Physico-chemical properties of GmCAMTA proteins
Various physico-chemical properties like number of amino acids, protein molecular weight (MW), pI (Isoelectric point), number of atoms, instability and aliphatic indexes and GRAVY (grand average of hydropathy) determined online with the ProtParam tool are given in table 1. GmCAMTA11 and GmCAMTA9 are the shortest polypeptides comprising of 910 and 911 aa, while GmCAMTA1 is the longest possessing 1122 aa. Overall, their average length is ~1004aa with a range of ~200 aa mutual difference. The molecular weight of GmCAMTAs ranges between 126989.43 and 102394.56 kDa with an average MW of 112848.3 and the number of atoms is proportional to the molecular weight of each protein. Similarly, GmCAMTA8 has the highest pI value of 7.64 showing that GmCAMTAs have relatively lower pI. Moreover, they are also hydrophilic in nature as the GRAVY ranges between -0.625 (GmCAMTA1) and -0.394 (GmCAMTA12). Almost all of the GmCAMTAs are thermally stable as their aliphatic indexes match with that of the other globular proteins (highest in GmCAMTA12). While none of them is stable in test tube as the instability index of all GmCAMTAs is higher than 40. Except GmCAMTA8, the number of Asp+Glu (negatively charged aa) is higher than Arg+Lys (positively charged aa).
Table 1. Physico-chemical properties of all GmCAMTA proteins
Proteins
|
No. of aa
|
MW (kDa)
|
pI
|
Asp+Glu
|
Arg+Lys
|
No. of atoms
|
II
|
AI
|
GRAVY
|
Glyma.05g178200 (GmCAMTA1)
|
1122
|
126989.43
|
5.79
|
159
|
130
|
17668
|
47.68
|
75.97
|
-0.625
|
Glyma.08g135200 (GmCAMTA2)
|
1102
|
124429.56
|
5.66
|
156
|
124
|
17322
|
43.50
|
77.79
|
-0.596
|
Glyma.15g053600 (GmCAMTA3)
|
1088
|
122478.36
|
5.69
|
151
|
118
|
16999
|
49.97
|
76.17
|
-0.552
|
Glyma.08g072100 (GmCAMTA4)
|
1079
|
121261.43
|
5.72
|
148
|
122
|
16874
|
45.99
|
77.27
|
-0.522
|
Glyma.05g117000 (GmCAMTA5)
|
1088
|
121941.02
|
5.58
|
150
|
123
|
16966
|
45.00
|
75.90
|
-0.514
|
Glyma.08g178900 (GmCAMTA6)
|
1081
|
121609.22
|
5.58
|
150
|
117
|
16890
|
50.60
|
77.12
|
-0.541
|
Glyma.17g038800 (GmCAMTA7)
|
999
|
112283.90
|
6.81
|
131
|
126
|
15695
|
40.79
|
80.39
|
-0.490
|
Glyma.15g143400 (GmCAMTA8)
|
911
|
103084.74
|
7.64
|
105
|
106
|
14419
|
42.07
|
80.72
|
-0.428
|
Glyma.09g038300 (GmCAMTA9)
|
911
|
102969.37
|
6.55
|
109
|
103
|
14385
|
43.48
|
80.42
|
-0.420
|
Glyma.05g148300 (GmCAMTA10)
|
983
|
109314.7
|
5.43
|
132
|
103
|
15159
|
49.09
|
73.92
|
-0.547
|
Glyma.18g005100 (GmCAMTA11)
|
962
|
108036.77
|
5.39
|
137
|
109
|
14981
|
46.47
|
76.13
|
-0.520
|
Glyma.17g031900 (GmCAMTA12)
|
922
|
103869.70
|
7.39
|
104
|
104
|
14503
|
41.29
|
81.3
|
-0.394
|
Glyma.07g242000 (GmCAMTA13)
|
921
|
104293.99
|
7.03
|
107
|
105
|
14545
|
41.60
|
79.60
|
-0.455
|
Glyma.11g251900 (GmCAMTA14)
|
910
|
102394.56
|
5.54
|
130
|
106
|
14222
|
45.31
|
78.58
|
-0.531
|
Glyma.08g105200 (GmCAMTA15)
|
965
|
107767.11
|
5.81
|
124
|
104
|
14945
|
50.69
|
73.30
|
-0.552
|
(aa- amino acids, MW- molecular weight, pI- isoelectric point, Asp- Aspartate, Glu- Glutamate, Arg- Arginine, Lys- Lysine, II- instability index, AI- aliphatic index)
Phylogenetics and structure of GmCAMTAs
As mentioned earlier, CAMTA TFs are multiple-stress responsive. Enrichment of cis motifs involved in signal response in the promoters of Medicago CAMTA genes hints that they are likely to respond variedly to various signals like other CAMTAs (15). We dissected the regulatory region of GmCAMTAs (~2 kb upstream) online with PLANTCare which detected stimulus-specific cis motifs in their promoters. Overall, there are light (G-box, MRE and AE-box), drought (MBS), salt (MYB), pathogen (TC-rich repeats), wound (WUN, WRE), low temperature (LTR), gibberellin (GARE, P-box), auxin (AUXRE) and abscisic acid (ABRE) responsive cis elements as shown in figure 1A. The presence of multiple cis motifs in GmCAMTA genes represent their responsivity to multiple stimuli. Moreover, the cis element (CG-box) is the binding site of CAMTA TFs, thus CG-box presence within GmCAMTA promoters indicate the interaction of one GmCAMTA TF with another. The promoter analysis of GmCAMTA12 revealed MYC, MYB, MBS and G-box indicating its potential role in drought stress.
Using the online GSDS tool, the gene structure of all the GmCAMTAs was visualized in order to mutually compare their structural diversity. The length of GmCAMTA genes lie in the range between 3196 bp (GmCAMTA14) to 3947 bp (GmCAMTA6) with an average length of 3607 bp. The exons (yellow), introns (black lines) as well as 5` and 3` UTR regions (blue) of each gene are shown in figure 2B. A close observation of the number of exon-intron reveals a similar pattern such as except 3 genes (GmCAMTA7, GCAMTA10 and GmCAMTA15) which are comprised of 12 exons, the rest have 13 exons and 12 introns of variable length, showing their close mutual evolutionary relationship. This fixed numbering of intron and exon is a conserved characteristic of CAMTA which is descended from ancestors and is also demonstrated in the CAMTA family of other species (23). Similarly, except the last one/two exons, an ascending order in the exon size from 5` to 3` UTR can be observed across all the genes.
The four domains (CG-1, ANK, IQ and CaMBD) is the common conserved characteristic of all CAMTA TFs (23). Scanning the amino acid sequences of GmCAMTAs with online protein domains illustrator tool showed the same 4 domains in all 15 members (figure 1C). GmCAMTA TFs through IQ (calcium-independent)/CaMBD (calcium-dependent) interact with Calmodulin while through CG-1 domain, they bind the target DNA in a sequence-specific manner (CGCG/CGTG) at their promoter region. “ANK repeats” mediate protein-protein interactions. All these conserved domains along with other properties make GmCAMTA proteins the “transcription factors”. The high sequence specificity is common in the Calmodulin binding domain of Arabidopsis and soybean CAMTAs as shown in figure 1D.
An ML (Maximum Likelihood) tree was constructed which traced the evolutionary relationship among the CAMTA gene families of soybean, Arabidopsis, maize and tomato (figure 1E). Using MEGAX, the evolutionary tree was constructed from the full length aligned CAMTA protein sequences of the four species. A total of 37 CAMTAs including 6 from Arabidopsis, 15 from soybean, 7 from tomato and 9 from maize clustered into 4 distinct groups, A, B, C and D. GmCAMTA1-6, AtCAMTA1-3, SlCAMTA1 and 2 and ZmCAMTA3, 6, 7a and 7b might have co-evolved and thus clustered together in group A, representing the largest clade. Similarly, GmCAMTA10, 11, 14, 15, AtCAMTA4, SlCAMTA3, 4 and ZmCAMTA1 clustered in group C showing their mutual high homology. GmCAMTA8, 9, 12 and 13 grouped with AtCAMTA5, 6, SlCAMTA5, 6 and ZmCAMTA5 making clade D. Two orthologs (GmCAMTA7 and SlCAMTA7) comprised group B, which is the smallest clade. Clustering in the phylogenetic reconstruction indicates more mutual similarity and probably weak homology to the members of the other three bigger clusters. It is noticeable that except B, in the rest of clades, CAMTAs of the same species are at the tips of the same branches and vice versa retaining their intraspecific homology.
miRNA targets in GmCAMTA transcripts
miRNA target prediction is important in finding their role in plant growth, development in normal as well as stress conditions (18). Keeping the Expectation score ≤ 5, the 639 miRNAs were scanned and top of the list miRNA (for each GmCAMTA transcript) with the minimum E. value was taken using the online psRNATarget tool (40). A total of 10 unique miRNAs were predicted which potentially target the GmCAMTA transcripts by inhibiting translation or through cleavage (Table 2). Their length ranges between 19 bp (gma-miR1533) to 24 bp (gma-miR343b). The accessibility of target site (UPE) which is associated with identification of target site and energy required to cleave transcript, varies from 11.8 (gma-miR1533) to 21.6 (gma-miR5780c). The translation of GmCAMTA1 and GmCAMTA2 transcripts is potentially inhibited by the common gma-miR5780c while gma-miR6299 cleaves GmCAMTA8, GmCAMTA9, GmCAMTA12 and GmCAMTA13. Similarly, GmCAMTA5 and GmCAMTA6 have potential targets for gma-miR1533 while GmCAMTA11 and GmCAMTA14 are predicted to be cleaved by gma-miR2111b. gma-miR9726 is predicted to cleave GmCAMTA3 and gma-miR1522 GmCAMTA15. These in silico predictions require experimental validation which would extend our understanding of the mechanisms of Ca-CaM-CAMTA-mediated stress tolerance in plants.
Table 2. miRNA potential targets in GmCAMTA transcripts
miRNA
vs
Transcript
|
Alignment
|
E.V.
|
UPE
Target Access.
|
Inhibition
|
M
|
gma-miR5780c
vs
GMCAMTA1,2
|
AGGUUCAGGGACUAUAUAAUUU
: : : : : : : : : : : : : : : . :
GGAAAGGCCCUAAUAUAUUAGA
|
3.0
|
21.668
|
Translation
|
1
|
gma-miR9726
vs
GMCAMTA3
|
CUUCUUUUUUUAUUACGGAUAU
. : : . : : : . : : : : : : . : : : :
AGAGGAAAUGAUACUGCUUAUA
|
4.0
|
18.124
|
Cleavage
|
1
|
gma-miR343b
vs
GMCAMTA4
|
AGGCUUAGUUGAACUAGACAUUCU
: : : : : : . : : : . : : : : : . . : :
AACCAAUCA--UUUGGUCUGUGGGA
|
4.0
|
19.255
|
Cleavage
|
1
|
gma-miR1533
vs
GMCAMTA5
|
AGUAAUAAUAAAAAUAAUA
: : : : : : : : : : : : : : :
ACAUGAUUAUUUUUAUUUA
|
4.5
|
12.197
|
Cleavage
|
1
|
gma-miR5369
vs
GMCAMTA6
|
ACUGUAGGAGGAAAAGAGU
: : : : . . : . . : : : : : : . .
UCACAUUUUUUUUUUCUUG
|
4.0
|
11.815
|
Cleavage
|
1
|
gma-miR5041-5p
vs
GMCAMTA7
|
AACUCGUUCAACUUCUACUUU
: : : : : : : : : : : : : : . : . CAGAACAAGUUGAAUAUGGAG
|
3.5
|
13.098
|
Cleavage
|
1
|
gma-miR6299
vs
GMCAMTA8,9, 12, 13
|
ACUGUUUAGUUAUUAAAAUUUA
: : : : . : . . : : : : : : : . : . :
AGCCAAGUUGUUAAUUUUGAGU
|
4.0
|
15.803
|
Cleavage
|
1
|
gma-miR5030
vs
GMCAMTA10
|
GGCCAUUUUGUGUUUAACAAGA
: : : : : : . : . . : : : : : : :
UUUGUAAAAUAUGAAAUGGUCU
|
4.5
|
18.387
|
Cleavage
|
1
|
gma-miR2111b
vs
GMCAMTA11,14
|
AUUUGGAGUCCUACGUCUAAU
. : : : : : : : : . : : . : : : : : :
AGAACCUCUGGGUGUAGAUUC
|
3.5
|
13.344
|
Cleavage
|
1
|
gma-miR1522
vs
GMCAMTA15
|
UAAAGUAAAAUUCGUUAUUU
: : : . : : : . : : : : : : : :
UAUUCGUUCUGAACAAUAAA
|
5.0
|
14.966
|
Cleavage
|
1
|
The potential miRNA target sites predicted with the online psRNATarget tool. Keeping a Max. Expectation value = 5, Max UPE= 25. E.V – Expectation Value, M – Multiplicity
Chromosomal distribution and regulatory network of GmCAMTAs
The genome browser tool in NCBI mapped GmCAMTA genes to their respective chromosomes. The 15 GmCAMTA genes are unequally distributed over 8 out of 20 chromosomes of soybean as shown in figure 2. The figure depicts the complete size of each chromosome with the exact position of genes. Chromosome 8 has the highest number of GmCAMTA genes, i.e; 4, while chromosome 7, 9, 11 and 18 have got one only in each. In prokaryotes, due to the absence of nucleus, transcription and translation occur simultaneously in coupling phase. On the other hand, translation of mRNA is always executed outside the nucleus in eukaryotic cytoplasm and the proteins that work in nucleus have a nuclear localization sequence (signal). Transcription factors also work in nucleus thus after their translation in cytoplasm, they are directed to the nucleus which is mediated through their NLS. To find NLS, protein sequences of each GmCAMTA were submitted to the online cNLS mapper tool at http://nls-mapper.iab.keio.ac.jp/cgi-bin/NLS_Mapper_form.cgi. Keeping a cut off score of 5, at least one NLS in all GmCAMTAs and even more than one in some proteins were detected. Likewise, all of the 6 Arabidopsis CAMTAs possess only one NLS in the CG-1 domain of each protein (12). In contrast, rice CAMTAs have one NLS in the C-terminal and another in N of CG-1 domain (41). Experimental evidence shows that these domains perform diverse functions in the regulation of gene expression (42). The nuclear localization sequence in each of the 15 GmCAMTAs and predict that all of these transcription factors are localized in the nucleus.
In order to find the interaction network of GmCAMTA proteins to relate them with other pathways, the protein sequences were individually put in STRING database, which predicted a number of interactors (43). Thus they can aid in linking proteins of interest to other pathways and could lead to the discovery of novel pathways as well. The String database displayed a network of ~10 interactors for each GmCAMTA protein among which some sets were redundant. Thus a total of 48 unique proteins were predicted for 15 GmCAMTAs (Figure S1, Additional File 4). This vast interaction network (experimentally determined and software predicted) indicates the complex regulatory upstream/downstream pathways of CAMTAs. However, these in silico predicted interactions require experimental validation. Besides these predicted interactions, their orthologues in other species such as Arabidopsis or other legumes could also be exploited to search other potential interactors in soybean. Information of the 48 proteins is mentioned in table S1, Additional File 3.
GmCAMTAs as early drought stress-responsive TFs
The spatiotemporal expression in roots and leaves under 0, 1, 3, 6, 9 and 12 hours of simulated drought stress is shown graphically in figure 3. In roots, GmCAMTA2 was highly expressed during 3 hours of drought followed by GmCAMTA7 and GmCAMTA10. In contrast, GmCAMTA14 was downregulated during all the 5 time points followed by GmCAMTA8, GmCAMTA9 and GmCAMTA11. Overall, these transcription factors upregulated abruptly during 1 and 3 hours of stress and downregulated afterwards (figure 3a). The expression profile of GmCAMTAs in leaves at different stress durations is different as compared to roots (figure 3b). In leaves, they look uniform except GmCAMTA4 which is the highest upregulated gene followed by GmCAMTA5, GmCAMTA11 and GmCAMTA12. Interestingly, like roots, majority of these 15 genes retained the 3 hours trend in the leaves as well. It is obvious that the expression is relatively the highest at 3 hours after which it decreased until 12 hours.
The differential expression of GmCAMTA family is a result of their tightly regulated transcription. We speculate that in stress conditions, though the calcium ions continuously convey the stress signals through calcium signatures to the cytoplasm as well as nucleus however the intensity/amount of these signals is weighed and adjusted by the next signal relaying molecules, such as CaM, before conveying to CAMTA transcription factors. At times, this signal transduction to CAMTA is continuous with the same intensity, however, after certain period (3 h in our case), CAMTAs response is not the same throughout the course and seems to be unconcerned and even downregulated as the stress period continues. From the control samples (o h) in leaves and roots, it is also obvious that the expression of all CAMTAs is active at all times. In brief, the spatiotemporal expression pattern revealed that GmCAMTAs are upregulated in the early phase of drought thus are early drought stress-responsive transcription factors.
Arabidopsis overexpressing GmCAMTA12 exhibited enhanced drought tolerance
To evaluate the contribution of GmCAMTA12 protein in drought stress, we engineered Arabidopsis plants (35s:GmCAMTA12) to constitutively express GmCAMTA12 gene under 35s promoter. Prior to Arabidopsis transformation, the Agrobacterium tumefacians strain EHA105 transformants harboring overexpression cassette were verified (through PCR). (Figure S2, Additional File 4). After floral dip, several overexpressing lines were obtained of which we selected two independent stable homozygous lines (OE5-Overexpressing GmCAMTA12-Line5 and OE12-Overexpressing GmCAMTA12-Line12) for functional analysis. The expression of GmCAMTA12 in two transgenic Arabidopsis lines was validated through qPCR.
For drought assay, the two lines of OE GmCAMTA12 (OE5 and OE12) and the wild type plants were subjected to drought stress by withholding water for two weeks and then re-watered as shown in figure 5A. Initially, all the plants were growing normally until water was withheld. However, upon encountering drought, nearly all the wild type and transgenic Arabidopsis stopped growth, wilted and started turning yellow afterwards. After 14 days of continuous drought treatment, most of the wild type plants were completely dried as obvious from their phenotype (dried leaves). Unlike wt, most of the OE lines retained life processes which was evident from chlorophyll they retained in their leaves. After re-watering, majority of transgenic Arabidopsis rejuvenated but most of the wild type did not. The plants were then allowed to grow under normal conditions until seed harvest. As expected, the seed yield of wt and transgenic lines was unequal and OE lines developed more seeds than wild type Arabidopsis. Under well-watered conditions, the survival rate in soil under drought was ~100%, however withholding the water for two weeks and then re-watering, less than 60% wt plants survived while OE5 and OE12 showed ~83% and ~87% survival rate as shown in figure 4D. Obviously, the constitutive overexpression of GmCAMTA12 had enhanced the drought survival efficiency of transgenic Arabidopsis leading to better growth and development.
In their root length assay on ½ MS-mannitol medium as shown in figure 4B, wt plants grew longer roots than OE plants at 0 mM concentration of mannitol, however, on MS-mannitol, OE plants specifically OE12 developed longer roots than wt at all three concentrations (50 mM, 100 mM and 200 mM) of mannitol. Interestingly, the roots of OE12 plants were the longest at 200 mM mannitol. Root length (cm) is shown graphically in figure 4E.
Between the two overexpression lines, OE12 performed better than OE5 in both drought assays (Figure A, B), thus we subsequently inoculated seeds of wt and OE12 on ½ MS with various concentrations of mannitol to evaluate the germination rate (Figure 4C). As expected, we observed higher germination rate of transgenic line OE12 than that of wt at all four concentrations of mannitol. The germination rate was ~30% higher than wt at 50 mM mannitol. At 100 mM, the germination rate decreased in both types at a similar pace (~75 % in OE and ~45 % in wt). As the mannitol concentration increased to 150 and 200 mM, we saw a dramatic decline in the germination efficiency of wt (30% and 25%) as compared to OE (>65% and >60%) as shown in figure 4F. We can say that the constitutive expression of GmCAMTA12 has enhanced the germination efficiency of transgenic seeds under drought.
In order to check the performance of wt and OE lines at physiological level, the physiological indexes like proline and MDA contents, CAT activity and relative electrolyte leakage were determined in all plants subjected to stress. Under well-watered conditions, we observed no significant difference in the level of proline contents which was quite low in wt and OE lines. In contrast, proline contents calculated in drought-stressed plants had significantly elevated. In wt, the average value of proline was ~400 µg/g, while in OE, it was recorded in the range of 850 to 900 µg/g (Figure 4G). Malodinaldehyde (MDA) is a well-known biomarker for sorting out stress-induced membrane damage due to oxidative stress. MDA contents in wt and OE plants during normal conditions matched mutually but its level was doubled in the drought-stressed wt plants compared with drought treated OE lines (figure 4H). Catalase (CAT) is a major antioxidant enzyme which accumulates during abiotic stresses and scavenges H2O2. The CAT activity in wt and OE plants under usual conditions was nearly equal (figure 4I) however, drought treatment enhanced CAT activity in transgenic plants as compare to wt. During stress, electrolytes, specifically K+ ions leak out of the cells through various channels and thus damage due ato stress could be monitored by comparing the electrolyte leakage in wt and transgenic lines. We determined relative electrolyte leakage (REL) in wt and OE lines during normal as well as drought conditions. REL % was nearly levelled under well-watered conditions but the leakage was more pronounced in wt as compared to OE lines during drought (figure 4J). Noticeably, the amino acid sequence analysis revealed high level of sequence similarity between GmCAMTA12 and AtCAMTA5. We can speculate that GmCAMTA12 being a transcription factor interacted with the downstream target genes (including AtCAMTA5 interactors) and modulated their expression in transgenic Arabidopsis which contributed to their better performance under drought (determined at molecular level in later section).
GmCAMTA12 overexpression regenerated more developed and drought-efficient hairy roots in soybean
For further functional validation of GmCAMTA12 in response to drought, the hairy roots system was exploited to overexpress the target gene in soybean. Prior to generating transgenic hairy roots, the Agrobacterium rhizogenes strain K599 cells harboring control vector (VC) and OEGmCAMTA12 were verified through gene-specific PCR (Figure S3, Additional File 4). After 1-2 weeks of infection of soybean seedlings with K599 (VC and OE), hairy roots had started regenerating with various frequency. Prior to subsequent experiments, we made sure that the hairy roots were transgenic by using gene-specific PCR from the genomic DNA of a small portion of hairy roots. Using qPCR, we validated the overexpression of the target gene in OE roots which was over 3 times higher than the VC hairy roots. After hairy roots had become long and strong enough by growing for two more weeks, the primary roots were excised and the chimeric plants (with transgenic roots and non-transgenic shoots) were shifted to fresh vermiculite.
When the transgenic roots were ~ 10 cm long, the VC and OE chimeric soybean plants were extirpated from vermiculite and transferred to Hoagland solution as shown in figure 5A. After a 3 days acclimation, both VC and OE chimeric plants were treated with 6% PEG6000. The plants started to wilt with leaves curling and shoot apexes drooping on encountering drought. However, the wilting was more apparent in VC than in OE chimeric plants as VC leaves and apexes were more wilted. GmCAMTA12 overexpression induced profuse hairy roots (figure 5B) due to which the aerial non transgenic part of the OE chimeric soybean plant was also more developed than the VC chimeric plants. The roots were analyzed using the root scanning (figure 4C). The OE hairy roots showed higher values with total root length (figure 5J), surface area (figure 5K), root volume (figure 5L), number of branches (figure 5M) and projected area (figure 5N).
The proline and MDA contents, CAT activity and relative electrolyte leakage were determined to check the impact of GmCAMTA12 overexpression at physiological level. In control samples (Hoagland), proline contents had nearly equal amount in VC and OE hairy roots, however under drought, proline content level was recorded significantly higher in OE hairy roots (figure 5F). MDA shows the level of membrane damage as it is the final product of lipid peroxidation. In the absence of stress, MDA contents in OE type were slightly less than VC hairy roots, however with PEG treatment, VC had substantial increase in MDA contents as compare to OE roots (figure 5G). In contrast, CAT activity was significantly higher in OE hairy roots than VC under drought stress (figure 5H). For REL, VC hairy roots had higher electrolyte leakage (%) than OE in response to drought (figure 5I). Comparative physiology of OE and VC hairy roots shows that the overexpression of GmCAMTA improved the drought tolerance of OE.
To analyze their growth efficiency under drought stress, 0.1 g of VC and OE hairy roots was weighed under sterile conditions and inoculated on GM media containing 4 various mannitol concentrations (50 mM, 100 mM, 150 mM and 200 mM) including a control (0 mM mannitol) with three replicates of each type (Figure S4, Additional File 4). All the plates were kept in dark in growth room with 28 ℃ for 10 days of culturing after which the fresh and dry weights were determined. The control samples of both types (VC and OE) hairy roots had nearly same fresh and dry weights however on stress media, OE hairy roots showed better performance as the weight of OE roots was more than VC roots at all 4 concentrations of mannitol (figure 5D, 5E). It further flaunted that the overexpression of GmCAMTA12 has improved the drought survival efficiency of OE roots.
Expression analysis of GmCAMTA12 orthologues’ regulatory network in Arabidopsis
In Arabidopsis, AtCAMTA5 is the orthologue of GmCAMTA12, thus the regulatory network of AtCAMTA5 was predicted with STRING database. To test our hypothesis whether the transcript abundance of GmCAMTA12 TF in transgenic Arabidopsis would interact with genes in the regulatory network of AtCAMTA5, we analyzed the expression of 10 interactors predicted with STRING database (figure 6A). Total RNA from drought treated WT and OE plants (figure 6B) was isolated and reverse transcribed into cDNA. Using cDNA and gene-specific primers, we conducted a qPCR which deciphered the differential expression pattern of the 10 genes in wt and OE plants (figure 6C). Among them, NIP30 (NEFA-interacting nuclear protein - AT3G62140) which in humans is involved in negative regulation of proteasomal protein catabolic process, is slightly upregulated in response to drought. AtWRKY14 (AT1G30650) encoding WRKY transcription factor14 possesses a DNA binding domain and specifically interacts with W box (a common elicitor-responsive cis element). WRKY14 is also nuclear localized transcription factor and regulates many important processes through gene regulation. GmCAMTA12 overexpression upregulated AtWRKY14 which indicates that along with other transcription factors including WRKY TFs, the spectrum of CAMTA TFs regulatory networks is much wider than we think. Thus the mutual interaction of CAMTA and WRKY should be further investigated. Peptdyl-prolyl cis-trans isomerase AtCYP59 (AT1G53720) mediates posttranslational modifications specifically protein folding. With GmCAMTA12 overexpression, the upregulation of AtCYP59 is almost equal to that of WRKY14. AtANN5 (AT1G68090) is a calcium binding protein, plays role in pollen development and is induced by cold, heat, drought and salt stresses. As expected, its expression is relatively the highest among the 10 interactors. SR (Serine racemase- AT4G11640) is involved in serine biosynthetic pathway. Its expression also seems to elevate with GmCAMTA12 overexpression in OE Arabidopsis. ELO3 (Elongator complex3- AT5G50320) is a part of Elongator multiprotein complex and regulates initiation and elongation of transcription. No apparent change in the expression level of ELO3 in wt and OE was observed. Like ANN5, CaMHSP (Calmodulin Binding Heat Shock Protein - AT3G49050) also called as Alpha/beta-Hydrolases superfamily protein exhibited higher transcript level. It is involved in lipid metabolic pathway. B120 (G-type lectin S-receptor-like serine/threonine-protein kinase - AT4G21390) is involved in protein kinase activity and recognition of pollen. Its expression level was downregulated in transgenic Arabidopsis. AT2G43110 (U3 containing 90S pre-ribosomal complex subunit) was also upregulated with the overexpression of GmCAMTA12. AT3G19850 (BTB/POZ domain-containing) mediates protein degradation by facilitating ubiquitination. Its expression level elevated with GmCAMTA12 upregulation. All of these genes possess CGCG/CGTG motif in their promoter sequences which also validates the sequence-specific binding of CAMTA TFs. Most of these interactions are based on text-mining and should be determined experimentally. The genomic, CDS and proteomic sequences of these 10 interactors can be found in (Table S4. Additional File 5)
GmCAMTA12 overexpression orchestrated downstream genes in transgenic hairy roots
In order to find whether the constitutive overexpression of GmCAMTA12 modulates the genes with which GmCAMTA12 TF interacts, we selected 8 genes in the GmCAMTA12 regulatory network in soybean predicted with STRING (figure 6D). In chimeric soybean plants possessing VC and OE (figure 6E) hairy roots treated with 6% PEG6000, we analyzed the expression pattern of the eight of the predicted interactors of GmCAMTA12 to know their GmCAMTA12-mediated regulation. Using gene-specific primers, qPCR of the selected interacting proteins in transgenic hairy roots as well as non-transgenic leaves was carried out. The genes displayed a differential expression profile in the control and drought treated chimeric soybean plants. In VC roots (figure 6H), GmNIP30 (NEFA interacting protein - GLYMA19G40910) was slightly upregulated in response to drought and its expression was indifferent to GmCAMTA12 overexpression in OE roots (figure 6I). GmPLA1-IId (Phospholipase A1-IId - GLYMA12G15430) involved in lipid metabolic process was upregulated in VC roots under drought stress while in OE roots, its upregulation was two-fold higher than that in VC. GmNAB (Nucleic Acids Binding - GLYMA18G40360) was downregulated in VC roots while upregulated with GmCAMTA12 overexpression in OE roots in response to drought. The expression of GmELO (Catalytic histone acetyltransferase subunit of the RNA polymerase II elongator complex - GLYMA06G18150) was the highest and it equally expressed in both VC and OE roots. GmSR1 (Serine racemase-1 involved in D-Serine biosynthetic process - GLYMA05G37930) was upregulated in VC roots however, with GmCAMTA12 overexpression, GmSRI was downregulated in OE roots under drought. Similarly, GmSR2 (Serine racemase-2 - GLYMA08G01670) was repressed in the absence of GmCAMTA12 overexpression, however in OE roots under drought, it was slightly upregulated. Simultaneously, GmUC1 (Uncharacterized but possessing a Myb-DNA binding domain - GLYMA20G32540) transcript was slightly upregulated in VC while it was significantly induced in OE hairy roots in response to drought. In contrast, GmUC2 (Uncharacterized2 - GLYMA17G07200) was positively regulated in VC and negatively regulated in OE roots. Unlike roots, the expression profile of these 8 genes was nearly similar in chimeric soybean leaves possessing VC (figure 6hF) and OE (figure 6G) hairy roots in response to drought. Five out of eight of these genes possess CGCG/CGTTG motif. The genomic, CDS and proteomic sequences of the 8 genes can be found in Table S5, Additional File 5.