Metabolomics and Transcriptomics Provide Insights into Lipid Biosynthesis in the Embryos of Walnut (Juglans regia L.)

Walnut (Juglans regia L.) is an important woody oilseed tree species due to its commercial value. However, the regulation mechanism of walnut oil accumulation is still poorly understood, which restricted the breeding and genetic improvement of high-quality oil-bearing walnuts. In order to explore the metabolic mechanism that regulates the synthesis of walnut oil, we used transcriptome sequencing technology and metabolome technology to comprehensively analyze the key genes and metabolites involved in oil synthesis of the walnut embryo at 60, 90, and 120 days after pollination (DAP). The results showed that the oil and protein contents increased gradually during fruit development, comprising 69.61% and 18.32% of the fruit, respectively, during ripening. Conversely, the contents of soluble sugar and starch decreased gradually during fruit development, comprising 2.14% and 0.84%, respectively, during ripening. Transcriptome sequencing generated 40,631 unigenes across 9 cDNA libraries. We identified 51 and 25 candidate unigenes related to the biosynthesis of fatty acid and the biosynthesis of triacylglycerol (TAG), respectively. The expression levels of the genes encoding Acetyl-CoA carboxylase (ACCase), long-chain acyl-CoA synthetases (LACS), 3-oxoacyl-ACP synthase II (KASII), and glycerol-3-phosphate acyl transfer (GPAT) were upregulated at 60 DAP relative to the levels at 90 and 120 DAP, while the stearoyl-ACP-desaturase (SAD) and fatty acid desaturase 2 (FAD2) genes were highly abundantly expressed during all walnut developmental periods. We found that ABSCISIC ACID INSENSEITIVE3 (ABI3), WRINKLEDl (WRI1), LEAFY COTYLEDON1 (LEC1), and FUSCA3 (FUS3) may be key transcription factors involved in lipid synthesis. Additionally, the metabolomics analysis detected 706 metabolites derived from 18 samples, among which, 4 are implicated in the TAG synthesis, 2 in the glycolysis pathway, and 5 in the tricarboxylic acid cycle (TCA cycle) pathway. The combined analysis of the related genes and metabolites in TAG synthesis showed that phospholipid:diacylglycerol acyltransferase (PDAT) genes were highly abundantly expressed across walnut fruit developmental periods, and their downstream metabolite TAG gradually accumulated with the progression of fruit development. The FAD2 gene showed consistently higher expression during fruit development, and its downstream metabolites 18:2-PC and 18:3-PC gradually accumulated. The ACCase, LACS, SAD, FAD2, and PDAT genes may be crucial genes required for walnut oil synthesis. Our data will enrich public databases and provide new insights into functional genes related to lipid metabolism in walnut.


Introduction
Walnut (Juglans regia L.) belongs to the Juglandaceae family, which is an important and widely cultivated woody oilseed tree species [1]. As a functional oil, walnut oil exerts various effects, including an overall health-promoting effect on the human body [2][3][4]. Its many health benefits include improving cognitive function and intelligence, lowering

Analysis of Walnut Fruit Morphological Characteristics and Oil Content
The morphological characteristics of the walnut fruits are shown in Figure 1A. We observed that the fruit diameter gradually increased at 50-70 DAP, thereafter remaining largely unchanged at 70-140 DAP. The embryo was just visible to the naked eye at 50 DAP. The endocarp had not yet formed at 50 DAP. The embryo began to solidify and gradually mature at 60-140 DAP. The dynamic changes in the oil and protein content during the two walnut oil accumulation periods (60-90 and 90-140 DAP) are shown in Figure 1B. The oil content increased significantly from 2.61% to 69.61% during the walnut oil accumulation periods. The oil content increased rapidly at 60-90 DAP and then increased slightly at 90-140 DAP. The protein content increased significantly from 10.25% to 21.12% at 50-90 DAP, while the protein content decreased gradually from 21.12% to 18.32% at 90-140 DAP. The dynamic changes in the soluble sugar and starch content during the walnut oil accumulation periods are shown in Figure 1C. During the period of rapid oil accumulation (60-90 DAP), the soluble sugar and starch content decreased rapidly, then reached a plateau of 90-140 DAP. The contents of soluble sugar and starch reached their highest levels at 50 DAP (18.62% and 13.64%, respectively) and lowest levels at 140 DAP (2.14% and 0.84%, respectively).

Analysis of the Principal Component Analysis (PCA)
According to the embryo morphology development and the oil content, walnut fruits at 60, 90, and 120 DAP were chosen for transcriptomics (each stage had three biological replicates) and metabolomics (each stage had six biological replicates) analysis, designated as L1, L2, and L3, respectively. PCA was conducted to reflect the characteristics of multi-dimensional omics data through several principal components. As shown in Figure  2A, the PCA of all 9 samples was conducted using RNA-seq FPKM. The PC1 and PC2

Analysis of the Principal Component Analysis (PCA)
According to the embryo morphology development and the oil content, walnut fruits at 60, 90, and 120 DAP were chosen for transcriptomics (each stage had three biological replicates) and metabolomics (each stage had six biological replicates) analysis, designated as L1, L2, and L3, respectively. PCA was conducted to reflect the characteristics of multidimensional omics data through several principal components. As shown in Figure 2A, the PCA of all 9 samples was conducted using RNA-seq FPKM. The PC1 and PC2 explained 98.09% and 1.47% of the variation, respectively. Furthermore, L1 was clearly distinguishable from the other two groups (L2 and L3) in the PC1 score plots, while L3 was clearly distinguishable from the other two groups (L1 and L2) in the PC2 score plots. As shown in Figure 2B, the PCA of all 18 samples was conducted based on the peak area. The PC1 and PC2 explained 70.89% and 20.4% of the variation, respectively. Furthermore, L1 was clearly distinguishable from the other two groups (L2 and L3) in PC1 and PC2 score plots. The repeated samples were compactly distributed, thus indicating that there was a high consistency between the biological replicates and that the experiment was reproducible and reliable.
Plants 2023, 12, x FOR PEER REVIEW 5 The PC1 and PC2 explained 70.89% and 20.4% of the variation, respectively. Furtherm L1 was clearly distinguishable from the other two groups (L2 and L3) in PC1 and score plots. The repeated samples were compactly distributed, thus indicating that was a high consistency between the biological replicates and that the experiment w producible and reliable.

Analysis of Differentially Expressed Genes (DEGs) across Different Developmental Per
To understand the potential molecular synthesis mechanisms involved in waln fruit development and ripening, nine libraries from walnut embryos were sequence total of 444,613,638 raw reads were obtained and 61.52 G clean bases were obtained quality control (Table S1). All of the clean reads were mapped to the walnut genom quence, and the proportion of the total map rate ranged from 94.77% to 96.87% (Tabl A cluster analysis of the DEGs based on FPKM is presented in Figure 3A. Nine ples were divided into two main groups: samples from the first stage (60 DAP) consti one group; samples from the last stage (120 DAP) formed the second group. These groups correspond to the beginning of oil accumulation and the maturity of oil accu lation. The result showed that there were more highly expressed genes at the begin of oil accumulation. The statistical results of the DEGs are shown in Figure 3B. The ber of DEGs in the L2 vs. L1, L3 vs. L1, and L3 vs. L2 comparison groups were 9978, and 3932, respectively; the corresponding number of those upregulated were 352, 965 3050, while those downregulated were 9626, 8393, and 882, respectively. The Venn gram in Figure 3C shows the overlap of DEGs among the comparison groups. There 897 common DEGs among all comparisons.

Analysis of Differentially Expressed Genes (DEGs) across Different Developmental Periods
To understand the potential molecular synthesis mechanisms involved in walnut oil fruit development and ripening, nine libraries from walnut embryos were sequenced. A total of 444,613,638 raw reads were obtained and 61.52 G clean bases were obtained after quality control (Table S1). All of the clean reads were mapped to the walnut genome sequence, and the proportion of the total map rate ranged from 94.77% to 96.87% (Table S2).
A cluster analysis of the DEGs based on FPKM is presented in Figure 3A. Nine samples were divided into two main groups: samples from the first stage (60 DAP) constituted one group; samples from the last stage (120 DAP) formed the second group. These two groups correspond to the beginning of oil accumulation and the maturity of oil accumulation. The result showed that there were more highly expressed genes at the beginning of oil accumulation. The statistical results of the DEGs are shown in Figure 3B. The number of DEGs in the L2 vs. L1, L3 vs. L1, and L3 vs. L2 comparison groups were 9978, 9358, and 3932, respectively; the corresponding number of those upregulated were 352, 965, and 3050, while those downregulated were 9626, 8393, and 882, respectively. The Venn diagram in Figure 3C shows the overlap of DEGs among the comparison groups. There were 897 common DEGs among all comparisons.

Analysis of Differentially Expressed Metabolites (DEMs) across Different Developmenta Periods
Multiomics analysis can more clearly identify the genes regulating walnut oil me olism, it was used in the present study. The samples collected from the walnut embr during oil accumulation periods (60, 90, and 120 DAP) were subjected to LC-MS analy The total ion chromatogram (TIC) of all the samples showed high stability, large p capacity, and good retention time ( Figure S1).
The cluster analysis of DEMs based on the peak area is shown in Figure 4A. All logical replicates clustered together, indicating the high reliability of the resulting me olomic data. Among them, most metabolites showed high levels of accumulation in early stage of fruit development (60 DAP), while some exhibited high levels of accum tion at the maturity stage (120 DAP). This indicates that the accumulation of these me olites was related to the development and maturation of walnut fruits. The statistica sults of DEMs are shown in Figure 4B. The number of DEMs in the L2 vs. L1, L3 vs. and L3 vs. L2 comparison groups were 316, 319, and 225, respectively; the correspond number of those upregulated were 156, 131, and 74, while those downregulated were 188, and 151, respectively. The Venn diagram analysis of the three comparison group shown in Figure 4C. There were 65 common DEMs in all comparisons.

Analysis of Differentially Expressed Metabolites (DEMs) across Different Developmental Periods
Multiomics analysis can more clearly identify the genes regulating walnut oil metabolism, it was used in the present study. The samples collected from the walnut embryos during oil accumulation periods (60, 90, and 120 DAP) were subjected to LC-MS analysis. The total ion chromatogram (TIC) of all the samples showed high stability, large peak capacity, and good retention time ( Figure S1).
The cluster analysis of DEMs based on the peak area is shown in Figure 4A. All biological replicates clustered together, indicating the high reliability of the resulting metabolomic data. Among them, most metabolites showed high levels of accumulation in the early stage of fruit development (60 DAP), while some exhibited high levels of accumulation at the maturity stage (120 DAP). This indicates that the accumulation of these metabolites was related to the development and maturation of walnut fruits. The statistical results of DEMs are shown in Figure 4B. The number of DEMs in the L2 vs. L1, L3 vs. L1, and L3 vs. L2 comparison groups were 316, 319, and 225, respectively; the corresponding number of those upregulated were 156, 131, and 74, while those downregulated were 160, 188, and 151, respectively. The Venn diagram analysis of the three comparison groups is shown in Figure 4C. There were 65 common DEMs in all comparisons.

Analysis of Transcription Factors
Transcription factors (TFs) are important proteins that act as upstream regulators of genes and play key roles in the regulation of plant development. A total of 384 differentially expressed TFs belonging to 27 families were identified across the walnut fruit development stages ( Figure 5A, Table S3). The classified results indicate that most of these TFs belonged to the bHLH (83 members), ERF (66 members), G2-like (30 members), and bZIP (28 members) families. As shown in Figure 5B, the TF family members were either significantly upregulated or downregulated to varying degrees.

Analysis of Transcription Factors
Transcription factors (TFs) are important proteins that act as upstream regulator genes and play key roles in the regulation of plant development. A total of 384 diffe tially expressed TFs belonging to 27 families were identified across the walnut fruit velopment stages ( Figure 5A, Table S3). The classified results indicate that most of th TFs belonged to the bHLH (83 members), ERF (66 members), G2-like (30 members), bZIP (28 members) families. As shown in Figure 5B, the TF family members were ei significantly upregulated or downregulated to varying degrees.

Analysis of Transcription Factors
Transcription factors (TFs) are important proteins that act as upstream regulato genes and play key roles in the regulation of plant development. A total of 384 diffe tially expressed TFs belonging to 27 families were identified across the walnut frui velopment stages ( Figure 5A, Table S3). The classified results indicate that most of t TFs belonged to the bHLH (83 members), ERF (66 members), G2-like (30 members), bZIP (28 members) families. As shown in Figure 5B, the TF family members were e significantly upregulated or downregulated to varying degrees.

Integrated Analysis of the Transcriptomic and Metabolomic Data Pertaining to Lipid Biosynthesis across Walnut Fruit Developmental Periods
To understand the differences in the oil accumulation across fruit developmental periods, transcriptomic and metabolomic data were integrated for analysis ( Figure 6, Table S4). ACCase consists of three subunits, including biotin carboxylase (BC), biotin carboxyl carrier protein (BCCP), and carboxyl transferase subunit alpha (CTα) [13]. Our analysis showed that the expression levels of the genes encoding three subunits (BC, BCCP, and CTα) were upregulated at 60 DAP relative to their levels at 90 and 120 DAP. This may partially explain why oil content increased from 12.38% (60 DAP) to 69.14% (120 DAP). There were three kinds of 3-oxoacyl-ACP synthase (KAS), including KASI, KASII, and KASIII, which can each catalyze the synthesis of different carbon chains in plastids. KASIII catalyzes acetyl-CoA and malonyl-ACP to form 4:0-ACP; KASI catalyzes the synthesis from 4:0-ACP to 16:0-ACP; KASII catalyzes the synthesis from 16:0-ACP to 18:0-ACP [32]. Our analysis revealed that the expression levels of the genes encoding KASI, KASII, and KASIII were upregulated at 60 DAP relative to their levels at 90 and 120 DAP. However, one of the five KASII-encoding genes was more highly expressed at 90 DAP than at 60 and 120 DAP. These expression patterns provided an explanation for the continuous and rapid oil accumulation during early fruit development (60 DAP).
SAD is a key enzyme that catalyzes the transformation of saturated fatty acids to unsaturated fatty acids and can catalyze the transformation of 18:0-ACP to 18:1-ACP [33]. Three SAD genes were detected during walnut oil accumulation, one of which had the highest expression level at 90 DAP (with an FPKM value of 580.05). The highly expressed SAD may rapidly catalyzed the synthesis of unsaturated fatty acids. FAT consists of two types, including FATA (fatty acyl-ACP thioesterase A) and FATB (fatty acyl-ACP thioesterase B) [23]. FAT has a large influence on the composition of fatty acids, the specificity of their carbon chain lengths, and their activity [34]. FATA has the highest activity for C18:1-ACP, and FATB mainly hydrolyzes the thiolipid bond of the saturated fatty ACP (C16:0-ACP and C18:0-ACP) [35], thereby releasing free fatty acids and ACP. The expression levels of the genes encoding FATA and FATB were upregulated at 60 DAP relative to their levels at 90 and 120 DAP. FATA was expressed at substantially higher levels than FATB, which is expected to have been beneficial to the accumulation of unsaturated fatty acids. LACS plays an important role in lipid synthesis. In this study, the LACS genes were found to be downregulated during walnut oil accumulation periods and had the highest expression levels at 60 DAP.
The FAD is localized to the endoplasmic reticulum, and can convert monounsaturated fatty acids to polyunsaturated fatty acids [36]. Two FAD2 were detected in this study; one was downregulated during walnut oil accumulation, while the other was characterized by consistently higher expression levels during walnut oil accumulation; it had the highest expression level at 90 DAP (with an FPKM value of 1877). The contents of the downstream metabolites 18:2-PC and 18:3-PC were more highly accumulated at 120 DAP compared to 60 and 90 DAP. This may be related to the persistently high expression of FAD2 gene. Previous studies showed that the high expression of PDAT was beneficial to the accumulation of polyunsaturated fatty acids [37,38]. Seven PDAT-encoding genes and one DGAT1-encoding gene were detected during walnut oil accumulation. Among them, six PDAT genes showed a downregulated expression pattern during walnut oil accumulation, while notably, the other PDAT gene showed the highest expression level at 90 DAP. The DGAT1 gene exhibited a low expression level at all developmental stages. The content of the downstream metabolite TAG increased gradually over with the progression of fruit development. Most of the genes showed higher expression levels in the early stage of oil accumulation and lower expression levels in the middle and late stages of oil accumulation, which corresponded to the oil accumulation rate. These crucial genes have similar expression profiles and may play a critical joint role in lipid formation.

Identification of TFs Related to Oil Biosynthesis
LEC1, LEC2, ABI3, and FUS3 play key roles in seed development and lipid synthesis [23]. Figure 7 shows that the expression level of LEC1, FUS3, ABI3, and WRI1 were more highly expressed at 60 DAP compared to 90 and 120 DAP. It is worth noting that the FPKM values of LEC2 were lower than 0.37 across the three stages. The average FPKM value of LEC1 was 439.59 at 60 DAP, which was much higher than that of the other transcription factors. We speculate that LEC2 plays a minor role in walnut oil accumulation, while LEC1 plays a more critical role.
The correlation analysis between lipid synthesis-related genes and key transcription factors is shown in Figure 8. Carbohydrate anabolism comprises much of the basic metabolism of plants, providing carbon sources and energy inputs for lipid synthesis. Pyruvate produced by glycolysis produces acetyl-CoA under the action of PDH, which is the raw material for fatty acid synthesis [39]. To understand the differences in the glycolysis pathway across fruit developmental periods, transcriptomic and metabolomic data were integrated for analysis ( Figure 9, Table S5). The expression levels of the genes encoding pyruvate decarboxylase (PDC), alcohol dehydrogenase (ADH), and aldehyde dehydrogenase (ALDH) were found to be downregulated during walnut oil accumulation. The content of the downstream metabolite acetaldehyde decreased with progressing fruit development, which suggests that these genes may directly affect acetaldehyde content. As the expression of the gene encoding pyruvate kinase (PK) decreased, so too did the content of its downstream metabolite, pyruvic acid. The upstream gene expression profile was consistent with the changes observed in the downstream metabolite content.

Analysis of the TCA Cycle Pathway
The TCA cycle represents the center of sugar and lipid metabolism. To understand the differences in the TCA cycle across fruit developmental periods, transcriptomic and metabolomic data were integrated for analysis ( Figure 10, Table S6). PDH catalyzes the conversion of pyruvate to acetyl-CoA and is composed of four subunits (E1α, E1β, E2, and E3) [39]. In this study, PDH-E1α and PDH-E1β were detected, and the expression levels of these genes were increasingly downregulated with continued fruit development and ripening. The expression levels of the genes coding ATP-citrate synthase (ACLY), aconitate hydratase (ACO), malate dehydrogenase (MDH), fumarate hydratase (FH), succinyl-CoA synthetase alpha subunit (LSC1), succinyl-CoA synthetase beta subunit (LSC2), succinate dehydrogenase (ubiquinone) iron-sulfur subunit (SDHB), succinate dehydrogenase (ubiquinone) flavoprotein subunit (SDHA), isocitrate dehydrogenase (NADP) (IDH1), and isocitrate dehydrogenase (NAD+) (IDH3) decreased gradually, and the contents of the downstream metabolites citric acid, s-malate, succinic acid, and oxoglutaric acid were found to decrease gradually with the progression of fruit development. The changes in the upstream gene expression and downstream metabolite content were consistent. accumulation and lower expression levels in the middle and late stages of oil accumulation, which corresponded to the oil accumulation rate. These crucial genes have similar expression profiles and may play a critical joint role in lipid formation.  values of LEC2 were lower than 0.37 across the three stages. LEC1 was 439.59 at 60 DAP, which was much higher than th factors. We speculate that LEC2 plays a minor role in walnut o plays a more critical role.

Analysis of the Glycolysis Pathway
Carbohydrate anabolism comprises much of the basic metabolism of plants, p ing carbon sources and energy inputs for lipid synthesis. Pyruvate produced by gly produces acetyl-CoA under the action of PDH, which is the raw material for fat synthesis [39]. To understand the differences in the glycolysis pathway across frui

RT-qPCR Validation of DEGs in Transcriptomic Data
In order to validate the transcriptomic data, 12 DEGs were selected for expression analysis by using qRT-PCR. Their expression levels were consistent between the qRT-PCR and RNA-seq data, thereby validating the reliability of the transcriptomic data ( Figure 11).

Discussion
In this study, as the walnut fruits developed and matured, their oil and protein contents increased rapidly at 50-90 DAP and remained relatively stable at 90-140 DAP. On the contrary, their soluble sugar content decreased rapidly at 50-90 DAP and remained relatively stable at 90-140 DAP. This may be attributable to the intermediate products of sugar metabolism providing the necessary raw materials and energy required for lipid synthesis, whereby the sugar content was gradually consumed as the oil content gradually accumulated. According to the changes in the physiological indicators during fruit development, we found that 50-90 DAP was the most critical period for oil synthesis. In this study, enzyme genes and metabolites involved in oil synthesis were detected by transcriptomic and untargeted metabolomic analysis during walnut oil accumulation.
In this study, 9 cDNA libraries for transcriptome sequencing of walnut embryos at three developmental stages were constructed, and more than 94% of clean reads in each library mapped to the walnut reference genome. Transcriptome analysis showed that the oil accumulation rate in walnut embryos was related to the specific expression of lipid biosynthesis genes. Previous studies on "Linzaoxiang" [23] and "Qingxiang" [40] walnut embryos indicated that KASIII, KASI, FATA, FATB, SAD, and FAD2 may play a critical role in lipid synthesis. These genes showed similar expression trends in different walnut varieties. In this study, by analyzing the expression of key enzyme genes involved in oil biosynthesis, it was found that the genes LACS and ACCase had higher transcript abundance in the oil accumulation periods, and so may play a role in accelerating the biosynthesis of fatty acids. The extension of fatty acid carbon chains, pre-desaturation of fatty acids, transport of fatty acids, and generation of other lipid derivatives from fatty acids all depend on LACS [41]. In this study, the persistently high expression of FAD2 could accelerate the accumulation of 18:2-PC and 18:3-PC metabolites. We speculated that the FAD2 gene could also accelerate the accumulation of oil by the TAG synthesis pathway related to the PDAT gene.
DGAT and PDAT represent the last step in the catalytic synthesis of TAG, which may be the key step in the synthesis of lipids [42]. The expression level of PDAT in walnut embryos is much higher than that of DGAT1 and DGAT2 [23]. In Torreya grandis kernels, PDAT showed a higher correlation with oil content than DGAT [43]. The loss of DGAT1 protein function can lead to a change in the composition of fatty acids [44], while its overexpression in Arabidopsis has been shown to cause seeds to become larger and have an increased oil content [45]. PDAT is another pathway involved in TAG synthesis, which codes for a key enzyme in the TAG synthesis pathway independent of acyl-CoA [46]. Further, in this study, the expression level of PDAT was much higher than that of DGAT1 during walnut oil accumulation. The high expression of PDAT may have more influence on the accumulation of walnut oil.
In addition, GPAT and LPAAT are also key rate-limiting enzyme genes in the TAG synthesis, and previous studies have shown that the overexpression of the yeast LPAAT gene in Arabidopsis thaliana can promote the accumulation of oil [47]. In this study, the expression pattern of these two genes was downregulated with fruit development and ripening. Therefore, the GPAT and LPAAT genes might increase oil content by increasing the flow rate of intermediates in the Kennedy pathway. WRI1 and FUS3 may be crucial in the regulation of lipid biosynthesis in Torreya grandis kernel [43]. The expression pattern of WRI1 aligned with the accumulation of walnut oil [23]. The expression patterns of TFs; such as LEC1, FUS3, ABI3, and WRI1; were consistent with the pattern of walnut oil accumulation, thus, these TFs may play a critical role in walnut lipid synthesis. In this study, these genes (Accase, FATA, TATB, KASI, KASIII, GPAT, PDAT, LEC1, and WRI1) showed similar expression trends during walnut oil accumulation, and these genes may be related to the development of the walnut embryo.
Sugar metabolism is highly important for the synthesis of plant lipids, and a series of intermediate products of sugar metabolism can provide the necessary raw materials and energy for lipid synthesis during fruit development [48]. The expression of key genes in glucose metabolism may affect fruit oil accumulation [49]. The genes ENO and PK were highly abundantly expressed and are involved in the glycolytic pathway. The upstream ENO and PK expression profiles were consistent with the changes observed in the content of downstream metabolite pyruvic acid. In the TCA cycle pathway, the expression of the genes PDH were the highest at 60 DAP; their products function to rapidly convert pyruvate to acetyl-CoA and provide sufficient raw materials for fatty acid synthesis. The molecular mechanism of walnut lipid synthesis was systematically studied by analyzing the regulatory relationship between key gene expression profiles and metabolite content during the walnut oil accumulation stage. The qPCR analysis confirmed that the expression data obtained by RNA-Seq was reliable. The findings provide a theoretical reference for the improved breeding of high-lipid varieties of walnut or other means of walnut lipid synthesis. Nevertheless, the regulating mechanism of oil accumulation in developing embryos is still unclear. Further studies on their functions in FA biosynthesis and oil accumulation are required.

Plant Materials
Walnut fruits were collected from a 19-year-old walnut cultivar "Lvling" trees from Hebei Lvling Fruit Industry Co., Ltd. (114 • 30 -114 • 33 E, 37 • 29 -37 • 32 N, elevation 90-135 m), which resides 6 km north of Lincheng County, Hebei Province, China. Walnut fruits with a similar growth state that was free of pests and diseases were selected for sampling. The trees were artificially pollinated on April 10 (0 DAP). Fruit samples were collected during the embryo development stage from June 1 to August 29, 2020. The fruits were harvested at 50 DAP (i.e., when the embryos were only just visible) and then every 10 days until 140 DAP (i.e., at embryo maturity). Twenty mixed fruits obtained from three trees were regarded as one biological replicate. Three biological replicates were collected for transcriptome and physiological indices analysis, and six biological replicates were collected for untargeted metabolome analysis. Walnut fruits at 50, 60, 70, 80, 90, 100, 110, 120, 130, and 140 DAP were selected for the analysis of physiological indices. Walnut fruits at 60, 90, and 120 DAP were selected as transcriptome and metabolome test materials, which were correspondingly labeled as L1, L2, and L3. After quickly removing the peel and seed coat, embryos were immediately frozen in liquid nitrogen and transferred to a −80 • C freezer for storage. The remaining fruits, which were used for the analysis of physiological indices, were placed in an oven and dried at 40 • C until their weight stabilized. Meanwhile, the fruit morphology was recorded using a camera.

Determination of Physiological Indices during Fruit Development
From the walnut embryos, the modified Soxhlet extraction method was used to determine their oil content [50]; the anthrone-sulfuric acid method was used to determine their soluble sugar and starch contents [51]; the Kjeldahl method was used to determine their protein content [52].

RNA Extraction and Library Construction
Total RNA was isolated and purified using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). The amount and purity of the RNA isolates from each sample were quantified using the NanoDrop ND-1000 (NanoDrop, Wilmington, DE, USA). The RNA integrity was assessed by the Bioanalyzer 2100 (Agilent, CA, USA), and was then confirmed by electrophoresis with denaturing agarose gel. After total RNA was extracted, mRNA was purified from total RNA using Dynabeads Oligo (dT). Then the cleaved RNA fragments were reverse-transcribed to create the cDNA by SuperScript™ II Reverse Transcriptase (Invitrogen, cat. 1896649, USA), which were next used to synthesize U-labeled secondstranded DNAs. Finally, a total of 9 libraries were constructed from three stages (each stage had three biological replicates). The average insert size for the final cDNA libraries was 300 ± 50 bp. We performed 2 × 150 bp paired-end sequencing (PE150) on an Illumina Novaseq™ 6000 (LC-Bio Technology Co., Ltd., Hangzhou, China).

Quantitative Analysis
Approximately 1 µg of the total RNA obtained from each sample was used to synthesize cDNA using HiFiScript gDNA Removal RT MasterMix for qPCR (CWBIO, Beijing, China). The qPCRs were performed using MagicSYBR Mixture (CWBIO, Beijing, China). Table S7 shows the gene-specific primers for 12 genes related to lipid biosynthesis and internal control primers; the walnut GAPDH gene was used as the reference gene. The relative gene expression levels were calculated by the 2 −∆∆CT method [56]. Three biological replicates were performed.

Metabolite Extraction
The collected samples were thawed on ice and 100 mg of powder was weighed from each sample using 1 mL of precooled 50% methanol. Then, the mixture of metabolites was vortexed for 1 min and incubated for 10 min at room temperature before being stored at −20 • C overnight. The next day, the mixture was centrifugated at 4000× g for 20 min. The samples were stored at −80 • C prior to the LC-MS analysis. Pooled quality control (QC) samples were also prepared by combining 10 µL of each extraction mixture [57]. Six biological replicates were executed for the extraction and subsequent analysis process.

LC-MS Analysis
All samples were analyzed using a TripleTOF 5600 Plus high-resolution tandem mass spectrometer (SCIEX, Warrington, UK) with both positive and negative ion modes. Chromatographic separation was performed using an ultra-performance liquid chromatography (UPLC) system (SCIEX, UK). An ACQUITY UPLC T3 column (100 mm × 2.1 mm, 1.8 µm; Waters, UK) was used for the reverse-phase separation. For the separation of metabolites, the mobile phase consisted of solvent A (water containing 0.1% formic acid) and solvent B (acetonitrile containing 0.1% formic acid). The gradient elution conditions were as follows: a flow rate of 0.4 mL/min: 5% solvent B for 0-0.5 min; 5-100% solvent B for 0.5-7 min; 100% solvent B for 7-8 min; 100-5% solvent B for 8-8.1 min; and 5% solvent B for 8.1-10 min. The column temperature was maintained at 35 • C.
The TripleTOF 5600 Plus system was used to detect metabolites eluted from the column. The curtain gas pressure was set at 30 PSI, while the ion source gas1 and gas2 pressure were set at 60 PSI. The interface heater temperature was 650 • C. For the positive-ion mode, the ion spray floating voltage was set at 5 kV; for the negative-ion mode, −4.5 kV. The MS data were acquired in the IDA mode. The TOF mass range was 60-1200 Da. Survey scans were acquired every 150 ms, and as many as 12 product ion scans were collected if the threshold of 100 counts/s was exceeded with a 1 + charge state. The total cycle time was fixed at 0.56 s. Four time bins were summed for each scan at a pulse frequency of 11 kHz by monitoring the 40 GHz multichannel TDC detector with four-anode/channel detection. Dynamic exclusion was set for 4 s. During the entire acquisition period, the mass accuracy was calibrated once every 20 samples. Furthermore, a QC sample was analyzed every 10 samples to evaluate the stability of the LC-MS.

Metabolomics Data Processing
The acquired LC-MS data pretreatment was performed using XCMS software. Each ion was identified using comprehensive retention time and m/z data. The intensity of each peak was recorded and a three-dimensional matrix containing arbitrarily assigned peak indices (retention time-m/z pairs), sample names (observations), and ion intensity information (variables) was generated. The open-access databases KEGG and HMDB were used to annotate the metabolites by matching the exact molecular mass data (m/z) to those from the database within a threshold of 10 ppm. The peak intensity data were further preprocessed using MetaX [58]. Those features that were detected in less than 50% of QC samples or 80% of biological samples were removed, the remaining peaks with missing values were imputed with the k-nearest neighbor algorithm and normalized using probabilistic quotient normalization method. In addition, the relative standard deviations of the metabolic features were calculated across all QC samples, and those with standard deviations >50% were removed.
The group datasets were normalized before analysis was performed. Data normalization was performed on all samples using the probabilistic quotient normalization algorithm. Then, QC-robust spline batch correction was performed using the QC samples. The pvalue was analyzed by the student t-test, which was then adjusted for multiple tests using the FDR (Benjamini-Hochberg) for the selection of differential metabolites. The filtering conditions for the significantly altered metabolites were as follows: ratio ≥ 2 or ratio ≤ 1/2, q value ≤ 0.05, VIP ≥ 1. The metabolites were analyzed using OmicStudio tools (https://www.omicstudio.cn/tool, accessed on 30 November 2022).

Statistical Analyses
All statistical analyses were conducted using Statistical Package for the Social Sciences (SPSS Version 21.0 for Windows). Significant differences were determined by ANOVA and Duncan's multiple range test (p < 0.05). All the data presented in this study were calculated from three independent biological replicates. Excel 2010 (Microsoft, United States) was applied to draw the bar graphs and broken line graphs. Each experiment was repeated either three or six times.

Conclusions
In conclusion, this study demonstrated that the oil and protein contents increased gradually during fruit development. Conversely, the contents of soluble sugar and starch decreased gradually during fruit development. The walnut oils reached the maximum value of 69.61% in mature embryos. We found that 50-90 DAP was the most critical period for oil synthesis. Combined metabolome and transcriptome is an effective analytical method for explaining the relationship between key genes and metabolites involved in biosynthesis pathways. The expression patterns of LEC1, FUS3, ABI3, and WRI1 were consistent with the pattern of walnut oil accumulation. These TFs may play a critical role in walnut oil synthesis. ACCase, KAS, LACS, FATA, FATB, and PDAT genes may be crucial genes required for walnut oil synthesis. FAD2 and SAD genes showed higher expression levels during walnut oil accumulation. Then, the contents of metabolites such as 18:1-PC, 18:2-PC, and 18:3-PC increased significantly with progressing fruit development. It is noteworthy that high expression levels of FAD2 and SAD genes may be conducive to the rapid accumulation of walnut unsaturated fatty acid, as they are key genes for lipid synthesis. The walnut lipid profile and the lipid metabolism pathway constructed here have important theoretical and practical value for further study of walnut lipid metabolism and functional development. In addition, further studies on their functions in FA biosynthesis and oil accumulation are required.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/plants12030538/s1, Figure S1: Total ion chromatogram; Table S1: Summary of sequencing data of walnut embryos and information of clean reads; Table S2: The mapping result of each sample; Table S3: Differentially expressed transcription factor families; Table S4: Identified related genes and metabolites involved in Lipid-Related; Table S5: Identified related genes and metabolites involved in glycolysis; Table S6: Identified related genes and metabolites involved in TCA cycle; Table S7: Primers used for qPCR analysis.