FoodStruc

ture14(20
17)8–16
Contents lists available at ScienceDirect

Food Structure
journal homepage: www.elsevier.com/locate/foostr

Effect of elevated temperature on the microstructure of full fat Cheddar cheese during ripening

Kevany Soodama,b,1, Lydia Onga,b,d, Ian B. Powellc,d, Sandra E. Kentisha,d, Sally L. Grasa,b,d,⁎
a
DEPARTMENT of CHEMICAL AND BIOMOLECULAR Engineering, The University of Melbourne, PARKVILLE, VICTORIA
3010, AUSTRALIA
b
The Bio21 MOLECULAR Science AND Biotechnology Institute, The University of Melbourne, PARKVILLE, VICTORIA
3010, AUSTRALIA
c
DAIRY INNOVATION AUSTRALIA Limited,180 Princes HIGHWAY, Werribee, VICTORIA 3030, AUSTRALIA
d
The ARC DAIRY INNOVATION Hub, DEPARTMENT of CHEMICAL AND BIOMOLECULAR Engineering, The University of
Melbourne, PARKVILLE, VICTORIA 3010, AUSTRALIA

ARTICLE INFO ABSTRACT

Keywords:
Cheddar Elevated temperatures have been widely studied as a route to accelerate cheese
cheese ripening and decrease energy and storage requirements but the impact of
Ripening temperature on the underlying microstructure of the cheese during prolonged
temperature periods of ripening is poorly understood. In this study, Cheddar cheese was matured
Cheese at four different ripening temperatures (8 °C, 15 °C, 20 °C or a combination of 8
microstructure °C and 15 °C) and the impact on cheese mi- crostructure assessed using
cryo-SEM confocal laser scanning microscopy, cryo scanning electron microscopy and quan-
titative image analysis of 3D images. An increase in ripening temperature was
Confocal laser scanning
shown to alter the microstructure of the cheese protein network after only a few
microscopy Biogenic
weeks of ripening. Incubation at 20 °C significantly reduced branching within the
amines
protein network, leading to thicker protein strands and larger pores after 33 days.
These structural changes coincided with increased proteolysis, consistent with
solubilisation of the protein network; they also led to a softer, less chewy and less
cohesive cheese. While the concentration of biogenic amines tryptamine and
tyramine were observed to increase with ripening temperature, the concentrations
were gen- erally low, confirming that biogenic amines do not represent a health
concern under the conditions examined. This study illustrates how 3D image
analysis can be used to observe and quantify the effect of process changes on
cheese structure, assisting our understanding of the link between structure and
function in Cheddar cheese.

1. Introduction
Ripening is a slow and lengthy process for rennet
coagulated cheeses, typically lasting from three weeks to
two or more years (Fox, 2002). During this period
microbiological and biochemical changes occur, leading to
changes in cheese texture and the development of flavour
(McSweeney, 2004). Accelerated ripening of Cheddar can
be accomplished by increasing the ripening temperature, the
use of addi- tional cultures (derived from starters or non-starter
lactic acid bacteria (NSLAB) in live or attenuated form), or
the use of enzymes (in many cases products of genetic
modification, where legislatively approved) or cultures with
high enzyme activities. Such acceleration can reduce the
energy required for temperature control during maturation
and de- crease storage requirements leading to increased
profitability. Among these options the application of a higher
ripening temperature is the easiest to implement.
Our understanding of the impact of higher temperatures
on the microstructure of Cheddar cheese is far from
complete. The microbial
and biochemical changes induced by high temperatures have
been well documented for Cheddar (Aston, Giles, Durward,
& Dulley, 1985; Cromie, Giles, & Dulley, 1987; Folkertsma,
Fox, & McSweeney, 1996) and the link between accelerated
flavour development and higher ri- pening temperatures
established (Aston et al., 1985; Folkertsma et al., 1996; Hannon
et al., 2005). Few studies, however, have probed the link
between elevated ripening temperature and Cheddar
microstructure. The link between this microstructure and
changes in cheese properties has also not been studied.
Industrial Cheddar cheese is conventionally ripened at a
constant temperature in the range of 6 °C–8 °C. Short
periods of 70 h at 25 °C were shown to have little impact on
the structure observed in Cheddar cheese by confocal laser
scanning microscopy (CLSM) (O’Reilly et al., 2003). The
structure observed was characteristic of Cheddar soon after
manufacture, which is typically described as containing a
continuous protein matrix with irregular shaped fat globules
embedded within the protein network; curd junctions may
also be present in the structure (Auty, Twomey, Guinee, &
Mulvihill, 2001).

⁎
Corresponding author at: The Bio21 Molecular Science and Biotechnology Institute, The University of Melbourne,
Parkville, Victoria 3010, Australia.
E-MAIL ADDRESS: sgras@unimelb.edu.au (S.L. Gras).
1
Present address: Moorepark Food Research Centre, Teagasc Moorepark, Fermoy, Co. Cork, Ireland.

http://dx.doi.org/10.1016/j.foostr.2017.05.003
Received 31 August 2016; Received in revised form 23 May 2017; Accepted 23 May 2017
Availableonline01June2017
2213-
3291/CrownCopyright©2017PublishedbyElsevierLtd.Allrig
htsreserved.
K. SOODAM et AL. FoodStructure1
4(2017)8–16

Proteolytic activity is known to be higher in Cheddar at
elevated ripening temperatures (Aston et al., 1985;
Folkertsma et al., 1996), leading to a softer cheese when
ripened at 16 °C (Folkertsma et al., 1996). Lipolysis can also
be accelerated (Folkertsma et al., 1996). Other textural
changes reported include a more brittle and less springy
Cheddar cheese when the temperature was increased to
20 °C (Fedrick & Dulley, 1984). These changes all indicate a
breakdown in the protein network and potential
rearrangement in the underlying mi- crostructure of the
cheese that could be visualised using microscopic
techniques.
Another potential consequence of ripening at elevated
tempera- tures, not linked directly to the microstructure, is
that changes in the bacterial population and proteolysis
may affect the concentration of amino acids and biogenic
amines, such as putrescine, cadaverine and tyramine. Both
amino acids and biogenic amines were observed to
significantly increase with temperature and storage in
Dutch type cheeses (Komprda et al., 2007; Pachlová,
Buňka, Flasarová, Válková, & Buňková, 2012), potentially
introducing a health hazard for consumers. The final
concentration of amines in Cheddar cheeses ri- pened at
high temperatures at the end of the accelerated ripening
period would therefore be valuable data to assess the relative
merits of incubation at high temperatures.
The aims of the current study were to investigate the
impact of elevated temperature on the microstructure, texture
and proteolysis of ripened Cheddar cheese, applying
quantitative image analysis to es- tablish structure-function
relationships during ripening. Four treat- ments were
applied: ripening at 8 °C, 15 °C, 20 °C, or a two-stage ri-
pening process involving ripening at 15 °C for 33 days
followed by ripening at 8 °C for the subsequent ripening
period. The presence of biogenic amines was also
assessed. This study provides new insights into the
microstructural changes that occur within Cheddar during
accelerated ripening and the link between changes in cheese
structure and functionality.
2. Materials and methods
2.1.Cheese TREATMENT
Three 20 kg blocks of commercial full fat Cheddar cheese,
made with pasteurised milk and a defined-strain LACTOCOCCUS
LACTIS bulk starter culture, were obtained from the same vat of
treatment were analysed using the methods described below,
except for the amine analysis which used six 500 g block
samples.
2.2.COMPOSITIONAL ANALYSIS of cheese
Cheeses were analysed in triplicate for fat, protein (%N x
6.38), salt and moisture using the gravimetric method
AS2300.1.3 (Australian Standard, 2008), the Kjeldahl
method AS2300.1.2.1 (Australian Standard, 1991), a
potentiometric method (AOAC, 2012) and the method
described in AS2300.1.1 (Australian Standard, 1988) respec-
tively. The pH of the fresh and mature cheese was
determined by blending 10 g of cheese with 10 g of water
purified to a resistivity of
18.2 MΩ (Millipore MilliQ, Billerica, MA, USA) and the pH
of the re-
sultant slurry was then measured using an electrode pH
meter (Orion 720A, Orion Pacific, Frankston, Australia).
Microscopic techniques
Cheese samples were analysed during ripening at days 1, 12,
33, 98, 194 by confocal laser scanning microscopy (CLSM,
Leica TCS SP2; Leica Microsystems, Heidelberg, Germany),
using a method described pre- viously (Soodam, Ong, Powell,
Kentish, & Gras, 2015). Stock solutions of Nile Red and Fast
Green FCF were prepared at a concentration of 1 mg/mL in
dimethyl sulfoxide and water respectively. These solutions
were then diluted 10 fold just prior to staining. A total of 6
images and 3 three-dimensional pictures were collected for
each cheese treatment at each time point during ripening.
Reconstruction and analysis of the three-dimensional images
was carried out using Imaris image proces- sing software
(Bitplane, South Windsor, CT, USA) as described pre-
viously (Ong, Dagastine, Kentish, & Gras, 2012) and yielded
parameters as described by Soodam et al. (2015).
Cheese samples were also analysed with cryo scanning
electron
microscopy (Cryo-SEM, Quanta; Fei, Hillsborro, OR, USA)
using a method described in a previous study (Ong, Dagastine,
Kentish, & Gras, 2011). A total of 6 images were taken at a
magnification of 2000×, 4000× or 8000× from one
representative cheese sample for each treatment at each
time-point (days 1, 12, 33, 98, 194) during ripening.
2.3.Proteolysis DETERMINATION

cheese and the blocks were then cut and packaged into at least
The proteolytic pattern of the cheese during ripening
84 ∼500 g block samples, which were then individually
was de- termined using sodium dodecyl sulfate
vacuum packed into commercial cheese barrier bags. The
polyacrylamide gel electro- phoresis (SDS-PAGE). The
cheese samples were then stored at 8 °C (T8), 15 °C followed by
cheese stock solution was prepared using the method of Ong,
8 °C (T15-8),15 °C (T15) or 20 °C (T20) for the ripening period
Henriksson, and Shah (2006), except that 25 mg sam- ples of
(330 days, 21 blocks of cheeses for each treatment). For one
cheeses were used. The stock solutions produced were
treatment (T15- 8), the cheese was stored at 15 °C for 33 days
then di- luted ∼4-fold with Trisaminomethane (Tris, 10
and then 8 °C for the remainder of the trial. At selected time
mmol/L)–Ethylene- diaminetetraacetic acid (EDTA 1 mmol/L)
points during ripening, three 500 g block samples of each

9
K. SOODAM et AL.
pH 8.0 buffer. Standards of α-casein (CN), β-CN and κ-CN
(Sigma-Aldrich, Castle Hill, Australia) were also prepared
at a concentration of 0.25 mg/mL. An aliquot of
6.5 μL was mixed with 2.5 μL of Bolt LDS Sample Buffer
(4×) and 1 μL Bolt Reducing Agent (10×) (Invitrogen, Mt
Waverley, Australia). The solution was heated at 70 °C for
10 min and then loaded into a 12% acrylamide Bis-Tris Plus
gel (Invitrogen). A set of molecular standards, SeeBlue Plus2
pre-stained (Invitrogen), was also loaded in each gel. Bolt 2-
(N-morpholino) ethanesulphonic acid SDS Running Buffer
(1×) (In- vitrogen) was used as buffer and the system
was run for 75 min at 125 V. The protein bands were then
stained and destained as described by Ong et al. (2006) except
7.5% v/v acetic acid was used and the de- staining
solution did not contain methanol. The gel was finally
scanned using a Fujifilm Intelligent Dark Box II with LAS-
3000 Lite V2.2 soft- ware and LAS-3000 Image Multi Gauge
(Fujifilm, Brookvale, Australia) to quantify the intensity of α-
and β-CN bands during ripening. The data presented are
representative of three gels.
2.4.DETERMINATION of AMINE content
2.4.1. EXTRACTION of SAMPLES
Hydrochloric acid (HCl, 1 mL, 0.1 M) was added to 0.1
g of grated cheese. The samples were then cryo-milled
(Precellys 24, Bertin Technologies, Île-de-France, France) at
−10 °C (program 2, 3 min total cycle; per cycle: 30 s spinning,
45 s waiting), followed by centrifugation for 10 min at 10,000
rpm (∼Th9e0f0a0t gla).yer was removed and the
supernatant was then collected (10 μL) for derivatization,
avoiding the casein fraction at the bottom of the tube.
2.4.2. DERIVATIZATION of SAMPLES
The derivatization was performed using a method
described by Boughton et al. (2011). Borate buffer (200
mM boric acid (Univar, Ajax Finechem, Seven Hills,
Australia), 35.7 μM13C,15N L-Valine, 10 mM Tris (2-
carboxyethyl)-phosphine (Sigma Aldrich), 1 mM ascorbic
acid (Sigma Aldrich) was prepared and adjusted to pH 8.8
using 2 M sodium hydroxide. The supernatant (10 μL) was
diluted 10 times using 0.1%formic acid and 10 μL of the
diluted sample was then mixed with 70 μL borate buffer
and 20 μL of 6-aminoquinolyl-N-hydroxysuccinimidyl
carbamate (AQC, Sigma-Aldrich, 10 mM in 100% acetonitrile)
deriva- tization reagent. The mixture was vortexed for 30 s
and mixed at 55 °C using a thermomixer (Eppendorf,
Macquarie Park, Australia) for 15 min. The samples were
then centrifuged and transferred into HPLC vials for LC–
MS analysis.
2.4.3. Liquid CHROMATOGRAPHY–MASS spectrometry
(LC–MS) ANALYSIS
Quantitative analysis was performed using a 1290 LC-
system cou- pled to a 6490 Electrospray Ionisation-Triple
Quadrupole MS (LC-QqQ- MS, Agilent, Santa Clara, CA,
USA) and a modified version of a method previously
described in the literature (Boughton et al., 2011). Samples
1
FoodStructure1
4(2017)8–16
and amine standards (2 μL) were injected into the system.
Ions were monitored in the positive mode using a dynamic
multiple reaction monitoring (DMRM) method optimized for
each analyte. The source, collision energies and fragmentor
voltages were optimized for each analyte by infusing a
derivatized standard with LC eluent. The source conditions
used were: sheath gas temperature 315 °C, gas flow 10 L/
min, nebulizer pressure 45 psi (310 kPa) and capillary
voltage 3800 V. Profiling experiments were performed by
multiple reaction monitoring (MRM) in electrospray ionisation
(ESI) positive mode by collecting transition of each selected
precursor ion to 171.1+ aminoquinoline (Amq) fragment ion
using collision energy of 20 V within the mass range of
100–1000 m/z.
The mobile phases used were (A) 0.1% formic acid (Sigma
Aldrich) in water and (B) 0.1% formic acid in acetonitrile.
Liquid chromato- graphy separation was achieved by
injecting samples into an Agilent Porshell 120 SB-C18 2.1 ×
100 mm, 2.7 μm column with a flow rate of 600 μL/min. The
temperature was kept at 65 °C, resulting in operating pressures
below 1000 bar with a 13 min run time. The gradient profile
used in the method is shown in Supplementary Table 1.
The limit of quantification, defined as the concentration at
which quantitative results can be reported with reasonable
confidence (Armbruster, Tillman, & Hubbs, 1994), was
0.001 μM.
2.4.4. Recovery experiments
The extraction was performed six times for the same
sample of cheese. Formic acid (30 μL) was added to the
supernatant (from the cryo-mill) of three of the samples (500
μL), acting as a negative control. The supernatant (500 μL
each) of the other three samples was spiked with a standard
solution containing a mixture of all the amines mon- itored
during this experiment (30 μL, 150 μM). The estimated recovery
for each amine was then calculated as shown below:
Recovery of amine = Concentration in spiked sample −
Concentration in non-spiked sample
concentration of additional amine
× 100
The recovery was 82–105% of the amines monitored during
the ex- periment.
2.5.Texture profiles
A texture analyser TA-XT Plus (Stable Micro Systems,
Surrey, England) was used to obtain the texture profile of
cheese samples using the method described by Soodam et
al. (2015). Texture analysis was performed at least three
times for each cheese treatment.
2.6.STATISTICAL ANALYSIS
Experimental analysis was performed at days 1, 12, 33, 98 and
194, except for the amine analysis, which was performed at
day 330. This delay was useful to determine whether the
ripening temperature had any effect on the amine content
after a longer ripening period, espe- cially as biogenic
amines are known to increase with ripening time and
temperature. For all analyses except for the amine analysis, all
K. SOODAM et AL.
cheese treatments were analysed in triplicate (three individual
∼500 g cheese blocks) and the results presented are the
mean of the three data points (n=3), where all replicate
chemical and physical analyses were averaged for each
cheese block. For the amine analysis, all cheese treatments
were analysed six times (six individual ∼500 g cheese blocks)
and the data are presented as a mean
of the six data points (n=6). Interactions be- tween ripening
time and ripening temperature were also analysed using a
general analysis of variance model in Genstat (VSN
International, Hertfordshire, UK) with the assumption that the
individual ∼500 g cheese blocks are considered
independent at the start of the study. If the effect of time or
temperature was significant, the data were further analysed
using Tukey’s paired comparison, with a significance level
of α = 0.05. In some cases, data at individual time points
may also be analysed using one way analysis of variance
(ANOVA) and Tukey’s test, with a significance level of
0.05, with respect to the temperature treatment.
Correlations were also analysed using Genstat (VSN
International).
FoodStructure1
4(2017)8–16
contain more large pores and thicker protein strands compared
to the T8 samples.
These qualitative observations are supported by image
analysis, which indicates a significantly lower number of
vertices in the rendered image of the T20 samples
compared to the T8 samples at day 33 (P < 0.05) (Fig.
2a). This measure may provide a comparative in-
dication of branching within the protein network and it
suggests that branching is more extensive in the protein
network of cheese incubated at lower temperatures and is
reduced at high temperatures where greater protein
solubilisation is expected to have occurred and the protein
appears compacted into thicker strands. The number of
vertices was statistically similar for both T15 treatments
and the T8 or T20 cheeses (P > 0.05), however, showing
that the change in protein mi- crostructure is gradual with
increasing temperature. No further major
changes to the cheese microstructure appear to occur with
further ri- pening for any of the four treatments, as the 3D
microstructure images of the protein network collected later
in ripening at day 194 appear similar to the images
collected at day 33 (Fig. 1).

3. Results and discussion

3.1.COMPOSITIONAL ANALYSIS of cheese

The mean compositions of the cheeses used for the

maturation study, determined one day after pressing, are
shown in Supplementary Table 2. The composition of the
cheese was within the parameters re- commended for
Cheddar ripening at elevated temperatures (Folkertsma et al.,
1996), which include a moisture less than 37%, a moisture in
fat- free-substance (MFFS) less than 55% and a fat in dry
matter (FDM) greater than 50%.
Some changes occurred in the cheese at the later stages
of ripening due to the visible release of a small volume of
whey from the T15-8 and T20 cheeses from day 98
onwards. This whey was collected from the cheese bag,
weighed and then expressed as a percentage of the weight of
the cheese block it originated from. More whey was released
from the T20 cheeses (4.3 ± 0.1% w/w for the T20
cheeses compared to
0.3 ± 0.2% w/w for the T15-8 cheeses at day 194). The fat
contents of the cheeses were again analysed at day 194.
The release in moisture resulted in an effective increase in
the fat concentration at day 194 (37.5 ± 0.2% w/w)
compared to day 1 (36.0 ± 0.4% w/w) (P < 0.05) for
the T20 cheeses.

3.2.Microstructure of CHEDDAR cheese throughout

ripening

Qualitative changes could be observed in the protein

network of the Cheddar as a function of temperature. These
changes are most apparent in the 3D microstructure of the
cheese samples at day 33 after some ripening (Fig. 1),
where the protein network within T20 samples ap- pears to

1
Fig. 1. Three dimensional (3D) confocal microscopy images of the microstructure of the protein network of the
ripened cheese at day 33 and day 194. Protein is stained green (FCF). The images are 3D reconstructions built from 40
layers, with each layer separated from the next by ∼0.25 μm and a total depth of 9.8 μm. The scale bars are 10 μm in
length. Images in colour can be viewed in the electronic version of this manuscript. In the black and white version of
this manuscript, protein is lighter coloured (For interpretation of the references to colour in this figure legend, the
reader is referred to the web version of this article).

However, the number of vertices in the protein network
was also observed to decrease significantly across the 194
day ripening period when analysed independently of
temperature treatment (P < 0.05). This observation is again
consistent with the protein solubilisation that is expected to
occur over the course of ripening. While the differences
between the temperature treatments are still visible
qualitatively after 194 days, there was no significant
difference in the number of vertices measured between
treatments at this time point (P > 0.05, Fig. 2b), possibly
reflecting the higher variability between replicate samples
towards the end of the ripening period.
Fig. 2. Number of vertices in the rendered protein network determined from 3D confocal microscopy images of the
ripened cheese at a) day 33 and b) day 194 for cheeses T8, T15-8, T15 and T20. The results are expressed as the mean
± the standard deviation of the mean (n=3). The number of vertices given is for a cheese volume of ∼140,000 μm3 with 40
layers, where each layer is separated by ∼0.25 μm, giving a total depth of 9.8 μm.
The distribution of the fat within the cheese did not
change during ripening, irrespective of the temperature
used for ripening. The number, sphericity and diameter of
the fat globules were not sig- nificantly affected by the
ripening temperature, ripening time or their interactions (P
> 0.05, data not shown). This observation is expected and
is consistent with previous observations that no further
clumping of fat globules occurs during the ripening of
Cheddar cheese due to the solid nature of the fat at the
typical ripening temperature of 7 °C (Guinee, Auty, &
Fenelon, 2000). At 20 °C, however, the proportion of milk
fat known to be liquid in milk is more than three times the
per- centage present at 8 °C (Norris, Gray, & Dolby, 1973). It
is likely that no further clumping or changes to fat were
observed in the context of cheese at 20 °C in this study,
due to the constraints of the protein net- work, although
whey was released by the T15 and T20 cheeses from day
98 onwards.
Only minor changes could be observed between the
different treatments at the different stages of ripening by cryo-
SEM (images not shown). Gaps were observed at the interface
between the protein net- work and fat towards the end of
ripening (day 194) when compared to images taken at day
1. These gaps are most likely an indirect result of
proteolysis, which causes a weakening of the protein network,
making the network more susceptible to changes during
sample preparation.
Low magnification cryo SEM images show that more
granules, possibly containing calcium salts, occur in the
cheeses matured at higher temperatures (T15 and T20
treatments, Fig. 3). These structures may occur as a result of
higher whey migration in these samples, which could result in
greater movement and concentration of salts, leading to
granule formation. Johnson (2004) suggested that the movement
of free whey within the cheese or to the surface of the
cheese helps to bring calcium lactate to nucleation sites,
from which granules (usually de- scribed as ‘crystals’) can
then develop. Higher ripening temperatures may also
increase the rate of calcium lactate deposition due to
lactate racemisation by cheese microflora (Chou,
Edwards, Luedecke, Bates, & Clark, 2003). The presence of
free whey and increased calcium lactate deposits may prove a
deterrent to the industrial use of these high temperatures for
long maturation periods.
3.3.Proteolysis in CHEDDAR cheese throughout ripening
Proteolysis is one of the primary reasons for changes in
cheese structure during ripening and the proteolytic
breakdown of the protein network was expected to coincide
with the structural changes observed by quantitative image
analysis. Proteolysis was measured using SDS- PAGE, as
shown in Fig. 4, and the protein bands were identified by
comparison to casein and molecular weight standards in
order to con- firm a correlation between the observed changes
in microstructure and the well-established patterns of protein
breakdown that occur during ripening.
Qualitatively there was a general decrease in residual
αs1-CN, αs2- CN and β-CN within the cheese with higher
ripening temperatures and with increased ripening time
(Fig. 4a). This trend could also be ob- served quantitatively
when the percentage of αs1-CN, αs2-CN and β-CN remaining
was determined by image intensity, with residual casein
significantly affected by both temperature and ripening
time
Fig. 3. Cryo-SEM microstructure of the ripened cheeses T8 (a), T15-8 (b), T15 (c) and T20 (d) at day 194. Images
were captured using a solid state detector at a magnification of 2000 x. The scale bars are 50 μm in length. Arrows in
white indicate possible crystalline salts.
Fig. 4. a) Image of a representative SDS-PAGE gel (4–12%) of day 1 cheese (Lane 1) and ripened cheeses T8
(Lanes 2, 6), T15-8 (Lanes 3, 7), T15 (Lanes 4, 8) and T20 (Lanes 5, 9) at day 33 (Lanes 2–5) and day 194 (Lanes
6–9), and changes in the proportion of b) αs1-CN, c) αs2-CN and d) β-CN, expressed as a percentage of the
concentration present at day 1 in T8 (■), T15-8 ( ), T15 ( ) and T20 ( ) cheeses at days 1, 33 and 194. Lane 1 was
used to indicate the cheeses at day 1, the starting point for all treatments. STD is the standard molecular weight
marker and α-CN, β-CN and κ-CN are pure standards of these compounds. The results are expressed as the mean ±
the standard deviation of the mean (n=3). The boxes contain the proteolytic products of αs-CN and β-CN, such as γ-
CN.

(P < 0.05, Fig. 4b–d). At the end of ripening at day 194, the
de- gradation of αs1-CN, αs2-CN and β-CN was significantly
higher for the T20 cheese incubated at the highest
temperature, compared to the control cheese T8 incubated
at the lowest temperature (P < 0.05).
This increase in proteolysis is consistent with changes in
Cheddar microstructure described above. Cheeses incubated
at lower tempera- tures (T8) had lower proteolysis, consistent
with the higher number of vertices (or network branches) in the
protein networks of these cheeses and lower number of large
pores. Cheeses incubated at higher tem- peratures (T20),
however, had greater proteolysis reducing the number of
vertices and leading to a compaction of the protein network,
with thicker protein strands and more large pores. The
increase in proteo- lysis over the period of ripening is also
consistent with the reduction in the number of protein
vertices in the cheese network during ripening as the protein
network is solubilised.
As expected, the trends in proteolysis are consistent with
reports in the literature where increased degradation of αs1-
CN and β-CN was observed in 9 month old Cheddar
cheese ripened at 16 °C compared to at 8 °C (Folkertsma et
al., 1996). While the cheeses were ripened for a shorter
interval here (∼ 6.4 months), this trend was already apparent
for the T20 cheeses and it is expected that an additional
ripening period of 3 months, typically employed in cheese
maturation, will further in- crease the differences between
the treatments.
The concentration of proteolytic products of αs-CN and β-
CN, such
as γ-CN (regions A, B), also qualitatively increased with the
increase in ripening temperature (Fig. 4a Lanes 2–9). This
observation is consistent with the decrease in the
concentration of CN observed here and past reports of
increases in proteolytic products. The relative amounts of α
and β-CN proteolysis were also as expected. The rate of β-
CN hydro- lysis, for example, was significantly lower than for
αs1-CN or αs2-CN at day 194 (P < 0.05, Fig. 4b–d), when
analysed independent of temperature treatment, consistent
with ∼ 50% of β-CN typically re- maining in mature
Cheddar (Banks, 2011).
The consistency between the proteolytic patterns observed
here and in the literature is valuable, as it allows comparison
between studies. It is likely that structural changes, similar
to those observed here, have also occurred and contributed
to the dynamics of cheese maturation in other studies
examining the effect of high temperature. It would be
interesting in future work, if appropriate labelling
techniques could be developed, to also map the
distribution of α and β-CN within the

structures observed by confocal microscopy, as the

differences in net- work structure and quantity of α-CN and
β-CN observed for the dif- ferent temperature treatments
suggest that the density of these proteins could also differ
within the protein network observed for the different
temperature treatments.

3.4.Texture profiles of CHEDDAR cheese throughout

ripening

A close relationship is known to exist between the

microstructure and texture of cheese; the structural changes
observed by microscopy were therefore expected to lead to
textural changes in the Cheddar cheeses. The hardness,
cohesiveness and chewiness were all sig- nificantly
affected by the ripening temperature, ripening time and
the interaction between these two variables (P < 0.05).
These observa- tions are consistent with the changes
detected in the microstructure of the cheeses at the early
stages of ripening (Fig. 1).
Towards the end of ripening at day 194, the T20
cheeses were significantly softer than the T8 and T15-8
cheeses (P < 0.05, Fig. 5a). This change is probably due
to accelerated proteolysis in T20, as cleaved casein has
less ability to contribute to the protein network (Lawrence,
Creamer, & Gilles, 1987). This suggests that the reduced
number of vertices and more compact protein network
with thicker
Fig. 5. Changes in the a) hardness, b) chewiness and c)
cohesiveness of T8 (■), T15-8 ( ), T15 ( ) and T20 ( )
cheeses at days 1, 12, 33, 98 and 194. The results are
ex- pressed as the mean ± the standard deviation of the
mean (n=3).
protein strands and greater number of large pores correlate
with a softer texture. In contrast, the protein network with a
higher number of vertices and thinner protein strands in T8 that
has been less subjected to proteolysis has a harder texture.
As expected, the cheeses also softened significantly from
day 1 to day 98 (Fig. 5a), when analysed independently of
temperature treat- ment. From day 98 to day 194, however,
the hardness of the cheeses increased significantly (P <
0.05). This behaviour has also been re- ported in the
literature (Chevanan & Muthukumarappan, 2007). One
possibility is that as the cheese ages, further proteolysis into
water so- luble peptides results in the exposure of new ionic
groups which then compete for the available free water in the
cheese. The water previously available for partial
solubilisation of the protein network becomes as- sociated
with the new ionic groups, making the cheeses harder
(Lawrence et al., 1987). Interestingly, changes in the
structure of the cheese could be observed by quantitative
analysis of microscopy images early in maturation but not at
the end point of 194 days, where such increases in hardness
occurred. The association of water with soluble peptides
released from the protein network may make detection of any
structural differences between the treatments more difficult at cheeses.
this time point. It is worthwhile to note that the T8 and the T15-8
The chewiness and cohesiveness of the cheeses were also cheeses have statistically similar concentrations of
observed amino acids at day 330 (P > 0.05), despite the 33 days
to decrease significantly with ripening time (P < 0.05, Fig. spent ripening at an elevated tem- perature for the T15-8
5b and c). The interaction between ripening time and
cheeses. While the impact of a higher con- centration of
ripening temperature was also significant for both
such amino acids on the flavour of aged Cheddar cheese is
cohesiveness and chewiness (P < 0.05). The rates of
change for both textural parameters differ between not known, previous studies did not find any changes in
temperature treatments and are particularly noticeable cheese flavor intensity when high concentrations of
during the first 33 days, where T8 cheeses appear to have a amino acids were present (McGarry et al., 1994),
slower decline in these two texturalproperties (Fig. 5b and c). suggesting the impact on flavor may be small. The total
This observation is again consistent with the lower degrees of concentration of free amino acids measured in this study
proteolysis occurring at lower temperatures and the highly (Fig. 6a,b) is of the order of ∼20 g/kg when ripened at the
branched protein network observed in these samples. standard temperature of 8 °C. While different methods were
used, Fenelon, Guinee, Delahunty, Murray, and Crowe
3.5.Amine CONCENTRATION in CHEDDAR cheese AFTER (2000) reported comparable total free amino acids of ∼15
ripening for 330 DAYS g/kg of cheese in retail mature Irish cheeses, while Guinee et
al. (2008) reported a concentration range of 27–57 g/kg
The effect of accelerated ripening on the concentration cheese in retail mature/vintage full fat Cheddar cheeses.
of amines was also assessed after an extended period of Conversely, others report significantly lower levels of these
ripening of 330 days, close to the typical period of 12 acids (e.g.,
months maturation. Free amino acids and amines occur as McCarthy et al. (2017) and Banks et al. (2001))
a result of bacterial growth and metabolism, processes The concentration of biogenic amines in cheese can
known to be significantly affected by the temperature used increase with both age and temperature and become toxic at
during ri- pening (Pachlova et al., 2012). The results do high concentrations; this potential link has been established
not provide direct cor- relations to the cheese structure but in Dutch type cheeses (Komprda et al., 2007; Pachlová et
are instead useful as amines pro- vide another measure of al., 2012) but there appears to be only one report of
proteolytic and biochemical activity at the end of extended tyramine levels in American Cheddar ripened at 4 °C, 10 °C
ripening. Further, particular amines, known as biogenic or 16 °C (Dahlberg & Kosikowsky, 1949). A set of biogenic
amines, are a potential safety concern in cheeses ripened amines typi- cally monitored during cheese ripening
at elevated temperatures. including putrescine, cada- verine, tyramine, tryptamine,
The temperature treatments significantly (P < 0.05) phenethylamine and spermidine were therefore assessed here,
as shown in Fig. 6c. Only tryptamine and tyr- amine showed
affected the concentration of the most common amino clear trends with respect to ripening temperature.
acids (methionine, gluta- mine, glutamate, leucine, lysine,
phenylalanine, valine, tyrosine, serine, alanine, aspartate,
isoleucine, glycine, histidine, proline, triptophan) at day 330,
as shown in Fig. 6a, b, consistent with an acceleration of
ri- pening at high temperatures. The level of methionine
measured in this study was higher than reported in the
literature (Guinee, Kilcawley, & Beresford, 2008;
McCarthy, Kelly, Wilkinson, & Guinee, 2017), although the
reason for this difference is unknown. The T20 cheeses
have significantly higher concentrations of all amino acids
compared to the T8 cheeses, with the exception of
glutamine, which is lower in these cheeses. The cause for
the decrease of glutamine as a function of temperature is
not known, although some lactobacilli bac- teria are known
to deamidate glutamine to raise the pH of their im-
mediate environment, which in turn allows the bacteria to
grow in an environment where pH is the limiting factor
(Vermeulen, Gänzle, & Vogel, 2007). A past study
showed that ripening of Cheddar cheese at 15 °C resulted
in a higher glutamine concentration than ri- pening at 6 °C
(Puchades, Lemieux, & Simard, 1989). Proteolysis,
however, is quite complex and the lower glutamine
concentration may be due to increased breakdown as a
result of secondary biochemical changes in the T20
Fig. 6. Concentration of a, b) amino acids and c) biogenic amines, expressed as mg/kg of cheese, in T8 (□), T15-8 (
), T15 ( ) and T20 ( ) cheeses at day 330. The results are expressed as the mean ± the standard deviation of the mean
(n=6, except for T15-8: n=5).

The T20 cheeses having significantly higher concentrations than the T8 cheeses.
The concentrations of biogenic amines determined here (∼ < 25 mg/kg cheese) were lower than the concentrations reported
in the literature for Cheddar cheese aged more than 12 months (Laleye, Simard, Gosselin, Lee, & Giroux, 1987), where
concentrations can reach 140–580 mg/kg, 140–330 mg/kg and 140–250 mg/kg for tyramine, putrescine and cadaverine
respectively (Laleye et al., 1987). In that study, some cheeses only contained traces of putrescine and cadaverine.
Differences in hygienic conditions can have a large impact on the concentration of biogenic amines present (Velíšek, 2014),
suggesting that hygiene may be more critical than the temperature of maturation in avoiding high levels of biogenic amines. This
study confirms that biogenic amines are not a concern in the industrial samples examined even when matured at higher
temperatures, although the situation could vary depending on the microbial composition of the starter cul- ture used and the
non-starter cheese flora.

4. Conclusion
The microstructure of Cheddar cheese was found to be altered by ripening at elevated temperatures. An increase in protein
solubilisation, from increased proteolysis, could be visualised by CLSM with fewer protein branches, thicker protein strands and
larger pores appearing in the protein network ripened at 20 °C. The number of vertices de- termined by quantitative image
analysis of the rendered protein surface was also significantly less for cheese incubated at 20 °C. These changes in structure lead
to changes in textural properties; cheeses ripened at 20 °C for 194 days were softer, less chewy and less cohesive, as could be
expected for a protein network degraded by greater proteolytic activity with fewer protein branches. The concentrations of
biogenic amines were also observed to be low, confirming these compounds do not re- present a health concern under these
conditions. This study provides new insights into the structural changes that occur during accelerated cheese ripening and
illustrates the role microscopy and quantitative image analysis can play alongside existing analytical techniques to monitor
and finely control the development of cheese ripening.'
Conflict of interest
The authors declare no conflict of interest.
Acknowledgements
The authors acknowledge the Australian Government for providing the Australian Postgraduate Award (APA) scholarship
and Dairy Innovation Australia Ltd and its member companies for financial and practical support for this project (project grant
08209C). We also thank the Particulate Fluids Processing Centre (PFPC) and the Bio21 Institute for access to equipment including
the Electron Microscopy Unit and the Biological Optical Microscopy Platform. Mr Roger Curtain also helped to operate the
scanning electron microscope in cryo mode. We also thank Dr Thusitha Rupasinghe, a research and development scientist of
Metabolomics Australia, School of BioSciences, the University of
Melbourne (Australia) who provided assistance in the
design and in- terpretation of the amine data and facilitated
access to the equipment. We acknowledge Ms Rachel Sore,
for her advice concerning the statis- tical analysis. Sally Gras
and Lydia Ong are supported by The ARC Dairy Innovation
Hub; a collaboration between The University of Melbourne,
The University of Queensland and Dairy Innovation
Australia Ltd sponsored by the Australian Research
Council's Industrial Transformation Research Program
(ITRP) funding scheme (project number IH120100005).
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at
http://dx.doi.org/10.1016/j.foostr.2017.05.003).
References
AOAC (2012). OffiCIAL methods of ANALYSIS of AOAC
INTERNATIONAL, method 971.27.
Gaithersburg, USA: AOAC International.
Armbruster, D. A., Tillman, M. D., & Hubbs, L. M. (1994).
Limit of detection (LQD)/limit of quantitation (LOQ):
comparison of the empirical and the statistical methods
ex- emplified with GC–MS assays of abused drugs.
CLINICAL Chemistry, 40(7), 1233–1238.
Aston, J. W., Giles, J. E., Durward, I. G., & Dulley, J. R.
(1985). Effect of elevated ripening
temperatures on proteolysis and flavour development in
Cheddar cheese. JOURNAL of DAIRY RESEARCH, 52(04),
565–572.
Australian Standard (1988). Methods of CHEMICAL AND
PHYSICAL testing for the DAIRYING in- dustry –
GENERAL methods AND principles – DETERMINATION of
TOTAL solids AND moisture. Australian Standard [AS
2300.1.1].
Australian Standard (1991). Methods of CHEMICAL AND
PHYSICAL testing for the DAIRYING in- DUSTRY –
GENERAL METHODS AND PRINCIPLES – DETERMINATION
OF NITROGEN – REFERENCE KJELDAHL method.
Australian Standard [AS 2300.1.2.1].
Australian Standard (2008). Methods of CHEMICAL AND
PHYSICAL testing for the DAIRYING in- dustry −
GENERAL methods AND principles − DETERMINATION of
FAT − GRAVIMETRIC method. Australian Standard [AS
2300.1.3].
Auty, M. A. E., Twomey, M., Guinee, T. P., & Mulvihill, D. M.
(2001). Development and application of confocal scanning
laser microscopy methods for studying the dis- tribution of
fat and protein in selected dairy products. JOURNAL of
DAIRY RESEARCH, 68(03), 417–427.
Banks, J. M., Yvon, M., Gripon, J. C., Miguel, A.,
Brechany, E. Y., Williams, A. G., & Muir,
D. D. (2001). Enhancement of amino acid catabolism in
Cheddar cheese using α- ketoglutarate: Amino acid
degradation in relation to volatile compounds and aroma
character. INTERNATIONAL DAIRY JOURNAL, 11(4), 235–
243.
Banks, J. M. (2011). Cheese | Cheddar-type cheeses. In J. W.
Fuquay, P. F. Fox, & P. L. H. McSweeney (Eds.),
ENCYCLOPEDIA of DAIRY Sciences (pp. 706–711). (2nd
ed.). San Diego: Academic Press.
Boughton, B. A., Callahan, D. L., Silva, C., Bowne, J.,
Nahid, A., Rupasinghe, T., ...
Roessner, U. (2011). Comprehensive
profianlidngquantitation of amine group containing
metabolites. ANALYTICAL Chemistry, 83(19),
7523–7530.
Chevanan, N., & Muthukumarappan, K. (2007). Effect of
calcium and phosphorus, re- sidual lactose, and salt-to-
moisture ratio on the melting characteristics and hardness of
Cheddar cheese during ripening. JOURNAL of Food
Science, 72(4), E168–E176.
Chou, Y. E., Edwards, C. G., Luedecke, L. O., Bates, M. P., &
Clark, S. (2003). Nonstarter lactic acid bacteria and aging
temperature affect calcium lactate crystallization in
Cheddar cheese. JOURNAL of DAIRY Science, 86(8),
2516–2524.
Cromie, S. J., Giles, J. E., & Dulley, J. R. (1987). Effect of
elevated ripening temperatures on the microflora of
Cheddar cheese. JOURNAL of DAIRY RESEARCH, 54(1),
69–76.
Dahlberg, A. C., & Kosikowsky, F. V. (1949). The influence
of temperature of ripening on the tyramine content and
flavor of American Cheddar cheese. JOURNAL of DAIRY
 Science, 32(4), 316–321.
Fedrick, I. A., & Dulley, J. R. (1984). The effect of elevated
storage temperatures on the rheology of Cheddar cheese.
New ZEALAND JOURNAL of DAIRY Science AND
Technology, 19(2), 141–150.
Fenelon, M. A., Guinee, T. P., Delahunty, C., Murray, J.,
& Crowe, F. (2000). Composition and sensory attributes of
retail Cheddar cheese with different fat contents. JOURNAL of
Food Composition AND ANALYSIS, 13(1), 13–26.
Folkertsma, B., Fox, P. F., & McSweeney, P. L. H. (1996).
Accelerated ripening of Cheddar cheese at elevated
temperatures. INTERNATIONAL DAIRY JOURNAL, 6(11-12),
1117–1134.
Fox, P. F. (2002). Cheese | overview. In H. Roginski, J.
W. Fuquay, & P. F. Fox (Eds.),
ENCYCLOPEDIA of DAIRY sciences (pp. 252–261).
Amsterdam: Academic Press.
Guinee, T. P., Auty, M. A. E., & Fenelon, M. A. (2000). The
effect of fat content on the rheology, microstructure and
heat-induced functional characteristics of Cheddar
cheese. INTERNATIONAL DAIRY JOURNAL, 10(4), 277–
288.
Guinee, T. P., Kilcawley, K. N., & Beresford, T. P. (2008).
How variable are retail vintage brands of cheddar cheese
in composition and biochemistry? The AUSTRALIAN
JOURNAL of DAIRY Technology, 63(2), 50–60.
Hannon, J. A., Wilkinson, M. G., Delahunty, C. M.,
Wallace, J. M., Morrissey, P. A., & Beresford, T. P.
(2005). Application of descriptive sensory analysis and
key chemical indices to assess the impact of elevated
ripening temperatures on the acceleration of Cheddar
cheese ripening. INTERNATIONAL DAIRY JOURNAL, 15(3),
263–273.
Johnson, M. (2004). Revisiting calcium lactate crystals in
cheese. DAIRY Pipeline, 16(1), 1–11.
Komprda, T., Smělá, D., Novická, K., Kalhotka, L.,
Šustová, K., & Pechová, P. (2007).
Content and distribution of biogenic amines in
Dutch-type hard cheese. Food Chemistry,
102(1), 129–137.
Laleye, L. C., Simard, R. E., Gosselin, C., Lee, B. H., &
Giroux, R. N. (1987). Assessment of Cheddar cheese
quality by chromatographic analysis of free amino acids
and biogenic amines. JOURNAL of Food Science,
52(2), 303–307.
Lawrence, R. C., Creamer, L. K., & Gilles, J. (1987). Texture
development during cheese ripening. JOURNAL of DAIRY
Science, 70(8), 1748–1760.
McCarthy, C. M., Kelly, P. M., Wilkinson, M. G., & Guinee,
T. P. (2017). Effect of fat and salt reduction on the
changes in the concentrations of free amino acids and
free fatty ACIDS IN CHEDDAR-STYLE CHEESES DURING
MATURATION. JOURNAL OF FOOD COMPOSITION AND
ANALYSIS, 59, 132–140.
McGarry, A., Law, J., Coffey, A., Daly, C., Fox, P. F., &
Fitzgerald, G. F. (1994). Effect of genetically modifying
the lactococcal proteolytic system on ripening and flavor
de- velopment in Cheddar cheese. Applied AND
ENVIRONMENTAL Microbiology, 60(12), 4226–4233.
McSweeney, P. L. H. (2004). Biochemistry of cheese ripening:
Introduction and overview. In P. F. Fox, P. L. H.
McSweeney, T. M. Cogan, & T. Guinee (Eds.), Cheese:
Chemistry, physics AND microbiology (pp. 347–360).
Academic Press.
Norris, G. E., Gray, I. K., & Dolby, R. M. (1973). Seasonal
variations in the composition and thermal properties of
New Zealand milk fat: II. Thermal properties of milk fat and
their relation to composition. JOURNAL of DAIRY
RESEARCH, 40(03), 311–321.
O’Reilly, C. E., Kelly, A. L., Oliveira, J. C., Murphy, P. M.,
Auty, M. A. E., & Beresford, T. P. (2003). Effect of
varying high-pressure treatment conditions on
acceleration of ri- pening of cheddar cheese. INNOVATIVE
Food Science & Emerging Technologies, 4(3), 277–
284.
Ong, L., Henriksson, A., & Shah, N. P. (2006). Development of
probiotic Cheddar cheese containing LACTOBACILLUS
ACIDOPHILUS, Lb. CASEI, Lb. PARACASEI and
BifiDOBACTERIUM spp. and the influence of these bacteria
on proteolytic patterns and production of organic acid.
INTERNATIONAL DAIRY JOURNAL, 16(5), 446–456.
Ong, L., Dagastine, R. R., Kentish, S. E., & Gras, S. L. (2011).
Microstructure of milk gel and cheese curd observed
using cryo scanning electron microscopy and confocal
microscopy. LWT − Food Science AND
Technology, 44(5), 1291–1302.
Ong, L., Dagastine, R. R., Kentish, S. E., & Gras, S. L. (2012).
The effect of pH at renneting ON THE MICROSTRUCTURE,
COMPOSITION AND TEXTURE OF CHEDDAR CHEESE. FOOD
RESEARCH INTERNATIONAL, 48(1), 119–130.
Pachlová, V., Buňka, F., Flasarová, R., Válková, P., &
Buňková, L. (2012). The effect of elevated
temperature on ripening of Dutch type cheese. Food
Chemistry, 132(4), 1846–1854.
Puchades, R., Lemieux, L., & Simard, R. E. (1989).
Evolution of free amino acids during the ripening of
Cheddar cheese containing added lactobacilli strains.
JOURNAL of Food Science, 8548(54), –888.
Soodam, K., Ong, L., Powell, I. B., Kentish, S. E., &
Gras, S. L. (2015). Effect of calcium chloride addition
and draining pH on the microstructure and texture of
full fat Cheddar cheese during ripening. Food
Chemistry, 181(0), 111–118.
Velíšek, J. (2014). ANTINUTRITIONAL, toxic AND other
BIOACTIVE compounds The Chemistry of Food.
Chichester, UK: John Wiley & Sons, 741–846.
Vermeulen, N., Gänzle, M. G., & Vogel, R. F. (2007).
Glutamine deamidation by cereal- associated lactic acid
bacteria. JOURNAL of Applied Microbiology, 103(4),
1197–1205.