REVIEW New Insight on the Role of Extrachromosomal Retroviral DNA a Cara1 and MS Reitz Jr2

Home , Extrachromosomal DNA, Murine leukemia virus, Nef (protein)

REVIEW New insight on the role of extrachromosomal retroviral DNA A Cara1 and MS Reitz Jr2

1Basic Research Laboratory, National Cancer Institute, National Institutes of Health, Bethesda, MD; and 2Institute of Human Virology, University of Maryland Biotechnology Institute and School of Medicine, Baltimore, MD 21201, USA

During infection with different retroviruses, high levels of unin- reverse-transcription process; circular with a single LTR tegrated extrachromosomal DNA accumulate in infected cells. (c1LTR), present in the nucleus of the infected cell and prob- While extrachromosomal linear DNA is the immediate precursor of the integrated provirus, the function, if any, of extra- ably derived from homologous recombination between the chromosomal circular DNA has been unclear. Several groups two viral LTR ends of linear E-DNA; and circular with two have attempted to address the possible function, activity, and tandem LTRs (c2LTR), present in the nucleus of the infected importance of this unintegrated DNA during the life cycle of cell derived either from the head to tail joining of the two viral retroviruses and the course of retroviral-associated diseases. LTR ends of linear E-DNA or from intramolecular autointegrat- This review summarizes recent work in this ﬁeld and tries to ive events. Lately, a number of reports have suggested that the analyze some aspects of extrachromosomal forms of retroviral DNA and their possible application as a molecular biological extrachromosomal forms of viral DNA and, more speciﬁcally, tool. circular E-DNA may have a role in the replication cycle of Keywords: HIV-1; extrachromosomal DNA; integrase; transcrip- retroviruses. tion vaccinia

Extrachromosomal DNA as a marker of viral replication Introduction Reverse transcription takes place in the cytoplasm of the All retroviruses enter target cells via interaction between the infected cell. In quiescent cells, the final product of reverse envelope glycoprotein present on the retroviral envelope and transcription, linear E-DNA is found only in the cytoplasm of the specific receptor(s) exposed on the outermost membrane the cell, which is bound to several proteins of the preinte- of the target cell. This interaction determines the fusion gration complex while it is absent inside the nuclei (Figure between the viral envelope and cellular membranes, thus 1).6–9 After the cell has been activated, the fate of linear inducing the release of the viral core into the cytoplasm of E-DNA differs between lentiretroviruses and oncoretroviruses. the infected cell. After the viral core has been internalized In lentiretroviruses, such as the human immunodeficiency in the target cell, reverse transcriptase mediates cytoplasmic virus type 1 (HIV-1), linear E-DNA is translocated upon cellu- reverse transcription of the viral genomic RNA into double- lar activation, both in dividing and nondividing cells, into the stranded linear complementary DNA by using the free nucleo- nucleus of the host cell. This translocation is an ATP-depen- tides and ions present in the cytoplasm of the cell and the dent active process involving the presence of several kary- tRNA primer bound to the primer binding site of the retroviral ophilic elements present in at least two proteins of the preinte- RNA. Once completed, double stranded linear complemen- gration complex (p17gag and vpr) (Figure 1a).3,10–12 In tary DNA enters into the nucleus as part of a preintegration oncoretroviruses, such as the Moloney murine leukemia virus complex which, at the very least, contains the viral enzyme (MMLV), linear E-DNA remains in the cytoplasm until nuclear integrase (IN). The linear viral DNA is integrated into host division during the M phase of the cell cycle breaks down the cellular DNA after both ends of linear DNA have undergone nuclear envelope and permits nuclear import of the preinte- loss of two bases at each end.1 Such integrated proviral DNA gration complex together with unintegrated linear E-DNA by is then competent for transcription of viral mRNAs that, after a passive mechanism (Figure 1b).9,13 It has been suggested that export into the cytoplasm, are translated into viral proteins to this difference may be attributable to a difference in the size produce new infectious viral particles. of the preintegration complexes of each retrovirus,9 since Typically, when reverse-transcribed linear extrachromo- translocation of preintegration complexes through the nuclear somal viral DNA (linear E-DNA) reaches the nucleus of the pore seems to be size dependent. However, the karyophilic infected cell, it can either integrate in the proviral form or elements encoded by HIV-1, which determine nuclear translo- circularize and remain as extrachromosomal DNA (circular E- cation of the viral preintegration complex, appear to be absent DNA).1–3 While it is known that linear E-DNA is the substrate in the MMLV preintegration complex. Thus, different retro- for integration,4 circular E-DNA is thought to be simply a side viruses may have developed different capabilities to infect tar- product of the reverse-transcription process.1,5 get cells in different phases of the cell cycle. Therefore, retro- E-DNA is found in infected cells in at least three different viral replication will either require cellular activation in the forms: linear, present in the cytoplasm and in the nucleus of absence of cell replication, as in the case of lentiretroviruses the infected cell and constituting the final product of the (HIV-1),3,8,12 or cell replication, as in the case of oncoretroviruses (MMLV).9,13 In either case, however, circular E-DNA forms are present only in the nucleus of the infected cells. For this reason, detection of circular E-DNA in the nuclei of the Correspondence: A Cara, Mount Sinai Medical Center, Division of Infectious Diseases, One Gustave Levy Place, Box 1090, Atran 6-620, infected cells is considered to be a marker for the nuclear New York, NY 10029, USA translocation process. In addition, circular E-DNA is always Received 4 December 1996; accepted 23 May 1997 present during the acute phase of infection in vitro and in Review A Cara and MS Reitz 1396

Figure 1 Schematic representation of the early events of replication in (a) lentiretroviruses and (b) oncoretroviruses.

vivo, and, for this reason, it is also considered a marker of an additional step to the simple amplification of DNA. Ampli- active viral replication.14–16 fication of circular E-DNA could easily be substituted for intra- Several studies of E-DNA have been performed in patients cellular RNA analysis as a measure of viral replication, thus infected with HIV-117–19 or human T cell leukemia virus type saving time and eliminating reverse transcription of RNA into I20,21 as well as in animals infected with simian immunodefi- cDNA. For these reasons, measurement of the level of circular ciency virus (SIV),22 feline leukemia virus23, MMLV,24 equine E-DNA during the course of infection, since it is a marker of infectious anemia virus,25 spleen necrosis virus,26 avian viral replication, appears to be a simple tool to follow the leukosis virus,27 avian reticuloendotheliosis virus,28 and bov- course of retroviral infections in vivo. However, due to the ine leukemia virus.29 In all cases, recovery of E-DNA was con- lower levels of circular E-DNA, with respect to integrated current with viral replication. Moreover, a more active viral DNA and viral RNA during the course of the infection and replication corresponded with an increased amount of E-DNA the consequent difficulties in detecting such smaller amounts and particularly circular E-DNA, suggesting a connection of viral DNA, we would argue against such an approach. between the stage of any associated diseases and the amount of E-DNA. With regard to HIV-1 infection, levels of extrachromosomal Extrachromosomal DNA is transcriptionally active viral DNA have been correlated with HIV-1 expression and CD4+ cell decline in vivo.18,19 Moreover, a decrease in HIV-1 During the past few years, several reports have suggested that unintegrated DNA recovery has been associated with antire- circular E-DNA is not simply a dead end product of the troviral therapy.30–32 It has been suggested that the level of reverse transcription process but rather participates actively in viral RNA is a better marker of HIV-1 replication and stage of the course of infection. Indeed, circular E-DNA has been the disease than is the decrease in CD4+ cell counts or the shown to be transcriptionally active, although at low levels level of total viral DNA.33 However, although amplification and only for a short period of time.34–40 In itself, this is not of extracellular RNA is necessary to quantitate the levels of surprising since both linear and circular E-DNA possess the serum-associated virus, testing for viral replication by analyz- promoter and termination signals essential for synthesis of ing the levels of intracellular RNA using RT-PCR introduces viral mRNA and consequently viral proteins. The fact that the Review A Cara and MS Reitz 1397 transcriptional potential of E-DNA appears to be quite low in the IN protein. With respect to the first possibility, it is very suggests that the structural conformation of this DNA does not unlikely that IN protein binds circular E-DNA. Indeed, while make it an efficient template for viral expression. several proteins of the preintegration complex are bound to Indeed, several lines of evidence suggest that the transcrip- linear E-DNA, only histones seem to bind circular E-DNA.41,42 tional activity of unintegrated circular c1LTR and c2LTR E- Possibly, when linear E-DNA circularizes to form either c1LTR DNA and linear E-DNA is much lower than that of its inte- or c2LTR E-DNA, the proteins of the preintegration complex grated counterpart.40 First, transcriptional activity of E-DNA are displaced from the LTR termini, and the circular forms are appears to be impeded by several proteins of the preinte- not recognized as suitable substrates for integration. Another gration complex which possibly displace or block the attach- explanation for the different transcriptional activity reported ment of transcriptional factors in the LTR.41,42 In addition, the for E-DNA using IN-defective mutants is that different short half-life of c1LTR and c2LTR, between 1 and 2 days as mutations introduced into the IN protein simply do not have has been determined in vivo,18 does not permit them to be the same effect on the amount of E-DNA produced. Since stable templates for viral mRNA transcription. Another reason recovery of RT activity after transient transfection of some IN- for the low contribution of E-DNA to the viral RNA pool is defective molecular clones of HIV-1 is lower than that of the likely due to phenomena of promoter interference or wild-type counterpart,34,36,39,43,44,48 it is likely that these occlusion because of close proximity of the two promoters in mutants produce a lower amount of E-DNA, thus reducing the the case of the c2LTR E-DNA, or the double use of the single amount of E-DNA template for mRNA transcription. LTR in c1LTR E-DNA to both start and terminate transcription. We have reported that in cell culture using the HIV-1 system, c1LTR is more active than c2LTR, while the activity of linear IN-defective viruses as a vaccine E-DNA is very low.40 It is not clear why linear E-DNA possesses a lower transcriptional activity than the circular forms. Since E-DNA is transcriptionally active (although at low In addition to being a more efficient template, integrated levels), at least with HIV-1 and possibly SIV as well, it may DNA possesses a much longer half-life, thus allowing it to be possible to exploit IN-defective viruses as vaccines. The produce the majority of viral mRNA necessary to produce expression of viral proteins for a short period of time and at viral proteins and, consequently, the majority of infectious low levels could still stimulate an immune response against viral particles. A lower amount of circular E-DNA is usually viral proteins. Viral proteins translated from mRNAs derived present both in vitro and in vivo at sites of viral replication18 from E-DNA would be processed by similar pathways as those further ensuring that the major producer of viral mRNA, and transcribed from integrated templates in natural infections and hence protein, is the integrated form of the retrovirus. How- would thus be in their most native form. Additionally, intra- ever, it is possible that E-DNA represents a way for the virus cellular production should ensure stimulation of a CTL to replicate, albeit inefficiently, under conditions which do response. The absence of IN would help to ensure safety not permit integration. because of the transient nature of E-DNA. The use of animal IN mutant or defective viruses can be used as tools to charac- models would of course be necessary to test this hypothesis. terize the transcriptional activity of circular E-DNA.34–40 IN- defective viruses are unable to integrate reverse-transcribed circular DNA, producing only linear E-DNA in the cytoplasm Conclusions and circular E-DNA in the nuclei of the infected cells. Although this system is the closest to the situation in vivo,a While the production of E-DNA during the life cycle of retro- great variability of transcriptional activity for circular E-DNA viruses is a well-established fact, their biological significance has been found when IN-defective viruses are used to infect is still a matter of controversy. Although it is possible that E- ␤ HeLa CD4 cells harboring a Tat-inducible construct (LTR- - DNA is only an accidental side product of the reverse tran- gal), peripheral blood lymphocytes or macrophages, or when scription process, its activity and potential uses seem signifi- used in transient transfection assays. This may depend upon cant. To investigate this possibility further, it will be important the particular mutation introduced, in that some mutations to produce new in vitro and in vivo models to test further may adversely affect gag-pol precursor polyprotein pro- possible functions of E-DNA and to see if the findings reported cessing, which may indirectly account for the low transcrip- for HIV-1 are applicable to other retroviruses. tional activity of E-DNA of IN mutants. In this regard, we and others have found that after transfection of molecular clones of HIV-1 containing a defective or mutated IN gene, the mor- Acknowledgements phology of the viral particle was aberrant and the release of viral proteins into supernatants of the transfected cells were We thank Kelli Carrington for editorial assistance. decreased relative to the wild-type. This indicates that the ability to sustain multiple rounds of infection is not only dependent on the transcriptional ability of unintegrated DNA References but also on the correct processing of viral proteins in the 34,36,39,43–48 absence of a functional IN. 1 Varmus H, Swanstrom R. Replication of retroviruses. In: Weiss R, The fact that different levels of transcriptional activity are Brown N, Teich PO, Varmus H, Coffin J (eds). RNA Tumor Viruses. associated with E-DNA of IN defective or mutated viruses sug- Cold Spring Harbor Laboratory: Cold Spring Harbor, NY, 1984, gests either that (1) IN protein remains bound to E-DNA and pp 75–134. mutations affect this binding to different degrees which in turn 2 Bowerman B, Varmus HE, Bishop JM. Correct integration of retro- results in different levels of transcriptional activity; or (2) IN viral DNA in vitro. Cell 1987; 49: 347–357. 3 Bukrinsky MI, Sharova N, Dempsey MP, Stanwick TL, Bukrinskaya protein possesses a yet to be discovered activity which either AG, Haggerty S, Stevenson M. Active nuclear import of human directly or indirectly modifies the transcriptional activity of immunodeficiency virus type 1 preintegration complexes. Proc unintegrated viral DNA depending upon the mutation present Natl Acad Sci USA 1992; 89: 6580–6584. Review A Cara and MS Reitz 1398 4 Lobel LI, Murphy JE, Goff SP. The palindromic LTR–LTR junction novel single stranded form in paralyzed mice. J Virol 1995; 69: of Moloney murine leukemia virus is not an efficient substrate for 348–356. proviral integration. J Virol 1989; 63: 2629–2637. 25 Rice N, Lequarre A, Casey J, Lahn S, Stephens R, Edwards J. Viral 5 Brown PO, Bowerman B, Varmus HE, Bishop JM. Retroviral inte- DNA in horses infected with equine infectious anemia virus. J gration: structure of the initial covalent product and its precursor, Virol 1989; 63: 5194–5200. and a role for the IN protein. Proc Natl Acad Sci USA 1989; 86: 26 Keshet E, Temin HM. Cell killing by spleen necrosis virus is corre- 2525–2529. lated with a transient accumulation of spleen necrosis virus DNA. 6 Polacino PS, Liang HA, Firpo EJ, Clark EA. T cell activation influ- J Virol 1979; 31: 376–388. ences initial DNA synthesis of simian immunodeficiency virus in 27 Robinson HL, Miles BD. Avian leukosis virus-induced osteo- resting T lymphocytes from macaques. J Virol 1993; 67: 7008– porosis is associated with the persistent synthesis of viral DNA. 7016. Virology 1985; 141: 130–143. 7 Zack JA, Arrigo SJ, Weitsman SR, Go AS, Haislip A, Chen IS. HIV- 28 Temin HM, Keshet E, Weller SK. Correlation of transient accumu- 1 entry into quiescent primary lymphocytes: molecular analysis lation of linear unintegrated viral DNA and transient cell killing reveals a labile, latent viral structure. Cell 1990; 61: 213–222. by avian leukosis and reticuloendotheliosis viruses. Cold Spring 8 Stevenson M, Stanwick TL, Dempsey MP, Lamonica CA. HIV-1 Harbor Symp Quant Biol 1979; 44: 773–778. replication is controlled at the level of T cell activation and pro- 29 Reyes RA, Cockerell GL. Unintegrated bovine leukemia virus viral integration. EMBO J 1990; 9: 1551–1560. DNA: association with viral expression and disease. J Virol 1996; 9 Roe T, Reynolds TC, Yu G, Brown PO. Integration of murine leu- 70: 4961–4965. kemia virus depends on mitosis. EMBO J 1993; 12: 2099–2108. 30 Bush CE, Donovan RM, Smereck SM, Strang D, Markowitz N, 10 Heinzinger NK, Bukrinsky MI, Haggerty SA, Ragland AM, Kawal- Saravolatz LD. Quantitation of unintegrated HIV-1 DNA in ramani V, Lee M-A, Gendelman HE, Ratner L, Stevenson M, Emer- asymptomatic patients in the presence of absence of antiretroviral man M. The vpr protein of human immunodeficiency virus type therapy. AIDS Res Hum Retrovir 1993; 9: 183–187. 1 influences nuclear localization of viral nucleic acids in nondi- 31 Donovan RM, Bush CE, Smereck SM, Baxa DM, Markowitz N, viding host cells. Proc Natl Acad Sci USA 1994; 91: 7311–7315. Saravolatz LD. Rapid decrease in unintegrated human immuno- 11 von Schwedler U, Kornbluth RS, Trono D. The nuclear localiz- deficiency virus DNA after the initiation of nucleoside therapy. J ation signal of the matrix protein of human immunodeficiency Infect Dis 1994; 170: 202–205. virus type 1 allows the establishment of infection in macrophages 32 Donovan RM, Bush CE, Smereck SM, Moore E, Cohen F, Saravol- and quiescent T lymphocytes. Proc Natl Acad Sci USA 1994; 91: atz LD. Antiretroviral therapy is associated with a decrease in 6992–6996. unintegrated HIV-1 DNA in pediatric patients. J Acq Imm Def Syn 12 Polacino PS, Liang HA, Clark EA. Formation of simian immuno- 1994; 7: 1237–1241. deficiency virus long terminal repeat circles in resting T cells 33 Haynes BF, Pantaleo G, Fauci AS. Toward an understanding of requires both T cell receptor and IL-2-dependent activation. J Exp the correlates of protective immunity to HIV-1 infection. Science Med 1995; 182: 617–621. 1996; 271: 324–328. 13 Lewis PF, Emerman M. Passage through mitosis is required for 34 Stevenson M, Haggerty S, Lamonica CA, Meyer CM, Welch S-K, oncoretroviruses but not for the human immunodeficiency virus. Wasiak AJ. Integration is not necessary for expression of human J Virol 1994; 68: 510–516. immunodeficiency virus type 1 protein products. J Virol 1990; 64: 14 Harris JD, Blum H, Scott J, Traynor B, Ventura P, Haase A. Slow 2421–2425. virus VISNA: reproduction in vitro of virus from extrachromosomal 35 Vogel M, Cichutek K, Norley S, Kurth R. Self-limiting infection of DNA. Proc Natl Acad Sci USA 1984; 81: 7212–7215. int/nef-double mutants of simian immunodeficiency virus. 15 Pauza CD, Galindo JE, Richman DD. Reinfection results in Virology 1993; 93: 115–123. accumulation of unintegrated viral DNA in cytopathic and persist- 36 Engelman A, Englund G, Orenstein JM, Martin MA, Craige R. Mul- ent human immunodeficiency virus type 1 infection of CM cells. tiple effects of mutations in human immunodeficiency virus type J Exp Med 1990; 172: 1035–1042. 1 integrase on viral replication. J Virol 1995; 69: 2729–2736. 16 Robinson HL, Zinkus DM. Accumulation of human immunodefi- 37 Wiskerchen M, Muesing M. Human immunodeficiency virus type ciency virus type 1 DNA in T cells: result of multiple infection 1: effects of mutations on viral ability to integrate, direct viral gene events. J Virol 1990; 64: 4836–4841. expression from unintegrated viral DNA templates and sustain 17 Pang S, Koyanagy Y, Miles S, Wiley C, Vinters HV, Chen I. High viral propagation in primary cells. J Virol 1995; 69: 376–386. levels of unintegrated HIV-1 in brain tissue of AIDS dementia 38 Cara A, Guarnaccia F, Reitz MS Jr, Gallo RC, Lori F. Self limiting, patients. Nature 1990; 343: 85–89. cell type-dependent replication of an integrase defective human 18 Pauza CD, Trivedi P, McKechnie TS, Richman DD, Graziano FM. immunodeficiency virus type 1 in primary macrophages but not 2-LTR circular viral DNA as a marker for human immunodefi- lymphocytes. Virology 1995; 208: 242–248. ciency virus type 1 infection in vivo. Virology 1994; 205: 470– 39 Ansari-Lari MA, Donehower L, Gibbs RA. Analysis of human 478. immunodeficiency virus type 1 integrase mutants. Virology 1995; 19 Jurriaans S, de Ronde A, Dekker J, Cornelissen M, Goudsmith J. 211: 332–335. Increased number of single-LTR HIV-1 DNA junctions correlates 40 Cara A, Cereseto A, Lori F, Reitz MS Jr. HIV-1 protein expression with HIV-1 antigen expression and CD4+ cell decline in vivo. J from synthetic circles of DNA mimicking the extrachromosomal Med Virol 1995; 45: 91–98. forms of viral DNA. J Biol Chem 1996; 271: 5393–5397. 20 Hiramatsu K, Yonemoto R, Yoshikura H. Epigenetic control and 41 Karageorgos L, Li P, Burrell C. Characterization of HIV replication reintegration of extrachromosomal proviral DNA in HL-60 cells complexes early after cell-to-cell infection. AIDS Res Hum chronically infected with human T cell leukemia virus type I. J Retrovir 1993; 9: 817–823. Gen Virol 1988; 69: 651–658. 42 Bukrinsky MI, Sharova N, Stevenson M. Human immunodefi- 21 Kitamura K, Besansky NJ, Rudolph D, Nutman TB, Folks TM, Lal ciency virus type 1 2-LTR circles reside in a nucleoprotein com- RB. Unintegrated two-long terminal repeat virus DNA accumu- plex which is different from the preintegration complex. J Virol lation during chronic HTLV infection. AIDS Res Hum Retrovir 1993; 67: 6863–6865. 1993; 9: 1167–1172. 43 Shin C-G, Taddeo B, Haseltine WA, Farnet CM. Genetic analysis 22 Hirsch VM, Zack PM, Vogel AP, Johnson PR. Simian immuno- of the human immunodeficiency virus type 1 integrase protein. J deficiency virus infection of macaques: end-stage disease is Virol 1994; 68: 1633–1642. characterized by widespread distribution of proviral DNA in 44 Taddeo B, Haseltine WA, Farnet CM. Integrase mutants of human tissues. J Infect Dis 1991; 163: 976–988. immunodeficiency virus type 1 with a specific defect in inte- 23 Mullis JJ, Ches CS, Hoover EA. Disease-specific and tissue-specific gration. J Virol 1994; 68: 8401–8405. production of unintegrated feline leukemia virus variant DNA in 45 Wiskerchen M, Muesing M. Identification and characterization of feline DNA. Nature 1986; 319: 333–336. a temperature sensitive mutant of human immunodeficiency virus 24 Szurek PF, Brooks BR. Development of physical forms of uninte- type 1 by alanine scanning mutagenesis of the integrase gene. J grated retroviral DNA in mouse spinal cord tissue during ts1- Virol 1995; 69: 597–601. induced spongiform encephalomyelopathy: elevated levels of a 46 Englund G, Theodore TS, Freed EO, Engelman A, Martin MA. Inte- Review A Cara and MS Reitz 1399 gration is required for productive infection of monocyte-derived replication of mutations at highly conserved residues. J Virol 1994; macrophages by human immunodeficiency virus type 1. J Virol 68: 4768–4755. 1995; 69: 3216–3219. 48 Cara A, Cereseto A, Fiorentini S, Gusella GL, Reitz MS Jr. Low 47 Cannon PM, Wilson W, Byles E, Kingsman SM, Kingsman AJ. levels of reverse transcriptase associated with one integrase defec- Human immunodeficiency virus type 1 integrase: effect on viral tive mutant of HIV-1. Fund Clin Immunol 1996; 4: 61–66.