Functional Diversification of Populus FLOWERING LOCUS D-LIKE3 Transcription Factor and Two Paralogs in Shoot Ontogeny, Flowering, and Vegetative Phenology

Abstract

Both the evolution of tree taxa and whole-genome duplication (WGD) have occurred many times during angiosperm evolution. Transcription factors are preferentially retained following WGD suggesting that functional divergence of duplicates could contribute to traits distinctive to the tree growth habit. We used gain- and loss-of-function transgenics, photoperiod treatments, and circannual expression studies in adult trees to study the diversification of three Populus FLOWERING LOCUS D-LIKE (FDL) genes encoding bZIP transcription factors. Expression patterns and transgenic studies indicate that FDL2.2 promotes flowering and that FDL1 and FDL3 function in different vegetative phenophases. Study of dominant repressor FDL versions indicates that the FDL proteins are partially equivalent in their ability to alter shoot growth. Like its paralogs, FDL3 overexpression delays short day-induced growth cessation, but also induces distinct heterochronic shifts in shoot development-more rapid phytomer initiation and coordinated delay in both leaf expansion and the transition to secondary growth in long days, but not in short days. Our results indicate that both regulatory and protein coding sequence variation contributed to diversification of FDL paralogs that has led to a degree of specialization in multiple developmental processes important for trees and their local adaptation.

Keywords: FLOWERING LOCUS D; FRUITFULL; FT; gene duplication; heterochrony; leaf development; phenology; secondary growth.

Figure 1

FLOWERING LOCUS D-LIKE (FDL) genes differ in regulation. Relative expression is fold change in transcript levels of (A) FDL1, (B) FDL2.1, (C) FDL3, and (D) FDL2.2 relative to the time point with the lowest expression within a tissue (n = 3, except for shoot apex where the three apices were pooled to provide sufficient sample for analysis). FDL expression was normalized against reference gene 18S rRNA. (E) FDL1 and FDL3 and (F) FDL2.1 and FDL2.2 expression in shoot apices is presented separately to allow comparison of circannual patterns among the different FDLs. Axillary reproductive bud flush began in late February with anthesis reached in March (April sample is newly initiated floral bud). From September to March, preformed leaves and shoots were dissected from terminal buds.

Figure 2

Phenotypic effects of dominant repressor versions of FDL genes. (A) FDL1rd transgenic shoots on shoot elongation medium compared to an unrelated transgenic regenerated at the same time that displays typical shoot elongation. Each clump of shoots corresponds to a single explant that was induced to form callus and then shoots. Bottom photos show shoots from one of the explants in the top photos. (B) FDL3rd_56 transgenics showed reduced shoot growth and set terminal buds within 2 months after potting under LD conditions, whereas WT continued to grow. (C) FDL3rd_52 trees showed reduced shoot elongation. Representative 6-month-old trees are shown and values are means ± SE for two WT and two FDL3rd_52 trees after 3 months of growth in a greenhouse. (D) Representative 6-month-old WT and FDL2.2rd trees and mean heights ± SE after 4 months of growth in a greenhouse. For WT, n = 12; For FDL2.2rd, n = 16 (eight events with two ramets/event). ^*p < 0.01 compared to WT.

Figure 3

Overexpression of FDL3 affects leaf size and shoot elongation but does not promote flowering. Representative plants showing opposite effects on shoot elongation in vitro of (A) FDL3 dominant repression (FDL3rd) vs. (B) overexpression (FDL3ox). In (A) WT plant is 6-week-old, whereas FDL3rd plant is 10-week-old and in (B), WT and FDL3ox were propagated at the same time and are 4-week-old in the photo. Shoot apices of (C) FDL2.2ox with consecutive axillary inflorescences, (D) WT, and (E) FDL3ox. White arrows in (C) point to axillary inflorescences with multiple female flowers. (F) A premature flowering FDL2.2ox plant with many inflorescences as shown in (C) on both the main shoot and branches. (G) A FDL3ox plant with a few branches, but no flowers. (C–G) All photos are of 6-month-old plants grown at the same time in a greenhouse under a 16-h photoperiod.

Figure 4

Overexpression of FDL3 accelerates leaf production, but represses leaf growth in long day (LD) conditions. Ramets of two FDL3ox events and WT were grown in a growth chamber under 16-h photoperiods for 2 months. (A) Shoot apices and young rolled leaves of a FDL3ox plant compared with that of a WT plant. Scale bars = 1 cm. (B) Number of young rolled leaves. (C) Representative WT and FDL3ox trees (D) Emergence of new leaves (leaf lamina longer than 1 cm) over time. Leaf number was counted weekly, beginning 3 weeks after transplantation. (E) Progression of leaf length with position on the shoot. Leaf position 1 is the youngest leaf whose lamina is longer than 1 cm. (B,D,E) Means ± SE (n = 6) for two FDL3ox events (33 and 40) and WT.

Figure 5

Overexpression of FDL3 synchronously inhibits leaf expansion and the transition to secondary growth in long days (LDs), but secondary growth is restored in short days (SDs). (A–C) Both FDL3ox and WT plants were grown in a LD greenhouse for 6 months and subsequently transferred to a SD growth chamber for 8 weeks. Leaves were counted from top to bottom according to leaf plastochron index (LPI). Internode (IN) number refers to the internode beneath the corresponding LPI. All panels show from top to bottom LPI2, LPI4, LPI6, and LPI10 leaf or corresponding IN. (A) Extremely slow growth of FDL3ox leaves compared to WT in LDs. Scale bars = 2 cm. (B) Severely inhibited secondary growth in FDL3ox plants (B1–4) compared to progressive transition to secondary growth in WT (B5–8) in LDs. (C) Secondary growth in IN2, IN4, and IN6 formed after exposure to SDs (images above dotted line) in FDL3ox (C1–3) and WT (C5–7) plants. Note that FDL3ox INs 4 and 6 (C2,3) now resemble the same INs of WT plants grown in LDs (B6,7). In contrast, IN10 (C4) formed in LDs before SD treatment remained underdeveloped in FDL3ox. After exposure to SDs, WT plants ceased elongation growth, IN2 transitioned to secondary growth (C5) and substantial secondary xylem accumulated in IN4 and IN6 of WT (C6,7). Transverse sections were 60 μm thick, Scale bars = 100 μm. Vb, vascular bundles; Pf, phloem fiber; and Xy, xylem. (D,E) Comparative expression analysis of leaf and stem developmental marker genes in WT and two events of FDL3ox (33 and 40) grown in LDs. Relative expression in LPI2 and LPI6 leaves (D) and in internodes IN4 and IN8 (E). Expression was normalized against reference gene ubiquitin gene (UBQ2).

Figure 6

Overexpression of FDL3 delays growth cessation and bud set in short days. Plants of WT and two FDL3ox events (33 and 40) were grown in long days for 2 months before exposure to SDs. (A) Apical bud development of FDL3ox plants compared to WT after 5 and 10 weeks in SD (Week 5 and Week 10). In Week 5, WT plants had formed buds. In contrast, FDL3ox plants maintained actively growing apex. By Week 10, FDL3ox plants formed buds. (B,C) Cumulative stem growth (B) and leaf formation (C) were measured weekly during the first 5 weeks in SDs. (B) Plant height and (C) leaf numbers are means ± SE (n = 6).

Figure 7

FDL3ox trees resume leaf development in short days. (A,B) Both FDL3ox and WT plants were grown in a long day (LD) greenhouse for 6 months and then were transferred into a SD growth chamber for 8 weeks. FDL3ox leaves formed after transfer to SDs (above the red arrows) showed leaf development similar to actively growing WT plants, in contrast to underdeveloped leaves formed on FD3ox plants in LDs (below the red arrows). Plants (A) and shoots (B) were imaged after 8 weeks exposure to SDs. (C) Shoots from ramets of the same FDL3ox event grown 8 weeks in SDs or LDs. (D,E) The changes in leaf expansion size of FDL3ox plants followed the changes of photoperiod duration. (D) FDL3ox and WT plants were grown for 2 months in LDs (below the red arrows), followed by 4 weeks of SDs (between red arrows and yellow arrows), and then 3 weeks of LDs (above the yellow arrows). (E) Fully expanded leaf length of WT and FDL3ox plants formed in SDs and LDs. Six fully expanded leaves were measured for each plant. Leaf length is mean ± SE (n = 4); different letters indicated significant differences, p < 0.0001, Tukey–Kramer’s test.

Figure 8

FDL3ox and daylength alter expression of FT2 and three AP1/FUL homologs with diverse vegetative and reproductive expression patterns. (A–D) Fully expanded leaves were collected from WT and two independent events of FDL3ox plants grown for 2 months in LDs, followed by 3 weeks in SDs. Relative fold changes in transcript levels of FT2 (A), LAP1a (B), FUL (C), and LAP1b (D). Expression of the other two members of the AP1/FUL family (Supplementary Figure 8), MADS14 and MADS28, was not detectable in either LDs or SDs. The expression was normalized against reference gene UBQ2. (E) Seasonal expression pattern of FUL in adult Populus deltoides. Relative expression is fold change in transcript levels relative to the time point with the lowest expression within a tissue (n = 3 biological replicates except that three technical replicates were assayed from a pool of three shoot apices). (F) In situ hybridization showing LAP1b expression in initiating floral meristems of an immature male Populus trichocarpa inflorescence. FM, floral meristem; B, bract. Scale bar = 100 μm. Additional LAP1a and LAP1b in situ hybridizations are provided in Supplementary Figure S12.