Introduction

The eukaryotic ribosome is the protein synthesis machinery that underlies the foundation of life. It is composed of a complex of four rRNA groups (18S, 5.8S, 25S, and 5S) and approximately 80 ribosomal proteins (Dörner et al. 2023). In plants, ribosome biogenesis undergoes a stepwise process concurrent with transcription by RNA polymerase I in the nucleolus, where numerous ribonucleoprotein factors progressively process the 45S pre-ribosomal RNA (rRNA). This process results in the maturation of 18S, 5.8S, and 25S rRNAs. 18S rRNA associates with ribosomal protein small subunits, whereas 5.8S and 25S rRNAs, along with the 5S rRNA transcribed by RNA polymerase III in the nucleoplasm, associate with ribosomal protein large subunits, to form small 40S and large 60S ribosomal subunits, respectively (Henras et al. 2015; Sáez-Vásquez and Delseny 2019). In mammals, yeast, and plant cells, the 40S ribosomal subunit contains approximately 30 ribosomal protein small subunit components, whereas the 60S ribosomal subunit contains 40–48 ribosomal protein large subunit components (Armache et al. 2010; Ben-Shem et al. 2011; Xiong et al. 2021). These manufacturing processes involve multiple stages and require highly intricate control of over 200 ribosome biogenesis factors (Henras et al. 2015; Sáez-Vásquez and Delseny 2019).

In animal cells, perturbations in ribosome biogenesis often lead to cell cycle arrest, senescence, and cell death, along with abnormalities in the structure of the nucleolus, which is known as ribosomal or nucleolar stress response. Owing to its direct link to cancer therapy, this response has been intensively studied, with the transcription factor p53 known to be a key factor in it (Lafita-Navarro and Conacci-Sorrell 2023; Lindström et al. 2022). However, research on this aspect in plants has been delayed, mainly because of the absence of a homologous gene for p53 (Rutkowski et al. 2010). The plant-specific transcription factor SRIW1/ANAC082 was found to be an important factor in ribosomal stress signaling in plants (Ohbayashi et al. 2017), and several reports on ribosomal stress have since been published (Choi et al. 2020; Hang et al. 2021; Maekawa et al. 2018; Wang et al. 2020). Additionally, prior to the keystone report (Ohbayashi et al. 2017), several characteristic phenotypes of ribosome-related mutants were reported, including delayed germination, growth defects, enhanced resistance to translation inhibitors, and abnormalities in vascularization and leaf shape (Abbasi et al 2010; Fujikura et al. 2009; Horiguchi et al. 2011; Weis et al. 2015; Zhu et al. 2016). However, it still needs to be determined whether these characteristic phenotypes of ribosome-related mutants are due to a response to ribosomal stress.

In addition to mutations in ribosome-related factors that affect leaf shape, the mutations are frequently reported to significantly enhance the leaf polarity defects found in asymmetric leaves 1 (as1) and as2 (Machida et al. 2015). AS1 and AS2 are transcription factor pairs that directly or indirectly repress the expression of abaxial-specific genes, respectively (Husbands et al. 2015; Iwasaki et al. 2013). The leaves of these single mutants have wavy surfaces with asymmetrically lobed, downward-curling leaf blades. Further mutations in ribosome-related factors in the as1 or as2 mutants frequently lead to defects in the development of the abaxial side of leaves, resulting in trumpet- or needle-like leaves (Horiguchi et al. 2011; Machida et al. 2015 and 2022). The genes that enhance the as1 or as2 leaf phenotypes are called modifier genes (Machida et al. 2022).

Recent studies have revealed that inosine monophosphate dehydrogenase (IMPDH), an enzyme involved in nucleotide metabolism, is also involved in ribosomal stress response in animals. IMPDH catalyzes a rate-limiting step in GTP de novo biosynthesis and oxidation of inosine monophosphate (IMP) to xanthosine monophosphate (XMP) (Witte and Herde 2020, Fig. 1). IMPDH has two domains: the catalytic domain, where IMP and NAD+ are converted to XMP and NADH; and a regulatory cystathionine beta synthase (CBS) domain, also called the Bateman domain (Burrell and Kollman 2022; Johnson and Kollman 2020). Because IMPDH is an essential enzyme for genome replication and cell proliferation, it has been intensively analyzed as an anti-cancer target in humans (Kofuji and Sasaki 2020). For instance, pharmacological or genetic inactivation of IMPDH, which is overexpressed in glioblastoma, induces ribosomal stress (nucleolar stress) responses, including nucleolus size reduction and inhibition of cell proliferation (Kofuji et al. 2019). However, despite its importance, IMPDH is yet to be studied in plants.

Fig. 1
figure 1

Model of the GTP biosynthesis pathway in Arabidopsis IMPDH catalyzes the conversion of IMP to XMP. This model was based on a previous study by Witte and Herde (2020). IMP inosine 5’-monophosphate, XMP xanthosine 5’-monophosphate, GMP guanosine 5’-monophosphate, GDP guanosine 5’-diphosphate, GTP guanosine 5’-triphosphate. The dotted arrows represent multiple steps

Full size image

This study analyzed a loss-of-function mutant of IMPDH genes in Arabidopsis thaliana (L.) Heynh. with ribosomal stress responses. The mutants not only displayed canonical ribosomal stress responses, such as growth defects and small nucleoli, but also plant-specific ribosomal stress responses, such as pale-green leaves, abnormally processed and reduced levels of plastid rRNA, and enhanced leaf morphological abnormalities due to multiple mutations in as2. Thus, this study reveals new aspects of ribosomal stress in plants.

Material and methods

Plant material, growth conditions, and leaf shape observation

A. thaliana accession Col-0 was used as the wild-type, and all T-DNA insertion lines and mutants used in this study were in the Col-0 background. Seeds of three T-DNA insertion lines, SAIL_5_B10C1 for impdh1-1, SALK_201269 for impdh2-1, and SALK_203056 for impdh2-2, were obtained from the Arabidopsis Biological Resource Center (Alonso et al. 2003). The T-DNA insertions were confirmed by means of PCR using the primers listed in Table S1. The seeds were sterilized and sown on solid plates containing half-strength Murashige and Skoog (1/2 MS) medium supplemented with 1% sucrose and 0.4% gellan gum. After cold treatment for two days, the plates were placed at 23 °C under diurnal light conditions (16 h light/8 h dark, ~ 70 μmol m–2 s–1). After two weeks of growth, the seedlings were transferred onto rock wool covered with powdered peat moss and grown at 23 °C under continuous light conditions (~ 70 μmol m–2 s–1). For MPA treatment experiments, MPA was used as a 10 mM concentration stock in 100% ethanol as a solvent. For the guanosine treatment experiments, guanosine was dissolved in 0.1 M NaCl as a solvent to a concentration of 200 mM and heated at 95 °C for 1 min, to dissolve completely to make a stock. In the MPA and guanosine treatment experiments, reagents were added to the 1/2 MS medium so that the amount of solvent retained a constant volume under each reagent concentration condition. For the translation inhibitor treatment experiments, 100% ethanol was used as the solvent for chloramphenicol, and water was used as the solvent for spectinomycin dihydrochloride pentahydrate and streptomycin sulfate. For each translation inhibitor, the concentration at which the wild type showed approximately 50% growth inhibition when grown for 7 days was determined, and root length was measured at that medium concentration. Healthy seeds harvested simultaneously from plants that had been grown together under the same growth conditions were used in all the physiological experiments.

Gene expression analysis

Total RNA was prepared from seven-day-old whole plants grown on 1/2 MS medium containing 1% (w/v) sucrose, using Trizol™ reagent (Thermo Fisher Scientific), and reverse transcription was performed using the PrimeScript™ RT Reagent Kit with oligo dT primers and random hexamers (Takara Bio). Semi-quantitative PCR was performed using normalized cDNAs. Quantitative PCR was performed with a QuantStudio™ 12 K Flex Real-time PCR System (Thermo Fisher Scientific), using TB Green™ Premix Ex Taq II (Takara Bio). Relative gene expression levels were calculated using the ΔΔCT method (Livak and Schmittgen 2001) and normalized to the expression levels of ct18S rRNA. The primers used for the expression analysis are listed in Table S1.

RNA gel-blot analysis

For RNA gel-blot analysis, 0.5 µg of total RNA from seven-day-old whole plants were separated by means of formaldehyde-agarose gel electrophoresis (1.2%, w/v) and transferred onto a positively charged nylon membrane (Roche) via capillary transfer. After ultraviolet crosslinking of the RNA to the membrane, hybridization was performed using a DIG Northern Starter Kit (Sigma). Probe DNAs were amplified using the appropriate oligonucleotide pairs (Table S1) and T7 RNA polymerase according to the instructions in the DIG Northern Starter Kit (Sigma). Hybridization signals were detected using a digital imaging system (LAS 4000mini, GE Healthcare).

Chlorophyll quantification

Chlorophyll quantification was performed using a batch of 30–40 cotyledons, as described by Porra et al. (1989).

Transient subcellular localization analysis

To generate plasmids for the transient expression assay, we obtained IMPDH1 cDNAs by RT-PCR using RNA from Col-0 seedlings as a template. The primers used are listed in Table S1. The resulting PCR products were cloned into the pENTR/D-TOPO vector (Thermo Fisher Scientific). To construct binary plasmids for the expression of IMPDH1 fused to the N-terminus of GFP (p35S::IMPDH1-GFP), IMPDH1 cDNA was introduced into the pH35WG vector (G.H. and Hirokazu Tsukaya, unpublished). All PCR products and inserts were verified by DNA sequencing.

The subcellular localization assay was performed by the co-infiltration of Agrobacterium into Nicotiana benthamiana leaves (D'Aoust et al. 2009; Kapila et al. 1997), as previously described (Maekawa et al. 2014). As a suppressor of RNA silencing, Agrobacterium with pBICp19 was simultaneously infected into the leaves (Takeda et al. 2002). Two days after infection, GFP fluorescence in the leaves was observed under an LSM 800 laser scanning microscope (Carl Zeiss).

Visualization and quantification of the nucleolus using ethynyl uridine

Nucleolus visualization of root cells was performed as previously described (Dvořáčková and Fajkus 2018), with minor modifications. Four-day-old plant seedlings grown vertically on 1/2 MS medium containing 1% (w/v) sucrose were incubated in 500 ml water with HYPONeX (0.5 g/l) and 100 µM ethynyl uridine for 3 h in the dark. After incubation, the seedlings were fixed in freshly prepared 4% formaldehyde/1 × PBS/0.5% Triton™ X-100 for 20 min, followed by incubation in 4% formaldehyde/1 × PBS/1% Triton™ X-100 for 25 min. Fixation for the first two minutes was performed under vacuum. After fixation, the seedlings were washed for 2 min in 1 × PBS, 2 min in 1 × PBS/135 mM glycine, and 2 min in 1 × PBS again, and proceeded to the click-iT™ reaction with Alexa Fluor™ 488 dye, using a kit (Thermo Fisher Scientific), as described previously (Dvořáčková and Fajkus 2018). The fluorescence was observed using an LSM 800 laser scanning microscope. To quantify the maximum area of the nucleolus in each cell, the captured Z-stacks were overlaid in ImageJ (Fiji) and their contours were traced and quantified. The nucleoli of root cells used for quantification were selected according to the following three criteria: 1. epidermal cells in the transitional zone; 2. the cells within five cells from the root cap side; and 3. the cells were trichoblasts (root hair cells with a large nucleolus), always flanked by atrichoblasts (non-root hair cells), and were vertically shorter than atrichoblasts.

Results

Homologous genes of IMPDH in A. thaliana

A previously cloned IMPDH gene in A. thaliana (IMPDH1, At1g79470) (Collart et al. 1996) and its homologous gene (IMPDH2, At1g16350) (Witte and Herde 2020), share high similarity in their amino acid sequences (identity: 84%; similarity: 91%) (Fig. S1). A comparison of the amino acid sequences of AtIMPDH1 and AtIMPDH2 with those of human HsIMPDH1 and HsIMPDH2 showed that the binding sites for NAD+ and IMP and the cysteine residues essential for enzyme activity were all conserved in AtIMPDH1 and AtIMPDH2 (Fig. S1). However, the N-terminal Bateman domain, which controls enzyme activity, is less conserved between human and Arabidopsis IMPDHs, suggesting that the regulatory mechanism of these two may differ (Fig. S1). Next, we compared the expression levels of AtIMPDH1 and AtIMPDH2 (hereafter referred to as IMPDH1 and IMPDH2, respectively) based on publicly available RNA-seq data and found that IMPDH2 was more highly expressed than IMPDH1 in most tissues. The only exception was mature pollen, where IMPDH2 expression was barely detectable (Fig. S2). These data suggested that IMPDH2, rather than IMPDH1, has a crucial role in most organs.

Embryonic lethality in impdh1 impdh2 double mutants

To investigate the physiological roles of IMPDH1 and IMPDH2, we established Arabidopsis T-DNA insertion lines containing impdh1-1 (SAIL_5_B10C1) with an insertion in the third exon, as well as impdh2-1 (SALK_201269) and impdh2-2 (SALK_203056), both with an insertion in the second exon (Fig. 2a, b). The expression levels of IMPDH1 and IMPDH2 across these T-DNA insertion sites were not detected in any of the mutants (primer pairs C-E and a-c, respectively). These expression levels were generally lower than those observed in the controls before (primer pairs A-B and a-b) and after (primer pair d-e) the T-DNA insertion sites, except for those of transcripts between primers F-G in impdh1-1 (Fig. 2a, c). Next, we attempted to generate double mutants of both the genes. The impdh1-1 and impdh2 single mutants showed no defects in seed development, but IMPDH1/impdh1-1 impdh2-1 and IMPDH1/impdh1-1 impdh2-2 (hereafter referred to as ‘impdh1(+ /–) impdh2-1’ and ‘impdh1(+ /–) impdh2-2’, respectively) developed aborted seeds, suggesting that impdh1 impdh2 double mutants are defective in gametogenesis and/or embryogenesis (Fig. 2d). We then tested the segregation ratios of impdh1(+ /–) impdh2-1 and impdh1(+ /–) impdh2-2, which confirmed the lethality of the impdh1 impdh2 double mutants (Table 1), with the ratio of IMPDH1/IMPDH1 to IMPDH1/impdh1-1 heterozygous plants clearly less than the expected ratio of 1:2 (Table 1). The segregation ratio of the reciprocal crossing between impdh1(+ /–) impdh2-2 and impdh2-2 showed that although both cases were negatively affected, the effect was more severe when the females were mutated in impdh1-1 impdh2-2 (Table 2). These results strongly suggested that both genes are redundant and are important for development.

Fig. 2
figure 2

Characterization of impdh1-1, impdh2-1, and impdh2-2 mutant plants and impdh1 impdh2 lethality. a Schematic representation of IMPDH1 and IMPDH2. White and black boxes indicate the untranslated and coding regions of the exons, respectively. The T-DNA insertion sites of the mutant alleles impdh1-1, impdh2-1, and impdh2-2 are indicated. The arrows indicate the positions of the primers used for genotyping or RT-PCR. b. PCR-based genotyping. Primers indicated in (a) were used for the PCR amplification of impdh1-1 (C-T), IMPDH1 (C-D), impdh2-1 and impdh2-2 (a-t), and IMPDH2 (a-c). c. RT-PCR analysis of the IMPDH1 and IMPDH2 transcripts. The primers used are shown in (a). Transcripts containing regions upstream (A-B) and (a-b) or downstream (F-G) and (d-e) or across the T-DNA insertion sites (C-E) and (a-c) of impdh1-1, impdh2-1, and impdh2-2. TUB4 was used as an internal control. d. Photographs of seeds in fruits. The arrowheads indicate the aborted embryos. Scale bar: 1 mm. i1(+ /–) i2-1 and i1(+ /–) i2-2 indicate IMPDH1/impdh1-1 impdh2-1 and IMPDH1/impdh1-1 impdh2-2, respectively

Full size image
Table 1 Segregation of self progeny
Full size table
Table 2 Segregation of reciprocal crosses
Full size table

Developmental defects in impdh mutant plants

Because the function of IMPDH is essential, quantitative sufficiency of IMPDH function may be important for growth. Therefore, we examined the phenotypes (shoot fresh weight, chlorophyll content in cotyledons, and primary root length) of the impdh1-1 and impdh2 mutants in detail. Although impdh1-1 showed no apparent difference from the wild-type, impdh2-1 and impdh2-2 had shorter primary root lengths and lighter shoot fresh weight after 7 days of growth (Fig. 3a–c). These results are consistent with the higher expression of IMPDH2 than that of IMPDH1 in most tissues (Fig. S2). Additionally, the impdh2 mutants displayed transient light green cotyledons and lower chlorophyll content than that of the wild-type on the 4 days of growth; the chlorophyll content of the impdh2 mutants recovered to roughly the same level as that of the wild-type by 7 days of growth (Fig. 3a, d). Next, we isolated impdh1(+ /–) impdh2 mutants to test their phenotypes at reduced IMPDH levels. Upon observation of seedlings derived from impdh1(+ /–) impdh2-1 and impdh1(+ /–) impdh2-2, some plants had severe growth defects and a light leaf color, whereas others did not (Fig. S3a). Therefore, we checked the genotypes of 20 seedlings of each phenotype and found that all the growth-defective plants were impdh1(+ /–) impdh2-1 or impdh1(+ /–) impdh2-2, whereas all others were impdh2-1 or impdh2-2 (Fig. S3a–c). Subsequently, in the following experiments, we tried to distinguish impdh1(+ /–) impdh2-1 and impdh1(+ /–) impdh2-2 from impdh2-1 and impdh2-2 based on differences in the post-germinative growth. The growth of impdh1(+ /–) impdh2-1 and impdh1(+ /–) impdh2-2 showed even stronger growth defects in both primary root length and shoot fresh weight than those of impdh2-1 and impdh2-2, respectively (Fig. 3a–c). Even after 7 days of growth, although the leaf chlorophyll content of the impdh2 mutants had recovered to the level observed in the wild-type, those of impdh1(+ /–) impdh2-1 and impdh1(+ /–) impdh2-2 remained low (Fig. 3a, d). In addition, the frequency of impdh1(+ /–) impdh2-1 and impdh1(+ /–) impdh2-2 was high in seedlings with three or four cotyledons, which was not the case for the impdh1-1 or impdh2 mutants (Fig. S4a-b). These results indicated that sufficient amounts of IMPDH are important for the integrity of chlorophyll biosynthesis, growth, and cotyledon development.

Fig. 3
figure 3

Phenotypes of the impdh mutant plants. a Photographs of shoots. Scale bars: 1 mm for days 4 and 7 and 10 mm for days 14, 21, and 28. b Photograph (left panel) and quantification (right panel) of the primary root length of seven-day-old seedlings. The horizontal lines were placed at the tip of the primary root. The vertical line indicates a scale bar of 1 cm. Error bars indicate S.D. (n = 11). c. Quantification of the shoot fresh weight of the seven-day-old seedlings. Error bars indicate S.D. (n = 11). d. Quantification of chlorophyll [chlorophyll a (Chl a) and chlorophyll b (Chl b)] content in 4- or 7-day-old leaves. Error bars indicate S.D. (n = 3). The asterisks indicate that the sample is undetectable because no cotyledons appeared from the seed coat. Statistical significance was determined by means of ANOVA followed by the Tukey–Kramer test. Means significantly different from each other (p < 0.05) have been indicated with different letters. i1(+ /–) i2-1 and i1(+ /–) i2-2 indicate IMPDH1/impdh1-1 impdh2-1 and IMPDH1/impdh1-1 impdh2-2, respectively

Full size image

Effect of mycophenolic acid (MPA) and guanosine treatments on impdh mutant plants

MPA, an inhibitor of IMPDH, causes ribosomal stress responses in animal cells, such as the suppression of rRNA synthesis, reduction of nucleolus size, and stabilization of p53 protein in specific cell types (Freedman et al. 2019Huang et al. 2008; Kofuji et al. 2019). In plants, MPA treatment caused the translocation of several nucleolar-localized proteins to the nucleoplasm (Ahn et al. 2016; Choi et al. 2020). To estimate if plant IMPDH is also affected by MPA, we tested its effect on the impdh mutants. MPA treatment of wild-type and impdh1-1 at a concentration of 3 nM had no effect on primary root elongation, whereas MPA at a concentration of 30 nM inhibited primary root elongation by approximately half (Fig. 4a). In contrast, the growth of impdh2-1 and impdh2-2 was significantly inhibited compared to the control condition upon treatment with 3 nM MPA, which had no effect on the growth of the wild-type or impdh1-1; in addition, the growth was drastically inhibited upon treatment with 30 nM MPA (Fig. 4a). These results suggested that the relatively low abundance of IMPDH in impdh2-1 and impdh2-2 inhibited root growth at lower concentrations of MPA than that observed in the wild-type and impdh1-1.

Fig. 4
figure 4

Effects of MPA and guanosine treatments on the impdh mutant plants. a MPA treatment assay. Photograph (left panel) and quantification (right panel) of primary root length of seven-day-old seedlings in the MPA-containing media. The horizontal lines were placed at the tip of the primary roots. The vertical line indicates a scale bar of 1 cm. Error bars indicate S.D. (n = 9). b Guanosine treatment assay. Quantification of primary root length (left panel) and photograph of representative shoots (right panel) of seven-day-old seedlings in the guanosine-containing media. The horizontal line indicates a scale bar of 1 cm. Error bars indicate S.D. (n = 7). Statistical significance was determined by means of ANOVA followed by the Tukey–Kramer test. Means significantly different from each other (p < 0.05) have been indicated with different letters

Full size image

There are two known pathways for GTP biosynthesis in animals and possibly in plants: de novo biosynthesis pathway involving IMPDH and a purine salvage pathway that does not involve IMPDH (Fig. 1; Kofuji and Sasaki 2020; Witte and Herde 2020). Accordingly, GTP can be synthesized from guanosine via the purine salvage pathway and does not require functional IMPDH. Therefore, we investigated whether the growth impairment caused by the impdh mutants used in this study due to insufficient GTP levels could be alleviated by guanosine treatment. The growth of impdh2-1 and impdh2-2 was significantly recovered upon treatment with 20 µM guanosine, whereas treatment with 20 µM guanosine had no effect on the growth of both the wild-type and impdh1-1 (Fig. 4b). Furthermore, in impdh1-1(+ /–) impdh2-1 and impdh1-1(+ /–) impdh2-2, treatment with 200 µM guanosine partially recovered the growth inhibition as well as the light color of the cotyledons (Fig. 4b). The results of these MPA and guanosine treatment experiments strongly suggested that the impdh1 and impdh2 mutants, especially the impdh1-1(+ /–) impdh2 mutants, lack IMPDH enzyme activity and reduce GTP levels.

Cellular localization of IMPDH1

We investigated the subcellular localization of GFP-fused IMPDH1 in tobacco leaves using a transient expression system, which revealed broad localization in the nucleoplasm and cytoplasm, similar to that of GFP alone (Fig. S5). Due to difficulties in cloning (causes unknown), it was not possible to verify the localization of IMPDH2. The results obtained for IMPDH1 were consistent with the reported cytoplasmic and nuclear localization of IMPDH in animals (Juda et al. 2014; Kozhevnikova et al. 2012) and the detection of IMPDH1 and IMPDH2 in the cytoplasmic proteome analysis of A. thaliana carried out previously (Ito et al. 2011). While IMPDH is thought to function primarily as an enzyme in the cytoplasm (Witte and Herde 2020), the possibility of nuclear localization has also been demonstrated, suggesting the existence of additional functions other than its enzymatic activity or unique regulatory mechanisms.

Involvement of ribosomal stress in impdh mutant plants

Since the inhibition of IMPDH function in animals causes ribosomal stress (Freedman et al. 2019Huang et al. 2008; Kofuji et al. 2019), we investigated whether this stress also occurred in the plant impdh mutants. Because several Arabidopsis ribosome biogenesis-associated mutants were more resistant to translation inhibitors than the wild-type (Hsu et al. 2014; Maekawa et al. 2018; Rosado et al. 2010; Zhu et al. 2016), we also tested the growth of the impdh mutants on a medium containing translation inhibitor (chloramphenicol, spectinomycin, or streptomycin) and measured the primary root length. These antibiotics used in this study target prokaryotes, thus mainly inhibiting plastid and mitochondrial translation. The primary root length of the impdh2 mutant in the control medium was lesser than that of the wild-type, which either became equivalent to or higher in the impdh2 mutants grown on the translation inhibitor-containing medium than that in the wild-type (Fig. 5). In other words, the impdh2-1 and impdh2-2 mutants were more resistant to the translation inhibitor than the wild-type and impdh1-1.

Fig. 5
figure 5

Translation inhibitor treatment assay. Photographs (upper panel) and quantification of primary root length (lower panel) of seedlings grown for seven days in translation inhibitor-containing media. The horizontal lines were placed at the tip of the primary roots. The vertical line indicates a scale bar of 1 cm. Error bars indicate S.D. (n = 10). Statistical significance was determined using a t-test followed by Bonferroni correction. Means significantly different from each other (p < 0.05) have been indicated with different letters

Full size image

Since human IMPDH inhibition-induced ribosomal stress reduces the nucleolus size, which is an important compartment for ribosome biogenesis (Kofuji et al. 2019), we tested whether the impdh mutants had a reduced nucleolus size. Detection and comparison of the nucleolus size near the root tip showed that while impdh1-1 did not differ from the wild-type, impdh2-1 and impdh2-2 showed a reduction of ~ 40% in size, as compared with the median value (Fig. 6). Owing to the experimental setup, the impdh1(+ /–) impdh2 mutants were too small to be used as test samples. These results indicated that a characteristic ribosomal stress response occurs in the impdh mutants, which depends on the level of IMPDH.

Fig. 6
figure 6

Quantification of the nucleolus size of root cells grown for four days. The upper panels show the representative nucleoli. The nucleolus was visualized by observing the green fluorescence using a click-IT™-based method (see Methods). Scale bar: 10 µm. The graph shows a box-plot of the quantification of the nucleolar size (n = 40). Statistical significance was determined by means of ANOVA followed by a Tukey–Kramer test. Means significantly different from each other (p < 0.05) have been indicated with different letters

Full size image

rRNA maturation in the impdh mutant plants

Because MPA treatment or suppression of IMPDH function also results in the inhibition of rRNA synthesis in animals (Huang et al. 2008; Kofuji et al. 2019), we attempted to detect rRNA profiles in impdh mutant plants grown for 7 days. First, quantitative PCR was used to quantify the amount of cytosolic, plastid, and mitochondrial rRNA relative to that of cytosolic 18S rRNA. The relative amounts of plastid 16S and 23S/cytosolic 18S rRNA in the impdh1(+ /–) impdh2 mutants were lower than those in the other samples, whereas the relative amounts of cytosolic 25S, mitochondrial 18S, and 26S/cytosolic 18S rRNAs were not different (Fig. 7a). Considering the reduced chlorophyll content and pale-green leaves (Fig. 3a, d), it is suggested that the relatively low amount of plastid rRNAs in impdh1(+ /–) impdh2 could be due not only to the potential shortage of GTP in the plastids but also to a reduced number of plastids or delayed plastid development.

Fig. 7
figure 7

Analysis of rRNA accumulation and processing. a Cytosolic and organellar rRNA levels were determined by means of RT-qPCR. ct, pt, and mt indicate cytosolic, plastid, and mitochondrial rRNAs, respectively. Each value was normalized to the ct18S rRNA levels. Error bars indicate S.D. (n = 3). Statistical significance was determined by means of ANOVA, followed by a Tukey–Kramer test. Means significantly different from each other (p < 0.05) have been indicated with different letters. b Diagram of the major processing pathway of cytosolic (left panel) and plastid (right panel) rRNA. The positions of the probes for the RNA-blot analysis (p1–p4) are indicated with gray boxes, while the regions amplified by means of RT-qPCR are indicated with arrows above the diagrams. c RNA gel-blot analysis of cytosolic (left panel) and plastid (right panel) rRNA processing. The probe positions are indicated in (b). Ethidium bromide (EtBr)-stained images of mature cytosolic 25S rRNA (as a loading control) are shown below each blot. i1(+ /–) i2-1 and i1(+ /–) i2-2 indicate IMPDH1/impdh1-1 impdh2-1 and IMPDH1/impdh1-1 impdh2-2, respectively

Full size image

The processing of cytosolic and plastid rRNA was detected using an RNA gel-blot with two probes (see Fig. 7b for rRNA processing steps and probe sites). There were no apparent differences in cytosolic rRNA processing patterns between the samples, along with the signals of cytosolic 25S rRNA, which was used as a control (Fig. 7c). For plastid rRNA, differences were observed in the ratio of mature 1.5 kb 16S rRNA to its precursor 1.7 kb 16S rRNA (1.5 kb/1.7 kb). The impdh2 mutant had a higher 1.5 kb/1.7 kb ratio than that of the wild type, and the impdh1( ±) impdh2 mutant showed an even more pronounced difference, while the impdh1-1 mutant did not differ from the wild type (Fig. 7c). The impdh1( ±) impdh2 mutant also had a generally lower ratio of rRNAs derived from rrn23 versus cytosolic 25S rRNAs, compared with other samples (Fig. 7c). These results indicated that impdh mutations have an inhibitory effect on the synthesis and processing of plastid rRNA, rather than those of cytosolic rRNA.

Adaxial-abaxial leaf polarity alterations in as2-1 impdh1(+ /–) impdh2 mutant plants

Mutations in plant ribosomal protein genes and ribosome biogenesis-related factors enhance the leaf abaxialization phenotype of the as2 mutant (Horiguchi et al. 2011; Machida et al. 2015). Genes that enhance the as2 phenotype are called ‘modifier genes’ (Machida et al. 2015). Therefore, we generated multiple mutant lines with as2-1 to examine whether impdh mutations had similar developmental effects. as2-1 rpl4d-3 showed a strong leaf abaxialization phenotype (Horiguchi et al. 2011) and was used as the positive control, which developed ~ 70% needle- and ~ 7% trumpet-like leaves (Fig. 8, Table 3). Observations of several multiple mutant lines revealed that the as2-1 impdh1(+ /–) impdh2 mutants developed trumpet- and needle-like leaves at a frequency of 3.4%–4.9% and 2.6%–3.0% respectively, which was not found in any other mutant line (Fig. 8, Table 3). These results indicated that the impdh1(+ /–) impdh2 mutations enhance the as2 mutation and that IMPDH serves as a 'modifier gene' of as2. Taken together, our results indicated that the impdh mutations causes both previously known and novel ribosomal stress responses, with a major impact on plastids.

Fig. 8
figure 8

Effects of impdh mutations on leaf development in the as2-1 background. Representative photographs of shoots grown for 25 days. Arrowheads indicate filamented leaves. Scale bar: 1 cm. Scale bar for each enlarged image = 1 mm

Full size image
Table 3 Frequencies of leaf shape phenotype
Full size table

Discussion

Analysis of impdh mutants in A. thaliana revealed that IMPDH is essential for plant viability and that plant-specific ribosomal stress occurs in these mutants. The impdh1 impdh2 double mutants were lethal (Fig. 2d, Table 1), while the impdh2 and impdh1(+ /–) impdh2 mutants exhibited growth defects (Fig. 3a) and a transient deficit in chlorophyll content (Fig. 3d). In addition, impdh2 mutants had a smaller nucleolus size (Fig. 6), while the impdh1(+ /–) impdh2 mutants showed abnormal processing of the 1.7 kb rRNA of the plastid rRNA, a decrease in the mature plastid rRNA (Fig. 7), and functioned as a modifier to enhance the abaxialization effect of as2 (Fig. 8). These results, especially the strong effects on plastids, provide a new perspective for ribosomal stress research in plants.

Although AtIMPDH1 and AtIMPDH2 have very similar amino acid sequences (identity: 84%; similarity: 91%), their expression patterns differ, suggesting that they have different roles. IMPDH2 is more strongly expressed than IMPDH1 in all organs, except in the pollen, as ascertained using previously published RNA sequencing data (Fig. S2), suggesting that IMPDH2 plays a more crucial role than IMPDH1. In support of this, the impdh1-1 mutant was indistinguishable from the wild-type, whereas the impdh2 mutants showed a phenotype of growth defects and low leaf chlorophyll content in the early growth stages (Fig. 3). However, the more severe phenotype of the impdh1(+ /–) impdh2 mutants than that of the impdh2 single mutants (Fig. 3), as well as the lethality of the impdh1 impdh2 double mutants (Fig. 2), strongly suggested that IMPDH1 and IMPDH2 are functionally redundant. Interestingly, among human IMPDH1 and IMPDH2, IMPDH2 is majorly expressed in most organs, whereas IMPDH1 is less expressed in general, but plays an important role in the retina (Bowne et al. 2002; Kofuji and Sasaki 2020; Mortimer et al. 2008). Similarly, Arabidopsis IMPDH1 may play some unique role in the pollen, where it is preferentially expressed.

This study did not confirm the enzymatic activity of IMPDH1 and IMPDH2, but strongly suggested that they have this function, while also suggesting additional functions other than this enzymatic activity. Treatment of the impdh2 mutants with MPA, an inhibitor of the IMPDH enzyme, was more effective in inhibiting growth than treatment of the wild-type (Fig. 4a). Additionally, treatment with guanosine, a material for the salvage pathway of GTP biosynthesis, reduced the growth inhibition in the impdh2 and impdh1(+ /–) impdh2 mutants, and the greener leaf color in the impdh1(+ /–) impdh2 mutants (Fig. 4b). These results are consistent with the idea that the proteins encoded by IMPDH1 and IMPDH2 have enzymatic activity. A comparison of the amino acid sequences of IMPDH1, IMPDH2, and human IMPDHs indicated that the amino acids required for activity were well conserved (Fig. S1), strongly supporting this idea. However, the limited alleviation of growth inhibition upon treatment of the impdh1(+ /–) impdh2 mutants with guanosine (Fig. 4b) suggested that the growth inhibition of the impdh mutants is not solely due to insufficient GTP biosynthesis, but also due to a unique function of IMPDH. As an example of another function of IMPDH, studies in Drosophila suggested that IMPDH accumulates in the nucleus during the G2 phase of the cell cycle or stresses and binds to DNA to regulate histone gene expression (Kozhevnikova et al. 2012). IMPDH1 was also found to localize to the nucleus (Fig. S5), suggesting that it functions as a transcription factor. Other studies in human cells have suggested that IMPDH may regulate translation by acting on polyribosomes through binding to either rRNA or mRNA (Mortimer et al. 2008). It is also possible that guanosine treatment alone did not fully restore the growth inhibition (Fig. 4b) due to functional defects other than those in IMPDH enzyme activity. Another possibility is that IMPDH influences plastid development by regulating the transcription or translation of nuclear coding genes in response to intracellular status, such as IMP.

As the GTP synthesis pathway includes not only the de novo pathway, in which IMPDH serves as the rate-limiting enzyme, but also the salvage pathway, the impdh mutant plants synthesized less GTP and the salvage pathway gradually relieved this deficiency. In support of this, the growth defects and pale-green leaf phenotype that appeared in the early growth stages of impdh2 mutants were not significantly different from those of the wild-type, after approximately 28 days of growth (Fig. 3a). This finding indicates that the phenotype of the small nucleolus size seen in the impdh2 mutants at 4 days of growth (Fig. 6) may no longer be observed in samples after increased growing days, and that the plastid 16S rRNA processing abnormalities and plastid rRNA deficits observed in the impdh1(+ /–) impdh2 mutants (Fig. 7) may also be observed in the impdh2 mutants after reduced growing days. Furthermore, the relatively low frequency of needle-like leaves, which were observed in multiple impdh mutations with as2 (Fig. 8, Table 3), may also be related to the complementation of GTP by the salvage pathway.

GTP not only provides material for DNA and RNA but also serves as an energy source and is responsible for signal transduction (Chakraborty and Raghuram 2023). Thus, the various defects, including the plastid abnormalities occurring in the impdh mutant, are not solely due to a lack of rRNA material, although rRNA is indeed the most demanding of GTP. A comprehensive examination of the various cellular activities involving GTP is important for elucidating the details of ribosomal stress in impdh mutants.

The impdh2 mutants, which exhibited pale-green leaves and reduced chlorophyll content (Fig. 3a, d), were relatively resistant to prokaryotic translation inhibitors. However, similar resistance to these inhibitors has been reported in cytoplasmic ribosome-related mutants that do not display plastid-associated phenotypes (Hsu et al. 2014; Maekawa et al. 2018; Rosado et al. 2010; Zhu et al. 2016). This suggests several possible explanations for the observed resistance in impdh2 mutants: 1. Plastid translation was less inhibited in the impdh2 mutants. 2. Conventional cytoplasmic ribosomal stress occurs in impdh2 mutants, contributing to the observed resistance. 3. Mutants exhibiting conventional cytoplasmic ribosome stress may be less susceptible to plastid translation inhibition due to an unknown mechanism. Further analysis of plastids in cytoplasmic ribosome-related mutants will likely provide insights into these aspects.

In this study, impdh mutants concurrently exhibit abnormal ribosome biosynthesis in the plastid (Fig. 7a, c) and smaller nucleolus, which is the site of cytoplasmic ribosome biosynthesis (Fig. 6). Although plastid-type rRNA processing occurs within the plastid and is not directly associated with the reduction in nucleolus size, some reports suggest that coordination occurs between plastid and cytoplasmic ribosome biosynthesis. For instance, LEFKOTHEA, an RNA-binding protein localized to both the nucleus and chloroplasts, plays a role in the pre-mRNA splicing of genes involved in cytoplasmic ribosome biogenesis within the nucleus and in the splicing of plastid ribosomal genes within the chloroplast (Daras et al. 2019). Thus, the observed changes in nucleolar size in the impdh mutants may be attributable to specific abnormalities in plastid ribosome biosynthesis. However, since no such abnormalities were detected in the impdh mutants, the physiological significance of the observed change in nucleolar size remains unclear. We hope that this study will serve as a basis for an improved understanding of the relationship between the biosynthetic integrity of the plastid ribosome and the regulatory mechanisms underlying nucleolar size regulation.

In this study, a plant-specific response to ribosomal stress was observed in impdh mutants. The impdh single and multiple mutants showed not only the previously reported ribosomal stress response phenotypes, such as growth retardation (Fig. 3), suppression of main root elongation (Fig. 3), resistance to translation inhibitors (Fig. 5), effect on nucleolus size (Fig. 6), and development of needle-like leaves on multiple mutations with as2 (Fig. 8), but also pale-green leaves due to low chlorophyll content (Fig. 3), defects in plastid rRNA/cytosolic 18S rRNA levels, and processing patterns (Fig. 7). Conversely, the previously reported phenotypes of abnormal cytosolic rRNA in human impdh mutants (Kofuji et al. 2019) were not identified (Fig. 7). Similar to the present study, the suppressed expression of CTPS2, a major player in the CTP synthase (CTPS) family of enzymes important for CTP synthesis, also results in growth retardation, reduced chlorophyll content, and reduced mature plastid rRNA (Bellin et al. 2021). The similarities between this report and the present study suggest that there may be meaningful reasons why defects in nucleotide metabolism in plants are more likely to affect plastids. Mutation of impdhs did not affect the relative levels of cytosolic or mitochondrial rRNA, suggesting a link between nucleotide biosynthesis and plastid function. Nucleotide transport to chloroplasts could be restricted with the depleted absolute nucleotide levels, or it is possible that nucleotides are more likely to be depleted within chloroplasts. However, ribosomal stress-like responses such as reduced rRNA levels and multiple mutations with as2 resulting in needle-like leaves have also been reported with mutations in the plastid ribosomal protein gene SCA1/PRPS5 (Mateo-Bonmatí et al. 2015), suggesting the presence of a sensing system for plastid ribosome biogenesis. The similarity of these results to those of the present study suggests overlapping signaling pathways between cytoplasmic and plastid ribosome biogenesis surveillance system. Further studies will provide us with a better understanding of whether these phenomena are caused by plant-specific plastid ribosomal stress.