of 30
UvrC Coordinates an O
2
-Sensitive [4Fe4S] Cofactor
Rebekah M. B. Silva
,
Michael A. Grodick
,
Jacqueline K. Barton
Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena,
California 91125, United States
Abstract
Recent advances have led to numerous landmark discoveries of [4Fe4S] clusters coordinated by
essential enzymes in repair, replication, and transcription across all domains of life. The cofactor
has notably been challenging to observe for many nucleic acid processing enzymes due to several
factors, including a weak bioinformatic signature of the coordinating cysteines and lability of the
metal cofactor. To overcome these challenges, we have used sequence alignments, an anaerobic
purification method, iron quantification, and UV–visible and electron paramagnetic resonance
spectroscopies to investigate UvrC, the dual-incision endonuclease in the bacterial nucleotide
excision repair (NER) pathway. The characteristics of UvrC are consistent with [4Fe4S]
coordination with 60–70% cofactor incorporation, and additionally, we show that, bound to UvrC,
the [4Fe4S] cofactor is susceptible to oxidative degradation with aggregation of apo species.
Importantly, in its holo form with the cofactor bound, UvrC forms high affinity complexes with
duplexed DNA substrates; the apparent dissociation constants to well-matched and damaged
duplex substrates are 100 ± 20 nM and 80 ± 30 nM, respectively. This high affinity DNA binding
contrasts reports made for isolated protein lacking the cofactor. Moreover, using DNA
electrochemistry, we find that the cluster coordinated by UvrC is redox-active and participates in
DNA-mediated charge transport chemistry with a DNA-bound midpoint potential of 90 mV vs
NHE. This work highlights that the [4Fe4S] center is critical to UvrC.
Graphical Abstract
Corresponding Author: Jacqueline K. Barton
– Division of Chemistry and Chemical Engineering, California Institute of
Technology, Pasadena, California 91125, United States; jkbarton@caltech.edu.
Complete contact information is available at:
https://pubs.acs.org/10.1021/jacs.0c01671
Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/jacs.0c01671
.
Additional experimental methods; protein sequence regions containing putative LYR motifs; affinity and size exclusion
chromatograms from purification; EPR spectra of ferric iron signals; SDS-PAGE, UV–vis spectra, and standard curve for molecular
weight verifying that O
2
targets the [4Fe4S] center; selected data from early studies of UvrC; induction trials of C
A mutants;
stability of UvrC in buffer containing a lower concentration of KCl; scan rate dependence of UvrC on DNA-modified Au electrodes;
table of primers and substrates used in this study (
PDF
)
The authors declare no competing financial interest.
HHS Public Access
Author manuscript
J Am Chem Soc
. Author manuscript; available in PMC 2020 July 30.
Published in final edited form as:
J Am Chem Soc
. 2020 June 24; 142(25): 10964–10977. doi:10.1021/jacs.0c01671.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
INTRODUCTION
Iron sulfur clusters are ubiquitous in proteins as inorganic cofactors found in all domains of
life and are essential in major cellular processes, including respiration, photosynthesis, and
nitrogen fixation.
1
5
In the past three decades, iron sulfur clusters have been found
coordinated by many DNA repair and replication enzymes essential for genome
maintenance.
6
11
In most of these proteins, however, the [4Fe4S] cofactor is not involved
directly in catalytic activity on DNA substrates, so identifying a general role for the metal
cofactor has become a growing area of interest.
12
Toward this end, we have investigated the
DNA-bound redox chemistry of several repair and replication proteins and demonstrated that
these proteins not only participate in DNA charge transport (DNA CT) chemistry but also do
so at a shared potential of ~80 mV vs NHE, making electron transfer energetically favorable.
We have found that the DNA-bound potentials of repair and replication proteins studied to
date are consistent with those of high potential iron–sulfur proteins (HiPIPs).
12
Models have
been developed to describe how redox signaling among repair proteins through DNA CT is
utilized for rapid lesion detection and to explain how the oxidation of the cluster can act as a
redox switch that modulates binding affinities to DNA substrates as well as processivity of
eukaryotic replication polymerases.
13
15
The observation of a [4Fe4S] cluster in repair and replication proteins has often occurred
many years after the first characterization of the gene and gene products due to several
reasons.
7
,
8
,
12
First, [4Fe4S] clusters, like other metal centers, can be sensitive to expression
and purification conditions, which has delayed isolation of protein with intact clusters.
16
18
Second, prediction of [4Fe4S] coordination sites in nucleic acid processing enzymes has
been particularly challenging because the spacing of the coordinating cysteines is atypical
across different protein families, leading to a weak bioinformatic signature.
8
Other
indications that a protein may bind an iron–sulfur cluster, such as the LYR sequence motif
identified first in human protein sequences as a recognition element for iron–sulfur
biogenesis machinery, are just emerging and could inform predictive tools for other
organisms in the future.
17
,
19
21
In light of these reasons, several research groups have
proposed that many more iron– sulfur cluster proteins involved in nucleic acid processing
remain to be discovered.
7
,
16
,
17
In
Escherichia coli
(
E. coli
), two repair glycosylases, Endonuclease III (EndoIII) and MutY,
and one Superfamily 2 (SF2) 5
3
repair helicase, DinG, coordinate a [4Fe4S] cluster (a
paralog of DinG, YoaA, has also recently been predicted to bind a [4Fe4S] cluster based on
sequence similarity) and are active in essential repair pathways in prokaryotes.
22
28
EndoIII
Silva et al.
Page 2
J Am Chem Soc
. Author manuscript; available in PMC 2020 July 30.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
and MutY primarily resolve oxidized pyrimidines and 8-oxoguanine mispaired with adenine,
respectively, in Base Excision Repair (BER), while DinG is involved in resolving R-loops
and D-loops (
RNA or
DNA or triple-stranded nucleic acid structures) which can occur at
collisions between transcription and replication machinery.
29
,
30
Inexplicably, DinG
overexpression and Δ
dinG
strains of
E. coli
are more sensitive to UV light, suggestive of
redox signaling between DinG and an unknown protein in Nucleotide Excision Repair
(NER; in eukaryotic and archaeal NER, XPD is known to a coordinate a [4Fe4S] center).
30
,
31
In prokaryotic NER, a relatively small pool of proteins recognize and repair structurally
and chemically diverse bulky DNA lesions that arise due to DNA damage by UV light,
cigarette smoke, small molecule chemotherapeutics, and protein–DNA adducts, among
others.
32
34
In the classic model of bacterial NER, the UvrAB complex scans and searches
the genome for bulky lesions, UvrC in complex with UvrB makes an incision 5
and 3
to
the site of damage, UvrD unwinds the damaged oligomer, and DNA polymerase 1 and DNA
ligase patch the damaged strand, completing the repair process (Figure 1).
32
34
UvrC is unique in many respects among the proteins in the NER pathway, most notably in
how challenging the protein has been to study, both
in vitro
and
in vivo
.
32
35
Insight on
UvrC has been difficult to gain because of limited sequence homology to other proteins.
UvrC comprises its own protein family of dual-incision endonucleases (also referred to as
excision nucleases or exinucleases) with two distinct active sites located at the N and C
termini of the same peptide.
33
,
36
UvrC has also been noted by many in the field to be
difficult to purify and, once purified, to be susceptible to degradation, aggregation, and loss
of activity over time, even when stored frozen.
34
36
Though first isolated in 1981, a full-
length or near full-length crystal structure has not been reported for UvrC (in contrast to
UvrA, UvrB, and UvrD), and cocrystal structures of UvrC in complex with a substrate or
other NER proteins are also not available, which has limited searches for structural
homology.
33
,
37
,
38
In vivo
regulation of UvrC expression is complex and has been reported to
be under the control of multiple promoters, which further distinguishes UvrC from the other
proteins in the NER pathway.
39
Expression of radio-labeled UvrC was first reported to be
10–20 copies per cell, though subsequent studies of UvrC mRNA levels, fluorescently
labeled UvrC, and active ribosomes have reported between 0.05 to 300 copies of UvrC per
cell (or per cell per generation).
39
41
Furthermore, the expression of UvrC is not inducible
by the SOS system, a cellular-wide response to DNA damage present in many bacteria.
42
This again is in contrast to UvrA, UvrB, and UvrD, which are induced during the SOS
response.
43
In the context of repair proteins bearing a [4Fe4S] cofactor, tight regulation of
copy number appears to be a common characteristic of these metalloproteins in
E. coli
.
12
Our initial examination of UvrC began with the sequence of the protein. Five main regions
of UvrC have been identified, which include the N-terminal GIY-YIG endonuclease domain
(3
incision), a cysteine rich region, a UvrBC interacting domain, the RNaseH endonuclease
domain (5
incision), and a helix-turn-helix motif (Figure 1).
33
,
34
In the N-terminal cysteine-
rich region, four highly conserved cysteine residues at positions 154, 166, 174, and 178 have
been observed. The function of the conserved cysteines has been unknown, but speculated to
facilitate interactions between UvrB and UvrC.
34
Because the four cysteines are atypically
spaced, close in proximity, and highly conserved through the Bacteria domain and up to
some archaeal species (CysX
6–14
CysX
7
CysX
3
Cys consensus sequence), we hypothesized
Silva et al.
Page 3
J Am Chem Soc
. Author manuscript; available in PMC 2020 July 30.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
that these cysteine residues coordinate a [4Fe4S] cluster (Figure 1). Two putative (LIV)(YF)
(RK) tripeptide motifs were also observed (Figure S1), one of which is located N-terminal to
the predicted coordination site and the other just after Cys154. Additionally, conserved
aromatic and proline residues are also located around Cys154, Cys166, Cys174, and Cys178
(Figure S1), one of the few common themes among the sequences of nucleic acid processing
enzymes that coordinate the [4Fe4S] cofactor.
8
More moderately conserved cysteine
residues can be found at positions 265, 398, and 413, but these were not hypothesized to
participate in Fe coordination because of the absence of other supporting sequence motifs
suggestive of a [4Fe4S] binding domain.
Here, we report the discovery and characterization of a HiPIP-like [4Fe4S] cluster that is
coordinated by UvrC. The holo form of UvrC is stabilized relative to its apo form; distinct
from what has been reported to date for other [4Fe4S] repair proteins in
E. coli
, the cluster
degrades in the presence of molecular oxygen. We also find, in its holo form, UvrC
independently forms a high-affinity complex with DNA, though DNA binding does not lead
to independent enzymatic activity on a damaged substrate. Finally, using DNA-modified Au
electrodes, we observe that UvrC participates in DNA CT chemistry and shares a DNA-
bound midpoint potential seen previously for EndoIII, MutY, and DinG. Based on these
observations, we discuss UvrC activity in its holo form in the context of the NER pathway
and the greater redox signaling network of [4Fe4S] repair proteins.
RESULTS AND DISCUSSION
Development of an Anaerobic Purification Method for UvrC.
UvrC has historically been difficult to express and purify.
32
35
Accordingly, we screened
new expression and purification strategies that would yield soluble and pure protein in large
enough quantities that a [4Fe4S] metal center could be detected spectroscopically. Included
in the screen was a pBAD overexpression plasmid under the control of the
L
-arabinose
promoter with a His
6
-Maltose Binding Protein (MBP) affinity/solubility tag encoded N-
terminally to UvrC (see Supporting Experimental Section).
44
47
The
L
-arabinose promoter
can be used to prevent leaky expression prior to induction with arabinose, which is an
advantage for overexpressing proteins that are potentially toxic to cells prior to induction.
48
51
Additionally, MBP tags are frequently used to enhance the expression and solubility of
target proteins, including for previous overexpression of the C-terminal half of UvrC as well
as repair and replication enzymes bearing the [4Fe4S] cluster.
52
58
A cooler overexpression
temperature (22 °C), longer induction time (16 h), and an anaerobic purification in buffers
that included a high concentration of potassium chloride were also chosen to minimize
accumulation of apo, aggregated, or degraded protein.
7
,
11
,
17
,
59
65
We found that the pBad overexpression system resulted in detectable overexpression of a
His
6
-MBP-UvrC fusion protein (110 kDa) in whole cell lysate (referred to as UvrC for the
remainder of this manuscript, Figure S2). Cell lysis, purification, and concentration of the
UvrC fusion protein using immobilized metal affinity and gel filtration chromatography was
performed under strict anaerobic conditions in an anaerobic chamber and using a standard
Schlenk line technique (Figure S2).
66
,
67
UvrC was concentrated between 20 and 30
μ
M to
prevent precipitation after flash freezing in liquid N
2
and stored at −80 °C. The expression
Silva et al.
Page 4
J Am Chem Soc
. Author manuscript; available in PMC 2020 July 30.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
and anaerobic purification strategy allowed for isolation of the UvrC fusion protein in high
purity with a broad and shallow absorption band centered at 410 nm and yellow-tan color
characteristic of a [4Fe4S] cluster (Figure 2).
3
,
68
70
Based on the ratio of absorbances at 410
and 280 nm, each purification yielded approximately 5 mg of the UvrC fusion protein per
liter of liquid culture (5 g of wet pellet) with 60–70% incorporation of the [4Fe4S] metal
center (Figure S2).
11
,
15
,
24
,
61
,
70
81
To characterize the nature of the Fe center further, we
used the ferene colorimetric assay to quantify the amount of protein-bound Fe.
80
We find
that UvrC coordinates about 3 Fe per protein on average (or 4.0 ± 0.3 Fe per cluster, Figure
2 and Figure S2), which is consistent with the substoichiometric levels of [4Fe-4S] cluster
incorporation seen with other repair and replication enzymes. A range from 2 to 4 Fe per
protein has been reported and has been attributed to incomplete incorporation or loss of a
labile Fe during the process of overexpression or purification, even when each step is
completed anaerobically.
25
,
27
,
53
,
61
,
64
,
82
85
Holo-UvrC Is Redox-Active.
We used X-band electron paramagnetic resonance (EPR) spectroscopy to assign oxidation
states of the [4Fe4S] cluster. A small signal centered at
g
= 2.01 can be seen from the native
UvrC EPR spectrum, which has been observed previously and found to be due to a small
percentage of the native protein population in the [3Fe4S]
1+
state (with the rest of the
population in the EPR-silent [4Fe4S]
2+
state) (Figure 2).
27
,
54
,
64
,
72
,
82
Two categories of
protein-bound [4Fe4S] clusters are known: (i) ferredoxins which cycle between the 2+/1+
oxidization states and (ii) HiPIPs which cycle between the 3+/2+ oxidation states.
68
,
69
,
86
To
classify the nature of the [4Fe4S] cluster, UvrC was treated with the oxidant potassium
ferricyanide and immediately frozen in liquid N
2
, which resulted in a large and sharp signal
at
g
= 2.01.
27
,
53
,
61
,
82
,
83
,
85
We assign this signal to a [3Fe4S]
1+
species derived from an
oxidized [4Fe4S]
3+
cluster, indicating that UvrC accesses the 3+/2+ redox couple.
Equivalent spectra have been observed for EndoIII and homologues, repair proteins with
[4Fe4S] clusters that are generally substantially more stable under aerobic conditions.
54
A
corresponding signal at
g
= 4.3 can also be observed after treatment with ferricyanide,
suggestive of a ferric species in solution (Figure S3) and consistent with the release of an
iron atom from the [4Fe4S]
3+
species following oxidization.
87
,
88
No clear evidence of the
[4Fe4S]
3+
species, characterized by a
g
= 2.1, before Fe loss was apparent.
3
,
68
,
69
Taken
together, the data for UvrC are consistent with other DNA repair proteins and coordinates a
HiPIP-like [4Fe4S] cluster.
The UvrC-Bound [4Fe4S] Cluster Is Susceptible to Oxidative Degradation.
Protein-bound Fe centers can serve multiple functions in the cell,
6
,
89
91
and the role(s) of
the [4Fe4S] center are still being defined for repair and replication proteins. The [4Fe4S]
cluster is not involved in active site chemistry on DNA substrates for the majority of repair
and replication proteins,
12
though a role for reactivity with dioxygen (and other reactive
species) and redox signaling are just emerging (
vide infra
). With regard to sensing reactive
species, particularly dioxygen, part of the challenge is that a range of stabilities under
aerobic conditions has been observed for repair and replication proteins, and more labile
[4Fe4S] centers which require anaerobic expression and/or purification conditions to remain
intact (and therefore detectable) have only recently been reported.
12
To investigate if the
Silva et al.
Page 5
J Am Chem Soc
. Author manuscript; available in PMC 2020 July 30.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
[4Fe4S] cluster of UvrC sensitive to O
2
at a physiologically relevant temperature, we
incubated UvrC in aerobic UvrC buffer (25 mM Tris-HCl, 0.5 M KCl, 20% v/v glycerol) at
37 °C and monitored the absorption band at 410 nm. We find that on the time scale of a
UvrC activity assay (1 h), 20% of the cluster degraded after incubation at 37 °C, while
complete degradation was observed after 4 h (Figure 3). We verified that the bleaching of the
absorbance at 410 nm was due only to degradation of the [4Fe4S] center and not the peptide
through gel analysis (Figure S4). Notably, anaerobic incubation of holo UvrC at 37 °C in the
absence of O
2
does not result in degradation of the [4Fe4S] cluster (Figure S4). We also
observe that binding to duplex DNA does not affect cluster degradation (see below for
studies of UvrC–DNA complexes) (Figure S4). We emphasize here that, in order to detect
the oxygen-driven degradation spectroscopically, concentrations of UvrC well above that of
what is commonly found in an activity assay needed to be used. Our observations highlight
the subtlety of handling a metalloprotein
in vitro
and the importance of carefully monitoring
cofactor stability.
Because apo and holo forms of metalloproteins can have different oligomeric states, we
compared holo- and apo-UvrC by analytical size exclusion chromatography.
6
,
89
Using a
standard curve, we determined that holo-UvrC elutes at a volume consistent with protein
migrating in dimeric form (Figure 3 and Figure S4). In contrast, apo-UvrC species elute at
the void volume of the column, which corresponds to aggregates that would be greater than
600 kDa in mass (Figure 3). Such a sensitivity to O
2
has not been seen previously in repair
proteins that coordinate a [4Fe4S] cluster.
The significance of our observations and findings of other investigators regarding how
[4Fe4S] cofactor in EndoIII, MutY, DinG, and UvrC are transformed by exposure to reactive
species remains to be explored fully
in vitro
and
in vivo
. As summarized above, EndoIII,
MutY, and DinG have not been reported to be similarly sensitive to O
2
. Moreover, we have
observed previously that EndoIII is not only highly soluble but also stable at room
temperature in atmosphere for many days. EndoIII, homologue MutY, and DinG have
furthermore been crystallized aerobically.
58
,
74
,
80
,
92
94
EndoIII does, however, react rapidly
with another diatomic signaling molecule, NO, causing loss of one iron atom per cluster and
formation of a mononuclear dinitrosyl iron complex and a dinuclear Roussin’s red ester in
the cluster binding domain.
80
Transformation of the iron center is reversible and does not
affect global protein structure or DNA binding, but does shift the redox potential of the
cluster and hinders enzymatic activity.
80
,
95
DinG also reacts with NO and is inactivated, but
surprisingly, is resistant to treatment with H
2
O
2
.
25
Cellular responses mediated by sensing of
reactive species by iron– sulfur transcription factors has been well-characterized, though in
contrast, the specificity and responses of repair proteins to reactive species are not as fully
understood.
6
Variations in the stability of the [4Fe4S] cofactor from repair proteins across
different bacterial species adapted to niche environments are also not fully appreciated.
33
,
96
99
Exploring how UvrC and other DNA repair enzymes are involved in detecting
changes in cellular environments and in the cellular response to endogenous and exogenous
stressors (separate or related to their repair activities in the genome) is an area that warrants
further investigation.
Silva et al.
Page 6
J Am Chem Soc
. Author manuscript; available in PMC 2020 July 30.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
As emphasized above, metalloproteins can be challenging to handle, and oxidative
degradation of the [4Fe4S] cofactor of UvrC was nontrivial to observe. Our early studies
were carried out with UvrC purified aerobically at 4 °C (Figure S5, top left).
100
The protein
was isolated from a peak that eluted after the void volume of the size exclusion column, was
yellow in color, and exhibited a temperature-dependent EPR signal, so it appeared that an
aerobic purification at 4 °C was sufficient for isolating the holoenzyme (Figure S5). This
sensitivity to O
2
could have easily been overlooked, but in the process of working with
aerobically purified UvrC, we noticed that the protein bleached over the course of
experiments. During a straightforward time course monitoring the absorbance band centered
at 410 nm, near-complete disappearance of the signal occurred after incubating at 37 °C
after 1 h, which is a common incubation temperature and time for many activity assays
(Figure S5, top right).
36
,
98
,
101
103
Furthermore, most activity assays use low concentrations
of enzyme such that the cofactor absorbance (and therefore integrity) is not generally
monitored.
Comparison of the UV–vis spectra of aerobically purified and anaerobically purified UvrC
further highlights the elusive nature of holo-UvrC. Briefly, the UV–vis absorption spectra of
both proteins are similar, though there is a small feature in the UV–vis spectrum of
aerobically purified UvrC at 325 nm, which could be due to the partially degraded cluster in
the [3Fe4S]
+
state (Figure S5).
69
Differences are also seen with the chromatograms of the
size exclusion column (Figure S5). Aggregated protein and what we thought was soluble
protein were not as easily separated under aerobic conditions, and the soluble protein
displayed an earlier elution volume. Based on the chromatogram, the yield of aerobically
isolated UvrC was also lower in comparison to anaerobically isolated UvrC. We also note
that postpurification, concentration steps were done aerobically at 4 °C during initial studies,
so it is likely that oligomeric species were generated
in situ
prior to freezing and storing.
These observations from early studies with UvrC highlight the necessity of handling this
protein in an anaerobic environment to minimize heterogeneity.
Mutation of Coordinating Cysteines Leads to Instability.
To explore the site of coordination, we generated Cys
Ala mutations of the cysteine
residues that are highly conserved at positions Cys154, Cys166, Cys174, and Cys178. Using
the conditions optimized for overexpression, we found the Cys154Ala (Cys154
Ala154)
and Cys166Ala mutants overexpress similarly to WT UvrC (Figure S6). In contrast,
expression of the Cys174Ala and Cys178Ala mutants were not detectable in whole cell
lysate. The Cys154Ala and Cys166Ala mutants precipitated or aggregated, eluting at the
void volume of the size exclusion column (data not shown). Attempts were made to isolate
any amount of the Cys174Ala and Cys178Ala mutants, but none could be harvested (data
not shown). The coordinating cysteine residues of repair and replication proteins have been
found to be involved in regulating protein expression, protein stability, subunit assembly, and
enzymatic activity.
64
,
79
,
104
,
105
The extent to which each protein, including homologous
protein, is disrupted by cysteine mutations (relative to WT) appears to vary widely
depending on which residue is mutated and what mutation is chosen. Overall, our
observations support the assignment of Cys154, Cys166, Cys174, and Cys178 as cluster-
Silva et al.
Page 7
J Am Chem Soc
. Author manuscript; available in PMC 2020 July 30.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
stabilizing residues, and examination of other regulatory aspects of the cysteines ligating the
[4Fe4S] cluster of UvrC warrants further investigation.
UvrC Independently Forms a Complex with DNA Substrates.
To continue characterizing UvrC in holo form, we examined how UvrC interacts with
radiolabeled, duplexed 30 base pair substrates using electrophoretic mobility shift assays
(EMSAs) completed in an anaerobic chamber.
78
,
80
,
106
,
107
UvrC along with the full UvrABC
exinuclease has been studied extensively
in vitro
with single-stranded DNA (ssDNA), well-
matched (WM) double-stranded DNA (dsDNA), damaged duplex substrates, and substrates
with nicks, gaps, bubbled regions, and overhangs derived from plasmid DNA and synthetic
oligomers.
33
,
34
The majority of previous work from chromatographic, optical, and gel-based
methods with UvrC from
E. coli
and thermophilic bacteria indicates that UvrC does not
form a complex with dsDNA independently of UvrA and UvrB.
108
112
(It should be noted
that UvrC, both truncated and full-length, has been seen to bind to ssDNA or single-stranded
regions of nicked, gapped, or bubbled substrates.
37
,
103
,
111
113
) Thus, it is widely accepted
that UvrC requires the action(s) of UvrA and UvrB in order to associate with substrates of
duplex character.
There have, however, been two reports of UvrC species forming a complex with dsDNA at
equilibrium (a tetramer in a gel-based assay) and nonequilibrium (single molecule assays)
conditions.
114
116
It is not clear how these data can be reconciled. What is clear is that UvrC
is sensitive to the conditions under which it is studied, exemplified by reports that UvrC has
a tendency to form precipitates with itself, with UvrAB, and with DNA substrates.
35
,
117
,
118
Since the majority of the genome in the cell is comprised of nondamaged dsDNA, we were
interested in how UvrC in its holo form interacts with WM dsDNA. We also selected a
fluorescein-modified substrate (F, Table S1), which is considered to mimic damage found in
the cell caused by polycyclic aromatic hydrocarbons, natural substrates of the UvrABC
system.
119
Once formation of WM and F duplexes was confirmed with annealing titrations
(Figure S7), UvrC was incubated anaerobically with duplexed substrates at a high KCl
concentration to avoid precipitation of UvrC (discussed above). Mixtures of UvrC and DNA
did not appear cloudy, and no scattering was observed by UV–vis (see Figure S7). Free and
complexed DNA was resolved on a native gel that was pre-equilibrated in degassed running
buffer. Band intensities were quantified, and the fraction of complexed DNA as a function of
free UvrC concentration was fit to the Hill equation. UvrC was found to form high affinity
complexes with duplexed substrates, with apparent dissociation constants of 100 ± 20 nM
and 80 ± 30 nM for WM and F substrates, respectively (
n
= 3 independent trials) (Figure 4).
The complexes not only appear to be high affinity but also stable, even at the high KCl
concentration used; band smearing was not observed over all three trials.
We note that the UvrC-DNA complex displayed a much lower mobility than free duplexes.
Low mobility of UvrC on native gels has been seen previously (with and without DNA
substrates) and has been attributed to the positive charge of UvrC (predicted pI of about 9) in
neutral buffers, which would cause migration to the positive electrode to be unfavorable.
111
Low migration has also been attributed to precipitation of protein in the gel,
116
but we do
not observe precipitation in solution (Figure S7). However, Hill coefficients >1 were found
Silva et al.
Page 8
J Am Chem Soc
. Author manuscript; available in PMC 2020 July 30.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
for both WM and F substrates, suggesting there is also some possibility of oligomerization
upon binding to DNA. Low mobility in our system could therefore be explained through the
predicted positive charge of UvrC or to formation of high molecular weight oligomers of
UvrC on DNA. In any case, we conclude that holo-UvrC forms a high affinity complex with
undamaged and damaged duplexed DNA.
DNA Binding Does Not Lead to Independent Enzymatic Activity.
Our observation that holo-UvrC forms complexes with dsDNA independently of other NER
proteins led us to investigate if holo-UvrC also exhibited independent or nonspecific
enzymatic activity, which has been observed previously.
120
,
121
UvrC uniquely contains two,
independent active sites, a GIY-YIG motif in the N-terminal domain which is responsible for
making the incision 3
to the site of damage and a second active site which is a structurally
conserved RNase H-like domain in the C-terminal end of the protein that achieves incision
5
to the damage site.
33
The 5
and 3
incisions can even be reconstituted
in vitro
with C-
terminal and N-terminal truncation products of UvrC, respectively, which both exclude the
Cys-rich region.
Standard activity assay conditions include Mg
2+
, ATP, DTT (or another common reducing
agent), and a KCl concentration of 0.1 M.
36
,
101
103
,
122
We first verified that buffer-
exchanging UvrC into a lower-salt buffer did not cause immediate destabilization of the
protein. The UV–vis spectrum indicated that [4Fe4S] cofactor was not lost in the process of
buffer exchanging (Figure S7). Furthermore, size exclusion chromatography confirmed that
UvrC in buffer containing 0.1 M KCl eluted at the same volume as UvrC in buffer
containing 0.5 M KCl, confirming that the oligomeric state was unchanged during buffer
exchange (Figure S7). We also verified that UvrC exhibited a similar binding profile to
dsDNA (both WM and F substrates) in buffer that contained 0.1 M KCl as in buffer
containing 0.5 M KCl (Figure S7 and see Figure 3 for comparison).
We tested the activity of UvrC at multiple concentrations on the F substrate as well as the
WM substrate as a control. Even up to a concentration of 1
μ
M UvrC by cluster (3:1
DNA:UvrC ratio), no evidence of substrate incision by UvrC under the conditions tested was
observed (Figure 5 and Figure S7). The absence of enzymatic activity in the presence of
DNA binding suggests that a complicated set of factors controls UvrC activity, which appear
to prevent spurious reactions from occurring even as UvrC is bound to dsDNA in holo form.
How the [4Fe4S] cofactor is involved in such regulation remains to be determined. For other
repair and replication proteins, the finely tuned roles of the [4Fe4S] cofactor have been
examined over many studies. These studies have been particularly informative for
understanding how disruption of the [4Fe4S] cofactor inhibits subunit assembly (and
therefore enzymatic activity)
or
enzymatic activity alone of the multisubunit B family
replication enzymes, polymerases
ε
and
δ
, respectively.
64
,
105
,
123
In the context of the
multisubunit exinuclease repair complex formed by UvrABC, we speculate that because the
[4Fe4S] domain is adjacent to the UvrBC interacting domain (see Figure 1), the [4Fe4S]
cluster may analogously be involved in the association of UvrC with other NER proteins or
the overall activity of the exinuclease complex.
7
Other roles may also exist for the [4Fe4S]
Silva et al.
Page 9
J Am Chem Soc
. Author manuscript; available in PMC 2020 July 30.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript