l-Isoaspartyl Methyltransferase Deficiency in Zebrafish Leads to Impaired Calcium Signaling in the Brain

Abstract

Isomerization of l-aspartyl and l-asparaginyl residues to l-isoaspartyl residues is one type of protein damage that can occur under physiological conditions and leads to conformational changes, loss of function, and enhanced protein degradation. Protein l-isoaspartyl methyltransferase (PCMT) is a repair enzyme whose action initiates the reconversion of abnormal l-isoaspartyl residues to normal l-aspartyl residues in proteins. Many lines of evidence support a crucial role for PCMT in the brain, but the mechanisms involved remain poorly understood. Here, we investigated PCMT activity and function in zebrafish, a vertebrate model that is particularly well-suited to analyze brain function using a variety of techniques. We characterized the expression products of the zebrafish PCMT homologous genes pcmt and pcmtl. Both zebrafish proteins showed a robust l-isoaspartyl methyltransferase activity and highest mRNA transcript levels were found in brain and testes. Zebrafish morphant larvae with a knockdown in both the pcmt and pcmtl genes showed pronounced morphological abnormalities, decreased survival, and increased isoaspartyl levels. Interestingly, we identified a profound perturbation of brain calcium homeostasis in these morphants. An abnormal calcium response upon ATP stimulation was also observed in mouse hippocampal HT22 cells knocked out for Pcmt1. This work shows that zebrafish is a promising model to unravel further facets of PCMT function and demonstrates, for the first time in vivo, that PCMT plays a pivotal role in the regulation of calcium fluxes.

Keywords: HT22 cells; calcium signaling; isoaspartyl; protein repair; zebrafish.

Figure 1

Sequence alignment of human PCMT1, mouse Pcmt1, and the zebrafish homologs Pcmt and Pcmtl. The sequences were aligned with the CLUSTAL Omega software (version 1.2.4) (Sievers et al., 2011) using default settings. Conserved methyltransferase domains are highlighted in yellow [AdoMet I–III, S-adenosyl-L-methionine-binding domains; pre-I and post-III, isoaspartyl methyltransferase-specific domains (Kagan et al., 1997a)]. Residues in red were found to directly interact with S-adenosylhomocysteine from crystallography data (Smith et al., 2002). The human polymorphic Ile120Val site is highlighted in green; the zebrafish Pcmt and Pcmtl sequences show an Ile to Val substitution in this same position. (*) Fully conserved residues, (:) residues with strongly similar properties, (.) residues with weakly similar properties. Sequence accession numbers: zebrafish Pcmt, NP_571540.1; zebrafish Pcmtl, NP_957062.1; human PCMT1, NP_001347385.1; mouse Pcmt1, NP_032812.2 (residues 59–283).

Figure 2

Developmental expression and adult tissue distribution of the pcmt and pcmtl transcripts in zebrafish. Total RNA was extracted from embryos (n = 3) at the indicated developmental stages (A,B) and from organs (n = 3) of 2-year old male (C,D) and female (E,F) zebrafish for quantification of the relative pcmt (black bars) and pcmtl (gray bars) transcript levels using qPCR. Relative expression levels (eef1a1l1 was used as reference gene) have been normalized to the expression at 7 hpf (A,B) or in the brain (C–F). Data shown are means ± SEMs for the 3 biological replicates.

Figure 3

pcmt and pcmtl morpholino-mediated knockdown efficiency. Splice-blocking MOs (5 ng) targeting pcmt (pcmt-i2e3 and pcmt-e5i5) or pcmtl (pcmtl-e1i1 and pcmtl-e4i4) or a non-targeting control MO (Ctrl MO) were microinjected into zebrafish embryos (1-cell stage) and total RNA was extracted at 1 dpf for cDNA synthesis and PCR amplification using primers spanning the targeted intron-exon boundaries. Agarose gel analysis showed very low efficiency for the pcmt-i2e3 and pcmtl-e1i1 MOs (asterisks), but high efficiency for the pcmt-e5i5 and pcmtl-e4i4 MOs leading to exon skipping as indicated by the smaller amplicon sizes (arrows).

Figure 4

Knockdown of pcmt and pcmtl leads to lower isoaspartyl methyltransferase activity and higher isoaspartyl levels in zebrafish larvae. Proteins were extracted from uninjected, control MO-injected and pcmt/l morpholino-injected larvae and isoaspartyl methyltransferase activity (A), as well as isoaspartyl levels (B), were assayed using the methanol vapor diffusion assay. The protein extracts were also analyzed after labeling in the presence of recombinant human PCMT1 (rhPCMT1) and tritiated SAM by SDS-PAGE (C) followed by fluorography (D). The values shown in panels (A) and (B) are means ± SEMs of 4 biological replicates (ns, not significant; *p-value < 0.05; ***p-value < 0.001). For the SDS-PAGE and fluorography analyses, one representative experiment out of 4 biological replicates is shown. M, molecular weight markers.

Figure 5

Knockdown of pcmt and pcmtl leads to morphological changes and developmental delay. The morphological changes were ranked from mild to severe in 2 dpf larvae [(A–D); trunk curvature and decreased pigmentation) and in 4 dpf larvae (E–H); lack of the swim bladder, body curvature, and strong malformation]. Relative quantification of these phenotypes was performed by observation of the indicated numbers of pcmt/l morphants compared to control and rescue larvae (exogenous expression of pcmt/l via mRNA or Tol2 transgenesis in the morphant background) at 2 dpf (I) and 4 dpf (J). Arrows indicate the internal brain structure showing smaller brain size in the pcmt/l morphant larvae (L) compared to control MO-injected larvae (K).

Figure 6

Knockdown of pcmt and pcmtl leads to disruption of calcium fluxes in the brain of zebrafish larvae. Calcium fluxes in response to PTZ exposure were imaged by fluorescence microscopy in the brain region of beta-actin:GCaMP6f larvae injected with control MO (A), pcmt/l MOs (B), and with co-injection of pcmt/l MOs and a plasmid for wild-type Pcmt and Pcmtl (C) or of pcmt/l MOs and a plasmid for catalytically inactive Pcmt and Pcmtl (D83V) (D). The 0 min time point coincides with the start of calcium imaging after a 30-min pre-incubation with PTZ. The traces shown are representative for 7 independent replicates.

Figure 7

Pcmt1 knockout in mouse hippocampal HT22 cells by CRISPR/Cas9 technology. Confirmation of the expected 40-bp deletion in the HT22 Pcmt1 KO cell line by PCR amplification of the Pcmt1 target site from genomic DNA (A). The generated KO cell line was also validated by Western blotting using primary antibodies against the Pcmt1 protein (B), isoaspartyl methyltransferase activity assay (C), and measurement of isoaspartyl levels in crude protein extracts (D). The values shown are means ± SEMs of 6 biological replicates (****p < 0.0001).

Figure 8

Impaired calcium response in Pcmt1 knockout HT22 cells. Control or Pcmt1 KO HT22 cells, stained with Fluo-4 AM, were exposed to 20 μM ATP, followed by a second stimulation with 40 μM ATP at the indicated times, and imaged by fluorescence microscopy in a controlled environment (A). The calcium spike observed after the first ATP stimulation was analyzed for peak area (B), peak amplitude (C), and peak width (D). The fluorescence traces shown are representative of 3 imaged replicate wells. The values shown are means ± SEMs of signals extracted from 150 cells (**p-value < 0.001; ****p-value < 0.0001).