The Passage of H2O2 from Chloroplasts to Their Associated Nucleus during Retrograde Signalling: Reflections on the Role of the Nuclear Envelope

The response of chloroplasts to adverse environmental cues, principally increases in light intensity, stimulates chloroplast-to-nucleus retrograde signalling, which leads to the induction of immediate protective responses and longer-term acclimation. Hydrogen peroxide (H2O2), generated during photosynthesis, is proposed to both initiate and transduce a retrograde signal in response to photoinhibitory light intensities. Signalling specificity achieved by chloroplast-sourced H2O2 for signal transduction may be dependent upon the oft-observed close association of a proportion of these organelles with the nucleus. In this review, we consider more precisely the nature of the close association between a chloroplast appressed to the nucleus and the requirement for H2O2 to cross both the double membranes of the chloroplast and nuclear envelopes. Of particular relevance is that the endoplasmic reticulum (ER) has close physical contact with chloroplasts and is contiguous with the nuclear envelope. Therefore, the perinuclear space, which transducing H2O2 molecules would have to cross, may have an oxidising environment the same as the ER lumen. Based on studies in animal cells, the ER lumen may be a significant source of H2O2 in plant cells arising from the oxidative folding of proteins. If this is the case, then there is potential for the ER lumen/perinuclear space to be an important location to modify chloroplast-to-nucleus H2O2 signal transduction and thereby introduce modulation of it by additional different environmental cues. These would include for example, heat stress and pathogen infection, which induce the unfolded protein response characterised by an increased H2O2 level in the ER lumen.


Introduction
Chloroplast-to-nucleus (retrograde) signalling is an important part of plants' capacity to sense and act upon changes in their environment, especially those that require eventual adjustments to photosynthetic capacity. The ability to coordinate immediate and longerterm responses to environmental perturbations occurs at the cellular, tissue and whole plant (systemic) level [1][2][3][4][5][6][7]. A particularly active area within this research sphere is the quest to identify the precise signalling routes between chloroplasts and the nucleus. Several signalling pathways and signal initiators and transducers have been identified and continue to attract attention, although there are undoubtedly many more to be uncovered [8][9][10][11][12][13].
The close association of a proportion of a cell's chloroplast complement with its nucleus is a feature of all plant species so far examined [11,14,15]. More recently, this relationship has received growing attention since the juxtaposition of a subset of chloroplasts with the nucleus is suggested to be a crucial feature in the communication and coordination of highly complex processes between these organelles in response to developmental and environmental cues. [4,11,[16][17][18]. Since some signalling molecules could originate from envelope membrane [11,[41][42][43]. The transient tethering of chloroplasts to the ER occurs at so-called membrane contact sites (MCS), which have been defined as "areas of close apposition between the membranes of two organelles" but crucially, the two organellar membranes do not fuse [38]. MCS are regarded as having specific functions, acting to concentrate protein-protein interactions to allow transfer of molecules between compartments [38]. The bidirectional exchange of lipids between the ER and chloroplasts via such MCS has been studied to some extent. Notably, transorganellar complementation experiments elegantly demonstrated the existence of metabolic continuity in biosynthetic pathways, which span both organelles [44,45]. These tethers between chloroplasts and the ER are such that a 400 pN force applied with optical tweezers could not separate them [39,46,47]. Various biophysical, genetic, biochemical and microscopy methodologies have begun to provide a picture of the complexity of these interactions and the reader is referred to the comprehensive review on this subject by Baillie et al. [39]. space between the inner and outer nuclear membrane is contiguous with the ER lumen [36]. Chloroplasts, like many other organelles that form physical interactions with the ER, are tethered to the outer ER/nuclear membrane typically at 10-30 nm distance [37][38][39][40]. The ER outer membrane is thus frequently in very close association with the outer chloroplast envelope membrane [11,[41][42][43]. The transient tethering of chloroplasts to the ER occurs at so-called membrane contact sites (MCS), which have been defined as "areas of close apposition between the membranes of two organelles" but crucially, the two organellar membranes do not fuse [38]. MCS are regarded as having specific functions, acting to concentrate protein-protein interactions to allow transfer of molecules between compartments [38]. The bidirectional exchange of lipids between the ER and chloroplasts via such MCS has been studied to some extent. Notably, transorganellar complementation experiments elegantly demonstrated the existence of metabolic continuity in biosynthetic pathways, which span both organelles [44,45]. These tethers between chloroplasts and the ER are such that a 400 pN force applied with optical tweezers could not separate them [39,46,47]. Various biophysical, genetic, biochemical and microscopy methodologies have begun to provide a picture of the complexity of these interactions and the reader is referred to the comprehensive review on this subject by Baillie et al. [39]. A long-observed phenomenon is the avoidance response of chloroplasts whereby they move away from high fluence blue light, which is controlled by phototropins and uses the actin cytoskeleton to guide movement [48,49]. Interestingly, the nucleus, which A long-observed phenomenon is the avoidance response of chloroplasts whereby they move away from high fluence blue light, which is controlled by phototropins and uses the actin cytoskeleton to guide movement [48,49]. Interestingly, the nucleus, which has no capacity to move independently, is towed by its attached chloroplasts [50]. Undoubtedly, many proteins are involved in the combined tethering of chloroplasts to nuclei and their repositioning in the cell, as well as being involved in other functions such as anchoring of plastids to the plasma membrane and chloroplast division. Examples include CHLOROPLAST UNUSUAL POSITIONING1 (CHUP1), KINESIN-LIKE PROTEIN FOR ACTIN-BASED CHLOROPLAST MOVEMENT1 (KAC1) and KAC2, PLASTID DIVISION1 (PDV1) and PDV2 and PARALOG OF ARC6 (PARC6) [15,[50][51][52][53][54][55][56].
From a structure-function perspective, CHUP1 currently is one of the best-understood proteins engaged in chloroplast relocation and positioning [57]. CHUP1 localises to the chloroplast envelope and to do this requires the first 25 N-terminal residues, which form a hydrophobic domain. The remainder of the protein protrudes outwards into the cytosol. A coiled-coil region (residues 65-276), an F-actin binding region (residues 350-360) and proline-rich region (residues 670-710) ensure the anchoring of the chloroplast to the plasma membrane and linking it to the actin cytoskeleton and/or its polymerisation. Completing the protein is a conserved C-terminal region (residues 720-1004) which binds profilin [58]. CHUP1 forms homodimers via leucine zippers contained within its N-terminal coiled-coil region [59] and has the effect of bringing the proline-rich and actin binding domains into close proximity [59]. Most recently, it has been shown that the conserved C-terminal region also forms dimers and is a novel plant-specific actin nucleator sharing structural homology, but not sequence homology, to the FH2 C-terminal domain dimers of formins that regulate actin polymerisation across the Eukarya [60,61]. It should be emphasised that the aforementioned studies did not specifically address nuclear-chloroplast connectivity having focussed instead on chloroplast-plasma membrane connectivity. Nevertheless, one important observation is that CHUP1 may be a negative regulator of stromule formation [16] and in addition there is, to our knowledge, no information on how or even if CHUP1 is part of chloroplast-ER/outer nuclear membrane MCS. It was suggested recently that direct contact between plastids, the nucleus, and the same connections involving stromules are a continuum of essentially the same process and may provide a means of distinguishing the role of CHUP1 in chloroplast-nuclear connections from that of chloroplast-plasma membrane association [15].
One further consideration in this inter-organellar communication is the role of nuclear pore complexes (NPCs) [62][63][64] which punctuate the nuclear membranes. This is the route for the trafficking of macromolecules, most commonly proteins and nucleic acids [62,63]. This could be a route for the transfer of proteins engaged in retrograde signalling such as WHIRLY1 [65]. However, it is not clear that small molecules enter the nucleus via NPCs. Therefore, while it is a theoretical route for trafficking H 2 O 2 or an oxidising equivalent as an oxidised protein there is no evidence of this and therefore no further consideration of NPCs will be undertaken here.

H 2 O 2 , Aquaporins and the Route to the Nucleus
From the above considerations, it can be proposed that there is close association between some of a cell's complement of chloroplasts and the nucleus, which would also involve both organelles tied into the cytoskeleton with the strength of the connections determined by tethering through MCS. More precisely, for H 2 O 2 to travel from the chloroplast stroma to the nucleus then it must not only cross the chloroplast double envelope, but also the outer and then inner nuclear membrane separated by the perinuclear space.
The movement of H 2 O 2 across membranes is considered to occur by diffusion down a concentration gradient facilitated by membrane intrinsic proteins (aquaporins; AQPs; reviewed by Bienert and Chaumont [66]). However, H 2 O 2 diffusion into red blood cells is not facilitated by AQPs but by an unknown membrane protein or through the lipid fraction [67] raising the possibility of AQP-independent means of transporting H 2 O 2 between cellular compartments. This is despite physico-chemical considerations concluding that simple diffusion of H 2 O 2 across membranes can be disregarded [66,67]. Instead, all AQPs that transport H 2 O may also transport H 2 O 2 , although there are differences in the efficiency of how individual AQP isoforms discriminate between these two molecules [66,68,69]. Assuming a uni-directional movement of signal-transducing H 2 O 2 to the nucleus from attached chloroplasts, then its journey would include crossing the chloroplast envelope membranes (Figure 2). Isolated chloroplasts exposed to high light intensities secrete H 2 O 2 into their medium [31] and this is blocked by the AQP inhibitor acetazolamide [70]. Of the 35 AQPs in Arabidopsis [71], up to 5 may be present in the chloroplast. Of these, at least two isoforms of the tonoplast intrinsic protein (TIP1;1 and TIP1;2) AQP family and one of the plasma membrane intrinsic proteins, PIP2a, may span the inner chloroplast envelope membrane [72][73][74]. Therefore, the current evidence strongly suggests that AQPs are the exit route out of the chloroplast for H 2 O 2 .
clear membrane, this would facilitate the transfer of H2O2 to the perinuclear space. Mitochondrial-ER MCS in animal cells form an environment where H2O2 does indeed concentrate in microdomains either side of the mitochondrial envelope [75]. It can be surmised that an analogous arrangement around chloroplast-outer nuclear/ER membrane could exist and certainly H2O2 microdomains have been observed associated with Nb epidermal chloroplasts [4]. Once in the perinuclear space, H2O2 would be in an oxidising environment (see following section) and therefore would have time to diffuse to the vicinity of any AQPs located on the inner nuclear membrane for its entry into the nucleus.

Figure 2. A proposed route for a transducing H2O2 retrograde signal.
In this case, the chloroplasts and nucleus are in close association linked by the nuclear envelope and possibly influenced by H2O2 produced in the ER lumen. The H2O2 generated by photosynthetic electron transport passes through membranes facilitated by aquaporins and arrives in the nucleus to transfer its oxidising equivalents to a redox relay network ultimately leading to the activation of a range of diverse regulatory proteins, which may act in the nucleus or migrate to other subcellular sites.

Figure 2. A proposed route for a transducing H 2 O 2 retrograde signal.
In this case, the chloroplasts and nucleus are in close association linked by the nuclear envelope and possibly influenced by H 2 O 2 produced in the ER lumen. The H 2 O 2 generated by photosynthetic electron transport passes through membranes facilitated by aquaporins and arrives in the nucleus to transfer its oxidising equivalents to a redox relay network ultimately leading to the activation of a range of diverse regulatory proteins, which may act in the nucleus or migrate to other subcellular sites.
The likelihood of very close contact between the chloroplast envelope and the outer nuclear envelope (see above) could include a localised increased concentration in microdomains at or near MCS and, if there is close proximity of further AQPs in the outer nuclear membrane, this would facilitate the transfer of H 2 O 2 to the perinuclear space. Mitochondrial-ER MCS in animal cells form an environment where H 2 O 2 does indeed concentrate in microdomains either side of the mitochondrial envelope [75]. It can be surmised that an analogous arrangement around chloroplast-outer nuclear/ER membrane could exist and certainly H 2 O 2 microdomains have been observed associated with Nb epidermal chloroplasts [4]. Once in the perinuclear space, H 2 O 2 would be in an oxidising environment (see following section) and therefore would have time to diffuse to the vicinity of any AQPs located on the inner nuclear membrane for its entry into the nucleus.
It should be emphasised that these considerations on the route from attached chloroplast to nucleus is informed speculation (Figure 2) based on the more complete information available from other eukaryotic cells. Whether this route for H 2 O 2 actually exists in plant cells awaits experimental investigation.

H 2 O 2 in the Perinuclear Space and ER Lumen and Its Impact on Retrograde Signalling
In animal cells, the ER lumen is regarded, along with mitochondria and peroxisomes, as a major source of H 2 O 2 for signalling [68,[76][77][78]. These organelles are often found in very close proximity to each other and may secrete H 2 O 2 into a shared microdomain in which proteins involved in further transducing the oxidising signal are also present. The cooperation between these three compartments to form a cytosol-located H 2 O 2 microdomain has been termed the "redoxosome" [78]. A redoxosome for these same organelles but also including chloroplasts has been suggested as possible in plant cells, but this suggestion remains unexplored [79]. It has been proposed that in animal cells, the directing of H 2 O 2 to the redoxosome ensures that it does not accumulate in the nucleus and cause oxidative damage there. However, plant cells subjected to environmental stress can accumulate chloroplast-sourced H 2 O 2 in their nucleus [4,16]. This suggests that the organisation of the spatial components of H 2 O 2 -mediated retrograde signalling may differ from those involving non-plastid organelles, which may share a degree of conservation across the Eukarya.
The midpoint redox potential of the reduced glutathione-glutathione disulphide (GSH-GSSG) couple (E GSH ) in the ER lumen is −208 ±4 mV, which is more oxidising than that of the cytosol at ca. −320 mV in animal cells [80]. However, very recent in vivo measurements conducted on Arabidopsis ER suggest a slightly more reducing E GSH of −241 mV [81]. Irrespective of these differences between animal and plant cells, the ER lumen environment allows the chaperone-catalysed oxidative folding of proteins to occur that requires molecular oxygen (O 2 ) and from which H 2 O 2 arises (Figure 3). This is a highly conserved process in all eukaryotic cells. Oxidative stress in the ER is caused when this protein folding activity exceeds the capacity of the lumen antioxidant system to remove the H 2 O 2 formed. GLUTATHIONE PEROXIDASE7 (GPX7), GPX8 and PEROXIREDOXIN4 (PRDX4) scavenge the H 2 O 2 generated by the ER oxidoreductase1 (ERO1)-catalysed oxidation of the PROTEIN DISULPHIDE ISOMERASE (PDI) isoforms ( Figure 3). Despite their names, GPX7 and GPX8 use reduced PDI isoforms as electron donors and not GSH [77,82]. There are also additional ERO1-independent means of generating H 2 O 2 [83].
as a major source of H2O2 for signalling [68,[76][77][78]. These organelles are often found in very close proximity to each other and may secrete H2O2 into a shared microdomain in which proteins involved in further transducing the oxidising signal are also present. The cooperation between these three compartments to form a cytosol-located H2O2 microdomain has been termed the "redoxosome" [78]. A redoxosome for these same organelles but also including chloroplasts has been suggested as possible in plant cells, but this suggestion remains unexplored [79]. It has been proposed that in animal cells, the directing of H2O2 to the redoxosome ensures that it does not accumulate in the nucleus and cause oxidative damage there. However, plant cells subjected to environmental stress can accumulate chloroplast-sourced H2O2 in their nucleus [4,16]. This suggests that the organisation of the spatial components of H2O2-mediated retrograde signalling may differ from those involving non-plastid organelles, which may share a degree of conservation across the Eukarya.
The midpoint redox potential of the reduced glutathione-glutathione disulphide (GSH-GSSG) couple (EGSH) in the ER lumen is -208 ±4 mV, which is more oxidising than that of the cytosol at ca. −320 mV in animal cells [80]. However, very recent in vivo measurements conducted on Arabidopsis ER suggest a slightly more reducing EGSH of -241 mV [81]. Irrespective of these differences between animal and plant cells, the ER lumen environment allows the chaperone-catalysed oxidative folding of proteins to occur that requires molecular oxygen (O2) and from which H2O2 arises (Figure 3). This is a highly conserved process in all eukaryotic cells. Oxidative stress in the ER is caused when this protein folding activity exceeds the capacity of the lumen antioxidant system to remove the H2O2 formed. GLUTATHIONE PEROXIDASE7 (GPX7), GPX8 and PEROXIREDOXIN4 (PRDX4) scavenge the H2O2 generated by the ER oxidoreductase1 (ERO1)-catalysed oxidation of the PROTEIN DISULPHIDE ISOMERASE (PDI) isoforms ( Figure 3). Despite their names, GPX7 and GPX8 use reduced PDI isoforms as electron donors and not GSH [77,82]. There are also additional ERO1-independent means of generating H2O2 [83]. . This H 2 O 2 may be scavenged by an ER glutathione peroxidase (GPX3), although the reductant for this enzyme is suggested to be protein disulfide isomerase (PDI) isoforms, which are members of the thioredoxin super-family. This proposed redox cycle is adapted from and available in more detail in the review by Meyer et al. [84].
The increased H 2 O 2 levels in the ER lumen can drive signalling, most notably the initiation of the Unfolded Protein Response (UPR), which acts to mitigate against the accumulation of unfolded or misfolded proteins in the ER lumen. One branch of the UPR is mediated by a pair of ER membrane-associated bZIP transcription factors-bZIP17 and bZIP28. UPR is also activated as a consequence of environmental perturbations including exposure to heat/chilling stress, oxidative stress, salt stress, induction of immunity and senescence [79,[85][86][87].

Suppression of the UPR by High Light Intensities
The transfer of H 2 O 2 from high light-exposed chloroplasts to their associated nucleus is an important step in the retrograde signalling mediated by this reactive oxygen species (ROS) [4,11,88]. Interestingly, exposure to high light suppresses the UPR, which is linked to the production of the ROS singlet oxygen ( 1 O 2 ) [87]. This is achieved by activation of the bZIP transcription factor LONG HYPOCOTYL5 (HY5), which competes with bZIP28 for binding to the promoters of UPR-activated genes and suppressing their induction [89]. The HY5-mediated negative regulation of the UPR involving 1 O 2 may be linked to the recent identification of HY5 as a positive regulator of high light acclimation [90]. This is because the relative levels of 1 O 2 and H 2 O 2 may be a good indicator of the type of physiological response a plant carries out when exposed to increased light intensities [88,[90][91][92][93].

Conclusions and Possibilities
The corollary of the above arguments is that during transit across the perinuclear space-an extension of the ER lumen-there could be the opportunity to modulate retrograde signalling mediated by H 2 O 2 from the chloroplast on its way to the nucleus. One can envisage two converse scenarios: (a) increased H 2 O 2 from the ER lumen augmenting H 2 O 2 coming from chloroplasts and amplifying a stress-responsive signal; or the opposite: (b) the attenuation of a retrograde signal at this point by increased and highly localised antioxidant activity. These possibilities now have the prospect of being tested with the advent of a novel GSH:GSSG redox biosensor that functions in the plant ER lumen [79] together with the possibility of using a modified Hyper, called Triper, to detect H 2 O 2 , which elegantly sidesteps problems of this biosensor's over-oxidation and its consequent non-responsiveness [77].
In conclusion, if the considerations in this essay are correct then this could provide a means of intervening in retrograde signalling to tailor a crop plant's response to environmental stress [13]. This may prove to be an easier option than trying to manipulate a H 2 O 2 signal once it has arrived in the nucleus considering the transfer of oxidising equivalents is likely through an extensive and highly mobile network of intermediate redox carriers [25,94,95] to a plethora of recipient redox sensitive regulatory proteins.