The fiber shaft of human adenoviruses (HAdVs) is essential for bringing the penton base into proximity to the secondary cellular receptor. Fiber shaft sequences of all 53 HAdV types were studied. Phylogeny of the fiber shaft revealed clustering corresponding to the HAdV species concept. An intraspecies recombination hot spot was found at the shaft/knob boundary, a highly conserved sequence stretch. For example, HAdV-D20 clustered with HAdV-D23 in the fiber shaft, but with HAdV-D47 in the fiber knob. Although all shafts exhibited the typical pseudorepeats, amino acid sequence identity was found to be as high as 92 % (interspecies) and 54 % (intraspecies). In contrast to a previous study, a flexibility motif (KXGGLXFD/N) was found in eight HAdV-D types, whereas the putative heparan sulfate-binding site (KKTK) was only found in species HAdV-C. Our results suggest that pseudotyping of gene-therapy vectors at the shaft/ knob boundary is feasible, but that flexibility data of shafts should be considered.

BackgroundHuman adenoviruses (HAdV) are causing a broad spectrum of diseases. One of the most severe forms of adenovirus infection is a disseminated disease resulting in significant morbidity and mortality. Several reports in recent years have identified HAdV-31 from species A (HAdV-A31) as a cause of disseminated disease in children following haematopoetic stem cell transplantation (hSCT) and liver transplantation. We sequenced and analyzed the complete genome of the HAdV-A31 prototype strain to uncover unique sequence motifs associated with its high virulence. Moreover, we sequenced coding regions known to be essential for tropism and virulence (early transcription units E1A, E3, E4, the fiber knob and the penton base) of HAdV-A31 clinical isolates from patients with disseminated disease.ResultsThe genome size of HAdV-A31 is 33763 base pairs (bp) in length with a GC content of 46.36%. Nucleotide alignment to the closely related HAdV-A12 revealed an overall homology of 84.2%. The genome organization into early, intermediate and late regions is similar to HAdV-A12. Sequence analysis of the prototype strain showed unique sequence features such as an immunoglobulin-like domain in the species A specific gene product E3 CR1 beta and a potentially integrin binding RGD motif in the C-terminal region of the protein IX. These features were conserved in all analyzed clinical isolates. Overall, amino acid sequences of clinical isolates were highly conserved compared to the prototype (99.2 to 100%), but a synonymous/non synonymous ratio (S/N) of 2.36 in E3 CR1 beta suggested positive selection.ConclusionUnique sequence features of HAdV-A31 may enhance its ability to escape the host's immune surveillance and may facilitate a promiscuous tropism for various tissues. Moderate evolution of clinical isolates did not indicate the emergence of new HAdV-A31 subtypes in the recent years.

Disseminated adenovirus infections cause significant mortality in stem cell transplanted patients and are suspected to originate from asymptomatic adenovirus persistence ("latency") in lymphocytes. The infection of three human T-lymphocyte lines (Jurkat, PM1, and CEM) with human adenovirus types of species A (HAdV-A31), B (HAdV-B3, -B11), and C (HAdV-C2, -C5) was investigated for 150 days in order to establish in vitro models for adenovirus persistence. HAdV-C5 persisted with continuous production of infectious virus progeny (about 10(7) TCID50 /ml) in PM1 cells. More than 100 copies of HAdV-C5-DNA per cell were detected by real-time PCR but hexon immunostaining showed that only 7.5% of the cells were infected ("carrier state infection"). Coxsackie and adenovirus receptor (CAR) expression was decreased in comparison to mock infected cultures suggesting selection of a semi-permissive subpopulation of PM-1 cells. By contrast, latency of HAdV-DNA (10(-3) -10(-4)  copies/cell) without production of infectious virus progeny was observed in HAdV-C2 infection of PM1 and Jurkat, HAdV-A31 infection of PM1, and HAdV-B3 infection of Jurkat cells. In addition, transcription of E1A, DNA polymerase and hexon mRNA was not detected by RT-PCR suggesting an equivalent of clinical "HAdV latency." Persistence of HAdV-DNA was not observed in abortive infections of PM1 cells with HAdV-B3 and -B11 and in productive, lytical infections of Jurkat cells with HAdV-C5, HAdV-B11, and HAdV-A31. In conclusion, lytic and persistent infections with and without production of infectious virus were observed depending on the type of adenovirus. Genetic determinants for viral persistence may be investigated using these newly established infection models.