RESEA
RCH
ARTICL
E
Genetically
dissecting
the
electron
transport
chain
of
a soil
bacterium
reveals
a
generalizable
mechanism
for
biological
phenazine-1-carboxyl
ic acid
oxidation
Lev
M.
Z. Tsypin
ID
1
¤
, Scott
H.
Saunders
ID
2
, Allen
W.
Chen
ID
3
, Dianne
K.
Newman
ID
1,4
*
1
Division
of Biology
and
Biological
Engineeri
ng,
Californi
a Institute
of Technolo
gy,
Pasade
na,
California,
United
States
of America,
2
Green
Center
for
Systems
Biology
—Lyda
Hill
Department
of Bioinform
atics,
University
of Texas
Southwester
n Medical
Center,
Dallas,
Texas,
United
States
of America,
3
Divisio
n of
Chemis
try
and
Chemic
al Engineering,
California
Institute
of Technolo
gy,
Pasade
na,
Californi
a, United
States
of America,
4
Division
of Geologic
al and
Planetary
Sciences,
Californi
a Institute
of Technolo
gy,
Pasade
na,
Californi
a, United
States
of America
¤
Current
address:
Departm
ent
of Pathology
, Leland
Stanford
Jr.
Universit
y, Palo
Alto,
Californi
a, United
States
of America
*
dkn@ca
ltech.edu
Abstract
The
capacity
for
bacterial
extracellular
electron
transfer
via
secreted
metabolites
is wide-
spread
in natural,
clinical,
and
industrial
environments.
Recently,
we
discovered
the
biologi-
cal
oxidation
of phenazine-
1-carboxylic
acid
(PCA),
the
first
example
of biological
regeneration
of a naturally
produced
extracellular
electron
shuttle.
However,
it remained
unclear
how
PCA
oxidation
was
catalyzed.
Here,
we
report
the
mechanism,
which
we
uncovered
by
genetically
perturbing
the
branched
electron
transport
chain
(ETC)
of the
soil
isolate
Citrobacter
portucalensis
MBL.
Biological
PCA
oxidation
is coupled
to anaerobic
res-
piration
with
nitrate,
fumarate,
dimethyl
sulfoxide,
or trimethylamine
-N-oxide
as
terminal
electron
acceptors.
Genetically
inactivating
the
catalytic
subunits
for
all
redundant
com-
plexes
for
a given
terminal
electron
acceptor
abolishes
PCA
oxidation.
In the
absence
of qui-
nones,
PCA
can
still
donate
electrons
to certain
terminal
reductases,
albeit
much
less
efficiently.
In
C
.
portucalensis
MBL,
PCA
oxidation
is largely
driven
by
flux
through
the
ETC,
which
suggests
a generalizable
mechanism
that
may
be
employed
by
any
anaerobically
respiring
bacterium
with
an
accessible
cytoplasm
ic membrane.
This
model
is supported
by
analogous
genetic
experiments
during
nitrate
respiration
by
Pseudomon
as
aeruginosa
.
Author
summary
Many
bacteria
have
extremely
flexible
metabolisms,
and
we
are
only
beginning
to
under-
stand
how
they
manifest
in
the
environment.
Our
study
focuses
on
the
role
of
phenazine-
1-carboxylic
acid
(PCA),
a molecule
that
some
bacteria
synthesize
and
secrete
into
their
surroundings.
PCA
is an
“extracellular
electron
shuttle,”
a molecule
that
readily
transfers
PLOS GENE
TICS
PLOS
Genetics
| https://
doi.org/10.13
71/journal.p
gen.101
1064
May
6, 2024
1 / 25
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
OPEN
ACCESS
Citation:
Tsypin
LMZ,
Saunders
SH, Chen
AW,
Newman
DK (2024)
Genetically
dissecting
the
electron
transpo
rt chain
of a soil bacterium
reveals
a generali
zable
mechanism
for biological
phenazine-1-
carboxylic
acid oxidation.
PLoS
Genet
20(5):
e1011064.
https://
doi.org/10.1371
/journal.
pgen.101106
4
Editor:
Jue D. Wang,
Univers
ity of Wisconsin-
Madison,
UNITED
STATES
Received:
November
14, 2023
Accepted:
March
25, 2024
Published:
May 6, 2024
Peer Review
History:
PLOS
recognize
s the
benefits
of transpar
ency
in the peer review
process;
therefore,
we enable
the publication
of
all of the content
of peer review
and author
response
s alongside
final,
published
articles.
The
editorial
history
of this article
is available
here:
https://doi.o
rg/10.1371/jo
urnal.pgen.1
011064
Copyright:
©
2024
Tsypin
et al. This is an open
access
article
distributed
under
the terms
of the
Creative
Commons
Attribution
License,
which
permits
unrestricte
d use, distribu
tion, and
reproduction
in any medium,
provided
the original
author
and source
are credited.
Data
Availabilit
y Statement:
The authors
confirm
that all data underlying
the findings
are fully
available
without
restrictio
n. All relevant
data are
electrons
between
cells
and
oxidizing/reducing
compounds
or
other
cells.
Until
our
inves-
tigation,
the
role
of
PCA
electron-shuttling
had
only
been
studied
in
one
direction:
how
it
takes
electrons
away
from
cells,
and
the
effect
this
has
on
their
viability.
Here
we
present
a
detailed
account
of
the
opposite
process
and
its
mechanism:
what
happens
when
PCA
delivers
electrons
to
cells?
Our
findings
indicate
that
this
previously
underappreciated
process
is generalizable
to
any
anaerobically
respiring
bacterium.
Consequently,
we
expect
that
electron
donation
by
PCA
is widespread
in
environments
where
PCA
is plentiful
and
oxygen
is sparse,
such
as
in
some
agricultural
soils.
The
universality
of
the
extracellular
electron
shuttle
oxidation
mechanism
we
describe
for
PCA
suggests
that
it should
also
occur
with
similar
small
molecules,
of
which
there
are
thousands,
deepening
the
implica-
tion
that
this
is a significant
process
in
the
environment
and
motivating
further
research
into
its
consequences.
Introduction
Phenazines
are
secreted
secondary
metabolites
produced
by
diverse
soil
bacteria
[1]
that
microbes
use
in
various
ways:
from
quorum
sensing
[2]
to
antimicrobial
warfare
[3–5]
to
energy
conservation
under
anoxia
[6].
Each
of
these
biological
roles
is connected
to
the
ability
of
phenazines
to
accept
and
donate
electrons
(i.e.,
their
redox
activity),
a process
that
has
been
studied
for
over
120
years.
Bacterial
phenazine
reduction
was
first
proposed
in
the
nineteenth
century
as
an
indicator
for
the
presence
of
enterics
in
water
supplies
[7].
Several
decades
later,
pyocyanin,
one
of
the
phenazines
produced
by
Pseudomonas
aeruginosa
,
was
described
as
an
“accessory
respiratory
pigment”
that
increased
the
rate
of
oxygen
consumption
by
Staphylococ-
cus
,
Pneumococcus
,
and
erythrocytes
by
shuttling
electrons
from
the
cells
to
oxygen
[8].
Once
it became
apparent
that
phenazines
can
have
cytotoxic
effects,
they
were
characterized
as
anti-
microbial
compounds
that
destructively
abstract
electrons
from
the
transport
chain
[3].
It was
then
discovered
that
reducing
phenazines
can
greatly
benefit
Pseudomonas
aeruginosa
by
1)
regulating
gene
expression
during
quorum
sensing
by
oxidizing
a transcription
factor;
2)
act-
ing
as
alternative
terminal
electron
acceptors
to
promote
anoxic
survival;
and
3)
facilitating
nutrient
acquisition
[2,6,9–11]
These
reports
paint
a complex
picture
of
the
multifarious
effects
phenazines
can
have,
but
in
each
case,
the
conceptual
model
ends
with
the
cell
reducing
the
phenazine,
which
raises
the
question:
how
are
phenazines
recycled?
The
first
answer
is:
abiotically.
Phenazines
are
broadly
reactive
molecules
and
can
be
oxi-
dized
by
a variety
of
oxidants,
including
molecular
oxygen
and
manganese
or
iron
minerals.
When
oxygen
serves
as
the
electron
acceptor,
superoxide
is produced,
harming
both
phena-
zine
producers
and
other
cell
types
[12].
In
contrast,
when
iron
minerals,
to
which
phosphate
is adsorbed,
serve
as
the
electron
acceptor,
ferrous
iron
and
phosphate
can
be
released,
alleviat-
ing
nutrient
limitation
[11,13,14]
However,
not
all
oxidants
of
higher
redox
potential
(e.g.,
nitrate
and
nitrite)
react
quickly
enough
with
phenazines
to
re-oxidize
them
abiotically
on
bio-
logically
relevant
timescales.
Nonetheless,
bacteria
with
versatile
respiratory
electron
transport
chains,
such
as
the
soil
isolate
C
.
portucalensis
MBL,
can
catalyze
biological
phenazine
oxida-
tion
during
anaerobic
respiration
[15,16].
Moreover,
the
thermodynamics
of
biological
phena-
zine
oxidation
with
a variety
of
terminal
electron
acceptors
are
favorable
(see
our
theoretical
treatment
of
this
subject
in
the
supplement,
S1
Text).
These
observations
provide
a second
possible
answer
to
the
question
of
how
phenazines
are
recycled:
it stands
to
reason
that
in
anoxic
microenvironments
where
phenazine-reactive
abiotic
oxidants
are
limited,
phenazine-
reducing
bacteria
may
benefit
from
the
presence
of
phenazine-oxidizing
bacteria.
PLOS GENE
TICS
A generaliz
able
mechanism
for
biologic
al PCA
oxidation
PLOS
Genetics
| https://
doi.org/10.13
71/journal.p
gen.101
1064
May
6, 2024
2 / 25
within
the paper,
its Supporting
Information
files
and, as described
in the Materials
and Methods
section,
the entire
dataset
is available
on the
CaltechDAT
A repository
, at the DOI https://do
i.org/
10.22002/
tdng7-twd27.
All data and code
for
analysis
are available
on GitHub
(https://gith
ub.
com/ltsypin/C
portucale
nsis_genetic
_mech.
Funding:
This work
was supported
by the NSF
Graduate
Research
Fellowship
to LMZT;
by the NIH
(1R01AI12
7850-01A1
to DKN)
and by the Doren
Family
Foundation
to DKN.
The funders
had no role
in study
design,
data collection
and analysis
,
decision
to publish,
or preparation
of the
manuscript.
Competing
interests
:
The authors
have declared
that no competing
interests
exist.
Abbreviati
ons:
PCA,
phenazine-1-c
arboxylic
acid;
ETC, electron
transport
chain;
TEA, terminal
electron
acceptor;
NO3-,
nitrate;
NO2-,
nitrite;
N
2
O,
nitrous
oxide;
NO, nitric
oxide;
Fum
2-
, fumarate;
Succ
2
, succinate;
DMSO,
dimethyl
sulfoxide;
DMS,
dimethyl
sulfide;
TMAO,
trimethylami
ne-N-oxide;
TMA,
trimethyla
mine;
UQ, ubiquinone
; UQH
2
,
ubiquinol;
MQ, menaquinone
; MQH
2
, menaquinol;
DMQ,
demethylme
naquinone;
DMQH
2
,
demethylmen
aquinol.
Furthermore,
we
hypothesize
that
phenazine
oxidation
itself
might
provide
a survival
advan-
tage
to
phenazine-oxidizing
cells
under
anaerobic
conditions
where
electron
donors
are
limited.
To
test
these
ideas,
we
used
a systematic
genetic
approach
to
dissect
how
C
.
portucalensis
oxidizes
phenazine-1-carboxylic
acid
(PCA),
which
is naturally
produced,
widespread,
and
environmentally
relevant
[17].
We
focused
on
potential
catalysts
within
the
C
.
portucalensis
electron
transport
chain
(Figs
1A
and
2,
and
Table
1).
In
addition
to
leading
us
to
a generaliz-
able
mechanistic
model
for
how
biological
PCA
works,
the
development
of
a genetic
system
in
this
recently
isolated
soil
organism
exemplifies
how
the
adaptation
of
existing
tools
can
be
used
to
rapidly
gain
new
insights
into
microbial
processes
of
environmental
interest.
Results
Inactivation
of
individual
nitrate
reductases
yields
mild
PCA
oxidation
defects
To
investigate
the
mechanisms
and
dynamics
of
PCA
oxidation
by
C
.
portucalensis
MBL,
we
adapted
Escherichia
coli
genetic
engineering
protocols
[18,19].
Given
that
C
.
portucalensis
har-
bors
three
functionally
redundant
nitrate
reductase
complexes
(whose
catalytic
subunits
Fig
1.
Models
and
thermody
namics
of
PCA
oxidati
on.
(A) Potential
models
of PCA oxidation
in Gram-negative
bacteria
capable
of respiration
.
In
aqueous
environmen
ts,
PCA
redox
reactions
are
two-electr
on,
two-proton
processes
. Reduced
PCA
is shown
in
green,
reflecting
its
true
color.
PCA
oxidation
can
theoreticall
y be
coupled
to
the
respiratory
electron
transport
chain
in
a couple
ways:
(Category
1, electron
transfer
to
terminal
reductase)
PCA
may
donate
electrons
to
the
terminal
reductase
(grey
shape)
for
a respirable
terminal
electron
acceptor
(T.E.A.),
thus
contributing
two
protons
to
the
periplasm;
(Category
2, electron
transfer
to
quinol
pool)
PCA
may
donate
two
electrons
and
two
protons
to
quinones
, thus
regenerating
the
quinol
pool.
(Category
3, electron
transfer
to
terminal
electron
acceptor)
Alternatively
, PCA
directly
reduces
the
terminal
electron
acceptor.
This
may
happen
externally
to
the
cell,
or
PCA
may
enter
the
cytoplasm
and
react
with
the
terminal
electron
acceptor
independen
tly
of
the
electron
transport
chain,
as
depicte
d.
In
this
illustrati
on,
the
arrows
may
represent
direct
reactions
or
ones
mediated
by
enzymes
or
other
factors.
Transferri
ng
electrons
to
the
terminal
reductase
or
quinol
pool
(Categories
1 and
2)
represents
scenario
s that
may
be
energetica
lly
beneficial
for
a respiring
bacteria
l cell,
whereas
transferring
electrons
to
the
terminal
electron
acceptor
(Category
3)
may
be
detriment
al.
This
would
require
the
transport
of
PCA
across
the
inner
membran
e because
its
carboxylic
acid
moiety
is negative
ly
charged
at
circumneut
ral
pH,
and
it cannot
passively
cross
the
membran
e. For
simplic
ity,
this
illustrati
on
does
not
show
potential
reactio
ns
with
a periplasmi
c reductase
,
but
the
logic
would
remain
the
same,
only
with
no
involvement
of
the
cytoplasmic
space.
(B) Electron
tower of relevant
half-reactions
.
Reactions
are
ordered
by
their
relative
standard
midpoin
t potentials
with
more
negative
values
on
top
and
more
positive
ones
on
the
bottom
(not
to
scale).
Thermody
namically
favorable
pairings
comprise
more
positive
half-reactions
with
more
negative
ones
in
reverse.
The
theoreti
cal
limit
for
energy
that
can
be
conserve
d from
a
pairing
correlates
with
the
magnitud
e of
the
difference
in
half-reaction
potentials.
PCA:
phenazine-1-c
arboxylic
acid.
MQ:
menaquinon
e. DMQ:
demethylme
naquinone.
UQ:
ubiquinone.
Fum
2-
: fumarat
e. Succ
2-
: succinate.
TMAO:
trimethyl
amine-N-oxi
de.
TMA:
trimethyla
mine.
DMSO:
dimethyl
sulfoxide
. DMS:
dimethyl
sulfide.
NO
2
-
: nitrite.
NO:
nitric
oxide.
NO
3
-
: nitrate.
https://do
i.org/10.1371/j
ournal.pge
n.1011064.g00
1
PLOS GENE
TICS
A generaliz
able
mechanism
for
biologic
al PCA
oxidation
PLOS
Genetics
| https://
doi.org/10.13
71/journal.p
gen.101
1064
May
6, 2024
3 / 25
Fig
2.
Explicit
conceptua
l model
of
PCA
oxidation
during
anaerobic
nitrate
respiratio
n.
To
interpret
PCA
oxidation
phenoty
pes
during
anaerobic
nitrate
respiration
, it is necessary
to
keep
in
mind
the
distinct
electron
pathways
in
the
respirato
ry
(A)
and
periplas
mic
(B)
nitrate
reductase
complexes.
These
models
are
made
assuming
a complete
analogy
to
the
arrangement
of
these
proteins
in
E
.
coli
[25,
29,
30].
The
paths
of
the
electrons
along
the
reductase
comple
xes
are
not
shown
for
simplicity,
but
they
flow
from
the
quinols
to
the
nitrate
in
each
case.
The
categorie
s
of
PCA
interactions
are
numbered
according
to
the
scheme
in
Fig
1A
and
Table
1.
In
the
case
of
the
respiratory
nitrate
reductase
comple
xes,
there
PLOS GENE
TICS
A generaliz
able
mechanism
for
biologic
al PCA
oxidation
PLOS
Genetics
| https://
doi.org/10.13
71/journal.p
gen.101
1064
May
6, 2024
4 / 25
comprise
NapA,
NarG,
or
NarZ),
we
hypothesized
that
inactivating
any
one
of
these
enzymes
would
be
insufficient
to
abolish
PCA-oxidation
activity.
To
start,
we
compared
the
phenotypes
of
strains
in
which
we
perturbed
nitrate
reductases
by
two
different
methods:
λ
Red
recombina-
tion
or
oligo-mediated
recombineering
to
make
whole
operon
deletions
or
targeted
transla-
tional
knockouts
of
catalytic
subunits,
respectively
[18,19].
Fig
2 shows
the
three
nitrate
reductases
we
mutagenized,
as
well
as
their
predicted
orientations
in
the
ETC.
According
to
this
model,
deleting
the
entire
operon
or
knocking
out
just
the
catalytic
subunit
should
have
the
same
effect:
all
three
categories
of
PCA
oxidation
reactions
would
be
abolished.
Disrupting
any
single
nitrate
reductase
by
either
genetic
engineering
method
reduced
but
did
not
eliminate
PCA
oxidation.
We
observed
no
difference
between
the
narZ
translational
knockout
(
narZ-tlKO)
and
the
operon
deletion
(
Δ
narZYWV
)
in
the
rate
and
dynamics
of
PCA
oxidation
(Fig
3A).
Comparing
narZ-tlKO
and
Δ
narUZYWV
,
we
found
that
Δ
narUZYWV
had
a greater
delay
before
PCA
oxidation
commenced,
but
the
maximum
rate
of
PCA
oxida-
tion
was
the
same
as
for
narZ-tlKO
and
Δ
narZYWV
(Fig
3A).
All
three
mutants
oxidized
PCA
later
and
more
slowly
than
the
wildtype
(Fig
3A).
The
lag
in
oxidation
by
the
Δ
narUZYWV
strain
compared
to
the
Δ
narZYWV
and
narZ-tlKO
strains
may
reflect
the
fact
that
narU
is an
inner
membrane
nitrate-nitrite
antiporter
[20]:
its
absence
may
delay
nitrate
entrance
to
the
cytosol
where
it can
encounter
NarZ
or
NarG
(Fig
2).
As
observed
for
the
narZ
mutants,
there
was
no
difference
between
the
narG-tlKO
and
Δ
narGHJI
PCA
oxidation
phenotypes,
and
there
was
only
mild
loss
of
PCA
reduction
compared
to
the
wildtype
(Fig
3B).
The
napA-tlKO
strain
appeared
to
have
a more
severe
PCA
oxidation
defect
than
Δ
napFDAGHBC
versus
the
wildtype,
yet
loss
of
PCA
oxidation
by
these
strains
still
only
minimally
slowed
down
PCA
oxi-
dation
(Fig
3C).
For
the
rest
of
this
study,
we
used
oligo
recombineering
rather
than
λ
Red
recombination
to
generate
strains
with
multiple
gene
disruptions.
PCA
oxidation
dynamics
with
nitrate
depend
on
which
nitrate
reductases
are
present
Given
the
presence
of
three
nitrate
reductases
in
the
genome,
we
speculated
that
our
ability
to
alter
PCA
oxidation
dynamics
via
their
deletion,
alone
or
in
combination,
would
depend
on
their
expression
under
different
growth
conditions.
Accordingly,
we
compared
nitrate-driven
are
two
redundant
homologs
in
C
.
portucalensis
:
NarGHI
and
NarZYV
[42].
NarI/V
can
accept
electrons
from
all
three
types
of
quinones
[25].
For
the
periplasmi
c nitrate
reductase
, there
are
two
distinct
quinone
interaction
sites
(NapH
for
ubiquinone
and
NapC
for
menaquinon
e);
notably,
demethylme
naquinone
does
not
appear
to
play
a role
in
periplasmi
c nitrate
reductase
activity
[30].
The
NapAB
complex
is soluble
in
the
periplasmi
c space,
and
the
electrons
from
the
NapHGC
complex
are
ferried
to
NapA
by
NapB,
which
is a cytochro
me
c-type
protein
[29].
Category
1 (electron
transfer
to
terminal
reductase
) PCA
interactions
are
depicted
as
occurring
at
quinol-ox
idizing
subunit
s of
the
protein
complexes
(NarI,
NarV,
NapH,
and
NapC)
to
illustrate
the
hypothesis
that
a reduced
PCA
molecule
may
replace
a quinol.
Auxiliary
and
chaperone
proteins
that
are
members
of
the
nitrate
reductase
operons
and
are
involved
in
complex
formation
but
not
activity
(NarJ,
NarW,
NapF,
and
NapD)
are
not
illustrated
[29].
Note:
P
.
aeruginosa
possesse
s only
one
set
of
homologs
for
the
respiratory
nitrate
reductase
(NarGHI)
and
the
periplasmi
c nitrate
reductase
.
https://do
i.org/10.1371/j
ournal.pge
n.1011064.g00
2
Table
1.
The
categorie
s of
PCA
oxidati
on
reactions.
Type
Description
Category
1
PCA
reduces
the
terminal
reductase.
Category
2
PCA
reduces
a quinone,
thereby
replenishing
the
quinol
pool.
Category
3
PCA
reduces
the
terminal
electron
acceptor,
either
internall
y or
externally
to
the
cell
(i.e.,
abiotic
oxidation
of
PCA
by
the
terminal
electron
acceptor)
.
https://d
oi.org/10.1371/j
ournal.pge
n.1011064.t00
1
PLOS GENE
TICS
A generaliz
able
mechanism
for
biologic
al PCA
oxidation
PLOS
Genetics
| https://
doi.org/10.13
71/journal.p
gen.101
1064
May
6, 2024
5 / 25
Fig
3.
Comparison
of
the
roles
of
the
three
termina
l nitrate
reductases
in
PCA
oxidation
.
(A-C)
Compar
ison
of
homologou
s recombinat
ion
knockouts
versus
translational
knockouts
for
each
individual
terminal
nitrate
reductase.
Each
graph
shows
the
oxidation
of
PCA
over
time
(measured
by
the
decay
of
PCAred,
which
is fluorescent)
with
the
abiotic
control
(nitrate
plus
reduced
PCA
in
reaction
medium
without
cells)
in
grey.
(A)
NarZ
comparison
: the
deletion
(
Δ
)
and
translati
onal
knocko
ut
(tlKO)
strains
all
have
a PCA
oxidatio
n deficit
relative
to
the
wildtype
(WT,
blue).
The
most
severe
phenoty
pe
is in
the
Δ
narUZYWV
strain
(green),
and
the
Δ
narZYWV
and
narZ
-tlKO
(yellow
and
orange,
respectively)
phenotype
s are
indistingui
shable.
(B)
NarG
comparison
: the
deletion
(green)
and
translati
onal
knockout
(orange)
strains
have
the
same
slight
PCA
oxidatio
n deficit
relative
to
the
wildtype
control
(WT,
blue).
(C)
NapA
comparison
s: the
translational
knockout
(orange)
has
a more
severe
PCA
oxidation
deficit
that
the
deletion
strain
(green),
relative
to
the
wildtype
control
(blue).
Each
thick
line
corresponds
to
the
mean
of
three
biological
replicates
plotted
in
semitranspa
rent
circles.
(D-E)
Comparisons
of
the
maximum
observed
PCA
oxidatio
n rate
for
each
combina
torial
nitrate
reductase
mutant
(and
abiotic
control).
(D)
The
phenot
ypes
after
shaking
(oxic)
pregrowth
. (E)
The
phenotype
s after
standing
(hypoxic)
pregrowth
. (F-G)
Compar
isons
of
the
time
until
half
of
the
PCA
was
oxidized
for
each
combinatori
al nitrate
reductase
mutant
(and
abiotic
control).
(F)
The
phenot
ypes
after
oxic
pregrowth
. (G)
The
phenoty
pes
after
hypoxic
pregrowth
. N.D.
stands
for
“not
detected,”
correspond
ing
to
strains
that
did
not
reach
the
50%
PCA
oxidized
threshold
over
the
48-hour
assay.
Squares
represent
the
means
of
techni
cal
triplicates
and
circles
are
independen
t biological
replicates
with
no
technical
replicates.
Double
asterisks
repres
ent
statistical
significan
ce
after
Bonferron
i correctio
n,
and
single
asterisks
represent
p
<
0.05
before
Bonferron
i correctio
n.
The
statistical
testing
is described
in
the
supplementa
ry
methods
(S1
Text),
and
the
complete
matrices
of
pairwise
comparison
s are
presented
in
S2
Fig.
The
conversion
of
PCA
oxidation
curves
to
rate
and
time
metrics
is described
in
S3
Fig.
https://do
i.org/10.1371/j
ournal.pge
n.1011064.g00
3
PLOS GENE
TICS
A generaliz
able
mechanism
for
biologic
al PCA
oxidation
PLOS
Genetics
| https://
doi.org/10.13
71/journal.p
gen.101
1064
May
6, 2024
6 / 25
PCA
oxidation
dynamics
after
two
different
pregrowth
conditions
in
lysogeny
broth
(LB):
slanted
shaking
overnight
tubes
and
standing
parafilm-sealed
overnight
tubes.
These
two
con-
ditions
permitted
fully
aerated
(oxic)
and
hypoxic
growth,
respectively.
Depending
on
which
cultivation
condition
was
used,
distinct
nitrate
reductases
dominantly
contributed
to
PCA
oxi-
dation
(Fig
3D–3G).
After
oxic
pregrowth
(Fig
3D
and
3F),
the
narZ
knockout
had
the
most
severe
phenotype
of
the
single
mutants,
and
eliminating
both
narG
and
narZ
was
sufficient
to
abolish
PCA
oxida-
tion,
implying
that
under
this
condition
napA
is irrelevant
(Fig
3D).
The
time
it took
for
the
different
strains
to
oxidize
50%
of
the
provided
reduced
PCA
was
correlated
with
the
average
maximum
PCA
oxidation
rate,
and
the
narZ
single
knockout
was
again
the
slowest
among
the
other
single
mutants
(Fig
3F).
The
abiotic,
narGnarZ
double
knockout,
and
napAnarZnarG
triple
knockout
conditions
never
reached
this
50%
threshold
(Fig
3F).
Relative
to
the
wildtype,
only
the
narZ
single
knockout
strain
had
a significantly
different
time
to
50%
oxidation
(Figs
3F,
and
S2A
and
S2C).
Cells
pre-grown
in
stationary
cultures
provided
a subtle
but
important
contrast
to
the
shak-
ing
pregrowth
results
(Fig
3E
and
3G).
Here,
each
of
the
double
knockouts
had
a detectable
PCA
oxidation
rate,
indicating
that
any
of
the
three
nitrate
reductases
can
drive
the
oxidation
(Fig
3E).
Moreover,
rather
than
narZ
,
the
greatest
loss
of
oxidation
in
a single
mutant
back-
ground
came
from
the
narG
(Fig
3E).
Once
again,
the
time
to
50%
oxidation
was
correlated
to
the
maximum
oxidation
rate
(Fig
3G),
but
in
this
condition,
the
narZ
single
mutant
did
not
delay
reaching
this
threshold
relative
to
the
wildtype
activity
(S2B
and
S2D
Fig).
These
results
imply
that
the
pregrowth
condition
determines
the
relative
presence
and
activity
of
the
differ-
ent
nitrate
reductases
prior
to
the
oxidation
assay,
leading
to
different
phenotypes
for
a given
strain.
Because
pre-growth
in
stationary
cultures
allowed
us
to
observe
the
contributions
of
each
of
the
three
nitrate
reductases,
we
employed
this
condition
for
the
rest
of
the
genetic
experiments,
including
complementation
assays
(S4
Fig).
When
the
nitrate
reductases
were
individually
overexpressed
in
the
triple
knockout
background
during
overnight
stationary
pre-growth
(S4A
Fig),
only
the
narZ
overexpression
strain
had
a statistically
significant
rescue
of
PCA
oxidation
(S4B
Fig).
The
rescue
effect
was
small,
and
the
napA
and
narG
overexpres-
sion
strains
did
shift
toward
higher
oxidation
rates,
so
this
is likely
due
to
the
overexpression
system
being
unoptimized
for
C
.
portucalensis
MBL.
Summarily,
PCA
oxidation
was
abolished
only
when
all
three
terminal
nitrate
reductases
were
knocked
out.
The
PCA
oxidation
dynamics
and
phenotypes
can
generally
be
summarized
by
the
time
it
takes
to
reach
a threshold
concentration
of
reduced
PCA
(e.g.,
S3A
Fig)
and
the
maximal
PCA
oxidation
rate
(e.g.,
S3B
Fig),
but
this
can
obscure
some
more
nuanced
phenotypes,
such
as
the
biphasic
nature
of
many
of
the
PCA
oxidation
curves
(e.g.,
Fig
3A,
narZ-tlKO
).
While
the
fit
and
oxidation
metrics
depend
on
the
parametrization
of
the
model,
we
were
able
to
identify
a
conservative
parametrization
that
robustly
represented
all
tested
strains
(S3C–S3F
Fig).
For
the
remaining
experiments,
we
report
only
the
maximum
PCA
oxidation
rates
as
the
pheno-
types
of
interest.
Loss of quinones
in
C.
portucalensis
MBL significantly
disrupts
PCA
oxidation
,
yet mild oxidation
is still achieved
by nitrate
reductases
in their
absence
Nitrate-driven
PCA
oxidation
by
C
.
portucalensis
is fully
abolished
only
when
all
three
termi-
nal
nitrate
reductases
are
knocked
out;
any
one
remaining
reductase
enables
PCA
oxidation
(Fig
4D).
This
observation
raises
the
question
whether
the
nitrate
reductases
oxidize
PCA
directly
(Fig
2 and
Table
1,
Category
1 reactions)
or
if PCA
contributes
electrons
to
an
PLOS GENE
TICS
A generaliz
able
mechanism
for
biologic
al PCA
oxidation
PLOS
Genetics
| https://
doi.org/10.13
71/journal.p
gen.101
1064
May
6, 2024
7 / 25
upstream
pool,
such
as
the
quinols
(Fig
2 and
Table
1,
Category
2 reactions).
PCA
oxidation
was
largely
lost
when
the
biosynthesis
of
all
three
quinones
(ubiquinone,
menaquinone,
and
demethylmenaquinone)
was
disrupted
in
the
menAubiC
-tlKO
strain,
indicating
that
PCA
oxi-
dation
by
quinones
is the
main
component
of
the
PCA
oxidation
rate
(Fig
4A).
However,
the
PCA
oxidation
rate
in
this
genetic
background
did
not
drop
to
abiotic
levels,
indicating
that
cellular
nitrate
reduction
can
drive
PCA
oxidation
at
a low
rate
without
quinones
as
intermediaries
(Fig
4A).
In
other
words,
PCA
oxidation
by
terminal
reductases
also
can
occur.
The
loss
of
ubiquinones
alone
in
the
ubiC-tlKO
background
did
not
affect
PCA
oxidation,
but
the
loss
of
(demethyl)menaquinon
es
in
the
menA
-
tlKO
strain
did
have
an
effect
(Fig
4A
and
4B).
In
the
quinone-null
background,
PCA
oxidation
persists
with
any
one
of
the
nitrate
reductases
(Fig
4A),
as
evidenced
by
the
positive
oxidation
rate
in
the
menAubiCnapAnarZ-
,
menAubiCnarGnapA-
,
and
menAubiCnarGnarZ-tlKO
strains.
Interestingly,
the
strains
with
no
quinones
and
only
intact
NarG
or
NapA
had
faster
PCA
oxidation
rates
than
the
menAubiC
strain
on
its
own
(Fig
4A
and
4B).
The
co-occurrence
of
PCA
transferring
electrons
to
both
terminal
reductases
and
quinones
(Category
1 and
2 reactions)
implies
that
PCA
oxidation
may
be
plastic:
it does
not
necessarily
depend
on
a specifically
evolved
enzyme
or
pathway
to
proceed.
We
were
unable
to
genetically
complement
the
loss
of
quinones
during
nitrate-driven
PCA
oxidation
(S5A
and
S5B
Fig).
Curiously,
overexpressing
menA
exacerbated
the
menAubiC-
tlKO
phenotype
(S5A
Fig).
Because
each
nitrate
reductase
retains
some
PCA
oxidation
activity
even
in
the
absence
of
quinones
(Fig
4A),
it may
be
the
case
that
exogenously
expressing
a low
level
of
quinones
is not
enough
to
give
a signal
over
the
independent
nitrate
reductase
activity.
However,
our
greater
degree
of
success
with
complementing
menA
for
the
other
terminal
Fig
4.
Nitrate-drive
n PCA
oxidati
on
rates
in
quinone
knockout
backgroun
ds,
including
single
nitrate
reductase
strains.
(A)
Maxim
um
PCA
oxidation
rates.
Squares
represent
the
means
of
techni
cal
triplicates
and
circles
represent
independen
t biological
replicat
es.
(B)
Pairwise
comparison
s between
differences
of
mean
maximu
m
PCA
oxidation
rates
for
the
strains
in
(A).
Given
28
comparison
s, the
Bonferroni-c
orrected
thresho
ld
for
significan
ce
is p
<
0.00179
.
https://do
i.org/10.1371/j
ournal.pge
n.1011064.g00
4
PLOS GENE
TICS
A generaliz
able
mechanism
for
biologic
al PCA
oxidation
PLOS
Genetics
| https://
doi.org/10.13
71/journal.p
gen.101
1064
May
6, 2024
8 / 25
electron
acceptors
indicates
that
the
system
works
in
principle
(see
the
alternative
TEA
section,
below).
Pseudomonas
aeruginosa
PA14 replicates
nitrate
reductase
knockout
phenotypes
from
C.
portucalensis
MBL
Given
the
apparent
generalizability
of
PCA
oxidation
to
the
three
nitrate
reductases
in
C
.
por-
tucalensis
MBL
and
the
observation
that
other
bacteria
perform
the
same
metabolism
(15,29),
we
assessed
whether
P
.
aeruginosa
PA14
nitrate
reductase
mutants
conform
to
the
same
mech-
anistic
model.
P
.
aeruginosa
is particularly
relevant
as
a point
of
comparison
because
it bio-
synthesizes
PCA,
has
been
studied
extensively
as
a PCA
reducer,
and,
like
C
.
portucalensis
MBL,
also
has
both
respiratory
and
periplasmic
nitrate
reductases
(albeit
only
one
homolog
of
the
respiratory
nitrate
reductase
complex,
NarGHI)
(Fig
2)
[6,21,22].
We
compared
wildtype,
Δ
narG
,
and
Δ
napAB
strains
of
P
.
aeruginosa
PA14
after
pregrowth
in
shaking
and
standing
conditions,
as
we
had
done
for
C
.
portucalensis
MBL.
We
observed
that
in
the
shaking
preg-
rowth
condition,
the
Δ
narG
strain
had
no
phenotype
versus
the
wildtype,
while
the
Δ
napAB
strain
had
a severe,
though
incomplete,
PCA
oxidation
defect
(Fig
5,
left).
In
contrast,
after
either
of
the
standing
pregrowth
treatments,
the
Δ
napAB
strain
did
not
substantially
differ
from
the
wildtype,
while
the
Δ
narG
strain
had
severe,
though
incomplete,
loss
of
PCA
oxida-
tion
(Fig
5,
right).
This
pattern
of
distinct
nitrate
reductases
dominating
PCA
oxidation
activ-
ity
depending
on
the
pre-growth
condition
corresponds
to
the
fact
that
napA
and
narG
are
regulated
by
distinct
systems:
RpoS
and
Anr,
respectively
[23,24],
meaning
that
the
P
.
aerugi-
nosa napA
gene
is regulated
like
the
C
.
portucalensis
narZ
gene.
The
P
.
aeruginosa
phenotypes
thus
correspond
to
the
difference
we
observed
between
C
.
portucalensis
MBL
narG
and
narZ
translational
knockouts
depending
on
pre-growth
conditions
(Fig
4C
and
4D),
which
implies
that
the
PCA
oxidation
mechanism
is not
specific
to
a given
organism
or
enzyme
but
rather
the
architecture
of
the
electron
transport
chain.
Fig
5.
Pre-growth
dependent
phenotypes
for
P.
aeruginosa
nitrate
reductase
mutants.
Left:
When
cultures
were
pre-grown
in
slanted
shaking
tubes,
providing
the
cultures
ample
oxygen,
only
the
Δ
napAB
strain
(orange)
had
a PCA
oxidation
deficit.
Right:
Pre-growi
ng
the
cultures
in
standing
tubes
without
nitrate
supplementati
on
also
leads
to
the
Δ
narG
strain
(green)
to
have
the
PCA
oxidation
deficit;
the
wildtype
(grey)
and
Δ
napAB
(orange)
strains
are
within
error
of
each
other.
Each
thick
line
correspond
s to
the
mean
of
three
biological
replicates
that
are
plotted
in
semitransp
arent
circles.
The
mediu
m
for
this
experime
nt
was
84%
LB
and
16%
basal
PCA
oxidation
assay
mediu
m
(see
material
s and
methods)
. Note:
due
to
its
relative
inability
to
grow
under
hypoxia
in
LB,
it was
not
feasible
to
grow
a
nap/nar
double
knockout
P.
aerugi
nosa
strain
to
compa
re
to
the
C
.
portucalensis
MBL
full
nitrate
reductase
knockout.
Similarly,
we
could
not
generate
any
quinone
knockouts
for
P
.
aeruginosa
because
it encodes
the
biosynthesis
of
only
ubiquinone
(24).
https://do
i.org/10.1371/j
ournal.pge
n.1011064.g00
5
PLOS GENE
TICS
A generaliz
able
mechanism
for
biologic
al PCA
oxidation
PLOS
Genetics
| https://
doi.org/10.13
71/journal.p
gen.101
1064
May
6, 2024
9 / 25