Introduction

ATP-binding cassette transporter family A member 12 (ABCA12, OMIM 607800) is a keratinocyte transmembrane lipid transporter. ABCA12 is expressed in the stratum spinosum and stratum granulosum of the skin, where it is localized in lamellar granules (LGs), the cellular organelles that contain the lipids, proteins and enzymes needed for formation of the stratum corneum1. The function of ABCA12 is critical for keratinocyte differentiation as well as for maintenance of the skin permeability barrier via the formation of intercellular lipid layers in the stratum corneum2,3,4. In normal skin and in cultured human keratinocytes, the expression of ABCA12 parallels the differentiation of the keratinocytes3. Accordingly, mutations in the ABCA12 gene underlie three distinct phenotypes of autosomal recessive congenital ichthyosis: harlequin ichthyosis (HI, OMIM 242500) and lamellar ichthyosis/congenital ichthyosiform erythroderma (LI/CIE, OMIM 601277)5,6,7. Since there are several causative genes whose mutation may cause congenital ichthyosis, identification of the pathogenic mutation in the patient's genome is essential for correct diagnosis8.

There are cases, however, in which the results of genetic diagnosis are inconclusive, such as for patients of HI, LI or CIE who possess a recessive pathogenic ABCA12 mutation in one allele but whose other allele is apparently intact7. In such cases, verification of decreased ABCA12 mRNA expression in keratinocytes taken from the patient's skin or the hair follicles confirms the diagnosis9. Apart from anomalous splicing events due to intronic mutations or large deletion mutations, one possible reason for the decreased expression is loss of the promoter function due to mutations in key promoter elements.

To date, however, detailed functional promoter analysis of ABCA12 has not been carried out and key genetic elements that regulate ABCA12 expression have not been described.

This study aims to identify the promoter region of ABCA12 and to locate the essential elements therein, thus providing the necessary information for genetic diagnostic screening of congenital ichthyosis. The results show that ABCA12 expression in differentiating cultured human keratinocytes is critically dependent on a palindromic motif that resides in the region of −2084 to −2078 from the transcription start site (TSS).

Results

Identification of the critical region of −2200 to −1934 in the upstream promoter of the ABCA12 gene

To locate the essential elements in the upstream region of ABCA12, we first cloned a −2980/+190 fragment (the base position +1 is the TSS of ABCA12, chr2: 216,003,151) and performed a dual luciferase reporter assay to assess its promoter activity with respect to keratinocyte differentiation status. We utilized cultured normal human epidermal keratinocytes (NHEK). It is well established that increased cellular calcium level is a key signal to promote the differentiation of keratinocytes in culture as well as in normal skin10,11. Accordingly, NHEKs differentiate under the high-calcium condition and ABCA12 expression increases in differentiated keratinocytes1,2. As expected, the −2980/+190 fragment showed 3.2 times as much activity when NHEKs were cultured under the high-calcium condition (1.2 mM) compared to the low-calcium condition (0.06 mM) (Fig. 1, Supplementary Fig. S1). We performed an insilico search of the ENCODE dataset using the UCSC genomic browser12 throughout this region and found increased ChIP-seq signals and peaks of H3K4me3 and H3K27ac and increased DNAseI-seq peaks in NHEK (Fig. 2). These results were considered to support the bona fide promoter function of this genetic region. Additionally, an in silico analysis using JASPAR CORE database (http://jaspar.genereg.net/)13 found 398 putative transcription factor binding sites within this region.

Figure 1
figure 1

Position −2980 to position −1994 of the ABCA12 gene has promoter activity.

Dual luciferase assay was performed using firefly luciferase reporter vectors that contain −2980/+190, −1994/+190 or −1039/+190 fragments of the ABCA12 gene. The culture medium contained either 0.06 mM or 1.2 mM calcium. The promoter activities are normalized using the empty vector as a standard. The bars represent the standard deviation (n = 3). *P < 0.01 compared with the −2980/+190 fragment.

Full size image
Figure 2
figure 2

Annotation of the ABCA12 upstream gene using the ENCODE dataset.

The ABCA12 upstream genomic region (chr2:215,993,140–216,017,439:GRCh37/hg19) is illustrated using the UCSC genome browser. ChIP-seq signals and peaks of H3K4me3 and H3K27ac and DNAseI-seq peaks in NHEK were extracted from the ENCODE dataset and annotated. The read density of ChIP-seq signals is indicated in the Y-axis. ChIP-seq peaks of c-Fos/c-Jun in 91 human cell types from the ENCODE dataset are also shown. c-Fos peaks for MCF10A-Er-Src cells stimulated with tamoxifen for 4 hr (MCF10AEr + T4 FOS), 12 hr (MCF10AEr + T12 FOS), or 36 hr (MCF10AEr + T36 FOS) were found in the database search. The −2084/−2078 palindromic motif is boxed and its genetic location is highlighted with a dotted line. The predicted −2133/−2144 Sp1 binding element is also indicated.

Full size image

We then aimed to narrow down the region that contains the essential promoter elements experimentally. For this purpose, increasing lengths of the upstream region of the ABCA12 gene were cloned and their promoter activities were assessed by a dual luciferase reporter assay (Fig. 1, Supplementary Fig. S1). Under the high-calcium condition, the highest promoter activity was detected with the −2980/+190 fragment, whereas −1994/+190 and −1039/+190 fragments showed drastically decreased promoter activity (−12.2 fold, P < 0.01 and −6.41 fold, P < 0.01 compared to the −2980/+190 fragment, respectively). These results suggest that one or more critical promoter elements may reside within the region from −2980 to −1994 of the ABCA12 promoter.

To further narrow the region containing the putative element(s), promoter activity was analyzed for five serially truncated fragments consisting of the −2980/+190 to −1934/+190 ABCA12 upstream region (Fig. 3a, Supplementary Fig. S2). Under the high-calcium condition, the 2700/+190 fragment showed roughly comparable promoter activity to the −2980/+190 fragment (−1.43 fold compared to the −2980/+190 fragment, not significant), whereas the −1994/+190 fragment showed markedly reduced promoter activity (−16.1 fold compared to the −2980/+190 fragment, P < 0.01). Promoter activities of the −2400/+190 to −2200/+190 fragments were decreased compared with that of the −2980/+190 fragment (−2.28 fold, P < 0.01 and 35.3%, P < 0.01, respectively). Therefore, we considered that the region from −2700 to −1994 is critical for the promoter activity of the ABCA12 upstream region in NHEK.

Figure 3
figure 3

Position −2200 to position −1934 of the ABCA12 gene exhibits potent promoter activity.

Promoter activities of −2980/+190, −2700/+190, −2400/+190, −2200/+190, or −1994/+190 fragments (a) and −2980/−2651, −2700/−2351, −2400/−2108, or −2200/−1934 fragments (b) were assayed. *P < 0.01 compared with the −2980/+190 fragment (a) or the −2980/−2651 fragment (b).

Full size image

To identify the most critical region between −2700 and −1994, four short overlapping fragments spanning the region were cloned and their promoter activities were assessed (Fig. 3b, Supplementary Fig. S2b). The −2200/−1934 fragment showed the highest promoter activity, suggesting that this region may contain the critical element(s) (Fig. 3b, Supplementary Fig. S2).

Interestingly, the promoter activity of this fragment was not dependent on the increased calcium concentration, suggesting that an unknown element outside the −2200/−1934 region may exist to suppress the activity under low-calcium conditions. The neighboring −2400/−2106 fragment also showed some promoter activity, although not as much as the −2200/−1934 fragment.

A palindromic motif in the −2084 to −2078 region is essential for the promoter activity

To pinpoint the critical bases that could be used for the genetic diagnostic screening of congenital ichthyosis patients, we searched for consensus sequences for transcription factor binding elements within the region from −2200 to −1934. An in silico analysis using the JASPAR CORE database found a predicted specificity protein 1 (Sp1) binding motif (−2133 to −2144, reverse strand, chr2:216,005,284–216,005,295) and a palindromic motif that matches an AP1 binding sequence (tgagtca, −2084 to −2078, chr2:216,005,235–216,005,228) within the −2200/−1934 fragment (Fig. 2, Supplementary Fig. S3).

The AP1 elements may be able to tether AP1 transcription factors to the promoter region of the genome and regulate downstream gene expression. In keratinocytes, AP1 regulates the expression of various genes, such as involucrin and transglutaminase 114; these AP1-regulated genes are involved in keratinocyte differentiation and are mostly expressed in the late stage of terminal differentiation15. When these motifs were mapped in the ENCODE dataset annotations using the UCSC genome browser, the palindromic −2084/−2078 motif was found to reside within the increased ChIP-seq signals of H3K4me3 and H3K27ac and within the DNAseI-seq peak in NHEK (Fig. 2).

These considerations lead us to hypothesize that the −2084/−2078 motif may regulate the expression of ABCA12. Therefore, we introduced one to three nucleotide deletion mutations into the wild-type sequence (tgagtca): mutant Δ1 (tga-tca), mutant Δ2 (tga--ca) and mutant Δ3 (tg---ca). We investigated the effect of the mutations on the promoter activity of the −2980/+190 fragment (Fig. 4, Supplementary Fig. S4). All three mutants showed strikingly reduced promoter activity compared to the wild type (−6.10 fold, P < 0.01, −9.71 fold, P < 0.01 and −11.7 fold, P < 0.01 for Δ1, Δ2 and Δ3, respectively) (Fig. 4, Supplementary Fig. S4). No significant differences in promoter activity were found between the three mutants.

Figure 4
figure 4

Mutants of the −2084/−2078 motif in ABCA12 show striking reduction of promoter activity.

Three different mutants of the −2084/−2078 motif were generated using the −2980/+190 fragment as a template and the promoter activity of each fragment was assessed. The mutants lacked 1–3 bases from the wild-type sequence (tgagtca): mutant Δ1 (tga-tca), mutant Δ2 (tga--ca) and mutant Δ3 (tg---ca). The promoter activities of the mutants were compared to that of the wild type. *P < 0.01 compared with the wild type.

Full size image

Discussion

The present results clearly demonstrate that disruption of the −2084/−2078 motif alone can critically decrease the promoter activity of the entire 3 kb upstream sequence of ABCA12. In addition, the 267 bp fragment containing this motif (−2200/−1934) is sufficient to establish potent promoter activity. This experiment using short genomic fragments may have the shortcoming that the assays are done without the context of their original core promoters or other minimal promoters and the interpretation of the results may be limited to the demonstration of a possible capability of each fragment to recruit transcription factors. The following mutagenesis experiment thus confirmed the essential role of the −2084/−2078 motif in its native genomic context. In support of these results, the ENCODE dataset search showed increased ChIP-seq signals for H3K4me3 and H3K27ac, which mark active promoters and a DNAseI-seq peak that demonstrates the open chromatin state in NHEK. Interestingly, in MCF10A-Er-Src cells stimulated with tamoxifen, the ChIP-seq tag containing this motif is associated with a peak for c-Fos, a component of AP1 (Fig. 2), showing that the motif may indeed recruit these transcription factors in certain cells, although it is not yet proven whether c-Fos binds the −2084/−2078 motif in NHEK. In agreement with the experimentally proven significance of this motif, a database search shows a high conservation of the −2084/−2078 sequence among multiple eutherian species (Supplementary Fig. S5). We searched for previously reported SNPs in the HAPMAP database16 in this genetic region, but we found no SNP in the −2084/−2078 sequence in populations with European or Asian ancestry.

Apart from the −2084/−2078 motif, the −2200/−1934 fragment harbors a predicted −2133/−2144 Sp1 binding element. It is reported that AP1 and Sp1 cooperatively regulate the expression of target genes in keratinocytes, such as LOR, which encodes loricrin17. Therefore, it may be assumed that the expression of ABCA12 is also regulated by the putative −2133/−2144 element. Nevertheless, in the actual context of the entire promoter sequences, the predicted element may play rather an adjunctive role in ABCA12 expression, because the disruption of the −2084/−2078 motif alone is sufficient to reduce the promoter activity of the −2980/+190 fragment to one-tenth of its original activity.

In the present study, we are aware of the limitation of depending solely on luciferase assays performed on NHEK, as they do not provide direct evidence for the function of the −2084/−2078 motif in its true genomic context. However, the existing ChIP-seq and DNaseI-seq results obtained from NHEK are in accordance with the proposed function of this motif as the essential promoter element and can reinforce the conclusions obtained from our experiments.

From these considerations, we propose that future routine genetic screening for the diagnosis of congenital ichthyosis patients include the sequencing of the −2084/−2078 upstream region of ABCA12, especially if ABCA12 mutation is detected in one of the patient's allele. We sequenced the −2200/−1934 ABCA12 upstream region of five autosomal recessive congenital ichthyosis patients. These patients showed decreased ABCA12 mRNA expression; however, known pathogenic mutations within ABCA12 exons or exon-intron boundaries were found in only one allele of each patient. Although no mutations were detected in this limited number of patients (data not shown), the screening of future patients could benefit from sequencing of the genetic region.

This report describes the −2084/−2078 palindromic motif upstream of the ABCA12 gene as a promoter element. However, it may be more precisely termed an enhancer element, because it is farther upstream of the core promoter that neighbors the TSS of ABCA12. Nevertheless, there is no widely accepted standard for differentiating between a promoter element and an enhancer element. Considering the comparative proximity of the motif (2.1 kb from TSS), we propose that it may be reasonably termed a promoter element.

It is of interest that the −2200/−1934 fragment exhibited potent promoter activity in the low-calcium condition as well as in the high-calcium condition. Since the −2980/+190 fragment shows a marked calcium-dependent upregulation of its promoter activity, it may be suggested that an unknown suppressor element(s) exists outside the −2200/−1934 region and functions specifically under low-calcium conditions. However, suppressor elements would not warrant the genetic sequencing of autosomal recessive congenital ichthyosis patients and is beyond the scope of the present study.

Methods

Ethics statement

This study was approved by the Medical Ethics Committee of Nagoya University (#1088) and performed according to the Declaration of Helsinki Principles. The participants gave written informed consent. Written informed consent was obtained from the guardians on behalf of the children enrolled. We recorded participant consent in paper. The ethic committee approved the consent procedure.

Construction of plasmids

To determine the promoter region of the ABCA12 gene, a series of luciferase reporter plasmids were generated. PCR fragments containing upstream regions of the human ABCA12 gene (NCBI Reference Sequence: NG_007074.1, http://www.ncbi.nlm.nih.gov/) of increasing lengths were amplified using specific primers (Table 1). The fragments were subcloned using the In-Fusion HD Cloning Kit (Clontech) into a pGL4.10 basic vector (Promega) that contains the firefly luciferase gene but does not contain any eukaryotic regulatory elements. All of the produced vectors were verified.

Table 1 PCR Primers for cloning of the promoter region of ABCA12
Full size table

Cell culture and transfection

Normal human epidermal keratinocytes (NHEKs) were cultured in EpiLife (Gibco) supplemented with human keratinocyte growth supplement (HKGS; Gibco) at 37°C under 5% CO2. For transfection of the plasmid vectors, the NHEKs were seeded on 24-well plates (1.0 × 105 cells/ml). At 50–70% confluence, the cells were washed once with PBS. Transfection was then performed by adding 0.5 μg of firefly luciferase reporter vector and 0.05 μg of pGL4.74 [hRluc/TK] renilla luciferase reporter vector (Promega) mixed with TransIT-Keratinocyte Transfection Reagent (Mirus) in EpiLife and incubating for 20 min at 25°C. The transfected NHEKs were incubated for 6 h at 37°C with the mixture and then the medium was changed to EpiLife/HKGS supplemented with either 0.06 mM or 1.2 mM calcium.

Dual-luciferase reporter assay

The NHEKs were incubated in either low (0.06 mM) or high (1.2 mM) calcium condition for 48 h after transfection. Then, the promoter activity was measured by Dual-Luciferase Reporter Assay System (Promega) using the GloMax-Multi Detection System (Promega) according to manufacturer's protocol. All experiments were performed in triplicate. The promoter activity was calculated as the ratio of firefly luciferase reporter expression to Renilla luciferase reporter expression that is driven by the Herpes simplex virus thymidine kinase promoter contained in the pGL4.74[hRluc/TK] vector.

In silico analysis of transcription factors

We performed in silico analysis to reveal potential transcription binding elements in the upstream region of the ABCA12 gene using the JASPAR CORE database (http://jaspar.genereg.net/)13. The sequence was scanned with JASPER CORE vertebrata matrix models with a relative profile score threshold of 90%.

ENCODE and HAPMAP data set search and comparative multiple alignments

A series of database searches was performed using the UCSC Genome Browser database; ENCODE dataset for ChIP-seq signals and peaks of H3K4me3 and H3K27ac in NHEK, DNAseI-seq peaks in NHEKs, transcription factor ChIP-seq peaks of c-Fos/c-Jun in 91 human cell types and SNPs from HAPMAP datasets12,16,18. All genomic annotations are mapped on genome assembly GRCh37/hg19. The multiple alignments of the conserved AP1 element among 55 vertebrae species were also obtained and the alignment and consensus logo was generated with CLC Main Workbench software (CLC Bio).

Mutation analysis

To confirm the promoter activity of the −2084/−2078 motif (TGAGTCA), three different mutants were constructed by inverse PCR using the pGL4.10 −2980/+190 reporter as a template (Table 1). The three mutants lacked 1–3 bases from the wild-type sequence (wild-type AP1, tgagtca; mutant Δ1, tga-tca; mutant Δ2, tga--ca; mutant Δ3, tg---ca). Promoter activities of reporter vectors containing the mutants were analyzed and compared with that of the wild type.

Statistics

Statistical analyses were performed with the GraphPad Prism software version 5.04 (GraphPad Software). One-way analysis of variance (ANOVA) with Tukey post hoc test was used to determine statistical significance. A P value of <0.05 was considered significant.