Extensive sequencing of cloned DNA isolated from the culture of an African green monkey simian cytomegalovirus (SCMV) derived stealth virus has identified multiple regions of highly significant homology to various bacterial genes. The apparent acquisition of bacterial sequences extends the potential role of stealth viruses as natural vectors in the transfer of genetic information. The findings highlight the dynamic interface between viral and bacterial genomes and the potential of this interaction in the emergence and spread of novel pathogens. The term viteria is proposed for microorganisms that contain both eukaryotic-viral and prokaryotic-bacterial genetic sequences

Introduction

Stealth-adaptation presumably involves the mutation and/or deletion of viral genes coding the major immunogenic components required for effective cellular immunity (1,7). This process would enable viruses to bypass cellular immune defenses and establish a persistent, active yet non-inflammatory, infection. The mechanisms that allow such stealth-adapted viruses to retain and/or regain the ability to infect and replicate in host cells are presently unknown.

The prototype stealth virus-1 was isolated from the blood of a patient diagnosed in 1991 as having chronic fatigue syndrome. The patient has never regained normal health. The virus induced a foamy vacuolated CPE in human and in animal cell lines (1), including insect cells (8). Electron microscopy of infected cultures showed herpesvirus-like particles (1). Genetic sequences, related to human cytomegalovirus (HCMV), and subsequently shown to be more closely related to African green monkey simian cytomegalovirus (SCMV), were amplified by PCR from infected cultures, and were also present in DNA extracted directly from the infected cultures (1,7,8). Several of the PCR generated products, and approximately one third of the cloned genetic sequences, could not, however be aligned to known herpesviruses (7). Moreover, even with many of the clones showing homology to either HCMV or SCMV sequences, the matching was often incomplete, with adjoining sequences seemingly unrelated to, or widely divergent from, known herpesviruses (7). The potential for sequence divergence was supported by directly comparing the sequences of clones that matched to similar regions of HCMV. Such comparisons showed significant sequence microheterogeneity, including nucleotide substitutions, insertions, deletions and recombination (7). Although sequences corresponding to widespread regions of the 235 kilobase (kb) cytomegalovirus genome were represented, DNA extracted from the cultures banded in agarose at a size of only approximately 20 kb (1).

Certain of the additional sequences contained in what appeared to be a genetically unstable, fragmented virus genome, were identified as cellular in origin (9). Moreover, the cellular genes frequently corresponded to repetitive and/or highly reiterated sequences in the human genome. A working model was proposed in which repeat sequences acted as attachment sites in the formation of a scaffold combining variously ordered viral and cellular sequences (9). This model system could similarly allow for the incorporation of bacterial sequences into the viral replicative process. This paper provides evidence for bacterial-derived sequences in at least 8% of the DNA clones derived from stealth virus-1 infected cells.

The specifics of the cloning of DNA from the stealth virus-1 infected MRC-5 cells into the pBluescript vector has been described previously (1,7,8). The 3B series of clones was obtained using EcoRI digestion of DNA extracted from the material, pelleted by ultracentrifugation, that was present in filtered supernatants of infected MRC-5 cells. The C16 series of clones was obtained using SacI digestion of agarose banded DNA extracted from the material, pelleted by ultracentrifugation, present in filtered supernatant of lysed virus-infected cells. The T3 and T7 promoter sites of the pBluescript vector were used in sequencing reactions to obtain partial sequence data from the ends of each of the inserts. Extended sequences of the inserts in selected clones were obtained using primers based on the T3 and T7 readouts. The sequencing services of the City of Hope Cancer Center, Durate CA; Midland Certified Reagent Company, Midland, Texas; and Lark Technology, Houston, Texas, were used with excellent correlation of the occasional duplicate and even triplicate testing of the same clones. The individual sequences were analyzed against GenBank entries using the gapped BlastN and unfiltered BlastX programs of the National Center for Biotechnology Information (NCBI) of the National Library of Medicine (10). A "p" value of < e -2.0 was used as a cutoff for a significant nucleotide or amino acid sequence homology.

The inserts in 180 clones of the 3B series and in 120 clones of the C16 series have been partially or completely sequenced. Of the 300 clones, the majority have at least a portion of their overall sequence that is homologous, by BlastN and/or by BlastX analysis, to the HCMV and/or SCMV genome. Other regions of the same clones and numerous additional clones, contain sequences that do not correspond to HCMV and/or to the limited known regions of the SCMV genome. Of these non-matching clones, 24 contained sequences that by BlastX analysis, could be partially matched to different protein sequences of bacterial origin. An additional clone had a nucleotide sequence that was similar to a bacterial ribosomal gene complex. The bacterial nucleotide and protein sequences that most closely matched to the sequences contained in the various clones derived from the stealth virus-1 infected culture are summarized in Table 1. For many of the matches, additional bacterial sequences were also identified, usually comprising groups of functionally similar entities. Several clones contained non-overlapping nucleotide sequences that, when translated by the BlastX program, could potentially encode amino acid sequences that corresponded to portions of quite distinct proteins, not known to be coded by any contiguous set of genes in bacterial genomes, and in some cases, seemingly derived from widely divergent bacteria. For several clones, the matching with a particular bacterial gene occurred in discrete segments separated by variable size gaps. The degree of identity and the corresponding statistical "p" values for each of the matching segments are shown in Table 1

The data presented in Table 1, indicate the existence of bacterial-related sequences in cloned DNA extracted from the stealth virus-1 infected culture. The diverse range of bacterial species identified using the BlastN and BlastX programs, argues strongly against the bacterial DNA being a contaminant of the cloning process. The high levels of identity also exclude the possibility of diverged herpesviral sequences being mistaken for bacterial sequences.

A flanking sequence in clone 3B313 matched to the human C-C chemokine receptor, which in turn matches to the G-protein coupled receptor coded by the US28 gene of HCMV (11). With this possible exception, the demonstration of a viral sequence followed by a bacterial sequence in the same clone has yet to be documented. This can clearly be inferred, however, from the information provided in the Footnote to Table 1. Specifically, near identity was observed between a stretch of clone 3B313 and a sequence contained in clone C16246 which aligned to the HCMV gene coding the US28 protein. Moreover, additional regions of clone C16246 show a strong homology with several HCMV coded genes. Similarly, the limited overlap between clones 3B513 and 3B525 is informative. Beyond the overlapping region, clone 3B525 shows highly significant matching to HCMV. The overall sequencing data are consistent with variable patterns of recombination between various sequences of viral, cellular and bacterial origins.

It could be argued that the bacterial sequences have, in fact, incorporated viral sequences, rather than the reverse. When dealing with obligate intracellular microorganisms, the distinction between virus and bacteria is somewhat irrelevant. This is especially so when several of the functions ascribed to the bacteria, are in fact, mediated by bacterial plasmids and/or subgenomic insertion elements. The predominance of virus-related sequences, and the lack of bacterial structures seen on detailed electron microscopy of stealth virus-1 cultures (1), strongly favor the notion of a virus with assimilated bacterial-related sequences. The term viteria is proposed for a eukaryotic virus which has incorporated genes of bacterial and/or bacterial plasmid origin.

Many of the matching bacterial sequences correspond to genes involved in rather unique energy generating and metabolic conversion reactions (Table 1). Particularly, noteworthy are sequences contained in genes that participate in the transport, activation and/or synthesis of uncommon metabolites. Given such a wide array of metabolic functions, it is conceivable that viteria could maintain a limited capacity to metabolize, and possibly even to replicate, outside the confines of a cell.

Certain bacteria could derive a competitive advantage by being infected with a viteria that encode functionally useful genes. The improved metabolic performance of a viteria infected bacteria, (or an infected fungus), could facilitate transmission of the underlying virus infection. Conversely, some of the pathogenicity of stealth viruses for human and animal hosts, could be mediated by toxic byproducts of the various metabolic pathways encoded by the assimilated bacterial genes. Toxic products have been detected in blood, urine and cerebrospinal fluids of stealth virus infected patients, and also in the supernatants of stealth virus cultures (unpublished). Supernatants from mixed bacterial cultures, obtained from a stool sample of the patient from whom stealth virus-1 was originally isolated, also induced a vacuolating CPE in cell culture (unpublished). In addition to one or more toxin, the supernatants contain a filterable cytopathic agent that can be passaged in tissue culture. Molecular studies on this agent have yet to be performed.

The presence of bacterial-related gene sequences in the stealth virus-1 culture has relevance to diagnostic microbiology. Positive PCR based assays, using primers reactive with various bacterial sequences, have been reported in patients with chronic fatigue syndrome, Gulf war syndrome, chronic Lyme disease, Alzheimer’s disease, multiple sclerosis, arteriosclerosis and other diseases (12-16). These reports may reflect the presence of limited bacterial-related nucleotide sequences contained within an essentially viral pathogen. Minor differences in the ribosomal sequences of bacterial species have also led investigators to use PCR to classify the postulated bacterial pathogens that were presumably being detected (17). The BlastX findings showed an even better matching of amino acid sequences than nucleotide sequences. Viteria encoded proteins might be expected, therefore, to evoke antibodies that could be misinterpreted as evidence for infection with conventional bacterial pathogens.

Acknowledgement. Ms. Susie Trang provided valuable assistance with compiling sequence data and preparation of the manuscript.