Patent Application
20110287044
The invention relates animal diseases. More specifically, the invention relates to a recombinant classical swine fever virus (CSFV) comprising a modified E2 protein. The invention further relates to a vaccine comprising said recombinant CSFV that allows differentiation of an infected animal from a vaccinated animal, and methods for generating said recombinant CSFV.
Classical swine fever virus (CSFV) is an enveloped, positive-strand RNA virus which, together with bovine viral diarrhoea virus (BVDV) and border disease virus (BDV), comprises the Pestivirus genus of the Flaviviridae family (Pringle, 1998. Arch Virol 143: 203-10). Introduction of CSFV in herds of domesticated pigs can result in huge economic losses (Terpstra and de Smit. 2000. Vet Microbiol 77: 3-15). Vaccination with a CSFV virus that has been attenuated by repeated passage in rabbits and cell culture, the so-called “Chinese” or “C”-strains, results in swift and livelong immunity against virulent CSFV. The C-strain virus is successfully used worldwide and is often referred to as the most effective veterinary vaccine ever produced. This vaccine, however, does not allow the serological differentiation between infected and vaccinated animals (DIVA). This is a major disadvantage, since the inability to detect CSF-infected animals in a vaccinated population can impose severe trade restrictions.
Diagnosis of CSF in the field can be performed by ELISAs that detect antibodies directed against either the structural glycoprotein Erns or E2. Several candidate vaccines have been developed that can fulfill the DIVA criterion when accompanied with the appropriate ELISA, varying from subunit vaccines (Hulst et al., 1993. J Virol 67: 5435-42) to live-attenuated viruses (van Gennip et al., 2000. Vaccine 19: 447-59; van Zijl et al., 1991. J Virol 65: 2761-2765) and replicon-based vaccines (van Gennip et al., 2002. Vaccine 20: 1544-56; Widjojoatmodjo et al., 2000. J Virol 74: 2973-2980).
The commercially available DIVA vaccine against CSF is based on baculovirus-produced E2, which is accompanied by a serological test that detects antibodies directed against Erns (Van Aarle, 2003. Dev Biol Stand Basel 114: 193-200). Although this vaccine provides protection against CSF, it is less efficacious than the C-strain vaccine with respect to both the onset and duration of immunity (van Oirschot, 2003. Vet Microbiol 96: 367-84). More importantly, Erns ELISAs that accompany this vaccine, also detect other members of the pestivirus genus (i.e. BVDV and BDV). Their use is, therefore, not recommended in regions were these viruses circulate (2003/265/EC; SANCO/10809/2003).
Furthermore, the sensitivity of Erns ELISAs was previously found to be insufficient to diagnose individual animals and should therefore only be used on a herd basis with sufficiently large numbers of animals (Blome et al., 2006. Rev Sci Tech 25: 1025-38; Floegel-Niesmann, 2003. Dev Biol (114: 185-91). This explains why, in general, E2 ELISAs are greatly preferred over Erns ELISAs in accompanying a DIVA vaccine.
The E2 protein contains two major antigenic domains, namely the B/C domain and the A domain (van Rijn et al., 1993. J Gen Virol 74: 2053-60; Wensvoort, 1989. J Gen Virol 70: 2865-76). Domain A, which is located between amino acids 766 and 866 of the CSFV polyprotein, is divided in subdomains A1, A2 and A3 (Wensvoort, 1989. J Gen Virol 70: 2865-76). Despite the fact that the A1 domain is a dominant target for neutralizing antibodies, it has been conserved throughout evolution. In fact, its sequence conservation and immunodominance have rendered it the dominant target in E2 ELISAs.
Although outbreaks of CSF are currently controlled by quarantine restrictions and slaughtering of suspected herds, there is an urgent need for the implementation of more humane and more economical intervention strategies to control future CSF outbreaks. Therefore, there is an urgent need for a DIVA vaccine that is accompanied by a robust and CSFV-specific ELISA.
The present invention provides a recombinant classical swine fever virus (CSFV), comprising a deletion of at least one amino acid in the “TAVSPTTLR” domain of the E2 protein, corresponding to position 829 to 837 of a parental CSFV polyprotein.
With polyprotein is meant the about 4000 amino acid hypothetical polyprotein that is formed upon translation of the viral RNA. Said polyprotein is processed to form the final cleavage products Npro-C-Erns-E1-E2-p7-N52-N53-NS4A-NS4B-NS5A-NS5B.
The CSFV E2 protein contains a recently identified epitope that comprises the amino acid sequence TAVSPTTLR (residues 829-837 of the CSFV polyprotein; using single letter code for amino acids) (Lin et al., 2000. J Virol 74: 11619-25). This epitope shares all characteristic features of the A1 domain, being immunodominant, evolutionarily conserved, specific for CSFV and a target for neutralizing antibodies. A comparison of the sequences around the TAVSPTTLR domain among E2 proteins from different strains of pestiviruses indicates that the sequence TAVSPTTLR is strongly conserved among strains of CSFV and is highly variable among strains of BVDV and BDV (Lin et al., 2000. J Virol 74: 11619-25).
Antibodies, especially monoclonal antibodies, used in E2-specific ELISAs that recognize the TAVSPTTLR domain do not cross react with other members of the pestivirus genus and can therefore be used in regions were these other viruses circulate. Said antibodies will not recognize a recombinant CSFV according to the invention, comprising a deletion in said domain of the E2 protein.
Thus, said recombinant virus enables to differentiate animals that are infected with the recombinant virus from animals that are infected with wild-type CSFV and from animals that are not infected and/or that are not vaccinated. Furthermore, the use of said recombinant virus will allow the use of a peptide-based diagnostic test to discriminate between these animals.
The present invention provides a recombinant classical swine fever virus (CSFV), comprising a deletion of at least one amino acid in the “TAVSPTTLR” domain of the CSFV, or an equivalent thereof. An equivalent domain is a domain within the E2 region in which one amino acid is substituted for another amino acid, such as an amino acid that belongs to the same group, i.e. an aromatic amino acid that is replace by another, aromatic amino acid, or an aliphatic amino acid that is replace by another aliphatic amino acid.
Said parental genome preferably comprises a substantially complete viral genome derived from a CSFV strain, preferably a naturally occurring or recombinant attenuated CSFV strain. The term parental genome comprises a nucleic acid molecule, such as a RNA molecule and/or a cDNA copy thereof.
With the term “deletion of at least one amino acid”, as used in the description, is meant the removal of at least one amino acid. The term “deletion” does not cover a removal of at least one amino acid, followed by insertion of another at least one amino acid at the same position. Therefore, the term “deletion” as used herein does not cover the substitution of an amino acid for another amino acid.
Said recombinant cDNA molecule preferably comprises a substantially complete recombinant classical swine fever virus (CSFV) genome, whereby said genome encodes an E2 protein that comprises a deletion of at least one amino acid in the conserved “TAVSPTTLR” domain corresponding to position 829 to 837 of the CSFV polyprotein. The term “substantially complete” indicates that said a virus generated by said genome is able to infect a suitable cell or cell line, and can be reproduced in said suitable cell or cell line. A “substantially complete” viral genome preferably is a replication-competent genome. More preferred is an infectious, replication-competent and packaging-competent viral genome, it is furthermore preferred that also the recombinant classical swine fever virus (CSFV) comprises a “substantially complete” viral genome, preferably a replication-competent genome, or, more preferred, an infectious, replication-competent and packaging-competent viral genome.
In a preferred embodiment, a deletion of at least one amino acid in the TAVSPTTLR domain of a recombinant CSFV according to the invention comprises a deletion of the central proline of said. “TAVSPTTLR,” domain.
Proline is unique among the 20 common amino acids in having the side-chain cyclized onto the backbone nitrogen atom. This restricts the conformation of said proline and also of the residue preceding the proline. Furthermore, proline can act as a conformational ‘switch’, allowing parts of proteins to adopt alternative conformations. Alteration of the central proline in the TAVSPTTLR domain, therefore, not only changes the primary sequence, but also changes the conformation of the immunogenic TAVSPTTLR domain. An E2 protein that comprises a deletion of the central proline of the TAVSPTTLR domain will not be recognized by antibodies that recognize said domain within the E2 protein.
In a further preferred embodiment, an altered E2 protein of a recombinant CSFV according to the invention comprises a deletion of at least two amino acids from said. TAVSPTTLR domain. A deletion of two amino acids comprises deletion of TA, AV, VS, SP, PT, TT, TL and LR, respectively, from the TAVSPTTLR domain of E2. A most preferred deletion of two amino acids is a deletion of PT at position 833-834 in the TAVSPTTLR domain.
In a further preferred, embodiment, an altered E2 protein of a recombinant CSFV according to the invention comprises a deletion of at least three amino acids from said TAVSPTTLR domain, such as three amino acids, four amino acids, five amino acids, six amino acids, seven amino acids, eight amino acids or nine amino acids. A further preferred E2 protein comprises a deletion of the amino acid sequences VSP, SPT, AVSP, or SPTTL.
The invention further provides a recombinant CSFV according to the invention, comprising at least one further alteration of the parental CSFV genome. Viruses comprising a deletion in said conserved TAVSPTTLR domain were less efficient in infecting cells, compared to the parental virus, and/or were less efficiently replicated in infected cells, resulting in a lower titer as compared to the parental virus. Continued passaging of cells infected with a virus comprising a deletion in the conserved TAVSPTTLR domain resulted in a rescued virus that more efficiently infected and/or replicated in cells, resulting in titers that were comparable to titers of the parental virus. The rescued virus had introduced one or more further alterations in the parental genome that compensated for the loss of fitness of the virus with a deletion in the conserved TAVSPTTLR domain.
In one embodiment, said at least one further alteration is a silent mutation that changes the genome of the virus but does not result in an alteration of an amino acid. Said silent mutation is present in a non-coding part of the viral genome, or a coding part of the viral genome, such as in a part of the viral genome encoding Npro, C, Erns, E1, and/or E2, and/or in a part encoding non-structural proteins. A silent mutation may contribute to the fitness recovery of the altered viruses at some point during the generation of these viruses. Without being bound by theory, a silent mutation, for example, results in an improved stability of the viral genome and/or an improved replication because said silent mutation alters a conformation of the viral genomic nucleic acid. Furthermore, said silent mutation may result in improved codon usage. A preferred silent mutation is provided by a U to C alteration at position 1549 in the ERNS gene.
In a further embodiment, said at least one further alteration is in a region encoding the E1 protein which is known to assemble into disulphide-linked heterodimers with the E2 protein. A deletion in the conserved TAVSPTTLR domain of E2 is thus, at least partially, compensated by an alteration of at least one amino acid in the E1 protein. In a further embodiment, said at least one further alteration is in a region encoding the Npro, C, Erns or E2 protein. Said at least one further alteration preferably comprises at least two alterations in different regions selected from the regions encoding the Npro, C, Erns E1, and/or E2 protein, or within one region encoding the Npro, C, Erns, E1, and/or E2 protein. Said at least two further alterations preferably comprise at least one silent mutation. A preferred silent mutation is provided by a U to C alteration at position 1549 in the ERNS gene.
In a preferred embodiment, said at least one further alteration results in the introduction of an additional N-linked glycosylation site in the E2 protein. An N-linked glycosylation site at an additional position in the E2 protein apparently compensates, either directly or by virtue of its function as an anchor site for a carbohydrate moiety, for the loss of fitness imposed by said deletion in the TAVSPTTLR epitope. In a preferred embodiment, said N-linked glycosylation site is introduced by alteration of the LFDGTNP domain from amino acid position 772-778, such as, for example, an alteration of D to N at position 774. Alternatively, or in addition, an N-linked glycosylation site is introduced by alteration of A to N at position 830 in the E2 protein.
Said further alteration preferably alters a codon at positions 1547-1549, encoding V392 within the Erns protein; a codon at positions 2273-2275, encoding E634 within the E1 protein; a codon at positions at position 2693-2695 encoding D774 within the E2 protein; and/or a codon at positions at position 2864-2866 encoding V831 within the E2 protein. Said alteration of a codon at the indicated positions comprises either a silent mutation, or comprises an alteration of the encoded amino acid. A preferred at least one further alteration comprises a substitution of valine (V) at position 831 for glycine (G).
In another preferred embodiment, said at least one further alteration of the parental genome results in a substitution of S to F at position 789 and/or a substitution of A to T at position 445. The serine at position 789 and the alanine at position 445 are present in all C-strain viruses and related lapinized CSFV strains, whereas phenylalanine at position 789 and threonine at position 445 are conserved in virulent CSFV strains. Although the history of the C-strain virus is not well documented, it is clear that the virus was attenuated by passage in rabbits hundreds of times, for which an S at position 789 and an A at position 445 may be beneficial. The alteration of S to F at position 789 and the alteration of A to T at position 445 are likely beneficial for propagation of a virus comprising a deletion in the TAVSPTTLR epitope, whereas it will be close to neutral in a virus comprising the TAVSPTTLR epitope.
In yet a further preferred embodiment, the invention provides a recombinant CSFV comprising a deletion of P at position 833 and a silent alteration. A preferred silent alteration is a U to C alteration at position 1549 in the ERNS gene. A further preferred recombinant CSFV comprises a deletion of P at position 833 and an alteration of S to F at position 789 and/or an alteration of A to T at position 445; a deletion of P at position 833, an alteration of D to N at position 774, a substitution of V to G at position 831, and a deletion of T at position 834, either or not in addition to a U to C alteration at position 1549. A most preferred recombinant CSFV comprises a deletion of proline and threonine at position 833 and 834, respectively, of the “TAVSPTTLR” domain of the E2 protein, a U to C alteration at position 1549, a substitution of a glutamic acid (E) for aspartic acid (D) at position 634, a substitution of aspartic acid (D) for asparagine (N) at position 774 in the E2 protein, and a substitution of valine (V) for glycine (G) at position 831.
In a further embodiment, the invention provides a recombinant classical swine fever virus (CSFV), comprising an alteration of at least one amino acid in a “TAVSPTTLR” domain from position 829 to 837 in E2 of a parental CSFV polyprotein, whereby said alteration comprises a substitution of the central proline in the TAVSPTTLR domain to asparagine. Said substitution was found to minimize or even inhibit binding of E2-specific monoclonal antibodies to the altered E2 protein.
Apart from changing the primary amino acid sequence, the introduction of said asparagine results in the introduction of an N-linked glycosylation site comprising the glycosylation consensus sequence [N-x-S/T], wherein x denotes any amino acid except for P or D (Kornfeld and Kornfeld, 1985. Annu Rev Biochem 54: 631-64). N-glycosylation of viral proteins has been implicated in immunogenicity, whereby the introduction of an N-linked glycosylation can limit both cellular and antibody response to a viral protein. Said substitution can thus be used to generate a recombinant virus that allows differentiation of animals infected with a wild type virus from animals infected with said recombinant virus.
In one embodiment, a recombinant CSFV, comprising a substitution of the central proline in the TAVSPTTLR domain to asparagine, comprises at least one further alteration in the genome. A preferred at least one further alteration results in an additional N-glycosylation site in Npro, C, Erns, E1, and/or E2, and/or in a part encoding non-structural proteins. A preferred additional N-glycosylation is present within E2, such as, for example, within the TAVSPTTLR domain or within a LFDGTNP epitope of the E2 protein.
A further preferred at least one further alteration is a silent mutation present in a non-coding part of the viral genome, or in a coding part of the viral genome, such as in a part of the viral genome encoding Npro, C, Erns, E1, and/or E2, and/or in a part encoding non-structural proteins. A preferred silent mutation is provided by a U to C alteration at position 1549 in the ERNS gene.
Yet a further preferred at least one further alteration comprises an alteration of at least one amino acid in Npro, C, Erns, E1, and/or E2, and/or in a part encoding non-structural proteins. Said at least one further alteration is preferably selected from a silent alteration, for example a U to C alteration at position 1549 in the ERNS gene, a alteration of S to F at position 789, an alteration of A to T at position 445, an alteration of D to N at position 774, an alteration of V to G at position 831, and/or a deletion of T at position 834. It is even more preferred that said at least one further alteration comprises at least two further alterations, selected from a silent mutation, an additional N-glycosylation site, and/or an amino acid alteration, in addition to substitution of the central proline in the TAVSPTTLR domain to asparagine. Said at least two further alterations are present in the same protein or in different proteins selected from Npro, C, Erns, E1, E2, and non-structural protein. It is preferred that said at least two further alterations are present in the same protein or in different proteins selected from Npro, C, Erns, E1, and E2.
In yet a further embodiment, the invention provides a recombinant classical swine fever virus (CSFV), comprising an addition of at least one amino acid in a “TAVSPTTLR” domain from position 829 to 837 in E2 encoded by a parental CSFV genome. Said insertion addition will minimize or even inhibit binding of E2-specific monoclonal antibodies to the altered E2 protein. An insertion of at least one amino acid in a “TAVSPTTLR” domain is preferably combined with at least one further alteration selected from a silent mutation, an additional N-glycosylation site, and/or an amino acid alteration, of a combination thereof, in Npro, C, Erns, E1, E2, and non-structural protein. Said at least one further alteration is preferably selected from a silent alteration, for example a U to C alteration at position 1549 in the ERNS gene, an alteration of S to F at position 789, an alteration of A to T at position 445, an alteration of D to N at position 774, and/or an alteration of V to G at position 831
The parental genome of a recombinant CSFV according to the invention preferably is the genome of an attenuated CSFV strain.
Attenuated CSFV strains can be generated by mutation of the Erns gene encoding a protein with RNase activity (Mayer et al., 2003. Virus Res. 98: 105-16); by deletion of Npro from CSFV virulent strains (Mayer et at. 2004. Vaccine 22: 31.7-328); by combining mutations in Erns, and E2 (van Gennip et al. 2004. J. Virol. 78: 8812-8823); by mutation of the E1 gene (Risatti et al. 2005. Virology 343: 116-127); and by mutation of the E2 gene (Risatti et al., 2007. Virology 364: 371-82).
A preferred attenuated. CSFV strain comprises an insertion in the 3′ terminal non-coding region. For example, an insertion of 12 nucleotides in the 3′ untranslated region leads to attenuation of CSFV (Wang et al., 2008. Virology 374: 390-8). Said insertion preferably comprises a sequence of 12 nucleotides consisting of 5′-CUUUUUUCUUUU.
A most preferred parental genome is the genome of C (Chinese)-strain. Even more preferred as a parental genome is the genome of the Cedipest C-strain, which is a C-strain virus that was adapted to suspension cultures of the swine kidney cell line SK6 (Terpstra et al., 1990. Dtsch Tierarztl Wochenschr. 97:77-9). Pigs inoculated with 400-600 TCID50 of the Cedipest strain are fully protected against challenge with greater than 100 pig LD50 of a virulent strain of CSFV at 7 days and at 6 month post vaccination. Therefore, the invention also provides a cDNA molecule comprising a copy of a recombinant classical swine fever virus (CSFV) genome encoding an altered “TAVSPTTLR” domain from position 829 to 837 in E2. Said cDNA molecule preferably comprises a substantially complete recombinant parental CSFV viral genome according to the invention, whereby said parental genome is derived from the C (Chinese)-strain, or, more preferred, from the Cedipest strain. A preferred cDNA molecule comprises a copy of a recombinant CSFV genome encoding an E2 protein comprising a deletion of at least one amino acid in a “TAVSPTTLR” domain. The invention also provides an antibody, preferably a monoclonal antibody, that specifically recognizes said altered “TAVSPTTLR” domain. The invention further relates to a construct encoding a protein comprising said altered, “TAVSPTTLR” domain. The invention further relates to a protein comprising said altered, “TAVSPTTLR” domain, and to a peptide comprising said altered, “TAVSPTTLR” domain, and to the use of said protein or peptide in an immuno-assay such as, for example, an ELISA.
In a further aspect, the invention provides a live CSF vaccine comprising a recombinant CSFV according to the invention.
A vaccine, or immuno active composition, comprising a recombinant CSFV according to the invention combines the rapid induction and length of protection induced by a live, attenuated virus with the possibility of differentiating between vaccinated animals and animals that are infected with wild type virus due to a difference, such as a deletion, in the conserved TAVSPTTLR domain of E2, which is an important immunodominant epitope of the E2 protein.
Several vaccines have been developed that lead to a marked reduction in the levels of circulating virus and a concomitant reduction of clinical cases. For example, a subunit vaccine against CSFV has been developed that is based on the envelope glycoprotein E2. This subunit vaccine allows discrimination between vaccinated and infected pigs on the basis of the detection of antibodies against ERNS K (Van Aarle, 2003. Dev Biol Stand (Basel) 114: 193-200). However, the vaccine is inferior to the C-strain vaccine with respect to both the onset and duration of immunity. More importantly. Erns ELISAs that accompany this vaccine, also detect other members of the pestivirus genus (i.e. BVDV and BDV). Their use is, therefore, not recommended in regions were these viruses circulate (2003/265/EC; SANCO/10809/2003).
A further vaccine is provided by a so-called virus replicon particle (VRP). A VRP particle contains a mutant genomic RNA which is able to replicate and to express the encoded viral proteins but which does not contain the complete information required for particle formation due to deletions in at least one of the genes encoding a viral structural protein. These virus particles are non-transmissible and, therefore, fulfill one of the requirements for a safe vaccine. However, VRP with a partial or complete deletion of the E2 gene induced only partial protection against lethal challenge with highly virulent CSFV (Maurer et al., 2005. Vaccine 23: 3318-28).
The invention further provides a method of protecting an animal against CSF, comprising administering to said animal an effective amount of a vaccine according to the invention.
An effective amount is defined as an amount of said vaccine that will induce an immunological response in the individual to which it is administered resulting in the development in the individual of a secretory, cellular and/or antibody-mediated immune response to the vaccine. Said secretory, cellular and/or antibody-mediated immune response to the vaccine is also effective against a challenge with a virulent CSFV strain.
Said effective amount is preferably administered orally or oronasally or intramuscularly. An immunogenic composition against classical swine fever, comprising a recombinant CSFV according to the invention, is preferably administered together with a pharmaceutically acceptable carrier.
A further method of protecting an animal against CSF comprises the provision an effective amount of a vaccine according to the invention as a bait vaccine, especially to protect wild animals such as wild boars.
Efficacious vaccines reduce or prevent clinical signs by preventing virus replication and/or reducing virus transmission. The term DIVA (differentiating infected from vaccinated, animals) is used for vaccines and their companion diagnostic tests which are based on mutants of wild-type viruses, in conjunction with a differentiating diagnostic test. This system makes possible the mass vaccination of a susceptible animal population without compromising the serological identification of convalescent animals.
The invention furthermore provides a method of differentiating animals infected with CSFV from non-infected animals or from animals vaccinated with the CSF vaccine according to the invention, comprising analyzing serum of an animal in a serological test. It is preferred that said CSFV vaccine comprises an alteration of the TAVSPTTLR domain according to the invention, and, more preferred, at least one further alteration in the E2 protein such as, for example, an alteration in LFDGTNP domain from amino acid position 772-778.
Said serological test preferably comprises one or more antibodies such as monoclonal antibodies that recognize an E2 protein comprising an intact TAVSPTTLR domain, and one or more antibodies such as monoclonal antibodies that recognize a further epitope encoded by CSFV, such as a further epitope within the E2 protein for example the LFDGTNP epitope.
In a preferred embodiment, said antibody enzyme-linked immunosorbent assays (ELISAs) allow diagnosis of CSF in live pigs. Preferred ELISAs are sandwich type ELISAs. A preferred ELISA is a competition ELISA based on E2, such as, for example, Ceditest 2.0 ELISA. By pre-incubating serum of an animal with a mutant E2 protein comprising an altered TAVSPTTLR domain, either or not in combination with an altered LFDGTNP domain, it is possible to deplete said serum for antibodies that are directed against said altered TAVSPTTLR domain and altered LFDGTNP domain.
A most preferred ELISA is a peptide-based ELISA, wherein peptides are cross-linked to micro-well assay plates. Said cross-linking preferably is performed through an anchor protein such as, for example, poly-L-lysine. ELISAs employing cross-linked peptides are in general more sensitive when compared to ELISAs employing passively coated peptides. The technique is relatively simple to perform, does not require tissue culture facilities, is suitable for automation and can provide results within half a day. Monoclonal and/or polyclonal antibodies can be used that unambiguous differentiate between field and vaccine strains of CSFV on the one hand, and between CSFV and other pestiviruses on the other. peptide) could be used to block unspecific cross-reactivity.
Said peptide-based ELISA preferably is a liquid-phase peptide ELISA (lp-ELISA). In said lp-ELISA for detection of antibodies against CSFV, a test serum is incubated with a mixture of a modified CSFV peptide and a heterologous control peptide, for example a BVDV peptide. Said modified CSFV peptide is preferably biotinylated, while the control peptide is not biotinylated. CSFV-specific antibodies will bind to said CSFV peptide and are captured by through said modification, for example by binding to avidin or streptavidin of the CSFV peptide is biotinylated. Antibodies that are complexed to the modified CSFV peptide can be detected, for example complexed swine antibodies are detected by an anti swine-peroxidase conjugate and subsequent incubation with a suitable substrate.
An alternative test for the differentiation of an animal that is infected with field CSFV or vaccinated with a recombinant virus of the invention is provided by a fluorescent antibody test (FAT) and reverse transcriptase followed by amplification of the cDNA, for example by the polymerase chain reaction, and analysis of the sequence of the amplified. DNA. FAT is used to detect CSFV antigen in cryostat sections of tonsils, spleen, kidney, lymph nodes or distal portions of the ileum isolated from a pig that is suspected to be infected with CSFV or that is vaccinated with a vaccine according to the invention. In an alternative test, CSFV is isolated from, for example, the tonsils of an animal that is infected or suspected to be infected with field CSFV and incubated with PK-15 cells. Replicated virus is subsequently detected in the PK-15 cells using antibodies that differentiate between a field virus and a recombinant virus according to the invention.
A further preferred serological test comprises a competitive ELISA to determine if said serum comprises antibodies that inhibit binding of an antibody directed against an E2 protein comprising an intact TAVSPTTLR domain, or if said serum comprises antibodies that inhibit binding of an antibody directed against an E2 protein comprising an altered TAVSPTTLR domain.
Yet a further preferred serological test comprises one or more antibodies, preferably monoclonal antibodies, which specifically recognize an E2 protein comprising an altered TAVSPTTLR domain according to the invention. In an even more preferred embodiment, said serological test comprises one or more antibodies such as monoclonal antibodies that specifically recognize an E2 protein comprising an altered TAVSPTTLR domain according to the invention and a further altered immunogenic domain of said E2 protein. Said marker vaccine requires the reliable induction and detection (sensitive and specific) of discriminatory antibodies after vaccination of animals. The presence of said discriminatory antibodies is detected by said serological test which is preferably selected from a fluorescent antibody virus neutralization test, a neutralizing peroxidase-linked assay, and an antibody ELISA.
The invention further provides a method for isolating an infectious recombinant virus that can be used in a marker vaccine which allows differentiating infected animals from vaccinated animals, the method comprising selecting an immuno-dominant domain of a protein encoded by said virus and which is conserved in at least 90% of the viruses, preferably in at least 95% of the viruses, more preferred in at least 99%, of the viruses; introducing an alteration, preferably a deletion, in a genome region of said virus that encodes said immunodominant domain such as to generate a starting virus; contacting the altered genome with a suitable cell or cell line; passaging the suitable cell or cell line to allow propagation of a virus; and isolating an infectious virus from said suitable cell or cell line which comprises one or more further alterations in the genome that compensate a loss of fitness of the starting virus.
Said alteration in a conserved, immuno-dominant domain of a protein allows the generation of antibodies, such as monoclonal and polyclonal antibodies, that discriminate a virus that expresses a wild type immuno-dominant domain from a virus that expresses an altered immuno-dominant domain. Therefore, the invention also provides an antibody, prefer ably a monoclonal antibody, that specifically recognizes said altered, conserved, immuno-dominant domain.
The invention further relates to a construct encoding a protein comprising said altered, conserved, immuno-dominant domain. The invention further relates to a protein comprising said altered, conserved, immuno-dominant domain, and to a peptide comprising said altered, conserved, immuno-dominant domain.
A vaccine comprising an alteration in a conserved, immuno-dominant domain furthermore does not induce the generation of antibodies in a vaccinated individual that recognize an unaltered immuno-dominant domain, A diagnostic test directed at recognizing said unaltered conserved, immuno-dominant domain allows to discriminate an animal infected with a virus that expresses a wild type immuno-dominant domain from a vaccinated animal.
Said alteration of a conserved, immuno-dominant domain may result in a virus that is less efficient in infecting cells, compared to the wild type virus, and/or that is less efficiently replicated in infected cells. Passaging of an infected cell or cell line enables the introduction of one or more second site genomic alterations that rescue the virus. Said rescued virus more efficiently infects and/or replicates in cells, compared to the altered virus, or has antagonized another disadvantage that correlates with the alteration of a conserved, immuno-dominant domain. The rescued virus therefore comprises one or more further alterations in the parental genome that compensate for a loss of fitness introduced by the alteration in a conserved, immuno-dominant domain. A further advantage of a method according to the invention is that the additional one or more further alterations in the parental genome will hamper the generation of a revertant virus comprising a parental immuno-dominant domain.
In a preferred method according to the invention, said alteration in a genome region results in a deletion of at least one amino acid. A deletion of at least one amino acid, such as one amino acid, two amino acids, three amino acids or more than three amino acids, will further hamper the generation of a revertant virus comprising a parental immuno-dominant domain.
In a further preferred method according to the invention, a deletion of at least one amino acid is combined with at least one further alteration in the genome of the parental virus. Said at least one further alteration allows the passaging of an infected cell or cell line of the starting virus in the event said deletion of at least one amino acid is detrimental for replication of the virus in said cell or cell line. The introduction of said at least one further alteration restores fitness of the starting virus to some extent, allowing the starting virus to replication in said cell or cell line.
In a preferred embodiment, said virus is selected from negative stranded RNA viruses, such as rabies virus and Newcastle Disease Virus, and positive stranded RNA viruses such as Flaviviridae, preferably a pestivirus such as bovine viral diarrhoea virus, border disease virus, and CSFV. A most preferred virus for a method of the invention is CSFV.
The invention further provides an infectious recombinant virus, preferably CSFV, obtainable by the method of the invention.
The invention also provides the use of an infectious recombinant virus, preferably CSFV, obtainable by the method of the invention, such as the use in a vaccine for vaccinating one or more animals to protect said animal against infection with a virulent field virus.
Said use of the vaccine allows differentiating infected animals from not infected animals or vaccinated animals.
A major advantage of an efficacious marker vaccine is that it allows the detection of infected pigs in a vaccinated pig population, and thus offers the possibility to monitor the spread or re-introduction of CSFV in a pig population. Thus, it makes it possible to declare, with a certain level of confidence, that a vaccinated pig population is free of CSF on the basis of laboratory test results.
The latter could result in a shortening of the restriction period for the export of live pigs or pig products. The economic advantage of a rapid resumption of trade is obvious and therefore closely related to the use of any marker vaccine.
a Comparison of the amino acids from positions 772 to 791 and 823 to 842 of the recombinant viruses under study.
b Introduced amino acid substitutions are indicated in bold, adaptive mutations are indicated in bold italics.
c The TAVSPTTLR and LFDGTNP epitopes are boxed.
d++; Virus was rescued and growth was comparable to that of vFlc34 (wildtype), +; Virus with impaired growth compared to vFlc34, †; Virus was lost from the culture medium after a few passages, −; No virus was detected, nd; Not determined.
eOne of the newly introduced AAU codons was deleted during propagation of the virus.
Viruses and cells. Swine kidney cells constitutively expressing T7 RNA polymerase (SK6.T7) (van Gennip et al., 1999. J Virol Methods 78:117-28) were grown in K1000 medium supplemented with glutamine (0.3 mg/ml, Gibco), 5% fetal bovine serum and the antibiotics penicillin (100 U/ml, Gibco), streptomycin (100 U/ml, Gibco), amphotericin B (2.5 μg/ml, Gibco) and, when appropriate, with 10 mM L-histidinol dihydrochloride (Sigma). Unless indicated otherwise, virus stocks were produced by passaging the virus three to four times on SK6.T7 cells, followed by two successive freeze/thaw cycles of the infected monolayers. The latter was performed to maximize release of the C-strain virus, which is strongly cell-associated. Virus stocks were titrated on SK6.T7 cells in log 10 dilutions and were determined as TCID50/ml.
Construction of C-strain mutants. Plasmid pPRK-flc34, which contains a DNA copy of the “Cedipest” CSFV C-strain under T7-promoter control, was used as a template to introduce mutations by site-directed mutagenesis. The previously published DNA copy of the C-strain virus, named pPRKflc-133 (Moormann et al., 1996. J Virol 70:763-70) was found to lack a cytosine at the −10 position at the 3′-end of the genome. In pPRK-flc34, this error is corrected. Primers are described in Table 1. The name of the forward primer corresponds to the name of the constructed recombinant virus. Primer RV-r was used as a reverse primer for each construction. PCR amplification was performed using the Expand High-Fidelity PCR system (Roche). The PCR products were cloned into pGEM-T Easy vectors according to the instructions of the manufacturer (Promega) and sequenced using an ABI PRISM 310 genetic analyzer (Applied Biosystems). The PCR products were released from pGEM-T plasmids by digestion with ApaLI and used to replace the corresponding genome fragment of plasmid pPRc129, a pOK12-derived plasmid containing cDNA encoding the 5′ half of the C-strain virus. The pPRc129 plasmids were subsequently digested with NotI and SalI, and the resulting fragments were used to replace the corresponding segment of plasmid pPRK-flc34, resulting in plasmids pFlc-N1, pFlc-N2, pFlc-N3, pFlc-N4, pFlc-N5, pFlc-ΔP, pFlc-ΔPT, pFlc-ASP, pFlc-ΔSPT, pFlc-ΔVSP, pFlc-ΔAVSP and pFlc-ΔSPTTL.
Virus production of C-strain mutants. Plasmids containing full-length cDNA clones were linearized with XbaI and transfected into SK6.T7 cells as described previously (van Gennip et al., 1999. J Virol Methods 78:117-28). Four days after transfection, expression of viral proteins was determined by immunoperoxidase monolayer assays (IPMA) using mAb WB103, which is directed against the CSFV non-structural protein NS3 (Edwards et al., 1991. Vet Microbiol 29:101-8). Cells from another well were treated with trypsin, transferred to a 25-cm2 tissue culture flask and grown for three to four days. When necessary, cells were passaged repeatedly to support virus growth. Monolayers were freeze-thawed, centrifuged to remove cell debris and subsequently stored at −70° C. The cleared lysates were used to prepare seedlots of the vaccine candidates by infecting fresh SK6.T7 cells followed by harvest four days later. Growth of the viruses was always performed on SK6.T7 cells. Although the viruses described here also replicated normally on SK6 cells, production yields were more accurately reproducible when using SK6.T7 cells.
To study the growth kinetics of the rescued viruses, subconfluent monolayers in 25-cm2 tissue culture flasks were infected at a multiplicity of infection of 0.1. After 24, 48, 72, 96, 120, 144 and 168 hours post infection, the virus titers in cell lysates were determined. The material was freeze-thawed twice, clarified by centrifugation at 2,500×g at 4° C. and stored at −70° C. Virus titers (log TCID50/ml) were determined on SK6.T7 cells.
The E2 genes of the rescued viruses were sequenced. When appropriate, also the genes encoding the capsid (C), ERNS and E1 proteins (i.e. the structural proteins) were sequenced. To this end, viral RNA was isolated using the High Pure Total RNA isolation kit (Roche), and was used for cDNA synthesis using the Superscript First-Strand Synthesis system (Invitrogen) and a gene-specific primer. The cDNA was sequenced as described above.
Western blots. Lysates of infected SK6.T7 cells were prepared from confluent monolayers grown in 25-cm2 tissue culture flasks. To this end, cells were lysed in 0.5 ml phosphate-buffered saline (PBS) containing 1% Nonidet P40 (BDH) and protease inhibitor cocktail (Complete, Roche). Celdebris was removed by centrifugation of 4 min at 10,000×g at 4° C. Proteins were separated in 12% polyacrylamide gels (NuPAGE system, Invitrogen) and subsequently transferred to nitrocellulose paper (Protran, Schleicher and Schuell). After blocking with PBS containing 0.05% Tween- and 1% Protifar (Nutricia), the blots were incubated with C-strain specific mAb C2, which is directed against the B/C domain of E2 (Bognár and Mészáros, 1963. Acta Vet Acad Sci Hung 13:429-438) and subsequently with peroxidase-conjugated rabbit anti-mouse immunoglobulins (DAKO). Peroxidase activity was detected with the enhanced chemiluminescense system (ECL Plus, GE Healthcare) using a Storm 860 molecular imager (GE Healthcare).
Immunoperoxidase monolayer assay (IPMA). Monolayers were washed with D-PBS (Gibco), dried to the air, and frozen at −20° C. The monolayers were fixed with paraformaldehyde (4% w/v in PBS) for 15 min and subsequently washed with PBS. Peroxidase-conjugated A domain-specific mAbs b2, b3, b4, b7 (Wensvoort et al., 1989. J Gen Virol 70: 2865-76), c1, c4, c8, c11 (Bognár and Mészáros, 1963. Acta Vet Acad Sci Hung 13:429-438) and the mAb used in the Ceditest 2.0 ELISA, named mAb 18.4, were used in PBS containing 0.05% Tween-80 (PBS-T) and 5% horse serum. After incubation at 37° C. for 1 h, the plates were washed three times with PBS-T after which activity of peroxidase was detected using 3-amino-9-ethyl-carbazole (AEC, Sigma) as the substrate.
Inoculation of rabbits. New Zealand white rabbits of approximately 2 kg were housed in groups of two to four animals. Body temperatures were monitored daily, starting from three days before the inoculation until five days after. The normal body temperature of rabbits varies from 38.5 to 40.1° C. Accordingly, fever was defined as a body temperature above 40.1° C. Rabbits were inoculated via the marginal ear vein with 200 μl growth medium containing 103 TCID50 of virus. Every seven days, serum was collected. EDTA blood, to be used for virus isolation, was collected four days after the inoculation.
Virus isolation. Peripheral blood leukocytes (PBLs) were concentrated from the EDTA blood samples by ammoniumchloride precipitation (0.83% NH4Cl) as described (Terpstra and Wensvoort, 1988. Vet Microbiol 16:123-8). The PBLs were resuspended in PBS and frozen at −70° C. The next day, subconfluent SK6.T7 cell monolayers were incubated with the suspension of freeze-thawed PBLs for 1 h, after which the suspension was replaced by fresh growth medium, followed by an incubation period of four days. To produce sufficient virus for sequence analysis, the rescued viruses were passaged a few times on SK6.T7 cells.
PEPSCAN analysis. A complete set of overlapping 15 amino acid-long peptides derived from the CSFV strain Brescia E2 protein, spanning amino acids 690 to 851 of the CSFV polyprotein, were synthesized in credit-card format miniPEPSCAN cards as described previously (Geysen et al., 1984. Proc Natl Acad Sci USA 81:3998-4002). The binding of antibodies from sera to each peptide was tested in an ELISA based on pins as described by Slootstra et al. (Slootstra et al., 1996. Mol Divers 1:87-96).
Production and characterization of C-strain mutants with newly introduced potential N-linked glycosylation sites (PNGSs). Full-length cDNA constructs encoding mutant C-strain viruses with newly introduced N-linked glycosylation sites in the TAVSPTTLR epitope (
To produce recombinant virus, linearized plasmids were transfected into SK6.T7 cells, constitutively expressing T7 RNA polymerase. Four days after transfection, in all cases, the presence of infectious virus in the culture medium was demonstrated by infection of fresh SK6-T7 cells. However, large differences in viral fitness were noted. Although foci of infection of vFlc-N1 were somewhat smaller than those of vFlc34 (
Interestingly, the virus produced from pFlc-N5 (
To be suitable as a DIVA vaccine, the vaccine candidate must be incapable of inducing A domain-specific antibodies in vivo. However, to get a first idea if the modifications introduced in the current candidates affected the antigenic structure of the A domain, we determined if the recombinant viruses could be recognized in IPMAs by any of our A domain-specific mAbs. The A domain-specific mAbs used were either raised against CSFV strain Brescia, (i.e. mAbs b2, b3, b4, b7) or the C-strain (i.e. c1, c4, c8 and c11). Also the mAb that is used in the Ceditest 2.0 ELISA, named mAb 18.4, was included in these experiments. Since some of our antibodies only weakly stained the wildtype C-strain, the ability of the antibodies to recognize the mutant C-strain viruses was studied by staining foci of infection, produced by growing the virus under methyl cellulose overlay. Whereas foci of infection of the wildtype C-strain were clearly stained by all antibodies used, these experiments demonstrated that the Pro833→Asn substitution was sufficient to prevent recognition of the A domain by mAbs in vitro (data not shown).
Despite its exposure at the surface of the virion and the fact that it is a dominant target for neutralizing antibodies, the TAVSPTTLR epitope of the A domain has been conserved throughout evolution. This suggests that viruses that contain this exact sequence have a selective advantage over variants within the virus population that contain amino acids substitutions in this region. Although we were successful in recovering viruses with either one (i.e. vFlc-N1) or two (i.e. vFlc-N2) newly introduced PNGSs in this epitope, it was of obvious relevance to study the evolution of these mutants upon growth in tissue culture. To this end, vFlc-N1 was passaged thirty times in vitro. To study the phenotypic stability of the virus, infected monolayers were immunostained with an antibody directed against the non-structural protein NS3, named WB103, and duplicate monolayers were stained with a mixture of A domain-specific antibodies b3 and b4 to determine if phenotypic revertants (i.e. viruses of which the A domain could be stained with these antibodies) were present. After only three passages, a very small percentage of cells infected with vFlc-N1 were stained with the b3/b4 mixture, clearly indicating that phenotypic reversion occurred. The general competitive exclusion principle of population biology states that when two organisms coexist in an environment in which resources are limited, one will eventually overgrow the other (Gause, 1934. Science 79:16-17). Considering this, we wanted to determine if the phenotypic revertant would overgrow vFlc-N1 by repeated passage of the virus. Strikingly, the amount of the phenotypic revertant remained at a very low level, in most experiments represented by only a few cells that were stained with the b3/b4 mAb mixture. It thus appeared that the phenotypic revertant virus was continuously created, but was always maintained as a minority variant in the virus population. Since this finding suggested a lower fitness of the phenotypic revertant relative to that of vFlc-N1, we considered it unlikely that the revertant virus represented a true reversion to the wildtype C-strain sequence. To identify the mutation that was responsible for the revertant phenotype, the virus population needed to be enriched for the revertant virus in order to be able to sequence its genome. By seeding dilutions of the virus in 96-well plates, populations that were enriched for the phenotypic revertants could be selected. After repeating this selection, a population was obtained which contained sufficient amounts of the revertant virus to determine its genome sequence. In two individual experiments, the phenotypic revertant was shown to contain a Ser residue at the position where the Asn was introduced (i.e. position 833 of the CSFV polyprotein), resulting from a single transition in the Asn codon (from AAU to AGU). Since this mutation results in a loss of the PNGS, it is not surprising that it was accompanied by a restored ability of mAbs to recognize the A domain. In conclusion, the detection of the revertant subpopulation demonstrates that vFlc-N1 explored possibilities to increase its fitness by mutation. However, since the newly introduced AAU codon was maintained by the master genotype, the virus is considered genetically stable on the population level.
Analysis of vFlc-N2 revealed phenotypic variants with similar characteristics as those observed during experiments with vFlc-N1, but these were not further characterized.
Analysis of the relative electrophoretic mobilities of E2 proteins with one or multiple newly introduced PNGSs. To determine if the newly introduced PNGSs in the TAVSPTTLR epitope of vFlc-N1 and vFlc-N2 resulted in the attachment of new carbohydrate moieties to E2, the relative electrophoretic mobilities of the modified E2 proteins were studied by polyacrylamide gel electrophoresis (PAGE) under reducing conditions followed by Western blotting. Western blots containing separated proteins of cells infected with vFlc34, vFlc-N1 and vFlc-N2 demonstrated that the E2 protein of vFlc-N1 was of higher molecular weight than the corresponding protein of vFlc34 (
Western blots containing separated proteins of vFlc-N2 suggested that at least one of the two newly introduced glycosylation sequons is used as an anchor site for a carbohydrate moiety (
Production and characterization of C-strain mutants with targeted deletions in the TAVSPTTLR epitope. Full-length cDNA constructs encoding C-strain viruses with targeted deletions in the TAVSPTTLR epitope (
In contrast to vFlc-ΔP, viruses with deletions of more than one amino acid were highly debilitated. Similar as already described for glycosylation mutants vFlc-N3 and vFlc-N4, virus from the supernatant of cells transfected with plasmids encoding these viruses could be passaged a few times, but were subsequently lost. To provide the viruses with the opportunity to increase their fitness by mutation, cells transfected with these constructs were passaged repeatedly. Passage of cells transfected with pFlc-ΔSP, pFlc-ΔSPT, pFlc-ΔVSP, pFlc-ΔAVSP and pFlc-ΔSPTTL did not result in an increase in virus production.
During initial passages of cells transfected with pFlc-ΔPT, the results were similar to those obtained with the other deletion constructs, yielding foci of infection that were of constant small size (average of 10-20 cells). However, after a few additional passages, immune staining demonstrated a sudden improved growth of the virus, which suggested that the virus had introduced fitness compensating mutations. The resulting virus was named vFlc-ΔPTa1. To identify putative resuscitating mutations in vFlc-ΔPTa1, the consensus sequence of its E2 gene was determined. Remarkably, this demonstrated that the virus had retained the introduced deletion and had introduced two mutations in the E2 gene. Conveniently, a clear double peak in the sequence chromatogram of one of the mutations suggested that a transition mutation resulting in a codon change from GAC to AAC was the first to occur (data not shown). This mutation resulted in the substitution of Asp774 to Asn and, interestingly, introduced a new PNGS in the A domain of E2 (
Considering the possibility that adaptive mutations could also be present in genes encoding other structural proteins, also the consensus sequence of the C, ERNS and E1 genes were determined. Sequence analysis of the C gene and the ERNS gene revealed only a single silent mutation in the latter (U1549→C). Interestingly, in the E1 gene, a transversion mutation at position 2275 (A2275→U) was detected that resulted in the substitution of glutamic acid (Glu)-634 for an aspartic acid (Asp) (Table 2).
The in vitro growth characteristics of vFlc-ΔPTa1, relative to vFlc34 and vFlc-ΔP, is visualized by multistep growth curves (
Experimental evolution of viruses produced form plasmid pFlc-ΔPT. The results obtained from sequence analysis of vFlc-ΔPTa1 led us to suggest that the silent mutation in ERNS, the amino acid substitution in E1 and the two substitutions in E2 were responsible for fitness recovery. To test this hypothesis, ten independent transfections with pFlc-ΔPT were performed and the genes encoding the structural proteins were sequenced. The transfected cells were passaged repeatedly and the presence of virus was monitored by IPMAs using mAb WB103. After only two to three passages, the number of positive cells clearly increased, suggesting that the viruses had introduced fitness compensating mutations. After eleven passages, the genomes of the viruses were analyzed as described for vFlc-ΔPTa1. The results of this experiment are summarized in Table 2. The mutation that resulted in the Asp774→Asn substitution was found in six out of ten evolved viruses, clearly demonstrating parallel evolution. However, in one of these viruses, a mutation was found that results in a Asp774→Glu substitution and in the three remaining viruses, no mutation was found in the codon of amino acid 774. The mutation that was responsible for the Val831→Gly in vFlc-ΔPTa1 was detected in one virus (i.e. vFlc-ΔPTa8, Table 2). Interestingly, the silent mutation at position 1549 was found in all viruses. This result suggests that natural selection also operated at the RNA level and that additional adaptive mutations can possibly be found in regions of the genome that were not analyzed in this experiment. However, it was striking to find that three of four mutations detected in the genome of vFlc-ΔPTa1 were again introduced during evolution of the viruses in the current experiment. Of note, in virus vFlc-ΔPTa5 a mutation at a fifth position was detected. Besides the silent mutation in the ERNS gene and the Asp774→Asn substitution in E2, this virus had introduced a transition mutation in the Ala445 codon (from GCA to ACA) of the ERNS gene, which resulted in the substitution of Ala445 for Thr (Table 2).
Analysis of the antibody response against vFlc-ΔP and vFlc-N1. The inability of A domain-specific antibodies to recognize vFlc-ΔP and vFlc-N1 in vitro demonstrated that we were successful in modifying the antigenic structure of the A domain. However, to be suitable as a DIVA vaccine, the antibody response induced in vivo must be sufficiently dampened to be able to serologically differentiate infected from vaccinated animals. Although the vaccine candidates will ultimately be tested for their DIVA property and protective efficacy in pigs, in the current work, we preferred using rabbits for analysis of the humoral immune response, for two main reasons. First, we wanted to determine if the vaccine viruses under study are capable of a productive infection in vivo. In pigs, inoculation of C-strain viruses does not induce any clinical symptoms, whereas inoculation in rabbits induces a temporal febrile illness. Hence, using rabbits enabled us to confirm productive infection and furthermore allowed us to study potential differences in virulence of the selected candidates, which could provide some idea of fitness of the viruses in vivo (de Smit, et al., 2000. Vaccine 18:2351-8). A second advantage of using rabbits is that the C-strain virus can be isolated from the blood, which is often unsuccessful when using pigs. The isolation of the vaccine virus after replication in vivo allowed us to determine if the introduced genetic modifications were stably maintained. Groups of four rabbits were inoculated with vFlc-ΔP or vFlc-N1. Control animals were inoculated with either vFlc34 or culture medium. During the days of the acclimatization period, the average body temperatures of the rabbits were normal (39.2° C., SD±0.28, n=68). Fever was defined as a body temperature above 40.1° C. Fever was first noted in the groups inoculated with vFlc34 (40.7° C., SD±0.48, n=4) and vFlc-N1 (40.5° C., SD±0.32, n=4), both at two days after inoculation. An elevated body temperature in rabbits inoculated with vFlc-ΔP (40.0° C., SD±0.13, n=4) was noted at three days after inoculation.
Virus was isolated from PBLs of three of four rabbits inoculated with vFlc34, three of four rabbits inoculated with vFlc-N1 and two of four rabbits inoculated with vFlc-ΔP. Consensus sequencing demonstrated that the E2 genes of these viruses were not altered by the passage in rabbits. To determine if the C-strain mutants enable the differentiation between infected and vaccinated animals, the rabbit antisera were analyzed by the Ceditest 2.0 E2 ELISA (Prionics). This ELISA specifically detects antibodies against the A domain of E2. As a reference control for effective immunization, the CHEKIT CSF ERNS ELISA (IDEXX laboratories) was used.
The ERNS responses induced by vFlc34, vFlc-N1 and vFlc-ΔP were comparable (
Analysis of the antibody response against vFlc-ΔPTa1. In the second animal trail, the antibody response against vFlc-ΔPTa1 was compared with that induced against vFlc34. The average body temperature of the rabbits prior inoculation was normal (39.2° C., SD±0.37, n=28). In this experiment, only one of two rabbits inoculated with vFlc34 experienced fever at two days after inoculation (40.3° C.). One rabbit (rabbit 2.1) that was inoculated with vFlc-ΔPTa1 did not show any elevation in body temperature. In contrast, fever was noted in the remaining two animals inoculated with vFlc-ΔPTa1 (rabbits 2.2 and 2.3). In one of these animals a temperature was noted of 40.6° C. at five days post inoculation, whereas in the other rabbit a temperature was noted of 40.4° C. at seven days post inoculation. Considering the fact that vFlc-ΔP did not induce fever, this finding was highly unexpected. It is important to note, however, that shortly after inoculation, rabbits 2.2 and 2.3 were involved in a quarrel that resulted in the affliction of wounds. Considering that these wounds induced an infection of the skin and the fact that the symptoms of this infection coincided with the occurrence of fever, it is plausible to assume that the fever resulted from the skin infection, rather than being a consequence of the virus inoculation. At four days after inoculation, virus was isolated from both animals inoculated with vFlc34. No virus was isolated from PBLs of the three rabbits inoculated with vFlc-ΔPTa1, which is likely related to the attenuation of this virus.
The sera of the two rabbits that were inoculated with vFlc34 (animals 1.1 and 1.2;
To analyze the antibody responses more thoroughly, we made use of PEPSCAN analysis using previously constructed 15 amino acid-long overlapping peptides derived from the E2 protein of CSFV strain Brescia. This analysis revealed that the sera obtained at day 36 from animals inoculated with the wildtype C-strain recognized two epitopes of CSFV strain Brescia (
In the past two decades, several experimental DIVA vaccines were developed that need either be accompanied by an E2 or an ERNS ELISA. It is often stated in literature that both diagnostic tests are suitable for the detection of CSF in the field. Although this is in principal correct, it is important to note that ERNS ELISAs have two major disadvantages over E2 ELISAs. First, although ERNS ELISAs can be used to reliable detect infected herds, these ELISAs were previously found to be insufficiently sensitive to diagnose individual animals. Second, and much more importantly, ERNS ELISAs are not CSF-specific (2003/265/EC). SANCO/10809/2003. European Commission, Directorate-General for Health and Consumer Protection, Brussels). Consequently, use of ERNS ELISAs in regions where BVDV and/or BDV viruses circulate is not recommended.
The objective of the current work was to construct a C-strain-based DIVA vaccine that is compatible with an E2 ELISA. We aimed to achieve this by introducing judicious mutations in the TAVSPTTLR epitope of the A domain, which is the dominant target in these ELISAs. In the first of two approaches, the central Pro (i.e. Pro833) residue of the TAVSPTTLR epitope was substituted for Asn, thereby introducing a PNGS in the center of the TAVSPTTLR epitope. Although N-linked glycans are well known for their ability to shield immunogenic domains from the humoral immune system, they can be remarkably well tolerated at or near functionally important domains, two features that are attributed to the structural flexibility of the carbohydrate chain. Although we were successful in producing a C-strain mutant, named vFlc-N1, that stably maintains a newly introduced N-linked glycan anchored to the center of the TAVSPTTLR epitope, this modification seemed to have an only minor effect on the A domain-specific antibody response. For this reason, we chose not to continue this approach. In an alternative approach, we aimed to alter the antigenic structure of the A domain by deleting amino acids from the TAVSPTTLR epitope. Deletion of the central Pro residue (i.e. Pro833) yielded a virus, named vFlc-ΔP, that stably maintained the introduced deletion, although the virus eventually acquired an adaptive mutation, resulting in the substitution of Ser789 to phenylalanine (Phe). We found this nonconservative substitution interesting, since Ser789 is present in all C-strain viruses and related lapinized CSFV strains, whereas Phe789 is completely conserved in virulent CSFV strains and even highly conserved among other members of the pestivirus genus (van Rijn et al., 1997. Virology 237:337-48). Although the history of the C-strain virus is not well documented, it is clear that the virus was attenuated by passage in rabbits hundreds of times (van Oirschot, 2003. Vet Microbiol 96:367-84). It could therefore be hypothesized that Ser789 is advantageous for replication of the C-strain in rabbits. If this assumption is correct, it is interesting to reason why the C-strain does not revert to Phe789 upon growth on porcine cells in vitro. This can be explained by the fitness landscape concept of Wright (Wright, 1931. Genetics 16:97-159). According to this concept, advantageous mutations arise slowly when a genotype exists at the top region of a fitness peak, as likely applies for the wildtype C-strain virus, whereas these arise more quickly when the genotype exists at a lower level of a fitness peak, as is the case for vFlc-ΔP.
Analysis of sera obtained from rabbits inoculated with vFlc-ΔP demonstrated that the deletion of Pro833 is already sufficient to considerably dampen the immunogenicity of the A domain. Although this result demonstrated proof of principle of our approach, it also made clear that the antigenic structure of the A domain needed to be more extensively modified to render the resulting virus suitable as a DIVA vaccine. Although plasmids encoding C-strain mutants with deletions of two to four amino acids from the TAVSPTTLR epitope (ΔSP, ΔPT, ΔSPT, ΔVSP, ΔAVSP) produced infectious virus, these were highly debilitated and incapable of sustained growth. After a few passages of cells containing vFlc-ΔPT, however, a remarkable increase in fitness was noted. The rescued virus was named vFlc-ΔPTa1. We considered it most likely that adaptive mutations in the E2 gene were responsible for fitness-recovery. Indeed, two adaptive mutations were detected in the E2 gene (
The rescue of a C-strain mutant with a PNGS in the A domain (i.e. vFlc-ΔPTa1,
Coming back to virus vFlc-ΔPTa1, it is interesting to note that the newly introduced PNGS in vFlc-ΔPTa1 at position 774 might also be located in proximity of the TAVSPTTLR epitope in the native E2 structure. Namely, the Asp774→Asn substitution in this virus is located at the center of a recently identified epitope comprising amino acids 772LFDGTNP778 (Peng et al., 2008. Virus Res 135: 267-72). Like the 829TAVSPTTLR837 epitope, the 772LFDGTNP778 epitope shares all three features that define the A1 domain, being CSFV-specific, evolutionarily conserved and a target for neutralizing antibodies. Based on the aforementioned and our experimental findings, we hypothesize that the TAVSPTTLR and the LFDGTNP epitope are co-localized in the native structure of E2 and, possibly, together make up the A1 domain.
In summary, we hypothesize that a mutation of the central Pro residue of the TAVSPTTLR epitope can be compensated by N-linked glycans located at one of three possible positions in the E2 protein, and that these positions are located in proximity in the native E2 structure. The first is located within the 829TAVSPTTLR837 epitope itself, the second in the 772LFDGTNP778 epitope and the third domain includes the amino acid sequence 858TTT860 (
Although it is plausible to assume that the adaptive mutations in the E2 gene play an important role in fitness recovery, we reasoned that also mutations in other genes encoding structural proteins could play a role. Therefore, also the C, ERNS and E1 genes of vFlc-ΔPTa1 were sequenced. Sequencing of this part of the genome indeed revealed two additional mutations. The first was detected in the E1 gene. This mutation resulted in the substitution of the completely conserved Glu634 residue for Asp. Considering that the E2 protein is known to assemble into disulphide-linked heterodimers with the E1 protein (Thiel, et al., 1991. J Virol 65:4705-12), it is conceivable that this mutation contributed to the fitness recovery of vFlc-ΔPTa1. Although no mutations were detected in the C gene, a single silent mutation was detected in the ERNS gene. Bearing in mind that the consensus sequence of wildtype CSFV is exceptionally stable, accumulating virtually no mutations upon passage in vitro or in vivo (Vanderhallen, et al., 1999. Arch Virol 144:1669-77), the detection of such a silent mutation by consensus sequencing demonstrated that natural selection also operated at the RNA level, and thereby that also synonymous mutations in other regions of the genome could have contributed to the fitness recovery of vFlc-ΔPTa1.
To gain insight into the molecular pathway of fitness recovery, an evolution experiment was performed. This experiment demonstrated parallel evolution of the silent mutation in the ERNS gene, the mutation that resulted in the Val831→Gly substitution and the mutation that introduced the new PNGS at position 774 (Table 2). Although additional studies are clearly required to fully elucidate the molecular mechanism of fitness recovery, this finding supports the notion that the aforementioned mutations were indeed involved in the fitness recovery of vFlc-ΔPTa1. It is very well possible that these mutations can restore the fitness of mutants with larger deletions in the TAVSPTTLR epitope.
Worth mentioning, in one of the viruses studied in the evolution experiment (i.e. vFlc-ΔPTa5), a mutation was detected in the ERNS gene that resulted in the substitution of Ala445 to Thr (Table 2). Like the previously described Ser789 to Phe substitution found upon the evolution of vFlc-ΔP, the amino acid that is altered in vFlc-ΔPTa5, in this case Ala445, is changed into an amino acid that is conserved among CSFV field strains, in this case Thr445. Although this falls beyond the scope of the current work, this finding again demonstrates that attenuation of CSFV can provide valuable insights into its molecular evolution.
Previous studies with MAR mutants performed by Van Rijn et al., suggested that Pro833 and Thr834 of the TAVSPTTLR epitope are very important for the integrity of conserved epitopes of the A domain (van Rijn, et al., 1994. J Virol 68:3934-42). The successful production of a well growing C-strain mutant that lacks both these amino acids and, in addition, contains a Val831→Gly substitution, suggested that the antigenic structure of the A domain of this virus was changed dramatically. We therefore expected that that the immune response induced by vFlc-ΔPTa1 would be distinguishable from that induced by wildtype CSFV using the Ceditest 2.0 E2 ELISA. It was surprising to find that vFlc-ΔPTa1 can actually be quite potent in inducing A domain-specific antibodies. A possible explanation for this finding is that the disruption of the otherwise immunodominant TAVSPTTLR epitope can result in a refocusing of the antibody response towards epitopes of the A domain that are normally subdominant. In an attempt to identify these epitopes, the antisera were analyzed by PEPSCAN analysis. Although this analysis did not reveal the newly recognized epitopes, it did demonstrate that none of the antisera recognized peptides containing the TAVSPTTLR epitope. Interestingly, this suggests that an ELISA based on the TAVSPTTLR epitope can be used as a DIVA test to accompany the vFlc-ΔPTa1 vaccine virus.
In summary, the current work demonstrates that forced virus evolution can be a powerful tool to genetically modify the CSF virus. Here, this method was successfully employed to produce a genetically stable C-strain mutant that can be serologically differentiated from wildtype CSFV.
The C-strain deletion mutant virus vFlc-ΔPTa1 was produced as described in Example 1. SK6-T7 cells are swine kidney cells that constitutively express T7 RNA polymerase (van Gennip et al., 1999. Vaccine 19(4-5), 447-59). These cells were grown in K1000 medium supplemented with glutamine (0.3 mg/ml, Gibco, Invitrogen, Breda, The Netherlands), 5% fetal bovine serum (FBS) and the antibiotics penicillin (100 U/ml, Gibco), streptomycin (100 U/ml, Gibco) and amphotericin B (2.5 μg/ml, Gibco). Virus stocks were titrated on SK6.T7 cells and are reported as 50% tissue culture infective doses(TCID50).
Thirteen conventional pigs free of antibodies against pestiviruses were divided into two groups of five pigs and one group of three pigs. Pigs from group 1 (nos. 3166-3170) were vaccinated once on day 0 via the intramuscular route with 1 ml of culture medium containing 2% fetal bovine serum (FBS) and 105 TCID50 of the recombinant C-strain virus vFlc34 (see Example 1). Pigs from group 2 (nos. 3171-3175) were vaccinated, by the same protocol, with the vFlc-ΔPTa1 vaccine. Pigs from group 3 (nos. 3176-3178) were inoculated once with culture medium containing 2% FBS.
On day 28, pigs from groups 1, 2 and 3 were challenged intranasally with 1 ml (0.5 ml each nostril) containing 200× the 50% lethal doses of CSFV strain Brescia. All pigs were observed daily for disease and their body temperatures were measured starting at day −3. The severity of clinical symptoms was scored using a previously defined list of 10 CSF-specific criteria (Mittelholzer et al., 2000. Vet Microbiol 74(4), 293-308).
Serum samples were prepared weekly from all pigs. EDTA samples and oropharyngeal fluids were collected from all pigs of groups 1, 2 and 3 at the day of challenge (day 28) and subsequently on days 30, 32, 35, 37, 39, 42, 44, 46, 49, 52 and 56. EDTA samples were used to determine the number of leukocytes and thrombocytes. This analysis was performed with the Medonic CA 620 coulter counter (Boule Medical AB). In conventional reared pigs, leukocyte and thrombocyte concentrations in the blood range between 11×109 and 23×109 and between 320×109 and 720×109/liter, respectively. Consequently, leukopenia was defined as <8×109 celles/l blood and thrombocytopenia as <200×109 thrombocytes/1 blood.
Virus isolation and quantitative real-time reverse-transcription PCR (qRRT-PCR) using peripheral blood leukocytes (PBLs) or throat swabs were performed essentially as described (Weesendorp et al., 2009. Vet Microbiol 135(3-4), 222-30).
Serum samples were analyzed using a commercial E2 ELISA (PrioCHECK CSFV Ab 2.0 E2 ELISA, Prionics, Lelystad, The Netherlands) and a commercial ERNS ELISA (Chekit CSF ERNS ELISA, IDEXX laboratories, Hoofddorp, The Netherlands), according to the instructions of the manufacturers.
We compared the protective efficacy of the vFlc-ΔPTa1 virus with its parental virus, vFlc34, against a lethal challenge with the highly virulent Brescia strain. Two groups of five pigs were vaccinated on day 0 with either vFlc34 (group 1) or with vFlc-ΔPTa1 (group 2). A control group (group 3), consisting of three pigs, was inoculated with cell culture medium only. Pigs from groups 1, 2 and 3 were challenged with the highly virulent Brescia strain on day 28.
All pigs in the control group developed typical clinical signs of CSF, including high fever (
None of the pigs vaccinated with the recombinant C-strain vaccine (i.e. vFlc34) developed clinical signs after challenge (
All pigs that were vaccinated with vFlc-ΔPTa1 developed fever for 3 to 6 days and displayed mild clinical signs after challenge (
All sera were analyzed for the presence of anti-E2 and anti-ERNS antibodies with the PrioCHECK CSFV Ab 2.0 E2 ELISA (Prionics) and the Chekit CSF ERNS ELISA (IDEXX laboratories), respectively. All pigs vaccinated with vFlc34 were positive in the E2 ELISA at 21 days p.i. (
Thus, it is concluded that all pigs vaccinated with vFlc-ΔPTa1 remained below the cut-off of the PrioCHECK E2 ELISA until challenge, showing that vFlc-ΔPTa1 can be used as a DIVA vaccine also in pigs.
In conclusion, we demonstrate that a single vaccination with vFlc-ΔPTa1 provided protection against a highly virulent challenge in all vaccinated pigs and evidence is provided that very little or no virus shedding occurred.
aApaLI restriction sites (underlined) and nucleotide substitutions (bold) are indicated.
GAC → AAC
GAC → AAC
GCA → ACA
GAC → AAC
GAC → AAC
GAC → AAC
GAC → AAC
GAC → AAC
aIdentified mutations are indicated in bold.
bThe in vitro growth characteristics of virus vFlc-ΔPTa1 are depicted in FIGS. 2 and 3b. The growth characteristics of the other viruses were not determined.
| Number | Date | Country | Kind |
|---|---|---|---|
| 08172874.3 | Dec 2008 | EP | regional |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/NL09/50801 | 12/23/2009 | WO | 00 | 8/11/2011 |
