The goal of this screen was to identify novel 5′–3′ exonucleases that degrade linear DNA and endonucleases that cleave branched DNA structures expected to formed during DNA repair or replication. At the time the screen was initiated, the identity of exonucleases involved in the 5′–3′ resection of DSB ends was unknown, and the only activities known to cleave DNA substrates containing a Holliday junction were the mitochondrial protein Cce1 and the Mus81-Mms4 complex.
From the screen of pooled strains we identified several known nucleases, including Rad27, Apn1, Pso2, Rex2 and Rat1. Although Rat1 and Rex2 are exoribonucleases, they appear to have activity on DNA substrates, similar to the Rat1-related protein Kem1/Sep1 
. The 5′–3′ exonuclease Exo1 was not detected in the screen, but weak activity was found for the fusion protein individually purified (data not shown). The low activity could be because the protein is rapidly cleaved during extract preparation and the nuclease activity resides in the N-terminal region of the protein, not the C-terminal region that would have been recovered using the tag 
. The failure to detect some known nucleases by the pooling strategy raises the possibility that unknown nucleases might have been missed in the screen if they have very low activity, or are inactive in the presence of 1 mM Mn2+
included in the reaction buffer. It is also possible that heteromeric nucleases would not be identified because only one component is over-expressed; however, we did identify a Rat1-associated protein, Rai1, in this screen. Of more concern is the lack of representation by nucleases in the library. The Dna2, Kem1 and Yen1 nucleases, as well as the catalytic subunits of DNA polymerase δ (Pol δ) and Pol ε, which have intrinsic 3′–5′ exonuclease activity, are absent from the MORF library. Dna2 has recently emerged as a candidate for 5′–3′ resection of the ends of DSBs by genetic and biochemical assays 
, and Yen1 was identified as a HJ resolvase using the TAP-tagged strain collection 
. Because the MORF library requires amplification of plasmids in E. coli
it is possible that nucleases expressed at low levels in E. coli
are toxic to the organism and thus are under-represented in the library. Most plasmid libraries are propagated in E. coli recA
strains and low expression of a nuclease might be particularly toxic to recA
mutants that have no capacity for homology-dependent repair.
Although no new nuclease activities were identified from this screen, we did find phosphatase activity associated with a protein of unknown function, YOR283w, and with the recently characterized protein Det1. Our preliminary biochemical characterization of these two proteins suggest they are able to hydrolyze phosphate from several substrates, including DNA, ATP, 3-PGA, α-napthyl phosphate and ρ-nitrophenyl phosphate, and that Det1 is active at acid pH whereas the phosphatase encoded by YOR282w has a broad pH range. Both proteins contain the conserved RHG motif characteristic of members of the histidine phosphatase superfamily, which includes phosphoglycerate mutases and phosphatases 
. The conserved histidine residue is transiently phosphorylated during catalysis. Members of this family identified by sequence homology are frequently referred to as mutases, but most members are in fact phosphatases 
was recently identified as a gene required for non-vesicular transport of sterols in both directions between the endoplasmic reticulum and plasma membrane 
. The precise function in sterol transport is unknown, and our biochemical characterization does not identify potential substrates. A recent study identified the insect enzyme ecdysteroid phosphate phosphatase and the related human protein Sts-1 as members of the histidine phosphatase superfamily, and demonstrated phosphatase activity using ecdysteroid and steroid phosphate substrates 
. Intriguingly, YOR283w was identified as essential for the viability of the sec14 kes1
double mutant 
. Sec14 is the major phosphatidylcholine (PC)/phosphatidylinositol (PI) transfer protein in budding yeast and is essential for vesicular transport from the Golgi apparatus. Kes1 is an oxysterol binding protein family member that binds to sterols and phosphatidylinositol 4-phosphate and regulates Golgi apparatus-derived vesicular transport. This finding raises the possibility that the physiological substrate for YOR283w could be a PI phosphate. Analysis of the activity of YOR283w and Det1 on specific phosphorylated lipids might be informative, as well as genetic studies to investigate the interaction with other components of the vesicular and non-vesicular transport pathways.