An iron (II) dependent oxygenase performs the last missing step of plant lysine catabolism

Article abstract of DOI:10.1038/s41467-020-16815-3

Despite intensive study, plant lysine catabolism beyond the 2-oxoadipate (2OA) intermediate remains unvalidated. Recently we described a missing step in the D-lysine catabolism of Pseudomonas putida in which 2OA is converted to D-2-hydroxyglutarate (2HG)

Full text of DOI:10.1038/s41467-020-16815-3

A
R
T
I
C
L
E
https://doi.org/10.1038/s41467-020-16815-3
OPEN
A
m
n
i
s
r
i
o
n
n
g
(
s
I
t
I
)
e
d
e
o
p
f
e
p
n
l
d
a
e
n
n
t
t
l
o
x
y
g
e
e
c
n
a
a
t
s
a
e
b
p
e
r
f
m
o
r
m
s
t
h
e
l
as
t
i
s
p
y
s
i
n
o
l
i
s
1 , 2 , 3 , 1 5
1 , 2 , 4 , 1 5
1
,
5
,
1
5
M i t c h e l l G . T h o m p s o n
, J a c q ue l y n M . B la k e - He d g e s
, J o s e H e n r i q u e P e re i r a
,
4
1 , 2 , 3
1 , 2 , 3
1 , 2 , 6
J oh n A . H a n g a sk y
, M i c ha e l S . B e l c h e r
, W il l i a m M . M o o r e , J e s u s F . B a ra j a s
,
1 , 2
1 , 2 , 3
1 , 2
1 , 2 , 7
P a b l o C r u z - M o ra l es , L or e n z o J . W a s hi n gt o n , R o b e rt W . H a u s h a l t e r , C h r is t o p h e r B . E i b e n
,
1
,
2
4
1
,
2
3
1 , 2
Y
u
z
h
o
n
g
L
i
u
,
W
i
l
l
S
k
y
r
u
d
,
V
e
r
o
n
i
c
a
T
.
B
e
n
i
t
e
s
,
T
y
l
e
r
P
.
B
a
r
n
u
m
,
E
d
w
a
r
d
E
.
K
.
B
a
i
d
o
o
,
1 , 2 , 3
4 , 8
1
,
2
,
9
,
1
0
,
1
1
1
,
5
,
7
H e nr i k V . S c h e l l e r
J a y D . K e a s li n g
,
M
i
c
h
a
e
l
A
.
M
a
r
l
e
t
t
a
,
P
a
t
r
i
c
k
M
.
S
h
i
h
,
P
a
u
l
D
.
A
d
a
m
s
&
✉
1
,
2
,
7
,
1
2
,
1
3
,
1
4
D e s p i t e i nt e n si ve s t u d y , p l a n t l y si n e c a ta b o l i s m b e y o n d t h e 2 - o x oa di pa t e ( 2 OA ) i n t e r m e di a t e
r e m a i n s u n va li d a te d . R e ce n t l y w e d e s c r i b e d a m i s s i n g s t e p i n t h e D - l y s i n e c a t a b o l i s m o f
Pseudomonas putida i n w h i c h 2 O A i s c o n v e r t e d t o D - 2 - h y d ro x y g l u t a r a te ( 2 H G ) v i a h y d ro -
x y g l u t a ra t e s yn t h a se ( H g l S) , a D U F 1 3 3 8 f a m i l y p r o t e i n . H e r e w e s o l v e t h e s t r u c t ur e o f H g l S
t o 1 . 1 Å r e so l u t i o n i n s u b s tr a te - f r e e f o r m a n d i n c o m p l ex w i t h 2 O A . W e p ro p os e a s u c c e s s i v e
d e c a r b o x yl a t i on a nd i n tr a m o l ec u l a r h y d r o x y l a t i o n m e c ha ni s m f or m i ng 2 H G i n a F e( I I )- a n d
O - d ep e n d e n t m an n e r . S p e c i fic i ty i s m e d i a te d b y a s i n g l e a r g i n i n e , h i g h l y c o n s e rv e d a c r o s s
2
m o s t D U F1 3 38 p ro t e i n s . A n Arabidopsis thaliana H g l S h om o l o g c o e xp re s s e s w i t h k n o wn l y s i n e
c a t a b o l is m e n z ym e s , a n d m u t a n ts s h o w p h e n o ty p e s c o n s i s t en t w i t h d i s r u pt e d l y s i n e c a t a -
b o li sm . S t ru c t u r a l a nd b i o c h e m i c a l a n a l y s i s o f Oryza sativa h om o l og F L O7 r ev e a l s i d e n t i c a l
a c t i v i t y t o H g l S d e s p i t e l o w s e q u e n c e i d e n t i t y . O u r r es ul t s s u g g e s t D U F 1 33 8 - c o nt a i ni ng
e n z y m e s c a t a ly z e t h e s a m e b i o c h e m i c a l r e a ct i o n , e x e r t i n g t h e s a m e p hy s i o l o g i c a l f un c t i o n
a c ro s s b ac t e r ia a nd e u k a r y o t e s .
1 Joint BioEnergy Institute, Emeryville, CA, USA. 2 Biological Systems & Engineering Division, Lawrence Berkeley National Laboratory, Berkeley, CA, USA.
3 Department of Plant and Microbial Biology, University of California-Berkeley, Berkeley, CA, USA. 4 Department of Chemistry, University of California-
Berkeley, Berkeley, CA, USA. 5 Molecular Biophysics and Integrated Bioimaging, Lawrence Berkeley National Laboratory, Berkeley, CA, USA. 6 Department of
Energy Agile BioFoundry, Emeryville, CA, USA. 7 Department of Bioengineering, University of California-Berkeley, Berkeley, CA 94720, USA. 8 Department of
Molecular and Cellular Biology, University of California-Berkeley, Berkeley, CA, USA. 9 Department of Plant Biology, University of California-Davis, Davis, CA,
USA. 10 Genome Center, University of California-Davis, Davis, CA, USA. 11 Environmental Genomics and Systems Biology Division, Lawrence Berkeley
National Laboratory, Berkeley, CA, USA. 12 Department of Chemical and Biomolecular Engineering, University of California-Berkeley, Berkeley, CA, USA.
13 The Novo Nordisk Foundation Center for Biosustainability, Technical University of Denmark, Lyngby, Denmark. 14 Center for Synthetic Biochemistry,
Shenzhen Institutes for Advanced Technologies, Shenzhen, China. 15These authors contributed equally: Mitchell G. Thompson, Jacquelyn M. Blake-Hedges,
✉
Jose Henrique Pereira. email: jdkeasling@lbl.gov
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
(
2
0
2
0
)
1
1
:
2
9
3
1
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
8
1
5
-
3
|
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s
1

A
R
T
I
C
L
E
N A T U R E C O M MU N ICA T IO N S | h t t p s : / /d o i .o r g / 1 0 .1 0 3 8/ s 41 46 7 - 0 2 0- 16 8 1 5 - 3
ysine is an essential amino acid, and due to its low abun- followed by the hydroxymandelate synthase (HMS) and 4-hydroxy-
dance in cereals and legumes it is produced on a scale of one phenyl pyruvate dioxygenase (HPPD) structures (Supplementary
million tons a year to supplement food supply needs1–3. To Data 1)11,12. HglS and HMS structure comparison revealed the two
L
thwart malnutrition in the developing world, significant work has enzymes share a similar central β-sheet fold common to the
been done to engineer rice, maize and other plants to produce DUF1338 structures (Supplementary Fig. 2)11,13. More importantly,
greater quantities of lysine3,4. Increasing lysine levels in cereal the β-sheet domain of HMS contains the enzyme active site, two
grains requires overexpression of lysine-producing enzymes and histidines and a glutamate that bind the metal cofactor in nearly the
concurrent disruption of lysine catabolism3. Thus, mutants such same orientation as in HglS. The similarity of the HglS, HMS, and
as opaque2 in maize have received considerable attention for their HPPD folds and active sites suggested that HglS is likely an
ability to accumulate lysine within their endosperm5,6. However, additional member of the vicinal oxygen chelate (VOC) enzyme
despite worldwide importance, the full plant lysine catabolism superfamily and could act via a similar mechanism to HMS and
pathway remains unknown, with no consensus in the steps HPPD14,15
beyond 2-oxoadipate (2OA) formation7.
To determine metal cofactor identity within the HglS structure,
.
Recently we described a novel D-lysine catabolic route in we conducted a fluorescent scan of a HglS crystal. We detected
the bacterium Pseudomonas putida which also contains a a photon emission near 7475 eV, the K-alpha emission of
2OA intermediate, similar to plant L-lysine catabolism8. In nickel (Supplementary Fig. 3). In addition, a less intense peak
the P. putida pathway, 2OA is converted to 2-oxoglutarate corresponding to the K-alpha emission of iron was also observed.
(2OG) via three enzymes, one of which catalyzes a unique We therefore assigned the co-crystallized metal as Ni(II) with a
decarboxylation–hydroxylation step. In this step, 2OA is con- small percent Fe(II) occupancy. While the metal cofactors present
verted to D-2-hydroxyglutarate (2HG) in a reaction catalyzed by in the HMS and HglS structures differ, we previously determined
the Fe(II)-dependent DUF1338 family enzyme hydroxyglutarate that HglS utilizes Fe(II), not Ni(II), for catalysis. The nickel
synthase (HglS) (Supplementary Fig. 1)8. Homologs of this bound to HglS is likely derived from the nickel affinity
enzyme are broadly distributed across multiple domains of life, chromatography protein purification.
including nearly every sequenced plant genome8,9. However,
only one study describing DUF1338 enzymes in plants has been the VOC superfamily, mandating
The HglS domain architecture suggests the enzyme belongs to
a
chemical mechanism
reported. In Oryza sativa FLO7 (a DUF1338 family member) employing the bidentate coordination of vicinal oxygen atoms
mutants, abnormal starch formation was observed in the endo- to a divalent metal center. Given the established HMS substrate-
sperm similar to maize opaque2 mutants that are known to binding mode, we hypothesized that the metal in HglS would
accumulate high levels of lysine9. The widespread DUF1338 bind the vicinal oxygen atoms of the α-keto group of 2OA11. We
family abundance in plants and the FLO7 and opaque2 pheno- therefore soaked HglS crystals with 2OA and solved the structure
types encouraged us to further investigate whether enzymes of the resulting complex. The noncatalytic active site nickel
throughout the family displayed similar activity to HglS.
prevented enzyme turnover, yielding a substrate-bound structure,
Conversion of 2OA directly to D-2HG requires two discrete and no observed product density. As hypothesized, the substrate
chemical steps, a decarboxylation and hydroxylation, making the carboxylate and α-keto group oxygens are coordinated to the
chemical mechanism of the enzyme puzzling8. Here, we leverage metal (Fig. 1b), classifying HglS as a VOC superfamily member.
structural and biochemical analyses to postulate a chemical
We had previously noted the similarity between the set of
mechanism for HglS and characterize its substrate specificity. We reactions catalyzed by HglS and HMS8,11,13. Both enzymes perform
show that critical residues involved in catalysis are highly con- a decarboxylation–hydroxylation reaction on a α-ketoacid substrate.
served across nearly all DUF1338 family proteins. We further HPPD catalyzes a similar overall reaction with the hydroxylation
show that despite very low sequence identity, plant homologs also occurring at a different position. The structural and biochemical
catalyze the conversion of 2OA to 2HG and adopt the same similarity of HglS to HMS and HPPD led us to propose a similar
structural fold as the bacterial enzyme, suggesting that DUF1338 chemical mechanism for HglS (Fig. 1c). More specifically, the
family proteins catalyze the last unknown step of plant lysine catalytic cycle begins when the substrate carboxylate and α-keto
catabolism.
group oxygens and molecular oxygen bind to the three open Fe(II)
coordination sites. A radical rearrangement results in a decarbox-
ylation and the formation of a Fe(IV)-oxo and an α-radical species.
Finally, continued radical rearrangement produces the hydroxylated
Results
Structural analysis reveals the catalytic mechanism of HglS. To product 2-hydroxyglutarate. If HglS follows this mechanism, it
better understand the unusual HglS reaction, we obtained crystal should consume 1 mol of O2 per mole of 2OA, and two oxygens
structures of the enzyme both with and without a bound substrate. in the product should derive from molecular O2. To support
Initially, Hg1S was crystallized without substrate, and the Hg1S this mechanism, we determined the stoichiometry of 2OA to O2
structure was solved at 1.1 Å resolution (Fig. 1a). The enzyme consumption using a dissolved oxygen probe. In the presence of
possesses a central β-sheet motif resembling a partially-closed β- 100 or 200 µM 2OA, the enzyme consumed approximately an
barrel consisting of seven β-sheets, which is conserved in the three equimolar concentration of O2 (Fig. 1d). No feedback inhibition
other DUF1338 structures deposited in the Protein Data Bank was observed with 1 mM D-2HG, or L-2HG (Supplementary
(PDB IDs 3LHO, 3IUZ, and 2RJB). Further analysis also revealed Fig. 4a), and kinetic parameters determined by monitoring oxygen
the presence of a metal cofactor bound within a conserved metal consumption were similar to those we previously reported using a
cofactor-binding motif—consisting of the residues His 70, His 226, different enzyme assay (Supplementary Fig. 4b)8. We next tested
and Glu 294—common to the DUF1338 structures. However, our hypothesis that two oxygen atoms in the 2HG product derive
none of the available DUF1338 structures have been biochemically from molecular O2. To determine the source of oxygen atoms,
characterized or solved in complex with a substrate, and conse- additional enzyme assays were performed under an 18O2 atmo-
quently no chemical reaction mechanism for the family has been sphere. High resolution LC-MS product analysis revealed a species
proposed.
exhibiting a m/z of 151.022 that co-eluted with a 2HG standard
Therefore, we performed a search for characterized proteins with (m/z 147.029), corresponding to the expected m/z of 2HG
similar structures using the Vector Alignment Search Tool (VAST)10. containing two 18O-labeled oxygens (Fig. 1e). These results strongly
The top VAST hits were the three reported DUF1338 structures, suggest HglS proceeds via a mechanism similar to HMS and is
2
N A T UR E C O M M U N I C A TI O N S |
(
2
0
2
0
)
1
1
:
2
9
3
1
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
8
1
5
-
3
|
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s

N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
8
1
5
-
3
A
R
T
I
C
L
E
a
180
b
c
O
O
O
O
O
O
O
O
H
H
O
O
O
O
H226
E294
O
H226
E294
O
H226
E294
O2
Fe
Fe
Fe
O
O
O
O
O
O
H226
E294
H70
H70
H70
H70
CO2
Ni
R74
O
O
O
O
O
O
HO
OH
Fe
O
H
2OA
O
H226
E294
O
H226
E294
O
H226
E294
H
H
H
Fe
Fe
O
O
O
H70
H70
H70
e
d
250
200
150
100
m/z 147.029
2OA added
100 mM
m/z 151.022
200 mM
Ambient O2
18O2
50
0
αHG standard
4.5
5
5.5
6
4.5
5
5.5
6
10
20
Time (s)
30
40
50
Time (min)
Time (min)
g
f
h
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
E294
H70
0.7
H226
Ni
0.6
0.5
0.4
0.3
0.2
0.1
0.0
2OA
Enzyme
WT
V402
R74
V402P
R74A
S403
V402
S403
0.0
0.2
0.4
0.6
0.8
1.0
2-Oxoadipate (mM)
therefore the third identified member of the two-substrate α- approximately 10 Å from the holo state to the enzyme-bound
ketoacid-dependent oxygenase (2S-αKAO) family14.
state, bringing it into proximity with the substrate. We
Further analysis of the co-crystal structure revealed several investigated the importance of Arg74 in determining enzyme
other substrate-binding residues. Specifically, arginine 74 forms a specificity by assaying enzyme activity against a panel of α-
salt bridge with the distal carboxylate of the 2OA substrate. In ketoacids (Fig. 1g). HglS exhibited activity only with 2OA, and
addition, valine 402 and serine 403 form hydrogen bonds with the displayed no detectable and statistically significant activity above
carboxylate keto group and distal carboxylate group, respectively background (a boiled enzyme control) on any other substrate.
(Fig. 1f). The loop containing residues 402 and 403 shifts We therefore concluded that Arg74 participates in favorable
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
(
2
0
2
0
)
1
1
:
2
9
3
1
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
8
1
5
-
3
|
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s
3

A
R
T
I
C
L
E
N A T U R E C O M MU N ICA T IO N S | h t t p s : / /d o i .o r g / 1 0 .1 0 3 8/ s 41 46 7 - 0 2 0- 16 8 1 5 - 3
Fig. 1 Structural and Biochemical Analyses of P. putida HglS. a R i b b o n d i a gr a m o f t h e H g l S c r y s t a l s t r u ct u r e . T h e l e f t i m a ge s h o w s t h e a c t i v e s i t e
e n t r a n c e , c o n t ai n i n g t h e m et a l c o fa c t o r ( r e d s p h e r e ) a n d m e t a l - c o o r d i n at i n g r e s i d u e s ( o r a n g e ) l o c a t e d i n t h e c e n t r a l β- s h e et p r o t e i n d o m a i n . A 1 8 0 °
r ot a ti o n o n t h e r i g h t s h o ws t h e o v e r a l l p r o t e i n s t r u c t u r e . b O ve r l a y o f holo ( c y a n ) a n d s u b s t r a t e - b ou n d ( g r e e n ) s t r u ct u r e s o f H g l S d i s p l a y i n g t h e e n z y m e
a c ti v e s i t e a n d t h e i o n i c i n t e r a c t i o n b e t we e n A r g 74 a n d t h e d i s t a l c a r b o x y l a t e g r o u p o f t h e 2 O A s u b s t r a t e . T h e n i c k e l b o u n d w i t h i n e a c h s t r uc t ur e i s
d
i
s
p
l
a
y
e
d
a
s
a
r
e
d
(
h
o
l
o
s
t
r
u
c
t
u
r
e
)
o
r
p
u
r
p
l
e
(
s
u
b
s
t
r
a
t
e
-
b
o
u
n
d
s
t
r
u
c
t
u
r
e
)
s
p
h
e
r
e
.
c
P
r
o
p
o
s
e
d
r
e
a
c
t
i
o
n
m
e
c
h
a
n
i
s
m
o
f
H
g
l
S
.
O
-
d
e
r
i
v
e
d
o
x
y
g
e
n
s
a
r
e
s
h
o
w
n
i
n
2
r e d . d O x y g en t o 2 O A s t o ic h i o m et r y o f t h e H gl S r e a c t io n. D is s o l v e d o x y g en c o n c e n t ra t i o n ( y- a x i s ) w a s m e a s u r e d w i t h a C l a r k - t y p e o x y ge n p r ob e . E i t he r
2 00 µM ( r ed l i n e ) o r 1 0 0 µM ( b lu e l in e) o f 2 OA w a s a d d e d t o i n i t i a t e r e a ct i o n , r e s u l t i n g i n e q u i m o l a r o x y ge n c o n s u m pt i o n . D o t t e d l i n e s r e p r e s e n t t h e
e
x
p
e
c
t
e
d
fi
n
a
l
O
c
o
n
c
e
n
t
r
a
t
i
o
n
w
h
e
n
f
1
0
0
µ
M
(
b
l
u
e
l
i
n
e
)
o
r
2
0
0
µ
M
(
r
e
d
l
i
n
e
)
2
O
A
i
s
c
o
n
s
u
m
e
d
.
e
L
C
-
H
R
M
S
e
x
t
r
a
c
t
e
d
i
o
n
c
h
r
o
m
a
t
o
g
r
a
m
s
(
E
I
C
s
)
2
1
8
s ho w in g l a b el ed O i n c o r p o r a tio n i n t o t h e 2 - h y d r o x y gl u ta r a t e p r o d u c t o f H g l S . O n t h e l e f t ( b l a c k l i n e s) a r e E IC s f o r i o n s w i t h m/ z 1 4 7 .0 2 9 , r e pr e s e nt i ng
1
6
1 8
2 - h y d ro xyg l u t ar a te c o n ta i n in g O . O n t h e r i gh t ( r e d l in e s ) a r e E IC s f o r i o n s w i t h m / z 1 5 1 .0 2 2 , r e p r e s e nt i n g 2 - h y d r o x yg l u t a r a t e c o n t a i ni n g t w o O a t om s .
1
8
A
c
o
n
t
r
o
l
r
e
a
c
t
i
o
n
p
e
r
f
o
r
m
e
d
u
n
d
e
r
a
m
b
i
e
n
t
O
c
o
n
d
i
t
i
o
n
s
(
n
=
3
)
i
s
c
o
m
p
a
r
e
d
w
i
t
h
a
r
e
a
c
t
i
o
n
p
e
r
f
o
r
m
e
d
u
n
d
e
r
a
O
a
t
m
o
s
p
h
e
r
e
(
n
=
3
)
a
n
d
a
2
2
2 - h y d ro xyg l u t ar a te s t a n d a rd . f O ve r l a y o f holo ( g r e en ) a n d s u b s t r a t e - b ou n d ( c y a n ) s t r u ct u r e s o f H g l S d i s p l a y i n g t h e e n z y m e a c t i v e s i t e a n d t h e i nt e r a c t i on
o f V a l 4 0 2 a n d S er 4 0 3 w i t h t h e 2 -o xo a d i p a t e s u b s t r a t e. T h e l o o p c o n t a i n i n g V a l 4 0 2 a n d S e r 40 3 i s s h o w n i n t h e holo ( d a r k g r e e n ) a n d s u b s t r at e -b ou nd
( d a r k b l u e ) s t a te s . g R ea c t i o n r a t e s o f H gl S w it h d i f f e r en t 2 -o xo a c i d s u b s t r a t e s : 2 - o x o g l u t a r at e ( 2 O G ) , 2 - o x o a d i p at e ( 2 O A ) , 2 - o x o p i m e l a te ( 2 O P ) , 2 -
o x o b ut y r at e ( 2 O B) , 2 - o x o v a l e r a te ( 2 O V ) , 2 -o xo h e xa n o a te ( 2 O H ) , 2 - o x o o c t a n o at e ( 2 O O) . E r r o r b a r s r e p r e s e n t 9 5 % c o nfid e n c e i n t e r v a l s , b l a c k d ot s
r e p re s e nt i n d i vi d u a l m e a s u r e m en t s , n = 3 . h R e a c t i o n r a t e s o f W T H g l S ( b l a c k d a s h e s) , R 7 4 A ( b l u e d a s h e s) , a n d V 4 0 2 P ( o r a n g e d a s h e s) m ut a nt s
m e as u r e d b y a n e n z ym e c o u p l e d d e c a r b o x y la t io n a s s a y w it h 2 O A a s a s u b s t r a t e , n = 3 . E r r o r b a r s r e p r e s e n t 9 5 % c o nfid e n c e i n t e r v a l s , c o l o r e d d ot s
c o rr e s p on d t o i n d i vi d u a l m e a s u r e me n t s .
substrate-binding interactions, but it also appears to influence the used to solve the substrate-bound crystal structure at 1.85 Å
strict enzyme substrate specificity. Furthermore, an R74A resolution. The FLO7 crystal structure, like the other DUF1338
mutation abolished enzymatic activity (Fig. 1h). We additionally proteins, displayed the conserved central VOC fold containing
probed the importance of the loop bearing Val402 and Ser403 on the active site and metal-binding center (Fig. 2a). Comparison of
substrate binding. Mutating Val402 to a proline residue to disrupt the HglS and FLO7 structures revealed that orientation of the
the hydrogen bond observed between 2OA and the amide metal-coordinating residues and 2OA in both structures were
backbone of Val402 significantly decreased enzyme activity, nearly identical (Fig. 2b) even though the proteins display low
suggesting that while this residue is not essential for turnover, it (~15%) sequence identity (Supplementary Fig. 6). In addition, the
likely contributes to substrate-binding affinity (Fig. 1h). However, FLO7 structure also contains a substrate-binding arginine, Arg64,
it is also possible that the observed decrease in activity is an located in nearly the same position as Arg74 of HglS. However,
artefact of disrupted protein folding due to the introduction of a unlike Val402 and Ser403 of HglS, FLO7 does not possess
proline residue. Further experiments with more conservative any other residues interacting with the substrate. The results of
mutations at positions 402 and 403 will be required to fully our biochemical and structural studies of FLO7 manifest the
understand the role(s) of these residues in substrate binding and/ remarkable conservation of the fold, active site architecture, and
or catalysis.
mechanism among DUF1338 proteins across different domains
of life. Consequently, we predict that all DUF1338 proteins
containing the conserved 2OA-interacting arginine likely catalyze
the decarboxylation and hydroxylation of 2OA to form 2HG.
These hydroxyglutarate synthases form a new two-substrate α-
ketoacid-dependent oxygenase (2S-αKAO) subfamily lacking
significant sequence identity to HMS or 4-HPPD and acting on
a distinct substrate.
DUF1338 biochemistry and structure is conserved in homo-
logs. Based on the mechanistic information gleaned from our
HglS biochemical and structural characterization, we sought to
propose potential functions for other DUF1338 family members.
Previously, we showed that DUF1338 family proteins are widely
distributed across several domains of life, while others have
demonstrated that DUF1338 protein coding sequences are pre-
sent in the majority of sequenced plant genomes8,9. In plants, the Conserved residues suggest a common biochemical role. The
catabolism of lysine is known only until 2OA, with further conserved structural features, substrate, and biochemical activity
catabolic steps having only been hypothesized7. Furthermore, D- of plant and bacterial DUF1338 proteins suggest a common
2HG has also been identified as an intermediate in plant lysine physiological role across domains of life. Therefore we analyzed
catabolism, though the mechanism of its formation has yet to be all known DUF1338 domain-containing protein amino acid
proven. We therefore hypothesized that plant homologs perform sequences for the key catalytic and substrate-binding residues we
the same reaction as HglS, converting 2OA to 2HG. To test this, identified in HglS. First, we queried all DUF1338 proteins within
we biochemically characterized the DUF1338 homologs from the Pfam database for the catalytic and metal-binding residues
Arabidopsis thaliana and O. sativa as well as an additional dis- identified in HglS. Of the 2417 unique DUF1338-containing
tantly related Escherichia coli homolog. Soluble variants of plant proteins found in the Pfam PF07063 family, 86% possess the
proteins were constructed by removing the predicted N-terminal conserved “HHE” metal-binding triad (Supplementary Table 1).
localization peptide. As expected, the E. coli homolog YdcJ and This strong conservation is consistent across domains of life, with
the plant homologs AT1G07040 and FLO7 catalyzed the con- 85% of DUF1338 proteins in bacteria, 92% in fungi, and 81% in
version of 2OA to that was confirmed by in vitro assays analyzed plants possessing the metal-binding residues. Further analysis
using high resolution LC-MS (Supplementary Fig. 5a). In addi- revealed arginine 74, which dictates HglS substrate specificity, is
tion, the FLO7 homolog displayed kinetic parameters similar to also highly conserved across the family. Of the homologs with
HglS, with a Km of 0.55 mM and a Vmax of 0.89 mM/min/µM conserved HHE triads, 100% of fungal proteins, 99.7% of bacterial
enzyme (Supplementary Fig. 5b).
proteins, and 93.4% of plant proteins maintain an arginine at this
In addition, the structural conservation between bacterial and position (Supplementary Table 1). Therefore, though there is little
eukaryotic DUF1338 proteins was compared by obtaining a primary sequence identity between bacterial homologs and plant
crystal structure of FLO7. Initial crystallization screens resulted in homologs (Supplementary Figs. 6 and 7), the critical residues
crystals that were soaked with 2OA, producing diffraction data coordinating the Fe(II) cofactor and the carboxylate-coordinating
4
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
(
2
0
2
0
)
1
1
:
2
9
3
1
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
8
1
5
-
3
|
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s

N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
8
1
5
-
3
A R T I C L E
a
180
b
E268
H226
E294
H60
Ni
H196
R74
H70
R64
2OA
Fig. 2 Structural comparison of FLO7 and HglS. a F L O 7 c r y s t a l s t r u ct u r e d i s p l a y e d a s c a r t o o n s h e e t s a n d h e l i ce s . M e t a l - co o rd i n a t i n g a n d s u bs t r at e -
b i nd i n g r e si d u es a r e s h o wn a s s t i c k s , w h il e t h e 2 - o xo a d i p at e s u b s t r a t e i s s h o w n a s g r e e n s t i ck s . T h e n i c k e l b o u n d w i t h i n t h e a c t i v e s i t e i s s h o w n a s a t a n
s ph e r e . b O v er l a y o f H g l S ( t e a l) a n d F L O 7 ( y e ll o w) s t r u c t u r e s s h o w i n g m e t a l - c o o r d i n at i n g r e s i d u es , s u b s t r a t e -b i n d i n g r e s i d u e s , t h e a c t i v e s i t e m e t a l , a nd
t h e 2 - o x o a d i p at e s u b st r a t e . R i g h t i n s e t : H gl S ( t e a l) a n d F L O 7 ( y e l l o w) a c t i v e s i t es d i s p l a y i n g c o n s e r v ed r e s i d u e s a n d s u b s t r a t e -b i n d i n g m o d e . T he 2 O A
s ub s t r a t e s a r e c o l o r e d l i g h t b l u e ( H gl S) a n d w h it e ( F L O7 ) , a n d t h e a c t i v e s i t e b o u n d n i c k e l s h o w n a s s p h e r e s a n d c o l o r e d b l u e ( H g l S) a n d o li v e ( F L O 7 ).
arginine are highly conserved. Given the substrate specificity proteins are isoforms of loci that also have putatively localized
exhibited by the P. putida DUF1338 protein HglS, we hypothe- isoforms, though phylogenetic analysis also revealed a distinct non-
sized that the remaining uncharacterized homologs maintaining localized DUF1338 protein clade predominantly within Brassica-
the metal-binding residues and conserved arginine would also use ceae (Supplementary Fig. 9). Notably, this non-localized subclade
2OA as a substrate.
was enriched for proteins maintaining the HHE coordination triad
Nearly all plant and algal genomes within GenBank encode but that possessed either a glutamine or methionine at the position
DUF1338 family proteins (Supplementary Table 1). Previous work corresponding to Arg74 of HglS (Supplementary Fig. 9). As most
to understand plant lysine catabolism suggested the missing plant homologs contain a localization tag, we predicted these
reactions between 2OA and 2HG proceed through the 2- homologs would likely function similar to rice FLO7.
ketoglutarate dehydrogenase complex (2KGD), forming glutaryl-
We therefore tested whether mutants of the DUF1338 protein
Coenzyme A (Supplementary Fig. 8)7. Transcript correlation AT1G07040 from the model plant A. thaliana exhibited a
analysis in A. thaliana, however, revealed that the transcription of phenotype resembling other plant lysine-accumulating mutants.
known lysine catabolic enzymes was not correlated with 2KGD When grown on germination media, Salk_103299C mutants
expression; rather, their transcription is highly correlated with a (which have confirmed homozygous T-DNA insertions in the
DUF1338 protein AT1G07040 (Supplementary Table 2). In AT1G07040 locus) displayed significantly delayed germination
addition, in O. sativa, mutants of the DUF1338 protein FLO7 compared with wild-type seedlings (Supplementary Fig. 10a).
display a floury starch phenotype that resembles the maize AT1G07040 mutants that germinated also displayed compro-
opaque2 phenotype which is known to cause lysine accumulation mised development and appeared cholortic with impaired growth
in mutant kernels3,5,6,9. These transcriptomic and phenotypic (Supplementary Fig. 10a). This phenotype was recapitulated when
correlations suggest that, like HglS in P. putida, DUF1338 proteins seeds were germinated in soil. After a 14 day incubation, no
are involved in the plant lysine degradation pathway.
Salk_103299C seeds germinated whereas 95% of wild-type seeds
During seed maturation, lysine catabolism is required for had germinated (n = 96) (Supplementary Fig. 10c). After 45 days,
normal amyloplast development16. In rice, FLO7 localizes to the only 7% of mutant seeds had germinated and after 60 days, only
amyloplast via a N-terminal chloroplast localization signal 15%. This phenotype is consistent with previous work showing
peptide9. While the A. thaliana homolog AT1G07040 also has a lysine accumulation in A. thaliana seeds resulted in significantly
predicted chloroplast localization sequence, only 117 of 202 plant delayed germination16,17. Toluidine blue staining of sectioned
proteins had predicted chloroplast or mitochondria localization mutant seeds also suggested that oil body formation was altered
tags (Supplementary Table 3). Many predicted non-localized (Supplementary Fig. 10b). The altered morphology of the oil
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
(
2
0
2
0
)
1
1
:
2
9
3
1
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
8
1
5
-
3
|
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s
5

A
R
T
I
C
L
E
N A T U R E C O M MU N ICA T IO N S | h t t p s : / /d o i .o r g / 1 0 .1 0 3 8/ s 41 46 7 - 0 2 0- 16 8 1 5 - 3
body, serving as the primary carbon storage unit in A. thaliana, is plants8,9. Our work here shows that the arginine residue that
analogous to the altered starch granule formation observed in rice mediates the 2OA specificity of HglS is highly conserved across
FLO7 mutants9. The aberrant phenotype in both A. thaliana and these homologs. It is therefore likely that all DUF1338 proteins
rice mutants suggest that DUF1338 homologs play a critical role with the conserved arginine catalyze the decarboxylation and
in embryo development both in monocots and dicots. The hydroxylation of 2OA to form 2HG. The enzyme’s substrate,
phenotypes in A. thaliana observed in this preliminary work will 2OA, is primarily known as a lysine catabolism intermediate,
need to be validated using multiple confirmed T-DNA lines that suggesting that DUF1338 enzymes participate in similar bio-
disrupt the function of AT1G07040.
chemical processes across domains of life (http://modelseed.org/
biochem/compounds/cpd00269). Previous plant phenotypic stu-
dies support this hypothesis; a rice study showed DUF1338-
containing protein FLO7 disruption produced a crystalline starch
Discussion
Here, we present the structural and mechanistic analysis of HglS, phenotype within the amyloplast9, similar to the lysine-
the first DUF1338 protein family member to be assigned a accumulating opaque2 maize mutants19.
function. In a previous report, we showed that HglS converts 2OA
DUF1338 protein AT1G07040 expression in A. thaliana is
to 2HG. The HglS chemical mechanism remained ambiguous highly correlated with other known lysine catabolism enzymes. In
however, as this conversion involves two apparent enzymatic this work, we further support this claim by showing that when
steps: decarboxylation and hydroxylation. Here we report the AT1G07040 is disrupted, A. thaliana seedlings have significantly
high resolution crystal structures of holo- and substrate-bound compromised germination ability. Previous work showed A.
enzyme aiding the proposal of an enzyme mechanism. In addi- thaliana lysine-accumulating mutants have delayed germination
tion, we present the substrate-bound structure of the plant rates due to unfavorable TCA cycle effects, suggesting that
homolog, FLO7.
AT1G07040 could be involved in lysine degradation16,17. Fur-
Several HglS unique features were used to propose an enzymatic thermore, by assaying the A. thaliana and rice enzymes in vitro,
reaction mechanism. A VAST structural similarity search revealed we show that AT1G07040 and FLO7 catalyze the transformation
the central VOC superfamily β-sheet fold, and metal-binding of 2OA to 2HG. While both compounds were previously iden-
residues of HglS were positioned similarly to the corresponding tified as plant lysine catabolism intermediates, no clear link
HMS and HPPD residues (Supplementary Data 1). Both HMS and between the two molecules had been demonstrated20. Finally,
HPPD catalyze an intramolecular decarboxylation–hydroxylation histopathological examination of mutant AT1G07040 seeds
reaction but share little sequence identity with HglS10. Further- showed aberrant oil body formation. While previous experiments
more, a 2OA-bound HglS co-crystal structure revealed bidentate showed that the disruption of lysine utilization compromises the
coordination of the 2OA α-keto group by the metal center as a key ability of seeds to store carbon in monocots such as rice and
step in the enzymatic mechanism. These features were remarkably maize, our results show this phenomenon also occurs in dicots
conserved in the FLO7 co-crystal structure, while the α-helices such as Arabidopsis. Given their near ubiquitous conservation
surrounding the central domain diverge.
and highly specific biochemical function we find it likely that
The VOC fold is conserved across the protein superfamily, yet DUF1338 proteins localized to chloroplasts catalyze the last
the specific orientation of β-sheets in HglS, FLO7, HMS, and missing step in lysine catabolism of all green plants.
HPPD appears to be most conserved in the intramolecular 2S-
While the majority of plant HglS homologs have chloroplast
αKAO subfamily. With only seven β-sheets composing the central localization tags, some lack any predicted signal peptide. This is
barrel-like domain, the HglS structure diverges slightly from the especially prevalent in the Brassicaceae, where many species
canonical VOC fold which contains eight β-sheets.
appear to have non-localized HglS paralogs. The majority of these
The central VOC domains of HglS, FLO7, HMS, and HPPD are paralogs retain arginine as their specificity residue, but they may
distinct from that of the mechanistically similar α-ketoglutarate have roles in pathways beyond lysine catabolism as the A. thali-
(αKG) dependent dioxygenases which possess a double-stranded ana paralog AT1G27030 (lacking a localization sequence) shows
β-helix (jelly roll) fold containing the metal-binding center and no expression correlation with known lysine catabolic genes
active site18. However, the orientation of metal-binding residues (Supplementary Table 2).
and bound substrate is understandably homologous given the
Recent studies show that the lysine-derived intermediates
similar chemical mechanisms of the enzyme families, which we pipecolate and N-hydroxy-pipecolate can initiate systemic
verified through several biochemical experiments. Whether the acquired resistance (SAR) in plants21,22. SAR, a global response,
shared general mechanism of the 2S-αKAOs and αKG-dependent grants lasting broad-spectrum disease protection in uninfected
dioxygenases is a product of convergent evolution remains tissue23. In bacteria, pipecolate is often catabolized to 2OA,
unclear.
suggesting that HglS homologs should receive future attention
While HMS and HPPD show significant sequence identity and when studying pipecolate metabolism in plants. In addition to the
share the same substrate, the DUF1338 family proteins are not non-localized paralogs retaining the conserved arginine, multiple
homologous to HMS or HPPD, nor do they share significant Brassicaceae paralogs have altered residues at this position, har-
sequence identity with each other (~15% identity between HglS boring either methionine or asparagine. These paralogs likely
and FLO7). Low sequence conservation is a general VOC catalyze the decarboxylation and hydroxylation of substrates
superfamily feature, but the low sequence identity between other than 2OA. Further research discerning the physiological
members catalyzing the same reaction is less common14. Despite and biochemical functionality of “mutant” paralogs is warranted.
low primary sequence conservation, the biochemical assays of
DUF1338 proteins are also widely distributed in both Ascomy-
mutants, structural analysis, and bioinformatics reported herein cota and Basidiomycota fungi, though the model fungi Sacchar-
show that very few residues are essential for enzyme turnover in omyces cerevisiae or Neurospora crassa lack homologs. Within the
the hydroxyglutarate synthase family. It is possible that the α- fungal homologs examined here, all proteins containing a HHE
helices surrounding the central VOC domain have been selected metal-binding triad also maintained the arginine specificity residue
for other attributes, such as mediating protein–protein interac- suggesting that all function on a 2OA substrate. Unfortunately,
tions as observed in the αKG-dependent dioxygenases18.
almost no fungal catabolic lysine pathways are fully characterized
Previous work has shown DUF1338 family proteins are wide- genetically or biochemically24. Multiple studies, however, suggested
spread in bacteria and fungi, but they are especially prevalent in 2OA is a likely lysine catabolism intermediate in Pyriculuria oryzae
6
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
(
2
0
2
0
)
1
1
:
2
9
3
1
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
8
1
5
-
3
|
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
o
m
m
u
n
i
c
a
t
i
o
n
s

N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
8
1
5
-
3
A R T I C L E
and Candida albicans25,26. Moreover, while P. oryzae possesses a imidazole concentrations (100, 200, and 400 mM). Fractions were collected and
analyzed via SDS-PAGE. Purified protein was dialyzed overnight at 4 °C against
50 mM HEPES pH 7.5, 5% glycerol.
DUF1338 homolog, C. albicans does not. This implies C. albicans
and other fungi may utilize a catabolic route similar to mammals in
which 2OA is converted to glutaryl-CoA via 2OA dehy-
Crystallization. An initial crystallization screen was set up using a Phoenix robot
(Art Robbins Instruments, Sunnyvale, CA) using the sparse matrix screening
drogenase27. Future studies are required to examine whether fungal
HglS homologs also play a role in fungal lysine catabolism.
Of the over 2000 bacterial HglS homologs examined here that
retain the HHE triad, over 99% maintained arginine as the
specificity-conferring residue, suggesting widespread DUF1338
enzymatic activity conservation in prokaryotes. Previous work in
P. putida and published fitness data in Pseudomonas fluorescens
method45. Purified HglS was concentrated to 20 mg/mL and Flo7 was concentrated
to 10 mg/mL prior to crystallization using the sitting drop method in 0.4 µL drops
containing a 1:1 ratio of protein sample to crystallization solution. For HglS, the
crystallization solution consisted of 0.2 M Ammonium Fluoride and 20% PEG
3,350, while the crystallization solution for Flo7 contained 0.01 M Magnesium
chloride hexahydrate, 0.05 M MES monohydrate pH 5.6 and 1.8 M Lithium sulfate
monohydrate. Crystals were transferred to crystallization solution containing 20%
and Sinorhizobium meliloti provide evidence of a conserved glycerol prior to flash freezing in liquid nitrogen.
physiological function as well8,28,29. E. coli was recently shown to
possess a non-ketogenic lysine catabolic route via a glutarate
hydroxylase, supplementing degradation routes to cadaverine30,31
X-ray data collection and model refinement. X-ray diffraction data for HglS were
collected at the Berkeley Center for Structural Biology on beamline 5.0.2 and 8.2.2
of the Advanced Light Source at Lawrence Berkeley National Lab. Diffraction data
for Flo7 were collected at the Stanford Synchrotron Radiation Lightsource on
beamline 12-2. The HglS and Flo7 structures were determined by the molecular-
replacement method with the program PHASER46 using uncharacterized protein
YdcJ (SF1787) from Shigella flexneri (PDB ID: 2RJB) and the putative hydrolase
(YP_751971.1) from Shewanella frigidimarina (PDB ID: 3LHO) as the search
models, respectively. Structure refinement was performed by phenix.refine pro-
gram47. Manual rebuilding using COOT48 and the addition of water molecules
allowed for construction of the final model. The R-work and R-free values for the
final models of all structures are listed in Supplementary Table 6. Root-mean-
square deviations from ideal geometries for bond lengths, angles, and dihedrals
were calculated with Phenix49. The overall stereochemical quality of the final
models was assessed using the MolProbity program50. Structural analyses were
performed in Coot48, PyMOL (https://pymol.org/2/)51, and UCSF Chimera52. All
structural data has been submitted to the Protein Database with the following PDB
IDs: HglS: 6W1G, HglS-2OA: 6W1H, Flo7-2OA:6 W1K.
.
E. coli also possesses a HglS homolog, YdcJ, which we showed has
identical activity to HglS. In addition, the E. coli enzyme structure
was solved (PDB ID 2RJB) and shows the same conserved VOC
fold, metal-binding motif, and arginine-binding residues as the
HglS and FLO7 structures reported herein. Future work will elu-
cidate whether, like P. putida, E. coli possesses multiple lysine
catabolism routes or whether 2OA functions in other physiological
processes.
Low cereal and legume lysine content produces protein-energy
malnutrition in 30% of the developing world population3,32–35
.
Increasing lysine content in staple crops will require both lysine
overproduction and catabolic pathway elimination34. However,
lysine metabolism changes in maize, rice, and soybeans resulted
in low germination rates, abnormal endosperm, and reduced
grain weights3,19,36–38. We hope that the more complete under-
standing of lysine catabolism elucidated here will help resolve
causes of pleiotropic effect in plants and aid in the development
of stable high-lysine crops to combat malnutrition globally.
Enzyme kinetics and O2 consumption. Enzyme coupled decarboxylation assays
were carried out as previously described53. Reaction mixtures contained 100 mM
Tris-HCl (pH 7), 10 mM MgCl2, 0.4 mM NADH, 4 mM phosphoenol pyruvate
(PEP), 100 U/mL pig heart malate dehydrogenase (Roche), 2 U/mL microbial PEP
carboxylase (Sigma), and 10 mM 2OA. Reactions were initiated by the addition of
purified HglS or boiled enzyme controls, and absorbance at 340 nm was measured
via a SpectraMax M4 plate reader (Molecular Devices, USA).
Initial rate measurements were directly recorded monitoring the consumption
of O2 using a Clarke-type electrode (Hansatech Oxygraph). Reaction mixtures
containing FeSO4 (10 µM) and 2-oxoadipic acid (10–200 µM) in 100 mM Tris pH
7.0 were allowed to equilibrate to room temperature determined by a stable O2
concentration reading of 240 µM. Addition of purified apo enzyme (100 nM) to the
sealed reaction vial initiated the reaction. Initial rates were determined from the
linear portion of consumption of O2 corresponding to up to 10% consumption of
the limiting reactant. No burst or lag phases were observed. All assays were
performed in triplicate.
Methods
Media, chemicals, and strains. Routine bacterial cultures were grown in Luria-
Bertani (LB) Miller medium (BD Biosciences, USA). E. coli was grown at 37 °C.
Cultures were supplemented with carbenicillin (100 mg/L, Sigma Aldrich, USA).
All compounds with the exception of 2-oxohexanoic acid were purchased through
Sigma Aldrich. All bacterial strains and plasmids used in this work are listed in
Supplementary Table 4 and are available through the public instance of the JBEI
registry. (https://public-registry.jbei.org/). A. thaliana mutants were obtained from
the Salk collection.
O2 consumption measurements used to determine the reaction stoichiometry
were also measured using a Clarke-type electrode. Reactions mixtures containing
FeSO4 (10 µM) and 2-oxoadipic acid (100 or 200 µM) in 100 mM Tris pH 7.0 were
equilibrate to room temperature ([O2] = 240 µM) and monitored for at least 2 min
prior to the addition of apo enzyme (1 µM). Oxygen consumption was monitored
until the signal plateaued and the O2 concentration was stable. The observed rate
was determined by fitting the data to a single exponential decay model. The
reaction stoichiometry was determined by taking the ratio of the moles of O2
consumed and the concentration of the 2-oxoadipic acid present in the reaction.
Each reaction condition was performed in triplicate.
DNA manipulation. All plasmids were designed using Device Editor and Vector
Editor software, while all primers used for the construction of plasmids were
designed using j5 software39–41. All primers used in this study can be found in
Supplementary Table 5. Plasmids were assembled via Gibson Assembly using
standard protocols42, or Golden Gate Assembly using standard protocols43. Plas-
mids were routinely isolated using the Qiaprep Spin Miniprep kit (Qiagen, USA),
and all primers were purchased from Integrated DNA Technologies (IDT, Coral-
ville, IA). Site directed mutants were created by incorporating desired mutations
into PCR primers. PCR fragments were then re-assembled into the mutant plasmid
using Golden Gate assembly44. The geneblock for the E. coli codon optimized O.
sativa was purchased through IDT (Coralville, IA). Arabidopsis cDNA was used to
amplify AT1G07040.
Oxygen labeling experiments. All reagents were exhaustively purged with argon
on a Schlenk link to remove 16O2. Anaerobic buffer was subsequently saturated
with 18O2 via gentle bubbling with 18O2 (Sigma, 99 atom % 18O). Reaction mix-
tures containing enzyme (1 µM), 2-oxoadipic acid (1 mM), FeSO4 (10 µM), and
18O2 saturated buffer (1.2 mM) were mixed in anaerobic sealed reaction vials using
gastight syringes. The reaction was initiated by the addition of 2-oxoadipic acid.
The headspace was filled with 18O2 gas. Reactions were incubated at room tem-
perature for 2 h before being quenched with an equal volume of methanol. Control
experiments, replacing 18O2 with 16O2, were run in parallel. Quenched reaction
mixtures were analyzed by LC-MS.
Protein purification. A 5 mL overnight culture of E. coli BL21 (DE3) containing
the expression plasmid was used to inoculate a 500 mL culture of LB. Cells were
grown at 37 °C to an OD of 0.6 then induced with Isopropyl β-D-1-
thiogalactopyranoside to a final concentration of 1 mM. The temperature was
lowered to 30 °C and cells were allowed to express for 18 h before being harvested
via centrifugation. Cell pellets were stored at −80 °C until purification. For pur-
ification, cell pellets were resuspended in lysis buffer (50 mM sodium phosphate,
300 mM sodium chloride, 10 mM imidazole, 8% glycerol, pH 7.5) and sonicated to
lyse cells. Insolubles were pelleted via centrifugation (30 min at 40,000 × g). The
supernatant was applied to a fritted column containing Ni-NTA resin (Qiagen,
USA), which had been pre-equilibrated with several column volumes of lysis buffer.
The resin was washed with lysis buffer containing 50 mM imidazole, then the
protein was eluted using a stepwise gradient of lysis buffer containing increasing
LC-MS analysis. All in vitro reactions to be analyzed via LC-MS were quenched
with an equal volume of ice cold methanol and stored at −80 °C until analyses.
Detection of 2OA and 2HG were described previously8. Briefly, HILIC-HRMS
analysis was performed using an Agilent Technologies 6510 Accurate-Mass Q-TOF
LC-MS instrument using positive mode and an Atlantis HILIC Silica 5 µm
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
(2 0 2 0 ) 1 1 : 2 9 3 1 | h t t p s :/ / d o i . o r g / 1 0 . 1 0 3 8 / s 4 1 4 6 7 - 0 2 0 - 1 6 8 1 5 - 3 | w ww . n a t u r e . c o m / n a t u r e c o m m u n ic a t io n s
7

A
R
T
I
C
L
E
N A T U R E C O M MU N ICA T IO N S | h t t p s : / /d o i .o r g / 1 0 .1 0 3 8/ s 41 46 7 - 0 2 0- 16 8 1 5 - 3
column (150 × 4.6 mm) with a linear of 95 to 50% acetonitrile (v/v) over 8 min in
water with 40 mM ammonium formate, pH 4.5, at a flow rate of 1 mL min−1
Reporting summary. Further information on research design is available in
the Nature Research Reporting Summary linked to this article.
.
Data availability
Plant growth. A. thaliana DUF1338 mutant, Salk_103299C, seeds were ordered
from the Arabidopsis Biological Resource Center (Columbus, OH, USA). Seeds
were surface sterilized with 70% EtOH for 3 min, followed by a ten-minute sub-
mersion in 10% bleach solution, then rinsed with sterile water three times. For soil
germination, seeds were planted in Premier Pro-Mix mycorise pro soil, two seeds
per well in a twenty-four well tray. Trays with seeds were then stratified at 4 °C for
3 days, then grown in a Percival-Scientific growth chambers at 22 °C in 10/14-h
light/dark short-day cycles with 60% humidity. After the formation of the full
rosette, plants were genotyped to confirm homozygosity for the mutation of
interest. Primers were designed using the automated method from http://signal.
salk.edu/cgi-bin/tdnaexpress, and genotyping was done in accordance with a pre-
viously described protocol and diagrammed in Supplementary Fig. 1154. After
confirmation of homozygous mutants, plants were moved into individual wells and
transferred to long-day conditions, 22 °C in 16/8-h light/dark cycles with 60%
humidity, to induce flowering for seed collection. For germination on synthetic
media, sterilized seeds were plated on solid media supplemented with ½ Murashige
and Skoog media base, 5% sucrose, 0.8% agar, and 10 μM gibberellic acid to induce
germination. Plates with seeds were stratified at 4 °C for 3 days then transferred to
a Percival-Scientific growth chambers at 22 °C in 10/14-h light/dark short-day
cycles with 60% humidity.
The atomic coordinates and structural factors of HglS without substrate, HglS in complex
with substrate, and Flo7 in complex with substrate have been deposited in the Worldwide
Protein Data Bank (https://www.wwpdb.org/) with PDB ID codes of 6W1G, 6W1H, and
6W1K, respectively. Bacterial strains and plasmids are available upon request from:
https://registry.jbei.org/. The source data underlying Figs. 1d, g, h and Supplementary
Figs. 3, 4a, b, 5b, and 10c are provided as a source data file. Source data are provided with
this paper.
Code availability
All code used in data analysis will be made available upon request. Source data are
provided with this paper.
Received: 17 March 2020; Accepted: 21 May 2020;
References
1. Galili, G. New insights into the regulation and functional significance of lysine
metabolism in plants. Annu. Rev. Plant Biol. 53, 27–43 (2002).
2. Galili, G. & Höfgen, R. Metabolic engineering of amino acids and storage
proteins in plants. Metab. Eng. 4, 3–11 (2002).
3. Galili, G. & Amir, R. Fortifying plants with the essential amino acids lysine
and methionine to improve nutritional quality. Plant Biotechnol. J. 11,
211–222 (2013).
4. Yang, Q.-Q. et al. Biofortification of rice with the essential amino acid lysine:
molecular characterization, nutritional evaluation, and field performance. J.
Exp. Bot. 67, 4285–4296 (2016).
5. Hartings, H., Lauria, M., Lazzaroni, N., Pirona, R. & Motto, M. The Zea mays
mutants opaque-2 and opaque-7 disclose extensive changes in endosperm
metabolism as revealed by protein, amino acid, and transcriptome-wide
analyses. BMC Genom. 12, 41 (2011).
Synthesis of 2-oxohexanoic acid. Synthesis of 2-oxohexanoic acid was carried out
as described previously55. Briefly, a solution of the Grignard reagent, prepared from
1-bromobutane (500 mg, 3.68 mmol) and the suspension of magnesium (178 mg,
7.33 mmol) in THF (5 mL) was added dropwise under N2 atmosphere to a solution
of diethyloxalate (487 mg, 3.33 mmol) in THF (4 mL) at −78 °C. After the addition
was complete, the reaction mixture was stirred at –78 °C for an additional 5 h. The
reaction was quenched with 2 N HCl, the aqueous layer was extracted with ether,
and the combined organic layer was washed with brine, dried over MgSO4, and
evaporated. The crude product was dissolved in acetic acid (20 mL) and conc. HCl
(5 mL). After 11 h, the reaction was concentrated directly, and the residue was
purified by distillation under reduced pressure to give the pure product (203 mg,
43%) as a colorless oil. 1H NMR (400 MHz, CDCl3) 2.92 (t, 2H), 1.81–1.54 (m,
2H), 1.48–1.28 (m, 2H), 0.92 (t, 3H) (Supplementary Fig. 12).
6. Locatelli, S., Piatti, P., Motto, M. & Rossi, V. Chromatin and DNA
modifications in the Opaque2-mediated regulation of gene transcription
during maize endosperm development. Plant Cell 21, 1410–1427 (2009).
7. Hildebrandt, T. M., Nunes Nesi, A., Araújo, W. L. & Braun, H.-P. Amino acid
catabolism in plants. Mol. Plant 8, 1563–1579 (2015).
8. Thompson, M. G. et al. Massively parallel fitness profiling reveals multiple
novel enzymes in Pseudomonas putida lysine metabolism. MBio 10, e02577-18
(2019).
9. Zhang, L. et al. FLOURY ENDOSPERM7 encodes a regulator of starch
synthesis and amyloplast development essential for peripheral endosperm
development in rice. J. Exp. Bot. 67, 633–647 (2016).
10. Madej, T. et al. MMDB and VAST+: tracking structural similarities between
macromolecular complexes. Nucleic Acids Res. 42, D297–D303 (2014).
11. Brownlee, J., He, P., Moran, G. R. & Harrison, D. H. T. Two roads diverged:
the structure of hydroxymandelate synthase from Amycolatopsis orientalis
in complex with 4-hydroxymandelate. Biochemistry 47, 2002–2013
(2008).
Histopathology. Seeds were imbibed in distilled water for 2 h with gently shaking
and a small hole was cut in the seed coat with a surgical scalpel to aid resin
infiltration. Seeds were fixed in 4% formaldehyde (Electron Microscopy Sciences)
in 50 mM PIPES buffer (pH 7) with gently pulling under vacuum for 10 min. Seeds
were left in fixative overnight at 4 °C, dehydrated in an ethanol gradient, and
infiltrated with Technovit 7100 plastic resin (Electron Microscopy Sciences)
according to manufacturer’s instructions. Resin was infiltrated under gentle
vacuum for 20 min followed by rotation at 4 °C overnight. Seeds were embedded in
beam capsules and 4 μm thick sections were cut on an MR2 manual rotary
microtome (RMC Boeckeler) using a glass knife. Sections were stained with 0.02%
(w/v) Toluidine blue-O for 30 s, then rinsed and mounted in distilled water. Images
were captured on a Leica DM6B microscope (Leica Biosystems Inc. Buffalo Grove,
IL) equipped with a Leica DMC 4500 color camera using Leica Application Suite X
(LASX) software.
12. Serre, L. et al. Crystal structure of Pseudomonas fluorescens 4-
hydroxyphenylpyruvate dioxygenase: an enzyme involved in the tyrosine
degradation pathway. Structure 7, 977–988 (1999).
13. Di Giuro, C. M. L. et al. Chiral hydroxylation at the mononuclear nonheme Fe
(II) center of 4-(S) hydroxymandelate synthase–a structure-activity
relationship analysis. PLoS ONE 8, e68932 (2013).
14. He, P. & Moran, G. R. Structural and mechanistic comparisons of the metal-
binding members of the vicinal oxygen chelate (VOC) superfamily. J. Inorg.
Biochem. 105, 1259–1272 (2011).
15. Armstrong, R. N. Mechanistic diversity in a metalloenzyme superfamily.
Biochemistry 39, 13625–13632 (2000).
16. Angelovici, R., Fait, A., Fernie, A. R. & Galili, G. A seed high-lysine trait is
negatively associated with the TCA cycle and slows down Arabidopsis seed
germination. N. Phytol. 189, 148–159 (2011).
17. Zhu, X. & Galili, G. Increased lysine synthesis coupled with a knockout of its
catabolism synergistically boosts lysine content and also transregulates the
metabolism of other amino acids in Arabidopsis seeds. Plant Cell 15, 845–853
(2003).
18. Aik, W., McDonough, M. A., Thalhammer, A., Chowdhury, R. & Schofield, C.
J. Role of the jelly-roll fold in substrate binding by 2-oxoglutarate oxygenases.
Curr. Opin. Struct. Biol. 22, 691–700 (2012).
Bioinformatics. For mining of DUF1338 homologs in plants all the proteins from
completed genomes with protein predictions available by september 2019 were
retrieved from the GenBank FTP site. The homologs were searched in each pro-
teome with BlastP using Flo7 as query with a bit score cutoff of 150 and an e-value
cutoff of E-12. The retrieved sequences were aligned using muscle56, and trimmed
using jalview57. The multiple sequence alignment was used for phylogenetic
reconstruction suing IQ tree, the best amino acid substitution model was selected
with ModelFinder implemented in IQtree58, branch support was calculated using
10000 bootstrap generations
Sequences of DUF1338 homologs were downloaded from Pfam (https://pfam.
xfam.org/family/PF07063). To compare the 3D structure of HglS with other
protein structures we used VAST10. All alignments were done using the MAFFT-
LINSI algorithm59, and alignments were compared with secondary structures and
visualized using Easy Sequencing in PostScrip (http://espript.ibcp.fr)60. Molecular
graphics and analyses were performed with UCSF Chimera, developed by the
Resource for Biocomputing, Visualization, and Informatics at the University of
California, San Francisco, with support from NIH P41-GM10331152. Python
scripts were developed to calculate conservation of “HHE” metal-binding triads
and R74 residues. Calculation of Shannon-Entropy for DUF1338 sequence
conservation was carried out using the python library Protein Dynamics and
Sequence Analysis (ProDy)61,62. Co-expression analysis of A. thalina transcripts
was performed using the ATTED-II database63
.
8
N A T UR E C O M M U N I C A TI O N S |
(2 0 2 0 ) 1 1 : 2 9 3 1 | h t t p s : / / d o i . o r g / 1 0 . 1 0 3 8 / s 41 4 6 7 - 0 2 0 - 1 6 8 1 5 - 3 | w ww . n a t u r e . c o m / n a t u re c o m m u n i c a t i o n s

N A T UR E C O M M UN I CA T I O NS | h t t p s : //d o i . o r g / 10 . 1 0 38 / s 4 1 4 6 7 - 02 0 - 1 6 8 1 5 - 3
A R T I C L E
19. Mertz, E. T., Bates, L. S. & Nelson, O. E. Mutant gene that changes protein
composition and increases lysine content of maize endosperm. Science 145,
279–280 (1964).
49. Adams, P. D. et al. PHENIX: a comprehensive Python-based system for
macromolecular structure solution. Acta Crystallogr. D Biol. Crystallogr. 66,
213–221 (2010).
20. Engqvist, M. K. M. et al. Plant D-2-hydroxyglutarate dehydrogenase
participates in the catabolism of lysine especially during senescence. J. Biol.
Chem. 286, 11382–11390 (2011).
21. Chen, Y.-C. et al. N-hydroxy-pipecolic acid is a mobile metabolite that induces
systemic disease resistance in Arabidopsis. Proc. Natl Acad. Sci. USA 115,
E4920–E4929 (2018).
22. Wang, C. et al. Pipecolic acid confers systemic immunity by regulating free
radicals. Sci. Adv. 4, eaar4509 (2018).
23. Návarová, H., Bernsdorff, F., Döring, A.-C. & Zeier, J. Pipecolic acid, an
endogenous mediator of defense amplification and priming, is a critical
regulator of inducible plant immunity. Plant Cell 24, 5123–5141
(2012).
24. Zabriskie, T. M. & Jackson, M. D. Lysine biosynthesis and metabolism in
fungi. Nat. Prod. Rep. 17, 85–97 (2000).
25. Wade, M., Thomson, D. M. & Miflin, B. J. Saccharopine: an Intermediate of L-
Lysine biosynthesis and degradation in Pyricularia oryzae. Microbiology 120,
11–20 (1980).
26. Hammer, T., Bode, R. & Birnbaum, D. Occurrence of a novel yeast enzyme, L-
lysine epsilon-dehydrogenase, which catalyses the first step of lysine
catabolism in Candida albicans. J. Gen. Microbiol. 137, 711–715 (1991).
27. Hallen, A., Jamie, J. F. & Cooper, A. J. L. Lysine metabolism in mammalian
brain: an update on the importance of recent discoveries. Amino Acids 45,
1249–1272 (2013).
28. Wetmore, K. M. et al. Rapid quantification of mutant fitness in diverse
bacteria by sequencing randomly bar-coded transposons. MBio 6,
e00306–e00315 (2015).
50. Davis, I. W. et al. MolProbity: all-atom contacts and structure validation for
proteins and nucleic acids. Nucleic Acids Res. 35, 375–383 (2007).
51. The PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC. at
https://pymol.org/2/.
52. Pettersen, E. F. et al. UCSF Chimera—a visualization system for exploratory
research and analysis. J. Comput. Chem. 25, 1605–1612 (2004).
53. Witkowski, A., Joshi, A. K. & Smith, S. Mechanism of the β-ketoacyl synthase
reaction catalyzed by the animal fatty acid synthase. Biochemistry 41,
10877–10887 (2002).
54. Alonso, J. M. et al. Genome-wide insertional mutagenesis of Arabidopsis
thaliana. Science 301, 653–657 (2003).
55. Rapf, R. J. et al. Photochemical synthesis of oligomeric amphiphiles from alkyl
oxoacids in aqueous environments. J. Am. Chem. Soc. 139, 6946–6959 (2017).
56. Edgar, R. C. MUSCLE: multiple sequence alignment with high accuracy and
high throughput. Nucleic Acids Res. 32, 1792–1797 (2004).
57. Waterhouse, A. M., Procter, J. B., Martin, D. M. A., Clamp, M. & Barton, G. J.
Jalview Version 2–a multiple sequence alignment editor and analysis
workbench. Bioinformatics 25, 1189–1191 (2009).
58. Nguyen, L.-T., Schmidt, H. A., von Haeseler, A. & Minh, B. Q. IQ-TREE: a fast
and effective stochastic algorithm for estimating maximum-likelihood
phylogenies. Mol. Biol. Evol. 32, 268–274 (2015).
59. Katoh, K. & Standley, D. M. MAFFT multiple sequence alignment software
version 7: improvements in performance and usability. Mol. Biol. Evol. 30,
772–780 (2013).
60. Robert, X. & Gouet, P. Deciphering key features in protein structures with the
new ENDscript server. Nucleic Acids Res. 42, W320–W324 (2014).
61. Bakan, A. et al. Evol and ProDy for bridging protein sequence evolution and
structural dynamics. Bioinformatics 30, 2681–2683 (2014).
29. Price, M. N. et al. Mutant phenotypes for thousands of bacterial genes of
unknown function. Nature 557, 503–509 (2018).
30. Soksawatmaekhin, W., Kuraishi, A., Sakata, K., Kashiwagi, K. & Igarashi, K.
Excretion and uptake of cadaverine by CadB and its physiological functions in
Escherichia coli. Mol. Microbiol. 51, 1401–1412 (2004).
31. Knorr, S. et al. Widespread bacterial lysine degradation proceeding via
glutarate and L-2-hydroxyglutarate. Nat. Commun. 9, 5071 (2018).
32. Grover, Z. & Ee, L. C. Protein energy malnutrition. Pediatr. Clin. N. Am. 56,
1055–1068 (2009).
62. Bakan, A., Meireles, L. M. & Bahar, I. ProDy: protein dynamics inferred from
theory and experiments. Bioinformatics 27, 1575–1577 (2011).
63. Obayashi, T., Aoki, Y., Tadaka, S., Kagaya, Y. & Kinoshita, K. ATTED-II in
2018: a plant coexpression database based on investigation of the statistical
property of the Mutual Rank Index. Plant Cell Physiol. 59, e3 (2018).
Acknowledgements
33. Wenefrida, I., Utomo, H. S., Blanche, S. B. & Linscombe, S. D. Enhancing
essential amino acids and health benefit components in grain crops for
improved nutritional values. Recent Pat. DNA Gene Seq. 3, 219–225
(2009).
We would like to thank Johan Jaenisch for generously providing A. thaliana cDNA.
Python code to analyze kinetics data was provided by Sam Curran. This work was part of
the DOE Joint BioEnergy Institute (https://www.jbei.org) supported by the US Depart-
ment of Energy, Office of Science, Office of Biological and Environmental Research, and
was part of the Agile BioFoundry (http://agilebiofoundry.org) supported by the US
Department of Energy, Energy Efficiency and Renewable Energy, Bioenergy Technologies
Office, through contract DE-AC02-05CH11231 between Lawrence Berkeley National
Laboratory and the US Department of Energy. J.M.B.H. was supported by the National
Science Foundation Graduate Research Fellowship Program under Grant No. DGE
1106400. The Advanced Light Source is a Department of Energy Office of Science User
Facility under Contract No. DE-AC02-05CH11231. The Berkeley Center for Structural
Biology is supported in part by the Howard Hughes Medical Institute. The ALS-ENABLE
beamlines are supported in part by the National Institutes of Health, National Institute of
General Medical Sciences, grant P30 GM124169. Use of the Stanford Synchrotron
Radiation Lightsource, SLAC National Accelerator Laboratory, is supported by the US
Department of Energy, Office of Science, Office of Basic Energy Sciences under Contract
No. DE-AC02-76SF00515. The SSRL Structural Molecular Biology Program is supported
by the DOE Office of Biological and Environmental Research, and by the National
Institutes of Health, National Institute of General Medical Sciences (P41GM103393). The
contents of this publication are solely the responsibility of the authors and do not
necessarily represent the official views of NIGMS or NIH. The views and opinions of the
authors expressed herein do not necessarily state or reflect those of the United States
Government or any agency thereof. Neither the United States Government nor any
agency thereof, nor any of their employees, makes any warranty, expressed or implied, or
assumes any legal liability or responsibility for the accuracy, completeness, or usefulness
of any information, apparatus, product, or process disclosed, or represents that its use
would not infringe privately owned rights.
34. Le, D. T., Chu, H. D. & Le, N. Q. Improving nutritional quality of plant
proteins through genetic engineering. Curr. Genom. 17, 220–229 (2016).
35. WHO. Protein and Amino Acid Requirements in Human Nutrition. https://
www.who.int/nutrition/publications/nutrientrequirements/WHO_TRS_935/
en/ (World Health Organization, 2007).
36. Kawakatsu, T., Wang, S., Wakasa, Y. & Takaiwa, F. Increased lysine content in
rice grains by over-accumulation of BiP in the endosperm. Biosci. Biotechnol.
Biochem. 74, 2529–2531 (2010).
37. Falco, S. C. et al. Transgenic canola and soybean seeds with increased lysine.
Biotechnology 13, 577–582 (1995).
38. Betrán, F. J., Bockholt, A., Fojt, F., Rooney, L. & Waniska, R. Registration of
Tx802. Crop Sci. 43, 1891-a (2003).
39. Ham, T. S. et al. Design, implementation and practice of JBEI-ICE: an open
source biological part registry platform and tools. Nucleic Acids Res. 40, e141
(2012).
40. Chen, J., Densmore, D., Ham, T. S., Keasling, J. D. & Hillson, N. J.
DeviceEditor visual biological CAD canvas. J. Biol. Eng. 6, 1 (2012).
41. Hillson, N. J., Rosengarten, R. D. & Keasling, J. D. j5 DNA assembly design
automation software. ACS Synth. Biol. 1, 14–21 (2012).
42. Gibson, D. G. et al. Enzymatic assembly of DNA molecules up to several
hundred kilobases. Nat. Methods 6, 343–345 (2009).
43. Engler, C., Kandzia, R. & Marillonnet, S. A one pot, one step, precision
cloning method with high throughput capability. PLoS ONE 3, e3647
(2008).
44. Yan, P., Gao, X., Shen, W., Zhou, P. & Duan, J. Parallel assembly for
multiple site-directed mutagenesis of plasmids. Anal. Biochem. 430, 65–67
(2012).
45. Jancarik, J. & Kim, S. H. Sparse matrix sampling: a screening method for
crystallization of proteins. J. Appl. Crystallogr. 24, 409–411 (1991).
46. McCoy, A. J. et al. Phaser crystallographic software. J. Appl. Crystallogr. 40,
658–674 (2007).
47. Afonine, P. V. et al. Towards automated crystallographic structure
refinement with phenix.refine. Acta Crystallogr. D Biol. Crystallogr. 68,
352–367 (2012).
48. Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics.
Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132 (2004).
Author contributions
Conceptualization, M.G.T., J.M.B.H., and J.H.P.; Methodology, M.G.T., J.M.B.H., J.H.P.,
J.F.B., P.C.M., E.E.K.B., J.A.H., W.M.M., M.S.B., and Y.L.; Investigation, M.G.T., J.M.B.H.,
J.H.P., V.T.B., J.A.H., M.S.B., W.M.M., L.J.W., T.P.B., W.S., R.W.H., C.B.E., E.E.K.B., and
Y.L.; Writing – Original Draft, M.G.T, J.M.B.H., and J.H.P.; Writing – Review and Editing,
All authors; Resources and supervision H.V.S., P.M.S., M.A.M., P.D.A., and J.D.K.
Competing interests
J.D.K. has financial interests in Amyris, Lygos, Demetrix, Napigen, Maple Bio, Apertor
Labs and Ansa Biotechnology. All other authors declare no competing interests.
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
(2 0 2 0 ) 1 1 : 2 9 3 1 | h t t p s :/ / d o i . o r g / 1 0 . 1 0 3 8 / s 4 1 4 6 7 - 0 2 0 - 1 6 8 1 5 - 3 | w ww . n a t u r e . c o m / n a t u r e c o m m u n ic a t io n s
9

A
R
T
I
C
L
E
N
A
T
U
R
E
C
O
M
M
U
N
I
C
A
T
I
O
N
S
|
h
t
t
p
s
:
/
/
d
o
i
.
o
r
g
/
1
0
.
1
0
3
8
/
s
4
1
4
6
7
-
0
2
0
-
1
6
8
1
5
-
3
Additional information
Open Access This article is licensed under a Creative Commons
Supplementary information is available for this paper at https://doi.org/10.1038/s41467-
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons license, and indicate if changes were made. The images or other third party
material in this article are included in the article’s Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not included in the
article’s Creative Commons license and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this license, visit http://creativecommons.org/
licenses/by/4.0/.
020-16815-3.
Correspondence and requests for materials should be addressed to J.D.K.
Peer review information Nature Communications thanks Tatjana Hildebrandt, Graham
Moran, and Jing-Ke Weng for their contribution to the peer review of this work. Peer
review reports are available.
Reprints and permission information is available at http://www.nature.com/reprints
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
© The Author(s) 2020
10
N A T UR E C O M M U N I C A TI O N S |
(2 0 2 0 ) 1 1 : 2 9 3 1 | h t t p s : / / d o i . o r g / 1 0 . 1 0 3 8 / s 41 4 6 7 - 0 2 0 - 1 6 8 1 5 - 3 | w ww . n a t u r e . c o m / n a t u re c o m m u n i c a t i o n s