A review of genetic differences between limited and extensively passaged human cytomegalovirus strains.

The complete genetic content of human cytomegalovirus (HCMV) has been difficult to determine, since most strains studied in the laboratory have been extensively passaged in human fibroblast cultures which can change the genetic content as well as the biological properties of the virus. Approximately 13 kb of novel DNA sequences located near the right edge of the unique long (UL) component of the genome has been discovered in Toledo, clinical isolates and certain stocks of Towne. This region of novel sequence, designated the UL/b' region, encodes several interesting proteins including vCXC-1, a potent IL-8 homologue, and UL144, a member of the TNF receptor family. This region is missing from the prototypic laboratory variants of Towne and AD169. In contrast to Toledo and other low passage isolates which have relatively small repeats bracketing the UL component, the Towne and AD169 laboratory variants contain large (>10 kb) b/b' repeats. The large size of these repeats in AD169 and Towne appear to have arisen as compensation for the loss of sequences from the UL/b' region that existed in less passaged variants of these strains. Consequently, many of the haploid genes at the left edge of the prototypic wild-type (wt) UL component are diploid in AD169 and Towne. We hypothesise that this plasticity of the genome at the right edge of the UL component results from extensive passage and adaptation to replication in fibroblasts in vitro. Further work will be required to understand the complete genetic content of wt HCMV.

Open reading frame S/L of varicella-zoster virus encodes a cytoplasmic protein expressed in infected cells.

We report the discovery of a novel gene in the varicella-zoster virus (VZV) genome, designated open reading frame (ORF) S/L. This gene, located at the left end of the prototype VZV genome isomer, expresses a polyadenylated mRNA containing a splice within the 3' untranslated region in virus-infected cells. Sequence analysis reveals significant differences between the ORF S/Ls of wild-type and attenuated strains of VZV. Antisera raised to a bacterially expressed portion of ORF S/L reacted specifically with a 21-kDa protein synthesized in cells infected with a VZV clinical isolate and with the original vaccine strain of VZV (Oka-ATCC). Cells infected with other VZV strains, including a wild-type strain that has been extensively passaged in tissue culture and commercially produced vaccine strains of Oka, synthesize a family of proteins ranging in size from 21 to 30 kDa that react with the anti-ORF S/L antiserum. MeWO cells infected with recombinant VZV harboring mutations in the C-terminal region of the ORF S/L gene lost adherence to the stratum and adjacent cells, resulting in an altered plaque morphology. Immunohistochemical analysis of VZV-infected cells demonstrated that ORF S/L protein localizes to the cytoplasm. ORF S/L protein was present in skin lesions of individuals with primary or reactivated infection and in the neurons of a dorsal root ganglion during virus reactivation.

Cytomegalovirus encodes a potent alpha chemokine.

Cytomegalovirus is a widespread opportunistic pathogen affecting immunocompromised individuals in whom neutrophils may mediate virus dissemination and contribute to progression of disease. Recent sequence analysis suggests that genes absent or altered in attenuated strains may influence pathogenesis. We have found two genes, UL146 and UL147, whose products have sequence similarity to alpha (CXC) chemokines. UL146 encodes a protein, designated vCXC-1, that is a 117-aa glycoprotein secreted into the culture medium as a late gene product, where its presence correlates with the ability to attract human neutrophils. Recombinant vCXC-1 is a fully functional chemokine, inducing calcium mobilization, chemotaxis, and degranulation of neutrophils. High-affinity vCXC-1 binding is shown to be mediated via CXCR2, but not CXCR1. vCXC-1 exhibits a potency approaching that of human IL-8. As the first example of a virus-encoded alpha chemokine, vCXC-1 may ensure the active recruitment of neutrophils during cytomegalovirus infection, thereby providing for efficient dissemination during acute infection and accounting for the prominence of this leukocyte subset in cytomegalovirus disease.

A deletion mutant in the human cytomegalovirus gene encoding IE1(491aa) is replication defective due to a failure in autoregulation.

Human cytomegalovirus (CMV) replication begins with the expression of two regulatory proteins, IE1(491aa) and IE2(579aa), produced from differentially spliced transcripts under control of the ie1/ie2 promoter-enhancer. A deletion mutation removing all 406 IE1(491aa)-specific amino acids was engineered into the viral genome and this mutant (RC303 delta Acc) was propagated on an IE1(491aa)-expressing human fibroblast cell line (ihfie1.3). RC303 delta Acc failed to replicate on normal human fibroblasts at low multiplicities of infection (mois). At mois > 3 plaque-forming units per cell, virus replication and production of progeny were comparable to wild type. However, at mois between 0.01 and 1, mutant virus replicated slowly on normal fibroblasts, a pattern that suggested initiation of productive infection required multiple hits. Replication of RC303 delta Acc correlated with the ability to express IE2(579aa), consistent with a role for IE1(491aa) in positive autoregulation of the ie1/ie2 promoter-enhancer and with data suggesting that virion transactivators compensate for the lack of IE1(491aa) under high moi conditions. ie1-deficient CMV should be completely avirulent, suggesting its utility as a gene therapy vector for hematopoietic progenitors that are normal sites of CMV latency.

Human cytomegalovirus clinical isolates carry at least 19 genes not found in laboratory strains.

Nucleotide sequence comparisons were performed on a highly heterogeneous region of three human cytomegalovirus strains, Toledo, Towne, and AD169. The low-passage, virulent Toledo genome contained a DNA segment of approximately 13 kbp that was not found in the Towne genome and a segment of approximately 15 kbp that was not found in the AD169 genome. The Towne strain contained approximately 4.7 kbp of DNA that was absent from the AD169 genome, and only about half of this segment was present, arranged in an inverted orientation, in the Toledo genome. These additional sequences were located at the unique long (UL)/b' (IRL) boundary within the L component of the viral genome. A region representing nucleotides 175082 to 178221 of the AD169 genome was conserved in all three strains; however, substantial reduction in the size of the adjacent b' sequence was found. The additional DNA segment within the Toledo genome contained 19 open reading frames not present in the AD169 genome. The additional DNA segment within the Towne genome contained four new open reading frames, only one of which shared homology with the Toledo genome. This comparison was extended to five additional clinical isolates, and the additional Toledo sequence was conserved in all. These findings reveal a dramatic level of genome sequence complexity that may explain the differences that these strains exhibit in virulence and tissue tropism. Although the additional sequences have not altered the predicted size of the viral genome (230 to 235 kbp), a total of 22 new open reading frames (denoted UL133 to UL154), many of which have sequence characteristics of glycoproteins, are now defined as cytomegalovirus specific. Our work suggests that wild-type virus carries more than 220 genes, some of which are lost by large-scale deletion and rearrangement of the UL/b' region during laboratory passage.

Recombinant cytomegaloviruses for study of replication and pathogenesis.

Systematic mutagenesis of large viral genomes such as those of the cytomegaloviruses requires strategies for identifying relevant functions as well as for detailed analysis of particular genes. A number of genetic markers that have been developed in other biological systems have been useful for insertion mutagenesis in these viruses. Thus far, 57 of the over 227 genes carried by wild-type human cytomegalovirus have been found to be dispensable for growth in cultured cells. Because of the limitations on studying human cytomegalovirus in an animal host, the closely related murine cytomegalovirus has been used as a surrogate for pathogenesis, tissue tropism and latency studies in the laboratory mouse. Genetic analysis of this virus has paralleled work on human cytomegalovirus, and an understanding of genes that specifically impact viral growth in particular organs has emerged from these studies. Strategies for generation of permissive cell lines able to complement essential human cytomegalovirus replication functions have been described, and sets of cosmid clones have been used to generate recombinant viruses. These methods will enable a systematic functional analysis of the genomes of human and animal cytomegaloviruses.

Lipid-anchored influenza hemagglutinin promotes hemifusion, not complete fusion.

It has been proposed that membrane fusion events such as virus-cell fusion proceed through a hemifusion intermediate, a state where lipids but not contents of the fusing compartments mix. We engineered the influenza hemagglutinin (HA) such that it would be anchored in membranes via a glycosylphosphatidylinositol (GPI) tail. GPI-anchored HA forms a trimer that can bind red blood cells (RBCs) and change conformation under fusion-inducing conditions. Using RBCs labeled with fluorescent lipid or fluorescent soluble content probes, we found that GPI-anchored HA mediated lipid mixing with similar time course and efficiency as wt-HA, yet did not mediate transfer of soluble contents. Hence, GPI-anchored HA appears to initiate, but not complete, a fusion reaction. We interpret our results as evidence for uncoupling a physiological fusion reaction, for trapping a hemifusion intermediate, and for assigning a role to a transmembrane domain in a fusion event.

GPI- and transmembrane-anchored influenza hemagglutinin differ in structure and receptor binding activity.

We investigated the influence of a glycosylphosphatidylinositol (GPI) anchor on the ectodomain of the influenza hemagglutinin (HA) by replacing the wild type (wt) transmembrane and cytoplasmic domains with a GPI lipid anchor. GPI-anchored HA (GPI-HA) was transported to the cell surface with equal efficiency and at the same rate as wt-HA. Like wt-HA, cell surface GPI-HA, and its ectodomain released with the enzyme PI-phospholipase C (PI-PLC), were 9S trimers. Compared to wt-HA, the GPI-HA ectodomain underwent additional terminal oligosaccharide modifications; some of these occurred near the receptor binding pocket and completely inhibited the ability of GPI-HA to bind erythrocytes. Growth of GPI-HA-expressing cells in the presence of the mannosidase I inhibitor deoxymannojirimycin (dMM) abrogated the differences in carbohydrate modification and restored the ability of GPI-HA to bind erythrocytes. The ectodomain of GPI-HA produced from cells grown in the presence or absence of dMM underwent characteristic low pH-induced conformational changes (it released its fusion peptides and became hydrophobic and proteinase sensitive) but at 0.2 and 0.4 pH units higher than wt-HA, respectively. These results demonstrate that although GPI-HA forms a stable trimer with characteristics of the wt, its structure is altered such that its receptor binding activity is abolished. Our results show that transmembrane and GPI-anchored forms of the same ectodomain can exhibit functionally important differences in structure at a great distance from the bilayer.

Intermonomer disulfide bonds impair the fusion activity of influenza virus hemagglutinin.

At a low pH, the influenza virus hemagglutinin (HA) undergoes conformational changes that promote membrane fusion. While the critical role of fusion peptide release from the trimer interface has been demonstrated previously, the role of globular head dissociation in the overall fusion mechanism remains unclear. To investigate this question, we have analyzed in detail the fusion activity and low pH-induced conformational changes of a mutant, Cys-HA, in which the globular head domains are locked together by engineered intermonomer disulfide bonds (L. Godley, J. Pfeifer, D. Steinhauer, B. Ely, G. Shaw, R. Kaufmann, E. Suchanek, C. Pabo, J. J. Skehel, D. C. Wiley, and S. Wharton, Cell 68:635-645, 1992). In this paper, we show that Cys-HA expressed on the cell surface is predominantly a disulfide-bonded trimer. Cell surface Cys-HA is impaired in its membrane fusion activity, as demonstrated by both content-mixing and lipid-mixing fusion assays. It is also impaired in its ability to change conformation at a low pH, as assessed by proteinase K sensitivity. The fusion activity and low pH-induced conformational changes of cell surface Cys-HA are, however, restored to nearly wild-type levels upon reduction of the intermonomer disulfide bonds. By using a set of conformation-specific monoclonal and anti-peptide antibodies, we found that purified Cys-HA trimers are impaired in changes that occur in the globular head domain interface. In addition, changes that occur at a great distance from the engineered intermonomer disulfide bonds, notably release of the fusion peptides, are also impaired. Our results are discussed with respect to current views of the fusion-active conformation of the HA trimer.

A host cell protein binds to a highly conserved sequence element (pac-2) within the cytomegalovirus a sequence.

The human cytomegalovirus (CMV) a sequence has significant homology to two regions, pac-1 and pac-2, within the a sequence of herpes simplex virus type 1 (HSV-1). Both regions have been shown to be important cis-acting signals in HSV-1 genome maturation. We have demonstrated that a small fragment from within the CMV a sequence, containing the pac-1 and pac-2 motifs, carries all of the signals necessary for generation of genomic termini and for inversion. These observations indicated that the function of these highly conserved sequence motifs was similar in CMV and HSV-1. We have identified and partially purified a host cell protein with affinity for the sequence 5'-GGCGGCGGCGCATAAAA-3' within CMV pac-2. This partially purified protein has an apparent molecular weight of 89,000 under denaturing conditions and could be renatured after sodium dodecyl sulfate-polyacrylamide gel electrophoresis, suggesting that the capacity to bind DNA was the property of a single polypeptide chain. This activity was found in a wide variety of human cell lines, including those that are permissive as well as those that are nonpermissive for CMV growth, but not in cell lines from monkey, mouse, or drosophila origins. Our work implicates a host cell protein in a sequence function.

A cytomegalovirus protein with properties of herpes simplex virus ICP8: partial purification of the polypeptide and map position of the gene.

We demonstrated the presence of a single-stranded DNA-binding protein in human cytomegalovirus (CMV)-infected cells with properties analogous to those of herpes simplex virus (HSV) ICP8. Using monoclonal antibody specific for the CMV protein, we analyzed its fluorescence pattern and time of synthesis, mapped the gene encoding it by using a lambda gt11 library of CMV DNA fragments, and monitored its purification by phosphocellulose and DNA-Sepharose chromatography. In all characteristics we examined, the CMV protein behaved analogously to HSV ICP8. Our results are consistent with a functional role of CMV ICP8 in viral replication that is similar to that of HSV ICP8 and with the evolutionary conservation of the gene of interest in two divergent herpesviruses.