Epidemic transmission of human immunodeficiency virus in renal dialysis centers in Egypt.

In 1993 an epidemic of human immunodeficiency virus (HIV) infection occurred among 39 patients at 2 renal dialysis centers in Egypt. The centers, private center A (PCA) and university center A (UCA) were visited, HIV-infected patients were interviewed, seroconversion rates at UCA were calculated, and relatedness of HIV strains was determined by sequence analysis; 34 (62%) of 55 patients from UCA and 5 (42%) of 12 patients from PCA were HIV-infected. The HIV seroconversion risk at UCA varied significantly with day and shift of dialysis session. Practices that resulted in sharing of syringes among patients were observed at both centers. The analyzed V3 loop sequences of the HIV strain of 12 outbreak patients were >96% related to each other. V3 loop sequences from each of 8 HIV-infected Egyptians unrelated to the 1993 epidemic were only 76%-89% related to those from outbreak strains. Dialysis patients may be at risk for HIV infection if infection control guidelines are not followed.

Phylogenetic analysis of hepatitis E virus isolates from Egypt.

Hepatitis E virus (HEV) genome was detected by reverse transcriptase-polymerase chain reaction (RT-PCR) in fecal samples of two sporadic cases of hepatitis E in Cairo Egypt. Sequence of the complete putative structural region [open reading frame (ORF)-2] and complete region of unknown function (ORF-3) was determined for the two HEV isolates. Phylogenetic analysis of the nucleotide sequences was performed using neighbor joining or maximum parsimony methods of tree reconstruction. Direct correspondence between the HEV evolutionary trees and geographic origin of the HEV isolates was observed. Three genotypes of HEV were identified: genotype I (Asia-Africa), genotype II (US), and genotype III (Mexico). Genotype I was further divided into two subgenotypes (Asia and Africa). In the Asian subgenotype, three smaller genetic clusters were observed (China-like sequences, Burma-like sequences, and sequence from a fulminant case of HEV). The segregation of all these genetic clusters was supported by the high level of bootstrap probabilities. Four regions of the HEV genome were used for phylogenetic analysis. In all four regions, Egyptian HEV isolates were grouped in a separate African clade.

Desialylation of peripheral blood mononuclear cells promotes growth of HIV-1.

Activation of peripheral blood CD4+ helper T lymphocytes establishes a permissive state for growth of HIV-1. Activated T lymphocytes expressed increased sialidase (neuraminidase) activity and were hyposialylated. Treatment of freshly isolated peripheral blood mononuclear cells (PBMCs) with microbial neuraminidase (NANase) or phytohemagglutinin (PHA) prior to infection at low multiplicity with T cell line-adapted HIV-1IIIB resulted in production of large amounts of p24 antigen and reverse transcriptase. In contrast, neither viral component was detected in the medium of mock-treated cells infected at a similar multiplicity through 21 days in culture. The titer of a stock solution of HIV-1IIIB was 1.4 +/- 0.18 log10 greater in NANase-treated PBMCs than in mock-treated cells; the titer was similarly raised 1.5 to 1.76 +/- 0.18 log10 in PHA-treated cells. Growth of the primary isolate HIV-1(91/US/056) was also enhanced in NANase-treated PBMCs; the titer of a stock solution of HIV-1(91/US/056 was 1.0 +/- 0.16 log10 greater in NANase-treated PBMCs than in mock-treated cells 7 days after infection. No enhancement of viral growth in PBMCs was detected when NANase was heat-inactivated or specifically inhibited with 2,3-dehydro-2-desoxy-N-acetyl-neuraminic acid prior to use. Treatment of PBMCs with NANase did not alter the distribution of lymphocyte subsets nor change the density of CD4 antigen per cell after 7 days in culture. Whereas PHA treatment of PBMCs was mitogenic, pretreatment with NANase was not; the amount of [3H]thymidine incorporated into DNA and culture growth characteristics were similar for NANase- and mock-treated cells. Thus, desialylation of PBMCs promoted a permissive state for growth of HIV-1 without affecting the rate of DNA synthesis or relative number of target CD4+ cells.

Seroprevalence survey of Egyptian tourism workers for hepatitis B virus, hepatitis C virus, human immunodeficiency virus, and Treponema pallidum infections: association of hepatitis C virus infections with specific regions of Egypt.

Blood samples from 740 Egyptian Nationals working in the tourism industry at two sites in the South Sinai governorate were screened for markers of infection with hepatitis B virus (HBV), hepatitis C virus (HCV), human immunodeficiency virus (HIV), and Treponema pallidum. Study subjects included 467 individuals from a rural seashore tourist village and 273 persons at two hotels in a well-established resort town. Subjects' ages ranged from 15 to 70 years; 99.3% were male. The prevalence of serologic markers for currently asymptomatic or past HBV infection alone was 20.7% (n = 153), of markers for past or chronic HCV infection alone was 7.4% (n = 55), and of markers for both HBV and HCV was 6.9% (n = 51). Of the 204 individuals positive for anti-HBV core antibody, 12 (5.9%) were also positive for hepatitis B surface antigen. Two individuals (0.3%) had a serologic market suggestive of an active syphilitic infection. No subject was found to be HIV-seropositive. History of prior injections and number of injections were associated with infection with HCV. Primary residence in the Nile delta and valley areas where schistosomiasis is highly endemic, was also a statistically significant risk factor for HCV, but not HBV infection.

PIP: In June 1994, in Egypt, a physician, a laboratory technician, and a recorder surveyed 740 nationals aged 15-70 years, 99.3% of whom were male, who worked in the local tourist industry of the South Sinai governorate (a rural seashore tourist village and a well-established tourist town). Researchers aimed to determine the prevalence of past or chronic infections with hepatitis B virus (HBV), hepatitis C virus (HCV), HIV, and Treponema pallidum (syphilis) in tourist workers and to identify risk factors for infection with these pathogens. Condoms were used and safer sex was practiced in about 90% of casual sexual encounters. No tourist worker tested positive for HIV-1 or HIV-2 infection. 0.3% had active syphilis. 27.6% of the tourist workers tested positive for HBV. 1.6% (5.9% of HBV-positive workers) were positive for hepatitis B surface antigen, indicating an asymptomatic HBV infection. 14.3% of all tourist workers tested positive for HCV. 6.9% tested positive for both HBV and HCV. Rural residence was a significant risk factor for HBV infection (odds ratio [OR] = 1.6; p = 0.02). Significant risk factors for HCV infection included residence in a region highly endemic for schistosomiasis (i.e., Nile delta and valley areas) (OR = 3.2; p 0.01), rural residence (OR = 2.3; p = 0.01), and more than 10 lifetime injections (OR = 2.6; p = 0.02).

Acute primary human immunodeficiency virus type 1 infection in a patient with concomitant cytomegalovirus encephalitis.

We report what we believe is the first case of primary human immunodeficiency virus type 1 (HIV-1) infection and simultaneous cytomegalovirus (CMV) encephalitis, which was confirmed by detection of CMV DNA in the patient's cerebrospinal fluid with use of the polymerase chain reaction. This coinfection had an unusual course, and the patient's clinical status deteriorated despite administration of combination antiretroviral therapy. The patient responded clinically only after therapy for CMV infection was added to his combination antiretroviral regimen. An atypical course and duration of symptomatic primary HIV-1 infection should suggest a possible coincident infection with other opportunistic agents that are normally expected to cause disease later in the course of HIV-1 infection. Current recommendations from the Centers for Disease Control and Prevention list CMV encephalitis as an AIDS-defining event.

Relative inefficiency of soluble recombinant CD4 for inhibition of infection by monocyte-tropic HIV in monocytes and T cells.

Macrophages are major viral reservoirs in the brain, lungs, and lymph nodes of HIV-infected patients. But not all HIV isolates infect macrophages. The molecular basis for this restrictive target cell tropism and the mechanisms by which HIV infects macrophages are not well understood: virus uptake by CD4-dependent and -independent pathways have both been proposed. Soluble rCD4 (sCD4) binds with high affinity to gp 120, the envelope glycoprotein of HIV, and at relatively low concentrations (less than 1 microgram/ml) completely inhibits infection of many HIV strains in T cells or T cell lines. HTLV-IIIB infection of the H9 T cell line was completely inhibited by prior treatment of virus with 10 micrograms/ml sCD4: no p24 Ag or HIV-induced T cell syncytia were detected in cultures of H9 cells exposed to 1 x 10(4) TCID50 HTLV-IIIB in the presence of sCD4. Under identical conditions and at a 100-fold lower viral inoculum, 10 micrograms/ml sCD4 had little or no effect on infection of monocytes by any of six different HIV isolates by three different criteria: p24 Ag release, virus-induced cytopathic effects, and the frequency of infected cells that express HIV-specific mRNA. At 10- to 100-fold higher concentrations of sCD4, however, infection was completely inhibited. Monoclonal anti-CD4 also prevented infection of these same viral isolates in monocytes. The relative inefficiency of sCD4 for inhibition of HIV infection in monocytes was a property of the virion, not the target cell: HIV isolates that infect both monocytes and T cells required similarly high levels of sCD4 (100 to 200 micrograms/ml) for inhibition of infection. These data suggest that the gp120 of progeny HIV derived from macrophages interacts with sCD4 differently than that of virions derived from T cells. For both variants of HIV, however, the predominant mechanism of virus entry for infection is CD4-dependent.

Role of mononuclear phagocytes in the pathogenesis of human immunodeficiency virus infection.

We have presented evidence in this review for the following: (a) Macrophages are likely the first cell infected by HIV. Recovery of HIV from macrophages has been documented in the early stages of infection in which virus isolation in T cells is unsuccessful and detectable levels of antibodies against HIV are absent. (b) Macrophages are major reservoirs for HIV during all stages of infection. Unlike the lytic infection of T cells, HIV-infected macrophages show little or no virus-induced cytopathic effects. HIV-infected macrophages persist in tissue for extended periods of time (months) with large numbers of infectious particles contained within intracytoplasmic vacuoles. (c) Macrophages are a vector for the spread of infection to different tissues within the patient and between individuals. Several studies suggest a "Trojan horse" role for HIV-infected macrophages in the dissemination of infectious particles. The predominant cell in most bodily fluids (alveolar fluid, colostrum, semen, vaginal secretions) is the macrophage. In semen, for example, the numbers of macrophages exceed those of lymphocytes by more than 20-fold. (d) Macrophages are major regulatory cells that control the pace and intensity of disease progression in HIV infection. Macrophage secretory products are implicated in the pathogenesis of CNS disease and in control of viral latency in HIV-infected T cells. This litany of events in which macrophages participate in HIV-infection in humans parallels similar observations in such animal lentivirus infections as visna-maedi or caprine arthritis-encephalitis viruses. HIV interacts with monocytes differently than with T cells. Understanding this interaction may more clearly define both the pathogenesis of HIV disease and strategies for therapeutic intervention.

Binding to selected regions of reovirus mRNAs by a nonstructural reovirus protein.

When assembled into 13--19S particles, the reovirus nonstructural protein sigma-NS selectively binds single-stranded RNAs. Sedimentation analyses combined with binding to nitrocellulose membrane filters showed that 1--2 pmol of reovirus mRNAs from the large, medium, or small size classes saturated in vitro the binding site(s) on 13--19S particles containing 100 pmol of sigma-NS. All mRNA segments in each size class bound to particles, and no mRNAs in one size class excluded the binding of mRNAs in any other class. In competition experiments, the maximal binding of all reovirus mRNAs to particles of sigma-NS was achieved when medium and small mRNAs were bound before the large mRNAs. This preferred order of addition of mRNAs to sigma-NS resulted in a marked increase in the size of some of the complexes. This finding suggests that the addition of large mRNAs last to particles promoted the formation of complexes with more than one RNA segment bound per particle. The 13--19S particles of sigma-NS protected 20- to 40-nucleotide RNA fragments from nuclease digestion. At least one of the protected fragments from mRNAs of each size class included the 3' terminus; the remaining were from internal regions of the mRNAs. The protected RNA fragments rebound to particles during a second or third cycle of binding in a configuration in which they were fully protected from nuclease digestion. We conclude that binding of particles of sigma-NS to reovirus mRNAs was not at random sites but was to specific regions unique for members of each size class.

Small reovirus particle composed solely of sigma NS with specificity for binding different nucleic acids.

We reported previously that polycytidylate [poly(C)]-dependent RNA polymerase activity was a property of small spherical or triangular reovirus-specific particles which sedimented at 13 to 19S and were composed solely of the reovirus protein, sigma NS. Depending on the fraction of cellular extracts from which they were obtained, these particles exhibited marked differences in stability. Most 13 to 19S particles from a particular fraction repeatedly disaggregated into smaller 4 to 5S subunits with no enzymatic activity. Disruption of many particles could be prevented and polymerase activity retained after these particles had bound different single-stranded (ss) RNAs. Our previous results indicated that there was heterogeneity among the 13 to 19S particles in that possession of poly(C)-dependent RNA polymerase activity was a property of only some. Support for this heterogeneity was derived from the demonstration in this report that there were at least three types of binding sites present within particles in any purified preparation: (i) those binding only poly(C); (ii) those binding only reovirus ss RNAs; and (iii) those binding one or the other, but not both at the same time. It is suggested that only those particles able to bind either poly(C) or reovirus ss RNAs had poly(C)-dependent RNA polymerase activity, as reovirus ss RNAs markedly inhibited the polymerase activity. All three size classes of reovirus ss RNAs were equally effective in binding, but once bound, they were not copied. It is possible that heterogeneity in binding capacity of different particles comprised of only one protein, sigma NS, could result from the ability of subunits containing this protein to assemble into slightly different 13 to 19S particles with specificity of binding or polymerase activity conferred by the configuration of the assembled particles. The high capacity of sigma NS to bind many different nucleic acids with some specificity suggests that these particles may act during infection as condensing agents to bring together 10 reovirus ss RNA templates in preparation for double-stranded RNA synthesis.

Small reovirus-specific particle with polycytidylate-dependent RNA polymerase activity.

We previously reported that virus-specific particles with polycytidylate [poly(C)]-dependent RNA polymerase activity accumulated at 30 degrees C in reovirus-infected cells. These particles sedimented heterogeneously from 300 to 550S and traversed through a 40% glycerol cushion to the pellet in 3 h at 190,000 x g. In the present report, we found that smaller particles with poly(C)-dependent RNA polymerase activity remained in the glycerol cushion. These smaller, enzymatically active particles, when purified, sedimented at 15 to 1S. They were spherical or triangular with a diameter of 11 to 12 nm. They were comprised mostly, and likely solely, of one reovirus protein, sigma NS. No particles with poly(C)-dependent RNA polymerase activity were found in mock-infected cells. Chromatography on the cation exchanger, CM-Sephadex, ascertained that sigma NS was the poly(C)-dependent RNA polymerase and showed its existence in two forms. In one form, it was enzymatically active and eluted from the column at 0.5 M KCl. In the enzymatically inactive state, it did not bind to the column. Our results suggest that the enzymatically active form of sigma NS carries a greater net positive charge than the inactive form. They also suggest that both forms of sigma NS are associated with a particle which has poly(C)-dependent RNA polymerase activity.

Semliki Forest virus replication complex capable of synthesizing 42S and 26S nascent RNA chains.

A complex synthesizing Semliki Forest virus (SFV)-specific RNAs was purified from infected HeLa cells. During purification, the RNA-synthesizing complex was monitored by the presence of RNA chains synthesized during a 1 min pulse in vivo and the ability to synthesize 42S and 26S RNAs in vitro. Finally, the protein composition of the replication complex was analysed. Thirty to 40% of the pulse-labelled RNAs and 10 to 25% of the polymerase activity present in the postnuclear supernatant were recovered in smooth membranes. At this stage of purification single stranded 42S and 26S RNA were synthesized and released from the replication complex in vitro. After treatment of the smooth membrane fraction with Triton X-100 the replication complex was solubilized. When analysed by sucrose gradient centrifugation, the solubilized replication complex distributed heterogeneously. It had reduced RNA polymerase activity, but was still able to synthesize both 42S and 26S nascent RNA chains which were not released from RIs and RFs. The non-structural protein ns70 was the major virus-specified component associated with the replication complex.

Cleavage defect in the non-structural polyprotein of Semliki Forest virus has two separate effects on virus RNA synthesis.

When Semliki Forest virus ts-4 mutant infected cultures are grown at the permissive temperature (28 degrees C) and shifted to the restrictive temperature (39 degrees C), two different defects in RNA synthesis are manifested: (i) the synthesis of 26S RNA is stopped within 60 min (Saraste et al. 1977) and (ii) the increase in RNA synthesizing activity ceases, in contrast to cultures maintained at 28 degrees C, indicating that no new active RNA polymerase is formed at 39 degrees C. Accumulation of a non-structural precursor protein with an apparent mol. wt. of about 220 000 (ns220) was demonstrated in ts-4 infected cultures shifted to 39 degrees C. NS220 was labelled during short pulses given immediately after release of protein synthesis from hypertonic initiation block, suggesting that genes coding for ns220 are located near the initiation site at the 5'-end of the 42S RNA. The viral specificity of ns220 was shown by its disappearance after a shift to 28 degrees C and by labelling in the presence of sucrose, when no host cell protein synthesis is detectable. The two functional defects can be explained if the polypeptides responsible for the RNA polymerizing activity and that responsible for the synthesis of 26S RNA are components of the same non-structural polyprotein. A mutation in the latter polypeptide which prevents cleavage of the polyprotein would thereby prevent the further formation of active RNA polymerase. If cleavage of the polyprotein has taken place at the permissive temperature, the RNA polymerase would remain active also at 39 degrees C, whereas the polypeptide responsible for 26S RNA synthesis would become inactive due to the mutation.

Mechanism for control of synthesis of Semliki Forest virus 26S and 42s RNA.

When cells infected with the Semliki Forest virus (SFV) mutant ts-4 were shifted to the nonpermissive temperature, synthesis of 26S RNA ceased, whereas synthesis of 42S RNA continued normally. These two single-stranded SFV RNAs are synthesized in two types of replicative intermediate (RI), 26S RNA in RI(b) and 42S RNA in RI(a). Cessation of 26S RNA synthesis after shift up in temperature was accompanied by loss of RI(b). When infected cells were shifted back down to 27 degrees C, 26S RNA synthesis resumed, coincident with the reappearance of RI(b). In both types of RI, the 42S minus-strand RNA is template for synthesis of plus-strand RNA. In pulse-chase experiments, we obtained RIs labeled only in their minus-strand RNA, and thus could follow the fate of RIs assembled at 27 degrees C when they were shifted to 39 degrees C. Our results show that, after shift up to 39 degrees C, there was a quantitative conversion of RIs in which 26S RNA had been synthesized to RIs in which 42S RNA was synthesized. This conversion of RI(b) to RI(a) was reversible, since RIs in which 26S RNA was synthesized reappeared when the infected cultures were shifted back down to 27 degrees C. We propose that, associated with RI(b), in which 26S RNA is synthesized, there is a virus-specific protein that functions to promote initiation of 26S RNA transcription at an internal site on the 42S minus-strand RNA and to block transcription on the minus strand in this region by the SFV RNA polymerase that had bound and was copying the minus-strand RNA from its 3' end. A ribonuclease-sensitive region would thus result in the sequence adjacent to the one that was complementary to 26S RNA. This virus-specific protein is not a component of the SFV RNA polymerase that continues to transcribe 42S RNA, and it is temperature sensitive in ts-4 mutant-infected cells. When this virus-specific protein is not present on RIs, the SFV polymerase transcribes the whole 42S minus-strand RNA and yields 42S plus-strand RNA.

Reovirus-specific enzyme(s) associated with subviral particles responds in vitro to polyribocytidylate to yield double-stranded polyribocytidylate-polyriboguanylate.

In reovirus-infected cells, virus-specific particles accumulate that have associated with them a polyribocytidylate [poly(C)]-dependent polymerase. This enzyme copies in vitro poly(C) to yield the double-stranded poly(C).polyriboguanylate [poly(G)]. The particles with poly(C)-dependent polymerase were heterogeneous in size, with most sedimenting from 300S to 550S. Exponential increase in these particles began at 23 h, and maximal amounts were present by 31 h, the time of onset of exponential growth of virus at 30 degrees C. Maximal amounts of particles with active transcriptase and replicase were present at 15 and 18 h after infection. Thereafter, there was a marked decrease in particles with active transcriptase and replicase until base line levels were reached at 31 h. Thus, the increase in poly(C)-responding particles occurred coincident with the decrease in particles with active transcriptase and replicase. The requirement for poly(C) as template was specific because no RNA was synthesized in vitro in response to any other homopolymer, including 2'-O-methyl-poly(C). Synthesis was optimal in the presence of Mn(2+) as the divalent cation, and no primer was necessary for synthesis. In contrast, the dinucleotide GpG markedly stimulated synthesis in the presence of 8 mM Mg(2+). The size of the poly(C).poly(G) synthesized in vitro was dependent on the size of the poly(C) used as template. This suggested that the whole template was copied into a complementary strand of similar size. The T(m) of the product was between 100 and 130 degrees C. Hydrolysis of the product labeled in [(32)P]GMP with alkali or RNase T2 yielded GMP as the only labeled mononucleotide. This does indicate that the synthesis of the poly(G) strand in vitro did not proceed by end addition to the poly(C) template, but proceeded on a separate strand.

Replication of semliki forest virus: polyadenylate in plus-strand RNA and polyuridylate in minus-strand RNA.

The 42S RNA from Semliki Forest virus contains a polyadenylate [poly(A)] sequence that is 80 to 90 residues long and is the 3'-terminus of the virion RNA. A poly(A) sequence of the same length was found in the plus strand of the replicative forms (RFs) and replicative intermediates (RIs) isolated 2 h after infection. In addition, both RFs and RIs contained a polyuridylate [poly(U)] sequence. No poly(U) was found in virion RNA, and thus the poly(U) sequence is in minus-strand RNA. The poly(U) from RFs was on the average 60 residues long, whereas that isolated from the RIs was 80 residues long. Poly(U) sequences isolated from RFs and RIs by digestion with RNase T1 contained 5'-phosphorylated pUp and ppUp residues, indicating that the poly(U) sequence was the 5'-terminus of the minus-strand RNA. The poly(U) sequence in RFs or RIs was free to bind to poly(A)-Sepharose only after denaturation of the RNAs, indicating that the poly(U) was hydrogen bonded to the poly(A) at the 3'-terminus of the plus-strand RNA in these molecules. When treated with 0.02 mug of RNase A per ml, both RFs and RIs yielded the same distribution of the three cores, RFI, RFII, and RFIII. The minus-strand RNA of both RFI and RFIII contained a poly(U) sequence. That from RFII did not. It is known that RFI is the double-stranded form of the 42S plus-strand RNA and that RFIII is the experimetnally derived double-stranded form of 26S mRNA. The poly(A) sequences in each are most likely transcribed directly from the poly(U) at the 5'-end of the 42S minus-strand RNA. The 26S mRNA thus represents the nucleotide sequence in that one-third of the 42S plus-strand RNA that includes its 3'-terminus.

Murine virus-induced proteins synthesized by hamster tumor cells transformed by, but not producing, murine sarcoma virus.

The 8303 hamster tumor cells transformed by Moloney strain of murine sarcoma virus (M-MSV), but which do not produce virus, do contain murine virus-induced proteins. The virus-induced proteins within the cell were identified either as free proteins or in association with membranous material, including the plasma membrane. In addition, some were excreted by the 8303 hamster tumor cells into the growth medium. Most virus-induced proteins were larger than 68,000 daltons, and they did not dissociate into components of smaller size in the presence of detergent and a reducing agent. A small amount of virus-induced protein with a molecular weight of less than 20,000 was also found in the hamster tumor cells. No virus-specific proteins with the identical antigenic specificity or size of the major internal group specific antigen (molecular weight about 30,000) of the murine leukemia viruses were present in these cells. There is a common cell surface antigen present in three other tumor cell lines, both virus-producing and non-virus-producing, identical in reactivity to that of the murine virus-induced antigen of the 8303 hamster tumor cell. This antigen is not present on the cell surface of normal mouse embryo cells.

Semliki forest virus-specific RNAs synthesized in vitro by enzyme from infected BHK cells.

When Semliki Forest virus (SFV)-infected BHK cells were disrupted 4 h after infection, 75 to 90% of the total virus-specific RNA synthesizing enzyme was found in the large particle fraction, along with 75 to 90% of the in vivo-synthesized double-stranded RNAs. The RNA products of this enzyme-template complex in an in vitro system were double-stranded RNAs sedimenting predominantly at 18S, and single-stranded RNAs sedimenting at 42S, 26S, and 22S. The various virus-specific SFV RNAs synthesized in vitro were associated with different sized structures, and thus each was separable by differential centrifugation. Kinetic and pulse-chase experiments showed that the double-stranded RNAs were the precursors to the single-stranded RNAs. There were several double-stranded RNAs identified both in the in vitro product and also in extracts from infected cells. The major replicative form had a molecular weight of 4.4 x 10(6).