Molecular analysis of the human placental cysteine dioxygenase type 1 gene.

Sulfate is essential for healthy fetal growth and development. Cysteine dioxygenase type 1 (CDO1) plays an important role in the catabolism of cysteine to sulfate. Cdo1 knockout mice exhibit severe and lethal fetal phenotypes but the involvement of CDO1 gene variants in human development is unknown. We searched the NCBI and Ensembl gene databases and identified four alternatively spliced CDO1 coding mRNA transcripts, as well as 148 validated CDO1 gene variants, including 138 missense, 6 nonsense, 1 frameshift, 1 in-frame deletion, and 2 splice site variants. In silico analyses predicted 68 of the missense variants to be deleterious to CDO1 protein structure and function. We examined the relative abundance of the four CDO1 coding mRNA transcripts in human term placentas using qRT-PCR. CDO1 mRNA variant 2 was the most abundant transcript, with intermediate levels of variant 4 and lower levels of variants 1 and 3. Using in situ hybridization, we localised CDO1 mRNA expression to the syncytiotrophoblast layer of human term placenta. To investigate the regulation of CDO1 gene expression, we analysed the transcriptional activity of the human CDO1 5'-flanking region in the JEG-3 placental cell line using luciferase reporter assays. Transcriptional activities were identified in the regions -5 to -269 and - 269 to -1200 nucleotides upstream of the CDO1 transcription initiation site. Mutational analyses of a single nucleotide polymorphism -289C > G that is common in the general population (allele frequency = 0.37) and a putative transcription factor binding motif (CCAAT enhancer binding protein beta) did not alter transcriptional activity of the CDO1 5'-flanking region. Collectively, this study provides an overview and analysis of human CDO1 for future investigations of this gene in human health.

Introduction

Sulfate is an obligate nutrient for numerous metabolic and cellular processes during fetal growth and development [1]. During pregnancy in mice, sulfate is supplied from maternal circulation to the fetus via the placenta [2]. Our previous studies showed that either reduced levels of sulfate in maternal circulation during pregnancy or disruption of sulfate transport through the placental syncytiotrophoblast layer leads to fetal demise in mice [3,4]. Remarkably, little is known about the physiology of maintaining sulfate supply to the developing human fetus, or the clinical consequences of human fetal sulfate deficiency.

In adults and children, approximately one third of sulfate requirements are obtained from the diet and the remaining two thirds are obtained from the catabolism of the sulfur-containing amino acids methionine and cysteine [5]. A higher proportion of absorbed free inorganic sulfate (SO42−) from the diet, together with increased renal reabsorption of filtered sulfate, helps to maintain body sulfate homeostasis when dietary methionine and cysteine intake is reduced (i.e. low protein) [5]. Methionine is converted to cysteine via the transsulfuration pathway, and cysteine is oxidised to sulfate via a major pathway involving cysteine dioxygenase type 1 (CDO1, EC 1.13.11.20) [6]. Human CDO1 is abundantly expressed in the adult liver, whereas negligible levels were found in first-trimester fetal liver [7]. The gestational age when CDO1 expression increases in the human fetus is not known but this most likely occurs in late gestation, which was found to be the case in fetal mice [8]. The expression of CDO1 from late gestation may be related to the absence of methionine to cysteine conversion in the human fetus because cystathionine γ-lyase, the last enzyme of the transsulfuration pathway, is not typically expressed until the early neonatal period [9]. Early studies also reported abundant expression of CDO1 mRNA in human term placenta [10]. The potential role of placental CDO1 in sulfate generation and supply to the fetus has not been considered.

As an initial step towards understanding the potential role of CDO1 in contributing to sulfate homeostasis in the human placenta and fetus, we provide an update on: CDO1 gene structure and its mRNA variants; CDO1 protein isoforms; CDO1 gene variants and their predicted impact on CDO1 protein function; and identify the cellular distribution of CDO1 mRNA expression in human placenta, as well as the minimal promotor sequence of the human CDO1 5′-flanking region that confers transcriptional activity.

Materials and methods

CDO1 gene, cDNA, protein and 5′-flanking sequences

We searched the NCBI Gene, Nucleotide, Protein, UniGene and SNP databases (https://www.ncbi.nlm.nih.gov/) using the term “CDO1” and “Homo sapiens” within the date range 19 to 22 January 2019. For this study, we used the reference CDO1 gene and 5′-flanking region (NC_000005.10), CDO1 mRNA (NM_001323565.1, NM_001801.2, NM_001323566.1, NM_001323567.1) and CDO1 protein (NP_001310494.1, NP_001792.2, NP_001310495.1, NP_001310496.1) sequences. Validated CDO1 gene variants were obtained from the Ensembl gene browser database (http://asia.ensembl.org/index.html) within the date range 24 to 30 July 2019. The predicted impact of each missense variant on CDO1 protein function was obtained using Sorting Intolerant From Tolerant (SIFT), Polymorphism Phenotyping (PolyPhen), Combined Annotation Dependent Depletion (CADD), Rare Exome Variant Ensemble Learner (REVEL), Methodfor deleterious missense mutations using Logistic Regression (MetaLR), and Mutation Assessor (MA) scores from the Ensembl Variant Effect Predictor (http://www.ensembl.org/Tools/VEP). Amino acid sequences of all four CDO1 protein isoforms were aligned using ClustalW software [11]. A phylogenetic tree of 20 species, with CDO1 orthologues reported in the NCBI database, was generated using the Interactive Tree of Life program (http://itol.embl.de/). Putative transcription factor binding motifs within the first 1500 nucleotides of the 5′-flanking region of CDO1 were identified using MatInspector software [12] and then a multiple species alignment of those motifs was generated using the DiAlign TF program (http://www.genomatix.de).

Placental tissues, RNA isolation, PCR and in situ hybridization

The research protocol was approved by the Mater Human Research Ethics Committee. Human placentae (n = 10 male and 6 female babies) were obtained from uncomplicated pregnancies ≥37 weeks gestation at elective caesarean section between 0900 and 1200 h, and sampled within 20 min of delivery. Placental weights and neonatal birth weights were within normal ranges [13,14]. Total RNA was isolated from placentae using previously described methods [15]. Four μg RNA was reverse transcribed by using random hexamers and an Omniscript RT kit (Qiagen) as recommended by the manufacturer. PCR was performed as previously described [16] using 200 nM forward and reverse primers (Table 1) in a Rotor-Gene 6000 thermal cycler (Corbett Research, Sydney, Australia). CDO1 variant mRNA-specific primers were used to determine the relative abundance of the 4 coding CDO1 mRNA variants V1-V4 (Table 1). The thermal cycling protocol was: 50 °C for 2 min; 94 °C for 2 min; 45 cycles of 94 °C for 1 s, 60 °C for 10 s, and 72 °C for 15 s. RNA expression levels and absolute threshold cycle values (Ct values) of each gene were normalized to those of GAPDH RNA with the Rotor-Gene 6000 series software (Corbett Research). Amplification specificity was confirmed by melting curve analysis. For in situ hybridization, we synthesized digoxigenin (DIG) labelled probes as previously described [17], using CDO1 forward (P17) and reverse (P18) primers (Table 1), which can bind to all 4 coding CDO1 variant mRNAs. Tissues were dissected into 4% paraformaldehyde in PBS and fixed overnight at 4 °C, embedded in paraffin, sectioned at 7 μm, probed, washed and then processed for antibody detection of DIG-labelled probes and colour development as previously described [18].

Table 1

Primers used for PCR, mutagenesis, DNA sequencing, and in situ probes.

Primer	gDirection	Sequence (5′ to 3′)		hPrimer location
aPrimers used for PCR-amplifying CDO1 5′-flanking regions
P1	F	CCCGGGATCCACCGGTACATCACCAAGGCTGGCTTC		−1200
P2	F	CCCGGGATCCACCGGTTACGTCCCAGCGTCGCGAAC		−269
P3	F	CCCGGGATCCACCGGTGGTACATTCCTAGTGACTCC		−5
P4	R	CATGGTGGCGACCGGTCTCGTGGGGAGCTGGCTG		+311
bPrimers used for site directed mutagenesis
P5	F	TCAGTCCCCGCAGCCATGTCCTCCGACCCTTTTTG		−306
P6	R	GGCTGCGGGGACTGACGCTGAGTAAAGGAGGAAAA		−292
P7	F	GCATTTCTTCCACCTTTGGCATCTTTCTCATTTAAA		−915
P8	R	AGGTGGAAGAAATGCAAGTGTAGATTTCTTCCTTC		−901
cPrimers used for DNA sequencing
P9	F	CTGTGGATAACCGTATTAC
P10	R	CAGATGTCGATGTTGGGG
dPrimers used for quantitative PCR
P11	F	GTTCGACCAGTACAGTCGTG	V1 c.156
eP12	F	GGTGAAGGACATGGCAGCAG	V2 c.232
P13	F	GGGGTGAAGGACATGGCAGT	V3 c.230
P14	F	TGCCAGGGGCCTGGGGGTAT	V4 c.-62
P15	R	AGCATCTTCAGAAAGCAGTG	V1 c.350,	V4 c.77
P16	R	GTAAGCCAATGGAATCATTG	V2 c.418,	V3 c.415
fPrimers used for PCR-amplifying in situ probes
P17	F	aattaaccctcactaaagggACCTGCTGTGAGCCTTCACT	V1 c.815,	V2 c.758
			V3 c.755,	V4 c.351
P18	R	taatacgactcactatagggCCATGGGACACAGACATGAG	V1 c.*1605,	V2 c.*1548
			V3 c.*1545,	V4 c.*1141

Underlined sequence homologous to the pMetLuc2 vector to allow recombination cloning.

Underlined nucleotides introduce the -289C > G SNP (P5) and mutate the core sequence of the CEBPB motif (P7).

Primers located in the pMetLuc2 vector upstream (P9) or downstream (P10) of the multiple cloning site.

CDO1 variant mRNA-specific primers enabled amplification of V1 (P11 + P15), V3 (P13 + P16) and V4 (P14 + P15).

Since primer combination P12 + P16 amplified variants 1, 2 and 4, we calculated variant 2 mRNA abundance as [(amplicon abundance using P12 + P16) – (V1 + V4 amplicon abundance)].

Lower case represents T3 (P17) and T7 (P18) RNA polymerase binding sites, and uppercase represents CDO1 sequences in exon 7 (P17) and the 3′-UTR of exon 8 (P18).

F, forward primer; R, reverse primer.

Number denotes the position of the first 5′-nucleotide in the CDO1 5′-flanking region (P1-P8) or CDO1 variant V1, V2, V3 or V4 mRNAs (P11-P18).

Sub-cloning and mutagenesis of the CDO1 5′-flanking region

Fragments of the CDO1 5′-flanking region with 5′-ends at −1200, −269 and −5, and with a common 3′-end at +311 (+1 is the transcription start site of CDO1 variant 2 mRNA NM_001801.2), were amplified by PCR using 1 μM primers (Table 1), and 1.25 U LA TAQ polymerase (TakaRa). The thermal cycling protocol was: 95 °C for 1 min; 35 cycles of 95 °C for 30 s, 60 °C for 30 s, and 72 °C for 3 min, followed by 1 cycle of 72 °C for 10 min. Each fragment was cloned upstream of a luciferase reporter gene in the multiple cloning site of the pMetLuc2 expression vector (Clontech) using In-Fusion reagent as described by the manufacturer (Clontech). Sequence variants were introduced into the cloned −1200 CDO1 5′-flanking region by PCR using 1 μM forward and reverse primers (Table 1), 6% DMSO, 1 U Phusion DNA Polymerase and thermal cycling parameters: 95 °C for 2 min; 35 cycles of 95 °C for 20 s, 59 °C for 20 s, and 72 °C for 3 min, followed by 1 cycle of 72 °C for 5 min. These PCR products were then treated with In-Fusion regent (Clontech) to generate circularized vectors, as previously described [19]. Nucleotide sequences of all cloned CDO1 sequences were verified by DNA sequence analysis using 9.6 pmol forward and reverse primers (Table 1) as previously described [19].

Cell culture and luciferase assays

The JEG-3 placental cell line was cultured in DMEM medium containing l-glutamine and glucose (Invitrogen) supplemented with 10% heat-inactivated fetal calf serum, 1% penicillin–streptomycin and 0.1% plasmocin (InvivoGen). Cultures were maintained at 37 °C in a humidified atmosphere of 5% CO2 and 95% air. JEG-3 cells were transiently co-transfected using Lipofectamine2000 (Invitrogen) with pMetLuc2 containing individual cloned CDO1 fragments or pMetLuc2 that lacks cloned fragments (negative control), and pSEAP2 alkaline phosphatase expression vector used to normalise luciferase activity. Cells (60% confluent in 48 well plates) were incubated with 2 μl Lipofectamine2000 and plasmids (3.2 ng each pMetLuc2 and pSEAP2) in DMEM containing 10% fetal calf serum (total volume 100 μl) for 7 h, and then the media was replaced with 800 μl OptiMEM® (Invitrogen) containing 10% fetal calf serum for an additional 70 h at 37 °C. Luciferase and alkaline phosphatase activities in the cell culture media were assayed using protocols and reagents (Ready-To-Glow™ Dual secreted reporter assay) purchased from Clontech, and measured using a PolarStar Omega plate reader (BMG Labtech).

Statistical analyses

Statistical significance of CDO1 mRNA levels between placentae from male and female babies was evaluated using the unpaired 2-tailed Student's t-test. The statistical significance of the differences of luciferase activities between each clone and control vector was evaluated using a one-way ANOVA, followed by a Dunnett multiple comparisons test. p < 0.05 considered significant.

Results and discussion

CDO1 gene, mRNA, protein and variant sequences

The human CDO1 gene contains 8 exons, spanning approximately 12 kb at chromosomal location 5q22.3 (Fig. 1A). Earlier investigations reported a CDO1 mRNA of approximately 1.5 kb encoding a 200 amino acid CDO1 protein [20], corresponding to transcript variant 2 (NM_001801.2) and protein isoform 2 (NP_001792.2) sequences in the current NCBI database. More recently, additional mRNA variants have been added to the NCBI database, including a total of 4 protein coding variants V1 to V4 (Fig. 1B), and 4 non-coding variants (not shown): V5 (NR_136618.1), V6 (NR_136619.1), V7 (NR_136620.1) and V8 (NR_136621.1). These eight mRNA variants arise from alternative splicing of exons and differential use of exon-intron junction sequences, as well as transcription initiation from 2 sites. Variants 1, 2, 3, 5, 6 and 7 are transcribed from 311 nt upstream of the ATG start codon, whereas variants 4 and 8 have a transcription start site at 260 nt upstream of ATG which was previously identified in an early study that investigated human CDO1 mRNA using primer extension analysis [10]. Variant 2 and its encoded 200 amino acid isoform 2 are the reference sequences in the current literature. However, the physiological roles of the other 3 coding CDO1 mRNA variants (Fig. 1B) and CDO1 protein isoforms (Fig. 1C) awaits further investigation.

Fig. 1

Human CDO1 gene structure, mRNA variants and protein isoforms. (A) Exon-intron organization showing exons (vertical lines) and introns (horizontal lines) spread over approximately 12 kb. Reference sequence NC_000005.10. (B) Schematic showing exons 1–8 (boxes) and protein coding sequences (white portions) for CDO1 mRNA variants V1 (NM_001323565.1), V2 (NM_001801.2), V3 (NM_001323566.1) and V4 (NM_001323567.1). (C) Aligned human CDO1 protein isoform 1 (NP_001310494.1), 2 (NP_001792.2), 3 (NP_001310495.1) and 4 (NP_001310496.1) sequences. Alignments were generated using the Clustal W program [11]. *Amino acids present in all 4 isoforms. Sequences accessed from the NCBI from 19 to 22 January 2019. (D) Validated CDO1 missense (letters), nonsense (*) and in-frame single amino acid deletion (Δ) variants (Ensembl database) are shown below the CDO1 isoform 2 protein sequence. Location of each α-helix (H1-H2, waved lines) and β-sheet (S1-S13, horizontal arrows), as well as the amino acids interacting with cysteine substrate (bar above CDO1 sequence) and/or Fe2+, in CDO1 isoform 2 protein (PDB ID: 2IC1) from the NCBI structure database (https://www.ncbi.nlm.nih.gov/structure/). Highlighted missense variants are predicted to be detrimental (red) or moderately disruptive (blue) to the CDO1 protein structure as determined using a combination of tools (SIFT, PolyPhen, CADD, REVEL, MetaLR, MA). Serine 83 (vertical arrow) is absent in the alternatively spliced CDO1 variant 3 mRNA transcript (as indicated by ΔCAG in panel B).

To date, 148 validated non-synonymous CDO1 sequence variants have been added to the Ensembl variant database. These include 138 missense, 6 nonsense, 1 frameshift, 1 in-frame deletion and 2 splice site variants (Supplemental Tables 1 and 2). The nonsense and frameshift variants are most likely deleterious to CDO1 function, whereas the impact of missense variants on CDO1 function has not been extensively investigated. Using the SIFT, Poly-Phen, CADD, REVEL, MetaLR and MA tools, 68 missense variants are predicted to be deleterious to CDO1 structure and function (Supplemental Table 1, Fig. 1D). These variants are located throughout the CDO1 protein, with most of the predicted damaging variants clustering in 2 amino acid regions that contain β-sheets 1–5 and 8–13 (Fig. 1D). Substitution of amino acids (Q34P, Y58C, R60G and R60Q, W77C, H88N, C93R, C130Y, Y157S and Y157C, F161V) that directly interact with cysteine substrate, are predicted to be damaging to CDO1 protein function (Fig. 1D), suggesting the likely importance of these cysteine-binding residues for CDO1 function. In addition, the missense variant H88N that alters an Fe2+ interacting amino acid is predicted to be detrimental, suggesting the importance of this combined cysteine and Fe2+ ion-interacting amino acid for CDO1 function.

All of the non-synonymous CDO1 variants are relatively rare: 3 have an allelic frequency > 1 × 10−4 (intron 3 splice donor variant rs201589147 ~1.0 × 10−3, and missense variants D64Y ~2.6 × 10−4 and S42R ~1.0 × 10−3), whereas the other 135 variants have allelic frequencies <1 × 10−4 (Supplemental Tables 1 and 2). The low abundance of more common CDO1 missense variants (allelic frequency > 0.001) in the general population may suggest a harmful physiological effect from disrupting the CDO1 protein sequence. This was found to be the case for the Cdo1 knock-out mouse which has severe developmental defects and high postnatal mortality [6]. Taken together, these observations may explain why non-synonymous CDO1 variants have yet to be linked to any human pathology, which has led us to investigate the CDO1 gene and its mRNA expression as a step towards understanding its role in early development.

CDO1 mRNA expression and localization in placental tissue

Early studies used Northern blot analyses to show abundant CDO1 mRNA expression in term placental and adult human liver tissue, lower levels in heart and whole brain, and undetectable levels in lung, skeletal muscle and kidney [10]. Using the NCBI database, we confirmed these findings and extend the CDO1 mRNA expression profile to 11 additional adult tissues, with intermediate CDO1 mRNA levels in cerebellum and lower levels in prostate, adrenal gland, thyroid, trachea, uterus, thymus, stomach, spleen, salivary gland and small intestine (Fig. 2A). In addition, the NCBI database shows moderate CDO1 mRNA levels in fetal liver (22 to 40 weeks gestation) and fetal whole brain (20 to 33 weeks gestation). Since these data are derived from pooled fetal samples from a range of gestational ages, it is not possible to determine when fetal CDO1 mRNA expression increases in gestation. Our studies in mice showed negligible Cdo1 mRNA levels in whole fetal tissues at embryonic day 10.5 (E10.5) with levels increasing from E12.5 [8], suggesting an important role of CDO1 in the later gestational stages of fetal development. This is also relevant to the strong expression of CDO1 mRNA in term placenta (Fig. 2A). In the present study, we localised CDO1 mRNA to the syncytiotrophoblast layer (Fig. 2B), the site of nutrient and waste exchange between maternal and fetal circulation. However, this finding was unexpected as our previous study localised Cdo1 mRNA to the decidua but not the syncytiotrophoblast layers of the mouse placenta [8]. The reason for these observed differences between human CDO1 and mouse Cdo1 expression in the placenta are unknown but could reflect species differences in the organization of the trophoblast cell types, as well as the much longer gestational length for humans. The next phase of our research will investigate CDO1 mRNA expression in placentas obtained from preterm births to determine the gestational age when CDO1 mRNA abundance increases and whether its cellular localization changes during gestation.

Fig. 2

Tissue expression of CDO1 mRNA. (A) Relative CDO1 mRNA levels in 20 human tissues using data obtained from the online NCBI database http://www.ncbi.nlm.nih.gov/gene, which was accessed 19 January 2019. RPKM, Reads Per Kilobase of transcript, per Million mapped reads. (B) Cellular localization of CDO1 mRNA in human term placenta. Low (top panels) and high (bottom panels) magnification images of placental sections probed with antisense and sense (control) CDO1 probes. Images show antisense probe staining of CDO1 mRNA within the syncytiotrophoblast layer. Black scale bar = 100 μM. (C) Relative abundance of each placental CDO1 mRNA variant (V1-V4) abundance. (D) Comparison of each CDO1 variant (V1-V4) mRNA abundance in placentas of male (M) and female (F) babies (p > 0.05 for each variant). Data are mean ± SEM with n = 4–5 males and n = 3 females.

As mentioned above, there are 4 protein coding CDO1 mRNA variant sequences reported in the NCBI database. Using qPCR, we show that variant 2 is the most abundant CDO1 mRNA transcript in the term placenta, with lower levels of variant 4 and negligible levels of variants 1 and 3 (Fig. 2C), and no sex-specific differences (p > 0.05) in the abundance of each variant CDO1 mRNA (Fig. 2D). The higher abundance of variant 4 CDO1 mRNA when compared to variants 1 and 3 was unexpected as variant 4 encodes a truncated CDO1 protein (Fig. 1C) that is most likely unstable and inactive. Overall, these findings suggest that the splicing of exon 2 in variant 1, the alternative splicing at the 5′-end of exon 4 (ΔCAG) in variant 3, and the differential splicing of exon 1 to exon 3 in variant 4, are either inefficient when compared to variant 2 and/or that variant 1, 3 and 4 mRNA transcripts are unstable leading to their rapid mRNA decay. Further studies, such as expressing each variant CDO1 mRNA in cultured cells, are required to confirm whether mRNA stability contributes to the different level of each CDO1 mRNA transcript. Nonetheless, variant 2 appears to be the predominant CDO1 mRNA transcript, implying that isoform 2 is the major form of CDO1 protein in the placenta. To further understand the biological relevance of the CDO1 mRNA transcripts, the next phase of our research will investigate the relative abundance and activity of the placental CDO1 proteins, which is relevant to the post-translational regulation of CDO1 [21,22].

Phylogenetic comparison of CDO1 protein and the CDO1 5′-flanking region

Alignment of human CDO1 amino acid sequences with 19 orthologues showed the highest identity (88 to 93%) to hemochorial species, intermediate identity (64 to 90%) with endotheliochorial and epitheliochorial species, and lowest identity (49 to 59%) with non-placental species (Fig. 3A). To identify conserved DNA sequences that are potentially involved in the transcriptional control of human CDO1, the first 3500 nucleotides of the CDO1 5′-flanking region was compared among all 20 species. Of the 387 putative transcription factor binding motifs in the human CDO1 5′-flanking region (data not shown), 6 motifs exhibited similar spatial conservation among placental species as determined by a multiple species alignment (Fig. 3B): Kruppel like factor (KLF1) at position −63 to −81; myeloid zinc finger 1 (MZF1) at position −91 to −101; zinc finger protein 300 (ZNF300) at position −174 to −196; zinc finger protein 239 (ZNF239) at position −272 to −292; distal-less homeobox 1 (DLX1) at position −741 to −759; and CCAAT enhancer binding protein beta (CEBPB) at position −890 to −904 (Fig. 4A). The conserved location of these 6 putative transcription factor binding motifs in the 5′-flanking region of orthologous CDO1 genes of placental species led us to further investigate these DNA sequences as potential regulators of placental CDO1 mRNA expression.

Fig. 3

Phylogenetic analysis of CDO1 and its 5′-flanking region. (A) Phylogenetic tree with distances representing the evolutionary divergence of species that carry a CDO1 gene. Species above the dotted line represent the placental types: hemochorial (red), endotheliochorial (orange) and epitheliochorial (green). Image adapted from a tree generated online using the Interactive Tree of Life (ITOL, http://itol.embl.de/). (B) Multiple species alignment of conserved putative transcription factor binding motifs in the CDO1 5′-flanking region. Thirteen placental species share six highly conserved domains with human CDO1: CEBPB, DLX1, ZNF239, ZNF300, MZF1 and KLF1. Scale bar at top is relative to the +1 of the transcription start site.

Fig. 4

Sequence and transcriptional activity of the human CDO1 5′-flanking region. (A) CDO1 nucleotide sequence from −1030 to +70, with +1 and +52 (arrows) denoting the transcription initiation sites of CDO1 mRNA variants 1, 2, 3, 5, 6 and 7, and variants 4 and 8, respectively. Putative transcription factor binding motifs are boxed, with core sequences shaded. *Location of the -289C > G single nucleotide polymorphism (rs34869) reported in the NCBI SNP database. (B) Approximate mRNA expression level of each transcription factor in human placenta and the JEG-3 cell line using data obtained from the online NCBI UniGene EST database http://www.ncbi.nlm.nih.gov/est and the Gene Expression Atlas database https://www.ebi.ac.uk/gxa/home, respectively, which were accessed between 30 and 31 January 2019. TPM, Transcripts Per Million. (C) Transcriptional activity of the 5′-truncated CDO1 5′-flanking region assessed by luciferase activity in JEG-3 cells. Data are shown as fold-induction to luciferase activity of the pMetLuc2 reporter, n = 3. *P < 0.05 and ***P < 0.001 when compared to constructs containing the −1200 fragment. (D) Transcriptional activity of the pMetLuc2–1200 vector containing wild-type sequence (filled symbols) and individually mutated CEBPB motif and introduction of the SNP rs34869 (unfilled symbols). Data are shown as fold-induction to luciferase activity of the pMetLuc2 reporter, n = 3.

Analysis of the human CDO1 5′-flanking region

For the 6 putative motifs in the CDO1 5′-flanking region (Fig. 4A) to be involved in regulating placental CDO1 mRNA expression, the transcription factors that bind to these DNA sequences need to be expressed in the placenta. Indeed, we found this to be the case for CEBPB which is abundantly expressed in human placenta and the JEG-3 placental cell line (Fig. 4B). The other 5 transcription factors have either lower abundance (MZF1 and ZNF300) or undetectable (DLX1, ZNF239 and KLF1) levels in human term placenta (Fig. 4B), suggesting that these transcription factors may possibly not regulate CDO1 mRNA expression. To further localise the DNA sequences which are important for CDO1 promoter activity, a series of CDO1 5′-flanking regions were cloned upstream of the luciferase gene, transfected into JEG-3 cells and then assayed for luciferase activity (Fig. 4C). The highest luciferase expression was obtained with an upstream end of −1200, when compared to the construct with a 5′-end at −269 which had intermediate expression, and to the construct with a 5′-end at position −5 which had negligible luciferase expression. This finding suggests the possible presence of positive regulatory elements between regions −269 to −1200, and −5 to −269. Given the abundant expression of CEBPB in human term placenta and the JEG-3 cell line (Fig. 4B), and the location of a putative CEBPB binding motif within the CDO1 5′-flanking region that produced the highest luciferase expression, we mutated the core sequence of the CEBPB motif and compared luciferase activities to the control sequence (Fig. 4D). However, luciferase activity from the mutated CEBPB sequence was similar to the control sequence, suggesting that this site is unlikely to be responsible for the transcriptional activity of CDO1. We also searched the NCBI SNP database for CDO1 5′-flanking region genetic variants that occur in the general population, and which may potentially disrupt CDO1 mRNA expression. Of the 40 validated SNPs reported within 1200 nt of the CDO1 5′-flanking region (Supplemental Table 3), only 6 variants are located within a putative transcription binding motif reported in the present study (Fig. 4A): −70G > T (KLF1), −97G > A and -99G > T (MZF1), −193G > A (ZNF300) and -283C > T,G and -289C > G (ZNF239). The latter variant is relatively common (allelic frequency = 0.367) in the general population (Supplemental Table 3). Accordingly, we tested the luciferase activity from the mutated -289C > G sequence and found similar luciferase levels when compared to the control sequence (Fig. 4D) suggesting that this commonly occurring variant is unlikely to alter CDO1 mRNA expression. Whilst this study did not identify the specific transcription factor binding motifs that regulate CDO1 mRNA expression, our findings do suggest that regulatory sequences are most likely located within 1200 nucleotides of the CDO1 5′-flanking region.

Summary

The importance of sulfate in fetal growth and development cannot be overestimated. While interest in sulfate physiology continues to expand, there is still much to learn about the genes that maintain sulfate homeostasis in early development. It is remarkable that CDO1 has not yet been linked to any human pathology, despite its role in a major pathway of sulfate generation, as well as its link to severe developmental pathologies in laboratory mice. More than 2 decades have passed since the last description of the human CDO1 gene and its expression in human tissues. Accordingly, it is timely that we now provide an update on its gene, mRNA, promoter region and protein isoform structures. Our finding of CDO1 mRNA expression in the syncytiotrophoblast layer of the human term placenta warrants further studies to determine the physiological contribution of CDO1 to placental and fetal physiology. This study also collated a list of validated gene variants and assessed their predicted impact on CDO1 protein structure and function, providing valuable reference information for future genetic studies of CDO1 in human health.