Abstract
The non-photosyntetic alga Euglena longa harbours a cryptic plastid of unknown function. By a combination of bioinformatic and biochemical approaches we found out that this organelle houses a surprising set of metabolic processes. Biosynthesis of isoprenoid precursors and fatty acids is absent and the tetrapyrrole pathway is incomplete, whereas phospholipids and glycolipids are still being produced in the E. longa plastid. Unprecedented among non-photosynthetic plastids is the ability of this organelle to make tocopherols and a phylloquinone derivative. The most striking attribute is the presence of a linearized Calvin-Benson (CB) pathway including RuBisCO, together with ferredoxin-NADP+ reductase and the ferredoxin/thioredoxin system. We hypothesize that the linear CB pathway is regulated by the redox status of the E. longa cell, in effect functioning as a redox valve bypassing the glycolytic oxidation of glyceraldehyde-3-phosphate to 3-phosphoglycerate. Altogether, the E. longa plastid defines a new class of relic plastids.
Introduction
Plastids are organelles that evolved from a cyanobacterial endosymbiont in the ancestor of Archaeplastida and later found their way into other eukaryotic lineages by secondary or even higher-order endosymbioses (Keeling 2013; McFadden 2014; Ponce-Toledo et al. 2017). Photosynthetic harvesting of solar energy is supposedly the primary metabolic function and evolutionary advantage of plastid acquisition. However, plastids also host a variety of other metabolic pathways, such as biosynthesis of amino and fatty acids, isopentenyl pyrophosphate (IPP) and its derivatives (isoprenoids), and tetrapyrroles (Neuhaus and Emes 2000; Oborník and Green 2005; Van Dingenen et al. 2016). Hence, reversion of photosynthetic lineages to heterotrophy typically does not imply plastid loss, and non-photosynthetic plastids are found in many eukaryotic lineages (Wilson et al. 1996; Sanchez-Puerta et al. 2007; Slamovits and Keeling 2008; Janouškovec et al. 2015; Kamikawa et al. 2017; Hadariová et al. 2018). In these cases, metabolic integration of the plastid presumably resulted in the host biochemistry being dependent on compound(s) supplied by the organelle (Oborník et al. 2009; Lim and McFadden 2010; Janouškovec et al. 2015; Hadariová et al. 2018).
The most extensively studied relic plastid is undoubtedly the apicoplast of apicomplexan parasites (especially of Plasmodium falciparum and Toxoplasma gondii). The knowledge of the apicoplast has expanded tremendously since its discovery two decades ago (McFadden and Yeh 2017). The apicoplast is ultimately derived from a red alga (Williamson et al. 1994; Moore et al. 2008; Janouškovec et al. 2010), with an ochrophyte alga being a possible direct donor of the plastids in apicomplexans and their photosynthetic relatives, chromerids (Ševčíková et al. 2015; Füssy and Oborník 2017). Except for the recently characterized corallicolid apicomplexans (Kwong et al. 2019), the apicoplast genome contains only genes related to gene expression, protein turnover, and assembly of FeS clusters, not considering a few short hypothetical open reading frames of unknown function (Wilson et al. 1996; Janouškovec et al. 2015). The essentiality of the apicoplast for parasite survival has attracted much attention, partly because this organelle is a promising target for parasite-specific inhibitors (e.g. Miller et al. 2013; McFadden and Yeh 2017). So far, three plastid pathways (encoded by the nuclear genome) seem to condition the apicoplast retention: non-mevalonate IPP synthesis, haem synthesis, and type II fatty acid synthesis (FASII). IPP biosynthesis in particular is vital for the bloodstream form of P. falciparum (Yeh and DeRisi 2011). On the other hand, the mosquito and liver stages of P. falciparum are dependent on haem synthesis (Nagaraj et al. 2013), while FASII is indispensable for pellicle formation in Toxoplasma gondii (Ke et al. 2014; Martins-Duarte et al. 2016).
Less is known about the actual metabolic functions of plastids in other non-photosynthetic algal lineages. Many of them have a similar metabolic capacity as the apicoplast (Sanchez-Puerta et al. 2007; Slamovits and Keeling 2008; Fernández Robledo et al. 2011), and some house an even more complex metabolism that includes amino acid biosynthesis and carbohydrate metabolism pathways (Borza et al. 2005; Pombert et al. 2014; Smith and Lee 2014). Until recently, IPP synthesis seemed to be a process conserved even in the most reduced relic plastids, such as the genome-lacking plastids of certain alveolates (Matsuzaki et al. 2008; Janouškovec et al. 2015), but non-photosynthetic plastids lacking the characteristic plastidial (MEP or DOXP) pathway of IPP biosynthesis are now known (Kamikawa et al. 2017; Graupner et al. 2018; Dorrell et al. 2019). Thus, there is generally a metabolic reason for a plastid retention, although the cases of plastid dependency differ between lineages.
An interesting group to study non-photosynthetic plastids are the euglenophytes. Like their prime representative Euglena gracilis, most euglenophytes are mixotrophs containing a complex three-membranes-bound photosynthetic plastid derived from a green alga belonging to Pyramimonadales (Turmel et al. 2009; Leander et al. 2017; Jackson et al. 2018). However, non-photosynthetic mutants of E. gracilis, induced by an antibiotic or mutagenic treatment, are often capable of heterotrophic living even after presumed plastid loss (reviewed in Krajčovič et al. 2002; Hadariová et al. 2018). This might be enabled by metabolic independence of the E. gracilis cell on the plastid. For instance, E. gracilis was shown to possess two parallel haem synthesis pathways, one located in the mitochondrion/cytosol and another in the plastid (Weinstein and Beale 1983). In the long run, this redundancy might predestine one of the pathways for loss, provided that the other can efficiently supply the end-product to all compartments requiring it (Kořený and Oborník 2011; Cihlář et al. 2016). On the other hand, localization of an essential pathway or its part exclusively into the plastid will impose essentiality of the organelle as such.
Several lineages of euglenophytes independently became secondarily heterotrophic, but evidence for the presence of a plastid has been provided only for Euglena longa (originally called Astasia longa), a close relative of E. gracilis (Marin et al. 2003; Nudelman et al. 2003). The documentation of the organelle at the cytological level is spurious (Webster et al. 1967; Kivic and Vesk 1974; Hachtel 1996), but the complete plastid genome sequence was reported nearly two decades ago (Gockel and Hachtel 2000). As expected, it lacks all the photosynthesis-related genes, except for rbcL encoding the large subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO). We recently documented the existence of a nucleus-encoded small subunit (RBCS) of the E. longa RuBisCO enzyme synthesized as a precursor polyprotein, but its processing into monomers could not be demonstrated (Záhonová et al. 2016). The physiological role of the E. longa RuBisCO and the whole plastid remains unknown, although indirect evidence suggests that the plastid is essential for E. longa survival (Siemeister et al. 1990a, b; Gockel et al. 1994; Hadariová et al. 2017).
To provide a key resource for investing the biology of E. longa and its plastid, we recently generated a transcriptome assembly for this species and demonstrated that it is complete and directly comparable to the transcriptome assemblies reported for E. gracilis (Záhonová et al. 2018). Evaluation of a set of high-confidence candidates for plastid-targeted proteins enabled us to conclude that nucleus-encoded plastidial proteins in E. longa employ N-terminal targeting presequences of the same two characteristic classes, as known from E. gracilis. The E. longa transcriptome revealed various unusual features of the plastid biogenesis and maintenance machinery shared with photosynthetic euglenophytes, but also supported the lack of the photosynthesis-related machinery and suggested specific reductions of several plastidial house-keeping functions presumably reflecting the loss of photosynthesis (Záhonová et al. 2018). However, the repertoire of anabolic and catabolic pathways localized to the E. longa colourless plastid has not been investigated and is the subject of the present paper.
To chart the main paths of the metabolic map of the E. longa plastids, we searched for homologs of enzymes underpinning pathways known from plastids of other species. The reconstruction was greatly facilitated by the recent characterization of the E. gracilis plastid metabolic network based on a proteomic analysis of the organelle (Novák Vanclová et al. 2019). N-terminal regions of the candidates were evaluated for characteristics of presequences predicting a specific subcellular localization to distinguish those likely representing plastid-targeted proteins from enzymes located to other parts of the cell. Some of the bioinformatics predictions were further tested by biochemical analyses. Our study provides the first comprehensive view of a non-photosynthetic secondary plastid of green-algal origin and shows that the metabolic capacity of the E. longa plastid is strikingly different from those of the apicoplast and other relic plastids characterized in sufficient detail.
Results and Discussion
E. longa plastid lacks the MEP pathway of IPP biosynthesis, yet has kept the production of tocopherol and a phylloquinone derivative
Two parallel pathways of IPP biosynthesis exist in E. gracilis (Kim et al. 2004): the mevalonate (MVA) pathway localized to the mitochondrion (first three enzymes) and the cytosol (the rest), and the plastid-localized 2-C-methyl-D-erythritol (MEP) pathway, but only enzymes of the MVA pathway were found in E. longa (Table S1, Fig. 1A). E. longa thus joins the group of recently discovered plastid-bearing eukaryotes lacking the MEP pathway, namely the colourless diatom Nitzschia sp. NIES-3581 (Kamikawa et al. 2017) and various colourless chrysophytes (Graupner et al. 2018; Dorrell et al. 2019). In contrast, the plastid-localized MEP pathway in apicomplexans and related alveolates (i.e. Myzozoa) is essential as a source of precursors for the synthesis of all cellular isoprenoids, since the cytosolic MVA pathway was lost from this group (Janouškovec et al. 2015; Waller et al. 2016). The retention of the MEP pathway in the colourless plastids of diverse non-photosynthetic chlorophytes (Figueroa-Martinez et al. 2015) is similarly explained by the loss of the MVA pathway in this group (Lohr et al. 2012).
A: Schematic comparison of the localization and evolutionary origin of enzymes (see colour-coding graphical legend below the “cells”). Abbreviations, IPP synthesis: ACAT – acetyl-CoA acetyltransferase, CDP-ME – 4-(cytidine 5’-diphospho)-2-C-methyl-D-erythritol, CDP-MEP – 2-phospho-CDP-ME, CMK – CDP-ME kinase, CMS – CDP-ME synthase, DMAPP – dimethylallyl diphosphate, DXP – 1-deoxy-D-xylulose 5-phosphate, DXR – DXP reductase, DXS – DXP synthase, FPP – farnesyl siphosphate synthase, GGPS – geranylgeranyl-diphosphate synthase, HDR – HMB-PP reductase, HDS – HMB-PP synthase, HMB-PP – 4-hydroxy-3-methylbut-2-en-1-yl diphosphate, HMG-CoA – 3-hydroxy-3-methylglutaryl-CoA, HMGCR – HMG-CoA reductase, HMGCS – HMG-CoA synthase, IDI – isopentenyl-diphosphate delta-isomerase, MCS – MEcPP synthase, MDD – mevalonate-diphosphate decarboxylase, MEcPP – 2-C-methyl-D-erythritol 2,4-cyclodiphosphate, MEP – 2-C-methyl-D-erythritol 4-phosphate, MVK – mevalonate kinase, PMVK – phosphomevalonate kinase, PPS – unspecified polyprenyl-diphosphate synthase, ? – unclear substrate; Terpenoid-quinone synthesis: 4HPP – 4-hydroxyphenylpyruvate, HPD – hydroxyphenylpyruvate dioxygenase, HPT – homogentisate phytyltransferase, MMT – MPBQ/MPSQ methyltransferase, TAT – tyrosine aminotransferase, TC – tocopherol cyclase, TMT – tocopherol-O-methyltransferase, VTE5 – phytyl kinase, VTE6 – phytyl-phosphate kinase. B: MS/MS spectrum record of E. longa α-tocopherol and the proposed fragmentation pattern in positive ionization mode (inset). Monoisotopic masses of particular fragments were obtained by simulation in Xcalibur software. C: Semiquantitative comparison of tocopherol species in E. longa, heterotrophically (dark) grown E. gracilis and autotrophic E. gracilis. D and E: MS/MS spectrum record of E. longa 5-hydroxyphylloquinone and the proposed fragmentation pattern in positive ionization mode (inset); semiquantitative comparison of 5-hydroxyphylloquinone.
The MEP pathway in E. gracilis provides precursors for the synthesis in the plastid of terpenoid compounds connected to photosynthesis, namely carotenoids and plastoquinone (Kim et al. 2004; Novák Vanclová et al. 2019). As expected, the respective enzymes are all missing from E. longa. However, we were surprized to find out that the E. longa plastid appears still involved in terpenoid metabolism, specifically in its phytol branch. Photosynthetic eukaryotes generally produce three types of phytol derivatives, tocopherols (vitamin E), phylloquinone (PhQ; vitamin K1) and chlorophyll (its phytyl chain), starting with a common precursor phytyl-PP, which is (directly or indirectly via salvage of phytol liberated by chlorophyll degradation) made by reduction of geranylgeranyl-PP synthesized in the plastid by the MEP pathway (Gutbrod et al. 2019). E. gracilis proved to be unusual not only because of lacking the conventional geranylgeranyl-PP reductase (Novák Vanclová et al. 2019), but also for making phytol from precursors provided by the MVA pathway (Disch et al. 1998; Kim et al. 2004). The route of phytol synthesis from the cytosolic isoprenoids is currently unknown, though it was proposed that phytyl-PP is synthesized in the E. gracilis plastid exclusively by step-wise phosphorylation of phytol by phytol kinase (VTE5) and phytyl phosphate kinase (VTE6), which are enzymes normally employed for recycling phytol from chlorophyll degradation (Novák Vanclová et al. 2019). This scheme is supported by the fact that E. longa has retained VTE5 as well as VTE6, both proteins being highly similar to their E. gracilis orthologs and exhibiting putative plastid targeting presequences (Fig. S1; Table S1). Since E. longa lacks chlorophyll, these two enzymes must have a function independent of phytol recycling.
In addition of its role in chlorophyll synthesis, phytyl-PP is used by E. gracilis to make tocopherols and a PhQ derivative, 5’-monohydroxyphylloquinone (OH-PhQ; Ziegler et al. 1989; Watanabe et al. 2017; Novák Vanclová et al. 2019). All four enzymes mediating synthesis of α-tocopherol from phytyl-PP and homogentisate were identified in E. gracilis and are localized to its plastid (Novák Vanclová et al. 2019). Interestingly, orthologs of all four enzymes are found in E. longa as well, all with a typical plastid-targeting presequence or at least with the N-terminal region being highly similar to the E. gracilis counterpart (Table S1), consistent with their presumed plastidial localization (Fig. 1A). Homogentisate itself is apparently made outside the plastid, most likely in the mitochondrion, as the enzyme responsible for its synthesis (4-hydroxyphenylpyruvate dioxygenase) is not found in the E. gracilis plastid proteome and the respective proteins have a predicted mitochondrial transit peptide in both E. gracilis and E. longa (Table S1). Our analysis thus predicts that like its photosynthetic cousin, E. longa produces α-tocopherol. To test this directly, we analysed extracts from E. longa by means of HPLC-MS/MS. For comparison, we employed samples from E. gracilis grown at two different conditions (in light and in darkness). Tocopherols were detected in both species (Fig. 1B), with α-tocopherol being the dominant form present in equivalent amounts in all three samples (Fig. 1C). The signals of β- and/or γ-tocopherol (indistinguishable by our method) and of δ-tocopherol suggest that tocopherol cyclase, and possibly also tocopherol O-methyltransferase, of both Euglena species can process substrates with or without the 3-methyl group on the benzene ring (Fig. S2).
The synthesis of OH-PhQ in E. gracilis is understood only partially, with only three enzymes of the pathway previously identified at the molecular level: the large multifunctional protein PHYLLO, apparently localized to the cytosol and catalysing the first four steps leading to o-succinylbenzoate; MenA catalysing phytylation of dihydroxynaphthoate localized in the plastid; and MenG (demethylnaphthoquinone methyltransferase), possessing a typical N-terminal plastid-targeting presequence but not directly confirmed as plastidial by proteomics (Novák Vanclová et al. 2019). Strikingly, E. longa expresses homologs of these three E. gracilis proteins, although at low levels resulting in a low RNA-seq read coverage and thus incomplete or inaccurate assembly of two of the respective transcripts (the PHYLLO sequence split into several contigs, the MenA sequence with a frameshift corrected by RT-PCR; see Materials and Methods). Nevertheless, N-terminal parts of both sequences were intact and confirmed the same subcellular localization as in E. gracilis (Fig. 1A, Table S1). In agreement with these insights, OH-PhQ could be detected in an extract from the E. longa culture (Fig. 1D, Fig. S3), although its abundance was smaller compared to that in E. gracilis by an order of magnitude (Fig. 1E).
These findings document another unexpected function of the E. longa plastid, OH-PhQ synthesis, albeit the plastid-associated part of the pathway cannot be presently reconstructed in full detail. One uncertainty concerns the middle steps of the pathway, since like E. gracilis (see Novák Vanclová et al. 2019), E. longa also lack homologs of the conventional enzymes that are responsible for converting o-succinylbenzoate to dihydroxynaphthoate and localized (at least in eukaryotes studied in sufficient detail) in the peroxisome (Cenci et al. 2018). In contrast, tentative evidence was presented for the association of the respective enzyme activities (presumably corresponding to enzymes non-orthologous to the conventional ones) with the plastid envelope in E. gracilis (Seeger and Bentley 1991), raising a possibility of a similar arrangement in E. longa. Secondly, the molecular identity of the putative hydroxylase catalysing the final step of OH-PhQ synthesis is unknown, so its plastidial localization in E. gracilis or E. longa cannot be ascertained. It is, nevertheless, likely given the fact that OH-PhQ is primarily needed in the plastid, at least in photosynthetic eukaryotes (Ziegler et al. 1989). Thirdly, a previously unknown step – reduction of the naphthoquinone ring – was recently demonstrated as a prerequisite for the reaction catalysed by MenG to proceed in plants and cyanobacteria (Fatihi et al. 2015). The respective reductase is well conserved among diverse plant and algal groups as well as cyanobacteria (Cenci et al. 2018), but we did not identify its close homologs in any of the euglenophytes transcriptome assemblies, suggesting that euglenophytes employ an alternative enzyme that yet needs to be characterized.
E. longa seems to be the first eukaryote with a non-photosynthetic plastid documented to have retained the pathways (or final parts thereof) for tocopherols and OH-PhQ synthesis localized in the organelle. The presence of tocopherols in E. longa is, however, not that surprising, as their function is not restricted to photosynthetic tissues in plants and were detected also in bleached (i.e. non-photosynthetic) E. gracilis mutants (Maeda and DellaPenna 2007; Watanabe et al. 2017). As potent lipophilic antioxidants involved especially in membrane protection from damage caused by lipid peroxidation, tocopherols might be used by E. longa as part of its protective mechanisms against reactive oxygen species generated by the action of mitochondria and peroxisomes. The retention of OH-PhQ synthesis in E. longa may seem more puzzling, given the fact that the best-established role of (OH-)PhQ in plants and algae is its functioning as an electron carrier within the photosystem I (Ziegler et al. 1989; Brettel 1997). PhQ was additionally proposed to serve as an electron acceptor in the process of disulfide bond formation in thylakoid lumenal proteins required for proper function of photosystem II (Furt et al. 2010; Karamoko et al. 2011). A homolog of the respective oxidoreductase (LTO1) exists in E. gracilis (Table S1), but not in the transcriptome data from E. longa, consistent with the lack of photosystem II, and perhaps thylakoids entirely in the latter species. Interestingly, PhQ was detected in the plasma membrane in plant tissues and proposed to be involved in photosynthesis-unrelated redox processes at the cell surface (Lochner et al. 2003; Schopfer et al. 2008). PhQ may even be synthesised directly in the plasma membrane, as a recent report from plants has documented the existence of alternative forms of the terminal enzymes of PhQ biosynthesis that result from alternative splicing and are localized to the plasma membrane rather than the plastid (Gu et al. 2018). However, the protein sequences of both MenA and MenG enzymes in E. longa carry typical plastid-targeting presequences, so they are unlikely to operate in the plasma membrane. We thus propose that OH-PhQ in E. longa is involved in a hitherto uncharacterized, photosynthesis-unrelated plastid-resident process.
E. longa plastid plays a limited role in the metabolism of nitrogen-containing compounds
Some of the apparent oddities of the E. longa plastid do not stem from the loss of photosynthesis in this species but reflect unusual features of the plastid in euglenophytes in general. These particularly concern plastid functions in the metabolism of nitrogen-containing compounds. Plastid is commonly involved in nitrogen assimilation due to housing nitrite reductase (Giordano and Raven 2014; Sanz-Luque et al. 2015), but it was established a long time ago that E. gracilis cannot assimilate nitrate or nitrite (Oda et al. 1979; Kitaoka et al. 1989) and, accordingly, no nitrite reductase can be identified in the transcriptome data from this species and E. longa. The plastids of both Euglena species apparently also lack the enzymes working immediately downstream, i.e. glutamine synthetase and glutamine oxoglutarate aminotransferase (the GS/GOGAT system common in plastids of other groups; Fernandez and Galvan 2008; Dagenais-Bellefeuille and Morse 2013), indicating that the plastids rely on the import of organic nitrogen from other parts of the cell, like recently proposed for the plastid in chromerids (Füssy et al. 2019).
One of the most surprising insights of the recent proteomics-aided analysis of the E. gracilis plastid metabolism was the paucity of pathways concerning amino acids (Novák Vanclová et al. 2019). E. longa is apparently even more extreme in this regard, because it lacks counterparts of the (predicted or proteomically verified) plastid-targeted forms of serine biosynthesis enzymes found previously in E. gracilis, i.e. phosphoglycerate dehydrogenase and phosphoserine phosphatase. Thus, we could localize only two elements of amino acid biosynthesis pathways to the E. longa plastid (Fig. S4): serine/glycine hydroxymethyltransferase, whose obvious role is to provide the one-carbon moiety for formylmethionyl-tRNA synthesis required for the plastidial translation; and one of the multiple isoforms of cysteine synthase A, which (like in E. gracilis) apparently relies on O-acetyl-L-serine synthesized outside the plastid, due to the absence of a discernible plastid-localized serine O-acetyltransferase (see Novák Vanclová et al. 2019, and Table S2). This is not due to a general reduction of amino acid metabolism in E. longa or the incompleteness of the sequence data, as its transcriptome assembly includes homologs of enzymes required for the synthesis of all 20 proteinogenic amino acids, but the respective proteins have predicted localization in compartments other than the plastid (Table S2).
Amino acids in the plastid are not only substrates of protein synthesis, but also serve as precursors or nitrogen donors for the synthesis of various other compounds (Moffatt and Ashihara 2002; Gerdes et al. 2012). One such pathway, described in detail in the subsequent section, leads to tetrapyrroles. In contrast, the spectrum of reactions related to the metabolism of other nitrogen-containing cofactors or their precursors (B vitamins) is very limited in the plastids of both Euglena species. We identified only one such candidate in E. longa – the reaction of vitamin B6 salvage catalysed by pyridoxamine 5’-phosphate oxidase, whereas E. gracilis additionally expresses two plastid-targeted isoforms of pyridoxine 4-dehydrogenase (Table S3). De novo synthesis or salvage of purines and pyrimidines is also absent from the plastid of both Euglena species, except for one step present in E. gracilis but apparently not E. longa: the former exhibits two forms of CTP synthase, one presumably cytosolic and another with a plastid-targeting presequence and found in the plastid proteome reported by Novák Vanclová et al. 2019, whereas E. longa expresses an ortholog of only the cytosolic version (Table S3). The lack of a separate plastidial CTP source in E. longa may reflect the presumably lower magnitude of RNA synthesis (and possibly also less extensive phospholipid synthesis requiring CTP in the reaction catalysed by CDP-diacylglycerol pyrophosphatase) in its plastid. Finally, E. longa possesses an ortholog of an enzyme involved in the synthesis of polyamines (spermidine synthase) found in the plastid proteome of E. gracilis (Novák Vanclová et al. 2019), but (in contrast to the E. gracilis protein) its N-terminus does not fit the characteristics of a plastid-targeting presequence, suggesting that its subcellular localization may be outside of the plastid (Fig. S4). However, we found out that both E. longa and E. gracilis have another version of this enzyme with an obvious plastid-targeting-like presequence in both species (Table S3), so we cannot rule out the possibility that polyamines are produced in the E. longa plastid after all.
A residual tetrapyrrole biosynthesis pathway of unclear function is retained in the E. longa plastid
Most plastid-bearing eukaryotes synthesize protoporphyrin IX, the common precursor of haem and chlorophyll, via a single pathway wholly or mostly localized to the plastid (Oborník and Green 2005; Cihlář et al. 2016; Füssy and Oborník 2017). E. gracilis is one of the known exceptions, because it possesses two independent protoporphyrin synthesis pathways, a mitochondrial-cytosolic and a plastid one (Fig. 2A; Weinstein and Beale 1983; Kořený and Oborník 2011; Lakey and Triemer 2016). Whereas all enzymes of the plastid pathway were identified at the sequence level, previous studies left some gaps in the enzyme assignment to the different steps of the mitochondrial-cytosolic pathway (Kořený and Oborník 2011; Lakey and Triemer 2016). The new sequence data from E. gracilis has now enabled us to identify all three missing enzymes of the mitochondrial-cytosolic pathway; specifically novel, apparently cytosolic, isoforms of uroporphyrinogen-III synthase (UROS) and uroporphyrinogen decarboxylase (UROD), and two variants of oxygen-independent coproporphyrinogen oxidase (CPOXi), one cytosolic and one perhaps localized to the mitochondrion. In addition, we found a third UROS homolog that carries a plastid-targeting presequence and constitutes a divergent sister group to the previously known euglenophyte plastidial UROS isoform (Fig. S5; Table S4). Indeed, it was detected in the experimentally determined plastid proteome, together with the previously known plastidial UROS isoform (protein IDs 16898 and 15143 in Novák Vanclová et al. 2019), but the functional significance of the plastidial UROS duplication remains unclear.
A: Schematic comparison of the localization and evolutionary origin of enzymes (see colour-coding graphical legend). Abbreviations: ALAD – delta-aminolevulinate dehydrogenase, ALAS – delta-aminolevulinate synthase, CobA – uroporphyrinogen-III C-methyltransferase, CPOX – coproporphyrinogen III oxidase, CPOXi – oxygen-independent CPOX, CysG – trifunctional enzyme of sirohaem synthesis (see text), FeCH – ferrochelatase, GSA-AT – glutamate semialdehyde-aminomutase, GTR – glutamyl-tRNA reductase, PBGD – porphobilinogen deaminase, PPOX – protoporphyrinogen oxidase, UROD – uroporphyrinogen decarboxylase, UROS – uroporphyrinogen-III synthase. B-C: Sequence alignment of UROD (B) and PPOX (C) presequences, showing the loss of the plastid-targeting motifs in E. longa. These targeting motifs represented by transmembrane domains are marked with frames in other euglenophyte sequences.
Our transcriptome data from E. longa revealed orthologs of all E. gracilis enzymes of the mitochondrial-cytosolic pathway with the same predicted subcellular localization, barring a single enzyme with an incomplete N-terminal sequence precluding confident prediction (Fig. 2A; Table S4). In contrast, orthologs of only six of the E. gracilis plastid-targeted enzymes could be identified in the transcriptome of E. longa: aminolevulinic acid dehydratase (ALAD), the two UROS isoforms, one of the UROD isoforms, one isoform of CPOXi, and protoporphyrinogen oxidase (PPOX). The first three proteins have putative plastid-targeting presequences, much like their E. gracilis orthologs. However, the respective CPOXi protein has a strong mitochondrial targeting presequence, whereas the UROD protein is devoid of any presequence and the presequence of the PPOX protein is markedly shorter and lacks the characteristics of a plastid-targeting signal (Fig. 2B; Table S4). This suggests that only ALAD and the two UROS isoforms are targeted to the plastid in E. longa, whereas the other enzymes have been retargeted to the cytosol or mitochondrion in the evolution of the E. longa lineage. Strikingly, a putative E. longa homolog of the E. gracilis plastid-localized porphobilinogen deaminase (PBGD), an enzyme for the reaction between those catalysed by ALAD and UROS, was previously detected at both the RNA (by a Northern blot) and protein level (by an immunoblot; Shashidhara and Smith 1991), but we could not identify the respective transcript in the E. longa transcriptome data even when the raw reads were searched.
Intact isolated E. gracilis plastids are capable of chlorophyll synthesis from externally supplied 5-aminolevulinic acid (ALA; Gomez-Silva et al. 1985), so it is possible that the E. longa plastid can analogously import ALA from the cytoplasm, thus providing a substrate for the plastid-localized ALAD isoform (Fig. 2A). The surprising absence of the previously documented plastid-localized PBGD in our transcriptome data may be potentially accounted for by the lack of expression of the respective enzyme at culture conditions employed by us. Hence, the pathway may theoretically proceed up to uroporphyrinogen III (the product of the UROS enzyme) in the E. longa plastid, but the absence of enzymes for further processing of this compound towards haem and/or chlorophyll (consistent with the lack of photosystems and the cytochrome b6f complex) indicates that it is consumed by another process.
Uroporphyrinogen III is indeed at the beginning of a separate branch converting it in three steps to sirohaem (Dailey et al. 2017), which serves as a cofactor of several enzymes, including nitrite reductase and sulfite reductase (Tripathy et al. 2010). The former enzyme is missing from the Euglena spp. (see above), but sulfite reductase is present in both E. longa and E. gracilis (Table S4) and was detected in the plastid proteome of the latter species (Novák Vanclová et al. 2019). Hence, sirohaem synthesis in the plastid to support the function of sulfite reductase is an attractive hypothesis that might explain the retention of some of the tetrapyrrole synthesis enzymes in the E. longa plastid. However, analyses of the transcriptome data alone do not provide any clear answer as to whether this hypothesis is valid. Of the various alternative enzymes mediating sirohaem synthesis in various taxa (Tripathy et al. 2010), only two have confidently identified homologs in both Euglena spp. (Table S4). One of them is the multifunctional protein CysG, which catalyses all reactions from uroporphyrinogen III to sirohaem, but the respective Euglena homolog is devoid of any presequence in either species and is thus most likely cytosolic (Table S4). The other is a stand-alone version of the enzyme (uroporphyrinogen-III C-methyltransferase) catalysing the first step of sirohaem synthesis, called CobA or SirA (Table S4). The Euglena homologs exhibit a predicted signal peptide, which is however immediately followed by a region belonging to the mature enzyme, so it is unclear whether this protein is routed to the plastid (indeed it is not among the proteomically confirmed plastid protein in E. gracilis). Hence, the role of the enzymes of the tetrapyrrole biosynthesis pathway in the E. longa plastid remains unexplained.
E. longa plastid does not make fatty acids but maintains phospholipid and glycolipid synthesis
Eukaryotes synthesize even-chain fatty acids by a single large multi-modular fatty acid synthase I (FASI) in the cytosol or by a multi-enzyme type II fatty acid synthesis complex in the plastid. E. gracilis possesses both systems (Zimorski et al. 2017). In contrast, E. longa encodes only a homolog of the FASI enzyme (Table S5), whereas enzymes of the type II fatty acid synthesis are absent (Fig. 3A) except for the acyl carrier protein (ACP). The loss of plastid-localized fatty acid synthesis in E. longa is not without precedent, as it has been also reported for the apicoplast of Theileria parva (which is fully dependent on fatty acid supply from host), the plastid of Perkinsus marinus (Janouškovec et al. 2015) and the chrysophyte “Spumella” sp. NIES-1846 (Dorrell et al. 2019). Nevertheless, the E. longa plastid has kept plastid-targeted enzymatic steps downstream of fatty acid synthesis. These include ACP and 4’-phosphopantetheinyl transferases (or holo-ACP synthase) crucial for the synthesis of an active form of ACP, which serves as a carrier of acyl chains in phospholipid and glycolipid biosynthesis (Lambalot and Walsh 1995). Next, E. longa possesses predicted plastid-targeted homologs of acyl-ACP synthetases (presumably activating fatty acids imported into the plastid from outside) and all enzymes required for the synthesis of phosphatidic acid (PA) and its subsequent conversion to phosphatidylglycerol (PG) (Fig. 3A; Table S5). It is worth noting that E. longa also has a parallel, plastid-independent, route of phosphatidylglycerol synthesis (Table S6).
A: Schematic comparison of the localization and evolutionary origin of enzymes. Abbreviations, fatty acid synthesis: ACC – acetyl-CoA carboxylase, ACS – acetyl-CoA synthetase, ENR – enoyl-CoA reductase, Fas1 – malonyl-CoA/acetyl-CoA: ACP transacylase, FASI – type I fatty acid synthase, FAT – fatty acyl-ACP thioesterase, HD – hydroxyacyl-ACP dehydratase, KAR – ketoacyl-ACP reductase, KAS – ketoacyl-ACP synthase, TE – fatty acid thioesterase, TRX – thioredoxin-regulated enzyme; glycolipid synthesis: AAS – acyl-ACP synthase, ACPS – holo-ACP synthase, AGP-AT – acylglycerophosphate acyltransferase, G3P-AT – glycerol-3-phosphate acyltransferase, G3PDH – glycerol-3-phosphate dehydrogenase, MGDG/DGDG – mono-/digalactosyl diacylglycerol, MGDGS/DGDGS – MGDG/DGDG synthase, PAP – phosphatidic acid phosphatase, SQD1 – UDP-sulfoquinovose synthase, SQD2/SQDX – sulfoquinovosyl diacylglycerol (SQDG) synthase, UGE/PHD1 – UDP-glucose epimerase, UGP3 – UDP-glucose pyrophosphorylase 3; phospholipid synthesis: CDS – CDP-diacylglycerol synthase, PGP1 – phosphatidylglycerophosphate synthase, PGP-P – phosphateidylglycerophosphate phosphatase. B: Semiquantitative comparison of glycolipids present in E. longa and autotrophic E. gracilis. Note the logarithmic scale of the quantification units (peak area). Peak area is an arbitrary unit expressing the intensity of the signal of a particular lipid species, recalculated according to their respective ionization promptitude. As each lipid species have different ionization promptitude, note that direct comparison can be done only within lipid class (for details, see Tomčala et al. 2017). C-E: Immunofluorescence micrographs using anti-DGDG antibody (C), DAPI (D) and differential interference contrast (E). Autotrophic E. gracilis represents a positive control, while the aplastidic euglenozoan R. costata was used as negative control.
No other reactions of phospholipid synthesis or decomposition beyond PG synthesis seem to operate in the E. longa plastid, but interestingly, we found homologs of the enzymes of the synthesis of galactolipids monogalactosyldiacylglycerol (MGDG) and digalactosyldiacylglycerol (DGDG) in the E. longa transcriptome (Fig. 3A, Table S5). All the proteins have a predicted N-terminal plastid-targeting presequence, consistent with the plastidial localization of galactolipid synthesis in all other plastid-bearing eukaryotes studied so far (Yuzawa et al. 2012). In support of the bioinformatic predictions, both MGDG and DGDG could be detected in lipid extracts from E. longa and E. gracilis, although galactolipid levels were significantly lower in E. longa than in the control sample of E. gracilis (Fig. 3B). The presence of DGDG was further confirmed by immunofluorescence using an anti-DGDG antibody, which showed DGDG to be present in small foci in the E. longa cells (white arrowheads in Fig. 3C), presumably representing individual small plastids. In comparison, most of the volume of the photosynthetic E. gracilis cells was stained, whereas the negative control, the primary osmotrophic (i.e. plastid-lacking) euglenoid Rhabdomonas costata, did not stain at all.
The presence of galactolipids in plastids is generally being explained by their essentiality for the proper functioning of the photosynthetic apparatus, but this view has been challenged by demonstration that cyanobacterial mutants lacking galactolipids can photosynthesize normally (Awai et al. 2001). A photosynthesis-independent role of galactolipids in plastid biology was proposed, too. Transit peptides of plastid-targeted proteins exhibit affinities for MGDG and DGDG in the plastid envelope (Pinnaduwage and Bruce 1996), suggesting a direct role of these lipids in plastid protein import. The photosynthesis-independent role of galactolipids is indicated not only by their presence in the E. longa plastid documented here, but also by previous reports from other non-photosynthetic algae, including the diatom Nitzschia alba (Anderson et al. 1978) and the chlorophyte Prototheca wickerhamii (Borza et al. 2005), and from non-photosynthetic tissues of plants (Awai et al. 2001; Kobayashi 2016). On the other hand, the apicoplast (Botté et al. 2008; Botté et al. 2013) and most likely also the relic plastid of Helicosporidium (based on our analysis of the respective genome sequence data generated by Pombert et al. 2014) lack galactolipid synthesis completely. The reason for the differential retention of galactolipids in different colourless plastids remains unclear.
In addition to MGDG and DGDG, we identified in samples from both Euglena spp. another common glycolipid characteristic for plastids, sulfoquinovosyldiacylglycerol (SQDG) (Fig. 3B) (Hori et al. 2016). The presence of SQDG in E. longa is also not unprecedented among non-photosynthetic plastid-bearing eukaryotes; see, e.g., its documented occurrence in the diatom N. alba or the dinoflagellate Oxyrrhis marina (Anderson et al. 1978; Goddard-Borger and Williams 2017; Yoon et al. 2017). Of the three enzymes of SQDG biosynthesis normally localized to the plastid, we found in both Euglena species only that catalysing the final step (sulfoquinovosyltransferase; Fig. 3A). Interestingly, the standard eukaryotic version of the enzyme, SQD2, is present only in E. gracilis, but both species proved to share another isoform phylogenetically affiliated to bacterial SqdX version of the enzyme (Fig. 4). To the best of our knowledge, this is the first encounter of SqdX in any eukaryote. The presence of SQD2 only in E. gracilis probably relates to specific needs of its photosynthetic plastid. Indeed, E. gracilis contains a much larger amount of SQDG compared to E. longa (Fig. 3B) and the profile of esterified fatty acids differs between the two species (E. longa lacks SQDG forms with unsaturated longer chains; Table S7).
The maximum-likelihood tree was inferred with IQ-TREE using the LG+F+G4 substitution model and ultra-fast bootstrapping. The UFboot support values are indicated at branches when higher than 75%.
Details of the synthesis of the saccharide moieties of glycolipids in E. longa are also worth considering (Fig. 3A). E. longa exhibits an ortholog of the E. gracilis UDP-glucose epimerase previously identified in the plastid proteome (Table S5), explaining the source of UDP-galactose for galactolipid synthesis. This enzyme seems to have been acquired by euglenophytes from a bacterial source (Fig. S6), but, interestingly, E. gracilis encodes another enzyme, which corresponds to the plastidial UDP-glucose epimerase, also called PHD1, known from plants and various algae (Li et al. 2011). The E. gracilis PHD1 possesses a predicted plastid-targeting presequence (Table S5) and is thus also likely plastidial (although this is not confirmed by proteomic data reported by Novák Vanclová et al. 2019). This putative redundancy in UDP-galactose is apparently not shared by E. longa, possibly because of a presumably much lower need for galactolipid synthesis (Fig. 3B). The origin of the SQDG precursor UDP-sulfoquinovose in E. longa remains obscure, because as noticed before (Novák Vanclová et al. 2019), euglenophytes in general lack the conventional UDP-sulfoquinovose synthase SQD1/SqdB and probably employ an alternative, unrelated enzyme. UDP-glucose, i.e. the common precursor of both UDP-galactose and UDP-sulfoquinovose, is most likely imported from the cytoplasm, owing to the absence in E. longa of a candidate plastid-targeted version of any of the relevant enzymes (UDP-glucose/UDP-sugar pyrophosphorylase).
A linearized Calvin-Benson pathway in the E. longa plastid
The presence of genes encoding both subunits of the enzyme RuBisCO in E. longa (Záhonová et al. 2016) raises the question whether the Calvin-Benson cycle (CBC) as a whole has been preserved in this organism to provide ribulose-1,5-bisphosphate, the RuBisCO substrate. A putative E. longa plastid triose-phosphate isomerase was described previously (Sun et al. 2008). We confirmed not only this candidate, but additionally identified homologs with putative plastid-targeting presequences for almost all remaining CBC enzymes (Table S8). Our identification was confirmed by phylogenetic analyses (supplementary Dataset S1), which showed specific relationships of the E. longa proteins to the previously characterized CBC enzymes from other euglenophytes (Markunas and Triemer 2016). However, two key CBC enzymes are apparently missing from the E. longa transcriptome: phosphoglycerate kinase (ptPGK) and glyceraldehyde-phosphate dehydrogenase (ptGAPDH). The homologs of these enzymes present in E. longa are not orthologous to the plastid-targeted isoenzymes from other euglenophytes and all clearly lack a plastid-targeting presequence (Table S8). Hence, these are presumably cytosolic enzymes involved in glycolysis/gluconeogenesis. The lack of ptPGK and ptGAPDH in E. longa means that the product of the RuBisCO carboxylase activity, 3-phosphoglycerate (3PG), cannot be converted (via 2,3-bisphosphoglycerate; 2,3-BPG) to glyceraldehyde-3-phosphate (GA3P) in the plastid and the cycle becomes a linear pathway (Fig. 5).
The Calvin-Benson cycle (CBC) resident to this organelle is central to the plastid carbon metabolism, regulated by the ferredoxin/thioredoxin (Fd/Trx) system. Reduction of disulfide bonds by the Fd/Trx system activates FBP and PRK. FTR and FD of the Fd/Trx system require for their function a post-translationally added Fe-S prosthetic group provided by the Fe-S assembly system. GapN apparently mediates shuttling of reducing equivalent (NADPH) through the exchange of DHAP/GA3P and 3PG, reflecting the cytosolic NADPH/NADP+ ratio and thus an overall metabolic state of the cell. In contrast, E. gracilis plastid is an energy-converting organelle, harvesting light into chemical energy bound as NADPH and ATP and subsequently using this bound energy to fix CO2 into organic carbohydrates via the CBC. Enzyme abbreviations are colour-coded according to their inferred evolutionary origin, see the graphical legend. Abbreviations, CBC: ALDO – aldolase, DHAP – dihydroxyacetone-phosphate, FBP – fructose-1,6-bisphosphatase, GAPDH – glyceraldehyde-3-phosphate dehydrogenase, PGK – 3-phosphoglygerate kinase, PGP – phosphoglycolate phosphatase, PLGG1 – plastid glycolate/glycerate transporter, PRK – phosphoribulokinase, RBCL/RBCS – RuBisCO large/small subunit, RCA – RuBisCO activase, RPE – ribulose-5-phosphate epimerase; RPIA – ribulose-phosphate isomerase A, SBP – sedoheptulose-1,7-bisphosphatase, TKTL – transketolase, TPI – triose-phosphate isomerase, TPT – triose-phosphate translocator; Fd/Trx system: FD – ferredoxin; FNR – FD/NADP+ oxidoreductase, FTR – FD/TRX oxidoreductase, TRX – thioredoxin, ATPS – ATP synthase, ATPC – ADP/ATP translocase, LHC – light-harvesting complex.
Assuming that the reactions catalyzed by fructose bisphosphatase, phosphoribulokinase, and RuBisCO are not reversible (Raines and Lloyd 2001), the flux through this linearized CB pathway most likely goes from GA3P to 3PG, with a net production of six molecules of 3PG from five molecules of GA3P due to fixation of three CO2 molecules catalysed by RuBisCO. We thus need to define the origin of GA3P entering the pathway. Euglenophytes do not store starch in the plastid (instead they have cytosolic paramylon as a storage polysaccharide; Kiss et al. 1987), and indeed we did not find any glucose metabolism-related enzymes in the predicted E. longa plastid proteome. Hence, GA3P cannot be produced by a glycolytic route in the E. longa plastid. The presence of the plastid-targeted glycerophosphate dehydrogenase (Table S5) in principle allows for generation of GA3P from glycerol-3-phosphate (via dihydroxyacetone phosphate; DHAP; Fig. 3), which could possibly come from the degradation of glycerolipids in the plastid. However, no phospholipid-degradation enzymes (phospholipases) appear to localize to the plastid in E. longa. Hence, the primary function of glycerophosphate dehydrogenase is perhaps to operate in the reverse direction, i.e. to provide glycerol-3-phosphate for the plastid phospholipid and glycolipid synthesis (see above). The E. longa plastid thus most likely imports GA3P or DHAP from the cytosol (Fig. 5). This assumption is supported by the presence of several members of the plastid phosphate translocator (pPT) family (Fig. S7; Facchinelli and Weber 2011), including one phylogenetically closest to a previously characterized cryptophyte transporter with a preference for DHAP (Haferkamp et al. 2006). Concerning the opposite end of the hypothesized linear CB pathway, we did not identify any candidate E. longa plastid-targeted enzyme that would metabolize 3PG further (see the absence of 3PG dehydrogenase discussed above), so this intermediate is most likely exported from the plastid into the cytosol, probably also by one of the members of the pPT family of transporters (Fig. 5).
The operation of CBC is inherently linked with the oxygenase side-activity of RuBisCO, which converts ribulose-1,5-bisphosphate into 3PG and phosphoglycolate instead of two molecules of 3PG that are produced by the regular carboxylase activity (Tabita et al. 2007). Phosphoglycolate is metabolized in the photorespiration pathway, initiated by phosphoglycolate phosphatase yielding glycolate. We did find a candidate plastid-targeted phosphoglycolate phosphatase in E. longa (Table S8), orthologous to a protein detected in E. gracilis plastid proteome, suggesting that the E. longa RuBisCO has the oxygenase activity and phosphoglycolate is metabolized in this species. However, we did not find in our E. longa transcriptome assembly any discernible homolog of the recently characterized transporter PLGG1 mediating glycolate export from the plastid in the canonical photorespiratory pathway (Pick et al. 2013), although E. gracilis does have it (Table S8). Since there is no obvious candidate for a plastid-targeted glycolate-metabolizing enzyme (glycolate oxidase, glyoxylate reductase, glycolaldehyde dehydrogenase) in E. longa, it is unclear how glycolate is removed from the plastid of this species. It is possible that the amount of glycolate produced in the E. longa plastid is low and that it can be exported by an alternative (PLGG1-independent) route, whose existence has been proposed also for plant plastids (Walker et al. 2016) and which might be sufficient for glycolate recycling in the semi-parasitic plant Cuscuta campestris capable of low-efficiency photosynthesis (Vogel et al. 2018).
The system of redox regulation of the Calvin-Benson pathway is conserved in E. longa
Although the photosynthetic machinery, including all photosystem I subunits, is missing from E. longa (Záhonová et al. 2018), we found homologs (with clear plastidial localization) of the typical “photosynthetic” (PetF-related) ferredoxin (Fd) and ferredoxin-NADP+ reductase (FNR) (Table S9). These two proteins are primarily involved in passing electrons from an activated photosystem I to NADP+, which is thus reduced to NADPH, the main donor of reducing equivalents for anabolic reactions in the plastid (above all, CO2 fixation). Phylogenetic analysis of FNR homologs from euglenophytes revealed the existence of two different, yet related, clades affiliated to FNR from green algae (Fig. 6). One clade comprises the E. longa FNR plus its orthologs from photosynthetic euglenophytes (E. gracilis and two Eutreptiella strains), whereas the second clade is restricted to the photosynthetic species. Two different FNR forms also exist in plants, one functioning in photosynthesis (production of NADPH dependent on the function of the photosystem I; marked with “P” in Fig. 6) and the other being a “non-photosynthetic” homolog (“NP” in Fig. 6) that allows the electron flow in the reverse direction, from NADPH to Fd (Vollmer et al. 2001). In analogy with plants, we suggest that the two euglenophyte FNR clades functionally differ, with one (that lacking a representative in E. longa) serving in photosynthesis and the other (present in E. longa) mediating light-independent production of reduced Fd.
The maximum-likelihood tree was inferred with IQ-TREE using the LG+F+G4 substitution model and ultra-fast bootstrapping. The UFboot support values are indicated at branches when higher than 75%. Euglenophyte species are in bold, and their putative photosynthetic and non-photosynthetic homologs are depicted. P, photosynthetic; NP, non-photosynthetic.
Multiple plastid enzymes depend on reduced Fd as an electron donor, namely glutamate synthase, lipid desaturases, nitrite reductase, and sulfite reductase (Neuhaus and Emes 2000). As discussed in previous sections, glutamate synthase and nitrite reductase are missing from E. longa, whereas all identified lipid desaturases are predicted as targeted to the mitochondrion or the ER (Table S5). A two-subunit sulfite reductase seems to be present in the plastid of E. gracilis (Novák Vanclová et al. 2019) and both subunits have highly similar homologs in E. longa (Table S4), but this form of the enzyme utilizes NADPH rather than ferredoxin as electron donor (Patron et al. 2008).
Another crucial role of Fd in plastids is to provide electrons to ferredoxin:thioredoxin reductase (FTR) mediating reduction of the protein thioredoxin (Trx). The Fd/Trx system regulates several plastid CBC enzymes in response to the redox status in the stroma and, in extension, to the photosynthetic activity of the plastid (Fig. 5). An excess of NADPH leads to electrons being relayed from reduced ferredoxin to the Fd/Trx system, which eventually reduces certain disulfide bonds in the target enzymes, thus changing their activity (Schürmann and Buchanan 2008). This ensures activation of the CBC only when the photosynthetic machinery works properly. Notably, FTR and Trx homologs with evident plastid-targeting presequences are both present in E. longa (Table S9). Specific motifs necessary for the function of the Fd/Trx system (Schürmann and Buchanan 2008) are conserved in the respective E. longa proteins (Fig. S8), consistent with the Fd/Trx system being functional in this species. In addition, two E. longa CB pathway enzymes, fructose bisphosphatase (two of the three isoforms present) and phosphoribulokinase, exhibit the conserved Trx regulatory cysteine motifs similar to their orthologs in E. gracilis (Fig. S8, Table S10). Thus, we suggest that the CB pathway in E. longa is sensitive to the redox status in the plastid, specifically to the concentration of NADPH (Fig. 5).
May the E. longa plastid be involved in keeping the redox balance of the cell?
How to interpret in functional terms the retention of the linearized CB pathway and its putative redox regulation in the non-photosynthetic plastid of E. longa? The key role of the pathway is supported by the fact that production of the large RuBisCO subunit seems to be the raison d’être for the preservation of the plastid genome in E. longa (Záhonová et al. 2016). The presence of CB enzymes in a non-photosynthetic plastid is not without precedent, as it has been reported from a set of unrelated colourless algae and plants. Some of them, e.g. the dinoflagellate Crypthecodinium cohnii, the dictyochophytes Pteridomonas danica and Ciliophrys infusionum, the cryptophyte Cryptomonas paramecium, and some parasitic or mycoheterotrophic land plants, are known to encode RuBisCO (Sekiguchi et al. 2002; Sanchez-Puerta et al. 2007; Donaher et al. 2009; Wicke et al. 2013; Hadariová et al. 2018), but how complete is the complement of other CBC enzymes in these species is unknown. In contrast, transcriptomic or genomic analyses of other colourless plastid-bearing taxa, such as the dinoflagellate Pfiesteria piscicida, the green alga Helicosporidium sp. ATCC50920, the diatom Nitzschia sp. NIES-3581, and the non-photosynthetic chrysophytes, revealed the presence of a subset of CB enzymes, including ptPGK and ptGAPDH, but not of RuBisCO (Kim et al. 2013; Pombert et al. 2014; Kamikawa et al. 2017; Graupner et al. 2018). Hence, the constellation of the CB enzymes retained in the E. longa plastid seems to be unique.
The CBC enzymes retained in various non-photosynthetic eukaryotes obviously do not serve to sustain autotrophic growth, as ATP and the reducing power generated by photosynthesis are unavailable. The CB pathway in Nitzschia sp. NIES-3581 was proposed to serve as a source of erythrose-4-P for the synthesis of aromatic amino acid via the shikimate pathway in the plastid (Kamikawa et al. 2017). Although not discussed in the respective report (Pombert et al. 2014), the CB pathway in the Helicosporidium plastid may likewise serve to feed the co-localized shikimate pathway with erythrose-4-P. However, such a rationalization of the CB pathway in the E. longa plastid would not work, since enzymes for aromatic amino acid biosynthesis in this species are apparently localized to the cytosol (Table S2) and thus have access to erythrose-4-P produced by the pentose phosphate pathway. In addition, the need to produce erythrose-4-P in the E. longa plastid would not explain the retention of RuBisCO (absent in both Nitzschia and Helicosporidium). A photosynthesis- and CBC-independent role of RuBisCO was described in oil formation in developing seeds of Brassica napus, where refixation of CO2 released during carbohydrate-to-fatty acid conversion increases carbon use efficiency (Schwender et al. 2004). A similar explanation is unlikely to hold for RuBisCO retention in E. longa, given the lack of fatty acid synthesis in its plastid and the apparently much smaller significance of oil as a reserve substance in E. longa compared to B. napus.
We believe that the identification of the putatively functional Fd/Trx system, despite the absence of the photosynthetic electron transport chain in this species, provides one of the key hints to understanding the physiological role of the linear CB pathway in the E. longa plastid. The second potentially important piece of the puzzle is provided by the proteomic data from E. gracilis (Novák Vanclová et al. 2019), which indicated the presence of a unique form of GAPDH, the so-called non-phosphorylating GAPDH also referred to as GapN, in the plastid (Table S8). This enzyme uses NADP+ to directly oxidize GA3P to 3PG, skipping the intermediate 2,3-BPG (and hence not leading to ATP generation) and producing NADPH rather than NADH (Iddar et al. 2003). In plants, this enzyme is localized to the cytosol and is involved in shuttling of reducing equivalents from the plastid by the exchange of GA3P and 3PG between the two compartments (Rius et al. 2006). E. longa possesses a GapN homolog highly similar to the E. gracilis protein, including an N-terminal presequence (Table S8), consistent with its presumed plastidial localization. It thus appears that in Euglena spp. GapN mediates shuttling of reducing equivalents in the opposite direction than in plants, i.e. from the cytosol to the plastid (Fig. 5). In case of E. longa this may be the main (if not the only) mechanism of providing NADPH for the use in the plastid, whereas E. gracilis would utilize it when photosynthetic production of NADPH is shut down. At the same time, the shuttle provides a mechanism of linking the level of NADPH in the plastid with the cytosolic concentration of GA3P.
Taken together, we propose that in E. longa (at specific circumstances possibly also in E. gracilis) the plastidial NADPH/NADP+ ratio is directly influenced by the redox status of the cell, i.e. that it rises in an excess of reducing power that slows down the glycolytic oxidation of GA3P in the cytosol. This stimulates the linear CB pathway via the Fd/Trx system, effectively decreasing the level of GA3 by converting it to 3PG without further increasing the reducing power in the cell. This conclusion is apparent from considering the overall stoichiometries of the two alternative pathways from GA3 to 3PG:
The key difference is that the CB pathway does not produce NADH that needs to be reoxidized to keep the glycolytic pathway running, since the fixed CO2 effectively serves as an electron acceptor. Hence, turning the CB bypass on may help the cell keep the redox balance when reoxidation of NADH is not efficient, e.g. at hypoxic (or anoxic) conditions (although this happens at the expense of ATP). Indeed, euglenophytes in their natural settings are probably often exposed to the shortage of oxygen, and anaerobiosis in E. gracilis has been studied to some extent (Tucci et al. 2010; Zimorski et al. 2017). The anaerobic heterotrophic metabolism of E. gracilis relies on fermentative degradation of the reserve polysaccharide paramylon leading to production of wax esters (Yoshida et al. 2016). It is likely that E. longa exhibits a similar metabolic adaptation to low levels of ambient oxygen as E. gracilis. However, details of the euglenophyte anaerobic metabolism need to be worked out yet, and we propose that the plastid may be involved in it as a kind of a redox valve. Work is ongoing to test this hypothesis and to illuminate further details of physiological role of the linear CB pathway in the E. longa plastid.
Conclusions
Endosymbiotic organelles have proven to be extremely evolutionarily versatile. One of the manifestations is the recurrent loss of key metabolic functions of the canonical forms of both mitochondria and plastids, resulting in anaerobic mitochondria-related organelles (such as hydrogenosomes and mitosomes) and non-photosynthetic plastids distributed across diverse eukaryotic branches. A lot of attention has been paid to various mitochondrial derivatives and it is now well documented that they vary substantially in the complement of functional pathways they have retained (Roger et al. 2017). The variation in functional profiles of non-photosynthetic plastids is less well known, as the only example studied in detail is the apicoplast of apicomplexan parasites. Nevertheless, a picture is emerging that independently evolved colourless plastids may also exhibit a surprising degree of diversity in terms of their metabolic capacity.
Our analyses of the E. longa plastid stretch the breadth of variation among non-photosynthetic plastids even further. The combination of pathways present (tocopherol and phylloquinone synthesis, glycolipid synthesis and a linearized CB pathway including RuBisCO), absent (fatty acid, amino acid, and isoprenoid precursor synthesis), and potentially residual (tetrapyrrole synthesis) makes the E. longa plastid unlike any of the previously investigated non-photosynthetic plastids, including the apicoplast. However, further work, combining additional in silico analyses (aimed, e.g., at potential plastid membrane transporters mediating metabolite exchange with the cytosol) with biochemical and cytological investigations is needed to achieve a more precise idea about the protein composition of the E. longa plastid and a better understanding of its physiological roles.
Materials and Methods
Identification and annotation of plastid-targeted proteins
The analyses reported in this study were done using the E. longa transcriptome assembly reported previously (Záhonová et al. 2018). Protein models for annotation were generated by a custom Geneious 8.1.6 (Kearse et al. 2012) script that extracted all open reading frames longer than 297 bp, translated the sequences and then filtered the protein models by a local BLAST+ ver.2.2.30 (Altschul et al. 1997) search against the Swiss-Prot database (version 10/5/15, max E-value=10). Transcript models containing at least a partial spliced leader sequence (TTTTTCG) at their 5’-end or 3’-end (within the first or last 35 nt) were translated in the forward or reverse direction only, respectively.
Candidates for plastid-targeted proteins were identified using criteria for prediction of plastid-targeting presequences described in detail by Záhonová et al. (2018). In the first step protein sequences were gathered that fulfilled at least one of the following requirements: (i) the signal peptide was predicted by the PrediSi v.2004 (Hiller et al. 2004) or PredSL v.2005 (Petsalaki et al. 2006) standalone programs; (ii) one or two transmembrane domains at the N-terminus of the protein were predicted by standalone TMHMM 2.0c (online version where graphical output was considered) (Krogh et al. 2001). The resulting set was then filtered by checking for the presence of a plastid transit peptide, which was predicted by standalone MultiLoc2.5 (Blum et al. 2009) after an in silico removal of the signal peptide or the first transmembrane domain. Finally, protein models with a putative plastid-targeting presequence were automatically annotated using InterProScan 5.21 (Jones et al. 2014) and the annotations were manually scrutinized to identify proteins with an assignable specific metabolism-related function (enzymes of the metabolism of nucleic acids and proteins were ignored in this study).
In parallel we searched with BLAST v.2.2.30 (including tBLASTn against the transcriptome assembly and, in special cases, even against the whole set of raw RNAseq reads) for putative plastid proteins by direct identification and evaluation of homologs of enzymes of specific biochemical pathways potentially localized to the plastid; as queries we used respective protein sequences from E. gracilis (as identified by Novák Vanclová et al. 2019), reference sequences from the KEGG PATHWAY Database (https://www.genome.jp/kegg/pathway.html), or sequences identified by literature searches. In some cases, a more sensitive homology detection algorithm HMMER 3.0 (Mistry et al. 2013) to identify homologs of poorly conserved enzymes (e.g., UROS). For comparative purposes we used the same approach to identify plastid-targeted proteins encoded by the transcriptome assemblies from E. gracilis reported by (Yoshida et al. 2016) (accession GDJR00000000.1) and (Ebenezer et al. 2017) (accession GEFR00000000.1). Where accessions are given for E. gracilis sequences, those starting with GDJR and GEFR belong to the former and latter dataset, respectively.
For MenA cDNA resequencing, RNA was isolated using TRI Reagent (Thermo Fisher Scientific, San Jose, USA) and mRNA was then extracted using the Dynabeads mRNA Purification kit (Thermo Fisher Scientific). Reverse-transcription was performed with random hexamers and StrataScript III Reverse Transcriptase (Thermo Fisher Scientific). For cDNA amplification, forward 5’-GGTGCTGTTCTGCTCTCACT-3’ and reverse 5’-CAGTGGGGATCAGAGATGCG-3’ primers, and the Q5 High-Fidelity DNA polymerase in a standard buffer (New England Biolabs) were used. Amplicons were purified on MinElute PCR Purification columns (Qiagen, Hilden, Germany) and sequenced at the GATC sequencing facility (Konstanz, Germany). The MenA cDNA sequence is deposited in GenBank with the accession number MK484704.
Phylogenetic analyses
Phylogenetic analyses were employed to establish orthologous relationships among E. longa and E. gracilis genes or to illuminate the origin of the euglenophyte proteins of special interest. Homologs of target proteins were identified by BLAST v.2.2.30 searches in the non-redundant protein sequence database at NCBI (www.ncbi.nlm.nih.gov) and among protein models of selected organisms from JGI (Joint Genome Institute, jgi.doe.gov) and MMETSP (Marine Microbial Eukaryote Transcriptome Sequencing Project, marinemicroeukaryotes.org; Keeling et al. 2014). Sequences were aligned using the MAFFT v7.407 (Multiple Alignment using Fast Fourier Transform) tool with L-INS-I setting (Katoh and Standley 2013) and poorly aligned positions were eliminated by the trimAL v1.4.rev22 tool with “-automated1” trimming (Capella-Gutierrez et al. 2009). For presentation purposes, alignments were processed using the CHROMA software (Goodstadt and Ponting 2001). Maximum likelihood (ML) trees were inferred from the alignments using the LG+F+G4 model of IQ-TREE v1.6.9 (Nguyen et al. 2015), and employing the strategy of rapid bootstrapping followed by a “thorough” ML search with 1,000 bootstrap replicates (-m TEST -bb 1000). The list of species, and the number of sequences and amino acid positions are present in Tables S11-S22 for each phylogenetic tree.
Culture conditions
Euglena gracilis strain Z (“autotrophic” conditions) were cultivated statically under constant illumination at 26 °C in Cramer-Myers medium supplemented with ethanol (0.8% v/v) as a carbon source (Cramer and Myers 1952). E. longa strain CCAP 1204-17a (a gift from W. Hachtel, Bonn, Germany) and heterotrophic E. gracilis strain Z were cultivated in identical medium without illumination. The cultures of E. longa were not completely axenic, but the contaminating bacteria were kept at as low level as possible. Rhabdomonas costata strain PANT2 was provided by Vladimír Hampl (Department of Parasitology, Faculty of Science, Charles University in Prague, Czech Republic). It was isolated from a freshwater body in Pantanal (Brasil) and grown with an uncharacterised mixture of bacteria in Sonneborn’s Paramecium medium (pH 7.4; Sonneborn 1950) at room temperature (inoculated every three to four weeks).
Mass spectrometry of structural lipids and terpenoids
For analysis of structural lipids, extracts from E. longa and autotrophic E. gracilis cellular pellets (four biological samples of different culture age) were obtained with chloroform and methanol solution (ratio – 2:1) following the method of (Folch et al. 1957) as modified by (Koštál and Šimek 1998). Samples were homogenized in extraction solution with glass beads using TissueLyser LT mill (Qiagen). Homogenates were dried, weighted, and resolved in 500 µl of chloroform and methanol (1:2) with internal standard PC 17:0/17:0 (Sigma Aldrich). Aliquots from each sample extract were used for lipid determination by HPLC using a liquid chromatograph and autosampler Accela (Thermo Fisher Scientific). The samples (5 µL) were injected and separated on the Gemini column 250 × 2 mm; i.d. 3 µm (Phenomenex, Torrance, USA). A linear ion trap LTQ-XL mass spectrometer (Thermo Fisher Scientific) was used in both positive and negative ion ESI mode. The settings of the system followed the methodology published earlier (Tomčala et al. 2017). Data were acquired and processed using Xcalibur software version 2.1 (Thermo Fisher Scientific). Particular compounds were determined based on m/z value, retention time, behaviour in positive and negative ionization mode, and characteristic fragmentation pattern of target molecules (for details see Tomčala et al. 2017).
Terpenoids were extracted from an autotrophic culture of E. gracilis, a heterotrophic culture of E. gracilis, and a culture of E. longa of the same age in three repetitions. The same extraction procedure as for lipid analysis was used. Sample aliquots were injected into the high-resolution mass spectrometry system powered by Orbitrap Q-Exactive Plus with Dionex Ultimate 3000 XRS pump and Dionex Ultimate 3000 XRS Open autosampler (all by Thermo Fisher Scientific) and followed the settings described in (Tomčala et al. 2017). Data were acquired and processed using Xcalibur software version 2.1. Identification of OH-PhQ was achieved by considering the m/z value, fragmentation pattern, and high-resolution data. Tocopherols (α, β/γ, and δ) were determined by the same characteristic as OH-PhQ and results were then compared with commercially purchased standards (α-tocopherol from Sigma Aldrich, the other three variants from SUPELCO).
Immunofluorescence assay
Immunofluorescence was performed as described previously (Botté et al. 2011). Briefly, cells were fixed in 4% paraformaldehyde for 30 minutes, cellular membranes were permeabilized for 10 minutes on ice with 0.1% non-ionic detergent Igepal CA-630 (Sigma-Aldrich) in PHEM buffer pH 6.9 (60 mM PIPES, 25 mM HEPES, 10 mM EGTA, 2 mM MgCl2), and background was masked with 3% BSA in PHEM buffer. DGDG was detected using a polyclonal rabbit anti-DGDG antibody (1:25) that was a kind gift from Cyrille Y. Botté (ApicoLipid Team, Laboratoire Adaptation et Pathogenie des Microorganismes, University of Grenoble I, France), followed by incubation with a secondary Cy3-labeled polyclonal goat anti-rabbit antibody (AP132C, 1:800, Merck Millipore, Billerica, USA). To improve the fluorescence lifetime, Fluoroshield™ with DAPI mounting medium (Sigma-Aldrich) was used. Cells were mounted on slides and observed with a fluorescent microscope Olympus BX53 (Olympus, Tokyo, Japan). Photosynthetic E. gracilis served as a positive control, and the primary osmotroph (i.e. aplastidic) R. costata as a negative control.
Competing interests
The authors declare that they have no financial or non-financial competing interests.
Acknowledgements
We thank Vladimír Hampl (Charles University) for providing the culture of Rhabdomonas costata and Cyrille Y. Botté (University of Grenoble I) for providing the anti-DGDG antibody. We acknowledge computation resources provided by CERIT-SC and MetaCentrum, Brno, Czech Republic, and the infrastructure grant “Přístroje IET” (CZ.1.05/2.1.00/19.0388). We thank Laboratory of Analytical Biochemistry and Metabolomics (Biology Centre, Czech Academy of Sciences) for free access to LC–MS instruments. This study was supported by the Czech Science Foundation grants 17-21409S (to ME) and 16-24027S (to MO), the Czech Academy of Sciences fellowship (to ZF), and the Scientific Grant Agency of the Slovak Ministry of Education (grant VEGA 1/0535/17 to JK). This work was also part of the project “Centre for research of pathogenicity and virulence of parasites”, supported by the European Regional Development Fund, within the Operational programme for Research, Development and Education (CZ.02.1.01/0.0/0.0/16_019/0000759).
Footnotes
Abbreviations: ACP, acyl carrier protein; CBC, Calvin-Benson cycle; FAS, fatty acid synthesis; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; IPP, isopentenyl pyrophosphate; MEP, 2-C-methyl-d-erythritol 4-phosphate; MGDG/DGDG, mono-/digalactosyldiacylglycerol; MVA, mevalonate; OH-PhQ, 5’-monohydroxyphylloquinone; PhQ, phylloquinone; SQDG, sulfoquinovosyldiacylglycerol; SP, signal peptide; TMD, transmembrane domain
References
