Impact of nanodisc lipid composition on cell-free expression of proton-coupled folate transporter

The Proton-Coupled Folate Transporter (PCFT) is a transmembrane transport protein that controls the absorption of dietary folates in the small intestine. PCFT also mediates uptake of chemotherapeutically used antifolates into tumor cells. PCFT has been identified within lipid rafts observed in phospholipid bilayers of plasma membranes, a micro environment that is altered in tumor cells. The present study aimed at investigating the impact of different lipids within Lipid-protein nanodiscs (LPNs), discoidal lipid structures stabilized by membrane scaffold proteins, to yield soluble PCFT expression in an E. coli lysate-based cell-free transcription/translation system. In the absence of detergents or lipids, we observed PCFT quantitatively as precipitate in this system. We then explored the ability of LPNs to support solubilized PCFT expression when present during in-vitro translation. LPNs consisted of either dimyristoyl phosphatidylcholine (DMPC), palmitoyl-oleoyl phosphatidylcholine (POPC), or dimyristoyl phosphatidylglycerol (DMPG). While POPC did not lead to soluble PCFT expression, both DMPG and DMPC supported PCFT translation directly into LPNs, the latter in a concentration dependent manner. The results obtained through this study provide insights into the lipid preferences of PCFT. Membrane-embedded or solubilized PCFT will enable further studies with diverse biophysical approaches to enhance the understanding of the structure and molecular mechanism of folate transport through PCFT.


Preparation of plasmid DNA
The open reading frame for full-length human proton-couple folate transporter (PCFT), Uni-ProtKB accession number Q96NT5, was cloned into the pEXP5-NT/TOPO vector to obtain PCFT-pEXP5-NT. The DNA sequence of the resulting PCFT construct contains an N-terminal 6X-His tag followed by a Tobacco Etch Virus (TEV) protease recognition site. The sequence was confirmed by DNA sequencing (Genewiz). Plasmid DNA was isolated with the EndoFree Plasmid Maxi Kit (Qiagen) following the manufacturer's protocol.

Cell-free expression of PCFT-protein synthesis reaction
The S30 T7 High-Yield Protein Expression System was used for cell-free PCFT protein expression in the presence and the absence of LPNs, lipids, and detergents. This expression system is an Escherichia coli extract-based cell-free protein synthesis system, which consists of T7 RNA polymerase for transcription and all necessary components for translation. PCFT-pEXP5-NT was used as template, which contains a T7 promoter, a ribosome binding site, and the DNA sequence encoding for full-length human PCFT protein (Fig 1). The typical reaction volume was 5 μL using the template, T7 S30 Extract for circular templates, and S30 Premix Plus as recommended by the manufacturer (Fig 2). Reactions were incubated in a thermomixer at 1,200 rpm and 37˚C for 1 hour.
After the expression, each sample was centrifuged at 22,000 g for 10 min at 4˚C to separate insoluble PCFT in the pellet from soluble PCFT in the supernatant. Each pellet fraction was resuspended in 50 μL of 1x Laemmli sample buffer containing 1% SDS, with or without 2.5% of 2-mercaptoethanol. Each supernatant fraction was diluted using the same sample buffer in a total volume of 50 μL.

Immunoblot analysis
Samples from the pellet and supernatant fractions were heated at 70˚C for 10 minutes to enhance protein denaturation. Ten microliters of each sample, which corresponds to 1 μL of the initial reaction volume, were separated on Mini-PROTEAN TGX gels (BioRad) and transferred to PVDF membranes (BioRad) using the Bio-Rad trans-blot turbo system following the manufacturer's protocol. PVDF membranes were briefly immersed in methanol and then equilibrated in western transfer buffer (20% of Bio-Rad 5x transfer buffer, 20% ethanol) for 3-5 minutes. Proteins were transferred at 2.5 Amperes and up 25 Volts for 3 minutes at room temperature. Polyvinylidene fluoride (PVDF, Biorad) membranes were blocked under agitation in 5% nonfat milk in tween-containing tris-buffered saline buffer (TTBS buffer: 0.1% Tween-20, 100 mM Tris, 0.9% NaCl, pH 7.5) for one hour or overnight. Afterwards, the blot was incubated with 6X-His tag antibody conjugated with HRP in a dilution of 1:5000 for one hour under gentle agitation. Subsequently, the blot was washed five times, each for 5 minutes with 5 ml of TTBS buffer. After an additional wash with tris-buffered saline (100 mM Tris, 0.9% NaCl, pH 7.5), SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Scientific) was used for imaging with a digital imaging system (ImageQuant TM LAS 4000, GE Healthcare).

Data analysis
PCFT quantification was performed to determine if PCFT solubility depends on the lipid composition and concentrations of LPNs. The soluble PCFT bands on western blot images were selected, the band intensities were plotted as peaks, and the area under a peak was analyzed using ImageJ software. The resulting values were normalized to the highest value obtained from soluble PCFT expressed in LPNs. Statistical significance in PCFT solubility, depending on lipid concentrations and compositions, was determined using one-way ANOVA with Tukey's multiple-comparisons test in Prism 6 Software (GraphPad Prism, La Jolla, CA).
Data charts were generated using Microsoft Excel and Prism 6 Software (GraphPad Prism, La Jolla, CA). Figures were created using ApE plasmid editor, Adobe Illustrator 2020, and Adobe Photoshop 2020.

Results and discussion
A detailed molecular understanding of proton-coupled folate transporter (PCFT) is desirable and essential in the structure-based design of drugs and therapies effectively targeting folaterelated neurological disorders and cancer. Biochemical and biophysical techniques routinely applied to study the structure and function of proteins frequently require soluble or solubilized (membrane) proteins. For integral membrane proteins, a critical step in making protein  available for biophysical approaches such as for example cryo-electron microscopy, circular dichroism, isothermal titration calorimetry, or X-ray crystallography is the solubilization of the protein from the native lipid bilayer or the integration into a stable bilayer (mimetic) or detergent micelle [46]. To produce soluble PCFT, a PCFT construct was designed as part of this study, containing full-length human PCFT cloned into the pEXP5-NT/TOPO vector ( Fig  1). The cell-free transcription and translation system was then used to examine the expression and solubility of PCFT in the presence (Fig 2) or the absence of LPNs containing different lipids.

PCFT expression is observed in cell-free expression system
Transcription/translation reactions contained PCFT-pEXP5-NT as a template. After incubation reactions were separated by centrifugation into pellet and supernatant fractions. Since PCFT is a transmembrane protein, in the absence of detergent or an otherwise solubilizing milieu, the translated protein is expected to be in the pellet, whereas soluble PCFT would be in the supernatant fraction (Fig 3). Proteins in both fractions were separated on SDS-PAGE and detected using western blotting with an antibody directed against the N-terminal 6X-His tag present in the PCFT construct.
In the absence of the plasmid PCFT-pEXP5-NT template, no band, and consequently no PCFT protein, was detected in either the pellet or supernatant fractions ( Fig 3A, lanes 'Control'), whereas the presence of PCFT-pEXP5-NT template in the reaction yielded a band at approximately 37 kDa ( Fig 3A, pellet fractions). The 6X-His-tagged full-length human PCFT construct in this work consists of 493 amino acids with a theoretical molecular weight of 52 kDa. When expressed in Spodoptera frugiperda (Sf9) insect cells [47], Chinese hamster ovary cells [34], or Xenopus laevis oocytes [34] monomeric PCFT migrates in two bands between 35 and 55 kDa. The electrophoretic mobility of PCFT depends on the post-translational modification of N-linked glycosylation [48]. Removing the glycosylation with PNGase F results in PCFT that mainly migrates at~35 kDa [48]. The migration of PCFT observed here (~37 kDa) is consistent with non-glycosylated PCFT as expected in the cell-free expression system where post-translational modifications do not occur.
This data indicates that the PCFT protein was expressed in the cell-free expression system using PCFT-pEXP5-NT as a template. As would be expected in the absence of detergents or bilayers, the expressed PCFT is exclusively insoluble, and PCFT protein is only found in the pellet but not the supernatant fraction.

Cell-free expressed and soluble PCFT is detected in the presence of LPNs
Next, the cell-free expression of PCFT was examined in the presence of LPNs of differing lipid compositions. LPNs were stabilized by the membrane scaffold protein MSP1D1 (genetically modified apolipoprotein A1 with N-terminal globular region replaced by 6X-His and Tobacco Etch Virus protease site). The~10-nm size of MSP1D1 LPNs was used to be tailored to the diameter of a transmembrane protein with 12 α-helical transmembrane segments. Fig 3B shows PCFT translation using cell-free expression in the presence of MSP1D1 LPNs containing zwitterionic POPC with a saturated palmitoyl-and unsaturated oleoyl chain (16:0-18:1 PC; 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine), zwitterionic saturated DMPC (14:0 PC; 1,2-dimyristoyl-sn-glycero-3-phosphocholine) or anionic saturated DMPG (14:0 PG; 1,2-dimyristoyl-sn-glycero-3-phospho-(1'-rac-glycerol)) (Cube Biotech) with PCFT-pEXP5-NT (6X-His-PCFT) as a DNA template. In the absence of LPNs, once again only insoluble PCFT was observed. In the presence of 20 or 40 μM LPNs, PCFT was found in both insoluble form in the pellet and soluble form in the supernatant (Fig 3B). More soluble PCFT was detected for reactions with LPNs containing the saturated lipids DMPG or DMPC as compared to those containing unsaturated POPC.

DMPG and DMPC LPNs produce soluble PCFT
To further assess the impact of LPNs, their concentrations, and lipid compositions on the PCFT solubility, the experiment in Fig 3B was repeated with higher concentrations of LPNs, i.e., 80 µM LPNs, containing POPC, DMPG, and DMPC lipids (Fig 4A). 'PCFT' lanes  The data show the mean ± SEM from n = 3 independent experiments. The highest level of solubilized PCFT was found in LPNs containing DMPC lipid. LPNs containing POPC lipid yielded minimal levels of solubilized PCFT. Statistical significance was determined using was determined using one-way ANOVA with Tukey's multiple-comparisons test in Prism 6 Software (GraphPad Prism). Significance is indicated vs "PCFT" sample without LPNs as �� p � 0.01, ��� p � 0.001, and p � ���� 0.0001). Nonsignificant p value is shown as ns.
https://doi.org/10.1371/journal.pone.0253184.g004 represent PCFT expressed in the absence of LPNs, and the lack of soluble PCFT again confirms that without LPNs, only insoluble PCFT was obtained in the cell-free expression. In contrast, depending on the specific lipid within LPNs, both insoluble and soluble PCFT were observed when LPNs were present during transcription and translation in the expression reaction. No or minimal PCFT was incorporated into POPC LPNs. For all three POPC LPN concentrations used, 20, 40, and 80 μM, when the band intensity was assessed using one-way ANOVA with Tukey's multiple-comparisons test, the PCFT band intensities were not significantly different from PCFT expressed in the absence of any solubilizing agent. At a concentration of 20 µM, DMPG LPNs yielded approximately three times the soluble PCFT intensity as compared to POPC LPNs. The soluble PCFT in DMPG reactions was significantly above the reactions without LPNs (p � 0.001, p � 0.01, p � 0.01 for 20, 40, and 80 µM, respectively). However, the concentration of the DMPG LPNs itself did not impact the soluble PCFT in a concentrationdependent manner.
Similar to 20 µM DMPG LPNs, the same concentration of 20 µM DMPC yielded a threefold increase in soluble PCFT as compared to POPC LPNs. Contrary to DMPG LPNs, DMPC LPNs led to a concentration-dependent increase in soluble PCFT. The highest concentration of DMPC LPNs of 80 µM, produced the highest amount of soluble PCFT, approximately twofold higher than all DMPC concentrations.
For downstream applications for which purified PCFT/nanodisc complexes are required, PCFT would be expressed as described here with the N-terminal His-tag in the presence of nanodiscs stabilized by MSP without His-tag. Metal affinity chromatography against the Histag of PCFT while using untagged MSP nanodiscs and/or a gel filtration would provide scalable pathways to purified PCFT-LPNs complexes for downstream assays. Some biophysical measurements, for example single-molecule fluorescence resonance energy transfer (smFRET) measurements using Cys pairs labeled with appropriate donor and acceptor fluorophores within PCFT do not require purification [49][50][51][52].
Overall, incorporation of PCFT into LPNs containing DMPG and DMPC (the more ordered or condensed lipids compared to POPC [53]) was higher than compared to POPC LPNs with an overall increase in solubility from POPC < DMPG < DMPC. PCFT has been found in lipids rafts. Reduced levels in lipid rafts have been associated with folate malabsorption in chronic alcoholism [54]. Changes to the lipid raft composition in cancer have been implicated in regulating cell proliferation, apoptosis, and cell migration [55]. Lipid rafts are microdomains found in the plasma membrane that are enriched in sphingolipids and cholesterol. High levels of cholesterol in membranes lead to directional organization of the lipid bilayer due to the rigid sterol [56]. The packing of cholesterol in between lipid acyl chains consequently results in a closer packing or compaction of the lipid bilayer, a state of the bilayer also termed as "liquid-ordered". Therefore, the preference of PCFT to integrate into LPNs containing ordered lipids is comparable to its native localization within lipid raft microdomains.

Conclusion
PCFT was identified as the main pathway for dietary folate uptake in the proximal small intestine in 2006 [5]. Detailed studies of human PCFT using PCFT constructs modified with amino acids engineered at specific positions, including the use of the substituted cysteine accessibly method (SCAM) [57] as well as recent cryo-EM structures of chicken PCFT have provided insights into the topology and conformational transitions during substrate translocation [15, 17, 33-40, 45, 58-61]. However, much remains unknown about PCFT's detailed molecular structure, substrate translocation mechanism, and lipid dependence. Structures of chicken PCFT were published in a substrate free state and in complex with the antifolate drug pemetrexed [45]. PCFT adopts the outward-facing conformation in both structures. The PCFT cryo-EM structure obtained in the absence of substrate (PDB entry: 7BC6) in the outward-open state is largely similar to our published human PCFT 3D homology model based on the YajR experimental structure (PDB entry: 3WDO), with a root-mean-square deviation (RMSD) of 4.0Å [33,35]. The substrate binding site as well as positions involved in functional aspects of PCFT are remarkably similar between our model and the experimental structure. In particular, the engagement of motif A (Asp109-Arg113-Asp170) is observed in both the cryo-EM structure and our 3WDO-based model, as would be expected in the outward-facing conformation. The cryo-EM structures were obtained using a nanobody that binds to the long extracellular loop between TM1 and TM2. The high affinity (KD of 8 nM) of the nanobody indicates its tight interactions with PCFT and at the same time provides the potential for structural disturbances based on the nanobody's interactions with extracellular loops and its insertion into the outward-facing substrate translocation pathway. These highly novel experimental structures further support our models but lack information pertaining to structural and functional changes occurring with disease causing mutations or conformational transitions caused by interaction with different substrates (folates vs antifolates). In particular, dynamic insights are lacking that could be obtained using smFRET experiments obtained with PCFT expressed as described in the present study. In addition, detailed structural information about the inward-open conformation, as well as occluded intermediate conformations in the transport cycle are still lacking.
The structure and activity of integral membrane proteins heavily depends on lipid environments [62][63][64][65][66][67][68][69] as demonstrated by the following select MFS transporter examples. For purine transporters, UapA and AzgA, it was shown that ergosterol, sphingolipids or phosphoinositides negatively affect biogenesis and stability of the transporters [70]. In the case of the secondary active transporter XylE, phosphatidylethanolamine has been demonstrated to disrupt conserved charge networks leading to conformational changes between different states and to favor an inward-facing state [71]. It was found that the transport activity of lysine transporter Lyp1 depends on the concentration of phosphatidylserine and ergosterol, while the transport rate of aspartate transporter relies on lipid composition [72,73].
In this study, PCFT was produced in a soluble form using cell-free synthesis in the presence of membrane-mimicking LPNs. Soluble PCFT was obtained with a continuous one-step reaction and within one hour with an overall increase in solubility from POPC < DMPG < DMPC. In preliminary experiments, detergent micelles alone or mixed detergent-lipid micelles did not support soluble PCFT synthesis (S1 Fig). The described soluble PCFT expression directly into LPNs facilitates the expression of engineered constructs with site-specific modifications. In addition, the lipid composition can be systematically modified further. The expression level is amenable for diverse biophysical approaches to study PCFT's structure and function, such as applications using fluorescence (tryptophan fluorescence or smFRET), cryo-EM, isothermal titration calorimetry or circular dichroism.