The role of carotenoids as a source of retrograde signals: impact on plant development and stress responses

Abstract

Communication from plastids to the nucleus via retrograde signal cascades is essential to modulate nuclear gene expression, impacting plant development and environmental responses. Recently, a new class of plastid retrograde signals has emerged, consisting of acyclic and cyclic carotenoids and/or their degradation products, apocarotenoids. Although the biochemical identity of many of the apocarotenoid signals is still under current investigation, the examples described herein demonstrate the central roles that these carotenoid-derived signals play in ensuring plant development and survival. We present recent advances in the discovery of apocarotenoid signals and their role in various plant developmental transitions and environmental stress responses. Moreover, we highlight the emerging data exposing the highly complex signal transduction pathways underlying plastid to nucleus apocarotenoid retrograde signaling cascades. Altogether, this review summarizes the central role of the carotenoid pathway as a major source of retrograde signals in plants.

Apocarotenoids, carotenoids, development, environmental responses, plastids, retrograde signals, signaling

Introduction

Plastids are a family of heterogeneous semi-autonomous organelles that carry out specialized functions essential for plants. These functions include, but are not limited to, the photosynthetic process (Archibald, 2015; López-Juez and Pyke, 2005) and synthesis of a variety of metabolic compounds (hormones, amino acids, vitamins, lipids, protective compounds, etc.) essential for plant growth, development, and environmental responses (Jarvis Lopez-Juez, 2013). In this regard, plastids can be described as metabolic hubs and environmental sensors critical for the plant’s development and survival (Neuhaus and Emes, 2000). The correct regulation of the operation and synthesis of the diverse metabolites produced within these organelles is essential to avoid overproduction and release of potentially toxic compounds, such as reactive oxygen species (ROS).

A small proportion of the proteins required for the plastid-localized pathways are encoded in the plastid genome, while the vast majority are nuclear-encoded and post-translationally imported. Consequently, the correct functioning of the different plastid types relies on a constant supply of nuclear-encoded proteins (Schwenkert et al., 2018; Chu et al., 2020). Thus, plastid homeostasis depends on strict coordination of the expression of the organellar and nuclear genomes accomplished through signaling from the nucleus to the plastid (referred to as anterograde communication). Plastids also signal their metabolic status and requirements back to the nucleus (referred to as retrograde communication) (Woodson and Chory, 2008).

A diverse set of retrograde signaling pathways have been documented in plants, performing crucial roles regulating plant development and acclimation to environmental fluctuations. Some of these signaling cascades, referred to as operational control pathways, operate in fully differentiated plastids in response to environmental fluctuations and adjust the organelle’s operation. In contrast, signals that are produced during the differentiation process of the plastids (referred to as biogenic) adjust the expression of nuclear genes required for the biogenesis of particular plastid types in coordination with plant development (Pogson et al., 2008; de Souza et al., 2017).

Retrograde signaling pathways comprise specific signaling molecules used to transduce information from plastids to the nucleus via signaling cascades, ultimately reprogramming nuclear gene expression, protein translation, and post-transcriptional regulatory responses (Tokumaru et al., 2017; Wu et al., 2019). Recently, progress has been made in the discovery of retrograde signaling molecules, yet the identity of most signals and the components of their corresponding signaling cascades remain to be defined.

Some of the best characterized retrograde signals are those derived from the tetrapyrrole biosynthetic pathway, including the chlorophyll precursor Mg-protoporphyrin IX and heme (Nott et al., 2006; Zhang et al., 2011; Terry and Smith, 2013). Many metabolic processes in plastids generate ROS, such as singlet oxygen (¹O₂) and hydrogen peroxide (H₂O₂), each of which plays diverse signaling roles in plant development and stress responses, either directly through diffusion of ROS from the plastid or indirectly via intermediate molecules (Pfannschmidt et al., 2003; Lee et al., 2007; Exposito-Rodriguez et al., 2017).

Additional metabolites such as 3´-phosphoadenosine-5´-phosphate dinucleotide (Estavillo et al., 2011), the methylerythritol 4-phosphate (MEP) pathway intermediate methyl erythritol cyclopyrophosphate (Xiao et al., 2012), and dihydroxyacetone phosphate (Vogel et al., 2014) act as retrograde signals in response to different stresses like wounding, drought, and high light. Further, retrograde information can also potentially be transmitted by plastid and nuclear dual-targeted proteins, which in response to certain signals retro-translocate from the plastid to the nucleus to regulate gene transcription. An increasing number of proteins have been shown to have dual plastid and nuclear localization (i.e. WHIRLY1, HEMERA, regulator of chloroplast biogenesis (RCB), and nuclear control of PEP activity (NCP)), and these proteins are required for proper chloroplast development and have been shown to mediate in the nuclear phytochrome-dependent regulation of diverse photosynthetic genes (Krupinska et al., 2014; Yang et al., 2019; Yoo et al., 2019). At least in the case of WHIRLY1 and HEMERA, translocation from the plastids to the nucleus has been demonstrated (Isemer et al., 2012; Nevarez et al., 2017). Analysis of the mechanisms underpinning this translocation and their role in retrograde signaling remains to be done.

Over the past decade apocarotenoids, breakdown derivatives of carotenoids, have emerged as an important class of retrograde signals. Initially, this was limited to β-cyclocitral, a non-enzymatic breakdown product of β-carotene by ROS that accumulates in response to high light stress (Ramel et al., 2012b). Interestingly, emerging data demonstrate that other carotenoids and/or apocarotenoids also elicit retrograde signals that regulate the expression of a specific set of nuclear genes and profoundly impact plant developmental programs and alternative stress responses (Avendaño-Vazquez et al., 2014; Van Norman et al., 2014; D’Alessandro et al., 2019; Jia et al., 2019; Cazzonelli et al., 2020). These findings support a central role of carotenoids as the source of retrograde signals in plants not only in the mature organelle, but also during the differentiation of particular plastid types. Excellent reviews on this topic have been published (Alagoz et al., 2018; Jiang and Dehesh, 2021; Moreno et al., 2021), but herein we take the opportunity to not only showcase recent advances, but also present the implications of the diverse carotenoid and carotenoid-derived signals for environmental responses and major developmental transitions (including meristem identity) that extend beyond photosynthesis-associated responses. Finally, we also aim to highlight the remaining areas requiring further exploration to better understand how these carotenoid-derived signals are transduced or translocated into the nucleus, ultimately leading to transcriptional reprogramming.

Carotenogenesis in plants

Carotenoids’ unique chemical nature, including enhanced susceptibility to oxidative modifications, makes them excellent targets to signal the functional and developmental status of the plastid to the nucleus, within the cell, throughout the entire plant, and between plants.

Carotenoids are isoprenoid-derived lipophilic compounds synthesized by diverse organisms (Nisar et al., 2015). In plants, they are essential photosynthetic pigments that quench excess light energy and scavenge ROS generated during the photosynthetic process (Sun et al., 2018). Carotenoids also play strategic roles as precursors to vital phytohormones (abscisic acid (ABA) and strigolactones), as well as signaling molecules, such as mycorradicin (Umehara et al., 2008; Walter et al., 2010; Finkelstein, 2013; Al-Babili and Bouwmeester, 2015).

In plants, carotenoids are synthesized within plastids and derived from the condensation of five-carbon isoprenoid building blocks, isopentenyl diphosphate and its isomer dimethylallyl diphosphate (Fig. 1), both synthesized through the MEP pathway (Bouvier et al., 2005; Hemmerlin et al., 2012). Subsequently, the condensation of one isopentenyl diphosphate with three dimethylallyl diphosphates produces the 20-carbon carotenoid precursor geranylgeranyl diphosphate (GGPP). Phytoene synthase (PSY) carries out the first committed and rate-limiting step of carotenogenesis through the condensation of two GGPPs producing 15-cis-phytoene (Moise et al., 2014). 15-cis-Phytoene is subsequently converted to all-trans-lycopene through a poly-cis-transformation by six consecutive reactions (Fig. 1) including desaturations catalysed by phytoene desaturase (PDS) and ζ-carotene desaturase (ZDS) and isomerizations by ζ-carotene isomerase (ZISO) and carotene cis–trans-isomerase (CRTISO) or non-enzymatically by photoisomerization to generate all-trans-lycopene (McQuinn et al., 2015).

Fig. 1.

Carotenoid biosynthetic pathway and derived signals. (A) The structures of the different acyclic and cyclic carotenoids and their names for each step. The isopentenyl diphosphate (IPP), dimethylallyl diphosphate (DMAPP), and geranylgeranyl diphosphate (GGPP) precursors, derived from the methyl-d-erythritol 4-phosphate (MEP) pathway, are indicated in the upper part of the figure. The enzymes catalysing each step in the pathway (bold) are as follows: geranylgeranyl diphosphate synthase (GGPPS), phytoene synthase (PSY), phytoene desaturase (PDS), ζ-carotene isomerase (ZISO), ζ-carotene desaturase (ZDS), carotene cis–trans isomerase (CRTISO), ε-cyclase (CRTL-E), carotenoid hydroxylases (CYP97A3 and CYP97C1), β-cyclases (CRTL-B1(2)), zeaxanthin epoxidase (ZEP), and neoxanthin synthase (NSY). In bacteria the entire desaturation steps are catalysed by CRTI. The chemical inhibitors norfluorazon (NF) and 2-(4-chlorophenylthio)-triethylamine hydrochloride, which target PDS3 and lycopene cyclases, respectively, are indicated. Carotenoids are cleaved by carotenoid cleavage dioxygenases (CCDs) and by lipoxygenase 2 (AtLOX2 of Arabidopsis or TomLOXC of tomato). The carotenoid substrates for different CCDs and AtLOX2/TomLOXC are indicated by the corresponding colored stars. (B–G) Phenotypic defects associated with the accumulation or deficiency (indicated with an arrow) of acyclic carotenoids as a source of retrograde signals. (B) The pds3 mutant shows accumulation of phytoene and reprograming of nuclear-encoded genes specific for this mutant, leaving open that phytoene could be a source of a retrograde signal (apocarotenoid signal 5 (ACS5)). (C) Accumulation of phytoene by overexpression of the bacterial PSY enzyme (crtB) results in the redifferentiation of chloroplasts to chromoplasts. (D) Accumulation of phytofluene and ζ-carotene in the clb5 mutant results in albino plants with finger-like leaf morphology and a reprogramming of the shoot apical meristem (SAM) to a flower meristem (FM). These defects are partially reverted in the double clb5 ccd4 mutant, demonstrating the apocarotenoid nature of the signal (apocarotenoid signal 1 (ACS1)). (E) The Arabidopsis CRTISO mutant (ccr2) displays a virescent phenotype and defects in etioplast development with impaired prolamellar body (PLB) formation. The recovery of greening and PLB formation in the double mutant ccr2 ziso155 (lacking ZISO and CRTISO) and with CCD inhibitors demonstrate that an apocarotenoid derived from pro-neurosporene and/or di-cis-ζ-carotene is implicated in this regulation. (F) In tomato, the epistatic relationship of the CRTISO mutant tangerine (r2997) over the PSY mutant yellow flesh (t3002) observed in the double r2997 t3002 mutant uncovers the role of prolycopene and/or tri-cis-neurosporene as a source of an apocarotenoid signal that regulates the transcription of PSY. (G) Signaling molecules are also derived from cyclic carotenoids including β-cyclocitral and β-cyclocitric acid derived from β-carotene enhanced during high light, drought stress, and herbivore attack. Also, retinal synthesized from β-carotene and probably corresponding to the previous signal referred to as apocarotenoid signal 3 (ACS3) regulates the correct establishment of lateral roots. Other signaling molecules derived from cyclic carotenoids including zaxinone and anchorene have been documented.

At all-trans-lycopene the carotenoid pathway bifurcates into the α- and β-branches, ultimately generating the photoprotective carotenes (e.g. β-carotene) and xanthophylls (Fig. 1), from which the phytohormones strigolactone and ABA, respectively, are also derived. The α-branch includes α-carotene (ε- and β-rings) and lutein (Kim et al., 2009). The β-branch includes β-carotene (two β-rings), zeaxanthin, antheraxanthin, violaxanthin, and neoxanthin (Cazzonelli and Pogson, 2010).

Interestingly, in bacteria the six reactions that convert phytoene to all-trans-lycopene are carried out by a single enzyme, CRTI (Fig. 1; Sandmann, 2009). It is believed plants evolved and maintained a more complex poly-cis-transformation pathway to link carotenogenesis with photosynthetic electron transport via the plastoquinone pool (Carol and Kuntz, 2001). Recent research suggests that the increase in the number of enzymes required for the synthesis of all-trans-lycopene provides additional control points that can be modulated in a spatio-temporal manner, an aspect that is further supported by available transcriptomic data (see Box 1). This ultimately enables the accumulation of specific carotenoid intermediates, from which apocarotenoid signals are derived that subsequently reprogram gene expression, thereby regulating various aspects of plant development (Avendaño-Vazquez et al., 2014; Cazzonelli et al., 2020; Escobar-Tovar et al., 2021).

Box 1. Multiple approaches shed light on mechanisms of novel signaling molecules

The identification of carotenoid-derived retrograde signaling has substantially advanced in past years, but the molecular characterization of most of the signaling molecules remains a challenge due to their low abundance and the complexity of their synthesis. Identifying the sites of synthesis of some of these signaling molecules might shed light on the strategies required for causal apocarotenoid identification and improved understanding of their physiological roles. The analysis of detailed transcriptional profiles for genes throughout the carotenoid pathway, as well as protein levels, in the cell and during different developmental stages, might expose the site of synthesis of specific carotenoid-derived signals and their potential functions. An example that illustrates this aspect is the tissue- and cell-specific expression pattern of the carotenogenic genes reported in roots, which correlates with the role of the retinal and ACS3 apocarotenoids in lateral root specification (Van Norman et al., 2014). The carotenoid biosynthetic genes in the shoot apical meristem domains, based on existing transcriptomic data (Tian et al., 2019), also show differential expression. These data support the potential accumulation of particular carotenoids in specific regions of the shoot.

(A) Heat map illustrating the expression levels in the epidermal layer (1), central zone (2), enlarged peripheral zone (3), and rib meristem (4) from the SAM and leaf abaxial (5) and adaxial (6) domains of the carotenogenic genes. PSY, phytoene synthase; PDS3, phytoene desaturase; ZIC1, ζ-carotene isomerase; ZDS, ζ-carotene desaturase; CRISTO, cis–trans isomerase; LYC, lycopene cyclase β, CRTL-E; LUT2, lycopene cyclase εCRTL-B1(2); BCH1, β-carotene hydroxylase 1; BCH2, β-carotene hydroxylase 2; LUT5, carotenoid hydroxylases CYP97A3; CYP97B3, carotenoid hydroxylases CYP97B3; LUT1, CYP97C. The leaf primordia display moderate to high expression levels compared with the SAM. (B) Diagrammatic cross-sections exemplifying the different expression levels of PSY, PDS3, ZDS, and LUT5 genes in SAM and leaf primordia (LP). The color code in the diagram correspond to the Z-score shown for the heat map. The data are from Tian et al. (2019) and the heat map was elaborated with heatmapper (Babicki et al., 2016).

Carotenoid modifications

Carotenoids are subject to oxidations and subsequent cleavage that can potentially occur non-enzymatically or enzymatically at each of the double bonds present in these molecules, resulting in a diverse array of aldehyde and ketone compounds known as apocarotenoids (Ramel et al., 2012a; Hou et al., 2016). Non-enzymatic cleavage occurs via ROS, which accumulate in response to environmental stressors (e.g. light, temperature, and drought), randomly at any C=C double bond on the carotenoid backbone thereby producing a multitude of apocarotenoids (Ramel et al., 2012b; Hou et al., 2016). However, enzymatic oxidation and cleavage also occurs via a family of non-heme iron-dependent carotenoid cleavage dioxygenases (CCDs), grouped into two independent subfamilies (CCDs and 9-cis-epoxycarotenoid dioxygenases (NCEDs)) that in Arabidopsis include nine members. CCDs vary in carotenoid substrate specificity and the specific C=C double bond they cleave, determined by the enzyme’s active site size and shape (Auldridge et al., 2006). The NCED subfamily (five members in Arabidopsis) participate in the cleavage of 9-cis-isomers of epoxycarotenoids (i.e. violaxanthin and neoxanthin) primarily involved in the biosynthesis of ABA (Tan et al., 2003). In contrast, the CCD subfamily is composed of four members in Arabidopsis (CCD1, CCD4, CCD7, and CCD8) displaying a broader cleavage specificity and localization (Hou et al., 2016).

CCD1, present in diverse species including Arabidopsis (CCD1) and tomato (CCD1A and CCD1B), lacks a chloroplast targeting peptide yet is still associated with the chloroplast membrane (Simkin et al., 2004). CCD1 cleaves at 9,10 and 9ʹ,10ʹ positions of acyclic and cyclic carotenoids (Fig. 1), generating volatile apocarotenoids providing flavors and aromas in fruits and flowers, and non-volatile apocarotenoids such as mycorradicin in roots (Fester et al., 2002; Simkin et al., 2004; Ibdah et al., 2006; Ilg et al., 2014).

Of the plastid-localized CCD members, CCD7 and CCD8 are involved in the synthesis of strigolactones through the cleavage of 9-cis-β-carotene and 9-cis-β-apo-10ʹ-carotenal, respectively (Umehara et al., 2008; Al-Babili and Bouwmeester, 2015; Bruno et al., 2016). In addition to strigolactone biosynthesis, CCD7 presents a broad substrate specificity cleaving phytoene and ζ-carotene in Escherichia coli (Schwartz et al., 2004). Also, CCD8 cleaves other long-chain non-volatile apocarotenoids including all-trans-β-apo-10ʹ-carotenal producing different apocarotenoids (Alder et al., 2012).

CCD4 is the largest subclass of the plant CCD family, only present in flowering plants. It displays broad specificity cleaving different substrates including all-trans-β-carotene, β-apo-10ʹ-carotenal, and a vast array of xanthophylls, including lutein (Latari et al., 2015; Bruno et al., 2016). CCD4 has also been associated with the production of aroma and flavor volatiles in Solanum tuberosum and Crocus sativa (Rubio-Moraga et al., 2014; Bruno et al., 2015), turnover of some carotenes during dark-induced leaf senescence, synthesis of the pigment β-citraurin in citrus, and is present in seeds (Gonzalez-Jorge et al., 2013; Rodrigo et al., 2013). Interestingly, the knockdown of CCD4 by RNAi in potato causes elongated shapes in the tubers and higher green stolon production supporting a role in tuber development (Campbell et al., 2010).

The increased number of available genomic sequences has allowed the identification of a novel, evolutionarily conserved CCD that belongs to an ancestral clade, CCD10, consisting of three members: zaxinone synthase (ZAS), CCD-like, and CCD10a from maize (Wei et al., 2016; Wang et al., 2019; Zhong et al., 2020). ZAS is highly conserved in land plants and synthesizes the apocarotenoid zaxinone in vitro, a potential new growth regulator (Wang et al., 2019). In contrast, the maize ZmCCD10a generates different C8 and C13 apocarotenoids when expressed in E. coli. Further, altered levels of this enzyme in plants such as maize and Arabidopsis are linked to changes in the plant’s tolerance to phosphate-limiting conditions and improved phosphate acquisition during mycorrhization (Zhong et al., 2020).

Recently, researchers have exposed novel carotenoid-associated roles for protein families thought to exclusively react with lipid substrates, like the discovery and characterization of the GDSL esterase/lipase (XAT1), which catalyses the esterification of xanthophylls in wheat (Watkins et al., 2019). Accordingly, it is intriguing to consider the possibility that additional enzymes contributing to the production of apocarotenoid signals in plants remain to be discovered. Lipoxygenases (LOXs) represent such proteins.

Despite their most recognized function in the synthesis of oxylipins (Mosblech et al., 2009), LOXs have the capacity to cleave carotenoids (Ramadoss et al., 1978; Siedow, 1991). This family of non-heme-iron-containing enzymes yield non-specific carotenoid oxidation as a secondary reaction of polyunsaturated fatty acid oxidation (Liavonchanka and Feussner, 2006). Most plants contain multiple LOXs that display different efficiencies in cleaving carotenoids (Bannenberg et al., 2009). The role of LOXs in the synthesis of apocarotenoids including flavors and aroma volatiles is emerging. A LOX from Capsicum annuum is responsible for the synthesis of α- and β-ionone volatiles with a potential antibacterial function in response to bacterial infection (Rodriguez-Bustamante and Sanchez, 2007). A pan-genome study in tomato allowed the identification of TomLOX C, a 13S-lipogenase participating in apocarotenoid production important for tomato fruit flavor (Gao et al., 2019). Lastly, the repression of a 9S-lipoxygenase gene, LOX1, in Golden Rice improved β-carotene stability during storage (Gayen et al., 2015). While no evidence was provided directly linking LOX1 with β-carotene cleavage in Golden Rice, it was suggested that the enhanced β-carotene stability resulted from reduced reactive carbonyls produced by LOX1-mediated lipid peroxidation (Gayen et al. 2015). The direct interaction between LOX1 and β-carotene needs to be further explored. Moreover, it remains clear that we have only scratched the surface regarding the potential of LOX enzymes for the synthesis of apocarotenoids in plants, and therefore this subject merits future exploration.

The emerging picture of the diversity of potential apocarotenoid molecules generated by non-enzymatic and enzymatic cleavage with the participation of CCDs and LOXs presents a complex scenario. Similar to ABA and strigolactones, the synthesis of biochemically undefined apocarotenoid retrograde signaling molecules might require multiple enzymatic and non-enzymatic steps, and thus represents an important area of future research.

Contribution of acyclic carotenoids as a source of retrograde signals

Phytoene

Since the early studies, carotenoids have been linked with plastid retrograde signaling (Rodermel, 2001). Blocking carotenogenesis at PDS through genetic mutations or via the application of chemical inhibitors such as norflurazon (NF) has been used for over three decades to analyse plastid tetrapyrrole-mediated retrograde signaling responses (Oelmüller, 1989; Susek and Chory, 1992). An aspect that is worth remarking on is that the retrograde signals generated by changes in the steady state levels of specific carotenoids probably do not directly cause phototoxic tetrapyrrole signals, in contrast to those that result from alterations in the upstream biosynthetic steps, as in the MEP pathway. Alteration of the MEP pathway flow results in a metabolic imbalance of the tetrapyrrole/phytol ratio due to the shortage of GGPP, which causes the accumulation of tetrapyrrole intermediates and generates ¹O₂ (Kim et al., 2013; Bergman et al., 2021). The accumulation of diverse carotenoids should in principle not generate phototoxic signals resulting from tetrapyrrole accumulation. Retrograde responses are also observed for mutations in the IMMUTANS gene, encoding a plastoquinol terminal oxidase required for the desaturation activity of PDS (Carol and Kuntz, 2001; Rosso et al., 2009). Both means of PDS inhibition result in albino phenotypes caused by chloroplast photooxidation due to the lack of downstream photoprotective carotenoids (e.g. β-carotene and xanthophylls). The retrograde signal(s) generated under these conditions drive the transcriptional reprograming of nuclear-encoded genes associated with photosynthesis (PhANGS) and defects in the mesophyll palisade cells (Aluru et al., 2009). While the primary, causal signaling molecule remains elusive, experimental evidence suggests that changes in the redox state of the plastoquinone pool resulting from photodamage generates the primary signal initiating a tetrapyrrole-derived GENOMES UNCOUPLED 1 (GUN1)-mediated retrograde signaling cascade (Pfannschmidt et al., 2009). This hypothesis is supported by similarities in the gene expression profiles between NF-treated and immutans plants, including genes involved in ¹O₂ retrograde responses such as FLU, EXECUTER1 and EXECUTER3 (Aluru et al., 2009; Rosso et al., 2009). Further, analysis of the PDS crystal structure demonstrated that NF occupies the quinone binding site within the active site of PDS, disrupting the redox balance of the plastoquinone pool while inhibiting PDS activity (Brausemann et al., 2017).

Intriguingly, comparative analysis of genes affected in both immutans and NF-treated plants and those from pds3, with a null mutation of PDS, displays only a minor overlap (Qin et al., 2007; Aluru et al., 2009; Foudree et al., 2010). Genes down-regulated specifically in pds3 include those involved in the synthesis of isoprenoid precursors resulting in the dwarf phenotype associated with reduced gibberellin biosynthesis (Qin et al., 2007). Although differences in the transcriptomic profiles might reflect fluctuations in the redox state of the plastoquinone pool (Pfannschmidt et al., 2009; Foudree et al., 2010), other interpretations are plausible. Alternatively, accumulation and subsequent cleavage of phytoene in pds3 may initiate a retrograde signaling cascade in a spatio-temporal manner promoting the reprogramming of nuclear genes unique to pds3 and not shared with the NF-treated or immutans plants. While it remains enigmatic whether phytoene is a substrate of known CCDs, in Arabidopsis evidence is emerging supporting this possibility. In E. coli phytoene is cleaved by CCD7 and by ZmCCD10a resulting in the synthesis of geranylacetone (Schwartz et al., 2004; Zhong et al., 2020). Together, these results support the possibility of phytoene accumulation and subsequent cleavage as a source of unknown retrograde signal(s) (apocarotenoid signal (ACS) 5; Fig. 1).

The potential role of phytoene in retrograde responses is also supported by the ectopic differentiation of chloroplasts to chromoplasts in leaves as a result of phytoene accumulation (Llorente et al., 2020). Arabidopsis and tobacco plants overexpressing the bacterial phytoene synthase enzyme (crtB) in the leaves overaccumulate phytoene beyond a particular threshold. This enhanced accumulation of phytoene resulted in an initial drop in the quantum yield of photosystem II in existing chloroplasts, followed by transcriptional reprogramming of nuclear-encoded genes associated with carotenogenesis and chromoplast differentiation (Fig. 1; Llorente et al., 2020).

While the signaling molecule(s) mediating these responses remains undefined, this adds to a putative function of phytoene or a phytoene derivative(s) in chloroplast reprogramming through modulating nuclear gene expression in response to metabolic cues. Future studies to enhance understanding of phytoene’s role in the morphological changes required for organelle redifferentiation competence and the associated nuclear transcriptional reprogramming remain to be performed.

Phytofluene and ζ-carotene

A developmental control point at ZDS was exposed through the characterization of the CHLOROPLAST BIOGENESIS 5 (clb5) mutant of Arabidopsis, with evidence of specific developmental regulation, further highlighting the role of retrograde communication as a mechanism to regulate central transitions in plant development. Arabidopsis mutants containing a lesion in the ZDS gene, clb5 and spc1-2, are albino plants with major defects in chloroplast development (Dong et al., 2007; Avendaño-Vazquez et al., 2014). While the above phenotypes are shared with other carotenoid mutants (psy and pds3), clb5 displays unique developmental alterations resulting in finger-like leaf morphology and more subtle defects in the primary root (Avendaño-Vazquez et al., 2014; Escobar-Tovar et al., 2021). Transcriptomic analysis demonstrated that clb5 developmental defects are associated with a massive deregulation of a unique set of nuclear genes (Escobar-Tovar et al., 2021) resulting from the overaccumulation of 9,15-di-cis-phytofluene, 9,15,9ʹ-tri-cis- and 9,9ʹ-di-cis-ζ-carotene, immediately upstream of ZDS.

Genetic interactions demonstrated that morphological and expression defects associated with zds mutants are caused by an unidentified apocarotenoid(s), referred to as ACS1, derived from the cleavage of phytofluene and/or ζ-carotene isomers by CCD4 (Avendaño-Vazquez et al., 2014; Hou et al., 2016; Escobar-Tovar et al., 2021). Accordingly, the clb5 ccd4 double mutant displays partial reversion of phenotypic and gene expression defects. However, in vivo and in vitro studies have been unable to confirm the most common cis-ζ-carotenes as direct targets of CCD4 (Huang et al., 2009; Bruno et al., 2016). It is important to note that multiple independent studies have demonstrated increased complexity at the ZDS step of carotenogenesis, increasing the number of ζ-carotene isomers beyond the four most common ζ-carotenes previously tested (Fantini et al., 2013; McQuinn et al., 2020). Another tempting possibility to reconcile these data is that the generation of the ACS1 molecule(s) might involve additional modifications including non-enzymatic and/or enzymatic cleavages, which is consistent with the phenotypic reversion reported under low light fluences (Escobar-Tovar et al., 2021). Together with the variable reports regarding CCD4 substrates (Gonzalez-Jorge et al., 2013; Latari et al., 2015; Bruno et al., 2016), it is imperative to further investigate CCD4’s function and substrate specificity in planta to conclusively determine its role in ACS1 accumulation.

Molecular analysis confirmed ACS1 accumulation results in the inhibition of chloroplast translation, which subsequently drives altered leaf development (Escobar-Tovar et al., 2021). The physiological relevance of such a regulation is intriguing. The L2 layer of the shoot apical meristem (SAM) presents an opportune location for ACS1 accumulation given the predominance of undifferentiated plastids and thus a reduced need for photoprotective carotenoids and xanthophylls (Charuvi et al., 2012). Further, within the SAM important developmental transitions (i.e. vegetative to reproductive development) are strictly regulated by master transcriptional regulators, and are associated with changes in leaf development. In fact, analysis of altered gene regulatory networks in clb5 demonstrate genes for key regulators of meristematic function, floral meristem identity, and floral organ differentiation (i.e. WUSCHEL, SEPALLATA 3, and APETALA 3, respectively) are among the highest up-regulated genes associated with ACS1 accumulation (Jenik and Irish, 2001; Wu et al., 2012; Escobar-Tovar et al., 2021; Lopes et al., 2021). Therefore, the potential role of ACS1 as a retrograde signal regulating the transition to reproductive development is an area of interest for further investigation.

Interestingly, recent data support that the signal mediated by ζ-carotenes might be conserved as the null ZDS mutants in maize (vp-wl2 and alb1) also display morphological aberrations and a massive reprogramming of nuclear-encoded genes (Wang et al., 2020). Similar to clb5, the gene expression defects are alleviated by blocking ζ-carotene overaccumulation. A comparative analysis of the genes affected in Arabidopsis and maize zds mutants could be very useful to further understand the contributions of ζ-carotene isomers in plant development.

Neurosporene and prolycopene

In tomato the epistatic relationship of the null CRTISO mutant tangerine over the PSY mutant yellow flesh uncovered the role of prolycopene and/or tri-cis-neurosporene as potential signals that regulate the expression level of PSY (Kachanovsky et al., 2012). The yellow flesh tomato mutant (r²⁹⁹⁷) displays a characteristic yellow ripe fruit phenotype reminiscent of a lack of carotenoids due to the loss of PSY1 expression. However, ripe fruit of the double mutant of r²⁹⁹⁷ and tangerine (t³⁰⁰²) exhibits elevated neurosporene and prolycopene consistent with the tangerine phenotype (Fig. 1) despite the mutation in PSY1, whose protein acts upstream of CRTISO (Kachanovsky et al., 2012). Molecular characterization of the r²⁹⁹⁷/t³⁰⁰² double mutant demonstrated that partial restoration of the downstream carotenoid levels results from the up-regulation of PSY1 expression in response to the accumulation of prolycopene and/or tri-cis-neurosporene. Whether these cis-carotenoids themselves or an apocarotenoid derivative (ACS2) is responsible for the induction of PSY1 is currently unknown. Considering PSY1 is a critical enzyme controlling the rate of carotenoid biosynthesis, this regulation might be important for fine-tuning this metabolic pathway in response to environmental fluctuations. Further, whether the above phenomenon is limited to PSY1 or additional genes are impacted will merit future analysis.

Apocarotenoid-derived retrograde signals not only regulate PSY transcription but also have been proposed to modulate the levels of the PSY protein post-transcriptionally. In several Brassicaceae species PSY is encoded by a single gene with two splice variants containing different 5ʹ-untranslated regions (5ʹUTR). The presence of a long 5ʹUTR in one variant decreases PSY translation in response to high pathway fluxes as a result of a conformational change in a hairpin loop structure present in the 5ʹUTR (Alvarez et al., 2016). The participation of a carotenoid or a carotenoid-derived molecule has been proposed as a direct interactor with the PSY 5ʹUTR adjusting the pathway flux in response to the environmental changes such as seedling de-etiolation. Even though evidence of a direct interaction between carotenoids or apocarotenoids and the PSY 5ʹUTR is still required, this regulation highlights a novel mechanism potentially involved in (apo)carotenoid retrograde signaling cascades that can mediate the regulation of different proteins, an aspect that should be analysed in the future.

Mutations in plants that affect the expression of CRTISO in tomato, melon, rice, and Arabidopsis lead to a unique virescent leaf phenotype due to perturbed chloroplast differentiation (Fig. 1; Isaacson et al., 2002; Park et al., 2002; Chen et al., 2010; Galpaz et al., 2013; Cazzonelli et al., 2020). Analysis of the Arabidopsis CRTISO mutant (ccr2) demonstrated that this virescent phenotype results from the accumulation of specific cis-carotenes that disrupts etioplast development and prolamellar body (PLB) formation during leaf greening (Cazzonelli et al., 2020). ccr2 also displays altered expression of a specific set of nuclear genes associated with photomorphogenesis (e.g. PHYTOCHROME INTERACTING FACTOR3, PIF3; and ELONGATED HYPOCOTYL5, HY5), photosynthesis (PhANGs) and epigenetic processes (Cazzonelli et al., 2020). Although the identity of the causal retrograde signal has not been defined, the partial recuperation of the PLB observed using the CCD chemical inhibitor D15 suggests it is apocarotenoid in nature (ACS4). Further, phenotypic suppression in the ccr2 ziso-155 double mutant demonstrates that the apocarotenoid controlling etioplast development and PLB formation in ccr2 would be derived from neurosporene and/or di-cis-ζ-carotene (Cazzonelli et al., 2020). This regulation was shown to act in parallel to the repressor of skotomorphogenesis, the transcription factor DEETIOLATED 1 (DET1), controlling etioplast development during the skotomorphogenic process (Cazzonelli et al., 2020).

The examples described above provide important evidence that supports the synthesis of different retrograde signals from acyclic carotenoids that modulate the expression of diverse nuclear genes regulating plant development and differentiation. However, the signaling molecules involved in these regulatory processes remain unidentified.

Contribution of cyclic carotenoids as a source of retrograde signals

Considering the status of the chloroplast as a central environmental sensor and the elevated abundance of β-carotene in close proximity to major sources of ROS within chloroplasts, it is no wonder β-carotene-derived apocarotenoids are entrenched in stress-induced retrograde signaling cascades (Chan et al., 2016; Havaux, 2020). β-Cyclocitral and β-cyclocitric acid are derived from the cleavage of β-carotene as a result of ¹O₂ attack produced by diverse stresses including high light, drought, and herbivore attack (Ramel et al., 2012b; Prasad et al., 2017; D’Alessandro et al., 2019; Havaux, 2020; Mitra et al., 2021). That said, β-cyclocitral may also be synthesized enzymatically by lipoxygenases (AtLOX2 and TomLOXC) and CCD4b as observed in citrus fruits (Rodrigo et al., 2013; Hayward et al., 2017; Gao et al., 2019).

β-Cyclocitral treatment results in massive transcriptional reprogramming with a strong overlap with that of ¹O₂ signaling, and thus it was inferred that β-cyclocitral acts downstream of ¹O₂ sharing several protein components involved in the signal transduction pathway (Ramel et al., 2012b). The β-cyclocitral retrograde signaling cascade splits into three main branches, based on current data. The first branch promotes enhanced phototolerance dependent on the translocation of the zinc finger protein METHYLENE BLUE SENSITIVITY 1 (MBS1) from the cytoplasm to the nucleus, while the second and third branches improve plant defense by enhancing xenobiotic detoxification (reviewed in Havaux, 2020).

The xenobiotic detoxification process is made up of three phases: (i) modification of reactive side groups on toxic reactive carbonyls, (ii) conjugation of modified reactive carbonyls to sugar moieties or glutathione by glycosyl transferases or glutathione-S-transferases, respectively, and (iii) sequestration of the resulting conjugates into the vacuole or the apoplast (Sandermann, 1992). The β-cyclocitral-mediated ‘modification’ phase is dependent on the redistribution of the transcriptional co-activator SCARECROW-LIKE 14 (SCL14) into the nucleus. In the nucleus SCL14 binds with TGACG sequence-specific binding transcription factor (TGAII) resulting in activation of multiple genes encoding enzymes that target reactive carbonyls accumulating during (a)biotic stresses (CHLOROPLASTIC ALDEHYDE REDUCTASE, SHORT-CHAIN DEHYDROGENASE/REDUCTASE 1, ALKENAL REDUCTASE, CHLOROPLASTIC ALDO-KETO REDUCTASE, and GLUTAREDOXIN) via the sequential induction of the ATAF-NAC transcription factors ANAC102 and subsequently ANAC002, ANC032, and ANAC081 (D’Alessandro et al., 2018; Havaux, 2020). Interestingly, β-cyclocitric acid, appears to share the same SCL14-dependent retrograde signaling cascade in its enhancement of drought tolerance in Arabidopsis (D’Alessandro et al. 2019). β-Cyclocitral treatment also promotes the ‘conjugation’ phase of xenobiotic detoxification through the translocation of ENHANCED DISEASE SUSCEPTIBILITY 1 (EDS1), a lipase like protein, into the nucleus where it enhances salicylic acid biosynthesis via the induction of ISOCHORISMATE SYNTHASE 1. Elevated salicylic acid accumulation results in the induction of glutathione-S-transferase genes GST5 and GST13 via the translocation of NON-EXPRESSOR PATHOGENESIS RELATED GENES 1 into the nucleus (Lv et al., 2015; Havaux, 2020).

β-Cyclocitral and/or its derivative β-cyclocitric acid regulates the expression of a large set of nuclear ¹O₂ stress-responsive and defense genes enhancing tolerance to several (a)biotic stresses (Ramel et al., 2012b; D’Alessandro et al., 2018, 2019). Due to the volatile nature of β-cyclocitral and improved solubility of β-cyclocitric acid in the cytosol, these molecules can diffuse from the chloroplast to the cytosol and the nucleus to perform their signaling function. This export initiates a signaling cascade in the cytoplasm that exemplifies the complexity and diversification of the retrograde signals. Given the level of complexity and the resources dedicated to the β-cyclocitral (and possibly β-cyclocitric acid) retrograde signaling cascades and β-cyclocitral’s enhancement of xenobiotic detoxification, it is clear how integral β-carotene and its derivative(s) are during (a)biotic stress responses. Considering this leads one to contemplate whether the plant might not have evolved a more regulated mechanism for β-cyclocitral synthesis rather than random non-enzymatic cleavage by ¹O₂ in response to a particular stress that is extremely time-sensitive. This is a research area that requires further exploration.

Recent data demonstrate that β-cyclocitral can also initiate retrograde signals by direct interaction with a key biosynthetic enzyme in response to herbivore attack (Mitra et al., 2021). Herbivore attack on Arabidopsis plants increases ROS in the plastids resulting in the accumulation of β-cyclocitral. β-Cyclocitral inhibits flux in the MEP pathway, a pathway that provides the precursors for crucial metabolites for photosynthesis, by inhibiting the activity of the first and rate-limiting enzyme of the pathway, 1-deoxy-d-xylulose-5-phosphate synthase (DXS) (Fig. 2; Lange et al., 1998; Mitra et al., 2021). The inhibition of DXS activity apparently results from the direct interaction of β-cyclocitral with the active site of DXS. Consistent with the decrease of DXS activity, reduced accumulation of the downstream products (chlorophylls, carotenoids, and tocopherols) is observed (Mitra et al., 2021). ROS accumulation from herbivory also initiates a parallel retrograde defense response cascade mediated by 2-C-methyl-d-erythritol 2,4-cyclodiphosphate. These findings further uncover the complex dialogue that exist between diverse plastid retrograde signaling pathways.

Fig. 2.

Schematic representation of signaling pathways of apocarotenoids signals produced in plastids. (A) The apocarotenoid retinal synthesized in root plastids from β-carotene regulates the lateral root clock in plants through a signaling cascade that involves its binding to the TEMPERATURE INDUCED LIPOCALIN (TIL) protein. It is unclear if this interaction occurs in the plastid or in the cytoplasm, denotated by the question marks. Retinal–TIL complex relocates to the nucleus where it regulates the expression of genes involved in lateral root organogenesis. Apocarotenoid signal 1 (ACS1) is synthesized from the accumulated phytofluene and/or ζ-carotene in the clb5 mutant with the participation of CAROTENOID CLEAVAGE DIOXYGENASE 4 (CCD4) and light. ACS1 through a yet unknown signaling cascade represses the expression (50%) of nuclear-encoded chlororibosomal proteins (ChRP). ChRP reduction in turn causes a decrease in the plastid translation and a major reduction of a diverse range of plastid-encoded proteins (e.g. ClpP protease subunit, ClpP; ribulose-1,5-bisphosphate large subunt, RbcL; Cytochrome F subunit, PetA). The inhibition of plastid translation generates a second unknown signal (S2) that initiates a second GUN1-mediated signaling cascade that through a yet unknown mechanism reprograms the expression of a set of nuclear genes required in shoot apical meristem (SAM) and flower meristem (FM) identity and leaf development. (B) β-Cyclocitral is synthesized in chloroplasts from β-carotene in response to ROS accumulation produced by different abiotic and biotic stresses such as high light and herbivore attack. The signaling cascade that transduces the β-cyclocitral signal involves three branches. One triggered by high light involves β-cyclocitral driving METHYLENE BLUE SENSITIVITY 1 (MBS1) translocation from the cytosol to the nucleus and the up-regulation of singlet oxygen response genes (SORGs). The other two branches participate in the two phases of the xenobiotic detoxification mechanism to improve plant defense. β-Cyclocitral modulates the ‘modification’ phase by binding the transcriptional co-activator SCARECROW-LIKE 14 (SCL14) leading to its relocation to the nucleus where in conjunction with TGAII it induces the expression of the ATAF-type NAC transcription factor gene ANAC102. ANAC102 activates the expression of ANAC002, ANAC032, and ANAC081 transcription factor genes to induce the expression of genes such as CHLOROPLASTIC ALDEHYDE REDUCTASE (ChlADR), SHORT-CHAIN DEHYDROGENASE/REDUCTASE 1 (SDR1), ALKENAL REDUCTASE (AER), CHLOROPLASTIC ALDO-KETO REDUCTASE (AKR4C9), and GLUTAREDOXIN (GRX480/ROXY19). β-Cyclocitral also promotes the ‘conjugation phase’ of xenobiotic detoxification by driving ENHANCED DISEASE SUSCEPTIBILITY 1 (EDS1) relocalization from the cytosol to the nucleus where it induces the expression of the ISOCHORISMATE SYNTHASE 1 (ICS1) resulting in enhanced accumulation of salicylic acid (SA). Subsequently, SA promotes the translocation of the transcriptional co-activator NON-EXPRESSOR OF PATHOGENESIS RELATED GENES 1 (NPR1) to the nucleus where upregulation of genes encoding GLUTATHIONE-S-TRANSFERASE 5 (GST5) and GLUTATHIONE-S-TRANSFERASE 3 (GST13) proteins follows. β-Cyclocitral accumulates in response to herbivore attack, decreasing the flux of the isoprenoid methylerythritol 4-phosphate (MEP) pathway through a direct inhibition of the rate-limiting 1-deoxy-d-xylulose 5-phosphate (DXS) enzyme activity. Inhibition of the MEP flux results in a reduction of the MEP pathway final products (chlorophyll, carotenoids, and tocopherols). Dashed lines indicate that the exact mechanism is still unknown.

Inhibition of carotenoid biosynthesis genetically or by chemical inhibitors such as NF or 2-(4-chlorophenylthio)-triethylamine hydrochloride affects lateral root number in Arabidopsis. Similar defects were observed with the CCD inhibitor D15, demonstrating that an apocarotenoid molecule (ACS3) synthesized from the β-carotene branch is responsible for the correct establishment of the lateral roots (Van Norman et al., 2014; Hou et al., 2016). Interestingly, recent work supports that ACS3 corresponds to retinal, a β-carotene-derived apocarotenoid commonly found in animals, which accumulates naturally in the roots of Arabidopsis where it binds to Temperature Induced Lipocalin (TIL) regulating root clock oscillations, predicting prebranch site formation in the developing root (Dickinson et al., 2021).

The impact of apocarotenoid retrograde signaling is not limited to lateral root branching. Development of anchor roots, specialized roots produced below the hypocotyl–root junction in plants like Arabidopsis, is regulated by a carotenoid-derived signal molecule known as anchorene (Jia et al., 2019). Anchorene is a diapocarotenoid (C₁₀) initially identified from in vitro analysis of apocarotenoid products from the cleavage at C11–C12 and C11ʹ–C12ʹ bonds of the most commonly known plant carotenoids (Ilg et al., 2014). In support of the biological function of anchorene, the Arabidopsis carotenoid-deficient mutant psy lacks anchor roots, and this defect is fully rescued by the exogenous application of anchorene (Jia et al., 2019). While the specific carotenoid precursor of anchorene remains elusive, inhibition of carotenogenesis either genetically or chemically in multiple carotenoid biosynthetic steps suggests that the β-carotene branch may contain the precursor of anchorene (Jia et al., 2019). The presence of anchorene alters the expression of a wide set of genes, including several related to auxin metabolism, supporting that anchorene modulates auxin homeostasis.

The phylogenetic characterization of the CCD gene family allowed the discovery of a new CCD, zaxinone synthase (ZAS), present in multiple plants, except in some Brassicaceae (Wang et al., 2019). In vitro, ZAS synthesizes a C18-ketone apocarotenoid named zaxinone by cleaving 3-OH-apocarotenals. The zas mutant in rice displays defects in overall plant growth, elevated strigolactone accumulation, defects in mycorrhization, and higher susceptibility to parasitic plant infestation (Striga) (Wang et al., 2019). The exogenous application of zaxinone reverts most of the phenotypes observed in the zas rice mutant, demonstrating that zaxinone is a positive regulator of plant development and biotic interactions and acts as a negative regulator of strigolactone biosynthesis (Wang et al., 2019). In contrast, recent data have shown that zaxinone is also synthesized in Arabidopsis through a ZAS-independent pathway where it promotes ABA and strigolactone biosynthesis in roots, but it does not increase growth (Ablazov et al., 2020). These findings support that zaxinone is a conserved metabolite among plants, but its function has diversified across different plant species.

What is next? Retrograde signal transduction/translocation remains largely unexplored

Recent advances have pinpointed novel apocarotenoid signal sources, whether it be the carotenoid precursor or the actual causal apocarotenoid signal, and the targeted developmental or physiological modulation including transcriptional alterations. Meanwhile, how the apocarotenoid signal is transduced or translocated from inside the plastid to the nucleus to drive transcriptional reprogramming remains largely unexplored. It has been suggested that the volatile nature of some apocarotenoids or their improved solubility in the cytoplasm provides the means for translocation. However, as the length of the carbon chain of the apocarotenoid signal increases so does its hydrophobicity, reducing its volatility, solubility, and mobility through the aqueous phase of the cellular compartments. Further, numerous developmental regulatory mechanisms and environmental stress responses are time dependent and therefore require well-synced accumulations of both the signal and the protein-signaling components for timely transcriptional reprogramming.

Investigation into the apocarotenoid retrograde signaling cascades has been limited to a few β-carotene-derived signals (i.e. β-cyclocitral and retinol) and the undefined ACS1, exposing the level of complexity, enhanced allocation of protein resources, and as described, integration with other retrograde and hormone signaling pathways. The β-cyclocitral retrograde signaling cascades appear to begin with the translocation of various transcriptional proteins from the cytoplasm to the nucleus, suggesting the primary site of β-cyclocitral perception. However, the exact means of perception is enigmatic. One protein of interest is EDS1, translocated to the nucleus in response to β-cyclocitral treatment (D’Alessandro et al., 2018). It is intriguing to speculate that EDS1 may directly bind and/or react with β-cyclocitral (Fig. 2), based on the recent discoveries demonstrating secondary carotenoid-associated roles of lipid-specific enzymes (Gao et al., 2019; Watkins et al., 2019). Regarding retinol, given the length of the carbon chain and resulting reduction in solubility and presumed mobility, a potential means for its facilitated translocation was discovered in TIL (Dickinson et al., 2021). TIL is closely related to the human RETINOL BINDING PROTEIN 4 (RBP4) and such proteins have been demonstrated to be integral in translocation of retinol/retinal/retinoic acid necessary to initiate transcriptional reprogramming (Steinhoff et al., 2021). The chloroplast localization of TIL (Abo-Ogiala et al., 2014) suggests it may bind retinol at the site of production facilitating its translocation to the cytosol or further to the nucleus.

As for ACS1, researchers have exposed the sequences of events connecting ACS1 accumulation to aberrant leaf development, dependent on GUN1, a prominent protein in multiple retrograde signaling pathways (Tadini et al., 2016; Wu et al., 2019; Escobar-Tovar et al., 2021). The ACS1 signal initiates a retrograde signaling cascade resulting in the massive reprogramming of nuclear genes, including the severe repression of around 50% of the nuclear-encoded 70S chlororibosomal proteins (Escobar-Tovar et al., 2021). Given this transcriptional repression is specifically associated with the accumulation of the ACS1 signal and is not observed in other carotenoid pathway mutants or in plastid translation-inhibited plants, it can be inferred that these genes represent primary targets of the apocarotenoid ACS1 signal (Escobar-Tovar et al., 2021). The severe reduction in chlororibosomal proteins negatively impacts overall plastid translation resulting in low or even non-existent accumulation of diverse plastid-encoded proteins resembling the effects of treatment with lincomycin, an inhibitor of plastid translation (Koussevitzky et al., 2007; Ruckle et al., 2012; Tameshige et al., 2013; Escobar-Tovar et al., 2021).

This work exposed the epistatic nature of plastid translation inhibition initiating a second signaling cascade mediating the transcriptional regulation of nuclear genes related to the developmental processes in a GUN1-dependent manner underpinning the aberrant leaf and root morphology in clb5 and lincomycin-treated plants alike (Fig. 2; Escobar-Tovar et al., 2021). Altogether, these findings exemplify the complexity in circuit networks established between different plastid retrograde signaling cascades with the nuclear genome. Further, this highlights the necessity of a detailed molecular analysis to properly dissect the signal transduction pathways pivotal for the identification of integral protein components.

Elucidation of the additional requisite proteins in the apocarotenoid signal transduction pathways may rely on what is known in animal systems. Given the demonstrated similarities between apocarotenoid production (i.e. CCDs and BCOs) and the resulting apocarotenoids’ roles as important signaling molecules across plant and animal kingdoms, it is logical to infer the signal transduction and/or translocation mechanisms may also be conserved (McQuinn et al., 2015). If in fact this is the case, an excellent model apocarotenoid signaling pathway exists in the well-known retinoic acid (RA), vitamin A, signaling pathway characterized in animals (Iskakova et al., 2015). In animals, RA first binds to lipid binding proteins (i.e. Cellular Retinoic Acid Binding Protein II (CRABP II) and Fatty-Acid Binding Protein 5 (FABP5)) in the cytosol, which facilitates transport to the nucleus of RA, where it binds to nuclear hormone receptors (Retinoic Acid Receptor (RAR) and PPARb/d). RA-bound RAR and PPARb/d subsequently form multimeric complexes to drive the desired transcriptional modulations (Iskakova et al., 2015).

Given what is known of retinal/retinol/RA signaling pathways in animals, the Calycin protein superfamily emerges as an interesting area to explore for novel plant members with apocarotenoid binding capabilities. The Calycin superfamily, includes the CRABPs and FABPs, as well as lipocalins, of which retinol binding proteins RBP4 and its closely related homologue in plants, TIL, are members. These lipid binding proteins could provide the necessary ‘raft’ to transport the more hydrophobic apocarotenoid signals from the plastid to the cytosol and/or nucleus. Sequence similarity networks represent an exciting tool to explore the diversification and conservation of protein functions within a large protein family based on amino acid sequence similarities shared with structurally and functionally characterized proteins within the same family (Atkinson et al., 2009; Ahmed, 2016). In this way, sequence similarity networks may allow researchers to ascertain whether other apocarotenoid binding proteins integral in plant apocarotenoid signaling exist.

Conclusion

In this review we have described the ample evidence demonstrating that different acyclic and cyclic carotenoids are a source of retrograde signals. These signals have been shown to impact organelle differentiation, basic aspects of plant development, and essential environmental responses. While the identification of new carotenoid-derived retrograde signals has substantially increased in past years, there are still many tasks pending. First, defining the chemical structure of the apocarotenoid signals and their biosynthetic pathway, in particular those derived from the acyclic carotenoids, remains a major challenge. This challenge is exacerbated by the apocarotenoid signals’ low abundance and the complexity of their synthesis, potentially involving the participation of unique, novel carotenoid cleaving enzymes as well as non-enzymatic modifications. This is nicely exemplified by the in vivo CCD4 potential targets observed, the putative role of light in the synthesis of ACS1, and by the participation of new carotenoid cleaving enzymes extending beyond the constraints of the CCD family. Advances in these particular aspects will probably require the implementation of new methodologies such as the ones described for zaxinone and anchorene.

Understanding the molecular mechanisms that participate in the signaling cascade that ultimately leads to regulation of molecular and cellular responses also remains a major task. The emerging scenarios from a limited number of carotenoid-derived signaling pathways demonstrate the high order of complexity within the signaling cascades. The concept of retrograde signal pathways has evolved to included complex scenarios in which an initial signal may not even leave the organelle to initiate its signaling cascade, as illustrated by β-cyclocitral. Also, major interconnections with other signaling cascades (e.g. hormones and light responses), as well as downstream bifurcations, are emerging as common scenarios for many apocarotenoid signaling pathways. Detailed phylogenetic analysis and sequence similarity networks might provide insight into the structural and functional conservation of some protein components within apocarotenoid signaling cascades in plants, based on what is known for animal signaling pathways.

Finally, defining biological functions and impact on plant development and environmental responses is also a major challenge that merits important efforts in upcoming years.

Acknowledgements

We thank Arihel Hernández-Muñoz and Lina María Escobar-Tovar for their comments and suggestions.

Author contributions

All authors wrote and discussed each section of the article, JSC and RPMQ prepared figures, and JSC performed expression analysis.

Conflict of interest

The authors declare they have no conflicts of interest.

Funding

This research was supported by Consejo Nacional de Ciencia y Tecnología [FCI 316070] and Programa de Apoyo a Proyectos de Investigación e Innovación Tecnológica-Dirección General de Asuntos del Personal Académico Universidad Nacional Autónoma de México [IN208620] grants to PL. JSC was supported by a PhD fellowship from CONACYT.

Author notes

These authors contributed equally to this work.