Enzymes that release 5'-deoxyribose-5-phosphate (dRP) residues from preinci
sed apurinic/apyrimidinic (AP) DNA have been collectively termed DNA deoxyr
ibophosphodiesterases (dRPases), but they fall into two distinct categories
: the hydrolytic dRPases and AP lyases. In order to resolve a number of con
flicting reports in the dRPase literature, we examined two putative hydroly
tic dRPases (Escherichia coli exonuclease I (exo I) and RecJ) and four AP l
yases (E. coli 2,6-dihydroxy-5N-formamidopyrimidine (Fapy) DNA glycosylase
(Fpg) and endonuclease III (endo III), bacteriophage T4 endonuclease V (end
o V), and rat polymerase beta (beta-pol)) for their abilities to (i) excise
dRP from preincised AP DNA and (ii) incise AP DNA. Although exo I and RecJ
exhibited robust 3' to 5' and 5' to 3' exonucleolytic activities, respecti
vely, on appropriate substrates, they failed to demonstrate detectable dRPa
se activity. All four AP lyases possessed both dRPase and traditional AP ly
ase activities, albeit to varying degrees. Moreover, as best illustrated wi
th Fpg, AP lyase enzymes could be trapped on both preincised and unincised
AP DNA using NaBH4 as the reducing agent. These results further support the
assertion that the catalytic mechanism of the AP lyases, the beta-eliminat
ion reaction, does proceed through an imine enzyme-DNA intermediate and tha
t the active site residues responsible for dRP release must contain primary
amines. Further, these data indicate a biological significance for the bet
a-elimination reaction of DNA grycosylase/AP lyases in that they, in concer
t with hydrolytic AP endonucleases, can create appropriate gapped substrate
s for short patch base excision repair (BER) synthesis to occur efficiently
. (C) 2000 Elsevier Science B.V. All rights reserved.