Secondary Mutation-Induced Alternative Splicing Suppresses RNA Splicing Defect of the jhs1 Mutant¹

Abstract

jing he sheng1 (jhs1) is a mutant of the DNA2 homolog in Arabidopsis (Arabidopsis thaliana), which was previously identified as being involved in DNA damage repair, cell cycle regulation, and meristem maintenance. A mutation at the 3′ intron splice site of the 11th intron causes alternative splicing of this intron at two other sites, which results in frame shifts and premature stop codons. Here, we screened suppressors of jhs1 to further study the function and regulatory networks of JHS1. Three suppressors with wild-type–like phenotypes were obtained. Sequencing analysis results showed that each of the suppressors has a second mutation in jhs1 that causes further alternative splicing of the intron and corrects the shifted reading frame with small insertions. Precursor mRNA sequence analysis and intron splice site evaluation results suggested that intron splicing was disturbed in the suppressors, and this switched the splice site, resulting in small insertions in the coding regions of JHS1. Structural analysis of JHS1 suggested that the insertions are in a disordered loop region of the DNA2 domain and do not seem to have much deleterious effect on the function of the protein. This work not only has implications for the evolution of protein sequences at exon junctions but also provides a strategy to study the mechanism of precursor mRNA splicing.

Jing he sheng1 (jhs1) is a mutant identified in our laboratory with a defect in the plant meristem (Jia et al., 2016). Longitudinal sections of the shoot apical meristem indicated that the shoot apical meristem in jhs1 is widened and consists of enlarged cells in an abnormal arrangement. jhs1 mutants also showed a variety of phenotypes, including retarded growth, oblate stem, and an irregular meristem cell arrangement without distinct layers. The mutation of jhs1 occurred at the 3′ splice site of the gene’s 11th intron and resulted in the splice site being shifted 63 bp upstream, causing premature termination of the gene product. Treatment with DNA-damaging agents showed that jhs1 is sensitive to DNA damage stress. Furthermore, the frequency of homologous recombination was increased in jhs1. This mutation could affect the cell cycle progression through a DNA damage response. Thus, these results suggest that JHS1 plays a vital role in DNA replication and damage repair, meristem maintenance, and development in plants (Jia et al., 2016).

Human and yeast DNA Replication Helicase/Nuclease2 (DNA2) and JHS1 are homologous. DNA2 was initially identified in Saccharomyces cerevisiae by screening for DNA replication mutants (Budd and Campbell, 1995). DNA2 is a multifunctional enzyme with ATPase, helicase, and nuclease activities (Budd et al., 2000) and is involved in the process of Okazaki fragment maturation (Budd and Campbell, 1997). In addition, DNA2 plays a crucial role in processing DNA double-strand breaks (Zhu et al., 2008) and telomere and mitochondrial DNA maintenance (Duxin et al., 2009; Budd and Campbell, 2013). DNA2 is conserved in eukaryotes from yeast to human and contains three essential domains: N-terminal domain, nuclease domain, and helicase domain (Kumar and Burgers, 2013). On the whole, the three-dimensional structure of DNA2 is a cylinder with a central tunnel through which the single-stranded DNA can thread (Zhou et al., 2015). The nuclease domain is located at the base of this cylinder and forms a doughnut-like structure with the active site embedded in the central tunnel (Zhou et al., 2015). The helicase domains, which extend the cylinder, are on the top of the nuclease doughnut and lengthen the central tunnel (Zhou et al., 2015). An oligonucleotide/oligosaccharide-binding fold domain, which belongs to the N-terminal structural domain of DNA2, decorates the exterior of the nuclease domain and does not have a role in DNA binding (Zhou et al., 2015). The combination of single-stranded DNA and DNA2 is directional, with its 5′ end at the helicase domain and 3′ end at the nuclease domain, and DNA crosses over from the helicase to the nuclease domain (Zhou et al., 2015).

RNA splicing is a posttranscriptional process that is necessary to produce mature mRNA. It was first discovered by Chow et al. (1977). There are two different modes of splicing: constitutive and alternative (van den Hoogenhof et al., 2016). Constitutive splicing is the process through which introns are removed and exons are joined together to form a mature mRNA. Most plant exonic mRNA are flanked by canonical GU/AG splicing signals. Disrupting normal splice sites can damage function of the wild-type splice junctions and create de novo splice junctions. They can also alter splice enhancers or splice silencers and produce aberrant splicing. RNA splicing is carried out by the spliceosome, a large and dynamic ribonucleoprotein machine that consists of multiple accessory proteins and five small nuclear ribonucleoproteins (U1, U2, U4, U5, and U6; Wahl et al., 2009). The assembled spliceosome excises the introns in two chemical steps: in the first, the 5′ splice site is cleaved and the lariat formed; in the second, the lariat intron is excised and both exons ligated (Jurica and Moore, 2003). During this process, the association of U1 small nuclear ribonucleoprotein with the 5′ splice site is important for determining the efficiency of the spliceosome assembly (Seraphin and Rosbash, 1989).

Suppressor screening is a widely used method of genetic screening in a diversity of organisms, such as Arabidopsis (Arabidopsis thaliana; Daszkowska-Golec et al., 2013), S. cerevisiae (Hartwell, 1991), and Drosophila melanogaster (Nüsslein-Volhard and Wieschaus, 1980). Suppressor screening can identify new signaling components and factors that characterize the pathway; it also facilitates the discovery of interactions between the components that are already known in signaling pathways. Suppressor mutants can be used to further study gene functions and analyze signaling transduction pathways. Suppressor screening has been adopted in Arabidopsis to identify gene/mutant function in a large number of pathways, including morphogenesis (Krishnakumar and Oppenheimer, 1999), jasmonates (Xiao et al., 2004), auxin (Parry et al., 2006), abscisic acid (Daszkowska-Golec et al., 2013), photosynthesis (Barkan et al., 2006), and biotic stress responses (Kwon et al., 2004).

In most cases, suppressor screening generates mutations outside the target gene, but in this study, we obtained mutations inside the target gene. We screened for suppressors of the jhs1 mutant through ethyl methane sulfonate (EMS) mutagenesis and identified three different suppressors that displayed phenotypic recovery of the jhs1 mutant. Moreover, these recovered plants can partly reduce the sensitivity of jhs1 to DNA-damaging reagents. Sequencing results showed that there was another mutation within the JHS1 gene for each of these three mutants that made JHS1 produce functional proteins that were active in vivo. Structural analyses revealed that the protein structures of JHS1 in the three recovered mutants were similar to that in the wild type, which explains how these three suppressors restored the phenotypes of the jhs1 mutant.

RESULTS

Phenotypic Analysis of the jhs1 Suppressors

Three suppressors of jhs1, k28-5 j, k48-1-1 j, and k109-6 j, were screened from the M2 generation by EMS mutagenesis of jhs1 mutants. The growth of true leaves was improved in the seedlings of the suppressor and was similar to that of the wild type (Fig. 1A). For 8-d–old seedlings grown on half-strength Murashige and Skoog (MS) medium, root length did not differ between wild type and the suppressors, while the root length of jhs1 mutants was significantly shorter than that of wild type plants (Fig. 1, C and D). A previous study showed that the most notable characteristics of the jhs1 mutant are its flat stem and irregular arrangement of flowers (Jia et al., 2016). However, these notable phenotypes of jhs1 mutants were recovered to normal in the three suppressors (Fig. 1B).

Identification of jhs1 Suppressors

In the beginning, it was not clear if k28-5 j, k48-1-1 j, and k109-6 j were true suppressors of jhs1 or only seed contamination of the wild type. So we sequenced the JHS1 genes of these plants. The results showed that these candidate suppressors had another mutation besides the jhs1 mutation (Supplemental Fig. S1). The k28-5 j mutant had a C-to-T transition mutation in At1g08840, which was 65 bp upstream of the jhs1 mutation site, a G-to-A transition mutation at 3,377th bp of JHS1. The k48-1-1 j mutant had a C-to-T transition mutation in At1g08840, 58 bp upstream of the jhs1 mutation site. The k109-6 j mutant plant had a G-to-A transition mutation in At1g08840-1, 51 bp upstream of the jhs1 mutation site (Supplemental Fig. S1). Hence, k28-5 j, k48-1-1 j, and k109-6 j could be suppressors of jhs1.

Genetic Confirmation of Suppressors

Genetic transformation experiments were carried out to prove that these mutations were responsible for the phenotypic recovery of jhs1. The JHS1 genes in these suppressors were each amplified and cloned into a vector, which was then transformed into jhs1 mutants. PCR amplification was used to verify that the target gene had been integrated into the genome of these transgenic plants. The results showed that extraneous gene fragments could be detected in k28-5-T j, k48-1-1-T j, and k109-6-T j (Supplemental Fig. S2A). Furthermore, genomic DNA in wild-type, jhs1, and transgenic plants was extracted and sequenced. Doublet peaks could be seen in the sequencing results of k28-5-T j, k48-1-1-T j, and k109-6-T j (Supplemental Fig. S2B). These data suggest that these transgenic plants are true transformants and can therefore be used in subsequent studies.

Phenotypic observation of wild-type, mutant and transgenic plants showed that k28-5-T j, k48-1-1-T j, and k109-6-T j rescued the slow growth in true leaves of jhs1 seedlings (Fig. 1A). Moreover, k28-5-T j, k48-1-1-T j, and k109-6-T j recovered the root length of jhs1 to be close to that of the wild type in 10-d–old plants (Fig. 1, C and D). For bolted plants, these transgenic plants also alleviated the phenotype of fasciated stems and abnormal silique arrangements in jhs1 (Fig. 1B). These results suggested that the second mutations largely restored the growth phenotype of jhs1.

Analysis of the Physiological Function of the Suppressors

Our previous studies showed that jhs1 is sensitive to DNA-damaging agents, and it is unclear whether suppressors or transgenic plants can reverse this characteristic. Therefore, hydroxyurea (HU) and zeocin were used to treat the wild type, jhs1, suppressors, and transgenic plants. Compared with the wild-type control, transgenic plants could recover the short root phenotype of jhs1 when treated with HU or zeocin. Similarly, suppressors could partly change the sensitivity characteristic to HU or zeocin (Fig. 2). k28-5 j was more sensitive to zeocin treatment than the other two mutants, but k28-5 T j grew better and looked like the wild type when treated with zeocin. This is probably because there is another mutation in k28-5 j that affects root growth. These results indicated that suppressors and transgenic plants can partly restore the physiological function of JHS1.

cDNA Sequence Analysis of Suppressors

Our previous study showed that the mutation in jhs1 resulted in a 63-bp shift of the splice site upstream, which led to premature termination of the protein (Jia et al., 2016). Genomic DNA sequencing indicated that a second mutation occurred in k28-5 j, k48-1-1 j, and k109-6 j (Supplemental Fig. S1). To investigate the impact of the second mutation on these three suppressors, cDNAs of the suppressors were amplified by reverse transcription (RT)-PCR and analyzed by PAGE, and different bands were seen as a result of alternative splicing (Fig. 3A). Meanwhile, JHS1 and its mutant versions (jhs1, k28-5 j, k48-1-1 j, and k109-6 j) with the native promoter and 3′ UTR were expressed in young leaves of Nicotiana benthamiana transiently, and similar results were observed (Fig. 3B). Sequencing results of the alternative spliced cDNA in the suppressors showed that the splice sites were also shifted upstream, like in jhs1 (Fig. 3C). Specifically, in k28-5 j and k48-1-1 j, the 3′ splice site of the 11th intron was shifted upstream, causing 51 additional bases to be added to the front of the 5′ end of the 12th exon. In k109-6 j, the 3′ splice site of the 11th intron was also shifted upstream, causing 36 additional bases to be added to the front of the 5′ end of the 12th exon (Figs. 3C and 4, A and B).

To determine whether the additional cDNA in the suppressors results in out-of-frame or premature termination of the protein, we analyzed the encoded sequences. Unlike jhs1, the translation reading frames in suppressors did not change and no stop codons were produced within the added bases (Fig. 4C). Therefore, the JHS1 protein in suppressors was successfully translated due to a second mutation, and 17 amino acids were inserted into k28-5 j and k48-1-1 j and 12 amino acids into k109-6 j (Fig. 4C).

Evaluation of 3′ RNA Splicing Modes

To reveal the possible recovery mechanism of suppressors, we statistically analyzed the 20 bases before the splice sites in the Arabidopsis genome and evaluated the effect of the mutation on the probability of the splice sites (Fig. 5). In general, the splice sites preferred C or T at the -1 position and were biased against C at the -2 position. For the sequences of -3 and upstream, T was preferred. In jhs1, the wild-type splice site S4 was abolished, and two additional splice sites, S1 and S5, were selected (Figs. 3, A and B, and 5, A and D). In k28-5 j, the second mutation turned C to T right before the S1 (Fig. 5, A and B), which greatly reduced the splicing at this site and consequently increased the splicing at S2 (Figs. 3, A and B, and 5, C and D). In k48-1-1 j, a C-to-T transition mutation occurred 5 bp before S2 (Fig. 5, A and B), which increased the splicing at S2 and reduced the splicing at S1 either directly or indirectly (Figs. 3, A and B, and 5, C and D). In k109-6 j, a G-to-A transition mutation abolished S2 (Fig. 5, A and B), which increased the splicing at S3 and reduced the splicing at S1 (Figs. 3, A and B, and 5, C and D). It seems the “A/T”-rich sequence upstream of the splice site will increase the chance of splicing. This can be explained by the selection of S1 but not S3 in jhs1 and reduction of the splicing of S1 in k48-1-1 j and k109-6 j due to the downstream C-to-T or G-to-A transitions. Overall, mutations in all three suppressors reduced the splicing efficiency at S1 and increased the splicing efficiency at S2 or S3, either directly or indirectly.

Analysis of Protein Structure and Function in Suppressors

JHS1 has a conserved DNA2 domain ranging from the 340th to the 543th amino acid, and the added protein segments in the suppressors were located within this domain. In Figure 6A, the structures of this conserved domain in the wild type and suppressors were predicted by homology modeling, and this prediction was based on the structure of a known template of DNA2 in mice (Zhou et al., 2015). The results suggested that a small segment of protein was added to the 496th to 497th amino acid of JHS1 in k28-5 j, k48-1-1 j, and k109-6 j. To verify whether this insertion changes the function of JHS1 in suppressors, we compared the structures of the DNA2 domain in the model of the suppressors with that in mice (5EAW in the Protein Data Bank database, https://www.rcsb.org/) and found that the counterpart of suppressors was located in the nuclease domain in the mouse model (Zhou et al., 2015). The DNA2 nuclease domain, which is responsible for resecting DNA, contains αβββαβ folds (Zhou et al., 2015). However, the insertions were not located in these key structures, but in an outside loop region; the addition of small insertions in these suppressors seemed to have little effect on the function of JHS1, which is the main reason for the phenotypic recovery. To further prove this explanation for the phenotypic recovery, we analyzed DNA damage in the root tips of the suppressors by TUNEL assay (Fig. 6B). DNA damages were found in many jhs1 mutant cells, but very few suppressor cells, suggesting the mutant JHS1 proteins in the suppressors are functional. Therefore, the functional recovery of JHS1 in the suppressors is a key factor of the phenotype restoration of the jhs1 mutant.

DISCUSSION

In this study, we adopted a genetic screening method for suppressors to further investigate the function of JHS1. We initially obtained three jhs1 suppressors. These suppressors were mutated in the JHS1 gene, and we studied the mechanism of them. The results indicate that, through a second mutation in the jhs1 gene, these suppressors circumvented the aberrant splicing of jhs1 to produce functional JHS1 proteins. Despite the addition of a small insertion in the random coil, the function of JHS1 was not affected in suppressors (Fig. 6). Therefore, these suppressors enabled jhs1 to regain wild type-like phenotypes of DNA damaging agent low-sensitive root elongation, normal true leaf growth, and root length (Figs. 1 and 2).

AG at the 3′ splice site of the intron is essential for the early steps of spliceosome assembly (Voith von Voithenberg et al., 2016). Intronic and exonic nucleotides adjacent to the conserved nucleotides are critical for splice site selection (van den Hoogenhof et al., 2016). To reveal the possible recovery mechanism in the suppressors, sequences upstream of the 3′ splice sites in the Arabidopsis genome were analyzed to learn the probability of finding each kind of base at each site and some patterns were found (Fig. 5, C and D). These patterns can be used to help explain the selection of the splice sites in the mutants used in this study.

In k28-5 j, the C right before the S1 splice site was mutated to T, which is less favorable according to our statistical analysis and experimental results (Figs. 3 and 5C). This increased the chance of splicing at S2 and S5. A small amount of mRNA from the S2 site could produce enough functional protein to rescue the mutant phenotype.

In k48-1-1 j, a C-to-T mutation occurred downstream of the S1 splice site and upstream of the S2 splice site. This mutation lead to the S2 site having more T upstream, increasing the chance of splicing at this site and rescuing the mutant phenotype.

The splicing at S2 increased in both k28-1 j and k48-1-1 j but through different mechanisms. k28-5 j made the S2 splice site more favorable by greatly reducing the splicing at S1, while k48-1-1 j seemed to increase splicing at S2 directly.

k109-6 j had no S2 site. However, the splicing at S3 was greatly increased to produce a dominant band, and the splicing at S1 was significantly reduced to produce a minor band, suggesting that this change to A is actually favorable for the splicing of S3. In jhs1, S1 was favored over S4, which is close to S5, the splice site in the wild type. These mutations probably affected the structure of the original splice sites and made them much less favorable.

Overall, all of these mutations disturbed the original splicing mode by reducing splicing at certain sites and increasing it at others (Fig. 5D). Our data and analysis provide useful information on the intron splice site selection. In the future, with the integration of our data and other data, a computer program may be developed to predict the probability of this kind of alternative splicing, caused by either artificial or naturally occurring mutations in introns.

DNA2 was first discovered in yeast, and homologs have since been discovered in other species, including human, Pan troglodytes, rat, and Arabidopsis. Phylogenetic analysis showed that DNA2 in different species is relatively conserved (Jia et al., 2017). Many unstructured or partially structured proteins are inclined to be degraded by protease (Wright and Dyson, 1999). Why is JHS1 in suppressors not degraded, but instead restores the phenotype of jhs1? On one hand, the DNA2 domain of JHS1 is similar to the 78th to the 283th amino acid region of DNA2, which is localized to the N terminus of a nuclease domain, containing the core αβββαβ-fold of the PD-(D/E)XK nuclease superfamily (Zhou et al., 2015). Our results showed that the insertion was located in the disordered region of JHS1, which has no known critical function (Fig. 6A). The ATP binding sites, the sequences (MPGTGKTT) that encode the endonuclease and helicase activities, were highly conserved, and their counterparts in suppressors were not altered. On the other hand, although DNA2 and its homologs were conserved in other species in evolutionary terms, there still exist some divergences among them, including the loop region mentioned above. Therefore, even though a small segment was inserted in JHS1, the mutant form of JHS1 is at least partially functional and could rescue the mutant phenotype of jhs1 (Fig. 6B). In terms of evolution, small insertions or deletions often occurred at the junctions of exons. It could be caused by a mechanism similar to those in the suppressor mutants in this study.

Suppressor screening is a common method for investigating the functions of target genes. Many research groups have used this method to obtain essential members in signaling pathways of target genes (Hua and Meyerowitz, 1998; Xiao et al., 2004). Therefore, suppressor screening can be adopted to help find JHS1-related genes. However, we identified three suppressors, all of which were mutated in JHS1; these are different from other suppressor screenings, which often gain new genes. This may be due to the nature of the mutation. The phenotypes of the internal mutation of JHS1 are more similar to those of the wild type than those of the suppressor mutants in some signaling pathways, which may have other additional phenotypes. Therefore, these suppressors are more likely to be detected during screening. On the other hand, because our suppressor screening identified mutations that caused alternative splicing of the mutant gene, it could be a strategy for screening suppressors of other splicing mutants and then studying the mechanism of mRNA splicing.

MATERIALS AND METHODS

Plant Materials and Growth Conditions

All Arabidopsis (Arabidopsis thaliana) plants used in this study were of the Columbia (Col) ecotype. Seeds were sowed on half-strength MS medium containing 1% (w/v) Suc and 0.8% (w/v) agarose. After stratification at 4°C for 2 ∼ 3 d, they were cultivated in a growth chamber (22°C) with 16-h light/8-h dark cycles.

For floral organ phenotypic experiments, 8-d–old seedlings were transferred to soil and cultivated for 27 d in a growth chamber before being photographed. For root phenotypic experiments, the seedlings were grown in half-strength MS or half-strength MS supplemented with different concentrations of HU or zeocin for 7 d before they were photographed.

Mutants Screening

Seeds of jhs1 mutants were mutated with EMS. The M2 generation seedlings with normal apical growth were selected as lines of interest. Meanwhile, their genetic backgrounds were identified by PCR and sequencing to guarantee the existence of the jhs1 mutation site.

Constructions, Transformation, and Transient Expression

JHS1 gene and its mutants (jhs1, k28-5 j, k48-1-1 j and k109-6 j) with the native promoter and 3′ UTR were amplified with primers JHS1-41 and JHS1-4 and cloned into a binary vector 3302Y2 between the SmaI and MluI sites. Then the constructed vectors containing a 10.259-kb–long DNA fragment were identified by sequencing with a series of primers (Supplemental Table S1). The designed constructs were introduced into jhs1 by Agrobacterium tumefaciens-mediated transformation. The transformants were identified by phenotypic analysis, PCR with primers BL17 and JHS1-44, and DNA sequencing analysis. Agrobacteria with constructs of various versions of the JHS1 gene were injected into young leaves of Nicotiana benthamiana for transient expression.

RT-PCR and cDNA Sequencing

Total RNAs were extracted from 11-d–old Arabidopsis seedlings grown on half-strength MS medium or N. benthamiana leaves with transient expression as described in the RNAPure Isolation Kit (AidLad). First-strand cDNA was synthesized from 3 μg of total RNA using RevertAid Premium First Strand cDNA Synthesis Kit (Thermo Fisher Scientific).

For cDNA analysis of suppressors, cDNA was amplified with primers JHS1-1 and JHS1-38 (Supplemental Table S1). The PCR products were analyzed by gel electrophoresis and the bands of interest were excised for sequencing. The sequencing results were analyzed by BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi) and BioEdit Sequence Alignment Editor Software (https://bioedit.software.informer.com/).

DNA Silver Staining

DNA Silver staining was performed based on Hu et al. (2009) with some modifications. PAGE gels (12%) were used to separate bands of interest. The gel was washed twice with double-distilled water after electrophoresis and transferred to 1 g/L AgNO₃ solution and shaken gently for 8 min in the dark. After being washed with water, the gel was immersed in chromogenic solution (1.5% NaOH [w/v], 1.25% formaldehyde [v/v]) until the DNA bands became clear.

Evaluation of 3′ RNA Splice Sites

To calculate the 3′ RNA splice site probability in the Arabidopsis genome, intron sequences were retrieved from The Arabidopsis Information Resource (www.arabidopsis.org), the last 20 bp of introns were retrieved by a Python program (https://www.python.org/) and the probability of the bases at each site was analyzed with the software WebLogo (Schneider and Stephens, 1990). The effect of a mutation on the probability of certain splice site was evaluated by looking whether it will increase or decrease the probability.

Homologous Modeling of the DNA2 Domain of Suppressors

The JHS1 gene has a conserved DNA2 domain from the 340th to the 543th amino acid, where the second mutation sites in three suppressors were located; this may affect the functions of the DNA2 domain. To make homologous models for the DNA2 domain in the wild type and suppressors, the corresponding amino acid sequences were submitted to the SWISS-MODEL databank (https://swissmodel.expasy.org/; Waterhouse et al., 2018). The predicted results were analyzed and annotated with the software PyMOL v2.0.6 (https://pymol.org/2/).

TUNEL Assay

To detect DNA damage in the wild type, jhs1, and suppressors, a TUNEL assay was carried out with the One Step TUNEL Apoptosis Assay Kit (Beyotime) using the method from Jia et al. (2016).

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Genomic DNA sequencing analysis of the wild type, jhs1, and suppressors.

• Supplemental Figure S2. Identification of transformants.

• Supplemental Table S1. Primers used in this study.

ACKNOWLEDGMENTS

We thank professor Zhi Lu of Tsinghua University for the assistance in RNA sequence analysis.

LITERATURE CITED

Author notes