The PCR itself may be a source of bias in molecular studies of food samples. Differential orpreferential amplification of rDNA genes by PCR was reported by Reysenbach et al. (1992). It wasfound that preferential amplification might be caused by reannealing of the template DNA, whichcompromises the hybridisation of the primers (Suzuki and Giovannoni, 1996). Attempts to overcomethese problems have been carried out; solvents or other substances capable of enhancing thedenaturation of the DNA have been added to the PCR mixtures (Reysenbach et al., 1992; Varadarajand Skinner 1994). Because of preferential amplification, a mixture of bacterial DNA from a complex community may be only partially amplified by PCR, thereby leading to a product where some of the original members of the community are missing. Preferential amplification may represent a problem for the PCR-DGGE analysis of microbial communities from food. In fact, the number of species detected may not be real because of a lack of amplification by PCR of a specific DNA template. Therefore, the choice of the primer couple and the fragment to target is fundamental. It has been shown that, sometimes, targeting different 16S variable regions may lead to different results in species composition of the same sample (Ercolini et al.,2003a). In fact, the presence of Leuconostoc in Stilton cheese could be revealed only by analysing the region V4–V5 of the 16S rDNA while the species was not detected when the region V3 was targeted (Ercolini et al., 2003a). Do we need to target more variable regions in every analysis? This would make the analysis time-consuming. However, even though one region is targeted, other experiments should be done to ascertain whether other microbial species are present but not detected. Other problems are the formation of chimeric (Liesack et al., 1991; Kopczynski et al., 1994) or heteroduplex molecules (Ferris et al., 1997), whichcan affect the distribution of the bands in the DGGE profile. These points were already discussed in a previous review on the application of PCR-DGGE in microbial ecology (Muyzer and Smalla, 1998). The fragments to be resolved by DGGE cannot be longer than 500 bp. This represents a limiting factor for the sequence analysis and eventually probe design. It also makes it difficult, sometimes, to achieve a reliable identification of the microbial species because the sequences from DGGE bands to be compared in the databases are relatively small 16S rDNA fragments not always different within the same genus. Moreover, it is not always possible to resolve DGGE fragments when the difference in sequence is not that wide; this is strongly affected by the electrophoretic conditions and the amplicon-specific formation ofmelting domains. Furthermore, comigration of DNA fragments can be a problem for retrieving clean sequences from individual bands. In fact, even being different in sequences, the 16S rDNA fragments might have identical melting behaviour and therefore they cannot be separated in DGGE.Another problem in the study of community diversity on the basis of 16S rRNA genes using DGGE isthe presence of multiple copies of the 16S rDNA gene with sequence microheterogeneity. A single species with multiple rRNA copies can display, indeed, a DGGE profile characterised by multiple bands, which overestimates the community diversity detected by DGGE (Nu¨bel et al., 1996). For example, multiple bands are displayed in DGGE profiles of the 16S rDNA V1 region by some species of lactobacilli and staphylococci of food origin (Cocolin et al., 2001c); moreover, multiple bands were also found for species of staphylococci by DGGE analysis of the region V3 (Cocolin et al., 2001b; Blaiotta et al., 2003). Attempts have been carried out in order to establishthe detection limit of the PCR-DGGE. Indeed, it is worthwhile to define the concentration of the microbial species, which is needed in the food matrix to reveal a band in a DGGE fingerprint. However, for only one or a few species, dilution series have been applied in PCR-DGGE after DNA extraction. Detection limits have been indicated, ranging between 104 and 108 cfu ml 1 (Dewettinck et al., 2001; Ogier et al., 2002; Fasoli et al., 2003; Temmerman et al., 2003;Ercolini et al., unpublished). As a matter of fact, the detection limit depends on the species and perhaps even the strain considered. Moreover, the number and the concentration of the other members of the microbial community, along with the nature of the foodmatrix, all represent variables influencing the detection limit of DGGE by affecting both the efficiency of DNA extraction and the PCR amplification due to the possible competition among templates.
đang được dịch, vui lòng đợi..
