Critical Level of 13C Enrichment for the Successful Isolation of 13C Labeled DNA

In the stable-isotope probing (SIP) technology, the 13C-labelled substrate is normally used to incorporate 13C into nucleic acids. A proper abundance of 13C in substrate is critical to the success of SIP, because the 13C level not only determines whether 13C in nucleic acids is sufficient to be detective, but also affects the enrichment bias in the labeled microbes. However, such information is very rare. In this study, a serial of 13C-labelled glucose from 0 to 50 atom% 13C was used to incubate with Escherichia coli and then performed DNA-SIP. Our results showed that the detective level of 13C-DNA could be reduced to 2 atom% 13C of glucose (1.30 atom% 13C in DNA extract), while the ideal level was 10 atom% 13C glucose (2.25 atom% 13C in DNA). The critical level of 13C for the separation of 13C-DNA provides a new reference of DNA-SIP in order to trace active microbial populations utilizing specific C substrates in environments.


Introduction
Stable-isotope probing (SIP) of nucleic acids has become a focal method in microbial ecology since it can identify microorganisms being involved in the metabolism of specific substrates [1][2][3][4][5][6][7]. The 13 C-labelled substrate has been typically used in this technique to incorporate 13 C into deoxyribonucleic acid (DNA), ribonucleic acid (RNA), and phospholipid-derived fatty acid (PLFA) [5,7]. The isolation of the 13 C-labeled DNA or RNA can provide the solid information of nutrient cycling in relation to phylogeny and function of uncultivated microorganisms in the natural environment [8].
Two requirements have to be satisfied in a successful isolation of the 13 C-labeled DNA ( 13 C-DNA). The first one is that the labeled DNA should be distinguished above a background of abundant unlabeled molecules, which depends on the minimum substrate concentration and duration of the incubation. The other is the avoidance of too much substrate that may lead to cross-feeding of the substrate and enrichment bias [5,9]. Therefore, identifying the proper concentration of substrate is the key for an acceptable SIP experiment that can reflect the nature of microorganisms incorporating the substrate turnover.
However, various concentrations of 13 C labeled substrate were used in the SIP studies. For example, Radajewski et al. [8] stated that approximately 15-20 atom% 13 C was necessary for separating 13 C-DNA from 12 C-DNA and analyzing the link between identity and function of microorganisms, while around 30 atom% 13 C substrate was recommended for DNA-SIP in other studies [10,11]. Fan et al. [10] reported that the maize residue with 31 atom% 13 C was necessary to effectively identify the active microorganisms which decomposed the residues. In a study of commercial strains isolation, 32 atom% 13 C of biphenyl was used to successfully probe polychlorinated biphenyls (PCB)degrading populations in PCB-contaminated river sediment [11]. Although the protocol of the DNA-SIP technology per se has been well documented [4,12], the critical level of the 13 C enrichment in substrate for differentiating 13 C-labeled DNA has not been quantitatively tested. Narrowing such knowledge gap will be helpful to precisely assessing the metabolic "functional" genes on the specific substrates. In the present study, we incubated the Escherichia coli (E.coli) with a serial of 13 C enrichments in glucose to identify the critical level of 13 C enrichment in the substrate for the effective separation of 13 C-DNA. It is predicted that the minimum enrichment of 13 C labeled substrate for the

Incubation of the E.coli with 13 C-glucose
Six levels of 13 C enrichment in glucose were designed in this study. They were 0, 2, 5, 10, 20 and 50 atom % 13 C in glucose (Sigma-Aldrich) as substrate. Each level comprised 3 replications. Using M9 medium with 4% of glucose [13,14], 15 µl of E. coli solution (optical density at the 600 nm, OD600 = 0.8) was added to the medium and incubated at 37°C for 96 hrs on an end-over-end shaker. The OD600 values of the medium were determined at the beginning and the end of incubation with a spectrometer.

Extraction of E.coli DNA
DNA of E.coli was extracted from each treatment according to Casas and Rohwer [15]. DNA samples were visualized by electrophoresis in a 1% agarose gel. The concentration of DNA was measured using a NanoDrop 1000 (Thermo Scientific, Wilmington, DE) and the δ 13 C of the total E.coli DNA was determined by the Mat 253 isotope ratio mass spectrometer (Thermo Fisher, Germany). 13 C enrichments were expressed relative to Pee Dee Belemnite standard as either δ 13 C or atom fraction 13 C excess.

DNA-SIP
Cesium trifluoroacetate (CsTFA, Amersham Pharmacia Biotech) density gradient centrifugation was performed to separate the 13 C-labled DNA from total E.coli DNA [16]. Control gradients were run with the DNA from the unlabeled E.coli for each course to calibrate the centrifugation system. Approximately 7000 ng of each DNA sample was loaded into a centrifuge solution with a starting buoyant density of 1.60 g ml -1 . This centrifugation medium consisted of 3.2 ml of a 1.99 g ml -1 the CsTFA solution and 1.8 ml of gradient buffer (GB). The CsTFA solution with DNA was transferred to 4.9 ml Ultracrimp tubes (Beckman, USA) using 5 ml syringes. The tubes were centrifuged in a vertical rotor (Vti 65.2, Beckman) at 179,000 g (43,500 rpm) for 40 hrs at 20°C. The gradients were fractionated into 14 fractions by being displaced with water at a flow rate of 11.3 µl/s using a syringe pump (New Era Pump Systems, Inc. New York, USA). An AR200 digital refractometer (Reichert Inc., Depew, New York, USA) in the nD-TC mode was used to measure the buoyant density of gradient fractions. DNA in each fraction was precipitated with isopropanol (885 µl) and 1/10 volume (30 µl) of 3 M sodium acetate (pH 5.2). The DNA pellets were then washed and redissolved in ddH 2 O. The redissolved DNA was amplified using a universal primer pair for the 16S rDNA gene, i.e. 357f (5'-CCT ACG GGA GGC AGC AG -3') and 517r (5'-ATT ACC GCG GCT GCT GG -3') [17]. The cycling profile was 95°C for 10 min and 28 cycles of 95°C for 15 s, 60°C for 10 s and 72°C for 20 s, and then 72°C for 10 min. The PCR products from the 1 st to 10 th fractions were run on a 1% agarose gel.
Using SAS (SAS Institute, 1994), data were analyzed statistically. The difference of 13 C enrichment in the extract DNA between treatments was assessed with protected ANOVA tests at P< 0.05 level [20].
Since ultracentrifuge creates gradient fractions, the stability of buyout densities between treatments greatly affects the quality of the 13 C labeled DNA fractionation [12]. This stability was determined by calculating the coefficient of variability (CV) of buyout densities across treatments. The smaller the CV, the more consistent the fractionation of 13 C-DNA was between treatments. Figure 1: Optical densities at the 600 nm (OD600) for Escherichia coli that were grown in the M9 medium [1,18] for 96 hrs. The dash line indicates the initial OD600 value at the beginning of incubation.

E.coli growth supplied with 13 C substrate and DNA extraction
After 96 hrs of incubation, there was no significant difference in optical density (OD600) of M9 medium between treatments (Figure 1). Relatively pure DNA was extracted from the E.coli and the concentrations of DNA were in a range from 441 to 760 ng/ µl (Figure 2). The δ 13 C of DNA increased significantly with the 003 Agricultural Research & Technology: Open Access Journal increase of 13 C enrichment in glucose. The F values had the same trend (Table 1).   The buyout density (Table 2) ranged from 1.582 g ml -1 to 1.622 g ml -1 . The density in each fraction was consistent across treatments with small CV. This suggested that the performance of ultracentrifugation was consistent and successful between courses.

PCR products of the fractions
For the unlabeled control, the DNA band was visible from the 8 th to 10 th fraction ( Figure 3). However, appearance of DNA band started in the 7 th fraction under the 2 atom% 13 C treatment. The distinguishable DNA band moved towards smaller fractions with the increase of 13 C enrichment. At 50 atom% 13 C, DNA band was clearly observed in the 2 nd -10 th fraction.

Discussion
The insignificant OD600 values between treatments ( Figure  1) suggested that the growth of E.coli was not affected by the addition of glucose with different 13 C enrichments. Furthermore, the relatively high concentration of DNA and the clear bands of total DNA indicated that the extracted DNA (Figure 2) was appropriate for the DNA-SIP process.
In this study, the DNA band in the 7 th fraction appeared for the 2 atom% 13 C-glucose treatment, while non DNA band in the same fraction for the unlabeled control, revealing that the detectable level of 13 C substrate for 13 C-DNA-SIP under the pure culture condition can be reduced to 2 atom% 13 C in glucose. However, the better separation of 13 C-DNA was achieved in the ≥10 atom% 13 C treatment since the DNA band was visible in the 5 th , 6 th and 7 th fraction under that treatment.
The detectable level of 13 C substrate for 13 C-DNA-SIP in this study is lower than that in a study by Radajewski et al. [8], reporting that 20 atom% 13 C in substrate would be feasible to incorporate 13 C into DNA, and resolve 13 C-DNA from 12 C-DNA. The difference is probably because the carbon source of 13 CH 3 OH used in that study could only label the methylotrophic bacterium i.e. Methylobacterium extorquens and this bacterium was not dominant in the soil microbial communities, while the 13 C-labeled glucose was the sole C source in the pure incubation solution in this study. Thus, the threshold of 13 C enrichment in glucose for the separation of 13 C-DNA in this study was not as high as other studies. In addition, compared with 15 N substrate for 15 N-DNA-SIP, the minimum requirement of 13 C enrichment in substrate was much lower even under the similar pure culture condition [1]. The minimum level of a 15 N-labelled substrate, i.e. NH 4 NO 3 has been quantified as low as 40 atom% 15 N for the clear separation of 15 N-DNA of Pseudomonas putida [1]. The difference is attributed to the fact that the C concentration of DNA is theoretically in a range of 41.2 to 46.1%, while the N concentration of DNA varies only between 13.9 and 15.8% [1]. This is also reflected in the shifts in buoyant density in the CsCl gradients. For 13 C-labeled DNA this shift is approximate 0.036 g/ ml, while it is only 0.013-0.016 g/ml for 15 N-labeled DNA [21,22].
Nevertheless, the 13 C enrichment in the extracted DNA for the minimum requirement of DNA-SIP is probably more practicable than that in substrates when this technology are used in the environmental experiments, especially in the soil experiments, because a number of uncertain factors such as carbon-conversion efficiency and growth rate of the target organisms [5] may lower the convert efficiency of 13 C from substrate to DNA. A number of Figure 3: Aliquots of gradient fractions of DNA-SIP were run on 1% agarose gels and a 1-kb ladder is included as a marker. DNA samples were extracted from the M9 mediums of Escherichia coli supplied with 0, 2, 5, 10, 20 and 50 atom% 13 C in glucose.

Agricultural Research & Technology: Open Access Journal
studies have also stated that the requirement for supplemental nutrient addition for C assimilation of microbes and the amount of carbon incorporated into nucleic acid mostly depend on the targeted microorganisms and the characteristics of the samples being analyzed [4,12]. On the point of these views, the 13 C enrichment in the extracted DNA is likely to be more accurate for predicting the successful SIP. In this study, the detectable level of 13 C enrichment for DNA-SIP in the extracted DNA was 1.3 atom% 13 C (Table 1).
It is worth to note that 13 C-DNA bands were successively distinguishable from light to heavy fractions with the increase of 13 C enrichment in substrate (Figure 3). This indicates that 13 C gradually and proportionally incorporates into the C frame of the DNA molecular. However, in some DNA-SIP studies [5,23], the 13 C DNA bands were observed in the heavy fractions (the 4 th and 5 th fractions) rather than the fractions in between such as the 6 th and 7 th fractions. The difference may be attributed to the 13 C enrichment of substrate, incubation time and microbial specificity in utilizing 13 C substrates.

Conclusion
In this study, we confirmed that the level of 13 C enrichment in glucose for detectable 13 C-labeled DNA could be reduced to 2 atom% 13 C (1.30 atom% 13 C in DNA extract). The critical level of 13 C for the isolation of 13 C-DNA provides a new reference of DNA-SIP in order to trace active microbial communities utilizing specific C substrates in environments.