Investigation of HIV-1 Gag binding with RNAs and Lipids using Atomic Force Microscopy
Shaolong Chen, Jun Xu, Mingyue Liu, A.L.N. Rao, Roya Zandi, Sarjeet S. Gill, Umar Mohideen
11 Investigation of HIV-1 Gag binding with RNAs and Lipidsusing Atomic Force MicroscopyShaolong Chen , Jun Xu , Mingyue Liu , A.L.N. Rao , Roya Zandi , Sarjeet S. Gill , Umar Mohideen * Department of Physics & Astronomy, University of California, Riverside, California, USA Department of Plant Pathology & Microbiology, University of California, Riverside, California, USA Department of Cell Biology & Neuroscience, University of California, Riverside, California, USA*For correspondence: [email protected] Force Microscopy was utilized to study the morphology of Gag, ΨRNA, and their bindingcomplexes with lipids in a solution environment with 0.1 Å vertical and 1nm lateral resolution. TARpolyARNA was used as a RNA control. The lipid used was phospha-tidylinositol-(4,5)-bisphosphate (PI(4,5)P2).The morphology of specific complexes Gag-ΨRNA, Gag-TARpolyA RNA, Gag-PI(4,5)P2 andPI(4,5)P2-ΨRNA-Gag were studied. They were imaged on either positively or negatively charged micasubstrates depending on the net charges carried. Gag and its complexes consist of monomers, dimers andtetramers, which was confirmed by gel electrophoresis. The addition of specific ΨRNA to Gag is found toincrease Gag multimerization. Non-specific TARpolyA RNA was found not to lead to an increase in Gagmultimerization. The addition PI(4,5)P2 to Gag increases Gag multimerization, but to a lesser extent thanΨRNA. When both ΨRNA and PI(4,5)P2 are present Gag undergoes comformational changes and an evenhigher degree of multimerization.Keywords: AFM, HIV-1 Gag, ΨRNA, TARpolyA RNA, PI(4,5)P2, Gag multimerization NTD ) and the C-terminal domain (CA
CTD ) connected bya flexible linker. The arrowhead-like shaped CA
NTD containing seven alpha helices is essential for theformation of a conical outer shell of the capsid core [16, 17]. The CA
NTD forms hexameric rings with anapproximate spacing of around 8 nm as observed with cryo-electron microscopy (cET) [18, 19]. The CA
NTD also forms pentameric rings, which are not necessary for the formation of immature virions but areindispensable for the formation of the closed shell of the core that adopts a fullerene cone structure in themature virus. The CA domain plays a vital role in the formation of both immature and mature virions. Inthe mature virus, the mature capsid core consists of 1000-1500 copies of the CA protein assembled into a hexameric lattice with a spacing of 10 nm rather than 8nm which is the spacing of CA hexamers in theimmature virions [3]. Therefore the CA hexamers must be rearranged during the final stage of maturation.Although the NC domain is the smallest component of HIV Gag compared to MA and CA domains, it iscritical for the genomic viral RNA recognition, interaction and dimerization. The NC domain tethers Gag tothe RNA genome through nonspecific interaction as well as specific binding to the stem-loop 3 (SL3) in thepackaging signal Ψ (ΨRNA) [20]. In the inner core of the mature virus, the viral genomic RNA is wrappedaround 1500-2000 copies of NC proteins [21]. Within the NC domain, there are two CCHC type zincfingers, which are crucial for specific ΨRNA binding and genomic viral RNA packaging [22, 23]. Thespacing distance between two zinc fingers is highly conserved which is essential for NC functions [24].HIV-1 genome is principally a sequence of RNA that consists of 9173 nucleotides (~9kb) [9, 25]. It isinvolved in many activities throughout the virus replication cycle, such as expressing transcription,facilitating genomic dimerization and transportation, signaling polyadenylation, mediating HIV genomepackaging, initiating reverse transcription etc. HIV-1 genome is always packed into the virion as an RNAdimer, i.e. two copies of full length ~9kb RNA, which is 5’ capped and 3’ polyadenylated. At the 5’untranslated region (UTR) of HIV-1 genomic RNA, there are a variety of critical regions which areconsidered necessary for genome dimerization and binding with Gag: the transactivation responsestem-loop (TAR), the polyadenylation stem-loop (polyA), the prime binding site (PBS) and the packagingsignal domain Ψ. Of the 104 nucleotides of HIV viral RNA, the first 1-57 nucleotides is TAR with 58-104nucleotides being polyA [26, 27]. The mass of the TARpolyA RNA is around 34kDa. The TARpolyA RNAsequence used in the experiments is shown in Fig.1(a) [22]. It plays an essential role in HIV genomepackaging and reverse transcription [26]. Small-angle X-ray scattering (SAXS) studies have shown TARand polyA hairpins extend into a stable coaxially stacked helices [26]. In addition to the dimerizationinitiation site (DIS) in Ψ, the TAR may also facilitate HIV genomic RNA dimerization only when the NCprotein is present. This is because the exposure of the hidden palindromic sequence in the TAR hairpinrequires the NC protein to form TAR-TAR dimers [28]. The packaging signal Ψ contains approximately109 nucleotides [26, 29]. The mass of the ΨRNA is about 36kDa. The ΨRNA sequence used in our work isshown in Fig.1(b) where most of nucleotides are paired with each other [26]. It contains 3 stem loops SL1,SL2, and SL3. SAXS studies have revealed ΨRNA adopts an unfolded conformation where all stem loopsare open for later interaction with both viral and host elements. The ΨRNA binds with the NC protein andis crucial for packaging of HIV genomic RNA into the virus. Within the ΨRNA, SL1 contains apalindromic sequence DIS that is responsible for HIV genomic RNA dimerization and Gag binding at theearly stages of HIV virus replication cycle [30-32]. SL2 includes the splice donor site (SD) that is used toproduce all subsequent spliced message RNAs (mRNA) for production and translation of viral accessoryproteins. SL3 is required for both viral RNA dimerization and packaging [33]. In addition to SL1, SL3 isalso an independent high affinity Gag binding site [34].PI(4,5)P2 belongs to the negatively charged lipid family known as phosphoinositide [35]. It consists ofa glycerol backbone with one saturated fatty acid chain at position 1’ and one unsaturated fatty acid chain atposition 2’ and a phosphoinositol headgroup at position 3’. PI(4,5)P2 is an important component of the hostcell membrane and is enriched at the cytoplasmic leaflet of the plasma membrane. It also serves as a raft forHIV-1 Gag targeting to the plasma membrane and regulates the HIV assembly [35, 36]. More specifically,
PI(4,5)P2 directly binds with the myristylated MA domain of Gag through a highly specific interaction.First, the phosphoinositol headgroup and the 2’ unsaturated fatty acid chain of PI(4,5)P2 insert into ahydrophobic pocket in the MA domain. This then triggers the exposure of the myristylated group forinsertion into the lipid bilayer membrane [37-40]. The 1’ saturated fatty acid chain serves as an anchor forGag targeting to plasma membrane although it does not interact with the MA domain. Therefore, there aretwo saturated fatty acid chains anchoring Gag-membrane binding, and one is the myristylated group fromthe MA domain and the other is the 1’ saturated fatty acid chain from PI(4,5)P2 [1].Previous research has suggested that Gag has distinct binding modes with different RNAs [22]. Fornon-specific TARpolyA RNA, both the NC and MA domains are bound to TARpolyA RNA. While in thecase of specific ΨRNA, only the NC domain was found to bind with ΨRNA [22]. In the latter case, MAdomain is left free for later interaction with lipid PI(4,5)P2. The motivation of this work is to utilize ahigh-resolution technique such as Atomic Force Microscopy (AFM) to explore HIV-1 Gag binding withdifferent RNAs and lipids. First, we studied the morphology of individual components of Gag, ΨRNA, andTARpolyA RNA separately as controls. Next, to see the effect of the addition of specific and non-specificRNAs, we investigated the effect on the complex formation by adding either ΨRNA or TARpolyA RNA toGag. Then, the effect of adding lipid PI(4,5)P2 into Gag was also studied. Finally, the effect of the additionof both ΨRNA and PI(4,5)P2 was examined to understand their collective effect on Gag.2 Materials and methods2.1 MaterialsHIV-1 Gag used in our experiments was obtained from Dr. Alan Rein and Dr. S.A.K. Datta. It lacks themyristylated group and p6 and thus usually is referred to as GagΔP6 [22]. Both ΨRNA and TARpolyARNA were obtained from Dr. Karin Musier-Forsyth and Dr. E.D. Olson [36]. Brain PI(4,5)P2(L-α-phosphatidylinositol-4,5-bisphosphate) was purchased from Avanti Polar Lipids (Alabaster, AL, US).2.2 Preparation of samplesHIV-1 GagΔP6 was originally at 30μM in the buffer containing 20 mM Tris-HCl (pH 7.5), 0.5 M NaCl,10 % (v/v) glycerol, 5 mM DTT (dithiothreitol). It was diluted to 0.5μM before AFM imaging with HEPESbuffer, which contained 20mM, HEPES (pH 7.5), 1mM MgCl , 50mM NaCl, 10μM TCEP(tris-2-carboxy-ethyl phosphine) 5 mM βME (β-mercaptoethanol). ΨRNA and TARpolyA RNA wereoriginally in the same HEPES buffer at a concentration of 74.18μM and 119μM, respectively. Both RNAsneed to be refolded before use. The protocol followed for refolding RNA was as follows. First, 22.2μL74.18μM ΨRNA (or 13.8uL 119μM TARpolyA RNA) was added to a clean vial. Next, 2.5μL 1M(PH 7.5)HEPES was added into the vial. Then, 20.3μL DEPC-H O (or 28.7μL for TARpolyA RNA) was added. Thetemperature of the mixture was then raised to 80 C for 2 minutes, followed by 60 C for another 2 minutesusing a water bath. Finally, 5μL 0.1M MgCl was added into the vial. Next, the mixture was kept at 37 Cfor 5 minutes followed by 0 C with ice for 30 minutes. The RNA could then be used immediately or storedat 4 C for later use up to a week. After applying the refolding protocol, 30μM RNAs were diluted to 0.5μMusing the same HEPES buffer before AFM imaging. Brain PI(4,5)P2 purchased in powder form was dissolved in distilled water to 1 mM concentration before use. For mixtures, the mixed solutions wereobtained such that the final concentration of each component was 0.5μM.2.3 Mica SubstratesAtomically smooth mica substrates were used in all AFM imaging experiments reported here. Micasubstrates with 1cm diameter were obtained from Ted Pella Inc. (Redding, CA, USA). The scotch tapetechnique was used to obtain freshly cleaved surfaces used in all experiments. The surface roughness wasmeasured to be 0.1~0.2nm. The clean mica surface is negatively charged with charge density σ = -0.33C/m in air and σ = -2.5mC/m in water [41-43]. Therefore, freshly cleaved raw mica is a perfect substrate forAFM imaging of HIV GagΔP6 because of its positive charge. However, as both ΨRNA and TARpolyARNA are negatively charged they cannot be observed directly on the raw mica substrate. To overcome therepulsion between mica and RNAs, the mica surface was made positive by functionalization with APTES(3-aminopropyltriethoxy silane). The procedure of preparing APTES-treated mica was as follows. First,double-sided tape was used to stick a 10mm in diameter mica upon an AFM metal specimen disc with adiameter of 15mm. Second, Scotch tape was used to cleave mica until the mica surface was complete andflat. Then, 100μL APTES was added into a small plastic petri dish and put at the bottom of a desiccatorwith a plastic net onto which the freshly cleaved mica was placed. The dessicator was next evacuated witha mechanical vacuum pump. The vacuum suction was maintained for 30 minutes to allow APTES toevaporate. APTES treated mica was ready to use immediately or can be stored in a covered petri dish forlater use. 30~50μL of the desired sample solution was next deposited on the freshly cleaved mica (orAPTES treated mica for RNAs) before AFM imaging .2.4 AFMCalibration of the AFM probes had to be done before imaging. The size of the AFM cantilever tip wassimilar to that of the proteins or protein complexes studied. The comparable tip size will lead to featurebroadening, which is a common type of widely known convolution effect [44]. The major factors withrespect to the feature broadening are the pyramidal geometry and curvature radius of the tip. The AFMprobe used (HI’RES-C19/CR-AU, MikroMasch USA, Watsonville, CA, USA) was 125μm long and22.5μm wide with a spring constant of 0.5N/m. it had a nominal resonant frequency of 65kHz in air and a~32.36kHz in liquid. The special tips used had high aspect ratio and small tip radius. Nevertheless, the tipsize still had to be taken into account when analyzing the sizes of HIV GagΔP6, RNAs, lipids and theirmixed complexes. The calibration was done as follows. 2nm gold spheres were used to calibrate AFMprobes due to the comparability of the heights of HIV GagΔP6 and two RNAs. Because the measuredsample size also depends on the height of the sample, the actual tip size is given by equation (1) (seeSupplementary Material for more details): λ = L – 1.46D (1)Where λ is the actual diameter of the AFM cantilever tip, L is the measured size of the sample, and D is theheight of the sample. The effective tip diameter t for any sample is the difference between measured size ofthe calibration standard sample and the height of the sample, as given by equation (2) (see Supplementary Material for more details): t = L – D = λ + 0.46D (2)For the 2nm gold spheres, after fitting to a normal distribution, the mean measured size
L = 7.56 ± 0.09nm ,the mean height
D = 2.10 ± 0.02nm which is the actual diameter of the 2nm gold particle, as shown in Fig.2.The total number of samples was 419 and the experiment was repeated twice. Therefore, the actualdiameter is λ = 4.5nm according to equation (1). The effective tip size for HIV GagΔP6 and other GagΔP6complexes, including GagΔP6-ΨRNA, GagΔP6-PI(4,5)P2, PI(4,5)P2-ΨRNA-GagΔP6, is t Gag = 5.4nm given that the measured height of HIV GagΔP6 is 1.9nm according to equation (2). Similarly, the effectivetip size for ΨRNA and TARpolyA RNA is t RNA = 5.0nm given that the measured height of both RNAs is1.1nm.The AFM Nanoscope IIIa (Veeco Metrology, Santa Barbara, CA, USA) was used in tapping mode .The procedure of the AFM operation in tapping mode in liquid environment is as follows [45]. First, afreshly cleaved mica (or APTES treated mica for RNAs) was mounted on the AFM metal specimen holderusing double-sided tape. Next, a 30~50μL drop of following sample solutions used in the experiment wasdeposited on the mica: (I) ΨRNA (0.5μM), (II) TARpolyA RNA (0.5μM), (III) GagΔP6 (0.5μM), (IV)mixture of PI(4,5)P2-DPhPC-POPC (0.5μM : 5μM : 5μM) complex, (V) mixture of GagΔP6-ΨRNA(0.5μM : 0.5μM) complex, (VI) mixture of GagΔP6-TARpolyA RNA (0.5μM : 0.5μM) complex, (VII)mixture of GagΔP6-PI(4,5)P2 (0.5μM : 0.5μM) complex, (VIII) and mixture of PI(4,5)P2-ΨRNA-GagΔP6(0.5μM : 0.5μM : 0.5μM) complex. Next the cantilever probe was mounted into the fluid cell. Care wastaken to make sure there are no bubbles. The tip is completely immersed in the solution. This is critical forlaser alignment when operating the AFM in a liquid environment. Next the laser signal was aligned andthen the piezo was oscillated through a range of frequencies till the resonance frequency was found. Nextthe best resolution was obtained by adjusting the following scanning parameters: the vertical range,samples/line, scan size, scan rate, integral gain, proportional gain and amplitude setpoint etc. Additionalchecks were made after engaging the cantilever. During imaging the amplitude setpoint was adjusted suchthat the trace and retrace lines were matched. The force exerted on the sample should be as small aspossible to prevent samples from being damaged. All the aforementioned parameters had to be adjustedcollectively to achieve the best resolution.3 Results and DiscussionAs RNAs have been well studied with the AFM [46-50], they were used to benchmark the experimentsreported here. The self-assembly of the CA domains in Gag and its mechanical properties have beenrecently investigated on a mica substrate with the AFM [51]. The CA domain was found to assemble in a hexagonal lattice and have self repair capacity after damage was induced [51]. Inthe AFM experiments here, we started with size and morphology measurements of the RNAs (ΨRNA andTARpolyA RNA) to benchmark and validate the measurement and analysis software developed.Independent calibration of the measurement resolution was done using 2nm Au spheres as discussed above.After confirmation of the validity of the technique we performed measurements on GagΔP6 and itscomplexes with RNAs and PI(4,5)P2 lipid. The summary of the results is provided in Table 1. All theexperiments were repeated twice. Below we present the results from each individual experiment and alsodiscuss the effect of the RNA and PI(4,5)P2 lipid interaction with GagΔP6.3.1 ΨRNA Size and Morphology MeasurementsΨRNA (0.5μM) being negatively charged was imaged on positively charged APTES treated mica.Another motivation for measuring the ΨRNA by itself is to understand its individual morphology for futurecomparison with that observed in the various GagΔP6 complexes. A typical AFM image of ΨRNA wasshown in Fig.3(a). In Fig.3(a), the AFM image of ΨRNA shows that most of ΨRNA molecules seem tohave inverted “L” shape. As shown in Fig.3 (b) and in Table 1, the mean height is 1.10 ± 0.01nm. Thisheight is in between the values 0.5nm, 2.5, 2.6nm found from double-stranded RNA [28,48,52] and similarto 0.9~1.2nm that found by Hansma et al. [53]. The lateral size of the image was analyzed using specially developed software analysis (see Supplementary Material for more details). The statistics of the length(longest dimension) and width (longest perpendicular dimension to the length) and height were plotted inFig. 3(b). As can be observed there are two distinct peaks for the length and similarly two distinct peaks forthe width. The three dimensional smooth histogram of the same length and width population distribution isshown in Fig.3(c). Two distinct populations can be observed. The size of the first distribution with a meanlength of 17.9 ± 0.2nm and width 1.01 ± 0.02nm and height of 1.10 ± 0.01nm corresponds to the ΨRNAmonomer. The second peak with a mean length of 34.6 ± 0.3nm and width 3.8 ± 0.1nm and height of 1.10 ±0.01nm corresponds to the ΨRNA dimer. Typical images of the monomer and dimers are enclosed in redand green boxes respectively.The expected size of the 109-nucleotide ΨRNA monomer can be calculated from the literature. RNAsmost commonly adopt either A-form or A'-form conformation [54, 55]. A-form RNA has 11 nucleotide perhelical pitch and A'-form has 12 nucleotide per helical pitch [55]. The rise per base pair for the A-form andA'-form are 0.38nm and 0.27nm, respectively [56]. Given that most of ΨRNA nucleotides are self-pairedwith each other as shown in Fig.1(b), ΨRNA should be double-stranded with 55 base pairs. This leads to alength of 20.9nm and 14.9nm for the A-form and A'-form respectively. These values are consistent with themean measured monomer length 17.9nm. The width of ΨRNA monomer is 1.01 ± 0.02nm consistent withthe height of ΨRNA. For the monomer the width and heights will be the same as it is cylindrical in shape.The mean measured length of ΨRNA dimer is 34.6 ± 0.3nm, which is approximately twice as long as thatof ΨRNA monomer. This means ΨRNA dimer consists of two ΨRNA monomers connecting head to head.This assumption is reasonable given that DIS is located in the SL1 loop of ΨRNA as shown in Fig.1(b).The width of ΨRNA dimer is roughly 3.8 ± 0.1nm that is much larger than two times the width of ΨRNAmonomer. This is probably because at the junction of two ΨRNA monomers the width is larger thanexpected. Next the population distribution between monomers and dimers was analyzed. The length andwidth histograms in Fig. 3(b) were fit to a normal distribution. Both histograms lead to a populationdistribution of 74% monomer and 26% dimer respectively3.2 TARpolyA RNA Size and Morphology MeasurementsTARpolyA RNA (0.5μM) being negatively charged was also imaged on positively charged APTEStreated mica. Similar to ΨRNA, the motivation for measuring TARpolyA RNA by itself is to benchmark theexperiments and analysis protocol as well as understand its individual morphology for comparison with thatobserved in the various GagΔP6 complexes. The observed typical AFM image of TARpolyA RNA in thebuffer solution is shown in Fig.4. In Fig.4(a), TARpolyA RNA image shows that most of TARpolyA RNAmolecules seem to have inverted “L” shape just like ΨRNA. Some examples are shown enclosed in redboxes. The size distribution of the length (longest dimension), width (longest perpendicular dimension tothe length) and the height are shown in Fig. 4(b). As shown in Fig.4(b) and in Table 1, the mean height ofTARpolyA RNA is 1.10 ± 0.01nm. This height is consistent with that of ΨRNA and expectations fromstructure. In Fig. 4(b) and (c) we observe only one peak in the length and width distributions. The meanlength of the TARpolyA RNA monomer is 17.1 ± 0.2nm. This value is slightly less than that of the ΨRNAmonomer observed earlier. Similar to ΨRNA, TARpolyA RNA should also be doubled-stranded as shownin Fig.1(a). Thus the slightly smaller length of 0.8nm is reasonable from the 5 fewer nucleotides, given that
ΨRNA contains 109 nucleotides while TARpolyA RNA has only 104 nucleotides. It is noteworthy that thewidth histogram distribution of TARpolyA RNA is 1.10 ± 0.05nm which is the same as the height andconsistent with the cylindrical structure for TARpolyA RNA. As the length and width distribution have onlyone observable peak, it is concluded that the TARpolyA exists in solution predominantly as a monomer.This is unlike that observed with ΨRNA where 26% dimers were found.3.3 GagΔP6 Size and Morphology MeasurementsThe morphology of GagΔP6 (0.5μM), being net positive charge, was measured using the AFM on afreshly negatively charged mica(-) substrate in solution. A typical AFM image is shown in Fig.5(a). Themotivation for measuring GagΔP6 is to serve as a control before addition of RNAs and lipids. Somecharacteristics of common shapes observed were shown in red, green and blue boxes. In Fig.5(a), the AFMimage of GagΔP6 shows that most of GagΔP6 molecules have ellipsoidal shape rather than the expectedrod-like shape of the extended molecule.As shown in Fig.5(b) and Table 1, the mean height is 1.93 ± 0.01nm which is consistent with theexpectation that the diameter of Gag is around 2~3nm as reported in Ref. [4]. The length and width of theobserved images are plotted in Fig.5(b). As observed three peaks were observed in the length distributionbut only two peaks in the width distribution. Normal distributions were fit to the length histogram to findpeak mean values of 10.3 ± 0.1nm, 20.0 ± 0.2nm and 29.0 ± 0.3nm. Similarly the mean values of the twowidth distributions were found to be 6.2 ± 0.1nm and 12.9 ± 0.2nm respectively.It might seem counterintuitive to have three peaks for length while having only two peaks for thewidth. To have more insight, a three dimensional smooth histogram of the same length-width data wasplotted as shown in Fig.5(c). The shortest length and width distribution would correspond to that of themonomer and the distribution with the next larger length would be expected to correspond to that of thedimer. From Fig.5(c) the monomer and dimer have the same width. Thus in the width distributionhistogram they are represented together and correspond to the first peak at 6.2 ± 0.1nm. The statisticalanalysis of the population also confirmed this assumption as the first peak of the width histogram is thesum of monomer and dimer populations in the two peaks of the length histogram as given in Table 1. Thisanalysis was done by fitting each length and width peak to a normal distribution. From the lengthdistribution, the percentages of monomer, dimer, and tetramer are 59%, 35%, and 6%, respectively. Thepopulation observed in the monomer and dimer from the length adds up to that observed in the first peak ofthe width distribution.Previous studies using hydrodynamic and neutron scattering measurements showed HIV GagΔP6 issupposed to adopt a compact conformation such that MA and NC domains of GagΔP6 are close to eachother even though both of them are positively charged [3-5, 57]. The hydrodynamic radius, R h , of GagΔP6given by three different hydrodynamic tests are 3.6nm, 3.8nm, and 4.1nm, respectively. The radius ofgyraton, R g , of GagΔP6 is best estimated to be 3.4nm from small angle neutron scattering (SANS). The R g of GagΔP6 when it is a 25nm straight rod is supposed to be 7.2nm [57]. The average R g of GagΔP6 insolution measured by SANS is also a monotonically increasing function of the GagΔP6 concentration, with maximum of R g = 5nm at extremely high concentration, which means GagΔP6 molecules are inmonomer-dimer equilibrium [57].Based on the AFM studies presented above, we can project approximate confirmations based on thevarious lengths and widths of the populations observed. For the case of the monomer given a length of10.3nm and width of 6.2nm a potential “C” like shape can be conjectured. This is consistent with the MAdomain being close to the NC domain as has been reported earlier for monomers in solution [4]. Aschematic of a potential monomer structure is shown in Fig.6. Based on the AFM measured dimensions of alength of 20.0 nm and width of 6.2nm (same as monomer) a model of the GagΔP6 dimer would have twomonomers connecting back to back through their CA-CA interaction [4]. The CA-CA interaction leadingto dimerization has been reported in the literature [4]. The last population size distribution of GagΔP6 inFig.5 (c) has a length of 29.0nm and width is 12.9nm. This population from gel electrophoresis correspondsto that of a tetramer (see Supplemental Material). From the measured size of the tetramers here, they couldbe potentially formed by the interaction of two dimers.3.4 GagΔP6-ΨRNA Complex Size, Morphology and InteractionAFM measurements of the GagΔP6-ΨRNA (0.5μM : 0.5μM) complex were done on negativelycharged mica(-). The motivation for measuring GagΔP6-ΨRNA complex is to investigate the effect ofaddition of specific ΨRNA to GagΔP6. According to current models the NC domain binds with ΨRNA andthis interaction is specific and is a critical step in the formation of HIV [22]. Fig.7(a) is a typical AFMimage of GagΔP6-ΨRNA complex. As shown in Fig.7(b), the mean height of GagΔP6-ΨRNA complex is1.90 ± 0.01nm which is roughly the same as the height of just GagΔP6 discussed earlier. This height isconsistent with expectation given that the height of GagΔP6 and ΨRNA are 1.93nm and 1.10nm,respectively. The length and width histograms of the complexes observed are shown in Fig.7(b). Similar toGagΔP6, GagΔP6-ΨRNA also has three peaks for length and two peaks for width. Normal distributionswere fit to the length histogram to find peak mean values of 10.6 ± 0.2nm, 22.4 ± 0.1nm and 31.2 ± 0.2nm.Similarly the mean values of the two width distributions were found to be 6.8 ± 0.1nm and 13.8 ± 0.1nmrespectively. The size of the length and widths are slightly larger than that found with only GagΔP6. This isconsistent with attachment of ΨRNA of around 1nm to the GagΔP6.To understand the role of the ΨRNA addition the three-dimensional population distribution of thelength and width as shown in Fig.7(c) was analyzed. As with GagΔP6, three peaks corresponding tomonomer, dimer and tetramer complexes were observed. However the population distributions of themonomer, dimer and tetramer are different for the GagΔP6- ΨRNA complex. Here we observed 35%monomer, 49% dimers and 16% tetramers. In contrast with only GagΔP6 we observed 59% monomers 35%dimers and 6% tetramers. The monomer decreased by 24%, the dimer increased by 14% and the tetramerincreased by 10%. Thus the addition of ΨRNA promotes multimerization of the GagΔP6 leading to higherpopulations of dimers and tetramers.The length of monomer remained roughly the same as prior to the addition of ΨRNA. The lengths ofdimer and tetramer were increased by about 2nm. The width increased by about 0.5nm~1nm for the monomer, dimer, and tetramer. This is consistent with the addition of ΨRNA of around 1nm to thisdimension. The most important conclusion of the effect of the addition of ΨRNA to GagΔP6 is that ΨRNAcan bind with GagΔP6 and facilitate GagΔP6 multimerization given the increases in percentages of dimerand tetramer population.3.5 GagΔP6-TARpolyA RNA Complex Size, Morphology and InteractionGagΔP6-TARpolyA RNA (0.5μM : 0.5μM) complex was measured on a negatively chargedmica(-) surface. The motivation for measuring GagΔP6-TARpolyA RNA complex is to verify the effect ofthe addition of non-specific TARpolyA RNA to GagΔP6 and compare it to the addition of specific ΨRNAdiscussed above. Fig.8(a) is a typical AFM image of GagΔP6-TARpolyA RNA complex. As shown inFig.8(b), the mean height of GagΔP6-ΨRNA complex is 1.93 ± 0.01nm.To understand the role of the TARPolyA RNA interaction with GagΔP6 the three dimensionalpopulation distribution of the length and width as shown in Fig.8(c) was analyzed. As with GagΔP6, threepeaks corresponding to monomer, dimer and tetramer complexes were observed. However in contrast to thecase of the GagΔP6-ΨRNA complex in Fig.7(c) the population distribution of monomer, dimer andtetramer are very similar to that of GagΔP6 alone observed in Fig.5(c). The sizes of GagΔP6-TARpolyARNA complex monomer, dimer and tetramer were found slightly larger than that of GagΔP6 alone. Thusthe addition of TARPolyA RNA might interact with GagΔP6 but does not promotes multimerization of theGagΔP6 as observed with ΨRNA. These experiments were repeated and the results were alwaysreproducible. Webb et al. reported HIV Gag can bind with both ΨRNA and TARpolyA RNA but withdistinct binding mechanisms [22]. They proposed that HIV GagΔP6 binds with TARpolyA RNA throughboth MA and NC domains whereas ΨRNA binds only through NC domain and leaves MA domain free tolater interact with the lipid membrane. Other studies also showed that HIV GagΔP6 can bind with bothΨRNA and non-Ψ RNAs but the selective binding with ΨRNA is more energetically favorable than othernon-Ψ RNAs for HIV virus assembly [58-60]. The role of the RNA in multimerization of the Gag was notaddressed in these studies. The conclusion based on our data is that it is highly likely that HIV GagΔP6interacts with ΨRNA and TARpolyA RNA through different mechanisms such that ΨRNA facilitatesGagΔP6 multimerization while TARpolyA RNA does not.3.6 GagΔP6-PI(4,5)P2 Complex Size, Morphology and InteractionThe GagΔP6-PI(4,5)P2 (0.5μM : 0.5μM) complex was measured on negatively charged mica(-). The Gagand lipid were diluted to 1μM before mixing. Then, equal amounts of Gag and lipid were mixed. The AFMmeasurement was performed after the mixture was incubated for 3 hours.. A typical AFM image is shownin Fig.9(a). The motivation for measuring GagΔP6-PI(4,5)P2 complex is to explore the effect of addition oflipid PI(4,5)P2 to GagΔP6. The current understanding is that both MA and NC domains of Gag can bind toPI(4,5)P2 through electrostatic forces.As shown in Fig.9(b), the mean height of GagΔP6-PI(4,5)P2 complex is 1.93 ± 0.01nm that isroughly the same as the height of just GagΔP6. This height is consistent with expectation given the smallsize of PI(4,5)P2 in comparison to GagΔP6. The length and width histograms of the complexes observed are shown in Fig.9(b). Similar to GagΔP6, GagΔP6-PI(4,5)P2 also has three peaks for length and two peaksfor width. Normal distributions were fit to the length histogram to find peak mean values of 11.2 ± 0.1nm,21.1 ± 0.1nm and 30.4 ± 0.2nm. Similarly the mean values of the two width distributions were found to be6.7 ± 0.1nm and 14.0 ± 0.2nm respectively. The size of the length and widths are slightly larger than thatwith only GagΔP6. In comparison to the GagΔP6-ΨRNA complex the GagΔP6-PI(4,5)P2 complex hasmonomers of slighter larger length while the dimers and tetramers are of slightly smaller length.To understand the role of PI(4,5)P2 addition to GagΔP6 the three dimensional populationdistribution of the length and width as shown in Fig.9(c) was analyzed. As with GagΔP6, three peakscorresponding to monomer, dimer and tetramer complexes were observed. However the populationdistribution of monomer, dimer and tetramer are different from that observed with GagΔP6 alone or for theGagΔP6- ΨRNA complex. Here we observed 47% monomer, 41% dimers and 12% tetramers. In contrastwith only GagΔP6 we observed 59% monomers 35% dimers and 6% tetramers and for the GagΔP6- ΨRNAcomplex we observed 35% monomer, 49% dimers and 16% tetramers. Thus the addition of PI(4,5)P2promotes multimerization of the GagΔP6. The increases in percentages of dimer and tetramer indicate thatPI(4,5)P2 can bind with GagΔP6 as reported in other studies using confocal microscopy, nuclear magneticresonance and equilibrium flotation assay [35, 40, 61-65]. But the increase in GagΔP6 multimerization withthe addition of PI(4,5)P2 is less than that observed with ΨRNA. The length and width of monomer, dimer,tetramer all increased by about 1nm. This is probably because of the size of PI(4,5)P2 attached to the endsof both MA and CA domains of HIV GagΔP6.3.7 PI(4,5)P2-ΨRNA-GagΔP6 Complex Size, Morphology and InteractionThe PI(4,5)P2-ΨRNA-GagΔP6 (0.5μM : 0.5μM : 0.5μM) complex was measured on a negativelycharged mica(-). According to the prevailing model the lipid interacts with the MA domain ofGagΔP6-ΨRNA complex leading to a conformational change [22]. In these experiments to study thePI(4,5)P2-ΨRNA-GagΔP6 complex, PI(4,5)P2 and ΨRNA were first mixed together in solution followedby the addition of GagΔP6. This mixture was then used to measure the size and size distribution using theAFM. Fig.10(a) is a typical AFM image of PI(4,5)P2-ΨRNA-GagΔP6 complex. As shown in Fig.10(b),the mean height of PI(4,5)P2-ΨRNA-GagΔP6 complex is 1.91 ± 0.01nm that is roughly the same as theheight of just GagΔP6. This height is consistent with expectation given that the height of GagΔP6 andΨRNA-GagΔP6 discussed earlier. The height of PI(4,5)P2 is much smaller by comparison. The length andwidth histograms of the complexes observed are shown in Fig.10(b). In contrast to GagΔP6,GagΔP6-ΨRNA and the GagΔP6- PI(4,5)P2 complexes there are only two peaks for length and two peaksfor width. Normal distributions were fit to the length histogram to find peak mean values of 10.9 ± 0.3nm,and 23.8 ± 0.2nm. Similarly the mean values of the two width distributions were found to be 7.4 ± 0.2nmand 22.1 ± 0.2nm respectively. In comparison to the binary complexes studied earlier the monomer lengthis approximately similar, while the monomer width is slightly larger. For the dimer, the length is almost60% larger than that of GagΔP6-ΨRNA and the GagΔP6- PI(4,5)P2 binary complexes . To understand the PI(4,5)P2-ΨRNA-GagΔP6 complex the three dimensional populationdistribution of the length and width as shown in Fig.10(c) was analyzed. In contrast to all others above,here the dimer and tetramer have the same length. The width of the tetramer is much larger than allprevious cases, increasing from 13-14nm to 22nm. Thus in the PI(4,5)P2-ΨRNA-GagΔP6 complex theGagΔP6 undergoes a dramatic conformational change. This is consistent with the GagΔP6 MA and NCdomains moving further away from each other and taking on a rod-like confirmation. From the populationdistribution in Fig.10(c), we observed 17% monomer, 51% dimers and 32% tetramers. In contrast, withonly GagΔP6 were we observed 59% monomers 35% dimers and 6% tetramers, the multimerization hasincreased considerably. Comparing to the populations in the binary mixtures of ΨRNA-GagΔP6 andPI(4,5)P2-GagΔP6 were we observed 35% and 47% monomer, 49% and 41% dimers and 16% and 12%tetramers respectively, here in particular we observe much higher percentage of tetramers. Thus theaddition of PI(4,5)P2 along with ΨRNA promotes not only the dimerization of the GagΔP6 but inparticular the multimerization to tetramers and higher order complexes. In addition, the significantchange of the size indicates that GagΔP6 undergoes some conformational changes when both ΨRNA andPI(4,5)P2 are present [38, 66]. Based on the sizes measured a schematic of the dimer and tetramer structureis proposed in Fig.11. From the size of the tetramers the spacing between the GagΔP6 molecules is around7 nm close to the value of 8nm reported in studies using cET [18,19]. The substantial increases in thepercentages of dimer and tetramer indicate that both ΨRNA and PI(4,5)P2 can bind with GagΔP6 andcollectively facilitate HIV GagΔP6 assembly as reported in other studies [57, 63].4 ConclusionThe AFM technique was utilized to study the morphology of GagΔP6 (0.5μM), ΨRNA (0.5μM), andtheir binding complexes with the lipid PI(4,5)P2 in HEPES buffer with 0.1 Å vertical and 1nm lateralresolution. For the calibration 2nm diameter Au spheres were used. TARpolyA RNA was used as a negativeRNA control. The morphology of specific complexes GagΔP6-ΨRNA (0.5μM : 0.5μM),GagΔP6-TARpolyA RNA (0.5μM : 0.5μM), GagΔP6-PI(4,5)P2 (0.5μM : 0.5μM) and PI(4,5)P2-ΨRNA-GagΔP6 (0.5μM : 0.5μM : 0.5μM) were studied. They were imaged on either positively charged ornegatively charged mica substrates depending on the net charges carried by the respective materials. Thesize and morphology of both ΨRNA and TARpolyA RNA measured was used to validate the technique incomparison with literature. For the ΨRNA, from the measured size two distinct populations correspondingto monomers and dimers were observed. In the case of TARpoly A RNA only the monomer population wasfound for the concentrations studied. The morphology of GagΔP6, being net positively charged, wasmeasured using the AFM on a freshly cleaved negatively charged mica(-) substrate in HEPES solution witha low salt concentration of 50mM NaCl. Three distinct size populations were found. They were found tocorrespond to 59% monomer, 35% dimer and 6% tetramer form of GagΔP6. The presence of the multimerswas confirmed by gel electrophoresis. The addition of ΨRNA to 0.5μM GagΔP6 was observed to promotemultimerization of the GagΔP6 leading to higher populations of dimers (14% increase) and tetramers (10%increase). The small change in size of the complexes confirmed the binding of the ΨRNA to GagΔP6.The addition TARPolyA RNA to GagΔP6 did not modify the GagΔP6 population distribution of monomers,dimers and tetramers. The interaction of PI(4,5)P2 with GagΔP6 complex was next measured onnegatively charged mica(-). From the population distribution of the monomers (47%), dimers (41%) and tetramers (12%), it was concluded that PI(4,5)P2 promotes multimerization of GagΔP6 but not to the extentobserved with ΨRNA.The PI(4,5)P2-ΨRNA-GagΔP6 ternary complex was next studied using a negatively chargedmica(-) surface. Both the population and size distribution of the GagΔP6 was completely different fromthat of the GagΔP6-ΨRNA and the GagΔP6- PI(4,5)P2 binary complexes. The population distribution ofthe monomers (17%), dimers (51%) and tetramers (32%) was significantly different. Given the largefraction of dimers and tetramers it was concluded that the presence of both PI(4,5)P2 and ΨRNA promotesextensive multimerization of GagΔP6. In addition, the significant change of size indicates that GagΔP6undergoes a conformational change to a 23.8nm rod like shape when both ΨRNA and PI(4,5)P2 are present.From the size of the tetramers the spacing between the GagΔP6 molecules is around 7 nm which isconsistent with studies using cET in Ref [18, 19]. The substantial increases in the percentages of dimerand tetramer indicate that both ΨRNA and PI(4,5)P2 can bind with GagΔP6 and collectively facilitate HIVGagΔP6 assembly as reported using other techniques.Author Contributionsdesigned research: S. C., U. M.performed research: S. C.contributed analytic tools: S. C., M. L.analyzed data: S. C., J. X., A.L.N. R., R. Z., S.S. G., U. M.wrote the manuscript: S. C., U. M.AcknowledgementsWe would like to acknowledge discussions with Karin Musier-Forsyth, Erick D. Olson and Ioulia Rouzina.We would like to thank Erick Olson and Karin Musier-Forsyth for providing the in vitro transcribed RNAs.We also acknowledge discussions with Alan Rein and thank him and S.A.K. Datta for providing theGagΔP6 molecules.5
ΨRNA contains 109 nucleotides while TARpolyA RNA has only 104 nucleotides. It is noteworthy that thewidth histogram distribution of TARpolyA RNA is 1.10 ± 0.05nm which is the same as the height andconsistent with the cylindrical structure for TARpolyA RNA. As the length and width distribution have onlyone observable peak, it is concluded that the TARpolyA exists in solution predominantly as a monomer.This is unlike that observed with ΨRNA where 26% dimers were found.3.3 GagΔP6 Size and Morphology MeasurementsThe morphology of GagΔP6 (0.5μM), being net positive charge, was measured using the AFM on afreshly negatively charged mica(-) substrate in solution. A typical AFM image is shown in Fig.5(a). Themotivation for measuring GagΔP6 is to serve as a control before addition of RNAs and lipids. Somecharacteristics of common shapes observed were shown in red, green and blue boxes. In Fig.5(a), the AFMimage of GagΔP6 shows that most of GagΔP6 molecules have ellipsoidal shape rather than the expectedrod-like shape of the extended molecule.As shown in Fig.5(b) and Table 1, the mean height is 1.93 ± 0.01nm which is consistent with theexpectation that the diameter of Gag is around 2~3nm as reported in Ref. [4]. The length and width of theobserved images are plotted in Fig.5(b). As observed three peaks were observed in the length distributionbut only two peaks in the width distribution. Normal distributions were fit to the length histogram to findpeak mean values of 10.3 ± 0.1nm, 20.0 ± 0.2nm and 29.0 ± 0.3nm. Similarly the mean values of the twowidth distributions were found to be 6.2 ± 0.1nm and 12.9 ± 0.2nm respectively.It might seem counterintuitive to have three peaks for length while having only two peaks for thewidth. To have more insight, a three dimensional smooth histogram of the same length-width data wasplotted as shown in Fig.5(c). The shortest length and width distribution would correspond to that of themonomer and the distribution with the next larger length would be expected to correspond to that of thedimer. From Fig.5(c) the monomer and dimer have the same width. Thus in the width distributionhistogram they are represented together and correspond to the first peak at 6.2 ± 0.1nm. The statisticalanalysis of the population also confirmed this assumption as the first peak of the width histogram is thesum of monomer and dimer populations in the two peaks of the length histogram as given in Table 1. Thisanalysis was done by fitting each length and width peak to a normal distribution. From the lengthdistribution, the percentages of monomer, dimer, and tetramer are 59%, 35%, and 6%, respectively. Thepopulation observed in the monomer and dimer from the length adds up to that observed in the first peak ofthe width distribution.Previous studies using hydrodynamic and neutron scattering measurements showed HIV GagΔP6 issupposed to adopt a compact conformation such that MA and NC domains of GagΔP6 are close to eachother even though both of them are positively charged [3-5, 57]. The hydrodynamic radius, R h , of GagΔP6given by three different hydrodynamic tests are 3.6nm, 3.8nm, and 4.1nm, respectively. The radius ofgyraton, R g , of GagΔP6 is best estimated to be 3.4nm from small angle neutron scattering (SANS). The R g of GagΔP6 when it is a 25nm straight rod is supposed to be 7.2nm [57]. The average R g of GagΔP6 insolution measured by SANS is also a monotonically increasing function of the GagΔP6 concentration, with maximum of R g = 5nm at extremely high concentration, which means GagΔP6 molecules are inmonomer-dimer equilibrium [57].Based on the AFM studies presented above, we can project approximate confirmations based on thevarious lengths and widths of the populations observed. For the case of the monomer given a length of10.3nm and width of 6.2nm a potential “C” like shape can be conjectured. This is consistent with the MAdomain being close to the NC domain as has been reported earlier for monomers in solution [4]. Aschematic of a potential monomer structure is shown in Fig.6. Based on the AFM measured dimensions of alength of 20.0 nm and width of 6.2nm (same as monomer) a model of the GagΔP6 dimer would have twomonomers connecting back to back through their CA-CA interaction [4]. The CA-CA interaction leadingto dimerization has been reported in the literature [4]. The last population size distribution of GagΔP6 inFig.5 (c) has a length of 29.0nm and width is 12.9nm. This population from gel electrophoresis correspondsto that of a tetramer (see Supplemental Material). From the measured size of the tetramers here, they couldbe potentially formed by the interaction of two dimers.3.4 GagΔP6-ΨRNA Complex Size, Morphology and InteractionAFM measurements of the GagΔP6-ΨRNA (0.5μM : 0.5μM) complex were done on negativelycharged mica(-). The motivation for measuring GagΔP6-ΨRNA complex is to investigate the effect ofaddition of specific ΨRNA to GagΔP6. According to current models the NC domain binds with ΨRNA andthis interaction is specific and is a critical step in the formation of HIV [22]. Fig.7(a) is a typical AFMimage of GagΔP6-ΨRNA complex. As shown in Fig.7(b), the mean height of GagΔP6-ΨRNA complex is1.90 ± 0.01nm which is roughly the same as the height of just GagΔP6 discussed earlier. This height isconsistent with expectation given that the height of GagΔP6 and ΨRNA are 1.93nm and 1.10nm,respectively. The length and width histograms of the complexes observed are shown in Fig.7(b). Similar toGagΔP6, GagΔP6-ΨRNA also has three peaks for length and two peaks for width. Normal distributionswere fit to the length histogram to find peak mean values of 10.6 ± 0.2nm, 22.4 ± 0.1nm and 31.2 ± 0.2nm.Similarly the mean values of the two width distributions were found to be 6.8 ± 0.1nm and 13.8 ± 0.1nmrespectively. The size of the length and widths are slightly larger than that found with only GagΔP6. This isconsistent with attachment of ΨRNA of around 1nm to the GagΔP6.To understand the role of the ΨRNA addition the three-dimensional population distribution of thelength and width as shown in Fig.7(c) was analyzed. As with GagΔP6, three peaks corresponding tomonomer, dimer and tetramer complexes were observed. However the population distributions of themonomer, dimer and tetramer are different for the GagΔP6- ΨRNA complex. Here we observed 35%monomer, 49% dimers and 16% tetramers. In contrast with only GagΔP6 we observed 59% monomers 35%dimers and 6% tetramers. The monomer decreased by 24%, the dimer increased by 14% and the tetramerincreased by 10%. Thus the addition of ΨRNA promotes multimerization of the GagΔP6 leading to higherpopulations of dimers and tetramers.The length of monomer remained roughly the same as prior to the addition of ΨRNA. The lengths ofdimer and tetramer were increased by about 2nm. The width increased by about 0.5nm~1nm for the monomer, dimer, and tetramer. This is consistent with the addition of ΨRNA of around 1nm to thisdimension. The most important conclusion of the effect of the addition of ΨRNA to GagΔP6 is that ΨRNAcan bind with GagΔP6 and facilitate GagΔP6 multimerization given the increases in percentages of dimerand tetramer population.3.5 GagΔP6-TARpolyA RNA Complex Size, Morphology and InteractionGagΔP6-TARpolyA RNA (0.5μM : 0.5μM) complex was measured on a negatively chargedmica(-) surface. The motivation for measuring GagΔP6-TARpolyA RNA complex is to verify the effect ofthe addition of non-specific TARpolyA RNA to GagΔP6 and compare it to the addition of specific ΨRNAdiscussed above. Fig.8(a) is a typical AFM image of GagΔP6-TARpolyA RNA complex. As shown inFig.8(b), the mean height of GagΔP6-ΨRNA complex is 1.93 ± 0.01nm.To understand the role of the TARPolyA RNA interaction with GagΔP6 the three dimensionalpopulation distribution of the length and width as shown in Fig.8(c) was analyzed. As with GagΔP6, threepeaks corresponding to monomer, dimer and tetramer complexes were observed. However in contrast to thecase of the GagΔP6-ΨRNA complex in Fig.7(c) the population distribution of monomer, dimer andtetramer are very similar to that of GagΔP6 alone observed in Fig.5(c). The sizes of GagΔP6-TARpolyARNA complex monomer, dimer and tetramer were found slightly larger than that of GagΔP6 alone. Thusthe addition of TARPolyA RNA might interact with GagΔP6 but does not promotes multimerization of theGagΔP6 as observed with ΨRNA. These experiments were repeated and the results were alwaysreproducible. Webb et al. reported HIV Gag can bind with both ΨRNA and TARpolyA RNA but withdistinct binding mechanisms [22]. They proposed that HIV GagΔP6 binds with TARpolyA RNA throughboth MA and NC domains whereas ΨRNA binds only through NC domain and leaves MA domain free tolater interact with the lipid membrane. Other studies also showed that HIV GagΔP6 can bind with bothΨRNA and non-Ψ RNAs but the selective binding with ΨRNA is more energetically favorable than othernon-Ψ RNAs for HIV virus assembly [58-60]. The role of the RNA in multimerization of the Gag was notaddressed in these studies. The conclusion based on our data is that it is highly likely that HIV GagΔP6interacts with ΨRNA and TARpolyA RNA through different mechanisms such that ΨRNA facilitatesGagΔP6 multimerization while TARpolyA RNA does not.3.6 GagΔP6-PI(4,5)P2 Complex Size, Morphology and InteractionThe GagΔP6-PI(4,5)P2 (0.5μM : 0.5μM) complex was measured on negatively charged mica(-). The Gagand lipid were diluted to 1μM before mixing. Then, equal amounts of Gag and lipid were mixed. The AFMmeasurement was performed after the mixture was incubated for 3 hours.. A typical AFM image is shownin Fig.9(a). The motivation for measuring GagΔP6-PI(4,5)P2 complex is to explore the effect of addition oflipid PI(4,5)P2 to GagΔP6. The current understanding is that both MA and NC domains of Gag can bind toPI(4,5)P2 through electrostatic forces.As shown in Fig.9(b), the mean height of GagΔP6-PI(4,5)P2 complex is 1.93 ± 0.01nm that isroughly the same as the height of just GagΔP6. This height is consistent with expectation given the smallsize of PI(4,5)P2 in comparison to GagΔP6. The length and width histograms of the complexes observed are shown in Fig.9(b). Similar to GagΔP6, GagΔP6-PI(4,5)P2 also has three peaks for length and two peaksfor width. Normal distributions were fit to the length histogram to find peak mean values of 11.2 ± 0.1nm,21.1 ± 0.1nm and 30.4 ± 0.2nm. Similarly the mean values of the two width distributions were found to be6.7 ± 0.1nm and 14.0 ± 0.2nm respectively. The size of the length and widths are slightly larger than thatwith only GagΔP6. In comparison to the GagΔP6-ΨRNA complex the GagΔP6-PI(4,5)P2 complex hasmonomers of slighter larger length while the dimers and tetramers are of slightly smaller length.To understand the role of PI(4,5)P2 addition to GagΔP6 the three dimensional populationdistribution of the length and width as shown in Fig.9(c) was analyzed. As with GagΔP6, three peakscorresponding to monomer, dimer and tetramer complexes were observed. However the populationdistribution of monomer, dimer and tetramer are different from that observed with GagΔP6 alone or for theGagΔP6- ΨRNA complex. Here we observed 47% monomer, 41% dimers and 12% tetramers. In contrastwith only GagΔP6 we observed 59% monomers 35% dimers and 6% tetramers and for the GagΔP6- ΨRNAcomplex we observed 35% monomer, 49% dimers and 16% tetramers. Thus the addition of PI(4,5)P2promotes multimerization of the GagΔP6. The increases in percentages of dimer and tetramer indicate thatPI(4,5)P2 can bind with GagΔP6 as reported in other studies using confocal microscopy, nuclear magneticresonance and equilibrium flotation assay [35, 40, 61-65]. But the increase in GagΔP6 multimerization withthe addition of PI(4,5)P2 is less than that observed with ΨRNA. The length and width of monomer, dimer,tetramer all increased by about 1nm. This is probably because of the size of PI(4,5)P2 attached to the endsof both MA and CA domains of HIV GagΔP6.3.7 PI(4,5)P2-ΨRNA-GagΔP6 Complex Size, Morphology and InteractionThe PI(4,5)P2-ΨRNA-GagΔP6 (0.5μM : 0.5μM : 0.5μM) complex was measured on a negativelycharged mica(-). According to the prevailing model the lipid interacts with the MA domain ofGagΔP6-ΨRNA complex leading to a conformational change [22]. In these experiments to study thePI(4,5)P2-ΨRNA-GagΔP6 complex, PI(4,5)P2 and ΨRNA were first mixed together in solution followedby the addition of GagΔP6. This mixture was then used to measure the size and size distribution using theAFM. Fig.10(a) is a typical AFM image of PI(4,5)P2-ΨRNA-GagΔP6 complex. As shown in Fig.10(b),the mean height of PI(4,5)P2-ΨRNA-GagΔP6 complex is 1.91 ± 0.01nm that is roughly the same as theheight of just GagΔP6. This height is consistent with expectation given that the height of GagΔP6 andΨRNA-GagΔP6 discussed earlier. The height of PI(4,5)P2 is much smaller by comparison. The length andwidth histograms of the complexes observed are shown in Fig.10(b). In contrast to GagΔP6,GagΔP6-ΨRNA and the GagΔP6- PI(4,5)P2 complexes there are only two peaks for length and two peaksfor width. Normal distributions were fit to the length histogram to find peak mean values of 10.9 ± 0.3nm,and 23.8 ± 0.2nm. Similarly the mean values of the two width distributions were found to be 7.4 ± 0.2nmand 22.1 ± 0.2nm respectively. In comparison to the binary complexes studied earlier the monomer lengthis approximately similar, while the monomer width is slightly larger. For the dimer, the length is almost60% larger than that of GagΔP6-ΨRNA and the GagΔP6- PI(4,5)P2 binary complexes . To understand the PI(4,5)P2-ΨRNA-GagΔP6 complex the three dimensional populationdistribution of the length and width as shown in Fig.10(c) was analyzed. In contrast to all others above,here the dimer and tetramer have the same length. The width of the tetramer is much larger than allprevious cases, increasing from 13-14nm to 22nm. Thus in the PI(4,5)P2-ΨRNA-GagΔP6 complex theGagΔP6 undergoes a dramatic conformational change. This is consistent with the GagΔP6 MA and NCdomains moving further away from each other and taking on a rod-like confirmation. From the populationdistribution in Fig.10(c), we observed 17% monomer, 51% dimers and 32% tetramers. In contrast, withonly GagΔP6 were we observed 59% monomers 35% dimers and 6% tetramers, the multimerization hasincreased considerably. Comparing to the populations in the binary mixtures of ΨRNA-GagΔP6 andPI(4,5)P2-GagΔP6 were we observed 35% and 47% monomer, 49% and 41% dimers and 16% and 12%tetramers respectively, here in particular we observe much higher percentage of tetramers. Thus theaddition of PI(4,5)P2 along with ΨRNA promotes not only the dimerization of the GagΔP6 but inparticular the multimerization to tetramers and higher order complexes. In addition, the significantchange of the size indicates that GagΔP6 undergoes some conformational changes when both ΨRNA andPI(4,5)P2 are present [38, 66]. Based on the sizes measured a schematic of the dimer and tetramer structureis proposed in Fig.11. From the size of the tetramers the spacing between the GagΔP6 molecules is around7 nm close to the value of 8nm reported in studies using cET [18,19]. The substantial increases in thepercentages of dimer and tetramer indicate that both ΨRNA and PI(4,5)P2 can bind with GagΔP6 andcollectively facilitate HIV GagΔP6 assembly as reported in other studies [57, 63].4 ConclusionThe AFM technique was utilized to study the morphology of GagΔP6 (0.5μM), ΨRNA (0.5μM), andtheir binding complexes with the lipid PI(4,5)P2 in HEPES buffer with 0.1 Å vertical and 1nm lateralresolution. For the calibration 2nm diameter Au spheres were used. TARpolyA RNA was used as a negativeRNA control. The morphology of specific complexes GagΔP6-ΨRNA (0.5μM : 0.5μM),GagΔP6-TARpolyA RNA (0.5μM : 0.5μM), GagΔP6-PI(4,5)P2 (0.5μM : 0.5μM) and PI(4,5)P2-ΨRNA-GagΔP6 (0.5μM : 0.5μM : 0.5μM) were studied. They were imaged on either positively charged ornegatively charged mica substrates depending on the net charges carried by the respective materials. Thesize and morphology of both ΨRNA and TARpolyA RNA measured was used to validate the technique incomparison with literature. For the ΨRNA, from the measured size two distinct populations correspondingto monomers and dimers were observed. In the case of TARpoly A RNA only the monomer population wasfound for the concentrations studied. The morphology of GagΔP6, being net positively charged, wasmeasured using the AFM on a freshly cleaved negatively charged mica(-) substrate in HEPES solution witha low salt concentration of 50mM NaCl. Three distinct size populations were found. They were found tocorrespond to 59% monomer, 35% dimer and 6% tetramer form of GagΔP6. The presence of the multimerswas confirmed by gel electrophoresis. The addition of ΨRNA to 0.5μM GagΔP6 was observed to promotemultimerization of the GagΔP6 leading to higher populations of dimers (14% increase) and tetramers (10%increase). The small change in size of the complexes confirmed the binding of the ΨRNA to GagΔP6.The addition TARPolyA RNA to GagΔP6 did not modify the GagΔP6 population distribution of monomers,dimers and tetramers. The interaction of PI(4,5)P2 with GagΔP6 complex was next measured onnegatively charged mica(-). From the population distribution of the monomers (47%), dimers (41%) and tetramers (12%), it was concluded that PI(4,5)P2 promotes multimerization of GagΔP6 but not to the extentobserved with ΨRNA.The PI(4,5)P2-ΨRNA-GagΔP6 ternary complex was next studied using a negatively chargedmica(-) surface. Both the population and size distribution of the GagΔP6 was completely different fromthat of the GagΔP6-ΨRNA and the GagΔP6- PI(4,5)P2 binary complexes. The population distribution ofthe monomers (17%), dimers (51%) and tetramers (32%) was significantly different. Given the largefraction of dimers and tetramers it was concluded that the presence of both PI(4,5)P2 and ΨRNA promotesextensive multimerization of GagΔP6. In addition, the significant change of size indicates that GagΔP6undergoes a conformational change to a 23.8nm rod like shape when both ΨRNA and PI(4,5)P2 are present.From the size of the tetramers the spacing between the GagΔP6 molecules is around 7 nm which isconsistent with studies using cET in Ref [18, 19]. The substantial increases in the percentages of dimerand tetramer indicate that both ΨRNA and PI(4,5)P2 can bind with GagΔP6 and collectively facilitate HIVGagΔP6 assembly as reported using other techniques.Author Contributionsdesigned research: S. C., U. M.performed research: S. C.contributed analytic tools: S. C., M. L.analyzed data: S. C., J. X., A.L.N. R., R. Z., S.S. G., U. M.wrote the manuscript: S. C., U. M.AcknowledgementsWe would like to acknowledge discussions with Karin Musier-Forsyth, Erick D. Olson and Ioulia Rouzina.We would like to thank Erick Olson and Karin Musier-Forsyth for providing the in vitro transcribed RNAs.We also acknowledge discussions with Alan Rein and thank him and S.A.K. Datta for providing theGagΔP6 molecules.5 References[1] Ganser-Pornillos, B.K., Yeager, M., Pornillos, O.: Assembly and architecture of HIV. Adv. Exp.Med. Biol., 726, 441-465 (2012)[2] Rein, A., Datta, S.A.K., Jones, C.P., Musier-Forsyth, K.: Diverse interactions of retroviral Gagproteins with RNAs. Trends. Biochem. Sci., 36, 373-380 (2011)[3] Briggs, J.A.G., Simon, M.N., Gross, I., Kräusslich, H.G., Fuller, S.D., Vogt, V.M., Johnson, M.C.: Thestoichiometry of Gag protein in HIV-1. Nat. Struct. Mol. Biol., 11, 672–675 (2004)[4] Datta, S.A.K., Curtis, J.E., Ratcliff, W., Clark, P.K., Crist, R.M., Lebowitz, J., Krueger, S., Rein A.:Conformation of the HIV-1 Gag protein in solution. J. Mol. Biol., 365, 812-824 (2007)[5] Paul M. S. and Beatrice H. H.: Origins of HIV and the AIDS Pandemic. Cold Spring Harb Perspect Med.1, 1, a006841 (2011)[6] Sundquist, W.I. Kräusslich, H.G.: HIV-1 Assembly, budding, and maturation. CSH Perspect. Med., 2, 7,a006924 (2012)[7] Freed E.O.: HIV-1 assembly, release and maturation. Nat. Rev. Microbiol., 13, 484–496 (2015)[8] Bell, N.M., Lever, A.M.L.: HIV Gag polyprotein: processing and early viral particle assembly. TrendsMicrobiol., 21, 136-144 (2013)[9] Briggs, J.A.G., Kräusslich, H.G.: The Molecular Architecture of HIV. J. Mol. Biol., 410, 491-500(2011)[10] Pettit, S.C., Sheng, N., Tritch, R., Erickson-Viitanen, S., Swanstrom, R.: The regulation of sequentialprocessing of HIV-1 Gag by the viral protease. Adv. Exp. Med. Biol., 436, 15-25 (1998)[11] Massiah, M.A., Starich, M.R., Paschall, C., Summers, M.F., Christensen, A.M., Sundquist, W.I.:Three-dimensional structure of the human immunodeficiency virus type 1 matrix protein, J. Mol. Biol., 244,198-223 (1994)[12] Chukkapalli, V., Ono, A.: Molecular determinants that regulate plasma membrane association of HIV-1Gag. J. Mol. Biol., 410, 512-524 (2011)[13] Hill, C.P., Worthylake, D., Bancroft, D.P., Christensen, A.M., Sundquist, W.I.: Crystal structures of thetrimeric human immunodeficiency virus type 1 matrix protein: implications for membrane association andassembly. PNAS, 93, 7, 3099-3104 (1996)[14] Göttlinger, H.G., Sodroski, J.G., Haseltine, W.A.: Role of capsid precursor processing andmyristoylation in morphogenesis and infectivity of human immunodeficiency virus type 1. PNAS, 86, 15,5781-5785 (1989)[15] Spearman, P., Wang, J.J., Heyden, N.V., Ratner, L.: Identification of human immunodeficiency virustype 1 Gag protein domains essential to membrane binding and particle assembly. J. Virol., 68, 5,3232-3242 (1994)[16] Bharat, T.A.M., Davey, N.E., Ulbrich, P., Riches, J.D., de Marco, A., Rumlova, M., Sachse, C., Ruml,T., Briggs, J.A. G. (2012). Structure of the immature retroviral capsid at 8 Å resolution by cryo-electronmicroscopy. Nature, 487, 385–389.[17] Kelly, B.N., Howard, B.R., Wang, H., Robinson, H., Sundquist, W.I., Hill, C.P.: Implications for viralcapsid assembly from crystal structures of HIV-1 Gag(1-278) and CA(N)(133-278). Biochemistry, 45, 38,11257–11266 (2006)[18] Wright, E.R., Schooler, J.B., Ding, H.J., Kieffer, C., Fillmore, C., Sundquist, W.I., Jensen, G.J.: Electron cryotomography of immature HIV-1 virions reveals the structure of the CA and SP1 Gag shells.EMBO J., 26, 2218-2226 (2007)[19] Briggs, J.A., Riches, J.D., Glass, B., Bartonova, V., Zanetti, G., Kräusslich, H.G.: Structure andassembly of immature HIV. PNAS, 106, 27, 11090-11095 (2009)[20] De Guzman, R.N., Wu, Z.R., Stalling, C.C., Pappalardo, L., Borer P.N., Summers, M.F.: Structure ofthe HIV-1 nucleocapsid protein bound to the SL3 psi-RNA recognition element. Science, 16, 279,5349,384-388 (1998)[21] Godet, J., Boudier, C., Humbert, N., Ivanyi-Nagy, R., Darlix, J.L., Mély, Y.: Comparative nucleicacid chaperone properties of the nucleocapsid protein NCp7 and Tat protein of HIV-1. Virus Res., 169, 2,349-360 (2012)[22] Webb, J.A., Jones, C.P., Parent, L.J., Rouzina, I., Musier-Forsyth, K.: Distinct binding interactions ofHIV-1 Gag to Psi and non-Psi RNAs: implications for viral genomic RNA packaging. RNA, 19, 8,1078-1088 (2013)[23] Kafaie, J., Song, R., Abrahamyan, L., Mouland, A.J., Laughrea, M.: Mapping of nucleocapsid residuesimportant for HIV-1 genomic RNA dimerization and packaging. Virology, 375, 2, 592-610 (2008)[24] Thomas, J.A., Gorelick, R.J.: Nucleocapsid protein function in early infection processes. Virus Res.,134, 1-2, 39-63 (2008)[25] Watts, J.M., Dang, K.K., Gorelick, R.J., Leonard, C.W., Bess, Jr J.W., Swanstrom, R., Burch, C.L.,Weeks, K.M.: Architecture and secondary structure of an entire HIV-1 RNA genome. Nature, 460, 711–716(2009)[26] Jones, C.P., Cantara, W.A., Olson, E.D., Musier-Forsyth, K.: Small-angle X-ray scattering-derivedstructure of the HIV-1 5' UTR reveals 3D tRNA mimicry. PNAS, 111, 9, 3395-3400 (2014)[27] Wilkinson, K.A., Gorelick, R.J., Vasa, S.M., Guex, N., Rein, A., Mathews, D.H., Giddings, M.C.,Weeks, K.M.: High-throughput SHAPE analysis reveals structures in HIV-1 genomic RNA stronglyconserved across distinct biological states. PLoS Biol. 6, 4, e96 (2008)[28] Andersen, E.S., Contera, S.A., Knudsen, B., Damgaard, C.K., Besenbacher, F., Kjems, J.: Role of thetrans-activation response element in dimerization of HIV-1 RNA. J. Biol. Chem., 279, 21, 22243-22249(2004)[29] Frankel, A.D., Young, J.A.: HIV-1: fifteen proteins and an RNA. Annu. Rev. Biochem., 67, 1-25(1998)[30] Moore, M.D., Hu, W.S.: HIV-1 RNA dimerization: It takes two to tango. AIDS Rev., 11, 2, 91-102(2009)[31] Skripkin, E., Paillart, J.C., Marquet, R., Ehresmann, B., Ehresmann, C.: Identification of the primarysite of the human immunodeficiency virus type 1 RNA dimerization in vitro. PNAS, 91, 11, 4945-4949(1994)[32] Lu, K., Heng, X., Summers, M.F.: Structural determinants and mechanism of HIV-1 genomepackaging. J. Mol. Biol., 410, 2, 609-633 (2011)[33] Russell, R.S., Hu, J., Bériault, V., Mouland, A.J., Laughrea, M., Kleiman, L., Wainberg, M.A., Liang,C.: Sequences downstream of the 5' splice donor site are required for both packaging and dimerization ofhuman immunodeficiency virus type 1 RNA. J. Virol., 77, 1, 84-96 (2003)[34] Clever, J.L., Parslow, T.G.: Mutant human immunodeficiency virus type 1 genomes with defects in RNA dimerization or encapsidation. J. Virol., 71, 5, 3407-3414 (1997)[35] 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. PNAS, 101, 41, 14889-14894 (2004)[36] Zhou, W., Parent, L.J., Wills, J.W., Resh, M.D.: Identification of a membrane-binding domain withinthe amino-terminal region of human immunodeficiency virus type 1 Gag protein which interacts withacidic phospholipids. J. Virol., 68, 4 2556-2569 (1994)[37] Zhou, W., Resh, M.D.: Differential membrane binding of the human immunodeficiency virus type 1matrix protein. J. Virol., 70, 12, 8540-8548 (1996)[38] Spearman, P., Horton, R., Ratner, L., Kuli-Zade, I.: Membrane binding of human immunodeficiencyvirus type 1 matrix protein in vivo supports a conformational myristyl switch mechanism. J. Virol., 71, 9,6582-6592 (1997)[39] Ono, A., Freed, E.O.: Binding of Human Immunodeficiency Virus Type 1 Gag to Membrane: Role ofthe Matrix Amino Terminus. J. Virol., 73, 5, 4136-4144 (1999)[40] Saad, J.S., Miller, J., Tai, J., Kim, A., Ghanam, R.H., Summers, M.F.: Structural basis for targetingHIV-1 Gag proteins to the plasma membrane for virus assembly. PNAS, 103, 30, 11364-11369 (2006)[41] Rojas, O.J.: Adsorption of polyelectrolytes on mica. Ency. Surf. Coll. Sci., 517-535 (2002)[42] Müller, D.J., Amrein, M., Engel ,A.: Adsorption of biological molecules to a solid support for scanningprobe microscopy. J. Struct. Biol., 119, 2, 172-188 (1997)[43] Müllerab, D.J., Engela, A., Amreinc, M.: Preparation techniques for the observation of nativebiological systems with the atomic force microscope. Biosens. Bioelectron., 12, 8, 867-877 (1997)[44] Raposo, M., Ferreira, Q., Ribeiro, P. A.: A guide for Atomic Force Microscopy Analysis ofSoft-Condensed Matter. Mod. Res. Educ. Top. Microsc., 758-769 . (2007)[45] Pfreundschuh, M., Martinez-Martin, D., Mulvihill, E., Wegmann, S., Muller, D.J.: Multiparametrichigh-resolution imaging of native proteins by force-distance curve–based AFM. Nat. Protoc., 9, 1113–1130(2014)[46] Lyubchenko, Y.L., Shlyakhtenko, L.S., Ando, T.: Imaging of nucleic acids with atomic forcemicroscopy. Methods, 54, 2, 274-283 (2011)[47] Pan Y., Sun Z., Maiti A., Kanai T., Matsuo H., Li M., Harris R.S., Shlyakhtenko, L.S., Lyubchenko,Y.L.: Nanoscale Characterization of Interaction of APOBEC3G with RNA. Biochemistry, 56 (10),1473-1481 (2017)[48] Ares, P., Fuentes-Perez, M.E., Herrero-Galán, E., Valpuesta, J.M., Gil, A., Gomez-Herrero, J.,Moreno-Herrero, F.: High resolution atomic force microscopy of double-stranded RNA. Nanoscale, 8, 23,11818-11826 (2016)[49] Pastré, D., Piétrement, O., Fusil, S., Landousy, F., Jeusset, J., David, M.O., Hamon, L., Le Cam, E.,Zozime, A.: Adsorption of DNA to mica mediated by divalent counterions: a theoretical and experimentalstudy. Biophys. J., 85, 4, 2507-2518 (2003)[50] Kienberger, F., Costa, L.T., Zhu, R., Kada, G., Reithmayer, M., Chtcheglova, L., Rankl, C., Pacheco,A.B., Thalhammer, S., Pastushenko, V., Heckl, W.M., Blaas, D., Hinterdorfer, P.: Dynamic forcemicroscopy imaging of plasmid DNA and viral RNA. Biomaterials, 28, 15, 2403-2411 (2007)[51] Valbuenaa, A., Mateu, M.G.: Quantification and modification of the equilibrium dynamics andmechanics of a viral capsid lattice self-assembled as a protein nanocoating. Nanoscale. 7(36), 14953-14964 (2015)[52] Bussiek, M., Schöne, A. Nellen, W.: Atomic force microscopy imaging and force spectroscopy ofRNA. Handbook of RNA Biochemistry: Second, Completely Revised and Enlarged Edition, 527-546(2014)[53] Hansma, H.G., Oroudjev, E., Baudrey, S., Jaeger, L.: TectoRNA and 'kissing-loop' RNA: atomic forcemicroscopy of self-assembling RNA structures. J. Microsc., 212, 3, 273-279 (2003)[54] Watson, J.D. Crick, F.H.C.: Molecular structure of nucleic acids: a structure for deoxyribose nucleicacid. Nature, 171, 737-738 (1953)[55] Tanaka Y., Fujii S., Hiroaki H., Sakata T., Tanaka T., Uesugi S., Tomita K., Kyogoku Y.: A'-form RNAdouble helix in the single crystal structure of r(UGAGCUUCGGCUC). Nucleic Acids Res, 27, 4, 949–955(1999)[56] Arnott S., Hukins D.W., Dover S.D.: Optimised parameters for RNA double-helices. Biochem BiophysRes Commun. 48, 6,1392-1399 (1972)[57] Datta, S.A., Heinrich, F., Raghunandan, S., Krueger, S., Curtis, J.E., Rein, A., Nanda, H.: HIV-1Gag extension: conformational changes require simultaneous interaction with membrane and nucleic acid. J.Mol. Biol., 406, 2, 205-214 (2011)[58] Comas-Garcia, M, Davis, S.R., Rein, A.: On the selective packaging of genomic RNA by HIV-1.Viruses, 8, 9, 246 (2016)[59] Didierlaurent, L., Racine, P.J., Houzet, L., Chamontin, C., Berkhout, B., Mougel, M.: Role of HIV-1RNA and protein determinants for the selective packaging of spliced and unspliced viral RNA and host U6and 7SL RNA in virus particles. Nucleic Acids Res., 39, 20, 8915-8927 (2011)[60] Houzet, L., Paillart, J.C., Smagulova, F., Maurel, S., Morichaud, Z., Marquet, R., Mougel, M.: HIVcontrols the selective packaging of genomic, spliced viral and cellular RNAs into virions through differentmechanisms. Nucleic Acids Res., 35, 8, 2695-2704 (2007)[61] Carlson, L.A. Hurley J.H.: In vitro reconstitution of the ordered assembly of the endosomal sortingcomplex required for transport at membrane-bound HIV-1 Gag clusters. PNAS, 109, 42, 16928-16933(2012)[62] Chukkapalli, V., Hogue, I.B., Boyko, V., Hu, W.S., Ono A.: Interaction between the humanimmunodeficiency virus type 1 Gag matrix domain and phosphatidylinositol-(4,5)-bisphosphate is essentialfor efficient gag membrane binding. J. Virol., 82, 5, 2405-2417 (2008)[63] Shkriabai, N., Datta, S.A., Zhao, Z., Hess, S., Rein, A., Kvaratskhelia, M.: Interactions of HIV-1 Gagwith assembly cofactors. Biochemistry, 45, 13, 4077-4083 (2006)[64] Alfadhli, A., Barklis, R.L., Barklis, E.: HIV-1 matrix organizes as a hexamer of trimers on membranescontaining phosphatidylinositol-(4,5)-bisphosphate. Virology, 387, 2, 466-472 (2009)[65] Freed, E.O.: HIV-1 Gag: Flipped out for PI(4,5)P2. PNAS, 103, 30, 11101-11102 (2006)[66] Ono, A., Demirov, D., Freed, E.O.: Relationship between human immunodeficiency virus type 1 Gagmultimerization and membrane binding. J. Virol., 74, 11, 5142-5150 (2000) Fig.1 HIV-1 genomic RNA partial sequence constructs used in this work. (a) TARpolyA RNA. (b) ΨRNA. Fig.2 AFM tip resolution calibration with 2nm diameter Au Sphere. (a) A typical AFM image of 2nm Ausphere on mica in tapping mode in liquid. The scan size is 250nm×250nm. The height color bar scale is2.5nm. (b) Histogram of 2nm diameter Au sphere for the measured size on the plane of the substrate (top)and the measured height (bottom). Shown in red are normal distribution fits to the peaks. The meanmeasured size on the plane of the substrate is 7.56 ± 0.09 nm, and the mean height is 2.10 ± 0.02nm. Theheight is consistent with the sphere diameter. The size in the plane of the substrate reflects the role of thetip size. Please see text and supplemental materials section for more details. The total number of sampleswas 419 and the experiment was repeated twice. Fig.3 0.5μM ΨRNA on positively charged mica(+). (a) A typical AFM image with a scan size of500nm×500nm. The height color bar scale is 2.0 nm. A few characteristic ΨRNAs are shown in boxes:monomer (red) and dimer (green). (b) Histogram for length (left), width (middle), and height (right).Shown in red are normal distribution fits to the peaks. (c) Three dimensional smooth histogram, where redarrows indicate monomer and dimer. The mean height is 1.10 ± 0.01nm. The first peak with a mean lengthof 17.9 ± 0.2nm and width 1.01 ± 0.02nm corresponds to the ΨRNA monomer. The second peak with amean length of 34.6 ± 0.3nm and width 3.8 ± 0.1nm corresponds to the ΨRNA dimer. The total number ofsamples was 551 and the experiment was repeated twice. Fig.4 0.5μM TARpolyA RNA on positively charged mica(+). (a) AFM image with a scan size of500nm×500nm. The height color bar scale is 2.0nm. A few characteristic TARpolyA RNAs are boxed in red.(b) Histogram for length (left), width (middle), and height (right). Shown in red are normal distribution fitsto the peaks for length and height. Width is fit to a gamma distribution due to its non-negativity andskewness. (c) Three dimensional smooth histogram, where the red arrow indicates the monomerdistribution. The peak corresponding to the TARpolyA RNA monomer has a mean length of 17.1 ± 0.2nm,width of 1.10 ± 0.05nm and height of 1.10 ± 0.01nm. The total number of samples was 504 and theexperiment was repeated twice. Fig.5 0.5μM GagΔP6 on negatively charged mica(-). (a) A typical AFM image with a scan size of500nm×500nm. The height color bar scale is 2.5nm. A few characteristic GagΔP6s are boxed: monomer(red), dimer (green) and tetramer (blue). (b) Histograms for length (left), width (middle), and height (right).Shown in red are normal distribution fits to the peaks. (c) Three dimensional smooth histogram, where redarrows indicate the monomer, dimer, and tetramer distributions. The mean height is 1.93 ± 0.01nm for allthree. The first peak with a mean length of 10.3 ± 0.1nm and width 6.2 ± 0.1nm corresponds to the GagΔP6monomer. The second peak with a mean length of 20.0 ± 0.2nm and width 6.2 ± 0.1nm corresponds to theGagΔP6 dimer. The third peak with a mean length of 29.0 ± 0.3nm and width 12.9 ± 0.2nm corresponds tothe GagΔP6 tetramer. The total number of samples was 858 and the experiment was repeated twice. Fig.6 Rough model of the GagΔP6 monomer based on the measured mean values of the length and width.MA domain is in red, CA domain is in yellow, and NC domain is in green.5
ΨRNA contains 109 nucleotides while TARpolyA RNA has only 104 nucleotides. It is noteworthy that thewidth histogram distribution of TARpolyA RNA is 1.10 ± 0.05nm which is the same as the height andconsistent with the cylindrical structure for TARpolyA RNA. As the length and width distribution have onlyone observable peak, it is concluded that the TARpolyA exists in solution predominantly as a monomer.This is unlike that observed with ΨRNA where 26% dimers were found.3.3 GagΔP6 Size and Morphology MeasurementsThe morphology of GagΔP6 (0.5μM), being net positive charge, was measured using the AFM on afreshly negatively charged mica(-) substrate in solution. A typical AFM image is shown in Fig.5(a). Themotivation for measuring GagΔP6 is to serve as a control before addition of RNAs and lipids. Somecharacteristics of common shapes observed were shown in red, green and blue boxes. In Fig.5(a), the AFMimage of GagΔP6 shows that most of GagΔP6 molecules have ellipsoidal shape rather than the expectedrod-like shape of the extended molecule.As shown in Fig.5(b) and Table 1, the mean height is 1.93 ± 0.01nm which is consistent with theexpectation that the diameter of Gag is around 2~3nm as reported in Ref. [4]. The length and width of theobserved images are plotted in Fig.5(b). As observed three peaks were observed in the length distributionbut only two peaks in the width distribution. Normal distributions were fit to the length histogram to findpeak mean values of 10.3 ± 0.1nm, 20.0 ± 0.2nm and 29.0 ± 0.3nm. Similarly the mean values of the twowidth distributions were found to be 6.2 ± 0.1nm and 12.9 ± 0.2nm respectively.It might seem counterintuitive to have three peaks for length while having only two peaks for thewidth. To have more insight, a three dimensional smooth histogram of the same length-width data wasplotted as shown in Fig.5(c). The shortest length and width distribution would correspond to that of themonomer and the distribution with the next larger length would be expected to correspond to that of thedimer. From Fig.5(c) the monomer and dimer have the same width. Thus in the width distributionhistogram they are represented together and correspond to the first peak at 6.2 ± 0.1nm. The statisticalanalysis of the population also confirmed this assumption as the first peak of the width histogram is thesum of monomer and dimer populations in the two peaks of the length histogram as given in Table 1. Thisanalysis was done by fitting each length and width peak to a normal distribution. From the lengthdistribution, the percentages of monomer, dimer, and tetramer are 59%, 35%, and 6%, respectively. Thepopulation observed in the monomer and dimer from the length adds up to that observed in the first peak ofthe width distribution.Previous studies using hydrodynamic and neutron scattering measurements showed HIV GagΔP6 issupposed to adopt a compact conformation such that MA and NC domains of GagΔP6 are close to eachother even though both of them are positively charged [3-5, 57]. The hydrodynamic radius, R h , of GagΔP6given by three different hydrodynamic tests are 3.6nm, 3.8nm, and 4.1nm, respectively. The radius ofgyraton, R g , of GagΔP6 is best estimated to be 3.4nm from small angle neutron scattering (SANS). The R g of GagΔP6 when it is a 25nm straight rod is supposed to be 7.2nm [57]. The average R g of GagΔP6 insolution measured by SANS is also a monotonically increasing function of the GagΔP6 concentration, with maximum of R g = 5nm at extremely high concentration, which means GagΔP6 molecules are inmonomer-dimer equilibrium [57].Based on the AFM studies presented above, we can project approximate confirmations based on thevarious lengths and widths of the populations observed. For the case of the monomer given a length of10.3nm and width of 6.2nm a potential “C” like shape can be conjectured. This is consistent with the MAdomain being close to the NC domain as has been reported earlier for monomers in solution [4]. Aschematic of a potential monomer structure is shown in Fig.6. Based on the AFM measured dimensions of alength of 20.0 nm and width of 6.2nm (same as monomer) a model of the GagΔP6 dimer would have twomonomers connecting back to back through their CA-CA interaction [4]. The CA-CA interaction leadingto dimerization has been reported in the literature [4]. The last population size distribution of GagΔP6 inFig.5 (c) has a length of 29.0nm and width is 12.9nm. This population from gel electrophoresis correspondsto that of a tetramer (see Supplemental Material). From the measured size of the tetramers here, they couldbe potentially formed by the interaction of two dimers.3.4 GagΔP6-ΨRNA Complex Size, Morphology and InteractionAFM measurements of the GagΔP6-ΨRNA (0.5μM : 0.5μM) complex were done on negativelycharged mica(-). The motivation for measuring GagΔP6-ΨRNA complex is to investigate the effect ofaddition of specific ΨRNA to GagΔP6. According to current models the NC domain binds with ΨRNA andthis interaction is specific and is a critical step in the formation of HIV [22]. Fig.7(a) is a typical AFMimage of GagΔP6-ΨRNA complex. As shown in Fig.7(b), the mean height of GagΔP6-ΨRNA complex is1.90 ± 0.01nm which is roughly the same as the height of just GagΔP6 discussed earlier. This height isconsistent with expectation given that the height of GagΔP6 and ΨRNA are 1.93nm and 1.10nm,respectively. The length and width histograms of the complexes observed are shown in Fig.7(b). Similar toGagΔP6, GagΔP6-ΨRNA also has three peaks for length and two peaks for width. Normal distributionswere fit to the length histogram to find peak mean values of 10.6 ± 0.2nm, 22.4 ± 0.1nm and 31.2 ± 0.2nm.Similarly the mean values of the two width distributions were found to be 6.8 ± 0.1nm and 13.8 ± 0.1nmrespectively. The size of the length and widths are slightly larger than that found with only GagΔP6. This isconsistent with attachment of ΨRNA of around 1nm to the GagΔP6.To understand the role of the ΨRNA addition the three-dimensional population distribution of thelength and width as shown in Fig.7(c) was analyzed. As with GagΔP6, three peaks corresponding tomonomer, dimer and tetramer complexes were observed. However the population distributions of themonomer, dimer and tetramer are different for the GagΔP6- ΨRNA complex. Here we observed 35%monomer, 49% dimers and 16% tetramers. In contrast with only GagΔP6 we observed 59% monomers 35%dimers and 6% tetramers. The monomer decreased by 24%, the dimer increased by 14% and the tetramerincreased by 10%. Thus the addition of ΨRNA promotes multimerization of the GagΔP6 leading to higherpopulations of dimers and tetramers.The length of monomer remained roughly the same as prior to the addition of ΨRNA. The lengths ofdimer and tetramer were increased by about 2nm. The width increased by about 0.5nm~1nm for the monomer, dimer, and tetramer. This is consistent with the addition of ΨRNA of around 1nm to thisdimension. The most important conclusion of the effect of the addition of ΨRNA to GagΔP6 is that ΨRNAcan bind with GagΔP6 and facilitate GagΔP6 multimerization given the increases in percentages of dimerand tetramer population.3.5 GagΔP6-TARpolyA RNA Complex Size, Morphology and InteractionGagΔP6-TARpolyA RNA (0.5μM : 0.5μM) complex was measured on a negatively chargedmica(-) surface. The motivation for measuring GagΔP6-TARpolyA RNA complex is to verify the effect ofthe addition of non-specific TARpolyA RNA to GagΔP6 and compare it to the addition of specific ΨRNAdiscussed above. Fig.8(a) is a typical AFM image of GagΔP6-TARpolyA RNA complex. As shown inFig.8(b), the mean height of GagΔP6-ΨRNA complex is 1.93 ± 0.01nm.To understand the role of the TARPolyA RNA interaction with GagΔP6 the three dimensionalpopulation distribution of the length and width as shown in Fig.8(c) was analyzed. As with GagΔP6, threepeaks corresponding to monomer, dimer and tetramer complexes were observed. However in contrast to thecase of the GagΔP6-ΨRNA complex in Fig.7(c) the population distribution of monomer, dimer andtetramer are very similar to that of GagΔP6 alone observed in Fig.5(c). The sizes of GagΔP6-TARpolyARNA complex monomer, dimer and tetramer were found slightly larger than that of GagΔP6 alone. Thusthe addition of TARPolyA RNA might interact with GagΔP6 but does not promotes multimerization of theGagΔP6 as observed with ΨRNA. These experiments were repeated and the results were alwaysreproducible. Webb et al. reported HIV Gag can bind with both ΨRNA and TARpolyA RNA but withdistinct binding mechanisms [22]. They proposed that HIV GagΔP6 binds with TARpolyA RNA throughboth MA and NC domains whereas ΨRNA binds only through NC domain and leaves MA domain free tolater interact with the lipid membrane. Other studies also showed that HIV GagΔP6 can bind with bothΨRNA and non-Ψ RNAs but the selective binding with ΨRNA is more energetically favorable than othernon-Ψ RNAs for HIV virus assembly [58-60]. The role of the RNA in multimerization of the Gag was notaddressed in these studies. The conclusion based on our data is that it is highly likely that HIV GagΔP6interacts with ΨRNA and TARpolyA RNA through different mechanisms such that ΨRNA facilitatesGagΔP6 multimerization while TARpolyA RNA does not.3.6 GagΔP6-PI(4,5)P2 Complex Size, Morphology and InteractionThe GagΔP6-PI(4,5)P2 (0.5μM : 0.5μM) complex was measured on negatively charged mica(-). The Gagand lipid were diluted to 1μM before mixing. Then, equal amounts of Gag and lipid were mixed. The AFMmeasurement was performed after the mixture was incubated for 3 hours.. A typical AFM image is shownin Fig.9(a). The motivation for measuring GagΔP6-PI(4,5)P2 complex is to explore the effect of addition oflipid PI(4,5)P2 to GagΔP6. The current understanding is that both MA and NC domains of Gag can bind toPI(4,5)P2 through electrostatic forces.As shown in Fig.9(b), the mean height of GagΔP6-PI(4,5)P2 complex is 1.93 ± 0.01nm that isroughly the same as the height of just GagΔP6. This height is consistent with expectation given the smallsize of PI(4,5)P2 in comparison to GagΔP6. The length and width histograms of the complexes observed are shown in Fig.9(b). Similar to GagΔP6, GagΔP6-PI(4,5)P2 also has three peaks for length and two peaksfor width. Normal distributions were fit to the length histogram to find peak mean values of 11.2 ± 0.1nm,21.1 ± 0.1nm and 30.4 ± 0.2nm. Similarly the mean values of the two width distributions were found to be6.7 ± 0.1nm and 14.0 ± 0.2nm respectively. The size of the length and widths are slightly larger than thatwith only GagΔP6. In comparison to the GagΔP6-ΨRNA complex the GagΔP6-PI(4,5)P2 complex hasmonomers of slighter larger length while the dimers and tetramers are of slightly smaller length.To understand the role of PI(4,5)P2 addition to GagΔP6 the three dimensional populationdistribution of the length and width as shown in Fig.9(c) was analyzed. As with GagΔP6, three peakscorresponding to monomer, dimer and tetramer complexes were observed. However the populationdistribution of monomer, dimer and tetramer are different from that observed with GagΔP6 alone or for theGagΔP6- ΨRNA complex. Here we observed 47% monomer, 41% dimers and 12% tetramers. In contrastwith only GagΔP6 we observed 59% monomers 35% dimers and 6% tetramers and for the GagΔP6- ΨRNAcomplex we observed 35% monomer, 49% dimers and 16% tetramers. Thus the addition of PI(4,5)P2promotes multimerization of the GagΔP6. The increases in percentages of dimer and tetramer indicate thatPI(4,5)P2 can bind with GagΔP6 as reported in other studies using confocal microscopy, nuclear magneticresonance and equilibrium flotation assay [35, 40, 61-65]. But the increase in GagΔP6 multimerization withthe addition of PI(4,5)P2 is less than that observed with ΨRNA. The length and width of monomer, dimer,tetramer all increased by about 1nm. This is probably because of the size of PI(4,5)P2 attached to the endsof both MA and CA domains of HIV GagΔP6.3.7 PI(4,5)P2-ΨRNA-GagΔP6 Complex Size, Morphology and InteractionThe PI(4,5)P2-ΨRNA-GagΔP6 (0.5μM : 0.5μM : 0.5μM) complex was measured on a negativelycharged mica(-). According to the prevailing model the lipid interacts with the MA domain ofGagΔP6-ΨRNA complex leading to a conformational change [22]. In these experiments to study thePI(4,5)P2-ΨRNA-GagΔP6 complex, PI(4,5)P2 and ΨRNA were first mixed together in solution followedby the addition of GagΔP6. This mixture was then used to measure the size and size distribution using theAFM. Fig.10(a) is a typical AFM image of PI(4,5)P2-ΨRNA-GagΔP6 complex. As shown in Fig.10(b),the mean height of PI(4,5)P2-ΨRNA-GagΔP6 complex is 1.91 ± 0.01nm that is roughly the same as theheight of just GagΔP6. This height is consistent with expectation given that the height of GagΔP6 andΨRNA-GagΔP6 discussed earlier. The height of PI(4,5)P2 is much smaller by comparison. The length andwidth histograms of the complexes observed are shown in Fig.10(b). In contrast to GagΔP6,GagΔP6-ΨRNA and the GagΔP6- PI(4,5)P2 complexes there are only two peaks for length and two peaksfor width. Normal distributions were fit to the length histogram to find peak mean values of 10.9 ± 0.3nm,and 23.8 ± 0.2nm. Similarly the mean values of the two width distributions were found to be 7.4 ± 0.2nmand 22.1 ± 0.2nm respectively. In comparison to the binary complexes studied earlier the monomer lengthis approximately similar, while the monomer width is slightly larger. For the dimer, the length is almost60% larger than that of GagΔP6-ΨRNA and the GagΔP6- PI(4,5)P2 binary complexes . To understand the PI(4,5)P2-ΨRNA-GagΔP6 complex the three dimensional populationdistribution of the length and width as shown in Fig.10(c) was analyzed. In contrast to all others above,here the dimer and tetramer have the same length. The width of the tetramer is much larger than allprevious cases, increasing from 13-14nm to 22nm. Thus in the PI(4,5)P2-ΨRNA-GagΔP6 complex theGagΔP6 undergoes a dramatic conformational change. This is consistent with the GagΔP6 MA and NCdomains moving further away from each other and taking on a rod-like confirmation. From the populationdistribution in Fig.10(c), we observed 17% monomer, 51% dimers and 32% tetramers. In contrast, withonly GagΔP6 were we observed 59% monomers 35% dimers and 6% tetramers, the multimerization hasincreased considerably. Comparing to the populations in the binary mixtures of ΨRNA-GagΔP6 andPI(4,5)P2-GagΔP6 were we observed 35% and 47% monomer, 49% and 41% dimers and 16% and 12%tetramers respectively, here in particular we observe much higher percentage of tetramers. Thus theaddition of PI(4,5)P2 along with ΨRNA promotes not only the dimerization of the GagΔP6 but inparticular the multimerization to tetramers and higher order complexes. In addition, the significantchange of the size indicates that GagΔP6 undergoes some conformational changes when both ΨRNA andPI(4,5)P2 are present [38, 66]. Based on the sizes measured a schematic of the dimer and tetramer structureis proposed in Fig.11. From the size of the tetramers the spacing between the GagΔP6 molecules is around7 nm close to the value of 8nm reported in studies using cET [18,19]. The substantial increases in thepercentages of dimer and tetramer indicate that both ΨRNA and PI(4,5)P2 can bind with GagΔP6 andcollectively facilitate HIV GagΔP6 assembly as reported in other studies [57, 63].4 ConclusionThe AFM technique was utilized to study the morphology of GagΔP6 (0.5μM), ΨRNA (0.5μM), andtheir binding complexes with the lipid PI(4,5)P2 in HEPES buffer with 0.1 Å vertical and 1nm lateralresolution. For the calibration 2nm diameter Au spheres were used. TARpolyA RNA was used as a negativeRNA control. The morphology of specific complexes GagΔP6-ΨRNA (0.5μM : 0.5μM),GagΔP6-TARpolyA RNA (0.5μM : 0.5μM), GagΔP6-PI(4,5)P2 (0.5μM : 0.5μM) and PI(4,5)P2-ΨRNA-GagΔP6 (0.5μM : 0.5μM : 0.5μM) were studied. They were imaged on either positively charged ornegatively charged mica substrates depending on the net charges carried by the respective materials. Thesize and morphology of both ΨRNA and TARpolyA RNA measured was used to validate the technique incomparison with literature. For the ΨRNA, from the measured size two distinct populations correspondingto monomers and dimers were observed. In the case of TARpoly A RNA only the monomer population wasfound for the concentrations studied. The morphology of GagΔP6, being net positively charged, wasmeasured using the AFM on a freshly cleaved negatively charged mica(-) substrate in HEPES solution witha low salt concentration of 50mM NaCl. Three distinct size populations were found. They were found tocorrespond to 59% monomer, 35% dimer and 6% tetramer form of GagΔP6. The presence of the multimerswas confirmed by gel electrophoresis. The addition of ΨRNA to 0.5μM GagΔP6 was observed to promotemultimerization of the GagΔP6 leading to higher populations of dimers (14% increase) and tetramers (10%increase). The small change in size of the complexes confirmed the binding of the ΨRNA to GagΔP6.The addition TARPolyA RNA to GagΔP6 did not modify the GagΔP6 population distribution of monomers,dimers and tetramers. The interaction of PI(4,5)P2 with GagΔP6 complex was next measured onnegatively charged mica(-). From the population distribution of the monomers (47%), dimers (41%) and tetramers (12%), it was concluded that PI(4,5)P2 promotes multimerization of GagΔP6 but not to the extentobserved with ΨRNA.The PI(4,5)P2-ΨRNA-GagΔP6 ternary complex was next studied using a negatively chargedmica(-) surface. Both the population and size distribution of the GagΔP6 was completely different fromthat of the GagΔP6-ΨRNA and the GagΔP6- PI(4,5)P2 binary complexes. The population distribution ofthe monomers (17%), dimers (51%) and tetramers (32%) was significantly different. Given the largefraction of dimers and tetramers it was concluded that the presence of both PI(4,5)P2 and ΨRNA promotesextensive multimerization of GagΔP6. In addition, the significant change of size indicates that GagΔP6undergoes a conformational change to a 23.8nm rod like shape when both ΨRNA and PI(4,5)P2 are present.From the size of the tetramers the spacing between the GagΔP6 molecules is around 7 nm which isconsistent with studies using cET in Ref [18, 19]. The substantial increases in the percentages of dimerand tetramer indicate that both ΨRNA and PI(4,5)P2 can bind with GagΔP6 and collectively facilitate HIVGagΔP6 assembly as reported using other techniques.Author Contributionsdesigned research: S. C., U. M.performed research: S. C.contributed analytic tools: S. C., M. L.analyzed data: S. C., J. X., A.L.N. R., R. Z., S.S. G., U. M.wrote the manuscript: S. C., U. M.AcknowledgementsWe would like to acknowledge discussions with Karin Musier-Forsyth, Erick D. Olson and Ioulia Rouzina.We would like to thank Erick Olson and Karin Musier-Forsyth for providing the in vitro transcribed RNAs.We also acknowledge discussions with Alan Rein and thank him and S.A.K. Datta for providing theGagΔP6 molecules.5 References[1] Ganser-Pornillos, B.K., Yeager, M., Pornillos, O.: Assembly and architecture of HIV. Adv. Exp.Med. Biol., 726, 441-465 (2012)[2] Rein, A., Datta, S.A.K., Jones, C.P., Musier-Forsyth, K.: Diverse interactions of retroviral Gagproteins with RNAs. Trends. Biochem. Sci., 36, 373-380 (2011)[3] Briggs, J.A.G., Simon, M.N., Gross, I., Kräusslich, H.G., Fuller, S.D., Vogt, V.M., Johnson, M.C.: Thestoichiometry of Gag protein in HIV-1. Nat. Struct. Mol. Biol., 11, 672–675 (2004)[4] Datta, S.A.K., Curtis, J.E., Ratcliff, W., Clark, P.K., Crist, R.M., Lebowitz, J., Krueger, S., Rein A.:Conformation of the HIV-1 Gag protein in solution. J. Mol. Biol., 365, 812-824 (2007)[5] Paul M. S. and Beatrice H. H.: Origins of HIV and the AIDS Pandemic. Cold Spring Harb Perspect Med.1, 1, a006841 (2011)[6] Sundquist, W.I. Kräusslich, H.G.: HIV-1 Assembly, budding, and maturation. CSH Perspect. Med., 2, 7,a006924 (2012)[7] Freed E.O.: HIV-1 assembly, release and maturation. Nat. Rev. Microbiol., 13, 484–496 (2015)[8] Bell, N.M., Lever, A.M.L.: HIV Gag polyprotein: processing and early viral particle assembly. TrendsMicrobiol., 21, 136-144 (2013)[9] Briggs, J.A.G., Kräusslich, H.G.: The Molecular Architecture of HIV. J. Mol. Biol., 410, 491-500(2011)[10] Pettit, S.C., Sheng, N., Tritch, R., Erickson-Viitanen, S., Swanstrom, R.: The regulation of sequentialprocessing of HIV-1 Gag by the viral protease. Adv. Exp. Med. Biol., 436, 15-25 (1998)[11] Massiah, M.A., Starich, M.R., Paschall, C., Summers, M.F., Christensen, A.M., Sundquist, W.I.:Three-dimensional structure of the human immunodeficiency virus type 1 matrix protein, J. Mol. Biol., 244,198-223 (1994)[12] Chukkapalli, V., Ono, A.: Molecular determinants that regulate plasma membrane association of HIV-1Gag. J. Mol. Biol., 410, 512-524 (2011)[13] Hill, C.P., Worthylake, D., Bancroft, D.P., Christensen, A.M., Sundquist, W.I.: Crystal structures of thetrimeric human immunodeficiency virus type 1 matrix protein: implications for membrane association andassembly. PNAS, 93, 7, 3099-3104 (1996)[14] Göttlinger, H.G., Sodroski, J.G., Haseltine, W.A.: Role of capsid precursor processing andmyristoylation in morphogenesis and infectivity of human immunodeficiency virus type 1. PNAS, 86, 15,5781-5785 (1989)[15] Spearman, P., Wang, J.J., Heyden, N.V., Ratner, L.: Identification of human immunodeficiency virustype 1 Gag protein domains essential to membrane binding and particle assembly. J. Virol., 68, 5,3232-3242 (1994)[16] Bharat, T.A.M., Davey, N.E., Ulbrich, P., Riches, J.D., de Marco, A., Rumlova, M., Sachse, C., Ruml,T., Briggs, J.A. G. (2012). Structure of the immature retroviral capsid at 8 Å resolution by cryo-electronmicroscopy. Nature, 487, 385–389.[17] Kelly, B.N., Howard, B.R., Wang, H., Robinson, H., Sundquist, W.I., Hill, C.P.: Implications for viralcapsid assembly from crystal structures of HIV-1 Gag(1-278) and CA(N)(133-278). Biochemistry, 45, 38,11257–11266 (2006)[18] Wright, E.R., Schooler, J.B., Ding, H.J., Kieffer, C., Fillmore, C., Sundquist, W.I., Jensen, G.J.: Electron cryotomography of immature HIV-1 virions reveals the structure of the CA and SP1 Gag shells.EMBO J., 26, 2218-2226 (2007)[19] Briggs, J.A., Riches, J.D., Glass, B., Bartonova, V., Zanetti, G., Kräusslich, H.G.: Structure andassembly of immature HIV. PNAS, 106, 27, 11090-11095 (2009)[20] De Guzman, R.N., Wu, Z.R., Stalling, C.C., Pappalardo, L., Borer P.N., Summers, M.F.: Structure ofthe HIV-1 nucleocapsid protein bound to the SL3 psi-RNA recognition element. Science, 16, 279,5349,384-388 (1998)[21] Godet, J., Boudier, C., Humbert, N., Ivanyi-Nagy, R., Darlix, J.L., Mély, Y.: Comparative nucleicacid chaperone properties of the nucleocapsid protein NCp7 and Tat protein of HIV-1. Virus Res., 169, 2,349-360 (2012)[22] Webb, J.A., Jones, C.P., Parent, L.J., Rouzina, I., Musier-Forsyth, K.: Distinct binding interactions ofHIV-1 Gag to Psi and non-Psi RNAs: implications for viral genomic RNA packaging. RNA, 19, 8,1078-1088 (2013)[23] Kafaie, J., Song, R., Abrahamyan, L., Mouland, A.J., Laughrea, M.: Mapping of nucleocapsid residuesimportant for HIV-1 genomic RNA dimerization and packaging. Virology, 375, 2, 592-610 (2008)[24] Thomas, J.A., Gorelick, R.J.: Nucleocapsid protein function in early infection processes. Virus Res.,134, 1-2, 39-63 (2008)[25] Watts, J.M., Dang, K.K., Gorelick, R.J., Leonard, C.W., Bess, Jr J.W., Swanstrom, R., Burch, C.L.,Weeks, K.M.: Architecture and secondary structure of an entire HIV-1 RNA genome. Nature, 460, 711–716(2009)[26] Jones, C.P., Cantara, W.A., Olson, E.D., Musier-Forsyth, K.: Small-angle X-ray scattering-derivedstructure of the HIV-1 5' UTR reveals 3D tRNA mimicry. PNAS, 111, 9, 3395-3400 (2014)[27] Wilkinson, K.A., Gorelick, R.J., Vasa, S.M., Guex, N., Rein, A., Mathews, D.H., Giddings, M.C.,Weeks, K.M.: High-throughput SHAPE analysis reveals structures in HIV-1 genomic RNA stronglyconserved across distinct biological states. PLoS Biol. 6, 4, e96 (2008)[28] Andersen, E.S., Contera, S.A., Knudsen, B., Damgaard, C.K., Besenbacher, F., Kjems, J.: Role of thetrans-activation response element in dimerization of HIV-1 RNA. J. Biol. Chem., 279, 21, 22243-22249(2004)[29] Frankel, A.D., Young, J.A.: HIV-1: fifteen proteins and an RNA. Annu. Rev. Biochem., 67, 1-25(1998)[30] Moore, M.D., Hu, W.S.: HIV-1 RNA dimerization: It takes two to tango. AIDS Rev., 11, 2, 91-102(2009)[31] Skripkin, E., Paillart, J.C., Marquet, R., Ehresmann, B., Ehresmann, C.: Identification of the primarysite of the human immunodeficiency virus type 1 RNA dimerization in vitro. PNAS, 91, 11, 4945-4949(1994)[32] Lu, K., Heng, X., Summers, M.F.: Structural determinants and mechanism of HIV-1 genomepackaging. J. Mol. Biol., 410, 2, 609-633 (2011)[33] Russell, R.S., Hu, J., Bériault, V., Mouland, A.J., Laughrea, M., Kleiman, L., Wainberg, M.A., Liang,C.: Sequences downstream of the 5' splice donor site are required for both packaging and dimerization ofhuman immunodeficiency virus type 1 RNA. J. Virol., 77, 1, 84-96 (2003)[34] Clever, J.L., Parslow, T.G.: Mutant human immunodeficiency virus type 1 genomes with defects in RNA dimerization or encapsidation. J. Virol., 71, 5, 3407-3414 (1997)[35] 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. PNAS, 101, 41, 14889-14894 (2004)[36] Zhou, W., Parent, L.J., Wills, J.W., Resh, M.D.: Identification of a membrane-binding domain withinthe amino-terminal region of human immunodeficiency virus type 1 Gag protein which interacts withacidic phospholipids. J. Virol., 68, 4 2556-2569 (1994)[37] Zhou, W., Resh, M.D.: Differential membrane binding of the human immunodeficiency virus type 1matrix protein. J. Virol., 70, 12, 8540-8548 (1996)[38] Spearman, P., Horton, R., Ratner, L., Kuli-Zade, I.: Membrane binding of human immunodeficiencyvirus type 1 matrix protein in vivo supports a conformational myristyl switch mechanism. J. Virol., 71, 9,6582-6592 (1997)[39] Ono, A., Freed, E.O.: Binding of Human Immunodeficiency Virus Type 1 Gag to Membrane: Role ofthe Matrix Amino Terminus. J. Virol., 73, 5, 4136-4144 (1999)[40] Saad, J.S., Miller, J., Tai, J., Kim, A., Ghanam, R.H., Summers, M.F.: Structural basis for targetingHIV-1 Gag proteins to the plasma membrane for virus assembly. PNAS, 103, 30, 11364-11369 (2006)[41] Rojas, O.J.: Adsorption of polyelectrolytes on mica. Ency. Surf. Coll. Sci., 517-535 (2002)[42] Müller, D.J., Amrein, M., Engel ,A.: Adsorption of biological molecules to a solid support for scanningprobe microscopy. J. Struct. Biol., 119, 2, 172-188 (1997)[43] Müllerab, D.J., Engela, A., Amreinc, M.: Preparation techniques for the observation of nativebiological systems with the atomic force microscope. Biosens. Bioelectron., 12, 8, 867-877 (1997)[44] Raposo, M., Ferreira, Q., Ribeiro, P. A.: A guide for Atomic Force Microscopy Analysis ofSoft-Condensed Matter. Mod. Res. Educ. Top. Microsc., 758-769 . (2007)[45] Pfreundschuh, M., Martinez-Martin, D., Mulvihill, E., Wegmann, S., Muller, D.J.: Multiparametrichigh-resolution imaging of native proteins by force-distance curve–based AFM. Nat. Protoc., 9, 1113–1130(2014)[46] Lyubchenko, Y.L., Shlyakhtenko, L.S., Ando, T.: Imaging of nucleic acids with atomic forcemicroscopy. Methods, 54, 2, 274-283 (2011)[47] Pan Y., Sun Z., Maiti A., Kanai T., Matsuo H., Li M., Harris R.S., Shlyakhtenko, L.S., Lyubchenko,Y.L.: Nanoscale Characterization of Interaction of APOBEC3G with RNA. Biochemistry, 56 (10),1473-1481 (2017)[48] Ares, P., Fuentes-Perez, M.E., Herrero-Galán, E., Valpuesta, J.M., Gil, A., Gomez-Herrero, J.,Moreno-Herrero, F.: High resolution atomic force microscopy of double-stranded RNA. Nanoscale, 8, 23,11818-11826 (2016)[49] Pastré, D., Piétrement, O., Fusil, S., Landousy, F., Jeusset, J., David, M.O., Hamon, L., Le Cam, E.,Zozime, A.: Adsorption of DNA to mica mediated by divalent counterions: a theoretical and experimentalstudy. Biophys. J., 85, 4, 2507-2518 (2003)[50] Kienberger, F., Costa, L.T., Zhu, R., Kada, G., Reithmayer, M., Chtcheglova, L., Rankl, C., Pacheco,A.B., Thalhammer, S., Pastushenko, V., Heckl, W.M., Blaas, D., Hinterdorfer, P.: Dynamic forcemicroscopy imaging of plasmid DNA and viral RNA. Biomaterials, 28, 15, 2403-2411 (2007)[51] Valbuenaa, A., Mateu, M.G.: Quantification and modification of the equilibrium dynamics andmechanics of a viral capsid lattice self-assembled as a protein nanocoating. Nanoscale. 7(36), 14953-14964 (2015)[52] Bussiek, M., Schöne, A. Nellen, W.: Atomic force microscopy imaging and force spectroscopy ofRNA. Handbook of RNA Biochemistry: Second, Completely Revised and Enlarged Edition, 527-546(2014)[53] Hansma, H.G., Oroudjev, E., Baudrey, S., Jaeger, L.: TectoRNA and 'kissing-loop' RNA: atomic forcemicroscopy of self-assembling RNA structures. J. Microsc., 212, 3, 273-279 (2003)[54] Watson, J.D. Crick, F.H.C.: Molecular structure of nucleic acids: a structure for deoxyribose nucleicacid. Nature, 171, 737-738 (1953)[55] Tanaka Y., Fujii S., Hiroaki H., Sakata T., Tanaka T., Uesugi S., Tomita K., Kyogoku Y.: A'-form RNAdouble helix in the single crystal structure of r(UGAGCUUCGGCUC). Nucleic Acids Res, 27, 4, 949–955(1999)[56] Arnott S., Hukins D.W., Dover S.D.: Optimised parameters for RNA double-helices. Biochem BiophysRes Commun. 48, 6,1392-1399 (1972)[57] Datta, S.A., Heinrich, F., Raghunandan, S., Krueger, S., Curtis, J.E., Rein, A., Nanda, H.: HIV-1Gag extension: conformational changes require simultaneous interaction with membrane and nucleic acid. J.Mol. Biol., 406, 2, 205-214 (2011)[58] Comas-Garcia, M, Davis, S.R., Rein, A.: On the selective packaging of genomic RNA by HIV-1.Viruses, 8, 9, 246 (2016)[59] Didierlaurent, L., Racine, P.J., Houzet, L., Chamontin, C., Berkhout, B., Mougel, M.: Role of HIV-1RNA and protein determinants for the selective packaging of spliced and unspliced viral RNA and host U6and 7SL RNA in virus particles. Nucleic Acids Res., 39, 20, 8915-8927 (2011)[60] Houzet, L., Paillart, J.C., Smagulova, F., Maurel, S., Morichaud, Z., Marquet, R., Mougel, M.: HIVcontrols the selective packaging of genomic, spliced viral and cellular RNAs into virions through differentmechanisms. Nucleic Acids Res., 35, 8, 2695-2704 (2007)[61] Carlson, L.A. Hurley J.H.: In vitro reconstitution of the ordered assembly of the endosomal sortingcomplex required for transport at membrane-bound HIV-1 Gag clusters. PNAS, 109, 42, 16928-16933(2012)[62] Chukkapalli, V., Hogue, I.B., Boyko, V., Hu, W.S., Ono A.: Interaction between the humanimmunodeficiency virus type 1 Gag matrix domain and phosphatidylinositol-(4,5)-bisphosphate is essentialfor efficient gag membrane binding. J. Virol., 82, 5, 2405-2417 (2008)[63] Shkriabai, N., Datta, S.A., Zhao, Z., Hess, S., Rein, A., Kvaratskhelia, M.: Interactions of HIV-1 Gagwith assembly cofactors. Biochemistry, 45, 13, 4077-4083 (2006)[64] Alfadhli, A., Barklis, R.L., Barklis, E.: HIV-1 matrix organizes as a hexamer of trimers on membranescontaining phosphatidylinositol-(4,5)-bisphosphate. Virology, 387, 2, 466-472 (2009)[65] Freed, E.O.: HIV-1 Gag: Flipped out for PI(4,5)P2. PNAS, 103, 30, 11101-11102 (2006)[66] Ono, A., Demirov, D., Freed, E.O.: Relationship between human immunodeficiency virus type 1 Gagmultimerization and membrane binding. J. Virol., 74, 11, 5142-5150 (2000) Fig.1 HIV-1 genomic RNA partial sequence constructs used in this work. (a) TARpolyA RNA. (b) ΨRNA. Fig.2 AFM tip resolution calibration with 2nm diameter Au Sphere. (a) A typical AFM image of 2nm Ausphere on mica in tapping mode in liquid. The scan size is 250nm×250nm. The height color bar scale is2.5nm. (b) Histogram of 2nm diameter Au sphere for the measured size on the plane of the substrate (top)and the measured height (bottom). Shown in red are normal distribution fits to the peaks. The meanmeasured size on the plane of the substrate is 7.56 ± 0.09 nm, and the mean height is 2.10 ± 0.02nm. Theheight is consistent with the sphere diameter. The size in the plane of the substrate reflects the role of thetip size. Please see text and supplemental materials section for more details. The total number of sampleswas 419 and the experiment was repeated twice. Fig.3 0.5μM ΨRNA on positively charged mica(+). (a) A typical AFM image with a scan size of500nm×500nm. The height color bar scale is 2.0 nm. A few characteristic ΨRNAs are shown in boxes:monomer (red) and dimer (green). (b) Histogram for length (left), width (middle), and height (right).Shown in red are normal distribution fits to the peaks. (c) Three dimensional smooth histogram, where redarrows indicate monomer and dimer. The mean height is 1.10 ± 0.01nm. The first peak with a mean lengthof 17.9 ± 0.2nm and width 1.01 ± 0.02nm corresponds to the ΨRNA monomer. The second peak with amean length of 34.6 ± 0.3nm and width 3.8 ± 0.1nm corresponds to the ΨRNA dimer. The total number ofsamples was 551 and the experiment was repeated twice. Fig.4 0.5μM TARpolyA RNA on positively charged mica(+). (a) AFM image with a scan size of500nm×500nm. The height color bar scale is 2.0nm. A few characteristic TARpolyA RNAs are boxed in red.(b) Histogram for length (left), width (middle), and height (right). Shown in red are normal distribution fitsto the peaks for length and height. Width is fit to a gamma distribution due to its non-negativity andskewness. (c) Three dimensional smooth histogram, where the red arrow indicates the monomerdistribution. The peak corresponding to the TARpolyA RNA monomer has a mean length of 17.1 ± 0.2nm,width of 1.10 ± 0.05nm and height of 1.10 ± 0.01nm. The total number of samples was 504 and theexperiment was repeated twice. Fig.5 0.5μM GagΔP6 on negatively charged mica(-). (a) A typical AFM image with a scan size of500nm×500nm. The height color bar scale is 2.5nm. A few characteristic GagΔP6s are boxed: monomer(red), dimer (green) and tetramer (blue). (b) Histograms for length (left), width (middle), and height (right).Shown in red are normal distribution fits to the peaks. (c) Three dimensional smooth histogram, where redarrows indicate the monomer, dimer, and tetramer distributions. The mean height is 1.93 ± 0.01nm for allthree. The first peak with a mean length of 10.3 ± 0.1nm and width 6.2 ± 0.1nm corresponds to the GagΔP6monomer. The second peak with a mean length of 20.0 ± 0.2nm and width 6.2 ± 0.1nm corresponds to theGagΔP6 dimer. The third peak with a mean length of 29.0 ± 0.3nm and width 12.9 ± 0.2nm corresponds tothe GagΔP6 tetramer. The total number of samples was 858 and the experiment was repeated twice. Fig.6 Rough model of the GagΔP6 monomer based on the measured mean values of the length and width.MA domain is in red, CA domain is in yellow, and NC domain is in green.5 Fig.7 The mixture of GagΔP6-ΨRNA (0.5μM : 0.5μM) complex on negatively charged mica(-). (a) Atypical AFM image with a scan size of 500 nm×500 nm. The height color bar scale is 2.5nm. A fewcharacteristic GagΔP6-ΨRNA complexes are boxed: monomer (red), dimer (green) and tetramer (blue). (b)Histogram for length (left), width (middle), and height (right). Shown in red are normal distribution fits tothe peaks. (c) Three dimensional smooth histogram, where red arrows indicate monomer, dimer, andtetramer. The mean height is 1.90 ± 0.01nm for all three complexes. The first peak with a mean length of10.6 ± 0.2nm and width 6.8 ± 0.1nm corresponds to the GagΔP6-ΨRNA monomer. The second peak with amean length of 22.4 ± 0.1nm and width 6.8 ± 0.1nm corresponds to the GagΔP6-ΨRNA dimer. The thirdpeak with a mean length of 31.2 ± 0.2nm and width 13.8 ± 0.1nm corresponds to the GagΔP6-ΨRNAtetramer. The total number of samples was 895 and the experiment was repeated twice. Fig.8 Mixture of GagΔP6-TARpolyA RNA (0.5μM : 0.5μM) complex on negatively charged mica(-). (a) Atypical AFM image with a scan size of 500nm×500nm. The height color bar scale is 2.5nm. A fewcharacteristics TARpolyA RNA complexes are boxed: monomer (red), dimer (green) and tetramer (blue). (b)Histogram for length (left), width (middle), and height (right). Shown in red are normal distribution fits tothe peaks. (c) Three dimensional smooth histogram, where monomer, dimer, and tetramer are indicated byred arrows. The mean height is 1.93 ± 0.01nm for all three complexes. The first peak with a mean length of10.8 ± 0.1nm and width 6.2 ± 0.1nm corresponds to the GagΔP6-TARpolyA RNA monomer. The secondpeak with a mean length of 20.7 ± 0.2nm and width 6.2 ± 0.1nm corresponds to the GagΔP6-TARpolyARNA dimer. The third peak with a mean length of 29.9 ± 0.3nm and width 13.6 ± 0.2nm corresponds to theGagΔP6-TARpolyA RNA tetramer. The total number of samples was 766 and the experiment was repeatedtwice. Fig.9 Mixture of GagΔP6-PI(4,5)P2 (0.5μM : 0.5μM) complex on negatively charged mica(-). (a) A typicalAFM image with a scan size of 500nm×500nm. The height the color bar scale is 2.5nm. A fewcharacteristic GagΔP6-PI(4,5)P2 complexes are boxed: monomer (red), dimer (green) and tetramer (blue).(b) Histogram for length (left), width (middle), and height (right). Shown in red are normal distribution fitsto the peaks. (c) Three dimensional smooth histogram, where monomer, dimer, and tetramer are indicatedby red arrows. The mean height is 1.93 ± 0.01nm for all complexes. The first peak with a mean length of11.2 ± 0.1nm and width 6.7 ± 0.1nm corresponds to the GagΔP6-PI(4,5)P2 monomer. The second peakwith a mean length of 21.1 ± 0.1nm and width 6.7 ± 0.1nm corresponds to the GagΔP6-PI(4,5)P2 dimer.The third peak with a mean length of 30.4 ± 0.2nm and width 14.0 ± 0.2nm corresponds to theGagΔP6-PI(4,5)P2 tetramer. The total number of samples was 903 and the experiment was repeated twice. Fig.10 Mixture of PI(4,5)P2-ΨRNA-GagΔP6 (0.5μM : 0.5μM : 0.5μM) complex on negatively chargedmica(-). (a) A typical AFM image with a scan size of 500nm×500nm. The height color bar scale is 2.5nm. Afew characteristic PI(4,5)P2-ΨRNA-GagΔP6 complexes are boxed: monomer (red), dimer (green) andtetramer (blue). (b) Histogram for length (left), width (middle), and height (right). Shown in red are normaldistribution fits to the peaks. (c) Three dimensional smooth histograms, where monomer, dimer, andtetramer are indicated by red arrows. The mean height is 1.91 ± 0.01nm for all three complexes. The firstpeak with a mean length of 10.9 ± 0.3nm and width 7.4 ± 0.1nm corresponds to thePI(4,5)P2-ΨRNA-GagΔP6 monomer. The second peak with a mean length of 23.8 ± 0.2nm and width 7.4 ±0.1nm corresponds to the PI(4,5)P2-ΨRNA-GagΔP6 dimer. The third peak with a mean length of 23.8 ±0.2nm and width 22.1 ± 0.2nm corresponds to the PI(4,5)P2-ΨRNA-GagΔP6 tetramer. The total number ofsamples was 616 and the experiment was repeated twice. Fig.11 Rough models of PI(4,5)P2-ΨRNA-GagΔP6 complexes based on the measured mean height andwidth. (a) Dimer complex, (b) Tetramer complex. MA domain is in red, CA domain is in yellow, NCdomain is in green, ΨRNA is in cyan, and PI(4,5)P2 is in purple. Supplemental Material: Investigation of HIV-1 Gag binding with RNAs and Lipidsby Atomic Force MicroscopyS1 AFM cantilever tip calibrationAu nanospheres of 2nm diameter were used for the calibration of the AFM cantilever tips. Asshown in Fig.S1(a), λ is the actual AFM cantilever tip size, L is the measured size of the calibration sample,and D is the height of the same calibration sample. α and β are the effective front and back angles of the tipwhich are related to the actual front angle FA and back angle BA of the tip FA through the tilt angle, θ, ofthe AFM probe holder (fluid cell if imaging in liquid) as shown in Fig.S1(b).Fig.S1 AFM tip calibration. (a) Schematic representation of the tip and a 2nm diameter Au calibrationsphere. The red curve is the trajectory of the tip when it scans from left to right. (b) The configuration of thetip after it is placed into the AFM probe holder (fluid cell if imaging in liquid). α = FA + θβ = BA - θHN = λOP = PS = OP sinβOS = OP cosβMQ = OS = cosβNM = PM tanβ = (PS + SM) tanβ = tanβ (sinβ + 1)NQ = NM + MQ = tanβ (sinβ + 1) + cosβ = (secβ + tanβ) Similarly,
QZ = (secα+ tanα)L = HN + NQ + QZ = λ + (secα+ tanα + secβ + tanβ) For θ = 10 , FA = BA = 20 , then α = 30 and β = 10 , hence, L ≈ λ +1.46D
Therefore, the actual tip size is λ = L – 1.46D (1)And the effective tip size t for a sample with the height of D is t = L – D = λ + 0.46D (2)S2 Method for computation of the sample sizeThe size of the sample is computed by a MATLAB script. First, a proper threshold value isselected to compute the height of the sample. The threshold value is so chosen to avoid the backgroundnoise. The height is the distance from the highest point of the sample to the threshold value cutoff plane.After the cutoff plane is set, the boundary of the sample can be obtained from the intersection of the sampleand the cutoff plane. The length of the sample is defined as the distance between the two furthest points onthe boundary. The width is defined as the distance between the two parallel lines restricting the objectperpendicular to the direction of the length defined above. Fig.S2 Schematic diagram of a sample with defined length and width.S3 Gel electrophoresisS3.1 Protocol for gel electrophoresisSDS-PAGE usually comprises of acrylamide, bisacrylamide, SDS, and a buffer with the proper pH.In the experiments reported here, tris-glycine SDS-PAGE was used with a 6% resolving gel and 5%stacking gel. The exact composition used in the Tris-Glycine SDS-PAGE is given in Table S1. Here SDS issodium dodecyl sulfate, APS is ammonium persulfate, and TEMED isN,N,N’,N’-tetramethylethylenediamine. In general, one SDS molecule is approximately bound to twoamino acids regardless of the polypeptide sequence. Therefore, the migration of SDS bound proteins isproportional to the molecular weight of proteins. The procedure of SDS-PAGE gel electrophoresis used wasas follows. First, 10mL 6% resolving gel and 4mL 5% stacking gel were prepared based on the compositionfor Tris-Glycine SDS-PAGE as given in Table S1. APS and TEMED were added later for the 5% stackinggel. Then about 7mL of 6% resolving gel solution was added into the Mini-PROTEAN II Cell (BIO-RAD,Hercules, CA, USA). Next 1mL isopropanol was added on top to remove air bubbles at the surface of theresolving gel solution. After a wait of 20 minutes, the isopropanol was removed. Next, 40μL 10% APS and4μL TEMED was added to the stacking gel solution. About 3mL of the 5% stacking gel solution wasplaced above the resolving gel solution. Finally a 10-well Mini-Protein comb (BIO-RAD, Hercules, CA,USA) was inserted on top of the stacking gel. The arrangement was allowed to form for 30 minutes beforeproceeding. The sample solutions to be used were mixed with the dye bromophenol blue (Sigma-Aldrich,Merck KGaA, Darmstadt, Germany) to the desired concentration. Then the sample solutions were boiledfor 5 minutes. Care was taken to seal all the sample solutions using parafilm. Next 1X SDS-PAGEloading buffer (25 mM Tris, 192 mM glycine, 0.1% SDS) was added into the gasket. The comb wasgently removed. The dyed sample solutions were added into each comb slot carefully. Enough 1XSDS-PAGE loading buffer was added into the chamber that surrounded the gasket. It was confirmed thatbubbles were generated at the bottom of the gel electrophoresis equipment. The following settings V =120V, I = 100mA, T = 75minutes, and K in “Volts” mode were used. The gel run was started and continueduntil the color had almost reached the bottom. The equipment was disassembled to get the gel with thedistinct bands located at the different positions. A plastic wedge plate was used to cut the top part of the gel.The gel was put into the transfer buffer (48 mM Tris, 39 mM glycine, 1.3mM SDS, 20%(v/v) methanol)using a clean container and then placed on a rotator for 20 minutes. A piece of Amersham Hybond 0.45μmPVDF blotting membrane (GE Healthcare Life Sciences, Pittsburgh, PA, USA) was cut to the same size asthe gel plate. A corner was cut to distinguish the orientation. Two pieces of the blot paper were soaked inthe transfer buffer for 20 minutes. The blotting membrane was soaked with pure methanol for 5 minutes.Next the methanol was removed and the transfer buffer was added and allowed to soak for 20 minutes. Theblotting paper was placed on the semi-dry transfer cell. The blotting membrane was next positioned atopthe blotting paper followed by the gel plate. A second blotting paper was placed on top of the gel. Airbubbles were removed by rolling a tube over the blotting paper. The following parameters V = 24V, I =100mA, T = 28minutes, and K in “Volts” mode, were set to run the semi-dry transfer cell. Next 0.5g drymilk was added to 10mL 1X Tris-buffered saline (TBS, 50 mM Tris-Cl, pH 7.5 150 mM NaCl) buffer to get5% milk TBS. The blotting membrane was put into a clean container, and 5% milk TBS buffer was addedand incubated for 1 hour on a rocking platform in a freezer. The 5% milk TBS was then removed. Next10mL reusable antibody (goat anti-HIV P24 in 5% milk TBS, 1:500) was added and incubated overnight ona rocking platform in a freezer. The blotting membrane was washed 3 times with 1X TBS buffer with 10min intervals. Next 10mL 5% milk TBS was prepared with the addition of 5μL of another antibody (rabbitanti-goat secondary antibody). This was added into the blotting membrane and incubated for 90 minutes. The milk TBS was removed and the blotting membrane was washed 3 times with 1X TBS buffer. Enoughdeveloping solution (10mL 1M Tris-HCl(pH 9.5), 2mL 5M NaCl, 0.5mL 1M MgCl2, 33μL 50mg/mL NBT(nitro-blue tetrazolium chloride), and 16.5μL 50mg/mL BCIP (5-bromo-4-chloro-3'-indolyphosphatep-toluidine salt) was added to completely cover the blotting membrane. After the bands appeared, theblotting membrane was washed with 1X TAE (40mM Tris, 20mM acetic acid, and 1mM EDTA(Ethylenediaminetetraacetic acid)) buffer. The blotting membrane was dried with nitrogen gas. A picture ofthe blotting membrane was taken and then analyzed using the software ImagJ.S3.2 Gel electrophoresis resultThe motivation for implementing gel electrophoresis measurement was to confirm the identity of thehigher order multimer complexes found in the AFM measurements as corresponding to trimers or tetramers.The results from Tris-Glycine SDS-PAGE is shown in Fig.S3. The 50kDa protein marker indicates thebottom band corresponding to the 50kDa monomer. This is because the mass of the complete Gagmonomer is about 55kDa and that of GagΔP6 is around 50kDa [8]. The middle band is just above the75kDa calibration marker and closely aligns with the 100kDa calibration marker, corresponding to that ofthe Gag dimer as the mass of the GagΔP6 dimer should be 100kDa. The top band is in between thecalibration mark for 150kDa and 250kDa, which means they are tetramers with approximate molecularweight of 200kDa for GagΔP6. For the Gag-ΨRNA complex the tetramer would have a mass around236kDa if including one ΨRNA. These results confirmed that the complexes found in AFM measurementswere tetramers rather than trimers. As shown in Fig.S3(a) and (b), the percentages of dimer and tetramersincreased as the concentration of GagΔP6 increased from 0.5μM to 2μM. This result is consistent with theconclusion that the average radius of gyration Rg of GagΔP6 in solution measured by SANS is amonotonically increasing function of GagΔP6 concentration. This is reasonable because the size of thetetramer is greater than the size of the dimer, which in turn is again greater than the size of the monomer[57]. As shown in Fig.S3(b) and (c), or (a) and (d), the percentages of dimer and tetramers increased withthe addition of ΨRNA for the same concentration of GagΔP6. This was also confirmed in the AFMmeasurements where ΨRNA can bind with GagΔP6 and facilitate GagΔP6 multimerization. As shown inFig.S3(c) and (f) and (g), the percentages of dimer and tetramers increased even more when both ΨRNAand PI(4,5)P2 were added. Similar to ΨRNA, increasing PI(4,5)P2 lipid lead to increased GagΔP6multimerization.5