ARTICLE
Exploi ing gene ic di e si y and gene syn hesis o
iden i y supe io ni ogenase Ni H p o ein a ian s
o enginee N
2
-fixa ion in plan s
Xi Jiang 1,2, Lucía Payá-To mo 1,2, Diana Co oian1, Inés Ga cía-Rubio 3, Rocío Cas ellanos-Rueda 1,4,
Ál a o Ese e i 1,2, Gema López-To ejón1,2, S e an Bu én 1✉& Luis Manuel Rubio 1,2✉
Enginee ing ni ogen fixa ion in euka yo es equi es high exp ession o unc ional ni ogenase
s uc u al p o eins, a goal ha has no ye been achie ed. He e we build a knowledge-based
lib a y con aining 32 ni ogenase ni H sequences om p oka yo es o di e se ecological
niches and me abolic ea u es and combine wi h apid sc eening in obacco o iden i y
supe io Ni H a ian s o plan mi ochond ia exp ession. Th ee Ni H a ian s ou pe o m in
obacco mi ochond ia and a e u he es ed in yeas . Hyd ogenobac e he mophilus (Aquifi-
cae) Ni H is isola ed in la ge quan i ies om yeas mi ochond ia and ulfills Ni H p o ein
equi emen s o e ficien N
2
fixa ion, including elec on ans e o subs a e educ ion, P-
clus e ma u a ion, and FeMo-co biosyn hesis. H. he mophilus Ni H exp essed in obacco
lea es shows lowe ni ogenase ac i i y han ha om yeas . Howe e , ans e o [Fe
4
S
4
]
clus e s om Ni U o Ni H in i o inc eases 10- old he ac i i y o he obacco-isola ed Ni H,
e ealing ha plan mi ochond ia [Fe-S] clus e a ailabili y cons i u es a bo leneck o
enginee plan ni ogenases.
h ps://doi.o g/10.1038/s42003-020-01536-6 OPEN
1Cen o de Bio ecnología y Genómica de Plan as, Uni e sidad Poli écnica de Mad id, Ins i u o Nacional de In es igación y Tecnología Ag a ia y Alimen a ia,
Pozuelo de Ala cón, 28223 Mad id, Spain. 2Depa amen o de Bio ecnología-Biología Vege al, Escuela Técnica Supe io de Ingenie ía Ag onómica, Alimen a ía
y de Biosis emas, Uni e sidad Poli écnica de Mad id, 28040 Mad id, Spain. 3Cen o Uni e si a io de la De ensa, C a. de Huesca s/n, 50090 Za agoza, Spain.
4Depa men o Biosys ems Science and Enginee ing, ETH Zü ich, 4058 Basel, Swi ze land. ✉email: s e an.bu [email protected];lm. [email protected]
COMMUNICATIONS BIOLOGY | (2021)4:4 | h ps://doi.o g/10.1038/s42003-020-01536-6 | www.na u e.com/commsbio 1
1234567890():,;
Ni ogen (N) e ilize s used o inc ease c op p oduc i i y
in in ensi e ag icul u e p ac ices pollu e g oundwa e and
elease g eenhouse gasses1. On he o he hand, sub-
sis ence ag icul u e p ac ices including poo N e iliza ion p o-
duce low and inconsis en yields causing malnu i ion and
po e y2,3. The e is la ge in e es in enginee ing ce eal c op
a ie ies capable o acqui ing hei own N4. One app oach o his
ou come elies on unc ional exp ession o a ni ogenase enzyme
by he ce eal plan 5. Ni ogenases a e p oka yo ic, O
2
-sensi i e,
wo-componen me allop o eins ha con e ine N
2
in o bio-
logically use ul NH
3
6–8. The mos e ficien and widesp ead a -
ian , he molybdenum ni ogenase, is composed o an Fe p o ein
(ni H-encoded) and a MoFe p o ein (encoded by ni D and ni K).
The Fe p o ein (Ni H) dona es elec ons o he MoFe p o ein
(Ni DK) ha in u n educes N
2
. Nascen Ni H and Ni DK
polypep ides need o acqui e p ope qua e na y s uc u e and o
ecei e me al clus e s, one [Fe
4
S
4
] clus e pe Ni H homodime
and wo pai s o P-clus e and FeMo-co pe Ni DK he e o-
e ame , o unc ionali y. We ha e ecen ly e iewed he
mechanisms and gene ic equi emen s o assemble hese co ac o s
and o ma u e Ni H and Ni DK in o ac i e Mo ni ogenase9. The
la ge numbe o ni ogen fixa ion (ni ) genes in ol ed, and he
sensi i i y o mos o he p o ein p oduc s owa ds O
2
, makes
ni ogenase enginee ing a daun ing ask wi h issues ha need o
be sol ed s epwise.
To da e, unc ional Ni H, Ni U, and Ni B ha e been pu ified
om mi ochond ia o ae obically cul u ed Saccha omyces ce e i-
siae cells10,11, while ac i e Ni U and Ni H we e isola ed om
chlo oplas s o Nico iana ben hamiana a he end o he da k
pe iod12. Also, he epo ed low s abili y o he Ni D p o ein13 has
now been imp o ed in wo ecen s udies ha iden ified key
esidues in he Ni D sequence as suscep ible o clea age upon
mi ochond ia impo 14,15. No wi hs anding hese achie emen s,
de ailed analysis o yeas mi ochond ia- a ge ed Azo obac e
inelandii Ni H has been hampe ed by low p o ein solubili y
esul ing in subop imal yields10. Accumula ion o mos ly insoluble
Ni H was also epo ed when Klebsiella oxy oca Ni H was a ge ed
o he obacco mi ochond ia16. The di ficul y o exp essing high
le els o soluble and unc ional Ni H in yeas and obacco poses a
majo p oblem o euka yo ic ni ogenase enginee ing as i is he
mos abundan Ni p o ein du ing N
2
fixa ion17. The p oblem is
exace ba ed because, in addi ion o se ing Ni DK wi h elec ons
o subs a e educ ion, Ni H is equi ed o ma u e P-clus e s on o
Ni DK and o he final s eps o FeMo-co biosyn hesis in complex
wi h Ni EN9. Fo hese easons i is essen ial o iden i y a Ni H
a ian ha is highly soluble and s able when exp essed a e y
high le els in a plan cell, and ha can pe o m all h ee Ni H-
dependen ac i i ies. One app oach o achie e his ou come would
be p o ein enginee ing o well-s udied Ni H om model diazo-
ophs (e.g., A. inelandii o K. oxy oca) aimed o in oduce
sequences ha imp o e s abili y in he mi ochond ia18. P o ein
enginee ing has been ex ensi ely employed o ob ain glyphosa e
esis ance19, ano he impo an ai o c ops. Al e na i ely,
mining o phylogene ically di e se ni H sou ces can be unde aken
in o de o find na u al Ni H p o eins wi h supe io p ope ies, a
s a egy ha was success ul o Ni B11 and o inc easing ca -
o enoid le els in “Golden Rice”20.
He e, 32 dis inc ni H genes we e sc eened o exp ession le el
and solubili y in mi ochond ia o N. ben hamiana. The ni M,
ni U, and ni S genes we e co-exp essed because hei p o ein
p oduc s a e in ol ed in Ni H olding and in he biosyn hesis and
deli e y o i s [Fe
4
S
4
] clus e 9. The Hyd ogenobac e he mophilus
Ni H was iden ified as as ly supe io o he A. inelandii Ni H in
e ms o exp ession le els, solubili y, and unc ionali y bo h in
obacco and yeas mi ochond ia. Mi ochond ia- a ge ed H. he -
mophilus Ni H sa isfied all unc ional and spec oscopic
equi emen s o a ni ogenase Fe p o ein when pu ified om
yeas . The sc eening also pinpoin ed he plan mi ochond ia [Fe-
S] clus e assembly as a bo leneck o u he enginee ing.
Resul s
Lib a y design and s a egy o exp ession o mi ochond ia-
a ge ed Ni H in N. ben hamiana. A lib a y o 32 ni H sequences
om phylogene ically di e se p oka yo es was designed con-
side ing one o se e al o he ollowing c i e ia: (i)ni H genes
ound in confi med diazo ophs; (ii)ni H genes om pho o ophs
o plan -associa ed bac e ia; (iii)ni H genes om ae obic o gan-
isms; (i ) g ow h empe a u e o he ni H hos ; ( )ni H genes
om a chaeal ep esen a i es (Supplemen a y Da a 1). O ganized
by phyla, he selec ion included genes om 1 Aquificae, 4
Fi micu es, 1 Ac inobac e ia, 15 P o eobac e ia, 6 Cyanobac e ia, 1
Chlo obi, 1 Chlo oflexi, and 3 Eu ya chaeo a (Fig. 1a).
The wo kflow o his s udy is desc ibed in Fig. 1b. The gene
sequences encoding he 32 Ni H a ian s we e cloned in o plan
ec o s o Ag obac e ium ume aciens infil a ion-media ed Ni H
exp ession in N. ben hamiana lea es (Supplemen a y Table 1, see
Me hods sec ion o de ails). The ni H sequences we e codon-
op imized o S. ce e isiae because codon-usage is simila o
obacco21 and he wo kflow included downs eam exp ession o
obacco-selec ed Ni H a ian s in yeas o biochemical cha ac e -
iza ion. The genes we e unde con ol o he s ong and
cons i u i e E35S p omo e . Amino- e minal COX4-TS ex ensions
we e added o Ni H p o eins. COX4 is he 29 amino acid ansi
pep ide o he S. ce e isiae mi ochond ia p o ein cy och ome c
oxidase subuni IV (MLSLRQSIRFFKPATRTLCSSRYLLQQKP),
whe eas TS deno es he 28 amino acid Twin-S ep-Tag pep ide
(WSHPQFEKGGGSGGGSGGSAWSHPQFEK)22. COX4 a ge ed
Ni H p o eins o he mi ochond ia ma ix and TS was used o
enable a ian -independen immunoblo de ec ion o Ni H and o
acili a e i s pu ifica ion. Impo an ly, he TS- ag has been shown
o no significan ly a ec Ni H unc ionali y12. COX4-TS-Ni H
a ian s a e he ea e deno ed as NbNi HXx whe e Nb s ands o
he hos N. ben hamiana,Xx deno es a ian s collec i ely, and
o he supe sc ip s indica e he species om which Ni H sequence
was ob ained. Vec o s wi h NbNi HXx cons uc s addi ionally
con ained a ansc ip ional uni o exp ession o he g een
fluo escen p o ein (GFP) ha was used as indica o o success ul
lea infil a ion (Supplemen a y Table 1).
An auxilia y ec o was cons uc ed o co-exp ess A. inelandii
ni M,ni U, and ni S and a ge hei p o ein p oduc s o
mi ochond ia ia N- e minal SU9 ex ensions. Simila o COX4,
he mi ochond ial p esequence o subuni 9 o he Neu ospo a
c assa F
0
-ATPase23 (SU9) has been shown o deli e Ni p o eins
o N. ben hamiana mi ochond ia24. Ni U and Ni S assemble [Fe-
S] clus e s des ined o Ni p o eins in A. inelandii25. While no
essen ial o exp ession o unc ional Ni HA in S. ce e isiae
mi ochond ia10 hey we e equi ed o gene a e high amoun s o
ac i e Ni B in yeas 11. As we aimed o iden i y Ni H a ian s
accumula ing a highe le els han Ni HA , Ni UA and Ni SA
we e included in his s udy. In A. inelandii and o he well-
s udied diazo ophs Ni M is in ol ed in Ni H olding o
dime iza ion p io [Fe
4
S
4
] clus e acquisi ion9,26. Despi e ni M
no being p esen in o ganisms o some selec ed ni H a ian s
(Supplemen a y Da a 1), his gene was always included in
infil a ion expe imen s o consis ency.
Iden ifica ion o Ni H p o eins sui able o exp ession in N.
ben hamiana.N. ben hamiana lea es we e co-infil a ed wi h a
1:1:1 mix u e o h ee dis inc A. ume aciens cul u es o
exp ession o , espec i ely, one NbNi HXx a ian plus GFP, he
auxilia y p o eins NbNi MA ,NbNi UA , and NbNi SA , and he
ARTICLE COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-020-01536-6
2COMMUNICATIONS BIOLOGY | (2021)4:4 | h ps://doi.o g/10.1038/s42003-020-01536-6 | www.na u e.com/commsbio
RNA silencing supp esso p19 o enhance he ni ansgene
exp ession (Fig. 1b)27. P o ein ex ac s we e p epa ed om he N.
ben hamiana lea es h ee days a e infil a ion and analyzed o
accumula ion o soluble NbNi HXx using an ibodies ecognizing
he TS- ag. Only wo Ni H a ian s we e consis en ly de ec ed
among expe imen s (Fig. 1c, Supplemen a y Fig. 1a), namely
hose o igina ing om Me hanocaldococcus in e nus (NbNi HMi)
and Hyd ogenobac e he mophilus (NbNi HH ). A hi d Ni H
a ian om Me hano he mobac e ma bu gensis (NbNi HMm)
was de ec ed a low le els a one occasion. In con as , analysis o
o al ex ac s p epa ed om he infil a ed obacco lea es showed
ha , al hough accumula ion le els o he NbNi HXx p o eins
a ied significan ly, 25 o he 32 a ian s could be de ec ed
(Supplemen a y Fig. 1b). Only NbNi H exp ession o a ian s
om B ady hizobium japonicum,Rhizobium leguminosa um b .
i olii,He baspi illum se opedicae,Gloeo hece sp. KO68DGA,
Rhodopseudomonas palus is,Me hano he mobac e he mau o-
ophicus, and F ankia sp. (s ain FaC1) could no be demon-
s a ed. Sequence alignmen s and 3D-modeling o Ni HMi,
Ni HH , and Ni HMm a e shown in Supplemen a y Fig. 2. The
3D-models did no e eal any specific ea u e ha would explain
hei supe io accumula ion as soluble p o ein in obacco mi o-
chond ia, bu all h ee p o eins o igina e om he mophilic
o ganisms (Supplemen a y Da a 1) which could possibly explain
hei s abili y and solubili y.
Ac i i y o Ni H a ian s isola ed om mi ochond ia o
ae obically cul u ed S. ce e isiae.N. ben hamiana sc eening-
iden ified a ian s and Ni HA we e exp essed in S. ce e isiae and
pu ified by S ep- ag a fini y ch oma og aphy (STAC) o e alua e
unc ionali y when a ge ed o mi ochond ia. Fo his, genes
encoding COX4-TS-Ni H cons uc s we e ans e ed o exp es-
sion ec o s oge he wi h su9-ni MA ,su9-ni UA , and su9-ni SA
unde he con ol o galac ose-inducible GAL1 o GAL10 p o-
mo e s (Supplemen a y Table 2, Supplemen a y Fig. 3a). These
COX4-TS-Ni H a ian s exp essed in ae obic S. ce e isiae cul-
u es a e he ea e deno ed ScNi HMm,ScNi HMi,ScNi HH , and
ScNi HA (ScNi HXx collec i ely).
While ScNi HMm,ScNi HMi, and ScNi HH we e pu ified o
nea homogenei y (Fig. 2a), SDS-PAGE analysis o ScNi HA
showed addi ional slowe mig a ing co-elu ing p o eins. Mass
spec ome y confi med ha hese we e con aminan s (Fig. 2a).
ScNi HA solubili y was low and much p o ein was los o he
pelle ac ion when p epa ing he soluble cell- ee ex ac (CFE)
explaining i s poo pu ifica ion yield (abou 11 mg pe kg o S.
ce e isiae cells) (Supplemen a y Fig. 3b–e, Supplemen a y
Table 3). The yield o ScNi HMm was also ela i ely low, in line
wi h he in e io esul in he N. ben hamiana sc eening. In
con as , he yields o ScNi HMi and ScNi HH we e ca. 20 imes
highe . I on (Fe) quan ifica ion o pu ified samples was a iable
bu indica ed ha ScNi HH was isola ed la gely as holo-p o ein
con aining one [Fe
4
S
4
] clus e pe dime (Supplemen a y Table 3).
Consis en ly, immunoblo analysis showed ha ScNi HH ,
ScNi MA ,ScNi UA , and ScNi SA had been e ficien ly a ge ed
o he mi ochond ia (Supplemen a y Fig. 4).
Ac i i ies o pu ified ScNi HXx a ian s we e de e mined
in i o using he ace ylene educ ion assay (ARA) and compa ed
Euka ya
Eu ya chaeo a (3)
C ena chaeo a
Nanoa chaeo a
Thauma chaeo a
Ko a chaeo a
P o eobac e ia (15)
(8) (2)
(2)
(3)
Ac inobac e ia (1)
Fi micu es (4)
Bac e oides
Aqui icae (1)
Chlo o lexi (1)
Chlo obi (1)
Cyanobac e ia (6)
LUCA
Bac e ia (29)
The mo ogae
To al (32)
ab
48 -
35 -
Ponceau
28 -
α-GFP
48 -
35 -
α-Ni S
35 -
α-Ni U
35 -
α-Ni M
35 -
kDa
C l
S ep-HRP
(NbNi HXx)
: Ni H
: Ni U/Ni S/Ni M
c+++++
-A Mm Mi H
Ni H lib a y design
(N. ben hamiana)
s nai a H iN 3s nai a H iN 23
Pu i ica ion
Scale up
(I)
(III)
Lea in il a ion
(A. ume aciens)
Ni H pu i ica ions
(S. ce e isiae)(II)
Ni H + GFP (pN2XJ81-pN2XJ112)
Ni U/Ni S/Ni M (pN2XJ165)
p19 (GB1203)
M. ma bu gensis
M. in e nus
H. he mophilus
Solubili y sc eening
(N. ben hamiana)
H. he mophilus
Ni H
Ni H cha ac e iza ion
(
H. he mophilus
)
In i o complemen a ion
(A DJ77)
(V)
3 Ni H a ian s
Ni H + Ni U/Ni S/Ni M
H. he mophilus (XJ1Y)
M. ma bu gensis (XJ2Y)
M. in e nus (XJ3Y)
Ace ylene educ ion analysis
(yeas pu i ied Ni H)
H. he mophilus
Ni H unc ional
N
i H inac i e
Ni H + Ni U/Ni S/Ni M
+ GFP + p19
(pAE382)
Ac i i y analysis
( obacco pu i ied
H. he mophilus
Ni H)
Fe e iliza ion
Ex ended da k pe iod
(N. ben hamiana)
(VIII)
N
i H ac i e
Ni H + Ni U/Ni S/Ni M
+ GFP + p19
(pAE382)
H. he mophilus
Ni H pu i ica ion
(N. ben hamiana)
(VII)
(IV)
[Fe
4
S
4
] analysis
(yeas Ni H)
i) P-clus e ma u a ion
ii) FeMo-co syn hesis
iii) N
2
educ ion
(yeas Ni H)
(VI)
Fig. 1 Ni H lib a y design and sc eening. a Phylogene ic ee indica ing he dis ibu ion and numbe o he es ed Ni H a ian s among bac e ial phyla
(blue) and a chaeal lineage (magen a). bExpe imen al wo kflow o his s udy. cImmunoblo s o N. ben hamiana soluble p o ein ex ac s de eloped wi h
an ibodies agains TS, GFP, Ni U, Ni S, and Ni M (s ep I o panel b). GFP de ec ion was used as con ol o no malize e ec i eness o infil a ion
expe imen s. Ponceau s aining panel is shown as loading and memb ane ans e con ol. Do ed line indica es di e en exposu es o he same memb ane.
Cell- ee ex ac s o A. inelandii DJ ( o Ni U and Ni S) o E. coli Rose a (DE3) o e exp essing Ni MA ( o Ni M, as Ni M exp ession le els in A. inelandii is
low) we e used as size con ols. Unc opped immunoblo s a e shown in Supplemen a y Fig. 8.
COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-020-01536-6 ARTICLE
COMMUNICATIONS BIOLOGY | (2021)4:4 | h ps://doi.o g/10.1038/s42003-020-01536-6 | www.na u e.com/commsbio 3
o ha o Ni H pu ified om A. inelandii (deno ed Ni HA ). In
all cases Ni DK pu ified om A. inelandii (deno ed Ni DKA )
was used as MoFe p o ein componen . ScNi HA ac i i y was 85%
o Ni HA (Fig. 2b), suppo ing p e ious obse a ions ha STAC
is sui able o pu ifica ion o me al-clus e con aining Ni p o eins
exp essed in yeas 11,28.ScNi HH specific ac i i y was abou hal
o ScNi HA , while ScNi HMm and ScNi HMi showed e y low
ac i i ies (Fig. 2b). The assay did no de e mine whe he lowe
ac i i ies we e due o Ni H a ian de ec s, o o incompa ibili y
wi h ScNi USA in i o ( esul ing in apo-Ni H p o ein wi h low
[Fe
4
S
4
] clus e occupancy) o Ni DKA in i o ( esul ing in poo
elec on dona ion). Recons i u ion o ScNi HMm [Fe
4
S
4
] clus e s
in i o by ei he mixing wi h Fe, L-cys eine, DTT, and EcNi SA
(di ec econs i u ion) o by incuba ing wi h [Fe
4
S
4
] clus e -
loaded EcNi UA (Ni U-media ed econs i u ion) did no ac i a e
he p o ein (Supplemen a y Fig. 5), indica ing ha his Ni H
a ian is no compa ible wi h Ni DKA . In con as , ScNi HMi
was ac i a ed o some ex en by Ni UA , and u he by di ec
econs i u ion, indica ing ha he A. inelandii Ni US machine y
is no op imal o Ni HMi (Supplemen a y Fig. 5). Howe e ,
ac i i ies we e e y low compa ed o he as-isola ed ScNi HH
p o ein (Fig. 2b). This could be explained by Ni HH ha bo ing
mo e o he conse ed amino acid esidues known o be
impo an o he in e ac ion wi h Ni DKA (Supplemen a y
Fig. 2).
Impo an ly, soluble accumula ion o ScNi HH in mi ochon-
d ia was 20- old highe han ScNi HA (Supplemen a y Table 3),
which ansla es in o a leas 10- old highe in i o ac i i y and
ulfills Ni H quan i y equi emen s o ni ogenase enginee ing.
Thus, ScNi HH was u he cha ac e ized.
ScNi HH exhibi s Ni H-cha ac e is ic spec oscopic signals
and is unc ional in i o. Pu ified ScNi HH p o ein p esen ed
ul a iole – isible (UV– is) abso p ion spec a ypical o O
2
-
sensi i e [Fe-S] clus e -con aining p o eins (Fig. 3a). Amino-
e minal sequencing e ealed ha amino acid esidues EQKP
emained a e COX4 p ocessing (Fig. 3b), whe e con e sion o
glu amine (Q) o glu amic acid (E) could be due o deamina ion
pe o med by he mi ochond ial ma ix N- e minal amidase
NTA129. Elec on pa amagne ic esonance (EPR) confi med ha
ScNi HH p o ein con ained an [Fe
4
S
4
] clus e wi h simila signal
in ensi y and g- alues as Ni HA (Fig. 3c), sugges i e o success ul
ma u a ion in o unc ional Fe p o ein.
The Ni H a ian chosen o enginee N
2
-fixing plan s mus
pe o m P-clus e ma u a ion and FeMo-co biosyn hesis in
addi ion o se e as elec on dono o subs a e educ ion. We
he e o e es ed whe he H. he mophilus Ni H could e e he
Ni −pheno ype o A. inelandii DJ77 (Δni H s ain)30. Fo his,
s-ni HH was in oduced by ans o ma ion in o DJ77 and he
esul ing s ain UW481 was es ed o diazo ophic g ow h and
in i o ace ylene educ ion ac i i y. UW481 showed diazo ophic
g ow h bo h in solid and liquid media (Supplemen a y Fig. 6a, b),
and immunoblo analysis demons a ed sus ained A Ni HH
exp ession and ace ylene educing ac i i y indica i e o ac i e
ni ogenase (Supplemen a y Fig. 6c, d). These da a s ongly
indica e ha Ni HH can eplace he unc ions o na i e A.
inelandii Ni H o some ex en , which equi es p oduc i e
in e ac ions wi h a leas apo-Ni DKA , Ni DKA , and Ni ENA
p o eins.
ScNi HH is ac i e in subs a e educ ion, P-clus e o ma ion
and FeMo-co syn hesis. Each indi idual Ni H-dependen ac i i y
was hen analyzed in i o using pu e ScNi HH p epa a ions
(Fig. 3d). P-clus e ma u a ion was de e mined by supplemen ing
CFE o A. inelandii DJ77 (Δni H) wi h ScNi HH . The DJ77
ex ac is de oid o FeMo-co and con ains inac i e apo-Ni DKA
wi h imma u e P-clus e s. The P-clus e ma u a ion assay using
DJ77 CFE elies on posi i e ou comes o h ee dis inc ac i i ies
pe o med in wo sequen ial eac ions (Fig. 3d). In he fi s
eac ion (S ep I +II) pu e Ni H and FeMo-co a e added o DJ77
CFE esul ing in Ni H-dependen educ i e coupling o he wo
[Fe
4
S
4
] P-clus e p ecu so s o o m ma u e P-clus e s (S ep I),
ollowed by FeMo-co inse ion in o P-clus e con aining apo-
Ni DKA o gene a e ac i e Ni DKA (S ep II) (Fig. 3d). Te a-
hiomolybda e is hen added o p e en u he FeMo-co inse -
ion, sepa a ing he ma u a ion (S ep I +II) and ac i i y (S ep III)
eac ions. Ac i a ion o DJ77 apo-Ni DKA by ScNi HH
demons a ed i s P-clus e ma u a ion ac i i y (Fig. 3e).
In i o FeMo-co syn hesis (Fig. 3d, S ep II)9was de e mined
by combining pu ified p epa a ions o ScNi HH , apo-Ni DKA
con aining P-clus e s bu de oid o FeMo-co31, apo-Ni ENA
con aining pe manen [Fe
4
S
4
] clus e s bu lacking FeMo-co
p ecu so 32, Mo, homoci a e, and ei he he FeMo-co p ecu so
ba
kDa
130 -
100 -
75 -
63 -
48 -
35 -
28 -
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
1
2
3
4
5
Coomassie
ScNi HA
ScNi HMi
ScNi HMm
ScNi HH
A Mm Mi H
-Ni H
0
500
1000
1500
2000
2500
E hylene p oduc ion
(nmol C
2
H
4
·min
-1
·mg
-1
o Ni DK
A
)
ScNi H
A Mm Mi H
-Ni H
0
10
20
30
O e
O e
Fig. 2 Ac i i y o selec ed Ni H a ian s pu ified om mi ochond ia o ae obically cul u ed S. ce e isiae enginee ed s ains. a STAC-pu ified ScNi HA
(s ain XJ4Y), ScNi HMm (s ain XJ2Y), ScNi HMi (s ain XJ3Y), and ScNi HH (s ain XJ1Y) p o eins. The low solubili y o ScNi HA p omo ed binding o
con amina ing p o eins (as column was no sa u a ed wi h ScNi HA ) ha we e iden ified by pep ide mass finge p in ing as: 1) Acc1p (N1P4Q3), 2 and 3)
py u a e ca boxylase (N1P377), and 4) HSP70 (D2J4C2). No iden ifica ion was possible o band numbe 5. Unc opped gels a e shown in Supplemen a y
Fig. 9. bARA o STAC-pu ified ScNi HXx a ian s. Ac i i y using Ni HA and Ni DKA (posi i e con ol) was 2406 ± 53 uni s (nmol e hylene o med pe min
and mg o Ni DKA ). Da a ep esen mean alues (n=2 echnical eplica es).
ARTICLE COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-020-01536-6
4COMMUNICATIONS BIOLOGY | (2021)4:4 | h ps://doi.o g/10.1038/s42003-020-01536-6 | www.na u e.com/commsbio
Ni B-co bound o he ca ie p o ein Ni XA 33 o Ni B p o ein
supplemen ed wi h Fe and S34. As o he P-clus e ma u a ion
assay, e a hiomolybda e was added be o e he ARA (Fig. 3d,
S ep III). Figu e 3 shows ha ScNi HH suppo ed FeMo-co
syn hesis in i o. Impo an ly, ScNi HH and ScNi BM (Me ha-
no he mobac e he mau o ophicus Ni B isola ed om S. ce e i-
siae)11 ac ed oge he in he Ni B-dependen in i o FeMo-co
syn hesis assay in which Ni B-co was concomi an ly syn hesized
by ScNi BM a he han added in pu ified o m. This esul
p o ed compa ibili y o wo essen ial p o eins o N
2
fixa ion,
ScNi HH and ScNi BM , when p oduced in yeas mi ochond ia. I
also showed in e species compa ibili y wi h Ni DKA and
Ni ENA , al oge he cons i u ing he conse ed biochemical co e
o ni ogenase.
ScNi HH ac i i y in subs a e educ ion was demons a ed by
he ARA and by educ ion o N
2
in o NH
3
. ARA i a ion was
ca ied ou wi h a fixed quan i y o Ni DKA and inc easing
amoun s o ScNi HH . Maximum Ni DKA ac i i y was achie ed
a mola ScNi HH o Ni DKA a ios la ge han 40 (Fig. 3g),
simila o eac ions wi h he na u al coun e pa Ni HA 35. This
esul sugges s ha he maximum ac i i y ha can be achie ed
combining ScNi HH wi h Ni DKA is 1000 uni s (i.e., hal o he
ac i i y wi h Ni HA ). In addi ion, ScNi HH suppo ed N
2
educ ion in o NH
3
by Ni DKA . Impo an ly, he a io o NH
3
o e hylene p oduced by Ni DKA was simila independen ly o
using Ni HA o ScNi HH (Fig. 3h).
As-isola ed NbNi HH was inac i e bu could be ac i a ed by
[Fe
4
S
4
] clus e econs i u ion.NbNi HH was pu ified om A.
ume aciens-infil a ed lea es o N. ben hamiana. Plan s we e
g own unde long-day condi ions (16 h ligh /8 h da k) and lea es
we e p ocessed a he end o he da k pe iod. Genes encoding
K
K
D
D
K
K
D
D
FeMo-co
K
K
D
D
P-clus e
[Fe
4
S
4
]
Open FeMo-co si e
HH
H
H
MoFe p o ein
Fe p o ein
ac
S ep I
P-clus e ma u a ion
S ep II
FeMo-co biosyn hesis
N
2
C
2
H
2
HH
S ep III
Ni ogenase ac i i y
d
300 400 500 600 700 800
0
0.2
0.4
0.6
0.8
1.0
Wa eleng h (nm)
No malized abso bance
As-isola ed
Ai -exposed
YLL EQKPSAWSH
TS- ag
COX4
300 400390380370360350340330320310
B (mT)
EPR in ensi y (a.u.)
ScNi HH
Ni HA
g = 2.04 g = 1.94
g = 1.84
g = 2.05 g = 1.94
g = 1.86
NH
3
C
2
H
4
hg e
E hylene p oduc ion
(nmol C2H4 · min-1 · assay-1)
S ep I+II:
S ep III:
+
+
Ni HA ScNi HH
-
-+
-+
+
-
-
+
-
0
5
10
15
20
b*
0
100
200
300
E hylene p oduc ion
(nmol C2H4 · min-1 · mg-1 o Ni DKA )
Ni B-co + Ni HA
Ni B-co + ScNi HH
Ni BA + Ni HA
ScNi BM + ScNi HH
ScNi BM
ScNi HH
5
10
20
40
60
80
100
120
140
160
0
250
500
750
1000
1250
Ra io ScNi HH : Ni DKA
E hylene p oduc ion
(nmol C2H4·min
-1·mg-1o Ni DKA )
Ni H
A
ScNi H
H
0
300
600
900
1200
1500
1800
2100
0
150
300
450
600
E hylene p oduc ion
(nmol C
2
H
4
·m
in
-1
·mg
-1
o Ni DK
A
)
Ammonia p oduc ion
(nmol NH
3
·mi
n
-1
·m
g
-1
o Ni DK
A
)
Fig. 3 Cha ac e iza ion o ScNi HH .aUV– is abso p ion spec a o as-isola ed and ai -exposed ScNi HH .bP ocessing si e (black a ow) o COX4
mi ochond ia a ge ing signal (blue) as de e mined by N- e minal sequencing o ScNi HH . The con e sion o Q o E (ma ked by *) could be due o
deamina ion p ocesses. cEPR signal o as-isola ed ScNi HH (190 μM) compa ed o Ni HA (71.2 μM). Expe imen al da a (black lines) and simula ions (blue
lines) o each p o ein a e shown oge he wi h g alues. dSchema ic ep esen a ion o Ni H-dependen ac i i ies es ed in e–h. Figu e adap ed om Bu én
e al.9. Copy igh 2020 ACS unde CC BY 4.0 h p://c ea i ecommons.o g/licenses/by/4.0.eIn i o P-clus e ma u a ion o apo-Ni DKA p esen in CFE
o A. inelandii DJ77 (Δni H) a e addi ion o ScNi HH o Ni HA as indica ed (±). Te a hiomolybda e was added ( ollowing s ep I +II) o inhibi u he
apo-Ni DKA ac i a ion du ing he ARA (s ep III). Da a ep esen mean alues (n=2 echnical eplica es). Ni B-co and Ni B-dependen in i o FeMo-co
syn hesis using ScNi HH o Ni HA . Te a hiomolybda e was added ( ollowing s ep II) o inhibi u he apo-Ni DKA ac i a ion du ing he ARA (s ep III). A
20:1 mola a io o ScNi HH o Ni DKA was used in he ARA (s ep III). Da a ep esen mean alues (n=2 echnical eplica es (Ni B-co +Ni HA ), n=3
echnical eplica es (Ni B-co +ScNi HH ), n=4 echnical eplica es (Ni BA +Ni HA ), n=5 echnical eplica es (ScNi BM +ScNi HH ), n=3 echnical
eplica es (ScNi BM ), n=4 echnical eplica es (ScNi HH )). Blue and ed do s co espond o independen expe imen s. gTi a ion o Ni DKA ac i i y wi h
ScNi HH . Posi i e con ol eac ions pe o med wi h Ni HA and Ni DKA a 40:1 mola a io ga e 1692 ± 4 uni s (nmol e hylene o med pe min and mg o
Ni DKA ). Reac ions lacking Ni H (nega i e con ol) ga e 2.5 ± 0.8 uni s. Da a ep esen mean alues (n=2 echnical eplica es). hARA ( ed do s, le
y-axis) and N
2
- educ ion assay (blue squa es, igh y-axis) using ScNi HH and Ni DKA (s ep III). Ni HA was used as con ol. Da a ep esen mean alues
(n=4 echnical eplica es).
COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-020-01536-6 ARTICLE
COMMUNICATIONS BIOLOGY | (2021)4:4 | h ps://doi.o g/10.1038/s42003-020-01536-6 | www.na u e.com/commsbio 5

NbNi HH ,NbNi MA ,NbNi UA , and NbNi SA ( oge he wi h
p19 and GFP) we e piled up in a single plan -exp ession ec o
o co-exp ession (Me hods sec ion and Supplemen a y Table 1).
Pu ified NbNi HH did no exhibi b own colo o [Fe-S] clus e s
and was inac i e in he ARA when combined wi h Ni DKA
(Fig. 4a). The e o e, we econs i u ed NbNi HH [Fe
4
S
4
] clus e
in i o ei he by mixing wi h Fe, L-cys eine, DTT, and EcNi SA
(di ec econs i u ion), o by incuba ing wi h [Fe
4
S
4
] clus e -
loaded EcNi UA (Ni U-media ed econs i u ion). Bo h me hods
ac i a ed he NbNi HH as de e mined by he ARA (Fig. 4a),
demons a ing ha he p o ein was co ec ly olded bu lacked i s
[Fe
4
S
4
] clus e . This esul sugges ed ha inse ion and/o s abi-
li y o NbNi HH [Fe
4
S
4
] clus e was poo in mi ochond ia o
lea es.
cb
kDa
QQKPS AWSHPQ FEK
TS- agCOX4
130 -
100 -
75 -
63 -
48 -
35 -
28 -
180 -
*
Coomassie
d
S ep -HRP (NbNi HH )
kDa
8 h 16 h
Ligh Da k
Ponceau
LighLighLighLighLigh Da kDa kDa kDa k
NbNi HH
kDa
180 -
α-GFP
QQKPSA WSHPQ FEKGG
TS- agCOX4
130 -
100 -
75 -
63 -
48 -
35 -
28 -
Coomassie
e
12345
a
As-isola ed
Di ec
EcNi UA
- Ni H
NbNi HH pu .:
NbNi UA
kDa
48 -
35 -
48 -
35 -
48 -
35 -
S ep-HRP
(NbNi UA )
α-Ni S
Ponceau
Fe e iliza ion
+-
E hylene p oduc ion
(nmol C2H4 · min-1 · mg-1 o Ni DKA )
Recons .
h
g
28 -
48 -
28 -
35 -
35 -
63 -
100 -
75 -
48 -
35 -
63 -
75 -
E hylene p oduc ion
(nmol C2H4 · min-1 · mg-1 o Ni DKA )
0
10
20
30
Da k
0
100
200
300
400
500
ARTICLE COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-020-01536-6
6COMMUNICATIONS BIOLOGY | (2021)4:4 | h ps://doi.o g/10.1038/s42003-020-01536-6 | www.na u e.com/commsbio
Fe e iliza ion o he soil inc eases soluble Ni U in mi o-
chond ia o N. ben hamiana. One explana ion o he low
[Fe
4
S
4
] clus e con en o NbNi HH could be insu ficien Fe
a ailabili y in he soil. We obse ed ha accumula ion o NbNi-
UA , bu no o NbNi SA , inc eased when he wa e used o
i iga e he A. ume aciens-infil a ed plan s was supplemen ed
wi h Fe (Fig. 4b). Sul u was no supplemen ed in soil as he
infil a ion solu ion con ained Mg
2
SO
4
. Al hough Fe e iliza ion
ipled he yield o STAC-isola ed NbNi UA , he a e age Fe
con en o 2 Fe a oms pe p o ein was no a ec ed (Supple-
men a y Fig. 7a, b, Supplemen a y Table 4). This could be due o
he loss o ansien Ni U [Fe-S] clus e s du ing pu ifica ion, and
i is no a su p ising ou come as isola ion o EcNi UA con aining
only he pe manen [Fe
2
S
2
] clus e s has been p e iously
obse ed36. Immunoblo s de ec ed wo di e en ly mig a ing
NbNi UA species in pu ifica ions om obacco lea es (Fig. 4c).
Amino- e minal sequencing showed ha bo h species we e
clea ed ei he one o se en amino acids in o he TS- ag (Fig. 4d).
As bo h NbNi UA species showed he same N- e mini p oces-
sing, we concluded ha he as e mig a ing polypep ide was
unca ed a he C- e minus.
Ex ended da k pe iod combined wi h Fe e iliza ion p oduced
ac i e NbNi HH in mi ochond ia o N. ben hamiana lea es.
The soil o N. ben hamiana plan s exp essing NbNi HH was
e ilized wi h Fe o inc ease Fe a ailabili y. In addi ion, he da k
pe iod p eceding lea ha es was ex ended om 8 h o 16 h
hypo hesizing ha longe da kness would lowe in acellula O
2
and s abilize NbNi HH [Fe
4
S
4
] clus e . Da k pe iod ex ension did
no inc ease NbNi HH accumula ion (Fig. 4e) bu allowed o
isola ion o ac i e p o ein as shown below. Abou 6 mg o
NbNi HH was consis en ly isola ed pe kg o N. ben hamiana
lea es (Fig. 4 , Supplemen a y Fig. 7c, Supplemen a y Table 5).
Amino- e minal sequencing showed ha mi ochond ia- a ge ed
NbNi HH accumula ed as wo species (simila o NbNi UA ), one
in which wo amino acid esidues om he TS- ag we e emo ed
wi h he COX4 signal and ano he ha was p ocessed fi e amino
acid esidues u he in o he TS- ag (Fig. 4g).
Func ionali y o NbNi HH isola ed om lea es o Fe e ilized
obacco plan s ollowing 16 h o da kness was de e mined using
ARA. NbNi HH p epa a ions consis en ly showed ac i i ies bu
hese we e low compa ed o hose o [Fe
4
S
4
] clus e - econs i u ed
NbNi HH (Fig. 4h). This esul sugges ed ha NbNi HH
accumula ed as wo species in obacco mi ochond ia, whe e
inac i e p o ein likely lacking [Fe
4
S
4
] clus e was mo e abundan
han unc ional and [Fe
4
S
4
] clus e -con aining NbNi HH . Con-
sis en ly, he Fe con en o pu ified NbNi HH p epa a ions we e
below de ec ion limi (Supplemen a y Table 5). Al oge he he
esul s indica e ha while soluble NbNi HH accumula es in good
quan i y in mi ochond ia o N. ben hamiana lea es, enginee ing
o addi ional p o ein componen s o biosyn he ic pa hways will
be equi ed o imp o e [Fe
4
S
4
] clus e acquisi ion o s abili y.
Discussion
The fi s s udy epo ing p oduc ion o ac i e Ni H in yeas
p o ed ha mi ochond ia is a sui able o ganelle o hos ing O
2
-
sensi i e Ni p o eins unde ae obic g ow h condi ions10. Despi e
being a alid p oo -o -concep , u he de elopmen s wi h A.
inelandii Ni H we e limi ed by low yields as only a small po ion
was soluble in he mi ochond ial ma ix. Simila solubili y issues
we e la e epo ed o K. oxy oca Ni H a ge ed o N. ben-
hamiana mi ochond ia16 and a e confi med in his s udy using
immunoblo sc eening and STAC. Iden i ying he bes possible
Ni H p o ein o euka yo ic (plan ) exp ession was he e o e o
u e mos impo ance. Ni H is he mos abundan Ni p o ein
equi ed o N
2
fixa ion in A. inelandii17. Besides being he Fe
p o ein componen o Mo ni ogenase, Ni H is essen ial o he
assembly o bo h Ni DK co ac o s, namely he P-clus e and he
FeMo-co9.
Ni H p o eins o ni ogenase enginee ing in plan s should: (i)
be s able and soluble a high le els in he mi ochond ial ma ix,
and (ii) be compa ible wi h he Ni DK componen om a well-
s udied model-diazo oph i hei own Ni DK componen s a e
no a ailable in pu ified o m. Compa ibili y is impo an when
e alua ing unc ion o candida e Ni H a ian s. In ou case i
mean ha any selec ed Ni HXx should be compa ible wi h Ni -
MA (i Ni HXx is no Ni M-independen ), Ni USA o ma u a-
ion and [Fe
4
S
4
] clus e syn hesis/inse ion, and Ni DKA o
ni ogenase ac i i y measu emen s. We no e ha his equi e-
men in oduces a selec ion bias and ha he sc eening could
ha e o e looked Ni H a ian s ha we e supe io o ha o H.
he mophilus i combined wi h di e en Ni DK.
The Ni H a ian s es ed in his s udy we e selec ed om a
cu a ed da ase o hund eds o Ni H sequences by a o ing
ae obic o plan -associa ed o igins, o o e come he inhe en O
2
-
sensi i i y o Ni H, and unc ionali y a mode a e empe a u es.
We also hypo hesized ha Ni H a ian s om a chaea could
unc ion be e in a euka yo ic en i onmen as his domain o li e
is belie ed o be mo e closely ela ed o he Euka yo a37, and
because ou p e ious wo k exp essing a chaeal Ni B a ian s in
yeas had shown hem o be supe io o hose o bac e ial o igin11.
We expec ed ha mos Ni H a ian s would be pa ly soluble
in obacco mi ochond ia when exp essed oge he wi h he
accesso y p o eins Ni UA , Ni SA , and Ni MA . Howe e , only
Ni H om M. in e nus and H. he mophilus we e consis en ly
de ec ed in soluble obacco ex ac s, in addi ion o M. ma bu -
gensis ha was occasionally de ec ed a lowe le els. Two o hese
Ni H p o eins o igina ed om a chaea and he hi d om a
bac e ium. One possibili y could be ha he Ni MA p o ein was
no exp essed a su ficien le els in he obacco mi ochond ia and
ha only hese h ee Ni H a ian s did no equi e Ni M o
ma u a ion. Howe e , low le els o Ni M exp ession appea o be
enough o Ni H ma u a ion in K. oxy oca38,39. A mo e plausible
explana ion can be ound in he he mophilic na u e o M.
in e nus,H. he mophilus, and M. ma bu gensis. I has ecen ly
Fig. 4 Cha ac e iza ion o he NbNi HH p o ein. a Ac i a ion o as-isola ed NbNi HH p o ein wi h [Fe
4
S
4
] clus e s ei he by di ec chemical syn hesis o
by EcNi UA -media ed econs i u ion. Ac i i y using Ni HA (posi i e con ol) was 1773 ± 10 uni s (nmol e hylene o med pe min and mg o Ni DKA ). Da a
ep esen mean alues (n=2 echnical eplica es). bImmunoblo s showing he e ec o Fe e iliza ion on NbNi UA and NbNi SA o al p o ein exp ession.
Ni S an ibody was p obed on he same memb ane a e incuba ion wi h S ep ac in-HRP. cNbNi UA p o ein pu ified om obacco plan s. The lowe band
(ma ked by *) indica es a as e mig a ing NbNi UA polypep ide. dP ocessing si es (black a ows) o he COX4 mi ochond ia a ge ing signal (blue) as
de e mined by N- e minal sequencing o he ull-leng h and he as e mig a ing NbNi UA .eImmunoblo s showing accumula ion o NbNi HH a he end o
8 h o 16 h nigh (da k pe iod). NbNi HH p o ein pu ified om obacco lea es. gP ocessing si es (black a ows) o COX4 signal (blue) as de e mined by
N- e minal sequencing o NbNi HH .hARA o fi e independen NbNi HH STAC-pu ifica ions (1–5). Measu ed ac i i ies using Ni HA (posi i e con ols) and
wi hou Ni H (nega i e con ols) we e, espec i ely, 2406 ± 53 and 4.6 ± 1.3 (pu ifica ion 1), 1773 ± 10 and 3.2 ± 2.2 (pu ifica ions 2 and 3), and 1692 ± 3.9
and 2.5 ± 0.8 (pu ifica ions 4 and 5). All ac i i ies a e in nmol e hylene o med pe min and mg o Ni DKA . Da a ep esen mean alues (n=2 echnical
eplica es). Unc opped immunoblo s and gels a e shown in Supplemen a y Fig. 10.
COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-020-01536-6 ARTICLE
COMMUNICATIONS BIOLOGY | (2021)4:4 | h ps://doi.o g/10.1038/s42003-020-01536-6 | www.na u e.com/commsbio 7
been epo ed ha he empe a u e inside espi ing mi ochond ia
o cul u ed human cells is a ound 50 °C40, e en when he ex e nal
medium is main ained a 38 °C. Whe he he same d as ic e ec
on empe a u e holds ue o mi ochond ia o a lea cell is no
known o us, bu i could explain in pa he ou come o ou Ni H
sc eening. None o he wo highes exp essed Ni H p o eins
o igina ed om p o en diazo ophs. We a e no awa e o any
s udy in es iga ing diazo ophy in he a chaeon M. in e nus.
Howe e , Ni BMi cu ed he Ni −pheno ype o an A. inelandii
ni B mu an s ain41 and, as Ni B has no o he known unc ion
han biosyn hesis o ni ogenase ac i e-si e co ac o s, i is likely
ha M. in e nus is in ac a diazo oph. On he o he hand, N
2
-
fixa ion has been es ed bu no obse ed in H. he mophilus TK-
642. In e es ingly only six Ni H a ian s in ou lib a y o igina ed
om o ganisms ha ing genes wi h high simila i y o A. inelandii
ni M. Pe haps o he p olyl isome ases could subs i u e o Ni M
in hese o ganisms. Whe he Ni M (and he Ni US machine y) is
equi ed o ma u a ion o he h ee selec ed Ni H p o eins
(especially Ni HH ) in mi ochond ia will be in es iga ed in
u u e wo k.
Mi ochond ia-exp essed ScNi HH was he only a ian ha
suppo ed ele an ni ogenase ac i i y when combined wi h
Ni DKA . I s ac i i y co esponded o oughly hal o ha using
ScNi HA e en i he ScNi HH o Ni DKA mola a io was
inc eased well abo e 40 no mally used o ARA. Eme ich and
Bu is showed ha Ni H p o eins can unc ion wi h Ni DK om
o he o ganisms35, bu his s udy only combined p o eins om
bac e ia. An op imal g ow h empe a u e o 72 °C has been
epo ed o H. he mophilus TK-643, which could explain lowe
ScNi HH ac i i y in subs a e educ ion assays. Howe e , ou
p edic ion om his s udy and p e ious wo k on Ni B is ha
subop imal wo king empe a u e o Ni p o eins om he mo-
philes is a p ice wo h paying when enginee ing ni ogenase in
euka yo es, as solubili y and s abili y o hese a ian s is so much
imp o ed.
One obse a ion o his s udy was ha he specific ac i i y o
as-isola ed NbNi HH p o ein was lowe han ScNi HH . We hink
his was caused by poo [Fe
4
S
4
] clus e a ailabili y –and hence
ine ficien inco po a ion –o poo Ni H [Fe
4
S
4
] clus e s abili y
wi hin he lea cell mi ochond ia. In his con ex , i is no known
how Fe e iliza ion inc eased accumula ion o soluble NbNi UA .
Mo e a ailable Fe could inc ease mi ochond ia [Fe-S] clus e s
biosyn hesis and [Fe
2
S
2
] clus e occupancy in NbNi UA which, in
u n, would p o ide s abili y o he p o ein. A compa ibili y issue
be ween NbNi HH and NbNi UA and NbNi SA is unlikely since
EcNi UA could e ec i ely ac i a e NbNi HH in i o. NbNi HH
mis olding in mi ochond ia is also unlikely as i was e ficien ly
ac i a ed by econs i u ion o i s [Fe
4
S
4
] clus e . I is howe e
likely ha p o ec ion by espi a o y O
2
-consump ion in lea is
lowe han in yeas . NbNi HH exposu e o O
2
du ing lea p o-
cessing is also a possibili y making his a pu ely echnical p o-
blem. While lea es we e kep in liquid ni ogen and lysis and
pu ifica ion we e pe o med inside an anae obic glo e box, i is
di ficul o comple ely ule ou ha some O
2
apped wi hin he
lea was eleased du ing issue dis up ion.
In conclusion, his s udy shows ha gene ic di e si y can be
exploi ed o iden i y, om a e y la ge pool o sequences, he
mos adequa e Ni p o ein componen s o enginee a euka yo ic
ni ogenase. Modula cloning echniques, gene syn hesis wi h
codon op imiza ion, and o he syn he ic biology ools pe mi
building mul i-p o ein pa hways wi h componen s o e y di e se
o igin. In his case he Ni H p o ein om H. he mophilus was
iden ified as soluble in mi ochond ia o bo h S. ce e isiae and N.
ben hamiana accumula ing a much highe le els han he A.
inelandii homologue. This example is ele an no only because
he iden ified a ian pe o med all h ee Ni H-essen ial eac ions,
namely P-clus e ma u a ion, FeMo-co biosyn hesis, and Ni D-
KA educ ion, bu also because Ni HH o med unc ional
in e species in e ac ions wi h Ni B, Ni EN, and Ni DK p o eins,
al oge he ep esen ing he ou p o eins cons i u ing he co e o
diazo ophy.
Me hods
Design, assembly, and cloning o he ni H lib a y. A cu a ed da ase o diazo-
ophs41 was used o collec ni H candida es and design he lib a y. Genes encoding
ni H a ian s we e codon op imized o exp ession in S. ce e isiae and he sequence
encoding pE35S::cox4- wins ep was codon op imized o exp ession in obacco
(Supplemen a y Da a 1). All gene ic pa s we e op imized using he GeneOp imize
ool (The moFishe ) and syn hesized by The moFishe ia he Enginee ing
Ni ogen Symbiosis o A ica (ENSA) p ojec . The ni H genes we e syn hesized
and cloned in o pMA cloning ec o wi h BamHI and Bs EII es ic ion si es
flanking each gene. The pE35S::cox4- wins ep sequence was flanked by HindIII
and BglII es ic ion si es.
pGFPGUSplus (plasmid #64401, Addgene) and he pMA ec o con aining
pE35S::cox4- wins ep we e diges ed wi h HindIII and BglII and used o gene a e
he pa en al ec o pN2SB41, con aining a pE35S::cox4- wins ep-gus- NOS
ansc ip ional uni in which gus was flanked by BamHI and Bs EII es ic ion si es.
The pa en al ec o pN2SB41 and all pMA ec o s con aining ni H a ian s we e
diges ed wi h BamHI and Bs EII and used o gene a e ec o s pN2XJ81-pN2XJ112
(Supplemen a y Table 1).
pGFPGUSplus was used o gene a e ec o pN2XJ165 con aining ansc ip ional
uni s o mi ochond ia- a ge ed accesso y Ni p o eins (A. inelandii Ni U, Ni S,
and Ni M). The su9-ni UA (AAAAGGATCCAATGGCCTCCACTCGTGTCCT
CG, AAAAAAGGTCACCTTAGACTTCCATTTGGGCGTGTGCG) and su9-
ni SA (AAACTAGTATGGCCTCCACTCGTGTCCTCG, AAAAGAGCTCTTAAC
CATAGACAGGAGCAAAGGCTTTACC) genes we e amplified by PCR om he
yeas ec o pN2GLT410. Amplifica ion eac ions added flanking BamHI and Bs EII
( o su9-ni U)o SpeI and SacI ( o su9-ni SA ) si es. The DNA agmen con aining
he su9-ni MA sequence was c ea ed by o e lapping PCR using p ime s in oducing
sequences homologous o hose flanking he XhoI si e o pGFPGUSplus (ATTATG
GAGAAACTCGAGTTAACCATGTGCTAAGTTTTCC, TACAAATCTATCTCT
CTCGAGATGGCCTCCACTCGTG, CTTTCTGAGGCCATGGAAGAGTAGGCG
CGCTTCTGG, CGCGCCTACTCTTCCATGGCCTCAGAAAGATTAGCTGA
TG). pGFPGUSplus was fi s diges ed wi h BglII and Bs EII o inse su9-ni UA ,
hen wi h XbaI and SacI o inse su9-ni SA , and finally diges ed wi h XhoI o inse
su9-ni MA by homologous ecombina ion44.
All DNA diges ions we e pe o med using enzymes om New England Biolabs.
Liga ed p oduc s (T4 ligase, P omega) we e in oduced in o E. coli DH5α
chemically compe en cells and selec ed on LB (Lysogenic b o h) supplemen ed
wi h app op ia e an ibio ics. Plasmid ex ac ion was pe o med using Qiap ep Spin
Minip ep ki (QIAGEN) and co ec cloning was confi med by Sange sequencing
(Mac ogen).
G ow h o S. ce e isiae, mi ochond ia isola ions, and ScNi H pu ifica ions.S.
ce e isiae o galac ose-induced exp ession o ScNi HMm,ScNi HMi,ScNi HH , and
ScNi HA oge he wi h SU9-Ni UA , SU9-Ni SA , and SU9-Ni MA (XJ1Y-XJ4Y,
Supplemen a y Table 3) we e cul u ed in 4-l e men e s unde ae obic condi ions
(0.625 l o ai pe minu e and l o cul u e, 250 pm s i ing) and used o mi o-
chond ia isola ions o Ni H pu ifica ions as p e iously desc ibed11. P epa a ion o
CFE and STAC pu ifica ions we e pe o med a O
2
-le els below 1 ppm in anae-
obic chambe s (Coy sys ems o MB aun). Typically, cells we e esuspended in
lysis bu e (100 mM T is-HCl (pH 8.6), 200 mM NaCl, 10% glyce ol, 2 mM
sodium di hioni e (DTH), 1 mM PMSF, 1 μg/ml leupep in, 5 μg/ml DNAse I) a a
a io o 1:2 (w/ ). To al ex ac s (TE) we e p epa ed by lysis o he cell suspensions
unde anae obic a mosphe e using an EmulsiFlex-C5 homogenize (A es in Inc.)
ope a ing a 20,000 psi. The TE was ans e ed o cen i uge ubes equipped wi h
sealing closu es (Beckman Coul e ) and cen i uged a 50,000 g o 1 h a 4 °C
(A an i J-26 XP). The supe na an was fil e ed using fil e ing cups wi h a po e size
o 0.2 μm, ende ing cell- ee ex ac (CFE) o soluble p o eins ha was loaded a
2.5 ml/min in o a 5 ml S ep-Tac in XP column (IBA Li eSciences) a ached o an
ÄKTA FPLC (GE Hea hca e). The column was washed using 75 ml washing bu e
(100 mM T is-HCl pH 8.0, 200 mM NaCl, 10% glyce ol, 2 mM DTH). S ep-Tac in
XP column-bound p o eins we e elu ed wi h 15 ml washing bu e supplemen ed
wi h 50 mM bio in (IBA Li eSciences). The elu ion ac ion was concen a ed, and
bio in emo ed, by passing he p o ein h ough PD-10 desal ing columns (GE
Heal hca e). Desal ed elua e was u he concen a ed using cen i ugal fil e s
(Amicon, Millipo e) wi h 30 kDa cu o . Finally, he concen a ed p o ein was
snap- ozen in c yo ials (Nalgene) and s o ed in liquid N
2
.
Soil Fe e iliza ion, p epa a ion o anae obic N. ben hamiana lea cell- ee
ex ac s, and pu ifica ion o NbNi HH and NbNi UA .N. ben hamiana plan s
we e g own unde long day condi ions (16 h ligh /8 h da k) wi h suppo ing ligh
om 17:00 o 00:00 o 4 weeks. Fo Fe e iliza ion expe imen s, plan s we e
i iga ed (2l pe week) wi h ap wa e supplemen ed wi h 1 g/l Seques ene G100
ARTICLE COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-020-01536-6
8COMMUNICATIONS BIOLOGY | (2021)4:4 | h ps://doi.o g/10.1038/s42003-020-01536-6 | www.na u e.com/commsbio
(Syngen a). Lea es ha es ed a e ex ended da k pe iod (16 h) we e kep in
da kness om 17:00 (p e ious day) un il sample collec ion (09:00 ollowing
mo ning).
Pu ifica ions o NbNi HH and NbNi UA we e pe o med a O
2
-le els below 1
ppm inside anae obic chambe s (Coy sys ema o MB aun). Typically, 200 g o lea
ma e ial was ha es ed and ozen in liquid N
2
. Lea ma e ial was ans e ed in o
an anae obic chambe in ozen condi ion and dis up ed in equal amoun (w/ ) o
lysis bu e (100 mM T is-HCl pH 8.6, 200 mM NaCl, 10% glyce ol, 2 mM DTH, 1
mM PMSF, 1 μg/ml leupep in, 5 μg/ml DNAseI) using a blende (Os e Classic
4655) ope a ing a maximum powe and main ained a 4 °C using a ci cula ing
wa e ba h. TE was fil e ed h ough cheese clo h o emo e la ge deb is.
P epa a ion CFE by cen i uga ion, S ep-Tac in a fini y ch oma og aphy, p o ein
elu ion, concen a ion, and s o age was iden ical as o yeas -exp essed ScNi H
p o eins. The pu ifica ion p ocedu e o NbNi UA only di e ed in ha no DTH
was p esen in he bu e s.
P o ein me hods, an ibodies, UV– is abso p ion spec um, and elec on
pa amagne ic esonance. P o ein concen a ions we e measu ed using he BCA
p o ein assay (PIERCE) in combina ion wi h iodoace amide o elimina e he
in e e ing e ec o DTH45. Colo ime ic Fe de e mina ion was pe o med as
epo ed46, and he N- e minal amino acid sequences we e de e mined by Edman
deg ada ion (P o eome Fac o y AG).
An ibodies used in his s udy and hei dilu ions o immunoblo ing we e as
ollows: polyclonal an ibodies de ec ing Ni UA (used a 1:2,000 in 5% BSA), Ni SA
(used a 1:1,000 in 5% BSA), Ni HA (used a 1:5,000 in 5% BSA), Ni MA (used a
1:2,000 in 5% BSA) we e aised agains pu ified p epa a ions o he co esponding
A. inelandii p o eins (gene a ed in house). S ep- ag II (“S ep-MAB”, 2-1507-001,
IBA Li esciences, 1:2,000 in 5% BSA), S ep-Tac in conju aged o HRP (“S ep-
HRP”, 2-1502-001, IBA Li esciences, 1:50,000 in TBS- T), GFP (sc-9996, San a
C uz Bio echnology, 1:2,000 in 5% BSA), HSP60 (LK-2, ab59458, Abcam, 1:1,000
in 5% BSA), and Tubulin (3H3087, sc-69971, San a C uz Bio echnology, 1:500 in
5% BSA) specific an ibodies a e comme cially a ailable.
The UV– is abso p ion spec a we e eco ded a e emo al o he DTH om
he p o ein samples using PD-10 desal ing columns (GE Heal hca e) equilib a ed
wi h he co esponding p o ein bu e wih ou DTH. DTH- ee p o ein samples
we e hen dilu ed in he same bu e s and ans e ed o a Q6 spec oscopy cu e es
wi h sealing closu es. Abso p ion (280 nm o 800 nm) was eco ded using a UV-
2600 spec opho ome e (Shimadzu).
EPR measu emen s we e pe o med in a B uke E500 spec ome e equipped
wi h a esona o ope a ing in he TE
102
mode a 9.47 GHz. Tempe a u e was se
and s abilized o 10 K by an Ox o d empe a u e con olle egula ing a gas-flow
c yos a e ige a ed wi h helium. Fo measu emen s, a mic owa e powe o 2.5
mW and a magne ic field modula ion ampli ude o 1 mT was used. Expe imen al
condi ions we e ca e ully moni o ed o a oid o e -modula ion o sa u a ion e ec s.
Simula ions o he EPR spec a we e pe o med using he Ma lab oolbox
Easyspin47.
In i o Ni H ac i i y. Ni H ac i i y was de e mined as desc ibed by Shah e al.
wi h sligh modifica ions48. Reac ions we e p epa ed inside anae obic chambe s.
Pu ified Ni H p o eins we e analyzed by ARA a e addi ion o Ni DKA and ATP-
egene a ing mix u e (1.23 mM ATP, 18 mM phosphoc ea ine, 2.2 mM MgCl
2
,
3 mM DTH and 46 μg/ml o c ea ine phosphokinase, 22 mM T is-HCl pH 7.5) in a
final olume o 600 μl inside 9 ml se um ials unde A a mosphe e con aining
500 μl o ace ylene (1 a m). The a io o Ni H o Ni DK in he assays was 40:1
unless o he wise indica ed. The ARA we e pe o med a 30 °C in a shaking wa e
ba h o 15 min. Reac ions we e s opped by adding 100 μl o 8 M NaOH. Posi i e
con ol eac ions o ace ylene educ ion we e ca ied ou wi h Ni HA . E hylene
o med was measu ed in 50 μl gas phase samples using a Po apak N 80/100 column
in a gas ch oma og aph (Shimadzu).
Reduc ion o N
2
o NH
3
was de e mined in eac ion mix u es p epa ed as o
he ARA bu con aining 100 mM 3-(N-mo pholino)p opanesul onic acid (MOPS),
pH 7.8, as bu e . Mix u es we e p epa ed in olumes o 750 μl, om which 100 μl
was emo ed a assay s a o se e as backg ound (
o
) o NH
3
measu emen s.
A e exchanging ial a mosphe e o N
2
, mix u es we e incuba ed a 30 °C o
30 min, and eac ions we e s opped by addi ion o 100 μl 5 M EDTA. Twen y-fi e
μl o he blank (
o
) and he eac ion (
30
) we e added in duplica es o 200 μlo-
ph haldialdehyde eagen solu ion (The moFishe Scien ific) in 96-well mic opla e
o fluo escence-based assays (Nunc). Fluo escence (Ex 390 nm, Em 472 nm) was
measu ed using a Va ioskan LUX pla e eade (The moFishe Scien ific). NH
3
p oduc ion was de e mined om he inc ease in fluo escence
(
30
-
o
) agains s anda ds p epa ed wi h NH
4
Cl and eco ded in he same pla e.
In i o P-clus e ma u a ion. P-clus e ma u a ion assays we e pe o med inside
anae obic chambe s. The in i o assay combined isola ed Ni H o be es ed (50 μg)
wi h A. inelandii DJ77 (Δni H) CFE (4.34 mg o al p o ein) and an excess o pu e
FeMo-co (0.85 μM) in 500 μl ATP- egene a ing mix u e as desc ibed abo e.
Reac ions we e incuba ed a 30 °C o 30 min. Fo y μl o 1 mM (NH
4
)
2
MoS
4
( e a hiomolybda e) we e hen added and mix u es we e incuba ed o 10 min a
oom empe a u e o p e en u he FeMo-co inco po a ion in o Ni DKA du ing
he ARA.
Apo-Ni DKA ac i a ion a e P-clus e ma u a ion and FeMo-co inse ion was
analyzed by ARA a e addi ion o an excess o he same Ni H species (100 μg) and
ATP- egene a ing mix u e in a final eac ion olume o 1 ml. ARA was ca ied ou
in 9 ml se um ials con aining A and 500 μl o ace ylene (1 a m) in he headspace
o 15 min a 30 °C. Posi i e con ol eac ions o in i o P-clus e ma u a ion and
ARA con ained pu ified Ni HA . E hylene o med was measu ed in 50 μl gas phase
samples using a Po apak N 80/100 column in a gas ch oma og aph (Shimadzu).
In i o FeMo-co syn hesis and apo-Ni DKA econs i u ion. Ni B-co-dependen
FeMo-co syn hesis assays we e pe o med inside anae obic chambe s as desc ibed
by Cu a i e al., wi h sligh modifica ions34. One hund ed μl eac ions con ained
3.0 μM Ni H, GST-Ni X-Ni B-co (20.4 μM Fe), 1.5 μM apo-Ni ENA , 0.6 μM apo-
Ni DKA , 17.5 μMNa
2
MoO
4
, 175 μMR-homoci a e, 1 mg/ml BSA, and ATP-
egene a ing mix u e (1.23 mM ATP, 18 mM phosphoc ea ine disodium sal , 2.2
mM MgCl
2
, 3 mM DTH, 46 μg/ml c ea ine phosphokinase, final concen a ions in
22 mM T is-HCl (pH 7.5) bu e a 30 °C o 60 min.
Ni B-dependen FeMo-co syn hesis assays we e pe o med as he abo e
desc ibed Ni B-co-dependen assay eplacing GST-Ni X-Ni B-co by 10.0 μM Ni B
monome , 125 μM FeSO
4
, 125 μMNa
2
S, and 125 μM SAM.
Following in i o syn hesis o FeMo-co, 17.5 μM (NH
4
)
2
MoS
4
was added o
p e en u he FeMo-co inco po a ion in o apo-Ni DKA , and incuba ed o 10
min a 25 °C. Ac i a ion o apo-Ni DKA was analyzed by addi ion o 500 μl ATP-
egene a ing mix u e and ScNi HH (2.0 μMfinal concen a ion) in 9 ml ials
con aining A and 500 μl ace ylene. The ARA we e pe o med a 30 °C o 20 min.
Posi i e con ol eac ions o ARA con ained Ni DKA and Ni HA . E hylene
o med was measu ed in 50 μl gas phase samples using a Po apak N 80/100 column
in a gas ch oma og aph (Shimadzu).
In i o [Fe-S] clus e econs i u ion and Ni H ac i i y. In i o [Fe-S] clus e
econs i u ions o Ni H and Ni UA pu ified om E. coli (EcNi UA )10 we e pe -
o med in anae obic chambe s as desc ibed by Zheng and Dean49 wi h sligh
modifica ions. Ni H o Ni U (20 μM) was added o 22 mM T is-HCl (pH 7.5)
bu e supplemen ed wi h 8 mM 1,4-di hio h ei ol (DTT) in a final olume o 100
μl and incuba ed a 37 °C o 30 min. Then, eac ions we e supplemen ed wi h 1
mM L-cys eine, 1 mM DTT, 400 μM (NH
4
)
2
Fe(SO
4
)
2
, and 225 nM Ni SA pu ified
om E. coli (EcNi SA )10, and incuba ed a 37 °C o e nigh . Finally, he p o eins
we e dilu ed 1000- old in 22 mM T is-HCl (pH 7.5) bu e , and hen concen a ed
using cen i ugal fil e s (Amicon, Millipo e) wi h 30 kDa cu o o emo e excess
eagen s.
Fo “di ec econs i u ion”ac i i y assays, he ac i i y o [Fe
4
S
4
] clus e
econs i u ed Ni H p o ein was de e mined using ARA. Fo “Ni U-media ed
econs i u ion”, as-isola ed Ni H p o ein was mixed wi h [Fe-S] clus e
econs i u ed EcNi UA , and hen immedia ely used o ARA.
S a is ics and ep oducibili y. Dis inc samples we e used o in i o ac i i y
measu emen s and sample sizes a e indica ed by n, whe e each dis inc sample was
measu ed a leas wo imes. Mean o measu ed ac i i ies a e shown. The da a
p esen ed in he figu e g aphs a e lis ed in Supplemen a y Da a 2.
Da a a ailabili y
The au ho s decla e ha he da a suppo ing he findings o his s udy a e a ailable
wi hin he a icle, i s supplemen a y in o ma ion and da a, and upon eques .
Recei ed: 15 June 2020; Accep ed: 25 No embe 2020;
Re e ences
1. E isman, J. W. e al. Ni ogen: Too much o a i al esou ce. (Science B ie .
WWF Ne he lands, Zeis , The Ne he lands, 2015).
2. Bo laug, N. E. Feeding a wo ld o 10 billion people: he mi acle ahead. In Vi o
Cell De . Biol. 38, 221–228 (2002).
3. Sibha u, K. T. & Qaim, M. Ru al ood secu i y, subsis ence ag icul u e, and
seasonali y. PLoS One 12, e0186406 (2017).
4. Bloch, S. E., Ryu, M. H., Ozaydin, B. & B oglie, R. Ha nessing a mosphe ic
ni ogen o ce eal c op p oduc ion. Cu . Opin. Bio echnol. 62, 181–188
(2020).
5. Bu én, S. & Rubio, L. M. S a e o he a in euka yo ic ni ogenase enginee ing.
FEMS Mic obiol. Le . 365, nx274– nx274 (2018).
6. Bulen, W. A. & LeCom e, J. R. The ni ogenase sys em om Azo obac e : wo-
enzyme equi emen o N2 educ ion, ATP-dependen H2 e olu ion, and
ATP hyd olysis. P oc. Na l Acad. Sci. USA 56, 979–986 (1966).
COMMUNICATIONS BIOLOGY | h ps://doi.o g/10.1038/s42003-020-01536-6 ARTICLE
COMMUNICATIONS BIOLOGY | (2021)4:4 | h ps://doi.o g/10.1038/s42003-020-01536-6 | www.na u e.com/commsbio 9