Molecular basis of the key regulator WRINKLED1 in plant oil biosynthesis.

Vegetable oils are not only major components of human diet but also vital for industrial applications. WRINKLED1 (WRI1) is a pivotal transcription factor governing plant oil biosynthesis, but the underlying DNA-binding mechanism remains incompletely understood. Here, we resolved the structure of Arabidopsis WRI1 (AtWRI1) with its cognate double-stranded DNA (dsDNA), revealing two antiparallel β sheets in the tandem AP2 domains that intercalate into the adjacent major grooves of dsDNA to determine the sequence recognition specificity. We showed that AtWRI1 represented a previously unidentified structural fold and DNA-binding mode. Mutations of the key residues interacting with DNA element affected its binding affinity and oil biosynthesis when these variants were transiently expressed in tobacco leaves. Seed oil content was enhanced in stable transgenic wri1-1 expressing an AtWRI1 variant (W74R). Together, our findings offer a structural basis explaining WRI1 recognition and binding of DNA and suggest an alternative strategy to increase oil yield in crops through WRI1 bioengineering.

INTRODUCTION

Numerous plants produce and accumulate triacylglycerol (TAG; most frequently recognized as vegetable oils) in seeds, which is an important carbon and energy resource supporting seedling development. Vegetable oils are essential for the human diet and have important industrial applications, such as the production of lubricants, detergents, and biodiesel fuels (–). In developed countries, vegetable oils provide approximately one-fourth of dietary calories. The global demand for vegetable oil is increasing very rapidly, which necessitates increased oil production (, , ).

Bioengineering to increase fatty acid (FA) biosynthesis in plant cells to improve oil yield depends on the characterization of the TAG biosynthetic pathway, which has been found to be conserved in various plant species. However, in bioengineering TAG biosynthesis, researchers generally encounter two main challenges, namely, de novo FA production by plastids and TAG assembly reactions in the endoplasmic reticulum, which are limiting factors that limit TAG yield (, ). Converting acetyl-CoA into malonyl-CoA through acetyl-CoA carboxylase (ACCase) activity is a pivotal rate-limiting step that regulates FA production in plastids. However, efforts to manipulate ACCase activity have achieved only moderate success, which might be due to regulation of the carbon flux needed for FA generation through multiple steps (, ). Thus, manipulation of a transcriptional regulator that can control the expression of multiple genes involved in de novo FA biosynthesis has become an ideal strategy for bioengineering increases in FA generation to improve plant oil yield (–).

To date, a wealth of information has suggested the pivotal role played by WRINKLED1 (WRI1) in governing plant oil biosynthesis. WRI1 was discovered as a member of the APETALA2 (AP2)/EREBP (ethylene-responsive element-binding protein) transcription factor family (, , ), and the Arabidopsis WRI1 (At3g54320) loss-of-function mutant wri1-1 exhibited an approximately 80% decrease in seed oil accumulation compared to the wild type (WT) (). Transcriptome comparisons between WT and wri1-1 in developing seeds have indicated that the genes with reduced expression in wri1-1–mutant mainly encode late glycolytic and FA biosynthetic enzymes (). Subsequent studies validated a variety of genes in glycolytic and FA biosynthesis pathways as AtWRI1 direct targets; for example, the pyruvate kinase PI-PKβ1 (which encodes a subunit of pyruvate kinase) yields pyruvate by transferring a phosphate group from phosphoenolpyruvate to ADP, and the ACCase BCCP2 (which encodes a biotin carboxyl carrier subunit of ACCase complex) converts acetyl-CoA into malonyl-CoA in the rate-limiting step of FA synthesis (, , , ). WRI1 is therefore considered a key regulator in the transcriptional control of plant oil production (, ). Consistent with this notion, AtWRI1 orthologs have been identified in various plant species, both monocots and dicots, such as Brassica napus, Brachypodium distachyon, Camelina sativa, Glycine max, Elaeis guineensis, Helianthus annuus, and Zea mays (, –). Transgenic plants overexpressing AtWRI1 or WRI1 orthologs exhibit elevated seed oil accumulation (, , , , , , ). In addition, ectopic expression of AtWRI1 or WRI1 orthologs leads to increased oil accumulation in leaves of stable transgenic plants or tobacco plants (via transient expression) (, –). AtWRI1 is posttranslationally regulated and subjected to 26S proteasome–mediated proteolysis (). AtWRI1 displays hallmarks of intrinsic disorder and harbors three intrinsically disordered regions (IDRs) (). The IDR3-PEST motif [an amino acid sequence enriched in proline (P), glutamic acid (E), serine (S), and threonine (T)] in AtWRI1 affects the stability of AtWRI1 and sequentially mediates oil production (). Recently, 14-3-3 proteins were discovered to interact with AtWRI1 and modulate AtWRI1 transcriptional activity and oil production (). AtWRI1 phosphorylation triggered by the sucrose-non–fermentation1 (SNF1)-related kinase KIN10 is important for AtWRI1 degradation (). Further study indicated that trehalose 6-phosphate plays a role in stabilizing AtWRI1 to elevate FA biosynthesis by repressing KIN10 activity (). Recent work has also revealed that AtWRI1 interacts with the teosinte branched 1/cycloidea/preoliferating cell factor 4 (TCP4) transcription factor, and TCP4 represses AtWRI1-regulated oil production and the transcriptional activity of AtWRI1 ().

In Arabidopsis, the AtWRI1-binding cis-element AW-box ([CnTnG](n)7[CG]; n, any nucleotide) has been recently identified in the promoters of AtWRI1 target genes (). However, how AtWRI1 specifically binds to its cognate DNA to function as a transcriptional regulator remains unclear. Given the significance of WRI1 action in regulating plant oil production, structural information of WRI1 and its interaction with DNA is greatly needed. Here, we report the crystal structure of AtWRI1 in complex with its cognate DNA at 2.35 Å resolution. On the basis of our structural information, a mutagenesis assay on AtWRI1 revealed a series of residues key to DNA binding as well as an engineered AtWRI1 variant (AtWRI1W74R) that increased plant oil yield.

RESULTS

AtWRI1 binds its cognate dsDNA with high affinity in vitro

Because WRI1 is a pivotal regulator in plant oil biosynthesis, structural information to characterize the mechanism of its interaction with target DNA is of great interest for a better understanding of WRI1 regulatory mechanism and for using protein engineering to boost oil yield. To this end, we sought to biochemically characterize WRI1 in the model organism Arabidopsis thaliana. Therefore, a construct with residues from A58 to A307 of AtWRI1, including two AP2 domains and an IDR (S227-G302), was designed, and the recombinant protein was prepared on a large scale with the Escherichia coli expression system (Fig. 1A and fig. S1A). For the cognate DNA of AtWRI1, DNA oligos containing the aforementioned AW-box in the BCCP2 (-34/-11) promoter region (5′-TACTTCCTCGGTTTCATCGTCCAC-3′) were synthesized, and double-stranded DNA (dsDNA) was experimentally prepared (Fig. 1B) (). To test whether the purified recombinant AtWRI1 can form a stable complex with the prepared dsDNA in vitro, analytical size exclusion chromatography was carried out with the AtWRI1 protein, dsDNA, and AtWRI1-DNA complex, respectively (Fig. 1C). As expected, the elution volume was ~13.3 ml after mixing the dsDNA with AtWRI1, with excess DNA eluting at ~15.2 ml and with AtWRI1 eluting only at ~14.7 ml, which suggested that AtWRI1 bound to the dsDNA. To further characterize the interaction between AtWRI1 and dsDNA, an isothermal titration calorimetry (ITC) assay was performed. The ITC results demonstrated that AtWRI1 displayed a strong binding affinity for its target dsDNA, with a dissociation constant (Kd) value of ~0.48 ± 0.05 μM. Furthermore, the stoichiometry of AtWRI1 to DNA was revealed to be 1:1 (Fig. 1D).

Fig. 1.

AtWRI1 binds its cognate dsDNA with high affinity.

(A) Domain organization of AtWRI1. The domain boundaries are numbered with the first AP2 domain in violet and the second AP2 domain in lime. (B) The 24 base pairs of the dsDNA sequence used in this study with the conserved motif are in green and represent the AW-box ([CnTnG](n)7[CG]). (C) Analytical size exclusion chromatograph of AtWRI1, dsDNA, and the AtWRI1-dsDNA complex. The elution volumes were 14.7, 15.2, and 13.3 ml for these three samples as indicated. mAU, milli-Absorbance Unit. (D) ITC measurement of AtWRI1 binding to DNA with the fitted dissociation constant (Kd) and binding ratio shown. DP, differential power.

Crystal structure of the AtWRI1-DNA complex

AP2 domain–containing transcription factors have been widely found in plants and play essential roles (–). Given that AtWRI1 contains two AP2 domains and forms a stable complex with dsDNA, we subsequently sought to determine the crystal structure of the AtWRI1-DNA complex through x-ray crystallography. The complex was crystallized with the space group C222, and the structure was resolved through the selenium (Se)–single-wavelength anomalous diffraction (SAD) phasing method (). Then, the resolution was extended with a native crystal, and the final model was built and refined to 2.35 Å resolution (table S1). All 24 base pairs of dsDNA and the region from residues S59 to I230 were unambiguously modeled. We were unable to observe sufficient density to model the C-terminal region from residues D231 to A307, which was probably degraded during the crystallization process or due to its flexibility. The overall AtWRI1-DNA complex was assembled with the two AP2 domains interacting with the dsDNA at the two major adjacent grooves (Fig. 2A). The two AP2 domains, colored violet and lime to indicate the first and second domains, respectively, and structurally similar, as indicated by a root mean square deviation value of 0.687 for all the main chain atoms, were found to be composed of a β sheet (antiparallel β strands β1 to β3 and β4 to β6) at the N terminus, followed by an α helix (α1 and α4) at the C terminus (Fig. 2A and fig. S1, B and C). This α helix was packed next to the β strands, connecting mainly through hydrophobic interactions that stabilize the domain structure. The α1 and α4 helices were enriched with the hydrophobic residues Ala with a small side chain, which may have led to the tight packing of the helices with the β strands that was observed (fig. S1, D and E). The residues E107, H112, Y114, D115, and Y121 in α1 in the first AP2 domain interacted with R67, F78, and D84 in the β sheet (β1 to β3) and Y135 in the connecting loops (fig. S1F). Similarly, residues Y208, D209, and Y215 in α4 in the second AP2 domain interacted with R186 on the β4 to β6 strands and R216 and Y229 in its loops, respectively (fig. S1G). The linker region between two AP2 domains consisted of two helices, α2 and α3, forming a “V”-shaped arrangement to tightly secure the first AP2 domain in close proximity to helix α1. Notably, an evident kink at Y135 in helix α2 was observed owing to the strong interactions with residues H112, D115, and L116 in helix α1 (fig. S1F). In addition, helix α3 and its following loop in the linker region (involving residues L151, L154, R155, R156, S159, F161, and R163) formed bilateral interactions with both AP2 domains and DNA, thereby stabilizing the entire structure (fig. S1H). In particular, R163 in the domain linker formed a perfect salt bridge with D105 in the first AP2 domain and further stabilized the packing of the two AP2 domains (fig. S2A).

Fig. 2.

Structure of the AtWRI1-DNA complex.

(A) Overall structure of the AtWRI1-DNA complex. The AtWRI1 protein and DNA are shown as cartoons. The first AP2 domain is violet, the second AP2 domain is lime, and the linker between the two AP2 domains is wheat (the color code for AtWRI1 is the same throughout the paper unless otherwise stated). The secondary structural elements in AtWRI1 are labeled numerically. (B and C) The specific recognition of DNA by AtWRI1 at the first motif (5′C6T8G10) in the AW-box ([CnTnG](n)7[CG]) is shown in (B) and the second motif [5′C18G19] in the AW-box ([CnTnG](n)7[CG]) is shown in (C). The protein residues that contribute to the specificity of DNA binding are labeled, and the interacting nucleotides are shown in stick mode. The hydrogen-bonding and Pi-Pi interactions are indicated by dashed lines and a double-ended arrow, respectively. (D) Illustration of the detailed interactions of AtWRI1 with DNA. The conserved nucleotides (AW-box motif sequences) are marked by red stars, and the interactions are shown with dashed lines and the interacting amino acids labeled on the two sides. The DNA bases are labeled (T, thymine; A, adenine; C, cytosine; G, guanine), and the pentose sugars are numbered following the 5′ to 3′ direction. The two dsDNA chains are colored chocolate and green.

The two β sheets, with antiparallel strands (β1 to β3 and β4 to β6) in the two AP2 domains, intercalated into the two adjacent major grooves of dsDNA to establish extensive interactions (Fig. 2, A to C). The residues, including T70, H72, R73, W74, T75, E79, H81, W83, Q92, K94, Q98, and Y100, located in strands β1 to β3 and the connecting loops, formed interactions with dsDNA (Fig. 2, B and D). Our structure demonstrated that Y100 interacted with the DNA phosphate backbone to further enhance WRI1 binding to dsDNA, in line with a previous report showing that mutation of this residue resulted in a decrease in oil yield (). Notably, residues H72, E79, W83, and Q98 inserted deep into the major groove and specifically recognized the DNA sequence 5′C6T8G10 mainly through H-bonding and Pi-Pi (π-π) interactions (Fig. 2B and fig. S2B). Note that these three nucleotides represent the first conserved motif in the AW-box ([CnTnG](n)7[CG]), which has been found to be critical for AtWRI1 binding (). The residues R62, S63, and R67 in the N-terminal loop before the antiparallel strands β1 to β3 as well as R155 in the domain linker contributed to DNA binding by interacting with its phosphate backbone (Fig. 2D). Similarly, residues A172 (via a water molecule), H174, H175, E181, R183, G185, V187, Y192, Y194, and Y208 in strands β4 to β6 and the connecting loops in the second AP2 domain formed extensive contacts with dsDNA in the major groove. In particular, residues H175, E181, and R183 specifically interacted with nucleobase 5′C18G19, which accounted for the sequence recognition of the second motif CG in the AW-box ([CnTnG](n)7[CG]) (Fig. 2C). The guanidinium group in R183 formed strong bidentate interactions with O6 and N7 in the G7 nucleobase. Notably, Y192 interacted with 5′A16T17 without affecting the recognition motif (fig. S2C). In addition, the loop residue R169 before β4 wedged into the minor groove that connected the two major grooves where the two AP2 domains established dominant interactions with DNA (Fig. 2A and fig. S1H).

Sequence alignment of WRI1 from diverse plant species indicated that the two AP2 domains have been extensively conserved, although variations in the amino acid sequences and lengths of both the N and C termini were identified (fig. S3). All the residues involved in DNA binding have been highly conserved, perhaps demonstrating a universal binding mechanism for WRI1 proteins to DNA. In particular, the residues H72, E79, W83, and Q98 in the first AP2 domain and H175, E181, and R183 in the second AP2 domain, which were involved in specific interactions with DNA nucleobases to sequence motif recognition, have been completely conserved. Together, the data show that our structure observed at high resolution offers atomic insights into how AtWRI1 binds to the target DNA. In particular, the structure shows the rationality of the previously reported DNA-binding specificity of AtWRI1, at the AW-box ([CnTnG](n)7[CG]), which has been characterized, and a single mutation in the five conserved nucleotides has been shown to result in markedly reduced DNA binding ().

Structural comparison of AtWRI1 with other AP2 domain–containing proteins

Comparison of our structure with that of the DNA-bound Plasmodium falciparum transcription factor SIP2, which contains two AP2 domains [Protein Data Bank (PDB) ID: 6SY0], revealed that SIP2 uses an α-β turn, rather than the β sheet observed in AtWRI1, for specific DNA recognition that is quite common as indicated in our recently described SghR (salicylic acid β-glucoside hydrolase repressor)–DNA structure (Fig. 3A) (). The domain organization of SIP2 was also significantly different from that of AtWRI1, particularly in the domain linker region (Fig. 3A). There are two single AP2 domain–containing structures that show similar DNA-binding features through β sheet recognition. The first one is the Arabidopsis protein AtERF1, in which a region containing ~60 residues forms a GCC-box (AGCCGCC)–binding domain (GBD), and this GBD is quite similar to the second AP2 domain (Fig. 3B). The other structure, which has been very recently published, is the transcriptional repressor TEM1 from Arabidopsis, which is composed of an AP2 domain that binds DNA (). The overall structure of the first AP2 domain in AtWRI1 is similar to that in TEM1, whereas the loop connecting β2 and β3 identified in this study is much longer, enabling its greater involvement in DNA binding compared with that in the DNA binding of TEM1 (Figs. 2A and 3C). Except for these reported structures, no similar structure (protein- or DNA-bound form) was found through the Dali server (). Collectively, our data show a structure with a novel DNA-binding mode, in which a tandem AP2 repeat specifically recognizes the AW-box ([CnTnG](n)7[CG]) and its β sheet is inserted into the consecutive major grooves of DNA.

Fig. 3.

Structural comparison of AtWRI1 with other known AP2 domain–containing proteins.

(A) Structural comparison of the AtWRI1-DNA complex with that of the P. falciparum transcription factor SIP2 comprising two AP2 domains [Protein Data Bank (PDB) ID: 6SY0; gray]. Left: The overall DNA-binding modes of the two complexes. Right: The first AP2 domain in SIP2 was aligned to the first AP2 domain in AtWRI1 to compare the organization of the two AP2 domains. Both the DNA-binding mode and the protein domain arrangement (assembly of the two AP2 domains) are very different. (B) Structural comparison of the AtWRI1-DNA complex with that of AtERF1 (PDB ID: 1GCC). The 1GCC is gray. (C) Structural comparison of the AtWRI1-DNA complex with the AtTEM1-DNA complex (PDB ID: 7ET4). The latter is gray. Note that the structure of AtWRI1 is shown from the similar perspective in (A) to (C) and in Fig. 2A.

Effect of structural-based AtWRI1 residue mutation on DNA binding

To investigate the importance of the amino acids (H72, E79, Q98, H175, E181, and R183) that potentially contribute to the specificity of binding to the AW-box ([CnTnG](n)7[CG]) motif, we next generated all possible constructs with an amino acid mutation in each residue (Table 1). Following a similar protein purification procedure as that used with native AtWRI1, we prepared mutant proteins and subsequently measured their binding affinity for the dsDNA by ITC. Compared to the AtWRI1 native protein that binds to the DNA with a Kd value of 0.48 ± 0.05 μM, the binding affinity for the sequence with the single mutation E79A was decreased by 2-fold, the single-residue mutations of H72A and H175A reduced the affinity by ~4-fold, and the Q98A and E181A mutations reduced it by 10-fold (Table 1). Notably, the single mutation R183A completely abolished the binding of WRI1 to dsDNA. Together, in line with our structural information showing that residues H72, E79, Q98, H175, E181, and R183 specifically interacted with dsDNA, the data obtained by binding analysis of these mutated AtWRI1 variants corroborated the crucial roles played by these residues in the DNA binding of AtWRI1.

Table 1.

The binding affinities of wild-type (WT) AtWRI1 and its mutants to DNA as determined by ITC assay.

n.b., not binding.

AtWRI1	Kd (μM)
WT	0.48 ± 0.05
H72A	1.74 ± 0.17
E79A	0.94 ± 0.14
Q98A	4.37 ± 0.95
H175A	2.00 ± 0.79
E181A	4.62 ± 0.58
R183A	n.b.
Y192E	n.b.
S90E/I91G/Q92K	0.24 ± 0.05
H174Q/H175Q/H176K	0.16 ± 0.20
W74R	0.05 ± 0.01

Given that WRI1 has been demonstrated to play a pivotal role in plant seed oil production (, , , ) and because the wri1-1 (an AtWRI1 loss-of-function mutant) has been shown to display significant decrease in seed oil content (), we postulated that plant seed oil yield can be boosted by enhancing the binding affinity of WRI1 for its target DNA. The aforementioned residues H72, E79, Q98, H175, E181, and R183 that specifically recognized the AW-box ([CnTnG](n)7[CG]) motif were crucial for DNA binding and were maintained without changes (Fig. 2, B to D, and Table 1). In contrast, the residues that were in proximity to the phosphate group of the dsDNA in our structure were appropriate targets for mutation as they show great potential to establish additional interactions with the DNA backbone. Furthermore, these newly established interactions did not affect the DNA-binding specificity of AtWRI1. On the basis of our structural information, we selected a variety of residues for mutation to introduce additional interactions; namely, W74R, Y192E, S90E/I91G/Q92K, and H174Q/H175Q/H176K mutants were generated and tested (Table 1). Residue W74 is a surface-exposed residue that is located in the β1-β2 loop facing the major groove of DNA, and it is located in close proximity to the backbone of dsDNA at a distance of ~3.5 Å (fig. S2D). In addition, its neighboring residue R73 is linked with the dsDNA backbone phosphate through strong electrostatic interactions. Hence, mutant AtWRI1 proteins were prepared for the DNA-binding assay performed by ITC. Notably, the AtWRI1 (W74R) mutant bound to DNA with a Kd of ~0.05 μM, which was almost 10-fold higher than that of the WT AtWRI1 protein with DNA (Kd: 0.48 ± 0.05 μM; Table 1).

Functional analysis of AtWRI1 variants that displayed reduced DNA binding in planta

Recent studies have confirmed that the transient production of plant oil regulators in Nicotiana benthamiana leaves assessed by quantification analysis of the TAG level is a vigorous and rapid method to investigate the regulatory mechanism of plant oil biosynthesis (–, ). Therefore, we assessed the effects of the aforementioned mutations in the key AP2 domain residues in AtWRI1 on AtWRI1-mediated oil accumulation using the established N. benthamiana transient expression system (–). As shown previously, transient production of AtWRI1 stimulated oil biosynthesis in N. benthamiana leaves; nevertheless, four AtWRI1 variants (H72A, R183A, Y192E, and E181A) showed significantly decreased (72.5, 89.2, 89.3, and 56.6%) oil production compared to the AtWRI1 native form (Fig. 4). In addition, two AtWRI1 variants (Q98A and H175A) showed moderately reduced TAG production compared to that produced by the AtWRI1 native form, but the difference was not significant (Fig. 4). Together, the data showed that the reduced FA content in the N. benthamiana leaves after transient expression of certain AtWRI1 variants (H72A, R183A, Y192E, and E181A) suggested a correlation between the DNA-binding affinity and oil accumulation in planta.

Fig. 4.

Functional analysis of AtWRI1 and AtWRI1 variants as determined with the N. benthamiana transient expression system.

Triacylglycerol (TAG) content in N. benthamiana leaves transiently producing AtWRI1 and AtWRI1 variants, which displayed reduced DNA binding affinity. The results are shown as the means ± SE (n = 2 to 4). Empty vector (EV) was used as the control. “*” and “**” indicate significant differences (P < 0.05 and P < 0.01, respectively, Student’s t test) between AtWRI1 and AtWRI1 variants.

The AtWRI1W74R variant enhanced plant oil biosynthesis

We also compared TAG production after transient expression of AtWRI1 and multiple AtWRI1 variants that showed increased DNA-binding affinity. Notably, the TAG accumulation in the presence of the AtWRI1 (W74R) variant showed an approximately 60% increase compared to the AtWRI1 native form in N. benthamiana leaves (Fig. 5A). In contrast, two AtWRI1 variants (H174Q/H175Q/H176K and S90E/I91G/Q92K) did not notably alter TAG accumulation, in line with the marginally enhanced DNA-binding affinity of these two variants as detected by ITC (Table 1 and fig. S4).

Fig. 5.

Functional analysis of AtWRI1W74R.

(A) TAG content in N. benthamiana leaves transiently producing AtWRI1 or an AtWRI1 variant (W74R). The results are shown as the means ± SE (n = 2 to 3). “**” indicates significant differences (P < 0.01, Student’s t test) between AtWRI1 and AtWRI1W74R. (B) TAG content in the N. benthamiana leaves transiently producing diverse WRI1 orthologs and WRI1 variants (with W-to-R substitution as indicated). The results are shown as the means ± SE (n = 4). “*,” “**,” and “***” indicate significant differences (P < 0.05, P < 0.01, and P < 0.001, respectively; Student’s t test) between WRI1 and WRI1 variants. (C) Total FA content of the seeds of WT, wri1-1, and transgenic wri1-1 lines expressing AtWRI1 and AtWRI1 driven by the native promoter of AtWRI1 (proAtWRI1). T2 transgenic seeds were used in the experiments. Difference between values of seed oil content in transgenic wri1-1 lines expressing AtWRI1 and AtWRI1 was assessed by Student’s t test. “**” indicates significant difference (P < 0.01).

Sequence examination of various WRI1 orthologs indicated that W74 is conserved in diverse plant species (fig. S3). To validate the importance of the W74 residue of AtWRI1 in various plant species, we correspondingly generated the same W to R mutants with B. napus, C. sativa, G. max, and Z. mays WRI1 proteins. As shown in Fig. 5B, the W-to-R substitution in the WRI1s resulted in significantly increased TAG production in N. benthamiana leaves compared to that produced by the corresponding WRI1 native forms.

To further substantiate the effect of W74R of AtWRI1 in planta, we generated stable transgenic wri1-1 expressing AtWRI1 and AtWRI1 [driven by the native promoter of AtWRI1 (proAtWRI1)]. The WT and transgenic lines were grown in the same chamber under identical conditions. Compared to transgenic wri1-1 expressing AtWRI1, the seed oil content in multiple transgenic lines was, on average, higher for the transgenic wri1-1 expressing AtWRI1 (fig. S5). A statistical analysis revealed that the distribution of total FA content in the seeds of the transgenic wri1-1 expressing AtWRI1 and AtWRI1 was significantly different (P < 0.01, Student’s t test) (Fig. 5C). In summary, the enhanced FA content in the transiently expressed genes in N. benthamiana leaves, as well as in the seeds of stable transgenic Arabidopsis plants, confirmed that the W74 residue was vital for the function of AtWRI1, and the W74R mutation led to enhanced AtWRI1-regulated oil production in plant cells, in line with our structure and in vitro DNA-binding assays.

DISCUSSION

As an essential transcription factor governing plant oil accumulation, WRI1 has become a key candidate for bioengineering to improve vegetable oil yield. In addition, WRI1 is composed of two tandem AP2 domains that bind to a dsDNA region called the AW-box (Fig. 1), and WRI1 orthologs have been discovered in a variety of plant species (, –, ). In contrast to the acknowledged significance of WRI1, essential questions regarding how these two AP2 domains concurrently fold and how the WRI1 protein specifically recognizes its cognate DNA remain largely unanswered. Therefore, the structure of WRI1 with its DNA at high resolution was urgently needed and may be of extremely high value for biotechnological applications to improve plant oil yield.

After achieving stable complex formation of AtWRI1-DNA in vitro, we managed to crystallize the complex and determined its structure at 2.35 Å resolution using the SAD phasing method (Figs. 1 and 2). The two structurally similar AP2 domains (with anti–β sheets flanked by an α helix) were linked by a V-shaped arrangement of two α helices, which subsequently tightly clamped the first AP2 domain (Fig. 2). The two anti–β sheets in the AP2 domains projected into the adjacent major grooves of dsDNA, making extensive contact that determined the sequence recognition specificity for the AW-box ([CnTnG](n)7[CG]). In addition to interacting with both AP2 domains, which facilitated structural folding and stabilization of AtWRI1, the domain linker region was involved in DNA binding (Fig. 2). A structural comparison of AtWRI1 with the P. falciparum SIP2 protein (PDB ID: 6SY0) revealed that these proteins are quite different in structural folding and domain organization, as well as structural motifs (an anti–β sheet and α-β turn in the former and latter, respectively) involved in DNA recognition. Therefore, AtWRI1 exhibits a novel structural fold and DNA-binding mode, and the difference between these proteins implies that the widely distributed AP2/EREBP transcription factors use diverse structural folds and DNA recognition mechanisms (Fig. 3).

Our high-resolution structure clearly showed how the two tandem AP2 domains in AtWRI1 interacted with cognate dsDNA (Fig. 2). In particular, residues H72, E79, W83, and Q98, and H175, E181, and R183 in the first and second AP2 domains specifically recognized the DNA sequences 5′C6T8G10 and 5′C18G19, respectively, which are the two motifs in the AW-box ([CnTnG](n)7[CG]) (Fig. 2, B and C). On the basis of our structure, residue mutagenesis and DNA-binding assays were carried out with all the key residues, although certain constructs were found to be either unexpressed or expressed as an insoluble protein that could not be solubilized through refolding (). Nevertheless, the results corroborated the structural data showing that disruption of the aforementioned crucial residues in AtWRI1 can markedly decrease its binding affinity for DNA (Table 1). Most variants with reduced DNA-binding affinity caused a decrease in oil accumulation, as determined via transient assays with N. benthamiana leaves (Fig. 4). Notably, the sequence alignment of AtWRI1 and its orthologs in various plant species indicated that these DNA-interacting residues have been highly conserved, suggesting a common regulatory mechanism underlying the transcriptional regulation of plant oil biosynthesis (fig. S3).

Elucidation of the structure of the complex formed by the DNA-binding domain in AtWRI1 with the cis-element, AW-box, provided us with an opportunity to examine the effects of mutations in various residues in AtWRI1 by measuring the DNA-binding affinity of these variants. Notably, one AtWRI1 variant (AtWRI1W74R) showed considerably enhanced DNA-binding affinity, which was ~10-fold greater than that of the native form (Table 1). Notably, this residue, W74, was found to be highly conserved among diverse WRI1 orthologs from other plant species (fig. S3). Further functional analysis revealed that the engineered AtWRI1 variant (AtWRI1W74R) exhibited increased oil production compared with that of the AtWRI1 native form, as determined via a transient expression assay with N. benthamiana leaves (Fig. 5A). Moreover, the similarly engineered WRI1s (with the same mutation at the conserved W residue) also displayed increased plant oil production (Fig. 5B). The increase in the AtWRI1 variant (AtWRI1W74R)–mediated plant oil production was verified in the seeds of stable transgenic plants (Fig. 5C and fig. S5). The W74R mutation in AtWRI1 led to increased DNA-binding affinity, which suggests that substitution of W74 to increase AtWRI1 DNA binding is connected to enhanced AtWRI1-mediated oil production in plant cells.

In addition to the two AP2 domains involved in DNA binding, as indicated herein, both AtWRI1 termini include IDRs, with the C-terminal region IDRs comprising more than 200 residues (Fig. 1A). A prior study showed that the C terminus of AtWRI1 includes a transactivation domain and a PEST motif that mediates AtWRI1 protein degradation (). AtWRI1 variants, with either a deleted IDR3-PEST motif or phosphorylation-deficient mutations within the IDR3-PEST motif, showed improved protein stability and higher oil production than the AtWRI1 native form (). Hence, phosphorylation is postulated to mediate the stability and activity of AtWRI1 (). In addition, recent advances have led to the identification of a variety of AtWRI1-interacting partners, which showed physical interaction with AtWRI1 and modulation of AtWRI1 activity. For instance, 14-3-3 proteins (a family of phosphopeptide-binding proteins) can physically interact with AtWRI1 in yeast and plant cells (). Transient coproduction of 14-3-3 with AtWRI1 leads to enhanced oil production in N. benthamiana leaves by augmenting the protein stability of AtWRI1 (). KIN10 (SNF1-related protein kinase) physically interacts with AtWRI1 and triggers AtWRI1 phosphorylation, which results in AtWRI1 degradation (). These studies suggest that molecular actions involving posttranslational modifications (e.g., phosphorylation), in association with other interacting regulators, might synergistically modulate the stability and activity of WRI1, thus fine-tuning WRI1-mediated oil biosynthesis. In these regulatory modules, the effect of the key residue modifications on the DNA-binding domain is unclear; is the DNA-binding ability of WRI1 affected or is the protein complex assembly involving WRI1 and its interacting regulators altered? Thus, these fascinating questions are worth exploring in the future, and the structural information provided in this study can potentially shed new light on the answers to these intriguing questions.

In conclusion, our work provides biochemical, structural, and genetic evidence to elucidate the molecular mechanism underlying WRI1 binding to its cis-element, the AW-box. Vegetable oils exhibit a broad range of applications, from being vital in the human diets to serving as renewable feed stocks in industry. The knowledge gained from this study has valuable implications, because it offers an alternative strategy for boosting oil content in plant cells, which is an essential objective to meet the increasing demand for vegetable oil globally.

MATERIALS AND METHODS

Plasmid construction

The coding region of AtWRI158–307 was codon-optimized, synthesized, and subcloned into the pOPT–maltose-binding protein (MBP) vector encoding a hexahistidine-MBP tag. The mutagenesis was performed using a site-directed mutagenesis kit (New England Biolabs). All insertions and mutagenesis were confirmed by sequencing. For plant expression constructs, the coding sequence of WRI1 genes (the native form and variants) was amplified by polymerase chain reaction and cloned into pENTR4 (Thermo Fisher Scientific) or synthesized into a pTwist ENTR by Twist Bioscience to obtain entry constructs. The WRI1 entry constructs were introduced into pEarleyGate binary vectors (pEarleyGate100) () by LR Clonase II (Thermo Fisher Scientific). To obtain the proAtWRI1:AtWRI1 construct, the CaMV 35S promoter was replaced by proAtWRI1 (2 kb upstream of the start codon). Table S2 provides a list of the primers used for plasmid construction in this study.

Protein expression and purification

The plasmid encoding AtWRI158–307 was transformed into E. coli BL21 (DE3). A single colony was picked and inoculated into 10 ml of 2xYT medium (tryptone, 16 g/liter; yeast extract, 10 g/liter; and NaCl, 5 g/liter) supplemented with a final concentration of ampicillin (100 μg/ml) and cultured overnight at 37°C. The bacteria were then inoculated into 1 liter of 2x YT medium and grown until the OD600 (optical density at 600 nm) reached ~0.6. The cells were then cooled in a cold room for 1 hour before isopropyl-β-d-thiogalactopyranoside was added to a final concentration of 0.2 mM and cultured overnight at 16°C to induce protein expression. The bacteria were harvested by centrifugation (Beckman Coulter, Avanti-26, JA8.1, 4000 rpm) for 10 min at 4°C and then resuspended in lysis buffer [25 mM tris (pH 7.5) and 150 mM NaCl]. The cells were lysed by sonication, and the crude lysate was clarified by centrifugation (Beckman Coulter, Avanti-26, JA25.5, 20,000 rpm) for 1 hour at 4°C. The supernatant was incubated with 5 ml of nickel–nitrilotriacetic acid beads (Bio Basic, Singapore) for 1 hour at 4°C. Then, the beads were collected by centrifugation and washed with buffer A [25 mM tris (pH 7.5), 150 mM NaCl, and 30 mM imidazole]. The protein was eluted by buffer B [25 mM tris (pH 7.5), 150 mM NaCl, and 200 mM imidazole]. The eluted protein was incubated with tobacco etch virus protease overnight at 4°C to cleave the hexahistidine-MBP tag. The cleaved protein was separated by anion ion exchange, collected, concentrated, and loaded for size exclusion chromatography on a HiLoad Superdex 200 16/60 preequilibrated with buffer [25 mM tris (pH 7.5) and 150 mM NaCl). The fractions eluted from the size exclusion chromatography column were collected, concentrated to 20 mg/ml, snap-frozen, and stored at −80°C for future use. The selenomethionine (Se-Met)–substituted WRI1 and the WRI1 mutants were purified similarly.

Analytical gel filtration chromatography

Analytical gel filtration chromatography was performed using Superdex 200 Increase 10/300 GL (GE Healthcare) in buffer [25 mM tris (pH 7.5) and 150 mM NaCl] for the following samples: WRI1, WRI1 and dsDNA in a 1:1 ratio, and dsDNA. WRI1 and its binding dsDNA were mixed and incubated on ice for 1 hour before loading.

Isothermal titration calorimetry

ITC measurements were performed using a MicroCal PEAQ-ITC (Malvern Instruments) at 20°C. Two hundred micromolar WRI1 binding dsDNA (5′-TACTTCCTCGGTTTCATCGTCCAC) was injected into the sample cell containing 20 μM WRI1 or its mutants in a total of 19 injections. The injection time was 4 s, with each 2-μl volume injected in 150-s intervals. DNA titration with buffer was performed following the same procedures. The raw data were processed and analyzed with MicroCal PEAQ-ITC analysis software.

Crystallization and structure determination

AtWRI1 (7.5 mg/ml) was mixed with equimolar AtWRI1-binding dsDNA BCCP2(-34/-11) (5′-TACTTCCTCGGTTTCATCGTCCAC) and incubated at 4°C for 1 hour. Crystallization screening was performed with Crystal Screen, Index, PEG Rx, and PEG/Ion Screen (Hampton Research) using the sitting drop vapor diffusion method at 20°C on an Intelli-Plate 96-3 LVR (Art Robbins Instruments). Crystals appeared during crystal screening under the A9 condition [0.2 M ammonium acetate, 0.1 M sodium citrate tribasic dihydrate (pH 5.6), and 30% w/v polyethylene glycol 4000]. The final optimized condition for crystal growth was 0.2 M ammonium acetate, 0.1 M MES (pH 6.5), and 28% w/v PEG 4000.

The native dataset of AtWRI1-DNA crystals was collected from the Swiss Light Source (SLS). SAD datasets for Se-Met–labeled AtWRI1-DNA crystals were collected at an inflection wavelength of 0.9795 Å with an Australian light source and SLS. All the datasets were processed using XDS (x-ray detector software) (). The Phenix AutoSol program was used for phasing, and four Se atoms were found in the substructure solution (figure of merit, 0.23). The structure served as a template to resolve native AtWRI1-DNA structures through molecular replacement. The models were built and refined using Phenix and Coot (, ). All the structure-related figures were generated with PyMOL ().

Plant materials

Arabidopsis and N. benthamiana plants were grown in a growth chamber at 23°C with a photoperiod of 16 hours of light (100 to 150 μmol m−2 s−1 illumination)/8 hours of dark. Arabidopsis WT (Columbia ecotype) and wri1-1 () were used in this study. Arabidopsis transformation, seed sterilization, and germination were performed followed methods described previously ().

Transient expression in N. benthamiana

Agrobacterium tumefaciens–mediated transient expression in N. benthamiana leaves was conducted as previously described ().

FA analysis

FA analysis of the Arabidopsis seeds and N. benthamiana leaves was performed as previously described (, ).