Transcriptome and Metabolome Analyses Provide Insights Into the Composition and Biosynthesis of Grassy Aroma Volatiles in White-Fleshed Pitaya.

Aroma is one of the major inherent quality characteristics in fruits. Understanding the composition of aroma volatiles and their biosynthesis mechanism is crucial to improving fruit quality. However, the biosynthesis mechanism of aroma volatiles has not been characterized yet in white-fleshed pitaya (Hylocereus undatus). This study was performed to investigate aroma volatiles and related gene expression patterns in the pulp of "mild grassy" and "strong grassy" aroma cultivars. Analysis of volatile composition and concentration showed that aldehydes, alcohols, esters, and alkenes were predominant in both cultivars. However, comparative analysis revealed a significant difference in the concentration of several metabolites, particularly hexanal and 1-hexanol. The results of the comparative transcriptome identified a large number of aroma-related differentially expressed genes. The majority of these genes were enriched in fatty acid and isoleucine degradation pathways. According to integrative analyses, changes in the expression of lipoxygenase pathway genes, specifically FAD, LOXs, HPLs, and ADHs, probably lead to the difference in strength of "grassy" aroma between both cultivars. The qRT-PCR of 18 aroma-related genes was performed to validate the transcriptome analysis. Our results identified key genes and pathways connected with the biosynthesis of aroma volatiles in white-fleshed pitaya. These results will be useful to dissect the genetic mechanism of fruit aroma in white-fleshed pitaya.

Introduction

Dragon fruit or pitaya is an exotic fruit with extreme nutritional and health benefits.[1,2] The genus Hylocereus has 16 species; however, red-fleshed and white-fleshed pitayas are the widely commercially produced species, in Vietnam, China, Mexico, the United States, Taiwan, and the Philippines.[3,4] The pitaya fruit has a short shelf life, and its storage causes changes in aroma, which ultimately reduces the fruit quality.[5,6] Besides, limited knowledge about the fruit-quality standards is the major constraint in large-scale export or marketing. Market competitiveness of fruits is generally regulated by fruit-quality traits such as appearance, texture, aroma, and flavor.[7] Fruit aroma is an essential nonvisual quality parameter that affects consumer perception and choice.[8] Its composition is altered with the difference in species, cultivars, growing environment, and postharvest conditions.[9,10] Fruit aroma is produced by different volatile compounds. Predominant aroma volatiles present in fruits include aldehydes, alcohols, esters, phenols, alkenes, and ketones.[11−13] To date, many studies have already identified a diversity of volatile compounds in fruits. For example, 400 aroma compounds in mango,[14] 350 in apple,[15] 240 in melon,[16] and 400 in tomato[17] have been identified in different cultivars. However, knowledge about aroma volatiles in pitaya is limited, especially grassy aroma, which is unpleasant to consumers. A previous research identified 121 volatile compounds in yellow dragon fruit.[18] Another study reported only 34 volatiles in the juice of different cultivars of pitaya.[4]

Even though the profile and composition of aroma volatiles in fruits change with species and cultivars, they are mainly synthesized through the metabolism of the fatty acid, amino acid, terpenoids, and carotenoid pathways.[19] Research evidence showed that fatty acids and amino acids are the major precursors of aroma biosynthesis in most fruit species.[20] Specifically, fatty-acid-regulated lipoxygenase (LOX) and β-oxidation processes are the major determinants for the biosynthesis of aldehydes, alcohols, and straight-chain esters.[21,22] The aroma derivatives formed through the LOX pathway are primarily synchronized by fatty acid desaturase (FAD), LOX, hydroperoxide lyase (HPL), alcohol dehydrogenase (ADH), and alcohol acyltransferase (AAT) coding genes.[23,24] In recent years, many candidate genes involved in this pathway have been identified in apple,[25] mango,[26] grapes,[27] pear,[28] peach,[29] tomato,[30] strawberry,[31] and kiwifruit.[32] β-oxidation, as well as amino acids such as alanine, valine, leucine, isoleucine, and phenylalanine, is a potential precursor of acyl-CoAs.[8,33] Later on, acyl-CoAs leads to the biosynthesis of aroma volatiles through the functional mechanism of acyltransferase, decarboxylation, oxidation, and reduction-related enzymes.[34] However, genes involved in β-oxidation and amino acid metabolism pathways are less characterized in fruits.

For particular importance, the differences in aroma components and concentrations between cultivars lead to different aromas and flavors in fruits. Therefore, characterization of aromatic volatiles is necessary to discriminate different species and even cultivars of a species.[35] For instance, methyl-1-butanol, 1-hexanol, and butyl acetate are the major contributors to the aroma in the “Pink Lady” apple cultivar.[36] In contrast, hexyl-methylbutyrate, hexyl 2-methylbutyrate, α-farnesene, and (E)-2-hexenal are key components of aroma in the “Honeycrisp” apple cultivar.[37] Moreover, the compositions of aroma volatiles significantly not only alter the mechanism of fruit maturity, ripping, and taste but also influence the postharvest handling, storage conditions, and final fruit-quality parameters. Advance research with gas chromatography-mass spectrometry and RNA sequencing tools is noteworthy to reveal the regulatory mechanism of aroma in fruits.[13] However, previous research in pitaya was usually focused on health benefits,[1] disease,[38] stress,[39] storage conditions,[5] and pulp or peel color formation.[40] To our knowledge, no research has yet been performed in white-fleshed pitaya to investigate the composition and concentration of aroma volatiles. Furthermore, the biosynthesis mechanism of aroma volatiles in white-fleshed pitaya is largely unknown. To fill these gaps, an integrative analysis of volatile profiles and transcriptomes was performed by selecting contrasting cultivars of white-fleshed pitaya with different strengths of grassy aroma. This study aims at identifying the key aroma volatiles and to further uncover their biosynthesis mechanism. Our results will be important to understand the genetic mechanism of the aroma of white-fleshed pitaya in the future.

Results

Volatile Profiles in the Pulp of White-Fleshed Pitaya

Fruit aroma is a key determinant that affects fruit flavor, customer acceptance, and competitiveness in the market. To investigate the aroma volatiles of white-fleshed pitaya, two contrasting cultivars including a mild grassy aroma 6–17 cultivar (hereafter NX) and a strong grassy aroma 16–41 cultivar (X) were utilized to detect organic volatiles in the pulp (Figure ). These cultivars have no obvious phenotypic difference except the aroma. A total of 60 organic volatiles were putatively detected in both cultivars (Table S1). The majority of the accumulated volatiles belonged to different kinds of aldehydes, esters, alcohols, and alkanes. An alkane named n-hexane was predominant in both cultivars. Furthermore, alcoholic compounds like 1-hexanol, 2,4-di-tert-butylphenol, and 1,3-dioxol-2-one had very high accumulation levels. The major aldehydes in both cultivars were benzeneacetaldehyde, hexanal, and 2-isocyanatopyridine. Among the ester compounds, the concentrations of formic acid-octyl ester, 1,2,5-oxadiborolane, 2,3,3,4,5-pentaethyl, and octanoic acid-ethyl ester were the highest (Table S1).

Figure 1

Appearance of contrasting aroma cultivars used in this study. Here, 6–17 represents mild grassy aroma cultivar (NX) and 16–41 stands for strong grassy aroma cultivar (X).

Besides, straight and branched alkanes including hexadecane, benzene-1-ethyl-4-methoxy, and pentadecane were dominant. The comparative analysis identified differences in the concentration of volatiles in the pulp of NX and X. There were two aldehydes, two esters, three alcohols, one alkane, one acid, and one other compound among the top 10 differentially accumulated volatiles in cultivars (Figure ). In the aldehyde group, hexanal and benzeneacetaldehyde had higher concentrations in X than NX. However, alcohols such as 1-hexanol, terpinen-4-ol, and 1,2-benzenediol-o-(thiophen-2-acetyl) exhibited dynamic differences between X and NX. In particular, terpinen-4-ol and 1,2-benzenediol-o-(thiophen-2-acetyl) had manyfold low levels of contents in X as compared to NX. Contrasting results were obtained for dodecanoic acid-ethyl ester and methyl 2-hydroxy-4-methylbenzoate. These two ester compounds showed a higher accumulation in X than NX. The pentanoic acid, tridecane, and oxacycloheptadecan-2-one were the other volatiles that showed a high difference between X and NX. The relative contents of pentanoic acid and tridecane were higher in the X cultivar. However, oxacycloheptadecan-2-one showed an increased accumulation in NX than X. In short, the results of the comparative volatile profiling in NX and X suggested that different compositions and concentrations of aldehydes, esters, and alcohols confer the different strengths of the grassy aroma between X and NX cultivars. Moreover, volatiles such as 1-hexanol and hexanal mainly affect the strength of the “grassy” aroma in the white-fleshed pitaya.

Figure 2

Heatmap of top 10 (fold changed) aroma volatile contents identified in the comparison of NX and X. Here, NX denotes mild grassy aroma cultivar and X denotes strong grassy aroma cultivar.

RNA Sequencing and De Novo Assembly

The pulp samples from the two contrasting aroma cultivars NX and X were used for RNA sequencing in triplicate. This helped unveil the gene expression profile associated with aroma volatiles in white-fleshed pitaya. The total raw read means were 62 018 490 for NX and 62 064 848 for X, whereas the means of clean reads were 60 165 751 and 60 515 777 for NX and X, respectively (Table ). The raw reads into clean reads transformation mean was almost 99% in our sequencing. After filtering the low-quality reads, on average 8.5 Gb clean data was acquired for NX and X. The Q20 score of all samples was above 97.1%, with a Q30 mean score of over 93%. The guanine–cytosine (GC) content % in all samples showed a mean value of 49. The nuclear genome sequence of the white-fleshed pitaya was unavailable when this study was conducted; hence, Trinity software was used to assemble the clean reads into unigenes. In total, the de novo assembly generated 109 738 unigenes with a mean length of 1447 bp, N50 length of 2070 bp, and N90 length of 693 bp (Table S2). The length of assembled unigenes showed a great diversity ranging from 200 to ≥2000 bp. The highest portion (24.2%) of unigenes had ≥2000 bp assembled length, followed by 400–500 bp unigene lengths. The unigenes with 1900–2000 bp length were less represented (Figure A). The mapping percentages were about 83 and 93% in NX and X samples, respectively.

Table 1

Details of the Transcriptome Sequencing Quality among NX and X Cultivars

sample	raw reads	clean reads	clean base (G)	error rate (%)	Q20 (%)	Q30 (%)	GC content (%)
NX1	53 389 434	51 399 455	8.02	0.02	97.55	93.51	49.76
NX2	56 185 988	54 272 754	8.14	0.02	97.61	93.03	48.73
NX3	67 850 992	66 058 748	9.91	0.02	97.57	93.02	49.88
X1	72 288 518	70 428 758	10.56	0.02	97.52	92.84	50.11
X2	51 841 178	50 602 796	7.59	0.02	97.76	93.27	49.89
X3	68 302 763	66 522 903	8.34	0.02	97.12	93.43	48.75

Figure 3

Statistics for length distribution and functional annotations of unigenes. (A) Length distribution of the sequence with the percentage of unigene. (B) Percentage of functional annotations for unigenes in seven different databases.

Unigene Functional Annotation and Gene Expression

The annotation information for the unigenes was retrieved with BLAST search against public databases including KEGG, NR, SwissProt, Trembl, KOG, GO, and Pfam. Approximately 70% of the total assembled unigenes had sequence similarity in the Nr and Trembl databases (Figure B). There were only 56, 49, 44, 58, and 54% unigenes that showed unique annotations in the KEGG, SwissProt, KOG, GO, and Pfam databases, respectively. As for the plant species homologous sequence similarity index, most of the unigene sequence similarities matched with Beta vulgaris (29.4%), Chenopodium quinoa (25.5%), and Spinacia oleracea (17.3%) (Figure S1). The annotated results in the GO database revealed that the majority of the unigenes are involved in cellular and binding functional activities (Figure S2). The expression level of each unigene was calculated in the form of fragments per kilobase of transcript per million fragments mapped (FPKM) value. The FPKM-based correlation coefficient analysis determined that the correlation among different repeats for NX was 0.88 and between repeats for X was 0.99 (Figure S3). In contrast, weaker correlation scores were obtained among samples of NX and X. For example, NX showed a correlation coefficient mean of 0.2 with X.

Differential Gene Expression between Contrasting Aroma Cultivars of White-Fleshed Pitaya

The difference in gene expression is highly interconnected with the difference in phenotypic characteristics. To understand the transcripts responsible for the grassy aroma strength in white-fleshed pitaya, we compared gene expression profiles of the “mild grassy” aroma NX cultivar with the “strong grassy” aroma X cultivar. The differentially expressed genes (DEGs) were identified among the pulp libraries of these cultivars. The screening criteria for differential genes were log2 Fold Change ≥ 1, and the false discovery rate was <0.05. The comparative analysis showed a total of 9527 DEGs between NX and X. Of all, 3322 DEGs were downregulated in NX as compared to X, whereas 6205 DEGs were upregulated in NX compared to X (Figure S4). Theoretically, a large number of DEGs with undulating expression trends caused a genetic difference in aroma compounds between NX and X cultivars of white-fleshed pitaya.

Functional and Pathway Enrichment Analyses of DEGs

Functional characterization of DEGs was obtained through GO and KEGG enrichment analyses. The GO enrichment analysis categorized functional annotations of DEGs into different classes of molecular and biological functions. In particular, the majority of DEGs showed functional enrichment associated with the biological mechanism of secondary metabolic biosynthesis and the phenylpropanoid metabolism. In the category of molecular functions, RNA polymerase and methyltransferase-related DEGs had more significance (Figure ). The KEGG pathway enrichment analysis based on P-value < 0.05 determined that the major portion of DEGs was enriched into 20 different pathways (Table S3). In particular, the phenylpropanoid biosynthesis, pyruvate metabolism, brassinosteroid metabolism, and nitrogen metabolism pathways were the most enriched ones (Figure ). These results predict that different genes involved in various metabolism pathways are critical for the strength of grassy aroma in white-fleshed pitaya.

Figure 4

GO enrichment analysis for DEGs of NX and X cultivars of white-fleshed pitaya.

Figure 5

Pathway enrichment analysis for DEGs of NX and X cultivars of white-fleshed pitaya.

Identification of DEGs Involved in the Biosynthesis of Aroma Volatiles

Our analysis identified 83 aroma-volatile-related core DEGs in NX and X cultivars. Aroma volatiles like aldehydes, alcohols, and straight-chain esters are generally formed by the metabolism of fatty acids.[13,29] This metabolic pathway comprises several kinds of reactions, produces or breaks down molecules within the cell, and is regulated by many enzyme encoding genes. Fatty acid metabolism produces volatiles in fruits through catalytic activities of different enzymes involved in the lipoxygenase pathway (Figure A). In our study, 28 genes related to fatty acid biosynthesis and degradation showed significant expression differences in the comparison of NX and X (Figure B). In particular, the expression profiles of four genes (unigene 24528, unigene 2779, unigene 36390, and unigene 17220) encoding acetyl-CoA carboxylase 1(ACC1) showed significantly higher levels in the NX than the X cultivar. Similarly, acyl carrier proteins (ACPs) such as palmitoyl acyl carrier protein thioesterase (FATB) and enoyl acyl carrier protein reductase (ENR) genes had significant expression differences in the comparison of both cultivars. Phospholipase (PL) activity converts membrane phospholipids into free polyunsaturated fatty acids. Four genes PL unigene 32816, unigene 36211, unigene 36212, and unigene 39810 were downregulated in NX. However, three genes denoted unigene 12636, unigene 16304, and unigene 30368 were upregulated in NX. Fatty acid desaturase (FAD) is an enzyme that plays a key role in the synthesis of unsaturated fatty acids like linoleic and α-linolenic acids.[29] Eight FAD-related genes denoted unigene 29976, unigene 43322, unigene 57295, unigene 59928, unigene 62255, unigene 63409, unigene 7753, and unigene 7766 were significantly upregulated in NX. However, unigene 26676 and unigene 43108 were downregulated in NX (Figure B). The dynamic expression changes in genes associated with carboxylase, phospholipase, reductase, and desaturase mechanisms indicate their important roles in fruit aroma in white-fleshed pitaya.

Figure 6

Expression profiles of aroma volatile genes associated with fatty acid metabolism and lipoxygenase pathways in white-fleshed pitaya. (A) Simplest diagram of the fatty-acid-metabolism-derived volatile pathway; (B) fatty-acid-metabolism-related DEGs; (C) lipoxygenase-pathway-related DEGs.

Lipoxygenase (LOX) catalyzes linoleic and α-linolenic acid into hydroperoxides, which are utilized as substrates to produce aldehydes, alcohols, and straight-chain esters.[41] Our results identified 28 DEGs involved in this pathway (Figure C). Especially, the expression of LOX-related unigene 629, unigene 40168, and unigene 160 was significantly lower in the NX compared to the X cultivar. In contrast, LOXunigene 6060 and unigene 21576 showed a higher expression in NX. The aldehydes produced by the hydroperoxide lyase (HPL) enzyme are reduced into different types of alcohol by alcohol dehydrogenase (ADH). Finally, alcohols are converted into esters through the functional activity of alcohol acyltransferase (AAT) enzymes. Interestingly, three HPL-associated genes showed significant expression difference between NX and X, of which unigene 34920 and unigene 34924 were downregulated in NX. The significant changes in expression profiles of HPL genes in both cultivars are probably linked with different concentrations of aldehydes, specifically hexanal and benzeneacetaldehyde compounds. Our results further identified that 10 ADHs encoding genes including unigene 35255, unigene 37754, unigene 38406, unigene 35256, unigene 51849, unigene 28518, unigene 30358, unigene 30106, unigene 2653, and unigene 44321 were upregulated in NX (Figure C). However, only unigene 44320 and unigene 57971 had a higher expression in X than NX. The greater expression of ADH genes in NX may cause a high accumulation of terpinen-4-ol and 1,2-benzenediol-o-(thiophen-2-acetyl) alcoholic volatiles in this cultivar. The expression levels of all seven AAT enzyme encoding genes (unigene 44283, unigene 37310, unigene 12202, unigene 17569, unigene 32857, unigene 38845, and unigene 50844) were higher in NX than X. It is noteworthy that significant expression differences in AAT enzyme encoding genes likely produce different concentrations of volatile esters in both cultivars.

β-oxidation is another fatty acid degradation mechanism involved in the production of esters in fruits[42] (Figure A). The expression trends of 10 key DEGs involved in this mechanism are presented in Figure B. The six unigenes (unigene 22538, unigene 2832, unigene 22100, unigene 22101, unigene 22105, and unigene 50813) related to β-oxidation enzyme 3-ketoacyl-CoA thiolase 2 (KAT2) showed a higher expression in NX. On the other hand, the unigenes associated with the β-oxidation multifunctional protein (MFP: unigene 19902 and unigene 19903) and acyl-CoA oxidase (AOX: unigene 22919 and unigene 48984) were downregulated in NX than in X. Different expressions of KAT2, MFP, and AOX indicate their dynamic role in acyl-CoAs biosynthesis and finally for aldehyde, alcohol, and ester biosynthesis in both cultivars. The isoleucine is an amino acid, and its degradation generally produced branched-chain esters.[13] The critical enzymes for isoleucine degradation include branched-chain amino acid aminotransferase (BCAT), pyruvate decarboxylase (PD), aldehyde dehydrogenase (ALDH), and carboxylesterase (CXE) (Figure C). This study identified 17 DEGs allied to these enzymatic reactions (Figure D). In particular, two BCAT interconnected genes (unigene 53860 and unigene 33329) were upregulated in NX compared to X, while unigene 1968 was downregulated in NX. The enzyme PD-associated unigene 11766, unigene 11767, and unigene 59816 showed a higher expression level in NX than in X. It was observed that five ALDH unigenes such as unigene 31927, unigene 22877, unigene 14643, unigene 36534, and unigene 28447 had significant difference in expression levels between NX and X. The CXE-linked unigene 1098 and unigene20621 were upregulated in NX compared to X. However, CXE encoding unigene 20622, unigene 28242, unigene 20618, and unigene 20620 exhibited a higher expression in X than in NX. These differential expression profiles of isoleucine degradation genes may affect the altered accumulation of volatile esters in both cultivars. Together, our results suggest that the differential expression patterns of FADs, LOXs, HPLs, ADHs, AATs, and ALDHs may underlie the differential concentrations of different aromatic volatiles observed between the two cultivars. In particular, higher expressions of LOXs and HPLs are in agreement with the higher amount of hexanal in the X cultivar. In contrast, the high expression of LOXs and ADHs agrees with a higher amount of 1-hexanol in the NX cultivar.

Figure 7

Expression profiles of aroma volatile genes associated with β-oxidation and isoleucine degradation pathways in white-fleshed pitaya. (A) Simplest diagram of β-oxidation-derived volatile pathway, (B) β-oxidation-related DEGs, (C) simplest diagram of isoleucine-degradation-derived volatile pathway, and (D) isoleucine-degradation-related DEGs.

Quantitative Real-Time PCR (qRT-PCR) Validation

To confirm the accuracy of our transcriptome data analysis, 18 aroma biosynthesis-related genes were randomly selected to perform a qRT-PCR analysis. These genes showed a significant differential expression between “strong grassy” aroma and “mild grassy” aroma cultivars. All gene expression profiles obtained by the qRT-PCR had a strong expression fold change between both cultivars (Figure A). Remarkably, the Pearson correlation coefficient revealed that the gene expression data generated with RNA-seq and qRT-PCR had similar trends (R2 > 0.8) (Figure B). Thus, these findings confirm the validity of our transcriptome analysis.

Figure 8

qRT-PCR-validated aroma-related genes in white-fleshed pitaya. (A) qRT-PCR relative expression levels among NX and X cultivars. (B) Pearson’s correlation of RNA-seq and qRT-PCR data. Analysis was performed in triplicate biological samples and technical triplicates. A relative expression of more than 2 or lower than 0.5 indicates significant change between the two cultivars.

Discussion

Pitaya is an ornamental fruit with tremendous health benefits. The fruit demand is increasing over recent years. It is a short-shelf-life nonclimacteric fruit. Storage most likely modifies aroma volatiles, which influences the postharvest fruit quality. Thus, the characterization of aroma volatiles is a prerequisite to improve the postharvest fruit quality, processed pitaya products, and storage techniques. This study mainly discusses the profiles of aromatic volatiles and their biosynthesis mechanism in two white-fleshed pitayas.

Composition and Concentration of Aroma Volatiles in White-Fleshed Pitaya

Composition and concentration of aromatic volatiles in fruits alter the texture, flavor, and nutrient contents. Moreover, characterization of aroma volatile profiles provides important data to evaluate fruit quality. According to our aroma volatile data analysis, n-hexane, 1,3-dioxol-2-one, pentadecane, octanoic acid-ethyl ester, benzeneacetaldehyde, hexanal, and 1-hexanol are major compounds in the pulp of white-fleshed pitaya. These results indicate that aroma in pitaya is a complex mixture of many volatile compounds, and most belong to alcohols, aldehydes, esters, and alkenes. In apples, alcohols, aldehydes, and esters are key components of fruit aroma.[15] Major aroma components in cherry include carbonyls, alcohols, acids, and esters.[43] The composition and concentration of aroma volatiles are specific to species and cultivars. Futhermore, it is altered with the genetic makeup of fruit, cultural practices, environmental conditions, harvest maturity, postharvest handling, and storage conditions.[8] Besides, many studies reported that volatile esters are the primary components of aroma in many fruits such as peer,[44] banana,[45] pineapple,[46] and melon.[47] Olfactometry is one of the important oldest methods to assess the sensory nature of aroma in food and beverages. This method defines that the “grassy” aroma in fruits is associated with compounds of aldehydes and alcohols.[48] In particular, major aldehyde compounds with the “grassy” aroma include hexanal, (E)-2-hexenal, and benzaldehyde, whereas alcohols like benzyl alcohol, 1-hexanol, and (E)-2-hexen-1-ol are major compounds with a “grassy” note.[43,49] Interestingly, our comparative analysis for volatile profiling showed that the “strong grassy” aroma cultivar had a significantly higher concentration of hexanal as compared to the “mild grassy” aroma cultivar. In contrast, 1-hexanol showed a significantly higher concentration in the “mild grassy” aroma cultivar. In agreement with a previous study, the differential concentration of 1-hexanol and hexanal most probably caused the difference of strength in “grassy” aroma in both cultivars of white-fleshed pitaya.[50] Earlier research stated that storage at 10 °C enhanced hexanal but reduced hexanol in different cultivars of pitaya.[4] Our study detected a significantly higher concentration of dodecanoic acid-ethyl ester and methyl 2-hydroxy-4-methylbenzoate-like esters in the “strong grassy” aroma cultivar. However, the level of “grassy” aroma in bananas reduced with an increased level of volatile esters at ripening.[51]

Biosynthesis Mechanism of Aromatic Volatiles in White-Fleshed Pitaya

The transcriptome and metabolomics analysis is widely utilized to explore regulatory genes and pathways involved in the biosynthesis of aroma volatiles.[13,52] The fatty acids and amino acids are the major precursors of aldehydes, alcohols, and esters in fruits. Among fatty acids, the lipoxygenase metabolic pathway is critical to synthesizing “grassy notes” volatiles in plants.[53] The FAD encoding genes in the initial step synthesized polyunsaturated fatty acids. The PL is another enzyme that formed free polyunsaturated fatty acids for the lipoxygenase pathway.[54] The altered expressions of FAD and PL encoding genes in our results predict their role in the breakdown of membrane lipids in pitaya. A previous study found that FAD mutants showed reduced desaturation activity in Arabidopsis.[55] The polyunsaturated fatty acids formed by desaturases enzymes are further converted into hydroperoxides by LOX genes. Then, enzyme HPL-associated genes played a role in producing C6 or C9 types of aldehydes from hydroperoxides. The aldehydes are further metabolized into alcohols through the functional activity of ADH genes. In the last step, the AAT enzyme encodes genes that synthesize esters from alcohols.[54] Our comparative analysis between “strong grassy” and “mild grassy” aroma cultivars identified 28 core genes involved in the LOX pathway. The higher expression of LOX and HPL encoding genes may regulate the higher concentration of aldehydes in the “strong grassy” aroma cultivar. On the other hand, the higher expression of ADH genes in the “mild grassy” aroma cultivar is most likely responsible for the increased level of alcoholic volatiles. These findings are in agreement with the results of the volatile profile of both cultivars. It was reported that LOX- and HPL-related genes influence volatile production in fruits.[56,57] Specifically, a transgenic study stated that reduction in LOX activity not only reduced C6 aldehydes but also reduced “grassy” volatiles.[58] The LOX genes caused a “beany” flavor in soybean seeds, and the identified single nucleotide polymorphism marker linked with LOX2 can be useful for molecular breeding.[59] The HPL1 gene exhibited a high expression in immature green tomatoes and a reduced expression in ripened fruits.[60] The ADH utilized C6 aldehydes as substrates to generate volatile alcohols. In peach, the ADH gene activity regulates “grassy note” volatiles mainly through the biosynthesis of hexanols. Moreover, the overexpression of ADH improved the concentration of volatile alcohols and ultimately changed the tomato flavor.[61] These results together suggest that LOX pathway genes mediate the “grassy” aroma in pitaya.

Straight-chain esters are produced through the fatty acid metabolism, while amino acid isoleucine degradation leads to the biosynthesis of branched-chain esters. The enzyme ATT is a rate-limiting step in ester biosynthesis. Our results showed significantly modified expressions of AAT genes in both cultivars of pitaya. These findings suggest their role in the biosynthesis of volatile esters in pitaya. The functional role of AAT in the biosynthesis of volatile esters has already been identified in apple,[62] banana,[63] and melon.[64] The transgenic AAT1 knockdown in apples significantly reduced the total concentration of esters.[65] The β-oxidation pathway enzymes catalyzed fatty acids into free acyl-CoAs, which are used by ADH to form alcohols. Finally, alcohols are converted into volatile esters by ATT.[42] In our study, many genes putatively involved in β-oxidation showed significant differences among both cultivars of pitaya. Although β-oxidation plays an important role in yielding aromatic volatiles, the characterization of the complete pathway with specific enzymes is not well elucidated in fruits. Amino acids such as alanine, phenylalanine, leucine, and isoleucine mediate the biosynthesis of aromatic alcohols, acids, and esters in fruits. In particular, the branched-chain isoleucine degradation pathway is responsible for the biosynthesis of branched-chain volatiles in apple,[66] banana,[63] and strawberry.[67] In this pathway, the BCAT enzyme first causes degradation, and ultimately, functional activities of ADH, ALDH, and CEX produce branched-chain volatile aldehydes, alcohols, and esters.[13] Our results identified many differentially expressed genes involved in the isoleucine degradation pathway in both cultivars. This may also be contributing to the difference in “grassy” aroma in both cultivars of pitaya. Based on our results and those of previous studies, we propose a simple scheme of “grassy” aroma biosynthesis in pitaya (Figure ), in which LOX pathway genes, especially FAD, LOX, HPL, and ADH, through interaction mechanisms altered the concentration of volatiles like hexanal and 1-hexanol. This ultimately may lead to a “grassy” fruit aroma. Our comprehensive analyses not only characterized aroma volatiles but also facilitated understanding of their biosynthesis mechanism. However, future research is mandatory to explore the molecular mechanism and genetic variants associated with the strength of the “grassy” aroma of white-fleshed pitaya.

Figure 9

Proposed scheme of lipoxygenase-pathway-derived “grassy” aroma formation in white-fleshed pitaya. Here, the genes involved in each step are in italic, red color stands for a significantly high expression level, and the blue color stands for a low expression level in the “strong grassy” aroma cultivar compared to the “mild grassy” aroma cultivar.

Materials and Methods

Plant Materials and Analysis of Volatile Aroma Substances

Fruit samples were selected from white-fleshed pitaya 4-year-old plants grown in Guangxi Modern Agricultural Science and Technology Demonstration Park, China, in 2019. The pulps of two contrasting cultivars, including Shuibaijing 6–17 (NX) with a “mild grassy” aroma and 16–41 (X) with a “strong grassy” aroma, were utilized as test materials at the fruit ripened stage. Sampling was performed in three independent biological replicates from three different fruit plants for each cultivar. The samples were placed immediately in liquid nitrogen after harvesting and stored in an ultralow-temperature refrigerator at −80 °C for metabolomics analysis and RNA sequencing. Fully automatic headspace solid-phase microextraction (HS–SPME) combined with gas chromatography-mass spectrometry (GC–MS) was used to detect and analyze volatile aroma substances in the pulp. The volatile substances were extracted and quantified for each sample by following the descriptions of Wang et al.[68] In brief, 1 g of pulp was taken for liquid nitrogen grinding and 2 mL of saturated NaCl solution and 10 uL of internal standard solution (n-Hexane Solution analytical, S-12550M4-1M) were added in 20 mL headspace vials. The vials were sealed using crimp-top caps with TFE-silicone headspace septa (Agilent). Later on, each vial was put at 60 °C for 10 min, extracted with a 65 μm-diameter and 1 cm-length divinylbenzene/carboxen/poly(dimethylsiloxane) fiber (Supelco, Bellefonte, PA), and finally exposed to the headspace of the sample for 20 min at 60 °C. After extraction, the pitaya pulp volatile identification and quantification were retrieved with an Agilent Model 7890B GC and a 7000D mass spectrometer equipped with a 30 m × 0.25 mm × 1.0 μm DB-5MS (5% phenyl-polymethylsiloxane) capillary column and helium as a carrier gas at a linear velocity of 1.0 mL/min. The injector and detector temperatures were 250 and 280 °C, respectively. At first, the furnace temperature was set at 40 °C for 5 min and increased to 280 °C for 5 min. The mass spectra electron impact (EI) ionization mode was 70 eV, and scanning was performed in the range m/z 30–350. Finally, the quality and quantity of aroma volatiles were analyzed by comparing them with the NIST14 library using MassHunter software (Agilent). Volatiles were considered to be significantly different among cultivars with a log 2-fold change difference of ≥1 or ≤1.

Total RNA Extraction, Sequencing, and Analysis

Total RNA was extracted from pulp samples of both cultivars with the CTAB method.[69] The integrity of RNA and the presence of DNA contamination were assessed with agarose gel electrophoresis. The Qubit 2.0 fluorescent meter was used to measure the high accuracy of the RNA concentration. Furthermore, the accurate detection of RNA integrity was achieved with Bioanalyzer Agilent 2100 (Agilent Technologies) before constructing a sequenced library. After libraries were constructed with standard quality, the Illumina HiSeq platform with the remanded protocol was used for sequencing de novo pair-end RNA sequencing. Reads having adapters and of low quality were removed from original sequencing data and filtered to obtain high-quality reads. The high reads were then assembled with Trinity 2.6.6 to obtain the unigene sequence.[70] The functional annotations for assembled unigenes were acquired with blastx from the NCBI nonredundant nucleotide and protein database (Nr), EuKaryotic Orthologous Group (KOG), Kyoto Encyclopedia of Genes and Genomes database (KEGG), Gene Ontology (GO) Swiss-Prot, Trembl, and Pfam. The expression level of each unigene was determined with the FPKM score. The R package DESeq. 2 was used to obtain DEGs among any two samples.[71] The log 2 (fold change) >1 or <−1 and P-value <0.05 were threshold criteria to identify DEGs. The functional and pathway enrichment analyses for DEGs were executed with Goatools[72] and KOBAS software,[73] respectively. The P-value ≤0.05 was the threshold criterion for significant enrichment.

qRT-PCR Analysis

To perform qRT-PCR, the first standard cDNA for each sample (three biological replicates) was synthesized with MonScript RTIII All-in-One Mix with the dsDNase kit. Gene-specific primers were obtained with Oligo 7 and are detailed in Table S4. The qRT-PCR reaction mixture in three technical repeats was prepared with the QuantiNova SYBR Green PCR kit. Each reaction mixture had a final volume of 10 μL and contained 5 μL of 2× SYBR Green PCR Master Mix, 0.05 μL of QN ROX reference dye, 0.7 μL of each primer, 2.55 μL of RNase-free water, and 1 μL of template cDNA. The running protocol for qRT-PCR included initial heating at 95 °C for 2 min, 40 cycles of denaturation at 95 °C for 15 s, and combined annealing/extension at 60 °C for 30 s. The generated raw data were analyzed with the 2–ΔΔCt method to obtain the final expression of each gene.[74]

Conclusions

This study revealed that aldehydes, alcohols, esters, and alkanes are key aromatic volatiles in the pulp of white-fleshed pitaya. The concentrations of hexanal and 1-hexanol showed a significant difference between the “mild grassy” aroma and “strong grassy” aroma cultivars. In addition, their regulatory genes exhibited significant expression differences in both cultivars. The integrative volatile and transcriptome data analyses identified that the lipoxygenase pathway gene most likely regulates the “grassy” aroma. Our results provide knowledge about aroma formation in pitaya. In-depth functional analysis of the candidate genes is necessary to fully elucidate the genetic mechanism of the “grassy” aroma formation and its strength in white-fleshed pitaya.