Evolutionary divergence and functional insights into the heteromeric cis‐prenyltransferase of Paramecium tetraurelia
Agnieszka Onysk, Kamil Steczkiewicz, Mariusz Radkiewicz, Paweł Link‐Lenczowski, Przemysław Surowiecki, Karolina Sztompka, Kariona A. Grabińska, Jacek K. Nowak, Liliana Surmacz
TL;DR
This study identifies a unique heteromeric cis-prenyltransferase complex in Paramecium tetraurelia, essential for dolichol synthesis and protein glycosylation.
Contribution
The first experimental evidence of a heteromeric CPT complex in Paramecium tetraurelia, including a novel CPT-accessory subunit POC1.
Findings
POC1 is a unique CPT-accessory subunit in Paramecium tetraurelia with preserved structural and functional features.
Loss of POC1 or CPT1a leads to reduced dolichol production and cell death in Paramecium tetraurelia.
POC1 in complex with CPT subunits can synthesize polyprenyl chains in a heterologous yeast system.
Abstract
The biosynthesis of polyprenyl/dolichyl phosphate, an essential lipid carrier in protein glycosylation, occurs across all domains of life. Eukaryotic heteromeric enzymes involved in polyprenyl chain elongation consist of a highly conserved catalytic cis‐prenyltransferase subunit (CPT‐CS) and a less conserved CPT‐accessory subunit (CPT‐AS). Here, we present the first experimental evidence that dolichol biosynthesis in Paramecium tetraurelia is mediated by a heteromeric CPT complex. Using a multidisciplinary experimental approach, we identified two highly homologous catalytic CPT subunits, CPT1a and CPT1b, which exhibit high sequence similarity to other eukaryotic CPTs, along with a unique CPT‐AS, named POC1 (partner of CPT1), which is a structural and functional relative of the human dehydrodolichyl diphosphate synthase complex subunit NUS1 (also known as NgBR) and yeast Nus1 CPT‐AS.…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10- —The National Science Centre of Poland
- —The qLIFE Priority Research Area
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsPlant biochemistry and biosynthesis · Microbial Natural Products and Biosynthesis · Algal biology and biofuel production
Introduction
Dolichol biosynthesis is a fundamental biological process crucial for the life of eukaryotic cells. Dolichol plays a pivotal role as a lipid carrier for sugar moieties during the initial phases of protein N‐glycosylation, a post‐translational modification essential for cell survival, in particular critical for protein folding, stability, translocation, and protein–protein interactions. N‐glycosylation is a multistep process that commences with the formation of lipid‐linked oligosaccharide (LLO) by attaching sugar residues, N‐acetylglucosamine, mannose, and glucose, to dolichyl diphosphate (Dol‐PP). The initial stage occurs on the cytoplasmic side of the endoplasmic reticulum, giving rise to an oligosaccharide with a Man_5_GlcNAc_2_ structure. Subsequently, this structure is translocated into the lumen of the endoplasmic reticulum, where it is further elongated to the Glc_3_Man_9_GlcNAc_2_ oligosaccharide. The resulting tetradecasaccharide is then transferred from the lipid carrier (Dol‐PP) by oligosaccharide transferase (OST) onto the growing polypeptide chain, where it forms an N‐glycosidic bond with the asparagine residue of the protein molecule. Additionally, dolichyl phosphate (Dol‐P) is involved in the formation of activated monosaccharides Dol‐P‐Man and Dol‐P‐Glc, which serve as substrates for glycosyltransferases participating in N‐glycosylation, O‐glycosylation, C‐mannosylation, and glycosylphosphatidylinositol (GPI) anchor biosynthesis [1]. In humans, deficiencies in early steps of the protein glycosylation process are categorized as type I glycosylation disorders [2].
The key step in dolichol biosynthesis is the formation of a polyprenyl backbone via subsequent condensations of an initiator—an allylic farnesyl diphosphate (FPP) or geranylgeranyl diphosphate (GGPP) with various numbers of isopentenyl diphosphate molecules (IPP) performed by cis‐prenyltransferases (CPTs). In eukaryotes, polyprenol chains are synthesized as a mixture of several different‐length isoprenologues, with one predominant length that varies among species.
Following the tissue requirement, specific groups of polyprenols are finally converted to dolichols in the three‐step reaction [3]. CPTs exhibit enzymatic activity only as homo‐ or heteromeric complexes. Homodimers are composed of two identical CPT subunits containing both catalytic and C‐terminal RXG motifs [4]. Such architecture is typical for bacterial enzymes including Escherichia coli ditrans, polycis‐undecaprenyl‐diphosphate synthase (EcUPPS) [4], the majority of archaeal CPTs, for example, tritrans,polycis‐undecaprenyl‐diphosphate synthase (Tk‐IdsB) from Thermococcus kodakaraensis [5], some plant CPTs, such as dehydrodolichyl diphosphate synthases: DDPS1 (also called AtCPT1), DDPS3 (AtCPT6) and DDPS2 (AtCPT7) from Arabidopsis [6, 7, 8], SlCPT5 from tomato [9] and some Protozoa enzymes including ditrans,polycis‐undecaprenyl‐diphosphate synthase (Gl‐UPPS) from Giardia lamblia [10] acquired through horizontal gene transfer. However, most eukaryotic CPTs consist of a highly conserved catalytic subunit (CPT‐CS) lacking essential substrate binding and a C‐terminal RXG motif, and an accessory CPT‐binding protein (CPT‐AS) which is catalytically inactive but possesses the C‐terminal RXG motif. Both subunits are indispensable to form an active enzyme. One heteromeric CPT (DHDDS/NgBR), composed of dehydrodolichyl diphosphate synthase complex subunit DHDDS and dehydrodolichyl diphosphate synthase complex subunit NUS1 (also called NgBR), has been identified in humans [11, 12]. Yeast possess two heteromeric CPTs (Rer2/Nus1 and Srt1/Nus1), both containing dehydrodolichyl diphosphate synthase complex subunit NUS1 in complex with either dehydrodolichyl diphosphate synthase complex subunit RER2 or dehydrodolichyl diphosphate synthase complex subunit SRT1 [12, 13, 14, 15]. In plants, several such CPTs have been found. In Arabidopsis thaliana, the AtCPT3/Lew1 complex is composed of dehydrodolichyl diphosphate synthase 6 (also called AtCPT3) in complex with ditrans,polycis‐polyprenyl diphosphate synthase (Lew1) [16]. In tomato (Solanum lycopersicum), the SlCPT3/SlCPTBP complex consists of dehydrodolichyl diphosphate synthase CPT3 and CPT‐binding protein [17]. In lettuce (Lactuca sativa), the CPT3/CPTL2 complex includes transferase CPT3 and ditrans,polycis‐polyprenyl diphosphate synthase CPTL2 [18]. While the majority of Archaea possess single‐subunit enzymes, a subgroup within Euryarchaeota encodes two‐component CPT. This has been experimentally confirmed for Methanosarcina sp., with CPT complexes composed of tritrans,polycis‐undecaprenyl‐diphosphate synthase and undecaprenyl diphosphate synthase MA3723/MA4402 in Methanosarcina acetovorans and MM_0618/MM_1083 in M. mazei [19, 20] as well as for the Archaeoglobaceae sp., where the complex is formed by tritrans,polycis‐undecaprenyl‐diphosphate (Af1219) and ditrans,polycis‐decaprenylcistransferase (Af0707) in Archaeoglobus fulgidus [21]. The majority of Lokiarchaeota, the nearest relative of eukaryotes [22] do have predicted single‐subunit CPT [4]. Importantly, the first cultivated Lokiarchaeota with a single‐subunit CPT shares an ecological niche with Methanogenium sp. that possess heteromeric CPT [23]. Recent phylogenetic analysis suggests the presence of CPT‐AS, identified as ditrans,polycis‐decaprenylcistransferase called CLaUPPS4 (GenBank: JAGXNS010000099.1), in the C. Lokiarchaeota archaeon [24]. These observations support the hypothesis of the origin of heteromeric CPT in Euryarchaeota, subsequently acquired through horizontal gene transfer by a subgroup of Archaea ancestral to eukaryotic cells. Most of them are localized on the membrane of the endoplasmic reticulum (ER). Well established is also lipid droplets localization of CPT in yeast [13, 25, 26].
The polyisoprenoid content and mechanisms of their formation are well understood in various eukaryotes such as yeast [12, 13, 14, 15, 26], humans [11, 12], and plants [27]; whereas, in protozoa, they are studied to a lesser extent. Giardia lamblia, the parasite that causes the diarrheal disease giardiasis, accumulates dolichols containing 11 and 12 isoprene units (i.u.) which are products of the only protist CPT (Gl‐UPPS) purified to homogeneity [10]. Gl‐UPPS is a homomeric enzyme with higher homology toward Eubacteria CPTs than other Metamonada, suggesting the occurrence of horizontal gene transfer, which is a major force in the evolution of Giardia. In Leishmania amazonensis, the parasite responsible for the disease leishmaniasis, the profile of accumulated polyisoprenoids also depends on the developmental stage. Promastigotes synthesize dolichols of 11 and 12 i.u., while amastigotes mainly synthesize polyprenols of 9 i.u. which do not undergo reduction to dolichol [28, 29]. The putative of Leishmania sp. and closely related Trypanosoma sp. also belong to bacteria‐type single‐subunit enzymes [4, 10]. The presence of the same dolichols has also been shown in Plasmodium falciparum [30] responsible for the most virulent form of human malaria. However, a more recent report has shown that Plasmodium accumulates significant amounts of a mixture of polyprenols and dolichols composed of 15 to 19 i.u. The profile of accumulated polyisoprenoids is more complicated, and changes throughout the intraerythrocytic asexual cycle, as well as between asexual stages and gametocytes, were observed. Also, the active cis‐prenyltransferase Plasmodium (PfCPT) has been identified [31]. However, the PfCPT was never studied as a purified enzyme in a heterologous system. Phylogenetic analysis suggests that PfCPT is in fact CPT‐CS, which needs CPT‐AS to form an active enzyme [4]. The presence of dolichol 16 involved in the synthesis of dolichyl phosphate glucose was established in the parasitic protists Trichomonas vaginalis. The size of the dolichol correlates well with CPTs active product in vitro [32]. The phylogenetic analysis predicts the presence of a two‐subunit enzyme in Trichomonas [4].
Still less is known about the mechanism of polyisoprenoid synthesis in non‐pathogenic protozoa. Tetrahymena pyriformis accumulates a dolichol family composed of 13 to 16 units, with Dol‐14 predominating [33]. In Paramecium, the content and mechanism of dolichol synthesis have not been investigated so far. It is only known that in Paramecium primaurelia, the synthesis of Dol‐P‐Man is regulated by heterotrimeric GTP binding proteins involving the adenylyl cyclase/cAMP system and protein phosphorylation [34], while the identity of dolichols forming the pool of Dol‐P‐Man has not been determined.
This study addresses the gap in our knowledge of protozoan polyisoprenoid biosynthesis and their role in the cell. It is found that cells of Paramecium tetraurelia accumulate dolichols composed of 16 to 18 i.u., with Dol‐17 dominating regardless of the phase of the life cycle. The genome of P. tetraurelia encodes two catalytic subunits of CPTs, CPT1a and CPT1b, the former being expressed at higher levels compared to the latter. The silencing of CPT1a leads to disruption of protein glycosylation and consequently to cell death. Both CPT1a and CPT1b interact with a specific CPT‐AS, a newly identified protein remotely homologous to previously identified CPT‐ASs, called POC1 (Partner of CPT1), which is essential for the formation of an enzymatically active heteromeric CPT complex in P. tetraurelia. Our studies in free‐living protist P. tetraurelia can serve as a model to better understand dolichol synthesis in clade Alveolata, which includes parasitic organisms such as Plasmodium falciparum and Toxoplasma gondii. It also supports the assumption that the two‐component enzyme is present in Trichomonas sp.
Results
Identification and phylogenetic analysis of Paramecium
CPT
Until now, the content and profile of polyisoprenoids in Paramecium cells as well as enzyme(s) responsible for their formation have not been identified. Paramecium tetraurelia 51 proteome contains two CPT representatives, namely CPT1a (PTET.51.1.P0120334 in ParameciumDB [35]; XP_001428023.1 in NCBI) and CPT1b (PTET.51.1.P0040087; XP_001446058.1). Both Paramecium CPTs are 238 amino acid‐long with molecular masses of 28.41 and 28.27 kDa, respectively. They exhibit a ~ 35% sequence identity to Saccharomyces cerevisiae Rer2 and retain five highly conserved regions, as well as the membrane sensor domain found in other cis‐prenyltransferases (Fig. 1). However, neither of the Paramecium CPTs has the C‐terminal RXG motif required for the enzymatic activity of known homomeric cis‐prenyltransferases (Fig. 1). Enzymes lacking this motif require an auxiliary partner to provide such, like human NgBR, yeast Nus1, or Arabidopsis Lew1 [11, 12, 16, 36]. By using standard protein sequence comparison tools (like BLASTP or even PSI‐BLAST), we could not identify any CPT‐ASs within the Paramecium proteome. However, the remote homology detection method, HHSEARCH [37], which is based on comparisons of sequence meta‐profiles, yielded a previously uncharacterized protein (PTET.51.1.P0200327; XP_001439101) as remotely homologous to known CPT‐ASs (human NgBR, pdb|6z1n_B, hh‐score 33.6, probability 92; yeast Nus1, pdb|6jcn_A, hh‐score 29.5, probability 71.7) (Fig. 1). This new, putative Paramecium CPT‐AS, here named POC1 (Partner of CPT1), consists of 162 amino acids (calculated molecular mass is 19.58 kDa) and is considerably shorter compared to known CPT‐ASs (NgBR – 293 aa, Nus1–375 aa), but nevertheless contains all the necessary structural elements along with the C‐terminal RXG motif.
Multiple sequence alignment of proteins with the cis‐prenyltransferase homology domain. Homomeric cis‐prenyltransferases (blue) are represented by EcUPPS (E. coli, GenBank™ accession no. WP_000979579.1), GlUPPS (G. lamblia, EDO82194.1), and AtCPT7 (A. thaliana, NP_200685.1). Heteromeric catalytic cis‐prenyltransferase subunits (CPT‐CS, green) are represented by ScRer2 (S. cerevisiae, NP_009556.1), ScSrt1 (S. cerevisiae, AJS89026.1), HsDHDDS (H. sapiens, BAB14439.1), AtCPT3 (A. thaliana, NP_565420.1), MaUPPS‐A (M. acetovorans, WP_226990666.1), PtCPT1a (P. tetraurelia, XP_001428023.1), and PtCPT1b (P. tetraurelia, XP_001446058.1). Heteromeric accessory cis‐prenyltransferase subunits (CPT‐AS, orange) are represented by HsNgBR (H. sapiens, NP_612468.1), ScNus1 (S. cerevisiae, NP_010088.1), AtLew1 (A. thaliana, NP_001077518.1), MaUPPS‐B (M. acetovorans, WP_011024281.1), and PtPOC1 (P. tetraurelia, XP_001439101.1). Residue conservation is highlighted according to the following scheme: polar – gray, hydrophobic – yellow. Predicted secondary structure elements are shown above the corresponding alignment blocks (orange arrow for β‐strand; blue helix for α‐helix). The N‐terminal α‐helix that functions as a membrane sensor (purple) domain in heteromeric CPT‐CSs (DHDDS orthologs) is indicated. Black boxes indicate the positions of five conserved regions originally described among CPTs. The blue box highlights the conserved C‐terminal RXG motif in homomeric CPTs and heteromeric CPT‐ASs (NgBR orthologs). Numbers in parentheses represent residues omitted from the alignment for clarity. For maximum accuracy, this structure‐driven multiple sequence alignment was made by manually identifying corresponding residues based on manually assessed SPDBV [38] structural superimposition of alphafold3 models for all presented proteins.
Both CPT‐CS proteins identified in Paramecium tentatively assigned as CPT1a and CPT1b belong to the phylogenetic clade of heteromeric CPTs, and POC1 clusters together with other CPT‐ASs (Fig. 2, Table S1). Although CPT‐ASs have similar functions, they display much higher sequence diversity comparing to CPTs, which might reflect their auxiliary, non‐catalytic function. All analyzed Paramecium proteins stand out from their clades, suggesting either an ancient evolutionary events of acquiring cis‐prenyltransferases or a hyper divergence in Paramecium evolution.
Maximum likelihood phylogenetic tree of CPT proteins. The proteins (tree leafs) are tagged with protein name, NCBI identifier, and an organism's taxonomic name. Proteins from P. tetraurelia are marked with bold black font. Bacterial sequences were used as an outgroup. The figure was partially prepared using itol [39]. Phylogenetic tree was constructed using iq‐tree [40] (LG + G4 model) based on multiple sequence alignment calculated using mafft [41]. Accession numbers of the analyzed sequences are listed in Table S1.
Homology modeling of CPT1a, CPT1b, and POC1 and their complexes
Structures of all discussed proteins: CPT1a, CPT1b, and POC1 were successfully modeled with the use of alphafold3 (Fig. 3). The model of POC1 is less reliable (pTM = 0.35) than the models of CPT1a (pTM = 0.94) and CPT1b (pTM = 0.93), likely due to the lack of appropriate structural templates for highly diverged CPT‐ASs. Despite being considerably shorter than other CPT‐ASs, POC1 retains a central β‐sheet and three α‐helices, including the C‐terminal helix harboring the RXG motif important for substrate binding and forming the catalytic cavity [19, 42], whereas it lost many helices from both flanks.
alphafold3 models of P. tetraurelia CPT subunits. Structural models for (A) CPT1a, (B) CPT1b, and (C) POC1 superimposed on their corresponding human homologs (pdb|6z1n, rendered in light pink): DHDDS (CPT1a and CPT1b) and NGBR (POC1). The models colored according to the estimated confidence score are shown in panels (D), (E) and (F).
When modeled in complex with either CPT1 protein (Fig. 4), the reliability of the POC1 structural model raises (pTM = 0.63 in complex with CPT1a, pTM = 0.64 with CPT1b). Modeled heterodimer structures resemble the overall structure of the human heterotetrameric cis‐prenyltransferase complex (pdb|6z1n) [44] and demonstrate catalytic site accessibility for POC1 arginine R159 from the RXG motif. In the modeled POC1, the C‐terminal α‐helix seems to be much longer and bent away from the transferase's active site compared to NgBR. However, the prediction confidence scores are clearly decreased for this helix, suggesting its conformational flexibility and thus its potency for delivering the arginine to the catalytic site.
Predicted structure of CPT complexes. (A) X‐ray structure for human heterotetrameric cis‐prenyltransferase complex (pdb|6z1n, chains A and B corresponding to DHDDS and NGBR, respectively; C‐terminal helices are not shown for clarity) (B) and (C) structural predictions for complexes of POC1 with CPT1a and CPT1b, respectively. Farnesyl diphosphate (FPP, rendered in purple) shown in the models was naively copied from X‐ray structure to highlight the catalytic site area. All models were generated with the use of alphafold3 [43].
CPT1a and CPT1b are very similar in their amino acid sequences (89% sequence identity), and the differences localize to the protein's surface, mostly away from the interface with POC1 (Fig. 5 and Fig. S1). Only two residues differing between CPT1a and CPT1b are presented toward POC1: 156 (Asp in CPT1a, Glu in CPT1b) and 208 (Lys in CPT1a, Asn in CPT1b). The next four residues line the area where the polyprenol – CPT product is being released: 93 (Ile in CPT1a, M in CPT1b) and 110 (Phe in CPT1a, Val in CPT1b) located at the β‐sheet, and 55 (Gln in CPT1a, Glu in CPT1b) as well as 105 (Lys in CPT1a, Gln in CPT1b) at the C‐termini of both α‐helices surrounding the pocket. These four amino acid differences might alter the conformational dynamics of the pocket, which in consequence might affect the length of released products. Residue 55 interacts with the CPT1 N‐terminal amide so that its substitution to an amino acid of different polarity (here Glu to Gln) might additionally change the dynamics of this region.
Amino acid residues differing between CPT1a and CPT1b, and mutated in this study, visualized in the context of interaction with POC1. Please mind that CPT1a and CPT1b models are nearly identical, and their superimposition looks like one protein. alphafold models of CPT1a and CPT1b in complex with POC1 are superimposed, and sidechains of differing residues are rendered as magenta and green sticks for CPT1a and CPT1b, respectively. POC1 is shown as a white surface for clarity. Catalytic aspartate 26 is shown as orange sticks. All models were generated with the use of alphafold3 [43].
Cis‐Prenyltransferase heterocomplex formation – experimental evidence
To experimentally assess the interaction between Paramecium CPTs and POC1, we conducted a yeast two‐hybrid (Y2H) assay. Constructs encoding CPT1a or CPT1b fused to the activation domain of the transcription factor Gal4 (AD) and POC1 fused to the Gal4 binding domain (BD) were introduced into S. cerevisiae AH109 cells, while in the second line of experiments, opposite fusions, CPT‐BD and POC1‐AD, were tested. The Y2H analysis revealed that both CPT1a‐AD and CPT1b‐AD interacted with POC1‐BD, leading to the activation of the GAL4 reporter gene. The same interaction pattern was observed when CPTs were fused with the BD domain and POC1 with the AD domain (Fig. 6A). As a result, AH109 yeast cells were able to grow on a selection medium lacking histidine (−HIS) and containing an inhibitor of the yeast HIS3 gene (−HIS + 1 mm 3‐AT). No yeast cell growth occurred due to either nonspecific auto‐activation of the GAL4 reporter gene when transformed with single CPT‐AD or POC1‐BD constructs, or when co‐transformed with CPT‐AD and an empty DB domain or POC1‐BD and an empty AD domain. This result confirms the formation of a complex between each of Paramecium CPTs and the POC1 protein.
Analysis of POC1 and CPT1a or CPT1b complex formation and their enzymatic activity. (A) Interaction analysis of POC1 with CPT1a and CPT1b in a yeast two‐hybrid (Y2H) assay. Positive interaction between POC1 and both CPTs is indicated by the growth of yeast on the plates lacking leucine (Leu), tryptophan (Trp), and histidine (His) additionally supplemented with 1 mm 3‐AT (3‐Amino‐1,2,4‐triazole), a competitive inhibitor of the HIS3 gene product (histidine synthase) as the reporter gene in the Y2H system. Representative images are shown. (B, C) Functional complementation of S. cerevisiae rer2Δ srt1Δ nus1Δ triple‐deletion mutant by the Paramecium cis‐prenyltransferase complexes. The triple‐deletion strain, expressing G. lamblia CPT (Gl‐UPPS) from the URA3 plasmid, was co‐transformed with LEU2 and MET15 plasmids bearing wild‐type or mutated variants of CPT1a or CPT1b and POC1, as indicated. (B) Cells transformed with an empty plasmid served as the negative control, while cells transformed with LEU2 Gl‐UPPS, LEU2 Tc‐CPT, LEU2 114 Tc‐CPT, or co‐transformed with RER2 (LEU2) and NUS1 (MET15) were used as positive controls. Cells co‐transformed with the mutated POC1G161A and CPT1a or CPT1b were used to demonstrate the essential role of the glycine residue in the RXG motif of the heteromeric accessory cis‐prenyltransferase subunit. (C) Cells co‐transformed with POC1 and RER2, CPT1a and NUS1, or CPT1b and NUS1 serve as controls to assess the interaction of Paramecium proteins with yeast proteins. The transformed cells were streaked onto complete plates (YPD) or synthetic complete medium containing 1% 5‐fluoroorotic acid (FOA). The Ura3 protein, expressed from the URA3 marker in the plasmids, converts FOA to the toxic 5‐fluorouracil, leading the cells to lose the Gl‐UPPS plasmid. Cell growth was monitored for 6 days to assess phenotypic differences. Transformants that exhibit growth on FOA medium are marked in red. (D) Profile of polyisoprenoids isolated from the S. cerevisiae rer2Δ srt1Δ nus1Δ triple‐deletion mutant transformed with the indicated constructs. Note that all studied yeast cells accumulated dolichols (Dol‐11 and Dol‐12), the products of Gl‐UPPS, while cells expressing pairs of CPT1a/POC1 and CPT1b/POC1 additionally accumulated a mixture of polyprenols and dolichols containing from 14 to 16 isoprene residues (Pren/Dol‐14, Pren/Dol‐15, and Pren/Dol‐16) synthesized by these complexes. The chain length and identity of lipids were determined by comparison with external standards of a polyprenol mixture (Pren‐9 to Pren‐25). Representative High‐Performance Liquid Chromatography with Ultraviolet Detection (HPLC/UV) chromatograms are shown. (E) Relative amounts of polyprenols and dolichols synthesized by CPT1a and CPT1b in yeast cells, normalized to the total content of Dol‐11 and Dol‐12 derived from Gl‐UPPS, which supports the survival of the triple mutant. Data are presented as mean ± standard deviation (SD) from three independent experiments (n = 3), each performed in triplicate.
Next, to determine whether both CPT1a and CPT1b, either individually or in complex with POC1, exhibit functional cis‐prenyltransferase activity, we conducted a complementation assay using a Saccharomyces rer2Δ srt1Δ nus1Δ triple‐deletion mutant (KG405) which lacks yeast CPT activity [12]. The survival of mutant cells was supported by the expression of G. lamblia CPT (Gl‐UPPS) [10]. The yeast were transformed with MET15 and LEU2 plasmids bearing the cDNAs encoding for analyzed CPTs and POC1 as indicated (Fig. 6B) and their growth on the FOA plates was monitored over 6 days. The Ura3 protein, expressed from the URA3 marker within the plasmid carrying Gl‐UPPS, converts FOA (5‐fluoroorotic acid) to the toxic 5‐fluorouracil, suppressing the growth of Gl‐UPPS‐positive cells. The ability of a triple‐deletion mutant to grow on a FOA plate depends on the functionality of the expressed proteins, whether they can complement the activity of the lost yeast CPT subunits to produce polyprenols that are subsequently converted to dolichols. As observed, the co‐expression of CPT1a/POC1, as well as CPT1b/POC1, supported cell growth, but neither CPT1a, CPT1b, nor POC1 alone is sufficient. This means that CPT1a as well as CPT1b possess CPT activity only when co‐expressed with POC1. Co‐expression of NUS1/RER2, as well as expression of putative homomeric CPT of Trypanosome cruzi, was performed as a positive control of yeast growth. It is noteworthy that the co‐expression of NUS1/CPT1a and NUS1/CPT1b does not support cell growth, which means that Nus1 does not form a functional complex with Paramecium CPTases (Fig. 6B). Furthermore, to confirm that the C‐terminal RXG motif in POC1 is critical for the prenyltransferase activity of studied complexes, we tested the effect of the POC1^G161A^ substitution in the C‐terminal RXG motif. This modification resulted in a severe growth delay indicating the critical role of the RXG motif present in the POC1 for the prenyltransferase activity of both CPT1/POC1 complexes, as observed in other CPT/Nus1‐like complexes (Fig. 6C) [4, 12, 19].
To study the profile of polyisoprenoids synthesized by Paramecium CPT complexes in yeast, a Saccharomyces rer2Δ srt1Δ nus1Δ triple‐deletion mutant (Degron) supported by the expression of Gl‐UPPS was transformed with the pESC‐URA plasmid containing both the sequences encoding CPT1a or CPT1b and POC1 under the control of the galactose‐inducible GAL1/GAL10 promoter. Control cells were transformed with constructs containing single sequences encoding either the catalytic subunit CPT1 or POC1, as well as with an empty pESC‐URA vector (EV). Next, polyisoprenoids were isolated from these cells, and their profiles were analyzed using the HPLC/UV technique. Analysis of the obtained profiles demonstrated that a triple‐deletion yeast mutant expressing CPT1a or CPT1b individually, as well as POC1 or the empty vector (EV), synthesized only dolichols Dol‐11 and Dol‐12, which are products of Gl‐UPPS (Fig. 6D). Mutant cells co‐expressing either CPT1a or CPT1b with POC1 acquired the ability to biosynthesize a mixture of polyprenols and dolichols with chain lengths ranging from 14 to 17 isoprene units, with Pren/Dol‐15 being the dominant product (Fig. 6D). The first striking observation was that the polyprenol‐to‐dolichol ratio shifted depending on whether CPT1a or CPT1b was involved in their biosynthesis. For CPT1a/POC1, the polyprenol‐to‐dolichol ratio was 1:3, whereas for CPT1b/POC1, it was 1:0.4. This indicates that approximately 75% of polyprenols produced by CPT1a were converted into dolichols, compared to only ~ 30% in the case of CPT1b. Interestingly, cells expressing CPT1b accumulated three times more polyprenol than those expressing CPT1a, whereas dolichol levels were comparable in both cases (Fig. 6E). This can suggest that, regardless of the amount of polyprenol produced, only a defined portion is converted into dolichol.
Yeast expressing complexes CPT1a/POC1 ^ G161A ^ and CPT1b/POC1 ^ G161A ^ are incapable of synthesizing even trace amounts of Pren/Dol other than Gl‐UPPS products, clearly indicating the necessity of the intact RXG sequence for enzymatic activity of the CPT1/POC1 complex (Fig. S2A,B). This result unequivocally confirms that, in accordance with their classification as heteromeric enzymes, CPT1a and CPT1b possess cis‐prenyltransferase enzymatic activity only in complex with POC1 and are capable of synthesizing polyisoprenoid chains.
To explain why, despite the high similarity between the amino acid sequences of CPT1a and CPT1b, we observed differences in the amount of accumulated tested compounds and the ratio of polyprenols to dolichols depending on the enzyme involved, we investigated whether specific residues that differ between CPT1a and CPT1b might influence the efficiency of the synthesis of polyprenols and their conversion to dolichols. Directional mutagenesis was performed, aiming to substitute selected amino acids between CPT1a and CPT1b. We generated triple‐deletion mutant cells (Degron‐Gl‐UPPS) expressing mutated CPT1b/POC1 constructs with single or double substitutions, specifically E55Q, Q105K, M93I, V110F, E156D, N208K, E55Q‐Q105K, M93I‐V110F, and E156D‐N208K. Despite these targeted mutations, the content and profile of polyisoprenoids synthesized by the modified variants of CPT1b were the same as that of wild‐type CPT1b (Fig. S2C–E), with the polyprenol‐to‐dolichol ratio still ranging from 1 : 0.3 to 1 : 0.5 for the individual variants (Fig. S2). These results suggest that the observed differences in polyisoprenoid profiles between CPT1a and CPT1b are likely due to more complex factors beyond the direct influence of these single amino acid residues.
Localization of CPT1s and POC1 proteins in the ER of Paramecium cells
To determine the localization of all three studied proteins in Paramecium, cells were injected with constructs expressing full gene sequences of CPT1a, CPT1b, and POC1 protein under the control of their native promotors. CPT1a and CPT1b were tagged at the C‐terminus with sequences encoding the fluorescent proteins GFP and Tomato, respectively, while POC1 was tagged with a FLAG peptide at the N‐terminus. In the latter case, FLAG was used instead of GFP due to the higher molecular mass of GFP (27 kDa) compared to POC1 (19 kDa), which could lead to nonspecific cellular localization of POC1. For the immunolocalization of POC1‐FLAG, the anti‐FLAG antibody conjugated with Alexa Fluor 488 was used. Protein localization was subsequently tracked using confocal microscopy. The green fluorescence signal of CPT1a‐GFP was detected in the endoplasmic reticulum (ER) like structures (Fig. 7A). To confirm that this localization pattern corresponds to ER, we co‐stained cells with ER‐Tracker™ Red, a highly selective endoplasmic reticulum (ER) dye. As expected, the red fluorescence of the ER marker co‐localized with the green fluorescence of CPT1a‐GFP (Fig. 7B). Similarly, the red fluorescence of CPT1b‐Tomato (Fig. 7C) and the green fluorescence of POC1‐FLAG‐AlexaFluor™488 (Fig. 7D) showed a similar pattern of localization in ER, as observed for CPT1a. These results demonstrate that CPT1a, CPT1b, and POC1 proteins are localized in the endoplasmic reticulum similarly to subunits forming heteromeric complexes implicated in dolichol biosynthesis in other eukaryotes.
Subcellular localization of Paramecium CPT1a, CPT1b, and POC1 by confocal microscopy. (A) The green fluorescence signal of CPT1a‐GFP is localized to endoplasmic reticulum‐like structures. (B) The green fluorescence pattern of CPT1a‐GFP mostly co‐localized with the red fluorescence of ER‐Tracker, a marker of the endoplasmic reticulum (ER). (C) The red fluorescence signal of CPT1b‐Tomato is observed similarly to CPT1a‐GFP in ER structures. (D) The green fluorescence of POC1‐FLAG‐Alexa488 was observed in the ER structures. Scale bars = 10 μm Representative images from two biologically independent and reproducible experiments are shown for each analyzed protein.
Paramecium polyisoprenoid profile
The RNA‐seq analysis of the P. tetraurelia transcriptome revealed that during vegetative growth (VEG), POC1 and CPT1b are expressed at similar levels, while CPT1a is expressed at twice the level of CPT1b and POC1. During intermediate development (Dev2/3), the expression of CPT1a is the highest, being 4.5 and 6.5 times more than POC1 and CPT1b, respectively (Fig. S3 according to ParameciumDB). Our attempts to independently analyze the CPT1a and CPT1b expression using RT‐PCR did not succeed due to the very high sequence similarity between the two genes (86.9% identity). Several primer pairs designed to differentiate between CPT1a and CPT1b isoforms recognized both with equal efficiency. Consequently, based on the published RNA‐seq analysis [45] we decided to focus on the VEG and DEV2/3 stages of Paramecium growth for the analysis of polyisoprenoid content.
Liquid Chromatography Mass Spectrometry (LC–MS) analysis of total polyisoprenoids isolated from Paramecium revealed that during the vegetative and developmental stages, cells accumulated a family of dolichols containing 16 to 18 isoprenoid units (i.u.), with the dominating homolog composed of 17 i.u. (Dol‐17). Based on the retention time of polyprenol and dolichol standards (Collection of Polyprenols, IBB PAS) (Fig. 8C,D), we identified prominent ion peaks at 25.98, 28.35, and 29.69 min, corresponding to the [M + NH_4_] + ions of dolichols containing 16, 17, and 18 isoprene units (C80, C85, and C90), with mass‐to‐charge ratios of m/z 1127.06, m/z 1195.12, and m/z 1263.18, respectively (Fig. 8A). Isotopic profiles of the identified compounds correspond to those calculated from known isotope abundances for C_80_H_132_ONH_4_, C_85_H_140_ONH_4_, and C_90_H_148_ONH_4_, respectively (Fig. 8B). The dolichol profile in Paramecium cells remains consistent across different cell growth stages, while the dolichol accumulation level varies from 1.17 μg of dolichol per 1 g of total protein in the VEG stage, whereas during the DEV2/3 stage, it increases to 2.75 μg of dolichol per 1 g of total protein (Fig. 8E).
Qualitative and quantitative determination of dolichols accumulated in Paramecium cells. (A) Liquid Chromatography Mass Spectrometry (LC–MS) chromatogram of dolichols isolated from Paramecium at DEV2/3 stages. (B) Mass spectra of Dol‐16, ‐17, and ‐18, showing Electrospray Ionization Mass Spectrometry (ESI‐MS) signals corresponding to ammoniated [M + NH4] + ions of dolichol isotopologues. (C) LC‐MS chromatogram of a dolichol mixture isolated from ram testes, which is used as a reference standard (Collection of Polyprenols, IBB PAS). (D) Mass spectra of the standard dolichol mixture (Dol‐16 to Dol‐22), showing ESI‐MS signals corresponding to ammoniated [M + NH4] + ions of dolichol isotopologues. (E) Total content of dolichols in Paramecium cells at vegetative growth (VEG) and intermediate development (DEV2/3) stages. Data are presented as mean ± standard deviation (SD) from three independent experiments (n = 3), each performed in triplicate.
CPT1s and POC1 are essential for the survival of Paramecium
Having identified a family of dolichols in Paramecium cells, we opted to explore which of the complexes, CPT1a/POC1, CPT1b/POC1, or both, plays a role in their biosynthesis. For this purpose, we generated three independent RNAi silenced cell lines targeting the genes CPT1a, CPT1b, or POC1. The expression of each studied gene was individually silenced by feeding Paramecium cultures with bacteria overproducing dsRNA to induce RNA interference. The effect of each RNAi treatment on Paramecium phenotypes, including division rate and survival, was examined at specific time points: 0, 11, 24, 32, and 48 h after initiating feeding with RNAi‐bearing bacteria. At 11 h, the division rate of all RNAi silenced cells was similar to the control cells. After 24 h, the rate of division of all RNAi‐treated cells was delayed to varying extents. Silencing of CPT1a or POC1 genes resulted in a significant delay in divisions after 32 h and ultimately cell death after 48 h (Fig. 9A). This indicates that these genes are essential for the survival of Paramecium. In contrast, silencing the CPT1b gene slightly delayed cell division but did not lead to cell death during 48 h of treatment. Due to high similarity between CPT1a and CPT1b genes, our RNAi constructs shared two 100% identical stretches of DNA – 26 and 33 nt long. Taking into consideration the fact that siRNA are 23 nt long in Paramecium, it is possible that silencing of CPT1a may partially downregulate CPT1b expression and vice versa. Therefore, partial silencing of the second CPT1 from the CPT enzyme pair of Paramecium cannot be excluded.
*Effect of CPT1a, CPT1b, and POC1 silencing on cell survival and dolichol content. (A) The rate of cell division in P. tetraurelia cells after 11, 24, 32, and 48 h of feeding bacteria carrying CPT1a‐RNAi, CPT1b‐RNAi, and POC1‐RNAi constructs. (B) The total content of dolichols (Dol‐16 to Dol‐18) was estimated in Paramecium cells after 24 h of silencing CPT1a, CPT1b, and POC1 genes, respectively, using Liquid Chromatography Mass Spectrometry (LC–MS). Data are presented as mean ± standard deviation (SD) from three independent experiments (n = 3), each performed in triplicate. Asterisks indicate significant differences: *0.01 < P < 0.05 and *0.001 < P < 0.01 (Student's t‐test).
In order to verify whether the observed phenotype is a result of disturbances in the synthesis of dolichols (Dol‐16 to Dol‐18), we performed LC–MS analysis on polyisoprenoids extracted from cells at the 24th hour of silencing. We observed a significant decrease in dolichol content in cells in response to silencing of the tested genes. Specifically, in cells with CPT1a or POC1 silenced, the dolichol content plummeted by 97% compared to the control. Intriguingly, in cells with CPT1b silencing, dolichol content decreased by 87% compared to the control, highlighting that a 13% dolichol pool is sufficient for the cell's survival (Fig. 9B). These observations indicate that CPT1a is responsible for dolichol biosynthesis in Paramecium and its absence is probably not complemented by CPT1b. CPT1a and CPT1b genes might be functionally redundant, but due to the significantly higher expression level, most of the catalytic activity in the cell seems to be attributed to CPT1a.
Impaired dolichol synthesis affects protein glycosylation
To investigate whether impaired dolichol synthesis in Paramecium cells influences protein glycosylation, we isolated proteins from both control cells and cells in which the CPT1a and CPT1b genes were independently silenced. Subsequently, we determined the glycosylation status of these proteins by examining their binding with selected lectins: Concanavalin A (Con A recognizes internal and nonreducing terminal α‐d‐mannosyl and α‐d‐glucosyl groups) and Wheat germ agglutinin (WGA identifies N‐acetyl‐d‐glucosamine and sialic acid). The effect of each RNAi treatment on Paramecium protein N‐glycosylation status was examined at the 11th, 24th, and 32nd hours of gene silencing. The control cells contained glycoproteins of various molecular masses that displayed reactivity with Con A (Fig. 10A) and WGA (Fig. 10B), and these glycosylation patterns remained consistent across all time points assessed. No changes in the glycosylation status of proteins were observed for any variant tested at the 11th hour of silencing. Significant alterations in protein glycosylation patterns were observed for CPT1a‐RNAi cells at the 24th and 32nd hours. Proteins in these cells exhibited notably reduced binding to Con A, while the binding of WGA was considerably inhibited compared to the proteins from control cells. Silencing CPT1b did not induce any alterations in the glycosylation profile of proteins at any time point, regardless of the lectin used in the study (Fig. 10A,B).
*Alterations in protein glycosylation in CPT1a‐ or CPT1b‐silenced cells. Total proteins were isolated from control and CPT1a‐ or CPT1b‐silenced cells at 11, 24, and 32 h of feeding with RNAi‐carrying bacteria. (A, B) SDS/PAGE and western blotting of glycoproteins with lectins Concanavalin A (A) and WGA (B). Ponceau S staining was used as a loading control. (C) Liquid chromatography mass spectrometry (LC–MS) analysis of the total content of oligomannose glycan residues released from Paramecium proteins. The sum of peak areas corresponding to identified glycans was normalized to protein content. Data are expressed as the mean percentage (%) of control values (RNAi untreated cells). Results are presented as means (n = 3) ± standard deviation (SD), and significant differences were assessed using the Student's t‐test (P < 0.05).
To get a deeper insight into the effect of dolichol deficiency resulting from CPT1 silencing on protein glycosylation, we performed a qualitative and quantitative analysis of N‐glycans. To achieve this, glycans were released from total proteins obtained from all Paramecium cell lines under investigation by using PNGase F (N‐Glycosidase F) and subsequently analyzed with HPLC/MS. We identified the signals originating from oligomannose N‐glycans containing 1 to 8 mannose residues (M1‐M8) released from Paramecium proteins. In the cell, mannose residues are added to the formed LLO at the initial stages of N‐linked glycosylation. After 24 h, cells with the silenced CPT1a gene exhibited a significant decrease in total oligomannose glycan content, reaching 12% at 32 h compared to the control. In contrast, cells with the silenced CPT1b gene maintained a level of oligomannose glycans consistent with that of the control line over time. The content of individual oligomannose residues is presented (Fig. S4). Obtained results indicate that protein N‐glycosylation is clearly affected in CPT1a‐RNAi cells in which the level of dolichols is significantly reduced. Interestingly, the decreased content of dolichol, approx. 13% of control observed for CPT1b‐RNAi line (Fig. 9B), is sufficient to support cellular glycosylation machinery, at least during 32 h of silencing (Fig. 10C). In conclusion, efficient silencing of CPT1a inhibits the synthesis of dolichols, which leads to disturbances in protein glycosylation and, consequently, to cell death. In contrast, partial silencing of CPT1b does not deteriorate the protein glycosylation cycle.
Taken together, this report demonstrates that CPT1a, due to its high expression in Paramecium cells, appears to be responsible, when in complex with POC1, for the biosynthesis of dolichols involved in protein glycosylation; whereas the role of CPT1b remains unclear.
Discussion
The major aim of this work is to explore the evolution of the roles of polyisoprenoid lipids in cellular metabolism, specifically by unraveling the mechanism of polyisoprenoid biosynthesis in the ancient unicellular, free‐living eukaryote P. tetraurelia.
Here, we have shown, through enzyme characterization, genetic manipulation, molecular modeling, and polyisoprenoid analysis, that POC1 is a functionally distant orthologue of the human protein NgBR; and when in complex with CPT1a or CPT1b, it is accountable for dolichol biosynthesis in Paramecium cells.
The starting point for our research was the results of a bioinformatics analysis which revealed that both CPTs from Paramecium, CPT1a and CPT1b, belong to the heteromeric CPT‐CS group and are related to each other by duplication. The genome of P. tetraurelia has undergone at least three successive whole‐genome duplications (WGDs) [46, 47]. Gene duplication, a result of WGD, often leads to gene amplification, which can result in the retention of both gene copies, loss of one copy, or differentiation of copies to acquire new functions (neofunctionalization) or subfunctionalization, where each copy retains a subset of the original gene function [46]. Gene expression studies in P. tetraurelia have shown that in the majority of genes after WGD, one of the duplicates is typically lost during evolution. However, retained duplicates may undergo asymmetric evolution, which can lead to pseudogenization (loss of function) or acquisition of new functions [46, 47]. CPT1a, as the more evolutionarily advantageous form, undergoes expression, while CPT1b may, in the future, become a non‐functional pseudogene that is not expressed. In other species from the Paramecium aurelia complex [48], either two CPT encoding genes (e.g., P. pentaurelia, P. biaurelia, P. octaurelia, and P. decaurelia), a single gene (e.g., P. jenningsi, P. sexaurelia, and P. tredecaurelia), or one functional gene and one possible shorter pseudogene (P. sonneborni) (Table S2) are found. All proteins classified into the large CPT family are well conserved evolutionarily and originate from a common ancestor [4]. POC1 in turn belongs to CPT‐AS proteins and is the smallest protein identified so far in this group. However, even smaller, as long as POC1 retains structural elements necessary to perform its functions, it still appears to be a genuine CPT‐AS. CPT‐ASs display much more diversity compared to CPT‐CSs, which might be a consequence of their auxiliary, helper function not requiring maintaining a more rigid structure scaffolding for the catalytic site. CPT‐ASs divergence might also suggest their regulatory roles specific for particular organisms. Although all proteins from this group perform the same function in the heterodimeric complex with CPT‐CS, they have highly divergent amino acid sequences. After all, protein sequences evolve faster than their structures, and even minor sequence similarities often indicate significant structural similarities among proteins [49, 50].
Based on the amino acid sequences of the CPT1a, CPT1b, and POC1 proteins, as well as information on the structures of the yeast orthologues, Rer2 and Nus1, we generated spatial models of the Paramecium proteins and demonstrated that both CPT1s can interact with POC1; that small differences in amino acid sequences between CPT1a and CPT1b do not affect the functionality of the complexes with POC1, as we showed experimentally.
Interaction and activity of POC1 with CPT1a and CPT1b in yeast system
Using the yeast two‐hybrid system (Y2H) and complementation assays with Saccharomyces rer2Δ, srt1Δ, nus1Δ triple‐deletion mutants, we demonstrated that POC1 interacts with CPT1a and CPT1b, forming a two‐component enzyme complex that, upon heterologous expression, synthesizes polyisoprenoid alcohols in vivo.
The length of prenyl chains synthesized in yeast by both Paramecium enzymatic complexes is 2 isoprene residues shorter (Dol‐15 dominating) than those synthesized in the native organism (with Dol‐17 dominating). Similar differences in the length of the dominant prenologue upon overexpression of plant CPTases in a heterologous yeast system have already been observed [6, 16] and may result from different tuning mechanisms regulating the specificity of these enzymes, different lipid compositions of the membranes, different pools of available precursors, and/or differences in access to precursors compared to the homologous system. The subunits, POC1, CPT1a, and CPT1b, are unable to independently synthesize prenyl chains and complement the yeast triple mutation. This is consistent with data for other heteromeric CPT complexes, such as NgBR‐DHDDS, Nus1‐Rer2, Nus1‐Srt1, or Lew1‐AtCPT3 [12, 16, 36, 42], where their full enzymatic activity is also strictly dependent on the mutual interaction of the two components. We have further demonstrated that POC1 function is not complemented by NUS1 overexpression in yeast, and this observation is in line with the lack of detectable sequence similarity between these two proteins beside their C‐terminal fragments. It is different in the case of NgBR, which shares 45% sequence similarity with yeast Nus1 and whose function is complemented by Nus1, compatible with the human DHDDS in a yeast complementation assay [12]. On the other hand, NgBR does not form an active enzyme either with Rer2 or Srt1.
The observation that upon expression in a heterologous system, polyprenols accumulate to a much greater extent in CPT1b than in CPT1a‐expressing cells, while dolichol levels remain comparable between the two, suggests that even subtle sequence differences between the CPT variants may have functional consequences. Our efforts, including substitutions of selected CPT1b amino acids with those present in CPT1a, did not clarify this phenomenon. Therefore, further studies are necessary to elucidate the structural and biochemical determinants underlying polyprenol biosynthesis and conversion of polyprenols to dolichols by polyprenol reductase and to clarify how these differences impact downstream dolichol‐dependent pathways. In this context, an interesting question arises regarding how the conversion of polyprenol to dolichol is regulated to ensure the maintenance of appropriate dolichol levels regardless of fluctuations in polyprenol abundance. The persistence of normal dolichol levels despite an excess of polyprenol has previously been observed in fibroblasts from a patient with a non‐functional polyprenol reductase (SRD5A) [51], suggesting not only the existence of an alternative pathway for dolichol biosynthesis but also the presence of a regulatory mechanism that limits dolichol production. The use of CPT1a‐POC1 and CPT1b‐POC1 complexes as experimental models for studying dolichol deficiency or conditions characterized by a high polyprenol/dolichol ratio [52], despite normal dolichol levels, may provide valuable insights into the molecular basis of glycosylation disorders, particularly in DHDDS‐CDG and NUS1‐CDG—a congenital disorder of glycosylation resulting from mutations in either the DHDDS or NUS1 gene [12, 53, 54, 55].
This work also prompts an important question regarding the enzymatic efficiency of the conversion of polyprenols to dolichols, specifically in the context of products derived from CPT1a and CPT1b.
Role of the ‐RXG‐ motif
The POC1 protein, despite its remnant similarity to CPT‐AS from other organisms, possesses a highly conserved RXG motif at its C‐terminus, presence of which is essential for the enzymatic activity of all CPTases, both homo‐ and heteromeric. This is consistent with the hypothesis of the significance of the RXG motif for the enzymatic activity of the CPT1a‐POC1 and CPT1b‐POC1 complexes. The G161A substitution in the RXG motif of POC1 completely inhibited the biosynthesis of CPT1‐POC1‐derived polyisoprenoids in triple‐deletion mutant which was confirmed by FOA test and analysis of lipid products. The G292A substitution in the RXG motif of human NgBR similarly resulted in the severe inhibition of the enzymatic activity of the DHDDS‐NgBR complex and lowering affinity for IPP [19]. Studies on the crystal structure of the human NgBR/DHDDS complex have shown that NgBR stabilizes DHDDS through dimerization and participates in the enzyme's active site through its C‐terminal RXG motif, clearly stimulating the activity of CPTase. NgBR itself lacks key catalytic residues characteristic of CPTases and has a distorted ‘CPT active site’ incapable of substrate binding [42]. Similarly, studies on the crystal structure of the homomeric bacterial CPT from Thermobifida fusca, responsible for the synthesis of dodecaprenyl diphosphate, revealed a dimeric architecture where the C‐terminal RXG motif is critical for subunit interactions and plays a role in the catalytic activity of the enzyme [56]. Comparison of NgBR/DHDDS with homomeric CPTase structures revealed that subunits of heteromeric CPTases involved in the synthesis of long‐chain polyprenols evolved from their homomeric ancestors [42].
Subcellular localization of CPT1a, CPT1b and POC1
We have shown that all three proteins investigated in this study are localized in the endoplasmic reticulum (ER). These results are consistent with literature data, where most of the previously described eukaryotic CPTs are localized in the ER. The presence of CPT‐CSs and their CPT‐ASs in ER membranes has been observed in humans [11], yeast [12], and plants, for example, Arabidopsis [16], lettuce [18], tomato [17], and rubber tree [57], as well as parasites G. lamblia [10], and Trichomonas vaginalis [32].
Interestingly, in silico analyses clearly indicated that neither CPT1a and CPT1b nor POC1 have predicted transmembrane domains (TM) that could anchor the two‐component complex in the ER membrane. This raises the question of how CPT1a/b‐POC1 complexes localize in the ER membrane and participate in dolichol biosynthesis. The idea that the TM domain in some CPT‐AS (NgBR/Nus1/HRBP) serves as docking subunits for the complex to the ER membrane is debatable. It has been postulated that human NgBR can exist in two different conformations; when the C‐terminal part of the protein is oriented toward the ER lumen, it interacts with NPC2, and when oriented toward the cytosol, it binds DHDDS (named also hCIT) and regulates its CPTase activity [11, 58]. However, the loss of NgBR does not lead to changes in the relative amount of DHDDS associated with ER membrane fractions [11]. Similarly, in the rubber tree, the role of HRBP in docking cytoplasmic CPT to the ER membrane has not been confirmed. The TM domain is not present in the Methanosarcina heteromeric CPT or homomeric Gl‐UPPS. Both of those enzymes are present in the membrane fractions and are able to complement yeast strains lacking endogenous CPT subunits [10, 19]. In fact, the membrane sensor at the N‐terminus of DHDDS and its orthologues may be crucial for the anchoring of the CPT complex to the ER membrane [42]. Although Hevea CPT‐CS interacts with HRBP (CPT‐AC) on the ER membrane, it then moves to the cell membrane [59], indicating that other proteins may be required to maintain the CPT‐HRBP complex on the ER membrane [57]. Our studies on the role of POC1 in the CPT enzymatic complex show that it is a structural component of the complex necessary for the biosynthesis of the prenyl chain. Based on predictions of the absence of TM domains, Paramecium CPT may be docked to the ER via the N‐terminal lipid sensor domain in CPT1. However, it is tempting to speculate that interaction with at least one more protein partner is plausible. It is also worth noting that the TM domain in Nus1 and its orthologues is present in the enzymes with dual localization both to the ER and lipid droplets [26, 60]. Simply, the synthesis of dolichol in Paramecium may be limited to the ER.
Characterization of polyisoprenoid biosynthesis in Paramecium cells
In Paramecium cells, regardless of the developmental stage, we identified a family of long‐chain dolichols dominated by Dol‐17. Long‐chain polyisoprenoids composed of 15 to 19 i.u. are also accumulated in Plasmodium falciparum cells, but their profile depends on the parasite's developmental stage [31]. The potential PfCPT does not have an RXG motif and is therefore classified into the group of heteromeric CPT‐CSs. All known heteromeric CPTs preferentially synthesize a family of long‐chain‐length polyprenol diphosphates, which are reduced to dolichols (e.g., human DHDDS‐NgBR, dominating Dol‐19; Arabidopsis AtCPT3‐Lew1, Dol‐16; tomato SlCPT3‐SlCPTBP, Dol‐16; yeast Rer2‐Nus1 and Srt1‐Nus1, respectively Dol‐15 and Dol‐21). Other pathogenic protozoa accumulate medium‐chain polyisoprenoids: G. lamblia and Trypanosoma brucei dolichols composed of 11 and 12 i.u. [4, 61], T. cruzi a single dolichol with 13 i.u. [62], and Crithidia fasciculata a single polyprenol with 11 i.u. [63]. Both Gl‐UPPS and potential Tc‐CPT belong to the group of homomeric CPTs with high homology to bacterial undecaprenyl diphosphate synthase (UPPS).
We have also found that in the synthesis of the long‐chain prenyl chain in Paramecium, mostly CPT1a in complex with POC1 is involved. RNAi silencing of the studied genes has shown that silencing of POC1 leads to a significant decrease in the dolichol content in Paramecium cells, which consequently results in lethality, similar to the loss of Nus1 in S. cerevisiae [64] and Lew1 in A. thaliana [65]. RNAi silencing of CPT1a also substantially reduces the dolichol content in Paramecium cells and is lethal for them. In CPT1b‐RNAi cells, the level of dolichols is also decreased, but it does not affect the viability of the cells, at least within the observation time window. This reduction in dolichol content might be due to nonspecific silencing of CPT1a by the CPT1b silencing construct, because of the very high similarity between the nucleotide sequences of both CPTs. This provides evidence that CPT1a is the primary CPT in Paramecium. Single deletions of yeast CPT‐CS, RER2, responsible for the synthesis of dolichols for glycosylation reactions, result in severe growth defects and glycosylation deficiency. However, the cells remain viable [13, 14, 15, 66] due to the induction of transcription of the SRT1 gene. Srt1 normally is expressed during sporulation and is responsible for the synthesis of polyprenols in specialized lipid droplets crucial for the synthesis of chitin [26]. CPT1b may have glycosylation‐independent function, which remains to be elucidated.
Under optimal culture conditions, CPT1b in Paramecium cells is expressed at a very low level, suggesting that it may function as a pseudogene and/or that its expression occurs only under specific conditions, such as environmental stress. In A. thaliana, deletion of the plastidial CPT7, which synthesizes polyprenols (Pren‐9, ‐10, and ‐11) present in thylakoid membranes, does not affect the morphology or functioning of the plants [8]. However, under elevated temperature conditions, its expression becomes essential for plant survival, as we have demonstrated that the presence of polyprenols in thylakoids is crucial for plant adaptation to heat stress [67].
Finally, we provided evidence that silencing of CPT1a, consistent with dolichol deficiency, leads to defects in protein N‐glycosylation. Defective protein glycosylation is considered the major reason for Paramecium cell lethality. Dolichyl phosphate is essential in various glycosylation pathways for transferring oligosaccharide and monosaccharide residues. The appropriate pool of Dol‐P in cells comes from both de novo synthesis and the recycling pathway. After the transfer of the oligosaccharide onto the appropriate protein, Dol‐PP is released into the lumen of the ER, where it is dephosphorylated to Dol‐P. Dol‐P then returns to the cytoplasmic side of the ER, where it can be reused for LLO biosynthesis [52, 68]. The mechanism by which Dol‐P translocates across the ER membrane in new rounds of Dol‐PP‐glycan synthesis is unknown [2]. Silencing of CPT1a in Paramecium cells results in inhibition of de novo dolichol synthesis, but the pool of Dol‐P from the recycling pathway, even though gradually decreased, is likely functional, delaying disruptions in protein glycosylation. Probably that is why we observe disturbances in the glycosylation of proteins only after 1 day of feeding RNAi‐bearing bacteria.
Future directions for POC1 identification
We hypothesize that identification of POC1 in Paramecium may facilitate the search for functional orthologues of NgBR/Nus1 in pathogenic protists, for example, Plasmodium falciparum, Toxoplasma gondii, Trichomonas vaginalis. In the proteome of the malaria parasite, a protein with high homology to DHDDS/Rer2 has been identified, whereas the homolog of NgBR/Nus1 is absent. In this case, it may be crucial to focus on structural similarities rather than amino acid sequences. Detecting CPT‐AS proteins in Plasmodium and other pathogenic protists, which could differ significantly from human NgBR and its orthologues in animals, may provide highly specific targets for antiparasitic therapy by selectively inhibiting polyprenol synthesis. Polyprenol synthesis is critical for the fluidity of the apicoplast membrane and its biogenesis [69]. Furthermore, the search for new antimalarial therapies is crucial in light of recent reports of Plasmodium falciparum resistance to artemisinin, the key antimalarial drug, in children with severe malaria in Africa. This, along with earlier studies indicating the emergence of partial resistance to artemisinin, suggests that the parasite is evolving mechanisms to evade the established first‐line treatment for malaria [70].
Conclusion
In summary, we identified POC1 in P. tetraurelia as an evolutionary distant functional orthologue of human NgBR and yeast Nus1. The interaction between POC1 and CPT1a is essential for dolichol biosynthesis, protein glycosylation, and the viability of Paramecium cells. Our studies pave the way for the identification and characterization of two‐component CPTs in other protists.
Materials and methods
Bioinformatic methods
Protein sequence searches were performed with HHSEARCH [37], a meta‐profile comparison method capable of identifying similarities even between remotely homologous proteins. 3D structure modeling was done with the use of alphafold3 [43], the newest version of the protein structure prediction method alphafold, which uses deep neural networks for reliable protein structure prediction. The phylogenetic tree of representative proteins was constructed using iq‐tree [40] (LG + G4 model) estimated by automatic model selection [71], 1000 replicates for the SH‐like approximate likelihood ratio test [72] based on multiple sequence alignment calculated using mafft [41] (localpair, 1000 iterations).
Biological material
- Paramecium tetraurelia 51new nd7‐1 (mt 7) [73]
- Escherichia coli HT115 (DE3): F‐, mcrA, mcrB, IN(rrnD‐rrnE)1, rnc14::Tn10(DE3 lysogen: lacUV5 promoter‐T7 polymerase) (IPTG‐inducible T7 polymerase) (RNAse III minus) [74]
- Escherichia coli SURE: F′[proAB+ lacIq lacZΔM15 Tn10(TetR)] endA1 glnV44 thi‐1 gyrA96 relA1 lac recB recJ sbcC umuC::Tn5(KanR uvrC e14–(mcrA–) Δ(mcrCB‐hsdSMR‐mrr)171) (Stratagene, La Jolla, CA, USA);
- Escherichia coli MH1: araD lacX74 galU hsdR hsdM rpsL [75]
- Escherichia coli DH5α: fhuA2 lac(del)U169 phoA glnV44 Φ8' lacZ(del)M15 gyrA96 recA1 relA1 endA1 thi‐1 hsdR17 (ATCC–LGC Standards, Lomianki, Poland)
- Klebsiella pneumoniae (kindly donated by Dr Eric Meyer, Institut de Biologie de l'Ecole Normale Superieure IBENS, Paris, France)
- Saccharomyces cerevisiae KG405 (rer2Δ srt1Δ nus1Δ/pNEV‐GlcisPT)
- Saccharomyces cerevisiae Degron‐GlcisPT (nus1Δ rer2Δ srt1Δ): MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 lys2Δ0; rer2ΔkanMX4, srt1Δhis3MX6 P^NUS1^‐Degron‐GlcisPT‐hphMx4
- Saccharomyces cerevisiae AH109: MATa, trp1‐901, leu2‐3, 112, ura3‐52, his3‐200, gal4Δ, gal80Δ, LYS2::GAL1 UAS‐GAL1 TATA‐HIS3, GAL2 UAS‐GAL2 TATA‐ADE2, URA3::MEL1 UAS‐MEL1 TATA–lacZ (Clontech Laboratories/Takara Bio, Palo Alto, CA, USA).
Construction of CPT1a, CPT1b, and POC1 yeast expression vectors
Paramecium TAA and TAG triplets code for glutamine (Q), and only the TGA codon is recognized as a termination signal [76]. The coding sequence (CDS) of POC1 contains 14 TAA codons, while both CPT1a and CPT1b contain 11 TAA codons each. To express POC1, CPT1a, and CPT1b in S. cerevisiae cells, three appropriate genetic constructs in the pDONR221 vector were synthesized by GeneArt (Thermo Scientific, Regensburg, Germany). The TAA codons were converted to CAA or CAG according to yeast‐preferred codon usage. Additionally, in all constructs, every codon was optimized for yeast‐preferred codon usage (Fig. S5).
Yeast two‐hybrid assay
The coding sequence of POC1, optimized for yeast, was ligated into the pGBKT7 vector as bait, while the coding sequences of CPT1a or CPT1b, also optimized for yeast, were ligated into the pGADT7 vector as the prey. Additionally, reverse cloning was performed, with POC1 being ligated into the pGADT7 vector and CPT1a or CPT1b into the pGBKT7 vector. Plasmids of the prey and bait were mated and transformed into the yeast strain AH109, and the mating cultures were spread on selective medium. Y2H assay was performed as described previously [16].
5‐Fluoroorotic Acid (5‐FOA) assay
To analyze the putative activity of the heteromeric CPT1‐POC1 complex, S. cerevisiae strain KG405 (nus1Δ rer2Δ srt1Δ), containing the Gl‐UPPS gene on a URA3‐marked plasmid [12], was transformed with pKG‐GW1 vectors carrying CPT1a or CPT1b (leucine selection) and pKG‐GW2 vectors carrying WT POC1 or POC1^G161A^ (methionine selection), or with the corresponding empty vectors as negative controls. As positive controls for the assay, the plasmids containing WT CPT from T. cruzi or its mutated variant ^114^Tc‐CPT, as well as Rer2 or Nus1 from S. cerevisiae, were used. Transformed yeast cells were selected on synthetic medium lacking uracil and leucine, uracil and methionine, or uracil, methionine, and leucine. They were then streaked onto synthetic medium containing all amino acids, nucleotide supplements, and 1% (w/v) 5‐FOA (Zymo Research, Irvine, CA, USA) as well as onto YPD plates. Plates were incubated for up to 7 days at 30 °C as previously described [12].
Functional complementation of the yeast
rer2Δ srt1Δ nus1Δ triple‐deletion mutant
To simultaneously express CPT1a and POC1 or CPT1b and POC1 in yeast mutant cells Degron‐GlUPPS (nus1Δ rer2Δ srt1Δ), coding sequences of POC1, CPT1a, and CPT1b, optimized for yeast, were cloned into the pESC‐URA yeast dual expression vector (Agilent Technologies, Santa Clara, CA, USA) in appropriate order according to the manufacturer's protocol. Transformant selection and growth were performed as described previously [6].
Construction of GFP or tomato or FLAG fusion transgenes
Plasmids pCPT1a::CPT1a‐GFP and pCPT1b::CPT1b‐Tomato were obtained by an overlapping PCR method [77] in pCRscript vector (Novagen, Madison, WI, USA). Constructs contain putative promoter regions (215 and 171 bp, respectively), open reading frame, EGFP coding sequence optimized for Paramecium codon usage [78] or double Tomato sequence (kindly provided by E. Meyer & B. Saudemont) fused directly before the stop codon and putative terminator (46 and 101 bp); (genomic coordinates of cloned fragments: CPT1a: nucleotides from 437362 to 438349 of the acc. no. CAAL01001200.1, CPT1b: nucleotides from 24708 to 25742 of the acc. no. CAAL01000854.1).
Construct pPOC1::FLAG‐POC1 containing the putative POC1 promoter (216 bp), a sequence encoding 3xFLAG, and the POC1 open reading frame as well as putative terminator region (93 bp) in the pMK‐RQ plasmid was synthesized by GeneArt (Thermo Scientific, Regensburg, Germany). Paramecium regulatory sequences were shown to be very short, sometimes even significantly shorter than 100 bp [79] and even sequences of 26 bp were shown to drive the expression of the genes.
The schemes of the constructs are shown in (Fig. S6). Primers used to prepare the appropriate constructs and PCR reaction conditions are presented in Table S3.
Paramecium transformation
Plasmids pCPT1a::CPT1a‐GFP and pCPT1b::CPT1b‐Tomato were digested with Psp1406I, whereas pPOC1::FLAG‐POC1 was digested with Alw44I. Digested plasmids mixed with the plasmid directing the expression of the ND7 gene (designed to complement the Nd7‐1 mutation) cut with Psp1406I were precipitated, resuspended to 5 μg·μL^−1^ and introduced into the vegetative macronucleus of P. tetraurelia 51 nd7‐1 cells. Paramecium cells were microinjected in Dryl solution containing 0.2% bovine serum albumin under an oil film. Transformants were screened first for their ability to discharge trichocysts and then for the fluorescence of GFP or Tomato or FLAG‐tag expression. Clones expressing the fusion protein and showing a growth rate similar to untransformed cells were chosen for further analyses.
Gene silencing in Paramecium by feeding
Plasmids used for T7Pol‐driven dsRNA production in silencing experiments were obtained by cloning PCR products from each gene using plasmid L4440 and E. coli strain HT115, as previously described [80]. Sequences used for silencing of CPT1a, CPT1b, and POC1 were segments 238–722 of PTET.51.1.P0120334 (CPT1a), 258–709 of PTET.51.1.P0040087 (CPT1b), and 231–489 of PTET.51.1.P0200327 (POC1). Silencing media were prepared basically as described in [81], by dissolving LB precultures of the appropriate bacterial strains 10 times into WGP medium containing 0.1 mg·mL^−1^ ampicillin. Following 6–8 h of shaking at 37 °C, bacterial cultures were diluted six‐fold into the same medium containing 0.4 mm IPTG to induce dsRNA synthesis. After overnight induction at 37 °C, all silencing media were supplemented with 0.8 μg·mL^−1^ β‐sitosterol before use.
Paramecium cell culture
All experiments were carried out with the entirely homozygous strain 51 new of P. tetraurelia. Cells were grown in wheat grass powder (WGP) (Pines International, Lawrence, KS, USA) infusion medium bacterized the day before use with K. pneumoniae, unless otherwise stated, and supplemented with 0.8 mg·mL^−1^ β‐sitosterol (EMD Millipore Corporation, Billerica, MA, USA). Cultivation and autogamy were carried out at 27 °C as described [82, 83].
Subcellular localization of CPT subunits – confocal microscopy
For CPT1a‐GFP and CPT1b‐Tomato localization, Paramecium cells were transformed and cultured as described above.
For POC1‐FLAG localization, Paramecium cells were premeabilized and fixed in two steps. First, with a buffer containing 1% formaldehyde, 2.5% Triton X‐100, 4% sucrose, and PHEM (10 mm EGTA, 25 mm HEPES, 2 mm MgCl_2_, 60 mm PIPES pH 6.9) for 30 min, and next with a buffer containing 4% formaldehyde, 1.2% Triton X‐100, 4% sucrose, and PHEM for 10 min [84]. Subsequently, cells were washed three times with TBS‐0.1% Tween 20/BSA and incubated for 2 h with the primary antibody (Mouse Monoclonal ANTI‐FLAG® M2 antibody, Sigma‐Aldrich, St. Louis, MO, USA) diluted 1 : 200. After washing three times the secondary antibody (Goat anti‐Mouse IgG (H + L), Superclonal™ Recombinant Secondary Antibody, Alexa Fluor™ 488, Thermo Fisher Scientific, Waltham, MA, USA) diluted 1 : 500 was applied for 45 min in the dark.
Additionally, for the proper assessment of the subcellular localization of studied proteins, cells were stained with ER‐Tracker (an endoplasmic reticulum marker, Thermo Fisher Scientific, Eugene, OR, USA) or DAPI (a nuclei marker, Thermo Fisher Scientific, Eugene, OR, USA).
To confocal analysis, all samples were mounted in VECTASHIELD® PLUS (Vector Laboratories, Burlingame, CA, USA).
Confocal Images were taken under a Nikon C1 confocal system built on TE2000E with 408, 488, and 543 nm laser excitations for DAPI (450/35 nm emission filter), GFP/Alexa Fluor 488 (515/30 nm emission filter), and Tomato/Oil Red O (605/75 emission filter), respectively.
Polyisoprenoid extraction and purification
Yeast polyisoprenoid
Yeast cells were harvested by centrifugation of 100 mL of yeast culture at 3900 ** g ** for 10 min at room temperature, washed with water, and suspended in 10 mL of a solution containing 25% KOH and 65% ethanol supplemented with 10 μg of the internal standard Pren‐21 (The Collection of Polyprenols, IBB PAS, Warsaw, Poland) and hydrolyzed for 1 h at 95 °C. Nonsaponifiable lipids were extracted with hexane, purified on silica gel 60 columns using isocratic elution with 10% diethyl ether in hexane, evaporated to dryness, and dissolved in isopropanol. Polyisoprenoids were analyzed by high‐performance liquid chromatography with ultraviolet detection (HPLC/UV) according to the method described previously [42].
Paramecium polyisoprenoid
Paramecium cells were harvested by centrifugation of growth medium containing 1 000 000 cells (~ 1000 cells per mL) at 600 ** g ** for 1 min at room temperature and washed twice with 10 mm Tris/HCl (pH 7.4). Cell pellets were frozen in liquid nitrogen and stored at −80 °C. Paramecium cell pellets were thawed on ice and suspended in 12.5 mL of a solution containing 25% KOH and 65% ethanol supplemented with 1 μg of the internal standard Dol‐13 (The Collection of Polyprenols, IBB PAS, Warsaw, Poland), hydrolyzed for 1 h at 95 °C. Nonsaponifiable lipids were extracted with hexane, evaporated to dryness, and dissolved in isopropanol. Polyisoprenoids were analyzed by Liquid Chromatography Mass Spectrometry (LC–MS).
Polyisoprenoid analysis by liquid chromatography mass spectrometry (LC–MS)
LC–MS analysis of polyisoprenoids content was performed using Acquity UPLC chromatograph (Waters, Milford, MA, USA) coupled with Xevo TQ mass spectrometer (Waters). Chromatographic separation of polyprenols and dolichols was acquired using reversed phase Accucore C30 column (150 × 2.1 mm, 2.6 μm) from Thermo Fisher Scientific Waltham, MA, USA. Column oven was heated to 60 °C. Mobile phase A consisted of 10 mm ammonium formate with 0.1% formic acid in ACN/MeOH/H_2_O (2 : 1 : 1, v/v/v) and mobile phase B consisted of 10 mm ammonium formate with 0.1% formic acid in IPA/ACN (9 : 1, v/v). The total run time of chromatographic separation was 47.0 min. Exact gradient parameters were as given: time = 0 min (t = 0.0), 90% A, 0.5 mL·min^−1^; t = 3.0, 40% A, 0.6 mL·min^−1^; t = 8.0, 35% A, 0.6 mL·min^−1^; t = 13.0, 30% A, 0.6 mL·min^−1^; t = 18.0, 25% A, 0.6 mL·min^−1^; t = 23.0, 20% A, 0.6 mL·min^−1^; t = 28.0, 15% A, 0.6 mL·min^−1^; t = 33.0, 10% A, 0.6 mL·min^−1^; t = 38.0, 8% A, 0.6 mL·min^−1^; t = 42.0, 5% A, 0.6 mL·min^−1^; t = 42.1, 90% A, 0.5 mL·min^−1^. Mass spectrometer operated in Selected‐Ion Monitoring mode (SIM) with positive Electrospray Ionization (ESI^+^). Capillary voltage was set to 2.0 kV, cone voltage 25.0 V, desolvation temperature 450 °C, desolvation gas flow 800 (L·h^−1^) and cone gas flow 150 (L·h^−1^). Data was collected for ammonium adducts of polyprenols and dolichols. Four isotope masses were analyzed for each respective analyte to cover the isotope distribution spectrum. Ammonia adduct masses were quantified using enviPat Isotope Pattern Calculator [85]. Acquired spectra were processed using targetlynx xs 4.1 (Waters) software.
Analysis of protein glycosylation status – lectin blotting
Total proteins were isolated from control and RNAi silenced Paramecium cells by homogenization on ice using a Potter‐Elvehjem homogenizer with 2.5 mL of RIPA buffer. The homogenates were centrifuged at 16 000 ** g ** for 30 min at 4 °C to remove cell debris. Protein concentrations were determined using a DC Protein Assay (Bio‐Rad Laboratories, Hercules, CA, USA). Protein samples (25 μg per lane) were resolved on a 10% SDS polyacrylamide gel and transferred to an ECL nitrocellulose membrane by wet transfer (Mini Trans Blot, Bio‐Rad Laboratories). The efficiency of the transfer was assessed by staining the membrane with 0.5% Ponceau Red in a 3% TCA solution.
For Concanavalin A, the membranes were incubated in PBS buffer containing 2% Tween20 for 2 min, washed 2 times in PBS, and probed overnight with Concanavalin A labeled with horseradish peroxidase (final concentration 0.75 μg·mL^−1^; Sigma‐Aldrich, St. Louis, MO, USA). The membranes were then washed 6 times with PBS and developed for 5 min with an ECL reagent (SuperSignal West Pico Chemiluminescent Substrate; Thermo Scientific, Rockford, IL, USA) according to the manufacturer's instructions. Signals were visualized on X‐ray film (Amersham).
For WGA (Wheat Germ Agglutinin), the membranes were blocked in RIPA buffer TBS‐T for 1 h and then incubated with WGA labeled with horseradish peroxidase (final concentration 0.1 μg·mL^−1^; Sigma‐Aldrich, St. Louis, MO, USA) for 2 h. After washing 5 times with RIPA buffer, membranes were developed for 5 min with an ECL reagent. Signals were visualized as described above.
Liquid chromatography mass spectrometry (LC–MS) glycan analysis
Glycoproteins were de‐glycosylated according to the method described by Link‐Lenczowski et al. [86]. Briefly, equal volumes (100 μL) of total protein extracts were precipitated using chloroform/methanol and centrifuged. Protein pellets were solubilized in 40 μL of denaturing buffer (0.5% SDS, 40 mm DTT) and incubated at 90 °C for 5 min. After that, the deglycosylation buffer was added (2% NP‐40, 100 mm Sodium Phosphate pH 7.5, PNGase F 500 000 U·mL^−1^; New England Biolabs, Ipswich, MA, USA) and the samples were incubated overnight at 37 °C. The released N‐glycans were purified using graphitized carbon solid phase extraction (SPE) columns (Supelclean™ ENVI‐Carb™, Sigma‐Aldrich, Darmstadt, Germany) according to Packer et al. [87]. The glycans were eluted from the columns using 0.05% TFA in 25% acetonitrile and dried down in a vacuum centrifuge. Dried glycans were labeled with a fluorescent tag (2‐aminobenzamide, 2‐AB) by a 3 h incubation with labeling buffer (5% 2‐AB, 6% sodium cyanoborohydride in acetic acid/DMSO 3 : 7) at 65 °C. After that, 1 mL of 97% acetonitrile was added to each sample. Labeled glycans were purified on solid‐phase extraction columns (speed Amide‐2, Applied Separations, Allentown, PA, USA) as previously described [86] and eluted with HPLC pure water. Prior to UPLC–MS, the labeled N‐glycans were dried down by lyophilization. 2‐AB‐labeled N‐glycans were analyzed by liquid chromatography mass spectrometry (LC–MS) on an Aquity I‐Class Plus UPLC system with an in‐line fluorescent detector coupled through an electrospray ion source to the Vion® IMS‐QToF high resolution mass spectrometer (Waters) according to the method described previously [88]. Briefly, N‐glycans were separated on HILIC‐UPLC on an ACQUITY UPLC Glycan BEH Amide Column, 130 Å, 1.7 μm, 2.1 × 150 mm (Waters) at 60 °C with the following gradient conditions: solvent A was 50 mm ammonium formate pH 4.4, solvent B was 100% acetonitrile (Chemsolv, Roanoke, VA, USA); time = 0 min (t = 0.0), 25% A, 0.4 mL·min^−1^; t = 35.0, 46% A, 0.4 mL·min^−1^; t = 36.0, 100% A, 0.2 mL·min^−1^; t = 39.5, 100% A, 0.2 mL·min^−1^; t = 43.5, 25% A, 0.2 mL·min^−1^; t = 47.6, 25% A, 0.4 mL·min^−1^; t = 55.0, 25% A, 0.4 mL·min^−1^. The fluorescence detector was configured with excitation and emission wavelengths of 330 and 420 nm, respectively. Electrospray ionization settings included a capillary voltage of 3.0 kV, a source temperature of 120 °C, a desolvation temperature of 350 °C, and a desolvation gas flow rate of 800 L·h^−1^. Mass spectrometry was performed in positive ion ToF MS mode, recording ion signals within the m/z range of 600 to 2000. External calibration of the UPLC runs was carried out using a 2‐AB‐labeled glucose homopolymer standard (catalog number 186006841, Waters). Chromatograms were processed through automatic integration, followed by manual refinement to delineate distinct peaks. Glycan structures were identified based on their glucose unit (GU) values and exact masses. Chromatographic and mass spectrometric data were evaluated using the Waters UNIFI scientific information system, in conjunction with the integrated Waters Glycan GU Scientific Library. For quantification purposes, the cumulative area of all glycan peaks was normalized relative to the protein concentration.
Conflict of interest
The authors declare no conflict of interest.
Author contributions
AO designed and performed research, analyzed data; KSt performed bioinformatics analysis, wrote manuscript; KSz performed research; PS analyzed data; PL‐L performed research and analyzed data; KAG designed and performed research and contributed tools; MR analyzed data; JKN designed and performed research, analyzed data; LS designed and performed research, analyzed data, and wrote manuscript. All authors read, contributed to the editing of, and approved the final submitted version of the manuscript.
Supporting information
Fig. S1. Amino acid residues differences between CPT1a and CPT1b, in the context of interaction with POC1. Fig. S2. Profile of polyisoprenoids isolated from the S. cerevisiae triple‐deletion mutant transformed with plasmid expressing wild‐type or mutated genes encoding the CPT1a/POC1 or CPT1b/POC1 complexes. Fig. S3. Expression of POC1, CPT1a, and CPT1b during live cycle of Paramecium tetraurelia. Fig. S4. Alterations in the protein glycosylation profile in control and CPT1a‐ or CPT1b‐silenced cells. Fig. S5. DNA sequences of native (NA) Paramecium CPT1a, CPT1b and POC1 and optimized for yeast‐preferred codon usage (YC). Fig. S6. Schematic diagrams of the construction of CPT1a‐GFP and CPT1b‐Tomato in the pCR‐Script SK+ vector and FLAG‐POC1 in the pMK‐RQ vector, which were used for the localization test. Table S1. Names and accession numbers of all CPT sequences used for phylogenetic analysis in Fig. 2. Table S2. Potential CPTs in various Paramecium species. Table S3. Primers used for localization, yeast complementation and mutagenesis construct generation. Data S1. The supplementary dataset is available on Zenodo at http://doi.org/10.5281/zenodo.16360136.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Hartley MD & Imperiali B (2012) At the membrane frontier: a prospectus on the remarkable evolutionary conservation of polyprenols and polyprenyl‐phosphates. Arch Biochem Biophys 517, 83–97.22093697 10.1016/j.abb.2011.10.018PMC 3253937 · doi ↗ · pubmed ↗
- 2Buczkowska A , Swiezewska E & Lefeber DJ (2015) Genetic defects in dolichol metabolism. J Inherit Metab Dis 38, 157–169.25270028 10.1007/s 10545-014-9760-1PMC 4281381 · doi ↗ · pubmed ↗
- 3Wilson MP , Kentache T , Althoff CR , Schulz C , de Bettignies G , Mateu Cabrera G , Cimbalistiene L , Burnyte B , Yoon G , Costain G et al. (2024) A pseudoautosomal glycosylation disorder prompts the revision of dolichol biosynthesis. Cell 187, 3585–3601.e 22.38821050 10.1016/j.cell.2024.04.041PMC 11250103 · doi ↗ · pubmed ↗
- 4Grabińska KA , Park EJ & Sessa WC (2016) Cis‐Prenyltransferase: new insights into protein glycosylation, rubber synthesis, and human diseases. J Biol Chem 291, 18582–18590.27402831 10.1074/jbc.R 116.739490 PMC 5000101 · doi ↗ · pubmed ↗
- 5Yamada Y , Fukuda W , Hirooka K , Hiromoto T , Nakayama J , Imanaka T , Fukusaki E & Fujiwara S (2009) Efficient in vitro synthesis of cis‐polyisoprenes using a thermostable cis‐prenyltransferase from a hyperthermophilic archaeon Thermococcus kodakaraensis . J Biotechnol 143, 151–156.19583987 10.1016/j.jbiotec.2009.06.026 · doi ↗ · pubmed ↗
- 6Surowiecki P , Onysk A , Manko K , Swiezewska E & Surmacz L (2019) Long‐chain Polyisoprenoids are synthesized by At CPT 1 in Arabidopsis thaliana . Molecules 24, 2789.31370240 10.3390/molecules 24152789 PMC 6695881 · doi ↗ · pubmed ↗
- 7Surmacz L , Plochocka D , Kania M , Danikiewicz W & Swiezewska E (2014) Cis‐Prenyltransferase At CPT 6 produces a family of very short‐chain polyisoprenoids in planta. Biochim Biophys Acta 1841, 240–250.24291644 10.1016/j.bbalip.2013.11.011 · doi ↗ · pubmed ↗
- 8Akhtar TA , Surowiecki P , Siekierska H , Kania M , Van Gelder K , Rea KA , Virta LKA , Vatta M , Gawarecka K , Wojcik J et al. (2017) Polyprenols are synthesized by a Plastidial cis‐Prenyltransferase and influence photosynthetic performance. Plant Cell 29, 1709–1725.28655749 10.1105/tpc.16.00796 PMC 5559739 · doi ↗ · pubmed ↗