A Mechanistic Investigation of HIV-1 Gag Association with Lipid

Subscriber access provided by University of Rochester | River Campus & Miner Libraries

B: Biophysics; Physical Chemistry of Biological Systems and Biomolecules

A Mechanistic Investigation of HIV-1 Gag Association with Lipid Membranes Renee J. Tran, Matthew S. Lalonde, Krystal L. Sly, and John C. Conboy J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.9b02655 • Publication Date (Web): 14 May 2019 Downloaded from http://pubs.acs.org on May 16, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

A Mechanistic Investigation of HIV-1 Gag Association with Lipid Membranes Renee J. Tran1†, Matthew S. Lalonde2†, Krystal L. Sly1, and John C. Conboy1* 1Department

of Chemistry, University of Utah, 315 South 1400 East RM. 2020, Salt Lake City, Utah 84112

2Department

of Biochemistry, University of Utah, 15 N Medical Drive East RM. 4100, Salt Lake City, Utah 84112

†Contributed equally to the study *To whom correspondence should be addressed: [email protected]

1 ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 46

Abstract An extensive investigation into the initial association of HIV-1 Gag with lipid membranes was conducted with second harmonic generation (SHG). The roles of lipid phase, PI(4,5)P2, the presence of the myristoyl group on Gag, the C-terminus of Gag, and the presence of tRNA on Gag-membrane association were examined using the most physiologically relevant full-length Gag protein studied thus far. The tighter packing of a bilayer composed of gel phase lipids was found to have a lower relative amount of membrane-bound Gag in comparison to its fluid phase counterpart. Rather than driving membrane association of Gag, the presence of PI(4,5)P2 and the myristoyl group were found to anchor Gag at the membrane by decreasing the rate of desorption. Specifically, the interaction with PI(4,5)P2 allows Gag to overcome electrostatic repulsion with negatively charged lipids at the membrane surface. This behavior was verified by measuring the binding properties of Gag mutants in the matrix domain of Gag which prevented anchoring to the membrane either by blocking interaction with PI(4,5)P2 or by preventing exposure of the myristoyl group. The presence of tRNA was found to inhibit Gag association with the membrane by specifically blocking the PI(4,5)P2 binding region, thereby preventing exposure of the myristoyl group and precluding subsequent anchoring of Gag to the membrane. While Gag likely samples all membranes, only the anchoring provided by the myristoyl group and PI(4,5)P2 allow Gag to accumulate at the membrane. These quantitative results on the kinetics and thermodynamics of Gag association to lipid membranes provide important new information about the mechanism of Gagmembrane association.


Page 3 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Introduction HIV-1 group-specific antigen (Gag) is a 56 kDa protein which is integral in the assembly of the HIV-1 virus particle and the only viral protein required to initiate and complete the budding process.1-2 Gag is an N-myristoylated polyprotein comprising from the N- to C-terminus, the MA domain (which is cleaved and forms the matrix assembly in the mature virus), the CA domain (forms the capsid assembly in the mature virus), the NC domain (forms the nucleocapsid assembly in the mature virus), and the p6 domain.3 The MA domain also contains a highly basic region (HBR), which is believed to facilitate interaction with lipid membranes. More specifically, the HBR is thought to interact with the charged phospholipid 1,2-dioleoyl-sn-glycero-3-phospho-(1-myo-inositol-4,5-bisphosphate) (PI(4,5)P2), which is present in the plasma membrane at a concentration of 1-2 mol %.1, 4 The interaction of PI(4,5)P2 with the HBR is thought to induce a conformational change in the Gag protein that exposes the myristoyl group at the N-terminus to a linearized conformation, which then facilitates Gag association with the plasma membrane.5-6 Following association, Gag co-localizes to the plasma membrane where it oligomerizes with other Gag polyproteins, HIV-1 genomic RNA (gRNA), and the C-termini of gp41 envelope glycoprotein trimers, to form a nascent viral particle.7 While envelope glycoproteins and gRNA are required to form an infectious particle, they are not essential for the budding process. There have been several studies examining the mechanism of MA association with the plasma membrane and its role in viral particle budding.6,

8-9

However, the many activities of full-length myristoylated HIV-1 Gag that make it an

interesting protein (membrane binding, nucleic acid binding, oligomerization, etc.) also make it difficult to study. There have been few studies on the binding kinetics and energetics of full-length, myristoylated Gag to lipid membranes in the absence of other cellular components or the modulation of these properties due to membrane composition in the same context. Determining the roles of the myristoyl group, PI(4,5)P2, and MA on Gag membrane association, is crucial to understanding of the mechanism of Gag plasma membrane localization and ultimately how HIV-1 virus particles are generated. 3 ACS Paragon Plus Environment


Page 4 of 46

Gag-membrane interactions have been extensively studied using liposome flotation assays, which have determined the relative binding efficiency of only the MA domain, as the interaction of MA with the membrane is thought to drive Gag assembly at the plasma membrane.1, 10-11 The membranes in these studies mainly contained 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) and 1-palmitoyl-2oleoyl-sn-glycero-3-[phosphor-L-serine] (POPS) in a 2:1 ratio, doped with 7.25 mol % PI(4,5)P2. These studies found that membrane binding of wild-type myristoylated MA was impaired when PI(4,5)P2 was absent from the membrane, which suggested a specific interaction between MA and PI(4,5)P2.1,

7

Additionally, it was found that non-myristoylated MA did not bind to the liposome membrane, even in the presence of PI(4,5)P2, indicating that the myristoyl group was required for MA membrane association.8 The binding affinity of MA to bilayers composed of DOPC, DOPS, cholesterol, and brain PI(4,5)P2 was previously measured using surface plasmon resonance (SPR).12 However, the isolated MA may not accurately represent membrane affinity for the full-length polyprotein,13 highlighting the need for binding studies of the intact Gag protein. Previous studies have also implicated transfer ribonucleic acid (tRNA) interactions with Gag in the viral budding process.10, 14 tRNA is present in the cytosol of mammalian cells at a concentration of about 1 mg/mL.10 Concentrations of tRNA ranging from 0.01 to 0.1 mg/mL were observed to decrease the binding efficiency of myristoylated Gag (mGag) to liposomes, but the presence of PI(4,5)P2 in the membrane restored mGag binding in the presence of tRNA.10-11 This was thought to be due to competition of PI(4,5)P2 with tRNA binding to the HBR of Gag.14 Other studies of mutations in the matrix domain, such as V7R, were also found to impair binding by blocking the exposure of the myristoyl group.15 NMR studies of the V7R mutant indicated a conformational change in the protein that sequestered the myristoyl group, making it inaccessible for membrane intercalation.16 However, it is important to note that the NMR studies were also conducted using only the MA of Gag. While the V7R mutant impaired MA-membrane binding, the L21K mutant was found to restore binding of the V7R mutant to levels comparable to wild


Page 5 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


type (non-mutated) MA, possibly due to prevention of the conformational changes associated with the V7R mutant that lead to sequestration of the myristoyl group.17 Although most of these previous studies point to the importance of the myristoyl group in facilitating the Gag-membrane interaction, the binding efficiencies of the Gag variants were determined using SDS-PAGE and only able to provide relative liposome binding efficiencies, not binding affinities or the kinetics of the association/dissociation of Gag to the membrane. Additionally, the Gag concentrations used in these previous studies was not provided, making it difficult to relate the measured behavior to conditions when biologically relevant concentrations of Gag are present. These studies were also complicated by the lipid composition used in the liposome binding assay, which contained a mixture of POPC, POPS, and PI(4,5)P2. Since PI(4,5)P2 and POPS are both negatively charged, there is still debate whether the surface charge provided by PS is involved in the specific Gag-membrane binding that is observed in the presence of PI(4,5)P2.12, 18 The aim of the current study is to quantitatively determine the binding affinities and kinetics of full-length Gag to model membranes, to more thoroughly investigate the mode of action of the initial association of Gag with lipid membranes. Specifically, the dependence of Gag membrane association on myristoylation and PI(4,5)P2, as well as the effect of Gag point mutations have all been investigated here. The role of the C-terminus of Gag and the regulation of Gag-membrane binding by tRNA were also examined. Additionally, the role of the phase state of the membrane in the absence and presence of negatively charged lipids, PS and PG, were studied. In order to gain a quantitative understanding of the Gag-membrane interaction, the label-free technique of second harmonic generation (SHG) was used. SHG is a nonlinear spectroscopic technique which involves the up-conversion of light at frequency ω to 2ω. SHG is a surface specific and highly sensitive spectroscopic technique capable of quantitatively measuring protein adsorption to membranes, protein-ligand interactions at membranes and smallmolecule partitioning into membranes.19-22



Page 6 of 46

The surface specificity of SHG is ideal for studying Gag association to a membrane as the interaction of interest occurs at the surface of the membrane and can be performed in a true equilibrium state with free Gag present in solution. Furthermore, SHG is a label-free technique which does not require the attachment of an external fluorescent label or fluorescent protein tag, allowing direct measurement of surface interactions without alterations to the native conformation and binding properties of the protein.2324

SHG also has the sensitivity to detect sub-monolayer quantities down to fg/cm2.21 This sensitivity is

106 times greater than the SDS-PAGE assays previously used to investigate Gag-membrane interactions, which has a ng/cm2 sensitivity.25 All these properties make SHG ideal for providing a quantitative assessment of Gag-membrane association. Materials and Methods Materials: 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), 1,2-distearoyl-sn-glycero-3phosphocholine

(DSPC),

1,2-dioleoyl-sn-glycero-3-phospho-(1-myo-inositol-4,5-bisphosphate)

[PI(4,5)P2], and 1,2-dioleoyl-sn-glycero-3-phospho-(1’-rac-glycerol) (DOPG) were purchased from Avanti Polar Lipids (Alabaster, Alabama). 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), potassium hydroxide, sodium chloride, sodium phosphate monobasic, and Alexa Fluor 647-labeled ssDNA (d(TG)25-AF647) were purchased from ThermoFisher Scientific (Waltham, Massachusetts). Dmyo-inositol 1,4,5-tris-phosphate trisodium salt (Ins(1,4,5)P3), myristoyl-CoA, B-mercaptoethanol (BME), glycerol, imidazole, MgCl2, Tris, phenylmethylsulfonyl fluoride (PMSF), Pepstatin, Leupeptin, Aprotinin,

imidazole,

nonyl

phenoxypolyethoxylethanol

(NP-40),

dithiothreitol

(DTT),

and

polyethyleneimine (PEI) were obtained from Sigma Aldrich (St. Louis, Missouri). The Gag peptide with sequence SLFGNDPSSQ was obtained from the Bandarian Group (University of Utah). Yeast tRNA and DNAseI was obtained from Roche Diagnostics (Risch-Rotkreuz, Switzerland). GST-tagged Prescission Protease and GST-sepharose beads were obtained from GE Life Sciences (Pittsburgh, Pennsylvania). Bilayer Preparation: Planar supported lipid bilayers (PSLBs) were composed of DOPC or DSPC. These bilayers contained 2 mol % PI(4,5)P2 or 6 mol % DOPG. Each PSLB was prepared in water 6 ACS Paragon Plus Environment

Page 7 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


on a prism surface using the Langmuir-Blodgett/Langmuir-Schaefer method on a KSV Instruments minitrough, as described previously 26. After preparation, the PSLBs were mounted in a custom Teflon flow cell. The flow cell allowed protein solution to be flowed above the PSLBs and Gag-membrane binding was monitored in real-time using SHG. All flow cells and prisms were cleaned in a 70:30 solution of sulfuric acid: hydrogen peroxide before use. (Caution: This solution reacts volatilely with organic solvents. Extreme caution must be exercised when handling this solution.) The flow cells were flushed with 50 mM HEPES buffer containing 250 mM NaCl and adjusted to pH 7.4 to keep the PSLBs in a biologically relevant media. SHG Theory: The SHG intensity produced from the surface at 2ω is proportional to the square of the second-order susceptibility tensor (𝜒(2) 𝑖𝑗𝑘 ), composed of a nonresonant (NR) and resonant (R) portion: 2 (2) (2) 2 𝐼𝑆𝐻𝐺 ∝ |𝜒(2) 𝑖𝑗𝑘 | ∝ |𝜒𝑁𝑅 + 𝜒𝑅 | .

(1)

The nonresonant portion describes the off-resonance response from the interface (or surface) and is typically negligible for amorphous dielectric substrates, such as the silica prism used in this study. The primary contribution to the SH signal comes from the resonant portion, which describes the interaction of the incident electric field with the molecules at the surface such that ⟨𝑎│𝜇𝑖│𝑐⟩⟨𝑎│𝜇𝑗│𝑏⟩⟨𝑏│𝜇𝑘│𝑐⟩

𝜒(2) 𝑅 ∝ 𝑁∑𝛼,𝛽,𝛾(2ℎ𝜔 ― 𝐸𝑐𝑎 ― 𝑖𝛤𝑐𝑎)(ℎ𝜔 ― 𝐸𝑏𝑐 ― 𝑖𝛤𝑏𝑐)

(2)

where N is the number of molecules, h is Planck’s constant, µ is the dipole operator, Γ is the transition linewidth and a, b, and c are the initial, intermediate, and final electronic states of the molecules, proteins, or peptides at the interface, respectively. Examination of Eq. 2 shows that when the molecules have optical transitions at ω or 2ω, there is an enhancement in the SHG intensity. When studying proteins, such as HIV-1 Gag, an excitation wavelength of 532 nm produced by a Nd:YAG (neodymium-doped yttrium aluminum garnet) laser is ideal as the resulting SHG at 266 nm is on resonance with the π-π* transitions of the tryptophan, tyrosine, and phenylalanine amino acids present in the protein.27 Combining equations 1 and 2 yields the relationship: 𝐼𝑆𝐻𝐺 ∝ 𝑁2. 7 ACS Paragon Plus Environment

(3)


Page 8 of 46

which provides a direct means to quantify the number density of Gag at the membrane surface (N) to the measured SHG intensity (𝐼𝑆𝐻𝐺). SHG Measurements: Counter-propagating SHG was employed, which has previously been described in detail.28 This geometry directs the incident 532 nm beam (second harmonic output of a 10 Hz Nd:YAG laser from Continuum Surelite) at the prism/water interface, under total internal reflection with an incident angle of 67°. The reflected beam was directed back on itself using a mirror to temporally and spatially overlap with the incident beam, resulting in a SHG signal at 266 nm which was emitted normal to the surface. The signal was then collected by a solar-blind photomultiplier tube (Hamamatsu), after the fundamental light at 532 nm was removed with optical filters. The signal was processed with a boxcar integrator (Stanford Research Systems) and recorded on a computer running LabView. Preparation of Human N-Myristoyltransferase 2: The human N-myristoyltransferase 2 coding sequence (hNMT2, NM_001308295) was cloned into the pET151 N-terminal 6xHis tag expression construct using standard methods. Rosetta cells (Novagen) were transformed with pET151-hNMT2 and grown overnight on LB (lysogeny broth) agar supplemented with 50 g/mL ampicillin. Approximately 100 colonies were grown in 3 liters autoinduction medium29 for 6 hours at 37C, and then overnight at 23C shaking at 120 RPM. Bacteria were recovered by centrifugation at 4800 x g for 10 minutes at room temperature and subjected to three cycles of freezing and thawing on ice. Freeze-cycled pellets were resuspended in 50 mM HEPES (pH 7.5), 300 mM NaCl, 5 mM BME, 10% glycerol, 10 mM imidazole, 2 mM MgCl2, 0.5% NP-40, 60 µg/mL PMSF, 300 ng/mL Pepstatin, 1.6 µg/mL Leupeptin, and 3.2 µg/mL Aprotinin on ice via intermittent stirring over 45 minutes. Approximately 50 mg DNaseI was stirred into the lysate and incubated on ice for 15 minutes. The lysate was pelleted at 35,000 x g and the supernatants were bound to Ni-NTA beads (~30 mL slurry pre-equilibrated in lysis buffer) on a nutating mixer at 4C for 30 minutes in a gravity column. The column was washed with 1500 mL wash buffer comprising 20 mM HEPES (pH 7.5), 300 mM NaCl, 10 mM BME, and 20 mM imidazole. The column was eluted with 40 mL 20 mM HEPES (pH 7.5), 300 mM NaCl, 10 mM BME, and 250 mM imidazole. Eluate was 8 ACS Paragon Plus Environment

Page 9 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


dialyzed overnight against 2 liters of 20mM Tris (pH 8.0), 50 mM NaCl, and 10 mM BME at 4C. The protein was bound to a 5 mL Q-column (HP) at 3 mL/min, and eluted in escalating NaCl at 4C. hNMT2containing fractions were pooled and dialyzed against 2 liters 20 mM Tris (pH 7.5), 150 mM NaCl, and 10 mM BME. The dialysate volume was reduced to ~5 mL in an Amicon 30k concentrator and eluted through Sephadex S75 gel filtration resin at 1 mL/min. hNMT2-containing fractions were pooled and dialyzed into 30 mM Tris (pH 7.5), 150 mM NaCl, 0.5 mM EDTA, 0.5 mM EGTA, 5 mM BME, and 50% glycerol. Preparation of Full-Length HIV-1 Gag: The HIV-1 Gag coding sequence (a gift from Jeremy Luban and Cagan Gurer) harbors 72 silent mutations, (see Supporting Material Table S1, some reported in30) relative to reference strain HXB2 (NCBI accession K03455), to both optimize for protein expression, and disable the so-called “slippery sequence” which otherwise induces ribosomal frameshifting in the Gag p6 domain. This protein coding sequence was cloned into a pET11a vector (Novagen) modified to express a C-terminal protease-cleavable GST tag using standard methods. Rosetta cells were transformed with pET11a-Gag-PP-GST (wild-type or harboring the required mutations), spread on LB agar supplemented with 50 g/mL ampicillin and grown overnight at 37C. Approximately 100 colonies were used to inoculate 2 liters autoinduction medium,29 and grown for 6 hours at 37C, and then overnight at 23C shaking at 120 RPM. Bacteria containing Gag proteins were recovered in pellets after centrifugation at 4800 x g for 10 min at room temperature. Bacterial pellets were subjected to three cycles of overnight freezing and subsequent thawing on ice. The freeze-thaw-cycled pellets were then further lysed with 100 mL buffer comprising of 50 mM HEPES (pH 8.0), 500 mM NaCl, 2 mM MgCl2, 0.1% NP-40, 5% glycerol, 40 mM DTT, 60 µg/mL PMSF, 600 ng/mL Pepstatin, 1.6 µg/mL Leupeptin, and 3.2 µg/mL Aprotinin. Approximately 50 mg lyophilized DNAseI was manually stirred into the lysate, which was then homogenized via 5-10 strokes in a Dounce homogenizer. Supernatants containing Gag protein were recovered after centrifugation at 35,000 x g for 20 minutes, and slowly combined with 0.11 volumes 10% PEI with stirring, then stirred for an additional 30 minutes at 4C. Supernatants were recovered after a 209 ACS Paragon Plus Environment


Page 10 of 46

minute centrifugation at 9800 x g, and slowly combined with 0.3 volumes saturated ammonium sulfate at 4C with stirring, then stirred for an additional 30 minutes. Pellets were recovered after 20-minute centrifugation at 9800 x g, and resuspended in 75 mL resuspension buffer comprising 50 mM HEPES (pH 8.0), 500 mM NaCl, 5% glycerol, and 40 mM DTT via 5-10 strokes in a Dounce homogenizer and filtered (0.45 um pore cellulose acetate). The resuspended pellets were then filtered (0.45 m pore size) and bound to a GST-sepharose beads (~15 mL slurry pre-equilibrated with resuspension buffer) at 4C in a gravity column for 1 hour on a nutating mixer. The column was drained and the resin-bound Gag was washed with 400 mL resuspension buffer. Gag protein was cleaved from the column in 75 mL wash buffer supplemented with 150 g GST-tagged Prescission Protease (HRV 3C protease fused to GST) at 4C overnight on a nutating mixer. The column eluate was collected and dialyzed overnight against 50 mM HEPES, 250 mM NaCl, 40 mM DTT, and 5% glycerol. HIV-1 Gag Myristoylation: Each Gag variant was myristoylated by combining Gag protein, hNMT2, and myristoyl-CoA in a molar ratio of 3:1:4 and rocking on a nutating mixer at room temperature for 24 hours. Myristoylation was verified by intact electrospray-ionization (ESI) mass spectrometry after adsorption to hydrophobic resin (C4 Zip Tip) and elution in 100% methanol, wherein the myristoylated species had ~210 Da more mass than the unmyristoylated protein. For every myristoylation reaction, spectra for unmyristoylated protein were obtained on the same day to compensate for instrument calibration drift. Microdialysis was run for two hours with stirring and exchanging with HEPES buffer at pH 7.4, using a 10 kDa molecular weight filter, to remove any unreacted myristoyl groups. The mGag and Gag solutions were used after microdialysis. Gag-ssDNA Binding Assay: Gag protein concentrations were estimated in Coomassie-stained denaturing polyacrylamide gels relative to bovine serum albumin reference standards run in the same gel. Nucleic acid binding activity was measured using a gel shift assay adapted from Berkowitz et al.31 Gag polyproteins were combined with a 40 nM single-stranded oligomer comprising 25 TG repeats labeled at the 3’-end with Alexa Fluor 647 (d(TG)25-AF647) in 20 mM Tris (pH 7.5), 100 mM NaCl, and 0.3 mM 10 ACS Paragon Plus Environment

Page 11 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


MgCl2. Binding reactions were incubated for 1 hour at 37C, then transferred to ice and combined with loading buffer (7% glycerol (final) and bromophenol blue). Reactions were resolved in (1.5 mm thick) 6% polyacrylamide/1% glycerol gel in Tris-Borate buffer at 0.3 W/cm. Gels were visualized on a Typhoon fluorescence reader (GE Healthcare) using appropriate filters for Alexa Fluor 647. The free probe band was measured in all lanes using ImageJ. Following the unbound single-stranded DNA (ssDNA) was the most practical since the signal from the fully-bound oligonucleotide was not apparent in the gel (possibly due to a net-neutral or net-positive charge of the fully-bound species), and apparent dissociation constants (Kd) were calculated by fitting a four-parameter logistic curve 𝐼=

𝐼𝑢𝑛𝑏𝑜𝑢𝑛𝑑 ― 𝐼𝑏𝑜𝑢𝑛𝑑 1+(

𝐺

ℎ

𝐾𝑑)

+ 𝐼𝑏𝑜𝑢𝑛𝑑

(4)

where I is the integrated density of the unbound ssDNA band in the gel image at some Gag concentration (G), Iunbound is the integrated density when the ssDNA is 0% bound, Ibound is the asymptotic integrated density of the unbound band when Gag is maximally bound, Kd is the apparent dissociation constant, and h is the Hill coefficient. Gag-Membrane Binding Assays: Increasing concentrations from 50 nM to 1600 nM of mGag, Gag, the Gag peptide, or one of the mutants (mGagV7R, GagV7R-L21K, GagL21K) in 50 mM HEPES buffer were injected into the flow cell and allowed to reach steady-state equilibrium with the PSLBs. Steady-state equilibrium was assumed when the SHG signal showed no further increase upon repeated incubation of the membrane at a specific concentration. For specific studies on the effect of the PI(4,5)P2 headgroup, a solution of 0.1 mg/mL Ins(1,4,5)P3 was prepared and incubated with mGag for 90 minutes. For studies involving tRNA, mGag and Gag were incubated with a 0.01 mg/mL solution of yeast tRNA for 1 hr. Each protein solution was freshly diluted before injection. The SHG intensity as a function of the bulk protein concentration was collected in order to obtain binding isotherms. The isotherms were fit using the Langmuir adsorption model, 𝐼1/2 𝑆𝐻𝐺 𝐼1/2 𝑚𝑎𝑥

𝐾𝑎[𝑃]

= 1 + 𝐾𝑎[𝑃] 11

ACS Paragon Plus Environment

(5)


Page 12 of 46

where ISHG is the SH intensity, Imax is the SH intensity at binding saturation, Ka is the binding affinity, and [P] is the protein concentration. The dissociation constant Kd can be obtained by taking the inverse of Ka. The recorded SHG intensities for mGag, Gag, and the mutants were normalized from experiment to experiment using a two point calibration by recording the SHG intensity of a solution from a 10 mM KOH and a phosphate buffered saline (PBS) solution composed of 100 mM NaCl and 50 mM NaH2PO4, which were injected into the flowcell after each experiment. The kinetics of mGag and Gag dissociation were obtained by rinsing the flow cells with excess HEPES buffer after the protein had reached equilibrium saturation to the membrane and the SHG intensity was measured as it decreased over time, followed by baseline subtraction for any electronic offset. The desorption rate, koff, was obtained by fitting the SHG intensity decay curves to 𝐼𝑆𝐻𝐺 = 𝑎𝑒 ― 𝑘𝑜𝑓𝑓𝑡

(6)

where a is the amplitude at surface saturation, and t is time. The adsorption rate (kon) was then obtained using the relation 1

𝐾 𝑑 = 𝐾𝑎 =

𝑘𝑜𝑓𝑓 𝑘𝑜𝑛

(7)

where Kd was determined from the steady-state adsorption isotherms, discussed above.

Results and Discussion HIV-1 Gag Purification, Myristoylation, and ssDNA Binding. After protein expression, purification, and cleavage from GST resin, purified HIV-1 Gag protein preparations had expected sizes in Coomassie-stained polyacrylamide gel with no obvious truncation products (Figure 1A). This was also confirmed by the intact-ESI measurements performed to confirm myristoylation (Figure 1C). For in-vitro myristoylation, we found that the ratio of myristoyl-CoA to Gag protein could not exceed ~1.3, else Gag protein rapidly precipitated from the solution (data not shown), so the reagent concentrations were kept near equimolar in slight favor of myristoyl-CoA. When the unmyristoylated 12 ACS Paragon Plus Environment

Page 13 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


species was no longer detectable by intact-ESI (typically 24-30 hours, Figure 1D), reactions were aliquoted and flash frozen in liquid nitrogen. N-myristoyltransferase was sometimes apparent in the mass spectra as a very low-intensity peak at ~60,730 Da, but was often absent, despite that its concentration was 1/3 that of HIV-1 Gag. This is not surprising since the desalting and solubilization conditions (adsorption to C4 hydrophobic resin, elution in methanol) were developed to detect Gag protein, and we did not expect hNMT2 to have similar solubility in methanol, or similar affinity for hydrophobic resin. The HIV-1 Gag NC domain contains two zinc finger motifs which must be in the reduced and metal-bound state to support viral replication via recruitment of HIV genomic RNA to the budding site.32 Early work by Fisher et al.33 showed that NC preferentially binds to single-stranded DNA oligonucleotides comprising alternating Ts and Gs (d(TG)n), and that binding requires intact zinc fingers. In earlier attempts to purify full-length Gag, purified Gag proteins had no detectable affinity for d(TG)25 (data not shown), so this property was used as a quality metric for developing a better expression/purification protocol. To confirm that the newly-developed purification protocol gave protein that approximated known nucleic acid binding properties of the isolated NC domain,34 we determined binding affinities for d(TG)25. The ssDNA was kept constant in binding reactions containing escalating concentrations of each Gag protein. Bound and unbound ssDNA were separated in native polyacrylamide gels and visualized by fluorescence imaging (Figure 1B). Several species observed in the native gel (Fig 1B “a” and “b”) were interpreted as lower-order (Fig 1B “a”), and higher-order Gag:d(TG)25 complexes (Fig 1B “b”). At the highest Gag concentrations, the ssDNA did not enter the gel, indicating that the largest Gag:d(TG)25 complexes had a non-negative overall charge, or were otherwise too large to enter the gel. Due to the difficulty in tracking the ssDNA probe in higher-order complexes, apparent Kd values were calculated by tracking the disappearance of the free probe in the gel images (values in Table 1). Apparent Kd values were both measurable and roughly consistent with published values for purified NC domain binding to ssDNA,34 indicating that the NC domain zinc fingers were properly folded and able to coordinate divalent metal ions. Moreover, all Gag 13 ACS Paragon Plus Environment


Page 14 of 46

variants bound ssDNA cooperatively, with Hill coefficients that were >2 (Table 1), consistent with what would be expected from a protein that oligomerizes via protein-protein interactions on nucleic acid as a viral particle assembles.

Figure 1. Characterization of purified HIV-1 Gag variants. (A) Coomassie gel of the wild-type Gag protein. (B) Electrophoretic mobility shift assay (EMSA) to detect d(TG)25-AF647 mobility with escalating Gag protein. Labels “a” and “b” indicate two discrete Gag:ssDNA complexes observed at intermediate Gag concentrations. We consistently observed that the ssDNA did not enter the gel when the Gag concentration was high, possibly due to a net positive charge on complexes with 3 or more Gag proteins. (C) Intact-ESI spectrum of wild-type HIV-1 Gag protein, theoretical mass = 56528.7 Da. (D) 14 ACS Paragon Plus Environment

Page 15 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Intact-ESI for wild-type Gag protein after 30 hours myristoylation. The expected mass of the myristoylated wild-type protein was 56738.7 Da (that of the unmyristoylated +210 Da). Table 1. Gag/ssDNA Hill coefficients and apparent dissociation constants.

Protein

Apparent Kd (Gag:d(TG)25, nM)

Hill Coefficient

Gag(wt)

311.2

2.5

mGag(wt)

253

4.2

mGagV7R

192.6

3.1

GagV7RL21K

165.8

2.4

GagL21K

270.3

2.8

Gag Binding to Neutral Membranes. As the association of Gag at the plasma membrane is integral to viral particle assembly, the present study aims to dissect the interactions necessary for Gagmembrane association. The direct interaction of Gag to a neutral lipid membrane has not been investigated in detail, as the previous membrane flotation studies used vesicles composed of a mixture of charged and neutral lipids.1, 7, 10 Measuring this interaction is crucial as it establishes the nonspecific binding propensity of the protein in the absence of electrostatic or specific ligand interactions. Another study investigated Gag binding to giant unilamellar vesicles (GUVs) containing 1,2-diphytanoyl-sn-glycero-3phosphocholine (DPhPC) found that Gag formed punctate structures on the GUV membrane, indicative of Gag binding and oligomerization on the neutral membrane; however, the binding affinity and surface density of that interaction was not measured.3 As a control, the association of mGag to neutral lipid bilayers in the gel and liquid crystalline (l.c.) states was measured by SHG. 1,2-dioleoyl-sn-glycero-3phosphocholine (DOPC), a phospholipid which is neutral at pH 7.4 with a transition temperature of 17°C, is in the l.c. fluid phase at room temperature and was utilized in the present study. In order to investigate the effects of lipid phase, Gag association to 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) membranes was also measured. DSPC has the same headgroup and chain length as DOPC, but 15 ACS Paragon Plus Environment


Page 16 of 46

has saturated acyl chains with a higher transition temperature of 55°C, resulting in a solid-like gel phase at room temperature.35 This allows comparison of Gag-membrane binding to fluid and gel phase membranes at the same temperature. The measured dissociation constant of mGag for a DOPC bilayer was 16 ± 7 nM which was significantly lower than the dissociation constant of mGag for a DSPC bilayer (143 ± 41 nM), Figure 2. At a surface pressure of 30 mN/m used to create the membrane, the mean molecular area for DOPC is approximately 65 Å2/molecule, whereas DSPC has a mean molecular area of

approximately 45

Å2/molecule, which means the DSPC membrane is tighter packed than DOPC at the same surface pressure.36 In addition, the surface density of mGag at membrane saturation was more than 50% lower for DSPC than DOPC, indicating that the tighter packing of the gel phase lipids hindered the adsorption of mGag to the membrane. Although the membrane flotation studies did not investigate the role of lipid phase state,1, 7, 11 the current study shows that there is a considerable propensity for mGag to bind to a neutral membrane in the absence of PI(4,5)P2 and other charged lipid components such as PS. It is also apparent that the phase of the lipids in the membrane plays a role in Gag-membrane association, with a lower dissociation constant of the protein found for the fluid phase membranes of DOPC. The SHG results are consistent with a previous fluorescence study where Gag was found to mainly associate with domains consisting of fluid phase lipids.37 The binding of mGag to a DOPC membrane serves here as a baseline for comparing the effect of PI(4,5)P2 and other charged lipids, Gag myristoylation, and Gag mutations on the affinity and saturation coverage of the protein, in the data that follows.


Page 17 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2. Relative surface density of mGag as a function of bulk concentration for both DOPC (circles) and DSPC (diamonds) bilayers. The solid and dashed lines are fits to Eq. 5. Error bars represent the standard deviation from three independent experiments. Effect of Myristoylation and PI(4,5)P2 on Gag-Membrane Binding. With the baseline of Gag binding to a pure DOPC membrane established, the effect of adding PI(4,5)P2 to the membrane was examined. Proper targeting of Gag to the plasma membrane has been shown to be dependent on the presence of PI(4,5)P2.38 In order to elucidate the role of PI(4,5)P2 on Gag-membrane association, the binding characteristics of mGag and Gag to DOPC membranes containing PI(4,5)P2 were measured by SHG. Membrane association was measured as a function of mGag and Gag concentrations ranging from 50 to 1600 nM (Figure 3), and the data were fit to Eq. 5 to obtain the dissociation constants. The dissociation constant of mGag to bilayers containing PI(4,5)P2, listed in Table 2, was 2.6-fold larger than the Kd of mGag for a neutral DOPC membrane, indicating the presence of PI(4,5)P2 does not appear to significantly enhance membrane association. These results are contradictory to the previous SDS-PAGE studies where a large enhancement in mGag binding was observed when 7.25 mol % PI(4,5)P2 was incorporated in the membrane.1, 10-11 However, one of those SDS-PAGE studies also showed that the binding efficiency was less than 10% for both 0 and 2.1 mol % PI(4,5)P2.1 In addition to the use of fulllength Gag, the concentration of PI(4,5)P2 chosen for the current study more closely represents the biological concentration found in cells (1-2 mol %).4 The similar dissociation constants in the presence and absence of PI(4,5)P2 indicate that at biologically relevant quantities of PI(4,5)P2, Gag-membrane association is not significantly increased, consistent with the small increase in binding efficiency observed in an SDS-PAGE study.1 Previous membrane flotation studies also indicated that the myristoyl group was required for membrane association.1, 11 The association of mGag and Gag to DOPC and DOPC + PI(4,5)P2 bilayers were compared to determine the role of the myristoyl group on Gag-membrane association. The Kd values of both mGag and Gag (Table 2) were the same order of magnitude regardless of the presence of PI(4,5)P2 17 ACS Paragon Plus Environment


Page 18 of 46

in the bilayer, indicating Gag can still associate with the membrane without the myristoyl group. Although the relative binding efficiency of Gag was slightly lower than mGag, this result is consistent with previous studies which measured MA binding even with the removal of the myristoyl group in liposome flotation assays.1,

9

Although results from the liposome flotation study has shown that the myristoyl group is

somehow involved in membrane association,9 the similar dissociation constant of nonmyristoylated Gag to that of mGag obtained by SHG suggests the myristoyl group plays a role outside the enhancement of the binding affinity. The affinity of Gag for the membrane is only part of the story, as the relative amount of bound Gag is also believed to be crucial for the assembly of viral particles.1,

7

The SH intensity at protein

saturation can be used to compare the relative surface densities of Gag or mGag on the membrane. Shown in Figure 3, although the pI of Gag is 8.1 and the net charge at pH 7.4 is positive, the relative amount of both membrane-bound mGag and Gag did not change in the presence or absence of PI(4,5)P2, which is in agreement with a previous study where a concentration of 2.1% PI(4,5)P2 enhanced binding efficiency of MA by less than 10%.1 However, the largest enhancement in the relative amount of bound MA was for 7.25% PI(4,5)P2 in the membrane flotation studies, where the authors concluded that the observed enhancement in binding efficiency was due to the increased negative charge of the liposomes, rather than specificity for PI(4,5)P2.1 While the surface density of mGag measured by SHG was not significantly different in the presence and absence of PI(4,5)P2, the surface densities of mGag were 10% higher compared to the surface densities of Gag to the same bilayers. This is in line with the membrane flotation studies where the relative binding efficiency was increased in the presence of the myristoyl group.11 As stated earlier, the myristoyl group, consisting of a 14-carbon saturated fatty acid, is thought to intercalate into the membrane, interacting with the hydrophobic acyl chains of the lipids upon Gag-membrane association. Intercalation of the myristoyl group may play a role in the kinetics of membrane association that results in a lower Kd of mGag at the membrane surface.


Page 19 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. Relative surface density as a function of bulk concentration of mGag (squares) and Gag (diamonds) binding to a DOPC + 2 mol % PI(4,5)P2 bilayer. mGag (circles) and Gag (triangles) binding to a pure DOPC bilayer are shown for comparison. Solid and dashed lines are fits to Eq. 5. Error bars represent the standard deviation of three independent experiments. Inset shows relative surface density upon addition of mGag to a DOPC bilayer in the presence of Ins(1,4,5)P3 (gray bar) and in the absence of Ins(1,4,5)P3 (striped bar). The Kinetics of mGag and Gag Association. In addition to the thermodynamics of mGag and Gag association to lipid membranes, the kinetic rates of association and dissociation reveal important information about the protein-membrane interaction. The adsorption (kon) and desorption (koff) rates for mGag and Gag binding to a DOPC bilayer and a DOPC + PI(4,5)P2 bilayer were measured in order to examine the effects of PI(4,5)P2 and the myristoylation of Gag on the binding kinetics. The measured koff rate for Gag desorption from a DOPC + PI(4,5)P2 bilayer was half the rate of desorption of Gag from a DOPC bilayer, shown in Figure 4, suggesting that the presence of PI(4,5)P2 prolongs the interaction of the protein at the surface of the membrane. However, the adsorption rates of Gag to a DOPC bilayer and DOPC + PI(4,5)P2 bilayer were similar (Table 2). The decrease in koff could be attributed to the specific interaction of PI(4,5)P2 with the protein, which prevents dissociation after adsorption occurs. The dissociation of mGag from a DOPC bilayer was 2.6-fold slower than the koff of Gag from a DOPC bilayer as well, Figure 4A. However, the kon of mGag to a DOPC bilayer was within error of the rate of adsorption 19 ACS Paragon Plus Environment


Page 20 of 46

of Gag. This is consistent with the myristoyl group anchoring Gag to the membrane, as the hydrophobic chain prefers to interact with the acyl chains in the core of the bilayer. Although Gag still associates with the membrane, it dissociates faster than mGag as it is not anchored by the myristoyl group, resulting in the lower surface density observed in the adsorption isotherm.

Figure 4. Desorption data for (A) mGag from a DOPC bilayer, (B) Gag from a DOPC bilayer, and (C) Gag from a DOPC + 2 mol % PI(4,5)P2 bilayer. The solid and dashed lines are fits to Eq. 6. The exposure of the myristoyl group through interaction with PI(4,5)P2 is thought to allow anchoring of the protein to the membrane.2,

5

To test this hypothesis, Ins(1,4,5)P3, a water-soluble

analogue of the PI(4,5)P2 lipid headgroup, was used to examine whether binding of mGag was enhanced 20 ACS Paragon Plus Environment

Page 21 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


by exposing the myristoyl group prior to membrane association by incubating mGag with 0.1 mg/mL Ins(1,4,5)P3, before association to a bilayer. The membrane associated mGag was quantified for a single mGag concentration of 400 nM. The surface density of mGag pre-incubated with Ins(1,4,5)P3 was not significantly different from the surface density of mGag bound to a pure DOPC bilayer, Figure 3 inset. However, pre-exposure to Ins(1,4,5)P3 did not significantly increase the mGag density at the membrane surface, indicating that the mGag can anchor to the membrane without first interacting with PI(4,5)P2. The results observed here are consistent with the kinetics of adsorption/desorption, which suggest that the myristoyl group is not required for initial membrane adsorption. Although PI(4,5)P2 does not appear to play a role in exposing the myristoyl group on Gag, it has been proposed that enhanced binding to PI(4,5)P2 was non-specific electrostatic interaction between Gag and the charged membrane.1 Specifically, the negatively charged lipid PI(4,5)P2 headgroups are thought to interact with the positively charged HBR of Gag, stabilizing the association of the protein with the membrane.7 DOPG, another charged lipid with the same acyl chain length as PI(4,5)P2 was used to investigate the effect of electrostatics on Gag-membrane binding. The binding of Gag and mGag that were injected over DOPC bilayers containing 6 mol % DOPG is shown in Figure 5. At pH 7.4, the charge of PI(4,5)P2 is -3, whereas the charge of DOPG is -1, therefore 6 mol % DOPG has the equivalent net charge of 2 mol % PI(4,5)P2.39-40 The dissociation constants of both mGag and Gag for bilayers containing DOPG were within error, shown in Table 2. These values were also twice that of the Kd values measured to bilayers with 2 mol % PI(4,5)P2. However, the relative surface densities at saturation of mGag and Gag binding to 6 mol % DOPG were respectively 60% and 50% less than those measured for PI(4,5)P2, which indicates the protein is repelled by the presence of DOPG. The decrease in surface density strongly indicates the association of Gag (or mGag) is not driven by electrostatic interactions. This is consistent with previous SDS-PAGE studies, where binding efficiency decreased for bilayers containing a mixture of PC and the negatively charged PS in the absence of PI(4,5)P2.1 Although 6 mol % DOPG and 2 mol % PI(4,5)P2 have the same overall charge in the bilayer, there is apparently a specificity of Gag for PI(4,5)P2 21 ACS Paragon Plus Environment


Page 22 of 46

that overcomes the electrostatic repulsion observed for bilayers containing DOPG. This is consistent with a SPR study which showed that MA binding to membranes containing PI(4,5)P2 is more specific than a purely electrostatic interaction.12

Figure 5. Relative surface density as a function of concentration of mGag (circles) and Gag (triangles) bound to a DOPC + 6 mol % DOPG bilayer. The short light dashed line shows binding of mGag to a DOPC + 2 mol % PI(4,5)P2 bilayer for comparison. The darker solid and dashed lines are fits to the Eq. 5. Error bars represent standard deviation determined from independent experiments. The binding kinetics for Gag in the presence of DOPG was compared to the kinetics of Gag in the presence of a membrane containing PI(4,5)P2 to examine the effect of charged lipids in the absence of the myristoyl group, results shown in Figure 6. The koff of Gag from 6 mol % DOPG containing bilayers was 2-fold faster than the desorption rate of Gag from a 2% PI(4,5)P2 + DOPC bilayer, suggesting a specific interaction between Gag and PI(4,5)P2 which occurs after association with the membrane and which overcomes the electrostatic repulsion between Gag and the negatively charged membrane surface, prolonging Gag association to the bilayer. This is evident from the observation that the kon measured for Gag to a DOPC + 6 mol % DOPG bilayer was within error of the adsorption rate of Gag to a DOPC + 2 mol % PI(4,5)P2 membrane, strongly suggesting that Gag-PI(4,5)P2 binding occurs after membrane association, slowing dissociation due to a specific lipid-protein interaction, that is not present when PI(4,5)P2 is replaced by DOPG. These results also suggest that electrostatics are not the driving force for Gag association with the membrane, hinting that the HBR may play a different role in the Gag-membrane 22 ACS Paragon Plus Environment

Page 23 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


interaction, namely stabilization of the protein at the membrane via a specific interaction with PI(4,5)P2 that increases the residence time of Gag at the membrane surface.

Figure 6. Desorption data for Gag from a DOPC + 6 mol % DOPG bilayer. The solid line is a fit to Eq. 6. Table 2. Measured dissociation constants (Kd), adsorption rates (kon), and desorption rates (koff) of mGag and Gag to lipid bilayers of various compositions. N/A indicates no data available. mGag

Gag

Bilayer Composition

Kd (nM)

kon × 105 (M-1 s-1)

koff × 10-2 (s-1)

Kd (nM)

kon × 105 (M-1 s-1)

koff × 10-2 (s-1)

DOPC

16 ± 7

3±1

0.5 ± 0.1

42 ± 7

3.1 ± 0.9

1.3 ± 0.4

2% PI(4,5)P2 + DOPC

35 ± 7

N/A

N/A

42 ± 5

1.7 ± 0.3

0.7 ± 0.1

6% DOPG + DOPC

71 ± 20

N/A

N/A

83 ± 35

3±2

2±1

Effect of Matrix Domain Mutations on Gag-Membrane Binding. As the HBR in the matrix domain is thought to be responsible for specific interaction with PI(4,5)P2, mutations in this region may have significant effects on the affinity of the protein to the membrane. SDS-PAGE studies have demonstrated that the V7R mutation resulted in a decreased binding efficiency of MA to liposomes compared to wild type MA.17 The V7R mutation was found to induce a conformational change in Gag, 23 ACS Paragon Plus Environment


Page 24 of 46

allowing the arginine side chain to interact with the glutamate at position 52, preventing the myristoyl group from being exposed upon binding to bilayers with PI(4,5)P2.16-17 The L21K mutation was suggested to compensate for the V7R mutation by allowing the myristoyl group to be exposed in the absence of PI(4,5)P2, resulting in a higher membrane binding efficiency and anchoring of MA to the membrane.16-17 mGagV7R Mutant: In the current study, the binding properties of the V7R, V7R-L21K, and L21K mutants of Gag to membranes in the presence and absence of PI(4,5)P2 were examined. Gag coverage to these membranes was measured for bulk concentrations of each of the Gag mutants ranging from 50-1600 nM, Figure 7. The relative surface densities of mGagV7R to bilayers with and without PI(4,5)P2 were reduced by 35% and 50%, respectively, compared to the surface density of wild type mGag, consistent with the reduced MA-membrane binding observed in the SDS-PAGE study.17 The Kd of mGagV7R in the presence of PI(4,5)P2 was 4-fold higher than the Kd of wild type mGag (Table 3). Conversely, the dissociation constant of mGagV7R in the absence of PI(4,5)P2 was 2-fold higher than wild type mGag. The decrease in surface density and increase of Kd could be due to the mutation sequestering the myristoyl group, in addition to blocking the HBR and preventing PI(4,5)P2 from interacting specifically with the protein.41 GagV7R-L21K Mutant: The double mutant GagV7R-L21K, lacking the myristoyl group, had a measured dissociation constant which was half that of the Kd for the single mutant mGagV7R, suggesting that the L21K mutation restores the binding affinity of Gag. The addition of the second mutation also was observed to restore some binding to a bilayer containing PI(4,5)P2 (Figure 7). However, there was still 22% less surface density at saturation of GagV7R-L21K compared to the wild type Gag. This is consistent with the L21K mutation compensating for the V7R mutation, as seen in previous studies,16 although the compensation observed presently is not sufficient to completely restore binding when both mutations are present. The partial compensation in the presence of PI(4,5)P2 suggests the double mutant may expose the myristoyl group, but still obstructs the HBR, preventing the specific interaction of Gag with PI(4,5)P2.


Page 25 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


GagL21K Mutant: As described earlier, the binding efficiency of the L21K mutant was found to restore binding comparable to the wild type Gag in previous studies.17 In the current study, the dissociation constants of GagL21K were within error of the Kd values measured for wild type Gag in the presence and absence of PI(4,5)P2. However, the surface density of GagL21K to a bilayer containing PI(4,5)P2 was still 15% lower compared to the surface density of wild type Gag, indicating that the conformational change due to L21K might prevent efficient interaction with PI(4,5)P2, consistent with the partial compensation of the double mutant observed in the presence of PI(4,5)P2. Conversely, there was a 45% enhancement in membrane-bound GagL21K compared to wild type Gag bound to a DOPC bilayer, meaning the association with a bilayer containing only DOPC was enhanced by the L21K mutation, which was also observed in a previous SDS-PAGE study that observed increased binding efficiency of MA with mutations in the HBR and in the absence of PI(4,5)P2.17

Figure 7. Relative surface density as a function of concentration of (A) mGagV7R (circles) and GagL21K (squares) binding to a DOPC bilayer and (B) mGagV7R (circles), GagV7R-L21K (triangles), and GagL21K (squares) binding to a DOPC + 2 mol % PI(4,5)P2 bilayer. The short light dashed lines represent the binding of wild type mGag to a 2 mol % PI(4,5)P2 doped DOPC bilayer. Solid and long dashed lines are fits to Eq. 5. Error bars represent the standard deviation between independent experiments.



Page 26 of 46

A comparison of the V7R and L21K mutant adsorption/desorption kinetics provides further evidence of the compensatory effects of these mutations. The desorption rate of mGagV7R from pure DOPC was within error of the koff of mGagV7R from a DOPC + PI(4,5)P2 bilayer, Figure 8. These rates were 4.8-fold faster than the rate of desorption for wild type mGag, indicating the myristoyl group of the mutant is unable to intercalate into the membrane. The lack of myristoyl intercalation is consistent with a conformation change induced by the V7R mutation which sequesters the myristoyl group, thereby preventing anchoring after adsorption of the protein to the membrane surface.16 The kon of mGagV7R is within error of the adsorption rate of wild type mGag in the absence of PI(4,5)P2, confirming the hypothesis that PI(4,5)P2 interacts with Gag after the initial adsorption of the protein to the membrane. However, the desorption rate of GagV7R-L21K from DOPC membranes containing PI(4,5)P2 was 3-fold faster than the wild type Gag from the same bilayer composition. The faster rate of desorption suggests a lack of anchoring to the membrane by the myristoyl group and PI(4,5)P2 being unable to specifically interact with the HBR. There was no measurable desorption of GagL21K from a DOPC bilayer, which could be responsible for the 45% increase in surface density in comparison to wild type Gag. However, the rate of desorption for GagL21K from a bilayer containing PI(4,5)P2 was within error of both the mGagV7R and GagV7R-L21K mutants, signifying mGagV7R is not anchored by the myristoyl group, as both GagV7R-L21K and GagL21K lack the myristoyl group. In addition, the rate of adsorption of GagL21K to a bilayer containing PI(4,5)P2 was the same order of magnitude as both of the other mutants, indicating the blockage of the PI(4,5)P2 binding site by the L21K mutation. This alteration does not affect adsorption, which is consistent with the hypothesis that PI(4,5)P2 binding to the HBR occurs after the primary membrane adsorption event. Although the L21K mutation can compensate for the V7R mutation, which prevents exposure of the myristoyl group, specific interaction with PI(4,5)P2 is still blocked, thereby resulting in a lower surface density than wild type Gag.


Page 27 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 8. Desorption data for mGagV7R from (A) a DOPC and (B) DOPC + 2 mol % PI(4,5)P2 bilayer. Desorption of (C) GagV7R-L21K and (D) GagL21K from a DOPC + 2 mol % PI(4,5)P2 bilayer. The solid and dashed lines are fits to Eq. 6. Table 3. Measured dissociation constants (Kd), adsorption rates (kon), and desorption rates (koff) of the Gag mutants to lipid bilayers of various compositions. N/A indicates no data available and ND indicates no observed desorption. DOPC

2% PI(4,5)P2 + DOPC

Mutant

Kd (nM)

kon × 105 (M-1 s-1)

koff × 10-2 (s-1)

Kd (nM)

kon × 105 (M-1 s-1)

koff × 10-2 (s-1)

mGagV7R

39 ± 9

6±2

2.4 ± 0.7

143 ± 61

3±2

4±2

GagV7RL21K

N/A

N/A

N/A

71 ± 15

5±1

3.2 ± 0.5

GagL21K

46 ± 6

N/A

ND

46 ± 2

5±1

2.2 ± 0.5



Page 28 of 46

The Role of the Gag C-terminus in Membrane Association. While the SHG results have implicated PI(4,5)P2 and the myristoyl group in affecting kinetics of Gag dissociation, no specific mutation or membrane components appear to affect membrane adsorption kinetics. One portion of the protein which could be involved in Gag-membrane association is the C-terminus of the protein, which consists of 10 amino acids with the sequence SLFGNDPSSQ. The C-terminus is part of the p6 domain of Gag, which is involved in recruiting proteins to release the budding viral particles from the plasma membrane.42 In a basic local alignment search, the Gag C-terminus displayed 80% similarity to sequences found in antimicrobial peptides (short chains consisting of 12-50 residues).43-45 Antimicrobial peptides are typically amphipathic and have been shown to associate with cell membranes, influencing a variety of cellular functions, suggesting the C-terminus of Gag may interact with the membrane as well.46-48 In order to examine the membrane interaction of the peptide representing the C-terminus of Gag, its surface density and dissociation constant to bilayers consisting of DOPC, DOPC + 2 mol % PI(4,5)P2, or DOPC + 6 mol % DOPG was determined, Figure 9. The Kd of the peptide to a DOPC bilayer was within error of the dissociation constant of the peptide to a bilayer containing 2 mol % PI(4,5)P2, shown in Table 4. These values were also similar to the dissociation constant of the peptide to DOPC containing 6 mol % DOPG. However, all of the measured dissociation constants for the C-terminus peptide are more than 4-fold lower than the Kd values of the full-length protein to the various membrane compositions, indicating the C-terminus of Gag has a high affinity for membranes, consistent with previous studies that observed lower binding efficiency when only the MA (lacking the C-terminus) was present in comparison to full-length Gag.13, 49 Additionally, the surface density of the peptide to DOPC containing DOPG was 37% lower in comparison to a pure DOPC bilayer. The decrease in surface density could be a result of the calculated isoelectric point (pI) of the peptide being 3.3, meaning it is negatively charged at pH 7.4, which would be repelled by the negatively charged DOPG. Although repulsion of the peptide by PI(4,5)P2 would be expected as well, the surface density of peptide bound to a DOPC + PI(4,5)P2 bilayer was not significantly different than the surface density of peptide bound to pure DOPC, suggesting there are 28 ACS Paragon Plus Environment

Page 29 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


interactions such as hydrogen bonding between the phosphate groups of PI(4,5)P2 and the polar side chains in the peptide that overcomes repulsion. Repulsion by DOPG, but not PI(4,5)P2, was also observed for the full-length Gag protein, suggesting the C-terminus may play a role in facilitating the initial association of Gag to the membrane.

Figure 9. Relative surface density of the Gag peptide as a function of bulk concentration to a DOPC bilayer (circles), a DOPC + 2 mol % PI(4,5)P2 bilayer (diamonds), and a DOPC + 6 mol % DOPG bilayer (triangles). Lines are fits to Eq. 5. Error bars represent standard deviation between individual experiments. The adsorption and desorption kinetics of the C-terminal peptide of Gag were measured to further examine its role in membrane association. Desorption of the peptide from all bilayer compositions, Figure 10, occurred over a much longer timescale (hours rather than minutes) in comparison to the full-length protein, indicating the peptide is intercalating into the membrane. The adsorption rate of the peptide from a DOPC bilayer, listed in Table 4, was comparable to the rate of adsorption of the peptide from DOPC + 2 mol % PI(4,5)P2, indicating PI(4,5)P2 does not drive adsorption. The comparable desorption rates in the presence and absence of PI(4,5)P2 indicate the presence of PI(4,5)P2 is also not prolonging the presence of peptide at the membrane, consistent with PI(4,5)P2 specifically interacting with the HBR, not the Cterminus. While the adsorption rate of the peptide to DOPC + 6 mol % DOPG appears to be faster than the kon for the other bilayers, it was calculated from the koff, which was 3-fold faster than the rates of desorption of the peptide from DOPC or DOPC containing PI(4,5)P2, consistent with repulsion due to the presence of DOPG in the membrane observed in the adsorption isotherms. The low Kd and intercalation 29 ACS Paragon Plus Environment


Page 30 of 46

of the peptide into the membrane suggest that the C-terminus of Gag is involved with initial membrane association. Although study of membrane association of a Gag mutant that has the C-terminus removed would clearly be ideal, it is outside the scope of the current study.

Figure 10. Desorption of the 10-mer Gag peptide from (A) a DOPC bilayer, (B) a DOPC + 2 mol % PI(4,5)P2 bilayer, and (C) a DOPC + 6 mol % DOPG bilayer. Solid and dashed lines are fits to Eq. 6. Table 4. Measured dissociation constants (Kd), adsorption rates (kon), and desorption rates (koff) of the Gag peptide to lipid bilayers of various compositions. Bilayer Composition

Gag C-terminal Peptide Kd (nM)

kon × 104 (M-1s-1)

koff × 10-5 (s-1) 30

ACS Paragon Plus Environment

Page 31 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


DOPC 2% PI(4,5)P2 + DOPC 6% DOPG + DOPC

3.7 ± 0.5

1.2 ± 0.2

4.3 ± 0.1

3.2 ± 0.2

2.5 ± 0.2

7.9 ± 0.1

2.7 ± 0.2

8.9 ± 0.8

24 ± 1

Effect of tRNA on Gag-Membrane Binding. In addition to the composition of the membrane itself, it has been suggested that tRNA present in cells may also modulate the Gag-membrane interaction. A suggested mechanism involves binding of tRNA to Gag, inhibiting membrane association by blocking the interaction with the HBR.14, 50 Since it was previously found that tRNA concentrations of 0.01 to 0.1 mg/mL hindered Gag binding,10 the effect of tRNA on Gag-membrane association was examined with the SHG assay. As a control, there was no binding of the tRNA to either a DOPC bilayer or a DOPC + PI(4,5)P2 bilayer (data not shown). mGag and Gag were incubated with 0.01 mg/mL tRNA prior to exposure to the membrane surface. The incubated proteins were then injected over DOPC and DOPC + 2 mol % PI(4,5)P2 bilayers. Listed in Table 5, the dissociation constants of mGag in the presence of tRNA to a DOPC bilayer and a DOPC containing PI(4,5)P2 bilayer were an order of magnitude higher and 3.6fold higher respectively than the dissociation constants of mGag in the absence of tRNA, suggesting tRNA modulates the interaction of mGag with the membrane, as reported previously.10 The Kd values of Gag in the presence of tRNA and in the presence and absence of PI(4,5)P2 (Table 5) were within error of the dissociation constants of Gag in the absence of tRNA, indicating tRNA only affects the Kd of mGag. As tRNA was shown to bind to the MA of Gag,11 the increase in Kd could be due to the sequestration of the myristoyl group, similar to the effect observed earlier for the V7R mutant. In addition to altering the Kd of mGag for the membrane, the relative surface density of mGag in the presence of tRNA was 26% and 57% lower than in the absence of tRNA for DOPC and DOPC + PI(4,5)P2 bilayers respectively, consistent with tRNA preventing exposure of the myristoyl group.14 However, the surface density of Gag + tRNA bound to the membrane in the presence and absence of PI(4,5)P2 was 42% lower than in the absence of tRNA as well, shown in Figure 11. This suggests tRNA 31 ACS Paragon Plus Environment


Page 32 of 46

also blocks the MA of Gag from interacting with the membrane, regardless of the presence of PI(4,5)P2. SDS-PAGE studies showed that tRNA inhibits binding, however, the presence of 7.25% PI(4,5)P2 was found to restore binding efficiency.10-11 In the present study, 2 mol % PI(4,5)P2 was not sufficient to restore binding after incubation of tRNA with mGag or Gag, suggesting the negatively charged tRNA repels PI(4,5)P2. The decrease in surface density of both mGag and Gag, and increased Kd of only mGag, suggests tRNA inhibits binding of mGag by preventing exposure of the myristoyl group and blocking interaction of the MA with the membrane.

Figure 11. Relative surface density as a function of bulk concentration of (A) mGag + tRNA binding a DOPC + 2 mol % PI(4,5)P2 bilayer (squares) and a DOPC bilayer (circles) and (B) Gag + tRNA binding a DOPC + 2 mol % PI(4,5)P2 bilayer (squares) and a DOPC bilayer (circles). The short light dashed lines represent the binding of mGag in the absence of tRNA binding a DOPC + 2 mol % PI(4,5)P2 bilayer. Solid and long dashed lines are fits to Eq. 5. Error bars represent the standard deviation from triplicate experiments. To confirm whether tRNA inhibits binding by preventing exposure of the myristoyl group through blockage of the PI(4,5)P2 binding site, the adsorption and desorption rates for mGag and Gag incubated with tRNA were measured. In the presence of PI(4,5)P2, the desorption rates for mGag + tRNA and Gag + tRNA, shown in Figures 12A and 12C, were 2.5-fold faster than the koff for Gag in the absence of tRNA and in the presence of PI(4,5)P2, suggesting tRNA affects the binding kinetics of mGag, consistent with 32 ACS Paragon Plus Environment

Page 33 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


tRNA blocking the PI(4,5)P2 binding site and preventing exposure of the myristoyl group. The kon of mGag + tRNA and Gag + tRNA were similar to the rates of adsorption in the absence of tRNA, indicating tRNA only affects dissociation. Furthermore, the desorption rates of both mGag + tRNA and Gag + tRNA in the absence of PI(4,5)P2 (Figures 12B and 12D) were within error of the koff of Gag in the absence of tRNA incubation, confirming tRNA prevents anchoring of Gag by sequestering the myristoyl group.

Figure 12. Desorption data for mGag + tRNA from a (A) DOPC + 2 mol % PI(4,5)P2 membrane and (B) DOPC membrane. Desorption data for Gag + tRNA from a (C) DOPC + 2 mol % PI(4,5)P2 membrane and (D) DOPC membrane. Solid and dashed lines are fits to Eq. 6. Table 5. Measured dissociation constants (Kd), adsorption rates (kon), and desorption rates (koff) of mGag and Gag incubated with tRNA to lipid bilayers of various compositions. mGag + tRNA

Gag + tRNA

Bilayer Composition

Kd (nM)

kon × 105 (M-1 s-1)

koff × 10-2 (s-1)

Kd (nM)

kon × 105 (M-1 s-1)

koff × 10-2 (s-1)

DOPC

250 ± 63

0.4 ± 0.2

1.0 ± 0.3

40 ± 6

3.3 ± 0.9

1.3 ± 0.3



2% PI(4,5)P2 + DOPC

125 ± 31

1.6 ± 0.8

2.0 ± 0.9

45 ± 12

Page 34 of 46

4±3

2±1

Gag-Membrane Association Mechanism. The current study provides a comprehensive mechanism, depicted in Figure 13, of the membrane association of Gag that builds upon previous SDSPAGE and NMR investigations. The initial adsorption of the protein to the membrane was suggested to involve the C-terminus of Gag. Association of the MA with the bilayer follows, allowing the specific interaction between the HBR and PI(4,5)P2, which slows down dissociation of Gag from the membrane. In addition, the binding of PI(4,5)P2 allows the insertion of the myristoyl group into the bilayer, further anchoring Gag to the membrane. This mechanism is supported by the rates of desorption measured for the V7R, V7R-L21K, and L21K mutations in the MA which increase the rate of desorption by sequestering the myristoyl group and preventing interaction of the HBR with PI(4,5)P2. In line with this mechanism is the observation that tRNA also inhibits membrane association by binding to the HBR of Gag and preventing exposure of the myristoyl group, resulting in faster dissociation and ultimately a lower surface density of Gag.


Page 35 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 13. Cartoon of the initial Gag-membrane association process (not to scale): (A) In solution, the MA domain of Gag with the myristoyl group (black coil) is folded over towards the C-terminus (black line). (B) The C-terminus intercalates into the bilayer, followed by (C) adsorption of the MA domain of Gag. (D) Specific interaction between PI(4,5)P2 (light gray circle) and the HBR in the MA domain, along with the insertion of the exposed myristoyl group prolongs the presence of Gag at the membrane. (E) Accumulation of Gag at the membrane surface and desorption of the C-terminus results in (F) the linear conformation of Gag.

Conclusions The binding mechanism of HIV-1 Gag to lipid membranes has been investigated by measuring the thermodynamics and kinetics of association for myristoylated and nonmyristoylated Gag, along with various point mutations in the matrix domain of Gag to bilayers in the presence and absence of PI(4,5)P2. 35 ACS Paragon Plus Environment


Page 36 of 46

A workflow was developed for producing soluble, full-length and myristoylated HIV-1 Gag proteins that recapitulate known ssDNA-binding properties. This innovation allowed us to investigate membrane interactions using the most physiologically-relevant purified Gag protein to date. Both Gag protein myristoylation and PI(4,5)P2 in membranes accelerated the Gag-membrane association by slowing the rate of desorption, rather than decreasing the Kd. The myristoyl group does not drive Gag-membrane association, but rather plays a role in anchoring Gag to the membrane, by decreasing the desorption rate. Although PI(4,5)P2 and the myristoyl group influenced dissociation, the C-terminus of Gag was found to facilitate initial membrane association. The SHG data suggest that a specific Gag interaction with PI(4,5)P2 allows Gag to overcome electrostatic repulsion observed when the same membrane surface charge is achieved using DOPG, indicating electrostatics are not the driving force for membrane association. The L21K mutation increases the rate of desorption by blocking this specific interaction with PI(4,5)P2. The V7R mutation also results in an increased rate of desorption, as the sequestered myristoyl group cannot anchor Gag to the membrane. The packing of the lipids in the membrane was also shown to play a role, as the tightly packed gel phase DSPC bilayer hindered Gag binding in comparison to the fluid phase DOPC bilayer. The incubation with soluble Ins(1,4,5)P3 essentially replicated binding of mGag to membranes lacking PI(4,5)P2, indicating that pre-exposure of Gag to the headgroup of PI(4,5)P2 does not drive membrane association. Consistent with previous SDS-PAGE studies, tRNA was observed to inhibit Gag-membrane association. However, the presence of PI(4,5)P2 in the bilayer did not restore binding, indicating the interaction of tRNA with Gag blocks specific interaction of the MA with the membrane and also blocks exposure of the myristoyl group. The binding behavior of Gag under all of these conditions indicates that it samples and adsorbs to membranes, before accumulating on membranes which allow insertion of the myristoyl group and which contain PI(4,5)P2 due to the slow desorption rate. The present study has provided a quantitative understanding of the roles of lipid packing, the presence of PI(4,5)P2 and the Gag myristoyl group, the C-terminus of Gag, and tRNA in Gag-membrane interactions. From the investigation of these roles, the dissociation constant of Gag was quantified, adding more detail to the 36 ACS Paragon Plus Environment

Page 37 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


membrane association process, and has helped elucidate the many mechanisms by which Gag associates with the plasma membrane.

Acknowledgments: The authors acknowledge the financial support from the National Science Foundation (#1402901). Any opinions, findings, conclusions, or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. The authors would like to thank Prof. Wes Sundquist, Department of Biochemistry, University of Utah School of Medicine for hosting Dr. Matthew S. Lalonde. The authors also acknowledge and thank Jeremy Luban and Cagan Gurer for providing the codon-optimized Gag ORF, Steve Alam for cloning the hNMT2 and wild-type HIV-1 Gag protein expression vectors, and Will Kincannon and the Bandarian Group for providing the C-terminus Gag peptide. Mass spectrometry analysis was performed at the Mass Spectrometry and Proteomics Core Facility at the University of Utah on instruments obtained through Shared Instrumentation Grant 1S10OD018210-01A1. The authors declare there are no conflicts of interest.

Author Contributions: R.J.T. performed, analyzed, and designed SHG experiments. M.S.L. performed and designed protein purification, derivation, and characterization. K.L.S. designed SHG experiments. R.J.T., M.S.L., and J.C.C. wrote the manuscript.

Supporting Information. List of silent mutations for the HIV-1 Gag coding sequence.

References 1.

Chukkapalli, V.; Hogue, I. B.; Boyko, V.; Hu, W. S.; Ono, A., Interaction between the

Human Immunodeficiency Virus Type 1 Gag Matrix Domain and Phosphatidylinositol-(4,5)37 ACS Paragon Plus Environment


Page 38 of 46

Bisphosphate Is Essential for Efficient Gag Membrane Binding. Journal of Virology 2008, 82 (5), 2405-2417. 2.

Modrow, S.; Kattenbeck, B.; von Poblotzki, A.; Niedrig, M.; Wagner, R.; Wolf, H., The

Gag Proteins of Human Immunodeficiency Virus Type 1: Mechanisms of Virus Assembly and Possibilities for Interference. Med. Microbiol. Immunol. 1994, 183, 177-194. 3.

Gui, D.; Gupta, S.; Xu, J.; Zandi, R.; Gill, S.; Huang, I. C.; Rao, A. L. N.; Mohideen, U., A

Novel Minimal in Vitro System for Analyzing Hiv-1 Gag-Mediated Budding. J. Biol. Phys. 2015,

41, 135-149. 4.

Ferrell Jr, J. E.; Huestis, W. H., Phosphoinositide Metabolism and the Morphology of

Human Erythrocytes. Journal of Cell Biology 1984, 98, 1992-1998. 5.

Saad, J. S.; Miller, J.; Tai, J.; Kim, A.; Ghanam, R. H.; Summers, M. F., Structural Basis

for Targeting Hiv-1 Gag Proteins to the Plasma Membrane for Virus Assembly. Proc. Natl. Acad.

Sci. 2006, 103 (30), 11364-11369. 6.

Mercredi, P. Y.; Bucca, N.; Loeliger, B.; Gaines, C. R.; Mehta, M.; Bhargava, P.; Tedbury,

P. R.; Charlier, L.; Floquet, N.; Muriaux, D., et al., Structural and Molecular Determinants of Membrane Binding by the Hiv-1 Matrix Protein. J Mol Biol 2016, 428 (8), 1637-1655.


Page 39 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

7.


Olety, B.; Ono, A., Roles Played by Acidic Lipids in Hiv-1 Gag Membrane Binding. Virus

Research 2014, 193, 108-115. 8.

Dalton, A. K.; Ako-Adjei, D.; Murray, P. S.; Murray, D.; Vogt, V. M., Electrostatic

Interactions Drive Membrane Association of the Human Immunodeficiency Virus Type 1 Gag Ma Domain. J Virol 2007, 81 (12), 6434-6445. 9.

Charlier, L.; Louet, M.; Chaloin, L.; Fuchs, P.; Martinez, J.; Muriaux, D.; Favard, C.;

Floquet, N., Coarse-Grained Simulations of the Hiv-1 Matrix Protein Anchoring: Revisiting Its Assembly on Membrane Domains. Biophys J 2014, 106 (3), 577-585. 10.

Chukkapalli, V.; Inlora, J.; Todd, G. C.; Ono, A., Evidence in Support of Rna-Mediated

Inhibition of Phosphatidylserine-Dependent Hiv-1 Gag Membrane Binding in Cells. Journal of

Virology 2013, 87, 7155-7159. 11.

Chukkapalli, V.; Oh, S. J.; Ono, A., Opposing Mechanisms Involving Rna and Lipids

Regulate Hiv-1 Gag Membrane Binding through the Highly Basic Region of the Matrix Domain.

Proc. Natl. Acad. Sci. 2010, 107 (4), 1600-1605. 12.

Barros, M.; Heinrich, F.; Datta, S. A. K.; Rein, A.; Karageorgos, I.; Nanda, H.; Losche, M.,

Membrane Binding of Hiv-1 Matrix Protein: Dependence on Bilayer Composition and Protein Lipidation. J Virol 2016, 90 (9), 4544-4555. 39 ACS Paragon Plus Environment


13.

Page 40 of 46

Yandrapalli, N.; Lubart, Q.; Tanwar, H. S.; Picart, C.; Mak, J.; Muriaux, D.; Favard, C.,

Self Assembly of Hiv-1 Gag Protein on Lipid Membranes Generates Pi(4,5)P2/Cholesterol Nanoclusters. Sci Rep 2016, 6, 1-13. 14.

Gaines, C. R.; Tkacik, E.; Rivera-Oven, A.; Somani, P.; Achimovich, A.; Alabi, T.; Zhu,

A.; Getachew, N.; Yang, A. L.; McDonough, M., et al., Hiv-1 Matrix Protein Interactions with Trna: Implications for Membrane Targeting. J Mol Biol 2018, 430 (14), 2113-2127. 15.

Paillart, J.-C.; Gottlinger, H. G., Opposing Effects of Human Immunodeficiency Virus

Type 1 Matrix Mutations Support a Myristyl Switch Model of Gag Membrane Targeting. Journal

of Virology 1999, 73 (4), 2604-2612. 16.

Saad, J. S.; Loeliger, E.; Luncsford, P.; Liriano, M.; Tai, J.; Kim, A.; Miller, J.; Joshi, A.;

Freed, E. O.; Summers, M. F., Point Mutations in the Hiv-1 Matrix Protein Turn Off the Myristyl Switch. J. Mol. Biol. 2007, 366, 574-585. 17.

Ono, A.; Freed, E. O., Binding of Human Immunodeficiency Virus Type 1 Gag to

Membrane: Role of the Matrix Amino Terminus. Journal of Virology 1999, 73 (5), 4136-4144. 18.

Dick, R. A.; Goh, S. L.; Feigenson, G. W.; Vogt, V. M., Hiv-1 Gag Protein Can Sense the

Cholesterol and Acyl Chain Environment in Model Membranes. Proc. Natl. Acad. Sci. 2012, 109 (46), 18761-18766. 40 ACS Paragon Plus Environment

Page 41 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


19.

Shen, Y. R., The Principles of Nonlinear Optics. Wiley: New York, 1984.

20.

Salafsky, J. S.; Eisenthal, K. B., Protein Adsorption at Interfaces Detected by Second

Harmonic Generation. J. Phys. Chem. B 2000, 104, 7752-7755. 21.

Nguyen, T. T.; Sly, K. L.; Conboy, J. C., Comparison of the Energetics of Avidin,

Streptavidin, Neutravidin, and Anti-Biotin Antibody Binding to Biotinylated Lipid Bilayer Examined by Second-Harmonic Generation. Analytical Chemistry 2012, 84 (1), 201-208. 22.

Stokes, G. Y.; Conboy, J. C., Measuring Selective Estrogen Receptor Modulator (Serm)–

Membrane Interactions with Second Harmonic Generation. Journal of the American Chemical

Society 2014, 136 (4), 1409-1417. 23.

Sun, Y. S.; Landry, J. P.; Fei, Y. Y.; Zhu, X. D.; Luo, J. T.; Wang, X. B.; Lam, K. S., Effect

of Fluorescently Labeling Protein Probes on Kinetics of Protein-Ligand Reactions. Langmuir 2008, 24 (23), 13399-13405. 24.

Song, X.; Swanson, B. I., Direct, Ultrasensitive, and Selective Optical Detection of Protein

Toxins Using Multivalent Interactions. Anal. Chem. 1999, 71 (11), 2097-2107. 25.

Chen, J.-R.; Liao, C.-W.; Mao, S. J. T.; Chen, T.-H.; Weng, C.-N., A Simple Technique

for the Simultaneous Determination of Molecular Weight and Activity of Superoxide Dismutase Using Sds-Page. J. Biochem. Biophys. Methods 2001, 47, 233-237. 41 ACS Paragon Plus Environment


26.

Page 42 of 46

Liu, J.; Conboy, J. C., Structure of a Gel Phase Lipid Bilayer Prepared by the Langmuir-

Blodgett/Langmuir-Schaefer

Method

Characterized

by

Sum-Frequency

Vibrational

Spectroscopy. Langmuir 2005, 21, 9091-9097. 27.

Havel, H. A.; Chao, R. S.; Haskell, R. J.; Thamann, T. J., Investigations of Protein

Structure with Optical Spectroscopy: Bovine Growth Hormone. Anal. Chem. 1989, 61, 642-650. 28.

Kriech, M. A.; Conboy, J. C., Counterpropagating Second-Harmonic Generation: A New

Technique for the Investigation of Molecular Chirality at Surfaces. J. Opt. Soc. Am. B 2004, 21 (5), 1013-1022. 29.

Studier, F. W., Protein Production by Auto-Induction in High-Density Shaking Cultures.

Protein Expression and Purification 2005, 41 (1), 207-234. 30.

Gurer, C.; Berthoux, L.; Luban, J., Covalent Modification of Human Immunodeficiency

Virus Type 1 P6 by Sumo-1. J Virol 2005, 79 (2), 910-917. 31.

Berkowitz, R. D.; Luban, J.; Goff, S. P., Specific Binding of Human Immunodeficiency

Virus Type 1 Gag Polyprotein and Nucleocapsid Protein to Viral Rnas Detected by Rna Mobility Shift Assays. Journal of Virology 1993, 67 (12), 7190-7200.


Page 43 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

32.


Rice, W. G.; Supko, J. G.; Malspeis, L.; Buckheit Jr., R. W.; Clanton, D.; Bu, M.; Graham,

L.; Schaeffer, C. A.; Turpin, J. A.; Domagala, J., et al., Inhibitors of Hiv Nucleocapsid Protein Zinc Fingers as Candidates for the Treatment of Aids. Science 1995, 270 (5239), 1194-1197. 33.

Fisher, R. J.; Rein, A.; Fivash, M. J.; Urbaneja, M. A.; Casas-Finet, J. R.; Medaglia, M.

V.; Henderson, L. E., Sequence-Specific Binding of Human Immunodeficiency Virus Type 1 Nucleocapsid Protein to Short Oligonucleotides. Journal of Virology 1998, 72 (3), 1902-1909. 34.

Fisher, R. J.; Fivash, M. J.; Stephen, A. G.; Hagan, N. A.; Shenoy, S. R.; Medaglia, M.

V.; Smith, L. R.; Worthy, K. M.; Simpson, J. T.; Shoemaker, R., et al., Complex Interactions of Hiv-1 Nucleocapsid Protein with Oligonucleotides. Nucleic Acids Res 2006, 34 (2), 472-484. 35.

Marsh, D., Handbook of Lipid Bilayers. CRC Press: Boca Raton, 1990.

36.

Marsh, D., Lateral Pressure in Membranes. Biochimica et Biophysica Acta (BBA) 1996,

1286, 183-223. 37.

Keller, H.; Krausslich, H. G.; Schwille, P., Multimerizable Hiv Gag Derivative Binds to the

Liquid-Disordered Phase in Model Membranes. Cell Microbiol 2013, 15 (2), 237-247. 38.

Ono, A.; Ablan, S. D.; Lockett, S. J.; Nagashima, K.; Freed, E. O., Phosphatidylinositol

(4,5) Bisphosphate Regulates Hiv-1 Gag Targeting to the Plasma Membrane. Proc. Natl. Acad.

Sci. 2004, 101 (4), 14889-14894. 43 ACS Paragon Plus Environment


39.

Page 44 of 46

Kooijman, E. E.; King, K. E.; Gangoda, M.; Gericke, A., Ionization Properties of

Phosphatidylinositol Polyphosphates in Mixed Model Membranes. Biochemistry 2009, 48 (40), 9360-9371. 40.

Sanden, T.; Salomonsson, L.; Brzezinski, P.; Widengren, J., Surface-Coupled Proton

Exchange of a Membrane-Bound Proton Acceptor. Proc. Natl. Acad. Sci. 2010, 107 (9), 41294134. 41.

Tang, C.; Loeliger, E.; Luncsford, P.; Kinde, I.; Beckett, D.; Summers, M. F., Entropic

Switch Regulates Myristate Exposure in the Hiv-1 Matrix Protein. Proc. Natl. Acad. Sci. 2003,

101 (2), 517-522. 42.

Adamson, C. S.; Jones, I. M., The Molecular Basis of Hiv Capsid Assembly--Five Years

of Progress. Rev Med Virol 2004, 14 (2), 107-121. 43.

Bussolati, B.; Grange, C.; Tei, L.; Deregibus, M. C.; Ercolani, M.; Aime, S.; Camussi, G.,

Targeting of Human Renal Tumor-Derived Endothelial Cells with Peptides Obtained by Phage Display. J Mol Med 2007, 85, 897-906. 44.

Verdon, J.; Falge, M.; Maier, E.; Bruhn, H.; Steinert, M.; Faber, C.; Benz, R.; Hechard,

Y., Detergent-Like Activity and Alpha-Helical Structure of Warnericin Rk, an Anti-Legionella Peptide. Biophys J 2009, 97 (7), 1933-1940. 44 ACS Paragon Plus Environment

Page 45 of 46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

45.


Yang, X.; Lee, W. H.; Zhang, Y., Extremely Abundant Antimicrobial Peptides Existed in

the Skins of Nine Kinds of Chinese Odorous Frogs. J Proteome Res 2012, 11 (1), 306-319. 46.

Sato, H.; Feix, J. B., Peptide–Membrane Interactions and Mechanisms of Membrane

Destruction by Amphipathic Α-Helical Antimicrobial Peptides. Biochimica et Biophysica Acta

(BBA) - Biomembranes 2006, 1758 (9), 1245-1256. 47.

Ganz, T., The Role of Antimicrobial Peptides in Innate Immunity. Integr. Comp. Biol.

2003, 43, 300-304. 48.

Epand, R. M.; Vogel, H. J., Diversity of Antimicrobial Peptides and Their Mechanisms of

Action. Biochimica et Biophysica Acta (BBA) - Biomembranes 1999, 1462, 11-28. 49.

Chukkapalli, V.; Ono, A., Molecular Determinants That Regulate Plasma Membrane

Association of Hiv-1 Gag. Journal of Molecular Biology 2011, 410 (4), 512-524. 50.

Olety, B.; Veatch, S. L.; Ono, A.; Sundquist, W. I., Phosphatidylinositol-(4,5)-

Bisphosphate Acyl Chains Differentiate Membrane Binding of Hiv-1 Gag from That of the Phospholipase Cδ1 Pleckstrin Homology Domain. Journal of Virology 2015, 89 (15), 7861-7873.



TOC Graphic


Page 46 of 46

A Mechanistic Investigation of HIV-1 Gag Association with Lipid

Recommend Documents