The infectious clone pSP6-SFV4 showing some of the restriction sites which can be used for genetic manipulation. The SFV genome (11445 nt) has been cloned under the control of the SP6 promoter ; p62 is the precursor to the E2 and E3 envelope proteins which is cleaved during virus maturation. To produce infectious virus, the plasmid is linearized with Spe I, then transcribed with SP6 polymerase to produce infectious RNA, which is then transfected into BHK cells by electroporation. 

Context in source publication

Context 1

... nonstructural proteins are translated from the genomic 42S RNA. The structural and nonstructural proteins are formed from precursors by separate post-translational cleavage pathways (Strauss & Strauss, 1994). By historical accident, most work on SFV has been carried out in Europe, whereas a different model alphavirus, Sindbis virus, has been studied in the USA. Here we intend to concentrate on reviewing recent advances in the molecular pathogenesis of SFV, but to refer to work on Sindbis virus where relevant. We last reviewed the molecular pathogenesis of SFV in 1985 (Atkins et al ., 1985), and since then infectious clones of SFV have been constructed and developed into vectors, and there have been advances in understanding immune mechanisms in SFV infection and virus–cell inter- actions. The original isolate of SFV is designated L10, and is neurovirulent for mice, causing lethal encephalitis by infection of the central nervous system (CNS). Subsequently, an avirulent strain designated A7 was isolated from mosquitoes in Mozambique (McIntosh et al ., 1961). Most SFV strains used for laboratory studies are derived from these two isolates. The avirulent A7[74] strain was derived from A7 by further selection for avirulence (Bradish et al ., 1971). The prototype strain, from which the original infectious clone of SFV was derived, appears to be derived from the L10 strain. However, it has lost some of its virulence, possibly due to an indeterminate number of passages in cell culture (Glasgow et al ., 1991). The original infectious clone of SFV, constructed from the prototype strain, is designated pSP6-SFV4 ; the virus produced by transcription of this infectious clone is labelled SFV4. SFV does infect humans and although an outbreak of a mild febrile illness among French soldiers serving in the Central African Republic was shown to be due to SFV (Mathiot et al ., 1990), SFV is considered less of a hazard to laboratory personnel than many other togaviruses (Winkler & Blenden, 1995). However, there has been one death suspected to be due to laboratory infection by SFV (Willems et al ., 1979). For this reason, SFV is classified as a Biosafety Level 3 virus in the USA, but with the caveat that most activities can be carried out at level 2 (U. S. HHS Publication no. (CDC) 93-8395, 1993). In the European Union, SFV is classified as a Biosafety Level 2 virus (E. C. Council Directive 93 \ 88 \ EEC, 1993). The SFV expression vectors, which lack the structural genes, are classified as Biosafety Level 2 in the EU and USA (Federal Register, 1993). The original L10 isolate of SFV shows complete virulence in inbred mouse strains such as BALB \ c. Following peripheral infection, all infected mice die, and the LD is 1 p.f.u. ; benign &! multiplication resulting in immune protection against virulent challenge does not occur. However, the virulent virus may be attenuated by mutation. The first such mutations studied were induced by chemical mutagenesis (Barrett et al ., 1980) and affected defined stages in virus multiplication ; for example, the M9 mutant, derived from the virulent L10 strain, has a defect in efficiency of viral RNA synthesis (Atkins & Sheahan, 1982). These defects allowed the survival of the majority of mice following peripheral infection, and allowed the consequences of avirulent virus multiplication to be detected (mainly CNS demyelination and teratogenesis). The avirulent A7 strain, however, multiplies in standard BHK cells at least as efficiently as the virulent L10 strain. All adult mice survive infection with A7, but neonatal mice are killed by infection with either L10 or A7. Infection with A7 does, however, induce CNS demyelination and foetal death in pregnant mice. As indicated above, initial attempts to analyse the virulence of SFV utilized chemically induced attenuated mutants of the virulent L10 strain. Such mutants are partially defective in the efficiency of virus multiplication (Barrett et al ., 1980 ; Atkins & Sheahan, 1982 ; Atkins et al ., 1985). Perhaps more interesting is analysis of the avirulent A7 strain, which is attenuated but multiplies at least as efficiently as L10 in standard cultured cells such as BHK cells (Atkins, 1983 ; Glasgow et al ., 1997). Initial studies analysed virulence following intraperitoneal (i.p.) infection, but it was found subsequently that intranasal (i.n.) infection gives more consistent results, and is a more direct route of infection to the CNS. Intranasal infection also targets the olfactory bulb, allowing analysis of early events following CNS infection (Sheahan et al ., 1996). It is a more sensitive indicator of virulence than i.p. infection, although a higher dose is required to produce initial infection. Male mice appear to be more sensitive to SFV infection than female mice (Santagati et al ., 1998). Analysis of the molecular basis of SFV virulence was facilitated by the construction of an infectious cDNA clone of SFV, derived from the prototype strain (Fig. 1 ; Liljestro $ m et al ., 1991). The SFV4 virus, produced by transcription of the plasmid encoding the SFV sequence, kills about 60–70 % of BALB \ c mice when 10 ' p.f.u. is given i.p., and 100 % when the same dose is given i.n. (Glasgow et al ., 1991). The A7 strain has been completely sequenced and the sequence compared to the more virulent prototype strain (Glasgow et al ., 1994 ;Santagati et al ., 1994, 1995 ; Tarbatt et al ., 1997). The most striking feature of the A7 sequence is the presence of a long untranslated sequence containing multiple repeats at the 3 h end of the genome. In A7 this region is 334 nucleotides longer than that of the prototype (and L10) strains (Glasgow et al ., 1994 ; Santagati et al ., 1994). There are multiple mutations throughout the A7 genome compared to the prototype strain, in all virus genes and in the 5 h untranslated region. A large proportion of the nucleotide substitutions in the translated region result in amino acid substitutions. Analysis of chimeras constructed between the avirulent A7 and A7[74] strains, and the more virulent SFV4 strain, shows that determination of virulence is polygenic. However, three regions of the genome appear to be important : the E2 gene, the nsP3 gene and the 5 h untranslated region (Santagati et al ., 1995, 1988 ; Santagati, 1998 ; Tarbatt et al ., 1997). Surprisingly, the 3 h untranslated region is unimportant in the determination of virulence (Tarbatt et al ., 1997 ;Santagati et al ., 1998). Control of virulence by the E2 gene has been described for Sindbis virus (Pence et al ., 1990 ; Tucker et al ., 1993 ; Ubol et al ., 1994) and Ross River virus (Meek et al ., 1989), and by the 5 h untranslated region in combination with the E2 gene for Venezuelan equine encephalitis virus (Kinney et al ., 1993). For the A7[74] strain of SFV, which has been further selected for avirulence compared to A7 (Bradish et al ., 1971), further changes have occurred in the nsP3 gene. First, a 21 nucleotide in-frame deletion has occurred in the portion of the nsP3 gene corresponding to the C-terminal domain of the nsP3 protein ; this deletion is flanked by repeated sequences. Secondly, an opal codon is present near the 3 h end of the nsP3 gene (Santagati, 1998). These changes are attenuating, and a chimeric virus containing the nsP3 and nsP4 genes from A7[74] and the rest of the genome from SFV4 is attenuated when given by the i.p. route. However, when given by the i.n. route, virus having the nsP3 deletion is still virulent. Changes in other regions of the genome appear to be required to fully attenuate the virus (G. Atkins, unpublished results). For the virulent SFV4 strain, a nuclear targeting signal is present in the nsP2 gene, which has the sequence P '%) RRRV (Rikkonen et al ., 1992 ; Kujala et al ., 1997), and whose function is unknown. Modi- fication of this signal by site-directed mutagenesis does not affect the multiplication of the virus in cell culture, but does affect its pathogenicity for mice (Rikkonen, 1996). In the avirulent A7[74] strain, the nuclear targeting function is abrogated by mutation to the sequence PQRKF (G. Atkins, unpublished results). Thus this appears to be a further pathogenicity determinant which contributes to the avirulence of the A7[74] strain. Early work on SFV virulence indicated that virulent strains induced more extensive damage to neurons in the CNS than avirulent strains (Atkins et al ., 1985). Multiplication of avirulent strains in the CNS is slower than virulent strains, and this is linked to the ability to multiply in neurons, which is partially restricted in avirulent strains (Gates et al ., 1985 ; Balluz et al ., 1993 ; Fazakerley et al ., 1993). In cultured rat cerebellar granule cells (a type of neuron, Fig. 2 c , d ), virulent strains multiply faster and to higher titre than avirulent strains. However, the restricted multiplication is more marked at low m.o.i., when the virus has to undergo several rounds of multiplication to infect every cell in the culture (Atkins et al ., 1990). Thus the crucial difference between virulent and avirulent strains is the rapidity of spread of neuronal damage in the CNS (Balluz et al ., 1993), leading to a lethal threshold in the case of virulent strains before the immune system can intervene. It is clear that virulent strains of SFV spread rapidly in the CNS, probably by axonal transport (Fig. 2 a ; Kaluza et al ., 1987). However, following i.p. infection, the avirulent strain A7[74] of SFV crosses the blood–brain barrier by infection of vascular endothelial cells (Dropulic & Masters, 1990 ; Soiluha $ nninen et al ., 1994), but then fails to spread rapidly in neurons (Fazakerley et al ., 1993). Thus virulent and avirulent strains of SFV differ in their ability both to multiply in neurons and to cause extensive neuronal damage by rapid spread. As yet, the virus properties which control these functions are unknown. Mice surviving infection by avirulent SFV show ...