Genome-wide analysis of HAK/KUP/KT potassium transporter genes in banana (Musa acuminata L.) and their tissue-specific expression profiles under potassium stress

Potassium is one of the most essential inorganic cations for plant growth and development. The high affinity K+ (HAK)/K+ uptake (KUP)/K+ transporter (KT) family plays essential roles in the regulation of cellular K+ levels and the maintenance of osmotic balance. However, the roles of these genes in the responses of bananas to low-potassium stress are unclear. In this study, 24 HAK/KUP/KT (MaHAK) genes were identified from banana. These genes were further classified into four groups based on phylogenetic analysis, gene structure and conserved domain analysis. Segmental duplication events played an important role in the expansion of the MaHAK gene family. Transcriptome analysis revealed the expression patterns of MaHAKs in various tissues under different K+ conditions. MaHAK14b was upregulated under both short- and long-term K+-deficient conditions, suggesting that it plays crucial roles in K+ uptake at low K+ concentrations. Furthermore, MaHAK14b mediated K+ uptake when it was heterologously expressed in the yeast mutant R5421 on low K+ medium. Collectively, these findings provide a foundation for further functional analysis of MaHAK genes, which may be used to improve potassium stress resistance in bananas.


Introduction
Potassium ions (K + ) are one of the most essential plant macronutrients and are involved in various plant physiological and biochemical processes, such as osmoregulation, enzyme activation, stomatal regulation, and maintenance of pH (Hasanuzzaman et al. 2018). Plants are often subjected to potassium deficiency, which adversely impacts plant growth by influencing photosynthesis, carbohydrate metabolism and transpiration rates (Pettigrew 2008). Potassium uptake by plant roots and its translocation inside plants rely on a complex transport system consisting of numerous transporters and channels (Epstein and Kim 1971). Potassium transporters in plants can be classified into four families: the high affinity K + (HAK)/K + uptake (KUP)/K + transporter (KT) family, the Trk/Ktr/HKT family, the K + efflux antiporter (KEA) family, and the cation/H + exchanger (CHX) family (Maser et al. 2001;Gierth and Maser 2007).
HAK/KUP/KT is the largest potassium transporter family and has been further divided into four groups: groups I, II, III, and IV (Maser et al. 2001;Li et al. 2021). The HAK/ KUP/KT gene family has been largely investigated in many Communicated by Luca Sebastiani .
Siwen Liu and Bangting Wu have contributed equally to this work. plants, e.g., Arabidopsis thaliana (Maser et al. 2001), Oryza sativa (Gupta et al. 2008), Zea mays (Zhang et al. 2012), Prunus persica (Song et al. 2015), Cicer arietinum (Azeem et al. 2018), Manihot esculenta (Ou et al. 2018), Triticum aestivum (Cheng et al. 2018), and in Rosaceae family (Li et al. 2018a, b). The number of HAK genes varies among species, ranging from 13 to 56, with the lowest number in A. thaliana and the highest number in T. aestivum (Maser et al. 2001;Cheng et al. 2018). Expression analysis of HAK/ KUP/KT genes from a variety of species with different tissues suggests the functional diversity of these genes. In Arabidopsis, most AtKT/KUP/HAK genes are expressed in the root, siliques, leaves and flowers (Ahn et al. 2004). In rice, five genes were shown to be expressed in all tissues (Gupta et al. 2008). In wheat, most genes from groups I and II were strongly expressed in all tissues (Cheng et al. 2018). In addition, the expression profiles of HAK/KUP/KT genes in response to different abiotic stresses have also been investigated in many plant species (Cheng et al. 2018).
Generally, the main functions of HAK/KUP/KT transporters are to regulate plant growth and development and to conduct high-affinity K + uptake to maintain cellular K + homeostasis under low K + conditions (Wang and Wu 2013;Li et al. 2018a, b). In Arabidopsis, AtKUP4 is related to cell expansion of root hairs (Rigas et al. 2001). AtHAK5 is the most important transporter involved in K + uptake in roots, and it is highly upregulated after K + starvation (Pyo et al. 2010). In addition, AtKUP7 is also confirmed to be involved in K + uptake and upregulated under K + -deficient conditions (Han et al. 2016). In rice, OsHAK1 and OsHAK5 are functionally characterized. Overexpressing OsHAK5 or OsHAK1 in rice increased K + uptake under K + -limited conditions, and they are all mainly expressed in roots, where they mediate K + uptake and K + translocation (Yang et al. 2014;Chen et al. 2015Chen et al. , 2017. ZmHAK5 shares a similar function with AtHAK5 and plays a role in K + uptake in maize roots (Qin et al. 2019). Loss of function of ZmHAK5 led to defective growth and K + uptake under low K + conditions. ZmHAK1 has also been confirmed as a high-affinity K + transporter, but its activity is weaker than that of Zam-HAK5 (Qin et al. 2019).
Banana (Musa acuminata L.) is one of the most important fruits and a staple crop in the world with high economic value. Compared to most of the other crops, banana plants have a very high demand for potassium. The yield and fruit quality of banana are directly related to K + availability; therefore, a large amount of K + fertilizer is usually applied in banana-planting fields (Rufyikiri et al. 2003). Characterization of HAK/KUP/KT family members would be helpful for improving K + utilization efficiency in banana. In this study, we identified 24 HAK/KUP/KT transporters from the genome of banana and analyzed their phylogenetic relationships, conserved motifs, gene structures, and expression profiles in response to different K + stresses using our transcriptome data and publicly available data. In addition, we cloned the MaHAK14b gene, which was highly upregulated in K + -deficient conditions, and demonstrated its roles in facilitating K + transport in yeast.

Identification of HAK/HUP/KT genes in banana
The hidden Markov model (HMM) profile of the HAK/HUP/ KT transporter family (Pfam ID: PF02705) was obtained from the Pfam protein family database (http:// pfam. xfam. org/). HMMER 3.3 was used to search the HAK/HUP/KT genes from the banana genome database, which were downloaded from Banana Genome Hub (https:// banana-genomehub. south green. fr/). Proteins with E-value < 0.01 was obtained. Then, all the candidate MaHAK protein sequences were further submitted to CDD databases and SMART to verify the presence of the K + transporter domain.

Phylogenetic analysis of MaHAKs
All the HAK/KUP/KT gene sequences of A. thaliana, O. sativa and Z. mays were retrieved from NCBI (https:// www. ncbi. nlm. nih. gov/). Multiple alignment of all HAKs of different plant species was carried out using ClustalW, and the phylogenetic tree was drawn using the neighbor-joining (NJ) method of MEGA v. 7.0, tested by 1000 Bootstrap replicated (Kumar et al. 2016).

Structural analysis of MaHAKs
The intron and exon distribution patterns of each MaHAK gene were illustrated using Gene Structure Display Server v. 2.0 . The MEME online program (http:// meme. nbcr. net/ meme/ intro. html) was used to identify conserved motifs in MaHAK proteins.
Conserved motifs of the MaHAKs proteins.

Chromosome localization and syntenic relationship analysis
MaHAK genes were mapped to banana chromosomes based on physical location information from the banana genome database using Circos (Krzywinski et al. 2009). The gene duplication events were analyzed by the Multiple Collinearity Scan toolkit (MCScanX), and the syntenic analysis map was constructed using Tbtools (Chen et al. 2020).

Plant growth condition
One-month-old "Cavendish" banana (AAA) cv "Brazilian" plantlets with six to seven leaves (c. 25 cm height) were cultured in a greenhouse with temperature of 28-30 °C, light of 16 h/day, and 60-70% relative humidity. The plants were first cultured in water for 1 day, divided into normal and low-K + groups, and cultured in full nutrient MS medium and potassium-free MS medium for 4 h, respectively. Then, the roots, pseudostems, and leaves of each sample were collected.

Transcriptome analysis
Total RNA was extracted from each sample using an Illumina standard library preparation kit. Each treatment had three biological replicates. RNA-seq reads were obtained by Illumina paired-end sequencing on the Illumina HiSeq2500 platform. Paired-end clean reads were mapped to the banana reference genome (M. acuminate DH Pahang v2) using HISAT2.2.4 with "-rna-strandness RF. Fragments per kilobase of transcript per million mapped reads (FPKM) were used to calculate gene expression levels.
RNA-seq data of long-term potassium stress in banana roots from a publicly available database (PRJNA399649) was further analyzed. The expression levels of the MaHAK genes were normalized and represented as FPKM values, which were used to generate a heatmap with Tbtools (Chen et al. 2020).

Subcellular location of MaHAK14b
The coding sequence of MaHAK14b was amplified and inserted into the pCAMBIA1300:GFP vector. The vector harboring GFP, MaHAK14b-GFP, and plasma membrane markers was transferred into Agrobacterium tumefaciens. Then, the Nicotiana benthamiana leaves were filtered with A. tumefaciens cells. The leaves were analyzed for fluorescence 2 days after transformation using a confocal laser scanning microscope (LSM 710; Carl Zeiss, Oberkochen, Germany).

Yeast complementation
The coding sequence of MaHAK14b was cloned into the pYES2 vector. pYES2-MaHAK14b and empty vector were transformed into the potassium uptake-deficient yeast strain R5421 (trk1∆, trk2∆). The yeast transformants were grown in YPDA medium supplemented with 100 mM KCl at 30 °C until the OD 600 reached 0.8. Cells were harvested by centrifugation at 5000×g for 1 min, washed twice with sterile water and then resuspended in sterile water to an OD 600 of 0.8. The yeast cells were diluted tenfold, 100-fold, and 1000-fold and cultured in AP plates containing 1 mM, 10 mM and 100 mM KCl (Li et al., 2014). The plates were cultured at 30 °C for 3 days before taking pictures.

Identification of the MaHAKs in banana
A total of 26 HAK genes were originally obtained from the banana A genome using the HMM search. Two erroneously predicted genes (Ma02_04550, Ma07_18090) were manually removed after confirming the presence of the K + transporter domain. Finally, 24 genes were identified as banana HAK genes and were renamed according to their corresponding homologs from rice and maize based on phylogenetic relationship analysis. Fourteen family members and the letters a, b, c and d were added when more than one gene clustered together with the specific rice gene, for instance, MaHAK10a, MaHAK10b, MaHAK10c and MaHAK10d (Table 1). The composition of HAK members in the banana genome was very different from that in rice, maize or Arabidopsis. No

Phylogenetic analysis of MaHAKs in banana
The evolutionary relationship of HAKs in banana, rice, maize, and Arabidopsis was analyzed by constructing a phylogenetic tree (Fig. 1). Based on previous studies of HAK proteins in Arabidopsis and rice (Maser et al. 2001;Gupta et al. 2008), the MaHAK proteins were classified into four groups (I-IV). Only MaHAK5 was placed in Group I; MaHAK1, 2, 7, 8, 9, 10, 13 were placed in Group II; MaHAK11, 12, 14, 18 and 23 were placed in Group III; and only MaHAK17 was placed in Group IV. Group II and III

Gene structure and conserved motif examination of MaHAKs in banana
To further investigate the evolution of the HAK family in banana, the exon-intron organizations of all identified MaHAK genes were examined. As shown in Fig. 2, the coding sequences of all MaHAK genes were disrupted by introns, with numbers ranging from 5 to 10. Homologous genes usually have similar structures; for example, the three MaHAK11 genes contained 9 exons and 8 introns.
To assess the conservation and diversification of MaHAK proteins, protein motifs were further analyzed using the online program MEME. Motifs 1, 2, 3, 4, 6, 9 and 10 were conserved in all MaHAK proteins. It was also obvious that the motif composition was similar among homologous genes. Interestingly, the length of introns between the two homologous genes MaHAK7a and MaHAK7b were quite different, while the motif distributions were highly similar.

Expression profile of MaHAKs under different potassium treatments
To study the expression profiles of MaHAK genes in different tissues under different potassium stresses, transcriptome analyses of the roots, pseudostems and leaves in the Baxi cultivar under normal (+ K + ) and potassium-deficient (− K + ) conditions were performed. As shown in Fig. 4a, most MaHAK genes exhibited moderate expression in different tissues under the different treatments, and the expression of MaHAK13a was very low or undetectable. MaHAK1, 2b and 10d showed low expression levels in roots, and MaHAK17 was expressed at low levels in both roots and pseudostems under both K + treatment conditions. MaHAK14b showed very low expression in roots under normal K + conditions but was upregulated under K + -deficient conditions. In addition, MaHAK14b was normally expressed at low levels in leaves in the K + treatment conditions but was highly expressed in the pseudostem. MaHAK7b was indeed upregulated under K + -deficient treatment in roots and leaves. To analyze the expression pattern of MaHAK genes when plants displayed evident symptoms (60 days after treatment) Six MaHAK genes were differentially expressed in normal-K + roots (NKRs) versus low-K + roots (LKRs) (Fig. 4b).
MaHAK10c and MaHAK10d were strongly upregulated and downregulated after low-K + treatment, respectively. These results implied the different roles of these MaHAK genes in different tissues under different K + conditions.

Functional characterization of MaHAK1 in potassium uptake
According to the increased expression of MaHAK14b induced by K + deficiency in banana roots, MaHAK14b might be a high-affinity K + uptake transporter from banana. First, the subcellular location of MaHAK14b was examined by generating a construct expressing a fusion protein of MaHAK14b-GFP (green fluorescent protein). The recombinant construct and control vector were introduced into the leaves of Nicotiana benthamiana. As shown in Fig. 5a, the results indicate that MaHAK14b was solely located in the plasma membrane, while the green fluorescence of the GFP control was located in the nucleus, cytoplasm and plasma membrane.
Several reported HAK proteins could complement the growth defect of potassium-deficient yeast strain R5421, e.g., OsHAK1, OsHAK5 and ZmHAK5 (Yang et al. 2014;Chen et al. 2015;Qin et al. 2019). To examine the potassium uptake activity of MaHAK14b, R5421 yeast were transformed with MaHAK14b and the vector as a control. The growth of transformants was compared on AP medium containing various K + concentrations. Yeast R5421 expressing MaHAK14b grew in 1 mM KCl, while no growth was observed in R5421 transformed with empty vector (Fig. 5b). This result indicates that MaHAK14b is a high-affinity K + transporter with K + uptake activity.

Discussion
The HAK/KUP/KT gene family is the largest K + transporter family and plays important roles in K + transport across membranes in bacteria, fungi, and plants (Li et al. 2018a, b). Currently, the HAK family has been reported in several plant species, such as Arabidopsis, rice, maize, wheat, cassava and peach (Maser et al. 2001;Gupta et al. 2008;Zhang et al. 2012;Song et al. 2015;Cheng et al. 2018;Ou et al. 2018). However, a systematic analysis of the HAK gene family in banana has not yet been conducted. In this study, we identified 24 HAK genes in the banana genome, which were distributed into four groups based on their evolutionary relationships with Arabidopsis, rice and maize. The classification of MaHAKs was also supported by gene structure and conserved motif analyses. The exon number of MaHAK genes ranged from 6 to 11, which was different from that of rice (2 to 10) and maize (3 to 10) (Gupta et al. 2008;Zhang et al. 2012) and similar to that of peach (7-12) and cassava (6-10) (Song et al. 2015;Ou et al. 2018). The reason for the difference may lie in the absence of homologs of HAK4 (5 exons) and HAK22 (3 exons) in banana.
The HAK gene number of banana is higher than that of Arabidopsis (13 members) and relatively lower than that of rice (27 members) and maize (27 members) (Gupta et al. 2008;Zhang et al. 2012). However, the gene diversification of MaHAKs was lower than that of OsHAKs and ZmHAKs, with only 14 kinds of HAK genes. The high number of members of MaHAKs indicates an event of succession of expansion, and extensive duplication occurred in the banana genome. In particular, MaHAK10 and MaHAK11 are significantly expanded in banana compared with other species, while banana orthologs of 13 OsHAKs of rice could not be identified. These results indicate that there might be gene Gene duplication events are essential for the expansion and evolution of gene families (Maher et al. 2006). Six segmental duplication events with 12 MaHAK genes, which are all from groups II and III, were identified with MCScanX methods. Remarkably, these genes tend to be located in distal telomeric segments of chromosomes, which are evolutionary hotspots consisting of many fastevolving genes Schilling et al. 2020). These results are in agreement with previous reports that groups II and III of HAK/KUP/KTs have substantial functional diversity (Yang et al. 2009). Further studies are required to characterize the functions of these duplicated genes in the adaptability of banana to different K + conditions.
To further understand the functions of MaHAK genes, their expression profiles were analyzed in various banana tissues under different K + conditions. Generally, K + transporter genes are upregulated under low-K + conditions. In Arabidopsis, AtHAK5 was strongly induced by low-K + stress and was the most important AtKT/KUP/HAK involved in K + uptake in roots (Ahn et al. 2004;Pyo et al. 2010). OsHAK1 and OsHAK5 in rice were upregulated under K + -deficient conditions, with increased K + uptake in rice roots (Yang et al. 2014;Chen et al. 2015Chen et al. , 2017. We found that the expression pattern of many MaHAK genes is different from that of close homologs in other species. MaHAK5 did not display a similar expression pattern in banana roots. In both the short-and long-term low-K + treatments, the transcript levels of MaHAK1 and MaHAK5 did not change significantly in banana roots, whereas that of MaHAK5 increased in leaves. These findings suggest that MaHAK transcription is regulated by a different mechanism than homologs in other plants, or perhaps the expression pattern of MaHAKs varies during different periods of low K + conditions. Additionally, we found that MaHAK14b was rapidly induced after 4 h of K + deficiency and that none was upregulated after 60 days of K + starvation in banana roots. A yeast complementation assay was conducted to further demonstrate that MaHAK14b functions in K + uptake at a 1 mM K + concentration. These results provide insight into the potential roles of MaHAK14b in K + uptake under low K + conditions in banana roots.

Conclusion
In summary, a total of 24 MaHAK genes were identified from the banana genome. The MaHAK gene family was comprehensively analyzed by assessing phylogenetic relationships, gene structures, chromosomal distribution, synteny and expression patterns. Furthermore, functional characterization of MaHAK14b demonstrated that it is associated with K + uptake under K + stress. The present study provides fundamental knowledge for further studying the functional mechanism of potassium transporters in banana development and abiotic stress responses.