sign in sign up
<<
Home , Bitter orange, Blood orange, Mangshanyegan, Ponkan

b summa- 4 et al. ) and one satsuma Several studies have 5 although most of them The aroma volatiles of sweet 13–18 ), three sour oranges ( 2,4,5 Citrus clementina Elizabeth A Baldwin b Citrus reticulata as well as in fresh fruits, 6–12 ), one ( Correspondence to: FG Gmitter, Citrus Research andsity Education of Center, , Univer- Lake Alfred, FL, USA. E-mail: fgmitter@ufl.edu Citrus Research and Education Center, University of Florida, Lake Alfred, FL, USA Horticultural Research Laboratory, ARS, USDA, Fort Pierce, FL, USA Southeastern Fruit and Tree NutUSA Research Laboratory, ARS, USDA, Byron, GA, Taste, aroma, color and texture are the most important qual- c a ∗ b Anne Plotto, ity attributes of citrus fruits.ture Citrus of aroma monoterpenes, sesquiterpenes, constituents alcohols, are aldehydes,esters, a acids, ketones mix- and other volatiles. and grapefruitthey have are major been commercial juice sources. extensively Overhave 300 studied been aroma volatiles reported because for fresh . been performed onessence, aroma volatiles in mandarin oil and have studied only one or a few genotypes. Tietel rized 37 consensus aroma volatiles in mandarin juice and nine c Citrus limonia 5years ≥ )produced ), one ( a* Citrus paradisi Chunxian Chen, The identification of molec- b 3 Citrus sinensis ), mandarins (or in the : 1124–1131 www.soci.org © 2017 Society of Chemical Industry ) and ( Jinhe Bai, 98 Citrus sinensis Traditional citrus breeding based on crossing and a ). 2018; 1,2 aroma; fruit quality; lime; mandarin; sour orange; volatiles ), one (

Citrus reticulata

Citrus unshiu ular markers linked togenetic citrus improvement and fruit the taste release of and new superior aroma varieties. can facilitate selection tends to be costly because hybrids usually take before fruiting, and the capacity to grow hybrid populations, which is essential for selectionthe of large superior individuals, sizeselection is (i.e. (MAS), restricted a which by tree canseedlings crop). at help an Molecular early stage, cull could marker-assisted be a poor potentiallythese useful performing challenges solution to in progeny citrus breeding. J Sci Food Agric Keywords: Supporting information may be found in the online version of this article. ( RESULTS: Large differences were observedorange contained with the respect greatest number to ofthe volatile volatile compounds smallest compositions and number the among largest of the volatile‘Moro’ volatiles production citrus level. blood and ‘’ genotypes. mandarin ‘Owari’ orange ‘Goutou’ had satsuma were sour yieldedeven clearly though the distinguished they lowest were from assigned volatile other into production one citrus single level. genotypes group ‘Goutou’ with based sour twoCONCLUSIONS: other on orange The sour the and clustering oranges by analysis analysis the based of molecularand on marker volatile natural the profiles. compositions, sub-groups, aroma and volatile was compositions alsoanalysis was supported of citrus by able juice the to aroma molecular volatiles differentiate can marker mandarin becitrus study. used varieties breeding The as programs. gas a The chromatography–mass tool aroma to volatile spectrometry distinguishvolatile profiles citrus compounds genotypes of and and the among assist different citrus in citrus the genotypes genotypes assessment willselection of and provide in future inter-relationships fundamental citrus detected information breeding. on among the development© of 2017 marker-assisted Society of Chemical Industry USA, INTRODUCTION Sweet oranges ( in Florida are importantThe production commodities and in consumption theincreasing of continuously mandarin US in fruits citrus the have market. attributes USA been relating as to a a resulting, of pleasant and high their aroma vitamin desirable and C, flavonoids flavor, andthe ease human carotenoids. health Awareness associated of of benefits derived peel- from phytonutrients is becoming a drivertion of of consumer fruits. purchases and the consump- BACKGROUND: Citrus fruit flavor is an important attributevolatiles prioritized compositions in from variety improvement. 13 The selected present study citrusaurantium compared genotypes, juice including six mandarins ( Abstract other citrus genotypes Yuan Yu, and Frederick G Gmitter selected mandarins, mandarin relatives and Comparative analysis of juice volatiles in (wileyonlinelibrary.com) DOI 10.1002/jsfa.8563 Research Article Received: 16 February 2017 Revised: 14 July 2017 Accepted article published: 21 July 2017 Published online in Wiley Online Library: 6 September 2017

1124 1125 ), C ∘ m, –1 𝜇 mass 27 from the to 260 (green to C, respec- ∘ –1 a –1 and ionized z / ) and titratable . Inlet, ionizing ratio (indicating 1 m –1 − Cmin b Cmin ∘ ∘ C. The GC equipment over C for 30 min, and then a ∘ ∘ a C until analyzed. For analy- ∘ wileyonlinelibrary.com/jsfa (black to white), L 26 m; J&W Scientific, Folsom, CA, USA), 𝜇 C. After exposure, the SPME fiber was ∘ , presented as a percentage of ). –1 C, and then ramped at 100 L of 1000 ppm 3-hexanone as an internal stan- ∘ 𝜇 (blue to yellow), as well as the and used to genotype the citrus samples with the Gold- C to 230 )and6 ∘ b C, and then thawed under tap water and loaded into an 28 –1 ∘ -alkanes was run at the beginning of each day separately to n DVB/CAR/PDMS; Supelco, Bellefonte, PA, USA) was exposedheadspace to the for 1 hinjected at into 40 the gas chromatographnologies, (Model Santa 6890; Clara, Agilent CA, Tech- release USA) the injector compounds port for 15 (splitless min mode) at to 250 and settings were: DB-50.25 mm, film columns thickness (length 1.00 60 m, inner diameter coupled with a 5973 N MS detector (Agilent Technologies).umn The col- oven was programmed to increase at 4 orange color). Then, thedomly 15 assigned fruits into three from pools each with fivepool genotype fruits served per were pool, ran- as and each apeeled biological replicate. and Citrus gently fruitscitrus juiced were juicer manually using (Hamilton Beach, a Glen Allen,oil Hamilton VA, USA) from Beach to prevent entering 932 peel thethe manual data juice. on Juice the samples number of were seeds, used juice to percentage (mL collect 100 g dard were mixed andwith a placed magnetic crimp in cap with a aseptum silicone/polytetrafluoroethylene 20-mL (Gerstel glass Inc., vial Linthicum,at MD, and USA). –20 capped The vialsauto-sampler were stored (Model MPS2; Gersteling tray Inc.) holder equipped (Laird Tech, with Gothenburg,Peltier Thermostat a Sweden) (CTC controlled Analytics cool- AG, by Zwingen, a Switzerland). Sam- ples were usually held for 2–20 h at 4 sis, the juice samples were incubated at 40 2-cm tri-phase solid phase microextraction (SPME) fiber (50/30 as the Hunter color space value: red) and Volatile extraction and gas chromatography (GC)-mass spectrometry (MS) analysis For this, 3 mL of juice,(359 3 g mL L of saturated sodium chloride solution Volatile compound identification Volatile compounds were identified by comparingtra the with mass three library spec- entries: NIST05.L (NIST/EPA/NIH Mass SpectralLibrary; National Institute of Standardsburg, and MA, Technology, USA), Gaithers- as well as Wiley HPCH 2205.L and ADAMS.L juice color, soluble solids content (SSC) (g 100 mL calculate retention indices (RI). at 70 eV. Data were collected withsystem the ChemStation (Hewlett-Packard, G1701AA Palo data Alto, CA,C-18 USA). A mixture of C-5 to source and transfer line were kepttively. at Mass 250, units 230 were monitored and from 280 30 to 250 acidity (TA) (g 100 mL enGate genotyping array platformInterdisciplinary in Center for the Biotechnology Universitydata Research. of were Genotyping collected Florida’s and analyzed using(Illumina Inc., Genome San Studio Diego, software CA, USA). spectral databases. In addition, the RIspared with for published compounds RIs were on the com- same column.aroma The amount volatile of each was expressedpeak as area relative over internal content standard (aroma peak area). volatile Genomic DNA extraction and SNP genotyping Genomic DNA was extractedmethod from fresh leaves using the CTAB and held for 11.70 minused for a as total carrier run gas time at of 60 a min. flow Helium rate was of 1.5 mL min initial 40 C. cv. Cit- C. sinensis family mem- Citrus cv. ‘Goutou’, ‘Zhu- each of 15 fruits per 25 , et al. cv. ‘’), one Clementine 24 Citrus aurantium population and construct mandarin 1 We used different analysis tools aiming 25 identified 49 aroma active compounds in ) compared to common mandarins such as : 1124–1131 © 2017 Society of Chemical Industry where mandarin hybrids produced volatiles Citrus limonia that detected more volatiles, including more 21 98 20 22 cv. ‘Nules’) and one satsuma ( detected 225 volatile compounds in juice sam- Liberibacter asiaticus, the presumed causal agent and the experiment on two mandarin-pummelo 19 2018; 23 cv.‘Fortune’, ‘King’. ‘’, ‘Pimpled’, ‘Ponkan’ and C. sinensis ), et al. × Candidatus Overall, these studies and others identified aroma volatiles in Citrus clementina C. reticulata ( Evaluation of fruit quality attributes In total, 15 citrus fruitscommercial free maturity of any (identical peel to2014 defect season industry were and harvested harvest taken at to time)typic the laboratory evaluation. in within As the 1–2 described h by for Yu pheno- ‘Owari’). The citrus selections were plantedunder in the the UF-CREC same groves field soil,conditions. irrigation, At the illumination time and of fertilization thewith study, those trees were not infected of Huanglongbing (aka HLB or ). luan’ and an unknown named ‘Sour’), one bloodcv. orange ‘Moro’), ( one lime ( genetic linkage maps. sesquiterpenes and esters, in ‘Temple’ andreticulata ‘Ortanique’ ( ‘Ora’ and ‘Ponkan’. Thefour volatile varieties, profile including was mandarins, able orange to differentiate and pummelo ( rus maxima mandarin juice, and comparative analyses ofperformed aroma with volatiles respect were to mandarins and other bers, such as sweetwe orange investigated and aroma pummelo. In volatilesatives the in and present 13 study, mandarins, otherlime, mandarin citrus which rel- have not genotypes, previously been includingaroma reported in sour volatile exhaustive juice studies. orange Meanwhile,rus and we selections in genotyped the all University of the Florida1536-single breeding nucleotide 13 polymorphism program (SNP) with cit- assay, which a was used to genotype a mandarin F to understand relationships among differentdarins, genotypes mandarin of relatives man- anddance with other their aroma citrus volatile profile and genotypes, molecular marker intypic geno- data. accor- The research wouldon provide development fundamental of information MAS for citrus fruit quality improvement. J Sci Food Agric Plant material Thirteen citrus genotypes were studied,( including six mandarins MATERIALS AND METHODS ‘Sunki’), three sour oranges ( Comparative analysis of juice volatilesof them were consideredKerbiriou to be mandarin core aroma volatiles. www.soci.org hybrids revealed transgressive behaviourduction of compared to aroma their parents. volatile pro- five mandarin hybrids. Those resultscomparative were study confirmed by a recent genotype was used to measure surfaceand color, fruit fruit weight diameter (g) (mm) individually. Fruitusing surface a color CR-331 was colorimeter measured (Minolta, Tokyo, Japan) and described ples of 56 mandarinrus Research hybrids and Education from Center (UF-CREC) the citrusgram, breeding University and pro- of found that Floridain hybrids Cit- esters. of Another an study orange onbythesameteam, parentage 20 were mandarin rich hybrids was performed in fewer number andhydes, lower whereas amounts, hybrids although with withgenerated more more sweet sesquiterpenes and alde- orange esters. In addition, genetic a follow-up background olfactometry study et al. -pinene, 𝛼 : 1124–1131

98 Number of volatiles of Number 2018; 140 120 100 80 60 40 20 0

J Sci Food Agric r u

SD of three biological replicates.

ngp

a ±

R

i

r

a

Ow

s

ule

N

ne

u

rt

Fo

Moro

n

ulua

Volatile profile A total ofsamples 167 of 13 aroma citrusalcohols genotypes volatiles (including by seven were GC–MS, monoterpene and detected whichalcohols), three comprised 12 from sesquiterpene 16 aldehydes the (including(including juice one five monoterpene), monoterpenes), 16 esters twomatic ethers, hydrocarbons, one 10 furan,and ketones one five (including sesquiterpene), aro- one five phenol,(among 62 monoterpenes putative which sesquiterpenes 25other were volatiles. identified), Theples 16 relative from contents monoterpenes the 13 of and citruscodes volatiles genotypes 26 and and in descriptive their juice statistics correspondinginformation peak are sam- (Table provided S1). Large innumbers differences the (quantity) were Supporting and observed content incitrus (quality) total genotypes. of The number volatiles of aroma among volatilesgenotypes the detected ranged in from citrus 28 in ‘Ponkan’ mandarin tosour 119 in ‘Goutou’ orange (Fig. 1;S1). see The also terpene-derived compounds thesatsuma, varied Supporting ‘Nules’ Clementine from and information, ‘Ponkan’ one mandarin Table to in 14blood in ‘Owari’ ‘Moro’ orange. The relative contentin of ‘Owari’ volatiles satsuma ranged to fromcategorization 499.59 1.50 of in the ‘Goutou’ aroma sour volatilesclasses, according orange to ‘Goutou’ (Fig. their 1). sour chemical By orangealcohols, contained aromatic the hydrocarbons, sesquiterpenes largestpenes, ‘Moro’ blood amount and orange produced of monoter- the largest amountand of esters ketones, whereasamount ‘Murcott’ of aldehydes mandarin amongproportions yielded the of citrus chemical the classes genotypes within largest (Fig.calculated each S2). by citrus The selection dividing were thesame number chemical of class by volatiles the belonging total‘Pimpled’ number to mandarin, of ‘Goutou’ the volatiles sour from orange, Table ‘Sour’ 1. ‘Moro’ sour blood orange orange and had relatively highpenes proportions (26.19–47.9%), of sesquiter- whereasproportions others of contained monoterpenes relatively (21.43–30.77%).also high had ‘Sunki’ relatively high proportions mandarin of aldehydes (aroundA 21.05%). total of 11types, consensus including ethanol, volatiles acetone, 2-methyl-2-propanol, was present in all citrus geno- h

Z ), Sour –1

a/b, www.soci.org Y Yu Goutou

respectively, Sunki ), which was © 2017 Society of Chemical Industry

–1

a/b kan

n

Po Pimpled -transformed relative con-

10

Murcott

ing K 0

700 600 500 100 Relative content Relative ) for ‘Owari’ satsuma. The three sour oranges also ) for ‘Rangpur’ lime. However, for the acid content, Volatile profile in juice samples from 13 citrus genotypes. The data of relative content are the mean –1 –1 respectively. whereas ‘Owari’ satsumathe and largest ‘Zhuluan’ (0.24) sour and orange smallest exhibited (–0.62) value of juice color had high TA,darin as and well ‘Goutou’ sour assmallest orange the (–0.01) exhibited mandarin the value ‘Sunki’. largest of (0.8) ‘Ponkan’ and fruit man- surface color ‘Rangpur’ lime contained the largest TA value (5.44 g 100 mL Figure 1. wileyonlinelibrary.com/jsfa RESULTS Fruit and juice characteristics The fruit qualitygenotypes, attributes and were(Fig. are S1). analyzed shown The forluan’) three in the were sour 13 the the(239.25–296.17 oranges g), citrus largest Supporting (‘Sour’, whereas inin ‘Sunki’ ‘Goutou’ information size mandarin size and was (82.3–92.06 ‘Zhu- (41.05contained mm) mm) the and the smallest and weight weight highest (30.82 g). SSC ‘Murcott’ (14.27 g mandarin 100 mL Statistical analysis JMP Genomics, version 7NC, and USA) were SAS, used to version analyzeficients 9.4 the data. (SAS of Pearson Corp., correlation fruit coef- Cary, |0.5| traits was and used aroma as(PCA) volatiles a based were cut-off on calculated, point.tiate Pearson’s and correlations citrus Principal genotypes was component using analysis applied the to log differen- which was 10100 mL times the amount of smallest TA value (0.53 g tent of volatiles.absence/presence Then (0/1) the format, volatile and(HCA) data hierarchical was were performed cluster transformed using analysis method the into with unweighted arithmetic mean pair-group to average observecitrus similarities between genotypes. the The HCA wasfruit also characteristics. The performed genetic using relatedness the oftypes the data 13 was of citrus calculated geno- from the genotypicon data the of Identity-by-state 1536 (IBS) SNPs values. based approximately twice100 as mL much as the smallest value (7.47 g

1126 1127 00 00 00 00 00 00 00 00 00 00 and ...... b / a Rangpur (%) , a -calacorene 33 1 2 33 1 2 67 11 22 33 5 10 67 12 24 67 2 4 00 7 14 33 7 14 67 2 4 𝛽 ...... Owari (%) 23 1 3 23 1 3 35516 03826 13413 45 2 6 90310 45 2 6 23413 ...... Nules (%) )-carvyl acetate (118), ethyl wileyonlinelibrary.com/jsfa E 63 1 3 63 1 3 42619 42516 95929 26 2 6 16412 26 2 6 26 1 3 -ylangene (130), RI1460 (160) )-cadina-1,4-diene (184; code ...... 𝛼 E Fortune (%) 6712 33 7 18 71 7 18 86 11 28 3325 33 5 13 5725 1925 ...... Moro (%) ratio displayed negative correlation coefficients 5600 00 00 00 24 56 0 0 1 2 13 14 16 95 7 8 26 9 10 77 15 17 13 7 8 38 7 8 69 3 3 56 22 26 ...... b ratio displayed a positive correlation coefficient / a b / Zhuluan (%) a . Soluble solids content and titratable acidity positively b -cubebene (152), 6,9-guaiadiene (154), RI1455 (155), )-dihydrocarvone (103), dihydrocarvone (101), citronellyl 89 2 5 53 7 17 26 4 10 11 12 30 63 2 5 53 6 15 63 3 7 42 1 2 ...... 𝛽 E -cadinene (185) (Fig. 2C; see also the Supporting infor- 𝛼 Sour (%) When ‘Moro’ blood orange, ‘Goutou’ and ‘Sour’ sour oranges, The pairwise correlation coefficient for each primary chemical -bourbonene (134), RI1548 (191), (187), 𝛽 and mation, Table S4).(172), ( The volatileacetate compounds, (119), ethyl including octanoate (96),2-butenoate RI1489 ( (33), ethylimportant hexanoate for the separation (54) of citrus and genotypes across RI1155 PC2. (83),Murcott’ were and ‘King’analysis, mandarin the were first five excluded principaland in components 8.5% accounted a of for the second 13.4% total PCA variance (Fig. 2B). ‘Rangpur’ lime, ‘Zhuluan’ correlated with aldehydes and alcohols, respectively. Fruitand juice surface color group against every fruit traittive was and also negative calculated, correlation coefficients and were both observedporting posi- (see information, Sup- Table S3). The total volatile contentcorrelated positively with fruit weightalthough (0.53) it and negatively number correlated of with seeds surface (0.54), color with all groups, although theJuice aldehyde color group was an exception. with aldehydes, even though it was small (0.2). PCA PCA analysis was employed as anmine exploratory how technique to samples deter- groupedand based on second volatile principal content.52.3% components The and first 14.8% (PC1 of and the13 citrus total PC2) genotypes variance, were represented respectively distributed (Fig. unevenlyformed 2A). and The some clear of them clusters.orange ‘Moro’ were blood clearly orange separatedcitrus from genotypes, and in each ‘Goutou’ agreement other withmuch sour and the more from fact volatiles that other than theysour produced the orange others. ‘Murcott’ and mandarin,each ‘King’ ‘Sour’ other mandarin and could fromthe be other citrus distinguished genotypes citrus from were genotypes.hardly clustered The closely be remainder together of differentiated andthe from could citrus genotypes each primarilyseveral other. based sesquiterpenes, including on The ( the PC1as relative shown separated content in of the Supporting information, Table S1), juice color 0467 29 8 10 3645 61 13 17 5222 61 8 10 6822 90 33 43 ...... Goutou (%) 6300 00 12 6300 00 12 6365 42 17 14 0543 95 15 12 2632 16 15 12 2621 ...... Sunki (%) 57 1 2 57 1 2 86 7 18 86 8 21 43 11 28 14 2 5 71 5 13 71 2 5 14 0 0 57 47 ...... The pairwise corre- Ponkan (%) 20 0.05). Of the 6858 pairs, 26 3 10 13 0 0 1 2 13 1 3 77 5 17 64 5 17 53 6 21 26 2 7 51 3 10 79 2 7 ...... < -cymene and acetaldehyde P p Pimpled (%) 2009200 0 00 0085 2 00 4 00 00 00 00 00 00 92 1 2 85 14 29 92 1 2 77 0 0 1 3 54 6 12 31 5 10 08 12 25 31 2 4 54 4 8 ...... Murcott (%) 6323 4223 6311 5335 16 6 11 63 9 17 58 12 23 26 9 17 16 6 11 : 1124–1131 © 2017 Society of Chemical Industry ...... 98 King (%) 2018; -terpinene, p-cymene, , terpinolene, linalool Number of juice volatiles in 11 chemical classes detected in the 13 citrus genotypes 𝛼 –0.58) and between RI0947 and several other volatiles. = Aromatic hydrocarbon 1 2 Furan 0 0 1 1 Ether 0 0 1 1 Sesquiterpene 7 18 Phenol 1 2 The 167 volatile compounds were categorized into 11 groups Ester 4 10 Total 38 52 47 28 38 119 76 39 84 38 31 30 50 Other 5 13 Aldehyde 1 2 Monoterpene 12 31 Ketone 2 5 Class Alcohol 5 13 Table 1. r J Sci Food Agric Correlation analysis The pairwise correlation analysisother for volatile each was volatilepounds. performed against Pearson correlation every was to employed for investigate juice samples of13 co-regulated the citrus com- genotypes. Of thein significant 13 correlation 861 coefficients pairs ( analysed, 6858 resulted and an unknown compound, RI1641 (see SupportingTable information, S1). Several genotype specificexample, volatiles eight, were five detected.in For and ‘Moro’ four blood orange, volatilesrespectively. ‘Murcott’ were Forty-three mandarin specifically and volatiles ‘Rangpur’ found ‘Goutou’ were sour lime, orange, specifically and more detected thanpenes. half of in The them volatile were sesquiter- butanalFive was carotenoid-derived only aroma volatiles found were in detected,6-methyl-5-hepten-2-one, ‘Sunki’ including neryl mandarin. acetate, citronellylanyl acetate acetate, and geranyl ger- acetone. Citronellyl‘Moro’ acetate was blood found orange, in 6-methyl-5-hepten-2-one andtone were geranyl observed ace- in ‘Murcott’ mandarin,and whereas neryl geranyl acetate acetate(geranyl acetate were only), ‘Sour’ sour found orange and ‘Moro’ in blood orange. ‘King’ mandarin, ‘Goutou’ myrcene, 6623 and 235 exhibited positivecients, respectively. and The negative strong correlation positive coeffi- correlationsily were found primar- betweenpenes, volatiles with correlation from coefficients close a to 1.correlations group The were strong observed of negative between several( sesquiter- based on theirester, chemical structure ether, includingsesquiterpene, alcohol, furan, aldehyde, monoterpene and aromatic other. hydrocarbon, ketone, phenol, Comparative analysis of juice volatiles www.soci.org lation coefficient forgroup each was calculated chemical for juice group samples(see against of Supporting the every 13 information, citrus Table other genotypes S2).ficients Positive were correlation observed coef- for(with most more of than three the compounds), primary although thewas chemical aldehyde group groups an exception. Thecorrelation coefficients aldehyde with group esters only andwith ketones. coefficients displayed greater The positive than correlations 0.8 weregroups observed (e.g. sesquiterpenes between and several monoterpenes). .(D) et al. : 1124–1131 98 2018; J Sci Food Agric genotypes (Fig. 3). The HCAclusters: A assigned and citrus B. genotypes Clusterwhereas into A cluster only two B contained contained ‘Goutou’ the sourdivided other into orange, citrus two sub-clusters: genotypes, B1 and and B2. was orange B1 included and ‘Moro’ ‘Sour’ blood sour orangedarin and, and within ‘Nules’ the Clementine B2, were closely ‘Fortune’‘Owari’ clustered man- together, satsuma and and ‘Ponkan’tered together. mandarin When were the HCA alsotics, used the closely 13 the citrus data clus- genotypes of were assignedSupportig fruit into characteris- information, three clusters Fig. (see S3B).oranges Cluster including A ‘Goutou’, contained ‘Zhuluan’ and threeB ‘Sour’, sour included whereas ‘Moro’ cluster blood orange,darin. ‘Rangpur’ Cluster lime C and comprised the ‘Sunki’ remainder man- with of most the of citrus them genotypes, being mandarins. Genotypic analysis The 13 citrus selections were genotyped usingand the several consensus 1536-SNP SNPs array (i.e. that exhibited identicalfigurations) allelic con- were found among citrusthe selections 1536 (Table 2). SNPs, Out 995oranges of ‘Goutou’, were ‘Sour’ and observed ‘Zhuluan’. More to thanSNPs 1000 have were consensus also consensus observed for betweendarins sour any two ‘King’, selections ‘Murcott’, from ‘Pimpled’ man- blood orange and and ‘Ponkan’, sour oranges between ‘Goutou’‘Rangpur’ and ‘Moro’ ‘Sour’, lime and and between ‘Sour’were sour grouped into orange. three The clustersvalues 13 A, calculated B citrus and from genotypes C the accordingCluster genotypic to A data the of included IBS 1536 sour SNPs oranges (Fig. 4). ‘Goutou’, ‘Sour’, ‘Zhuluan’ and www.soci.org Y Yu , fruit weight © 2017 Society of Chemical Industry a/b and a, b -phellandrene (64), myrcene (50), terpinolene 𝛽 -phellandrene (59) (Fig. 2D; see also Supporting infor- 𝛼 PCA of volatile profile in juice samples from citrus genotypes. (A) PCA scores of 13 citrus genotypes. (B) PCA scores of eight citrus genotypes. (C) In addition, a PCA analysis was applied to study the 13 citrus -cymene (62), -elemene (149), were important for the differentiation of citrus Cluster analysis HCA using relativewas volatile content performed of to the investigate 13 the citrus relationship genotypes between citrus mation, Table S5).(42), The 2,4-dimethyl-1-heptene (32), volatile RI1270𝛾 compounds, (108), RI0754 including (25) RI0947 genotypes and across PC2. genotypes based on fruit qualitycipal traits. components The (PC1 first and and PC2) second representedthe prin- 37.6% total and variance (see 17.5% Supporting of information, Fig. S3A).sour The three oranges were clearlytypes, distinguished as from was other ‘Sunki’ citruspur’ geno- mandarin. lime ‘Moro’ were clustered blood almost orange together.‘Fortune’ ‘Murcott’ and mandarin mandarin could ‘Rang- and be distinguished fromThe other mandarins. fruit traits,and including size, soluble juice solids content/titratable color aciditysolids ratio and content, soluble were important fortypes across the PC1. separation The PC2 of separated the citrus citrus genotypesbasedonfruitsurfacecolorandtitratableacidity. geno- primarily (73) and wileyonlinelibrary.com/jsfa sour orange andeach ‘Pimpled’ mandarin other could andof be from separated the from other citrusother. citrus genotypes The were genotypes. PC1based hardly The on the differentiated differentiated remainder relative content from the ofp several each monoterpenes, citrus including genotypes primarily Figure 2. Loading plot of volatile profile for (A), with each number corresponding to a volatile compound, as indicated in the Supporting information (Table S1) Loading plot of volatile profile for (B).

1128 1129 as 32 These and this 15 ‘Orlando’ -terpinene 31 × 𝛾 -terpinene. 𝛾 23 Our present findings sug- -terpinene in ‘Ponkan’ man- wileyonlinelibrary.com/jsfa 𝛾 22 ‘Fortune’ and ‘Nules’ Clementine were -terpinene. 31 𝛾 reported that PCA based on the aroma The absence of -terpinene was not detected in juice sam- 𝛾 34 23,35 -terpinene (0.05–0.41%), whereas mandarins ‘Pim- 𝛾 supported this conclusion and further found that both and‘Murcott’wasderivedfromacrossbetweenman- ‘King’ was suggested to be a natural 22 33 31 et al. In previous studies, the Mediterranean ‘Willowleaf’ mandarins Previous studies -terpinene, whereas ‘Commune’ had a relatively ‘Ponkan’ contained fewer esters thancott’ mandarins ‘King’ (Table 1), and ‘Mur- fewerfewer sesquiterpenes overall volatiles than than ‘King’ mandarinscott’. mandarin, ‘Fortune’, ‘Ponkan’ ‘King’ and and is ‘Mur- avariety. traditional, old and widely grown mandarin (4.27%). However, ples of ‘Ponkan’ mandarin, evenwith though ‘Pimpled’ it mandarin was based grouped on together plotted their SNP close genotypic to data, ‘Willowleaf’ and genome sequence. mandarin by PCA using the whole volatile profiles was able toorange effectively juice distinguish samples, and mandarin the and experimentberg performed by Golden- produced a relatively low𝛾 level of limonene andhigh a level high of level limonene of and a low level of was confirmed by subsequent genetic diversitydom studies using amplification ran- of polymorphic DNA molecular markers, observations were supported bythat most a tangors, subsequent studyrelatively and reporting low ‘King’ mandarin levels also of contained closely clustered togetheraroma in volatile compositions the between them. study, explaining the similar darin and sweet orange. darin, which might beprobably be a explained by result our subsequent of conductedpolymerase quantitative variety chain reaction level experiments on variability, the could genes expression that level encode of the key enzymespathway. in the terpenoid biosynthetic gest the same conclusion (seeMandarins Supporting ‘King’, information, Table ‘Murcott’, S6). ‘Fortune’ anda ‘Nules’ Clementine relatively had highamount level of of limonenepled’ (81.43–92.29%) and and ‘Ponkan’(79.4-81.07%) a and had ‘Pimpled’ low a contained high relatively levels of low amount of limonene well as by SNP molecular markers‘King’ in and the ‘Murcott’ present were closely study. clustered Mandarins together basedjuice aroma on volatile their profiles and SNP marker genotypicand data (Figs 4) 3 with 1186 outthus of revealing the 1536 close SNPs genetic being backgroundtune’ between in was them. consensus ‘For- supposed (Table 2), to be derived from Clementine 24 ‘Moro’ blood 5 The present study 20,22,30 : 1124–1131 © 2017 Society of Chemical Industry 98 2018; King Murcott Pimpled Ponkan Sunki Goutou Sour Zhuluan Moro Fortune Nules Owari 29 Number of consensus SNPs among the 13 citrus genotypes HCA of volatile profile in juice samples from 13 citrus genotypes. It was previously observed that mandarin hybrids with a sweet MurcottPimpled 1186 Ponkan 1030Sunki 1002Goutou 1078 Sour 1093 896 823ZhuluanMoro 906 828 957 1150 Fortune 692Nules 698 784 758Owari 708 987Rangpur 633 793 617 738 907 882 721 614 1034 780 600 846 622 798 801 702 597 549 920 564 804 781 918 584 583 983 274 899 1180 1215 707 797 809 1117 1009 870 712 874 1039 712 727 834 978 991 718 718 1014 853 747 936 688 896 786 686 658 942 806 836 753 612 868 Table 2. orange genetic background produced more esterspenes, and as sesquiter- well as overall aroma volatiles. J Sci Food Agric because ethyl esters wereethanol synthesized and from acyl-CoAs the generated esterification frommetabolism. fatty of acid and amino acid Esters and aldehydes are considered toof be the major aroma constituents active compounds in orange juice. DISCUSSION blood orange ‘Moro’. Cluster‘King’, B ‘Murcott’, comprised ‘Pimpled’, mandarins, ‘Ponkan’sisted including of and diverse members, ‘Sunki’. including ‘Nules’ Clementine, Clustermandarin, ‘Fortune’ ‘Owari’ C satsuma and con- ‘Rangpur’ lime. Figure 3. Comparative analysis of juice volatiles www.soci.org is in agreement with these conclusions. The typical contained 14 esters (includingwhich five was monoterpene the esters), largest(Table 1; number see among also the SupportingOut 13 information, of citrus Fig. the genotypes S2ethyl and 14 Table esters, butanoate, S1). five ethyl2-butenoate, ethyl were hexanoate, esters, much more including ethyl abundantcompared ethyl in to octanoate the ‘Moro’ acetate, other blood citrus and orange genotypes, andwas ethyl ‘Moro’ blood also orange observedagreement to be with relatively the rich previous in finding ethanol. This in was ‘Fortune’ in mandarin et al. : 1124–1131 98 J Agr Food Chem :611–626 (2012). 2018; :14–23 (2011). 8 91 :24–41 (2013). 25 :681–695 (2008). J Sci Food Agric 48 J Sci Food Agr Tree Genet Genomes . However, juice aroma volatile profiles a :1290–1292 (1972). 20 J Int Food Agribus Mark Crit Rev Food Sci Nutr , Citrus genomics. J Agr Food Chem :8115–8124 (2009). et al. fresh and stored mandarins. review. oil. substitution. citrus flavor research during57 the Past 100 Years. Y A great genetic diversity in fruit characteristics and juice aroma 4 TietelZ,PlottoA,FallikE,LewinsohnEandPoratR,Tasteandaromaof 5 Perez-Cacho PR and Rouseff R,6 Fresh squeezed Coleman orange RL juice and Shaw odor: PE, Analysis a of essence oil and aroma 1 Baldwin KL and Jones KG, US citrus import demand: seasonality and 2 Rouseff RL, Perez-Cacho PR and Jabalpurwala F, Historical review3 of Gmitter FG, Chen C, Machado MA, de Souza AA, Ollitrault P, Froehlicher ‘Zhuluan’ generated a lower(Fig. level 1), of even though volatile 1180 compounds outthe of (23.59) two 1536 sour SNPs oranges were in (Table 2). consensus for volatile constituents was observed, especiallyvolatiles. On concerning a genetic aroma level, the 13 citrusinto genotypes three groups, were which assigned fit the traditional classification. Theically genet- related citrus genotypes, such as the three sourerated oranges, similar gen- fruit characteristics, such asthey acidic were juice. In clustered addition, relatively closepared to to ‘Moro’ the blood others orange com- because,small values for of example, juice color they carriedwere not relatively able to distinguish ‘Zhuluan’ sour orange fromin mandarins the PCA study‘Goutou’ sour that orange, was clustered included relatively close to all ‘Moro’orange blood the and ‘Sour’ 13 sour genotypes, orange. and even ACKNOWLEDGEMENTS We thank Yanzi Zhang, NancyQibin Owens, Yu Marjorie and Wendell, Misty Xuwork Holt Wei, was principally for funded providing by the technicalManagement New Corporation, Varieties assistance. Development the This & Citrus ResearchFoundation, and and Development the University of Florida Plant MolecularInitiative. Breeding SUPPORTING INFORMATION Supporting information may be found inarticle. the online version of this REFERENCES CONCLUSIONS In conclusion, the analysis byof GC–MS of 13 the citrus juice aromacompounds. genotypes, volatiles All allowed the theand 13 quantitative detection citrus variations of in genotypesPCA 167 aroma and exhibited volatile volatile HCA qualitative compositions. baseddistinguish The mandarin on varieties the and aromasupported natural by volatile groups, the levels analysis and using were this SNPanalysis of able was genotypic juice data. aroma to volatiles The from GC–MS citrus fruitsstrated has been to demon- beused a for tool the forbreeding selection differentiating programs. genotypes, of The present and appropriatein data can parents aroma reveal be for volatile the composition wide futurenatural and variations citrus groups, content indicating among/within their citrus important genetic information for relationships, understanding providing gesting citrus aroma, that and manipulation sug- ofbe aroma possible in volatile future citrus constituents breeding would programmes using MAS. g × 45 𝜇 www.soci.org Y Yu consid- and they 38 39–41 © 2017 Society of Chemical Industry and a SNP marker study Gmitter and Hu ) in Tunisia, although the The essential oils obtained –1 37 36 . 42–45 gmL 𝜇 Our data revealed that the true sour 46 Citrus medica and Dendrogram of 13 citrus genotypes based on IBS calculated from C. reticulata ) than sour orange (99.52 –1 In previous studies, mitochondrial marker analysis indicated ‘Goutou’ and ‘Zhuluan’ were probably sour orange-pummelo suggested that the ‘Rangpur’cross lime of probably arose from a direct wileyonlinelibrary.com/jsfa Blood orange yielded a higher level of juice volatiles (157.55 mL PCA and HCA, based on aromaguish volatile mandarin levels, varieties were able and tostudy, natural distin- ‘Moro’ blood sub-groups. orange, In ‘Sour’ and the ‘Goutou’distinguished sour present oranges from were otherall genotypes 13 by the citrus(Fig. PCA genotypes 2A), that whereas based included ‘Zhuluan’ onthose sour aroma four orange were volatile was closely compositions data. not, clustered Mandarins based even ‘Murcott’ and on though ‘King’ their wereother also SNP distinguished citrus from genotypic genotypes,‘Orlando’ although tangelo, ‘Fortune’, and as thetered typical a with other mandarin Clementine mandarins ‘Ponkan’ (Figs 2A were and clus- 3). that ‘Rangpur’ lime shared the(‘Sunki’, cytoplasm of ‘Cleopatra’ rootstock and mandarins ‘Shekwasha’), Figure 4. data of 1536 SNP markers. varieties were unknown. orange ‘Sour’ contained a similar level of(91.68) juice volatile compared compounds to ‘Moro’two blood sour orange orange hybrids (89.75). exhibited‘Moro’: However, different ‘Goutou’ the results produced compared a to much higher level (499.59), whereas from leaves of ‘Zhuluan’those sour from orange ‘Goutou’ were sour more orangeseasons, abundant from and than both linalool dry was and wet the harvest most representative compound. ered ‘Sunki’ mandarin and ‘Rangpur’ limeChina. to be In natives of the Yunnan, presenttered study, with the ‘Owari’ ‘Rangpur’ satsuma, lime andbased was mandarins on SNP ‘Pimpled’ closely genotypic and data, clus- juice ‘Sunki’, aroma volatile profilescharacteristics and (Figs fruit 3 and 4; seeS3). also PCA analysis Supporting using information, aroma Fig. volatile levelsguish was ‘Rangpur’ not able lime to from distin- other mandarinsgenotypes were when included. all the 13 citrus hybrids used as rootstocks introduced from China, were closely clustered with thethe true SNP sour genotypic orange data, ‘Sour’ asalso based well Supporting on as information, Fig. fruit S3). The characteristics composition (Fig.oils of 4; extracted essential from see sour orange leaves, flowers andstudied peel have extensively been becausefood, they aromatherapy and are medicine. widely used in cosmetic,

1130 1131 , , Nat ,ed. L. on et al. et al. Phyto- :49–61 Proc Fla 7 :267–277 :128–141 ,Sequenc- 44 11 Acta Horticul- JEconEntomol et al. Biol Pharm Bull spp.) using RAPD J Essent Oil Bear Pl Econ Bot :497–501 (1997). Citrus aurantium L.). Citrus :174–181 (2004). 30 :84–89 (2006). :3968–3972 (1997). 117 41 : e125628 (2015). Tree Genet Genomes aurantium L. 45 10 wileyonlinelibrary.com/jsfa Plant Mol Biol Report Citrus Citrus aurantium Hortscience PloS ONE :11779–11785 (2011). . Allured Publishing Corporation, Carol (Psocoptera: Liposcelididae). 59 :133–139 (1998). J Agr Food Chem Lebensm Wiss Technol 102 :56–60 (1992). , Nuclear -diagnostic SNP markers mined Proc Fla State Hort Soc :1904–1912 (2012). 105 26 :656–662 (2014). et al. :1283–1289 (2003). :449–456 (2014). 32 62 Euphytica Identification of Components by Gas Chromatog- J Agr Food Chem , New universal mitochondrial PCR markers reveal new infor- 1065 :1629–1633 (2002). :1296–1301 (2001). :426–435 (2015). et al. from 454 amplicon sequencing reveal admixtureof genomic modern structure citrus varieties. of contemporary citrus species (). (1990). KD, Evaluation of new citrus rootstocks forsouthern ‘Tahiti’lime Florida. production in Rootstocks similar to sourState Hort orange Soc for Florida citrus trees. F mation on maternal citrus phylogeny. (2011). Garcia-Lor A Biotechnol mandarin juices and itsmultivariate use analysis. for comparison with orange juices by turae ing of diversecomplex mandarin, pummelo history and of orange genomes admixture reveals during citrus domestication. orange juice drinks andmultivariate its analysis. use for comparison with pure juices by parentage of ‘Orri’ and ‘Fortune’ mandarin hybrids. by Reuther W, Webber HJ andDivision Batchelor of LD. Agricultural University Sciences, of Berkeley, California, CA (1967). of the genetic diversity amongmarkers. mandarins ( sour orange accessions and evaluation of theirmolecular genetic marker variability analyses. by extracts and essential oil25 from plant oils alone and in combinationLiposcelisbostrychophila with controlled atmosphere on raphy/Mass Spectrometry Stream, IL (2007). reproducibility using Klen Taq1gel DNA purification polymerase, of Chelex genomic 100, DNA. (1993). and ethanol fermentation metabolismdarins. in causing off-flavors in man- 94 Suppressive effect of flavonoids from Korean the expression of inflammatory mediators in L6Phytother skeletal muscle Res cells. Vieira RF, Chemical characterization ofaccessions leaf of essential sour oil orange from18 seven ( orange, sweet orange, ,chemistry bergamot and . 40 Castle WS, Pelosi RR, Youtsey CO, Gmitter Jr FG, Lee RF, Powell CA 38 Gmitter FG Jr and Hu XL, The possible role of Yunnan, China, in the39 ori- Castle WS, Grosser JW, Schnell RJ, Ayala-Silva T, Crane JH and Bowman 36 Froelicher Y, Mouhaya W, Bassene J-B, Costantino G, Kamiri M, Luro 37 Curk F, Ancillo G, Ollitrault F, Perrier X, Jacquemoud-Collet J-P, 35 Moshonas MG and Shaw PE, Quantitation of volatile constituents in 31 Hodgson RW, Horticultural varieties of citrus, in 34 WuGA,ProchnikS,JenkinsJ,SalseJ,HellstenU,MuratF 30 Shaw PE and Moshonas MG, Quantification of volatile constituents in 33 Barry G, Gmitter F, Chen C, Roose M and Federici C, Investigating the 32 ColettaHD,MachadoMA,TargonM,MoreiraMandPompeuJ,Analysis 41 Siragusa M, De Pasquale F, Abbate L and Tusa N, Identification of 42 Carvalho-Freitas MIR and Costa M, Anxiolytic and sedative effects43 of Wang JJ, Tsai JH, Ding W, Zhao ZM and Li LS, Toxic effects of six 27 Adams RP, 28 Aldrich J and Cullis CA, RAPD analysis in flax: optimization of yield and 29 Tietel Z, Lewinsohn E, Fallik E and Porat R, Elucidating the roles of 44 Kim J-A, Park H-S, Kang S-R, Park K-I, Lee D-H, Nagappan A 45 da Hora Almeida LA, Santos JZ, Soares-Filho WdS, Bizzo HR, Silva JP and 46 Moufida S and Marzouk B, Biochemical characterization of blood J J Agr Tree Citrus Proc Fla Lauriro) : e58411 :275–281 8 :995–1007 18 ,Chemotax- 90 LWT – Food Sci- et al. J Agr Food Chem Int J Food Sci Tech genotypes based :57–65 (2015). Eur Food Res Technol Citrus nobilis PLoS ONE , Study on aroma com- Blanco cv. Clementine 96 :163–170 (2005). Flavour Frag J et al. J Sci Food Agr 28 Fortunella :282–284 (1974). Blanco (Ponkan) cold-pressed Lauriro). 22 :e22016 (2011). and 6 J Sci Food Agr Citrus reticulata :2617-2628 (2012). J Food Qual :1690–1697 (2004). Poncirus 60 , :951–958 (2008). PLoS ONE 68 Blanco cv. clementine). Citrus nobilis Citrus reticulata 41 :382–390 (2009). J Agr Food Chem :1–16 (2016). :267–275 (2007). : 1124–1131 © 2017 Society of Chemical Industry Citrus :449–460 (2011). 116 12 98 :727–735 (2012). :2404–2411 (2006). 120 91 92 86 :1487–1493 (2007). 6 ) fruit characters using high-throughput SNP markers. 2018; J Agr Food Chem Citrus reticulata Food Chem :11–14 (2003). Biosci Biotech Bioch :11312–11322 (2014). :2225–2231 (2011). A, Comparative analysis ofdifferent citrus the species. volatile fraction of fruit juice from Sci Food Agr ter FGgas-chromatography-olfactometry Jr, and Aroma sensory characterization of evaluation. R, tangerine Diversity among hybrids mandarinaroma volatiles varieties compositions. by and natural sub-groups in Granell A, Fruitmatically different volatile from profiles62 those of of two their parents. citrus hybridsreticulata are dra- tangerine peel oil. Genet Genomes and DB-Wax for flavor and fragrance compounds. ence and Technology tion of the most odour-active compoundstines in ( the peel oil of clemen- 216 using gas chromatography-olfactometry. (2003). odor components of (2010). compounds in fresh and stored46 ’Mor’ mandarins. ponents in fruit from three different Satsuma mandarinSci varieties. China oil. ants found in the peel oil of on peel oilof volatile compounds – ( deciphering(2013). the genetic origin tivars (Satsuma, Bodrum, Clemantine)profile analysis by techniques. using GC/MS and flavor nents and sensory quality of freshSci and Food Agr processed mandarin juices. State Hort Soc Costa J, Volatile composition ofsolid-phase hybrids micro extraction/gas citrus chromatography/mass juices spectrom- byetry. headspace wax coatings and postharvest storage on sensoryvolatile quality composition and of aroma ’Mor’ mandarins. tion of aroma volatiles in a population of tangerine hybrids. onomic study of peel oil. constituents of wild citrus Mangshanyegan ( volatile compounds in tangerine hybrids andJ proposed Sci inheritance. Food Agr 7 Moshonas MG and Shaw PE, Quantitative and qualitative analysis of 8 Buettner A, Mestres M, Fischer A, Guasch J and Schieberle P, Evalua- 9 Chisholm MG, Jell JA and Cass DM, Characterization of the major odor- 23 Gonzalez-Mas MC, Rambla JL, Alamar MC, Gutierrez A and Granell J Sci Food Agric 21 Miyazaki T, Plotto A, Baldwin EA, Reyes-De-Corcuera JI and22 Gmit- Goldenberg L, Yaniv Y, Doron-Faigenboim A, Carmi N and Porat 24 Rambla JL, Gonzalez-Mas MC, Pons C, Bernet GP, Jose Asins M and 25 Yu Y, Chen C and Gmitter FG, QTL mapping of mandarin ( Comparative analysis of juice volatiles www.soci.org 26 Goodner KL, Practical retention index models of OV-101, DB-1, DB-5, 10 Sawamura M, Tu NTM, Onishi Y, Ogawa E and Choi HS, Characteristic 17 Tietel Z, Porat R, Weiss K and Ulrich D, Identification of aroma-active 18 Qiao Y, Xie B, Zhang Y, Zhou H, Pan S, Qiao Y 12 Liu CH, Jiang D, Cheng YJ, Deng XX, Chen F, Fang L 13 Elmaci Y and Altug T, characterization of three14 mandarin cul- Perez-Lopez AJ and Carbonell-Barrachina AA, Volatile odour compo- 15 Barboni T, Luro F, Chiaramonti N, Desjobert J-M,16 Muselli Tietel Z, Bar A E, Lewinsohn E, and Feldmesser E, Fallik E and Porat R, Effects of 19 Kerbiriou P, Plotto A, Goodner K, Baldwin EA and Gmitter FGJ, Distribu- 11 Liu CH, Cheng YJ, Zhang HY, Deng XX, Chen F and Xu J, Volatile 20 Miyazaki T, Plotto A, Goodner K and Gmitter FG Jr, Distribution of aroma