Phosphorylations of Serines 21/9 in Glycogen Synthase Kinase 3α/β Are Not Required for Cell Lineage Commitment or WNT Signaling in the Normal Mouse Intestine

The WNT signalling pathway controls many developmental processes and plays a key role in maintenance of intestine renewal and homeostasis. Glycogen Synthase Kinase 3 (GSK3) is an important component of the WNT pathway and is involved in regulating β-catenin stability and expression of WNT target genes. The mechanisms underpinning GSK3 regulation in this context are not completely understood, with some evidence suggesting this occurs through inhibitory N-terminal serine phosphorylation in a similar way to GSK3 inactivation in insulin signaling. To investigate this in a physiologically relevant context, we have analysed the intestinal phenotype of GSK3 knockin mice in which N-terminal serines 21/9 of GSK3α/β have been mutated to non-phosphorylatable alanine residues. We show that these knockin mutations have very little effect on overall intestinal integrity, cell lineage commitment, β-catenin localization or WNT target gene expression although a small increase in apoptosis at villi tips is observed. Our results provide in vivo evidence that GSK3 is regulated through mechanisms independent of N-terminal serine phosphorylation in order for β-catenin to be stabilised.


Introduction
The small intestine is comprised of the crypts of Lieberkühn, invaginations of the epithelium containing pluripotent stem cells, and of villi, finger-like protrusions containing differentiated cells. The intestinal epithelium renews every 3-5 days in adults and this is achieved through the activity of the crypts [1]. The intestinal stem cells (ISCs), which reside at crypt bottoms, are the driving force behind intestinal renewal [2]. Their proliferative progenitors, transit amplifying (TA) cells, undergo 4-5 divisions before terminally differentiating [3]. During this process, the TA cells migrate upwards towards the villus base and differentiate into absorptive or secretory cells, namely enterocytes, goblet cells and enteroendocrine cells. These differentiated cells continue to move upwards and, upon reaching the villus tip, they undergo apoptosis and are shed into the gut lumen. The Paneth cell represents a fourth cell type, also derived from the ISCs, but this cell type migrates in the opposite direction into the crypt base [4,5].
At the crypt base is the 'stem cell zone' created by ISCs or crypt base columnar (CBC) cells interspersed between the Paneth cells [6]. These stem cells can be identified by their expression of key marker genes including Lgr5, Ascl2 and Olmf4 [7][8][9]. More recently a second population of slow cycling/quiescent cells has been noted to reside at the fourth position from the crypt base (+4) [10,11] and Bmi1, Tert and Hopx amongst other genes have been proposed as markers of this +4 population [12][13][14][15][16][17]. However the +4 model has been challenged with a number of studies demonstrating the expression of these markers also within the Lgr5+ population of CBCs [15,18]. A more general model of crypt plasticity attempts to reconcile both of these observations and suggests the existence of two populations of stem cells; one actual and one in reserve that can be called upon in response to tissue damage [16].
The WNT pathway plays a crucial role in intestinal proliferation and ISC maintenance as evidenced by ablation of the major WNT target gene T cell factor 4 (TCF4), which causes a loss of proliferative crypts [19]. A primary point of regulation of the WNT pathway is at the level of the destruction complex. In the absence of WNT proteins, the kinase GSK3 along with β-catenin, Adenomatous Polyposis Coli (APC), Casein Kinase 1 (CK1) and the scaffolding protein AXIN form this complex. In the basal state, GSK3 is active and phosphorylates β-catenin leading to its ubiquitination and subsequent degradation [20,21]. Upon WNT binding to Frizzled receptors and low-density lipoprotein receptor-related protein (LRP) co-receptors and subsequent engagement of Dishevelled, the destruction complex is antagonised and cytoplasmic βcatenin avoids phosphorylation by GSK3. These events result in β-catenin accumulation, nuclear translocation and the engagement of TCF/LEF (T-cell factor/lymphoid enhancer-binding factor) to activate the expression of downstream genes.
GSK3 has two isoforms, GSK3α and GSK3β, encoded by different genes and these kinases share over 97% sequence homology within their catalytic domains [22]. GSK3α and GSK3β are redundant in the WNT pathway [23] but the mechanisms underpinning WNT-induced inhibition of β-catenin phosphorylation by GSK3 are incompletely understood. Initially, parallels were drawn to the inactivation of GSK3 in the insulin signalling pathway whereby GSK3α/β is inactivated through inhibitory phosphorylation at residues S21/S9 by Protein Kinase B (PKB) [24]. Phosphorylation of GSK3α/β in this manner was shown to cause the N-terminal tail of GSK3 to associate in the substrate-binding pocket, preventing the binding of primed substrates [25]. However, subsequent analysis has shown that GSK3 regulation occurs through distinct mechanisms in the two pathways and that WNT signalling does not lead to a detectable change in S9 phosphorylation of GSK3β [26]. More recent investigations have suggested alternative models to explain how WNTs inhibit GSK3 [27,28]. One hypothesis is that phosphorylated PPPSPxS motifs in the cytoplasmic tail of LRP act as a competitive pseudosubstrate in the GSK catalytic pocket, thus preventing substrate phosphorylation through GSK3 inhibition [29,30]. A second proposes that GSK3 and β-catenin are spatially separated by the uptake of GSK3 into multivesicular bodies (MVBs), preventing cytoplasmic β-catenin phosphorylation [31]. Another model posits disruption of the destruction complex through APC/Axin dissociation from GSK3/β-catenin [32,33].
To investigate GSK3 regulation, and the specific role of N-terminal serine phosphorylation, in a biologically relevant tissue, we have analysed the intestinal phenotype of knockin mice in which serines 21 and 9 of GSK3α and GSK3β respectively have been mutated to alanine [34]. In vivo, we find no role for this phosphorylation event in WNT signalling or gut homeostasis although a small effect on intestinal apoptosis is observed. These data confirm that GSK3 regulation in the context of the WNT pathway must occur through mechanisms independent of Nterminal serine phosphorylation.

Experimental animals
Animal experiments were performed under Home Office project license authority. The GSK3α S21A/S21A /β S9A/S9A (GSK3 KI/KI ) mice used for this study have been previously reported [34] and were kindly provided by Dario Alessi, MRC Phosphorylation Unit, Dundee, UK. Mice were maintained on the C57BL6 background and male and female mice were randomly selected for all experiments. For the survival data, animals were kept on study until they became moribund and were sacrificed humanely according to Home Office regulations. For all other experiments, all mice used were 9-10 weeks of age and none of these became ill or died before this time point. For proliferation analysis, mice were injected with 1mg of BrdU by intraperitoneal injection and harvested 3 hours later [35].

Processing of tissue and staining
The small intestine was removed, flushed with PBS and cut into 6 equal segments that were cut longitudinally and "swiss-rolled" as described previously [36]. Each roll was placed in 4% (w/v) paraformaldehyde (PFA) in PBS and rocked gently for 16-18 hr. The PFA was decanted and the tissue was washed and stored in 70% (v/v) ethanol at 4°C. Tissue was processed and embedded in paraffin for sectioning. H&E staining was performed as previously described [35]. PAS staining was carried out according to the manufacturer's instructions using the PAS Kit (Thermo Scientific, 87007). For AB-PAS staining, sections were de-paraffinized, re-hydrated and Alcian Blue was added for 30 min (1% Alcian blue pH2.5 in 3% glacial acetic acid). Sections were washed for 5 min before continuing with the PAS kit instructions. Alkaline phosphatase staining was carried out using the BCIP/NBT Liquid Substrate System according to the manufacturer's instructions (Sigma, B1911). Stained sections were analysed with a Leica DFC420 light microscope and photographed using LAS V4.5.

Crypt and villi measurements
Villi and crypt length measurements were taken using the Leica light microscope as above. For crypt cell counts, only well-orientated crypts were counted and a minimum of 50 crypts per animal was scored for each experiment. All counting and measurements were undertaken blind and conducted on at least 3 animals per genotype.

Expression analysis
RNA extraction and cDNA synthesis were performed using the RNeasy Mini Kit (Qiagen, 74104) and SuperScript III Reverse Transcriptase (Invitrogen 18080-093) kits respectively following the manufacturers' instructions. qRT-PCR was performed using Bioline Sensifast SYBR No-ROX (BIO-98020) on a Roche Light Cycler 480 real-time cycling machine as outlined by the manufacturer and described in [37]. Each sample was analysed in triplicate and three biological replicates were utilised for each experiment. For stem cell marker genes, the expression level was normalized against the housekeeping gene Aprt1 and the expression values of KI/KI samples relative to that of WT/WT mice were expressed as fold change. Primers used for stem cell marker genes were: Aprt1 (For: ). For analysis of the Notch pathway, ready-made individual primers encoding regulatory genes of differentiation were used (QuantiTect Primer Assay, Qiagen). The mRNA expression level of these genes was standardized against the geometric mean of three housekeeping genes: Aprt1, Hprt1 and Gusb.

RT2 Profiler PCR Array
The Mouse WNT Signalling Targets (PAMM-243Z SABiosciences) was utilised to investigate a panel of 84 WNT specific genes in mouse intestinal tissue as per manufacturer's instructions.
Briefly, 500 ng of RNA was converted to cDNA using RT 2 First Strand Kit (SABiosciences). The resultant cDNA product was immediately amplified by qPCR using RT 2 SYBR Green qPCR Master Mix (SABiosciences) and a Roche Lightcycler 480. The Ct values (threshold cycle) for both KI/KI and WT/WT samples were evaluated using the provided web-based portal (http://pcrdataanalysis.sabiosciences.com/pcr/arrayanalysis.php) and normalised to housekeeping genes present on the array. This analysis generated fold-change, which represents the normalized gene expression in the KI/KI samples compared to the normalized gene expression in the WT/WT samples. Following the portal guidelines, a significant difference in expression was set at greater than or less than 3-fold difference in fold-change.

Statistical analysis
Comparison between any two groups was performed by an unpaired t-test assuming Gaussian distribution.

Phenotype of GSK3α/β KI/KI mutant mice
To obtain GSK3α/β single and double knockin mutant mice, GSK3α +/S21A animals were first intercrossed with GSK3β +/S9A animals. GSK3α +/S21A ;GSK3β +/S9A double heterozygous mice in the offspring were then subjected to a second round of breeding by intercrossing (Mating 1 in Table 1). An additional round of breeding was performed to generate further double   Table 1). Consistent with the results of McManus et al [34], double homozygous GSK3α S21A/S21A ;GSK3β S9A/S9A mice were born alive in both matings and survived to adulthood. Protein lysates were generated from the intestine of GSK3α S21A/S21A ;GSK3β S9A/S9A (called KI/KI) animals and immunoblot analysis with antibodies for phosphorylated forms of S21 of GSK3α and S9 of GSK3β confirmed inheritance of mutated knockin alleles (Fig 1A). Consistent with previous reports [34], the single and double homozygous and heterozygote knockin animals demonstrated no significant difference in survival compared to controls ( Fig 1B) and appeared overtly normal. Male and female weights at 10 weeks were indistinguishable from that of wild-type controls (Fig 1C). A decrease in villus number and loss of the proliferative compartment has been previously reported in mice with defective WNT signalling [19,38]. However the length of the small intestine was not significantly different in the GSK3 knockin mice compared to wild type controls ( Fig 1D) and neither was the length of the villi (Fig 1E). Furthermore, Haematoxylin & Eosin (H&E) staining of small intestine sections demonstrated normal tissue architecture and integrity (Fig 1F).

Intestinal cell proliferation and apoptosis
To further assess whether the GSK3 mutations affect intestinal homeostasis, we assessed proliferation and apoptosis in the crypts and villi. Total numbers of cells per intestinal crypt were not significantly different in the KI/KI animals compared to wild-type controls (Fig 2A). To investigate proliferation in more detail, BrdU incorporation was assessed, but demonstrated no difference between the knockin animals and controls (Fig 2B and 2C). Immunohistochemical staining for phosphohistone-H3 also demonstrated no difference in staining pattern (Fig 2B). In both KI/KI and WT/WT mice, cycling cells were confined to the mid crypt region with there being no difference in the localisation of these cells in the mutant mice ( Fig 2D).
To assess apoptosis, immunostaining was performed using an antibody for cleaved caspase 3 and this demonstrated a moderate increase in staining in the villi tips of the KI/KI animals compared to controls (Fig 2E). An increase in levels of apoptosis was supported by immunoblotting, demonstrating increased levels of cleaved PARP and cleaved caspase 3 in the majority of the KI/KI intestinal samples compared to controls (Fig 2F).
To further confirm that patterns of cell differentiation are not affected by the GSK3α/β knockin mutations, we undertook quantitative gene expression analysis of key genes known to be involved in intestinal epithelial cell differentiation namely Atoh1, Dll1, Elf3, Gata4, Hes1, Hnf4a, Ngn3 and Spdef1. The expression of these genes was not significantly different between KI/KI intestinal samples and controls (Fig 3C).

WNT pathway activation and stem cell homeostasis
GSK3 is a key point of control of the canonical WNT signalling pathway, with the output of this pathway involving nuclear β-catenin translocation and transactivation of WNT target genes. To address the involvement of N-terminal serine phosphorylation of GSK3 in the canonical WNT signalling pathway in the intestine, immunohistochemical staining for β-catenin was performed using KI/KI samples and controls (Fig 4A). β-catenin localisation was observed in crypt base cells in tissue derived from both sets of mice and there was no difference in the number or pattern of nuclear β-catenin staining (Fig 4A). To further examine the consequence of GSK3 mutation on the WNT pathway we assessed the expression of 84 WNT target genes using a real-time PCR array. The difference in expression of these genes in the KI/KI intestine compared to control intestine is shown in the scatter plot in Fig 4B. The data show no significant difference in the expression of any of these genes.
The WNT signalling pathway is known to play a key role in stem cell maintenance and fate in the mouse small intestine and ISCs have been identified at the +4 position as well as in the crypt base [39]. The +4 stem cells and CBCs have been distinguished based on expression of key genes with Lgr5 and Ascl2 identifying CBCs and Bmi1, Tert and Hopx thought to distinguish +4 cells [39]. To investigate if GSK3 phosphorylation plays a role in stem cell maintenance, the expression of these genes and other genes known to be responsible for the proliferative effects of the WNT pathway namely Axin2, Ccnd1 and c-Myc [40,41] were examined using quantitative RT-PCR. We found no significant difference in the expression of these genes in the KI/KI intestine compared to controls (Fig 4C).

Discussion
GSK3 is a conserved serine-threonine kinase that was originally identified as an important regulator of insulin-dependent glycogen synthesis [42]. GSK3, and specifically its two isoforms GSK3α and GSK3β, has since been shown to have a wide range of substrates and be involved in multiple physiological processes including apoptosis, cell differentiation and proliferation. As a consequence GSK3 has been implicated in many human pathologies including cancer, diabetes and Alzheimer's disease [28,43]. Given its pleiotropic functions and therapeutic potential, the mode of regulation of GSK3 under different cellular contexts and conditions has been the subject of intensive investigation. statistically significant difference was observed using the Student's t-test. D: The position of BrdU-positive cells within the WT/WT and KI/KI crypts was quantified. Position 0 represents the base of the crypt. A minimum of 50 crypts were quantified in three mice of each genotype. No statistically significant difference was observed using the Student's t-test. E: Immunohistochemical staining of WT/WT and KI/KI villi using an antibody for cleaved caspase 3. Scale bars, 50μm. F: Western blot analysis of small intestinal protein samples from WT/WT or KI/KI mice using antibodies for phosphorylated GSK3α/β, cleaved PARP or cleaved caspase-3. Immunoblotting for ERK2 and GAPDH was used to determine equal protein loading.
doi:10.1371/journal.pone.0156877.g002 In the bottom two panels, sections were subjected to PAS and AB-PAS staining to detect neutral and acid mucin-producing goblet cells GSK3 kinase activity is inhibited through phosphorylation of serine residues within its Nterminal region and, specifically, serine 21 in GSK3α and serine 9 in GSK3β. A number of kinases have been shown to phosphorylate these residues including PKB/AKT, p70 S6K and p90 RSK, with PKB/AKT targeting GSK3 phosphorylation in response to insulin signalling. Within the WNT signalling pathway, GSK3 is also subject to regulation and phosphorylation of serines 21/9 has been suggested to play a role in this [44,45]. However, this mode of regulation has been contradicted by more recent observations [26,43], including the demonstration that embryonic stem (ES) cells carrying homozygous serine-alanine mutations at residues 21/9 of GSK3 α/β do not have disrupted WNT signalling [34]. In this manuscript we have extended the analysis to a more physiologically relevant setting, and specifically through analysis of the intestine, which is known to be dependent on the WNT pathway for renewal and tissue homeostasis [19]. Our data show that, like ES cells, the activity of the WNT pathway is not dependent on N-terminal serine phosphorylation of GSK3 and that these phosphorylation events play only a minor role in maintenance of intestinal integrity. Thus, alternative mechanisms must underpin GSK3 regulation in canonical WNT signalling [27][28][29][30][31][32][33].
Homozygous knockin mice with the S21/9A GSK3 mutations survive the normal lifespan of a mouse (Fig 1), are fertile with very little alteration in the anatomical and molecular characteristics of the tissue upon which we have focussed-the small intestine (Figs 1, 2, 3 and 4). However, a small increase in apoptosis levels in the villi tips was observed using immunohistochemical analysis ( Fig 2E) and this was validated using western blot analysis ( Fig  2F). In line with these data, GSK3β has previously been identified as an anti-apoptotic kinase as evidenced by analysis of Gsk3β knockout mice that undergo embryonic lethality due to failure to activate the TNF α/NF-κB survival pathway [46]. Conversely, GSK3 has also been shown to induce apoptosis in response to a number of stimuli including DNA damage, hypoxia and heat shock with multiple mechanisms being reported including modulation of TP53, eIF2B and BCL-2 family members [43,47]. The opposing roles of GSK3 in apoptosis regulation are likely related to its intricate mode of control and the fact that it has pleiotropic functions and multiple substrates. The specific role of N-terminal GSK3 serine phosphorylation in suppressing or promoting apoptosis has not previously been investigated although a recent in vitro study suggests a role of S9 GSK3β phosphorylation in antagonising apoptosis in a NF-κBdependent manner [48]. Whether this is the case for the intestine requires further experimental investigation.
There is currently considerable interest in targeting GSK3 for therapeutic benefit in cancer. However, this is complicated by the fact that GSK3 has both pro-and anti-tumourigenic functions [28,47]. Examination of the specific mechanisms by which GSK3 functions in different cancers may allow therapies to be tailored for optimal benefit. To this end, N-terminal serine phosphorylation of GSK3 occurs in response to growth factor signals and GSK3 regulation by this mechanism has been observed in cells bearing oncogenic mutations in growth factor pathway components including PIK3CA, RAS and RAF [35,[49][50][51]. GSK3 has been shown to mediate oncogenic effects in these contexts including through the stabilisation of Cyclin D1 respectively. Scale bars, 50μm. B: Quantification of goblet cell number. PAS+ and AB-PAS+ cells were quantified in the villi of WT/WT and KI/KI mice. A minimum of 50 villi were quantitated per animal from at least five mice of each genotype and mean values are shown + SEM. Cell counts were adjusted according to villi length. No statistically significant difference was observed using the Student's t-test. C: Expression of genes involved in intestinal differentiation. Real-time PCR analysis was used to determine transcript levels in RNA specimens of the small intestine derived from three WT/WT or three KI/KI mice for a panel of 8 genes involved in epithelial differentiation. Shown are the fold-changes in transcript levels between KI/KI and WT/WT mice. No statistically significant difference in the expression of any of the eight genes was observed using the Student's t-test (P values for each gene are shown).  Real-time PCR analysis was used to determine transcript levels in RNA specimens of the small intestine derived from three WT/WT or three KI/KI mice for a panel of 84 WNT target genes. The Scatter plot shows the differential expression patterns for each gene between the WT/WT [49]. Therefore, strategies that reverse N-terminal GSK3 serine phosphorylation may be effective in cancers with the relevant oncogenic mutations. The fact that the S21/9A GSK3α/β homozygous knockin mice survive normally with relatively few pathologies, as shown in this report, is encouraging that any therapy targeting this phosphorylation event would have little effect on normal cells or generate adverse reactions at the tissue or whole organism level.