Dynamic chromatin association of IκBα is regulated by acetylation and cleavage of histone H4

IκBs exert principal functions as cytoplasmic inhibitors of NF‐kB transcription factors. Additional roles for IκB homologues have been described, including chromatin association and transcriptional regulation. Phosphorylated and SUMOylated IκBα (pS‐IκBα) binds to histones H2A and H4 in the stem cell and progenitor cell compartment of skin and intestine, but the mechanisms controlling its recruitment to chromatin are largely unknown. Here, we show that serine 32–36 phosphorylation of IκBα favors its binding to nucleosomes and demonstrate that p‐IκBα association with H4 depends on the acetylation of specific H4 lysine residues. The N‐terminal tail of H4 is removed during intestinal cell differentiation by proteolytic cleavage by trypsin or chymotrypsin at residues 17–19, which reduces p‐IκBα binding. Inhibition of trypsin and chymotrypsin activity in HT29 cells increases p‐IκBα chromatin binding but, paradoxically, impaired goblet cell differentiation, comparable to IκBα deletion. Taken together, our results indicate that dynamic binding of IκBα to chromatin is a requirement for intestinal cell differentiation and provide a molecular basis for the understanding of the restricted nuclear distribution of p‐IκBα in specific stem cell compartments.

Nuclear IκBα preferentially binds the acetylated N‐terminal tail of histone H4 in vivo, specifically in the skin and intestine stem cell compartments. N‐terminal cleavage of histone H4 facilitates IκBα dissociation and cellular differentiation.

Introduction

Cytoplasmic IκB proteins play an essential role as negative regulators of the NF‐κB pathway, which controls immune responses and inflammation in hydra up to mammals (Hoffmann & Akira, 2013). Canonical activation of NF‐κB is initiated by stimuli such as TNFα, IL1β, or LPS and requires IκB kinase (IKK)‐induced phosphorylation and subsequent ubiquitination‐mediated degradation of the IκB inhibitors. Alternative nuclear functions for various IκB and IκB‐like family members including IκBα (Arenzana‐Seisdedos et al, 1997; Mulero et al, 2013; Marruecos et al, 2020), IκBβ (Rao et al, 2010), IκBNS (Hirotani et al, 2005), or Bcl3 (Bours et al, 1993) have been reported. The molecular mechanisms that dictate these moonlighting nuclear functions remain understudied.

Previous studies from different groups (Desterro et al, 1998; Culver et al, 2010; Mulero et al, 2013; Hendriks et al, 2014; Marruecos et al, 2020) demonstrated that a fraction of the IκBα protein is SUMOylated at the same K21 residue that, when ubiquitinated, triggers IκBα degradation. Thus, SUMOylated‐IκBα is primarily protected from degradation even in its phosphorylated form and independent of stimuli such as TNFα (Mulero et al, 2013). Phosphorylated and SUMOylated IκBα (pS‐IκBα) is localized in the nucleus of basal layer keratinocytes (Mulero et al, 2013) and in the intestinal crypt compartment (Marruecos et al, 2020), where stem and progenitor cells of both tissues reside, to regulate cytokine‐dependent activation of a subset of polycomb repression complex (PRC) 2 target genes (Mulero et al, 2013; Marruecos et al, 2020). SUMOylated IκBα shows reduced association with NF‐kB factors (Culver et al, 2010; Mulero et al, 2013), but efficiently binds the N‐terminal tail of histones, in particular H2A and H4. In fact, binding of IκBα to histones precludes its subsequent association with NF‐kB factors (Mulero et al, 2013).

Histones are an essential structural and regulatory element of the chromatin and they are organized in a highly stable structure called nucleosome. Nucleosomes are formed by several subunits of histone H2A, H2B, H3, and H4 tightly associated with DNA. Importantly, the N‐terminal tail of histones is flexible and protrudes outward the nucleosome, which makes histones accessible to editing enzymes that decorate their tails with a high variety of post‐translational modifications (PTMs). PTMs can be dynamically added and removed thus modulating gene activation, silencing, chromatin accessibility, replication, and DNA repair, in part by association with a plethora of non‐histone proteins including transcription factors. Prototypical examples of proteins that specifically recognize PTMs in histones are the bromodomains (BRDs) and the BET (bromodomains and extra‐terminal) family of proteins. BETs bind specific (Ac) K residues of histones in active regulatory domains such as promoters and enhancers (Filippakopoulos et al, 2012).

Here, we have studied the molecular requirements for pS‐IκBα binding to chromatin, the residues important for the pS‐IκBα‐histone H4 interaction, and the mechanisms that define restricted pS‐IκBα distribution in the stem cell compartments of the skin and the intestine.

Results and Discussion

Hydrophobic interactions mediate p‐IκBα binding to histone H4

Histones are basic proteins, and their positive charges determine association with the DNA. We investigated the possibility that electrostatic forces exert a major role in the observed association of IκBα with histones. We found that full‐length (FL) and two different amino‐terminal IκBα fragments (aa1–200; aa20–206) robustly interacted with histone H4 even at high salt concentrations (300 mM NaCl) (Fig EV1A), indicating the involvement of hydrophobic forces in histones‐IκBα binding. We aimed to confirm the stability of the interaction by Fast Protein Liquid Chromatography (FPLC) analysis of IκBα and histone H4 complexes. However, histone H4 was highly degraded in this assay and was replaced by H2A (previously found to similarly bind IκBα). Chromatography analysis demonstrated that both IκBα and H2A eluted at fraction 15–17 when run together, whereas IκBα or H2A alone eluted at fractions 29–30 and 18–20, respectively. Importantly, IκBα was never detected in fractions below 28 in two independent experiments performed. This shift strongly suggested the existence of stable complexes that were composed of equimolar amounts of IκBα and H2A proteins as determined by Coomassie Blue staining of the eluted fractions (Fig EV1B). Then, we performed in vitro phosphorylation of IκBα and SUMOylated IκBα (Fig EV1C) and tested the capacity of different IκBα species to bind to reconstituted nucleosome core particles (NCP) (instead of individual histones). We observed a significantly higher capacity of in vitro phosphorylated IκBα to form stable NCP‐IκBα complexes compared with non‐phosphorylated IκBα (Fig 1A). Of note that band retardation imposed by p‐IκBα and pS‐IκBα was not significantly different likely due to loss of the SUMO chain during the experimental procedure. As an additional control, NCP‐IκBα complexes were prevented or displaced by the addition of anti‐IκBα in the binding reaction (Fig 1B). By pull‐down (PD) assays, we confirm that the phosphorylation‐deficient IκBα_(S32‐36A) mutant shows reduced histone binding capacity when compared with IκBα wild type (WT) or the phosphor‐mimetic mutant IκBα_(S32‐36E) (Fig EV1D).

Figure EV1. In vitro interactions and phosphorylation of IκBα.

1. PD experiments under different salt concentrations using GST‐H4 as bait and the indicated IκBα constructs expressed in HEK‐293T cells.
2. Coomassie staining analysis of the indicated fractions recovered in the Fast Protein Liquid Chromatography (FPLC) analysis of IκBα and histone H2A complexes.
3. Western blot analysis of IκBα phosphorylated in vitro by addition of active IKKβ kinase with anti‐p‐IκBα _(S32‐36) antibody.
4. PD experiment using GST‐H4 as bait and the indicated IκBα mutants. Quantification of the interaction from 3 biological replicates performed. Bars represent mean values ± standard deviation of the replicates. Statistical analysis of different was obtained by t‐test; *P < 0.05, **P < 0.01.

Figure 1. Specific interaction of nucleosome particles with phosphorylated IκBα.

1. Electrophoretic analysis (under non‐denaturing conditions) of the association between the indicated IκBα species association and reconstituted nucleosome core particles (NCP).
2. Electrophoretic analysis in agarose gels of in vitro generated p‐IκBα and reconstituted nucleosome in absence or presence of anti‐IκBα antibodies.

Our results indicate that amino acids 1–200 of IκBα are required for H4 binding, which is independent of electrostatic forces and favored by IκBα phosphorylation.

Preferential association of IκBα with acetylated histone H4 in vivo

We previously demonstrated that pS‐IκBα binds the N‐terminal tail of histones H2A and H4 and identified acetylated K12 and K16 of H4 as preferential motifs for pS‐IκBα binding in vitro (Mulero et al, 2013). To further investigate pS‐IκBα binding specificity, we generated biotinylated peptides of the N‐terminal H4 sequence (aa1–23) including specific PTM combinations and used them to pull‐down (PD) IκBα from cell extracts (Fig 2A and B). Both pS‐IκBα (≈60 kDa) and non‐SUMOylated p‐IκBα (≈37 kDa) bands were consistently detected in the precipitates from all but the hyperacetylated H4 tail and the K > A mutant. Comparable inhibition of IκBα binding was detected in peptides containing methylated K12 and K20 (Fig 2B). Binding affinity was confirmed in PD assays using GST‐SUMO‐IκBα as bait and serial dilutions of histone‐enriched cell extracts (see methods) with the highest signal corresponding to the H4K16Ac and H4K20Ac marks and the lowest to H4K20me2,3 (Fig 2C). These results indicate that hyperacetylation or methylation of K residues reduces IκBα to histone H4 binding, whereas K12, K16, and K20 acetylated histone binds to IκBα similar to the non‐acetylated histone.

Figure 2. Preferential association of IκBα with N‐terminal acetylated histone H4.

1. PD experiment with biotinylated peptides of H4 (1–23aa) and total lysates of HCT‐116 cells.
2. Quantification of the IκBα 60 kDa band (upper panels) and the 37 kDa band (lower panels) is shown as average and s.d. of three independent experiments.
3. PD experiments with GST‐SUMO IκBα fusion protein and serial dilutions of histone‐enriched HEK‐293T lysates. Quantification of three independent replicates is shown.
4. Venn diagrams representing the overlap between IκBα and acetylated H4 peaks obtained in ChIP‐seq experiments from HCT‐116 cells.
5. Peak distribution of acetylated H4, H4K12ac and total histone H4 obtained in IκBα target genes and non‐IκBα targets relative to the transcription start site (TSS).

Data information: WB in A and C are representative of three biological replicates performed. In B, P values were derived from unpaired two‐tailed t‐test of triplicates, ***P < 0.001, **P < 0.01, *P < 0.05, n.s. no significant.

Source data are available online for this figure.

To study whether H4 acetylation favors IκBα binding to chromatin in vivo, we performed chromatin immunoprecipitation (ChIP) assay from HCT‐116 colorectal cancer cells using the antibody against IκBα (IκBα, sc‐307) and 2 antibodies against H4KAc (pan‐H4KAc, Abcam ab177790 and H4K12ac, Active Motif 61527). Analysis of data demonstrated significant enrichment of IκBα peaks in acetylated chromatin regions (P < 0.01) compared with random distribution of peaks from both precipitations (Fig 2D). Moreover, we detected higher accumulation of H4KAc around the TSS of IκBα‐bound genes compared with non‐selected gene promoters (labeled as All genes) (Fig 2E, upper panels) as well as a significant overlap between IκBα and H4KAc peaks at specific genomic loci (Fig 2E, lower panels). As expected, total H4 was distributed all along the genome with no enrichment at the TSS of IκBα targets or random genes.

These results indicate that IκBα similarly binds acetylated histone H4 in vitro, which is precluded by K methylation, and preferentially binds acetylated H4 in vivo.

Acetylated H4 species are restricted to stem cell compartments and lost in differentiated cells associated with histone N‐tail cleavage

Previous studies from our group demonstrated that nuclear pS‐IκBα is localized in the stem cell compartments of skin and intestine (Mulero et al, 2013; Marruecos et al, 2020). Our results suggest that pS‐IκBα preferentially bind acetylated histone H4 in vivo (see Fig 2D and E). Thus, we studied the possibility that nuclear IκBα was restricted to areas containing specific H4KAc marks. Different antibodies against acetylated histone H4 labeled cells localized in the intestinal crypt compartment colocalizing with p‐IκBα (Fig 3A) and including the canonical Lgr5+ ISCs (Fig 3B). Similarly, H4K12Ac was primarily detected in the keratinocytes of the basal layer of skin and the hair follicles, where progenitors and stem cells reside (Fig EV2A). In contrast, H4K20me2,3, which showed the lower affinity for IκBα binding (see Fig 1A–C), was exclusively present in differentiated cells of the mouse intestinal villi and skin (Figs 3A and EV2A). Comparable distribution of acetylated and methylated histone H4 species was observed in human colonic tissue (Fig EV2B). Of note, the few intestinal crypt cells that contained H4K20me2,3 mark were identified as terminally differentiated Paneth cells based on their morphology and localization (Fig 3A and B). Compartmentalization of H4KAc and H4K20me marks was progressively reached during embryonic development (Fig EV2C), which parallels the progressive restriction of nuclear IκBα distribution in the developing intestinal tissue (Marruecos et al, 2020).

Figure 3. Stem cell compartments contain acetylated H4 species that are lost following histone cleavage.

1. Immunofluorescence (IF) analysis with the indicated antibodies in sections from murine small intestine of 2‐month‐old WT mice.
2. Double IF analysis with the indicated antibodies in sections from murine small intestine of 2‐month‐old Lgr5‐GFP transgenic mice.
3. Western blot analysis of chromatin extracts from isolated intestinal villus and crypt cells.
4. Western blot analysis of chromatin extracts from isolated intestinal villus and crypt cells in several 2‐month‐old WT mice (each # represents a mouse).
5. Western blot analysis of soluble (Sol) and chromatin (Chr) extracts from isolated intestinal villus and crypt cells in 2‐month‐old mCherry‐H4 transgenic mice.
6. Western blot analysis of chromatin fraction from mCherry‐H4 transfected HCT‐116 cells.
7. Pull‐down assay using GST‐SUMO IκBα protein and chromatin lysates from 2‐month‐old WT mice villi.

Data information: Scale bars in A and B, 25 μm. Red arrowheads in C‐G are showing the location of truncated histone H4.

Source data are available online for this figure.

Figure EV2. IκBα is localized in stem cell compartments.

1. IF analysis with the indicated antibodies in sections from skin of 2‐month‐old mice.
2. IF analysis of human colonic tissue with the indicated antibodies.
3. IF analysis of sections from small intestine of mice at different stages of development.

Data information: The proximo‐distal axis is indicated in B. Scale bars, 25 μm.

Specific distribution of acetylated histone H4 restricted to the intestinal crypts was confirmed by Western blot of crypts and villus‐enriched fractions (Fig 3C). Remarkably, WB analysis of total histone H4, H4K20me2,3 and H4K20KAc revealed the presence of a low molecular weight (LMW) band (red arrowheads in Fig 3C) specifically in the villus‐enriched fractions. This result was confirmed in independent villus‐ and crypt‐enriched extracts obtained from several mice (Fig 3D). The fact that the LMW band that was not recognized by the H4K5, K8, K12, and K16Ac antibodies (see Fig 3C) strongly suggested that the cleavage site of H4 mapped between K16 and K20 residues comprising the sequence KRHRK. Using a transgenic mouse line carrying mCherry fused to H4 (see Materials and Methods), we confirmed that histone cleavage occurs in the intestinal villi in vivo and involved the N‐terminal tail of histone H4 leading to the release of mCherry from the H4 protein (Fig 3E). Comparable results were obtained by transfection of this construct in HCT‐116 CRC cells (Fig 3F). PD experiments using chromatin extracts from intestinal villus demonstrated that IκBα specifically bound full‐length histone H4 but failed to associate with truncated histone H4 lacking the N‐terminal tail (Fig 3G).

Our results indicated that intestinal cell differentiation correlates with N‐terminal truncation of histone H4, resulting in a cleaved histone H4 that is unable to bind IκBα. Notably, nuclear IκBα preferentially binds acetylated H4 species that are restricted to the stem cell compartment of the intestine and skin where nuclear p‐IκBα is also localized.

Cleavage of the N‐terminal tail of H4 is mediated by chymotrypsin and trypsin present in the intestinal villus

To address the possibility that histone H4 cleavage was due to the activity of a protease specifically contained in the differentiated intestinal cells, we incubated crypt‐derived chromatin extracts (containing intact histone H4) with soluble lysates from either villus‐derived or crypt‐derived cells. Cleavage of histone H4 was specifically induced by incubation with villus‐derived lysates (Fig 4A) further indicating that one or more proteases present and active in the differentiated cells induce histone H4 cleavage. Then, we obtained villus‐ and crypt‐enriched cell fractions (by cell sorting based on EPHB2 levels), purified the RNA, and performed RNA‐seq analysis. We identified several proteases that were specifically expressed in the villus‐derived RNA (Fig 4B). Bioinformatic analysis using the PeptideCutter‐ExPASy (Wilkins et al, 1999) and PROSPER (Song et al, 2012) tools identified the serine proteases trypsin and chymotrypsin C as the most likely enzymes capable to cleave histone H4 in the region involving residues K16 to K20, which we have experimentally identified as the sites for cleavage. By point mutation of specific H4 residues in mCherry‐H4 fusion protein, we recognized amino acids 17–19 as essential for H4 cleavage (Fig 4C), which is consistent with trypsin or chymotrypsin as responsible for this activity. However, we still detected some cleavage of the mutant carrying all K residues mutated to A, indicating the presence of non‐specific cleavage of H4 tail.

Figure 4. Cleavage of the N‐terminal tail of H4 is mediated by chymotrypsin and trypsin activity present in the intestinal villus.

• A
  Western blot analysis of the crypt‐derived chromatin fractions incubated with soluble lysates from villi or crypts for the indicated times.
• B
  Table showing the proteases differentially expressed in the villus compartment as determined by RNA‐seq analysis of intestinal crypt (purified EphB2 high cells) and villus (EphB2 negative/low) cells. Predicted cleavage sites for each protease in the histone H4 sequence were determined using the PeptideCutter‐ExPASy and PROSPER software.
• C
  Western blot analysis of mCherry‐H4 with the indicated mutations transfected in HCT‐116 cells. Quantification of relative amounts of truncated H4 relative to the WT.
• D, E
  Western blot analysis of mCherry‐H4 transfected in HCT‐116 cells treated for 16 hours with TLCK (D) or AdaAhx₃L₃VS (E) at the indicated concentrations.
• F
  Western blot analysis of a cleavage experiment incubating modified H4 peptides (P1‐P9) with soluble lysates from villi or crypts in the presence (+) or absence (−) of the commercial protease inhibitors cocktail. Lanes indicated as P correspond to the control peptide without lysate incubation. Notice the electrophoretic shift of peptides incubated with villus lysates (compared with control peptides) suggestive of post‐translational modifications or binding to proteins absent from crypt lysates.
• G
  Quantification of the relative cleavage of specific H4 peptides incubated with villus‐derived lysates from 3 independent experiments performed.

Data information: Red arrowheads in A, C–F indicate truncated histone H4.

Source data are available online for this figure.

To further test whether trypsin or chymotrypsin execute histone H4 cleavage, we treated mCherry‐H4 expressing HCT‐116 cells with the specific inhibitors of trypsin‐like proteases TLCK and AdaAhx₃L₃VS for 16 h. WB analysis indicated that TLCK (Fig 4D) or AdaAhx₃L₃VS (Fig 4E) imposed a dose‐dependent inhibition of mCherry‐H4. Treatment with protease inhibitors also reduced endogenous histone H4 cleavage in post‐confluent cells and increased chromatin retention of p‐IκBα (≈37 kDa) and pS‐IκBα (≈60 kDa) as determined by WB of chromatin extracts (Fig EV3A).

Figure EV3. Association between chromatin‐bound IκBα and histone H4 cleavage.

1. WB analysis of chromatin extracts from HT29 cells treated as indicated. The red arrowhead indicates truncated histone H4.
2. WB analysis of chromatin extracts from IκBα KO HT29 cells carrying doxycycline‐inducible IκBα expression construct untreated or treated with doxycycline for 16 h.

It is established that trypsin and trypsin‐like proteases digest substrates containing unmodified K residues but this activity can be affected by K acetylation due to both steric effects and loss of the positive charge (Huang et al, 2015). We used different histone H4 peptides (aa1–23) to determine whether protease activity from villus‐derived extracts was modified by specific K modifications. Our data indicated that villus extracts (but not with crypt‐derived extracts) induced the degradation of histone H4 peptides with the only exception of the K to A mutant and the peptide with all K residues acetylated (K5,8,12,16Ac) (Fig 4F and G).

This result, together with the observation that crypt‐derived histone H4 is cleaved after incubation with soluble villus lysates, indicated that histone H4 in the crypts is not hyperacetylated but different H4Ac species (i.e., H4K5Ac, H4K8Ac, H4K12Ac) may coexist in this cellular compartment.

We considered the possibility that IκBα association with specific histone H4Ac species may prevent (by competing protease binding) or facilitate (by recruitment of specific protease activity) their cleavage. We investigated this possibility using a model of IκBα KO cells (generated by CRISPR‐Cas9) carrying a doxycycline‐inducible IκBα expression vector, we found that histone cleavage at post‐confluence was not affected by IκBα expression and chromatin binding (Fig EV3B). However, we cannot exclude that binding of IκBα to chromatin modifies histone H4 cleavage at specific genomic regions or genes, which should be further investigated. Adding further complexity to the system, we here showed that histone H3 is also cleaved during intestinal cell differentiation of HT29 cells, which is in agreement with a recent publication on mouse intestine (Ferrari et al, 2021).

Histone cleavage and their functional impact represent an extremely difficult field of study (reviewed in Dhaenens et al, 2015; Azad et al, 2018). In general, cleavage of histone tails may result in the localized clearing of multiple repressive signals, other than IκBα, during (or leading to) the induction of gene expression. Moreover, not only binding to activators or repressors but altered nucleosome structure and dynamics have been found associated with histone tail cleavage (Nurse et al, 2013), thus highlighting the extreme complexity of this regulation. N‐terminal cleavage of histones, mainly H3, was previously reported in other systems associated with organism development and cell differentiation (Duncan et al, 2008; Vossaert et al, 2014; Kim et al, 2016; Melo et al, 2017).

We propose that N‐terminal loss at specific histones might represent both a general and a specific mechanism of gene regulation during development and tissue differentiation, imposed by levels and activity of proteases recognizing specific histones and histone codes. IκBα dissociation from particular gene promoters (i.e., during cell differentiation) as a result of histone H4 cleavage may represent an additional element contributing to gene regulation at specific gene loci, which could be, in turn, controlled by acetylation or other PTMs. This dynamics of IκBα binding and dissociation would then trigger a switch in the repertoire of activators and repressors bound at specific gene promoters leading to a whole transcriptional response. Whether or not histone H4 cleavage takes place at specific gene sets (i.e., IκBα target genes) at particular differentiation stages, and it is linked to specific histone acetylation codes is a relevant issue that we will further investigate.

Functional impact of IκBα chromatin binding dynamics on cell differentiation

To further investigate whether histone H4 acetylation and cleavage correlates with the dynamics of IκBα binding to chromatin, we obtained chromatin extracts from HT29 CRC cells, which differentiate into the Goblet cell lineage at confluence. By WB analysis, we found IκBα and p‐IκBα chromatin binding parallels H4K12 acetylation, reaching a maximum around days 3–5 of post‐confluence, and then being released at days 6–7. Histone H4 cleavage started early at post‐confluence and reached a maximum at day 5, as detected with the H4K20me2,3 antibody. We also detected some levels of histone H3 cleavage associated with cell differentiation (Fig 5A).

Figure 5. Functional impact of IκBα chromatin binding dynamics in cell differentiation.

• A
  Western blot analysis of soluble and chromatin extracts from HT29 cells at different differentiation stages (days post‐confluence).
• B, C
  qPCR (B) and WB analysis (C) of the indicated genes or proteins in parental or IκBα KO HT29 cells obtained at pre‐confluence or 7‐day post‐confluence. Bars represent mean values ± standard deviation of 3 technical replicates from 3 biological replicates performed.
• D, E
  qPCR (D) and WB analysis (E) of MUC5AC in HT29 cells treated with the indicated protease inhibitors. Bars represent mean values ± standard deviation from 3 biological replicates performed.
• F, G
  qPCR (F) and WB analysis (G) of MUC5AC in HT29 cells transduced with sh‐RNAs against trypsin (T1, T2) or chymotrypsin (C1, C2). Bars represent mean values ± standard deviation of 3 technical replicates from 3 biological replicates performed.
• H
  Western blot analysis of cells transfected with mCherry‐H4 and the indicated sh‐RNAs.
• I
  Model for gene regulation by chromatin‐associated IκBα. In brief, dynamic dissociation of IκBα from the chromatin at specific genetic loci promotes transcription of genes involved in stem cell maturation or differentiation (upper panel). Absence of dynamic binding (middle panel) or cells IκBα deficient (lower panels) fail to activate IκBα‐dependent transcription thus imposing a differentiation/maturation blockage.

Data information: Images in A–F are representative of three independent replicates performed. P values were derived from unpaired two‐tailed t‐test, ***P < 0.001, **P < 0.01, *P < 0.05, n.s. no significant.

Source data are available online for this figure.

IκBα deficiency results in altered stem cell maturation and defective intestinal and skin differentiation in mice (Mulero et al, 2013; Marruecos et al, 2020). We used HT29 cells to further investigate the impact of IκBα association to chromatin in intestinal differentiation. We found that IκBα deletion (KO) precluded goblet cell differentiation of human HT29 CRC cells at 7 days of post‐confluence as determined by qPCR (Figs 5B and EV4A) and WB analysis (Fig 5C) of the terminal differentiation markers MUC5AC and SPDEF.

Figure EV4. Inhibition of goblet cell differentiation by trypsin or chymotrypsin knock‐down.

• A, B
  Analysis by qPCR of SPDEF in HT29 cells WT or IκBα KO (A) or transduced with specific shRNA against trypsin (T) or chymotrypsin (C) at pre‐confluence or 7 days after confluence (B). Bars represent mean values ± standard deviation from 3 biological replicates performed.
• C
  qPCR analysis of the muco‐secretory differentiation marker MUC2 in mouse intestinal organoids treated as indicated. Bars represent mean values ± standard deviation from 3 biological replicates performed.
• D
  qPCR analysis of the stem cell marker Lgr5 in HT29 cells treated with the indicated inhibitors. Bars represent mean values ± standard deviation from 3 biological replicates performed.

Statistical differences were determined by t‐test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

We then tested whether inhibition of histone cleavage, which favors pS‐IκBα chromatin retention (see Fig EV3A), affected HT29 differentiation. Treatment with the trypsin‐like inhibitors TLCK or AdaAhx₃L₃VS (Fig 5D and E) or knocking down trypsin or chymotrypsin with different shRNA (Figs 5F and G, and EV4B) both precluded goblet cell differentiation of HT29 cells, comparable to the effects imposed by IκBα deletion. Inhibition of mucin expression by AdaAhx₃L₃VS was also observed in the murine intestinal organoid model (Fig EV4C). Notably, the degree of differentiation inhibition imposed by specific shRNA against trypsin and chymotrypsin mainly parallels their inhibitory effects on histone H4 cleavage (Fig 5H). These results suggest that histone cleavage, likely by inducing local dissociation of IκBα from the chromatin, is required for intestinal cell differentiation. Interestingly, we found that both AdaAhx₃L₃VS treatment and IκBα KO inhibited expression of the canonical ISC marker Lgr5 in pre‐confluent HT29 cells (Fig EV4D), suggestive of defective stem cell maturation, which is similar to that observed in the mouse organoids (Marruecos et al, 2020). Because of the significant impact of nuclear IκBα in cell differentiation and stem cell maturation, and our previous data obtained in squamous cell carcinoma (Mulero et al, 2013), we speculate that IκBα chromatin binding and histone cleavage could similarly play a relevant contribution to tumorigenesis and tumor metastasis, which mainly depend on the stem cell‐like capacity of tumor cells.

Together our results support a model in which dynamic binding/dissociation of IκBα to/from chromatin regulates specific gene transcription, which is essential for stem cell maturation and subsequent progression toward tissue‐specific mature lineages (Mulero et al, 2013; Brena et al, 2020; Marruecos et al, 2020 and this work). Thus, in the absence of chromatin‐bound IκBα or in case of irreversible IκBα chromatin binding (i.e., after trypsin or chymotrypsin inhibition), tissue stem cells and/or progenitors are retained into an immature state and fail to express differentiation markers such as epidermal Filaggrin and Keratin 10 (Mulero et al, 2013) or intestinal Muc2, Lyz1 (Marruecos et al, 2020), and Muc5 (this work) (see model in Fig 5I). Our prediction is that IκBα may represent a unique cellular tool for the integration of pro‐inflammatory signals provided by the tumor stroma into a transcriptional program that impacts in cellular stemness. This IκBα function that facilitates the maintenance of stem cell homeostasis in response to inflammation or damage, would also execute pro‐tumorigenic and pro‐metastatic activities linked to inflammatory tumor microenvironments thus counteracting the positive effect of the immune system.

Materials and Methods

Animal studies

C57B6 mice and Lgr5^GFP‐CreERT mice were from The Jackson Laboratories. In all procedures, animals were kept under pathogen‐free conditions, and animal work was conducted according to the guidelines from the Animal Care Committee at the Generalitat de Catalunya. The Committee for Animal Experimentation at the Institute of Biomedical Research of Bellvitge (Barcelona) approved these studies.

Transgenic mice carrying Cherry‐histone H4 were generated by pronuclear injection of B6CBA zygotes with 30 ng/μl hyPBase mRNA+ 30 ng/μl PB_CAG_Cherry‐histoneH4 plasmid and transfer into pseudo‐pregnant 0.5 dpc CD1 females. Positive pups were screened for Cherry fluorescence presence in blood and expanded thereafter.

Cell lines and reagents

All cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) [Invitrogen] supplemented with 10% fetal bovine serum (FBS) [Biological Industries]. Cells were grown in an incubator at 37°C and 5% CO₂. Cells used in these studies were HEK‐293T [ATCC Ref. CRL‐3216], HT29 [ATCC Ref. HTB‐38D], and HCT‐116 [ATCC Ref. CCL‐247]. Reagents used are the following: TLCK (Tosyl‐L‐lysyl‐chloromethane hydrochloride) [Abcam ab144542] and AdaAhx3L3VS [Sigma‐Aldrich 114802]. HT29 IκBα KO cells were generated by CRISPR‐Cas9 using guides targeting exon 1.

Villus/crypt‐enriched fractionation

Intestine of WT mice was extracted. Villus was separated from crypts mechanically, as described previously (Marruecos et al, 2020). Then, crypts were purified by mechanical disaggregation and filtration in a 70 µm cell strainer.

Cell transfection

We used Polyethylenimine (PEI) [Polysciences Inc. Ref. 23996] as a carrier vector following standard methods. In brief, we diluted 4 μl PEI per μg of DNA in serum‐free DMEM and incubated 5 min at room temperature. Then, we added the DNA and incubated the mix 20 min at RT. Finally, we incorporated the PEI/DNA solution to the cell cultures.

Pull‐down and peptide immunoprecipitation (IP) assays

PD assays were performed as previously described (Espinosa et al, 2003). Briefly, GST fusion proteins were incubated with lysates for 45 min in a rotary shaker at 4°C. When indicated, nuclear extracts were boiled at 98°C for 5 min in the presence of 1% SDS to disassemble pre‐existing protein complexes and then neutralized in 1% Triton X‐100. Precipitates were resolved in SDS–PAGE and analyzed by IB. For peptide IP, histone H4 peptides [Synpeptide CO LTD] were synthesized as biotinylated N‐terminal and C‐terminal amides. Peptides were incubated overnight at 4°C with the indicated cell extracts and precipitated with streptavidin–sepharose beads for 45 min.

Cell fractionation and Western blot (WB)

For soluble and chromatin separations, cells were lysed 1 mM EDTA, 0.1 mM Na‐orthovanadate (Na₃VO₄), 0.5% Triton X‐100, 20 mM β‐glycerol‐phosphate, 0.2 mM PMSF, protease inhibitor cocktail, in PBS for 20 min on ice and centrifuged at 18,000 g. Supernatants were recovered as the soluble fraction, and the pellets were lysed in Laemmli buffer (1× SDS–PAGE buffer plus β‐mercaptoethanol (BME) [Sigma, Ref. M‐3148]) or in 1%SDS PBS, sonicated, and treated with 1% Triton X‐100. For histone‐enriched fractions, cells were lysed in 10 mM HEPES pH7.9, 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM DTT, 1.5 mM PMSF, 100 mM sodium butyrate, protease inhibitor cocktail and 0.2 mM H₂SO₄ for 30 min on ice and centrifuged at 10,080 g. Then, supernatants were dialyzed against PBS. Lysates were analyzed by Western blotting using standard SDS–polyacrylamide gel electrophoresis (SDS–PAGE) techniques. In brief, protein samples were boiled in Laemmli buffer, run in polyacrylamide gels, and transferred onto polyvinylidene–difluoride (PVDF) membranes [Millipore Ref. IPVH00010]. Membranes were incubated overnight at 4°C with the appropriate primary antibodies, extensively washed, and then incubated with specific secondary horseradish peroxidase‐linked antibodies from Dako [Ref. P0260 and P0448]. Peroxidase activity was visualized using the enhanced chemiluminescence reagent [Biological Industries Ref. 20‐500‐120] and autoradiography films [GE Healthcare Ref. 28906835]. Gels were stained with Coomassie (Brilliant Blue G‐250 [Sigma Ref.6104‐58‐1].

Antibodies

All antibodies used in this work are listed in Table 1.

Table 1.

List of antibodies used.

Antibody	Company	Reference	Specie	Dilution
Histone H3	Abcam	ab1791	Rabbit	1:5,000
Histone H4	Abcam	ab10158	Rabbit	1:1,000
Histone H2A	Abcam	ab18258	Rabbit	1:1,000
H4K12ac	Active Motif	61527	Mouse	1:1,000
H4K16ac	Active Motif	61529	Mouse	1:1,000
H4K20me2,3	Abcam	ab78517	Mouse	1:1,000
H4K5ac	Active Motif	39169	Rabbit	1:1,000
H4K8ac	Active Motif	61103	Rabbit	1:1,000
H4K20ac	Active Motif	61531	Mouse	1:1,000
IκBα	Abcam	ab32518	Rabbit	1:1,000
P‐ IκBα	Cell Signalling	#9246	Mouse	1:1,000
Tubulin‐α (clone B‐5‐1‐2)	Sigma‐Aldrich	T6074	Mouse	1:10,000
Lamin B (C‐20)	Santa Cruz	sc‐6216	Goat	1:1,000
Cherry	Abcam	ab167453	Rabbit	1:1,000
Muc5	Previously published	de Bolos et al (2001)	Rabbit	1:5,000
Polyclonal Goat anti‐Rabbit Immunoglobulins/HRP	Dako	P0448	Goat anti‐Rabbit	1:2,000
Polyclonal Rabbit anti‐Mouse Immunoglobulins/HRP	Dako	P0260	Rabbit anti‐Mouse	1:2,000
Polyclonal Rabbit anti‐Goat Immunoglobulins/HRP	Dako	P0449	Rabbit anti‐goat	1:2,000

Immunofluorescence (IF) analysis

Tissues were fixed in 4% formaldehyde overnight at room temperature and embedded in paraffin. 4 μm paraffin‐embedded sections were first deparaffinized in xylene. IHC was performed following standard techniques with EDTA‐ or citrate‐based antigen retrieval and developed with the Envision⁺ System HRP Labelled Polymer anti‐Rabbit [Dako Ref. K4003] or anti‐Mouse [Dako Ref. K4001] and developed with TSA^TM Plus Cyanine 3/Fluorescein System [PerkinElmer Ref. NEL753001KT] and mounted in ProLong™ Diamond Antifade Mountant plus DAPI [Thermo Scientific Ref. P36971]. Images were taken in an SP5 upright confocal microscope (Leica).

Size exclusion chromatography

GST‐H2A or His‐IκBα 1‐200 proteins were expressed in BL21 bacteria and lysed as explained above. Then, proteins were purified using glutathione‐Sepharose [Amersham Biosciences] or Nickel‐NTA [Qiagen] resins, respectively. After elution, proteins were buffered exchanged using desalting columns [GE Healthcare] to a buffer containing 25 mM Tris pH 7.4, 150 mM NaCl, 0.2 mM EDTA pH 8.0, 1 mM PMSF, and 1 mM DTT. Then, individual proteins (GST‐H2A and His‐IκBα 1–200) or a mix of both proteins at a ratio of 1:1 incubated at room temperature for 1 h to form complexes prior size exclusion chromatography were analyzed. 150 µl of each sample was loaded onto an analytical Superdex SD200 10/300 column [GE Healthcare] pre‐equilibrated with the sample buffer and resolved at a flow rate of 0.5 ml/min on an AKTA Purifier [GE Healthcare] automated liquid chromatography system. 0.5 ml size fractions were collected, and an equal volume aliquot of each fraction was analyzed by SDS–PAGE followed by Coomassie Blue staining.

Recombinant histone octamer formation

pET11a‐H2A, pET11a‐H2B, and pET11a‐H3‐H4 plasmids were a kind gift from Dr. James Kadonaga, UCSD. The original pET11a‐H2B plasmid was mutated to pET11a‐Ile‐H2B (inserting an Isoleucine residue before the H2B sequence) to increase its protein yield. All histones were expressed in Escherichia coliBL21(DE3) by growing cells to A600 0.5‐0.6 followed by induction with 0.2 mM isopropyl β‐D‐1‐thiogalactopyranoside (IPTG) for 3 h at 37°C. Cells were pelleted by centrifugation at 950 g for 20 min at 4°C and washed once with phosphate‐buffered saline (4 mM Na₂HPO₄, 1 mM KH₂PO₄, 137 mM NaCl, and 3 mM KCl) (100 ml/liter bacterial cell culture). The H2A‐H2B dimers and H3‐H4 tetramers were purified following the detailed protocol previously published (Levenstein & Kadonaga, 2002).

Histone octamer formation was performed following the protocols from Dyer et al (2004) with some modifications. Briefly, histones molar ratio was quantified using A280 (NanoDrop) and applying an extinction coefficient for H2A/H2B,10.240 cm⁻¹ M⁻¹ and for H3/H4, 9080 cm^‐1 M^‐1, respectively. Mixed histones were unfolded with Unfolding Buffer (20 mM Tris–HCl pH 7.5, 10 mM DTT and 7 mM guanidinium hydrochloride) for 2 h at 4°C. Then, the sample was dialyzed against the Refolding Buffer (10 mM Tris–HCl pH 7.5, 1 mM EDTA, 5 mM BME and 2 M NaCl) at 4°C two times 3 h each and then ON. The day after, the sample was concentrated using a Centriprep 30‐kDa cutoff membrane concentrator unit [Millipore] at 4°C and injected in a gel filtration analytical Superdex SD200 column. Fractions containing the histone octamer were pooled and dialyzed against Refolding Buffer with 50% glycerol (10 mM Tris–HCl pH 7.5, 1 mM EDTA, 5 mM BME, 2 M NaCl and 50% (v/v) glycerol) at 4°C ON. Histone octamers were aliquoted and stored at −80°C.

Nucleosome reconstitution

Original 601‐DNA 12 copies plasmid was a kind gift from Dr. Karolyn Luger. Nucleosome core particles (NCP) were reconstituted following the classical salt gradient method. Briefly, we mixed 1 μg of DNA with different volumes of histone octamers in a buffer containing 10 mM Tris–HCl pH 8.0, 2 M NaCl and 1 mM EDTA. Samples were dialyzed using a 3.5 kDa cutoff Slide‐A‐Lyzer Mini dialysis device [Thermo Fischer Scientific]. We dialyzed the NaCl concentration from the initial 2 M to 1.5 M, 1 M, 800 mM, 600 mM in the same buffer changing the dialysis every 3 h at 4°C. Finally, we left the dialysis going overnight at 2.5 mM NaCl. The day after we changed the buffer one more time 3 h at 4°C in presence of 2.5 mM NaCl. Successful NCP reconstitution was confirmed by running a native electrophoretic mobility shift assay (EMSA) loading 10 ng of NCPs in a 4% native acrylamide gel (19:1) in the cold room. Before loading, samples were mixed with 10% (v/v) glycerol in absence of bromophenol blue to preserve the integrity of the NCPs.

Non‐radiolabeled electrophoretic mobility shift assay

Purified full‐length His‐IκBα protein was incubated with PKA in the kinase buffer 20 mM HEPES pH 7.9, 150 mM NaCl, 10 mM MgCl₂, 10 mM NaF, 0.2 mM sodium orthovanadate, and 20μM ATP for 30 minutes at 30°C. Immediately after, NCP were incubated with increasing concentrations of phosphorylated IκBα (p‐IκBα) for 1 h at 4°C. p‐IκBα was obtained by incubation of recombinant IκBα with active IKKß in the presence of ATP. When indicated, anti‐IκBα antibody was added to the mix to block the NCP‐IκBα interaction. After, NCP‐IκBα complexes were run in a 4% native acrylamide gel (19:1) in the cold room. Gels were stained using GelRed [Biotium] for 30 min and destained using TBE 0.5 × 3 times for 10 min in the cold room. Gels were imaged using a GelDoc system [Bio‐Rad].

Radiolabeled electrophoretic mobility assay

601‐DNA was radiolabeled with ³²P using T4‐polynucleotide kinase and [γ‐³²P] ATP for 1 h at 37°C and then incubated with the histone octamer in presence of 2 M NaCl. Here, NCP formation was achieved by sequential dilution instead of dialysis as described above. Radiolabeled NCPs were incubated with the proteins under study for 20 min at room temperature in binding buffer 10 mM Tris–HCl (pH 7.5), 50 mM NaCl, 10% (v/v) glycerol, 1% (v/v) NP‐40, 1 mM EDTA, and 0.1 mg/mL PolydIdC. Samples were run in TGE buffer (24.8 mM Tris base, 190 mM glycine, and 1 mM EDTA) at 200V for 1 h, and the gel was dried. Protein complexes were analyzed by native electrophoresis on a 4% (w/v) native acrylamide gel.

Knock‐down assays

MISSION shRNA for trypsin and chymotrypsin were obtained from Sigma. For lentiviral production, HEK293T cells were transfected with the lentiviral vectors of interest using standard protocols. One day after transfection, media was refreshed. Virus was collected 24 h later and then concentrated using Ultracentrifuge Optima™ XPN‐100 ‐ IVD (Biosafe) [Beckman Coulter]. Cells were infected and selected with puromycin.

Chromatin Immunoprecipitation (ChIP) and ChIP‐sequencing analysis

Human colon cell lines were subjected to chromatin immunoprecipitation (ChIP) as previously described (Aguilera et al, 2004; Mulero et al, 2013). Briefly, formaldehyde crosslinked cell extracts were sonicated, and chromatin fractions were incubated for 16 h with anti‐IκBα [abcam ab32518], anti‐P‐IκBα [Cell Signaling #9246], anti‐Acetylated H4 [Abcam ab177790], and anti‐H4K12ac [Active Motif 61527] antibodies in RIPA buffer and then precipitated with protein A/G‐sepharose [GE Healthcare, Refs. 17‐0618‐01 and 17‐0780‐01]. Crosslinkage was reversed and 6–10 ng of precipitated chromatin was directly sequenced in the genomics facility of Parc de Recerca Biomèdica de Barcelona (PRBB) using Illumina HiSeq platform.

Raw single‐end 50‐bp sequences were filtered by their quality (Q > 30) and length (length > 20‐bp) with the Trim Galore software (available at: http://www.bioinformatics.babraham.ac.uk/projects/download.html#trim_galore). Filtered sequences were aligned against the reference genome (mm10 release) with Bowtie2. MACEV2 software was run first for each replicate, and then by combining all replicates, using unique alignments (q‐value < 0.1). Broad peaks calling was set. Peak annotation was performed with ChIPseeker package and functional enrichment analysis with enrichR using the latest version of GO annotations. ChIP‐sequencing data are submitted to GEO database.

Cell sorting

Villus and crypt cells were obtained after mechanical disaggregation and 40 µm filtration. Cells were incubated with APC‐EphB2 antibody [BD Pharmingen Ref. 564699] for 20 min and stained with DAPI and the sorted in an Influx™ Sorter [BD Biosciences].

qRT–PCR analysis

Total RNA was extracted with the RNeasy Mini Kit [Qiagen Ref. 74004] and the RT‐First Strand cDNA Synthesis Kit [GE Healthcare Life Sciences Ref. 27‐ 9261‐01] and was used to produce cDNA. qRT–PCR was performed in LightCycler480 system using SYBR Green I Master Kit [Roche Ref. 04887352001]. The primers used are listed in Table 2.

Table 2.

Primers used in the qPRC analysis.

Target	Specie	Forward	Reverse
MUC5AC	Human	CTGGTGCTGAAGAGGGTCAT	CAACCCCTCCTACTGCTACG
SPDEF	Human	CTGTGGACAGAGCACCAATACC	GGTCGAGGCACAGTAGTGAATC

RNA‐sequencing experiments and data analysis

We extracted total RNA using RNeasy Micro Kit [Qiagen Ref. 74004]. The RNA concentration and integrity were determined using Agilent Bioanalyser [Agilent Technologies]. Libraries were prepared at the Genomics unit of PRBB (Barcelona, Spain) using standard protocols, and cDNA was sequenced using Illumina^® HiSeq platform, obtaining ˜25 to 30 million 50 bp single‐end reads per sample. Adapter sequences were trimmed with Trim Galore. Sequences were filtered by quality (Q > 30) and length (> 20 bp). Filtered reads were mapped against the latest release of the mouse reference genome (mm10) using default parameters of TopHat (v.2.1.1) (Kim et al, 2013), and expressed transcripts were then assembled. High‐quality alignments were fed to HTSeq (v.0.9.1) (Anders et al, 2015) to estimate the normalized counts of each expressed gene. Differentially expressed genes between different conditions were explored using DESeq2 R package (v.1.20.0) (Love et al, 2014). Plots were done in R. RNA‐sequencing data are deposited at the GEO database with accession number GSE131187.

Quantification and statistical analysis

Statistical parameters, including number of events quantified, standard deviation, and statistical significance, are reported in the figures and in the figure legends. Statistical analysis has been performed using GraphPad Prism6 software (GraphPad) and P < 0.05 is considered significant. Two‐sided Student’s t‐test was used to compare differences between two groups. Each experiment shown in the manuscript has been repeated at least twice.

Author contributions

AB and LE conceptualized the study, designed the experiments, and wrote the manuscript. LM performed experiments and wrote the manuscript. JB, DA‐V, MCM, MF, AV, LGP, SA‐G, IP, and LB performed biochemical assays and in vitro and in vivo experiments. JV‐F and GG analyzed data and interpreted results. YG analyzed ChIP‐sequencing and RNA‐sequencing data.

Conflict of interest

The authors declare that they have no conflict of interest.

Supporting information

Expanded View Figures PDF

Review Process File

Source Data for Figure 2

Source Data for Figure 3

Source Data for Figure 4

Source Data for Figure 5

EMBO reports (2021) 22: e52649.

Data availability

RNA‐seq and ChIP‐seq data are available at the GEO databases with the accession codes GSE131187 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE131187) and GSE167087 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE167087), respectively.