Introduction

Plants are usually grown by the cell elongation through water absorption besides the cell division. Water transport comprises a combination of apoplastic (cell wall and extracellular space) and symplastic (plasmodesmata and transmembrane) pathways [1]. Transcellular movement of water is controlled by aquaporins (AQPs), which are present in all living organisms and their distribution greatly varies with cellular complexity. AQPs, belonging to the ancient major intrinsic protein (MIPs) family, are known to play a major role in maintaining water and hydraulic conductivity balance in the plant system [2]. Recent studies show that AQPs constitute a large gene family in plants. For instance, 41 AQP homologues were identified in potato [1], 34 in O. sativa [3], 65 in Populus trichocarpa [4], 53 in Chinese cabbage [5], and 41 in common bean [6], respectively. Based on the amino acid sequences, membrane localization and substrate specificity, AQPs are classified into five major subfamilies, including plasma membrane intrinsic proteins (PIPs), tonoplast intrinsic proteins (TIPs), nodulin 26-like intrinsic proteins (NIPs), small basic intrinsic proteins (SIPs) and uncharacterized X intrinsic proteins (XIPs). Though XIPs have been shown to exist in different species like potato, common bean, poplar, and Chinese cabbage [1, 4–6], they have not yet been found in monocots, and certain plant species in dicots, such as Arabidopsis thaliana [7].

All AQPs share a conserved structure with an α-helical bundle forming six transmembrane (TM) helices (TM1–TM6), which are connected by five interhelical loops (loop A–loop E), and both the amino and the carboxyl termini were located on the cytoplasmic side of the membrane. Two highly conserved Asn-Pro-Ala (NPA) motifs in the pore regions of the channel, located in loop B and loop E, have the function of transport selectivity [8]. In addition, another pore constriction known as aromatic/arginine (ar/R) selectivity filter, is formed by four residues from TM2, TM5, and loop E at the extracellular side of the pore [9]. Differences in ar/R selectivity filter result in different pore sizes, which thought to transport a great diversity of solutes [10, 11]. AQPs are known to increase the permeability of membranes to water and other small molecules, such as silicon, boron, arsenic and lactic acid [9, 12]. Recent studies suggest that plant AQPs play a central role in numerous physiological processes, such as plant-soil water relations, biotic interaction, micro-nutrient uptake, seed germination, stomatal regulation, and leaf water transport, embolism repair and various abiotic stresses [2].

Moso bamboo (Phyllostachys edulis) is one of the most fast-growing plants in the world and represents a significant ecological, economic and cultural value in China. It can grow up to more than 100 cm in only 24 h and reach to its maximum height of about 24 m during one and a half months at growth spurt in spring [13]. Hydraulic conductivity is regulated to maintain the morphological changes of the cell. Both cell division and cell elongation in individual internodes of bamboo dominate the development of culm at different growing stages [14]. Apparently, water supply is essential for maintaining normal physiological function during the rapid growth of bamboo. Previous researches suggest that AQPs play an important role in cell elongation and may be a vital part in the fast growth of bamboo shoot via the coregulation of auxin and gibberellin [15, 16]. Although large numbers of AQPs in monocots have been extensively studied through genome-wide analysis [3, 17–19], our knowledge of AQPs in bamboo is still limited, and the expression profile of AQP genes in different tissues and those under stress conditions are unknown in bamboo.

In this study, 26 AQP genes (PeAQPs) were identified from Ph. edulis genome by homologous analysis, which encoded different AQPs classified into four groups (PIPs, TIPs, NIPs, and SIPs) based on the phylogenetic analysis. We performed a considerable analysis of the PeAQPs, such as gene structure, conserved motif and scaffold location. We also analyzed the transcriptional levels of PeAQPs in different tissues, including root, stem, leaf and sheath. Additionally, the expression profiles of PeAQPs in root and leaf under drought, water and NaCl stresses were also investigated. These results provided helpful information for further studies on function of PeAQPs, which would facilitate to select genes for bamboo improvement through genetic engineering in future.

Materials and methods

Identification and structure analysis of AQP genes in moso bamboo

Due to the successful implementation of Ph. edulis genome sequencing, the CDS and putative protein sequences of Ph. edulis have become available [13]. BLASTN, BLASTX and BLASTP were performed to identify putative AQP genes in Ph. edulis using the bamboo genome database (BambooGDB, http://www.bamboogdb.org/) [20] with the AQPs of Z. mays, O. sativa and P. trichocarpa as query sequences, the aligned result was filtered with a cut-off e-value of 10−10 [4, 7, 17]. From each homology search, the top four–six sequences were selected as candidate genes. The gene structure was analyzed using the bamboo gene models (BambooGDB) or experimentally determined EST sequences [21]. Gene structure display server (GSDS; http://gsds.cbi.pku.edu.cn) was used to show the number and the position of exons [22].

Protein characteristics, scaffold location, and phylogenetic analysis

The AQP sequences of Ph. edulis were downloaded from BambooGDB. Molecular weight (MW) and isoelectric point (pI) of the AQP proteins in Ph. edulis were analyzed using ProtParam (http://web.expasy.org/protparam/). Two pore regions of the AQPs were manually checked, and proteins without these two motifs were discarded from the analysis. Genomic locations of bamboo AQP genes were searched in BambooGDB.

In order to investigate the phylogenetic relationship and the presence of XIPs AQPs in bamboo, the clustering analysis of AQPs including three monocots (Ph. edulis, Z. mays and O. sativa) and one dicot (P. trichocarpa) were constructed in MEGA6.0 using the following parameters: Neighbor-Joining tree method, complete deletion, and bootstrap with 1000 replicates [23].

In silico prediction of subcellular localization and transmembrane domains

The subcellular localization of AQPs was predicted using the online tool WoLF PSORT algorithms (http://wolfpsort.seq.cbrc.jp) and Plant-mPLoc algorithms with default parameters [24, 25]. TM helices were predicted with the tool of TMHMM Server v.2.0 (http://www.cbs.dtu.dk/services/TMHMM/) [26].

Plant materials and growth conditions

One-year-old Ph. edulis seedlings were potted in our laboratory under long-day conditions (16 h light/8 h dark) at 28 °C, with a light intensity of 200 μmol m−2 s−1. The air relative humidity was about 50 %. To simulate stresses, the seedlings were cleaned with sterile water, and then stress treatments were conducted. The seedlings were exposed to the air for drought stress, those were incubated in sterile water for water stress, and the salinity stress was performed by incubating the seedling roots into the solution of 400 mmol L−1 NaCl [3, 27]. All treatments were conducted for two hours under the same temperature, air relative humidity and light condition as those of normal cultivation.

Total RNA isolation and expression analysis

For the tissue specific expression analysis of genes, roots, stems, leaves and sheaths of 1 year-old bamboo seedlings were collected and quickly frozen into liquid nitrogen for RNA isolation, respectively. Besides, roots and leaves of different treatments as mentioned above were also harvested for RNA isolation respectively using trizol reagent solution (Invitrogen, USA) followed by the manufacturer’s protocol. All samples collected from each treatment for at least three individuals (biological replicates). The total RNA was treated with RNase-free DNase (Promega, USA) for 30 min at 37 °C and the absence of DNA was verified by PCR using PeEF1 (Elongation factor 1, GenBank: FP095360.1) specific primers (PeEF1-F: 5′-GAATCCGAACTCCCGCCGTCTCT-3′) and (PeEF1-R: 5′-GCGGTCCTGAAGTAGAACTGTGG-3′). The integrity of total RNA was verified through agarose gel electrophoresis as well as the purity and concentration of total RNA was determined by spectrophotometer (Nanodrop2000). First-strand of cDNA was synthesized by reverse transcription system (Promega, USA). For each 20 μl reaction 1000 ng of total RNA was used, the synthesis was carried out at the 42 °C for 15 min and the final cDNA product was diluted 5-fold prior to use.

Based on the multiple alignments, specific primers for different AQP genes of Ph. edulis were designed and empirically adjusted for gene expression analysis (Supplementary Table 1). Besides, all primers showed a clear specific melting peak by real-time melting curve analysis, which agreed with the results of agarose gel electrophoresis for specific PCR product respectively, were used for further analysis. Quantitative real time RT-PCR (qRT-PCR) was performed on qTOWER2.2 (Analytik Jena, Germany) with Roche LightCycler®480 SYBR Green 1 Master kit, according to the procedure of 95 °C for 5 min; 95 °C for 10 s, 60–66 °C for 10 s, 40 cycles. The reaction volume was 10 μl containing 5.0 μl of 2× SYBR Green 1 Master, 0.8 μl cDNA, 0.2 μl of forward primer and reverse primer (5 μmol L−1, each) and 3.8 μl ddH2O. Calibration curves were obtained by amplification using a 10-fold serial dilution of each cDNA sample, the correlation coefficients (R2) of curves were also ranged from 0.990 to 0.999 and the PCR efficiency values were between 0.95 and 1.05 (Supplementary Table 1). For each condition, the qRT-PCR experiments were performed in biological triplicate with three technical replicates. Gene expression was calculated with 2−ΔΔCt method [28] using nucleotide tract-binding protein gene (NTB, GenBank: FP091427.1) as the reference gene [29]. To ensure the reliability of the data, another reference of tonoplast intrinsic protein gene (TIP41, GenBank: FP092936.1) was also used to validate the expression patterns in different tissues and those under different stress treatments. Both NTB and TIP41, found to be homogeneously expressed and were adequate for normalization purposes, were used for the gene expression analysis in Ph. edulis [29, 30].

Results

Identification and classification of AQP genes

Based on the AQP homologous sequence analysis, a total of 33 putative genome sequences and 40 putative full length cDNA sequences of Ph. edulis were mined from BambooGDB. While not always one AQP identified on the basis of EST evidence could be integrated into one genome sequence. To avoid confusion we decided to regard several AQP genes mapped to one genome sequence as the same one. Genes encoding partial aquaporin-like sequences, which were truncated and lack any of the conserved domains, were considered as non-functional pseudogenes and were excluded from the sequence analysis. For instance, the seven sequences (PH01000012G2420, PH01000021G0560, PH01001074G0030, PH01003142G0020, PH01003305G0130 PH01000157G0470 and PH01031653G0010) showed visibly error or loss of conserved domain, resulting in incomplete function of AQPs, were discarded.

A Neighbor-Joining phylogenetic tree was constructed based on the amino acid sequences of AQPs from Ph. edulis, Z. mays, O. sativa and P. trichocarpa (Fig. 1). The clustering data clearly showed that no AQPs sequence of Ph. edulis was clustered into the clade of XIPs, indicating that XIPs is not presence in Ph. edulis. Besides, the phylogenetic pattern of bamboo AQPs was in a similar manner to those of Z. mays and O. sativa counterparts, which clustered into four distinct subfamilies including PIPs (10 AQPs), TIPs (6 AQPs), NIPs (8 AQPs) and SIPs (2 AQPs). In addition, the PIPs, TIPs, NIPs and SIPs were further clustered into two subgroups (PIP1 and PIP2), three subgroups (TIP1, TIP2 and TIP4), three subgroups (NIP1, NIP2 and NIP3), and two subgroups (SIP1 and SIP2), respectively. Homologous AQPs corresponding to OsPIP2;8, OsTIP3;1, OsTIP5;1 and OsNIP4;1 were also not found in Ph. edulis (Fig. 1). Consequently, 26 full-length genes encoding AQPs were identified in the genome of Ph. edulis, which were given names (PeAQPs) correspondingly based on their orthologs to Z. mays and O. sativa (Table 1).

Fig. 1
figure 1

Phylogenetic tree of AQP amino acid sequences from Phyllostachys edulis, Zea mays, Oryza sativa and Populus trichocarpa. Predicted amino acid sequences were aligned using ClustalW2 program and the phylogenetic tree was constructed using bootstrap Neighbor-Joining tree method (1000 replicates) in MEGA6.0 software. Triangle, square, round and rhombus represent AQPs of Ph. edulis, Z. mays, O. sativa and P. trichocarpa, respectively

Full size image
Table 1 Nomenclature and protein information of AQPs in Ph. edulis
Full size table

Structure analysis and scaffold localization of PeAQPs

The distribution of the introns and exons were analyzed by comparing the CDS or EST sequences to the reference genome sequences. The results showed the complex gene structure with varying intron positions and lengths (Fig. 2). The number of introns in PeAQPs ranged from one (PeTIP1;1, PeTIP1;2 and PeTIP4;2) to four (PeNIP1;3, PeNIP2;1 and PeNIP2;2). With some exceptions, the number and size of the exons were conserved within each subfamily. Most of the PePIPs were characterized by four exons, except for PePIP1;3, PePIP2;1, PePIP2;2, and PePIP2;4, which contained three exons including a long 1st exon. Half of PeTIPs showed two exons, including PeTIP1;1, PeTIP1;2 and PeTIP4;2, and others exhibited three ones. Besides, most members of PeNIPs showed four exons, except for PeNIP1;3, PeNIP2;1 and PeNIP2;2 with five ones. Two members of the PeSIPs showed two exons of similar length.

Fig. 2
figure 2

Intron and exon organization of AQP genes in Ph. edulis. Round-corner rectangle and line represent exon and intron respectively

Full size image

All nonredundant PeAQPs were mapped and found to be distributed in the 26 scaffolds of the Ph. edulis genome (Fig. 3). Each scaffold only harbored one gene, which was unlike those in potato [1], and common bean [6] varied widely. This might be explained by the uncompleted genome and low abundance of genetic variability of bamboo [13].

Fig. 3
figure 3

Distribution of AQP genes on scaffolds in Ph. edulis genome

Full size image

Basic physical and chemical characteristics of PeAQPs

Furthermore, the MW and isoelectric point (pI) were calculated to analyze the protein characteristics of PeAQPs. The results showed that the MW of PeAQPs ranged from 23.5537 to 33.0378 kDa, and those of pI were 5.45 to 9.94 (Table 1). Most PeAQPs shared a similar MW and pI in the same subfamily, which were consistent with those of Chinese cabbage [5]. Comparative analysis of different subfamilies indicated that most PePIPs were alkaline, with MW of approximately 30 kDa. Besides, all PeTIPs were acidic and showed lower MW. However, PeNIPs showed a wider fluctuation in both MW and pI, and those of PeSIPs were alkaline and had lower MW than those of PePIPs and PeNIPs (Table 1).

Subcellular localization, transmembrane domains of PeAQPs

Based on the two different tools for sub-cellular localization prediction, all the members of PePIP subfamily were localized in the plasma membranes (Table 1). All PeTIPs were localized in the vacuole, PeNIPs and PeSIPs were localized in plasma membranes predicted by Plant-mPLoc algorithms. However, the results predicted by WoLF PSORT were diverse, in which included plasma membrane, tonoplast, chloroplast, cytosol, cytoskeleton, ER and nuclear (Table 1). Among this, PeTIP subfamily was not always localized in the tonoplast, which told us the names of AQPs were not always correspondent with the localization [6]. Prediction of TM helical domains showed that the numbers of TM helices in PeAQPs ranged from three (PeNIP1;4) to seven (PePIP2;7), and among which six TM helices (H1–H6) were the majority in PeAQPs (20 of 26, 77 %) (Table 1). PeTIPs and PePIPs, with the exception of PePIP2;5 and PePIP2;7, had six TM helices.

Analysis of conserved, substrate-specific residues of PeAQPs

The feature of substrate specificity determines that AQPs have their specific conserved amino acid domains. Both pore diameter and hydrophobicity are the vital factors that determine the specificity of substrate [8–11]. To analyze the conserved domains, multiple alignments were carried out using the sequences of PeAQPs. The results showed that all PePIPs and PeTIPs had two typical NPA in loop B and loop E (Table 2). However, some members of PeNIPs, such as PeNIP1;4 and PeNIP3;1, showed a variable first residue in the second NPA motif, in which the asparagine was replaced by aspartic acid or leucine, the third residue of the first NPA motif in PeNIP3;1 was serine. The third residue of the first NPA motif in PeSIPs was threonine and leucine, respectively. Besides the NPA motifs, the ar/R selectivity filter of AQPs is another indispensable structure determining channel selectivity. The ar/R in the AQPs was illustrated by the human AQP1 protein (F56-H180-C189-R195) [6]. The results showed different combination of ar/R selectivity filters were existed in different PeAQP subfamilies (Table 2). The ar/R selectivity filter in PeTIPs was diverse, all the TM2 residues of the ar/R filter region were H, except in PeTIP4;1 where Q was found. The major of TM5 position was I, except for PeTIP4;1 which was replaced by T. PeTIP4;1 had a specific ar/R selectivity filter, suggesting that it probably had function of unconventional solutes permeation. Besides, The loop E1 and E2 position were mostly A and R respectively, except for PeTIP1(V) in E2 and PeTIP2(G) in E2. Both PeNIPs and PeSIPs showed wide variability in their ar/R selectivity filter. PeNIP1 showed the residues of W/V/A/R, except for W/P/V/V in PeNIP1;4. PeNIP2 residues were G/S/G/R. PeNIP3;1 presented a new combination (A/I/D/A). Due to PeNIP3;2 was lack of the second NPA, it did not work to find the ar/R in loop E1 and E2. The PeSIPs exhibited two different types of residues, L/V/P/N (PeSIP1;1) and S/H/G/S (PeSIP1;2) respectively (Table 2). These diverse characteristics of PeAQPs mean that they might have a variety of functions involved in numerous physiological processes during the growth and development of bamboo.

Table 2 Conserved dual NPA motifs and ar/R selectivity filter (TM2, TM5, loop E1 and loop E2) in AQPs of Ph. edulis
Full size table

Tissue specific expression of PeAQPs

The expression patterns of 26 PeAQPs in four different tissues were analyzed by qRT-PCR (Fig. 4). The results indicated that the 26 PeAQPs were all detected in the tissues with diverse patterns. Several genes (PePIP1;1, PePIP1;2, PePIP1;3, PePIP2;2, PeTIP1;1 and PeSIP1;1) seemed to be constitutively expressed, indicating they might play important roles in maintaining transport processes throughout the plant. Beside, some genes expressed specifically among one or two tissues were also detected. PePIP2;3, PePIP2;4, PeTIP2;1, PeTIP2;2 and PeNIP2;1 showed a relatively high-level expression in root, among these two PeTIPs (PeTIP2;1 and PeTIP2;2) were expressed almost exclusively in root and two PePIPs (PePIP2;3 and PePIP2;4) still displayed lower expression level in leaf. PePIP2;7 was expressed almost exclusively in leaf and rarely in other tissues. PePIP2;1 and PeTIP1;3 were mainly expressed in root and leaf, PeNIP2;2 showed higher expression level in sheath and stem. The second reference gene (TIP41) was also used to validate the expression patterns of PeAQPs (Supplementary Fig. 1), which demonstrated very consistent results with those of the former using NTB as native control. These results indicated that PeAQPs maybe play vital roles in cellular homeostasis of different tissues for maintaining bamboo growth and development.

Fig. 4
figure 4

Tissue specific expression analysis of PeAQPs in Ph. edulis using qRT-PCR. 1 Root; 2 stem; 3 leaf; 4 sheath. The expression values in various tissues were normalized to that of in root. NTB was used as the reference gene. Error bars represent standard deviation of biological triplicate with three technical replicates

Full size image

Expression of PeAQPs in response to stresses

Abiotic stresses are known to induce osmotic stress in plants and disturb the plant water balance. Since AQPs are important to maintain water balance in plant cell, the expression patterns of PeAQPs in root and leaf were further analyzed after 2 h treatment of drought, water, and NaCl stresses using qRT-PCR with NTB as reference gene, respectively. The results demonstrated that PeAQPs have a varied expression patterns in response to different abiotic stresses. We identified 8 PeAQPs that were significantly up-regulated in response to drought stress (>2-fold, p < 0.05), including three genes (PeTIP4;1, PeTIP4;2 and PeNIP1;1) in root (Fig. 5) and 6 genes (PePIP2;1, PePIP2;4, PePIP2;7, PeTIP1;2, PeTIP4;1 and PeNIP2;2) in leaf (Fig. 6). Among these, PeTIP4;1 was categorized as commonly induced genes in both root and leaf, which was up-regulated for 4.7 and 17.5 times of the control respectively. Interestingly, PePIP2;4 was up-regulated in leaf but down-regulated in root by drought treatment. In leaf, the number of induced genes (six as above) were greater than that of repressed ones (PePIP2;3 and PePIP2;6) under drought stress (Fig. 6). However, in root, the opposite trend occurred with the number of induced genes were less than that of repressed ones (PePIP2;4, PeTIP2;1, PeTIP2;2, PeNIP1;4 and PeNIP2;1) (Fig. 5).

Fig. 5
figure 5

qRT-PCR analysis of PeAQPs in root of Ph. edulis under stresses. 1 Expression profiles in untreated root; 2, 3 and 4 expression profiles in root after stress treatments of drought, water and 400 mmol L−1 NaCl for 2 h, respectively. The gene expression values were normalized to that of untreated root. NTB was used as the reference gene. Error bars represent standard deviation of biological triplicate with three technical replicates

Full size image
Fig. 6
figure 6

qRT-PCR analysis of PeAQPs in leaf of Ph. edulis under stresses. 1 Expression profiles in untreated leaf; 2, 3 and 4 expression profiles in leaf after stress treatments of drought, water and 400 mmol L−1 NaCl for 2 h, respectively. The gene expression values were normalized to that of untreated leaf. NTB was used as the reference gene. Error bars represent standard deviation of biological triplicate with three technical replicates

Full size image

After water stress, the expression levels of all PeNIPs (except PeNIP1;4), two PePIPs (PePIP2;2, PePIP2;5), one PeTIPs (PeTIP1;1) and one PeSIPs (PeSIP1;1) were up-regulated significantly (p < 0.05) in root, among which PeNIP3;2 was induced for 25 times of the control (Fig. 5). The qRT-PCR analysis of PeAQPs in leaf also established the relationship with water stress (Fig. 6). Three PePIPs (PePIP2;4, PePIP2;5 and PePIP2;7), three PeTIPs (PeTIP1;1, PeTIP1;2 and PeTIP4;1), five PeNIPs (PeNIP1;1, PeNIP1;3, PeNIP1;4, PeNIP2;1 and PeNIP2;2) and one PeSIPs (PeSIP1;1) were significantly (p < 0.05) induced in leaf, among which the highest expression level of PeTIP4;1 was up-regulated to 82 times of the control. In addition, PeNIP2;1, PePIP2;7 and PeNIP2;2 were also up-regulated for 18, 18 and 17 times, respectively. It can be speculated that these genes are responsive to water stress and may be involved in the anti-waterlogging.

In response to NaCl stress, 9 PeAQPs (PePIP1;2, PePIP1;3, PePIP2;1, PePIP2;4, PePIP2;5, PePIP2;7, PeTIP1;2, PeTIP2;1 and PeNIP2;1) were up-regulated significantly (p < 0.05) and none was down-regulated significantly in root compared to those under water stress (Fig. 5). In leaf PePIP1;3 and PePIP2;4 were up-regulated while PePIP1;2 and PeTIP2;1 were down-regulated significantly (p < 0.05), respectively (Fig. 6). Besides, PePIP1;3 and PePIP2;4 were categorized as commonly induced genes in both root and leaf, while PePIP1;2 and PeTIP2;1 showed opposite expression patterns in root compared with those in leaf under NaCl stress.

Overall, these findings demonstrated that the expression of PeAQPs were differentially regulated by various abiotic stresses, which might be explained as a part of adaptation mechanism in bamboo by minimizing water losses through efflux to survive under drought and NaCl stresses and preventing moisture from entering under water stress.

Meanwhile, all PeAQPs were validated by using TIP41 as a second reference gene, and similar results were obtained (Supplementary Fig. 2, Fig. 3). The results of differentially expressed genes strengthened the data produced by using NTB, which will be helpful to select candidate genes for genetic engineering in future.

Discussion

With the rapid development of technologies in plant genome sequencing, an increasing number of plant genomes have been successfully sequenced, which facilitate the genome-wide genes identification. AQPs, involved in regulating plant growth and development, are abundant, diverse and widely distributed in plant cell. Genome-wide analysis of AQPs has been reported in some monocots, such as maize, rice, wheat and barley [3, 7, 18, 19]. The first investigation of AQPs in bamboo was performed in this study, the number of PeAQPs is slightly less than those found in other monocots [3, 17–19] (Table 1). It can be speculated that the genome is still a draft and assembly covers ~95 % of the whole genomic region [13]. Besides, unpredictable flowering and lacking sexual reproduction are expected to negatively affect the diversity of genetics [31, 32].

Furthermore, PeAQPs were clustered into four subfamilies with different subgroups, which was in agreement with current literatures [3, 7, 17–19] (Fig. 1). PePIPs were divided into two subgroups based on N- and C-terminal motifs, which seemed to have different water transport activities in oocyte experiments [33, 34]. TIPs in some monocots, such as maize, rice and barley, were divided into five subgroups of TIP1, TIP2, TIP3, TIP4 and TIP5. While PeTIPs were only consisted of three subgroups, the members of TIP3 and TIP5 were not found in Ph. edulis. Otherwise, PeNIPs included three subgroups, and PeSIPs were classified into two subgroups in which PeSIP1 were found to form a clade distinct from PeSIP2. In addition, Ph. edulis had close genetic relationship with O. sativa, followed by Z. mays, and distant with P. trichocarpa. This result is consistent with the findings in the stated phylogeny and classification of grass subfamilies [13].

To seek further insights into gene structure, the splicing pattern of PeAQPs were analyzed. Similar to many eukaryotic nuclear introns, PeAQPs showed the canonical dinucleotides, AG and GT in their acceptor and donor splice sites respectively. Most PeAQPs in each subfamily had the same size for the homologous exons, with the difference only in the size of introns (Fig. 2). In PePIPs, all the number and size of exons were conserved, in which the last two exons had the same size, with the differences of intron length and the position of first intron splicing junction. Furthermore, with some exceptions, the number and size of exons (but not of the introns) in most members of PeNIPs and PeTIPs are conserved in each subgroup. This finding further validated the nomenclature proposed by cluster analysis.

Not only involved in the transport and storage of water, AQPs also play a key role in transportation of some small molecules, such as silicon, boron, arsenic and lactic acid [9, 12]. An essential feature of AQPs is having highly conserved NPA and ar/R selectivity filter, which is indispensable in determining transport specificity of channel [9, 35, 36]. In addition, the ar/R selectivity filter also had highly conserved and typically domain (F/H/T/R), which was shared by the PIP subfamily in some plants such as Arabidopsis, maize, wheat and barley [7, 17–19]. PeTIPs exhibited four different combinations such as H/I/A/V (PeTIP1), H/I/G/R (PeTIP2), Q/T/A/R (PeTIP4;1) and H/I/A/R (PeTIP4;2), confirming the variability of TIPs in plants [7, 17–19]. Widely distributed in plasma membrane and tonoplast, PIPs and TIPs were the main channels controlling water balance in plant cell [2, 37]. The NIPs subfamily had a lower conductivity of transport water and had function of unconventional solutes like formamide, glycerol and ammonia [12, 37]. Except PeNIP1;4, all members of PeNIP1 and PeNIP2 showed conserved NPA and ar/R selectivity filter sequence shared with other monocots [17–19]. The residues of other NIPs presented different motifs led to different pore sizes of the channel, which might further result in the transportation of unconventional solutes in plants [11, 12]. SIPs had the fewest members and smallest MW in AQPs. SIP1 possessed the ability of transporting water across the ER membrane and SIP2 might act as an ER channel for other small molecules or ions [38, 39]. This study shows the third residue in the first NPA of SIPs is variable in bamboo (Table 2), which might more or less influence the function of transporting water. Besides, the variety of ar/R selectivity filter suggested that they could transport some novel substrates [39].

AQPs are widely distributed in plant tissues and their expression levels are correlated with the developmental stages and environmental conditions [2, 40]. The expression levels of AQPs in different tissues as well as under different stresses may be a cue to reflect their physiological functions. The qRT-PCR results showed that the transcripts of PeAQPs were detected in all tissues, but the levels varied greatly in different tissues (Fig. 4), indicating that PeAQPs might function differently. For instance, PePIP2;7 showed higher expression in leaf and was obviously induced by drought and water stresses, similar to those findings of the corresponding homologs in P. balsamifera and Jatropha curcas [41, 42], indicating PePIP2;7 might have important role in leaf water transport. PePIP2;4 displayed a higher expression level in root, which was up-regulated by salinity stress and down-regulated by drought stress, suggesting that PePIP2;4 might be involved in a part of adaptation mechanism to maintain water balance under stress. PeNIP2;1, predominantly expressed in the root, was obviously induced under salinity and water stresses, but suppressed by drought stress, indicating that PeNIP2;1 might be a channel transporter of lactic acid induced by anoxic stress [12]. Besides, PePIPs demonstrated higher expression levels than other three subfamilies, similar to those cases in Arabidopsis, rice and maize [3, 7, 17]. With some exceptions, PeTIPs were expressed more abundantly in root than other tissues. For example, PeTIP2;1 and PeTIP2;2 were tissue-specific and expressed higher in root than in other tissues, similar to those findings of the corresponding homologs in Arabidopsis [43]. It suggests that PeTIPs may play important roles in transporting inorganic solutes and metabolites from the soil. Previous studies have been reported that AtTIP1;1 is indispensable under environmental stress [44, 45]. In this study, PeTIP1;1 was constitutively expressed and up-regulated in both root and leaf under water stress, indicating PeTIP1;1 may be involved in water stress tolerance. In Arabidopsis, SIP1;1 and SIP1;2 were found take part in water transportation, while SIP2;1 was found transporting other small molecules or ions in the ER [38, 39]. It suggests that PeSIPs may be a channel in ER of bamboo requiring further confirmation. In short, the different expression patterns of PeAQPs indicated that they may have functions with varying degrees for maintaining water homeostasis in bamboo.

Abiotic stresses such as drought, water and salinity caused serious losses in plant production throughout the world [46]. Water is vital for plant production and retaining water is essential for plants to combat drought stress [47]. AQPs functioned in rapid transmembrane water flow during plant growth and development. Since AQPs in roots are important in taking up water from the soil [2], we firstly investigated the expression levels of PeAQPs in root under drought conditions. Three PeAQPs (PeTIP4;1, PeTIP4;2 and PeNIP1;1) in root were induced by drought stress, among which PeTIP4;2 was up-regulated for 18 times of the control, suggesting that it may play a potential role in response to drought stress. PIPs are known to play important roles in water transport through the plasma membrane in plants [33, 34, 40, 48, 49]. Overexpressing VfPIP1 in Arabidopsis resulted in improved drought tolerance by inducing stomatal closure and reducing the loss of transpiration under drought conditions [50]. Interestingly, in response to drought, most PIPs and TIPs showed down-regulated trend in root, which were also found in A. thaliana [40]. Water stress was another factor limited plant growth. As a transport unconventional solutes channel, almost all PeNIPs in root were induced under water stress, meaning PeNIPs might be a channel transporter of lactic acid induced by anoxic stress [12]. In general, this might be a self-protection for the plant during drought stress by maintaining the balance of water.

However, compared with drought stress, PeAQPs demonstrated an opposite expression trend in root under NaCl stress, which was consistent with those in barley [19]. Among this, PePIP1;3 and PePIP2;4 were obviously induced in both root and leaf. It has been reported that some PIPs heterogeneous expression increased salinity stress tolerance [10, 51, 52]. In leaf, the expression of most PeTIPs was down-regulated by drought stress and up-regulated by salinity stress (Fig. 6). The highly expressed genes may be involved in the regulation of high-salinity stress responses and play an important role in osmotic regulation under salinity stress. It is obviously that PeAQPs were expressed differentially in root and leaf, indicating that PeAQPs may play roles with varying degrees in bamboo for the tolerance of drought, water, NaCl and other abiotic stresses. Nonetheless, further researches are in need to help us gain a better understanding about the molecular mechanisms of the PeAQPs in bamboo.