ASSESSING THE QUALITY OF HUANGLONGBING SYMPTOMATIC AND ASYMPTOMATIC ORANGE OILS FROM FLORIDA HAMLIN AND VALENCIA ORANGES

Transcription

1 ASSESSING THE QUALITY OF HUANGLONGBING SYMPTOMATIC AND ASYMPTOMATIC ORANGE OILS FROM FLORIDA HAMLIN AND VALENCIA ORANGES By BRITTANY MARTIN HUBBARD A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2017

2 2017 Brittany Martin Hubbard

3 ACKNOWLEDGMENTS I would first and foremost like to thank my Major Advisor Dr. Renée Goodrich- Schneider and my Supervisory Committee Members Drs. George Baker, Charles Sims, Ed Etxeberria, and Paul Sarnoski. I appreciate the advice and motivation you have given me as well as the use of your books, equipment, lab space, fields, and labor. I would also like to thank Roy Sweeb, Jason Taylor, and Tonya Horner, who are all currently or formerly of the University of Florida Citrus Research and Education Center in Lake Alfred, FL. I appreciate all of your help in coordinating the orange oil extraction. I would also like to thank Troy Gainey for his maturity assessments and for allowing me the use of his harvesting and transport equipment in Lake Alfred. I wish to thank Kaipeng Xu, Caitlyn Soriano, Galo Chuchuca Moran, and Sophie Zhu for helping me pick oranges. An especial thank you to Dr. Nian Wang for allowing me the use of his Hamlin oranges and Dr. Ed Etxeberria for the use of his Valencia oranges. I would like to thank Dr. Charles Sims and Sara Marshall for allowing me to conduct taste and aroma panels in the UF Food Science and Human Nutrition Sensory Panel. Finally, I would like to thank all of my panelists who participated in the taste and aroma panels. 3

4 TABLE OF CONTENTS page ACKNOWLEDGMENTS... 3 LIST OF TABLES... 6 LIST OF FIGURES... 8 LIST OF ABBREVIATIONS... 9 ABSTRACT CHAPTER 1 INTRODUCTION LITERATURE REVIEW Oranges and the Florida Citrus Industry Introduction to Citrus Economics of Florida Citrus Industry Orange Composition Orange Oil Introduction to Orange Oil Cold Press Processing Methods Distilled and Orange Essence Oil Orange Seed Oil Large Scale Study on Florida Cold Pressed Orange Oils Quality of Cold Pressed Orange Oil Flavor Chemistry Introduction to Flavor Chemistry Flavor Volatiles in Citrus Huanglongbing Introduction to Huanglongbing Effect of Huanglongbing on Orange Juice Taste Aroma Physiological qualities Chemical differences Metabolomic study Mitigation and Control of Huanglongbing Removal of infected trees and quarantine Foliar nutrition Antibiotics Thermotherapy Tamarixia radiata

5 Genetic modification of citrus trees MATERIALS AND METHODS Orange Trees and Harvesting Cold Pressed Oil Production US Pharmacopeia Tests for Orange Oil Quality Aldehyde Content Refractive Index Ultraviolet Absorbance Optical Rotation Specific Gravity Juice Quality Tests Brix Titratable Acidity Color ph Sensory Analysis Flavor Panels Aroma Panels Gas Chromatography Statistical Analysis RESULTS AND DISCUSSION US Pharmacopeia Tests Flavor Panels Aroma Panels Gas Chromatography Qualitative Gas Chromatography Quantitative Gas Chromatography Effect of disease stage Effect of seasonality Comparison to literature Orange Juice Study CONCLUSIONS LIST OF REFERENCES BIOGRAPHICAL SKETCH

6 LIST OF TABLES Table page 3-1 Compound equations for quantitative GC Mean US Pharmacopeia physicochemical test results for all samples Differences for US Pharmacopeia tests between asymptomatic and symptomatic oils Differences for US Pharmacopeia tests between early and late oils Differences for US Pharmacopeia tests between and oils p-values for flavor panels with AS and SY oils in model solution and orange juice Panelist preference for AS or SY Valencia oil in flavor panels Results of aroma panels with AS and SY oils Panelist preference for AS or SY Valencia oil in aroma panels Hamlin oils linear retention indices and percent peak areas Valencia oils linear retention indices and percent peak areas Qualitative list of volatile compounds in Hamlin and Valencia oils List of compounds which had 20% concentration difference between asymptomatic and symptomatic samples Hamlin early asymptomatic vs. symptomatic compound concentration Hamlin late asymptomatic vs. symptomatic compound concentration Valencia early asymptomatic vs. symptomatic compound concentration Valencia late asymptomatic vs. symptomatic compound concentration Hamlin asymptomatic early vs. late compound concentration Hamlin symptomatic early vs. late compound concentration Valencia asymptomatic early vs. late compound concentration Valencia symptomatic early vs. late compound concentration

7 4-21 Hamlin peak area comparison to literature data Valencia peak area comparison to literature data Means for orange juice physicochemical tests for all samples Differences for orange juice test mean values between early and late juices Differences for orange juice tests between asymptomatic and symptomatic juices Differences for orange juice tests between and juices

8 LIST OF FIGURES Figure page 3-1 HLB AS (left) and SY (right) Valencia oranges USP Aldehyde content calculation Example ultraviolet absorption plot Calculation for acid correction Titratable acidity measurement for single-strength juice Color number calculation from Tristimulus values Equation for calculation of Kovat s indices Correct panelist descriptor key word summary for flavor panels

9 C Degrees Celsius µl Microliter µm Micrometer LIST OF ABBREVIATIONS 1 H NMR proton nuclear magnetic resonance, a method to determine the structure of molecules ACP AS CLaf CLam CLas cm CPOO FDA FID GC GC/GC-MS GC-MS The Asian citrus psyllid Diaphorina citri Kuwayama. The vector for huanglongbing in America. Asymptomatic, oranges which are huanglongbing positive but do not show severe symptoms Candidatus liberibacter africanus. The most common form of huanglongbing found in Africa. Candidatus liberibacter americanus. A form of huanglongbing found in South America and occasionally in China. Candidatus liberibacter asiaticus. Most common bacteria associated with huanglongbing in Asia, North America, and South America Centimeter Cold pressed orange oil Food and Drug Administration, federal agency responsible for promoting and protecting public health Flame ionization detector, commonly used detector in gas chromatography Gas chromatography Dual gas chromatography mass spectrometry. Two gas chromatographers are used in conjunction with one mass spectrometer Instrumentation which combines gas chromatography with mass spectrometry 9

10 GRAS HLB JBT/FMC KPa KV L mg min ml mm MS nm NPR 1 ph ppm SY US USP Generally Recognized as Safe, a food additive which is considered safe by experts Huanglongbing, or citrus greening disease. Bacterial disease of citrus fruit which causes significant changes to fruit health and quality John Bean Technologies, formerly Food Machinery Corporation. A company which produces commercial orange juice extractors Kilopascal, a unit of pressure Kilovolt, a unit of voltage Liter Milligrams Minute Milliliters Millimeters Mass spectrometry Nanometers Non-expressor of pathogenesis related genes 1, a gene from the Brassicaceae family which can be inserted into citrus trees to confer huanglongbing resistance Potential of hydrogen, measures the acidity or basicity of an aqueous solution Parts per million Symptomatic, orange which are HLB positive and show severe symptoms The United States United States Pharmacopeia, compendium of food and drug information published annually 10

11 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ASSESSING THE QUALITY OF HUANGLONGBING SYMPTOMATIC AND ASYMPTOMATIC ORANGE OILS FROM FLORIDA HAMLIN AND VALENCIA ORANGES Chair: Renée Goodrich-Schneider Major: Food Science By Brittany Martin Hubbard August 2017 Florida grown Hamlin and Valencia oranges were harvested twice each in the and harvest years. During each harvest, both huanglongbing (HLB) asymptomatic and symptomatic fruit were harvested. Cold pressed orange oil was extracted, purified, and analyzed via several quality measurements including the US Pharmacopeia mandated physicochemical tests (aldehyde content, UV absorbance, optical rotation, specific gravity, and refractive index), flavor panels, aroma panels, and qualitative and quantitative gas chromatography. Both symptomatic and asymptomatic Hamlin oils had aldehyde contents below the US Pharmacopeia minimum (1.2%) for both harvest years. For Hamlin Early , there were significant differences between symptomatic and asymptomatic oils for aldehyde content, specific gravity, UV absorbance, and optical rotation. There were no significant differences between solutions made with Hamlin asymptomatic and symptomatic oils in flavor panels; however, aroma panels showed significant differences between Hamlin Early asymptomatic and symptomatic oils. 11

12 Several differences were seen between asymptomatic and symptomatic Valencia oils for the US Pharmacopeia tests for both the and harvest years. Additionally, late season symptomatic Valencia oils exceeded the US Pharmacopeia maximum for specific gravity for both years. There were no significant differences between solutions made with Valencia asymptomatic and symptomatic oils in flavor panels; however, aroma panels showed significant differences between Late Valencia asymptomatic and symptomatic oils. For both Hamlin and Valencia oils, several compounds important to orange aroma had significantly different volatile concentrations between asymptomatic and symptomatic samples. These volatiles included decanal, linalool, citronellal, citronellol, neral, geranial, perillaldehyde, carvone, citronellyl acetate, α-terpinyl acetate, and dodecanal. While most volatiles were found at lower concentrations in symptomatic oil, citronellol was found at significantly higher concentrations in both Hamlin and Valencia symptomatic oil while the esters citronellyl acetate and α-terpinyl acetate were found at significantly higher concentrations in symptomatic Hamlin oil. Research on the effect of HLB on the volatile composition of cold pressed orange oil has not previously been published. This work is important to those working in the flavor and fragrance industries, for whom cold pressed orange oil and/or its chemical fractions are important products used in the formulation of many beverages, flavors and perfumes. 12

13 CHAPTER 1 INTRODUCTION Cold pressed orange oil is a by-product of orange juice processing. An oil emulsion is commonly extracted from the fruit simultaneously with orange juice. Orange oil contains numerous volatile compounds, many of which are important for flavor or aroma. Cold pressed orange oil (CPOO) is a valuable product which is used in the food and beverage industries, as well as the fragrance industry (Braddock 1999). Huanglongbing, also known as HLB or citrus greening disease, is a phloem limited, gram negative bacterial affliction of citrus (Bove 1974). The causative agent is Candidatus Liberibacter spp., and in the United States huanglongbing is vectored by the Asian citrus psyllid (ACP) Diaphorina citri Kuwayama. This disease negatively effects orange production and quality; the total economic impact of the citrus industry on Florida s economy is roughly $9 billion (Hodges and Spreen 2012). Prior to noticeable changes in tree productivity, there are changes in the roots, leaves, and fruit of the citrus tree (da Graca 1991). While it is known that HLB affects the physiology (da Graca and others 1991) of the orange fruit and flavor of orange juice (Baldwin and others 2010), the effect upon cold pressed orange oil is unknown. The purpose of this study is to determine the effect HLB has on the quality of cold pressed orange oil. To determine quality, both Hamlin and Valencia oranges were harvested twice each per season for two years. US Pharmacopeia (USP) mandated tests (aldehyde content, ultraviolet absorbance, optical rotation, specific gravity, and refractive index), flavor and aroma sensory panels, and quantitative and qualitative gas chromatography were utilized to determine if there were differences between huanglongbing symptomatic and asymptomatic Hamlin and Valencia oranges. The gas 13

14 chromatography data was also compared to historical data to determine if huanglongbing affected oranges appeared to be different from control, or unaffected fruit. This study is beneficial to citrus processors as it will assess the quality of huanglongbing asymptomatic and symptomatic CPOO. This was done by comparing these oils to one another, as well as to historical standards. By knowing the quality of oil from both Hamlin and Valencia oranges, orange juice processors can better understand the limitations of huanglongbing affected oils. This may ultimately help processors make decisions as to which orange varieties have optimal aroma compounds in their peel oil for manufacturing standards. Additionally, as this is the first published study reporting on huanglongbing affected cold pressed orange oil, this research will be of interest to those studying the effects of huanglongbing on orange byproducts. 14

15 Introduction to Citrus CHAPTER 2 LITERATURE REVIEW Oranges and the Florida Citrus Industry Citrus is a subgenus of fruit thought to have originated in the Malay Archipelago several thousand years ago. References to citrus fruits date back almost three thousand years, when mandarin oranges were mentioned in 8 th century BC Sanskrit writings (Jahoda 1976). Citrus was introduced to the Western Hemisphere when Columbus brought orange and lemon seeds to Haiti in Oranges were first brought to Florida for cultivation in The first groves in Florida were cultivated in the early 19 th century in Pinellas County (Jahoda 1976). Citrus is grown mainly from N and S within Earth s latitudes. Within these latitudes, the characteristics of planting areas most suited for producing quality citrus are periods of rainfall and drought, periodic freezes and frosts, and seasonal changes, including cool nights (Braddock 1999). Sweet oranges (Citrus sinensis (L.) Osbeck) are consumed both as fresh fruit and juice, and comprise 2/3 of citrus fruit produced in the world (Kimball 1999). There are four types of sweet orange: the common orange, blood orange, acidless orange, and navel orange. The common sweet orange is the variety grown most in Florida (Kimball 1999). Sweet common oranges will be referred to as oranges from henceforth. The two most common Florida orange varieties are Valencia and Hamlin. In Florida, Hamlin oranges are harvested September-March while Valencia oranges can be harvested January-June, with most common harvesting times being roughly in the middle of these time frames (NASS 2016). 15

16 Orange trees are generally propagated by grafting (or budding) a twig from a tree to rootstock grown from seeds. The new tree will combine the desired traits of the rootstock and the original tree. Seedlings are grown in a nursery and transferred to a commercial grove at 1-2 years of age. Most orange varieties mature ten months after blossoming, but Valencia can take twelve months or more (Albrigo and Sauco 2004). Oranges for the fresh market are usually handpicked and placed into bags carried by individual harvesters. Oranges for the processed market may be hand harvested or could be mechanically harvested by a canopy shaker which shakes fruit to the ground, which is then packed onto trailers for transport to citrus processing plants. In Florida, 94% of oranges are grown for processing and the remainder are sold whole on the fresh market to consumers (Braddock 1999). Economics of Florida Citrus Industry Florida is the leading producer of citrus in the United States, producing 56% of the nation s citrus in the season (NASS 2016). Florida s citrus production was million boxes (90 pounds of fruit per box) for the season, which was a decrease of 9% from the previous year and a decrease of 61% from the season (NASS 2016). Citrus is a $1.35 billion dollar industry in Florida and accounts for 76,000 jobs. The total economic impact of the citrus industry on Florida s economy is roughly $9 billion (Hodges and Spreen 2012). There are approximately 40 packing houses and 15 juice processers within the state. While most citrus produced in Florida is consumed in the US, approximately 7.0 million ¾ bushels of fruit was exported in , mainly to Canada and Japan (NASS 2016). 16

17 Orange Composition Orange peels are composed of two parts: the flavedo and the albedo. The flavedo is the orange colored outer peel and the albedo is the white inner peel. Orange peels have oblate-shaped oil glands in the flavedo. These cells also contain compounds which are the result of metabolic processes in the fruit as the fruit matures. The primary flavedo chemical class are terpenes, with d-limonene being the most prevalent compound (Braddock 1999). The presence of limonene and oil in the flavedo deters insects from damaging the orange fruit (Braddock 1999). Hesperidin is a bioflavonoid with antioxidant properties. The flavedo and albedo are both rich in this flavanone glycoside, which can be commercially extracted with hot methanol (Kesterson and Braddock 1976). This process can obtain approximately 8-10 pounds of hesperidin per ton of peel residue. Citrus fruits contain greater than 85% water. In oranges, the remaining dry solids fraction is composed of sugars (10%), fiber (2%), protein and amino acids (1%), organic acids (1%), minerals (0.7%), and lipids (0.3%) (USDA 1982). The most common sugars are sucrose, fructose and glucose. Very small amounts of rhamnose and xylose are also present (Ting and Deszyck 1961). The fiber fraction contains pectin, cellulose, hemicellulose, and some lignin. The most common organic acid in the orange fruit is citric acid (Kimball 1999). Malic, oxalic, and malonic acids are also present. Malonic and oxalic are the major peel organic acids. (Sinclair and Eny 1947). Lipids are found in the seed and peel and are composed mainly of triglycerides, phospholipids, and fatty acids. Many minerals exist in oranges, as virtually every soluble mineral found in the soil is found in oranges and their juice (Braddock 1999). The most significant mineral in 17

18 orange juice is potassium ( mg/l), followed by phosphorus ( mg/l), calcium ( mg/l), and magnesium ( mg/l). The most abundant vitamin is Vitamin C (ascorbic acid) which has a concentration of mg/100ml, followed by significant but smaller amounts of folic acid and thiamin. Orange juice also contains significant levels of the flavonoid hesperidin ( mg/l) and pectin ( mg/l) (Braddock 1999). Citrus peel is composed mainly of carbohydrates. The water soluble fraction contains sucrose, glucose, fructose, and xylose. The insoluble fraction contains pectin, cellulose, and hemicellulose. Polyunsaturated lipids, including linoleic and linolenic acid are present in the peel. Oil can be extracted from the orange peel and marketed for several purposes including as an additive to foods and beverages and an aroma for perfumes, cosmetics, and personal care products. Orange Oil Introduction to Orange Oil Cold pressed orange oil (CPOO) is a valuable commodity extracted from orange peels, a product which would otherwise become waste or animal feed (Kesterson and Braddock 1976). After juice is extracted, approximately 45% of the orange weight remains as refuse (Kesterson and others 1971). This includes the peel, rag, seeds, and pulp. The quality of orange oils is dependent upon a number of factors including climate, soil, weather, fruit maturity, and extraction method (Kesterson and others 1971). Hood (1916) found wide variation in the oil yield of Florida oranges, reporting values of % oil per total fruit weight. Oil content reaches a maximum when the orange is fully mature. However, commercial quantities of oil are present before the fruit is ready for 18

19 harvest (Kesterson and others 1971). Oil content decreases after a period of rainfall and will increase after a period of drought. Cold Press Processing Methods The most popular commercial model for cold pressing orange oil is the JBT (formerly FMC) in-line extractor; the Brown oil extractor is the second most common model. (Braddock 1999). These two extractors differ mainly in the way the orange is treated for extraction. The JBT in-line extractor juices the orange and extracts oil simultaneously. This method uses pressure and a shredding mechanism to extract oil from the peel. The main drawback of this system is a higher oil content in the juice as well as more damaged pulp (Tetra Pak 1998). The JBT extractor separates oranges into four parts: an oil emulsion, peel, core, and juice. The extractor is comprised of an upper and lower cup. A mechanism pushes through the lower cup to cut a hole in the fruit, through which the core will be removed. The upper cup presses down on the lower cup. Pressure causes the juice to squeeze out of the fruit, which is strained through the lower cup, along with some pulp. The pressure also causes the peel to break apart and the peel and core are caught in the strainer cup. As the peel is forced through the fingers of the cup, oil is extracted from the peel, which is washed with water to form an oil emulsion (Tetra Pak 1998). The Brown oil extractor is a reamer-type extractor which releases oil before juice extraction. The peel is punctured with sharp needles, and is washed with water to create an oil emulsion. Because the flavedo is pricked rather than abraded as it is with the JBT extractor, the oil emulsion is cleaner, centrifuges well, and the orange juice contains less oil (Tetra Pak 1998). Downstream, oranges are cut in half, then reamers 19

20 penetrate fruit, squeezing juice through one outlet and rag material through another. The juice yield of Brown and JBT extractors is similar (Kesterson and others 1971). The following general processing procedure for extracting oil from orange peel is followed, regardless of extraction method (Kesterson and others 1971). Oil cells in the flavedo are abraded or ruptured by pressure. The oil is then washed away from the fruit in a stream of water. Excess water is necessary so the oil is not reabsorbed by the peel. The crude oil emulsion is put through a finisher with a screen to recover oil present in the slurry. The finisher pressure should not be excessive to avoid including too much pectin or insoluble solids in the emulsion. This would create high emulsion viscosity and make it difficult to separate the oil by centrifugation. This finished emulsion is sent to the desludger, a centrifuge which produces a product which should be at least 70% oil. This emulsion is sent to a polisher centrifuge, which is the final step before the oil is blended and dewaxed in large stainless steel drums. Dewaxing is dependent upon temperature and time. For example, oil can be dewaxed for five days at -10 F or for three weeks at 25 F. For storage, oxygen is displaced in these drums by carbon dioxide or nitrogen gas. The ideal storage temperature is F. Distilled and Orange Essence Oil Commercial oil is obtained from oranges in two ways other than peel extraction: from the deoiler and from the evaporator in the juicing process. Oil obtained from the deoiler is called distilled orange oil. High levels of oil (>0.035%) are considered a defect in orange juice, so oil generally needs to be removed from the juice prior to packaging (Braddock 1999). To do so, juice is flash heated in a deoiler. Decanting or centrifugation follows. Distilled oils have a higher aldehyde and lower ester content than the best quality cold pressed oils (Braddock 1999). 20

21 Oil obtained from the evaporator when juice is made into concentrate is called essence oil. Essence oil contains flavor and aroma enhancing compounds and has a fruity aroma characteristic of fresh juice (Kesterson and others 1971). Essence oils contain valencene, a sesquiterpene not appreciably present in cold pressed oils. Valencene can be recovered and converted to nootkatone and used as a flavor enhancer. Good quality essence oils may be added back to juice or other beverages, or added to perfumes or chewing gum (Braddock 1999). Orange Seed Oil Significant quantities of oil can be extracted from orange seeds. Orange seed oil was a commercial product in the past, however, it is no longer produced due to the intense bitter taste of the crude oil (Kesterson and Braddock 1976). This bitter taste is due to the presence of excessive limonin. To produce seed oil, orange seeds are first dried then passed through a roller mill to loosen the hulls. The dehulled seeds are pressed by an oil expeller at high pressure. The oil is clarified in a plate and frame press which reduces the bitter taste. The highest oil content generally coincides with the optimum maturity for juicing (Kesterson and Braddock 1976). This type of oil may never be commercialized again. The high quantity of linolenic acid in orange seed oil means partial hydrogenation is necessary to produce a product with a sustained shelf life. This is no longer acceptable since the Food and Drug Administration (FDA) revoked the generally recognized as safe (GRAS) status of trans fats (FDA 2015). Large Scale Study on Florida Cold Pressed Orange Oils Kesterson and others (1971) performed an analysis on more than 800 samples of Florida cold pressed and distilled citrus oils from commercial processing plants. This group used methods of analysis including optical rotation, refractive index, specific 21

22 gravity, aldehyde determination, and evaporation residue. These properties are standards of purity determined by the US Pharmacopoeia (USP) (Food Chemicals Codex 2010). The factor which influenced the chemical and physical properties of cold pressed orange oil to the greatest extent was the yield of oil from the peel. Overall, the differences in physical properties of orange oils from different fruit varieties by any particular process were not significant. Differences were seen in oils produced from oranges of disparate maturity and storage time as well as between different oil extraction methods. As yield increased specific gravity, evaporation residue, and refractive index increased. Optical rotation decreased. As yield increased more high-boiling point, high molecular weight constituents were extracted. The presence of a higher percentage of these compounds causes reduction in the percentage of d-limonene. This results in the observed lower optical rotation values, as d-limonene is the most optically active component in the oil. The oil yield obtained by various processing methods varied from pounds per ton of peel. Many plants have more peel than it is possible to process, so oil is partially extracted from a large quantity of peel rather than producing the maximum recovery of oil from a smaller quantity of peel. Analyses of expressed orange oils indicate oil produced by some manufacturing processes at certain times during the season did not meet USP specifications because yields were too low or too high. An estimated yield of pounds of oil per ton of peel from midseason oranges or an extract of 45-60% of the total amount of oil in a peel of any variety of mature fruit generally results in oil that will meet the USP specifications. Oil extracted in mid-may 22

23 from Valencia oranges past peak maturity did not meet USP standards. This is because the peel had become soft and pliable, making oil extraction more difficult. The aldehyde content of orange oil is important to flavor quality. In this study, the average aldehyde content varied from %. The best method to produce CPOO with a high aldehyde content is to reduce the amount of water that comes into contact with the oil during processing. When the aqueous phase volume is increased, the amount of insoluble solids in the extraction mixture also increases. These are primarily pulp particles, which act as ion-exchange resins and absorb oil constituents such as aldehydes. Storage time of fresh oranges prior to processing had an effect on the chemical but not the physical properties of the resulting oil. Oranges stored in bins for three to five days prior to extraction and oranges extracted for oil the same day as harvest showed no differences in physical properties. However, the stored oranges did show significant differences in chemical properties, including increased ester content and evaporation residue and decreased aldehyde content. Differences in maturity levels also led to significant differences. The aldehyde content of oranges increased as the fruit matured. The maximum aldehyde content was seen in early season fruit, which was just past maturity standards. Ester content was lowest in oil produced from early season fruit and increased as the growing season progressed. Annual rainfall has an effect on the properties of refractive index and aldehyde content in orange oils. Aldehyde content is an important criterion related to the flavor and aroma of orange oils. Oranges with a low aldehyde content are generally 23

24 considered inferior. Aldehyde content and total rainfall show a positive correlation. It was suggested by the authors that this effect could be mitigated by irrigation during low rainfall years. Quality of Cold Pressed Orange Oil The ultraviolet absorption of citrus oil is an important evaluation criterion (Sale 1953). The ultraviolet absorption is dependent upon both the type of citrus the oil is extracted from as well as the type of machine used to extract the oil (Kesterson and others 1971). Thus, it can be used to identify the origin of a citrus oil and determine if adulteration may have occurred, but it is not a quality measurement. Reaction of orange oil with oxygen creates undesirable flavors and odors described as turpentine in character (Kesterson 1971). Barrels of orange oil have their oxygen headspace displaced with nitrogen or carbon dioxide, but inclusion of oxygen is a concern when orange oil is used in a product development application. For this reason, antioxidants are often added to formulated products as stabilizers. Some effective antioxidants include hydroquinone, α-tocopherol, and nordihydroguaiaretic acid (Kesterson 1971). Bacterial degradation of orange oil can also be problematic. Orange oil emulsions are good mediums for microbial growth (Murdock and Hunter 1968). Microbes form α-terpineol in citrus oil, by degradation of limonene (Marostica and Pastore 2007). This process creates unfavorable odors and flavors which result in unmarketable oil. It is important that oil slurries not be held prior to the desludging process, as this is when bacterial growth is likely to occur. 24

25 Flavor Chemistry Introduction to Flavor Chemistry Flavor chemistry is the study of compounds which elicit flavor, a complex sensation composed of taste, olfaction, and somatosensory inputs including mouthfeel, temperature, burning, and tingling sensations in the mouth or nasal cavity. The known tastants are sweet, sour, salty bitter, and umami. The combination of basic taste and retronasal olfaction creates the sensation of flavor. Retronasal olfaction occurs when volatiles released from masticated food particles are forced behind the palate and enter the nasal cavity from the rear. Conversely, olfaction occurs when small volatile molecules are inhaled through the nose. This leads to the sensation generally referred to as smell, which is separate from taste and flavor. However, the olfaction of volatiles is an important component of flavor chemistry and thus smell will be discussed in relation to this discipline alongside taste. Flavor compounds can be assessed in two main ways: by instrument, or by man. Humans can evaluate flavor by smelling or tasting a food product or a solution in which a compound has been isolated. Flavor tests by humans can take several forms. Some basic sensory tests, such as likeability tests, do not evaluate flavor compounds directly but instead ask sensory panelists how much they enjoy a product on a set scale. For example, if a researcher is testing how much HLB affected juice can be blended with unaffected juice to create an acceptable product, a likeability test may be performed. Similarly, panelists may be asked on a just-about-right scale how they feel about the orange flavor, sweetness, bitterness, or any other component of a product. These tests may give a big picture assessment of potential issues with the flavor of a product, but 25

26 flavor compounds are not directly evaluated. Often, human sensory testing is correlated to instrumental testing of volatiles. Gas chromatography combined with mass spectrometry (GC-MS) is the most common instrumental assessment of volatile flavor molecules. Volatiles can be captured for analysis by a number of methods; which method a researcher chooses is dependent upon several factors. These include whether the volatiles need to be concentrated, whether quantitative or qualitative analysis is desired, and the polarity of the compounds of interest. Solid phase microextraction (SPME) with dynamic headspace can be used to absorb volatiles for instrumental analysis. Dynamic headspace concentrates volatiles present in low concentration so they can be assessed by gas chromatography (Reineccius 2006). Orange oil can be diluted with a polar solvent and directly injected into a gas chromatographer (Dugo and others 2011). A flame ionization detector (FID) is the most common detector used to analyze citrus flavor compounds (Dugo and others 2011). After analysis by a GC detector, the resulting gas chromatogram will show compound peaks with retention times. These retention times can be compared to standard compounds, literature values, or library databases of flavor compounds run on the same column under the same conditions. Mass spectrometry (MS) is additionally used to identify compounds because mass fragmentation data can be used to interpret the exact molecular weight of compounds analyzed by mass spectrometry. Flavor Volatiles in Citrus The primary class of volatile compounds present in citrus oils are terpenes. Over 95% of the volatile compounds in sweet orange oils are terpenes (Reineccius 2006). The primary terpene present is d-limonene, which has a slightly citrusy aroma. 26

27 Although terpenes are the largest component of orange oil, they are thought to have only minimal flavor impact. Many processed citrus oils have their terpene portion removed for quality purposes; these are referred to as terpeneless oils (Reineccius 2006). Other volatiles present in appreciable quantities in citrus oils are α-pinene, sabinene, β-myrcene, octanal, linalool, decanal, neral, and geranial (Mitiku 2000). α- pinene and sabinene are terpenes which have aromas reminiscent of pine and spruce trees, respectively. β-myrcene is a terpene with an aroma reminiscent of hops. Octanal and decanal are aldehydes, octanal smells fruity while decanal has a characteristic orange aroma. Linalool is a terpene with a fruity, flowery aroma. Neral and geranial are aldehydes which smell lemony and flowery, respectively (UF CREC 2006). The major general classes of compounds found in sweet orange oils are terpenes, aldehydes, esters, and alcohols. Approximately 80 compounds are frequently identified in CPOO form sweet oranges (Dugo and others 2011). These compounds include 34 terpenes, 17 aldehydes, 9 esters, 16 alcohols, 2 ketones, and 4 oxides. Twenty-two compound were identified as being aroma-active in CPOO (Qiao and others 2008). The most odor active of these were α-pinene, β-myrcene, linalool, citronellal, α- terpineol, and decanal (Qiao and others 2008). Huanglongbing Introduction to Huanglongbing Huanglongbing (yellow dragon disease), also known by its initialism HLB, is a bacterial disease of citrus trees caused by Candidatus Liberibacter spp. and in the United States vectored by the Asian citrus psyllid (ACP) Diaphorina citri Kuwayama (da Graca 1991). There is no known cure for HLB once a tree is infected (Bove 2006). 27

28 Before noticeable changes in the productivity of a tree occurs, there are noticeable changes in the fruit, leaves, and roots (da Garca 1991). The root system begins to decay and leaves display asymmetrical leaf chlorosis, commonly called yellow blotchy mottle. Symptomatic (SY) fruit are small, somewhat green in color, misshapen, and contain aborted seeds (da Garca 1991). HLB is present in many citrus growing regions of the world including Asia, Africa, South America, and North America (Invasive Species Compendium 2016). There are three different species of Candidatus Liberibacter: Candidatus Liberibacter asiaticus (CLas), Candidatus Liberibacter africanus (CLaf), and Candidatus Liberibacter americanus (CLam). CLas is the most prevalent variety; it is found throughout citrus growing regions in Asia as well as in North and South America (Gottwald 2010). CLam is most commonly found in South America, but it has been reported in China as well (Gottwald 2010). Brazil, the world s largest producer of oranges, initially had a higher population of CLam but CLas is now the more prevalent subspecies (Gottwald 2010). CLaf is present in Africa as well as the Middle East (Gottwald 2010). While D. citri predominately transmits CLas and CLam, CLaf is transmitted mainly by the African citrus psyllid, Trioza erytreae. Unlike CLas and CLam, CLaf is heat labile and is susceptible to temperatures higher than approximately 30 C (Bove and others 1974). The earliest known description of trees with HLB symptoms was from India in the 1700s (Capoor 1963). The disease was well characterized in India and China by the early 1960s (Capoor 1963, Lin 1963) but the disease was unknown in the US until relatively recently. The vector D. citri was first discovered on orange jasmine (Murraya paniculata) hedges in Palm Beach County in 1998 (Halbert and Manjunath 2004), but 28

29 HLB was not observed until Recently, the ACP has spread west to Texas and California (Stokstad 2012). The first HLB infected tree was found in California in 2012; several dozen dooryard trees in residential neighborhoods have been identified as HLB positive since then (UC IPM 2016). More than 1700 trees in Texas have become infected since HLB was first discovered in the state in 2012, this has led to a citrus quarantine in six Texas counties (Texas Department of Agriculture 2016). Compared to other vectored crop diseases, HLB moves relatively slowly (Gottwald 2010). A worst case scenario may occur when a grove with young trees (less than three years of age) is infected and there is no control of the vector population or removal of infected citrus. The disease will then likely reach an incidence of 50% in 3-5 years (Gottwald 2010). In older groves under the same conditions, a 50% infection incidence will generally not occur until five years post infection. Severe symptoms resulting in crop loss usually occur 1-5 years after the onset of HLB (Aubert 1992). Fruit quality and yield decrease with increased symptom severity. An infected grove will likely be economically infeasible 7-10 years after infection (Aubert 1990). In studies comparing fruit unaffected (control) and affected (SY) by HLB, the determination of which group a tree belongs in is commonly done one of two ways. This first is visual inspection, where a researcher inspects a tree for the physical defects accompanying HLB and determines whether or not a tree is infected. HLB will often not affect the entire tree, so some sections of an infected tree may be SY and others may be asymptomatic (AS). The AS fruit appears similar to healthy control fruit. The second way a determination can be made is by polymerase chain reaction. In addition to the trees identified as being HLB positive by polymerase chain reaction, there is an 29

30 additional assumed population of trees whose disease stage is subclinical at the time of monitoring (Irey and others 2006). Effect of Huanglongbing on Orange Juice In addition to having undesirable physical differences from unaffected fruit, oranges affected by HLB also show differences in taste, aroma, physiological, and chemical properties. Taste Research showed SY fruit had a Brix/acid 62% lower than control fruit (Dagulo and others 2010). Thus, the HLB SY fruit was more sour and less sweet than unaffected fruit. Limonin, the primary bitter compound in oranges, was found to have a concentration as much as 400% greater in SY fruit than unaffected fruit. This observed concentration is lower than the bitterness threshold for most individuals, however, more sensitive individuals may be able to perceive a difference at this level (Dagulo and others 2010). In a sensory panel analyzing fruit from the 2007 harvest year, experienced panelists found flavor differences between SY and control Hamlin juices in a difference from control test (Plotto and others 2010). In their study, some panelists described SY fruit as fermented, sweeter, overripe, sour, or bitter. There was a positive correlation (r=0.860) between difference ratings and limonin, which as discussed above may elicit a bitter taste in orange juice to the most sensitive panelists. The same panel also found juice from SY Midsweet fruit to be different from the control, and participants also described the SY fruit as fermented, sweeter, and overripe. Chemical analysis showed fruit from SY trees had a higher ethanol content than control fruit whereas Valencia fruit did not show significant differences between SY and control fruit. Compared to the 30

31 2007 season, somewhat different results were seen for the 2008 season. Similar to the 2007 season, juice from symptomatic Hamlin fruit showed significant differences compared to juice from control fruit in a difference from control test (Plotto and others 2010). Panelists found symptomatic Hamlin juice to be bitter, sour, grapefruit-like, peppery, and metallic. A trained panel described the same SY juice as fatty, fermented, salty, or musty/earthy. Like juice from Valencia oranges harvested in the 2007 season, Valencia oranges harvested in the 2008 season did not show significant differences between control and symptomatic juice in taste panels (Plotto and others 2010). Overall, SY Hamlin juice had more off-flavors than Valencia or Midsweet juice, and was most commonly described as bitter, sour, or fermented. Juice from SY and control fruit can be blended to create a palatable product. Raithore and others (2015) investigated the extent by which blending was possible without making the blended juice easily distinguishable from the control. At 25 ml of juice from SY fruit per 100 ml of total juice, 27% of panelists could recognize a taste difference. At 6.25 and 12.5 ml of SY juice per 100 ml juice blend, less than 9% of panelists could recognize taste differences. These results show that blends exceeding 25% SY juice could not likely be used, while 12.5% blends would be very acceptable to most consumers. Aroma Juice made from SY fruit has a lower concentration of esters, aldehydes, and total sesquiterpenes than juice extracted from HLB unaffected fruit. Juice from SY fruit had increased levels of alcohols and terpenes. Overall, these changes are comparable to juice made from immature fruit, which has a less favorable aroma profile than juice made from mature HLB unaffected fruit (Dagulo 2010). Juice made from SY and control 31

32 Hamlin oranges were shown to have significantly different aromas in difference from control tests (Plotto and others 2010). There were no significant differences for Midsweet and Valencia juices. Physiological qualities Baldwin and others (2010) investigated the quality of three orange varieties (Hamlin, Midsweet, and Valencia) for two seasons to determine if there were differences in the physiological qualities of juices made from SY and control fruit. Hamlin and Midsweet, an early and midseason variety respectively, did not show differences for soluble solids content (SSC), titratable acidity (TA), SSC/TA, total sugars, individual sugars, or galacturonic acid content for the 2007 season. SY and control Valencia oranges did show some differences in chemical properties for different months of harvest. For example, SY oranges had significantly lower SSC in the May and June harvest months than in the March and April harvest months, although there was no significant difference in the SSC/TA ratio. SY oranges also showed differences in total sugars for the April, May, and June harvest months, although no significant differences were seen in February. These differences were for glucose and sucrose but not fructose. Galacturonic acid, which represents the amount of pectin in a juice, was significantly higher for the March harvest in SY fruit but showed no differences in the other harvest months. SY and control Valencia fruit harvested in the 2008 season showed no significant differences for any month in SSC and SSC/TA, which is different from what was seen in The researchers believe this may be because different groves of Valencia oranges had to be used in this study for the 2007 and 2008 seasons (Baldwin and others 2007). The 2007 grove had many trees eradicated due to the prevalence of HLB 32

33 and could not be used for the 2008 season. These results show that differences between SY and control HLB infected fruit may differ depending on the grove. Chemical differences The same study by Baldwin and others (2010) investigated chemical properties of orange juice from SY and control trees. Changes in concentrations of secondary metabolites can reflect stressful conditions in plants. In both 2007 and 2008, many secondary metabolites were higher in SY than control juices. These included the chemical classes of hydroxycinnamic acids and sesquiterpenoid limonoids. Specifically, SY Hamlin and Valencia oranges showed an increase in sesquiterpenoid limonoids while Midsweet oranges showed increases in both sesquiterpenoid limonoids and hydroxycinnamic acids for both the 2007 and 2008 season. Differences in volatile concentrations were also seen between SY and control fruit (Baldwin and others 2010). During the 2007 season, hexanol levels were higher in juice from SY Hamlin from than juice from control fruit. Ethanol was higher in Midsweet SY fruit, while hexanal, cis-3-hexenol, and linalool were higher in control fruit. Valencia fruit did not show any variation in volatiles by disease status but did show differences by harvest times. During the 2008 season, no differences were seen in the Midsweet fruit. Hamlin showed higher levels of acetaldehyde, octanal, ethanol, cis-3-hexenol, and sabinene in control fruit and higher levels of β-myrcene and ethyl hexanoate in SY fruit. SY Valencia oranges showed higher levels of ethanol and sabinene and control fruit showed higher levels of hexanol and ethyl acetate. These data show there can be season to season and cultivar differences in how HLB affects volatile concentrations in juices from these fruit. 33

34 Metabolomic study Metabolomics is the study of the small molecules which are the intermediate or end products of metabolism in biological systems. This area of research can give insight into both the natural equilibrium state of an organism as well as an organism s response to environmental, genetic, biotic, and abiotic changes (Metabolomic Society 2014). Metabolomics is a relevant research tool for evaluating the physiological and chemical changes which occur in oranges as a result of HLB. Slisz and others (2012) conducted a study investigating the differences in the metabolomic profiles of orange juice procured from HLB SY, AS, and control (unaffected) fruit. Specifically, 1 H NMR coupled with multivariate analysis was used to quantify the metabolites in the juices. Juice from SY fruit had a phenylalanine concentration three times higher than juice from control fruit, and a citrate concentration twice as high as the control. Juice from control fruit had a sucrose concentration three times higher than that of juice from SY fruit. Additionally, juice from SY fruit had decreased concentrations of amino acids (leucine, proline, valine, alanine, isoleucine, threonine, and arginine), glucose and fructose, adenosine, malate, and choline. Juice from SY fruit had increased concentrations of limonin, limonin glucoside, formate, ascorbate, synepherine, histidine, and asparagine. Very similar results were obtained when comparing SY to AS fruit, suggesting AS fruit is metabolomically similar to SY fruit (Slisz and others 2012). Decreased proline and arginine found in SY fruit were unexpected results, as these amino acids normally increase when a plant is under stress. The authors suggested the decrease in these compounds may be due C. Liberibacter suppressing the orange fruit s defense mechanisms (Slisz and others 2012). 34

35 Mitigation and Control of Huanglongbing Removal of infected trees and quarantine There is no cure for HLB. However, techniques for controlling or mitigating HLB are under investigation. The best method during early infection is removal of affected trees to prevent the spread of HLB to other nearby trees (Spann and others 2010). This method is currently being employed with success in California, where only a few dozen HLB positive trees have been identified (University of California IPC 2016). In California, a strict monitoring system is in place to detect the ACP. If psyllids are detected, the area is quarantined to prevent further spread of the vector. HLB is more widespread in Texas than in California, but several counties are quarantined to try to prevent further spread of HLB (Texas Department of Agriculture 2016). Rigorous tree monitoring is necessary because eliminating inoculum via removing infected trees is more efficacious than attempting to manage the ACP population (Spann and others 2010). Foliar nutrition Foliar nutrition is a method which has been used to reduce the symptoms of HLB in citrus trees and fruit. The purpose of foliar nutrition is not to prevent or cure HLB but to produce high yields of healthier fruit than what is usually observed with HLB infection. Foliar nutrition allows leaves to acquire essential nutrients they may otherwise be unable to obtain due to HLB induced root damage (Tansey and others 2016). Foliar nutrition is generally applied three times per year (spring, summer, and fall) when new leaf flush is fully expanded, but not hardened (Stansly and others 2014). A common foliar treatment is a mixture of macronutrients (for example, with K-Phite and fertilizer) and micronutrients (magnesium sulfate, zinc sulfate, 35

36 sodium molybdate, and magnesium sulfate) (Stansly 2014). Foliar nutrition was first used by Maury Boyd in 2006 (Spann and others 2011). In a time when the prevailing wisdom was to remove HLB affected trees, Boyd believed his infection rate was too high to eradicate trees without removing nearly his entire grove. He instead chose foliar nutrition to help maintain the health and yield of his trees. Tansey and others (2016) reported positive results for foliar treatment. This study investigated the effects of foliar nutrition for the years In , foliar nutrition led to significantly increased yields; yield was not increased in Additionally, this study showed that insecticides were more important than foliar nutrition in promoting fruit yield, as insecticides control the ACP population. The authors of this study ultimately concluded that the cost of foliar nutrition may outweigh its benefits. There is some controversy as to whether foliar nutrition is effective. Gottwald and others (2012) conducted field trials from and determined foliar nutrition had no significant effects on yield or fruit quality. Stansly (2014) studied foliar nutrition beginning in For the harvest years , foliar nutrition only showed increased yield for In 2010, there were no differences between trees receiving foliar nutrition and control trees. Yield increases were most significant compared to control when foliar nutrition was used in conjunction with insecticides (Stansly 2014). Costs per hectare for foliar nutrition were $1499 for the season (Tansey and others 2016). Insecticides, which may be more efficacious in increasing yield compared to foliar nutrition, cost only $689 per hectare for the 2012 season (Stansly 2014). Additionally, foliar nutrition did not have any effect on the Brix/acid ratio, an orange quality parameter, for (Stansly 2014). It should be noted that the 36

37 authors of this study state that adverse conditions in 2010 and 2011 may have led to the poor results in comparison to 2012, when growing conditions were more ideal. Antibiotics Bacterial plant diseases are difficult to control. This is particularly the case when the bacteria resides in the phloem (Slinski 2016). An emergency exemption passed in March of 2016 allows growers to use antibiotics to treat citrus trees affected by HLB. Under the Environmental Protection Agency s Federal Insecticide, Fungicide, and Rodenticide Act growers are now allowed to use oxytetracycline calcium, oxytetracycline hydrochloride, and streptomycin sulfate to treat HLB (Federal Register 2016). Oxytetracycline and streptomycin are the most common antibiotics used on agricultural crops (Slinski 2016). Streptomycin is bactericidal while oxytetracycline is bacteriostatic. Alternating bactericidal and bacteriostatic treatments is common, and may be especially pertinent for an antibiotic such as streptomycin which is important in treating human conditions. The University of Florida IFAS Extension Service published a helpful information sheet detailing which antibiotic is best administered in which month depending on the citrus variety (Dewdney and Graham 2016). Suggested volumes of product, maximum number of treatments per year, minimum retreatment intervals, and necessary personal protective equipment for product handlers are also detailed in this information sheet. A 2014 study by Zhang and others evaluated 29 antibiotics for effectiveness in treating HLB. It was ascertained that 6 antibiotics were highly effective, 11 were noneffective, and 12 somewhat effective. The six highly effective antibiotics were: ampicillin, carbenicillin, penicillin, cephalexin, rifampicin and sulfadimethoxine. This study did evaluate streptomycin sulfate and oxytetracycline hydrochloride, which are 37

38 two of the antibiotics approved to treat HLB. Oxytetracycline had a significant level of phytotoxicity, with less than 6.3% of scions surviving the treatment process. In contrast, for all but one other antibiotic, scion survival rates after treatment were % (Zhang and others 2014). Due to poor survival rate, oxytetracycline hydrochloride was eliminated from further analysis in this study. Streptomycin sulfate was deemed noneffective. Beta-lactam antibiotics, including penicillin, were the most successful in this study, with scions no longer being HLB positive after six months of treatment. Additionally, there was no transmission of HLB to rootstocks (Zhang and others 2014). It should be noted that penicillin is currently not used in plant agriculture and would likely take many years to register due to its importance in treating human conditions (Slinski 2016). A recent field study investigated the effectiveness of oxytetracycline hydrochloride by trunk injection (Hu and Wang 2016). CLas populations were reduced by 99% in the leaves and 95% in the roots. While higher fruit yield and lower juice acidity were seen compared to control untreated trees injected with water, these differences were not significant. Several older studies have shown the efficacy of using oxytetracyclines to treat HLB (Capoor and others 1974, Schwartz and others 1974, Kapur 1996). Applications can be made either by spraying the foliage or via direct trunk injection, the most efficient method being trunk injection. Trunk injection is not often used because it is not cost effective (Slinski 2016). Additionally, in South Africa where greening has been established for many years, use of tetracyclines is uncommon due to phytotoxicity, residues in fruit, and non-sustained disease suppression (Buitendag and Von 38

39 Broembsen 1993). It should be noted that the variety of HLB found in South Africa, CLaf, is not heat tolerant. Many of the issues with HLB in Africa have been solved by moving orange groves to warmer regions. Unfortunately, CLas is not heat susceptible so the same method of moving growing areas cannot be used to solve the issues faced in the US. Thermotherapy Thermotherapy is an emerging method of combating HLB symptoms in citrus. This method has been used for nearly a century to eliminate or prevent plant diseases, beginning with a report in the 1930s on using thermotherapy to cure a viral yellowing disease in peaches (Kunkel 1936). Thermotherapy is commonly used to treat diseases of sugarcane, including grassy shoot disease (Candidatus phytoplasma) and ratoon stunting disease (Leifsonia xyli subsp. xyli) (Viswanathan 2001). Additionally, thermotherapy has previously proven effective in inactivating the Citrus tristeza virus, which causes citrus quick decline disease (Nyland and Goheen 1969). To utilize thermotherapy for citrus, a single tree is covered with a plastic shelter. Heated air or steam can then be introduced to the shelter to treat each tree. Alternatively, solar radiation can be used to heat the trees (Ehsani and others 2013). In Florida, solar radiation can generally only be used May-September when there is enough solar radiation to bring the shelter to an appropriate temperature. A preliminary field trial using solar radiation on highly visually symptomatic trees showed that trees treated in mid-summer for a five hour heating period visually looked healthier than untreated trees. Treated trees also had fruit yield and juice quality similar to control trees unaffected by HLB. Trees treated in late summer did not achieve the same results. Due to cooler temperatures and cloud cover, the shelter did not reach the 39

40 time and temperature (45 C for 4 hours) most successful in eliminating HLB (Ehsani and others 2013). While solar radiation is more cost effective than injecting steam into a tent structure, it is not the optimum method because it is dependent on factors such as temperature, rain, wind, and cloud cover. Thus, solar radiation cannot produce consistent results over the course of a growing season. Hoffman and others (2013) demonstrated that the high temperatures used in thermotherapy can cause non-permanent damage to citrus trees. Grapefruit (C. paradisi) trees were treated by thermotherapy for cycles of 42 C for 19 hours then 30 C for 5 hours for 2 days. No CLas was detected 30, 60, or 270 days after treatment. These trees showed initial defoliation but had healthy new growth 30 days after treatment (Hoffman and others 2013). In the same study, 42 C continuously for 10 days also eliminated HLB but caused scorching of leaves. These plants eventually recovered and continued to grow without symptoms of HLB. The authors explained heat stress may be the mechanism by which thermotherapy eliminates HLB. Changes evoked by heat stress include increased levels of secondary metabolites, hormones, stress-related proteins, and reactive oxygen species (Hoffman and others 2013). Both root growth and a strong flush were observed after thermotherapy treatment in this study; it is probably that heat stress induced metabolomics changes are the impetus for these differences. Additionally, heat stress directly on CLas could be the source of the titer reduction. Two prophages present in CLas can alter the lytic cycle (Zhang and others 2011). In other bacteria it has been proven that the lytic cycle can be altered due to heat stress (Chu and others 2011). If this is true for HLB, then the heat stress induced by thermotherapy may be directly killing CLas in citrus trees. 40

41 A recent study in China (Fan and others 2016) evaluated treating Mandarin oranges (C. reticulata) with thermotherapy at 45 C and 48 C for 4 hour intervals once a week for 3 weeks. Plants had more than a 90% decrease in CLas titers in leaf flush and a 56-60% decrease in CLas titers in mature leaves 8 weeks after initial treatment. While there were still detectable levels on CLas in plants, all plants appeared healthy six months after treatment with no signs of yellowing or mottle (Fan and others 2016). Tamarixia radiata The population of the HLB vector D. citri can be controlled by the ectoparasitoid Tamarixia radiata Waterson, a species which feeds on D. citri. After the discovery of D. citri in 1998, Tamarixia radiata was experimentally introduced to Florida around the year 2000 in hopes it would prevent the spread of D. citri (Qureshi and others 2008). Females of T. radiata feed on young instars of D. citri and lay eggs on the older instars. In laboratory experiments, T. radiata killed > 90% of instars (Aubert 1992). Despite great hope for this method and reports of success with T. radiata elsewhere (Chien and Chu 1996, Etienne 2001), this biological control species had 10-21% kill rates in Florida groves during spring and summer, and up to 56% kill rates during fall (Qureshi and others 2008). At the time of this study, the authors concluded that these results did not show much promise for control of D. citri by T. radiata. However, a more recent study by the same group in South Florida showed field parasitism averaged 50% in May and 80% in November of 2011 (Quereshi and others 2012). These results raise the question why T. radiata has proven much more effective in places such as Taiwan (Chien and Chu 1996) and Guadeloupe (Etienne 2001) compared to Florida. It may be that T. radiata preyed upon coccinellid species native to 41

42 Florida (Michaud 2004) which would have neither been present in the field studies on other continents nor in the controlled laboratory experiment. It is also possible that T. radiata does not work as well in Florida because 80% of insecticides which can be used on citrus trees are considered harmful to T. radiata (Beloti and others 2015). Thus, most insecticide treatments cannot coincide with ectoparasitoid releases. Genetic modification of citrus trees A recent study (Dutt and others 2015) demonstrated that a gene from Arabidopsis, a flowering plant in the Brassicaceae family, can be used to enhance resistance of citrus trees to HLB. Transgenic sweet orange trees with overexpressed Arabidopsis non-expressor of pathogenesis related genes 1 (NPR1) were created in this study. Plants which over-express NPR1 had previously demonstrated enhanced resistance to several pathogens (Cao and others 1998). Dutt and others (2015) used two different methods to produce transgenic sweet orange plants: both phloem specific and constitutive promoters were investigated. These trees were exposed to CLas positive ACPs for two years in a greenhouse environment; trees were evaluated for HLB once every six months. Additionally, a second set of trees were planted in a field site where the HLB infection rate was > 90%; trees planted in this field were also evaluated for HLB every six months for two years. After two years, 45% of trees with the phloem specific promoter were negative for HLB while 27% of the trees with the constitutive promoter remained HLB negative. No abnormal phenotypic changes were observed; phenotypic abnormalities were previously reported as issues with Arabidopsis NPR1 when using this gene in strawberries (Silva and others 2015) and rice (Chern and others 2005). It was concluded that transgenic trees showed reduced severity of HLB and 42

43 noted that some lines were disease free 36 months after being planted in the high HLB infection rate field. Currently, transgenic trees are allowed for research purposes only. Even if a highly successful, HLB resistant orange tree were developed, it is possible that transgenic oranges could be rejected by the public due to low public acceptance of genetically modified foods (Frewer and others 2013). 43

44 CHAPTER 3 MATERIALS AND METHODS Orange Trees and Harvesting Florida grown Hamlin and Valencia oranges were harvested twice per year each for two years. Fruit were harvested from established groves at the University of Florida Citrus Research and Education Center in Lake Alfred, Florida. For both Hamlin and Valencia oranges, fruit were harvested early and late each year in the respective variety s season. The maturity and disease stage of the fruit were determined by a trained horticulturist using established protocols (Achor and others 2010). The time of maturity varied slightly year to year based on flowering time. Oranges were harvested for the and seasons. Hamlin Early oranges were harvested on December 2, 2014 and November 15, Hamlin Late oranges were harvested on January 20, 2015 and January 21, Valencia Early oranges were harvested February 3, 2015 and February 11, Valencia Late oranges were harvested March 24, 2015 and March 31, These harvests were earlier than historical early or late season harvests for both Hamlin and Valencia oranges (Kesterson and others 1971). However, orange maturity dictated the harvest times. Early season fruit was tested before harvest to ensure it met minimum quality standards including Brix/acid ratio. Late season fruit was harvested as late as possible without having an insufficient number of fruit to harvest due to fruit drop. Approximately 30 trees per variety per year were harvested. The trees harvested were in adjacent rows (n = 2-5). Approximately 700 pounds each of SY and AS oranges were picked each harvest. SY oranges were noted to be mature but with characteristic outward greening symptoms including small fruit and reversed ripening 44

45 with the stylar end remaining green. While some trees contained both SY and AS fruit, many trees contained nearly 100% SY fruit. These trees commonly displayed yellow blotchy mottle on the leaves and overall sparse foliage on the tree. AS fruit appeared outwardly healthy. AS fruit were overall larger and did not display the reversed ripening common with HLB SY fruit. Figure 3-1 shows HLB AS (left) and HLB SY (right) oranges for the Valencia Late harvest on February 11, There are no longer significant quantities of HLB unaffected fruit in the state of Florida, as HLB infection rates at the grove level is 100% (Browning 2015). Given these circumstances, only HLB AS and SY fruit were utilized for this study. Figure 3-1. HLB AS (left) and SY (right) Valencia oranges (Photos courtesy of the author) Cold Pressed Oil Production After harvest, oranges were stored in a cold room at 4 C until processing. Processing of the raw oranges into CPOO was performed in the Pilot Plant at the Citrus Research and Education Center in Lake Alfred, Florida according to typical industry practices (Braddock 1999). Prior to processing, the oranges were sanitized with Fruit Cleaner 395 (JBT FoodTech, Lakeland, FL) and surface dried. Juice and an oil emulsion were extracted simultaneously with a commercial JBT extractor (JBT FoodTech, Lakeland, FL). The oil emulsion was collected in 190 liter tanks and put into 45