ABC1K10a, an atypical kinase, functions in plant salt stress tolerance

ABC1K (Activity of BC1 complex Kinase) is an evolutionarily primitive atypical kinase family widely distributed among prokaryotes and eukaryotes. The ABC1K protein kinases in Arabidopsis are predicted to localize either to the mitochondria or chloroplasts, in which plastid-located ABC1K proteins are involved in the response against photo-oxidative stress and cadmium-induced oxidative stress. Here, we report that the mitochondria-localized ABC1K10a functions in plant salt stress tolerance by regulating reactive oxygen species (ROS). Our results show that the ABC1K10a expression is induced by salt stress, and the mutations in this gene result in overaccumulation of ROS and hypersensitivity to salt stress. Exogenous application of the ROS-scavenger GSH significantly represses ROS accumulation and rescues the salt hypersensitive phenotype of abc1k10a. ROS overaccumulation in abc1k10a mutants under salt stress is likely due to the defect in mitochondria electron transport chain. Furthermore, defects of several other mitochondria-localized ABC1K genes also result in salt hypersensitivity. Taken together, our results reveal that the mitochondria-located ABC1K10a regulates mitochondrial ROS production and is a positive regulator of salt tolerance in Arabidopsis.

which fine-tunes the calcium signal and modulates the SOS pathway in response to salt stress [8].
Salt stress induces the overaccumulation of reactive oxygen species (ROS) such as hydrogen peroxide (H 2 O 2 ), singlet oxygen ( 1 O 2 ), hydroxyl radical (•OH), and superoxide anion radical (O 2˙ˉ) [3,9]. Under salt stress, ROS can be generated from multiple cellular locations including chloroplasts, mitochondria, peroxisomes and plasma membrane NADPH oxidases [10]. The mutation of rice gene WSL12 that encodes chloroplast nucleoside diphosphate kinase 2 (NDPK2), results in enhanced ROS accumulation and sensitivity to salt stress, suggesting that chloroplast is a source of damaging ROS elicited by salt stress [11]. High salinity also impairs mitochondria electron transfer rates between different respiratory chain complexes and causes ubiquinone (UQ) over-reduction, and the excess electrons are transferred to molecular oxygen or nitrate, giving rise to O 2˙ˉo r nitric oxide (NO) [12]. The mitochondrial protein AtMT2b (a type 2 metallothionein) is involved in plant salt stress response by interacting with the voltage-dependent anion channel 3 (AtVDAC3) to regulate ROS homeostasis and mitochondrial membrane potential (MMP) [13].
Aerobic organisms have evolved both non-enzymatic and enzymatic antioxidant defense mechanisms to protect plants against oxidative stress [14]. Non-enzymatic antioxidants include vitamin C, vitamin E, alkaloids, carotenoids, tripeptide glutathione (GSH) and flavonoids, while antioxidant enzymes such as superoxide dismutases (SODs), catalase (CAT), ascorbate peroxidase (APX) and glutathione peroxidase (GPX) can effectively detoxify cellular O 2˙ˉa nd H 2 O 2 [10,15]. Numerous studies have reported that overexpression of SODs (Cu/Zn-SOD, Mn-SOD, or Fe-SOD) enhances salt tolerance of the transgenic plants [15]. The transgenic rice overexpressing the Escherichia coli CAT encoded gene, katE, shows improved tolerance to salt stress [16]. Furthermore, as one of the most important metabolites, GSH functions in mitigating oxidative stress, and cellular GSH concentrations directly affect plant stress tolerance [15]. Therefore, maintaining cellular ROS homeostasis under stress conditions is important in plant stress response.
The bc1 complex kinases (ABC1Ks) are atypical kinases which lack many of the characteristics of eukaryotic protein kinases. The founder member of ABC1Ks is ABC1/COQ8 (now called ScCOQ8) from Saccharomyces cerevisiae, which functions to maintain the activity of the mitochondrial bc1 complex by regulating UQ synthesis [17]. As homologues of ScCOQ8, UbiB in Escherichia coli, aarF in Providencia stuartii and CABC1 in Homo sapiens also function in UQ biosynthesis [18,19]. The Arabidopsis ABC1K family contains 17 members, eight of which (AtABC1K1-8) belong to the photosynthetic-specific clade, six (AtABC1K11-15) belong to the mitochondrial clade, and three (AtABC1K9, AtABC1K10a and AtABC1K10b) are ancestral clade members [20,21]. AtABC1K8 (AtOSA1) is the first identified member of this family localized in chloroplasts and participates in mitigating cadmium-induced oxidative stress [22]. ABC1K1 and its homolog ABC1K3 phosphorylate VTE1, a major limiting factor in tocopherol synthesis, and participate in plant response to photooxidative stress [23][24][25][26]. The chloroplastic ABC1Ks have been extensively studied, but the role of mitochondrial ABC1Ks remains unclear.
In this study, we demonstrate that the mitochondrial ABC1K10a plays an important role in salt stress tolerance. The ABC1K10a knock-out mutants accumulate excessive H 2 O 2 and O 2˙ˉa nd are hypersensitive to salt stress. Inhibition of ROS synthesis alleviated the salt stress hypersensitivity of abc1k10a mutants. Further analysis showed that the excessive ROS accumulation in abc1k10a under salt stress is likely to be resulted from irregularities of the respiratory complex in mitochondria. Other mitochondrial ABC1K members are also involved in the response to salt stress by regulating ROS accumulation, indicating the important roles of ABC1K family in salt stress response. Taken together, the outcome of our results sheds new light on the positive role of the mitochondria-located ABC1K10a in the regulation of salt tolerance in Arabidopsis.

Results
Arabidopsis mitochondria-abc1k mutants accumulate higher ROS under salt and osmotic stresses The members of Arabidopsis ABC1K family can be divided into three clades according to their evolutionary origins and subcellular localizations: chloroplast clade, mitochondrial clade and ancestral clade (Fig. 1a). Previous studies demonstrated that ABC1K proteins located in chloroplast are involved in oxidative stress response [22,24]. To determine whether the ancestral clade and mitochondrial clade of ABC1Ks are also involved in ROS metabolism, we detected the ROS levels in several abc1k mutants under NaCl or mannitol treatment. ROS levels in most abc1k mutants increased significantly under both treatments, suggesting a role for these proteins in ROS detoxification under abiotic stress conditions. The results revealed the highest ROS accumulation in two independent abc1k10a mutants (k10a-1 and k10a-2) and this gene was chosen for further characterization (Fig. 1b, c).

The expression pattern of ABC1K10a (K10a) gene
The expression patterns of ABC1K10a (named K10a herewith) were first analyzed by using proK10a::GUS transgenic plants. GUS staining showed that K10a was expressed in various tissues, with relatively higher expression in siliques and cotyledons (Fig. 2a). These results were further confirmed by qRT-PCR analysis, which showed the strongest expression levels in roots, flowers and siliques (Fig. 2b). Bioinformatics analysis indicated a putative mitochondrial signal peptide at the N-terminus (Fig. 2c), suggesting a mitochondrial localization of K10a. To verify this, the K10a-YFP fusion proteins were expressed in Nicotiana benthamiana leaves, and the fluorescence of K10a-YFP was exclusively detected in the mitochondria, overlapping with the mitochondria marker Mito-tracker red (Fig. 2d). This result indicates that the K10a is indeed a mitochondrial protein.
To obtain support for a putative role for the mitochondrial K10a protein in abiotic stress we determined the expression patterns under salt stress treatment. The results of qRT-PCR and GUS staining showed that the transcript levels of K10a were upregulated by NaCl treatment (Fig. 2e, f). Furthermore, analysis of transgenic lines expressing a K10a-GFP fusion protein under the control of the K10a native promoter (proK10a::K10a-GFP) confirmed that NaCl Fig. 1 The abc1k mutants accumulate more ROS than the wild type under salt and osmotic stresses. (a) Phylogenetic tree of ABC1K proteins from Arabidopsis. The tree was constructed by comparing the amino acid sequences of all 17 ABC1K proteins encoded in the Arabidopsis thaliana genome. Three different clades were highlighted with different colors (green indicates the plastids clade; purple indicates the ancestral clade, and light pink indicates the mitochondria clade) according to previous reports [20,21]. (b) NBT staining of the roots of wild type, abc1k11, abc1k10a, abc1k13, abc1k14 and abc1k15 mutant roots treated with 200 mM NaCl or 300 mM mannitol. Bars = 400 μm. k, abc1k. (c) Quantification of NBT staining intensity. Asterisks indicate statistically significant difference with wild type plants in each treatment. Error bars indicate ± SD (n = 15) (**P < 0.01, *P < 0.05, t-test) . Asterisks indicate statistically significant differences between treated and untreated samples at the different treatment times (**P < 0.01, t-test). (f) For GUS staining, 5 day-old transgenic seedlings expressing ABC1K10apro::GUS were treated with 200 mM NaCl for 0, 3 or 6 h. Bars = 200 μm. (g) To extract the total proteins, 5 day-old transgenic seedlings expressing ABC1K10a pro::ABC1K10a-GFP were treated with 200 mM NaCl for 0, 3 or 6 h. Protein extracts were analyzed by western blotting with α-GFP antibody (g). Actin was used as a control treatment induces a marked increase in K10a protein levels ( Fig. 2g; Additional file 2: S2). These results indicate that K10a is a salt inducible gene and may play a role in salt stress response.
The abc1k10a mutants are hypersensitive to salt stress To study the putative role of K10a during salt stress response, we characterized the behavior of two independent T-DNA mutants (10a-1 and 10a-2) showing negligible K10a expression (Fig. 3a, b). Under normal growth conditions, the 10a mutants showed similar growth with the wild type. Under NaCl treatment, however, the 10a mutants displayed more severely inhibited growth than the wild type ( Fig. 3c, d, e).
Effects on cotyledon greening were clearly observed on 10a mutants at lower NaCl concentrations than WT (Fig. 3d). After growing for 10 days on MS medium containing 100 mM NaCl, all the wild type seedlings had green cotyledons, while only 11% of the 10a-1 and 10% of the 10a-2 mutants had green cotyledons (Fig. 3e). Expression analysis of the saltresponsive marker genes of Responsive to Desiccation 29A (RD29A) [27], Kinase 1 (KIN1) [27] and Cold Regulated 15b (COR15b) [28] showed significantly higher expression in the 10a mutants than in wild type after salt treatment for 6 h (Fig. 3f). Collectively, these results indicate that K10a is important in plant salt stress response and tolerance. Transcript abundance of ABC1K10a in the wild type and abc1k10a mutants. Total RNAs isolated from 5 day-old wild type and 10a mutant seedlings were used for real-time qRT-PCR. Three biological replicates were conducted (means ± SD, n = 3). 10a, abc1k10a. Asterisks indicate statistical difference between 10a mutants and wild type (***P < 0.001, t-test). (c) Salt sensitivity of wild type and 10a mutants. Wild type, 10a-1 and 10a-2 mutants seeds were germinated and grown for 8 days on MS medium containing different concentrations of NaCl. Bar = 500 μm. (d) Cotyledon greening of seedlings grown on MS medium containing different concentrations of NaCl was scored after 10 days and represent averages of 100 seedlings from at least four independent experiments. (means ± SD; n = 100). Asterisks indicate statistically significant differences (**P < 0.01, t-test) between 10a mutants and wild type. (e) Cotyledon greening rates in wild type and 10a mutants in response to NaCl treatment. Wild type, 10a-1 and 10a-2 seeds were sown on MS medium or MS medium with 100 mM NaCl. Cotyledon greening rates were calculated at the indicated days. Error bars indicate ± SD (n = 100). Asterisks indicate statistically significant differences (***P < 0.001, t-test) between the 10a mutants and wild type. (F) The transcript levels of RD29A, KIN1 and COR15b in wild-type, 10a-1 and 10a-2 mutants at different times after treatment with NaCl. 5-day-old seedlings were treated without or with 200 mM NaCl at the indicated times, and transcript levels were analyzed by qRT-PCR. Error bars indicate ± SD (n = 3). Asterisks indicate statistically significant differences between the 10a mutants and wild type (**P < 0.01, ***P < 0.001, t-test)

Arabidopsis ABC1K10a positively regulates salt stress tolerance
To further confirm that the salt hyper-sensitive phenotype of the 10a mutants is caused by the defect of K10a, two independent complementation transgenic lines were generated by introducing the wild type K10a gene driven by the native K10a promoter into the 10a-2 mutant. The expression of K10a in the 10a-2 mutant fully restored the salt hypersensitive phenotype of the 10a-2 mutant to wild type levels, further supporting that K10a is required for salt tolerance in Arabidopsis (Fig. 4a). Consistently, the expression levels of RD29A, KIN1 and COR15b in the complementation lines were also restored to the levels of wild type in the presence or absence of salt stress (Fig. 4b). In addition, we generated transgenic lines overexpressing K10a-Myc under the control of 35S promoter (K10a-OE), and the K10a-OE lines displayed enhanced salt tolerance when compared with the wild type (Additional file 1: S1). Overall, our results indicate that K10a positively regulates plant salt tolerance.

ROS overaccumulation results in the salt hypersensitivity of 10a mutants
Our results showed that NaCl treatment resulted in higher ROS accumulation in the 10a mutants, therefore we wondered whether the increased ROS is the main reason of the observed salt hypersensitive phenotypes (Fig. 1b, c). Methyl viologen (MV) induces the production of ROS in plant chloroplasts and mitochondria [29]. When plants were grown in MS media supplemented with 0.1 μM ROS-generating chemical methyl viologen MV, the 10a mutants displayed hypersensitivity to MV, similar phenotypes to those observed under salt stress ( Fig. 3c; Fig. 5a). In the presence of MV, the 10a mutants had only 11% green cotyledons, compared with 70% green cotyledons in the wild type (Fig. 5b). These results suggest that the salt hypersensitive phenotype of 10a mutants may be resulted from ROS overaccumulation.
We used the ROS-scavenger GSH to determine the involvement of ROS in the salt stress response of 10a mutants. After germination in MS medium containing 100 mM NaCl for six days, 40% of wild type seedlings grew green cotyledons, while only 10% of the 10a-1 and 6% of the 10a-2 mutants grew green cotyledons (Fig. 5c, d). However, the addition of 300 μM GSH to the medium neutralized the effect of salt with WT, 10a-1 and 10a-2 seedling cotyledon greening reaching 94, 90 and 88%, respectively, indicating that the growth inhibition by NaCl treatment was resulted from ROS accumulation and the salt hypersensitivity of the 10a mutants was due to overaccumulation of ROS (Fig. 5c, d). ROS were detected by NBT and 2′7'dichlorofluorescin diacetate (H 2 DCF-DA) staining, which indicate O 2˙ˉa nd H 2 O 2 , respectively [30]. Under normal growth conditions, no differences were observed between wild type and 10a mutants. However, NaCl treatment produced a significantly stronger staining in the 10a mutants compared to wild type, while GSH reduced the excess ROS accumulation in all genotypes (Fig. 5e, f). These results further support that ROS overaccumulation under salt stress accounts for the salt hypersensitivity of 10a mutants.
The 10a mutations cause dysfunction of mitochondria under salt stress Mitochondria are important sources of ROS production, especially under various abiotic stresses which restrain Fig. 4 Complementation of 10a mutants restores salt sensitivity to wild type levels. (a) Arabidopsis seedlings grown in the absence or presence of 100 mM NaCl for 12 days. Complementation lines 10a::10a-1 and 10a::10a-2 were obtained by introducing a ProABC1K10a::ABC1K10a-GFP construct into the 10a mutant background (Bar = 500 μm). Asterisks indicate statistically significant difference compared to the wild type in the same treatment (**P < 0.01, t-test). (b) Relative expression levels of RD29A, KIN1 and COR15b in wild-type, 10a-1, 10a-2 and 10a::10a-2-GFP seedlings analyzed by qRT-PCR. Asterisks indicate statistically significant differences with WT (**P < 0.01, t-test) ETC activity, disturb electron transfer and promote ROS production [10]. Under salt stress, the classic cytochrome c oxidase (COX) pathway is inhibited, and the alternative oxidase (AOX) pathway plays an important role transferring electrons from ubiquinone to oxygen and limiting ROS production [31,32]. To determine whether the high level of ROS induced by salt stress in 10a mutants resulted from mitochondrial dysfunction, we analyzed the expression of AOX genes and two antioxidant genes. Under NaCl treatment, the expression of SOD2 and CAT2 was significantly increased in the 10a mutants compared to the wild type, suggesting that Rates of cotyledon greening were calculated in (b) (means ± SD, n = 100). Asterisks indicate statistically significant differences with the corresponding wild type (**P < 0.01, t-test). (c) Alleviation of salt stress symptoms by GSH treatment in 10a mutants. Seedlings grown on the indicated media for ten days were photographed and the rates of cotyledon greening were calculated (d). (Bar = 500 μm. means ± SD, n = 100). Asterisks indicate statistically significant differences with the corresponding wild type (**P < 0.01, t-test). (e and f) NBT staining (e) and H 2 DCF-DA staining (f) of the cotyledons of wild type and 10a mutants after NaCl treatment or supplemented with 100 μM GSH. Bars = 2 mm (e), bars = 300 μm (f) the10a mutants suffered more severe oxidative stress than the WT (Fig. 6a). However, lower expression of AOX gene in the 10a mutants than in the wild type implied that AOX pathway was severely restrained, resulting in ROS over-production (Fig. 6a).
We performed Blue Native (BN)-PAGE analyses to determine the assembly of the complexes in the mitochondrial respiratory chain. Similar levels of the mitochondrial complexes were detected between wild type and 10a mutants under normal growth conditions. However, the complex III levels were more severely inhibited in the 10a mutants under NaCl treatment when compared with the WT (Fig. 6b; Additional file 3: S3). Since the mitochondrial complex I and III are the key sites resulting in ROS production [12], the abnormality of the complex III assembly under salt stress may account for the overaccumulation of ROS in the 10a mutants. These results also suggest that K10a may play a role in maintaining the integrity and stability of the complex III under salt stress in plants.

Defects in other ABC1K genes also result in salt hypersensitivity
We also examined the possible role of other mitochondrial localized ABC1Ks in salt stress tolerance. There are no differences between k11, k13, k14 and k15 mutants and wild type when grown on MS medium, however, the growth phenotypes of k11, k13 and k15 mutants were severely altered under 100 mM NaCl treatment (Fig. 7a, b; Additional file 4: S4). Since these mutants accumulated higher levels of ROS when exposed to salt stress (Fig. 1b,  c), we determined whether the addition of the ROS scavenger GSH could revert the effects of salt stress. Consistent with our previous results using 10a mutants, the salt hypersensitive phenotypes of these mutants were restored by the addition of GSH (Fig. 7a, b). Under NaCl treatment, the percentages of green cotyledons of k11, k10a, k13 and k15 were about 47, 16, 30 and 27% of that of the wild type, while the ratios were increased to 95, 95, 95 and 78%, respectively, when GSH was exogenously applied (Fig. 7b). These results demonstrate that the Arabidopsis mitochondria-localized ABC1K proteins play prominent roles in plant salt tolerance through maintaining ROS homeostasis.

Discussion
As an atypical protein kinase, the yeast ancestral member ABC1K was reported to be required for coenzyme Q biosynthesis and thus mitochondrial oxidation metabolism [33]. In higher plants, the abc1k mutants were found to be highly sensitive to oxidative stress caused by high intensity light, Cd 2+ or H 2 O 2 [23][24][25]. In this study, we found that the T-DNA insertion mutants of the mitochondria ABC1K members are hypersensitive to salinity stress ( Fig. 1b; Fig. 7). Detailed phenotypic characterization of two null mutants of 10a, a member of the ancestral clade, revealed markedly stunted de-etiolation process and growth inhibition under salinity stress (Fig. 3c, d. e). In consistent with the dramatic response phenotype, the expression of the salt-inducible . Asterisks indicate statistically significant differences with the corresponding wild type (**P < 0.01 and ***P < 0.001, t-test). (b) The abundance of the mitochondrial complexes in wild type and the 10a mutants. Crude mitochondrial membranes were extracted from 5-day-old seedlings grown on MS medium or MS medium supplemented 100 mM NaCl. The mitochondrial complexes were separated by Blue Native-PAGE and stained with coomassie brilliant blue marker genes was also significantly higher in the 10a mutants than in wild type (Fig. 3f), suggesting a sensitiveresponse of the mutants to salt stress. In contrast to mature plants, seedlings during de-etiolation period are more sensitive to various abiotic stresses like drought, salinity, light and temperature stress [34,35]. The 10a mutants show a strong defect in cotyledon greening, which may be associated with its relatively high expression in siliques and cotyledons (Fig. 2a, b; Fig. 3c). This also indicates that ABC1K10a may play a prominent role in early seedling development under salinity stress. Our lossof-function and gain-of-function analyses revealed that ABC1K10a is a positive regulator of salinity tolerance in Arabidopsis (Fig. 3c; Additional file 1: S1).
Abiotic stresses often result in the accumulation of cellular ROS and thus cause oxidative stress as a secondary stress. Although ROS have been recognized as signaling molecules, excess ROS induced by stress conditions damage cellular components such as DNA, RNA, and proteins [36]. Mitochondria are a major source of ROS and play an essential role in regulating important cellular metabolic processes [37]. Our results suggest that the mitochondrial ABC1K10a play a prominent role in mitochondria function under salt stress condition (Fig. 6a, b), and defective of this gene results in overaccumulation of ROS (Fig. 5e,  f), which is likely originated from mitochondria. The salt hypersensitive phenotype of 10a mutants can be rescued by GSH (Fig. 5c, d), a non-enzymatic antioxidant that counteracts ROS accumulation [15], indicating that the salt hypersensitivity of 10a mutants is caused by overaccumulation of mitochondrial originated ROS that result in oxidative stress damage. ROS has also been reported to regulate various transcriptional regulation, for instance, when Arabidopsis thaliana seedlings were exposed to H 2 O 2 , 459 genes expression are enhanced more than 2-fold [38]. Previous research indicated that the deficiency of ROS scavengers catalase caused redundant ROS accumulation and diverse genes induced expression [39]. The similar proofs were observed in abc1k10a mutants, more ROS accumulated, while enhanced genes expression of RD29A, KIN1 and COR15b, which may be triggered by endogenous ROS accumulation (Fig. 3f).
Plant cellular respiration incorporates COX pathway and AP pathway. When COX pathway is inhibited under stress conditions, AOX functions in preventing UQ over-reduction, alleviating ROS production, and maintaining cellular respiration status [39]. In the 10a mutants, the expression of AOX gene was reduced under salt stress, suggesting that the AP pathway was also restrained, which resulted in enhanced ROS production in mitochondria and oxidative damage in the mutants. Mitochondria are important for the maintenance of cellular redox equilibrium and the mitochondrial complex III has been proved to be a major site for ROS production [9]. The 10a mutants showed diminished mitochondrial complex III, suggesting that K10a may modulate the accumulation of mitochondrial complex III to finetune ROS level in response to salinity stress (Fig. 6b).
ABC1K proteins share conserved motifs including the catalytic loop, NTP-binding pockets, and Mg 2+ chelation Fig. 7 Mitochondrial ABC1K protein participate in salt stress response. The cotyledon greening phenotype of the abc1k mutants in response to NaCl treatment and the recovery of this phenotype by GSH. The seeds of the abc1k mutants were planted on MS medium without or with 200 mM NaCl for 10 days. The recovery of the cotyledon greening phenotype was test with 100 μM GSH (a). Bars = 5 mm. The rates of cotyledon greening (b). Error bars indicate means ± SD (n = 100). Asterisks indicate significant difference with the corresponding wild type (**P < 0.01, ***P < 0.001, t-test) site, which suggests similar functions of these proteins in plants [21,40]. Our study indicates that, in addition to the 10a mutants, several other abc1k mutants also show salt hypersensitive phenotype and overaccumulation of ROS under salt stress ( Fig. 1b; Fig. 7), revealing a conservative role of ABC1K family proteins in plant response to salinity stress. Meanwhile, the plastid clade members ABC1K1, ABC1K3 and ABC1K8 have also been shown to play important roles in oxidative stress response [25,26]. Therefore, the ABC1K family proteins might be evolutionarily conserved proteins functioning in modulating cellular ROS production in response to abiotic stresses in plants.

Conclusions
Here, we found the mitochondria-localized ABC1K10a functions in plant salt stress tolerance by regulating ROS accumulation. Our data suggest that ABC1K10a expression is induced by salt treatment, and the mutation of this gene results in significant sensitive to salt stress and severe oxidative stress. Further studies reveal that the defects of mitochondria electron transport chain are responsible for ROS overaccumulation in the 10a mutants. Other ABC1K proteins in mitochondria also regulate ROS production and salt stress tolerance.

Plasmid construction
The coding sequence of K10a was inserted into pCAM-BIA1302 to create a YFP tagged 10a protein at the C terminal. The K10a coding sequence was also cloned into the pDONR™221 vector by using the Gateway BP Clo-nase™ II Enzyme Mix (Invitrogen). And through LR reaction (Invitrogen) the CDS sequence was cloned into the binary vector pYL436 to generate a 10a protein with nine Myc-tags at its C-terminal [41].
The genomic sequence including 1500 bp upstream of K10a transcription start codon and the entire coding sequence of K10a without stop codon was amplified and cloned into pCAMBIA1300 to generate proABC1K10a:: ABC1K10a-GFP vector. The promoter used in the vector proABC1K10a::ABC1K10a-GFP was the same as in proABC1K10a::GUS. Primers used for vector constructs have been listed in the Additional file 5: Table S1.
The transgenic plants were generated by using the Agrobacterium tumefaciens-mediated floral dip method [42,43]. The constructs of 35S::ABC1K10a-YFP/Myc and proABC1K10a::ABC1K10a-YFP were introduced into wild type and the 10a mutant, producing transgenic lines for K10a overexpression and mutant complementation, respectively. Transgenic seeds were screened on 1/ 2 MS medium with appropriate antibiotics, and the T3 plants were used for relevant experiments.

Abiotic stress treatments
For the cotyledon greening assay, seeds harvested from the same batch were sterilized for 10 min with 10% (v/v) sodium hypochlorite, rinsed five times with sterile-deionized water, and then grown on 1/2 MS medium with different concentrations of NaCl or GSH. The seeds were kept at 4°C for 2 days, and then incubated at 23°C under long-day photoperiod (16-h-light/8-h-dark) illumination for 8-10 days before photographed and statistics analyzed described [44].
For salt stress treatment analysis, 5-day-old seedlings grown on normal medium were immersed in 1/2 MS liquid medium with or without 200 mM NaCl for the indicated times [45,46]. The plants were collected, frozen in liquid nitrogen and preserved at − 80°C for further use.

GUS staining assay
Histochemical analysis of promoter-GUS was performed as described [30]. 5-day-old transgenic seedlings of proABC1K10a::GUS grown on 1/2 MS medium were soaked in 1/2 MS liquid medium with 200 mM NaCl for 3 or 6 h and then used for GUS staining. Different organs of mature plants were collected for GUS staining. The samples were submerged in cold 90% acetone for 20 min, washed with rinse solution (50 mM phosphate buffer, pH 7.2, 0.5 mM K-ferrocyanide, 0.5 mM K-ferricyanide), and then stained in the pre-prepared staining solution (rinse solution containing 2 mM X-Gluc) in the dark for a consistent incubation time. After removing the staining solution, samples were washed with 15, 35, 50, 70% ethanol for 15 min each, and then 100% ethanol for 10 min to remove the chlorophyll. The seedlings and organs were photographed with an Olympus MVX10 stereomicroscope.

Confocal microscopy
To determine the subcellular localization of K10a, the vector pro35S::ABC1K10a-YFP was transformed into Agrobacterium strain GV3101 [43,44]. After 12 to 16 h culturing, the Agrobacteria cells were collected, resuspended to the transformation buffer with OD600 ≈ 1.0, and injected into Nicotiana benthamiana young leaves. After 48 h co-culturing, the infiltrated leaves were stained with Mito-Tracker orange as a mitochondrial fluorescence marker. Fluorescent signals were observed with a Carl Zeiss LSM710 META laser scanning microscope.

Quantitative RT-PCR
For qRT-PCR, 5-d-old wild type and 10a mutant seedlings were treated with 1/2 MS liquid medium containing 200 mM NaCl for 3 or 6 h. Total RNAs were extracted using TRIzol reagent (Invitrogen), treated with RNase-free DNase set (Qiagen), and reverse transcribed with M-MLV reverse transcriptase (Promega). For realtime PCR assays, reactions were performed with ChamQ Universal SYBR qPCR Master Mix (Vazyme) using three independent biological replicates. The relative real-time PCR amplification was described, with ACTIN 2/8 as the internal control [47]. Primers used for qRT-PCR have been listed in the Additional file 5: Table S1.

Protein gel blotting
For detecting the protein levels of K10a-GFP, 5-day-old transgenic seedlings were treated with 1/2 MS liquid medium containing 200 mM NaCl at the indicated times, and then ground into fine powder in liquid nitrogen. The proteins were extracted in extraction buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 1 mM PMSF, 5 mM DTT, 0.5% SDS) and quantified with BCA Protein Assay Kit (TaKaRa) as described [48]. Equal amount of proteins for each sample was separated with 12% SDS-PAGE gel and transferred onto a nitrocellulose membrane. The membrane was blocked with 5% non-fat milk, incubated with polyclonal anti-GFP (1/ 2000) or anti-Actin (1/5000) antibodies. The signals were visualized with an X-ray film (Kodak).

Measurement of ROS in plants
For the measurement of ROS, 5-day-old wild type and abc1k mutants were treated with 1/2 MS medium containing 200 mM NaCl at the indicated times and stained with the dyes. The ROS levels were analyzed as previously described [49].
During NBT staining, the seedlings were incubated in reaction buffers for 15 mins which were composed by 20 mM K-phosphate buffer (pH 6.2), 0.1 M NaCl and 1 mM NBT. The reaction was performed in the dark in order to inhibit the decomposition of NBT. The seedlings were washed with water for three times to remove excess dyes, then incubated in acidified buffer (10 mL of methanol, 2 mL of HCl, 38 mL of water) and basic buffer (7% NaOH in 60% ethanol) for 15 min, respectively. The mazarine can be observed, and the pigments were removed by with gradient concentration of ethanol treatment. The cotyledons or roots were photographed using Olympus MVX10 stereomicroscope.
For DCFH-DA staining, the seedlings were incubated in reaction buffer (20 mM K-phosphate, pH 6.0, 50 mM DCFH-DA) in darkness for 10 min. The cotyledons were photographed using Carl Zeiss LSM710 META laser scanning microscope at 488 nm.

Blue native-PAGE and complex activity assay
To detect the accumulation of mitochondrial complexes, 5-day-old wild type and the 10a-1, 10a-2 mutants grown on MS medium or MS medium containing 100 mM NaCl were used for the preparation of mitochondria membrane proteins as previously described [30,50,51]. For the analysis of mitochondrial protein abundance, 50 μg mitochondria proteins from each sample were separated by a 4.5 to 16% gradient Blue-Native PAGE (Invitrogen, BN1002BOX), stained with coomassie brilliant blue (CBB), faded with 10% methanol and 10% acetic acid (v/v) and then imaged.

Bioinformatics and phylogenetic analysis
Arabidopsis ABC1Ks protein sequences obtained from the UniProt (https://www.uniprot.org/) were used for homologous alignment. Sequence alignment and the drawing of evolutionary trees were carried out by using the software MEGA4.

Quantification and statistical analysis
Quantitative data is expressed as mean ± SD and mean ± SE. Statistical significance among various treatments and materials was performed by one-way ANOVA and twoway ANOVA. Statistical significance was set at * P < 0.05, ** P < 0.01, *** P < 0.001.

Availability of data and materials
All data generated or analysed during this study are included in this published article and its supplementary information files. The plant materials during the current study available from the corresponding author on reasonable request.
Additional file 1: S1 The ABC1K10a overexpression lines are more tolerant to salt stress. (A) Seedlings from different genotypes, wild type, 10a-1 mutant, 10a-2 mutant and two independent overexpressing lines