Grifonin-1: A Small HIV-1 Entry Inhibitor Derived from the Algal Lectin, Griffithsin

Background Griffithsin, a 121-residue protein isolated from a red algal Griffithsia sp., binds high mannose N-linked glycans of virus surface glycoproteins with extremely high affinity, a property that allows it to prevent the entry of primary isolates and laboratory strains of T- and M-tropic HIV-1. We used the sequence of a portion of griffithsin's sequence as a design template to create smaller peptides with antiviral and carbohydrate-binding properties. Methodology/Results The new peptides derived from a trio of homologous β-sheet repeats that comprise the motifs responsible for its biological activity. Our most active antiviral peptide, grifonin-1 (GRFN-1), had an EC50 of 190.8±11.0 nM in in vitro TZM-bl assays and an EC50 of 546.6±66.1 nM in p24gag antigen release assays. GRFN-1 showed considerable structural plasticity, assuming different conformations in solvents that differed in polarity and hydrophobicity. Higher concentrations of GRFN-1 formed oligomers, based on intermolecular β-sheet interactions. Like its parent protein, GRFN-1 bound viral glycoproteins gp41 and gp120 via the N-linked glycans on their surface. Conclusion Its substantial antiviral activity and low toxicity in vitro suggest that GRFN-1 and/or its derivatives may have therapeutic potential as topical and/or systemic agents directed against HIV-1.

In addition to high potency [24], GRFT showed stability over a satisfactory pH and temperature range, caused minimal toxicity, and did not induce the release of proinflammatory cytokines that might recruit potential HIV-susceptible target cells to the target mucosa. These desirable properties make GRFT an excellent candidate microbicide [28], as well as an intriguing starting point for the design of smaller peptide-based antiviral minilectins directed against high mannose sugars.
As only two entry inhibitors, FuzeonH (T20, Enfuvirtide) and Maraviroc, are currently in clinical use, there is a need for new entry inhibitors [29][30][31][32][33][34] that could be used topically to prevent infection, or systemically to treat patients with drug-resistant HIV.
Since the systemic use of non-human proteins, including griffithsin or cyanovirin-N, may be time-limited by their immunogenicity, smaller griffithsin-derived peptides may present a more suitable alternative for such applications.

Ethics Statement
Venous blood samples (uncoded) were collected in the Department of Medicine at UCLA, by a qualified person under the approved Institutional Review Board protocol (IRB #92-11-596-41) after obtaining written consent from participants.
Primary vaginal epithelial cells (VECs) were purchased from MatTek Corp., (Ashland, MA USA) and provided with no identifiers not allowing direct identification of the donors.
Disulfide bond formation. Peptides were dissolved at a final concentration of 0.25 mg/ml in 50% DMSO in H 2 O and stirred overnight at room temperature. Subsequently, the peptides were lyophilized and re-purified on a preparative C18 Symmetry-Shield TM RP-HPLC column to .95% homogeneity and their purity was evaluated by MALDI-MS and analytical RP-HPLC.
Analytical HPLC. Analytical RP-HPLC was performed on a Varian ProStar 210 HPLC system equipped with ProStar 325 Dual Wavelength UV-Vis detector with the wavelengths set at 220 nm and 280 nm (Varian Inc., Palo Alto, CA). Mobile phases consisted of solvent A, 0.1% TFA in water, and solvent B, 0.1% TFA in acetonitrile. Analyses of peptides were performed with an analytical reversed-phase C18 SymmetryShield TM RP18 column, 4.66250 mm, 5 mm (Waters Corp., Milford, MA) applying a linear gradient of solvent B, from 0 to 100%, over 100 min with a flow rate of 1 ml/min.
All biological and biophysical experiments presented in the article were performed using unfractionated heterogeneous mixture of oligomers of GRFN peptide(s) (please see Analytical ultracentrifugation section).
Binding studies. Surface plasmon resonance (SPR) studies were performed on a BIAcore 3000 system (BiaCore AB, Piscataway, NJ), using BIAcore CM5 sensor chips and HBS-EP running buffer (pH = 7.4) containing 10 mM HEPES, 150 mM NaCl, 3 mM EDTA, and 0.005% polysorbate 20. Proteins/ peptides were immobilized on a CM5 sensor chip using the amine coupling method. The chip was activated by mixing 400 mM Nethyl-N-(3-dimethylaminopropyl)-carbodiimide hydrochloride and 100 mM N-hydroxysuccinimide. Residual reactive groups on the chip surface were blocked using 1.0 M ethanolamine/HCl (pH = 8.5). The flow cell-1 chamber, which served as a control, lacked immobilized protein but was treated with N-ethyl-N-(3-dimethylaminopropyl)-carbodiimide hydrochloride, N-hydroxysuccinimide, and ethanolamine/HCl. Binding signals were corrected for nonspecific binding by subtracting the flow cell-1 signal. To regenerate chip surfaces, bound ligands were removed with 10 mM HCl. Data were analyzed with BIAevaluation 4.1 software (Biacore, Piscataway, NJ).

Antiviral activity studies
Luciferase reporter assay (TZM-bl assay). [36] The ability of the analogs to inhibit HIV infection was determined with an assay that determines viral infectivity with an LTRmediated luciferase reporter construct [37]. In brief, TZM-bl reporter cells are HeLa-derived JC53-BL cells that express high levels of CD4, CXCR4, and CCR5. They also contain reporter cassettes for luciferase and b-galactosidase, both driven by the HIV-1 long terminal repeat [38]). 24 hr after these reporter cells were exposed to R5 lab-adapted strains of HIV-1 (BaL) in the presence of peptides, protection was measured as % reduction in luciferase activity (relative light units) compared to cells infected with virus and peptide vehicle alone.
p24 gag antigen release assay. As previously described [15], immortalized CD4+ lymphoblastic PM1 cells, (1.5610 5 /100 ml) were incubated with virus (BaL) (MOI = 10 22 ) in the presence or absence of 0.078-5 mM of GRFN-1 for 3 hrs at 37uC/5% CO 2 . The cells were washed and resuspended in fresh growth media containing vehicle or peptides for 7 days. Supernatants were collected at days 3, 5, and 7 post-infection, and on days 3 and 5 the cells were resuspended in 1 ml of growth medium containing the appropriate concentration of vehicle or peptide. HIV-1 infection was quantified by measuring the amount of p24 gag in the cell supernatants (Perkin Elmer p24 ELISA, Waltham, MA). Day 5 supernatants were determined to represent the peak infection day in PM1 cells. Cell viability and cytotoxicity. Cell viability experiments were carried out using CytoTox-Glo TM Assay (Promega Corp., Madison, WI) and cytotoxicity was analyzed using an MTT-based cell proliferation kit (Boehringer Mannheim, Indianapolis, IN) Both tests were performed according to manufacturer's instructions. MTT is 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide. Experiments utilizing vaginal epithelial cells (VECs) were carried out as previously described [49].
Human red blood cells (hRBCs). Heparinized fresh blood (,3 ml) was collected from an anonymous donor. A 200 ml aliquot was removed and washed 46 with 800 ml of cold PBS at RT for 3 min at ,5006g. After each centrifugation, 780 ml of supernatant was removed. After the last wash, the supernatant was completely removed and the packed red blood cells were diluted 1:20 in PBS to prepare a 5% v/v stock suspension of hRBCs.
Primary vaginal epithelial cells. (VECs) were purchased from MatTek Corp., (Ashland, MA USA) and maintained according to provided protocol.
Hemolysis assay. As previously described [39], 2.5% v/v normal human RBCs were exposed to various concentrations of GRFN-1 at 37uC for 30 min. All experiments were carried out in triplicate using 96-well microplate (Costar 3596) and the OD at 700 nm was monitored every 30 s. employing the SpectraMAX 250 microplate reader (Molecular Devices, Sunnyvale, CA). Controls included PBS, untreated human RBCs, and human RBCs +2.5% Triton (100% lysis). The hemolytic effect was calculated as follows: Inflammatory response was assessed using Bio-Plex Human Cytokine Multiplex Assays (Cat# M50-0KCAF0Y & MF0-005KMII, Bio-Rad Laboratories, Inc., Hercules, CA). Briefly, human peripheral blood mononuclear cells and primary vaginal epithelial cells (VECs) were cultured with 0-100 mM of GRFN-1 in appropriate media. After 24 h supernatants were collected and assayed according to manufacturer's protocol. Only data for factors with readable, ''in range'' values are presented.
Circular dichroism (CD) analyses of secondary structure. CD spectra from 185-260 nm of GRFN-1 were examined in different solution environments using a JASCO 720 spectropolarimeter (Jasco, Easton, MD) calibrated for wavelength and optical rotation with 10-camphorsulphonic acid [40,41]. Peptide was scanned at 20 nm per minute in a 0.01 cm path-length cell at 25uC with a sample interval of 0.2 nm. The peptide concentration was determined by UV absorbance at 280 nm. After baseline correction, spectra were expressed as the Mean Residue Ellipticity [h] MRE . Quantitative estimates of secondary structural contributions were made with Selcon [42] using the spectral basis set for proteins in the Olis Global Works TM software package (Olis Inc., Bogart, GA).
Fourier transform infrared (FTIR) spectroscopy. Infrared spectra were recorded at 25uC using a Bruker Vector 22 TM FTIR spectrometer with a deuterated triglycine sulfate (DTGS) detector, and averaged over 256 scans at a gain of 4 with a resolution of 2 cm 21 . Peptide samples were initially freeze-dried several times from 10 mM HCl in D 2 O to remove any interfering counter ions and residual H 2 O. Solution spectra of peptides were made in deuterated 10 mM phosphate buffer, pD = 7.4 (pD = pH+0.4) and in structure promoting mixed solvent-buffer solutions (trifluoroethanol (TFE) or hexafluoroiso propanol (HFIP). Spectra were acquired using a temperature controlled, demountable liquid cell with calcium fluoride windows fitted with a 50 mm thick spacer (Harrick Scientific, Pleasantville, NY). The relative proportions of a-helix, b-turn, b-sheet, and disordered conformations of solu-  Table 1. Sequences of Grifonins (GRFNs).

GRFN-1
Cha   [43]. Molecular dynamics modeling. The starting structure for GRFN-1 was obtained by using Hyperchem 7.5 (http://www. hyper.com) to template the peptide structure in a beta hairpin conformation derived from segments of the griffithsin protein (PDB 2GTY, residues 18-31). These monomeric starting structures were placed in a periodic 56Å 3 box of TIP4P water and the ensemble was neutralized with counter ions to simulate the environment used for the experimental CD measurements. The peptide in the solution box was conjugate-gradient minimized using the Polak-Ribiere approach implemented in Hyperchem. The minimized monomeric GRFN-1 ensemble was ported to the Gromacs program suite, version 4.0.4 (http://www.gromacs.org), and subjected to the steepest descent method using the OPLS_AA option [44].
The system was subjected to 20 psec of pre-run molecular dynamics at 300uK allowing the solvent to equilibrate while restraining the peptide. After pre-run solvent equilibration, the peptides were subjected to 50 nsec of free MD simulations at 300uK without any experimental constraints, utilizing Berendsen temperature and pressure coupling and the Particle Mesh Ewald method for evaluating long-range electrostatic interactions. The time-dependent evolution of the peptide secondary structure (i.e., analyzed using the DSSP criteria [45] for the peptide in the water environment indicated when equilibrium was reached. Molecular model illustrations were rendered using PyMOL v0.99 (http://www.pymol.org).
Analytical ultracentrifugation. Sedimentation equilibrium was performed at 20uC in a Beckman Optima XL-A analytical ultracentrifuge using absorption optics at 280 nm. Twelve mm path length double sector cells were used for all samples. Samples (OD 280 = 0.7 and OD 280 = 0.15)) were in 20 mM NaH 2 PO 4 . Sedimentation equilibrium profiles were measured at speeds of 3000, 7000, 11000, 24000 and 36000 rpm. The data were fitted with a nonlinear least-squares exponential fit for a single ideal species using the Beckman Origin-based software (Version 3.01 A partial specific volume of 0.739 was estimated from the amino acid composition [46] using 0.9 (the partial specific volume of a leucine residue) for the partial specific volume of cyclohexylalanine and cyclohexylglycine. It was then corrected to 20uC [47].

Design and synthesis of Grifonins (GRFNs)
The remarkable reported activity of griffithsin against laboratory strains and primary isolates of T-and M-tropic HIV-1 prompted us to perform a detailed structural analysis of this protein as the first stage of our attempt to design small peptidic analogs that potentially could mimic the anti-viral properties of the native griffithsin.
Structurally, three homologous b-sheet domains (residues 20-34, 58-76 and 96-120) form the protein's core (Figures 1 & 2) and contain the functional repeats responsible for its carbohydrate-binding [23]. Since the protein lacks disulfide bonds, its integrity is most likely maintained by hydrophobic and ionic interactions. We theorized that each b-sheet domain can either act independently as an entry inhibitor and/or assemble into the higher order structures resembling those of native griffithsin. Consequently, some of our peptide analogues contained unusual amino acids and/or structural features not present in native griffithsin to impose secondary structural features, which we hypothesized might be important for activity and multimerization. Specifically, we introduced varying numbers of cyclohexylalanine (Cha) and cyclohexylglycine (Chg) residues that might ''dial-up'' hydrophobic interactions between monomers and possibly increase assembly of higher order structures (desirably trimers). Some analogs also contained a disulfide bridge to stabilize bhairpin-monomer structure. After re-evaluating the initial group of compounds with respect to their composition, hydrophobicity and ease of synthesis, we concluded that 3 analogues shown in Table 1 were feasible for further development. An attempt to obtain folded (oxidized) GRFN-1 directly on the solid support using thalium trifluoroacetate ((CF 3 COO) 3 Tl) protocol [48] was unsuccessful. Consequently, after solid-phase synthesis, the linear peptides were oxidized in 50% DMSO to form the desired b-hairpin GRFNs. Analytical data for synthesized peptides are presented in Table 2 & Figure 3.

Antiviral activity of GRFNs
Initially, grifonins 1-3 were screened in a standard HIV-1 reporter assay ( Figure 4A). Dose response experiments revealed that GRFN-1 was the most potent analog and that it maintains high activity at low mM concentrations. Consequently we chose GRFN-1 as our lead compound and re-tested a broader range of concentrations in the TZM-bl assay as well as in the more stringent p24 gag antigen release assay using immortalized CD4+ lymphoblastic PM1 cells ( Figure 4B). This 18 amino acid long peptide showed considerable activity in both assays with EC 50 values 190.8611.0 nM and 546.6666.1 nM for TZM-bl and p24 gag antigen release assays respectively. GRFN-1 was also more potent than retrocyclin (RC)-101, which had an EC 50 of 3,404691 nM in the TZM-bl assay. RC-101, which served as positive control, is a humanized h-defensin, which is being developed as a prospective topical microbicide [49]. However, GRFN-1 was significantly less active than the parental protein griffithsin ( Figure 4C) that in our hands showed potent activity in low picomolar range (EC 50 = 19.661.9 pM). Both, griffithsin and GRFN-1 maintained their antiviral activity in PBMC based p24 gag antigen release assays toward CXCR4 (HIV-1 IIIB ) as well as CCR5 (HIV-1 BAL ) strains ( Figure 4D).

GRFN-1 binds viral glycoproteins
As N-linked glycans are the molecular target(s) of griffithsin, we sought to determine whether GRFN-1 acts via similar mechanism. To ascertain this, we performed two types of surface plasmon resonance (SPR) binding experiments. In the first set, we established that GRFN-1 binds gp41, gp120 BAL and gp120 LAV in a dose dependent manner ( Figure Figure 5D shows that the K D (0.5060.13 mM) for the selfassociation of GRFN-1 was considerably below its K D for binding to the aforementioned viral glycoproteins. This finding suggested that GRFN-1 acts as a multimer rather than as a monomer. Binding to N-linked glycans was additionally confirmed in competition experiments employing various saccharide components of such glycans: mannose (a-D-mannopyranose, Man), fucose (L-fucopyranose, Fuc), galactose (a-D-galactopyranose Gal), sialic acid (N-acetyl-neuraminic acid, Neu5Ac), N-acetyl-glucosamine (GlcNAc) and common core pentasaccharide Man 3 -GlcNAc 2 ( Figure 6 & Supporting Information S1). Although GRFN-1 interacted with all of these sugars, except fucose, in the range of concentrations tested, the most effective competitors were GlcNAc and mannose.

Cytotoxicity and pro-inflammatory properties
Since systemic uses would typically require either intravenous or subcutaneous administration of the GRFN-1, we performed hemolysis assays using various concentrations of the peptide and human red blood cells (hRBCs) ( Table 3 & Figure 7). The compound showed low hemolytic activity (,10%) toward hRBC at the highest concentration tested (20 mM) and was not hemolytic at concentrations below 2.5 mM. Toxicity studies of GRFN-1 with TZM-bl cells (Figure 8) showed no effect on viability, although  the peptide appeared to inhibit cellular dehydrogenase activity (as gauged by MTT reduction) in a dose dependent manner. Studies with primary vaginal epithelial cells (VEC), peripheral blood mononuclear cells and PM1 cells showed similar effects (Figure 9) although actual level of metabolic inhibition was strongly dependent on cell type.
One of the important properties of antivirals is lack of proinflammatory properties. We tested GRFN-1 for such properties using human PBMCs and primary VECs. Cells were treated with the peptide for 24 hr and subsequently supernatant was analyzed using Bio-Plex Human Cytokine Multiplex Assays (Bio-Rad Laboratories, Inc., Hercules, CA) for various pro-inflammatory cytokines and growth factors. Results are presented in Supporting Information S1. We found a number of factors to be either unchanged or decreased after treatment with GRFN-1. These include IL-5, IL-8, IL-10, IL-13, VEGF, IFN-c, TNF-b, GM-CSF, MIP-1a and others. Notably, reverse effect of GRFN-1 treatment was also observed for limited number of proinflammatory factors.

Secondary structure analysis of GRFN-1
Analysis by FTIR spectroscopy. Analysis of the secondary structure of GRFN-1 in solvent systems of varying polarity are shown in Figure 10A. In aqueous buffer the peptide has a dominant b-sheet structure ( Table 4). In less polar environments such as the amphipathic TFE:buffer solvent system and the more hydrophobic HFIP:buffer environment, there was a shift from bsheet to more helical conformations with greatest helical propensity in the more hydrophobic environment ( Table 4).
These observations suggest that the GRFN-1 peptide can assume different conformations, depending on the polarity of the solvent.
Analysis by CD spectroscopy. The conformational plasticity of GRFN-1 was further studied by CD spectroscopy. In 10 mM phosphate buffer, pH = 6.5, GRFN-1 shows a concentration dependent change in secondary structure ( Figure 10B). At 250 mM, almost equal proportions of a-helix, b-sheet, turn and disordered conformations are present. However, at 500 mM the bsheet structure increases at the expense of a-helix ( Table 5). In contrast to the concentration-dependent secondary structural changes in aqueous buffer, the conformation of GRFN-1 in structure-promoting solvent systems such as TFE:buffer and HFIP:buffer ( Figure 10C) had dichroic minima at 222 and 208 nm with a maximum near 193 nm. These features, characteristic of peptides with more helical conformations [50], were not concentration dependent over the range of 50 to 500 mM (data not shown). Analysis of the CD spectra with curve-fitting algorithms ( Table 5) reveal that GRFN-1 had a dominant helical conformation in HFIP:buffer and a mix of a-helix, b-sheet, turns in TFE:buffer suggesting that the peptide has a polarity dependent polymorphism.

Molecular dynamics simulations
To further investigate similarities between their structural elements, molecular dynamics simulations were performed using the conformation of griffithsin residues 18-31 as a homology template for the starting structure of GRFN-1. The templated structure was then subjected to 50 nsec of dynamics without constraints to refine the structure. The DSSP plot ( Figure 11) indicates that the peptide assumes a stable secondary structure with well defined loop and b-sheet segments. This disposition of secondary structure in the peptide construct is very similar to that observed in the parent griffithsin crystal structure ( Figure 12) and is consistent with the type and amount of secondary structure observed experimentally in the FTIR and CD measurements of the GRFN-1 peptide in aqueous medium.

Analytical ultracentrifugation
Sedimentation equilibrium suggests that GRFN-1 in diluted salt is present as high molecular weight soluble aggregates with some monomer present. When initially dissolved in 20 mM NaH 2 PO 4 the peptide gave a slightly turbid solution. Very little (less than 5%) of the OD 280 was lost on centrifugation at 3000 rpm. Early scans (after 16 hours of centrifugation) suggested the sample was quite heterogeneous, with a weight-average molecular weight of about 150,000. The samples were then examined at higher speeds. Approximately two-thirds of the OD 280 was lost at 7000 rpm, with the remaining material being heterogeneous (as seen by the nonrandom residuals from a single-exponential fit visible in Supporting Information S1 with a weight-average molecular weight of about 69000. Further scans at 11000, 24,000 and 36,000 rpm gave successively lower weight-average molecular weights as the higher molecular weight complexes were removed by centrifugation ( Table 6). By 36,000 rpm the remaining material (approximately one quarter of the original OD) had a weight-average molecular weight, 2200, and random residuals from a single-exponential fit consistent with the remaining material being reasonably homogeneous monomer.

Discussion
Although 30 years of concerted research have led to impressive progress in the therapy of patients infected with HIV-1, therapy remains imperfect and chemoprevention of HIV infections remains an unmet challenge. In this report we present data for a novel HIV-1 entry inhibitor, grifonin-1 (GRFN-1), that was obtained by modifying and truncating the naturally occurring lectin, griffithsin. GRFN-1 peptide is over 6 times smaller than the original protein (18 residues vs. 121), and it is only half the size of FuzeonH (18 vs. 36 residues), a peptidic entry inhibitor in clinical use. These features make GRFN-1 an attractive compound for further development.
We initially synthesized 3 closely related analogues (GRFNs 1-3) that were engineered to form stable b-hairpin structures that simulated structural features found in the native protein. The properties of each monomer were further modified to enhance self-assembly into higher order structure(s) by increasing hydro-     Figure 4A) was considerably less active than GRFN-1. Asn 31 is not implicated in interactions responsible for ligand(s) binding [23,[25][26][27], neither it is responsible for interactions between functional domains. Nonetheless, in GRFN-1 it may promote oligomerization since the Asn side chain may serve as both a donor and acceptor of hydrogen bonds [51]. Theoretically, inserting a positively charged Arg in the same position could create a repulsive force based on electrostatic interactions. Oligomerization of GRFN-1 in various solvents (water, DMSO and their mixtures) was suggested in our limited NMR studies (data not shown). Therefore we decided to research this phenomenon in a greater detail. SDS-PAGE analysis in nonreductive conditions was inconclusive (Supporting Information S1), however it suggested that dimers of GRFN-1 (,4.5 kDa) may be stable enough to persist in this relatively ''hostile'' environment. In addition, SPR experiments demonstrated that GRFN-1 can self-associate at low micromolar concentrations, (K D = 0.5060.13 mM) strengthening the likelihood that GRFN-1 oligomers contribute substantially to the activities demonstrated in our antiviral assays.
Analytical ultracentrifugation experiments provided insight into the size distribution of GRFN-1 oligomers and/or aggregates, showing that GRFN-1 forms ensembles that range in size from 2.2 (monomer)2150 kDa ( Table 6). Whether these higher order aggregates of GRFN-1 form any sort of regular structure(s) is difficult to predict from our data, however a ''multimer-based'' mode of action in in vitro/in vivo settings is highly probable. This may also account for the relatively high biological activity of GRFN-1, considering that the potency of its parental molecule, griffithsin, seems to be strongly associated with its multivalency [25]. Indeed, native griffithsin forms a domain swapped dimer with three almost identical carbohydrate-binding sites in each monomer. Such a mode of action may be advantageous since multivalent aggregates are likely to form more stable complexes with viral glycoproteins. In addition, they may be more resistant to proteolysis and more persistent in the bloodstream due to size imposed delay in renal excretion. Notably, based on EC 50 obtained in various antiviral assays, our leading compound (GRFN-1) is approximately 100 to 1000 times less effective than the parental protein griffithsin. Such a result might be explained by formation of imperfect oligomers that only partially mimic spatial arrangement of griffithsin dimer and its carbohydrate binding centers. In addition, GRFN-1 forms various size oligomers as illustrated by our ultracentrifugation studies, that is in contrast with very stable dimers formed by griffithsin. Similarly, stability of the GRFN-1 oligomers might also impose a detrimental effect on biological activity, since self-association of the peptide is rather moderate (K D = 0.5060.13 mM) and complexes may not be   ''stable enough'' leading to substantially lower biological activity outcome. From our SPR binding studies, GRFN-1 like its parental molecule, likely acts by binding N-linked carbohydrates on viral glycoproteins gp41 and gp120. However, competition experiments revealed certain binding preferences (Supporting Information S1). Although each of the sugars that we tested, except fucose, competed with the bindin of GRFN-1 to viral glycans, the inhibitory effects depended on the carbohydrate's structure and concentration. Inhibition of binding by GlcNAc and Gal was immediate and ''saturated'' at a molar ratio of ,4:1. D-Mannose was an especially potent inhibitor that could almost completely abrogate binding of GRFN-1 to the viral glycoproteins. Inhibition by sialic acid (Neu5Ac) was less profound and analysis was complicated by the accumulation of sialic acid, most likely due to ionic interactions with the peptide.
Given the substantial similarity between the tested carbohydrates, it is interesting to speculate why only fucose failed to compete against GRFN-1's binding to viral glycoproteins. The most striking structural difference between fucose and the other group members is absence of an equatorial hydroxymethyl (-CH 2 OH, in red in Figure 13)) in position 5 of the pyranose ring, which is occupied by a methyl group in fucose. This small difference seems to be pivotal for binding by GRFN-1. It is noteworthy that crystallographic analyses of complexes between griffithsin and various carbohydrates [23,26] demonstrated that the same hydroxymethyl group plays an important role in the hydrogen bond network that underlies its carbohydrate interac-tions-a finding that also underlines mechanistic similarities between grifithsin and its peptide derivative, GRFN-1.
The presented study identifies a novel 18-residue peptide, GRFN-1, that manifests potent anti-HIV-1 activity. Its low toxicity, limited hemolytic/proinflammatory properties, activity against CCR5 and CXCR4 HIV-1 strains and relatively small size identifies it as a strong lead candidate for further development as HIV-1 entry inhibitor for topical or systemic applications.