WO2001039802A1

WO2001039802A1 - Chimeric flavivirus vaccines

Info

Publication number: WO2001039802A1
Application number: PCT/US2000/032821
Authority: WO
Inventors: Thomas J. Chambers; Thomas P. Monath; Farshad Guirakhoo
Original assignee: Oravax, Inc.; St. Louis University
Priority date: 1999-12-01
Filing date: 2000-12-01
Publication date: 2001-06-07
Also published as: WO2001039802A9; AU1813901A

Abstract

A chimeric live, infectious, attenuated virus containing a yellow fever virus, in which the nucleotide sequence for a prM-E protein is either deleted, truncated, or mutated, so that functional prM-E protein is not expressed, and integrated into the genome of the yellow fever virus, a nucleotide sequence encoding a prM-E protein of a second, different flavivirus, so that the prM-E protein of the second flavivirus is expressed.

Description

CHTMERTC F AVTVTRUS VACCINES Background of the Invention

This invention relates to infectious, attenuated viruses useful as vaccines against diseases caused by flaviviruses.

Several members of the flavivirus family pose current or potential threats to global public health. For example, Japanese encephalitis is a significant public health problem involving millions of at risk individuals in the Far East. Dengue virus, with an estimated annual incidence of 100 million cases of primary dengue fever and over 450,000 cases of dengue hemorrhagic fever worldwide, has emerged as the single most important arthropod-transmitted human disease. Other flaviviruses continue to cause endemic diseases of variable nature and have the potential to emerge into new areas as a result of changes in climate, vector populations, and environmental disturbances caused by human activity. These flaviviruses include, for example, St. Louis encephalitis virus, which causes sporadic, but serious, acute disease in the midwest, southeast, and western United States; West Nile virus, which causes febrile illness, occasionally complicated by acute encephalitis, and is widely distributed throughout Africa, the Middle East, the former Soviet Union, and parts of Europe; Murray Valley encephalitis virus, which causes endemic nervous system disease in Australia; and Tick-borne encephalitis virus, which is distributed throughout the former Soviet Union and eastern Europe, where its Ixodes tick vector is prevalent and responsible for a serious form of encephalitis in those regions.

Hepatitis C virus (HCV) is another member of the flavivirus family, with a genome organization and replication strategy that are similar, but not identical, to those of the flaviviruses mentioned above. HCV is transmitted mostly by parenteral exposure and congenital infection, is associated with chronic hepatitis that can progress to cirrhosis and hepatocellular carcinoma, and is a leading cause of liver disease requiring orthotopic transplantation in the United States.

The Flaviviridae family is distinct from the alphaviruses (e.g., WEE, VEE, EEE, SFV, etc.) and currently contains three genera, the flaviviruses, the pestiviruses, and the hepatitis C viruses. Fully processed mature virions of flaviviruses contain three structural proteins, envelope (E), capsid (C), and membrane (M), and seven non-structural proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5). Immature flavivirions found in infected cells contain pre-membrane (prM) protein, which is the precursor to the M protein.

After binding of virions to host cell receptors, the E protein undergoes an irreversible conformational change upon exposure to the acidic pH of endosomes, causing fusion between the envelope bilayers of the virions and endocytic vesicles, thus releasing the viral genome into the host cytosol. PrM-containing tick-borne encephalitis (TBE) viruses are fusion-incompetent, indicating that proteolytic processing of prM is necessary for the generation of fusion-competent and fully infectious virions (Guirakhoo et al, J. Gen. Virol. 72(Pt. 2):333-338, 1991). Using ammonium chloride late in the virus replication cycle, prM-containing Murray Valley encephalitis (MVE) viruses were produced and shown to be fusion incompetent. By using sequence-specific peptides and monoclonal antibodies, it was demonstrated that prM interacts with amino acids 200-327 of the E protein. This interaction is necessary to protect the E protein from the irreversible conformational changes caused by maturation in the acidic vesicles of the exocytic pathway (Guirakhoo et al, Virology 191 :921-931, 1992).

The cleavage of prM to M protein occurs shortly before release of virions by a furin-like cellular protease (Stadler et al, J. Virol. 71:8475- 8481, 1997), which is necessary to activate hemagglutinating activity, fiisogenic activity, and infectivity of virions. The M protein is cleaved from its precursor protein (prM) after the consensus sequence R-X-R/K-R (X is variable), and incorporated into the virus lipid envelope together with the E protein.

Cleavage sequences have been conserved not only within flaviviruses, but also within proteins of other, unrelated viruses, such as PE2 of murine coronaviruses, PE2 of alphaviruses, HA of influenza viruses, and pl60 of retroviruses. Cleavage of the precursor protein is essential for virus infectivity, but not particle formation. It was shown that, in case of a TBE-dengue 4 chimera, a change in the prM cleavage site resulted in decreased neurovirulence of this chimera (Pletnev et al, J. Virol. 67:4956-4963, 1993), consistent with the previous observation that efficient processing of the prM is necessary for full infectivity (Guirakhoo et al, 1991, supra; Guirakhoo et al, 1992, supra; Heinz et al, Virology 198:109-117, 1994). Antibodies to prM protein can mediate protective immunity, apparently due to neutralization of released virions that contain some uncleaved prM. The proteolytic cleavage site of the PE2 of VEE (4 amino acids) was deleted by site-directed mutagenesis of the infectious clone (Smith et al. , ASTMH meeting, December 7-11, 1997). Deletion mutants replicated with high efficiency and PE2 proteins were incorporated into particles. This mutant was evaluated in lethal mouse and hamster models and shown to be attenuated; in non-human primates it caused 100% seroconversion and protected all immunized monkeys from a lethal challenge.

Summary of the Invention

The invention features chimeric, live, infectious, attenuated viruses that are each composed of:

(a) a first yellow fever virus (e.g., strain 17D), representing a live, attenuated vaccine virus, in which the nucleotide sequence encoding the prM-E protein is either deleted, truncated, or mutated so that the functional prM-E protein of the first flavivirus is not expressed, and

(b) integrated into the genome of the first flavivirus, a nucleotide sequence encoding the viral envelope (prM-E) protein of a second, different flavivirus, so that the prM-E protein of the second flavivirus is expressed from the altered genome of the first flavivirus.

The chimeric virus is thus composed of the genes and gene products responsible for intracellular replication belonging to the first flavivirus and the genes and gene products of the envelope of the second flavivirus. Since the viral envelope contains antigenic determinants responsible for inducing neutralizing antibodies, the result of infection with the chimeric virus is that such antibodies are generated against the second flavivirus.

A preferred live virus for use as the first yellow fever virus in the chimeric viruses of the invention is YF 17D, which has been used for human immunization for over 50 years. YF 17D vaccine is described in a number of publications, including publications by Smithburn et al. ("Yellow Fever Vaccination," World Health Org., p. 238, 1956), and Freestone (in Plotkin et al, (Eds.), Vaccines, 2^nd edition, W.B. Saunders, Philadelphia, 1995). In addition, the yellow fever virus has been studied at the genetic level (Rice et al, Science 229:726-733, 1985) and information correlating genotype and phenotype has been established (Marchevsky et al, Am. J. Trop. Med. Hyg. 52:75-80, 1995). Specific examples of yellow fever substrains that can be used in the invention include, for example, YF 17DD (GenBank Accession No. U17066), YF 17D-213 (GenBank Accession No. U17067), YF 17D-204 France (X15067, X15062), and YF- 17D-204, 234 US (Rice et al, Science 229:726-733, 1985; Rice et al, New Biologist 1 :285-296, 1989; C 03700, K 02749). Yellow Fever virus strains are also described by Galler et al, Vaccine 16 (9/10):1024-28, 1998.

Preferred flaviviruses for use as the second flavivirus in the chimeric viruses of the invention, and thus sources of immunizing antigen, include Japanese Encephalitis (JE, e.g., JE SA14-14-2), Dengue (DEN, e.g., any of Dengue types 1-4; for example, Dengue-2 strain PUO-218) (Gruenberg et al, J. Gen. Virol. 67:1391-1398, 1988) (sequence appendix 1; nucleotide sequence of Dengue-2 insert; Pr-M: nucleotides 1-273; M: nucleotides 274-498; E: nucleotides 499-1983) (sequence appendix 1; amino acid sequence of Dengue-2 insert; Pr-M: amino acids 1-91; M: amino acids 92-166; E: amino acids 167-661), Murray Valley Encephalitis (MVE), St. Louis Encephalitis (SLE), West Nile (WN), Tick-borne Encephalitis (TBE) (i.e., Central European Encephalitis (CEE) and Russian Spring-Summer Encephalitis (RSSE) viruses), and Hepatitis C (HCV) viruses. Additional flaviviruses for use as the second flavivirus include Kunjin virus, Powassan virus, Kyasanur Forest Disease virus, and Omsk Hemorrhagic Fever virus. As is discussed further below, the second flavivirus sequences can be provided from two different second flaviviruses, such as two Dengue strains. It is preferable to use attenuated inserts, for example, in the case of inserts from neurotropic viruses, such as JE, MVE, SLE, CEE, and RSSE. In the case of non-neurotropic viruses, such as dengue viruses, it may be preferable to use unmodified inserts, from unattenuated strains. Maintenance of native sequences in such inserts can lead to enhanced immunogenicity of the proteins encoded by the inserts, leading to a more effective vaccine.

In a preferred chimeric virus of the invention, the prM-E protein coding sequence of the second flavivirus is substituted for the prM-E protein coding sequence of the live yellow fever virus. Also, as is described further below, the prM portion of the protein can contain a mutation or mutations that prevent cleavage to generate mature membrane protein. Finally, as is discussed in detail below, the chimeric viruses of the invention include the prM signal of yellow fever virus. Also included in the invention are methods of preventing or treating flavivirus infection in a mammal, such as a human, by administering a chimeric flavivirus of the invention to the mammal; use of the chimeric flaviviruses of the invention in the preparation of medicaments for preventing or treating flavivirus infection; nucleic acid molecules encoding the chimeric flaviviruses of the invention; and methods of manufacturing the chimeric flaviviruses of the invention. Further, the invention includes use of the chimeric flaviviruses and of the invention in the preparation of medicaments for administering therapeutic gene products to patients, such as cancer patients. The invention provides several advantages. For example, because they are live and replicating, the chimeric viruses of the invention can be used to produce long-lasting protective immunity. Also, because the viruses have the replication genes of an attenuated virus (e.g., Yellow Fever 17D), the resulting chimeric virus is attenuated to a degree that renders it safe for use in humans.

Other features and advantages of the invention will be apparent from the following detailed description, the drawings, and the claims.

Brief Description of the Drawings Fig. 1 A is a schematic representation of processing events at the C/prM junction of parental viruses that can be used in the invention.

Fig. IB is a schematic representation of the sequences in the capsid, prM signal, and prM regions of flaviviruses that can be used in the invention (SEQ ID NOs:54-70).

Fig. 2 is a schematic representation of the approach to making chimeric flaviviruses at the prM signal region used (SEQ ID NOs:71 and 72) by C.J. Lai (WO 93/06214). Fig. 3 is a schematic representation of an attempt to use the method of C.J. Lai (WO 93/06214) with a yellow fever backbone (SEQ ID NOs:73 and 74).

Fig. 4 is a schematic representation illustrating that the viability of flavivirus chimeras depends on the choice of signal. Fig. 5 is a schematic representation of the cloning method used in the present invention, at the prM signal region (SEQ ID NOs:75-77). Fig. 6 is a schematic representation of the C, prM, E, and NS1 regions and junction sequences of a YF/JE chimera of the invention. The amino acid sequences flanking cleavage sites at the junctions are indicated for JE, YF, and the YF/JE chimera (SEQ ID NOs:78-85).

Fig. 7 is a schematic representation of genetic manipulation steps that were carried out to construct a Yellow-Fever/Japanese Encephalitis (YF/JE) chimeric virus of the invention. Fig. 8 is a set of growth curves for chimeric YF/JE viruses of the invention in cell cultures of vertebrate and mosquito origin.

Fig. 9 is a growth curve of RMS (Research Master Seed, YF/JE SA14-14-2) in Vero and LLC-MK2 cells. Fig. 10 is a graph showing a growth comparison between RMS

(YF/JE SA14-14-2) and YF-Vax in MRC-5 cells.

Fig. 11 A is a graph showing the effects of indomethacin (IM) or 2- aminopurine (2-AP) on growth kinetics of YF/JE SA14-14-2 (0.01 MOI) in FRhL cells. Fig. 1 IB is a graph showing the effects of indomethacin (IM) or 2- aminopurine (2-AP) on growth kinetics of YF/JE SA14-14-2 (0.1 MOI) in FRhL cells.

Fig. 12 is a graph and a table showing the results of a mouse neurovirulence analysis carried out with a YF/JE chimeric virus of the invention.

Fig. 13 is a graph showing the neutralizing antibody response of mice immunized with a YF/JE SA14-14-2 chimeric vaccine of the invention. Three week old mice were immunized, and samples for testing were taken at 6 weeks. Fig. 14A is a graph showing the results of neurovirulence testing of

YF-Vax in 4 week old ICR mice by the i.e. route.

Fig. 14B is a graph showing the results of neurovirulence testing of YF/JE SA14-14-2 in 4 week old ICR mice by the i.e. route.

Fig. 15 is a set of graphs showing the results of PRNT analysis of neutralizing antibody titers in mice inoculated s.c. with graded doses of YF/JE vaccine. The results in the top graph are 3 weeks post immunization, and the results in the bottom graph are 8 weeks post immunization. Fig. 16 is a series of graphs showing the serological responses of mice immunized with a single dose of the live viruses indicated in the figure.

Fig. 17 is a set of graphs showing viremia and GMT of viremia in 3 rhesus monkeys inoculated with ChimeriVax or YF-Vax by the i.e. route.

Fig. 18 is a graph showing the PRNT neutralizing antibody titers (50%) in rhesus monkeys 2 and 4 weeks post inoculation with a single dose of YF-Vax or ChimeriVax vaccines by the i.e. route.

Fig. 19 is a graph showing the results of neurovirulence testing of YF/JE SA14-14-2 (E-138 K— > mutant).

Fig. 20 is a schematic representation of a two plasmid system for generating chimeric YF/DEN-2 virus. The strategy is essentially as described for the YF/JE chimeric virus.

Fig. 21 is a schematic representation of the structure of modified YF clones designed to delete portions of the NSl protein and/or express foreign proteins under control of an internal ribosome entry site (IRES). The figure shows only the E/NS1 region of the viral genome. A translational stop codon is introduced at the carboxyl terminus of the envelope (E) protein. Downstream translation is initiated within an intergenic open reading frame (ORF) by IRES-1, driving expression of foreign proteins (e.g., HCV proteins El and/or E2). The second IRES (IRES-2) controls translational initiation of the YF nonstructural region, in which nested, truncated NSl proteins (e.g., NSldel-1, NSldel-2, or NSldel-3) are expressed. The size of the NSl deletion is inversely proportional to that of the ORF linked to IRES- 1.

Fig. 22 is a graph showing the neurovirulence phenotype of ChimeriVax-Den2 in outbred (CD-I) suckling mice inoculated by the I.C. route with 10,000 PFU/0.02 ml. Fig. 23 is a graph showing the neurovirulence phenotype of 17D vaccine (YF-Vax) in outbred (CD-I) suckling mice inoculated by the I.P. route with 1000 PFU/0.02 ml.

Figs. 24 A-C are graphs showing the growth of JE SA14, JE SA 14- 14-2, ChimeriVax™- JE and YF 17D intrathoracically inoculated mosquitoes: (A) Cx. tritaeniorhynchus mosquitoes, (B) Ae. albopictus mosquitoes, and (C) Ae. aegypti. Mean titer = geometric mean of the titers of three individual mosquitoes; log₁₀ pfu/mosquito.

Figs. 25 A-C are graphs showing the growth of JE SA14, JE SA14-14-2, ChimeriVax™- JE and YF 17D IT orally exposed mosquitoes: (A) Cx. tritaeniorhynchus mosquitoes, (B) Ae. albopictus mosquitoes, and (C) Ae. aegypti mosquitoes. Mean titer = geometric mean of the titers of three individual mosquitoes; log₁₀ pfu/mosquito.

Figs. 26 A and B are graphs showing the growth of virus in IT inoculated Ae. aegypti (A) andAe. albopictus (B) mosquitoes.

Fig. 27 is a schematic representation of an overview of construction of a YF/DEN1 chimera of the invention.

Fig. 28 is a schematic representation of a plasmid and fragment map relating to construction of a YF/DEN1 chimera of the invention. Fig. 29 is a schematic representation of RT-PCR amplification of the prM-E region of the PaH881/88 DEN3 virus genome. The virus genome is shown on the top diagram. Regions encoding hydrophobic signals for corresponding downstream proteins are shadowed. The prM-E region was amplified in two fragments (black solid lines). Restriction sites introduced for subsequent in-frame in vitro ligation into YF backbone (BstBI and Narl) and cloning (Nhel) are indicated.

Fig. 30 is a schematic representation of construction of a YF/DEN3 chimera of the invention. YF- and DEN3-specific sequences are shown as shadowed and black boxes, respectively. The chimeric YF/DEN3 genome was reconstituted by in vitro ligation of three fragments: the large BstBI- Aatll portion of 5'37Den3/DXho plasmid, a PCR fragment containing the DEN3-specific part of 5.2/Den3 without the one nucleotide deletion (Dl) digested with BstBI and Ehel (an isoschizomer of Narl), and the large Ehel-Aatll fragment of YFM5.2 JE SA14-14-2. Ligation products were linearized with Xhol and then transcribed in vitro with SP6 RNA polymerase. Vero PM cells were transfected with in vitro RNA transcripts to recover the chimeric virus. Fig. 31 is a schematic representation of an overview of construction of a YF/DEN4 chimera of the invention.

Fig. 32 is a schematic representation of a plasmid and fragment map relating to construction of a YF/DEN4 chimera of the invention.

Detailed Description

The invention provides chimeric flaviviruses that can be used in vaccination methods against flavivirus infection. Construction and analysis of chimeric flaviviruses of the invention, such as chimeras of yellow fever virus and Japanese Encephalitis (JE), Dengue types 1-4 (DEN 1-4), Murray Valley Encephalitis (MVE), St. Louis Encephalitis (SLE), West Nile (WN), Tick-borne Encephalitis (TBE), and Hepatitis C (HCV) viruses are described as follows.

Yellow fever (YF) virus is a member of the Flaviviridae family of small enveloped positive-strand RNA viruses. Flavivirus proteins are produced by translation of a single long open reading frame to generate a polyprotein, and a complex series of post-translational proteolytic cleavages of the polyprotein by a combination of host and viral proteases, to generate mature viral proteins (Amberg et al, J. Virol. 73:8083-8094, 1999; Fields, "Flaviviridae," In Virology, Fields (ed.), Raven-Lippincott, New York, 1995, Volume I, p. 937). The virus structural proteins are arranged in the order C-prM-E, where "C" is capsid, "prM" is a precursor of the viral envelope-bound M protein, and "E" is the envelope protein. These proteins are present in the N-terminal region of the polyprotein, while the non-structural proteins (NSl, NS2A, NS2B, NS3, NS4A, NS4B, and NS5) are located in the C-terminal region of the polyprotein. The amino termini of prM, E, NSl, and NS4B are generated by host signalase cleavage within the lumen of the endoplasmic reticulum (ER), while most cleavages within the non-structural region are mediated by a viral protease complex known as NS2B-NS3 (Fields, "Flaviviridae," In Virology, Fields (ed.), Raven-Lippincott, New York, 1995, Volume I, p. 937). In addition, the NS2B-NS3 protease complex is responsible for mediating cleavages at the C terminus of both the C protein and the NS4A protein ((Amberg et al, J. Virol. 73:8083-8094, 1999).

Several research efforts suggest a regulatory role for NS2B-NS3-mediated cleavage at the C terminus of the capsid protein. This site is the only site in the structural region of polyprotein that is cleaved by the NS2B-NS3 protease and, in addition, it includes a highly conserved dibasic-site motif of flaviviruses, which indicates a functional role (Amberg et al, J. Virol. 68:3794-3802, 1994; Yamshchikov et al, J. Virol. 68:5765-5771, 1994). In vitro data from various flavivirus models suggest that efficient generation of the prM protein, an early step in proper release of structural proteins, is dependent on the function of the viral protease at the capsid protein site (see summary in Amberg et al, J. Virol. 73:8083-8094, 1999).

Maintenance of this mechanism of coordinate cleavages by NS2B-NS3 at the C terminus of the capsid protein and signalase at the N terminus of prM in the chimeras described below is central to the present invention. In particular, in the chimeras of the present invention, the length of the so-called "prM signal," which separates the two cleavage sites by 20 amino acids in YF (Figs. 1 A and IB), is substantially maintained, to ensure polyprotein proteolytic processing and subsequent growth of chimeric viruses that are created in a YF backbone. A hydrophobic domain within this signal serves to direct the translocation of prM into the ER lumen, where efficient signalase cleavage occurs only after cleavage at the NS2B-NS3 site in the capsid protein (Amberg et al, J. Virol. 73:8083-8094, 1999; Figs. 1A and IB).

In the chimeras of the present invention, only the regions encoding the membrane and envelope proteins (i.e., the prME region) of a non- yellow fever flavivirus are used to replace the corresponding genes in a yellow fever virus clone. The prM signal of the yellow fever virus backbone is maintained. Another method, described in a patent application by C.J. Lai, WO 93/06214, suggests a universal approach to constructing chimeric flaviviruses, involving cloning the prME region of a donor virus into the backbone of an acceptor virus, such that the prM signal sequence is contributed by the incoming prM protein gene. This approach was illustrated using dengue 4 virus as the backbone (acceptor) and tick-borne encephalitis as the donor prME gene. As is illustrated in Fig. 2, the approach described in WO 93/06214 suggests that variability in this cloning strategy, with other chimeric models using flaviviruses as backbone, will have no effect on proper processing of the resulting polyprotein. That is, that flavivirus prM signals are exchangeable when producing viable chimeric viruses. However, attempts to use this approach with YF as a backbone for the insertion of prME genes of dengue 2 virus to create a chimera in which dengue 2 sequences were inserted at the 5' NS2B/NS3 cleavage site (e.g., by R. Galler, see Fig. 3) failed to produce a viable virus, and the use of the 14 amino acid signal of the dengue 2 virus prM in this construct explains why. When the cloning strategy was altered to maintain the YF prM signal of 20 amino acids (i.e., only the Den 2 prME region was replaced), it was successful. The failure to make a viable chimera in YF using the shorter signal of dengue prM demonstrates that the approach described in WO 93/06214 does not work for all chimeras, such as those including a YF backbone.

The explanation of the success of the approach described in WO 93/06214, using a dengue virus backbone, is that both the viral protease and the prM signal of dengue viruses were maintained. The dengue prM signal is 6-8 amino acids shorter than that of other flaviviruses, such as YF, TBE, MVE, and JE. Dengue, and chimeric flaviviruses with a dengue backbone, rely on dengue NS2B-NS3 protease complex for eventual signalase cleavage at the prM signal. Possibly dengue strain evolution favors a short signal sequence for optimum cascade-event processing of structural viral proteins, proper assembly, and virus production. If a longer signal is cloned in a dengue chimera, the amino acid additions favors translocation of prME, and perhaps cleavage, but not necessarily optimal viral growth. On the other hand, YF, TBE, MVE, and JE have evolved using a long prM signal, and cloning of a shorter signal in any of these backbones obliterates C-prM-E processing and viral growth (Fig. 4). Thus, central to the present invention is the length of the prM signal (Fig. 5). The cloning of the prME region of any flavivirus into a YF backbone according to the invention always takes place after the prM signal sequence (i.e., LMTGG/VTL for yellow fever); in this way, all prME chimeras encode the yellow fever prM signal, thus ensuring proper processing of the polyprotein. In addition, it is preferable to maintain the length and sequence of the YF prM signal in the chimeras of the invention. That is, preferably, the length of the prM signal is 20 amino acids. Less preferably, the length of the prM signal is 15, 16, 17, 18, 19, or more than 20 amino acids in length. Also, it is preferable that the amino acid sequence of the YF prM signal is maintained in the chimeras of the invention, although this sequence can be modified using, for example, conservative amino acid substitutions. Preferably, the sequence of the prM signal is 100%, less preferably, 90%, 80%, 70%, 60%, 50%, or 40% identical to the YF prM signal. As an example of construction of a chimera of the invention, Fig. 6 illustrates a YF/JE chimera in which the YF NS2B-NS3 protease recognition site is maintained. Thus, the recognition site for cleavage of the cytosolic from membrane-associated portions of capsid is homologous for the YF NS2B-NS3 enzyme. At the C/pr-M junction, the portion of the signalase recognition site upstream of the cleavage site is that of the backbone, YF, and the portion downstream of the cleavage site is that of the insert, JE. At the E/NS1 junction, the portion of the signalase recognition site upstream of the cleavage site is similar to that of the insert, JE (four of five of the amino acids are identical to those of the JE sequence), and the portion downstream of the cleavage site is that of the backbone, YF. It is preferable to maintain this or a higher level of amino acid sequence identity to the viruses that form the chimera. Alternatively, at least 25, 50, or 75% sequence identity can be maintained in the three to five amino acid positions flanking the signalase and NS2B-NS3 protease recognition sites.

Also possible, though less preferable, is the use of any of numerous known signal sequences to link the C and pre-M or E and NSl proteins of the chimeras (see, e.g., von Heijne, Eur. J. Biochem. 133:17-21, 1983; von Heijne, J. Mol. Biol. 184:99-105, 1985) or, for example, using the known sequences for guidance, one skilled in the art can design additional signal sequences that can be used in the chimeras of the invention. Typically, for example, the signal sequence will include as its last residue an amino acid with a small, uncharged side chain, such as alanine, glycine, serine, cysteine, threonine, or glutamine. Other requirements of signal sequences are known in the art (see, e.g., von Heijne, 1983, supra; von Heijne, 1985, supra).

Following the approach described above, we have succeeded in making viable YF chimeric constructs for JE (section 1) and dengue serotypes 1 through 4 (sections 2-5, respectively). Construction and characterization of these, as well as other, constructs is described further below.

1.0 Construction of cDNA Templates for Generation of YF/JE Chimeric Virus

The derivation of full-length cDNA templates for YF/JE chimeras of the invention described below employed a strategy similar to that earlier workers used to regenerate YF 17D from cDNA for molecular genetic analysis of YF replication. The strategy is described, e.g., by Nestorowicz et al. (Virology 199:114-123, 1994).

Briefly, derivation of a YF/JE chimera of the invention involves the following. YF genomic sequences are propagated in two plasmids (YF5'3TV and YFM5.2), which encode the YF sequences from nucleotides 1-2,276 and 8,279-10,861 (YF5'3TV) and from 1,373-8,704 (YFM5.2) (Rice et al, The New Biologist 1 :285-296, 1989). Full-length cDNA templates are generated by ligation of appropriate restriction fragments derived from these plasmids. This method has been the most reliable for ensuring stable expression of YF sequences and generation of RNA transcripts of high specific infectivity.

Our strategy for construction of chimeras involves replacement of YF sequences within the YF5'3'IV and YFM5.2 plasmids by the corresponding JE sequences from the start of the prM protein (nucleotide 478, amino acid 128) through the E/NS1 cleavage site (nucleotide 2,452, amino acid 817). In addition to cloning of JE cDNA, several steps were required to introduce or eliminate restriction sites in both the YF and JE sequences to permit in vitro ligation. The structure of the template for regenerating chimeric YF (C)/JE (prM-E) virus is shown in Fig. 7. A second chimera, encoding the entire JE structural region (C-prM-E) was engineered using a similar strategy. The second chimera was not able to generate RNA of high infectivity.

1.1 Molecular Cloning of the JE Virus Structural Region

Clones of authentic JE structural protein genes were generated from the JE SA14-14-2 strain (JE live, attenuated vaccine strain), because the biological properties and molecular characterization of this strain are well- documented (see, e.g., Eckels et al, Vaccine 6:513-518, 1988; JE SA14- 14-2 virus is available from the Centers for Disease Control, Fort Collins, Colorado and the Yale Arbovirus Research Unit, Yale University, New Haven, Connecticut, which are World Health Organization-designated Reference Centers for Arboviruses in the United States). JE SA14-14-2 virus at passage level PDK-5 was obtained and passaged in LLC-MK₂ cells to obtain sufficient amounts of virus for cDNA cloning. The strategy used involved cloning the structural region in two pieces that overlap at an Nhel site (JE nucleotide 1,125), which can then be used for in vitro ligation. RNA was extracted from monolayers of infected LLC-MK₂ cells and first strand synthesis of negative sense cDNA was carried out using reverse transcriptase with a negative sense primer (JE nucleotide sequence 2,456-71) containing nested Xbal and Nαrl restriction sites for cloning initially into pBluescript II KS(+), and subsequently into YVM.5.2(Narl), respectively. First strand cDΝA synthesis was followed by PCR amplification of the JE sequence from nucleotides 1,108-2,471 using the same negative sense primer and a positive sense primer (JE nucleotides sequence 1,108-1,130) containing nested Xbal and Ny I restriction sites for cloning into pBluescript and YFM5.2(NαrI), respectively. JE sequences were verified by restriction enzyme digestion and nucleotide sequencing. The JE nucleotide sequence from nucleotides 1 to 1,130 was derived by PCR amplification of negative strand JE cDΝA using a negative sense primer corresponding to JE nucleotides 1,116 to 1,130 and a positive sense primer corresponding to JE nucleotides 1 to 18, both containing an EcoRl restriction site. PCR fragments were cloned into pBluescript and JE sequences were verified by nucleotide sequencing. Together, this represents cloning of the JE sequence from nucleotides 1-2,471 (amino acids 1-792).

1.2 Construction of YF5 '3 'IV/JE and YFM5.2/JE Derivatives

To insert the C terminus of the JE envelope protein at the YF E/ΝS 1 cleavage site, a unique Narl restriction site was introduced into the YFM5.2 plasmid by oligonucleotide-directed mutagenesis of the signalase sequence at the E/ΝS1 cleavage site (YF nucleotides 2,447-2,452, amino acids 816-817) to create YFM5.2(NαrI). Transcripts derived from templates incorporating this change were checked for infectivity and yielded a specific infectivity similar to the parental templates (approximately 100 plaque-forming units/250 nanograms of transcript). The JE sequence from nucleotides 1,108 to 2,471 was subcloned from several independent PCR-derived clones of pBluescript/JE into YFM5.2(Narϊ) using the unique Nsil and Narl restriction sites. YF5'3TV/JE clones containing the YF 5' untranslated region (nucleotides 1-118) adjacent to the JE prM-E region were derived by PCR amplification.

To derive sequences containing the junction of the YF capsid and JE prM, a negative sense chimeric primer spanning this region was used with a positive sense primer corresponding to YF5'3TV nucleotides 6,625- 6,639 to generate PCR fragments that were then used as negative sense PCR primers in conjunction with a positive sense primer complementary to the pBluescript vector sequence upstream of the EcoRl site, to amplify the JE sequence (encoded in reverse orientation in the pBluescript vector) from nucleotide 477 (N-terminus of the prM protein) through the Nhel site at nucleotide 1,125. The resulting PCR fragments were inserted into the YF5'3TV plasmid using the Notl and EcoRl restriction sites. This construct contains the SP6 promoter preceding the YF 5'-untranslated region, followed by the sequence: YF (C) JE (prM-E), and contains the Nhel site (JE nucleotide 1,125) required for in vitro ligation.

1.3 Engineering YFM5.2 and YF5'3'IVto Contain Restriction Sites for in vitro Ligation

To use the NTzel site within the JE envelope sequence as a 5' in vitro ligation site, a redundant Nhel site in the YFM5.2 plasmid (nucleotide 5,459) was eliminated. This was accomplished by silent mutation of the YF sequence at nucleotide 5,461 (T→C; alanine, amino acid 1820). This site was incorporated into YFM5.2 by ligation of appropriate restriction fragments and introduced into YFM5.2(NαrI)/JE by exchange of an Nsil/Narl fragment encoding the chimeric YF/JE sequence.

To create a unique 3' restriction site for in vitro ligation, a BspEl site was engineered downstream of the ^ tll site normally used to generate full-length templates from YF5'3TV and YFM5.2. (Multiple Aatll sites are present in the JE structural sequence, precluding use of this site for in vitro ligation.) The BspEl site was created by silent mutation of YF nucleotide 8,581 (A→C; serine, amino acid 2,860), and was introduced into YFM5.2 by exchange of appropriate restriction fragments. The unique site was incorporated into YFM5.2/JE by exchange of the

XballSphl fragment, and into the YF5'3TV/JE(prM-E) plasmids by three- piece ligation of appropriate restriction fragments from these parent plasmids and from a derivative of YFM5.2 (BspEl) deleting the YF sequence between the EcoRl sites at nucleotides 1 and 6,912.

1.4 Exchange of JE Nakayama cDNA into YF/JE Chimeric Plasmids

Because of uncertainty about the capacity of the PCR-derived JE SA14-14-2 structural region to function properly in the context of the chimeric virus, we used cDΝA from a clone of the JE Nakayama strain that has been extensively characterized in expression experiments and for its capacity to induce protective immunity (see, e.g., Mclda et al, Virology 158:348-360, 1987; the JE Nakayama strain is available from the Centers for Disease Control, Fort Collins, Colorado, and the Yale Arbovirus Research Unit, Yale University, New Haven, Connecticut). The Nakayama cDNA was inserted into the YF/JE chimeric plasmids using available restriction sites (Hindlϊl to Pvull and Bpml to Mun ) to replace the entire prM-E region in the two plasmid system except for a single amino acid, serine, at position 49, which was left intact in order to utilize the Nhel site for in vitro ligation. The entire JE region in the Nakayama clone was sequenced to verify that the replaced cDNA was authentic (Table 1).

7.5 Generation of Full-Length cDNA Templates, RNA Transfection, and Recovery of Infectious Virus

Procedures for generating full-length cDNA templates are essentially as described in Rice et al. (The New Biologist 1 :285-96, 1989; also see Fig. 7). In the case of chimeric templates, the plasmids YF5'3^,IV/JE(prM-E) and YFM5.2/JE are digested with Nhel/BspEl and in vitro ligation is performed using 300 nanograms of purified fragments in the presence of T4 DNA ligase. The ligation products are linearized with Xhol to allow run-off transcription. SP6 transcripts are synthesized using 50 nanograms of purified template, quantitated by incorporation of ³H- UTP, and integrity of the RNA is verified by non-denaturing agarose gel electrophoresis. Yields range from 5 to 10 micrograms of RNA per reaction using this procedure, most of which is present as full-length transcripts. Transfection of RNA transcripts in the presence of cationic liposomes is carried out as described by Rice et al. (supra) for YF 17D. In initial experiments, LLC-MK₂ cells were used for transfection and quantitation of virus, since we have determined the permissiveness for replication and plaque formation of the parental strains of YF and JE. Table 2 illustrates typical results of transfection experiments using Lipofectin (GIBCO/BRL) as a transfection vehicle. Vero cell lines have also been used routinely for preparation of infectious virus stocks, characterization of labeled proteins, and neutralization tests.

Amplification products from Vero cells were sent to the FDA (CBER) for preparation of the RMS in diploid, Fetal Rhesus lung cells. Fetal rhesus lung cells were received from the ATCC as cultured cells and were infected with YF/JE SA14-14-2 (clone A-l) at an MOI of 1.0. After 1 hour of incubation at 37 °C, the inoculum was aspirated and replaced with 50 ml of EMEM, containing 2% FBS. Virus was harvested 78 hours later, aliquoted into 1 ml vials (a total of 200 vials) and frozen at -70 °C. Virus titers were determined in Vero, LLC MK2, and CV-1 cells using a standard plaque assay. Titers (pfu/ml) were 1.6 x 10⁶ in Vero cells, 1.25 x 10⁶ in LLC MK2 cells, and 1.35 x 10⁵ in CV-1 cells.

1.6 Nucleotide Sequencing of Chimeric cDNA Templates Plasmids containing the chimeric YF/JE cDNA were subjected to sequence analysis of the JE portion of the clones to identify the correct sequences of the SA14-14-2 and Nakayama envelope protein. The nucleotide sequence differences between these constructs in comparison to the reported sequences (McAda et al, supra) are shown in Table 1. Five amino acid differences at positions 107, 138, 176, 264, and

279 separate the virulent from the attenuated strains of JE virus. Amino acid differences map to three subregions of Domains I and II of the flavivirus E protein model (Rey et al, Nature 375:291-298, 1995). These include the putative fusion peptide (position 107), the hinge cluster (positions 138, 279), the exposed surface of Domain I (positions 176 and 177), and the alpha-helix located in the dimerization Domain II (position 264). Changes at position 107, 138, 176, and 279 were selected early in the passage history, resulting in attenuation of JE SA14-14-2, and remained stable genetic differences from the SA₁₄-14-2 parent (Ni et al, J. Gen. Virol. 75: 1505-1510, 1994), showing that one or more of these mutations are critical for the attenuation phenotype. The changes at positions 177 and 264 occurred during subsequent passage, and appear to be genetically unstable between two SA 14- 14-2 virus passages in PHK and PDK cells, showing that this mutation is less critical for attenuation. The nucleotide sequence of the E protein coding region of the RMS was determined to assess potential sequence variability resulting from viral passage. Total RNA was isolated from RMS-infected Vero cells, reversed transcribed, and PCR amplified to obtain sequencing templates. Several primers specific for SA14-14-2 virus were used in individual sequencing reactions and standard protocols for cycle sequencing were performed.

Sequence data revealed two single nucleotide mutations in the RMS E protein, when compared to the published SA14-14-2 JE strain sequence data. The first mutation is silent, and maps to amino acid position 4 (CTT to CTG); the second is at amino acid position 243 (AAA to GAA) and introduces a change from lysine to glutamic acid. Both mutations identified are present in the sequence of the JE wild type strains Nakayama, SA14 (parent of SA14-14-2), and JaOArS982 (Sumiyoshi et al, J. Infect. Dis. 171 : 1144-1151, 1995); thus, they are unlikely to contribute to virulence phenotype. We conclude that in vitro passage in FRhL cells to obtain the RMS did not introduce unwanted mutations in the E protein.

1.7 Structural and Biological Characterization of Chimeric YF/JE Viruses The genomic structure of chimeric YF/JE viruses recovered from transfection experiments was verified by RT/PCR-based analysis of viral RNA harvested from infected cell monolayers. These experiments were performed to eliminate the possibility that virus stocks were contaminated during transfection procedures. For these experiments, first-pass virus was used to initiate a cycle of infection, to eliminate any possible artifacts generated by the presence of residual transfected viral RNA. Total RNA extracts of cells infected with either the YF/JE (prM-E)-SA14-14-2 or YF/JE (prM-E)-Nakayama chimera were subjected to RT/PCR using YF and JE-specific primers that allowed recovery of the entire structural region as two PCR products of approximately 1 kilobase in size. These products were then analyzed by restriction enzyme digestion using the predicted sites within the JE SA14-14-2 and Nakayama sequences that allow differentiation of these viruses. Using this approach, the viral RNA was demonstrated to be chimeric and the recovered viruses were verified to have the appropriate restriction sites. The actual C-prM boundary was then verified to be intact at the sequence level by cycle sequence analysis across the chimeric YF/JE C-prM junction.

The presence of the JE envelope protein in the two chimeras was verified by both immunoprecipitation with JE-specific antisera and by plaque reduction neutralization testing using YF and JE-specific antisera. Immunoprecipitation of ³⁵S-labeled extracts of LLC-MK₂ cells infected with the chimeras using a monoclonal antibody to the JE envelope protein showed that the JE envelope protein could be recovered as a 55 kDa protein, while the same antisera failed to immunoprecipitate a protein from YF-infected cells. Both JE and YF hyperimmune sera demonstrated cross- reactivity for the two envelope proteins, but the size difference between the proteins (YF=53 kDa, unglycosylated; JE=55 kDa, glycosylated) could reproducibly be observed. Use of YF monoclonal antibodies was not satisfactory under the immunoprecipitation conditions, thus, the specificity was dependent on the JE monoclonal antibodies in this analysis.

Plaque reduction neutralization testing (PRNT) was performed on the chimeric viruses and the YF and JE SA 14- 14-2 viruses using YF and JE-specific hyperimmune ascitic fluid (ATCC) and YF-specific purified IgG (monoclonal antibody 2E10). Significant differences in the 50% plaque reduction titer of these antisera were observed for the chimeras when compared to the control viruses in these experiments (Table 3). The YF/JE SA14-14-2 chimeric vaccine candidate, as well as the Nakayama chimera and SA14-14-2 viruses, were neutralized only by JE ascitic fluid, whereas YF 17D was neutralized in a specific fashion by YF ascites and the monoclonal antibody (Table 3). Thus, epitopes required for neutralization are expressed in the infectious chimeric YF/JE viruses, and are specific for the JE virus.

1.8 Growth Properties in Cell Culture

The growth capacity of the chimeras has been examined quantitatively in cell lines of both primate and mosquito origin. Fig. 8 illustrates the cumulative growth curves of the chimeras on LLC-MK₂ cells after low multiplicity infection (0.5 plaque-forming units/cell). In this experiment, YF5.2iv (cloned derivative) and JE SA14-14-2 (uncloned) viruses were used for comparison. Both chimeric viruses reached a maximal virus yield of approximately one log higher than either parental virus. In the case of the YF/JE S A ₁₄- 14-2 chimera, the peak of virus production occurred 12 hours later than the YF/JE Nakayama chimera (50 hours vs. 38 hours). The YF/JE Nakayama chimera exhibited considerably more cytopathic effects than the YF/JE SA14-14-2 chimera on this cell line. A similar experiment was carried out in C6/36 cells after low multiplicity infection (0.5 plaque-forming units/cell). Fig. 8 also illustrates the growth kinetics of the viruses in this invertebrate cell line. Similar virus yields were obtained at all points used for virus harvest in this experiment, further substantiating the notion that chimeric viruses are not impaired in replication efficiency.

Additional experiments showing the growth properties of RMS are shown in Fig. 9. Briefly, Vero cells were grown in EMEM, 1% L- Glutamine, 1% non-essential amino acid, and 10% FBS buffered with sodium bicarbonate. LLC-MK2 cells were purchased from the ATCC (CLL-7.1, passage 12) and were grown in the same medium as Vero cells. Cells were inoculated with the RMS virus at an MOI of 0.1. Supernatant fluid was sampled at 24 hour intervals for 7 days and frozen at -70 °-C for subsequent plaque assay. Plaque assays were performed in 6-well plates. The RMS reached more than 8 log₁₀ pfu/ml in 5 days. In LLC-MK2 cells, the RMS grew slower and peaked (6 log₁₀ pfu/ml) at about 6 days.

1.9 Comparison of Growth Kinetics of the RMS (YF/JE SA14-14-2) with YF 17D Vaccine in MRC-5 Cells

An experiment was performed to assess the ability of the vaccine candidate to propagate in a cell line acceptable for human vaccines. Commercial Yellow Fever 17D vaccine (YF-Vax®) was obtained from Connaught Laboratories, Swiftwater, PA. MRC-5 (diploid human embryonal lung cells) were purchased from ATCC (171 -CCL, Batch#: F- 14308, passage 18) and grown in EMEM, 2 mM L-Gln, Earle's BSS adjusted to contain 1.5 g/L sodium bicarbonate, 0.1 mM non-essential amino acids, and 10% FBS. To compare growth kinetics of RMS (sequence appendices 2 and 3;

Research Master Seed, YF/JE SA14-14-2; nucleotide sequence of ORF; C: nucleotides 119-421; Pr-M: nucleotides 422-982; E: nucleotides 983-2482; and Non-structural proteins: 2483-10381); (amino acid sequence of ORF; C: amino acids 1-101; Pr-M: amino acids 102-288; E: amino acids 289- 788; and Non-structural proteins: amino acids 789-3421); (nucleotide sequence of RMS; the coding sequence is from nucleotide 119 to nucleotide 10381)) with YF-Vax®, cells were grown to 90% confluency and infected with RMS or YF-Vax® at an MOI of 0.1 pfu. Since MRC-5 cells generally grow slowly, these cells were kept for 10 days post inoculation. Samples were frozen daily for 7-10 days and infectivity determined by plaque assay in Vero cells. YF-Vax® and the YF/JE chimera grew to modest titers in MRC-5 cells (Fig. 10). The peak titer was ~4.7 log₁₀ pfu for YF-Vax® achieved on the second day and was slightly lower, 4.5 log₁₀ pfu, for the RMS after 6 days.

1.10 Growth Curve of YF/JE SA14-14-2 in FRhL cells with and without IFN-inhibitors

Fetal rhesus lung cells were obtained from the ATCC and propagated as described for MRC-5 cells. Growth kinetics of the RMS were determined with and without interferon inhibitors.

Double-stranded RNA appears to be the molecular species most likely to induce interferon (IFN) in many virus infected cells. Induction of interferon apparently plays a significant role in the cellular defense against viral infection. To escape cellular destruction, many viruses have developed strategies to down-regulate induction of interferon-dependent activities. Sindbis virus and vesicular stomatitis virus have been shown to be potent IFN inducers. Using chick embryo cells, mouse L cells, and different viral inducers of IFN, it was shown that 2-aminopurine (2AP) and indomethacin (IM) efficiently and reversibly inhibit IFN action (Sekellick et al, J. IFN Res. 5:651, 1985; Marcus et al, J. Gen. Virol. 69:1637, 1988). To test whether inhibition of IFN (if present) in FRhL cells will increase the virus yield, we added 2AP at a concentration of 10 mM or IM at a concentration of 10 mg/ml to the FRhL cells at the time of infection with 0.1 or 0.01 MOI of RMS. Samples were taken daily and frozen for determination of virus infectivity by plaque assay. As shown in Fig. 11 A, virus titers peaked on day 4 in the presence or absence of inhibitors. When cells were infected at 0.01 MOI (Fig. 11 A), virus titer reached 2.65 x 10 ⁷ pfu/ml on day 4 in the absence of inhibitors. In cells infected in the presence of IM, virus titer was increased about 2-fold, to 5.95 x 10⁷ pfu/ml on day 4. This increase was more dramatic (4- fold) when 2AP was used (9.7 x 10⁷ pfu/ml). Addition of IM did not increase virus yield when cells were infected at a higher MOI (0.1). A titer of 5.42 x 10⁷ was reached without inhibitor and 3.45 x 10⁷ was achieved in the presence of IM. Addition of 2AP increased virus yields to 1.1 x 10⁸ pfu/ml by day 4 and only 1 log₁₀ pfu was lost in the following 3 days (9.5 x 10⁶ pfu/ml on day 7) (Fig. 1 IB). We conclude from this experiment that the YF/JE SA14- 14-2 vaccine candidate replicates to titers of ~7.5 log₁₀/ml in an acceptable cell substrate. The addition of interferon inhibitors can result in a modest increase in yields, but is not a requirement for vaccine production.

7.77 Neurovirulence Testing in Normal Adult Mice

The virulence properties of the YF/JE SA14-14-2 chimera was analyzed in young adult mice by intracerebral inoculation. Groups of 10 mice (4 week old male and female ICR mice, 5 each per group) were inoculated with 10,000 plaque-forming units of the YF/JE S A 14- 14-2 chimera, YF 17D 5.2iv, or the Chinese vaccine strain JE SA14-14-2 and observed daily for 3 weeks. The results of these experiments are illustrated in Fig. 12. Mice receiving the YF5.2iv parent succumbed by approximately one week post-inoculation. No mortality or illness was observed among mice receiving either the JE SA14-14-2 parent or the chimera. The inocula used for the experiments were titered at the time of injection and a subgroup of the surviving mice were tested for the presence of neutralizing antibodies to confirm that infection had taken place.

Among those tested, titers against the JE SA14-14-2 virus were similar for animals receiving either this strain or the chimera.

The results of additional experiments investigating the neurovirulence of the YF/JE SA 14- 14-2 chimera in mice are illustrated in Table 4. In these experiments, all of the mice inoculated with YF5.2iv died within 7-8 days. In contrast, none of the mice inoculated with YF/JE SA 14- 14-2 died during two weeks of post-inoculation observation.

The results of experiments investigating the neuroinvasiveness and pathogenesis of YF/JE chimeras are illustrated in Table 5. In these experiments, the chimeric viruses were inoculated into 3 week old mice at doses varying between 10,000 and 1 million plaque-forming units via the intraperitoneal route. None of the mice inoculated with YF/JE Nakayama or YF/JE SA14-14-2 died during three weeks of post- inoculation observation, indicating that the virus was incapable of causing illness after peripheral inoculation. Mice inoculated with YF/JE SA14-14-2 developed neutralizing antibodies against JE virus (Fig. 13).

In additional experiments testing the neurovirulence phenotype and immunogenicity of the RMS, 4- week old ICR mice (n=5) were inoculated by the i.e. route with 0.03 ml of graded doses of the RMS or YF-Vax® (Table 6). Control mice received only diluent medium by this route. Mice were observed daily and mortality rates were calculated.

Mice inoculated with YF-Vax® started to die on day 7 (Fig. 14A). The icLD₅₀ of unpassaged YF-Vax®, calculated by the method of Reed and Muench, was 1.62 log₁₀ and the average survival time (AST) at the highest dose (4.2 log₁₀ pfu) was 8.8 days. In contrast, all mice receiving the RMS survived challenge at all doses (Fig. 14B), indicating that the virus is not neuro virulent for mice. None of the mice inoculated with YF- Vax® or the RMS by the peripheral (subcutaneous) route (as shown in Table 6) showed signs of illness or death. Thus, as expected, yellow fever 17D virus was not neuroinvasive.

7.72 Comparison of immunogenicity of YF/JE RMS with YF 17D vaccine The immunogenicity of the of the RMS was compared with that of the YF 17D vaccine in outbred ICR mice. Groups of five 4 week-old mice received graded doses of the vaccines shown in Table 6. Mice were inoculated with 100 μl of each virus dilution by the s.c. route. For comparison, two groups of mice received two weekly doses of commercial inactivated JE vaccine prepared in mouse brain tissue (JE-Vax®) at 1 :30 and 1 :300 dilution, representing lOx and lx the human equivalent dose based on body weight, respectively. Animals were bled 3 and 8 weeks later and neutralizing antibody titers were measured in heat-inactivated sera against homologous viruses by PRNT. End-point titers were the highest dilution of sera which reduced the number of viral plaques by 50% compared to a normal mouse serum control.

The highest N antibody titers were observed 8 weeks after immunization in mice receiving 5 log₁₀ pfu of the RMS (Fig. 15 and Table 7). The geometric mean N antibody titer in these mice was 5,614. N antibody responses induced by YF/JE SA14-14-2 vaccine against JE were higher than N antibody responses against YF induced by YF 17D vaccine. Interestingly, the highest concentration of the YF 17D vaccine did not induce significant titers of neutralizing antibodies 3 or 8 weeks post immunization, but antibodies were elicited at lower doses.

Very low doses (1.4-2.4 log₁₀ PFU) of YF 17D vaccine elicited an immune response in mice 8 weeks after inoculation (Table 7). This result may indicate delayed replication of the vaccine in mice receiving low virus inocula. In contrast, the YF/JE SA14-14-2 chimeric vaccine in this dose range was not immunogenic. It is likely that the chimeric vaccine is somewhat less infectious for mice than YF 17D. However, when inoculated at an infective dose, the chimera appears to elicits a higher immune response. This may be due to higher replication in, or altered tropism for, host tissues. Animals that received two doses of JE-Vax® did not mount a significant antibody response. Only one animal in the 1 :30 dose group developed a neutralizing titer of 1 : 10 eight weeks after immunization. This might be due to the route (s.c.) and dilution (1 :30) of the vaccine.

7.73 Protection of YF/JE SA14-14-2 RMS immunized mice against challenge with virulent JE

The YF/JE SA14-14-2 RMS and other viruses were evaluated for immunogenicity and protection in C57/BL6 mice in collaboration with Dr. Alan Barrett, Department of Pathology, University of Texas Medical Branch, Galveston. Experimental groups are shown in Table 8. Ten- fold dilutions (10²-10⁵) of each virus were inoculated by the s.c. route into groups of 8 mice. Mice were observed for 21 days, at which time surviving animals were bled from the retro-orbital sinus and serum frozen for neutralization tests. The 50% immunizing dose (ID₅₀) for each virus and GMT was determined (see below). Surviving mice that received viruses by the s.c. route were challenged on day 28 by i.p. inoculation of 158 LD₅₀ (2,000 PFU) of JE virus (JaOArS982, IC37). Animals were observed for 21 days following challenge. Protection is expressed as the proportion of mice surviving challenge (Table 9).

As expected, YF 17D virus afforded minimal cross-protection against JE challenge. The YF/JE SA 14- 14-2 RMS chimera was protective at doses >10³ PFU. The 50% protective dose of the chimeric vaccine was 2.32 log₁₀ PFU. Animals that received 3 doses of JE-Vax were solidly protected against challenge. Mice given a single dose of the SA 14- 14-2 vaccine were poorly protected. Wild-type Nakayama virus was lethal for a proportion of animals, in a dose-dependent fashion; survivors were poorly protected against challenge indicating that the lethal dose was close to the infecting dose for this virus. The F/ E_j^ _jna chimeric virus was somewhat more virulent than the Nakayama strain, in that all mice given 10⁵ of the chimera died after inoculation. This is in contrast to earlier studies in outbred mice, in which this virus was not neuroinvasive, confirming the increased susceptibility of C57/BL6 mice to peripheral challenge with JE viruses. Survivors were fully protected against challenge, showing that the infection established by the chimeric virus was more active (immunogenic) than infection by Nakayama virus without the YF replication background. These results show that the combination of viral envelope determinants of a neurovirulent strain (Nakayama) with a replication-efficient virus (YF 17D) can enhance virulence of the recombinant, emphasizing the need for genetic stability of the mutations conferring attenuation in the

YF/JENa ayama ChimeiH. 1.14 Serological response

Sera from mice in groups shown in Table 8 were tested 21 days after immunization for neutralizing antibodies. N tests were performed as follows. Six- well plates were seeded with Vero cells at a density of 10⁶ cells/well in MEM alpha containing 10% FBS, 1% nonessential amino acids, buffered with sodium bicarbonate. One hundred μl of each test serum (inactivated at 60 °C for 30 minutes) diluted two-fold was mixed with an equal volume of virus containing 200-300 PFU. The virus-serum mixtures were incubated at 4°C overnight and 100 μl added to each well after removal of growth medium. The plates were overlaid after 1 hour incubation at 37 °C with 0.6% agarose containing 3% fetal calf serum, 1% L-glutamine, 1% HEPES, and 1% pen-strep-amphotericin mixed 1 :1 with 2x M199. After 4 days of incubation at 37°C, 5% C0₂, a second overlay containing 3% Neutral red was added. After appearance of plaques, the monolayer was fixed with 1% formaldehyde and stained with crystal violet. The plaque reduction titer is determined as the highest dilution of serum inhibiting > 50% of plaques compared with the diluent- virus control.

Results are shown in Table 10 and Fig. 16. NT antibody responses in mice immunized with the YF/JE SA14-14-2 chimera showed a dose response and good correlation with protection. At doses of 4-5 logs, the chimeric vaccine elicited higher N antibody responses against JE than either SA14-14-2 virus or wild-type Nakayama virus. Responses were superior to those elicited by YF-Vax® against YF 17D virus. No prozone effect was observed in animals receiving the chimera or infectious-clone derived YF 5.2iv; responses at the highest vaccine dose (5 logs) were higher than at the next lower dose (4 logs). In contrast, mice that received SA14-14-2, Nakayama, and YF-Vax® at the highest dose responded less well than animals inoculated with diluted virus.

7.75 Safety and Immunogenicity of ChimeriVax^'™ in Monkeys The safety of RMS was tested in monkeys, essentially as described in WHO Biological Standards for YF 17D vaccine with minor modifications (see below). Two groups (N=3) of rhesus monkeys were bled and shown to be free from HI antibodies to YF, JE, and SLE. Group 1 received undiluted ChimeriVax™ (Vero-passage 2) by the I.C. route (frontal lobe). Group 2 (N=3) received 0.25 ml of 1 : 10 diluted commercial YF 17D vaccine (YF-Vax®) by the same route. The virus inocula were frozen, back titrated, and shown to contain 7.0 and 5.0 log₁₀ pfu /0.25 ml of YF/JE SA14-14-2 and YF-Vax®, respectively.

Monkeys were observed daily for clinical signs and scored as in WHO standards. Sera were collected daily for 7 days after inoculations and tested for viremia by plaque assay in Vero cells. Blood collected 2 and 4 weeks post inoculation and tested for NT antibodies to the homologous viruses. None of the monkeys showed sign of illness. Monkeys were euthanized on Day 30, and brains and spinal cords were examined for neuropathology as described in the WHO standards. A sample of the brain and spinal cord from each animal was collected and stored frozen for virus isolation attempts and immunocytochemistry experiments.

As shown in Fig. 17, a low level viremia was detected in all animals in both groups, and lasted for 2-3 days for the RMS and 1-2 days for YF-Vax. All viruses were cleared from the blood by Day 4. According to the WHO standards, monkeys receiving 5,000-50,000 (3.7- 4.7 log₎₀) pfu should not have viremia greater than 165,000 pfu/ml (approximately 16,500 mLD₅₀). None of the monkeys in the experiments had viremia of more than 15,000 pfu/ml, despite receiving 6 log₁₀ pfu of the RMS. Neutralizing antibody titers were measured at 2 and 4 weeks post inoculation (Fig. 18). All monkeys seroconverted and had high titers of neutralizing antibodies against the inoculated viruses. The level of neutralizing antibodies in 2 of 3 monkeys in both groups exceeded a titer of 1 : 6,400 (the last dilution of sera tested) at 4 weeks post inoculation. The geometric mean antibody titers for ChimeriVax were 75 and 3,200 after 2 and 4 weeks respectively and were 66 and 4971 for the YF-Vax® for the same time points (Table 11).

Histopathological examination of coded specimens of brain and spinal cord were performed by an expert neuropathologist (Dr. I. Levenbook, previously CBER FDA), according to the WHO biological standards for yellow fever vaccine. There were no unusual target areas for histopathological lesions in brains of monkeys inoculated with ChimeriVax™-JE. Mean lesion scores in discriminator areas were similar in monkeys inoculated with YF-Vax® (0.08) and monkeys inoculated with a 100-fold higher dose of ChimeriVaxTM- JE (0.07). Mean lesion scores in discriminator + target areas were higher in monkeys inoculated with YF-Vax® (0.39) than in monkeys inoculated with a 100-fold higher dose of ChimeriVax™- JE (0.11). These preliminary results show an acceptable neurovirulence profile and immunogenicity for ChimeriVax™- JE vaccine. A summary of the histopathology results is provided in Table 22. 1.16 Efficacy of YF/JE Chimera in Protecting Monkeys Against Intracerebral Challenge

The YF/JE chimera were given to adult rhesus monkeys without pre-existing flavivirus immunity by the subcutaneous route. Three monkeys received 4.3 log pfu and three monkeys received 5.3 log pfu of YF/JE SA14-14-2 virus. All 6 monkeys developed very low level (1-2 log/ml) viremias. All animals developed neutralizing antibodies by day 15 (earliest time tested) and titers rose by day 30. Five of six animals survived a very severe intracerebral challenge with a highly virulent JE virus (100,000 mouse LD50 were injected IC 60 days after immunization). None of 4 sham immunized monkeys survived; all died between days 8-10 after challenge. The single death in the immunized group was a pregnant female; pregnancy could have suppressed the cellular immune response to the vaccine. The results show the immunogenicity and protective efficacy of the vaccine, while validating safety with respect to low vaccine viremia. The results of these experiments are illustrated in Tables 12-15.

7.77 Genetic stability of the RMS

The E protein of the attenuated SA 14- 14-2 virus used to construct the YF/JE chimera differs from its virulent parent (SA14 or Nakayama) at 6 positions; 107, 138, 176, 177, 264, and 279. Because the presence of a single residue controlling virulence would be a disadvantage for any vaccine candidate because of the potential for reversion, studies are being undertaken to determine which residue(s) are responsible for attenuation and in particular whether a single residue is responsible for the difference. 7.75 Position 138 on the E protein

A single mutation of an acidic residue glutamic acid (E) to a basic residue, lysine (K) at position 138 on the E protein of JE virus results in attenuation (Sumy'oshi et al, J. Infect. Dis. 171 :1144, 1995). Experiments were carried out to determine whether the amino acid at position 138 of the JE envelope protein (K in the vaccine chimera and E in the virulent Nakayama chimera) is a critical determinant for neurovirulence in mice. Chimeric YF/JE SA14-14-2 (K 138 — > E) virus containing the single reversion of K — >E at position 138 was generated from an engineered cDNA template. The presence of the substitution and the integrity of the entire E protein of the resulting virus was verified by RT/PCR sequencing of the recovered virus. A standard fixed-dose neurovirulence test of the virus was conducted in 4- week-old outbred mice by i.e. inoculation with 10⁴ pfu of virus. The YF/JE SA14-14-2 and YF/JE Nakayama chimeric viruses were used as controls. The virulence phenotype of YF/JE SA14-14-2 (K— ->E) was indistinguishable from that of its attenuated parent YF/JE SA14- 14-2 in this assay, with no morbidity or mortality observed in the mice during the observation period (Fig. 19). We conclude that the single mutation at position 138 to the residue found in the JE-Nakayama virus does not exert a dominant effect on the neurovirulence of the YF/JE SA 14- 14-2 chimera, and that one or more additional mutations are required to establish the virulent phenotype.

7.79 Other putative attenuation loci

Additional experiments to address the contributions of the other 6 residues (mentioned above) using the format described here were conducted. The mutant viruses constructed by site directed mutagenesis of the YF and JE infectious clones are listed in Table 16. The E proteins of these viruses were sequenced and confirmed to contained the desired mutations. Upon inoculation into weanling mice by the I.C. route it is possible to determine those residues involved in attenuation of the vaccine.

Additional experiments to address the contributions of other residues are underway. The mutant viruses constructed to date by site-directed mutagenesis of the YF and JE infectious clones are listed in Table 16. The methodology is as described above. Results to date confirm that at least two and possibly more than 2 mutations are responsible for the attenuation phenotype of YF/JE SA14-14-2 virus (Table 23).

1.20 Stability of the RMS in Tissue Cultures: Characterization of Genetic Changes, Neurovirulence and Immunogenicity, Serial Passages In Vitro

The RMS was used to inoculate a T75 flask of FRhL2 cells at an m.o.i. of 0.1. Subsequent passages were carried out in T75 flasks and harvested 3 days post-inoculation. At each passage, the culture supernatant was assumed to hold 10⁷ pfu/ml and an aliquot corresponding to an moi of approximately 0.1 was added to a fresh flask of cells. The remainder of the culture supernatant was stored at -80°C for later characterization.

7.27 Quasispecies and DNA sequencing The chimeric JE vaccine is an RNA virus. Selective pressure can cause rapid changes in the nucleic acid sequences of RNA viruses. A mutant virus that invades FRhL cells more rapidly, for example, may gain a selective advantage by competing more effectively with the original vaccine virus and take over the culture. Therefore, mutant strains of the vaccine that grow better than the original vaccine may be selected by subculturing in vitro. One concern that addressed experimentally is whether such selective pressures might lead to mutant vaccine viruses with increased virulence.

In theory, molecular evolution should occur more rapidly for RNA viruses than DNA viruses because viral RNA polymerases have higher error rates than viral DNA polymerases. According to some measurements, RNA virus mutation rates approach one mutation per replication event. This is why an RNA virus can be thought of as a family of very closely related sequences (or "quasispecies"), instead of a single unchanging sequence (a "classical species").

Two different approaches can be taken to determine the sequence of an RNA virus:

1) purify viral genomic RNA from the culture supernatant, reverse- transcribe the RNA into cDNA and sequence this cDNA. This is the approach we have taken. It yields an averaged, or consensus sequence, such that only mutations which represent a large proportion (roughly, >20%) of the viruses in the culture can be detected.

2) Alternatively, cDNA can be cloned and individual clones sequenced. This approach would reveal the quasispecies nature of the vaccine by identifying individual mutations (deviations from the consensus sequence) in some proportion of the clones. 7.22 Biological Characterization of Serially Passaged RMS

As stated above, we demonstrated experimentally that the selective pressures exerted by serial passaging of the RMS does not lead to mutant vaccine viruses with increased virulence. Here, three biological properties of Passages 10 and 18 (P10 and PI 8) were examined. First, neurovirulence was tested by inoculating mice i.e. with graded doses of PI as well as P10 and PI 8. Second, immunogenicity was compared by inoculating mice s.c. with graded doses of the RMS, P10 and PI 8. Blood was drawn from these mice 30 days post inoculation and serum neutralizing titers were determined and compared. Finally, the growth kinetics of the RMS and of P10 and PI 8 were compared by inoculating FRhL cells at moi's of 0.1 and 0.01 and collecting samples of culture supernatant daily. The titers in each flask were plotted as a function of time and compared.

1.23 Stability of prM and E genes

The M and E genes of P10 and PI 8 were sequenced completely from base 642 to base 2454. Both sequences were identical and carried only one mutation (A— >G) resulting an amino acid substitution from H to R at position 394 on the E protein. This means that selective pressures did not lead to the loss of any of the attenuating mutations of the E gene. Codon H394 (CAC) encodes a Histidine in the RMS but we have found that the second base of this codon is mutated to a G in a significant proportion of the viruses, leading to the expression of Arginine. It is important to emphasize that a mixture of A and G are observed at this position in the sequence data. The ratio of A to G (A/G) was also determined for PI, P4, and P8. Interestingly, the ratio decreases steadily from PI to PI 0, but at PI 8 it is back to the value seen at P8. One possible explanation for this observation is that a mutant bearing the H394R mutation gradually became as abundant as the original virus but was then out-competed by a new mutant bearing other mutations not present in the M or E genes and therefore, only detected as a rebound in the A/G ratio. We are reproducing these results by doing a second passaging experiment under identical conditions. It must also be noted that duplicate samples of viral genomic RNA were isolated, reverse-transcribed, amplified, and sequenced in parallel for each passage examined. Reported results were seen in both duplicate samples, arguing against any RT-PCR artifacts obscuring the data.

These observations show that minor genetic changes (one nucleotide substitution in the entire envelope E and M genes) have occurred in the JE sequences of the chimeric vaccine upon passaging, but that selective pressures did not lead to the loss of any of the attenuating mutations of the E gene.

7.24 Neurovirulence phenotype of passages 10 and 18

Groups of five female ICR mice, 3 to 4 weeks-old, received 30 μl i.e. of undiluted, PI, PI 0, or PI 8, as well as 30 μl of 10-fold dilutions. None of the mice injected with PI, P10, or P18 (doses > 7 log₁₀ pfu) showed any sign of illness over a five week period. As determined by back-titration, the doses administered (pfu) were measured as shown in Table 17. 7.25 Immunogenicity of passages 10 and 18

Groups of five female ICR mice were injected subcutaneously (s.c.) with 100 μl of undiluted virus stock of either the RMS or P10 or PI 8, as well as with doses of 10⁵ and 10⁴ pfu (see Table 18, results of back- titration).

7.2<5 Growth kinetics of passages 10 and 18

Monolayers (90% confluent) of FRhL cells were infected with an moi of 0.1 or 0.01 of RMS, P10, or P18. Time points were then taken daily for seven days and the titer of each time point was determined by plaque assay. Visual observation of cytopathic effects (CPE) on FRhL cells used in this growth curve experiment show that later passages of the RMS have different growth properties than the RMS itself. CPE is clearly greater for PI 8 and P10 than for the RMS at 4 days postinfection showing that these viruses might replicate much faster than the RMS.

Other observations also show that the growth properties of P10 and PI 8 differ from those of the RMS. The titers of PI, P10, and PI 8 are ~2xl0⁷, 2xl0⁸, and 3x10⁸, respectively. The relative yields of RT-PCR products suggest higher titers of P10 and PI 8 compared to PI. Although the PCR data are not necessarily quantitative, they are consistent with the observed titers.

These results raise the possibility that we have discovered a variant of the vaccine that is immunogenic, attenuated for mouse neurovirulence, and that grows to titers ten-fold higher than the original vaccine (RMS) in tissue culture. Such a mutant may have value for manufacturing. Finally, the sequences of the entire genomes of the RMS and pi 8 were determined and found to be identical, except for the E-H394 mutation (Table 25). There are 6 nucleotide (NT) differences (NT positions are shaded) between the published YF 17D sequences and RMS shown in bold letters. Changes in positions 5461, 5641, 8212, and 8581 are silent and do not result in amino acid substitution, whereas changes in positions 4025 (ns2a) and 7319 (ns4b) result in amino acid substitutions from V to M and from E to K, respectively. Amino acid Methionine (M) at position 4025 is unique for RMS and is not found in any other YF strains, including parent Asibi virus and other yellow fever 17D strains (e.g., 204, 213, and 17DD), whereas Lysine (K) at position 7319 is found in 17D204F, 17D213, and 17DD, but not in 17D204US or Asibi strain. Since the RMS is more attenuated than YF 17D with respect to neurovirulence, and thus has better biological attributes as a human vaccine, it is possible that the amino acid differences at positions 4025 and 7319 in the nonstructural genes of the yellow fever portion of the chimeric virus contribute to attenuation. Other workers have shown that the nonstructural genes of yellow fever virus play an important role in the attenuation of neurovirulence (Monath, "Yellow Fever," in Plotkin et al, (Eds.), Vaccines, 2^nd edition, W.B. Saunders, Philadelphia, 1998).

7.27 Experiment to Identify Possible Interference Between YF 17D and YF/JE SA 14- 14-2

It is well-established that yellow fever virus encodes antigenic determinants on the NSl protein that induce non-neutralizing, complement- fixing antibodies. Passive immunization of mice with monoclonal anti-NSl antibodies confers protection against challenge. Active immunization with purified or recombinant NSl protects mice and monkeys against lethal challenge. The mechanism of protection is presumed to involve antibody-mediated complement-dependent cytotoxicity. In addition to protective determinants on NSl, CTL epitopes on other nonstructural proteins, including NS3, NS2a, and possibly NS5 may be involved in protection. Thus, infection with the YF/JE chimeric virus may stimulate humoral or cellular anti-yellow fever immunity. It is possible, therefore, that use of the chimeric vaccine may interfere with subsequent immunization against YF 17D, or that prior immunization with YF 17D may interfere with seroconversion to YF/JE SA14-14-2. Against this hypothesis is a substantial body of data showing that reimmunization with YF 17D results in a boost in yellow fever N antibodies. Those data show that it should be possible to successfully immunize against JE in an individual with prior YF immunity and vice versa.

To investigate possible interference effects, the experiment shown in Table 19 was initiated. Mice are immunized with one vaccine and subsequently boosted with the heterologous vaccine. Mice are bled every 30 days and sera tested for neutralizing antibodies against heterologous and homologous viruses.

1.28 Seroconversion Rate and Antibody Titers After Primary Immunization

Three groups (n=8) of 3-4 weeks old female outbred ICR mice were immunized with a single dose (5.3 log₁₀ pfu) of ChimeriVax™ JE (YF/JE SA 14- 14-2), three groups (n=8) were immunized with two doses of JE-Vax® (0.5 ml of a 1 :5 dilution of reconstituted vaccine) and three groups (n=8) were immunized with a single dose of YF-Vax® (0.1 ml of a 1 :2 dilution of reconstituted vaccine, containing 4.4 log₁₀ pfu, previously determined to induce the highest immune response to YF virus). Six groups (n=4) of mice (similar age, 3-4 weeks old) were kept as controls for booster doses at 3, 6, and 12 months post primary immunization.

All mice were bled 4 and 8 weeks after primary immunization and their neutralizing antibody titers were measured against homologous viruses in a plaque assay. 21/24 (87.5%) of the animals immunized with a single dose of ChimeriVax™-JE developed anti-JE neutralizing antibodies 1 month after immunization; at 2 months, 18/24 (75%) were seropositive. Geometric mean increased somewhat between 1 and 2 months post inoculation. In contrast, only 25%-33% of the mice immunized with YF-Vax® seroconverted and antibody responses were low. These results show that YF 17D virus and chimeric viruses derived from YF 17D are restricted in their ability to replicate in the murine host; however, when the envelope of JE virus is incorporated in the chimeric virus, the ability to replicate in and immunize mice is apparently enhanced. Mice receiving two doses of JE-Vax® developed high neutralizing titers against parent Nakayama virus, and titers increased between 1 and 2 months post immunization.

7.29 Secondary Immunization of ChimeriVax™ JE and JE-Vax® Immunized Mice With YFVax®

Three months and six months after primary immunization with ChimeriVax- JE, mice were inoculated with YF-Vax® (1 :2 dilution of a human dose containing 4.4 log₁₀ pfu). Control mice not previously immunized and of identical age received ChimeriVax™ JE only or YF-Vax® (Groups 10-13). One month later, mice were tested for presence of YF-specific neutralizing antibodies.

At the 3 month time point, none of the control mice or mice previously immunized with ChimeriVax- JE or JE-Vax seroconverted to YF-Vax®, again confirming the poor immunogenicity of YF-Vax® at the dose used. However, all mice immunized with YF-Vax® 6 months after primary immunization with ChimeriVax- JE and 7/8 mice previously immunized with JE-Vax, seroconverted after immunization with YF-Vax® (Table 24). There was no difference in seroconversion rate or GMT in mice with and without prior immunization with either JE vaccine.

1.30 Secondary Immunization of YF- Vox® Immunized Mice with ChimeriVax™ JE

All mice previously immunized with YF-Vax® and reimmunized with ChimeriVax- JE 3 months later developed neutralizing antibodies to JE (group 7, Table 10). None of the controls seroconverted. Five of 6 mice (83%) previously immunized to YF-Vax® and. reimmunized with ChimeriVax- JE 6 months later seroconverted to JE (group 8, Table 10, as did all controls (group 13)), and the GMTs were similar across these groups.

There was no evidence for cross-protection between YF and JE viruses or limitation of antibody response to sequential vaccination with these viruses. Yellow fever 17D vaccine elicits a poor antibody response in the mouse; while this limited interpretation of the data somewhat, it provided a sensitive test of any restriction in replication and immunogenicity of YF 17D virus in mice previously immunized with ChimeriVax-JE. The fact that all mice immunized with ChimeriVax- JE responded 6 months later to immunization with YF-Vax® and that the GMT and range of neutralizing antibody titers were similar to controls suggests that the chimeric vaccine imposed no significant barrier to yellow fever immunization.

2.0 Construction ofcDNA Templates for Generation of Yellow Fever/Dengue (YF/DEN) Chimeric Viruses

Derivation of chimeric Yellow Fever/Dengue (YF/DEN) viruses is described as follows which, in principle, is carried out the same as construction of the YF/JE chimeras described above. Other flavivirus chimeras can be engineered with a similar strategy, using natural or engineered restriction sites and, for example, oligonucleotide primers as shown in Table 20.

2.7 Construction of YF/DEN Chimeric Virus

Although several molecular clones for dengue viruses have been developed, problems have commonly been encountered with stability of viral cDNA in plasmid systems, and with the efficiency of replication of the recovered virus. We chose to use a clone of DEN- 2 developed by Dr. Peter Wright, Dept. of Microbiology, Monash University, Clayton,

Australia, because this system is relatively efficient for regenerating virus and employs a two-plasmid system similar to our own methodology. (See Table 21 for a comparison of the sequences of Dengue-2 and YF/Den-2₂₁₈ viruses; YF/Den-2₂₁₈ contains the nucleotide and amino acid sequences of PUO-218. The NGC and PR- 159 strains, which are also listed in Table 21, are other wild strains of dengue that differ from PUO-218 and can be used in the chimeras of the invention.) The complete sequence of this DEN-2 clone is available and facilitated the construction of chimeric YF/DEN templates because only a few modifications of the YF clone were required. The relevant steps are outlined as follows.

Similar to the two plasmid system for YF5.2iv and YF/JE viruses, the YF/DEN system uses a unique restriction site within the DEN-2 envelope protein (E) as a breakpoint for propagating the structural region (prM-E) within the two plasmids, hereinafter referred to as YF5'3'IV/DEN (prM-E') and YFM5.2/DEN (E'-E) (see Fig. 20). The two restriction sites for in vitro ligation of the chimeric template are Aatll and Sphl. The recipient plasmid for the 3' portion of the DEN E protein sequence is

YFM5.2(Narl[+]Sphl[-]). This plasmid contains the N rl site at the E/ΝSI junction, which was used for insertion of the carboxyl terminus of the JE E protein. It was further modified by elimination of an extra Sphl site in the ΝS5 protein region by silent site-directed mutagenesis. This allowed insertion of DEN-2 sequence from the unique Sphl site to the Narl site by simple directional cloning. The appropriate fragment of DEN-2 cDNA was derived by PCR from the DEN-2 clone MON310 furnished by Dr. Wright. PCR primers included a 5' primer flanking the Sphl site and a 3' primer homologous to the DEN-2 nucleotides immediately upstream of the signalase site at the E/NSI junction and replacing the signalase site by substitutions that create a novel site, but also introduce a Nαrl site. The resulting 1,170 basepair PCR fragment was then introduced into YFM5.2(NαrI[+]_S/>/?I[-]).

The 5' portion of the YF/den-2 clone, including the C/prM junction was engineered by PCR. The C/prM junction was created by incorporating a Tfil restriction site at the junction using synthetic oligonucleotides. A 5' PCR fragment encompassing the flanking YF sequence 5' untranslated and capsid sequence and a 3' Tfil site, together with a 3' PCR fragment beginning with a Tfil site at the amino terminus of the dengue-2 prM protein and the flanking dengue-2 prM protein sequence, were ligated into the YF5'3'IV plasmid after intermediate construction in pBluescript. Screening with Tfil was used to confirm correct assembly of the chimeric junction in the final plasmid YF5'3TV/DEN(prM-E).

2.2 Construction of Chimeric YF/DEN Viruses Containing Portions of Two DEN Envelope Proteins

Since neutralization epitopes against DEN viruses are present on all three domains of the E protein, it is possible to construct novel chimeric virus vaccines that include sequences from two or more different DEN serotypes. In this embodiment of the invention, the C/prM junction and gene encoding the carboxyl terminal domain (Domain III) of one DEN serotype (e.g., DEN-2) and the N-terminal sequences encoding Domains I and II of another DEN serotype (e.g., DEN-1) are inserted in the YF 17D cDNA backbone. The junctions at C/prM and E/NS1 proteins are retained, as previously specified, to ensure the infectivity of the double-chimera. The resulting infectious virus progeny contains antigenic regions of two DEN serotypes and elicits neutralizing antibodies against both.

2.3 Transfection and Production of Progeny Virus Plasmid YF5'3 V/DEN(prME) and YFM5.2/DEN(E'-E) were cut with Sphl and Aatll restriction enzymes, appropriate YF and dengue fragments were isolated and ligated in vitro using T4 DNA ligase. After digestion with Xhol to allow run-off transcription, RNA was transcribed (using 50 ng of purified template) from the SP6 promoter and its integrity was verified by non-denaturing agarose gel electrophoresis. Vero cells were transfected with YF/Den-2 RNA using Lipofectin (Gibco/BRL), virus was recovered from the supematants, amplified twice in Vero cells, and titrated in a standard plaque assay on Vero cells. The virus titer was 2x10⁶ PFU/ml.

2.4 Nucleotide sequencing of YF/Den-2 Chimera Vero cells were infected with YF/DEN-2 (clone 5.75) at an MOI of

0.1. After 96 hours, cells were harvested with Trizol (Life Technologies, Inc.). Total RNA was primed with a YF-5' end NSl minus oligo, and reverse transcribed with Superscript II RT following a long-RT protocol (Life Technologies, Inc.). Amplification of cDNA was achieved with XL- PCR kit (Perkin Elmer). Several primers specific for dengue type 2 strain PUO-218 were used in individual sequencing reactions and standard protocols for cycle sequencing were performed. Sequence homology comparisons were against the PUO-218 strain prME sequence (GenBank accession number D00345). Sequencing showed that the YF/DEN-2 chimera prME sequence is identical to that of PUO-218 (Gruenberg et al, J. Gen. Virol 69:1391- 1398, 1988). In addition, a Narl site was introduced at the 3' end of E, resulting in amino acid change Q494G (this residue is located in the transmembrane domain and not compared in Table 21). In Table 21, amino acid differences in the prME region of YF/Den2 is compared with prototype New Guinea C (NGC) virus and the attenuated dengue-2 vaccine strain PR-159 SI (Hahn, et al, Virology 162:167-180, 1988). 2.5 Growth Kinetics in Cell Culture

The growth kinetics of the YF/Den-2 chimera were compared in Vero and FeRhL cells (Fig. 16). Cells were grown to confluency in tissue culture flask (T-75). FeRhL cells were grown in MEM containing Earle's salt, L-Glu, non-essential amino acids, 10% FBS and buffered with sodium bicarbonate, and Vero cells were grown in MEM- Alpha, L-Glu, 10% FBS (both media purchased from Gibco/BRL). Cells were inoculated with YF/Den2 at 0.1 MOI. After 1 hour of incubation at 37°C, medium containing 3% FBS was added, and flasks were returned to a C0₂ incubator. Every 24 hours, aliquots of 0.5 ml were removed, FBS was added to a final concentration of 20%, and frozen for determination of titers in a plaque assay. Forty eight hours post infection CPE was observed in FeRhL cells and reached 100% by day 3. In Vero cells, CPE was less dramatic and did not reached 100% by the completion of the experiment (day 5). As shown, the YF/Den2 reached its maximum titer (7.4 log₁₀ pfu/ml) by day 3 and lost about one log (6.4 log₁₀ pfu/ml) upon further incubation at 37 °C, apparently due to death of host cells and virus degradation at this temperature. The maximum virus titer in Vero cells was achieved by day 2 (7.2 log₁₀ pfu/ml) and only half log virus (6.8 log₁₀ pfu/ml) was lost on the following 3 days. This higher rate of viable viruses in Vero cells may be explained by incomplete CPE observed in these cells. In sum, the chimera grows well in approved cell substrate for human use. 2.6 Neurovirulence Phenotype in Suckling Mice

Although wild-type unpassaged dengue viruses replicate in brains of suckling mice and hamsters inoculated by the intracerebral route (Brandt et al, J. Virol 6:500-506, 1970), they usually induce subclinical infection and death occur only in rare cases. However, neurovirulence for mice can be achieved by extensive passage in mouse brain. Such neuroadapted viruses can be attenuated for humans. For example, the New Guinea C (NGC), the prototype dengue 2 virus isolated in 1944 and introduced into the Americas in 1981, is not neurovirulent for suckling mice; however after sequential passage in mouse brain it became neurovirulent for mice, but was attenuated for humans (Sabin, Am. J. Trop. Med. Hyg., 1 :30-50, 1952; Sabin et al, Science 101 :640-642, 1945; Wisseman et al, Am. J. Trop. Med. Hyg.12:620-623, 1963). The PUO- 218 strain is a wild type dengue 2 virus isolated in 1980 epidemic in Bangkok. It is closely related to the NGC strain by nucleotide sequencing (Gruenberg et al, J. Gen. Virol 69:1391-1398, 1988). When the prME genes of the PUO-218 strain were inserted into the neuroadapted NGC backbone, the chimeric virus was attenuated for 3 -days old mice inoculated by the I.C. route (Peter Wright, Xth International Congress of Virology, Jerusalem, Israel, 1996). The PU0218 virus differs from NGC in one amino acid in prM (residue 55 is F in NGC and is L in PU0218) and 6 amino acids in the E protein (71 D->E, 126K->E, 141I->V, 164 I- >V, 402I->F, and 484 V->I) (see Table 21). All amino acid differences (except residue E-126) are also present in PR SI strain (attenuated vaccine strain), indicating that they may not be involved in attenuation. Only residue 126 on the E protein is different between these viruses. This residue was shown to be responsible for the neurovirulent phenotype of the mouse adapted NGC (Bray et al, J. Virology 72: 1647-1651, 1998). Although mouse neurovirulence does not predict virulence/attenuation of dengue viruses for humans, it is important to determine the neurovirulence of a YF/Den-2 chimeric virus. YF 17D retains a degree of neurotropism for mice, and causes (generally subclinical) encephalitis in monkeys after IC inoculation. For vaccine development of a den/YF chimera it will be necessary to show that the construct does not exceed YF 17D in neuroinvasiveness and neurovirulence. Ultimately safety studies in monkeys will be required. In initial studies, we determined if insertion of the prME of the PU0218 into YF 17D vaccine strain will affect its neurovirulence for suckling mice (Table 24). Groups of 3, 5, 7, and 9 days old suckling mice were inoculated by the I.C. route with 10,000 pfu of YF/Den-2 or YF/JE SA14-14-2 chimera and observed for paralysis or death for 21 days. For controls similar age groups were inoculated either sham with medium (I.C. or I.P.) or with 1,000 pfu of unpassaged commercial YF vaccine (YF-Vax) by the I.P. route (it is not necessary to inoculate suckling mice with YF-Vax by the I.C. route because we have previously shown that this vaccine is virulent for 4-weeks old mice by this route).

As shown in Fig. 22, all suckling mice (3 to 7 days old) inoculated by the I.C. route with the YF/Den2 chimera died between 11 and 14 days post inoculation, whereas 8 out of 10 suckling mice (9 days old) survived. Similarly, all suckling mice (3-5 days old) inoculated with YF-Vax by the I.P. route, with a dose which was 10-fold lower than the YF/Den2 chimera, died between 11 to 13 days post inoculation (Fig. 23). All nine day old, as well as 8 out of 9 seven day old, mice inoculated with the YF- Vax survived. Similar results to the YF/Den2 chimera obtained with suckling mice inoculated with the YF/JE SA14-14-2 chimera. As is mentioned above, when prME genes of the PU0218 strain were inserted into the NGC backbone the chimeric virus was not neurovirulent for 3 days old suckling mice inoculated by the I.C. route. In contrast, when these genes were inserted into the 17D backbone, the resulting YF/Den2 chimera demonstrated a neurovirulence phenotype (for suckling mice) similar to the YF/ JE SA14-14-2. This experiment also demonstrated that the replacement of the prME genes of the YF 17 D with prME genes of the Dengue 2 PU0218 resulted in a chimeric virus which was less neurovirulent than the 17D parent strain. Unlike most flaviviruses, there is no correlation between neurovirulence of dengue viruses in mice and humans. Currently the most suitable animal models for dengue infection are Old World monkeys, New World monkeys, and apes that develop subclinical infection and viremia. There is, however, no animal model for the most severe illness (DHF) in humans, which occurs when individuals become infected with a heterologous serotype due to antibody dependent enhancement of infection. Today it is generally accepted that a tetravalent vaccine is required to induce protective immunity in human beings against all four serotypes to avoid sensitizing vaccinee to more severe illness DHF. For the last fifty years, many approaches have been undertaken to produce effective dengue vaccines and although dengue viruses have been satisfactory attenuated (e.g., PR-159/S-1 for Dengue 2) in many cases in vitro or in vivo correlation of attenuation were not reproducible in humans. A current strategy is to test selected live virus vaccine candidates stepwise in small numbers of human volunteers. Many laboratories around the world are exploring various strategies to produce suitable vaccine candidates. These range from subunit vaccines including prME (protein vaccine or DNA vaccine) of dengue viruses to live attenuated whole viruses (produced by tissue culture passage or recombinant DNA technology). Although some of these candidates have shown promise in preclinical and human volunteers, development of a successful dengue vaccine remained to implemented. Evaluating the immunogenicity and protective efficacy of the

YF/Den2 chimera in monkeys should shed light on selection of appropriate prME genes (form wild type or attenuated strain) for construction of all 4 serotypes of chimeric dengue viruses.

2.7 Stability of prME genes of ChimeriVax™ -D2 virus in vitro

The ChimeriVax™-D2 virus at passage 2 post transfection was used to inoculate a 25 cm² flask of Vero cells. Total RNA was isolated and the complete nucleotide sequence of the ChimeriVax™-D2 was determined (P3) and compared to the published sequence of the YF 17D virus (Rice et al, Science 229:726-733, 1985). There was one nucleotide difference: at position 6898 there was an A in the chimera (P3), which was a C in the 17D nucleotide sequence. No difference in the prME region was found when the sequence of ChimeriVax™-D2 was compared to its parent dengue 2 virus (PU0218 strain). Also, no mutations were found in the prME genes of the chimera upon 18 passages in VeroPM cells. Within the YF genes, however, there was one silent mutation in position 6910 (C to A), and at position 3524 the PI 8 virus appeared to be heterozygous (both parent nucleotides, G and mutant A, were present). This would translate into a mixture of E and K amino acids at position 354 of the NSl protein.

Similar to the passage 3 virus, the passage 18 virus was not neurovirulent for 4 week old outbred mice inoculated by the IC route (5 log₁₀ pfu was the highest dose tested). Passage 3, passage 5, passage 10, and passage 18 of ChimeriVax™-D2 were inoculated into mice by SC and IC routes, and antibody responses were compared. There were no significant differences in production of anti-dengue 2 neutralizing antibodies across 18 passages (Table 26).

2.8 Viremia, immunogenicity, and protective efficacy of ChimeriVax™ -D 2 in YF immune monkeys

Because ChimeriVax™- viruses contain core and NS genes of the YF 17D virus, it is important to determine if preimmunity to the 17D vaccine interferes with vaccination with ChimeriVax™-D2 virus. As is discussed above, in the case of ChimeriVax™- JE, there was no significant interference between YF17D immunity and ChimeriVax™- JE virus, measured by production of neutralizing antibodies in mice. Since YF 17D vaccine and both ChimeriVax™- JE and

ChimeriVax™-D2 were not highly immunogenic in mice inoculated by the SC route (see also table 26), non-human primates were used, which are more susceptible/relevant for evaluation of flavivirus vaccines for humans. Sixteen rhesus monkeys (some of which were previously immunized with YF 17D vaccine) received ChimeriVax™-D2, YF 17D vaccine, or a wild type dengue 2 virus (strain SI 6803). As is shown in Table 27, all YF immune monkeys seroconverted to the YF/dengue 2 or wild type dengue 2 virus, demonstrating a lack of vector immunity. These monkeys were also protected from viremia after challenge with wild type dengue 2 virus. In contrast, YF 17D immunized monkeys, as well as non- immunized animals, became viremic after challenge with wild type dengue 2 virus. Wild type dengue 2 viruses produce a high level of viremia (3-5 logs) in rhesus monkeys, which lasts between 3-6 days. Attenuation of dengue 2 viruses can therefore be estimated by comparing the level and duration of viremia with reference wild-type strains. These experiments clearly showed that core and non- structural proteins of YF 17D virus present in ChimeriVax™-D2 do not interfere with ChimeriVax™-D2 immunization.

2.9 Dose response effectiveness of ChimeriVax™ -D 2 in monkeys.

The goals of this experiment were to (i) determine the viremia profile of the vaccine candidate, using YF 17D and wild type dengue 2 virus controls, (ii) compare neutralizing antibody responses to the vaccine candidate and wildtype virus, and (iii) determine minimum dose required for protection against challenge with wild type dengue-2 virus. It was anticipated that these experiments would define the viremia profile of the ChimeriVax™-D2 virus in non-YF immune monkeys, and would determine whether immunization with a single dose results in protection of animals against challenge with a wild type dengue 2 virus. Protection in these experiments is defined as reduction of viremia in test monkeys compared to control viruses.

As is shown in table 28, all monkeys became viremic, and the duration of viremia was dose-dependent. The peak level of viremia for ChimeriVax™-D2 was between 1.3 to 1.6 log₁₀ pfu, which was significantly lower than that of the wild type dengue virus (3.6 log₁₀ pfu).

All monkeys developed anti-dengue 2 neutralizing antibodies by day 15. Lower dose of the vaccine resulted in lower GMTs, however, by day 30 post-immunization, all monkeys developed high titers of neutralizing antibodies, independent of the dose they received. Upon challenge, no viremia was detected in any immunized monkeys, demonstrating that even at its lowest dose (2 log₁₀ pfu), ChimeriVax™ -D2 had protected these animals from dengue infection (Table 29).

2.70 Growth characteristics of ChimeriVax™ -JE virus in Culex tritaeniorhynchus, Aedes albopictus, and Aedes aegypti mosquitoes

As is described above, ChimeriVax™- JE, which consists of a YF 17D virus backbone containing the prM and E genes from the JE vaccine strain SA14-14-2, exhibited restricted replication in non-human primates, producing only a low level viremia following peripheral inoculation. Although this reduces the likelihood that hematophagous insects could become infected by feeding on a vaccinated host, it is prudent to investigate the replication kinetics of the vaccine virus in mosquito species that are known to vector the viruses from which the chimera is derived. In this study, ChimeriVax™-JE virus was compared to its parent viruses (YF 17D and JE S A 14- 14-2), as well as to wild type JE S A 14 virus, for its ability to replicate in Culex tritaeniorhyncus, Aedes albopictus, and Aedes aegypti mosquitoes. Individual mosquitoes were exposed to the viruses by intrathoracic (IT) virus inoculation or by oral ingestion of a virus-laden blood meal.

2.77 Intrathoracic inoculation

Mosquitoes were inoculated 7-10 days post-emergence with 0.34 ml of approximately 6.0 log₁₀ pfu/ml virus suspension (-5.5 log₁₀ pfu/mosquito). This route of inoculation was chosen to avoid variables, such as threshold titer, that might limit midgut infection and subsequent dissemination of the viruses. Three mosquitoes per virus were collected either at 24 hour intervals for 5-10 days or at 72 hour intervals for 18 days. Individual mosquitoes were triturated in 1 ml of Ml 99 media (Gibco BRL, Grand Island, New York) supplemented with 5% fetal calf serum, clarified by brief centrifugation, and then titrated in Vero cells to monitor virus replication. Both JE SA14 and JE SA14-14-2 viruses replicated in Cx. tritaeniorhynchus following IT inoculation, reaching titers at day 14 of 6.7 and 6.0 log_]0 pfu/mosquito, respectively (Figure 24A). Additionally, IFA conducted on head squashes from JE SA14 and JE SA14-14-2-inoculated Cx. tritaeniorhynchus mosquitoes was positive for detection of JE virus antigen. In contrast, YF 17D and ChimeriVax™- JE did not replicate in Cx. tritaeniorhynchus mosquitoes. Virus titers declined rapidly following inoculation, and no virus was detectable by plaque titration assay in YF 17D or ChimeriVax™- JE-inoculated mosquitoes by days 1 and 2, respectively (Figure 24 A). IFA analysis of head squashes from Cx. tritaeniorhynchus mosquitoes inoculated with ChimeriVax™- JE or YF 17D was negative for JE or YF virus antigens, supporting our observation that neither the chimera nor YF 17D replicate in this mosquito species.

ChimeriVax™-JE did replicate in IT-inoculated Ae. albopictus mosquitoes, reaching a titer of 5.2 log₁₀ pfii/mosquito at day 18 (Figure 24B) and IFA results were weakly positive for both JE virus and YF virus antigens. The JE SA14 and JE SA14-14-2 viruses also replicated i Ae. albopictus mosquitoes, reaching maximum titers of 6.3 and 6.0 log₁₀ pfu/mosquito, respectively. YF 17D virus did not replicate to high titers in Ae. albopictus mosquitoes, however, a low level of detectable virus was maintained (3.8 log₁₀ pfu/mosquito at day 18) (Figure 24B) and

IFA-stained head squashes were weakly positive for YF virus antigen. ChimeriVax™- JE and YF 17D inoculated IT into Ae. aegypti mosquitoes replicated at low levels over the 18 day incubation period (Figure 24C). Peak titers of 3.6 and 4.4 log₁₀ pfu/mosquito, respectively, were reached on day 15. IFA staining of head tissues from ChimeriVax™- JE and YF 17D IT-inoculated mosquitoes was weakly positive. Both JE SA14 and JE SA14-14-2 viruses proliferated in Ae. aegypti mosquitoes, reaching peak titers of 6.3 and 6.1 log₁₀ pfu/mosquito on days 9 and 18, respectively.

2.72 Oral infection

Seven to ten day old Ae. albopictus and Ae. aegypti mosquitoes were orally exposed to an artificial virus-containing blood meal that was prepared from equal parts of washed calf red blood cells (Colorado Serum Company, Denver, Colorado) and freshly harvested virus. The blood/virus mixture was heated to 37 °C immediately prior to feeding. Mosquitoes were starved for 48-72 hours prior to feeding on virus/blood soaked cotton pledgets. Cx. tritaeniorhynchus mosquitoes are reluctant to feed from blood-soaked pledgets, and were therefore fed using a membrane feeder. Mosquitoes were allowed to feed for 15-30 minutes, after which fully engorged mosquitoes were collected. Three mosquitoes per virus were harvested at 48-72 hour intervals over a 15-18 day period, or, in a second experiment, all mosquitoes were harvested at 22 days after feeding.

Figure 25 A illustrates growth of the viruses in orally exposed Cx. tritaeniorhynchus mosquitoes. Individuals that were fed a blood meal containing 6.9 log₁₀ pfu/ml ChimeriVax™- JE virus did not become infected. Similar results were observed in mosquitoes that had ingested a blood meal containing YF 17D virus. In contrast, high virus titers were detected in Cx. tritaeniorhynchus mosquitoes that had ingested JE SA14 or JE SA14-14-2 viruses. Figures 25B and 25C illustrate growth of the viruses in orally exposed Ae. albopictus and Ae. aegypti mosquitoes, respectively. Only JE SA14 and JE SA14-14-2 viruses successfully infected and replicated in these species. For example, in Ae. aegypti mosquitoes on day 15, the titers of JE SA14 and JE SA14-14-2 viruses were 5.4 and 5.5 log₁₀ pfu, respectively. In contrast, mosquitoes that had ingested 4.7 log₁₀ pfu/mosquito of YF17D virus or 4.5 log₁₀ pfu/mosquito of ChimeriVax™- JE virus failed to become infected.

In a separate experiment, Ae. aegypti and Ae. albopictus mosquitoes were orally exposed to JE SA 14- 14-2, YF 17D, and ChimeriVax™- JE viruses and processed after 22 days extrinsic incubation to permit growth to maximum virus titers. The results of this experiment are summarized in Table 30. Only JE SA14-14-2 virus was detectable in mosquitoes. Because ChimeriVax™- JE did not grow in any of the mosquito species tested, transmission studies were not performed.

Viruses recovered from Ae. Albopictus after IT or oral inoculation, or from Ae. Aegypti after IT inoculation, were identical to their parent ChimeriVax™- JE virus (Vero2FrhLl) in the prME region.

2.73 Amplification and sequencing of the "late replicating" ChimeriVax™ -JE viruses isolated from mosquitoes

Ae. albopictus mosquitoes inoculated with ChimeriVax™- JE by IT or oral routes and Ae. aegypti inoculated with ChimeriVax™-JE by IT route, were harvested on day 15 post-inoculation. After triturating in 1 ml of Ml 99 (supplemented with 5% fetal calf serum), samples were clarified by centrifugation, filtered through a 0.2 micron filter, and used to inoculate a T-25 cm² flask of VeroPM cells, passage 144 (0.5 ml/flask). After 1 hour, virus adsorption at 37°C, 5 ml MEM-containing 5% FBS was added and flasks returned to the 37 °C C0₂ incubator. Viruses were harvested from supematants 4 days later (at 2+ CPE) and kept frozen at -70 °C. Total RNA was extracted from infected monolayers by the use of the Trizol® reagent (Gibco/BRL), reverse-transcribed, amplified by XL PCR (Perkin-Elmer), and the prME region was sequenced. Viruses recovered from Ae. Albopictus after IT or oral inoculation, or from Ae. Aegypti after IT inoculation, were identical to their parent ChimeriVax™- JE vims (Vero2FrhLl) in the prME region.

2.14 Growth characteristics of ChimeriVax™ -D2 virus in Aedes albopictus and Aedes aegypti mosquitoes

Similar experiments were carried out in Ae. albopictus and Ae. aegypti mosquitoes with ChimeriVax™-D2 vims. For controls, the YF17D vaccine and a dengue 2 wild type vims were used. Dengue 2 wild type vims grew to more than 5 log₁₀ pfu/ml in both mosquito species inoculated by IT or oral routes. The growth of YF17D vaccine was lower than the wild type dengue 2 vims, and did not exceed 4 log₁₀ pfu/ml. Interestingly, the ChimeriVax™-dengue 2 vims revealed the most restricted growth in both mosquito species inoculated by either route (its titer did not exceed 3 log₁₀ pfu/ml) (Figure 26).

2.75 Summary ofJE andDen2 Experiments in Mosquitos

In summary, ChimeriVax™- JE virus did not replicate following ingestion by any of the three mosquito species. Additionally, replication was not detected after IT inoculation of ChimeriVax™- JE in the primary JE vims vector, Cx. tritaeniorhynchus. ChimeriVax™- JE exhibited moderate growth following IT inoculation into Ae. aegypti and Ae. albopictus mosquitoes, reaching titers of 3.6-5.0 log₁₀ pfu/mosquito. There was no change in the vims genotype associated with replication in mosquitoes. Similar results were observed in mosquitoes of all three species that were IT inoculated or had orally ingested the YF 17D vaccine virus. In contrast, all mosquitoes either IT inoculated with, or orally fed, wild type and vaccine JE vimses became infected, reaching maximum titers of 5.4-7.3 log₁₀ pfu/mosquito. The growth of ChimeriVax™-D2 in both Ae. albopictus and Ae. aegypti mosquitoes inoculated by IT or oral routes was also significantly lower than its parent wild type dengue 2 and YF17D vaccine vimses.

These results showed that ChimeriVax™- JE and ChimeriVax™-D2 vimses are restricted in their abilities to infect and replicate in these mosquito vectors. The low viremia caused by the viruses in primates and poor infectivity for mosquitoes are safeguards against secondary spread of the vaccine vims.

3.0 Construction of ChimeriVax™ YF/DEN-1

A yellow fever/dengue 1 (YF/DEN-1) chimeric vims was constructed using a novel technology, which differs from the approaches used to construct Yellow fever/Japanese encephalitis (YF/JE) chimeric vimses as described by Chambers et al (J. Virol. 73:3095-4101, 1999; see above), and the construction of YF/DEN-4 chimera (see below). We used the same two plasmid system used to create YF/DEN-4. These plasmids first encoded the yellow fever (YF) genome as created by Rice et al. (New Biol. 1:285-296, 1989). Later, the structural membrane precursor and envelope protein genes, i.e., the prME region, of the YF genome plasmids was replaced with those of the JE SA14-14-2 sequence, and the resulting plasmids were used to produce RNA in vitro, which was then transfected into cells to produce live YF/JE chimeric vims. Although the two-plasmid system was suitable for the constmction of JE, DEN-2, and DEN-4 chimeras in a YF backbone, the marked instability of one of the plasmids created with DEN-1 sequences was such that we opted for a PCR alternative to replace it.

Here we describe in detail the procedures for constmction of the YF/DEN-1 chimera (Figs. 27 and 28). Dengue 1 (strain PU0359 isolated in 1980 in Thailand) was passed once in C6/36 and total RNA was isolated. The dengue 1 prME region was first amplified and sequenced using primers derived from consensus sequences (Genbank). The sequence data created was applied to primer design and was used, with the cDNA produced earlier, as a PCR starting point for assembly of chimeric YF/DEN-1 vims. A dengue 1 PCR product encoding prM and the 5' end of E was then used as a template, along with template encoding the capsid (C) of yellow fever derived from plasmid pYF5'3TV/JE SA 14-14-2, in an overlap extension PCR to result in a single fusion product, which was then cloned into a vector fragment of pYFM5'3TV in which JE sequences were deleted. In contrast to the constmction of the YF/DEN-4 (see below) and YF/JE SA 14-14-2 (see above) plasmid systems, the 3' end of the DEN-1 envelope was fused to the YF non- structural genes normally present in the pYFM5.2/JE SA 14-14-2 plasmid using an overlap extension PCR similar to that used to construct the fusion of YF capsid to the DEN-1 prM and envelope gene 5' end. Next, the pYDl-5'3' plasmid was transformed into E. coli strain MCI 061 (RecA-) for amplification, followed by midi-scale plasmid purification, while the DEN-lEnv/YF5.2 (Fragment H) was gel purified. In vitro ligation of the plasmid to the PCR product resulted in full-length vims cDNA template for RNA transcription. All steps involving cDNA fragments, plasmids, and PCR products were carried out in a BL-2 lab designated for recombinant DNA work. Steps involving manipulations of infectious RNA and virus were carried out in a limited access BL-2+ vims lab.

3.7 Amplification of Dengue 1 sequence

Dengue 1 cDNA was synthesized from RNA using the Superscript II™ method. All primers for this experiment were synthesized by Life Technologies and are listed in Table 31. Upon arrival as lyophilized material, they were dissolved to 250 μM stock solutions using RODI-water. From this, 25 μM working solutions were made. The fragment encoding the SP6 promoter and the yellow fever capsid (Fragment A) was amplified using XL-PCR Reaction Kit TM (Perkin-Elmer Part#N808-0192), with 0.5 μl (250 ng) of pYF5'3'IV plus 3.5 μl RODI-water as template and primers 1 and 2 (see Table 31). The fragment encoding dengue 1 prM and 5' end of E (Fragment B) was amplified using the XL-PCR Reaction Kit™ (Perkin-Elmer Part#N808-0192) and primers 3 and 4. The fragment encoding the 3' end of the Dengue 1 envelope gene (Fragment F) was amplified using the same protocol, but with primers 5 and 7. The fragment encompassing the YF portion of pYFM5.2 (Fragment G) was amplified using the same protocol, but with primers 8 and 9 and 1 μl of pYFM5.2/2 with 39 μl water. The PCR for fragments F and G required an annealing temperature of 50 °C and an extension time of 6.5 minutes. The PCR reaction was performed using the following master mixes for each reaction. Upper Mix (UM)

Lower Mix (LM)

cDNA Mix

The LM was added to a Perkin-Elmer thin-walled 0.2 ml tube. Next, Ampliwax 100 (Perkin-Elmer) was added to the tube, which was then placed in a Perkin-Elmer 2400 Thermal Cycler and heated to 80 °C for 5 minutes, and then cooled to 4°C. The cDNA and UM were then added to the top of the wax layer. The tube was then cycled in a Perkin-Elmer 2400 as follows: 94°C, 1 minute; repeat 30x (94°C, 15 seconds; 53 °C, 15 seconds; 68°C, 3 minutes), 72°C, 4 minutes; 4°C, hold. The expected sizes for the fragments are as follows.

Forty μl of each fragment was then separated on a 1% Agarose/TAE gel and purified using the QIAquick Gel Extraction Kit (Qiagen cat#28704). Next, the concentrations of the purified fragments were determined by UV absoφtion using 1 :40 dilutions in RODI-water.

3.2 Recombinant PCR

To create a fusion between the yellow fever capsid and DEN-1 prM, a recombinant PCR technique known as overlap-extension PCR was used to create Fragment E. The same basic UM and LM were used, and primers 1 and 4 replaced earlier primers. The same approach was used to create a fusion between fragment F and G, resulting in fragment H. For this, primers 5 and 9 were used. The cDNA mixes were as follows:

Fragment E Fragment B Fragment A control control

Fragment H Fragment F Fragment G control control

The same protocol that was used for creation of fragments A, B, F, and G was used, except that only ^lA the cDNA, UM, and LM were used for the control reactions. The tubes were cycled in a Perkin-Elmer 2400 as follows: 94°C, 1 minute; repeat 30x (94°C, 15 seconds; 55 °C, 15 seconds; 68 °C, 2 minutes; 72 °C, 7 minutes; 4°C, hold, for Fragment C and its controls. For Fragment H and its controls, the annealing temperature was 50 °C and the extension time was 6.5 minutes. The expected sizes were as follows:

Forty μl of Fragment E and 50 μl of Fragment H was then separated on a 1% Agarose/TAE gel and purified using the QIAquick Gel Extraction Kit (Qiagen cat#28704). Next, the concentration of the purified fragment was determined by UV absoφtion using 1 :40 dilutions in RODI-water.

3.3 Cloning of Fragment E into the Yellow Fever Vector The capsid-prME fusion was cloned into the yellow fever plasmid needed, and after digestion of the purified Fragment E, as well as pYF5'3TV, with the appropriate enzymes. The digested plasmid resulted in two bands. Lower bands seen contain a fragment of Japanese encephalitis vims equivalent to Fragment E. All restriction enzymes, buffers, and lOOx BSA were from New England Biolabs. All the digestions were incubated in a Perkin-Elmer 480 cycler set to hold at 37 °C overnight. Fragment E Digest

pYFMIV5'3' Digest

3.4 Vector Dephosphorylation

Calf Intestinal Phosphatase (CIP) from New England Biolabs (cat#290S) was diluted 1:10 in lx CIP Buffer. One μl of this dilution was then added to the pYFMIV5'3' digest, and incubated for 1 hour at 37 °C. Then, 0.8 μl 125 mM EDTA was added to the tube, which was placed at 75 °C for 10 minutes to deactivate CIP.

3.5 Gel Excision

The digested fragment E and the digested plasmid were separated on a 1.0% Agarose/TAE gel, and were purified using the QIAquick Gel Extraction Kit (Qiagen cat#28704).

3.6 Ligations

The digested Fragment E and pYF5'3TV were ligated using T4 DNA Ligase (New England Biolabs cat#202S) to create pYDl-5"3\ All ligation reactions were incubated in a Perkin-Elmer 480 cycler set to hold at 16°C overnight. PYDl-5'3' Ligation

pYFM-5*3' Control Ligation

3.7 Transformations Ligation reactions were individually transformed into E. coli strain

MCI 061 (recA-). Briefly, an aliquot of MCI 061 was removed from storage at -80 °C and allowed to thaw on ice for one to two minutes. 0.9 ml of cold 0.1 M CaCl₂ was added to the cells. One hundred μl of cells was aliquoted into three 12 ml culture tubes on ice. Ten μl of each ligation reaction was added to each culture tube, leaving the third tube as a no DNA control. Culture tubes were left on ice for 30 minutes. The tubes were heat shocked in a water bath at 42 °C for 45 seconds, and then were put back on ice for 2 minutes. 0.9 ml SOC medium was added to each culture tube and incubated at 225 pm in a shaking incubator at 37° C for 1 hour. Each transformation mix was aliquoted into 1.5 ml microcentrifuge tubes. One hundred μl of each mix was spread onto LB/Agar-Amp (100 μg/ml) plates and labeled as "neat." Each tube was spun at 14,000 φm in a microcentrifuge for 2-3 seconds to pellet the cells. The supernatant was poured into the waste container and the pellet resuspended in the residual broth by pipetting up and down. This material was plated (approximately 100 μl) onto LB/Agar-Amp (100 μg/ml) plates and labeled as lOx. All plates were inverted in a 37° C incubator overnight. 3.8 Transformant Screening

Resulting bacterial colonies were patch-plated onto fresh LB/Agar-Amp (100 μg/ml) and placed inverted in a 37 °C incubator overnight. The following day, 50 μl of RODI-water was aliquoted into 0.5 ml tubes. Using a sterile plastic pick, a small amount of each patch was scraped into one of the 0.5 ml tubes containing water. These were then placed at 95 °C for 10 minutes, and spun at 14,000 φm for 10 minutes in a microcentrifuge. To identify the proper insert in pYDl-5'3', colonies were screened by PCR using Taq Polymerase (Promega) and primers 4 and 10. The PCR reaction was performed using the following master mix.

Forty eight μl of each master mix was added to a 0.5 ml tube along with 5 μl of the template prepared from the colony patches. Additionally, a tube using RODI-water as a template was also made as a negative control. These tubes were then placed in the Ericomp Delta Cycler and cycled as follows: 96°C, 4 minutes; 30x (94°C, 30 seconds; 50°C, 1 minute; 72 °C, 1 minute 25 seconds), 72 °C, 4 minutes; 4°C, Hold. The PCR products were then run on 1.5% Agarose/TAE gels to check for positive colonies. 3.9 Glycerol Stocks

One hundred twenty ml of LB- Amp (100 μg/ml) was then inoculated from a patch pYDl-5'372 and shaken at 225 ipm overnight at 37°C. Two x 1 ml of this culture was then spun at 14 Kφm for 2-3 seconds to pellet the cells. These were resuspended in LB-Glycerol (30%) and frozen at -80 °C.

3. 0 MIDI Plasmid Preparation

Qiagen Midi-Prep was performed on the remaining culture using the following modified protocol.

1. Spin 150 ml of each culture at 7 Kφm in GSA rotor for 10 minutes to pellet.

2. Decant Supernatant

3. Resuspend pellet in 4 ml PI buffer; Transfer to 50 ml Falcon tube. 4. Rinse centrifuge bottle with 1 ml PI buffer and transfer to the Falcon tube.

5. Add 5 ml P2 buffer; invert gently; incubate 5 minutes at room temperature or until lysed (no more than 12 minutes).

6. Add 5 ml P3 buffer; mix as above; incubate 10 minutes on ice. 7. Transfer supernatant to Qiagen Syringe Filter; Let sit for 10 minutes.

8. Equilibrate Q-100 tip with 4 ml QBT.

9. Gently push plunger to filter supernatant onto Q-100 tip.

10. Allow to drain by gravity.

11. Wash with 10 ml 2x QC. 12. Elute into 50 ml polypropylene centrifuge tube with 5 ml QF.

13. Add 14 ml 100% EtOH (or 8 ml isopropanol); Place in dry ice/ethanol bath for 10 minutes (or at -20 °C for 2 hours or at 4°C overnight).

14. Spin at 15 K x G for 30 minutes at 4°C. 15. Wash with 5 ml 70% ethanol; Spin 15 K x G for 20 minutes.

16. Air Dry.

17. Resuspend in 150 μl EB.

DNA concentration was measured as before.

3.77 In Vitro Ligation to Create Full-length cDNA Chimeric Template and RNA Production Digestion with Aatll and BstBI

The plasmid and fragment H were then digested with Aat II and BstB I (NEB) in a sequential digest as follows. pYDl-5'372 (Aatll digest)

Both reactions were incubated in a 37° C block overnight. Five μl of each digestion was mn out on a 1.5% Agarose/TAE gel to check for complete digestion. The digestion was then incubated at 65 °C for 20 minutes to inactivate the enzyme. 2.5 μl Bst BI (NEB) was added to each reaction and placed at 65 °C overnight. The expected results of the digest are as follows:

The largest band from each reaction was gel excised and the UV concentration was determined (as previously described).

There was not enough fragment H for the ligation. Another 50 μl of fragment H was cleaned over a Qiagen Qiaquick column and digested with Aat II and Bst BI as described previously. The digested fragment was then gel excised as before and the UV concentration determined.

3.72 Precipitation of Fragments

The following was prepared in a 1.5 ml tube.

The fragments were then ethanol precipitated and resuspended in 10 43.5 μl of water to facilitate the ligation reaction.

3.73 Ligation

The following ligation reaction was setup using high concentration T4 DNA ligase (NEB). The ligations were incubated at 16°C overnight.

15

0 * heat reaction at 37° for 5 minutes, then briefly chill on ice before adding ligase

3.14 Linearization of Template

The ligation was heat inactivated at 65 °C for 10 minutes. The 5 ligated material was then linearized at the 3' end of the Yellow Fever sequence to allow proper RNA transcription. 5.5 μl of Buffer 2 (NEB) was added to the ligation, followed by 1.5 μl Xhol (NEB), and then the reaction was put at 37° C for 2 hours. 3.75 Linear cDNA extraction (RNAse free phase)

1. Add H₂0 to 100 μl total volume.

2. Add 1/10^th volume 3 M Sodium Acetate

3. Add 100 μl Phenol/Chloroform/Isoamyl Alcohol and spin at 14 Kφm 5 for 5 minutes in a microcentrifuge. Extract upper layer into RNAse-free

1.5 ml tube. Repeat once.

4. Add 100 μl RNAse-free Chloroform. Spin at 14 Kφm for 5 minutes in a microcentrifuge. Extract upper layer into RNAse-free 1.5 ml tube. Repeat once.

10 5. Add 200 μl 100% RNAse-free ethanol.

6. Place on dry-ice/ethanol bath for 10 minutes.

7. Spin at 14 K m for 20 minutes in a microcentrifuge.

8. Wash with 200 μl 70% ethanol (RNAse-free).

9. Repeat 70% ethanol wash two more times.

15 10. Dry in Speed- Vac for 8 minutes (or until no more ethanol is present). 11. Resuspend in 22 μl nuclease free water from the SP6 kit listed below.

3.16 SP6 transcription

The following reaction was setup using the SP6 transcriptase kit 20 (Epicentre) and Rnasin (Promega) in an RNAse-free 1.5 ml tube using RNAse-free tips in a BL-2 hood. The reaction was then placed in a 40° C water bath for 1 hour.

3. 7 Transfection with RNA

Three six well tissue culture plates were seeded with Vero-PM cells (p#162 from Cell Culture Facility) (2 plates at 1 x 10⁶ cells/well and 1 plate at 2 x 10⁶ cells/well) in growth media (Minimum Essential Media, Sodium Pymvate, Non-Essential Amino Acids, Penicillin/Streptomycin, and 5% Fetal Bovine Semm) and placed in a 37 °C C0₂ incubator until confluent.

The following transfection reactions were made using Lipofectin (Life Technologies) and RNAse-free PBS (Sigma). YF/DEN-1

Total RNA control

1. Allow reactions to sit at room temperature for 10 minutes, and then remove Media from the six well plates.

2. Wash 3 times with PBS. 3. Remove last of PBS.

4. Overlay with each lipofectin reaction (add the YF/DEN-1 RNA to the 2 x 10⁶ cells/well plate). Add 280 μl media to the remaining wells.

5. Rock for 10 minutes at room temperature.

6. Wash 2 times with media. 7. Add 2 ml of media to each well and place in the 37 °C C0₂ incubator for 4 days or more.

3.18 Harvest of the first Vero-PM passage (PI)

The supernatant from YF/DEN-1 was harvested on day 6 by splitting the 2 ml of supernatant between two cryovials (each containing 1 ml FBS) that were labeled YF/DEN-1 (PCR) (PI). The cell monolayer was harvested with 1 ml Trizol into a 1.5 ml tube. All vials and tubes were then placed at -80 °C.

3.19 Amplification of YF/DEN-1 Vero-PM Passage #2

1. Three T-25 flasks containing Vero-PM cells (p#166) were obtained from the Cell Culture Facility. A frozen aliquot of YF/DEN-1 (PCR) (PI) was removed from the -80 °C freezer, thawed, and then placed on ice. The same was done for an aliquot of YF/JE (frozen stock from the P 1 control transfection) to be used as a control.

2. Media was removed from each T-25 flask.

3. One ml of YF/DEN-1 (PCR) (PI) was added to the first flask, 1 ml of media only (the same as was used in the transfection) was added to the second flask, and 1 ml of YF/JE(P1) was added to the third flask.

4. Each flask was then put in the 37 °C, 5% C0₂ incubator for 1 hour with rocking every 10 minutes.

5. Four ml of media was then added to each flask. 6. The flasks were then placed in the 37°C, 5% C0₂ incubator for 4 days. Harvest P2 and Passage #3

1. The supernatant from YF/DEN-1 flask was harvested on day 7.

2. Five hundred ml of YF/DEN- 1(P2) was harvested and subsequently overlayed onto a T-25 flask containing Vero-PM cells (p#168 from the Cell Culture Facility). 3. Five hundred ml of media (same as used for transfection) was added to the monolayer.

4. One ml of media only was added to a control flask.

5. The flasks were placed in a 37 °C C0₂ incubator and rocked every 15 minutes for 1 hour.

6. Meanwhile, the remaining YF/DEN- 1(P2) was harvested into 4 cryovials containing 1 ml FBS and 1 cryovial containing 0.5 ml FBS and labeled as YF/DEN- 1(P2). The cell monolayer was harvested with 3 ml Trizol into 1.5 ml tubes. All vials were placed at -80 °C in a box labeled YF/DEN-1.

7. After infection (Step 5), 4 ml of media was added to each flask and were transferred to the incubator for 4 or more days.

Harvest ofP3

The supernatant from YF/DEN-(P3) was harvested on day 5 by splitting the 5 ml of supernatant between five cryovials (each containing 1 ml FBS), which were labeled YF/DEN- 1(P3). The cell monolayer was harvested with 3 ml Trizol into 1.5 ml tubes. All vials and tubes were then placed at -80°C.

3.20 Virus Identification

The RNA from P3 was extracted using Trizol methods according to the manufacturer's protocol, RT-PCR was performed followed by sequencing of the YF/DEN-1 prME region 5', 3' junctions, inclusive. The expected sequence of the prME region was confirmed. 4.0 Construction of ChimeriVax™ YF/DEN3

A viable yellow fever/dengue type 3 chimera (YF/DEN3) was constructed that contains the pre-membrane (prM) and envelope (E) genes of dengue type 3 vims (DEN3) replacing the corresponding prM-E region of the genome of the attenuated 17D yellow fever vims (YF). Virion RNA of wild-type DEN3 (strain PaH881/88) was used as a template to synthesize by RT-PCR two cDNA fragments that cover the DEN3 prM-E region. These fragments were cloned and sequenced. A modified protocol was used to prepare infectious YF/DEN3 in vitro RNA transcripts in which three appropriate DNA fragments were ligated in vitro followed by linearization with Xhol and in vitro transcription with SP6 RNA polymerase (standard ChimeriVax protocol employs two-fragment ligation). Following transfection of Vero PM cells with the RNA transcripts, vims-specific CPE was detectable as early as on day 5 post-transfection (and on day 3 postinfection in subsequent passages). The presence of the chimeric vims in the post-transfection (postinfection) media and the DEN3 -specificity of its prM-E region were confirmed by RT-PCR and sequencing.

These results demonstrate that a chimeric YF virus containing DEN3-specific envelope is readily recoverable. The obtained YF/DEN3 chimera is a candidate part of our proposed terra val ent vaccine directed against all four dengue vims serotypes (dengue types 1-4).

The puφose of these experiments was to determine whether it is possible to create a viable YF/DEN3 chimera containing DEN3 -specific envelope. At the same time, the proposed chimera was designed to contain the prM-E region from a pathogenic wild type strain of DEN3 (a prerequisite for high immunogenicity) in a backbone of the 17D vaccine strain of YF that includes the 5' and 3' UTRs, the C gene, and the nonstructural protein genes, NSl -5, (a prerequisite for safety).

To engineer a YF/DEN3 chimera containing the prM-E cassette from DEN3 in place of the prM-E cassette of YF we first wanted to use the two-plasmid approach that was successful in previous studies where 17D YF vims (Rice et al, New Biol. 1:285-296, 1989) and the YF/JE chimera (Chambers et al, J. Virol. 73:3095-3101, 1999) were recovered following in vitro transcription and transfection. The DEN3 (strain PaH881/88) prM-E region was RT-PCR amplified in two adjacent fragments (Fig. 29). To determine consensus sequence of this region of the parental virus, the RT-PCT fragments were directly sequenced in both directions. Since oligonucleotide primers used to synthesize these fragments were designed based on the published sequence of the H87 reference strain of DEN3 (Osatomi et al, Virology 176:643-647, 1990), actual viral sequences in the primer areas (at the beginning of prM, nucleotides 437-459; at the junction between the two fragments, nucleotides 1079-1131; and at the end of E, nucleotides 2385-2413) could not be determined. A total of 83 nucleotides changes were found compared to H87 strain. The rate of nucleotides differences, 4.44%, was similar to that (4.5%) reported previously by Delenda et al. (J. Gen. Virol. 75:1569-1578, 1994) who sequenced roughly 80% of PaH881/88 E gene. Although the majority of nucleotides differences in the 80% E area coincided with those found by Delenda et al. (V. Deubel, personal communication) (53 changes coincided), there were 4 additional changes that were not found by Delenda et al. In addition, we did not observe 3 of the changes reported by these authors. The PaH881/88 vims (a starting material in our experiments) was isolated from a patient by single amplification in mosquito AP61 cells. We propagated this vims in C6/36, another mosquito cell line. Sequences of both our PI and P3 vimses were found to be identical indicating that there was no selective pressure on the vims in C6/36 cells. Therefore, it is more likely that the few discrepancies were due to sequencing mistakes in (Delenda et al, J. Gen. Virol. 75:1569-1578, 1994). They resulted in three discrepancies in the amino acid sequence of the corresponding region of E protein (C-terminally truncated E; compare Table 32 below with Table 1 in Delenda et al, J. Gen. Virol. 75:1569-1578, 1994). A complete list of amino acid differences we found in the prM-E region between the H87 and PaH881/88 strains is given in Table 33 (a total of 12 changes).

The RT-PCR fragments were used to replace corresponding JE-specific sequences in YFM5'3TV JE SA14-14-2 and YFM5.2 JE SA 14-14-2 plasmids, which resulted in 5'37Den3 and 5.2/Den3 plasmids (Fig. 30). Inserts of both plasmids were sequenced. Since an extra Xhol site was found in the DEN3-specifϊc region of 5'37Den3, the site was ablated by silent mutagenesis that resulted in 5'37Den3/DXho plasmid. The insert of this plasmid was also sequenced to ensure the absence of mutations introduced by PCR.

Difficulties were encountered in obtaining high quality 5.2/Den3 plasmid without mutations within its DEN3-specific region. Different growth conditions of plasmid-containing cultures and modifications in the extraction protocol (e.g., reduction of alkali concentration in the lysis buffer), as well as growth of the plasmid in different E. coli strains were examined to overcome these difficulties. Finally, a clone of the plasmid (#26) was selected (propagated in ABLE C cells) that was of high quality and contained no nucleotide changes except for a single one nucleotide deletion at the 3' end of the DEN3 insert which could be easily eliminated by PCR (other sequenced clones contained more mutations). Therefore, the standard ChimeriVax procedure for preparation of infectious in vitro RNA transcripts that employs two fragment ligation prior to in vitro transcription was modified. According to the standard protocol, the large BstBI- Aatll fragment from 5.2/Den3 would be ligated with the large BstBI- Aatll fragment of 5'37Den3/DXho (see in Fig. 30). Instead, to correct the deletion, three- fragment ligation was done (Fig. 30). The DEN-3 part of 5.2/Den3 was PCR-amplified on the #26 clone template with high-fidelity LA Taq polymerase and digested with BstBI and Ehel (isoschizomer of Narl). The opposite PCR primer was expected to correct the deletion. Second fragment, corresponding to the Narl- Aatll part of 5.2/Den3, was derived by digestion of YFM5.2 JE SA14-14-2 with Ehel and Aatll. The two fragments were ligated with the large BstBI- Aatll fragment of 5'37Den3/ΔXho. Ligation products were digested with Xhol and transcribed in vitro with SP6 RNA polymerase.

Vero PM cells (at passage 149) grown in 6 well plates were transfected with the in vitro RNA transcripts. A first indication that the expected YF/DEN3 chimera was present was the appearance of CPE characteristic of other chimeras created to date based on the YF backbone. It was first noticeable on day 5 post-transfection and became apparent (~10% of detached and rounded cells) on day 7 when vims-containing medium was harvested (PI). Subsequent P2 and P3 vimses were obtained by infecting fresh monolayers of Vero PM cells (at passages 150 and 151, respectively) with the PI and P2 vimses (1 and 0.5 ml of the vimses were used for each infection, respectively) and harvesting the virus when apparent CPE (-10%) was observable (on days 3 and 4 for P2, and day 3 for P3). The presence and DEN3-specificity of the YF/DEN3 chimera was confirmed by RT-PCR with YF- and/or DEN3-specific primers using PI and P2 virion and intracellular RNAs as templates. All these reactions yielded specific RT-PCR products of expected sizes. As was expected, when DEN3 -specific primers were used for RT-PCR on a control YF/JE RNA template, no product was recovered. The entire prM-E region of the P2 vims was sequenced. This also confirmed that the chimera contained the DEN3 envelope glycoprotein genes. In addition to the silent mutation introduced to ablate the Xhol site in the E gene, only one more silent nucleotide change was detected in the vims (A→G at nucleotide 2341 in the chimeric genome that corresponds to nucleotide 2296 in the parental DEN3). Since one of the DNA fragments used in the three-fragment ligation was synthesized by PCR (albeit on a plasmid template with known sequence), it is possible that the PI -3 viruses described here contain minor subpopulations with mutations introduced by PCR. To ensure homogeneity, these vimses can be plaque-purified and then sequenced. In addition, we are currently developing alternative cloning techniques that, if necessary, will allow recreation of the YF/DEN3 genome without using the intermediate PCR step. For instance, the DEN3-specific BstBI-Narl fragment of 5.2/Den3 plasmid was recently cloned without any mutations in low-copy number vectors (pCL and pACYC series). This fragment can be excised from the new plasmids and used instead of the PCR fragment in the three- fragment ligation to regenerate the chimera.

In conclusion, the prM-E region of the PaH881/88 DEN3 was sequenced and cloned. We demonstrate that a recombinant flavivirus genome (e.g., YF/DEN3 in this study) can be reconstituted in vitro by using three- fragment ligation (instead of two-fragment ligation used previously to create other YF chimeras). This approach can be helpful in overcoming technical difficulties that are often encountered during cloning of genetic material from many flaviviruses in E. coli (especially dengue vimses). A viable 17D YF/DEN3 chimeric vims was recovered which is yet another successful example of the usefulness of the approach developed by Chambers et al. (Chambers et al, J. Virol. 73:3095-3101, 1999; see above), in which the prM-E cassette of a heterologous flavivirus is inserted into the YF backbone such that the hydrophobic signal for prM remains YF-specific.

The materials and methods used to make and characterize the YF/DEN3 chimera are described as follows.

4.1 Virus and cells

DEN3 strain PaH881/88 was isolated from a patient by single amplification in AP61 (mosquito) cells. C6/36 cells were maintained in MEM (Gibco, Cat.# 11095-072) supplemented with 10% FBS (HyClone, Cat.# SH30070103) and lx non-essential amino acids (Sigma, Cat.# M7145) (OraVax ML-8 medium, Lot.# 108H2308) at 28°C under 5% C0₂. DEN3 was passaged two times by infecting monolayers of C6/36 at an unknown MOI and harvesting virus-containing growth media on day 7 post-infection (PI and P2) and one time by infecting C6/36 cells with the P2 vims at an MOI of- 0.01 pfu/cell and harvesting the medium on day 6 (P3; pronounced virus-specific CPE was observed in P3). Vims-containing media were mixed with an equal volume of FBS, aliquoted and stored at 70 °C. Following transfection/infection, Vero PM cells were maintained in MEM (Gibco, Cat.# 11095-080, Lot.# 1017611) supplemented with 5% heat-deactivated FBS (OraVax Lot.# AGE6578) and penicillin/streptomycin (100 U/0.1 mg per ml; Sigma, Cat.# P-0781, Lot.# 78H2386) at 37°C under 5% C0₂.

4.2 RNA extraction DEN3 virion RNA was extracted from 0.5 ml of clarified P3 vims-containing medium using TRI Reagent-LS (Molecular Research Center, Inc., Cat.# TS-120) according to the manufacturer's procedure and redissolved in 10 μl H₂0. Alternatively, (e.g., to confirm the presence of YF/DEN3 chimera), intracellular RNA from infected cells was extracted using TRI Reagent (Molecular Research Center, Cat.# TR- 118).

4.4 Reverse transcription, PCR, and sequencing

First strand cDNA syntheses were done in a total volume of 20 μl using 2.5 μl of DEN3 virion RNA as a template, indicated oligonucleotide primers (see below) and Superscript II reverse transcriptase (Gibco, Cat.# 18064-014) according to the manufacturer's procedure. Prior to PCR, RT products were treated with RNAse H (Promega). High-fidelity PCR on RT products or indicated plasmid DNAs as templates was done using the GeneAmp XL PCR kit (Perking Elmer, Cat.# N808-0192), or the TaKaRa LA Taq polymerase kit (PanVera, Cat.# TAK RR002M) according to the protocols provided by the kit manufacturers with a GeneAmp 2400 thermocycler (Perkin Elmer). Sequencing of indicated PCR products or plasmid DNAs was done using the ABI Prism dRhodamine Terminator Cycle Sequencing kit (Perkin Elmer, Cat.# 403042). Sequencing reaction products were resolved with ABI Prism 310 automated sequencer (Perkin Elmer). Data were analyzed using Sequencher 3.0 software and stored on the Internet (ORAVAX/VOLTEMP/GROUPS/LABTECH/KOSTIA/folder "KP sequencing data"/Exp.##). With each area of interest, both DNA strands were sequenced and analyzed. Oligonucleotide primers are listed in Table 32. Primers were ordered from Custom Primers (Life

Technologies/GibcoBRL). In the names of primers, numbers indicate approximate localization of oligos on the DEN3 genome and "+/-" indicates orientation of each primer, with the following exceptions: oligo 5 is colinear with a region of YFM5'3' series of plasmids upstream from the Notl cloning site; oligos 6 (opposite) and 7 (direct) are YF-specific; the former corresponds to the end of YF C gene; oligos 15 (direct) and 16 (opposite) were designed for amplification and sequencing of inserts in the YFM5.2 series of plasmids and correspond to regions of the plasmids located within - 60 nucleotides upstream and downstream from the corresponding inserts, respectively; oligo 8 (direct) was used to mutate the Xhol site at nucleotide 1052 of the recombinant YF/DEN3 genome (within 5'3VDen3 plasmid); and oligo 17 is colinear with the SP6 promoter and a few of the 5' terminal nucleotides from YF.

4.3 DNA manipulations

Standard molecular biology techniques were in accordance with Maniatis et al, Molecular Cloning: a Laboratory Manual, 2^nd Edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 1992. All restriction enzymes, except for Ehel (Fermentas) and T4 DNA ligase, were from New England Biolabs. 4.4 Construction of 5'37Den3/DXho plasmid

The 5' terminal part of the DEN3 prM-E region was synthesized on purified virion RNA of the P3 vims by RT-high-fidelity PCR (XL PCR) using oligonucleotide primers 1 and 2. It starts precisely at the beginning of the coding region for prM protein (at nucleotide 437; DEN3 nucleotide numbering is according to the sequence of H87 reference strain (Osatomi et al, Virology 176:643-647, 1990)) that is generated by signalase and ends at nucleotide 1106 and thus contains the entire prM and approximately one-seventh of the E gene. In addition to DEN3 sequences, the resulting RT-PCR product contains the last 23 nucleotides of the YF C gene for subsequent overlapping PCR (at its 5' end). The last six nucleotides of DEN3 sequence (nucleotides 1101-1106) are changed to a BstBI site by introduction of three silent nucleotide changes for subsequent in vitro ligation, which is followed by a Nhel site for cloning. To combine the RT-PCR product with the upstream YF-specific sequences, a fragment of YFM5'3TV JE SA14-14-2 plasmid (an analog of 5'37Den3 plasmid used to generate a similar YF/JE chimera (Chambers et al, J. Virol. 73:3095-3101, 1999; see above)) containing SP6 RNA polymerase promoter followed by the 17D YF 5' UTR and C gene (first 481 nucleotides of YF genome) was amplified by XL PCR with oligos 5 and 6. For overlapping PCR, the resulting DNA fragment was mixed with the RT-PCR product and XL PCR amplified with oligos 2 and 5. Consensus sequence of the dengue type 3 region was determined by sequencing the RT-PCR and the overlapping PCR products in both directions using oligos 1, 2, 5, 7, 9, and 10.

The overlapping PCR product was used to replace the short Notl-Nhel fragment in YFM5'3TV JE SA14-14-2. The replaced region of the resulting 5'37Den3, which is a pBR322-based plasmid maintained in E. coli MC1061RecA- cells, was sequenced using oligos 1, 2, 9, 10, and 17, and a correct clone (#3) was selected, which does not have any mutations compared to the consensus sequence. Sequencing revealed that the DEN3-specific portion of 5'37Den3 contains an additional Xhol site located in the beginning of E gene (nucleotides 1007-1012 in DEN3 genome). Another Xhol site used for linearization prior to in vitro transcription (see below) is located at the end of YF sequence in 5'3VDen3. The additional site was destroyed by silent oligonucleotide-directed mutagenesis (LA PCR; DEN3-specific C at nucleotide 1009 was changed to G) using oligo 8, resulting in a plasmid 5'37Den3/DXho. The entire region of the plasmid replaced during mutagenesis was sequenced with oligos 1, 2, 9, 10, and 17 and a clone (#10) was selected that does not have any mutations, except for the desired C to G nucleotide change.

4.5 Construction of 5.2/Den3 plasmid

The 3' terminal part of DEN3 prM-E region was RT-PCR amplified (XL PCR) on the P3 virion RNA template using primers 3 and 4. It starts with BstBI site introduced at nucleotides 1101-1106 for in- frame ligation with 5'3'/Den3/DXho plasmid and ends with a Narl site introduced precisely at the 3' end of E gene (nucleotides 2408-2413) for in-frame ligation with YF NSl. The Narl site that leads to Q to G change of the penultimate amino acid residue in the DEN3 E was used previously to generate YF/JE chimera (Chambers et al, J. Virol. 73:3095-3101, 1999; see above). An Nhel cloning site was placed upstream from the BstBI site. The consensus sequence of this DEN3 region was determined by sequencing of the RT-PCR product in both orientations using oligos 3, 4, 11, 12, 13, and 14.

The PCR product was cloned in place of the short Nhel-Narl fragment in YFM5.2 JE SA14-14-2 plasmid (Chambers et al, J. Virol. 73:3095-3101, 1999; see above), resulting in the 5.2/Den3 plasmid.

Originally it was propagated in E. coli MC1061RecA- cells, and the DNA extracted from these cells was of poor quality (partially denatured), which hindered restriction digestions. Subsequently, several of the extracted plasmid clones were propagated in E. coli ABLE C cells (Stratagene), and good quality DNAs were recovered. Six of these clones were thoroughly analyzed by restriction analysis and then sequenced. For sequencing, the DEN3 -specific insert with adjacent regions of the vector was PCR amplified using oligos 15 and 16, and the product was sequenced in both orientations with oligos 11, 12, 13, 14, 15, and 16. Clone #26 was chosen for subsequent manipulations. It contains no mutations, compared to the consensus sequence, except for a single nucleotide deletion (nucleotide 2407, in the region of the opposite PCR primer 4).

4.6 In vitro transcription and transfection

These techniques were essentially as described by Rice et al. (New Biol. 1:285-296, 1989). Approximately 1 μg total of equimolar amounts of appropriate gel-purified DNA fragments were ligated overnight in 20 μl volume at 40°C. T4 DNA ligase was heat-inactivated (10 minutes, 65 °C). Ligation products were digested with Xhol to provide run-off transcription, phenol-chloroform extracted, and precipitated with ethanol. The resulting DNA was transcribed in vitro with SP6 RNA polymerase in the presence of m7G(5')ppp(5')G cap analog and Rnasin in a 20 μl reaction (AmpliScribe SP6 Kit With Cap Analog; Epicentre Technologies; Cat.# AS2606C2). RNA transcripts were analyzed by electrophoresis of 2 μl aliquots in 1% agarose gel. Monolayers of Vero PM cells grown in 6 well tissue culture plates were transfected with RNA transcripts using Lipofectin reagent (Gibco, Cat.# 18292-011). Following transfections, cells were incubated as is described above, and vims-containing media were harvested on indicated days post-transfection, mixed with equal volume of FBS, aliquoted and stored at -70°C.

5.0 Construction of ChimeriVaxTM YF/DEN-4 The puφose was to generate yellow fever/dengue 4 (YF/DEN-4) chimeric vims as a dengue vaccine candidate (see Figs. 31 and 32). To attain this, we used a technology derived from the construction of Yellow fever/ Japanese encephalitis (YF/JE SA 14-14-2) chimeric vims (Chambers et al, J. Virol. 73:3095-3101, 1999). It consists of a two plasmid system which originally encoded the yellow fever (YF) genome. These YF plasmids were created by Charlie Rice (Rice et al, New Biol. 1 :285-296, 1989). The structural membrane precursor and envelope protein genes, i.e., the prME portion, of the YF genome plasmids with that of JE SA14-14-2 sequence and used the resulting plasmids to produce RNA in vitro, which was then transfected into cells to produce live YF/JE chimeric virus. The system seemed suitable to construct other flavivirus chimeras using YF as backbone and here we describe the use of dengue 4 as a start point. The dengue 4 strain, #1228 isolated in 1978 in Indonesia and passaged twice in Mosquitoes, was passed once in C6/36 and total RNA was isolated to synthesize cDNA for PCR of the prME region as needed for cloning. Here we describe in detail the procedures for constmction of the YF/DEN-4 chimera. The dengue 4 prME region was first amplified and sequenced using primers derived from consensus sequences (Genbank). The sequence data created was applied to primer design which were used, with the cDNA produced earlier, as PCR starting point for assembly of the two-plasmid system for dengue 4 (i.e., by replacing the corresponding prME JE sequences in each plasmid). A PCR product encoding dengue prM and 5' end of E was used as template, along with template encoding the capsid (C) gene of yellow fever derived from the plasmid pYF5'3'IV/JE SA 14-14-2, in an overlap extension PCR to result in a single fusion product which was then cloned into a fragment of pYFM5'3TV where JE sequences were deleted. The 3' end of dengue 4 envelope protein gene was also amplified and then cloned into a vector fragment of pYFM5.2/JE SA 14-14-2, resulting in replacement of JE sequence with that of dengue 4. Both plasmids were then transformed into E. coli strain MCI 061 (RecA-) and midi-scale plasmid cultures were grown. In vitro ligation of the two plasmids resulted in full-length vims DNA template of YF/DEN-4 for RNA transcription. All steps involving cDNA fragments and plasmids were carried out in a BL-2 lab designated for recombinant DNA work. Steps involving manipulations of infectious RNA and vims were carried out in a limited access BL-2+ vims lab.

5.7 Amplification of Dengue 4 Sequence

Dengue 4 cDNA was synthesized from RNA using the Superscript II™ method. All primers for this experiment were synthesized by Life Technologies and are listed in Table 34. Upon arrival as lyophilized material, primers were dissolved to 250 mM stock solutions in

RODI-water. From this, 25 mM working solutions were made. The fragment encoding the SP6 promoter and yellow fever capsid (Fragment A) was amplified using the XL-PCR Reaction Kit TM (Perkin-Elmer Part# N808-0192 ), 0.5 ml (250 ng) of template pYF5'3TV plus 3.5 ml RODI-water, and primers 1 and 2. The fragment encoding dengue 4 prM and the 5' end of E (Fragment B) was amplified using the XL-PCR Reaction Kit TM (Perkin-Elmer Part#N808-0192 ) and primers 3 and 4. The fragment encoding the 3' end of dengue 4 envelope (Fragment C) was amplified using the same protocol but using primers 5 and 6. Each PCR reaction was performed as indicated in master mixes (see section 3.1, above). For each reaction, the lower mix (LM) was added to a Perkin-Elmer thin- walled 0.2 ml tube. Next, Ampliwax 100 (Perkin-Elmer) was added to the tube, which was then placed in a Perkin-Elmer 2400 Thermal Cycler and heated to 80°C for 5 minutes, and then cooled to 4°C. The cDNA and UM were then added to the top of the wax layer. The tube was then cycled in a Perkin-Elmer 2400 as follows: 94°C, 1 minute; repeat 30 x (94°C, 15 seconds; 53 °C, 15 seconds; 68°C, 3 minutes), 72 °C, 4 minutes; 4°C, hold. The expected sizes of the PCR fragments for cloning were as follows:

Forty μl of each fragment was then separated on a 1% Agarose/TAE gel and purified using the QIAquick Gel Extraction Kit (Qiagen cat#28704). Next, the concentration of the purified fragments was determined by UV absoφtion using 1 :40 dilutions in RODI-water.

5.2 Recombinant PCR

To create a fusion between the yellow fever capsid gene and the 5' end of the dengue 4 prM, a recombinant PCR technique, Overlap-extension PCR was used to create Fragment E. The same basic UM and LM were used and primers 1 and 4 replaced earlier primers; the cDNA templates used were mixes shown below:

Fragment E Fragment A control Fragment B control

The same protocol that was used for creation of Fragments A, B, and C was used except that only ^lA of the cDNA, UM, and LM were used for the control reactions. The tube was then cycled in a Perkin-Elmer 2400 cycler as follows: 94°C, 1 minute; repeat 30x (94°C, 15 seconds; 55 °C, 15 seconds; 68 °C, 2 minutes; 72 °C, 7 minutes; 4°C, hold. The expected sizes were as follows.

Forty μl of Fragment E was then separated on a 1% Agarose/TAE gel and purified using the QIAquick Gel Extraction Kit (Qiagen cat#28704). Next, the concentration of the purified fragment was determined by UV absoφtion using 1 :40 dilutions in RODI-water.

5.3 Cloning of Fragments C and E into Yellow Fever Vectors

The fragments were then cloned into the yellow fever two-plasmid system by digestion of the purified Fragments C and E as well as the plasmids pYF5'3'IV and pYFM5.2/2 with appropriate restriction enzymes as shown below. The digested plasmids resulted in two bands. The smaller bands contain a fragment of Japanese encephalitis corresponding to either Fragment C or Fragment E for the new dengue 4 constructs. All restriction enzymes, buffers, and lOOx BSA were from New England Biolabs. All the digestions were carried in a Perkin-Elmer 480 cycler set to hold at 37 °C overnight.

Fragment £ digest

pYF5.2 digest

5.4 Vector Dephosphorylation

Calf Intestinal Phosphatase (CIP) from New England Biolabs (cat#290S) was diluted 1 :10 in lx CIP Buffer. One μl of this dilution was then added to the pYFMIV5'3' digest. 0.62 μl of stock CIP was added directly to the pYF5.2 digest. Both were incubated for 1 hour at 37°C. Then, 0.8 μl 125 mM EDTA was added to the two tubes and placed at 75 °C for 10 minutes to inactivate CIP

5.5 Gel Excision The digested PCR fragments were separated on a 1.0%

Agarose/TAE gel, while the digested plasmids were separated on a 0.8% Agarose/TAE gel. All were purified using the QIAquick Gel Extraction Kit (Qiagen cat#28704).

5.6 Ligations

The digested Fragment E and pYF5'3'IV were ligated using T4 DNA Ligase (New England Biolabs cat#202S) to create pYD4-5'3'. The digested Fragment C and pYFM5.2 were ligated to create pYD4-5.2. All ligation reactions were incubated in a Perkin-Elmer 480 cycler set to hold at 16 °C overnight.

PYD4-5'3' Ligation

pYF5.2 Control Ligation

5.7 Transformations All four ligation reactions were transformed into E. coli strain

MCI 061 (recA-). An aliquot of MCI 061 (OraVax Notebook 661-4) was removed from storage at -80 °C and allowed to thaw on ice for one to two minutes. 0.9 ml of cold 0.1 M CaCl₂ was added to the cells. One hundred μl of cells was aliquoted into five 12 ml culture tubes on ice. Ten μl of each ligation reaction was added to each culture tube, leaving the fifth tube as a negative (no DNA) control. Culture tubes were left on ice for 30 minutes. The tubes were heat shocked in a water bath at 42 °C for 45 seconds. The tubes were put back on ice for 2 minutes. 0.9 ml SOC medium was added to each culture tube and incubated at 225 pm in a shaking incubator at 37° C for 1 hour. Each transformation mix was aliquoted into 1.5 ml microcentrifuge tubes. One hundred μl of each was spread onto LB/Agar-Amp (100 μg/ml) plates and labeled as "neat." Each tube was spun at 14 Kφm in a microcentrifuge for 2-3 seconds to pellet the cells. The supernatant was poured into the waste container and the pellet resuspended in the residual broth by pipetting up and down. This material was plated (approximately 100 μl) onto LB/Agar-Amp (100 μg/ml) plates and labeled as lOx. All plates were inverted in a 37° C incubator overnight. 5.8 Transformant Screening

The resulting bacterial colonies were patch-plated onto fresh LB/Agar-Amp (100 μg/ml) and placed inverted in a 37° C incubator overnight. The following day, 50 μl of RODI-water was aliquoted into 0.5 ml tubes. Using a sterile plastic pick, a small amount of each patch was scraped into one of the 0.5 ml tubes containing water. These were then placed at 95 °C for 10 minutes and spun at 14,000 rpm for 10 minutes in a microcentrifuge. To identify the proper insert in pYD4-5'3', colonies were screened by PCR using Taq Polymerase (Promega) and primers 4 and 7. The pYD4-5.2 was screened using primers 5 and 6. The PCR reaction was performed using the following master mix.

Forty eight μl of each master mix was added to a 0.5 ml tube along with 5 ml of the template prepared from the colony patches. Additionally, a tube using RODI-water as a template was also made as a negative control. These tubes were then placed in the Ericomp Delta Cycler and cycled as follows: 96°C, 4 minutes; 30x (94°C, 30 seconds; 50°C, 1 minute; 72°C, 1 minute 25 seconds), 72°C, 4 minutes; 4°C hold. The PCR products were then mn on 1.5% Agarose/TAE gels to check for positive colonies. 5.9 Glycerol Stocks

Five ml of LB- Amp (100 μg/ml) was then inoculated from a patch pYD4-5'372 or ρYD4-5.2/l and shaken at 225 φm overnight at 37°C. One ml of this culture was then spun at 14 Kφm for 2-3 seconds to pellet the cells. This was then resuspended in LB-Glycerol (30%) and frozen at -80°C.

5.10 MIDI Plasmid Preparation

Fifty μl of each glycerol stock was added to 150 ml LB- Amp (100 μg/ml) in separate 4 L flasks and shaken at 225 φm overnight at 37 °C. Qiagen Midi-Prep (Qiagen) was performed using the following modified protocol.

1. Spin 150 ml of each culture at 7 Kφm in GSA rotor for 10 minutes to pellet. 2. Decant Supernatant.

3. Resuspend pellet in 4 ml PI Buffer; transfer to 50 ml Falcon tube.

4. Rinse centrifuge bottle with 2 ml PI buffer and transfer to the Falcon tube.

5. Add 6 ml P2 buffer; invert gently; 5 minutes at room temperature or until lysed (no more than 12 minutes).

6. Add 6 ml P3; mix as above; 10 minutes on ice.

7. Transfer supernatant to Qiagen Syringe Filter; let sit for 10 minutes.

8. Equilibrate Q-100 tip with 4 ml QBT.

9. Gently push plunger to filter supernatant onto Q-100 tip. 10. Allow to drain by gravity. 11. Wash with 2 x 10 ml QC.

12. Elute into 50 ml polypropylene centrifuge tube with 5 ml QF.

14. Spin at 15 K x G for 30 minutes at 4°C.

15. Wash with 5 ml 70% ethanol; Spin 15 K x G for 20 minutes.

16. Air Dry.

17. Resuspend in 150 μl EB. DNA concentrations were measured as before.

5.11 In Vitro Ligation to Create Full-length cDNA Chimeric Template and RNA Production

Digestion with Aatll and BstBI

Plasmids pYD4-5*372 and pYD4-5.2/l were digested with Aatll and BstBI (NEB) in a sequential digest as follows. YD4-5'372 (Aatll digest)

YD4-5.2/1 (Aatπ digest)

Both reactions were incubated in a 37 °C block for 2 hours. Five μl of each digest was mn out on a 1.5% Agarose/TAE gel to check for complete digestion. The pYD4-5'372 digest did not cut completely so the reaction was cleaned over a Qiaprep spin column (Qiagen). The digest was repeated using this material and 3 μl of Aat II. In addition, 3 μl of Aat II was added to the existing pYD4-5.2/l reaction. Both were incubated in a 37 °C block, overnight. After confirmation of digest on another gel (as previously described), 2.5 μl Bst BI (NEB) was added to each reaction and placed at 65 °C for 3 hours. The results of the digest were as follows.

The largest band from each reaction was gel excised as and the UV concentration was determined (as previously described).

5.72 Ligation

*Heat reaction at 37°C for 5 minutes, then briefly chill on ice before adding ligase

5.73 Linearization of Template

The ligation was heat inactivated at 65 °C for 10 minutes. The ligated material was then linearized at 3' end of the yellow fever sequence to allow proper RNA transcription, as follows: 5.5 ml Buffer 2 (NEB) was added to the ligation, followed by 1.5 μl Xhol (NEB), and this reaction mixture was placed at 37° C for 2 hours.

5.14 Linear cDNA extraction (RNAse free phase) 1. Add H₂0 to 100 μl total volume. 2. Add 1/10^th volume 3 M Sodium Acetate.

3. Add 100 μl Phenol/Chloroform/Isoamyl Alcohol and spin at 14 Kφm for 5 minutes in a microcentrifuge. Extract upper layer into RNAse-free 1.5 ml tube. Repeat once.

4. Add 100 μl RNAse-free Chloroform. Spin at 14 Krpm for 5 minutes in a microcentrifuge. Extract upper layer into RNAse-free 1.5 ml tube.

Repeat once.

5. Add 200 μl 100% RNASE-free ethanol. 6. Place on dry-ice/ethanol bath for 10 minutes.

7. Spin at 14 Kφm for 20 minutes in a microcentrifuge.

8. Wash with 200 μl 70% ethanol (RNAse-free).

9. Repeat 70% ethanol wash two more times. 10. Dry in Speed- Vac for 8 minutes (or until no more ethanol is present). 11. Resuspend in 22 μl nuclease free water from the SP6 kit listed below.

5.75 SP 6 transcription

The following reaction was setup using the SP6 transcriptase kit (Epicentre) and Rnasin (Promega) in an RNAse-free 1.5 ml tube using RNAse-free tips in a BL-2 hood. The reaction was then placed in a 40° C water bath for 1 hour.

5.16 RNA Transfection

Two six well tissue culture plates were seeded with Vero-PM (ρ#153 OraVax notebook#743-7) cells at 7.4 x 10⁵ cells/well in growth media (Gibco MEM; Sodium Pymvate; NEAA; Penicillin/Streptomycin; 5% fetal bovine semm), and placed in a 37° C C0₂ incubator until confluent.

The following transfection reactions were made using Lipofectin (Life Technologies) and RNAse-free PBS (Sigma), and allowed to sit at room temperature for 10 minutes.

YF/DEN-4

Total RNA control

Lipofectin control

1. Remove Media from the 6 well plates. 2. Wash 3 times with PBS.

3. Remove last of PBS.

4. Overlay with each lipofectin reaction. Add 280 μl PBS for the remaining wells. 5. Rock for 10 minutes at room temperature.

6. Wash 2 times with media (MEM, Sodium Pymvate, NEAA, P/S, 5% FBS).

7. Add 2 ml of media to each well and place in the 37°C C0₂ incubator for 4 days or more.

5. 7 Chimeric Virus Harvest

The supernatant from YF/DEN-4 was harvested on day 6 by splitting the 2 ml of supernatant between two cryovials (each containing 1 ml FBS), which were labeled YF/DEN-4 (PI). The cell monolayer was harvested with 1 ml Trizol into a 1.5 ml tube. All vials and tubes were then placed at -80°C.

5.75 Amplification of YF/DEN-4 Passage #2

1. Three T-25 flasks containing Vero-PM cells (p#149) were obtained from the Cell Culture Facility. A frozen aliquot of YF/DEN-4 (PI) was removed from the -80 °C freezer, thawed, then placed on ice. The same was done for an aliquot of YF/JE (frozen stock from the PI control transfection).

2. Media was removed from each T-25 flask.

3. Five hundred μl of YF/DEN-4(P1) was added to the first flask, 500 μl of media (MEalphaM, NEAA, Sodium Pymvate, 5% FBS, P/S) was added to the second flask, and 500 μl of YF/JE(P1) was added to the third flask. 4. Two hundred μl of the media was added to each flask to ensure complete coverage.

5. Each flask was then put in the 37 °C C0₂ incubator for 1 hour with rocking every 15 minutes. 6. 4.5 ml of media was then added to each flask.

7. The flasks were then placed in the 37°C C0₂ incubator for 4 days.

Harvest P2

The supernatant from YF/DEN-4 was harvested on day 4 by splitting the 5 ml of supernatant between five cryovials (each containing 1 ml FBS), which were labeled YF/DEN-4 (P2). The cell monolayer was harvested in 3 ml Trizol and aliquoted into 3 tubes. All vials and tubes were then placed at -80 °C.

Passage#3 (Titration ofP2)

1. A 12 well tissue culture plate containing Vero-PM cells (p#163) was obtained from the Cell Culture Facility. A frozen aliquot of YF/DEN-4

(P2) was removed from the -80 °C freezer, thawed, and then placed on ice.

2. Serial dilutions were made in a 24 well tissue culture plate ending at 10^"4 in a total volume of 300 μl. The plate was then placed on ice.

3. The media was removed from each well of the 12 well plate and 100 μl of vims stock, as well as each dilution, was added to the wells in duplicate.

One hundred μl of media only was added to the last set of wells as a negative control.

4. The plate was then put in the 37 °C C0₂ incubator for 1 hour with rocking every 20 minutes. 5. The 1 ° overlay was made by preheating 25 ml M199(2X), 1.5 ml FBS, and 0.5 ml PSA at 42°C in a 50 ml Falcon tube (tube #1). Additionally, 23 ml Agarose (0.6% in water) was heated at 42 °C in a 50 ml Falcon tube (tube #2). At the end of the 1 hour incubation, tube #1 was added to tube #2 and mixed thoroughly.

6. One ml of this overlay was then added to the edge of each well. 7. The plate was then put in the 37° C C0₂ incubator for 4 days.

8. The 2° overlay was made by preheating 25 ml M199(2X), 1.5 ml FBS, 1.5 ml Neutral Red, and 0.5 ml PSA at 42° C in a 50 ml Falcon tube (tube #1). Additionally, 21.5 ml Agarose (0.6% in water) was heated at 42 °C in a 50 ml Falcon tube (tube #2). Tube #1 was added to tube #2 and mixed thoroughly.

9. One ml of this overlay was added to the center of each well.

10. It was then placed in the 37 °C C0₂ incubator

Titration ofP2 results

Instead of titer determination, plaques were picked for purification to segregate a mixed population of large and small plaques observed. The RNA from P2 was extracted using Trizol methods according to the manufacturer's protocol, RT-PCR was performed followed by sequencing of the YF/DEN-4 prME region 5', 3' junctions, inclusive. The expected sequence of the prME was confirmed.

6.0 Construction of Chimeric Templates for Other Flaviviruses

Procedures for generating full-length cDNA templates encoding chimeric YF/MVE, YF/SLE, YF/WN, YF/TBE vimses are similar to those described above for the YF/DEN-2 system. Table 20 illustrates the features of the strategy for generating YF 17D-based chimeric vimses. The unique restriction sites used for in vitro ligation, and the chimeric primers for engineering the C/prM and E/NSI junctions are also shown. Sources of cDNA for these heterologous flavivimses are readily available (MVE: Dalgamo et al, J. Mol. Biol. 187:309-323, 1986; SLE: Trent et al, Virology 156:293-304, 1987; TBE: Mandl et al, Virology 166: 197-205, 1988; Dengue 1: Mason et al, Virology 161 :262-267, 1987; Dengue 2: Deubel et al, Virology 155:365-377, 1986; Dengue 3: Hahn et al, Virology 162:167-180, 1988; Dengue 4: Zhao et al, Virology 155:77-88, 1986).

An alternative approach to engineering additional chimeric viruses is to create the C/prM junction by blunt end ligation of PCR-derived restriction fragments having ends that meet at this junction and 5' and 3' termini that flank appropriate restriction sites for introduction into YF5'3TV or an intermediate plasmid such as pBS-KS(+). The option to use a chimeric oligonucleotide or blunt-end ligation will vary, depending on the availability of unique restriction sites within the envelope protein coding region of the vims in question.

6.1 Construction ofYF Viruses Encoding HCV Antigens

Because the structural proteins El and E2 of HCV are not homologous to the structural proteins of the flavivimses described above, the strategy for expression of these proteins involves insertion within a nonessential region of the genome, such that all of these proteins are then co-expressed with yellow fever proteins during viral replication in infected cells. The region to be targeted for insertion of the proteins is the N terminal portion of the NSl protein, since the entire NSl protein is not required for viral replication. Because of the potential problems with stability of the YF genome in the presence of heterologous sequence exceeding the normal size of the genome (approximately 10,000 nucleotides), the detection strategy described below can be used. In addition, deletion of NSl may be advantageous in the chimeric YF/Flavivims systems described above, because partial deletion of this protein may abrogate the immunity to YF associated with antibodies against NSl, and thus avoid problems with vector immunity if more than one chimeric vaccine was to be needed in a given recipient, or if a YF vaccine had been previously given or needed at a future point.

The strategy involves creating a series of in- frame deletions within the NSl coding region of the YFM5.2 plasmid, in conjunction with engineering a translational termination codon at the end of E, and a series of two IRESs (internal ribosome entry sites). One IRES is immediately downstream of the termination codon and allows for expression of an open reading frame within the region between E and NSl. The second IRES initiates translation from truncated NSl proteins, providing expression of the remainder of the YF nonstructural polyprotein. These derivatives are tested for recovery of infectious vims and the construct with the largest deletion is used for insertion of foreign sequences (e.g., HCV proteins) in the first IRES. This particular construct can also serve as a basis for determining whether deletion of NSl will affect vector-specific immunity in the context of YF/Flavivims chimeric vimses expressing prM-E, as described above.

The insertion of nucleotides encoding El, E2, and/or El plus E2 HCV proteins is limited by the size of the deletion tolerated in the NSl protein. Because of this, truncated HCV proteins can be used to enhance stability within the modified YF clone. The HCV proteins are engineered with an N-terminal signal sequence immediately following the IRES and a termination codon at the C terminus. This construction will direct the HCV proteins into the endoplasmic reticulum for secretion from the cell. The strategy for this construction is shown schematically in Fig. 21. Plasmids encoding HCV proteins of genotype I can be used for these constructions, for example, HCV plasmids obtained from Dr. Charles Rice at Washington University (Grakoui et al, J. Virology 67: 1385-1395, 1993), who has expressed this region of the vims in processing systems and within a replication-complement full-length HCV clone.

6.2 PrM cleavage deletion mutants as attenuating vaccine candidates for flaviviruses

Additional chimeric vimses included in the invention contain mutations that prevent prM cleavage, such as mutations in the prM cleavage site. For example, the prM cleavage site in flavivirus infectious clones of interest, such as dengue, TBE, SLE, and others can be mutated by site-directed mutagenesis. Any or all of the amino acids in the cleavage site, as set forth above, can be deleted or substituted. A nucleic acid fragment containing the mutated prM-E genes can then be inserted into a yellow fever virus vector using the methods described above. The prM deletion can be used with or without other attenuating mutations, for example, mutations in the E protein, to be inserted into the yellow fever vims. These mutants have advantages over single substitution mutants as vaccine candidates, because it is almost impossible to revert the deleted sequence and restore vimlence. The following chimeric flaviviruses of the invention were deposited with the American Type Culture Collection (ATCC) in Rockville, Maryland, U.S.A. under the terms of the Budapest Treaty and granted a deposit date of January 6, 1998: Chimeric Yellow Fever 17D/Dengue Type 2 Vims (YF/DEN-2; ATCC accession number ATCC VR-2593) and Chimeric Yellow Fever 17D/Japanese Encephalitis SA14-14-2 Vims (YF/JE A1.3; ATCC accession number ATCC VR-2594). Table 1

Virus E E E E E E E E E E

107 138 176 177 227 243 244 264 279 315

JE SAl₄-l₄-₂ F K V T S K G H M V

YF/JE SA 14- F K V A S E G H M V

14-2

YF/JE L E I T P E E Q A

Nakayama

JE L E I T P E E Q A

Nakayama

JE SA14 L E I T s E G Q V

Table 2

Characterization of YF/JF, chimeras

Clone Yield (μg) Infectivity PBS RNAse DNAse plaques/100 ng log titer log titer log titer

LLC-M 2 VERO VERO VERO

YF5 21v 5 5 15 7 2 0 7 YF/JE-S 7 6 ₅₀ 6 2 0 6 2 YF/JE-N 7 60 5 0 5 4 Table 3

Plaque reduction neutralization titers on YF/IE chimeras

Virus non-immune YF ascitic JE ascitic non-immune YF I ascitic fluid fluid fluid IgG

YF5.2iv <1.3 3.7 <1.3 <2.2 >4.3

JE SA14-14-2 <1.3 <1.3 3.4 <2.2 <2.2

YF/JE SA₁₄- <1.3 <1.3 3.1 <2.2 <1.9

14-2

YF/JE <1.3 <1.3 3.4 <2.2 <2.2

Nakayama

JE Nakayama <1.3 <1.3 3.4 not not

V1TUS determinedSe ptember 15, 1998

Table 4

Neurovirulence of YF/.TE SA 14- 14-2 Chimera week old male ICR mice

log dose I.C. % Mortality

YF5.2iv 4 100 (7/7)

YF/JE SA14-14-2 4 0 (0/7)

YF/JE SA14-14-2 5 0 (0/7)

YF/JE SA14-14-2 6 0 (0/8) Table 5

Neuroinvi-siveness of YF/.TE Chimeras 3 week old male ICR mice

log dose (intraperitoneal) % mortality

YF/JE Nakayama 4 0 (0/5)

YF/JE Nakayama 5 0 (0/4)

YF/JE Nakayama 6 0 (0/4)

YF/JE SA14-14-2 4 0 (0/5)

YF/JE SA14-14-2 5 0 (0/4)

YF/JE SA14-14-2 6 0 (0/4)

Table 6

Doses and routes of virus inoculation into groups of 4-week-old ICR mice Group YF/JE s.c. YF/JE i.e. YF-VAX s.c. YF-Vax i.e. Total # log₁₀pfϊι log.opfu log₁₀pfu log₁₀pfu mice

Table 7

Geometric mean neutralizing antibody titers 3 and 8 weeks after immunization, onthred mice inoculated with graded doses of vaccines by the s.c. route

Vaccine Dose Antibody titer (GMT ± SD) vs. Iog₁₀PFU

IE YF17D

3w 8w 3w 8w

YF/JE 5.0 151 ±93 5,614 ±3514 4.0 38 ±60 127 ± 247 3.0 19 ±65 43 ± 560 2.0 7 ±12 3± 71 1.0 2 ±8 0

YF17D 4.7 2 ±4 18 ± 13 4.4 35 ±24 250 ± 109 3.4 9 ±20 54 ± 179 2.4 l± 0 53 ±22 1.4 0 46 ±18

Table 8

Immunogenicity and protection vs. challenge

Mice were immunized on Day 0 with live vaccines and on days 0, 7, and 20 with JE-Vax, bled on day 21 and challenged on day 28.

Virus No. /group Dose (pfu) Route Total no. mice

* YF/JE SA 14- 14-2 vaccine candidate

** Chinese live vaccine, passed once in BHK cells

# Chimeric YF/JE virus, with prM-E insert of wild-type JE Nakayama

Tnble ?. Prolcciioπ of C57/II 6 mice by n single SC inoctihim of grndcd doses of live vims vncciπes ngninsl IP chnllengc 158 LD50 of wild-fype JU vims (IC-37). Mice clinllcngcd 28 days nfler immunization.

^• Not tested

II Some mice died ns a result of inoculation of the wild-type vims nt high doses, llnis fewer mire rf.ιι.nin for challenge

• ^• I luce doses nl I week intci vnls

^Λ No mice survived inilinl inoculnlion nt this dose

7β Corrected from previous report

Table 10

Geometric mean neutralizing antibody titers, C 7/RT.6 mice 21 days after immunization with a single SC inoculum of graded doses of live vims vaccines and 1 day after the third dose of inactivated .TE-Vax.

* only 2/8 mice survived immunization with this virus; the low antibody titers in these animals probably reflect low level virus replication consistent with survival.

** 3 doses on days 0, 7, 20; animals were bled on day 21, 1 day after their 5 third immunization. The day 20 boost was performed with a higher dose of vaccine, thus antibody titers pre-challenge are expected to be higher than those shown here.

Table 11

10 Geometric mean neutralizing antibody titers (GMT) in 3 monkeys 2 and 4 weeks post inoculation with a single dose of YF-Vax® or ChimeriVax™ by the I.C. route

Vaccine Dose πogl O fii) GMT

IE YE

2W 2_

15 4W

ChimeriVax™ 7.0 75 3200

YF-Vax® 5.0 66 4971

Table 12 Immunization and protection: rhesus monkeys

Screening HI test for flavivirus antibodies: negative

Group N Virus Dose, route (log₁₀PFU/0.5 ml) JE Challenge Day 60

1 3 YF/JE SA14-14-2 4.3 SC 5.0 IC

2 3 YF/JE SA14-14-2 5.3 SC 5.0 IC

3 4 Saline/sham - SC 5.0 IC

• Viremia days 1-7 after immunization and challenge

• Neutralization test days 0, 15, 30, 45, and 60 after immunization and days 15 and 30 after challenge £

• Necropsy day 30 post challenge

Table 13 Viremia, rhesus monkeys immunized with ChimeriVax by the SC route

Monkey Dose Day post- inoculation log₁₀PFU 0 1 2 3 4 5 6

R423 4.3 <1.0* <1.0 <1.0 1.1 1.7 1.0 <1.0

R073 <1.0 <1.0 <1.0 1.0 1.0 <1.0 <1.0

R364 <1.0 1.0 <1.0 1.0 1.0 <1.0 <1.0

R756 5.3 <1.0 1.0 1.0 1.6 1.0 <1.0 <1.0 l<

R174 <1.0 1.3 1.8 1.6 1.1 <1.0 <1.0

R147 <1.0 2.0 1.6 1.0 1.0 <1.0 <1.0

* log_l0PFU/ml

Table 14 JE neutralizing antibody responses, rhesus monkeys immunized with ChimeriVax by the SC Route

0% PRNT titers, heat-inactivated serum, no added complement

Monkey Dose Day post-inoculation log₁₀PFU Baseline 15 30

R423 4.3 <10 160 2560

R073 <10 80 640

R364 <10 160 320 l< U

R756 5.3 <10 20 320

R174 <10 640 2560

R147 <10 160 2560

Table 15 Protection against IC challenge, rhesus monkeys immunized with ChimeriVax by the SC route

Monkeys challenged IC on Day 60 with 100,000 pfu/mouse LD50

Vaccine Dose log_]0PFU No. survived/No. tested

4.3 2/3 (67%)*

5.3 3/3 (100%)

Sham 0/4 (0%)

1 monkey that died was a pregnant female

Table 16

List of chimeric YF/IE mutants (1 to 9) constnicted to identify residues involved in attenuation of the ChimeriVax™. Mutated amino acids on the F-proteins are shown in bold letters.

Mutant Vimses

Table 17 Dose administered i.e. (pfu) Group P1 P10 P18

Neat >6xl0⁴ lxlO⁶ 2xl0⁷

10^"1 >6xl0³ lxlO⁵ 2xl0⁶

Table 18

Dose administered s.c. (pfu)

Group RMS P10 P18

Neat 2xl0⁵ 2xl0⁷ 3xl0⁷

10⁵ lxlO⁵ 5xl0⁵ 5xl0⁴

10⁴ lxlO⁴ 5x10⁴ 5xl0³ Table 19

Design of an experiment to determine cross-protection/interference between YF 17D and YF/JE SA14- 14-2

f: One dose of YF/JE SA14-14-2 , 5.3 log₁₀ pfu/mouse, sc.

One dose of YF-Vax®, 4.4 log₁₀ pfu mouse, sc.

Two doses of JE-Vax® (PMC), 0.5 ml of 1 :5 dilution administered ip at 1 week intervals.

Table 20

Engineering of YF/Flaviγinjs chimeras

7 Virus Chimeric C/prM Chimeric E/NSl 5* 3' Sites⁵ junction¹ junction² ligation³ ligation⁴ eliminated or (created)

YF/WN X-cactgggagagcttgaaggtc aaagccagttgcagccgcggtttaa Aatll Nsil

(SEQ ID NO: 1) (SEQ ID NO:2) YF/DEN-1 X-aaggtagactggtgggctccc gatcctcagtaccaaccgcggtttaa Aatll Sphl Sphl in DEN

(SEQ ID NO:3) (SEQ ID NO:4) YF/DEN-2 X-aaggtagattggtgtgcattg aaccctcagtaccacccgcggtttaa Aatll Sphl

(SEQ ID NO:5) (SEQ ID NO:6) YF/DEN-3 X-aaggtgaattgaagtgctcta acccccagcaccacccgcggtttaa Aatll Sphl Xhol in DEN

(SEQ ID NO:7) (SEQ ID NO:8) (Sphl in DEN) YF/DEN-4 X-aaaaggaacagttgttctcta acccgaagtgtcaaccgcggtttaa Aatll Nsil

(SEQ ID NO:9) (SEQ ID NO: 10) YF/SLE X-aacgtgaatagttggatagtc accgttggtcgcacccgcggtttaa Aatll Sphl Aatll in SLE

(SEQ ID NO: 11) (SEQ ID NO: 12) YF/MVE X-aatttcgaaaggtggaaggtc gaccggtgtttacagccgcggtttaa Aatll Agel (Agel in YF)

(SEQ ID NO: 13) (SEQ ID NO: 14) YF/TBE X-tactgcgaacgacgttgccac actgggaacctcacccgcggtttaa Aatll Agel (Agel in YF)

(SEQ ID NO: 15) (SEQ ID NO: 16)

1,2: The column illustrates the oligonucleotide used to generate chimeric YF Flavivirus primers corresponding to the C/prM or E/NSl junction. (See text). X = carboxyl terminal coding sequence of the YF capsid. The underlined region corresponds to the targeted heterologous sequence immediately upstream of the Narl site (antisense - ccgcgg). This site allows insertion of PCR products into the Yfm5.2 (Narl) plasmid required for generating full-length cDNA templates. Other nucleotides are specific to the heterologous virus. Oligonucleotide primers are listed 5' to 3'.

3,4: The unique restriction sites used for creating restriction fragments that can be isolated and ligated in vitro to produce full-length chimeric cDNA templates are listed. Because some sequences do not contain convenient sites, engineering of appropriate sites is required in some cases (footnote 5).

5: In parentheses are the restriction enzyme sites that must be created either in the YF backbone or the heterologous virus to allow efficient in vitro ligation. Sites not in parentheses must be eliminated. All such modifications are done by silent mutagenesis of the cDNA for the respective clone. Blank spaces indicate that no modification of the cDNA clones is required. Sequence coniparsion of Dengιιc-2 and YF/Dcπ-2_2J8 viruses

Virus prM

28 31 55 57 125 152 161

YF/D2_m E V L R r Λ V

PUO- E V L R r Λ V

218

NGC E V F R T A V

PR-159 K T F K T V I

(SJ)

Virus ENVELOPE

71 81 126 129 139 141 162 16'! 202 203 335 352 390 402 4M

18

Nt;e D K V J 1 J I li N 1 I N ϊ V

PR-159 D T E I V I V I K D T T D F 1

(SI)

Table 22

Summary of histopathology results, monkeys inoculated with YF-Vax or YF/JE SA 14- 14-2 by the IC route

Table 23

List of initial chimeric YF/JE mutants constructed to identify residues involved in attenuation of the ChimeriVax TM. Reverted amino acids on the E-proteins are shown in BOLD

Table 24

Experiment to determine neurovirulence and neuroinvasiveness phenotypes of vaccine candidates in suckling mice

*: PFU/0.02 ml of inoculum jblc

Table 26. Immunogenicity of ChimeriVax^T -D2

s

B: eometric Mean Titers measured as t e ast iution of sera which resulted in 50% reduction in number of virus plaques.

C: Titers less than 1:10

Table 27. Immunization and challenge of yellow fever immune monkeys

Table 28. Viremia in rhesus monkeys inoculated SC with graded doses of ChimeriVax™-Den-2 virus.

Table 29. Neutralizing antibody responses in monkeys inoculated with graded doses of ChimeriVax™— Den-2 and protection against wild-type dengue-2 challenge

Dose No. GMT bv day No.Viremic Log_!0 PFU seroconverted 0 15^' 30 after

Challenge

(D60)~

2.0 4/4 <10 47 316 | 0/4

3.0 5/5 <10 112 275 i 0/5

4.0 | 5/5 <10 269 316 i 0/5

5.0 | 4/4 <10 631 376 | 0/4

Sham | 0/4 i <10 <10 <10 | 4/4

Table 30. Mean virus titer in orally exposed mosquitoes 22 days post feeding

Mosquito Virus # Infected/ ean Species S Fed Titer^*

JE SA14-14-2 20/20 6.0

Ae. aeg\-pti

YF 17D 0/30 < 1.0

1 ChimeriVax™-JE 0/40 < 1.0

Ae. JE SA 14- 14-2 3S/3S 5.S albopictus

YF 17D 0/44 < 1.0

logιopfu/ml Table 31

Primers (restriction sites are underlined)

#1) YFM5'3'(4.56)+

(GTGAGCATTGAGAAAGCGCCACGCTTC)(SEQ ID NO: 17)

#2) YF0.481-

(TCCACCCGTCATCAACAGCATTCCCAAAATTAG)(SEQ ID NO: 18)

#3) IDE 0.42+

(GAATGCTGTTGATGACGGGTGGATTTCATCTGACCACACGAGGG) (SEQ ID NO: 19)

#4) IDE 1.095-

(MιeI/Bs i)(GCCGCIAG( TTllCGΔΔGGACGGCAGGGTTTGTGACT TC)(SEQ ID NO:20)

#5) IDE 1.102+ (BstBI)(GCCATGCATIICGAAAACTGTGCATCGAAGCTAAAATAT

C)(SEQ ID NO:21)

#7) IDE 2.409FUSE-

(GGCGCATCCTTGATCGGCGCCAACCATGACTCCTAGGTACAG)(SE Q ID NO:22)

#8) YF Narl+

(GGCGCCGATCAAGGATGCGCCATC)(SEQ ID NO:23)

#9) YF 8.545-

(CCAAGAGGTCATGTACTCAG)(SEQ ID NO:24)

#10) SP6YFa+

(ATTTAGGTGACACTATAGAGTAAATCCTGTGTGCTAATT)(SEQ ID

NO:25)

Table 32

Oligonucleotide primers. No. Name

Sequence

1 3D 0.432+

5*-GAATGCTGTTGATGACGGGTGGATTCCACTTAACTTCACGAG ATGG (SEQ ID NO:26)

2 3DE 1.095-

5'-GCCGCTAGCCTTTCGAAGGGTCGCCAGCTGAGTGGCCTC (SEQ ID NO:27)

3 3DE 1.102+ 5'-GCCGCTAGCTTCGAAAGCTATGCATTGAGGGAAAAATTAC (SEQ ID NO:28)

4 3DE 2.409-

5*-GCCGCCGGCGCCCACCACGACCCCCAGATAGAGTG (SEQ ID NO:29)

5 YFM5'3'(4.65)+ 5'-GTGAGCATTGAGAAAGCGCCACGCTTC (SEQ ID NO:30)

6 YF 0.481 -

5'-TCCACCCGTCATCAACAGCATTCCCAAAATTAG (SEQ ID NO:31)

7 YF 0.2+ 5'-ATGGTACGACGAGGAGTTCGC (SEQ ID NO:32)

8 Kpd3/-Xho/+

5'-GGTTGATGTGGTGCTGGAGCACGGTGGGTGTG (SEQ ID

NO:33)

9 PsD3/ .7+ 5'-TACATCGACATGGGTGAC (SEQ ID NO:34)

10 KPsD3/ .75- 5'-GACATGGGGAGCTAACGC (SEQ ID NO:35)

11 KPsD3/ 1.5- 5'-CCCGAGGGTTCCATATTCAGG (SEQ ID NO:36)

12 KPsD3/ 1.6+ 5'-GGAACAGGAAAGAGCTTC (SEQ ID NO:37)

13 KPsD3/ 1.9- 5'-GAGTATTGTCCCATGCTG (SEQ ID NO:38)

14 KPsD3/ 2.1+ 5'-GGAATTGGAGACAAAGCC (SEQ ID NO:39)

15 KPs5.2 0.23+ 5'-TGGATAGTGGACAGACAGTGG (SEQ ID NO:40)

16 KPs5.2 1.66- 5'-CTCTAAATATGAAGATACCATC (SEQ ID N0.41)

17 SP6-yfa

5-.ATTTAGGTGACACTATAGAGTAAATCCTGTGTGCTAATT (SEQ ID

NO:42)

Table 33

Amino acid differences in the prM-E region of DEN3 H87 and PaH881/88 strains

Position H87(PaH881/88 Position H87(PaH881

(nt/gene)¹ (codon change) (a.a.)² (a.a. change)

599/ prM CAC→CTC 168 H→L

605/ prM ACC→GCC 171 T→A

932/ prM ACA→GCA 280 T→A

1373/ E GAC→AAC 427 D→N

1394/ E GAA→GAC 434 E→D

1424/ E TCC→CCC 444 S→P

1439/ E GCT→GTT 449 A→V

1607/ E AAA→GAG 505 K→E

1742/ E ACC→AAC 550 T→N

1805/ E AAA→GAA 571 K→E

2105/ E AGG→AAG 671 R→K

2366/ E ATT→GTT 758 I →V

1 Positions of nucleotides in the DEN3 genome; numbering is according to (Osatomi et al, Virology 176:643-647, 1990)

2 Positions of amino acid residues in the DEN3 polyprotein; numbering is according to (Osatomi et al, Virology 176:643-647, 1990).

Table 34

#1) YFM5'3'(4.56)+ (GTGAGCATTGAGAAAGCGCCACGCTTC)(SEQ ID NO:43)

#2) YF0.481- (TCCACCCGTCATCAACAGCATTCCCAAAATTAG)(SEQ ID NO:44)

#3) 4DE 0.432+

(GAATGCTGTTGATGACGGGTGAATTTCACCTGTCAACAAGAGACGG) (SEQ ID NO:45)

#4) 4DEE 1.095-

(GCCGCTAGCGGTTCGAAATAGAGCCACTTCCTTGGCTGT)(SEQ ID

NO:46)

#5) 4DE 1.102+ (GCCGCTAGCTTCGAACCTATTGCATTGAAGCCTCGATATC)(SEQ ID

NO:47)

#6) 4DE 2.409- (GCCGCCGGCGCCAACTGTGAAACCTAGAAACACAG)(SEQ ID NO:48)

#7) sp6YFa+(ATTTAGGTGACACTATAGAGTAAATCCTGTGTGCTAATT)(SE Q ID NO:49)

Other Embodiments

Other embodiments are within the following claims. For example, the prM-E protein genes of other flaviviruses of medical importance can be inserted into the yellow fever vaccine virus backbone to produce vaccines against other medically important flaviviruses (see, e.g., Monath et al, "Flaviviruses," In Virology, Fields (ed.), Raven-Lippincott, New York, 1995, Volume I, 961-1034). Examples of additional flaviviruses from which genes to be inserted into the chimeric vectors of the invention can be obtained include, e.g., Kunjin, Central European Encephalitis, Russian Spring-Summer Encephalitis, Powassan, Kyasanur Forest

Disease, and Omsk Hemorrhagic Fever viruses. In addition, genes from even more distantly related viruses can be inserted into the yellow fever vaccine virus to construct novel vaccines.

Vaccine Production and Use The vaccines of the invention are administered in amounts, and by using methods, that can readily be determined by persons of ordinary skill in this art. The vaccines can be administered and formulated, for example, in the same manner as the yellow fever 17D vaccine, e.g., as a clarified suspension of infected chicken embryo tissue, or a fluid harvested from cell cultures infected with the chimeric yellow fever virus. Thus, the live, attenuated chimeric virus is formulated as a sterile aqueous solution containing between 100 and 1,000,000 infectious units (e.g., plaque- forming units or tissue culture infectious doses) in a dose volume of 0.1 to 1.0 ml, to be administered by, for example, intramuscular, subcutaneous, or intradermal routes. In addition, because flaviviruses may be capable of infecting the human host via the mucosal routes, such as the oral route (Gresikova et al, "Tick-borne Encephalitis," In The Arboviruses, Ecology and Epidemiology, Monath (ed.), CRC Press, Boca Raton, Florida, 1988, Volume IV, 177-203), the vaccine virus can be administered by a mucosal route to achieve a protective immune response. The vaccine can be administered as a primary prophylactic agent in adults or children at risk of flavivirus infection. The vaccines can also be used as secondary agents for treating flavivirus-infected patients by stimulating an immune response against the flavivirus.

It may be desirable to use the yellow fever vaccine vector system for immunizing a host against one virus (for example, Japanese

Encephalitis virus) and to later reimmunize the same individual against a second or third virus using a different chimeric construct. A significant advantage of the chimeric yellow fever system is that the vector will not elicit strong immunity to itself. Nor will prior immunity to yellow fever virus preclude the use of the chimeric vaccine as a vector for heterologous gene expression. These advantages are due to the removal of the portion of the yellow fever vaccine E gene that encodes neutralizing (protective) antigens to yellow fever, and replacement with another, heterologous gene that does not provide cross-protection against yellow fever. Although YF 17D virus nonstructural proteins may play a role in protection, for example, by eliciting antibodies against NSl, which is involved in complement-dependent antibody mediated lysis of infected cells (Schlesinger et al, J. Immunology 135:2805-2809, 1985), or by inducing cytotoxic T cell responses to NS3 or other proteins of the virus, it is unlikely that these responses will abrogate the ability of a live virus vaccine to stimulate neutralizing antibodies. This is supported by the facts that (1) individuals who have been previously infected with JE virus respond to vaccination with YF 17D similarly to persons without previous JE infection, and (2) individuals who have previously received the YF 17D vaccine respond to revaccination with a rise in neutralizing antibody titers (Sweet et al, Am. J. Trop. Med. Hyg. 11 :562-569, 1962). Thus, the chimeric vector can be used in populations that are immune to yellow fever because of prior natural infection or vaccination, and can be used repeatedly, or to immunize simultaneously or sequentially with several different constructs, including yellow fever chimeras with inserts from, for example, Japanese Encephalitis, St. Louis Encephalitis, or West Nile viruses. For vaccine applications, adjuvants that are known to those skilled in the art can be used. Adjuvants that can be used to enhance the immunogenicity of the chimeric vaccines include, for example, liposomal formulations, synthetic adjuvants, such as saponins (e.g., QS21), muramyl dipeptide, monophosphoryl lipid A, or polyphosphazine. Although these adjuvants are typically used to enhance immune responses to inactivated vaccines, they can also be used with live vaccines. In the case of a chimeric vaccine delivered via a mucosal route, for example, orally, mucosal adjuvants such as the heat-labile toxin of E. coli (LT) or mutant derivations of LT are useful adjuvants. In addition, genes encoding cytokines that have adjuvant activities can be inserted into the yellow fever vectors. Thus, genes encoding cytokines, such as γ-interferon, GM- CSF, IL-2, IL-12, IL-13, or IL-5, can be inserted together with heterologous flavivirus genes to produce a vaccine that results in enhanced immune responses, or to modulate immunity directed more specifically towards cellular, humoral, or mucosal responses.

In addition to vaccine applications, as one skilled in the art can readily understand, the vectors of the invention can be used in gene therapy methods to introduce therapeutic gene products into a patient's cells and in cancer therapy. In these methods, genes encoding therapeutic gene products are inserted into the vectors, for example, in place of the gene encoding the prM-E protein. Yellow fever 17D virus targets cells of the lymphoid and reticuloendothelial systems, including precursors in bone marrow, monocytes, macrophages, T cells, and B cells (Monath, "Pathobiology of the Flaviviruses," pp. 375-425, in Schlesinger & Schlesinger (Eds.), The Togaviridae and Flaviviridae, Plenum Press, New York 1986). The yellow fever 17D virus thus naturally targets cells involved in antigen presentation and immune stimulation. Replication of the virus in these cells, with high-level expression of heterologous genes, makes yellow fever 17D vaccine virus an ideal vector for gene therapy or immunotherapy against cancers of the lymphoreticular system and leukemias, for example. Additional advantages are that (1) the flavivirus genome does not integrate into host cell DNA, (2) yellow fever virus appears to persist in the host for prolonged periods, and (3) that heterologous genes can be inserted at the 3' end of the yellow fever vector, as described above in the strategy for producing a Hepatitis C vaccine. Yellow fever 17D virus can be used as a vector carrying tumor antigens for induction of immune responses for cancer immunotherapy. As a second application, yellow fever 17D can be used to target lymphoreticular tumors and express heterologous genes that have anti- tumor effects, including cytokines, such as TNF-alpha. As a third application, yellow fever 17D can be used to target heterologous genes to bone marrow to direct expression of bioactive molecules required to treat hematologic diseases, such as, for example, neutropenia; an example of a bioactive molecule that can be used in such an application is GM-CSF, but other appropriate bioactive molecules can be selected by those skilled in the art.

An additional advantage of the yellow fever vector system is that flaviviruses replicate in the cytoplasm of cells, so that the virus replication strategy does not involve integration of the viral genome into the host cell (Chambers et al, "Flavivirus Genome Organization, Expression, and Replication," In Annual Review of Microbiology 44:649-688, 1990), providing an important safety measure.

All references cited herein are incorporated by reference in their entirety.

What is claimed is:

Claims

1. A chimeric live, infectious, attenuated virus, comprising: a yellow fever virus in which the nucleotide sequence encoding a prM-E protein is either deleted, truncated, or mutated so that functional yellow fever virus prM-E protein is not expressed, and integrated into the genome of said yellow fever virus, a nucleotide sequence encoding a prM-E protein of a second, different flavivirus, so that said prM-E protein of said second flavivirus is expressed.

2. The chimeric virus of claim 1, wherein said second flavivirus is selected from the group consisting of a Japanese Encephalitis (JE) virus, a Dengue virus selected from the group consisting of Dengue types 1, 2, 3, and 4, a Murray Valley Encephalitis virus, a St. Louis Encephalitis virus, a West Nile virus, a Tick-borne Encephalitis virus (i.e., a Central European Encephalitis virus or a Russian Spring-Summer Encephalitis virus), a Hepatitis C virus, a Kunjin virus, a Powassan virus, a Kyasanur Forest Disease virus, and an Omsk Hemorrhagic Fever virus.

3. The chimeric virus of claim 1, wherein said second flavivirus is a Dengue virus, and nucleotide sequences derived from said Dengue virus are derived from two or more different Dengue strains.

4. The chimeric virus of claim 1, wherein the nucleotide sequence encoding the prM-E protein of said second, different flavivirus replaces the nucleotide sequence encoding the prM-E protein of said yellow fever virus or comprises a mutation that prevents prM cleavage to produce M protein.

5. The chimeric virus of claim 1, wherein the prM signal of said chimeric virus is that of yellow fever virus.

6. The chimeric virus of claim 1, wherein the NS2B-NS3 protease recognition site and the signal sequences and cleavage sites at the C/prM and E/NSl junctions are maintained in construction of said chimeric flavivirus.

7. Use of a chimeric, live, infectious attenuated virus in the preparation of a medicament for preventing or treating flavivirus infection in a patient, said virus comprising: a yellow fever virus in which the nucleotide sequence encoding a prM-E protein is either deleted, truncated, or mutated so that functional yellow fever virus prM-E protein is not expressed, and integrated into the genome of said yellow fever virus, a nucleotide sequence encoding a prM-E protein of a second, different flavivirus, so that said prM-E protein of said second flavivirus is expressed.

8. The use of claim 7, wherein said second flavivirus is selected from the group consisting of a Japanese Encephalitis (JE) virus, a Dengue virus selected from the group consisting of Dengue types 1, 2, 3, and 4, a Murray Valley Encephalitis virus, a St. Louis Encephalitis virus, a West Nile virus, a Tick-borne Encephalitis virus (i.e., a Central European Encephalitis virus or a Russian Spring-Summer Encephalitis virus), a Hepatitis C virus, a Kunjin virus, a Powassan virus, a Kyasanur Forest Disease virus, and an Omsk Hemorrhagic Fever virus.

9. The use of claim 7, wherein second flavivirus is a Dengue virus, and nucleotide sequences derived from said Dengue virus are derived from two or more different Dengue strains.

10. The use of claim 7, wherein the nucleotide sequence encoding the prM-E protein of said second, different flavivirus replaces the nucleotide sequence encoding the prM-E protein of said yellow fever virus or comprises a mutation that prevents prM cleavage to produce M protein.

11. The use of claim 7, wherein the prM signal of said chimeric virus is that of yellow fever virus.

12. The use of claim 7, wherein the NS2B-NS3 protease recognition site and the signal sequences and cleavage sites at the C/prM and E/NSl junctions are maintained in construction of said chimeric flavivirus.

13. A nucleic acid molecule encoding a chimeric live, infectious, attenuated virus comprising: a yellow fever virus in which the nucleotide sequence encoding a prM-E protein is either deleted, truncated, or mutated so that functional yellow fever virus prM-E protein is not expressed, and integrated into the genome of said yellow fever virus, a nucleotide sequence encoding a prM-E protein of a second, different flavivirus, so that said prM-E protein of said second flavivirus is expressed.

14. The nucleic acid molecule of claim 13, wherein said second flavivirus is selected from the group consisting of a Japanese Encephalitis (JE) virus, a Dengue virus selected from the group consisting of Dengue types 1, 2, 3, and 4, a Murray Valley Encephalitis virus, a St. Louis Encephalitis virus, a West Nile virus, a Tick-borne Encephalitis virus (i.e., a Central European Encephalitis virus or a Russian Spring-Summer Encephalitis virus), a Hepatitis C virus, a Kunjin virus, a Powassan virus, a Kyasanur Forest Disease virus, and an Omsk Hemorrhagic Fever virus.

15. The nucleic acid molecule of claim 13, wherein second flavivirus is a Dengue virus, and nucleotide sequences derived from said Dengue virus are derived from two or more different Dengue strains.

16. The nucleic acid molecule of claim 13, wherein the nucleotide sequence encoding the prM-E protein of said second, different flavivirus replaces the nucleotide sequence encoding the prM-E protein of said yellow fever virus or comprises a mutation that prevents prM cleavage to produce M protein.

17. The nucleic acid molecule of claim 13, wherein the prM signal of said chimeric virus is that of yellow fever virus.

18. The nucleic acid molecule of claim 13, wherein NS2B-NS3 protease recognition site and the signal sequences and cleavage sites at the C/prM and E/NSl junctions are maintained in construction of said chimeric flavivirus.

19. Use of a yellow fever virus vector comprising a gene encoding a gene product in the preparation of a medicament for producing said gene product in a cell of a patient.

20. The use of claim 19, wherein said cell is a cell of the lymphoid system or the reticuloendothelial system, or a precursor thereof.

21. The use of claim 19, wherein said patient has cancer, such as leukemia.

22. The use of claim 19, wherein said gene product is a tumor antigen or a cytokine.