1. Introduction
Papaya (
Carica papaya L.) is a tropical fruit crop cultivated worldwide. In Australia, papaya was first introduced as a commercial crop in the 1900s, with currently >90% of production in the tropical regions of North Queensland [
1]. Since its introduction, Australian papaya production has steadily increased, with 6.3 thousand tonnes produced in 2020 [
2]. Initial breeding efforts for Australian papaya have focused on traits such as resistance to ringspot virus and increased yield rather than flavour or other consumer-preferred characteristics [
3]. However, the commercial Australian papaya varieties vary greatly in their flavour profiles, likely impacting consumer demand and, hence, sales. These range from sweet and fruity to others with bitter aftertastes and unpleasant aromas [
1]. Therefore, a major objective of the national papaya breeding program (PP18000, funded by Hort Innovation) is to develop varieties with improved flavours.
Flavour is a multifactorial trait that includes both mouth perception and aroma [
4], contributed by the interactions between sugar, organic acids, and volatile organic compounds (VOCs) [
3,
5]. Several studies have illustrated that combinations of different amounts and ratios of sugar types and VOCs directly impact papaya consumer liking [
3,
6,
7,
8]. Therefore, the underpinning biochemical compounds and their expression levels are key selection targets for improvement towards preferred flavours. Accordingly, attempts have been undertaken over the past 50 years to uncover the key discriminant compounds that produce the unique aromas and flavours of papaya. Methods have included headspace, odour olfactometry, and gas chromatography mass spectrometry (GCMS) technology, resulting in the discovery of more than 400 VOCs that are proposed to be related to papaya aroma and flavour. The accurate identification and quantification of the VOCs was reliant on reference standards, of which a limited and varying number were used [
4,
6,
8,
9,
10,
11]. In previous studies, papaya aroma and flavour compounds have generally fallen into ester- or terpene-rich chemotypes, with high concentrations of linalool and benzyl isothiocyanate commonly detected [
6,
8,
9,
12,
13]. Meanwhile, papaya sweetness is largely derived from three soluble sugars: glucose, fructose, and sucrose [
4]. During fruit ripening, glucose and fructose accumulate and are converted to sucrose under multiple enzyme metabolisms, and sucrose is the predominant sugar that contributes to flesh sweetening [
14]. To evaluate papaya sweetness, it is important to assess the total sugar content and the percentage of each type of soluble sugar present [
15]. In addition to the precise evaluation of biochemical metabolites involved in aroma and flavour production, sensory panel testing and consumer acceptance surveying are commonly used to determine flavour attributes preferred by consumers [
8,
16,
17]. Blind preference surveying may reveal the range of consumer liking levels among diverse papaya samples. The outputs may then be used to assess for correlation with biochemical profiles to identify preferred aromas, flavours and, hence, varieties.
Moreover, genomic-based studies may be undertaken to identify, characterise, and validate sequences and putative alleles related to aroma, flavour, and sweetness. This approach may also aid with understanding the metabolic pathways of the underpinning sugar and VOCs. In a previous study, enzymes involved in the production, translocation, and storage of sugars in papaya fruit included the soluble acid invertase, the insoluble acid invertase, neutral invertase, sucrose phosphate synthase (SPS), and sucrose synthase (SS) [
18]. Accordingly, several genes related to these functional enzymes have been identified in previous studies for use in the selection of papaya varieties with higher sugar content [
19,
20,
21]. In addition, the isoprenoid biosynthesis, the shikimic acid, and the acyl lipid catabolism pathways were associated with the production of papaya fruit aroma and flavour VOCs [
4]. Further research is required to identify the genes and alleles that encode enzymes and regulatory sequences in these pathways and that govern VOC synthesis.
This study aimed to characterise the flavour profiles of the major commercial papaya varieties in Australia through sensory descriptive profiling, consumer acceptability study, VOC profiling, and sugar component determination. Furthermore, the differential expression of flavour-related gene sequences in papaya was evaluated, and their underpinning metabolic pathways are proposed.
3. Discussion
Flavour is a major factor influencing consumer purchasing decision, with a drop-off in flavour quality leading to consumer dissatisfaction [
16,
22]. Therefore, recent breeding has concentrated on improving fruit flavour quality traits to expand the market [
1,
4]. Meanwhile, human flavour perception includes sugars, acids, and a group of volatile compounds as well as quantitative information from diverse sensory systems [
22]. To improve the flavour of Australian papaya varieties, objective standards of good taste and aroma must be set, and the key underpinning metabolism pathways are needed for in-depth investigation. In this study, by using a combination of biochemical, sensorial and consumer acceptability evaluation, we identified the key biochemical variables which significantly correlated to consumer liking and show their application in a selective breeding program.
Papaya sweetness is mainly contributed by glucose, fructose and sucrose [
4,
15]. During fruit development, glucose is the predominant sugar that starts accumulating since seed maturation. When fruit is harvested and starts ripening, sucrose becomes the major sugar, which results in the sweet taste in the fruit flesh [
14]. Fruit sweetness could be evaluated as the unit Degrees Brix (◦Brix), which indicates the total soluble solids in solution, which is a traditional method used by the fruit juice and fresh produce industry [
4]. For papaya, Brix is considered as the percentage of glucose, fructose, and sucrose in the solution [
23]. The minimum Brix level requirement for mature papaya fruit in Hawaii should be 11.5% to meet market grade [
24]. Red papayas have overall higher Brix levels than yellow papayas, which may indicate that red-flesh varieties contain higher sugar contents than yellow varieties. Based on the correlation analysis between sugar determination and sensory panel outputs, the sweetness level evaluated in Brix correlated positively with the perception of sweetness by humans (
p < 0.05). ‘Skybury’ fruit was found to be the sweetest, due to that fact of its higher sweet caramelised flavour score (74.8 ± 9.6) and sweet aftertaste score (54.8 ± 18.2). Compared to other cultivars with 19.2–64.2 for
sweet caramelised flavour and 9.2–45.2 for
sweet aftertaste, the lowest scores came from ‘H13’. The concentrations of three sugar components showed no significant difference among the five papaya varieties, while the percentage of glucose in total sugar was significantly different among varieties (
p < 0.05). Glucose was, on average, 25% of total sugars in all varieties, while sucrose was approximately 40% of total sugars from our study, which were similar to those described by Nantawan et al. [
19] for Australian papaya cultivars ‘Sunrise Solo’ and ‘RB2’. A previous study illustrated that each sugar type has a different contribution to the perception of flavour, and sucrose has the most contribution to the sweetness taste, while glucose sweetness is only 55 to 60% of that of fructose or sucrose and has the presence of a slight bitterness [
25,
26]. The correlation between sensory panel results and sugar components indicated that the percentage of glucose in total sugars was negatively correlated to the
sweetness perception and positively correlated to
bitterness flavour (
p < 0.05), while sucrose showed an opposite result. In general, yellow papaya varieties ‘1B’ and ‘H13’ had lower TSS levels and higher %Glucose values, which leads to the bitterness flavour perceived by human.
A total of 14 volatiles were precisely identified from five papaya varieties through GCMS analysis. Of them, 12 volatiles have been identified in various papaya varieties from previous studies [
6,
8,
9,
10,
12,
27]. Among these 12 volatiles, linalool and benzyl isothiocyanate have been reported as the key impact odorants in papaya [
9]. The outcome of our study is consistent with this theory; the concentrations of linalool and benzyl isothiocyanate were abundant in all varieties without statistical differences (
p < 0.05). The remaining two compounds were citronellol and eucalyptol, which have been reported as fruit and floral constituents in grape, apple, peach, rose (
Rosa spp.), and Tulip (
Tulipa spp.) [
28,
29,
30], but to the best of our knowledge, these two volatiles have not been assigned in papaya fruit. The results from the GCMS analysis indicated that yellow-fleshed variety ‘1B’ had significantly higher concentrations of linalool oxide and terpinolene than the red-fleshed variety ‘Skybury’. Terpinolene is known for its pleasant sweet-pine, citrus aroma, which is considered the major VOC responsible for the characteristic flavour of mango cultivars [
31,
32]. Linalool oxide has a fresh floral citrus scent and is identified in most tropical fruits [
28,
32,
33]. Due to the low odour threshold of these two VOCs, they play key roles in ‘citrus’, ‘sweet’, and ‘floral’ aromas characteristic of papaya [
33]. This also corresponded to the correlation analysis between the sensory panel results and VOCs in our study, which showed that the concentrations of linalool oxide and terpinolene were significantly positively correlated to the
citrus aroma perceived by the trained panel (
p < 0.05). It is also interesting to find that these two VOCs were also positively correlated to the
musty off-note aroma but negatively correlated to the
fishy aroma. Further investigation is needed to determine the key VOCs that lead to the production of musty off-notes and fishy aroma. The multiple linear regression model was generated using three major biochemical variables: percentage of glucose in total sugar and concentrations of linalool oxide and terpinolene, which could explain over 99% of the variability in consumer liking. However, the consumer-liking score from our study has limitations. Most of the people involved in the consumer acceptability study were from the elder age group (50–80 years old), and over 50% of them had a papaya consumption habit in which they consumed at least one papaya per month; this could lead to bias in the means of the overall liking score. Moreover, only 40 panellists were included in this study, and broader consumer groups are required to generate a more reliable linear regression model.
Moreover, the genomic-based study to identify and validate genes related to sugars and VOCs synthesis metabolism pathways is required to assist in future DNA marker development to select premium papaya cultivars that align with consumer acceptance and demand. Under such conditions, a time efficiency with a high accuracy method to evaluate the candidate genes in a large papaya population is required. NanoString nCounter technology is a simple, robust, and highly reproducible method for the quantifying expression levels of multiple genes in a single reaction [
34,
35]. In our study, the nCounts from nanostring were firstly normalised using reference genes, then a
t-test was applied to determine the significant differences between samples. This comparison was allowed due to the lack of amplification steps involved during Nanostring analysis. Two new papaya varieties were included in this experiment: ‘Sunshine’ and ‘Holland’. ‘Sunshine’ is a new red-flesh papaya variety from the Australian breeding line. ‘Holland’ is also a red-flesh variety from Thailand, and this variety is also called ‘Plak Mai Lai’ in Thailand or ‘Sekaki’ elsewhere. In addition, ‘RB4’ was excluded from the gene expression analysis due to the shortage of fruit supplements.
Among the 10 candidate genes assessed, four of them were related to glucose production.
cpGPT2 is predicted to be a glucose 6-phosphate (Glc6P) transporter (GPT2), which functions in transporting Glc6p into plastids of heterotrophic tissues to start starch biosynthesis [
36]. The accumulation of GPT2 is positively correlated with the amount of total soluble sugars, especially glucose [
36,
37]. The expression level of
cpGPT2 was significantly higher in ‘1B’ than in ‘RB1’, ‘Skybury’, and ‘H13’, which corresponded to the sugar concentration results from our biochemical evaluation, in which ‘1B’ had the highest %Glucose level. Furthermore,
cpBGH3B,
cpBGLU42, and
cpBGLU31 are three predicted beta-glucosidase-related genes, which have been identified to release glucose from polysaccharides [
38,
39]. Among them,
cpBGLU42 and
cpBGLU31 shared similar expression patterns, while the yellow-flesh varieties ‘1B’ and ‘H13’ had higher expression levels than red-flesh varieties ‘RB1’ and ‘Skybury’.
cpBGH3B is predicted to be a beta-glucosidase BoGH3B-like gene. ‘RB1’ and ‘Holland’ had the highest expression levels of
cpBGH3B, while the lowest expression came from ‘Sunshine’. The DE analysis also demonstrated the significant difference between ‘Sunshine’ and ‘RB1’. Beta-glucosidase BoGH3B has been identified as an enzyme involved in the xyloglucan pathway, which functions in the degradation of cellulose to glucose in peach [
39,
40]. However, a BoGH3B-like gene has not been enzymatically characterised to date.
Except for glucose-related genes,
cpRFS2 is predicted to be a transglycosidase, which functions in a ping-pong reaction mechanism to catalyse the transfer of alpha-D-galactosyl-(1->3)-1D-myoinositol and sucrose to myo-inositol and raffinose [
41]. Raffinose is a trisaccharide composed of galactose, glucose, and fructose. ‘Skybury’ had significantly higher expression of
cpRFS2 than other varieties. However, the sugar determination outputs from the previous experiment showed ‘Skybury’ had the lowest concentrations of glucose and fructose but the highest sucrose. This could be the result of the involvement of other enzymes which function in catalysing glucose and fructose to sucrose more actively than
cpRFS2. In addition,
cpPFP is predicted to be involved in the subpathway of glycolysis that synthesizes D-glyceraldehyde 3-phosphate and glycerone phosphate from D-glucose [
42]. The amount of this enzyme activity was proved to be negatively correlated to the sucrose content in sugarcane [
43]. In our research, ‘H13’ had the lowest expression level in
cpPFP, while ‘Skybury’ and ‘Sunshine’ had higher expression levels. However, the sucrose content in ‘H13’ was lower than ‘Skybury’, which is opposite to the conclusion from sugarcane. Two sugar transport-related genes were also involved in this study, both of them functioned in mediating the uptake of hexoses by sugar/hydrogen symport [
44]. Previous research illustrated that
cpSTP14 specifically transports glucose and galactose, and
cpSTP1 shows a significant impact on sugar uptake during plant seedling development in
Arabidopsis thaliana [
44,
45,
46]. The hexoses included in papaya fruit are glucose and fructose [
47]. ‘RB1’ expressed significantly highest in
cpSTP14 but lowest in
cpSTP1, while ‘Holland’ expressed lowest in
cpSTP14 but highest in
cpSTP1. Although sugar transporter proteins are essential to plants’ sugar transport, growth, and development, the functional differentiation of these proteins in papaya is still unclear.
The remaining two genes are related to VOCs synthesis.
cpGES is predicted to be an (E, E)-geranyl linalool synthase involved in the terpenoid biosynthesis [
48]. The cooperation of (E, E)-geranyl linalool synthase initiates the catalysis of 4,8,12-trimethyltrideca-1,3,7,11-tetraene (TMTT), which is a floral odour compound identified from flowering plants such as Magnoliaceae family [
48,
49].
cpBEBT is predicted to be a benzyl alcohol O-benzoyl transferase, which is potentially involved in the production of volatile ester benzyl benzoate [
50]. Benzyl benzoate is an aroma compound with a fresh floral scent identified from flowers and fruit including
Clarkia breweri, cranberries, and papaya [
50,
51,
52]. Both genes are related to the synthesis of floral scents, while ‘Skybury’ was significantly highly expressed in
cpGES, and ‘H13’ was significantly highly expressed in
cpBEBT. ‘Holland’ expressed the lowest in two genes. From sensory panel results, ‘Skybury’ had a dominating floral flavour, and ‘H13’ had high citrus and sweet fruit aromas. It is worth adding that floral aroma is a new descriptor in the sensory panel testing attributes list, and the sensory profiles of ‘Holland’ and ‘Sunshine’ are also required in the next step of research. Furthermore, the aroma compounds TMTT and benzyl benzoate have not been validated from our GCMS study and are worth adding to the future volatile compound analysis.