Transcriptome analysis indicates dominant effects on ribosome and mitochondrial function of a premature termination codon mutation in the zebrafish gene psen2

PRESENILIN 2 (PSEN2) is one of the genes mutated in early onset familial Alzheimer’s disease (EOfAD). PSEN2 shares significant amino acid sequence identity with another EOfAD-related gene PRESENILIN 1 (PSEN1), and partial functional redundancy is seen between these two genes. However, the complete range of functions of PSEN1 and PSEN2 is not yet understood. In this study, we performed targeted mutagenesis of the zebrafish psen2 gene to generate a premature termination codon close downstream of the translation start with the intention of creating a null mutation. Homozygotes for this mutation, psen2S4Ter, are viable and fertile, and adults do not show any gross psen2-dependent pigmentation defects, arguing against significant loss of γ-secretase activity. Also, assessment of the numbers of Dorsal Longitudinal Ascending (DoLA) interneurons that are responsive to psen2 but not psen1 activity during embryogenesis did not reveal decreased psen2 function. Transcripts containing the S4Ter mutation show no evidence of destabilization by nonsense-mediated decay. Forced expression in zebrafish embryos of fusions of psen2S4Ter 5’ mRNA sequences with sequence encoding enhanced green fluorescent protein (EGFP) indicated that the psen2S4Ter mutation permits utilization of cryptic, novel downstream translation start codons. These likely initiate translation of N-terminally truncated Psen2 proteins lacking late endosomal/lysosomal localization sequences and that obey the “reading frame preservation rule” of PRESENILIN EOfAD mutations. Transcriptome analysis of entire brains from a 6-month-old family of wild type, heterozygous and homozygous psen2S4Ter female siblings revealed profoundly dominant effects on gene expression likely indicating changes in ribosomal, mitochondrial, and anion transport functions.

Introduction 72 73 PRESENILIN 2 (PSEN2) was first identified as a candidate locus for mutations causing 74 familial Alzheimer's disease (AD) with early onset (EOfAD) when a point mutation 75 resulting in the substitution of an isoleucine residue for an asparagine residue (N141I) 76 was found in a Volga German AD family in 1995 (1). PSEN2 is similar in structure to 77 the major EOfAD gene PRESENILIN 1 (PSEN1) and the two genes encode proteins 78 with 62% amino acid sequence identity (2). The age of onset of Alzheimer's disease 79 (AD) caused by mutations in PSEN2 ranges from 39 to 75 years, which overlaps both 80 with PSEN1 EOfAD-associated mutation disease onset ages and with late onset, 81 sporadic AD (3). The later mean onset age of AD caused by PSEN2 mutations 82 compared to mutations in PSEN1 is still unexplained, but some studies suggest that it 83 may be caused by the partial replacement of PSEN2 function by PSEN1 (4). However, 84 the functions of PSEN1 and PSEN2 have not yet been determined comprehensively. 85 Moreover, despite the partial functional redundancy between PSEN1 and PSEN2, in 86 vitro studies have shown that the protein products of the two genes also play divergent 87 roles in cellular physiology ((5, 6) and reviewed in (7)).

89
Both PSEN1 and PSEN2 proteins are components of γ-secretase complexes. The 90 absence of PSEN1 is thought to reduce γ-secretase activity in mammalian cells (8,9), 91 while the absence of both PSEN1 and PSEN2 is thought to eliminate it completely (10, 92 11) although some data does not agree with this (reviewed in Jayne et al. (12)). In mice,      As part of an effort using zebrafish to identify the specific cellular changes caused by 149 EOfAD-like mutations in these genes, we wished to examine null mutations so that 150 their effects could be excluded from consideration. In this paper we describe an 151 unsuccessful attempt to generate a null mutation of the zebrafish orthologue of the 152 human PSEN2 gene, psen2, by introduction of a premature termination codon 153 downstream of the asssumed translation start codon. Unexpectedly, the mutation 154 appears to force utilization of downsteam methionine codons for translation initiation.

155
This generates N-terminally truncated proteins that act dominantly in an EOfAD-like 156 manner.  The target sequence of Ps2Ex3 sgRNA is 5'-CAGACAGTGAAGAGGAC TCC-3'. 169 This target sequence was cloned into the plasmid pDR274 (Addgene plasmid # 42250)   To test whether the CRISPR/Cas9 system had functioned in the injected G0 embryos,   When a G0 injected fish had grown to sufficient size (>2 cm in length, 2 to 3 months 217 old), the tip of its tail (~2 mm in length) was biopsied (clipped) under Tricaine 218 (1.68μg/mL) anesthesia for genomic DNA extraction. The clipped tail was placed in 219 100 μL of 50 mM NaOH and then heated to 95°C for 15 min to extract genomic DNA.

220
The sample was then cooled to 4°C, and a 1/10th volume of 1 M Tris-HCl, pH 8.0 was 221 then added to each sample to neutralize the basic solution (31). The same T7 222 Endonuclease I assay used previously for mutation detection in G0 embryos was then 223 applied to the genomic DNA extracted from the G0 adult fish biopsy. However, since 224 each G0 mutation-carrying fish was probably mosaic for several different mutations at 225 the target site, each G0 fish was outbred to a wild type Tübingen fish, to produce the 226 F1 progeny, some of which could be heterozygous for single mutations. The F1 fish 227 were biopsied and screened using the T7 Endonuclease I assay when large enough. For   Breeding of psen2 S4Ter mutant fish 250 The initial F1 fish carrying the psen2 S4Ter mutation was outbred to a wild type fish to 251 generate a population of F2 progeny that was 50% heterozygous mutants and 50% wild 252 type fish. Two F2 heterozygous mutant fish were then inbred to generate a family of F3 253 fish consisting of (theoretically) 50% heterozygous mutants, 25% homozygous mutants 254 and 25% wild type fish. This F3 family was raised to six months of age before brain 255 removal and total brain RNA extraction for RNA-seq and other analyses.  For dqPCR tests, six wild type, six heterozygous and six homozygous fish from the F3 265 family were selected. Three of each genotype were then exposed to acute hypoxia 266 (dissolved oxygen content of the water was ~1.0 mg/L) for ~2.5 h, while the remaining 267 three of each genotype were exposed to normoxia. Fish were then euthanized and brains 268 removed (as above) for total RNA was extracted using the RNeasy Mini Kit (QIAGEN, 269 Venlo, Netherlands, 74104). cDNA was synthesised from brain RNAs using the       Raw data is given in supplementary data file S1 Table 3.

569
Translation initiation at S1 (S1-EGFP) does not permit initiation at S2-4. However, in Laboratory zebrafish become sexually mature at between 3 and 5 months of age.

589
Therefore, to examine the transcriptome of young adult zebrafish brains we identified 590 individuals of the desired genotype using PCRs specific for the mutant and wild type 591 alleles on DNA from tail biopsies ("tail clips") and then removed brains from fish of 592 the desired genotypes at 6 months of age. Total RNA was then purified from these and 593 subjected to either RNA-seq analysis (described below) or digital quantitative PCR 594 (dqPCR) as shown previously in Figure 2). The psen2 S4Ter allele shows dominance and may be EOfAD-like 731 Transcriptome analysis of 6-month-old entire brains from wild type, and psen2 S4Ter 732 heterozygous and homozygous zebrafish showed relatively few differences between the 733 heterozygous and homozygous fish centered around mitochondrial function while 734 extensive changes from wild type were caused by any presence of the psen2 S4Ter allele.

735
The differences that were seen between heterozygous and homozygous brains were in 736 similar functions to those seen for any presence of psen2 S4Ter compared to wild type, In conclusion, the psen2 S4Ter mutation is not the null allele we had hoped to isolate and 772 probably results in production of N-terminally truncated Psen2 proteins. These