Guayule plant seeds (PI478663) were obtained from the United States Department of Agriculture (USDA). Seeds were soaked in fresh water for 4 days before germination, and potted to a mixture of 35% autoclaved sands and 65% potting soil. The germinated plants were grown in a growth chamber under conditions of 16 h light, 8 h dark, at 25°C.

Two months old guayule plants were ground in liquid nitrogen using mortar and pestle, and 100 mg of powder was mixed with 1 mL of Trizol (Invitrogen), and total RNA was extracted according to the manufacturer’s protocol. First strand synthesis of cDNA was performed using 1 μg of the extracted RNA (leaf and stem) as template, combined with the M-MuLV reverse transcriptase (NEB) and anchored oligo dT₂₂ (IDT). The synthesis of cDNA was performed according to NEB’s provided protocol. PaCPT1 was amplified from guayule cDNA by primers 1/2 and cloned into the pGEMT-easy vector (Promega) using TA/blunt ligase mix (NEB). PaCPT2/3 and PaCBP were synthesized after yeast codon optimization (SynPaCBP, MF688939; SynPaCPT2, MF688937; SynPaCPT3, MF688939). Phylogenetic and molecular evolutionary analyses were conducted using MEGA ver. 6 (Tamura et al., 2013). All sequences were retrieved from TAIR or NCBI database. The CPT sequences are from AtCPT1 (At2g23410), AtCPT2 (At2g23400), AtCPT3 (At2g17570), AtCPT4 (At5g60510), AtCPT5 (At5g60500), AtCPT6 (At5g58780), AtCPT7 (At5g58770), AtCPT8 (At5g58782), AtCPT9 (At5g58784), CeCPT (NP_001023351), EcCPT (WP_077899378), HbCPT1 (BAB92023), HbCPT2 (BAB83522), HsCPT (NP_995583), LsCPT1 (XP_023749459), LsCPT2 (XP_023743551), LsCPT3 (XP_023762616), PaCPT1 (ATD87115), PaCPT2 (ATD87118), PaCPT3 (ATD87116), RER2 (NP_009556), SaCPT (CEH25875), SlCPT1 (NP_001234633), SlCPT2 (AFW98426), SlCPT3 (AFW98427), SlCPT4 (AFW98428), SlCPT5 (AFW98429), SlCPT6 (AFW98430), SlCPT7 (AFW98431), SRT1 (NP_013819), TbCPT1 (AGE89403), TbCPT2 (AGE89404), TbCPT3 (AGE89405). The CBP sequences are from AtLew1 (At1g11755), CeCBP (NP_495928), HaCBP (OTG12218), HbCBP (XP_021659156), HtCBP (GHBN01046996), LsCBP1 (AIQ81186), LsCBP2 (AIQ81187), NogoBR (NP_612468), NUS1 (NP_010088), PaCBP (ATD87120), PtCBP (XP_002305421), SlCBP (NP_001333078), ToCBP1 (DY820019), ToCBP2 (DY832694), VvCBP (XP_003631635). The guayule sequences were deposited at NCBI: PaCBP (MF688938), PaCPT1 (MF688933), PaCPT2 (MF688936), PaCPT3 (MF688934).

Details of yeast complementation, in vitro enzyme assay, and reverse-phase thin layer chromatography were previously described (Kwon et al., 2016). All primer sequences are given in Supplementary Table S1. PaCPT1-3 were amplified using primers 3/4, 5/6 or 7/8, respectively, and the amplicons were cloned into p414-GPD using SpeI/SalI, SpeI/XhoI or BamHI/ClaI sites, respectively. PaCBP was cloned into p415-GPD using SpeI/XhoI sites. The cloned vectors were introduced to a yeast strain deficient in both copies of CPTs, rer2Δ::HIS3 srt1Δ::KanMX [pRS316-RER2], described in Kwon et al. (2016). This strain is referred to as rer2Δ srt1Δ hereafter, and Ro lab yeast stock number is ROY13. The constructed PaCBP/CPT pairs were transformed in this yeast strain. To serve as a positive control of complementation, p414-GPD ScRER2-TRP1 (previously cloned in Kwon et al., 2016) with empty p415-GPD was also transformed in the rer2Δ srt1Δ strain. The transformants were streaked on both SC-Leu-Trp-Ura-His plates (positive control) and SC-Leu-Trp-His plates containing 5FOA (5-fluoroorotic acid) and uracil to allow for counter-selection of pRS316 ScRER2-URA3. Growth on 5FOA/Ura-containing media indicates functional complementation of PaCBP/CPTs in place of yeast CPTs. Growth on both media was observed after a 7-day incubation period at 30°C.

The yeast strains, rescued from the 5FOA selection in the complementation assay, were cultured overnight and inoculated into 100 mL of SC-Leu-Trp-His media. The cells were harvested at late logarithmic phase by centrifugation at 3,000 × g for 3 min. The pelleted cells were then resuspended in 1 mL of TES-B buffer (0.6 M sorbitol, 1 mM EDTA, and 50 mM Tris-HCl, pH 7.4) and homogenized by bead-beating. The yeast lysate (1 mL) was centrifuged at 10,000 × g for 10 min to remove large cellular debris. The supernatant was diluted with 10 mL of TES-B and centrifuged again similarly. The final supernatant was centrifuged at 100,000 × g for 1 h, 4°C. The pellet was re-suspended in TG buffer [50 mM Tris-HCl and 20% (v/v) glycerol, pH 7.4] to a concentration appropriate for in vitro biochemical assays.

The in vitro biochemical assays of CPB/CPT were performed in 50 μL reaction mixtures containing 50 μg of prepared microsomal proteins, 20 μM FPP, 80 μM of IPP (¹⁴C-IPP PerkinElmer, 50–60 mCi mmol^-1), 50 mM HEPES, 5 mM MgCl₂, 2 mM NaF, 2 mM Na₃VO₄, and 2 mM DTT, pH 7.5. The reaction mixture was incubated at 30°C for 2 h. For each reaction, the extraction of polyisoprenoid products was performed in 1 mL of chloroform/methanol (2:1, v/v) and 0.4 mL of 0.9% NaCl, by vortexing for 20 s followed by centrifuged at 10,000 × g for 1 min at room temperature. The chloroform phase was transferred to a new tube and washed twice with 0.5 mL of methanol/water/chloroform mixture (48:47:3, v/v/v) by vortexing and centrifuging as above to remove unincorporated IPP. The lower chloroform phase was transferred to a glass vial and allowed to evaporate. Dephosphorylation of the product to polyprenol was carried out by incubating the dried product in 300 μL of 1 M HCl at 85°C for 1 h, followed by benzene extraction. The concentrated products in benzene were used for thin layer chromatography (TLC). The concentrated products were spotted on a C18 reverse-phase silica plate (Analtech cat. P90011, 20 × 20 cm) and separated by acetone/water (39:1, v/v) mobile phase. The TLC plate was then exposed to a BAS Storage Phosphor Screen (GE Healthcare) and the screen was scanned by a phosphorimager (Bio-Rad, Molecular Imager FX). The R_f value of known RER2 products (Brasher et al., 2015) was confirmed in our TLC system, and the R_f values were used to determine the carbon numbers.

MYTH assays were performed according to the published methods (Gisler et al., 2008; Snider et al., 2010). The bait plasmid (PaCBP-Cub-LexA-VP16) was generated by ligating the PaCBP amplified with primers 9/10 into pAMBVα using StuI/NcoI sites. For the prey plasmids, PaCPT1-3 were amplified with primers 11/12, 13/14 or 15/16, and cloned into pPR3C-NubG in SpeI/BamHI (PaCPT1) or BamHI/ClaI (PaCPT2 and 3) to yield PaCPT-HA-NubG constructs. PaCPT1-3 were also amplified with primers 17/18, 19/20 or 21/22 and cloned into pPR3N-NubG in BamHI/SalI (PaCPT1) or BamHI/ClaI (PaCPT2 and 3) to generate NubG-HA-PaCPT constructs. The bait and prey plasmid pair was co-transformed into the Y2H reporter yeast strain, THY.AP4. The control prey vectors pOST1-NubG and pOST1-NubI were also co-transformed with the bait vector. Individual transformed-yeast colonies were cultured in 30°C until OD₆₀₀ reached to 3.0–3.5. The cell densities were then normalized to 0.05 (OD₆₀₀ value) by dilution. The normalized cells were diluted to produce dilution of 1/10 and 1/100. Ten μL of cells was spotted on SC-Leu-Trp (SC-LT) and on SC-Leu-Trp-Ade-His (SC-LTAH) containing 3-amino-1,2,4-triazole (3-AT) at concentrations of 5, 10, or 25 mM. Spotted plates were incubated for 3 days at 30°C to observe cell growth.

The strength of protein interactions was quantified using β-galactosidase assay with O-nitrophenyl-β-D-galactoside (ONPG) as the substrate. From SC-LTAH selection plates, the positive colonies were cultured overnight in SC-LTAH medium without 3-AT. Cells from the overnight culture were inoculated to 4 mL of SC-LTAH medium and cultured for 5 h. The cells were harvested at 9,000 × g for 3 min at 4°C and resuspended in 1 mL of Z buffer (60 mM Na₂HPO₄, 40 mM NaH₂PO₄, 10 mM KCl, 1 mM MgSO_4, and 50 mM β-mercaptoethanol, pH 7.0), 20 μL of chloroform and 20 μL of 0.1% SDS. After vortexing, the mixture was pre-incubated for 5 min at 28°C. The reaction was initiated by adding 0.2 mL of ONPG (4 mg mL^-1 in Z buffer) and terminated by the addition of 1 M of Na₂CO₃. The mixture was centrifuged at 20,000 × g for 10 min, and the absorbance at 420 nm and 550 nm were measured to determine the Miller units (Miller, 1972).

DNA constructs for in vitro translation were cloned into pT7CFE1-3xFLAG (constructed using pT7CFE1-NHa) for PaCPT1-3 and pT7CFE1-NHa (Thermo Fisher Scientific) for PaCBP. pT7CFE1-3xFLAG was generated by replacing HA tag in pT7CFE1-NHa with 3xFLAG. The 3xFLAG (atggactacaaagaccatgacggtgattataaagatcatgatatcgattacaaggatgacgatgacaag) was amplified using primer sets 23/24, 23/25, and 23/26 sequentially using the IRES region in pT7CFE1-NHa as a template. The final PCR product containing partial IRES sequence and 3xFLAG was digested by KpnI and BamHI and ligated to corresponding restriction sites in pT7CFE1-NHa. Synthetic PaCPT1 and PaCPT2 (GenScript, Piscataway, NJ, United States) were ligated to BamHI and XhoI sites in pT7CFE1-3xFLAG. PaCPT3 was amplified from P. argentatum leaf cDNA using a primer set, 27/28. The PaCPT3 PCR product was cloned into pT7CFE1-3xFLAG via BamHI and XhoI restriction sites. Synthetic PaCBP (GenScript, Piscataway, NJ, United States) was amplified using a primer set 29/30 and digested with BamHI and XhoI, followed by ligation to corresponding sites in pT7CFE1-NHa. A control construct, pT7CFE1-NHa-GFP for co-immunoprecipitation described below, was generated by cloning GFP sequence into pT7CFE1-NHa. GFP was amplified from pCFE-GFP (Thermo Fisher Scientific) using a primer set 31/32, containing BamHI and XhoI, respectively, followed by ligation to corresponding sites in pT7CFE1-NHa. The purified plasmid preparations of all IVT constructs were cleaned up further by ethanol precipitation, followed by resuspension in nuclease-free water to a concentration of 500 ng μL^-1 prior to IVT. IVT was carried out using 1-Step Human Coupled IVT Kit (Thermo Fisher Scientific), following the manufacturer’s procedure except for reducing the reaction to half of the recommended amounts. Immediately following incubation of the IVT reaction mixture at 30°C for 4 h, 1 μL of the 12.5 μL IVT mixture was saved as input for Western blotting by adding SDS-PAGE loading buffer and storage in -20°C until use. The remaining IVT mixture was used immediately for immunoprecipitation as described below.

For pre-binding of protein G magnetic beads (NEB, ON, Canada) to monoclonal mouse anti-FLAG M2 antibody (Sigma, ON, Canada) or monoclonal rat anti-HA 3F10 antibody (Roche, Laval, QC, Canada), for each co-immunoprecipitation reaction, 10 μL of the bead suspension was washed (invert five times/vortex 5 s) in 100 μL of room temperature PBS buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄, pH 7.4) four times and incubated with 0.5 μL of the antibody in 100 μL of cold PBST (PBS, pH 7.4, 0.1% Tween 20) at 4°C for 16 h on a Labquake^TM Tube Rotator. The separation of magnetic beads from solution was achieved by applying magnetic field for 30 s on a magnetic stand. To remove unbound antibody, the antibody-immobilized beads were washed (invert five times/vortex 5 s) three times, each in 100 μL of cold IP buffer (150 mM NaCl, 100 mM Tris-HCl pH 7.4, 1 mM EDTA, 10% glycerol, 1% Triton X-100, 0.5% NP-40, 1 mM PMSF). The immobilized beads were resuspended in 500 μL of cold IP buffer, followed by addition of the 11.5 μL of the completed IVT reaction mixture. Co-immunoprecipitation reaction was incubated at 4°C for 1 h on the rotator. To remove unbound proteins, the mixture was washed (invert five times/vortex 5 s) five times, each in 700 μL of cold IP buffer. Immunoprecipitated proteins were eluted from the beads by resuspending in 20 μL of 2X SDS-PAGE loading buffer [100 mM Tris-HCl pH 6.8, 4% (w/v) SDS, 20% glycerol, 0.01% (w/v) bromophenol blue, 200 mM β-mercaptoethanol] and incubating at 95°C for 5 min with intermittent gentle flicks. This elution also liberates anti-FLAG and anti-HA antibodies from the magnetic beads as the heavy (50 kDa) and light (25 kDa) chains of the IgG antibodies. After bringing the eluted mixture to room temperature, the mixture was spun down by a brief centrifugation, followed by transfer of the eluate while on a magnetic stand. Half of the eluate was used to run 12% SDS-PAGE followed by Western blotting with mouse anti-FLAG M2 antibody and the other half with rat anti-HA 3F10 antibody. The IVT reaction mixture (1 μL of 12.5 μL total IVT reaction) saved as input was also divided in half to run in SDS-PAGE and Western blotting with anti-FLAG and anti-HA antibodies. SDS-PAGE gel was transferred to PVDF membrane using a Tris-glycine-methanol buffer system, and the antibody hybridization was performed in 5% skim milk TBST (5% skim milk, 50 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.05% Tween 20) at 1:2000 for both anti-FLAG and anti-HA antibodies by overnight incubation at 4°C. HRP-conjugated anti-mouse and anti-rat secondary antibodies were also hybridized in 5% skim milk TBST at 1:10,000 at room temperature for 1 h.

For de novo transcriptome assembly and differential gene expression analysis, 4-month-old greenhouse-grown guayule plants (AZ-2 cultivar) were transferred to a simulated summer condition in a growth chamber (16 h daylight, 25°C daytime and 15°C night time temperatures). At 6 months of age, plants were transferred to a simulated winter condition (13 h daylight, 25°C daytime and 5°C nighttime temperatures) for induction of rubber production. Plants were harvested for analysis at 8 months of age. RNA was extracted from three biological replicates each for leaf and stem tissues from induced and control plants using Trizol (Invitrogen) following the manufacturer’s instructions. Following Trizol extraction, RNA samples were treated with DNAase using the Turbo DNA-free kit (Ambion) and further purified using RNeasy column purification (Qiagen). Directional, bar-coded Illumina sequencing libraries were prepared using NEBNext^® Ultra^TM Directional RNA Library Prep Kit (NEB) with the suggestions for size selection of an average insert size of 300–450 bp. Libraries were then pooled and sequenced on Illumina Miseq and HiSeq2500 platforms producing 28 million 300 bp paired end reads and 155 million 150 bp paired end reads, respectively, with an average of 2.4 million MiSeq and 12.9 million HiSeq reads per sample. RNA-seq data was quality-controlled using Trim Galore (Babraham Bioinformatics group) to eliminate adapter sequence contamination and to trim data with a quality score below Q30. Miseq and Hiseq reads were pooled for transcriptome assembly and differential gene expression analysis. The transcriptome was assembled using Trinity assembler version 2.04 set to minimum Kmer coverage of 2 (Haas et al., 2013). The initial transcriptome assembly was filtered using a utility included within the Trinity package to eliminate transcript isoforms with low abundance and low coverage. Differential gene expression analysis was performed using EdgeR (Robinson et al., 2010). The assembly data are available in the NCBI short read archive under submission number SRP107961.

Commercially cultivated guayule (Parthenium argentatum) plants for NR production are polyploids, making the analysis of gene family difficult due to gene redundancy. However, the guayule transcriptome assembly from the diploid accession PI478640 is accessible in the public NCBI database (Hodgins et al., 2014), and we used the data set from diploid guayule in this work. In this transcriptome data, 51,947 unigenes were assembled from 983,076 reads (454 GS FLX Titanium) from four different guayule tissues (root, leaf, flower, and stem). We screened this assembled transcriptome using lettuce CPT and CPT-Like (CPTL) in our previous study in lettuce (Qu et al., 2015). CPTL was renamed as CBP (CPT-Binding Protein) in the previous work (Kwon et al., 2016) and will be referred to as CBP in this work. From this analysis, one guayule CBP and three CPT cDNAs were identified, and accordingly these cDNAs were named as PaCBP and PaCPT1-3. Unlike lettuce that has two CBP isoforms, one of which displays a laticifer-specific gene expression, guayule does not have laticifer cells and has only a single copy of CBP.

To understand phylogenetic relatedness of PaCPT1-3 to other known eukaryotic and prokaryotic CPTs, a phylogenetic tree was reconstructed using the deduced amino acid sequences from PaCPT1-3 cDNAs. It has been established that two clades of CPTs (clade I and clade II) are present in plants (Figure 1A). Clade I is comprised of chloroplastic CPTs that likely originated from prokaryotic homo-dimeric CPTs, whereas clade II consists of cytosolic/ER CPTs that interact with CBPs to form hetero-protein complexes, which are then recruited to ER. The CPT clade II is further divided into two sub-clades – one type of CPT for dolichol biosynthesis in the primary metabolism and the other type for NR biosynthesis in the specialized secondary metabolism.

In the phylogenetic analysis, all three PaCPTs belong to the cytosolic/ER clade II CPT. At the sub-clade level, with strong statistical supports, PaCPT1 is clustered with LsCPT1 (lettuce), SlCPT3 (tomato), and AtCPT1 (Arabidopsis), all of which are known to synthesize dehydrodolichyl diphosphate (Brasher et al., 2015; Qu et al., 2015; Kwon et al., 2016). On the other hand, PaCPT2 and PaCPT3 form a close cluster with LsCPT3 and TbCPT1-3 (dandelion), known to have implications in NR biosynthesis (Post et al., 2012; Qu et al., 2015). This phylogenetic tree suggested that guayule possesses two sub-clades of diverged CPTs within the CPT clade II. It can be inferred from these results that PaCPT1 is likely to be involved in the dolichol metabolism, while PaCPT2 and PaCPT3 are involved in NR metabolism.

The phylogenetic tree for CBPs was also constructed. CBP is exclusively found in eukaryotes, and thus the phylogenetic structure of CBP shows a monophyletic group (Figure 1B). The duplications of CBP (i.e., two CBP isoforms) are restricted to lettuce and dandelion in the Cichorieae tribe (within Cichorioideae sub-family under Asteraceae family), and other plants with high quality genome sequences encode a single copy of CBP. Interestingly, PaCBP and other single copy CBPs from the Asteroideae sub-family (sunflower HaCBP and Jerusalem artichoke HtCBP) are tightly clustered with the laticifer-specific CBPs (lettuce LsCBP2 and common dandelion ToCBP2), even though there is no developed laticifer in sunflower and artichoke. Thus, the CBP gene duplication event likely occurred prior to the divergence of the two sub-families and the advent of laticifer in Cichorioideae subfamily.

In eukaryotes, dolichol is an indispensable carrier molecule of sugar moieties for post-translational protein modifications. Therefore, all eukaryotes have at least one set of CPT and CBP to synthesize dehydrodolichyl diphosphate (i.e., dolichol precursor), and the absence of either CPT or CBP is lethal. Yeast (Saccharomyces cerevisiae) has one homolog of CBP (NUS1) and two homologs of CPTs (RER2 and SRT1). We previously developed a yeast strain that has double deletions in RER2 and SRT1, in which its lethality is rescued by expressing a plasmid copy of RER2 (Kwon et al., 2016). To examine the ability of PaCPT and/or PaCBP to rescue the dolichol-deficiency of rer2Δ srt1Δ, these guayule genes were first transformed to this strain individually or in PaCPT/PaCBP pairs, and then the URA-selective plasmid containing RER2 was removed by 5-fluoroorotic acid (5FOA) selection (Figure 2). This enabled us to assess the PaCPT activity in the absence of both of the yeast CPTs.

In the absence of CPT enzymatic activity, the rer2Δ srt1Δ lethal phenotype will not be rescued. In all cases tested, the rer2Δ srt1Δ strain transformed with empty plasmids or a single plasmid that contains either PaCBP or one of the PaCPTs failed to grow upon removal of the RER2-URA plasmid by 5FOA selection. In contrast, co-expression of PaCBP and one of the three PaCPTs could effectively complement the lethality of the yeast CPT-knockouts upon removal of the RER2-URA plasmid. Our results indicate that each PaCPT can produce isoprenoid molecules analogous to dehydrodolichyl diphosphate only in the presence of PaCBP but not together with endogenous yeast CBP, Nus1.

The surviving yeast strains after 5FOA selection should be able to bring about necessary CPT activities for the yeast survival in the rer2Δsrt1Δ yeast. To assess the guayule CPT activities, microsomes were prepared from the rescued yeast strains from the complementation assay, and radiolabeled ¹⁴C-isopentenyl diphosphate (¹⁴C-IPP) and the priming molecule, FPP, were co-incubated with the isolated microsomes. When the incorporations of ¹⁴C-IPP to hydrophobic polymers were measured, significant IPP-incorporation activities ranging from 14.4 to 50.8% incorporation were measured in 2-h assays (Table 1). Their specific activities were comparable to (in the cases of PaCPT1/PaCBP1or PaCPT3/PaCBP activity) or exceeded (in the case of PaCPT2/PaCBP activity) the CPT activities previously reported from lettuce CPT and CBP (Qu et al., 2015). The CPT activity of the microsomes from the rescued yeast strain transformed with RER2 as a complementation control was substantially lower (4.5- to 15.6-fold) than those from guayule PaCPT/PaCBP co-expression. The lower CPT activity of RER2 control strain is attributable to the imbalance between overexpressed RER2 and its binding partner NUS1, expressed at a native level in yeast.

The resulting ¹⁴C-labeled products were separated by the reverse-phase thin layer chromatography (RP-TLC) optimized for the separation of dolichol-sized polymers. In the RP-TLC analysis, we detected no obvious differences among the radiolabeled isoprene polymers produced from any of the PaCPT and PaCBP pairs (Figure 3A), and their polymer size was slightly smaller than, but still comparable to, the dehydrodolichols produced by RER2 (Figure 3B). We could not find evidence that the cis-isoprene polymer, significantly longer than dolichol, is synthesized by PaCPT and PaCBP in yeast microsomes.

The complementation of rer2Δ srt1Δ yeast strain by co-expression of PaCPT and PaCBP, but not by a single PaCPT or PaCBP expression, implied that both proteins are required for dolichol metabolism. Thus, possible interactions between each PaCPT and PaCBP were examined by split-ubiquitin yeast 2-hybrid assay developed for protein interactions of membrane-bound proteins (Stagljar et al., 1998). In this assay, the restoration of ubiquitin by N-terminal ubiquitin (Nub) and C-terminal ubiquitin (Cub) allows a release of transcriptional regulator to nucleus to activate the auxotrophic marker genes and the β-galactosidase reporter gene. A single residue mutation (I to G) of natural Nub protein (i.e., NubI to NubG) prevents the auto-assembly of the ubiquitin, unless NubG and Cub binding is mediated by the interaction of the proteins of interest each fused to NubG and Cub. Therefore, interactions between PaCPT and PaCBP proteins translationally fused to Cub and NubG can be assessed by yeast growth on selective medium and β-galactosidase activity.

As a control, co-expression of PaCBP-Cub fusion and OST1 (α-subunit of the oligosaccharyltransferase complex localized on the ER)-NubI fusion allows restoration of ubiquitin, thereby presenting efficient yeast growth and strong β-galactosidase activity (Figure 4A). In contrast, changing NubI to NubG in the same experiment did not permit yeast growth, suggesting PaCBP and OST1 do not interact with each other and cannot bring the two ubiquitin subunits together as expected as a negative control (Figure 4A). Similarly, PaCPTs 1–3 were individually fused to NubG, and each fusion construct was co-expressed with PaCBP-Cub to assess whether PaCPT-PaCBP interaction occurs to restore ubiquitin. The interaction between each of the PaCPTs and PaCBP could be clearly inferred by the efficient yeast growth and the by β-galactosidase activities. Interestingly, the interactions were only observed in one orientation of the PaCPT fusion protein (Figure 4B). The NubG-PaCPT orientation allowed activations of the growth markers and reporter gene, whereas the PaCPT-NubG orientation did not show evidence of protein–protein interactions. Activations of marker genes only when NubG is located in the N-terminus of prey protein have been previously reported (Yao et al., 2017) and have occasionally happened in split-ubiquitin yeast 2-hybrid assays (personal communication with Igor Stagljar, University of Toronto). Wherever interactions were observed, the HIS3 marker gene was activated very efficiently as yeast could grow at up to 25 mM of AT (3-amino-1,2,4-triazole), a competitive inhibitor of HIS3. In addition, β-galactosidase activities were comparable to that from the positive control. Therefore, the split-ubiquitin Y2H provided molecular evidence that each PaCPT and PaCBP interact with each other to form a hetero-protein complex in yeast.

In addition to the split-ubiquitin Y2H approach, PaCPT1-3 and PaCBP interaction were independently examined by co-immunoprecipitation assays (Co-IP). PaCPT1-3 and PaCBP were tagged by FLAG- and HA-epitope, respectively, and their proteins were prepared by in vitro transcription and translation. Green fluorescent protein (GFP), which has a similar size to PaCBP and the PaCPTs, was used as a negative control in Co-IP experiments. All in vitro co-transcribed and co-translated PaCPT/PaCBP proteins were detectable in input by immunoblotting with either FLAG or HA antibody (Figure 5A,B input; Supplementary Figure 1). Incubation of FLAG-antibody immobilized magnetic beads resulted in the co-IP of FLAG-PaCPTs with HA-PaCBP as detected by FLAG-antibody and HA-antibody immunoblotting, respectively, while HA-GFP control protein could not be pulled down by FLAG-PaCPT2 (Figure 5A and Supplementary Figure 1). Similarly, immunoprecipitation with HA-antibody immobilized magnetic beads resulted in the co-IP of HA-PaCBP with FLAG-PaCPTs, while FLAG-PaCPT2 control protein could not be pulled down by HA-GFP (Figure 5B). These results showed that each of the PaCPT1-3 and PaCBP interact with each other to organize hetero-protein complexes when co-expressed in vitro.

Cold temperature is known to induce NR biosynthesis in guayule (Madhavan et al., 1989). Based on the positive effect of cold-stress on rubber synthesis, Illumina sequencing of guayule stem and leaf tissues after cold-stress was recently performed, and the sequencing data are publicly accessible at NCBI (short read archive number: SRP107961). Using these data, we conducted a RNA-seq analysis to examine the transcript abundance of PaCPT1-3 and PaCBP (Figure 6).

PaCPT1 transcripts were maintained at a relatively constant basal level in leaves and stems under both normal and cold-treated conditions. These data suggest the basal metabolic role of PaCPT1 in dolichol metabolism and are congruent to the phylogenetic analysis, in which PaCPT1 is clustered with other CPTs involved in the dolichol metabolism. PaCPT2 transcripts were hardly detectable in stem tissues while detectable at moderate levels in leaves. Of the three PaCPTs examined, PaCPT3 showed the highest level of expression in stem (2- to 3-orders of magnitude higher levels than PaCPT1/2), and its transcript abundance increased 1.8-fold by cold-treatment in stem. As NR is primarily synthesized in the parenchyma cells of the guayule stem (Kajiura et al., 2018), these data suggest that PaCPT3 is the primary CPT that contributes to the NR biosynthesis in guayule stem. On the other hand, transcript levels of PaCBP showed a lower degree of fluctuation, compared to PaCPT1-3 gene family. Its expression levels closely matched those of PaCPT2 and PaCPT3 in leaves. Noticeably, the PaCBP transcript abundance is highest in stem and is increased by 2.2-fold by cold-treatment, precisely mirroring the expression pattern of PaCPT3 in stem.

Average values of PaCBP and PaCPT3 expression increased by cold treatment between 1.8- and 2.2-fold in leaves and stem, but only PaCBP1 induction in leave was statistically significant (p-value < 0.05) in the triplicate Illumina transcript data. Systematic statistical analyses further showed that expression levels of PaCPT3 in comparison to PaCPT1 and PaCPT2 are significant in stem (p-value < 0.01) and leaves (p-value < 0.05). Preferential expression of PaCPT3 in stem in comparison to leaves was also statistically significant (p-value < 0.05).