies [51] in light-off jump habituation. We found that they were not able to suppress their escape behavior as a result of deficient habituation learning (). This finding is in line with the recently published habituation deficit of the complete knock-out of OGT ortholog in C. elegans [61] and shows that altering O-GlcNAc transferase activity is sufficient to induce this deficit. We thus demonstrate the importance of O-GlcNAcylation in habituation learning.

Proper development and maintenance of synapses is an important aspect of neuronal function and cognition. The synaptic connection between motor neurons and muscle cells, termed the neuromuscular junction (NMJ), represents an excellent model system to study the molecular mechanisms of synaptic development in Drosophila [62]. Because NMJ defects were found in several Drosophila disease models with defective habituation [48,54-56], we investigated the synaptic architecture of the sxc larvae. We found that NMJs of the sxc larvae are characterized by an increased number of synaptic boutons, recognizable structures that contain the synaptic vesicles (). Larvae with a stronger homozygous catalytic mutation, sxc, also show an increase in NMJ length, area, and perimeter. We conclude that O-GlcNAcylation is important for control of synaptic size and synaptic bouton number.

Appropriate O-GlcNAc cycling is required for habituation learning and maintenance of the synaptic size

We recently showed that increased protein O-GlcNAcylation in homozygous Oga knockout flies causes a habituation deficit [53]. Here we show that heterozygous Oga knockout can restore the habituation deficit of sxc flies (). This indicates that habituation learning depends on O-GlcNAc cycling. Because the loss of one Oga allele does not significantly affect total O-GlcNAc levels [53], we presume that subtle changes in O-GlcNAcylation dynamics rather than gross loss of O-GlcNAc transferase activity inhibits habituation learning.

It is known that postsynaptic expression of OGT in excitatory synapses is important for synapse maturity in mammals [24]. Here we show that presynaptic O-GlcNAc transferase also has role in synapse growth. At the NMJ, the synapses of larvae with neuronal overexpression of sxc are shorter, and the number of synaptic boutons is decreased (). Both length and bouton number are normalized when sxc is overexpressed in neurons of the sxc larvae (). This phenotype was not observed in Oga knockout larvae with increased O-GlcNAcylation. Knockout of Oga can correct the increased bouton number in larvae with sxc mutation but not the NMJ size ().

Our data suggest that the NMJ defects associated with decreased O-GlcNAc transferase function are of neuronal origin and that O-GlcNAcylation controls the number of synaptic boutons and partially also synaptic size. Absence of synaptic size defects in Oga knockout larvae and failure of Oga to rescue the NMJ size defects caused by decreased O-GlcNAcylation indicates that other, non-catalytic O-GlcNAc transferase functions may be involved in the control of synaptic size. Levine et al. recently demonstrated that non-catalytic activities of OGT are necessary for its function in some cellular processes, such as proliferation [38].

Drosophila NMJ as a model for the O-GlcNAc-related synaptopathy

The sxc catalytic hypomorph mutations (sxcH537A, sxcH596F) as well as the OGT-CDG-patient equivalent mutations (R284P, A319T, L254F) that we introduced with the CRISPR/Cas-9 gene-editing technology in the Drosophila sxc gene (sxcR313P, sxcA348T, sxc), lead to an increase in the number of synaptic boutons, and in some cases also to an increase in synaptic size. NMJ size and the number of synaptic boutons in our model is determined by the level of sxc activity. Dependence of these parameters on gene activity/dosage was previously established in Fmr1 (the Drosophila model of Fragile X Syndrome) [59] and other Drosophila models of neurodevelopmental or neurological disorders, including Prosap/SHANK mutants (modelling Phelan-McDermid Syndrome caused by mutations in SHANK3, characterized by ID and ASD), Neuroligin 4 (ID and ASD caused by mutations in NLGN4), VAP33 (model of Amyotrophic Lateral Sclerosis caused by mutations in VAP-33A) and highwire (potential therapeutical target in traumatic brain injury) [63-66]. The synaptic phenotypes associated with impaired sxc catalytic activity may be linked to increased microtubule polymerization, since it has been shown that O-GlcNAcylation of tubulin negatively regulates microtubule polymerization and neurite outgrowth in mammalian cell lines [67] and Fmr1 and VAP-33A control synaptic growth and bouton expansion through presynaptic organization of microtubules [59,66].

Increased number of synaptic boutons has been also associated with increased excitability at the NMJ [68-70] although not consistently [70-72]. An interesting future direction could involve electrophysiological assessment of NMJ activity to determine whether O-GlcNAc cycling and the patient-related sxc mutations go beyond determining synapse development and affect synapse excitability and/or plasticity. However, these investigations would need to test various aspects of physiology and would still leave the impact of O-GlcNAc on cognition undetermined. For this reason, we assessed habituation as a highly cognition-relevant parameter.

Mutations implicated in OGT-CDG affect habituation via modulation of O-GlcNAc transferase activity

We assessed the effect of OGT-CDG missense mutations on habituation. We found that sxcR313P and sxcA348T inhibit habituation in the light-off jump habituation assay (). sxcL283F could not be investigated as these mutants displayed a non-performer phenotype in the light-off jump response. While the full spectrum of ID-related phenotypes in an individual with R284P mutation has been attributed to OGT, the A319T mutation segregates with an uncharacterized missense mutation in another gene implicated in ID, MED12 (G1974H) [25]. It was not known which of the mutations is responsible for ID in the affected individuals. We provide evidence that the Drosophila equivalent of the A319T mutation in the TPR domain of OGT causes a cognitive deficit and support a causal role of A319T in OGT-CDG.

Consistent with no detectable O-GlcNAc changes in patient samples and cellular models of the non-catalytic OGT mutations [26,28], no appreciable reduction in protein O-GlcNAcylation was observed in sxc and sxc flies. However, habituation learning was restored by increasing O-GlcNAcylation through blocking Oga activity (). This argues that the mechanism by which sxcR313P and sxcA348T inhibit habituation is defective O-GlcNAc transferase activity, paralleling impaired O-GlcNAc transferase activity and significant reduction of protein O-GlcNAcylation demonstrated in the catalytic OGT-CDG mutations [30,31]. It is worth noticing that we have previously shown that mutations in Oga also cause habituation deficits [53]. Our finding that genetic combination of loss of OGA with loss of OGT activity rescues the cognitive readout argues that OGA inhibition using available inhibitors may represent a viable treatment strategy. The R284P and A319T reside in the TPR domain (), which is responsible for recognition and binding of OGT substrates [38,73,74]. All OGT-CDG mutations investigated in this study were shown to impair the substrate interaction properties and the glycosyltransferase kinetics [27,36]. The observed habituation deficits may thus be caused by impaired O-GlcNAcylation dynamics towards a specific set of substrates that cannot be captured by standard O-GlcNAc detection assays. Identification of these substrates may pinpoint the underlying defective mechanisms and additional treatment targets.

Potential downstream effectors of O-GlcNAcylation in cognition and cognition-relevant processes

Our explorative analysis found that of 43 established O-GlcNAcylated proteins [60], nine are orthologs of human proteins implicated in ID: ATP1A2, ATN1, HCF1, LAMA2, PTPN23, CHRNA7, NID1, NUP62 and CENPJ. These proteins represent potential downstream effectors and can be investigated in future studies. Particularly the transcriptional co-regulator HCF1 (Host Cell Factor 1) emerges as a top candidate. In mammals, OGT mediates glycosylation and subsequent cleavage of HCF1, which is essential for its maturation [75]. Recombinant OGT with an R284P amino acid substitution is defective in HCF1 glycosylation [28] and HCF1 processing was shown to be completely abrogated by a catalytic OGT-CDG mutation [30].

Drosophila sxc is a member of the polycomb group (PcG), a conserved set of chromatin and transcriptional modifiers that initially have been identified by phenotypic similarity of their mutant phenotypes: homeotic transformations. They are required for maintenance of transcriptional repression (of non-lineage genes) during embryonic development and cell proliferation [76,77]. Missing O-GlcNAcylation of PcG component Polyhomeotic (Ph) is responsible for misexpression of HOX genes and homeotic transformations in sxc null mutants [78]. A recent study has shown that chromatin redistribution induced by interaction between sxc and PcG member Polycomb like (Pcl) controls plasticity of sensory taste neurons [79]. It is not known whether the regulation of PcG activity by sxc/OGT is important for cognitive function, but it can be noted that a series of PcG genes are associated with ID [80,81], and some of them are subject to regulation by OGT in the context of development or cancer. These include PHC1 –human ortholog of Drosophila Ph [82], RING1B (Drosophila Sce) [83,84], EZH2 (Drosophila E(z)) [85-87], YY1 (Drosophila pho) [88,89] and ASXL1 (Drosophila Asx) [90,91]. In addition, OGT regulates expression of PcG genes by O-GlcNAcylation of PcG transcriptional regulators, for example ATN1 (Drosophila Gug), which was identified in the embryonic O-GlcNAc proteome [60]. The encoded proteins represent interesting candidate targets that may link cognitive deficits of OGT-CDG mutations to PcG function.

We propose that in depth clinical phenotyping of patients with mutations in OGT and the above listed genes may give additional hints to the most crucial downstream targets of OGT-mediated O-GlcNAcylation.

In summary, we show that OGT-CDG mutations in the TPR domain negatively affect habituation learning in Drosophila via reduced protein O-GlcNAcylation. The data support a causal role of A319T in OGT-CDG and demonstrate that Drosophila habituation can be used to analyze the contribution of OGT mutations to cognitive deficits. This important aspect of ID has to date not been addressed for any of the OGT-CDG mutations. Moreover, our genetic approach points to a key role of O-GlcNAc transferase activity in ID-associated cognitive deficits and identifies blocking O-GlcNAc hydrolysis as a treatment strategy that can ameliorate cognitive deficits in OGT-CDG patients. Thanks to its high-throughput compatibility, the light-off jump habituation assay can be used with high efficiency for future identification of the downstream effectors and novel therapeutic targets for OGT-CDG.

Materials and methods

Cloning of the guide RNA and repair template DNA vectors for Drosophila CRISPR/Cas9 editing

Novel mutant Drosophila lines, sxc, sxc, sxc, and sxc, were generated via CRISPR/Cas9 gene editing, following a previously described protocol [51]. Briefly, guide RNA sites were selected using an online tool (crispr.mit.edu) and the annealing primer pairs with appropriate overhangs for BpiI restriction digestion were cloned into pCFD3-dU63gRNA plasmid [92]. Vectors coding for repair template DNA of roughly 2 kb were generated from Drosophila Schneider 2 cell genomic DNA by PCR using GoTaq G2 Polymerase (Promega) and primer pairs appropriate for the desired region (). The PCR products were digested with BpiI and inserted into the pGEX6P1 plasmid. The intended mutation, as well as silent mutations required to remove the gRNA sequence (), were incorporated by either site-directed mutagenesis (H596F) using the QuikChange kit (Stratagene) or restriction-free cloning (R313P, A348T and L283F) [93]. The four sets of mutations–H596F, L283F, R313P, and A348T removed restriction sites for HinfI, BfmI, MnlI and BsqI, respectively. DNA products of cloning and mutagenesis were confirmed by sequencing. All primer sequences are listed in .

Generation of sxc, sxc, sxc, and sxc Drosophila lines

Vasa::Cas9 Drosophila embryos (strain #51323 from Bloomington Drosophila Stock Center; bdsc.indiana.edu) were injected with a mixture of CRISPR/Cas9 reagents, 100 ng/μl guide RNA plasmid and 300 ng/μl repair template DNA vector (University of Cambridge fly facility). Injected male flies were crossed with an in-house Sp/CyO balancer stock for two generations, allowing for the elimination of the vasa::Cas9 carrying X chromosome. Candidate F1 males were genotyped exploiting restriction fragment length polymorphism. All lines were validated by sequencing the region approximately 250 base pairs upstream and downstream of the mutations and sequencing the areas outside the repair templates. In addition, all of the predicted off-target sites were PCR-amplified and checked for the presence of any lesions compared with the genomic DNA from the BL51323 line. None of the predicted off-target sites were found to have mutations. To eliminate any other potential off-target mutations introduced during CRISPR, all lines were backcrossed into the w control genetic background for six generations.

Restriction fragment length polymorphism assay

To assess and confirm the presence of the H596F, L283F, R313P, and A348T mutations in the sxc gene, DNA of candidate individual adult flies was extracted using 10–50 μl of DNA extraction buffer containing 10 mM Tris-HCl pH 8, 1 mM EDTA, 25 mM NaCl and 200 μg/ml freshly added Proteinase K (Roche). The solution was subsequently incubated at 37°C for 30 min, followed by inactivation of Proteinase K at 95°C for 3 min, and centrifuged briefly. 1 μl of the crude DNA extract was used per 25 μl PCR reaction with the relevant diagnostic primers, using a 2x GoTaq G2 Green premix (Promega). 5 μl of the PCR products were used for restriction fragment length polymorphism assay with the appropriate enzymes, followed by agarose gel electrophoresis of the digested products. Reactions which showed the presence of an undigested full-length PCR product resistant to the expected restriction enzyme cleavage indicated CRISPR/Cas9 gene editing event and were sequenced. Precise incorporation of the repair template into the right position of the genome was confirmed by sequencing a second round of PCR products obtained from potential homozygous CRISPR mutants with mixed diagnostic and line-check primer pairs. Primer sequences are listed in .

Fly stocks and maintenance

Drosophila stocks and experimental crosses were reared on a standard Drosophila diet (sugar/cornmeal/yeast). An RNAi strain to knockdown sxc (#18610) and a genetic background control strain (#60000) were obtained from the Vienna Drosophila Resource Center (VDRC; www.vdrc.at). In-house sxc [51] and UAS-sxc [40] strains, and the generated sxc, sxc, sxc, and sxc strains, were crossed into the VDRC w control genetic background (#60000) for six generations. The sxc/CyO; UAS-sxc strain was assembled using the isogenic strains. Oga and Oga lines were also crossed to this background as described earlier [53]. #60000 was used as isogenic control for the mutant alleles. In the neuromuscular junction analysis of sxcR313P, sxcA348T and sxcL283F alleles, the control flies were derived by crossing the flies from the stock used for microinjection (Bloomington Stock: BL51323) and same crossing scheme as that used to derive the sxcR313P, sxcA348T and sxcL283F homozygotes and eliminate the Cas9 transgene were used. To induce neuronal knockdown and overexpression, a w; 2xGMR-wIR; elav-Gal4, UAS-Dicer-2 driver strain was used. This strain contains a double insertion of an RNAi construct targeting the gene white specifically in the Drosophila eye (2xGMR-wIR) to suppress pigmentation, as required for an efficient light-off jump response [54,55]. Progeny of the crosses between the driver, RNAi/UAS-sxc and #60000 strain were used as controls for knockdown and overexpression experiments. All crosses were raised at 25°C, 70% humidity, and a 12:12h light-dark cycle.

Western blotting

Protein lysates for Western blotting were prepared from adult male (1–4 days old) fly head samples or 0–16 h embryo collection and snap frozen in liquid nitrogen. Samples were homogenized in lysis buffer containing 2x NuPAGE LDS Sample Buffer, 50 mM Tris- HCl (pH 8.0), 150 mM NaCl, 4 mM sodium pyrophosphate, 1 mM EDTA, 1 mM benzamidine, 0.2 mM PMSF, 5 μM leupeptin, and 1% 2-mercaptoethanol. Crude lysates were then incubated for 5 min at 95°C, centrifuged at 13000 rpm for 10 min, and supernatants were collected. Pierce 660 nm protein assay supplemented with Ionic Detergent Compatibility Reagent (Thermo Scientific) was used to determine protein concentration. 20–30 μg of protein samples were separated on RunBlue 4–12% gradient gels (Expedeon) using MOPS running buffer, before being transferred onto nitrocellulose membranes. Western blot analysis was carried out with anti-O-GlcNAc (RL2, Abcam, 1:1000) and anti-actin (Sigma, 1:5000) antibodies. Membranes were incubated overnight with selected primary antibodies in 5% BSA at 4°C. Blots were visualized via Li-Cor infrared imaging with Li-Cor secondary antibodies (1:10000) Signal intensities were quantified using ImageStudioLite software. Significance was calculated using one-way ANOVA with Tukey’s multiple comparisons test (padj).

Developmental survival

Stage 11–16 embryos (25 embryos per vial, 100 per genotype per experiment, n = 3) were cultured at 25°C and assessed for lethality by counting the number of pupae and adults derived. Significance was calculated using Student’s t-test with Holm-Sidak’s correction for multiple comparisons when appropriate.

Light-off jump habituation

The light-off jump reflex habituation assay was performed as previously described [49,94]. Briefly, 3- to 7-day-old individual male flies were subjected to the light-off jump reflex habituation paradigm in two independent 16-chamber light-off jump habituation systems. Male progeny of the appropriate control genetic background was tested simultaneously on all experimental days. Flies were transferred to the testing chambers without anesthesia. After 5 min adaptation, a total of 32 flies (16 flies/system) were simultaneously exposed to a series of 100 short (15 ms) light-off pulses with 1 s interval. The noise amplitude of wing vibration following every jump response was recorded for 500 ms after the start of each light-off pulse. A carefully chosen automatic threshold was applied to filter out background noise and distinguish it from jump responses. Data were collected by a custom-made Labview Software (National Instruments). Initial jump responses to light-off pulse decreased with the increasing number of trials and flies were considered habituated when they failed to jump in five consecutive trials (no-jump criterion). Habituation was quantified as the number of trials required to reach the no-jump criterion (Trials To Criterion (TTC)). All experiments were done in triplicates (N = 96 flies). Main effects of genotype on log-transformed TTC values were tested using a linear model regression analysis (lm) in the R statistical software (R version 3.0.0 (2013-04-03)) [95] and corrected for the effects of testing day and system. Bonferroni-Holm correction for multiple testing [96] was used to calculate adjusted p-values (padj).

Fatigue assay

Each genotype that was tested in light-off jump habituation was subsequently subjected to fatigue assay. The fatigue assay was used to evaluate whether the lower TTCs in the rescue experiments were not a result of increased fatigue rather than improved habituation/non-associative learning. The assay was performed as previously described [49]. The interval between light-off pulses was increased to 5 seconds, an intertrial interval that is sufficiently long to prevent habituation. The light-off pulse was repeated 50 times. Fatigue was concluded when log-transformed TTC values of the rescue were significantly smaller than log-transformed TTC values of the control (based on lm analysis and Bonferroni-Holm correction; padj < 0.05).

Analysis of Drosophila neuromuscular junction

Wandering male L3 larvae were dissected with an open book preparation [97], and fixed in 3.7% paraformaldehyde for 30 minutes. Larvae were stained overnight at 4°C with the primary antibodies against synaptic markers Discs large (anti-dlg1, mouse, 1:25, Developmental Studies Hybridoma Bank) and synaptotagmin (anti-Syt, rabbit, 1:2000, kindly provided by H. Bellen). Secondary antibodies anti-mouse Alexa 488 and anti-rabbit Alexa 568 (Invitrogen) were applied for 2 hours at room temperature (1:500). Projections of type 1b neuromuscular junctions (NMJs) at muscle 4 from abdominal segments A2-A4 were assessed. Individual synapses were imaged with a Zeiss Axio Imager Z2 microscope with Apotome and quantified using in-house developed Fiji-compatible macros [98,99]. Anti-dlg1 (4F3 anti-discs large, DSHB, 1:25) labeling was used to analyze NMJ area, length, number of branches and branching points. Anti-Syt (kind gift of Hugo Bellen, 1:2000) labeling was used to analyze the number of synaptic boutons. Secondary antibodies goat anti-mouse Alexa Fluor 488 (1:200) and goat anti-rabbit Alexa Fluor 568 from Life Technologies were used for visualization. Parameters with a normal distribution (area, length, number of boutons) were compared between the mutants and controls with one-way ANOVA (p) and Tukey’s test for multiple comparisons (padj). Parameters without normal distribution (number of branches and branching points) were compared with non-parametric Wilcoxon test (p, single comparisons) and Kruskal-Wallis test with Wilcoxon pairwise test for multiple comparisons (padj) in the R statistical software (R version 3.0.0 (2013-04-03)) [95].

Island assay

Locomotor behaviour of 3–6 days old male flies was assessed with the island assay as described previously [100,101]. Each trial was performed using 15 flies. 3–4 repeats were carried out on each test day, and data was collected on 3 consecutive days. In total, data from 11–16 trials were collected per genotype. The percentage of flies on the island platform over time was plotted and area under curve (AUC) was determined for each run. Groups were compared using one-way ANOVA with Holm-Sidak’s multiple comparisons (padj) of means for AUC.

Negative geotaxis test

The negative geotaxis assay was performed as described previously [102]. The climbing ability of 3–6 days old male flies was evaluated on groups of 10 animals. Prior to the measurement, flies were transferred into 150 x 16 mm transparent plastic test tubes without anesthesia. Test tubes were secured into a frame that allowed for monitoring of climbing behavior of up to 10 vials at once. Upon release, the frame is dropped from a fixed height onto a mouse pad, thereby tapping the flies to the bottom of the tubes. The climbing assay was repeated 4 times for each loaded frame providing data from 4 runs. The experiment was video-recorded with a Nikon D3100 DSLR camera. ImageJ/FIJI software was used to analyse the resulting recordings. First, images were converted to an 8-bit grey scale TIFF image sequence (10 frames per second) file format. Background-subtraction and filtering were then applied, and the image pixel values were made binary. The MTrack3 plug-in was used for tracking of flies. Mean climbing speed (mm/s) was quantified for each genotype in 2nd, 3rd and 4th runs, between 17–89 data points were collected per run. Groups were compared using one-way ANOVA with Tukey’s multiple comparisons (padj) of means on mean climbing speed values calculated for each run.

Enrichment analysis

The O-GlcNAc proteome data was extracted from Selvan et al. (Supplementary dataset 3) [60]. Phenotype annotations of Drosophila gene alleles were extracted from Flybase (Flybase.org, downloaded in April 2016). Human genes implicated in Intellectual disability (ID + ID candidate genes) were extracted from sysid database (https://sysid.cmbi.umcn.nl/, downloaded in April 2016). Enrichment was calculated as follows: (a/b)/((c-a)/(d-b)), whereby a = genes in O-GlcNAc proteome and associated with the phenotype term/human ID gene orthologs, b = genes in O-GlcNAc proteome, c = genes associated with the phenotype term/human ID gene orthologs, d = background/all Drosophila genes. Significance was determined using two-sided Fisher’s exact test in R [95]. p-values were adjusted for multiple testing (padj) with Bonferroni-Holm correction [96].

Generation and characterization of sxc, sxc, sxc and sxc alleles.

(A) Schematic representation of Drosophila sxc protein showing the location of H596F, L283F, R313P and A348T mutations; purple tetratricopeptide repeat (TPR) domain, green glycosyl transferase (GT) domain. (B)–(E) Sequences of genomic DNA of wild type, sxcH596F, sxcL283F, sxcR313P and sxcA348T Drosophila alleles. The missense mutation and additional silent mutations are highlighted. The restriction digestion sites used for genotyping are shown.

(TIF)

Click here for additional data file.

Jump responses in the fatigue assay.

In the fatigue assay, jump responses were induced with 50 light-off pulses with 5 s interval between pulses that prevents habituation. The jump response is presented as % of jumping flies in each light-off trial. The mean number of trials that flies needed to reach the no-jump criterion (Trials To Criterion, TTC) ± SEM is presented. (A) Jump response of the sxc; Oga flies (N = 85, mean TTC ± SD: 27.4 ± 8, in cyan) remains high throughout the entire course of the experiment, similar to control flies (+/+, N = 85, mean TTC ± SD: 34.5 ± 5, padj = 0.14, in blue) demonstrating that restored habituation in sxc; Oga flies () is not confounded by fatigue. (B) Jump response of the sxc; elav-Gal4>UAS-sxc flies (N = 52, mean TTC ± SD: 25.1 ± 7.9, in green) remains high throughout the entire course of the experiment, similar to control flies (+/+, N = 55, mean TTC ± SD: 28.3 ± 2.3, padj = 0.128, in blue) demonstrating that restored habituation in sxc; elav-Gal4>UAS-sxc flies () is not confounded by fatigue. (C) Jump response of the sxc; Oga flies (N = 73, mean TTC ± SD: 28.9 ± 7.6, in cyan) remains high throughout the entire course of the experiment, similar to control flies (+/+, N = 84, mean TTC ± SD: 26.1 ± 5.1, padj = 1, in blue) demonstrating that restored habituation in sxc; Oga flies () is not confounded by fatigue. (D) Jump response of the sxc; Oga flies (N = 78, mean TTC ± SD: 26.9 ± 5.1, in cyan) remains high throughout the entire course of the experiment, similar to control flies (+/+, N = 85, mean TTC ± SD: 34.5 ± 5, padj = 0.31, in blue) demonstrating that restored habituation in sxc; Oga flies () is not confounded by fatigue. (E) Jump response of the sxc; Oga flies (N = 83, mean TTC ± SD: 25.5 ± 9.2, in cyan) remains high throughout the entire course of the experiment, similar to control flies (+/+, N = 84, mean TTC ± SD: 26.1 ± 5.1, padj = 1, in blue) demonstrating that restored habituation in sxc; Oga flies () is not confounded by fatigue. (F) Jump response of the sxc; Oga flies (N = 63, mean TTC ± SD: 24.4 ± 5.1, in cyan) remains high throughout the entire course of the experiment, similar to control flies (+/+, N = 85, mean TTC ± SD: 34.5 ± 5, padj = 0.15, in blue) demonstrating that restored habituation in sxc; Oga flies () is not confounded by fatigue. * padj<0.1, based on lm analysis with Bonferroni-Holm correction for multiple comparisons. Complete list of p-values and summary statistics is provided in .

(TIF)

Click here for additional data file.

NMJ of sxc mutant and NMJ branching morphology.

(A) Number of synaptic branches and branching points in sxc larvae (N = 29, in red) is not significantly different from the genetic background control larvae (+/+, N = 24, branches: p = 0.1439, branching points: p = 0.05648, in blue). P-values are based non-parametric Wilcoxon test analysis. (B) Number of branches and branching points is not affected in elav-Gal4>UAS-sxc larvae (N = 29, in dark blue) compared the elav-Gal4/+ larvae (N = 30, branches: padj = 0.8702, branching points: padj = 0.8488, in light blue) and to UAS-sxc/+ larvae (N = 26, branches: padj = 0.8294, branching points: padj = 0.2689, in grey). (C) Branches and branching points are not affected in sxc; UAS-sxc/+ larvae (N = 28, in red) compared to the control larvae (+/+, N = 28, branches: padj = 0.7121, branching points: padj = 0.2979, in blue). sxc; elav-Gal4>UAS-sxc larvae (N = 29, in green) do not show any changes in number of branches and branching points compared to the sxc; UAS-sxc/+ larvae (branches: padj = 0.4097, branching points: padj = 0.5928) and control larvae (+/+; branches: padj = 0.4301, branching points: padj = 0.6927). (D) sxc larvae have significantly increased NMJ length (N = 25, p =) and perimeter (N = 21, p =, in red) compared to their genetic background control (+/+, N = 26, in blue) but not significantly different number of boutons (p = 0.085), NMJ area (p = 0.618), number of branches (p = 0.691) and branching points (p = 0.371). * p<0.05, *** p<0.001. P-values for boutons, length, area and perimeter are based on one-way ANOVA. P-values for branches and branching points are based on non-parametric Wilcoxon test analysis. (E) Branches and branching points are not affected in sxc larvae (N = 31, branches: padj = 1, branching points: padj = 0.5, in brown), Oga larvae (N = 30, branches: padj = 0.75, branching points: padj = 0.51, in purple), and sxc; Oga larvae (N = 30, branches: padj = 0.75, branching points: padj = 0.5, in cyan) compared to the genetic background control larvae (+/+, N = 28, in blue). sxc; Oga larvae do not show a significant change in number of branches and branching points compared to the sxc larvae (branches: padj = 0.75, branching points: padj = 1). Data presented as individual data points with mean ± SD. P-values are based on Kruskal-Wallis test with Wilcoxon pairwise test for multiple comparisons. Complete list of p-values and summary statistics is provided in . (D’) Representative NMJs of genetic background control (+/+) and sxc wandering third instar larvae labeled with anti-discs large 1 (Dlg, magenta) and anti-synaptotagmin (Syt, green). Scale bar, 20μm. The quantitative parameter values of the representative images (+/+ | sxc): #Boutons (26 | 30), Length (96.2 | 169.8), Area (415.9 | 464.3), Perimeter (258.4 | 445).

(TIF)

Click here for additional data file.

Western Blot and developmental survival of sxc flies.

(A) Embryos from either wildtype, sxc, sxc homozygotes were assessed for levels of global O-GlcNAc using a pan-O-GlcNAc antibody RL2. The blot was normalized to actin. This blot is a representative of three experiments. (B) Reduced total O-GlcNAc levels in sxc and sxc homozygotes are not associated with developmental lethality. Data presented as percentage of pupae and adults derived from stage 11–16 embryos (100 per genotype per experiment, n = 3). Based on Student’s t-test with Holm-Sidak’s correction for multiple testing. Complete list of p-values and summary statistics is provided in

(TIF)

Click here for additional data file.

Developmental and locomotor characterization and NMJ branching morphology of sxc, sxc and sxc.

(A) Control, sxc, sxc and sxc embryos (stage 11–16, 80–100 per experiment) were transferred to fresh food at 25°C, and the numbers of pupae formed and adults eclosed were counted. Development from embryo to pupae or from pupae to adulthood was not significantly affected in sxc (pupae: N = 4 repeats, p = 0.9, adults: N = 3, p = 0.29, in red), sxc (pupae: N = 3, p = 0.152, adults: N = 3, p = 0.345, in blue) and sxc mutants (pupae: N = 4, p = 0.108, adults: N = 3, p = 0.727, in green). Data presented as individual data points with mean ± SD. P-values are based on Student’s t-test. (B) Flight escape performance was assessed in the island assay. 15 flies per measurement were thrown on a white platform surrounded with water. Data was collected over 3 days of measurement (Control: N = 23, sxc: N = 16, sxc: N = 15 and sxc: N = 14 repeats). Floating bars depict mean ± SD area under curve (AUC), a parameter that is derived from data plotted as % flies on the platform over time. One-way ANOVA with Tukey’s multiple comparisons was used to compare the mean AUC between genotypes. Flight escape performance of sxc, sxc and sxc flies revealed no defects in locomotion or fitness. (C) Number of synaptic branches is increased in sxc (N = 21, p = 0.049, in red) and sxc larvae (N = 20, in green) compared to the genetic background control (+/+, N = 41, in grey). Number of branching points is not significantly different. Data presented as individual data points with mean ± SD. * p< 0.05. P-values are based on Kruskal-Wallis test with Wilcoxon pairwise test for multiple comparisons. Complete list of p-values and summary statistics is provided in

(TIF)

Click here for additional data file.

Primer sequences.

(XLSX)

Click here for additional data file.

Light-off jump habituation parameters summary.

(XLSX)

Click here for additional data file.

Summary statistics.

(XLSX)

Click here for additional data file.

Enrichment in Drosophila phenotypes and orthologs of human genes implicated in intellectual disability.

(XLSX)

Click here for additional data file.

2 Sep 2021

Dear Dr van Aalten,

Thank you very much for submitting your Research Article entitled 'Intellectual disability-associated disruption of O-GlcNAcylation impairs neuronal development and cognitive function in Drosophila' to PLOS Genetics.

The manuscript was fully evaluated at the editorial level and by four independent peer reviewers who are experts in the field. As you can see from the reports below, the reviewers acknowledged the importance of this work and appreciated the authors' efforts in generating mutant fly lines for OGT and designing experiments for testing neuronal and cognitive effects. However, the reviewers also raised substantial concerns, including comments related to analysis of null allele, testing patient-specific mutants for synaptic assays, clarity and criteria for terminology used in the paper, and conclusions drawn from bioinformatic analysis towards neural processes and function. Based on the reviews, we will not be able to accept this version of the manuscript, but we would be willing to review a much-revised version. We cannot, of course, promise publication at that time.

Should you decide to revise the manuscript for further consideration here, your revisions should address the specific points made by each reviewer. We will also require a detailed list of your responses to the review comments and a description of the changes you have made in the manuscript.

If you decide to revise the manuscript for further consideration at PLOS Genetics, please aim to resubmit within the next 60 days, unless it will take extra time to address the concerns of the reviewers, in which case we would appreciate an expected resubmission date by email to plosgenetics@plos.org.

If present, accompanying reviewer attachments are included with this email; please notify the journal office if any appear to be missing. They will also be available for download from the link below. You can use this link to log into the system when you are ready to submit a revised version, having first consulted our Submission Checklist.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please be aware that our data availability policy requires that all numerical data underlying graphs or summary statistics are included with the submission, and you will need to provide this upon resubmission if not already present. In addition, we do not permit the inclusion of phrases such as "data not shown" or "unpublished results" in manuscripts. All points should be backed up by data provided with the submission.

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool.  PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org.

PLOS has incorporated Similarity Check, powered by iThenticate, into its journal-wide submission system in order to screen submitted content for originality before publication. Each PLOS journal undertakes screening on a proportion of submitted articles. You will be contacted if needed following the screening process.

To resubmit, use the link below and 'Revise Submission' in the 'Submissions Needing Revision' folder.

[LINK]

We are sorry that we cannot be more positive about your manuscript at this stage. Please do not hesitate to contact us if you have any concerns or questions.

Yours sincerely,

Santhosh Girirajan

Associate Editor

PLOS Genetics

Gregory P. Copenhaver

Editor-in-Chief

PLOS Genetics

Reviewer's Responses to Questions

Comments to the Authors:

Please note here if the review is uploaded as an attachment.

Reviewer #1: The manuscript by Fenckova et al explores cognitive mechanisms in the fly focusing on mutations identified in humans as causative of intellectual disability in the gene O-GlcNAc transferase (OGT). The habituation studies are carefully crafted with well controlled genetic tools. The authors offer important genetic evidence supporting a model where by disruption of the O-GlcNAc addition-removal cycle is causative of cognitive dysfunction in Drosophila models of human mutations. The most clear cut and conclusive evidence is emerging of the Drosophila habituation assays presented in Figures 1 and 4. In my view, these genetic-behavioral experiments provide a solid foundation for the assertion in the title that “Intellectual disability-associated disruption of O-GlcNAcylation impairs […] cognitive function in Drosophila”. However, the paper falls short to provide a comprehensive understanding of processes affected by the diverse mutations tested either individually or collectively. It is clear that global defects on O-GlcNAc-sylation do not account for the habituation phenotypes described in figure 4. Yet, the identity of neural processes and targets are insufficiently explored with a limited data-mining analysis of the Drosophila O-GlcNAc proteome. This last part of the paper is truncated and inconclusive.

Concerning figures 2 and 3, in my view they need some reconsideration. Figure 3 is a collection of important controls that can go into supplementary materials.

Figure 2, while I can see the graphs depicting the reported differences in boutons number and other NMJ parameters, I cannot see in the images the phenotypes reported by the graphs. Thus, I cannot concur with the authors that ”Intellectual disability-associated disruption of O-GlcNAcylation impairs neuronal development […]”. This assertion needs better NMJ evidence and/or new approach to document this conclusion.

Appendix has references at the end not related to the content of the paper.

In summary, only two figures provide uncontestable insight (Fig. 1 and 4) making this manuscript a solid yet rather discrete contribution.

Reviewer #2: Review of PGenetics-D-21-00961

Summary

Overall I find this is an interesting study using Drosophila as a model system to understand the function of O-GlyNAcylation on neuronal development and function. I further like the use of CRISPR to generate specific mutant alleles found in human clinical subjects and thus explore the mechanism of specific allelic variants in human disease.

I also think the analysis well planned and executed. My comments and criticisms are minor but I offer those for consideration of the author and editor.

Questions and concerns (not in any particular order)

1. I do think that labeling habituation responses as cognitive function will not be well received by most human geneticists. There has always been a tension between model organism researchers and human geneticists regarding when invertebrate models reflect human pathological mechanisms. I would advise not using this term, but simply stating that they have found affects on habituation and synapse development.

2. Have they done any CNS morphological assessments of their mutants?

3. I am a bit puzzled by their results with the sxc[H537A] allele. As heterozygotes this allele affects bouton number but not other parameters of synapse size. Yet, when the homozygote is analyzed it displays phenotypes for NMJ length and perimeters but not bouton number. Likewise I noticed that the effect of this allele as a heterozygote on bouton number is a bit modest, and show some variation within experiments shown (for example see the difference between Fig. 2 Panel A, and Figure 2 Panel C. The difference between sxc/+ and +/+ is pretty modest in Panel C, affected largely by a few outliers. I think some further discussion of these findings is warranted. In my mind it tells me the effect of synapse morphology is modest and makes me wonder if there are any observable electrophysiological changes.

4. I found the allele specific analysis very interesting and followed their stepwise arrival at the conclusion that while overall GlycNAcylation was not changed in some alleles, specific protein targets do likely see differences in modification since the phenotypes can be rescued by altering Oga function. However, I am not a big fan of making biological conclusions from multi-step bioinformatics analyses, and while their enrichment analysis seems valid, it is a far cry from doing an experiment. Their ability to look at the patterns of GlycNAcylation by western seems a logical place to begin to see if different alleles give different patterns (ie intensities) of modification. I would be intrigued if this sort of analysis was attempted and if there were any indications of differential modification of protein targets.

Reviewer #3: In this manuscript, Fenckova et al., studies the role of OGT (O-GlcNAc transferase) and disease associated variants in this gene using Drosophila melanogaster. First the authors studied a previously reported hypomorphic allele of sxc (OGT fly ortholog with H537A mutation) that has diminished enzymatic activity and showed that heterozygous mutants show defects in a light-induced habituation assay (and the homozygous are defective in jumping in response to light stimuli). This phenotype can be rescued by re-introducing sxc in neurons, or via removing one copy of Oga, a gene that encodes the O-GlcNAc hydrolase. This indicates that this gene is required in the neurons for proper habituation and also suggests that proper O-GlcNAc cycling (or dynamics) is required for this behavior. The authors also showed that hypomorphic sxc mutants (tested H537A and H596F) show altered synaptic morphology at the larval neuromuscular junction and that this phenotype can also be rescued by the same genetic manipulations. Interestingly, over-expression of sxc also showed synaptic defects, suggesting that tight control of protein O-GlcNAc modification is critical for proper synaptic morphology. Next, the authors generated knock-in flies that carry mutations that corresponds to one of three previously reported OGT alleles found in patients (R313P, A348T, L283F) and subjected them to biochemical, behavioral and histological assays. Through western blotting of O-GlcNACylated proteins, the authors found that these mutants do not seem to have a general/global defect in protein O-GlcNACylation. The authors showed that A348T and L283F mutants did not show a major motor defect but the R313P did show some motor coordination defect through a negative geotaxis assay. Because mutant alleles of sxc that show strong enzymatic defect do not show this defect, the authors interpreted this as a sign that the R313P variant causes some sort of an enzymatic activity independent function. Next, they tested the light-induced habituation in the three patient alleles and found that two of them show a significant phenotype which can be corrected by reducing Oga (L283F line were not good responders to light stimuli, despite their normal motor behaviors in other tests). Finally, the authors performed a bioinformatics analysis on the Drosophila O-GlcNAc proteome and found that proteins encoded by genes involved in nervous system function and homologs of human ID genes are enriched in this group. The authors list 9 high-confidence mapped O-GclNAc genes that are homologous to human ID genes and propose that these maybe potential targets that may be responsible for the phenotypes described in this study.

I feel the authors have put in significant effort to generate a number of fly knock-in strains to functionally characterize the role sxc/OGT and disease associated variants in a cognition related behavior and synaptic morphology. I feel this work will be of interest to the readership of PLoS Genetics, especially to fly researchers interested in neuronal function, glycobiologists who study physiological significance of certain sugar modifications, and clinicians and human geneticists who study disease associated variants. I would be happy to recommend this study for publication, if the following points are addressed.

Major Points

1) The author performs the habituation assay on a number of sxc mutants but do not show data for the null allele. I acknowledge the homozygous null flies are pupal lethal but the authors should be able to test this in a heterozygous fly. Since the null allele is an important reference point to interpret the function of missense alleles (e.g. if the mutant phenotype of certain missense allele is stronger than the null, they maybe antimorphic or neomorphic alleles, rather than a hypomorph), it would be valuable to show what the null allele looks like in this assay (which have been generated by the authors in PMID: 29588363. In addition, multiple alleles are available from stock centers). Also, note that the habituation data for the strong hypomorphic allele (H596F) they generated is also not provided here, which would make this paper more complete.

2) The authors explored the synaptic morphology phenotype of the known enzymatic defective mutants (H537A and H596F), but have not examined this for the patient derived variants (R313P, A348T, L283F). Was this because they looked at these phenotypes and did not see any relevant phenotype? Or was there a reason this experiment cannot be carried out? Since the authors are trying to make the point that these synaptic phenotype has some relevance to human ID, it would make sense to show how the NMJs of the flies with mutations that are analogous to those found in the patient mutations look like. Also, data from the heterozygous or homozygous null allele can also be provided here to strengthen this paper.

3) The authors attribute the negative geotaxis defect of the R313P allele to this variant affecting a non-enzymatic function of Sxc. While this is an interesting idea, alternatively this phenotype may be due to some sort of a 2nd site mutation that is unique to this strain (off targeting by CRISPR or some floating mutation that is closely linked that was not eliminated through back crossing). The authors should perform a rescue experiment (e.g. neuronal rescue experiment using UAS/GAL4 as they do for other alleles or genomic rescue) if they wish to claim that the negative geotaxsis defect in this allele is due to sxc and not other genes.

4) The section subtitle “Patient-related sxc mutant alleles have normal levels of O-GlcNAcylation” seems a bit misleading, especially since they later argue that phenotypes seen in some of the patient mutant alleles can be corrected by removing one copy of Oga. As they discuss in the text that there may be some minor O-GlcNAcylatiojn defect that is not obvious when performing a bulk O-GlcNAcylation assay by western blot (e.g. a few substrates that are more sensitive to slight reduction in O-GlcNAcylation is responsible for the habituation and synaptic phenotypes). The authors may want to consider an alternative title like “Patient-related sxc mutant alleles do not affect global protein O-GlcNAcylation”.

Minor Points:

1) This last sentence of the abstract (“This study establishes a critical role for O-GlcNAc cycling and disrupted O-GlcNAc transferase activity in cognitive dysfunction and intellectual disability and points to potential treatment strategy for OGT-CDG.”) is a bit length with many “and” being used. Better to split into two sentences or rephrase.

2) The fly ortholog of OGT (sxc) has been reported to be a member of the Polycomb group as the authors discuss. Considering that many members of Polycomb genes are associated with human diseases that affect the nervous system including ID (reviewed in papers such as PMID: 34426021), the author may want to discuss the potential connection between OGT, Polycomb genes and O-GlcNAcylation, if there may be a link.

3) There is one prior study that found a role for OGT in the nervous system, specifically in the context of circadian rhythm regulation using timeless-GAL4 and UAS-RNAi. Although this study may not be directly relevant to this work, the author may want to consider citing this it as one previous evidence in flies that this gene has critical functions in the nervous system (note that this reviewer is not the author of this study nor have any conflict of interest).

Reviewer #4: Manuscript #: PGENETICS-D21-00961

Title: Intellectual disability-associated disruption of O-GlcNAcylation impairs neuronal development and cognitive function in Drosophila.

Summary: Fenckova et al. present an impressive series of behavioral, genetic, and molecular experiments to uncover phenotypes caused by increased and decreased O-GlcNAc modification. In addition to the use of existing genetic tools, they develop and molecularly characterize new strains that both further decrease transferase activity and mimic the intellectual disability (ID)-associated missense mutations found in humans. The focus of their experimental work centers on a simple learning assay, an habituated jump in response to repeated lights-off stimuli. The authors then extend these findings by hypothesizing a mechanistic basis, synaptic change, and by using additional behavioral assays. The experimental findings are, for the most part, clear and the text is well-written. While I recommend it for publication, I strongly suggest a number of major and minor revisions that should provide for more clarity and a greater over-all impact.

Major Revisions:

1. In both the title and throughout the Introduction and Discussion, the authors use the term ‘cognitive function’ lightly, basing this on the habituation defects observed in their mutants. However, cognition generally describes more complex, executive functioning and its use may misrepresent the experimental findings. A more fitting word choice would simply be ‘habituation’.

2. In the learning and memory literature, habituation as a form of learning is accepted once a number of criteria are met. First, the decreased response should not be fatigue (which the authors effectively rule out) and sensory adaptation should not explain the results (this is unaddressed by the authors). To suggest learning, the habituated response should also show a) recovery by dishabituating stimuli (an air puff, for example), b) that the jump response recovers spontaneously in a stimulus-dependent fashion, and c) that the habituation is stimulus specific. The authors should address (experimentally or through literature review) at least one of these criteria (a-c) before claiming that learning is being measured. If they or others have shown that sensory adaptation does not explain the habitutation, that too should be documented.

3. While the Introduction was informative as to the importance of O-GlcNAcylation, the authors make frequent interpretations of their habituation assay and of sxc function in the Results that are only explained after reading the Discussion, and then only briefly. In particular, the reader would benefit from a short introduction to a) habituation learning, b) the function of the TPR domain of sxc in relation to the glycosyl transferase catalytic region, and c) a structure-function analysis of ID-specific mutations (through crystallography), and their known effects (if any) on substrate interactions.

4. Regarding the light-off jump habituation, the data indicates that not only do genetic manipulations of O-GlcNAcylation alter the rate of habituation, but they also influence the post-habituation baseline activity level (indicated by higher baselines in Figure 1 and Figure 4). Furthermore, the rescue of the trials to criterion (TTC) throughout the paper correlates with the return of this baseline activity back to control levels. The authors should explain their interpretation of this higher baseline at its first occurrence in the Results. I am also wondering if, since the authors are measuring wing vibration and the actual jump frequency, there may be a change in the animal’s excitability. Given the change in NMJ bouton number (increased in some genotypes), a change in neuronal excitability might also be hypothesized in the Discussion.'

Minor Revisions:

1. Throughout the text there numerous grammatical errors (missing ‘the’, ‘for’, ‘a’, for example).

2. On page 6, lines 123 and 124, and then again on page 7, lines 148 and 149, the authors present data without showing the n or any statistics. These additions will strengthen the author’s conclusions and should be included.

3. The first sentence of the paragraph at the bottom of page 12 (lines 288-291) is a run-on and should be broken into two.

4. At the end of page 15 and the top of page 16 of the Discussion the authors discuss the link between O-GlycNAcylation and microtubule polymerization, but they don’t indicate the details that may be relevant – mechanisms of the modulation, for example, which may be relevant to the NMJ length measurements.

5. In Figure 2 A’-D’, larval NMJ synapses are labeled with a pre- and post-synaptic marker. There are also arrowheads to distinguish the type 1b synapses from other synapses. However, there appear only type 1b boutons in the images, making the arrowheads unnecessary.

**********

Have all data underlying the figures and results presented in the manuscript been provided?

Large-scale datasets should be made available via a public repository as described in the PLOS Genetics data availability policy, and numerical data that underlies graphs or summary statistics should be provided in spreadsheet form as supporting information.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

Reviewer #4: No

2 Mar 2022

Submitted filename: Point_to_point_response_PloSGenet_v7.docx

Click here for additional data file.

21 Mar 2022

Dear Dr van Aalten,

We are pleased to inform you that your manuscript entitled "Intellectual disability-associated disruption of O-GlcNAcylation impairs neuronal development and cognition-relevant habituation learning in Drosophila" has been editorially accepted for publication in PLOS Genetics. Congratulations!

Please note that Reviewer 1 has suggested to use the short title as the main title of the paper.  If the authors are in favor of this suggestion, we ask that this change is made in the final version of the manuscript.

Before your submission can be formally accepted and sent to production you will need to complete our formatting changes, which you will receive in a follow up email. Please be aware that it may take several days for you to receive this email; during this time no action is required by you. Please note: the accept date on your published article will reflect the date of this provisional acceptance, but your manuscript will not be scheduled for publication until the required changes have been made.

Once your paper is formally accepted, an uncorrected proof of your manuscript will be published online ahead of the final version, unless you’ve already opted out via the online submission form. If, for any reason, you do not want an earlier version of your manuscript published online or are unsure if you have already indicated as such, please let the journal staff know immediately at plosgenetics@plos.org.

In the meantime, please log into Editorial Manager at https://www.editorialmanager.com/pgenetics/, click the "Update My Information" link at the top of the page, and update your user information to ensure an efficient production and billing process. Note that PLOS requires an ORCID iD for all corresponding authors. Therefore, please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field.  This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager.

If you have a press-related query, or would like to know about making your underlying data available (as you will be aware, this is required for publication), please see the end of this email. If your institution or institutions have a press office, please notify them about your upcoming article at this point, to enable them to help maximise its impact. Inform journal staff as soon as possible if you are preparing a press release for your article and need a publication date.

Thank you again for supporting open-access publishing; we are looking forward to publishing your work in PLOS Genetics!

Yours sincerely,

Santhosh Girirajan

Associate Editor

PLOS Genetics

Gregory P. Copenhaver

Editor-in-Chief

PLOS Genetics

www.plosgenetics.org

Twitter: @PLOSGenetics

----------------------------------------------------

Comments from the reviewers (if applicable):

Reviewer's Responses to Questions

Comments to the Authors:

Please note here if the review is uploaded as an attachment.

Reviewer #1: The authors have done a thorough and comprehensive job to amend the paper. Now it reads nicely and presents a solid case for the conclusions reached by the authors.

The short title is much better than the long title, I suggest the authors use the short title as the main title of this paper.

Reviewer #3: The authors have made significant efforts to address my earlier concerns and questions. I am happy to support the publication of this paper in PLoS Genetics.

Reviewer #4: No response

**********

Have all data underlying the figures and results presented in the manuscript been provided?

Large-scale datasets should be made available via a public repository as described in the PLOS Genetics data availability policy, and numerical data that underlies graphs or summary statistics should be provided in spreadsheet form as supporting information.

Reviewer #1: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Victor Faundez

Reviewer #3: No

Reviewer #4: No

----------------------------------------------------

Data Deposition

If you have submitted a Research Article or Front Matter that has associated data that are not suitable for deposition in a subject-specific public repository (such as GenBank or ArrayExpress), one way to make that data available is to deposit it in the Dryad Digital Repository. As you may recall, we ask all authors to agree to make data available; this is one way to achieve that. A full list of recommended repositories can be found on our website.

The following link will take you to the Dryad record for your article, so you won't have to re‐enter its bibliographic information, and can upload your files directly:

http://datadryad.org/submit?journalID=pgenetics&manu=PGENETICS-D-21-00961R1

More information about depositing data in Dryad is available at http://www.datadryad.org/depositing. If you experience any difficulties in submitting your data, please contact help@datadryad.org for support.

Additionally, please be aware that our data availability policy requires that all numerical data underlying display items are included with the submission, and you will need to provide this before we can formally accept your manuscript, if not already present.

----------------------------------------------------

Press Queries

If you or your institution will be preparing press materials for this manuscript, or if you need to know your paper's publication date for media purposes, please inform the journal staff as soon as possible so that your submission can be scheduled accordingly. Your manuscript will remain under a strict press embargo until the publication date and time. This means an early version of your manuscript will not be published ahead of your final version. PLOS Genetics may also choose to issue a press release for your article. If there's anything the journal should know or you'd like more information, please get in touch via plosgenetics@plos.org.

Submitted filename: Re-review of Fenchova et al.docx

Click here for additional data file.

25 Apr 2022

PGENETICS-D-21-00961R1

Intellectual disability-associated disruption of O-GlcNAc cycling impairs habituation learning in Drosophila

Dear Dr van Aalten,

We are pleased to inform you that your manuscript entitled "Intellectual disability-associated disruption of O-GlcNAc cycling impairs habituation learning in Drosophila " has been formally accepted for publication in PLOS Genetics! Your manuscript is now with our production department and you will be notified of the publication date in due course.

The corresponding author will soon be receiving a typeset proof for review, to ensure errors have not been introduced during production. Please review the PDF proof of your manuscript carefully, as this is the last chance to correct any errors. Please note that major changes, or those which affect the scientific understanding of the work, will likely cause delays to the publication date of your manuscript.

Soon after your final files are uploaded, unless you have opted out or your manuscript is a front-matter piece, the early version of your manuscript will be published online. The date of the early version will be your article's publication date. The final article will be published to the same URL, and all versions of the paper will be accessible to readers.

Thank you again for supporting PLOS Genetics and open-access publishing. We are looking forward to publishing your work!

With kind regards,

Agnes Pap

PLOS Genetics

On behalf of:

The PLOS Genetics Team

Carlyle House, Carlyle Road, Cambridge CB4 3DN | United Kingdom

plosgenetics@plos.org | +44 (0) 1223-442823

plosgenetics.org | Twitter: @PLOSGenetics