- Transformation is a major bottleneck for genetic engineering and gene editing in forest tree species. This includes most genotypes of Populus and Eucalyptus, which are some of the world’s most widely-cultivated genera of plantation forest trees. To provide new tools for transformation, I tested the transcription factor-protein chimera consisting of GROWTH REGULATING FACTOR 4 (GRF4) & GRF-INTERACTING FACTOR 1 (GIF1), which gave large increases in regeneration of transgenic shoots in citrus and other crop species. Using an Agrobacterium organogenic system, I tested various configurations of the chimera, including different levels of miRNA sensitivity and promoter strengths.
I transformed two hybrid poplar genotypes with binary vectors containing an empty vector control and either a 2x35S:miRNA-resistant GRF4-GIF1 chimera from Citrus provided by the Dubcovsky laboratory at the University of California at Davis, or a 1x35S:miRNA-sensitive GRF4-GIF1 chimera from Populus that I created. The poplar GRF4 and GIF1 genes were inferred from the Populus genome following a phylogenetic analysis comparing all Populus GRF and GIF proteins with those from Citrus, Vitis, Eucalyptus, and Oryza, and then I removed the GRF4 stop codon and fused it to a synthetically-derived linker with four amino acids (AAAA) and then to the GIF coding region. A red fluorescent protein-encoding gene was added to both constructs to enable rapid quantification of transgenic regeneration rates. The chimeric GRF-GIF sequences were cloned into an expression vector and poplar as well as eucalypt tissue explants were transformed using Agrobacterium.
I found that the effects of overexpressing the Citrus miRNA-resistant GRF4-GIF1 chimera were highly genotype-specific. A significant negative regeneration phenotype was observed in a P. tremula x tremuloides genotype (‘353-53’) while a P. tremula x alba genotype (‘717-1B4’) had no significant difference in regeneration versus an empty vector control (mean reduction of 95% and 4%). The negative regeneration phenotype in 353-53 and the interaction between genotype and overall effect of Citrus GRF-GIF overexpression were statistically significant (P < 0.01). There was also a statistically significant decrease in ectopic root formation observed on leaf disc explants from 717-1B4 (P < 0.01), suggesting that Citrus GRF4-GIF1 overexpression may suppress rooting in poplar.
Results from overexpression of the Populus miRNA-sensitive GRF4-GIF1 chimera (PtGRF-GIF) were less genotype-dependent. Transgenic shoot regeneration nearly doubled in genotype 717-1B4 over the empty vector control, but the effect did not reach statistical significance (P = 0.07). In genotype 353-53, there was no appreciable change in transformation rate nor did effects approach statistical significance. Interestingly, PtGRF-GIF reduced transgenic callus formation in both genotypes (mean reduction of 65%), which was statistically significant in 353-53 (P = 0.03). All transgenic Populus GRF4-GIF1 shoots from genotype 353-53 died when placed on elongation medium (EM), but no 717-1B4 Populus GRF4-GIF1 shoots died. This difference in shoot viability between genotypes is statistically significant (Fisher’s exact test, P = 0.02).
Non-replicated exploratory experiments with additional GRF-GIF chimeras suggested that GRF4-GIF1 transcript stability (miRNA396-resistance) level was inversely correlated with transgenic shoot regeneration rate. Constructs which were mutated to be miRNA-resistant consistently reduced shoot regeneration in two hybrid poplar genotypes more than a miRNA-sensitive construct from Citrus. Additionally, a dexamethasone-activated miRNA-resistant chimera from Vitis reduced shoot regeneration in both poplar genotypes when dexamethasone was applied. A non-replicated experiment which tested the Populus GRF4-GIF1 in a recalcitrant P. alba genotype (‘6K10’) resulted in a high transgenic shoot regeneration rate (37%), compared to a complete lack of regeneration in an empty vector control (Appendix Figures C.4 & C.5).
Some GRF-GIF constructs were also tested in hybrid eucalypt, however many of the experiments suffered from poor transformation efficiency, tissue necrosis, and contamination. Overexpression of the Citrus miRNA-resistant GRF4-GIF1 chimera in two hybrid eucalypt genotypes increased the prevalence of transgenic callus (mean increase of 35%), but the effect did not reach statistical significance in either genotype (P = 0.12 & 0.9). Additionally, a Eucalyptus grandis GRF5-GIF1 chimera which I produced had unclear effects on transformation and regeneration in a non-replicated experiment. The dexamethasone-activated miRNA-resistant chimera from Vitis increased the prevalence of transgenic callus and shoot primordia in two eucalypt genotypes (mean increase of 28%).
My results suggest that certain GRF-GIF chimeras may increase the regeneration rate of transgenic shoots in select genotypes of Populus, while having a negative impact on the root and callus development pathway in other genotypes. A larger study with more tree genotypes, GRF-GIF sources and miRNA sensitivities, and a wider array of promoter strengths and/or an inducible or excision system, may help shed light on the wider potential of GRF-GIF to promote regeneration of transgenic poplars and eucalypts.