The pip1s Quintuple Mutants Demonstrate the Essential Roles of PIP1s in the Plant Growth and Development of Arabidopsis

Plasma membrane intrinsic proteins (PIPs) transport water, CO2 and small neutral solutes across the plasma membranes. In this study, we used the clustered regularly interspaced short palindromic repeats (CRISPR) and CRISPR-associated protein 9 system (CRISPR/Cas9) to mutate PIP1;4 and PIP1;5 in a pip1;1,2,3 triple mutant to generate a pip1;1,2,3,4,5 (pip1s−) quintuple mutant. Compared to the wild-type (WT) plant, the pip1s− mutants had smaller sized rosette leaves and flowers, less rosette leaf number, more undeveloped siliques, shorter silique and less seeds. The pollen germination rate of the pip1s− mutant was significantly lower than that of the WT and the outer wall of the pip1s− mutant’s pollen was deformed. The transcriptomic analysis showed significant alterations in the expression of many key genes and transcription factors (TFs) in the pip1s− mutant which involved in the development of leaf, flower and pollen, suggesting that the mutant of PIP1s not only directly affects hydraulics and carbon fixation, but also regulates the expression of related genes to affect plant growth and development.


Introduction
Water movement is a key physiological process of vegetative and reproductive growth in plants that must be tightly regulated under different conditions [1][2][3][4][5]. The transport of water is controlled by both symplastic and apoplastic pathways. Aquaporins (AQPs) play a central role in the symplastic pathway which is efficient in transporting water across membranes [6][7][8]. Aquaporins transport water, CO 2 and small neutral solutes through the plasma membranes and intracellular membranes of cells in many physiological and developmental processes, including cell elongation, stomatal regulation, seed germination, reproductive growth and stress responses in plants [1,5,7,8].
According to the conserved amino acid sequences and intron positions, AQPs are divided into five groups including plasma membrane intrinsic proteins (PIPs), tonoplast intrinsic proteins (TIPs), nodulin-like plasma membrane intrinsic proteins (NIPs), small intrinsic proteins (SIPs), and X intrinsic proteins (XIPs) [9][10][11][12]. PIPs are located to the plasma membrane and regulate water movement between cells to maintain the water balance [7,13]. Sequences of the PIP1s and PIP2s subgroups are highly conserved, however, the two isoforms are different in many aspects including their transport properties and subcellular localization.
The PIP2s subgroup is an efficient water channel, but the water transport capacity of PIP1s is inconclusive. Although some PIP1 has been proven to be an effective water channel in some species [14][15][16], the water transport efficiency of other PIP1s is relatively low or even none [17][18][19][20]. Subsequent studies have shown that this may be due to singly expressed PIP1s are retained intracellularly and fail to traffic to the plasma membrane [21,22]. Coexpression studies in Xenopus oocytes or maize protoplasts have clearly elucidated that PIP2s physically interact with PIP1s to form a heterotetramer [21,22]. PIP2s promote the

The pip1s -Quintuple Mutants Are Defective in Plant Vegetative Growth
The WT, pip1s --1, and pip1s --2 plants were grown on soil under 16/8 h light/dark conditions to investigate the phenotypes of the pip1s -. Compared to the WT plant, the pip1smutants had smaller rosette leaves and fewer rosette leaves during the vegetative stage (Figure 2A-D). Early bolting was observed in the pip1s -mutants under normal growth conditions (Figure 2A). The bolting time in the pip1s -mutants was significantly advanced (up to 3 days) compared to the WT plants ( Figure 2E).

The pip1s − Quintuple Mutants Are Defective in Plant Vegetative Growth
The WT, pip1s − -1, and pip1s − -2 plants were grown on soil under 16/8 h light/dark conditions to investigate the phenotypes of the pip1s − . Compared to the WT plant, the pip1s − mutants had smaller rosette leaves and fewer rosette leaves during the vegetative stage (Figure 2A-D). Early bolting was observed in the pip1smutants under normal growth conditions ( Figure 2A). The bolting time in the pip1s − mutants was significantly advanced (up to 3 days) compared to the WT plants ( Figure 2E). The area of the seventh rosette leaf of the WT, pip1s --1 and pip1s --2 plants after transplanted for 3 weeks. The area was determined with ImageJ software. (E) Days to start bolting of the WT, pip1s --1 and pip1s --2 plants. Bars represent the means ± SD of three biological replicates. Significant differences were determined by one-way analysis of variance (ANOVA) followed by Duncan's multiple range test (*** p < 0.0001).

Mutation of PIP1s Affects Flower Growth and Development
Compared to the WT plants, significantly reduced lengths and widths of mature flowers were observed in pip1s --1 and pip1s --2 mutants ( Figure 3A,B). Many flower buds of pip1s -mutants began to wither before flowering ( Figure 3C). Eventually, the sepals, petals, stamens, and all carpels of the flower withered ( Figure 3D). The area of the seventh rosette leaf of the WT, pip1s − -1 and pip1s − -2 plants after transplanted for 3 weeks. The area was determined with ImageJ software. (E) Days to start bolting of the WT, pip1s − -1 and pip1s − -2 plants. Bars represent the means ± SD of three biological replicates. Significant differences were determined by one-way analysis of variance (ANOVA) followed by Duncan's multiple range test (*** p < 0.0001).

Mutation of PIP1s Affects Flower Growth and Development
Compared to the WT plants, significantly reduced lengths and widths of mature flowers were observed in pip1s − -1 and pip1s − -2 mutants ( Figure 3A,B). Many flower buds of pip1s − mutants began to wither before flowering ( Figure 3C). Eventually, the sepals, petals, stamens, and all carpels of the flower withered ( Figure 3D). Bars represent the means ± SD of three biological replicates. Significant differences were determined by one-way ANOVA followed by Duncan's multiple range test (*** p < 0.0001).

Mutation of PIP1s Decreases the Number and Length of Silique
Unfertilized or withered flowers and undeveloped siliques were observed in both WT and pip1s -mutant plants ( Figure 4A). Undeveloped siliques turned yellow, wilted, and even moldy during silique development ( Figure 4B). The number of siliques per plant was significantly reduced in the pip1s -mutants compared to the WT plants (p < 0.01) (Figure 4C). We divided these siliques into three major categories: well-developed, undeveloped, and unfertilized ( Figure 4D). The percentages of undeveloped and unfertilized siliques were substantially increased in the pip1s -mutants. Bars represent the means ± SD of three biological replicates. Significant differences were determined by one-way ANOVA followed by Duncan's multiple range test (*** p < 0.0001).

Mutation of PIP1s Decreases the Number and Length of Silique
Unfertilized or withered flowers and undeveloped siliques were observed in both WT and pip1s − mutant plants ( Figure 4A). Undeveloped siliques turned yellow, wilted, and even moldy during silique development ( Figure 4B). The number of siliques per plant was significantly reduced in the pip1s − mutants compared to the WT plants (p < 0.01) ( Figure 4C). We divided these siliques into three major categories: well-developed, undeveloped, and unfertilized ( Figure 4D). The percentages of undeveloped and unfertilized siliques were substantially increased in the pip1s − mutants. Siliques were divided into three categories and the percentage was calculated. Bars represent the means ± SD of three biological replicates. Significant differences were determined by one-way ANOVA followed by Duncan's multiple range test (*** p < 0.0001).
When comparing the developed siliques of the WT and pip1s -mutant plants, we found that the mutant siliques were nicked ( Figure 5A). The pip1s -mutants had shorter silique lengths than the WT plants at 14 days after flowering (DAF) ( Figure 5B). The average seed number per silique of the WT was 46.7, which was significantly greater than that of the pip1s --1 (23.8) and pip1s --2 (23.1) (p < 0.01) ( Figure 5C). We divided the silique length into three categories: >10 mm, 6-10 mm, and <6 mm. The silique lengths of the pip1s --1 and pip1s --2 mutants were largely reduced compared to the WT plants ( Figure 5D). Siliques were divided into three categories and the percentage was calculated. Bars represent the means ± SD of three biological replicates. Significant differences were determined by one-way ANOVA followed by Duncan's multiple range test (*** p < 0.0001).
When comparing the developed siliques of the WT and pip1smutant plants, we found that the mutant siliques were nicked ( Figure 5A). The pip1smutants had shorter silique lengths than the WT plants at 14 days after flowering (DAF) ( Figure 5B). The average seed number per silique of the WT was 46.7, which was significantly greater than that of the pip1s − -1 (23.8) and pip1s − -2 (23.1) (p < 0.01) ( Figure 5C). We divided the silique length into three categories: >10 mm, 6-10 mm, and <6 mm. The silique lengths of the pip1s − -1 and pip1s − -2 mutants were largely reduced compared to the WT plants ( Figure 5D). Bars represent the means ± SD of three biological replicates. Significant differences were determined by oneway ANOVA followed by Duncan's multiple range test (*** p < 0.0001).

Mutation of PIP1s Affects Seed Development
Compared to the WT plants, some aborted and malformed seeds were observed in the pip1s -mutants ( Figure 6A,B). The proportions of the aborted seeds in the pip1s --1 and pip1s --2 mutants were 31.2% and 31.6%, respectively ( Figure 6C). Bars represent the means ± SD of three biological replicates. Significant differences were determined by one-way ANOVA followed by Duncan's multiple range test (*** p < 0.0001).

Mutation of PIP1s Affects Seed Development
Compared to the WT plants, some aborted and malformed seeds were observed in the pip1s − mutants ( Figure 6A,B). The proportions of the aborted seeds in the pip1s − -1 and pip1s − -2 mutants were 31.2% and 31.6%, respectively ( Figure 6C). . Bars represent the means ± SD of three biological replicates. Significant differences were determined by one-way ANOV followed by Duncan's multiple range test (** p < 0.0005, *** p < 0.0001).

Mutation of PIP1s Affects Pollen Morphology and Activity
Scanning electron microscopy indicated that both the WT and pip1s -mutant had regular honeycomb outer wall structures. However, the outer wall of the muta len was deformed and shrunk ( Figure 7A). Pollen germination experiments conduc vitro revealed that the pollen germination rate of the pip1s -mutant was significantly than the WT plant (p < 0.01) ( Figure 7B,F). The pollen germination experiment cond in vivo produced similar results, and the pollen-stigma binding ability of the pip1s -m was impaired compared to the WT plant ( Figure 7D). Alexander staining experimen played a lower number of active pollen grains per anther in the pip1s -mutant than WT plant ( Figure 7C). Bars represent the means ± SD of three biological replicates. Significant differences were determined by one-way ANOVA followed by Duncan's multiple range test (** p < 0.0005, *** p < 0.0001).

Mutation of PIP1s Affects Pollen Morphology and Activity
Scanning electron microscopy indicated that both the WT and pip1s − mutant pollen had regular honeycomb outer wall structures. However, the outer wall of the mutant pollen was deformed and shrunk ( Figure 7A). Pollen germination experiments conducted in vitro revealed that the pollen germination rate of the pip1smutant was significantly lower than the WT plant (p < 0.01) ( Figure 7B,F). The pollen germination experiment conducted in vivo produced similar results, and the pollen-stigma binding ability of the pip1s − mutant was impaired compared to the WT plant ( Figure 7D). Alexander staining experiments displayed a lower number of active pollen grains per anther in the pip1s − mutant than in the WT plant ( Figure 7C). vitro revealed that the pollen germination rate of the pip1smutant was significantly lower than the WT plant (p < 0.01) ( Figure 7B,F). The pollen germination experiment conducted in vivo produced similar results, and the pollen-stigma binding ability of the pip1s -mutant was impaired compared to the WT plant ( Figure 7D). Alexander staining experiments displayed a lower number of active pollen grains per anther in the pip1s -mutant than in the WT plant ( Figure 7C). Bars represent the means ± SD of three biological replicates. Significant differences were determined by one-way ANOVA followed by Duncan's multiple range test (*** p < 0.0001).
Further analysis of DEGs found that many were related to photosynthesis, pollen wall formation, pollen development, pollen hydration and pollen tube growth (Table 1).

GO and KEGG Analyses of pip1s − Mutants and WT plants
The gene Ontology (GO) analysis of DEGs in the group muleaf vs. WTleaf revealed 176 GO terms that were significantly altered at FDR < 0.05, including 130 biological processes, 22 cellular components and 24 molecular functions ( Figure 5A). While in the group musili vs. WTsili, 162 terms with FDR < 0.05 were identified, including 112 biological processes, seven cellular components, and 43 molecular functions ( Figure 5B). The top significantly altered terms were mainly related to metabolic processes, cellular processes, responses to stimulus and biological regulation, suggesting that the five PIP1 genes played an important role in these processes.
In the Kyoto Encyclopedia of Genes and Genomes (KEGG) category, 881 transcripts were assigned to 117 pathways, of which 4 pathways were significantly enriched in group muleaf vs. WTleaf (q ≤ 0.05). Another 616 transcripts were assigned to 109 pathways, of which 7 pathways were significantly enriched in group musili vs. WTsili ( Table 2). The KEGG pathway analysis showed that 47 hormone-related genes were significantly differentially expressed (including 26 up-regulated and 21 down-regulated) in the pip1s − mutant rosette leaves compared to the WT rosette leaves. The analysis of secondary metabolism pathways using KEGG showed that the phenylpropanoids and flavonoid biosynthesis pathway were significantly altered in the rosette leaves and silique of the pip1s − mutant compared to the WT plants. Some genes related to starch and sucrose metabolism and cutin, suberin and wax biosynthesis were significantly altered in the pip1s − mutant silique compared to the WT siliques ( Figure 9C).

Transcription Factor Analysis of the DEGs
A total of 150 and 69 transcription factor genes were significantly differentially expressed in the rosette leaves and siliques of the pip1s -mutant compared to the WT plant, respectively. Among the identified TF genes, many were involved in leaf, flower and seed development (Tables 3 and 4). In rosette leaves, the TF genes related to leaf development covered many TF families, such as TCP, AP2, B3, MYB, NAC and WRKY. The expression

Transcription Factor Analysis of the DEGs
A total of 150 and 69 transcription factor genes were significantly differentially expressed in the rosette leaves and siliques of the pip1s − mutant compared to the WT plant, respectively. Among the identified TF genes, many were involved in leaf, flower and seed development (Tables 3 and 4). In rosette leaves, the TF genes related to leaf development covered many TF families, such as TCP, AP2, B3, MYB, NAC and WRKY. The expression of seven AP2 genes (ERF1A, ERF2, ERF6, ERF11, ERF018, ERF104, and ERF 109) were all upregulated in the pip1s − mutant compared to the WT plants, and the GO analysis indicated that these genes were involved in cell division [54,55]. Four NAC genes, including NAC59, NAC29, NAC42, and NAC84, involved in leaf senescence and cell division and expansion were all promoted in the pip1s − mutant compared to the WT plants [56]. In the silique, the TF genes related to flower and seed development included the B-box domain proteins, zinc finger proteins, MYB, B3 and NAM TF families. The expression of four B-box domain protein genes (MIP1B, BBX32, AT5G54470, and AT4G27310) were all down-regulated in the pip1s − mutant compared to the WT and the GO analysis indicated that these genes were involved in the regulation of flower development. Four zf-Dof genes, including CDF1, CDF2, CDF3, and CDF5, involved in flower development, were all repressed in the pip1smutant compared to the WT plants.

Verification of Transcriptome Sequencing
Eight genes were selected for quantitative real-time PCR (Q-PCR) to confirm the reliability of the RNA-seq data ( Figure 10). LHCA1 and LHCB1.1 encode chlorophyll a/b binding protein, as a light harvesting protein complex in photosystem (PS) I or II, which converts light energy into unstable chemical energy. RAV1 was identified as a negative regulator of leaf development in Arabidopsis. WRKY22 was reported to be related to leaf senescence and response to chitin. CALS7 and CALS8 are callose-synthesizing genes, and their deletion hinders the normal development of pollen outer walls. MYB103 was reported to regulate the expression of CALS7 and CALS8, and subsequently regulate the development of pollen tapetum. PMEI1 was identified as pectin methyl esterase inhibitor, which was located at the top of pollen tube cells and negatively regulated catalysis. LHCA1, LHCB1.1, CALS7, CALS8 and MYB103 were significantly down regulated, whereas RAV1, WRKY22 and PEMI1 were significantly up regulated in pip1s − mutant compared to the WT ( Figure 10). Bars represent means ± SD of three biological replicates. Significant differences were determined by one-way ANOVA followed by Duncan's multiple range test (* p < 0.05).

Mutation of PIP1s Influences the Vegetative Growth of Arabidopsis
PIPs are usually located to the plasma membrane, but recent evidence suggests that some isoforms may also be located to the chloroplast envelope in at least minute amounts [57][58][59]. The co-localization of the PIPs and TIPs in chloroplast and thylakoid membranes may reflect their key roles in water supplying and CO2 transport for photosynthesis [60]. Some PIP1s appear to play a dual role in water and CO2 transport of Arabidopsis and tobacco [14,[61][62][63]. The CO2 permeability of the chloroplast envelope purified from tobacco leaves is five times lower than that of plasma membrane vesicles. After antisense inhibition of NtAQP1 in transgenic tobacco, its CO2 permeability is reduced by 90% [64]. In addition to NtAQP1, AtPIP1;2 has also been shown to promote CO2 transmembrane transport after heterologous expression in yeast cells [65]. The genetic changes intheir functions in tobacco or Arabidopsis plants, using antisense suppression or overexpression, revealed a positive correlation between their expression and CO2 assimilation rate [65]. To date, the aquaporins that all belong to the subclasses of PIP1 and PIP2 have been shown to promote the membrane diffusion of CO2 in several plant species.
In our study, compared to the WT plants, the pip1s -mutants had smaller rosette leaf size and fewer rosette leaf numbers which exhibited overall defects in vegetative growth. The vegetative growth defect of the mutant may be caused by the mutation of PIP1s which limits the transportation of CO2 in plants, thereby affecting the second stage of photosynthesis in plants (Figure 11). On the other hand, seven down-regulated LHCAs and LHCBs, which encode the chlorophyll a/b binding protein, may affect the first stage of photosynthesis and block the conversion of light energy into unstable chemical energy (Figure 11). In addition, according to our previous studies, single pip1 -mutants, double mutants, or Bars represent means ± SD of three biological replicates. Significant differences were determined by one-way ANOVA followed by Duncan's multiple range test (* p < 0.05).

Mutation of PIP1s Influences the Vegetative Growth of Arabidopsis
PIPs are usually located to the plasma membrane, but recent evidence suggests that some isoforms may also be located to the chloroplast envelope in at least minute amounts [57][58][59]. The co-localization of the PIPs and TIPs in chloroplast and thylakoid membranes may reflect their key roles in water supplying and CO 2 transport for photosynthesis [60]. Some PIP1s appear to play a dual role in water and CO 2 transport of Arabidopsis and tobacco [14,[61][62][63]. The CO 2 permeability of the chloroplast envelope purified from tobacco leaves is five times lower than that of plasma membrane vesicles. After antisense inhibition of NtAQP1 in transgenic tobacco, its CO 2 permeability is reduced by 90% [64]. In addition to NtAQP1, AtPIP1;2 has also been shown to promote CO 2 transmembrane transport after heterologous expression in yeast cells [65]. The genetic changes intheir functions in tobacco or Arabidopsis plants, using antisense suppression or overexpression, revealed a positive correlation between their expression and CO 2 assimilation rate [65]. To date, the aquaporins that all belong to the subclasses of PIP1 and PIP2 have been shown to promote the membrane diffusion of CO 2 in several plant species.
In our study, compared to the WT plants, the pip1s − mutants had smaller rosette leaf size and fewer rosette leaf numbers which exhibited overall defects in vegetative growth. The vegetative growth defect of the mutant may be caused by the mutation of PIP1s which limits the transportation of CO 2 in plants, thereby affecting the second stage of photosynthesis in plants ( Figure 11). On the other hand, seven down-regulated LHCAs and LHCBs, which encode the chlorophyll a/b binding protein, may affect the first stage of photosynthesis and block the conversion of light energy into unstable chemical energy ( Figure 11). In addition, according to our previous studies, single pip1mutants, double mutants, or even triple mutants had no significant effect on vegetative growth, implying that in addition to AtPIP1;2, other PIP1s genes are also involved in the active transport of CO 2 . The phenotype of plants with genetically altered aquaporins is usually difficult to decipher, because it integrates far more than the direct effects of aquaporins on tissue hydraulics or carbon fixation, but mutations in PIP1s may also regulate the expression of other AQPs or transcription factors and hormone signal transduction. According to our work, a hypothesis has been proposed that the mutation of PIP1s defect the vegetative growth of mutants by regulating the expression of chlorophyll a/b binding protein genes and the transmembrane transport of CO 2 ( Figure 11). even triple mutants had no significant effect on vegetative growth, implying that in addition to AtPIP1;2, other PIP1s genes are also involved in the active transport of CO2. The phenotype of plants with genetically altered aquaporins is usually difficult to decipher, because it integrates far more than the direct effects of aquaporins on tissue hydraulics or carbon fixation, but mutations in PIP1s may also regulate the expression of other AQPs or transcription factors and hormone signal transduction. According to our work, a hypothesis has been proposed that the mutation of PIP1s defect the vegetative growth of mutants by regulating the expression of chlorophyll a/b binding protein genes and the transmembrane transport of CO2 ( Figure 11).

Mutation of PIP1s Influences the Reproductive Growth of Arabidopsis
Two desiccated forms of higher plant life, pollen and seeds, play an important role in the plant life cycle. Pollen and seeds express TIP5 and TIP3 specific aquaporin subclasses, respectively [66][67][68]. This specificity may result from the highly specialized growth or germination processes observed in pollen and seeds. In flowers, tissue desiccation involving aquaporins at various stages is required during reproductive growth. For example, dehydration of anthers is necessary for dehiscence and the release of mature pollen and is hindered by reduced expression of PIP2s in tobacco plants [69]. The maturation of pollen grains is accompanied by gradual dehydration, and their germination is induced by the rapid growth of pollen tubes, which also involves water transduction via PIPs. Previous research showed that after mutating two pollen TIPs specific to vegetative and sperm cells, Arabidopsis showed reduced fecundity in the presence of a limited water or nutrient supply [68].
According to our research, the number of viable pollen grains per anther in the pip1smutants was significantly less than the WT plants, potentially because the mutation of PIP1s affects the normal dehydration and maturation process of pollen grains. Additionally, the pollen from both the WT and the pip1s -mutants had regular honeycomb outer wall structures. However, pollen outer wall of the mutant was deformed and shrunken, which resulted in the impaired pollen-stigma binding ability of the pip1s -mutant compared to the WT plants. Additionally, the transcriptional sequence analysis showed that

Mutation of PIP1s Influences the Reproductive Growth of Arabidopsis
Two desiccated forms of higher plant life, pollen and seeds, play an important role in the plant life cycle. Pollen and seeds express TIP5 and TIP3 specific aquaporin subclasses, respectively [66][67][68]. This specificity may result from the highly specialized growth or germination processes observed in pollen and seeds. In flowers, tissue desiccation involving aquaporins at various stages is required during reproductive growth. For example, dehydration of anthers is necessary for dehiscence and the release of mature pollen and is hindered by reduced expression of PIP2s in tobacco plants [69]. The maturation of pollen grains is accompanied by gradual dehydration, and their germination is induced by the rapid growth of pollen tubes, which also involves water transduction via PIPs. Previous research showed that after mutating two pollen TIPs specific to vegetative and sperm cells, Arabidopsis showed reduced fecundity in the presence of a limited water or nutrient supply [68].
According to our research, the number of viable pollen grains per anther in the pip1s − mutants was significantly less than the WT plants, potentially because the mutation of PIP1s affects the normal dehydration and maturation process of pollen grains. Additionally, the pollen from both the WT and the pip1s − mutants had regular honeycomb outer wall structures. However, pollen outer wall of the mutant was deformed and shrunken, which resulted in the impaired pollen-stigma binding ability of the pip1s − mutant compared to the WT plants. Additionally, the transcriptional sequence analysis showed that some key DEGs, including SYN3, KIN7A, PIN5, PS1, CYP704B1, CYP703A2, CALS7 and CALS8, were involved in exine formation and pollen development ( Figure 11).
In addition, the pollen germination rate of the pip1s − mutant was significantly lower than the WT plants in vitro and in vivo. Mutations in PIP1s also regulate the expressions of many genes related to pollen tube growth and may block pollen grains from absorbing water and germinating ( Figure 11).
The pip1s − mutants exhibited more undeveloped siliques, shorter siliques and fewer seeds. The lower numbers of siliques and seeds in the pip1s − mutant may be partly attributed to the withering of many flower buds of pip1s − before fertilization. Compared the developed siliques of the WT and pip1s − mutant plants, we found that the mutant siliques were nicked. In addition, compared to the WT plants, some aborted seeds and malformed seeds were observed in the pip1s − mutants. In summary, the change in pollen morphology and the reduced pollen viability may lead to a decrease in pollen adhesion and germination rate, that the siliques of the pip1s − mutants were shorter and nicked, finally leading to a significant decrease in the yield of the mutant.
The KEGG analysis showed that plant hormone signal transduction, phenylpropanoids and flavonoid biosynthesis, pentose and glucuronate interconversions, starch and sucrose metabolism and cutin, suberin and wax biosynthesis were significantly altered in pip1s − mutant siliques compared to WT siliques. The GO analysis of DEGs in the pip1s − mutant compared to the WT plants revealed that they were mainly related to metabolic processes, cellular processes, responses to stimulus and biological regulation, suggesting that PIP1s may be involved in regulating growth and development regulation and the stress response in plants. TF genes accounted for a large proportion of the DEGs identified in the pip1s − and consisted of many TF families, such as TCP, AP2, B3, MYB, NAC, NAM, WRKY, B-box domain proteins and zinc finger proteins. Most of the TF genes were related to cell proliferation, cell division, leaf development, leaf senescence, flower development, pollen development, pollen maturation, pollen sperm cell differentiation, pollen tube guidance, embryo sac development and embryo development ending in seed dormancy, suggesting that these TF genes play various important roles in plant growth and development [54][55][56].

Plant Materials and Growth Conditions
Arabidopsis ecotype Columbia (Col-0) was used as the WT control in the present study. The triple T-DNA insertion mutant used in this study was obtained through the hybridization method. Seeds were sterilized with 0.1% (w/v) HgCl 2 for 10 min, washed five times with sterile water, sown on Murashige and Skoog (MS) medium [3% (w/v) sucrose, 0.7% (w/v) agar] and vernalized at 4 • C for 3 days in the dark. Then, 10-day-old seedlings were transferred to pots filled with a mixture of soil and sand (3:1) and grown in a chamber set at 22 • C; 110 µmol·m −2 ·s −1 light intensity; 16-h light/8-h dark cycle; and 70% relative humidity.

Generation of the pip1s − Mutant
To generate the pip1s − -1 and pip1s − -2 mutant, sgRNA targets in the PIP1;4 and PIP1;5 genes were selected and cloned into the pVK004-15 vector (Viewsolid biotech, Beijing, China), and then were transformed into the triple mutant by floral dip method. The transgenic T1 seeds were collected and screened on 1/2 medium containing 50 mg/L hygromycin. The fragments covering the mutation sites were amplified from the T1 transgenic lines by PCR and sequenced to identify the successfully mutated ones. The homozygous mutants were screened from the T2 generation and the seeds were harvested from individual lines to obtain T3 plants, of which non-hygromycin resistant plants were obtained.

Identification of the pip1s − T-DNA Mutant
Homozygous T-DNA insertional mutant plants were confirmed by conducting two consecutive PCR assays. The first assay involved the use of two gene-specific primers: LP and RP. The second assay used one gene-specific primer, RP, and one T-DNA-specific primer (LB).

High-Throughput mRNA Sequencing Analysis
Total RNA was extracted from thirty-day-old rosette leaves and 14 DAF siliques and 3 µg of RNA from each sample were used for library construction and subsequent RNAdeep sequencing on the Illumina HiSeq 2500 platform. RNA-seq data were collected from two independent experiments. The adaptor sequences and low quality sequences were removed. Approximately 4.0 GB of clean reads were generated from each sample. The clean reads were mapped to the Arabidopsis reference genome using TopHat with TAIR10 gene annotation as the transcript index. The minimum and maximum intron lengths were set to 40 and 5000, respectively. Cufflinks was used to assemble the new transcripts. HTSeq was used to calculate the raw read counts for each gene. Gene expression was normalized among samples using DESeq. The differental gene expression data were collected from the comparison with a fold change ≥2 and a false discovery rate of 0.01.

RNA Extraction and qRT-PCR
Total RNA was extracted from leaves and siliques at different developmental stages using the RNeasy Plant Mini Kit (Qiagen, Amsterdam, The Netherlands). Total RNA (1 µg) from each sample was converted into cDNA by reverse transcription using the RNA PCR Kit (TaKaRa, Dalian, China) according to the manufacturer's instructions. qRT-PCR was conducted on an ABI 7500 system (Applied Biosystems, New York, NY, USA) using the TransStart™ Green qRT-PCR SuperMix Kit (TransGen, Beijing, China). Actin2 was used as a reference gene to normalize the relative transcriptional abundance and to minimize differences in the copy numbers of cDNA templates. The control sample was conferred a value of 1. All data were calculated and analysed from three independent samples based on the 2 −∆∆Ct method.

Statistical Analysis
The data are presented as means ± SD and were compared using SPSS software with one-way ANOVA followed by Duncan's multiple range test at a significance level of p < 0.05.

Conclusions
The pip1s − quintuple mutant displayed severe growth defects in rosette leaves, flowers, siliques and seeds. Compared to the WT plant, the pip1s − mutants showed smaller and fewer rosette leaves, smaller flowers, shorter silique and fewer seeds under physiological conditions. Further studies on the pollens exhibited that pollen exine shape was abnormal and pollen vitality and germination rate are significantly reduced in the pip1s − mutants. In summary, PIP1s play a very important role in plant vegetative growth and reproductive growth and are important potential genetic resources in agronomic and crop science. Our research also provides a theoretical basis for agricultural production to improve crop traits and yield.