OUP user menu

Gene transfer using human papillomavirus pseudovirions varies according to virus genotype and requires cell surface heparan sulfate

Alba Lucia Combita , Antoine Touzé , Latifa Bousarghin , Pierre-Yves Sizaret , Nubia Muñoz , Pierre Coursaget
DOI: http://dx.doi.org/10.1111/j.1574-6968.2001.tb10883.x 183-188 First published online: 1 October 2001


Artificial viruses consisting of DNA plasmid packaged in vitro into virus-like particles (VLPs) are new vehicles for gene transfer. We therefore investigated the ability of nine human papillomavirus (HPV) VLPs to interact with heterologous DNA and transfer genes. HPV 16, 18, 31, 33, 39, 45, 58, 59, and 68 VLPs were able to bind heterologous DNA and to transfer genes into Cos-7 cells. Inhibition of gene transfer by preincubation of the pseudovirions with heparin confirmed that heparan sulfate on the cell surface plays a role as cell receptor for HPVs. As HPV neutralizing antibodies are mainly type-specific, gene transfer with different HPV pseudovirions offers the possibility of their sequential use in vivo for a greater efficacy.

Key words
  • Human papillomavirus
  • Virus-like particle
  • Pseudovirion
  • Gene transfer
  • Direct interaction
  • Glycosaminoglycan

1 Introduction

The use of human papillomavirus (HPV) or polyomavirus virus-like particles (VLPs) as carriers for gene transfer has recently been described[17]. These pseudovirions, consisting of plasmid DNA packaged in vitro into VLPs, mimic viral methods of DNA delivery such as DNA packaging, endosomal escape, and nuclear transport, while the absence of viral genome eliminates the risks. Pseudovirions suffer from the same disadvantages as replication-deficient recombinant viruses in that they induce the production of neutralizing antibodies. More than 100 human papillomavirus types have been described, and recombinant VLPs induce a virus-neutralizing antibody response that has been shown to be predominantly type-specific [8,9]. Thus, if artificial viruses could be obtained from different HPVs, their sequential use would offer the possibility of repeated induction of efficient gene transfer.

Papillomaviruses are small non-enveloped tumorigenic DNA viruses that infect the basal cells of the epithelium. HPVs infecting the cutaneous epithelium such as types 1, 2, and 4 mainly cause skin warts. Those infecting the mucosal epithelium such as types 6 and 11 cause benign condyloma and low grade squamous intraepithelial lesions, and types 16, 18, 31, 33, 39, 45, 58 and 59 are strongly associated with premalignant lesions and genital carcinomas [10].

The viral capsid is composed of major L1 and minor L2 structural proteins. The major capsid protein alone can self-assemble into VLPs when expressed in eukaryotic systems [11]. These VLPs represent the antigen of choice for both the development of serological assays and prophylactic vaccines. Attempts to grow HPVs efficiently in cell culture have to date been unsuccessful, and consequently surrogate systems using HPV VLPs have been developed to test for viral infectivity [14,7]. These assays have led to the identification of putative receptors. The binding receptor for HPV 6 was suggested to be the α-6 integrin [12,13]. On the other hand, gene transfer by HPV 11, 16 and 31 pseudovirions has been shown to be inhibited by heparin or by removing heparan sulfates from the cell surface [14].

We have previously demonstrated that HPV VLP 16 self-assembled from L1 protein can encapsidate heterologous DNA in vitro. The packaged DNA could be transferred to cells as shown by the expression of a reporter gene into mammalian cells [3]. Gene transfer has also been reported for HPV 11, 16, 18 and 33 L1/L2 VLPs [1,2,4,15]. To dispose of a range of vectors that could be used sequentially in vivo, we investigated five other types of HPVs, namely types 31, 39, 45, 58, 59 and 68, for gene transfer. In addition, the nine pseudovirions produced were used to study the role of proteoglycans in papillomavirus cell entry.

2 Materials and methods

2.1 Protein expression and purification

Recombinant HPV L1 VLPs were prepared as previously described for HPV 16 [17]. Briefly, HPV L1 coding sequences from HPV 18, 31, 33, 39, 58 and 59 were amplified by PCR from cervical biopsies from patients with invasive cervical cancer from Thailand and Colombia [18,19]. HPV 45 L1 was amplified from a PCR product from a patient from the Ivory Coast with squamous cell carcinoma [20] and HPV 68 L1 from a HPV genome cloned from a French patient with low grade SIL [21]. For amplification of HPV 31 forward primer 5′-ATA TTT TGG ATC CGA TGT CTC TGT G-3′ and reverse primer 5′-CAT ACA CAA GCT TTA CTT TTT AGT TTT T-3′ were used. For amplification of HPV 18 and 45, forward primers 5′-CTT ATT AGA TCT CAG ATG GCT TTG TGG CGG-3′ and 5′-CCT ATT AGA TCT CAG ATG GCT TTG T-3′ and reverse primers 5′-ATA CAC AGA TCT ATA TTA CTT CCT GGC ACG-3′ and 5′-GCT AAG CTT TTA TTT CTT ACT ACG TAT-3′ were used, respectively. For the amplification of HPV 33 and 39 the forward and reverse primers were 5′-CGT TTT GGA TCC TTT TTT ACA GAT GTC CGT-3′ and 5′-AAC AAC AAG CTT ACA CAA TTA CAC AAA GTG-3′, and 5′-C CTA TTT GGA TCC AGA TGG CTA TGT G and 5′-GCA TAC ACA AGC TTT ATT TAG ACA CAC GTT-3′, respectively. Forward primers 5′-ATA TTT GGA TCC AGA TGT CCG TGT GGC-3′ and 5′-CC TAT TTT GGA TCC GAT GGC TCT ATG-3′, and reverse primers 5′-CCA AAG CTT TTA TTT TTT AAC CTT TTT GCG-3′ and 5′-CAT AAC AAG CTT CAC TAT TTT CTG GAA GAC-3′ were used to amplify HPV 58 and 59 L1 genes, respectively. Finally, primers 5′-AGA TCT GAT GGC ATT GTG GCG CTC TAG-3′ and 5′-GAA TTC TTA CTT TGA CAC ACG TTT ACG TTT G-3′ were used to subclone the HPV 68 L1 gene.

Following amplification PCR products were cloned into the pCR2.1 vector (TOPO TA cloning, Invitrogen). The L1 genes cloned into pCR2.1 plasmid were sequenced using an ABI PRISM 310 automated sequencing system using Big Dye terminators (Applied Biosystems, Courtaboeuf, France), and M13 forward and M13-21 primers and four type-specific L1 gene internal primers. The L1 genes from HPV 16, 18, 31 and 39 were subcloned into pBlueBacIII vector (Invitrogen), downstream from the polyhedrin promoter. The resulting transfer vectors were cotransfected into Sf21 cells along with linearized Autographa californica multiple nuclear polyhedrosis virus DNA (Linear AcMNPV transfection module, Invitrogen). The HPV 33, 45, 58, 59 and 68 L1 sequences were subcloned into pFastBacI (Life Technologies, Cergy Pontoise, France). Escherichia coli DH10Bac (Life Technologies) were transformed with pFastBacI 33, 45, 58 and 59 L1, respectively. Recombinant baculoviruses were generated and selected as recommended by the respective manufacturers.

Recombinant baculoviruses were used to infect Sf21 cells, and VLPs were purified by ultracentrifugation in CsCl gradients according to a previously described procedure [17]. Each preparation was tested for the presence of virus-like particles by electron microscopy. For this purpose, VLP preparations were applied to 400-mesh carbon-coated grids, negatively stained with 1.5% uranyl acetate and then examined at a nominal magnification of 50 000 with a JEOL 1010 electron microscope. For quantification of the VLPs, each preparation was mixed with an equal volume of a standard suspension of 10-nm diameter gold beads conjugated to anti-rabbit IgG (British BioCell International, Cardiff, UK) and then examined by electron microscopy. The results were expressed as the ratio of the number of VLPs per gold bead in four randomly chosen fields.

Relative amounts of L1 protein were assessed by SDS–PAGE. After Coomassie blue staining the results were analyzed by densitometry using Molecular Analyst software (Bio-Rad, Ivry/Seine, France).

2.2 Determination of DNA binding activity

DNA binding of the different L1 was assessed by South-Western blot assay as described previously [22]. Briefly, purified VLPs were subjected to 10% SDS–PAGE before transfer to BA 83 nitrocellulose by electroblotting (Hoefer Semifor; Pharmacia). The DNA binding assay was performed by 30 min incubation at room temperature with a digoxigenin-labeled DNA probe. This probe was obtained by PCR in the presence of DIG-11 dUTP (Roche Diagnostics, Meylan, France). The membranes were then washed with binding buffer and bound DNA was revealed with an anti-digoxigenin alkaline phosphatase-conjugated antibody (Roche Diagnostics) using NBT and BCIP as substrates.

VLP/DNA interaction was further characterized by gel retardation assays. Complexes obtained by direct interaction (see below) were analyzed by electrophoresis on 1% agarose gel followed by ethidium bromide staining. Hepatitis B core particles obtained by expression in insect cells of the Core gene deleted of its 39 C-terminal amino acids (HBcΔ) was used as control [23].

2.3 VLP/DNA complex formation

Three methods were used for the formation of VLP/DNA complexes: disassembly–reassembly, osmotic shock and direct interaction. The disassembly–reassembly of VLPs was performed according to a previously described procedure with some modifications [3]. Ten micrograms of purified VLPs were incubated in 50 mM Tris–HCl buffer (pH 7.5) containing 150 mM NaCl, 20 mM EGTA and 5 mM DTT at a final volume of 50 μl at room temperature for 30 min. At this stage, 1 μg of pCMV-Luciferase plasmid (7.2 kb) in 50 mM Tris–HCl buffer (pH 7.5) and 150 mM NaCl was added to the disrupted VLPs in the presence of 1% DMSO. The preparation was then diluted in 50 mM Tris–HCl buffer (pH 7.5) and 150 mM NaCl. CaCl2 molarity was increased stepwise from 0 to 6 mM with an increment of 1 mM h−1 at 20°C to obtain a final volume of 125 μl.

VLP/DNA complexes were also obtained by osmotic shock according to the method described by Barr et al. [24] for polyomavirus, with some modifications. Ten micrograms of VLPs and 1 μg of plasmid DNA were mixed in 150 mM NaCl, 10 mM Tris–HCl (pH 7.5) and 0.01 mM CaCl2. After 10 min at 37°C, the mixture was subjected to osmotic shock by dilution in 4 vols. of distilled water and incubated for 20 min at 37°C. In the direct interaction method 10 μg of VLPs and 1 μg of DNA were mixed in 80 mM NaCl (pH 6) and incubated for 30 min at room temperature.

In order to evaluate the proportion of packaged DNA, the pseudovirions were treated with 10 IU benzonase (Merck, Darmstadt, Germany) for 1 h at 20°C. After benzonase treatment, the mixture was incubated in the presence of 3% SDS and 1 mg ml−1 of proteinase K (Appligene, Illkirch, France) for 2 h at 56°C. The same experiments, without benzonase treatment, were conducted as controls. Plasmid DNA was then phenol-extracted and ethanol-precipitated. Purified DNA was linearized using EcoRI restriction enzyme and electrophoresed in a 1% agarose gel. The percentage of protected plasmid was determined by densitometry as above.

2.4 Transfection experiments

Cos-7 cells grown in monolayer in D-MEM/Glutamax (Life Technologies) supplemented with 10% fetal calf serum (FCS), 100 IU ml−1 penicillin and 100 μg ml−1 streptomycin were seeded in 48-well plates (Nunc, Life Technologies) and grown to 80% confluence. Cells were washed twice with D-MEM/Glutamax. VLP/DNA complexes (10 μg/1 μg) were diluted to 200 μl in culture medium and added to each well. After incubation for 3 h at 37°C, 2 ml of D-MEM/Glutamax supplemented with 10% FCS were added. The cells were then incubated for 48 h at 37°C. Luciferase activity was determined using the luciferase reporter gene assay with constant light signal (Roche). The results were expressed as count per second (cps) per well.

3 Results and discussion

High relative levels of L1 protein were observed with HPV 16, 31, 33, 45 and 68. In contrast, low levels of L1 were detected with HPV 18, 39, 58 and 59 (Table 1). Virus-like particles with a diameter of approx. 50 nm were observed by electron microscopy for all the nine HPV types (16, 18, 31, 33, 39, 45, 58, 59 and 68) expressed in insect cells (data not schown). Numerous VLPs were observed for HPV 16, 18, 31, 58 and 68, but few VLPs were seen for HPV 33, 39, 45 and 59. It should be noted that for HPV 33 only very few VLPs were observed in comparison to the high level of L1 protein detected. Moreover, well-formed and regular VLPs were observed for types 16, 18, 31, 45 and 68, and mainly irregular VLPs for types 33, 39, 58 and 59. All HPV L1 proteins were able to bind DNA as determined by South-Western blotting (Fig. 1, Table 1) and gel retardation assay (Table 1).

View this table:

Quantification of VLPs and L1 proteins, DNA binding and gene transfer for nine HPV VLPs after direct interaction between VLPs and DNA

VLPs (VLP/gold beads)L1 protein (relative amount)DNA bindingGene transfer (luciferase activity, cps/well)
South-WesternDNA retardation assay
HPV 162.04.1++22 637
HPV 181.23.6++807
HPV 311.73.4++19 963
HPV 330.23.4++45
HPV 390.61.0++74
HPV 451.02.0++7 963
HPV 580.75.2++6 205
HPV 590.43.9++2 366
HPV 681.94.0++10 923

Detection of DNA binding to HPV 16, 18, 31, 39, 58 and 59 L1 proteins by South-Western blotting using digoxigenin-labeled DNA probe. BSA, bovine serum albumin.

We investigated gene transfer with HPV 16, 31, 45, 58 and 59 VLPs according to three methods: disassembly–reassembly, osmotic shock or direct interaction. The results indicate that high levels of gene transfer were observed for all types with osmotic shock and direct interaction compared to the disassembly–reassembly technique (Fig. 2). The highest level of gene transfer was always obtained with pseudovirions generated by direct interaction of DNA and VLPs. No gene transfer was detected with HPV pseudovirion types 45 and 59 generated by disassembly–reassembly of VLPs. Thus, the direct interaction method was used for subsequent investigations.


Gene transfer with HPV 16, HPV 31, HPV 45, HPV 58 and HPV 59 pseudovirions obtained by disassembly–reassembly (black bar), osmotic shock (hatched bar) or direct interaction (dotted bar).

All nine oncogenic HPV pseudovirions investigated were able to transfer the luciferase gene into Cos-7 cells when generated by the direct interaction method (Table 1 and Fig. 3). The highest levels of gene expression were detected with HPV 16, 31 and 68 pseudovirions. Lower efficiency was observed for types 45, 58, 59 and 18. Only weak gene transfer, but statistically different, from DNA alone was detected after transfection with HPV type 33 and 39 pseudovirions (P=0.002 and P=0.025, respectively; F test). Variation in gene transfer could be due to differences in the level of pseudovirions generated, but also to different efficiencies linked to the VLP type or to the presence of mutations in the L1 protein. To date, VLPs have been interpreted as having the ability to bind and encapsidate DNA, to express type-specific epitopes [14,8,9] and to bind to cell receptors [1214]. However, it has been observed that mutations can exist or be introduced in the process to obtain recombinant vectors which abrogate type-specific epitopes [25] but do not affect VLP formation. Similarly, one can imagine that point mutations which do not affect VLP formation, would inhibit or reduce gene transfer capacities. However, no amino acid change was observed in HPV-33 and HPV-39 L1 sequences compared to the corresponding prototypes.


Gene transfer with pseudovirions obtained by direct interaction of DNA with different oncogenic HPVs with DNA.

In order to investigate if the DNA is internalized within the VLPs, we have treated HPV 16 and 31 pseudovirions with benzonase according to the three methods used. The results (Table 3) indicate that protection was observed with all three methods and that the highest levels of protection were observed with the direct interaction method and that they correlate with the highest levels of gene transfer. However, it has been shown by Stokrova et al. [26] that with the osmotic shock method DNA could be observed by electron microscopy at the exterior of the polyomavirus VLPs, and these authors concluded that only a proportion of the DNA molecule is internalized. In addition the specificity of the gene transfer with HPV-16 and HPV-31 pseudovirions generated by the direct interaction method was established by preincubation with mouse anti-HPV-16 and anti-HPV-31 VLP antisera (diluted 1/1000) before Cos-7 cell transfection. Under these conditions, gene transfer is totally abolished by homotypic antibodies but not by heterotypic antibodies.

View this table:

DNA protection from benzonase treatment according to methods used to generate HPV 16 and HPV 31 pseudovirions

VLPsDisassembly–reassemblyOsmotic shockDirect interaction
HPV 1631%45%55%
HPV 3145%53%67%

Glycosaminoglycans have been proposed as cell receptors for HPV 11, 16, and 33 [14,16], and virus sequences interacting with these cell components have been identified at the C-terminus of the L1 protein for HPV 11 and 16 [14]. To analyze whether VLPs of other HPVs exhibit the same binding to glycosaminoglycans, we investigated the effect of heparin on HPV pseudoinfection. Pseudovirions were preincubated with heparin and subsequently added to Cos-7 cells. Heparin at a concentration of 5 μg ml−1 (around 120 heparin molecules/VLP) completely inhibited the infectivity of HPV 16, 18, 31, 45, 58, 59 and 68 pseudovirions (Table 2). Sodium chlorate has been shown to reduce proteoglycan sulfation and to affect gene transfer by HPV 11, 16 and 33 [14,16]. We observed similar results for HPV 16 and 31 (data not shown), confirming that heparan sulfate present on the cell surface is required for cellular entry of HPVs.

View this table:

Inhibition of gene transfer into Cos-7 cells by HPV pseudovirions 16, 18, 31, 45, 58, 59 and 68 by preincubation with heparin (0.5 U)

VLPsGene transfer inhibition by heparin (%)
HPV 1699
HPV 1896
HPV 31100
HPV 4592
HPV 5898
HPV 5997
HPV 6894

In conclusion, the production of artificial virus vectors for various types of human papillomavirus VLPs offers the possibility of repeated induction of gene transfer by the sequential use of these different vectors. This possibility of performing numerous injections has important advantages compared to the use of recombinant virus vectors for which limitation in the number of injections is one of the major drawbacks. Our results suggest that papillomavirus VLPs are promising vehicles to deliver genes into cells.


We thank G. Orth (Institut Pasteur, Paris) for the HPV 45 L1 gene and HPV 68 clone. This work was supported by grants from Biotechnocentre, the Association pour la Recherche contre le Cancer (No. 9977) and from the Association ‘Vaincre la Mucoviscidose’. A.L.C. was the recipient of a fellowship from COLCIENCIAS and L.B. from the Région Centre.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
View Abstract