GhZFP1, a novel CCCH-type zinc finger protein from cotton, enhances salt stress tolerance and fungal disease resistance in transgenic tobacco by interacting with GZIRD21A and GZIPR5


Author for correspondence:
Cheng-Chao Zheng
Tel: +86 538 8242894


  • • Zinc finger proteins are a superfamily involved in many aspects of plant growth and development. However, CCCH-type zinc finger proteins involved in plant stress tolerance are poorly understood.
  • • A cDNA clone designated Gossypium hirsutum zinc finger protein 1 (GhZFP1), which encodes a novel CCCH-type zinc finger protein, was isolated from a salt-induced cotton (G. hirsutum) cDNA library using differential hybridization screening and further studied in transgenic tobacco Nicotiana tabacum cv. NC89. Using yeast two-hybrid screening (Y2H), proteins GZIRD21A (GhZFP1 interacting and responsive to dehydration protein 21A) and GZIPR5 (GhZFP1 interacting and pathogenesis-related protein 5), which interacted with GhZFP1, were isolated.
  • • GhZFP1 contains two typical zinc finger motifs (Cx8Cx5Cx3H and Cx5Cx4Cx3H), a putative nuclear export sequence (NES) and a potential nuclear localization signal (NLS). Transient expression analysis using a GhZFP1::GFP fusion gene in onion epidermal cells indicated a nuclear localization for GhZFP1. RNA blot analysis showed that the GhZFP1 transcript was induced by salt (NaCl), drought and salicylic acid (SA). The regions in GhZFP1 that interact with GZIRD21A and GZIPR5 were identified using truncation mutations.
  • • Overexpression of GhZFP1 in transgenic tobacco enhanced tolerance to salt stress and resistance to Rhizoctonia solani. Therefore, it appears that GhZFP1 might be involved as an important regulator in plant responses to abiotic and biotic stresses.


High salinity and fungal pathogen attack are major stresses that are commonly encountered by plants growing in their native environments. Upon exposure to these stresses, many genes are induced and their products either directly protect plants against stresses or further control the expression of other target genes (Kamalay & Goldberg, 1980; Skriver & Mundy, 1990; Waditee et al., 2002; Gao et al., 2003), including those encoding transcription factors, enzymes, molecular chaperons, ion channels and transporters, or alter the activities of their products (Maathuis et al., 2003; Miao et al., 2006; Sakuma et al., 2006). To understand the process of development in plants and their responses to environmental stresses, it is imperative to know about the function of crucial genes and their regulation during different stress responses (Cui et al., 2002; Ronald & Leung, 2002).

Zinc finger proteins are a superfamily involved in many aspects of plant growth and development. They have been classified into nine types according to their structural and functional diversities: C2H2, C8, C6, C3HC4, C2HC, C2HC5, C4, C4HC3 and CCCH (C and H represent cysteine and histidine, respectively) (Berg & Shi, 1996; Takatsuji, 1998; Moore & Ullman, 2003; Jenkins et al., 2005; Schumann et al., 2007). Many zinc finger proteins have been confirmed to be involved in abiotic and biotic stresses (Sakamoto et al., 2004; Ham et al., 2006; Oh et al., 2006; Ciftci-Yilmaz et al., 2007). The C2H2-type zinc finger protein Zat12 (responsive to high light zinc finger protein 12 of Arabidopsis thaliana) plays a central role in reactive oxygen and abiotic stress signalling in Arabidopsis (Davletova et al., 2005). Overexpression of the zinc finger protein gene OsISAP1 (Oryza sativa subspecies indica stress-associated protein gene 1) from rice (Oryza sativa) confers tolerance to cold, dehydration and salt stress in transgenic tobacco (Nicotiana tabacum var. Xanthi) (Mukhopadhyay et al., 2004). Among the different types of zinc finger proteins, those containing a tandem zinc finger domain characterized by three cysteines followed by one histidine are poorly characterized (Varnum et al., 1991; Blackshear, 2002). Until now, only a few of the CCCH-type proteins associated with RNA metabolism have been characterized in eukaryotes (Shimada et al., 2002; Brown, 2005). In animals, the best known CCCH-type zinc finger protein is tristetraprolin (TTP) in mice, which binds to AU-rich elements in the 3′ untranslated regions of tumour necrosis factor-α and granulocyte–macrophage colony-stimulating factor mRNAs for degradation (Carballo et al., 1998, 2000). Another protein, ZAP (Zinc-finger Antiviral Protein), isolated from Rat-2 fibroblasts was shown to inhibit retroviral RNA production (Gao et al., 2002). In plants, PEI1 (an Arabidopsis Embryo-Specific Zinc Finger Protein gene 1), HUA1 (a nuclear RNA binding protein that is regulating of stamen and carpel), FES1 (FRIGIDA-ESSENTIAL 1), OsDOS (Oryza sativa delay of the onset of senescence) and AtCPSF30 (Arabidopsis thaliana gene encodes the 30-kD subunit of the cleavage and polyadenylation specificity factor) were shown to be required for heart-stage embryo formation, floral reproductive organ identity, promotion of winter-annual habit, delay of leaf senescence and calmodulin-mediated RNA processing in Arabidopsis, respectively (Li & Thomas, 1998; Li et al., 2001; Schmitz et al., 2005; Delaney et al., 2006; Kong et al., 2006). Recently, the zinc finger proteins AtSZF (salt stress-inducible zinc finger protein 1) 1 and AtSZF2 were revealed to be involved in the regulation of salt responses in Arabidopsis (Sun et al., 2007). To our knowledge, only two predicted CCCH-type zinc finger proteins that confer tolerance to stresses in transgenic plants have been identified.

Cotton (Gossypium hirsutum) is one of the oldest and most important fibre and oil crops. Its growth and yield are severely inhibited in high-salinity conditions, especially at the germination and emergence stages (Timpa et al., 1986; Garratt et al., 2002; Burke, 2007). Our attention has therefore been focused on identifying genes induced by salt stress. To this end, a cDNA library of salt-induced seedlings of ZMS19, a salt-tolerant cotton cultivar, was constructed and screened using the differential hybridization method. In this paper, we present the isolation and functional characterization of a novel CCCH-type zinc finger protein designated Gossypium hirsutum zinc finger protein 1 (GhZFP1) in cotton. We provide evidence that GhZFP1 plays important roles in improving biotic and abiotic stress tolerance in plants.

Materials and Methods

Plant material and treatments

Seeds of cotton (Gossypium hirsutum L. cultivar ZMS19) were provided by the Chinese Academy of Agricultural Sciences. Seedlings were grown in a growth chamber for 10 d under a 16-h light : 8-h dark photoperiod, with 300 µmol m−2 s−1 light intensity and a day:night temperature of 25°C. For the low-temperature treatment, uniformly developed seedlings were treated at 4°C for given time periods. For other treatments, uniformly developed seedlings were cultured in solutions containing 100 µM abscisic acid (ABA), 5 mM salicylic acid (SA), 1 mM ethephon, 100 mM CaCl2, 300 µM CuSO4, 25% polyethylene glycol (PEG) (Verslues et al., 2006) and 200 mM NaCl for given time periods (Fig. 3). The cotyledons, leaves, stems and roots were harvested directly into liquid nitrogen and stored at −80°C for later use.

Figure 3.

RNA gel blot analysis of Gossypium hirsutum zinc finger protein 1 (GhZFP1) gene expression induced by abiotic and biotic stresses in cotton (G. hirsutum). (a) Expression patterns of the GhZFP1 gene induced under different treatments. Total RNA was prepared from cotton seedling leaves treated with distilled water, 300 µM CuSO4, 100 µM abscisic acid (ABA), 4°C, 100 mM CaCl2, 200 mM NaCl, 25% polyethylene glycol (PEG) (drought), 1 mM ethephon or 5 mM salicylic acid (SA) for 12 h. (b–d) Time-course accumulation of GhZFP1 mRNA following water, NaCl, drought or SA treatment. (e) Organ-specific expression of the GhZFP1 gene. In each lane, 20 µg of total RNA prepared from roots, stems and leaves was loaded. Ethidium bromide staining of the RNA gel was used to show equal loading.

cDNA library construction and screening

Poly(A) RNA (0.5 µg) isolated from cotyledons of ZMS19 seedlings treated with 200 mM NaCl for 12 h was used to synthesize first-strand cDNA, which was then amplified by long-distance PCR according to the manufacturer's protocol (SMARTTM cDNA Library Construction Kit; Clontech, Palo Alto, CA, USA). The original cDNA library was estimated to contain 5 × 105 independent recombinants and was amplified with Escherichia coli XL-Blue before screening. The cDNA library was screened by differential hybridization (one hybridization with an NaCl-untreated cotyledon cDNA probe, and another hybridization with an NaCl-treated cotyledon cDNA probe) as described previously (Wu et al., 2004). One cDNA clone, GhZFP1, is described in this report.

DNA sequence analysis

DNA sequencing was performed using the dideoxy chain termination method at the Huada Gene Research Center (Beijing, China) using M13 universal primers. The nucleotide sequences were determined from overlapping clones; 100% of the sequence was determined from both strands. Nucleotide and predicted amino acid sequence analysis and database homology searching were performed using the NCBI (National Center for Biotechnology Information) programs. All constructs were confirmed by DNA sequencing.

Subcellular localization prediction was performed using SubLoc v1.0 (Institute of Bioinformatics of Tsinghua University, Beijing, China). Nuclear export sequence (NES) or other sequence motifs were identified by multiple sequence alignment performed with CLUSTALX (Wang et al., 2008). Evolutionary distances were calculated by the neighbour joining method, and the phylogenetic tree was drawn with the program mega3.1 (Center for Evolutionary Functional Genomics of the Biodesign Institute, Tempe, AZ, USA).

RNA blot hybridization

Total RNA was extracted from plant tissue using an RNeasy Plant Mini Kit (Qiagen) according to the manufacturer's instructions. 20 µg of total RNA per sample was separated by electrophoresis on 1.2% agarose gels containing 6% formaldehyde and blotted onto nylon membranes (Hybond-N; Amersham Pharmacia Biotech). The procedure of hybridization was performed using the same method as described for cDNA library screening. The GhZFP1 3′-specific cDNA labelled with [α-32P] dCTP by Priming, a gene labelling system from Promega, was used for the hybridization probe. The blots were autoradiographed at −80°C for up to 7 d. The ethidium bromide-stained rRNA band in the agarose gel is shown as a loading control.

Subcellular localization of GhZFP1

Plasmid pBI121-GhZFP1-GFP (green fluorescent protein) driven by the CaMV35S (cauliflower mosaic virus) promoter was constructed to investigate the subcellular localization of GhZFP1 in onion (Allium cepa) epidermal cells. The termination codon of the GhZFP1 cDNA was removed after PCR using a specific oligonucleotide primer and fused to the N terminus of GFP. Agrobacterium tumefaciens strain LBA4404 carrying pBI121-GFP and pBI121-GhZFP1-GFP expression vectors was cultured from a single colony in YEP (yeast extract peptone) medium. Transient expression assays were performed using the A. tumefaciens-mediated transformation method as described previously (Zhao et al., 2005). The onion epidermal cells were visualized with a fluorescence microscope (BX51, model 7.3; Olympus, Japan).

Transformation of tobacco

To overexpress cotton GhZFP1 in tobacco (Nicotiana tabacum cv. NC89), the cDNA GhZFP1 was cloned into a pBI121-based binary Ti empty vector. The resulting vector, pBI121-GhZFP1, has the CaMV35S promoter driving expression of the GhZFP1 gene. This construct was transferred into A. tumefaciens strain LBA4404 and then transformed into tobacco via the leaf-disc method as described previously (Zhang et al., 2007). The integration of the transgene into different transgenic lines was confirmed by PCR and RNA blot analysis.

Analysis of T2 transgenic tobacco for salt stress tolerance

Fifteen-day-old wild-type (WT) and homozygous T2 transgenic seedlings were transferred to half-strength Murashige and Skoog medium (MSH) supplemented with 100, 200 or 300 mM NaCl, under culture room conditions (16 h light : 8 h dark; 300 µmol m−2 s−1 light intensity; 25°C). The phenotype of seedlings was photographed after 30 d of growth. After 60 d of growth, the root length and fresh weight of seedlings were measured. Subsequently, the seedlings of T2 transgenic tobacco that had been exposed to 200 mM NaCl were washed briefly in sterile Milli-Q water, and then allowed to grow on MSH without salt stress for a 15-d recovery period.

Sixty-day-old WT and T2 transgenic seedlings grown in soil were watered with 0, 100, 200 or 300 mM NaCl solution every other day for 4 wk. The experiment was repeated at least three times with different transgenic lines and > 10 independent lines were used in each experiment.

The ion content and chlorophyll fluorescence of seedlings treated with 200 mM NaCl were measured using an atomic absorption spectrophotometer (Z-8000; Hitachi) and FMS2 fluorescence (Hansatech, Cambridge, UK), respectively. The shoots and roots of WT and T2 transgenic seedlings salt-treated for 30 d were used to determine K+ and Na+ contents, respectively. Fv/Fm (the maximum photochemical efficiency of photosystem II in the dark-adapted state) values were determined for the two youngest fully expanded leaves from the transgenic and WT plants, which were at the same developmental stage after exposure to 200 mM NaCl for 3, 6 and 9 d. Leaves were kept in darkness for 30 min before the fluorescence parameters were recorded. Each data point represents the mean ± SE of triplicate experiments (n = 6). All experiments were repeated at least three times, and data in the form of the mean of three experiments with absolute variation are given.

Yeast two-hybrid assays

The full-length encoding region, a fragment carrying residues 1–237 (mN237/m1) of GhZFP1 and a fragment carrying residues 238–339 (mC102) of GhZFP1 were fused to the plasmid pGBKT7 and transformed into yeast strain AH109 with the empty pGADT7-Rec vector. The transformants of GhZFP1 and mC102 with the empty pGADT7-Rec vector can grow on SD/-Trp-Leu-His-Ade medium, and therefore the full-length GhZFP1 and mC102 were not subjected to the yeast two-hybrid screening (Y2H) assay. We used mN237 as bait to screen a salt-induced cotton seedling cDNA library constructed in the pGADT7-Rec vector. The cDNA library was transformed into yeast strain AH109 with mN237. Y2H assays were performed according to standard methods (Chien et al., 1991). Positive colonies were selected on SD/-Trp-Leu-His-Ade medium. After confirmation using the 5-bromo-4-chloro-3-indoyl-alpha-D-galactoside (X-α-Gal) test and retransformation, the inserts were sequenced. Moreover, pGADT7-Rec, pGADT7-GZIRD21A and pGADT7-GZIPR5 were transformed into yeast strain AH109 with the empty pGBDT7 vector as controls. The expression of the third reporter gene lacZ was followed by measuring the accumulation of the product metabolized by β-galactosidase with o-nitrophenyl β-D-galactopyranoside (o-NPG; Sigma) as substrate at OD420.

Bimolecular fluorescence complementation (BiFC)

For BiFC studies, the cDNA without a termination codon encoding mN237 was cloned into pSPYNE-35S and the cDNAs encoding wild-type GZIRD21A (GhZFP1 interacting and responsive to dehydration protein 21A) and GZIPR5 (GhZFP1 interacting and pathogenesis-related protein 5) were cloned into pSPYCE-35S, respectively. Both the cDNAs encoding pwHAP5 (pwNF-YC/pwCBF-C) in pSPYCE-35S and the empty vector 35S-pSPYCE were used as negative controls, and the bZIP63-pSPYNE-35S and bZIP63-pSPYCE-35S vectors were used as a positive control (Walter et al., 2004). These vectors were introduced into the A. tumefaciens strain GV3101.

For infiltration of Nicotiana benthamiana, the p19 protein of tomato bushy stunt virus was used to suppress gene silencing. The A. tumefaciens strains were grown overnight in YEB media, containing appropriate antibiotic selections. Cells were pelleted at 4000 g and resuspended in infiltration medium (10 mM MgCl2, 10 mM MES, and 150 mM acetosyringone) and incubated for at least 2 h at room temperature. Co-infiltration of A. tumefaciens strains containing the BiFC constructs and the p19 silencing plasmid was carried out at OD600 of 0.7 : 0.7 : 1.0. Resuspended cells were infiltrated into leaves of 4-wk-old N. benthamiana plants as described previously (Voinnet et al., 2003; Walter et al., 2004). Epidermal cell layers of tobacco leaves were assayed for fluorescence with a fluorescence microscope (BX51, model 7.3; Olympus) after 2 d.

Construction of various GhZFP1 deletion mutants and quantitative β-galactosidase activity assays

The deletion mutants of GhZFP1 were generated by PCR and fused to the plasmid pGBKT7. The deletions m2–8 encode GhZFP1 abridged proteins with 1–79, 80–237, 1–40, 41–79, 80–185, 186–237 and 41–237 amino acids, respectively. The fusion plasmids were transformed into yeast strain AH109 with pGADT7-Rec, pGADT7-GZIRD21A and pGADT7-GZIPR5, respectively. The transformants were selected by growth on SD/-Trp-Leu-His-Ade medium at 30°C for 3 d. The liquid culture assay using o-NPG as a substrate was performed to determine the β-galactosidase activity for the different deletions and AD vectors.

Rhizoctonia solani resistance test

The fungal pathogen Rhizoctonia solani AG-4 was provided and identified by the College of Plant Protection of Shandong Agricultural University. Cultures of R. solani AG-4 were incubated in the dark at 30°C for 48 h on potato dextrose agar (PDA) plates and maintained at 23°C for 2 wk before use in the experiment. Fungal cultures grown on PDA plates were homogenized and suspended in sterile water and mixed with sterile soil (five plates for 3 l of soil) (Zoubenko et al., 1997). Eighty independent seedlings of 40-d-old transgenic and WT tobacco were transplanted into soil inoculated with R. solani AG-4. Plants were maintained under the same conditions as those before inoculation except that the relative humidity was increased to 99%. The development of disease symptoms was observed for 20 d and the seedling mortality was calculated.


Isolation and characterization of GhZFP1

A cDNA clone, GhZFP1 (NCBI accession number: AY887895), was isolated from a cotton cotyledons by differential hybridization screening to identify genes involved in salt tolerance. The full-length sequence of GhZFP1 consisted of 1377 nucleotides, encoding a polypeptide of 339 amino acids with a molecular weight of c. 38.4 kDa. Database searches also revealed that it contained two typical CX8CX5CX3H and CX5CX4CX3H zinc finger motifs spaced 18 amino acids apart (Fig. 1a) and belonged to an unusual CCCH-type zinc finger protein family. Moreover, we found a potential nuclear localization signal (NLS; between amino acids 14 and 28) at the N-terminal region and a putative leucine-rich nuclear export sequence (NES; between amino acids 331 and 339) (Fig. 1c) at the C-terminal region of the protein (la Cour et al., 2003; Fujita et al., 2007; Hawkins et al., 2007). These data suggest that the novel CCCH-type zinc finger protein GhZFP1 might function as a regulator in cotton.

Figure 1.

Gossypium hirsutum zinc finger protein 1 (GhZFP1) is predicted to encode a CCCH-type zinc finger protein. (a) Gene structure analysis of GhZFP1. The grey box indicates the potential nuclear localization signal (NLS), black boxes indicate the predicted zinc finger motifs (ZF1 and ZF2) and the hatched box indicates the nuclear export sequence (NES). Numbers indicate the positions of the initial amino acid of the NLS, ZF1, ZF2 and NES motifs, respectively. (b) Alignment of the zinc finger domain of GhZFP1 with several putative or well-known proteins. The conserved cysteine (C) and histidine (H) residues that may form two zinc fingers in the CCCH-type proteins are shown. The gene names, accession numbers and sources of zinc finger proteins are as follows: AT1G03790 and AT5G07500 (PEI1 (an Arabidopsis Embryo-Specific Zinc Finger Protein gene 1)), Arabidopsis thaliana; Os01G53650, Os05G45020 and Os01G09620 (OsDOS (Oryza sativa delay of the onset of senescence)), Oryza sativa; X52590 (CMG1 (mitogen-induced gene)), rat; P26651 (tristetraprolin (TTP)) and X78992 (TIS11d), Homo sapiens. (c) A comparison of the NES-like sequence of the GhZFP1 protein with those of the stress-responsive zinc finger protein (SRZFP) and TTP families. These alignments show the spatial conservation of the four key hydrophobic residues of the leucine-rich type NES (shaded). The accession numbers for the proteins listed are as follows: GhZFP1: AY887895, AT1G03790, AT2G19810, AT5G44260, AT2G40140, AT5G55980, AT5G58620, AT5G07500 and AT4G29190; HumTTP: P26651; RatCMG1: X52590; MouIRF-3: AF036341; and HumE2F4: XM_054397. (d) Subcellular localization of the GhZFP1 protein in onion epidermal cells. The GhZFP1::GFP fusion gene and the green fluorescent protein (GFP) gene were fused in the plant transformation vector pBI121. Expression of the introduced genes was examined after 48 h by fluorescence and light microscopy. Bar, 10 µm.

A more detailed comparison of GhZFP1 zinc finger sequences with several CCCH proteins from other organisms revealed a high degree of conservation of the cysteine and histidine residues of the CCCH motifs, whereas the spacing between the cysteine and histidine residues and the length of the linker between the two zinc finger motifs were variable (Fig. 1b). Further analysis of the phylogenetic relationship between GhZFP1 and CCCH-type zinc finger proteins in Arabidopsis revealed that 11 zinc finger proteins and GhZFP1 were significantly clustered together on a single branch of the tree; in particular, GhZFP1 had a high degree of similarity to AT1G03790 and AT5G44260 (Fig. 2). Structural analysis of the GhZFP1 protein demonstrated that it possesses most of the characteristics of this subfamily. These features, together with our recent analysis (Wang et al., 2008), suggest that the GhZFP1 protein belongs to a putative CCCH-type zinc finger protein subfamily involved in responses to abiotic and/or biotic stress. We thus designated this subfamily the stress-responsive zinc finger protein (SRZFP) subfamily.

Figure 2.

Phylogenetic analysis of Gossypium hirsutum zinc finger protein 1 (GhZFP1) and CCCH-type zinc finger proteins in Arabidopsis. The accession numbers of zinc finger protein genes and GhZFP1 are shown in the phylogenetic tree. The vertical line indicates the 11 members of the stress-responsive zinc finger protein (SRZFP) subfamily.

The N-terminal region of the putative GhZFP1 protein contained a potential NLS sequence, KKX10RKK, that resembled an Arabidopsis DREB1A (dehydration-responsive element binding protein 1A) NLS (Liu et al., 1998). To examine the subcellular location of the GhZFP1 protein in plant cells, we performed an in vivo targeting experiment using green fluorescent protein (GFP) as a marker (Koroleva et al., 2005; Dixit et al., 2006; Shaw, 2006). As shown in Fig. 1(d), the fusion protein GhZFP1::GFP was localized to the nucleus of onion epidermal cells, whereas the control GFP was uniformly distributed throughout the cell, indicating that GhZFP1 encodes a nuclear-localized protein.

Stress-induced expression of GhZFP1 in cotton

To examine the expression of GhZFP1 under stress, RNA gel blot analysis was performed. GhZFP1 mRNA in cotton seedlings was not detected under normal growth conditions but was significantly up-regulated following treatment with NaCl, PEG (drought) and exogenous SA. However, there was no significant accumulation of GhZFP1 mRNA in seedlings treated with cold, CuSO4, CaCl2, ABA and ethephon (Fig. 3a), indicating that the expression of GhZFP1 was only induced by some abiotic and biotic stresses. Expression of GhZFP1 in response to NaCl, PEG and SA was further analysed to investigate time-dependent induction patterns. As shown in Fig. 3(b,d), the GhZFP1 transcripts accumulated within 6 h following NaCl and SA treatment, and were strongly expressed after 12 h. In comparison, expression of the GhZFP1 gene was induced within 6 h, peaked at 12 h and then decreased at 24 h after PEG treatment (Fig. 3c). To determine the organ-specific expression of the GhZFP1 gene, we also assayed RNA from the roots, stems and leaves of the cotton seedlings. RNA gel blot analysis showed that the expression of GhZFP1 was up-regulated following NaCl stress in both leaves and stems, but was not detected in roots (Fig. 3e).

Overexpression of GhZFP1 in T2 transgenic plants improves tolerance to high salinity

To investigate whether overexpression of GhZFP1 in plants enhances salt tolerance, we generated transgenic tobacco plants expressing the GhZFP1 gene (Fig. 4a). Seedlings at different developmental stages from T2 homozygous lines and WT plants were treated separately with different concentrations of NaCl. When 15-d-old seedlings were transferred to medium supplemented with NaCl, the WT seedlings showed severe chlorosis and stunted phenotypes and ultimately died, whereas the T2 transgenic seedlings continued to grow. Moreover, the growth rates of the T2 and WT seedlings were affected in a dose-dependent manner, and the S2 and S4 seedlings grew better than S1 seedlings as a result of higher GhZFP1 expression (Fig. 4b). When 60-d-old transgenic plants were exposed to a continuous stress consisting of 200 mM NaCl added to the soil for 3 wk, they showed a slightly slower growth rate initially compared with similar plants growing in the absence of stress (Fig. 4c), but they continued to grow, reached maturity, flowered and set seed. However, the growth of WT plants was severely affected and they failed to survive under these conditions (Fig. 4e; representative plants shown). In addition, when 15-d-old seedlings were transferred to medium supplemented with 200 mM NaCl for 60 d followed by 15 d of recovery, a distinct difference was observed for root length and fresh weight. As shown in Fig. 4(d), the transgenic seedlings gained 81% fresh weight, whereas WT seedlings gained only 30% fresh weight compared with unstressed seedlings. The enhanced salt tolerance of the transgenic plants was further confirmed by measuring changes in chlorophyll fluorescence. Reductions in the maximum photochemical efficiency of photosystem II (PSII) in the dark-adapted state (Fv/Fm) were considerably larger in WT plants than in transgenic plants throughout the time course (Fig. 4f), thereby confirming the increased tolerance to salt stress.

Figure 4.

RNA gel blot analysis and representative image showing the relative salinity tolerances of wild-type (WT) and T2 transgenic tobacco (Nicotiana benthamiana) plants. (a) RNA gel blot analysis of WT and transgenic T2 tobacco plants. S1–S6 are six lines of T2 plants. (b) Comparative images of 15-d-old WT and T2 transgenic (S1, S2 and S4) seedlings growing on medium supplemented with 0, 100, 200 and 300 mM NaCl for 30 d. (c) Comparative images of 60-d-old WT and T2 transgenic (S2) seedlings treated with or without 100, 200 and 300 mM NaCl solution every other day for 4 wk. (d) Root length comparative images of 15-d-old WT and T2 transgenic (S2) seedlings growing on medium supplemented with 200 mM NaCl for 60 d, followed by 15 d of recovery in water. (e) Representative phenotype of WT and T2 transgenic (S2) seedlings growing in soil pots supplied with 200 mM NaCl solution. Note that the WT plant could not sustain growth under salinity stress. (f) The Fv/Fm of T2 transgenic (S2) and WT plants was measured following different times under 200 mM NaCl stress using FMS2 fluorescence. (g) Total K+/ Na+ratio for the shoots and roots of WT and T2 transgenic (S1, S2 and S4) lines growing in 200 mM NaCl. The values represent the average of six independent samples. Error bars indicate ± SE. Within each set of experiments, bars with different letters are significantly different at P < 0.05 according to Duncan's multiple range test.

To examine the mechanism resulting in the phenotype of salt tolerance in lines overexpressing GhZFP1, we measured the Na+ and K+ contents of the T2 and WT plants before and after exposure to NaCl stress, because these ions are known to play important roles in plants under NaCl stress. As expected, NaCl treatment increased cellular Na+ and decreased K+ contents in both types of plant. The Na+content in T2 plants was significantly lower than in WT plants, whereas the K+ content in T2 plants was higher than in WT plants (data not shown). Interestingly, the ratio of K+ to Na+ in transgenic lines was significantly higher than that in WT plants but the difference in this ratio between transgenic and WT plants was smaller in roots than in shoots (Fig. 4g). These results indicated that the increased salt tolerance of transgenic plants was probably a result of the ability of Na+homeostasis or K+ acquisition.

The GhZFP1 protein interacts with GZIRD21A and GZIPR5

We used Y2H screening to identify proteins that interact with GhZFP1, anticipating that such interactions might provide a new insight into the function of the GhZFP1 protein. For this purpose, the mN237/m1 was used as bait to screen a salt-induced cDNA library from cotton leaves (Fig. 5a). We screened 6 × 106 colonies and identified 35 positive clones corresponding to eight cDNAs (data not shown). Five cDNA clones (1, 7, 15, 17 and 27) encoded the same overlapping sequence, with clone 27 having the longest sequence. Another three cDNA clones (9, 19 and 22) encoded the same peptide. The database search using the deduced amino acid sequence showed that they shared high sequence similarities to RD21A (responsive to dehydration protein 21A) and PR5 (pathogenesis-related protein 5) in other plant species, respectively. The former belongs to a cysteine-type proteinase family responding to water deprivation, and the latter belongs to the pathogenesis-related thaumatin/PR5-like family of proteins, which function as part of the plant defence system; we thus designated clones 27 and 9 as GZIRD21A and GZIPR5, respectively. Protein interactions between mN237 and GZIRD21A or GZIPR5 were further confirmed by analysing growth on selective medium, followed by measuring the true β-galactosidase activities. Growth was observed for both mN237-GZIRD21A and mN237-GZIPR5 combinations, but no growth was observed for the control combinations (Fig. 5a). The β-galactosidase activities of the mN237 fusion proteins were 10 times higher than those of the controls (Fig. 5b). These results confirmed the interaction between the mN237 and GZIRD21A or GZIPR5 proteins.

Figure 5.

Yeast two-hybrid (Y2H) assays and analysis of GZIRD21A (GhZFP1 interacting and responsive to dehydration protein 21A) and GZIPR5 (GhZFP1 interacting and pathogenesis-related protein 5) expression. (a) The interactions were assayed in the GAL4 (a regulator of galactose-induced genes) Y2H system to retest the interactions of GZIRD21A and GZIPR5 proteins with the pGBDKT7 vector alone (BD) and the N-terminal 237 amino acid domain of Gossypium hirsutum zinc finger protein 1 (GhZFP1) (mN237/m1), monitored by growth on medium-selective plates (SD/-Leu-Trp-His-Ade). As a control, combinations for the pGADT7-Rec vector (AD) with the pGBDKT7 vector (BD), the full-length GhZFP1 (GhZFP1), mC102 and mN237 are also shown. (b) Liquid β-galactosidase assay using 2-nitrophenyl β-D-galactopyranoside o-NPG as a substrate. β-galactosidase activity is expressed in U (= nmol min−1). The values displayed are the average β-galactosidase activities for three individual double transformants, with standard deviations indicated by error bars. (c) Expression patterns for GZIRD21A and GZIPR5 induced following treatment with 200 mM NaCl, 25% polyethylene glycol (PEG) and 5 mM salicylic acid (SA) for the designated times. Total RNA (20 µg) prepared from cotton ZMS19 leaves was loaded. The blotted membrane was hybridized with the GZIRD21A and GZIPR5 probes, respectively. (d) Bimolecular fluorescence complementation (BiFC) visualization in Agrobacterium tumefaciens-infiltrated tobacco (Nicotiana benthamiana) leaves. GhZFP1 interacts with GZIRD21A and GZIPR5 in the cytoplasm and the nucleus, and does not interact with the negative control pwHAP5 or empty vector. The yellow fluorescent protein (YFP) fluorescence of the positive control vectors, bZIP63-pSPYNE-35S and bZIP63-pSPYCE-35S, was detected only in the nucleus. Bar, 25 µm.

To confirm the interactions of GhZFP1 with GZIRD21A or GZIPR5 detected in yeast cells, we employed a BiFC system. The results showed that yellow fluorescent protein (YFP) fluorescence was detected throughout the cytoplasm and the nucleus in transformed epidermal cells transfected with mN237-YFPN and GZIRD21A-YFPC or GZIPR5-YFPC, and was especially strong in the nucleus (Fig. 5d). In contrast, cells transfected with mN237-YFPN and empty vectors or pwHAP5-YFPC produced no or only background fluorescence, indicating that GhZFP1 interacts with GZIRD21A and GZIPR5 in both compartments.

To determine the relationship between GhZFP1 and GZIRD21A or GZIPR5, their expression patterns were analysed by RNA blot hybridization. The GZIRD21A transcript was induced very rapidly following treatment with PEG and SA, and the GZIPR5 mRNA was induced by treatment with NaCl and SA in cotton seedlings (Fig. 5c). These results are consistent with the RD21A and PR5 expression profiles identified in previous studies (Singh et al., 1989; Koizumi et al., 1993; Avrova et al., 1999; Beyene et al., 2006; Thompson et al., 2006). These findings, together with the fact that up-regulation of the GhZFP1 transcript was observed under all of the above stresses (Fig. 3a), suggest that the functions of GhZFP1 are associated with GZIRD21A and GZIPR5 in response to abiotic and biotic stresses.

Interaction domains for GZIRD21A and GZIPR5 in GhZFP1

To identify the GhZFP1 binding domains required for interaction with GZIRD21A and GZIPR5, a series of GhZFP1 deletions were generated by PCR and cloned into BD bait vectors, and these constructs were then used for cotransformation with pGADT7-Rec vectors carrying GZIRD21A or GZIPR5 into yeast. By monitoring growth on selective medium and measuring β-galactosidase activities, we observed that mN237 provided the strongest binding activity for both GZIRD21A and GZIPR5. All further mN237 truncations resulted in a reduction or loss of binding activity, indicating that multiple sites within mN237/m1 contribute to their differential interactions with GZIRD21A or GZIPR5 in yeast (Fig. 6). Moreover, m1, m3, m6 and m8, which harbour two zinc finger motifs, showed higher binding strength compared with m2, m5 and m7, suggesting that the zinc finger motif in GhZFP1 is necessary and sufficient for interaction with GZIRD21A or GZIPR5. When comparing m2, m4 and m5, we observed that the binding activity of m4 was significantly stronger than that of m2, and m5 failed to interact with GZIRD21A or GZIPR5, implying that the N-terminal region consisting of amino acids 1–40 (m4) might contain another binding domain and the region consisting of amino acids 41–79 has a negative effect on the interaction of GhZFP1 with GZIRD21A or GZIPR5. In addition, although m7 showed a very low binding activity (< 10%), m3 displayed a much stronger interaction than m6, revealing that the region consisting of amino acids 186–237 plays a positive role in the interaction of GhZFP1 with GZIRD21A or GZIPR5.

Figure 6.

A series of deletions of Gossypium hirsutum zinc finger protein 1 (GhZFP1) significantly changed their ability to interact with GZIRD21A (GhZFP1 interacting and responsive to dehydration protein 21A) and GZIPR5 (GhZFP1 interacting and pathogenesis-related protein 5). A series of deletions of GhZFP1 and cotransformant growth of GZIRD21A or GZIPR5 with deletions of GhZFP1 were shown. Solid grey boxes indicate the DNA-binding domain (BD) in the pGBKT7 vector, and black boxes indicate the predicted zinc finger motifs (ZF1 and ZF2). Numbers in the open boxes represent the beginning and end positions of GhZFP1 protein fragments. Each deletion construct was transformed into the yeast strain AH109 containing pGADT7-GZIRD21A and pGADT7-GZIPR5, respectively. Three individual transformants were used to measure the β-galactosidase activity. Results are shown as the mean ± SEM from three different colonies. No β-galactosidase activity was detected in negative control colonies carrying the empty vector pGADT7-Rec.

Overexpression of GhZFP1 in T2 transgenic tobacco enhances resistance to R. solani

To determine whether overexpression of GhZFP1 enhances resistance to pathogens, we inoculated WT and T2 transgenic tobacco with R. solani, a fungal pathogen that causes damping-off, seedling blight, and stem rot. After seedlings had been transplanted into infested soil, the development of disease symptoms in transgenic and control plants was observed for 20 d and the seedling mortality rate was calculated. A significant difference between WT and transgenic plants was found, as shown in Fig. 7(a): the mortality rate of WT plants inoculated with R. solani reached 78% by day 7, whereas only c. 38% of transgenic seedlings were affected by R. solani. Disease progressed rapidly in WT tobacco plants, which exhibited severe disease symptoms before they died. By contrast, although several lesions developed on the leaves of the transgenic tobacco plants, the lesions were much smaller in size and lower in density than those on WT tobacco plants after 2 wk (Fig. 7b,c). Transgenic lines exhibited a delay in the appearance of disease symptoms and a significantly lower mortality rate compared with controls. These results demonstrated that the enhanced resistance to R. solani may be conferred by GhZFP1 overexpression in transgenic plants.

Figure 7.

Comparison of the mortality rate and phenotype of wild-type (WT) and T2 transgenic tobacco Nicotiana tabacum cv. NC89 seedlings upon infection with Rhizoctonia solani. (a) Seedling mortality rate in WT and T2 transgenic seedlings (S1, S2 and S4) upon infection with Rhizoctonia solani. The percentage of seedling mortality was calculated and plotted against days post-inoculation. Each data point represents the mean ± SE of triplicate experiments (n = 20). Bars with different letters indicate significant differences at each stage at P < 0.05 according to Duncan's multiple range test. (b) Representative seedlings from WT and T2 lines (S2) were photographed after transplantation into R. solani-infected soil for 2 wk. Left, WT seedlings; right, T2 transgenic seedlings (S2). (c) The phenotype of WT and T2 transgenic seedling leaves and stems upon infection with R. solani.


In this study, we isolated a gene, GhZFP1, encoding a novel CCCH-type zinc finger protein from a salt-induced cotton cDNA library using differential hybridization screening. Phylogenetic analysis indicated that GhZFP1 belongs to a poorly studied CCCH-type zinc finger protein subfamily (SRZFP). All members of this SRZFP subfamily have several characters in common, including two typical zinc finger motifs (CX8CX5CX3H and CX5CX4CX3H) and a putative NES-like motif. In addition, they are all intronless genes that have been identified via gene structural analysis (Wang et al., 2008). The putative NES-like amino acid sequences in the SRZFP subfamily showed high similarity to those in mammalian TTP, CMG1 (mitogen-induced gene 1), IRF-3 (interferon regulatory factor 3) and E2F4 (transcription factor 4 of E2 promoter binding factor) zinc finger proteins (Fig. 1c), indicating that this subfamily may act as nuclear export proteins in plants (Phillips et al., 2002). Database searches failed to reveal any proteins corresponding to the SRZFP subfamily in lower plant species, indicating that it may be unique to higher plant species. More recently, two CCCH proteins, AtSZF1 and AtSZF2, have been reported to negatively regulate the expression of salt-responsive genes and play important roles in modulating plant tolerance to salt stress in Arabidopsis (Sun et al., 2007). In the light of the present results for GhZFP1 and the expression profiles of other SRZFP members obtained from a microarray database (Zimmermann et al., 2004; Wang et al., 2008), we propose the hypothesis that members of the SRZFP subfamily might be involved in the response to multiple stresses and may play important roles in different signal transduction pathways. To our knowledge, GhZFP1 is the first CCCH-type zinc finger protein from the SRZFP subfamily to be identified and functionally characterized in cotton. Therefore, more studies exploring the functional diversity of the SRZFP subfamily responses to stresses in plants need to be carried out.

Biotic and abiotic stresses induce overlapping sets of genes in plants (Zhu et al., 1995; Ingram & Bartels, 1996). This suggests the hypothesis that the biotic and abiotic signal pathways may interact to activate or repress biotic and abiotic response genes in plants. Cross-coupling of transcription factors is thought to play an important role in mediating responses to various signalling events. In this study, GhZFP1 was induced following high-salt, drought and SA treatments in cotton seedlings (Fig. 3A), and overexpression of GhZFP1 in transgenic tobacco plants resulted in salt tolerance and resistance to fungal pathogens (Figs 4, 7). These results indicate that the GhZFP1 gene is likely to be regulated by at least two signalling pathways that are activated by high-salt stress and SA, respectively. It will be important to determine whether GhZFP1 proteins can interact with other regulation proteins that are involved in distinct signal transduction pathways.

To elucidate the precise molecular mechanism of GhZFP1 increasing drought tolerance and disease resistance in transgenic plants, two proteins – GZIRD21A and GZIPR5 – that interact with GhZFP1 were isolated and identified by Y2H screening and further confirmed by BiFC. Previous studies demonstrated that RD21A belongs to a cysteine-type proteinase family responding to water deprivation and that PR5 is one class of pathogenesis-related proteins that functions as part of the plant defence system (Koizumi et al., 1993; Stintzi et al., 1993; Wang et al., 1996). Although the roles of the RD21A and PR proteins in plant drought tolerance and disease resistance have not been established, their enzymatic functions indicate that they are well suited to increase drought tolerance and defence against pathogens (Zhu et al., 1995; Ingram & Bartels, 1996; Yoshihiro et al., 2003). Transgenic tobacco plants carrying the 35S::GhZFP1 fusion gene showed tolerance to high-salt stress and fungal pathogens. We also revealed that the induced expression patterns of GZIRD21A and GZIPR5 showed a resemblance to that of GhZFP1 under biotic and abiotic stresses (Fig. 5). This may be a result of the increased expression of stress-inducible genes induced by the overexpression of GhZFP1. The increased abundance of GhZFP1 transcripts probably up-regulated the transcription of several stress-inducible genes, which in turn contributed to increased endurance under the stress conditions. It has been reported that overexpression of the tobacco Tsi1 (Tobacco stress-induced gene 1) gene encoding an EREBP/AP2 (ethylene-responsive element binding protein)-type transcription factor enhanced resistance against pathogen attack and osmotic stress by inducing the expression of several pathogenesis-related genes in tobacco (Park et al., 2001). These results suggest that Tsi1 might be involved as a positive trans-acting factor in two separate signal transduction pathways under abiotic and biotic stresses. The true roles of the RD21A and PR5 genes under abiotic/biotic stresses need to be fully explored using transgenic plants in future studies.

Further analysis of the GhZFP1 binding domains using Y2H assays demonstrated that the two zinc finger motifs and the N-terminal region consisting of amino acids 1–40 are necessary and sufficient for mediating interactions with GZIRD21A or GZIPR5, respectively. A positive domain (amino acids 186–237) and a negative domain (amino acids 41–79) for protein interaction were also identified. The fact that the interaction domains for GZIRD21A and GZIPR5 are similar implies that GhZFP1 may regulate the expression of GZIPR5 and GZIRD21A in a similar or different manner, and subsequently regulate biotic or abiotic stress responses, respectively, in plants. However, GhZFP1 may play multiple roles in different signal transduction pathways relying on different protein–protein interactions via activated or repressed responses. For instance, the extensively studied animal transcription factor CREB (for cAMP response element binding protein) is phosphorylated and activated by protein kinases from a number of different signalling pathways (Hill & Treisman, 1995). Recent studies have suggested that some CCCH-type zinc finger proteins also act as RNA-binding proteins that interact with target RNA involved in many aspects of plant growth and development (Ganss & Jheon, 2004; Hall, 2005). Therefore, it might be also possible that GhZFP1 functions as an RNA-binding protein and regulates the RNA metabolism of stress-induced target genes by transmitting distinct sets of intracellular signals in plants.

However, we have confirmed that GhZFP1 is a nuclear-localized protein, but Arabidopsis RD21A and tobacco PR5, the homologues of GZIRD21A and GZIPR5, were shown to be predominantly cytoplasmic proteins (Stintzi et al., 1993; ). Thus, how can a nuclear-localized protein be exported from the nucleus to the cytoplasm and interact with cytoplasmic proteins? Previous studies have demonstrated that TTP could be rapidly phosphorylated and translocated from the nucleus to the cytoplasm in response to mitogens (Taylor et al., 1996). Deletion and mutation analyses revealed functional nuclear export signals in the amino terminus of TTP and in the carboxyl termini of other members (CMG1and TIS11D) of the TTP family, and these proteins exhibited CRM1 (exportin 1)- and NES sequence-dependent nucleocytoplasmic shuttling (Taylor et al., 1996; Phillips et al., 2002). A comparison of the amino acid sequence of GhZFP1 with that of SRZFP and other members of the TTP family revealed that they show high levels of similarity, not only in the zinc finger domains but also in the putative NES peptides (Fig. 1c). In plants, the formation of the RD19 and PopP2 (Pseudomonas outer protein P2) complex was shown to be essential for the targeting of RD19 and its transfer from mobile vacuole-associated compartments to the nucleus (Bernoux et al., 2008). In addition, in the BiFC assay, interactions of GhZFP1 with GZIRD21A or GZIPR5 were detected in both the cytoplasm and the nucleus (Fig. 5d). These findings led to the reasonable assumption that the GhZFP1 protein might transfer from the nucleus to the cytoplasm and facilitate the formation of higher order complexes by interacting with GZIRD21A and GZIPR5 proteins or carry the complex into the nucleus to activate regulatory pathways under stress, and thus confer tolerance to transgenic tobacco plants. Further genetic and biochemical studies are required to test this assumption.

In conclusion, our results strongly suggest that GhZFP1 acts as a novel positive regulator to confer salt tolerance and fungal pathogen resistance to plants. Elucidation of possible GhZFP1 targets and mechanisms of interaction will help to elucidate how these processes are related. Given the significant functional correlation between GhZFP1 and its interacting proteins, GZIRD21A and GZIPR5, it will also be useful to examine previously undescribed cross-talk between signal transduction pathways in the response to abiotic and biotic stresses in higher plants.


We thank Dr J. Haseloff (MRC Laboratory of Molecular Biology, Cambridge, UK) for the GFP construction pBINmGFP5-ER. This work was supported by the National Basic Research Program (Grant No. 2006CB1001006), the Program for Changjiang Scholars and Innovative Research Team in University (Grant IRT0635) and the Early Stage of Key Development Project for Basic Research (Grant No. 2007CB116208) in China.