PCR restriction analysis of genome composition and stability of cold-adapted reassortant live influenza vaccines

Journal of Virological Methods

LSEVIER

Journal of Virotogicat Methods 52 (1995) 41-49

PCR restriction analysis of genome composition and stability of cold-adapted reassortant live influenza vaccines Alexander I. Klimov a,b,*, Nancy J. Cox a a Influenza Branch, G-16, Division of Viral and Rickettsial Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, GA 30333, USA h Research Institute for Viral Preparations, Russian Academy of Medical Sciences, Moscow 109088, Russia Accepted 30 August

1994

Abstract Using cold-adapted master donor strains of influenza virus as a model, an approach was developed that exploits unique nucleotide differences between the donor strains and wild-type influenza viruses for rapidly and simply determining the genome composition and genetic stability of live attenuated vaccine reassortants. The approach is based on PCR amplification of approximately HO-300-nucleotide-long regions of individual RNA segments that include the unique nucleotide positions, followed by restriction nuclease treatment of the DNAs obtained with specific restriction endonucleases. Restriction sites recognized by chosen nucleases either existed or were created during PCR in the genome of one (but not the other) parent strain. The technique requires a minimal amount of infectious virus (approx. 100 ~1 of allantoic or tissue culture fluid with a haemagglutination titre 1:4-l% or less) and allows rapid (within about 10 h) determination of the origin of the RNA segment or the presence of a mutation. The method is beneficial for genome composition analysis of reassortant vaccine strains as well as for investigation of the genetic stability of live attenuated vaccines during replication in vaccinees. Keywords: analysis

Influenza

* Corresponding

virus; Vaccine

reassortants;

Genome

composition;

author at address a. Fax + I (404) 6392334.

0166-0934/95/$09.50 0 1995 Elsevier Science B.V. All rights reserved .SSDf 0166-0934(94)00133-2

Genetic

stability;

PCR restriction

42

A.I. Klimot,,

NJ. (‘ax /.lournul

of Virolo~mzl

Methodr

52 (199.51 31-39

1. Introduction Periodic changes in the antigenic specificity of influenza virus surface glycoproteins (i.e., the haemagglutinin and neuraminidasel require that the antigenic composition of influenza vaccines be updated almost every year. Genetic reassortment between the current epidemic viruses and a carefully characterized laboratory donor strain is used to new vaccine variants. Reassortment between the laboratory-adapted produce A/PR/8/34 (HlNl) strain with desirable growth properties and currently circulating viruses with desirable antigenic properties is used to produce new versions of inactivated vaccine. For live attenuated (att) influenza vaccines that have been widely used in Russia (Zhdanov, 1986) and have also been investigated in the USA (Kendal et al., 1981; Wright and Karzon, 19871, cold-adapted (ca) strains (i.e., strains that can replicate at low (25°C) temperature) are in use as donors to produce attenuated vaccine reassortants. To ensure the maximum possible uniformity of high yield (hy) or attenuation properties, the genome composition of a candidate vaccine reassortant should be 6:2. This means that the six genes coding for non-glycosylated virus proteins (‘internal’ genes) should be obtained from the hy or art donor, while the other two genes encoding the haemagglutinin and neuraminidase should be inherited from the epidemic virus. Thus, determining the genome composition of reassortants is an important step in preparing vaccine candidates. Haemagglutination inhibition (HI) and neuraminidase inhibition (NI) tests can be used to determine the origin of genes coding for surface glycoproteins; however, internal genes must be identified using other methods. Several approaches have been developed for determining the genome composition of influenza virus reassortants (Scholtissek, 1984). The most popular one is based on differences in mobilities of single-stranded RNA segments of different influenza virus strains during polyacrylamide gel electrophoresis (PAGE) (Palese and Shulman, 1976; Ritchey et al., 1976). One disadvantage of this method is that mobilities of some cognate RNA segments (especially the largest ones) of different strains sometimes cannot be differentiated. Another disadvantage is that the order of migration during PAGE may be reversed for certain RNA segments; for example, under conditions in which the neuraminidase gene of A/Hong Kong/ l/68 (H3N2) virus migrates as RNA segment 5, the same gene of the A/PR/8/34 (HlNl) strain appears in the same gel as RNA segment 6 (Palese and Shulman, 1976). Another widely used approach is based on hybridization of labelled complementary RNA (isolated from infected cells incubated in the presence of cycloheximide and [ 3Hluridine) with an excess of unlabelled homologous or heterologous virion RNA, followed by treatment of double-stranded RNAs with nuclease Sl and analysis by PAGE (Hay et al., 1977; Ghendon et al., 1979, 1981). Although this technique is more specific, it is time consuming and requires a considerable amount of viral RNA. Alternative methods for determining the genome composition of reassortants are more complicated and/or less specific (Scholtissek, 1984). Accumulation of sequence data for many influenza genes as well as widespread use of the polymerase chain reaction (PCR) technique to obtain DNA copies of individual genes has allowed the development of a practical approach for genome analysis that combines the high sensitivity of the PCR technique with the high specificity of

A.I. Klimou, N.J. Cox/Journal

of Virological Methods 52 (1995) 41-49

43

restriction analysis. Here we describe an approach we have developed that uses live attenuated vaccine viruses as a model. Recently we identified sequence changes compared to the parent wt virus in the internal genes (PB2, PBl, PA, NP, M and NS) of two ca master strains, A/Leningrad/l34/17/57 (H2N2) (A/Len/l7) and A/Leningrad/134/47/57 (H2N2) (A/Len/47), that are used in Russia to prepare reassortant live attenuated vaccines for adults and children, respectively (Klimov et al., 1992). Ten nucleotide differences (eight coding) were detected between the sequences of the A/Len/l7 and A/Leningrad/l34/57 (H2N2) wild-type (A/Len/W) viruses. Four additional changes (three coding) were found in the genome of the A/Len/47 ca virus. All but one of these nucleotide changes are unique among published influenza gene sequences. This makes it possible to differentiate exactly between these two vaccine strains and other strains, as well as to evaluate the genetic stability of these defined mutations during replication of the live att vaccine in vaccinees. The approach described here exploits unique nucleotide differences between vaccine donor strains and wr viruses for the analysis of the genome composition and the genetic stability of reassortant influenza vaccines. This approach is based on PCR amplification of regions near distinctive positions in each internal RNA segment, followed by treatment of DNAs obtained with restriction nucleases specific to nucleotide sequences in the genome of one, but not the other parent strain.

2. Materials and methods 2.1. RNA extraction

Virion RNAs were extracted from allantoic fluids (100 ~1) by RNAzol accordance with the manufacturer’s protocol (Biotecx Laboratories, Inc.).

B in

2.2. PCR amplification

The regions including unique nucleotide positions in each internal RNA segment (Klimov et al., 1992) were amplified by the PCR method. Forward primers were PB2-1374 (5’-d GGGAATTGAACATATCGA), PBl-740 (5’-d GAGCAA’ITGCAACACCCG), PA-900 (5’-d TGAGGACCCAAGTCACGA), NP-886 (5’-d TATGGACCTGCCGTAGCC), M-785 (5’-d CCCTCTTGTI’GTTGCCGC), and NS-673 (5’-d CAAAACAGAAACGGAAAA). Reverse complementary primers were PB2rc-1614 (5’-d TATTGTCAGTITCTCTGT), PBlrc-1044 (5’-d TGCGATGCTCAGGACG’IT), PArc1218 (5’-d AGGTTCATCACTATCATA) or PArc-1077a (5’-d CTCATTCTCAATGTCCIGCAG’ITCTGCCA’ITA), NPrc-1200 (5’-d GTACCTGCTTCTCAGTTC), Mrc-1001 (5’-d CAGCTCTATGCTGACAAA), and NSrc-830a (5’-d TGTI’CCACTTCAAATAGTAGCTGTAAGGCATG).

A.I. Klimoc~, NJ. Cox / Journal

44

2.3. Restriction

of Viroiogical

Methods 52 (1995) 41-49

analysis

Restriction enzymes Mt~l or Tru91, Hindlll, Seal or Asnl, S&II or EcoRl, F&l or Tsp5091 and Nsil were used for restriction analysis of PCR-amplified DNA fragments of the PB2 (nucleotides 1374-1614), PBl (740-1044), PA (900-1218 or 900-1077), NP (886-12001, M (785-lOOl), and NS (673-830) genes, respectively (Figs. 1 and 2). Restriction enzymes (2-5 U) (Boehringer Mannheim or New England Biolabs) were added directly to 10 ~1 aliquots of PCR product and incubated for 2 h at 37°C (along with control DNA& followed by electrophoresis in 1.5% agarose gels.

3. Results 3.1. Analysis of genetic stability of the liue attenuated

ca reassortant

caccine

Comparison of sequences of internal genes of A/Len/wt and A/Len/47 (Klimov et al., 1992; GenBank accession numbers M81570-M81587) reveals that unique nucleotide changes in genes PB2 (C-1497-A), PBl (G-819-T), PA (G-1045-T), NP (C-1066-A), and M (G-969-A) of the A/Len/47 donor lead to elimination of recognition sites for restriction enzymes Mrlnl, Hindlll, Seal, Bglll, and F&l, respectively, which exist at corresponding positions of the A/Len/wt and other viruses (Fig. 1). In contrast, the nucleotide change (G-1459-T) in the PB2 gene creates a Tru91 restriction

Len/wt

Len/47

2

$1

GTAACATT-

CGCG GCGF

Mvn I AA -AGC TT-

PB2

TCG

4

PBZrc-1614 m m

P

P61 K-1 044

i

P

PA-900 AGTACF-------

AL;TACY TAATGm6pA

NP-886

Bgl II i

-

$

N

AGATCTCTAG

NP NPrc-1200 5 i

Fok I

GGATG CCTAi-

NPrc-1200 8 ID

M-786 GGATA CCTAT-

M

Mm-1 001 g

NSa Fig. 1. PCR-amplified genes of A/Len/

1

Nsi I

ATG CATTACIGT NSrc-830a

Mm-1 001 i ATACAT TA-lIGTd

~ NS NSrc-830~1

areas, nucleotidc differences, restriction enzymes and restriction sites in the internal

WI and A/Len/47

strains used for analysis of live attenuated vaccine genetic stability.

A.I. Klimou, NJ. Cox/ Journal of Virological Methods 52 (I 995) 41-49

Lenlwt

45

Lent47

PB2

GCG PB2rc-1614 P

t

P AAGCTF------TTCGAm44

I

g,Sca

i

AGTAC-

ATTACTAATGA---%??8

‘*

NP NPrc-1200 ID

z

M-785 GGAT CCTA

M

Mm-1 001 :

1

Nsi I

ATG CATTACKiTW NSrc-830a

Mrc-1001

z ATACAT TAlIGT”-

NS NSrc-830a

Fig. 2. PCR-amplified areas, nucleotide differences, restriction enzymes and restriction sites in the internal genes of A/Len/ca strains and other (wt) viruses used for analysis of reassortants’ genome composition.

site in the A/Len/47 (A/Len/l7) strain; this site is absent in the PB2 gene of other viruses including the A/Len/wt. Thus, these enzymes can be used to examine the conservation of mutations related to attenuation in genomes of ca reassortant vaccine strains during their preparation and during their replication in vaccinees. For the analysis of the NS gene, reverse primer NSrc-830a (see Materials and methods) was designed for the PCR amplification to create a NsiI recognition site in the situation in which a strain has A-798 (wt viruses), but not G-798 (A/Len/ca donors) (Fig. 1). Fig. 3 represents an example of the application of this PCR restriction analysis to examine the genetic stability of live attenuated influenza vaccine during replication in vaccinees. The vaccine strain A/Zak(R) (H3N2) was obtained by reassortment of the epidemic virus A/Zakarpatje/354/89 (H3N2) (A/Shanghai/l1/87_like) with the A/Len/47 donor and inherited the haemagglutinin and neuraminidase genes from the epidemic virus and all other genes from the ca master strain (see below). This reassortant vaccine strain was used in 1990 in Novgorod, Russia, for immunization of schoolchildren in a large comparative vaccine trial (Rudenko et al., 1993). PCR restriction analysis of this vaccine reassortant was done along with A/Len/wt virus (control) and a strain 2/8 which was isolated from a child on day 8 after vaccination. As seen in Fig. 3, the vaccine reassortant A/Zak(R) and isolate 2/8 both conserved all investigated mutations in their internal genes. None of the PCR-amplified DNAs of these two strains was cleaved by restriction enzymes, in contrast to the corresponding DNAs of the A/Len/wt virus. Treatment of PB2 DNA with Tru9I verified the conservation of the G-1459-T substitution in this gene as well (not shown). These findings confirm our sequence analysis of Novgorod isolates (Klimov et al., 1994) and

A.I. Klimoc,, N.J. Cox / Journal of Vwological Methods 52 (1995) 41-49

46

PB2 (Mvn Llwt 210 --f-+---t

I) Al2

PBl

NP (Bgl II) Llwt

218

Al2

(Hind III)

Liwt 210 -+-+-+

AIZ

PA (Sea I) Llwt 218 -+-+-_$

M (Fok I) Llwt 218 --f-+--f

A/Z

A/Z

NS (Nsi I) Llwt 218 -+-+-+

AIZ

Fig. 3. Restriction analysis of art-associated mutations in genomes of A/Len/ wf virus, A/Zak(R) (H3N2) vaccine (A/Z) and isolate 2/g. I pl (2-S U) of Nsil, FokI, BglII, ScaI, Hind111 or MwrI was added directly to 10 ~1 of PCR-amplified DNAs NS (673~830), M (785-lOOl), NP (886-12001, PA (900-1218) PB1 (740-1044) and PB2 (1374-1614), respectively. After incubation for 3 h at 37°C (along with control DNAs), electrophoresis in 1.5% agarose gel (containing 0.5 pg/ml of ethidium bromide) was carried out. -, control DNA; + , DNA treated with the restriction enzyme; Hue111 digested @X174 RF DNA was used as a molecular weight marker (outside slots).

demonstrate the high degree of genetic stability of A/Len/47-based cines, even after 8 days of replication in the vaccinee. 3.2. Analysis of genome composition

of lice attenuated

ca reassortant

reassortant

vac-

vaccine

The approach described above is based on the existence of natural restriction sites in all internal genes except the NS gene of A/Len/wt virus compared with the A/Len/47 ca donor. For the NS gene, such a restriction site is created artificially using a specially designed reverse PCR primer. This approach can also be used for determining the genome composition of reassortants between the A/Len/47 master strain and other viruses, unless those other viruses have the same restriction sites as A/Len/wt virus. The GenBank sequences and our own experience have shown that the MunI, HindIII, and FokI restriction sites exist in the PB2, PBI, and M genes of all wt viruses sequenced to date. However, the PA genes of modern epidemic strains have lost the ScaI recognition site near position 1045. For example, the recent epidemic strain

A.I. Klimou, NJ. Cox / Journal of Virological Methods 52 (I 995) 41-49

PB2 (Tru 91) T HI 47 H3 Z

PBl (Hind III) T HI 47 H3 Z

PA (Asn I) T HI 47 H3 Z

NP (Eco RI) T HI 47 H3 Z

M (Fok I) T HI 47 H3

NS (Nsi I) T HI 47 H3 Z

Z

47

Fig. 4. Restriction analysis of the genome composition of reassortants A/Zak(R) (H3) and A/Taiw(R) (Hl) between A/Len/47 (L) and A/Zakarpatje/354/39 (H3N2) (Z) and A/Taiwan/l/86 (HlNl) (T) viruses, respectively. 1 ~1 (2-S U) of NsiI, FokI, EcoRI, Asnl. Hind111 or Tru91 was added directly to 10 ~1 of PCR-amplified DNAs NS (673-830), M (785lOOl), NP (886-1200), PA (900-1077), PBl (740-1044) and PB2 (1374-16141, respectively. After incubation for 2 h at 37°C electrophoresis in 1.5% agarose gel (containing 0.5 pg/ml of ethidium bromide) was carried out. HaeIII digested +X174 RF DNA was used as a molecular weight marker (left side slots).

A/Texas/36/91 (HlNl) has the sequence 1044-GGTACT-1049 (Kbmov, unpublished data), which is different from the sequence AGTACT in the A/Len/wt virus (ScaI site) and from the sequence ATTACT in the PA gene of both A/Len/ca master strains. Thus, the ScaI restriction enzyme is unable to differentiate between PA genes of A/Len/cu and recent epidemic viruses. This observation made it necessary to design reverse primer PArc-1077a for determining the origin of the PA gene of reassortants between A/Len/cu and other strains. This primer creates an AsnI restriction site during PCR if a strain has T-1045 (A/Len/ca viruses), but not G-1045 (A/Len/wt and other wt viruses) (Fig. 2). Another approach to the PCR restriction analysis of the origin of reassortants’ genes is based on the existence of restriction sites unique for all A/Len variants (A/Len/w& A/Len/l7, A/Len/47). A unique EcoRI restriction site (913-GAATTC-918) exists in the NP gene of our A/Len strains (GAC’ITC in all other viruses) (Fig. 2) and we used this restriction enzyme to determine the genome composition, since it can be used for both the A/Len/47 and A/Len/l7 donors. Note that the A/Len/l7 ca strain has no &t-related mutations in the NP gene (Kbmov et al., 19921, and hence BglII (Fig. 1)

48

A.I. Klimoc~. N.J. Cox/Journal

of Virological Method.7 52 (19951 41-49

could not differentiate between NP genes of this donor and other viruses. In addition, a unique restriction site for Tsp5091 (864-AATT-867) exists in the M gene of all A/Len variants (864-ATTT-867 in other viruses) (Fig. 2), and therefore this restriction enzyme could be used for the genome composition analysis. The complete set of PCR primers and restriction enzymes suitable for the genome composition analysis of A/Len/cabased reassortants is shown in Fig. 2. For illustration, this set of primers was used to analyse the genome composition of two reassortant vaccine strains used in the Novgorod vaccine study: A/Zak(R) (H3N2) (see above) and A/Taiw(R) (HlNl). The latter strain was obtained by reassortment of epidemic virus A/Taiwan/l/86 (HlNl) with the A/Len/47 donor (Rudenko et al., 1993). Comparison of restriction patterns for parent and reassortant viruses demonstrates (Fig. 4) that both vaccine reassortants inherited all internal genes from the ca parent. It should be noted that our PBl primers were initially designed for the A/Len viruses and that they do not work well with non-H2N2 influenza virus strains because of the considerable sequence diversity between the PBl genes of H2N2 and non-H2N2 virus subtypes (Kawaoka et al., 1989). These data (Fig. 4) confirm previous results obtained using the RNA:RNA hybridization technique, concerning the genome composition of live attenuated reassortant vaccines used in the Novgorod study (Rudenko et al., 1993).

4. Discussion The method described here combines the high specificity of restriction analysis with the high sensitivity of PCR and yet is technically quite simple. Very small amounts of infectious viruses are required for the test (the quantity of RNA isolated from 100 ~1 of infected allantoic or tissue culture fluid with a haemagglutination titre 1:4-1:8 or less is enough for the analysis of all the genes). This is especially advantageous during the initial stages of screening candidate vaccine strains, when many reassortants must be analysed. The restriction maps of the DNAs amplified by PCR are clearly registered after electrophoresis in routine 1.5-1.7% agarose gel, even for small (150-200 nucleotides) DNA fragments (see Figs. 3 and 4). Using this method, it is possible not only to determine the genome composition of reassortant viruses (in this case ca live attenuated reassortants), but also to determine if mutations related to attenuation (or other properties) are conserved during preparation of reassortants and during their replication in vaccinees. This approach allows the analysis of the genome composition of virtually any influenza virus reassortant on the basis of the existence of unique nucleotide differences in genes between any two strains. In particular, specific sets of primers and restriction enzymes could be designed for the genome composition analysis of reassortant strains obtained with the high-yielding A/PR/8/34 (HlNl) virus (candidates for inactivated influenza vaccines) or attenuated A/Ann Arbor/6/60-ca strain (candidates for live attenuated vaccines), since both these donor viruses have unique nucleotide substitutions in all the genes. Further, this approach can be easily adjusted for surveillance of influenza viruses allowing, in particular, for detection of natural reassortants between

A.I. Klimou, N.J. Cox / Journal of Vimlogical Methods 52 (1995) 41-49

human epidemic strains of A(HlN1) and A(H3N2) subtypes non-human (e.g., avian, swine, equine) influenza viruses.

or between

49

human

and

Acknowledgements We thank Prof. G.I. Alexandrova and her colleagues (Institute for Experimental Medicine, St. Petersburg, Russia) for providing us with attenuated reassortant vaccine strains and reisolates from vaccinees. This study was conducted in part under the auspices of the Joint Health Sciences Agreement between Russia and the USA in the area of viral infections other than AIDS.

References Ghendon, Y., Klimov, A., Blagoveshenskaya, 0. and Genkina, D. (1979) Investigation of recombinants of human influenza and fowl plague viruses. J. Gen. Virol. 43, 183-191. Ghendon, Y.Z., Klimov, A.I., AIexandrova, G.I. and Polezhaev, F.I. (1981) Analysis of genome composition and reactogenicity of recombinants of cold-adapted and virulent virus strains. J. Gen. Virol. 53, 215-224. Hay, A.J., Bellamy, A.R., Abraham, G., Skehel, J.J., Brand, C.M. and Webster, R.G. (1977) Procedure for characterization of genetic material of candidate vaccine strains. Dev. Biol. Stand. 39, 15-24. Kawaoka, Y., Krauss, S. and Webster, R. (1989) Avian-to-human transmission of the PBl gene of influenza A viruses in the 1957 and 1968 pandemics. J. Virol. 63, 4603-4608. Kendal, A.P., Maassab, H.F., Alexandrova, G.I. and Ghendon, Y.Z. (1981) Development of cold-adapted recombinant live, attenuated influenza vaccines in the U.S.A. and U.S.S.R. Antivir. Res. 1, 339-365. Klimov, AI., Cox, N.J., Yotov, W.V., Rocha, E., Alexandrova, G.I. and Kendal, A.P. (1992) Sequence changes in the live attenuated, cold-adapted variants of influenza A/Leningrad/l34/57 (H2N2) virus. Virology 186, 795-797. Klimov, AI., Egorov, A.Y., Gushchina, MI., Medvedeva, T.E., Gamble, W.C., Rudenko, L.G., Alexandrova, G.I. and Cox, N.J. (1994) Genetic stability of cold-adapted A/Leningrad/l34/47/57 (H2N2) influenza virus genes: sequence analysis of live cold-adapted reassortant vaccine strains before and after replication in children. J. Gen. Virol., in press. Palese, P. and Shulman, J.L. (1976) Mapping of the influenza virus genome: identification of the hemagglutinin and neuraminidase genes. Proc. Natl. Acad. Sci. USA 73, 2142-2146. Ritchey, M.B., Palese, P. and Shulman, J.L. (1976) Mapping of the influenza virus genome: identification of genes coding for nucleoproteins, membrane protein, and nonstructural protein. J. Virol. 20, 307-313. Rudenko, L.G., Slepushkin, A.N., Monto, AS., Kendal, A.P., Grigorieva, E.P., Buttseva, E.P., Rekstin, A.R., Beljaev, A.L., Bragina, V.E., Cox, N.J., Ghendon, Y.Z. and Alexandrova, G.I. (1993) Efficacy of live attenuated and inactivated vaccines in schoolchildren and their unvaccinated contacts in Novgorod, Russia. J. Infect. Dis. 168, 881-887. Scholtissek, C. (1984) Genetic relatedness of influenza viruses (RNA and protein). In: P. Palese and D.W. Kingsbery (Eds.), Genetics of Influenza Viruses. Springer, Vienna, pp. 99-126. Wright, P.F. and Karzon, D.T. (1987) Live attenuated influenza vaccines. Prog. Med. Virol. 34, 70-88. Zhdanov, V.M. (1986) Live influenza vaccines in USSR: development and practical application. In: A. Kendal and P. Patriarca (Eds.), Options for the Control of Influenza: UCLA Symposia on Molecular and Cellular Biology No. 36, A.R. Liss, New York, pp. 193-205.