https://orcid.org/0000-0002-1134-9238
Abstract
Plants stand out among eukaryotes due to the large variety of sterols and sterol derivatives that they can produce. These metabolites not only serve as critical determinants of membrane structures, but also act as signaling molecules, as growth-regulating hormones, or as modulators of enzyme activities. Therefore, it is critical to understand the wiring of the biosynthetic pathways by which plants generate these distinct sterols, to allow their manipulation and to dissect their precise physiological roles. Here, we review the complexity and variation of the biosynthetic routes of the most abundant phytosterols and cholesterol in the green lineage and how different enzymes in these pathways are conserved and diverged from humans, yeast, and even bacteria. Many enzymatic steps show a deep evolutionary conservation, while others are executed by completely different enzymes. This has important implications for the use and specificity of available human and yeast sterol biosynthesis inhibitors in plants, and argues for the development of plant-tailored inhibitors of sterol biosynthesis.
Introduction
Sterols are a class of triterpenoid lipids that consist of a hydrated phenanthrene group and a cyclopentane ring. These molecules have a crucial impact on membrane fluidity and transmembrane export and import processes, and some sterols can even act as second messengers or signaling molecules during developmental and cellular signaling processes (Boutté and Grebe, 2009; Vriet et al., 2013; Valitova et al., 2016; Boutté and Jaillais, 2020).
The ancient rise in atmospheric O2 levels enabled the evolution of the oxygen-dependent sterol biosynthesis pathways found in all eukaryotes (Mouritsen, 2005; Galea and Brown, 2009), and is often used as an indicator of eukaryotic life. This is contrasted by the occurrence of hopanoids in prokaryotes, which are ring-structured molecules that look similar to sterols and exert analogous functions in the membranes to cholesterol, but do not require O2 for their biosynthesis and lack a 3β-hydroxyl group (Sáenz et al., 2015). Notably, some bacteria also produce sterols, a feature that was presumably gained via horizontal gene transfer (Bode et al., 2003; Rivas-Marin et al., 2019).
While sterols occur in all eukaryotic organisms, the types and amounts of the main sterols vary considerably between the different kingdoms. For example, the major sterol produced in animals is cholesterol, whereas fungi mainly produce ergosterol. Plants, on the other hand, produce a wide variety of sterols (or phytosterols), with >200 variants known to date (Guo et al., 1995; Benveniste, 2004; Schaller, 2004). Among the phytosterols, campesterol, stigmasterol, and β-sitosterol make up the predominant molecules of the sterol profile in plants (Schaeffer et al., 2001; Benveniste, 2004). Cholesterol and derivatives are typically 100- to 1000-fold less abundant in plants than in humans, but in some species they make up a large fraction of the total sterol content, as seen in canola, Solanaceae, Liliaceae, and Scrophulariaceae (Behrman and Gopalan, 2005). In addition to free sterols, plants also accumulate a variety of sterol conjugates such as fatty acid acyl sterol esters (SEs), steryl glucosides (SGs), and acyl steryl glucosides (ASGs) (Ferrer et al., 2017; Zhang et al., 2020). A survey of different diatoms also revealed a very complex sterol landscape, again with a completely different set of specific sterols being the most abundant, even varying between different diatom lineages (Rampen et al., 2010). Also, the green algae display a markedly distinct sterol profile compared with land plants, being mainly dominated by cholesterol, 24-methylene cholesterol, and 28-isofucosterol (Lopes et al., 2011; Li et al., 2017). This diversity in sterol composition is sufficient to allow for chemotaxonomic classification of algae (Taipale et al., 2016).
The common evolutionary origin means that many conserved enzymatic activities of sterol biosynthesis across the eukaryotic kingdoms are sensitive to the same pharmacology. This allows the use of a part of the vast sterol pharmacology, which was originally designed to inhibit human or fungal sterol biosynthesis, to interfere with corresponding enzymes in plants. Here, we review our current understanding of phytosterol and cholesterol biosynthesis in plants, and highlight available pharmacology and mutants, mainly using Arabidopsis (Arabidopsis thaliana) as an example.
Early sterol biosynthesis: building the universal terpene precursors
Squalene is the universal precursor of triterpenes; in prokaryotes for hopanoids and in eukaryotes of steroids (including phytosterols, lanosterol, and cholesterol). For details about its biochemistry, we refer the reader to specialized literature (Vranová et al., 2013; Henry et al., 2018). In brief, squalene is formed by the condensation of two farnesyl pyrophosphate (FPP) units by squalene synthase (SQS), and FPP is an assembly of two isopentenyl pyrophosphates (IPPs) and one molecule of dimethylallyl pyrophosphate (DMAPP), catalyzed by FPP synthase (FPPS) (Fig. 1). The cytoplasmic IPP and DMAPP levels are subject to a highly complex metabolic regulation in plants, possibly involving at least two distinct mechanisms. The best characterized, and most important pathway for cytosolic IPP production is the mevalonate (MVA) pathway, which is largely conserved across eukaryotes and archaea. Central to this pathway is the activity of 3-hydroxy-3-methylglutaryl-CoA reductase (HMGR) that produces MVA, a precursor of IPP and DMAPP (Fig. 1). An alternative MVA pathway generates isopentenyl phosphate (IP) and dimethylallyl phosphate (DMAP), which can be phosphorylated to form IPP and DMAPP (Henry et al., 2015, 2018). The methylerythritol phosphate (MEP) pathway produces IPP and DMAPP in plastids (Banerjee and Sharkey, 2014), of which a fraction feeds into the cytoplasm (Kasahara et al., 2002). The contribution of the MEP pathway to the cytoplasmic IPP pool varies greatly between species, which in some green algae led to the complete loss of the MVA pathway (Lohr et al., 2012).
Key steps in squalene biosynthesis with an indication of inhibitor targets. In the cytoplasm, IPP and DMAPP mainly derive from MVA, which depends on HMGR activity. Additional contributions to the cytoplasmic IPP and DMAPP pools are phosphorylation of IP and DMAP, and IPP and DMAPP that was synthesized in the plastids via the MEP pathway. Inhibitors are indicated in red. DMAP, dimethylallyl phosphate; DMAPP, dimethylallyl pyrophosphate; FPP, farnesyl pyrophosphate; FPPS, farnesyl pyrophosphate synthase; GA-3P, glyceraldehyde 3-phosphate; GGPP, geranylgeranyl pyrophosphate; GPP, geranyl pyrophosphate; GPPS, geranyl pyrophosphate synthase; HMG, 3-hydroxy-3-methylglutaryl; HMGR, HMG-CoA reductase; HMGS, HMG-CoA synthase; IP, isopentenyl phosphate; IPP, isopentenyl pyrophosphate; MEP, methylerythritol phosphate; MVA, mevalonate.
Despite the complex regulation of cytoplasmic IPP and DMAPP levels, HMGR activity has been shown to be rate limiting for sterol biosynthesis in plants (Shimada et al., 2019). Also in humans, HMGR activity is rate limiting for cholesterol biosynthesis, making it a major therapeutic target for treating hypercholesterolemia (Davies et al., 2016). This medicinal application has led to the development of statins that compete with the endogenous substrate for the HMG-CoA-binding site in HMGR (Istvan and Deisenhofer, 2001). Given the deep evolutionary functional conservation of HMGR activity, statins, such as lovastatin and mevastatin, are also potent inhibitors of plant HMGRs. Consistently, the Arabidopsis hmgr mutants (hmg1 and hmg2) are hypersensitive to lovastatin (Suzuki et al., 2004). Moreover, lovastatin treatment had a fast, but transient, negative effect on sterol content in Arabidopsis seedlings, suggesting compensatory IPP and DMAPP flux from the plastids into the cytosol (Laule et al., 2003).
The subsequent assembly of IPP and DMAPP into FPP can be impaired by inducible gene silencing of both Arabidopsis FPSs, resulting in a reduction of sterol levels, but also defects in plastidial terpene metabolism (Manzano et al., 2016). The complete loss of FPS function was embryonic lethal (Closa et al., 2010). While human FPS can be potently inhibited by nitrogen-containing bisphosphonates (Drake et al., 2008), these inhibitors are currently not commonly used in plants.
The subsequent condensation of FPP into squalene depends on a single functional SQS in Arabidopsis (SQS1; Busquets et al., 2008). This SQS can complement a Saccharomyces cerevisiae erg9 null mutant and displayed the expected biochemical activity in vitro (Kribii et al., 1997; Busquets et al., 2008). No mutants have been characterized to date. Squalestatins (also called zaragozic acids) are highly potent and specific competitive inhibitors of rat SQS, with apparent subnanomolar Ki values (Baxter et al., 1992; Bergstrom et al., 1993). Also in plants, squalestatins are highly potent inhibitors that inhibit SQS in BY-2 cell suspensions, with an IC50 value of 5.5 nM, possibly via an irreversible inhibition mechanism (Hartmann et al., 2000; Wentzinger et al., 2002). Interestingly, exogenous application of squalestatins to Arabidopsis plants impairs the plants’ fertility and induces transcriptional responses that are also induced in lovastatin-treated plants (Suzuki et al., 2004). The transcriptional changes seen after squalestatin or lovastatin treatment were proposed to reflect a response to defective isoprenoid or non-brassinosteroidal steroid accumulations, via an unknown mechanism rather than an effect of reduced cytokinin biosynthesis (Suzuki et al., 2004). One possible explanation is that sterol deficiency differentially regulates the activities of homeodomain-containing transcription factors that contain a lipid/sterol-binding StAR-related lipid transfer (START) domain (Schrick et al., 2004, 2014). An interesting consequence of inhibiting SQS by squalestatin is the increased availability of FPP for sesquiterpene biosynthesis (Kobayashi et al., 2017).
Committing to sterol biosynthesis: squalene epoxidase
The subsequent, oxygen-dependent epoxidation of squalene into 2,3-oxidosqualene by squalene epoxidases (SQEs) is a key step in eukaryotic sterol biosynthesis (Thimmappa et al., 2014). Of three functional SQEs in Arabidopsis that can rescue SQE-deficient yeast (Rasbery et al., 2007; Laranjeira et al., 2015), SQE1 seems to play the most predominant function, as a single sqe1 mutant already displays pleiotropic phenotypes in the root and shoot (Rasbery et al., 2007; Posé et al., 2009). However, these phenotypes were not due to the reduced sterol content of the mutant, but because of hyperaccumulation of its substrate, squalene (Doblas et al., 2013).
Docking analyses on modeled SQE suggest that the allylamine fungicide, terbinafine, causes a conformational change that blocks one mode of substrate binding, while changing the geometry of another (Nowosielski et al., 2011). Although conventional plant SQEs can complement yeast SQE-deficient mutants (Rasbery et al., 2007), they are not highly sensitive to these inhibitors (Wentzinger et al., 2002). This is not surprising as single amino acid substitutions in SQE are sufficient to confer terbinafine resistance to yeast mutants (Leber et al., 2003). Yet, the sqe1-5 mutant is hypersensitive to terbinafine (Posé et al., 2009). On the other hand, some organisms, such as the diatom Phaeodactylum tricornutum, are completely insensitive to terbinafine (Fabris et al., 2014; Pollier et al., 2019). This can be easily explained by the presence of an alternative SQE, that belongs to the fatty acid hydroxylase superfamily, instead of to the common flavoprotein monooxygenase-type SQEs (Pollier et al., 2019), thus requiring a specific pharmacology.
Committing to sterol biosynthesis: oxidosqualene cyclases
The biosynthetic step that forms the branch point between primary sterol biosynthesis and specialized triterpene metabolism is the cyclization of oxidosqualene catalyzed by specialized oxidosqualene cyclases (OSCs) (Fig. 2). These OSCs evolved from bacterial squalene/hopane synthases, and the Arabidopsis genome contains 13 of them. In contrast to animals and yeast that use lanosterol as the first cyclic intermediate in cholesterol and ergosterol biosynthesis, respectively, plants use cycloartenol as the main cyclic intermediate in phytosterol and cholesterol biosynthesis (Fig. 2). This 2,3-oxidosqualene cyclization product is produced by the enzyme cycloartenol synthase 1 (CAS1) in Arabidopsis (Gas-Pascual et al., 2014; Thimmappa et al., 2014). In addition to CAS, Arabidopsis also encodes lanosterol synthase (LAS), lupeol synthase (LUP1), and β-amyrin synthase (bAS), OSCs that show a widespread occurrence throughout the plant kingdom (Thimmappa et al., 2014; Miettinen et al., 2018) with a suggested role in phytosterol biosynthesis (LAS) or plant defense (LUP1 and bAS). In addition to these ‘universal’ plant OSCs, Arabidopsis also encodes a set of OSCs that are specific to Brassicaceae or even to Arabidopsis as a species. These OSCs include thalianol synthase (THAS), marneral synthase (MRN1), and arabidiol synthase (PEN1), OSCs with a suggested role in plant defense and modulation of the root microbiome (Huang et al., 2019).
Conversion of squalene into cycloartenol and lanosterol. SQE, squalene epoxidase; alt. SQE, alternative SQE (found in diatoms); LAS, lanosterol synthase; CAS, cycloartenol synthase, THAS, thalianol synthase. Terbinafine inhibits SQE, but not alt. SQE. RO 48-8071 inhibits the 2,3-oxidosqualene cyclases LAS and CAS.
While strong cas1 mutant alleles are male gametophytic lethal, hypomorphic cas1-1 mutants displayed organ-specific albinism (Babiychuk et al., 2008). Strikingly, although these hypomorphic mutants accumulate 2,3-oxidosqualene, the total phytosterol levels, albeit with a reduction in β-sitosterol, remained unchanged in cas1-1 mutants, suggesting the existence of alternative pathways contributing to phytosterol biosynthesis (Babiychuk et al., 2008). With the exception of a significant reduction in β-sitosterol content, similar results were obtained by overexpression of cytokinin-induced F-box (CFB) that stimulates CAS1 turnover (Brenner et al., 2017). Interestingly, cas1-1 also had reduced cholesterol levels (Babiychuk et al., 2008), consistent with the recent identification of cycloartenol as a precursor in cholesterol biosynthesis in plants (Fig. 3) (Sonawane et al., 2016).
Scheme of C4 and C24 metabolism in cycloartenol-derived phytosterol and cholesterol biosynthesis. SMO1/2/3/4, sterol-4α-methyl oxidase 1/2/3/4; C4-DMC, C4-sterol demethylation complex; SMT1/2/3, C-24 sterol methyltransferase 1/2/3; SSR1/2; sterol side chain reductase 1/ 2. Blue circles indicate the target sites of SMTs and SSRs at distinct positions in the sterol biosynthesis pathway. Green circles indicate the target sites of C4-DMCs. C4-(di)methylated sterol intermediates are indicated by the dashed-line box.
A possible alternative route for phytosterol biosynthesis is LAS dependent (Fig. 2). The Arabidopsis genome codes for one functional LAS gene that can complement a LAS-deficient yeast strain (Kolesnikova et al., 2006; Sawai et al., 2006; Suzuki et al., 2006). However, the las1 knock-out, besides a reduction in sitostanol, did not display gross changes in total endogenous sterol content (Suzuki et al., 2006). Tracer feeding confirmed that the LAS pathway accounts for ±1.5% of β-sitosterol biosynthesis in Arabidopsis, which could be increased to 4.5% by overexpression of LAS1 (Ohyama et al., 2009). How lanosterol is converted into phytosterols currently remains unknown.
In humans, the benchmark OSC inhibitor is the competitive inhibitor RO 48-8071 (Thoma et al., 2004). Similar to cas1-1 mutants, this inhibitor causes the accumulation of 2,3-oxidosqualene in Arabidopsis seedlings and in plant cell suspensions (Gas-Pascual et al., 2015). Only after prolonged treatments (6 d) did it also cause reductions of β-sitosterol and stigmasterol, and an increase of campesterol, that was probably due to a transcriptional down-regulation of SMT2 (see further) (Gas-Pascual et al., 2015). The stronger effect on phytosterol content by the inhibitor treatment than in the cas1 mutant might be explained by additional inhibition of other OSCs that contribute to phytosterol biosynthesis, such as LAS1. Recently, lauryldimethylamine oxide (LDAO) was characterized as a potent CAS1 inhibitor in Arabidopsis, but also of the more downstream CPI activity (Darnet et al., 2020).
Diversification of the sterol end-products: side chain metabolism
Cycloartenol is a precursor for the biosynthesis of phytosterols as well as of cholesterol in plants, by methylation at the C24-position by the C-24 sterol methyltransferase SMT1, or by Δ24 reduction via Sterol Side Chain Reductase 2 (SSR2), respectively (Fig. 3) (Sawai et al., 2014; Sonawane et al., 2016).
In Arabidopsis, SMT1 catalyzes the formation of Δ 5 C-24 alkyl sterols that are precursors for campesterol, β-sitosterol, and stigmasterol. Its enzymatic activity was demonstrated via biochemical assays and via an ERG6-deficient yeast complementation assay (Diener et al., 2000). At the level of the sterol content, smt1 mutants showed a reduction in β-sitosterol, but not in campesterol content (Diener et al., 2000; Schrick et al., 2002; Willemsen et al., 2003). The normal campesterol levels in smt1 are surprising as campesterol should derive from the same SMT1-dependent pathway as β-sitosterol. One explanation could be that the diverged SMT2 and SMT3 can partially substitute for the lack of SMT1 at the expense of their normal, more downstream function in producing ethylsterols, such as β-sitosterol, resulting in a higher flux of precursors into campesterol biosynthesis.
Later in the pathway, from 24-methylenelophenol onwards, the pathway bifurcates via two separate branches, to form the 24-ethylsterols, β-sitosterol and stigmasterol, or the 24-methylsterol, campesterol, as end-products, respectively (Figs 3, 4). The bifurcation is due to the formation of an ethyl side chain at the C-24 position, via a second type of SMT [in Arabidopsis, SMT2/COTYLEDON VASCULAR PATTERN 1 (SMT2/CVP1) and SMT3] (Schaeffer et al., 2001; Carland et al., 2010). Therefore, the activity of these SMT enzymes controls the balance of ethylsterol and methylsterols. Consistently with the proposed position of SMT2/3 in the pathway, the cvp1 single and cvp1smt3 double mutants accumulate 24-methylenelophenol, and campesterol biosynthesis at the expense of stigmasterol and β-sitosterol. Intriguingly, the sterol profile of the cvp1smt3 double mutant showed marked resemblances to that of smt1, such as increased cholesterol and 24-methylenecholesterol levels, suggesting partially overlapping biochemical functions at the first methylation step (Carland et al., 2010).
Parallel processing of C4-methyl sterols in phytosterol and cholesterol biosynthesis. CPI, cycloeucalenol cycloisomerase; CVP1, cotyledon vascular pattern 1; CYP51G1, cytochrome P450 51G1; CYP710A1, cytochrome P450 710A1; DIM, DIMINUTO; DWF1/5/7, DWARF1/5/7; FK, FACKEL; HYD, HYDRA; SMO, sterol-4α-methyl oxidase. C14-R, C14 reductase, C14-DM, C14 demethylase; 8,7SI C-8,7 sterol isomerase (green-yellow background corresponds to Fig. 3)
In contrast to mammals that synthesize cholesterol via a lanosterol intermediate, plants mainly initiate cholesterol biosynthesis using cycloartenol as a precursor (Sawai et al., 2014; Sonawane et al., 2016). The responsible gene was identified in a homology blast using the human DHCR24 amino acid sequence as a query (Sawai et al., 2014). While two homologs with different specificities were identified in potato and tomato, SSR1 and SSR2, Arabidopsis contains only SSR1, previously identified as DWF1, an enzyme that was characterized in a more downstream step of phytosterol biosynthesis. Gene silencing of SSR2 in potato caused a reduction in cholesterol and cholesterol-derived steroidal glycoalkaloids, and an increase in campesterol and β-sitosterol levels (Sawai et al., 2014). Conversely, cholesterol became the most abundant sterol in the Arabidopsis smt1 alleles (Diener et al., 2000; Schrick et al., 2002; Willemsen et al., 2003), a phenomenon that is also observed in a bermudagrass mutant with reduced SMT1 expression (Chen et al., 2018). The increase in cholesterol levels is consistent with SMT1 competing with SSR2 for the precursor cycloartenol. Notably, Arabidopsis contains no known SSR2 enzyme that could explain the cholesterol accumulation in smt1 mutants, suggesting that its SSR1 displays some SSR2 activity.
Several side chain azasterols, carbocationic transition state analogs of the substrates of SMT, were previously identified as SMT inhibitors in yeast (Renard et al., 2009). Feeding experiments allowed for the identification of several azasterols that selectively inhibit either SMT1 or SMT2/SMT3 activities (Darnet et al., 2020). The SMT1-selective azasterols caused sterol profiles typically observed in smt1 mutants, including accumulation of cholesterol and cycloartenol. The homology of SSR to human DHCR24 suggests that DHCR24 inhibitors could be used as templates to develop selective SSR1 and SSR2 inhibitors, similarly to what was done for SMT1 and SMT2/3.
A double C4-demethylation via distinct sets of sterol-4α-methyl oxidases
While in mammals and yeast a double demethylation at the C4 position is executed consecutively by a single multienzyme complex, plants have up to four distinct C4-demethylation complexes, acting at different positions in the pathway (Fig. 3) (Darnet and Schaller, 2019). The highly conserved C4-demethylation complex (C4-DMC) consists of sterol-4α-methyl oxidase (SMO), 4α-carboxysterol-C3-dehydrogenase/C4-decarboxylase (CSD), and sterone ketoreductase (SR), tethered together via ERG28 (Mialoundama et al., 2013; Sonawane et al., 2016).
Plant SMOs can largely complement the 4,4-dimethyl sterol accumulation in the erg25 mutant yeast strains (Darnet et al., 2001). The first family, SMO1, is specifically involved in removal of a single methyl group from the C4 position of the 4,4-dimethyl sterol, 24-methylenecycloartenol, generating the 4-methyl sterol, cycloeucalenol. The second family (SMO2) is specifically active in removing the second methyl at the C4 position, at more downstream steps of phytosterol biosynthesis (Fig. 4). The tomato cholesterol biosynthetic pathway involves another pair of SMOs (SMO3 and SMO4 that evolved from SMO1 and SMO2, respectively) at analogous positions in the cholesterol pathway (Fig. 4) (Sonawane et al., 2016).
Using tissue cultures of smo1-1smo1-2/+ roots, and virus-induced gene silencing (VIGS), it was shown that smo1 activity is required for biosynthesis of β-sitosterol, stigmasterol, campesterol, and cholesterol, which was accompanied by accumulation of the SMO1 substrate 24-methylenecycloartenol and its precursor cycloartenol (Darnet and Rahier, 2004; Song et al., 2019). Moreover, VIGS of SMO2 genes led to accumulation of 4α-methylsterols in Nicotiana benthamiana leaves (Darnet and Rahier, 2004), which could be confirmed by analyzing smo2-1smo2-2/+ and smo2-1/+ smo2-2 (Zhang et al., 2016). Jointly, these knockdown strategies provide in planta support for their biochemical function.
Plant CSDs have been shown to complement ergosterol biosynthesis in erg26 yeast mutants (Rahier et al., 2006), putatively acting in conjunction with the distinct plant SMO types in phytosterol and cholesterol biosynthesis (Rahier et al., 2006; Sonawane et al., 2016). Additionally, silencing of a putative SR gene in tomato further supported a role for distinct C4-demethylation complexes involved in phytosterol and cholesterol biosynthesis (Sonawane et al., 2016).
Importantly, despite their ability to complement the erg25 mutant, plant SMOs display marked differences in substrate specificity, and are 500-fold less sensitive to the antifungal SMO inhibitor APB (6-amino-2-n-pentylthiobenzothiazole) (Darnet and Rahier, 2003). This suggests that new and subtype-specific SMO inhibitors need to be developed to target these steps in phytosterol and cholesterol biosynthesis. Additionally, inhibitors of CSD and SR activity that were recently identified remain to be evaluated for interference with C4-demethylation in plants (Darnet and Schaller, 2019).
Processing of C4-methylsterols: cycloeucalenol cycloisomerase (CPI1)
The opening of the cyclopropane ring of cycloeucalenol by cycloeucalenol cycloisomerase1 (CPI1) subsequently leads to the production of obtusifoliol (Fig. 4). Consistent with its biochemical function, the cpi1 mutant is characterized by increased levels of cycloeucalenol and a strong decrease in β-sitosterol and stigmasterol (Men et al., 2008). Additionally the mutant accumulates some abnormal sterols, such as 24-methylpollinastanol. Silencing of the single copy of CPI in tomato revealed an analogous biochemical function in cholesterol biosynthesis, as indicated by the accumulation of the 31-norcycloartenol substrate (Sonawane et al., 2016).
Currently, no specific CPI inhibitors are available. CPI1 catalyzes similar reactions to C14 sterol reductase and C8,7 sterol isomerase, making them shared targets for molecular inhibition by morpholines such as fenpropimorph (Taton et al., 1987). Moreover CPI was found to be a secondary target of the CAS1 inhibitor LDAO (Darnet et al., 2020).
Processing of C4-methylsterols: obtusifoliol 14α-demethylase (CYP51G1)
Obtusifoliol then undergoes demethylation and reduction at its C14 position, via an obtusifoliol 14α-demethylase to form 4α-methyl-5α-ergosta-8,24(28)-trien-3β-ol (Fig. 4). This enzymatic activity is exerted by CYP51, a cytochrome P450-dependent monooxygenase that is conserved across phyla (Lepesheva and Waterman, 2007). Arabidopsis contains a single, functional sterol C14-demethylase, CYP51 (renamed CYP51G1; Bak et al., 2011), that can complement the erg11 yeast mutant, deficient in lanosterol C14-demethylase (Kushiro et al., 2001).
Consistent with its metabolic function in phytosterol biosynthesis, a cyp51 mutant accumulated its substrate obtusifoliol, at the expense of downstream phytosterols, campesterol, β-sitosterol, and stigmasterol (Kim et al., 2005). CYP51 also catalyzes cholesterol biosynthesis at an analogous step in Arabidopsis, tomato, and N. benthamiana (Fig. 4) (Kim et al., 2005; Sonawane et al., 2016).
CYP51 belongs to the cytochrome P450s, a large group of enzymes that are often sensitive to two types of azoles: the imidazoles, such as clotrimazole, oxiconazole, and ketoconazole; and the triazoles, such as triadimenol, voriconazole, and fluconazole. These are nitrogen-containing heterocyclic compounds that non-competitively bind to the ferric ion of the heme group of the cytochrome P450, thus preventing it from binding its substrate (Warrilow et al., 2013). Despite being primarily used as fungicides, several azoles also cause growth inhibition in plants, which may be due to interference with downstream brassinosteroid biosynthesis (Rozhon et al., 2013). The fungicidal activity of azoles derives from a combination of depleting ergosterol and accumulating the toxic obtusifoliol metabolite 14α-methyl ergosta 8,24(28)-dien-3β-6α-diol (Martel et al., 2010a, b). One of the potential problems of working with azoles is their promiscuity towards diverse CYPs, and thus their potential for off-target effects, especially at higher concentrations, which are often needed to obtain a significant inhibitory effect on phytosterol biosynthesis in plants and diatoms (Fabris et al., 2014). Indeed, uniconazole does not only inhibit CYP707A, that is involved in abscisic acid catabolism, it can also inhibit CYP90B1/DWARF4, that is involved in brassinosteroid biosynthesis (Fujiyama et al., 2019).
Processing of C4-methylsterols: sterol C14-reductase (FACKEL /FK) and C-Δ8,Δ7 sterol isomerase (HYDRA1/HYD1)
In the subsequent steps, a sterol C14 reductase generates 4α-methylfecosterol that is converted into 4α-methylenelophenol by a C-Δ8,Δ7 sterol isomerase (Fig. 4). Both enzymatic steps are encoded by single genes in Arabidopsis and in tomato, FACKEL/HYDRA2 (FK/HYD2) and HYDRA1 (HYD1), respectively (Sonawane et al., 2016). Both types of enzymes can complement the ergosterol biosynthesis in corresponding mutant yeast strains, erg24 (Schrick et al., 2000) and erg2 (Grebenok et al., 1998; Sonawane et al., 2016), respectively. Moreover, the sterol profiles are markedly similar, showing defects that are more prominent in actively dividing mutant calli than in the corresponding mutant seedlings (Schrick et al., 2000, 2002; Souter et al., 2002). Calli of fk and hyd1 are mainly affected in their campesterol levels, and to a lesser extent in their β-sitosterol levels, while stigmasterol levels remain largely unaffected and cholesterol levels even increase (Schrick et al., 2000, 2002). The largely specific effect on campesterol is difficult to explain as the other phytosterols and cholesterol all depend on precursors generated via these single gene-encoded enzymatic activities (Sonawane et al., 2016).
Together with CPI, FK/HYD2 and HYD1 are targeted by a largely overlapping pharmacology (Fig. 4). An important class of inhibitors that target these enzymes are the morpholine fungicides, such as fenpropimorph. These compounds inhibit C8,7 sterol isomerases (nanomolar concentrations) and/or C14 sterol reductases (micromolar concentrations) in fungi and yeast (Mercer, 1993). In plants, morpholines inhibit HYD1, FK, and CPI (Rahier et al., 1986; Taton et al., 1987; He et al., 2003). While fenpropimorph is the most commonly used morpholine in plants, it requires relatively high concentrations to function (30–100 μM), it is unstable, and is relatively expensive. Plants treated with morpholines have a disturbed sterol profile and growth impairments, similar to those observed in mutants defective in the targeted enzymes (He et al., 2003).
A strong, more specific inhibitor of C-14 sterol reductases is the antifungal agent 15-aza‐24-methylene-d-homocholesta-8,14-dien-3β-ol (15-azasterol) (Fig. 4), which causes similar changes in the sterol profiles to those observed in the fk mutant (Schrick et al., 2002).
Final processing steps in phytosterol and cholesterol biosynthesis
After the second C4-demethylation, the subsequent production of campesterol and stigmasterol is effected by three enzymes acting in both branches of the pathway at corresponding steps: Δ 7-sterol-C5-desaturase (C5-SD1) [in Arabidopsis DWARF7/STEROL1 (DWF7/STE1)], sterol Δ 7-reductase (7-DR1) [in Arabidopsis, Dwarf5 (DWF5)], and Δ 24-STEROL-Δ 24-REDUCTASE/SIDE CHAIN REDUCTASE1 (SSR1) [in Arabidopsis DIMINUTO/DWARF1 (DIM/DWF1)] (Fig. 5). Cholesterol biosynthesis in tomato involves the diverged SMO4, C5-SD-2, and 7-DR-2 (Sonawane et al., 2016) (Fig. 5).
Schemes of the final enzymatic conversions to generate stigmasterol, campesterol, and cholesterol. DIM, DIMINUTO; DWF1/5/7, DWARF1/5/7; HYD, HYDRA; STE1, STEROL1; C5-SD1/2; 7-DR1/2; SSR1; C22-SD; CYP710A1. Left: ethylesterol branch; middle, methyl-sterol branch; right, cholesterol branch (light and dark green-yellow backgrounds correspond to Fig. 3)
The genes DWF7/STE1, DWF5 and DIM/DWF1 encode the C5-SD1, 7-DR1, and SSR1 enzymes in Arabidopsis, and are required for the production of campesterol and β-sitosterol. The dwf7/ste1 mutants accumulate Δ7-sterols, and are strongly defective in campesterol, β-sitosterol, and stigmasterol (Silvestro et al., 2013). The dwf5 mutant also accumulates Δ7-, Δ5,7-, and Δ8-sterols and has strongly reduced levels of campesterol, β-sitosterol, and stigmasterol (Silvestro et al., 2013). In dim/dwf1 mutants, campesterol, β-sitosterol, and stigmasterol were reduced, while the β-sitosterol and campesterol precursors isofucosterol and 24-methylene cholesterol, respectively, were increased (Klahre et al., 1998). Interestingly, besides acting as a sterol C24-reductase, DWF1 was found to also display brassinosteroid C24-reductase activity (Youn et al., 2018). Currently, no inhibitors are available for this part of the phytosterol biosynthesis pathway.
In the ethylsterol branch, β-sitosterol undergoes C-22 desaturation by the C22 sterol desaturase CYP710A1, resulting in the end-product of this pathway: stigmasterol (Morikawa et al., 2006). However, not many details are known about this desaturation reaction in higher plants. Interestingly, in Arabidopsis, a second CYP710 enzyme (CYP710A2) is also able to produce stigmasterol from β-sitosterol, and can also produce brassicasterol from 24-epi-campesterol (Benveniste, 2002; Morikawa et al., 2006). Given that CYP710 belongs to the cytochrome P450 family, it is likely that this enzyme is sensitive to azoles, as indicated in the CYP51 section.
Sterol conjugation and oxidation
As well as through biosynthesis, sterol levels can be regulated via conjugation, to form fatty acid acyl SEs, SGs, and ASGs. The SEs are major constituents of cytosolic lipid droplets (Shimada et al., 2019) and are proposed to serve as storage forms that can be rapidly mobilized. SGs and ASGs contribute to the membrane organization, for example as important constituents of plasma membrane microdomains. For details about the biosynthesis and physiological function of SEs, SGs, and ASGs, we refer the reader to some excellent recent reviews (Ferrer et al., 2017; Korber et al., 2017; Mamode Cassim et al., 2019; Zhang et al., 2020)
While the hydrolases and glycosidases remain elusive, several other key biosynthetic enzymes have been identified and characterized. In brief, SEs are formed by activities of two distinct types of acyltransferases: acylCoA:sterol acyltransferases (ASATs) and phospholipid:sterol acyltransferases (PSATs), using long-chain acyl-CoA and unsaturated fatty acyl groups from phospholipids, respectively, as donors (Chen et al., 2007). Although belonging to a family of six related proteins, Arabidopsis is thought to have only one PSAT (PSAT1) (Ferrer et al., 2017). Similarly, only one member of the membrane-bound O-acyltransferases (MBOATs) was found to display ASAT activity (Chen et al., 2007). Mutants in psat1 have early leaf senescence, but asat1 mutants lack an obvious morphological phenotype (Bouvier-Navé et al., 2010). At the level of SE formation, analysis of these mutants showed major contributions of PSAT1 over ASAT1.
Several inhibitors of human ASAT enzymes have been identified (Ohtawa et al., 2018; Ohshiro et al., 2020a, b), but their activity and specificity in plants remain to be demonstrated. The recently resolved structure of human sterol O-acyltransferase 1 revealed that the small molecule mode CI-976 blocks the accessibility of several active site residues (Guan et al., 2020), and may serve as a template for docking analyses to assess the activity of such inhibitors on plant enzymes.
Two steryl glycosyl transferases (UGT80A2/B1), both related to the yeast steryl glycosl transferase UGT51A1, are jointly required for SE formation in Arabidopsis (Stucky et al., 2015), by adding a sugar moiety, usually glucose, to the C3-hydroxyl group of the sterol. However, while steryl glycosyl acyltransferase activity could be detected, its molecular constituents remain to be identified. Mutants defective in the steryl glycosyl transferase enzymes display moderate phenotypes, ranging from some seed coat and embryo phenotypes to a reduced root elongation (Stucky et al., 2015). Recently, reduced root hair production in ugt80b1 mutants was connected to mislocalization of the root hair cell fate regulator SCRAMBLED (Pook et al., 2017). The residual SG and ASG accumulation in these mutants suggests that additional, redundant SGT enzyme activities remain to be identified (Stucky et al., 2015). Also for UDP-glycosyl transferases, several inhibitors can be found (Oda et al., 2015; Lv et al., 2019), that have not yet been used or evaluated to inhibit steryl glycosyl transferase activities in plants.
Conclusions
Deriving from an ancestral pathway, each of the three eukaryotic kingdoms gave sterol biosynthesis its own flavor, with an end result that is tailored to the specific needs inherent to the adopted life style. In plants, animals, and fungi, the sterol composition is a key determinant of the physico-chemical properties and spatial organization of membranes. Additionally, sterols have a serious impact on physiology and signaling processes as precursors of potent hormones, such as brassinolide (Vriet et al., 2013). However, it is becoming more and more obvious that sterols are key components of plant development, in addition to their role in brassinolide biosynthesis (Boutté and Jaillais, 2020). Currently, most work has been done using Arabidopsis as a model. It is, however, easy to envisage that much remains to be learnt about the roles of the multitude of diverse sterols that exist within the green lineage. The most straightforward way to achieve this is through interference with the biosynthetic pathway using inhibitors. Several enzymatic steps are sufficiently conserved to justify the use of inhibitors that were developed in yeast and mammals, while others are not. In combination with a tremendous diversification of the enzymes, and even novel enzymes being recruited, as seen in diatoms, it is clear that a more targeted pharmacology is needed to meet specific needs. On the other hand, the existence of plant-specific, or even species-specific enzymatic activities represents attractive opportunities for the development of highly selective herbicides.
Most of our inferences about the physiological and cell biological roles of phytosterols derive from studies in mutants that accumulate C4-dimethylated sterol biosynthesis intermediates. These mutants, such as smt1 (Diener et al., 2000; Schrick et al., 2002; Willemsen et al., 2003), smo1-1smo1-2 (Song et al., 2019), cyp51 (Kim et al., 2005), fk/hyd2 (Schrick et al., 2000), hyd1 (Topping et al., 1997), cpi1 (Men et al., 2008), smt2smt3 (Souter et al., 2002; Carland et al., 2010; Short et al., 2018), and smo2-1smo2-2 (Zhang et al., 2016), display severe developmental defects, and are often associated with misregulated auxin and cytokinin homeostasis, and with defects in endocytosis and cellular polarization mechanisms. However, mutants defective in later steps of sterol biosynthesis and not accumulating C4-(di)-methylated sterol biosynthesis intermediates, dwf7/ste1, dwf5 and dim/dwf1, mainly display typical brassinosteroid deficiency phenotypes, that can be partially reversed by exogenous application of brassinolide (Vriet et al., 2013).This discrepancy in phenotypes indicates that the observed, strong, phenotypes reflect a specific toxicity, or signaling function related to the accumulation of the sterol biosynthesis intermediates, rather than defects in phytosterols. This calls for caution when inferring physiological roles of sterols in plants, as most available inhibitors result in accumulation of C4-(di)methylated sterol biosynthesis intermediates that have a severe impact on biological processes in plants, animals, and yeast (Darnet and Schaller, 2019).
Acknowledgements
KDV is funded by the Special Research Fund Ghent University (01D25813).
Author contributions
KDV and SV conceptualized the review; all authors contributed to the writing and revision of the manuscript.
Conflict of interest
The authors declare that they have no conflicts of interest.
Comments