Promoter-like sequences regulating transcriptional activity in neurexin and neuroligin genes

Abstract

Synapse function requires the cell-adhesion molecules neurexins (Nrxn) and neuroligins (Nlgn). Although these molecules are essential for neurotransmission and prefer distinct isoform combinations for interaction, little is known about their transcriptional regulation. Here, we started to explore this important aspect because expression of Nrxn1-3 and Nlgn1-3 genes is altered in mice lacking the transcriptional regulator methyl-CpG-binding protein2 (MeCP2). Since MeCP2 can bind to methylated CpG-dinucleotides and Nrxn/Nlgn contain CpG-islands, we tested genomic sequences for transcriptional activity in reporter gene assays. We found that their influence on transcription are differentially activating or inhibiting. As we observed an activity difference between heterologous and neuronal cell lines for distinct Nrxn1 and Nlgn2 sequences, we dissected their putative promoter regions. In both genes, we identify regions in exon1 that can induce transcription, in addition to the alternative transcriptional start points in exon2. While the 5'-regions of Nrxn1 and Nlgn2 contain two CpG-rich elements that show distinct methylation frequency and binding to MeCP2, other regions may act independently of this transcriptional regulator. These data provide first insights into regulatory sequences of Nrxn and Nlgn genes that may represent an important aspect of their function at synapses in health and disease.

Keywords: CpG island; MeCP2; neurexin; neuroligin; promoter; transcriptional activity.

Figure 1

Expression of Nrxn and Nlgn genes in methyl-CpG-binding protein2 (MeCP2) knockout (KO) brains. (a) RT-qPCR analysis of α-Nrxn and β-Nrxn variants from Nrxn1-3 in WT and MeCP2 KO brains at post-natal days (P) 7 (light green) and P20 (dark green). mRNA levels are expressed as a ratio of KO relative to WT (= 1, dashed line), revealing a decrease at day P7 and an increase at day P20 in KO. (b) Similar RT-qPCR analysis as in (a) for Nlgn1-3 genes showing a similar tendency as for Nrxn (P7, light red; P20, dark red). Data are means ± sem (n = 3–6 brains/genotype and age). Significance was determined by Student's t-test, significance levels are indicated as *p < 0.05, **p < 0.01, ***p < 0.001, ns = non-significant.

Figure 2

Nlgn protein levels in methyl-CpG-binding protein2 (MeCP2) knockout (KO) brains. (a) Immunoblots of brain lysates from WT and MeCP2 KO mice at P7 and P20 probed with antibodies specific to Nlgn1, -2 and -3, and HSP70 as control. (b) Quantification of protein levels of the three Nlgn variants in MeCP2 KO brains relative to WT samples (= 1, dashed line; 3–4 brains/genotype and age) at P7 (light red) and P20 (dark red). Bar graphs represent normalized protein levels from at least five independent experiments as shown in (a). Data are means ± SEM. Significance was determined by Student's t-test, significance levels are indicated as *p < 0.05, **p < 0.01, ***p < 0.001, ns = non-significant.

Figure 3

Regulatory sequences in Nrxn1-3. (a) Scheme showing exon/intron organization of all three murine Nrxn genes. Representation is to scale with the exception of size of exons (vertical black bars). Arrowheads indicate translation start of α-Nrxn (left) and β-Nrxn (right) variants. Regulatory sequences were predicted by in silico analysis and test regions cloned (red bars, GS1-11), including potential promoter regions upstream of transcriptional starting sites and/or CpG-rich sequences. (b) Results of a dual luciferase reporter gene assay in HEK293 (white bars) and PC12 cells (dark gray bars) used to determine regulatory activity of genomic test sequences (GS numbers as in a). Data are means ± SEM; values for bar graphs are in relation to control vector and blotted logarithmically (> 1 = activation, < 1 = inhibition).

Figure 4

Dissection of transcriptionally active sequences in Nrxn1. Black arrowheads (left) represent position and length of genomic sequences cloned to identify the putative Nrxn1α promoter region. Fragments tested are shown in relation to exon1 and exon2 of the Nrxn1 gene depicted above. Red arrow = translation start in exon2, red box in exon1 = novel transcription start site. Three groups of bar graphs show quantitative results of luciferase assays in HEK293 (white), undifferentiated PC12 (dark gray), and differentiated PC12 cells (light gray). Increased translational activity is restricted to sequences containing exon1. Data are means ± SEM (units: relative luminescence); values calculated in relation to control vector (= 1).

Figure 5

Methylation frequency and methyl-CpG-binding protein2 (MeCP2) binding at exon1 and exon2 sequences of Nrxn1. Schematic representation of the 5′-region of Nrxn1 (black arrowhead = translation start point, black boxes = UTR). Bioinformatics predict that exon1 and exon2 contain different numbers of CpG-dinucleotides (numbered boxes). Green lines represent PCR products used for methylation-specific PCRs (MSP) on bi-sulfite converted genomic DNA (red shapes indicate the relative methylation frequency observed). Blue lines denote position of PCR products amplified after chromatin-immunoprecipitation with anti-MeCP2 (ChIP-PCR), shown on sample gel images below (B = brain, L = liver tissue as templates). Positive signals of ChIP-PCR could only be found for MSP-B sequences (right lanes) that are more methylated than MSP-A sequences.

Figure 6

Regulatory sequences in Nlgn1-3. (a) Scheme showing exon/intron organization of Nlgn1-3 genes. Representation is to scale and exons indicated by black bars (arrowheads = translation start, small boxes = UTR). Putative regulatory sequences upstream of transcription start point (TSS) in first exons were predicted by in silico analysis (red bars, GS12-15); GS14 marks the TSS of an alternatively spliced Nlgn2 transcript. (b) Quantitative results from dual luciferase assays in HEK293 (white bars), PC12 (dark gray bars), and nerve growth factor-differentiated PC12 cells (light gray bars) used to probe the regulatory activity of genomic sequences tested (see a). Note that exon1 sequences of Nlgn2 (GS13) exert differential effects on transcriptional activity in HEK293 compared to PC12 cells. Data are means ± SEM; values for bar graphs are in relation to control vector and blotted logarithmically (> 1 = activation, < 1 = inhibition).

Figure 7

Bidirectional activity driving expression of Nlgn2 and RIK18100 genes. The 5′-region of Nlgn2 is organized in a head-to-head arrangement with an uncharacterized gene (RIK18100) on mouse chromosome 11. Their transcription start point (TSS) (red arrows) point in opposite directions and are separated by 751 bp. Arrowheads (left) indicate length, position, and orientation of genomic fragments that were cloned and tested in luciferase assays. Bar graphs show quantitative results for their regulatory activity in HEK293 (white), PC12 (dark gray), and differentiated (light gray) PC12 cells. Strongest transcriptional activity in Nlgn2 orientation resides upstream of exon1 overlapping with RIK18100 sequences, whereas optimal activity in RIK18100 direction is confined to a small fragment upstream of Nlgn2 exon1 (red arrowheads). Data are means ± SEM (units: relative luminescence); values calculated in relation to control vector (= 1).

Figure 8

Methylation frequency and methyl-CpG-binding protein2 (MeCP2) binding of Nlgn2 promoter sequences affect transcriptional activity. (a) Schematic representation of the Nlgn2/18100RIK head-to-head arrangement (black arrowheads = translation start, black boxes = UTR). Bioinformatics predict that exon1 and exon2 of Nlgn2 contain different numbers of CpG-dinucleotides (numbered boxes). Green lines represent PCR products used for methylation-specific PCRs (MSP1-3) on bi-sulfite converted genomic DNA from PC12 cells (red shapes indicate relative methylation frequency observed). Blue lines denote position of PCR products amplified after chromatin-immunoprecipitation with anti-MeCP2 (ChIP-PCR), shown on sample gel images below (lanes 1, 4: PC12 cells; lanes 2, 5: differentiated PC12; lanes 3, 6: brain DNA as control). (b) Results of luciferase assays using in vitro-methylated constructs that correspond to MSP1 and MSP2 fragments. In vitro methylation has an effect on MSP1 but not on exon2-containing MSP2, reflecting its strong native methylation shown in (a). Data are means ± SEM; values calculated in relation to control vector (= 1). Significance was determined by Student's t-test, significance levels are indicated as *p < 0.05, ns = non-significant.