Analogs of LDL Receptor Ligand Motifs in Dengue Envelope and Capsid Proteins as Potential Codes for Cell Entry
It is established that cell entry of low density lipoprotein particles (LLPs) containing Apo B100 and Apo E is mediated by receptors and GAGs. Receptor ligand motifs, XBBBXXBX, XBBXBX, and ΨBΨXB, and mono- and bipartite NLS sequences are abundant in Apo E and Apo B100 as well as in envelope and capsid proteins of dengue viruses 1–4 (DENV1–4). Synthetic, fluorescence-labeled peptides of sequences in DENV2 envelope protein, and DENV3 capsid that include these motifs were used to conduct a qualitative assessment of cell binding and entry capacity using HeLa cells. DENV2 envelope peptide, Dsp2EP, 0564Gly-Gly0595, was shown to bind and remain at the cell surface. In contrast, DENV3 capsid protein peptide, Dsp3CP, 0002Asn-Gln0028, readily enters HeLa cells and accumulates at discrete loci in the nucleus. FITC-labeled dengue synthetic peptides colocalize with low density lipoprotein-CM-DiI and Apo E-CM-DiI to a degree suggesting that dengue viruses may utilize cell entry pathways used by LLPs.
The Flaviviridae viruses and low density lipoparticles (LLPs), low density lipoproteins (LDL), intermediate density lipoproteins (IDL) and very low density lipoproteins (VLDL), are unrelated biological entities. Nevertheless, there are many striking similarities between viral particles and LLPs that have been largely overlooked. Flaviviruses and LLPs are similar in a general sense as both include lipids and proteins arranged in roughly spherical structures. Flaviviruses are approximately 500 Å in diameter  while LLPs range in size from 250 to 600 Å [2, 3]. Intriguing, proteins of Flaviviridae and LLP have similar capacities to enclose, protect, transport, and deliver nucleic acids to cytoplasm and/or cell nucleus [4, 5]. Lipids in both are essential for conformational integrity and functionality of the proteins.
The Flaviviridae family includes three genera of positive, single-strand RNA viruses, Flavivirus, Hepacivirus, and Pestivirus . These arthropod-vectored human pathogenic viruses are emerging threats to first world nations . Flaviviruses are characterized by an outer “envelope” that protects and delivers the encapsulated genome. The outer envelope is comprised of a bilayer lipid membrane, derived from the host cell endoplasmic reticulum , and two glycoproteins, M and envelope glycoprotein E . In dengue viruses 1–4 (DENV1–4), the outer envelope contains 90 dimers of glycoprotein E each associated with a protein M . Many viruses, including the Flaviviridae, use receptor-mediated pathways for cell entry . Clathrin-mediated endocytosis of dengue has been suggested [10–13]. Domain III of the DENV envelope protein is thought to be essential for cell entry; however, specific receptor ligand motifs in dengue structural proteins have not been identified [6–14].
LLPs are characterized by a single molecule of apolipoprotein B100 (Apo B100), multiple copies of apolipoprotein E (Apo E), and other apolipoproteins such as apo C-I, C-II, C-III, A-I, A-II, A-IV, and M [15, 16]. LLPs include mainly phospholipids, cholesterol, cholesterol esters, and triglycerides [17, 18], arranged in pseudomicelles with a monolayer of phospholipid on the surface and a hydrophobic core. Endocytosis, which involves cell surface receptors and clathrin-coated endosomes, was elucidated for LDL and other LLPs first by Goldstein and Brown . Apo B100 and Apo E contain motifs that bind LDL B/E receptors and other proteins of the LDL receptor-related superfamily (LRPs) [20–23]. Motifs, XBBBXXBX, XBBXBX, ΨBΨXBX, and XXBXXBXXXBB (Types I, II, III, and IV, resp.), where B can be Arginine or Lysine, Ψ is a nonpolar amino acid, and X is any amino acid, are multifunctional motifs that impart the capacity to bind nucleic acids and other proteins, that is, receptors [21, 22, 24]. These motifs give LLPs a high affinity for GAGs, such as heparin and heparan sulfate. Binding of heparan sulfate proteoglycans (HSPG) may involve both GAG moieties and ligand-binding modules of the HSPG core protein .
In a previous report , we described structural similarities between Apo E and dengue capsid protein, PDB: 1LPE and 1R6R, respectively. Although their primary structures are not significantly similar, their secondary and tertiary structures are alike, that is, helix-loop-helix; and it has been confirmed experimentally that both proteins possess nucleic acid-binding and nuclear transit capacity [26, 27]. The Apo E receptor ligand motifs and receptor binding domain are well established ; however, the exact cell entry “codes” have yet to be elucidated in Flavivirus envelope and capsid proteins.
Our present studies are based on the hypothesis that proteins of the LDL receptor gene family ubiquitously expressed in human and mosquito may be used by DENV for cell entry. Our goals were to identify potential receptor ligand motifs in dengue envelope and capsid proteins by sequence comparison to the motifs in Apo B100 and Apo E; use fluorescence-labeled synthetic peptides to assess cell binding/entry capacity of the viral proteins; and assess effects of LDL and Apo E on binding and uptake of viral peptides using live HeLa cells in culture. Our findings may have implications for intervention strategies in viral infections in general.
DENV polyprotein sequences used in this study were obtained from NIH NCBI PUBMED Protein and Structure databases.
Phosphate-buffered saline (PBS) and cell culture media were from Media Tech, Inc., Herndon, VA. Penicillin G-Sodium, streptomycin sulfate, Fetal Bovine Serum, Trypsin, Corning Costar 6- and 12-well culture plates, and 8-well Lab-Tek Chamber glass slides with cover were from Thermo Fisher Scientific, Inc. Mevastatin was from Sigma-Aldrich, Inc. Trypan Blue (modified) 0.4% solution in PBS was from MP Biomedicals, LLC, Solon, OH.
2.2. Fluorescence-Labeled Peptides
Fluorescence-labeled synthetic peptides (Tables 1–3), of 95% or greater purity by HPLC analysis, were from NEO Group, Inc. Cambridge, MA. Peptides were received in lyophilized form in sealed 1 mL vials and solubilized in PBS-10 mM MgCl2 to yield 1 mg/mL.
2.3. Protein Labeling
Purified human Apo E (>95%) was from Athens Research & Technology, Athens GA. LDL was isolated as described previously  from pooled normal human plasma from Innovative Research, Inc., Movi, MI. For covalent labeling of the protein, 250 µL of ethanol was added to the vial containing 50 µg CM-DiI (7000), a thiol-reactive fluorescence compound from Invitrogen, Inc., to make a stock solution. Next, 1 µL of CM-DiI stock solution was added to 150 µg of Apo E in 500 µL or to 1 mg of purified LDL in 1 mL of PBS-5 mM MgCl2. These reaction mixtures were allowed to stand for 30 minutes at ambient temperature then dialyzed in 1 liter PBS-MgCl2 with three changes overnight.
HeLa cells (human cervix epithelial adenocarcinoma cell line, CCL-2) were from American Type Culture Collection (ATCC). Typically, cells were stored in liquid Nitrogen until needed and then seeded and propagated according to ATCC protocols in DMEM supplemented with 10% FBS, 100 units Penicillin G-Sodium, and 100 units/mL streptomycin sulfate (DMEM–P/S-FBS) at 37°C in an atmosphere of 5% CO2 in a humidified incubator. Cells at ~70% confluence were incubated overnight in DMEM–P/S-3% FBS medium containing 40 μM mevastatin, after which medium was replaced with 1 mL PBS. Labeled peptides (alone or with a potential competitor for receptor binding, Apo E, Apo E-derived peptide, or LDL) were added to each well then incubated under same conditions as above. Time-course and peptide concentration studies were conducted using FITC-labeled peptides to determine minimal concentration required to detect fluorescence signal, incubation period required for binding and/or uptake, and potential cytotoxicity. For Dsp2EP-FITC, 0.5–45 µg/mL of peptide (0.2–10.6 µM) was used, and the cells were observed for 5–120 min. Assays for Dsp3CP-FITC binding to HeLa cells were conducted using increasing concentrations of the peptide: 6.75 µM, 13.5 µM, 20.25 µM, 27 µM, 33.75 µM, 40.5 µM, and 47.25 µM. Cells were monitored at 15-minute intervals for 60 minutes and experiment was terminated at 90-minute time point.
To reduce background fluorescence the assay mixture was removed and live cells were rinsed thrice with PBS. In quenching experiments cells were incubated at ambient temperature for 10 minutes using PBS with 0.2% Trypan Blue. Live cells were then rinsed with PBS sufficiently to visualize cells to obtain bright-light images as described below.
2.5. Dual Label Experiments
In these experiments, Dsp2EP-FITC was used in combination with either Apo E-CM-DiI or LDL-CM-DiI. Cells in 8-chamber slides were preconditioned as routine and rinsed with PBS-5 mM MgCl2, incubated for 15 minutes at ambient temperature in 100 µL PBS-MgCl2 containing 5 µg Dsp2EP-FITC mixed with 1.5 µg Apo E-CM-DiI or 5.0 µg LDL-CM-DiI. Binding assay mix was removed and cells rinsed in PBS prior to microscopy analysis in 200 µL PBS-MgCl2. Images of live cells were obtained as before using green and red filters. Overlay images of Dsp2EP-FITC (green fluorescence image) and Apo E-CM-DiI or LDL-CM-DiI (red fluorescence image) were created using Adobe Photoshop CS 4 as described below. Representative regions showing signals green, red, and yellow were enlarged and subjected to digital analysis using Image J. Color was split into channels red, blue, and green, and raw integrated densities for each color were rendered in 3D.
In similar experiments, 50 µg Dsp3CP-FITC was combined with 75 µg Apo E-CM-DiI in 300 µL PBS-MgCl2. Increasing volumes, 8, 16, 24, 32, and 40 µL, of this mixture were added to cells preconditioned as above in an 8-chamber slide in PBS-MgCl2 in 200 µL assay. These were then incubated at 37°C for 15 and 90 minutes.
2.6. Fluorescence Microscopy
All images were obtained using live cells in PBS-MgCl2 in culture trays or 8-well slides. Microscopy was performed using a Zeiss Axiovert 25 equipped with Zeiss objectives and FITC/GFP (Zeiss HQ470/40x and HQ525/50m) and red (Chroma AT540/25x and AT605/55m) filter cubes. Images were captured through a 0.5x C-mount adapter using an Optronics Microfire CCD camera and Picture Frame software.
Adobe Photoshop CS Version 4 was used to enhance images. Full-frame images of field of view equal in size and dimensions were used to create overlay images . Analysis and assessment of raw integrated density values were obtained using Image J, which was also used to render 3D images.
Apo E binds all members of the LDL receptor superfamily [19, 23, 29]. The receptor ligand sequence in Apo E spans residues 0130Thr-Ala0159 [30, 31] and contains Type I and Type II ligand motifs and both types of NLS sequences (Table 1). Apo E also contains two Type III motifs, at 071Leu-Ser076 and 091Ala-Glu096 and additional NLS (Table 1). Peptide E0131-FITC, spanning Glu0131-Gly0173, contains the receptor ligand sequence LRKLRKRLLR. Synthetic peptides of the Apo E LDLR ligand region have been used to substantiate a variety of functions for this apolipoprotein [31–33], including anti-infective activity [34–36]. Lysines in this sequence are essential in binding to the LDL B/E receptors . Interaction of E0131-FITC with HeLa cells was evaluated in the presence or absence of LDL. Figure 1 shows images obtained with E0131-FITC (frames (a) and (b)) and E0131-FITC plus highly purified human LDL ((c) and (d)). E0131-FITC appears to bind to the cell surface (a) and form clusters (b) but does not appear to enter the cell. In contrast, the peptide signal is seen inside the cell when LDL is added to the assay (frame (c)) as verified by quenching of fluorescence using Trypan Blue (not shown). Hence, synthetic peptides of the Apo E molecule retain cell binding properties of the intact molecule (not necessarily the capacity for internalization) and may be used to identify functional LDLR ligand motifs in DENV proteins.
The LDLR ligand motifs in Table 1 were used to identify sequences with similar function potential in DENV capsid and envelope proteins. The first 800 N-terminal residues of the DENV polyproteins including capsid (0001Met-Arg0100), precursor membrane protein (0115Phe-Thr0280), and envelope glycoprotein E (0281Met-Ala0775) contain numerous potential receptor/GAG ligand motifs and NLS sequences (Tables 2 and 3).
3.1. Dengue Virus 2 Envelope Glycoprotein
Peptide Dsp2-EP with two ligand motifs, ΨBΨXBX, and one NLS sequence, XBXBXBX (Table 2), was evaluated for cell binding capacity. A comparison of similar motifs present in Apo E and Apo B200 is included in Table 2. Only Type III ligand motifs are present in DENV envelope protein (E protein) sequences that are rich in Glycine, Lysine, and Arginine residues, approximately 10%, 7%, and 3%, respectively. Figure 2 images (A1)–(A4) show that Dsp2EP-FITC binds HeLa cells but does not appear in the cytoplasm. Images (A2) and (A4), enlarged areas of (A1) and a similar field of view, clearly demonstrate the appearance of clusters on the cell surface. Mutated synthetic peptides Dsp2EP (B)- and (C)-FITC in which Aspartates replace basic residues in one of the ligand motifs (Table 2) also bind cells and do not appear in cytoplasm. Images (A3), (B3), and (C3) show that addition of Trypan Blue dye quenches fluorescence and confirms that Dsp2-EP and mutated forms bind to the cell surface but lack cell entry capacity. Raw integrated densities for images (A1), (B1), and (C1) obtained using Image J are shown in Figure 2(b). Reduced integrated densities are observed for the mutated peptides indicating that each ligand motif may be functional but not as efficient in binding as their combination in Dsp2EP.
HeLa cells were used to test cell binding properties of Dsp2EP-FITC in presence of Apo E and LDL. Increasing amounts of unlabeled Apo E were included with 10 µg Dsp2EP-FITC in 500 µL PBS/5 mM MgCl2 and placed at 4°C for 18 hours. In parallel, unlabeled human LDL was added to HeLa cells and similarly incubated. Overlays of bright-light and fluorescence images are shown in the lower section of Figure 3. Image (A) shows fluorescence obtained for 10 µg Dsp2EP-FITC (positive control). Images (B) and (C) were obtained with cultures containing 10 µg Dsp2EP-FITC in presence 400 µg of unlabeled Apo E (B) and 300 µg of unlabeled LDL (C) in the assay. Enlarged areas in (A1), (B1), and (C1), respectively, show that fluorescence appears both diffused on the cell surface and in clusters.
Graph in Figure 3(b) shows the number of loci of Dsp2-EP-FITC bound to HeLa cells in presence of unlabeled Apo E-2 peptide and LDL (10 µg of the peptide in all wells; effects of 40 µg, 160 µg, 400 µg of Apo E-2, and 300 µg, 600 µg, 1200 µg of LDL). Addition of unlabeled Apo E-2 peptide enhances Dsp2EP binding significantly as is shown by the increase in the number of loci with fluorescence (Graph (b), blue columns) and the increase in total raw integrated density (Graph (c), blue columns). Also, results in image (B), Figure 3(a), indicate that addition of Apo E-2 makes receptors available almost ubiquitously across the field of view. In comparison, the presence of LDL reduces the number of Dsp2EP-FITC fluorescing loci as shown in image (C) and Graph (b) (orange columns). These data also indicate that receptors for the Dsp2EP-FITC are made available in a limited number of cells. Our observations suggest a mechanism in which increased concentration of Apo E-2 triggers a conformational change in the receptor making more binding modules available for Dsp2EP-FITC binding, a mechanism reported previously for LDL receptors [23, 37, 38].
Dual label experiments using Dsp2EP-FITC with either Apo E-CM-DiI or LDL-CM-DiI, as described in Section 2, were performed to assess colocalization of Dsp2EP-FITC with Apo E and LDL. Images were obtained as described above. Overlay in frame (A) of Figure 4 shows combined fluorescence of Dsp2EP-FITC and Apo E-CM-DiI. Enlarged representative regions showing green, red, and yellow signals are indicated by 1–8. Each image was subjected to digital analysis using Image J and split into channels red, blue, and green, and raw integrated densities were then rendered in 3D. Enlarged images 1, 2, 5, and 6 are shown at bottom two rows of Figure 4 as 1G, 1R, and so forth. 3D images reveal that Dsp2EP-FITC binds solo as well as colocalized within the red clusters formed by Apo E-CM-DiI. These observations suggest that Dsp2EP-FITC and Apo E-CM-DiI may share receptors, binding to either the same or adjacent receptor molecules. Apo E-CM-DiI signal appears uniformly in clusters with similar maxima in all enlarged images, 1R, 2R, 5R, and 6R, in Figure 4. This may indicate saturation in the area density of receptors occupied by Apo E, which is consistent with the finding that each LDLR provides two modules for Apo E binding . This mechanism may leave sufficient number of unoccupied modules for binding to other ligands, such as Dsp2EP-FITC.
Results of parallel dual label experiments using Dsp2EP-FITC and LDL-CM-DiI are shown in Figure 5. A section of an overlay image presented in frame (A) shows separate loci of LDL-CM-DiI and Dsp2EP-FITC, respectively, while a yellow signal indicates colocation. Areas 1, 2, and 3 in frame (A) are shown enlarged as images 1, 2, and 3 in the row beneath. Image 1 shows loci of Dsp2EP-FITC alone and collocated with LDL-CM-DiI, as pinkish and yellow signals (also evident in 1G and 1R). Image 2 shows loci indicating LDL-CM-DiI predominantly (2R) except for two green dots identified as Dsp2EP-FITC (2G). Image 3 suggests a stronger presence of Dsp2EP-FITC versus LDL-CM-DiI that is supported by 3R and 3G. LDL-CM-DiI appears to form clusters similar to those seen for Apo E-CM-DiI discussed above. These results suggest that in some regions of the cell surface Dsp2EP-FITC may interact with the same receptor molecules as do Apo E and LDL.
3.2. Dengue Virus 3 Capsid Protein
The DENV nucleocapsid, encasing the viral genome, is enclosed by the bilayer lipid membrane in which the envelope glycoprotein is anchored . The nucleocapsid is comprised of multiple molecules of the capsid protein (C protein) associated with lipid. In DENV capsid proteins (residues 0001–0100 of the viral polyprotein), basic amino acids K and R represent approximately 26% of the sequence and are present in approximately equimolar concentrations. Ligand motifs, I, II, and III, are abundant (Table 3) and may impart the capsid protein with multiple options in binding to both receptor proteins and glycosaminoglycans. Such binding capacities may be functionally significant (see Section 4). Type I ligand motifs are present in DENV1, 3, and 4 C proteins, motif Types II and III occur in the capsids of all four viral versions; there is also a surplus of potential mono- and bipartite NLS.
Dsp3CP-FITC (0002Asn-Gln0028 of the DENV3 capsid protein, Table 3) contains a type I, QRKKTGKP, and a type II, LKRVRNR motifs (Table 3), that may function as monopartite and as components of a bipartite NLS. Dsp3CP-FITC was tested for cell entry and nuclear translocation potential in mevastatin-treated HeLa cells. Results are shown in Figure 6, (A1) and (A2) (an enlarged area of (A1)). Fluorescence appears located in cytoplasm in apparent endosomes, and clusters are seen in the nuclear space. This distribution of fluorescence is seen in most images obtained with Dsp3CP-FITC. Similar patterns were observed in presence of increasing concentrations of unlabeled Apo E-2. Examples are shown in Figure 6 (B1), with an enlarged area in (B2), which shows Dsp3CP-FITC signal in loci suggesting endosomes, at loci proximal to the nuclear envelope, and possibly in the nuclear space. Hence, Apo E-2 peptide at 1.5 molar excess does not appear to influence cell entry or nuclear translocation of Dsp3CP-FITC.
In dual label experiments, Apo E-CM-DiI was added with Dsp3CP-FITC to live HeLa cells in FBS-depleted EMEM and then incubated at ambient temperature for 15 and 90 minutes. Results are shown as overlays in Figure 6 rows (C) and (D), respectively. At 15 minutes (row (C)), colocalized signals are evident across the field of view with most appearing at the cell surface. This observation suggests a sharing of receptors and/or clustering of receptors carrying different signals. Enlarged overlay (C)1 shows a cell with colocalized signals at the membrane, in the cytoplasmic space, and apparently in the nuclear space. Enlarged overlays (C)2 and (C)3 show colocalized signals at discrete loci on cell surface. Dsp3CP-FITC represents ~71% raw integrated density; this includes diffused green, whitish green clusters indicating colocalized signals in the cytoplasm, and possibly in the nuclei. At the 90 minute point (row (D)), intracellular separation of Dsp3CP-FITC and Apo E-CM-DiI is evident. Dsp3CP-FITC signal appears both at discrete loci and diffused in cytoplasm. Elevated green fluorescence background may indicate binding of Dsp3CP-FITC to extracellular debris on culture plate. In overlays (D)1–(D)3 Apo E-CM-DiI is clearly separated from the diffused Dsp3CP-FITC; red signal seems to fill the nuclear space, where colocalization of the two signals is also apparent (yellow dots). These results in toto suggest Dsp3CP-FITC, which contains receptor ligand motifs and NLS sequences, may share LDL B/E receptors in cell entry and then translocate to nucleus.
Humans, sylvatic nonhuman primates, and mosquitoes are major reservoirs of DENV and other Flaviviridae . LDLR and/or LRPs are expressed in most cell types in humans. Analogue proteins in Aedes aegypti [39–41] and Culex quinquefasciatus  include apolipophorins and lipophorin receptor [43–46]. There are 17 different known members in the LDL receptor gene family (LRPs) including LDLR, VLDLR, LRPs 1–6, 8, 10, 11, and 12, Megalin (LRP2), and ST7p. All members have five distinct structural components, three are extracellular, one transmembrane sequence, and a cytoplasmic, carboxyl region . Relevant to the present report are the numerous clusters of ligand-binding, cysteine-rich modules contained in the extracellular domain (also known as complement-like repeats and ligand adhesion type A units). There are a total of 170 known renditions of ligand-binding modules that recognize at least 30 different ligands belonging to a variety of protein families. All LRPs are thought to bind Apo E .
DENVs are known to employ multiple mechanisms to obtain cell entry [11, 48]. There is evidence that DENV interacts initially with heparan sulfate proteoglycans essential in anchoring at the cell surface [49, 50]. Several target proteins on the cell surface have been identified and clathrin-mediated endocytosis has been shown to play role in DENV internalization in both mammalian and insect cells [48, 51–53]. It is possible that viruses have evolved to use abundant LDL and LRP receptors to gain ubiquitous access to human tissues. These receptors provide classical examples for the cargo internalization via clathrin-mediated endocytosis; they are also glycosylated and hence may provide GAG-mediated initial attachment of the virion which facilitates specific interaction with the receptor.
DENVs are known to infect a variety of tissues  in vivo and numerous human cell types in vitro including epithelial cells  and HeLa cells . Dendritic cells (DCs) and macrophages are the primary targets for DENV infection in human blood [57, 58]. Permissive monocytes, antigen presenting cells, ingest and process the viral particle prior to presenting its components, proteins and nucleic acid, to immune response cells, T-cells, in the lymph nodes [59, 60]. In DCs, dengue viruses are recognized by a C-type lectin, DC-specific intercellular adhesion molecule-3 grabbing nonintegrin (DC-SIGN) [57, 58]. DC-SIGN also recognizes the arenavirus Lassa virus  and the Enterovirus 71 .
When the mosquito injects its beveled stinger (proboscis) into the epidermis it deposits a mixture of salivary proteins along with the DENV particles. Keratinocytes that can express the LDLR  and LRP1a  and make up 95% of the epidermis may represent the first line of defense against arboviruses. Only a small fraction of the DENV injected by the mosquito enters the blood compartment and is addressed by the dendritic cells . The rest may remain with the tissues at the injection site where the keratinocytes stimulate inflammation and activate Langerhans cells . By comparison, only chemically modified forms of LDL, for example, oxidized-LDL, are engulfed and removed from the blood by DC-scavenger receptors ; native LDL is removed from circulation by various tissues expressing LDLR and LRPs.
The hypothesis of DENV specific interaction to LDL or related receptor is strongly supported by sequence comparison of viral proteins to known LDLR/LRP ligands. In the Apo E receptor ligand site (0130Thr-Ala0159; PDB: 1LPE), Lysine residues in positions 0143 and 0146, 0141LRKLRKRLLR0150, are essential for binding LDL B/E receptors  and Arginine residues located between 0170Ala-Leu0181 are involved in binding GAGs and heparan sulfate proteoglycans, HSPGs . Our examination of the DENV polyprotein sequences revealed that potential ligand analogues occur in structural as well as in nonstructural proteins. In DENV, the envelope glycoproteins that span residues 0281Met-Ala0775 make the initial contact with receptor molecules on the cell surface . We posited that this interaction involves the LDLR Lysine/Arginine-rich ligand motifs contained in the DENV envelope and capsid proteins.
In DENV envelope proteins, the region essential for cell entry, Domain III, spans 0580Ser-Trp0671 . Studies by Falconar  showed that a neutralizing monoclonal 3A8.1 has a high binding avidity for several DENV2 E protein-derived peptides, including 0568RMDKLQLKG0576 and 0584GKFKIVKEI0592 located in E protein Domains I and III, respectively . Both peptides contain LDLR Type III ligand motifs (Table 1); analogues are present as 0567LKMDKL in DENV1 and 3, as 0567L/VRMD/EKL0572 in DENV2 and DENV4, and as 0586FKLE/VKE0591 in DENV1 and 2 (Table 4). The latter motif is absent in DENV3 and 4. Instead, DENV4 E protein has an analogue of the motif as 0478LMKMKKKT0485. In the PBD model 1OAN of the DENV2, both motifs in tandem appear in a loop located on the surface of the molecule.
We used fluorescence-labeled synthetic peptides of DENV sequences containing LDLR/LRP ligand motifs in dual label experiments with Apo E CM-DiI and LDL CM-DiI. This approach presents advantages over using an intact virus including cost issues, safety issues, and steric hindrance. As illustrated in Results, our Dsp2EP-FITC peptide that spans DENV2 region 0564Lys-0595 and contains the two Type III ligand motifs binds HeLa cells but is not internalized. Dsp2EP-FITC also binds HepG2 cells (shown as a Supplemental Figure 1 available online at http://dx.doi.org/10.1155/2015/646303). Dsp2EP-FITC (B) and (C), mutated forms of Dsp2EP-FITC, bind HeLa cells with lower integrated density (Figure 2(b)). Similar binding pattern (decreased binding of the mutated peptides compared to the nonmodified version) was observed with dextran sulfate cellulose beads (Supplemental Figure 2). These results confirm that residues Arg0568, Lys0571, Lys0587 and Lys0590 in LDLR ligand motifs play a role in binding to the cell surface. Colocalization of Apo E-CM-DiI and Dsp2EP-FITC suggests that they may bind the same receptor or adjacent molecules. The LDL receptor contains 7 copies of the ligand modules, and Apo E binds only LA4 and 5 modules . Hence, presence of Apo E-FITC may not hinder binding of Dsp2EP-FITC to the remaining modules and would result in colocalized red and green signals. Further, involvement of other LRPs should not be ruled out.
Sequence similarities exhibited by human Apo E, Apo B, and DENV capsid proteins were first described in our earlier report . The 3D models of Apo E  and DENV2 capsid protein  show that both are helix-loop/turn-helix molecules with amphipathic features that encourage homodimerization in a pure state and perhaps heterodimerization in a mixture. These common properties may account for DENV capsid protein specific interaction with VLDL . All three proteins associate with lipids, bind nucleic acids, possess cell entry capacity, and translocate to the cell nucleus.
The primary structures of DENV capsid proteins suggest them to be multifunctional proteins with analogues of LDLR and LRP ligand and NLS motifs at both termini [72, 73]. Based on our studies, DENV3 capsid may have inherent cell entry capacity. LRP ligand analogues occur within the first 40 residues of the N-termini of DENV capsid proteins. In DENV3 capsid protein two ligand motifs, RKKTGK and LKRVRN, two mono-NLS (0005Arg-Lys0010, 0017Lys-Arg0020), and one bi-NLS (0006Lys-Arg0020) motifs likely provide the capsid not only with cell entry capacity but also with potential to transit into the nucleus. DENV2 C protein, Type II ligand analogue, 0004QRKKAKN0010, is also a high affinity LRP ligand. DENV1 and 4 C proteins, 0004QRKKTGRP0011 and 0003QRKKVVRP0010, respectively, are Type I ligand motifs and may impart both LRP/GAG binding and nuclear translocation potentials. Our results and those reported by others  support the premise that the intact DENV capsids have inherent infectivity capacity and may be an important player in the second phase of the DENV infection process as an unfinished/immature viral particle.
Considering that LDLR/LRP receptors occur ubiquitously in humans, sylvatic mammals, and mosquitoes, Aedes sp. and Culex sp., and that the receptor ligand motifs are mimicked by DENVs , it is possible that DENV possesses a “pass” into virtually any human cell type. Then vascular leakage of the virus in DHF syndrome  could lead to infection of vital organs such as liver, spleen, lungs, heart, and kidneys. High level of infectivity provided by the LDLR/LRP mechanism may be a major factor in dengue shock syndrome and dengue hemorrhagic fever.
We conclude that dengue viruses may have the capacity to achieve cell binding and entry via the receptors utilized by low density lipoproteins. Our results indicate that dengue capsid and envelope proteins contain Lysine-based LDL receptor ligand and NLS motifs. Synthetic peptides of DENV proteins, representing potential LDL receptor binding sites, colocalize with LDL and Apo E on the cell surface. In humans and other mammals LDL receptors are expressed ubiquitously throughout the organism, a factor that may greatly enhance viral infectivity. The processes for cell entry for dengue virus have not been fully elaborated, and our report provides an additional intriguing possibility for molecular mechanisms of viral entry.
|CP:||C protein, capsid protein|
|Cm-DiI:||Cell Tracker C7000|
|EP:||E protein, envelope protein|
|HSPG:||Heparin sulfate proteoglycan|
|IDL:||Intermediate density lipoprotein|
|LDL:||Low density lipoprotein|
|LDLR:||Low density lipoprotein receptor|
|LLP(s):||Low density lipoprotein particle(s)|
|NLS:||Nuclear localization signal|
|VLDL:||Very low density lipoprotein.|
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
Studies were supported in part by Grants F49620-99-1-0327 (AFOSR, PI, Juan Guevara Jr.), FA 9550-05-1-0472 (AFOSR, PI, A. Hanke), SC2GM081218 and SC3GM099637 (NIH, PI, Natalia Valentinova Guevara). Mr. Troy McWhorter donated his time and effort on this project. Mr. Jamie Romo Jr. was supported by R25GM083755 (NIH, PIs, J. Facelli, S. Nair).
Basic amino acid clusters in motifs, XBBBXXBX, XBBXBX, and ΨBΨXB, in apo E and apo B100 of the low-density lipoproteins, are known to bind the LDL receptors, heparin, and highly sulfated polysaccharides also occur in the polyproteins sequences of the Dengue viruses 1-4. The ΨBΨXB motif occurs in Domain III, the putative receptor binding region of the DENV envelope glycoproteins, twice in DENV 1 and 2, and once in DENV 3 and 4. The DENV2 synthetic peptide Dsp2EP-FITC that encompasses residues 0564Gly-Gly0595 in the envelope protein region of the polyprotein and DENV3 Dsp3CP-FITC, 0002Asn-Gln0028, in the N-terminal capsid region of the DENV3 polyprotein were used to assess binding to Hep G2 cells and Liposorber Dextran Sulfate Cellulose Affinity beads of the Lipopheresis System.
E. V. Orlova, M. B. Sherman, W. Chiu, H. Mowri, L. C. Smith, and A. M. Gotto Jr., “Three-dimensional structure of low density lipoproteins by electron cryomicroscopy,” Proceedings of the National Academy of Sciences of the United States of America, vol. 96, no. 15, pp. 8420–8425, 1999.View at: Publisher Site | Google Scholar
B. D. Lindenbach, H.-J. Thiel, and C. M. Rice, “Flaviviridae: the viruses and their replication,” in Fields Virology, D. M. Knipe and P. M. Howley, Eds., chapter 33, pp. 1101–1152, Lippincott Williams & Wilkins, Philadelphia, Pa, USA, 2006.View at: Google Scholar
D. J. Gubler, G. Kuno, and L. Markoff, “Flaviviruses,” in Fields Virology, D. M. Knipe and P. M. Howley, Eds., chapter 34, pp. 1153–1252, Lippincott Williams & Wilkins, Philadelphia, Pa, USA, 2006.View at: Google Scholar
J. Callow, L. K. M. Summers, H. Bradshaw, and K. N. Frayn, “Changes in LDL particle composition after the consumption of meals containing different amounts and types of fat,” The American Journal of Clinical Nutrition, vol. 76, no. 2, pp. 345–350, 2002.View at: Google Scholar
J. L. Goldstein and M. S. Brown, “History of discovery: the LDL receptor,” Arteriosclerosis, Thrombosis, and Vascular Biology, vol. 29, pp. 431–438, 2009.View at: Google Scholar
C. A. Dyer, D. P. Cistola, G. C. Parry, and L. K. Curtiss, “Structural features of synthetic peptides of apolipoprotein E that bind the LDL receptor,” The Journal of Lipid Research, vol. 36, no. 1, pp. 80–88, 1995.View at: Google Scholar
A. V. Hospattankar, S. W. Law, K. Lackner, and H. B. Brewer Jr., “Identification of low density lipoprotein receptor binding domains of human apolipoprotein B-100: a proposed consensus LDL receptor binding sequence of apoB-100,” Biochemical and Biophysical Research Communications, vol. 139, no. 3, pp. 1078–1085, 1986.View at: Publisher Site | Google Scholar
A. D. Murdoch, G. R. Dodge, I. Cohen, R. S. Tuan, and R. V. Iozzo, “Primary structure of the human heparan sulfate proteoglycan from basement membrane (HSPG2/perlecan). A chimeric molecule with multiple domains homologous to the low density lipoprotein receptor, laminin, neural cell adhesion molecules, and epidermal growth factor,” The Journal of Biological Chemistry, vol. 267, no. 12, pp. 8544–8557, 1992.View at: Google Scholar
M. Zaiou, K. S. Arnold, Y. M. Newhouse et al., “Apolipoprotein E-low density lipoprotein receptor interaction: influences of basic residue and amphipathic α-helix organization in the ligand,” Journal of Lipid Research, vol. 41, no. 7, pp. 1087–1095, 2000.View at: Google Scholar
C. A. Dyer and L. K. Curtiss, “A synthetic peptide mimic of plasma apolipoprotein E that binds the LDL receptor,” The Journal of Biological Chemistry, vol. 266, no. 34, pp. 22803–22806, 1991.View at: Google Scholar
I. R. Nikoulin and L. K. Curtiss, “An apolipoprotein E synthetic peptide targets to lipoproteins in plasma and mediates both cellular lipoprotein interactions in vitro and acute clearance of cholesterol-rich lipoproteins in vivo,” The Journal of Clinical Investigation, vol. 101, no. 1, pp. 223–234, 1998.View at: Publisher Site | Google Scholar
S. D. Roosendaal, J. M. van Doorn, K. M. Valentijn, D. J. van der Horst, and K. W. Rodenburg, “Delipidation of insect lipoprotein, lipophorin, affects its binding to the lipophorin receptor, LpR: implications for the role of LpR-mediated endocytosis,” Insect Biochemistry and Molecular Biology, vol. 39, no. 2, pp. 135–144, 2009.View at: Publisher Site | Google Scholar
H.-M. Cheon, W. S. Sang, G. Bian, J.-H. Park, and A. S. Raikhel, “Regulation of lipid metabolism genes, lipid carrier protein lipophorin, and its receptor during immune challenge in the mosquito Aedes aegypti,” The Journal of Biological Chemistry, vol. 281, no. 13, pp. 8426–8435, 2006.View at: Publisher Site | Google Scholar
K. W. Rodenburg and D. J. van der Horst, “Review. Lipoprotein-mediated lipid transport in insects: analogy to the mammalian lipid carrier system and novel concepts for the functioning of LDL receptor family members,” Biochimica et Biophysica Acta, vol. 1736, no. 1, pp. 10–29, 2005.View at: Publisher Site | Google Scholar
P. W. Atkinson, J. Hemingway, B. M. Christensen et al., “Annotation of Culex pipiens quinquefasciatus,” PUBMED: Accession: XP_001864948, 2009.View at: Google Scholar
H. M. Cheon, S. J. Seo, J. Sun, T. W. Sappington, and A. S. Raikhel, “Molecular characterization of the VLDL receptor homolog mediating binding of lipophorin in oocyte of the mosquito Aedes aegypti,” Insect Biochemistry and Molecular Biology, vol. 31, no. 8, pp. 753–760, 2001.View at: Publisher Site | Google Scholar
B. A. Kumar and K. P. Paily, “Up-regulation of lipophorin (Lp) and lipophorin receptor (LpR) gene in the mosquito, Culex quinquefasciatus (Diptera: Culicidae), infected with the filarial parasite, Wuchereria bancrofti (Spirurida: Onchocercidae),” Parasitology Research, vol. 108, no. 2, pp. 377–381, 2011.View at: Publisher Site | Google Scholar
A. Cabrera-Hernandez and D. R. Smith, “Mammalian dengue virus receptors,” Dengue Bulletin, vol. 29, pp. 119–135, 2005.View at: Google Scholar
P.-Y. Lozach, L. Burleigh, I. Staropoli et al., “Dendritic cell-specific intercellular adhesion molecule 3-grabbing non-integrin (DC-SIGN)-mediated enhancement of dengue virus infection is independent of DC-SIGN internalization signals,” Journal of Biological Chemistry, vol. 280, no. 25, pp. 23698–23708, 2005.View at: Publisher Site | Google Scholar
M. L. Williams, A. M. Mommaas-Kienhuis, S. L. Rutherford, S. Grayson, B. J. Vermeer, and P. M. Elias, “Free sterol metabolism and low density lipoprotein receptor expression as differentiation markers of cultured human keratinocytes,” Journal of Cellular Physiology, vol. 132, no. 3, pp. 428–440, 1987.View at: Publisher Site | Google Scholar
M.-F. Galliano, E. Toulza, N. Jonca, S. L. Gonias, G. Serre, and M. Guerrin, “Binding of α2ML1 to the low density lipoprotein receptor-related protein 1 (LRP1) reveals a new role for LRP1 in the human epidermis,” PLoS ONE, vol. 3, no. 7, Article ID e2729, 2008.View at: Publisher Site | Google Scholar
Y. Chen, T. Maguire, and R. M. Marks, “Demonstration of binding of dengue virus envelope protein to target cells,” Journal of Virology, vol. 70, no. 12, pp. 8765–8772, 1996.View at: Google Scholar
A. K. I. Falconar, “Use of synthetic peptides to represent surface-exposed epitopes defined by neutralizing dengue complex- and flavivirus group-reactive monoclonal antibodies on the native dengue type-2 virus envelope glycoprotein,” Journal of General Virology, vol. 89, no. 7, pp. 1616–1621, 2008.View at: Publisher Site | Google Scholar
S.-H. Wang, W.-J. Syu, K.-J. Huang et al., “Intracellular localization and determination of a nuclear localization signal of the core protein of dengue virus,” Journal of General Virology, vol. 83, no. 12, pp. 3093–3102, 2002.View at: Google Scholar
C. A. Balinsky, H. Schmeisser, S. Ganesan, K. Singh, T. C. Pierson, and K. C. Zoon, “Nucleolin interacts with the dengue virus capsid protein and plays a role in formation of infectious virus particles,” Journal of Virology, vol. 87, no. 24, pp. 13094–13106, 2013.View at: Publisher Site | Google Scholar
D. T. Trung and B. Wills, “Systemic vascular leakage associated with dengue infections—The Clinical Perspective,” in Dengue Virus, A. L. Rothman, Ed., vol. 338 of Current Topics in Microbiology and Immunology, pp. 57–66, Springer, Berlin, Germany, 2010.View at: Google Scholar