Shiga-like toxin I is a polynucleotide:adenosine glycosidase.

Abstract

Sir, Shiga toxin from Shigella dysenteriae type I and shiga-like toxins produced by several serotypes of Escherichia coli are involved in bacillary dysentery, haemorrhagic colitis and haemolytic uraemic syndrome (O'Brien et al., 1992, Current Topics Microbiol Immunol180: 65–94), although their exact role in the pathogenesis is not fully understood. These cytotoxins are composed of an enzymatic A subunit linked to a multimeric receptor-binding B subunit. The enzymatic activity of both shiga and shiga-like toxins so far described is surprisingly similar to that identified for ricin and other plant ribosome-inactivating proteins (Barbieri et al., 1993, Biochim Biophys Acta1154: 237–282): adenine is removed from the major rRNA at a specific adenosine residue thus blocking ribosome function (Endo et al., 1988, Eur J Biochem171: 45–50). For over a decade it has been common belief that ribosome-inactivating proteins could act only on rRNA within ribosomes. Recently, however, it has been shown that all ribosome-inactivating proteins from plants also depurinate other polyribo- and polydeoxyribonucleotides besides rRNA in ribosomes (Barbieri et al., 1994, Nature372, 624; Barbieri et al., 1997, Nucleic Acids Res25, 518–522), and the denomination polynucleotide:adenosine glycosidases has been proposed. This prompted us to verify if shiga-like toxin I has polynucleotide:adenosine glycosidase activity. Here we report that indeed activated shiga-like toxin I extensively depurinates DNA from two different sources. The adenine release by activated shiga-like toxin I is reported in Table 1 and Fig. 1. The depurination rate observed with herring-sperm DNA as substrate is similar to the highest reported for plant ribosome-inactivating proteins (Barbieri et al., 1997, ibid ), and in particular is much higher than the rates observed for plant ribosome-inactivating proteins consisting of an A and a B subunit (ricin and related toxins). Activity was detected only with the activated toxin and not with the whole toxin, at odds with the observation made with the above-cited plant toxins, for which removal of the receptor-binding B chain has no effect on the depurination rate on DNA (Barbieri et al., 1997, ibid ). DNA from calf thymus was less sensitive than herring-sperm DNA to enzymatic depurination by shiga-like toxin I, possibly because of the different preparation procedures used, which may yield DNA with non-identical physicochemical properties. The possibility that the activity observed is due to a contaminant and not to the shiga-like toxin itself is ruled out by (i) the high grade of purification of the toxin prepared by receptor–ligand chromatographic interaction, and (ii) the fact that a significant activity was observed only after activation of the toxin. Shiga-like toxin I also depurinated naked rRNA, tRNA and poly(A), albeit less efficiently. . Adenine released from herring-sperm DNA by activated shiga-like toxin I. Reaction conditions were as described in the table legend except for the amount of shiga-like toxin I added. The results described here indicate that shiga toxins may be classified as polynucleotide:adenosine glycosidases. The high depurination rate of shiga toxins on various substrates and particularly DNA may be relevant to the mechanism of action of this medically important toxin.