Wenjun Ma · Yao Xiao · Yuan Li · Pan Hu ·Zhi Wang · Guijuan Yang · Junhui Wang

Abstract Catalpa fargesii is an important economic tree species used for furniture and timber production because of its high density and hardness. Its survival and growth are severely affected and primarily limited by drought stress.Thus, to better understand the mechanism of drought resistance in C. fargesii, we used qRT-PCR to reveal significantly different expression of three plasma membrane intrinsic protein genes: CfPIP1- 1, CfPIP1- 2 and CfPIP1- 4. We then cloned their full-length cDNA sequences and characterized the encoded proteins. We analyzed the genes phylogenetically and predicted conserved motifs, domains, and secondary and tertiary structures. To verify the function of the CfPIP1 genes further, we ectopically expressed CfPIP1 transgenes in Arabidopsis thaliana. The results showed that CfPIP1-1, CfPIP1-2 and CfPIP1-4 had several characteristics of aquaporins. The transgenic plants grew better than the WT plants did under drought stress, and overexpression of the CfPIP1 genes increased the plant water content and resistance to drought. Thus, CfPIP1-1, CfPIP1-2 and CfPIP1-4 of C. fargesii play key roles in regulating the intracellular and extracellular water balance and in mediating the plant response to drought.

Keywords Catalpa fargesii · Drought tolerance · PIP1 ·Aquaporin · Heterologous transformation

Introduction

Water is one of the most important factors for plant growth and development. Drought promotes changes in forest structure and reduces tree growth in forests (Bunker and Carson 2005; Dulamsuren et al. 2010). Efforts to protect natural forest resources have increased, and plantation areas have increased worldwide (Payn et al. 2015). To meet the demand for commercial timber, some plantations are also established in semiarid areas, consequently affecting survival and timber stocks and severely hindering agricultural and forestry production. Therefore, understanding the drought resistance mechanism of forest species and improving poor tree growth caused by drought are important. Plants can adapt to or resist drought in various ways; for example, stomatal closure can be induced to reduce both transpiration and stomatal conductivity to prevent excessive water loss when water shortage contents reach a certain level (Blake and Li 2003; Gindaba et al. 2004). However, stomatal closure is only a short-term mechanism of plants to adapt to soil water scarcity (Pereira and Chaves 1993). The activities of various antioxidant enzymes in plants can be regulated to remove excess free radicals and maintain their dynamic balance (Jiang and Zhang 2001; Parvin et al. 2012). With the development of molecular biology techniques, aquaporins(AQPs), which promote water transport across membranes,have been found to exist widely in plants and participate in the regulation of osmotic balance. The expression of AQP genes is upregulated in response to abiotic stress, and studies have shown that root growth of Arabidopsis thaliana overexpressing PIP from banana increased in nonstressed and stressed conditions (Xu et al. 2014). When A. thaliana plants were transformed with TaNIP were able to regulate cell water potential by altering the concentrations of mineral elements in their cells (Gao et al. 2010). Therefore,AQPs constitute a key factor in plant water-use and drought response. AQPs of important economic plant species, such as Norway spruce (Hakman and Oliviusson 2002), banana(Xu et al. 2014), maize (Gaspar et al. 2003) and bamboo(Sun et al. 2017), have been cloned and shown to be involved in the tolerance to drought stress.

Species of Catalpa are semievergreen or deciduous trees or shrubs and mainly distributed in North America, the West Indies and East Asia. They differ in their growth adaptability and thus in their economic and ecological value. Most can be used horticulturally as garden and street trees for their large heart-shaped leaves and showy flowers. In addition, the extracts of C. ovata bark and leaves can be used as Chinese medicines (Wang 1990). The straight stems of C. bungei and C. fargesii from China produce of dense, hard wood wood with high bending strength valued for building, appliances and furniture (Ma et al. 2013). As the demand for highgrade wood increases, improving the variable growth and wood properties of C. bungei has become a primary goals for breeders. Toward this goal, Xiao et al. (2019) selected the best clones of C. bungei for high volume and stable yield.Among interspecific and intraspecific Catalpa hybrids that have also been created (Jia et al. 2010), some had obvious heterosis at an early stage, which greatly accelerated the genetic improvement of C. bungei. Although selective breeding has improved the wood yield of C. bungei, however, most plantations with this species were established in semiarid mountainous and hilly areas, which has severely limited their survival and growth because C. bungei has poor drought resistance. Thus, breeding or engineering fast-growing, drought-resistant varieties is needed to study the drought resistance mechanism of C. bungei. The drought response of potted seedlings is known to depend on adequate N nutrition(Shi et al. 2017), but the applicability of this result to large plantation areas is not known. Unlike C. bungei distributed in the Yellow River and Yangtze River regions, C. fargesii is distributed in northwestern China, which consists of both arid and semiarid regions. Therefore, C. fargesii may have a stronger drought resistance than C. bungei and may be more suitable for studying the drought resistance mechanism of Catalpa species. By using gene chip technology,we identified three putative PIP1 genes from C. fargesii as candidate factors that respond early to drought stress. To further conf irm the function of these genes, we conf irmed three full-length cDNAs encoding CfPIPs, analyzed the sequences in detail, then assessed their functions and effects under drought stress using overexpression experiments in A. thaliana. The results increase our understanding of the mechanisms of the Catalpa drought stress response and lay a strong foundation for drought resistance-based genetic engineering and cross breeding of Catalpa species.

Materials and methods

Plant material and stress treatment

Seeds of C. fargesii were germinated in a phytotron (75-80%relative humidity; 12,000 lx) with 16 h days at 26 °C/8 h nights at 23 °C. Seedlings were then transplanted into soil matrix (perlite:Vermiculite:peat to 1:1:8) in plastic containers and grown in a greenhouse (45-65% relative humidity;23 °C average temperature).

When seedlings were approximately 15 cm tall, several seedlings with a relatively consistent shape, size and growth were selected, and their roots were washed with tap water three times to prevent damage to the root system during the cleaning process. The seedlings were then placed in a basin covered with a sheet of foam, after which the seedlings were secured with a sponge to a hydroponic system. During hydroponic growth, the root systems were protected from light, and the oxygen supply was sufficient. The water was changed every 2 days, and three treatments were applied 1 week later to 15 seedlings each: (1) water culture control(CK) group, (2) 200 g L ?1 polyethylene glycol (PEG) 6000,(3) 200 mg L ?1 abscisic acid (ABA). Leaf and root samples from three groups of five seedlings for each treatment were collected after 1 day of treatment (3 biological replicates per organ sample per treatment). The samples were frozen immediately in liquid N and then placed in a ? 80 °C cryogenic refrigerator.

Cloning of CfPIP1 genes

Total RNA was extracted from the leaves of C. fargesii via an RNA extraction kit (TIANGEN, Beijing, China).First-strand cDNA was synthesized from 2 μg of DNase I-treated RNA using oligo(dT)-18 adaptor primers and Moloney murine leukemia virus (M-MLV) reverse transcriptase (TaKaRa). We used the full RACE Core Set Kit version 2.0 (TaKaRa) for rapid amplification of cDNA ends (RACE) to obtain the full-length CfPIP s. The primers for the full-length CfPIP sequences (Table S1) were designed based on the transcriptome results. The detailed full-length CDS sequence of CfPIPs was showed in Table S5.

Phylogenetic analysis

The Translate tool of the ExPASy (SIB Bioinformatics Resource Portal; https://web.expas y.org/trans late/) was used to translate the coding sequences of genes and to infer the amino acid sequences. Aquaporin (AQP) sequences from a variety of model plant species annotated in the UniProt database (https://www.unipr ot.org/unipr ot/) were downloaded,and selected protein information is shown in Table S2.

Phylogenetic trees were constructed with MEGA X software (Kumar et al. 2018) using the neighbor-joining method(Saitou and Nei 1987) and 1000 bootstrap replications, the Poisson model, uniform rates among sites, and gaps/missing data treatment with pairwise deletion.

Protein sequence characteristics

The basic characteristics of the protein sequence were described using the ExPASy ProtParam tool (https://web.expas y.org/protp aram/), and protein transmembrane domains were predicted by TMHMM Server v. 2.0 (http://www.cbs.dtu.dk/servi ces/TMHMM/). The Ar/R motifs of the tested proteins were predicted by comparison against standard AQP proteins (Table S3). The secondary and tertiary structures of the proteins were predicted using the SOPMA method(http://pbil.ibcp.fr/) and SWISS-MODEL (https ://swiss model.expas y.org/), respectively.

Expression of CfPIP genes

Total RNA was extracted from plant tissues using an RNA extraction kit (TIANGEN). Total RNA was converted into cDNA using a FastQuant RT Kit (with gDNAase) (TIANGEN). Optimal primers (Table S4) with high efficiency and specificity were selected via agarose gel electrophoresis detection. UBC was used as a reference gene for normalizing transcript levels of the target genes and verified as constitutively expressed and suitable for use as internal controls.The expression of CfPIP1- 1, CfPIP1- 2 and CfPIP1- 4 in C.fargesii roots and leaves after various treatments was quantified using the SYBR Premix Ex Taq Kit (TaKaRa, Japan)and Roche LightCycler 96 for qRT-PCR according to the manufacturers’ protocols. Relative expression of the genes was calculated using the 2?ΔΔCtformula.

Genetic transformation and transgenic plants

The seeds of wild-type (WT) A. thaliana (ecotype Columbia) were vernalized at 4 °C for 3 days, then disinfected with 75% alcohol on the benchtop and seeded into 1/2-strength Murashige and Skoog (MS) agar. The seedlings at the 4-leaf stage were transplanted into small pots (1:1:8 perlite-vermiculite-peat), then transferred to an artificial-climate incubator(20 °C temperature; 16 h/day illumination). An appropriate fusion construct plasmid (pCAMBIA1301) was inserted into Agrobacterium tumefaciens (GV3101), the floral dip method was used to insert the construct into seedlings of A. thaliana(Clough and Bent 1998). Seeds produced by the transformed plants were collected and sown on 1/2-strength MS agar supplemented with 600 mg L ?1 hygromycin. Individual plants (T 0 generation) that grew normally were selected for cultivation,and T3-generation plants were used for gene function verification experiments.

Root growth of transgenic A. thaliana plants

The seeds of the WT and transgenic A. thaliana plants were washed with sterile water and surface-sterilized with 75% alcohol and 1% sodium hypochlorite. The seeds were subsequently sown on 1/2-strength MS agar (control) or 1/2-strength MS agar supplemented with 6% PEG6000 (treatment). The phenotypes were photographed, and root lengths measured after 7 days.

Phenotype characteristics and moisture content of transgenic plants

The WT and transgenic plants were transplanted into pots that contained vermiculite and 1/4-strength MS nutrient solution.After 7 days, plants were treated to one of three treatments: (1)CK, which involved watering with 1/4-strength MS solution every 3 days; (2) low drought (LS), which involved watering with 1/4-strength MS solution every 10 days; (3) strong drought (SS) treatment, which involved no watering. After 27 days, plants were photographed, plant height and basal leaf height were measured, and leaves were counted; at least three biological replicates were evaluated per trait per treatment. The plants were then rewatered, and traits were measured again 7 days later. Fresh mass (FW) and dry mass (DW, after ovendrying 76 °C for 48 h) ofleaves were determined to calculate water content as:

where RMC is the relative water content and AMC is the absolute water content.

Results

Phylogenetic analysis and gene characteristics

The full-length cDNAs of the CfPIP1- 1, CfPIP1- 2 and CfPIP1- 4 genes were 864 bp, 861 bp and 858 bp long,respectively. In the phylogenetic tree based on the selected plant AQPs from the UniProt database, the three predicted AQPs, classified into the PIP1 and PIP2 subclasses of PIP lineages, clustered with the PIP1 lineage (Fig. 1). The genes were therefore named CfPIP1- 1, CfPIP1- 2 and CfPIP1- 4.

When the CfPIP1- 1, CfPIP1- 2 and CfPIP1- 4 sequences were aligned with other PIP1 subclass proteins, the amino acid sequences encoded by the three target genes were highly similar to those of the PIPs in the model plant species (Fig. 2). Sequence similarity ranged from 85%(CfPIP1- 1 vs. VfPIP1- 1 and CfPIP1- 2 vs. SIRAMP) to 91% (CfPIP1-2 vs. OsPIP1-2). Although C. fargesii is a tree species and is distantly evolutionarily related to herbaceous plant species, the AQP sequences were highly similar; hence, this finding further conf irmed that AQPs have been highly conserved during evolution.

The results of the subcellular prediction (Table 1)showed that the tested genes had the highest scores for the plasma membrane. The scores for other subcellular structures were all lower than 1. Many studies have conf irmed that PIPs are intrinsic plasma membrane proteins;therefore, the three putative aquaporin genes of C. fargesii belong to the PIP family.

Fig. 1 Phylogenetic tree of AQPs from different plants

Fig. 2 Sequences alignment of CfPIP1s with other PIP1s in model plants

Protein sequence characteristics

CfPIP1- 1, CfPIP1- 2 and CfPIP1- 4 encode 287, 286, and 285 amino acids, respectively (Table 2). The numbers of amino acids were highly consistent with those of PIPs encoded by other model plant species (286-289). The variation in the isoelectric point of these proteins was small, ranging from 8.97 to 9.31, their GRAVY range was 0.350 ? 0.373, and all of these proteins were hydrophobic. These results indicated that the characteristics of CfPIP1-1, CfPIP1-2 and CfPIP1-4 were highly similar to those of PIP proteins in model plant species.

The amino acid sequence alignment results showed that the selected PIPs have six transmembrane domains with completely consistent sequences (Fig. 3). Each sequence has a conserved NPA motifin LoopB and LoopE. The NPA motifis a highly conserved motif sequence in the AQP family and is directly involved in the binding and selectivity of water molecules. In addition, the sequences have two ar/R motifs (F and H) on TM2 and TM5, and another two ar/R motifs (T and R) are present on LoopE. Ar/R motifs play an important role in the selectivity and transmembrane transport of substrates, and the combination of motifs determines the type of substrate transported. This sequence information further conf irmed that CfPIP1-1, CfPIP1-2 and CfPIP1-4 belong to the AQP family and may function in transporting water in C. fargesii cells through the membrane system.

The secondary structures of CfPIP1-1, CfPIP1-2 and CfPIP1-4 were also highly similar to those of PIPs from several model plant species. The secondary structures were dominated by α-helices (26.92-33.45%) and random curls(43.60-50.70%), while extended strands accounted for 18.06-21.68%, and β-turns accounted for only 2.11-4.84%(Fig. 4). We predicted the three-dimensional structure of CfPIP1-1, CfPIP1-2 and CfPIP1-4 (Fig. 5), and the three proteins were up to 70% similar to the spinach AQP template 4jc6.2. D (Frick et al. 2013). The three proteins were homologous tetramers and included many random curls and α-helices, and most of the sequences consisted of highly hydrophobic structures. In addition, the GMQE value was between 0.75 and 0.77 (this value can range from 0 to 1; a larger value indicates a more accurate prediction), and the QMEA values of CfPIP1-1, CfPIP1-2 and CfPIP1-4 were? 3.15, ? 2.89 and ? 3.40, respectively, all of which were greater than ? 4. Together, these results showed that the predicted results were highly reliable and that the tested proteins had the tertiary structure characteristics of AQPs.

Expression of CfPIP genes in different organs under various treatments

The expression patterns of CfPIP1- 1, CfPIP1- 2, and CfPIP1- 4 in the leaves of C. fargesii differed under PEG and ABA treatments (Fig. 6 a). Compared with that under the control treatment, the expression of the CfPIP1- 1 gene in the leaves under the ABA treatment was downregulated, but the expression of CfPIP1- 2 under the PEG and ABA treatments was significantly greater than under the control treatment (by 1.4- and 1.9-fold, respectively). In addition, the expression level of the CfPIP1- 4 gene under the PEG and ABA treatments was similar to that of the CfPIP1- 1 gene.

PIP1-2 Zm 0 9.43 00 0.07 0 0.03 0.03 0.26 0.18 VfPIP1-1 0 9.47 00 0.08 0 0.03 0.01 0.24 0.18 MP SlRA 0 9.52 0 00.08 00.02 0 0.21 0.17 OsPIP1-2 0 9.4 00 0.09 0 0.03 0.02 0.27 0.19 OsPIP1-1 0 9.43 00 0.07 0 0.03 0.04 0.25 0.18-4 ays 0 9.41 AtPIP1 00 0.07 0 0.03 0.03 0.27 0.19 m Zea m a; Z f Vicia fab m diff erent plants AtPIP1-3 09.39 00 0.08 0 0.03 0.03 0.28 0.2 AtPIP1-1 0 9.47 00 0.08 0 0.03 0.01 0.24 0.18 m lycopersicum; V orin proteins fro CfPIP1-4 lanu l So 00 0.07 0 0.03 09.44 0.02 0.25 0.18 s Oryza sativa; S of subcellular localization of aquap CfPIP1-2 0 9.45 00 0.07 0 0.03 0.02 0.24 0.18 liana; O CfPIP1-1 00 0.07 0 0.03 0.03 opsis tha 0.27 0.18 t Arabid e em Prediction bran0 9.42 Table 1 Protein Nuclear Plasma m Extracellular plasmic ondrial lpa fargesii; A CytochMitoplasm reticulum Endo Peroxisomal i roplast Golg Chloolar Vacu Cf Cata

The expression levels of the CfPIP1- 1 and CfPIP1- 2 genes in the roots under the ABA treatment were significantly greater than those under the control treatment (by approximately 1.6-fold). The expression level of CfPIP1-1 under the PEG treatment was significantly greater than that under the control treatment (by 2.2-fold). Moreover, the expression level of the CfPIP1- 4 gene in the roots was similar to that in the leaves. Together, these results implied that these genes may be involved in the regulation of resistance to stress, especially drought stress, but that they are governed by different mechanisms.

Root growth of transgenic A. thaliana plants under PEG treatment

The root growth of the WT and transgenic A. thaliana plants grown on 1/2-strength MS medium supplemented with 6%PEG was significantly different from that of the plants grown on 1/2-strength MS control medium (Fig. 7). The roots of both the WT and transgenic plant grew well in 1/2-strength MS medium, and there was no significant difference in root length between the WT and transgenic plants (Fig. 7 b). In the 6% PEG treatment, roots of the WT plants ceased growing, those of the transgenic plants grew slower; however,the roots of the transgenic plants were significantly longer than those of the WT plants. The root lengths of the plants containing inverted CfPIP1- 1, CfPIP1- 2 and CfPIP1- 4 were 69.24%, 75.63% and 49.46% longer than those of WT plants, respectively (Fig. 7 b). These results indicated that these three genes may activate or directly participate in the osmotic regulatory mechanism of plants, improving water transport and use and enhancing drought resistance.

Phenotypic characteristics of transgenic A. thaliana plants under drought stress

The phenotypes of the WT and transgenic plants did not significantly differ under the CK and LS conditions. However,the WT plants did not grow normally under the SS conditions. Although growth of plants overexpressing the CfPIP genes was inhibited, the plants survived (Fig. 8 a).

Under the LS conditions, the plants transformed with CfPIP1- 1 and CfPIP1- 2 were significantly taller (27.93%and 43.92%, respectively) than the WT plants. The basal leaf height of plants overexpressing CfPIP- 1 was significantly greater (47.95%) than that of the WT plants, and the overexpressing plants also had significantly more leaves than on the WT plants. Thus, CfPIP1- 1, CfPIP1- 2 and CfPIP1-4 may improve water absorption, water-use efficiency and drought resistance.

Under the SS conditions, plants overexpressing CfPIP1-1, CfPIP1- 2 and CfPIP1- 4 were significantly taller (4.8,4.1 and 3.5 times, respectively) than the WT plants, and the basal leaf heights were significantly greater (76.40%,78.61% and 65.63%, respectively) than those of the WT plants. There were no significant differences in growthtraits between the WT plants and the transgenic plants after rewatering (Fig. 9 d-f). These results indicated that the function of these three genes is enhanced under drought stress and that the expression of these genes improved the drought resistance and growth plants under drought stress.

Table 2 Basic information on PIP proteins from different plant species

Fig. 3 Analysis of the PIP1 protein domains. TM transmembrane domain

Water content in A. thaliana overexpressing CfPIPs

Fig. 4 Analysis of the PIP1 protein secondary structures

Fig. 5 Prediction of PIP1 protein tertiary structures

Fig. 7 Effect of 6% PEG treatment on root growth of transgenic and wild-type A. thaliana plants. Different lowercase letters indicate a significant difference in growth between the plant lines

Fig. 8 Growth phenotypes of transgenic and wild-type A. thaliana plants in response to drought stress. WT wild type; CK control check;LS light drought stress; SS strong drought stress

To further verify that the functions of the CfPIP genes were related to water transport, we measured and compared the water content of the WT plants and the transgenic plants under the different treatments (Fig. 10). The results revealed no significant difference between the WT plants and the transgenic plants in terms of RMC and AMC under the LS treatment. However, the RMC of plants overexpressing CfPIP1- 1, CfPIP1- 2, and CfPIP1- 4 was 26.87%, 26.22%, and 20.92% greater than that of the WT plants under the SS treatment, and the AMC was 4.4, 4.2 and 2.8 times that of WT plants, respectively. There was no significant difference in water content after rewatering.These results suggest that the CfPIP genes improve the ability of the plants to absorb and transport water and thus enhance plant resistance to drought stress.

Discussion

Drought has become the main abiotic stress limiting crop and forest resources (Dulamsuren et al. 2010; Beyene et al.2016; Sserumaga et al. 2018; Zhao et al. 2019). Thus,improving the drought resistance of plants and breeding drought-resistant plant varieties has become a focus for plant scientists. Many genes are involved in the drought response of plants (Zhao et al. 2019; Wang et al. 2016). Previous studies have conf irmed the close relationship between AQPs and drought resistance in plants (Mahdieh et al. 2008; Xu et al.2014; Li et al. 2015). AQPs constitute a superfamily, and PIPs are the largest subfamily and are highly concentrated in the plasma membrane. The PIP subfamily includes two subclasses: PIP1 and PIP2. We found that the three CfPIP proteins were categorized within the PIP1 subfamily according to our phylogenetic tree.

AQPs, highly conserved integrins, function on the basis of their structure. Therefore, analyzing the characteristics and structure of CfPIPs would be helpful to understand protein function more accurately. Our phylogenetic results showed that all three CfPIPs belong to the PIP1 subfamily.CfPIP1-1, CfPIP1-2 and CfPIP1-4 have six transmembrane domains, one ar/R motif on TM2 and TM5, and two ar/R motifs on LoopE. In addition, they have a conserved NPA motif on LoopB and LoopE. These conserved motifs play key roles in the transport of water molecules; for example,the Van der Waals force of the proline in the NPA motif folds LoopB and LoopE into an “hourglass model” hydrophilic channel (Jung et al. 1994), and the ar/R aromatic amino acid filters water molecules by hydrophobicity and molecular size, thereby maintaining specificity to transport water across the plasma membrane (Wang et al. 2005; Hub and de Groot 2008). Previous studies have shown that the transmembrane domains of AQPs exist as α-helices and form tetramers embedded in the lipid bilayer (Sui et al. 2001;Wang et al. 2005; Gomes et al. 2009). The prediction of the secondary and tertiary structures of the three PIP proteins in this study also indicated a tetrameric structure with multiple helixes, which was consistent with the characteristics of AQPs. These data suggest that CfPIP1- 1, CfPIP1- 2 and CfPIP1- 4 may function in the transport of water molecules.

Fig. 9 Effect of drought stress on growth of transgenic and wild-type A. thaliana plants. WT wild type; CK control check; LS light drought stress; SS strong drought stress

Fig. 10 Moisture content in transgenic and wild-type A. thaliana plants a, b before rehydration and c, d under strong drought stress

The large scale of the family also leads to differences in expression patterns and functions, and it is generally accepted that PIP2 subfamily proteins have a significantly stronger ability to regulate transmembrane water transport than do PIP1 subfamily proteins (Chaumont et al. 2000;Yaneffet al. 2015). For instance, ?urbanovski et al. (2013)reported that wild strawberry (Fragaria vesca) FvPIP1 s and FvPIP2 s exhibited different tissue-specific expression patterns under drought stress. When we quantified the relative expression of the three CfPIP1 genes, they exhibited different expression patterns, and the expression of CfPIP1- 1 and CfPIP1- 2 was upregulated under PEG and ABA treatments.The above results showed that the functions of the three genes are closely related to drought stress. These findings are similar to those of Aroca et al. (2006), who determined that the Phaseolus vulgaris PvPIP s, encoding AQPs, were highly expressed under drought and ABA treatments and that some PIP1 s may encode transporters of small solutes or gas molecules in the membrane of mesophyll cells or act as water channels in root cells via modifications or interactions with other AQPs (Kaldenhoffand Fischer 2006). CfPIP1-1, CfPIP1- 2 and CfPIP1- 4 are expected to function more efficiently in leaves than in roots.

The results of the heterogeneous transformation conf irmed that the functions of the CfPIP1- 1, CfPIP1- 2 and CfPIP1- 4 genes were closely related to drought resistance.The phenotype of A. thaliana overexpressing these three genes was not significantly different from that of the WT plants under normal water and moderate drought conditions;however, the transgenic plants survived and grew under strong drought stress, but the WT plants did not. Furthermore, the water content of the transgenic plants was significantly greater than in the WT plants under drought stress.These results indicated that these three genes of C. fargesii regulate both the use and absorption of water. Consistent with our results, A. thaliana overexpressing SpPIP1 (Chen et al. 2018) and CfPIP2- 1 (Jang et al. 2007) had greater survival and better phenotypes under drought conditions than did WT plants. However, the upregulated expression of all PIP genes does not improve plant stress resistance. Aharon et al. (2003) reported that although overexpressing the PIP1 gene in tobacco promoted growth in ideal environments, it reduced the drought resistance and salt tolerance of plants.Similarly, CsPIP1- 1 and GsPIP2- 1 in transgenic A. thaliana may negatively affect stress tolerance by regulating water potential (Jang et al. 2007; Wang et al. 2014), suggesting that there is a high diversity of PIP functions in different species and environmental conditions.

Conclusions

The CfPIP1-1, CfPIP1-2 and CfPIP1-4 protein sequences,conserved motifs, transmembrane domains, and secondary and tertiary structures demonstrated that these proteins had characteristics consistent with those of AQPs. The results of this study revealed that these three genes and the encoded proteins promoted transmembrane transport of external water, improved the use rate of internal water and played a key role in regulating the intracellular and extracellular water balance when the plants were drought-stressed. The results of this study will aid in the cultivation of highly drought-resistant C. fargesii varieties. However, the functions of the CfPIP1- 1, CfPIP1- 2 and CfPIP1- 4 genes and the encoded proteins should be further verified.