Functional analysis of GbAGL1, a D-lineage gene from cotton (Gossypium barbadense)

Abstract

Cotton fibres originate from the outer ovule integument and D-lineage genes are essential for ovule development and their roles can be described by the ‘ABCDE’ model of flower development. To investigate the role of D-lineage genes during ovule and fibre development, GbAGL1 (GenBank accession number: FJ198049) was isolated from G. barbadense by using the SMART RACE strategy. Sequence and phylogenetic analyses revealed that GbAGL1 was a member of the D-lineage gene family. Southern blot analysis showed that GbAGL1 belonged to a low-copy gene family. Semi-quantitative RT-PCR and RNA in situ hybridization analyses revealed that the GbAGL1 gene in G. barbadense was highly expressed in whole floral bud primordia and the floral organs including ovules and fibres, but the signals were barely observed in vegetative tissues. GbAGL1 expression increased gradually with the ovule developmental stages. Over-expression of GbAGL1 in Arabidopsis caused obvious homeotic alternations in the floral organs, such as early flowering, and an extruded stigma, which were the typical phenotypes of the D-lineage gene family. In addition, a complementation test revealed that GbAGL1 could rescue the phenotypes of the stk mutant. Our study indicated that GbAGL1 was a D-lineage gene that was involved in ovule development and might play key roles in fibres development.

Introduction

Cotton is an important crop and fibres are initiated from the ovule epidermal cells (Ramsey and Berlin, 1976; Basra and Malik, 1984; Tiwari and Wilkins, 1995). With each developing ovule, approximately 15–25% (13 000–21 000) of the epidermal layer cells, differentiate into the commercially important lint fibres (Kim and Triplett, 2001; Lee et al., 2007). Hence, cotton ovule size and development undoubtedly affect the yield and quality of cotton fibres.

Ovule development has been widely studied in model plants and has yielded the ‘ABCDE’ model (Angenent and Colombo, 1996; Pelaz et al., 2000). According to this model, floral organ identity and specification are controlled by a series of homeotic genes, and ovule development is closely related to the activities of the D-lineage genes (Theissen and Saedler, 2001; Favaro et al., 2003). Given their critical roles in shaping ovule development, a number of D-lineage genes have been cloned and their roles in ovule development have been well documented. FLORAL BINDING PROTEIN 7 (FBP7) and FBP11 were the first D-lineage genes identified in plant species (Angenent et al., 1995; Colombo et al., 1995; Angenent and Colombo, 1996). These two genes are transcribed in ovular primordial layers, and ectopic expression of FBP7 or FBP11 in petunia induced the formation of ovules on the sepals and petals. In Arabidopsis, SPEEDSTICK (STK), which shares high homology with FBP genes and acts redundantly with SHATTERPROOF (SHP) in promoting ovules, is another typical D-lineage gene (Pinyopich et al., 2003). Similar to FBP7, FBP11, and STK, OsMADS13 is also a key regulator of ovule identity determination in rice, and ovules of the osmads13 knock-out mutant are converted into a reiteration of ectopic carpels or into more amorphous structures that have carpel identity (Dreni et al., 2007). In G. hirsutum, two D-lineage genes, GhMADS5 and GhMADS6, have been cloned. RT-PCR analysis indicates they are expressed in flowers, ovules, and fibres (Lightfoot et al., 2008). In brief, the expression patterns of D-lineage genes are highly expressed in developing ovules, suggesting their function is in regulating ovule development. Sequence analysis indicates that each of the cloned D-lineage genes belongs to the AG-subfamily and shows high degree of conserved regions including MADS-box, K-box, and AG-motifs (Kramer et al., 2004).

Until now, there is no report on the cloning of D-lineage genes from G. barbadense, and the role of D-lineage genes in G. barbadense is unclear. So, to clone and characterize the D-lineage genes may provide us with very useful information on the regulatory mechanism of floral and fibre development.

In this study, the isolation and characterization of GbAGL1, a D-lineage gene from G. barbadense using SMART RACE strategy are reported. Sequence and phylogenetic analyses revealed that GbAGL1 falls into the D-lineage of THE AG-subfamily. The expression patterns of GbAGL1 were investigated via semi-quantitative RT-PCR and RNA in situ hybridization in cotton. Furthermore, ectopic and complementation tests in Arabidopsis were also performed. Our results showed that GbAGL1 was highly expressed in the ovules and fibres. In addition, constitutive expression of GbAGL1 in Arabidopsis caused homeotic alternation in floral organs, and GbAGL1 was able to rescue the phenotype of the stk mutant, indicating that GbAGL1 is a D-lineage gene and is involved in ovule development. It was suggested that GbAGL1 might be involved in fibre development.

Materials and methods

Plant materials

Cotton (G. barbadense) cultivar PIMA-90 was grown in the field in Shanghai Jiao Tong University. The developmental stages of cotton ovules were judged according to Hasenfratz et al. (1995). Wild-type Arabidopsis thaliana and stk-2 mutant (ecotype Columbia) plants were grown in the greenhouse under long-day conditions (22 °C, 16/8 h light/dark). The criteria of Arabidopsis flower bud developmental stages were judged according to Smyth et al. (1990).

Total RNA and genomic DNA extraction

For cotton RNA extraction, samples of various plant tissues (bracts, petals, sepals, stamens, gynoecium, and ovules) were taken from cotton (G. barbadense) plants and immediately immersed in liquid nitrogen. Total RNA was isolated using a CTAB solution (2% CTAB, 0.1 M TRIS, 0.2 M EDTA, 1.4 M NaCl, 1% mercaptoethanol, 0.1% spermidine, pH 9.5) according to a modified procedure (Reid et al., 2006). DNase I (Tiangen, Shanghai) was added to remove genomic DNA and RNase-free columns (Tiangen, Shanghai) were used for purifying total RNA. Arabidopsis total RNA was isolated using the RNAprep pure Plant Kit (Tiangen, Shanghai). Arabidopsis and cotton genomic DNA was isolated by using the DNAquick Plant System (Tiangen, Shanghai).

Isolation and sequence analysis of GbAGL1

To obtain the main fragment of the D-lineage genes, degenerated primers AG1 and AG2 (Table 1) derived from the conservative regions of the AG proteins were designed. The 5′- and 3′-RACE were performed according to the SMART RACE cDNA Amplification Kit (Clontech, USA). Two micrograms of total RNA, isolated from the +3 DPA ovule of G. barbadense, was used for reverse transcription with MMLV Reverse Transcriptase with the 5′-CDS primer A and SMART II A oligo or 3′-CDS primer A from the amplification kit. The primers GPS1, GPS2, and GPS3 (Table 1) were used in combination with the UPM primer to generate the 5′-, 3′-fragment and the full-length cDNA, respectively. All RACE PCR reactions were performed using the Advantage 2 PCR Kit (Clontech, USA) and the conditions were as follows: 25 cycles of 94 °C for 30 s, followed by 68 °C for 30 s and by 72 °C for 3 min.

Table 1.

Primers used in SMART RACE

Primers	Sequence (5′–3′)
3′-CDS primer A	AAGCAGTGGTATCAACGCAGAGTAC (T)30V N-3′, (N=A, C, G, or T; V=A, G, or C)
5′-CDS Primer	(T)25V N-3’ (N=A, C, G, or T; V= A, G, or C)
SMART II A oligo	AAGCAGTGGTATCAACGCAGAGTACGCGGG
UPM	CTAATACGACTCACTATAGGGCAAGCAGTGGTATCAACGCAGAGT
AG1	CNYTNGCNWSNMGNAAYTTYTT
AG2	NCCNARRTGNARDATYTTYTTRTC
GPS1	CTACCCAAGATGGAGAAT
GPS2	CTCGCAATTTCTTTAGCC
GPS3	GATTTTCCGAAAGCTGAATT

The amplified fragments were purified with the DNA Gel Extraction Kit (Waston Shanghai), cloned into the pMD18-T vector (Takara, Japan), and verified by sequencing.

Alignment and phylogenetic analysis

The amino acid sequences of the proteins used in this study were obtained from the GenBank database. Multiple sequence alignment was performed using ClustalX 1.83 (http://www.ebi.ac.uk). The phylogenetic tree was constructed using Molecular Evolutionary Genetics Analysis (MEGA) software 3.1 (Kumar et al., 2004). The Neighbor–Joining tree method was adopted and the tree nodes were evaluated by bootstrap analysis for 1000 replicates. The isoelectric point (pI) and molecular weight (MW) were predicted online (http://www.swiss-prot.com). The 20 MADS proteins used in this study, and their accession numbers, are listed in Table 2.

Table 2.

Proteins used in this study

Gene	Protein accession	Species source	Gene	Protein accession	Species source
AG	NP_567569	A. thaliana	HoMADS1	AAF08830	Hyacinthus orientalis
STK	Q38836	A. thaliana	LAG	AAD38119	Liquidambar styraciflua
AVAG2	BAD83772	Asparagus virgatus	LMADS2	AAS01766	Lilium longiflorum
CUM10	AAC08529	Cucumis sativus	MASAKO D1	BAA90743	Rosa rugosa
EgMADS1	AAS01765	Eustoma grandiflorum	OsMADS3	Q40704	Oryza sativa
FBP7	CAA57311	P. hybrida	PhalAG2	BAG69624	Phalaenopsis hybrid cultivar
FBP11	CAA57445	P. hybrida	PLENA	AAB25101	P. hybrida
GbAGL1	ACI23560	G. barbadense	SHP1	NP_001078311	A. thaliana
GhMADS5	ABM69043	G. hirsutum	SHP2	NP_850377	A. thaliana
GhMADS6	EF202090	G. hirsutum	TAG1	AAM33099	Lycopersicon esculentum

Semi-quantitative RT-PCR analysis

One microgram of total RNA isolated from different G. barbadense organs was used for generating the first strand of cDNA according to the instructions of the Reverse Transcriptase XL (AMV). PCR was carried out as follows: 94 °C for 3 min followed by 30 cycles of amplifications (94 °C for 30 s, 56 °C for 30 s, 72 °C for 1 min) and by 72 °C for 5 min. The specific primers GbAGL1-3 (5′-GAAATCAATGCTCAGTATTATCAAC-3′) and GbAGL1-4: (5′-CTACCCAAGATGGAGAATCTTC-3′) were used for RT-PCR amplification. The internal control Ubiquitin gene was amplified by using primers Ub1 (5′-AAGACCTACACCAAGCCCAA-3′) and Ub2 (5′-AAGTGAGCCCACACTTACCA-3′) under the same conditions to estimate if equal amounts of RNA among the samples were used in the reaction. Three μl of the amplified products were separated on a 1% agarose gel with 20 mM ethidium bromide (EB), visualized under UV light, and photographed.

In situ hybridization

Tissue samples (G. barbadense flower buds and ovules) for in situ hybridization were treated by using a modified version of the procedure described by Coen et al. (1990). G. barbadense flower buds and ovules at different developmental stages (–3, –1, 0, and +3 DPA) were fixed in 4% paraformaldelyde PBS solution with 0.1% Triton-X-100 and 0.1% Tween-20, dehydrated through an ethanol series (50%, 70%, 85%, 95%, and 100%), washed with 100% dimethylbenzene three times, embedded in 100% paraffin, and cut with a microtome into 8 μm thick sections. For RNA probe preparation, a 423 bp fragment (3′-end) with a low degree of sequence identity to other AG-subfamilies from cotton was generated by PCR with primers (GbAGL1-3 and GbAGL1-4) and cloned into the pGEM-T-Easy vector (Promega, USA). The plasmid was linearized and then used as a template for in vitro transcription. The sense and antisense probes were generated by using a Digoxigenin SP6/T7 Labeling Kit (Roche, Germany). RNA in situ hybridization was performed as described previously (Drews et al., 1991). The sections were observed and photographed under the light microscope Olympus BX51 (Olympus, Japan).

Genomic Southern blot analysis

Southern blot was carried out to study the genetic organization of GbAGL1 in the G. barbadense genome. Sixty micrograms of genomic DNA was digested completely with DraI, XbaI, and EcoRV (Takara, Japan) respectively, separated on a 1% agarose gel, and transferred to a Hybond-N⁺ nylon membrane (Amersham Biosciences, USA) as previously described (Sambrook and Russell, 2002). The fragment of 3′-end GbAGL1 (the relatively divergent region) was amplified with gene-specific primers (GbAGL1-3 and GbAGL1-4) and used as the probe. Probe labelling and hybridization was performed according to the Amersham Gene Images AlkPhos Direct Labelling and Detection System (GE, UK). After hybridization, the membrane was exposed to X-ray film at room temperature for 4 h.

Over-expression and complementation analysis

To determine the effect of GbAGL1 in flower and ovule development, the open-reading frame (ORF) of GbAGL1 was cloned into the pCAMBIA1304 (pre-digested by BglII and PmlI) downstream of the 35S promoter to generate the recombinant plasmid pCAMBIA1304-35S::GbAGL1::NOS. The construct was transferred into Agrobacterium tumefaciens GV3101, and then introduced into A. thaliana (ecotype Columbia) plants using the Floral dip method (Clough and Bent, 1998). Fully-mature seeds were collected and screened on a 2.63% phytagel (Sigma) supplemented with 0.5 Murashige and Skoog medium (MS, Sigma) and 20 μg ml hygromycin. The germinated seedlings were transplanted into pots with a soil mixture containing 20% of vermiculite and 20% of perlite and placed in a greenhouse for further growth. PCR was performed to verify the transgenic status of the screened plants.

On condition that GbAGL1 showed high homology to the Arabidopsis D-lineage gene SPEEDSTICK (STK/AGL11), the complementation test was carried out to verify whether GbAGL1 was able to rescue the stk phenotype. The fragment of the promoter of the STK gene was amplified with PCR and the pCAMBIA1304-pSTK::GbAGL1::NOS construct was generated. The Arabidopsis transformation and screening was performed as described above, and characters including plant height, branch number, silique etc. were recorded.

The phenotypic effects of GbAGL1 in transgenic plants was analysed in the T₂ generation through photographing with a digital camera. To compare the phenotypes of the transgenic and wild-type plants, 20 transgenic plants from independent lines 4 and 15 and 20 wild-type plants were grown under the same conditions. Plant height, silique length, seed number per silique, and 10 000-seed weight were recorded.

Results

Characterization and sequence analysis of GbAGL1

A cDNA of 951 bp was isolated from the cDNA pool (+3 DPA ovules) using SMART RACE strategy and designated as GbAGL1 (Accession number: FJ198049). GbAGL1 consisted of a 67 bp 5′-UTR, a 212 bp 3′-UTR, and a 672 bp ORF. GbAGL1 encoded a polypeptide of 223 amino acids (protein accession: ACI23560) with a calculated molecular weight of 25.72 kDa and pI of 9.44. The deduced GbAGL1 protein contained the characteristic conserved motifs: a MADS-box (Fig. 1A) from residues 2 to 56, a less conserved K-box from residues 91 to 157, and two AG motifs (AG motif I: 194–206; AG motif II: 211–223), suggesting that GbAGL1 was a typical AG-subfamily gene.

Fig. 1.

(A) Multiple sequence alignment analysis of GbAGL1. Multiple sequence alignment among GbAGL1 (G. barbadense), GhMADS6 (G. hirsutum), STK (A. thaliana), and FBP11 (P. hybrida). The conserved structures, MADS-box, K-box, AG motif I and II, are marked with black boxes, respectively. (B) Phylogenetic analysis of GbAGL1. Twenty of the complete coding regions of MADS proteins were used and the Neighbor–Joining tree was generated by GeneDoc and MEGA, and the numbers next to each node give bootstrap values from 1000 replicates; The GbAGL1 is marked with a black square.

Comparison of the deduced GbAGL1 amino acid sequence with sequences of other members of D-lineages showed that GbAGL1 exhibited a high degree of similarity to these AG-subfamily proteins. GbAGL1 shared high similarity with GhMADS6 in G. hirsutum (identity of 100%), STK in Arabidopsis (identity of 77.06%), and FBP11 in Petunia (identity of 70.18%).

To look more closely at the relationships between GbAGL1 and other members of the AG-subfamily, a phylogenetic tree of selected members of AG homologues was generated. The phylogenetic analysis grouped GbAGL1 into the D-lineage genes (Fig. 1B). GbAGL1 (emphasized by a black square) was located in the same clade with GhMADS6 and GhMADS5, which had been identified earlier to be D-lineage genes in G. hirsutum. This further implies that GbAGL1 is highly conserved in the cotton species G. barbadense and G. hirsutum.

Multiple alignment and phylogenetic analysis showed that GbAGL1 belonged to D-lineage genes and might be involved in the development of ovules in cotton.

Southern blot was performed with the specific probe of GbAGL1. The result (Fig. 2) showed that there were two strong band signals in lanes 1 and 3 (digested with DraI and XbaI, respectively) and only one band in lane 2 (digested with EcoRV), indicating that GbAGL1 belongs to low-copy-number gene family in the G. barbadense genome.

Fig. 2.

Southern blot analysis of the GbAGL1 gene in the G. barbadense genome. Each lane contains 60 μg genomic DNA digested by DraI, EcoRV, and XbaI, respectively. The λDNA marker (digested by HindIII) is used, and the bands on the lane from top are 23.1, 9.4, 6.6, 4.4, 2.3, and 2.0 kp.

Expression pattern of GbAGL1 gene

Semi-quantitative RT-PCR and RNA in situ hybridization were performed to investigate GbAGL1 spatial expression patterns in cotton. The semi-quantitative RT-PCR results revealed that GbAGL1 expression was restricted in reproductive tissues. As shown in Fig. 3, GbAGL1 was mainly expressed in ovule tissues, while no signals were detected in roots, stems, leaves, bracts, carpels, petals, and stamens. Furthermore, the transcriptional signals were increased throughout ovule development (from –3 DPA to +8 DPA), indicating GbAGL1 is preferentially expressed in fibres.

Fig. 3.

RT-PCR analysis of the GbAGL1 gene in G. barbadense. Expression analysis of GbAGL1 gene in vegetative tissues (roots, stems, leaves, and bracts) and reproductive tissues (petals, carpels, gynoecia, stamens, and ovules in –3, –1, 0, +5, and +8 DPA) by RT-PCR. The Ubiquitin gene is used as the internal control.

Since GbAGL1 expression was restricted to ovules and regularly showed dynamics with developmental stage, RNA in situ hybridization was adopted to determine the precise location of GbAGL1 mRNA in floral buds and ovules at different stages. As shown in Fig. 4, GbAGL1 transcription showed variation and, to a great extent, was distributed in the reproductive organs. A high level of GbAGL1 expression was observed in floral buds and ovules at different stages (–3, 0, and +3 DPA). Strong GbAGL1 activities were detected in the area of ovarian primordia, androecium primordia, and young petals (Fig. 4A), ovular primordia (Fig. 4B), the outer ovule integuments (Fig. 4C, D, E, F), and in fibres (Fig. 4D, E, F), but was weak in the ovary septum (Fig. 4C), and no signals were detected in the mature petals and bracts (Fig. 4B). Indeed, the expression of GbAGL1 was shown to be variational spatio-temporal. GbAGL1 was expressed in the whole floral bud primordia, but was limited mainly to ovules and fibres when the ovules and fibres developed to maturity. On condition that ovules are initiated from the placental region of the inner surface of the ovary (Angenent and Colombo, 1996), weak GbAGL1 signals were detected in the ovary septum, indicating that GbAGL1 and other D-lineage genes might share common conserved expression regions.

Fig. 4.

Expression analysis of the GbAGL1 gene in G. barbadense flower buds and ovules at different stages (–3, 0, and +3 DPA) by RNA in situ hybridization. Sections are subjected to in situ hybridization using the digoxigenin-labelled GbAGL1 cDNA probe. (A) Longitudinal section of a floral bud (about –15 DPA). Hybridization signals are visible in the ovarian primordia (ow), stamen primordia (sp), and young petals (yp). (B) Longitudinal section of a floral bud (about –8 DPA). Hybridization signals are visible in the ovular primordia (ovp) ovarian primordia (op), and stamen (s). (C) Transverse sections of a –3 DPA ovary. Ovules are housed in the ovary, hybridization signals are visible in the ovary septum (os) and outer ovule integument (oi), but barely in the ovary wall (ow). (D) Longitudinal section of a +3 DPA ovule. Strong signals are visible in the outer integument (oi) and fibres (f). (E, F) Longitudinal section of a +5 DPA ovule. Strong signals are visible in the outer ovule integument and fibres (f). s, Stamen; F, fibres; os, ovary septum; op, ovarian primordial; ovp, ovular primordial; ovarian primordial, ow; stamen primordial, sp; young petals, yp.

These expression patterns suggested that GbAGL1 was involved in the development of reproductive floral organs since it was expressed in floral bud and ovule tissues, but remained weak in vegetative tissues. These findings were consistent with the presumption that GbAGL1 had a similar function as other D-lineage genes in floral development. Furthermore, the GbAGL1 transcriptional signals were also detected in fibres, suggesting its role in fibre development.

GbAGL1 affects floral organ identity and rescues stk mutant

To gain further insight into the function of GbAGL1 in the Arabidopsis floral organ development, a recombinant plasmid (pCAMBIA1304-CaMV35S::GbAGL1::NOS) with a hygromycin-resistance gene was constructed (Fig. 5A) and transferred into the Arabidopsis genome. Ten independent transformants were obtained through PCR and hygromycin screening (results not shown). Among the 10 independent transgenic plants, eight showed severe homeotic alteration in the floral organs, which were similar to those of the agamous-like gene transgenic plants, but the other two showed normal phenotypes (results not shown).

Fig. 5.

Vector construction. (A) The construct of transforming vector pCAMBIA-1304-35S::GbAGL1::NOS. The GbAGL1 coding sequence (672 bp) is digested with BglII and PmlI and cloned into the pCAMBIA1304 vector. (B) The construct of vector pCAMBIA-1304-pSTK::GbAGL1::NOS.

The transgenic 35S::GbAGL1 lines 4 and 5 with stronger phenotypes were selected for further analyses. As shown in Fig. 6, the transgenic plants exhibited alterations in both vegetative and reproductive tissues. The leaves of all transgenic plants had curled filamentous profiles and were smaller (Fig. 6A, B), but no carpel tissues were exhibited on leaves. The branch number and the number of inflorescence buds of the transgenic plants were severely decreased (Fig. 6B; Table 3). The young flower buds developed from wild-type plants were completely enclosed by sepals (Fig. 6D). By contrast the flower buds from 35S::GbAGL1 plants opened prematurely, some even before the appearance of the fourth rosette leaf (Fig. 6A; Table 3). The sepals was reduced in size and showed carpel-like profiles. The flower buds opened prematurely, and the stigma extended more severely (Fig. 6C, D; the length ratios of stigma/sepal in stages 13 and 15 were about 2.0, and in the wild type, the ratio was about 1.0). The floral organ feature was further testified by scanning electron microscope (SEM) (see Supplementary Fig. S1 at JXB online). These alterations were similar to those described in Arabidopsis plants transformed with the AG-subfamily from other plant species (Mizukami and Ma, 1992; Kater et al., 1998; Rutledge et al., 1998). Furthermore, the morphological features of transgenic plants were also investigated (Table 3). The height of 35S::GbAGL1 plants was 20.3±3.3 cm, which was shorter than that of Wt (25.8±4.7 cm). The silique was shorter (the average silique length of 35S::GbAGL1 plants was 11.6±3.3 mm versus 14.4±2.5 mm of Wt), and the number of seeds per silique decreased slightly (the average seed number per silique of 35S::GbAGL1 plants was 37.6±5.2 versus 48.9±6.7 of Wt). The seed size also decreased slightly (the 10 000-seed weight of transgenic plants was 0.226 g versus 0.238 g of Wt).

Fig. 6.

Phenotypes of transgenic Arabidopsis plants expressing 35S::GbAGL1 and pSTK::GbAGL1. (A–D) Transgenic Arabidopsis plants expressing the 35S::GbAGL1 gene. (E, F) Transgenic Arabidopsis plants expressing pSTK::GbAGL1. (A) Compared with the Wt Arabidopsis (right), the GbAGL1 transgenic Arabidopsis (left) shows reduced size, early flowering, and small and curled leaves. (B) The GbAGL1 transgenic Arabidopsis (left) shows decreased flowers. (C) The transgenic buds are prematurely opened. (D) The transgenic (left) and Wt (right) floral organs at different stages. The transgenic floral buds are prematurely open. In stages 9 and 11, the stigma lengths of stages 13 and 15 are twice that of the sepal lengths. The Wt floral organs of differential stages. In stages 9 and 11, the floral bud is just open. The stigmas of stages 13 and 15 are level with the sepals. (E) Compared with the stk mutant (left), the pSTK::GbAGL1 transgenic plants have longer siliques. (F) Enlarged view of (E) showing that the pSTK::GbAGL1 transgenic plants have longer siliques than those of the stk mutants.

Table 3.

Morphological features of wild type (Wt) and 35S::GbAGL1

Plants	Plant heighta (cm)	Branch numberb (n)	Leaf number before floweringc (n)	Sepal number (n)	Silique lengthd (mm)	Seed number/siliquee (n)	10 000-seed weight (g)
35S::GbAGL1	20.3±3.3	1∼3	4∼5	4	11.6±3.3	37.6±5.2	0.226
Wt	25.8±4.7	3∼6	8∼10	4	14.4±2.5	48.9±6.7	0.238

^aValues obtained from 20 plants from Wt and 20 35S::GbAGL1 plants.

^bValues obtained from 20 Wt and 20 35S::GbAGL1 plants.

^cValues obtained from 20 Wt and 20 35S::GbAGL1 plants.

^dValues obtained from 20 siliques from Wt and 35S::GbAGL1 plants.

^eValues obtained from 20 siliques from Wt and 35S::GbAGL1 plants.

In addition, the complementarity test was performed and it revealed that GbAGL1 can rescue the stk mutant. There were seven independent pSTK::GbAGL1 plants, and most of the transgenic plants showed little change in the vegetative tissues. For example, the pSTK::GbAGL1 transgenic vegetative traits including plant height, leaf size, branch number etc. showed no significant difference (Fig. 6E, F; Table 4). Compared with the stk mutant, the sizes of the floral organs and silique developed from the pSTK::GbAGL1 plants were obviously enhanced (Fig. 6E, F; Table 4). The flower number of the pSTK::GbAGL1 transgenic plants was more than those of the stk mutant (Fig. 6E, F). In addition, the floral size was larger, and pollen fertility was better than those of the stk mutant (Fig. 6E, F). In the later stages, the silique from the pSTK::GbAGL1 transgenic plants was much longer (the average silique length of pSTK::GbAGL1 plants was 10.5±4.5 mm versus 3∼12 mm of stk), and the seeds per silique increased slightly (the average seed number per silique of pSTK::GbAGL1 plants was 30.6±7.8 versus 0∼35 of stk). The seed size was increased slightly (the 10 000-seed weight of pSTK::GbAGL1 plants was 0.226 g versus 0.184 g of stk).

Table 4.

Morphological features of stk and the transgenic Arabidopsis expressing GbAGL1

Plants	Plant heighta (cm)	Branch number (n)	Leaf number before flowering (n)	Sepal number (n)	Silique lengthb (mm)	Seed number/ siliquec (n)	10 000-seed weight (g)
pSTK::GbAGL1	42.5±6.5	4∼6	7∼9	4	10.5±4.5	30.6±7.8	0.266
stk	40.5±4.5	4∼6	7∼9	4	3∼12	0∼35	0.184

^aValues is obtained from 20 stk and 20 pSTK::GbAGL1 plants.

^bValues is obtained from 20 stk and 20 pSTK::GbAGL1 plants.

^cValues is obtained from 20 stk and 20 pSTK::GbAGL1 plants.

Discussion

Compared with other cotton species, G. barbadense has relatively long and fine fibres. The identification of genetic factors responsible for fibre quality and yield improvement from G. barbadense germplasm is meaningful and attractive. According to the characterization of the C-terminal end of the AG subfamily, it was possible to clone GbAGL1. Sequence analysis suggests GbAGL1 falls within the D-lineage genes, and shows a high degree of similarity with GhMADS6 of G. hirsutum, suggesting that GbAGL1 is highly conserved in the cotton species G. barbadense and G. hirsutum. The deduced peptide of GbAGL1 contains two short and highly conserved regions (AG motifs I and II, which have been defined by Kramer) showing that GbAGL1 belongs to the AG-subfamily and is a putative D-lineage gene (Kramer et al., 2004).

According to the ABCDE model, AG-subfamily genes have been divided into C- and D-lineage genes. However, evolutionary conservation analysis has revealed that they evolved from a duplication event during early angiosperm evolution (Kramer et al., 2004). Data have shown that both C- and D-lineage genes are expressed and have functional roles in ovule development in Arabidopsis and other plants (Kyozuka and Shimamoto, 2002; Pinyopich et al., 2003). Constitutive expression of GbAGL1 has induced homeotic changes in the flower, which are very similar to the phenotypes caused by C-lineage genes. These findings have further testified their functional conservations.

Earlier studies have shown that the activities of D-lineage genes are restricted to the floral organs and they are involved in ovule development. For instance, AVAG2 has been cloned from ornamental asparagus and its transcriptional signals were detected in the ovule only at the later developmental stages (Yun et al., 2004). In this study, it was found that GbAGL1 was transcribed in the ovule primordia (op), ovary septum (os), and ovule integument (oi) (Fig. 4C), which was very similar to STK (Pinyopich et al., 2003; Brambilla et al., 2007, 2008). However, it was also found that this gene transcribed a little in other tissues including the androecium primordia and young petals, which indicated GbAGL1 may have wider expression patterns. Furthermore, GbAGL1 was expressed in specific tissues in the flowers and ovules, which was very similar to those of other D-lineage genes like HoMADS1 from hyacinth, OsMADS13 from rice, AVAG2 from asparagus, and DthyrAG2 from orchid (Xu et al., 2004; Yun et al., 2004; Skipper et al., 2006; Dreni et al., 2007). These findings revealed that the expression patterns of D-lineage genes were conserved in plant fields.

In cotton, fibres are distinct from the seed trichomes in Arabidopsis in that they are unbranched and extremely elongated (Kim and Triplett, 2001). However, new data suggest that cotton and Arabidopsis use similar transcription factors in regulating trichomes (Wang et al., 2004). In our study, it was found that GbAGL1 exhibited temporal and spatial expression patterns in G. barbadense, which were similar to other homologous D-lineage genes, such as GhMADS5 and GhMADS6 in G. hirsutum (Lightfoot et al., 2008). These findings suggested that GbAGL1 could be highly conserved in cotton species and might be involved in a series of different developmental processes, including ovule development, fibre initiation, and elongation. The RNA in situ hybridization indicated GbAGL1 was expressed highly in the outer ovule integuments and fibres. These findings suggested that GbAGL1 might not only associate with ovule development, but also with fibre development. Recent studies indicated that the underlying mechanisms controlling fibre development are extremely complex (Hovav et al., 2008), and our observations have provided reasonable explanations for the roles of D-lineage genes in the processes of cotton ovule and fibre development.

The function of AG-subfamily genes in flower and fruit development have shown that floral development was conserved among divergent species, and ectopic expression caused homeotic alteration in floral organs (Ng and Yanofsky, 2001). LMADS2 and EgMADS1 are D-lineage genes characterized from the species lily and lisianthus, respectively, and both of them can cause similar homeotic conversion of the sepals and petals in Arabidopsis (Tzeng et al., 2002). In this study, ectopic expression of GbAGL1 also caused homeotic changes in the floral organs, and the transgenic phenotypes were in accordance with reports on the over-expression of LMADS2 and EgMADS1 in Arabidopsis. Besides, the observed modifications were entirely in accordance with reports on AG or AG orthologues DAL2, SAG1, MASAKO C1 and D1 over-expression in Arabidopsis (Mizukami and Ma, 1992, 1997; Rutledge et al., 1998; Tandre et al., 1998; Kitahara and Matsumoto, 2000). However, the 35S::GbAGL1 Arabidopsis had significant differences with those observed in the FBP11 transgenic petunia in which the sepals and petals ectopically produced ovule-like structures, whereas the leaf morphology and flowering time were normal (Colombo et al., 1995). This difference might be due to the diversity in the D-lineage gene functions or the difference between backgrounds of those plants. In addition, the complementation analysis revealed that GbAGL1 was probably able to rescue the stk mutant. The silique length and seed of pSTK::GbAGL1 plants were more enhanced than those of the stk mutant, which revealed that GbAGL1 was a STK homologous gene in cotton. However, whether the functions of GbAGL1 in G. barbadense and A. thaliana share common principles and GbAGL1 affects ovule size in cotton plants needs to be further investigated in the future.

Glossary

Abbreviations

Bp

base pair

CaMV

Cauliflower mosaic virus

DPA

days post anthesis

LB

Luria–Bertani medium

MEGA

Molecular Evolutionary Genetics Analysis

min

minute

MS

Murashige and Skoog medium

MW

molecular weight

pI

isoelectric point

RACE

Rapid Amplification of cDNA End

s

second

SEM

scanning electron microscopy

Wt

wild type