Litcius/Paper detail

Flowering Locus T chimeric protein induces floral precocity in edible citrus

Judith P. Sinn, Jeremy B. Held, Chad Vosburg, Sara M. Klee, Vladimir Orbović, Earl L. Taylor, T. R. Gottwald, Ed Stover, Gloria A. Moore, Timothy W. McNellis

2020Plant Biotechnology Journal29 citationsDOIOpen Access PDF

Abstract

Citrus (family Rutaceae) is one of the most important fruit crops, with world production exceeding 150 million metric tons in 2018 and an international gross production value of 37.5 billion US dollars in 2016 (http://www.fao.org/faostat/en/#data). Citrus breeding programmes seek improvements in areas ranging from fruit quality (e.g. seedlessness, acid content) to resistance against emergent diseases (e.g. huanglongbing). Unfortunately, breeding efforts are hampered by a long juvenility period of 6 or more years, characterized by thorniness, lack of flowering and vertical as opposed to spreading growth form (Spiegel-Roy and Goldschmidt, 1996). Genetic transformation of citrus rootstock varieties has produced early flowering. ‘Carrizo’ citrange (Citrus sinensis × Poncirus trifoliata) constitutively expressing Arabidopsis thaliana floral identity genes LEAFY (LFY) or APETALA1 (AP1) flowered and set fruit 12–20 months after transfer to the greenhouse (Peña et al., 2001). In trifoliate orange (P. trifoliata), ectopic expression of Satsuma mandarin (C. unshiu) FLOWERING LOCUS T (CiFT) led to flowering as early as 12 weeks after transfer to the greenhouse (Endo et al., 2005). However, ectopic expression of flowering genes in citrus has been associated with aberrant phenotypes such as dwarfing, curled leaves and a weeping growth form (Endo et al., 2005; Peña et al., 2001). In contrast, expression of FT using a Citrus leaf blotch virus (CLBV) vector inoculated into the non-commercial C. excelsa and hybrids produced as part of a citrus breeding programme resulted in early flowering with minimal effects on growth form (Velázquez et al., 2016). Efforts to develop transgenic edible citrus constitutively expressing floral-inducing genes have encountered difficulty. Transgenic ‘Duncan’ grapefruit (Citrus × paradisi) and ‘Hamlin’ sweet orange (C. sinensis) constitutively expressing FT genes produced flower buds in tissue culture and failed to produce shoots or whole plants (Moore et al., 2016). The authors hypothesized that reducing FT activity in transgenic citrus using non-constitutive or inducible promoters might allow regeneration of edible citrus plants with a precocious blooming phenotype. Translational fusions of FT to GFP produced a milder early flowering phenotype in A. thaliana compared to unfused FT (Corbesier et al., 2007). This raises the possibility that chimeric FT proteins could be expressed constitutively in transgenic edible citrus to produce plants with a precocious blooming phenotype. Here, we expressed P. trifoliata FT1 (PtFT1) as a translational fusion with a single-chain variable fragment antibody (scFv; Pack and Plückthun, 1992) that is part of a separate study. A constitutive expression cassette was used that included the Cauliflower mosaic virus (CaMV) 35S promoter with a double enhancer region (CaMV 35Sp), the Tobacco etch virus 5′ untranslated region and the CaMV 35S polyadenylation signal (35St; Restrepo et al., 1990). The PtFT1-scFv-coding region included the PtFT1 cDNA sequence, a flexible linker sequence ([gly4ser]4), the scFv sequence and a C-terminal cMyc epitope tag (Figure 1a). Agrobacterium tumefaciens was used to transform ‘Duncan’ grapefruit with the PtFT1-scFv expression construct. Fifteen successful transformants were grafted onto ‘Carrizo’ citrange rootstocks and maintained in a growth chamber at The Pennsylvania State University (PSU; lines A–D, F–H) and a greenhouse at the United States Horticultural Research Laboratory (USHRL; lines I–P). Transgenic control line E, also maintained at PSU, was transformed with a construct lacking PtFT1-scFv. Full-length PtFT1-scFv protein was detected in growth chamber- and greenhouse-grown plants (Figure 1b). PtFT1-scFv levels varied among the lines, an observation consistent with previous FT overexpression studies in trees (Endo et al., 2005; Srinivasan et al., 2012). PtFT1-scFv protein was detected in both shoot tips and mature leaves, and was detectable over the course of the four-year experiment period (data not shown). Grapefruit trees transformed with PtFT1-scFv displayed varying reductions in juvenile characters that generally correlated with PtFT1-scFv protein level. For example, line L, which did not have detectable PtFT1-scFv protein (Figure 1b), largely resembled a juvenile tree, with an upright growth habit and long thorns in the leaf axils (Figure 1c,d). Moderate expressor lines, such as J (Figure 1b), retained an upright growth habit, but had reduced thorn size and occasional moderate leaf curling (Figure 1c). Severe growth phenotypes were observed in the highest expressing lines, H and O, and to a lesser extent in line G. These trees displayed a highly branched, prostrate, dwarfed growth form (Figure 1c) with greatly reduced thorniness or a complete absence of thorns (Figure 1d). Control transgenic line E displayed no reduction in thorniness or other juvenile characters (Figure 1d). Endo et al. (2005) similarly reported that higher expression of CiFT in transgenic P. trifoliata correlated with reduced prevalence of thorns and shorter tree stature. Precocious flowering was observed in PtFT1-scFv transgenic lines, with the earliest and most frequent flowering being observed in high-expressing lines H and O. Moderate- and high-expressing lines generally flowered within six months of transfer to soil, with H and O blooming within ~ 3 months. Transgenic lines at PSU were carefully monitored and fell into three blooming phenotype categories: strongly precocious, with large blooming flushes nearly continuously; mildly precocious, blooming 1–4 times a year with isolated flowers; and non-precocious, which did not bloom during the four-year monitoring period (Figure 1e). These phenotypic categories generally correlated with FT-scFv protein levels (Figure 1b) and agreed with previous studies showing that constitutive FT expression in trees can break down seasonality of flowering (Endo et al., 2005; Srinivasan et al., 2012). Transgenic line G had strongly reduced juvenility but did not flower continuously, possibly due to having somewhat lower FT-scFv levels compared to line H (Figure 1b). In lines H and O, flushes arose on multiple branches (Figure 1f), usually in leaf axils on leafy inflorescences. Leafless inflorescences were also observed on some shoots. Inflorescences displayed normal development patterns, with the terminal flower usually being the first to reach anthesis (Figure 1f; Spiegel-Roy and Goldschmidt, 1996). Flowers produced by PtFT1-scFv trees were fragrant and morphologically normal, with large petals flanking pollen–laden stamens and a central pistil (Figure 1g). Pollen grains were germinated in Brewbaker's medium (Figure 1h; Brewbaker and Kwack, 1963). Hand pollination of flowers with pollen from the same tree resulted in fruit production and seed set (Figure 1i). Fruit development was also observed on several lines without manual pollination, including lines H (Figure 1c) and O, within the first year of transfer to the growth chamber or greenhouse. These fruits ripened normally but were seedless (Figure 1j). Clonal propagation of low- and moderate-expressing lines via rooted cuttings was highly successful. While line H cuttings never produced roots, rooting was achieved for line O, albeit at a lower rate than moderate and low expressors. Relative PtFT1-scFv levels in clonal propagates were consistent with those in the original transformants (Figure 1b). Rooted cuttings of PtFT1-scFv transgenic lines maintained the blooming precocity phenotype (Figure 1k). Our results document the successful use of a chimeric FT protein to reduce flowering time in an edible citrus cultivar. In addition, we developed eight ‘Carrizo’ FT-scFv transformants, of which three had a precocious blooming phenotype (Figure 1l), and one transformant of a hybrid of ‘Jackson’ grapefruit, which had a precocious blooming phenotype (Figure 1m). While very high expression of PtFT1-scFv was associated with phenotypes that may alter agronomic fitness, moderate expression of the protein resulted in precocious blooming largely without negative effects. FT fusion proteins may have attenuated flowering promotion activity relative to native FT (Corbesier et al., 2007), possibly accounting for the success of our approach. The reduced juvenility offered by transgenic expression of chimeric FT proteins may provide an additional valuable tool for rapid-cycle citrus breeding (Moore et al., 2016). Continued studies are underway to evaluate how C. × paradisi PtFT1-scFv lines perform in a grove setting. This research was supported by Citrus Research and Development Foundation grants 11-125-424 and 18-016 (T.W.M.) and 573 (E.S. and G.A.M.), Citrus Research Board grant 5200-146 (G.A.M. and E.S.), a travel grant from the PSU Plant Pathology and Environmental Microbiology Department (S.M.K.), and United States Department of Agriculture National Institute of Food and Agriculture Hatch Appropriations under project no. PEN04649 and accession no. 1016093 (T.W.M.). We thank Nancy Wenner (PSU) for assistance with pollen photography, and Joanne Hodge and Spencer Marshall (USHRL) for technical assistance. The authors declare no conflicts of interest. T.W.M. designed the study. J.B.H., J.P.S., C.V. and T.W.M. wrote the paper with contributions from all authors. J.P.S., T.W.M., S.M.K., J.B.H., C.V., V.O., E.L.T. and E.S. performed experiments and analysed data. All authors participated in data interpretation for the manuscript.

Topics & Concepts

BiologyLocus (genetics)GeneBotanyFusion proteinGeneticsRecombinant DNAPlant Molecular Biology ResearchPlant Reproductive BiologyCocoa and Sweet Potato Agronomy