Introduction

Dominant microbial populations of microbial communities in wastewater bio-treatment systems are often emphasized because of their possible responsibilities for pollutants removal or sludge characters (Akarsubasi et al. 2006; Ivnitsky et al. 2007). It is generally accepted nowadays that the traditional culture-based methods are usually invalid in analysis of microbial communities because of unculturability of most microbes. The molecular biological technologies have developed some feasible culture-independent approaches at present to analyze complex microbial communities and monitor dominant microbial populations successfully (Sanz and Köchling 2007). These techniques are based on the RNA of the small ribosomal subunit (16S rRNA for prokaryotes) or their corresponding genes, due to their universality and abundance in all living beings as well as high conservation. Direct amplification of bacterial 16S rRNA genes from extracted total DNA and analysis amplified products by DGGE provide the most comprehensive and flexible means of sampling microbial communities. DGGE analysis will generate band patterns that directly reflect the genetic biodiversity of the sample and the bands generally corresponds to the dominant species. Coupled with sequencing and phylogenetic analysis of the bands, this method can give a good overview of the composition of a given microbial community (Lyautey et al. 2005). The main methodological difficulties of PCR-DGGE include reproducible and efficient extraction of DNA from environmental samples containing complex and diverse microbial communities, PCR bias, the co-migration of DNA from different species in the same band, the formation of multiple bands in the amplification of genes from single genomes, and so on (Dahllöf 2002). For example, Calábria de Araújo and Schneider (2008) demonstrated that PCR-DGGE was affected by differences in DNA extraction efficiency and 16S rDNA regions targeted by primer sets.

In spite of these defects, PCR-DGGE has been accepted as a feasible and available method to analyze complex microbial communities and identify dominant populations (Kozdrój and van Elsas 2001; Siddique et al. 2006; Kolehmainen et al. 2008). The bacterial diversity of microbial communities as well as dominant populations in different wastewater bio-treatment systems, including anaerobic treatment systems, have been investigated and determined by PCR-DGGE (Buzzini et al. 2006; Liu et al. 2007). The species or genus of dominant microbial populations, as well as their shifts attract most attention when compositions and structures of complex microbial communities were investigated (Rowan et al. 2003; Calli et al. 2005; Li et al. 2008; Mertoglu et al. 2008).

In fact, the sizes of dominant microbial populations are also significant for understanding their functions in complex microbial communities, especially in wastewater bio-treatment systems. At present, the dominant population sizes in complex microbial communities are still little known due to lack of suitable approaches probably. Although PCR-DGGE has been widely used to analyze microbial communities, its application to estimation of the dominant population sizes are still absent. Moreover, it is also still unclear that what level of microbial population size can be detected by PCR-DGGE as a dominant population. Except the defects of PCR-DGGE mentioned above, rRNA operon copy number is another key influencing factor for quantification of dominant population sizes analyzed by PCR-DGGE. The existence of multiple operons in one organism has been described for genes coding for the 16S rRNA and the number of 16S rRNA genes per cell could be between one (many organisms) and 15 (Clostridium paradoxum). Then PCR may over- or under-represent species with multiple copies (MacGregor 1999; Sharma et al. 2007). In addition, the efficiency of PCR amplification by universal primers between different phylotypes will also lead to quantification bias. The comparison between intra-phylotypes has been proven to be valid and reliable in the PCR-DGGE analysis (Ihalin and Asikainen 2006), but inter-phylotype comparisons are uncertain. Therefore it seems more difficult to estimate the dominant microbial population sizes in complex microbial communities by PCR-DGGE alone.

Then in this study, PCR-DGGE, coupled with an inoculated strain (Escherichia coli) with known population sizes as an internal standard, was used in an attempt to estimate dominant microbial population sizes in the anaerobic granular sludge treating streptomycin wastewater. The feasibility and limitation of this approach were investigated and discussed. The possible further improvement was also mentioned.

Materials and methods

The anaerobic granular sludge sample used in this study came from a full-scale UASB with effective volume of 200 m3 treating streptomycin wastewater in Shijiazhuang City, China, with treatment capacity of 125 m3/day and average volumetric COD loading of 3.7 kg/(m3/day). The average COD concentration and streptomycin residue in the influent were 7,100 mg/l and 130 mg/l, respectively. The anaerobic granular sludge sample was obtained from the bottom of the UASB, and then was centrifuged and rinsed before used in the following experiments.

E. coli DH 5α was cultured in LB medium for overnight at 37°C on a rotary shaker (140 rpm). After harvested by centrifugation, washed twice by a sterile phosphate buffer (pH 7.0) and re-suspended by sterile water, the strain suspension was obtained and the cell density was determined by counting colony formation units (CFU) grown on LB plates. Then the strain was inoculated into sludge samples (1.5 g wet weight) with population sizes of 106, 107, 108 and 109 CFU.

The total DNA of sludge samples were extracted immediately after inoculating the strain, according to the method from Zhang et al. (2008). Glass beads with diameter of 2–3 mm were added and the sample blended in a bead-beater for 3 min before rinsed twice by centrifugation with 5 ml TE buffer (100 mM Tris–HCl; 100 mM EDTA, pH 8.0). Lysozyme was mixed to a final concentration of 5 mg/ml and incubated for 30 min on a rotary shaker at 225 rpm and 37°C. Then, 1% SDS was added and incubated at 65°C for 3 h. After, proteinase K (10 mg/ml) 6 μl was added and incubated at 55°C for 1 h. Then, the suspension was incubated with a mixture of 10 M cetyltriethylammnonium bromide (CTAB) in 0.7 M NaCl at 65°C for 10 min. After centrifugation, the supernatant was extracted with chloroform. Total DNA was precipitated from the extracted aqueous with same volume iso-propanol. The DNA pellet was rinsed with 80% (v/v) ethanol, dried, and re-suspended in sterile double-distilled water and then stored at −20°C. The concentration of the extracted total DNA was determined by absorbance at 260 nm (A 260) and the purity was determined by A 260/A 280.

The total DNA preparation of each sample was amplified by PCR with a 2720 Thermal Cycler (Applied Biosystems, USA). The variable V3 region of 16S rDNA was amplified by PCR with primers to conserved regions of the 16S rRNA genes. The nucleotide sequences of the primers were as follows: primer BSF338 (5′-ACT CCT ACG GGA GGC AGC AG-3′) together with a GC clamp (5′-CGC CCG CCG CGC CCC GCG CCC GTC CCGCCG CCC CCG CCC G-3′) and primer BSR534 (5′-ATT ACC GCG GCT GCT GGC-3′) (Xing et al. 2008). The PCR mixture used contained 1 μl of DNA template (50 ng), 20 pM of each primer, 1 μl of each dNTP (10 mmol/l), 5 μl of 10 × PCR buffer (SBS Genetech, China), 5 U of Taq enzyme (SBS Genetech, China) and sterile water to a final volume of 50 μl. PCR temperature programme began with 3 min of activation of the polymerase at 94°C. Twenty PCR cycles were then conducted, with the first cycle consisting of 1 min denaturation at 94°C, 0.5 min annealing at 65°C and 0.75 min synthesis at 72°C. While the temperatures for denaturation and synthesis remained the same, the annealing temperature was subsequently decreased by 0.5°C after each cycle until it reached 55°C. Then another fifteen PCR cycles were conducted, with each cycle consisting of 1 min denaturation at 94°C, 0.5 min annealing at 55°C and 0.75 min synthesis at 72°C. Finally, after the last 15 PCR cycles, an 8 min extension step at 72°C was performed.

The PCR-amplified DNA products were separated by DGGE on 8% polyacrylamide gels with a linear gradient of 35–60% denaturant (100% denaturant = 40% [vol/vol] formamide plus 42% [wt/vol] urea) using the DCode™ Universal Mutation Detection System (Bio-Rad, Hercules, CA, USA). Gels were run for 12.5 h at 60 V in 1× TAE buffer maintained at 60°C. Gels were then stained with ethidium bromide (0.5 mg/l) for 20 min and visualised on a UV illuminator at 254 nm. The DGGE image was then acquired using the ChemiDoc (Bio-Rad) gel documentation system. The scanned gels containing DNA band profiles were analyzed to determine the intensity of each band using Quantity One analysis software (Bio-Rad). These bands revealed the dominant microbial populations and the intensity of each band implied its relative abundance. So the dominant microbial population sizes were estimated according to the intensities of these bands, compared with the band of the inoculated strain.

The Shannon index (H′), Shannon evenness (E), Margalef index (D Mg) and Berger-Parker dominance index (d) of the microbial community in the sludge sample were calculated based on band numbers and band intensities of the DGGE profiles (Hill et al. 2003; Susumu and Makoto 2008).

Results and discussion

The DGGE profiles of sludge samples with the inoculated strain and the control (without the inoculated strain) were shown in Fig 1. The bands in the DGGE profiles reflected the dominant populations in the sludge sample and the inoculated strain. The Shannon index (H′), Shannon evenness (E), Margalef index (D Mg) and Berger-Parker dominance index (d) of the microbial community in the sludge sample were 2.3765, 0.9911, 1.0036 and 0.1198, respectively, excluding the inoculated strain. The band of the inoculated strain in the DGGE profiles was distinguishable, comparing the band profiles of sludge samples in the presence of the inoculated strain with the control (Fig 1). Moreover, the intensity of this band increased obviously along with an increasing of the inoculated strain population size, whereas other bands were invariant in the band profiles of sludge samples (Fig 1). So it was reasonable to deduce that this band reflected the inoculated strain.

Fig. 1
figure 1

DGGE band patterns of sludge samples in the presence of the inoculated strain and the control

Full size image

A definite positive correlation between the band intensities and the population sizes for the inoculated strain was shown in Fig 1. Furthermore, a good linear relationship between the band intensities and the logarithms of the population sizes was also obtained for the inoculated strain (Fig 2), implying the feasibility of detection and quantification of the inoculated strain by PCR-DGGE. Another similar attempt to determine the proportions of some similar strains in a pure culture mixture by PCR-DGGE was reported and demonstrated that the detection level of a strain was 5% in cell mixture corresponding to 1.5 × 106 cells (Ihalin and Asikainen 2006). Then it could be concluded that the detection and quantification of a certain known strain by PCR-DGGE were reliable even if in a complex microbial community. The inoculated strain with known population sizes could be considered as an internal standard in the microbial community of the sludge sample. So the indigenous dominant microbial population sizes could be estimated by means of comparing their band intensities with the inoculated strain and referring to the linear relationship. Therefore PCR-DGGE coupled with inoculated strain as an internal standard provided a possible approach to estimating the dominant microbial population sizes in the sludge sample.

Fig. 2
figure 2

The relationship between band intensity and population size for the inoculated strain

Full size image

The intensities of all detectable bands in the DGGE profiles were measured and the dominant microbial population sizes were calculated according to the established linear relationship of the inoculated strain (Fig 3). The results indicated that the dominant microbial populations in the sludge sample showed various size levels. The four major dominant microbial populations showed the size level of 107–108 CFU/g (if culturable). The three secondary dominant microbial populations showed the size level of 105–106 CFU/g. Other dominant microbial populations showed the size level of 103–104 CFU/g although the members of this group were largest (19 in this study). The results also implied that the microbial populations with the size level lower than 103 CFU/g in the sludge sample were undetectable by PCR-DGGE.

Fig. 3
figure 3

The estimated dominant microbial population sizes in the sludge sample

Full size image

However, the estimated results should be amended due to the difference in rRNA copy number between the inoculated strain and indigenous species. Many indigenous species have probably one or two copies, less than the inoculated strain of E. coli with seven copies. Then these indigenous dominant population sizes might be under-estimated. The differences in DNA extraction efficiency and PCR amplification efficiency between the inoculated strain of E. coli and indigenous species also resulted in estimation bias. One feasible improvement was to optimize the inoculated strains in order to construct comprehensive and modified comparison standard. For example, if several strains with similar species to indigenous dominant populations and with different known rRNA copy numbers were selected and inoculated together as internal standards, the influences of rRNA copy number and PCR amplification efficiency could be evaluated. Then the comparison could be modified and become more specific. As a result, the estimated sizes of indigenous dominant populations could be more reliable.

In the previous studies, PCR-DGGE was considered to be suitable for detection of numerically important organisms but not for identification of the most abundant organisms (Calábria de Araújo and Schneider 2008). As a result, only several major dominant microbial populations were mentioned and other dominant microbial populations were always ignored in the analysis of PCR-DGGE (Pedro et al. 2003; Díaz-Ramírez et al. 2008; Li et al. 2008). Considering the size level of least dominant populations was 2–4 orders of magnitude lower than that of major dominant populations, it seemed probably reasonable to ignore them in general case.

Conclusions

This study attempted to estimate dominant microbial population sizes in the anaerobic granular sludge sample by PCR-DGGE, coupled with an inoculated strain with known population sizes as an internal standard. The inoculated strain in the sample could be detected and quantified well by PCR-DGGE. The DGGE band intensities of the inoculated strain showed a good linear relationship with its population sizes. Then it was possible to estimate indigenous dominant population sizes in the sludge sample by means of comparing their band intensities with the inoculated strain. The estimated sizes of major dominant microbial populations were at the level of 107–108 CFU/g and the microbial populations with the size level lower than 103 CFU/g were not detected by PCR-DGGE. Then a potential approach to estimation of dominant microbial population sizes in a complex microbial community was provide and some valuable information about dominant microbial populations in the sludge sample was obtained in this study.