Previously, the purified recombinant 2A proteases (2A(pro)) of coxsack
ievirus B4 (CVB4) and human rhinovirus type 2 (HRV2) were shown to cle
ave synthetic peptides derived from human or rabbit eIF4G as well as e
IF4G protein purified from rabbit reticulocytes. These results were in
contrast to previous evidence which supported the view that eIF4G cle
avage activity in poliovirus-infected HeLa cells required a cellular f
actor(s) activated by poliovirus (PV) 2A(pro). In the present study, r
ecombinant PV 2A(pro) was shown to cleave either rabbit or human eIF4G
or their derived peptides in direct cleavage reactions, but cleaved t
he 4G-derived peptides with 100-fold lower efficiency than with a pept
ide derived from the poliovirus polyprotein. in these experiments, up
to 25-fold molar excess of 2A(pro) over eIF4G protein was required to
cause greater than 50% cleavage. CVB4 2A(pro) was also tested in pepti
de cleavage assays under the same conditions as PV 2A(pro) and was fou
nd to cleave all eIF4G substrates with efficiencies similar to PV 2A(p
ro). Finally, cleavage reactions utilizing recombinant eIF4G containin
g a G486E substitution at the cleavage site for CVB4 and HRV2 protease
s resulted in drastically reduced cleavage by PV 2A(pro), similar to t
he reduction previously seen with HRV2 and CVB4 2A(pro), confirming th
at all three viral 2A proteases recognize the same cleavage site on eI
F4G. These data show that PV 2A(pro) can directly cleave eIF4G in vitr
o with efficiencies similar to those of CVB 2A(pro), but cleavage effi
ciency of eIF4G is approximately 1000-fold lower than cleavage of a pe
ptide derived from the authentic 2A cleavage site on the poliovirus po
lyprotein. (C) 1998 Academic Press.