lncRNA LINC00941 modulates MTA2/NuRD occupancy to suppress premature human epidermal differentiation

Tissue culture

Pooled primary human keratinocytes from different juvenile donors were obtained from PromoCell. The tissue used by PromoCell for the isolation of human cells is derived from donors who have signed an informed consent form (by the donor, an authorized agent, or a legal agent) that outlines the purpose of the donation and the procedure for processing the tissue. PromoCell does not accept or use any tissue without prior signing of the consent document. The present analyses were conducted with pooled cell samples without personally identifiable information and deemed exempt from the requirement for IRB approval.

Keratinocytes were grown in a 1:1 mixture of KSF-M (Gibco) and Medium 154 for keratinocytes (Gibco). The media were supplemented with Human Keratinocyte Growth Supplement, bovine pituitary extract, human epidermal growth factor, and 0.5x antibiotic–antimycotic (all supplements were obtained from Gibco). Cells were cultured at 37°C in a humidified chamber with 5% CO₂. In vitro keratinocyte differentiation was induced by the addition of 1.2 mM calcium to the media and seeding cells at full confluency.

RNA knockdown

siRNA pools of 11–30 different siRNAs were synthesized and obtained from siTools (Hannus et al, 2014). For siRNA-mediated knockdown, 5–6 million primary human keratinocytes were electroporated with 1 nmol siPools using the Amaxa human keratinocyte Nucleofector kit (Lonza) according to the manufacturer’s instructions and the T-018 program of the Amaxa Nucleofector II device (Lonza). After nucleofection, cells recovered for 24 h.

Organotypic human epidermal tissue

For the generation of organotypic human epidermal tissues, 550,000 human keratinocytes nucleofected with siRNAs were seeded onto a devitalized dermal matrix and raised to the air–liquid interface to initiate stratification and differentiation, as described previously (Truong et al, 2006; Sen et al, 2010).

Immunofluorescence and tissue analysis

Seven-micrometer-thick paraffin-embedded cross sections of human organotypic skin cultures were deparaffinized and rehydrated by two changes of xylene and washed with descending ethanol concentrations. For antigen retrieval, the cross sections were boiled twice in demasking buffer (10 mM sodium citrate, pH 6.0, with 0.05% Tween-20) followed by blocking in PBS with 10% BCS for 20 min at RT. Antibodies were diluted in PBS with 1% BCS and incubated with the sections for 1 h. Primary antibodies included the following: filaggrin (sc-66192; Santa Cruz) at 1:50 dilution, MTA2 (ab8106; Abcam) at 1:500 dilution, and keratin 10 (MS-611-P; NeoMarkers) at 1:400 dilution. Alexa Fluor 555–conjugated goat anti-mouse (A21422, 1:300; Life Technologies) was used as a secondary antibody diluted in PBS with 1% BCS. Unbound antibodies were washed off after 1 h of incubation with PBS (three times, 5 min, RT). Slides were mounted in DAPI Fluoromount-G (Biozol). Finally, cross sections were analyzed with a BZ-X800 fluorescence microscope and BZ-X800 Analyzer software (Keyence).

qRT–PCR analysis

Total RNA from organotypic skin cultures was isolated with the RNeasy Plus mini kit (QIAGEN) according to the manufacturer’s instructions. Total RNA from keratinocytes was isolated with TRIzol (Invitrogen) as stated by the manufacturer. Quantified by NanoDrop, 1 μg total RNA was subjected to reverse transcription with the iScript cDNA synthesis kit (Bio-Rad). In case of RNA isolated with TRIzol, genomic DNA was removed using DNase I (Thermo Fisher Scientific) before cDNA synthesis. For qRT–PCR measurements, the Takyon Mix (Eurogentec) was used with CFX96 Touch Real-time PCR Detection System (Bio-Rad). Samples were run at least in biological duplicates and normalized to ribosomal protein L32 mRNA according to the ΔΔC_t method (Livak & Schmittgen, 2001). The following primer sequences were used: CHD3_fwd: 5′-CCATCTCCTGAACTTCCTCACC-3′, CHD3_rev: 5′-CTTGAGTCTCCGCAGCATGTGT-3′; CHD4_fwd: 5′-CTGTTGCTGACTGGGACACCAT-3′, CHD4_rev: 5′-TGGTCCTCCTTGGCAATGTCAG-3′; EGR3_fwd: 5′-ACGCCAAGATCCACCTCAAG-3′, EGR3_rev: 5′-GGAAAAGTGGGGATCTGGGG-3′; EGR3_ChIP_fwd: 5′-TCCCAAGTAGGTCACGGTCT-3′, EGR3_ChIP_rev: 5′-GGCTCAGCGTTCCTCTTAC-3′; Filaggrin_fwd: 5′-AAAGAGCTGAAGGAACTTCTGG-3′, Filaggrin_rev: 5′-AACCATATCTGGGTCATCTGG-3′; GAPDH_fwd: 5′-GAAGAGAGAGACCCTCACTGCTG-3′, GAPDH_rev: 5′-ACTGTGAGGAGGGGAGATTCAGT-3′; Keratin1_fwd: 5′-TGAGCTGAATCGTGTGATCC-3′, Keratin1_rev: 5′-CCAGGTCATTCAGCTTGTTC-3′; L32_fwd: 5′-AGGCATTGACAACAGGGTTC-3′, L32_rev: 5′-GTTGCACATCAGCAGCACTT-3′; LCE1A_fwd: 5′-GAAGCGGACTCTGCACCTAGAA-3′, LCE1A_rev: 5′-AGGAGACAGGTGAGGAGGAAATG-3′; LCE6A_fwd: 5′-GTCCTGATCTCTCCTCTCGTCT-3′, LCE6A_rev: 5′-CAAGATTGCTGCTTCTGCTGT-3′; LINC00941_fwd: 5′-GACCTTTTCAGGCCAGCATT-3′, LINC00941_rev: 5′-ACAATCTGGATAGAGGGCTCA-3′; MTA1_fwd: 5′-CCAGGACCAAACCGCAGTAACA-3′, MTA1_rev: 5′-GTCAGCTTCGTCGTGTGCAGAT-3′; MTA2_fwd: 5′-TGTACCGGGTGGGAGATTAC-3′, MTA2_rev: 5′-GCCTCCACATTTCCATTTGC-3′; MTA3_fwd: 5′-AGCCCACTTACGGATCGACAGA-3′, MTA3_rev: 5′-CAAACTAGGCTGCCTCACAGAAC-3′; SPRR4_fwd: 5′-AGCCTCCAAGAGCAAACAGA-3′, SPRR4_rev: 5′-GCAGGAGGAGATGTGAGAGG-3′. Statistical significance was tested by an unpaired t test and corrected for multiple testing after Bonferroni (*adj. P < 0.05, **adj. P < 0.01, ***adj. P < 0.001, and ns = not significant).

In vitro RNA pull-down with subsequent protein detection

For in vitro RNA transcription of LINC00941, the Sp6 RNA polymerase was used according to the manufacturer’s instructions with the exception that LINC00941 was labeled with biotin-16-UTP (Roche). 2 μg of biotin-16–labeled LINC00941 was incubated with a protein lysate of 3 × 10⁷ keratinocytes. The protein lysate was prepared beforehand by collecting the cells in lysis buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 5% glycine, 1 mM DTT, 1 mM EDTA, 0.1 U/μl RiboLock RNase Inhibitor [Thermo Fisher Scientific], 1 mM AEBSF, 1x cOmplete Protease Inhibitor Cocktail [Roche]) and clearing the lysates for 15 min, at full speed at 4°C. After 16 h at 4°C, 5 μl of MyOne Streptavidin C1 Dynabeads (Invitrogen) was added, and the mixture was incubated for 1 h at RT and subjected to three wash cycles of 10 min, each using 500 μl wash buffer (150 mM NaCl, 50 mM Tris, pH 7.5, 10% glycine, 5 μM EDTA, 0.5% Igepal, and 1x cOmplete Protease Inhibitor Cocktail [Roche]). After the final wash, magnetic beads were resuspended in 12 μl 1x Laemmli buffer with 10% 2-mercaptoethanol following the Western blot analysis.

Western blot analysis

Protein lysates from calcium-induced differentiated keratinocytes were obtained by scraping cells into a suitable amount of lysis buffer (50 mM Tris, pH 7.5, 150 mM sodium chloride, 5% glycerin, 1 mM DTT, 1 mM EDTA, 0.1 U/μl RiboLock RNase Inhibitor [Thermo Fisher Scientific], 1 mM AEBSF, 1x cOmplete Protease Inhibitor Cocktail [Roche]). Lysates were cleared for 15 min, at full speed at 4°C, and the protein concentration was determined via the Bradford assay. 30 μg of total protein was separated on a 10% SDS–PAGE and transferred onto the Amersham Hybond ECL membrane (Cytiva Life Science) by semi-dry blot. Blocking and antibody dilution were done in 5% milk powder in TBS-T. Primary antibodies included MTA2 (ab8106; Abcam) at 1:1,000 dilution, EGR3 (ab232820; Abcam) at 1:500 dilution, CHD4 (ab72418; Abcam) at 1:5,000 dilution, and β-actin (ab6276; Abcam) at 1:10,000 dilution. Secondary antibodies used were IRDye 800 goat anti-rabbit (926-32211; LI-COR Biosciences) and IRDye 680 goat anti-mouse (926-32220; LI-COR Biosciences), both at a 1:15,000 dilution. Western blots were analyzed with Odyssey Infrared Imager (LI-COR Biosciences).

Lentiviral-mediated LINC00941 overexpression

For HEK293T transfection, Lipofectamine 3000 (Thermo Fisher Scientific) was used according to the manufacturer’s instruction. The lentiviral transfer vector pLARTA-LINC00941 and packaging construct vectors (pUG-MDG and pCMV-ΔR8.91) were applied in equimolar quantities. Viral particles were harvested 48 and 72 h after transfection. For lentiviral transduction, 50,000 keratinocytes per well were seeded in a six-well plate the day before infection. The next day, an appropriate amount of viral particles diluted in keratinocyte medium and 5 μg/ml polybrene were added to the cells. Infection was accomplished in a centrifugation step at RT with 250 rcf for 1 h. Afterward, cells were recovered in keratinocyte medium for at least 24 h before their cell lysate was used for RNA-IP.

MS analysis

MS analysis was essentially performed as previously described (Hoffmeister et al, 2023). Proteins were separated on a NuPAGE 4–12% Bis-Tris gel (Invitrogen) according to the manufacturer’s instructions, and gel lanes were cut into consecutive slices. The gel slices were washed with 50 mM NH₄HCO₃, 50 mM NH₄HCO₃/acetonitrile (3/1), and 50 mM NH₄HCO₃/acetonitrile (1/1) while shaking gently in an orbital shaker (VXR basic Vibrax, IKA). Gel pieces were lyophilized after shrinking by 100% acetonitrile. To block cysteines, reduction with DTT was carried out for 30 min at 57°C followed by an alkylation step with iodoacetamide for 30 min at room temperature in the dark. Subsequently, gel slices were washed and lyophilized again as described above. Proteins were subjected to in-gel tryptic digest overnight at 37°C with ∼2 μg trypsin per 100 μl gel volume (Trypsin Gold, MS grade; Promega). Peptides were eluted twice with 100 mM NH₄HCO₃ followed by an additional extraction with 50 mM NH₄HCO₃ in 50% acetonitrile. Before LC-MS/MS analysis, combined eluates were lyophilized and reconstituted in 20 μl of 1% formic acid. Separation of peptides by reversed-phase chromatography was carried out on UltiMate 3000 RSLCnano System (Thermo Fisher Scientific), which was equipped with a C18 Acclaim PepMap 100 preconcentration column (100 μm i.D. × 20 mm, Thermo Fisher Scientific) in front of an Acclaim PepMap 100 C18 nano column (75 μm i.d. × 150 mm, Thermo Fisher Scientific). A linear gradient of 4–40% acetonitrile in 0.1% formic acid over 90 min was used to separate peptides at a flow rate of 300 nl/min. The LC system was coupled on-line to maXis plus UHR-QTOF System (Bruker Daltonics) via a CaptiveSpray nanoflow electrospray source (Bruker Daltonics). Data-dependent acquisition of MS/MS spectra by CID fragmentation was performed at a resolution of minimum 60,000 for MS and MS/MS scans, respectively. The MS spectrum rate of the precursor scan was 2 Hz processing a mass range between m/z 175 and m/z 2,000. Using Compass 1.7 acquisition and processing software (Bruker Daltonics), a dynamic method with a fixed cycle time of 3 s and a m/z-dependent collision energy adjustment between 34 and 55 eV was applied. Raw data processing was performed in Data Analysis 4.2 (Bruker Daltonics), and Protein Scape 3.1.3 (Bruker Daltonics) in connection with Mascot 2.5.1 (Matrix Science) facilitated database searching of the Swiss-Prot Homo sapiens database (release-2020_01, 220420 entries). Search parameters were as follows: enzyme specificity trypsin with one missed cleavage allowed, precursor tolerance 0.02 D, MS/MS tolerance 0.04 D. Carbamidomethylation or propionamide modification of cysteine, oxidation of methionine, deamidation of asparagine and glutamine were set as variable modifications. Mascot peptide ion-score cutoff was set to 25. Search conditions were adjusted to provide an FDR of less than 1%. Protein list compilation was done using the Protein Extractor function of Protein Scape. EmPAI values (exponentially modified protein abundance index), which can be used for an approximate relative quantitation of proteins in a mixture, were extracted from Mascot. MS analysis resulted in n = 627 LINC00941 interaction partners. GO term analysis was carried out using the PANTHER overrepresentation test (Thomas et al, 2022). GO terms with an FDR of less than 0.05 and chromatin association were considered for further analysis.

RNA-IP

For RNA-IP, the cell lysate of in vitro differentiated keratinocytes overexpressing LINC00941 was obtained by scraping cells into a suitable amount of RIPA buffer. Lysates were cleared for 10 min, at full speed at 4°C. 5 μg of either α-MTA2 antibody (ab8106; Abcam), α-CHD4 (ab263025; Abcam), or IgG control (I5006; Santa Cruz) was added to the lysate and incubated at 4°C for 2 h. 40 μl Protein G Dynabeads (Invitrogen) were transferred to the lysate and incubated for 1 h at 4°C. After three washing steps with RIPA buffer, the beads were resuspended in TRIzol (Invitrogen), and RNA was isolated.

Chromatin immunoprecipitation (ChIP) sequencing

For ChIP, primary keratinocytes were grown to a confluency of about 80% and dethatched from cell culture dishes. Following a cross-linking step adding 1% formaldehyde at RT for 10 min, the reaction was quenched with 125 mM glycine at RT for 5 min. Cells were pelleted and then lysed in swelling buffer (100 mM Tris, pH 7.5, 10 mM potassium acetate, 15 mM magnesium acetate, 1% Igepal, 1x cOmplete Protease Inhibitor Cocktail [Roche], and 1 mM AEBSF). Chromatin fragmentation was subsequently performed in RIPA buffer for 30 min using the S220 Focused-ultrasonicator (Covaris) in the Freq sweeping mode, intensity 8, 20% duty cycle, and 200 cycles per burst. 100 μg fragmented chromatin underwent IP with 5 μg α-MTA2 antibody (ab8106; Abcam) for 16 h at 4°C. 40 μl Protein G Dynabeads (Invitrogen) were added to the IP and incubated for 45 min at RT. The beads were washed three times with wash buffer (100 mM Tris, pH 9.0, 0.5 M lithium chloride, 1% Igepal, 1% sodium deoxycholate, and 1 mM AEBSF) followed by the elution with elution buffer (50 mM sodium bicarbonate and 1% SDS). For reverse cross-linking, 0.2 M sodium chloride was added, and the chromatin was incubated at 67°C for 4 h following a purification step with the QIAquick PCR Purification kit (QIAGEN) according to the manufacturer’s instructions. For library preparation, the NEBNext Ultra II DNA Library Preparation kit for Illumina (New England Biolabs) and NEBNext Multiplex Oligos for Illumina (New England Biolabs) were used following the manufacturer’s instructions. The libraries were pooled equimolar, and the pool was quantified using the KAPA Library Quantification kit—Illumina (Roche). The libraries were sequenced on an Illumina NextSeq 2000 instrument (Illumina) controlled by NextSeq Control Software, v1.4.1.39716, using a 100 cycles P2 Flow Cell with the single-index, paired-end (PE) run parameters. Image analysis and base calling were done by Real Time Analysis Software, v3.9.25. The resulting .cbcl files were converted into .fastq files with bcl2fastq v2.20 software. ChIP sequencing was performed at the Genomics Core Facility “KFB–Center of Excellence for Fluorescent Bioanalytics” (University of Regensburg, Germany).

ChIP-sequencing data analysis

Initially, quality control of the raw sequence reads was conducted using FastQC (v0.11.8) (Andrews, 2010). Subsequently, reads were mapped to the reference genome (GRCh38) using bowtie2 (v2.4.4). The following options were used to optimize the alignment process: --very-sensitive-local, --no-discordant, --no-mix, --dovetail. Aligned reads were filtered for MAPQ >= 30 using SAMtools (Li et al, 2009). Reads mapping to blacklisted genomic regions (ENCODE accession: ENCFF356LFX) were removed using BEDTools (Quinlan & Hall, 2010). Peak calling was executed on all samples relative to input samples using MACS2 software (Zhang et al, 2008). The specific options used for this analysis were as follows: −f BAMPE, −g 2.7e9, --keep-dup auto. After peak calling, the resultant peaks were filtered based on a log q-value threshold >20 for subsequent analyses. This stringent cutoff was used to ensure a high level of confidence in the identified peaks, minimizing the potential for false positives. The number of fragments in each sample falling under these peaks was quantified using the featureCounts function of the Subread package (Liao et al, 2014).

To identify changes in MTA2 binding upon LINC00941 knockdown, the generated count table was processed in R using the Bioconductor package DESeq2 (Love et al, 2014). 33 MTA2 binding sites were found to significantly change their binding abundance upon LINC00941 knockdown based on an FDR of 0.05.

RNA-sequencing data analysis

Publicly available RNA-sequencing data of LINC00941 knockdown during keratinocyte differentiation (GSE118077) were reanalyzed using a Nextflow RNA-sequencing pipeline (Di Tommaso et al, 2017).

Initially, quality control of the raw sequence reads was conducted using FastQC (v0.11.8) (Andrews, 2010). Subsequently, reads were mapped to the reference genome (GRCh38) and the corresponding gene annotation (Ensembl version 106) using Spliced Transcripts Alignment to a Reference software (v2.7.8a) (Dobin et al, 2013). The following options were used to optimize the alignment process: --outFilterType BySJout, --outFilterMultimapNmax 20, --alignSJoverhangMin 8, --alignSJDBoverhangMin 1, --outFilterMismatchNmax 999, --alignIntronMin 10, --alignIntronMax 1,000,000, --outFilterMismatchNoverReadLmax 0.04, --runThreadN 12, --outSAMtype BAM SortedByCoordinate, --outSAMmultNmax 1, and --outMultimapperOrder Random.

Post-mapping quality control was performed using the rnaseq mode of Qualimap (v2.2.1) (García-Alcalde et al, 2012). The level of PCR duplication was assessed using Picard MarkDuplicates (v2.21.8) (Broad Institute, 2019) and dupRadar (v1.15.0) (Sayols et al, 2016). Gene expression quantification was carried out using featureCounts (v1.6.3) (Liao et al, 2014).

Finally, differential gene expression analysis between LINC00941 knockdown and control samples was conducted using the Bioconductor DESeq2 package (Love et al, 2014). The normal shrinkage method was applied for scaling log₂ fold changes (Zhu et al, 2019). Genes with an FDR of less than 0.05 were considered significantly differentially expressed. This analysis resulted in the identification of 385 down-regulated and 1,532 up-regulated genes in the LINC00941 knockdown keratinocytes at day 3.

Downstream analysis

ChIP-sequencing coverage tracks were generated using the bamCoverage function from the deepTools package (Ramírez et al, 2016). For normalization, the size factors derived from DESeq2 analysis were used. Data track visualizations were generated using the Bioconductor package Gviz (Hahne & Ivanek, 2016). The Bioconductor package ChIPseeker was used to annotate the genomic regions of the MTA2 binding sites (Yu et al, 2015). To link the MTA2 binding sites to the next gene, only protein-coding genes were considered. Gene set overrepresentation analysis was carried out on MTA-associated genes with a maximum distance to the TSS of 3 kb using the Bioconductor package clusterProfiler (Wu et al, 2021).

External datasets used in this study

The 15-state core chromatin model of male keratinocytes (roadmap accession: E057) was obtained from the roadmap repository (Roadmap Epigenomics Consortium et al, 2015). The corresponding histone modification and RNA-sequencing data tracks were obtained from ENCODE repository (The ENCODE Project consortium, 2012; Luo et al, 2020): poly(A) RNA-sequencing plus/minus strand (ENCFF283IQC/ENCFF804BRH), H3K4me3 (ENCFF517FHJ), H3K27me3 (ENCFF400FLX), H3K4me1 (ENCFF319BIJ).

Data on in vitro differentiation of foreskin keratinocytes (at day 0/day 2.5/day 5.5) were obtained from the ENCODE portal: ChromHMM 18-state model (ENCFF571LVQ/ENCFF385FJV/ENCFF058GJN), H3K4me1 (ENCFF539WZS/ENCFF546FWF/ENCFF821VJZ), H3K27ac (ENCFF786TBH/ENCFF870GTM/ENCFF988UOV), H3K27me3(ENCFF492WYD/ENCFF846IJG/ENCFF902OOG), poly(A) RNA-sequencing plus (ENCFF646IJP/ENCFF786AVM/ENCFF533VFT), poly(A) RNA-sequencing minus (ENCFF699EVT/ENCFF701XBY/ENCFF366JLV), poly(A) RNA-sequencing quantification (ENCFF423MWU/ENCFF137YHI/ENCFF379PNP).

RNA-sequencing data of LINC00941 knockdown after days 2 and 3 were obtained from Gene Expression Omnibus (GEO; GSE118077) (Ziegler et al, 2019).