Proc. Nati. Acad. Sci. USA Vol. 88, pp. 10342-10346, November 1991 Medical Sciences Construction of intertypic chimeric dengue viruses by substitution of structural protein genes (chimera/viral proteins/neurovirulence) MICHAEL BRAY AND CHING-JUH LAI Molecular Viral Biology Section, Laboratory of Infectious Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD 20892 Communicated by Robert M. Chanock, August 28, 1991 ABSTRACT Dengue virus contains an li-kilobase posi- temperature sensitive, and/or (iii) were adapted to cell cul- tive-strand RNA genome that codes for, in one open reading tures derived from an unnatural host (i.e., host range mu- frame, three structural proteins (capsid, premembrane, and tants) have been selected and evaluated as candidates for envelope), followed by seven nonstructural proteins. The struc- inclusion in a live attenuated virus vaccine (6-8). However, tural protein genes ofa full-length cDNA done of type 4 dengue despite 25 years of such efforts, safe, effective dengue virus were replaced with the corresponding genes of dengue 1 vaccines are still not available for general use. or dengue 2 to create intertypic chimeric cDNA. The RNA There are four serotypes of dengue viruses (types 1-4), transcripts made from these templates were infectious when which are distinguishable by plaque reduction neutralization transfected into permissive cells in culture. Progeny ofchimeric using serotype-specific monoclonal antibodies (mAbs) and cDNA produced apparently authentic dengue 1 or dengue 2 by less specific tests using polyclonal sera (9, 10). The structural proteins, together with dengue 4 nonstructural existence of serotypes was first discovered during early proteins, and as a consequence exhibited type 1 or type 2 studies in human volunteers, which showed that infection serological specificity. Both of the chimeras ultimately grew to with one dengue serotype induced durable homotypic immu- the same titer as their type 1 or type 2 parent, but the type nity, whereas heterotypic immunity lasted only 3-6 months 2/type 4 chimera grew very slowly. This chimera also pro- (2). It has been suggested that an effective dengue vaccine duced small plaques; in contrast, the type 1/type 4 chimera should contain all four serotypes in order to induce broad produced normal size plaques. The type 2/type 4 chimera immunity that would preclude the occurrence of DHF/DSS. retained the mouse neurovirulence ofthe dengue 2 virus, which The complete nucleotide sequences have been determined was the source of its structural protein genes. Each of the mice for dengue virus types 3 and 4 and several strains of type 2 inoculated intracerebrally with the chimera died, but survival virus, including the mouse-neurovirulent New Guinea C time was prolonged. The retardation of replication of the type (NGC) strain; however, only the 5' portion ofthe type 1 virus 2/type 4 chimeric virus suggests that this virus and possibly genome has been sequenced (11-16). The results of these other intertypic dengue virus chimeras with similar properties studies indicate that the four dengue virus serotypes share a should be examined for attenuation in primates and possible common genome organization. The genome of the dengue usefulness in a live dengue virus vaccine for humans. type 4 Caribbean strain 814669 was found to contain 10,646 nucleotides (11, 12). The first 101 nucleotides at the 5' end Dengue is a mosquito-borne viral disease that occurs in and the last 384 at the 3' end are noncoding. The remaining tropical and subtropical regions throughout the world. The sequence codes for a 3386 amino acid polyprotein, which dengue virus subgroup causes more human disease than any includes the three structural capsid (C), other member of the flavivirus family. Dengue is character- proteins-namely, ized by fever, rash, severe headache, and joint pain. Its pre-membrane (pre-M), and envelope (E), at its N terminus- mortality rate is low. However, over the past few decades, a followed by seven nonstructural proteins in the order NS1- more severe form of dengue, characterized by hemorrhage NS2A-NS2B-NS3-NS4A-NS4B-NS5. The polyprotein is and shock (dengue hemorrhagic fever/dengue shock syn- processed by cell signal peptidase(s) and by viral proteases to drome, or DHF/DSS) has been observed with increasing generate 11 or more viral proteins (17-20). frequency in children and young adults. DHF/DSS occurs We recently succeeded in cloning a stable, full-length most often during dengue virus infection of individuals pre- cDNA copy of type 4 dengue virus (designated p2A), which viously infected with another dengue virus serotype, and this can serve as a template for in vitro transcription of infectious has led to the suggestion that immune enhancement of viral RNA (21). Thus, it is now possible to introduce specific replication plays a role in pathogenesis of the more severe mutations into a full-length dengue cDNA and recover viable form of the disease (1). dengue viruses bearing those mutations. Mutants produced in Soon after their isolation in 1944, dengue viruses were this manner should be useful for biologic studies and for passaged repeatedly in mouse brain, resulting in the selection construction of live virus vaccines. We now report the of mouse neurovirulent mutants (2). Interestingly, studies construction of two viable chimeric dengue viruses through performed in volunteers showed that mouse brain-adapted replacement of the structural protein genes of the dengue neurovirulent mutants ofthree strains oftype 1 or type 2 virus type 4 cDNA clone with the corresponding genes oftype 1 or were attenuated, but still immunogenic for humans (2-5). The type 2 dengue virus. These chimeric viruses exhibit the mutants were not developed further as candidate vaccine serologic specificity of type 1 or type 2 but retain the strains because of concern for mouse brain antigens in the nonstructural proteins of type 4. The chimeric type 2/type 4 vaccine preparations. Since that time, virus mutants that (i) exhibited the small plaque size phenotype, and/or (ii) were Abbreviations: DHF/DSS, dengue hemorrhagic fever/dengue shock syndrome; pfu, plaque-forming unit(s); E protein, envelope protein; C protein, capsid protein; pre-M protein, pre-membrane protein; The publication costs of this article were defrayed in part by page charge IFA, immunofluorescence assay; HMAF, hyperimmune mouse as- payment. This article must therefore be hereby marked "advertisement" citic fluid; WP strain, Western Pacific strain; NGC strain, New in accordance with 18 U.S.C. §1734 solely to indicate this fact. Guinea C strain; mAb, monoclonal antibody. 10342 Downloaded by guest on September 30, 2021 Medical Sciences: Bray and Lai Proc. Natl. Acad. Sci. USA 88 (1991) 10343 virus retains the neurovirulence phenotype of dengue 2 NGC I-Xho I fragment was used to replace the corresponding strain, from which its structural genes were derived. fragment of p2A(Xho I) to create p2A(D2 NGC). RNA Transcription, Transfection, and Recovery of Virus. MATERIALS AND METHODS Transfection of cells with full-length RNA transcripts was performed as described (22). Ten days after transfection, Viruses. Dengue type 4 virus strain 814669 was used for cells were trypsinized and transferred to a 6-well plate and to construction offull-length cDNA that served as the source of a chamber slide. Two days later, cells on the slide were tested full-length infectious RNA transcripts (11, 12). A preparation by an immunofluorescence assay (IFA) for evidence of of dengue type 1 virus Western Pacific (WP) strain (D1 WP), dengue viral antigens. If IFA showed that most cells were fetal rhesus lung cell passage level 9, was kindly provided by infected, the cells in the 6-well plate were trypsinized, mixed K. Eckels (Walter Reed Army Institute of Research, Wash- with a 6-fold excess of uninfected cells, and incubated until ington) (21). A mouse brain preparation of mouse neurovir- cytopathic effects (appearance of numerous dead cells in the ulent dengue type 2 virus New Guinea C (NGC) strain (D2 medium) became evident, usually after 6 or 7 days. The NGC), mouse brain passage level 38, was kindly provided by infected cells were then harvested by removing the medium, D. Dubois (Walter Reed Army Institute of Research) (2). scraping the cells, resuspending them in a standard volume of Dengue 1 and dengue 2 viruses were amplified by one passage 50% Eagle's minimal essential medium/50%o serum, and then in C6/36 mosquito cells. Each ofthese three viruses was then freezing. If a small percentage of cells was positive, the cells passaged once in LLC-MK2 simian kidney cells, and the were in and to resulting virus suspension was used to study plaque mor- trypsinized, diluted 1:3 fresh medium, allowed phology and mouse neurovirulence. grow without addition of uninfected cells. The percentage of Cloning Vectors. Plasmid p5'-2, which contains the 5' half infected cells was then estimated on a weekly basis, and cell of the dengue 4 genome cDNA, was modified to facilitate lysates were harvested and titered for virus at intervals. replacement of the three structural protein genes (22). First, Detection of Dengue Virus Antigens. Infected cells were a unique Xho I site was introduced through site-directed analyzed by IFA using serotype-specific mAbs iF1 (dengue mutagenesis at nucleotide 2342 (A -+ G) of the dengue 4 1), 3H5 (dengue 2), 5D4 (dengue 3), and 1H10 (dengue 4); sequence, near the site encoding the C terminus of E protein, nonstructural protein NS1-specific mAb 1G6 (dengue 4), creating p5'-2(Xho I). This nucleotide change did not alter the originally produced by M. K. Gentry and E. Henchal (10); amino acid sequence. This vector was then digested at the and fluorescein-conjugated anti-mouse antiserum (Kierke- unique BstBI site within the dengue sequence and at the gaard & Perry Laboratories, Gaithersburg, MD).