The study, published in Plant Cell Reports by Ninh Hai Ho, Marcus Linde and Thomas Debener of Leibniz University Hannover, focused on transient expression, the short-lived production of a protein from DNA that has been delivered into cells but not stably integrated into the genome. Transient assays are prized in plant research because they allow scientists to test promoters, reporters and candidate genes within days rather than the many months required to regenerate a fully transgenic plant. In roses, however, even this shortcut is complicated by the fact that different varieties respond very differently to infiltration, and until now nobody had examined that variation at the level of the whole genome. Stable transformation of rose remains laborious, genotype-dependent and unavailable as a routine tool for most cultivars, making fast transient systems especially valuable for a species whose genome harbors genes controlling flower colour, scent, senescence, stress responses and plant architecture.

To dissect the genetic basis of this variation, the researchers assembled an association panel of 96 rose genotypes, 87 of them tetraploid, alongside eight triploid and one diploid variety, grown under semi-controlled greenhouse conditions at the Federal Plant Variety Office in Hannover. From flowers at the initial and full bloom stages, the developmental windows previously shown to be most amenable to petal agroinfiltration, the team harvested petals from the middle of each flower and infiltrated them from the abaxial side using a needleless syringe. The bacterial strain carried a T-DNA containing an intron-containing green fluorescent protein reporter driven by the Arabidopsis ubiquitin 10 promoter, a construct designed so that any detectable fluorescence would reflect successful delivery, nuclear import and expression of the transgene inside rose cells. The petals were then incubated and scored for fluorescence at three and five days after infiltration.

The phenotypic results were striking. Using a five-class ordinal scale ranging from no detectable expression to very strong expression, the researchers found that mean scores at five days post-infiltration ranged from zero to 3.98 across the panel, an almost complete sweep of the possible range. At three days, 33 of the 96 genotypes showed no GFP signal at all, but by five days only 13 remained dark, and the number of genotypes showing strong or very strong expression increased over time. Statistical testing confirmed that genotypic effects were highly significant at both time points, and the scores from the two time points correlated strongly, with a Pearson coefficient of 0.90. Cultivars such as Sebastian Kneipp, Friesia and Comtessa AL emerged as the most permissive genotypes, with Sebastian Kneipp reaching a mean score of 3.98 at the five-day endpoint. These varieties, the authors suggest, are immediate candidates for researchers seeking reliable petal-based transient assays in rose.

With the phenotype quantified, the team turned to genome-wide association analysis, using genotyping data from the WagRhSNP 68K Axiom SNP array. After quality filtering, 37,161 high-quality single-nucleotide polymorphisms remained for analysis. The researchers employed GWASpoly, software designed for autopolyploid species that tests different allele-dosage models, and controlled for population structure and relatedness with a kinship matrix and principal components. The analysis revealed a major association on chromosome 2, in a region spanning roughly four million base pairs from 69 to 73 megabase pairs. The strongest marker, at 70.13 megabase pairs, reached a significance of minus log10 P equals 7.85 under a simplex dominance model, well beyond the genome-wide threshold of 5.54. A second peak appeared on so-called chromosome 0 contigs, sequences not yet assigned to any of the seven rose pseudochromosomes, but follow-up sequence comparisons indicated that these unanchored fragments most likely belong to the same chromosome 2 region.

Within the associated interval, the researchers identified a collection of candidate genes whose annotated functions read like a checklist of the host processes known to matter during Agrobacterium-mediated transformation. Among them are an ABC transporter B family member 19 homologue involved in auxin transport, a Cullin 3 homologue representing the ubiquitin-mediated protein degradation machinery, DExH-box RNA helicases and an RNA-binding protein that could influence the processing of delivered transcripts, a hypersensitive-induced response protein 1-like membrane protein tied to defence signalling, monodehydroascorbate reductase 4, and several nuclear pore complex protein GP210-like sequences. Each of these annotations is biologically plausible. The ubiquitin-proteasome system has been implicated in the turnover of T-complex-associated proteins during bacterial DNA transfer, nuclear import is an essential step between T-DNA delivery and reporter expression, and plant immune signalling is well known to restrict transformation efficiency, as illustrated by the MKK4/5-MPK3/6 cascade that modulates Agrobacterium transformation in Arabidopsis.

The allele-dosage patterns at representative SNPs added further weight to the association. For two markers near 70.13 to 70.16 megabase pairs, genotypes carrying a low allele dosage showed substantially higher median GFP scores than other dosage classes, while for two other markers the opposite homozygous or quadruplex dose class presented the highest expression. Among cultivars in the high-expression dosage classes for two of these SNPs, 25 of 46, or 54.3 percent, reached GFP scores of two or better. The authors are careful, however, to frame the finding appropriately. With only 96 genotypes in the panel, mapping resolution is limited, and local linkage disequilibrium in the panel decays over megabase-scale distances. The chromosome 2 interval should therefore be treated as a high-priority candidate region containing multiple potential host factors rather than as evidence for a single causal gene, and larger panels combined with candidate-gene sequencing will be needed to pinpoint the responsible variants.

An intriguing wrinkle emerged when the team compared their GFP expression data with fragrance scores previously collected from the same association panel. The chromosome 2 region overlaps with marker clusters previously linked to fragrance-related variation, yet direct comparison revealed only a weak positive correlation of 0.27, explaining less than eight percent of the phenotypic variance. This suggests that the two traits are likely governed by different but linked loci within the same broad genomic region rather than by a single gene with pleiotropic effects, a distinction that future fine-mapping work will need to resolve.

The authors also emphasize what their phenotype does and does not measure. The score reflects visible transient GFP expression, not bacterial migration, T-DNA copy number or stable transformation efficiency. Low expression could stem from many causes, including reduced bacterial entry, impaired T-DNA transfer, rapid silencing, tissue necrosis, cuticle properties or defence activation, while high transient expression does not guarantee efficient regeneration of stable transgenic lines. Future work, they suggest, should incorporate quantitative image analysis of fluorescence, histological assessment of petal tissue and direct measurement of bacterial load, alongside sequencing and functional testing of the strongest candidate genes in contrasting genotypes.

Even with those caveats, the study delivers practical value immediately. The permissive cultivars identified in the screen give rose researchers a shortlist of genotypes for rapid petal-based gene function tests, potentially accelerating work on petal abscission, dehydration tolerance and pigmentation before any stable transformation is attempted. More broadly, the finding that transformation responsiveness behaves as a genetically tractable trait, echoing earlier quantitative trait locus work in Brassica oleracea and a recent GWAS of hairy root formation in rose, opens a new frontier in understanding the host side of Agrobacterium-mediated gene transfer. If the chromosome 2 interval can be narrowed and its candidate genes validated, the humble rose may yet teach biotechnologists a great deal about why some cells say yes to foreign DNA and others say no.

Subject of Research: Genetic control of Agrobacterium-mediated transient GFP expression in rose petals

Article Title: A chromosome 2 locus controls transient Agrobacterium-mediated GFP expression in rose petals

Article References: A chromosome 2 locus controls transient Agrobacterium-mediated GFP expression in rose petals. (n.d.). https://doi.org/10.1007/s00299-026-03961-z

Image Credits: AI Generated

DOI: 10.1007/s00299-026-03961-z

Keywords: rose, Agrobacterium tumefaciens, transient expression, GFP, genome-wide association study, chromosome 2, petal agroinfiltration, tetraploid rose, candidate genes, plant transformation, functional genomics, host susceptibility