Blueberry estimated harvest from seven new cultivars: Fruit and anthocyanins

This study compares the yields, weights and anthocyanin contents of fruit from a group
of seven new cultivars released from the New Zealand blueberry breeding programme
and selected for the longest possible combined harvest season. The measured factors
were primarily influenced by cultivar, and seasonal variations had relatively minor
effects. The late-ripening cultivars ‘Velluto Blue’ and ‘Centra Blue’ had the highest fruit
yields, anthocyanin contents and estimated total anthocyanin harvestable from a given
area. ‘Blue Moon’ and ‘Sky Blue’ had the largest fruit sizes. The early-ripening cultivars
‘Blue Bayou’, ‘Blue Moon’ and ‘Sunset Blue’ had the lowest anthocyanin contents. The
yield, fruit size and total anthocyanin content results obtained from any single year were
highly correlated with the average of the three years, which makes pursuing the
evaluation for these traits from a single year and at an early stage of plant development
a practical proposition.

1. Introduction
Cultivated blueberries (Vaccinium corymbosum, V. corymbosum hybrids and V.
virgatum) are commercially produced in New Zealand and in the last few years the total
area planted with blueberries increased from 239 ha in 2000 to 522 ha in 2009 (Hort
Research, 2000; Plant & Food Research 2009). In New Zealand, V. corymbosum is
commercially known as the ‘‘northern highbush’’ blueberry, V. corymbosum hybrids are
known as ‘‘southern highbush’’ blueberry and V. virgatum is known as ‘‘rabbiteye’’
blueberry. Within the Vaccinium species internationally, there is a large range
of cultivars, but relatively few are available in New Zealand. The cultivars that are
readily available from nurseries may come from overseas breeding programmes, where
they have been selected for environments that may be considerably different from that
in New Zealand.

Plant & Food Research (PFR) started a blueberry breeding programme in the 1980s
with the aim of producing blueberry cultivars that were tailored to New Zealand
conditions. Breeding objectives were specific agronomic and fruit traits such as high
yield, disease resistance, winter chilling adaptability, seasonality range, hand-
harvestable fruit of large size and light blue coloured fruit. A blueberry plantation of well-
adapted cultivars has the potential to produce a good crop for many years and thus
planting the right genotype is crucial. Growing two or more different cultivars in the
same commercial block is a popular choice among New Zealand growers, to satisfy
pollination requirements and to extend the production window. There are reported
differences in plant yields among blueberry cultivars, in seasonality changes of the yield
and in plant management techniques affecting yield and fruit size (Celik, 2009; Finn,
Strik, & Wagner, 2009; Gaskell, 2009; Lyrene & Williamson, 1997).

Blueberry fruit available from the market come from cultivars that have been selected
for specific agronomic and fruit traits such as high yield and large fruit size and not
necessarily for high concentrations of health-promoting phytochemicals in the fruit. For
this reason, it was important to measure the phytochemical content and composition of
the fruit from the new cultivars.
Blueberry fruit are believed to be good for health because of their anthocyanins and
other polyphenolic compounds. Anthocyanins provide blueberries with their
characteristic colour and have been shown to contribute to the antioxidant capacity of
berryfruit (Connor, Finn, McGhie, & Alspach, 2005; Connor, Luby, Tong, Finn, &
Hancock, 2002; Currie et al., 2006; Deighton, Brennan, Finn, & Davies, 2000; Ehlenfeldt
& Prior, 2001; Kalt, McDonald, Ricker, & Lu, 1999; Kalt et al., 2001; Moyer, Hummer,
Finn, Frei, & Wrolstad, 2002; Prior et al., 1998; Proteggente et al., 2002; Wang, Cao, &
Prior, 1997). The health value of anthocyanins has been reviewed (Beattie, Crozier, &
Duthie, 2005; Kong, Chia, Goh, Chia, & Brouillard, 2003) and as well as antioxidant
capacity (Stintzing, Stintzing, Carle, Frei, & Wrolstad, 2002; Wang et al., 1997),
anthocyanins are reported to have a role in improving circulation (Matsumoto et al.,
2005), preventing stroke (Keli, Hertog, Feskens, & Kromhout, 1996), providing benefits
to vision (Lee et al., 2005; Matsumoto, Nakamura, Tachibanaki, Kawamura, &
Hirayama, 2003; Matsumoto et al., 2005), and their anti-inflammatory and anti-oxidative
effects are extensively reported (Ghosh, McGhie, Zhang, Adaim, & Skinner, 2006; Lyall
et al., 2009; Wang et al., 1999).

The anthocyanins present in blueberry are galactosides, glucosides and arabinosides of

the anthocyanidins delphinidin, cyanidin, etunidin, peonidin and malvidin, and their
concentrations vary greatly between genotypes (Scalzo, Miller, Edwards, Meekings, &
Alspach, 2009; Wang, Chen, Camp, & Ehlenfeldt, 2012a). Additionally, these glycosides
may also be acylated (Wu & Prior, 2005). Blueberries are the richest sources of the
more hydrophobic malvidins and petunidins among a wide selection of fruits and
vegetables (Wu et al., 2006). Blueberries are also rich in delphinidin- 3-galactoside and
petunidin-3-glucoside (Kader, Rovel, Girardin, & Metche, 1996). Research shows that
there is interest in the anthocyanin content of blueberry fruit and that changes are to be
expected between fruit of different cultivars and between seasons (Connor, Luby, Tong,
Finn, & Hancock, 2002; Scalzo et al., 2009; Wang et al., 2012a, 2012b). There is,
however, little knowledge of how much anthocyanin can be harvested from a blueberry
plantation, i.e., which cultivars have the best combination of high anthocyanin content
and high yield.

In this study, a combination of seven new cultivars released by he PFR blueberry

breeding programme in New Zealand was considered. Altogether, the new seven
cultivars in combination offer an extended production window, with an early harvest in
mid November (late spring in New Zealand) and the last harvest around the beginning
of April (mid autumn). We investigated and compared plant yields, fruit weights, total
anthocyanin contents and polyphenol composition of the fruit for all the seven cultivars
and over three years. Sources of the observed variations are discussed.

2. Materials and methods

2.1. Chemicals
Solvents and general chemicals were obtained from local suppliers. Reference standards were
obtained as follows: cyanidin- 3-glucoside from Extrasynthese (Genay, France), chlorogenic
acid and rutin from Sigma–Aldrich.
2.2. Plants and fruit sampling

The blueberry plants under evaluation were planted in the ground in winter 2007 (June–
July) at the Ruakura Research Centre – New Zealand (37_480S 175_170E) and grown
in soil modified with additional organic material at pH _4.3. Details of the species and
type of the seven cultivars and harvest dates are shown in Table

1. The site has a mean annual rainfall of about 1200 mm, with moderate temperatures
(min./max. 0 C/29 C). Berries
were harvested from plants growing in a replicated plot trial of 20 plants (four plots
five plants of each cultivar) and for three consecutive years
(summers 2009–2010, 2010–2011 and 2011–2012). In this work, summer seasons are
indicated as follows: 2009– 2010 = 2010, 2010–2011 = 2011 and 2011–2012 = 2012.
For the yield assessments, fully ripe fruit were collected from each plant, starting when
about 50% of the fruit per plant were estimated to be blue. Multiple harvests were
required to collect the total yield per plant. Total yield recording started in year 3 after
planting. Little crop was produced in years 1 and 2 and this was removed
from the plants to promote growth. For the fruit weight determination, 50 uniform fully
ripe fruit per plot were selected from the harvest at the stage when 50% of fruit were
blue, and the mean fruit weight recorded within 2 h of harvesting. For the total
anthocyanin content and the polyphenol assessments, a subsample of 100 g of uniform
fruit was selected from the harvest at the stage when 50% of fruit were blue and stored
at

20 C until analysed.

2.3.2. HPLC analysis

Analyses were carried out on a Shimadzu 20-series analytical high performance liquid
chromatograph (HPLC) with a column oven, auto-sampler, vacuum solvent degas
module and diode-array detector. The column used was a 150
2 mm, Synergi Polar-RP, 4l particle size, 80 Å pore size,
fitted with a Security-Guard 3 2 mm Polar RP guard
cartridge (Phenomenex, Auckland, New Zealand). Flow rate was 0.6 ml/min and column
temperature 50 C. Solvents were (A) methanol and (B) 2%
aqueous formic acid, and the initial mobile phase was 12% A and 88% B. The time
programme of pump B concentration was set up as 86% at 2.5 min, 78% at 5 min, 62%
at 8.5 min, 37% at 10.50 min, 30% from 10.5 to 12.5 min, returning to 88% at 12.75 min
and staying at that concentration until the end of the run at 15 min. Sample injection
volume was 10 ll. UV/visible spectra were recorded from 250 to 600 nm in 1.2-nm steps.

Quantification of anthocyanins was carried out at 520 nm, incomparison with standard
solutions of cyanidin glucoside. Chlorogenic acid and quercetin glycosides were
quantified at 350 nm compared with standards of chlorogenic acid and rutin,
respectively. Results were expressed as mg cyanidin glucoside equivalent, chlorogenic
acid, or rutin equivalents/100 g fruit. Concentrations in the HPLC samples were
converted to fruit equivalents using dilution factors, fruit weight and extract volume.

2.3.3. Confirmation of identity of compounds

To assist identification of compounds observed during HPLC runs, some samples were
rerun on the same HPLC, with the addition of a Shimadzu LCMS 2020, single
quadrupole mass spectrometer, fitted with an electrospray (ESI) interface. Instrument
parameters used were the manufacturer’s default settings. MS scans were carried out in
both positive and negative mode in the same run, using a mass range of 140–650 Da.
Confirmation of identity was achieved both though comparison of expected and
observed molecular ion masses and ‘‘neutral losses’’ of sugars during ionisation to
leave aglycones or partially glycosylated ions. For example, cyanidin glucoside
exhibited a positive molecular ion at 449 Da and an ion at 287 Da, corresponding to loss
of the glucose residue (162 Da) to leave the cyanidin fragment ion. Compounds were
identified by comparison with previously reported data (Borges, Degeneve, Mullen, &
Crozier, 2010; Lohachoompol, Mulholland, Srzednicki, & Craske, 2008).

2.4. Statistical design

The field trial was established as a randomised complete blockdesign with four
replications of five plants each. Data were analysed using linear mixed effects models,
testing effects of year (i.e., harvest season), cultivar and their interaction (fixed effects)
against the variability between blocks, plot within block, and block by year and plot by
year interactions (random effects). Data were log transformed to stabilise the variance;
the means presented have been back-transformed. The least significant differences for
the log-scale data have also been back-transformed to give least significant ratios
(LSR); two means are significantly different if the larger is more than LSR times the
smaller. The analysis was done with GenStat (version 14, 2011, VSNi Ltd, Hemel
Hempstead, UK).

3. Results and discussion

3.1. Fruit yield, fruit weight and total anthocyanin content (ACY)

Little crop was produced in years 1 and 2 and this was removed from the plants early in
the season to improve their growth and establishment of root systems. The yield in the
first year (2010) was lower than that in years 4 and 5, for most of the cultivars; however,
it did appear to give an indication of future yield potential.

3.1.1. The fruit yield

The fruit yield showed significant variation with year, cultivar and the interaction of these
two factors (Table 2A). The F ratios for year and cultivar, however, are both much
higher than that for the interaction year X cultivar, suggesting that the cultivars reacted
similarly to seasonal climate variations. The F ratios also suggest that seasonal
variations were much greater than variations between cultivars, although the seasonal
variations were confounded by the marked increase in fruit yield per plant over the three
years in all cultivars except ‘Dolce Blue’ and ‘Sky Blue’. A yield increase
with year was to be expected, since the plants were still growing at a strongly rate.
Significant differences in fruit yield were found between the three years when cultivars
were individually compared, and among cultivars within each year. The cultivars with
the highest mean yield over the three years were ‘Velluto Blue’ and ‘Centra Blue’ (Table
3), but the former’s fruit yield increased only slightly (by 7%) as the plants matured. In
contrast, ‘Centra Blue’ had the second highest mean yield, but increased by 77% as the
plant matured, suggesting that this cultivar may produce the highest yields over the
longer term. The cultivars with the lowest mean yields were ‘Blue Moon’ and ‘Blue
Bayou’. ‘Blue Moon’ was consistently the cultivar with the lowest yield within each year.

3.1.2. The individual fruit weight

The individual fruit weight varied significantly with cultivar and the year
cultivar interaction, with cultivar being the major influence
(Table 2A). This suggests that seasonal yield variations primarily result from changes in
the number of fruit, rather than
their size. ‘Blue Moon’ and ‘Sky Blue’ had the highest mean fruit weights and no
significant variation was found between years (Table 3). The cultivar with the lowest
average fruit weight was ‘Blue Bayou’, and its fruit size was consistently small across
the years.

3.1.3. The total anthocyanin content

The total anthocyanin content varied significantly with cultivarand the year x cultivar
interaction (Table 2A). The F ratio for year x cultivar was relatively lower than that for
cultivar, meaning that although there were some variations in the pattern of anthocyanin
content variation from year to year, they were smaller than the general trends for
cultivars. The cultivar with the highest mean anthocyanin content over the three years of
evaluation was ‘Dolce Blue’ (303 mg/g) (Table 3). ‘Dolce Blue’ fruit had the highest
anthocyanin content of these cultivars in 2010 and 2012, but not in 2011. ‘Sunset Blue’
fruit had the lowest mean anthocyanin content (101 mg/g) and were the lowest in each
year. The cultivars appear to have responded differently to weather variations within
each season. The fruit from the first three cultivars listed in Table 3 were harvested in
December and had markedly lower anthocyanin contents than the remaining four
cultivars, which were harvested in January–March. Examination of weather records for
the three growing seasons (not shown) detected no significant differences in either
temperature or solar radiation intensity between December and January–March. There
was, however, a markedly lower temperature and solar intensity all through the 2012
season, which had no detectable effect on the anthocyanin contents of the fruit
for the seven cultivars considered in this study (Table 3). If anything, this parameter was
slightly higher in 2012. It therefore appears that the observed differences are related to
fundamental differences between the early ripening V. corymbosum cultivars and the
later ripening V. virgatum cultivars.

Blueberry cultivars with high yield are desirable and always popular with growers,
whereas cultivars with large fruit are particularly advantageous to the majority of New
Zealand growers who hand-harvest their crop. Blueberry cultivars with high anthocyanin
contents can have an advantage in marketing to consumers because anthocyanins
have been reported to have health benefits, as previously discussed. This study agrees
with previous reports (Clark, Howard, & Talcott, 2002; Connor, Luby, & Tong, 2002;
Connor, Luby, Tong, Finn, et al., 2002; Pranprawit, Molan, Heyes, & Kruger, 2009;
Scalzo et al., 2009; Wang et al., 2012a, 2012b; You et al., 2011) in finding considerable
variation in fruit anthocyanin contents between cultivars (Table 3). The fruit yield also
varies, independently of anthocyanin content, between cultivars and between years,
and it generally increases with plant maturity (Finn et al., 2009). There is little
knowledge of whether the anthocyanin content in blueberry increases with the plant
age, similarly to the plant yield; however, there is reported evidence that it varies
between years (Connor, Luby, Tong, Finn, et al., 2002; Wang et al., 2012b). Our results
suggest that, in the New Zealand environment, total fruit anthocyanin content does not
significantly increase as blueberry plants mature.

3.2. Polyphenol composition

Compositional analysis of the polyphenols revealed that malvidin (as the sum of the
three glycosidic forms) was consistently the major anthocyanin component, comprising
30–47% of the mean total anthocyanin content (Table 4). The most variable
anthocyanin components were the acylated anthocyanins (AA) (Cho, Howard, Prior, &
Clark, 2004, 2005), present in low percentages in rabbiteye blueberry fruit, and the
anthocyanins chlorogenic acid (CA) and quercetin (Q) (in the form of 2–3 different
glycosyl derivatives) also exhibited significant variability (Table 2B).

Total malvidin (TM) showed significant variation with year, cultivar and their interaction
(Table 2B). The F ratio for cultivar was much higher than that for the other main factors
and the differences in cultivar content are shown in Table 5. Similarly to TM, AA showed
significant variation for all the main factors, with cultivar being the predominant factor
(Table 2B). Significant differences in AA were found between the three years when
cultivars were individually compared, and among cultivars within each year (Table 5),
with ‘Blue Moon’ being the cultivar with the highest AA content in each of the three
years.

CA was more abundant in V. virgatum cultivars than in V. corymbosum. AA and Q were

predominant in early-ripening cultivars such as ‘Blue Bayou’, ‘Sunset Blue’ and ‘Blue
Moon’, rather than in the late-ripening rabbiteye cultivars. CA showed significant
variation with year, cultivar and their interaction (Table 2B). The F ratio for year,
however, was much higher than those for cultivar and the interaction of year
 cultivar. Large and significant differences in CA were found
between the three years when cultivars were individually compared, and among
cultivars within each year (Table 5). Similarly to CA, Q showed significant variation for
all the main factors, and the F ratio for year was much higher than those for cultivar and
the interaction of year cultivar (Table 2B). In particular in
year 2010, the Q content was the lowest in each cultivar,
compared with values in 2011 and 2012. In summary, CA, AA and Q showed much
larger seasonal variations than the variation in total anthocyanin content, but these are
relatively minor components, so the importance of these variations is much less than
that of the anthocyanins.

3.3. Estimated anthocyanin harvest per hectare and seasonality changes

If blueberries were to be grown for processing, particularly for manufacture of high-

anthocyanin polyphenol extract powders, then the most important factor would be the
overall yield of anthocyanins per hectare. This factor might be optimised by high
anthocyanin content berries, high fruit yield, and preferably both. To estimate the
potential anthocyanin yield for each cultivar and each year, the average anthocyanin
content per plant was multiplied by the plant density (Table 6). By comparing the mean
values for the three years, ‘Velluto Blue’ was the cultivar with the highest estimated
anthocyanin yield, at an estimated plant density of 4125 plants per hectare. Under these
conditions, we estimated that the potential total anthocyanin harvest from ‘Velluto Blue’
fruit was 1.1-fold higher than that of ‘Centra Blue’, 1.9-fold higher than that of ‘Dolce
Blue’, 3.1-fold higher than that of ‘Sky Blue’, 6.1-fold higher than those of ‘Sunset Blue’
and ‘Blue Bayou’, and 9.2-fold higher than that of ‘Blue Moon’. By comparing the
estimated total harvest of anthocyanin in the last year of evaluation (2012), the mature
plants of ‘Centra Blue’ were the ones producing the highest potential harvest, because
of that cultivar’s high fruit yield in 2012.

Modifying plant yield and fruit size by changing the plant densityhas been reported
(Gaskell, 2009; Lyrene & Williamson, 1997; Strik & Buller, 2002): however, we have
found no report relating the anthocyanin content of fruit from different cultivars to
changes in plant density in the field. According to our results, late-ripening cultivars
such as rabbiteye blueberries have higher anthocyanin contents than early ripening
highbush blueberries (‘Sunset Blue’, ‘Blue Bayou’ and ‘Blue Moon’). As discussed
above, these differences appear to relate to the different species involved, rather than to
weather variation through the season. One might expect smaller fruit to contain more
anthocyanins because these compounds are found only in the skin and small fruit have
a relatively larger surface area in relation to volume. Inspection of Table 3, however,
shows this not to be the case: when comparing ‘Blue Bayou’, the smallest fruit, and
‘Blue Moon’, the biggest, their anthocyanin contents are very similar.

3.4. Correlation between traits and between years

Evaluating a multitude of cultivars in test trials is common practice in breeding

programmes and it is associated with considerable cost related to the cultivation and
maintenance of large numbers of bushes and data collection from the trial. Amongst all
the traits included in our work, determining the yield during the three years for all the
cultivars was the most expensive part of the research. According to our results, the yield
from a single year collected from all the cultivars was highly correlated to the average
yield across the three years (Table 7). Therefore, assessing this trait from a single year
of data should be mostly reliable for individual cultivars. Our 21 yield measurements
show only two outliers, ‘Sky Blue’, the yield of which decreased markedly between 2011
and 2012, and ‘Dolce Blue’ the yield of which in year 2012 was slightly lower than that
of year 2011. In addition, the relative fruit yields of the cultivars under evaluation appear
to be reasonably predictable from the third year of cultivation (i.e., the first
year of meaningful yield determination). Similar results were obtained previously: (Finn
et al., 2009), who monitored the harvest of different cultivars over a 3–9 year time
frame, found that the harvest in years 4–6 was strongly correlated to the cumulative
yield over all years and could predict genotypic differences.

Similarly to the yield, the single-year results obtained for fruit weights and total
anthocyanin contents were highly correlated to the averages across the three years
(Table 7), which makes evaluation of these traits from a single year and at an early
stage of plant development reasonably achievable. It therefore appears that it would be
practical to select promising crosses from a breeding programme in the third year of
plant growth, with a low probability of missing high-performing plants that happened to
perform unusually poorly in year 3. As an example, ‘Centra Blue’ had by far the highest
yield of the cultivars in year 5 (2012), and this was a considerable increase over its year
3 yield. Its year 3 yield (2010), however, was still the second highest, so it would still
have been selected as a high yielding cultivar, based solely on year 3 results. Only the
promising selections from year 3 would need to be further monitored, thereby freeing up
most of the plots in which new crosses could then be planted.

The correlation of the individual polyphenol compounds with the total anthocyanin
contents in each year and over the three-year evaluation showed some significant
differences (Table 8). Only TM showed consistently positive and significant correlations
with ACY, each year and over the three-year period. CA had high positive correlations
with ACY, but not in 2011. Overall, TM and CA were highly and positively correlated to
the total anthocyanin content over the three years of evaluation. The correlations
between AA and ACY, and Q and ACY, were negative over the entire evaluation, but
not significantly so.

4. Conclusions

Year effect was the main factor that influenced fruit yield and polyphenols such as CA
and Q. As the plants matured, the average total yield per plant increased. The average
CA was significantly reduced in 2012, while the Q content was the lowest for all the
cultivars in 2010.
The FW and ACY did not significantly change between the years of evaluation.
Although there were some variations of fruit weight and in the pattern of anthocyanin
content from year to year, they were much smaller than the general trends for cultivars.
The yield, fruit size and total anthocyanin contents obtained from any single year were
highly correlated to the average across the three years, which makes pursuing the
evaluation for these traits from a single year and at an early stage of plant development
a practical proposition. This finding could permit large savings in time and resources
during selection of new blueberry cultivars, by cutting the time required from 5 to 3
years.
Traditionally, blueberries have been bred primarily to optimize the production and
consumer attributes of taste and appearance for fresh eating fruit. Little or no attention
has been paid to optimizing polyphenol content for perceived heath attributes, or for
making polyphenol extracts for use in formulated functional foods. We have introduced
a different approach by additionally selecting for a high polyphenol content and
particularly for anthocyanins that have potential health benefits. If blueberries were to be
grown for processing, particularly for manufacture of high-anthocyanin polyphenol
extract powders, then the most important factor would be the overall yield of
anthocyanins per hectare. Agronomic traits would still be important, but consumer
attributes much less so. Our results show that for a given plant density, the total
anthocyanin content harvestable from a blueberry cultivation varies greatly among
cultivars. We estimated that the potential total anthocyanin harvest from rabbiteye
cultivars was higher than those for southern highbush and northern highbush. For
processing purposes, ‘Velluto Blue’ and ‘Centra Blue’ would be clearly much better than
the other cultivars studied, since each has botha high fruit yield and high anthocyanin
content.

Acknowledgements

The authors would like to acknowledge Dawei Deng for the polyphenol analysis, Judith
Rees, Carolyn Edwards and Shirley Miller for assisting with the fruit harvest.