Abstract
There are very few studies detailing the aroma, astringency, and flavor of rabbiteye blueberry [RAB (Vaccinium ashei)] fruit typically grown in the southeastern United States. The objectives were to investigate the rapid and qualitative solid-phase microextraction gas chromatographic–mass spectrometry volatile composition of several local RAB cultivars with an overall goal to build a database of possible flavor and aroma compounds. Volatile profiles were obtained in five Louisiana-grown RAB cultivars (Brightwell, Climax, Premier, Powder Blue, and Tifblue) assayed at four maturities. The method routinely captured 53 volatiles, including 12 aldehydes, six alcohols, 11 esters, four ketones, 17 terpenoids, one furan, and two aromatics. Of the 33 compounds considered important in blueberries, 17 were recovered in the RAB cultivars assessed. Herein, 10 compounds were recovered for the first time in blueberry (Vaccinium sp.) and five of those compounds were confirmed with standards [2-ethylfuran, (E)-2-pentenal, (Z)-dehydroxylinalool oxide, (E)-dehydroxylinalool oxide and 1,4-cineole]. In general, terpenoids and their subclass linalools were the most significant volatiles followed closely by esters, aldehydes, and then alcohols. Terpenoids and linalools displayed the greatest significant differences in ‘Powder Blue’ and ‘Premier’. Esters and aldehydes were the most significant compound classes based on cultivar effect per maturity in firm-ripe fruit. From the suite of 17 of the 33 important compounds in upright blueberry, 10 were recovered across the five cultivars at four maturities that displayed a high level of significance. These were linalool, methyl 3-methylbutanoate, 1,8-cineole, (E)-2-hexanal, (Z)-3-hexenal, (Z)-3-hexenyl acetate, limonene, hexyl acetate, hexanal, and α-terpineol. These data will be useful to evaluate aroma volatiles in RAB and changes in processed and value-added byproducts.
Heightened consumer awareness of the health benefits of consuming phytonutrient-rich fruits and aggressive marketing have resulted in expanding markets for blueberries, fruit juices, a large number of nutritional supplements, and value-added foods containing berries. Diets rich in blueberries deliver anti-inflammatory, anticarcinogenic, and antimutagenic components that help protect the brain, cardiovascular and central nervous system as well as reduce cancer, obesity, and type 2 diabetes (Basu et al., 2010; Kalt et al., 2007; Miller and Shukitt-Hale, 2012; Seeram, 2008; Soto-Vaca et al., 2012). “Superfruits” such as blueberries have become popular as scientists, marketers, and consumers push forward knowledge and demand for high-antioxidant, healthier diets.
Over the last decade in the southeastern United States there has been an almost 4-fold increase in blueberry production and use from 12.5 million kilograms in 1999 to 96.5 million kilograms in 2009 [U.S. Department of Agriculture (USDA), 2010a]. More important than production, however, during this same period, the national value of the used blueberry crop has increased over 3-fold from $154.0 million to $501.5 million (USDA, 2010b). In 2008, the southeastern U.S. blueberry industry had ≈9632 ha of rabbiteye blueberry and southern highbush blueberry [SHB (Vaccinium darrowii × V. corymbosum)] production with a market value of $72.8 million (Braswell et al., 2009). North American blueberry production acreage has increased significantly over the past few years in response to recent high blueberry demand and prices. Through 2005, the most common RAB cultivars grown in Louisiana and Mississippi were Climax, Premier, and Tifblue (Marshall et al., 2006). ‘Tifblue’ has been considered the most productive of all RAB and is generally the standard to which RAB are compared.
Volatile Compounds in Rabbiteye Blueberries
An extensive literature search indicates that volatile and semivolatile compounds in RAB appear in only three papers from the same laboratory for three cultivars (Horvat et al., 1983, 1996; Horvat and Senter, 1985). The majority of cultivars currently in production and on the market have little to no volatile information reported. Studies remain sparse regarding flavor volatiles and higher molecular weight (MW) bitter/astringent compounds, processing, and sensory attributes related to RAB (Makus and Morris, 1993). The first report documenting volatile data in SHB was recently published, yet there were general volatile differences indicated among SHB, RAB, and northern highbush blueberry [NHB (V. corymbosum)] types (Du et al., 2011). Over several years, many reports have indicated that volatile composition varies markedly between locations, within season, years, maturities, and cultivars (Baloga et al., 1995; Du et al., 2011; Horvat and Senter, 1985; Rohloff et al., 2009). The objective of this study was to investigate the qualitative volatile composition of several local RAB cultivars because berries are often culled and/or comingled for juicing and value-added products. Thus, a rapid qualitative solid-phase microextraction (SPME) method was used with the overall goal to build a database of possible flavor and aroma compounds in local blueberries that are consumed fresh and as processed products.
Materials and Methods
Fruit harvest and maturity designation.
Fresh hand-picked RAB were harvested from Louisiana during Summer 2010. Cultivars Premier and Powder Blue were harvested from a local U-pick operation in Bush, LA, on 19 June 2010. Cultivars Climax, Brightwell, and Tifblue were harvested from the Louisiana State University, Bob R. Jones Idlewild Research Station in Clinton, LA, on 24 June 2010. A stalk separator/blower with a conveyor belt (BEI Inc., South Haven, MI) was used to de-stalk and sort ripe ‘Tifblue’ fruit, and all other cultivars were manually sorted after harvest. Cultivars were analyzed as fresh fruit at four distinct maturities. Maturity was designated as mature green [MG (white/green)], light pink or “blue-red” (≈75% blue and ≈25% red), firm-blue or firm-ripe [FR (100% blue)] (Eck, 1986), and soft-blue or soft-ripe (SR). A replication of each treatment was a bag of randomly harvested fruit coming from several bushes for the same cultivar all harvested on the same date. Bushes for a specific cultivar were interspersed in separate rows with other cultivars (as a result of self-incompatibility pollination issues) and three separate samples/rows were used to attain randomization within the field. All maturities came from the same harvest (divided into maturity groups based on color) and therefore date of harvest is not confounded with maturity and not part of the definition of the experimental unit.
Juice sample preparation.
To derive a rapid headspace (HS) volatile juice sample, whole fresh berries were homogenized because MG and pink fruit would not readily juice as a result of low moisture content and firmness. Ten grams of berries of each color/firmness in each cultivar were placed in a 20-mL SPME vial and homogenized (Tissumizer SDT-1810; Tekmar, Cincinnati, OH) for ≈1.5 min using an 8-mm-diameter probe. After homogenization, 2.2 g of NaCl was added, vials were sealed and vortexed for 30 s, and placed on an autosampling tray at 1 °C on gas chromatography-mass spectrometry (GC/MS), similar to previously described methods used in fruit juice (Beaulieu, 2006; Beaulieu and Grimm, 2001) and wine (Torrens et al., 2004).
Headspace solid-phase microextraction gas chromatography–mass spectrometry.
Volatile sample vials (20 mL) were equilibrated 10 min by oscillation in a 35 °C autosampler (MPS2 XL; Gerstel, Baltimore, MD); then the HS was exposed to 1-cm 50/30 μm triple-fiber (divinylbenzene carboxen polydimethylsiloxane) SPME (StableFlex; Supelco, Bellefonte, PA) for 15 min at 35 °C. Vials were continuously swirled at 100 rpm during SPME adsorption before injection into GC/MS (HP6890/HP5973; Agilent Technologies, Santa Clara, CA) with a crosslinked 5% phenyl methyl silicone column with a 30 m × 0.25 mm × 0.25-μm film thickness (HP-5; Agilent Technologies). The injection port was operated in splitless mode and subjected to a pressure of 172.4 kPa of ultrahigh purity helium (99.9995%) for 1 min, and flow velocity was constant at 40 cm·s−1 for the remainder of the GC run. Initial oven temperature was 50 °C, held 1 min, ramped 5 °C·min−1 to 100 °C, then 10 °C·min−1 to 250 °C and held 9 min. The HP5973 quadrupole MS was operated in the electron ionization mode at 70 electron volts at 200 °C with a continuous scan from 33 to 300 mass to charge ratio (m/z).
Data analysis.
MS data were collected with HP ChemStation software (A.03.00; Agilent Technologies) and searched against the Wiley7th/NIST02 registry of mass spectral data (McLafferty, 2000). Library identification was confirmed by standard comparisons, GC retention time (RT), MS ion spectra, and a retention index (RI), per Table 1. An RI was calculated from the averaged RT from triplicated series of straight chain alkanes (C6 – C18). Integrated target ion response was examined carefully by selected unique target ions per compound, qualifying ion (Q-ion) ratio, and GC/MS grade (greater than 95% to 99.9% purity) standards purchased from commercial sources {Aldrich Chemical Co. [Milwaukee, WI], Bedoukian Research [Danbury, CT], Fluka [Buchs, Switzerland (now Sigma-Aldrich, St. Louis, MO)], Fisher [Pittsburgh, PA], JT Baker [Phillipsburg, PA], Mallinckrodt [St. Louis, MO], Sigma [Sigma-Aldrich], and Sigma-Aldrich Fine Chemicals, Flavors and Fragrances Products [St. Louis, MO]}, according to Table 1. Compound classes were defined and comprised of aldehydes (ALD), alcohols (OLS), esters (EST), ketones (ONE), furans (FUR), terpenoids (OID), and aromatics (ARO) according to Table 1. OID were further subdivided into terpenes (ENE) and linalools (LOO).
Volatile compounds qualitatively recovered by headspace solid-phase microextraction using divinylbenzene carboxen polydimethylsiloxane fibers and gas chromatography–mass spectrometry in five rabbiteye blueberry cultivars.
Statistical analysis.
The experimental design was a completely randomized design with a factorial treatment structure for five cultivars at four maturities and three replicates per treatment. Volatile measurements were taken on 53 compounds and these were analyzed individually and as aforementioned compound classes (Table 1). The treatments had a factorial structure of five cultivars and four maturities resulting in 20 treatments. However, compounds were not detected in all treatment combinations. Omission of zeros for no detection from the data (Table 2) was performed because transformations of the responses including too many zeros is faulty (Sokal and Rohlf, 2012) and constructing a nonparametric test for treatment differences will not work with abundant zeros (Conover, 1971). Analysis of variance (ANOVA) was used to examine cultivar and harvest treatment effects for each compound individually and in groups with SAS (Version 9.1; SAS Institute, Cary, NC).
Percent of observations and total number of possible observations for volatiles by class, cultivar, and maturity for the 53 compounds comprising seven classes of aroma compounds recovered by headspace solid-phase microextraction gas chromatography–mass spectrometry in five rabbiteye blueberry cultivars at four maturities.
For any given compound, there were often small measurements and few with much larger measurements, resulting in a mean distribution that was not symmetric but skewed to the right. Volatile measurements did not meet the ANOVA assumptions required for the dependent variable because they were often not normally distributed. Using a single volatile measurement (integrated unique Q-ion MS response) as a percent of the total volatiles for a given sample was considered to correct for this problem. However, percentages were not used because not all compounds were measured in each treatment combination and therefore there was no way to consistently obtain a denominator (i.e., total volatiles in all compounds) for every sample. Therefore, a log transformation was applied to raw volatile measurement to make it fit the assumption of normality required for ANOVA. This allowed distribution to approximate normality and balance the disproportionate high abundance vs. low abundance volatiles and to apply F-test inferences.
To measure how each volatile compound was affected by maturity and cultivar, three contrast statements (resulting in F tests) were constructed from the ANOVA to address the factorial treatment structure with missing treatment combinations (Table 2). Overall F value measures the combined effect of both maturity and cultivar on the volatile’s response (contrast H0: all treatments are equal). Maturity within cultivar F value measures treatment effects resulting from maturity only (contrast H0: all maturity treatments for cultivar Brightwell are equal + all maturity treatments for ‘Climax’ are equal + ..). Cultivar within maturity F value measures treatment effects resulting from cultivar only (contrast H0: all cultivar treatments for MG maturity are equal + all cultivar treatments for pink maturity are equal + ..).
Results and Discussion
Using StableFlex triple-fiber SPME, the five cultivars (Brightwell, Climax, Premier, Powder Blue, and Tifblue) sampled at four maturities contained 53 volatiles (Table 1). Seven of the 53 compounds were only tentatively identified based on MS ion fragmentation patterns because standards were not obtained. With FR and SR fruit, a suite of 32 to 39 and 36 to 40 compounds, respectively, in a given cultivar were recovered. With pink and MG fruit, respectively, a suite of 28 to 33 and 31 to 34 compounds in a given cultivar was recovered. In the five RAB cultivars analyzed with this method, the volatiles recovered were principally OID (17), ALD (12), EST (11), OLS (six), and ONE (four).
Ten compounds not previously reported in blueberry were recovered (Table 1). These were: 2-ethylfuran, (E)-2-pentenal, 2-methyl-4-pentenal (tentative), isopropyl 3-methylbutanoate (tentative), (Z)-dehydroxylinalool oxide, (E)-dehydroxylinalool oxide, 1,4-cineole (isocineole), δ-elemene (tentative), α-caryophellene [also known as α-humulene (tentative)], and γ-caryophellene (tentative). Five were confirmed by standards and designated “first known observation” and the remainder are “tentative.” (Z)-dehydroxylinalool oxide and (E)-dehydroxylinalool oxide (MS library hit “herboxide,” MW 152.237), as confirmed by RT with standards and MS ion fragmentation patterns, were present (Table 1). These could possibly be the same two compounds labeled as (Z)-dihydrolinalool oxide and (E)-dihydrolinalool oxide (MW 156.27) as tentatively identified previously (Du et al., 2011). However, both oxidized forms of linalool may exist because there are slightly different ring unfolding structures that cause differing MWs, RTs, and RIs.
Over 3 years sampling blueberry volatiles, the composition varied markedly between locations, within season, years, harvest dates, and between maturities for the same cultivar (data not shown). Four of the five cultivars were repeated in a separate year harvested in the same two locations. Because locations varied and our methods varied as analytical approaches were improved (as more standards were acquired), data were not combined. During these and other blueberry evaluations of three SHB vs. three RAB harvested twice from the same plot in a single season, several additional volatiles were recovered including: 2-heptanol, γ-terpineol, hotrienol, β-damasenone, (Z)-geranylacetone, (E)-geranylacetone, geraniol, p-menth-1-en-9-al, and p-cymen-8-ol and the apparently “important” geraniol (data not shown). Regardless, within- and between-field variability yields incredible volatile differences, as documented in the literature (Du et al., 2011), and this is not easy to overcome. Hence, the use of other berries, like grapes (Vitis sp.), used to prepare juices to make wines, is considered “an art,” not a science (Lund and Bohlmann, 2006).
Because dissimilarity in volatile trends has been reported in blueberries (Baloga et al., 1995; Du et al., 2011; Horvat and Senter, 1985; Rohloff et al., 2009), data were analyzed by omitting zeros for no detection. We purposely chose to highlight compounds detected, treating their presence, as opposed to considering their absence, as a zero. This effectively precluded occasional absence from negatively affecting compound distribution in the analysis. The typical two-way ANOVA with main effects and interactions could be misleading for compounds that had missing treatment combinations (Conover, 1971; Sokal and Rohlf, 2012). For example, performing the ANOVA with zeros included resulted in the highest overall F values generated in the FUR (one compound) and EST (11 compounds) classes (data not shown). Only 2-ethylfuran was recovered in very few treatments (Table 2), and the esters tended to be either absent while immature or highly abundant when FR or SR, and individual esters also varied by maturity and slightly by cultivar (Table 2). ALD and OID ranked as the third and fourth most significant volatile classes, respectively (in the ANOVA with zeros included, data not shown). As a result of zeros, low df, and variability in individual compounds, these appraisals were not realistic. Subsequently, ANOVA F tests were obtained for: 1) overall treatment effect (including effects of both cultivar and maturity treatments) where df indicates number of treatments minus one (Tables 3 and 5); 2) effects of maturity treatments for each cultivar group (Tables 4 and 6) and combined into one joint F test for maturity treatment effects within all cultivar groups (Tables 3 and 5); and 3) effect of cultivar treatments for each maturity group (Tables 4 and 6) combined into one joint F test for cultivar/treatment effects within all maturity groups (Tables 3 and 5).
Analysis of variance F values indicating differences in maturity/cultivar treatments in volatile compound classes recovered in five rabbiteye blueberry cultivars at four maturities using headspace solid-phase microextraction with divinylbenzene carboxen polydimethylsiloxane fibers and gas chromatography–mass spectrometry.
F values for additional analysis of variance tests to assess impact between compound classes recovered by headspace solid-phase microextraction gas chromatography–mass spectrometry in five rabbiteye blueberry cultivars at four maturities.
Analysis of variance F values indicating differences in maturity/cultivar treatments for individual headspace solid-phase microextraction gas chromatography–mass spectrometry volatile compound recovered in five rabbiteye blueberry cultivars at four maturities.
F values for additional analysis of variance tests to assess impact between individual compounds recovered by headspace solid-phase microextraction gas chromatography–mass spectrometry in five rabbiteye blueberry cultivars at four maturities.
Most volatiles displayed highly significant F values (P < 0.001) for all three contrast statement tests. This is similar to previously published quantitative volatile appraisal in SHB where an ANOVA was used (Du et al., 2011). However, herein, compounds were ranked to sensitivity to maturity and cultivar effects based on the magnitude of the F values. Table 2 is a summary of data used in the analysis by indicating all treatments and combinations, thereof reported as the percentage of observations delivering results (i.e., volatile measurements > 0 or y > 0). For example, in the ALD compound group, ‘Tifblue’ at the pink stage had 100% observations with y > 0 (based on 12 ALD × three replications = 36 observations) and 92% for ‘Brightwell’ at MG, where compounds were not recovered in some treatments and combinations (e.g., 33 of 36 with y > 0).
Variation by cultivar.
When overall F values were sorted to display the greatest significance per compound class, OID and the subclass LOO were most significant followed closely by EST, ALD, and then OLS (Table 3). As expected based on previous maturity-related cantaloupe (Cucumis melo var. reticulatus) volatile data using a similar SPME method (Beaulieu, 2006), maturity differences delivered the most significance in compound classes generally believed to be aroma-important in blueberry such as OID, LOO, EST, OLS, and ARO. In general, blueberries have appreciable terpenoid content (Du et al., 2011; Hirvi and Honkanen, 1983a; Horvat et al., 1983; von Sydow and Anjou, 1969). Several compounds recovered and denoted in the literature are terpenoids and terpenoid alcohols (such as α-terpineol and linalool), which are common flavor and aroma compounds in blueberry (Du et al., 2011; Horvat and Senter, 1985; Parliment and Kolor, 1975; von Sydow and Anjou, 1969) and other berries such as grapes (Marais, 1983). As a result of diversity and range of terpenoid concentrations reported by cultivar in SHB, it has been speculated that terpenoids might be the group of compounds that help determine the characteristic aroma per cultivar (Du et al., 2011). OID also yielded the second greatest significant difference based on cultivar, which was very close in ranking compared with ALD followed by EST (Table 3). Little emphasis is placed on FUR, because only one compound was recovered and only present in some cultivars in MG and pink stages (Table 2). ONE also had a more significant cultivar than maturity effect. Although a subclass of OID, the ENE category (non-oxygenated terpenes) displayed more significance per cultivar (Table 3). This illustrates again the relative importance of OID, LOO, and one terpene alcohol (α-terpineol), which had the opposite and most pronounced significance related with maturity.
Aldehydes were reported and recovered long ago in lowbush blueberry [LOW (Vaccinium angustifolium and Vaccinium myrtillus)] and considered as important aroma components in highbush types (Hirvi and Honkanen, 1983a; Rohloff et al., 2009; von Sydow et al., 1970; von Sydow and Anjou, 1969). ALD displayed the most significance by cultivar effect. ALD also comprised the vast majority of volatiles (54% to 77%) in SHB, even using an extraction method that was not considered oxidative (Du et al., 2011). Although methods used herein and in previous RAB reports may favor oxidation (Horvat et al., 1983; Horvat and Senter, 1985), aldehydes dominated the volatile profiles. Unlike the methods of Du et al. (2011) where salt was added during sample preparation to reduce oxidation, we added salt only to drive volatiles into the HS, after blending. Many C6 and C9 unsaturated aldehydes are generally considered to be products of fatty acid oxidation. Nonetheless, appreciable and similar aldehydes in the cultivars of RAB tested here were recovered. Like in cucumber (Cucumis sativus) and tomato (Solanum lycopersicum), C6 aldehydes and alcohols seem to be flavor-important (Parliment, 1977), regardless of whether they are generated by oxidation (Fleming et al., 1968; Parliment and Kolor, 1975; Riley and Thompson, 1998).
Additional F values were used to measure the specific effect of maturity within cultivar and presented separately for each cultivar as well as to measure cultivar effect within maturity (Table 4). The most differences in compound classes based on maturity effects per cultivar occurred in ‘Climax’ and ‘Tifblue’ compared against ‘Brightwell’, ‘Powder Blue’, and ‘Premier’ (Table 4). This could be because these former two cultivars do not share the same relative proportion of the original four land races as the latter three cultivars that share 25% background from ‘Ethyl’, ‘Myers’, ‘Clara’, and ‘Black Giant’. This is logical as a result of common lineage tracing back to the same four parental land races (Aruna et al., 1995; Lyrene, 1983; Rowland et al., 2010). Diploid parental lineages from the putative progenitor species of cultivated LOW, NHB, and RAB contain many of the same aroma volatile patterns among species and the groupings of nine quantified flavor compounds correlated with the taxonomic allozyme groupings (Baloga et al., 1995). This coincides well with principle component analysis of the original seven progenitors of cultivated U.S. diploid blueberry (Bruederle and Vorsa, 1994). However, ‘Brightwell’ and ‘Powder Blue’ are siblings (Aruna et al., 1995) and the volatile profile for compound classes based on maturity effects per cultivar in ‘Brightwell’ was rather dissimilar compared with ‘Powder Blue’ and ‘Premier’. A rational explanation for this trend eludes interpretation without several seasons and/or locations compared.
F values again indicated the relative importance and significance of OID (Table 4). OID and LOO displayed the greatest significant differences in maturity per cultivar in ‘Powder Blue’ and ‘Premier’. Herein, ARO variation over maturity was most significant in the cultivar Powder Blue, yet the ARO weight or importance is likely not as strong because only two compounds (benzaldehyde and phenylacetaldehyde) comprised this class. At lower concentrations, these aromatic aldehydes are generally associated with positive aroma attributes such as almond, bilberry, floral, hyacinth, sweet, and fruity-like (Latrasse, 1991). OLS and ALD variation over maturity was most pronounced in ‘Tifblue’. Aside from the difference in FUR (2-ethylfuran), ‘Powder Blue’ and ‘Premier’ displayed rather similar trends.
Variation by maturity.
EST and ALD were the most significant classes of compounds based on cultivar effect per maturity in FR fruit (Table 4). This is when berries are at their horticultural peak and just attaining optimal eating quality. These compound classes have specific compounds that are flavor-important in all blueberry types, and their peak at FR coincides nicely with the horticultural optimum stage for eating quality. Similarly, OID compounds were the next most significant across ripeness from pink through FR to SR. The significance of LOO compounds also peaked at the FR stage. This finding of high relative significance again lends credence to the likely importance of terpenoids and terpene alcohols within OID and LOO in RAB aroma, similar to other blueberry types.
Terpenes, C6 unsaturated aldehydes, and unsaturated alcohols have also been reported to be the predominant compounds identified in the volatile extracts of RAB (Horvat et al., 1983, 1996; Horvat and Senter, 1985). Each cultivar displayed an increase in mid- to high MW volatiles (linalool, 154.25 MW and larger) with increasing maturity, whereas the lower MW compounds like hexanol (102.17 and lower MW) decreased (Horvat and Senter, 1985). Similar trends regarding linalool and (Z)-linalool oxide were found, and the whole OID class generally increased with increasing maturity (data not shown). OID compounds displayed the greatest significant F values in FR fruit (Table 4). According to weighted F values, (E)-2-hexenal, (E)-linalool oxide, and α-terpineol displayed a greater significance (magnitude) in FR and SR compared with MG and pink fruit (Table 6).
On a relative percentage basis (data not shown), EST (11 compounds) were exceedingly low at MG and pink stages and increased massively, similar to cantaloupe esters (Beaulieu, 2006), in all five cultivars with SR having the greatest integrated ion content, oftentimes 5- to 10-fold relative increases from FR to SR. However, on data transformation, the means only reflected significant increases in EST in FR and SR fruit in ‘Powder Blue’ and ‘Premier’. OID (17 compounds) were generally lowest in MG fruit and increased steadily through the SR stage in all five cultivars (data not shown). FUR (one compound) generally declined from MG to pink and was not observed in FR and SR maturity stages, and ONE (four compounds) displayed no clear trends (data not shown).
Compound variation by maturity and cultivar.
Based on consensus from only upright blueberry (NHB, RAB, and SHB) volatile literature, there are 33 compounds recovered in substantial levels or believed to have aroma impact (J.C. Beaulieu, unpublished data). These are: ethyl acetate, (E)-2-hexanal, (Z)-3-hexenal, hexanal, heptanol (E,Z)-2,6-nonadienal, (E)-2-hexenol, (Z)-2-hexenol, (Z)-3-hexenol, hexanol, ethyl butanoate, methyl 2-methylbutanoate, methyl 3-methylbutanoate, ethyl 2-methylbutanoate, ethyl 3-methylbutanoate, (Z)-3-hexenyl acetate, (E)-3-hexenyl acetate, hexyl acetate, linalool, p-cymene, cinerolone, β-ionone, 1,8-cineole, terpene-4-ol, limonene, α-terpineol, carveol, geraniol, eugenol, geranylacetone, 2-undecanone, 2-butanone, and damascenone (Du et al., 2011; Hirvi and Honkanen, 1983a, 1983b; Horvat et al., 1983; Horvat and Senter, 1985; Parliment, 1977; Parliment and Kolor, 1975; Zhang et al., 2009). From that group of compounds, 17 were recovered here by SPME across the five cultivars at four maturities [ethyl acetate, (E)-2-hexanal, (Z)-3-hexenal, hexanal, (E)-2-hexenol, (Z)-3-hexenol, hexanol, methyl 3-methylbutanoate, ethyl 2-methylbutanoate, ethyl 3-methylbutanoate, (Z)-3-hexenyl acetate, hexyl acetate, linalool, p-cymene, 1,8-cineole, limonene, and α-terpineol]. Presentation of individual compounds (Tables 5 and 6) will therefore be focused within this group that delivered significant overall and treatment combination effects in the ANOVA (Table 5). Effectively, this caused ethyl 3-methylbutyrate, ethyl 2-methylbutyrate, and (Z)-3-hexenol to be removed from consideration. Once compound F values decreased and were predominantly insignificant, compounds were not tabled. Because OID and LOO classes were demonstrated to be the most F significant compound classes, compounds within those classes with highly significant F values are also tabled.
Only one laboratory has reported volatiles in the hexaploid (2n = 6x = 72) RAB and they found the major volatiles were ethyl acetate, p-cymene, hexanol, (Z)-2-hexenol, heptanol, cinerolone, β-ionone, terpene-4-ol, 2-undecanone, α-terpineol, carveol, nerol, and eugenol (Horvat et al., 1983; Horvat and Senter, 1985). Although volatile constituents are different among different species of blueberry, (E)-2-hexenal (E)-2-hexenol, (Z)-3-hexenol, linalool, and geraniol have been found in most types and are considered typical aroma compounds for blueberry (Du et al., 2011; Hirvi and Honkanen, 1983a, 1983b; Horvat et al., 1996; Horvat and Senter, 1985; Parliment and Kolor, 1975). Aside from geraniol (not recovered or resolved herein), the other four compounds were isolated but (Z)-3-hexenol was not significant. Surprisingly, a patent claims that combinations of just four compounds [linalool, (E)-2-hexanal, (E)-2-hexenol and (Z)-3-hexenol] mimic blueberry flavor (Parliment, 1977). However, this was based on knowledge obtained for NHB cultivars, which are tetraploids (2n = 4x = 48). Furthermore, within the SHB types with a markedly diverse parental lineage (ploidy varies as a result of complicated interbreeding among diploids, tetraploids, and hexaploids), there were often marked volatile differences between cultivars (Du et al., 2011).
Individual compounds in Tables 5 and 6 have been ranked by their overall F values. In the top 16 highest ranked compounds, 10 of the compounds (bold font) are considered aroma-important or highly abundant in blueberry (Table 5). Hence, the following previously reported compounds were determined to be significant based on their F values: linalool, methyl 3-methylbutyrate, 1,8-cineole, (E)-2-hexenal, (Z)-3-hexenal, (Z)-3-hexenyl acetate, limonene, hexyl acetate, hexanal, and α-terpineol. Compounds recovered that displayed significance that may contribute to aroma in RAB, which have not previously been delineated as such in the literature, were: methyl butyrate, 2-methyl-4-pentenal, 2-heptanone, (E,E)-2,4-hexadienal, and (Z)-linalool oxide. Linalool and geraniol were considered to be major aroma active terpenoids in RAB (Horvat and Senter, 1985) and have also been reported to contribute to floral flavor (Rychlik et al., 1998). Aldehydes [e.g., hexanal and (E)-2-hexenal] have been commonly reported in several NHB blueberry (di Cesare et al., 1999; Hirvi and Honkanen, 1983a; Parliment and Kolor, 1975) and denoted in a patented blueberry (NHB) flavor (Parliment, 1977). Hexanal and (E)-2-hexenal were the major “green compounds,” and (Z)-3-hexenol and (E)-2-hexenol were believed to be contributors to aroma as well (Horvat and Senter, 1985).
(E)-2-hexenal was a predominant compound reported in several SHB cultivars (Du et al., 2011) and had very high F values in ‘Brightwell’, ‘Climax’, and ‘Powder Blue’ (Table 6). However, the other generally dominant ALD, (Z)-3-hexenal, was only prevalent in ‘Brightwell’, ‘Climax’, and especially so in ‘Tifblue’. Six carbon compounds such as hexanal, (Z)-3-hexenal, and (E)-2-hexenal have characteristic green/grassy odors, octanal and nonanal have citrus-like odor, (E,Z)-2,6-nonadienal has a characteristic cucumber odor, and (E,E)-2,4-nonadienal and (E,E)-2,4-decadienal have fatty/waxy odors (Rychlik et al., 1998; Schieberle et al., 1990). In RAB, (E)-2-hexenal, (E)-linalool oxide, α-terpineol, and β-caryophyllene decreased with increasing maturity, whereas linalool and (Z)-linalool oxide increased with maturity (Horvat et al., 1996; Horvat and Senter, 1985). With the exception of geraniol (not recovered) and (Z)-3-hexenol, the remaining compounds yielded significant F values regarding overall magnitude in their F tests and both maturity and cultivar interaction effects (Tables 5 and 6).
The weight of compounds significance can also be viewed based on the df. The higher the df (e.g., 19), the more confidence in the delivered F value and significant contribution of a compound regarding a maturity or cultivar effect. This is largely the result of inference being based on more information because more treatments are being represented. However, this does not negate the fact that an individual compound may be highly significant even when not recovered in certain cultivars or maturities. For example, β-myrcene displayed the greatest overall significance in addition to maturity and cultivar interaction effects (Table 5). However, it was only recovered in half the treatment combinations (Tables 2 and 6). β-Myrcene was not recovered in MG or pink fruit in any cultivar, yet this illustrates how one compound with pleasant or desirable aroma attributes (e.g., grape, fruity, peach, must, rosy, citrus, mango) may indeed have important significance, especially because it only appeared in FR and SR fruit, similar in nature to non-oxygenated terpenes recovered in FR and SR fresh-cut mangos [Mangifera indica (Beaulieu and Lea, 2003)]. The same exact relationship occurred with methyl butanoate, which has been ascribed positively as well; e.g., as fruity, ethereal, apple, sweet (Beaulieu, 2005).
Compounds displaying the highest overall significance with all or almost all treatment combinations were generally in the OID and ALD classes (Tables 2 and 5). Again, linalool was the most significant OID and it conferred high significance across maturity within cultivar (Table 5) and displayed significant differences in ‘Climax’, ‘Powder Blue’, and ‘Premier’ (Table 6). Linalool also displayed the most significant differences in cultivar by maturity in FR and SR, similar to several other OID and ALD compounds denoted in Table 6. In general, compounds generally had greater significance for FR and SR maturity, pooled across cultivar, than MG or pink. Aside from two of the top three compounds ranked by overall F value (β-myrcene and methyl butyrate), the relative ranking (weight) of maturity effects per cultivar was significantly greater than cultivar differences pooled across maturity (Table 6). As noted earlier, this is somewhat expected because inbreeding depression has occurred in RAB as a result of a limited breeding background used during cultivar development (Aruna et al., 1995; Lyrene, 1981, 1983; Rowland et al., 2010; Stringer et al., 2009). Most current RAB cultivars originated from a narrow genetic base involving only four natively collected cultivars (i.e., Ethyl, Myers, Clara, and Black Giant) and their continued parental use resulted in inbreeding depression (Lyrene, 1983). For example, one of the standard cultivars assessed, Tifblue (‘Ethel’ × ‘Clara’) has common parental crosses also resulting in ‘Brightwell’ (‘Tifblue’ × ‘Menditoo’), ‘Powder Blue’ (‘Tifblue’ × ‘Menditoo’), and ‘Premier’ (‘Homebell’ × ‘Tifblue’) (Aruna et al., 1993; Eck, 1988; Okie, 2004). Nonetheless, minor volatile differences and profile dissimilarity resulting from growing location, year, harvest dates, and cultivar are common to several fruits, including SHB and RAB (Du et al., 2011; Horvat and Senter, 1985).
Across the five cultivars, based on relative percentage recovery (data not shown), the relative compound abundance in FR fruit was ALD (75.2%) > OLS (11.2%) > EST (6.9%) > OID (4.3%) > LOO (2.7%). On the other hand, in SR fruit, order of compound abundance was generally ALD (52.1%) > EST (32.3%) > OLS (8.2%) > OID (5.7%) > LOO (3.7%). Nonetheless, in general, there was a very similar pattern of dominant volatiles recovered in all five cultivars with 11 of the top 15 compounds (based on relative percentage) being similar: (E)-2-hexenal, ethyl acetate, hexanal, acetaldehyde, hexanol, (E,E)-2,4-hexadienal, 1-penten-3-ol, linalool, (Z)-3-hexenal, (E)-2-hexenol, and methyl 3-methylbutyrate (data not shown). The five most abundant compounds in most SR and FR cultivars were ethyl acetate, (E)-2-hexenal, acetaldehyde, hexanal, and (E,E)-2,4-hexadienal. The top 15 compounds in FR fruit comprised 97.5%, 97.1%, 96.6%, 96.5%, and 96.1% of the relative volatile contribution in cultivars Brightwell, Premier, Tifblue, Climax, and Powder Blue, respectively.
Herein, ANOVA provided F values measuring treatment differences for several compounds and compound groups in selected cultivars of RAB. Because those volatiles displaying no response were omitted rather than treated as zeros, for some compounds or compound groups, this resulted in entire treatments being eliminated. Given a factorial treatment structure of five cultivars × four maturity groups, a missing treatment combination resulted in a missing cell. Subsequently, ANOVA using the typical two-way with interactions sources of variance was not appropriate for all compounds. As a result of volatiles occasionally not being recovered, it was not wise to pool data and attempt to deliver horticultural-like quantitative or statistical appraisals. Fortunately, this was not the intended purpose for our analysis.
Industrially, many juices are made from concentrates and those fruit are often culls, residuals, and off-grades that are collected, refrigerated, and eventually bulk-transported to a limited number of concentrating plants nationwide. Subsequently, it is not uncommon to expect a vast mixture of cultivars, quality, and volatile attributes emanating from immature, ripe, overripe, and possibly fermenting product, which gets concentrated. Subsequently, because aldehydes are commonly reported as being important volatiles in blueberry, the method herein possibly deriving excess aldehydes is not considered undesirable. Indeed, ALD compounds were significant across cultivar and maturity. However, the data here indicate more importantly, similar to (Du et al., 2011), that OID and LOO compounds, along with ALD, likely confer the aroma importance based on recovery significance, trends, and previous literature. A sensory lexicon was recently published that will be used in further analysis to help gauge the relative importance of these compounds in blueberries and processed products (Bett-Garber and Lea, 2013). The qualitative SPME volatile data method reflecting a diverse array of mainly authenticated compounds recovered in RAB will be a useful database geared toward viewing changes in blueberry types, processed juice, byproducts, and value-added blueberry products.
Literature Cited
Aruna, M., Austin, M.E. & Ozias-Akins, P. 1995 Randomly amplified polymorphic DNA fingerprinting for identifying rabbiteye blueberry (Vaccinium ashei Reade) cultivars J. Amer. Soc. Hort. Sci. 120 710 713
Aruna, M., Ozias-Akins, P., Austin, M.E. & Kochert, G. 1993 Genetic relatedness among rabbiteye blueberry (Vaccinium ashei) cultivars determined by DNA amplification using single primers of arbitrary sequence Genome 36 971 977
Baloga, D.W., Vorsa, N. & Lawter, L. 1995 Dynamic headspace gas chromatography–mass spectrometry analysis of volatile flavor compounds from wild diploid blueberry species, p. 235–247. In: Rouseff, R.L. and M.M. Leahy (eds.). Fruit flavors: Biogenesis, characterization, and authentication. No. 596. Amer. Chem. Soc., Washington, DC
Basu, A., Rhone, M. & Lyons, T.J. 2010 Berries: Emerging impact on cardiovascular health Nutr. Rev. 68 168 177
Beaulieu, J.C. 2005 Within-season volatile and quality differences in stored fresh-cut cantaloupe cultivars J. Agr. Food Chem. 53 8679 8687
Beaulieu, J.C. 2006 Volatile changes in cantaloupe during growth, maturation and in stored fresh-cuts prepared from fruit harvested at various maturities J. Amer. Soc. Hort. Sci. 131 127 139
Beaulieu, J.C. & Grimm, C.C. 2001 Identification of volatile compounds in cantaloupe at various developmental stages using solid phase microextraction J. Agr. Food Chem. 49 1345 1352
Beaulieu, J.C. & Lea, J.M. 2003 Volatile and quality changes in fresh-cut mangos prepared from firm-ripe and soft-ripe fruit, stored in clamshell containers and passive MAP Postharvest Biol. Technol. 30 15 28
Bett-Garber, K.L. & Lea, J.M. 2013 Development of flavor lexicon for freshly pressed and processed blueberry juice J. Sens. Stud. 28 161 170
Braswell, J.H., Stringer, S., Sampson, B. & Ingram, D. 2009 Establishment and maintenance of blueberries. Mississippi State Univ. Ext. Serv. Publ. No. 1758
Bruederle, L.P. & Vorsa, N. 1994 Genetic differentiation of diploid blueberry, Vaccinium sect. Cyanococcus (Ericaceae) Syst. Bot. 19 337 349
Conover, W.J. 1971 Practical nonparametric statistics. Wiley, New York, NY
di Cesare, L.F., Nani, R., Proietti, M. & Giombelli, R. 1999 Volatile composition of the fruit and juice of some blueberry cultivars grown in Italy Industrie Alimentari 38 277 282
Du, X., Plotto, A., Song, M., Olmstead, J. & Rouseff, R.L. 2011 Volatile composition of four southern highbush blueberry cultivars and effect of growing location and harvest date J. Agr. Food Chem. 59 8347 8357
Eck, P. 1986 Blueberry, p. 75–85. In: Monselise, S.P. (ed.). CRC handbook of fruit set and development. CRC Press, Boca Raton, FL
Eck, P. 1988 Blueberry science. Rutgers Univ. Press, New Brunswick, NJ
Fleming, H.P., Cobb, W.Y., Etchells, J.L. & Bell, T.A. 1968 The formation of carbonyl compounds in cucumbers J. Food Sci. 33 572 576
Hirvi, T. & Honkanen, E. 1983a The aroma of blueberries J. Sci. Food Agr. 34 922 996
Hirvi, T. & Honkanen, E. 1983b The aroma of some hybrids between high-bush blueberry (Vaccinium corymbosum, L.) and bog blueberry (Vaccinium uliginosum, L.) Z. Lebensm. Unters. Forsch. 176 346 349
Horvat, R.J., Schlotzhauer, W.S., Chortyk, O.T., Nottingham, S.F. & Payne, J.A. 1996 Comparison of volatile compounds from rabbiteye blueberry (Vaccinium ashei) and deerberry (V. stamineum) during maturation J. Essential Oil Res. 8 645 648
Horvat, R.J. & Senter, S.D. 1985 Comparison of the volatile constituents of rabbiteye blueberries during ripening J. Food Sci. 50 429 436
Horvat, R.J., Senter, S.D. & Dekazos, E.D. 1983 GLC-MS analysis of volatile constituents in rabbiteye blueberries J. Food Sci. 48 278 279
Kalt, W., Joseph, J.A. & Shukitt-Hale, B. 2007 Blueberries and human health: A review of current reseach J. Amer. Pomol. Soc. 61 151 160
Latrasse, A. 1991 Fruits III, p. 329–387. In: Maarse, H. (ed.). Volatile compounds in foods and beverages. Marcel Dekker, New York, NY
Lund, S.T. & Bohlmann, J. 2006 The molecular basis for wine grape quality: A volatile subject Science 311 804 805
Lyrene, P.M. 1981 Recurrent selection in breeding rabbiteye blueberries (Vaccimium ashei Reade) Euphytica 30 505 511
Lyrene, P.M. 1983 Inbreeding depression in rabbiteye blueberries HortScience 18 226 227
Marais, J. 1983 Terpenes in the aroma of grapes and wines: A review S. African J. Enol. Viticult. 4 49 58
Makus, D.J. & Morris, J.R. 1993 Highbush vs. rabbiteye blueberry: A comparison of fruit quality J. Food Qual. 16 417 436
Marshall, D.A., Spiers, J.M. & Braswell, J.H. 2006 Splitting severity among rabbiteye (Vaccinium ashei Reade) blueberry cultivars in Mississippi and Louisiana Intl. J. Fruit Sci. 6 77 81
McLafferty, F. 2000 Wiley registry of mass spectral data. 7th Ed. Wiley 7th/NIST02 Software. Palisade Corp., Newfield, NY
Miller, M.G. & Shukitt-Hale, B. 2012 Berry fruit enhances beneficial signaling in the brain J. Agr. Food Chem. 60 5709 5715
Okie, W.R. 2004 Register of new fruit and nut varieties: List 42 HortScience 39 1509 1523
Parliment, T.H. 1977 Blueberry flavor. Patent US 4041185. U.S. Patent Trademark Office, Washington, DC
Parliment, T.H. & Kolor, M.G. 1975 Identification of the major volatile components of blueberry J. Food Sci. 40 762 763
Riley, J.C.M. & Thompson, J.E. 1998 Ripening-induced acceleration of volatile aldehyde generation following tissue disruption in tomato fruit Physiol. Plant. 104 571 576
Rohloff, J., Nestby, R., Nes, A. & Martinussen, I. 2009 Volatile profiles of European blueberry: Few major players, but complex aroma patterns Latvian J. Agron. 12 98 103
Rowland, L.J., Ogden, E.L. & Ehlenfeldt, M.K. 2010 EST-PCR markers developed for highbush blueberry are also useful for genetic fingerprinting and relationship studies in rabbiteye blueberry Sci. Hort. 125 779 784
Rychlik, M., Schieberle, P. & Grosch, W. 1998 Compilation of odor thresholds, odor qualities and retention indices of key food odorants. Deutsche Forschungsanstalt fur Lebensmittelchemie and Institut fur Lebensmittelchemie der Technischen Universitat Munchen, Garching, Germany
Schieberle, P., Ofner, S. & Grosch, W. 1990 Evaluation of potent odorants in cucumbers (Cucumis sativus) and muskmelons (Cucumis melo) by aroma extract dilution analysis J. Food Sci. 55 193 195
Seeram, N.P. 2008 Berry fruits for cancer prevention: Current status and future prospects J. Agr. Food Chem. 56 630 635
Sokal, R.R. & Rohlf, F.J. 2012 Biometry: The principles and practice of statistics in biological research. Freeman, New York, NY
Soto-Vaca, A., Gutierrez, A., Losso, J.N., Xu, Z. & Finley, J.W. 2012 Evolution of phenolic compounds from color and flavor problems to health benefits J. Agr. Food Chem. 60 6658 6677
Stringer, S.J., Draper, A.D., Spiers, J.M. & Marshall, D.A. 2009 Potential new blueberry cultivars for the Gulf Coast region of the U.S Acta Hort 810 87 91
Torrens, J., Riu-Aumatell, M., Lopez-Tamames, E. & Buxaderas, S. 2004 Volatile compounds of red and white wines by headspace–solid-phase microextraction using different fibers J. Chromatogr. Sci. 42 310 316
U.S. Department of Agriculture 2010a U.S. blueberry industry. U.S. blueberry production and utilization (cultivated and wild), selected states, 1980–2009 (table02.xls). 10 Dec. 2013. <http://usda.mannlib.cornell.edu/MannUsda/viewDocumentInfo.do?documentID=1765>
U.S. Department of Agriculture 2010b U.S. blueberry industry. Cultivated blueberries: Commercial acreage, yield per acre, production, and season-average grower price in the United States, 1980–2009 (table08.xls). 10 Dec. 2013. <http://usda.mannlib.cornell.edu/MannUsda/viewDocumentInfo.do?documentID=1765>
von Sydow, E., Andersson, J., Anjou, K. & Karlsson, G. 1970 The aroma of bilberries (Vaccinium myrtillus L.). II. Evaluation of the press juice by sensory methods and by gas chromatography and mass spectrometry Lebensmittle-Wissenschaft und Technologie 3 11 17
von Sydow, E. & Anjou, K. 1969 The aroma of bilberries (Vaccinium myrtillus L.). I. Identification of volatile compounds Lebensmittle-Wissenschaft und Technologie 2 78 81
Zhang, C.Y., Li, Y.D., Zhang, Z.D., Liu, H.G., Wu, L. & Wang, J.Y. 2009 Changes of aroma components of different developmental fruits of high-bush blueberry Scientia Agr. Sinica 42 3216 3223