A transposon in Comt generates mRNA variants and causes widespread expression and behavioral differences among mice

Abstract

Background: Catechol-O-methyltransferase (COMT) is a key enzyme responsible for the degradation of dopamine and norepinephrine. COMT activity influences cognitive and emotional states in humans and aggression and drug responses in mice. This study identifies the key sequence variant that leads to differences in Comt mRNA and protein levels among mice, and that modulates synaptic function and pharmacological and behavioral traits.

Methodology/principal findings: We examined Comt expression in multiple tissues in over 100 diverse strains and several genetic crosses. Differences in expression map back to Comt and are generated by a 230 nt insertion of a B2 short interspersed element (B2 SINE) in the proximal 3' UTR of Comt in C57BL/6J. This transposon introduces a premature polyadenylation signal and creates a short 3' UTR isoform. The B2 SINE is shared by a subset of strains, including C57BL/6J, A/J, BALB/cByJ, and AKR/J, but is absent in others, including DBA/2J, FVB/NJ, SJL/J, and wild subspecies. The short isoform is associated with increased protein expression in prefrontal cortex and hippocampus relative to the longer ancestral isoform. The Comt variant causes downstream differences in the expression of genes involved in synaptic function, and also modulates phenotypes such as dopamine D1 and D2 receptor binding and pharmacological responses to haloperidol.

Conclusions/significance: We have precisely defined the B2 SINE as the source of variation in Comt and demonstrated that a transposon in a 3' UTR can alter mRNA isoform use and modulate behavior. The recent fixation of the variant in a subset of strains may have contributed to the rapid divergence of inbred strains.

Figure 1. Cis modulation of Comt expression by B and D alleles in different tissues.

(A) Orientation of probe sets relative to the Comt gene based on the UCSC Genome Browser Mouse July 2007 Assembly (mm9). Probe sets 1418701_at and 1449183_at are associated with mRNA containing distal 3′ UTR and the last two coding exons and proximal 3′ UTR of Comt mRNA, respectively. (B) Cis-modulation of Comt expression visualized using the QTL heatmap tool available at www.genenetwork.org. Coloration at 18.408 Mb indicates association between genotype at this locus and gene expression. No color means there is not an association., whereas blue and red shading indicates whether B6 (B) or D2 (D) alleles, respectively, are associated with higher expression. Yellow arrowheads indicate the physical location of Comt. Among the BXD family the D allele is associated with higher expression of the distal 3′ UTR of Comt in all tissues. However, expression of the common coding mRNA probe set is associated with higher expression of the B allele in whole brain (WB), cerebellum (CER), hippocampus (HIP) and striatum (STR). In the kidney (KID) and the prefrontal cortex (PFC), expression of coding Comt mRNA is also associated with higher expression of the D allele. (C) Multiple QTL maps for Comt coding exons. Strong cis-regulation of expression is observed for all tissues except for the EYE (eye), VTA (ventral tegmental area), and NAC (nucleus accumbens). Purple arrowheads indicate the approximate location of Comt. (D) Multiple QTL maps for Comt distal 3′ UTR. Significant cis-regulation of Comt is observed across all tissues.

Figure 2. Expression of Comt coding exon and distal 3′ UTR in the BXD population across tissues.

Probe sets representing either coding exon or distal 3′ UTR were used to measure Comt expression in parental and BXD strains. Strain allelic expression for both Comt probe sets was determined based on the genotype (B6 or D2) of nearby marker rs4165069 (located on chromosome 16 at 17.577 Mb). Average expression for B6 (B) and D2 alleles (D) was determined for each probe set across tissues. (Top panel)There is less variation between alleles for Comt exon expression across multiple tissues as assayed by probe set 1449183_at. (Bottom Panel) In contrast, there is much greater variation when assayed using the distal 3′UTR (1418701_at). Higher expression of the distal 3′ UTR is consistently associated with the D allele while higher expression is associated with the B allele for Comt coding exons in the whole brain (WB), cerebellum (CER), hippocampus (HIP), and striatum (STR). Higher expression of coding exons is associated with the D allele in the prefrontal cortex (PFC). Expression of coding exons in the ventral tegmental area (VTA), nucleus accumbens (NAC), eye (EYE), and kidney (KID) is not strongly dependent on genotype.

Figure 3. C57BL/6J and DBA/2J are identical by descent for a large region on chromosome 16 surrounding Comt.

(A) Next-generation sequencing (red) reveals previously unknown SNPs between B6 and D2. The original SNP panel (blue) was obtained at www.ncbi.nlm.nih.gov/projects/SNP/. A large interval surrounding Comt from ∼15 to 23 Mb is devoid of SNPs. (B) There are no known SNPs between B6 and D2.

Figure 4. A 230 bp insertion in the proximal 3′ UTR of Comt is present in multiple strains.

PCR amplification of genomic DNA using primers flanking the 3′UTR (18,407,151 to 18,407,758 Mb) generates two different products. Strains containing the B2 SINE, such as B6, generate a larger PCR product (approximately 600 bp) then do strains lacking the SINE, such as D2 (approximately 400 bp). Genomic DNA from reciprocal hybrids (B6D2F1 and D2B6F1), which are heterozygous for the B2 SINE indel, generates both a short and long product due to inheritance of both B6 and D2 alleles, respectively.

Figure 5. Insertion of a B2 SINE containing an alternative polyadenylation site into the Comt gene leads to the production of mRNA containing a shorter 3′ UTR in B6.

(A) Gel electrophoresis of 3′ RACE products shows that D2 produces mRNA containing a 3′ UTR that is ∼200 nts longer compared to B6. This is consistent with the truncation of B6 mRNA due to the introduction of the B2 SINE polyadenylation signal (A). Expression data indicate that a small amount of the long 3′ UTR produced in most tissues of strains with the insertion, likely due to read-through of the B2 SINE polyadenylation signal. (B) Black shaded text indicates identical proximal 3′ UTR DNA sequence in B6 and D2 strains. Bold text and grey shaded text indicate the position of a poly (A) tail and polyadenylation signals, respectively. The B2 SINE (sequence in center box) contains both A and B box RNA polymerase III promoters (asterisks and underlined text, respectively), a consensus RNA polymerase III termination signal (underlined text), a polyadenylation signal arranged in an overlapping array (bold text, shaded in grey), and an adenine- (A-) rich 3′ flanking region (bold text). Arrows indicate the approximate position of both a custom Comt forward primer (right arrow) and a generic reverse primer (left arrow) used to amplify 3′ RACE products from cDNA for each strain. The exact length of the poly-A tail for each mRNA species from B6 and D2 is not known. Genomic positions based on the B6 reference strain (UCSC Genome Browser on Mouse July 2007 Assembly mm9).

Figure 6. The level of COMT protein differs between the B6 and D2 strains.

Lanes 1 through 3 and 4 through 6 represent biological replicates from B6 and D2, respectively. Upper and lower panels show protein levels in the prefrontal cortex and hippocampus, respectively. Actin levels are measured as a loading control. Two protein isoforms of COMT are expressed in mice and recognized by the COMT antibody; the longer membrane bound (MB)-COMT and the smaller soluble (S)-COMT. Both isoforms have higher expression in B6 compared to D2 in the prefrontal cortex (PFC) and hippocampus (HIP). Based on the amount of actin detected in the hippocampus it appears that there may be a slightly higher amount of protein in the D2 samples, however, the expression of MB-COMT and S-COMT is still higher in B6 compared to D2.

Figure 7. Comt regulatory model.

Insertion of a B2 SINE causes 3′ UTR length isoforms with important functional consequences. Two concentric rings show the downstream effect Comt sequence variation has on the gene expression (white ring) and phenotypes (yellow ring). Variation in Comt influences several neurotransmitter systems, including GABA and dopamine, and modulates the expression of genes involved in numerous biological processes, including receptor trafficking at excitatory synapses (Palm, Sqstm1, Nsg1, Akap9, and Apba1).