Scholarly article on topic 'Metatranscriptome analysis of fungal strains Penicillium camemberti and Geotrichum candidum reveal cheese matrix breakdown and potential development of sensory properties of ripened Camembert-type cheese'

Metatranscriptome analysis of fungal strains Penicillium camemberti and Geotrichum candidum reveal cheese matrix breakdown and potential development of sensory properties of ripened Camembert-type cheese Academic research paper on "Biological sciences"

Share paper
Academic journal
BMC Genomics
OECD Field of science

Academic research paper on topic "Metatranscriptome analysis of fungal strains Penicillium camemberti and Geotrichum candidum reveal cheese matrix breakdown and potential development of sensory properties of ripened Camembert-type cheese"



Metatranscriptome analysis of fungal strains Pénicillium camemberti and Geotrichum candidum reveal cheese matrix breakdown and potential development of sensory properties of ripened Camembert-type cheese

Marie-Hélène Lessard1, Catherine Vie!1, Brian Boyle2, Daniel St-Gelais1,3 and Steve Labrie1*


Background: Camembert-type cheese ripening is driven mainly by fungal microflora including Geotrichum candidum and Penicillium camemberti. These species are major contributors to the texture and flavour of typical bloomy rind cheeses. Biochemical studies showed that G. candidum reduces bitterness, enhances sulphur flavors through amino acid catabolism and has an impact on rind texture, firmness and thickness, while P. camemberti is responsible for the white and bloomy aspect of the rind, and produces enzymes involved in proteolysis and lipolysis activities. However, very little is known about the genetic determinants that code for these activities and their expression profile over time during the ripening process.

Results: The metatranscriptome of an industrial Canadian Camembert-type cheese was studied at seven different sampling days over 77 days of ripening. A database called CamemBank01was generated, containing a total of 1,060,019 sequence tags (reads) assembled in 7916 contigs. Sequence analysis revealed that 57% of the contigs could be affiliated to molds, 16% originated from yeasts, and 27% could not be identified. According to the functional annotation performed, the predominant processes during Camembert ripening include gene expression, energy-, carbohydrate-, organic acid-, lipid- and protein- metabolic processes, cell growth, and response to different stresses. Relative expression data showed that these functions occurred mostly in the first two weeks of the ripening period. Conclusions: These data provide further advances in our knowledge about the biological activities of the dominant ripening microflora of Camembert cheese and will help select biological markers to improve cheese quality assessment.


Camembert cheese is a soft, mold-ripened cheese. The mold Penicillium camemberti and the yeast Geotrichum candidum are the two major Fungi that give the white coated characteristic of this cheese variety. Their association is crucial not only for appearance, but also for typical sensory characteristics of Camembert cheese [1,2]. Previous studies considered aroma production in pure

* Correspondence:

department of Food Sciences and Nutrition, Institute of Nutrition and FunctionalFoods (INAF), STELA Dairy Research Centre, Université Laval, 2425 rue de l'Agriculture, G1V 0A6 Québec City, QC, Canada Fulllist of author information is available at the end of the article

Bio Med Central

culture, on culture media or on model cheese medium [3], biochemical pathways potentially involved in the development of sensory properties and even microbiological succession during Camembert ripening [4-10]. Surprisingly, only limited genetic information is available for these Fungi, since fewer than 30 different genes of each organism have been deposited in public databases.

Molecular biology techniques were recently used to evaluate several aspects of sensory characteristics of cheese. For example, multispecies DNA microarrays combined with biochemical analysis (HPLC and SPME-GCMS) has been a useful tool to evaluate L-methionine catabo-lism, production of volatile sulfur compounds (VSC) and

© 2014 Lessard et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, anc reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated.

lactose/lactate consumption during yeast growth [11,12]. Even though microarrays provide information about gene expression under various conditions, their utility is limited to organisms for which genetic information is available [13]. Next-Generation sequencing (NGS) methods are now widely used for de novo- and re- sequencing of genomes, transcriptomes, epigenomes and metagenomes [14-19]. The first metagenomic analysis using 454 pyrosequenc-ing was performed on bacterial communities in mines [20] and since then, high quality information is available about ecosystems from soil [21,22], sea water [23,24], humans [25,26], and even cheese [27], most of them identifying microorganisms and establishing their phylogenetic relationships [28]. Genome and metagenome sequencing are powerful tools, but massive transcriptome sequencing using NGS provides a more dynamic and functional view of microbial activity under particular conditions by accumulating data on RNA and its expression profile.

Several studies used NGS technologies to compare the transcriptomic response of a single organism exposed to different conditions [29-35]. In multiple-organism environments, establishing the metatranscriptome reveals the activity of a community, but only rare and very recent papers selected this approach [36-39]. This study is the first comprehensive metatranscriptome analysis of the Camembert cheese complex fungal ripening ecosystem. Here, the fungal metatranscriptome was sequenced using a Roche 454 pyrosequencing NGS strategy, without prior knowledge of the Penicillium camemberti and Geo-trichum candidum genome sequences. The longer reads produced by the 454 instruments enabled the discovery and characterization of new genetic information for these Fungi and simultaneously established their activity profile. Many fungal activities were identified using this strategy, including the central metabolism and the response to environmental stresses and nutrient availability in the cheese matrix. This semi-quantitative gene expression profiling revealed the adaptation of G. candidum and P. camem-berti during the 77-day ripening period of a commercial Canadian Camembert-type cheese.

Results and discussion

Cheese characteristics and fungal growth

Commercial Camembert-type cheeses made from pasteurized milk were obtained from a processing plant located in Canada. Cheeses used in the present study developed no obvious defects during the ripening period and met the high quality criteria of the company who provided the cheeses for the characteristics of cheese texture, fat matter, salt and water content (confidential data, not presented). Also, the measured pH increase fit the normal alkalinisation of the rind over time observed for similar Canadian mold ripened cheeses (Figure 1) [40]. When fungal strains selected for this cheese were

quantified using a TaqMan-based qPCR method [5,41], G. candidum and P. camemberti had similar growth profiles with an active phase in the first 5 days of ripening. Their maximum cell density was 6.45 x 109 and 4.69 x 1010 gene copies/cm2, respectively, at the end of ripening (Figure 1).

Sequencing and assembly of the Camembert cheese transcriptome

Since only scarce genetic information is available for G. candidum and P. camemberti, the metatranscriptomic approach using massive parallel sequencing had the advantage of simultaneously identifying new genes and determining their expression profile during cheese ripening. A de novo assembly performed using all 1,019,060 reads generated 8,909 contigs (length > 99 nt, average length of 916 nt). After sorting data for a minimum contig length of 200 nt and a minimum of 6 assembled reads, 8,318 contigs were conserved in the original cheese database. Reads were mapped back to the de novo assembly to enable semi-quantitative analysis and quality control of the assembly. De novo assembly and mapping data were compared to remove artefacts, such as duplicated transcript models, resulting in the exclusion of 402/8,318 contigs. The assembly contigs were free of fungal rDNA and mt-rDNA contamination as revealed by local BLAST search. This high quality dataset of 7,916 contigs (average length of 988 nt; Table 1) represents the fungal metatranscriptome of the Canadian Camemberttype cheese selected and was called CamemBank01, henceforth compensating the absence of available sequenced genomes for ripening species Penicillium camemberti and Geotrichum candidum.

Identification and functional annotation of contigs found in CamemBank01

All 7,916 contigs were analyzed using the Blast2GO platform [42]. Because no genome of the yeast G. candidum and the mold P. camemberti are currently available in public databases, sequence analysis was performed with caution. Therefore, contigs were assigned according to their similarity to mold or yeast relatives if sequences had a >70% identity with known proteins in GenBank. Globally, 56,7% contigs originated from molds (M, n = 4,491 contigs) and 16,4% from yeasts (Y, n = 1,299 contigs). The other 26,9% was defined as of uncharacterized origin (U), either because the Blastx protein similarity was under 70% or because they had no significant homology. Over the 563,733 reads assembled, 275,586 reads (48.89%) were confidently assigned to molds and 105,017 reads (18.63%) to yeasts, while 183,130 reads are still unassigned. The average expression was 71 reads/contig, or 71 transcripts/ gene (Table 1). At each sampling time, the majority of expressed contigs originated from molds and the

Ripening time (days)

Figure 1 Evolution of pH and fungal growth during Camembert cheese ripening. The ripening culture was a mixture of (□) G. candidum LMA-1028 and (A) P. camemberti LMA- 1029. Each strain was quantified individually using a TaqMan real-time qPCR method [41], over 77 days of ripening. pH (x) measures were taken weekly untilday 50.

average proportions of M, Y and U transcripts were similar over time.

Information on the metabolic pathways active in the tran-scriptome library was obtained from the crossed-analysis of the Gene Ontology (GO) annotation, Kyoto Encyclopedia of Genes and Genomes (KEGG) ontology (KO), and functional classification of clusters of euKaryotic Orthologous Groups (KOG database) [42-44]. The KOG database delivered the most informative analysis, providing 10% more affiliation of transcripts to a category than GO and KEGG [45,46]. Genes belonging to KOG categories D (Cell cycle control and mitosis), M (Cell wall/membrane/envelope biogenesis), Z (Cytoskeleton) and B (Chromatin structure and dynamics) were expressed at least 10-fold less than genes belonging to other KOG categories (Figure 2).

Table 1 Sequencing statistics and expression data in CamemBank01

Total number of reads in CamemBank01

1,019,060 reads

Totalnumber of filtered contigs in CamemBank01 7916 contigs

Average length per contig 988 bp

Minimum length 202 bp

Maximum length 4994 bp

Average expression per contig 71 reads/contig

Minimum expression 6 reads/contig

Maximum expression 10,928 reads/contig Number of contigs with an expression

< 71 reads/contig 6644 contigs (84%)

> 71 reads/contig 1272 contigs (16%)

Overall metatranscriptomic expression shows that, aside from translation (KOG category J; Figure 2A) and energy metabolism (KOG category C; Figure 2B), yeast transcripts dominated the early stages of ripening (day 5 and 9), while mold contigs experienced higher levels of expression around day 15. These transcription data matched the active growth phase of G. candidum and P. camemberti, as quantified by qPCR (Figure 1) [5,41,47-49].

Central metabolism

According to KOG annotation, energy metabolism (KOG category C) was mainly expressed in the early stage of ripening (Figure 2). We identified numerous gene functions related to energy metabolism, including all enzymes in the glycolysis/gluconeogenesis, pentose phosphate (PP) pathways, tricarboxylic acid (TCA) cycle and oxidative phos-phorylation, for both yeasts and molds. P. camemberti and G. candidum are, therefore, aerobic microorganisms capable of complete pyruvate degradation to CO2 and ATP production through carbohydrate, lipid and protein breakdown. For both fungal species, we identified 111 different contigs related to oxidative phosphorylation, including all five major complexes (NADH dehydrogenase, fumarate reductase, cytochrome bc1, cytochrome c oxidase and ATP synthase). Actually, energy metabolism was the dominant biological process in CamemBank01 (31% of all reads). Moreover, key enzymes in the glyoxylate bypass, namely isocitrate lyase (ICL; EC, respectively 489 and 86 reads found for yeasts and molds) and malate synthase (MAS; EC, respectively 194 and 208 reads), were found in high numbers [50]. In

Figure 2 Functional classification in yeast and mold genes expressed during Camembert cheese ripening. Functional classification of clusters of euKaryotic Orthologous Groups (KOG database) in yeasts (black) and molds (grey) during Camembert cheese ripening. Scales were adjusted to fit categories with read numbers (A) generally over 1000 reads and (B) generally above 100 reads. Seven time points were taken (Day 5, 9, 15, 21, 35, 56 and 77), corresponding to key times in the ripening period. Read numbers were normalized to 100,000 reads/ripening day. The KOG categories presented belong to "Cellular processes and signaling" (D: cycle control, celldivision, chromosome partitioning; M: Cellwall/membrane/envelope biogenesis; O: Post-translationalmodifications, protein turnover, chaperone functions; T: Signaltransduction mechanisms; Y: Nuclear structures; Z: Cytoskeleton), "Information storage and processing" (B: Chromatin structures and dynamics; J: Translation); and "Metabolism" (C: Energy production and conversion; E: Amino acid transport and metabolism; G: Carbohydrate transport and metabolism; I: Lipid transport and metabolism).

CamemBank01, most transcripts coding for those two enzymes are present at day 9 (Figure 3). Therefore, P. camemberti and G. candidum seem to be able to grow in a two-carbon source environment (acetate, ethanol, fatty acids), when other more complex carbon sources are unavailable [51,52].

Lactose and lactate utilization in dairy Fungi

The presence of lactose and galactose influence microbial and fungal community development in the cheese matrix. Once |3-galactosidase (LAC4, EC hydrolyses lactose to form galactose and glucose, the latter is metabolized through the glycolysis, TCA cycle and PP pathways. Contigs related to lactose and galactose transport and utilization were expressed by molds only at the very beginning of cheese ripening (Figure 3), which is consistent with the negligible concentration of lactose in the rind after six days of ripening [5]. As expected, no evidence of lactose utilization was found in yeast contigs, confirming the well-known incapacity of G. candidum to assimilate lactose [53].

Lactate generated by lactic acid bacteria during cheese making is a major carbon source for surface fungal microflora in Camembert-type cheese. Its metabolism contributes to fungal growth and alkalinisation of the cheese surface [9]. For this purpose, a specific lactate transporter (JEN1) and two distinct lactate dehydrogenases, DLD1 (EC and CYB2 (EC [54,55], are essential. In CamemBank01, contigs coding for these enzymes were found for yeasts and molds (Figure 3). All non-fermentable

carbon sources, such as lactate, are metabolized into sugars through the gluconeogenesis pathway and then redirected into central metabolism. Phosphoenolpyruvate kinase (PEPCK, EC and fructose-1,6-bisphosphatase (FBP, EC are two essential enzymes in this pathway. For yeasts and molds FBP and PEPCK are mainly expressed at days 9 and 15. PEPCK is massively expressed in both yeasts and molds, especially in the latter where it is among the top 1% of the most expressed contigs in CamemBank01 (Figure 3). This finding is consistent with the early expression of lactate metabolism related contigs, as well as ICL and MAS enzyme expression profiles in the early ripening stage, because of the possible depletion of glucose and lactose (Figure 3). At this stage, lactate, caseins and milk lipids are the dominant remaining energy sources [5] which explains the high transcription rate of the gluco-neogenesis pathway. Considering its importance in fungal metabolism in relation to cheese production and its high expression in CamemBank01, the PEPCK transcript could be a useful biomarker to ensure the normal progression of the Camembert-type cheese ripening process.

Protein metabolism

Proteolytic activity of fungal ripening cultures was proposed to be a key contributor to cheese flavor but only limited information is available. Analysis using the MER-OPS peptidase database [56] ( identified 226 peptidases and five peptidase inhibitors in the CamemBank01 metatranscriptome. From this number,

Figure 3 Gene expression related to sugar and organic acid metabolism and transport. For each gene function, totalread number and relative expression during ripening are presented. On this heat map, relative expression is represented by a greyscale, between high (white) and low (black) expression levels.

89 (origin: 52 M; 19 Y; 18 U) were linked to the extracellular protein digestion category of the proteolysis activity. MEROPS analysis revealed that Metallopeptidase (MP) and Serine peptidases (SP) are the most abundant peptid-ase families expressed in yeasts and molds. Global expression profiles show that protease and peptidase transcripts are mainly detected in the first 21 days of the ripening period, supporting other findings indicating that proteoly-sis occurs mostly in the first two weeks of the ripening time [5,53,57,58].

In the cytoplasm, peptides and amino acids are catabo-lized by different enzymes that lead to the formation of aroma compounds [59-62]. Widely used ripening yeasts including Kluyveromyces, Debaryomyces, Yarrowia and Geotrichum are known for their volatile sulfur compound (VSC) biosynthesis through methionine degradation [63-66]. Most contigs involved in VSC production [11,12,67,68] were clustered in the KOG category E in CamemBank01 (Table 2, Figure 2). Methionine catabol-ism and the corresponding VSC production can occur in one (elimination pathway) or two steps (transamination or Ehrlich pathway) enzymatic reactions [69] (Figure 4).

Cystathionine y-lyase (CGL, EC and cystathionine P-lyase (CBL, EC (Figure 4) are two potential lyase candidates in the one-step generation of VSC through methionine catabolism [70]. In CamemBank01, cgl and cbl transcripts were found in molds, but only cgl transcripts were found in yeasts. The expression of both cgl and cbl was observed to be higher in yeasts throughout ripening (Figure 4). In G. candidum, cgl expression

is linked to cabbage and sulfur aroma development in smear cheeses through methanethiol (MTL) production [71-73]. At an expression level of 366 reads, cgl is among the top 5% of expressed contigs in CamemBank01 and is a good candidate for producing the cabbage and sulphur notes G. candidum is known for [73]. These data suggest that G. candidum could be more involved in aroma and ammonia production through methionine catabolism than P. camemberti, considering that these enzymes are also linked to ammonia and a-ketobutyric acid production in G. candidum [74].

Transamination of methionine leading to MTL formation can be initiated by aminotansferases (Figure 4) [75]. In dairy Fungi the proposed pathway includes branched-chain (BcAT) and aromatic aminotransferases (ArAT) essential for flavor formation in K. lactis, G. candidum and Yarrowia lipolytica [63,64,66,76]. The next step of transamination is responsible for ammonia generation and is catalyzed by the NAD-glutamate dehydrogenase enzyme (NAD-GDH, EC (Figure 4) [12,66]. In CamemBank01, BcAT, ArAT and gdh contigs were retrieved for yeasts and molds. The NAD-gdh contig was only found in yeasts (Figure 4). This observation confirms that G. candidum uses peptides and amino acids for energy metabolism and cellular growth, which contributes greatly to ammonia production and pH increase in cheese, while P. camemberti uses lactate [58,77-79]. According to the transcription data in Camem-Bank01, ammonia production and amino acid metabolism appear after the first week of ripening. Formation of a-keto-y-methylthio butyric acid (KMBA) and MTL through

Table 2 Functional annotation statistics and expression data of contigs in CamemBank01

KOG category

General function Metabolic pathway

Yeasts [Y]

Molds [M]

(Nb contigs/Nb reads)


Metabolism and transport


Carbohydrates (sugars and organic acids)

Amino acids


Cellular processes and signaling

Post-translationalmodifications Signaltransduction Information storage and processing

Transcription Translation

RNA processing and modification Poorly characterized

Generalfunction prediction only Unknown function or no annotation

122/9,155 58/4,141 102/10,139 43/2,239

122/5,678 55/5,952

40/3,933 190/37,614 35/1,250


182/14,014 152/9,778 183/12,707 131/4,974

147/9,153 168/4,859

102/4,126 240/41,001 136/3,787

402/23,419 1887/108,840

the Ehrlich pathway may need an enzyme called KMBA demethiolase. Such a gene was not found in Camem-Bank01 and suggests, as others have previously stated, that the conversion of KMBA in MTL could be spontaneous and non-enzymatic [80,81]. In light of these observations, CamemBank01 outlines the need and provides the ability to investigate these metabolic pathways in depth, and to correlate these data with biochemical analysis.

Lipid metabolism

Lipids have major roles in Camembert-type cheeses since they modulate the texture, act as the carrier for aroma compounds and are the major precursor for flavor compounds such as methylketones, lactones, esters and alcohols [2,62,82,83]. The lipid metabolism KOG category (I) is divided in two groups: fatty acid metabolism and cell wall-related lipid metabolism. Functional annotation of all contigs in CamemBank01 showed that fatty acid transport and metabolism counted for more than half of all of lipid metabolism (KOG I) contigs found in CamemBank01 (Table 2). Lipolysis pathways are expressed at the beginning of the ripening period; gene expression is limited at day 5 but increased at days 9 and 15 (Figure 2B). Seven transporters were also found, which had the same expression profile as all other lipid-related contigs.

Yeasts and molds that participate in the ripening of Camembert-type cheeses are known to possess lipases (EC that hydrolyse triglycerides into di- and mono-glycerides, free fatty acids (FA) and glycerol. Only a few lipase transcripts were found in CamemBank01. According to GO annotation, all three lipases found have triglyceride lipase activity and, for G. candidum, two

such enzymes were previously identified in the literature [84-88]. In both yeasts and molds, the contigs encoding lipase genes were expressed during the entire ripening period, but at a very low rate (under 71 reads/contig), which is consistent with the globally low expression of the lipolysis pathway genes compared to those of other metabolic pathways (Table 2).

Yeasts such as Saccharomyces and Candida appear to possess only the peroxisomal version of the P-oxidation pathway [89,90], while Aspergillus and Podospora possess both peroxisomal and mitochondrial pathways [91-93], consistent with CamemBank01 expression data. Camem-Bank01 expression data does not indicate the presence of a mitochondrial P-oxidation pathway in G. candidum but both pathways were identified in P. camemberti. Each cycle of P-oxidation produces one molecule of acetyl-coA that can be redirected into the TCA cycle to generate energy or transformed in ketone bodies (aroma precursors), and one molecule of acyl-coA that can go through other P-oxidation cycles (Figure 5). In Fungi, a peroxi-somal multifunctional enzyme (MFE) is also responsible for the P-oxidation of fatty acids [91,94]. This enzyme combines the two middle steps (EC and EC of the P-oxidation cycle (Figure 5). In Camem-Bank01, we found the four enzymatic functions, including the MFE. The MFE's expression profile is very different for yeasts and molds: in molds, it is expressed for most of the 2.5 month period of ripening, whereas in yeasts, it is clearly over-expressed at day 21 (Figure 5). Interestingly, 69% of all transcripts related to the P-oxidation cycle in yeast-related contigs coded for the MFE, suggesting that this enzyme could have a central biological role. In yeasts,

Figure 4 Volatile sulfur compound formation through methionine catabolism. Figure inspired from [67,68,76]. Expression data is represented by a greyscale, between high (white) and low (black) expression levels, for each enzymatic reaction. Legend: Y: Yeasts; M: Molds;

FA: fatty acids; MTL: methanethiol; MTA: S-methylthioacetate; KMBA: 2-keto-4-methylthio butyric acid; HMBA: 4-methylthio hydroxybutyric acid; a-hb: a-hydroxybutyrate; VSC: volatile sulfur compounds; DMDS: dimethyldisulfide; DMTS: dimethyl trisulfide; DMQS: dimethyltetrasulfide; PA: propionaldehyde. Solid lines: enzymatic reactions; double solid line: generalmetabolic pathways; simple dotted lines: potentialenzymatic reactions; double dotted lines: chemicalreactions.

AT: aminotransferase; a-kg: a-ketoglutarate; Glu: glutamate; gdh: glutamate deyhrogenase; a-kb: a-ketobutyrate; NH4: ammonia; AA: aminoacids

MFE is the second most highly expressed of all lipolysis-related contigs, after the acyl-coA synthase (ACS, EC (Figure 5). The acyl-coA synthase accounted for 44% of the total lipolysis-related transcripts. In molds, approximately 21% of transcripts coded for these two enzymes combined. From the perspective of finding potential biomarkers for Camembert-type cheese ripening, the multi-functional enzyme could be one of interest, given its expression over time in both microorganisms.

In the last degradation step of fatty acids, 3-ketoacyl-coA is redirected in the TCA cycle through a 3-ketoacyl-coA thiolase (KAT, EC activity [95]. The high expression level in molds (706 reads in molds compared to 32 in yeasts) at the very end of the ripening period suggests that fatty acids are late energy sources for molds and that this gene could be an interesting biomarker to follow this activity. Finally, some fatty acids are only partially P-

oxidized. Thioesterases, decarboxylases and reductases are then responsible for the potential production of methylke-tones and secondary alcohols, which are important aroma compounds in Camembert-type cheese [82]. During the ripening period of a Camembert-type cheese, fatty acids may be entirely degraded for energy production by P. camemberti and G. candidum. In fact, very few transcripts related to partial ^-oxidation were found only in molds in CamemBank01, (30 reads total for a thioesterase gene; Figure 5). However, these findings confirm the hypothesis that P. camemberti has a higher lipolytic potential than G. candidum and its gene expression should be investigated more extensively [2].


Overall, 7916 new contigs have been identified related to the metabolism of yeasts and molds that develop at the

■2 -1.9 -1 -0,5 0 0.S 1 15 2

Figure 5 Fatty acid metabolism and [3-oxidation with associated expression data. Figure adapted from [2,83]. Expression data is represented by a greyscale, between high (white) and low (black) expression levels, for each enzymatic reaction. Solid lines: enzymatic reactions; simple dotted lines: enzymatic reactions not found in CamemBank01.

surface of a commercial Canadian Camembert-type cheese, increasing our knowledge about fungal metabolism. Considering that this cheese ecosystem was composed of two fungal strains, these data suggest that the transcripts associated with yeasts and molds potentially reflect the activity of Geotrichum candidum and Pénicillium camemberti. CamemBank01 permitted us to simultaneously determine the sequence of a large part of the genetic information encoded by these two microorganisms and detail the expression of these putative genes. Since the previous genetic information available was mostly riboso-mal DNA, CamemBank01 provides a data mining resource for the dairy Fungi scientific community. Whole genome sequencing improves knowledge of the genetic structure of an organism [96], while the comparison between genome sequences allows understanding the evolutionary structure of populations [14,16,97,98]. We demonstrated that NGS approach for transcriptome analysis is a powerful tool for acquiring massive genetic information in a given biological condition. Therefore, CamemBank01 can now contribute

to the structural annotation of the genomic sequences of P. camemberti and G. candidum, when they will be available. Moreover, this new database has shown the genomic determinants responsible for the enzymatic and biochemical reactions occurring during soft cheese ripening, previously described by other authors. This metatranscriptome analysis helped to both demonstrate the presence and the expression of these genes in the cheese ripening process. Globally, for yeasts and molds, the same general functions (KOG categories C, G, E and I) seem to be participating in fungal metabolism during Camembert-type cheese ripening. These pathways are not only the most expressed in CamemBank01, but also the most relevant in terms of sensory properties. Selection and study of biological markers should be the next step in understanding the real contribution of individual fungal strains and the consortium. It is crucial to carry out a more in-depth study of their biochemical activity during cheese ripening, which will provide key information about their implication in the development of cheese flavor.


Cheese production and sampling

Commercial Camembert-type cheeses were provided by a producer of Canadian premium specialty cheeses. All cheeses were sampled from a regular production of an 870 g format pasteurized-milk Camembert from a high capacity cheesemaking facility for which the process is confidential. A commercial starter culture, containing thermophilic and mesophilic lactic acid bacteria, was used in combination with a ripening starter containing only P. camemberti LMA-1029 and G. candidum LMA-1028. Inoculation of the fungal strains provided an initial count of approximately 8 x 101 CFU of G. candidum LMA-1028 and 6 x 103 CFU of P. camemberti LMA-1029 per ml of milk. No other yeasts or molds were used as ripening agents to produce a cheese characterized by a mild proteolysis. Cheeses were ripened for the first 9 days at 13°C, 98% relative humidity, then wrapped and ripened at 4°C for up to 77 days. The total 77 day ripening period included the first 9 days prior to wrapping. Samples were analyzed at days 5, 9, 15, 21, 35, 56, and 77, which corresponds, chronologically to the appearance of the mycelium, through ripening, to the consumption period.

DNA extraction and quantification of fungi

For each sampling time, mycelium from a 50 cm2 area (25 cm2 of both flat sides of the cheese) was recovered from cheese triplicates, frozen in liquid nitrogen and ground using a mortar and pestle. DNA extraction was performed according to Al-Samarrai et al. [99] using 2025 mg of ground mycelium. Quantitative real-time PCR (qPCR) was performed as described in our previous work to detect and quantify two major fungal ripening cultures: Geotrichum candidum and Penicillium camemberti [41].

RNA extraction, quality assessment and cDNA synthesis

To reduce possible sampling bias during the metatran-scriptomic analysis, at each sampling time, the total RNA was extracted from three cheeses and each extraction was performed in triplicate. Total RNA was purified from 75 mg of frozen ground rind powder using the RNAqueous RNA isolation kit (Ambion) combined with the Plant RNA Isolation Aid solution (Ambion) in a 12:1 ratio, according to the manufacturer's instructions. The quality of total RNA was evaluated using the RNA 6000 Nano Chip Kit (Agilent Technologies) and an Agilent 2100 Bioanalyzer (Agilent Technologies). For each sampling day, the three RNA extraction replicates were pooled at equal concentrations (1 ^g/^L) and a 5 ^L aliquot was incubated at 37°C for 2 h and analyzed again using the Agilent 2100 Bioanalyzer to ensure that no degradation had occurred. Reverse transcription was carried out using 1 ^g of total RNA. cDNA was synthesized using the

SMARTer PCR cDNA synthesis kit (Clontech) according to manufacturer's instructions. Freshly synthesized cDNA samples were purified using the Wizard SV PCR Cleanup system (Promega) to remove residual nucleotides, enzymes and primers.

Metatranscriptomic library preparation and cDNA sequencing

A metatranscriptomic library was created for each sampling day. Each library originated from three cheeses, from which RNA was extracted in three replicates, resulting in a pool of nine samples per ripening day. cDNA was fragmented using a Rapid library nebulizer (Roche/454 Sequencing) to obtain 750 bp fragments. The seven libraries (ripening days 5, 9, 15, 21, 35, 56, and 77) were prepared using the GS FLX Titanium Rapid Library preparation kit (454 Life Sciences). Each library was tagged with a unique barcode to be traced for analysis. Libraries were clonally amplified on beads by emulsion PCR using the GS FLX Titanium LV emPCR kit (454 Life Sciences). Beads with amplified libraries were loaded onto GS FLX Titanium PicoTiter-Plate. Sequencing reactions were carried out using FLX Genome Sequencer (454 Life Sciences) with GS FLX Titanium reagents (454 Life Sciences). Data were initially processed using the GS Run processor software provided by 454 Life Sciences with default settings for image acquisition, base calling and quality estimation. Metatranscriptomic library synthesis and massive parallel sequencing was performed at Institut de Biologie Intégrative et des Systèmes (IBIS) at Université Laval (

Sequence assembly, mapping and quality assessment

A de novo assembly step was done using all 1,019,060 reads obtained from the seven time points and was named CamemBank01. This de novo assembly was performed using the gsAssembler module of Newbler (v2.5.3, 454 Life Sciences) with default parameters except for identity (95%) and overlapping length (40 nt). A trimming database was used to remove reverse transcription adapters (from SMARTer kit) from the sequencing reads prior to assembly. Newbler is an overlap-layout-consensus (OLC) assembler that merges short reads into non-redundant sequences without gaps (contigs) to obtain full transcript sequences. Data were manually filtered with the specific criteria of length (min. 200 nt) and read numbers were assigned to each contig (min. 6 assembled reads/contig), which reduced the number of contigs to 8,318. As an assembly validation step and to measure transcript numbers, we used the Newbler v2.5.3 gsMapper module (454 Life Sciences) to map individual sequencing reads back to the de novo database generated with gsAssem-bler, with an approach similar to what was done for the shrimp Pandalus latirostris transcriptome [100]. Default

parameters were used for the mapping process except for 97.5% identity with existing contigs in CamemBank01, over a minimum of 20 nt. To remove assembly artefacts, such as redundancy, only the contigs showing 85- to 117% variation between number mapped and number assembled reads percentage were retained as high quality con-tigs in CamemBank01. This resulted in the exclusion of 402 contigs out of 8,318 (4.85%).

Sequence identity and annotation

All 7,916 high quality contigs were submitted to automated Blastx annotation using Blast2GO software v2.5.0, with default parameters (e-value <0.0001) [42]. Subsequently, gene ontology (GO) was determined by using Blast2GO. Gene ontology terms corresponding to either one or all GO categories: biological processes (P), molecular functions (F) and cellular components (C), were assigned to each contig. This study focused on P and F categories because of their higher relevance in the description of fungal metabolism. Again, the annotation step was performed with default parameters except that the e-value parameter was set to <1e-6 to increase stringency. InterProScan analysis was performed with default parameters to find functional motifs, and then annotation refinement was performed with the Augment Annotation tool ANNEX in the Blast2GO software. Finally, Enzyme Code (EC) numbers were assigned.

A second annotation step was performed using a different database. The functional classification of clusters of euKaryotic Orthologous Groups (KOG database) [43] was preferred because it was globally more informative for CamemBank01. The NCBI KOG database containing 112,920 protein sequences from seven eukaryotic genomes was uploaded, and sequence comparison using Blastx against the database (e-value <0.0001) allowed the retrieval of KOG categories for each transcript. Data were sorted for each KOG group, at each day of ripening, for yeasts, molds and transcripts of uncharacterized origin. For this purpose, mapped reads were manually normalized to 100,000 reads per library.

Finally, nucleotide sequences of all 7,916 contigs were submitted to Kyoto Encyclopedia of Genes and Genomes (KEGG; [101] through KEGG Automated Annotation Server tool (KAAS; for further functional annotation [102]. Using a single-directional best hit (SBH) blast method, KAAS compares nucleotide sequences to KEGG GENES database, allowing KEGG Orthology (KO) identifiers to be attributed to most contigs. Each contig with a KO identifier could then be mapped on KEGG metabolic pathways using KEGG mapper (http:// Finally, we performed manual crossed-annotation using KOG, GO, KEGG databases and EC numbers.

Sequence accession numbers

The initial reads data reported here have been submitted to NCBI sequence read archive (SRA, http://www.ncbi.nlm. under accession number SRP030470. All contig sequences are available in the Transcriptome Shotgun Assembly Sequence database (TSA, http:// This TSA project has been deposited at DDBJ/EMBL/GenBank under the accession GAQB00000000. The version described in this paper is the first version, GAQB01000000.

Semi-quantitative gene expression profiling in yeasts and molds for the identification of biological markers of the camembert cheese ripening period

Raw mapping data was used to visualize gene expression profiles during the ripening period. The fold change in expression for each transcript (Rxi) at each ripening time (yi) was calculated using a Serial Analysis of Gene Expression (SAGE) approach [103,104], according to the following formula (Eq. 1)

Rxi = log2[(n2 + f )/(m + f )] (1)

+ log2[(ti-ni + f )/(t2-n2 + f )]

where n1 is the average number of mapped reads for contig xi, n2, is the number of mapped reads for contig xi at ripening day yi, ti is the total of the average number of reads (sum of all x), t2 is the total number of reads at day yi and f is the 0.5 correction factor.

Competing interests

The authors declare that they have no competing interests. Authors' contributions

MHL carried out allexperiments and data analysis, except for cDNA sequencing. CV previously optimized RNA extractions and quality assessment protocols. BB was responsible for the preparation and sequencing of cDNA libraries and also contributed to data analysis. DSG and SL conceived of the study, and participated in its design and coordination. MHL, SL, BB and DSG draft the manuscript. Allauthors read and approved the finalmanuscript.


This project was supported by the "Chaire de recherche CRSNG-industrie laitière en technologie et typicité fromagère" (Chair holder Dr. Denis Roy). We are thankfulto the CRSNG and the industrial partners supporting this grant: Agropur, Dairy farmers of Canada, Damafro, Novalait, Parmalat, Saputo, and Université Laval. The authors thank Jérôme Laroche (IBIS) for his insights and technicalhelp in bioinformatics analysis, and Dr. Barb Conway for her assistance in editing the manuscript.

Author details

department of Food Sciences and Nutrition, Institute of Nutrition and FunctionalFoods (INAF), STELA Dairy Research Centre, Université Laval, 2425 rue de l'Agriculture, G1V 0A6 Québec City, QC, Canada. 2Institut de Biologie Intégrative et des Systèmes (IBIS), Pavillon Charles-Eugène-Marchand, 1030, rue de la Médecine, Université Laval, G1V 0A6 Québec City, QC, Canada. 3Agriculture and Agri-Food Canada, Food Research and Development Centre, 3600 Casavant Blvd. West, Saint Hyacinthe J2S 8E3 QC, Canada.

Received: 12 December 2013 Accepted: 11 March 2014 Published: 26 March 2014


1. Marcellino N, Beuvier E, Grappin R, Gueguen M, Benson DR: Diversity of Geotrichum candidum strains isolated from traditional cheesemaking fabrications in France. Appl Environ Microbiol 2001, 67(10):4752-4759.

2. Spinnler HE, Gripon JC: Surface mould-ripened cheeses. In Cheese: Chemistry, Physics and Microbiology. 2nd edition. Edited by Fox PF, McSweeney PL, Cogan TM, Guinee TP. Oxford, UK: Academic Press; 2004:157-174.

3. Berger C, Khan JA, Molimard P, Martin N, Spinnler HE: Production of sulfur flavors by ten strains of Geotrichum candidum. Appl Environ Microbiol 1999, 65(12):5510-5514.

4. Molimard P, Lesschaeve I, Issanchou S, Brousse M, Spinnler HE: Effect of the association of surface flora on the sensory properties of mould-ripened cheese. Lait 1997, 77(1):181 -187.

5. Leclercq-Perlat MN, Buono F, Lambert D, Latrille E, Spinnler HE, Corrieu G: Controlled production of Camembert-type cheeses. Part I: Microbiological and physicochemical evolutions. J Dairy Res 2004, 71(3):346-354.

6. Leclercq-Perlat MN, Corrieu G, Spinnler HE: Controlled production of Camembert-type cheeses. Part III: Role of the ripening microflora on free fatty acid concentrations. J Dairy Res 2007, 74(2):218-225.

7. Leclercq-Perlat MN, Latrille E, Corrieu G, Spinnler HE: Controlled production of Camembert-type cheeses. Part II: Changes in the concentration of the more volatile compounds. J Dairy Res 2004, 71 (3):355-366.

8. Leclercq-Perlat MN, Oumer A, Bergere JL, Spinnler HE, Corrieu G: Behavior of Brevibacterium linens and Debaryomyces hansenii as ripening flora in controlled production of smear soft cheese from reconstituted milk: growth and substrate consumption dairy foods. J Dairy Sci 2000,83(8):1665-1673.

9. Leclercq-Perlat MN, Oumer A, Bergere JL, Spinnler HE, Corrieu G: Growth of Debaryomyces hansenii on a bacterial surface-ripened soft cheese. J Dairy Res 1999, 66(02):271-281.

10. Leclercq-Perlat MN, Picque D, Riahi H, Corrieu G: Microbiological and biochemical aspects of Camembert-type cheeses depend on atmospheric composition in the ripening chamber. J Dairy Sci 2006,89(8):3260-3273.

11. Cholet O, Henaut A, Casaregola S, Bonnarme P: Gene expression and biochemical analysis of cheese-ripening yeasts: focus on catabolism of L-methionine, lactate, and lactose. Appl Environ Microbiol 2007, 73(8):2561 -2570.

12. Cholet O, Henaut A, Hebert A, Bonnarme P: Transcriptional analysis of L-methionine catabolism in the cheese-ripening yeast Yarrowia lipolytica in relation to volatile sulfur compound biosynthesis. Appl Environ Microbiol 2008, 74(11):3356-3367.

13. Roh SW, Abell GC, Kim KH, Nam YD, Bae JW: Comparing microarrays and next-generation sequencing technologies for microbial ecology research. Trends Biotechnol 2010, 28(6):291 -299.

14. DiGuistini S, Liao NY, Platt D, Robertson G, Seidel M, Chan SK, Docking TR, Birol I, Holt RA, Hirst M, Mardis E, Marra MA, Hamelin RC, Bohlmann J, Breuil C, Jones SJ: De novo genome sequence assembly of a filamentous fungus using Sanger, 454 and Illumina sequence data. Genome Biol 2009,10(9):R94.

15. Losada L, Varga JJ, Hostetler J, Radune D, Kim M, Durkin S, Schneewind O, Nierman WC: Genome sequencing and analysis of Yersina pestis KIM D27, an avirulent strain exempt from select agent regulation. PLoS One 2011, 6(4):e19054.

16. Haridas S, Breuill C, Bohlmann J, Hsiang T: A biologist's guide to de novo genome assembly using next-generation sequence data: a test with fungal genomes. J Microbiol Methods 2011, 86(3):368-375.

17. Zhou X, Ren L, Li Y, Zhang M, Yu Y, Yu J: The next-generation sequencing technology: a technology review and future perspective. Sci China Life Sci 2010, 53(1):44-57.

18. Zhou X, Ren L, Meng Q, Li Y, Yu Y, Yu J: The next-generation sequencing technology and application. Protein Cell 2010,1 (6):520-536.

19. Nowrousian M, Stajich JE, Chu M, Engh I, Espagne E, Halliday K, Kamerewerd J, Kempken F, Knab B, Kuo HC, Osiewacz HD, Pöggeler S, Read ND, Seiler S, Smith KM, Zickler D, Kück U, Freitag M: De novo assembly of a 40 Mb eukaryotic genome from short sequence reads: Sordaria macrospora, a model organism for fungal morphogenesis. PLoS Genet 2010, 6(4):e1000891.

20. Edwards RA, Rodriguez-Brito B, Wegley L, Haynes M, Breitbart M, Peterson DM, Saar MO, Alexander S, Alexander EC Jr, Rohwer F: Using pyrosequencing to shed light on deep mine microbial ecology. BMC Genomics 2006, 7:57.

21. Buee M, Reich M, Murat C, Morin E, Nilsson RH, Uroz S, Martin F: 454 Pyrosequencing analyses of forest soils reveal an unexpectedly high fungal diversity. New Phytol 2009, 184(2):449-456.

22. Lim YW, Kim BK, Kim C, Jung HS, Kim BS, Lee JH, Chun J: Assessment of soil fungal communities using pyrosequencing. J Microbiol 2010,48(3):284-289.

23. Frias-Lopez J, Shi Y, Tyson GW, Coleman ML, Schuster SC, Chisholm SW, Delong EF: Microbial community gene expression in ocean surface waters. Proc Natl Acad Sci USA 2008, 105(10):3805-3810.

24. Burke C, Steinberg P, Rusch D, Kjelleberg S, Thomas T: Bacterial community assembly based on functional genes rather than species. Proc Natl Acad Sci USA 2011, 108(34):14288-14293.

25. Reyes A, Semenkovich NP, Whiteson K, Rohwer F, Gordon JI: Going viral: next-generation sequencing applied to phage populations in the human gut. Nat Rev Microbiol 2012, 10(9):607-617.

26. Kuczynski J, Lauber CL, Walters WA, Parfrey LW, Clemente JC, Gevers D, Knight R: Experimental and analytical tools for studying the human microbiome. Nat Rev Genet 2012, 13(1):47-58.

27. Alegria A, Szczesny P, Mayo B, Bardowski J, Kowalczyk M: Biodiversity in Oscypek, a traditional Polish cheese, determined by culture-dependent and -independent approaches. Appl Environ Microbiol 2012, 78(6):1890—1898.

28. Zaneveld JR, Parfrey LW, Van Treuren W, Lozupone C, Clemente JC, Knights D, Stombaugh J, Kuczynski J, Knight R: Combined phylogenetic and genomic approaches for the high-throughput study of microbial habitat adaptation. Trends Microbiol 2011,19(10):472-482.

29. Alagna F, D'Agostino N, Torchia L, Servili M, Rao R, Pietrella M, Giuliano G, Chiusano ML, Baldoni L, Perrotta G: Comparative 454 pyrosequencing of transcripts from two olive genotypes during fruit development. BMC Genomics 2009, 10:399.

30. Mizrachi E, Hefer CA, Ranik M, Joubert F, Myburg AA: De novo assembled expressed gene catalog of a fast-growing Eucalyptus tree produced by Illumina mRNA-Seq. BMC Genomics 2010,11:681.

31. Natarajan P, Parani M: De novo assembly and transcriptome analysis of five major tissues of Jatropha curcas L. using GS FLX titanium platform of 454 pyrosequencing. BMC Genomics 2011, 12:191.

32. Zhou Y, Gao F, Liu R, Feng J, Li H: De novo sequencing and analysis of root transcriptome using 454 pyrosequencing to discover putative genes associated with drought tolerance in Ammopiptanthus mongolicus. BMC Genomics 2012, 13:266.

33. Parchman T, Geist K, Grahnen J, Benkman C, Buerkle C: Transcriptome sequencing in an ecologically important tree species: assembly, annotation, and marker discovery. BMC Genomics 2010, 11:180.

34. Chen ZF, Matsumura K, Wang H, Arellano SM, Yan X, Alam I, Archer JA, Bajic VB, Qian PY: Toward an understanding of the molecular mechanisms of barnacle larval settlement: a comparative transcriptomic approach. PLoS One 2011, 6(7):e22913.

35. Di Guistini S, Wang Y, Liao NY, Taylor G, Tanguay P, Feau N, Henrissat B, Chan SK, Hesse-Orce U, Alamouti SM, Tsui CK, Docking RT, Levasseur A, Haridas S, Robertson G, Birol I, Holt RA, Marra MA, Hamelin RC, Hirst M, Jones SJ, Bohlmann J, Breuil C: Genome and transcriptome analyses of the mountain pine beetle-fungal symbiont Grosmannia clavigera,

a lodgepole pine pathogen. Proc Natl Acad Sci USA 2011, 108(6):2504-2509.

36. Urich T, Lanzen A, Qi J, Huson DH, Schleper C, Schuster SC: Simultaneous assessment of soil microbial community structure and function through analysis of the meta-transcriptome. PLoS One 2008, 3(6):e2527.

37. Gifford SM, Sharma S, Rinta-Kanto JM, Moran MA: Quantitative analysis of a deeply sequenced marine microbial metatranscriptome. ISME J 2011, 5(3):461 -472.

38. Gosalbes MJ, Durban A, Pignatelli M, Abellan JJ, Jimenez-Hernandez N, Perez-Cobas AE, Latorre A, Moya A: Metatranscriptomic approach to analyze the functional human gut microbiota. PLoS One 2011, 6(3):e17447.

39. Damon C, Lehembre F, Oger-Desfeux C, Luis P, Ranger J, Fraissinet-Tachet L, Marmeisse R: Metatranscriptomics reveals the diversity of genes expressed by eukaryotes in forest soils. PLoS One 2012, 7(1):e28967.

40. Champagne CP, Soulignac L, Marcotte M, Innocent J-P: Texture et évolution du pH de fromages de type Brie entreposés en atmosphère contrôlée. Lait 2003, 83(2):145-151.

41. Lessard MH, Belanger G, St-Gelais D, Labrie S: The composition of Camembert cheese ripening cultures modulates both mycelial growth and appearance. Appl Environ Microbiol 2012, 78(6):1813-1819.

42. Conesa A, Götz S, García-Gómez JM, Terol J, Talón M, Robles M: Blast2GO: a universal tool for annotation, visualization and analysis in functional genomics research. Bioinformatics 2005, 21 (18):3674-3676.

43. Tatusov R, Fedorova N, Jackson J, Jacobs A, Kiryutin B, Koonin E, Krylov D, Mazumder R, Mekhedov S, Nikolskaya A, Rao BS, Smirnov S, Sverdlov AV, Vasudevan S, Wolf YI, Yin JJ, Natale DA: The COG database: an updated version includes eukaryotes. BMC Bioinforma 2003, 4:41.

44. Mao X, Cai T, Olyarchuk JG, Wei L: Automated genome annotation and pathway identification using the KEGG Orthology (KO) as a controlled vocabulary. Bioinformatics 2005, 21 (19):3787-3793.

45. Chen F, Mackey AJ, Vermunt JK, Roos DS: Assessing performance of orthology detection strategies applied to eukaryotic genomes. PLoS One 2007, 2(4):e383.

46. Altenhoff AM, Dessimoz C: Phylogenetic and functional assessment of orthologs inference projects and methods. PLoS Comput Biol 2009, 5(1):e1000262.

47. Wouters JTM, Ayad EHE, Hugenholtz J, Smit G: Microbes from raw milk for fermented dairy products. Int Dairy J 2002, 12(2-3):91-109.

48. Larpin S, Mondoloni C, Goerges S, Vernoux JP, Gueguen M, Desmasures N: Geotrichum candidum dominates in yeast population dynamics in Livarot, a French red-smear cheese. FEMS Yeast Res 2006, 6(8):1243-1253.

49. Gente S, Larpin S, Cholet O, Gueguen M, Vernoux JP, Desmasures N: Development of primers for detecting dominant yeasts in smear-ripened cheeses. J Dairy Res 2007, 74(2):137-145.

50. Kornberg HL: The role and control of the glyoxylate cycle in Escherichia coli. Biochem J 1966, 99(1):1 -11.

51. Lorenz MC, Fink GR: The glyoxylate cycle is required for fungal virulence. Nature 2001, 412(6842):83-86.

52. Sandeman RA, Hynes MJ, Fincham JR, Connerton IF: Molecular organisation of the malate synthase genes of Aspergillus nidulans and Neurospora crassa. Mol Gen Genet 1991, 228(3):445-452.

53. Boutrou R, Gueguen M: Interests in Geotrichum candidum for cheese technology. Int J Food Microbiol 2005, 102(1):1 -20.

54. Lodi T, Alberti A, Guiard B, Ferrero I: Regulation of the Saccharomyces cerevisiae DLD1 gene encoding the mitochondrial protein D-lactate ferricytochrome c oxidoreductase by HAP1 and HAP2/3/4/5. Mol Gen Genet 1999, 262(4-5):623-632.

55. Casal M, Paiva S, Andrade RP, Gancedo C, Leao C: The lactate-proton symport of Saccharomyces cerevisiae is encoded by JEN1. J Bacteriol 1999, 181 (8):2620-2623.

56. Rawlings ND, Waller M, Barrett AJ, Bateman A: MEROPS: the database of proteolytic enzymes, their substrates and inhibitors. Nucleic Acids Res 2014, 42(D1):D503-D509.

57. Engel E, Nicklaus S, Septier C, Salles C, Le Quere JL: Evolution of the taste of a bitter Camembert cheese during ripening: characterization of a matrix effect. J Agric Food Chem 2001, 49(6):2930-2939.

58. Boutrou R, Aziza M, Amrane A: Enhanced proteolytic activities of Geotrichum candidum and Penicillium camemberti in mixed culture. Enzyme Microb Technol 2006, 39(2):325-331.

59. Kubickova J, Grosch W: Evaluation of potent odorants of camembert cheese by dilution and concentration techniques. Int Dairy J 1997, 7:65-70.

60. McSweeney PLH, Sousa MJ: Biochemical pathways for the production of flavor compounds in cheese during ripening: a review. Lait 2000, 80:293-324.

61. Weimer B, Seefeldt K, Dias B: Sulfur metabolism in bacteria associated with cheese. Antonie Van Leeuwenhoek 1999, 76(1 -4):247-261.

62. Sable S, Cottenceau G: Current knowledge of soft cheeses flavor and related compounds. J Agric Food Chem 1999, 47(12):4825-4836.

63. Bondar DC, Beckerich JM, Bonnarme P: Involvement of a branched-chain aminotransferase in production of volatile sulfur compounds in Yarrowia lipolytica. Appl Environ Microbiol 2005, 71 (8):4585-4591.

64. Arfi K, Landaud S, Bonnarme P: Evidence for distinct L-methionine catabolic pathways in the yeast Geotrichum candidum and the bacterium Brevibacterium linens. Appl Environ Microbiol 2006, 72(3):2155-2162.

65. Arfi K, Tache R, Spinnler HE, Bonnarme P: Dual influence of the carbon source and L-methionine on the synthesis of sulphur compounds in the cheese-ripening yeast Geotrichum candidum. Appl Microbiol Biotechnol 2003, 61 (4):359-365.

66. Kagkli DM, Bonnarme P, Neuveglise C, Cogan TM, Casaregola S: L-methionine degradation pathway in Kluyveromyces lactis: identification and functional analysis of the genes encoding L-methionine aminotransferase. Appl Environ Microbiol 2006, 72(5):3330-3335.

67. Landaud S, Helinck S, Bonnarme P: Formation of volatile sulfur compounds and metabolism of methionine and other sulfur compounds in fermented food. Appl Microbiol Biotechnol 2008, 77(6):1191 -1205.

68. Lopez Del Castillo-Lozano M, Delile A, Spinnler HE, Bonnarme P, Landaud S: Comparison of volatil sulphur compound production by cheese-ripening yeasts from methionine and methionine-cysteine mixtures. Appl Microbiol Biotechnol 2007, 75:1447-1454.

69. Martin N, Berger C, Le Du C, Spinnler HE: Aroma compound production in cheese curd by coculturing with selected yeast and bacteria. J Dairy Sci 2001, 84(10):2125-2135.

70. Demarigny Y, Berger C, Desmasures N, Gueguen M, Spinnler HE: Flavour sulphides are produced from methionine by two different pathways by Geotrichum candidum. J Dairy Res 2000, 67(3):371 -380.

71. Bruinenberg PG, De Roo G, Limsowtin G: Purification and characterization of cystathionine (gamma)-lyase from Lactococcus lactis subsp. cremoris SK11: possible role in flavor compound formation during cheese maturation. Appl Environ Microbiol 1997, 63(2):561 -566.

72. De Angelis M, Curtin AC, McSweeney PL, Faccia M, Gobbetti M: Lactobacillus reuteri DSM 20016: purification and characterization of a cystathionine gamma-lyase and use as adjunct starter in cheesemaking. J Dairy Res 2002, 69(2):255-267.

73. Gente S, La Carbona S, Gueguen M: Levels of cystathionine gamma lyase production by Geotrichum candidum in synthetic media and correlation with the presence of sulphur flavours in cheese. Int J Food Microbiol 2007, 114(2):136-142.

74. Yvon M, Rijnen L: Cheese flavor formation by amino acids catabolism. Int

Daily J 2001, 11:185-201.

75. Bonnarme P, Arfi K, Dury C, Helinck S, Yvon M, Spinnler HE: Sulfur compound production by Geotrichum candidum from L-methionine: importance of the transamination step. FEMS Microbiol Lett 2001, 205(2)247-252.

76. Bonnarme P, Lapadatescu C, Yvon M, Spinnler HE: L-methionine degradation potentialities of cheese-ripening microorganisms. J Dairy Res 2001, 68(4):663-674.

77. Adour L, Couriol C, Amrane A: The effect of lactate addition on the growth of Penicillium camemberti on glutamate. J Biotechnol 2004, 114(3):307-314.

78. Brennan NM, Cogan TM, Loessner M, Scherer S: Bacterial surface-ripened cheeses. In Cheese: Chemistry, Physics and Microbiology. 2nd edition. Edited by Fox PF, McSweeney PL, Cogan TM, Guinee TP. Oxford, UK: Academic Press; 2004:199-225.

79. Aziza M, Adour L, Amrane A: Assimilation of peptides and amino acids and dissimilation of lactate during submerged pure cultures of Penicillium camemberti and Geotrichum candidum. J Microbiol Biotechnol 2008, 18(1):124-127.

80. Perpete P, Duthoit O, De Maeyer S, Imray L, Lawton AI, Stavropoulos KE, Gitonga VW, Hewlins MJ, Dickinson JR: Methionine catabolism in Saccharomyces cerevisiae. FEMS Yeast Res 2006, 6(1):48-56.

81. Hazelwood LA, Daran JM, van Maris AJ, Pronk JT, Dickinson JR: The Ehrlich pathway for fusel alcohol production: a century of research on Saccharomyces cerevisiae metabolism. Appl Environ Microbiol 2008, 74(8):2259-2266.

82. Molimard P, Spinnler HE: Review: compounds involved in the flavor of surface mold-ripened cheeses: origins and properties. J Dairy Sci 1996, 79(2):169-184.

83. Kinsella JE, Hwang DH, Dwivedi B: Enzymes of Penicillium roqueforti involved in the biosynthesis of cheese flavor. CRC Cr Rev Food Sci 1976, 8(2):191 -228.

84. Shimada Y, Sugihara A, Iizumi T, Tominaga Y: cDNA cloning and characterization of Geotrichum candidum lipase II. J Biochem 1990,107(5):703-707.

85. Shimada Y, Sugihara A, Tominaga Y, Iizumi T, Tsunasawa S: cDNA molecular cloning of Geotrichum candidum lipase. J Biochem 1989, 106(3):383-388.

86. Sugihara A, Hata S, Shimada Y, Goto K, Tsunasawa S, Tominaga Y: Characterization of Geotrichum candidum lipase III with some preference for the inside ester bond of triglyceride. Appl Microbiol Biotechnol 1993, 40(2-3):279-283.

87. Sugihara A, Shimada Y, Nakamura M, Nagao T, Tominaga Y: Positional and fatty acid specificities of Geotrichum candidum lipases. Protein Eng 1994, 7(4):585-588.

88. Sugihara A, Shimada Y, Tominaga Y: Separation and characterization of two molecular forms of Geotrichum candidum lipase. J Biochem 1990, 107(3):426-430.

89. Kunau WH, Buhne S, de la Garza M, Kionka C, Mateblowski M, Schultz-Borchard U, Thieringer R: Comparative enzymology of beta-oxidation. Biochem Soc Trans 1988,16(3):418-420.

101. 102.

Smith JJ, Brown TW, Eitzen GA, Rachubinski RA: Regulation of peroxisome size and number by fatty acid beta -oxidation in the yeast Yarrowia lipolytica. J Biol Chem 2000, 275(26):20168-20178. Maggio-Hall LA, Keller NP: Mitochondrial beta-oxidation in Aspergillus nidulans. Mol Microbiol 2004, 54(5):1173-1185.

Hynes MJ, Murray SL, Khew GS, Davis MA: Genetic analysis of the role of peroxisomes in the utilization of acetate and fatty acids in Aspergillus nidulans. Genetics 2008, 178(3):1355-1369.

Boisnard S, Espagne E, Zickler D, Bourdais A, Riquet AL, Berteaux-Lecellier V: Peroxisomal ABC transporters and beta-oxidation during the life cycle of the filamentous fungus Podospora anserina. Fungal Genet Biol 2009, 46(1):55-66.

Trotter PJ: The genetics of fatty acid metabolism in Saccharomyces cerevisiae. Annu Rev Nutr 2001, 21:97-119.

Carrie C, Murcha MW, Millar AH, Smith SM, Whelan J: Nine 3-ketoacyl-CoA thiolases (KATs) and acetoacetyl-CoA thiolases (ACATs) encoded by five genes in Arabidopsis thaliana are targeted either to peroxisomes or cytosol but not to mitochondria. Plant Mol Biol 2007, 63(1):97-108.

Wheeler DA, Srinivasan M, Egholm M, Shen Y, Chen L, McGuire A, He W, Chen YJ, Makhijani V, Roth GT, Gomes X, Tartaro K, Niazi F, Turcotte CL, Irzyk GP, Lupski JR, Chinault C, Song XZ, Liu Y, Yuan Y, Nazareth L, Qin X, Muzny DM, Margulies M, Weinstock GM, Gibbs RA, Rothberg JM: The complete genome of an individual by massively parallel DNA sequencing. Nature 2008, 452(7189):872-876.

Nowrousian M: Next-generation sequencing techniques for eukaryotic microorganisms: sequencing-based solutions to biological problems.

Eukaryot Cell 2010, 9(9):1300-1310.

Unterseher M, Jumpponen A, Opik M, Tedersoo L, Moora M, Dormann CF, Schnittler M: Species abundance distributions and richness estimations in fungal metagenomics-lessons learned from community ecology. Mol

Ecol 2011, 20(2):275-285.

Al-Samarrai TH, Schmid J: A simple method for extraction of fungal genomic DNA. Lett Appl Microbiol 2000, 30(1):53-56. Kawahara-Miki R, Wada K, Azuma N, Chiba S: Expression profiling without genome sequence information in a non-model species, Pandalid shrimp Pandalus latirostris, by Next-Generation sequencing. PLoS One 2011, 6(10):e26043.

Kanehisa M, Goto S: KEGG: Kyoto Encyclopedia of Genes and Genomes.

Nucleic Acids Res 2000, 28(1):27-30.

Moriya Y, Itoh M, Okuda S, Yoshizawa AC, Kanehisa M: KAAS: an automatic genome annotation and pathway reconstruction server. Nucleic Acids Res 2007, 35:W182-W185. Web Server issue.

Beissbarth T, Hyde L, Smyth GK, Job C, Boon WM, Tan SS, Scott HS, Speed TP: Statistical modeling of sequencing errors in SAGE libraries.

Bioinformatics 2004, 20(Suppl 1):i31-i39.

Old WM, Meyer-Arendt K, Aveline-Wolf L, Pierce KG, Mendoza A, Sevinsky JR, Resing KA, Ahn NG: Comparison of label-free methods for quantifying human proteins by shotgun proteomics. Mol Cell Proteomics 2005, 4(10):1487-1502.


Cite this article as: Lessard et al.: Metatranscriptome analysis of fungal strains Penicillium camemberti and Geotrichum candidum reveal cheese matrix breakdown and potential development of sensory properties of ripened Camembert-type cheese. BMC Genomics 2014 15:235.

Submit your next manuscript to BioMed Central and take full advantage of:

• Convenient online submission

• Thorough peer review

• No space constraints or color figure charges

• Immediate publication on acceptance

• Inclusion in PubMed, CAS, Scopus and Google Scholar

• Research which is freely available for redistribution

Submit your manuscript at

I Central