We have shown that SUB1 is required for NHEJ repair of DNA breaks in plasmids, but not in chromosomes. The fact that deletion of SUB1 does not reduce accuracy of NHEJ indicates that Sub1 functions differently from the KU complex, which is required for accurate ligations. In fact, the dispensable role of Sub1 in chromosomal repair suggests that Sub1 is not a core component of NHEJ, since KU and other NHEJ factors, but not Sub1, are absolutely required for NHEJ repair of chromosomal breaks.
Previously a microarray-based screen has been performed by Ooi et al
to identify new genes that are required for repairing plasmid DSBs 
. While the screen has been proven useful, successfully identifying NEJ1
as a novel component in NHEJ, SUB1
was not reported. However, several mutants including rad9Δ
, rad17 Δ
, rad24 Δ
, and srs2Δ
that are known to have reduced NHEJ efficiencies were also not identified in the microarray-based screen 
. Furthermore, the authors reported that 13% of haploid mutants were not analyzed due to high signal noise. Thus, it is possible that SUB1
was missed in this screen.
It is interesting to note that the sub1Δ
mutant, unlike the yku70Δ
mutant, is deficient in repairing the blunt-end DSB in plasmid DNA. It has long been found that blunt-end DSB is repaired at a low efficiency and that knockout mutations of Yku70 or Yku80 can increase the efficiency of repair 
. Furthermore, Westmoreland et al
found that yeast cells inefficiently survive PvuII-induced chromosomal breaks and this survival is not affected by deletion of RAD52
, suggesting repair of blunt-end DSBs is independent of HR or canonical NHEJ. It remains elusive how blunt-end DSBs are repaired independently of KU. Our results show that deletion of SUB1
reduces the efficiency of the already-inefficient repair of blunt-end DSBs. However, in the absence of YKU70
, repair of blunt-end DSBs is higher than wild type, regardless of whether Sub1 is functional or not. We speculate that when repair is channeled into the KU-independent pathway, Sub1 no longer plays a role.
Given the pronounced effect of SUB1
on repair all types of DSBs in plasmids, it is surprising to find that Sub1 is dispensable for repair of chromosomal breaks, whether they are induced by HO or I-SceI. While the reasons remain to be determined, this is reminiscent of the work by Batta et al.
They found that PC4 enhances DNA ligation in vitro
but human cells with PC4 knocked down are not sensitive to DSB inducing reagents 
. In this study it was shown by atomic force microscopy that PC4 has activity that bridges DNA ends 
. Thus the proposed model has been that PC4 facilitates ligation of plasmid DSBs by bridging the free DNA ends. Indeed, in yeast the ends of chromosomal breaks have been shown to be held in place by chromatin structures after chromosomal breaks are induced 
, whereas the ends of plasmid DNA are produced prior to transformation and therefore not held in close proximity. This distinct difference between ends of chromosomal breaks and plasmid breaks may underlie the differential requirement of SUB1
in repair of plasmid DSBs versus chromosomal DSBs.
We do not rule out other possible reasons that may explain the differential requirement of SUB1 in repairing plasmid DNA and chromosomal DNA. For example, as a DNA binding protein, Sub1 could potentially affect DNA resection. DNA end resection can have different outcomes in long chromosomes versus short DNA of plasmids: resection at both ends of the plasmid will destroy it, resulting in an apparent low repair efficiency. Alternatively, the effects of Sub1 can be indirect as Sub1 is a transcription factor. However, the plausible targets of transcriptional regulation are not expected to be any known NHEJ factors that are required for repair of chromosomal breaks, since the consequences of SUB1 deficiency is different and unique compared to the consequences due to loss of known NHEJ factors.
The plasmid repair assay has been widely used to test the ability of the cells to repair DNA breaks 
. Often the conclusions are extended to implicate the ability of the cells to repair chromosomal breaks. Here we provide an example that such implications may not be warranted, since the strong genetic requirement for Sub1 in NHEJ of transformed plasmid DNA does not extend to repair of endonuclease induced chromosomal double-strand DNA breaks. In summary, our results clearly demonstrate that repair of transformed DNA and chromosomal or chromatin associated DNA is quite different in their requirement for Sub1. While many of the genetic requirements are identical, the ability to genetically separate the two methods of assessing NHEJ clearly demonstrates that the two substrates are not repaired in an identical manner.