GIGANTEA orthologs, E2 members, redundantly determine photoperiodic flowering and yield in soybean

Abstract

Soybean (Glycine max L.) is a typical photoperiod‐sensitive crop, such that photoperiod determines its flowering time, maturity, grain yield, and phenological adaptability. During evolution, the soybean genome has undergone two duplication events, resulting in about 75% of all genes being represented by multiple copies, which is associated with rampant gene redundancy. Among duplicated genes, the important soybean maturity gene E2 has two homologs, E2‐Like a (E2La) and E2‐Like b (E2Lb), which encode orthologs of Arabidopsis GIGANTEA (GI). Although E2 was cloned a decade ago, we still know very little about its contribution to flowering time and even less about the function of its homologs. Here, we generated single and double mutants in E2, E2La, and E2Lb by genome editing and determined that E2 plays major roles in the regulation of flowering time and yield, with the two E2 homologs depending on E2 function. At high latitude regions, e2 single mutants showed earlier flowering and high grain yield. Remarkably, in terms of genetic relationship, genes from the legume‐specific transcription factor family E1 were epistatic to E2. We established that E2 and E2‐like proteins form homodimers or heterodimers to regulate the transcription of E1 family genes, with the homodimer exerting a greater function than the heterodimers. In addition, we established that the H3 haplotype of E2 is the ancestral allele and is mainly restricted to low latitude regions, from which the loss‐of‐function alleles of the H1 and H2 haplotypes were derived. Furthermore, we demonstrated that the function of the H3 allele is stronger than that of the H1 haplotype in the regulation of flowering time, which has not been shown before. Our findings provide excellent allelic combinations for classical breeding and targeted gene disruption or editing.

Full text

J
IPB Journal of Integrative
Plant Biology Research Article
https://doi.org/10.1111/jipb.13398
GIGANTEA orthologs, E2 members, redundantly
determine photoperiodic flowering and yield
in soybean
Lingshuang Wang
1
†
, Haiyang Li
1,2
†
, Milan He
1
†
, Lidong Dong , Zerong Huang
1
, Liyu Chen
1
,
Haiyang Nan
1
, Fanjiang Kong
1
, Baohui Liu
1
* and Xiaohui Zhao
1
*
1. Guangdong Key Laboratory of Plant Adaptation and Molecular Design, Guangzhou Key Laboratory of Crop Gene Editing, Innovative
Center of Molecular Genetics and Evolution, School of Life Sciences, Guangzhou University, Guangzhou 510006, China
2. National Key Laboratory of Crop Genetics and Germplasm Enhancement, National Center for Soybean Improvement, Jiangsu
Collaborative Innovation Center for Modern Crop Production, Nanjing Agricultural University, Nanjing 210095, China
†
These authors contributed equally to this work.
*Correspondences: Baohui Liu (liubh@gzhu.edu.cn); Xiaohui Zhao (zhaoxh@gzhu.edu.cn, Dr. Zhao is responsible for the distribution of
the materials associated with this article)
Lingshuang Wang Xiaohui Zhao
ABSTRACT
Soybean (Glycine max L.) is a typical photoperiod‐
sensitive crop, such that photoperiod determines
its flowering time, maturity, grain yield, and
phenological adaptability. During evolution, the
soybean genome has undergone two duplication
events, resulting in about 75% of all genes being
represented by multiple copies, which is asso-
ciated with rampant gene redundancy. Among
duplicated genes, the important soybean maturity
gene E2 has two homologs, E2‐Like a (E2La)and
E2‐Like b (E2Lb), which encode orthologs of Ara-
bidopsis GIGANTEA (GI). Although E2 was cloned a
decade ago, we still know very little about its con-
tribution to flowering time and even less about the
function of its homologs. Here, we generated single
and double mutants in E2,E2La,andE2Lb by
genome editing and determined that E2 plays major
roles in the regulation of flowering time and yield,
with the two E2 homologs depending on E2 func-
tion. At high latitude regions, e2 single mutants
showed earlier flowering and high grain yield. Re-
markably, in terms of genetic relationship, genes
from the legume‐specific transcription factor family
E1 were epistatic to E2. We established that E2 and
E2‐like proteins form homodimers or heterodimers
to regulate the transcription of E1 family genes,
with the homodimer exerting a greater function
than the heterodimers. In addition, we established
that the H3 haplotype of E2 is the ancestral allele
and is mainly restricted to low latitude regions, from
which the loss‐of‐function alleles of the H1 and H2
haplotypes were derived. Furthermore, we demon-
strated that the function of the H3 allele is stronger
than that of the H1 haplotype in the regulation of
flowering time, which has not been shown before.
Our findings provide excellent allelic combinations
for classical breeding and targeted gene disruption
or editing.
Keywords: E2,E2‐Like,flowering time, GIGANTEA, natural
variation, redundancy, yield
Wang, L., Li, H., He, M., Dong, L., Huang, Z., Chen, L., Nan, H.,
Kong, F., Liu, B., and Zhao, X. (2023). GIGANTEA orthologs, E2
members, redundantly determine photoperiodic flowering and
yield in soybean. J. Integr. Plant Biol. 65: 188–202.
INTRODUCTION
Photoperiod‐mediated flowering is a critical stage of
plant growth and development. Compared to other crops,
soybean (Glycine max. L) is a typical short‐day (SD) crop that
is extremely sensitive to photoperiod and flowers early when
exposed to SD conditions. Moreover, a specific soybean
germplasm is generally only suitable for planting in areas with
© 2022 Institute of Botany, Chinese Academy of Sciences.
17447909, 2023, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jipb.13398 by Shanghai Jiao Tong University, Wiley Online Library on [06/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

a small latitudinal span of no more than 1.5° (Garner
and Allard, 1920,1923). Therefore, photoperiod largely
determines the length of maturity and the adaptability of
soybean. If the vegetative growth period is too short, the yield
will be greatly reduced due to an insufficient accumulation of
nutrients during the vegetative stage. By contrast, especially
at high‐latitude areas, an overly long growth period will result
in a failure of soybean to mature before the end of the frost‐
free season. Therefore, it is important that crops bloom and
mature at the right time.
Several major flowering and maturity loci have been
identified and functionally characterized in soybean,
including maturity genes and long juvenile genes, such as
E1‐E11,J,Time of Flowering 5 (Tof5), Tof11,Tof12, and
Tof16 (Bernard, 1971;Buzzell, 1971;Buzzell and Voldeng,
1980;McBlain and Bernard, 1987;Ray et al., 1995;Bonato
and Vello, 1999;Cober and Voldeng, 2001;Cober et al.,
2010;Kong et al., 2014;Samanfar et al., 2016;Wang et al.,
2019;Lu et al., 2020;Dong et al., 2021;Dong et al., 2022). Of
these, the soybean E2 locus was shown to be an ortholog of
Arabidopsis thaliana GIGANTEA (GI) using a map‐based
cloning strategy; further research demonstrated that E2 is
involved in soybean flowering and maturity (Bernard, 1971),
thus sharing the same function as its Arabidopsis ortholog.
However, since the soybean genome has undergone two
duplication events, about 75% of all soybean genes are
present in multiple copies, resulting in high genetic
redundancy (Schmutz et al., 2010). In fact, in the soybean
genome there are two homologs in addition to E2, one of
which is located on chromosome 20 and encodes proteins
with 94%–97% sequence identity with E2, while the other
homolog maps to chromosome 16 but is much more distant
from E2 based on phylogenic analysis (Watanabe et al.,
2011). E2 and the two E2‐like proteins localize to the nucleus
and physically interact with each other based on yeast two‐
hybrid assays in a light‐independent manner (Li et al., 2013).
An allele e2 carrying a single nucleotide polymorphism (SNP)
within exon 10 of E2 introduced a premature stop codon and
showed an earlier flowering phenotype, in agreement with
the elevated transcript levels of GmFT2a, one of the soybean
FLOWERING LOCUS T (FT) genes, thus leading to the early
flowering phenotype (Watanabe et al., 2011). The effects of
the e2 mutant allele on flowering time were similar in regions
of high and middle latitudes in Japan (Watanabe et al., 2011),
indicating that E2 may control photoperiod insensitivity in
soybean. Besides photoperiod‐mediated flowering time,
genome‐wide association studies showed a correlation be-
tween the genotype at the E2 locus and plant height and the
ratio of linolenic acid to linoleic acid, which is associated
with seed yield and quality (Fang et al., 2017). However, the
function of the two closely related E2‐Like genes remains
unknown.
In soybean, members from the legume‐specifictran-
scription factor E1 gene family function as the most im-
portant maturity genes in the control of flowering time and
the length of the maturity period (Xia et al., 2012). E1
encodes a B3 superfamily transcription factor whose tran-
scription is regulated by the phytochrome A (phyA) genes E3
and E4. E1 family genes control flowering by repressing the
expression of two key FT genes, GmFT2a and GmFT5a (Xia
et al., 2012;Xu et al., 2015). The soybean circadian evening
complex has been reported to act upstream of E1 and
suppress E1 transcription through the binding of LUX AR-
RHYTHMO (LUX) to its cognate binding sites (Lu et al.,
2017;Bu et al., 2021). Another set of components influ-
encing the regulation of flowering time are orthologs to
Arabidopsis LATE ELONGATED HYPOCOTYL (LHY), which
directly bind to the AATATC motif in the E1 promoter and
repress its transcription (Lu et al., 2020;Dong et al., 2021).
Therefore, the soybean‐specific factor E1 is at the core of
the soybean photoperiod regulatory network and illustrates
a distinct regulatory mechanism from that represented by
the classic Arabidopsis GI‐CONSTANS (CO)‐FT photo-
periodic flowering pathway.
Despite the decade that has elapsed since the map‐
based cloning of E2,andalthoughE2 has been selected by
nature and used by farmers and breeders, we still know very
little about this gene and even less about the function of its
homologs. Here, we used clustered regularly interspaced
short palindromic repeats (CRISPR)/CRISPR‐associated
nuclease 9 (Cas9)‐mediated gene editing to generate
single and double mutants in E2 family members to inves-
tigate their function in flowering time and grain yield. We
also aimed to clarify their genetic and regulatory relationship
with the core photoperiodic flowering factor E1 and explore
the relationship between E2 homologs. In addition, we in-
vestigated the natural variation, origin, and geographical
distribution of E2 and E2‐Like genes. This study reveals the
asymmetric redundancy among E2 family members in the
regulation of flowering time and yield and provides a
valuable reference point to obtain excellent genotype
combinations for different ecological adaptations to breed
high‐yielding soybean.
RESULTS
Generation of the e2 and e2‐like mutants
To explore the function of the three E2 soybean homologs,
we generated knockout mutants for E2,E2La,andE2Lb by
CRISPR/Cas9‐mediated gene editing in the Williams 82
(W82) background. To this end, we selected two regions in
the coding sequences of the three genes as target sites for
the cloning of single guide RNAs (sgRNAs) in a Cas9 ex-
pression vector (Figure 1A). We identified heterozygous
edited transgenic plants for each of the three genes and
harvested their progeny for screening transgene‐free plants
in the next generation. We thus isolated two homozygous e2
single mutants, designated e2‐1(harboring a 1‐bp deletion
in the target 1 site) and e2‐9(with a 13‐bp deletion and
multiple mutations around the target 1 site); one homo-
zygous e2la single mutant, named e2la‐15 (with a 17‐bp
GI orthologs regulate soybean flowering time and yieldJournal of Integrative Plant Biology
www.jipb.net January 2023
|
Volume 65
|
Issue 1
|
188–202 189
17447909, 2023, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jipb.13398 by Shanghai Jiao Tong University, Wiley Online Library on [06/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

deletion at the target 1 site); and two homozygous e2lb
single mutants, named e2lb‐1(with a 1‐bp insertion at target
2) and e2lb‐10 (with a 45‐bp deletion at target 2, including a
9‐bp deletion in the previous intron and a 36‐bp deletion
in the exon) (Figure 1B). For all five single mutants, the
mutations caused frameshifts, leading to the introduction of
premature stop codons (Figure 1C). In addition, we checked
the transcription levels of the E2 members in their respective
mutants, all mutants with normal transcripts but at a very
low level (Figure S1).
Figure 1. Generation of the e2 and e2‐like mutants by CRISPR/Cas9‐mediated gene editing in soybean
(A) Schematic diagrams of the E2 family genes E2,E2La, and E2Lb and the selected target sites for gene editing via CRISPR/Cas9 gene editing, as
indicated by the red arrows. E2 and E2La share the target 1. (B) Sequence alignment of the single guide RNA (sgRNA) target region in the indicated
homozygous mutant lines. The underlined nucleotides represent the sgRNA target, and the red boxes indicate the protospacer‐adjacent motif (PAM) NGG.
Red letters denote mutated nucleotides, and the gray letters indicate nucleotides in introns. Red dashes indicate deleted nucleotides. Δindicates the
mutant line with multiple mismatches compared to the wild type. (C) Predicted changes in the amino acid sequence of E2 from the wild‐type cultivar
Williams 82 (W82) and the E2 mutants. Underlined nucleotides, sgRNA target sites; red boxes, PAM; blue letters show the amino acids that are different
from W82; asterisk, termination of translation.
GI orthologs regulate soybean flowering time and yield Journal of Integrative Plant Biology
190 January 2023
|
Volume 65
|
Issue 1
|
188–202 www.jipb.net
17447909, 2023, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jipb.13398 by Shanghai Jiao Tong University, Wiley Online Library on [06/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

E2‐Like genes genetically depend on E2 for
flowering time
To examine the consequences of the loss of E2,E2La, and
E2Lb function in the context of flowering time, we grew the
wild‐type W82 and the five single mutants in an artificial
growth incubator under long‐day (LD) and SD conditions. We
determined that the two e2 single mutants flower 8–10 d
earlier than the wild type, while the e2la and e2lb single
mutants showed no obvious effect on flowering time under
LD conditions (Figure 2A, C). We also observed no significant
difference in the flowering time of W82 or any of the mutants
under SD conditions (Figure S2A).
To test the phenotype of these mutants under natural
conditions, we grew W82 and all e2‐1,e2la‐15, and e2lb‐10
single mutants (hereafter referred to as e2,e2la, and e2lb,
respectively) in the field at a natural environment at three sites
in China, Hefei (31°51′N, 117°15′E), Shijiazhuang (37°27′N,
113°30′E), and Harbin (45°55′N, 126°96′E), and scored
flowering phenotype and grain yield. In the North of China in
Shijiazhuang, the e2 single mutant showed an early flowering
phenotype, along with a distinctly increased yield perform-
ance relative to W82 (Figure 2E). By contrast, the e2la and
e2lb single mutants flowered at the same time as W82, which
was similar to their observed phenotype under artificial
growth conditions (Figure 2E). In the Northeast of China in
Harbin, W82 plants failed to reach maturity before frost,
prompting us to move the W82 and e2 mutant plants into
pots outdoors before transfer to the greenhouse at the later
stage for seed harvesting. The e2 mutant plants still flowered
earlier and exhibited a greater total grain yield than W82,
suggesting an important role in enhancing yield for the e2
mutant allele in soybean (Figure S2C). However, in the middle
of China in Hefei, the e2,e2la, and e2lb single mutants
flowered at the same time as the wild‐type W82, confirming
the dispensable role of E2 family genes in the regulation of
flowering time under SD conditions (Figure 2F). In conclusion,
the three E2 family members tested here displayed locus‐
specific effects on photoperiod‐regulated flowering, with E2
playing a more dominant role under LD conditions than its
homologs.
To further explore how E2 homologs control flowering
time, we crossed the e2,e2la, and e2lb single mutants to
generate the e2 e2la,e2 e2lb, and e2la e2lb double mutants.
We then assessed the phenotype of all materials under arti-
ficial LD and SD conditions. Unlike the e2la or e2lb single
mutants, we observed a significantly earlier flowering phe-
notype in the e2 e2la,e2 e2lb, and e2la e2lb double mutants
in artificial LD conditions relative to W82 (Figure 2B, D).
Notably, the loss of E2 function in the e2 e2la and e2 e2lb
double mutants was accompanied by a more pronounced
early flowering phenotype than that seen in the e2 single
mutant or in the e2la e2lb double mutant. In addition, the e2
e2la and e2 e2lb double mutants flowered almost at the same
time, indicating a greater contribution to the regulation of
flowering time by E2 than by E2la and E2lb combined, with
E2La and E2Lb participating in flowering time mainly in an
E2‐dependent manner (Figure 2B). All single mutants and
double mutants flowered at the same time as the wild‐type
W82 under SD conditions (Figure S2B). We conclude that E2
family members control flowering time in soybean via asym-
metric redundancy.
E1 family genes are genetically epistatic to E2
The E1 gene family encodes legume‐specificflowering re-
pressors that play a central role in photoperiodic flowering
and is represented by three members (E1,E1La, and E1Lb)in
soybean. To examine the relationship between E2 and E1
members, we crossed the e2 single mutant with the e1 e1la
e1lb triple mutant previously developed in the W82 back-
ground (Lin et al., 2022), from which we obtained four higher‐
order mutant combinations: e1la e1lb,e1 e2,e1la e1lb e2,
and e1 e1la e1lb e2 lines (Figure 3A). We then grew all gen-
otypes in the field in Shijiazhuang under natural LD conditions
to investigate their phenotypes, which revealed that all mu-
tants promoted flowering compared to W82, with the higher‐
order mutants carrying the e2 mutation flowering earlier than
the e2 single mutant, reflecting the dependence of E2 on E1
family members (Figure 3B). Notably, the e1 e1la e1lb triple
mutant and the e1 e1la e1lb e2 quadruple mutant exhibited
an extremely early flowering phenotype at 23 days after
emergence (DAE) under LD conditions, confirming that E2 is
genetically fully dependent on E1 family members to regulate
flowering in soybean (Figure 3B). This dependence was not
only reflected in the flowering phenotype, but also in multiple
yield traits such as plant height, branch number, pod number,
and grain weight per plant, as the performances of the e1
e1la e1lb e2 quadruple mutant were all consistent with those
of the e1 e1la e1lb triple mutant (Figure 3C–I). The flowering
phenotype observed under artificial LD conditions was also
completely consistent with that seen in the natural environ-
ment (Figure S3).
We also determined the phenotypes of all genotypes at
the mid‐latitudinal location of Hefei, where the flowering time
was identical across all genotypes, although the E2 gene
appeared to be equally dependent on the E1 family members
in terms of yield traits (Figure S4). Together, we propose
that E2 depends on the family of legume‐specific core tran-
scription factor E1 in the regulation of both flowering and
yield in soybean.
E2 family members regulate the transcription of E1
genes asymmetrically and redundantly
To determine the transcriptional regulation of E2 family
members on E1s, we measured the relative transcript levels
of E1,E1La,andE1Lb in W82 and in the mutants with a
flowering time phenotype under LD conditions. All three E1
members showed the same expression patterns in the e2
single mutant and the e2 e2la,e2 e2lb,ande2la e2lb double
mutants, which generally reached lower expression levels
relativetoW82overadiurnaltimecourse(
Figure 4A). To
validate the regulatory relationship between the three E2
proteins and E1 transcription, we performed a transient
GI orthologs regulate soybean flowering time and yieldJournal of Integrative Plant Biology
www.jipb.net January 2023
|
Volume 65
|
Issue 1
|
188–202 191
17447909, 2023, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jipb.13398 by Shanghai Jiao Tong University, Wiley Online Library on [06/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

Figure 2. Flowering phenotype of e2 and e2‐like single and double mutants
(A) Flowering time phenotype of the indicated e2 and e2‐like single mutants under long‐day conditions (16 h light/8 h dark), reported as days after
emergence (DAE). (B‐D) Representative phenotype of single and double mutants in E2 family members. The plants were grown in an incubator in long‐day
conditions (16 h light/8 h dark). The images below panels (C) and (D) are enlarged views of the red boxes in the upper panels, showing the axils of trifoliate
leaves (C, D). (E, F) Field phenotypes of e2 and e2‐like single mutants grown in Shijiazhuang (37°27′N) and Hefei (31°51′N) under natural conditions,
showing flowering time (left) and grain weight per plant (total seed weight; right). Flowering time was scored at the R1 stage (days from emergence to the
appearance of the first open flower on 50% of the plants). W82, wild‐type cultivar Williams 82. At least eight plants were scored for each phenotype; data
are shown as means ±SD, with all individual plants shown as a dot. P<0.05, as determined by multiple comparison testing by one‐way analysis of
variance. Different lowercase letters indicate significant differences between the genotypes.
GI orthologs regulate soybean flowering time and yield Journal of Integrative Plant Biology
192 January 2023
|
Volume 65
|
Issue 1
|
188–202 www.jipb.net
17447909, 2023, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jipb.13398 by Shanghai Jiao Tong University, Wiley Online Library on [06/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

infiltration assay in Nicotiana benthamiana leaves
(Figure 4B). Accordingly, we placed the transcription of the
firefly luciferase (LUC) reporter gene under the control of the
E1 promoter, using E2 and E2‐Like effector constructs.
Although the co‐infiltration of the E1pro:LUC reporter with
individual E2 effector constructs induced E1 transcription,
we observed a further promotion of transcription when
effector constructs from two E2 family members were
co‐infiltrated, especially when E2 was involved (Figure 4C).
These results suggested that E2 proteins might interact
with the E1 promoter to varying degrees to activate its
transcription, although whether this interaction is direct or
indirect is unknown.
In view of the redundancy between E2 homologs, we
wondered whether each E2 family member might compen-
sate for a loss of function in the other members by raising
their own transcript levels. To explore this possibility, we
measured the relative transcript levels of the remaining genes
in each of the three e2 or e2‐like single mutants (Figure 4D)
and three e2 double mutants (Figure 4E). We observed that
each E2 homolog is expressed at comparable levels in W82
and the single and double mutants (Figure 4D, E). We thus
assessed the protein–protein interaction potential between the
three E2 proteins via the yeast two‐hybrid and bimolecular flu-
orescence complementation (BiFC) assay. We established that
all three E2 and E2‐Like proteins interact with each other
Figure 3. The flowering time and grain yield phenotypes regulated by E2 are dependent on E1 family genes
(A) Phenotypes of the indicated homozygous progenies of mutants between E2 and E1 family members. Because E1La and E1Lb are linked, we
characterized eight genotypic combinations. (B–I) Flowering time (B), scored as days after emergence (DAE), and the agronomic traits of plant height (C),
measured from the cotyledonary node of the main stem to the apex in centimeters, branch number per plant (D), number of nodes (E), pod number per
plant (F), total grain number per plant (G), hundred‐seed weight (H), and grain weight per plant (I). The plants were grown in a standard field in Shijiazhuang
(37°27′N) under natural conditions in 2021. W82, wild‐type Williams 82. At least six plants were scored for each phenotype; data are means ±SD, with all
individual values shown as a dot. P<0.05, as determined by multiple comparison testing by one‐way analysis of variance. Different lowercase letters
indicate significant differences between genotypes.
GI orthologs regulate soybean flowering time and yieldJournal of Integrative Plant Biology
www.jipb.net January 2023
|
Volume 65
|
Issue 1
|
188–202 193
17447909, 2023, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jipb.13398 by Shanghai Jiao Tong University, Wiley Online Library on [06/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

Figure 4. E2 family members activate the transcription of E1 and E1‐Like genes in different degrees
(A) Relative transcript levels of E1 family genes in e2 single and double mutants. The plants were grown under long‐day (16 h light/8 h dark) conditions in a
plant growth chamber. Fully developed trifoliate leaves of wild‐type and e2 family mutants at the 20‐days after emergence stage. Data are means ±SE from
three biological replicates. Student's t‐test. **P<0.01 for E1 and E1Lb transcription level in e2,e2 e2la,e2 e2lb,e2la e2lb at ZT 4 and ZT 4, ZT 8, ZT 16,
respectively; for E1La in e2 e2la,e2 e2lb,e2la e2lb at ZT 4 and ZT 8. (B) Schematic diagrams of the effector constructs for E2 family members and the
effector construct consisting of the E1 promoter driving firefly luciferase transcription used for the transient infiltration assay in Nicotiana benthamiana
leaves. (C) E2 family members promote E1 transcription in a dual luciferase reporter assay in N. benthamiana.(D, E) Relative transcript levels of E2 and
E2‐Like family members in single and double mutants of E2 family members. The plants were grown under long‐day (16 h light/8 h dark) conditions in a plant
growth chamber from fully open third trifoliate leaves from different plants. β‐Tubulin (TUB) was used as a reference transcript. Data are mean ±SE from
three biological replicates. Student's t‐test. **P<0.01 for E2 transcription level in e2la and e2la e2lb at ZT 12 compared with W82. Zeitgeber time 0 (ZT 0) is
the point of start to give light at 6:00 a.m. every day.
GI orthologs regulate soybean flowering time and yield Journal of Integrative Plant Biology
194 January 2023
|
Volume 65
|
Issue 1
|
188–202 www.jipb.net
17447909, 2023, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jipb.13398 by Shanghai Jiao Tong University, Wiley Online Library on [06/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

(Figure S5), which was consistent with a previous report
(Li et al., 2013). Taken together, our data suggest that E2 family
members do not affect the transcription of their encoding
homologs but form homodimers and heterodimers, which may
explain the observed functional redundancy seen in
photoperiod‐regulated flowering and yield in soybean.
The H3 haplotype of E2 is the ancestral allele and is
adapted to low latitude locations
To explore the evolutionary history of E2, we assessed the
extent of natural variation in the coding sequences of E2
genes relative to the W82 reference genome among a panel
of 2,345 worldwide soybean accessions (1,272 cultivars, 624
landraces, and 449 wild accessions) whose genomes were
sequenced (Lu et al., 2020;Dong et al., 2021;Dong et al.,
2022;Kou et al., 2022). We identified 14 haplotypes in E2,of
which haplotypes H1‐H3 were the main alleles present in
most varieties (Figure S6A). A valine residue at position 220 in
the predicted protein, present in the H3 haplotype, was highly
conserved across E2 orthologs, while the H1 haplotype was
characterized by a single G‐to‐A SNP leading to a sub-
stitution from valine to isoleucine at this position (from V220I)
(Figure S6A, B). We established that the H1 and H2 originated
from H3, which first occurred in wild soybeans, suggesting it
may be the ancestral allele for E2 (Figures 5B,S6C).
Population genetic association analysis of flowering time in a
1094‐accession diversity panel under LD conditions in Harbin
showed that the H3 allele is associated with the latest flow-
ering, followed by H1, which is present in W82, suggesting
that the W82 background carries a weak loss‐of‐function E2
allele (Figure 5A). The H2 haplotype (an A‐to‐T SNP at
position 1,582 resulting in a stop codon leading to a non‐
functional e2 allele) was mainly distributed in high‐latitude
regions of China and the United States, which contributed to
early flowering time in these accessions (Figure S6D). The
relative dominant H3 allele appeared at low frequency in
landraces and cultivars, which were substantially present in
low‐latitude regions such as Brazil due to the extreme late
flowering time conferred by H3 (Figure S6D).
To validate the functional significance of the H1 and H3
haplotypes to flowering time, we cloned the E2 coding
sequences from accessions harboring the H1 or H3 alleles to
test complementation of e2 single mutants when placed under
the control of the E2 promoter from W82. We obtained two
positive transgenic lines for each of the two haplotypes (Figure
S7). Under LD conditions, the four transgenic lines all largely
rescued the early flowering phenotype of the e2 mutant (Figure
5G). Importantly, both E2pro:H3 lines displayed a significantly
later flowering than the two E2pro:H1 lines, indicating a stronger
effect of the H3 haplotype on regulating flowering time in soy-
bean compared to the H1 haplotype, which was consistent with
the population genetic association analysis (Figure 5A). Under
SD conditions, the E2pro:H3 lines still flowered slightly later
than the E2pro:H1 lines, further supporting the notion that the
H3 haplotype exerts a strong effect on flowering time than the
H1 haplotype (Figure 5H).
We also assessed the natural variation and distribution
pattern of the two E2‐like genes among the same 2,345‐
accession panel. The H3 haplotype of E2La was strongly en-
riched among wild accessions (Figure S8); based on the very
late flowering time measured in Harbin for these accessions
(Figure 5C), we hypothesize that the H3 haplotype of E2La is a
fully functional allele. Accessions carrying the two loss‐of‐
function H1 and H2 alleles of E2La flowered earlier and were
widely distributed in landraces and cultivars, which were
mostly enriched in Northern China, the United States, and
Brazil (Figures 5C, D,S8B, C). Similarly, the frequency of the
H1 haplotype of E2Lb contributing to early flowering gradually
increased among landraces and cultivars and conferred a
fitness advantage to higher latitudes (Figures 5E,S9). The
H2 haplotype of E2Lb harbored a 3‐bp insertion that might
represent the ancestral allele of the H1 allele and was asso-
ciated with later flowering under LD conditions (Figure 5E, F).
In summary, we identified the dominant alleles of E2 and its
homologs among natural varieties; these observations support
the crucial roles of diversity in E2 homologs in the variation of
soybean flowering time.
Selection of superior alleles of E2 family genes
improves soybean adaptation
The results above confirmed that E2 family members control
flowering time in soybean with an asymmetric redundancy,
as evidenced by the flowering time of single and double
mutants, which led us to determine whether selection of E2
family genes might also improve soybean adaptation. To this
end, we investigated the contribution of natural alleles of E2
and its homologs to plant adaptation by designating 21
allelic combinations representing all pairwise combinations
between the main alleles for E2,E2La, and E2Lb (E2‐H1
E2La‐H1,E2‐H2 E2Lla‐H1,E2‐H3 E2La‐H1;E2‐H1 E2La‐H2,
E2‐H2 E2La‐H2,E2‐H3 E2La‐H2,E2‐H1 E2La‐H3,E2‐H2
E2La‐H3,E2‐H3 E2La‐H3;E2‐H1 E2Lb‐H1,E2‐H2 E2Lb‐H1,
E2‐H3 E2Lb‐H1;E2‐H1 E2Lb‐H2,E2‐H2 E2Lb‐H2,E2‐H3
E2Lb‐H2;E2La‐H1 E2Lb‐H1,E2La‐H2 E2Lb‐H1,E2La‐H3
E2Lb‐H1;E2La‐H1 E2Lb‐H2,E2La‐H2 E2Lb‐H2,E2La‐H3
E2Lb‐H2) and analyzed the geographic distribution of these
allelic combinations in the 2,324 accessions covering various
latitudes (Figure 6). We determined that the allele pairs E2‐H2
E2La‐H1 and E2‐H2 E2La‐H2 are mainly distributed in high‐
latitude regions (Figure 6A), while the pairs E2‐H3 E2La‐H1,
E2‐H3 E2La‐H2, and E2‐H3 E2Lb‐H2 were restricted to low‐
latitude regions (Figure 6A, C). The E2‐H2 E2Lb‐H1 genotype
was mostly enriched in Northern China and was also widely
distributed in the United States (Figure 6C). The E2La‐H2
E2Lb‐H1 and E2La‐H2 E2Lb‐H2 genotypes were widely dis-
tributed in Northern China, with E2La‐H1 E2Lb‐H1 being
mostly enriched in the Huanghuai area of China and in the
United States (Figure 6E). Finally, to explore the functional
significance of E2 family genes, we measured the flowering
time associated with each of the 21 allelic combinations
across accessions. We observed that accessions carrying
weak or non‐functional alleles at E2,E2La, and E2Lb flower
GI orthologs regulate soybean flowering time and yieldJournal of Integrative Plant Biology
www.jipb.net January 2023
|
Volume 65
|
Issue 1
|
188–202 195
17447909, 2023, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jipb.13398 by Shanghai Jiao Tong University, Wiley Online Library on [06/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

Figure 5. Natural variation in E2 family genes
Variation in flowering time in 1,094 accessions carrying the E2 (A), E2La (C),orE2Lb (E) alleles grown in Harbin in 2019. Proportions of E2 (B),E2La (D),andE2Lb
(F) alleles in wild soybean, landraces, and improved cultivars. Data are combined from diversity panels with 1,295 (Lu et al., 2020), 329 (Dong et al., 2021), 372
(Dong et al., 2022), and 349 (Kou et al., 2022) accessions. (G–H) E2 encoded by haplotype H3 has a stronger activity than that encoded by haplotype H1.Thee2
mutant was transformed with a construct comprising the E2 promoter and the E2 allele from the H1 or H3 haplotypes. Two positive transgenic lines were
phenotyped per construct for flowering time. DAE, days after emergence. Plantsweregrowninaplantgrowthchamberinlong‐day (16 h light/8 h dark) (G) and
short‐day (12 h light/12h dark) (H) conditions. Data are mean ±SD, with all individual plants shown as a dot. At least six plants were phenotyped per line. P<0.05,
as determined by multiple comparison testing by one‐way analysis of variance. Different lowercase letters indicate significant differences between genotypes.
GI orthologs regulate soybean flowering time and yield Journal of Integrative Plant Biology
196 January 2023
|
Volume 65
|
Issue 1
|
188–202 www.jipb.net
17447909, 2023, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jipb.13398 by Shanghai Jiao Tong University, Wiley Online Library on [06/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

earlier than accessions carrying functional alleles at the
same loci, with the exception of E2La‐H3 E2Lb‐H1 vs E2La‐
H3 E2Lb‐H2 (Figure 6B, D, F). These results were consistent
with our double mutant analysis (Figure 2B) and confirmed
that E2 family genes control flowering time with asymmetric
redundancy. Latitudinal correlation analysis of 21 allelic
combinations across accessions displayed a further evi-
dence that accessions carrying weak or non‐functional
alleles mainly distributed at higher latitude areas than that
functional alleles, which were consistent with the flowering
time (Figure S10). Therefore, the selection of natural alleles
for E2 family members has played an important role in
expanding the soybean cultivation area into high‐latitude re-
gions and improving soybean adaptation.
DISCUSSION
Roles of E2 and E2‐Like proteins in regulating
photoperiod‐mediated flowering
GI is a circadian clock–controlled gene that is involved in the
control of flowering in Arabidopsis and in crop plants such as
garden pea (Pisum sativum,LATE BLOOMER1), bread wheat
(Triticum aestivum,TaGI1), barley (Hordeum vulgare,HvGI),
Figure 6. Geographical distribution of the main haplotype combinations for E2 family genes
Geographical distribution (A, C, E) and flowering time (B, D, F) associated with the three main haplotypes at E2, the three haplotypes at E2La, and the two
haplotypes at E2Lb. The genotype data represent 2,345 accessions, and the phenotypic data comprise 1,094 accessions, including wild soybean,
landraces, and improved cultivars. The size of each pie chart represents the proportion of accessions. DAE, days after emergence. P<0.05, as determined
by multiple comparison testing by one‐way analysis of variance. Different lowercase letters indicate significant differences between genotypes. DAE, days
after emergence.
GI orthologs regulate soybean flowering time and yieldJournal of Integrative Plant Biology
www.jipb.net January 2023
|
Volume 65
|
Issue 1
|
188–202 197
17447909, 2023, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jipb.13398 by Shanghai Jiao Tong University, Wiley Online Library on [06/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

and rice (Oryza sativa,OsGI)(Park et al., 1999;Huq et al.,
2000;Dunford et al., 2005;Zhao et al., 2005;Hecht et al.,
2007). As the focus of this work, we characterized the phe-
notypes of mutants in E2 and E2‐like genes when grown in an
artificial incubator and in the field under different natural
photoperiod conditions. We established that only the e2
single mutant flowered earlier than the wild‐type cultivar W82,
while e2la and e2lb single mutants displayed no obvious
flowering time phenotype. However, the three double
mutants between E2 family members accelerated the tran-
sition to the reproductive stage, with e2la e2lb flowering later
than the two e2 e2la and e2 e2lb double mutants, even
later than the e2 single mutant (Figure 2A, B). These results
therefore indicated that E2 family members regulate flowering
time redundantly and that E2 is the main contributor to
flowering time in soybean, with E2La and E2Lb only affecting
flowering time in the absence of E2 function. Interestingly, in
high‐latitude regions like Harbin and Shijiazhuang, the e2
single mutant exhibited an early flowering phenotype that
shortened the time to maturity but increased yield. We also
observed that the yield of the e2la single mutant decreased,
while that of e2lb showed no change in Shijiazhuang (Figures
2E,S1C). However, neither the e2 nor e2la single mutant
suffered from a change in yield in the Hefei region, while that
of the e2lb single mutant increased, although the flowering
time of all genotypes was the same (Figure 2F). It is possible
that the three E2 members differentially regulate grain yield in
a photoperiod‐dependent manner. Future work will include a
detailed investigation of the yield structure components of
single and double mutants between E2 family members, such
as plant height, branch number per plant, number of nodes,
pod number per plant, total grain number per plant, and
hundred‐seed weight, to determine which factor(s) contrib-
utes specifically to yield. In addition, the phenotypes of
complete loss‐of‐function mutants in E2 family members (e2
e2la e2lb) are worth exploring.
E2 is genetically dependent on E1s in regulating
flowering and yield
In Arabidopsis, the classic photoperiodic flowering pathway
consists of the GI‐CO‐FT module, which is conserved in
many plants. Soybean is an ancient tetraploid plant, with a
rather complex genome. It was reported that soybean has
26 CO homologs, actually they are CO‐Like genes, and not
play a central role in the photoperiod response mechanism,
thus they are distinct from the CO and from Arabidopsis
(Wu et al., 2014;Cao et al., 2015). In soybean, classical
genetic analyses have identified several genes contributing
to the regulation of flowering, which have been assembled
into models reflecting the underlying gene interactions (Lin
et al., 2021). In this study, we focused on the role of the
legume‐specific E1 transcription factor in photoperiodic
flowering in relation to the other soybean maturity gene E2
and its homologs. We established that E1 family members
are epistatic to E2 both for flowering time and grain yield; in
particular, the effect on flowering time mediated by E2 was
fully dependent on E1 family genes (Figures 3,S3). In Hefei,
we determined that the yield of the e2 e1 e1la e1lb quad-
ruple mutant was higher than that of the e1 e1la e1lb triple
mutant with functional E2, although they both had the same
flowering time. We observed the opposite effects for the e2
single mutant relative to the W82 wild‐type background
(Figure S3). In Hefei, the effect of E2 on yield was largely,
but not completely, dependent on E1 family members, with
an apparent similar trend in Shijiazhuang, although these
results were not significantly different. These results sug-
gested that although E2 may control soybean yield, the E1
pathway is clearly the main player. In addition to genetic
analysis, we explored the transcriptional activation of E1
and E1‐Like genes by E2 and its homologs; we observed
that they were indeed downregulated in mutants with a
change in flowering time. Similar results were obtained in
the transient infiltration assays using a dual luciferase re-
porting assay in N. benthamiana leaves (Figure 4A–D).
Taken together, we propose a model whereby E2 regulates
flowering time and grain yield in soybean through the
transcriptional activation of the legume‐specificE1 tran-
scription factor gene family, which is not present in Arabi-
dopsis. One question arises from these results: how can E2
activate the transcription of E1 family genes? And the
activation is direct or indirect? The exploration of the
genetic interactions between E2 and E1 will be key to
answering this question.
The E2 loss‐of‐function H1 and H2 haplotypes are
derived from H3
Analyzing natural variation and deploying excellent alleles to
agricultural production practice are fast and effective ways to
improve crop breeding. We analyzed the haplotypes of E2
family members among 2,345 resequencing worldwide soy-
bean accessions, from which we identified three major E2
haplotypes, namely, H1,H2, and H3. Of these, H2 harbored a
premature stop codon in exon 10, which was characterized
as e2 and was widely distributed in both Northern and
Southern China (Figure S5). And previously confirmed the
early flowering phenotype of this allele in near isogenic lines
(Watanabe et al., 2011;Fang et al., 2017). In addition to its
early flowering phenotype, the loss of E2 function via genome
editing was associated with high yield at high‐latitude regions
(Figures 2E,S1C). Thus, the natural variation of E2‐H2 likely
was selected by the early arrival of frost‐free seasons in high‐
latitude areas for complete maturity and was artificially se-
lected by local farmers and breeders for high yield, in addition
to being amenable for multiple cropping. As a result, the allele
was selected and widely deployed.
The E2 alleles represented by the H3 and H1 haplotypes
carry an A‐to‐G transition in exon 6 that introduces a non-
synonymous amino acid substitution (V220I) (Figure S5).
Previous studies identified two E2 haplotypes and speculated
a lack of functional divergence between the two isoforms
(Wang et al., 2016;Kim et al., 2018). In this study, we iden-
tified a third haplotype in wild accessions: the H3 allele, from
GI orthologs regulate soybean flowering time and yield Journal of Integrative Plant Biology
198 January 2023
|
Volume 65
|
Issue 1
|
188–202 www.jipb.net
17447909, 2023, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jipb.13398 by Shanghai Jiao Tong University, Wiley Online Library on [06/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

which the loss‐of‐function H1 and H2 haplotypes were
derived (Figures 5,S5). We thus speculate that H3 is the
ancestral allele with the strongest function in regulating
flowering time. To test this idea, we performed com-
plementation assays by introducing the full‐length E2 coding
sequence from the H1 and H3 haplotypes driven by the E2
promoter into the e2 mutant. All H1 and H3 transgenic lines
rescued the early floral phenotype of e2 under LD conditions,
although H3 lines flowered later than H1 lines. Even under SD
conditions, an inducible photoperiodic condition for soybean,
the H3 transgenic lines also flowered later (Figure 5G, H).
Therefore, we conclude that the H3 haplotype is important
and should not be ignored. H3 was mainly limited to low‐
latitude regions like Southern China and Brazil, even
in combination with the functional alleles E2La and E2Lb
(Figures 6A, C,S5), thus perfectly aligning the selection rules
and geographical distribution with the function of the H3
haplotype.
The phenotypes associated with the main E2La and E2Lb
haplotypes showed that E2La is also present as three main
haplotypes, with H3 being the ancestral type with later
flowering, from which the loss‐of‐function haplotypes H1 and
H2 were derived and led to earlier flowering (Figures 5C,
D,S6). These results suggested that E2 and E2La are a pair
of more closely related homologs than E2Lb. Importantly, the
gene‐edited e2la and e2lb single mutants generated here
showed no change in their flowering time compared to W82
(Figure 2A, B). There may be some unknown genes masking
the function of E2‐Like genes due to the genetic background
of the variety W82 used for gene editing.
Finally, we propose a working model for the asymmetric
redundancy of E2 family members in flowering and adapta-
tion in soybean according to their geographical distribution
and molecular experiments (Figure 7). The diversified allele
combinations of E2 family members allow the adaptation to
different eco‐regions, which should contribute to soybean
improvement.
MATERIALS AND METHODS
Soybean accessions, growth conditions, and
phenotyping
The soybean (Glycine max L.) cultivar Williams 82 was
used for expression analysis and as the receptor back-
ground for transgenic lines. Plants were grown under SD
(12 h light/12 h dark) or LD (16 h light/8 h dark) conditions
in artificial incubators (Conviron, Canada) with a relative
humidity of 70% (±10%) and a constant temperature of
25°C. The mutant materials for phenotyping were planted
at the experimental station in the field of Shijiazhuang and
Hefei in 2020. The 1,094‐accession panel was planted
under natural daylength conditions in Harbin, China, in
2019 to evaluate flowering time, which was recorded as
thenumberofDAEwhenthefirst flower opened (Fehr and
Cavines, 1977).
Vector construction and genetic transformation of
soybean
The target sequence adapters for the three E2 family genes
were designed and subcloned into sgRNA expression cas-
settes and then moved into the pYLCRISPR/Cas9‐DB vector
according to previously described methods (Bu et al., 2021).
For complementation assay vectors, the coding sequences
of the H1 and H3 haplotypes of E2 were individually cloned
from their corresponding varieties, while the promoter
sequence was amplified from W82 genomic DNA and in-
serted into the vector pTF101. The resulting CRISPR/Cas9
plasmid and overexpression plasmid were transformed into
the Agrobacterium (Agrobacterium tumefaciens) strain
EHA101 for soybean transformation into Williams 82. Trans-
formation was performed using the cotyledonary node
method as described previously (Flores et al., 2008). The
primers are listed in Table S1.
RNA extraction and RT‐qPCR
Total RNA was extracted from leaves of the W82 cultivar and
mutants following the extraction kit instructions (cat. no.
CW0581; Cowin Biotech). Total RNA was reverse transcribed
into cDNA using a first‐strand cDNA synthesis kit (cat. no.
RR047; Takara). Reverse transcription quantitative PCR
(RT‐qPCR) was performed using a real‐time PCR kit (cat. no.
RR430; Takara) on a Roche Light Cycler 480 instrument
(Roche Molecular Biochemicals, USA). Each experiment was
performed using three biological replicates from RNA sam-
ples extracted from three independent plants, each with three
technical replicates. The expression levels of the target genes
were normalized to those of Tubulin. All primers used for
RT‐qPCR are listed in Table S1.
Transient infiltration assays
An approximately 3‐kb E1 promoter fragment was amplified
from W82 and introduced into pGreen0800‐LUC/REN vector
to generate the 35S:REN‐E1pro:LUC reporter construct. The
plasmids 35S:E2‐HA,35 S:E2La‐HA,35S:E2Lb‐HA, and 35S:
HA were used as effector constructs. The constructs were
introduced into Agrobacterium strain GV3101. Positive col-
onies were grown overnight before being resuspended in
infiltration buffer at an OD
600
of 0.4–0.6. The cell suspensions
were mixed equally to co‐infiltrate into fresh N. benthamiana
leaves. At least three leaves from independent N. ben-
thamiana plants were infiltrated. The LUC and REN activities
were measured using a Luciferase 1000 Assay System (cat.
no. E4550; Promega) and a Renilla Luciferase Assay System
(cat. no. E2820; Promega), respectively. The final transcrip-
tional activity was expressed as the ratio between LUC
and REN activities. The primers used in this assay are listed
in Table S1.
Yeast two‐hybrid assay
The full‐length coding sequences of E2,E2La, and E2Lb were
amplified from W82 and cloned into pGBKT7 and pGADT7.
The resulting plasmids were co‐transformed as pairs into
GI orthologs regulate soybean flowering time and yieldJournal of Integrative Plant Biology
www.jipb.net January 2023
|
Volume 65
|
Issue 1
|
188–202 199
17447909, 2023, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jipb.13398 by Shanghai Jiao Tong University, Wiley Online Library on [06/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

yeast strain Y2H Gold (cat. no. 630498; Clontech). Positive
yeast colonies were selected on synthetic defined (SD)
medium lacking Trp and Leu (SD–Leu–Trp) according to the
manufacturer's instructions. Protein interaction was then
tested on SD medium lacking Ade, His, Leu, and Trp
(SD–Ade–His–Leu–Trp). Colonies showing a positive signal
were subsequently assessed for the activation of the lacZ
reporter gene.
Bimolecular fluorescence complementation assays
For the BiFC experiment, the CDS of E2,E2La, and E2Lb
cloned from W82, then introduced into pSPYCE(M) and
pSPYNE173 containing either N‐terminal or C‐terminal en-
hanced yellow fluorescence protein (YFP) fragments. The
constructs were transformed into A. tumefaciens strain
GV3101. Equal volumes of each culture were mixed together
for injection. Different combinations were co‐infiltrated into
4‐week‐old N. benthamiana leaves. YFP fluorescence was
observed under a confocal laser‐scanning microscope
(LSM800; Zeiss) after 48–72 h. For visualizing nuclei, leaves
were stained with 2 mg/mL 4’,6‐diamidino‐2‐phenylindole
(DAPI) for 2 h before observation.
Variation calling and annotation analysis
Paired‐end resequencing reads of the 2,345 accessions were
previously mapped to the soybean reference genome Glycine
max Wm82.a2.v1 (https://phytozome.jgi.doe.gov/pz/portal.
html#!info?alias=Org_Gmax)(Schmutz et al., 2010). The
VCF files used in this study were obtained from Lu et al.
(2020), Dong et al. (2021), Dong et al. (2022), and Kou et al.
(2022) and further filtered using VCF tools software (v.0.1.16)
(Danecek et al., 2011) with the parameters “‐‐min‐alleles
2‐‐max‐missing 0.5 ‐‐maf 0.002”. Annotation was carried out
based on the reference genome Glycine max Wm82.a2.v1
with Annovar (Wang et al., 2010).
ACKNOWLEDGEMENTS
This work was funded by the National Natural Science
Foundation of China (Grant No. 32072013, 31801383 to X. Z.),
and the National Key Research and Development Program
(Grant no. 2021YFF1001203 to X. Z.).
CONFLICTS OF INTEREST
The authors declare no conflict of interests.
AUTHOR CONTRIBUTIONS
X.Z., B.L., and F.K. designed the research, organized the
data, and wrote the manuscript; M.H., L.D., L.C., and H.N.
generated the genetic experiments and investigated agro-
nomic traits in the field; H.L. and L.W. performed the bio-
informatics analysis; L.W. and Z.H. performed RT‐qPCR,
Figure 7. Proposed seesaw model to explain the asymmetric redundancy between E2 family genes in flowering and adaptation in
soybean
Functional E2 and E2‐Like combinations activate the transcription of E1 family genes intensely and are mainly distributed in low‐latitude regions. By
contrast, single or multiple loss‐of‐function alleles prevent the transcriptional activation of E1 family genes and are mainly distributed at high‐latitude
regions. E2Ls indicates E2La and/or E2Lb. E1s indicates E1,E1La, and E1Lb. The red arrows represent activation.
GI orthologs regulate soybean flowering time and yield Journal of Integrative Plant Biology
200 January 2023
|
Volume 65
|
Issue 1
|
188–202 www.jipb.net
17447909, 2023, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jipb.13398 by Shanghai Jiao Tong University, Wiley Online Library on [06/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

transient transformation assays, and yeast two‐hybrid as-
says. All authors read and approved the manuscript.
Edited by: Zhixi Tian, Institute of Genetics and Developmental Biology,
CAS, China.
Received Jul. 17, 2022; Accepted Oct. 25, 2022; Published Oct. 26, 2022
REFERENCES
Bernard, R.L. (1971). Two major genes for time of flowering and maturity
in soybeans. Crop Sci. 11: 242–244.
Bonato, E.R., and Vello, N.A. (1999). E6, a dominant gene conditioning
early flowering and maturity in soybeans. Genet. Mol. Biol. 22:
229–232.
Bu, T.T., Lu, S.J., Wang, K., Dong, L.D., Li, S.L., Xie, Q.G., Xu, X.D.,
Cheng, Q., Chen, L.Y., Fang, C., Li, H.Y., Liu, B.H., James, L.W., and
Kong, F.J. (2021). A critical role of the soybean evening complex in the
control of photoperiod sensitivity and adaptation. Proc. Natl. Acad. Sci.
U.S.A. 118: e2010241118.
Buzzell, R.I. (1971). Inheritance of a soybean flowering response to
fluorescent‐daylength conditions. Can. J. Genet. Cytol. 13: 703–707.
Buzzell, R.I., and Voldeng, H.D. (1980). Inheritance of insensitivity to long
daylength. Soybean Genet. Newsl. 7: 26–29.
Cao,D.,Li,Y.,Lu,S.J.,Wang,J.L.,Nan,H.Y.,Li,X.M.,Shi,D.N.,
Fang,C.,Zhai,H.,Yuan,X.H.,Anai,T.,Xia,Z.J.,Liu,B.H.,and
Kong, F.J. (2015). GmCOL1a and GmCOL1b function as flowering
repressors in soybean under long‐day conditions. Plant Cell Physiol.
56: 2409–2422.
Cober, E.R., Molnar, S.J., Charette, M., and Voldeng, H.D. (2010). A
new locus for early maturity in soybean. Crop Sci. 50: 524–527.
Cober, E.R., and Voldeng, H.D. (2001). Low R:FR light quality delays
flowering of E7E7 soybean lines. Crop Sci. 41: 1823–1826.
Danecek, P., Auton, A., Abecasis, G., Albers, C.A., Banks, E., DePristo,
M.A., Handsaker, R.E., Lunter, G., Marth, G.T., and Sherry, S.T.
(2011). The variant call format and VCFtools. Bioinformatics 27:
2156–2158.
Dong, L.D., Cheng, Q., Fang, C., Kong, L.P., Yang, H., Hou, Z.H., Li, Y.
L., Nan, H.Y., Zhang, Y.H., Chen, Q.S., Zhang, C.B., Kou, K., Su, T.,
Wang, L.S., Li, S.C., Li, H.Y., Lin, X.Y., Tang, Y., Zhao, X.H., Lu, S.J.,
Liu, B.H., and Kong, F.J. (2022). Parallel selection of distinct tof5
alleles drove the adaptation of cultivated and wild soybean to high
latitudes. Mol. Plant 15: 308–321.
Dong, L.D., Fang, C., Cheng, Q., Su, T., Kou, K., Kong, L.P., Zhang, C.
B., Li, H.Y., Hou, Z.H., Zhang, Y.H., Chen, L.Y., Yue, L., Wang, L.S.,
Wang, K., Li, Y.L., Gan, Z.R., Yuan, X.H., James, L.W., Lu, S.J.,
Kong, F.J., and Liu, B.H. (2021). Genetic basis and adaptation tra-
jectory of soybean from its temperate origin to tropics. Nat. Commun.
12: 5445.
Dunford, R.P., Griffiths, S., Christodoulou, V., and Laurie, D.A. (2005).
Characterisation of a barley (Hordeum vulgare L.) homologue of the
Arabidopsis flowering time regulator GIGANTEA. Theor. Appl. Genet.
110: 925–931.
Fang, C., Ma, Y., Wu, S., Liu, Z., Wang, Z., Yang, R., Hu, G., Zhou, Z.,
Yu, H., Zhang, M., Pan, Y., Zhou, G., Ren, H., Du, W., Yan, H., Wang,
Y., Han, D., Shen, Y., Liu, S., Liu, T., Zhang, J., Qin, H., Yuan, J.,
Yuan, X., Kong, F., Liu, B., Li, J., Zhang, Z., Wang, G., Zhu, B., and
Tian, Z. (2017). Genome‐wide association studies dissect the genetic
networks underlying agronomical traits in soybean. Genome Biol.
18: 161.
Fehr, W.R., and Cavines, C.E. (1977). Stages of Soybean Development.
Special Report 80.. Ames: Iowa Agricultural Experiment Station, Iowa
Cooperative External Series, Iowa State University.
Flores, T., Karpova, O., Su, X., Zeng, P., Bilyeu, K., Sleper, D.A.,
Nguyen, H.T., and Zhang, Z.J. (2008). Silencing of GmFAD3 gene by
siRNA leads to low α‐linolenic acids (18:3) of fad3‐mutant phenotype in
soybean [Glycine max (Merr.)]. Transgenic Res. 17: 839–850.
Garner, W.W., and Allard, H.A. (1920). Effect of the relative length of day
and night and other factors of the environment on growth and re-
production in plants. J. Agric. Res. 18: 553–606.
Garner, W.W., and Allard, H.A. (1923). Further studies in photoperiodism,
the response of the plant to relative length of day and night. Agric. Res.
23: 871–920.
Hecht, V., Knowles, C.L., Vander Schoor, J.K., Liew, L.C., Jones, S.E.,
Lambert, M.J., and Weller, J.L. (2007). Pea LATE BLOOMER1 is a
GIGANTEA ortholog with roles in photoperiodic flowering, deetiolation,
and transcriptional regulation of circadian clock gene homologs. Plant
Physiol. 144: 648–661.
Huq, E., Tepperman, J.M., and Quail, P.H. (2000). GIGANTEA is a
nuclear protein involved in phytochrome signaling in Arabidopsis. Proc.
Natl. Acad. Sci. U.S.A. 97: 9789–9794.
Kim, S.K., Kim, E.S., Kim, K.H., Ju, N.J., Lee, S., and Kang, S. (2018).
Genetic variance for flowering time conferring E2 gene in photoperiod‐
insensitive early‐maturing soybean accessions and topological dis-
tribution in Korea peninsula. Mol. Breed. 38: 148.
Kong, F., Nan, H., Cao, D., Li, Y., Wu, F., Wang, J., Lu, S., Yuan, X.,
Cober, E.R., Abe, J., and Liu, B. (2014). A new dominant gene E9
conditions early flowering and maturity in soybean. Crop Sci. 54:
2529–2535.
Kou, K., Yang, H., Li, H.Y., Fang, C., Chen, L.Y., Yue, L., Nan, H.Y.,
Kong, L.P., Li, X.M., Wang, F., Wang, J.H., Du, H.P., Yang, Z.Y., Bi,
Y.D., Lai, Y.C., Dong, L.D., Cheng, Q., Su, T., Wang, L.S., Li, S.C.,
Hou, Z.H., Lu, S.J., Zhang, Y.H., Che, Z.J., Yu, D.Y., Zhao, X.H., Liu,
B.H., and Kong, F.J. (2022). A functionally divergent SOC1 homologue
improves soybean yield and latitudinal adaptation. Curr. Biol. 32: 1–15.
Li, F., Zhang, X.M., Hu, R.B., Wu, F.Q., Ma, J.H., Meng, Y., and Fu, Y.F.
(2013). Identification and molecular characterization of FKF1 and GI
homologues genes in soybean. PLoS ONE 8: e79036.
Lin, X., Liu, B., Weller, J.L., Abe, J., and Kong, F. (2021). Molecular
mechanisms for the photoperiodic regulation of flowering in soybean.
J. Integr. Plant Biol. 63: 981–994.
Lin, X.Y., Dong, L.D., Tang, Y., Li, H.Y., Cheng, Q., Li, H., Zhang, T., Ma,
L.X., Xiang, H.L., Chen, L.N., Nan, H.Y., Fang, C., Lu, S.J., Li, J.G.,
Liu, B.H., and Kong, F.J. (2022). Novel and multifaceted regulations of
photoperiodic flowering by phytochrome A in soybean. Proc. Natl.
Acad. Sci. U.S.A. 119: 41.
Lu, S.J., Dong, L.D., Fang, C., Liu, S.L., Kong, L.P., Cheng, Q., Chen, L.
Y., Su, T., Nan, H.Y., Zhang, D., Zhang, L., Wang, Z.J., Yang, Y.Q.,
Yu, D.Y., Liu, X.L., Yang, Q.Y., Lin, X.Y., Tang, Y., Zhao, X.H., Yang,
X.Q., Tian, C.G., Xie, Q.G., Li, X., Yuan, X.H., Tian, Z.X., Liu, B.H.,
Weller, J.L., and Kong, F.J. (2020). Stepwise selection on homologues
PRR genes controlling flowering and maturity during soybean domes-
tication. Nat. Genet. 52: 428–436.
Lu, S.J., Zhao, X.H., Hu, Y.L., Liu, S.L., Nan, H.Y., Li, X.M., Fang, C.,
Cao, D., Shi, X.Y., Kong, L.P., Su, T., Zhang, F.G., Li, S.C., Wang, Z.,
Yuan, X.H., Cober, E.R., Weller, J.L., Liu, B.H., Hou, X.L., Tian, Z.X.,
and Kong, F.J. (2017). Natural variation at the soybean Jlocus im-
proves adaptation to the tropics and enhances yield. Nat. Genet. 49:
773–779.
McBlain, B., and Bernard, R. (1987). A new gene affecting the time of
flowering and maturity in soybeans. J. Hered. 78: 160–162.
Park, D.H., Somers, D.E., Kim, Y.S., Choy, Y.H., Lim, H.K., Soh, M.S.,
Kim, H.J., Kay, S.A., and Nam, H.G. (1999). Control of circadian
rhythms and photoperiodic flowering by the Arabidopsis GIGANTEA
gene. Science 285: 1579–1582.
Ray, J.D., Hinson, K., Mankono, J., and Malo, M.F. (1995). Genetic‐
control of a long‐juvenile trait in soybean. Crop Sci. 35: 1001–1006.
GI orthologs regulate soybean flowering time and yieldJournal of Integrative Plant Biology
www.jipb.net January 2023
|
Volume 65
|
Issue 1
|
188–202 201
17447909, 2023, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jipb.13398 by Shanghai Jiao Tong University, Wiley Online Library on [06/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

Samanfar, B., Molnar, S.J., Charette, M., Schoenrock, A., Dehne, F.,
Golshani, A., Belzile, F., and Cober, E.R. (2016). Mapping and iden-
tification of a potential candidate gene for a novel maturity locus, E10,
in soybean. Theor. Appl. Genet. 130: 377–390.
Schmutz, J., Cannon, S.B., Schlueter, J., Ma, J., Mitros, T., Nelson, W.,
Hyten, D.L., Song, Q., Thelen, J.J., Cheng, J., Xu, D., Hellsten, U.,
May, G.D., Yu, Y., Sakurai, T., Umezawa, T., Bhattacharyya, M.K.,
Sandhu, D., Valliyodan, B., Lindquist, E., Peto, M., Grant, D., Shu, S.
Q., Goodstein, D., Barry, K., Griggs, M.F., Abernathy, B., Du, J.C.,
Tian, Z.X., Zhu, L.C., Gill, N., Joshi, T., Libault, M., Sethuraman, A.,
Zhang, X.C., Shinozaki, K., Nguyen, H.T., Wing, R.A., Cregan, P.,
Specht, J., Grimwood, J., Rokhsar, D., Stacey, G., Shoemaker, R.
C., and Jackson, S.A. (2010). Genome sequence of the palae-
opolyploid soybean. Nature 463: 178–183.
Wang, F., Nan, H., Chen, L., Fang, C., Zhang, H., Su, T., Li, S., Cheng,
Q., Dong, L., Liu, B., Kong, F., and Lu, S. (2019). A new dominant
locus, E11, controls early flowering time and maturity in soybean. Mol.
Breed. 39: 70.
Wang, K., Li, M., and Hakonarson, H. (2010). ANNOVAR: functional an-
notation of genetic variants from next‐generation sequencing data.
Nucleic Acids Res. 38: e164.
Wang, Y., Gu, Y.Z., Gao, H.H., Qiu, L.J., Chang, R.Z., Chen, S.Y., and
He, C.Y. (2016). Molecular and geographic evolutionary support for the
essential role of GIGANTEAa in soybean domestication of flowering
time. BMC Evol. Biol. 16: 79.
Watanabe, S., Xia, Z., Hideshima, R., Tsubokura, Y., Sato, S.,
Yamanaka, N., Takahashi, R., Anai, T., Tabata, S., Kitamura, K., and
Harada, K. (2011). A map‐based cloning strategy employing a residual
heterozygous line reveals that the GIGANTEA gene is involved in
soybean maturity and flowering. Genetics 188: 395–407.
Wu, F., Price, B.W., Haider, W., Seufferheld, G., Nelson, R., and Yoshie,
H. (2014). Functional and evolutionary characterization of the CON-
STANS gene family in short‐day photoperiodic flowering in soybean.
PLoS ONE 9: e85754.
Xia , Z. , Wa tan abe, S., Yamada, T., Tsubokura, Y., Nakashima, H., Zhai,
H., Anai, T., Sato, S., Yamazaki, T., Lu, S., Wu, H., Tabata, S., and
Harada, K. (2012). Positional cloning and characterization reveal the
molecular basis for soybean maturity locus E1 that regulates pho-
toperiodic flowering. Proc. Natl. Acad. Sci. U.S.A. 109: E2155–
E2164.
Xu, M.L., Yamagishi, N., Zhao, C., Takeshima, R., Kasai, M.,
Watanabe, S., Kanazawa, A., Yoshikawa, N., Liu, B.H., Yamada, T.,
and Abe, J. (2015). The soybean‐specific maturity gene E1 family of
floral repressors controls night‐break responses through down‐
regulation of FLOWERING LOCUS T orthologs. Plant Physiol. 168:
1735–1746.
Zhao, X.Y., Liu, M.S., Li, J.R., Guan, C.M., and Zhang, X.S. (2005). The
wheat TaGI1, involved in photoperiodic flowering, encodes an Arabi-
dopsis GI ortholog. Plant Mol. Biol. 58: 53–64.
SUPPORTING INFORMATION
Additional Supporting Information may be found online in the supporting
information tab for this article: https://onlinelibrary.wiley.com/doi/10.1111/
jipb.13398/suppinfo
Figure S1. The transcription levels of E2 members in their respective
mutants
Figure S2. Phenotypes of e2 and e2‐like mutants
Figure S3. Phenotypes of mutants in E2 and E1 family genes in the in-
cubator
Figure S4. Phenotypes of mutants in E2 and E1 family genes in Hefei
Figure S5. E2 family members can form homodimers and heterodimers
Figure S6. Origin of E2 haplotypes and their geographical distribution
Figure S7. The relative transcript levels of E2 in transgenic complementary
lines
Figure S8. Origin of E2La haplotypes and their geographical
distribution
Figure S9. Origin of E2Lb haplotypes and their geographical
distribution
Figure S10. Association test between latitude and flowering time in the
main haplotype combinations for E2 family genes
Table S1. Primers used in this study
GI orthologs regulate soybean flowering time and yield Journal of Integrative Plant Biology
202 January 2023
|
Volume 65
|
Issue 1
|
188–202 www.jipb.net
17447909, 2023, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jipb.13398 by Shanghai Jiao Tong University, Wiley Online Library on [06/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License