The role of genetic and molecular characterization of viruses in relation to influenza surveillance and epidemiology: a WHO Memorandum*

Bulletin of the World Health Organization, 59 (6): 875-879 (1981) The role of genetic and molecular characterization of viruses in relation to influenza surveillance and epidemiology: a WHO Memorandum* A number ofrecently developed analytical techniques have been used in many labora- tories in attempts to define more fully the genetic and antigenic structures of influenza viruses. The applicability ofsome of these techniques and the value of the results obtained by their use are discussed in this Memorandum. The techniques considered range in complexity from comparative polyacrylamide gel electrophoresis of virus RNAs to deter- mination ofthe nucleic acidsequences ofthe virusgenes and the amino acidsequences ofthe corresponding proteins. The isolation and identification of influenza A and B viruses is the major function of the WHO inter- national influenza surveillance programme. Viruses isolated in national influenza centres are identified locally in most cases with reagents supplied by the World Health Organization. Any suspected influenza virus that cannot be readily identified is potentially a member of a new subtype or a variant with epi- demiological significance and is sent immediately to one of the two WHO Collaborating Centres for Reference and Research on Influenza in London and Atlanta for further antigenic analysis. At the same time national laboratories make every effort to obtain as much data as possible about the circumstances of such virus isolations, including epidemiological and serological evidence of human infection. Viruses received in the WHO collaborating centres will usually be characterized by traditional methods such as haemagglutination-inhibition and neuramini- dase-inhibition with a range of specific antisera. However, in addition, there is now available a wide range of molecular techniques that can be used for further characterization of selected isolates. This Memorandum discusses the relative merits and ease of applicability of some of these analytical procedures and assesses the significance for surveil- lance programmes of the information they reveal. POLYACRYLAMIDE GEL ELECTROPHORESIS OF VIRION RNAS Influenza viruses of types A and B contain eight RNA genome segments which carry the information for at least eight virus proteins designated P1, P2, P3, * This Memorandum was drafted by the signatories listed on page 878, on the occasion of a WHO meeting held in Geneva on 17- 19 February 1981. A French translation will appear in a later issue of the Bulletin. Requests for reprints should be addressed to Chief, Virus Diseases, World Health Organization, 1211 Geneva 27, Switzerland. H (haemagglutinin), NP (nucleoprotein), N (neur- aminidase), MP (matrix protein), and NS (non- structural proteins). Polyacrylamide gel electrophoresis of the virion RNA (vRNA) segments is a valuable tool in compar- ing influenza A and B viruses. The RNA for these experiments may be labelled by growing virus in tissue culture in the presence of 32p_ or 3H-uridine, or may be labelled in vitro with 125I or by enzymatic addition of 32p to the 5' or 3' termini; an alternative way of identifying influenza virus RNAs after electrophor- etic separation is to stain with ethidium bromide. These methods have been reviewed by Palese (1). Comparison of the RNA-migration patterns of different viruses often reveals differences in the migration rates of corresponding RNA segments. This makes it possible to determine fairly rapidly the genomes of recombinant viruses derived from known parents by simply comparing the migration rates of the RNAs from the recombinants with those of the RNAs from the parents. Differences in migration rates between equivalent RNA segments of different viruses probably depend on conformational differ- ences between the RNAs and therefore may not reflect the degree of sequence relatedness of different RNA segments. Electrophoretic analysis of vRNA is thus unlikely to yield conclusive evidence about the genetic similarity of field isolates. This has been confirmed on several occasions when closely related viruses have been observed to have different vRNA-migration patterns. Therefore, the value of the procedure for epidemiological studies lies in the demonstration of possible resemblances of unusual isolates to labora- tory reference viruses and in identification of recom- binant viruses, which can then be confirmed by more definitive methods. Electrophoresis is also used for the isolation ofRNA segments for use as reagents and probes in nucleic acid hybridization, oligonucleotide mapping, or sequencing procedures. 4129 -875 - WHO MEMORANDUM PROTEIN COMPARISON BY ELECTROPHORETIC AND CHROMATOGRAPHIC PROCEDURES Influenza virus proteins may be compared by poly- acrylamide gel electrophoresis (PAGE) in the pres- ence of the ionic detergent sodium dodecyl sulfate (SDS). When the method was originally applied with detergent-disrupted viruses there were some discrep- ancies in the findings with different strains studied in different laboratories. These, however, have been largely eliminated and higher resolution has been attained by the use of discontinuous buffer systems (2). Difficulties are still encountered, however, in resolving the P1, P2, and P3 components, in identify- ing the neuraminidase polypeptide, and in detecting the small NS2 polypeptide. Although the different polypeptides are dis- tinguished in SDS-PAGE primarily on the basis of differences in relative molecular mass, relative migration rates can be affected by factors such as acrylamide concentration or the type of buffer system used. This implies that polypeptide charge or confor- mation may also affect migration rates and because these properties may be affected by relatively few changes in amino acid sequences, or changes in the number or location of glycosylation sites, comparison of relative migration rates of polypeptides for different virus isolates is unlikely to be a reliable way of comparing their structural similarity. However, SDS-PAGE is valuable in comparative studies for indirect determination of the genome composition of laboratory-derived recombinants (also referred to as reassortants) and for isolation of polypeptides for additional comparison by peptide mapping. Although analysis of the genome of recom- binants is not generally undertaken in influenza surveillance programmes, the occurrence of natural recombination between influenza A(H1N1) and (H3N2) strains during their co-circulation in 1978 - 79 has made this an area of interest. A simple method such as this for genotyping isolates may there- fore be useful. Similarly, the method may be useful for the initial comparison of prototype laboratory viruses with field isolates when there is a possibility of laboratory con- tamination, because the procedure requires less sophisticated resources than RNA analysis. Peptide mapping studies can be facilitated by the use of SDS-PAGE to isolate proteins. Either stained, or radiolabelled virus-specific polypeptides identified by SDS-PAGE, can be recovered in a semi-purified state by excising the gel fragment containing the poly- peptide, and may then be partially or completely digested with proteolytic enzymes either in situ or after elution from the gel. Peptides obtained by this or other methods may be analysed by procedures such as high-performance liquid chromatography, reverse- phase column chromatography, and paper or thin- layer two-dimensional electrophoresis and chromato- graphy. Peptide mapping procedures, particularly those involving complete, rather than partial proteolytic digestion, produce results that directly relate to amino acid sequence and have, for example, been useful in comparing the haemagglutinins of antigenically different viruses (3). Advantages of the procedures include (a) the ability to compare a large portion of any polypeptide (all the soluble peptides), (b) the possibility of comparing proteins without the use of radioactive isotopes, for example, in the case of polypeptides such as H, NP, and MP, which can be purified in large amounts from egg-grown viruses and then detected by staining. Disadvantages include the difficulty of obtaining certain polypeptides, particularly P1, P2, and P3, in sufficient purity and quantity. NUCLEIC ACID HYBRIDIZATION Hybridization is a rapid method for the evaluation of base sequence homologies between virus RNAs. Four different procedures have been used. Electrophoresis of double-stranded RNA Labelled complementary RNA (cRNA) from cyclo- heximide-treated and infected cells is hybridized to non-labelled virion RNA (vRNA) followed by treatment of the cRNA/vRNA double-stranded hybrids with S1-nuclease and analysis by polyacryl- amide gel electrophoresis. This allows the detection of very slight differences, i.e., point mutations, in the genes under comparison. In case of complete hom- ology of vRNA and cRNA molecules, the double- stranded complexes are completely resistant to, and their electrophoretic mobilities are not affected by S1- nuclease treatment. If the homology is incomplete the electrophoretic mobility of the hybridized segment may change or alternatively the segment may not be detectable at all (4, 5). Direct RNA - RNA hybridization Non-labelled cRNA is extracted from cells infected with influenza virus and is used to protect labelled vRNA segments, by means of hybridization, against digestion by RNase. In the case of homologous hybridization protection approaches 100%; when hybridization is heterologous the degree of protection depends on the extent of base sequence homology as well as on the distribution of homologous sequences in the RNA molecules (6). 876 GENETIC AND MOLECULAR CHARACTERIZATION OF INFLUENZA VIRUSES Competitive RNA - RNA hybridization Radiolabelled vRNA is hybridized with hom- ologous cRNA isolated from infected cells. The hom- ologous vRNA is displaced by heating the double- stranded RNA with an excess of non-labelled vRNA (7). As in previous methods the extent of competition depends on the sequence homology and its molecular distribution. DNA -RNA hybridization Radiolabelled cDNA is prepared by transcribing virus RNA using reverse transcriptase in the presence of actinomycin D. The labelled single-stranded cDNA is hybridized with non-labelled vRNA and single- stranded regions are digested by S1-nuclease. Applications of these hybridization methods have included the demonstration of the close genetic relatedness of H IN 1 viruses isolated in 1950 and 1977, and estimation of the degree of relatedness of viruses of different subtypes. All four methods have been used to compare the genetic composition of influenza viruses, but they may all miss point mutations and only the last three procedures yield quantitative data. The base sequence homologies determined may be underestimates when compared with those deduced from sequence data. As a rule, the results obtained by DNA- RNA hybridization are lower than those of RNA - RNA hybridization. OLIGONUCLEOTIDE MAPPING Oligonucleotide mapping (RNA fingerprints) of radiolabelled RNAs involves specific enzymatic cleavage of the RNA segments followed by two- dimensional separation of the cleavage products (8). The radiolabelling may be done in vitro to high specific activity. Limitations of the technique include the fact that the analysis focuses on the identification of unique oligonucleotides that contain a minimum of about 7 bases. Consequently, only a small part of the RNA is represented in the oligonucleotides that are com- pared. Another limitation of the technique arises when two RNAs of less than about 900/o sequence homology are compared. In such cases, and if the changes are distributed throughout the genomes, the oligonucleotide maps of the two RNAs will share few, if any, unique oligonucleotides despite the existence of up to about 900o base-sequence homology (9). Applications of the method include the analysis of the relationship between genomes of closely related viruses (10). For example, it was shown by oligo- nucleotide mapping that the eight RNAs tested that were obtained from the 1977 isolates of the HI NI sub- type were all very similar to those of strains circulating in 1950 (11) and that some viruses of the HINI subtype isolated since 1977 were recombinants between viruses of the HlNI and H3N2 subtypes (12). NUCLEIC ACID AND PROTEIN SEQUENCING The most definitive procedures for comparing influenza viruses involve determination of the nucleic acid sequences of the virus genes and the amino acid sequences of the corresponding proteins. By classical procedures, extensive amino acid sequences of rep- resentative haemagglutinins from the H2 and H3 subtypes have been determined (13, 14). Analyses of the terminal nucleotide sequences of virus RNAs have been made directly by procedures involving partial digestion of terminally labelled RNAs by specific nucleases, or cleavage of terminally labelled RNAs following limited chemical modi- fication. In addition, the complete nucleotide sequences of several virus RNAs have been derived either by sequencing DNA copies of the genes cloned in bacteria or by the use of specific fragments of such DNA as primers in nucleotide sequence analyses involving DNA synthesis. The two methods of DNA sequencing used involve either cleavage of specifically modified terminally labelled DNA molecules and separation of the resulting fragments by polyacryl- amide gel electrophoresis (15), or the use of DNA polymerase, single-stranded DNA primers, and specific chain-terminating inhibitors to generate similar catalogues of oligodeoxynucleotides of differ- ent length which again are analysed by polyacrylamide gel electrophoresis. In addition to complete nucleo- tide sequences, terminal sequence data have also been obtained by using modifications of the latter method. Using these procedures, the complete nucleotide sequences of the haemagglutinin genes of four sub- types (H 1, H2, H3, and H7) have been determined including three representatives of the H3 subtype (16). Similarly, the nucleotide sequences of the seventh and eighth largest virus RNAs, which code for theMP and NS proteins, respectively, have been determined from viruses of the H I subtype in the former case and from three viruses of the H1, H3, and H7 subtypes for the latter. Clearly, although providing definitive information, nucleic acid sequencing requires the application of several techniques and is more complex than the other procedures described. 877 878 WHO MEMORANDUM ANTIGENIC ANALYSIS USING MONOCLONAL ANTIBODIES The ability to clone antibody-producing cells pro- vides the opportunity to use monoclonal antibodies in the detailed analysis of influenza viruses by individual antigenic sites, and such preparations have been used to characterize the H, N, and NP antigens of a number of viruses (17). For the past few years, monoclonal antibodies have been evaluated by the WHO Collaborating Centres for Reference and Research on Influenza in London and Atlanta for the identification of influenza viruses, by comparing them with post-infection ferret anti- sera. At present they are not used as a primary means of strain characterization, but selected preparations may be useful as an adjunct to specific animal antisera, particularly for the differentiation of very closely related variants. CONCLUSIONS In research on the structure and replication of influ- enza viruses there is no doubt that application of the procedures described here have been of enormous value; of particular interest have been the determi- nations of the total nucleotide sequences of haemag- glutinin genes, as these have provided the basis for an understanding of antigenic variation. Similarly, these methods have been of immediate epidemiological value in determining the genetic composition of the viruses of the HINI subtype that reappeared in 1977 and in comparing it with that of the 1950 strains, in detecting reassorted genomes in the HINI isolates of 1978-79, and in demonstrating the relatedness of viruses isolated in the 1933 - 57 period, which are all now recognized as members of the HINI subtype. Such applications are expected to continue. Apart from the nucleic acid and protein sequencing procedures, all the methods described may be used to compare the genetic composition of virus isolates with that of prototype strains and they all require consider- able experience both for their performance and for interpretation of the results. For routine comparison, the procedures involving polyacrylamide gel electro- phoresis of the specific polypeptides of virus particles or infected cells and of virion RNAs may be the simplest to apply. However, the information they give is often of limited value. Similarly, comparisons involving the analysis of RNA -RNA hybrids by polyacrylamide gel electrophoresis or involving oligo- nucleotide mapping can indicate differences between the RNAs of viruses being compared but the infor- mation obtained is qualitative in nature. Qualitative data can be obtained by the other nucleic acid hybrid- ization methods described. As a consequence of these limitations, different combinations of the methods are used for the analysis of viruses of interest, according to the experience of the investigators concerned and the availability of reagents. The WHO Collaborating Centres for Reference and Research on Influenza in London and Atlanta have been active in the application of molecular tech- niques to the WHO influenza surveillance pro- gramme. It is hoped that publication of this Memor- andum will stimulate further exchange of infor- mation, viruses, and reagents between the WHO collaborating centres and the individual investigators working on these genetic and molecular methods. The value of such exchanges both to theWHOprogramme and to the scientists working in this field cannot be overestimated. * * F. Assaad, Virus Diseases, World Health Organiz- ation, Geneva, Switzerland P. Bres, Virus Diseases, World Health Organization, Geneva, Switzerland F. Ennis, Bureau of Biologics, National Institutes of Health, Bethesda, MD, USA K. Esteves, Virus Diseases, World Health Organiz- ation, Geneva, Switzerland Y. Z. Ghendon, Laboratory of Genetics and Bio- chemistry of Viruses, Institute of Viral Prepar- ations Research, Moscow, USSR A. P. Kendal, Deputy Director, WHO Collaborating Centre for Reference and Research on Influenza, Centers for Disease Control, Atlanta, GA, USA P. Palese, Department of Microbiology, Mount Sinai School of Medicine, The City University of New York, New York, NY, USA M. S. Pereira, Co-Director, WHO Collaborating Centre for Reference and Research on Influenza, Virus Reference Laboratory, Central Public Health Laboratory, London, England F. Perkins, Biologicals, World Health Organization, Geneva, Switzerland G. C. Schild, Director, WHO Collaborating Centre for the Standardization of Viral Products, Division of Viral Products, National Institute for Biological Standards and Control, London, England C. Scholtissek, Institut fiUr Virologie der Justus- Liebig-Universitat, Giessen, Federal Republic of Germany J. Skehel, Co-Director, WHO Collaborating Centre for Reference and Research on Influenza, National Institute for Medical Research, Division of Virology, Mill Hill, London, England GENETIC AND MOLECULAR CHARACTERIZATION OF INFLUENZA VIRUSES 879 REFERENCES 1. PALESE, P. The genes of influenza virus. Cell, 10: 1 - 10 (1977). 2. LAEMMLI, U. K. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London), 227: 680-685 (1970). 3. LAVER, W. G. ET AL. Studies on the origin of pandemic influenza. II. Peptide maps of the light and heavy poly- peptide chains from the hemagglutinin subunits of A2 influenza viruses isolated before and after the appear- ance of Hong Kong influenza. Virology, 48: 445 - 455 (1972). 4. HAY, A. J. ET AL. Transcription of the influenza virus genome. Virology, 83: 337 - 355 (1977). 5. GHENDON, Y. ET AL. Genome analysis of influenza virus strains isolated during an epidemic of 1979- 1980. Journal ofgeneral virology, (in press, 1981). 6. SCHOLTISSEK, C. ET AL. Correlation between RNA frag- ments of fowl plague virus and their corresponding gene functions. Virology, 74: 332 - 344 (1976). 7. BEAN, W. J. JR ET AL. Recombination of human influenza A viruses in nature. Nature (London), 284: 638 -640 (1980). 8. DE WACHTER, R. & FIERS, W. Preparative two-dimen- sional polyacrylamide gel electrophoresis of 32P-labeled RNA. Analytical biochemistry, 49: 184 - 97 (1972). 9. YOUNG, J. F. ET AL. Advantages and limitations of the oligonucleotide mapping technique for the analysis of viral RNAs. In: D. H. L. Bishop & R. W. Compans, ed., Replication of negative strain viruses. New York, Academic Press, 1981, pp. 209 - 215. 10. ORTIN, J. ET AL. Genetic variability of Hong Kong (H3N2) influenza viruses: spontaneous mutations and their location in the viral genome. Gene, 11: 319 - 331 (1980). 11. NAKAJIMA, K. ET AL. Recent human influenza A(H IN I) viruses are closely related genetically to strains isolated in 1950. Nature (London), 274: 334 - 339 (1978). 12. YOUNG, J. F. ET AL. Evolution of human influenza A viruses in nature: sequential mutations in genomes of new HlI1. Cell, 18: 73 - 83 (1979). 13. WARD, C. W. & DOLPHEIDE, T. A. The Hong Kong (H3) haemagglutinin. Complete amino acid sequence and oligosaccharide distribution for the heavy chain of A/Memphis/102/72. In: W. G. Laver&G. M. Air, ed., Structure and variation in influenza virus. Amsterdam, Elsevier, 1980, pp. 27 - 38. 14. WATERFIELD, M. ET AL. Disulphide bonds of haemag- glutinin of Asian influenza virus. Nature (London), 289: 422-424 (1981). 15. MAXAM, A. M. & GILBERT, W. A new method for sequencing DNA. Proceedings oftheNationalAcademy of Sciences of the United States of America, 74: 560- 564 (1977). 16. LAVER, W. G. & AIR, G. M., ed., Structure and vari- ation in influenza virus. Amsterdam, Elsevier, 1980. 17. WEBSTER, R. G. ET AL. Analysis of antigenic drift in recently isolated influenza A(HINI) viruses using monoclonal antibody preparations. Virology, 96: 258 - 264 (1979).