Introduction

Since the first successful production of genetically engineered (GE) tobacco via Agrobacterium tumefaciens inoculation1, stable plant transformation has served as an important means to uncover the genetic control of plant growth and development, and to realize crop improvement targets, such as enhanced productivity, nutrition, and stress tolerance2,3. Entering the era of plant synthetic biology (SynBio), where engineering principles are applied to plant biology, A. tumefaciens serves as a foundational tool for introducing genetic circuits to plants and testing the efficacy of genetic modification on modifying existing functions or introducing novel genes and pathways4,5. Driven by the engineering concept of an iterative use of design-build-test-learn (DBTL) cycles to improve genetic circuits6, plant SynBio requires efficient transformation methods that ensure GE plant production within a short timeframe and are feasible in a wide number of plant species, especially non-model species with commercial or ecological importance.

Although successful transformation has been reported for many plant species, A. tumefaciens-mediated stable transformation is slow and costly and remains intractable in a wide-range of plant species. With the exception of floral dip transformation method such as in Arabidopsis7,8, in all other tissue culture-based plant transformation, existing protocols share similar procedures and rely on the occurrence of three consecutive events to successfully recover GE plants, where the first event is delivery of foreign or modified native DNA into plant cells. This entails co-culturing excised plant parts, typically leaves, called leaf explants with A. tumefaciens that carry the DNA of interest, washing and transferring of the infected explants onto phytohormone-containing media for organogenesis. The second event that ensures recovery of GE plants, occurs in the early organogenesis stage and, entails stable integration of the introduced DNA fragment into the plant genome9. The final event is regeneration of GE plants by controlling the relative levels of phytohormones, auxin and cytokinin, in the media, where a high auxin to cytokinin ratio induces callus and root formation, while low auxin to cytokinin ratio facilitates shoot formation10. From initiation of A. tumefaciens inoculation, it generally takes several months to obtain rooted GE plants in vitro. This turnaround time is plant type-dependent, for example, the average transformation timeline for poplar plants is at least six months11,12,13. Compounding the slow turnaround time in obtaining rooted transformant is the low recovery rate of stably transformed GE plants, which can be lower than 10%, even in plant species with well-established transformation protocols14. Because many plant species and genotypes are recalcitrant to in vitro regeneration15,16, production GE varieties via A. tumefaciens-mediated methods have been restricted to a small number of well-studied plant species. These bottlenecks in plant transformation and regeneration prompt the desire for improving existing transformation protocols and developing new routes towards more efficient GE plant production.

Even in plant species where a transformation system has been established, there is still increasing need for improving GE plant production. As the woody model organisms of economic and ecological importance17,18,19, many species and hybrids from the genus Populus, particularly those from sections Populus (aspens and white poplar), Aigeiros (cottonwood), and Tacamahaca (balsam poplar), have been frequently transformed with A. tumefaciens. Explants from leaves, stems, and petioles are used singly or in combinations in established transformation protocols, depending on the species. For example, in aspens, leaf explants are commonly used as they tend to generate the highest success rate among different types of explants. In cottonwoods, however, petioles and stems perform better than leaves, and therefore are the preferred sources for explants20. Yet few existing protocols employ root explants. In contrast, in the herbaceous model plant, Arabidopsis, root has been employed in stable transformation21,22. Transformations with root explants tend to show higher percentage of single T-DNA insertions and a lower rate of polyploidy, compared with leaf-disc transformation23. There have been previous reports of root derived explants from other plant species, such as eggplant, spinach, and Medicago truncatula24,25,26,27. Given that Populus root cells are competent for regeneration, as demonstrated in organogenesis-focused studies28,29,30,31,32, employing explants from the entire Populus plant, including the above- and below-ground tissues, in transformation is a logical approach to maximize the utility of plant material and improve the transformation productivity per unit plant form. However, reports on systematic, side-by-side comparative assessment of explant type-tracked tissue culture and genetic transformation outcomes are lacking. Last but not the least, in recent years, modification of morphogenic and growth regulators to overcome tissue culture regeneration of transformants is heralded as a new route to tackling regeneration recalcitrance and maximizing efficiency of stable plant transformation16,33,34. Progress in gaining a molecular level understanding of the tissue culture and stable transformation processes is far advanced in model plants such as rice and tobacco compared that in bioenergy model species such as poplar, switchgrass and miscanthus35,36,37.

Here we report results from a study exploring the feasibility of using a wider range of explant sources from the bioenergy crop model, Populus, as a route to enhance transformation efficiency. Specifically, we evaluated feasibility and performance of variable tissue explants (leaf, stem, petiole, and root) derived from Populus clone 717-1B4 (P. tremula × P. alba clone INRA 717-1B4) in stable transformation assays. We demonstrated that roots, together with above-ground tissues from in vitro plants can serve as viable explant sources for A. tumefaciens-mediated stable transformation. We complemented the systematic characterization of regeneration and transformation efficiency with transcriptome profiling of leaf and root explant types to shed light on the molecular changes underlying the observed morphogenic changes. The collective information provides potential new avenues to maximize tissue culture productivity and throughput in combination with future technology development to meet increasing demands for plant variant lines and to, ultimately, accelerate SynBio DBTL cycles and crop improvement efforts.

Materials and methods

Tissue culture, plant transformation, and transgenic plant production

The eYGFPuv plasmid38 was firstly transformed into A. tumefaciens strain ‘GV3101’ using electroporation, and transformed into P. tremula × P. alba clone INRA 717-1B4, following published methods39. The hybrid aspen 717-1B4 tissue culture stock plant material was provided by Dr. Steven Strauss (Oregon State University). Briefly, explants were collected from 6 to 8 week-old in vitro 717-1B4 plants using scalpels. All plant parts, including leaf, petiole, stem (internode), node, and root, were included as explants. These explants, regardless of their source, were co-cultivated with A. tumefaciens culture carrying eYGFPuv in dark for two days before being washed with sterile water and transferred on to callus induction medium. After three weeks of callus induction in dark, explants were cultivated on shoot induction medium for eight weeks, followed by cultivation on shoot elongation medium for four weeks. Elongated shoots were harvested from individual explants and rooted on rooting medium. All growth media used in this study were MS (Murashige and Skoog medium) based, supplemented with different phytohormones (i.e., 10 μM NAA and 5 μM 2ip for callus induction, 0.2 μM thidiazuron (TDZ) for shoot induction, 0.1 μM 6-benzylaminopurine (BAP) for shoot elongation, and 0.5 µM Indole-3-butyric acid (IBA) for rooting). Starting from the callus induction stage, 50 mg/L kanamycin was used to select for transgenic tissue, and 200 mg/L timentin was used to select against A. tumefaciens during the entire regeneration process. All plant materials were maintained in a tissue culture room with a temperature of 25(± 1) °C. Apart from the initial callus induction stage when explants were recovered in dark, plant materials were cultured with a 16/8 h light–dark cycle. The eYGFPuv was monitored using a hand-held 365 nm UV flashlight. Images were recorded with a Nikon camera (model D7500; Nikon Corporation, Tokyo, Japan).

Greenhouse study of transgenic plants

Growth and phenotyping of transgenic plants was undertaken under greenhouse conditions set at 25 ℃, 50–70% humidity and a 16 h photoperiod according to prior published methods40. Prior to propagation of tissue culture plants in the greenhouse, tissue culture plants were subjected to a four-week acclimation phase consisting of removal of the tissue culture container lid and enclosure of the container in a loosely closely one gallon Ziploc bag. Plants were slowly acclimated from high humidity closed container conditions to greenhouse conditions in a walk-in a growth chamber set at growth conditions (25 ℃, 80% humidity and 16 h photoperiod) similar to the greenhouse. This was followed by transplantation of plants into potting mix and transfer to a greenhouse misting bench for the final two-week phase of acclimation.

Statistical analysis of regeneration and transformation efficiency

We collected one shoot per independent callus (event) for further downstream propagation and regeneration of transgenic plants to ensure the regenerated shoots/plantlets are independent from each other. We followed this standard approach while collecting samples for RNAseq and calculating transformation efficiency. We defined regeneration rate as the ratio of the total number of explants with regenerated shoots at the end of shoot induction stage to the total number of explants, and transformation rate as the ratio of the total number of independent transgenic events to the total number of explants co-cultured with A. tumefaciens. We performed ANOVA in R to examine whether there were significant differences (p < 0.05) in regeneration rate, transformation rate, and physiological data among different types of explants. Error bars represent standard errors.

RNA-Seq library preparation and sequencing

Whole leaf and root explants were collected at four different time points (i.e., freshly excised from in vitro plants, three weeks after callus induction, one week after shoot induction, and two weeks after shoot induction) were collected for RNA extraction (Supplemental Fig. S1). RNA from frozen ground tissue samples was extracted using a Plant Total RNA extraction kit (Sigma, St Louis, MO) with modifications to the kit protocol. Briefly, 100 mg of frozen ground tissue was incubated at 65 ℃ in 850 µl of a 2% CTAB + 1% βme buffer for 5 min. This was followed by the addition of 600ul of chloroform:isoamylalcohol (24:1 v/v). The mixture was spun at full speed in a centrifuge for 8 min at which point the supernatant in the top layer was carefully removed and passed through a filtration column included in the kit. The filtered elutant was diluted with 500 µl of 100% EtOH and passed through a binding column. This was repeated until all the filtered elutant/EtOH mixture was passed through the binding column. Further steps including on-column DNase digestion (DNase70, Sigma), filter washes, and total RNA elution were followed as per the manufacturer’s protocol. The total RNA concentration, 260/280 and 260/230 values were checked on a NanoDrop 2000™ (Thermo Scientific). RNA samples were then shipped to BGI Americas where RNAseq was performed on a DNBseq platform after DNA nanoball library synthesis.

Analysis of RNA-Seq data and identification of differentially expressed genes (DEGs)

The clean RNA-Seq reads from BGI were mapped onto the Populus trichocarpa genome (v4.1)41 using STAR (version 2.7.8a)42. The RNA-seq alignments in BAM format were used to count the number of reads per annotated gene and calculate transcripts per million (TPM) using TPMCalculator43. Differential gene expression analysis was performed using DESeq244 implemented in Trinity45.

Results

Root explants showed comparable transformation efficiency to other explants types

To evaluate the regeneration capacity and amenability to A. tumefaciens of root explants, we performed three independent transformation experiments. In each transformation, we co-cultivated a total of ~ 100 explants derived from leaf, petiole/stem, and roots (Supplemental Table S1) using the A. tumefaciens strain GV3101 carrying the eYGFPuv gene. We observed callus and shoot formation among all tested explant types (Fig. 1a), with an average regeneration rate of 94.3% in leaf, 64.7% in petiole/stem, and 52.4% in root (Fig. 1b). While regenerated shoots from root explants were visible, the shoot regeneration appeared to be slower compared with leaf and petiole/stem explants at the given temporal stage (Fig. 1b). By visualizing the eYGFPuv expression (Fig. 2a), we obtained GE shoots from leaf, petiole/stem, and root explants at an average rate of 25.0, 17.6, and 15.0%, respectively (Fig. 2b). Despite the apparent lower average regeneration rate and transformation efficiency in root-derived explants, the differences in regeneration capacity and transformation amenability were not statistically significant among the tested explant types (p = 0.0905 for regeneration rate and p = 0.356 for transformation rate) based on one-way analysis of variance (ANOVA). We also undertook Student’s t-test analysis on regeneration and transformation efficiencies between any two explant types and found only one instance of significant difference (p = 0.029) i.e., when regeneration rates of leaf and root explant types were compared (Supplemental Table S1). We followed the growth and establishment of regenerated plants in the greenhouse and observed that all plants continued showing the eYGFPuv expression under greenhouse conditions and had comparable growth with no significant differences in plant height, stem diameter, leaf fresh weight or stem fresh weight properties of transgenic plants regenerated from leaf, petiole/stem, and root explants (Supplemental Fig. S2).

Fig. 1
figure 1

Regeneration capacity among different types of explants. (a) Calli and shoots regenerated from leaf, petiole/stem, and root explants, respectively. (b) Regeneration efficiency among different types of explants. Error bars denote standard errors over three independent transformation experiments.

Full size image
Fig. 2
figure 2

Recovery of genetically engineered (GE) plants from different types of explants. (a) Detection of GE calli and shoots, indicated by the GFP signal, from root explants. (b) Transformation efficiency among different types of explants. Error bars denote standard errors over three independent transformation experiments.

Full size image

Leaf and root explants shared overlap in transcriptome changes during early organogenesis

To explore the similarity and divergence among different types of explants at the molecular level, we performed RNA-sequencing (RNA-Seq) and examined transcriptome changes in leaf and root explants during callus induction and early shoot regeneration. We built and sequenced a total of 24 RNA-Seq libraries with three biological replicates per collection timepoint per explant type. For each library, we obtained 20.4 million to 26.1 million pairs of clean paired-end (2 × 100 bp) reads from Illumina sequencing, and 18.3 million to 23.3 million pairs of mapped reads from STAR-based mapping (Supplementary Table S2). StringTie-based transcript assembly and quantification revealed that a total of 34,083 gene models were expressed in the collected samples originating from leaf, root, and their differentiated and regenerated tissue samples. All the sequencing data were deposited in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) database with BioProject ID: PRJNA1093501.

Principal component analysis (PCA) of regularized-log-transformed read counts and heatmap of sample-to-sample distances, both of which were generated using the Bioconductor R package DESeq2, clustered the tissue culture samples into three major groups: leaf explants that were directly excised from in vitro plants, root explants that were directly excised from in vitro plants and explants with regenerated callus or shoots (Fig. 3a,b). Pairwise comparisons revealed that 14,164 (41.6%) and 8,794 (25.8%) of the expressed 34,083 genes were differentially expressed (FDR < 0.05, LFC cutoff = 1) in leaf explants and root explants, respectively, during the sample collection period (Supplementary Table S3). When focusing on the time-course comparisons (i.e., analyses performed with samples collected at two adjacent timepoints), we noticed that, in both leaf and root explants, the total numbers of DEGs decreased as organogenesis progressed. In leaf explants, for example, we discovered a total of 11,293 DEGs when comparing samples collated at the end of callus induction stage to samples just excised from in vitro plants, and only 6 DEGs when comparing samples that underwent two weeks of root induction to those that underwent one week of shoot induction. Similarly, in root explants, we identified a total of 6333 DEGs as we compared samples at the end of callus induction to those excised from in vitro plants; we were unable to identify any DEGs when comparing samples collected after two weeks of root induction and those after one week of shoot induction. To reveal the number of DEGs that are shared by and unique to leaf- and root-based analyses, we generated Venn diagrams with the DEGs obtained from the time-course comparisons (Supplemental Fig. S3). We found 9.2 to 40.8% of the DEGs were shared between leaf and root explants as compared to their respective samples collected at the previous timepoint.

Fig. 3
figure 3

RNA sequencing data suggested similarity and divergence in between leaf and root explants. (a) PCA lot of explant samples. (b) Heatmap of sample-to-sample distances. LA leaf explant just excised from in vitro plants, LC leaf explant after three weeks on callus induction medium, LS07D leaf explant after one week on shoot induction medium, LS14D leaf explant after two weeks on shoot induction medium, RA leaf explant just excised from in vitro plants, RC root explant after three weeks on callus induction medium, RS07D root explant after one week on shoot induction medium, RS14D root explant after two weeks on shoot induction medium.

Full size image

Expression of hormone signaling pathway genes in leaf and root explants during early organogenesis

Given auxin and cytokinin are two phytohormones that are commonly used in in vitro organogenesis, we examined genes involved in their biosynthesis and signaling pathways. In total, we found 73 auxin-related genes that were differentially expressed in leaf and/or root explants, with 35 of them identified in both types of explants (Fig. 4a,b, and Supplementary Table S4).

Fig. 4
figure 4

Identification of auxin-related gene among DEGs. (a) Number of auxin-related gene that are differentially expressed in leaf and/root. (b) Expression profile of the common 35 genes shared by leaf and root.

Full size image

Similarly, we found a total of 12 cytokinin-related genes that were differentially expressed during early organogenesis, and eight of them were shared by leaf and root explants (Fig. 5 and Supplementary Table S5), suggesting that for genes that were differentially expressed in leaf and root explants, a majority of them were shared.

Fig. 5
figure 5

Number of cytokinin-related gene that are differentially expressed in leaf and/root.

Full size image

Expression of developmental genes in leaf and root explants during early organogenesis

To explore the role of developmental genes in the poplar organogenesis system, we examined the expression profile of 17 Populus genes that are homologous to 11 Arabidopsis developmental genes (Table 1). Nine genes, homologous to PLT3/5/7, CUC1/2, WOX11/12, WOX5/7, and WUS, respectively, were differentially expressed in leaf and/or root explants (Table 1 and Fig. 6).

Table 1 List of developmental genes that regulate organogenesis.
Full size table
Fig. 6
figure 6

Expression profile of developmental genes.

Full size image

Discussion

It is widely recognized that the burgeoning needs in the research community for experimental exploration of predicted genomic information via advanced synthetic biology methods and faster Design-Build-Test-Learn (DBTL) cycles are ill-matched by the efficacy and productivity of current plant stable transformation methods. Long-standing issues of high tissue culture recalcitrance and low transformation productivity of a wide range of plant species, especially dedicated food and bioenergy crops, if addressed either through optimization of existing approaches or through new molecular or tissue culture approaches and development of game-changing new technologies based on automation and artificial intelligence (AI), can accelerate both fundamental and applied research with crop species. The present study explored the feasibility of a specific approach; that of maximizing plant transformation productivity using all plant parts as explants in tissue culture, followed by testing of stable transformant generation, and transcriptome profiling to establish both consistency of molecular makeup as well as identification of potential significantly differential candidate morphogenic factors. Our results show that the expansion of the range of tissue culture explants for Populus is not only feasible it also aids in maximizing utility of stock plant material to improve the transformation productivity per unit plant form. The transgenic plants regenerated from different explant sources showed similar morphology. Our transcriptome analyses suggest the reset in physiological status at the callus stage across varied source tissue is comparable and independent of initial explant organ type.

Transcriptome analyses further revealed the expression of the hormone signaling pathway such as auxin and cytokinin signaling genes in leaf and root explants during early organogenesis and identified expression of developmental genes such as PLT3/5/7, CUC1/2, WOX11/12, WOX5/7, and WUS in leaf and root explants during early organogenesis for which, there is extensive support from studies in model organisms, Arabidopsis and tobacco. For example, experimental testing of developmental regulators, such as LEAFY COTYLEDON 1, LEAFY COTYLEDON 2, WUSCHEL (WUS), BABY BOOM (BBM), and the GRF-GIF chimeric protein, via overexpression of native or modified genes has been shown to boost embryogenesis in monocots46,47,48,49, heralding a new route to tackling regeneration recalcitrance. Additional work in Arabidopsis has expanded the list of genes proposed to take part in regeneration pathways. Studies in Arabidopsis have shown that, on auxin-rich callus induction medium, PLETHORA 3 (PLT3), PLT5, and PLT7 activates the expression of root meristem genes PLT1, PLT2, CUP-SHAPED COTYLEDON 1 (CUC1) and CUC2. These regeneration-competent cells then express WUSCHEL RELATED HOMEOBOX 11 (WOX11) and WOX12, and form root funder cells, followed by direct activation of WOX5/7 and LATERAL ORGAN BOUNDARIES DOMAIN 16 (LBD16). The shoot meristem identity gene WUS, whose expression can be activated by a group of B-type ARABIDOPSIS RESPONSE REGULATORs (ARR-Bs), is expressed after transferring explants onto cytokinin rich shoot induction medium. The expression of SHOOT MERISTEMLESS (STM), which is activated by CUC1/2, has also been found to be required for shoot formation10,50,51. These prior studies attest to the value of transcriptome profiling to identify strong candidate genes52,53 to be tested for their functional roles in tissue culture morphogenesis, and the present study uniquely adds to that database with new information from Populus, expanding the opportunities for the bioenergy feedstock improvement community to create and deploy advanced genetic tools with improving transformation efficiency.

In summary, results from this study expand the type of explants that can be employed for transformation in Populus and present new routes to maximize productivity and efficiency of a stable transformation process. We additionally provide a rich transcriptome dataset collected from transformants at various tissue culture/developmental stages for further in-depth studies to correlate factors associated with higher transformability. In the future, combining these methods, with other emerging stable, transient and cell free approaches will provide an all-hands-on-the-deck approach to achieving higher efficacy and productivity in currently manually undertaken plant genetic transformation processes. Finally, leveraging the new frontier technologies in plant sciences, such as automation and AI, will usher in a truly transformative approach to maximizing throughput while simultaneously minimizing risk and enable SynBio-enabled, faster plant DBTL cycles.