Abstract
Forward genetic screening of mutants using firefly luciferase (LUC) reporter gene became a standard practice in plant research. Such screenings frequently identified alleles of CPL1 (Carboxyl-terminal Phosphatase-Like 1) regardless of promoters or pathways studied. Expression of the corresponding endogenous genes often shows the minimal difference between wild type and cpl1. Here we show that the LUC coding sequence is responsible for the high expression in cpl1, using a classical RD29a-LUC. Deletion of the LUC 3’-UTR did not change hyperactivation of LUC in cpl1. However, a codon-modified LUC (LUC2) produced similar expression levels both in wild type and in cpl1. These results indicate that the coding region of LUC is responsible for the cpl1-specific LUC overexpression uncoupled with the expression of the endogenous counterpart.
Use of the reporter genes to monitor gene expression has become a standard practice in the characterization of genes of interest. Various reporter genes with different benefits include bacterial β-glucuronidase (GUS), fluorescent proteins from jellyfish/coral organisms (XFPs), and luciferase from bacteria (LUX), firefly (LUC) and sea pansy (Renilla LUC). Of these, LUC from firefly has been widely used to conduct non-invasive monitoring of plant gene expression. Due to the short half-life of LUC mRNA (45 min) and protein (15.5 min with luciferin, 155 min without), LUC provides a low-background and highly sensitive way to monitor the plant gene expression in real time and to study the inducible gene expression in response to environmental stimuli 1. Several groups including ours took advantage of LUC system to identify genetic mutations in Arabidopsis thaliana, where LUC reporter lines were subjected to mutagenesis, and genetic mutations were identified based on the alteration of LUC expression profile 2. Typically, inducible promoters were fused to LUC, and plants that over- or under-express LUC upon stimulation were identified as potential mutants for regulators of gene expression.
Over the past years, it became evident that the LUC reporter-based forward genetic approaches repeatedly identify mutations in CPL1/FRY2 as well as HOS5/RCF3 genes regardless of the biological processes studied 3-8. These include salt/osmotic stress, low-temperature-stress, jasmonate signaling/wounding, miRNA expression. Interestingly, in these studies, cpl1/fry2 and hos5/rcf3 mutations typically yield 10-100-fold enhancement in LUC mRNA or activities, however, the expression of corresponding endogenous transcripts was enhanced only ∼2 fold. Because endogenous stress-responsive transcripts are often induced >10 fold, the impact caused by the mutations are relatively subtle. The typical explanation for this inconsistency was the likely presence of additional negative regulatory elements in the endogenous promoter, which was not included in the promoter fragments used in the reporter constructs.
CPL1/FRY2 encodes a protein with phosphatase activity that resembles eukaryotic TFIIF-interacting carboxyl terminal phosphatase (FCP1) and can dephosphorylate RNA polymerase II in vitro 3, 4, 9. As oppose to the expected function of CPL1 in the regulation of RNA polymerase II transcription, however, in vivo data are indicative that CPL1 functions in post-transcriptional metabolism of mRNA 5, 7. The first reported evidence that CPL1 affects the state of LUC reporter mRNA was that a cpl1 mutant showed higher capping efficiency and altered polyadenylation site of LUC mRNA. Another study demonstrated that CPL1/FRY2 regulate RNA decay pathway of abnormal transcripts, such as transcripts that retain one or more introns 10. These findings prompted me to re-investigate whether stress-inducible promoter/signaling or reporter gene structure are the determinants of the enhanced expression of stress-inducible reporter gene in the cpl1 mutant.
In addition to RD29a-LUC used in previous studies, we prepared RD29a-LUCΔ3’ and RD29a-LUC2. RD29a-LUCΔ3’ was made by removing 3’-UTR sequence of original firefly LUC mRNA from RD29a-LUC. The 3’-UTR of LUC mRNA contains two cryptic polyadenylation sites, usage of which were affected in cpl1 mutants 5. On the other hand, RD29a-LUC2 used codon-modified LUC (LUC2, Promega) coding sequence without firefly LUC 3’-UTR. These reporter cassettes were transformed into Arabidopsis thaliana Col-0 cpl1-6 mutant line, and then wild-type lines were prepared by backcrossing T1 plants showing 3:1 segregation ratio to the wild-type Col-0 plants. Homozygous T3 (cpl1-6) and F3 (wild type) lines were used to directly compare the reporter gene expression levels between wild type and the cpl1 mutant.
As shown in Figure 1, the cpl1-6 mutant host strongly enhanced RD29a-LUC as reported previously. Removal of LUC 3’-UTR did not alter hyper-induction of LUC in cpl1-6, suggesting LUC hyperexpression in cpl1-6 was not due to cryptic polyadenylation sites in LUC 3’-UTR. Notably, overall expression of RD29a-LUCΔ3’ was enhanced compared to RD29a-LUC. Expression of codon-modified RD29a-LUC2 substantially altered reporter gene expression profile. Compared to RD29a-LUC, RD29a-LUC2 activity was similarly increased in both wild type and cpl1-6. The induction pattern was close to that of endogenous RD29a mRNA level measured by RT-qPCR (Figure 1C), albeit RD29a-LUC2 response was slower than that of RD29a mRNA. This result suggested that hyperexpression of RD29a-LUC in cpl1 mutant are neither due to the enhancement of RD29a promoter activity, upstream osmotic-stress signaling events nor alternative polyadenylation in 3’-UTR. Instead, the data showed that LUC coding sequence strongly influences cpl1-dependent hyperexpression. This reporter-gene-specific effect by the cpl1 mutation explains why mutations in CPL1/FRY2 as well as in HOS5/RCF3 encoding a binding partner of CPL1 are repeatedly identified in LUC reporter gene-based forward genetic screening, and why there are discrepancies between LUC expression level and endogenous counterpart in cpl1/fry2. Because LUC mRNA has a short half-life in the plant cell 1, 11, the likely mechanism is LUC mRNA but not endogenous counterpart being a target of CPL1-dependent RNA decay pathway 10. This observation suggests that we need to be cautious about the impact of cpl1 on synthetic phenotype produced by artificial transgenes, because cpl1 may produce phenotype by specifically affecting the level of transgene expression.
I found the expression of codon-modified LUC2 reporter gene retains stress-regulation in wild type and is much less affected by the cpl1-6 mutation, therefore, LUC2 provides a promising alternative for reporter-based forward-genetic screening. It should be noted, however, that RD29a-LUC2 produced higher background than RD29a-LUC, and there was a delay in the timing of induction of RD29a-LUC2 activity above the background level after the onset of cold treatment relative to RD29a-LUC, which would require careful optimization of the assay system.
Materials and methods
Preparation and analysis of reporter lines
RD29a-LUCΔ3’ was prepared by inserting a PCR fragment containing 3’-end of LUC coding sequence (primer pair 1358 TAGTTGACCGCTTGAAGTCT/1359 GGAAAAGAGCTCTTACAATTTGGACTTTCCGC) into PacI-SacI sites of original RD29a-LUC plasmid 2. RD29a-LUC2 was prepared by inserting a PCR fragment encoding LUC2 (primer pair: 1471 CTATTTACAATTACAGTCGACATGGAAGATGCCAAAAACATT/1472 cagatcccccgggggtaccgagctcTTACACGGCGATCTTGCCGC) into SalI-SacI site of the RD29a-LUC plasmid. Resulting plasmids were introduced into Agrobacterium tumefaciens ABI, and used for transformation of cpl1-6 12 by flower transformation. Kanamycin-resistant plants were selected as described 13. Transformants whose luciferase expression phenotype were segregating with 3:1 ratio were backcrossed to Col-0. Homozygous T3 (cpl1-6 background) and F3 (Col-0 background) lines were established for each transgene and were used for the expression analysis. Luciferase imaging analysis and RT-qPCR analysis was performed as described 12.
Acknowledgement
This study was partially supported by USDA-NIFA (2015-67013-22816).
Footnotes
Conflict of interest: There is no conflict of interest for this study.