Figure 3. Functional mapping of the N-terminal domain of UPF1. (A) Schematic, non-proportional representation of the used constructs. (B-C) NMD competency test of the N-terminal UPF1 mutant constructs. As the N- and C-terminal domains of UPF1 are functionally redundant, the N-terminal domain was mapped in UPF1 mutant lacking the C-terminal domain (U1ΔC). Point mutants were generated from the U1ΔC construct by changing four (NP1-4), three (NP1-2A4A) or one (NP3A) N- terminal S/TQ sites to alanine. In NΔNn, the Nn region was deleted from the U1ΔC construct. The Nc region was deleted from the U1ΔC constructs (NΔNc), and then point mutants were generated by changing three, two or one S/TQ sites to alanine in the NΔNc construct (NΔNcP1-3A, NΔNcP1-2A and NΔNcP3A). (D) The Nc region of N-terminal domain of UPF1 is phosphorylated. (E) The NMD relevant S/TQ sites of the Nn region of N-terminal domain are phosphorylated. Note that while the NΔNc sample is phosphorylated, the NΔNcP1-3A mutant protein, in which the S/TQ sites were mutated, is not phosphorylated. C-term.region1 UPF1 phosphorylation is essential for Nonsense Mediated mRNA Decay (NMD) in plants Farkas Kerényi 1 , Izabela Wawer 2 , Joanna Kufel 2 , Dániel Silhavy 1 1 Plant RNA Biology Group, Institute of Genetics, Agricultural Biotechnology Center, Gödöllő, Hungary; 2 Institute of Genetics and Biotechnology, University of Warsaw, Poland ABSTRACT - In mammals, there is two important S/TQ phosphorylation site in UPF1 involved in NMD. One for SMG6 binding on the N-terminal and one for SMG7 binding on the C-terminal S/TQ-rich region. These sites are phosphorylated by SMG1 PIKK-kinase. Interestingly, in Arabidopsis there is no SMG1 and SMG6, but the N- and C-terminal S/TQ-rich region of Arabidopsis UPF1 are redundant. We examined if there is such important unique phosphorylation sites in AtUPF1 and whether these sites are phosphorylated. INTRODUCTION – Nonsense Mediated mRNA Decay (NMD) is a conserved eukaryotic quality control mechanism to identify and degrade mRNAs containing premature termination codon (PTC) thus preventing the accumulation of truncated and potentially harmful proteins. NMD also controls the expression of several wild-type genes. If the 3’ untranslated region (3’UTR) of a mRNA is too long (yeast and invertebrates) or there is an exon junction complex (EJC – formed after intron-excision) in 3’UTR (mammals) the translation termination factor eRF3 can’t interact with PABP so UPF1 can bind to eRF3. Then UPF1 binds UPF2 and UPF3 NMD factors forming the NMD complex. 1 Then, in mammals, the SMG1 PIKK-kinase phosphorylates a threonine (T28) in N-terminal S/TQ-rich region and a serine (S1096) in the C-terminal S/TQ-rich region allowing SMG6 and SMG7 to bind to these sites, respectively, followed by rapid degradation of the PTC-containing mRNA. 2 In plants, uniquely among eukaryotes, both types of NMD (long 3’UTR and intron in 3’UTR) works efficiently. 3 Interestingly, in Arabidopsis the N- and C-terminal S/TQ-rich region of UPF1 are redundant 4 , but there is no SMG1 and SMG6. To clarify the role of UPF1 phosphorylation in plant NMD, we created several UPF1 mutants deleting or point- mutating (alanine-substitution) the potential S/TQ phosphorylation sites. Then we studied the NMD competency and phosphorylation status of these mutants. METHODS – Earlier our group has elaborated an efficient agroinfiltration-based transient NMD test system combined with virus induced gene silencing (VIGS) to define and examine the cis-acting elements and trans factors of plant NMD. Our UPF1 mutants was examined in this VIGS-complementation system. In UPF1 VIGS plants the NMD doesn't work. Agroinfiltration with a functional UPF1 mutant that can complement the UPF1-deficency of VIGS plant leads to restoration of the NMD in the infiltrated patch. If we agroinfiltrate with a non-functional UPF1 mutant, it can't complement the UPF1 VIGS, thus the NMD is not restored in the infiltrated spot. These processes are observable by co- infiltrating an NMD-sensitive reporter gene (e.g. GFP containing an intron in its 3'UTR – Gc-I). If there is no NMD, the GFP will illuminate strong. If the NMD is restored, the GFP expression will be weak (Figure 1). The Gc-I mRNA level was examined by Northern blot. The Northern blots correlate very well with GFP fluorescence so here the Northern- blots are not shown. The N-terminal region of UPF1 is strongly conserved. It contains four potential S/TQ phospho-sites: three serines (S3, S13, S105) and a threonin (T29). The T29 is the homolog of human T28 which binds the SMG7. Therefore we examined whether this phospho-site is only needed for plant NMD. The results of our experiments show that the T29 is the most important phspho-site in the N-terminal region because it is able to complement the UPF1 VIGS alone. However, the others together are also able to do. Next, we examined whether these potential S/TQ phospho-sites are really phosphorylated. We phospho-stained our mutants with the following results: first, there is a highly phosphorylated region between T29 and S105 but this region is not needed for NMD. Second, the examined potential S/TQ phospho- sites are really phosphorylated. These results were verified by mass spectrometry. SUMMARY – Taken together, our results show that, in contrast to human UPF1, there is no strict site-specific phosphorylation to bind SMG7. Instead, there is a mass-phosphorylation on the N-terminal region and on the C-terminal region as well. There are important S/TQ sites and more or less important ones and there are S/TQ sites that are not needed for NMD in plants. UPF1 plays a key role in NMD in eukaryotes. However, its phosphorylation status is different amongst yeast (no phosphorylation), mammals (strict site-specific phosphorylation) and plants (mass-phosphorylation). Even among plants the phosphoregulation of UPF1 could be different, because SMG1is present in rice and in grapevine, while in Arabidopsis SMG1 is not present. As plant and animals diverged very early during eukaryotic evolution, the finding that phosphorylation of the N- and C- terminal S/TQ sites of UPF1 play an important role in both plant and animal NMD systems suggest that UPF1 phosphorylation by PIKK kinases already linked the NMD complex formation and target degradation steps of NMD in the last common ancestor of extant eukaryotes (stem eukaryotes). Presumably, the T28-T29 conserved N-terminal target site was one of the phosphorylated PIKK kinase target site in stem eukaryotes. ACKNOWLEDGEMENT – We are grateful to C. Lacomme (University of Edinburgh), D. Baulcombe (University of Cambridge) and SP. Dinesh-Kumar (Yale University) for TRV vectors. This research was supported by grants from the OTKA(K60102 and C77086) and ICGEB (CRP/HUN09-01). LITERATURE 1. Rewieved in: Hwang, Maquat: Nonsense-mediated mRNA decay (NMD) in animal embryogenesis: to die or not to die, that is the question. Current Opinion in Genetics & Development 2011, 21:422–430. 2. Okada-Katsuhata et al.: N- and C-terminal Upf1 phosphorylations create binding platforms for SMG-6 and SMG- 5:SMG-7 during NMD. Nucleic Acids Research, 2012, Vol. 40, No. 3 1251–1266. 3. Kerényi Z et al.: Inter-kingdom conservation of mechanism of nonsense-mediated mRNA decay. The EMBO Journal (2008) 27, 1585 – 1595. 4. Mérai et al.:The late steps of plant nonsense-mediated mRNA decay. The Plant Journal (2012) Volume 73, Issue 1, p50–62. CH Helicase ΔNU1C1 ΔNΔC Gc-I Ls intron 200nt GFP AAA S1013 T1056 S1076 S1085 C1P1-4A A1013 A1056 A1076 A1085 C1P1-2A A1013 A1056 S1076 S1085 C1P3-4A S1013 T1056 A1076 A1085 - ΔNΔC ΔNU1C1 C1P1-4A C1P3-4A C1P1-2A A ProQ Diamond AP ΔNΔC C - + - + - + ΔNU1C1 C1P1-4A B VIGS: U1 Gc-I + P14 λN HA Figure 1. The VIGS-complementation system. In VIGS silenced UPF1 plants, the NMD doesn’t work, so the NMD-target reporter construct GFP (Gc-I) express to high levels. GFP illumination will be weak, if the co-infiltrated UPF1 (Gc-I+UPF1) is functional and can complemens the NMD in the infiltrated patch. In this way we can examine the UPF1 mutant contructs (Gc- I+UPF1mut) whether they remain functional. TRV PDS UPF1 VIGS 8-10 days Gc-I Gc-I UPF1 silenced leaf Gc-I +UPF1 Gc-I +UPF1mut Gc-I Ls intron 200nt GFP AAA 3 days RESULTS – The C-terminal region of UPF1 is less conserved. The only conserved in different species' UPF1 that there are lots of S/TQ sites in this region. In plants, there are 18. Our group previously showed that the C-terminal region truncated from the 3'end can complement the UPF1 VIGS well enough. We examined these segments separately and we found that they are redundant (data not shown). So we examined further only the first four potential S/TQ phospho-site containing segment. Our result show that the first, the second and the fourth sites are not or less phosphorylated and are not involved in NMD, while the third one is phosphorylated and is needed for NMD (Figure 2). Figure 2. The S/TQ sites of the C1 region are functionally different. (A) Schematic, non-proportional representation of the C1 mutant constructs. The S/TQ sites of the C1 region (S1013-C1P1, T1056-C1P2, S1076-C1P3 and S1085-C1P4) were changed to alanine. In C1P1-4A construct all four sites were changed to alanine, while in C1P1-2A and C1P3-4A mutant two sites were mutated. (B) NMD competency tests of the C1 UPF1 mutant constructs. UPF1-silenced leaves (VIGS:U1) were infiltrated with Gc-I NMD reporter + P14 and with one of the C1 mutants or with ΔNU1C1 positive or ΔNΔC negative control constructs. Bright fluorescence shows that NMD is inactive as the co-infiltrated UPF1 mutant failed to complement the NMD deficiency of the UPF1 VIGS leaf, whereas weak fluorescence indicates that co-infiltrated UPF1 mutant complemented the UPF1-silenced leaf. (C) The S/TQ sites of the C1 region are phosphorylated. The UPF1 C-terminal mutant IP samples were treated with alkaline phosphatase (AP +) or with buffer (-), separated on SDS-PAGE, and the gels were stained with ProQ Diamond phospho-specific stain. Note that strong difference between the ProQ-diamond staining of a given protein phosophatase treated (+) and non-treated (-) samples indicates that the protein is heavily phosphorylated. The ΔNU1C1 protein is heavily phosphorylated, while the C1P1-4Amutant protein is weakly phosphorylated. A3 A3 U1ΔC B NP1-4A NP1-2A4A NP3A NΔNn VIGS: U1 Gc-I + P14 AP ΔNΔC U1ΔC NP1-4A NP1-2A4A NP3A D - + - + - + - + - + ΔNΔC U1ΔC NP1-4A NΔNn - + - + - + - + AP Nc-region U1ΔC NP1-4A NP1-2A4A NΔNn NP3A A S3 S13 T29 S105 A3 A13 A29 A105 T29 A29 ΔNΔC Nc-region (36-109 aa) Nn-region (1-35 aa) A3 A13 A105 S3 S13 S105 CH Helicase λN HA S105 CH Helicase λN HA CH Helicase λN HA N-term. VIGS: U1 Gc-I + P14 - NΔNc NΔNc-P1-2A ΔNΔC U1ΔC NΔNc-P1-3A NΔNc-P3A C ProQ Diamond AP ΔNΔC U1ΔC E - + - + - + - + ΔNΔC - + - + - + AP NΔNc NΔNcP1-3A NΔNcP1-2A NΔNcP3A Ls intron Gc-I 200nt GFP AAA Nn-region NΔNc S3 S13 T29 Δ36-109 aa. Nn-region NΔNcP1-3A A13 A29 Nn-region NΔNcP1-2A A13 T29 Nn-region NΔNcP3A S3 S13 A29 CH Helicase CH Helicase CH Helicase CH Helicase λN HA λN HA λN HA λN HA CH Helicase λN HA N-term. CH Helicase λN HA N-term. CH Helicase λN HA N-term. ProQ Diamond C1 C2 C3 C4 S T S S S S S T T TS S S SS S S T ΔNU1 CH Helicase λN HA C-term.region1 CH Helicase λN HA C-term.region1 CH Helicase λN HA C-term.region1 CH Helicase λN HA CH Helicase λN HA