DNA LIGASE IV STRUCTURALLY SUPPORTS END JOINING REPAIR OF DNA DOUBLE STRAND 
BREAKS. 

By 

Noah J. Goff 

A DISSERTATION 

Submitted to 
Michigan State University 
in partial fulfillment of the requirements   
for the degree of 

Cell and Molecular Biology – Doctor of Philosophy 

2024 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ABSTRACT 

DNA  Double  strand  breaks  (DSBs)  are  highly  genotoxic  lesions  induced  by  external  agents 

(ionizing  radiation  or  chemotherapeutic  drugs),  cellular  processes  (DNA  replication  defects, 

reactive  oxygen  species,  recombination  intermediates  in  B/T  cell  development),  or  through 

increasingly prevalent gene editing techniques (CRISPR/Cas9). Mammals have evolved two major 

pathways for repairing DSBs: Homologous Recombination (HR) and Non-Homologous End Joining 

(NHEJ). While NHEJ is frequently referred to as the “error prone” pathway, recent biochemical 

and  in  vitro  single-molecule  imaging  studies  have  revealed  mechanisms  to  protect  ends  from 

excessive end-processing through a two-stage synaptic model. First, DNA ends are synapsed by 

NHEJ  factors  in  a  long-range  complex  (LRC)  with  ends  held  ~115  Å  apart.  The  LRC  recruits 

downstream factors and only permits highly regulated end processing of bulky adducts. Second—

following  the  recruitment  of  the  NHEJ-associated  DNA  Ligase  IV  (L4)—ends  are  brought  into 

direct proximity for final end-processing and ligation. In this dissertation, I report that catalytically 

inactive L4 promotes significant amounts of end joining in cell models, supporting the recent two-

stage synaptic model for NHEJ. Furthermore, I characterize repair products from cells expressing 

catalytically  inactive  L4  showing  that  repair  is  significantly  more  mutagenic  than  in  cells 

expressing active L4. Finally, I identify multiple interfaces in L4’s DNA binding domain critical for 

maintaining  promoting  repair,  providing  insight  into  how  L4  structurally  supports  synapsis  of 

ends.   

 
 
This dissertagon is dedicated to my parents. It’s 
impossible to overstate the significance of your 
support over the past 5 years. 

iii 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ACKNOWLEDGMENTS 

The past five years of research and gme spent wrigng for my PhD would not have been possible 

without the generous and thoughiul support from dozens of people. It absolutely takes a village 

and I’m extremely grateful to the amazing community of people who have helped me over the 

past  ~5  years.  Special  thanks  to  my  commijee  for  your  valuable  advice  and  well-thought 

quesgons throughout my gme here at MSU; Dr. Daniel Vocelle from the MSU Flow Cytometry 

core for your dedicagon to teaching and technical know-how; the friends I’ve made through the 

BMS and CMB programs for their willingness to discuss science at any gme of day (preferably over 

a beer); Dr. Jens Schmidt’s lab for being so accommodagng with microscope gme (and room in 

their biosafety cabinets); the CMB program—specifically Dr. Margaret Petroff for being a fantasgc 

graduate director and Alaina Burghardt for being the best program coordinator at the university; 

Emily Durocher for her unerring ability to make me smile amer a long day; and finally my family 

for  being  there  since  day  one,  I’m  so  happy  with  how  this  journey  has  turned  out  and  It’s 

impossible to quangfy how much having you nearby has meant along the way (not to mengon all 

of the free food).  

iv 

 
 
 
 
TABLE OF CONTENTS 

Chapter 1: Evolving models of non-homologous end joining. ........................................................ 1 
REFERENCES ..................................................................................................................... 18 

Chapter 2: Catalygcally inacgve DNA Ligase IV promotes DNA repair in living cells.  ................. .25 
REFERENCES ..................................................................................................................... 51 

Chapter 3: New insight into how the DNA binding domain of DNA Ligase IV facilitates end-
joining, independent of its catalygc acgvity.  ............................................................................... 56 
REFERENCES ..................................................................................................................... 78 

Chapter 4: Discussing the two-stage synapgc model of NHEJ.  .................................................... 81 
REFERENCES ..................................................................................................................... 95 
APPENDIX: A list of authorship contribugons to other papers.  ....................................... 97 

v 

 
 
 
 
 
 
 
Chapter 1: Evolving models of non-homologous end joining.  

By 

Noah J. Goff 

Introducgon and literature review 

  1 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
INTRODUCTION 

DNA double strand breaks (DSBs) are extremely genotoxic lesions generated by a host of 

cellular  processes  and  exogenous  agents.  Repair  of  DSBs  is  essengal,  and  consequences  of 

unrepaired  breaks  can  impair  core  nuclear  funcgons  (transcripgon,  DNA  replicagon),  stall  cell 

cycle progression by triggering DNA damage checkpoints (DSB forming drugs form the backbone 

of  many  chemotherapeugc  treatments),  compromise  genomic  stability  through  aberrant 

genomic translocagons, and ulgmately an unrepaired break can lead to programmed cell death. 

Furthermore,  failures  during  repair  processes  can  further  compromise  genomic  integrity, 

including short insergons/delegons, copy number variagons, large intrachromosomal delegons, 

and  inappropriate  pairing  of  ends  can  exacerbate  large  chromosomal  delegons.  My  work  has 

focused  on  DNA  end  joining  pathways  and  the  structural  role  that  DNA  Ligase  IV  (L4)  plays  in 

supporgng end joining through the canonical non-homologous end joining pathway (NHEJ) and 

how loss of L4 catalygc acgvity promotes mutagenic alternagve end joining (alt-EJ).  

DSB Repair pathways aim to accurately resolve breaks. DSBs are generally repaired through two 

mechanisms:  NHEJ  and  homologous  recombinagon  (HR).  Briefly  (DSB  repair  pathways  are 

discussed in greater detail later in this chapter), HR uglizes a nearby homologous DNA sequence 

to perform high fidelity templated repair. NHEJ serves to resolve DSB repair through direct ligagon 

of two free DNA ends (hence “end joining”). First, a DNA end joining pathway must sense free 

DNA ends, then generate a ligagon substrate by modifying chemistry at the ends, and eventually 

catalygcally seal the break by ligagng both broken phosphodiester backbones. NHEJ solves a core 

challenge for cells: how to repair breaks that occur without a nearby homologous template (which 

is only present immediately following DNA replicagon). The flexibility of NHEJ (it can act on DNA 

ends through nearly the engre cell cycle) presents several challenges; namely, how to conserve 

sequence idengty (prevengng indels) and to conserve larger chromosomal structures (prevengng 

large genomic translocagons). Recent work has highlighted the remarkable fidelity of NHEJ and 

has sought to idengfy mechanisms to minimize errors generated by repair.  

Conserved NHEJ structural factors promote repair. A structural heterodimer formed by Ku70 and 

Ku80  serves  both  as  the  key  inigal  sensor  of  DNA  damage  as  well  as  an  anchor  point  for 

downstream  factor  recruitment.  Early  studies  into  Ku  revealed  that  each  factor  exists 

  2 

 
independently of each other but rapidly bind to DSBs and promote recruitment of other factors 

(citagons). The two “C” shaped proteins clamp onto a double strand break, forming a donut shape 

that mediates protein-protein interacgons with DNA-PKcs (forming the DNA-PK holoenzyme) and 

other repair factors (frequently though interacgons between Ku binding mogfs (KBM) and BRCT 

domains on other repair factors). In human cell strains, loss of either Ku gene is lethal, resulgng 

in telomere fusions and cell death (1, 2).   

Interesgngly,  the  heterodimeric  ring  formed  by  Ku70/Ku80  is  extremely  stable  and 

allowed to slide upstream of the break site—like beads on a string. As such, several mechanisms 

exist  to  prevent  overloading  of  Ku  onto  a  DNA  an  end  and  there’s  significant  discussion  and 

ongoing research into how these gghtly-bound DNA repair donuts are removed from repaired 

ends (3, 4).  Several structural features of Ku have been well characterized, including the Ku80 C-

Terminal Domain (in pargcular the final 14 residues) as being crigcal for funcgonal interacgons 

with DNA-PKcs to promote end joining (5–7).  

X  ray  cross  complimengng  protein  4  (XRCC4)  serves  both  as  a  key  structural  factor  in 

supporgng NHEJ synapsis and as the obligate binding partner of Ligase IV (L4). Two molecules of 

XRCC4 and one molecule of L4 form the ligagon complex LX4, together a core NHEJ factor that is 

necessary for robust end joining. Each XRCC4 homodimer is shaped with an N-terminal globular 

head  and  long  C-terminal  alpha-helix  stalk  that  mediates  its  interacgon  with  L4  (8,  9).  When 

homodimerized, the two XRCC4 head domains pair up with the C-terminal alpha-helices wound 

around  each  other  1-2  gmes  (Figures  1,  2,  4).  Loss  of  XRCC4  or  L4  results  in  extreme 

hypersensigvity to DSB inducing agents (both as a result of radiagon and drug induced damage) 

(10–16). Addigonally, XRCC4 possesses a disordered c-terminal “tail” region that has been shown 

to transiently interact with L4’s DBD and directly with DNA (17). As part of the LX4 complex, XRCC4 

facilitates interacgons with XLF and Ku, providing support for end synapsis observed in recent 

Cryo-EM structures (18, 19).  

  3 

 
Figure 1: LX4 Has been observed in the Ku80-Mediated Long-Range Complex when incubated 
with PAXX. PDB 8BHY: dimeric structure of DNA-PKcs (grey), Ku70/Ku80 (grey/black), L4 (green), 
XRCC4 (purple), XLF (orange), and PAXX (red) work together to synapse DNA (orange) in a long-
range  synapse.  Here,  the  LX4  complex  are  posigoned  away  from  the  break,  with  interacgons 
between  DNA-PK  and  PAXX,  along  with  trans  interacgons  between  Ku80  and  DNA-PKcs 
supporgng synapsis. 

Addigonal structural factors XRCC4-like factor (XLF) and paralog of XRCC4 and XLF (PAXX) 

also work to synapse ends. XLF is a central mediator of synapsis in recent Cryo-EM structures, 

with  the  “XLF-mediated  dimer”  consisgng  of  a  group  of  structures  where  an  XLF  homodimer 

serves as a major contact in bridging the end-bound DNA-PK holoenzyme and LX4 complex (18, 

19).  The  XLF-mediated  dimer  appears  to  fill  the  role  of  “end  protecgon  complex”  originally 

proposed as the primary role of the long-range complex (LRC) proposed by Loparo and colleagues 

(20, 21). While complete ablagon of XLF has variable impacts in different organisms, loss of XLF 

is remarkably synthegcally lethal with a host of other DNA factors, and many studies implicate its 

role in promogng high-fidelity repair (12, 22–28). While typically not considered a “core” NHEJ 

factor (PAXX is funcgonally redundant with XLF), PAXX has recently been observed supporgng LRC 

  4 

 
 
 
via Cryo-EM, including a structure very similar to the Ku80-mediated dimer that includes PAXX 

and LX4 (providing evidence that the NHEJ dimers may all be capable of recruigng core factors 

required for transigon into the SRC). Despite its apparent redundancy with XLF, PAXX has been 

idengfied in a wide range of species including plants (29).  

The two core enzymagc NHEJ factors, DNA-PKcs and L4, each play unique structural roles 

in supporgng NHEJ. DNA-PKcs is a large 465 kDa serine/threonine protein kinase that acts as the 

primary  structural  regulator  of  DNA  end  access.  Furthermore,  DNA-PKcs  autophosphorylagon 

plays  a  key  role  in  driving  conformagon  changes  in  NHEJ  complexes,  with  well-studied 

phosphorylagon mogfs directly regulagng access to DNA ends (11, 30–32). While not present in 

prokaryotes  (many  prokaryotes  have  end-joining  with  distant  orthologs  of  the  Ku  and  L4 

molecules)  (33,  34),  DNA-PKcs  is  found  in  a  wide  range  of  eukaryotes  with  many  of  its  key 

structural mogfs (and autophosphorylagon sites) conserved (35).   

Mechanism and domains of mammalian DNA ligases: There are three mammalian DNA ligase 

genes: Ligase I (L1), Ligase III (L3) and Ligase IV (L4). All three genes share a conserved catalygc 

core  consisgng  of  a  reacgve  Nucleogdyl  Transferase  Domain  (NTD)  flanked  by  a  DNA-Binding 

Domain (DBD) and Oligonucleogde-Binding-Fold (OBF). While all three ligases serve DNA repair 

funcgons in the nucleus, L3 also provides mitochondrial DNA repair funcgon through an alternate 

translagonal  start  site.  This  replaces  the  extreme  N-terminus  of  the  protein,  subsgtugng  the 

nuclear localizagon signal for a mitochondrial localizagon signal (36). Each mammalian ligase is 

ATP dependent, capable of sealing a single nick in a deoxyribose-phosphate backbone (37–39). 

The  general  reacgon  progresses  in  three  steps:  the  ligase  1)  hydrolyzes  ATP  to  adenylate  a 

conserved lysine residue in NTD acgve-site, 2) binds to a nick and transfers the AMP moiety from 

its  lysine  to  the  5’  phosphate,  then  3)  supports  a  nucleophilic  ajack  from  the  remaining  3’ 

hydroxyl end to the 5’ PO4-AMP bond resulgng in a free AMP, uncharged ligase, and sealed single-

strand break (37–41). Notably, the high degree of domain-conservagon and architecture across 

all three DNA ligases implies that these genes share a common evolugonary ancestor (37, 39–

41). Each Ligase NTD acts primarily through a conserved catalygc lysine that self-adenylated in 

the presence of ATP, providing energy to catalyze the break (37, 39).  

  5 

 
 
Ligase IV is the NHEJ-exclusive ligase and well adapted for DSB repair. While Ligases I and III are 

frequently referred to as redundant, L4 serves exclusively in NHEJ-mediated DSB repair. While 

each of the mammalian ligases share a conserved mechanism and degree of structural homology 

in their catalygc cores, L4 has evolved several unique structural features to promote ligagng ends 

from a DSB: 

First, L4 has evolved a pair of tandem BRCT (BRCA1 c-terminus) domains that facilitate its 

interacgons with NHEJ proteins. The tandem L4 BRCT domains are connected by a short, flexible 

linker termed the XRCC4 Interacgng Region (XIR) for its role intercalagng into the central alpha-

helix stalk of an XRCC4-homodimer (37–41). The XIR is both sufficient and necessary to mediate 

L4’s interacgon with XRCC4 (of note, L3 also has evolved n-terminal BRCT domains that mediate 

its interacgon with XRCC1). Second, unlike L1/L3, L4 rapidly adenylates following translagon, and 

nearly all the L4 molecules present within the nucleus are prepared to catalyze break repair (37). 

This readiness could serve as a mechanism to minimize the amount of gme required to perform 

the  inigal  ligagon  step—convergng  the  DSB  into  a  much  less  harmful  single-stranded  break. 

Interesgngly,  while  we  were  studying  a  catalygcally  inacgve  L4  (designed  by  mutagng  the 

conserved catalygc lysine), we observed residual ligagon acgvity in in vitro joining assays. This 

residual ligase acgvity is lost following mutagon of addigonal lysines posigoned close to the DNA 

end, indicagng that there is some flexibility within the catalygc core (10). To my knowledge, there 

have been no reports of alternate lysine adenylagon in L1 or L3.  

Third, L4 has a unique 8-amino acid loop structure in its DBD that promotes flexibility in 

DNA-substrate binding, including tolerance of mismatches, aberrant bases, and gaps in one end 

of the DNA substrate (42). This unique lesion tolerance is another representagon of the priority 

given to double strand break repair—with failure to catalyze repair resulgng in significant cell-

growth defects (and loss of L4 is generally lethal) (14, 43). Notably in cell line models, loss of L1 

or  nuclear  L3  generally  does  not  result  in  cell  fitness  defect  or  even  notable  drug-sensigvity 

phenotypes (15, 44). The two single-stranded ligases are funcgonally redundant and loss of both 

is synthegcally lethal; in contrast, L4 does not appear to complement DNA repair outside of NHEJ.  

DNA-PKcs: A Synopsis Beyond Synapsis. The following secgons have been reproduced from a 

review  that  I  co-wrote  with  another  graduate  student,  Maria  Mikhova,  alongside  our  advisors 

  6 

 
Katheryn Meek and Jens Schmidt (45). The reproduced secgons were primarily dramed by me and 

edited  as  a  group.  The  figure  referenced  in  this  secgon  was  originally  produced  by  Dr.  Meek, 

originally  produced  as  as  “Figure  2”  within  the  review.  I’ve  also  removed  chapter  numbers, 

updated citagon formats, citagon numbers, and abbreviagons to be consistent with the styles 

used in my dissertagon.   

A two-step model of DNA end synapsis in long and short-range NHEJ complexes. An emerging 

consensus in the field is a two-stage synapgc model for NHEJ driven by the presence and catalygc 

acgvity  of  DNA-PK.  This  model  was  inigally  proposed  by  Loparo  and  colleagues  who  have 

extensively employed X. Laevis egg extracts to measure NHEJ-mediated synapsis of two DNA ends 

using  single-molecule  fluorescence  resonance  energy  transfer  (smFRET,  simultaneously 

measuring  synapsis  and  proximity).  Graham  et  al.  provided  the  first  evidence  and  model  for 

disgnct  long-range  and  short-range  synapgc  complexes  [LRC,  SRC]  (46).  They  proposed  a  LRC 

which  notably  holds  DNA  ends  far  apart  (>100  Å);  these  LRCs  are  short-lived  (on  the  order  of 

several seconds), and only a subset progress to SRCs where ends are brought in direct proximity 

(as assessed by FRET). Formagon of LRCs requires only Ku and DNA-PKcs. Crigcally, in this study, 

acceptor  fluorescence  was  visualized 

independently  of  donor-mediated  FRET  allowing 

observagon of synapgc events at distances beyond the range of FRET detecgon (>100 Å). This 

study established that progression to the SRC required the presence of Ku, L4, XRCC4, XLF, and 

the catalygc acgvity of DNA-PKcs. Interesgngly, the catalygc acgvity of LX4, essengal for the final 

ligagon  step  in  NHEJ,  was  engrely  dispensable  for  SRC  formagon,  supporgng  an  important 

structural role for LX4 which had been proposed in previous studies (12, 47). Our recent cellular 

studies  confirmed that LX4’s catalygc acgvity is dispensable for progression from long-range to 

short-range complexes (10). Moreover, these findings imply that nuclear DNA ligase III (L3) can 

access DNA ends maintained in NHEJ complexes by the presence of catalygcally inacgve LX4,  a 

heretofore unappreciated level of cooperagvity between mammalian DNA ligases both in living 

cells (10) and in animal models (14). 

LX4 binding to DNA ends drives progression from long-range to short-range complexes. Work 

from Loparo and colleagues has recently addressed how NHEJ priorigzes ligagon (of compagble 

ends) over end processing, promogng error-free repair. A 3-color imaging system was uglized to 

  7 

 
independently assess LX4 binding at each end, the presence or absence of short-range synapsis, 

and  LX4  stoichiometry,  independent  of  its  end-binding  (48).  This  study  demonstrates  that  LX4 

dynamically binds DNA ends through residues in its DNA Binding Domain (DBD) prior to short-

range synapsis, potengally as a sensor for large end adducts that could impede ligagon or final 

end-processing. SRC assembly (as measured by FRET between DNA ends) occurs exactly at the 

gme when a single LX4 complex binds to the two DNA ends. Surprisingly, SRC formagon results 

in  evicgon  of  one  of  the  two  LX4  complexes  that  are  present  prior  to  short-range  synapsis. 

Immediately at the gme of short-range synapsis, loss of one LX4 complex was observed with the 

second LX4 residing for much longer consistent with the previously reported persistence of short-

range  complexes.  A  caveat  of  this  study  is  that  a  blunt-ended  DNA  substrate  was  uglized  that 

should be readily ligated, although LX4 end-binding progression to a short-range complex was 

similarly efficient with or without 5’ phosphates. An important quesgon remains as to whether 

the  change  in  LX4  stoichiometry  occurs  prior  to  strand  ligagon.  A  similarly  important  issue  is 

determining  how  LX4’s  dwell  gme  is  impacted  if  the  DNA  ends  require  fill-in  synthesis  or 

nucleolygc processing on one or both ends.  

An ajracgve hypothesis supported by robust previous studies (49, 50) would be that the 

immediate capacity to ligate paired ends of one of the DNA strands would funcgon even if the 

second strand requires substangal end-processing. Simply put, there is no bejer way to maintain 

synapsis of DNA ends than by sealing one strand. The unique characterisgc (among vertebrate 

DNA ligases) of ligase 4 as a single-turnover enzyme (51) is curious, and obviously infers that the 

ligase 4 molecule that resolves the first strand break does not (necessarily) resolve the second 

strand break (which could be recognized as a nick or single-stranded DNA gap). This may reflect 

LX4’s  important  structural  role  both  in  recruigng  other  enzymes,  and  in  promogng  stable 

synapsis. As noted above, emerging studies demonstrate not only that Lig3 can associate with 

NHEJ factors, but also that when LX4 is inacgvated by mutagon, Lig3 clearly funcgons to repair 

DSBs, in an inacgve LX4-dependent manner, in living cells and animals (10, 14).  

  8 

 
 
 
Figure 2: Numerous NHEJ complexes have been visualized by cryo-EM. Cartoons depicgng NHEJ 
long-range  complexes  (Ku80-medieated  and  XLF-dependent  dimer),  long-range  ATP  (LR-ATP) 
complex, DNA-PKcs homodimer, and SRCs. DNA-PKcs is colored pink, Ku-green, XLF-red, XRCC4-
blue, Lig4 (fuscia) (derived from PDB 6ZHE, PDB 7NFC, PDF 7LT3, PDB 8EZB, PDB 8BYH, PDB 7LSY, 
PDB 8EZ9).  

Cryo-EM structures of NHEJ synapZc complexes. This two-stage synapgc model of NHEJ has been 

strengthened  by  recent  structural  work  describing  Cryo-EM  structures  of  mulgmeric  NHEJ 

complexes posigoned around breaks. The published structures include descripgons of two major 

forms  of  NHEJ  complexes  that  synapse  DNA  ends  ~115Å  apart  (consistent  with  long-range 

synapgc complexes). These dimers are organized around two disgnct interfaces: one mediated by 

a trans interacgon between DNA-PKcs and Ku80’s extreme C-terminus, and the other dependent 

  9 

 
 
 
on a centrally placed XLF homodimer interacgng with Ku and the LX4 complex, which also involves 

a  large  interface  between  the  two  DNA-PKcs  molecules  (Figure  2).  These  dimers  have  been 

termed the Ku80-mediated dimer (or domain-swap dimer), and the XLF-dependent dimer. Inigal 

observagons of the Ku80-mediated dimer only included the components of the DNA-PK complex, 

which would be consistent with the LRC requiring only Ku and DNA-PKcs. However, further studies 

show that the Ku80-mediated dimer can also contain XLF, PAXX, and LX4, or XLF and LX4, or PAXX 

and LX4 (22). The XLF-dependent dimer was first observed with two DNA-PK protomers, two LX4 

complexes and one XLF homodimer. Recent studies have shown that PAXX can also assemble into 

this  complex  (22).  Figure  2  presents  cartoon  depicgons  of  the  dimers  (for  simplicity,  the 

complexes  that  include  PAXX  are  not  represented);  structural  models  of  these  complexes  are 

available in recent excellent reviews (52, 53)  

He and colleagues have shown that in the presence of ATP, the XLF-dependent dimer can 

progress  to  a  new  form,  termed  long-range  ATP  (LR-ATP)  (54).  In  this  complex,  ABCDE 

phosphorylagon has occurred (likely in trans) resulgng in a major conformagonal change: each 

DNA-PK protomer has a remarkably similar structure to the ABCDE phosphorylated monomeric 

complex with the DEB dissolved, and the phosphorylated ABCDE sites anchored by K/R residues 

in  M-HEAT.  These  authors  also  observed  a  DNA-PKcs  homodimer  that  shares  some  structural 

similarity  with  homodimers  of  other  PIKKs,  ATM  and  ATR.  They  hypothesized  that  this  dimer 

formed amer ATP-induced dissociagon of DNA-PKcs from Ku/DNA. Another idea proposed was 

that  this  homodimer  might  serve  as  a  DNA-PKcs  reservoir.  Notably,  although  these  DNA-PKcs   

dimers were observed without the other components of the XLF-dependent dimer, no complexes 

consistent with SRC were observed in the cryo-EM experiments with ATP. Cryo-EM structures of 

NHEJ complexes consistent with SRCs have only been obtained in experiments uglizing Ku, LX4, 

and XLF with a DNA substrate that facilitates end synapsis (18). The organizagon of this short-

range synapgc complex is remarkably like the organizagon observed in the XLF-dependent dimer, 

with  a  single  XLF  monomer  mediagng  the  synapsis  between  two  Ku-bound  DNA  ends,  and 

LX4/XLF/LX4 bridge promogng synapsis (Figure 2). Of note, in the short-range structure, although 

the catalygc domain of only one LX4 is structured and posigoned over the perfectly juxtaposed 

DNA  ends,  the  second  LX4  complex  is  clearly  present,  which  is  not  consistent  with  the 

 10 

 
 
stoichiometry discussed above. Sgnson et al. suggest that one LX4 complex is released at the gme 

of long-range to short-range transigon but noted that another LX4 may associate (48) (thus, two 

potengal  SRCs  are  depicted  in  Figure  2).  Finally,  a  DNA-PK  trimer  was  observed  (not  shown), 

consisgng  of  a  central  DNA-PK  protomer,  interacgng  with  another  to  form  an  XLF-dependent 

dimer, and concurrently interacgng with a third DNA-PK protomer via the Ku80-mediated domain 

swap interacgon. It is not engrely clear how this trimeric structure would funcgon, but at least 

two disgnct DSBs would be required (55). 

AlternaZve models of NHEJ synapsis. The observagon that DNA-PKcs catalygc acgvity is required 

for SRC formagon in the X. Laevis extract model has generated substangal discussion over the 

role of DNA-PKcs when presengng complete models for NHEJ synapsis. Although the idea that 

DNA-PK facilitates synapsis was proposed decades ago, originally by Chu and colleagues (56) and 

supported  by  others  (57),  in  other  elegant  smFRET  studies  using  purified  human  NHEJ 

components, stable synapsis of DNA ends is not dependent on DNA-PKcs6 (58–60) Sgll, other 

experimental systems using purified human proteins, demonstrate a clear role for DNA-PKcs in 

synapsis (61).  In the studies from Lieber and colleagues (58–60), adding DNA-PKcs decreased the 

total number of synapgc foci. These invesggators also examined the role of DNA end chemistry 

on  LX4  mediated  synapsis  and  ligagon;  their  studies  showed  a  strong  dependence  for 

5’phosphates for synapsis in the absence of DNA-PKcs (59). A caveat of this and other smFRET 

approaches is the inability to differengate between synapsis-mediated or covalent ligagon of the 

two oligonucleogdes. Further in vitro single-molecule work (58) provided evidence for a long-

range synapgc complex formed from purified human Ku70/80 and LX4. Addigon of purified XLF 

to  these  reacgons  induced  a  transigon  into  a  short-range  synapgc  complex,  a  consistent 

observagon with their earlier findings of DNA-PKcs-independent short-range complex formagon. 

Thus, addigonal work will be required to elucidate how disgnct NHEJ complexes facilitate repair.  

Sgll, it seems likely that there are complex, cell-wide effects of losing DNA-PKcs beyond 

its role in synapsing and regulagng access to DNA ends in NHEJ. There is ample evidence for a 

mulgfaceted role for DNA-PKcs in both directly synapsing ends and pargcipagng in the cell-wide 

DNA damage response; moreover, DNA-PK has been implicated to funcgon in numerous other 

cellular funcgons [reviewed in (62)]. One argument suggesgng that DNA-PKcs can be dispensable 

 11 

 
 
for  NHEJ  is  that  cells  lacking  DNA-PKcs  are  notably  less  sensigve  to  DSB  inducing  drugs  than 

isogenic cells lacking other core NHEJ factors (63, 64). However, this conclusion is complicated by 

the fact that cells lacking DNA-PKcs also lose significant expression of ATM, leading to an impaired 

apoptogc response (65, 66). In fact, loss of even one allele of ATM substangally rescues the severe 

phenotype of Lig4 deficient mice (67). 

There  are  also  obvious  examples  of  DNA-PKcs-independent  NHEJ.  Although  recent 

phylogenegc work from Lees-Miller et al. have established that DNA-PKcs is highly represented in 

many species far removed from vertebrates (35), there are sgll many species with funcgonal NHEJ 

pathways that lack DNA-PKcs homologues (for example, some insects, yeast, and bacteria) (68). 

Addigonally, the recently published cryo-EM structure of the putagve short-range complex was 

observed in cryo-EM experiments that lacked DNA-PKcs; thus, short-range complex assembly can 

occur in the absence of DNA-PKcs (18). Moreover, because of the iteragve nature of NHEJ, we 

must consider that there may be mulgple paths that lead to the same outcome: there is more 

than  one  way  to  ge  a  knot/ligate  an  end. In  sum,  while  recent  studies  and  technological 

advancements have provided new insights, substangal gaps in knowledge remain that must be 

resolved before a full understanding of how NHEJ funcgons in living cells to rapidly repair DSB 

while minimizing genome alteragon. 

Many,  but  not  all  NHEJ  end-processing  steps  occur  in  short-range  complexes.  Another 

transformagve study from Loparo and colleagues (21) provided compelling biochemical evidence 

that factors required for the formagon of SRCs [Ku, XLF, LX4, and the catalygc acgvity of DNA-PK] 

are also essengal to enable fill-in synthesis mediated by X-family polymerases [polƛ and polµ, 3’ 

adduct cleavage mediated by Tdp1, phosphorylagon of 5’ hydroxyls by PNKP, and the acgvity of 

an  unidengfied  5’>3’  exonuclease  on  5’  flaps  (21).  Moreover,  DNA  ends  were  observed  to  be 

protected from end-processing acgviges in the LRC.    

At first, these results might seem at odds with our conclusion that the two long-range 

complexes  promote  different  aspects  of  end-processing;  this  is  not  the  case.  The  mutagonal 

studies  suggest  that  the  SRC  derives  from  the  XLF-mediated  dimer;  the  enzymagc  acgviges 

absolutely  ascribed  to  the  SRC  by  Sgnson  et  al.,  (except  for  the  unidengfied  5'>3"  nuclease) 

should  all  promote  fill-in  end  processing.  So,  the  XLF-dependent  dimer  promotes  fill-in  end 

 12 

 
 
processing by promogng progression to the SRC. Part of the argument from Sgnson et al that 

nuclease  acgvity  is  limited  to  the  SRC  was  that  DNA-PK  inhibigon  blocked  nuclease  acgvity, 

presumably because this blocks progression to the SRC. However, DNA-PK inhibigon also blocks 

the ABCDE phosphorylated form of DNA-PK acgvagon that is requisite to Artemis acgvagon and 

potengally  required  for  the  acgvagon  of  other  nucleases  that  may  funcgon  in  NHEJ  (Mre11, 

Apollo)(1, 69).  Thus, what is less clear is what nucleases (besides Artemis) contribute to NHEJ 

and in what NHEJ complexes do these nucleases funcgon.  This concludes the work adapted from 

DNA-PK: A synopsis beyond synapsis.  

L4’s role in dimeric NHEJ complexes: In recent NHEJ cryo-EM structures, the XRCC4-L4 complex 

has  been  observed  in  variagons  of  both  classes  of  NHEJ  dimers,  with  the  catalygc  core  only 

observed  in  the  SRC  and  one  unpublished  version  of  the  Ku80-mediated  LRC  (4–6).  The  LX4 

complex  is  seen  to  interact  with  Ku70  and  Ku80  through  XRCC4’s  head  domain  and  L4’s  BRCT 

repeats  (Figure  1,  3)  (18,  19,  22,  71).  In  the  XLF-mediated  dimer,  LX4  instead  facilitates  an 

interacgon with an XLF homodimer spanning the gap between ends, bridging each Ku “donut” 

(there is also a conformagon change in DNA-PKcs to more gghtly bind ends with a trans PKcs-

PKcs  interface  supporgng  synapsis  in  addigon  to  LX4-XLF-LX4,  Figure  3).  The  Ku80-mediated 

dimer  was  inigally  observed  in  the  absence  of  LX4,  however  recent  studies  from  Chaplin  and 

colleagues incorporagng PAXX and LX4 have since resolved a version of the Ku80-mediated dimer 

with LX4 bound to DNA-PK, but not supporgng the synapse (in contrast to their role in supporgng 

the XLF-mediated dimer). The short-range complex structure contains two LX4 complexes, with 

one interacgng only though its LX4 structural domain and the other directly bound to the break 

(Figure 4)(18)—notably contradicgng recent smFRET work from Loparo and colleagues (48) that 

indicates one molecule of L4 leaves the break prior to SR synapsis.  

 13 

 
 
 
Figure 3: XLF-Mediated Long-Range Complex. PDB 7LSY: a second dimeric structure of DNA-
PKcs (grey), Ku70/Ku80 (grey/black), L4 (green), XRCC4 (purple), and XLF (orange), work 
together to synapse DNA (orange) in a long-range synapse. Here, LX4 synapses ends through 
interacgons with Ku80 and XLF, along with a DNA-PKcs trans interacgon in the head domain.  

 14 

 
 
 
 
 
Figure  4:  DNA  Ligase  IV  plays  a  pivotal  role  in  the  observed  CryoEM  Complex.  PDB  7LSY: 
Ku70/Ku80  (Black),  L4  (Green),  XRCC4  (Purple),  XLF  (Orange)  work  together  to  synapse  DNA 
(yellow)  in  a  ligagon-competent  state.  This  structure  features  two  LX4  complexes,  with  the 
rightmost L4 catalygc core bound to the DNA end. Only the structural domains (L4 BRCT, XRCC4 
homodimer) are observed for the second LX4 complex (lem).  

Diverse DNA substrates and the cell cycle dictate repair pathway choice: While it’s somewhat 

unconvengonal to think of DNA repair enzymes in terms of classical enzyme kinegcs—since the 

concentragon of a repair factor typically dwarfs the concentragon of DSBs—the model works well 

for considering DNA end chemistry as a key element in pathway choice. The diverse mechanisms 

that generate DSBs results in a very wide range of substrates—ranging from unique end adducts 

formed by many chemotherapeugc agents, secondary structures biased near the break, single-

ended DSBs formed at failed replicagon forks, to ends compromised by previous repair steps. To 

this end, eukaryogc organisms have evolved several mechanisms to sense breaks and engage the 

correct repair mechanism depending on substrate chemistry and cell cycle status.  

Homologous recombinaZon: The other major DSB repair pathway, homologous recombinagon 

(HR), serves as a slow, energy intensive, high-fidelity repair pathway that can repair nearly every 

variagon  of  DSB  chemistry  (with  the  notable  excepgon  of  programmed  DSBs  generated  in 

 15 

 
 
 
VDJ/CSR recombinagon).  The primary drawback to HR is gming: namely, it’s only available in the 

S/G2 phases of the cell cycle amer DNA replicagon generates a nearby homologous template for 

repair. This limits repair to acgvely dividing cells that encounter breaks behind replicagon forks. 

Furthermore, instead of directly repairing ends, HR first resects large stretches of DNA to generate 

long single-stranded ends. This resecgon (mediated by MRN and Exo1) serves to clear lesions and 

facilitate  homology  search  prior  to  synthesis  of  new  DNA—thereby  creagng  a  “high  fidelity” 

repaired  region  (assuming  the  homologous  template  matches  the  original  sequence).  The 

resecgon  and  homology  search  are  a  gme-intensive  processes,  requiring  mulgple  rounds  of 

processing (including resolugon of the DNA tetramer holiday juncgons) prior to finalizing repair.  

The  intense  processing  of  ends  directed  to  HR  results  in  intermediates  incompagble  with 

NHEJ;  as  such,  the  pathways  act  compeggvely  and  there  are  mulgple  mechanisms  to  restrain 

either pathway depending on context. These mechanisms act 1) to restrain NHEJ and prevent 

premature  ligagon  (possibly  by  prevengng  transigon  into  the  SRC)(21)  and  2)  acgvely  remove 

NHEJ factors from acgng on ends that are targeted towards HR (3, 72). Interesgngly, a recent live-

cell SMI study indicates the same aborgve mechanism engages in cells lacking L4 (73)—implying 

this mechanism may more broadly engage if NHEJ is incapable of rapidly transigoning ends into 

the SRC.  

AlternaZve end joining: (alt-EJ) serves as less well-defined collecgon of factors that’s available to 

join breaks not resolved by either of the “canonical” NHEJ or HR pathways. While the pathway is 

loosely defined, recent studies (including experiments reported in Chapter 3) suggest that DNA 

Polymerase Theta (Polθ) is a crigcal factor essengal for many alt-EJ repair events. Polθ is a large 

and  uniquely  contains  both  a  highly  error-prone  polymerase  domain  and  a  helicase  domain 

capable  of  bridging  ends  uglizing  1-5  bp  of  terminal  microhomology  (74).  The  dual  helicase-

polymerase  funcgon  is  generally  thought  to  facilitate  synthesis  through  extremely  difficult  to 

replicate stretches of DNA (g quadruplexes, other complicated secondary structures) in addigon 

to its role as a backup end-joining factor (74–76).   

Of note, Polθ prefers free DNA end substrates with long secgons of 3’ ssDNA (77, 78) and its 

end-joining  funcgons  likely  support  joining  of  failed  HR  intermediates  (where  resecgon  can 

produce long single-stranded ends that are poor substrates for canonical NHEJ). While further 

 16 

 
work  needs  to  be  done,  current  literature  supports  MMEJ—and  its  dependence  on  Polθ 

helicase/polymerase—as a key mediator of joining in the absence of NHEJ.  

Alt-EJ serves as an interesgng foil to NHEJ, being both lower fidelity and significantly slower. 

These qualiges reflect alt-EJ’s typical substrates: rare complex breaks that “fail” out of the two 

major  DSB  repair  pathways.  As  such,  there  is  lijle  evolugonary  pressure  for  alt-EJ  to  have 

developed  mechanisms  to  respond  to  many  breaks,  quickly.  This  is  in  stark  contrast  to  NHEJ, 

where factors have evolved to be both fast, highly regulated, and relagvely accurate. It may be 

more  accurate  to  characterize  alt-EJ  as  a  collecgon  of  backup  mechanisms  that  may  really  be 

secondary funcgons of many other DNA repair pathways working together to resolve DSBs. Many 

factors  hypothesized  to  resolve  breaks  in  the  absence  of  NHEJ  have  notable  roles  in  other 

pathways. For example, there is sgll some residual end joining funcgon in the absence of L4, but 

neither Ligase I nor Ligase III becomes essengal following loss of L4 (nuclear loss of both L1 and 

L3  is  synthegcally  lethal  regardless  of  L4  status).    Furthermore,  the  slower  kinegcs  of  MMEJ 

indicate  that  it  is  not  likely  in  direct  compeggon  to  NHEJ  or  HR  (which  do  act  in  direct 

compeggon),  and  instead  works  as  a  “catch-all”  mechanism  that  only  acts  if  higher-fidelity 

pathways are unable to resolve a pargcular break in a gmely manner.  

CONCLUSIONS 

I  would  like  to  leave  this  chapter  highlighgng  the  remarkable  advancements  in  our 

understanding  of  NHEJ  over  the  course  of  my  PhD.  When  I  began,  there  was  significant 

disagreement over DNA synapsis models—with arguments over the significance of DNA-PKcs vs 

XRCC4/XLF filaments as a primary synapgc mechanism. The applicagon of clever SMI experiments 

to address NHEJ factor kinegcs and dedicagon of Cryo-EM groups to generate 3D-maps of NHEJ 

complexes has truly deepened the field’s appreciagon for how flexible the pathway can be. I look 

at  this  project  through  the  lens  of  those  advancements.  From  March  1st,  2020,  I’ve  sought  to 

understand L4’s structural role in supporgng NHEJ—before we understood anything about the 

disgnct LR complexes or true organizagon of the SR complex. I’ve since focused my PhD work—

including  much  work  that  is  not  included  in  this  dissertagon—through  the  lens  of  structure-

funcgon relagonships between NHEJ’s structural factors (XRCC4, XLF, Ku70/80) and L4, advancing 

our understanding the structural role of L4 and how it promotes end joining beyond catalysis.  

 17 

 
 
1.  

2.  

3.  

4.  

5.  

6.  

7. 

8.  

9.  

REFERENCES 

Sonmez,C.,  Toia,B.,  Eickhoff,P.,  Matei,A.M.,  El Beyrouthy,M.,  Wallner,B.,  Douglas,M.E., 
de Lange,T. and Lojersberger,F. (2024) DNA-PK controls Apollo’s access to leading-end 
telomeres. Nucleic Acids Research, 10.1093/nar/gkae105. 

Wang,Y., Ghosh,G. and Hendrickson,E.A. (2009) Ku86 represses lethal telomere delegon 
events  in  human  somagc  cells.  Proceedings  of  the  Na9onal  Academy  of  Sciences,  106, 
12430–12435. 

Bossaert,M.,  Moreno,A.T.,  Peixoto,A.,  Pillaire,M.-J.,  Chanut,P.,  Frit,P.,  Calsou,P., 
Loparo,J.J. and Brijon,S. (2024) Idengficagon of the main barriers to Ku accumulagon in 
chromagn. Cell Reports, 43. 

Kragelund,B.B., Weterings,E., Hartmann-Petersen,R. and Keijzers,G. (2016) The Ku70/80 
ring  in  Non-Homologous  End-Joining:  easy  to  slip  on,  hard  to  remove.  Front  Biosci 
(Landmark Ed), 21, 514–527. 

Weterings,E.,  Verkaik,N.S.,  Keijzers,G.,  Florea,B.I.,  Wang,S.-Y.,  Ortega,L.G.,  Uematsu,N., 
Chen,D.J. and van Gent,D.C. (2009) The Ku80 Carboxy Terminus Sgmulates Joining and 
Artemis-Mediated Processing of DNA Ends. MCB, 29, 1134–1142. 

Inagawa,T.,  Wennink,T.,  Lebbink,J.,  Keijzers,G.,  Florea,B.,  Verkaik,N.  and  van  Gent,D. 
(2020) C-Terminal Extensions of Ku70 and Ku80 Differengally Influence DNA End Binding 
Properges. INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES, 21. 

Singleton,B.K., Torres-Arzayus,M.I., Ro~nghaus,S.T., Taccioli,G.E. and Jeggo,P.A. (1999) 
The C Terminus of Ku80 Acgvates the DNA-Dependent Protein Kinase Catalygc Subunit. 
Mol. Cell. Biol., 19, 3267–3277. 

Sibanda,B.L., Critchlow,S.E., Begun,J., Pei,X.Y., Jackson,S.P., Blundell,T.L. and Pellegrini,L. 
(2001) Crystal structure of an Xrcc4–DNA ligase IV complex. Nat Struct Mol Biol, 8, 1015–
1019. 

Grawunder,U., Zimmer,D. and Lieber,M.R. (1998) DNA ligase IV binds to XRCC4 via a mogf 
located between rather than within its BRCT domains. Current Biology, 8, 873–879. 

10.  

Goff,N.J.,  Brenière,M.,  Buehl,C.J.,  de Melo,A.J.,  Huskova,H.,  Ochi,T.,  Blundell,T.L., 
Mao,W.,  Yu,K.,  Modesg,M.,  et  al.  (2022)  Catalygcally  inacgve  DNA  ligase  IV  promotes 
DNA repair in living cells. Nucleic Acids Research, 10.1093/nar/gkac913. 

11.   Meek,K.  (2020)  Acgvagon  of  DNA-PK  by  hairpinned  DNA  ends  reveals  a  stepwise 

mechanism of kinase acgvagon. Nucleic Acids Research, 48, 9098–9108. 

12.  

Roy,S.,  de  Melo,A.,  Xu,Y.,  Tadi,S.,  Negrel,A.,  Hendrickson,E.,  Modesg,M.  and  Meek,K. 
(2015) XRCC4/XLF Interacgon Is Variably Required for DNA Repair and Is Not Required for 
Ligase IV Sgmulagon. MOLECULAR AND CELLULAR BIOLOGY, 35, 3017–3028. 

 18 

 
13.  

Yu,Y.,  Wang,W.,  Ding,Q.,  Ye,R.,  Chen,D.,  Merkle,D.,  Schriemer,D.,  Meek,K.  and  Lees-
Miller,S. (2003) DNA-PK phosphorylagon sites in XRCC4 are not required for survival amer 
radiagon or for V(D)J recombinagon. DNA REPAIR, 2, 1239–1252. 

14.   Medina-Suarez,D.,  Han,L.,  O’Reily,S.,  Liu,J.,  Wei,C.,  Brenière,M.,  Goff,N.J.,  Chen,C., 
Modesg,M., Meek,K., et al. (2024) DNA ligases 3 and 4 cooperate in vivo to facilitate DNA 
repair and organism viability. In Review. 

15.   Masani,S., Han,L., Meek,K. and Yu,K. (2016) Redundant funcgon of DNA ligase 1 and 3 in 
alternagve  end-joining  during  immunoglobulin  class  switch  recombinagon.  PNAS,  113, 
1261–1266. 

16.   Oh,S., Harvey,A., Zimbric,J., Wang,Y., Nguyen,T., Jackson,P.J. and Hendrickson,E.A. (2014) 
DNA ligase III and DNA ligase IV carry out genegcally disgnct forms of end joining in human 
somagc cells. DNA Repair, 21, 97–110. 

17.  

18.  

19.  

20.  

21.  

22.  

23.  

Vu,D.-D., Bonucci,A., Brenière,M., Cisneros-Aguirre,M., Pelupessy,P., Wang,Z., Carlier,L., 
Bouvignies,G.,  Cortes,P.,  Aggarwal,A.K.,  et  al.  (2024)  Mulgvalent  interacgons  of  the 
disordered regions of XLF and XRCC4 foster robust cellular NHEJ and drive the formagon 
of  ligagon-boosgng  condensates  in  vitro.  Nat  Struct  Mol  Biol,  10.1038/s41594-024-
01339-x. 

Chen,S., Lee,L., Naila,T., Fishbain,S., Wang,A., Tomkinson,A.E., Lees-Miller,S.P. and He,Y. 
(2021) Structural basis of long-range to short-range synapgc transigon in NHEJ. Nature, 
10.1038/s41586-021-03458-7. 

Chaplin,A.K.,  Hardwick,S.W.,  Stavridi,A.K.,  Buehl,C.J.,  Goff,N.J.,  Ropars,V.,  Liang,S., 
Oliveira,T.M.D., Chirgadze,D.Y., Meek,K., et al. (2021) Cryo-EM of NHEJ supercomplexes 
provides insights into DNA repair. Molecular Cell, 0. 

Buehl,C.J.,  Goff,N.J.,  Hardwick,S.W.,  Gellert,M.,  Blundell,T.L.,  Yang,W.,  Chaplin,A.K.  and 
Meek,K. (2023) Two disgnct long-range synapgc complexes promote different aspects of 
end processing prior to repair of DNA breaks by non-homologous end joining. Molecular 
Cell, 10.1016/j.molcel.2023.01.012. 

Sgnson,B.M., Moreno,A.T., Walter,J.C. and Loparo,J.J. (2020) A Mechanism to Minimize 
Errors during Non-homologous End Joining. Molecular Cell, 77, 1080-1091.e8. 

Seif-El-Dahan,M.,  Kefala-Stavridi,A.,  Frit,P.,  Hardwick,S.W.,  Chirgadze,D.Y.,  Maia  De 
Oliviera,T.,  Brijon,S.,  Barboule,N.,  Bossaert,M.,  Pandurangan,A.P.,  et  al.  (2023)  PAXX 
binding  to  the  NHEJ  machinery  explains  funcgonal  redundancy  with  XLF.  Science 
Advances, 9, eadg2834. 

Tadi,S.,  Tellier-Lebegue,C.,  Nemoz,C.,  Drevet,P.,  Audebert,S.,  Roy,S.,  Meek,K., 
Charbonnier,J. and Modesg,M. (2016) PAXX Is an Accessory c-NHEJ Factor that Associates 
with Ku70 and Has Overlapping Funcgons with XLF. CELL REPORTS, 17, 541–555. 

 19 

 
24.  

25.  

26.  

27.  

28.  

Roy,S.,  Andres,S.,  Vergnes,A.,  Neal,J.,  Xu,Y.,  Yu,Y.,  Lees-Miller,S.,  Junop,M.,  Modesg,M. 
and Meek,K. (2012) XRCC4’s interacgon with XLF is required for coding (but not signal) 
end joining. NUCLEIC ACIDS RESEARCH, 40, 1684–1694. 

Balmus,G., Barros,A.C., Wijnhoven,P.W.G., Lescale,C., Hasse,H.L., Boroviak,K., Sage,C. le, 
Doe,B.,  Speak,A.O.,  Galli,A.,  et  al.  (2016)  Synthegc  lethality  between  PAXX  and  XLF  in 
mammalian development. Genes Dev., 30, 2152–2157. 

Cisneros-Aguirre,M., Lopezcolorado,F.W., Tsai,L.J., Bhargava,R. and Stark,J.M. (2022) The 
importance of DNAPKcs for blunt DNA end joining is magnified when XLF is weakened. 
Nat Commun, 13, 3662. 

Graham,T.G.W., Carney,S.M., Walter,J.C. and Loparo,J.J. (2018) A single XLF dimer bridges 
DNA ends during nonhomologous end joining. Nat Struct Mol Biol, 25, 877–884. 

Bhargava,R., Sandhu,M., Muk,S., Lee,G., Vaidehi,N. and Stark,J.M. (2018) C-NHEJ without 
indels is robust and requires synergisgc funcgon of disgnct XLF domains. Nat Commun, 9, 
2484. 

29.   Ochi,T. and Khan,H. (2023) Plant PAXX has an XLF-like funcgon and sgmulates DNA end-

joining by the Ku-DNA ligase IV-XRCC4 complex. 10.1101/2023.05.26.542399. 

30.  

Yu,Y.,  Gupta,S., 

(2005) 
Cui,X., 
Autophosphorylagon of DNA-Dependent Protein Kinase Regulates DNA End Processing 
and May Also Alter Double-Strand Break Repair Pathway Choice. MCB, 25, 10842–10852. 

Lees-Miller,S.P. 

and  Meek,K. 

Cho,Y.-M., 

31.   Neal,J.A.,  Sugiman-Marangos,S.,  VanderVere-Carozza,P.,  Wagner,M.,  Turchi,J.,  Lees-
Miller,S.P.,  Junop,M.S.  and  Meek,K.  (2014)  Unraveling  the  Complexiges  of  DNA-
Dependent  Protein  Kinase  Autophosphorylagon.  Molecular  and  Cellular  Biology,  34, 
2162–2175. 

32.  

Liu,L., Chen,X., Li,J., Wang,H., Buehl,C.J., Goff,N.J., Meek,K., Yang,W. and Gellert,M. (2021) 
Autophosphorylagon  transforms  DNA-PK  from  protecgng  to  processing  DNA  ends. 
Molecular Cell, 0. 

33.   Öz,R.,  Wang,J.L.,  Guerois,R.,  Goyal,G.,  KK,S.,  Ropars,V.,  Sharma,R.,  Koca,F., 
Charbonnier,J.-B.,  Modesg,M.,  et  al.  (2021)  Dynamics  of  Ku  and  bacterial  non-
homologous end-joining characterized using single DNA molecule analysis. Nucleic Acids 
Research, 49, 2629–2641. 

34. 

 Manoj,F.  and  Kuhlman,T.E.  (2023)  Phylogenegc  Distribugon  of  Prokaryogc  Non-
in  Bacteria  and  Archaea. 
Homologous  End 
10.1101/2023.09.30.560322. 

Joining  DNA  Repair  Systems 

35.  

Lees-Miller,J.P., Cobban,A., Katsonis,P., Bacolla,A., Tsutakawa,S.E., Hammel,M., Meek,K., 
Anderson,D.W.,  Lichtarge,O.,  Tainer,J.A.,  et  al.  (2021)  Uncovering  DNA-PKcs  ancient 

 20 

 
phylogeny, unique sequence mogfs and insights for human disease. Progress in Biophysics 
and Molecular Biology, 163, 87–108. 

Lakshmipathy,U. and Campbell,C. (1999) The Human DNA Ligase III Gene Encodes Nuclear 
and Mitochondrial Proteins. Molecular and Cellular Biology, 19, 3869. 

Ellenberger,T.  and  Tomkinson,A.E.  (2008)  Eukaryogc  DNA  Ligases:  Structural  and 
Funcgonal Insights. Annual Review of Biochemistry, 77, 313–338. 

Sallmyr,A.,  Khajri  Bhandari,S.,  Naila,T.  and  Tomkinson,A.E.  (2023)  Mammalian  DNA 
ligases;  roles 
integrity.  Journal  of  Molecular  Biology, 
10.1016/j.jmb.2023.168276. 

in  maintaining  genome 

Sallmyr,A.,  Rashid,I.,  Bhandari,S.K.,  Naila,T.  and  Tomkinson,A.E.  (2020)  Human  DNA 
ligases in replicagon and repair. DNA Repair, 93, 102908. 

Doherty,A.J. and Suh,S.W. (2000) Structural and mechanisgc conservagon in DNA ligases. 
Nucleic Acids Research, 28, 4051–4058. 

Shuman,S. (2009) DNA Ligases: Progress and Prospects*. Journal of Biological Chemistry, 
284, 17365–17369. 

Conlin,M.P., Reid,D.A., Small,G.W., Chang,H.H., Watanabe,G., Lieber,M.R., Ramsden,D.A. 
IV  Guides  End-Processing  Choice  during 
and  Rothenberg,E.  (2017)  DNA  Ligase 
Nonhomologous End Joining. Cell Reports, 20, 2810–2819. 

Chisgakov,D.A., Voronova,N.V. and Chisgakov,A.P. (2009) Ligase IV syndrome. European 
Journal of Medical Gene9cs, 52, 373–378. 

Han,L., Masani,S., Hsieh,C. and Yu,K. (2014) DNA Ligase I Is Not Essengal for Mammalian 
Cell Viability. Cell Reports, 7, 316–320. 

Goff,N.J.,  Mikhova,M.,  Schmidt,J.C.  and  Meek,K.  (2024)  DNA-PK:  A  synopsis  beyond 
synapsis. DNA Repair, 141, 103716. 

Graham,T.G.W., Walter,J.C. and Loparo,J.J. (2016) Two-Stage Synapsis of DNA Ends during 
Non-homologous End Joining. Molecular Cell, 61, 850–858. 

36.  

37.  

38.  

39.  

40.  

41.  

42.  

43.  

44.  

45.  

46.  

47.   Meek,K.,  Lees-Miller,S.  and  Modesg,M.  (2012)  N-terminal  constraint  acgvates  the 
catalygc  subunit  of  the  DNA-dependent  protein  kinase  in  the  absence  of  DNA  or  Ku. 
NUCLEIC ACIDS RESEARCH, 40, 2964–2973. 

48.  

Sgnson,B.M.,  Carney,S.M.,  Walter,J.C.  and  Loparo,J.J.  (2024)  Structural  role  for  DNA 
Ligase IV in promogng the fidelity of non-homologous end joining. Nat Commun, 15, 1250. 

49.   Waters,C.A.,  Strande,N.T.,  Pryor,J.M.,  Strom,C.N.,  Mieczkowski,P.,  Burkhalter,M.D., 
Oh,S., Qaqish,B.F., Moore,D.T., Hendrickson,E.A., et al. (2014) The fidelity of the ligagon 

 21 

 
50. 

51.  

52.  

53.  

54.  

55.  

56.  

57.  

58.  

59.  

step determines how ends are resolved during nonhomologous end joining. Nat Commun, 
5, 4286. 

Gu,J., Lu,H., Tippin,B., Shimazaki,N., Goodman,M.F. and Lieber,M.R. (2007) XRCC4:DNA 
ligase IV can ligate incompagble DNA ends and can ligate across gaps. The EMBO Journal, 
26, 1010–1023. 

Chen,X., Ballin,J.D., Della-Maria,J., Tsai,M.-S., White,E.J., Tomkinson,A.E. and Wilson,G.M. 
(2009) Disgnct kinegcs of human DNA ligases I, IIIα, IIIβ, and IV reveal direct DNA sensing 
ability and differengal physiological funcgons in DNA repair. DNA Repair, 8, 961–968. 

Amin,H., Zahid,S., Hall,C. and Chaplin,A.K. (2024) Cold snapshots of DNA repair: Cryo-EM 
structures  of  DNA-PKcs  and  NHEJ  machinery.  Progress  in  Biophysics  and  Molecular 
Biology, 186, 1–13. 

Vogt,A., He,Y. and Lees-Miller,S.P. (2023) How to fix DNA breaks: new insights into the 
mechanism  of  non-homologous  end 
joining.  Biochemical  Society  Transac9ons, 
10.1042/BST20220741. 

Chen,S., Vogt,A., Lee,L., Naila,T., McKeown,R., Tomkinson,A.E., Lees-Miller,S.P. and He,Y. 
(2023)  Cryo-EM  visualizagon  of  DNA-PKcs  structural  intermediates  in  NHEJ.  Science 
Advances, 9, eadg2838. 

Hardwick,S.W.,  Stavridi,A.K.,  Chirgadze,D.Y.,  De  Oliveira,T.M.,  Charbonnier,J.-B., 
Ropars,V., Meek,K., Blundell,T.L. and Chaplin,A.K. (2023) Cryo-EM structure of a DNA-PK 
trimer: higher order oligomerisagon in NHEJ. Structure, 31, 895-902.e3. 

DeFazio,L.G., Stansel,R.M., Griffith,J.D. and Chu,G. (2002) Synapsis of DNA ends by DNA-
dependent protein kinase. The EMBO journal, 21, 3192–3200. 

Hammel,M.,  Yu,Y.,  Mahaney,B.L.,  Cai,B.,  Ye,R.,  Phipps,B.M.,  Rambo,R.P.,  Hura,G.L., 
Pelikan,M.,  So,S.,  et  al.  (2010)  Ku  and  DNA-dependent  protein  kinase  dynamic 
conformagons and assembly regulate DNA binding and the inigal non-homologous end 
joining complex. The Journal of biological chemistry, 285, 1414–1423. 

Zhao,B., Watanabe,G., Morten,M.J., Reid,D.A., Rothenberg,E. and Lieber,M.R. (2019) The 
essengal  elements  for  the  noncovalent  associagon  of  two  DNA  ends  during  NHEJ 
synapsis. Nat Commun, 10, 3588. 

Reid,D.A., Conlin,M.P., Yin,Y., Chang,H.H., Watanabe,G., Lieber,M.R., Ramsden,D.A. and 
Rothenberg,E.  (2017)  Bridging  of  double-stranded  breaks  by  the  nonhomologous  end-
joining ligagon complex is modulated by DNA end chemistry. Nucleic Acids Research, 45, 
1872–1878. 

60.  

Reid,D.A., Keegan,S., Leo-Macias,A., Watanabe,G., Strande,N.T., Chang,H.H., Oksuz,B.A., 
Fenyo,D.,  Lieber,M.R.,  Ramsden,D.A.,  et  al.  (2015)  Organizagon  and  dynamics  of  the 

 22 

 
nonhomologous  end-joining  machinery  during  DNA  double-strand  break  repair. 
Proceedings of the Na9onal Academy of Sciences, 112, E2575–E2584. 

61.   Wang,J.L., Duboc,C., Wu,Q., Ochi,T., Liang,S., Tsutakawa,S.E., Lees-Miller,S.P., Nadal,M., 
Tainer,J.A., Blundell,T.L., et al. (2018) Dissecgon of DNA double-strand-break repair using 
novel single-molecule forceps. Nature structural & molecular biology, 25, 482–487. 

62.  

63.  

64.  

65.  

Dylgjeri,E.,  McNair,C.,  Goodwin,J.F.,  Raymon,H.K.,  McCue,P.A.,  Shafi,A.A.,  Leiby,B.E., 
Leeuw,R., Kothari,V., McCann,J.J., et al. (2019) Pleiotropic impact of DNA-PK in cancer and 
implicagons for therapeugc strategies. Clinical cancer research : an official journal of the 
American Associa9on for Cancer Research, 25, 5623–5637. 

Beucher,A.,  Birraux,J.,  Tchouandong,L.,  Barton,O.,  Shibata,A.,  Conrad,S.,  Goodarzi,A.A., 
Krempler,A.,  Jeggo,P.A.  and  Lobrich,M.  (2009)  ATM  and  Artemis  promote  homologous 
recombinagon of radiagon-induced DNA double-strand breaks in G2. The EMBO journal, 
28, 3413–3427. 

Rooney,S.,  Alt,F.W.,  Lombard,D.,  Whitlow,S.,  Eckersdorff,M.,  Fleming,J.,  Fugmann,S., 
Ferguson,D.O.,  Schatz,D.G.  and  Sekiguchi,J.  (2003)  Defecgve  DNA  repair  and  increased 
genomic  instability  in  Artemis-deficient  murine  cells.  The  Journal  of  experimental 
medicine, 197, 553–565. 

Peng,Y.,  Woods,R.G.,  Beamish,H.,  Ye,R.,  Lees-Miller,S.P.,  Lavin,M.F.  and  Bedford,J.S. 
(2005) Deficiency in the catalygc subunit of DNA-dependent protein kinase causes down-
regulagon of ATM. Cancer Res, 65, 1670–1677. 

66.   Neal,J. and Meek,K. (2019) Deciphering phenotypic variance in different models of DNA-

PKcs deficiency. DNA REPAIR, 73, 7–16. 

67.  

68.  

69.  

70.  

Sekiguchi,J.,  Ferguson,D.O.,  Chen,H.T.,  Yang,E.M.,  Earle,J.,  Frank,K.,  Whitlow,S.,  Gu,Y., 
Xu,Y.,  Nussenzweig,A.,  et  al.  (2001)  Genegc  interacgons  between  ATM  and  the 
nonhomologous end-joining factors in genomic stability and development. In Proceedings 
of the na9onal academy of sciences of the united states of america.Vol. 98, pp. 3243–
3248. 

Bertrand,C., Thibessard,A., Bruand,C., Lecointe,F. and Leblond,P. (2019) Bacterial NHEJ: a 
never ending story. Mol Microbiol, 111, 1139–1151. 

Deshpande,R.A.,  Myler,L.R.,  Soniat,M.M.,  Makharashvili,N.,  Lee,L.,  Lees-Miller,S.P., 
Finkelstein,I.J. and Paull,T.T. (2020) DNA-dependent protein kinase promotes DNA end 
processing by MRN and CtIP. Sci Adv, 6, 0922. 

Chaplin,A.K.,  Hardwick,S.W.,  Liang,S.,  Kefala  Stavridi,A.,  Hnizda,A.,  Cooper,L.R.,  De 
Oliveira,T.M., Chirgadze,D.Y. and Blundell,T.L. (2021) Dimers of DNA-PK create a stage for 
DNA double-strand break repair. Nat Struct Mol Biol, 28, 13–19. 

 23 

 
71.  

72.  

Chen,X., Xu,X., Chen,Y., Cheung,J.C., Wang,H., Jiang,J., de Val,N., Fox,T., Gellert,M. and 
Yang,W. (2021) Structure of an acgvated DNA-PK and its implicagons for NHEJ. Molecular 
Cell, 81, 801-810.e3. 

Balmus,G.,  Pilger,D.,  Coates,J.,  Demir,M.,  Sczaniecka-Clim,M.,  Barros,A.C.,  Woods,M., 
Fu,B.,  Yang,F.,  Chen,E.,  et  al.  (2019)  ATM  orchestrates  the  DNA-damage  response  to 
counter toxic non-homologous end-joining at broken replicagon forks. Nat Commun, 10, 
87. 

73.   Mikhova,M.,  Goff,N.J.,  Janovič,T.,  Heyza,J.R.,  Meek,K.  and  Schmidt,J.C.  (2024)  Single-
molecule  imaging  reveals  the  kinegcs  of  non-homologous  end-joining  in  living  cells. 
10.1101/2023.06.22.546088. 

74.  

75.  

76.  

Schaub,J.M., Soniat,M.M. and Finkelstein,I.J. (2022) Polymerase theta-helicase promotes 
end  joining  by  stripping  single-stranded  DNA-binding  proteins  and  bridging  DNA  ends. 
Nucleic Acids Research, 10.1093/nar/gkac119. 

Beagan,K. and McVey,M. (2016) Linking DNA polymerase theta structure and funcgon in 
health and disease. Cell. Mol. Life Sci., 73, 603–615. 

Li,C., Zhu,H., Jin,S., Maksoud,L.M., Jain,N., Sun,J. and Gao,Y. (2023) Structural basis of DNA 
polymerase θ mediated DNA end joining. Nucleic Acids Research, 51, 463–474. 

77.   Wyaj,D.W.,  Feng,W.,  Conlin,M.P.,  Yousefzadeh,M.J.,  Roberts,S.A.,  Mieczkowski,P., 
Wood,R.D.,  Gupta,G.P.  and  Ramsden,D.A.  (2016)  Essengal  Roles  for  Polymerase  theta-
Mediated End Joining in the Repair of Chromosome Breaks. Mol. Cell, 63, 662–673. 

78.  

He,P. and Yang,W. (2018) Template and primer requirements for DNA Pol θ-mediated end 
joining. Proceedings of the Na9onal Academy of Sciences, 115, 7747–7752. 

 24 

 
 
 
 
 
 
Chapter 2: Catalytically inactive DNA ligase IV promotes DNA repair in living cells. 

By: 

Noah J. Goff1,3,4, Manon Brenière5, Christopher J. Buehl1,3,4, Abinadabe J. de Melo5, Hana 
Huskova5, Takashi Ochi6, Tom Blundell7, Weifeng Mao2,3,8, Kefei Yu2,3, Mauro Modesg5,, and 
Katheryn Meek1,3,4*. 

Data chapter 

1College of Veterinary Medicine, 2College of Human Medicine, 3Department of Microbiology & Molecular Gene;cs, 4Department 
of Pathobiology & Diagnos;c Inves;ga;on, Michigan State University, East Lansing, MI 48824, USA; and 5Centre de Recherche en 
Cancérologie de Marseille, CNRS UMR7258, INSERM U1068, Ins;tut Paoli-CalmeSes, Aix-Marseille Université, Marseille, France; 
6The Astbury Centre for Structural Molecular Biology, School of Molecular and Cellular Biology, Faculty of Biological Sciences, 
University of Leeds, Leeds LS2 9TJ, UK; 7Department of Biochemistry , University of Cambridge, Cambridge CB2 1GA, UK; 8current 
address,  Department of Biotechnology, Dalian Medical University, Dalian 116044, China.*Correspondence:  kmeek@msu.edu 

 25 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ABSTRACT 

DNA double strand breaks (DSBs) are induced by external genotoxic agents (ionizing radiation or 

genotoxins)  or  by  internal  processes  (recombination  intermediates  in  lymphocytes  or  by 

replication errors). The DNA ends induced by these genotoxic processes are often not ligatable, 

requiring potentially mutagenic end-processing to render ends compatible for ligation by non-

homologous  end-joining  (NHEJ).  Using  single  molecule  approaches,  Loparo  and  colleagues 

propose  that  NHEJ  fidelity  can  be  maintained  by  restricting  end-processing  to  a  ligation 

competent short-range NHEJ complex that “maximizes the fidelity of DNA repair”. These in vitro 

studies  show  that  although  this  short-range  NHEJ  complex  requires  DNA  ligase  IV  (Lig4),  its 

catalytic activity is dispensable. Here using cellular models, we show that inactive Lig4 robustly 

promotes  DNA  repair  in  living  cells.  Compared  to  repair  products  from  wild-type  cells,  those 

isolated  from  cells  with  inactive  Lig4  show  a  somewhat  increased  fraction  that  utilize  micro-

homology (MH) at the joining site consistent with alternative end-joining (a-EJ). But unlike a-EJ in 

the absence of NHEJ, a large percentage of joints isolated from cells with inactive Lig4 occur with 

no MH -- thus, clearly distinct from a-EJ. Finally, biochemical assays demonstrate that the inactive 

Lig4 complex promotes the activity of DNA ligase III (Lig3).  

 26 

 
 
 
 
INTRODUCTION 

A  growing  consensus  is  emerging  that  DNA  double-strand  break  repair  by  the  non-

homologous  end-joining  [NHEJ]  pathway  proceeds  through  distinct  steps  (3-10).  Loparo  and 

colleagues have shown that  a  long-range synaptic  complex  that  positions the DNA  ends ~115 

Angstroms (Å) apart is dependent on the catalytic subunit of the DNA dependent protein kinase 

[DNA-PKcs, DNA-PK] and its regulatory subunit the DNA end-binding factor Ku. Transition of these 

long-range complexes to short-range synaptic complexes requires the catalytic activity of DNA-

PK and the NHEJ ligase complex including XRCC4, XLF, and DNA ligase IV [Lig4]. These in vitro 

studies establish that whereas the long-range complex effectively blocks DNA end-processing, 

the short-range complex facilitates both end-processing and ligation (6,11). These studies are in 

good agreement from cellular studies from Ramsden and colleagues who propose that the ligase 

complex helps to limit end-processing to promote more error-free repair (12). Of note, in these 

in vitro studies, the catalytic activity of Lig4 is not required to either promote formation of the 

short-range  complex  or  to  facilitate  end-processing  that  occurs  in  the  complex.  This  result  is 

consistent with previous studies proposing various structural roles for the Lig4 complex including 

promotion  of  end-processing  (13),  promotion  of  end-synapsis  (2,14,15),  and  notably  a  report 

from  Chiruvella  et  al  who  show  that  in  yeast,  a  catalytically  inactive  Lig4  complex  increases 

chromosomal  end-joining  (16).  Support  for  this  emerging  two-step  model  of  NHEJ  has  been 

bolstered  by  recent  cryo-EM  studies  that  precisely  delineate  potential  long-range  and  short-

range NHEJ complexes (7-9).  

In these previous studies, although end-processing mediated by the X family polymerases 

poll and polm, as well as TDP1, and PNKP was clearly dependent on the short-range complex, 

end-processing by the Artemis nuclease was not (11). Artemis functions exclusively with DNA-

PKcs; its role in facilitating opening of the hairpin DNA termini associated with VDJ recombination 

[the  process  that  provides  for  the  generation  of  a  diverse  repertoire  of  antibodies  and  T  cell 

receptors] has been well-defined (17-19). However, Artemis also functions to repair a subset of 

DSBs  with  non-ligatable DNA ends  (20).  Work from  our laboratory also  strongly suggests that 

Artemis hairpin opening is not restricted to the short-range complex (21). Briefly, using episomal 

end-joining assays, we found that cells which lack the NHEJ ligase complex robustly open hairpin 

 27 

 
 
termini.  These  data  are  consistent  with  experiments  in  developing  mouse  lymphocytes  from 

NHEJ deficient mice. In these studies, opened hairpin coding joints are observed in mice with 

defects  in  the  ligase  complex  (22),  but  not  in  mice  with  defects  in  either  DNA-PK  or  Artemis 

(23,24); but, coding end-joining is dramatically impaired in mice with any of these defects in NHEJ 

(23).  Finally,  these  results  are  completely  consistent  with  recent  structural  studies  that 

demonstrate that DNA-PK and Artemis function in a monomeric complex that does not require 

the Lig4 complex that is apparently requisite for other end-processing activities (25).  

It has been known for decades that NHEJ defective cell lines robustly repair DNA DSBs on 

episomes using the alternative non-homologous end-joining pathway (a-EJ)(26,27). However, our 

recent study demonstrates that cells lacking the XRCC4/Lig4 complex are similarly defective in 

joining the opened hairpins as are cells lacking either DNA-PKcs or Artemis where the hairpins 

remain  sealed  (21).  This  suggests  that  the  DNA-PK/Artemis  complex  that  facilitates  hairpin 

opening also, via some undefined mechanism, shields the opened hairpin ends from a-EJ; this 

“undefined  mechanism”  would  be  consistent  with  the  function  of  the  long-range  complex 

(blocking end-processing and ligation) proposed by Loparo and colleagues (11).   

Here we explore the potential basis for Lig4's non-catalytic role in NHEJ. To test whether 

the  presence  of  the  ligase  complex  (but  not  its  ligase  activity)  is  sufficient  to  promote  end-

processing  in  living  cells,  we  generated  (via  Crispr  strategies)  both  Lig4  deficient  and 

enzymatically  inactive  Lig4  mutant  cell  strains.  We  find  that  whereas  Lig4  deficient  cells  are 

similarly  impaired  in  joining  DSBs  with  hairpin  termini  as  are  other  NHEJ  defective  cells,  cells 

expressing inactive Lig4 are remarkably proficient in rejoining these DSBs. We conclude that the 

DNA-PK complex and perhaps the long-range synaptic complex protects DSBs from other DNA 

ligases  and  end-processing  factors  (except  for  Artemis)  in  living  cells.  Moreover,  our  cellular 

experiments demonstrate that a catalytically inactive Lig4 complex can efficiently promote end-

joining in living cells, consistent with the model that NHEJ-mediated end-processing is limited to 

a short-range synaptic complex. It follows that either Lig3 or Lig1 must be capable of facilitating 

ligation of DNA ends processed by NHEJ's short-range complexes. Thus, we used in vitro assays 

and establish that the catalytically inactive Lig4 complex robustly promotes the activity of Lig3 

but not Lig1. Finally, repaired DSBs recovered from cells expressing catalytically inactive Lig4 have 

 28 

 
 
 
increased  utilization  of  short  sequence  micro-homologies  (MH)  at  the  joining  site,  a  well-

established characteristic of a-EJ (28).   

RESULTS 

Figure 5: Cells expressing catalytically inactive DNA ligase IV or completely deficient in DNA 
ligase IV are markedly sensitive to the DSB-inducing agents calicheamicin and teniposide. (A) 
Western blot of Ligase IV in wild-type 293T cells (Lig4+/+), Lig4 haplo-insufficient cells (Lig4+/), or 
293T cells deficient in Lig4 (Lig4-/-), deficient in DNA-PKcs (DNA-PKcs-/-) or with mutations in Lig4 
[K273A/-, and K273A*/-]. (These are two independent clones possessing one copy of catalytically 
inactive Lig4 and a frameshift mutation on the second Lig4 allele; the K273A clone containing an 
additional mutation, I270V is labelled K273A*/-, for brevity.) (B) Sensitivity of the same panel of 
cell strains to calicheamicin (B) or teniposide (C) was assessed. Briefly, cells were plated in 24 
well plates into complete medium with increasing doses of calicheamicin or teniposide. After 7 
days, MTT was added, and cell survival assessed by colorimetry. Survivakl assays were performed 
four times in duplicate (B) or three times in duplicate (C). 

Cells expressing catalytically inactive DNA ligase IV or completely deficient in DNA ligase IV are 

markedly sensitive to DSB-inducing agents. To begin to address whether Lig4 has a non-catalytic 

function in living cells, a CRISPR strategy was devised using a single gRNA that targets the codon 

AAG, encoding K273 [the well-conserved catalytic site (16,29,30) in Lig4 and a single stranded 

oligonucleotide to direct an HDR mediated K273A mutation at this site (as well as the introduction 

of a novel HhaI site). From these transfections in 293T cells, single clones were isolated, and DNA 

was  isolated  for  PCR/HhaI  digestion.  Amplicon  sequencing  was  performed  on  clones  that 

included addition of the HhaI site in the initial PCR screen. Four clones were chosen for further 

analyses (Supplemental Table 1). Sanger sequencing from one clone (clone 17, that lacks a wild-

 29 

 
 
 
 
type allele by initial PCR screen) revealed two alleles with frameshift mutations, and we conclude 

that this clone is completely deficient in Lig4. Amplicon sequencing of two clones (that have a 

novel HhaI site ascertained by the initial PCR screen) reveal that one (clone 144) has the targeted 

K273A mutation on one allele, and a frameshift mutation on the second allele. The second clone 

(clone 195) contains the K273A mutation as well as an additional mutation, I270V on one allele, 

and  the  second  allele includes a  frameshift mutation (Supplemental  Table 1).  A heterozygous 

clone (clone 104) was isolated that has an inactivating frameshift mutation on one allele, but the 

second allele is wild-type. 

Cells with defects in the NHEJ pathway are markedly sensitive to numerous different DNA 

damaging  agents  that  can  induce  DNA  DSBs  by  various  mechanisms.  We  chose  two  drugs, 

calicheamicin  [generating  DSBs  with  3'  overhangs  with  3'  phosphoglycolate  (3'PG)]  (31)  that 

generates DSBs throughout the cell cycle (31) and teniposide a topoisomerase II (topII) inhibitor 

that generates two-ended DSBs primarily in S phase [with 5' overhangs with 5' hydroxyls] (32). 

Cellular  calicheamicin  and  teniposide  sensitivity  was  assessed  for  wild-type  293T  as  well  as 

Lig4+/-, K273A/-, K273A+I270V/- (labelled K273A*/-, for brevity), and Lig4-/-,  clones as well as a 

previously described DNA-PKcs deficient clone (33) using MTT staining as a measure of cellular 

viability (Figure 5B + C). As can be seen, cells completely deficient in Lig4 are remarkably sensitive 

to calicheamicin and teniposide; the two clones expressing only catalytically inactive Lig4 are also 

sensitive to both drugs, but substantially more resistant than cells that are completely deficient 

in Lig4. In contrast, at these doses, the Lig4+/- heterozygous clone is similarly resistant to both 

drugs as wild-type 293T cells. Finally, the previously described DNA-PKcs-/- 293T cells are also 

hypersensitive to calicheamicin and teniposide, but more resistant than Lig4-/- cells consistent 

with previous studies (23). These data suggest that catalytically inactive Lig4 promotes survival 

(albeit inefficient) to calicheamicin and teniposide.  

 30 

 
  
Figure 6: Catalytically inactive DNA ligase IV promotes joining of DSBs in episomal substrates 
in 293T cells.  (A) Fluorescent substrates (depicted in top panels indicate promoter with arrow) 
were utilized to detect TelN, ISce-1, or V(D)J coding and signal joints in 293T cells of the indicated 
genotypes.  293T  clones  with  the  indicated  genotypes  were  co-transfected  with  TelN  or  I-SceI 
plasmid substrate, with and without enzyme (-/+) and analyzed for red/green fluorescence via 
flow cytometry (lower panels). (B)  Lig4-/- 293T cells were tested for TelN and I-SceI joining by 
co-transfecting  the  TelN/I-SceI  substrates  and  expression  plasmids  with  or  without  co-
transfection of expression plasmids encoding either WT or catalytically inactive human ligase IV 
as indicated. (C) 293T clones were co-transfected with wild-type (W) or hypermutant (M) RAG 
expression  plasmids  with  either  plasmid  substrates  to  detect  coding  or  signal  end-joining  as 
indicated by analysis of red/blue fluorescence via flow cytometry. Episomal V(D)J assays testing 
joining  of  coding  (hairpin)  and  signal  (blunt)  ends.  Cells  were  transfected  with  substrate  and 
either: no Rag2 (-), WT rags (W), or mutant rag2 (M) and analyzed via flow cytometry. In A, B, and 
C, student's T test comparing joining rates between Lig4-/- and either Lig4+/+, Lig4+/-, or K273A 
were performed;  ****P<0.0001; ***P<0.001; ns=not significant in two-tailed unpaired t test. 

 31 

 
 
Catalytically inactive DNA ligase IV promotes joining of DSBs on episomal substrates in 293T 

cells.   We have previously  developed plasmid  substrates  with recognition  sequences for  DNA 

endonucleases that can produce a range of DNA end structures. The cutting sites flank the coding 

sequence  for  a  red  fluorescent  protein  (Crimson,  RFP)  driven  by  a  CMV  promoter  and  by  co-

transfecting  substrate  with  appropriate  enzymes,  end-joining  efficiency  can  be  assessed  by 

measuring the fraction of cells that express GFP or CFP after excision of the RFP cassette. The 

TelN protelomerase and I-SceI homing endonuclease produce hairpin DNA ends or 4 nucleotide 

overhangs respectively. To test whether 293T cells that either lack Lig4 entirely or retain a single 

copy  of  a  catalytically  inactive  Lig4  gene  can  rejoin  TelN  or  I-SceI-induced  DSBs,  a  series  of 

episomal end-joining assays were performed (Figure 6). Our previous work has shown that cells 

with defects in either the DNA-PK complex, Artemis, or the Lig4 complex are all significantly and 

similarly  impaired  in  joining  TelN-induced  hairpin  ends,  even  though  the  hairpinned  ends  are 

opened in cells deficient in the Lig4 complex (that retain DNA-PK and Artemis), but are obviously 

sealed in cells lacking DNA-PK or Artemis (21). Consistent with this previous study, cells deficient 

in Lig4 have a marked decrease in TelN joining (Figure 6A). In contrast, rejoining I-SceI DSBs is 

only  ~50%  reduced  in  cells  deficient  in  Lig4,  because  restriction  enzyme-induced  DSBs  are 

available to the a-EJ pathway [consistent with previous studies, (26,27,34)]. These data suggest 

that engagement of the Artemis/DNA-PK complex to resolve the closed hairpins must in some 

undefined way restrict the opened hairpins to the NHEJ pathway. Strikingly, in contrast to Lig4 

deficient  cells,  the  two  independent  clones  expressing  only  catalytically  inactive  Lig4  rejoin 

significant levels of both I-SceI and TelN-induced DSBs (Figure 6A). These data suggest that the 

Lig4 complex (whether active or not) promotes end-joining by a-EJ presumably by relaxing the 

NHEJ restriction imposed by hairpin opening by DNA-PK/Artemis. 

To confirm that the observed differences in end-joining are the result of the targeted Lig4 

mutations, we utilized the Lig4 deficient 293T cells and expression vectors encoding wild-type or 

catalytically inactive human Lig4 (Figure 6B) in a complementation experiment. As can be seen, 

both wild-type and catalytically inactive Lig4 (but not empty vector) substantially reverse both 

TelN and ISce-1 joining in 293T cells completely deficient in Lig4.  

 32 

 
 
In  cells  that  lack  NHEJ,  DSBs  (including  those  induced  by  restriction  enzymes,  those 

introduced during class switch recombination, or by CRISPR or other gene editing strategies) can 

all be efficiently re-joined by the a-EJ pathway. In contrast, NHEJ-deficient cells do not join VDJ-

associated  DSBs  because  VDJ  recombination  intermediates  are  tightly  restricted  to  the  NHEJ 

pathway by mechanism(s) that are still incompletely understood. To corroborate the TelN joining 

experiments, we next assessed V(D)J coding and signal end-joining in the same panel of 293T 

cells.  In  these  experiments we  used  both  wild-type  RAG  expression  vectors  as  well  as a  well-

studied  hyper-RAG2  mutant  that  substantially  increases  VDJ  joining  by  de-stabilizing  the  RAG 

post-cleavage complex(s) that function to limit end-joining to the NHEJ pathway (35). As can be 

seen, 293T cells lacking Lig4 are effectively incapable of joining RAG-induced DSBs, either blunt-

ended signal ends or hairpinned coding ends. In contrast, cells expressing catalytically inactive 

Lig4  join  both  coding  and  signal  ends  at  levels  only  modestly  reduced  as  compared  to  levels 

observed  in  wild-type  293T  or  Lig4  haplo-insufficient  293T  cells.  The  hyper  RAG2  mutant 

substantially increases the level of both coding and signal joining in wild-type cells, but joining is 

minimal in cells completely deficient in Lig4. In contrast, both signal and coding end-joining is 

robust, and only modestly reduced in cells expressing catalytically inactive Lig4 in experiments 

utilizing  the  hyper-RAG  mutant  (Figure  6C).  It  is  interesting  to  note  that  cells  expressing 

catalytically  inactive  Lig4 have  a  more  severe  deficit  in  signal  end-joining than  in  coding  end-

joining. To our knowledge, this is unlike VDJ deficits in any other NHEJ mutant studied to date, 

and perhaps suggest that catalytically inactive Lig4 can promote joining of over-hanged DNA ends 

better  than  blunt  DNA  ends.  In  sum,  we  conclude  that  catalytically  inactive  Lig4  efficiently 

promotes rejoining of RAG-induced DSBs. These data reveal that once V(D)J intermediates are 

released  to  a  complete  Lig4  complex  (potentially,  the  short-range  NHEJ  complex),  the  strict 

requirement for NHEJ-only dependent joining of RAG-induced DSBs has been fulfilled, and the 

RAG-induced DSBs can be joined by the a-EJ ligases just as well as non-RAG induced DSBs.  

 33 

 
 
Figure 7: Catalytically inactive DNA ligase IV promotes joining of DSBs in episomal substrates 
in a variety of cell types.  (A) Lig4-/- HCT116 cells were tested for TelN and I-SceI joining by co-
transfecting the TelN/I-SceI substrates and expression plasmids with or without co-transfection 
of expression plasmids encoding either WT or catalytically inactive human ligase IV as indicated. 
(B) Lig4-/- U20S cells were tested for TelN and I-SceI joining by co-transfecting the TelN/I-SceI 
substrates  and  expression  plasmids  with  or  without  co-transfection  of  expression  plasmids 
encoding either WT or catalytically inactive mouse ligase IV as indicated. In A and B, student's T 
test comparing joining rates between vector and Lig4 or K273A were performed; ****P<0.0001; 
***P<0.001; **P<0.01; ns=not significant in two-tailed unpaired t test.  

Catalytically inactive DNA ligase IV promotes joining of DSBs on episomal substrates in diverse 

cell types. To extend these studies, a CRISPR/Cas9 strategy was utilized to ablate Lig4 from the 

U2OS cell strain, and a Lig4 deficient HCT116 cell strain was obtained from Dr. Eric Hendrickson 

(36). As can be seen, in these cell types, both wild-type and catalytically inactive Lig4 (but not 

empty  vector)  substantially  reverse  both  TelN  and  ISce-I  end-joining  (Figure  7),  completely 

 34 

 
 
 
analogous to the results observed in Lig4-/- 293T cells. These data substantiate our conclusion 

that the Lig4 complex facilitates end-joining of a variety of DSBs on episomal substrates.   

Figure 8:  Catalytically inactive Ligase IV is only moderately deficient in CSR.  (A) Western blot 
of murine Ligase IV expressed in Lig4-/- CH12 cells. (B) CSR efficiency (percentage of IgA-positive 
cells,  assessed  by  flow  cytometry)  in  mouse  CH12  cells  expressing  WT  ligase  IV,  catalytically 
inactive Ligase IV (dLig4), or no ligase IV. Error bars indicate SE of three independent experiments. 
(C) Sensitivity of the same panel of cell strains to calicheamicin was assessed. Briefly, cells were 
plated in 24 well plates into complete medium with increasing doses of calicheamicin. After 7 
days, MTT was added, and cell survival assessed by colorimetry. In B, student's T test comparing 
joining rates between vector and Lig4, or vector K273S were performed; ****P<0.0001; in two-
tailed unpaired t test. Survival assay was performed three times in duplicate.  

Catalytically inactive DNA ligase IV supports chromosomal rearrangement during class switch 

recombination. Previously we studied the role of Lig4 in the process of immunoglobulin class 

switch recombination (CSR) using a well-studied mouse B cell model, CH12 cells (37). Retroviral 

vectors encoding either wild-type or K273S murine Lig4 were prepared and used to transduce a 

previously described Lig4-/- Ch12 cell strain. When CH12 cells are stimulated with an anti-CD40 

antibody,  interleukin  4,  and  TGFβ1  (+CIT),  CSR  (from  IgM  to  IgA)  is  robustly  induced.  This 

produces clear populations of IgM+/IgA-, IgM+/IgA+, and IgM-/IgA+ cells that can be analyzed by 

flow cytometry. Like VDJ recombination, CSR is a lymphocyte specific DNA recombination event 

that proceeds through a double-strand break intermediate; but unlike VDJ recombination, a-EJ 

can facilitate CSR (albeit at reduced levels) in the absence of most NHEJ factors. Our previous 

study demonstrated that either Lig1 or Lig3 could facilitate a-EJ in a Lig4 deficient murine B cell 

line (CH12 cells) that were induced to undergo class switch recombination.  

 35 

 
 
 
 
We next assessed CSR in the CH12 Lig4-/- cells expressing wild-type Lig4, K273S, or vector 

only from lentivirus expression vectors. Consistent with our previous study, only ~10% of Lig4 

deficient CH12 cells expressing vector alone undergo CSR compared to ~60% of cells expressing 

wild-type murine Lig4. Cells expressing K273S Lig4 display an intermediate level of CSR (Figure 8). 

We  conclude  that  catalytically  inactive  Lig4  promotes  chromosomal  end-joining  of  cell-

programmed DSBs. We also assessed cellular resistance to calicheamicin in this panel of CH12 

cells. As can be seen, whereas cells expressing wild-type Lig4 are substantially more resistant to 

calicheamicin than cells expressing no Lig4, cells expressing catalytically inactive Lig4 are similarly 

sensitive to calicheamicin as are cells lacking Lig4. We conclude that catalytically inactive Lig4 

cannot restore cellular resistance to DNA damaging agents in all cell types.  

Figure  9:  CatalyZcally  inacZve  Lig4/XRCC4/XLF  promotes  Lig3-mediated  joining  of  both 
cohesive  and  blunt  DNA  ends  in  vitro.    The  capacity  for  catalygcally  inacgve  Lig4/XRCC4  to 
promote ligagon by Lig3 (A), or Lig1 (B) was assessed by incubagng a blunt DNA ligagon substrate 
(linearized  pUC19)  with  purified  recombinant  proteins  as 
indicated  at  the  following 
concentragons:  Ku,  250nM;  XLF,  200nM;  LIG4/XRCC4,  500nM;  LIG3/XRCC1,  40nM;  and 
LIG1,180nM. 10ul reacgons were incubated for 30 minutes at room temperature and analyzed by 
gel  electrophoresis  and  quangfied  with  Image  J.  Bar  graph  represents  three  independent 
experiments and error bars represent SD.  

 36 

 
 
 
Catalytically  inactive  Lig4/XRCC4/XLF  promotes  Lig3-mediated  joining  of  both  cohesive  and 

blunt  DNA  ends  in  vitro.  The  robust  joining  facilitated  by  catalytically  inactive  Lig4  in  cellular 

assays suggests that  the  NHEJ short-range complex might be  capable  of facilitating ligation of 

DNA  ends  by  either  Lig3  or  Lig1.  To  directly  test  this  possibility,  Lig1  and  Lig3/XRCC1  were 

expressed and purified from E. coli (Sup. Figure 1A) and tested for activity (Sup. Figure 1B). Then 

purified  Lig1  or  Lig3/XRCC1  were tested  in  ligase assays  that  included the  components of the 

NHEJ short range complex (Ku, X4/Lig4, and XLF) using catalytically inactive Lig4, on substrates 

with either blunt or cohesive DNA ends. In these assays, higher levels Lig4/XRCC4 were used as 

compared to either Lig1 (2.8 molar excess) or Lig3/XRCC1 (12.5 molar excess). As can be seen 

(Figure 9A), using a blunt ligation substrate, no ligation products are observed with just the NHEJ 

components, and no activity is observed with only Lig3/XRCC1. In contrast, when Lig3/XRCC1 is 

added to reactions containing catalytically inactive Lig4 (with or without Ku), robust ligation is 

observed.  This  activity  is  completely  dependent  on  both  Lig4/XRCC4  and  XLF,  but  Ku  is 

dispensable. In parallel assays, Lig1 activity is not significantly enhanced by catalytically inactive 

Lig4 complex (Figure 9B) under any of the conditions tested.   

This assay was repeated using DNA  with cohesive  ends (Sup. Figure2); as can  be seen, 

ligation of compatible ends is robust in the presence of catalytically inactive Lig4 complex and 

Lig3/XRCC1,  but  not  Lig1,  suggesting  that  catalytically  inactive  Lig4  complex  can  stimulate 

Lig3/XRCC1-mediated  joining  of  cohesive  ends.  However,  (to  our  surprise)  we  consistently 

observed very low levels of ligation products in reactions with XLF and XRCC4/Lig4 (catalytically 

inactive, K273A) without Lig3/XRCC1 or Lig1. The activity represents <5% the activity observed in 

assays with wild-type Lig4 complexes (compare Sup. Figure 1C and 2A). This residual activity could 

not be attributed to a contaminating ligase since the proteins are prepared in E. coli that has only 

NAD dependent ligase. Moreover, adenylation assays demonstrate minimal adenylation of the 

K273A mutant Lig4 (Sup. Figure 3). From the recent cryo-EM study of the short-range complex 

(7), we observed that 4 additional lysine residues are very close to the 5’ phosphate of the DNA 

end; whereas K273 is ~7Å away, lysines 449, 451, 352, and 345 are between 8Å and 9.5Å away 

from the 5’ phosphate. We considered that one of these four lysines might serve as a back-up 

adenylation site explaining the minimal ligase activity of the K273A mutant. These four residues 

 37 

 
were  substituted  with  arginine  in  the  K273A  mutant  construct.  As  can  be  seen,  this  5X  lysine 

mutant has no residual activity in ligase assays but retains the capacity to stimulate Lig3 activity 

towards both blunt and cohesive ends in vitro (Supplemental Figure 4A) to a similar extent as the 

K273A  mutant.  Moreover,  in  cellular  assays,  the  5X  lysine  mutant  joining  activity  that  is 

indistinguishable from that of the K273A mutant (Supplemental Figure 4B). From these data, we 

conclude that the catalytically inactive Lig4/XRCC4/XLF complex promotes end-joining of Lig3 in 

vitro,  potentially  explaining  the  robust  joining  activity  observed  in  living  cells  expressing  the 

catalytically 

inactive 

complex.  

Figure 10: DSBs repaired in cells expressing K273A Lig4 have characteristics of a-EJ. (A) 293T 
clones of the indicated genotypes were transfected with Fill-in joining substrate either uncleaved, 
or  cleaved  with  Eco53KI  and  PspOMI  (to  generate  blunt  and  5'  overhangs),  or  Eco53K1  and 
Apa1(to generate blunt and 3' overhangs). Cells were assessed for RFP and GFP expression 72- 

 38 

 
 
Figure 10 (cont’d) 
hours later. Student's T test comparing joining rates between Lig4-/- and either Lig4+/+, Lig4+/-, 
or K273A*/- were performed; ****P<0.0001; in two-tailed unpaired t test. (B)  293T clones of the 
indicated  genotypes  were  co-transfected  with  the  alt-VDJ  substrate,  a  ds-RED  expression 
plasmid,  and  hypermutant  RAG  expression  plasmids.  Restoration  of  the  GFP  reading  frame 
requires joining via a 9bp region of MH that occurs 10bp from the termini of both coding ends. 
Cells  were  assessed  for  RFP  and  GFP  expression  72  hours  after  transfection.  Student's  T  test 
comparing  joining  rates  between  Lig4-/-  and  either  Lig4+/+,  Lig4+/-,  or  K273A*/-    as  well  as 
betwee  Lig4+/+  and  either  Lig4+/-  or  K273A*/-  were  performed;  ****P<0.0001;  ***P<0.001; 
**P<0.01; in two-tailed unpaired t test.  

Cells expressing K273A Lig4 cannot perform fill-in end-processing, but robustly promote a-EJ in 

episomal assays. Clearly catalytically inactive Lig4 enhances end-joining of episomal substrates 

and  in  some  cell  types,  cellular  survival  after  exposure  to  agents  that  induce  DSBs.  To  gain 

knowledge as to how end-joining is enhanced by the catalytically inactive Lig4 complex, we used 

assays that clarify structural characteristics of repaired DSBs. More specifically, we focused on 

determining whether catalytically inactive Lig4 promotes end-joining that is similar to authentic 

NHEJ [i.e. generating joints ranging from perfect end-joining with no or minimal base pair loss, or 

joining at sites of MH (1-3bp)], or alternatively, if catalytically inactive Lig4 promotes joining that 

is  more  similar  to  a-EJ  [characterized  by  increased  loss  of  terminal  nucleotides,  and  a  strong 

dependence on the presence of MH (often longer than 3bp)]. 

Although NHEJ is generally characterized as an error-prone repair mechanism, numerous 

studies  have  documented  the  fidelity  of  NHEJ  when  joining  most  DSBs  (11,12).  For  example, 

perfect  joining  of  compatible  ends  are  highly  favored  and  incompatible  ends  are  rejoined  to 

minimize nucleotide loss (11,12). To assess the capacity of catalytically inactive Lig4 to promote 

joining that is similar to that of NHEJ, a joining substrate that measures fidelity of joining was 

generated (Figure 10A). Briefly, the substrate plasmid is restricted with two enzymes to generate 

a blunt end on one side and an over-hanged end (either 5' or 3') on the other side as illustrated. 

To restore the crimson open reading frame, the ends must be aligned so that the missing bases 

across from the overhangs are filled in. In both cellular and in vitro models of NHEJ, perfect fill-in 

of DNA ends like these is efficiently mediated by NHEJ (11,12). If uncut plasmid is transfected, 

crimson  is  not  expressed  because  of  the  disruption  in  the  open  reading  frame,  but  GFP  is 

expressed  by  use  of  its  own  ATG.  When  restricted  plasmids  are  transfected,  if  any  plasmid 

 39 

 
 
 
rejoining occurs, GFP is expressed. With this  assay, "Perfect rejoining" represents the percent 

GFP positive cells that also express crimson, and in these assays, GFP expression is robust in all 

cell  types.  As  would  be  expected,  cells  completely  lacking  Lig4-/-  are  completely  unable  to 

perfectly rejoin the transfected substrate promoting crimson expression, whereas in wild-type 

293T cells more than 30% of the cells that rejoin the plasmid, rejoin the plasmid perfectly and 

crimson is robustly expressed (Figure 10A). In contrast to other joining assays (Figure 6), cells 

expressing only catalytically inactive Lig4 are similarly deficient in perfect rejoining as are cells 

completely  lacking  Lig4.  We  conclude  that  catalytically  inactive  Lig4  cannot  promote  perfect 

rejoining of the fill-in episomal substrate. It follows that the catalytically inactive complex either 

does  not  efficiently  support  fill-in  end-processing,  or  that  the  cellular  ligase  facilitating  end-

joining (Lig3 or Lig1) does not efficiently join the filled-in ends.  

To assess whether a-EJ-like events facilitate repair in cells expressing catalytically inactive 

Lig4, we utilized an assay developed by Roth and colleagues (38). It is well-appreciated that short 

sequence homologies at opened coding-end termini can facilitate coding end-joining (39); these 

short  sequence  homologies  are  generally  small  (1-3  base  pair)  and  occur  close  to  the  DNA 

terminus. We utilized their a-EJ assay (Figure 10B, termed alt-VDJ, using mutant RAG expression 

constructs  that  destabilize  the  RAG  post-cleavage  complex(s)  to  promote  more  alt-VDJ)  to 

determine  whether  cells  expressing  catalytically  inactive  Lig4  depend  on  MH  for  joining.  This 

assay requires nucleotide deletions of 10bp from each coding end, and then use of 9 base pair of 

short sequence homology to restore the GFP open-reading frame. Whereas minimal alt-VDJ is 

observed in wild-type 293T cells or Lig4 haplo-insufficient cells where end-processing generally 

precludes loss of the required 10bp from each coding-end, alt-VDJ is readily detected in Lig4-/- 

293T cells (Figure 10B). Strikingly, alt-VDJ joining is robust in cells expressing catalytically inactive 

Lig4; these data suggest that K273A mutant Lig4 promotes end-joining that is similar to a-EJ. 

 40 

 
 
Figure 11: Rejoined episomal VDJ coding and signal joints in cells expressing K273A Lig4 have 
characteristics of  a-EJ mediated  joining. (A) Coding  (left)  and  signal  joints  (middle)  were PCR 
amplified  from  substrates  isolated  from  cells  of  the  indicated  genotypes  72  hours  after 
transfection. (*) indicates a PCR artifact. Half of the PCR reaction for signal joint amplification was 
digested with ApaLI prior to electrophoresis (right). (B) Isolated coding joints from the indicated 
cell strains were subjected to amplicon sequencing. Histograms depicting numbers of nucleotides 
deleted or extent of microhomology at site of joining. 

 41 

 
 
Rejoined episomal DSBs in cells expressing K273A Lig4 have characteristics of a-EJ mediated 

joining. To further characterize DSB joining events in cells expressing catalytically inactive Lig4, 

we  characterized  VDJ  joints  isolated  from  cells  expressing  the  inactive  complex.  In  cells  with 

intact NHEJ, VDJ coding end-joining generates a diverse array of rejoined opened hairpin coding 

ends.  In  contrast,  rejoining  of  the  blunt  signal  ends  is  usually  a  precise,  perfect  head-to-head 

joining  of  the  two  heptamers.  This  recapitulates  VDJ  joining  characteristics  in  developing 

lymphocytes.  It  has  been  shown  that  the  rare  VDJ  joints  generated  in  NHEJ  deficient  cells  or 

animals have characteristics of a-EJ, such that coding joints have increased levels of nucleotide 

loss and the joints often occur at regions of MH; whereas signal ends are not perfectly rejoined. 

As shown above (Figure 6C), in cells completely deficient in Lig4, the joining rate of both coding 

and  signal  ends  is  severely  reduced.  Coding  and  signal  joints  from  episomal  assays  were  PCR 

amplified and analyzed by gel electrophoresis. As can be seen (Figure 11A, left), coding joints 

amplified  from  wild-type  or  Lig4  haplo-insufficient  293T  cells,  generate  a  diverse  “smear”  of 

coding  joints,  whereas  coding  joints  amplified  from  293T  cells  that  express  only  catalytically 

inactive Lig4 are homogeneous, and slightly smaller than those recovered from wild-type cells.  

As expected, signal joints from wild-type or Lig4 haplo-insufficient 293T cells are uniform, 

but signal ends isolated from cells that express catalytically inactive Lig4 are diverse, generating 

a smear of joints (Figure 11A middle). Rejoining of signal ends (without nucleotide loss or gain) 

generates a novel ApaLI site; so fidelity of signal end-joining can be ascertained by restricting 

signal joints with this enzyme. Signal joints amplified from assays in wild-type cells are uniform 

and largely susceptible to ApaLI cleavage consistent with perfect rejoining of signal joints in NHEJ-

proficient cells (40). In contrast, signal joints amplified from cells that lack Lig4 or express only 

catalytically inactive Lig4 are completely resistant to ApaLI cleavage indicating nucleotide loss or 

addition from the signal ends prior to rejoining, suggesting catalytic inactive Lig4 complex is not 

proficient at promoting perfect blunt end-joining (Figure 11A, right). 

PCR  amplified  coding  joints  were  submitted  for  amplicon  sequencing;  as  can  be  seen 

(Figure  11B),  joints  isolated  from  wild-type  or  Lig4  haplo-insufficient  cells  are  virtually 

indistinguishable, and ~50% do not have sequence MH at the site of joining. In these assays, there 

is  a  bias  for  joining  at  particular  sites  of  microhomology  resulting  over-representation  of 

 42 

 
sequences with 3bp  and 5bp deletions; these over-represented joints can be appreciated in all 

three samples. In wild-type cells, the predominate 3bp and 5bp deletion products account for 

5.99% and 12.12% of all sequences, in L4 haplo-insufficient cells, 9.28% and 14.66%, but in K273A 

mutant cells these two particular joints account for 27.4% and 18.37% of all joints. These data 

support the conclusion that joints facilitated by catalytically inactive Lig4 have characteristics of 

a-EJ.   

Figure 12: Chromosomal DSBs repaired in cells expressing K273A Lig4 have characteristics of a-
EJ.  (A) Diagram of region on chromosome nine targeted by two different gRNAs and position of 
primers  utilized  to  detect  chromosomal  deletions  induced  by  Cas9  and  indicated  gRNAs.  (B) 
Summary of amplicon analyses of PCR amplified deletional joints from indicated cell strains.  (C) 
Histograms depicting numbers of nucleotides deleted, inserted, or extent of microhomology at 
site of joining.  

 43 

 
 
 
Chromosomal end-joining in cells expressing K273A Lig4 has characteristics of a-EJ. To extend 

these findings to chromosomal end-joining, we exploited CRISPR/Cas9 targeting of the FANCG 

gene on chromosome 9, which we found to be remarkably efficient in 293T cells (41). Although 

this strategy does not allow quantification of joining, the quality of joining can be assessed by 

sequencing. Briefly, cells of the indicated genotypes were transfected with two gRNA/cas9/puro 

plasmids that target sequences ~300bp apart; after 48 hours, cells were subjected to puromycin 

selection. After 72 hours, genomic DNA was isolated and PCR was utilized to assess deletional 

rejoining of the two DSBs. Isolated PCR fragments were submitted for amplicon sequencing. In 

wild-type cells, ~92% of recovered joints are perfect (of the blunt-ended Cas9 cleavage sites), 

whereas in K273A mutant cells, only ~80% of the joints are perfect. In wild-type cells, the average 

nucleotide  loss/joint  is  0.45bp,  and  ~4%  of  joints  occur  at  sites  of  MH.  In  contrast,  in  K273A 

mutant cells, the average loss/joint was 2.9bp, and ~19% occur at sites with MH. The range of 

nucleotide loss, nucleotide insertion, and utilization of short sequence homologies for joints from 

wild-type or  K273A mutant cells is  shown  in  Figure  12C. We conclude  that  joints  from K273A 

mutant cells are consistent with joining via a-EJ.   

DISCUSSION 

In other mutational studies that block enzymatic function of NHEJ factors (ie Artemis or 

DNA-PKcs), inactivation results in a cellular phenotype that is similar to (42-44) or in certain cell 

types worse than complete loss (33,45). Thus, it appears that the Lig4 complex is unique in that 

in its inactive form, Lig4 can promote the function of other repair factors.  

There have been several previous studies proposing various structural roles for the Lig4 

complex.  In  2007,  Chu  and  colleagues  reported  that  the  NHEJ  ligase  complex  promotes  end-

processing activities in vitro (13). This is consistent with more recent reports, where in an in vitro 

model of NHEJ, Stinson et al. demonstrated that Lig4 (either active or inactive) was necessary for 

formation  of  a  short-range  non-homologous  end-joining  complex  that  closely  juxtaposed  two 

DNA ends, promoting both end-processing and ligation (11). 

Besides the studies from Loparo and colleagues suggesting a two-stage model of NHEJ 

(6,11), several other reports support a role in DNA end synapsis for the Lig4 complex (2,14,15). 

Reid et al demonstrated that the Lig4 complex can promote end-synapsis in vitro (using purified 

 44 

 
proteins in a FRET assay), and the ability to synapse the ends was strongly impacted by DNA end 

structure,  especially  by  the  presence  of  the  5’  phosphate  (2).  Of  note,  a  K273A  mutant  Lig4 

complex was defective in synapsing DNA ends in their assays using purified proteins (2), whereas 

similar FRET assays from Graham et al showed that K273A using Xenopus extracts, fully support 

end synapsis (6). Conlin et al, extended the studies of end synapsis to show that a unique pocket 

in Lig4 allows synapsis of DNA ends with mis-matched termini (15). There are also cellular studies 

that support a role for the Lig4 complex in DNA end synapsis. Cottarel et al. demonstrated that 

cells  deficient  in  Lig4  are  defective  in  damage-induced  autophosphorylation  of  DNA-PKcs  at 

serine 2056 (S2056) (14). We first showed that S2056 autophosphorylation could occur in trans, 

and that S2056 phosphorylation occurs almost exclusively by autophosphorylation (46). Recent 

structural  studies strongly  support  the  conclusion  that  S2056 phosphorylation  occurs  in  trans 

(25). Cottarel et al. demonstrated that catalytically inactive Lig4 could restore the deficit in S2056 

phosphorylation in Lig4 deficient cells as well as wild-type Lig4. From these studies they proposed 

that 

the  catalytically 

inactive  Lig4  complex 

facilitates  synapsis,  promoting 

trans-

autophosphorylation of DNA-PKcs at S2056 (14). We extended this finding by demonstrating that 

XLF also participates in the capacity of the Lig4 complex to promote S2056 phosphorylation in 

living cells, and that only XLF that was proficient in interacting with XRCC4 (to generate DNA-

bridging filaments) could promote S2056 phosphorylation (47). Finally, Chiruvella et al reported 

that  a  catalytically  inactive  Lig4  complex  increases  not  only  synapsis,  but  chromosomal  end-

joining in yeast; this impact on end-joining was absolutely dependent on XLF (16). This result is 

entirely consistent with cellular assays presented here, and with the in vitro assays demonstrating 

that catalytically inactive Lig4/XRCC4 only stimulates Lig3/XRCC1 joining in the presence of XLF 

(Figure 9).    

Our  working  model  is  that  Lig4  itself  is  important  in  promoting  synapsis  and  that 

formation  of  a  short-range  synaptic  complex  (even  one  that  lacks  Lig4  activity):  1)  facilitates 

joining of many DSBs, 2) promotes cellular resistance to drugs that induce DSBs in some cell types, 

and 3) fulfills the RAG-induced restriction step of VDJ recombination so that even RAG-induced 

DSBs (both coding and signal ends) can be joined by non-NHEJ DNA ligases.   

 45 

 
In sum, data presented here extend these previous studies demonstrating a non-catalytic 

role for Lig4 that robustly stimulates both end-joining and cellular resistance to DNA damage. 

Joining that is facilitated by the inactive Lig4 complex shares characteristics with a-EJ (increased 

nucleotide  loss,  and  increased  use  of  microhomologies).  This  observation  presents  an 

unanticipated  possibility,  that  the  inactive  Lig4  complex  (likely  consistent  with  short-range 

complexes  proposed  recently)  may  facilitate  a-EJ  mediated  by  Lig3.  These  studies  present  an 

important  unanswered  question:  Do  complexes  with  catalytically  active  Lig4  utilize  Lig3  to 

facilitate joining in normal cells? Work is ongoing to address this possibility.   

MATERIALS AND METHODS 

Cell culture, genome editing, and survival assays. 293T, U2OS and HCT116 cells were cultured in 

Dulbecco’s  Modified  Eagle  Medium  (Life  Technologies)  supplemented  with  10%  fetal  bovine 

serum (Atlanta Biologicals, GA), 2 mM L-glutamine, 0.1  mM  non-essential amino acids,  1 mM 

sodium pyruvate, 100 U/ml penicillin, 100 µg/ml streptomycin (Life Technologies) and 10 µg/ml 

ciprofloxacin.  CH12F3  cells  were  cultured  in  RPMI  1640  medium  supplemented  with  10% 

(vol/vol) FBS and 50 μM β-mercaptoethanol. 

Cas9-targeted gene disruption was performed using methods similar to those reported 

by Mali et al. (48). Briefly, duplex oligonucleotides (Integrated DNA Technologies) generating a 

gRNA specific for a PAM site targeting the K273 codon were cloned into pCas2A-puro. Cells were 

transfected with 2ug plasmid and a 120 nucleotide oligonucleotide encoding a K273A mutation 

as well as a silent mutation generating a novel restriction site (HhaI) (see Supplemental Table I). 

48  hours  after  transfection,  cells  were  replated  at  cloning  densities  in  media  containing 

puromycin (1ug/ml). Puromycin was removed after 72 hours. Isolated clones were selected, and 

DNA isolated with DNAzol (Sigma) according to the manufacturer’s protocol. Restriction digestion 

of PCR products was utilized to detect the K273A mutation. Western blotting was used to confirm 

expression  and  genotypes  were  confirmed  by  Sanger  sequencing  or  Amplicon  sequencing 

(Genewiz) as depicted in Supplemental Table 1.  

MTT staining was performed to assess cell viability for both 293T cells and CH12. 30,000 

to  50,000  cells  were  plated  in  each  well  of  a  24-well  plate,  containing  medium  with  varying 

concentrations of zeocin. After 5 to 7 days of calicheamicin treatment, cells were treated with 1 

 46 

 
 
 
mg/ml MTT (Sigma) solution for 1 hr. Medium containing MTT was then removed and formazan 

crystals thus produced were solubilized in acidic isopropanol. Absorbance was read at 570 nm to 

determine relative survival. 

Episomal  end-joining  assays.  The  fluorescent  VDJ  coding,  signal,  and  alt-VDJ  substrates  have 

been described (21,38,49). The Fill-in substrate was generated by synthesizing (Integrated DNA 

Technologies) a fragment encoding crimson with the addition of restriction endonuclease sites 

Eco53KI, PspOMI, and Apa1 that disrupt the Crimson open-reading frame. Cleavage with these 

restriction  enzymes  generate  blunt  and  over-hanged  ends  as  described  in  Figure  11D.  The 

Crimson fragment included NheI and BamHI which were used to subclone this fragment between 

the promoter and GFP open reading frame in the substrate plasmids. Briefly, extrachromosomal 

fluorescent  joining  assays  were  performed  on  cells  plated  at  20-40%  confluency  into  24-well 

plates in complete medium. Cells were transfected with 0.125 µg substrate, and either 0.25 I-

Sce1,  TelN,  or  RAG  1+2  expression  plasmids  per  well  using  polyethylenimine  (PEI,  1  ug/mL, 

Polysciences) at 2 µL/1 µg DNA. Cells were harvested 72 hours after transfection and analyzed 

for GFP and RFP expression by flow cytometry. The percentage of recombination was calculated 

as the percentage of live cells expressing GFP divided by the percentage expressing RFP. Data 

presented  represents  at  least  three  independent  experiments,  which  each  includes  triplicate 

transfections.  

In figure 11, the coding joint substrate was modified so that the coding flanks included 

only A or T sequences; in addition, the 23RSS was inverted to provide for inversional joining which 

facilitates  PCR  amplification  of  coding  joints.  Transfected  plasmids  were  isolated  by  alkaline 

lysates 72 hours after transfection. Coding joints from each transfection were PCR amplified and 

analyzed by electrophoresis and amplicon sequencing by Genewiz.   

Class  switch  recombination  assays.  CSR  assays  were  performed  as  described  previously  (37). 

Briefly,  cells  were  seeded  in  the  presence  of  1  μg/mL  anti-CD40  antibody  (16-0402-86; 

bioscience), 5 ng/mL of IL-4 (404-ML; R&D Systems), and 0.5 ng/mL TGF-β1 (R&D Systems 240-B) 

and grown for 72 h. Cells were stained with a FITC-conjugated anti-mouse IgA antibody (BD Bio- 

sciences  559354)  and  analyzed  on  a  LSR  II  flow  cytometer  (BD  Biosciences).  CSR  efficiency  is 

determined as the percentage of IgA-positive cells.  

 47 

 
 
Amplicon sequencing of Crispr/Cas9-induced chromosomal DSBs. We have used a Crispr/Cas9 

strategy  to  sequence  DSBs  from  within  the  FancG  locus  previously  (41).  To  induce  deletional 

DSBs,  two  gRNA/Cas9/puro  plasmids  were  transfected  into  293T  cells  with  the  indicated 

genotypes. After 48 hours, cells were place in puromycin selection media. Cells were harvested 

and  DNA  prepared  72  hr  later  and  PCR  performed  to  detect  chromosomal  deletions.  PCR 

fragments were isolated and subjected to Amplicon sequencing provided by Genewiz.   

Mammalian  expression  vectors  and  recombinant  protein  expression  constructs  and 

purification. Expression plasmids and purification procedures for XRCC4-WT, XLF-WT, Lig4-WT, 

and Lig4-K273A  have been described (41).The LIG4/XRCC4 complex was produced as described 

(14). All proteins batches were dialyzed against 150 mM KCl, 20 HEPES pH8, 1 mM EDTA, 2 mM 

DTT and 10% (v/v) glycerol, flash frozen in liquid nitrogen and stored at -80 ̊C. T4 DNA ligase was 

obtained from New England Biolabs. LIG4 fragments 1-620 (WT or K273A), DBD, NTase, and OBD 

were prepared as described (50). Mammalian expression constructs for wild-type Lig4 and Lig4-

K273A  were  generated  by  subcloning  from  plasmids  described  above.  The  5XLys  mutant  was 

prepared by subcloning a geneblock (IDT) encoding K273A, K449R, K451R, K352R, and K345R via 

PflMI/BlpI subcloning.  

Ligation  Assay.  Ligation  assays  were  performed  in  a  volume  of  10  μL  containing  50  ng  of 

linearized pUC19 plasmid DNA (with XbaI for cohesive ends and SmaI for blunt ends), 1 mM ATP, 

1 mM DTT, 20 mM Tris-HCl pH8.0, 2 mM MgCl2, and 60 mM KCl, 1% (v/v) glycerol with addition 

of proteins at the indicated concentrations. Reactions were incubated at room temperature for 

30 min  before  addition of  Proteinase  K  at 1.4 mg/mL  final  concentration  and 1X of  a  loading 

buffer containing 0.01% SDS (NEB #B7024S) and incubated for 10 min at 55°C. Samples were next 

resolved  by 0.8% agarose gel electrophoresis in Tris-Borate-EDTA buffer.  Gels were stained in 

Tris-Borate-EDTA  buffer  supplemented  with  0.5  μg/mL  ethidium  bromide  and  destained  in 

deionized  water.  Images  where  captured  under  UV  light  using  a  Bio-Rad  chemidoc  and 

quantitatively analyzed with Image J. 

Oligonucleotides  and antibodies. Oligonucleotides used  in  this study are as  follows;  only one 

strand of the oligonucleotides used for targeting PAM sites are presented, and are without BMHI 

overhangs.  

 48 

 
For 293T Lig4 CRISPR experiments: 

gRNA Lig4: CATACGTTCACCATCTAGCT 

Lig4 K273A HDR: 

CTATTGCAGATATTGAGCACATTGAGAAGGATATGAAACATCAGAGTTTCTACATAGAAACAGCGCTAG

ATGGTGAACGTATGCAAATGCACAAAGATGGAGATGTATATAAATACTTCTCTCGAAATGG 

For U2OS CRISPR experiments:  

gRNA Lig4: GTTCAGCACTTGAGCAAAAG.  

The U20S clone used in the experiments had +1 frameshift mutations on both alleles. 

For 293T polQ CRISPR experiments:   

gRNA1 polQ:TGAAGCGGGTTTTGGAAATG 

gRNA2 polQ:TCTGATCAATCGCCTCATAG 

For 293T FANCG CRISPR experiments:   

gRNA1 FancG:GGGCCAGGCCTGGGTTCAAC 

gRNA2 FancG:GACTTAAGAGAAAGGGACTG 

5’ FancG PCR: CCCAAGATGTCCCGGCTGTGGG  

3’FancG PCR:CCATGGGCCTCTCTGTCCTTGCAC  

Oligonucleotides for substrate PCR:  

5’ coding joint PCR: CGGTGGGAGGTCTATATAAGCA 

3’ coding joint PCR:CTACACCGTGGTGGAGCAGTA 

5’ signal joint PCR:ACCTTGAAGCGCATGAAGGGC 

3’ signal joint PCR:TCCATGCGGTACTTCATGGTC 

Antibodies utilized in this study include: anti-DNA Lig 4 (Abcam, 26039), anti-Ku80 (Invitrogen, 

111), anti-DNA-PKcs (generous gift Tim Carter).  

Data  availability:  The  authors  confirm  that  the  data  supporgng  the  findings  of  this  study  are 

available within the argcle and its supplementary materials; the amplicon sequencing files from 

this study are available from the corresponding author.   

Funding:  USDA  Nagonal  Insgtute  of  Food  and  Agriculture  [1019208];  Public  Health  Service 

[AI048758, AI147634 to K.M.] and [AI 38345, AI 39039 to KY]; French Nagonal Research Agency 

 49 

 
(ANR-17-CE12-0020-01,  ANR-18-CE29-0003-04)  and  French  Nagonal  League  against  cancer 

(équipe labellisée) to M.M. 

Acknowledgements: We are pargcularly indebted to both Dale Ramsden and Adam Luthman for 

their  generous  and  invaluable  assistance  in  analyzing  amplicon  sequencing  projects  for 

differences in juncgonal diversity and uglizagon of short sequence homologies. 

ContribuZon  by  the  author  (CMB  dissertaZon  requirement):  This  work  contains  work  from 

mulgple authors. As first author, Noah Goff performed the bulk of the data collecgon, analysis, 

and contributed to the conceptualizagon/wrigng/edigng of this manuscript—with the excepgon 

of Figure 9 and supplemental figures 1-3. Note the supplemental figures are available in the online 

version of this manuscript at Nucleic Acids Research.   

 50 

 
 
 
1. 

2. 

3. 

4. 

5. 

6. 

7. 

8. 

9. 

REFERENCES 

Reid, D.A., Keegan, S., Leo-Macias, A., Watanabe, G., Strande, N.T., Chang, H.H., Oksuz, 
B.A., Fenyo, D., Lieber, M.R., Ramsden, D.A. et al. (2015) Organizagon and dynamics of the 
nonhomologous  end-joining  machinery  during  DNA  double-strand  break  repair. 
Proceedings of the Na9onal Academy of Sciences of the United States of America, 112, 
E2575-2584. 

Reid, D.A., Conlin, M.P., Yin, Y., Chang, H.H., Watanabe, G., Lieber, M.R., Ramsden, D.A. 
and  Rothenberg,  E.  (2017)  Bridging  of  double-stranded  breaks  by  the  nonhomologous 
end-joining ligagon complex is modulated by DNA end chemistry. Nucleic acids research, 
45, 1872-1878. 

Zhao, B., Watanabe, G., Morten, M.J., Reid, D.A., Rothenberg, E. and Lieber, M.R. (2019) 
The  essengal  elements  for  the  noncovalent  associagon  of  two  DNA  ends  during  NHEJ 
synapsis. Nature communica9ons, 10, 3588. 

Jackson, S.P. (2002) Sensing and repairing DNA double-strand breaks. Carcinogenesis, 23, 
687-696. 

Yoo, S. and Dynan, W.S. (1999) Geometry of a complex formed by double strand break 
repair proteins at a single DNA end: recruitment of DNA-PKcs induces inward translocagon 
of Ku protein. Nucleic acids research, 27, 4679-4686. 

Graham, T.G., Walter, J.C. and Loparo, J.J. (2016) Two-Stage Synapsis of DNA Ends during 
Non-homologous End Joining. Molecular cell, 61, 850-858. 

Chen, S., Lee, L., Naila, T., Fishbain, S., Wang, A., Tomkinson, A.E., Lees-Miller, S.P. and He, 
Y. (2021) Structural basis of long-range to short-range synapgc transigon in NHEJ. Nature, 
593, 294-298. 

Chaplin, A.K., Hardwick, S.W., Stavridi, A.K., Buehl, C.J., Goff, N.J., Ropars, V., Liang, S., De 
Oliveira,  T.M.,  Chirgadze,  D.Y.,  Meek,  K.  et  al.  (2021)  Cryo-EM  of  NHEJ  supercomplexes 
provides insights into DNA repair. Molecular cell, 81, 3400-3409 e3403. 

Chaplin,  A.K.,  Hardwick,  S.W.,  Liang,  S.,  Kefala  Stavridi,  A.,  Hnizda,  A.,  Cooper,  L.R.,  De 
Oliveira, T.M., Chirgadze, D.Y. and Blundell, T.L. (2021) Dimers of DNA-PK create a stage for 
DNA double-strand break repair. Nature structural & molecular biology, 28, 13-19. 

10.  Wang, J.L., Duboc, C., Wu, Q., Ochi, T., Liang, S., Tsutakawa, S.E., Lees-Miller, S.P., Nadal, 
M., Tainer, J.A., Blundell, T.L. et al. (2018) Dissecgon of DNA double-strand-break repair 
using novel single-molecule forceps. Nat Struct Mol Biol, 25, 482-487. 

11. 

Sgnson, B.M., Moreno, A.T., Walter, J.C. and Loparo, J.J. (2020) A Mechanism to Minimize 
Errors during Non-homologous End Joining. Molecular cell, 77, 1080-1091 e1088. 

12.  Waters, C.A., Strande, N.T., Pryor, J.M., Strom, C.N., Mieczkowski, P., Burkhalter, M.D., Oh, 
S., Qaqish, B.F., Moore, D.T., Hendrickson, E.A. et al. (2014) The fidelity of the ligagon step 

 51 

 
13. 

14. 

15. 

16. 

17. 

determines  how  ends  are  resolved  during  nonhomologous  end  joining.  Nature 
communica9ons, 5, 4286. 

Budman,  J.,  Kim,  S.A.  and  Chu,  G.  (2007)  Processing  of  DNA  for  nonhomologous  end-
joining  is  controlled  by  kinase  acgvity  and  XRCC4/ligase  IV.  The  Journal  of  biological 
chemistry, 282, 11950-11959. 

Cojarel,  J.,  Frit,  P.,  Bombarde,  O.,  Salles,  B.,  Negrel,  A.,  Bernard,  S.,  Jeggo,  P.A.,  Lieber, 
M.R., Modesg, M. and Calsou, P. (2013) A noncatalygc funcgon of the ligagon complex 
during nonhomologous end joining. The Journal of cell biology, 200, 173-186. 

Conlin, M.P., Reid, D.A., Small, G.W., Chang, H.H., Watanabe, G., Lieber, M.R., Ramsden, 
D.A.  and  Rothenberg,  E.  (2017)  DNA  Ligase  IV  Guides  End-Processing  Choice  during 
Nonhomologous End Joining. Cell reports, 20, 2810-2819. 

Chiruvella, K.K., Liang, Z., Birkeland, S.R., Basrur, V. and Wilson, T.E. (2013) Saccharomyces 
cerevisiae  DNA  ligase  IV  supports  imprecise  end  joining  independently  of  its  catalygc 
acgvity. PLoS gene9cs, 9, e1003599. 

Lu, H., Shimazaki, N., Raval, P., Gu, J., Watanabe, G., Schwarz, K., Swanson, P.C. and Lieber, 
M.R.  (2008)  A  biochemically  defined  system  for  coding  joint  formagon  in  V(D)J 
recombinagon. Molecular cell, 31, 485-497. 

18.  Ma, Y., Pannicke, U., Schwarz, K. and Lieber, M.R. (2002) Hairpin opening and overhang 
processing by an Artemis/DNA-dependent protein kinase complex in nonhomologous end 
joining and V(D)J recombinagon. Cell, 108, 781-794. 

19. 

20. 

Goodarzi, A.A., Yu, Y., Riballo, E., Douglas, P., Walker, S.A., Ye, R., Harer, C., Marche~, C., 
Morrice,  N.,  Jeggo,  P.A.  et  al.  (2006)  DNA-PK  autophosphorylagon  facilitates  Artemis 
endonuclease acgvity. The EMBO journal, 25, 3880-3889. 

Riballo, E., Kuhne, M., Rief, N., Doherty, A., Smith, G.C., Recio, M.J., Reis, C., Dahm, K., 
Fricke, A., Krempler, A. et al. (2004) A pathway of double-strand break rejoining dependent 
upon ATM, Artemis, and proteins locagng to gamma-H2AX foci. Molecular cell, 16, 715-
724. 

21.  Meek,  K.  (2020)  Acgvagon  of  DNA-PK  by  hairpinned  DNA  ends  reveals  a  stepwise 

mechanism of kinase acgvagon. Nucleic acids research, 48, 9098-9108. 

22. 

23. 

Helmink, B.A., Tubbs, A.T., Dorsej, Y., Bednarski, J.J., Walker, L.M., Feng, Z., Sharma, G.G., 
McKinnon, P.J., Zhang, J., Bassing, C.H. et al. (2011) H2AX prevents CtIP-mediated DNA 
end resecgon and aberrant repair in G1-phase lymphocytes. Nature, 469, 245-249. 

Rooney, S., Sekiguchi, J., Zhu, C., Cheng, H.L., Manis, J., Whitlow, S., DeVido, J., Foy, D., 
Chaudhuri, J., Lombard, D. et al. (2002) Leaky Scid phenotype associated with defecgve 
V(D)J coding end processing in Artemis-deficient mice. Mol Cell, 10, 1379-1390. 

 52 

 
24. 

25. 

26. 

27. 

28. 

29. 

30. 

31. 

Roth,  D.B.,  Menetski,  J.P.,  Nakajima,  P.B.,  Bosma,  M.J.  and  Gellert,  M.  (1992)  V(D)J 
recombinagon: broken DNA molecules with covalently sealed (hairpin) coding ends in scid 
mouse thymocytes. Cell, 70, 983-991. 

Liu, L., Chen, X., Li, J., Wang, H., Buehl, C.J., Goff, N.J., Meek, K., Yang, W. and Gellert, M. 
(2021) Autophosphorylagon transforms DNA-PK from protecgng to processing DNA ends. 
Molecular cell. 

Kabotyanski,  E.B.,  Gomelsky,  L.,  Han,  J.O.,  Stamato,  T.D.  and  Roth,  D.B.  (1998)  Double-
strand break repair in Ku86- and XRCC4-deficient cells. Nucleic acids research, 26, 5333-
5342. 

Cui,  X.  and  Meek,  K.  (2007)  Linking  double-stranded  DNA  breaks  to  the  recombinagon 
acgvagng  gene  complex  directs  repair  to  the  nonhomologous  end-joining  pathway. 
Proceedings of the Na9onal Academy of Sciences of the United States of America, 104, 
17046-17051. 

Sallmyr, A. and Tomkinson, A.E. (2018) Repair of DNA double-strand breaks by mammalian 
alternagve end-joining pathways. The Journal of biological chemistry, 293, 10536-10546. 

Buck, D., Moshous, D., de Chasseval, R., Ma, Y., le Deist, F., Cavazzana-Calvo, M., Fischer, 
A.,  Casanova,  J.L.,  Lieber,  M.R.  and  de  Villartay,  J.P.  (2006)  Severe  combined 
immunodeficiency  and  microcephaly  in  siblings  with  hypomorphic  mutagons  in  DNA 
ligase IV. European journal of immunology, 36, 224-235. 

Chen, S.H. and Yu, X. (2019) Human DNA ligase IV is able to use NAD+ as an alternagve 
adenylagon donor for DNA ends ligagon. Nucleic acids research, 47, 1321-1334. 

Dedon, P.C. and Goldberg, I.H. (1992) Free-radical mechanisms involved in the formagon 
of sequence-dependent bistranded DNA lesions by the angtumor angbiogcs bleomycin, 
neocarzinostagn, and calicheamicin. Chem Res Toxicol, 5, 311-332. 

32.  Wu, C.C., Li, T.K., Farh, L., Lin, L.Y., Lin, T.S., Yu, Y.J., Yen, T.J., Chiang, C.W. and Chan, N.L. 
(2011)  Structural  basis  of  type  II  topoisomerase  inhibigon  by  the  angcancer  drug 
etoposide. Science, 333, 459-462. 

33. 

34. 

35. 

Neal,  J.A.  and  Meek,  K.  (2019)  Deciphering  phenotypic  variance  in  different  models  of 
DNA-PKcs deficiency. DNA repair, 73, 7-16. 

Roth,  D.B.  and  Wilson,  J.H.  (1986)  Nonhomologous  recombinagon  in  mammalian  cells: 
role for short sequence homologies in the joining reacgon. Molecular and cellular biology, 
6, 4295-4304. 

Corneo, B., Wendland, R.L., Deriano, L., Cui, X., Klein, I.A., Wong, S.Y., Arnal, S., Holub, A.J., 
Weller,  G.R.,  Pancake,  B.A.  et  al.  (2007)  Rag  mutagons  reveal  robust  alternagve  end 
joining. Nature, 449, 483-486. 

 53 

 
36. 

Oh,  S.,  Harvey,  A.,  Zimbric,  J.,  Wang,  Y.,  Nguyen,  T.,  Jackson,  P.J.  and  Hendrickson,  E.A. 
(2014) DNA ligase III and DNA ligase IV carry out genegcally disgnct forms of end joining 
in human somagc cells. DNA repair, 21, 97-110. 

37.  Masani, S., Han, L., Meek, K. and Yu, K. (2016) Redundant funcgon of DNA ligase 1 and 3 
in alternagve end-joining during immunoglobulin class switch recombinagon. Proceedings 
of the Na9onal Academy of Sciences of the United States of America, 113, 1261-1266. 

38. 

Deriano, L., Stracker, T.H., Baker, A., Petrini, J.H. and Roth, D.B. (2009) Roles for NBS1 in 
alternagve nonhomologous end-joining of V(D)J recombinagon intermediates. Molecular 
cell, 34, 13-25. 

39.  Meek,  K.  (1990)  Analysis  of  Juncgonal  Diversity  during  Lymphocyte-B  Development. 

Science, 250, 820-823. 

40. 

41. 

42. 

43. 

44. 

45. 

Lewis, S. and Gellert, M. (1989) The mechanism of anggen receptor gene assembly. Cell, 
59, 585-588. 

Normanno,  D.,  Negrel,  A.,  de  Melo,  A.J.,  Betzi,  S.,  Meek,  K.  and  Modesg,  M.  (2017) 
Mutagonal phospho-mimicry reveals a regulatory role for the XRCC4 and XLF C-terminal 
tails in modulagng DNA bridging during classical non-homologous end joining. Elife, 6. 

Kienker, L.J., Shin, E.K. and Meek, K. (2000) Both V(D)J recombinagon and radioresistance 
require  DNA-PK  kinase  acgvity,  though  minimal  levels  suffice  for  V(D)J  recombinagon. 
Nucleic acids research, 28, 2752-2761. 

Kurimasa, A., Kumano, S., Boubnov, N.V., Story, M.D., Tung, C.S., Peterson, S.R. and Chen, 
D.J. (1999) Requirement for the kinase acgvity of human DNA-dependent protein kinase 
catalygc subunit in DNA strand break rejoining. Molecular and cellular biology, 19, 3877-
3884. 

Lu,  H.,  Saha,  J.,  Beckmann,  P.J.,  Hendrickson,  E.A.  and  Davis,  A.J.  (2019)  DNA-PKcs 
promotes chromagn decondensagon to facilitate inigagon of the DNA damage response. 
Nucleic acids research, 47, 9467-9479. 

Jiang, W., Crowe, J.L., Liu, X., Nakajima, S., Wang, Y., Li, C., Lee, B.J., Dubois, R.L., Liu, C., 
Yu,  X.  et  al.  (2015)  Differengal  phosphorylagon  of  DNA-PKcs  regulates  the  interplay 
between end-processing and end-ligagon during nonhomologous end-joining. Molecular 
cell, 58, 172-185. 

46.  Meek,  K.,  Douglas,  P.,  Cui,  X.,  Ding,  Q.  and  Lees-Miller,  S.P. 

(2007)  trans 
Autophosphorylagon at DNA-dependent protein kinase's two major autophosphorylagon 
site clusters facilitates end processing but not end joining. Molecular and cellular biology, 
27, 3881-3890. 

 54 

 
47. 

Roy, S., de Melo, A.J., Xu, Y., Tadi, S.K., Negrel, A., Hendrickson, E., Modesg, M. and Meek, 
K. (2015) XRCC4/XLF Interacgon Is Variably Required for DNA Repair and Is Not Required 
for Ligase IV Sgmulagon. Molecular and cellular biology, 35, 3017-3028. 

48.  Mali, P., Yang, L., Esvelt, K.M., Aach, J., Guell, M., DiCarlo, J.E., Norville, J.E. and Church, 

G.M. (2013) RNA-guided human genome engineering via Cas9. Science, 339, 823-826. 

49. 

50. 

Neal,  J.A.,  Xu,  Y.,  Abe,  M.,  Hendrickson,  E.  and  Meek,  K.  (2016)  Restoragon  of  ATM 
Expression in DNA-PKcs-Deficient Cells Inhibits Signal End Joining. Journal of immunology, 
196, 3032-3042. 

Ochi, T., Gu, X. and Blundell, T.L. (2013) Structure of the catalygc region of DNA ligase IV 
in complex with an Artemis fragment sheds light on double-strand break repair. Structure, 
21, 672-679. 

 55 

 
 
 
 
Chapter 3: New insight into how the DNA binding domain of DNA ligase IV facilitates end-
joining, independent of its catalyZc acZvity. 

By:  

Noah J. Goff, Saanjh Sundar, Addy Walia, Gargi Parkhi, Katheryn Meek 

Data Chapter 

 56 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ABSTRACT 

DNA double strand breaks (DSBs) are highly genotoxic lesions generated throughout the cell 

cycle by cellular processes (DNA tension, replicagon errors, metabolic stress) or exogenous agents 

(ionizing radiagon, chemotherapeugc drugs). Cells have several pathways to resolve DSBs and 

miggate their insult to genomic integrity. End-joining pathways, most notably non-homologous 

end  joining  (NHEJ),  are  the  most  readily  available  DSB  repair  mechanisms—being  available 

throughout the cell cycle and capable of repairing a wide range of damage adducted DNA-ends. 

Recent studies have revealed a two-stage synapgc mechanism for NHEJ, where DNA Ligase IV (L4) 

acts both catalygcally to ligate the broken strands and as an essengal structural factor that drives 

progression  from  long-range  synapgc  complexes  (LRCs)  into  a  ligagon-capable  short-range 

synapgc complex (SRC). Here we report a novel cryo-EM structure containing density for L4’s DNA 

binding domain (DBD) in the LRC, something not previously seen in any reported structures. The 

L4 side of that interface is shared with a trans interacgon between L4 and Ku70 in the SRC. We 

characterize how the L4 DBD promotes end-joining in the absence of L4 catalygc acgvity through 

interacgons in the SRC with DNA and Ku70. Each of these interacgons are essengal to the capacity 

of catalygcally inacgve L4 to promote joining. Moreover, we have discovered that catalygcally 

inacgve  L4  requires  a  specific  LRC  to  promote  end-joining.  Finally,  we  have  observed  several 

instances of a single structural interface in one protein mediagng interacgons with different NHEJ 

components, and we discuss potengal implicagons of this finding.  

 57 

 
 
 
INTRODUCTION 

Preservagon  of  genegc  material  is  paramount  to  long-term  health  and  survival  of  all 

organisms. Threats to genomic stability either result from common environmental agents or are 

byproducts of essengal cellular funcgons, including reacgve byproducts of metabolism, torsional 

stress within the nucleus, autonomous nuclear elements, and some viruses among other sources. 

Failure  of  DNA  repair  pathways  in  eukaryotes  omen  results  in  harmful  lesions  that  can  be 

converted  into  mutagons,  impair  cellular  funcgon,  contribute  to  development  of  cancers,  or 

eventually drive cell death. DNA double strand breaks (DSBs) are considered the most genotoxic 

lesions—with inappropriate repair resulgng in a range of outcomes from small delegons to large 

chromosomal rearrangements (79–81, 52, 45).  

Direct end-joining is the primary repair mechanism available throughout the cell cycle and 

is primarily mediated through the canonical non-homologous end-joining pathway (NHEJ) (82). A 

general model of NHEJ begins with sensing damage, synapsis of two ends, processing of chemical 

adducts  to  produce  compagble  3’  hydroxyl  and  5’  phosphate  groups  and  complegon  of  the 

process by direct ligagon of the DNA ends. The core NHEJ factors in mammals (and many other 

eukaryotes)(35) include the DNA dependent protein kinase [DNA-PK, consisgng of the DNA end-

binding heterodimer, (Ku70/Ku80), and the catalygc subunit DNA-PKcs] that acts as a damage 

sensor, structural factors XRCC4, XRCC4-Like Factor (XLF) and paralog of XRCC4 and XLF (PAXX), 

and the NHEJ-specific DNA ligase IV (L4) that performs the final ligagon step. NHEJ’s availability 

throughout  the  cell  cycle  addresses  a  central  challenge  of  cellular  DNA  repair:  how  to  repair 

lesions  when  a  homologous  template  is  rarely  available?  As  a  result,  NHEJ  is  both  remarkably 

flexible and fast, but increases the risk of small indels or larger chromosomal instability. Recent 

work in the field has developed a more detailed two-stage synapgc model of NHEJ, where ends 

are synapsed and protected from aberrant end processing in a “long-range” synapgc complex, 

essengal  end  processing  is  performed  and  required  downstream  factors  are  recruited  before 

ulgmately transigoning into a “short-range” complex where ends become accessible to either L4 

or  further  end-processing  by  X-family  DNA  polymerases  lambda  and  mu  (Poll  and  Polm),  the 

phosphodiesterase TDP1, and the nucleases Artemis, MRN, and others (21, 69, 83, 84).  

 58 

 
Work from Loparo and colleagues first proposed a two-stage model of DNA end synapsis 

by NHEJ factors (46). Their single molecule FRET studies revealed both short-range and long-range 

synapses where ends are held in either direct proximity (short-range complex) or ~115Å apart 

(long-range  complex)  termed  SRC  and  LRC  respecgvely.  Transigon  into  a  short-range  complex 

required the presence of Ku70, Ku80, XLF, XRCC4, L4, DNA-PKcs kinase acgvity (but not necessarily 

its presence), and crigcally L4, but not its catalygc acgvity (46).  Further FRET and biochemical 

studies from these authors revealed specific roles for Poll mediated fill-in synthesis and TDP1-

dependent phosphodiesterase in the SR complex (21).  

Interesgngly,  while  Loparo  and  colleagues  worked  on  single  molecule  imaging,  new 

structural data from several groups revealed diverse NHEJ complexes consisgng of DNA-PK bound 

to a single DNA end or in complex with other NHEJ factors (XLF, XRCC4/L4, PAXX) to tether two 

ends together in a dimeric structure. Remarkably, these new NHEJ structures revealed dimeric 

complexes with DNA ends synapsed in direct contact, in good agreement with the SR complex 

predicted  by  the  smFRET  approaches,  as  well  as  mulgple  disgnct  iteragons  of  LR  synapgc 

complexes, posigoning the two ends ~115Å apart (18, 22, 70).  

DNA  ends  synapsed  by  the  LR  complexes  can  be  classified  into  two  groups:  1)  Ku80-

mediated dimers dominated by trans interacgons between DNA-PKcs and the extreme C-terminal 

tail of Ku80, and 2) XLF-mediated dimers dominated by a pair of DNA-PKcs interfaces interacgng 

reciprocally in trans along with a centrally posigoned XLF dimer interacgng with two dimers of 

XRCC4 (XRCC4-XLF-XRCC4) providing mulgple, protein-protein interfaces that stabilize the DNA 

synapse.  Our  recent  structure-funcgon  mutagenesis  studies  suggest  that  LR  dimer  complexes 

have  disgnct  funcgons  in  promogng  NHEJ.  Specifically,  the  Ku80-mediated  dimer  promotes 

access  to  DNA  ends  for  nucleolygc  end  processing,  while  the  XLF-mediated  dimer  limits 

nucleolygc  processing—matching  the  “end  protecgon”  complex  suggested  by  Loparo  and 

colleagues  (20).  Our  current  limited  understanding  of  how  NHEJ  progresses  stems  from  all  of 

these studies (with all the unique approaches uglized); acgve areas of research now focus on 

understanding how LR complexes translate between the two dimer forms and what is required 

to progress from LR complexes to SR complexes.  

 59 

 
 It is now well-appreciated that L4 has an essengal role in promogng the transigon to the 

short-range complex (10, 21, 46, 73). Recent studies have provided more informagon as to how 

L4 contributes to the transigon from long-range to short-range end synapsis (20, 70). This funcgon 

is unique among mammalian DNA ligases. The catalygc domains of the three human DNA ligases 

contain three mogfs: a DNA binding domain (DBD), a nucleogdyltransferase domain (NTD), and 

an oligonucleogde oligosaccharide fold domain (OBD). DNA ligases III and VI, but not ligase I (L1, 

L3, and L4) also include BRCT repeats (39). In the case of L4, a small region of the polypepgde 

between its two BRCT domains mediates L4’s interacgon with its obligate co-factor XRCC4 (9). 

L4’s catalygc core has generally not been observed in most of the recent Cryo-EM structures, with 

the notable excepgon of the SR complex when the catalygc domain is seen bound to a DSB (18–

20). Of note, single molecule imaging experiments  revealed that L4 dynamically interacts with 

each DNA end independently in the LRC; these interacgons are mediated by posigvely charged 

residues in the DBD that were observed (in Cryo-EM structures of human L4 in the SRC) to make 

electrostagc interacgons with the DNA backbone, slightly internal to the DNA termini (48). Cortes 

and colleagues using engineered chimeras of human DNA ligases revealed that chimeric proteins 

including  the  BRCT  domain  of  L4  with  L1’s  NTD  and  OBD,  but  not  L1’s  DBD  produces  a  fusion 

protein that supports NHEJ-specific VDJ recombinagon (85). The finding that fusing L1’s engre 

catalygc core (including its DBD) with L4’s BRCT-mogfs was insufficient to promote ligagon via 

NHEJ, directly implicates L4’s DBD in facilitagng L4's non-catalygc funcgon in NHEJ.  

More experiments that promote the conclusion that L4 has an important structural role 

in  joining  comes  from  our  recent  single  molecule  imaging  study  showing  that  whereas  L4  is 

required for stable recruitment of DNA-PK to chromagn amer DNA damage, catalygcally inacgve 

L4 supports its recruitment. Of note, retengon of DNA-PK at DSBs in cells expressing catalygcally 

inacgve  L4  is  markedly  longer  than  acgve  L4;  these  data  suggest  that  repair  facilitated  by 

catalygcally inacgve L4 is slower than when L4 is acgve (73). 

In  this  study,  we  further  invesggate  L4’s  structural  role  in  promogng  end-joining  using 

NHEJ funcgonal assays in genegcally engineered cell strains and cryo-electron microscopy (Cryo-

EM, data not shown). We confirm that repair facilitated by catalygcally inacgve L4 is much slower 

than with wild type L4, and that repair is pargally dependent on DNA polymerase theta (Polθ). 

 60 

 
Moreover,  repair  promoted  by  inacgve  L4  is  highly  dependent  on  the  Ku80-mediated  LRC.  In 

addigon, we report a new version of the Ku80-mediated LRC with density for L4’s DBD interacgng 

with DNA-PKcs. Of note, the interface that mediates L4's interacgon with DNA-PKcs also facilitates 

a previously reported trans interacgon between L4's DBD and Ku70 that tethers DNA ends in the 

SR complex (18). Funcgonal assays establish the relevance of this L4 DBD/Ku70 interacgon for 

end joining, whereas the interacgon of L4's DBD with DNA-PKcs appears to be dispensable for 

NHEJ. This interface is disgnct from the two DNA/protein interfaces studied by Sgnson et al (48) 

that  impact  L4's  tethering  of  DNA  ends  in  single  molecule  FRET  experiments.  We  posit  that 

numerous interacgons of the DBD with Ku and the two DNA termini are essengal in promogng 

transigon from LR to SR complexes where end-joining occurs. 

A

RESULTS 

B

CCTGCAG
GGACGTCTTAAG

CGCGTCCTGCAG
AGGACGTC

Figure 13: Microhomology mediated end joining is mediated through Polθ independent of NHEJ 
deficiency.  A)  Schemagc  of  MMEJ  fluorescent  reporter  assay,  with  7  bp  of  microhomology 
highlighted  and  underlined.  Substrates  were  pre-cut  with  ApoI  and  MluI  (NEB)  to  generate 
noncomplementary  overhangs,  then  cotransfected  with  a  ECFP  plasmid  to  assess  transfecgon 
efficiency. Use of this microhomology is required to restore the full ds-red reading frame to allow 
MMEJ+  signal.  MMEJ+  signal  was  assessed  as  %  of  transfected  cells  that  expressed  DsRed.  B) 
MMEJ joining acgvity in edited 293T cell strains, independent of L4 presence or catalygc acgvity. 
Stagsgcs performed using an unpaired two-tailed Student’s t-test (**** p > 0.0001). 

 61 

 
 
 
Microhomology-mediated end joining (MMEJ) is dependent on Polθ. We recently showed that 

catalygcally  inacgve  L4  promotes  substangal  end-joining  and  radioresistance  in  cell  culture 

models,  bolstering  earlier  reports  documengng  a  non-catalygc  role  for  L4  in  NHEJ.  Our  study 

suggests that joining in cells expressing inacgve L4 [a K273A mutagon] is likely mediated by L3, 

but in the context of an undefined NHEJ complex (10). This cooperagon between L4 and L3 is 

strongly corroborated by a recent collaboragve study of catalygcally inacgve L4 in mice that lack 

nuclear L3 (14). In a separate publicagon, analyses of NHEJ factor recruitment to DNA DSBs in 

living  cells  revealed  that  whereas  in  cells  lacking  L4,  NHEJ  factor  recruitment  (using  Halo-Flag 

tagged Ku70, DNA-PKcs, and XRCC4) is transient (in the case of Ku70 and DNA-PKcs) and absent 

for  XRCC4;  but  in  cells  expressing  the  K273A  mutant,  although  inigal  NHEJ  factor  recruitment 

occurs similarly as in cells expressing acgve L4, dissociagon of NHEJ factors is markedly stalled 

(73).  

The increased retengon of NHEJ factors in cells expressing L4 K273A suggests that repair 

may occur less rapidly than in NHEJ-proficient cells where end-joining is rapid, and (when DNA 

ends are ligatable and compagble) generally accurate. Sequencing across rejoined DSBs in cells 

expressing  catalygcally  inacgve  L4  revealed  lower  fidelity  repair  outcomes  with  an  increased 

dependence  on  ligagng  ends  across  short  homologous  sequences  near  the  break  (1-5  bp, 

frequently referred to as microhomology). This is characterisgc of end-joining facilitated by Polθ, 

which is notably slower than NHEJ (and therefore only happens at low levels in NHEJ proficient 

cells) . Addigonally, rejoined DSBs from cells expressing L4 K273A share characterisgcs of joints 

mediated by Polθ (increased nucleogde loss and increased use of terminal microhomology) (15). 

To assess the role of Polθ in cells expressing L4 K273A, Polθ was ablated using a CRISPR/Cas9 

strategy; the impact of Polθ loss specifically on repair of DSBs via terminal microhomology was 

tested using a linearized plasmid substrate that requires use of a 7-bp short-sequence homology 

internal  to  the  break  site  to  restore  an  RFP  open  reading  frame.    All  cells  proficient  in  Polθ, 

regardless  of  NHEJ  proficiency,  support  basal  levels  of  MMEJ.  In  contrast,  MMEJ  joining  is 

remarkably reduced in Polθ ablated cells expressing either wild-type L4 or catalygcally inacgve L4 

(Figure  13B).  [This  joining  substrate  reports  only      on  MMEJ  mediated  joints,  explaining  the 

independence from NHEJ acgvity.] Consistent with previous reports (77, 86, 87), Polθ mediated 

 62 

 
 
end joining funcgons independently from end joining catalyzed by NHEJ. We conclude that Polθ 

is required for MMEJ mediated joining, and that catalygcally inacgve L4 cannot facilitate MMEJ 

in the absence of Polθ. 

A

B

C

Figure 14: CatalyZcally inacZve L4 supports end joining through mulZple alt-EJ pathways. A) 
Sensigvity of 293T cells to teniposide as assessed by MTT assay. B) Sensigvity of 293T cells to 
calicheamicin as assessed by MTT assay. C) Time-course version of a VDJ episomal assay, with 
fixed cells represengng joining amer 24, 48, or 72 hours as analyzed by flow cytometry.  

In  cells  expressing  catalyZcally  inacZve  L4,  resistance  to  DSB-inducing  agents  requires  Polθ. 

Since Polθ promotes substangal amounts of MMEJ, we reasoned that MMEJ may play a more 

prominent  role  in  the  alternagve  end  joining  promoted  by  cells  expressing  L4  K273A.  Cells 

deficient in Polθ were not sensigve to either calicheamicin (that generates DSBs with 3’ overhangs 

with 3’ phosphoglycolate adducts) or teniposide (a type II topoisomerase poison that leaves 5’-

pepgde adducted DSB overhangs primarily in S phase). However, L4 K273A expressing cells that 

 63 

 
 
 
lack Polθ, are markedly sensigve to both drugs, approaching the level of sensigvity observed in 

cells that completely lack L4 (Figure 14A-B). Because Polθ cannot facilitate sufficient end-joining 

in L4 deficient cells (as evidenced by a synthegc lethal interacgon between loss of Polθ and NHEJ), 

these data suggest that Polθ may promote end-joining in the context of a catalygcally inacgve L4 

complex.  

VDJ coding joining facilitated by catalyZcally inacZve L4 is slower than wild type NHEJ and is 

not dependent on Polθ. Time course assays were performed to assess the rate of VDJ coding end 

joining (Figure 14C). As shown previously, cells expressing K273A L4 join RAG-induced DSBs (in 

this case coding ends) much more efficiently than cells lacking L4 (10). However, joining in the 

K273A cells is delayed compared to joining from cells expressing acgve L4. Finally coding joint 

efficiency or rate is not impacted by loss of Polθ in cells expressing either wild type or K273A L4. 

Altogether,  these  data  suggest  that  K273A  L4  contributes  to  a  variety  of  repair  mechanisms, 

including  both  Polθ  mediated  alt-EJ  and  L3-mediated  joining  [as  shown  by  our  recent 

collaboragve study (14)]. 

End-joining facilitated by K273A L4 requires the Ku80-mediated long-range NHEJ complex.  As 

noted  above,  accumulagng  data  support  a  new  model  whereby  NHEJ  funcgons  by  inigally 

synapsing two DNA ends in LR complexes posigoning the ends ~115 Å apart. A fracgon of LRCs 

transigon to SR complexes that juxtapose ends close enough to facilitate ligagon (46). Emerging 

cryo-EM data have dramagcally bolstered and expanded this model providing structural evidence 

for a variety of disgnct LRCs, but only one SRC (18–20). What is lacking is a clear understanding 

of how complexes transigon from one form to another, or if catalygcally inacgve L4 is dependent 

on a specific complex. The slow rate of joining in cells expressing K273A L4 as well as the delayed 

dissociagon of NHEJ factors from chromagn in cells expressing K273A L4 might suggest stalling in 

a pargcular complex.  

To  test  the  impact  of  L4  inacgvity  when  DNA-PK  mutagon  disrupts  normal  transigon 

between LR and/or SR complexes, a series of episomal end-joining assays were performed in a 

293T K273A cells strain that was ablated for DNA-PKcs expression using a CRISPR strategy. In these 

assays, the interplay between DNA-PKcs and L4 can be ascertained by complemengng the DNA-

PKcs deficiency using mutant constructs that disrupt either the Ku80-mediated dimer (4A) or the 

 64 

 
XLF-dependent dimer (898/2569) or both (4A/2569, Figure 15). In these experiments VDJ coding 

and signal joining assays were performed using a previously described hyper-RAG2 mutant that 

induces substangally higher levels of recombinagon; this accentuates differences in end-joining 

capacity. As can be seen, K273A cells support substangal VDJ coding and signal end-joining in cells 

complemented with wild type DNA-PKcs (although sgll less than observed in cells with wild type 

L4). As reported previously, DNA-PKcs mutants disrupgng either of the two LR complexes have 

modest deficiencies in joining in cells with wild type L4 (20). When complemengng the DNA-PKcs 

deficiency with the 898/2569 mutant that disrupts the XLF-dependent dimer, signal and coding 

end-joining are proporgonally reduced in cells with K273A L4, as compared to cells with wild type 

L4. In contrast, both signal and coding end-joining in cells expressing inacgve L4 is virtually ablated 

in complementagon experiments with DNA-PKcs mutants that disrupt the Ku80-mediated dimer 

(4A and 4A/2569, Figure 15). These data suggest that interacgons mediated by the domain swap 

dimer are required to facilitate joining promoted by inacgve L4.  

Figure 15: End-joining facilitated by K273A L4 requires the Ku80-mediated long-range NHEJ 
complex. A-B) VDJ coding end (lem) and signal end (right) joining assays in 293T PKcs-/- L4+/+ 
(solid bars) and PKcs-/- L4K273A/- (striped bars). 

L4’s DNA binding domain (DBD) interacts with the M-HEAT region of DNA-PKcs  in  the  Ku80-

mediated  dimer.  During  refinement  of  recent  NHEJ  complexes,  a  density  previously  not 

appreciated was idengfied as a part or L4’s DBD (Chloe L. Hall, Steven W. Hardwick, and Amanda 

K. Chaplin. Data not shown). Of note, L4's DBD has not previously been observed in any of the 

 65 

 
 
 
different  forms  of  the  LRC  (18,  19).  This  interacgon  appears  to  be  mediated  by  two  highly 

conserved posigvely charged residues R136 and K140. L4 R136 interacts with D1440 in DNA-PKcs 

whereas  L4  K140  interacts  with  DNA-PKcs  at  residues  1495-1498.  There  is  also  a  potengal 

interacgon of a loop within the FAT domain (3197-3226) that interacts with L4's acgve site. Finally, 

there is a potengal interacgon of L4 K264 with Ku70 T307. This density for L4 and its interacgon 

with DNA-PKcs has only been observed in the Ku80-mediated dimer.  

The  DNA-PKcs  interface  that  interacts  with  L4’s  DBD  is  not  required  for  efficient  NHEJ.  To 

ascertain the funcgonal relevance of this DNA-PKcs/L4 interacgon, expression constructs ablagng 

the interacgng interfaces in the two proteins were generated. In DNA-PKcs, residues D1440 and 

E1497 were changed to arginine and lysine respecgvely (DE>RK). In L4, residues R136 and K140 

in the DBD were subsgtuted with aspargc acid (RK>DD). Mutant DE>RK DNA-PKcs was tested in 

transient episomal end-joining assays in the DNA-PKcs deficient 293T cell line described above. 

As  can  be  seen,  DNA-PKcs  DE>RK  restores  end-joining  at  or  above  WT  levels  (in  this  case  VDJ 

coding  joints)  as  well  as  wild  type  DNA-PKcs  (Figure  16B).  Because  we  considered  that  the 

tethering of L4 might be relevant to the LR to SR transigon (that is possibly altered in cells with 

inacgve L4), end-joining was also tested in the K273A expressing 293T cells with CRISPR-ablated 

DNA-PKcs (described above, Figure 16C). DE>RK mutant DNA-PKcs restores joining to a similar 

level as wild type DNA-PKcs, albeit at a reduced level compared to cells expressing acgve 293T 

cells. We next generated stable cell strains expressing either wild type, DE>RK, or no DNA-PKcs in 

the V3 CHO cell strain that lacks DNA-PKcs (Figure 16D). While cells lacking DNA-PKcs are hyper-

sensigve  to  both  calicheamicin  and  teniposide,  cells  expressing  DE>RK  mutant  DNA-PKcs  are 

similarly  resistant  to  these  DSB-inducing  agents  as  wild  type  DNA-PKcs.  We  conclude  that  the 

interface in DNA-PKcs that interacts with L4’s DBD in the Ku80-mediated dimer is not required for 

cellular resistance to DSB-inducing drugs or for rejoining DSBs in episomal end-joining assays.  

 66 

 
Figure 16: The DNA-PKcs interface that interacts with L4’s DBD is not required for efficient NHEJ.  
A)  293T  Western  with  transient  complementagon  of  Vect,  WT,  DE>RK  in  both  Lig4+/+  and 
Lig4K273A*/- B) 293T joining assays, VDJ coding PK22 lem c195-14 right C) Western assessing DNA-
PKcs expression in clonal populagons of CHO-V3 cells complemented with empty vector (vect), 
WT DNA-PKcs (WT) or DNA-PKcs DE>RK (DE>RK) D) Colony formagon assays tesgng sensigvity to 
Calicheamicin (lem) and teniposide (right). 

The L4 interface that interacts with DNA-PKcs is important for efficient NHEJ in cells with acZve 

L4, but essenZal for NHEJ in catalyZcally inacZve L4. Expression vectors encoding L4 with only 

the RK>DD subsgtugons were prepared; in addigon, the RK>DD subsgtugon was combined with 

addigonal L4 mutagons targegng its catalygc acgvity [including the K273A mutagon that ablates 

the major adenylagon site, and also K273A with five addigonal lysine>arginine subsgtugons (5xK) 

that ablate a small level of “back-up” adenylagon observed in in vitro experiments with the K273A 

mutant] (10). To test whether the DBD interface that can interact with DNA-PKcs impacts joining, 

episomal end-joining assays as described above were performed. Consistent with our previous 

study,  in  the  absence  of  L4,  virtually  no  VDJ  coding  end  joining  is  observed  (Figure  17A-C);  in 

contrast, wild type L4 (WT), as well as catalygcally inacgve L4 mutants (K273A, 5xK mutants) all 

facilitate  substangal  coding  end  joining.  When  L4  is  acgve,  the  RK>DD  mutant  promotes 

substangal coding end joining, albeit reduced compared to acgve L4.  However, when the RK>DD 

mutagon is combined with inacgve L4, joining is virtually ablated.  

 67 

 
 
 
It is well-established that the RAG1/2 endonuclease directs its DSBs into the NHEJ pathway 

by a poorly understood mechanism (reference). To address whether the RK>DD mutant impacts 

joining of DSBs that are not exclusively repaired by NHEJ, joining of DSBs induced by both TelN (a 

DNA  hairpin  forming  phage  protelomerase)  and  I-SceI  (an  endonuclease  that  produces 

noncomplementary,  4  base  pair  3’  overhangs)  were  studied.  We  previously  found  that  cells 

expressing inacgve L4 can join non-RAG induced DSBs more efficiently than RAG DSBs (10). Similar 

to VDJ-coding hairpins, cells expressing the RK>DD mutant are modestly deficient in joining TelN 

induced DSBs; the deficiency is exacerbated when the RK>DD subsgtugon is combined with loss 

of catalygc acgvity (Figure 17D). Interesgngly, cells expressing the mutant combining inacgve L4 

with the RK>DD subsgtugon are also severely deficient in joining non-hairpin I-SceI breaks despite 

the increased level of joining reported previously with inacgve L4 (12). 

Stable  complementagon  of  wild  type  and  mutant  L4  was  performed  uglizing  a  PiggyBac 

retrotransposon system to stably introduce L4 into L4 deficient 293T and U2OS cell lines. To test 

whether  cells  expressing  the  RK>DD  mutant  are  also  more  sensigve  to  agents  that  induce 

genomic  DSBs,  stable  cell  strains  expressing  wild  type  or  mutant  L4  were  treated  with 

calicheamicin and teniposide. Cells expressing inacgve L4 (either K273A or 5xK) or the RK>DD 

mutant L4 are more resistant to calicheamicin and teniposide than cells lacking L4. Consistent 

with the end-joining assays, L4 mutants that combine the RK>DD subsgtugon with mutagons that 

disrupt  catalygc  acgvity  are  as  sensigve  to  calicheamicin  as  cells  lacking  L4  (Figure  17B).  We 

conclude  that  R136  and  K140  are  crigcal  for  end-joining;  the  observagon  that  the  impact  of 

mutagon  of  these  residues  potengates  deficits  in  joining  mediated  by  catalygcally  inacgve  L4 

suggests that interacgons mediated by these residues are important for L4's non-catalygc role in 

repair via NHEJ. 

 68 

 
A

D

G

B

C

E2-Crimson

ZsGreen

RAG/TelN/I-SceI
Cut Sites

E

F

H

I

D
D
>
K
R
+
A
3
7
2
K

D
D
>
K
R
+
K
x
5

K
x
5

D
D
>
K
R

A
2
7
2
K

t
c
e
V

T
W

DNA-PKcs

Lig4

Figure 17: The L4 interface that interacts with DNA-PKcs is important for efficient NHEJ in acZve 
L4, but essenZal for NHEJ in catalyZcally inacZve L4. A) Diagram of the joining assay substrate, 
with cut sites . Substrates were cotransfected into cells along with and without enzyme (+/-) and 
appropriate L4 expression vector, then red/green fluorescence via flow cytometry at 72 hours. B) 
Results  from  the  VDJ  coding-  and  C)  signal-end  substrates  when  cotransfected  with  Rag1  and 
hyper mutant Rag2 along with the corresponding L4 expression vector. D-E) TelN/I-SceI substrates 
cotransfected  with  endonuclease  (-)  or  with  appropriate  endonuclease  and  corresponding  L4 
expression vector. F) VDJ coding-end joining assays as before tesgng addigonal mutagons in the 
L4 DBD G) Sensigvity of 293T L4-/- cells complemented with PiggyBacc-expressed L4 mutants to 
calicheamicin as assessed by MTT assay, compared to 293T L4+/+ (termed 293T WT). H) Sensigvity 
of PiggyBacc-edited U2OS L4 cell strains as assessed by clonogenic survival assays. I) Immunoblot 
showing transient L4 expression in a 293T L4 -/- cell strain with DNA-PKcs levels as a control. 

The  L4  interface  that  interacts  with  DNA-PKcs  in  the  Ku80-mediated  dimer  forms  a  highly 

conserved  salt-bridge  interacZon  with  Ku70  in  the  short-range  complex.  Mulgple  Sequence 

 69 

 
 
 
 
 
 
 
Alignments  performed  by  COBALT  revealed  substangal  conservagon  of  R136  and  K140,  from 

numerous  vertebrate  species  (Figure  18).  To  clarify  the  discrepancies  of  mutagng  these  two 

interacgng  interfaces  in  L4  and  DNA-PKcs,  we  searched  published  structures  (18,  19,  22)  of 

mulgmeric NHEJ complexes for interacgons between L4 and structural factors that could explain 

how disrupgon of R136 and K140 might impact NHEJ. R136 and K140 are only structured in two 

available structures of NHEJ complexes: the interacgon with DNA-PKcs described above, and in 

the short-range complex reported by He and colleagues (18). In this publicagon, L4 residues 134-

143 were observed interacgng with the von Willebrand domain of Ku70. Specifically, R136 in L4's 

DBD bridges directly with D192 in Ku70 in a trans interacgon (Figure 17). We posit that L4's ability 

to  tether  two  DNA  ends  is  mediated  by  numerous  interacgons  including  1)  protein/DNA 

interacgons  involving  two  patches  of  posigvely  charged  residues  described  by  Loparo  and 

colleagues  (48),  2)  protein/DNA  interacgons  with  L4's  catalygc  site,  3)  the  trans  interacgon 

between L4 R136 and K140 with Ku70 D192 (18), and finally 4) the capacity to form the Ku80-

mediated dimer. Whereas disrupgon of any one of these interacgons results in only a pargal loss 

of end-joining and radio-resistance, loss of two results in severe deficits in both.  

 70 

 
B

D

F

Homo sapiens
Mus musculus
Ra>us norvegicus
Danio rerio
Xenopus laevis
Equus caballus
Gallus gallus
Bos taurus
Canus lupus familiaris
Pan troglodytes 

Homo sapiens
Mus musculus
Ra>us norvegicus
Danio rerio
Xenopus laevis
Equus caballus
Gallus gallus
Bos taurus
Canus lupus familiaris
Pan troglodytes 

Homo sapiens
Mus musculus
Ra>us norvegicus
Danio rerio
Xenopus laevis
Equus caballus
Gallus gallus
Bos taurus
Canus lupus familiaris
Pan troglodytes 

Homo sapiens
Mus musculus
Ra>us norvegicus
Danio rerio
Xenopus laevis
Equus caballus
Gallus gallus
Bos taurus
Canus lupus familiaris
Pan troglodytes 

A

C

E

G

I

H

Homo sapiens
Mus musculus
Ra>us norvegicus
Danio rerio
Xenopus laevis
Equus caballus
Gallus gallus
Bos taurus
Canus lupus familiaris
Pan troglodytes 

Homo sapiens
Mus musculus
Ra>us norvegicus
Danio rerio
Xenopus laevis
Equus caballus
Gallus gallus
Bos taurus
Canus lupus familiaris
Pan troglodytes 

Figure 18: IdenZfying conserved shared interfaces between NHEJ factors. A) Cryo-EM structure 
of  the  Short-Range  NHEJ  dimer  (PDB:  7LSY)  highlighgng  an  interface  between  the  ligagon-
posigoned Lig4 molecule (green) and Ku70 (black). Lig4 residues R136 and K140 are highlighted 
in magenta and form the basic side of a salt bridge with Ku70. Also pictured Ku80 (black), … 

 71 

 
 
Figure 18 (cont’d)  
XRCC4 (Purple), XLF (pink), and the BRCT-structural domain of a second Lig4 molecule (gold). B) 
Species conservagon of Lig4 orthologs in 10 vertebrate species aligned to human Lig4 K134–L143. 
C)  Cryo-EM  structure  of  the  XLF-dependent  DNA-PK  long-range  NHEJ  dimer  highlighgng  an 
interacgon between Lig4 and Ku70/Ku80. Zoomed in diagram showing Lig4 N692, R708, N711 
(magenta)  interacgng  with  Ku80.  D)  Species  conservagon  of  Lig4  orthologs  in  10  vertebrate 
species aligned to human Lig4 K134–L143. E) Shared interface between Ku80 and Lig4-BRCT in 
the Ku80-mediated dimer with PAXX/XRCC4/L4 (PDB 8BHY, Lig4 N711 highlighted in magenta). F) 
Species conservagon of Ku70, with D192-D196 highlighted in red. G) Species conservagon of Ku80 
D327-Q330  ”DEEQ  mogf”  (red)  aligned  to  human  Ku80.  H)  Conservagon  of  DNA-PKcs  D1440 
(red). I) Conservagon of DNA-PKcs E1497 (red). 

DNA-interacZng residues in the L4 DBD work cooperaZvely with the Ku70 trans interface to 

join ends. A recent study by Loparo and colleagues revealed that L4 dynamically interacts with 

ends prior to SRC synapsis, mediated by 4 basic residues in its DBD that directly interact with 

DNA in cryo-EM structures. Mutagng these interacgons significantly impaired interacgons with 

DNA and disrupted stable SRC formagon. We hypothesized that combining the DNA-interface 

mutant with catalygcally inacgve L4 would impair its ability to promote end joining, similar to 

K273A + the RK>DD mutagon targegng the L4 DBD-Ku70 interface. To address this, we designed 

an expression vector mutagng the equivalent residues in human L4 (K28E, K30E, R32E, K162E, 

R163E, K164E, together termed mDBD) and tested the ability to join VDJ coding-end substrates 

(Figure 17F). In cells expressing catalygcally acgve L4, mDBD decreases joining by ~50%, 

indicagng that disrupgng DNA-binding significantly (but not completely) impacts the ability of 

NHEJ to promote joining. Combining mDBD with either RK>DD or L4 K273A completely ablates 

all joining. This suggests that 1) ligagon via L4 is structurally supported both by protein-DNA 

interacgons and protein-protein interacgons between L4’s DBD and Ku70’s VWA, and 2) that the 

alternagve end-joining sgmulated by catalygcally inacgve L4 is dependent on its structural 

interacgons that stabilize the SRC.  

DISCUSSION 

A  single  protein  interface  may  have  mulZple  funcZons,  potenZally  supporZng  flexibility  in 

complex formaZon or driving transiZons between disZnct complexes. Here we report funcgonal 

significance  of  a  L4  protein-protein  interacgon  recently  idengfied  in  a  new  form  of  the  Ku80 

mediated LRC. To address the funcgonal relevance of this interacgon, a mutagonal approach was 

 72 

 
 
uglized to ablate the interface by subsgtugng residues in both DNA-PKcs and L4. In contrast to 

many of our previous studies where mutagon of either side of a single interface generally results 

in similar cellular phenotypes, mutagons designed to ablate either the L4 or DNA-PKcs side of the 

interface had very different cellular impacts. Of note, ablagng the L4 surface impacted funcgon, 

and  this  impact  was  enhanced  when  L4’s  catalygc  acgvity  was  also  blocked.  Amer  examining 

published structures of NHEJ complexes, we observed that the L4 residues involved (R136/K140) 

also mediate a trans interacgon with Ku70 in the NHEJ short-range complex. This study suggests 

a model where the L4 interface that interacts with DNA-PKcs in the Ku80-mediated dimer may 

eventually transit (either directly or indirectly) into a more consequengal interacgon with Ku70 

that we posit contributes to tethering DNA ends prior to ligagon.  

Of note, we have observed a similar dual funcgon of another interface in DNA-PKcs. The 

four highly conserved lysine residues in the M-HEAT region of DNA-PKcs that facilitate both cis 

and  trans  interacgons  with  Ku80's  C-terminal  acidic  pepgde  (20),  also  interact  with  a  disgnct 

interface  in  the  recently  described  LR-ATP  complex  reported  by  He  and  colleagues  that  likely 

represents  a  transigonal  LR  complex  (54).  Finally,  we  have  recently  observed  via  cryo-EM 

experiments,  Ku-bound  to  chromagnized  DSBs  (unpublished  data,  Amanda  Chaplin).  In  this 

structure, we observe a specific interface in Ku80 that interacts with the H3 nucleosome subunit. 

However, this same interface in Ku80 also mediates its interacgon with the BRCT domains of L4 

observed  in  many  NHEJ  complexes  (both  long-  and  short-range).  One  explanagon  for  these 

observagons is that single interfaces that facilitate assembly of different complexes may dictate 

how transigons between NHEJ complexes proceed.  

Lig4 plays an essenZal structural role in tethering DNA ends prior to ligaZon.  Single molecule 

imaging studies from Loparo and colleagues demonstrate that interacgons of L4's DBD are crigcal 

for progression from long-range to short-range complexes (48). More specifically, crigcal DNA-

interacgng residues in the L4's DBD dramagcally reduced L4-DNA interacgons in the LR complex 

and subsequent SR complex formagon, indicagng that this noncatalygc interacgon between L4 

and  the  break  serves  as  an  essengal  step  preceding  SR  synapsis.  Their  data  suggests  a  model 

where L4 dynamically interacts with ends protected by DNA-PKcs, and L4's binding of the two 

ends promotes progression to the SR complex. Our recent report established that catalygcally 

 73 

 
 
inacgve L4 delays release of Ku, DNA-PKcs, and XRCC4 from chromagn amer DNA damage in living 

cells (73); from these data, we posit that the catalygc site of L4 also promotes progression to 

short-range complexes. The descripgon here of another L4 DBD interacgon (R136 and K140 in L4 

with  D192  in  Ku70)  that  funcgons  collaboragvely  with  the  catalygc  site  may  also  impact  L4's 

ability to promote progression to short-range complexes underscores the central role of L4 in this 

transigon.  These  data  prompted  an  examinagon  of  the  same  L4/DNA  end  interacgon  in  the 

cellular  assays  used  in  this  study.  Altogether,  these  studies  suggest  that  L4  may  serve  as  a 

molecular “sensor” in the long-range complex, promogng transigon into the ligagon-competent 

short-range complex only if both DNA ends are appropriate substrates for ligagon. This "sensor" 

funcgon is dependent not only on L4's interacgon with DNA ends, but also L4's interacgon with 

Ku70, and L4's catalygc site.   

L4  is  unique  among  eukaryogc  ligases  in  its  characterisgc  as  a  single-turnover  enzyme 

(51); thus, it is quite possible that the L4 molecule that seals one strand does not necessarily seal 

the second strand. One possibility is that rapid ligagon of one strand and NHEJ-complex mediated 

end-tethering act cooperagvely to support ends. Conversion of a DSB to a single strand break 

through one-step ligagon would dramagcally reduce the threat of the lesion and the conversion 

may  allow  NHEJ  factors  to  “make  way”  for  other  DNA  repair  pathways  to  resolve  the  second 

strand’s break.  In sum, these data underscore how the molecular dynamics of NHEJ complexes 

are highly dependent on L4 and its structural interacgons. 

InacZve L4 requires the Ku80-mediated dimer for funcZon. We have shown previously that the 

two forms of long-range synapgc complexes have disgnct funcgons and are in equilibrium. The 

observagon that cells expressing K273A L4 have prolonged DSB-induced chromagn associagon, 

suggested to us that long-range to long-range or long-range to short-range transigons might be 

impacted by loss of L4's catalygc acgvity. This prompted an examinagon of whether either of the 

two  LR  complexes  are  funcgonally  crigcal  in  cells  expressing  K273A  L4.  Our  data  demonstrate 

show that the Ku80-mediated dimer is essengal for K273A L4 to promote end-joining, and we 

suggest  that  K273A  L4  promotes  alternagve  end-joining  in  the  context  of  the  Ku80  mediated 

dimer.  

 74 

 
 
L4's essenZal non-catalyZc role that facilitates end-joining is mediated by disZnct interacZons 

in  NHEJ  short-range  complexes  and  involves  disZnct  alternaZve  joining  pathways.  NHEJ 

complexes in living cells expressing K273A L4 have significantly different characterisgcs than cells 

expressing wild type L4. NHEJ factors in complex with K273A L4 display prolonged associagon 

with  chromagn  in  response  to  damage,  suggesgng  that  although  catalygcally  inacgve  L4 

facilitates NHEJ complex formagon, progression of those complexes is delayed (73). 

Clearly,  L4’s  structural  interacgons  with  other  NHEJ  factors,  and  not  just  its  catalygc 

interacgons with DNA are crigcal for promogng end-joining. But how does catalygcally inacgve 

L4  promote  joining?  The  data  presented  here  and  previously  demonstrate  that  disgnct  end-

joining mechanisms are facilitated by the presence of inacgve L4. We have shown previously that 

K273A robustly facilitates L3 mediated joining in vitro; these data are strongly corroborated by 

our recent collaboragve study from Yu and colleagues demonstragng a synthegc lethal interacgon 

between catalygcally inacgve L4 and loss of nuclear L3 in mice (14). Here we show that unlike 

loss of L4, inacgvagon of L4 is not synthegcally lethal with Polq ablagon. Moreover, in K273A L4 

expressing  cells,  whereas  VDJ  joining  is  minimally  impacted  by  loss  of  Polq,  both  MMEJ  and 

resistance to DSB-inducing agents are strongly impacted by loss of Polq. From drug sensigvity and 

episomal joining assays, we show that alternagve end-joining sgmulated by catalygcally inacgve 

L4 is dependent on stable promogon of the SRC. Furthermore, since 293T L4K273A/- Polθ-/- and 

cells expressing L4 K273A + RK>DD are comparably sensigve to DSB inducing agents, we can infer 

that the high degree of SRC stability with catalygcally inacgve L4 reported by mulgple groups (73) 

is  responsible  for  promogng  alt-EJ  in  the  context  of  catalygcally  inacgve  L4.  An  emerging 

consensus in the end-joining field is that there are mulgple alternagve end-joining pathways. The 

data  presented  here  suggest  that  catalygcally  inacgve  L4  promotes  repair  by  several  of  these 

pathways, likely thorough its ability to promote sustained end-processing in the SRC. 

Conclusions: All together, these studies further bolster a perhaps incontrovergble conclusion that 

L4 has mulgple disgnct structural roles in NHEJ.  What is lacking is an understanding of 1) the 

molecular  mechanism  of  transigon  between  the  two  forms  of  long-range  complexes,  2)  the 

transigon from long- to short-range complexes and 3) the stoichiometry of DNA ligases in the 

short-range complex—pargcularly for difficult ends that are not rapidly ligated.  

 75 

 
MATERIALS AND METHODS 

Cell culture and genome ediZng 293T and U2OS cells were cultured in Dulbecco’s Modified 

Eagle Medium (DMEM, Life Technologies) and Roswell Park Memorial Insgtute 1640 Medium 

(RPMI), respecgvely, supplemented with 10% fetal bovine serum (Atlanta Biologicals, GA), 2 mM 

L-glutamine, 0.1 mM non-essengal amino acids, 1 mM sodium pyruvate, 100 U/ml penicillin, 

100 g/ml streptomycin and 25 ng/ml Gibco Amphotericin B (Life Technologies).  

TransfecZons of 293T and U2OS cell lines were performed using unsupplemented RPMI media, 

polyethyleneimine (PEI, 1 µg/mL Polysciences), and plasmid DNA at a rago of 100 µL RPMI:1 µg 

DNA:2 µg PEI. First, plasmid DNA and RPMI were well-mixed in a 1.5 mL microcentrifuge tube 

followed by addigon of PEI. Transfecgon mixes were shaken briefly to thoroughly mix contents, 

then incubated at room-temperature for up to 30 minutes to allow PEI:DNA complexes to form 

before adding the reacgon volume to freshly plated cells 

Genome ediZng 

PolQ delegon in 293T cell strains was performed via CRISPR/Cas9 by targegng cells with a pair 

of gRNAs spaced 2.1 kb apart in the polq locus to create a large delegon. Each gRNA was cloned 

into a pCas-2A-Puro (Addgene 62988), then 1 µg of each plasmid was transfected into a 293T L4-

/- and L4K273A/- cell strains. Clonal populagons were grown out for one week and screened via 

outside-outside PCR (to idengfy clones with large delegons at the polq locus). Candidates were 

then checked for presence of remaining WT allele.  

 pPB stable L4 complementagon: Transfected 1.0E6 cells with 1.5 µg pPB-Puro-Lig4 payload and 

1.5 µg pBase transposase plasmids. Cells were allowed to recover for 48 hours before selecgon 

with 1-2 µg/mL puromycin. Bulk populagons were cultured in the presence of puromycin for the 

duragon of the experiments.  

Cell Survival assays: U2OS L4-/- cell strains complemented with PiggyBacc L4 mutants were 

assessed for resistance to DSB inducing drugs using colony formagon assays. Briefly, 200-300 

cells of each mutant strain were plated into 6 cm gssue culture dishes with increasing 

concentragons of drug and allowed to grow for 9 days. Plates were stained with 2% crystal 

violet solugon (VWR) and counted. A minimum of 3 Replicates was performed for each strain at 

each concentragon.  293T+L4 mutant cell strains were assessed using MTT-metabolic assays as a 

 76 

 
marker for cell viability. 1.6E4 cells were plated in wells of a 24-well plate at increasing 

concentragons of calicheamicin in 1 mL of DMEM media. Amer 4 days of drug treatment, cells 

were treated with 1 mg/mL MTT (Goldbio) for 1 hr at standard culture condigons. MTT was 

removed and cells were dissolved in 200 µL 100% DMSO and absorbance was sampled on a 

plate reader at 570 nm wavelength. A well without cells treated with media + MTT was used as 

a background control subtracted from all absorbance values.  

Episomal Joining Assays: 293T L4-/- cells were transfected following our PEI protocol with a 

mixture of previously reported episomal joining substrates (10, 11, 88, 89), endonucleases, and 

plasmid expression vectors encoding L4 mutants in a 1:1:1 rago (by plasmid mass), then 

cultured for 72 hours before paraformaldehyde fixagon before running on an Ajune CytPix flow 

cytometer. MMEJ assays were similarly transfected into 293T cells with 1.0 µg linearized DNA 

substrate and 0.1 µg pECFP transfecgon marker, with 2.0 µg PEI, 100 µL RPMI and fluorescence 

analyzed 72 hours later gagng for CFP+ transfected cells, with the subpopulagon of dsRed+ cells 

represengng MMEJ+ outcomes.  

ACKNOWLEDGMENTS 

We would like to thank Daniel Vocelle and the MSU Flow Cytometry core for operagng and 

maintaining the Ajune CytPix, supported by the Equipment Grants Program, award #2022-

70410-38419, from the U.S. Department of Agriculture (USDA), Nagonal Insgtute of Food and 

Agriculture (NIFA). 

 77 

 
 
 
 
 
1.  

2.  

3.  

4. 

5.  

6.  

7.  

8.  

9.  

10.  

11.  

12.  

13.  

REFERENCES 

Waters,C.A., 
Nonhomologous end joining: A good solugon for bad ends. DNA Repair, 17, 39–51. 

Strande,N.T.,  Wyaj,D.W.,  Pryor,J.M. 

and  Ramsden,D.A. 

(2014) 

Loparo,J.J.  (2023)  Holding  it  together:  DNA  end  synapsis  during  non-homologous  end 
joining. DNA Repair, 130, 103553. 

Watanabe,G.  and  Lieber,M.R.  (2023)  The  flexible  and  iteragve  steps  within  the  NHEJ 
pathway. 
Biology, 
Biophysics 
10.1016/j.pbiomolbio.2023.05.001. 

Molecular 

Progress 

and 

in 

Amin,H., Zahid,S., Hall,C. and Chaplin,A.K. (2024) Cold snapshots of DNA repair: Cryo-EM 
structures  of  DNA-PKcs  and  NHEJ  machinery.  Progress  in  Biophysics  and  Molecular 
Biology, 186, 1–13. 

Goff,N.J.,  Mikhova,M.,  Schmidt,J.C.  and  Meek,K.  (2024)  DNA-PK:  A  synopsis  beyond 
synapsis. DNA Repair, 141, 103716. 

Rothkamm,K., Kruger,I., Thompson,L.H. and Lobrich,M. (2003) Pathways of DNA double-
strand break repair during the mammalian cell cycle. Mol Cell Biol, 23, 5706–5715. 

Lees-Miller,J.P., Cobban,A., Katsonis,P., Bacolla,A., Tsutakawa,S.E., Hammel,M., Meek,K., 
Anderson,D.W.,  Lichtarge,O.,  Tainer,J.A.,  et  al.  (2021)  Uncovering  DNA-PKcs  ancient 
phylogeny, unique sequence mogfs and insights for human disease. Progress in Biophysics 
and Molecular Biology, 163, 87–108. 

Sgnson,B.M.,  Moreno,A.T.,  Walter,J.C.  and  Loparo,J.J.  (2020)  A  Mechanism  to  Minimize 
Errors during Non-homologous End Joining. Molecular Cell, 77, 1080-1091.e8. 

Deshpande,R.A.,  Myler,L.R.,  Soniat,M.M.,  Makharashvili,N.,  Lee,L.,  Lees-Miller,S.P., 
Finkelstein,I.J.  and  Paull,T.T.  (2020)  DNA-dependent  protein  kinase  promotes  DNA  end 
processing by MRN and CtIP. Sci Adv, 6, 0922. 

Deshpande,R.A., Marin-Gonzalez,A., Barnes,H.K., Woolley,P.R., Ha,T. and Paull,T.T. (2023) 
Genome-wide analysis of DNA-PK-bound MRN cleavage products supports a sequengal 
model of DSB repair pathway choice. Nature communica9ons, 14. 

Chang,H. and Lieber,M. (2016) Structure-Specific nuclease acgviges of Artemis and the 
Artemis: DNA-PKcs complex. NUCLEIC ACIDS RESEARCH, 44, 4991–4997. 

Graham,T.G.W., Walter,J.C. and Loparo,J.J. (2016) Two-Stage Synapsis of DNA Ends during 
Non-homologous End Joining. Molecular Cell, 61, 850–858. 

Chen,S.,  Lee,L.,  Naila,T.,  Fishbain,S.,  Wang,A.,  Tomkinson,A.E.,  Lees-Miller,S.P.  and  He,Y. 
(2021) Structural basis of long-range to short-range synapgc transigon in NHEJ. Nature, 
10.1038/s41586-021-03458-7. 

 78 

 
14.  

15.  

16.  

17.  

Chaplin,A.K.,  Hardwick,S.W.,  Liang,S.,  Kefala  Stavridi,A.,  Hnizda,A.,  Cooper,L.R.,  De 
Oliveira,T.M., Chirgadze,D.Y. and Blundell,T.L. (2021) Dimers of DNA-PK create a stage for 
DNA double-strand break repair. Nat Struct Mol Biol, 28, 13–19. 

Seif-El-Dahan,M.,  Kefala-Stavridi,A.,  Frit,P.,  Hardwick,S.W.,  Chirgadze,D.Y.,  Maia  De 
Oliviera,T.,  Brijon,S.,  Barboule,N.,  Bossaert,M.,  Pandurangan,A.P.,  et  al.  (2023)  PAXX 
binding  to  the  NHEJ  machinery  explains  funcgonal  redundancy  with  XLF.  Science 
Advances, 9, eadg2834. 

Buehl,C.J.,  Goff,N.J.,  Hardwick,S.W.,  Gellert,M.,  Blundell,T.L.,  Yang,W.,  Chaplin,A.K.  and 
Meek,K. (2023) Two disgnct long-range synapgc complexes promote different aspects of 
end processing prior to repair of DNA breaks by non-homologous end joining. Molecular 
Cell, 10.1016/j.molcel.2023.01.012. 

Goff,N.J., Brenière,M., Buehl,C.J., de Melo,A.J., Huskova,H., Ochi,T., Blundell,T.L., Mao,W., 
Yu,K., Modesg,M., et al. (2022) Catalygcally inacgve DNA ligase IV promotes DNA repair 
in living cells. Nucleic Acids Research, 10.1093/nar/gkac913. 

18.   Mikhova,M.,  Goff,N.J.,  Janovič,T.,  Heyza,J.R.,  Meek,K.  and  Schmidt,J.C.  (2024)  Single-
molecule  imaging  reveals  the  kinegcs  of  non-homologous  end-joining  in  living  cells. 
10.1101/2023.06.22.546088. 

19.  

20.  

21.  

Sallmyr,A., Rashid,I., Bhandari,S.K., Naila,T. and Tomkinson,A.E. (2020) Human DNA ligases 
in replicagon and repair. DNA Repair, 93, 102908. 

Grawunder,U., Zimmer,D. and Lieber,M.R. (1998) DNA ligase IV binds to XRCC4 via a mogf 
located between rather than within its BRCT domains. Current Biology, 8, 873–879. 

Chaplin,A.K.,  Hardwick,S.W.,  Stavridi,A.K.,  Buehl,C.J.,  Goff,N.J.,  Ropars,V.,  Liang,S., 
Oliveira,T.M.D.,  Chirgadze,D.Y.,  Meek,K.,  et  al.  (2021)  Cryo-EM  of  NHEJ  supercomplexes 
provides insights into DNA repair. Molecular Cell, 0. 

22.  

Sgnson,B.M., Carney,S.M., Walter,J.C. and Loparo,J.J. (2024) Structural role for DNA Ligase 
IV in promogng the fidelity of non-homologous end joining. Nat Commun, 15, 1250. 

23.   Malashejy,V., Au,A., Chavez,J., Hanna,M., Chu,J., Penna,J. and Cortes,P. (2023) The DNA 
binding  domain  and  the  C-terminal  region  of  DNA  Ligase  IV  specify  its  role  in  V(D)J 
recombinagon. PLOS ONE, 18, e0282236. 

24.   Medina-Suarez,D.,  Han,L.,  O’Reily,S.,  Liu,J.,  Wei,C.,  Brenière,M.,  Goff,N.J.,  Chen,C., 
Modesg,M., Meek,K., et al. (2024) DNA ligases 3 and 4 cooperate in vivo to facilitate DNA 
repair and organism viability. 

25.  

Chan,S.H.,  Yu,A.M.  and  McVey,M.  (2010)  Dual  Roles  for  DNA  Polymerase  Theta  in 
Alternagve End-Joining Repair of Double-Strand Breaks in Drosophila. PLOS Gene9cs, 6, 
e1001005. 

 79 

 
26.   Wyaj,D.W.,  Feng,W.,  Conlin,M.P.,  Yousefzadeh,M.J.,  Roberts,S.A.,  Mieczkowski,P., 
Wood,R.D.,  Gupta,G.P.  and  Ramsden,D.A.  (2016)  Essengal  Roles  for  Polymerase  theta-
Mediated End Joining in the Repair of Chromosome Breaks. Mol. Cell, 63, 662–673. 

27.  

28.  

29.  

Yousefzadeh,M.J.,  Wyaj,D.W.,  Takata,K.,  Mu,Y.,  Hensley,S.C.,  Tomida,J.,  Bylund,G.O., 
Doublié,S.,  Johansson,E.,  Ramsden,D.A.,  et  al.  (2014)  Mechanism  of  Suppression  of 
Chromosomal Instability by DNA Polymerase POLQ. PLOS Gene9cs, 10, e1004654. 

Chen,S., Vogt,A., Lee,L., Naila,T., McKeown,R., Tomkinson,A.E., Lees-Miller,S.P. and He,Y. 
(2023)  Cryo-EM  visualizagon  of  DNA-PKcs  structural  intermediates  in  NHEJ.  Science 
Advances, 9, eadg2838. 

Chen,X., Ballin,J.D., Della-Maria,J., Tsai,M.-S., White,E.J., Tomkinson,A.E. and Wilson,G.M. 
(2009) Disgnct kinegcs of human DNA ligases I, IIIα, IIIβ, and IV reveal direct DNA sensing 
ability and differengal physiological funcgons in DNA repair. DNA Repair, 8, 961–968. 

30.   Meek,K.  (2020)  Acgvagon  of  DNA-PK  by  hairpinned  DNA  ends  reveals  a  stepwise 

mechanism of kinase acgvagon. Nucleic Acids Research, 48, 9098–9108. 

31.   Neal,J.A., Xu,Y., Abe,M., Hendrickson,E. and Meek,K. (2016) Restoragon of ATM Expression 

in DNA-PKcs–Deficient Cells Inhibits Signal End Joining. J.I., 196, 3032–3042. 

32.  

Deriano,L., Stracker,T., Baker,A., Petrini,J. and Roth,D. (2009) Roles for NBS1 in Alternagve 
Nonhomologous  End-Joining  of  V(D)J  Recombinagon  Intermediates.  MOLECULAR  CELL, 
34, 13–25. 

 80 

 
 
 
 
 
 
 
 
 
 
Chapter 4: Discussing the two-stage synapZc model of NHEJ.  

By 

Noah J. Goff 

Conclusions and Future Direcgons 

 81 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
INTRODUCTION 

My goal with this chapter is to place each of the previous chapters more firmly in the 

context of the evolving synapgc model of NHEJ, highlight some outstanding quesgons in the 

field, and to propose experiments involving DNA Ligase IV to address those quesgons.  The 

overarching themes of my dissertagon research center around how DSB repair pathways—in 

pargcular end joining pathways—use structural features to promote repair while minimizing 

mutagons. Of course, we are not the only group acgvely working on this problem, and I’ve 

highlighted several notable studies released over the course of my PhD studies that contribute 

to our most recent model of NHEJ and how it interacts with alternagve end joining pathways.  

DISCUSSION 

Chapter 1: Evolving models of non-homologous end joining.  The core of this chapter is a 

review of recent NHEJ literature that’s contributed to an updated model of end synapsis and 

processing. This work spans early single molecule imaging (SMI) studies performed by Loparo 

and colleagues (21, 46), early Cryo-EM work from Tom Blundell, Yuan He, and Amanda Chaplin 

(18, 19), and an introducgon to the role of DNA Ligase IV (L4).  I’ve also reproduced secgons of a 

review that I co-wrote with another graduate student, Mariia Mikhova, and our advisors Kathy 

Meek and Jens Schmidt. The engre review is quite large (too large to reproduce in total for this 

work) and focuses on the role of DNA-PKcs in the new synapgc models of NHEJ.  The secgons 

included were both primarily dramed by me and highly pergnent to understanding evolving 

models of NHEJ—a core throughline of both my dissertagon work and other contribugons 

during my PhD. There have been several other reviews from leading groups focused on other 

aspects of recent end joining research—including excellent reviews of recent Cryo-EM data (52, 

53), and end-synapsis (80).  

Chapter 2: CatalyZcally inacZve DNA ligase IV promotes DNA repair in living cells. This project 

began as a straighiorward CRISPR-Cas9 gene edigng experiment coupled with an easy set of 

lab-standard reporter assays to onboard a first-year graduate student. We never expected to be 

the core project of my PhD. The inigal 293T mutagenesis work in early spring of 2020, 

immediately before the onset of the COVID-19 pandemic. The central research quesgon—does 

catalygcally inacgve L4 promote end joining in cell cultures—was inspired by papers from 

 82 

 
Loparo and colleagues proposing a 2-stage synapgc model of NHEJ, where DNA ends were 

“protected” in a long range complex prior to recruitment of downstream factors (21, 46). 

Importantly, the authors discovered that L4’s catalygc funcgon was dispensable for this result in 

vitro, indicagng L4 held important noncatalygc funcgon that could support otherwise highly 

regulated end processing. While other studies had suggested a non-catalygc role for L4 both in 

vitro and in yeast cells, it was unclear whether DSB repair in more complex eukaryogc cells 

would support any NHEJ amer loss of L4’s core catalygc funcgon. Furthermore, the studies only 

tested the role of L4 in promogng processing mediated by DNA polymerase lambda and the 

phosphodiesterase TDP1.  

In  addigon  to  studying  NHEJ,  our  lab  also  has  research  projects  into  immunoglobulin 

development through VDJ recombinagon, which produces NHEJ-targeted DNA hairpin structures 

as  recombinagon  intermediates.    Given  the  addigonal  targegng  of  hairpins  to  NHEJ,  we  were 

quite surprised when the inigal 293T joining assays revealed robust hairpin joining in cells lacking 

L4 catalygc acgvity. Notably, the inigal in vitro work from Loparo and colleagues did not test DNA 

hairpin substrates in the biochemical xenopus with catalygcally inacgve L4. While revising the 

paper, we found a 2013 paper from Wilson and colleagues (90) that provided early evidence for 

catalygcally  inacgve  L4  supporgng  end  joining  in  yeast.  We  were  completely  unaware  of  this 

paper  at  the  gme,  but  it  suggests  a  remarkable  conservagon  of  NHEJ  phenotypes  across 

eukaryotes—even those lacking DNA-PKcs.   

The primary takeaways from my first paper were: 1) that L4’s structural presence does 

serve  as  a  regulatory  step  in  canonical  NHEJ,  2)  loss  of  L4  catalygc  acgvity  in  the  shortrange 

complex  decreases  the  fidelity  of  end  joining,  and  3)  that  the  L3-XRCC1  complex  acts 

collaboragvely with the core NHEJ factors in vitro to sgmulate end-joining. To date, the role of a 

L3-NHEJ connecgon serves as a major extant quesgon in the two-stage iteragve model of NHEJ. 

An in-review (at gme of wrigng) paper from Dr. Kefei Yu’s laboratory established a catalygcally 

inacgve  L4K273S/K273S  and  L3nuc-/-  mouse  lines,  with  each  being  viable  through  birth  and 

adolescence.  Notably,  L3nuc-/-  L4K273S/K273S  double  mutant  offspring  are  nonviable,  providing 

addigonal  in  vivo  evidence  that  there  is  compensagon  from  L3  following  loss  of  L4’s  catalygc 

acgvity.   

 83 

 
Chapter  3:  New  insight  into  how  the  DNA  binding  domain  of  DNA  Ligase  IV  facilitates  end 

joining independent of its catalyZc acZvity. This project originated both as a direct follow-up to 

my first publicagon, but also a structure-directed invesggagon of new data from Amanda Chaplin 

and Steve Hardwick’s group at the University of Leicester and Cambridge University to invesggate 

interfaces  discovered  in  new  Cryo-EM  data.  It  is  my  current  area  of  acgve  research.  We  have 

shown that disrupgng end-tethering substangally impairs L4 K273A-mediated end joining, with 

only modest impacts on joining in the presence of catalygcally acgve L4. I propose that tethering 

ends in the SRC occurs through four mechanisms: 1) the XRCC4:XLF:XRCC4 bridge linking each 

Ku70/80 heterodimer, 2) the trans L4 DBD:Ku70 interacgon (disrupted by the RK>DD mutagon) 

that anchors one end of the complex with the ligagon posigoned break, 3) the L4-DNA interacgon 

recently reported by Loparo and colleagues (48), and 4) rapid ligagon of one strand early following 

transigon  into  the  shortrange  complex  to  covalently  link  DNA  ends.  As  seen  in  chapter  3, 

disrupgon of any two of these interfaces appears to be sufficient to ablate end joining, either 

mediated  by  L4  catalygc  acgvity  or  through  promogon  of  alt-EJ  pathways  (as  is  the  case  with 

catalygcally inacgve L4).   

In addigon to invesggagng interfaces that could mediate L4’s structural role in NHEJ, we 

also sought to invesggate backup repair factors (in addigon to L3-mediated joining) that promote 

end joining amer failed NHEJ. Based on my observagon that cells expressing catalygcally inacgve 

L4  overuglizes  microhomology  in  L4  K273A  cells  (Chapter  2),  we  hypothesized  that  DNA-

Polymerase Theta (Polθ) may play an outsized role in supporgng NHEJ in cases where the ligagon 

complex fails. To test, I used a CRISPR/Cas9 genome edigng approach to knock out Polθ from my 

293T  L4K273A*/-  cells,  then  tested  their  response  to  a  MMEJ  specific  reporter  substrate,  NHEJ-

specific VDJ Coding joint substrate, and sensigvity to DSB inducing calicheamicin and teniposide. 

Of note, I was unable to isolate any 293T L4-/- Polθ-/- double knockout cell strains despite mulgple 

gRNA transfecgons. Based on previous experiments in the lab as well as other reports (77, 91, 

92), we believe that Polθ serves as the primary alternagve end joining factor following loss of 

NHEJ and as such double knockouts are nonviable—or nearly nonviable.   

This hypothesis is supported by my experiments indicagng that 293T L4K273A/- Polθ-/- cells 

are hypersensigve to DSB inducing agents calicheamicin and teniposide (Chapter 3). Anecdotally, 

 84 

 
these  cell  cultures  also  divide  much  slower,  comparable  to  L4-/-  cultured  under  the  same 

condigons. There is neither a growth disparity nor DSB drug sensigvity observed with loss of Polθ 

on its own. Despite the hypersensigvity to genomic DSBs, addigonal loss of Polθ appears to have 

no impact on VDJ coding, and only a modest deficiency in joining TelN and I-SceI substrates. A 

combinagon of two factors could explain this: 1) signal from episomal joining assays is measured 

as a binary on/off per cell and is not necessarily a fair measure of total break repair capacity for 

a given substrate so long as minimal joining is performed; 2) there may be an alternate repair 

mechanism that can join simple overhangs (like those generated by I-SceI and Artemis-opened 

hairpins) but not more complex adducts caused by calicheamicin and teniposide. I suggest that 

L3 may serve in this role based on Chapter 2 and the mouse models developed by Dr. Kefei Yu’s 

laboratory.   

AddiZonal contribuZons to work that haven’t been reproduced for my dissertaZon:  In addigon 

to the work presented in chapters 2 and 3, I have made several substangal contribugons to other 

projects that have resulted in authorship, both published and in press. I’ve chosen to focus my 

dissertagon  on  my  Ligase  IV  projects,  and  as  such  I’ve  not  reproduced  excerpts  from  these 

manuscripts (a full list of my contribugons to papers can be found in Appendix 1).  

Ku80 Linker Mutants: First, I’d like to highlight my contribugons to tesgng the Ku80 C-terminal 

domain mutagons published in Buehl et al. 2023 (20). This project began shortly amer returning 

from COVID-19 lockdowns, with our lab working collaboragvely with several structural biologists 

to use mutagonal strategies to test the funcgonal relevance of synapgc NHEJ complex 

structures recently idengfied using Cryo-EM. These structures revealed a trans interacgon 

between Ku80’s acidic extreme C-terminal domain (CTD) and a basic patch on DNA-PKcs, 

mediated by an unstructured flexible linker in Ku80 that is highly conserved in length (but not 

primary sequence). The Ku80 CTD had previously been idengfied as a mediator of end joining 

(5–7), and mutagng the DNA-PKcs residues that mediate this interacgon (termed 4A) revealed a 

significant defect in nucleolygc end processing. We hypothesized that the highly conserved 

length of the flexible linker (93) observed in the Cryo-EM structures may be funcgonally 

relevant in promogng the trans interacgon. It was observed by Dr. Wei Yang (NIH) that the Ku80 

CTD could also interact with DNA-PKcs in cis if the linker stretches to its maximum length, and 

 85 

 
she suggested that adding or removing residues may differengally impact funcgon by biasing 

formagon of the cis or trans interacgons.  We observed that lengthening the linker by 6-15 

residues substangally reduced the amount of joining (>2 fold).  Interesgngly, shortening the 

linker 3 or 6 residues significantly increased NHEJ funcgon both in episomal joining assays and 

in resistance to calicheamicin and etoposide (a Top2 poison). Further shortening the linker by 

10, 13, or 15 residues progressively impaired NHEJ, with the shortest linker length mimicking 

complete delegon of the CTD. We concluded that the hypothesized cis interacgon may have an 

inhibitory effect on synapsis, and there is a minimum linker length required to promote synapsis 

and NHEJ funcgon.  It’s currently unclear why the flexible linker is conserved at its current 

length (as opposed to 3-6 AA shorter), it may be due to lower repair fidelity, or some other 

deficiency not idengfied in our tesgng.   

Single Molecule Imaging of NHEJ Factors: The second contribugon I would like to highlight is to 

Mariia Mikhova’s fantasgc study into the nuclear dynamics of NHEJ factors using live-cell Single 

Molecule Imaging (73), in pargcular my work generagng L4 knockouts in their U2OS cell strains 

with halo-tagged NHEJ factors. These cells were then used to assess changes in Ku70, DNA-PKcs, 

XLF, and XRCC4 dynamics with complementagon of L4 expression vectors for the revisions of 

her (in press) paper. Notable findings include: 1) XRCC4 is primarily localized to the nucleus but 

does not interact with chromagn in the absence of L4, 2) There appears to be a mechanism to 

acgvely remove Ku and DNA-PKcs if L4/XRCC4 recruitment fails, and 3) in the presence of 

catalygcally inacgve L4 all four tagged NHEJ factors have a greatly elongated window (compared 

to WT L4) where they interact with chromagn—including DNA-PKcs. This final point is 

pargcularly surprising given that DNA-PKcs departs chromagn significantly earlier than other 

factors. We originally hypothesized that DNA-PKcs departure corresponded with transigon into 

the short-range complex—matching the single Cryo-EM structure reported by Yuan He and 

colleagues (18), but the in vitro work indicates that there is no defect in short-range complex 

formagon or in end processing in the presence of catalygcally inacgve L4 (21, 46). In my 

opinion, it’s likely that the Cryo-EM short-range complex structure is not necessarily the only 

possible conformagon of NHEJ factors in the experimentally observed short-range complex. To 

wit, I believe that it may be possible for DNA-PKcs to remain present in some short-range 

 86 

 
complexes, but it notably does depart the break earlier than other factors— possibly amer single 

strand ligagon to allow more room for processing the second strand.  

Ongoing argument over the role of DNA-PKcs and an alternate model of synapsis: DNA-PKcs 

had previously thought to primarily be dominant in NHEJ in vertebrates— pargcularly in those 

that predominantly use VDJ recombinagon for immunoglobulin diversificagon. Central to this 

hypothesis was the absence of DNA-PKcs from several model organisms: most notably budding 

yeast (S. cerevisiae), C. elegans, and fruit flies (D. melanogaster). A study from Lees-Miller and 

colleagues searched genomic sequencing data across a wide range of species for potengal DNA-

PKcs orthologs looking to build a deeper understanding of how NHEJ evolved (35). Remarkably, 

the authors discovered candidate DNA-PKcs orthologs in genomes sequenced from a broad 

range of plant, fungi, and animal species—with high conservagon of some well-studied features 

(including the ABCDE phosphorylagon patch). This startling result suggests that DNA-PKcs is not 

a unique regulator of NHEJ that arose due to programmed recombinagon—but rather a more 

universal regulator of core end joining across nucleated life-forms.   

There is an alternate model for NHEJ synapsis and ligagon—primarily championed by Eli 

Rothenberg (NYU) and Michael Lieber (USC). They conclude—through their imaging approaches 

using purified NHEJ proteins—that DNA-PKcs is dispensable for most NHEJ funcgon and tethering 

may occur through XRCC4-XLF filaments coagng intact secgons of DNA (they observe filaments 

of purified LX4 complexes coagng dsDNA near a DNA end with super resolugon microscopy). In 

their model, Ku70/80 recognizes one end, promotes filamentagon upstream of the break, then 

Ku binds to a trans filament (bound to the opposite DNA end) to inigate synapsis, then proceeds 

to bring the ends into a short-range synapse (measured as direct FRET+ signal in their assays).   

Notably, addigon of DNA-PKcs reduces the number of FRET+ foci (decreasing their “pairing 

efficiency”  quangficagon)  but  does  substangally  increase  the  size  of  each  focus.  The  authors 

claim that the presence of DNA-PKcs only results in pairing of “clusters” of DSBs, and that their 

relagvely  low  measurement  of  “normalized  pairing  efficiency”  points  to  DNA-PKcs  being 

dispensable for “simple” DSBs (60). I propose an alternate hypothesis where DNA-PKcs can act to 

synapse the DNA-stem loop structures used to cap their fluorescent oligonucleogdes (the authors 

confirmed that purified Ku70/80 + LX4 was unable to synapse these hairpin loops, but never ran 

 87 

 
the control in the presence of DNA-PKcs). The stem-loop structures closely resemble hairpin DNA 

intermediates present in VDJ recombinagon, and the resulgng aggregate of tagged DNA donors 

would perfectly match the larger foci seen in all reacgons included with DNA-PKcs. Furthermore, 

data from Mariia Mikhova’s paper indicate that there is likely no role for XRCC4-XLF filaments 

given XRCC4 does not associate with breaks in the absence of L4 (73).  

In  my  opinion,  most  of  their  reasoning  regarding  the  role  of  DNA-PKcs  is  circular;  the 

authors cite the data from Reid et al. 2015 (60) as jusgficagon to not test if DNA-PKcs promotes 

synapsis  in  subsequent  papers  (42,  58,  59).  They  then  use  their  engre  body  of  work  to  argue 

against  DNA-PKcs’s  involvement  in  most  NHEJ  events.  Their  experimental  design  is  unable  to 

independently  visualize  localizagon  of  the  acceptor  fluorophore  in  the  absence  of  a  donor  in 

proximity for FRET (less than 100 Å with their fluorophores), therefore making it impossible to 

clearly  observe  long-range  synapses.  This  approach  is  less  powerful  than  the  Loparo  group’s 

studies and misses synapsis consistent with the LRC (where ends are not held close enough for 

the  FRET+  signal  Rothenberg  and  colleagues  uglized  as  their  metric  for  synapsis).    In  their 

measurements, they observe that addigon of purified Ku70/80, XLF, and L4-XRCC4 is sufficient to 

synapse ends using a “Normalized Frequency” of high-FRET foci. The authors argue for a model 

of NHEJ resolving “simple DSBs” without presence or acgvity of DNA-PKcs. Instead, they argue 

synapsis occurs through long XLF-XRCC4-L4 filaments that they observe “coagng” DNA.  

Studies on the Zming of NHEJ are a major component of the new two-stage synapZc model: A 

recent  report  from  Loparo  and  colleagues  addressed  the  dynamics  of  L4  at  the  instant  of 

transigon  into  the  short-range  complex.  They  developed  a  three-color  imaging  system  to 

simultaneously DNA end synapsis (consistent with the SRC) as well as DNA-protein interacgons. 

Their data suggest that two molecules of L4 are recruited to a DSB prior to complex formagon 

and both dynamically interact with DNA ends. Prior to short-range synapsis, one LX4 complex 

vacates  the  LRC  seconds  before  they  observe  DNA-DNA  interacgons  consistent  with  SRC 

formagon (48). Unfortunately, the authors only tested substrates generated with blunt ligagon-

compagble  DNA  ends  that  require  no  addigonal  end-processing.  I  believe  it’s  likely  that  one 

molecule of L4 may act to catalyze single strand break repair—pargcularly given that L4 is pre-

adenylated  in  a  catalygcally  charged  state  it  seems  likely  that  this  reacgon  may  progress  very 

 88 

 
quickly. Tesgng a wider range of end joining substrates in the three-color smFRET experiment—

pargcularly  a  substrate  that  requires  short-range  complex  targeted  end  processing—would  be 

highly informagve.   

Addigonal open quesgons remain regarding the role of DNA-PKcs in supporgng the short-

range complex. The inigal report of a short-range complex observed without DNA-PKcs by Cryo-

EM (18), coupled with preliminary data showing DNA-PKcs leaving chromagn earlier than other 

NHEJ factors in a SMI study (16) suggested to us that DNA-PKcs dissociated from the break site 

amer inigal synapsis and end processing. One striking result: DNA-PKcs remains bound to breaks 

for significantly longer with catalygcally inacgve L4. It’s unclear whether DNA-PKcs’s retengon at 

breaks is mediated through a prolonged long range complex (something unsupported by the in 

vitro smFRET data) (21, 46) or if there is a conformagon of the short-range complex that retains 

DNA-PKcs. I suggest the following model to explain the discrepancy in data: DNA-PKcs remains to 

stabilize the short-range complex ungl L4 can catalyze repair of one end. Finalizing catalysis allows 

for L4 and DNA-PKcs to dissociate from the break, leaving behind a significantly less hazardous 

break  that  can  be  repaired  by  the  remaining  NHEJ  machinery  (XRCC4,  XLF,  and  Ku  all  remain 

bound  for  significantly  longer  than  DNA-PKcs).  This  turnover  of  DNA-PKcs  would  allow  it  to 

“supervise” repair, then either recycle to bind another DSB or allow more room for extensive end 

processing on the remaining SSB (pargcularly in a case following ligagon opposite another DNA 

lesion: gap, mismatch, aberrant base, etc.).   

 89 

 
Figure 19: NHEJ progresses through a two-stage synapZc mechanism. Normal NHEJ progresses 
from  break  recognigon  (DNA-PK  monomers)  to  LRC  formagon.  Depending  on  end  chemistry, 
either  the  end-protecgon  XLF  mediated  dimer  (for  ends  with  small  adducts/overhangs  ends, 
middle  lem)  or  the  Ku80-mediated  dimer  (ends  with  bulky  adducts,  middle  right).  The  NHEJ 
synapse transigons into a SRC following recruitment of LX4 and XLF (and possibly formagon of 
the XLF-mediated dimer). Loss of L4 catalygc acgvity promotes long-term stable SRC synapsis, 
leading to excessive end processing followed by Polθ/L3 mediated end joining (bojom right).  

 90 

 
 
Figure 19 (cont’d) 
Notably, loss of tethering acgvity (either of the mDBD or RK>DD mutagons discussed in chapter 
3)  in  the  SRC  in  combinagon  with  losing  L4  catalygc  acgvity  results  in  no  L4-mediated  alt-EJ. 
Furthermore, catalygcally acgve L4 combining DNA end binding mutagons (mDBD) with the L4-
Ku70 interface mutant (RK>DD) ablates nearly all NHEJ.  

Areas of future L4 research: While I believe the work in this dissertagon outlines an important 

contribugon to our understanding of NHEJ, there are several open quesgons regarding L4’s role 

in end joining that I would like to see addressed in the future. First, addigonal work uglizing many 

of the newly idengfied mutants in Chapter 3 could explain why targegng the shared interfaces 

impact  NHEJ  so  severely.  I  propose  performing  a  suite  of  single-molecule  imaging  studies— 

idengcal  to  the  calicheamicin  “gme  course”  studies  using  transiently  complemented  L4  that  I 

contributed to in Mariia Mikhova’s study (73)—pargcularly to study nuclear dynamics of XRCC4 

and DNA-PKcs in the short-range complex. I hypothesize that disrupgng residues important for 

stable long-term synapsis should shorten the gme that the XRCC4-L4 complex remains bound to 

breaks, resulgng in a shortened window for chromagn associagon. Conversely, the gme-window 

for DNA-PKcs chromagn associagon should increase as failed repair events require rebinding and 

synapsis before beginning the pathway again.   

To support this model, I suggest that future researchers should perform deep sequencing 

across  repaired  breaks  to  invesggate  repair  outcomes  in  a  disrupted  shortrange  complex.  Our 

current model of NHEJ predicts that the high fidelity of joining rougne breaks can be ajributed 

to highly regulated access to DNA ends prior to ligagon. It would be very interesgng to know if a 

disrupgon in short-range complex stability (which has a modest impact on joining of most breaks, 

chapter 3) also disrupts fidelity of repair outcomes. Furthermore, combining L4 K273A with the 

RK>DD mutant (disrupgng the L4-Ku70 interacgon in the SRC) ablates nearly all joining. In my 

2022 paper, we suggested that prolonged synapsis in the short-range complex could contribute 

to excessive end processing and the increased rate of delegons. Perhaps the repair outcomes in 

these combinagon mutants are both less efficient at repairing breaks but higher fidelity than the 

forms of alternagve end joining.   

I also suggest that a future graduate student uglize the Ku70/DNA-PK/XLF/XRCC4 halo + 

L4  to  test  mutagons  idengfied  by  other  groups.  The  SMI  approach  in  these  cells  allows 

 91 

 
 
conclusions  regarding  the  gming  and  regulagon  of  NHEJ  that  are  currently  unaddressed.  The 

recent L4 DBD-DNA interacgng mutants reported in Sgnson et al. 2024 (48) clearly disrupt short-

range  complex  formagon  in  vitro,  but  there  are  important  differences  between  biochemical 

systems  and  live-cell  imaging.  For  example,  deplegon  of  XRCC4-L4  complex  (LX4)  from  frog 

extracts has minor impacts in the frog extract system has no discernable impact on LR complex 

formagon (46); however, there are clear major impacts on DNA-PKcs and Ku70’s associagon with 

chromagn in live cells, hypothesized to be dissolugon of the LRC if LX4 is unable to associate with 

breaks (73). To this end, I propose uglizing the  DNA-PKcs-, Ku70-, and XRCC4-halo L4-/- cell strains 

to test the L4 DBD-DNA interacgng mutants published in S9nson et al. 2024 (48). I predict that 

targegng  the  proposed  L4-DNA  interacgon  may  trap  NHEJ  complexes  in  the  LR  complex  in  a 

manner that protects from NHEJ-aborgon, resulgng in prolonged associagon of all three tagged 

NHEJ factors.  

Beyond  imaging  canonical  NHEJ  factors  in  the  context  of  L4  mutants,  it  also  would  be 

interesgng to image alternagve end joining factors both 1) in the presence of catalygcally inacgve 

L4  to  assess  gming  of  alt-EJ  amer  failed  NHEJ  and  2)  in  the  presence  of  fully  funcgonal  L4  to 

determine  whether  tradigonal  “backup”  factors  respond  to  DSB  inducgon.  There  is  an  open 

quesgon from chapters 2 and 3 whether XRCC1-L3 can respond to DSBs and act directly within 

the short-range complex in vivo.  I briefly ajempted a version of this experiment by transiently 

transfecgng L4-/- U2OS cells with expression vectors containing L4 K273A and L3-Halo (data not 

shown).  Unfortunately,  there  were  inconsistencies  with  L3-Halo  expression  levels,  and  I  was 

unable  to  perform  any  experiments  to  test  response  to  calicheamicin.    A  more  careful 

experimental  setup—likely  generagng  a  genomically  edited  halo-L3  L4-/-  cell  strain  may  be 

required to fully address this quesgon.   

Beyond imaging approaches, I think there is significant work to be done understanding 

both  what  processes  promotes  NHEJ’s  high  fidelity  outcomes  and  what  factors  determine 

pathway  choice  between  NHEJ  and  mutagenic  alternagve  end  joining.  Experiments  with  a 

substrate  that  assesses  joining  via  7  bp  of  microhomology  (Chapter  3)  suggest  that  a  certain 

subset  of  DSBs  are  joined  via  a  highly  mutagenic  Polθ-dependent  mechanism—even  in  the 

presence of NHEJ. Using high-depth Illumina sequencing, both my work and work from other labs 

 92 

 
(21, 20, 94, 10) indicates that NHEJ+ cells generally repair breaks with extremely high fidelity—or 

at  least  consistently.  Further  experiments  looking  at  mutagonal  frequency  from  resolved 

chromosomal and episomal breaks in the Polθ +/- cells, pargcularly the L4K273A*/- cells to examine 

the  quality  of  repair  outcome  following  break  inducgon.  Preliminary  sequencing  results  are 

inconclusive, VDJ coding substrates were a poor choice.   

While several papers have addressed the fidelity of NHEJ in small scale sequencing studies, 

I believe decreasing sequencing costs will enable larger whole-genome sequencing projects to 

invesggate  how  mutagons  in  NHEJ  impact  repair  fidelity  genome-wide.  To  do  this,  I  propose 

sequencing whole-genomes of cells treated +/- DSB inducing agents in NHEJ mutant cell lines. 

This could become a valuable complementary approach to smaller targeted experiments like the 

presented targeted sequencing and would deepen our understanding of how mutagons in NHEJ 

factors impact genomic stability—pargcularly those which have been idengfied as funcgonally 

significant  in  lab  or  prevalent  in  the  clinic.  Addigonally,  newer  sequencing  techniques—

pargcularly Oxford Nanopore Technologies can inform not only per-base data but also provide 

more  accurate  representagons  of  genomic  translocagons  and  copy  number  variagons  for 

repeggve regions.   

FINAL THOUGHTS 

First and foremost, thank you for reading through this dissertagon, I hope you appreciate 

both  the  amount  of  work  that’s  been  done  across  the  field  over  the  past  five  years  and  my 

contribugons to understanding L4’s role in the evolving NHEJ model. Graduate school during the 

COVID pandemic was not easy, between lockdowns limigng access to the lab (a significant hurdle 

when  first  isolagng  and  screening  clones  from  my  inigal  L4  CRISPR  experiment)  to  supply 

shortages delaying experiments. Over the course of my PhD, I was most surprised by the literature 

revealing how prevalent DNA-PKcs across eukaryotes. When I first met Kathy in early September 

2019 (before nearly all of the most recent studies into NHEJ’s synapgc mechanism), I recall talking 

about how the mysteries of DNA-PKcs expression levels in humans compared to its absence in 

many other well studied organisms. This lem major doubts about how widely we could apply our 

studies beyond vertebrates. Learning the true breadth of cross species DNA-PKcs conservagon— 

from  plants  to  fungi  to  insects  and  beyond—points  to  some  deeper  throughline  of  highly 

 93 

 
regulated end joining repair and deepens my appreciagon for NHEJ’s role across eukaryotes. To 

that end, I hope that research into DNA Ligase IV’s role in promogng end-joining congnues in 

groups at MSU and beyond.     

 94 

 
 
REFERENCES 

1.  

2.  

3.  

4.  

5.  

6.  

7. 

8.  

9.  

Graham,T.G.W., Walter,J.C. and Loparo,J.J. (2016) Two-Stage Synapsis of DNA Ends during 
Non-homologous End Joining. Molecular Cell, 61, 850–858. 

Sgnson,B.M.,  Moreno,A.T.,  Walter,J.C.  and  Loparo,J.J.  (2020)  A  Mechanism  to  Minimize 
Errors during Non-homologous End Joining. Molecular Cell, 77, 1080-1091.e8. 

Chen,S.,  Lee,L.,  Naila,T.,  Fishbain,S.,  Wang,A.,  Tomkinson,A.E.,  Lees-Miller,S.P.  and  He,Y. 
(2021) Structural basis of long-range to short-range synapgc transigon in NHEJ. Nature, 
10.1038/s41586-021-03458-7. 

Chaplin,A.K.,  Hardwick,S.W.,  Stavridi,A.K.,  Buehl,C.J.,  Goff,N.J.,  Ropars,V.,  Liang,S., 
Oliveira,T.M.D.,  Chirgadze,D.Y.,  Meek,K.,  et  al.  (2021)  Cryo-EM  of  NHEJ  supercomplexes 
provides insights into DNA repair. Molecular Cell, 0. 

Amin,H., Zahid,S., Hall,C. and Chaplin,A.K. (2024) Cold snapshots of DNA repair: Cryo-EM 
structures  of  DNA-PKcs  and  NHEJ  machinery.  Progress  in  Biophysics  and  Molecular 
Biology, 186, 1–13. 

Vogt,A.,  He,Y.  and  Lees-Miller,S.P.  (2023)  How  to  fix  DNA  breaks:  new  insights  into  the 
mechanism  of  non-homologous  end 
joining.  Biochemical  Society  Transac9ons, 
10.1042/BST20220741. 

Loparo,J.J.  (2023)  Holding  it  together:  DNA  end  synapsis  during  non-homologous  end 
joining. DNA Repair, 130, 103553. 

Chiruvella,K.K.,  Liang,Z.,  Birkeland,S.R.,  Basrur,V.  and  Wilson,T.E.  (2013)  Saccharomyces 
cerevisiae  DNA  Ligase  IV  Supports  Imprecise  End  Joining  Independently  of  Its  Catalygc 
Acgvity. PLOS Gene9cs, 9, e1003599. 

Sgnson,B.M., Carney,S.M., Walter,J.C. and Loparo,J.J. (2024) Structural role for DNA Ligase 
IV in promogng the fidelity of non-homologous end joining. Nat Commun, 15, 1250. 

10.   Wyaj,D.W.,  Feng,W.,  Conlin,M.P.,  Yousefzadeh,M.J.,  Roberts,S.A.,  Mieczkowski,P., 
Wood,R.D.,  Gupta,G.P.  and  Ramsden,D.A.  (2016)  Essengal  Roles  for  Polymerase  theta-
Mediated End Joining in the Repair of Chromosome Breaks. Mol. Cell, 63, 662–673. 

11.   Merker,L., Feller,L., Dorn,A. and Puchta,H. Deficiency of both classical and alternagve end-
joining pathways leads to a synergisgc defect in double-strand break repair but not to an 
increase in homology-dependent gene targegng in Arabidopsis. The Plant Journal, n/a. 

12.  

13.  

Zhou,J.,  Gelot,C.,  Pantelidou,C.,  Li,A.,  Yücel,H.,  Davis,R.E.,  Färkkilä,A.,  Kochupurakkal,B., 
Syed,A.,  Shapiro,G.I.,  et  al.  (2021)  A  first-in-class  polymerase  theta  inhibitor  selecgvely 
targets homologous-recombinagon-deficient tumors. Nat Cancer, 2, 598–610. 

Buehl,C.J.,  Goff,N.J.,  Hardwick,S.W.,  Gellert,M.,  Blundell,T.L.,  Yang,W.,  Chaplin,A.K.  and 
Meek,K. (2023) Two disgnct long-range synapgc complexes promote different aspects of 
end processing prior to repair of DNA breaks by non-homologous end joining. Molecular 
Cell, 10.1016/j.molcel.2023.01.012. 

 95 

 
14.  

15.  

Inagawa,T., Wennink,T., Lebbink,J., Keijzers,G., Florea,B., Verkaik,N. and van Gent,D. (2020) 
C-Terminal  Extensions  of  Ku70  and  Ku80  Differengally  Influence  DNA  End  Binding 
Properges. INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES, 21. 

Singleton,B.K.,  Torres-Arzayus,M.I.,  Ro~nghaus,S.T.,  Taccioli,G.E.  and  Jeggo,P.A.  (1999) 
The C Terminus of Ku80 Acgvates the DNA-Dependent Protein Kinase Catalygc Subunit. 
Mol. Cell. Biol., 19, 3267–3277. 

16.   Weterings,E.,  Verkaik,N.S.,  Keijzers,G.,  Florea,B.I.,  Wang,S.-Y.,  Ortega,L.G.,  Uematsu,N., 
Chen,D.J.  and  van  Gent,D.C.  (2009)  The  Ku80  Carboxy  Terminus  Sgmulates  Joining  and 
Artemis-Mediated Processing of DNA Ends. MCB, 29, 1134–1142. 

17.  

Chen,S., Lees-Miller,J.P., He,Y. and Lees-Miller,S.P. (2021) Structural insights into the role 
of DNA-PK as a master regulator in NHEJ. GENOME INSTAB. DIS., 2, 195–210. 

18.  Mikhova,M.,  Goff,N.J.,  Janovič,T.,  Heyza,J.R.,  Meek,K.  and  Schmidt,J.C.  (2024)  Single-
molecule  imaging  reveals  the  kinegcs  of  non-homologous  end-joining  in  living  cells. 
10.1101/2023.06.22.546088. 

19.  

20.  

21.  

22.  

23.  

24.  

25.  

Lees-Miller,J.P., Cobban,A., Katsonis,P., Bacolla,A., Tsutakawa,S.E., Hammel,M., Meek,K., 
Anderson,D.W.,  Lichtarge,O.,  Tainer,J.A.,  et  al.  (2021)  Uncovering  DNA-PKcs  ancient 
phylogeny, unique sequence mogfs and insights for human disease. Progress in Biophysics 
and Molecular Biology, 163, 87–108. 

Reid,D.A., Keegan,S., Leo-Macias,A., Watanabe,G., Strande,N.T., Chang,H.H., Oksuz,B.A., 
Fenyo,D.,  Lieber,M.R.,  Ramsden,D.A.,  et  al.  (2015)  Organizagon  and  dynamics  of  the 
nonhomologous  end-joining  machinery  during  DNA  double-strand  break  repair. 
Proceedings of the Na9onal Academy of Sciences, 112, E2575–E2584. 

Reid,D.A.,  Conlin,M.P.,  Yin,Y.,  Chang,H.H.,  Watanabe,G.,  Lieber,M.R.,  Ramsden,D.A.  and 
Rothenberg,E.  (2017)  Bridging  of  double-stranded  breaks  by  the  nonhomologous  end-
joining ligagon complex is modulated by DNA end chemistry. Nucleic Acids Research, 45, 
1872–1878. 

Conlin,M.P., Reid,D.A., Small,G.W., Chang,H.H., Watanabe,G., Lieber,M.R., Ramsden,D.A. 
IV  Guides  End-Processing  Choice  during 
and  Rothenberg,E.  (2017)  DNA  Ligase 
Nonhomologous End Joining. Cell Reports, 20, 2810–2819. 

Zhao,B., Watanabe,G., Morten,M.J., Reid,D.A., Rothenberg,E. and Lieber,M.R. (2019) The 
essengal elements for the noncovalent associagon of two DNA ends during NHEJ synapsis. 
Nat Commun, 10, 3588. 

Luedeman,M.E.,  Stroik,S.,  Feng,W.,  Luthman,A.J.,  Gupta,G.P.  and  Ramsden,D.A.  (2022) 
Poly(ADP) ribose polymerase promotes DNA polymerase theta-mediated end joining by 
acgvagon of end resecgon. Nat Commun, 13, 4547. 

Goff,N.J., Brenière,M., Buehl,C.J., de Melo,A.J., Huskova,H., Ochi,T., Blundell,T.L., Mao,W., 
Yu,K., Modesg,M., et al. (2022) Catalygcally inacgve DNA ligase IV promotes DNA repair 
in living cells. Nucleic Acids Research, 10.1093/nar/gkac913. 

 96 

 
 
APPENDIX: A list of authorship contribugons to other papers. 

This  is  a  list  of  authorship  contribugons  to  academic  papers  as  described  in  Chapter  4.  All 

informagon updated as of 11-25-2024, presented in reverse chronological order of acceptance 

for publicagon.  

1.  Medina-Suarez, D., Han, L., O’Reily, S., Liu, J., Wei, C., Brenière, M., Goff, N.J., Chen, C., 
Modesg, M., Meek, K., Harrington, B., Yu, K. (2024). DNA ligases 3 and 4 cooperate in vivo 
to facilitate DNA repair and organism viability. Nucleic Acids Research, In Press. 

2.  Mikhova,  M.,  Goff,  N.J.,  Janovič,  T.,  Heyza,  J.R.,  Meek,  K.,  Schmidt,  J.C.  (2024)  Single-
molecule  imaging  reveals  the  kinegcs  of  non-homologous  end-joining  in  living  cells. 
Nature Communica9ons, in press.  

3.  Pascarella, G., Conner, K.N., Goff, N.J., Carninci,P., Olive,A.J. & Meek,K. (2024) Compared 
to other NHEJ factors, DNA-PK protein and RNA levels are markedly increased in all higher 
primates, but not in prosimians or other mammals. DNA Repair, 142, 103737. 

4.  Goff,  N.J.,  Mikhova,  M.,  Schmidt,  J.C.,  &  Meek,  K.  (2024).  DNA-PK:  A  synopsis  beyond 

Synapsis. DNA Repair, 141, 103716.  

5.  Buehl, C. J., Goff, N.J., Hardwick, S.W., Gellert, M., Blundell, T.L., Yang, W., Chaplin, A.K., & 
Meek, K. (2023). Two disgnct long-range synapgc complexes promote different aspects of 
end-processing prior to repair of DNA breaks by non-homologous end joining. Molecular 
Cell 83(5), 698-714.  

6.  Goff,  N.  J.,  Brenière,  M.,  Buehl,  C.  J.,  de  Melo,  A.  J.,  Huskova,  H.,  Ochi,  T.,  Blundell,  T, 
Weifeng, M., Yu, K., Modesg, M., & Meek, K. (2022) Catalygcally inacgve DNA ligase IV 
promotes DNA repair in living cells. Nucleic Acids Research, 50(19), 11058–11071. 

7.  Liu, L., Chen, X., Li, J., Wang, H., Buehl, C. J., Goff, N. J., Meek, K., Yang, W., & Gellert, M. 
(2022). Autophosphorylagon transforms DNA-PK from protecgng to processing DNA ends. 
Molecular Cell, 82(1), 177-189. 

8.  Chaplin, A. K., Hardwick, S. W., Stavridi, A. K., Buehl, C. J., Goff, N. J., Ropars, V., Liang, S., 
De Oliveira, T. M., Chirgadze, D. Y., Meek, K., Charbonnier, J.B., & Blundell, T. L. (2021). 
Cryo-EM  of  NHEJ  supercomplexes  provides  insights  into  DNA  repair.  Molecular  Cell, 
81(16), 3400-3409. 

 97