PLANT SCIENCE
Processing of NODULE INCEPTION controls the
transition to nitrogen fixation in root nodules
Jian Feng^1 , Tak Lee1,2, Katharina Schiessl^1 , Giles E. D. Oldroyd1,2*
Legume nodules create an environment for intracellular bacterial symbionts to fix atmospheric nitrogen.
The master regulator NODULE INCEPTION (NIN) controls many aspects of nodule initiation, and we
demonstrate that it also regulates the transition to nitrogen fixation via proteolytic processing by a signal
peptidase complex. Processing of NIN results in a carboxyl-terminal NIN fragment containing the DNA binding
motifs, which activates a suite of genes associated with symbiosome development and nitrogen fixation.
Similar NIN processing is observed inMedicago truncatulaandLotus japonicus, implying a conserved
mechanism of cell state transition. These findings explain how legume nodules transition to a nitrogen-fixing
state and a mechanism by which a single transcription factor can regulate many different developmental
processes necessary in the activation and regulation of nitrogen fixation.
S
ome angiosperms, including legumes,
associate with nitrogen-fixing bacteria
within nodules, which are modified
lateral roots ( 1 , 2 ). Nodules create the
low-oxygen environment for bacterial
nitrogenase to convert dinitrogen to ammo-
nium. Nodule initiation and promotion of in-
tracellular bacterial infection is facilitated by
NODULE INCEPTION(NIN)( 3 , 4 ), in part
through transcriptional activation ofNUCLEAR
FACTOR-Y SUBUNIT A1(NF-YA1),NODULE
PECTATE LYASE(NPL),CYTOKININ RESPONSE
1 (CRE1), andLATERAL ORGAN BOUNDARIES
DOMAIN 16(LBD16)( 5 , 6 ).NINalso controls
nodule number through mobile CLE peptides
that promote root-shoot-root signaling ( 7 , 8 ).
NINisoneofonlyafewgenesthatarecon-
sistently lost with losses of nodulation over
evolutionary time ( 9 , 10 ). Here we demonstrate
that NIN also controls maturation of the nod-
ule to the nitrogen-fixation state through pro-
teolytic processing by a nodule-specific signal
peptidase complex (SPC).
A series of functional NIN–green fluorescent
protein (GFP) fusions spanning the full length
of the protein (table S1) all showed nuclear lo-
calization (Fig. 1E), consistent with multiple
predicted nuclear localization signals (Fig. 1A).
An array of antibodies produced against NIN
peptides (Fig. 1A) revealed a predicted full-
length NIN and a second larger band, possibly
an oligomeric form, whenNINin transientlyexpressed inNicotiana benthamiana(Fig. 1B
and fig. S1). The same antibodies showed no
NIN-specific bands in total root and nodule
preparations (fig. S2) but did show clear NIN-
specific bands from nuclei preparations (Fig.
1C and figs. S2 and S3). However, the predicted
full-length product was only faintly observed
witha-RWP in wild type (Fig. 1C and fig. S3A),
as well as witha-internal region (IR) in hyper-
nodulatingsunnmutant ( 11 ) (fig. S4). In place
of the full-length product, we observed two
smaller products: a c40kDa product detected
bya-N terminus (NT) and a c55kDa product
witha-IR,a-RWP, anda-C terminus (CT) (Fig.
1C).a-RWP also revealed additional products
not detected by the other antibodies (Fig. 1C).
All of these bands were absent in thenin-4
(Fig. 1C) null mutant (fig. S5). The c55kDa pro-
duct must possess the DNA binding domain
(RWP-RK) and the protein-protein interaction
domain (PB1), because antibodies that detect
this product span these domains (Fig. 1A). To
validate these observations, we generated func-
tional (fig. S6) 3×FLAG NIN-fusions:a-FLAG
antibodies detected the c55kDa C-terminal and
predicted full-length products (Fig. 1D) but not
the c40kDa product.
The processing of NIN is not the result of
alternative splicing ofNINmRNA and is in-SCIENCEscience.org 29 OCTOBER 2021•VOL 374 ISSUE 6567 629
(^1) Sainsbury Laboratory, University of Cambridge, 47 Bateman
Street, Cambridge CB2 1LR, UK.^2 Crop Science Centre,
University of Cambridge, 93 Lawrence Weaver Road,
Cambridge CB3 0LE, UK.
*Corresponding author. Email: [email protected]
Fig. 1. NIN is processed during nodule formation.(A) A schematic of NIN
showing the RWP-RK and PB1 domains and predicted nuclear localization
signals (hatching). The position of peptides used for antibody generation and
mutant alleles are indicated. (B) Immunoblot analysis of NIN-3×FLAG proteins
expressed inN. benthamianaleaves usinga-IR,a-CT, anda-FLAG antibodies.
(C) Immunoblot analysis witha-NIN in the nuclear extracts of wild-type
(R108) andnin-4roots at indicated time points afterS. melilotiinoculation.
dpi, days post inoculation. (D) Different NIN-3×FLAG fusions expressed in
M. truncatula(R108) roots and detected witha-FLAG antibodies. EV, empty vector;
NIN, full-lengthNINandNIN-3×FLAGfusions (numbers refer to different positions
of3×FLAG)(fig.S6).(E) Nuclear localization of NIN-GFP fusions expressed in
nodules. Nuclei are labeled with 3×mCherry fused to a nuclear localization signal.
Numbers refer to different positions of the GFP tag, as shown in table S1. Red
asterisks in (B) to (D) indicate NIN-specific products. For (B) to (D), equal loading
was verified witha-histone H3. O, oligomeric form of NIN; FL, predicted full-length
NIN; C, c55kDa C-terminal product; N, c40kDa N-terminal product.
RESEARCH | REPORTS