PEBP Signaling Network in Tubers and Tuberous Root Crops

Tubers and tuberous root crops are essential carbohydrate sources and staple foods for humans, second only to cereals. The developmental phase transition, including floral initiation and underground storage organ formation, is controlled by complex signaling processes involving the integration of environmental and endogenous cues. FLOWERING LOCUS T (FT) and TERMINAL FLOWER 1/CENTRORADIALIS (TFL1/CEN), members of the phosphatidylethanolamine-binding protein (PEBP) gene family, play a central role in this developmental phase transition process. FT and FT-like proteins have a function to promote developmental phase transition, while TFL1/CEN act oppositely. The balance between FT and TFL1/CEN is critical to ensure a successful plant life cycle. Here, we present a summarized review of the role and signaling network of PEBP in floral initiation and underground storage organ formation, specifically in tubers and tuberous root crops. Lastly, we point out several questions that need to be answered in order to have a more complete understanding of the PEBP signaling network, which is crucial for the agronomical improvement of tubers and tuberous crops.


Introduction
Billions of people around the world consume tubers and tuberous root crops. Tuber crops, such as potato (Solanum tuberosum) and yam (Dioscorea spp.), and tuberous roots or starchy root crops, such as cassava (Manihot esculenta) and sweet potato (Ipomea batatas), are important carbohydrate sources and staple foods for humans, second only to cereals [1]. Tubers and tuberous crops give more energy per unit per day than cereals while also having a high amount of dietary fibers and serving as a good source of vitamins and minerals [1]. The global production of potato, yam, cassava, and sweet potato in 2020 was 359, 74, 302, and 89 megatons, respectively [2]. Potato is mainly cultivated in subtropical and temperate regions, such as China, India, Russia, Ukraine, and the USA. In contrast, yam, cassava, and sweet potato are mainly produced in tropical regions such as Southeast Asia, Central Africa, and Middle and South America [2]. Given their widespread cultivation and high nutritional values, tubers and tuberous root crops are valued to combat food insecurity, especially in developing countries [3,4].
Potato and yam tubers are initiated from specialized shoots called stolons. There are two main phases of tuber formation. The first phase is stolon initiation and growth, followed by the second phase of tuber induction and growth [5]. Potato tubers mostly grow from a stolon tip that forms scattered meristematic cells from a former vasculature tissue [6]. Meanwhile, tuberous roots mainly grow from the development of the primary cambium, followed by active cell division and cell expansion in the vascular cylinder [6]. The development of underground storage organ formation is initiated by a complex molecular biology process, which is controlled by environmental factors such as nutrient content, photoperiod, and temperature [7][8][9].
Underground storage organs such as tubers and tuberous roots contain high carbohydrate levels produced by photosynthesis. Tubers and tuberous roots are primary sink tissues to store excess assimilates. Sucrose, as a significant photo-assimilate, is transported to the underground storage organ by SUCROSE TRANSPORTERs (SUTs), such as the high-affinity sucrose/H + -symporter, StSUT1, and the low-affinity sucrose/H + -symporters, StSUT2 and StSUT4 [10][11][12]. Furthermore, the external application of sugar induces the formation of potato tubers and cassava tuberous roots [13,14], suggesting that the initiation of underground storage organ formation is well connected with plant growth and development.
The floral initiation and underground storage organ development are controlled by overlapping regulatory processes involving the Phosphatidylethanolamine-Binding Protein (PEBP) gene family [15]. In this review, we briefly summarize the function of PEBPs in floral initiation and underground storage organ formation of tuber and tuberous root crops, such as potato, yam, cassava, and sweet potato. We also describe the environmental and internal cues that modulate the function of PEBPs as regulators of flowering time and underground storage organ formation. Furthermore, the known upstream regulators and interacting partners of PEBPs in tubers and tuberous root crops are described and summarized into a PEBP signaling network.

PEBPs in Tuber and Tuberous Root Crops
PEBPs were first purified from the bovine brain and characterized as having affinities with several organic anions and phospholipids, e.g., phosphatidylethanolamine (PE) [16]. The overall protein structure of plant PEBPs shows a high degree of similarity with bovine PEBP. It comprises a large central antiparallel β-sheet flanked by a smaller β-sheet on one side and an α-helix on the other side [17,18]. These two central β-sheets are conserved and essential for the anion-binding pocket structure, the signature of PEBPs [17,18].
On the basis of the genome sequence, PEBPs can be classified into three main groups, MOTHER OF FT AND TFL1 (MFT)-like genes, FLOWERING LOCUS T (FT)-like genes, and TERMINAL FLOWER 1 (TFL1)/CENTRORADIALIS (CEN)-like genes [19]. Algae and nonvascular plants only contain MFT-like genes, whereas further duplication and divergence of FT/TFL1-like genes have occurred in vascular plants [19]. In angiosperms, the PEBP gene family plays an important role in plant development, including floral initiation and underground storage organ formation [20,21]. In general, the florigen FT-like protein acts as an activator to promote flowering and underground storage organ formation, while the anti-florigen, TFL1/CEN-like protein, has a function to repress flowering and underground storage organ formation [15,22]. In Arabidopsis, FT, TWIS SISTER OF FT (TSF), and MFT are known as floral activators, while BROTHER OF FT AND TFL1 (BFT), TFL1, and ARABIDOPSIS THALIANA CENTRORADIALIS (ATC) act as floral repressors [23]. Interestingly, the functional evolution of FT-like genes likely started in basal angiosperms, as the gymnosperm FT-like genes failed to promote flowering, in contrast with the gymnosperm TFL1-like genes, which were capable of repressing flowering when expressed in Arabidopsis (Arabidopsis thaliana) [24,25].
Several copies of FT-, TFL1/CEN-, BFT-, and MFT-like genes were identified in the selected tubers and tuberous root crops, including potato, yam, cassava, and sweet potato ( Figure 1A). On the basis of phylogenetic analysis, we found that potato has three FT-like genes, two TFL1/CEN-like genes, three BFT-like genes, and one MFT-like gene, consistent with a previous report [26]. We found that yam (Dioschorea rotundata) has eight FT-like genes, three TFL1/CEN-like genes, one BFT-like gene, and one MFT-like gene. Cassava has two FT-like genes, five TFL1/CEN-like genes, one BFT-like gene, and two MFT-like genes. Lastly, sweet potato has 10 FT-like genes, 11 TFL1/CEN-like genes, and seven MFT-like genes. Among the identified FT-like genes, until now, only PGSC0003DMG400016179 (StSP3D) and Manes_12G001600 (MeFT1) have been characterized as floral activators [21,27,28], while PGSC0003DMG400023365 (StSP6A) was found to be responsible for promoting potato tuberization [21]. Interestingly, StSP5G, the repressor of StSP6A [29], is seemingly not annotated by the Potato Genome Sequencing Consortium (PGSC). Furthermore, our phylogenetic analysis suggested that PGSC0003DMG400014322 (StCEN), which has been identified as a negative regulator of floral development and tuber formation [30,31], is more closely related to BFT than TFL1/CEN ( Figure 1A). seemingly not annotated by the Potato Genome Sequencing Consortium (PGSC). Furthermore, our phylogenetic analysis suggested that PGSC0003DMG400014322 (StCEN), which has been identified as a negative regulator of floral development and tuber formation [30,31], is more closely related to BFT than TFL1/CEN ( Figure 1A). The divergence of FT-like genes involves a positive selection of the fourth exon after the duplication event [33]. Among the thirty-nine nonconservative substitutions that distinguish FT and TFL1, four residues (E109, W138, Q140, and N152) in the divergent external loop segment in the fourth exon are critical for the opposite function of FT and TFL1 [18,34]. Several FT homologs undergo neofunctionalization after the duplication event to revert the function of FT from floral activator into floral repressor [35]. For example, an FT homolog from sugar beet (Beta vulgaris), BvFT1, acts as a floral repressor due to mutations that correspond to Y134N, G137Q, and W138Q in Arabidopsis FT [36], while barley (Hordeum vulgare), HvFT4, acts as a floral repressor likely due to W138A mutation [37]. Other residues such as V70, S76, and R83 are crucial for cell-to-cell movement of FT near the shoot apical meristem (SAM) region [38], while the D73, E84, Y85, P94, R119, and G171 residues are essential for the function of FT as the coactivator [20,34].
The majority of FT-like proteins from tubers and tuberous root crops have conserved critical residues in comparison with Arabidopsis FT ( Figure 1B). Interestingly, PGSC0003DMG400016180 (StSP5G-like) has two mutations corresponding to R83K and W138N in Arabidopsis FT which are crucial for FT movement [38] and neofunctionalization of FT as a floral repressor [34], respectively. A similar mutation, crucial for the neofunctionalization of FT as a floral repressor, is also seen in the yam. For instance, compared with Arabidopsis FT, DRNTG_29881 has a Y85H mutation, DRNTG_11086 has a W138M mutation, DRNTG_11087 has W138M and Q140H mutations, and DRNTG_20845 has a Q140P mutation. Furthermore, DRNTG_16474 has a V70I mutation, which is crucial for FT movement [38]. Therefore, only two FT-like proteins, DRNTG_28722 and DRNTG_10592, are likely responsible for promoting flowering and underground storage organ formation in yam. For tuberous root crops, two FT-like proteins in cassava and 10 FT-like proteins in sweet potato have similar critical residues to Arabidopsis FT. Nevertheless, functional studies of the remaining FT-, BFT-, and TFL1/CEN-like genes in potato, yam, cassava, and sweet potato are essential to understand the genetic redundancy and the signaling network of PEBPs in the regulation of flowering time and underground storage organ formation.

PEBP Signaling Network in Flowering Time Regulation
Internal and external cues strictly control the timing of flowering by regulating PEBP [39] (Figure 2A). Long-day (LD) plants, such as Arabidopsis, flower in response to increasing day length, whereas other plants, including rice and sorghum, flower under short-day (SD) conditions. Photoperiodic flowering is controlled by the integration of light perception and internal circadian rhythm [40]. In Arabidopsis, LD conditions induce high levels of FT and TSF expression, which eventually promote flowering [20]. Transcription factor complexes are also involved, such as the B-box transcription factor CONSTANS (CO) [41,42]. The rhythmic expression and protein stability of CO are controlled by the circadian clock and light, which eventually only allow FT and TSF expression at dusk under LD conditions [39,40]. In contrast, the expression of ATC is activated by SD conditions, especially at night time, likely to prevent flowering under noninductive SD conditions [43].
Temperatures also control the flowering time of Arabidopsis by flowering earlier under warm temperatures and later under cool temperatures [44] (Figure 2A). Warm temperatures promote the expression of FT and TSF through the activation of PHYTOCHROME INTER-ACTING FACTOR 4 (PIF4) [45]. Moreover, temperature regulates FT trafficking through the modulation of FT-phospholipid interactions [46]. Drought also accelerates flowering through the drought escape response by activating FT and TSF expression via phytohormone abscisic acid (ABA) [47,48]. Interestingly, early drought also activates BFT expression, which likely serves as a buffering mechanism to counteract FT and TSF activation [49]. The phytohormone gibberellic acid (GA) also promotes flowering through FT-dependent and -independent pathways [50]. Lastly, other internal cues, including sugars as photosynthesis products and age, regulate flowering through microRNA 156 and microRNA 172 via FT-dependent and -independent pathways [51] (Figure 2A). Temperatures also control the flowering time of Arabidopsis by flowering earlier under warm temperatures and later under cool temperatures [44] (Figure 2A). Warm temperatures promote the expression of FT and TSF through the activation of PHYTO-CHROME INTERACTING FACTOR 4 (PIF4) [45]. Moreover, temperature regulates FT trafficking through the modulation of FT-phospholipid interactions [46]. Drought also accelerates flowering through the drought escape response by activating FT and TSF expression via phytohormone abscisic acid (ABA) [47,48]. Interestingly, early drought also activates BFT expression, which likely serves as a buffering mechanism to counteract FT and TSF activation [49]. The phytohormone gibberellic acid (GA) also promotes flowering through FT-dependent and -independent pathways [50]. Lastly, other internal cues, including sugars as photosynthesis products and age, regulate flowering through mi-croRNA 156 and microRNA 172 via FT-dependent and -independent pathways [51] (Figure 2A).
FT is expressed in specific cells in the leaf vasculature called companion cells, in contrast with TFL1, which is expressed in the SAM [52,53] (Figure 2A). FT is then translocated to the SAM through interactions with several multiple C2 domain and transmembrane region proteins (MCTPs) transporters, including FT-INTERACTING PROTEIN1 (FTIP1/MCTP1), QUIRKY (QKY/MCTP15), and MCTP6 [54][55][56]. In the SAM, both FT and TFL1 compete to interact with chromatin-bound transcription factors FD to modulate the expression of floral identity genes, such as SUPPRESSOR OF OVEREXPRESSION OF CO 1 (SOC1), LEAFY (LFY), FRUITFULL (FUL), and APETALA1 (AP1) [57]. Furthermore, Arabidopsis BFT, TSF, and ATC were also reported to interact with FD [43,58,59], while the MFT ortholog from Adiantum capillus-veneris is also capable of interacting with FD [60]. The interaction of FT or TFL1 with FD is mediated by the 14-3-3 protein to form either a floral activator complex or a floral repressor complex [61]. Considering that all Arabidopsis PEBPs contain the 14-3-3-binding motif, this suggests a dynamic complex formation to fine-tune floral formation ( Figure 2B). Furthermore, FT and TFL1 also interact with different classes of TCP (for TEOSINTE BRANCHED1, CYCLOIDEA, and PCF) transcription factors to regulate flowering time [34,62]. FT is expressed in specific cells in the leaf vasculature called companion cells, in contrast with TFL1, which is expressed in the SAM [52,53] (Figure 2A). FT is then translocated to the SAM through interactions with several multiple C2 domain and transmembrane region proteins (MCTPs) transporters, including FT-INTERACTING PROTEIN1 (FTIP1/MCTP1), QUIRKY (QKY/MCTP15), and MCTP6 [54][55][56]. In the SAM, both FT and TFL1 compete to interact with chromatin-bound transcription factors FD to modulate the expression of floral identity genes, such as SUPPRESSOR OF OVEREXPRESSION OF CO 1 (SOC1), LEAFY (LFY), FRUITFULL (FUL), and APETALA1 (AP1) [57]. Furthermore, Arabidopsis BFT, TSF, and ATC were also reported to interact with FD [43,58,59], while the MFT ortholog from Adiantum capillus-veneris is also capable of interacting with FD [60]. The interaction of FT or TFL1 with FD is mediated by the 14-3-3 protein to form either a floral activator complex or a floral repressor complex [61]. Considering that all Arabidopsis PEBPs contain the 14-3-3-binding motif, this suggests a dynamic complex formation to fine-tune floral formation ( Figure 2B). Furthermore, FT and TFL1 also interact with different classes of TCP (for TEOSINTE BRANCHED1, CYCLOIDEA, and PCF) transcription factors to regulate flowering time [34,62]. Unlike in the model plant Arabidopsis, the molecular mechanism of flowering time regulation in tubers and tuberous root crops is still elusive [22]. Most tubers and tuberous root crops are mainly propagated through a vegetative method using clonal propagation from a tuber, stem, or vine [63][64][65][66]. Clonal propagation through the tuber or stem is the easiest way to multiply the plants as it does not require a complicated technique while maintaining the parental plants' characteristics. However, loss of genetic diversity, deleterious mutation, and pathogen accumulation are the major pitfalls of these clonal propagation methods [67]. Furthermore, the relatively long growth period of tubers and tuberous crops is a major obstacle in the breeding process of these plants [68][69][70]. Therefore, the complete understanding of molecular mechanism that control floral formation in tubers and tuberous root crops could aid in the breeding process of these crops.
Potato can be classified into two subspecies, the wild Andean varieties (S. tuberosum ssp. andigena), cultivated in the Andean highlands, and the modern potato (S. tuberosum ssp. tuberosum), originating from the lowlands of southern Chile [71]. The flowering time of both potato varieties is accelerated under LD conditions, although the plant still flowers under short day (SD) conditions [72,73]. Interestingly, unlike in Arabidopsis, StSP3D is expressed in the middle of the day instead of at dusk [74]. Furthermore, constitutive expression of Arabidopsis CO results in downregulation of StSP3D and late flowering in potato plants [21,75]. Potato StCO is encoded by three tandem genes, CONSTANS-like 1 (StCOL1), StCOL2, and StCOL3 [29]. The knockdown of StCOL1 using RNA interference accelerates potato flowering time [29], in contrast with the role of CO as a floral activator in Arabidopsis [42]. StCOL1 directly activates the expression of StSP5G through the TGTGGT motif in the StSP5G promoter [29]. StSP5G represses the expression of StSP3D and, thus, delays flowering [21,29] (Figure 3A). easiest way to multiply the plants as it does not require a complicated technique while maintaining the parental plants' characteristics. However, loss of genetic diversity, deleterious mutation, and pathogen accumulation are the major pitfalls of these clonal propagation methods [67]. Furthermore, the relatively long growth period of tubers and tuberous crops is a major obstacle in the breeding process of these plants [68][69][70]. Therefore, the complete understanding of molecular mechanism that control floral formation in tubers and tuberous root crops could aid in the breeding process of these crops.
Potato can be classified into two subspecies, the wild Andean varieties (S. tuberosum ssp. andigena), cultivated in the Andean highlands, and the modern potato (S. tuberosum ssp. tuberosum), originating from the lowlands of southern Chile [71]. The flowering time of both potato varieties is accelerated under LD conditions, although the plant still flowers under short day (SD) conditions [72,73]. Interestingly, unlike in Arabidopsis, StSP3D is expressed in the middle of the day instead of at dusk [74]. Furthermore, constitutive expression of Arabidopsis CO results in downregulation of StSP3D and late flowering in potato plants [21,75]. Potato StCO is encoded by three tandem genes, CONSTANS-like 1 (StCOL1), StCOL2, and StCOL3 [29]. The knockdown of StCOL1 using RNA interference accelerates potato flowering time [29], in contrast with the role of CO as a floral activator in Arabidopsis [42]. StCOL1 directly activates the expression of StSP5G through the TGTGGT motif in the StSP5G promoter [29]. StSP5G represses the expression of StSP3D and, thus, delays flowering [21,29] (Figure 3A).  The potato B-box 24 (stBBX24) acts as a negative regulator of the flowering time through the regulation of StSP3D ( Figure 3A), in contrast with the positive role of BBX24 in Arabidopsis flowering [74,76]. However, the overexpression of potato CYCLING DOF FACTOR 1 (stCDF1) results in a late flowering phenotype [77], consistent with the role of CDF1 as a floral repressor in Arabidopsis [78,79]. Interestingly, while Arabidopsis CDF1 prevents flowering through repressing CO and FT transcription [80,81], StCDF1 prevents  Figure 3A). StSP6A was found to prevent potato flowering, as StSP6A knockdown plants flowered early, while plants with high StSP6A levels flowered late [77]. The flowering time regulation by StSP6A seems unrelated to the regulation of StSP3D expression [21,77]. It is noteworthy to mention that StSP6A still maintains florigenic activity, as the overexpression of StSP6A, using constitutive 35S promoter, accelerated the flowering of Arabidopsis and potato [21]. Considering that both StSP6A and StSP3D are expressed in the leaves and stem, it is interesting to speculate whether StSP6A competes with StSP3D to interact with protein transporters such as the MCTPs [54] (Figure 3A).
LD conditions also promote the flowering of yam and cassava, while sweet potato flowering is accelerated under SD conditions [82][83][84][85]. The upregulation of MeFT1 and MeFT2 likely regulates the photoperiodic flowering of cassava under LD conditions [86] ( Figure 3B). However, the function of MeFT2 to promote flowering still needs to be discovered. Furthermore, we have yet to find any report describing the role of florigen or anti-florigen, as well as any molecular study of flowering time regulation in yam and sweet potato. Considering that the regulation of flowering time by photoperiod is likely caused by natural selection and domestication of crops [15,87], further research into this area may uncover novel insight into photoperiodic flowering in tubers and tuberous root crops.
In addition to the photoperiod, temperature is one of the seasonal cues that control the flowering time of many plant species [39]. A cold night temperature (5-6 • C) or GA treatment during SD conditions is reported to induce the flowering of LD potato species, including S. sparsipilum, S. acaule, S. punae, and S. demissum [72]. In Arabidopsis, the expression levels of the GA biosynthesis gene, AtGA20ox1, AtGA20ox2, and AtGA30x1 are induced under low temperatures [88]. Cold night temperature also reduces the expression of the GA catabolism gene, SlGA2ox, and promotes the accumulation of bioactive GA in tomato (S. lycopersicum) [89]. Furthermore, the flowering time of potato is also accelerated by high light, albeit in an StSP3D-independent manner [90]. In Arabidopsis, high light prevents the transcription of the floral repressor, FLOWERING LOCUS C (FLC), through nuclear accumulation of the PHD type transcription factor [91,92]. Considering that FLC also regulates temperature-responsive flowering of Arabidopsis and GA biosynthesis and signaling [93,94], it is tempting to speculate that temperature, high light, and GA also mutually regulate the flowering time of potato.
Cassava's flowering time is induced by LD conditions and cool temperature (early flowering at 22 • C and late flowering at 34 • C) [84,95] (Figure 3B). The temperatureresponsive flowering of cassava has the opposite pattern to several plant species, including Arabidopsis, which flower early in warm temperatures [44]. However, unlike in Arabidopsis, the expression levels of MeFT1 and MeFT2 are mostly unperturbed by temperature, while PIF4 expression increases in warm temperatures (similar to Arabidopsis) [84,86]. As shown in another study, the flowering time of field-grown cassava in Southeast Asia is promoted during the dry season in mountainous regions [96]. The drought and low night temperature of mountainous regions activate the expression of MeFT1 and MeFT2 [96] ( Figure 3B). The activation of florigen by drought is well documented in annual plants, including Arabidopsis and rice [47,48,97].
Overall, the mechanism underlying the long-distance transport and activation of the downstream targets of florigen in tubers and tuberous root crops is still elusive. However, a recent report suggested that StSP3D is able to form a floral activation complex with StABI5-like 1 (StABL1) to promote flowering [98]. However, whether StSP3D also makes a complex with StFDs such as StSP6A needs further verification. To initiate floral formation, the StSP3D/St14-3-3/StABL1 complex directly activates the expression of StSOC1 and StFUL to mediate meristem determinacy [98]. Furthermore, the anti-florigen StCEN is also capable of making a complex with StFD-like (StFDL1), probably to repress floral formation [31] (Figure 3A). Considering the sequence conservation of PEBPs, it is tempting to speculate that florigens and anti-florigens in other tubers and tuberous root crops may also make complexes with 14-3-3, ABI5-like, and FD-like proteins. Nevertheless, further analysis of upstream regulators of PEBPs, their transporters, and their downstream targets is essential to expand the PEBP signaling network in tubers and tuberous root crops.

PEBP Signaling Network in Underground Storage Organ Formation
Underground storage organ formation, including tubers and tuberous roots, is essential for plant survival under adverse environmental conditions. Day length regulates potato tuber formation in a genotype-dependent manner. For instance, the wild Andean varieties strictly require SD conditions to promote tuber formation, while the modern potato cultivar is more adapted to LD conditions [71]. SD conditions also promote the tuber formation of yams (D. rotundata, D. alata, and D. cayanensis), while LD conditions inhibit tuber formation and stimulate vine and leaf growth [99,100]. Interestingly, SD conditions also promote the tuberous root formation of cassava [101], suggesting an inverse correlation of photoperiod requirement for floral initiation and underground storage organ formation. Similar to flowering time, photoperiodic tuber formation is controlled by integrating light signaling and the circadian clock [71]. Light signals are perceived by multiple photoreceptors, including phytochromes, cryptochromes, phototropins, and F-box-containing flavin-binding proteins [102]. White-or red-light treatment of SD-grown plants in the middle of the night (known as night break) represses the tuber formation of Andigena potatos, while subsequent application of far-red light reverses the repression [103], suggesting the role of phytochrome in tuberization. Indeed, the knockdown of StPhyB induces tuber formation in noninductive conditions (LD and SD + night break) [104]. Furthermore, StPhyB interacts with StPhyF to repress tuber formation and flowering by stabilizing StCOL1 [29,105] (Figure 3A), in contrast to their function in Arabidopsis, where PhyB induces CO degradation [106]. Interestingly, Arabidopsis CONSTANS-like 7 is stabilized by PhyB [107], similar to StCOL1, suggesting that the function of PhyB depends on the CO-like protein sequence. StCOL1 then represses tuber formation by activating StSP5G transcription, a negative regulator of StSP6A [29]. Finally, StSP6A undergoes long-distance movement to initiate tuber formation [21]. Interestingly, the photoperiodic tuberization in water yam (D. alata) is also likely connected with FT-like genes, DaFT1 and DaFT2 [108].
The natural variation of StCDF1 confers the adaptation of European potato to tuberization under noninductive LD conditions [109]. StCDF1 represses the expression of StCOL1 and StCOL2 to activate tuber formation [29,109]. The early-maturing line, CE3130, has two truncated alleles, StCDF1.2 and StCDF1.3, which do not contain a carboxyl-terminal region that is important for interaction with FLAVIN-BINDING, KELCH REPEAT, F-BOX1 (FKF1) [109]. StFKF1 and GIGANTEA (StGI) promote the degradation of StCDF1 through direct interactions [109] (Figure 3A), suggesting the similarity of molecular components of the photoperiodic response between potato and Arabidopsis. Furthermore, the BELLRINGER-1-like transcription factor (StBEL5) and its interacting partner, the StKNOX transcription factor, have been identified as activators of StCDF1 and StSP6A [110,111].
The expression levels of StBEL5 are regulated by photoperiod through the PhyB/ microRNA miR172 cascade [112]. Similar to flowering time regulation in Arabidopsis, miR172 and miR156 have been shown to promote and inhibit tuber formation, respectively, partly through the regulation of StSP6A [112,113]. Both miR172 and miR156 are graft-transmissible and regulate phytohormone signaling, including GA, cytokinin, and strigolactone [112][113][114]. Considering that the sequential action of miR156 and miR172 define the age-dependent pathway of flowering time regulation in Arabidopsis [51] (Figure 2A), it is tempting to speculate whether a similar mode of action also regulates underground storage organ formation.
Temperature is another seasonal cue that controls underground storage organ formation. Warm temperatures inhibit potato tuber formation and sweet potato and Mexican turnip (Pachyrhizus tuberosus) tuberous root formation [112,115,116]. Warm temperature promotes the expression of a small RNA called Suppressing Expression of SP6A (SES), an upstream regulator of StSP6A [117]. SES reduces the accumulation of StSP6A transcripts to reduce the sink strength and delay tuber formation [117]. Interestingly, SES is likely a potato-specific small RNA, as no homolog of SES -pri-miRNA was found in the Solanaceae genome [117]. Furthermore, a circadian clock component, TIMING OF CAB EXPRESSION 1 (StTOC1), interacts with StSP6A to inhibit the autoactivation of StSP6A, especially in warm temperatures [118]. In Arabidopsis, the interaction between TOC1 and PIF4 mediates the circadian gating of thermomorphogenesis [119], suggesting that a similar mechanism may apply to underground storage organ formation. Further analysis of temperature-sensing components, including PhyB, EARLY FLOWERING 3, and florigen movement, could potentially elucidate the molecular mechanism of temperature-responsive underground storage organ formation [44,46,120].
A previous report suggested that the tuberigen signal, StSP6A, prevents floral formation in an StSP3D-independent manner [77]. However, the relationship between floral initiation and underground storage organ formation remains elusive. Interestingly, the removal of flowers increases the yield of potato and artichoke (Helianthus tuberosus) tubers [121][122][123], suggesting a competition between flower and tuber development for photosynthesis products. Furthermore, the overexpression of FT reduces the tuberous root growth and accelerated flowering of cassava, suggesting a relationship between both developmental processes [124] (Figure 3B). On a molecular level, the source-sink regulation of tuber formation is controlled by the interaction between StSP6A and sucrose efflux transporter StSWEET11 in the plasma membrane [125]. The binding of StSP6A to StSWEET11 prevents the leakage of sucrose to the apoplast, promoting symplastic sucrose transport and activating tuber formation [125]. The presence of StSP6A along the phloem also likely enables enhanced sucrose delivery toward tuber development [126] (Figure 3A). It will be interesting to see whether StSP3D also interacts with StSWEETs to compete in source-sink carbon partitioning for flower development. Furthermore, whether FT-like protein is also involved in sucrose export during tuberous root formation will be interesting to study ( Figure 3B). Similar to floral initiation, StSP6A forms a tuberigen activation complex by making a protein complex with St14-3-3 and StFD-like1 (StFDL1) to activate tuber formation [27]. Moreover, StCEN is also reported to interact with StFD and StFDL1a to prevent the activation of StFD and StFDL1 downstream targets, including StMADSs, StSP6A, and germin3 [31]. StSP6A and StCEN likely compete to form tuberigen activation or repression complex to control tuber formation [31] (Figure 3A), similar to the competition between FT and TFL1 during Arabidopsis floral formation [57] ( Figure 2B).
The phylogenomic approach suggests that StSP6A and its interacting partner, IDEN-TITY OF TUBER 1 (IT1), a plant-specific TCP transcription factor, are crucial for tuber formation in potato species [127]. Another TCP transcription factor, BRANCHED1b, interacts with StSP6A to prevent aerial tuber formation [128] ( Figure 3A). In Arabidopsis, BRANCHED1 interacts with FT but not TFL1 to prevent floral transition in the axillary meristem [129], suggesting a conserved regulatory network across plant species. StSP6A was also found to interact with Flowering-Promoting Factor 1, No Flowering in Short Days 1 (StNFL1), and StNFL2, which are likely crucial for tuber formation [130].
StSP6A also forms an alternative complex with St14-3-3 and StABL1 to promote tuber formation [98]. StABL1 binding regions are enriched with several binding motifs, including the bZIP/BHLH and TCP transcription factor-binding motifs, suggesting that the StSP6A/St14-3-3/StABL1 complex may form a larger protein complex with TCP transcription factors, such as IT1 [98,127]. StABL1 also directly activates the expression of StGA2ox1, a GA-catabolizing enzyme [98]. The fact that StABL1 is a component of ABA signaling probably explains the opposite role of GA and ABA as an inhibitor and an inducer of tuberization, respectively [98,131]. Interestingly, cassava tuberous root formation is inhibited by ABA [132], suggesting the opposite role of phytohormone signaling in (potato) tuber and (cassava) tuberous root development. Further functional studies of yam, cassava, and sweet potato PEBP are essential to expand our knowledge on the molecular mechanism of underground storage organ formation of these plants.

Conclusions and Future Perspectives
Floral initiation and underground storage organ formation are complex processes that integrate environmental cues and endogenous signals through the regulation of master regulators, i.e., the PEBP gene family [6,22]. FT and TFL1 act oppositely due to the external loop segment variation in the fourth exon [18]. Subtle amino-acid changes in the external loop segment during evolution and diversification cause neofunctionalization of PEBP, converting FT from a coactivator into a corepressor [15,35,36]. The PEBP signaling network of tuber and tuberous root crops is likely conserved, as upstream regulators and interacting partners of FT-and TFL1-like proteins are likely similar such as in Arabidopsis [6,15,22,39,71]. Several important issues, as listed below, could offer a future perspective to elucidate the PEBP signaling network in tubers and tuberous root crops.

1.
All Arabidopsis PEBPs are involved in flowering time regulation [15,23], while little is known in other plant species. Furthermore, whether PEBPs also regulate tuberous root formation remains elusive. Therefore, a functional study of the role of PEBP in flowering time and underground storage organ formation is essential to elucidate the PEBP signaling network in each tuber and tuberous root crop; 2.
Further in-depth studies of the molecular mechanism that controls the function of PEBP are needed. For example, the non-cell-autonomous function of FT requires an interaction with the MCTP transporter and phospholipids [24,46,55,56,133]. However, the mechanism underlying the long-distance transport of FT-like proteins in tubers and tuberous root crops is still elusive. Therefore, further verification of protein transporters and phospholipid interactors of the PEBP family from tubers and tuberous root crops is needed for a complete understanding of the PEBP signaling network in these plants. Furthermore, although StSP5G is a central regulator connecting the photoperiod with floral initiation and tuber formation by regulating StSP3D and StSP6A [21,29,109], it is not clear how StSP5G regulates its downstream targets. StSP5G may interact with other proteins, including St14-3-3s and StTCPs, to modulate floral initiation and tuber formation. Therefore, translating the knowledge from Arabidopsis could accelerate the characterization of the regulatory module of the PEBP-gene family in tubers and tuberous root crops ( Figure 2); 3.
Some plants develop underground storage organs to survive adverse environmental conditions [6]. Whole-genome analysis of the non-tuber-bearing Etuberosum, sister of the Petota (tuber-bearing) section of the Solanum genus, suggests that the nontuber-bearing phenotype of Etuberosum is caused by the deletion of the fourth exon of SP6A [127]. However, a more in-depth examination is needed to elucidate how PEBP, a flowering time regulator, evolved and diversified its function to regulate underground storage organ formation.