Phylogeny of nucleolar Pumilio proteins
Pumilio proteins are a family of evolutionarily conserved RNA-binding proteins found in all eukaryotes. Among the organisms whose whole genome sequences are available, higher plants have a higher number of Pumilio proteins than photosynthetic single-cell organisms and nonplant organisms; for example, 25 Pumilio proteins are found in Arabidopsis thaliana, 20 in Oryza sativa, 14 in Physcomitrella patens, 5 in Chlamydomonas reinhardtii, 11 in Caenorhabditis elegans, 7 in Saccharomyces cerevisiae, and 2 in humans [10]. Based on the similarity search using Arabidopsis nucleolar Pumilio proteins (APUM23 and APUM24) as queries and the existence of nucleolar localization signal(s) (NoLS) [28], viridiplantae species, including green algae, were examined and found to have two putative nucleolar Pumilio proteins. Using our transcriptome data obtained from PacBio iso-seq, we also identified two putative nucleolar Pumilio proteins, ChPUM2 and ChPUM3, in Chara corallina.
To gain insight into the evolutionary relationship of putative nucleolar Pumilio proteins, we constructed a phylogenic tree from the homologous proteins of representative viridiplantae species (Fig. 1a). All land plants and green algae analyzed in this study have a single protein homologous to APUM23 and APUM24. Chara ChPUM2 was closer to APUM23 than APUM24, whereas ChPUM3 was categorized as being in the APUM24 clade. Our phylogenetic tree showed that Chara ChPUM2 and ChPUM3 are more homologous to the Bryophyta counterparts than to the potential nucleolar Pumilios of other land plants.
We then examined the structural similarity and numbers of Puf domains in ChPUM2 and ChPUM3 and Arabidopsis nucleolar Pumilios using their primary sequences in a SMART web program (http://smart.embl-heidelberg.de) [29]. APUM23 and APUM24 have six and five Puf domains, respectively; ChPUM2 and ChPUM3 have one and five Puf domains, respectively. As the number of Puf domains may determine the RNA base numbers, this observation raises the possibility that ChPUM2 has a distinct structure-function relationship from APUM23 and that ChPUM3 may be an APUM24 homologue (Fig. 1b).
Structure of ChPUM2 and ChPUM3
Classical structural analysis of Pumilio proteins shows the typical motif of 8 Puf domains in the C-terminal end of the protein [30, 31]; however, a more detailed analysis using human Puf-A as a representative domain showed the existence of 3 additional Puf domains at the N-terminus [14]. Therefore, we analyzed the domain homology according to Qiu et al. (2014), which enabled us to identify cryptic Puf domains that had a low homology with typical amino acid sequences in the flanking regions of the Puf core domain. It was previously reported that APUM23 and its budding yeast homologue Nop1 have 10 Puf domains when analyzed using this tool [12, 32]. Consistently, all 14 plant proteins were categorized into the ChPUM2 and APUM23 families (Fig. 1a) and contained 10 Puf domains that are distributed at irregular intervals (Fig. 2a and Additional file 1a). Each domain of ChPUM2 showed an average 34.3% identity with the corresponding domains of APUM23 with a 12.9% gap. Despite low homology between the entire sequences of each domain of ChPUM2 and APUM23, a high degree of homology was found in the 1st, 2nd, and 5th residues of the 2nd α-helix, and these residues are pivotal for recognizing RNA (red box in Fig. 2a and Additional file 1a). Overall, the Puf domains among ChPUM2 relatives are well conserved throughout the viridiplantae species examined in this study when considering the RNA recognition residues.
Using the same methodology, all the proteins that belong to the ChPUM3 and APUM24 clades (Fig. 1a) possessed 11 Puf domains (N-R1 ~ N-R3 in the N-terminus and C-R1 ~ C-R8 in the C-terminus), as reported in human Puf-A and APUM24 [14]. ChPUM3 displayed an average of 47.9% identity with APUM24 in the N-terminal Puf domains but 34.4% identity and 14.7% gap in the C-terminus. Puf domains in the N-terminus of ChPUM3 showed high homology to APUM24 in the 1st, 2nd, and 5th residues of the 2nd α-helix, but those on the C-terminus of ChPUM3 had low homology to APUM24 (red box in Fig. 2b and Additional file 1b).
Although the comparison of amino acid sequences among these Puf domains showed that ChPUM2 and ChPUM3 are homologous nucleolar proteins to APUM23 and APUM24 (Fig. 2a-b), it cannot be ruled out that ChPUM2 has a distinct RNA-binding activity from APUM23 as ChPUM2, unlike APUM23, has only a single Puf domain in the classical domain analysis (Fig. 1b). Therefore, we used the SWISS-MODEL web server (https://swissmodel.expasy.org/) to predict the tertiary structure of these proteins. It is known that APUM23 has a long chain between the 2nd and 3rd α-helix of the 3rd Puf domain and has a twisted C-shaped structure [12]. Our prediction showed a high coverage of the ChPUM2 tertiary structure with the APUM23 reference and a twisted C-shaped structure that is similar to APUM23. Although ChPUM2 has many unfolded chains in the 3rd Puf domain compared with APUM23, it has nearly identical α-helix structures to APUM23 that contribute to RNA binding (Fig. 3a). ChPUM3 and APUM24 had an L-shape structure that was similar to that found in the human Puf-A reference [14] (Fig. 3b). However, ChPUM3 had a prominently longer random coil in the C-R5 domain than APUM24 (Fig. 2b and Fig. 3b) and contained negative (E265) and uncharged (Q305) amino acids in the N-R2 and N-R3 domains of patch 1B (Fig. 2b), unlike Puf-A and APUM24. Overall, ChPUM2 had a similar structure to APUM23 except for the unfolded structures positioned outside the core sequence of the 2th, 4th, and 8th Puf domains, and ChPUM3 had a longer side chain in the C-R5 domain than APUM24 but also contained distinct amino acids in the N-terminal Puf domains that were distinct from those in APUM24.
Subcellular localization of ChPUM2 and ChPUM3
Next, we examined the subcellular localization of ChPUM2-RFP and ChPUM3-RFP. Previously, the GFP fusions of APUM23 and APUM24 were known to preferentially localize in the nucleoli of the Arabidopsis root and tobacco leaf cells [10, 15]. We performed Agrobacterium-mediated coinfiltration into N. benthamiana leaf cells using 35S:APUM23-GFP and 35S:ChPUM2-RFP, and using 35S:APUM24-GFP and 35S:ChPUM3-RFP. All the GFP- and RFP-tagged Pumilio proteins were found in the nucleolus and weakly in the nucleoplasm, probably due to high constitutive expression under the 35S promoter (Fig. 4), indicating that ChPUM2 and ChPUM3 have nucleolar functions, such as pre-rRNA recognition and processing. Our results obtained from the structural prediction and localization assays suggested a strong possibility that ChPUM2 and ChPUM3 are Chara orthologues of Arabidopsis APUM23 and APUM24, respectively.
Expression of 35S:ChPUM2 restored apum23 mutant phenotype to normal
As APUM23 and APUM24 are closely related to ChPUM2 and ChPUM3, respectively, it is likely that apum23 and apum24 mutants could be complemented by the heterologous expression of ChPUM2 and ChPUM3. To address this possibility, we generated 35S:ChPUM2 and 35S:ChPUM3 plants in the apum23-2 mutant background that shows the phenotypes of delayed germination, light green leaves, short stem and roots, and streptomycin resistance. The 35S:ChPUM2/apum23-2 plants exhibited normal phenotypes, whereas the 35S:ChPUM3/apum23-2 plants showed an apum23-2 phenotype (Fig. 5a-d). The 35S:ChPUM2/apum23-2 plants recovered seed germination efficiency, root length, plant height, and color and shape of rosette leaves. Moreover, complemented plants were sensitive to streptomycin, making them similar to control Col-0 plants and suggesting that defects in ribosome biogenesis in the apum23-2 mutant might be recovered (Fig. 5D). The failure to restore the apum23-2 phenotype to the wild-type phenotype by 35S:ChPUM2 excluded the possibility of orthologues of ChPUM3 and APUM23.
In addition to the restoration of morphological phenotypes, 35S:ChPUM2/apum23-2 complemented the defects observed in the apum23-2 mutant that accumulates polyadenylated [poly(A)] 18S-ITS1 and 5.8S-ITS2 pre-rRNAs (Fig. 5e). We synthesized the oligo (dT)-primed cDNAs from 35S:ChPUM/apum23-2 and control Col-0 plants to measure the amount of poly(A) pre-rRNA as PCR templates. For the detection of unprocessed poly(A) rRNA, RT-PCR was performed using three primer sets for 5’ETS-18S, 18S-ITS1, and 5.8S-ITS2. In the apum23-2 mutant, all three quantitative RT-PCR products (5’ETS-18S, 18S-ITS1, and 5.8S-ITS2) were more accumulated than those in the Col-0 control. However, in 35S:ChPUM2/apum23-2 plants, the amounts of poly(A) 18S-ITS1 and 5.8S-ITS2 pre-rRNAs were greatly reduced compared with those of apum23-2, although that of poly(A) 5.8S-ITS2 was slightly decreased. In contrast to 35S:ChPUM2/apum23-2, 35S:ChPUM3/apum23-2 showed a nearly identical amount of poly(A) pre-rRNAs to the mutant (Fig. 5e, right panel). Hence, qRT-PCR results using poly(A) pre-rRNA indicated that ChPUM2 is involved in pre-rRNA processing that is similar to that in APUM23, although it was not fully functional for removing the 5.8S pre-rRNA byproducts.
Chara nucleolar Pumilio protein genes did not rescue the Arabidopsis apum24 mutant
APUM24, a homologue of ChPUM3, is involved in 5.8S rRNA processing and is essential for cell division and pattern formation in early embryogenesis [13, 15, 33]. To examine if nucleolar ChPUMs recovered the apum24 mutant to wild-type, we performed a complementation experiment using a heterozygous apum24-1+/− mutant line. Homozygous 35S:ChPUM2 and 35S:ChPUM3 transgenics on the apum24-1+/− background set the normal and aborted seeds at the same rate as the apum24+/− mutant (Fig. 6a-c and Table 1). Consistent with this result, 35S:ChPUM2+/+/apum24-1+/− and 35S:ChPUM3+/+/apum24-1+/− plants accumulated poly(A) 5.8S pre-rRNA similar to that in the apum24-1+/− plants (Fig. 6d), indicating that Chara ChPUM3 is not a functional orthologue of Arabidopsis APUM23 and APUM24.
Table 1
The ratio of normal seeds to aborted seeds.
Genotype | Total seeds (n*) | Normal seeds (n) | Aborted seeds + undeveloped ovules (n) | Ratio (N : A&U**) |
Control (pB2GW7) | 3016 | 2869 | 147 | 19.5 : 1 |
apum24-1+/− | 3043 | 2108 | 935 | 2.25 : 1 |
35S:ChPUM2 / apum24-1+/− | 3035 | 2050 | 985 | 2.08 : 1 |
35S:ChPUM3 / apum24-1+/− | 2988 | 1997 | 991 | 2.01 : 1 |
*n, numbers; **A&U, normal seeds : aborted seeds & undeveloped ovules. |
ChPUM2 restored the salt- and glucose-hypersensitive phenotypes of apum23-2, but ChPUM3 did not recover apum23-2 and apum24-1
The apum23 and apum24 have changes in the expression levels of their ribosomal biosynthetic genes, thereby resulting in a hypersensitivity to high concentrations of salt or glucose [13, 17] (Fig. 7). The 35S:ChPUM2/apum23-2 seedlings exhibited a similar degree of resistance to 150 mM NaCl and 200 mM glucose to that found in wild-type seedlings, while the 35S:ChPUM3/apum23-2 seedlings showed NaCl- and glucose-susceptibility that was similar to the apum23-2 seedlings (Fig. 7a). However, 35S:ChPUM2/apum24-1+/− and 35S:ChPUM3/apum24-1+/− failed to recover the salt sensitivity of apum24-1+/−. Unexpectedly, when 35S:ChPUM2 or 35S:ChPUM3 was overexpressed in apum24-1+/−, their transgenic seedlings showed a similar glucose resistance to that found in wild-type seedlings (Fig. 7a). It was previously reported that APUM24 gene expression was greatly increased in wild-type plants by exogenously supplying glucose [13, 15]. Our data showed similar expression levels of the APUM24 transcript in the control, apum24-1+/−, 35S:ChPUM2/apum24-1+/−, and 35S:ChPUM3/apum24-1+/− (Fig. 7b), indicating a haploid sufficiency of the APUM24 gene for the expression of the glucose-induced phenotype. Consistently, apum24-1+/− and transgenic plants reduced the amount of unprocessed 5.8S rRNA to normal control levels under high concentrations of glucose (Fig. 7c). Our results on the salt and glucose treatments support that ChPUM2 complements apum23-2 and ChPUM3 does not complement apum23-2 and apum24-1.