Professional Documents
Culture Documents
Cheese Rind Microbial Communities: Diversity, Composition and Origin
Cheese Rind Microbial Communities: Diversity, Composition and Origin
net/publication/272190884
CITATIONS READS
120 1,824
4 authors:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Eric Dugat-Bony on 01 September 2015.
doi: 10.1093/femsle/fnu015
Advance Access Publication Date: 4 December 2014
Minireview
M I N I R E V I E W – Food Microbiology
ABSTRACT
Cheese rinds host a specific microbiota composed of both prokaryotes (such as Actinobacteria, Firmicutes and Proteobacteria)
and eukaryotes (primarily yeasts and moulds). By combining modern molecular biology tools with conventional,
culture-based techniques, it has now become possible to create a catalogue of the biodiversity that inhabits this special
environment. Here, we review the microbial genera detected on the cheese surface and highlight the previously
unsuspected importance of non-inoculated microflora—raising the question of the latter’s environmental sources and their
role in shaping microbial communities. There is now a clear need to revise the current view of the cheese rind ecosystem
(i.e. that of a well-defined, perfectly controlled ecosystem). Inclusion of these new findings should enable us to better
understand the cheese-making process.
1
2 FEMS Microbiology Letters, 2015, Vol. 362, No. 1
Our knowledge of microbial diversity in cheese rinds is still cesses for these cheeses differ in terms of three critical steps
fragmented but has expanded significantly over the last decade, (namely coagulation, draining and ripening), which means that
due to the use of molecular biology techniques to profile mi- cheese varieties have very different formats and physicochem-
crobial communities. Here, we review and summarize recent ical characteristics (Almena-Aliste and Mietton 2014). Accord-
literature on the microbial diversity of cheese rinds. This analy- ingly, a great variety of species assemblages has been observed.
sis enabled us to establish an up-to-date catalogue of the micro- The picture that has emerged from the published data is that
bial taxa associated with this special environment. Furthermore, microbial populations may differ extensively from one cheese
we draw a distinction between technological and adventitious variety to another but also within a given variety, as a func-
microflora, and focus on the latter’s possible origins. This evalu- tion of the place of production or the season (Viljoen et al., 2003;
ation should be of immense help in exploring unstudied cheese Rea et al., 2007). Nonetheless, the specific signature of a cheese
environments and retrieving isolates belonging to taxa not pre- surface community is due to the presence of a set of different
viously observed on the cheese surface. microbial species, rather than a particular species. Many phy-
lotypes are common to most cheese varieties but have varying
populations and levels of complexity. The microbial communi-
THE DISTRIBUTION OF BACTERIAL AND ties of cheese rinds range from simple to complex assemblages
FUNGAL GENERA WITHIN RIND harbouring Firmicutes, Actinobacteria, Proteobacteria, Bacteroidetes,
yeasts and moulds. These microorganisms vary in terms of their
COMMUNITIES
abundance and diversity during the cheese ripening process, de-
The composition of the cheese surface microbiota has been pending on the type of rind (bloomy, washed or natural) and
studied for several decades via the application of conventional, the technology used (soft, hard or semi-hard). At the beginning
culture-based analyses (Valdés-Stauber et al., 1997; Maoz, Mayr of the cheese-making process, lactic acid bacteria (LAB) from
and Scherer 2003; Viljoen, Khoury and Hattingh 2003; Feurer starter cultures are dominant in terms of cell counts. Over the
et al., 2004a; Mounier et al., 2005, 2009; Callon et al., 2006; Flo- first few days of ripening, yeasts and/or moulds colonize the
rez and Mayo 2006;Larpin et al., 2006; Lopandic et al., 2006; cheese surface. The viable cell count for yeast reaches around 6–
Rea et al., 2007; Goerges et al., 2008; Bleicher et al., 2010; Roth 8 log10 cfu/cm2 and then remains fairly constant until the end of
et al., 2010; Larpin-Laborde et al., 2011; Amato et al., 2012; Lavoie ripening. The progressive deacidification of the cheese surface
et al., 2012; Panelli et al., 2012; Gori et al., 2013; Cogan et al., 2014; by the yeasts and/or moulds favours the establishment of an
Gkatzionis et al., 2014) and, since the end of the 20th century, acid-sensitive, salt-tolerant bacterial community. The final bac-
by using culture-independent molecular biology methods based terial cell count is 1–2 log10 units higher than the yeast cell count
on the direct extraction of DNA and/or RNA from the matrix. (Cogan et al., 2014). After reviewing 33 cheese rind studies, we
The first culture-independent methods to be applied to cheese identified 104 bacterial genera (1 Acidobacteria, 28 Actinobacteria,
microbial communities were based either on molecular finger- 5 Bacteroidetes, 24 Firmicutes and 46 Proteobacteria) and 39 fungal
printing (such as PCR-denaturing gradient gel electrophoresis, genera (21 moulds and 18 yeasts). On average, 10.2 bacterial gen-
PCR-temperature gradient gel electrophoresis and PCR-single era (ranging from 3 to 30, depending on the cheese variety) and
strand conformation polymorphism analysis) or cloning and se- 4.5 fungal genera (ranging from 1 to 11) were detected in cheese
quencing of target rRNA sequences after PCR amplification (Er- rinds (Tables 1–3). We did not take account of the overall diver-
colini, Hill and Dodd 2003; Feurer et al., 2004a; Ogier et al., 2004; sity studies performed by Wolfe et al. (2014) and Quigley et al.
Callon et al., 2006; Parayre et al., 2007; Dolci et al., 2009, 2013; (2012) using HTS technologies because the types and varieties
Mounier et al., 2009; Fontana et al., 2010; Roth et al., 2010; Fe- of cheese were not specified. Among the yeasts, Debaryomyces,
ligini et al., 2012; Gori et al., 2013, Gkatzionis et al., 2014; Her- Yarrowia, Candida and Geotrichum were the most frequently de-
met et al., 2014; Schornsteiner et al., 2014). The recent devel- tected genera (present in 86%, 57%, 54% and 49% of the cheeses,
opment of high-throughput sequencing (HTS) technologies has respectively), followed by Kluyveromyces (32%) and Pichia (22%). Of
dramatically changed our view of microbial communities in- the filamentous fungi, Penicillium was most frequently detected
habiting fermented food products. In particular, metabarcoding (in 19% of the cheeses), followed by Scopulariopsis and Fusar-
analysis (e.g. deep sequencing of phylogenetic biomarker ampli- ium (both 8%). Staphylococcus was the most frequently found
cons, such as variable regions of rRNA genes or intergenic tran- Firmicutes genus (present in 78% of the cheeses). Other abun-
scribed spacers amplified using universal PCR primers) now en- dant Firmicutes were the LAB Lactococcus, Enterococcus, Lactobacil-
able microbial communities to be profiled with unprecedented lus, Streptococcus and Vagococcus, present in 50%, 41%, 25%, 22%
depth (Bokulich and Mills 2012; Ercolini 2013). To the best of and 16% of the cheeses, respectively, and halophilic lactic bacte-
our knowledge, only four publications have described the ap- ria such as Marinilactibacillus and Facklamia (present in 22% and
plication of HTS techniques to cheese rind samples (Quigley 16% of the cheeses, respectively). Considering the Actinobacteria,
et al., 2012; Bokulich and Mills 2013; Delcenserie et al., 2014; Wolfe Brevibacterium, Corynebacterium and Arthrobacter were the most
et al., 2014). However, it is likely that greater use of HTS tech- frequent genera (present in 75%, 75% and 66% of the cheeses,
niques in the future will result in large-scale, highly detailed de- respectively), followed by Brachybacterium, Microbacterium, Agro-
scriptions of the microbial communities from cheese rinds. coccus and Micrococcus (present in 38%, 38%, 19% and 19% of
The literature on the microbial diversity of cheese rinds the cheeses, respectively). The genera Psychrobacter, Halomonas,
(studied with both culture-dependent and -independent ap- Pseudoalteromonas and Vibrio (which are all halotolerant Pro-
proaches) is summarized in Tables 1 and 2. Although this list teobacteria) are also major cheese rind microorganisms, since
is not exhaustive, it encompasses most of the results generated they were detected in 33%, 31%, 22% and 19% of the cheeses,
over the last 15 years. Four main cheese types have been con- respectively.
sidered: (i) unpressed, uncooked soft cheeses with bloomy or Similar profiles have been obtained in recent, large-scale HTS
washed rinds; (ii) pressed, uncooked, semi-hard cheeses with studies. For example, Wolfe et al. (2014) characterized the fungal
natural or washed rinds; (iii) pressed, cooked hard cheeses with and bacterial diversity of 137 different cheese rinds collected in
washed rinds and iv) blue-veined cheeses. The production pro- the United States and 9 European countries, and Quigley et al.
Irlinger et al. 3
Table 1. The presence of bacterial genera in cheese rinds, as determined in selected biodiversity studies1 .
4
Frequency (%)
Frequency (%)
Methods2 CD CD CD CD CD CD CI CI CD CI CD CD CD CD CD CI CD CI CI CI CD CD CD CI CI CI CI CI CI CI HTS HTS HTS HTS
Candidatus solibacter
5
1 3
A
Agrococcus 1 1 1 1 1 1 19
Arcanobacterium 1 3
Arthrobacter 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 66 1 1 33
Brachybacterium 1 1 1 1 1 1 1 1 1 1 1 1 38 1 1 69
Brevibacterium 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 75 1 1 79
Citricoccus 1 1 1 9
Corynebacterium 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 75 1 1 47
Actinobacteria
Curtobacterium 1 3
Kocuria 1 3
Leucobacter 1 1 1 1 13
Microbacterium 1 1 1 1 1 1 1 1 1 1 1 1 38
Micrococcus 1 1 1 1 1 1 19
Mycetocola 1 1 6
Nocardiopsis 0 1 13
Propionibacterium 1 3
Pseudoclavibacter 1 3
Rothia 1 3
Yaniella 1 3 1 7
Prevotella 1 3 1
Psychroflexus 1 1
6
6
B
Sphingobacterium 1 3 1 22
Alkalibacterium 1 1 1 9
Atopostipes 1 3
Bacillus 1 3
Bavariicoccus 1 1 1 9
Carnobacterium 1 3
Enterococcus 1 1 1 1 1 1 1 1 1 1 1 1 1 41
Facklamia 1 1 1 1 1 16 1
Geobacillus 1 3
Firmicutes
Lactobacillus 1 1 1 1 1 1 1 1 25 1
Lactococcus 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 50 1
Leuconostoc 1 1 1 9 1
Macrococcus 1 1 6
Marinilactibacillus 1 1 1 1 1 1 1 22
Oenococcus 1 3
Sporanaerobacter 1 3
Staphylococcus 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 78 1 1 67
Streptococcus 1 1 1 1 1 1 1 22 1
Tetragenococcus 0 1
Vagococcus 1 1 1 1 1 16
Virgibacillus 1 3
4 FEMS Microbiology Letters, 2015, Vol. 362, No. 1
Table 1. (Continued.)
Acinetobacter 1 1 6
Advenella 1 3
Aeromonas 1 3
Agrobacterium 1 3
Alcaligenes 1 1 6
Aliivibrio 1 3
Arcobacter 1 1 6
Brevundimonas 1 3
Citrobacter 1 3
Cobetia 1 3
Cronobacter 0 1
Enhydrobacter 1 3
Enterobacter 1 1 6
Escherichia 1 1 6
Ewingella 1 3
Flavobacterium 0 1 1 11
Hafnia 1 1 6 1 51
Proteobacteria
Halomonas 1 1 1 1 1 1 1 1 1 1 31 1
Idiomarina 1 3
Klebsiella 1 1 6
Kluyvera 1 1 6
Marinobacter 1 1 6
Marinomonas 1 3
Methylobacterium 1 3
Morganella 1 3
Proteus 1 1 1 9
Providencia 1 3 1 15
Pseudoalteromonas 1 1 1 1 1 1 1 22 1 1 11
Pseudomonas 1 1 6 1 1 43
Psychrobacter 1 1 1 1 1 1 1 1 1 1 1 34 1
Pusillimonas 1 3
Raoultella 1 3
Serratia 1 3 1 11
Sphingomonas 1 3
Stenotrophomonas 1 3
Vibrio 1 1 1 1 1 1 19 1
unidentified species 1 1 1 1 13 14
unidentified bacteria 1 1 1 1 1 1 19 1
Number of genera 8 3 3 3 9 25 9 6 17 15 11 6 3 7 9 7 7 8 3 8 10 18 9 13 7 6 6 22 3 5 19 30 20 14
A total of 77 bacterial genera are listed (1 Acidobacteria, 18 Actinobacteria, 3 Bacteroidetes, 19 Firmicutes and 36 Proteobacteria).
1
The bibliographic references for the data in this table are as follows: Cogan et al. (2014), Goerges et al. (2008), Valdés-Stauber et al. (1997); Larpin et al. (2006), Larpin-
Laborde et al. (2011), Mounier et al. (2005, 2009), , Ogier et al. (2004); Feurer et al. (2004a,b) Bleicher et al. (2010), Maoz et al. (2003), Parayre et al. (2007), Rea et al. (2007), Gori
et al. (2013), Fontana et al. (2010), Feligini et al. (2012); Bokulich and Mills (2013), Amato et al. (2012), Roth et al. (2010), Dolci et al. (2009, 2013), Schornsteiner et al. (2014),
Ercolini et al. (2003), Quigley et al. (2012), Wolfe et al. (2014) and Delcenserie et al. (2014).
2
Various methods were used to detect dominant genera in cheese rind samples:
r Culture-dependent (CD) methods (traditional plating and isolation of bacteria, followed by genotyping).
r Culture-independent (CI) methods (direct extraction of microbial DNA from cheese rinds, followed by identification via molecular biology methods such as PCR-
DGGE/TGGE, PCR-CE-SSCP and 16S rDNA cloning/sequencing.
r High-throughput sequencing [HTS, direct extraction of microbial DNA from cheese rinds, followed by identification via V4/V1-V3 16S rDNA amplicon sequencing
(454 pyrosequencing or Illumina)].
3
The frequency calculation takes account of 31 different cheese varieties from all separate studies except Quigley et al. (2012) and Wolfe et al. (2014).
4
The frequency calculation takes account of 137 different cheese rinds from a single study (Wolfe et al., 2014).
5
Acidobacteria.
6
Bacteroidetes.
(2012) studied bacterial communities from 11 Irish cheese rinds. biodiversity. It is noteworthy that representative isolates from 24
In Wolfe et al.’s study (2014), 14 bacterial and 10 fungal dominant dominant genera detected by sequencing were easily retrieved
genera (representing >1% of the overall dataset) were distributed from dilutions of rind samples (Wolfe et al., 2014). On average,
across three different rind biofilms (bloomy, washed and natu- there were 6.5 dominant bacterial genera and 3.2 dominant fun-
ral) with varying abundance. Of these, only two fungal genera gal genera per rind sample (Table 3). These values are slightly
(Aspergillus and Sporandonema) and one bacterial genus (Nocar- lower than those estimated in our review on the basis of either
diopsis) were not detected in the other studies shown in Tables 1 culture-dependent or culture-independent techniques—almost
and 2. This observation shows that conventional culture-based certainly because (i) only dominant genera were considered in
and culture-independent techniques detect the same dominant the calculation and (ii) Wolfe et al. (2014) excluded LAB gen-
populations as HTS, leading to the conclusion that all dom- era (Lactococcus, Streptococcus, Leuconostoc and Lactobacillus) from
inant microorganisms are cultivable and the selective media their analysis. Indeed, 54 subdominant genera (44 bacteria and
used for enumerating and isolation of cheese strains are appro- 10 fungi) were detected in the study by Wolfe et al. (2014) (Ta-
priate and effectively allow the highlighting of the whole rind ble 4), including genera commonly found on the cheese surface
Irlinger et al. 5
Table 2. The presence of fungal genera in cheese rinds, as determined in selected biodiversity studies1 .
such as the eukaryotes Mucor, Kluyveromyces, Yarrowia and Pichia (i.e. the thickness of the rind samples, or whether the rinds are
and the bacteria Leucobacter and Microbacterium. Moreover, Wolfe scraped or cut off), since this may result in inadvertent sam-
et al. (2014) also detected 4 fungal genera and 27 bacterial gen- pling of the cheese core and thus contamination of the rind by
era that had never been mentioned in other reports. This finding LAB. By combining the literature results, we found that culture-
reflects the powerful ability of HTS technologies to access the di- dependent methods and culture-independent molecular biology
versity of low-abundant taxa for the first time. In terms of LAB techniques are complementary, rather than contradictory or ex-
(often considered to be cheese curd contaminants), the current clusive. On one hand, culture-based techniques are easily able to
literature suggests that these bacteria are indeed found in the characterize the individual properties of dominant members of
cheese rind itself (Table 1). Furthermore, Quigley et al. (2012) es- the cheese rind microflora (e.g. genomic repertoires, metabolic
timated that LAB taxa account for between 2 and 4.8% of the to- capacities and growth kinetics). On the other hand, molecular
tal bacterial community in washed rinds and up to 98% in some biology methods (especially those based on HTS technologies)
natural rinds. The reported presence of LAB in cheese rinds may are valuable for highlighting the existence of previously unex-
also be due to interstudy differences in the sampling technique pected subdominant populations. In turn, this information will
6 FEMS Microbiology Letters, 2015, Vol. 362, No. 1
Table 3. Comparison of genus richness in different types of cheeses, using culture-dependent and culture-independent methods.
6.3
be very helpful in designing specific culture media for the isola- nology (Wolfe et al., 2014). Nevertheless, G. candidum’s growth
tion of novel members of the cheese microbiota. is inhibited in cheeses with a high salt content (Boutrou and
Gueguen 2005). Penicillium roqueforti is a major ripening species
in blue-veined cheeses such as Roquefort, Gorgonzola, Stilton,
RIPENING ADJUNCT CULTURES VERSUS Danish Blue and Cabrales. In Stilton, it has been shown that P.
HOUSE MICROBIOTA roqueforti grows within the veins and not on the surface (Gkatzio-
If we take into account Bourdichon et al.’s (2012) inventory of mi- nis et al., 2014). Scopulariopsis flava is abundant in the washed
crobial food cultures (an update of the 2002 International Dairy rinds of Danish Havarti cheese, Swiss Tilsit, Austrian Vorarl-
Federation/European Food and Feed Cultures Association list of berger Bergkäse hard cheeses, and French tommes from the
cultures with ‘technological beneficial use’), some of the above- Pyrenees and Ossau-Iraty (Amato et al., 2012; Ropars et al., 2012;
mentioned genera may be considered as starters or ripening cul- Gori et al., 2013; Schornsteiner et al., 2014). It is also used as
tures that are deliberately inoculated to produce cheeses un- a ripening culture (Bourdichon et al., 2012; Hermet et al., 2012;
der well-controlled ripening conditions. Thus, the filamentous Ropars et al., 2012). Wolfe et al. (2014) also observed that mem-
fungal species that are highly adapted to the cheese surface in- bers of the Scopulariopsis genus are frequently detected in dry
clude Penicillium camemberti (a species only found in cheese, de- natural rinds and that the genus’ presence is negatively cor-
rived from P. commune), the filamentous yeast Geotrichum can- related with surface moisture levels. Several Mucor species (M.
didum and Fusarium domesticum. These species are commonly mucedo, M. plumbeus and M. racemosus) can be used as ripen-
used and found in cheeses requiring a velvety appearance, such ing cultures in some cheeses with washed or natural rinds, in-
as Camembert, Brie, Taleggio, Reblochon, Saint-Nectaire, Tilsit, cluding Saint-Nectaire, Tomme de Savoie and Taleggio (Barrios
Limburger, Brick and Raclette (Arteau, Labrie and Roy 2010). et al., 1998; Hermet et al., 2012). The most common commercial
The species F. domesticum (also referred to as ‘Anticollanti’ and ripening yeasts are Kluyveromyces lactis, K. marxianus and Debary-
formerly assigned taxonomically to the Trichotecium domesticum omyces hansenii (Lavoie et al., 2012). Given that D. hansenii is ubiq-
species or the Cylindrocarpon genus) is able to favour drying of uitous and highly salt tolerant, it may be present at the surface
the cheese surface and thus a reduction in the latter’s stickiness of most cheese varieties—even in cases in which it has not been
(Bachmann et al., 2005; Ropars et al., 2012). The presence of G. added deliberately. This species is very common in blue-veined
candidum (an anamorph of Galactomyces candidus) does not nec- cheeses but is also found in washed and bloomy rinds (Valdés-
essarily result from deliberate inoculation. Indeed, G. candidum Stauber et al., 1997; Cocolin et al., 2009; Gkatzionis et al., 2014;
is known to be ubiquitous and appears rapidly on the cheese sur- Schornsteiner et al., 2014). Kluyveromyces lactis and K. marxianus
face, regardless of geographical distribution or the type of tech- are involved in the very early steps of ripening (thanks to their
Irlinger et al. 7
Table 4. The subdominant bacterial and fungal genera in cheese rinds, as determined by an HTS study of 137 European and American cheeses
(Wolfe et al., 2014).
1
Bacteroidetes
2
The frequency calculation takes account of 137 different cheese rinds from a single study (Wolfe et al., 2014).
A total of 44 subdominant bacterial genera (15 Actinobacteria, 2 Bacteroidetes, 6 Firmicutes and 21 Proteobacteria) and 10 subdominant fungal genera (6 moulds and
4 yeasts) were detected in the 137 cheese rinds.
∗A total of 23 genera (5 Actinobacteria, 1 Firmicute, 11 Proteobacteria, 2 moulds and 4 yeasts) were found to be dominant in other biodiversity studies.
8 FEMS Microbiology Letters, 2015, Vol. 362, No. 1
ability to ferment lactose) but generally disappear after a few genus have been detected in Swiss and Austrian washed rinds
days of ripening (Larpin et al., 2006; Arteau, Labrie and Roy 2010; (Schornsteiner et al., 2014; Wolfe et al., 2014). The fungal airborne
Larpin-Laborde et al., 2011). genus Cladosporium is commonly detected in the rind of blue-
When considering the Firmicutes, the LAB Lactococcus lactis veined cheeses (e.g. Taleggio) and the rind of mould-ripened
ssp. lactis and L. lactis ssp. cremoris are the major components cheeses (e.g. Camembert) (Arteau, Labrie and Roy 2010; Panelli
of the mesophilic lactic starter cultures widely used in cheese et al., 2012).
making. They are frequently combined with other mesophilic Furthermore, several studies have shown that the bacterial
(Leuconostoc mesenteroides, L. pseudomesenteroides) or thermophilic flora dominating cheese surfaces belong to well-known inocu-
(Streptococcus thermophilus) LAB. The non-starter LAB (NSLAB) lated species assigned to genera Staphylococcus, Corynebacterium,
mainly correspond to facultative heterofermentative lactobacilli Brachybacterium, Arthrobacter, Microbacterium, Brevibacterium, De-
(Lactobacillus paracasei, L. casei, L. rhamnosus, L. plantarum, L. cur- baryomyces and Geotrichum, but that the dominating strains tend
vatus and Pediococcus acidilactici). In some technologies, NSLAB to originate from the ‘house’ microflora rather than the added
are added as adjunct cultures to accelerate the ripening pro- ripening culture (Feurer et al., 2004b; Mounier et al., 2006; Rea
cess; they help to produce the free amino acids that improve et al., 2007; Goerges et al., 2008; Gori et al., 2013). These find-
cheese flavour and reduce harshness and bitterness. Beside ings emphasize that (i) the microbial composition of cheese
LAB, other components of commercial cheese-making cultures rind is strongly influenced by the environmental communities
include Propionibacterium freudenreichii (an anaerobic actinobac- present in the area of cheese production and (ii) microorgan-
terium involved in flavour and eye formation in Swiss-type isms selected as ripening cultures for their expression of defined
cheeses), coagulase-negative staphylococci (mostly S. xylosus technological functions often behave differently in complex mi-
and S. equorum) and other aerobic Actinobacteria (Brevibacterium crobial communities, probably due to their poor adaptation to
aurantiacum, B. linens, Arthrobacter arilaitensis, Arthrobacter sp. cheese-making processes and their lack of competitive advan-
Corynebacterium casei, C. variabile, Brachybacterium alimentarium tage over indigenous microbiota.
and Microbacterium gubbeenense) that contribute to the flavour
and colour of surface-ripened cheeses (Bockelmann 2010). The
only Gram-negative bacterium used in commercial cultures is
RESERVOIRS OF MICROBIAL DIVERSITY
the enterobacterium Hafnia alvei. However, this species can also The presence of adventitious microorganisms in rind raises sev-
be adventitious. When H. alvei is inoculated into smear-ripened eral questions: (i) how do they colonize the cheese-making envi-
or mould-ripened cheeses manufactured with pasteurized milk, ronment, (ii) how do they transfer to the cheese surface and (iii)
it increases the level of volatile sulphur compounds and there- which technological parameters favour or hamper their devel-
fore improves the cheese flavour (Irlinger et al., 2012). opment? Adventitious microorganisms may be present through
It is noteworthy that the strains from commercial ripening the entire production chain, from the farm to the final product.
cultures used for the manufacturing of cheeses made from pas- This is especially true of dairy environments such as raw milk,
teurized milk are not necessarily found to be the dominant sur- dairy utensils (e.g. wooden shelves, vat or brines) and the atmo-
face microorganisms (Feurer et al., 2004b; Rea et al., 2007; Go- sphere of farms or the ripening cellar—all of which act as po-
erges et al., 2008; Larpin-Laborde et al., 2011; Gori et al., 2013). tential vectors (Fig. 1). Although many studies have focused on
For example, the species B. aurantiacum (formerly assigned to microbial diversity in raw milk (reviewed in Quigley et al., 2013),
B. linens and frequently used for the manufacturing of smear- few studies have looked at the impact of cheese-processing envi-
ripened cheeses) is only sporadically found on the surface of ronments on microbial community assemblages in specific pro-
these cheeses (Brennan et al., 2002; Mounier et al., 2005; Rea duction units. However, all the available studies show that each
et al., 2007; Goerges et al., 2008; Larpin-Laborde et al., 2011). manufacturing unit has a specific house microflora, which is
According to most of the studies considered in Tables 1 and
2, the cheese surface is dominated by adventitious microflora
and by microorganisms from starter and secondary cultures.
In a study of the microbial diversity of 137 cheese rind com-
munities (Wolfe et al., 2014), it was estimated that 60% of the
bacterial genera and 25% of the fungal genera detected did
not originate from starter or secondary cultures. Many of the
‘house’ flora cited in Tables 1 and 2 correspond to environmental
genera, such as marine halotolerant Proteobacteria (Psychrobac-
ter, Pseudoalteromonas, Halomonas, Vibrio and Advenella), and ma-
rine halophilic and alkaliphilic LAB (Vagococcus, Facklamia and
Marinilactibacillus) (Ishikawa et al., 2006; Mounier et al., 2009; Ble-
icher et al., 2010; Roth et al., 2011; Amato et al., 2012; Quigley
et al., 2012; Bokulich and Mills 2013; Gori et al., 2013; Schorn-
steiner et al., 2014; Wolfe et al., 2014).
Among the yeasts, Yarrowia lipolytica is ubiquitous and grows
spontaneously on cheese surfaces. It has been shown that
other dairy yeast species (including D. hansenii and G. can-
didum) can be outgrown by Y. lipolytica (Viljoen et al., 2003;
Mounier et al., 2008). Several species of Candida and Pichia are
also frequently found on the rind of European smear-ripened
cheeses (Larpin et al., 2006; Mounier et al., 2009; Larpin-Laborde Figure 1. Schematic overview of potential reservoirs of the different microbial
et al., 2011; Cogan et al., 2014). Some species have been detected groups colonizing cheese rinds. LAB, lactic acid bacteria; HALAB, halophilic and
only in specific rinds. For example, members of the Yaniella alkalophilic LAB.
Irlinger et al. 9
dependent on the environmental conditions prevailing during are dominated by yeasts, moulds and Actinobacteria, and corre-
the cheese making and is characterized by a typical, stable and lations have been observed with the microbial groups in cheese
recurrent microbiota that drives the ripening and potentially has rinds (Mounier et al., 2006; Mariani et al., 2007; Soares et al., 2011;
a role in the development of the cheese’s organoleptic properties Didienne et al., 2012).
(Viljoen et al., 2003; Mounier et al., 2006; Feligini et al., 2012; Pan-
elli et al., 2012; Bokulich and Mills 2013; Schirmer et al., 2013).
The microbial diversity in raw milk is substantial, since a sin- CONCLUSIONS AND FUTURE PERSPECTIVES
gle milk sample can contain as many as 36 dominant microbial
Thanks to efforts over the past decade, we are starting to obtain
species [for a review, see Montel et al. (2014)]. The microbial di-
a reliable, in-depth picture of the microbial diversity of cheese
versity in milk varies from farm to farm and is strongly influ-
rinds. In this review, we emphasized that the cheese rind envi-
enced by the overall farm management system. The teat surface
ronment hosts a variety of microorganisms from several sources
is the main source of useful cheese-making bacteria present in
(including commercial ripening adjunct cultures and adventi-
milk (Vacheyrou et al., 2011; Verdier-Metz et al., 2012). The fungal
tious flora). Indeed, cheese production and aging environments
species most frequently present in raw cow’s milk sampled at
may be a rich source of microbes throughout the course of fer-
different locations and different farms have been assigned to P.
mentation. However, little attention has been paid to the char-
commune, Y. lipolytica, D. hansenii, K. marxianus and several species
acterization of facility-specific ‘house’ microbiota and how the
of Candida (Viljoen et al., 2003; Fleet 2007; Vacheyrou et al., 2011;
latter are selected, survive and colonize cheese rinds. This is a
Lavoie et al., 2012; Panelli et al., 2013). While no significant
topic that needs to be explored further.
geographical influence was established, there is a difference
We have summarized the literature data at the genus level,
in microbiota composition between farm milk and the dairy
and have thus shed light on a common set of a dozen genera
tank milk. The former is dominated by Gram-positive bacte-
distributed across many cheese varieties. These findings sug-
ria (Staphylococcus, Macrococcus, Corynebacterium, Kocuria, Lactococ-
gest the existence of a core microbiota that has adapted to the
cus, Lactobacillus and Enterococcus) and the latter is dominated by
cheese surface. However, large-scale studies of phylotypes at
Gram-negative bacteria (Pseudomonas, Acinetobacter, Chryseobac-
the species or strain level are now required. Recent advances
terium, Achromobacter, Halomonas and Psychrobacter) (Delbès, Ali-
in HTS technologies have made this type of study feasible—
Mandjee and Montel 2007; Rasolofo et al., 2010; Fricker et al., 2011;
providing that similar progress in the tools’ taxonomic resolu-
Raats et al., 2011; Vacheyrou et al., 2011). This disparity may
tion can also achieved. Given that HTS and cloning-dependent
be attributed to differences in storage temperature and stor-
approaches are only semi-quantitative, it might also be inter-
age time. During cheese making, the main shifts in rind mi-
esting to use quantitative PCR to more precisely quantify the
crobial composition occurred during curd production and ripen-
abundance of certain phylotypes. Lastly, there is a need to focus
ing. This observation shows that the two key driving forces for
on the species’ ecological adaptation, colonization capabilities,
microbial growth are pH (which decreases during curd produc-
ability to resist and/or adapt to disturbances (e.g. environmen-
tion and increases at mid-ripening) and the salt content (Fig. 1).
tal changes or contamination by a pathogen) and mutual inter-
These physical–chemical attributes exert selective pressure on
actions. In particular, the food industry is looking for powerful
the microbiota and consequently favour microbial species that
tools that enable the in-depth, reliable characterization of food
are specifically adapted to these environmental constraints (Ir-
microbial communities and thus provides better control of pro-
linger and Mounier 2009). For example, the deacidification of
duction processes.
the cheese surface (due to the consumption of lactate and the
production of ammonia by moulds and yeasts) stimulates the
growth of acid-sensitive aerobic bacteria, such as Actinobacteria ACKNOWLEDGEMENTS
and Proteobacteria. Although there are too few data on the pres-
ence of strains related to alkaliphilic LAB (e.g. Marinilactibacillus, The authors thank Julien Irlinger for help with graphism of
Vagococcus, Facklamia) to form an opinion on the latter’s practi- figure.
cal significance and possible origin, the environmental study by
Conflict of interest statement. None declared.
Bokulich and Mills (2013) has shown that γ -Proteobacteria (e.g.
Psychrobacter, Pseudoalteromonas, Halomonas and Vibrio) present in
cheese rinds also dominated the washed-rind maturation room
REFERENCES
surfaces (i.e. the wash bucket, draining table, brine tank, ag-
ing rack and drains) in one of the two dairy facilities studied. Almena-Aliste M, Mietton B. Cheese classification, characteri-
These genera are known to be halophilic (or halotolerant) and zation and categorization: a global perspective. In: Donnelly
psychrotolerant, which could explain their ubiquitous distribu- CW (ed). Cheese and Microbes. Washington, DC: ASM Press,
tion in food products (cheese, meat and fish) and in natural mi- 2014, chapter 3, 39–71.
lieus such as sea water and soil (Olofsson, Ahrne and Molin Amato L, Ritschard JS, Kurtz O, et al. Microbial composition of de-
2007; Nychas et al., 2008; Margesin and Miteva 2011). In this re- fect smear—a problem evolving during foil-prepacked stor-
spect, the salting step (which consists in rubbing salt on the age of red-smear cheeses. Int Dairy J 2012;27:77–85.
cheese surface, or soaking cheeses in a brine solution) gener- Arteau M, Labrie S, Roy D. Terminal-restriction fragment length
ates a high salt concentration that favours the growth of these polymorphism and automated ribosomal intergenic spacer
microorganisms or other halotolerant microorganisms such as analysis profiling of fungal communities in Camembert
Staphylococcus, Brevibacterium, Arthrobacter, Corynebacterium and cheese. Int Dairy J 2010;20:545–54.
Leucobacter (Mounier et al., 2005, 2006p; Ishikawa et al., 2006; Fe- Bachmann HP, Bobst C, Bütikofer U, et al. Occurrence and signif-
ligini et al., 2012; Bokulich and Mills 2013; Cogan et al., 2014). icance of Fusarium domesticum alias Anticollanti on smear-
Wooden processing surfaces (including aging shelves and vats) ripened cheeses. LWT-Food Sci Technol 2005;38:399–407.
are also rich sources of microorganisms, which generally pro- Barrios MJ, Medina LM, Lopez MC, et al. Fungal biota isolated
duce biofilms that are stable over several seasons. These biofilms from Spanish cheeses. J Food Safety 1998;18:151–7.
10 FEMS Microbiology Letters, 2015, Vol. 362, No. 1
Margesin R, Miteva V. Diversity and ecology of psychrophilic mi- Raats D, Offek M, Minz D, et al. Molecular analysis of bacterial
croorganisms. Res Microbiol 2011;162:346–61. communities in raw cow milk and the impact of refrigeration
Mariani C, Briandet R, Chamba JF, et al. Biofilm ecology of on its structure and dynamics. Food Microbiol 2011;28:465–71.
wooden shelves used in ripening the French raw milk Rasolofo EA, St-Gelais D, LaPointe G, et al. Molecular analysis
smear cheese Reblochon de Savoie. J Dairy Sci 2007;90: of bacterial population structure and dynamics during cold
1653–61. storage of untreated and treated milk. Int J Food Microbiol
Montel MC, Buchin S, Mallet A, et al. Traditional cheeses: rich and 2010;138:108–18.
diverse microbiota with associated benefits. Int J Food Micro- Rea MC, Goerges S, Gelsomino R, et al. Stability of the biodiversity
biol 2014;177:136–54. of the surface consortia of Gubbeen, a red-smear cheese. J
Mounier J, Gelsomino R, Goerges S, et al. Surface microflora Dairy Sci 2007; 90:2200–10.
of four smear-ripened cheeses. Appl Environ Microb Ropars J, Cruaud C, Lacoste S, et al. A taxonomic and ecologi-
2005;71:6489–500. cal overview of cheese fungi. Int J Food Microbiol 2012;155:199–
Mounier J, Goerges S, Gelsomino R, et al. Sources of the ad- 210.
ventitious microflora of a smear-ripened cheese. J Dairy Res Roth E, Schwenninger SM, Eugster-Meier E, et al. Facultative
2006;101:668–81. anaerobic halophilic and alkaliphilic bacteria isolated from
Mounier J, Monnet C, Jacques N, et al. Assessment of the micro- a natural smear ecosystem inhibit Listeria growth in early
bial diversity at the surface of Livarot cheese using culture- ripening stages. Int J Food Microbiol 2011;147:26–32.
dependent and independent approaches. Int J Food Microbiol Roth E, Schwenninger SM, Hasler M, et al. Population dynamics
2009;133:31–7. of two antilisterial cheese surface consortia revealed by tem-
Mounier J, Monnet C, Vallaeys T, et al. Microbial interactions poral temperature gradient gel electrophoresis. BMC Microbiol
within a cheese microbial community. Appl Environ Microb 2010;10:74–87.
2008;74:172–81. Schirmer BCT, Heir E, Møretrø T, et al. Microbial background flora
Nychas GJE, Skandamis PN, Tassou CC, et al. Meat spoilage dur- in small-scale cheese production facilities does not inhibit
ing distribution. Meat Sci 2008;78:77–89. growth and surface attachment of Listeria monocytogenes. J
Ogier JC, Lafarge V, Girard V, et al. Molecular fingerprinting of Dairy Sci 2013;96:6161–71.
dairy microbial ecosystems by use of temporal temperature Schornsteiner E, Mann E, Bereuter O, et al. Cultivation-
and denaturing gradient gel electrophoresis. Appl Environ Mi- independent analysis of microbial communities on Austrian
crob 2004;70:5628–43. raw milk hard cheese rinds. Int J Food Microbiol 2014;180:
Olofsson TC, Ahrne S, Molin G. The bacterial flora of vacuum- 88–97.
packed cold-smoked salmon stored at 7◦ C, identified by di- Soares JC, Marques MR, Tavaria FK, et al. Biodiversity and char-
rect 16S rRNA gene analysis and pure culture technique. J acterization of Staphylococcus species isolated from a small
Appl Microbiol 2007;103:109–19. manufacturing dairy plant in Portugal. Int J Food Microbiol
Panelli S, Brambati E, Bonacina C, et al. Diversity of fungal flora in 2011;146:123–9.
raw milk from the Italian Alps in relation to pasture altitude. Vacheyrou M, Normand AC, Guyot P, et al. Cultivable micro-
SpringerPlus 2013;2:405. bial communities in raw cow milk and potential transfers
Panelli S, Buffoni JN, Bonacina C, et al. Identification of moulds from stables of sixteen French farms. Int J Food Microbiol
from the Taleggio cheese environment by the use of DNA bar- 2011;146:253–62.
codes. Food Control 2012;28:385–91. Valdés-Stauber N, Scherer S, Seiler H. Identification of yeasts
Parayre S, Falentin H, Madec MN, et al. Easy DNA extraction and coryneform bacteria from the surface microflora of brick
method and optimisation of PCR-temporal temperature gel cheeses. Int J Food Microbiol 1997;34:115–29.
electrophoresis to identify the predominant high and low Verdier-Metz I, Gagne G, Bornes S, et al. Cow teat skin, a poten-
GC-content bacteria from dairy products. J Microbiol Meth tial source of diverse microbial populations for cheese pro-
2007;69:431–41. duction. Appl Environ Microb 2012;78:326–33.
Quigley L, O’Sullivan O, Beresford TP, et al. High-throughput se- Viljoen BC, Khoury AR, Hattingh A. Seasonal diversity of yeasts
quencing for detection of subpopulations of bacteria not pre- associated with white-surface mould-ripened cheeses. Food
viously associated with artisanal cheeses. Appl Environ Microb Res Int 2003;36:275–83.
2012;78:5717–23. Wolfe BE, Button JE, Santarelli M, et al. Cheese rind communities
Quigley L, O’Sullivan O, Stanton C, et al. The complex microbiota provide tractable systems for in situ and in vitro studies of
of raw milk. FEMS Microbiol Rev 2013;37:664–98. microbial diversity. Cell 2014;158:422–33.