Partitioning of semisynthetic lipidated N-Ras in lipid raft nanodomains determined by FRET to lipid domain markers

TL;DR: It is demonstrated that N-Ras preferentially populated raft domains when bound to Mant-GDP, while losing preference for rafts when it was associated with a GTP mimic, mant-GppNHp.

Abstract: Cellular membranes are heterogeneous planar lipid bilayers displaying lateral phase separation with the nanometer-scale liquid-ordered phase (aka “lipid rafts” or Lo) surrounded by the liquid-disordered phase (Ld). Many membrane-associated proteins were found to stably integrate in the rafts, which is critical for their biological function. Isoforms H and N of Ras GTPase possess a unique ability to switch their lipid domain preference depending on the type of bound guanine nucleotide (GDP or GTP). This behavior, however, has never been reproduced in vitro in model bilayers with recombinant proteins, and therefore has been attributed to action of other proteins binding Ras at the membrane surface. In this paper, we report the observation of the nucleotide-dependent switch of lipid domain preferences of the semisynthetic lipidated N-Ras in raft lipid vesicles in the absence of other proteins. To detect segregation of Ras molecules in raft and disordered lipid domains, we measured Forster Resonance Energy Transfer (FRET) between the donor fluorophore, mant, attached to the protein-bound guanine nucleotides, and the acceptor, rhodamine-conjugated lipid, localized to the liquid-disordered domains. We demonstrated that N-Ras preferentially populated raft domains when bound to mant-GDP, while losing preference for rafts when it was associated with a GTP mimic, mant-GppNHp. At the same time, the isolated lipidated C-terminal peptide of N-Ras was found localized outside of the liquid-ordered rafts, most likely—in the bulk disordered lipid.

Summary

Introduction

• Separation with the nanometer-scale liquid-ordered phase (aka "lipid rafts" or Lo) surrounded by the liquid-disordered phase (Ld).
• Their cellular counterparts are expected to be much smaller, nanometer-sized, making them only resolvable by electron and atomic force microscopy techniques11-14.
• In a cell, many membrane proteins permanently reside in raft membrane domains, which is essential for their function5, 15-19.
• Ras proteins are represented by three Ras isoforms with a high degree of homology and nearly 90% sequence identity in the N-terminal GTPase domain33.

1. Lipid membrane mimic with nano-scale lipid domains

• To create lipid bilayers that spontaneously forms nanometer-sized raft domains (approx. ranging from 4 to 15 nm), the authors followed Pathak and London50 and utilized a lipid mixture of sphingomyelin (SM), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), and cholesterol in the equimolar ratio (referred to in the following as the raft lipid mixture).
• Confocal fluorescence microscopy of a supported lipid bilayer made of the raft lipid mix confirmed that these bilayers do not form non-physiological micrometer-sized rafts .
• The Ro of 36Å allows for sensitive detection of formation and melting of nanoscopic raft nanodomains50.
• Increase of relative access of Rhod-DOPE acceptors (magenta stars) to the DPH donor molecules (green stars) upon heating.
• The light scattering by LUV is relatively temperature- independent, therefore, the F/Fo variation with temperature reflects the relative change of FRET from DPH to Rhod-DOPE.

2. Preferential localization of N-Ras C-terminal lipopeptide

• The N-methylanthranyl group (mant) was attached to the lipopeptide N-terminus to serve as a donor fluorophore.
• Heating led to increasing F/Fo indicating reduction of the FRET efficiency at higher temperatures, while homogeneous lipid showed relatively constant F/Fo values.
• In an analogous system, Fastenberg et al. explained such increasing pattern by hypothesizing that donor is present in the same disordered phase as the acceptor59.
• Intensity ratio, F/Fo, was calculated using mant emission of F and Fo samples, containing and lacking RhodDOPE, respectively.
• A similar increasing trend in the temperature dependence of F/Fo values was observed confirming their conclusion of the localization of the mant-labeled C-terminal N-Ras lipopeptide outside of lipid rafts—readily accessible by acceptor fluorophores.

3. Test of the raft-boundary localization of the C-terminal N-Ras peptide

• Experiments with mant-lipopeptide revealed that the lipopeptide is accessible to the acceptor fluorophore, Rhod-DOPE, at all times .
• The lineactant facilitates increase of the total length of the boundary thus promoting breaking the existing rafts into smaller ones (destabilization of large rafts).
• Reduction in raft size will be detectable in FRET experiments with DPH and Rhod-DOPE, because DPH will be more effectively quenched by Rhod-DOPE in smaller rafts.
• Reduction in F/Fo values upon heating due to melting of lipid rafts occurs in a similar temperature range both in the absence and the presence of the lipidated N-Ras peptide.
• This observation implies that the raft boundary does not significantly attract the lipopeptide.

4. Determination of the domain localization with time-domain fluorescence measurements

• Analysis of the FRET donor distribution among ordered and disordered lipid domains in the heating/cooling experiments described above relied on a measurement of relative fluorescence intensities in the two samples with and without acceptor (F and Fo), which required exactly matching concentrations of the donor.
• This is easy to accomplish for lipid mixtures that are made by taking accurate aliquots of fluorophore stocks, yet is very hard to achieve for the protein associated with LUV.
• The homogeneous and raft mixtures lacking acceptor (Fo samples; green and black symbols) exhibited relatively invariable lifetimes throughout the full temperature range.
• Containing mixtures in the presence and the absence of acceptor Rhod-DOPE.
• (B) FRET efficiency calculated using Eq. 1 (see Materials and Methods) from lifetimes of DPH in panel A.

5. Preferential localization of N-Ras bound with fluorescent GDP and GTP-mimics

• To mimic a full-length N-Ras with the native posttranslational lipidation pattern (one palmitoyl and one farnesyl chain), the authors prepared a semisynthetic protein following protocols developed by Herbert Waldman group68-70.
• To establish the predominant localization of the N-Ras bound to mant-nucleotides, the authors determined efficiency of FRET between mant group and Rhod-DOPE in homogeneous and raft LUV.
• N-Ras in its activated conformation (bound to GTP-mimic) was localized at the raft boundary or in a disordered lipid phase.

Protein constructs

• The full-length gene of the wild-type N-Ras was a gift of Dr. Robert Deschenes, University of South Florida.
• For bacterial expression, N-Ras gene was subcloned into the pET vector (EMD Millipore, Billerica, MA).
• The truncated N-Ras construct ending with the cysteine 181 was prepared by introducing a stop codon in place of the methionine 182 codon.
• The C118S mutation was introduced to avoid possible side reactions between the only exposed cysteine on the G domain and maleimido group of the lipidated peptide93.
• All mutagenesis steps were performed using the QuikChange Site-Directed Mutagenesis Kit (Life Technologies, Grand Island, NY).

Protein preparation

• Expression and isolation NRas-C118S-181 was performed as described earlier for a similar construct of H-Ras with little modifications92, 94.
• Final yield was approximately 0.5-2 mg of 95% pure protein from each liter of the expression medium.

Preparation of the lipidated peptides

• Fmoc-protected farnesylated cysteine was prepared as described95.
• The synthesized peptides were purified using the RP-HPLC-C4 column .
• This non-ionic detergent undergoes phase separation in aqueous solutions at temperatures above 30°C.
• The Triton X-114 solution was prepared prior to the reaction to achieve the final concentration of about 30 g/L as described98.
• Figures S5 and S6 show an expected increase in mass of 1315 Da indicating that Ras was successfully conjugates to the lipidated peptide.

LUVs preparation

• Lipids and their fluorescent derivatives were dissolved in chloroform (with the exception of DPH, which was dissolved in ethanol) and stored at -20°C.
• Lipid unilamellar vesicles, LUV, were prepared by extrusion following published protocols50, 99.
• To make the Fo samples, Rhod-DOPE was substituted by 2% DOPG to remove acceptor fluorophore but maintain the negative charge of the bilayer.
• Donor fluorophores DPH, dansyl-DOPE, and mant-lipopeptide, were added to F and Fo samples to 0.1% mol of total lipid, respectively.
• The temperature of the exturder block was maintained at 75 oC.

Preparation of Ras-LUV samples

• Fluorescent Ras-GDP and Ras-GTP complexes were prepared using the (2'-(or-3')-O-(N- methylanthraniloyl) guanosine 5'-diphosphate, mGDP, and the slowly hydrolysable GTP mimic 2'/3'-O-(N-methyl-anthraniloyl)-guanosine-5'-[(β,γ)-imido] triphosphate, mGppNHp, respectively.
• To prepare N-Ras-mGDP or N-Ras-mGppNHp associated with LUV, the lipidated Ras samples were subject to the nucleotide exchange followed by association with LUV and chromatographic separation as described in the following.
• The reaction mixtures were incubated for 2 hours at room temperature.
• Sizeexclusion elution profiles were monitored by tyrosine and mant fluorescence for protein, and rhodamine fluorescence for LUV.
• Confocal microscopy of supported lipid bilayers Supported lipid bilayers were created using raft LUV and observed with Nikon Perfect Focus Ti-E inverted research microscope using standard laser and filter sets.

Fluorescence spectroscopy

• Measurements of steady-state and time-resolved fluorescence were performed using the Photon Technology International QM40 QuantaMaster system equipped with Pico-Master 1 time- correlated single-photon counting unit (HORIBA Scientific, Edison, NJ).
• A four-position Peltier- based Turret 400 (Quantum Northwest, Shoreline, WA) allowed for simultaneous temperature control and observation of up to four replicates for each sample condition.
• Time-domain fluorescence decays were analyzed using DecayFit software (kindly shared by Søren Preus; available from www.fluortools.com).

Lipidated N-Ras protein

• Figure S5. SDS-PAGE of the C118S N-Ras protein before (lanes 2 and 4) and after conjugation with lipidated peptide (lane 3).
• The difference in masses of the major peaks is 1315 Da. Figure S7.
• Two-dimensional excitation-emission spectra of the full-length lipidated C118S N-Ras bound with mGDP (left) and mGppNHp .

Number of peptide molecules per LUV =

• Maximum possible number of rafts per LUV = LUV surface area / Area of raft = 400; Rafts are estimated to occupy approximately 10-40% of the membrane106-108 in the cell and 50% in the lipid mixture that the authors are using50, which results in ca. 200 rafts per LUV.
• Calculation of the protein surface density for Ras-LUV complex.
• The external surface area of LUV per liter of the sample is 6.8 m2.

Figures

Figure 1. Overlay of images of NBD-DPPE fluorescence (green) and Rhod-DOPE fluorescence (red) in supported lipid bilayers made of the raft lipid mix (molar ratio: SM/POPC/Chol = 1:1:1). Bright yellow spots correspond to aggregated LUV that were not removed during the wash phase. Black areas (in the middle of the image) are, likely, due to defects on the glass surface.

Figure 4. Disordered domain markers demonstrate increasing FRET upon reduction in raft size. (A) F/Fo temperature dependence for Dansyl-DOPE donor (0.1% mol) incorporated into the homogeneous and raftcontaining lipid bilayers containing Rhod-DOPE (2% mol). (B) Schematic drawing illustrating an increase in the average distance between donors (green stars) and acceptors (magenta stars) due to the melting of a raft phase (gray).

Figure 2. Presence of rafts in SM/POPC/cholesterol lipid bilayers detected by FRET between lipid domain markers. (A) Heating and cooling profiles of the homogeneous and raft LUV solutions with the DPH (0.1% mol) and Rhod-DOPE (2% mol) donor/acceptor pair at a scan rate of 0.5 oC/min. Each curve is an average of two independent samples. Fluorescence intensity ratio, F/Fo, is calculated using DPH emission of F and Fo samples, containing and lacking Rhod-DOPE, respectively. (B) A schematic drawing illustrating the

Figure 6. Raft stability in SM/POPC/cholesterol bilayers evaluated through time-domain fluorescence measurements. (A) Lifetimes of DPH fluorescence at different temperatures in homogeneous and raft-

Figure 5. Test of a boundary localization of N-Ras C-terminal lipopeptide: heating profiles for the raft LUV with DPH and Rhod-DOPE and increasing concentration of non-fluorescent lipopeptide. The curves were shifted along Y axis to facilitate the comparison of the transition region.

Figure 3. Non-raft localization of N-Ras C-terminal lipopeptide revealed by FRET to the disordered domain marker. Heating and cooling profiles of the homogeneous and raft LUV with Rhod-DOPE (acceptor; 2% mol) in the presence of the mant-labeled N-Ras C-terminal lipopeptide (donor; 0.1% mol). Fluorescence

Figure 7. Efficiency of FRET between mant and Rhod-DOPE in samples of N-Ras-mGDP and N-RasmGppNHp at 5 oC. Error bars indicate standard deviations from replicate lifetime measurements (for the numbers of replicates see Supporting Table 1). The raft LUV sample preparations were repeated to increase confidence in the result (indicated as prep #1 and #2, accordingly).

Full text

1
Partitioning of semisynthetic lipidated N-Ras in lipid
raft nanodomains determined by FRET to lipid
domain markers
Anna K. Shishina
1
, Elizaveta A. Kovrigina
1,2
, Azamat R. Galiakhmetov
1
, Rajendra Rathore
1,
*,
Evgenii L. Kovrigin
1,
*
1
) Chemistry Department, Marquette University, P.O. Box 1881, Milwaukee, Wisconsin 53201,
United States
2
) current address: Biochemistry Department, Medical College of Wisconsin, Milwaukee, WI
53226
* corresponding authors
RUNNING TITLE: N-Ras localization in model membranes
KEYWORDS: semi-synthetic lipidated N-Ras GTPase, FRET, lipid membranes, lipid rafts, time-
resolved fluorescence spectroscopy
ABSTRACT: Cellular membranes are heterogeneous planar lipid bilayers displaying lateral phase
separation with the nanometer-scale liquid-ordered phase (aka "lipid rafts" or L
o
) surrounded by
the liquid-disordered phase (L
d
). Many membrane-associated proteins were found to stably
.CC-BY-NC-ND 4.0 International licenseunder a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available
The copyright holder for this preprint (which wasthis version posted November 17, 2017. ; https://doi.org/10.1101/221382doi: bioRxiv preprint

2
integrate in the rafts, which is critical for their biological function. Isoforms H and N of Ras
GTPase possess a unique ability to switch their lipid domain preference depending on the type of
bound guanine nucleotide (GDP or GTP). This behavior, however, has never been reproduced in
vitro in model bilayers with recombinant proteins, and therefore has been attributed to action of
other proteins binding Ras at the membrane surface. In this paper, we report the observation of the
nucleotide-dependent switch of lipid domain preferences of the semisynthetic lipidated N-Ras in
raft lipid vesicles in the absence of other proteins. To detect segregation of Ras molecules in raft
and disordered lipid domains, we measured Förster Resonance Energy Transfer (FRET) between
the donor fluorophore, mant, attached to the protein-bound guanine nucleotides, and the acceptor,
rhodamine-conjugated lipid, localized to the liquid-disordered domains. We demonstrated that N-
Ras preferentially populated raft domains when bound to mant-GDP, while losing preference for
rafts when it was associated with a GTP mimic, mant-GppNHp. At the same time, the isolated
lipidated C-terminal peptide of N-Ras was found localized outside of the liquid-ordered rafts, most
likely—in the bulk disordered lipid.
.CC-BY-NC-ND 4.0 International licenseunder a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available
The copyright holder for this preprint (which wasthis version posted November 17, 2017. ; https://doi.org/10.1101/221382doi: bioRxiv preprint

3
INTRODUCTION
Lipid rafts, the nanoscale lipid domains, in a plasma membrane of living cells play a
crucial organizing role in cellular signaling and regulatory cascades
1-6
. Micrometer-sized lipid
domains with a liquid-crystal-like order are easily observable by optical fluorescence microscopy
in model membranes made of heterogeneous lipid mixtures
4, 7-10
. However, their cellular
counterparts are expected to be much smaller, nanometer-sized, making them only resolvable by
electron and atomic force microscopy techniques
11-14
. In a cell, many membrane proteins
permanently reside in raft membrane domains, which is essential for their function
5, 15-19
. Ras, a
small monomeric GTPase, provides an intriguing example of a membrane protein that dynamically
switches its nanodomain affinity upon transition between its active and inactive functional states
(bound to GTP and GDP, respectively)
20-23
.
Ras is a small monomeric GTPase involved in regulation of cell growth, proliferation and
differentiation
24
. Mutations in the Ras genes are observed in up to 25% of all human cancers, which
makes Ras one of the major targets for cancer therapy
25-28
. Ras consists of a GTPase catalytic
domain (G domain) binding guanine nucleotides and the C-terminal peptide anchored to the inner
leaflet of plasma membrane through a posttranslational lipidation motif
29-31
. Membrane attachment
is crucial to Ras function: most effector proteins can only be activated by Ras-GTP when it is
attached to the membrane surface
25, 32
.
Ras proteins are represented by three Ras isoforms with a high degree of homology and
nearly 90% sequence identity in the N-terminal GTPase domain
33
. The remaining C-terminal 22-
23 amino acids, known as the hyper-variable region, have no sequence similarity except for the
conserved CAAX motif necessary for membrane targeting
34
. The variability of the C-terminal
sequences of the Ras isoforms leads to different processing patterns in the cell. All Ras isoforms
.CC-BY-NC-ND 4.0 International licenseunder a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available
The copyright holder for this preprint (which wasthis version posted November 17, 2017. ; https://doi.org/10.1101/221382doi: bioRxiv preprint

4
are modified by attachment of a prenyl (farnesyl) chain at the extreme C-terminal cysteine. H-Ras
and N-Ras additionally get two and one palmitoyl chains, respectively, while K-Ras4B features a
polybasic domain as alternative membrane-anchoring mechanism
29, 35
. It was demonstrated that the
membrane-targeting region is responsible for partitioning of proteins between membrane
domains
36
.
Dynamic change in H-Ras localization from cholesterol-rich rafts to the disordered lipid
domains upon activation (GTP binding) was first observed using density gradients and immuno-
gold electron microscopy in native cellular membranes
22, 37-40
. Explanation of this behavior of H-
Ras was proposed when scaffolding protein galectin-1 was found to associate with activated H-
Ras nanoclusters in disordered lipid domains
41
. The K-Ras isoform was found residing in the
disordered phase irrespective of its activation status (bound GDP or GTP)
22, 38
. The lipid domain
preferences of N-Ras remain controversial as it was observed in a raft phase of COS-7 cell
membranes when in the GDP form
42
, while Roy reported that N-Ras-GDP was localized in the
disordered lipid phase of BHK cells and moved to raft domains upon GTP binding
43
. Experiments
in model membranes recapitulated none of these findings: N-Ras was found concentrated at the
raft/disordered domain boundary in model lipid bilayers irrespective of the bound nucleotide
44-47
.
The cited reports characterized N-Ras behavior in very different membranes systems: from natural
plasma membranes of BHK and COS-7 cells to synthetic lipid mixtures, which might be one of
the causes of difference. The dynamic shift from one phase to another upon activation of N-Ras
observed by Roy et al.
43
could, potentially, be due to binding to yet-unidentified protein scaffolds
(by analogy with H-Ras). In the present report, we make use of a full-length semi-synthetic
lipidated N-Ras to demonstrate that it is capable of changing its nano-domain localization in model
lipid membranes in nucleotide-dependent manner in the absence of any other proteins.
.CC-BY-NC-ND 4.0 International licenseunder a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available
The copyright holder for this preprint (which wasthis version posted November 17, 2017. ; https://doi.org/10.1101/221382doi: bioRxiv preprint

5
RESULTS
Our goal was to assess relative affinity of N-Ras lipoprotein to raft and disordered lipid
domains in a model lipid system, and determine whether raft affinity of N-Ras is dependent on the
nature of a bound nucleotide (hence, the biologically active/inactive protein conformation) in the
absence of "helper" proteins. Because of the nanoscale dimensions of rafts, we relied on
measurements of FRET between Ras-attached fluorophore and fluorescent lipid domain markers
48-
50
. H-Ras localization was previously probed by FRET to lipid domain markers but those reports
did not include N-Ras
51, 52
.
In the following subsections we
(1) evaluated the model lipid bilayers to confirm that they form nanometer
ordered domains mimicking size of cellular rafts,
(2) detected non-raft localization of the C-terminal lipidated peptide of N-Ras,
(3) evaluated a hypothesis that the C-terminal peptide may be attracted to the raft
boundary,
(4) established lifetime-based detection of nanodomain localization, and
(5) detected distinct nanodomain preferences of N-Ras in active and inactive
states (bound to GTP mimic or GDP).
1. Lipid membrane mimic with nano-scale lipid domains
To create lipid bilayers that spontaneously forms nanometer-sized raft domains (approx.
ranging from 4 to 15 nm), we followed Pathak and London
50
and utilized a lipid mixture of
sphingomyelin (SM), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), and cholesterol
in the equimolar ratio (referred to in the following as the raft lipid mixture). A pure POPC lipid
was used to make a homogeneous (non-raft) control bilayers. Confocal fluorescence microscopy
.CC-BY-NC-ND 4.0 International licenseunder a
not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available
The copyright holder for this preprint (which wasthis version posted November 17, 2017. ; https://doi.org/10.1101/221382doi: bioRxiv preprint

Frequently asked questions

Q1. What contributions have the authors mentioned in the paper "Partitioning of semisynthetic lipidated n-ras in lipid raft nanodomains determined by fret to lipid domain markers" ?

CC-BY-NC-ND 4. 0 International license under a not certified by peer review ) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.