Selectable marker for U. gibba transformation protocol
To identify an appropriate resistance gene for U. gibba transformation, we tested the effect of different concentrations of selective agents previously used for developing plant transformation systems, such as Hygromycin (Hyg, aminocyclitol antibiotic) [38], Kanamycin (Kan, aminoglycoside antibiotic) [39] and Glufosinate-ammonium PESTANAL (PPT) on the growth of U. gibba plants. Hyg and Kan inhibit protein synthesis, whereas PPT is an herbicide that inhibits glutamine synthetase, causing accumulation of toxic levels of ammonia in the cells [40]. U. gibba plants were grown for two weeks in liquid media containing different concentrations of these selective agents. U. gibba plants cultivated with 20, 40, 60 and 80 mg/l of Kanamycin survived all concentrations during their growth kinetics. Tissues treated with 5, 10, 20 and 30 mg/l Hygromycin showed a slight yellowing and sporadic dead tissue. However, even at the highest concentration of Hygromycin, most tissue remained green and viable. By contrast, after 9 days of exposure to all concentrations of PPT (6, 10, 15 and 20 mg/l), plants showed gradual whitening, and after two weeks, all tissues were dead. We selected 10 mg/l of PPT to supplement selective media for all subsequent transformation experiments (See Additional file 1: Figure S1).
Transformation protocol
To determine whether the cytokinin benzyl amino purine (BAP) could induce a greater number of transformants by increasing de novo shoot formation, U. gibba explants were grown for a week in MS medium supplemented with 0, 0.25 or 0.75 mg/l of BAP prior to co-cultivation. Explants were then co-cultivated for 72h with an Agrobacterium strain carrying a binary vector containing the p35S-GUS::GFP construct. They were then transferred to PPT selective media for two weeks, each explant returning to BAP concentrations as previously described. All treatments were then transferred to selective media without BAP. Four independent biological replicates were used for each treatment. For explants incubated in media without BAP, we obtained a transformation efficiency of 0.14 transformants per gram of tissue, 0.9 for explants incubated with 0.75 mg/l of BAP, and 2.24 for media containing 0.25 mg/l BAP. Although an ANOVA test showed that there was not a significant difference between the BAP treatments, we found reproducibly higher efficiencies using 0.25 mg/l BAP in the selective media (Fig.1B). The selection procedure was very effective, as no escape events were detected. All plantlets rescued after transformation with the p35S-GUS::GFP construct showed homogeneous GUS staining and were PCR positive (Figure 3; Additional File 1: Figure S4). Moreover, the results were reproducible using the protocol with 0.25 mg/l BAP and a vector carrying a reporter gene under the control of the promoter of a previously reported gene with trap-specific expression [41].
To better understand how the genetic transformation in U. gibba tissue takes place, the transformation process was monitored by GUS staining during the 45 subsequent days after co-culture with Agrobacterium. U. gibba tissue was sampled from two independent experiments at 7, 18, 31, 43 and 45 days after co-culture with Agrobacterium. Each tissue sample was stained with X-Gluc to observe GUS expression foci in co-cultivated tissue, and representative stained tissue was selected for photography. At 7 (Figure 2a-c) and 18 days after co-culture many small regions showing clear GUS staining were observed in both meristematic and vegetative tissues (Figure 2d-f). Approximately one month after Agrobacterium co-culture, we observed larger sectors with homogeneous GUS staining (Figure 2g-h). After six weeks in selective medium, clearly distinguishable green sectors were observed after staining that showed homogeneous GUS activity (Figure 2 i). Application of BAP did not result in the formation of callus tissues but rather it directly promoted a more rapid de novo formation of meristems that developed in new plant branches that could be dissected to establish a new plant. Between seven and eight weeks of growth in selective media, green sectors that were clearly visible were dissected and separated from the original dead tissue for further propagation. Initially, we propagated three independent transgenic lines from two independent transformation experiments for further analysis. Selected putative transgenic lines were stained to corroborate that they expressed the GUS reporter gene (Figure 2 j-m). To test whether the transgenic clones produced were stably transformed, three independent putatively transgenic clones containing p35S-GUS::GFP were sub-cultivated three times for one month in fresh media without PPT and then tested for GUS expression. We observed that the three clones maintained the expression of the GUS reporter gene after each of three subcultures (Figure 3). In all transgenic lines examined, the 35S promoter was found to directly express in stolons and leaf-like structures, as well as in the entire trap walls including antennae and trap hairs (Figure 4). No GUS or GFP expression was ever observed in wild type tissue even after long incubation times in histochemical assays (Figure 4). To confirm that GUS positive clones were indeed transgenic, we quantified the expression level of p35S-GUS::GFP in U. gibba tissue by Real Time PCR. Three different lines expressing GUS activity and one wild type line were analyzed. Transgenic lines displayed different levels of expression, as expected from independent lines in which T-DNA is probably inserted in different regions of the U. gibba genome (Figure 5a).
Plants transformed with the pRibonuclease-GUS::GFP construct show cell-type-specific expression
The development of plant transformation protocols facilitates the functional characterization of genes and their patterns of expression. To test the U. gibba transformation system for gene expression research, we analyzed the expression of an organ-specific gene. We previously reported the identification of genes that are specifically expressed in U. gibba traps, some of which could be implicated in P uptake from prey digestion [41]. Among the trap-specific genes, we selected one that was very strongly expressed (See Additional file 1: Figure S2). This gene, originally identified as unitig_26.g10301.t1 in the U. gibba genome, was renamed UgRibT2-1. It encodes a protein that is a member of the Ribonuclease-related protein family HOM04D000621. This family has a conserved Ribonuclease_T2 protein domain, and is encoded by 5 genes in A. thaliana, 7 in S. lycopersicum and 3 in U. gibba (https://bioinformatics.psb.ugent.be/plaza; See Additional file 1: Figure S2). Among the biological functions assigned to this RNAse family is phosphate and nitrogen nutrition. In order to compare the expression patterns of UgRibT2-1 and its putative orthologs in Arabidopsis (AT1G14210) and tomato (Solyc05g007940), we examined the expression data available in the developmental maps of the EFP browser (https://bar.utoronto.ca/efp/cgi-bin/efpWeb.cgi; http://bar.utoronto.ca/efp_tomato/cgi-bin/efpWeb.cgi). We found that AT1G14210 and Solyc05g007940 are expressed at low levels in different tissues at several developmental stages; however, both genes are strongly expressed in roots. The 5’-upstream intergenic region of UgRibT2-1 is 1514 nucleotides, which we selected as the promoter region. In silico characterization to search for DNA motifs showed that 2 copies of P1BS are present in the UgRibT2-1 5’ flanking sequences. P1BS is the cognate binding site for PHR1, one of the master regulators of the low-phosphate response in different plant species. Interestingly, the promoter region of AT1G14210 is small and lacks P1BS DNA motifs. However, we found other DNA motifs shared between the promoter regions of UgRibT2-1, AT1G14210 and Solyc05g007940: cognate binding sites for transcription factors of the MYB, bZIP and Homeobox gene families (See Additional file 1: Figure S3). To determine the expression patterns directed by UgRibT2-1 promoter, we created a transcriptional fusion between this promoter and a UidA (GUS)-GFP reporter gene, which we named pRib-GUS::GFP. Using the transformation system reported above, we generated 8 independent transgenic U. gibba lines harboring the UgRib-GUS-GFP construct. Transgenic UgRib-GUS-GFP plants were subjected to histochemical GUS analysis and confocal microscopy to detect the presence of GFP. We found that ribonuclease promoter expression was specific to the trap, more precisely to specialized quadrifid gland cells (Figure 4c, h, i). These cells are involved in transport, digestion and absorption of prey-derived nutrients inside the trap. Therefore, we propose that UgRibT2-1 has a role in nucleic acid degradation during prey digestion (Figure 5b).
To quantify the expression level of pRib-GUS::GFP and compare it with UgRibT2-1 in U. gibba plants, real time PCR analysis was performed. RNAs extracted from transgenic lines were analyzed after collecting traps and vegetative tissue separately. We found that the level of expression directed by the UgRibT2-1 promoter is over 2000 times higher in traps than in vegetative tissues, whereas pRib-GUS::GFP expressed 10 times greater in traps than vegetative tissue. Although the pRib promoter did direct the expected trap-specific expression pattern, its lower expression strength suggests that regulatory elements of the pRib promoter important to determining expression level are missing in the 1500 bp that we used as pRib promoter. Since this promoter includes all of the 5’ flaking intergenic region of RibT2-1, it is possible that an enhancer element is present inside the transcribed region of this gene (Figure 5b).
Detection of UidA (GUS) and phosphinothricin acetyl transferase (BAR) in U. gibba transgenic lines
To confirm that transgenes were integrated into the U. gibba genome, DNA from 5 lines each of p35S-GUS::GFP and pRib-GUS::GFP and 3 plants of wild type were isolated. In transgenic lines, 332pb and 428pb fragments were amplified corresponding to UidA (GUS) and BAR amplicons. In wild type lines, these fragments were not observed. As a positive control, a 239 bp fragment was synthesized corresponding to an endogenous Ubiquitine gene, which can be observed in both transgenic and untransformed control lines (Additional file 1: Figure S4).