Overexpression of Cassava MeAnn2 Enhances the Salt and IAA Tolerance of Transgenic Arabidopsis

Annexins are a superfamily of soluble calcium-dependent phospholipid-binding proteins that have considerable regulatory effects in plants, especially in response to adversity and stress. The Arabidopsis thaliana AtAnn1 gene has been reported to play a significant role in various abiotic stress responses. In our study, the cDNA of an annexin gene highly similar to AtAnn1 was isolated from the cassava genome and named MeAnn2. It contains domains specific to annexins, including four annexin repeat sequences (I–IV), a Ca2+-binding sequence, Ca2+-independent membrane-binding-related tryptophan residues, and a salt bridge-related domain. MeAnn2 is localized in the cell membrane and cytoplasm, and it was found to be preferentially expressed in the storage roots of cassava. The overexpression of MeAnn2 reduced the sensitivity of transgenic Arabidopsis to various Ca2+, NaCl, and indole-3-acetic acid (IAA) concentrations. The expression of the stress resistance-related gene AtRD29B and auxin signaling pathway-related genes AtIAA4 and AtLBD18 in transgenic Arabidopsis was significantly increased under salt stress, while the Malondialdehyde (MDA) content was significantly lower than that of the control. These results indicate that the MeAnn2 gene may increase the salt tolerance of transgenic Arabidopsis via the IAA signaling pathway.


Introduction
Annexins, similar to calmodulins, are multifunctional phospholipid-binding proteins that bind calcium ions and phospholipids in the membrane in response to external stimuli. Studies have shown that plant annexins are widely present in plant cells and have roles in plant growth, development, and environmental responses, amongst other processes. The structures of plant annexins are evolutionally conserved, whereby the N-terminal domain of each annexin is variable, while the C-terminal core domain is conserved. The N-terminal, as a variable region, may have different amino acid sequence lengths in each annexin, containing multiple protein-binding sites, such as for proteolysis and phosphorylation. Therefore, the N-terminal functionally serves as the regulatory region of the annexin molecule and the basis by which members of different annexin families can be distinguished [1][2][3][4]. The conserved C-terminal homologous domains are generally composed of four highly similar repeats of amino acids (Repeats I-IV), each containing about 70 amino acid residues. Generally, only Repeat I and Repeat IV contain the K-G-X-G-T-(38)-D/E motif, which is the type II Ca 2+ -binding site [1,3], while Repeat II and III do not have the above motif [1,3,4]. Each repeat has five α-helical folding (helix A-E) secondary structures, CaNAC4 transgenic Arabidopsis [35]. Overexpressing the CaNAC4 in Arabidopsis, the transgenic plants gain tolerance to salt stress, as their roots grew better than the control plants. Meanwhile, the root growth-related genes such as AtLBD18 and AtIAA4 were significantly up-regulated in transgenic plants [35].
Cassava (Manihot esculenta Crantz) is an important starch crop which has good tolerance to drought and barren soil [36,37]. In this study, we successfully cloned the cDNA sequence of an annexin gene MeAnn2, from cassava SC8 (South China No. 8) cultivar. It shares 70.9% amino acid sequence similarity with AtAnn1, and it contains domains specific to annexins. The overexpression of MeAnn2 reduced the sensitivity of transgenic Arabidopsis to various Ca 2+ , NaCl, and IAA concentrations. The expression of the stress resistancerelated genes in transgenic Arabidopsis was significantly increased under salt stress. These results provide new knowledge of the role of MeAnn2 gene in plant stress resistance.

Cloning and Characterization of MeAnn2 in Cassava
The cDNA of MeAnn2 (GenBank No. KM975563) was cloned from cassava ( Figure S1) based on its sequence homology to AtAnn1. The MeAnn2 coding sequence (CDS) was 951 bp and encodes 316 amino acids. The predicted molecular mass of MeAnn2 protein is 36.05 kDa, and its isoelectric point was 6.34. The hydrophilic index of MeAnn2 protein was calculated as −0.525, indicating that MeAnn2 is a stable hydrophilic protein. The protein instability index of MeAnn2 was calculated as 24.66, indicating that it is a stable protein.

Subcellular Localization of the MeAnn2 Protein
To explore the localization of the protein expressed by the MeAnn2 gene as an indicator of function, a fusion expression vector of pCAMBIA1300-35S-MeAnn2:GFP was constructed (Figure 2a), and transient expression was mediated by Agrobacterium tumefaciens in tobacco leaves. The empty vector pCAMBIA1300-35S-GFP was used as the control. After 3-5 days of co-culturing, laser confocal microscopy showed that the fluorescence signal corresponding to MeAnn2:GFP in transfected tobacco leaf cells could be observed in the cell membrane and cytoplasm. The fluorescence signal in cells transfected with control vector was observed in the cell membrane, cytoplasm, and nucleus ( Figure 2b). The dark and light blue shadows indicate the amino acid residues with greater than 50% identity. The black line marks the annexin repeat I-IV. The orange star indicates the conserved tryptophan required for Ca 2+ -independent membrane binding. The red line indicates the diacidic motif, an important signal for endoplasmic reticulum export to the plasma membrane. The orange boxes indicate the channel-function-related salt bridges, and the green boxes indicate the Ca 2+ -binding sequences. The red circle indicates the phosphatidylserine-binding motifs. (b) Phylogenetic relationships between MeAnn2 (Manihot esculenta, KM975563) and annexin proteins from other species. AtAnn1 (Arabidopsis thaliana, NP_174810), PeAnn1 (Populus euphratica, XP_011016329), OsAnn1 (Oryza sativa, XP_015644251), StAnn1 (Solanum tuberosum, XP_006355349.1), CkAnn (Cynanchum komarovii, GU067483), BjAnn1 (Brassica juncea, AAR10457).

Subcellular Localization of the MeAnn2 Protein
To explore the localization of the protein expressed by the MeAnn2 gene as an indicator of function, a fusion expression vector of pCAMBIA1300-35S-MeAnn2:GFP was constructed (Figure 2a), and transient expression was mediated by Agrobacterium tumefaciens in tobacco leaves. The empty vector pCAMBIA1300-35S-GFP was used as the control. After 3-5 days of co-culturing, laser confocal microscopy showed that the fluorescence signal corresponding to MeAnn2:GFP in transfected tobacco leaf cells could be observed in the cell membrane and cytoplasm. The fluorescence signal in cells transfected with control vector was observed in the cell membrane, cytoplasm, and nucleus ( Figure 2b).

Analysis of MeAnn2 Expression in Cassava
To further study the function of cassava MeAnn2, we analyzed the expression patterns in specific cassava tissues based on existing cassava transcriptome data. In the transcriptome data, mRNA was sequenced and quantified in 11 organs/tissues. The MeAnn2 gene expression data (FPKM values) were picked out to draw a histogram. The results show that the MeAnn2 gene was expressed in all tested organs/tissues, but the expression levels varied greatly. The expression level of the MeAnn2 gene was highest in the storage roots, followed by the stems, root apical meristems, petioles, and stem apical meristems, and was the lowest in the leaves ( Figure 3).

Analysis of MeAnn2 Expression in Cassava
To further study the function of cassava MeAnn2, we analyzed the expression patterns in specific cassava tissues based on existing cassava transcriptome data. In the transcriptome data, mRNA was sequenced and quantified in 11 organs/tissues. The MeAnn2 gene expression data (FPKM values) were picked out to draw a histogram. The results show that the MeAnn2 gene was expressed in all tested organs/tissues, but the expression levels varied greatly. The expression level of the MeAnn2 gene was highest in the storage roots, followed by the stems, root apical meristems, petioles, and stem apical meristems, and was the lowest in the leaves ( Figure 3).

Transformation of MeAnn2 in Arabidopsis
To reveal the function of MeAnn2 in abiotic stress resistance, the pCambia1300-35S-MeAnn2:GFP vector was transformed into Arabidopsis Columbia-0. Twenty independent transgenic lines were obtained, among which the identity of 18 lines had been previously confirmed by PCR analysis (Figure 4a), and these lines were designated as OE1-18. The transcription levels were determined by fluorescence detection, and three lines (OE1, OE4, and OE6) with the different fluorescence intensity were selected for subsequent function analysis (Figure 4b).

Transformation of MeAnn2 in Arabidopsis
To reveal the function of MeAnn2 in abiotic stress resistance, the pCambia1300-35S-MeAnn2:GFP vector was transformed into Arabidopsis Columbia-0. Twenty independent transgenic lines were obtained, among which the identity of 18 lines had been previously confirmed by PCR analysis (Figure 4a), and these lines were designated as OE1-18. The transcription levels were determined by fluorescence detection, and three lines (OE1, OE4, and OE6) with the different fluorescence intensity were selected for subsequent function analysis ( Figure 4b).

Effect of Different Ca 2+ Concentrations on Arabidopsis Seedlings
In terms of root length, no differences were found between the MeAnn2-overexpressing Arabidopsis lines (OE1, OE4, and OE6) and wild-type Arabidopsis (WT) when they were cultured on 1/2 MS medium containing 20 mM Ca 2+ . The growth of the lateral roots and leaves of WT was inhibited on Ca 2+ -free 1/2 MS medium, while the growth of the lateral roots and leaves of MeAnn2 overexpressing lines was promoted. At a higher Ca 2+ concentration (40 mM), the root length of WT was severely inhibited, with an average root length of about 1.5 cm, while the average root length of transgenic Arabidopsis ranged from 2.9 to 3.5 cm. This was significantly longer than that of WT ( Figure 5).       cultured on 1/2 MS medium containing 20 mM Ca 2+ . The growth of the lateral roots and leaves of WT was inhibited on Ca 2+ -free 1/2 MS medium, while the growth of the lateral roots and leaves of MeAnn2 overexpressing lines was promoted. At a higher Ca 2+ concentration (40 mM), the root length of WT was severely inhibited, with an average root length of about 1.5 cm, while the average root length of transgenic Arabidopsis ranged from 2.9 to 3.5 cm. This was significantly longer than that of WT ( Figure 5).

Effects of NaCl Stress on Growth in MeAnn2-Overexpressing Arabidopsis
The phenotypes of the transgenic Arabidopsis lines overexpressing MeAnn2 (OE1, OE4, and OE6) were similar to those of wild-type Arabidopsis (WT) on 1/2 MS without NaCl medium. On the medium with 100 and 150 mM NaCl added, the upper parts of the plant seedlings grew similarly between the transgenic lines and the WT, while the roots in the transgenic lines had significantly better development with more lateral roots than in the WT. In the medium containing 200 mM NaCl, growth of both the transgenic and WT Arabidopsis was inhibited, and all plants gradually yellowed and died. However, the yellowing speed of the transgenic plants was slightly slower than that of the control (Figure 6a,b).
Gene expression analysis indicated that the expression levels of the stress resistancerelated gene AtRD29B, and the auxin pathway-related genes AtLBD18 and AtIAA4 were

Effects of NaCl Stress on Growth in MeAnn2-Overexpressing Arabidopsis
The phenotypes of the transgenic Arabidopsis lines overexpressing MeAnn2 (OE1, OE4, and OE6) were similar to those of wild-type Arabidopsis (WT) on 1/2 MS without NaCl medium. On the medium with 100 and 150 mM NaCl added, the upper parts of the plant seedlings grew similarly between the transgenic lines and the WT, while the roots in the transgenic lines had significantly better development with more lateral roots than in the WT. In the medium containing 200 mM NaCl, growth of both the transgenic and WT Arabidopsis was inhibited, and all plants gradually yellowed and died. However, the yellowing speed of the transgenic plants was slightly slower than that of the control (Figure 6a,b).
Gene expression analysis indicated that the expression levels of the stress resistancerelated gene AtRD29B, and the auxin pathway-related genes AtLBD18 and AtIAA4 were significantly increased under 150 mM NaCl stress in the transgenic Arabidopsis overexpressing MeAnn2 (Figure 6c). The MDA levels were significantly lower in the transgenic lines compared with the wild type ( Figure 6d).

The Effects of IAA on MeAnn2 Overexpressing Arabidopsis
To study whether MeAnn2 directly responds to indole-3-acetic acid (IAA) signals, the growth of transgenic Arabidopsis lines overexpressing MeAnn2 (OE1, OE4, and OE6) was compared with that of control Arabidopsis when grown on 1/2 MS medium containing various IAA concentrations for 10 days. The results show that on the medium without IAA, the growth was similar between the transgenic and WT Arabidopsis. On the medium with 20 µM IAA, the taproot length of both the transgenic and wild-type Arabidopsis thaliana was inhibited, but the lateral roots were developed and the leaves were enlarged. The phenotypes of MeAnn2-overexpressing and wild-type Arabidopsis showed little difference. On the medium containing 50 and 100 µM IAA, the growth and development of the lateral roots and leaves of wild-type Arabidopsis were inhibited, while the transgenic lines had

The Effects of IAA on MeAnn2 Overexpressing Arabidopsis
To study whether MeAnn2 directly responds to indole-3-acetic acid (IAA) signals, the growth of transgenic Arabidopsis lines overexpressing MeAnn2 (OE1, OE4, and OE6) was compared with that of control Arabidopsis when grown on 1/2 MS medium containing various IAA concentrations for 10 days. The results show that on the medium without IAA, the growth was similar between the transgenic and WT Arabidopsis. On the medium with 20 μM IAA, the taproot length of both the transgenic and wild-type Arabidopsis thaliana was inhibited, but the lateral roots were developed and the leaves were enlarged. The phenotypes of MeAnn2-overexpressing and wild-type Arabidopsis showed little difference. On the medium containing 50 and 100 μM IAA, the growth and development of the lateral roots and leaves of wild-type Arabidopsis were inhibited, while the transgenic lines had abundant lateral root formation and larger leaves (Figure 7). The above results indicate that the MeAnn2 gene responds to IAA and participates in the IAA signaling pathway, and its transformation and expression in Arabidopsis resulted in reduced sensitivity of the transgenic plants to IAA.

Discussion
Plant annexins are Ca 2+ signal receptive proteins and are part of the first step in the signal transduction pathway in addition to playing important roles in plant biological development and responses to adversity and stress [1,5,[38][39][40]. Arabidopsis annexin mediates the calcium and potassium permeability of the root cell plasma membrane, and participates in an increase in [Ca 2+ ]cyt in roots and seedlings in response to hydrogen peroxide [14,16,41]. The AtAnn1 gene responds to drought, low temperature, and salt stresses and mediates the calcium-dependent system defense of Arabidopsis [14,15,42,43]. It is one of the key genes in Arabidopsis resistance to stress. In Arabidopsis, AtAnn8 participates in cell death and disease resistance through negative regulation of RPW8.1 [44]. Capsicum annuum annexin 24 (Ca32) acts as an ion channel to regulate calcium influx in liposomes [45]. Populus euphratica PeAnn1 promotes the accumulation of cadmium in transgenic Arabidopsis [46]. Wheat annexin, TdAnn6, participates in resistance to salt and osmotic stress by regulating antioxidant mechanisms [28]. OsAnn10 is a negative regulator of resistance to osmotic stress in rice (Oryza sativa) [47], while OsAnn1 enhances heat stress tolerance [48]. The cotton (Gossypium hirsutum) annexin GhANN8b participates in the regulating of Ca 2+ and Na + fluxes in plant salt tolerance. [29]. Tobacco seedlings heterologously expressing mustard (Brassica juncea) BjAnn1 and BjAnn2 have enhanced tolerance to salt damage [25,26]. AtAnn1 is the most studied annexin gene in Arabidopsis, and its key role in plant stress resistance has attracted our attention. As a tropical and subtropical plant, cassava has good drought tolerance. The roles of the annexin genes in this plant remain as yet unverified. In this article, we studied the characteristics and functions of the MeAnn2, whose protein has 70.9% amino acid sequence similarity with AtAnn1. In Figure 1a, from the protein sequence alignment, it is easy to determine that the third, fifth and seventh green box indicated that Ca 2+ binding motifs were different in MeAnn2 compared with AtAnn1. In MeAnn2, it is "LHWTL" in the third green box, while in PeAnn1, AtAnn1, BjAnn1, CkAnn1 and StAnn1 the motif is "LLWTL". In the fifth green box, the motif "KAYS" was the same in MeAnn2 and PeAnn1, but different to "KHYN" in AtAnn1, BjAnn1 and CkAnn1. In the seventh green box, the motif "NKRGT" was the same in

Discussion
Plant annexins are Ca 2+ signal receptive proteins and are part of the first step in the signal transduction pathway in addition to playing important roles in plant biological development and responses to adversity and stress [1,5,[38][39][40]. Arabidopsis annexin mediates the calcium and potassium permeability of the root cell plasma membrane, and participates in an increase in [Ca 2+ ] cyt in roots and seedlings in response to hydrogen peroxide [14,16,41]. The AtAnn1 gene responds to drought, low temperature, and salt stresses and mediates the calcium-dependent system defense of Arabidopsis [14,15,42,43]. It is one of the key genes in Arabidopsis resistance to stress. In Arabidopsis, AtAnn8 participates in cell death and disease resistance through negative regulation of RPW8.1 [44]. Capsicum annuum annexin 24 (Ca32) acts as an ion channel to regulate calcium influx in liposomes [45]. Populus euphratica PeAnn1 promotes the accumulation of cadmium in transgenic Arabidopsis [46]. Wheat annexin, TdAnn6, participates in resistance to salt and osmotic stress by regulating antioxidant mechanisms [28]. OsAnn10 is a negative regulator of resistance to osmotic stress in rice (Oryza sativa) [47], while OsAnn1 enhances heat stress tolerance [48]. The cotton (Gossypium hirsutum) annexin GhANN8b participates in the regulating of Ca 2+ and Na + fluxes in plant salt tolerance. [29]. Tobacco seedlings heterologously expressing mustard (Brassica juncea) BjAnn1 and BjAnn2 have enhanced tolerance to salt damage [25,26]. AtAnn1 is the most studied annexin gene in Arabidopsis, and its key role in plant stress resistance has attracted our attention. As a tropical and subtropical plant, cassava has good drought tolerance. The roles of the annexin genes in this plant remain as yet unverified. In this article, we studied the characteristics and functions of the MeAnn2, whose protein has 70.9% amino acid sequence similarity with AtAnn1. In Figure 1a, from the protein sequence alignment, it is easy to determine that the third, fifth and seventh green box indicated that Ca 2+ binding motifs were different in MeAnn2 compared with AtAnn1. In MeAnn2, it is "LHWTL" in the third green box, while in PeAnn1, AtAnn1, BjAnn1, CkAnn1 and StAnn1 the motif is "LLWTL". In the fifth green box, the motif "KAYS" was the same in MeAnn2 and PeAnn1, but different to "KHYN" in AtAnn1, BjAnn1 and CkAnn1. In the seventh green box, the motif "NKRGT" was the same in MeAnn2, PeAnn1 and CkAnn1, but it was "NKTGT" in AtAnn1 and BjAnn1, "KKLGT" in StAnn1, and "AGMGT" in OsAnn1. Whether the changed amino acid residues in the motifs affect the function of MeAnn2 needs further study.
The cellular location of genes is closely related to their functions. Most of the known plant annexins are mainly located in the cytoplasm, and a small number are located in the plasma membrane, inner membrane, and nuclear membrane [5]. For example, in Medicago truncatula, MtAnn1 is located in the nuclear membrane [6]. In the present study, we found that MeAnn2 was localized in the cell membrane, nucleus and cytoplasm. This indicates that annexins of the same family do not function at exactly the same locations.
Under normal conditions, the expression of annexins is dynamic in plant tissues and organs during plant growth processes [9][10][11][12][13]. The analysis of tissue-specific expression levels indicated that the MeAnn2 gene can be detected in all the tested tissues and organs of cassava, but the expression level is highest in the roots, followed by the stems and root tips, and is lowest in the leaves. The expression levels and locations of plant annexins determine their various functions, such as in plant growth, cell secretion, Ca 2+ metabolism, and regulating abiotic stress.
Ca 2+ is the most prevalent second messenger in plants and plays a crucial role in signaling pathways related to many abiotic stresses, such as in the response of plants to salt and in drought tolerance. It has been found that annexin can promote Ca 2+ influx in transgenic Arabidopsis protoplasts, thereby increasing the intracellular Ca 2+ concentration [14][15][16]. In this study, through the overexpression of MeAnn2 in Arabidopsis, we found that the root system of transgenic Arabidopsis was more developed than the wild type under no Ca 2+ (0 mM) and a high Ca 2+ concentration (40 mM), indicating that MeAnn2 can reduce the Ca 2+ sensitivity of plants and thus maintain plant growth.
There are many reports on the functions of annexins in plant stress resistance. For example, AtAnn1 overexpression in transgenic plants enhanced resistance to drought stress, while mutants in which AtAnn1 gene was silenced were more sensitive [43]. In addition, following salt stress, Arabidopsis AtAnn1 and AtAnn4 [21], mustard BjAnn3 and BjAnn7 [24,25], alfalfa MsAnn2 [49], and tobacco NtAnn12 had differential expression [50]. Tobacco seedlings overexpressing BjAnn1 showed enhanced tolerance to salt stress [25]. Our results preliminarily indicated that MeAnn2 overexpression in transgenic Arabidopsis lines led to improved salt stress tolerance compared to wild Arabidopsis. For example, under 150 mM NaCl stress conditions, the transgenic Arabidopsis had a longer root length and more lateral roots than WT. Meanwhile, the expression of the stress resistance-related gene AtRD29B and lateral root development-related auxin pathway genes AtIAA4 and AtLBD18 were meaningfully increased in transgenic Arabidopsis compared to the control plants, while the MDA content in transgenic Arabidopsis was significantly reduced. The AtRD29B is one of the key stress response genes always analyzed in gene function analysis. Its up-regulated expression is usually associated with the tolerance of transgenic plants to salt or other abiotic stresses [35,51,52]. In this study, we found that AtRD29B was significantly upregulated by overexpression MeAnn2 gene in Arabidopsis under salt stress. This result indicates that MeAnn2 as a positive regulator mediates plant tolerance to salt stress.
Previous studies have revealed that the elongation of primary roots, development of lateral roots, and changes in root gravity when plants are exposed to salt stress are very important adaptation strategies for plants to survive high levels of salt. Salt stress affects the lateral root number, root development, and root growth direction of plants by regulating the concentration gradient and redistribution of IAA [53]. The previous research indicated that the high expression of AtLBD18 and AtIAA4 was related to the salt tolerance of transgenic Arabidopsis [35]. In our analysis of transgenic Arabidopsis overexpressing MeAnn2 under salt stress, in which salt tolerance was enhanced compared to WT, it was found that the expression levels of the related genes AtIAA4 and AtLBD18 in the auxin signaling pathway were significantly increased. Therefore, we can speculate that the improved salt tolerance of transgenic Arabidopsis overexpressing MeAnn2 may be related to effects on auxin signaling. We then investigated whether MeAnn2 directly responds to IAA. The results show that, in response to exogenous IAA application, the lateral roots and leaves of transgenic Arabidopsis were more developed than those of the wild type, indicating that IAA had a stronger effect on the induction of lateral root formation in transgenic Arabidopsis. Above all, through the overexpression of the MeAnn2 gene in Arabidopsis, we confirmed the response function of MeAnn2 in Ca 2+ , NaCl and IAA. Additionally, MeAnn2 might improve the salt tolerance of transgenic Arabidopsis by improving the expression of the stress-related RD29B gene and participating in the response of IAA4 mediating the auxin signaling pathway, which would promote the LBD18 and IAA4 regulated root growth under salt stress (Figure 8). However, there are still a lot of deep regulatory mechanisms to be studied.
WT, but it was better compared with Arabidopsis overexpressed AtAnn1 or AtAnn1 mutants. Fortunately, previous studies have shown that AtAnn1 mutants inhibit root growth under salt stress, suggesting that AtAnn1 is involved in root morphogenesis [15]. Therefore, the overexpression of MeAnn2 in our study enhanced root growth under salt stress, which indicated that MeAnn2 in cassava was also involved in root morphogenesis to a certain extent. We analyzed the expression of the AtAnn1 gene in the MeAnn2 overexpression transgenic Arabidopsis OE1, OE4 and OE6 ( Figure S2). To our surprise, we found that endogenous AtAnn1 expression was inhibited, and the inhibition trend was consistent with the overexpression trend of MeAnn2, that is, the more overexpression of MeAnn2, the stronger the inhibition degree. This may be due to the high similarity between MeAnn2 and AtAnn1 protein sequences, and thus the high similarity in their functions, leading to functional redundancy when one of the genes is overexpressed.

Total RNA Extraction and PCR Amplification of the MeAnn2 Gene
One-month-old cassava cultivar SC8 seedlings were used as material for total cassava RNA extraction using the RNAprep Pure Plant Total RNA Extraction Kit (TianGen, Beijing, China) according to the manufacturer's instructions. After detection by 1% gel electrophoresis, the total RNA was first treated to remove the genomic DNA, then transferred In this study, MeAnn2 overexpressed transgenic Arabidopsis was only compared with WT, but it was better compared with Arabidopsis overexpressed AtAnn1 or AtAnn1 mutants. Fortunately, previous studies have shown that AtAnn1 mutants inhibit root growth under salt stress, suggesting that AtAnn1 is involved in root morphogenesis [15]. Therefore, the overexpression of MeAnn2 in our study enhanced root growth under salt stress, which indicated that MeAnn2 in cassava was also involved in root morphogenesis to a certain extent. We analyzed the expression of the AtAnn1 gene in the MeAnn2 overexpression transgenic Arabidopsis OE1, OE4 and OE6 ( Figure S2). To our surprise, we found that endogenous AtAnn1 expression was inhibited, and the inhibition trend was consistent with the overexpression trend of MeAnn2, that is, the more overexpression of MeAnn2, the stronger the inhibition degree. This may be due to the high similarity between MeAnn2 and AtAnn1 protein sequences, and thus the high similarity in their functions, leading to functional redundancy when one of the genes is overexpressed.

Total RNA Extraction and PCR Amplification of the MeAnn2 Gene
One-month-old cassava cultivar SC8 seedlings were used as material for total cassava RNA extraction using the RNAprep Pure Plant Total RNA Extraction Kit (TianGen, Beijing, China) according to the manufacturer's instructions. After detection by 1% gel electrophoresis, the total RNA was first treated to remove the genomic DNA, then transferred into the first strand of cDNA by using the PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, Dalian, China) according to the manufacturer's instructions. The first strand of synthesized cDNA was used as a template, and Me-Tubulin-F/R was used as upstream and downstream primers to detect the quality of reverse transcription.
To clone the cDNA of MeAnn2 from cassava, the primer pairs MeAnn2-F: GAGTC-TACACAGAGCAGAGAGCCT and MeAnn2-R: TATTCCCCGTTGAAACCACA, were designed using Primer Premier 5 outside the coding region, and the high-fidelity enzyme Prime STAR HS (Premix) (TaKaRa, Dalian, China) was used to perform the PCR reaction. The amplified product was recovered, directly ligated to the pMD-19T vector using the TA cloning method according to the operating instruction of pMD™19-T Vector Cloning Kit (TaKaRa, Dalian, China), and then transformed into DH-5α competent cells. The positive clones, identified by bacterial liquid PCR using the BcaBEST Sequencing Primers RV-M and M13-47 in the pMD-19T vector, were sent for DNA sequencing.

MeAnn2 Gene Bioinformatics Analysis and Gene Expression
The physical and chemical properties of the MeAnn2 encoded protein were predicted using the ProParam online software (https://web.expasy.org/protparam/, accessed on 10 April 2021). The sequences of annexin genes in Arabidopsis thaliana, Populus euphratica, Oryza sativa, Solanum tuberosum, Cynanchum komarovii, Brassica juncea, and Manihot esculenta were obtained from the NCBI database (https://blast.ncbi.nlm.nih.gov/Blast.cgi, accessed on 10 April 2021). The amino acid sequences translated from each annexin gene can be found in the Supplementary Materials (Table S1). The multiple protein alignment was performed using Jalview in Phytozome net. The phylogenetic tree was created using MEGA 7.0 software. The protein sequences were first aligned by the ClustalW in the MEGA 7.0 software with the default parameters. Then, the result was used to construct the phylogeny tree, by using the neighbor-joining statistical method, with the Bootstrap method as the test of phylogeny, and the number of Bootstrap replications was 1000, the model was a Poisson model, and the Gaps/Missing data treatment was Pairwise deletion. The expression level of MeAnn2 in different organs and tissues was calculated according to the publicly available cassava RNA sequencing data (Bioproject ID PRJNA324539) [54]. The transcriptome data were download from the NCBI web (https://www.ncbi.nlm.nih. gov/bioproject/PRJNA324539, accessed on 10 April 2021). The FPKM data of MeAnn2 gene expression in different organs and tissues, such as fibrous root, storage root, root apical meristem, stems, stem apical meristem, leaf, petiole, midvein, lateral bud, organized embryogenic structure, and friable embryogenic calli, were selected to analyze and draw the histogram using Excel software.

Subcellular Localization of MeAnn2:GFP
SnapGene software was used to design the primers p1300-MeAnn2-F: ACGCGTCGACATGTCTACCCTTATAG (where the SalI restriction site is underlined) and p1300-MeAnn2-R: CGCGGATCCACTCAATCCTCCTTATG (where the BamHI restriction site is underlined). Using the double-enzyme digestion method, the target gene was inserted into the plant overexpression vector pCAMBIA1300 to construct the pCAM-BIA 1300-MeAnn2:GFP vector. The primer pairs, designed on the pCAMBIA1300, 1300-F: GTTGATACATATGCCCGTCG and 1300-R: CTCGCCCTTGCTCACCAT, were used for the detection and verification of the recombinant vector. The recombinant plasmid and the pCAMBIA1300-GFP empty vector were transformed into competent LBA4404 Agrobacterium cells, which were injected into tobacco leaves. The MeAnn2:GFP protein localization was observed under a laser confocal microscope (OLYMPUS FV1000) 3-5 days after the injection.

Genetic Transformation and Identification of MeAnn2 Overexpression Transgenic Lines
Agrobacterium containing the recombinant pCAMBIA1300-MeAnn2 overexpression vector was transfected into Arabidopsis thaliana Col-0 inflorescences, and the infected mature seeds (T0) were collected and cultured on 1/2 MS (containing 100 mg/L hygromycin) solid medium to screen for positive transformants. The Arabidopsis seedlings of various lines capable of rooting on hygromycin-containing medium were identified by leaf PCR with 2× M5 HiPer Superluminal mix with blue dye (Mei5bio, Beijing, China) and the detection primers 1300-F and 1300-R (same in the Section 4.3). A Luyor-3415RG dualwavelength fluorescent protein excitation light source was used to observe whether the transgenic plants (T1) emitted green fluorescence, and the fluorescence intensity was used to qualitatively detect MeAnn2 expression. Three transgenic lines with strong green fluorescence were selected for subsequent experiments. To evaluate the expression levels of MeAnn2 and AtAnn1 in the transgenic Arabidopsis lines, the total RNA was isolated from the transgenic Arabidopsis lines OE1, OE4, OE6, and wild-type Arabidopsis using the RNAplant Plus reagent (TianGen, Beijing, China) based on the manufacturer's instructions. Additionally, the RNA first erased the genome DNA and then reversed to cDNA as described in 4.1. SYBR ® Premix Ex Taq TM II reagent (TaKaRa, Dalian, China) was used to analyze the expression level of each gene. The Arabidopsis actin gene was employed as an internal reference. The 2 −∆∆Ct method was used to calculate the relative expression level. Each sample of the experiment had three biological replicates, and the calculated results were plotted with Excel. The primer pairs used were as follows: MeAnn2 (QMeAnn2-F: CCCTTGAAGAGGATGTGGCG and QMeAnn2-R: TTCACCTCAGCCCC-CTCGTAT); AtAnn1 (QAtAnn1-F: TGAACCCGGTGAGCGTGATG and QAtAnn1-R: A-GCAGCTGCGTTGATGTCCT); AtActin (QAtActin-F: TTCCTCATGCCATCCTCCGTCT-T and QAtActin-R: CAGCGATACCTGAGAACATAGTG).

Transgenic Arabidopsis Stress Treatment
Under aseptic conditions, three T3 generation transgenic Arabidopsis lines, OE1, OE4, OE6, along with wild-type Col-0, were grown on 1/2 MS medium for about 6 days. Plants that had two true leaves were then moved onto 1/2 MS medium square plates to undergo different stress treatments and were cultured at 22 • C in 12 h light/12 h dark conditions. Ten days later, the phenotypes and root length were observed. The Ca 2+ treatments were performed by the inclusion of 0, 20, and 40 mM CaCl 2 into 1/2 MS medium. The NaCl treatments were performed by inclusion of 0, 100, 150, and 200 mM NaCl into the 1/2 MS medium. The IAA treatments were performed by inclusion of 0, 20, 50 and 100 µM IAA into the 1/2 MS medium. In this experiment, three biological replicates were performed.

Quantitative RT-PCR Analysis of Stress Resistance-Related Genes
Total RNA was isolated from the treated transgenic Arabidopsis lines OE1, OE4, OE6, and wild-type Arabidopsis using the RNAplant Plus reagent (TianGen, Beijing, China) based on the manufacturer's instructions. TaKaRa SYBR ® Premix Ex Taq TM II reagent (TaKaRa, Dalian, China) was used to analyze the expression level of related stress resistance genes in Arabidopsis. The Arabidopsis actin gene was employed as an internal reference. The 2 −∆∆Ct method was used to calculate the relative expression level. Each sample of the experiment had three biological replicates, and the calculated results were plotted with Excel. The detected genes and their specific primers are as follows: AtIAA4 (QAtIAA4-F: GCAGAGGAGGCAATGAGTAGTG and QAtIAA4-R: GAG-CATCCAGTCACCATCTTTG), AtRD29B (QAtRD29B-F: ATTCACCATCCAG AAGAA-GAGCATC and QAtRD29B-R: ACTTCTGGGTCTTGCTCGTCA), AtLBD18 (QAtLBD18-F: GTCGCTCACATCTTTGCTCTTC and QAtLBD18-R: GTCGCTCACATC TTTGCTCTTC), AtActin (QAtActin-F: TTCCTCATGCCATCCTCCGTCTT and QAtActin-R: CAGCGATAC-CTGAGAACATAGTG) [35].

Statistical Analysis
The data are presented as the mean ± SD. The significance of the data was evaluated by one-way ANOVA with Duncan's multiple range test. p ≤ 0.05 and p ≤ 0.01 were set as the significance levels.

Conclusions
In summary, the function of MeAnn2, the cassava homolog of AtAnn1, was studied in the context of stress tolerance. The overexpression of MeAnn2 reduced the Ca 2+ , NaCl, and IAA sensitivity of transgenic Arabidopsis. The enhanced salt tolerance of transgenic Arabidopsis was accompanied by the upregulation of the stress tolerance-related gene AtRD29B and the auxin signaling pathway-related genes AtIAA4 and AtLBD18, as well as a reduction in MDA content. These results suggest that MeAnn2 might improve the salt tolerance of these plants through the auxin signaling pathway.