Evidence of a cellulosic layer in Pandoravirus tegument and the mystery of the genetic support of its biosynthesis

Pandoraviruses are giant viruses of amoebae with 1 m-long virions. They have an ovoid morphology and are surrounded by a tegument-like structure lacking any capsid protein nor any gene encoding a capsid protein. In this work, we studied the ultrastructure of the tegument surrounding Pandoravirus massiliensis virions and noticed that this tegument is composed of a peripheral sugar layer, an electron-dense membrane, and a thick electron-dense layer consisting in several tubules arranged in a helicoidal structure resembling that of cellulose. Pandoravirus massiliensis particles were stained by Calcofluor white, a fluorescent dye of cellulose, and the enzymatic treatment of particles by cellulase showed the degradation of the viral tegument. We first hypothesized that the cellulose tegument could be synthesized by enzymes encoded by Pandoravirus. Bioinformatic analyses revealed in Pandoravirus massiliensis, a candidate gene encoding a putative cellulose synthase, with a homology with the BcsA domain, one of the catalytic subunits of the bacterial cellulose synthase, but with a low level of homology. This gene was transcribed during the replicative cycle of Pandoravirus massiliensis, but several arguments run counter to this hypothesis. Indeed, even if this gene is present in other Pandoraviruses, the one of the strain studied is the only one to have this BcsA domain and no other enzymes involved in the synthesis of cellulose could be detected, although we cannot rule out that such genes could have been undetected among the large proportion of Orfans of Pandoraviruses. As an alternative, we investigated whether Pandoravirus could divert the cellulose synthesis machinery of the amoeba to its own account. Indeed, contrary to what is observed in the case of infections with other giant viruses such as mimivirus, it appears that the transcription of the amoeba, at least for the cellulose synthase gene, continues throughout the growth phase of envelopes of Pandoravirus. Finally, we believe that this scenario is more plausible. If confirmed, it could be a unique mechanism in the virosphere.