Nrf2 Activation Attenuates Genetic Endoplasmic Reticulum Stress Induced by a Mutation in the Phosphomannomutase 2 Gene in Zebrafish

Home , Phosphomannomutase, PMM1, PMM2

Nrf2 activation attenuates genetic endoplasmic reticulum stress induced by a mutation in the phosphomannomutase 2 gene in zebrafish

Katsuki Mukaigasaa,1,2, Tadayuki Tsujitaa,b,1,3, Vu Thanh Nguyena,1,LiLia,1,4, Hirokazu Yagic, Yuji Fusea, Yaeko Nakajima-Takagia,5, Koichi Katoc,d,e, Masayuki Yamamotob,f, and Makoto Kobayashia,b,6

aDepartment of Molecular and Developmental Biology, Faculty of Medicine, University of Tsukuba, 305-8575 Tsukuba, Japan; bExploratory Research for Advanced Technology Environmental Response Project, Japan Science and Technology Agency, University of Tsukuba, 305-8575 Tsukuba, Japan; cGraduate School of Pharmaceutical Sciences, Nagoya City University, Mizuho-ku, 467-8603 Nagoya, Japan; dOkazaki Institute for Integrative Bioscience, National Institutes of Natural Sciences, Okazaki, 444-8787 Aichi, Japan; eInstitute for Molecular Science, National Institutes of Natural Sciences, Okazaki, 444-8787 Aichi, Japan; and fDepartment of Medical Biochemistry, Tohoku University Graduate School of Medicine, Aoba-ku, 980-8575 Sendai, Japan

Edited by Igor B. Dawid, The Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD, and approved January 31, 2018 (received for review August 9, 2017)

Nrf2 plays critical roles in animals’ defense against electrophiles the novel Nrf2-inducing cellular stresses and elucidate the mo- and oxidative stress by orchestrating the induction of cytoprotec- lecular basis of their regulation. tive genes. We previously isolated the zebrafish mutant it768,which In this study, we identified the gene responsible for it768, one displays up-regulated expression of Nrf2 target genes in an uninduced of the six mutant lines, and elucidated the molecular mechanism state. In this paper, we determine that the gene responsible for it768 for how the mutation activates the Keap1–Nrf2 system. The was the zebrafish homolog of phosphomannomutase 2 (Pmm2), gene responsible for it768 was phosphomannomutase 2 (pmm2), which is a key enzyme in the initial steps of N-glycosylation, and its which encodes the enzyme required for the biosynthesis of mutation in humans leads to PMM2-CDG (congenital disorders of gly- mannose-1-phosphate, an essential metabolite for an early step cosylation), the most frequent type of CDG. The pmm2it768 larvae in the N-glycosylation process. pmm2it768 homozygous larvae exhibited mild defects in N-glycosylation, indicating that the showed impaired and insufficient N-glycosylation, followed by an pmm2it768 mutation is a hypomorph, as in human PMM2-CDG pa- increase in the endoplasmic reticulum (ER) stress. We found tients. A gene expression analysis showed that pmm2it768 larvae that the ER stress induces the up-regulation of the Keap1– display up-regulation of endoplasmic reticulum (ER) stress, sug- Nrf2 system through the double-stranded RNA-activated protein gesting that the activation of Nrf2 was induced by the ER stress. Indeed, the treatment with the ER stress-inducing compounds up- Significance regulated the gstp1 expression in an Nrf2-dependent manner. Fur- thermore, the up-regulation of gstp1 by the pmm2 inactivation Nrf2 is a master regulator of the antioxidant response and xeno- was diminished by knocking down or out double-stranded RNA- biotic metabolism. In this paper, we demonstrate that Nrf2 also activated protein kinase (PKR)-like ER kinase (PERK), one of the main plays a critical role in the endoplasmic reticulum (ER) stress response ER stress sensors, suggesting that Nrf2 was activated in response to usingazebrafishmutantinwhichNrf2isspontaneouslyactivated. the ER stress via the PERK pathway. ER stress-induced activation of The gene responsible for this mutant was phosphomannomutase 2 Nrf2 was reported previously, but the results have been controver- (pmm2), the enzyme required for the N-glycosylation. Human sial. Our present study clearly demonstrated that ER stress can indeed PMM2 is known to be the gene responsible for PMM2-CDG activate Nrf2 and this regulation is evolutionarily conserved among (congenital disorders of glycosylation), which currently has no vertebrates. Moreover, ER stress induced in pmm2it768 mutants was therapeutic options. pmm2 mutant larvae showed up-regulated ameliorated by the treatment of the Nrf2-activator sulforaphane, in- ER stress and ER stress-dependent Nrf2 activation. Of note, the dicating that Nrf2 plays significant roles in the reduction of ER stress. ER stress in mutant larvae was attenuated following treatment with the Nrf2 activator sulforaphane, suggesting that the ER stress | Nrf2 | PMM2-CDG | sulforaphane | zebrafish mutant pathological conditions of ER stress-associated diseases may be improved by taking Nrf2-activating foods. he Keap1–Nrf2 system senses various environmental stresses Tand induces cytoprotective genes to protect cells from these Author contributions: K.M., T.T., V.T.N., L.L., M.Y., and M.K. designed research; K.M., T.T., stresses (1–3). The main components of the system are Nrf2 and V.T.N., L.L., H.Y., Y.F., Y.N.-T., K.K., and M.K. performed research; K.M., T.T., V.T.N., and Keap1; Nrf2 heterodimerizes with small Maf proteins and func- L.L. analyzed data; and M.K. wrote the paper. tions as a transcription activator while Keap1 represses Nrf2 The authors declare no conflict of interest. functions as an adaptor for Cul3-based E3 ligase to regulate This article is a PNAS Direct Submission. the proteasomal degradation of Nrf2. In addition to the Nrf2- Published under the PNAS license. inhibiting effects, Keap1 also functions as a sensor molecule for a 1K.M., T.T., V.T.N., and L.L. contributed equally to this work. variety of Nrf2-activating compounds. One of the most interesting 2Present address: Department of Neuroanatomy and Embryology, School of Medicine, points in the Keap1–Nrf2 system is that it can respond to cellular Fukushima Medical University, 960-1295 Fukushima, Japan. stresses, such as autophagy impairment and mitochondria stress 3Present address: Department of Applied Biochemistry and Food Science, Saga University, (4, 5), in addition to well-studied oxidative stress and electrophiles. 840-8502 Saga, Japan. The Keap1–Nrf2 system is well conserved among vertebrates, 4Present address: School of Life Science and Technology, Harbin Institute of Technology, including zebrafish (6–8), in light of its molecular basis (9–11), 150080 Harbin, China. activating chemicals and stress (12–15), target genes (16–18), 5Present address: Department of Cellular and Molecular Medicine, Graduate School of and physiological functions (15, 19). We previously isolated six Medicine, Chiba University, 260-8670 Chiba, Japan. mutant zebrafish lines in which Nrf2 is activated without treat- 6To whom correspondence should be addressed. Email: [email protected]. ment with Nrf2 activators (12). We concluded that the mutations This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. in these lines generate some endogenous cellular stresses that 1073/pnas.1714056115/-/DCSupplemental. can activate Nrf2. The analysis of these mutants will help clarify Published online February 22, 2018.

2758–2763 | PNAS | March 13, 2018 | vol. 115 | no. 11 www.pnas.org/cgi/doi/10.1073/pnas.1714056115 Downloaded by guest on September 29, 2021 kinase (PKR)-like ER kinase (PERK) pathway. The Nrf2 acti- type (5mis-pmm2MO). Conversely, the overexpression of pmm2 vation by ER stress has been demonstrated previously using cultured reduced the up-regulated expression of gstp1 in pmm2it768 homo- cells (20), but whether or not this is part of general regulation is zygous larvae (Fig. S2D). These results indicated that the mutation controversial, since negative data were also reported by other groups of pmm2 is indeed the reason why the gstp1 expression was up- (21–23). regulated in it768 mutants. Genomic sequencing of pmm2 demonstrated that there were Results no marked differences in the coding region between WT and pmm2 Is the Gene Responsible for it768. it768 mutant embryos and it768 mutants, but the G in the beginning of intron 5 was A in the larvae showed no obvious defects during embryogenesis and mutant (Fig. 1B). Since it is the first G in the GU–AG rule (25), hatching process in bright-field images while the gstp1 expression a splicing defect was suggested. We therefore next analyzed the in the gills was up-regulated in 5-d-postfertilization (dpf) larvae pmm2 mRNA in WT and mutant larvae by reverse transcription- (Fig. 1A). The up-regulation of other Nrf2 targets, such as PCR (RT-PCR) and found that the WT mRNA was a bit longer hmox1a, was also observed in the liver of 7-dpf larvae so Nrf2 may than that from the it768 larvae, which lost the corresponding have been activated constitutively in the mutants. The expression sequences to exon 5, including the EcoRI site (Fig. S2E). To of gstp1 and hmox1a in the mutant larvae was reduced when nrf2a investigate this splicing defect in detail, we subcloned cDNA was knocked down by morpholino oligonucleotide injection from 2-dpf embryos and 5-dpf larvae and determined their nu- (nrf2aMO) (Fig. 1A), suggesting that up-regulation of these genes cleotide sequences (Fig. S2F). Normal type was not observed in is indeed mediated by the Keap1–Nrf2 system. To confirm the the homozygous mutants; instead, four types of abnormal pmm2 Nrf2 dependence of the up-regulation, we crossed the it768 mu- cDNAs with different splicing types were identified, including fh318 “+3,” which is probably translated into enzymatic-active pro- tant line with the Nrf2 mutant line nrf2a (19) and observed the it768 gstp1 expression in pmm2it768;nrf2afh318 compound larvae. As teins. We therefore considered that pmm2 is not the null but expected, the nrf2a mutation decreased the it768-dependent gstp1 hypomorphic mutant. Noteworthy, most of the isolated cDNA were the 30-nucleotides-deleted type (Δ30) in the exon 5 (total up-regulation (Fig. S1), which reflected the result of the nrf2a- = Δ knockdown analysis. 94%, n 90). Since 30 was an in-frame deletion, the resulting To identify the gene responsible for it768, we restricted its it768-type Pmm2 protein was deduced to have an internal 10-aa chromosomal location to an ∼0.1-centimorgan interval on chro- deletion (Fig. 1B). Importantly, the amino acid residue corre- mosome 1 using about 2,000 meioses and found pmm2 as a can- sponding to Arg141 in human PMM2 was included in the deleted didate (Fig. S2A). Phosphomannomutase is a highly conserved 10 amino acids (Fig. 1B, green R). Arg141 is a crucial residue for enzyme that catalyzes the conversion of mannose-6-phospate substrate binding (26) and highly conserved among Pmm pro- to mannose-1-phosphate, and, in vertebrates, two homologs— teins, including invertebrate Pmm proteins (Fig. S2G). Its mu- Pmm1 and Pmm2—are present (24). The amino acid sequence of tation to histidine (R141H) is the most frequent mutation in the deduced zebrafish Pmm2 protein shared 70.4% identity with PMM2-CDG (congenital disorders of glycosylation) (previously human PMM2 protein (Fig. S2B). In the zebrafish genome, one CDG-Ia) patients (27). CDG is a subset of genetic disorders characterized by defective N-glycosylation of serum and cellular Pmm1 and one Pmm2 gene exist (www.ensembl.org/Danio_rerio/ – Info/Index). To confirm whether the up-regulation of gstp1 ex- proteins (28 30). Since the R141H mutation in the recombinant human PMM2 proteins was shown to disrupt its enzymatic ac- pression was based on the pmm2 mutation, we next performed a tivity (31, 32), this 10-aa deletion will also eliminate the activity gene knockdown analysis using morpholino oligonucleotides (Fig. of the zebrafish Pmm2 protein. S2C). As expected, gstp1 was up-regulated in the WT AB larvae injected with an antisense morpholino oligonucleotide against it768 pmm2 Larvae Displays N-Glycosylation Defects. Pmm2 homozy- pmm2 (pmm2MO), but not in those injected with the 5-mismatched gous knockout mice were reported to be embryonically lethal (33), and, in humans, the lack of patients homozygous for the R141H mutation was shown to be highly statistically significant (34, 35). To determine whether or not pmm2 is essential for A zebrafish development, the survival of larvae generated from heterozygous cross-breeding was examined (Fig. 2A). The results indicated that pmm2it768 homozygous larvae become weak after 7 dpf and die within 2 wk after fertilization whereas sibling WT and heterozygous mutants remain viable. Pmm2 is an essential enzyme for synthesizing oligosaccharides required for N-glycosylation (Fig. 2B) (28). To examine whether or not N-glycosylation was decreased in pmm2it768 larvae, the N- glycosylated protein extracted from 5-dpf larvae was analyzed. The total amount of N-linked oligosaccharides was reduced in B pmm2it768 homozygous larvae compared with WT; instead, free oligosaccharide levels were increased (Fig. 2C and Fig. S3A). N-glycans were processed and modified in the secretory pathway (28). To investigate which steps in this processing were affected by the pmm2 mutation, we analyzed the N-glycosylated chains using HPLC and found that all high-mannose-type glycans were down-regulated (Fig. 2 C and D), suggesting that defects do not Fig. 1. Up-regulated expression of Nrf2 target genes in it768 larvae that influence the processing steps in the Golgi body but much earlier CELL BIOLOGY have splicing defects in the pmm2 gene. (A) Expression of gstp1 and hmox1a it768 steps, probably at the ER. In addition, we determined the in WT or pmm2 larvae injected or not with 1 pmol nrf2aMO. The closed – and open arrowheads indicate the gills and liver, respectively. The numbers structures of pauci-mannose type glycans in peaks a, b, and c in Fig. 2D (Fig. 2E and Fig. S3 B and C) and found that abnormal- in each picture indicate the larvae exhibiting strong expression of gstp1 or it768 hmox1a/tested larvae. (B) Nucleotide and deduced amino acid sequences of type N-glycans were accumulated in pmm2 larvae (Fig. 2E). the major cDNA isolated from WT and pmm2it768 larvae. The red “a” indi- Given these findings, we hypothesized that an extraordinary it768 cates the G-to-A mutation site of the it768 mutant in the zebrafish pmm2 structure of N-glycans was generated in pmm2 larvae be- genome. The green “R” indicates the amino acid residue corresponding to cause of the shortage of GDP-mannose, resulting in the Arg141 in human PMM2. EcoRI indicates the EcoRI site only present in the underglycosylation and accumulation of aberrant processed WT cDNA. N-glycans.

Mukaigasa et al. PNAS | March 13, 2018 | vol. 115 | no. 11 | 2759 Downloaded by guest on September 29, 2021 AB results indicate that ER stress was induced in the cells in the mutant liver. Nrf2 has been shown to be activated by chemically induced ER stress in some mammalian cultured cells (44). The genetic ER stress induced by the pmm2 mutation may also activate zebrafish Nrf2. To test this hypothesis, we examined whether the Nrf2 target gene gstp1 is induced by treatment with the chemical ER stress inducers tunicamycin (TM) and thapsigargin (TG) in WT AB larvae (Fig. 3D). Tunicamycin is a competitive inhibitor of N-acetylglucosamine transferase that blocks protein N-glycosylation, similar to the pmm2 mutation (Fig. 2B), while thapsigargin + C is an inhibitor of the Ca2 -ATPase in the ER that induces per- turbation of ER calcium homeostasis (45). As expected, gstp1 induction was observed in both tunicamycin- and thapsigargin- treated larvae, indicating that Nrf2 is activated by ER stress

DE A 1/27 B C bip pmm2it768 b WT WT 6 m/+ m/m 4 a a bip 2 39/53 0 chop chop pmm2it768 4 b 2 a a ef1 Relative expression 0 Probe: bip WT m/+ m/m

Fig. 2. Lethality and glycosylation defects in pmm2it768 larvae. (A)Thesurvival D – nrf2aMO F – perkMO of pmm2it768 mutants calculated using the Kaplan–Meier method (78). 0/40 0/15 0/6 0/11 (B) Mannose-related metabolic pathways in N-glycan synthesis. On the ER mem- Control WT brane, 14 sugars [2 N-acetylglucosamine (GlcNAc), 9 mannose (Man), and 3 glucose (Glc)] from nucleotide-activated sugars (UDP-GlcNAc, GDP-Man, Dol- 38/40 3/32 5/5 0/3 P-Man, and Dol-P-Glc) are sequentially transferred to the lipid-like precursor TM pmm2it768 dolichol phosphate (Dol-P) to synthesize the lipid-linked oligosaccharides (LLO) required for the N-glycosylation of proteins. PMM2 is an essential enzyme for generating mannose-1-phosphate (Man-1-P), which is required for the forma- 36/36 3/18 Probe: gstp1 tion of both GDP-Man and Dol-P-Man. Tunicamycin (TM) blocks the transfer of TG G WT perkit312 GlcNAc to Dol-P. (C) Pyridylamino-glycans derived from the zebrafish larvae. it768 The glycan content in WT and pmm2 was calculated based on the peak Probe: gstp1 Control areas of the chromatograms on the amide column. (D) The HPLC profiles on the amide column of N-linked glycans from whole bodies of WT (black) or E WT nrf2afh318 it768/it768 pmm2 larvae (pink) at 5 dpf. MX (X = numbers) indicates ManxGlcNAc2- 0/14 0/10 pmm2MO asn, and a, b, and c indicate pauci-mannose–type glycans. (E)N-glycosylation Control profiles on the octadecylsilyl column of fraction c separated on the amide column. The N-glycan structures were identified using the HPLC mapping 13/13 0/12 Probe: gstp1 method, as described in Materials and Methods. circles, mannose; squares, Tested No. 20 2036 38 12 23 TG 100 N-acetylglucosamine. Of note, the peak c1 was not detected in WT larvae. * *

Probe: gstp1 50 gstp1 - it768 positive pmm2 larvae (%) Up-Regulation of Physiological ER Stress in the Liver of 0 Larvae. The expression profile of pmm2 mRNA during the em- pmm2MO ––++– + bryonic and larval stages was ubiquitous (Fig. S4A). Intriguingly, perkit312 WT m/+ m/m the pmm2 expression in the gills and liver became strong in it768 Fig. 3. PERK-dependent activation of Nrf2 in 5-dpf larvae by genetic- and pmm2 homozygous mutants compared with WT siblings after chemical-induced ER stress. The numbers in each picture indicate the larvae 3 dpf (Fig. S4B). Since the expression of the Pmm2 gene has exhibiting strong bip expression in the liver (open arrowheads)/tested larvae (A) been reported to be induced by ER stress in both mice and or strong gstp1 expression in the gills (closed arrowheads)/tested larvae zebrafish (36, 37), we hypothesized that the up-regulation of (D–G). (A) bip expression in WT sibling and pmm2it768 larvae. (B)Theex- + pmm2 expression in pmm2it768 larvae was due to the induction of pression of bip and chop was analyzed by RT-PCR in WT, pmm2it768/ (m/+), or it768/it768 ER stress by the pmm2 mutation through the unfolded protein pmm2 (m/m) larvae. The amount of cDNA used for RT-PCR was stan- dardized by the ef1α expression. (C) The relative expression levels of bip and chop response (UPR) pathway. The UPR is a cellular protective re- + to ef1α was evaluated by qPCR in WT, pmm2it768/ (m/+), or pmm2it768/it768 (m/m) sponse to ER stress, such as the transcriptional induction of the larvae. a and b indicate statistically significant differences (Tukey’stest,P < 0.001; chaperone gene BiP/Grp78/Hspa5 and transcription factor n = 3 for each genotype). (D) Induced expression of gstp1 after 12-h treatment of Chop/Ddit3 (38, 39). Given that ER stress is induced when ab- 5 μg/mL tunicamycin (TM) or 2 μM thapsigargin (TG) in WT AB larvae injected or normal N-glycosylation occurs (40), we speculated that ER stress not with 1 pmol nrf2aMO. (E) Induced expression of gstp1 after 12-h treatment was up-regulated in pmm2it768 larvae. We therefore examined of 1 μM thapsigargin (TG) in WT sibling and nrf2afh318 larvae. (F) gstp1 expression in WT sibling and pmm2it768 larvae injected or not with 1 pmol the expression of the ER stress/UPR marker bip/grp78/hspa5 (37, it312 41) in the mutant larvae by whole-mount in situ hybridization perkMO. (G) gstp1 expression in WT sibling and perk larvae injected or not with 1 pmol pmm2MO. The graph shows the percentages of gstp1- (WISH) and RT-PCR and found that the bip expression was up- it768 – positive larvae of the indicated genotypes. The numbers of larvae tested regulated in the liver in pmm2 homozygous larvae (Fig. 3 A are indicated above the bars. Asterisks denote statistical significance C). Another ER stress/UPR marker, chop/ddit3 (42, 43), was also (Fisher’s exact test; *P < 0.001). The data were obtained based on the re- up-regulated in the homozygous mutants (Fig. 3 B and C). These sults of the WISH analysis shown in the Upper panels.

2760 | www.pnas.org/cgi/doi/10.1073/pnas.1714056115 Mukaigasa et al. Downloaded by guest on September 29, 2021 regardless of its underlying mechanisms, although the induction Tested No. 21 82 35 139 20 66 44 50 94 85 39 43 was weak and required a longer time than that induced by the 100 negative * typical Nrf2 activator diethyl maleate (DEM) (Fig. S5A). The induction was demonstrated to be Nrf2-dependent since it was eliminated when nrf2a was knocked down by morpholino oligonucleotides (Fig. 3D). The Nrf2 dependence was further confirmed 50 by experiments using the Nrf2 mutant line nrf2afh318 (Fig. 3E). In positive contrast, the tunicamycin-induced bip expression was not reduced positives larvae (%)

bip- weak by nrf2a knockdown, suggesting that it was Nrf2-independent (Fig. 0 S5B). This idea is also supported by the result that no bip induction SF +++ +++

was observed by diethyl maleate treatment (Fig. S5B). it768 Cullinan and Diehl (44) reported that the ER stress-induced pmm2 +/+ m/+ m/m +/+ m/+ m/m strong activation of Nrf2 was mediated by the PERK pathway, out of nrf2afh318 +/+ m/m three major UPR pathways (46). We therefore performed a knockdown analysis of perk/eif2ak3, the only ortholog of mamma- Fig. 4. The Nrf2-dependent attenuation of the up-regulated ER stress by lian PERK in zebrafish (Fig. S6A), using splicing inhibition-type sulforaphane treatment. The graphs show the percentages of bip-positive 5-dpf larvae of the indicated genotypes and their expression after 12-h morpholino oligonucleotides. With nonspliced perk mRNA, the treatment of 40 μM sulforaphane (SF). The numbers of larvae tested are expression of the perk-dependent ER-stress marker chop (47) was indicated above the bars. An asterisk denotes statistical significance (Fisher’s drastically reduced while that of the perk-independent ER-stress exact test; *P < 0.001). The data were obtained based on the results of the marker bip was enhanced (Fig. S6B). This latter observation may WISH analysis shown in the Right panels. The strength of the bip staining suggest that the ER stress was up-regulated by perk knockdown. was determined by double-blind scoring. The arrowheads indicate the liver. When perkMO-injected larvae were treated with tunicamycin or thapsigargin, the induction of gstp1 expression was obviously weakened, compared with the control larvae (Fig. S6C). No effects of plays a critical role in reducing ER stress in zebrafish (Fig. S8). perk knockdown on the diethyl maleate-induced expression of gstp1 Some previous reports showed that ER stress activates Nrf2 (20, were noted (Fig. S6C). Similarly, the up-regulated expression of gstp1 48, 49) while others denied this regulation (21–23). In addition to in pmm2it768 homozygous larvae was reduced by perk knockdown the latter reports, there were no transcriptome papers, which (Fig. 3F), suggesting that the ER stress-induced activation of clearly showed the induction of Nrf2 target genes by ER stress. Nrf2 was mediated by the PERK pathway. To strengthen our con- The ER stress-induced activation of Nrf2 has therefore not been clusion, we generated a perk-knockout zebrafish line (perkit312)by widely recognized. In the present study, we clearly showed that the CRISPR-Cas9 system (Fig. S6D) and performed gene ex- ER stress activated Nrf2 in animals by genetic analysis and pression analysis of gstp1 using pmm2-knockdown perkit312 larvae. suggest that this regulation has been evolutionarily conserved As expected, the up-regulated expression of gstp1 in pmm2 mor- among vertebrates. Previous studies may have failed to observe phants was suppressed by perk knockout (Fig. 3G). Taken to- this regulation for one of two reasons: (i) The Nrf2 activation by gether, these results indicate that Nrf2 was activated by ER stress ER stress is weak and slow compared with that by electrophiles in zebrafish larvae, at least in part, in a PERK-dependent manner. or oxidative stress, as shown in Fig. S5A (TM, TG vs. DEM); and (ii) this regulation may depend on certain conditions, such as Attenuation of ER Stress by the Nrf2 Activation. The fact that the tissue specificities and developmental stages. Nrf1, a paralog Keap1–Nrf2 system is activated in response to ER stress implies protein of Nrf2, has been shown to exist in the ER and undergo that this cellular defense system can suppress ER stress and/or its N-glycosylation (50–52). It is possible that Nrf1 was activated in downstream adverse effects. To test this hypothesis, the expres- response to ER stress and cooperated with Nrf2 to induce the it768 expression of Nrf2 target genes, but this was not the case. The sion of the ER stress marker bip in pmm2 larvae was analyzed it768 in pmm2it768;nrf2afh318 compound larvae. Interestingly, the ex- expression of gstp1 in both WT and pmm2 larvae was up- it768/it768 fh318/fh318 regulated, not down-regulated, by double knockdown of nrf1a pression of bip in pmm2 ;nrf2a larvae was higher it768/it768 fh318/+ it768/it768 +/+ and nrf1b (53), zebrafish coorthologs of mammalian Nrf1, sug- than in pmm2 ;nrf2a and pmm2 ;nrf2a gesting that the gstp1 up-regulation in pmm2it768 larvae may not larvae and was elevated in accordance with the decrease in the be mediated by Nrf1 (Fig. S9). WT nrf2a allele, suggesting that Nrf2 attenuated the ER stress in Attenuation of ER stress by Nrf2 activation suggests the im- pmm2it768/it768 single mutants (Fig. S7A). We also examined the – it768 fh318 portance of the Keap1 Nrf2 system in the homeostasis of cells lethality of pmm2 ;nrf2a compound larvae (Fig. S7B). The producing a number of secreted factors. Indeed, it was reported it768 nrf2a mutation had no effect on the survival of pmm2 ho- that the tunicamycin-induced ER stress was worsened in an mozygous larvae, suggesting that Nrf2 is not directly related to insulinoma cell line by Nrf2 knockdown (49). In the present either the lethality or survival of pmm2it768 larvae. study, the treatment of sulforaphane, a cancer chemopreventive The finding of Nrf2-dependent suppression of ER stress suggested ingredient in broccoli sprouts (54–56), reduced the up-regulated that further activation of the Keap1–Nrf2 system by exogenous ER stress in mutant larvae. Since the conservation of ER stress chemical inducers might reduce the up-regulated ER stress in regulation among vertebrates has been validated using medaka pmm2it768 larvae. To examine this possibility, pmm2it768 larvae were fish and zebrafish (37, 57, 58), the attenuation of ER stress by treated with sulforaphane, a well-known Nrf2 activator, and the ex- Nrf2 activation may be an effective strategy, not only for human pression of bip was analyzed. Fig. 4 indicates that sulforaphane health, but also farming animals. The molecular basis of this treatment reduced the numbers of bip-positive larvae in pmm2it768 attenuation is unknown, but enhancement of ER-associated homozygous mutants, suggesting that Nrf2 activation can attenuate degradation (ERAD) of unfolded and misfolded proteins by the pmm2 mutation-induced up-regulation of ER stress. As expected, Nrf2-dependent up-regulation of proteasome subunits might CELL BIOLOGY the sulforaphane-dependent attenuation of bip up-regulation in the have contributed (18, 49, 59). pmm2it768 homozygous larvae was canceled by the Nrf2 mutation (Fig. Human PMM2 is known as the gene responsible for PMM2- 4), suggesting that the effect of sulforaphane depends on the activation CDG (60). In PMM2-CDG, the most common type of CDG, of Nrf2. Taken together, these results suggest that the Keap1– patients have a life-threatening liver insufficiency, with an overall Nrf2 system plays a critical role in the UPR to reduce ER stress. 20% mortality during the neonatal period, combined with psy- chomotor retardation, hypotonia, dysmorphic features, failure to Discussion thrive, coagulopathy, abnormal endocrine functions, and a pro- In this study, we demonstrated that Nrf2 is activated by the ge- nounced susceptibility to infection (61). No therapeutic options netic or chemical-induced ER stress via the PERK pathway and are available for PMM2-CDG patients at present (34). A trial to

Mukaigasa et al. PNAS | March 13, 2018 | vol. 115 | no. 11 | 2761 Downloaded by guest on September 29, 2021 develop a zebrafish model of PMM2-CDG has recently been Gene Expression Analysis. The WISH and RT-PCR analyses were carried out as started using pmm2-knockdown zebrafish (62), but the pheno- described previously (14). To prepare RNA probes for the WISH analysis, types of their morphants were not observed in our pmm2it768 pKSpmm2, pKSbip, and pKSgstp1N were digested with BamHI, and pSKhmox1a mutants: dysmorphic craniofacial cartilage and motility defects was digested with XhoI, mixed with DIG RNA labeling mix (Roche), and tran- in the embryonic period (Table S3). The cause of this discrep- scribed with T3 RNA polymerase (Roche). All pictures were taken using a Leica it768 ancy is unclear. Since pmm2 is a hypomorphic mutant with MZ16 microscope equipped with an Olympus DP73 digital camera. Genotyping splicing defects, it may display weaker symptoms than knock- of the stained embryos was carried out after taking photos. The specific primers down embryos. It is also possible that the phenotypic differ- used for the RT-PCR analysis are described in Table S1. A quantitative PCR it768 ence between pmm2 mutants and pmm2 morphants is due to (qPCR) was performed using a 7500 Fast Real-Time PCR System (Thermo Fisher off-target effects of morphant phenotypes (63) and/or genetic Scientific) with THUNDERBIRD SYBR qPCR Mix (Toyobo). The expression level of compensation by the deleterious mutation (64). As shown in Fig. each gene was normalized to the level in ef1α (or eef1a1l1). To visualize the it768 2, pmm2 larvae showed defects in N-glycosylation, as in splicing defects of pmm2, we amplified the region corresponding to exons it768 PMM2-CDG patients (65, 66), suggesting that pmm2 larvae 5 and 6 by RT-PCR using specific primers (Table S1) and digested the PCR can be used as a disease model for PMM2-CDG. We have products with EcoRI. therefore attempted to search for chemical compounds that can it768 suppress the lethality of pmm2 larvae, hoping to find drug Microinjection. Morpholino oligonucleotides (0.5 to 1 pmol) or capped mRNAs candidates for PMM2-CDG, but no compound has been shown (100 pg) were injected into single cell-stage embryos using an IM300 to be effective so far, including sulforaphane. We think that the microinjector (Narishige) as described previously (9). The nucleotide se- cause of larval lethality may be hypoglycosylation of critical quences of the morpholino oligonucleotides (Gene Tools) are listed in Table proteins for survival, not the up-regulated ER stress. Further- S2. The pmm2 mRNA was transcribed from pCS2FLpmm2 linearized by more, the up-regulation of mild ER stress, which was detected in Bsp120I using the mMESSAGE mMACHINE kit (Ambion). pmm2it768, has also been observed in some CDG patients (67). In contrast, the UPR was not detected in fibroblasts derived from Preparation and Identification of Oligosaccharides and N-Glycans. Ninety (WT) CDG patients (68), probably because those fibroblasts produce and 100 (pmm2it768) zebrafish larvae were lysed in hypotonic buffer, fol- fewer secreted proteins than liver cells, which expressed bip in lowed by homogenization and centrifugation to isolate free oligosaccharides, our study. Since the cerebellum, one of the major organs in cytosolic proteins, and membrane-associated proteins as described previously which CDG symptoms are observed, has been shown to be (71). The N-glycans were released from the lyophilized protein fraction (soluble sensitive to ER stress (69), ER stress may be the cause of CDG plus membrane-associated proteins) by hydrazinolysis and purified using a symptoms in some patients. For such patients, diet therapy using Nrf2-activating foods/drinks/supplements, such as broccoli sprouts, graphite carbon column (72). For the HPLC analysis, the reducing ends of the may improve their conditions. free and N-linked oligosaccharides were pyridylaminated and purified using a cellulose column (73, 74). The structures of N-glycans were identified by the Materials and Methods HPLC mapping method combined with mass spectrometry. All of the experimental procedures used, including the chromatographic conditions and gly- Zebrafish and Chemicals. AB and TL strains were used as WT zebrafish. For it768 cosidase treatments, have been described previously (75). The calculated HPLC genotyping pmm2 , the genomic DNA at the mutation sites was amplified by PCR using primers in Table S1 and digested with ClaI: pmm2it768 map based on the unit contribution values was used for the estimation of a – mutant alleles were digested; WT alleles were not. Genotyping of nrf2afh318 pauci-mannose type pyridylamino-oligosaccharides (76). was carried out as described previously (19). For the induction experiments and survival assays, larvae were placed in culture dishes with E3 medium plus Gene Knockout Using CRISPR-Cas9. A knockout line of zebrafish perk was 0.1 mg/L methylene blue (19), containing diethyl maleate (Wako), sulfor- generated using CRISPR-Cas9 technology as previously described (77). In brief, aphane (LKT Laboratories), tunicamycin (Sigma-Aldrich), and/or thapsigargin transactivating CRISPR RNA (100 pg), perk-specific CRISPR RNA (50 pg), and (Sigma-Aldrich). In the case of survival assays, the E3 medium was exchanged Cas9 nuclease (400 pg) were coinjected into the yolk of one cell-stage WT AB every 2 d. Dead embryos were sequentially collected each day and geno- embryos. All RNAs and protein were purchased from Integrated DNA Technol- typed as described above. ogies. Identification of knockout line was performed by heteroduplex mobility assay and/or DNA sequencing using PCR and primers shown in Table S1. Chromosome Mapping. The gene locus of the it768 mutant was roughly mapped by a bulked-segregant analysis using 333 polymorphic Z-markers ACKNOWLEDGMENTS. We thank M. Eguchi, M. Komeda, Y. Nakayama, (70). Fine mapping with additional polymorphic markers was carried out Y. Terashita, and Y. Wada for help with fish maintenance; T. E. Dever, ∼ using 2,000 homozygous mutant larvae in the AB/TL hybrid F1 generation. H. H. Freeze, T. Irimura, T. Ishikawa, A. Kawahara, A. Kobayashi, K. Mori, H. Yoshida, and Y. Wada for valuable suggestions; and C.-S. Andrea, Plasmid Construction. Full-length cDNA of pmm2 was synthesized by RT-PCR H. Nakajima, Miki Takeuchi, Miho Takeuchi, and J. Tamaoki for experimental using zebrafish larval RNA and specific primers (Table S1) and subcloned into help and discussions. This work was supported by grants from the Yamazaki the pBluescript II KS and pCS2FL vectors (10); the resulting plasmids were Spice Promotion Foundation (to M.K.), the Koyanagi Foundation (to M.K.), named pKSpmm2 and pCS2FLpmm2, respectively. To construct pKSbip, the the Japan Science and Technology Corporation (ERATO) (to M.Y.), and the bip cDNA was synthesized by RT-PCR using specific primers (Table S1) and Ministry of Education, Science, Sports and Culture of Japan [Grants 20059004, subcloned into the pBluescript II KS. The plasmids pSKhmox1a (17) and 21026003, 21659043, 24590340, 26116705, and 26520101 (to M.K.); Grant pKSgstp1N (16) have been described previously. 25102008 (to K.K.); and Grants 15K07935 and 26110716 (to H.Y.)].

1. Hayes JD, Dinkova-Kostova AT (2014) The Nrf2 regulatory network provides an 8. Kobayashi M, Yamamoto M (2005) Molecular mechanisms activating the Nrf2- interface between redox and intermediary metabolism. Trends Biochem Sci 39: Keap1 pathway of antioxidant gene regulation. Antioxid Redox Signal 7: 199–218. 385–394. 2. Suzuki T, Motohashi H, Yamamoto M (2013) Toward clinical application of the Keap1- 9. Kobayashi M, et al. (2002) Identification of the interactive interface and phylogenic Nrf2 pathway. Trends Pharmacol Sci 34:340–346. conservation of the Nrf2-Keap1 system. Genes Cells 7:807–820. 3. Kobayashi M, Yamamoto M (2006) Nrf2-Keap1 regulation of cellular defense mech- 10. Li L, et al. (2008) Molecular evolution of Keap1. Two Keap1 molecules with distinc- anisms against electrophiles and reactive oxygen species. Adv Enzyme Regul 46:113–140. tive intervening region structures are conserved among fish. J Biol Chem 283: 4. Itoh K, Ye P, Matsumiya T, Tanji K, Ozaki T (2015) Emerging functional cross-talk 3248–3255. between the Keap1-Nrf2 system and mitochondria. J Clin Biochem Nutr 56:91–97. 11. Takagi Y, et al. (2004) MafT, a new member of the small Maf protein family in ze- 5. Ichimura Y, et al. (2013) Phosphorylation of p62 activates the Keap1-Nrf2 pathway brafish. Biochem Biophys Res Commun 320:62–69. during selective autophagy. Mol Cell 51:618–631. 12. Kobayashi M, et al. (2009) The antioxidant defense system Keap1-Nrf2 comprises a 6. Fuse Y, Kobayashi M (2017) Conservation of the Keap1-Nrf2 system: An evolutionary multiple sensing mechanism for responding to a wide range of chemical compounds. journey through stressful space and time. Molecules 22:E436. Mol Cell Biol 29:493–502. 7. Hahn ME, Timme-Laragy AR, Karchner SI, Stegeman JJ (2015) Nrf2 and Nrf2-related 13. Tsujita T, et al. (2011) Nitro-fatty acids and cyclopentenone prostaglandins share proteins in development and developmental toxicity: Insights from studies in zebra- strategies to activate the Keap1-Nrf2 system: A study using green fluorescent protein fish (Danio rerio). Free Radic Biol Med 88:275–289. transgenic zebrafish. Genes Cells 16:46–57.

2762 | www.pnas.org/cgi/doi/10.1073/pnas.1714056115 Mukaigasa et al. Downloaded by guest on September 29, 2021 14. Fuse Y, Nakajima H, Nakajima-Takagi Y, Nakajima O, Kobayashi M (2015) Heme- 48. Liu XM, et al. (2005) Endoplasmic reticulum stress stimulates heme oxygenase-1 gene mediated inhibition of Bach1 regulates the liver specificity and transience of the Nrf2- expression in vascular smooth muscle. Role in cell survival. J Biol Chem 280:872–877. dependent induction of zebrafish heme oxygenase 1. Genes Cells 20:590–600. 49. Lee S, et al. (2012) Involvement of the Nrf2-proteasome pathway in the endoplasmic 15. Fuse Y, Nguyen VT, Kobayashi M (2016) Nrf2-dependent protection against acute reticulum stress response in pancreatic β-cells. Toxicol Appl Pharmacol 264:431–438. sodium arsenite toxicity in zebrafish. Toxicol Appl Pharmacol 305:136–142. 50. Zhang Y, Lucocq JM, Yamamoto M, Hayes JD (2007) The NHB1 (N-terminal homology 16. Suzuki T, et al. (2005) Pi class glutathione S-transferase genes are regulated by Nrf box 1) sequence in transcription factor Nrf1 is required to anchor it to the endo- 2 through an evolutionarily conserved regulatory element in zebrafish. Biochem J plasmic reticulum and also to enable its asparagine-glycosylation. Biochem J 408: 388:65–73. 161–172. 17. Nakajima H, et al. (2011) Tissue-restricted expression of Nrf2 and its target genes in 51. Zhang Y, Crouch DH, Yamamoto M, Hayes JD (2006) Negative regulation of the zebrafish with gene-specific variations in the induction profiles. PLoS One 6:e26884. Nrf1 transcription factor by its N-terminal domain is independent of Keap1: Nrf1, but 18. Nguyen VT, et al. (2016) Conservation of the Nrf2-mediated gene regulation of not Nrf2, is targeted to the endoplasmic reticulum. Biochem J 399:373–385. proteasome subunits and glucose metabolism in zebrafish. Oxid Med Cell Longev 52. Wang W, Chan JY (2006) Nrf1 is targeted to the endoplasmic reticulum membrane by 2016:5720574. an N-terminal transmembrane domain. Inhibition of nuclear translocation and 19. Mukaigasa K, et al. (2012) Genetic evidence of an evolutionarily conserved role for transacting function. J Biol Chem 281:19676–19687. Nrf2 in the protection against oxidative stress. Mol Cell Biol 32:4455–4461. 53. Williams LM, et al. (2013) Developmental expression of the Nfe2-related factor (Nrf) 20. Cullinan SB, et al. (2003) Nrf2 is a direct PERK substrate and effector of PERK-de- transcription factor family in the zebrafish, Danio rerio. PLoS One 8:e79574. pendent cell survival. Mol Cell Biol 23:7198–7209. 54. O’Connell MA, Hayes JD (2015) The Keap1/Nrf2 pathway in health and disease: From 21. Nair S, et al. (2007) Toxicogenomics of endoplasmic reticulum stress inducer tunica- the bench to the clinic. Biochem Soc Trans 43:687–689. mycin in the small intestine and liver of Nrf2 knockout and C57BL/6J mice. Toxicol Lett 55. Kumar H, Kim IS, More SV, Kim BW, Choi DK (2014) Natural product-derived phar- 168:21–39. macological modulators of Nrf2/ARE pathway for chronic diseases. Nat Prod Rep 31: 22. Palsamy P, Bidasee KR, Shinohara T (2014) Valproic acid suppresses Nrf2/ 109–139. Keap1 dependent antioxidant protection through induction of endoplasmic re- 56. Kensler TW, et al. (2013) Keap1-nrf2 signaling: A target for cancer prevention by ticulum stress and Keap1 promoter DNA demethylation in human lens epithelial cells. sulforaphane. Top Curr Chem 329:163–177. Exp Eye Res 121:26–34. 57. Ishikawa T, Taniguchi Y, Okada T, Takeda S, Mori K (2011) Vertebrate unfolded 23. Mozzini C, et al. (2014) Increased endoplasmic reticulum stress and Nrf2 repression in protein response: Mammalian signaling pathways are conserved in medaka fish. Cell peripheral blood mononuclear cells of patients with stable coronary artery disease. Struct Funct 36:247–259. Free Radic Biol Med 68:178–185. 58. Ishikawa T, et al. (2013) ATF6α/β-mediated adjustment of ER chaperone levels is es- 24. Quental R, Moleirinho A, Azevedo L, Amorim A (2010) Evolutionary history and sential for development of the notochord in medaka fish. Mol Biol Cell 24:1387–1395. functional diversification of phosphomannomutase genes. J Mol Evol 71:119–127. 59. Kwak MK, Wakabayashi N, Greenlaw JL, Yamamoto M, Kensler TW (2003) Antioxi- 25. Padgett RA, Grabowski PJ, Konarska MM, Seiler S, Sharp PA (1986) Splicing of mes- dants enhance mammalian proteasome expression through the Keap1-Nrf2 signaling senger RNA precursors. Annu Rev Biochem 55:1119–1150. pathway. Mol Cell Biol 23:8786–8794. 26. Silvaggi NR, et al. (2006) The X-ray crystal structures of human α-phosphomanno- 60. Matthijs G, et al. (1997) Mutations in PMM2, a phosphomannomutase gene on mutase 1 reveal the structural basis of congenital disorder of glycosylation type 1a. chromosome 16p13, in carbohydrate-deficient glycoprotein type I syndrome (Jaeken J Biol Chem 281:14918–14926. syndrome). Nat Genet 16:88–92. 27. Jaeken J, Matthijs G (2001) Congenital disorders of glycosylation. Annu Rev Genomics 61. Grünewald S (2009) The clinical spectrum of phosphomannomutase 2 deficiency Hum Genet 2:129–151. (CDG-Ia). Biochim Biophys Acta 1792:827–834. 28. Freeze HH (2006) Genetic defects in the human glycome. Nat Rev Genet 7:537–551. 62. Cline A, et al. (2012) A zebrafish model of PMM2-CDG reveals altered neurogenesis 29. Freeze HH, Schachter H (2009) Genetic disorders of glycosylation. Essentials of and a substrate-accumulation mechanism for N-linked glycosylation deficiency. Mol Glycobiology, eds Varki A, et al. (Cold Spring Harbor Lab Press, Cold Spring Harbor, Biol Cell 23:4175–4187. NY), 2nd Ed. 63. Kok FO, et al. (2015) Reverse genetic screening reveals poor correlation between 30. Jaeken J, Matthijs G (2007) Congenital disorders of glycosylation: A rapidly expanding morpholino-induced and mutant phenotypes in zebrafish. Dev Cell 32:97–108. disease family. Annu Rev Genomics Hum Genet 8:261–278. 64. Rossi A, et al. (2015) Genetic compensation induced by deleterious mutations but not 31. Kjaergaard S, Skovby F, Schwartz M (1999) Carbohydrate-deficient glycoprotein gene knockdowns. Nature 524:230–233. syndrome type 1A: Expression and characterisation of wild type and mutant PMM2 in 65. Mills P, et al. (2001) Congenital disorders of glycosylation type I leads to altered E. coli. Eur J Hum Genet 7:884–888. processing of N-linked glycans, as well as underglycosylation. Biochem J 359:249–254. 32. Pirard M, et al. (1999) Effect of mutations found in carbohydrate-deficient glyco- 66. Mandato C, et al. (2006) Cryptogenic liver disease in four children: A novel congenital protein syndrome type IA on the activity of phosphomannomutase 2. FEBS Lett 452: disorder of glycosylation. Pediatr Res 59:293–298. 319–322. 67. Lecca MR, Wagner U, Patrignani A, Berger EG, Hennet T (2005) Genome-wide analysis 33. Thiel C, Lübke T, Matthijs G, von Figura K, Körner C (2006) Targeted disruption of the of the unfolded protein response in fibroblasts from congenital disorders of glyco- mouse phosphomannomutase 2 gene causes early embryonic lethality. Mol Cell Biol sylation type-I patients. FASEB J 19:240–242. 26:5615–5620. 68. Shang J, Körner C, Freeze H, Lehrman MA (2002) Extension of lipid-linked oligosac- 34. Freeze HH (2009) Towards a therapy for phosphomannomutase 2 deficiency, the charides is a high-priority aspect of the unfolded protein response: Endoplasmic re- defect in CDG-Ia patients. Biochim Biophys Acta 1792:835–840. ticulum stress in type I congenital disorder of glycosylation fibroblasts. Glycobiology 35. Sharma V, Ichikawa M, Freeze HH (2014) Mannose metabolism: More than meets the 12:307–317. eye. Biochem Biophys Res Commun 453:220–228. 69. Sun L, et al. (2013) Insufficient ER-stress response causes selective mouse cerebellar 36. Oommen AM, Narayanan U, Jagannath MR (2012) An integrative network biology granule cell degeneration resembling that seen in congenital disorders of glycosyla- approach to evaluate the role of endoplasmic reticulum stress response in obese type tion. Mol Brain 6:52. 2 diabetes. ISRN Cell Biol 2012:278636. 70. Takeuchi M, et al. (2015) LSD1/KDM1A promotes hematopoietic commitment of 37. Hu MC, et al. (2007) XBP-1, a key regulator of unfolded protein response, activates hemangioblasts through downregulation of Etv2. Proc Natl Acad Sci USA 112: transcription of IGF1 and Akt phosphorylation in zebrafish embryonic cell line. 13922–13927. Biochem Biophys Res Commun 359:778–783. 71. Yanagida K, Natsuka S, Hase S (2006) Structural diversity of cytosolic free oligosac- 38. Yoshida H (2007) ER stress and diseases. FEBS J 274:630–658. charides in the human hepatoma cell line, HepG2. Glycobiology 16:294–304. 39. Mori K (2009) Signalling pathways in the unfolded protein response: Development 72. Tanabe K, Ikenaka K (2006) In-column removal of hydrazine and N-acetylation of from yeast to mammals. J Biochem 146:743–750. oligosaccharides released by hydrazionolysis. Anal Biochem 348:324–326. 40. Lehrman MA (2001) Oligosaccharide-based information in endoplasmic reticulum 73. Yamamoto S, Hase S, Fukuda S, Sano O, Ikenaka T (1989) Structures of the sugar quality control and other biological systems. J Biol Chem 276:8623–8626. chains of interferon-gamma produced by human myelomonocyte cell line HBL-38. 41. Schlombs K, Wagner T, Scheel J (2003) Site-1 protease is required for cartilage de- J Biochem 105:547–555. velopment in zebrafish. Proc Natl Acad Sci USA 100:14024–14029. 74. Shimizu Y, et al. (2001) Rapid and simple preparation of N-linked oligosaccharides by 42. Cinaroglu A, Gao C, Imrie D, Sadler KC (2011) Activating transcription factor 6 plays cellulose-column chromatography. Carbohydr Res 332:381–388. protective and pathological roles in steatosis due to endoplasmic reticulum stress in 75. Yagi H, et al. (2005) Development of structural analysis of sulfated N-glycans by zebrafish. Hepatology 54:495–508. multidimensional high performance liquid chromatography mapping methods. 43. Lee HC, et al. (2011) Transgenic zebrafish model to study translational control Glycobiology 15:1051–1060. mediated by upstream open reading frame of human chop gene. Nucleic Acids Res 76. Tomiya N, et al. (1991) Calculated two-dimensional sugar map of pyridylaminated

39:e139. oligosaccharides: Elucidation of the jack bean alpha-mannosidase digestion pathway CELL BIOLOGY 44. Cullinan SB, Diehl JA (2006) Coordination of ER and oxidative stress signaling: The of Man9GlcNAc2. Anal Biochem 193:90–100. PERK/Nrf2 signaling pathway. Int J Biochem Cell Biol 38:317–332. 77. Kotani H, Taimatsu K, Ohga R, Ota S, Kawahara A (2015) Efficient multiple genome 45. Schönthal AH (2012) Endoplasmic reticulum stress: Its role in disease and novel modifications induced by the crRNAs, tracrRNA and Cas9 protein complex in zebra- prospects for therapy. Scientifica (Cairo) 2012:857516. fish. PLoS One 10:e0128319. 46. Mori K (2000) Tripartite management of unfolded proteins in the endoplasmic re- 78. Kaplan EL, Meier P (1958) Nonparametric estimation from incomplete observations. ticulum. Cell 101:451–454. J Am Stat Assoc 53:457–481. 47. Harding HP, et al. (2000) Regulated translation initiation controls stress-induced gene 79. Schneider A, et al. (2011) Successful prenatal mannose treatment for congenital expression in mammalian cells. Mol Cell 6:1099–1108. disorder of glycosylation-Ia in mice. Nat Med 18:71–73.

Mukaigasa et al. PNAS | March 13, 2018 | vol. 115 | no. 11 | 2763 Downloaded by guest on September 29, 2021