Twenty-five years of photobodies: formation, composition, and the 2-compartment logic of phytochrome B signaling

Abstract

How nuclear condensates encode cell-signaling dynamics remains unclear. Photobodies (PBs) in Arabidopsis offer a genetically tractable paradigm. PBs are light- and temperature-sensing nuclear condensates organized by the thermosensitive photoreceptor phytochrome B (phyB). Recent advances have clarified PB composition and illuminated PB formation and function. phyB is the dominant component and provides scaffold-like determinants that recruit selective signaling partners as primary clients (direct phyB binders) and secondary clients (recruited via primaries), spanning transcription/splicing regulators, E3-ligase components, kinases/phosphatases, and chaperones. These clients connect phyB condensates to diverse environmental and hormonal pathways, positioning PBs as a central hub for signaling integration. PB assembly is driven by condensation encoded in phyB's output module and modulated by its photosensory module, coupling assembly/dissolution to photostate and temperature. PBs nucleate nonrandomly at preferred seeding sites, producing spatially distinct classes with different occurrence frequencies and thermosensitivities. PB formation partitions signaling between PBs and the surrounding nucleoplasm, establishing a 2-compartment photosensory system. Within this architecture, dynamic sequestration in PBs tunes nucleoplasmic transcription factor stability and activity to expand signaling dynamic range and extends phyB control into the night by stabilizing active phyB. We propose that PBs function as an autoregulatory rheostat, dialing nucleoplasmic light sensitivity in proportion to incident irradiance and thereby enabling continuous discrimination of light-intensity changes across multiple orders of magnitude. We suggest that this 2-compartment logic illustrates a general role of membraneless organelles in signaling: using dense-phase dynamics to adjust pathway sensitivity and output in the surrounding dilute phase.

Introduction

How membraneless subnuclear organelles encode cell-signaling dynamics remains unclear. Eukaryotic nuclei are spatially organized across scales—from chromosome territories and active/repressive compartments to gene positioning at subnuclear landmarks (Lieberman-Aiden et al. 2009; Cremer and Cremer 2010; Egecioglu and Brickner 2011; Tiang et al. 2012; Feng et al. 2014; Shachar and Misteli 2017; Liu et al. 2017a). The interchromatin space of both plant and animal nuclei hosts a variety of functionally distinct membraneless condensates (“nuclear bodies”) that spatiotemporally organize and regulate nuclear functions (Shaw and Brown 2004; Dundr and Misteli 2010; Mao et al. 2011; Banani et al. 2017; Shin and Brangwynne 2017; Emenecker et al. 2021; Belmont 2022). It is commonly thought that nuclear bodies form through phase separation—a concentration-dependent, thermodynamic process driven by multivalent interactions among proteins and nucleic acids (Banani et al. 2017; Shin and Brangwynne 2017). Although nucleoli, Cajal bodies, nuclear speckles, paraspeckles, histone locus bodies, and promyelocytic leukemia protein (PML) bodies have defined roles in gene expression and genome maintenance (Dundr and Misteli 2010; Mao et al. 2011; Belmont 2022), genetically tractable systems that directly couple condensate behavior to signaling outputs remain limited.

Photobodies (PBs) provide such a system. PBs are plant nuclear bodies defined molecularly by the temperature-sensitive red (R) and far-red (FR) photoreceptor phytochrome B (phyB) (Chen and Chory 2011; Van Buskirk et al. 2012). The reference plant species Arabidopsis thaliana (Arabidopsis) encodes five phytochromes (phyA to phyE); phyB is the principal detector of light intensity and quality (Chen et al. 2004; Rockwell et al. 2006). phyB photoconverts between a R-light-absorbing inactive Pr and an FR-light-absorbing active Pfr (Fig. 1a) (Rockwell et al. 2006; Burgie and Vierstra 2014). Photoconversion rates (k₁, k₂) scale with R and FR intensities, and Pfr thermally reverts to Pr with a temperature-dependent rate (k_r), making phyB a thermosensor in addition to a photoreceptor (Fig. 1a) (Jung et al. 2016; Legris et al. 2016; Fan et al. 2025).

Figure 1.

phyB conformation state determines the steady-state PB pattern. a) phyB photoconverts reversibly between an inactive red-absorbing Pr and an active far-red-absorbing Pfr. Pfr also reverts to Pr via temperature-dependent thermal reversion. Rates k₁, k₂, and k_r denote Pr→Pfr, Pfr→Pr (photoconversion), and thermal reversion, respectively. b) Light-dependent subcellular localization of phyB-GFP (green). In darkness, Pr is predominantly cytoplasmic. R light drives nuclear accumulation and sets the steady-state subnuclear pattern as a function of intensity: under dim R (<0.5 μmol m⁻² s⁻¹), phyB is diffuse in the nucleoplasm; under low R (0.5 to 2 μmol m⁻² s⁻¹), it partitions into tens of small PBs (sPBs) plus nucleoplasm; under strong R (≥2 μmol m⁻² s⁻¹), it assembles into a few large PBs.

At the cellular level, photoactivation drives phyB to the nucleus and into discrete subnuclear foci (Kircher et al. 1999; Yamaguchi et al. 1999; Zhao et al. 2023). In 1999, Akira Nagatani, Ferenc Nagy, and Eberhard Schäfer showed that upon photoactivation, fluorescent protein-tagged phyB (phyB-FP) accumulates in nuclei and further partitions into subnuclear speckles (PBs) (Kircher et al. 1999; Yamaguchi et al. 1999). Joanne Chory and colleagues later demonstrated that the steady-state pattern of these phyB-containing condensates tracks the Pfr:Pr equilibrium: phyB-FP is diffuse in the nucleoplasm under dim light and forms condensates at ≥0.5 μmol m⁻² s⁻¹ R light (Fig. 1b) (Chen et al. 2003). Chory coined the term “photobodies” for these light-responsive condensates (Chen and Chory 2011). Two size classes of phyB-FP PBs are observed (Fig. 1b): Under low R light (0.5 to 2 μmol m⁻² s⁻¹) or FR-enriched shade, tens of small PBs (sPBs) (0.1 to 0.7 μm in diameter) appear alongside nucleoplasmic signal; under intense R light (eg ≥8 μmol m⁻² s⁻¹), 2 to 10 large PBs (0.7 to 2 μm in diameter) dominate (Chen et al. 2003, 2010; Van Buskirk et al. 2014; Hahm et al. 2020). For clarity, we refer to the large condensates as PBs and the small ones as sPBs. Because PB assembly and dissolution closely track phyB activity, PBs provide a tractable paradigm for how nuclear bodies implement signaling (Chen 2008; Chen and Chory 2011; Van Buskirk et al. 2012; Ronald and Davis 2019; Hahm et al. 2020; Pardi and Nusinow 2021; Willige et al. 2024; Chang et al. 2025).

PB formation segregates nuclear phyB into 2 compartments—the PBs and the surrounding nucleoplasm. Although active phyB correlates with PB condensation, the mechanisms that generate PBs and the functional logic of segregating phyB into 2 spatial pools remain incompletely defined. Here we synthesize recent advances into PB formation, composition, and function in Arabidopsis phyB signaling and outline open questions and experimental directions. Note that blue-light-induced cryptochrome 2 (CRY2) condensates are distinct from phyB-containing PBs and are not covered here (Más et al. 2000; Wang et al. 2021b; Kim et al. 2023).

Photobody formation

phyB is the dominant scaffold that encodes light and temperature information

phyB is a homodimeric, bilin-containing photoreceptor; each protomer comprises an N-terminal photosensory module (PSM) and a C-terminal output module (OPM) (Rockwell et al. 2006; Nagatani 2010; Burgie and Vierstra 2014) (Fig. 2a). The PSM includes N-terminal extension (NTE), nPAS (N-terminal period-ARNT-single-minded), the GAF (cGMP phosphodiesterase/adenylate cyclase/FhlA) domain that binds phytochromobilin, and the PHY (phytochrome-specific) domain. At the nPAS–GAF interface lies the light-sensing knot-lasso (KL), a figure-8 knot implicated in signaling (Wagner et al. 2005; Essen et al. 2008; Oka et al. 2008; Kikis et al. 2009; Yang et al. 2009; Burgie et al. 2014). The PHY domain contributes a “tongue” hairpin that contacts the GAF bilin pocket and converts from β-strand to helix during Pr-to-Pfr photoactivation (Burgie et al. 2014). Thus, the PSM senses light/temperature and mediates interactions with signaling partners (Fig. 2a). The OPM contains tandem PAS1/PAS2 and a histidine kinase-related domain (HKRD) (Rockwell et al. 2006; Nagatani 2010; Burgie and Vierstra 2014). HKRD is similar to bacterial histidine kinases, which comprise a dimerization and histidine phosphotransfer (DHp) domain and a catalytic/ATP-binding (CA) domain (Schneider-Poetsch et al. 1991; Yeh and Lagarias 1998; Stewart 2010). Because phyB lacks the conserved histidine in the DHp region (Schneider-Poetsch et al. 1991; Yeh and Lagarias 1998), we refer to this DHp-like segment as the “D domain.” The phyB CA domain contains all the conserved N, G1, F, and G2 subdomains for ATP binding (Schneider-Poetsch et al. 1991; Yeh and Lagarias 1998). The OPM is required for dimerization, nuclear localization, PB condensation, and binding to signaling partners (Fig. 2b). In particular, PAS1/PAS2 mediate nuclear localization, and the full OPM is required for PB formation (Fig. 2b) (Matsushita et al. 2003; Chen et al. 2005, 2022).

Figure 2.

Structure and functional regulation of phyB. a) Domain architecture of phyB. The N-terminal photosensory module (PSM) comprises the N-terminal extension (NTE), nPAS, GAF (binding phytochromobilin), and PHY domains. The C-terminal output module (OPM) contains PAS1/PAS2 and the histidine kinase-related domain (HKRD). Ancillary features include the knot-lasso (KL), the PHY “tongue,” and the modulator loop (Mod). The PSM senses light/temperature and mediates partner interactions, whereas the OPM is required for dimerization, nuclear localization, PB condensation, and partner binding. In Pr, extensive PSM–OPM contacts autoinhibit both modules; in Pfr, these contacts are disrupted, releasing localization and signaling functions. Sites of Y276H (YHB) and D1040V (phyB-18) mutations are indicated. b) Conformational models for the Pr head-to-tail dimer and the Pfr head-to-head dimer.

Cryo-electron microscopy (cryo-EM) reveals extensive conformational rearrangements between Pr and Pfr (Li et al. 2022a; Wang et al. 2024; Jia et al. 2025) (Fig. 2b). In Pr, phyB adopts a head-to-tail dimer in which dimerization is mediated by two types of PSM–OPM contacts: (i) interprotomer interactions between the nPAS–GAF region of one protomer and the PAS2 of the other and (ii) intraprotomer interactions between the PHY domain and the modulator loop—a β-hairpin motif between the PAS1 and PAS2—of the same protomer (Fig. 2) (Li et al. 2022a). Photoactivation shifts phyB to a head-to-head Pfr dimer in which both types of PSM–OPM contacts are disrupted, permitting head-to-head PSM alignment via the GAF domain (Fig. 2) (Wang et al. 2024; Jia et al. 2025). The cryo-EM structures corroborate biochemical and genetic evidence that photoactivation weakens PSM–OPM interactions (Chen et al. 2005; Jeong et al. 2025). Thus, extensive PSM–OPM contacts act as a self-inhibitory mechanism in Pr; their disruption in Pfr releases localization and signaling functions, rendering phyB outputs—such as nuclear accumulation, PB formation, and partner interactions—light sensitive (Fig. 2) (Chen et al. 2005; Van Buskirk et al. 2012; Li et al. 2022a; Wang et al. 2024; Jeong et al. 2025; Jia et al. 2025).

Active phyB provides scaffold-like determinants for PB assembly and, by mass, is the dominant PB component; its abundance dictates PB size. PBs are ∼0.1 to 2 μm spherical condensates, whose steady-state size tracks active phyB: intense R light yields a few PBs (∼0.7 to 2 μm), whereas reduced photoactivation under low R light or shade yields sPBs (∼0.1 to 0.7 μm) (Fig. 1b) (Gil et al. 2000; Chen et al. 2003, 2022; Trupkin et al. 2014; Van Buskirk et al. 2014; Hahm et al. 2020). Upon light exposure of dark-grown seedlings, photoactivated phyB-FP first accumulates in tens of sPBs and then into several PBs (Kircher et al. 1999, 2002; Yamaguchi et al. 1999; Bauer et al. 2004). Conversely, far-red exposure or darkness redistributes phyB-FP from PBs to sPBs and then to the nucleoplasm (Trupkin et al. 2014; Van Buskirk et al. 2014; Huang et al. 2016b; Murcia et al. 2021). Increasing temperature from 12 to 27 °C destabilizes Pfr and incrementally reduces PB number (Hahm et al. 2020; Du et al. 2024). In heterologous HEK293T cells, phyB-FP alone forms light-regulated condensates with similar morphology and dynamics (Chen et al. 2022). The constitutively active allele YHB (Y276H)—which locks phyB in an active photostate lacking both photoconversion and thermal reversion (Fischer et al. 2005; Su and Lagarias 2007; Burgie et al. 2014; Wang et al. 2024)—forms PBs but loses light and temperature regulation, reinforcing the link between active phyB and assembly (Su and Lagarias 2007; Hahm et al. 2020). Increasing phyB abundance enlarges PBs, consistent with PB size as a function of active-phyB concentration (Kim et al. 2024). Each phyB-FP PB is estimated to contain ∼1,500 phyB molecules, with total non-phyB components <20% of phyB (Kim et al. 2023; Kwon et al. 2024). Together, these observations support a phyB-dominant scaffold model for PB condensates.

OPM possesses condensation activity

OPM drives PB condensation. Expressing PSM and OPM separately in Arabidopsis shows that OPM alone forms nuclear condensates even in darkness, whereas PSM alone is cytoplasmic; when PSM is fused with an ectopic nuclear localization signal (NLS), it enters the nucleus but remains evenly dispersed (Matsushita et al. 2003; Chen et al. 2005, 2022). A similar behavior is observed in HEK293T cells—only OPM, not PSM, forms cytoplasmic condensates (Chen et al. 2022). Thus, OPM provides the intrinsic determinants for oligomerization and PB formation (Matsushita et al. 2003; Chen et al. 2005, 2022). Truncation analyses indicate that PAS1/PAS2 contributes to nuclear localization and that the full OPM is required for condensation activity (Fig. 2a) (Chen et al. 2005).

HKRD supplies dimerization essential for condensation. Although HKRD was not resolved by cryo-EM in either Pr or Pfr, biochemical and genetic data support a critical role for HKRD in phyB dimerization, function, and OPM-driven condensation. Dimerization of Pfr is primarily provided by OPM HKRD: full-length phyB is dimeric in isolation, whereas both a truncation without HKRD and PSM alone are monomeric (Burgie et al. 2014, 2017; Wang et al. 2024). Correspondingly, PSM signaling outputs—such as partner binding and repression of hypocotyl growth (Oka et al. 2008; Kikis et al. 2009)—require dimerization and are observed only when PSM is fused to a heterologous dimerization domain (eg β-glucuronidase, GUS) (Matsushita et al. 2003; Yoo et al. 2021). The importance of HKRD dimerization is underscored by the strong loss-of-function allele phyB-18 (D1040V) (Krall and Reed 2000): D1040V disrupts HKRD dimerization and eliminates OPM-dependent nuclear accumulation and PB condensation (Qiu et al. 2017). D1040 is conserved among plant, algal, and bacterial phytochromes and lies in the linker between the D and CA domains (Fig. 2a), indicating that intact HKRD architecture is essential for phyB dimerization and PB condensation (Qiu et al. 2017).

PSM regulates OPM condensation activity

Although OPM alone forms condensates constitutively—even in darkness—its condensation activity is regulated by light via the PSM conformational state. In the inactive Pr state, the PSM inhibits both OPM-driven nuclear accumulation and condensation, likely through direct PSM–OPM contacts, including interprotomer nPAS–GAF and PAS2 interactions and intraprotomer PHY-mod contacts (Fig. 2a) (Van Buskirk et al. 2012; Li et al. 2022a). Consistent with this, the N-terminal truncation that removes NTE–nPAS but retains GAF–PHY–OPM shows light-dependent nuclear accumulation yet fails to form PBs, indicating that the contacts between GAF–PHY and PAS1/PAS2 are sufficient to repress OPM-driven condensation in Pr (Chen et al. 2005).

Light regulation of OPM condensation requires the entire PSM. Removing NTE alone (nPAS–GAF–PHY–OPM) permits photoactivation-dependent nuclear entry but still fails to form PBs (Chen et al. 2005, 2022). NTE stabilizes Pfr; its deletion strongly accelerates Pfr→Pr thermal reversion (Cherry et al. 1992; Burgie et al. 2014, 2017; Klose et al. 2020). Thus, NTE promotes PB formation by stabilizing Pfr and sustaining OPM condensation. In support, phosphorylation within NTE (eg S86 and Y104) accelerates thermal reversion; perturbing NTE phosphorylation alters thermal reversion and PB formation/stability (Ryu et al. 2005; Medzihradszky et al. 2013; Nito et al. 2013; Viczián et al. 2020; Liu et al. 2023). Although NTE has been suggested to act as an intrinsically disordered region (IDR) that drives promiscuous interactions during PB phase separation (Chen et al. 2022), structural data indicate that while NTE is unstructured in Pr, it folds into 3 helices in Pfr that contact the GAF chromophore binding pocket and the PHY tongue (Wang et al. 2024). This supports NTE's primary role in photostate stabilization rather than in generic IDR-mediated condensation.

Accordingly, mutations that shift PSM photostate (YHB, Y276H) or impair OPM condensation activity (D1040V) alter phyB condensation behavior (Fig. 2a) (Su and Lagarias 2007; Van Buskirk et al. 2012; Qiu et al. 2017).

Photobody formation requires nucleation mechanisms beyond phase separation

Membraneless organelles often behave as liquid- or gel-like droplets and are commonly treated as biomolecular condensates arising via concentration-dependent, energetically favorable phase separation (Banani et al. 2017; Emenecker et al. 2021). While phase separation explains how condensate size scales with scaffold concentration, it does not, by itself, specify how condensate number is precisely regulated in living cells. In general, condensate formation initiates by nucleation (Bracha et al. 2019; Martin et al. 2021; Shimobayashi et al. 2021). In nonliving systems, nucleation is typically homogeneous, with seeds forming randomly once a critical concentration is reached (Mao et al. 2011; Bracha et al. 2019; Martin et al. 2021; Shimobayashi et al. 2021). If cellular condensates relied solely on concentration-driven processes, they would form randomly within a compartment and exhibit homogeneous properties. Thus, two key questions for living cells are whether nucleation is random or site-specific and whether condensates within the same compartment are homogeneous or heterogeneous.

PB formation has been proposed to be driven by phase separation of active phyB (Chen et al. 2022). Because phyB forms apparently random condensates in mammalian cytoplasm, it was suggested that PBs also form randomly in plant nuclei (Chen et al. 2022). Testing this hypothesis in plants requires stable subnuclear landmarks to assess spatial randomness. The nucleolus is a natural landmark: two PB types were defined by proximity—nucleolar-associated PBs (Nuo-PBs) at the nucleolar periphery and nonnucleolar-associated PBs (nonNuo-PBs) positioned away from it (Hahm et al. 2020). The numbers of Nuo- and nonNuo-PBs vary by cell/tissue type and across Arabidopsis ecotypes (Hahm et al. 2020). Intriguingly, temperature increases from 12 to 27 °C selectively reduce the number of nonNuo-PBs, indicating intranuclear heterogeneity—at least two PB populations differing in location and biophysical properties (Hahm et al. 2020).

Defined chromatin landmarks distinguish individual PB classes. In Arabidopsis, interphase chromosomes form rosette-like chromosome territories; centromeres and pericentromeric heterochromatin cluster as discrete chromocenters (CCs), with euchromatin loops emanating outward (Fransz et al. 2002). CCs are excellent landmarks because sister centromeres in endopolyploid interphase nuclei are largely aligned, so CC number does not scale with ploidy (Fransz et al. 2002; Schubert et al. 2006). Conventional FISH (fluorescence in situ hybridization) with centromeric 178 bp satellite (CEN178) or pericentromeric BAC probes labels all CCs, limiting chromosome resolution (Fransz et al. 2002; Schubert et al. 2006). Du et al. overcame this barrier by establishing Oligopaints in Arabidopsis, which uses uniquely designed chromosome-specific pericentromeric probes to distinguish individual CCs (Beliveau et al. 2017; Du et al. 2024). Combining PB immunostaining with Oligopaints (immunoFISH) revealed that ∼half of PBs closely associate with a single CC (CC-PBs), subclassified as CC1 to CC5 PBs; PBs not associated with CCs are nC-PBs (Fig. 3a) (Du et al. 2024).

Figure 3.

Photobodies nucleate nonrandomly at preferred seeding sites. a) Confocal immunoFISH of cotyledon pavement-cell nuclei showing CC1 to CC5 PBs and nC-PBs. Pericentromeres are labeled with chromosome-specific Oligopaints probes (PC1 to PC5); chromocenters (CCs) are labeled with CEN178 FISH. PBs are detected by anti-GFP immunostaining. Nuclear and nucleolar boundaries are outlined with dashed blue and yellow lines, respectively. Scale bar, 2 µm. b) Two-step model for PB formation. Nucleation occurs at specific seeding sites and determines PB number; subsequent growth determines PB size. Thus, while growth can be explained by phase separation, nucleation requires additional site-specific mechanisms beyond homogeneous phase separation.

Crucially, the occurrence frequencies of CC-PBs are far higher than expected by random association, and individual CC-PB classes differ markedly, with CC2 and CC4 PBs occurring most frequent (Du et al. 2024). These data argue that PBs nucleate at preferred seeding sites—likely at specific chromatin loci—with distinct seeding efficacies. Using the nucleolus and identified CCs as landmarks, 12 spatial PB classes were resolved in cotyledon pavement-cell nuclei: 6 nucleolar-associated (Nuo-CC1/2/3/4/5 and Nuo-nC) and 6 nonnucleolar (nonNuo-CC1/2/3/4/5 and nonNuo-nC). The frequencies of the Nuo versus nonNuo classes mirror CC-positioning: CC1/3/5 often lie at the nuclear periphery away from the nucleolus, whereas CC2/4 preferentially contact the nucleolus, consistent with the organization of 45S rDNA arrays on these two chromosomes (Fransz et al. 2002; Du et al. 2024).

Together, these findings support a two-step PB formation model: nucleation at preferred seeding sites followed by growth (Fig. 3b). Mathematic modeling likewise predicts a rapid nucleation step—either initial formation of phyB macroparticles or binding to a molecular structure (seeding site)—followed by slower body growth (Grima et al. 2018). Tethering phyB to the nucleolar periphery can nucleate de novo condensates, further supporting the nucleation model (Liu et al. 2014). Nucleation likely depends on phyB-seeding-site interactions that concentrate active phyB locally, enabling condensation below the global critical-concentration threshold—ie beyond classical homogeneous phase separation. Subsequent growth can proceed via phase-separation-driven processes. As such, the occurrence frequencies of individual PBs can vary by local seeding environments, so the PB number per nucleus reflects the integrated nucleation efficacies across all available sites.

Thermosensitive photobody nucleation

Moderate temperature increases (10 to 30 °C) accelerate phyB thermal reversion and destabilize Pfr, rendering phyB a thermosensor (Jung et al. 2016; Legris et al. 2016). Unlike shade or FR—which shifts the photoequilibrium strongly toward Pr and shrinks PB size (presumably by lowering Pfr concentration)—raising temperature from 12 to 27 °C under intense R light reduces PB number incrementally without substantially changing PB size (Hahm et al. 2020; Du et al. 2024). At 16 °C, 9 of 12 spatial PB classes (based on nucleolar and CC association) occur frequent (present in ≥20% of cotyledon pavement-cell nuclei): Nuo-CC2/3/4, Nuo-nC, nonNuo-CC1/2/3/5, and nonNuo-nC (Fig. 4) (Du et al. 2024). Among these, 6 classes are thermosensitive—Nuo-CC4, Nuo-nC, nonNuo-CC1/2/3/5—showing ≥30% declines at 27 °C relative to 16 °C, indicating temperature-responsive nucleation (Du et al. 2024). Consequently, the 4 nonNuo-CC PBs at the nuclear periphery become rare at elevated temperature, driving the overall decrease in PB number (Fig. 4). In contrast, 3 classes—Nuo-CC2, Nuo-CC3, and nonNuo-nC—are thermo-insensitive, with frequencies unchanged or only minimally affected at 27 °C (Fig. 4) (Du et al. 2024). Thus, spatiotemporal control of PB nucleation represents a critical mechanism in signaling; nucleation of thermosensitive PB classes may act as a cellular thermosensor that links modest temperature shifts to discrete changes in specific PBs and, consequently, to downstream signaling output (Du et al. 2024).

Figure 4.

Temperature control of PB number via nucleation. Model for thermal regulation of PB nucleation. At 16 °C, 9 major PB classes are detected. Among them, 6 classes (magenta) are thermosensitive—their nucleation frequencies decline as temperature rises—and 3 classes (blue) are thermo-insensitive. Increasing the temperature to 27 °C reduces the nucleation frequencies of thermosensitive PBs, causing loss of 4 nuclear peripheral nonNuo-CC PBs (nonNuo-CC1/2/3/5) and leaving 5 major PB classes at the elevated temperature. Circle size denotes occurrence frequency; circle position indicates nucleolar-associated (Nuo) versus nonnucleolar (nonNuo) locations.

Photobody composition

Defining PB composition and client taxonomy

A long-standing obstacle to comprehensively defining PB composition has been the physical isolation of PBs. Giltsu Choi and colleagues overcame this by purifying phyB-FP-tagged PBs from Arabidopsis leaves with fluorescence-activated particle sorting (FAPS) coupled to liquid chromatography-tandem mass spectrometry (LC-MS/MS), identifying 23 constituents (Kim et al. 2023; Kwon et al. 2024). phyB-FP was the most abundant, followed by PHOTOPERIODIC CONTROL OF HYPOCOTYL 1 (PCH1) and TANDEM ZINC-FINGER/PLUS3 (TZP); total non-phyB proteins amounted to <20% of phyB (Kim et al. 2023; Kwon et al. 2024). These data support a scaffold-client architecture, in which phyB serves as the dominant scaffold to recruit client components to PBs (Fig. 5a) (Kim et al. 2023; Kwon et al. 2024). Protoplast colocalization assays further separated clients into primary clients (direct phyB binders recruited by phyB) and secondary clients (lacking strong direct binding and recruited via primaries) (Kim et al. 2023; Kwon et al. 2024).

Figure 5.

Photobody composition. a) Scaffold-client model integrating constituents from Choi and Nusinow studies. phyB serves as the dominant scaffold, recruiting primary clients (direct phyB interactors; blue) and, via these, secondary clients (green). Seven major primary-client groups are shown: phyA/C/D/E, PIF1/3/4/5/7 and PIL1 (PIF2), COP1, SPA1/2/3, ELF3, PCH1/PCHL, and TZP. b) Functional classes of PB components. Putative PB components fall into 4 functional categories: (i) transcription/splicing, (ii) protein degradation (E3 machinery and regulators), (iii) phosphorylation/dephosphorylation (kinases/phosphatases), and (iv) chaperones. Components’ roles in nonphytochrome pathways are indicated in parentheses: nodal, nodal regulators in light/temperature/hormone networks; light, light signaling; clock, circadian clock; ABA, abscisic acid signaling; JA, jasmonic acid signaling; GA, gibberellic acid signaling; strigolactone/karrikin, strigolactone/karrikin signaling; salt, salt responses; chloroplast, chloroplast biogenesis; nodal/multi, involved in multiple pathways.

Earlier work from the Nusinow lab, aimed at identifying proteins associated with the Evening Complex—a transcriptional repressor complex consisting of EARLY FLOWERING 3 (ELF3), EARLY FLOWERING 4 (ELF4), and the DNA-binding protein LUX ARRHYTHMO (LUX)—identified PCH1 and ELF3 as PB-associated proteins (Nusinow et al. 2011; Huang et al. 2016a, 2016b). PCH1/PCH1-LIKE (PCHL) preferentially bind to active phyB, slow thermal reversion, and thereby promote PB stability (Huang et al. 2016b; Enderle et al. 2017; Huang et al. 2019). R light promotes ELF3 nuclear accumulation and condensate formation in a phyB-dependent manner (Ronald et al. 2022). The prion-like domain of ELF3 supports thermo-inducible condensation, reducing Evening Complex occupancy at target promoters (Raschke et al. 2015; Jung et al. 2020; Silva et al. 2020; Hutin et al. 2023). The FAPS-defined constituents overlap substantially with PCH1/ELF3 pulldown partners, supporting their association with many PB components (Huang et al. 2016a, 2016b; Kim et al. 2023).

Integrating the Choi and Nusinow datasets yields 7 major primary-client groups (Fig. 5a): (i) other phys, phyA/C/D/E; (ii) PCH1/PCHL; (iii) TZP; (iv) PHYTOCHROME-INTERACTING FACTORS (PIFs), PIF1/3/4/5/7, and PIL1 (PIF2); (v) CONSTITUTIVELY PHOTOMORPHOGENIC 1 (COP1); (vi) SUPPRESSOR OF PHYTOCHROME A-105 1/2 (SPA1/2); (vii) ELF3 (Huang et al. 2016a, 2016b; Kim et al. 2023; Kwon et al. 2024). Recently, Chan Yul Yoo and colleagues used miniTurbo proximity labeling to recover phyB-associated proteins in sPBs during early light exposure; all 7 groups were detected, suggesting these components also associate with sPBs (Olson et al. 2025).

Almost all primary clients—beyond PCH1/PCHL and ELF3—are also supported by in planta colocalization. All phytochromes form light-dependent condensates (Kircher et al. 2002). During dark-to-light transition, R light drives phyA/phyB/phyD into sPBs, which recruit PIF3 (Bauer et al. 2004; Al-Sady et al. 2006; Sheerin et al. 2015). phyA condensates can also be stimulated by FR independently of phyB, indicating phyA can act as a scaffold (Bauer et al. 2004; Sheerin et al. 2015; Feng et al. 2024). TZP, a transcriptional regulator linking light signaling and flowering, interacts with phyA and phyB, uses an N-terminal IDR to promote its phase separation, and localizes to PBs and phyA-containing sPBs (Kaiserli et al. 2015; Perrella et al. 2018; Fang et al. 2022; Feng et al. 2024, 2025; Perrella et al. 2025). COP1 and SPAs (SPA1/2/3/4) form the substrate-receptor module of cullin4-RING E3 ubiquitin ligases (CRL4^COP1/SPA) (Saijo et al. 2003; Zhu et al. 2008, 2015; Pham et al. 2018a; Ponnu and Hoecker 2021). SPA1 also functions as a Ser/Thr kinase (Paik et al. 2019; Lee et al. 2020; Wang et al. 2021a; Chang et al. 2024). SPA1/2 and COP1 colocalize with phyA/phyB in PBs (Zheng et al. 2013; Lu et al. 2015; Sheerin et al. 2015). SPA1 can also form condensates in darkness, although it is not required for phyA condensation into sPBs during dark-to-light transitions (Sheerin et al. 2015). COP1 forms condensates in darkness and under R light; FR/blue/white light redistributes COP1 to the nucleoplasm and cytoplasm via phyA and CRY1, whereas shade (reduced blue) promotes COP1 partitioning into PBs (Osterlund and Deng 1998; Pacín et al. 2013; Balcerowicz et al. 2017).

Detecting FP-tagged PIFs in PBs under light is challenging because many PIFs are short-lived and fluorescent proteins mature slowly, leading to turnover before fluorescence (Balleza et al. 2018; Kim et al. 2024). For mCherry-PIF5, PB colocalization with endogenous phyB is observed by immunostaining but not by live fluorescence imaging, owing to rapid turnover (Kim et al. 2024). This likely explains why PIF4-FP partitions into PBs in transient assays (Chen et al. 2022; Kim et al. 2023) yet has not been reported in stable lines. PIF3-FP fluorescence is detected in sPBs during dark-to-light transition—consistent with greater preillumination stability (Bauer et al. 2004; Al-Sady et al. 2006). In contrast, PIF7, a relatively stable PIF, shows PB localization with both antibodies and fluorescent tags (Willige et al. 2021; Xie et al. 2023). Moreover, PIF3-LIKE 1 (PIL1, also called PIF2)—stabilized by phyB in the light—colocalizes with phyB in PBs when coexpressed transiently (Luo et al. 2014; Kim et al. 2023).

Among secondary clients (Fig. 5a), PHOTOREGULATORY PROTEIN KINASES (PPK1 to PPK4; MUT9-Like Kinases/MLKs) interact with phyB and PIF3 to mediate phyB-dependent PIF3 phosphorylation and degradation (Ni et al. 2017; Kim et al. 2023) and also function as CRY2-associated kinases in blue light (Liu et al. 2017b). PPKs preferentially occupy CRY2 condensates; coexpression of PIF3 redistributes PPK1/2/3 (but not PPK4) to PBs, indicating PIF3 acts as a partition switch for PPK1/2/3 (Kim et al. 2023). TZP enhances PPK1–phyA association at phyA sPBs during dark-to-light transition or under FR, promoting phyA phosphorylation (Feng et al. 2024); by analogy, TZP may also help recruit PPKs to phyB PBs (Fig. 5a). PIL1 recruits the HLH protein LONG HYPCOTYL IN FAR-RED 1 (HFR1) (Luo et al. 2014). PCH1 recruits HSP70-1/HSP70-2 and the Groucho-type corepressors TOPLESS (TPL) and TPL-RELATED (TRP1 to TRP4) (Huang et al. 2016b; Plant et al. 2021; Kim et al. 2023). HSP70s have been proposed to stabilize the PB by limiting nonspecific aggregation (Kim et al. 2023) and also function broadly in hormone signaling, quality control, and immunity (Berka et al. 2022). TPL/TPRs act as transcription corepressors and contribute to transcription initiation (Plant et al. 2021; Leydon et al. 2025). ELF3 likely recruits the Evening Complex (ELF3–ELF4–LUX) (Huang et al. 2016a, 2016b), although consistent colocalization of ELF4–LUX within PBs remains to be established (Kim et al. 2023).

SUMO1 (small ubiquitin-like modifier 1) is present in PBs (Kim et al. 2023), consistent with SUMOylation of phyB and PIF3 (Sadanandom et al. 2015; Bernula et al. 2021; Péter et al. 2021). The R2R3-MYB transcription factor LONG AFTER FAR-RED LIGHT1 (LAF1), a phyA-pathway component, is also SUMOylated and requires SUMOylation for condensate formation (Ballesteros et al. 2001). Abolishing phyB SUMOylation at K996 enhances signaling without markedly changing PB patterns, implying that SUMOylation dampens phyB activity but is not essential for phyB condensation (Sadanandom et al. 2015).

Other putative PB clients

The FAPS and PCH1/ELF3 pulldown catalogs so far are almost certainly incomplete (Huang et al. 2016a, 2016b; Pardi and Nusinow 2021; Kim et al. 2023; Kwon et al. 2024). Several phyB interactors that colocalize with PBs were not recovered by these approaches, including transcription factors, B-BOX CONTAINING PROTEIN 4/CONSTANS-LIKE 3 (BBX4/COL3) (Datta et al. 2005; Heng et al. 2019), ETHYLENE-RESPONSIVE FACTORS 55 and 58 (ERF55/58) (Li et al. 2022b), AUXIN RESPONSE FACTORS 6 and 8 (ARF6/8) (Mao et al. 2020), and WRKY DNA-BINDING PROTEINS 18 and 40 (WRKY18/40) (Geilen and Böhmer 2015); splicing factors SPLICING FACTOR FOR PHYTOCHROME SIGNALING (SFPS) (Xin et al. 2017), SUPPRESSOR-OF-WHITE-APRICOT/SURP RNA-BINDING DOMAIN CONTAINING PROTEIN 1 (SWAP1) (Kathare et al. 2022), and SWELLMAP2 (SMP2) (Yan et al. 2022); the E3-ligase-associated SALT OVERLY SENSITIVE 2 (SOS2) (Ma et al. 2023); the Type 5 protein phosphatase PAPP5 (Ryu et al. 2005); and HSP70-HSP90 ORGANIZING PROTEIN 1 (HOP1) (Fig. 5b) (Olson et al. 2025). Such discrepancies likely reflect sampling time/conditions, abundance cutoffs, or transient/weak interactions.

Forward genetic screens for PB-deficient mutants identified HEMERA (HMR) (Chen et al. 2010), REGULATOR OF CHLOROPLAST BIOGENESIS (RCB) (Yoo et al. 2019), and NUCLEAR CONTROL OF PEP ACTIVITY (NCP) (Yang et al. 2019). All 3 are dual-targeted to the nucleus and plastids (Pfalz et al. 2006; Chen et al. 2010; Nevarez et al. 2017; Yang et al. 2019; Qiu et al. 2021). In nuclei, HMR acts as a phyB- and PIF-interacting transcription activator (Galvão et al. 2012; Qiu et al. 2015, 2019, 2021); in plastids, HMR/PAP5/pTAC12 is an essential subunit of the ∼1.1 MDa plastid-encoded RNA polymerase (PEP) complex (Ruedas et al. 2022; do Prado et al. 2024; Vergara-Cruces et al. 2024; Wu et al. 2024). RCB and NCP are thioredoxin-like proteins required for PB formation in the nucleus and PEP assembly in plastids (Yang et al. 2019; Qiu et al. 2021; Lee et al. 2024). PAP8/pTAC6, another dual nuclear/plastid PEP subunit, is also required for PB formation (Liebers et al. 2020). These findings reveal an intimate coordination between phyB signaling and plastid transcription (Yoo et al. 2020; Flynn et al. 2024). Notably, these factors were not detected in FAPS or PCH1/ELF3 pulldowns (Huang et al. 2016a, 2016b; Kim et al. 2023), suggesting they are low-abundant/loosely associated or act early in PB formation.

Client functions position PBs as a central hub for signal integration

Across studies, PB constituents fall into 4 functional classes (Fig. 5b): (i) transcription/splicing, gene-specific transcriptional factors (PIFs, the Evening Complex, TZP, BBX4, ARF6/8, ERF55/58, and WRKY18/40), general transcription repressing/initiation factors (TPL/TPR1/3/4), and splicing factors (SFPS/SMP2/SWAP1); (ii) protein degradation, E3-associated factors (COP1, SPA1/2, PCH1/PCHL); (iii) protein phosphorylation and dephosphorylation, kinases (PPK1 to PPK3, SPA1/2, SOS2) and the phosphatase PAPP5; and (iv) chaperones, HSP70-1/2 and HOP1. This composition supports a model in which PBs coordinate transcriptional control with regulated proteolysis and reversible phosphorylation, integrating these processes within the phyB signaling network.

PB clients also link PBs to diverse environmental and developmental pathways (Fig. 5b). Many are nodal regulators in light, temperature, and hormone networks (PIFs, COP1, SPAs) (Ponnu and Hoecker 2021; Delker et al. 2022; Cai and Huq 2025; Zhou and Deng 2025), while others connect PBs to blue-light signaling (TZP, PPKs) (Loudet et al. 2008; Liu et al. 2017b), salt tolerance (SOS2) (Ma et al. 2023), the circadian clock (ELF3, PCH1/PCHL) (Nusinow et al. 2011; Huang et al. 2016b), thermomorphogenesis (HMR, ELF3) (Quint et al. 2023; Fan et al. 2025), chloroplast biogenesis (HMR) (Chen et al. 2010), auxin (ARF6/8, TPL/TPR1/3/4) (Mao et al. 2020; Leydon et al. 2025), abscisic acid (ERF55/58, WRKY18/40, SFPS) (Geilen and Böhmer 2015; Punzo et al. 2020; Li et al. 2022b), gibberellin (ERF55/58) (Li et al. 2022b), jasmonic acid (ARF6/8) (Nagpal et al. 2005), strigolactone/karrikin (PAPP5) (Struk et al. 2021), and plant immunity (HSP70-1/2) (Berka et al. 2022; Pierik and Ballaré 2022). Collectively, these connections position PBs as a central hub for signaling integration.

Photobody function

phyB signaling framework

phyB orchestrates plant development from germination through flowering and senescence, including seedling morphogenesis, vegetative growth, chloroplast biogenesis/photosynthesis, and immunity (Legris et al. 2019; Yoo et al. 2020; Delker et al. 2022). PB functions are best resolved during Arabidopsis seedling establishment, where phyB senses R-light intensity across ∼5 orders of magnitude (≈0.001 to 100 µmol m⁻² s⁻¹) to drive quantitative outputs, most notably hypocotyl growth (Fig. 6) (Chen and Chory 2011). Below, we outline the core signaling mechanisms and then synthesize PB functions within that framework.

Figure 6.

phyB signaling framework. phyB continually senses R-light intensity across ∼5 orders of magnitude (≈0.001 to 100 µmol m⁻² s⁻¹) to produce quantitative outputs such as progressively slower hypocotyl elongation at higher light. phyB regulates growth genes primarily by controlling PIF transcription factors: (i) Directly, phyB binds PIFs to repress transactivation and promote degradation; (ii) indirectly, phyB inhibits CRL4^COP1/SPA E3 ligases, which target PIF antagonists (eg HY5, HFR1, PIL1, and BBX4) and PIF1/5 for degradation. A subset of phyB-pathway components (indicated as underlined in the schematic) partitions between PBs and the nucleoplasm.

phyB restrains hypocotyl elongation as light intensity increases by quantitatively suppressing the abundance and activity of PIF bHLH transcription factors, which generally antagonize phyB signaling (Legris et al. 2019; Cheng et al. 2021; Cai and Huq 2025). PIF1/3/4/5/7 collectively promote elongation by activating growth-relevant genes, including those involved in auxin biosynthesis and signaling (Leivar et al. 2008b; Shin et al. 2009; Li et al. 2012; Legris et al. 2019; Cheng et al. 2021; Cai and Huq 2025). phyB binds PIFs and inhibits them via two direct mechanisms: (i) accelerating degradation (PIF1/3/4/5) (Shen et al. 2005, 2007; Al-Sady et al. 2006; Lorrain et al. 2008; Qiu et al. 2017) and (ii) repressing transactivation and attenuating DNA binding (PIF1/3/4/5/7) (Oka et al. 2008; Kikis et al. 2009; Park et al. 2012, 2018; Van Buskirk et al. 2014; Yoo et al. 2021; Jeong et al. 2025) (Fig. 6). Despite overlapping functions, PIF accumulation patterns diverge: PIF1/3 are largely dark-/shade-enriched (Shen et al. 2005; Al-Sady et al. 2006; Lorrain et al. 2009; Chen et al. 2010; Soy et al. 2012, 2014; Qiu et al. 2017), whereas PIF4/5 persist at appreciable levels in prolonged light and PIF7 accumulates to similar levels in light and dark (Lorrain et al. 2008; Leivar et al. 2008a; Nieto et al. 2015; Qiu et al. 2017, 2019; Pham et al. 2018a; Fiorucci et al. 2020; Fan et al. 2025), indicative of distinct regulation mechanisms (Cheng et al. 2021; Cai and Huq 2025).

In parallel, phyB indirectly curbs PIF activity and stability by inhibiting COP1/SPA, the substrate receptors of CRL4^COP1/SPA E3 ubiquitin ligases (Fig. 6) (Saijo et al. 2003; Lu et al. 2015; Sheerin et al. 2015). CRL4^COP1/SPA enhances PIF output by degrading PIF antagonists, including ELONGATED HYPOCOTYL 5 (HY5), HFR1, PIL1, and BBX4—thereby both increasing PIF activity and promoting PIF accumulation (Fig. 6) (Ang et al. 1998; Kim et al. 2002; Saijo et al. 2003; Duek et al. 2004; Jang et al. 2005; Yang et al. 2005; Pham et al. 2018a, 2018b; Heng et al. 2019; Lau et al. 2019; Wang et al. 2021a). HY5, a bZIP transcription factor lacking an intrinsic activation domain (AD), gains transactivation via BBX20/21/22 and antagonizes PIFs by competing for ACGT-class enhancers (Oyama et al. 1997; Toledo-Ortiz et al. 2014; Gangappa and Kumar 2017; Burko et al. 2020; Bursch et al. 2020; Cañibano et al. 2021). HFR1, PIL1, BBX4 dampen PIF DNA binding (Duek et al. 2004; Shi et al. 2013; Luo et al. 2014; Pham et al. 2018b; Heng et al. 2019), and HFR1 also promotes degradation of PIF1/4/5 (Xu et al. 2017; Hwang et al. 2025). COP1 can additionally act with DET1 (CRL4^COP1/DET1) to degrade HY5 (Cañibano et al. 2021), but DET1 has not been detected in PBs and is omitted from Fig. 6. Adding complexity, CRL4^COP1/SPA also directly targets PIF1/5 for degradation (Zhu et al. 2015; Pham et al. 2018a).

Only a subset of the core pathway components partitions between PBs and the surrounding nucleoplasm—notably phyB, PIFs, COP1/SPA, HFR1, PIL1, and BBX4. This raises several PB-focused questions within the pathway: (i) How do PBs contribute to repression of PIF activity? (ii) How do PBs influence PIF degradation? (iii) How do PBs modulate COP1/SPA and the stability of PIF antagonists? (iv) How does PB formation improve light sensing across R-light intensities? The following sections integrate PB dynamics with established phyB signaling mechanisms to address these questions.

phyB inhibits PIF activity via nucleoplasmic AD repression and PB-based sequestration

Current evidence supports two separable mechanisms by which phyB inhibits PIF activity: (i) PB-independent repression of PIF transactivation activity in the nucleoplasm and (ii) PB-based sequestration.

phyB directly binds PIFs and suppresses their activity without requiring PBs. PIFs carry a C-terminal bHLH (dimerization/DNA binding to G-box/PBE) and an N-terminal activation domain (AD) with a conserved ΦxxΦΦ core (Φ, bulky hydrophobic residues; x, any residue) flanked by acidic residues, akin to p53 and Gcn4 (Fig. 7a) (Martínez-García et al. 2000; Zhang et al. 2013; Yoo et al. 2021). The AD is conserved across the family (evident in PIF1/4/5/7/8; Possart et al. 2017; Yoo et al. 2021) and lies adjacent to the conserved active-phyB binding (APB) motif that contacts the knot-lasso (KL) in the phyB PSM; PIF1/3 also harbor an active-phyA binding (APA) motif that binds the OPM of phyA/phyB (Fig. 7a) (Khanna et al. 2004; Oka et al. 2008; Kikis et al. 2009; Nagatani 2010; Jeong et al. 2025). APB-PSM binding alone is sufficient to repress AD activity and reduce DNA binding. This was demonstrated with NGB, in which phyB's OPM is replaced by GFP, a GUS dimerization module, and an SV40 NLS (Matsushita et al. 2003). NGB is nuclear and light-responsive but cannot form PBs (Matsushita et al. 2003). Nevertheless, NGB inhibits PIF3 AD activity and DNA binding in a light-dependent manner without triggering PIF3 degradation, yielding short hypocotyls even under dim light (Matsushita et al. 2003; Park et al. 2012, 2018; Van Buskirk et al. 2014; Qiu et al. 2017; Yoo et al. 2021). Thus, nucleoplasmic AD repression constitutes a robust, PB-independent branch of phyB signaling (Fig. 7b).

Figure 7.

phyB inhibits PIF activity via nucleoplasmic AD repression and PB sequestration. a) Domain structures of PIF3, phyB, and NGB. PIF3 carries a C-terminal bHLH (dimerization/DNA binding) and an N-terminal activation domain (AD), containing a ΦxxΦΦ core flanked by acidic residues. The AD lies between two phyB binding motifs: APB (binds the knot-lasso, KL, in the phyB PSM) and APA (binds PAS1/PAS2 within the OPM). APB-PSM binding alone is sufficient to repress AD activity. In NGB, the phyB OPM is replaced by GFP, a GUS dimerization module, and an SV40 NLS, eliminating PB condensation. NGB remains nuclear and light-responsive via the PSM but cannot form PBs. b) Two signaling architectures. Full-length phyB (2-compartment system): Light titrates PIF activity through (i) nucleoplasmic AD repression and (ii) PB-based sequestration. NGB (single-compartment system): Light titrates PIF activity only via AD repression, which saturates at low intensity and loses sensitivity to R-light-intensity changes above ∼0.1 µmol m⁻² s⁻¹. Thus, PB formation adds a sequestration mechanism that expands dynamic range, preserving discrimination at mid-to-high R light intensities.

phyB also sequesters PIFs into PBs, limiting their availability to act on nucleoplasmic targets (Fig. 7b). For PIF7, strong light partitions a substantial fraction to PBs with minimal promoter occupancy and correlates with hyperphosphorylation; shade dissolves PBs, rapidly yields hypophosphorylated PIF7, restores enhancer binding, promotes H2B eviction, and activates target transcription (Willige et al. 2021). Whether phosphorylation drives PB entry or results from it remains unresolved. PIF5 likewise supports the sequestration model: Overexpressed mCherry-PIF5 in a phyB-CFP background fails to promote hypocotyl elongation under high light when concentrated in PBs but regains activity under low light as PB partitioning decreases (Kim et al. 2024). Analogous PB-mediated control likely applies to additional PB clients, including ARF6/8, ERF55/88, and WRKY18/40, where phyB binding can dampen DNA binding and PB sequestration further limits activity (Geilen and Böhmer 2015; Mao et al. 2020; Li et al. 2022b), and to the Evening Complex (ELF3–ELF4–LUX), reducing its transcription repressor output in the nucleoplasm (Jung et al. 2020; Ronald et al. 2022; Hutin et al. 2023).

Because phyB-FP lines implement both mechanisms, whereas NGB implements only AD repression, comparing them isolates the PB contribution (Fig. 7b). NGB responses saturate below ∼0.1 µmol m⁻² s⁻¹ (Matsushita et al. 2003), indicating a narrow sensing range set by single-compartment AD repression. In contrast, full-length phyB maintains discernible responses across ∼4 orders of magnitude in R light (≈0.001 to 10 µmol m⁻² s⁻¹) (Matsushita et al. 2003), because PB formation adds a sequestration branch, creating a 2-compartment system (PBs and nucleoplasm) that reduces effective nucleoplasmic phyB to preserve responsiveness at mid/high light. Consistently, steady-state phyB-FP PBs appear only above ∼0.5 µmol m⁻² s⁻¹ (Chen et al. 2003), aligning sequestration onset with the light regime where added dynamic range is required. Thus, PB formation enables a 2-compartment logic that expands the signaling dynamic range.

PB sequestration of COP1/SPA controls nucleoplasmic transcription factor degradation

PBs have long been linked to protein-turnover steps in phyB signaling. For example, during the dark-to-light transition, FP-tagged PIF3 transiently associates with PBs before degradation, and PIF3 reaccumulates as PBs disappear in the light-to-dark transition (Bauer et al. 2004; Al-Sady et al. 2006; Van Buskirk et al. 2014). PB-deficient mutants—hmr, rcb, ncp, and pap8—fail to form large PBs and are blocked for light-induced PIF3 degradation (Chen et al. 2010; Galvão et al. 2012; Qiu et al. 2015, 2021; Yang et al. 2019; Liebers et al. 2020). Conversely, overexpressing the phyB OPM alone, which constitutively forms PBs, drives PIF3 degradation even in darkness (Qiu et al. 2017). A central challenge, however, is assigning functions to the dense (PB) versus dilute (nucleoplasmic) phases as components exchange between them; the co-loss of PBs and a readout is correlative and does not prove that the reaction occurs inside PBs (Kim et al. 2024). Recent composition data help disentangle roles: Proteasome subunits were not recovered from isolated PBs or by PCH1/ELF3 pulldown, and cullin E3 cores appear nucleoplasmic, whereas COP1 and SPAs—substrate receptors for CRL4^COP1/SPA—do localize to PBs (Chen et al. 2006; Huang et al. 2016a, 2016b; Kim et al. 2023). Thus, CRL assembly and proteasomal degradation likely occur outside PBs. Together, current evidence supports two PB-based mechanisms that regulate nucleoplasmic degradation (Fig. 8).

Figure 8.

phyB controls transcription factor abundance via two PB-based mechanisms. a) PB-proximal phosphorylation primes nucleoplasmic PIF3 ubiquitylation and degradation. Within PBs, phyB promotes PIF3 phosphorylation by PPK1/2/3 (and SOS2), which enhances PIF3 degradation in the nucleoplasm via CRL1^EBF and phyB-PIF3 co-degradation via CRL3^LRB. PP2A dephosphorylates PIF3 in the nucleoplasm. This phosphorylation/dephosphorylation cycle sets nucleoplasmic PIF3 abundance and transcriptional output. b) PB sequestration titrates nucleoplasmic CRL4^COP1/SPA activity and substrate stability. Sequestering COP1 and SPA1/2 in PBs reduces the effective nucleoplasmic pool of CRL4^COP1/SPA, which targets HY5 and PIF5 for degradation. HY5 does not partition into PBs and is therefore controlled solely by nucleoplasmic CUL4^COP1/SPA levels, whereas PIF5, which is also recruited to PBs, is stabilized by PB sequestration that limits its nucleoplasmic degradation.

PB-proximal phosphorylation primes PIF3 for nucleoplasmic ubiquitylation (Fig. 8a). phyB binding promotes PIF3 phosphorylation by PPK1/2/3 (and also SOS2; SPA1/2 can also act as kinases) (Ni et al. 2017; Paik et al. 2019; Lee et al. 2020; Wang et al. 2021a; Ma et al. 2023). PBs may elevate local concentrations of kinases and substrates, enhancing phosphorylation. Phosphorylated PIF3 is then ubiquitylated in the nucleoplasm by CRL1^EBF and CRL3^LRB, leading to proteasomal degradation; neither EBFs nor LRBs have been detected as PB constituents, consistent with nucleoplasmic processing (Ni et al. 2014; Dong et al. 2017). PP2A reverses PIF3 phosphorylation in the nucleoplasm, tuning both PIF3 abundance and transcription output (Cai et al. 2024).

PB sequestration titrates CRL4^COP1/SPA activity and substrate stability (Fig. 8b). phyB inhibits CRL4^COP1/SPA by both disrupting COP1/SPA complexes in the nucleoplasm and by sequestering COP1/SPA in PBs (particularly under R light), thereby lowering the effective nucleoplasmic pool of active E3 (Osterlund and Deng 1998; Lu et al. 2015; Sheerin et al. 2015; Balcerowicz et al. 2017). By analogy to phyB-mediated repression of PIF activity (Fig. 7b), this creates a 2-compartment control that can extend the dynamic range of phyB regulation of CRL4^COP1/SPA. Titration of the E3 dosage indirectly sets the stability of CRL4^COP1/SPA substrates, including HY5, PIF5, HFR1, PIL1, BBX4, and ELF3 (Saijo et al. 2003; Yu et al. 2008; Luo et al. 2014; Nieto et al. 2015; Pham et al. 2018a; Heng et al. 2019). HY5 does not partition into PBs; its abundance must track nucleoplasmic CRL4^COP1/SPA levels. In contrast, PIF5 is recruited to PBs; sequestration stabilizes PIF5 by limiting its exposure to nucleoplasmic CRL4^COP1/SPA (Kim et al. 2024). Thus, nucleoplasmic PIF5 is controlled by dual sequestration of COP1/SPA and PIF5 itself. Similar logic likely applies to other CRL4^COP1/SPA substrates that are also PB resident, such as ELF3, PIL1, HFR1, and BBX4.

Taken together, PBs can regulate nucleoplasmic degradation of transcription regulators by (i) priming substrates via PB-proximal phosphorylation and (ii) dosing E3 activity via COP1/SPA sequestration. These mechanisms couple PB dynamics to protein-turnover kinetics across the phyB network.

PBs integrate daylight and temperature information to control night processes

A core PB function is stabilizing active phyB. Protein–protein networks within PBs—between phyB and primary clients—likely attenuate thermal reversion, prolonging phyB activity after lights-off (Van Buskirk et al. 2014). During the light-to-dark transition, PB formation sustains phyB signaling into the night; this is especially pronounced in short days, where PBs act as a reservoir that suppresses PIF1/3/4/5 accumulation and activity to restrain hypocotyl growth early at night, allowing growth mainly before dawn (Fig. 9) (Nozue et al. 2007; Soy et al. 2012, 2014; Trupkin et al. 2014; Van Buskirk et al. 2014; Huang et al. 2016b). PB sequestration of ELF3 may shield it from COP1-dependent degradation in the nucleoplasm, supporting Evening Complex repression of PIF4 transcription (Yu et al. 2008; Nusinow et al. 2011; Anwer et al. 2014; Ronald et al. 2022). Warm nights accelerate thermal reversion, driving thermomorphogenetic hypocotyl growth (Jung et al. 2016; Legris et al. 2016). PCH1/PCHL accumulate at dusk in short days and further stabilize Pfr within PBs, integrating circadian and light pathways; accordingly, pch1 and pch1/pchl mutants display elongated hypocotyls specifically under short days, whereas PCH1 or PCHL1 overexpression reduces thermal reversion and slows PB dissolution in darkness (Huang et al. 2016b; Enderle et al. 2017; Huang et al. 2019). Thus, daytime light regime and nighttime temperature jointly shape nocturnal growth by modulating the stable Pfr pool—PBs thereby encode short-term memory of light and temperature information at night (Trupkin et al. 2014; Van Buskirk et al. 2014; Murcia et al. 2021).

Figure 9.

PBs extend phyB signaling into nighttime. Protein–protein interaction networks likely stabilize active phyB in PBs. In short days, PCH1/PCHL accumulate at dusk (ZT8) and attenuates thermal reversion, maintaining Pfr in PBs. PBs therefore act as a reservoir that sustains phyB signaling into the early night, suppressing PIF accumulation and activity to restrain hypocotyl growth. PBs can persist for ∼18 h; subsequent PB dissolution permits hypocotyl elongation before dawn. Hence, dynamic PB patterns provide short-term memory, conveying daytime light and nighttime temperature information into nocturnal growth.

An integrated model for phyB light-intensity sensing

How essential are PBs and their dynamics for light sensing? The primary function of phyB is to continuously detect changes in light intensity across 5 orders of magnitude. To enable this function, a key feature of the photosensory system is to adjust its light sensitivity with intensity, analogous to adjusting the camera's ISO—ie increasing the camera's ISO under dim light and reducing it under bright light. We propose that phyB employs two distinct mechanisms—nuclear import and PB-based sequestration—to auto-adjust the effective nucleoplasmic phyB level (ie ISO sensitivity) to discriminate light changes in dark-to-low and low-to-high light ranges, respectively (Fig. 10). From dark to dim R light (below the PB threshold), light drives nuclear accumulation of phyB, elevating effective nucleoplasmic sensitivity. Because active phyB, its signaling partners, and target genes all function within a single output compartment (the nucleoplasm), signaling responses scale with active phyB, and signaling saturates at dim light (∼0.1 to 0.5 µmol m⁻² s⁻¹). From low to high R light (above ∼0.5 µmol m⁻² s⁻¹), PB formation shifts phyB signaling into a PB-nucleoplasm 2-compartment signaling mode. Increasing intensity partitions more active phyB and selected clients into PBs, which titrates their nucleoplasmic availability and rebalances antagonistic transcriptional drives (eg COP1/SPA substrates vs. PIFs), thereby preventing early saturation and preserving responsiveness at mid- to high-light levels. Analogous to an antomatic camera, the overall PB compartment (its number/size) acts as an auto-rheostat: Small PBs at low light confer high nucleoplasmic sensitivity; larger PBs at higher light reduce nucleoplasmic sensitivity (Fig. 10). The result is a photosensory system that maintains discrimination across 5 orders of magnitude. We suggest that this PB-nucleoplasm 2-compartment signaling logic illustrates a general role of membraneless organelles in signaling: using dense-phase dynamics to adjust pathway sensitivity and output in the surrounding dilute phase.

Figure 10.

Model for phyB light-intensity sensing. phyB discriminates light intensity via 2 modes. (i) Single-compartment (nucleoplasm) mode, dark→dim R: Light drives nuclear import of phyB, increasing effective nucleoplasmic sensitivity; this mode saturates at ∼0.1 to 0.5 µmol m⁻² s⁻¹. (ii) Two-compartment (PB-nucleoplasm) mode, low→high R: Above ∼0.5 µmol m⁻² s⁻¹, light promotes PB formation; sequestration of active phyB to PBs lowers effective nucleoplasmic sensitivity, preventing early saturation and preserving responsiveness at higher intensities. In both modes, sensitivity auto-adjusts with incident light—analogous to camera ISO (low→high ISO from dark to dim R; high→low ISO from low to high R). In the second mode, PBs act as an auto-rheostat, tuning nucleoplasmic signaling sensitivity and output to irradiation.

Perspectives

Recent advances have clarified PB formation, composition, and function in light signaling, revealed a 2-compartment signaling architecture, and mapped PB clients from diverse environmental and hormonal pathways—positioning PBs as a central hub for signaling integration. These insights open several directions. First, the identities of seeding factors and whether PB nucleation requires dedicated PB components or non-PB chromatin landmarks remain unknown. It is also unclear whether sPBs nucleate at distinct sites and how cells govern the sPB-to-PB transition. Second, the PB growth mechanism also requires careful consideration. Widely used diagnostic properties to infer phase separation (sphericity, fusion, FRAP dynamics) do not distinguish stochastic phase separation from ordered assembly (McSwiggen et al. 2019); rigorous tests are needed to resolve the underlying modalities. Third, PB formation should be concentration dependent, yet most dynamics have been defined in phyB-FP overexpression lines. Endogenous phyB PBs likely differ in number, size, and kinetics and warrant careful analysis. Moreover, most studies focused on epidermal nuclei; how cell type, tissue context, and developmental state tune PB nucleation, growth, and client composition remains largely unexplored. Fourth, with clients from almost all environmental and hormonal pathways spanning transcription/splicing, proteolysis, phosphorylation/dephosphorylation, and chaperon functions, PBs likely mediate cross-pathway coupling. Defining client-specific partitioning rules and kinetics—and how light and temperature rewire them—will be essential to understand PB-based integration. Together, PBs have emerged as an experimental paradigm for how nuclear condensates encode signaling logic. Future work should resolve nucleation principles, assembly modalities, and client-specific kinetics and functions, advancing a general framework for condensate-enabled, 2-compartment control in cellular signaling.

Advances box.

• Definitive PB composition: Fluorescence-activated particle sorting of Arabidopsis photobodies, coupled with LC-MS/MS, defines PB constituents, supports a scaffold-client architecture, and illuminates functions.

• Integration hub: PB clients connect phyB condensates to diverse environmental and hormonal pathways, positioning PBs as a central hub for signaling integration.

• phyB as the dominant scaffold: phyB alone forms light-regulated condensates in mammalian cells, indicating intrinsic scaffold capacity for PB condensation.

• Structural mechanism for control: Cryo-EM structures of Pr and Pfr states reveal how the C-terminal output module's condensation activity is gated by the N-terminal photosensory module, enabling light and temperature regulation.

• Nonrandom nucleation: PBs nucleate at preferred seeding sites marked by nucleolar and chromocenter landmarks, yielding spatially distinct PB classes with different occurrence frequencies and thermosensitivities.

• Two-compartment signaling logic: PB formation enables a 2-compartment system that broadens signaling dynamic range and extends phyB control to night processes.

Author contributions

J.D., D.F., J.H., and M.C. designed the figures. M.C. wrote the manuscript with contributions from all authors.

Funding

This study was supported by the National Institutes of Health grants R01GM087388 and R01GM132765 and the National Science Foundation grant MCB-2141560 to M.C.

Data availability

Not applicable.