The CpG island encompassing the promoter and first exon of human DNMT3L gene is a PcG/TrX response element (PRE)

Abstract

DNMT3L, a member of DNA methyltransferases family, is present only in mammals. As it provides specificity to the action of de novo methyltransferases, DNMT3A and DNMT3B and interacts with histone H3, DNMT3L has been invoked as the molecule that can read the histone code and translate it into DNA methylation. It plays an important role in the initiation of genomic imprints during gametogenesis and in nuclear reprogramming. With important functions attributed to it, it is imperative that the DNMT3L expression is tightly controlled. Previously, we had identified a CpG island within the human DNMT3L promoter and first exon that showed loss of DNA methylation in cancer samples. Here we show that this Differentially Methylated CpG island within DNMT3L (DNMT3L DMC) acts to repress transcription, is a Polycomb/Trithorax Response Element (PRE) and interacts with both PRC1 and PRC2 Polycomb repressive complexes. In addition, it adopts inactive chromatin conformation and is associated with other inactive chromatin-specific proteins like SUV39H1 and HP1. The presence of DNMT3L DMC also influences the adjacent promoter to adopt repressive histone post-translational modifications. Due to its association with multiple layers of repressive epigenetic modifications, we believe that PRE within the DNMT3L DMC is responsible for the tight regulation of DNMT3L expression and the aberrant epigenetic modifications of this region leading to DNMT3L overexpression could be the reason of nuclear programming during carcinogenesis.

Figure 1. Graphical representation of the alternate Transcription Start Sites for the human DNMT3L gene.

The human DNMT3L gene produces four transcripts (not drawn to scale) from three different Transcription start sites (TSS, raised arrows). Exons are shown as grey rectangles. Shown here in detail are two TSS which are responsible for the DNMT3L-001 & DNMT3L-201 (from +1) and DNMT3L-002 (+334) transcripts. Only DNMT3L-001 and 002 are shown in the representation. The most 5′ TSS has been taken as +1 position. Translation start site (shown as vertical line with ATG) is the same (at +952, in the second exon) for all these three DNMT3L transcripts. However, all the three differ in the length of the final DNMT3L protein. Protein made by DNMT3L-001 and DNMT3L-201 are 387 aa and 386 aa long respectively, DNMT3L-003 is 181 aa and DNMT3L-002 makes a 288 aa long protein. The fragments 3L-L and 3L-S from this region were analysed for their function by reporter gene assays in the present study. The approximate positions of the CpG dinucleotides within these fragments are shown with filled circles. All nucleotide positions mentioned are relative to the most 5′ TSS site.

Figure 2. Functional analysis of DNMT3L DMC by reporter gene assay in transiently transfected mammalian cells.

A. Graphical representation of the various constructs transfected into HEK293 cells. Shown for each construct is the reporter AcGFP gene, its CMV promoter, the selection Kan^R/Neo^R marker, the various DNMT3L DMC fragments, the control Chr1 fragment and the H19 ICR inserted upstream of the CMV promoter. B. Relative transcriptional level of the reporter AcGFP gene in the various constructs. The transcriptional levels were measured by Real-Time PCR and calculated relative to the AcGFP expression in the control CMV only construct. Kan ^R/Neo ^R expression level was used to control for transfection efficiency in the Real-Time PCR analysis. C. A representative Western Blot analysis showing AcGFP protein expression for HEK293 cells transfected with the various constructs mentioned in A. β-TUBULIN was used as a loading control. Luciferase activity for the pG5luc vector co-transfected with the above mentioned constructs is shown below the β-TUBULIN panel. D. Relative AcGFP protein expression levels for the cells transfected with the various constructs. Error bars represent Standard Deviation (S.D.). Asterisks indicate significant difference (Student's t test, * - p<0.05, ** - p<0.01, *** - p<0.001).

Figure 3. Identification of the minimal region within the DNMT3L DMC that can repress reporter gene transcription.

A. Graphical representation of the smaller fragments of the 3L-S region which were cloned into the AcGFP reporter gene vector and transfected into HEK293 cells (see figure 2A for the details of the vector). Base-pair position of each fragment with respect to the 3L-S fragment is given within brackets. Vertical lines denote the minimal region that can repress reporter gene transcription. Subfragments of 3L-S have been named s1, s2, s3 and s4. B & C- Quantitative RT-PCR was performed to determine relative transcriptional level of the reporter AcGFP gene in the various constructs. The relative transcriptional levels were measured with respect to the AcGFP expression in the CMV only construct. See figure 2A for information on the Control and H19 constructs. Error bars represent Standard Deviation (S.D.). Asterisks indicate significant difference (Student's t test, * - p<0.05, *** - p<0.001, *** - p<0.005).

Figure 4. Functional analysis of DNMT3L DMC by reporter transgene assay in Drosophila.

A. The three panels show comparison of eye color phenotype between representative DNMT3L DMC transgenic lines and their counterpart lines where the inserted 3L-L region had been flipped out using the surrounding loxp sites. P/+, heterozygous 3L-L transgenic lines; ΔP/+ their respective flipped-out counterparts. B. Graphical representation of the reporter gene construct pCaSPeR-3L-L and it's flipped out counterpart (Δ3LL). The construct contains the mini-white reporter gene under the control of hsp70 promoter. The 3L-L region (shown as unfilled arrow) was inserted upstream of the hsp70 promoter between the loxP sites (filled arrows). Only one loxP site remains in the flipped out counterpart. 5′P and 3′P refer to P element present 5′ and 3′ to the reporter gene. C. Comparison of eye color pigmentation between representative 3L-L transgenic lines and their flipped out counterparts. Numbers below the X-axis denote names of individual transgenic lines. White bars represent P/+ lines whereas black bars represent their ΔP/+ flipped out counterparts. Error bars represent Standard Deviation (S.D.). Asterisks indicate significant difference (Student's t test, * - p<0.05, ** - p<0.01, *** - p<0.001).

Figure 5. DNMT3L DMC interacts with Polycomb group of proteins.

A. The 15 panels show comparison of eye color phenotype between representative DNMT3L DMC transgenic lines (25.2.12) and their counterpart lines after crossing with the respective Polyomb (Pc¹, esc², Asx^XF53, Psc′, Pcl^T1, Scm^R5-13B, Phd, Pho), Trithorax (Ash2¹, Mor¹, Bbrm², Trl^R85, Trx¹) and Supressor of Variegation (Su(var)2-5⁰¹, Su(var)3-9⁰⁶) mutants. 3LL/+, heterozygous 3L-L transgenic lines; 3LL/−, their respective counterparts after crosses with the respective mutant (− is the name of the Polycomb, Trithorax or Suvar mutant). B. Comparison of eye color pigmentation between 3L-L transgenic lines (25.2.12 and 25.2.29) and their counterparts from crosses with the various mutant lines. Each bar represents eye color pigmentation for progeny from crosses of individual transgenic lines with a particular mutant, the details of which are provided below the X-axis. As the assays were done in batches, the eye pigmentation for the control 3L-L transgenic line was done for each batch and is shown as white bars (P/+ male with W¹¹¹⁸) in the graphs. Error bars represent Standard Deviation (S.D.). Asterisks indicate significant difference (Student's t test, * - p<0.05, ** - p<0.01).

Figure 6. Epigenetic profile of DNMT3L DMC and the associated promoter in the reporter gene assay.

A&B. Histone modifications associated with DNMT3L DMC and the hsp70 promoter in the transgene reporter assay in Drosophila. Histone ChIP analysis for the 3L-L region (A) and the CMV promoter (B) in the Drosophila transgene assay. Chromatin immunoprecipitation was carried out on the 25.2.12 transgenic line with the indicated histone H3 modifications, followed by quantitative Real-time PCR. Comparison of histone modifications associated with the hsp70 promoter in the 3L-L (3LL, black bars) transgenic lines and their flipped out counterparts (Δ3LL, white bars). The H3 histone modifications examined are mentioned below the X-axis. Enrichment in the bound fraction is represented as percentage of Input. IgG - control ChIP with rabbit IgG. Error bars represent Standard Deviation (S.D.). Asterisks indicate significant difference (Student's t test, * - p<0.05, ** - p<0.01, *** - p<0.005).

Figure 7. Epigenetic profile of the 3L-L region and CMV promoter in transient transfection assay.

A–C. Effect of DNMT3L DMC on the interaction between CMV promoter and Polycomb, Trithorax and Suvar proteins in mammalian cells. ChIP analysis for the CMV promoter in control (referred to as CMV) and 3L-L region (CMV-3LL) containing mammalian transient transfection constructs (see figure 2A for graphical representation of the constructs). Quantitation was done by Real-time PCR. Enrichment of the CMV region in the bound fraction is represented as percentage of Input. The names of the chromatin proteins examined are mentioned below the X-axis. D–G. Histone modifications and DNA methylation profile of the 3L-L region in the AcGFP reporter gene construct after transient transfection in HEK293 cells. C- ChIP analysis for the 3L-L region with the indicated histone H3 modifications. D- Comparison of histone modifications associated with the CMV promoter in CMV only (vector) and 3L-L (3LL) constructs after transfection into HEK293 cells. Enrichment in the bound fraction is represented as percentage of Input. IgG - control ChIP with rabbit IgG. Error bars represent Standard Deviation (S.D.). Asterisks indicate significant difference (Student's t test, * - p<0.05, ** - p<0.01, *** - p<0.005). F, G- Methylation profile of the 3L-L region (F) and CMV promoter (G) in the construct after bisulfite sequencing. Compare the methylation profile of the CMV promoter in CMV only (vector) and the 3L-L (3LL) constructs. Each horizontal line indicates a single clone from the 3L-L PCR products after bisulfite treatment. Circles denote CpG dinucleotides present within the sequence. The positions are not drawn to scale. Open circles indicate no methylation. Filled circles represent methylated cytosine.

Figure 8. Epigenetic profile of endogenous DNMT3L DMC.

A- Histone modifications; B. DNA methylation profile of the 3L-L region at the endogenous DNMT3L locus in HEK293 cells. C. ChIP analysis for the 3L-L region at the endogenous DNMT3L locus using antibodies to the various Polycomb, Trithorax and Suvar proteins mentioned below the X-Axis. Enrichment in the bound fraction is represented as percentage of Input. IgG - control ChIP with rabbit IgG. Error bars represent Standard Deviation (S.D.). Asterisks indicate significant difference (Student's t test, * - p<0.05, *** - p<0.001, *** - p<0.005).

Figure 9. Knock-down of Polycomb proteins and its effect on transcription of the DNMT3L gene from the endogenous locus.

A. siRNA mediated transcriptional repression of polycomb proteins. mRNA levels of the indicated Polycomb genes was assayed in presence of scrambled or specific siRNA in HEK293 cells. % mRNA level was calculated with respect to untransfected cells. B. DNMT3L gene expression from the endogenous locus in Polycomb siRNA transfected or untransfected HEK293 cells was quantitated by Real-time RT-PCR. Relative expression was calculated with respect to untransfected cells. The genes for which siRNA was used in our assay are mentioned below the X-axis. Error bars represent Standard Deviation (S.D.). Asterisks indicate significant difference (Student's t test, * - p<0.05, *** - p<0.001, *** - p<0.005).