Identification of HAK genes in sugarcane
Based on comparative genomics, 29 SbHAK genes were identified from sorghum (Sorghum bicolor, sugarcane’s nearest relative). Using the protein sequences of sorghum HAK genes as a reference, 30 distinct S. spontaneum HAK genes (Table 1), excluding alleles, were identified from the genome of tetraploid S. spontaneum AP85-441 [28]. Each of these genes contained one to four alleles, with an average of 3 (Additional file 1). The 30 SsHAK genes were distributed on seven S. spontaneum chromosomes: chromosome 1 contained six genes; chromosome 2 contained seven genes; chromosome 3 contained four genes; chromosome 4 contained two genes; chromosome 5 contained five genes; and chromosome 6 and 8 each contained three genes. No SsHAK genes were identified on chromosome 7 (Additional file 1).
All 30 predicted SsHAK proteins had a typical “K_trans” domain (PF02705), which is specific to HAK/KUP/KT potassium transporter family members. For consistency, these SsHAK genes were named based on the previously reported O. sativa HAK nomenclature and phylogenetic relationships [17]. If two SsHAK genes were equally close to a single OsHAK gene, then the same name was used, followed by the letters “a” and “b” (Table 1). Two paralogous SsHAK genes (SsHAK19a and SsHAK19b) were identified that corresponded to the same sorghum gene, Sobic.006G062100, which may be caused by gene loss in sorghum or gene duplication in sugarcane. The number of amino acids in the 30 identified SsHAKs ranged from 487 to 967, with an average of 758. The predicted isoelectric points (pI) of the SsHAKs varied from 5.88 to 9.26, and the average pI was 8.15. The molecular weight ranged from 55.84 kDa to 106.49 kDa, with an average of 84.47 kDa (Table 1). The prediction of transmembrane domains in the SsHAK proteins indicated that most contained 11 or 12 transmembrane helices, which was similar to the findings in sorghum. The subcellular locations of the SsHAK proteins predicted by WoLF PSORT were mainly the plasma membrane, which is most suitable for their roles as transporters to maintain K+ homeostasis in sugarcane. In addition, the SsHAK proteins were also located on some organelles, including the endoplasmic reticulum, vacuole, cytoplasm, Golgi body and chloroplast. Protein sequence alignment of SsHAKs with their orthologs in sorghum showed that S. spontaneum and Sorghum bicolor shared identities ranging from 81% to 98%, with an average of 92.5% (Table 1). Four hundred thirty-five pairwise protein sequence comparisons among these SsHAKs showed that SsHAK19a and SsHAK19b shared the highest identity (96%), while other gene pairs had protein sequence similarities ranging from 28% to 82% with an average of 46%, indicating that the SsHAKs are an ancient gene family with high sequence divergence (Additional file 2).
To investigate the possible evolutionary functional constraints after the split of sorghum and sugarcane, the nonsynonymous to synonymous substitution ratios (Ka/Ks) between SsHAKs and their orthologous genes in sorghum were calculated. The results showed that the Ka/Ks ratios were less than 0.5, except for SsHAK13, suggesting that purifying selection was the main force driving the evolution of HAK genes (Fig. 1).
Phylogenic analysis of HAK genes in S. spontaneum and representative angiosperms
To analyze the evolution of the HAK gene family in S. spontaneum and different plants, a total of 278 HAK genes from 14 representative angiosperms and a HAK member from Chlamydomonas reinhardtii as the outgroup were used to construct a phylogenetic tree using the neighbor-joining method (Fig. 2, Additional file 3). The 278 HAK genes included 6 from Amborella trichopoda, 8 from Solanum lycopersicum, 13 from Vitis vinifera, 8 from Carica papaya, 13 from Arabidopsis thaliana, 12 from Ananas comosus, 25 from Brachypodium distachyon, 27 from Oryza sativa, 28 from Setaria italica, 28 from Setaria viridis, 27 from Zea mays, 29 from Sorghum bicolor, 30 from Saccharum spontaneum and 24 from Saccharum hybrid R570 [29]. The amino acid sequence of the 279 HAK/KUP/KT transporters from 15 representative plant species is provided in the supplementary data (Additional file 4).
Table 1
Overview and comparison of HAK genes in Saccharum spontaneum and Sorghum bicolor
Sorghum bicolor
|
|
Saccharum spontaneum
|
Similarityf
|
Gene
|
AAa
|
pIb
|
Mwc (kDa)
|
TMSd
|
P.L.e
|
|
Gene
|
AAa
|
pIb
|
Mwc
(kDa)
|
TMSd
|
P.L.e
|
Sobic.006G061300
|
788
|
8.75
|
87.13
|
12
|
PM
|
|
SsHAK1
|
780
|
8.83
|
86.84
|
12
|
PM
|
94.42%
|
Sobic.003G418100
|
783
|
8.91
|
87.53
|
12
|
PM
|
|
SsHAK2
|
788
|
8.85
|
88.18
|
12
|
PM
|
94.61%
|
Sobic.003G164400
|
811
|
8.4
|
89.60
|
10
|
PM/ER
|
|
SsHAK3
|
785
|
8.69
|
86.79
|
11
|
PM
|
97.34%
|
Sobic.007G153001
|
706
|
8.37
|
78.02
|
9
|
PM/ER
|
|
SsHAK4
|
702
|
8.90
|
78.08
|
9
|
PM
|
92.92%
|
Sobic.003G413600
|
775
|
8.78
|
86.36
|
11
|
PM
|
|
SsHAK5a
|
705
|
8.39
|
78.76
|
11
|
PM
|
85.64%
|
Sobic.003G413700
|
775
|
8.54
|
86.42
|
11
|
PM
|
|
SsHAK5b
|
750
|
7.58
|
83.86
|
10
|
PM
|
93.35%
|
Sobic.002G411500
|
788
|
8.8
|
87.72
|
13
|
PM
|
|
SsHAK7
|
818
|
8.81
|
91.32
|
13
|
PM/Vacu
|
90.95%
|
Sobic.001G379900
|
805
|
7.36
|
89.80
|
12
|
PM/Cyto
|
|
SsHAK8
|
770
|
8.36
|
85.88
|
11
|
PM/ER
|
93.18%
|
Sobic.002G417500
|
792
|
6.96
|
87.53
|
12
|
PM/Cyto
|
|
SsHAK9
|
743
|
8.39
|
82.35
|
11
|
PM/ER
|
91.34%
|
Sobic.010G197500
|
820
|
8.37
|
91.15
|
10
|
PM/ER
|
|
SsHAK10
|
755
|
8.94
|
83.57
|
10
|
PM/Vacu
|
90.52%
|
Sobic.006G213500
|
805
|
8.33
|
89.66
|
13
|
PM/ER
|
|
SsHAK11
|
719
|
7.24
|
80.33
|
12
|
PM/ER
|
92.06%
|
Sobic.007G075100
|
790
|
8.21
|
88.50
|
14
|
PM
|
|
SsHAK12
|
509
|
8.54
|
57.87
|
8
|
PM
|
87.93%
|
Sobic.010G224400
|
779
|
8.97
|
85.92
|
12
|
PM/Cyto
|
|
SsHAK13
|
757
|
8.62
|
83.38
|
12
|
PM/ER
|
95.76%
|
Sobic.002G313900
|
843
|
5.71
|
93.38
|
12
|
PM/ER
|
|
SsHAK14
|
811
|
5.88
|
90.03
|
11
|
PM
|
91.12%
|
Sobic.006G210700
|
743
|
8.85
|
82.93
|
12
|
PM/ER
|
|
SsHAK15
|
852
|
6.00
|
95.04
|
12
|
PM/ER
|
90.12%
|
Sobic.001G184000
|
817
|
8.91
|
92.60
|
12
|
PM
|
|
SsHAK16a
|
487
|
9.26
|
55.84
|
8
|
PM/Cyto
|
81.06%
|
Sobic.001G184100
|
810
|
8.61
|
91.65
|
11
|
PM/ER
|
|
SsHAK16b
|
802
|
8.69
|
91.07
|
12
|
PM/ER
|
96.03%
|
Sobic.002G220600
|
708
|
8.77
|
78.15
|
12
|
PM
|
|
SsHAK17
|
701
|
9.06
|
78.01
|
12
|
PM
|
93.57%
|
Sobic.002G130800
|
787
|
8.69
|
88.61
|
14
|
PM/ER
|
|
SsHAK18
|
788
|
8.35
|
88.56
|
14
|
PM/ER
|
96.45%
|
Sobic.006G062100
|
746
|
7.29
|
83.31
|
12
|
PM/Golgi
|
|
SsHAK19a
|
767
|
7.00
|
85.62
|
10
|
PM/Golgi
|
94.78%
|
Sobic.006G062100
|
746
|
7.29
|
83.31
|
12
|
PM/Golgi
|
|
SsHAK19b
|
730
|
6.65
|
81.30
|
9
|
PM/Vacu
|
93.33%
|
Sobic.004G160000
|
735
|
8.46
|
80.43
|
12
|
PM/ER
|
|
SsHAK20a
|
730
|
8.81
|
80.09
|
12
|
PM/ER
|
97.01%
|
Sobic.006G061700
|
788
|
8.66
|
88.27
|
11
|
PM/Cyto
|
|
SsHAK20b
|
794
|
8.60
|
89.03
|
11
|
PM/Golgi
|
83.01%
|
Sobic.001G183700
|
828
|
8.51
|
92.29
|
11
|
PM/Cyto
|
|
SsHAK21
|
818
|
8.22
|
91.50
|
11
|
PM/ER
|
95.17%
|
Sobic.002G001800
|
931
|
8.61
|
102.07
|
12
|
PM/Chlo
|
|
SsHAK22
|
967
|
9.08
|
106.49
|
11
|
PM/Vacu
|
88.52%
|
Sobic.002G188600
|
852
|
6.78
|
93.82
|
12
|
PM/ER
|
|
SsHAK23
|
846
|
6.55
|
93.13
|
12
|
PM
|
98.00%
|
Sobic.010G112800
|
773
|
8.39
|
85.44
|
12
|
PM/Chlo
|
|
SsHAK24
|
698
|
7.62
|
77.44
|
10
|
PM/Chlo
|
96.94%
|
Sobic.004G250700
|
774
|
7.34
|
86.29
|
13
|
PM/ER
|
|
SsHAK25
|
800
|
7.13
|
89.27
|
14
|
PM/ER
|
94.62%
|
Sobic.007G209900
|
774
|
9.08
|
82.47
|
10
|
PM/Chlo
|
|
SsHAK26
|
744
|
8.98
|
82.93
|
10
|
PM/Chlo
|
89.63%
|
Sobic.001G184300
|
814
|
8.32
|
91.82
|
11
|
PM/ER
|
|
SsHAK27
|
812
|
8.44
|
91.41
|
11
|
PM/ER
|
97.67%
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
PM= plasma membrane, ER= endoplasmic reticulum, Vacu= vacuole, Cyto= cytoplasm, Golgi= Golgi body, Chlo= chloroplast
a Amino acid number in HAK protein sequences
b Isoelectric point (pI) predicted by ExPASy (https://web.expasy.org/compute_pi/)
c Molecular weight (Mw) predicted by ExPASy (https://web.expasy.org/compute_pi/)
d Number of transmembrane domains possessed by HAKs, as predicted by TMHHM Server v.2.0 (http://www.cbs.dtu.dk/ services/TMHMM/)
e Subcellular location of the HAK proteins predicted by WoLF PSORT (https://www.genscript.com/wolf-psort.html)
f Protein sequence similarity between sorghum and sugarcane calculated by BLASTP
These HAK genes could be divided into four clades (I, II, III, IV) based on previously reported OsHAKs [17]. In A. trichopoda, the earliest diverging angiosperm, there were only 6 HAK genes, while in dicots and monocots, the number of HAKs ranged from 8 to 30 (Fig. 2, Fig. 3), indicating that the ancient whole-genome duplication (WGD) contributed to the expansion of the HAK gene family in both dicots and monocots. Clade II and clade III included HAK genes from all 14 angiosperm genomes, indicating that the progenitors of these genes may have already existed prior to the split from angiosperms (Fig. 2, Fig. 3). Clade I and clade IV mainly contained HAK genes from monocotyledons. Eighty-three HAK genes were identified in clade I, in which only one HAK gene was from A. comosus (Aco006685, homologous with SsHAK5) and Arabidopsis (AtHAK5), and the other 81 HAK genes were from all eight examined Poaceae species (Fig. 2, Fig. 3). Twenty-nine HAKs were grouped into clade IV, and only 2 of them were from dicotyledons. These results indicated that the HAKs were unevenly distributed.
Based on the pairwise synonymous substitution rates (Ks) in Sorghum bicolor and S. spontaneum (Additional file 5), the divergence time among the four clades of the HAK family was estimated. The median values of pairwise Ks varied from 1.644 to 2.851, corresponding to a divergence time ranging from 134.8 to 233.7 Mya, suggesting that HAKs in the four clades were ancient and divergent. Moreover, the divergence time between two pairs of duplicated SsHAKs (SsHAK5a/5b and SsHAK16a/16b) ranged from 18.94 to 58.14 Mya (Additional file 6). These results suggested that the SsHAK family is an ancient gene family with recent gene duplication events.
Exon/intron organization of the HAK family in S. spontaneum and other angiosperms
To investigate the structural characteristics and evolution of the HAK gene family, the exon/intron organization in HAKs was mapped to the phylogenic tree, and the gene features and patterns were analyzed (Fig. 2). The exon number in the HAK family of the 15 examined plant species ranged from 2 to 16, with an average of 8.4, and 217 out of 279 (77.8%) HAK genes possessed 8 to 10 exons (Additional files 7 and 8). This result suggested that the last common ancestor (LCA) of angiosperm HAK genes had 8 to 10 exons.
The exon number of SsHAKs varied from 2 to 12, and half of the SsHAKs possessed 8 or 9 exons. The pattern of SsHAK gene structure was similar to that of HAK gene structure from sorghum and maize in the same clade, suggesting that the HAK gene structure in the Panicoideae was relatively conserved. In clade I, the exon number in HAK genes varied from 2 to 12, which was the most variation among these 4 clades. Notably, the HAK genes in the subfamily with SsHAK22 had only 2 to 4 exons; however, the protein size remained consistent, which was likely due to the loss of introns. Clade II had the largest number of HAK genes, with 60 out of 98 HAKs possessing 9 exons and 5 out of 9 SsHAKs harboring 8 exons. SsHAK3/8/10 had one less exon than their orthologous genes in sorghum; the first exon in SsHAK13 and seventh exon in SsHAK24 were smaller than the corresponding exons in sorghum, and both resulted in shorter amino acid sequences in S. spontaneum (Table 1, Fig. 2). In clade III, the exon number was relatively conserved, with 61 out of 68 HAK genes possessing 8 to 10 exons, while the gene size varied greatly, mainly due to the different sizes of introns. The exon number in clade IV ranged from 2 to 8, with an average of 7, which was smaller than that in other clades. Notably, the HAK genes in the subfamily with SsHAK4 had only 2 to 5 exons, which was likely caused by intron loss during the process of evolution. The results indicated that HAKs underwent gene structure reconstruction under different evolutionary dynamics in S. spontaneum and other angiosperms in this study.
Expression analysis of HAK genes in Saccharum species
To study the expression profiles and potential functions of HAKs in Saccharum, we compared the gene expression patterns according to 4 sets of RNA-seq data: 1) different developmental stages and tissues; 2) a leaf gradient; 3) the circadian rhythm; and 4) treatment under low-potassium stress. The FPKM values of HAK1, HAK7 and HAK20b in YT55 at 0 h, 6 h, 12 h, 24 h, 48 h and 72 h under K+-starvation conditions were verified by RT-qPCR. The relative expression level was positively correlated with the FPKM value (R² = 0.8419, Additional file 9), suggesting the reliability of gene expression based on the RNA-seq analysis.
Expression pattern of HAKs in different tissues at different stages
To study gene functional divergence among the Saccharum species, transcriptome profiles of HAKs between two Saccharum species, S. officinarum and S. spontaneum, were analyzed based on RNA-seq at three developmental stages (seedling, premature and mature stages) in five different tissues, 2 leaf (leaf and leaf roll) and 3 stalk (immature, maturing and mature) tissues (Fig. 4). Among the 30 HAK genes analyzed, 18 genes (HAK3/4/5a/5b/12/13/14/15/16a/16b/17/19a/19b/20a/20b/21/22/26) showed very low or undetectable expression levels in all examined tissues of the two Saccharum species. HAK1 and HAK2 had different expression patterns in the two Saccharum species. HAK1 had higher expression levels in S. spontaneum than in S. officinarum, and the expression level in leaves was higher than that in stems at three different stages. HAK2 had higher expression levels in S. officinarum than in S. spontaneum, and the expression level in stems was higher than that in leaves. HAK8 was mainly expressed in the upper stems, while the expression levels in the middle and lower stems were very low. HAK9 and HAK10 had higher expression levels in stems than in leaves. HAK18 was expressed in all examined tissues, with higher expression levels, especially in leaves at the seedling stage and in mature stems. Notably, HAK27 was highly expressed in leaves at all examined three stages, but the expression level in stems was very low or undetectable.
Expression pattern of HAKs across a leaf segment gradient
To further explore the functional divergence of HAK genes for photosynthesis in the source tissues, we studied the expression pattern of HAKs in continuously developing leaf segment gradients from S. officinarum and S. spontaneum (Fig. 5). Saccharum leaves were divided into four zones: the basal zone (sink tissue), transitional zone (undergoing sink-source transition), maturing zone and mature zone (fully differentiated zone with active photosynthesis), following the method described in maize [30]. Consistent with the expression pattern at different developmental stages, 18 HAK genes (HAK3/4/5a/5b/12/13/14/15/16a/16b/17/19a/19b/20a/20b/21/22/26) showed very low or undetected expression levels in all examined leaf segments, suggesting their limited roles in sugar transport (Fig. 5). HAK1 and HAK2 showed higher expression levels in the basal zone than in the other 3 zones. The expression level of HAK7 increased gradually from the base to the tip of the S. spontaneum leaf, while in S. officinarum, HAK7 displayed higher expression levels in the maturing zone than in the other 3 zones. The expression level of HAK8 decreased gradually from the base to the tip of the leaf in both S. officinarum and S. spontaneum. HAK9 showed different expression patterns in S. spontaneum and S. officinarum. In S. spontaneum, the expression level of HAK9 increased gradually from the basal zone to the maturing zone and then decreased in the mature zone. In S. spontaneum, the expression level of HAK9 decreased from the transition zone to the maturing zone and then increased in the mature zone, and the expression level was much higher in S. officinarum, suggesting gene functional divergence after the split of these two Saccharum species. HAK10 showed higher expression levels in the transition zone in S. spontaneum and higher expression levels in the mature zone in S. officinarum. HAK18 displayed higher expression levels in the maturing zone in both S. spontaneum and S. officinarum, while HAK23 showed higher expression levels in the basal zone in the two Saccharum species. HAK25 displayed higher expression levels in the maturing zone in S. officinarum but higher expression levels in the basal zone in S. spontaneum.
Expression pattern of HAKs during the circadian rhythm
Acting as an enzyme activator, potassium ions participate in a series of photosynthetic processes [31]. To analyze the expression pattern of HAKs during diurnal cycles, we investigated the transcriptome profiles of the mature leaves in the two Saccharum species at 2 h intervals over a 24 h period and at 4 h intervals over an additional 24 h. Consistent with the transcriptome profiles at different developmental stages and in the leaf segment gradient, 18 genes (HAK3/4/5a/5b/12/13/14/15/16a/16b/17
/19a/19b/20a/20b/21/22/26) displayed very low or undetectable expression levels in the two Saccharum species, further supporting their limited roles in growth and development (Fig. 6). In addition, HAK8 and HAK24 also showed low expression levels over the two 24 h periods. HAK1, HAK2, HAK7, HAK18 and HAK27 showed higher expression levels in S. officinarum than in S. spontaneum, while HAK9 and HAK10 displayed higher expression levels in S. spontaneum than in S. officinarum. HAK1 and HAK2 had no diurnal expression pattern in the two Saccharum species. HAK7 displayed a higher expression level at night than in the daytime and showed the lowest expression level at noon in S. officinarum but showed no diurnal expression pattern in S. spontaneum. HAK10 displayed a higher expression level at night than in the daytime in S. spontaneum but showed no diurnal expression pattern in S. officinarum. HAK9 displayed a higher expression level at night than in the daytime in both Saccharum species. HAK18 and HAK27 displayed higher expression in the morning in the two Saccharum species. These findings suggested the functional divergence of the HAK genes in diurnal rhythms.
Expression pattern of HAKs under K+-deficiency stress
To investigate the functional divergence of HAK genes in response to low-potassium stress in sugarcane, we studied the expression profiles of HAKs in roots from the Saccharum hybrid variety YT55 at 0 h, 6 h, 12 h, 24 h, 48 h and 72 h under low-K+ stress (Fig. 7). Among the 30 HAK genes analyzed, 14 genes (HAK3/4/5a/5b/11/13/16a
/16b/19a/19b/20a/22/26/27) showed very low or undetectable expression levels before and after exposure to low-K+ stress. Notably, HAK1 showed strong induction in roots under low-K+ conditions, reached the highest level at 24 h, and then decreased subsequently at 48 h and 72 h. HAK21 was strongly induced after exposure to low-K+ stress within 12 h but was subsequently downregulated to a low expression level. HAK20b was downregulated within 12 h and then upregulated to the highest level at 72 h. HAK7, HAK10, HAK18 and HAK24 were downregulated after exposure to low-K+ stress. Other HAKs, such as HAK12/14/15/25, were constitutively expressed.
Functional complementation validation of SsHAK1 and SsHAK21 in the yeast mutant strain R5421
SsHAK1 and SsHAK21 were selected for complementary validation in yeast because they were both induced in response to low-K+ stress. The transformed yeast strain carrying only the empty vector pYES2.0 was used as a control. There were no obvious growth differences between yeast transformed with pYES2.0 and yeast transformed with pYES2.0-SsHAK1 or pYES2.0-SsHAK21 in SC/-ura medium containing 100 mM KCl (Fig. 8). However, when the KCl concentration decreased to 10 mM, the growth of yeast transformed with SsHAK1 and SsHAK21 was better than that of yeast transformed with the empty vector. When the KCl concentration decreased to 1 mM, the growth of yeast transformed with the empty vector was significantly inhibited, while the growth of yeast transformed with SsHAK1 or SsHAK21 was almost unaffected (Fig. 8). These results suggested that both SsHAK1 and SsHAK21 could recover the K+ absorption function in the yeast mutant strain R5421, indicating that they had potassium transporter activity.