IL-7Rα glutamylation and activation of transcription factor Sall3 promote group 3 ILC development

Group 3 innate lymphoid cells (ILC3) promote lymphoid organogenesis and potentiate immune responses against bacterial infection. However, how ILC3 cells are developed and maintained is still unclear. Here, we show that carboxypeptidase CCP2 is highly expressed in common helper-like innate lymphoid progenitors, the progenitor of innate lymphoid cells, and CCP2 deficiency increases ILC3 numbers. Interleukin-7 receptor subunit alpha (IL-7Rα) is identified as a substrate of CCP2 for deglutamylation, and IL-7Rα polyglutamylation is catalyzed by polyglutamylases TTLL4 and TTLL13 in common helper-like innate lymphoid progenitors. IL-7Rα polyglutamylation triggers STAT5 activation to initiate transcription factor Sall3 expression in common helper-like innate lymphoid progenitors, which drives ILC3 cell differentiation. Moreover, Ttll4 −/− or Ttll13 −/− mice have reduced IL-7Rα polyglutamylation and Sall3 expression in common helper-like innate lymphoid progenitors. Importantly, mice with IL-7Rα E446A mutation have reduced Sall3 expression and ILC3 population. Thus, polyglutamylation and deglutamylation of IL-7Rα tightly controls the development and effector functions of ILC3s.

All ILC cells are derived from common lymphoid progenitors (CLPs), which also differentiate to T and B cells 13 . ILC3s, together with other ILCs, are derived from the earliest α-lymphoid progenitor cells (αLPs, CXCR6 + integrin α 4 β 7 -expressing CLPs) 2 , which differentiate into common helper-like innate lymphoid progenitor (CHILP) cells 14 . CHILPs generate all ILCs including LTi cells but not NK cells. Downstream of CHILP, ILC progenitors (ILCP), characterized by expression of the transcription factor (TF) PLZF, lose the ability to generate LTi cells and give rise to all ILC1, ILC2, and ILC3 subsets 15 . RORγt (encoded by Rorc) drives differentiation of ILC3s from their precursor ILCPs 16,17 . RORγt deletion causes a complete loss of ILC3s but not ILC1s or ILC2s. Of note, the cytokine receptor chain IL-7Rα (CD127) is constitutively expressed in CHILPs and all ILCs, and forms a heterodimer with the common γ-chain of IL-2R or thymic stromal lymphopoietin (TSLP) receptor to detect IL-7 and TSLP, respectively 14,18 . However, how IL-7Rα signaling regulates the ILC development and/or maintenance still remains elusive.
Protein post-translational modifications (PTM) such as phosphorylation, glycosylation, acetylation, and ubiquitination have critical functions in the regulation of activities of target proteins by changing their chemical or structural properties 19,20 . Another PTM, glutamylation, adds glutamate side chains onto the γ-carboxyl groups of glutamic acid residues in the sequence of target proteins [21][22][23] . Glutamylation is catalyzed by polyglutamylases, also called tubulin tyrosine ligase-like (TTLL) enzymes 24,25 . Glutamylation is a reversible modification that can be hydrolyzed by a family of cytosolic carboxypeptidases (CCPs) 26 . Misregulations of glutamylation contribute to several physiological abnormalities. CCP1 deficiency causes hyperglutamylation of tubulins resulting in Purkinje cell degeneration 26,27 . We previously demonstrated that CCP6 deficiency induces hyperglutamylation of Mad2, leading to underdevelopment of megakaryocytes and abnormal thrombocytosis 28 . In addition, we also show that glutamylation of the DNA sensor cGAS regulates its binding and synthase activity in antiviral immunity 29 , suggesting that glutamylation is involved in the regulation of immune response. However, how glutamylation regulates the development and/or maintenance of ILCs is unknown.
Here, we show that IL-7Rα can be glutamylated by TTLL4 and TTLL13, and deglutamylated by CCP2. IL-7Rα glutamylation enhances STAT5 activation and then promotes Sall3 transcription in CHILPs that drives the development of ILC3s. Therefore, IL-7Rα glutamylation has a critical function in ILC3 development.

Results
CCP2 deficiency increases ILC3 numbers. We previously demonstrated that deficiency in CCP5 or CCP6 leads to susceptibility to virus infection 29 . CCP5 and CCP6 are required for the activation of TF IRF3 and IFN induction. We, therefore, sought to explore whether glutamylation was involved in the development of ILCs and their defense against bacterial infection. We used previously established Ccp1-6 knockout (KO) mice and further validated deletion of these genes in mouse bone marrow (BM) (Supplementary Fig. 1a). We analyzed ILC3s (Lin − CD45 + RORγt + ) in the small intestine lamina propria in all six deficient mouse strains and found that the number of ILC3 cells was significantly increased in Ccp2 −/− mice, but not in other CCP KO mouse strains ( Fig. 1a and Supplementary Fig. 1b,  c). ILC3 cells can be divided into a set of subpopulations according to their expression of CD4 and NKp46 (encoded by Ncr1) receptors, such as CD4 + ILC3s, NKp46 + ILC3s, and CD4 − NKp46 − ILC3s (DN ILC3s) 6,30 . We then determined changes of NKp46 + ILC3s (Lin − CD45 + RORγt + NKp46 + ) and NKp46 − ILC3s (Lin − CD45 + RORγt + NKp46 − ) in Ccp2 −/− mice. We observed that both of NKp46 + ILC3s and NKp46 − ILC3s were markedly increased in Ccp2 −/− mice, but not in other CCP-deficient mouse strains ( Fig. 1b and Supplementary Fig. 1d). These observations were further verified by immunofluorescence staining (Fig. 1c). By contrast, CCP2-deficient mice displayed reduced numbers of ILC1s and ILC2s ( Supplementary Fig. 1e, f).
CCP2 deficiency potentiates ILC3 differentiation from CHILPs. We next analyzed expression patterns of CCP members in the mouse hematopoietic system. We found that Ccps displayed distinct expression profiles in different hematopoietic cell populations and their progenitors (Fig. 2a). Of note, Ccp2 was highly expressed in the CHILPs and ILC3s (Fig. 2a). Intriguingly, CCP2 deficiency led to reduced numbers of CHILPs, whereas more ILCPs in BM ( Fig. 2b and Supplementary Fig. 1h), suggesting CCP2 was involved in the development of ILC3s from the stage of CHILPs. We then conducted in vitro differentiation assays. We isolated CHILPs from Ccp2 +/+ and Ccp2 −/− mice and cultured them with OP9 feeder cells in the presence of murine IL-7 (25 ng/ml, Peprotech) and SCF (25 ng/ml, Peprotech). We noticed that Ccp2 −/− CHILPs generated more ILC3s compared to Ccp2 +/+ CHILPs (Fig. 2c, d and Supplementary Fig. 1i). Moreover, overexpression of CCP2 dramatically reduced the formation of ILC3s, indicating that CCP2 was implicated in the development of ILC3. CoCl 2 is an agonist for CCP family proteins 32 , and phenanthroline (Phen) is their pan inhibitor 26 .
IL-7Rα is a substrate of CCP2 in CHILPs. To further explore the molecular mechanism of CCP2-mediated ILC3 differentiation, we analyzed lysates of Ccp2 +/+ and Ccp2 −/− BM by immunoblotting with a glutamylation-specific antibody GT335. The antibody GT335 specifically recognizes the branch points of glutamate side chains and detects all glutamylation forms of target proteins 26 . After immunoblot analysis, one band around 60 kD appeared in the lane of CCP2-deficient BM lysates (Fig. 3a). This band was undetectable in the corresponding lane location from the littermate control BM lysates. Thus, this band could be a potential candidate substrate for CCP2. To identify the candidate substrates of CCP2, we generated an enzymatically inactive mutant of CCP2 (CCP2-mut) through H425S and E428Q mutations as previously described 26 . WT CCP2 (CCP2-wt) and CCP2-mut were immobilized with Affi-gel10 resin to go through mouse BM lysates for affinity chromatography. The eluted fractions were resolved by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), followed by silver staining. This band was present in the gel analyzing CCP2-mut and was cut for mass spectrometry, whose band was identified as IL-7Rα ( Fig. 3b and Supplementary Fig. 2a). We transfected Myc-tagged intracellular (amino acid: 265-459) or Myc-tagged extracellular segment (amino acid: 21-239) of IL-7Rα with Flag-tagged CCP2-wt or CCP2-mut into 293 T cells for co-immunoprecipitation assay. We found that the Flag-tagged enzymatic dead CCP2 (Flag-CCP2-mut) could pull down the Myc-tagged intracellular segment of IL-7Rα (hereafter we called kD IL-7Rα) (Fig. 3c). By contrast, Flag-CCP2-mut failed to precipitate Myc-tagged extracellular segment of IL-7Rα (Fig. 3c). Consistently, glutamylated GST-tagged intracellular segment of IL-7Rα protein could pull down MBP-CCP2-mut by a pulldown assay (Fig. 3d), suggesting the intracellular segment of IL-7Rα was deglutamylated by CCP2. Moreover, MBP-tagged mutant CCP2 (MBP-CCP2-mut) was able to pull down native IL-7Rα from BM lysates, whereas the enzymatic active CCP2 (MBP-CCP2-wt) could not precipitate IL-7Rα (Fig. 3e). These observations suggest that the intracellular segment of IL-7Rα undergoes deglutamylation by CCP2. With immunofluorescence staining, hyperglutamylation of IL-7Rα appeared in primary CCP2-deficient CHILPs (Fig. 3f). In parallel, IL-7Rα was highly polyglutamylated in BM lysates of Ccp2 −/− mice (Fig. 3g). Finally, BM cells treated with the CCP antagonist Phen increased substantial amounts of polyglutamylated IL-7Rα, whereas treatment with the CCP T t l l 1 T t l l 2 T t l l 4 T t l l 5 T t l l 6 T t l l 7 T t l l 9 T t l l 1 1 T t l l 1 3 WT a d e f g j k l agonist CoCl 2 abrogated the glutamylation of IL-7Rα (Fig. 3h). Collectively, we conclude that IL-7Rα is a novel substrate for CCP2.
IL-7Rα is polyglutamylated at Glu446 by TTLL4 and TTLL13. Nine polyglutamylases have been reported to catalyze protein glutamylation 21,24 . To determine the physiological polyglutamylases catalyzing IL-7Rα glutamylation, we examined expression patterns of all nine polyglutamylases in αLPs and CHILPs of mouse BM through quantitative real-time PCR. We observed that Ttll4 and Ttll13 were highly expressed in αLPs and CHILPs, with peak expression in CHILPs (Fig. 4a). Additionally, Ttll4 and Ttll13 were highest expressed in CHILPs among all the hematopoietic progenitor cells (Fig. 4b). We next incubated recombinant intracellular segment rGST-IL-7Rα with Flag-tagged TTLL4 or TTLL13 in vitro. We noticed that Flag-tagged TTLL4 and TTLL13 were able to precipitate rGST-IL-7Rα (Fig. 4c). Their interactions were further verified by co-transfection assays (Fig. 4d). Moreover, IL-7Rα was co-localized with TTLL4 and TTLL13 in CHILPs (Fig. 4e). We then conducted in vitro glutamylation assays by incubation of rGST-IL-7Rα with Flag-TTLL4 or Flag-TTLL13. We found that rGST-IL-7Rα was polyglutamylated by TTLL4 and TTLL13 (Fig. 4f). Importantly, TTLL4-and TTLL13-mediated polygutamylation of rGST-IL-7Rα was successfully removed by enzymatic active CCP2 (Fig. 4g). These data indicate that TTLL4 and TTLL13 are two polyglutamylases for IL-7Rα polyglutamylation. Glutamate-rich stretches and acidic environment at the acceptor sites have been reported to be important for glutamylation modification 33 . Based on the conservative aminoacid sequence analysis, only Glu446 and Glu447 were two conserved identical glutamic acid residues located on the loop region of intracellular domain of IL-7Rα ( Supplementary Fig. 2b), which might be potential acceptor site candidates for glutamylation. We then mutated Glu446 to Ala of IL-7Rα (E446A-IL-7Rα) and incubated recombinant intracellular E446A-IL-7Rα protein with Flag-TTLL4 or Flag-TTLL13 in vitro. We observed that E446A-IL-7Rα mutant abolished TTLL4-or TTLL13-mediated glutamylation (Fig. 4h), suggesting IL-7Rα is catalyzed by TTLL4 and TTLL13 at Glu446. We next explored the physiological relevance of IL-7Rα glutamylation in ILC3 differentiation. We silenced IL-7Rα by LMP retrovirus-carried short hairpin RNA (shRNA) infection in CHILPs and then rescued expression of WT-IL-7Rα or E446A-IL-7Rα, followed by BM transplantation assays. Eight weeks after transplantation, IL-7Rα knockdown with empty vector infection remarkably decreased ILC3 numbers ( Supplementary Fig. 2c). By contrast, WT-IL-7Rα restoration could rescue the normal number of ILC3s in recipient mice, whereas E446A-IL-7Rα mutant overexpression had no such effect ( Supplementary  Fig. 2c). Parallelly, these observations were further validated by in vitro differentiation assays ( Supplementary Fig. 2d, e).
IL-7Rα glutamylation promotes Sall3 expression by STAT5. IL-7Rα pairs with the common γ-chain of IL-2R or TSLP receptor to detect IL-7 and TSLP, respectively, for the activation of STAT proteins in DCs, CD4 + T as well as B cells 18,34,35 . However, how IL-7Rα glutamylation regulates the development of ILCs remains unclear. We then analyzed all STAT protein phosphorylation signals in Ccp2 +/+ and Ccp2 −/− CHILPs. We found that only STAT5 was hyperphosphorylated in Ccp2 −/− CHILPs compared to Ccp2 +/+ CHILPs with IL-7 stimulation (Fig. 5a). However, other STAT proteins were not activated (Fig. 5a). We thus used STAT3 as a negative control in the following experiments. These observations were further validated by flow cytometry and immunofluorescence staining (Fig. 5b, c). These results indicate that IL-7Rα glutamylation leads to STAT5 activation in CHILPs after IL-7 stimulation.

Discussion
ILCs are a distinct arm of the innate immune system, which can directly communicate with other hematopoietic and non-hematopoietic cells to regulate immunity, inflammation and tissue homeostasis 1 . However, how these ILC lineages develop and/or maintain remains unclear. In this study, we show that CCP2 deficiency causes increased numbers of ILC3s. With IL-7 engagement, IL-7Rα undergoes polyglutamylation in CHILPs. IL-7Rα polyglutamylation specifically activate STAT5 phosphorylation to initiate Sall3 expression for ILC3 development (Fig. 6k). In addition, Ttll4 −/− and Ttll13 −/− mice abrogate IL-7Rα polyglutamylation and Sall3 expression in CHILPs, leading to impaired ILC3 differentiation and more susceptibility to C. rodentium infection. Finally, E446A-IL-7Rα mutation mice indeed abrogates Sall3 expression and ILC3 development.
The earliest progenitor cells specific to ILCs are CXCR6 + integrin α 4 β 7 -expressing CLPs, referred to as α-lymphoid precursor (αLP) cells, which give rise to ILC1, ILC2, ILC3, and conventional NK cells (cNK) 40 . The common progenitor to all ILC lineages (CHILP) is identified as its Lin − IL-7Rα + Id2 + CD25 − α 4 β 7 + phenotype and differentiates to all ILC subsets, but not cNKs 14 . The common precursor to ILCs (ILCP) is defined by expression of TF PLZF and generates ILC1, ILC2, and ILC3 subpopulations 15 . In this study, we show that CCP2 is highly constitutively expressed in CHILPs and ILC3s, which blocks the deglutamylation of IL-7Rα to drive ILC3 development. CCP2 deficiency causes increased numbers of ILC3s, but reduced numbers of ILC1s and ILC2s, which augments clearance of C. rodentium. Given that CCP2 is also moderately expressed in other cells such as CD3 + T cells, we thus cannot exclude the potential involvement of other cells in the bacterial clearance of CCP2 deficiency. Of note, CCP2 deficiency does not impact cell deaths of CHILPs and all ILC lineages. A recent study showed that different ILC subsets are defined by distinct gene-expression patterns 41 . Of note, cytokines such as IL-7, IL-15, and IL-2 play major roles in the regulation of ILC development. However, how CCP2-mediated IL-7 signaling regulates the switch balance of ILC development still needs to be further investigated. We notice that CCP members are differentially expressed in the hematopoietic progenitors and lineages we checked. We previously demonstrated that CCP6 is mostly highly expressed in BM and megakaryocytes, and also exhibits different expression profiles in different tissues and cell types 28 . Our findings suggest that different tissue and cell type distributions of CCPs may exert unique roles in the modulation of different physiological and pathological processes.
Protein polyglutamylation is catalyzed by a family of polyglutamylases, also called TTLLs 24,25 . The well-known substrates of polyglutamylation are tubulins and nucleosome assembly proteins 33 . Through regulating the interaction of microtubules (MTs) and MT-associated proteins (MAPs), polyglutamylation may exert major effects on MT-related cellular processes, including stability of centrosomes 42 , motility of cilia and flagella 43,44 , neurite outgrowth 45 , as well as neurodegeneration 26 . A recent study delineates a structural MT recognition basis by catalysis with TTLL7 21 . TTLLs have different expression patterns in diverse tissues and their functions are not entirely redundant 43 .
We recently reported that TTLL4 and TTLL6 are most highly expressed in megakaryocytes 28 , both of whom catalyze polyglutamylation of Mad2 to modulate megakaryocyte maturation. Here we demonstrate that TTLL4 and TTLL13 are constitutively elevated in CHILPs, both of which can catalyze polyglutamylation of IL-7Rα to regulate the development of ILC3s. Deletion of TTLL4 or TTLL13 impairs ILC3 differentiation and their effector functions. Thus, IL-7Rα polyglutamylation mediated by TTLL4 or TTLL13 has a critical function in the regulation of ILC3 development from the stage of CHILPs.
IL-7Rα (CD127), encoded by Il7r gene, forms a receptor complex with the common cytokine receptor γ-chain of IL-2R or TSLP receptor to sense IL-7 and TSLP, respectively 18,46 . The IL-7-IL-7Rα ligand-receptor pair signaling is critical for proliferation and survival of T and B lymphocytes in a non-redundant fashion. Genetic aberrations of IL-7Rα signaling lead to immune deficiency syndromes and other immune diseases 47,48 . It has been reported that all ILC lineages express high levels of IL-7Rα 14 . Of note, the ILCP CHILPs also express IL-7Rα, which gives rise to all ILCs. However, the molecular mechanism by which IL-7Rα signaling regulates the development of ILCs remains elusive. In this study, we show that TTLL4 and TTLL13-mediated IL-7Rα polyglutamylation regulates the differentiation of ILC3s from CHILPs. Mechanistically, polyglutamylated IL-7Rα is able to activate STAT5 and phosphorylated STAT5 can directly bind to Sall3 promoter to initiate its transcription, which drives the development of ILC3s from CHILPs.
A CHILP cell has been defined that lacks expression of Flt3 and CD25 but expresses IL-7Rα and α 4 β 7 14 . CHILPs differ from α-LPs in that CHILPs express Id2. CHILPs generate all ILCs, including LTi cells, but they fail to give rise to conventional NK cells. Subsequently, their downstream precursor ILCPs (common precursor of ILCs), characterized by expression of the TF PLZF, lose the ability to generate LTi cells and produce all ILC1, ILC2, and ILC3 subsets 15 . RORγt (encoded by Rorc) drives differentiation of ILC3s from their precursor ILCPs 16 . RORγt deletion causes a complete loss of ILC3s but not ILC1s or ILC2s. Runx3 is also required for the development of ILC1s and ILC3s, but not for ILC2s 6 . GATA3 is also involved in the development of ILC3s, and It continues to exert a critical role in mature ILC3s 49,50 . These observations suggest that the development of different ILC subsets are controlled by TF networks 41 . Sall3 (Spalt-like transcription factor 3) belongs to the SAL family, which is implicated in embryonic development 36,37,51 . However, how Sall3 regulates the development of ILCs is still unknown. Here we define that Sall3 is a downstream target of IL-7 signaling, whose expression induced by IL-7Rα polyglutamylation drives CHILPs to differentiate ILC3s.
Glutamylation is highly conserved in all metazoans and protists, exerting critical roles in many physiological and pathological processes 52 . For example, TTLL7, the most abundantly expressed TTLLs in the mammalian nervous system, is conserved from acorn worm to primates, where it modulates neurite outgrowth and localization of dendritic MAPs 45 . ILC3s are enriched in Peyer's patches (PPs) and intestinal lamina propria 30 . Prior to the development of adaptive immunity, ILC3-induced IL-22 production has a critical function in priming innate immunity to eradicate C. rodentium 53,54 . IL-22-deficient mice displayed exaggerated intestinal inflammation and impairment of the epithelial barrier and rapidly succumbed to bacterial infection. Given that ILC3s produce large amounts of IL-22, we thus used C. rodentium infection as a readout for determining the physiological function of ILC3s in the knockout mouse responses. The host protective effects of ILC3s are not restricted to bacterial infection in the intestine. ILC3s are also implicated in the resistance to infections of Candida albicans and Mycobacterium tuberculosis in the lungs 55,56 . Thereby ILC3s may be targeted to enhance or block immune responses for inflammatory pathology and immunotherapy. In this study, we show that the glutamylation and deglutamylation of IL-7Rα mediated by CCP2 and TTLL4/13 controls the development and effector function of ILC3s. Therefore, we strongly believe that it is necessary to develop specific inhibitors or agonists for these related polyglutamylases and CCPs. Manipulating polyglutamylation profiles by using these compounds, we may potentially target ILC3s for future clinical applications. In sum, IL-7Rα polyglutamylation has a critical function in the regulation of ILC3 development and their effector function. Our findings provide new mechanistic insights into how polyglutamylation modulates ILC3 development. Generation of knockout mice and Il7r E446A mice. Ccp1 and Ccp6 knockout mice were described previously 28 , and Ttll13 −/− mice were generated through CRISPR-Cas9 approaches as described 29 . Gt(ROSA)26Sor tm1(CAG-xstpx-cas9,-EGFP)Fezh , Rorc(γt) +/GFP and Id2 +/GFP mice were purchased from the Jackson Laboratory. Stat3 flox/flox was kindly provided by Dr Shizuo Akira (Osaka University, Japan). Stat3 flox/flox ; MxCre + mice were obtained by crossing Stat3 f/f mice with MxCre + mice. To induce STAT3 deletion, 200 μg polyinosine-polycyticylic acid (poly(I:C)) was intraperitoneally injected to mice every other day for three times. Mouse experiments were performed according to the guidelines of the institutional animal care and use committees at the Institute of Biophysics, Chinese Academy of Sciences. For deletion of Sall3 in BM, B6;129-Gt(ROSA)26Sor tm1(CAG-xstpx-cas9,-EGFP)Fezh knockin mice were crossed with Vav-Cre transgenic mice to generate Rosa26-LSL-Cas9 + ;Vav-Cre + mice. In all, 2 × 10 6 BM cells were infected with lentiCRISPRv2 containing sgSall3 lentivirus. BM cells were then transplanted into lethally irradiated recipient mice (CD45.1 + ). Donor-derived ILC3s were analyzed 8 weeks post transplantation. Sall3 deletion was confirmed by immunoblotting. For generation of Il7r E446A mice, the genome locus of Il7r gene was knocked in with IL-7Rα-E446A mutation via a CRISPR-Cas9 approach. Mixture of Cas9 mRNA, single guide RNA (sgRNA), and IL-7Rα-E446A donor templates was microinjected into the cytoplasm of C57BL/6 fertilized eggs and transferred into the uterus of pseudopregnant ICR females. IL-7Rα-E446A mutations were identified by PCR screening and DNA sequencing. gRNA sequences are as follows: Ccp2: 5′-TAGAAATATTCTGGTTGATGTGG-3′; Ccp3: 5′-GGAGTATCAGCTAGGAAGATGGG-3′; Ccp4: 5′-AGCTCT-GAGCTGGTGCTCCCAGG-3′; Ccp5: 5′-GGTTCTACTTCAGTGTCCGGGG-3′; Ttll4: 5′-TTTGCCTCACGTTGGTGCGGCGG-3′; Ttll13: 5′-TTTCTTGGCTA-CAACCGATAAGG-3′; Il7r: 5′-TTCTTCTTGATTCAGTACTGAGG-3′; Sall3: 5′-CCAGCATCTCAAGTCGGACG-3′. Mice used in all experiments were 8-weeks old. And we performed three independent experiments of each mouse from at least three mice for each group. The background of mice was C57BL/6, and mice were grouped by the same age and gender. Animal use and protocols were approved by the Institutional Animal Care and Use Committees at Institute of Biophysics, Chinese Academy of Sciences.
Histology analysis. Mouse colons after C. rodentium infection were fixed in 4% PFA (Sigma-Aldrich) for 24 h, washed twice with phosphate-buffered saline (PBS) and stored using 75% ethanol before embedded in paraffin. Then colons in paraffin were sectioned and stained with hematoxylin and eosin (H&E) according to standard laboratory procedures.
Intestinal lymphocyte separation. Protocols for lymphocyte isolation from the intestine had been described 57 . With some modifications, intestines were dissected and cleaned, then PPs were removed. Intestines were cut longitudinally and washed with Dulbecco's Phosphate Buffered Saline (dPBS) five times. Then intestines were cut into pieces, and washed with solution I buffer (10 mM HEPES and 5 mM EDTA in Hank's Balanced Salt Solution (HBSS)) five times. For LPL isolation, the intestinal fragments were digested with solution II buffer containing DNaseI, 5% FBS, 0.2 mg/ml collagenase II and collagenase III there times at 37°C. Then the tissues were sifted through 70-μm strainers.
Immunofluorescence assay. Cells were isolated by fluorescence-activated cell sorting (FACS) and fixed with 4% PFA for 20 min at room temperature, then followed by 0.5% NP40 permeabilization and 10% donkey serum blocking. Cells were incubated with antibodies at 4°C overnight, and then incubated with fluorescence-conjugated secondary antibodies. DAPI was used for nucleus staining. Cells were visualized by laser scanning confocal microscopy (Olympus FV1000, Olympus, Japan).
Recombinant protein preparation. cDNAs were cloned from a BM cDNA library. CCP2 was subcloned into H-MBP-3c for MBP-tagged protein expression vectors. IL-7Rα was cloned into pGEX-6p-1 plasmid for GST-tagged protein expression. Plasmids were transformed into E. coli strain BL21 (DE3), followed by induction with 0.2 mM isopropyl-β-D-thiogalactoside (IPTG) at 16°C for 24 h. Cells were collected and lysed by supersonic, followed by purification through Amylose or GST resins.
EMSA assay. EMSA experiments were conducted according to the manufacturer's protocol with a Light Shift Chemiluminescent RNA EMSA Kit (Thermo Scientific). Briefly, Flag-STAT5 was incubated with or without unlabeled probe for competitive reaction and anti-STAT5 antibody for super shift at room temperature for 20 min in a reaction buffer. Then, Biotin-labeled probe was added into the reaction system and incubated for 20 min at room temperature. Samples were carried out in 4% polyacrylamide gel in 0.5 × TBE buffer. After transferred on a nylon membrane (Amersham Biosciences), the labeled DNA was cross-linked by ultraviolet, probed with streptavidin-HRP conjugate and then incubated with the detection substrate. The probe sequence for Sall3 was: 5′-CGGAGCCTAAAGCTGTTGCTTCGTG-GAACTTAGA CTAGCGGGAGAATTCAGTGTG-3′.
DNase I accessibility assay. DNaseI digestion assay has been described previously 28 . In brief, Nuclei were purified from CHILPs according to the manufacturer's protocol with the Nuclei isolating Kit (Sigma-Aldrich). Then nuclei were resuspended with DNase I digestion buffer and treated with indicated units of DNase I (Sigma, USA) at 37°C for 5 min. In all, 2 × DNase I stop buffer (20 mM Tris Ph 8.0, 4 mM EDTA, 2 mM EGTA) was added to stop reactions. DNA was extracted and examined by qPCR.
In vitro glutamylation assay. Detailed protocol for in vitro glutamylation assay was described as previously described 28 . In brief, CCP2, TTLL4, and TTLL13 were transfected into 293 T cells for 48 h. Cells were harvested and lysed. Supernatants were incubated with GST-IL-7Rα at 37°C for 2 h. GST-IL-7Rα was precipitated and tested for glutamylation with GT335 antibody.
Statistical analysis. An unpaired Student's t-test was used as statistical analysis in this study. Statistical calculation was performed by using Microsoft Excel or SPSS 13.
Data availability. All data generated or analyzed during this study are included in this published article and its Supplementary Information Files. Microarray data, are deposited in the Genebank as GSE97487.