This study provides the first Next-Generation Sequencing (NGS) survey of the bacterial community in Latin-style cheeses. The order Lactobacillales
was present in significant abundance in all Brand C replicates, which is expected since lactic acid bacteria are known for their role in the production of fermented foods including cheese (Table
). Renye et al. sampled queso fresco from Mexico, plated samples on selective agar, and subjected colonies to 16S rRNA sequencing
]. Lactococcus lactis
, of the order Lactobacillales
, was found in the highest numbers in both the cheeses made with raw milk and those made with pasteurized milk. Leuconostoc mesenteroides,
another member of the Lactobacillales order, was also abundant
The genus Exiguobacterium
of the order Bacillales
dominated all Brand B samples in this study; however, this genus has not been previously reported in cheese
]. Food matrices in which this genus has been identified include raw milk
], however, as well as potato processing effluent and water-boiled salted duck
have been identified in a wide variety of non-food matrices including surface and pond water, oral cancer tumors, hot springs in Yellowstone National Park, Siberian permafrost, coastal soil, and a saline Romanian lake
]. They have also been found to be useful in bioremediation efforts
Serum dextrose broth (SDB) was used in this study due to ongoing research efforts in our laboratory to enrich Brucella
species that might be associated with this type of soft cheese. However, SDB is not particularly selective and this rich nutrient source may have allowed uncommon bacteria to out-compete other components of the original metagenomic microflora. The Jameson Effect describes the phenomenon of low abundance microbial species ceasing growth in response to a dominant population’s arrival at stationary phase
]. Tran et al. explored microflora and pathogen dynamics by using selective broth and agar to isolate Listeria
from inoculated cheese. They found that ease of isolation was not correlated with concentration of inocula, which supports the theory that microbial community composition may play a bigger role in Listeria
inhibition than initial concentrations
]. Due to this potential effect of broth enrichment on the sample microflora, the selective agar employed in the next step in detection is all the more crucial and must be formulated taking into account the sample microflora after enrichment.
In this study, all replicates within each cheese brand clustered well, with the exception of Brand A_rep1 in Brand A. Perhaps bacterial DNA extraction was more efficient with this sample; however, there is not a clear reason for this discrepancy since all samples were processed identically and at the same time. Insufficient homogenization is also a possibility since enriched samples were not treated to stomaching between enrichment and aliquot collection. But if this were the case, it’s curious that other samples were not similarly affected.
While the three cheese brands used in this study were similar in style, color and texture, the bacterial abundance profiles of each were very different. The cheese manufacturers were contacted for information regarding manufacturing process to elucidate possible reasons for the observed differences (Table
). In the U.S., commercially available queso fresco is generally prepared with starter cultures; however, this may not be true for queso fresco made in other countries
]. Starter cultures are used in the manufacturing process for Brands A and B cheeses (use of starter culture to manufacture Brand C cheese could not be determined), although information about the specific cultures used could not be obtained. Other information obtained from Brands A and B included pH, % moisture, salt concentration, and % fat, but substantial differences were not noted between the two brands (Table
). Salt concentration was not available for Brand C cheese. Brand C does have the lowest pH (5.3 versus 6.2 - 6.7), however this alone may not account for the difference in microflora profiles between Brand C and the other brands. Further study would be required to discern the effect of these and similar parameters on the microflora of the cheese brands.
Manufacturer-provided parameters of Brands A, B, and C cheeses
The methods used in this study do not discern between live and dead cells because the amplification target, 16S ribosomal RNA-encoding genes, is highly conserved in bacteria regardless of viability. Efforts exist to manipulate sample preparation to detect only cells with intact membranes by sample treatment with propidium monoazide in combination with PCR amplification
] or the generation of transcriptomes. This will improve NGS as a tool for assessing microflora of cheese at different stages of the aging process. Additionally, Renye et al. found more variety in the types of bacteria isolated from cheeses made with raw milk versus those made with pasteurized milk
]; another public health risk best evaluated with tools that can distinguish between live and dead cells.
It is known that DNA extraction efficiency varies within and between laboratories, and that this can have an effect on subsequent microflora analysis
]. We addressed this in a variety of ways. First, the extraction kit used to perform the DNA extractions was chosen based on data collected in which the Qiagen DNeasy Blood and Tissue kit was compared to five other commercially-available kits for the extraction of Brucella neotomae
DNA from the same Latin-style cheeses used in this study (T. Lusk, E. Strain, and J.A. Kase, submitted for publication). The Qiagen DNeasy kit was found to produce the highest quality and quantity DNA from this matrix. All extractions were performed by a single person at one time. Lastly, four subsamples of each enriched cheese brand were extracted and sequenced, with all replicates producing similar bacterial profiles within each brand except for Brand A, in which 1 replicate showed more diversity than its counterparts.