Introduction

The London plane tree (Platanus acerifolia) is widely planted as a roadside and parkland tree in cities, and it is considered as an interspecific hybrid of the oriental plane (Platanus orientalis) and American sycamore (Platanus occidentalis) (Zhang et al. 2011a, b; Lu et al. 2012). It has a lot of desirable characters, such as rapid growth, huge crown for shading, attractive colorful leaves in fall season, and good adaptability to urban environment (Li et al. 2007). However, this species also has a few unfavorable traits, and the greatest negative impact on human life is the pollen shedding of male flowers in spring and fruit hair releasing in the late spring. It results in the pollution of environment, leading to some serious allergic symptoms (such as rhinitis and bronchial asthma, etc.) (Han et al. 2016). In order to solve the problems, breeding non-flowering or fruitless varieties is imperative.

The development of genetic engineering and molecular biology provides an effective new way to generate genetically modified (GM) plants including sterile plants (Mariani et al. 1990). And the utilization of tissue-specifc promoters has important practical application value by controlling the timing and precision of foreign gene expression and reducing the impact on plant growth, development, and metabolic pathways (Lv and Zhang 2016). Among these promoters, flower-specific promoters were proved to be applicable to generate complete sterility (Kobayashi et al. 2006; Liu and Liu 2008; Yang et al. 2010). A number of flower-specific promoters have been isolated and identified to engineer sterility. The anther tapetum-specific tobacco promoter (TA29) was fused to BARNASE or RNase T1 and transformed into tobacco, and the results showed that transgenic plants with BARNASE specifically expressed in anther and tapetum cells were dead (Mariani et al. 1990). Expression of BpFULL1: BARNASE in tobacco and Arabidopsis resulted in ablation of inflorescences (Lannenpaa et al. 2005). In addition, expression of BARNASE in tobacco reproductive tissues driven by the FBP6 promoter (pFBP6) from petunia also resulted in the ablation of flower buds (Liao et al. 2012). Above evidences suggested that flower-specific promoters can be used for engineering sterility in these species mentioned above.

According to the ABC model of flower development, B-class MADS-box transcription factors, APETALA3 (AP3) and PISTILLATA (PI), are key regulators of petal and stamen development (Jack et al. 1992; Goto and Meyerowitz 1994). In eudicots and monocots, loss of function of B-class genes exhibited similar phenotypes, i.e. the development of second and third whorl floral organs are disrupted (Jack et al. 1992; Schwarz-Sommer et al. 1992; Trobner et al. 1992; Goto and Meyerowitz 1994; Ambrose et al. 2000; Nagasawa et al. 2003; Vandenbussche et al. 2004; Drea et al. 2007; Kramer et al. 2007; Roque et al. 2013; Bartlett et al. 2015; Otani et al. 2016), suggesting that the function of B-class genes is largely conserved during evolution. Expression of B-class genes in core eudicots and gramineous monocots is relatively conserved, predominantly found in developing petals or lodicules (organs likely to be homologous to petals) and stamens (Jack et al. 1992; Ambrose et al. 2000; Nagasawa et al. 2003; Vandenbussche et al. 2004). In species outside the core eudicots, however, the B-class genes showed a higher degree of expression divergence (Kim et al. 2005; Kramer and Irish 1999; Zhao et al. 2010; Zhang et al. 2011a, b). Previously, we have isolated AP3 and PI homologues (PaAP3 and PaPI) from London plane tree. PaAP3 showed high expression in male inflorescences, low expression in female inflorescences, and weak expression in stems of 1-year-old seedlings and adult plants, but no expression was detected in roots, leaves and vegetative buds (Li et al. 2012). There are at least five PI homologous genes (PaPI1/2/3/4/5) that show different expression profiles and transcript abundance in London plane tree, among which PaPI2 may be the major functional PI homolog and expressed predominantly in male and female inflorescences (Zhang et al. 2011a, b).

In this study, the PaAP3 and PaPI2 (thereafter named PaPI for convenience) promoter were isolated from London plane tree, and their expression features were characterized by fusing to GUS report gene followed by introduction into tobacco plants. Histochemical staining and gene expression analysis showed that GUS activity of pPaAP3::GUS transgenic plants could be detected evidently in various floral organs while in pPaPI::GUS transgenic plants GUS activity predominantly existed in the inner three whorls of flowers. In addition, the expression of BARNASE driven by PaAP3 and PaPI promoter was also investigated.

Materials and methods

Plant materials

London plane tree grown at Huazhong Agricultural University was used for promoter cloning. Nicotiana tabacum ‘Xanthi’ was used for genetic transformation experiment. Transformed plants were grown in a containment greenhouse with a 14 h photoperiod and a 25 °C/17 °C day/night temperature regime.

Isolation of PaAP3 and PaPI promoters

Total DNA was extracted with CTAB method (Li et al. 2008). Thermal asymmetric interlaced-PCR (TAIL-PCR) was used to obtain the promoter region of the PaAP3 and PaPI (Liu et al. 1995). The eight universal primers used for TAIL-PCR were listed in Table 1. We designed three gene-specific reverse primers (AP3-GSP1/2/3) according to the base sequence of exon 1 and 5′UTR of PaAP3 (PaAP3: GenBank accession No. EF488452). In the meantime, we designed three gene-specific reverse primers of PaPI (PI-GSP1-1/2/3) following the same principle of PaAP3 (PaPI: GenBank accession no. EF488452). After the PaPI promoter sequences were isolated and analyzed, another three gene-specific reverse primers for TAIL-PCR (PI-GSP2-1/2/3) were designed to isolate longer promoter sequence. All the gene-specific primers are listed in Table 1 while the steps of TAIL-PCR program are listed in Table 2.

Table 1 Primer sequences used for TAIL-PCR
Full size table
Table 2 The modified protocol of TAIL- PCR used in study
Full size table

Analysis of PaAP3 and PaPI promoter sequences

Fundamental functions of the promoters were confirmed by the Promoter Scan program (http://www-bimas.cit.nih.gov/molbio/proscan/). PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) and PLACE (http://www.dna.affrc.go.jp/PLACE/index.html) were used to analyze potential cis-acting regulatory elements of PaPI and PaAP3 promoters.

Vector construction

Primers pPaAP3-F and pPaAP3-R were used to amplify a 2.9 kb PaAP3 promoter (pPaAP3) (Table 3); meanwhile, primers pPaPI-F and pPaPI-R were used to amplify a 2.6 kb PaPI promoter (pPaPI) (Table 3). Both of the reverse primers, pPaAP3-R and pPaPI-R, were exactly next to the initiation codon of gene coding region. Both of the amplified fragments were inserted into pMD-18T (Takara Biomedical Technology, Beijing, China) and sequenced (Sangon Biotech, Shanghai, China). Then, they were double digested by BamHI and PstI, and both were introduced into pCAMBIA1391Z vector (Fig. 1b, c). The empty pCAMBIA1391Z vector was used as a negative control construct (Fig. 1a), and the CaMV 35S promoter was introduced into pCAMBIA1391Z as a positive control (Fig. 1d). In addition, the pPaAP3 and pPaPI were introduced into BpFULL1::BARNASE binary construct (Lannenpaa et al. 2005), separately with SalI and BamHI digestion to produce the pPaAP3::BARNASE and pPaPI::BARNASE plasmids (Fig. 1e–g).

Table 3 Primers used for vector construction
Full size table
Fig. 1
figure 1

Schematic diagram of constructs utilized in this study. a The original vector of pCAMBIA1391Z containing a promoterless GUS reporter gene. b The pPaAP3::GUS construct, containing the 2.9 kb upstream promoter region of PaAP3 placed in front of the GUS reporter gene. c The pPaPI::GUS construct, containing the 2.6 kb upstream promoter region of PaPI. d The positive control vector pCAMBIA1391Z containing a 35S promoter placed in front of the GUS. e The original pBpFULL1::BARNASE construct, containing the promoter region of BpFULL1 placed in front of BARNASE and BARSTAR gene (from Dr. M. Lӓnnenpӓӓ of University of Joensuu, Finland). f The pPaAP3::BARNASE construct, with pBpFULL1 replaced with pPaAP3. g The pPaPI::BARNASE construct, with pBpFULL1 replaced with pPaPI

Full size image

Tobacco transformation

All plasmids were transferred into the Agrobacterium tumefaciens EHA105. The transformation method for tobacco was according to Ning et al. (2012). Transgenic tobacco lines were verified by PCR with the primer pairs GUS-F/GUS-R and BARNASE-F/BARNASE-R (Table 3).

Histochemical assays

The method of histochemical staining was described by Liao et al. (2012). various tissues of each transgenic line and control were immersed in a 1 mg/ml solution of X-Gluc, prepared with 50 mM sodium phosphate (pH 7.0), 0.5 mM potassium ferricyanide, 0.5 mM potassium ferrocyanide, 10 mM Na2-EDTA and 0.1% w/v Triton X-100, for 12–18 h under 37 °C.

RT-PCR analysis

Total RNA was extracted from leaves, stems and flowers with the RNA extraction kit (Trizol Reagent, Invitrogen). The NtEF1α was used as internal control. The primers used for RT-PCR are listed in Table 3. The PCR program was 3 min at 94 °C for initial denaturation, followed by 30 cycles of 94 °C for 30 s, 59 °C for 30 s and 72 °C for 50 s.

Results

Sequence information of pPaAP3 and pPaPI

In order to isolate the promoter sequence of PaAP3 and PaPI, TAIL-PCR strategy was adopted. After first three rounds of amplification, a 3.1 kb fragment upstream of PaAP3 gene and a 2.2 kb fragment upstream of PaPI gene were isolated and sequenced, using the AD primer and AP3-GSP1/2/3 or PI-GSP1-1/2/3 (Table 1). Based on the obtained sequence information, another three specific reverse primers PI-GSP2-1/2/3 were designed to isolate longer PaPI promoter, which resulted in a 700 bp extension to the promoter fragment, making its length up to 2.9 kb. Sequence accuracy was identified by repeatedly sequencing.

Bioinformatics analysis of the pPaAP3 and pPaPI

The sequence of pPaAP3 was submitted to the PlantCARE and PLACE, and confirmed as the “Highly likely prediction”. The result indicated that pPaAP3 fragment contains several basic motifs, including six putative TATA box (i.e., TATTAAT, TATATAA, TTATTT), ten CAAT box (CAAT), one MADS-box protein binding CArG box (CWWWWWWWWG) (Table 4) and a target sequence named LEAFYTAG (CCAATGT). When the sequence of pPaPI was analyzed, similar result was displayed, in which the detected motifs include one putative TATA box (TTATTT) and ten CAAT box (CAAT) (Table 5). In addition, a cis-acting regulatory element AGL3ATCONSENSUS, which was the target sequence of AGL3, was detected in both promoter fragments. Nevertheless, some different motifs were also observed between the two promoter fragments. For example, AGL2ATCONSENSUS was found in pPaAP3 but not in pPaPI (Tables 4, 5).

Table 4 Analysis of putative cis-acting regulatory elements in the pPaAP3 promoter
Full size table
Table 5 Analysis of putative cis-acting regulatory elements in the pPaPI promoter
Full size table

Spatial expression patterns of pPaAP3 and pPaPI

To investigate the tissue-specific expression patterns directed by the pPaAP3 and pPaPI, expression vectors containing a 2.9 kb length fragment of pPaAP3 and a 2.6 kb length fragment of pPaPI fused with GUS were constructed, separately, and used to transform tobacco plants (Fig. 2). 11 pPaAP3::GUS transgenic lines and 13 pPaAP3::GUS transgenic lines were obtained, and all of them were used for histochemical staining analysis. The results demonstrated that the transgenic lines of pPaAP3::GUS not only showed GUS activity in sepal, petal, stamen and pistil, but a certain level of activity was also detected in vegetative organs (Fig. 2c, d). In contrast, only petal, stamen and pistil were stained in the transgenic lines of pPaPI::GUS (Fig. 2e). As positive control, transgenic lines of 35S::GUS demonstrated GUS activity in all the tissues and organs (Fig. 2b). No GUS activity was detected in wild-type (WT) and transgenic tobacco plants with empty control construct (Fig. 2a).

Fig. 2
figure 2

Histochemical GUS staining of pPaAP3::GUS and pPaPI::GUS transgenic tobacco plants. a Histochemical GUS staining of non-transgenic tobacco plants. b Histochemical GUS staining of 35S::GUS transgenic tobacco plants. c, d Sepal, petal, stamen, anther, pistil, ovary, leaves and stem from pPaAP3::GUS transgenic tobacco plants. e Sepal, petal, stamen, anther, pistil and leaves from pPaPI::GUS transgenic tobacco plants

Full size image

Heterologous expression of BARNASE driven by pPaAP3 or pPaPI in tobacco

Considering temporal and spatial expression patterns of pPaAP3 and pPaPI, it is potential to create flower organ-specific ablation by using the promoter fragments fused with BARNASE gene. Thus, we transformed tobacco plants with pPaAP3::BARNASE and pPaPI::BARNASE construct, aiming to observe the effect of heterologous expression before applying them to P. acerforlia.

Eighteen transgenic lines of pPaPI::BARNASE were obtained, and it demonstrated that 76% plants formed depauperate flower buds—which abscised from pedicels in an early stage of floral development (Fig. 3b). Based on the morphology anatomy, we found only sepals were persistent, and all the petals, stamens, stigma and ovary were completely ablated (Fig. 3d) compared to the flowers of wild-type and transgenic tobacco plants with empty construct (Fig. 3e). Vegetative growth was not affected, except lateral branchs increased (Fig. 3a). In contrast, transgenic plants of pPaAP3::BARNASE showed the inhibition of vegetative growth. More than half of transgenic lines (ten lines) were either small or weak, and the flower buds were completely withered and browning in the early stage of development, resulting in abnormal structures of the flower buds (Fig. 3c).

Fig. 3
figure 3

The phenotype of tobaccos transformed with pPaAP3::BARNASE and pPaPI::BARNASE. a Compared to the wild-type tobacco (right), plant transformed with pPaPPI::BARNASE (left) showed complete bud abscission and lateral branches increased. b Tobacco transformed with pPaPI::BARNASE showed floral abscission. c Plants transformed with pPaAP3::BARNASE showed browning and withering of top meristem. d Longitudinal section of flower buds from the pPaPI::BARNASE transgenic plants revealed ablation of inner three whorls of floral organs. e Longitudinal section of flower buds from WT tobacco

Full size image

Semi-quantitative expression in transgenic tobacco

Using the RT-PCR analysis, the floral organ-specific ablation and the influence of vegetative growth were clearly proved to be related with the expression patterns directed by the promoter fragments. Total RNA was extracted from leaf, stem, and the floral organs in the transgenic lines of GUS constructs, and from leaf, stem and the residual floral organs (petal) in the transgenic lines of BARNASE constructs, which then were reversely transcribed into cDNA. Besides, a whole plant of 1-month old wild-type tobacco was also applied for the RT-PCR analysis as negative control. The concentration of all the cDNA samples used for detection of target gene was balanced by the internal control gene NtEF1α. The RT-PCR results demonstrated that GUS activity was only detected in the inner three whorls of floral organs in transgenic plants with pPaPI::GUS constructs (Fig. 4a), but in all the whorls of floral organs in transgenic lines of pPaAP3::GUS, including sepals (Fig. 4b). In line with this, the expression of BARNASE in transgenic lines of pPaPI::BARNASE was not detected in the sepals (floral organ left only), and it was detected a certain expression in the stem and leaf (Fig. 4c); but in the transgenic lines of pPaAP3::BARNASE, BARNASE could be detected obviously in stem and weakly in leaf (Fig. 4d). In wild-type tobacco plants, neither GUS nor BARNASE transcripts were detected in all the plant tissues.

Fig. 4
figure 4

RT-PCR analysis of GUS and BARNASE gene expression in transgenic tobacco. a Semi-quantitative RT-PCR analysis of GUS expression in WT and pPaPI::GUS transgenic tobacco. b Semi-quantitative RT-PCR analysis of GUS expression in WT and pPaAP3::GUS transgenic tobacco. c Semi-quantitative RT-PCR analysis of BARNASE expression in WT and pPaPI::BARNASE transgenic tobacco. d Semi-quantitative RT-PCR analysis of BARNASE expression in WT and pPaAP3::BARNASE transgenic tobacco. The NtEF1α gene of tobacco was used as an internal control

Full size image

Discussion

In this study, we mainly investigated the possibility of breeding sterile London plane trees by using the genetic engineering with the flower tissue-specific promoters of the B-class MADS-box genes. Compared to means of artificial microRNA or RNA interference of floral related genes, flower tissue-specific promoters fused with cytotoxic genes such as BARNASE may be more efficient approaches to create sterile plants (Lannenpaa et al. 2005), because the flowering and floral development progress is regulated by a large number of genes, which usually show overlapping or redundant functions leading to inefficient results by silencing a single gene.

At first, we isolated the promoter fragments upstream of the 5′UTR of PaAP3 and PaPI from P. acerifolia. Bioinformatics analysis confirmed that both fragments contain the basic promoter domains and regulation elements. Then, both of them were fused with GUS reporter gene and cytotoxic gene BARNASE, and transformed into tobacco, respectively. The expression of GUS and BARNASE demonstrated that the two promoter fragments have different expression specificity, with the PaPI promoter driving gene expression more particularly in floral organs, which can lead to flower organ-specific cell ablation and sterile plants as expected.

As we known, B-class MADS-box genes regulate the development of the second and third whorl floral organs, and their expression is primarily limited in petals and stamens, hardly detected in vegetative organs in the majority of plant species (Jack et al. 1992; Trobner et al. 1992; Vandenbussche et al. 2004). In our study, when pPaAP3::GUS was transformed into tobacco, the activity of GUS was detected in all four whorls of flower. RT-PCR analysis further confirmed the expression of GUS at various levels not obly in all floral organs, but also in vegetative organs including stems and leaves, although lower than that in floral organs (Fig. 4b). In contrast, the pPaPI::GUS transgenic lines exhibited a high level activity of GUS in petal, stamen and pistil, with no or low expression, if any, in sepal and vegetative tissues (Fig. 4a). Our previous qRT-PCR analysis indicated PaAP3 was expressed mainly in the male inflorescences of London plane tree, but that its expression was also detected in female inflorescences and even weakly in the stems of juvenile and adult plants (Li et al. 2012); PaPI was likewise expressed in both male and female inflorescences, as well as in mature embryos, but almost no expression could be detected in vegetative tissues (Zhang et al. 2011a, b). The expression patterns of PaAP3 and PaPI in London plane suggest that their promoters (pPaAP3 and pPaPI) may contain special and distinct cis-regulatory elements, which may explain the different activities and the broader expression region of pPaAP3 and pPaPI in London plane than that in most core eudicots. Alternatively, the expression pattern of heterologously introduced promoters may be regulated by the upstream transcription factors in receptor plants (Kusters et al. 2015), such as tobacco used here. In tobacco, differences were observed in the expression patterns of NTDEF and NTGLO (AP3 and PI ortholog, respectively). Expression of NTGLO was restricted to the regions of petals and stamens, butno expression was detected in stem, root, leaf, sepal and gynoecium. Whereas the expression of NTDEF was more widespread, which was easily detected in the fourth whorl and faintly detected in the first whorl, with relatively high expression in the second and third whorls (Davies et al. 1996; Hansen et al. 1993).

When pPaAP3::BARNASE construct was transformed into tobacco, the apical meristem and flower buds were browned to death in the early stage of development. The transcripts of BARNASE, in few survivals, were detected in stem and leaf, and the expression level was higher than the parallel group containing the pPaPI::BARNASE. The expression of BARNASE could be detected in all lines transformed with either of the two constructs, although at different expression levels. The reason of survivals may due to a small 35S core fragment fused with BARSTAR gene following the BARNASE gene in the constructs (Lannenpaa et al. 2005). Instead of inhibiting the transcription of BARNASE gene, BARSTAR gene encodes a kind of enzymatic inhibitor of RNase which is encoded by BARNASE gene on protein level. Therefore, the transgenic plants kept normal growth even though the transcript of BARNASE was detected in the vegetative organs. In this study, when BARSTAR was expressed at a low but ubiquitous level in the transgenic lines with pPaPI::BARNASE, the leakage of BARNASE gene at a certain level did not cause damage to vegetative organs and sepals. However, in the transgenic plants of pPaAP3::BARNASE, the expression of BARSTAR was not high enough to inhibit the toxicity of BARNASE. This result also confirmed the good specificity of pPaPI as well. Although the leakage of BARNASE was detected in the transgenic plants, the expression in leaf, stem and sepal was much lower than that in the inner three whorls of floral organs. In contrast, the poor specificity of pPaAP3 leading to the leakage of BARNASE in the transgenic plants resulted in the death of most transformants.

In summary, our study characterized the expression specificity and function of the promoters of PaAP3 and PaPI in transgenic tobacco plants. The results revealed that pPaPI has stronger specificity than pPaAP3, which can specifically control gene expression within the inner three whorls of floral organs, and so may be used efficiently in the gene engineering of sterility. Although the transgenic lines of pPaAP3 fused with BARNASE gene also resulted in floral abortion, the poor specificity limited its potential for application. Based on the present results and the expression pattern of PaPI in London plane, the pPaPI::BARNASE construct may have the potential to be applied in breeding programs for sterile lines of London plane tree. This will be benefit for the environment and human health by reducing the fruit hairs and pollens. In addition, this approach can also be used for obtaining sterile transgenic plants, thereby preventing transgene flow through pollens, fruits, or seeds.