VIROLOGY 221, 291–300 (1996) ARTICLE NO. 0378

Terminal Region Sequence Variations in Variola DNA

View metadata, citation and similar papers at core.ac.uk brought to you by CORE ROBERT F. MASSUNG,* VLADIMIR N. LOPAREV,* JANICE C. KNIGHT,* ALEXEI V. TOTMENIN,† provided by Elsevier - Publisher Connector VLADIMIR E. CHIZHIKOV,*,† JOSEPH M. PARSONS,* PAVEL F. SAFRONOV,† VALERY V. GUTOROV,† SERGEI N. SHCHELKUNOV,† and JOSEPH J. ESPOSITO*,1

*Poxvirus Section, Division of Viral and Rickettsial Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Public Health Service, U.S. Department of Health and Human Services, Atlanta, Georgia 30333; and †Department of Molecular Biology of Genomes, Institute of Molecular Biology, State Research Center of Virology and Biotechnology, ‘‘Vector,’’ Koltsovo, Novosibirsk Region, Russia 633159 Received January 24, 1996; accepted May 23, 1996

Genome DNA terminal region sequences were determined for a Brazilian alastrim variola minor virus strain Garcia-1966 that was associated with an 0.8% case-fatality rate and African strains Congo-1970 and Somalia-1977 associated with variola major (9.6%) and minor (0.4%) mortality rates, respectively. A base sequence identity of §98.8% was determined after aligning 30 kb of the left- or right-end region sequences with cognate sequences previously determined for Asian variola major strains India-1967 (31% death rate) and Bangladesh-1975 (18.5% death rate). The deduced amino acid sequences of putative proteins of §65 amino acids also showed relatively high identity, although the Asian and African were clearly more related to each other than to alastrim virus. Alastrim virus contained only 10 of 70 proteins that were 100% identical to homologs in Asian strains, and 7 alastrim-specific proteins were noted. ᭧ 1996 Academic Press, Inc.

INTRODUCTION strains in Africa to account for some of the low-range fatality rates. The variola smallpox virus was a strictly Interestingly, in the final eradication years in South human pathogen. During the smallpox eradication cam- Africa, and in Ethiopia and Somalia, where vaccine cover- paign, which culminated with the last naturally occurring age was minimal, isolates associated with alastrim-like case in Somalia in 1977, variola strains of graded patho- fatality rates were recovered. However, certain isolates genic and biologic features in laboratory tests were iso- tested were closely related to variola major virus by lated from human clinical samples. However, isolates assays such as reproductive ceiling temperature in cell had been characterized broadly as major or minor on culture, hemadsorption assay, and DNA restriction assay the basis of epidemiologic criteria, particularly the asso- (Dumbell and Huq, 1986; Nakano and Esposito, 1989); ciated case-fatality rate. such isolates were proposed ‘‘intermediate’’ strains (Bed- Variola major virus occurred worldwide, with mortality rates ranging from about 10 to 40%; rates in Asia were son et al., 1963; Dumbell and Huq, 1975, 1986). often higher than those in Africa. Variola minor outbreaks DNA sequencing, restriction maps, and mutation anal- had death rates of õ1%, and the frequency was typically yses have confirmed that the genome DNA of orthopoxvi- associated with isolation of alastrim virus, a subtype dis- ruses has a conserved central region of about 100 kilo- tinctive by several biologic tests. Alastrim predominated base pairs (kb) flanked by more varied terminal region in South America from the late 1800s and spread mainly sequences. The central region mainly encodes proteins through the Americas, Europe, and Australia (Chapin, essential for virus growth, and the terminal regions en- 1913; Chapin and Smith, 1932). On the other hand, partic- code proteins involved in virus survival, such as determi- ular well-defined outbreaks of smallpox in recent African nants of host range, transmissibility, and for modulating history showed mortality rates that ranged from 0.4 to the host immune response. Recently, this distribution 15% (Fenner et al., 1988); but, it is not fully resolved was confirmed for variola virus genes by (1)resolution of whether alastrim virus sometimes circulated with major the complete 186,103-base pair (bp) genome DNA se- quences of an isolate from the last infected person in the final outbreak of variola major in Bangladesh in 1975, Nucleotide sequence data from this article have been deposited which had 18.5% associated fatalities, and by (2) determi- with the GenBank Data Library under Accession Nos. U18337, U18338, nation of the complete 185,578-bp DNA coding region U18339, U18340, and U18341. 1 To whom correspondence and reprint requests should be ad- sequences of an isolate from a 1967 outbreak in India dressed. Fax: (404) 639-0049. E-mail: [email protected]. associated with a peak death rate that year of 31% for

0042-6822/96 $18.00 291 Copyright ᭧ 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.

AID VY 8025 / 6a1b$$$$81 06-14-96 22:36:02 viras AP: Virology 292 MASSUNG ET AL. the 84,902 smallpox victims identified (Fenner et al., 1988; Massung et al., 1993, 1994; Shchelkunov et al., 1993a,c, 1994a,b). In another report, we directly compared base sequences, open reading frames, and deduced amino acid sequences of proteins of the Bangladesh-1975 and India-1967 strains (Shchelkunov et al., 1995), including certain relevant features of sequences of the smallpox vaccine, virus (Goebel et al., 1990; Johnson et al., 1993). Because Bangladesh-1975 (BSH) and India-1967 (IND) are major strains isolated within a short time frame, it FIG. 1. Variola virus genome DNA HindIII maps. (A) Asian and African is reasonable to assume that the outbreaks in these major and African minor viruses have the same HindIII map compared neighboring nations derive from the same epidemic. with (B) alastrim virus which shows two more HindIII sites (Esposito Therefore, the sequence data from BSH and IND are and Knight, 1985; Mackett and Archard, 1979). The shaded SacI and limited in terms of assessing the extent of variation BstEII fragments on the map in (A) represent genome regions cloned among apparently different strains and identifying ge- for DNA sequences presently determined for the GAR, SOM, and CNG left-end and the GAR and SOM right-end. The black bars on the map netic elements that could confer high- or low-virulence in (B) represent alastrim-specific DNA segments. attributes. To further explore the spectrum of conservation and variation of strains of varying grades of virulence and to cent DNA reaction products were separated and col- gain insight into some of the genes contributing to the lected using an Applied Biosystems, Inc., Model 373A differences, we determined a total of Ç150 kb of DNA sequencer according to the manufacturer’s protocols. sequences that comprise a significant portion of the ge- Ambiguities and other problematic nucleotides were re- nome terminal regions of three more smallpox virus solved by sequencing directly from amplified genome strains: (1) Congo-1970, associated with a 9.6% mortality DNA using methods suggested by Applied Biosystems, rate; (2) Somalia-1977, from the last natural case of small- Inc. Raw nucleotide sequence data were edited and as- pox in the world related to a 0.4% mortality rate; and (3) sembled to achieve a minimum twofold redundancy for Garcia-1966, from a Brazilian alastrim outbreak with a each DNA strand using the Staden software package 0.8% death rate. Recently, we described the (noncoding) (Dear and Staden, 1991). inverted terminal repeat (ITR) sequences near the ge- Other plasmid inserts containing GAR left-end SacI nome DNA hairpin-ends of these and other strains (Mas- and right-end BstEII DNA regions were sequenced at sung et al., 1995). The present report extends our studies ‘‘Vector’’ by Maxam–Gilbert methods; sequences were by comparing the coding sequences adjacent to the ITRs assembled using NWASF software as described for IND of the following viral genetic subtypes: South American genome DNA sequences (Shchelkunov et al., 1993a,c, alastrim minor, Asian major, and African major and minor 1994a,b; EMBL Accession No. X69198). viruses. Data analysis MATERIALS AND METHODS Comparison of DNA sequences, determination of po- DNA sequencing tential open reading frames and translation products, and alignments and comparisons with analog proteins were At the CDC, sequences were determined from plasmid performed using programs in the Genetics Computer inserts containing African variola minor Somalia-1977 vi- Group, Inc., software package (Devereux et al., 1984), rus (SOM) genome left-end SacI-I, -J, -F, and part of - including FASTA. Database searches for analogue pro- M DNA fragments and right-end BstEII-D and -E DNA teins were done using BLAST software (Altschul et al., fragments; African variola major Congo-1970 virus (CNG) 1990). cognates of the SacI DNA fragments; and the alastrim Garcia-1966 virus (GAR) right-end BstEII-D DNA frag- ment. Sequencing was accomplished using Taq polymer- RESULTS ase in dye-terminated dideoxynucleotide reactions per- Nucleotide sequence homology formed in a Perkin–Elmer–Cetus, Inc., Model 9600 ther- mocycler according to the manufacturer’s protocols. Figure 1A shows the HindIII map of BSH, IND, SOM, Synthetic primers for polymerase chain reaction (PCR) and CNG with the sequenced regions highlighted. We amplifications and sequencing reactions were designed also sequenced corresponding regions of GAR; however, from BSH genome DNA sequences (Massung et al., because of HindIII map differences, we include here a 1993, 1994; GenBank Accession No. L22579). Fluores- clarification on GAR DNA restriction fragment names,

AID VY 8025 / 6a1b$$$$81 06-14-96 22:36:02 viras AP: Virology SMALLPOX VIRUS TERMINAL DNA REGIONS 293

TABLE 1 Variola Virus DNA Interstrain Sequence Homology

Percentage identity

Fragment Strain—Fragment size BSH CNG SOM GAR

SacI-I; BSH—8136 bp CNG—8139 bp 99.89 — — — SOM—8139 bp 99.85 99.91 — — GAR—8944 bp 99.59 99.63 99.59 — IND—8126 bp 99.35 99.41 99.37 98.83 SacI-J; BSH—6415 bp CNG—6471 bp 99.84 — — — SOM—6455 bp 99.84 99.91 — — GAR—6425 bp 99.48 99.41 99.45 — IND—6444 bp 99.84 99.84 99.81 99.49 SacI-F; BSH—14076 bp CNG—14051 bp 99.92 — — — SOM—14085 bp 99.86 99.89 — — GAR—14272 bp 99.57 99.59 99.53 — IND–14118 bp 99.73 99.75 99.67 99.54 BstEII-E; BSH—15446 bp SOM—15498 bp 99.85 — — — GAR—16058 bpa 99.68 — 99.68 — IND—15393 bp 99.86 — 99.84 99.68 BstEII-D; BSH—15633 bp SOM—15364 bp 99.89 — — — GAR—15503 bpa 99.53 — 99.50 — IND—15821 bp 99.44 — 99.39 99.12

a DNA fragments BstEII-E and -D of BSH, SOM, and IND correlate with GAR BstEII-D and -E, respectively (see text for details). which affects GAR reading frame enumeration. Figure 1B D fragments. The overall nucleotide sequence homology, shows the HindIII map of GAR DNA, including the loca- excluding gaps for insertions and deletions, was deter- tion of 898- and 627-bp segments (described later) that mined to be §98.83%. If GAR sequences are excluded, are distinctive of GAR and other alastrim strains exam- the overall sequence homology of the African isolates ined (Knight et al., 1995). Because of the 627-bp insert, (SOM and CNG) with corresponding Asian virus (BSH restriction fragment names (which are based on fragment and IND) sequences increases to §99.37%, indicating size) change so that GAR BstEII-D, for example, is the that GAR has diverged the most. cognate of BstEII-E of SOM, BSH, IND, and CNG. Alto- For the individual fragments examined, homologies gether, the highlighted region comprises about 150 kb: ranged from 99.41 to 99.68% identity in comparing the 30 kb of the left-end of SOM, GAR, and CNG and 30 kb GAR fragment sequences with BSH, SOM, and CNG cog- of the right-end SOM and GAR. nates. Sequence homologies ranged from 98.83 to The left-end sequences begin at a BstEII site within 99.68% identity in comparing GAR and IND correlate frag- the SacI-M fragment, the hairpin-end terminal fragment of ments. The two African strains gave identities of 99.91% several variola virus strains (Esposito and Knight, 1985). for both SacI-I and SacI-J. The lowest homology was From our ITR topography determination (Massung et al., 98.83%, noted in comparing the SacI-I fragments of IND 1995), we noted that the site is within the ITR and is and GAR. Sequence variations consisted mainly of differ- located about 500 bp from the hairpin-ends of SOM and ences in a single base or small deletions/insertions, GAR DNA, and about 700 bp from the hairpin-ends of most of which occurred in noncoding sequences. Certain CNG DNA. The first 92 bases of the GAR left-end se- variances caused frameshifts that markedly altered the quences determined are part of the ITR and thus are the deduced amino acid sequences of proteins. However, inverted complement of the last 92 bases of the GAR most differences within reading frames were single nu- right-end sequences presented here. Also, the first 557 cleotide substitutions that did not alter the encoded bases of the SOM left-end are the complement of the amino acid. final 557 bases determined for the SOM right-end; the increased complementarity derives from the larger size Left-terminal region open reading frames and putative of the SOM ITR described by Massung et al. (1995). proteins Table 1 shows the sequence homology for the left-end The percentage identity of amino acid sequences pre- SacI-I, -J, and -F fragments and right-end BstEII-E and - dicted from left-end open reading frames of SOM, CNG,

AID VY 8025 / 6a1b$$$$81 06-14-96 22:36:02 viras AP: Virology 294 MASSUNG ET AL.

TABLE 2 Comparison of Proteins Coded by Variola Virus Genome Left-End Region Open Reading Framesa

BSH GAR to BSH IND to BSH SOM to BSH CNG to BSH No. aa % identity % identity ORF % identity ORF % identity

D1L 91 — 100/57 (57 aa) B0.5L 100/57 (57 aa) D0.5L 96.5/57 (57 aa) D2L 143 100/122 (153 aa) 99.2/122 (153 aa) B1L 100/122 (153 aa) D1L 100/122 (153 aa) D3L 128 100 99.2 B2L 99.1/113 (113 aa) D1.5L 100 D4R 140 100 100 B3R 99.3 D2R 97.9 D5L 330 100 100 B4L 99.7 D3L 98.5 D6R 242 100 99.6 B5R 100 D4R 100 D7L 126 100 100 B6L 99.2 D5L 99.2 D8L 452 99.6 99.6 B7L 97.4/77 (97 aa) D6L 98.9 ‘‘ ‘‘ — — B8L 98.6/357 (357 aa) ‘‘ ‘‘ D9L 91 100 100 B9L 100/91 (98 aa) D6.5L 100 — —— — B10L 172 aa — — — —— — B11R 93 aa — — D10L 152 98.0/153 (153 aa) 98.0/153 (153 aa) B12L 97.7/130 (132 aa) D7L 98.0/153 (153 aa) — —— — B13R 83 aa — — D11L 150 100 100 B14L 100 D8L 100 D12L 156 99.4 99.4 B15L 100 D9L 100 D13L 134 100 100 B16L 100 D10L 100 D14L 316 99.7 99.7 B17L 99.4 D11L 99.7 D15L 263 100 99.6 B18L 99.6 D12L 99.6 D16L 221 99.5/201 (201 aa) 99.5/201 (201 aa) B19L 98.2/110 (154 aa) D13L 99.0/201 (201 aa) D17L 79 100 100 B20L 98.3/58 (65 aa) D13.5L 100 D18L 214 98.6 99.5 B21L 97.2/214 (212aa) D14L 99.5 P1L 117 100 99.6 R1L 99.1 P1L 99.1 P2L 177 100 100 R2L 98.3 P2L 100 O1L 446 100/422 (425 aa) 100/422 (425 aa) Q1L 99.8 O1L 100 O2L 220 99.5 100 Q2L 99.5 O2L 100 O3L 70 98.6 100 Q3L 100 O3L 100 C1L 66 100 100 P1L 100 C1L 100 C2L 373 99.5 100 P2L 99.7 C2L 99.2 C3L 87 100/87 (88 aa) 100/87 (88 aa) P3L 100/87 (88 aa) C3L 100/87 (88 aa) C4R 149 99.2/133 (136 aa) 99.3 P4R 100 C4R 100 C5L 237 96.3/243 (241 aa) 96.6/237 (231 aa) E1L 99.1/223 (312 aa) C5L 92.0/251 (251 aa) C6L 147 100 100 E2L 100 C6L 100 C7L 161 100 100 E3L 98.1/156 (179 aa) C7L 100/156 (179 aa) C8L 333 99.7/319 (319 aa) 99.7/319 (319 aa) E4L 99.4/319 (319 aa) C8L 98.5 C9L 348 99.7 100 E5L 99.7 C9L 99.4 C10L 72 100 100 E6L 98.6 C10L 100 C11L 79 100 100 E7L 97.5/79 (78 aa) C11L 100 C12L 65 100 98.5 E8L 98.5 C12L 98.5 C13L 212 100 100 E9L 100 C13L 100 C14L 439 99.8 99.8 E10L 99.1 rC14L 99.5

a FASTA analysis. The numbers of deduced amino acids (aa) in proteins coded by open reading frames (ORFs) that differ in size relative to BSH correlates are shown in parentheses. Enumerations of SOM and CNG ORFs are the same as BSH and enumeration of IND ORFs follows that of Shchelkunov et al. (1995). Dash (—) indicates that the cognate ORF is absent in the sequences determined. Quote marks (‘‘) under BSH D8L and IND D6L indicate a sequence variation that leads to two ORFs in GAR DNA.

IND, and GAR are compared with BSH reading frames were identical to GAR proteins. Thus, the left-end region in Table 2. As expected from similarities of BSH, SOM, of GAR DNA encodes more diverged proteins than the and CNG HindIII maps (Fig. 1A), relatively few novel SOM other strains. and CNG reading frames were observed. Of the 37 BSH Analysis of the open reading frames revealed several left-end open reading frames, 19 coded for proteins 100% truncated or elongated polypeptides. For example, as indi- identical in amino acid content and number of residues cated in Table 2, the first reading frame in BSH DNA, D1L, to predicted proteins encoded in corresponding SOM which encodes a 91-amino-acid protein, was absent in reading frames, 18 encoded proteins identical to corre- SOM DNA; a 251-base deletion in SOM DNA accounts lates in CNG, 17 were identical to IND correlates, and 9 for the difference. The corresponding region of SOM DNA

AID VY 8025 / 6a1b$$$$81 06-14-96 22:36:02 viras AP: Virology SMALLPOX VIRUS TERMINAL DNA REGIONS 295 showed a distinguishing 464-bp insert with no open reading -E. Similar to the left-end region, BSH, IND, and SOM frames for proteins §65 amino acids. Whereas CNG, GAR, showed a high degree of open reading frame and se- and IND DNAs contain correlates of BSH D1L, a 2-bp insert quence conservation, and GAR was more varied. Four- near the terminus causes a translation frameshift to provide teen of the 23 open reading frames of SOM encoded a 57-amino-acid protein. The next reading frame, BSH D2L, proteins identical to BSH proteins, and 12 IND proteins codes for a 143-residue product; however, the correlates, were 100% identical to those of BSH. However, only two SOM D2L, CNG D2L, GAR B1L, and IND D1L, encode an GAR reading frames, H8R and D6L, encoded proteins elongated protein of 153 amino acids due to a 23-base identical to BSH correlates. Of the right-end open reading insertion. The most varied left-end reading frame was C5L frames, B15L in BSH and SOM (GAR D6L, IND B18L), (GAR E1L), which encodes a different size protein for each provides the only protein 100% conserved in all four strain examined. Although no function has yet been as- strains. cribed C5L or its counterpart vaccinia virus F1L, the se- Despite the GAR sequence variations, several open quences might be useful for developing diagnostics. An- reading frames encoded putative proteins that are rela- other varied reading frame, D8L of BSH, SOM, and CNG tively well-conserved in the four strains examined and (IND D6L), encodes a 452-amino-acid truncated version of have homologs reported for other poxviruses, including the virus 668-residue host-range protein described vaccinia virus: B6R of BSH and SOM (GAR H7R, IND B7R) by Spehner et al. (1988). Cognates of the cowpox virus encodes an envelope glycoprotein analog (Englestad et reading frame are absent or highly fragmented, and the al., 1992; Isaacs et al., 1992; Takahashi-Nishimaki et al., corresponding DNA of GAR shows a 4-base deletion, ef- 1991); B8R of BSH and SOM (GAR H9R, IND B9R) pro- fecting a frameshift that fragments the reading frame into vides an interferon-K-receptor isolog (Alcami and Smith, two smaller frames, B7L and B8L. The cowpox virus protein 1995; Mossman et al., 1995); B12R (GAR D2R, IND B13R) is involved in permitting virus growth in Chinese hamster and B21R (GAR D13R, IND B25R) code for serine prote- ovary cell cultures. The function of the protein in vivo is ase inhibitor isologs (Kotwal and Moss, 1989; Pickup et unresolved as is the role of the truncated variola homologs, al., 1986); B17R (GAR D8R, IND B20R) codes for a soluble although database searches reveal that all are isologs of interferon-I/J-receptor isolog (Smith and Chan, 1991; human ankyrin. Ueda et al., 1990); G2R of BSH, SOM, IND, and GAR Only five reading frames coded for proteins with 100% provides a TNF-receptor isolog that is truncated in vac- amino acid identity between all five strains, including D11L cinia virus strains examined to date (Upton et al., 1991, (B14L in GAR, D8L in IND), D13L (GAR B16L, IND D10L), 1992); and G3R of BSH, SOM, IND, and GAR encodes a C1L (GAR P1L), C6L (GAR E2L), and C13L (GAR E9L). Three cognate of vaccinia virus 35-kDa secreted antigen (Patel (D11L, C6L, and C13L) of the five have correlates in vaccinia et al., 1990). virus strain Copenhagen, and the apparent function of two SOM B22R and GAR D14R encode proteins with strong has been reported: BSH D11L and cognates correspond homology to a 214-kDa protein encoded by BSH B22R to C7L, a Copenhagen vaccinia virus host-range protein (IND B26R); the reading frame is absent in vaccinia (Perkus et al., 1990), and C6L corresponds to F2L, the Co- strains examined but present in certain other orthopoxvi- penhagen vaccinia virus dUTPase (Goebel et al., 1990). No ruses (manuscript in preparation). Four reading frames function has been ascribed to vaccinia virus Copenhagen are conserved that encode ankyrin-like proteins: BSH F9L, the cognate of variola C13L (GAR E9L). and SOM B5R (GAR H6R, IND B6R), BSH and SOM B16R Several variola virus open reading frames provide pro- (GAR D7R, IND B19R), BSH and SOM B18R (GAR D9R, teins that are not identical in amino acid sequence but IND B21R), and G1R. The B5R and B16R are cognates are nevertheless highly conserved in other poxviruses as of vaccinia Copenhagen B4R and B18R, respectively, al- isologs of cellular proteins. These include D4R (GAR B3R, though the vaccinia homologs of BSH B18R and G1R are IND D2R) for a putative zinc-binding protein (Senkevich et truncated by frameshifts (Shchelkunov et al., 1993b). al., 1994), D6R (GAR B5R, IND D4R) for an epidermal growth factor homolog (Blomquist et al., 1984; Twardzik et al., 1985), Variations in repetitive sequences D15L (GAR B18L, IND D12L) for a complement-binding pro- tein (Kotwal and Moss, 1988a; Kotwal et al., 1990), O3L Variations caused by insertions, deletions, duplica- (GAR Q3L) for an interferon-resistance factor (Beattie et al., tions, and base changes provide a basis for evolution; 1991), and C2L (GAR P2L) for a serine protease inhibitor for poxviruses, repetitive sequences likely enable homol- isolog (Boursnell et al., 1988). ogous recombination to accomplish such changes. For example, within the sequences presently examined, di- Right-terminal open reading frames and putative rect repeats were noted in which the number of repeat proteins units varied for the different strains. Beginning at nucleo- Table 3 presents a comparison of putative proteins tide position 22,385 in the BSH DNA left-end sequence, encoded in the right-end DNA fragments, BstEII-D and the nonamer TATATCATC is repeated seven times; CNG,

AID VY 8025 / 6a1b$$$$81 06-14-96 22:36:02 viras AP: Virology 296 MASSUNG ET AL.

TABLE 3 Comparison of Proteins Coded by Variola Virus Genome Right-End Region Open Reading Framesa

BSH GAR to BSH IND to BSH SOM to BSH ORF No. aa % identity ORF % identity ORF % identity

— — — H3R 65 aa — — B4L 85 100 H4L 100/67 (67aa) B4L 100 — — — H5R 92 aa — — B5R 558 99.8 H6R 99.6 B6R 100 B6R 317 99.7 H7R 99.7 B7R 99.7 B7R 65 — H8R 100 B8R 100/45 (56 aa) B8R 266 100 H9R 98.9 B9R 99.6 B9R 74 100 H10R 97.4/76 (76 aa) B10R 97.2/71 (97 aa) — — — H11R 76 aa — — B10R 65 100 H12R 100/56 (138 aa) B11R 100 — — — H13L 53 aa — — B11R 104 100/76 (85 aa) D1R 100/104 (134 aa) B12R 100/104 (134 aa) B12R 344 100 D2R 99.1 B13R 100 B13R 149 100 D3R 99.3 B14R 100 B14L 69 100 D4L 98.6/69 (97 aa) B16L 98.6/69 (86 aa) — — B14.5R (69 aa) D5R 100 B17R 98.6 B15L 340 100 D6L 100 B18L 100 B16R 574 99.8 D7R 99.5 B19R 100 B17R 354 99.4 D8R 99.4/355 (355 aa) B20R 100 B18R 787 100 D9R 99.2 B21R 99.9 B19R 70 100 D10R 97.1 B22R 100 — — B19.5R (126 aa) D11R 99.2/127 (127 aa) B23R 100/81 (83 aa) B20R 88 100 D12R 98.9 B24R 98.9 B21R 372 100 D13R 99.5 B25R 100 B22R 1897 99.8/1897 (1896 aa) D14R 99.6/1897 (1896 aa) B26R 99.7/1897 (1896 aa) G1R 585 100 G1R 99.0 G1R 100 G2R 348 98.6/349 (349 aa) G2R 98.6/349 (349 aa) G2R 99.7/349 (349 aa) G3R 253 100 G3R 99.2 G3R 99.6 — — — G4R 165 aa — —

a FASTA analysis. The numbers of deduced amino acids (aa) in proteins coded by open reading frames (ORFs) that differ in size relative to BSH correlates are shown in parentheses. Enumerations of SOM ORFs are the same as BSH and enumeration of IND ORFs follows that of Shchelkunov et al. (1995). Dash (—) indicates that the cognate ORF is absent in the sequences determined.

SOM, IND, and GAR DNAs contain 5, 9, 12, and 31 copies thymidines, and GAR shows cytosine and guanine bases of the nonamer, respectively. The nonamer is within the between the thymidine clusters; however, there are no variola C5L (E1L in GAR) reading frame; in each variola intervening bases in the 12-thymidine repeat in BSH. The strain, the number of repeat units determines the number differences in the thymidine strings cause frameshifts in of Ile-Asp-Asp repeats in the C5L-encoded protein. Con- SOM, CNG, IND, and GAR correlates of BSH D16L, and sequently, the protein products of C5L (GAR E1L) vary GAR B19L shows a second frameshift because of a sin- because the number of deduced amino acid residues is gle base deletion coinciding with BSH base 13,794. Con- different: CNG, 231; BSH, 237; SOM, 241; IND, 251; and sequently, the methionine at position 21 in the BSH pro- GAR, 312. The BSH C5L-encoded protein has 88% amino tein encoded by D16L is predicted to be the amino-termi- acid identity with a 226-residue protein encoded by F1L nal methionine of the SOM, IND, and CNG homologs, of vaccinia virus Copenhagen; however, inspection of and the methionine at position 87 in the BSH protein vaccinia Copenhagen (and WR) shows only a single non- coincides with the amino-terminal residue of the GAR amer at positions 34,047–34,055 (Goebel et al., 1990). B19L product. The BSH D16L-encoded protein shows Further leftward, the BSH genome DNA contains 12 90% identity to the carboxy-terminal 200 residues of the thymidine residues beginning at position 13836, within vaccinia virus Copenhagen C2L-encoded 512-residue D16L; however, the corresponding SOM, CNG, IND, and protein, which has no known function (Goebel et al., GAR sequences contain 30, 41, 21, and 7 thymidine resi- 1990). dues, respectively. In SOM, IND, and CNG, the thymi- Near the left terminus of the BSH genome, within D2L, dines are followed by a single guanine and then 4 more there is a 21-bp direct repeat unit GTTGAAGACTCTTCC-

AID VY 8025 / 6a1b$$$$81 06-14-96 22:36:02 viras AP: Virology SMALLPOX VIRUS TERMINAL DNA REGIONS 297

AGAGAC with CG separating the units. SOM, CNG, IND, sponding to BSH and SOM reading frame B10R (IND and GAR contain a single 21-bp unit and no CG divider, B11R). The consequence of the insert is a modified B10R which provides a frameshift following codon 122 and a correlate, GAR H12R, and the alastrim-specific reading diverged translation product for these strains. Vaccinia frames H11R and H13L. GAR H12R initiates by coding virus Copenhagen also contains a single 21-bp unit for 82 alastrim-specific amino-terminal amino acids and within C16L (and B22R, a vaccinia ITR reading frame) for concludes by coding for the same carboxy-terminal 56 a 181-amino-acid protein of unknown function (Goebel amino acids coded by BSH B10R to yield a 138-amino- et al., 1990). acid GAR polypeptide. Queries of protein databases with BSH has 17 direct repeats of the dimer AT upstream amino acid sequences coded by H11R and H13L showed of B14L, a right-end reading frame for a putative protein no significant homologs. The H13L-encoded putative pro- with no correlate in vaccinia virus Copenhagen; SOM tein is particularly interesting because it is an excessively has 38 copies, GAR has 30 copies, and IND, 15 copies. hydrophobic 53-amino-acid product with virtually no hy- The B14L reading frames of BSH and SOM code for a drophilic sectors. Poxvirus late transcript leader consen- 69-amino-acid protein, and the repeated dimers are 9 bp sus sequences, TAAAT, adjoin the start codon, sug- upstream of the B14L initiating codon. However, in GAR gesting that H13L is expressed at a late time after infec- and IND the repeats are in frame with an upstream ATG tion. We noted that two other alastrim isolates, Sierra start codon, thereby providing 97- and 86-amino-acid ho- Leone-1968 (SLN) and Butler-1968 (BUT), contain the mologs, respectively (GAR D4L and IND B16L). 627-bp segment and reading frames therein. Last, we noted several different direct repeat arrays in An 898-bp segment of GAR-distinctive sequences was genome regions that do not appear to code for proteins. identified about 8 kb from the left-terminal loop (Fig. 1B). For example, in BSH and SOM DNA, between B8R and The bases, between GAR positions 7373 and 8269, were B9R, there are 16 direct repeats of TAATAC; but, IND, absent in nonalastrim variola virus DNAs examined. PCR between B9R and B10R, contains 8 copies of the hex- amplification and DNA sequencing indicated that the amer, and GAR contains 2 copies between H9R and segment is also present in SLN and BUT alastrim strains H10R. and that correlate sequences exist in cowpox and mon- keypox virus DNAs (data not shown). Of interest, the Alastrim-specific sequences segment is not fully alastrim-specific as is the 627-bp segment. A 72.5% base sequence identity was observed As indicated in Fig. 1, HindIII DNA restriction maps for between the 898-bp and cognate left-end sequences of alastrim variola minor virus differ slightly from maps for WR and Copenhagen vaccinia viruses, and an 89% iden- variola major and African variola minor strains (Esposito tity was noted in a comparison with cowpox virus cog- and Knight, 1985; Mackett and Archard, 1979). Two addi- nate sequences (cowpox and vaccinia cognate se- tional HindIII sites are present in alastrim DNA (Fig. 1B), quences are 79% identical). The greater homology of the one near the left-end separating HindIII-P and -E, and alastrim 898-bp sequences with cowpox virus sequences one near the right-end separating HindIII-H and -D. The is puzzling (see Discussion). additional left-end site is the result of a single base Superimposed on BSH DNA, the 898-bp GAR segment change, T to C, relative to BSH nucleotide position lies between BSH nucleotide positions 8042 and 8940, 21,997. Interestingly, this site aligns with a mapped Hin- an intergenic region between BSH D9L and D10L. Figure dIII site in the DNA of vaccinia, cowpox, ectromelia, ger- 2 (see also Table 2) shows the reading frames sur- bilpox, and camelpox viruses (Esposito and Knight, 1985). rounding the 898 bp in vaccinia and variola virus DNAs. The additional right-end HindIII site is within a 627-bp A 91-amino-acid protein coded by D9L of BSH, SOM, segment not present by cross-hybridization in DNAs of and CNG (IND D6.5L) and a 152-amino-acid protein (153 , cowpox, ectromelia, camelpox, Asian and residues in CNG, SOM, and IND) coded by D10L (IND African variola major, and African variola minor viruses D7L) each pair with a 636-amino-acid (74.7-kDa) protein (data not shown). This apparently alastrim-distinctive coded by vaccinia virus C9L. The D9L-encoded protein segment, which has AT content and codon usage similar matches the carboxyl segment, and the D10L-encoded to those of the rest of GAR DNA, shows no significant protein matches the amino segment of the vaccinia virus correlate in current GenBank sequences. Of note, the protein. hexamer AAGAAT flanks the segment in all alastrim Four GAR proteins, B9L, B10L, B11R, and B12L, are strains examined but appears as a single copy in variola encoded in the area of the 898-bp alastrim insertion. GAR virus DNAs in which the segment is absent. This sug- B9L and B12L are homologous to BSH D9L and D10L, gested to us that recombination in a common ancestor respectively. The GAR B10L and B11R reading frames lie likely deleted the 627 bp during evolution of major virus entirely within the 898-bp segment. GAR B10L encodes BSH, IND, and SOM. a 172-amino-acid protein that has 62.9% identity to a 167- The added sequences in GAR are in the region corre- amino-acid section in the center of the 74.7-kDa vaccinia

AID VY 8025 / 6a1b$$$$81 06-14-96 22:36:02 viras AP: Virology 298 MASSUNG ET AL.

1994). IND-1967 derives from an outbreak inflicting 20– 30% mortalities in India in the mid-1900s, with a peak death rate of 31% in 1967. BSH-1975 derives from an outbreak in Bangladesh in the 1970s that was associated with a death rate of 18.5% (Fenner et al., 1988). The strains examined presently show high sequence homology, suggesting a common ancestor. In general, the sequence data support epidemiologic information, FIG. 2. Open reading frames in the left-end alastrim-specific se- which maintains that a relatively high-virulence (Ç25% quences compared with reading frames from variola major BSH, CNG, and IND, African variola minor SOM, and vaccinia virus Copenhagen mortalities) Asian virus was the forerunner of the African (VAC). VAC C9L codes for a protein containing 636 amino acids. BSH, viruses whose lineage somehow diverged into strains SOM, CNG, and IND show a deletion in this region (dotted line) and inflicting Ç10% or õ1% mortality. We originally expected code for two open reading frames, D9L, which has homology with that IND and BSH would have higher sequence homol- initiating sequences in VAC C9L, and D10L, which has homology with ogy than that calculated because both were major strains terminating sequences in VAC C9L. The GAR sequences show four from bordering nations during overlapping outbreaks. reading frames coding for proteins of §65 amino acids; three are transcribed leftward and one rightward as indicated by the arrowheads. However, IND sequences showed a relatively lower per- The numbers of amino acids (aa) deduced from open reading frames centage identity to CNG and SOM than does BSH. The are also shown. marked variation in the IND fragments is difficult to ex- plain if one assumes that the percentage accuracy of BSH and IND sequencing is the same and barring un- protein. The GAR B10L protein also shows homology usual DNA-modifying adaptations of BSH or IND in the (30.5% identity in 154 amino acids) to a cowpox virus host or in culture. Epidemiologic data suggest that BSH host-range protein described by Spehner et al. (1988), did not likely arise from Africa because BSH-1975 was and both the GAR and the cowpox virus proteins show from the last case in the last variola major outbreak in limited homology to erythrocyte ankyrin. GAR B11R en- the world at a time when variola major disease was codes a 93-amino-acid protein with no homology to pre- recognized only in Asia. viously described vaccinia virus proteins or any other The sequence conservation between the two African current database protein. strains examined here suggests that SOM (and other Finally, another significant difference in GAR DNA low-virulence African isolates from the 1970s, e.g., Ethio- relative to Asian and African variola virus DNAs is a pian strains) likely descended from an existing African 69-bp deletion located 48 bases to the left of the 898- variola virus lineage. Such a scenario has been de- bp segment. In BSH, SOM, IND, and CNG the 69-bp scribed previously because of the recognition of African segment is flanked by the octamer ATAATATA; the oc- strains inflicting õ1% mortality in 1904 (de Korte, 1904; tamer exists as a single copy in GAR and likely reflects Bedson et al., 1963; Dumbell and Huq, 1975, 1986). The a deletion in an ancestral virus genome. The corre- relatively close relation of high- and low-virulence African sponding region in vaccinia virus DNA contains a dele- strains is further supported by the similarity of left-end tion similar to that in GAR. region open reading frames. These distinctive, closely related CNG and SOM proteins connect the ancestry DISCUSSION of the two strains despite the fact that CNG and SOM Epidemiologic factors associated with smallpox out- infections resulted in significantly different death rates. breaks and the examination of the biologic characteris- No protein stood out, particularly for homologs of vac- tics of individual variola virus isolates from several out- cinia proteins of described function, that provide clear- breaks have indicated that there existed a relatively cut insight into the reason for the low mortality rate asso- broad spectrum of strains. To investigate variation and ciated with SOM infection; completely sequencing SOM conservation at the genetic level, we analyzed the left- DNA might be more revealing. The current data lead us end near-terminal DNA of CNG from an outbreak in the to the impression that the relative attenuation of SOM is Congo (Zaire) in 1970 with a case-fatality rate of 9.6% associated with either (1) the expression or differential and the left- and right-near-terminal sequences of SOM expression of one or more SOM-distinct proteins arising from an outbreak in 1977 in Somalia with a death rate from the present sequences, (2) a variant protein coded of 0.4%. We also analyzed DNA sequences near each in the central region sequences not yet determined, or end of the genome of GAR from an alastrim outbreak in (3) a combination of gene alterations that appears to be Brazil in 1966 that had a case-fatality rate of 0.8%. The inconsequential individually but is significant cumula- base sequences were compared with previously deter- tively. mined sequences for IND-1967 (Shchelkunov et al., For the terminal regions examined here, the variance 1993a,c, 1994a,b) and BSH-1975 (Massung et al., 1993, between strains for any segment of DNA does not appear

AID VY 8025 / 6a1b$$$$81 06-14-96 22:36:02 viras AP: Virology SMALLPOX VIRUS TERMINAL DNA REGIONS 299 to correlate with the physical position of the particular Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. segment relative to the terminal loops, even if the 898- (1990). Basic local alignment search tool. J. Mol. Biol. 215, 403–410. Beattie, E., Tartaglia, J., and Paoletti, E. (1991). Vaccinia virus-encoded and 627-bp alastrim regions are excluded from such de- eIF-2alpha homolog abrogates the antiviral effect of interferon. Virol- terminations. The relative stability of certain genes com- ogy 183, 419–422. pared with others in the regions examined are likely the Bedson, H. S., Dumbell, K. R., and Thomas, W. R. G. (1963). Variola in result of selective pressures imposed on particular Tanganyika. Lancet 2, 1085–1089. genes to maintain virulence level, strict human host Blomquist, M. C., Hunt, L. T., and Barker, W. C. (1984). Vaccinia virus 19-kilodalton protein: Relationship to several mammalian proteins, range, and high aerosol transmissibility of variola vi- including two growth factors. Proc. Natl. Acad. Sci. USA 81, 7363– ruses. 7367. The origin of alastrim virus has been generally as- Boursnell, M. E. G., Foulds, I. J., Campbell, J. I., and Binns, M. M. (1988). cribed to the attenuation of a variola major strain with Non-essential genes in the vaccinia virus HindIII K fragment: A gene no further genetic detail. The first reports of alastrim out- related to serine protease inhibitors and a gene related to the 37K vaccinia virus major envelope antigen. J. Gen. Virol. 69, 2995–3003. breaks were in the United States in 1895 in Pensacola, Chapin, C. V. (1913). Variation in type of infectious disease as shown Florida (Chapin, 1913; Chapin and Smith, 1932). Above, by the history of smallpox in the United States. J. Infect. Dis. 13, 171– we noted that the left-end of GAR contains a HindIII-P/- 196. E site paralleling a HindIII site in the DNA of several Old Chapin, C. V., and Smith, J. (1932). Permanency of the mild type of World and that the 898-bp GAR se- smallpox. J. Prevent. Med. 6, 273–320. Dear, S., and Staden, R. (1991). A sequence assembly and editing quences left of the site are related, at reduced homology, program for efficient management of large projects. Nucleic Acids to sequences in vaccinia (72.5%), cowpox (89%), and Res. 19, 3907–3911. DNAs. Further, we noted that the right- de Korte, W. E. (1904). Amaas, or kaffir milk-pox. Lancet 1, 1273–1276. end 627-bp segment within the AAGAAT direct repeat is Devereux, J., Haeberli, P., and Smithies, O. (1984). A comprehensive alastrim-distinctive. The features suggested to us that set of sequence analysis programs for the VAX. Nucleic Acids Res. 12, 387–395. the left-end of alastrim virus DNA derived as a product Dumbell, K. R., and Huq, F. (1975). Epidemiological implications of the of recombination with another orthopoxvirus, and the typing of variola isolates. Trans. R. Soc. Trop. Med. Hyg. 69, 303– right-end may be maintained from an ancestral variola 306. strain. Dumbell, K. R., and Huq, F. (1986). The virology of variola minor. Correla- Alternatively, if the left-end 898-bp segment is from a tion of laboratory tests with the geographic distribution and human virulence of variola isolates. Am. J. Epidemiol. 123, 403–415. precursor virus, it would suggest that Asian and African Englestad, M., Howard, S. T., and Smith, G. L. (1992). A constitutively major and African minor strains that lack the sequences expressed vaccinia gene encodes a 42-kilodalton glycoprotein re- are a result of relatively recent evolutionary events and lated to complement control factors that forms part of the extracellu- that alastrim virus is a precursor to other variola strains. lar virus envelope. Virology 188, 801–810. However, epidemiologic information does not support Esposito J. J., and Knight J. C. (1985). Orthopoxvirus DNA: A comparison of restriction profiles and maps. Virology 143, 230–251. this scenario. Fenner, F., Henderson, D. A., Arita, I., Jezek, Z., and Ladnyi, I. D. (1988). Interestingly, when 500- to 1000-bp segments across ‘‘Smallpox and Its Eradication.’’ World Health Organization, Geneva. the genome of GAR DNA are paired with vaccinia virus Goebel, S. J., Johnson, G. P., Perkus, M. E., Davis, S. W., Winslow, DNA sequences, ú90% identity is observed, except for J. P., and Paoletti, E. (1990). The complete nucleotide sequence of the 72.5% identity of the region containing the 898-bp vaccinia virus. Virology 179, 247–266. Isaacs, S. N., Wolffe, E. J., Payne, L. G., and Moss, B. (1992). Character- segment. The higher percentage of homology with a cog- ization of a vaccinia virus-encoded 42-kilodalton class I membrane nate in cowpox virus suggested to us that that the intro- glycoprotein component of the extracellular virus envelope. J. Virol. duction of the 898-bp segment into alastrim did not in- 66, 7217–7224. volve a recombination with a current-day vaccinia virus. Johnson, G. P., Goebel, S. J., and Paoletti, E. (1993). An update on the Rather, it was a poxvirus more closely related to cowpox vaccinia virus genome. Virology 196, 381–401. Knight, J. C., Massung, R. F., and Esposito, J. J. (1995). Polymerase virus, possibly an early passage of the New York Board chain reaction identification of smallpox virus. In ‘‘Diagnosis of Hu- of Health strain of vaccinia virus, which represents the man Viruses by Polymerase Chain Reaction Technology,’’ (Y. Becker earliest known smallpox vaccine used in the New World and G. Darai, Eds.), 2nd ed., pp. 297–302. Springer-Verlag, Berlin. in 1856 (Fenner et al., 1988). Kotwal, G. J., Isaacs, S. N., McKenzie, R., Frank, M. M., and Moss, B. (1990). Inhibition of the complement cascade by the major secretory ACKNOWLEDGMENT protein of vaccinia virus. Science 250, 827–830. Kotwal, G. J., and Moss, B. (1988a). Vaccinia virus encodes a secretory We thank the technical staff of the CDC Biotechnology Core Facility polypeptide structurally related to complement control proteins. Na- Branch for oligonucleotides. ture 335, 176–178. Kotwal, G. J., and Moss, B. (1988b). Analysis of a large cluster of nones- REFERENCES sential genes deleted from a vaccinia virus terminal transposition mutant. Virology 167, 524–537. Alcami, A., and Smith, G. L. (1995). Vaccinia, cowpox, and camelpox Kotwal, G. J., and Moss, B. (1989). Vaccinia virus encodes two proteins viruses encode soluble gamma interferon receptors with novel broad that are structurally related to members of the plasma serine prote- species specificity. J. Virol. 69, 4633–4639. ase inhibitor superfamily. J. Virol. 63, 600–606.

AID VY 8025 / 6a1b$$$$81 06-14-96 22:36:02 viras AP: Virology 300 MASSUNG ET AL.

Mackett, M., and Archard, L. C. (1979). Conservation and variation in Shchelkunov, S. N., Blinov, V. M., Totmenin, A. V., Marennikova, S. S., Orthopoxvirus genome structure. J. Gen. Virol. 45, 683–701. Kolykhalov, A. A., Frolov, I. V., Chizhikov, V. E., Gutorov, V. V., Gashni- Massung, R. F., Esposito, J. J., Liu, L-i., Qi, J., Utterback, T. R., Knight, kov, P. V., Belanov, E. F., Belavin, P. A., Resenchuk, S. M., Andzhapari- J. C., Aubin, L., Yuran, T. E., Parsons, J. M., Loparev, V. N., Selivanov, dze, O. G., and Sandakhchiev, L. S. (1993c). Nucleotide sequence N. A., Cavallaro, K. F., Kerlavage, A. R., Mahy, B. W. J., and Venter, analysis of variola virus HindIII M, L, I genome fragments. Virus Res. J. C. (1993). Complete DNA sequence of the variola smallpox virus 27, 25–35. genome reveals potential virulence determinants within terminal re- Shchelkunov, S. N., Blinov, V. M., Resenchuk, S. M., Totmenin, A. V., gions. Nature 366, 748–751. Olenina, L. V., Chirikova, G. B., and Sandakhchiev, L. S. (1994a). Massung, R. F., Knight, J. C., and Esposito, J. J. (1995). Topography of Analysis of the nucleotide sequence of 53 kbp from the right terminus variola smallpox virus inverted terminal repeats. Virology 211, 350– of the variola major virus strain India-1967 genome. Virus Res. 34, 355. 207–236. Massung, R. F., Liu, L-i., Qi, J., Knight, J. C., Yuran, T. E., Kerlavage, Shchelkunov, S. N., Resenchuk, S. M., Totmenin, A. V., Blinov, V. M., A. R., Parsons, J. M., Venter, J. C., and Esposito, J. J. (1994). Analysis and Sandakhchiev, L. S. (1994b). Analysis of the nucleotide sequence of the complete genome of smallpox variola major virus strain Ban- of 48 kbp of the variola major virus strain India-1967 located on the gladesh-1975. Virology 201, 215–240. right terminus of the conservative genome region. Virus Res. 32, 37– Mossman, K., Upton, C., Buller, R. M. L., and McFadden, G. (1995). 55. Species specificity of ectromelia virus and vaccinia virus interferon- Shchelkunov, S. N., Massung, R. F., and Esposito, J. J. (1995). Compari- K binding proteins. Virology 208, 762–769. son of the genome DNA sequences of Bangladesh-1975 and India- Nakano, J. H., and Esposito, J. J. (1989). Poxviruses. In ‘‘Diagnostic 1967 variola viruses. Virus Res. 36, 107–118. Procedures for Viral, Rickettsial, and Chlamydial Infections’’ (N. J. Smith, G. L., and Chan, Y. S. (1991). Two vaccinia virus proteins structur- Schmidt and R. W. Emmons, Eds.), 6th ed., pp. 224–265. American ally related to the interleukin-1 receptor and the immunoglobulin Public Health Association, Washington, D. C. superfamily. J. Gen. Virol. 72, 511–518. Patel, A. H., Gaffney, D. F., Subak-Sharpe, J. H., and Stow, N. D. (1990). Spehner, D., Gillard, S., Drillien, R., and Kirn, A. (1988). A cowpox virus DNA sequence of the gene encoding a major secreted protein of gene required for multiplication in Chinese hamster ovary cells. J. vaccinia virus, strain Lister. J. Gen. Virol. 71, 2013–2021. Virol. 62, 1297–1304. Perkus, M. E., Goebel, S. J., Davis, S. W., Johnson, G. P., Limbach, K., Takahashi-Nishimaki, F., Funahashi, S-i., Miki, H., Hashizume, S., and Norton, E. K., and Paoletti, E. (1990). Vaccinia virus host range genes. Sugimoto, M. (1991). Regulation of plaque size and host range by a Virology 179, 276–286. vaccinia virus gene related to complement system proteins. Virology Pickup, D. J., Ink, B. S., Hu, W., Ray, C. A., and Joklik, W. K. (1986). 181, 158–164. Hemorrhage in lesions caused by cowpox virus is induced by a viral Twardzik, D. R., Brown, J. P., Ranchalis, J. E., Todaro, G. J., and Moss, protein that is related to plasma protein inhibitors of serine prote- B. (1985). Vaccinia virus-infected cells release a novel polypeptide ases. Proc. Natl. Acad. Sci. USA 83, 7698–7702. functionally related to transforming and epidermal growth factors. Senkevich, T. G., Koonin, E. V., and Buller, R. M. L. (1994). A poxvirus Proc. Natl. Acad. Sci. USA 82, 5300–5304. protein with a RING zinc finger motif is of crucial importance for Ueda, Y., Morikawa, S., and Matsuura, Y. (1990). Identification and virulence. Virology 198, 118–128. nucleotide sequence of the gene encoding a surface antigen induced Shchelkunov, S. N., Blinov, V. M., Resenchuk, S. M., Totmenin, A. V., by vaccinia virus. Virology 177, 588–594. and Sandakhchiev, L. S. (1993a). Analysis of the nucleotide sequence Upton, C., Macen, J. L., Schreiber, M., and McFadden, G. (1991). Myx- of a 43 kbp segment of the genome of variola virus India-1967 strain. oma virus expresses a secreted protein with homology to the tumor Virus Res. 30, 239–258. necrosis factor receptor gene family that contributes to viral viru- Shchelkunov, S. N., Blinov, V. M., and Sandakhchiev, L. S. (1993b). lence. Virology 184, 370–382. Ankyrin-like proteins of variola and vaccinia viruses. FEBS Lett. 319, Upton, C., Mossman K., and McFadden, G. (1992). Encoding of a homo- 163–165. log of the IFN-K receptor by myxoma virus. Science 258, 1369–1372.

AID VY 8025 / 6a1b$$$$81 06-14-96 22:36:02 viras AP: Virology