LWT - Food Science and Technology 107 (2019) 72–78

Contents lists available at ScienceDirect

LWT - Food Science and Technology

journal homepage: www.elsevier.com/locate/lwt

Lactobacillus rhamnosus GG improves the sensorial profile of Camembert- T

type cheese: An approach through flash-profile and CATA
Bruno Domingues Galli∗, Débora Parra Baptista, Flávia Giacometti Cavalheiro,
Mirna Lúcia Gigante
Department of Food Technology, School of Food Engineering, University of Campinas, UNICAMP, 13083-862, Campinas, SP, Brazil

A R T I C LE I N FO A B S T R A C T

Keywords: Camembert is a white-mould ripened cheese in which intense proteolysis, caused mainly by the fungus
Dairy Penicillium candidum, among other factors, leads to a substantial production of aromatic and flavor compounds.
Descriptive analysis The aim of this study was to evaluate the sensory profile of the Camembert-type cheese made from raw milk and
Food technology from pasteurized milk with or without the addition of Lactobacillus rhamnosus GG. The cheeses were analyzed
Adjunct culture
after 50 days of ripening through a combination of Flash-Profile (n = 15) and Check-All-That-Apply (CATA)
Lactobacillus rhamnosus GG
(n = 80) sensory methods with global acceptance. The 51 sensory attributes obtained allowed the separation of
the samples in three clusters. Through the CATA, a significant difference (p < 0.05) was observed for 21 of the
51 descriptors evaluated. Camembert-type cheese obtained from pasteurized milk with the addition of
Lactobacillus rhamnosus GG showed greater global acceptance when compared to the other cheeses, which did
not significantly differ between them. The addition of the adjunct culture was a good technological strategy to
improve the sensory characteristics of industrial Camembert-type cheese obtained from pasteurized milk,
bringing its sensory attributes closer to the ideal product and the sensorial characteristics of a cheese produced
with raw milk, but in a standardized and microbiologically safe way.

1. Introduction cheeses (Walstra, Wouters, & Geurts, 2006).

Camembert, a soft cylindrical cheese of high moisture content and
The sensory characteristics of foods are largely influenced by the covered with a velvet-like white layer originated by the growth of the
structural and biochemical changes that occur by both chemical reac- mould Penicilium candidum, presents an extremely complex ripening. Its
tions between compounds present in the food and by the action of proteolysis is affected by natural milk enzymes (especially when raw
microorganisms present environment in which it is inserted. The mi- milk is used), enzymes from the starter culture, NSLAB, residual coa-
croorganisms involved in such changes include either deteriorating gulant and also by the enzymes produced by the growth of the mould,
microorganisms that lead to the natural senescence of the food or even that have high proteolytic activity and gives the Camembert cheese its
microorganisms of technological interest such as fermenting bacteria or specific flavor and aroma characteristics (Shaw, 1981; Spinnler &
proteolytic moulds (Hutkins, 2006). Gripon, 2004).
Dairy products, especially cheeses, are susceptible to the enzymatic Thus, the sensorial differentiation caused by the use of raw or
hydrolysis of their protein chains during fermentation and ripening, a pasteurized milk on Camembert manufacturing has been extensively
process known as proteolysis. The enzymes involved in the proteolysis studied. These studies were conducted in order not only to determine
may be endogenous and exogenous from milk, residual coagulant, the organoleptic effect of the intrinsic characteristics of the milk com-
starter culture, secondary cultures, and non-starter lactic acid bacteria, position of each region and production system on cheese characteristics
the so-called NSLAB (Fox, Guinee, Cogan, & Mcsweeney, 2000; Guinee, but also to understand the unique and diversified microbiological
2017). Proteolysis is the most complex and important biochemical re- profile that the raw milk has compared to the standardized micro-
action that occurs during cheese ripening, being responsible for changes biological profile of a pasteurized milk (Chambers, 2005).
in hardness, elasticity and melting capacity, as well as substantially In this context, despite the widespread prohibition of the marketing
affecting the development of taste and aroma of different types of of cheeses obtained from raw milk with less than 60 days of ripening in

∗
Corresponding author.
E-mail address: galli.bd@gmail.com (B.D. Galli).

https://doi.org/10.1016/j.lwt.2019.02.077
Received 13 December 2018; Received in revised form 30 January 2019; Accepted 25 February 2019
Available online 02 March 2019
0023-6438/ © 2019 Elsevier Ltd. All rights reserved.
B.D. Galli, et al. LWT - Food Science and Technology 107 (2019) 72–78

Brazil and the United States, new comparative studies between cheeses University of São Paulo, Campus “Luiz de Queiroz” (ESALQ-USP). The
produced from milk with or without heat treatment are being per- milk was immediately cooled (4 °C) and processed within 2 h after
formed. Additionally, a growth interest in the use of adjunct cultures, milking. The raw milk (150 L) was divided into two portions containing
whether proteolytic, bioprotective or probiotic is also being observed. 50 and 100 L. The first milk portion (50 L) was kept raw and heated to
These cultures are used to modify the standardized and safe biota of 35 °C before the manufacture of Camembert-type cheeses. The second
pasteurized products leading to the production of new sensory profiles portion (100 L) was pasteurized (65 °C/30 min) in an electrical double
or even to modify their sensory characteristic in order to approximate jacketed milk pasteurizer, cooled to 35 °C and used to make
them with non-pasteurized products but in a controlled and micro- Camembert-type cheeses. The heat treatment efficiency was evaluated
biologically safe manner (Brasil, 2017; Fajardo, Ossa, & Hernández, by the activity of the enzymes alkaline phosphatase (AOAC, 2006) and
2016; FDA, 1949; Galli, Martin, Silva, Porto, & Spoto, 2016; Samelis peroxidase (Brasil, 1981).
et al., 2009; Van Hekken et al., 2008). Cheese production, ripening, and all analysis were performed at the
The thermophilic culture Lactobacillus helveticus has been frequently Department of Food Technology of the School of Food Engineering of
used in the manufacture of yoghurts and hard and semi-hard cheeses as the State University of Campinas (DTA/FEA - UNICAMP). All cheeses
an adjunct microorganism, while the mesophilic culture Lactobacillus were produced according to the same manufacturing protocol and on
rhamnosus has been increasingly used for the manufacture of soft and the same day.
fresh cheeses, due to the compatibility between the different techno- The cheeses were manufactured in stainless steel vats according to
logical processes and the viability of these strains and their ability to the traditional manufacturing method (Munk & Ferreira, 2008; Shaw,
modify the proteolysis behavior and the sensorial profile of the pro- 1981) with modifications. Initially, the pasteurized milk was further
ducts, while maintaining their compositional characteristics within the divided into two portions (50 L each). Lyophilized “O” type lactic acid
established standards (Azambuja et al., 2017; Baptista et al., 2018; culture consisting of Lactococcus lactis subsp. lactis and Lactococcus lactis
Chamba & Irlinger, 2004; Ibarra, Acha, Calleja, Chiralt-Boix, & Wittig, subsp. cremoris (DVS-R704 - Chr. Hansen, Valinhos, Brazil) was added
2012; Kocaoglu-Vurma, Harper, Drake, & Courtney, 2008). directly into the three manufacturing vats (0.03 g/L). The milk was pre-
The strain Lb. rhamnosus GG (ATCC 53103; LGG), also known as acidified until the acidity reached 22 °D (Dornic grades). After reaching
Lactobacillus GG (275), is one of the most known and referenced pro- this acidity, two doses (3 × 1010 CFU) of the lyophilized adjunct culture
biotic strains in the literature. This strain is classified as a Lb. rhamnosus consisting of Lactobacillus rhamnosus GG ATCC 53103 (Culturelle®) was
specie, a mesophilic culture with growth temperature between 2.6 °C added to only one vat of pasteurized milk.
and 52 °C and optimum growth temperature at 44.4 °C. The broad Then, calcium chloride (250 ppm) and chymosin coagulant (Ha La
growth temperature range of this bacterium makes it viable during the 1175, Chr. Hansen, Valinhos - SP, Brazil) sufficient to coagulate the
entire process of cheese manufacturing and also during refrigerated milk within 45 min were added. After reaching the cutting point (when
storage of the final product. In addition, its probiotic potential is a well- the acidity of the whey corresponded to 2/3 of the milk's initial acidity),
explored claim and its non-acidifying characteristic makes it techno- the curd was cut into 2 cm cubes and kept undisturbed for 3 min. Then
logically interesting adjunct culture because it does not affect the the slow stirring of the curd cubes was performed (stirring for 15 min at
acidification during the manufacturing process and the composition of minimum speed with rest of 3 min every 5 min). After the stirring, the
the products in which it is added (Mendes, Souza, Galletti Júnior, Leite, whey was drained off only until the curd cubes appeared on the surface.
& Penna, 2014; Saxelin & Kajander, 2008; Valík, Medveďová, & The curd was placed in cylindrical plastic forms (80 mm in diameter x
Liptáková, 2008). 100 mm in height) until filling the forms. After 1 h of rest, cheeses were
Based on these premises, the aim of this study was to evaluate the turned only once.
effect of the heat treatment and the use of adjunct culture in the de- After that, the cheeses were fermented for approximately 24 h at
velopment of the sensorial profile of Camembert-type cheese during controlled room temperature (25 °C) until the acidity of the exuded
ripening. The cheeses were made from raw and heat-treated milk, with whey reached 80 to 100 °D. After fermentation, the cheeses heights
and without the addition of Lactobacillus rhamnosus GG (ATCC 53103) were standardized at 3 ± 0.5 cm. Then, cheeses were salted in static
and ripened for 50 days. brine with 20% NaCl and 0.5% CaCl for 30 min. The volume of brine
used was 4 times greater than the volume of cheese and the brines were
2. Materials and methods kept at 4 °C throughout the salting period.
At the end of the salting, the cheeses were dried for 30 min (12 °C)
2.1. Experimental design and then homogenously pulverized with 0.1 g (2.0 × 109 CFU) of lyo-
philized P. candidum mould (Lyofast V5 – SACCO, Campinas, SP, Brazil)
Three batches of cheese were manufactured according to the fol- suspended in 500 ml of sterile saline solution (0.85% NaCl). After
lowing variations of heat treatment and lactic culture addition: 1) 30 min the cheeses were turned and subsequently pulverized with the
Camembert-type cheese made with raw milk using O type starter cul- same solution. The cheeses were then dried for 1 h. The cheeses were
ture, identified in this study as C1; 2) Camembert-type cheese made kept in the ripening chamber for 11 days (UR% 95/12 °C) with daily
with pasteurized milk with addition of O type starter culture, identified turns until the cheese surface was fully covered by the mould. Finally,
in this study as C2; 3) Camembert-type cheese made with pasteurized the cheeses were packed in aluminum foil and stored under refrigera-
milk with addition of O type starter culture and adjunct culture of tion (4 °C) for 50 days.
Lactobacillus rhamnosus GG, identified in this study as C3. The raw
material was characterized and the cheeses were analyzed for micro- 2.3. Sensory analysis
biological analysis required by the Brazilian legislation (Brasil, 2001).
The sensory methods Flash-Profile and “check all that apply” (CATA) The Flash-Profile and CATA (Check-all-that-apply) methods were
with global acceptance were carried out to obtain sensorial descriptors carried out after approval of the project by the Research Ethics
and to describe the sensory profile of cheeses after a ripening period of Committee of UNICAMP (CAAE 66734317.9.0000.5404) and carrying
50 days. out the microbiological analyzes required by the Brazilian legislation
(Brasil, 2001).
2.2. Cheese making Flash-Profile descriptive analysis followed the methodology de-
scribed by Dairou and Sieffermann (2002). The sensory analysis was
The milk used in the cheeses manufacturing was obtained from performed after 50 days of ripening. Fifteen assessors who did not have
Holstein cows from the herd of the Zootechnical Department of the any restrictions on the samples offered and who were familiar with

73
B.D. Galli, et al. LWT - Food Science and Technology 107 (2019) 72–78

Camembert cheese (monthly or higher frequency consumption) were In the statistical analysis, the data were tabulated as 55 columns (54
recruited. The group of tasters of the Flash-Profile analysis was com- corresponding to the attributes and 1 of the global acceptance) with 4
posed of 10 female tasters and 5 male tasters, aged between 20 and 50 data lines for each assessor, corresponding to each of the three samples
years. The Flash-Profile analysis was performed with 4 cheeses of each and the ideal virtual sample. The data related to the sensorial attributes
treatment which were obtained from the same process and the CATA were tabulated in a binary form, being assigned 1 when the attribute
analysis with 8 cheeses from each process batch. The tests were per- was mentioned and 0 when it was absent. Data on overall acceptance
formed in individual booths under white light. Samples (10 g wedges were tabulated from 0 to 9 as noted in the analysis sheet. When the
cut from the center of the pieces) were served at controlled ambient effects were significant, the differences were calculated by analysis of
temperature (25 ± 2 °C) in stainless steel trays encoded with three variance using the Cochran Q test. It was also performed analysis of
random digits. Mineral water at room temperature and cream-cracker multiple comparisons paired with the Marascuilo test, analysis of cor-
wafer were served so that the assessors could clean the palate between respondence to the attributes and the ideal sample through the chi-
the samples. square distances, besides analysis of main coordinates (PCA) applied to
Before the start of the analysis, a simplified explanation was given the correlation coefficients relative to acceptance. Additionally, the
to each assessor about the Flash-Profile method, describing the basic percentage of consumers who used each attribute to describe each
steps of the test. A protocol was also presented to be used during the sample differently and ideally was determined (Ares et al., 2014;
attribute survey and for the analysis, with instructions about how to Bruzzone et al., 2015).
evaluate the attributes of appearance (observing color and appearance),
aroma (smelling the sample twice), taste (tasting the sample), texture 3. Results and discussion
(chewing and evaluating the sensation in the mouth) and also any
sensation in the oral cavity after swallowing (residual). The milk used in all the processes was in accordance with the
For the survey of attributes, the samples were presented simulta- composition and microbiological safety parameters recommended by
neously, requesting that the assessors note the similarities and differ- the current Brazilian legislation (Brasil, 2001, 2011). The effectiveness
ences in a specific form. After the survey, in an individual interview of heat treatment applied to the pasteurized milk was demonstrated by
with each participant, a specific and personalized sample evaluation the positive activity of peroxidase and inactivation of alkaline phos-
form was set up for each assessor. Afterwards, the samples were again phatase.
presented, and the assessors were instructed to order the samples in All batches of cheese obtained were in accordance with the com-
ascending order of intensity for each of the previously defined attri- position parameters expected for this type of cheese and with the mi-
butes using the unstructured model scale for quantitative descriptive crobiological requirements required by the current legislation, without
analysis. The assessors were instructed that, if necessary, they could significant differences for discussion (Brasil, 1996; FAO/WHO, 2000;
change the evaluation form during the samples ranking procedure, re- FAO/WHO, 2011). Since the adjunct culture used is a non-acidifying
moving or including attributes, or changing the definitions in the strain, the acidification behavior during the process was identical for all
glossary. treatments, as well as the final pH of the cheese at the 1st day of ma-
To carry out the consumer perception and global acceptance, a turation. (Valík et al., 2008).
CATA sensory analysis was carried out with 80 assessors without re-
strictions or aversion to camembert cheese and with a minimal famil- 3.1. Flash-profile
iarity (having consumed this cheese in the last 6 months) with the
product, according to the methodology described by Ares et al. (2007). The tasters used different terms to characterize the cheeses ac-
The group of tasters of the CATA analysis was composed of 49 female cording to appearance, texture, aroma, and taste. The number of at-
tasters and 31 male tasters, aged between 18 and 60 years. Participants tributes varied between 5 and 9, with an average of 7 attributes for each
responded to a questionnaire containing sensory attributes generated participant and a total of 51 terms of differentiation (Appendix, Fig.
by the Flash-Profile analysis team to describe the samples, plus attri- A1). The residues by the configuration of the tasters were considered
butes added or grouped after group discussion in agreement with the 15 low, with a maximum value of 1.66 and an overall mean of 0.47. Re-
tasters of the Flash-Profile analysis. Also, a blank field was added to the sidues per sample object had a maximum value of 3.91 for sample C3
evaluation forms so that new attributes could be added if the CATA (cheese with adjunct culture), with a general mean of 2.37. These data
analysis assessors desired. show that there is a good consensus among the tasters and that the C3
In order to avoid tendencies, the position of the terms in the ques- sample was the one that generated the greatest difficulty among the
tionnaire was balanced and randomized for each listing according to tasters to survey attributes in a consensual way, probably due to their
the adapted methodology recommended by Ares, Dauber, Fernández, greater complexity of descriptors.
Giménez, and Varela (2014). Before the evaluation of all samples, as- The data obtained generated a total of 2 dimensions, that explained
sessors were asked to respond to an equal CATA form, marking the 100% of the total multivariate data. Fig. 1 represents, through the map
terms they considered to describe an ideal Camembert-type cheese, as of objects, the differentiation of the treatments in the two-dimensional
suggested by Bruzzone et al. (2015). In each analysis form corre- space. There is a clear separation between the samples, due to the
sponding to each of the samples, the assessors also evaluated the overall consistent differentiation between them, as well as little dispersion of
impression of the cheeses through a 9-point hedonic scale, which the results and good repeatability of the assessors. The consensus
showed terms ranging from “extremely disliked” to “extremely liked” among the team is confirmed by the low residual variance of the judges
(Stone & Sidel, 2004). in the two-dimensional solution, ranging from 14% to 27%. The good
The results of the Flash-Profile sensory analysis were analyzed by consensus among the evaluators demonstrated in the test may be due to
Generalized Procrustes Analysis through the Commandeur method. The the ease of ranking the product. Similar results were presented in the
Hedonic Acceptance (Global Impression of the sample) data from the description of commercial soy drinks by Flash-Profile, in which there
CATA analysis were analyzed considering the sample as a fixed source was a residual variance range of 27–60% per taster for a three-dimen-
of variation and the consumer as a random effect, at a significance level sional solution (Terhaag & Benassi, 2011).
of 5%. Table 1 lists the attributes obtained by the tasters, which were si-
For the CATA analysis, the Analysis of Penalties followed the one milar to the consumer's language, and their respective correlations with
proposed by Ares et al. (2014) and Meyners, Castura, and Carr (2013), dimensions 1 and 2. As a criterion for choosing the most relevant terms
this and other analyzes were also performed using XLSTAT 2015.6 for describing the sample, we chose to use coincident attributes be-
software. tween several tasters and with correlation, in the module, for each

74
B.D. Galli, et al. LWT - Food Science and Technology 107 (2019) 72–78

Fig. 1. Maps of the objects of Flash-Profile analysis ( = C1; = C2; = C3; = Consensus).

Table 1
Best correlated attributes (|r |) with the first two dimensions (D1, D2) for each of the 15 assessors in the Flash-Profile.
Assessors D1 D2

1 Moldy aroma (−0,60); Softness (0,60); Charac. taste (0,99); Uniform texture (0,99); Moldy taste (−0,99); Moldy aroma (0,80); Softness (−0,80); Ammoniacal
Salty (−0,99); Hardy rind (−0,99) taste (−0,92)
2 Creamy (−0,99); Ammoniacal taste (0,99); Soft texture (−0,99); Charac. taste. (0,60); Spicy (0,60); Firmness Charac. taste (−0,79); Spicy (−0,79)
(−0,99); Riccota taste (−0,99); Moldy taste (−0,99)
3 Ripening extension(0,99); Ammoniacal aroma (−0,92); Shiny (0,92); Bitterness (−0,99); Softness (0,99); Spicy (0,92); Ammoniacal taste (0,92)
Spicy (0,92); Ammoniacal taste (0,92)
4 Shiny (0,99); Ophaque (−0,99); Bitterness (0,99); Adhesivity (−0,99); Salty (−0,99); Firmness (−0,99) Soft rind (−0,92); Bitter rind (−0,92)
5 Softness (−0,99); Salty (−0,99); Bitterness (0,60; Acid aroma (0,99); Yellowness (−0,99); Rubbery (−0,99); Bitterness (−0,79); Hardy rind (0,92); Umami (0,79)
Umami (−0,60)
6 Shiny (0,99); Soft texture (−0,99); Buttery aroma(0,99); Buttery taste(0,99); Bitterness (−0,79); Acid taste Softness (−0,92); Bitterness (−0,60)
(−0,99); Acid aroma (−0,99)
7 Soft texture (−0,92); Shiny (0,99); Bitterness (−0,99); Creamy (−0,99) Bitter taste (0,92); Charac. Aroma (0,92)
8 Solid appearance (−0,99); Shiny (0,99); Pasty texture (0,99); Bitterness (−0,92); Salty (0,60); Whiteness Salty (−0,79)
(−0,99)
9 Creamy (−0,79); Softness (0,60); Bitterness (0,92); Quebradiço (−0,99); Sb. Amoniacal (−0,92) Creamy (−0,60); Softness (−0,79);
10 Uniform color (−0,99); Firmness (−0,99); Bitterness (−0,60); Sticky (0,99); Whiteness (−0,99); Acid aroma Bitterness (0,79); Acid Aroma (0,60)
(0,79)
11 Creamy (−0,79); Bitterness (−0,99); Soft core (0,99); Sweety aroma (0,93); Cheesy aroma (−0,99); Charac. Creamy (−0,60)
taste (0,99); Hardy rind (0,99)
12 Salty (0,60); Bitterness (0,60); Dry (−0,99); Creamy (−0,92); Shiny (0,99) Salty (−0,79); Bitterness (−0,79)
13 Firmness (−0,99); Creamy (−0,92); Acid taste (−0,99); Bitterness (−0,60); Salty (−0,99) Residual taste (−0,92); Bitterness (−0,79)
14 Ripened (−0,99); Milky taste (0,99); Firmness (0,99); Charac. aroma. (−0,92); Softness (−0,99); Uniform Bitterness (−0,92)
(−0,99)
15 Whiteness (0,92); Creamy (−0,92); Ammoniacal aroma (−0,92); Bitterness (−0,92); Spicy (−0,92); Residual Ripened taste (0,92); Buttery taste (0,92); Hardy rind
taste (−0,92) (0,99)

75
B.D. Galli, et al. LWT - Food Science and Technology 107 (2019) 72–78

Fig. 2. Symmetric plot of sensorial descriptors in CATA analysis ( = Sensorial descriptors; = Cheese samples).

taster greater than or equal to 0.5 (Terhaag & Benassi, 2011). The most which correlated with attributes such as ammoniacal taste and aroma
cited terms were sweet taste (negative correlation in D1), acid/sour and characteristic, spicy, earthy and bitter tastes. Camembert-type
taste (positive correlation in D1), sandy texture (negative correlation in cheese obtained from pasteurized milk without an adjunct culture was
D1 and positive correlation in D2), and characteristic aroma (negative the one that was most distant from the ideal product after 50 days of
correlation in D1, and positive correlation in D2). ripening, being associated with attributes such as rubbery, sandy and
firm texture, and opaque appearance. On the other hand, the Ca-
membert cheese made with raw milk was placed in the area with the
3.2. CATA analysis
greatest variety of descriptors, suggesting that despite the ideal sample
distance, this cheese has complex sensorial characteristics and quite
Significant differences in the frequencies of terms described by the
divergent to the other samples.
assessors were found for 21 terms of the 50 evaluated in the analysis,
According to Irlinger, Helinck, and Jany (2017), the bacteria pre-
suggesting that the CATA method was able to detect differences in
sent naturally in raw milk, especially the so-called NSLAB, play a cen-
participants' perceptions, as shown in the table of sensorial descriptors
tral role in the development of ketones, alcohols, aldehydes, and var-
(Appendix, Table A1). The ideal Camembert cheese was described as
ious flavor compounds, in addition to having proteolytic systems often
having characteristic taste and aroma, creaminess, ammoniacal aroma,
more potent and non-specific than traditional acidifying cultures.
bitter taste, white color, shiny and uniform texture in accordance with
Therefore, it is expected that the pasteurization of milk decreases the
the characteristics of this cheese variety as described by Mcsweeney
potentiality of production of sensorial descriptors related to this cheese,
(2004) and Engel, Tournier, Salles, and Le-Quéré (2001).
differentiating its sensorial characteristics and modifying the accep-
The result of the correspondence analysis across the chi-square
tance of the product as we can see in the results of the Flash-Profile and
distance can be seen in the Symmetric Plot (Fig. 2). As the p-value is
CATA analyzes.
lower than the significance level (p < 0.05), we conclude that it is very
In this same context, the addition of adjunct cultures in cheese-
likely that there are real differences between the products regarding
making can diversify the standardized biota of a pasteurized milk,
their sensory profiles.
bringing it closer to the diversity of raw milk with its original NSLAB
The cheeses obtained good separation of attributes distributed
biota and favoring the production of flavor compounds. The strain of
among the four quadrants, with Camembert cheese made with pas-
Lb. rhamnosus GG, for example, has a more efficient and diversified
teurized milk and Lb. rhamnosus was close to the ideal sample (Fig. 2),

76
B.D. Galli, et al. LWT - Food Science and Technology 107 (2019) 72–78

Fig. 3. Principal Coordinates Analysis plot of ac-

ceptance in CATA analysis ( = Sensorial de-
scriptors; X = Acceptance).

proteolytic system than bacteria commonly used as cheese starter cul- of the product. The attributes “Shiny” appearance and “ammoniacal”
tures. Its proteolytic system is composed by 2 proteases bound to the taste were classified as necessary attributes for this type of cheese. The
cell membrane and 23 internal peptidases with preferential hydrolysis attributes “creamy” texture and “bitter” taste were classified as inter-
sites at αs1-casein fractions and β-casein proline-rich peptides (Liu, esting. Attributes such as “uniform” texture and “salty” taste did not
Bayjanov, Renckens, Nauta, & Siezen, 2010; Savijoki, Ingmer, & influence the sensorial evaluation of the product. Interestingly, the
Varmanen, 2006). “sour or acid” taste attribute, a characteristic generally associated to
In addition, it is shown in the spatial position of the attributes in the fresh cheeses, with little or without ripening, that presents significative
Symmetric plot and the Principal Coordinate Analysis Plot (Figs. 2 and amount of organic acids, was classified as a negative attribute for
3) the antagonism, or in other words, the inverse correlation between Camembert-type cheese.
the attributes generated, suggesting that even attributes less cited as Finally, the global acceptance analysis applied to CATA also re-
being characteristics of an ideal Camembert-type cheese, such as earthy vealed which attributes had a greater correlation with the product ac-
and spicy taste, might be interesting for this type of cheese and in- ceptance (Fig. 3), besides showing the relation of these attributes with
tegrate at an important way the positive sensorial characteristic of this the ideal product described by the tasters and the average value in the
cheese since they are strongly opposed to negative terms such as sour hedonic scale of nine points for each cheese treatments (Table 2).
taste and unripened appearance. Likewise, the position of attributes All cheeses had good overall acceptance with scores between 6 and
cited by most tasters, but located in the spatial center of the graph, such 7 on the 9-point scale, which correspond to slightly liked and moder-
as the attributes dairy and salty taste suggests that although they are ately liked respectively. Despite this good overall acceptance of all
attributes that the evaluators consider to be characteristics of an ideal cheeses, the results showed that the addition of the adjunct culture Lb.
cheese, they are not attributes with great discriminative power or of rhamnosus GG significantly improved (p < 0.05) the sensory accep-
positive or negative relevance for Camembert-type cheese. These ob- tance of Camembert-type cheese, intensifying attributes of interest for
servations were also corroborated by the correlation matrix generated this cheese such as creaminess, uniform appearance, moderate bitter-
(Appendix, Table A2.). ness and other complex tastes and aromas.
The analysis of penalties allowed the selection of the essential or
interesting attributes for Camembert-type cheese, as well as the attri-
4. Conclusion
butes that do not influence or negatively affect the sensorial evaluation
The association between Flash-Profile and CATA analysis with
Table 2 global acceptance proved to be efficient in obtaining, differentiating
Global acceptance of Camembert-type cheeses (mean ± standard deviation) and evaluating sensorial descriptors of similar and complex products
(n = 80). such as Camembert-type cheese. All the cheeses had good sensorial
Treatment acceptance and wide variety of sensorial descriptors in consensus that
resulted in the separation of the cheeses according to the sensorial at-
C1 C2 C3
tributes that described them.
Global acceptance 6,2 ± 1,05b 6,1 ± 1,65b 7,5 ± 1,11a The use of adjunct culture Lb. rhamnosus GG proved to be an option
not only to approximate the Camembert cheese obtained from

77
B.D. Galli, et al. LWT - Food Science and Technology 107 (2019) 72–78

pasteurized milk of the ideal product but also a way of approaching its conventional profile and a quick original method, the Flash Profile. Journal of Food
sensorial characteristics to that of a cheese obtained with raw milk but Science, 67, 826–834.
Engel, E., Tournier, C., Salles, C., & Le-Quéré, J. L. (2001). Evolution of the composition
in a standardized and microbiologically safe way. Thus, the addition of of a selected bitter camembert cheese during ripening: Release and migration of taste-
the adjunct culture proved to be a viable technological strategy to im- active compounds. Journal of Agricultural and Food Chemistry, 49(6), 2940–2947.
prove its acceptance significantly. Fajardo, J. G. S., Ossa, T. I. P., & Hernández, E. J. (2016). Effect of pasteurization and
starter cultures on physicochemical and microbiological properties of costeño cheese.
Revista Facultad Nacional de Agronomía, Medellín, 69(2), 8007–8014.
Acknowledgements FAO/WHO (2000). Codex international individual standard for camembert: Codex STAN C-
33-1973. Codex alimentarius volume 12: Milk and milk products (2nd ed.). 94–96 Rome.
ISBN 92-5-104497-x.
This study was financed in part by the Coordenação de FAO/WHO (2011). Codex standard for camembert: Codex STAN 276-1973. Codex ali-
Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - mentarius: Milk and milk products (2nd ed.). 166–170 Rome. ISBN 978-92-5-105837-4.
Finance Code 001. The authors acknowledge the National Council for FDA (1949). Foods and drugs administration Code of federal regulations title 21, Part 133 -
cheeses and related cheese products12 C.F.R., §133. CFR.
Scientific and Technological Development (CNPq) for granted scholar-
Fox, P. F., Guinee, T. P., Cogan, T. M., & Mcsweeney, P. L. H. (2000). Fundamentals of
ships. cheese science. Gaithersburg: Aspen Publication (Chapter 2).
Galli, B. D., Martin, J. G. P., Silva, P. P. M., Porto, E., & Spoto, M. H. F. (2016). Sensory
Appendix A. Supplementary data quality of camembert-type cheese: Relationship between starter cultures and ripening
molds. International Journal of Food Microbiology, 234(3), 71–75.
Guinee, T. P. (2017). Pasteurized processed and imitation cheese products. In (4th ed.). P.
Supplementary data to this article can be found online at https:// L. H. McSweeney, P.F, P. D. Fox, Cotter, & D. W. Everett (Vol. Eds.), Cheese chemistry,
doi.org/10.1016/j.lwt.2019.02.077. physics & microbiology: Vol 2London-wall: Academic Press Elsevier (Chapter 46).
Hutkins, R. W. (2006). Microbiology and technology of fermented foods (1st ed.). Oxford:
Blackwell Publishing (Chapter 5).
References Ibarra, A., Acha, R., Calleja, M. T., Chiralt-Boix, A., & Wittig, E. (2012). Optimization and
shelf life of a low-lactose yogurt with Lactobacillus rhamnosus HN001. Journal of Dairy
Science, 95(7), 3536–3548.
AOAC (2006). Method V, method no. 979.13. In W. Horwitz (Ed.). Official methods of
Irlinger, F., Helinck, S., & Jany, J. L. (2017). Secondary and adjunct cultures. In (4th ed.).
analysis of AOAC International (pp. 37–38). (18th ed.). Gaithersburg: AOAC
P. L. H. McSweeney, P.F, P. D. Fox, Cotter, & D. W. Everett (Vol. Eds.), Cheese
International.
chemistry, physics & microbiology: Vol 1London-wall: Academic Press Elsevier
Ares, G., Dauber, C., Fernández, E., Giménez, A., & Varela, P. (2014). Penalty analysis
(Chapter 11).
based on CATA questions to identify drivers of liking and directions for product re-
Kocaoglu-Vurma, N. A., Harper, W. J., Drake, M. A., & Courtney, P. D. (2008).
formulation. Food Quality and Preference, 32, 65–76.
Microbiological, chemical, and sensory characteristics of Swiss cheese manufactured
Ares, G., Gonçalvez, D., Pérez, C., Reolón, G., Segura, N., Lema, P., et al. (2007). Influence
with adjunct Lactobacillus strains using a low cooking temperature. Journal of Dairy
of gelatin and starch on the instrumental and sensory texture of stirred yogurt.
Science, 91(8), 2947–2959.
International Journal of Dairy Technology, 60(4), 263–269.
Liu, M., Bayjanov, J. R., Renckens, B., Nauta, A., & Siezen, R. J. (2010). The proteolytic
Azambuja, N. C., Moreno, I., Gallina, D. A., Spadoti, L. M., Motta, E. M. P., Pacheco, M. T.
system of lactic acid bacteria revisited: A genomic comparison. BMC Genomics,
B., et al. (2017). Effect of adjunct culture Lactobacillus helveticus (B02) on the com-
11, 36.
position, proteolysis, free amino acids release and sensory characteristics of Prato
Mcsweeney, P. L. H. (2004). Biochemistry of cheese ripening. International Journal of
Cheese. Food and Nutrition Sciences, 8(5), 512–525.
Dairy Technology, 57(2–3), 127–144.
Baptista, D. P., Galli, B. D., Cavalheiro, F. G., Silva, F. N., Eberlin, M. N., & Gigante, M. L.
Mendes, D. P. G., Souza, M. R., Galletti Júnior, M., Leite, M. O., & Penna, C. F. A. M.
(2018). Lactobacillus helveticus LH-B02 favours the release of bioactive peptide during
(2014). Quality of fermented milks produced with Lactobacillus rhamnosus and
prato cheese ripening. International Dairy Journal, (87), 75–83. https://doi.org/10.
Lactobacillus fermentum isolated from artisanal cheeses. Arquivo Brasileiro de Medicina
1016/j.idairyj.2018.08.001.
Veterinária e Zootecnia, 66(4), 1291–1295.
Brasil, & Ministério da Agricultura Pecuária e Abastecimento – MAPA (1996). Portaria nº
Meyners, M., Castura, J. C., & Carr, B. T. (2013). Existing and new approaches for the
146 De 07 De Março De 1996: aprova os Regulamentos Técnicos de Identidade e
analysis of CATA data. Food Quality and Preference, 30(2), 309–319.
Qualidade dos Produtos Lácteos e Fixa a identidade e os requisitos mínimos de
Munk, A. V., & Ferreira, D. G. S. (2008). Produção de queijos finos: Morbier, Saint Paulin,
qualidade que deverão possuir os queijos de modo geral, com exceção dos Queijos
Gorgonzola e Camembert. Viçosa: Editora viçosa - centro de Produções técnicas(1st ed.).
Fundidos, Ralados, em Pó e Requeijão. Diário Oficial da União (Brasília, DF), section
(Chapter 3).
1, p. 3977.
Samelis, J., Lianou, A., Kakouri, A., Delbes, C., Rogelj, I., Bogovic-Matijasic, B., et al.
Brasil, & Ministério da Agricultura Pecuária e Abastecimento – MAPA (2011). Instrução
(2009). Changes in the microbial composition of raw milk induced by thermization
Normativa n°62, de 29 de setembro de 2011: altera a IN 51 de 2002 e aprova os
treatments applied prior to traditional Greek Hard Cheese processing. Journal of Food
Regulamentos Técnicos de Produção, Identidade e Qualidade do Leite tipo A, do Leite
Protection, 72(4), 783–790.
Cru Refrigerado, do Leite Pasteurizado e o Regulamento Técnico da Coleta de Leite
Savijoki, K., Ingmer, H., & Varmanen, P. (2006). Proteolytic systems of lactic acid bac-
Cru Refrigerado e seu Transporte a Granel. Diário Oficial da União (Brasília, DF),
teria. Applied Microbiology and Biotechnology, 71(4), 394–406.
section 1, pp. 6-1.
Saxelin, M., & Kajander, K. (2008). Commercially available human probiotic micro-
Brasil, & Ministério da Agricultura Secretaria Nacional de Defesa Agropecuária
organisms. In Y. Kun, & L. S. Salminen (Eds.). Handbook of probiotics and prebiotics
Laboratório Nacional de Referência Animal - LANARA (1981). Métodos analíticos
(pp. 441–532). (2nd ed.). New Jersey: John Wiley & Sons.
oficiais para controle de produtos de origem animal e seus ingredientes: Métodos Físico
Shaw, M. B. (1981). The manufacture of soft, surface mould, ripened cheese in France
Químicos. (Brasilia. DF).
with particular reference to Camembert. International Journal of Dairy Technology,
Brasil, & Ministério da Saúde Agência Nacional de Vigilância Sanitária (2001). Resolução
34(4), 131–138.
RDC nº 12, de 02 de Janeiro de 2001: aprova o regulamento técnico sobre padrões
Spinnler, H. E., & Gripon, J. C. (2004). Surface mould-ripened cheeses. In (3rd ed.). P. F.
microbiológicos para alimentos. Diário Oficial da República Federativa do Brasil
Fox, P. L. H. McSweeney, T. M. Cogan, & T. P. Guinee (Vol. Eds.), Cheese - chemistry,
(Brasília, DF), section 1, pp. 45-53.
physics and microbiology: Vol 2, (pp. 157–174). Oxford: Elsevier Academic Press.
Brasil Presidência da República - Casa Civil (2017). Decreto Nº 9.013, de 29 de março de
Stone, H., & Sidel, J. (2004). Sensory evaluation practices. (3ath). Cambridge: Academic
2017: regulamenta a Lei Nº 1.283, de 18 de dezembro de 1950, e a Lei Nº 7.889, de
Press, Elsevier Inc (Chapter 7).
23 de novembro de 1989, que dispõem sobre a inspeção industrial e sanitária de
Terhaag, M. M., & Benassi, M. T. (2011). Flash profile: An alternative for quick descriptive
produtos de origem animal. Diário Oficial da União (Brasilia. DF), section 1, p. 03.
analysis. Brazilian Journal of Food Technology, 14, 140–151.
Bruzzone, F., Vidal, L., Antúnez, L., Giménez, A., Deliza, R., & Ares, G. (2015).
Valík, L., Medveďová, A., & Liptáková, D. (2008). Characterization of the growth of
Comparison of intensity scales and CATA questions in new product development:
Lactobacillus rhamnosus GG in milk at suboptimal temperatures. Journal of Food and
Sensory characterisation and directions for product reformulation of milk desserts.
Nutrition Research, 47(2), 60–67.
Food Quality and Preference, 44, 183–193.
Van Hekken, D. L., Drake, M. A., Tunick, M. H., Guerrero, V. M., Molina-Corral, F. J., &
Chamba, J., & Irlinger, F. (2004). Secondary and adjunct cultures. In (3rd ed.). P. F. Fox,
Gardea, A. A. (2008). Effect of pasteurization and season on the sensorial and
P. L. H. McSweeney, T. M. Cogan, & T. P. Guinee (Vol. Eds.), Cheese - chemistry,
rheological traits of Mexican Chihuahua cheese. Dairy Science & Technology, 88(4–5),
physics and microbiology: Vol 1, (pp. 191–206). Oxford: Elsevier Academic Press.
525–536.
Chambers, J. V. (2005). The microbiology of raw milk. In R. K. Robinson (Ed.). Dairy
Walstra, P., Wouters, J. T., & Geurts, T. J. (2006). Dairy science and technology (2nd). Boca
microbiology handbook (pp. 39–90). (3rd ed.). New York: John Wiley & Sons.
Raton: Taylor & Francis Group (Chapter 25).
Dairou, V., & Sieffermann, J. M. A. (2002). Comparison of 14 jams characterized by

78