Identification and molecular characterization of the alternative spliced variants of beta carbonic anhydrase 1 (βCA1) from Arabidopsis thaliana

Plant material and stress treatment

All Arabidopsis plants generated in this study are of the Columbia-0 (Col-0) ecotype. The T-DNA mutants lines used in this study have been described previously: βca1 (Salk_106570) and βca2 (CS303346) (Huang et al., 2017). Seeds were surface sterilized and grown on plates with Murashige and Skoog (MS) plus 1% sucrose and 0.8% phyto agar at 22 °C under a long-day (16/8 h light/dark) photoperiod with a photon flux density of 180 μmol photons m-2 s-1. 7-d-old seedlings from the plates were transferred to the soil and grown in the growth chamber. Abiotic stress and phytohormone treatments were subjected to 2-week-old seedlings. For abscisic acid (ABA) 10 μM of ABA was supplied on the solid agar medium with seedling grown vertically. For salt stress, seedlings were vertically placed in plates with 150 mM NaCl for the indicated period of time. For heat treatment, seedlings were transferred to a new plate and placed within a bake oven set at 40 °C for the indicated period of time. For analysis of the seed germination after heat treatment, seeds were sown in the dark for 2 d at 4 °C, and then the seeds were heated at 50 °C for 3 h or 6 h. After this heat treatment, the seeds were grown at 22 °C for 7 days, and then the germination rate and seedling with green cotyledon were measured.

RNA extraction and reverse transcription-polymerase chain reaction (RT-PCR)

Total RNA was extracted and purified from the 1-week-old Arabidopsis seedlings using ReliaPrep™ RNA Miniprep System (Promega, Madison, WI, USA). For cDNA synthesis, the first-strand cDNA was synthesized from 2 μg total RNA with NovoScript®Plus All-in-one 1st Strand cDNA Synthesis SuperMix (Novoprotein, China) according to the instructions. For semi-quantitative RT-PCR analysis primers were designed manually after identification of unique sites in the cDNA of four spliced variants of βCA1 (βCA1.1, βCA1.2, βCA1.3 and βCA1.4). Specific primers for amplification of βCA1.2 and βCA1.3 were applied, respectively. The primers used were mentioned in Table S1. The ACTIN2 was used as the endogenous control.

Identification and analysis of βCA1 in plants

To identify ortholog(s) of βCA1 from Arabidopsis in the representative plant genome (including Oryza sativa, Glycine max, Populus trichocarpa, Brachypodium distachyon, Sorghum bicolor and Physcomitrella patens), the amino acid sequence of AtβCA1 was used as the guide sequence to perform a BLASTp search of the database whole genome sequences in the Phytozome database (https://phytozome.jgi.doe.gov). Coding sequences (CDS) and corresponding genomic DNA (gDNA) sequences of the βCA1s from the selected plant species were retrieved from the databases. The GSDS tool (GSDS v2.0, http://gsds.gao-lab.org/) was employed to analyze the exon-intron structures for plant βCA1 genes.

Plasmids construction and plant transformation

All DNAs and cDNAs were amplified using Tks Gflex™ DNA Polymerase (Takara, Kusatsu, Japan). The CDS of four βCA1 spliced variants (βCA1.1, βCA1.2, βCA1.3 and βCA1.4) were amplified and cloned into pDONOR221 using Gateway™ BP Clonase™ Enzyme Mix (Thermo Scientific, Waltham, MA, USA). Each βCA1 spliced variant was further cloned into the pGWB405 plant expression vector designed for the production of C-terminal GFP-tagged fusion proteins under the control of the 35S Cauliflower Mosaic Virus (35S CaMV) promoter using Gateway™ LR Clonase™ II Enzyme mix (Thermo Scientific, Waltham, MA, USA). To generate the plasmid for the native expression of βCA1, the genomic region including 1.5 kb upstream of ATG plus introns and exons were cloned. For insertion of specific tag including HA or Myc, overlapping PCR was applied by designed primers. These DNA fragments were cloned into pDONOR223 and pGWB4 vectors with the same method used for βCA1 spliced variants cloning. The constructs were confirmed by sequencing. All the plant expression constructs were introduced into the Agrobacterium tumefaciens GV3101 strain. Plant transformation was performed with floral dipping method (Clough & Bent, 1998). All the primers used for cloning were listed in Table S1.

Total protein extraction and western blotting

To analyze the GFP-tagged βCA1 spliced variants, seedlings of stable transgenic plants expressing the corresponding GFP fusions were freeze grounded into powder and homogenized in total protein buffer (20 mmo1/L Tris-HCl (pH 7.5), 150 mmol/L NaCl, 2 mmol/L EDTA) with protease inhibitor cocktail in DMSO (Yeasen, China). Lysates were incubated on ice for 20 min and clarified by centrifugation at 18,407 g for 15 min at 4 °C. For immunoblotting, samples were separated on 12% SDS polyacrylamide gel and transferred to PVDF membranes. The membranes were blocked with 5% (g/v) defatted milk in TBST buffer (10 mM Tris-HCl (pH7.4), 150 mM NaCl, 0.05% Tween 20) and probed with using appropriate antibodies including 1:6,000 dilution α-GFP conjugated with HRP (MBL, USA), 1:2,000 α-HA (MBL, Ottawa, IL, USA) and 1:2,000 α-Myc (MBL, Ottawa, IL, USA) overnight at 4 °C. Then the samples were washed with TBST buffer for three times and visualized by using the ECL (Amersham™, USA). Actin protein was used as the internal control.

Subcellular localization analysis

The protoplasts from the 4-week-old stable transgenic plants leaves with GFP-tagged βCA1 were isolated following the previous used methods (Yoo, Cho & Sheen, 2007). The subcellular localization of the GFP fusion proteins from both the protoplast and seedlings was determined by confocal laser scanning microscopy Leica SP8. The excitation wavelength was 488 nm for GFP and the emission window was set at 500–520 nm.

Co-immunoprecipitation (Co-IP) and MS analysis

The samples of 0.5 g 7-d-old seedlings tissue were freeze grounded to powder and homogenized in 2 ml IP buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM MgCl₂, 20% glycerol, 0.2% CA-360, 1× protease inhibitor cocktail in DMSO). Lysates were clarified by centrifugation at 14,000 rpm for 15 min at 4 °C and were incubated with Magnetic GFP beads (MBL, Ottawa, IL, USA) for overnight at 4°C in a top to end rotator. After incubation, the beads were washed five times with ice-cold washing buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM MgCl₂, 20% glycerol and 0.2% CA-360) and then eluted by boiling in reducing SDS sample buffer. Samples were separated by SDS-PAGE. For mass spectrometry analysis, the gel was stained with Coomassie blue after SDS-PAGE separation, and the visible stained protein bands were cut out for in-gel trypsin digestion, followed by tandem liquid chromatograph-mass spectrometry (LC-MS/MS) at Shanghai Applied Protein Technology Co. Ltd.

Proteins enrichment and protein-protein interaction analysis

Annotation and GO enrichment analysis for proteins identified by MS was performed using the Metascape tools, a free online platform for data analysis (Zhou et al., 2019). The protein interaction network analysis was applied in STRING Version 11.0 (minimum required interaction score with high confidence of 0.700) (Szklarczyk et al., 2019).

Gene IDs

Sequence data from this article can be found in The Arabidopsis In-formation Resource (http://www.arabidopsis.org/) under the following gene IDs: βCA1 (At3g01500), βCA1.1 (At3g01500.1), βCA1.2 (At3g01500.2), βCA1.3 (At3g01500.3), βCA1.4 (At3g01500.4).

AS analysis in Arabidopsis βCA1

A genome-wide analysis of the βCA1 gene family was performed based on the complete genome sequences. Using the Phytozome database and the βCA1 from Arabidopsis as the guide sequence for a BLASTp search, we first retrieved available βCA1 sequences from representative sequenced genomes, including three monocots (Oryza sativa, Brachypodium distachyon, Sorghum bicolor), two dicots (Glycine max, Populus trichocarpa), and Physcomitrella patens. Further analysis of genome annotation obtained from the Phytozome database revealed that all βCA1 genes from selected plant species possessed several transcripts resulting from AS (Fig. S1). At least four spliced variants were found for the βCA1 gene. These spliced variants seemed to result from pre-mRNA AS, mostly via the use of either alternative 5′ or 3′ splice site. By analyzing the latest Arabidopsis thaliana genome database, four spliced variants of βCA1 were found: βCA1.1, βCA1.2, βCA1.3, and βCA1.4 (Fig. 1A). A previous study showed that the full-length βCA1 (same as βCA1.2) contains the chloroplast transit peptide at the N-terminus (Hu et al., 2015), whereas βCA1.1 and βCA1.3 coded for truncated proteins, lacking the chloroplast transit peptide at the N-terminus and 11 amino acids at the C-terminus, respectively (Fig. 1B). Moreover, the shortest βCA1.4 was truncated at both ends of the protein.

To investigate these spliced variants of βCA1, the expression pattern of each variant was analyzed. It was not possible to obtain RT-PCR primers suitable for βCA1.1, so the corresponding transcripts of βCA1.1 and βCA1.2 were analyzed together. The same strategy was applied to βCA1.3 and βCA1.4. For instance, primers F1 and R1 were used for the total transcripts of βCA1.1 and βCA1.2, whereas F2 and R1 were used for the analysis of βCA1.2 (Fig. 1A). We investigated the transcript expression in four tissues from flowering plants: leaves, inflorescences, flowers, and fresh fruits (Fig. 1C). βCA1.1 showed the highest global expression levels, with more transcripts from leaves and inflorescences. On the contrary, the expression of βCA1.2 was higher in flowers and fruits than in leaves and inflorescences. The expression levels of βCA1.3 were weaker than the expression levels of βCA1.4, and βCA1.3 transcripts were rarely detected in flowers.

To further study the accumulation of each βCA1 spliced variant in vivo, two plasmids were generated. The first (βCA1:βCA1-HA) had an HA tag (YPYDVPDYA) directly attached to the last exon before the stop codon. The second (βCA1:βCA1-Myc) had a Myc tag (QILFRDEFLL) linked to the end of the second-to-last exon, which was designed for expression analysis of four individual spliced variants of βCA1 in vivo (Fig. 1D). The spliced variant of βCA1 was detected in the protein extracts from stable transgenic seedlings. The total translation accumulation of βCA1.1 and βCA1.2 were comparable, and protein levels of βCA1.4 were much higher than βCA1.3. However, it is not possible to compare protein levels of all four spliced variants of βCA1 due to the different genetic backgrounds. Taken together, these results suggest that βCA1 is regulated by AS.

Subcellular localization of spliced variants for βCA1

As the chloroplast targeting peptide was absent in βCA1.1 and βCA1.4, we speculated that the different spliced variants might have different subcellular locations. We then investigated the subcellular localization of each βCA1 spliced variant tagged with GFP in stable transgenic plants. In 1-week-old seedlings, βCA1.1 and βCA1.2 showed a similar distribution with a visible GFP fluorescence signal in leaf chloroplasts localized on stomata (Figs. 2A and B). βCA1.3 was also localized in leaf chloroplasts, however, it seemed to accumulate in the chloroplast envelope and was also observed within the protoplast (Fig. 2B). Surprising, βCA1.4 presented a very different pattern. Strong GFP fluorescence signals in the cytoplasm were observed in transgenic plants with βCA1.4-GFP, whereas no visible signal was detected in chloroplasts either from stomata or leaf protoplasts.

In the root tips, a strong fluorescent signal was detected only for βCA1.4, whereas βCA1.1, βCA1.2, and βCA1.3 were almost undetectable (Fig. 2C). Instead, the latter three variants seemed to accumulate only in small, punctuated structures in the cytoplasm of the root mature zone (Fig. 2D). Subsequent analysis showed that the signals from these small punctuated structures were dynamic (Fig. S2). Immunoblotting analysis confirmed the accumulation of the four spliced variants in the corresponding tissues and organs (Fig. S3).

Stress influences the transcript abundance of βCA1 spliced variants

The CA1 family is induced under multiple abiotic stresses and phytohormones; therefore, we investigated the effect of some abiotic stresses (salt, heat) and ABA on the expression pattern of βCA1 spliced variants. The results showed that none of the stresses affected the two βCA1 spliced variant pools (βCA1.1+βCA1.2, βCA1.3+βCA1.4), whereas salt and heat stresses significantly repressed the transcriptional expression of βCA1.2 and βCA1.3 (Fig. 3A). This indicated that salt and heat stresses significantly enhanced the induction of βCA1.1 and βCA1.4.

Previous studies had reported that a mature rice CA protein could confer the heat stress tolerance in E. coli recombinants and that CA1 proteins from poplar were induced under heat stress (Tianpei et al., 2015; Shi et al., 2017). Thus, we characterized the response of a T-DNA insertion mutant βca1 to heat stress (Fig. S4). For the heat stress germination assay, the vernalized Col-0 and βca1 mutant seeds were pretreated at 50 °C for 3 or 6 h before moving them to normal growth conditions. The results showed that the germination rate of Col-0 was almost unaffected after heat treatment for 3 h, while it decreased to 85% in the βca1 mutant (Fig. 3B). Under 6 h heat stress treatment, a more significant reduction in the germination rate of βca1 was observed, whereas 80% of Col-0 seeds still germinated. This suggests that the βca1 mutant is heat-sensitive during seed germination.

Protein-protein inertaction (PPI) network analysis for Arabidopsis βCA1

Knowledge of all direct and indirect interactions between proteins will provide new insights into the complex molecular mechanisms inside a cell. To better understand the roles of βCA1 in Arabidopsis, we used co-IP to identify proteins that interact with the most redundant spliced variant βCA1.4. The results showed with high confidence that 109 proteins interact with βCA1.4 directly or indirectly (Table S2).

Among the interacting proteins of βCA1.4, two other βCA family proteins, βCA2 and βCA4, were identified. This interaction was confirmed in vivo by bimolecular fluorescence complementation (BiFC) (Huang et al., 2017). To gain further insight into their protein functions, the 109 proteins interacting with βCA1.4 were analyzed using Gene Ontology (GO) classifications using Metascape. The analysis revealed that 19 clusters were significantly enriched (P value cutoff of 0.01). These proteins played various roles, most of which were related to the functions of βCA1, including photosynthesis and stress response (Fig. 4A). Surprisingly, 29 of these proteins were predicted in response to cadmium ions, which indicated that Arabidopsis βCA1 may also function as a cadmium enzyme. In addition, GO analysis showed that 17 proteins, including βCA1, were involved in the response to temperature, which suggests a potential molecular basis for βCA1 in the heat stress response (Fig. 4B).

Additional PPI networks for βCA1.4 associated proteins were constructed using Search Tool for the Retrieval of Interacting Genes/Proteins (STRING 11.0). The full set of 109 proteins produced an interactome map composed of 108 nodes (proteins) and 516 edges, with an average node degree of 9.56 (avg. local clustering coefficient = 0.7) (Fig. 5). The various PPI sub-networks are associated with photosystem (green), cellular response to stress (red), and ribosome pathway (cyan), which was consistent with the results from Metascape analysis. Further subcellular localization analysis by STRING showed that most of these proteins are localized in chloroplasts and plastids, where βCA1 proteins are also common. Moreover, several proteins, including βCA1, were predicted to be present in stromules (Hanson & Hines, 2018), implying that they may function in morphological maintenance and stromule regulation of non-mesophyll plastids (Fig. S5).