Analysis of Selected Biochemical Constituents in Black Tea ( Camellia sinensis ) for Predicting the Quality of Tea Germplasm in Sri Lanka

Twenty two accessions from tea (Camellia sinensis) germplasm representing high, moderate and low quality of made tea were used for black tea production in wet and dry seasons. Total polyphenol content, total free amino acid content, total liquor colour and brightness were analysed by spectrophotometric methods and correlated with organoleptic parameters viz infused leaf colour, colour of liquor, strength and quality. Strong positive correlations were identified between infused leaf colour and total liquor colour with total quality score. It was also found that the polyphenol content had a positive correlation with brightness which was responsible for liquor quality. The resultant dendrogram of cluster analysis enabled separation of these accessions into high, moderate and low quality categories. Principal Component analysis (PCA) revealed that, of the 16 parameters studied, infused leaf colour has the highest contribution towards clustering of accessions. Group III represented 5 well established high quality accessions (DT1, DUN7, NAY3, TRI 777 and PK2) and group I represented moderate quality accessions including two accessions (N2 and TRI 4079) which were previously categorized in the high quality group. Therefore, polyphenol content, total liquor colour, brightness and infused leaf colour could be used as reliable quality parameters, whereas total amino acid and crude fibre contents cannot be considered as useful quality parameters. Furthermore, fermentation rate was determined for the 22 accessions and four groups were identified based on the fermentation rate. Results revealed that the fermentation rate could be used as an early selection criterion for determination of fermentation properties of tea accessions at progeny level.


INTRODUCTION
The tea (Camellia sinensis) produced in Sri Lanka is popular as "Ceylon tea" and has a higher demand as 'best quality tea' in the international trade.Sri Lankan tea industry continues to maintain its creditability in terms of quality and cleanliness of the product.The quality of black tea refers to all the characters such as colour, brightness, appearance, liquoring properties, strength and aroma by which it may be judged for its market value.In general, quality of black tea is a summation of all its desirable attributes.Tea tasters while assessing the made tea using their palate usually measure the essential characteristics such as brightness, strength, flavour, colour of the brew and appearance of black tea as well as infused leaf (Keegel, 1959;Sanderson, 1972;Hazarika et al., 2002).
The tea manufacturing process is a biochemical one which is controlled to a certain extent by mechanical manipulation of the harvested shoots (tea flush).The traditional process of manufacture of black tea consisted of several stages viz.(a) harvesting, (b) withering, (c) leaf maceration or rolling, (d) fermentation, (e) drying and (f) sorting and packaging (Roberts, 2008;Sanderson, 1972).Most important chemical changes which result in development of desirable liquor characteristics such as liquor quality, strength and colour occur during fermentation.Polyphenols, the main group of chemical substances present in the tea flush, undergo oxidative changes during this stage.
In some countries biochemical constituents in tea have been used as discriminative markers to evaluate diversity and genetic potential warehoused in the germplasm.A review by Owuor and Obanda (1998) has suggested the possibility of using morphological features and various chemical constituents in selecting quality accessions.Therefore, it is timely to characterize and evaluate the available germplasm collection in Sri Lanka using biochemical methods to further support germplasm characterization to identify desired parents and elite germplasm for the tea breeding programme.Nonetheless, a pre-mediated selection among the germplasm of high quality has been lacking.Selection of accessions that can make good quality tea needs extensive evaluation of biochemical constituents which contribute towards the liquor characteristics.The most important chemical compounds that are directly responsible for tea quality are flavanols and their oxidation products (Saravanan et al., 2005).However, many chemical constituents which are responsible for the quality have not been studied in detail using local germplasm accessions to identify desirable genotypes for breeding purposes or even for commercial production of tea in Sri Lanka.
Most of the previous studies carried out at the Tea Research Institute of Sri Lanka were focused only on individual biochemical properties in fresh leaf as well as in processed tea, such as total polyphenol content and total amino acids (Wickremasinghe et al., 1966), composition of major fatty acids (Liyanage et al., 1988), volatile compounds (Herath et al., 1991), individual catechins content (Herath et al., 1993), TF (theaflavins) and TR (thearubigins) composition and chlorophyll content (Wickremasinghe et al., 1966;De Silva & Sivapalan, 1982).If a precise correlation can be established between chemical parameters in the fresh leaf and the final quality of the product, this could enhance the process of identifying quality tea accessions at an early stage in the tea breeding program.This would, in turn, facilitate germplasm characterization using biochemical markers to group the accessions into unique clusters.
In the present study, fermentation rate and crude fibre content of 22 tea accessions grown in high elevation (1400 m above) in Sri Lanka were investigated.Total polyphenol and amino acid content in the black tea manufactured from the 22 tea accessions were also determined.In addition, correlation between spectrophotometric quality parameters, such as total colour and brightness of liquor, and the tea taster scores were also determined.

MATERIALS AND METHODS
Twenty two germplasm accessions including TRI series, estate selections and introductions from other countries were selected based on their previously collected information as indicated in Table 1

Fermentation rate
The fermentation rate of 22 tea accessions was studied by conducting the chloroform test described initially by Sanderson (1963) and modified by Samaraweera and Ranaweera (1988).The tests were carried out in a glass tank fitted with a lid and chloroform-soaked cotton wool kept at the bottom.Second leaf of the tea shoot was hung on a wire and placed horizontally, so that the leaves were nearly equidistant from the soaked wool.The fermentation was considered to be complete when the leaf turned to a brown colour after a period of exposure to the chloroform vapour.The experiment was repeated six times.

Determination of crude fibre content
Crude fibre content of tea flush was determined and expressed as a percentage by mass of the sample on a dry basis according to the ISO standard (ISO -15598).

Black tea manufacturing
Tea shoots consisting of two leaves and an active bud (500 g) were harvested from 20 bushes of each accession and black tea was produced in an environmental-controlled manufacturing system designed by Teacraft ® .Harvested shoots were withered in a withering cabinet.
Withered leaves, after reaching the optimum withering stage (14 to 18 hours), were rolled for 15 minutes in a miniature roller and fed into a miniature spiral rotorvane to achieve a fine dhool.The macerated dhool was sieved through a 2.5 mm mesh and placed in a cabinet for a period of optimum fermentation time (2 ½ to 3 hours) at 20 o C. The inlet temperature of the miniature fluid bed dryer was set at 125 o C and fermented dhool was fed to it.The drying step was terminated when the outlet temperature reached 98 o C. The black tea production was repeated twice in wet (September to December, 2009) and dry seasons (January to April, 2010).

Sensory evaluation
Made tea samples (20 g from each accession) were sent for tasters' evaluation.Sensory evaluation made by two professional tea tasters with scores based on infused leaf colour, colour of liquor, strength and liquor quality was used to determine the parameters related to quality of made tea.

Determination of total colour and brightness of tea liquor
Nine grams of made tea samples were used for hot water tea extract and total colour and brightness of tea liquor was determined by measuring the absorbance at 380 nm and 460 nm using UV-VIS spectrophotometer (Roberts & Smith, 1963).

Preparation of tea stock extraction for biochemical analysis
Made tea samples were used for extracting tea infusion and 0.20 g of made tea was extracted in 70 % methanol and used as a stock extraction for all analysis as indicated below.

Determination of total polyphenols in black tea
The total polyphenol content in the tea extract was determined by measuring the colour development with Folin-Ciocalteu phenol reagent in alkaline medium (ISO 14502-1), at absorbance of 765 nm using UV-VIS spectrophotometer (CARY 510 Bio).

Determination of amino acids in black tea
Total free amino acid content in the tea extract was determined by measuring the absorbance at 570 nm using UV-VIS spectrophotometer (CARY 510 Bio) by the reaction with ninhydrin.The results were expressed as a percentage by mass of the black tea sample on a dry basis (Yemm & Cocking, 1955).
All the above experiments were repeated twice using 22 black tea samples processed in both wet and dry seasons.

Statistical analysis
The differences between mean values of each parameter were analyzed using Duncan's Multiple Range Test and cluster diagram was constructed using average linkage method.A principal component analysis was performed to determine the most contributory factors associated with grouping of accessions in the dendrogram.The Statistical analysis system (SAS, 1999) was used in the analysis.

Rate of fermentation
According to the results obtained, (Table 2) the selected 22 accessions could be categorized into four groups based on the fermentation rate.TRI 777 was the most rapidly fermenting accession and TRI 2016 and TRI 3019 were the slowest-fermenting accessions.Samaraweera and Ranaweera (1988) used 8 accessions for estimating fermenting period and observed that the fermentation rate of TRI 777 and DT 1 was significantly higher than TRI 2025.A similar pattern was observed in the present study too, which grouped DT 1, NAY 3, TRI 2043 and CY 9 in the fast fermenting category.In contrast, high yielding accessions such as TRI 2025, TRI 3013, TRI 4052, TRI 4071 and TRI 4078 were found to be moderate fermenting accessions.The period of fermentation widely practiced at present varies from 1 to 3 hours.During the fermentation period, polyphenols in tea oxidize into TF and TR.In general, fermentation is normally done for a shorter period to maximize the amount of TF.Prolonged oxidation results in deterioration of black tea quality.It was observed that TF reached a maximum at optimum oxidation time and showed a decline in quality if the processed dhool was overoxidized during the longer period of fermentation (Cloughley, 1980;Lopez et al., 2005).Therefore, optimum time of fermentation is a critical determinant of quality of black tea.
The results of the present study revealed that the rate of fermentation depends on the accession and facilitates to predict the optimum time for fermentation during black tea processing.These findings may be taken into consideration in planning, harvesting and processing of tea on a commercial scale to achieve best results.In general, 4-5 high yielding TRI accessions are frequently used for replanting or infilling purpose without considering the biochemical properties of the accessions.Therefore, attention should be paid to plant those accessions in separate blocks to avoid mixing of accessions between categories identified in this study.
Furthermore, fermentation rate could be used as a rapid means of evaluating the fermenting properties during the seedling tea selection programme and progeny trials at a very early stage of the breeding programme.This method was also used in early selection of potential quality accessions in tea breeding programs in Kenya (Seurei et al., 1998) and Pakistan (Waheed et al., 2001).On the other hand, green tea is a non-fermented form of tea and basically different from black tea.The present study indicated that those accessions with slow fermentation rate may be suitable for green tea production and hence, the suitability of accessions for green tea could be assessed in the selection programme of seedlings to generate potential tea accessions.

Polyphenol content of black tea
The present study revealed that the total polyphenol content largely varied and was in the range of 14.42 to 22.20 % in the wet season and 14.24 to 24.79 % in the dry season (Table 3).tasters' scores had increased (Table 3).In contrast, the total polyphenol content of lowquality accessions, except TRI 4052, showed a very little increment during the dry period.
During the processing of black tea, about 90-95 % of the polyphenols undergo enzymatic oxidation and the majority of the polyphenols would convert into theaflavins and thearubigins, which contribute to the quality of black tea brew (Roberts andSmith, 1963 &Sanderson, 1972).The level of polyphenols is therefore highly attributable to the final quality of black teas.

Free amino acids and crude fiber content
Total free amino acid content of black tea varied from 0.72 to 2.67 % in the wet season and 0.86 -2.00 % in the dry season.Range of the crude fiber content of the accessions was from 8.94 to 11.55 %.According to the results obtained in this study, a clear trend could not be discerned among the accessions in relation to quality and the crude fiber content (data not presented).As such, the total amino acid and crude fiber contents are not valid parameters for predicting the made-tea quality.Similar observations were noted by Sanderson and Kanapathipillai (1964), who stated that the total amino acid and crude fiber contents cannot be considered as useful quality parameters.Although those two parameters cannot be used to determine quality characteristics of an accession, the same could be used as biochemical descriptors in characterization of germplasm as there is a significant variation present among the accessions studied in relation to both parameters (crude fiber and amino acid content).Amino acids play an important role as precursors of some volatile flavor compounds in development of tea aroma (Sanderson, 1972).Therefore, further investigation on individual amino acid content and flavor profile of accessions would be more meaningful than total amino acid content.In addition, coarse plucking (mature leaf harvest) will have a tendency to produce off-grade teas with high values of crude fiber content.

Total color and brightness of tea liquor
Total liquor color and brightness of known quality accessions such as DT 1, N2, TRI 777 and TRI 4079 was higher than those of moderate quality accessions such as TRI 62/5, TRI 3019 and TRI 2025 in dry season.Present results also revealed that the accessions having higher values for brightness were considered as good quality accessions by tasters.Furthermore, TF contribute to brightness and long fermentation times produced more color in black tea at the expense of brightness (Owuor et al., 2008).It indicates that the good quality teas should have high values of brightness.All high quality accessions except TRI 4067 followed the above pattern.On the other hand, brightness and overall tasters' score of NAY 3 showed higher values indicating that this accession may possibly be considered as a high quality accession.Therefore, it is necessary to study this in-depth for a proper reevaluation of quality ranking of TRI 4067 and NAY3.
In addition to the biochemical parameters discussed above, liquor color and brightness were also considered as quality parameters by early scientists (Keegel, 1959;Kirthisinghe et al., 1968) for evaluating made tea quality.Therefore, liquor color and brightness can also be considered as good quality parameters for determining the final quality of black teas.
Principal component analyses (PCA) on the biochemical and tasters' evaluation parameters are presented in Table 4.It shows that the first five principal components (PCs) account for 82 % of total variability.According to the PCs loadings for characters, infused leaf color and total quality scores in dry season (0.35) show the highest value indicating that they are the main contributing characters to the clustering order in the dendrogram.This suggests that the infused leaf color is highly correlated with the quality parameter of made tea.On the other hand, the second PC was correlated with brightness of liquor in wet and dry seasons (0.37).Parameters with high loadings for the third PC were total polyphenols in dry season (0.39) and tasters values for color of liquor in wet (0.39) and dry (0.37) seasons.In addition, total liquor colour in dry season (0.71) shows the highest loading value in the fifth PC.
According to the results of the average linkage cluster analysis, accessions studied were clustered into three main groups (Fig. 1).The group III comprised of DT 1, DUN 7, NAY 3, TRI 777 and PK 2. All those accessions were considered as high quality accessions and ranked similarly by earlier attempts (Kirthisinghe et al., 1968;Anon, 2002).Most of the low quality accessions such as TRI 2016, TRI 2043, TRI 4052 and TRI 4071 were included in group II.Two moderate quality accessions (TRI 3019 and TRI 62/5) (Anon, 2002) also clustered in group II.Interestingly, the accession TRI 4067 has been clustered with the low quality group in the present study.In the current list of recommended tea accessions (Anon, 2002), TRI 4067 was rated as a high quality accession.Besides, most of moderate quality accessions such as DN, KEN 16/3, DT 95, and TRI 4078 (Kirthisinghe et al., 1968;Anon, 2002) were clustered in group I with two high quality accessions such as N2 and TRI 4079.Quality of tea accessions were frequently assessed by professional tea taster's by sensory evaluation and present results revealed that it is worthwhile to incorporate biochemical parameters to derive more meaningful conclusions about black tea quality.

Group I Group II Group III
The correlation coefficient of different biochemical parameters and tasters' scores are given in Table 5.There was a strong positive correlation (0.94) observed between infused leaf colour and total quality score in dry season.Therefore, infused leaf colour could be considered as a good made tea quality parameter.In addition, a positive correlation (0.79) could be observed between total liquor colour and total quality score in wet season.There was a significant negative correlation (-0.61) between tasters' values for colour of liquor and liquor strength in dry season.On the other hand, liquor quality was positively correlated with total quality scores (0.90, 0.88) in wet and dry seasons respectively.
Overall quality of black tea should be evaluated by professional tea tasters and the prices will depend on the tasters' score.High market value is guaranteed with the opinion of tea tasters, whereas low tasters' scores indicate the poor quality of made tea.According to the tasters' quality score, DT 1, DUN 7, NAY 3, PK 2, and TRI 777 can be categorized as good quality accessions whereas CY 9, KEN 16/3, TRI 2016, TRI 2043, TRI 3019 and TRI 62/5 as low quality accessions.Those categorizations are in agreement with the results of previous studies to a greater extent (Kirthisinghe et al., 1968;Anon, 2002), although certain accessions were not qualified to be in the same quality category.
In addition, several attempts have been made by various researchers to correlate biochemical compounds such as total polyphenols, amino acids etc., in assessing the quality of made tea, although the correlation between those compounds and the quality was found to be inconsistent in many cases (Wickremasighe et al., 1966;Roberts & Fernando, 1981;Owuor and Obanda, 1998) to be able to use them as chemical markers in screening for made tea quality.In contrast, theaflavins and thearubigins contents, total colour and brightness (Owuor et al., 2008), theaflavins and their fractions (Wright et al., 2002), catechin fractions and ratio between polyphenol oxidase and catechins (Lopez et al., 2005;Saravanan et al., 2005) have been considered as reliable black tea quality parameters.
The results obtained in the present investigation on total polyphenol content, liquor colour and brightness can be considered as good quality parameters for determining the final quality of black tea.Further investigations of other biochemical parameters are vital to establish reliable parameters to screen large number of progenies and germplasm accessions.

CONCLUSIONS
The present study reveals that the fermentation rate could be used as a selection criterion to determine fermentation properties and identify potential quality accessions during the early stages of the tea breeding program.Polyphenol content in black tea, liquor color and brightness could be considered as reliable quality parameters whereas total amino acid and crude fiber content cannot be considered as useful criteria to determine quality of made tea.Furthermore, it is suggested to extend the study including other tea growing areas in different seasons with more accessions.
thank two professional tea tasters (Mr. Lalith Ramanayake and Mr. Anil Cooke) for sensory evaluating of black tea.

Fig. 1 .
Fig. 1.Dendrogram of average linkage cluster analysis based on biochemical parameters

Table 1 . Origin and descriptive characters of selected tea accessions Accession Origin* Yield* Quality †, * Blister blight* Shot hole borer*
. Plants of those accessions which were conserved in the ex situ field gene bank at the Tea Research Institute of Sri Lanka, Talawakelle (latitude 6 o 54'N, longitude 80 o 42'E) were used to obtain the material.Average annual rainfall of the area is about 2500 mm with an annual average minimum, maximum, temperatures are 14.2 o C and 22.8 o C, respectively.Average elevation is 1394 m above sea level.

Table 3 . Mean ± SD values of the biochemical parameters in different tea accessions
accessions had increased during the dry season.The other 4 moderate quality accessions showed a similar pattern except NAY 3. It showed a slight decline of the total polyphenol content during the dry season, although other biochemical parameters such as brightness and 35 High quality †,* Low quality †,*