<i>In vitro</i> floral development in poplar: insights into seed trichome regulation and trimonoecy

Abstract

Woody perennials including Populus spp. (poplars) have a juvenile phase that ranges from several years to decades in length. This and the year-long floral development process are major impediments to breeding and to fundamental research of reproductive traits. Here, we report a CRISPR-empowered in vitro flowering system and demonstrate its application using three reproductive traits: sex, seed trichomes, and a previously undescribed potential for trimonoecy in poplar. Several regulators of floral development, such as LEAFY (LFY) and FLOWERING LOCUS-T (FT), have been used to induce precocious flowering in annual models and fruit trees (Weigel & Nilsson, 1995; Callahan et al., 2016). Translating these findings into the poplar system has met with various challenges, including dwarfism and sterility (Zhang et al., 2010). Using heat-inducible promoters for FT expression has circumvented many of the developmental anomalies; however, repeated application of heat treatments over weeks or months can be detrimental to microsporogenesis (Hoenicka et al., 2014). The method's efficacy is also season- and genotype-dependent (Zhang et al., 2010; Hoenicka et al., 2016), limiting its widespread adoption. Here, we present an induction-free, in vitro flowering system by targeting a negative regulator of floral initiation, CENTRORADIALIS (CEN, also called TERMINAL FLOWER), for knockout (KO). CEN antagonizes FT and LFY to regulate meristem determinacy and flowering (Bradley et al., 1997; Jaeger et al., 2013). RNAi silencing of CEN orthologs in poplar and pear (Pyrus communis) shortened their juvenile phase to under 3 yr (Mohamed et al., 2010; Freiman et al., 2012), and CRISPR-KO of CENs in kiwi (Actinidia chinensis) reduced flowering time to under 1 yr (Varkonyi-Gasic et al., 2019). For the present work, we adopted CRISPR/Cas9 to edit the CEN1/CEN2 paralogs in a female Populus tremula × Populus alba INRA 717-1B4 hybrid (hereafter 717) (Supporting Information Methods S1). In vitro flowering of rooted plantlets was observed under long-day tissue culture conditions within 4 months of Agrobacterium transformation. During vegetative propagation of the mutants, single flowers developed directly from axillary buds of subcultured stem segments and decapitated mother plants within 1–2 wk (Fig. 1a). Amplicon sequencing confirmed all 17 flowering events as cen1cen2 double-KOs (Dataset S1; Table S1). In vitro flowering was reproducible in nodal cultures where carpels with a cupular disk developed directly from axillary buds (Fig. 1b). This effectively fast-tracked multiseason floral organogenesis to a timeframe of days. The system represents a significant improvement, both in response time and in efficiency, over the female organ culture approach pioneered by Bawa & Stettler (1972) half a century ago. We next asked whether both CEN paralogs are involved in floral development as only CEN1 transcripts are detected in shoot tissues (Fig. S1). We generated 10 cen1 single-KO events with biallelic mutations (Dataset S1; Table S1) and observed in vitro flowering phenotypes similar to the cen1cen2 double-KOs (Fig. S2). The findings suggest that CEN1, but not CEN2, represses poplar floral development. Similar results were independently reported by another group (Sheng et al., 2022). A female-specific, type-A response regulator (ARR17) gene was recently identified as the sole sex determinant in poplars with either the XY (most Populus spp.) or ZW (P. alba) system (Müller et al., 2020). ARR17-KO in an early-flowering female P. tremula triggered male flower development (Müller et al., 2020). Interestingly, 717 has a hybrid sex configuration (♀, XZ) derived from P. tremula (♀, XX) and P. alba (♂, ZZ), which is supported by the detection of hemizygous ARR17 in the haplotype-resolved 717 draft genome (Phytozome). Monoallelic ARR17-KO is predicted to convert 717 from female to male, and this was tested using multiplex editing of ARR17 and CENs to determine the sex-switch outcome in vitro (Methods S1). We observed male flowers in all eight arr17cen1cen2 events (Fig. 1c,d) and confirmed targeted mutations in all cases (Dataset S1; Table S1). Our results support ARR17-dependent sex switch in a XZ background. The fast-track flowering system is more time- and labor-efficient than the heat-inducible-FT method (Zhang et al., 2010; Hoenicka et al., 2016; Müller et al., 2020) for studying floral traits in poplars. Soil-transplanted female and male mutants (8–10 events per group) flowered in terminal as well as axillary buds after acclimation (Fig. 1e,f). The indeterminate-to-determinate meristem conversion resulted in cessation of vegetative growth, accelerated maturation with thickening and darkening of preexisting leaves, and following repeated cutbacks, prolific root suckers not seen in WT (Fig. S3). Healthy suckers and sprouts grew vegetatively for several weeks before they terminated into flowers, whereas cutback triggered flowering from axillary buds within days (Fig. 1g,h). Unfertilized carpels still matured into seedpod-like capsules, which eventually opened and released cottony hairs (Fig. 1i,j). These phenotypes remained consistent across all mutant (♀) lines, over multiple rounds of transplanting, and for over a year. In comparison, male flower development of tissue culture and soil-grown mutants (♂) appeared constrained under the experimental conditions. Anther development was frequently aborted, though pollen shedding was occasionally observed (Fig. S4). Future optimization of growth conditions is necessary to improve pollen collection for fertility testing. To further exemplify the utility of in vitro flowering in reproductive trait investigation, we asked whether a group of MYB transcription factors recently shown to be essential for leaf and stem trichome initiation (Bewg et al., 2022) also regulate seed trichome development. Seeds with tufted hairs are characteristic of poplars and willows, with the cottony trichomes facilitating wind dispersal of seeds. In urban and plantation forestry, seed hairs are carriers of airborne allergens representing a potential health hazard (Hu et al., 2008). Multiplex-KO of CENs and trichome-regulating MYBs produced 11 early-flowering (♀) glabrous events (Fig. S5), with confirmed edits at all 12 (eight MYB and four CEN) target sites (Dataset S1; Table S1). We compared in vitro carpel development between trichome-bearing and trichomeless (♀) mutants. Ovules were already visible in the immature carpels of cen1cen2 mutants we bisected (Fig. 1k). Intra-ovarian trichomes were not observed until carpels reached c. 2 mm in length and gradually filled the ovary during maturation (Fig. 1l–n). An abundance of trichomes remained attached to ovules, which resembled comose seeds (Fig. 1m inset). In glabrous mutants, ovules but not intra-ovarian trichomes were observed throughout carpel development in all three events examined (Fig. 1o–r). The results suggest seed trichomes, like other aerial organ trichomes, are regulated by the same MYBs in poplar and provide a molecular basis for engineering hairless seeds for genetic confinement or for reducing allergen spread in urban/plantation forestry. Finally, monitoring of cen1cen2 nodal cultures revealed a remarkable diversity of sex morphs in 717 (♀, XZ), with male, female, and perfect flower development (trimonoecy) dictated by stem node position on the mother plant. In multiple cen1cen2 events, we observed male flowers with stamens only from subapical nodes (Fig. 2a–c), perfect flowers with carpels and stamens from upper nodes (Fig. 2d–f, see also Fig. 1m,n,p), female flowers only from middle nodes (Fig. 2g–i), and vegetative buds from older nodes (Fig. 2j–l). The unusual stamen appearance (♂ or ⚥) was transient, limited to just a few nodes near the top. Such a developmental gradient is difficult to capture in soil-grown cen1cen2 mutants because of the pleiotropic phenotypes as discussed previously. Nevertheless, carpellate flowers with stamens or stamen-carpel chimeras were occasionally observed and always near the top of the (cutback) plant (Fig. S6), consistent with transient male organ development. The appearance of perfect flowers in some poplars, while rare, has also been reported in the field (Stettler, 1971; Boes & Strauss, 1994; Zhang et al., 2010). Overall, the data suggest a role for ontogenic regulation on poplar sex determination that warrants further research. In sum, the in vitro flowering system bypasses the multiyear reproductive phase transition in poplar and fast-tracks the year-long floral development process to days and weeks. It offers a facile model for investigating floral traits and holds promise for rapid-cycle breeding and genomic selection in perennial trees. We thank Gilles Pilate of INRAE, France, for providing poplar clone 717, Makenzie Drowns and Brent Lieb for amplicon library assistance, Kimberly McInnes for plant care, and Georgia Genomics and Bioinformatics Core for Illumina sequencing. The work was funded in part by The Center for Bioenergy Innovation, a US Department of Energy Research Center supported by the Office of Biological and Environmental Research in the DOE Office of Science, and the Georgia Research Alliance-Hank Haynes Forest Biotechnology Endowment. None declared. C-JT and RZ conceived the study. C-JT and MAO designed the experiments. MSSC, MOA, WPB, RZ and BS conducted the experiments. C-JT wrote the paper with contributions from MAO and RZ. The data that support the findings of this study are available in the Supporting Information for this article. Dataset S1 Summary of mutation patterns determined by amplicon sequencing. Fig. S1 Expression of CEN1 and CEN2 in various poplar tissues. Fig. S2 In vitro flowering of representative cen1 poplar mutants. Fig. S3 Phenotypes of soil-grown poplar mutants. Fig. S4 Pollen shedding of soil-grown arr17cen1cen2 (♂) poplar mutants. Fig. S5 In vitro flowering of representative glabrous cen1cen2myb (♀) poplar mutants. Fig. S6 Chimeric and abnormal male structure in cen1cen2 (♀) poplar mutants. Methods S1 Materials and methods. Table S1 gRNA, oligo, and synthetic fragment sequences from Populus tremula × Populus alba. Please note: Wiley is not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts