The EMBO Journal Vol. 19 No. 13 pp. 3436±3445, 2000
Ef®cient intracellular retrotransposition of an exogenous primate retrovirus genome
Martin Heinkelein1, Thomas Pietschmann1, Gergely JaÂrmy1, Marco Dressler1, Horst Imrich1, Jana Thurow1, Dirk Lindemann1, Michael Bock1,2, Astrid Moebes1,3, Jacqueline Roy4, Ottmar HerchenroÈder4 and Axel Rethwilm1,4,5 1
Institut fuÈr Virologie und Immunbiologie, UniversitaÈt WuÈrzburg and Institut fuÈr Virologie, Medizinische FakultaÈt `Carl Gustav Carus', Technische UniversitaÈt Dresden, Gerichtsstrasse 5, 01069 Dresden, Germany 4
2 Present address: National Institute for Medical Research, Division of Virology, Mill Hill, London NW7 1AA, UK 3 Present address: Department Innere Medizin II/Molekularbiologie, UniversitaÈtsklinik Freiburg, Germany 5 Corresponding author e-mail:
[email protected]
The foamy virus (FV) subgroup of Retroviridae reverse transcribe their RNA (pre-)genome late in the replication cycle before leaving an infected cell. We studied whether a marker gene-transducing FV vector is able to shuttle to the nucleus and integrate into host cell genomic DNA. While a potential intracellular retrotransposition of vectors derived from other retroviruses was below the detection limit of our assay, we found that up to 5% of cells transfected with the FV vector were stably transduced, harboring 1 to ~10 vector integrants. Generation of the integrants depended on expression of functional capsid, reverse transcriptase and integrase proteins, and did not involve an extracellular step. PCR analysis of the U3 region of the 5¢ long terminal repeat and determination of proviral integration sites showed that a reverse transcription step had taken place to generate the integrants. Co-expression of a mutated envelope allowing particle egress and avoiding extracellular infection resulted in a signi®cantly increased rescue of cells harboring integrants, suggesting that accumulation of proviruses via intracellular retrotransposition represents an integral part of the FV replication strategy. Keywords: foamy viruses/hepadnaviruses/ retrotransposition/retroviruses/reverse transcription
Introduction The eukaryotic genome is largely shaped by sequences that arose by reverse transcription of RNA intermediates (Wilkinson et al., 1994; Boeke and Stoye, 1997). The genetic structure of elements coding for gag and pol and bearing ¯anking long terminal repeats (LTRs), class I retrotransposons, resembles remarkably proviruses of 3436
exogenous retroviruses (Wilkinson et al., 1994; Boeke and Stoye, 1997). However, the mechanism by which multicopy endogenous vertebrate retroviruses were acquired is not fully understood. While eukaryotic proviral ampli®cation is unlikely to arise by DNA-mediated transposition, a mechanism involving reverse transcription of an RNA intermediate could occur either by intracellular retrotransposition or could involve an extracellular step leading to the generation of an infectious retrovirus (Boeke and Stoye, 1997). Intracellular retrotransposition in eukaryotes has mainly been studied in yeast Ty and Drosophila copia retrotransposon class I elements (Wilkinson et al., 1994; Boeke and Stoye, 1997). Following transcription of the RNA from an integrated retrotransposon, a cDNA copy is generated by particle-associated reverse transcription and is delivered into the nucleus where it may eventually integrate at another locus (Boeke and Stoye, 1997). Intracellular retrotransposition is depicted in Figure 1, pathway C. In vertebrates there is no de®nite example of an LTR-¯anked gag/pol-encoding retrotransposon. The present day endogenous retroviruses with a comparable genetic structure are believed to be derived from exogenous retroviruses which once entered the germ line (Temin, 1992; Boeke and Stoye, 1997). However, in functional terms the discrimination between retrotransposons and endogenous retroviruses is arbitrary, since both genetic elements are able to move intracellularly (Wilkinson et al., 1994; Boeke and Stoye, 1997). As far as has been analyzed, intracellular retrotransposition of endogenous retroviruses is a relatively rare event (Heidmann and Heidmann, 1991). Exogenous retroviruses are believed to be derived from retrotransposons that acquired an envelope gene (Temin, 1992; Boeke and Stoye, 1997). They usually require an extracellular phase to establish offspring proviruses, since reverse transcription of the RNA genome is only activated following infection of another cell (Vogt, 1997). The replication strategy of exogenous retroviruses is shown in Figure 1, pathway A. Primate spuma or foamy viruses (FVs) are a subgroup of exogenous retroviruses which diverge signi®cantly from the classical retroviral replication pathway (Rethwilm, 1996; Linial, 1999). The most salient difference is that the FV (pre-)genomic RNA is to a considerable extent already reverse transcribed late in the replication cycle, which leads to almost full-length linear double-stranded DNA in the viral particle (Yu et al., 1996; Moebes et al., 1997). Furthermore, studies with the reverse transcriptase (RT) inhibitor zidovudine (AZT) indicated that the functionally relevant FV genome is DNA (Moebes et al., 1997; Yu et al., 1999). The FV replication strategy is depicted in Figure 1, pathway B. The ability to generate cDNA before budding of the virion opens the possibility for FVs to ã European Molecular Biology Organization
Foamy virus intracellular retrotransposition
Fig. 1. Schematic view of the classical exogenous retroviral (pathway A) and foamy viral (pathway B) replication pathways. Pathway C is used by retrotransposons. Note that only some retroviruses, FVs and retrotransposons form intracellular capsids (Swanstrom and Wills, 1997). As in the vectors used in this study, the initial viral transcription is shown to be directed by the CMV immediate early enhancer/promoter. Following reverse transcription of the viral RNA, a 5¢ LTR is reconstituted with a copy of the U3 region.
shuttle the viral genome into the nucleus and to integrate into the host cell genome, i.e. to retrotranspose intracellularly similar to yeast retrotransposons (Figure 1, pathway C). In the present study we demonstrate formal proof of this pathway.
Results Analysis of transposition frequencies of different retroviral vectors
It has been shown previously that an env gene-deleted murine leukemia virus (MLV) genome and a marker genetagged endogenous retrovirus genome derived from a murine intracisternal A particle (IAP) are, on very rare occasions, able to retrotranspose intracellularly (Heidmann et al., 1988; Heidmann and Heidmann, 1991). The frequency with which the IAP-derived element moved to a different genetic locus was estimated as 1 transposition per 106 proviruses per cell cycle (Heidmann and Heidmann, 1991). To compare the frequencies of potential retrotranspositions by different marker gene-tagged exogenous retroviruses in a transient transfection system, we constructed the FV vector plasmid pMH92, a human
immunode®ciency virus type 1 (HIV-1)-derived lentivirus vector pGJ2/U3E, and pcAMS/U3EN, an oncoretrovirus vector derived from MLV (Figure 2A). These vectors are replication de®cient, because they lack at least one essential gene (env). Expression of the vector genomes with the untranslated leader including R-U5 regions, gag± pol genes, heterologous elements, and 3¢ LTR is initiated by the human cytomegalovirus (CMV) immediate early gene enhancer/promoter. Furthermore, in all vectors the U3 region of the 3¢ LTR was internally deleted to abrogate transcription following cDNA generation and integration. Marker gene cassettes, consisting of a constitutively active retroviral U3 promoter from spleen focus forming virus (SFFV) either directing the expression of the gene for enhanced green ¯uorescent protein (EGFP) in the HIV vector or an egfp±neomycin (neo) fusion gene in the FV and MLV vectors, replaced the deleted genes. All three vectors proved to be functionally active, since following co-transfection with appropriate env expression constructs, the cell-free supernatants contained particles that ef®ciently transduced recipient human HT1080 ®broblastoid cells as measured by ¯ow cytometry for EGFP expression (Figure 2B). While all the vectors 3437
M.Heinkelein et al.
The detection limit of protein expression by ¯ow cytometry is ~0.1%, being 3±4 log orders of magnitude less sensitive than required to determine occasional retrotransposition of env gene-deleted MLV as reported previously (Heidmann et al., 1988). Therefore, our result indicated a signi®cant difference between FVs and the representatives of other retrovirus subfamilies and suggested that the FV vector genome had ef®ciently integrated into the cellular genome. Integration could either be due to an unusual high frequency of non-speci®c pMH92 plasmid DNA integration or to retroviral integration following intracellular reverse transcription of an RNA intermediate. FV intracellular transposition involves an RNA intermediate
Fig. 2. Basic retroviral vectors and transduction rates following co-transfection with suitable envelope constructs. (A) Schematic view of the foamy, lenti- and oncoretroviral vectors. Gene expression of all vectors is directed by the CMV immediate early enhancer/promoter and transcripts start at the native cap site of R within the respective hybrid LTRs. The vectors direct the expression of wild-type gag and pol genes and, in the case of the HIV vector, of some accessory genes. However, all vectors lack at least the env gene. The 3¢ LTR had been deleted in most of its U3 region, which avoids gene expression following reverse transcription and integration. The marker gene cassette consists of the SFFV U3 region either directing the expression of an EGFP±Neo fusion protein (FV and MLV vectors) or of EGFP (HIV vector). RRE denotes the rev response element. (B) The functionality of the vectors was shown in a classical retroviral transduction experiment. 293T cells were co-transfected with vector DNA and suitable env expression constructs (FV Env with the FV vector and VSV-G with the HIV and MLV vectors) and the cell-free supernatant (1 ml) was used to transduce 104 human HT1080 ®brosarcoma cells. The numbers indicate percentages of EGFP-positive target cells 6 SE 48 h following transduction as analyzed by ¯ow cytometry. In the absence of Env, none of the plasmids produced infectious cell-free vector-virus. n.d., not done.
require an envelope to produce infectious extracellular viruses, it is noted that envelope-de®cient FV capsids do not cross cellular membranes, and thereby are unable to form virus-like particles (Fischer et al., 1998; Wang et al., 1999). In addition, FV capsids require the cognate glycoprotein for particle egress and cannot be pseudotyped by foreign Env proteins, such as MLV-derived envelope or the vesicular stomatitis virus (VSV) G protein (Pietschmann et al., 1999). HeLa and 293 cells were transiently transfected with the vectors and EGFP expression was monitored over time by ¯ow cytometry without applying selective pressure (Figure 3A). While HIV and MLV vector-transfected cells became EGFP-negative by 2±4 weeks following transfection, ~2±6% of the initially FV vector-transfected cells remained EGFP-positive for up to 50 days (Figure 3B and data not shown). The supernatant of cells harvested 48 h following transient transfection failed to transduce the marker gene to recipient HeLa and 293 cells via the extracellular route (Figure 3A). 3438
To investigate whether FV capsid formation and active FV RT and integrase (IN) proteins were involved in the generation of the EGFP-positive cells transfected with pMH92, we analyzed the pMH92 derived vectors pMH96, pMH97, pMH98 and pMH117 (Figure 4). Since in FVs Pol protein is expressed independently of Gag from a spliced mRNA (Enssle et al., 1996; Jordan et al., 1996; Yu et al., 1996), the down-mutation of the pol gene ATG start codon abrogates protease, RT/RNaseH and IN functions in pMH96 (Enssle et al., 1996). In pMH97, the catalytic center of IN was inactivated by a point mutation (DD35E to DA35E). Since FV replication strictly depends on an active integrase, the DA35E mutation abolishes FV replication (Enssle et al., 1999). In pMH98, the active center of RT was mutated (YVDD to GAAA). This mutant, in the context of an infectious molecular clone, has been described previously to be de®cient in cDNA synthesis (Moebes et al., 1997). To abrogate Gag protein expression in pMH117, the gag gene ATG start codon was changed to TTG and an in-frame stop codon (TAA) was introduced two triplets downstream thereof. Viral protein analysis of transiently transfected 293T cells revealed that pMH92, pMH97 and pMH98 expressed Gag (pr74 and p70) and Pol proteins, while only the uncleaved Gag precursor protein (pr74) but no Pol was expressed by pMH96, and only Pol proteins were expressed by pMH117 (data not shown). In addition, we included cells that were transfected with pMH92 and incubated with 5 mM AZT. Since AZT at this concentration blocks FV reverse transcription (Moebes et al., 1997), any non-speci®c plasmid DNA integration of pMH92 should be detected under these conditions. The percentages of EGFP expression in unselected HeLa and 293 cells transiently transfected with the vectors pMH96, pMH97, pMH98 and pMH117, or with the vector pMH92 in the presence of AZT declined continuously to background levels within 2±3 weeks after transfection, while 2±6% of initially pMH92-transfected cells remained EGFP-positive 4 weeks after transfection and longer (Figure 5A and B and data not shown). Cell-free supernatant transferred to naive cells 48 h following transfection was unable to transduce the marker gene (Figure 5A). Characterization of the integrants generated by intracellular retrotransposition of FV genomes
Individual cell clones were established by limiting dilution and selection in G418 from an aliquot of HeLa and 293
Foamy virus intracellular retrotransposition
Fig. 3. Analysis of potential retrotransposition by different env-deleted retroviral vectors. HeLa and 293 cells were transiently transfected with the vectors shown in Figure 2 and EGFP expression was monitored over time by ¯ow cytometry without applying selective pressure. (A) Mean percentages of positive cells 6 SE. Two days following transfection, 3 ml of cell-free supernatant (SN) were transferred to 104 recipient HeLa or 293 cells. When scored for EGFP expression, no marker gene transfer could be detected 3 days later. (B) The level of positive cells at 48 h following transfection was arbitrarily set to 100% for the individual vectors. The lines indicate the relative percentages of marker gene-expressing cells over time to demonstrate the differences between the three vectors. d.p.t., days post-transfection.
cells with the selection starting 2 and 3 weeks posttransfection, respectively. DNA extracted from several cell clones was digested with EcoRI, which cleaves pMH92 once in the 3¢ pol region, and analyzed by Southern blot hybridization to a probe derived from the gag/pol overlap region. Thus, the probe would detect the 5¢ portion of the provirus of ~5.4 kb plus upstream extensions of 5¢ cellular genomic sequences of varying length. As shown in Figure 6, between 1 and ~10 individual proviruses were detected per cell clone. Three of the cell clones designated 293/MH92 were subcloned in a second round of limiting dilution. The hybridization pattern of nine resulting subclones remained identical to the parental cell clones (data not shown). Approximately 1 in 10 G418resistant HeLa cell clones did not hybridize to the 5¢ FV probe (data not shown). A likely explanation is that 2 weeks after transfection traces of plasmid DNA remained in the HeLa cells and integrated by irregular mechanisms when G418 selection pressure was applied. Therefore, further analyses focused on cell clones reacting positively in Southern blots. 293/MH92 cell clones were analyzed by PCR. As shown in Figure 7A and B, the PCR result with forward primer #534 binding in the U3 region of the LTR and gag reverse primer #327 indicated that a reverse transcription
step had taken place to generate the U3 region of the 5¢ LTR, since we exclusively detected the indicative amplicon of ~1 kb in DNA from pMH92-transfected cells. To show unequivocally that these proviruses were generated through retroviral integration, we sequenced the junctions of cellular DNA and the 5¢ provirus of some integrants (Figure 7C). All viral sequences started with the conserved TGTGGTGG motif of the U3 region of the LTR, as reported previously for primate FV integrants (Neves et al., 1998; Enssle et al., 1999; Meiering et al., 2000), and the proviruses had integrated at different cellular sites. Taken together, via intracellular particleassociated reverse transcription a FV DNA with a 5¢ U3 end was generated from the pMH92 vector plasmid, which was imported into the nucleus and had integrated in typical fashion into the cellular genome. Analysis of FV intracellular retrotransposition in the presence of Env protein
The shuttling of integration-competent FV cDNA into the nucleus of a cell already harboring a provirus, or an equivalent in plasmid form, implies that the virus must utilize speci®c ways to regulate both the intracellular and extracellular pathway (pathways B and C, respectively, in Figure 1). Furthermore, we observed FV retrotransposition 3439
M.Heinkelein et al.
Fig. 4. FV vector constructs used to investigate whether functional Gag and Pol proteins are required for intracellular retrotransposition. Initial transcription of all viral genomes is directed by the CMV immediate early enhancer/promoter. pcHSRV2 is the replication-competent parental plasmid from which the vectors were derived. pMH96, -97, -98 and -117 are derivatives of pMH92 that bear mutations in the gag or pol reading frames. pol cannot be expressed by pMH96 due to inactivation of the ATG start codon. The catalytic center of the integrase was mutated in pMH97 by changing the second aspartic acid residue of the DD35E motif to alanine. In pMH98, the active center (YVDD) of the RT was inactivated and in pMH117 translation of the gag gene was ablated by mutating the ATG start codon.
in only a minority of the transfected cells, in which, however, several proviruses per genome were generated. This may be attributable to the assay conditions. Speci®cally, transfected cells in which FV retrotransposition occurred could have escaped detection if too many new proviruses were generated, and/or the transiently high expression level of viral Gag and Pol proteins was incompatible with cell survival. In FV infection, the presence of envelope protein is required and suf®cient to execute the budding of viral particles into intracellular vesicles or the extracellular space (Fischer et al., 1998; Heinkelein et al., 1998, 2000; Pietschmann et al., 1999, 2000). To analyze whether co-expression of Env protein has an in¯uence on the frequency of detectable intracellular retrotransposition in our assay, we co-transfected the pMH92 vector plasmid with increasing amounts of an Env expression plasmid. We then monitored the percentages of stable EGFP-positive cells as before. To avoid the generation of infectious vector virus particles, which would lead to EGFP-positive cells by the classical extracellular retroviral transduction, we used the Env protein cleavage site mutant EM20 (Figure 8A). In EM20, the gp130 Env precursor is not processsed into the gp80 surface (SU) and the gp48 transmembrane (TM) 3440
Fig. 5. Analysis of retrotransposition by different foamy virus vectors. HeLa and 293 cells were transiently transfected with the vectors shown in Figure 4 and scored for EGFP expression over time as described in the legend to Figure 3. Cells were also transfected with pMH92 and maintained in the presence of 5 mM RT inhibitor AZT. (A) Percentages 6 SE of marker gene-expressing cells. Three milliliters of cell-free supernatant (SN) from transiently transfected cells did not transduce 104 recipient HeLa or 293 cells. n.d., not done. (B) Relative percentages of marker gene-expressing cells over time of the transiently transfected cells. d.p.t., days post-transfection.
subunits (Pietschmann et al., 2000). Although Env incorporation and viral particle release are only slightly impaired (Pietschmann et al., 2000), EM20 virions are non-infectious, since the fusion peptide cannot be exposed (Figure 8B). Interestingly, we observed a strong positive correlation between increasing amounts of EM20 cotransfected with the pMH92 vector and the percentage of stable EGFP-positive cells that were rescued over time (Figure 8C). Even after 70 days in culture HeLa cells
Foamy virus intracellular retrotransposition
Single cell clones were established from an aliquot of the HeLa cells co-transfected with pMH92 and 10 mg of EM20 by limiting dilution and G418 selection starting 2 weeks after transient transfection. Following cellular DNA extraction and EcoRI digestion, the proviral copy number was determined by Southern blot hybridization. As shown in Figure 9, the HeLa/MH92/EM20 cell clones harbored between 1 and ~10 independent MH92 proviruses. This result does not deviate from that found for cells transfected with pMH92 in the absence of the env expression plasmid (Figure 6). We conclude from these experiments that primate FV genomes retrotranspose intracellularly with high ef®ciency in the presence of envelope protein.
Discussion
Fig. 6. Southern blot of individual cell clones with MH92 integrants. Individual cell clones were established by limiting dilution of HeLa and 293 cells transfected with pMH92 and selection in G418 starting 2 or 3 weeks following transfection, respectively. Total DNA was extracted and 10 mg were digested with EcoRI. EcoRI linearizes pMH92, yielding a band of 13.9 kb. In the lanes denoted DNA + pMH92, 17 pg of plasmid were added to 10 mg of cellular 293 or HeLa DNA. This correponds to approximately one plasmid copy per cell genome. The blots were hybridized to a gag/5¢pol gene probe to detect individual proviruses. Between 1 (293/MH92#6 and HeLa/MH92#9) and 8±10 (293/MH92#5 and HeLa/MH92#8) integrants were detected in the cell clones. The marker size is given in kilobases.
primarily co-transfected with pMH92 and 10 mg of EM20 plasmid were up to ~90% marker gene-positive without applying any selection pressure (Figure 8C and E). Similar results were obtained with 293 cells, which, however, were only up to 16% stably transduced (Figure 8D and E). Control experiments of EM20 with pMH97 did not result in an increased rescue of marker gene-expressing cells, providing evidence that an active integrase was essential for obtaining such cells upon co-transfection of pMH92 and EM20 (Figure 8D).
The experiments presented in this study show that a primate FV has the ability to follow two different ways of replication: it either leaves the cell to become an infectious exogenous retrovirus, or it behaves as a retrotransposon and integrates into the host cell genome. To our knowledge, the FV vector pMH92 described here is one of the most ef®cient examples of movable genetic elements in mammalian cells. Being able to perform both an extra- and intracellular replication pathway, primate FVs functionally bridge the world of retrotransposons and exogenous retroviruses. The intracellular pathway can be regarded as a replicative shortcut, which is also believed to be part of the hepadnavirus replication strategy (Nassal and Schaller, 1993; Ganem, 1996). A similarity between FVs and hepadnaviruses with respect to intracellular replication has been suggested previously, when the nuclear translocation of foamy viral capsid proteins in infected cells was disclosed (Schliephake and Rethwilm, 1994). The recent description of multiple proviruses in persistently FVinfected cells supported this hypothesis, although an extracellular way to acquire multiple integrants could not be excluded (Meiering et al., 2000). Following reverse transcription in hepatitis B virus-infected cells, DNA genomes may be reimported into the nucleus to serve as additional templates for further rounds of transcription (Nassal and Schaller, 1993; Ganem, 1996). Since hepadnaviruses do not integrate into the host cell genome, the consistent internal replication cycle is required to maintain a persistent infection of this episomal virus (Nassal and Schaller, 1993; Ganem, 1996). FVs are integrationcompetent as well as integration-dependent (Neves et al., 1998; Enssle et al., 1999). The evolutionary advantage of generating more than one provirus per infected cell is not obvious. However, we can think of two explanations that are not mutually exclusive. First, FVs may gain an advantage from multiple proviruses similar to that which hepadnaviruses get from multiple nuclear templates. The ampli®cation of foamy proviruses may result in more transcripts and, thus, in superior viral replication. As shown above (Figures 6 and 9), the host cell can easily contain up to 10 proviruses and may even tolerate more, e.g. ~20 proviruses have recently been described in erythroleukemia cells persistently infected with FV (Meiering et al., 2000). This hypothesis is most attractive if FV integration has a speci®city for 3441
M.Heinkelein et al.
Fig. 7. Analysis of the proviral 5¢ ends of MH92 integrants in 293 cell clones. (A) Schematic view of the hybridization pattern of the primers #534 (forward in U3) and #327 (reverse in gag) in the parental plasmid and FV integrants. While the primers ¯ank a 1.5 kb fragment in the undeleted HSRV2 LTR, the indicative PCR fragment of 1.0 kb is formed following transfer of the deleted U3 region of pMH92 to the 5¢ LTR during reverse transcription. (B) PCR result of 293 single cell clones indicating transfer of the U3 region to the 5¢ LTR. pcHSRV2 is a pHSRV2-derived plasmid containing a hybrid 5¢ CMV-LTR promoter (Figure 4). The marker (M) is l-DNA cleaved with PstI. (C) Sequence analysis of the cellular DNA± provirus junctions of some MH92 integrants. All proviruses start with the conserved U3 sequence and had integrated at different genomic loci.
functionally irrelevant host cell genomic regions. If this is not the case, the generation of multiple proviruses before progeny virus has accumulated must be considered a very risky endeavor for viral spread. Additional proviruses may hit and destroy genes essential for cellular survival and, thus, abort the viral replication cycle before completion. The better survival of cells in which the pMH92 vector retrotransposed in the presence of Env compared with cells transfected solely with pMH92 suggests that the cytoplasmic accumulation of FV capsids and/or continued retrotransposition is highly toxic for the cells. In support of this view, any attempt to rescue cells either transfected or transduced with a vector that was designed to continuously retrotranspose intracellularly was unsuccessful (data not shown). Furthermore, even without integration the accumulation of episomal retroviral DNA forms can be toxic to cells (Temin, 1988; Moebes et al., 1997). An alternative explanation for a potential advantage of intracellular retrotransposition for FV replication would suggest that FV spread is enhanced by destroying the host cell via proviral bombardement of the cellular genome. In cell culture, primate FVs bud from the plasma membrane as well as into intracellular vesicles, which may be derived from the endoplasmic reticulum (ER) or post-ER compartments (Hooks and Gibbs, 1975; Hooks and Detrick3442
Hooks, 1981; Gelderblom and Frank, 1987; Fischer et al., 1998; Pietschmann et al., 1999; Wang et al., 1999). While cytoplasmic vesicles stuffed with virions can be easily viewed in electron micrographs of primate FV-infected cultures (Hooks and Gibbs, 1975; Hooks and DetrickHooks, 1981; Gelderblom and Frank, 1987), we never observed virus-loaded vesicles that fused with the plasma membrane to deliver virus. This raises the question of how and when the intracellularly trapped, fully infectious virions are released. One possibility is that infected cells disintegrate due to proviral overload late in the viral replication cycle. In our transient assay, intracellular retrotransposition occurred in the presence of env expression (Figure 9). This suggests that intracellular retrotransposition is an integral part of the wild-type FV replication strategy and probably occurs as an ongoing process that, over time, may accumulate high quantities of proviral integrants and, thus, contribute to cellular death. Following this hypothesis, the interaction of Env protein with capsids may prime them in an as yet unkown manner to deliver cDNA into the nucleus. As long as the cell remains intact, further proviruses can obviously serve as templates for additional viral transcripts. Host cell death induced by accumulating foamy proviruses would be remarkably similar to the bacteriophage Mu replication
Foamy virus intracellular retrotransposition
Fig. 9. Southern blot of individual HeLa cell clones with MH92 integrants following co-transfection of cells with pMH92 and EM20. Single cell clones were established by limiting dilution of HeLa cells transfected with pMH92 and EM20 (10 mg) and selection in G418 starting 2 weeks following transfection. Southern blot hybridization with a gag/5¢pol gene-derived probe was performed on 10 mg of EcoRI-digested DNA as described in Figure 6. The control lane (DNA + pMH92) contains 10 mg of HeLa cell DNA and 17 pg of plasmid. The cell clones harbored between 1 (clones #9 and #16) and ~10 (clones #4 and #7) integrants. The marker is in kilobases.
Fig. 8. FV env gene mutants, vector transduction rates on recipient cells and in¯uence of Env expression on pMH92 retrotransposition. (A) The env EM20 mutation abrogates the SU/TM cleavage from the Env precursor, rendering the respective virus non-infectious. However, it still allows export of FV capsids (Pietschmann et al., 2000). EM02 expresses the wild-type FV Env protein. In this plasmid, splice donor and acceptor sites (SD/SA) present in the genomic sequence were inactivated by silent mutagenesis (Lindemann and Rethwilm, 1998). (B) Rates of marker gene-positive recipient cells in cell-free vector transfer experiments. The vectors were produced by transient cotransfection of 293T cells with plasmid DNAs (10 mg each) and the transduction of 104 recipient cells with cell-free virus (1 ml) was carried out and determined as described previously (Heinkelein et al., 1998, 2000). (C) HeLa cells were co-transfected with pMH92 (10 mg) and increasing amounts of EM20. The total of transfected DNA was adjusted to 20 mg with pcDNA. The percentages of marker geneexpressing cells as determined by ¯ow cytometry were measured at the indicated number of days post-transfection (d.p.t.). Two days following transfection, transfer of 3 ml of cell-free supernatant (SN) did not yield EGFP-positive recipient cells. n.d., not done. (D) Results from a cotransfection experiment of 293 cells with pMH92 or integrase mutant pMH97 vector together with the env mutant EM20. (E) With increasing amounts of EM20 a higher percentage of the pMH92-transfected cells expressed the marker gene.
strategy (Mizuuchi and Craigie, 1986). However, phage Mu does not replicate through an RNA intermediate, and besides destruction of the host cell genome, the generation of multiple prophages serves to accumulate what later become progeny virion genomes (Mizuuchi and Craigie, 1986). FVs cause persistent infections in vivo, obviously in the absence of any pathology (Hooks and Gibbs, 1975; Neumann-Haefelin et al., 1993; Aguzzi et al., 1996). Retrovirus infections often cause tumors due to insertional mutagenesis (Rosenberg and Joliceur, 1997). There are various examples of pathological conditions or other biological effects induced by the movement of the rather stationary endogenous retroviruses (Wilkinson et al., 1994; LoÈwer, 1999). Provided multiple FV integrations occur in vivo, virus-induced cell death resulting from an internal replication cycle may be one explanation for why FV infections are not associated with malignancies. However, speci®c FV integration into functionally inert sites of the host cell genome could equally well explain a lack of tumorigenicity. Exogenous retroviruses are believed to have evolved from cellular retrotransposons that acquired an envelope gene (Temin, 1992; Wilkinson et al., 1994; Boeke and Stoye, 1997). Endogenous retroviruses later resulted from infection of the germ line by exogenous retroviruses (Temin, 1992; Wilkinson et al., 1994; Boeke and Stoye, 1997). Although FVs can behave as retrotransposons or as exogenous retroviruses, we do not regard them as a true evolutionary link. FV intracellular retrotransposition appears to be too ef®cient and probably too sophisticated to stand at the evolutionary edge between retrotransposons and exogenous retroviruses. However, FVs may have preserved and perfected features of an ancient retro3443
M.Heinkelein et al.
transposon and may represent the decendant of a genetic element that previously linked retrotransposons to exogenous retroviruses. FVs use a mixed replication strategy that is made up of mechanisms found in all three (retrotransposons, retroviruses and hepadnaviruses). On the basis of these considerations, we suggest that FVs represent a separate viral lineage.
Materials and methods Cells Human cervix epithelial carcinoma (HeLa), kidney [293 and 293T (DuBridge et al., 1987)] and ®brosarcoma (HT1080) cells were cultivated in Eagle's minimal essential medium (MEM) or Dulbecco's modi®ed MEM (DMEM) supplemented with 5 or 10% fetal calf serum and antibiotics. Transfections by calcium phosphate co-precipitation were carried out with 0.8 3 106 HeLa or 1.6 3 106 293 or 293T cells seeded the day before in 6 cm dishes as described previously (Lindemann et al., 1997a; Heinkelein et al., 1998). Plasmids Established molecular cloning techniques were used to generate recombinant plasmids (Ausubel et al., 1987). To produce the egfp±neo fusion gene, both the egfp marker and the neo resistance genes were separately ampli®ed by PCR from pEGFP (Clontech) and pcDNA3 (Invitrogen), respectively. The amplicons were ligated and subcloned. All FV constructs were derivatives of the infectious molecular clone pcHSRV2, which is in a pcDNA backbone (Moebes et al., 1997; Lindemann and Rethwilm, 1998). Plasmids pMH92, pMH96, pMH97 and pMH98 harbor a 2.0 kb cassette where the SFFV U3 promoter (from ±383 to +36 relative to the start of transcription) directs the expression of EGFP±Neo. The 0.42 kb NheI±KpnI SFFV U3 region promoter fragment originated from pSSS-cat (Baum et al., 1997). The U3 region of the FV 3¢ LTR in these plasmids was internally deleted of a 0.56 kb BstEII±XbaI fragment. While the gag±pol genes in pMH92 are of wild-type sequence, the pol ATG down-mutant pMH96 was constructed by exchange of a 1.87 kb SwaI±PacI fragment between pHSRV2/M54 (Enssle et al., 1996) and pMH92. The IN catalytic center (DD35E to DA35E) mutant pMH97 was made by replacing a the 3.17 kb pMH92 SwaI±EcoRI fragment with the corresponding sequence of pHSRV2/M73 (Enssle et al., 1999). Likewise, the exchange of a 5.82 kb MluI±EcoRI fragment derived from pcHSRV2/M69 (Moebes et al., 1997) for the respective fragment of pMH92 generated the RT active center mutant (YVDD to GAAA) pMH98. The gag mutant pMH117 with an altered initiation sequence (ATG/GCT/TCA to TTG/GCT/TAA) was generated by recombinant PCR (Higuchi, 1990) on a 1.05 kb NdeI±EcoRV fragment, which was then substituted for the pMH92 wild-type sequence. All point mutations were veri®ed by DNA sequence analysis. The MLV vector pcAMS/U3EN has a pBR322 backbone and was derived from pAMS (Miller and Buttimore, 1986). The CMV immediate early gene enhancer/promoter was analogously substituted for the 5¢ LTR U3 region as described for other MLV vectors (Soneoka et al., 1995). The SFFV U3/egfp±neo expression cassette was inserted instead of the env gene, which was deleted by recombinant PCR (Higuchi, 1990). This vector was made self-inactivating by incorporating the 3¢ LTR of pSFGlucECT2, which bears deletions in the U3 region from ±412 to ±63 and from ±30 to ±24 relative to the start of transcription (Lindemann et al., 1997b), for the wild-type pAMS 3¢ LTR. The HIV-1 vector pGJ2/U3E within pcDNA (Invitrogen) was derived from the pHXB-2D provirus (Fischer et al., 1985). The CMV immediate early gene enhancer/promoter was substituted for the U3 region of the 5¢ LTR as described for other HIV-1 vectors (Kim et al., 1998). Further deletions were a 1.44 kb A¯III env, a 0.15 kb Bpu1102I±Asp718 nef and a 0.4 kb EcoRV±PvuII fragment from within the the U3 region of the 3¢ LTR. Cellular sequences up to the XbaI site and adjacent to the 3¢ LTR in the HXB-2D provirus were eliminated by recombinant PCR (Higuchi, 1990). A 1.18 kb Esp3I±NotI fragment containing SFFV U3/egfp was inserted into the unique Esp3I site located upstream of the deleted nef reading frame. The CMV enhancer/promoter directed FV env and VSV-G expression constructs pcHFVenv-EM02 and pcVG-wt, respectively, and the FV env expression construct with a mutated surface-transmembrane protein cleavage site (pcHFVenv-EM20) have already been described (Lindemann and Rethwilm, 1998; Pietschmann et al., 1999, 2000).
3444
Vector transduction analysis following transient transfections All individual vector transduction and retrotransposition experiments were carried out 3±12 times. For vector transfer experiments, 293T cells were transfected with a total of 20 mg of plasmid DNA (10 mg of vector and 10 mg of env expression plasmid). Two days after transfection, the supernatants were passed through 0.45 mm ®lters (Schleicher & SchuÈll) and 1 ml was applied to 104 recipient HT1080, HeLa or 293 cells. Forty-eight hours following transduction, EGFPpositive cells were identi®ed and quantitated by ¯uorescence-activated cell sorting (FACS) on a FACScan using the LysisII and CellQuest Software package (Becton Dickinson), as described previously (Heinkelein et al., 1998, 2000). To analyze the intracellular retrotransposition of the vectors, 293 and HeLa cells were transfected with 15 mg of plasmid DNA unless stated otherwise in the Figure legends. Transfection ef®ciencies were determined by FACS analysis 2 days following transfection. The cells were further cultivated and the numbers of marker gene-expressing cells determined by weekly FACS analysis and expressed as percentages of total cells analyzed (usually 2 3 104). AZT was added to the cells in some experiments at a concentration of 5 mM during transfection and maintained at this concentration throughout the observation period. Analysis of individual cell clones pMH92-transfected EGFP-positive 293 and HeLa cells were selected in 1 mg/ml G418 (Gibco-BRL). Single cell clones were established by limiting dilution. Cellular DNA was extracted from individual clones by lysis in SDS-containing buffer, overnight proteinase K digestion, and isopropanol precipitation using established protocols (Ausubel et al., 1987). One microgram of the DNA was analyzed by PCR using the LTR forward primer #534 (5¢-CAAGGAGGAGAGTATTACAGGGAAGGA3¢) and the gag gene-speci®c reverse primer #327 (5¢-GTTACTGGTCTGTAAGAACTAC-3¢). The PCR conditions were 30 s at 95°C, 60 s at 55°C and 120 s at 72°C for 35 cycles. Southern blot analysis was performed on 10 mg of cellular DNA following EcoRI digestion, electrophoretic separation in 0.8% agarose gels and passive transfer to nylon membranes (Schleicher & SchuÈll) in alkaline buffer. The blots were hybridized under stringent conditions using standard protocols (Ausubel et al., 1987) to a 1.8 kb SwaI±PacI probe from the gag/pol genes, which was randomly labeled with [a-32P]dATP using the prime-it II kit (Stratagene). The ®nal wash was performed in 0.13 SSC, 0.1% SDS at 65°C, and following air drying the blots were exposed to X-ray ®lm for 2±3 days. The junctions between proviruses and cellular DNA were determined by inverse nested PCR with minor modi®cations of the published method (Enssle et al., 1999). Brie¯y, 2.5 mg of total cellular DNA were cleaved overnight with HaeIII. Following heat inactivation and ethanol precipitation, one-quarter of the DNA was self-ligated with phage T4 DNA ligase overnight. The ligated DNA was recut with AvrII to enhance the ampli®cation from linear templates and one-tenth of the material was used for the ®rst round PCR with primers #498 (5¢-AGGTTCTTCACCTCCTTCCCTG-3¢) and #188 (5¢-ACAATGGGTACCTCAGGAAGTAATGTTGAAGA-3¢). A nested PCR was performed with one®fth of the material from the ®rst round PCR using primers #571 (5¢AACTTGATGTTGAAGCTCTGG-3¢) and #572 (5¢-TCCTTCCCTGTAATACTCTCC-3¢). Following gel electrophoretic separation and puri®cation, the amplicons were sequenced in an ABI 310 automated DNA sequencing device (Perkin Elmer) with the same oligonucleotide primers as used for the second round of PCR ampli®cation.
Acknowledgements We thank Myra O.McClure and Dieter Neumann-Haefelin for critical review of the manuscript, Dusty Miller and Christian Jassoy for the gift of reagents, Roswitha LoÈwer for discussion, Volker Bellmann for artwork, and JoÈrg Enssle for his input into the initial phase of these experiments. This work was supported in part by grants from the Deutsche Forschungsgemeinschaft (SFB 479, Re627/6-1 and Li621/2-1), Bundesministerium fuÈr Bildung und Forschung (BMBF) (01KV9817/0), EU (BMH4-CT97-2010), Bayerische Forschungsstiftung (Forgen), SaÈchsisches Staatsministerium fuÈr Umwelt und Landwirtschaft, and by the Medical Faculty `Carl Gustav Carus', TU Dresden. G.J. and D.L. were supported by grants from the Hector Stiftung and the AIDSStipendienprogramm of the BMBF, respectively.
Foamy virus intracellular retrotransposition
References Aguzzi,A., Marino,S., Tschopp,R. and Rethwilm,A. (1996) Regulation of expression and pathogenic potential of human foamy virus in vitro and in transgenic mice. Curr. Top. Microbiol. Immunol., 206, 243± 273. Ausubel,F.M., Brent,R., Kingston,R.E., Moore,D., Seidman,J.G., Smith,J.A. and Struhl,K. (1987) Current Protocols in Molecular Biology. John Wiley, New York, NY. Baum,C., Ito,K., Meyer,J., Laker,C., Ito,Y. and Ostertag,W. (1997) The potent enhancer activity of the polycythemic strain of spleen focus forming virus in hematopoietic cells is governed by a binding site for Sp1 in the upstream control region and by a unique enhancer core motif, creating an exclusive target for PEBP/CBF. J. Virol., 71, 6323± 6331. Boeke,J.D. and Stoye,J.P. (1997) Retrotransposons, endogenous retroviruses and the evolution of retroelements. In Cof®n,J.M., Hughes,S.H. and Varmus,H.E. (eds), Retroviruses. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 343±435. DuBridge,R.B., Tang,P., Hsia,H.C., Leong,P.-M., Miller,J.H. and Calos,M.P. (1987) Analysis of mutation in human cells by using Epstein±Barr virus shuttle system. Mol. Cell. Biol., 7, 379±387. Enssle,J., Jordan,I., Mauer,B. and Rethwilm,A. (1996) Foamy virus reverse transcriptase is expressed independently from the gag protein. Proc. Natl Acad. Sci. USA, 93, 4137±4141. Enssle,J., Moebes,A., Heinkelein,M., Panhuysen,M., Mauer,B., Schweizer,M., Neumann-Haefelin,D. and Rethwilm,A. (1999) An active human foamy virus integrase is required for viral replication. J. Gen. Virol., 80, 1445±1452. Fischer,A.G., Collalti,E., Ratner,L., Gallo,R.C. and Wong-Staal,F. (1985) A molecular clone of HTLV-III with biological activity. Nature, 316, 262±265. Fischer,N., Heinkelein,M., Lindemann,D., Enssle,J., Baum,C., Werder,E., Zentgraf,H., MuÈller,J.G. and Rethwilm,A. (1998) Foamy virus particle formation. J. Virol., 72, 1610±1615. Ganem,D. (1996) Hepadnaviridae and their replication. In Fields,B.N., Knipe,D.M. and Howley,P.M. (eds), Virology. Vol. 2. LippincottRaven, New York, NY, pp. 2703±2737. Gelderblom,H. and Frank,H. (1987) Spumavirinae. In Nermut,M.V. and Steven,A.C. (eds), Animal Virus Structure. Elsevier, Amsterdam, The Netherlands, pp. 305±311. Heidmann,O. and Heidmann,T. (1991) Retrotransposition of a mouse IAP sequence tagged with an indicator gene. Cell, 64, 159±170. Heidmann,T., Heidmann,O. and Nicolas,J.-F. (1988) An indicator gene to demonstrate intracellular transposition of defective retroviruses. Proc. Natl Acad. Sci. USA, 85, 2219±2223. Heinkelein,M., Schmidt,M., Fischer,N., Moebes,A., Lindemann,D., Enssle,J. and Rethwilm,A. (1998) Characterization of a cis-acting sequence in the pol region required to transfer human foamy virus vectors. J. Virol., 72, 6307±6314. Heinkelein,M., Thurow,J., Dressler,M., Imrich,H., NeumannHaefelin,D., McClure,M.O. and Rethwilm,A. (2000) Complex effects of deletions in the 5¢ untranslated region of primate foamy virus on viral gene expression and RNA packaging. J. Virol., 74, 3141±3148. Higuchi,R. (1990) Recombinant PCR. In Innis,M.A., Gelfand,D.H. and White,T.J. (eds), PCR Protocols: A Guide to Methods and Applications. Academic Press, San Diego, CA, pp. 177±183. Hooks,J.J. and Detrick-Hooks,B. (1981) Spumavirinae: foamy virus group infections: comparative aspects and diagnosis. In Kurstak,K. and Kurstak,C. (eds), Comparative Diagnosis of Viral Diseases. Vol. 4. Academic Press, New York, NY, pp. 599±618. Hooks,J.J. and Gibbs,C.J. (1975) The foamy viruses. Bacteriol. Rev., 39, 169±185. Jordan,I., Enssle,J., GuÈttler,E., Mauer,B. and Rethwilm,A. (1996) Expression of human foamy virus reverse transcriptase involves a spliced pol mRNA. Virology, 224, 314±319. Kim,V.N., Mitrophanous,K., Kingsman,S.M. and Kingsman,A.J. (1998) Minimal requirements for a lentivirus vector based on human immunode®ciency virus type 1. J. Virol., 72, 811±816. Lindemann,D. and Rethwilm,A. (1998) Characterization of a human foamy virus 170 kD env±bet fusion protein generated by alternative splicing. J. Virol., 72, 4088±4094. Lindemann,D., Bock,M., Schweizer,M. and Rethwilm,A. (1997a) Ef®cient pseudotyping of murine leukemia virus particles with
chimeric human foamy virus envelope proteins. J. Virol., 71, 4815± 4820. Lindemann,D., Patriquin,E., Feng,S. and Mulligan,R.C. (1997b) Versatile retrovirus vector systems for regulated gene expression in vitro and in vivo. Mol. Med., 3, 466±476. Linial,M.L. (1999) Foamy viruses are unconventional retroviruses. J. Virol., 73, 1747±1755. LoÈwer,R. (1999) The pathogenenic potential of endogenous retroviruses: facts and fantasies. Trends Microbiol., 7, 350±356. Meiering,C.D., Comstock,K.E. and Linial,M.L. (2000) Multiple integrations of human foamy virus in persistently infected human erythroleukemia cells. J. Virol., 74, 1718±1726. Miller,A.D. and Buttimore,C. (1986) Redesign of retrovirus packaging cell lines to avoid recombination leading to helper virus production. Mol. Cell. Biol., 6, 2895±2902. Mizuuchi,K. and Craigie,R. (1986) Mechanism of bacteriophage Mu transposition. Annu. Rev. Genet., 20, 385±429. Moebes,A., Enssle,J., Bieniasz,P.D., Heinkelein,M., Lindemann,D., Bock,M., McClure,M.O. and Rethwilm,A. (1997) Human foamy virus reverse transcription that occurs late in the viral replication cycle. J. Virol., 71, 7305±7311. Nassal,M. and Schaller,H. (1993) Hepatitis B virus replication. Trends Microbiol., 1, 221±228. Neumann-Haefelin,D., Fleps,U., Renne,R. and Schweizer,M. (1993) Foamy viruses. Intervirology, 35, 196±207. Neves,M., PeÂrieÁs,J. and SaõÈb,A. (1998) Study of human foamy virus proviral integration in chronically infected murine cells. Res. Virol., 149, 393±401. Pietschmann,T., Heinkelein,M., Heldmann,M.A., Zentgraf,H., Rethwilm,A. and Lindemann,D. (1999) Foamy virus capsids require the cognate envelope protein for particle export. J. Virol., 73, 2613± 2621. Pietschmann,T., Zentgraf,H., Rethwilm,A. and Lindemann,D. (2000) An evolutionarily conserved positively charged amino acid in the putative membrane-spanning-domain of the foamy virus envelope protein controls fusion activity. J. Virol., 74, 4474±4482. Rethwilm,A. (1996) Unexpected replication pathways of foamy viruses. J. Acquir. Immune De®c. Syndr. Hum. Retrovirol., 13(Suppl. 1), S248± S253. Rosenberg,N. and Joliceur,P. (1997) Retroviral pathogenesis. In Cof®n,J.M., Hughes,S.H. and Varmus,H.E. (eds), Retroviruses. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 475± 585. Schliephake,A.W. and Rethwilm,A. (1994) Nuclear localization of foamy virus gag precursor protein. J. Virol., 68, 4946±4954. Soneoka,Y., Cannon,P.M., Ramsdale,E.E., Grif®th,J.C., Romano,G., Kingsman,S.M. and Kingsman,A.J. (1995) A transient three-plasmid expression system for the production of high titer retroviral vectors. Nucleic Acids Res., 23, 628±633. Swanstrom,R. and Wills,J.W. (1997) Synthesis, assembly and processing of viral proteins. In Cof®n,J.M., Hughes,S.H. and Varmus,H.E. (eds), Retroviruses. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 263±334. Temin,H.M. (1988) Mechanisms of cell killing/cytopathic effects by nonhuman retroviruses. Rev. Infect. Dis., 10, 399±405. Temin,H.M. (1992) Origin and general nature of retroviruses. In Levy,J.A. (ed.), The Retroviridae. Vol. 1. Plenum Press, New York, NY, pp. 1±18. Vogt,P.K. (1997) Historical introduction to the general properties of retroviruses. In Cof®n,J.M., Hughes,S.H. and Varmus,H.E. (eds), Retroviruses. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 1±25. Wang,G., Mulligan,M.J., Baldwin,D.N. and Linial,M.L. (1999) Endogenous virus of BHK-21 cells complicates electron microscopy studies of foamy virus maturation. J. Virol., 73, 8917. Wilkinson,D.A., Mager,D.L. and Leong,J.-A. (1994) Endogenous human retroviruses. In Levy,J.A. (ed.), The Retroviridae. Vol. 3. Plenum Press, New York, NY, pp. 465±535. Yu,S.F., Baldwin,D.N., Gwynn,S.R., Yendapalli,S. and Linial,M.L. (1996) Human foamy virus replication: a pathway distinct from that of retroviruses and hepadnaviruses. Science, 271, 1579±1582. Yu,S.F., Sullivan,M.D. and Linial,M.L. (1999) Evidence that the human foamy virus genome is DNA. J. Virol., 73, 1565±1572. Received February 7, 2000; revised May 2, 2000; accepted May 5, 2000
3445
