Bias problems in culture-independent analysis of environmental bacterial communities: a representative study on hydrocarbonoclastic bacteria

Abstract

Culture-dependent methods for bacterial community analysis are currently considered obsolete; therefore, molecular techniques are usually used instead. The results of the current study on hydrocarbonoclastic bacteria in various oily habitats in Kuwait showed however, that the bacterial identities varied dramatically according to the analytical approach used. For six desert and six seawater samples used in this study, the culture-independent and culture-dependent techniques each led to a unique bacterial composition. Problems related to the culture-dependent technique are well known. The results of the current study highlighted bias problems other than those already recorded in the literature for the molecular approaches. Thus, for example, in contrast to the culture-dependent technique, the primers used in the molecular approach preferentially amplified the 16S rDNAs of hydrocarbonoclastic bacteria in total genomic DNAs of all the studied environmental samples, and in addition, failed to reveal in any environmental sample members of the Actinobacteria. The primers used in the molecular approach also amplified certain "pure" 16S rDNAs, but failed to do so when these DNAs were in mixture. In view of these results, it is recommended that the two analytical approaches should be used simultaneously because their combined results would reflect the bacterial community composition more precisely than either of them can do alone.

Figure 1

Kuwait map showing the coastal water (CW) and desert soil (DS) sampling sites.

Figure 2

DGGE of 16S rDNA amplicons in total DNA extracts from six pristine (P) and nearby six oil-polluted (O) seawater samples collected from the Arabian Gulf coast of Kuwait. (a) DGGE gel; in contrast to the cases of the “closed” desert samples (Figure 3), the DGGE bands of the “open” seawater samples were similar as far as the band numbers and patterns are concerned. The total numbers of the bands for all samples were not much higher than the total number of species recorded in each sample using culture-based analysis (Al-Awadhi et al. 2012). DNAs in individual bands were amplified, sequenced and the sequences were compared with those (cultured and uncultured) of the closest species in the GenBank (results in Table 1, which presents also information related to the sequencing). (b) Cluster analysis of DGGE-results using Euclidean distances.

Figure 3

DGGE of 16S rDNA amplicons in total DNA extracts from six pristine (P) and nearby six oil-polluted (O) soil samples collected from the Kuwaiti desert. (a) DGGE gel, the 16S rDNA band numbers and patterns varied, not only according to the sampling sites, but also to whether the sample was pristine or oil-polluted. The bands were further processed as described in the legend to Figure 1, and the results are summarized in Table 2. (b) Cluster analysis of DGGE-results using Euclidean distances.

Figure 4

Phylogenatic tree of 16S rRNA genes of bacteria from Kuwaiti oil polluted coastal water, as analyzed by the culture-independent method. Values shown in each node of the phylogenetic tree are bootstrap value; 2,000 bootstrap replicates were performed.

Figure 5

Phylogenatic tree of 16S rRNA genes of bacteria from Kuwaiti oil polluted desert soil, as analyzed by the culture-independent method. Values shown in each node of the phylogenetic tree are bootstrap value; 2,000 bootstrap replicates were performed.