Relative and Absolute Quantitative Real-Time PCR-Based Quantifications of hcnC and phlD Gene Transcripts in Natural Soil Spiked with Pseudomonas sp. Strain LBUM300

Abstract

Transcriptional analysis of microbial gene expression using relative quantitative real-time PCR (qRT-PCR) has been hampered by various technical problems. One such problem is the unavailability of an exogenous standard robust enough for use in a complex matrix like soil. To circumvent this technical issue, we made use of a recently developed artificial RNA (myIC) as an exogenous “spike-in” control. Nonsterile field soil was inoculated with various concentrations of the test bacterium Pseudomonas sp. strain LBUM300, ranging from 4.3- to 8.3-log bacterial cells per gram of soil. Total soil RNA was extracted at days 0, 7, and 14 postinoculation, and using two-step TaqMan assays, phlD (encoding the production of 2,4-diacetylphloroglucinol) and hcnC (encoding the production of hydrogen cyanide) gene expression was monitored. For relative quantification, a defined quantity of in vitro-synthesized myIC RNA was spiked during the RNA extraction procedure. Absolute qRT-PCR was also performed in parallel. Both the absolute and relative quantifications showed similar transcriptional trends. Overall, the transcriptional activity of phlD and hcnC changed over time and with respect to the bacterial concentrations used. Transcripts of the phlD and hcnC genes were detected for all five bacterial concentrations, but the phlD transcript copy numbers detected were lower than those detected for hcnC, regardless of the initial bacterial concentration or sampling date. For quantifying a low number of transcripts, the relative method was more reliable than the absolute method. This study demonstrates for the first time the use of a relative quantification approach to quantifying microbial gene transcripts from field soil using an exogenous spike-in control.

Performance of a robust transcriptomic analysis of rhizospheric microbial populations under soil conditions has been a desirable, yet elusive, goal of microbial ecologists for the past 2 decades. Such a measurement would provide important clues on the exact functional role and hence the contribution of target bacteria under defined experimental conditions (4). Even though traditional DNA methods, such as the analysis of 16S ribosomal DNA and functional genes, can help determine microbial diversity and identify particular microbial abilities (20, 33), they provide relatively little information on the functional and, hence, ecological significance of the target microorganism(s). One major stumbling block in our inability to realize the aforementioned goal has largely been attributed to the inherent complex composition of the extraction matrix itself, i.e., soil. Due to this, RNA extraction from soil, the first step in any gene expression studies, is plagued with inefficiencies at each individual step of the extraction process (19). In practical terms, these individual inefficiencies add up, leading to an erroneous estimation of the target, unless there is a method to quantify and correct the final results, which in the case of soils has not been developed to date (4).

In soil gene expression studies, the key to quantitatively study bacterial gene transcripts present in situ relies on one's ability to discern the target signal from the background activity. In the case of natural soils, this is extremely high (3, 4). Being highly sensitive and specific, the quantitative real-time PCR (qRT-PCR) technique, a combination of reverse transcription (RT) and PCR, is ideally suited for this purpose (11, 12, 30). The technique is highly specific and allows one to detect very low transcript levels (34). Gene expression data generated using qRT-PCR can be analyzed by two different approaches, absolute or relative quantification. In absolute quantification, the expression data are compared to an external calibration curve, usually generated from dilutions of the cloned amplicon under study (10). The major limitation of the absolute method is its inability to account for any procedure that may introduce inter- or intrasample variability (11). The relative quantification method, however, is able to address these limitations by making use of an internal calibrator (e.g., a constitutively expressed transcript commonly called a “housekeeping” gene [37]) or an external calibrator (e.g., exogenously spiked RNA [5, 8, 31]). Among them, at least in practice, the exogenously “spike-in” strategy appears to be the most convenient option for many researchers, since no a priori knowledge about housekeeping genes is required (39). While used in other systems, to the best of our knowledge, there are no reports to date on the development or use of a specific exogenous spike-in standard in soil systems. This could be explained by the difficulties in identifying or developing a RNA calibrator stable and inert enough that it does not cross-react with any other coextracted soil components. Notwithstanding these drawbacks, there has been some headway recently in our ability to quantify bacterial gene transcripts under soil conditions using qRT-PCR and the absolute quantification approach (10).

The recent development of a synthetic RNA internal amplification control (IAC), called myIC (15, 17), specifically designed for quantitative PCR (qPCR)/qRT-PCR, has opened up unique possibilities for implementing the spike-in strategy in complex experimental systems such as soil. The 200-bp myIC is a hypothetical sequence, and therefore, no homologue exits in the GenBank (NCBI) database to date (17). In light of this development, in the present work, we have sought to examine the possibility of using the myIC IAC RNA as an exogenous spike-in to relatively quantify bacterial gene transcripts in a nonsterile soil previously inoculated with a test bacterium. In parallel, we have also quantified the target transcripts using the established absolute quantification method and explored the effects of normalization to total amounts of extracted RNA.

The target bacterium used in the present study was the previously characterized biocontrol agent (BCA) Pseudomonas sp. strain LBUM300 (24), which harbors coordinately regulated operons for the biosynthesis of two antimicrobial metabolites, namely, 2,4-diacetylphloroglucinol (2,4-DAPG) and hydrogen cyanide (HCN). These two compounds have been linked to the suppression of many soilborne plant diseases (36). Therefore, there is interest in understanding the in situ spatiotemporal expression patterns of these two operons under soil conditions (14).

MATERIALS AND METHODS

Soil samples.

The soil used in this study was collected in July 2007 from experimental plots located at the Agriculture and Agri-Food Canada S. H. J. Michaud Research Farm (Bouctouche, NB, Canada). Field soil was characterized as a gleyed podzolic gray luvisol, a subgroup of the Canadian System of Soil Classification (1), with a pH of 5.2, 62% sand, 25% silt, 13% clay, and 2.6% organic matter. The soil was stored at 4°C until use.

Bacterial strain, growth conditions, and soil inoculation.

Pseudomonas sp. strain LBUM300 was previously isolated on King's B medium from strawberry rhizosphere samples (24). A preliminary study has determined that Pseudomonas sp. strain LBUM300 carries the operons for the production of HCN and 2,4-DAPG and expresses the biosynthetic genes involved in HCN and 2,4-DAPG production for at least up to 14 days on synthetic media (24). Bacteria were grown at 25°C for 24 h with shaking at 250 rpm in 10 ml of tryptic soy broth (TSB). After a 24-h growth period, bacteria were separated from growth medium by centrifugation at 4,000 × g (4°C) and resuspended in an equivalent volume of 0.85% saline solution. The bacterial suspension was quantified using a spectrophotometric standard curve (optical density [OD] = 600 nm). The following five bacterial dilutions were prepared with 0.85% saline solution: 1 × 10⁹, 1 × 10⁸, 1 × 10⁷, 1 × 10⁶, and 1 × 10⁵ bacteria/ml. A total of 20 g of soil was added to 50-ml tubes and inoculated with 4 ml of the respective bacterial dilution or saline solution (nonspiked control). The tubes were manually shaken for 30 s. For each bacterial dilution or nonspiked control, 9 replicate soil samples were prepared as described for a total of 54 samples and incubated in the dark at 25°C until sampling at 0, 7, and 14 days. At each sampling day, 3 replicate samples per bacterial dilution and nonspiked control were used for bacterial RNA extraction from soil (destructive sampling).

Bacterial RNA extraction from soil.

Total bacterial RNA was extracted from 2 g of soil using the PowerSoil total RNA isolation kit (MoBio Laboratories, Carlsbad, CA), according to the manufacturer's protocol. To facilitate relative quantification, 300 ng of in vitro-synthesized myIC RNA template was spiked as an exogenous control at the post-phenol extraction step. This spiking amount was empirically determined to give a positive signal between threshold cycle (C_T) values of 15 and 20 of the qPCR amplification. The post-phenol spiking strategy was developed because our attempts to spike the myIC RNA to a set of test soil samples at the lysis step did not yield a consistent qRT-PCR signal.

To eliminate coextracted genomic DNA, DNA digestion reactions were carried out using 80 μl of extracted RNA. DNase treatments were carried out according to the manufacturer's protocol, using 6 U of Turbo DNA-free enzyme (Ambion, Austin, TX), with the following modifications: incubation was done for 45 min at 37°C, followed by RNA isolation by phenol-chloroform-isoamyl alcohol at 25:24:1 (Fisher Scientific, Mississauga, Ontario, Canada). RNA was then precipitated by ethanol for 30 min at −20°C. The RNA pellet was dried and resuspended in 28 μl of SR7 solution (MoBio Laboratories). The total RNA quantity and quality were evaluated by spectrophotometry using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE).

Primer design.

PCR primers and TaqMan probes for PCR, RT, and qPCRs were designed using Primer Express software version 3.0 (Applied Biosystems, Foster City, CA) based on sequences from partial hcnC and phlD genes of Pseudomonas sp. strain LBUM300 and homologous sequences retrieved from the GenBank database (Table 1 ). The myIC qRT-PCR primers and probes were previously described by González-Escalona et al. (17) (Table 1). The PCR primers and Cy5-TaqMan probes were custom synthesized and obtained from Integrated DNA Technologies (Coralville, IA), while the 6-carboxyfluorescein (FAM)-TaqMan probes were custom synthesized and obtained from Applied Biosystems. The specificity of each PCR primer set was verified by performing BLASTn searches in the NCBI database (2) and also by visualizing PCR amplicons using conventional agarose gel electrophoresis.

TABLE 1.

Primers and probes used in this study

Target	Primer or probe	Sequence (5′→3′)a	Position on target	Product size (bp)	Reference (GenBank accession no.)
hcnC	hcnC-fwd	CCTGCCCCAGTCGTTCTTT	423 to 483	60	Present work (DQ788990)
	hcnC-rev	TGCAACTGCGGATACATTGC
	hcnC-FAM	FAM-ATTTCGCCTTGCAGTCC-MGBNFQ
phlD	phlD-fwd	CGGCGGACGGAAAATTC	620 to 677	58	Present work (DQ788986)
	phlD-rev	CCGACCGGGTTCCAAGTC
	phlD-FAM	FAM-TGATGAACTGGTCCTGCAA-MGBNFQ
myIC	dd-myIC-f	CTAACCTTCGTGATGAGCAATCG	25 to 170	145	17 (FJ357008)
	dd-myIC-r	GATCAGCTACGTGAGGTCCTAC
	dd-myIC-Cy5	Cy5-AGCTAGTCGATGCACTCCAGTCCTCCT-IowaBlackRQ-Sp

^aMGBNFQ, minor groove-binding nonfluorescent quencher.

Production of standards for qPCR.

Pseudomonas sp. strain LBUM300 was grown in liquid culture as previously described, and genomic DNA was extracted using the UltraClean microbial DNA isolation kit (MoBio). The partial phlD and hcnC gene sequences were amplified by PCR using the appropriate primers (Table 1). Each PCR mix contained 5 μl of 10× PCR buffer, 1.5 mM MgCl₂, 5 μl (5 μM) of each primer, 1 μl (10 mM) of deoxynucleoside triphosphates (dNTPs), 1.25 units of Taq DNA polymerase (New England Biolabs, Mississauga, Ontario, Canada), 31.75 μl of double-distilled water (ddH₂O), and 2 μl genomic DNA for a final volume of 50 μl. Cycling conditions used for each gene were the following: 94°C for 2 min, 40 cycles of 94°C for 1 min, 60°C for 1 min, and 72°C for 1 min, followed by a final extension at 72°C for 10 min on a PTC200 Peltier thermocycler (MJ Research, Waltham, MA). Amplicon size was verified by agarose gel electrophoresis and compared to a 50-bp DNA ladder (Invitrogen, Burlington, Ontario, Canada). PCR fragments were cloned using the TOPO TA cloning kit, according to the manufacturer's manual (Invitrogen). Plasmids were extracted using the DirectPrep 96 miniprep plasmid extraction kit (Qiagen, Mississauga, Ontario, Canada), and cloned fragments were verified by PCR and agarose gel electrophoresis. The plasmid copy number was quantified by spectrophotometry using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies), and calculation of the gene copy number was performed according to the molar mass derived from the plasmid and amplicon sequences (24). Dilutions of plasmidic DNA containing cloned partial phlD and hcnC gene sequences were used to generate standard curves in quantities ranging from 5 × 10² to 5 × 10⁹ copies for phlD and from 2.5 × 10² to 2.5 × 10⁹ copies for hcnC. Purified RNA not submitted to the RT reaction was also submitted to qPCR as described above to confirm total elimination of coextracted DNA. All qPCRs were replicated three times.

In vitro synthesis of IAC RNA.

The exogenous RNA internal amplification control (IAC) myIC was synthesized in vitro using the MEGAshortscript high-yield transcription kit (Ambion). A PCR fragment (336 bp) containing the T7 promoter was obtained after PCR amplification of the plasmid pDMD801 using M13 primers as described previously (15). The PCR mixture was cleaned with the QIAquick PCR purification kit (Qiagen), and the results were quantified using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies). For in vitro transcription, 100 ng of the clean PCR template was used per reaction and incubated for 12 h at 37°C to increase the final yield, as per the manufacturer's recommendation. The myIC IAC RNA transcript reaction mixtures were cleaned using the MEGAclear kit (Ambion) and treated twice with Turbo DNase I (3 U) to remove the DNA template (24). The RNA yield was quantified using a NanoDrop ND-1000 spectrophotometer (NanoDrop technologies) and stored in aliquots at either −20°C (short-term storage) or −80°C (long-term storage).

cDNA synthesis.

The reverse transcription reactions were performed by using the TaqMan reverse transcription kit (Applied Biosystems). Each RT reaction mix contained 7.9 μl of extracted RNA, 2.0 μl of 10× RT buffer, 4.4 μl (25 mM) of MgCl₂, 4.0 μl (2.5 mM) of dNTPs, 0.8 μl (5 μM) of reverse primer (Table 1), 8 units of RNase inhibitor, and 2.5 units of MultiScribe RT enzyme for a final volume of 20 μl. The cycling conditions used with a PTC200 Peltier thermocycler (MJ Research) were the following: 48°C for 30 min, followed by 95°C for 5 min.

qPCR assays used for absolute and relative quantifications of hcnC and phlD.

qPCR for hcnC, phlD, and myIC were performed separately by using the TaqMan PCR core reagent kit (Applied Biosystems). Each qPCR mix contained 6 μl of cDNA, 2.5 μl of 10× TaqMan buffer, 5.5 μl (25 mM) of MgCl₂, 0.5 μl (10 mM) of dATP, dCTP, and dGTP, 0.5 μl (20 mM) of dUTP, 1.0 μl (5 μM) of probe, 2.5 μl (5 μM) of each primer, 0.25 units of AmpErase uracil N-glycosylase (UNG), 0.625 units of AmpliTaq Gold DNA polymerase, and 2.625 μl of diethyl pyrocarbonate (DEPC)-treated water (Ambion). Cycling conditions used were the following: 50°C for 2 min, 95°C for 10 min, and then 50 cycles of 95°C for 15 s and 60°C for 1 min. Fluorescence was detected after each cycle.

Data and statistical analysis.

For absolute quantification, absolute transcript copy numbers for each gene and replicate were calculated with the ABI 7500 system SDS software version 1.4 (Applied Biosystems). For absolute quantification with normalization to total RNA, absolute transcript copy numbers were related to total RNA of the sample to obtain the absolute copy number of gene transcripts/ng total RNA. The relative quantification of the hcnC and phlD gene transcripts was standardized to the myIC IAC based on a mathematical model in which no calibration curve is required (25). Dilutions were generated from hcnC, phlD, and myIC RNA samples diluted 10-fold in order to calculate individual reaction efficiency [E = 10^(−1/slope)] as described previously (24). In order to follow the fold change in gene expression over time, the C_T value of the corresponding qRT-PCR amplification of the lowest bacterial seeded number of 1 × 10⁵ bacteria/ml was used. For each gene and sampling date, the effect of bacterial concentration was analyzed by factorial analysis of variance (ANOVA). For factorial ANOVA, a posteriori comparisons of the means between dilutions were done using Tukey's Studentized Range tests at a 5% level of significance. Correlation between gene expression trends was analyzed by multivariate analysis of covariance (MANCOVA). All statistical analyses were performed using the SAS statistical analysis software, version 9.1 (SAS Institute, Cary, NC).

RESULTS

qRT-PCR primer efficiency and linearity.

phlD and hcnC qRT-PCR amplification products generated single bands of the appropriate size when run on gel electrophoresis (data not shown). When amplifying the phlD DNA standards, a linear relation (R² = 0.990 to 0.999) was observed between the log copy number and real-time PCR threshold cycles. A similar linear relation was also observed for hcnC (R² = 0.998 to 0.999). Amplification efficiency ranged from 83% to 95% for phlD and from 86% to 88% for hcnC.

Before assessing the relative expression levels of target genes by qRT-PCR, the efficiency of each primer set was determined by establishing 10-fold dilution curves of RNA samples extracted from strain LBUM300 and in vitro-synthesized myIC RNA. Linear regression revealed that all RNA dilution curves were highly linear, with a correlation coefficient (R²) of ≥0.98. The qRT-PCR efficiency for each primer set was calculated with slopes and is as follows: phlD (82%, 1.822), hcnC (86.3%, 1.863), and myIC (84%, 1.847).

RNA extraction and DNase treatment.

The chosen method for soil RNA extraction produced clear RNA pellets free from any brownish coloring. The purity of RNA was confirmed by spectrophotometric readings, resulting in an average A₂₆₀/A₂₈₀ value of 1.94. RNA yields obtained from samples were, on average, 5.6 μg RNA/g of soil. qPCR amplification of DNase-treated RNA samples not submitted to RT did not cross threshold cycle values before a minimum of 50 cycles, indicating sufficient elimination of coextracted DNA in qRT-PCRs.

Evaluation of the myIC RNA internal control.

A previously designed exogenous RNA IAC, myIC, was synthesized in vitro using the plasmid pDMD801. This plasmid has a T7 promoter close to the myIC sequence which facilitates in vitro transcription. Typical RNA yields post-in vitro transcription ranged from 90 to 120 μg per reaction. qPCR and qRT-PCR amplifications using the dd-myIC-f and dd-myIC-r primers on total DNA/RNA extracted previously from the test soil did not generate any detectable signal or give any amplification product with gel electrophoresis (data not shown).

Relative quantification of phlD and hcnC gene expression in soil.

The relative expression of the phlD and hcnC gene transcripts was analyzed over three time points, ranging from 0 to 14 days postinoculation of the rhizospheric soil. The C_T of the target was normalized to that of the exogenous synthetic RNA myIC, which was spiked at the post-phenol step of RNA extraction. This allowed the determination of the total fold change for both the phlD and hcnC genes over the three sampling time points. The C_T value obtained with the spike of 1 × 10⁵ bacteria/ml was used as the control value in order to calculate the fold change in gene expression over various bacterial dilutions for each of the three sampling time points.

In general, transcripts of the phlD and hcnC genes for all five bacterial dilutions used were detected, but the phlD transcript copy numbers detected (Fig. 1) were lower than those detected for hcnC (Fig. 2), regardless of initial bacterial concentration or sampling date. Overall, the transcriptional activity of phlD from LBUM300 changed over time (P ≤ 0.001) and with respect to dilution (P ≤ 0.0001), and there was a significant interaction between combinations of time and dilution (P ≤ 0.0001). The transcriptional activity of the hcnC gene also followed a similar trend, as observed for phlD, and was also significantly altered over time (P ≤ 0.0001) and with respect to the dilution (P ≤ 0.0001) of the initial inoculated concentration of bacterium (Fig. 2). There was a significant interaction between combinations of time and dilution (P ≤ 0.0001).

FIG. 1.

Relative expression of phlD in Pseudomonas sp. strain LBUM300 under soil conditions at day 0 (A), day 7 (B), and day 14 (C) postinoculation. Bars show the standard errors of the means. The x axes indicate the number of bacterial cells inoculated at day 0 of the experiment. Values followed by a different letter were significantly different using Tukey's Studentized Range test (P < 0.05).

FIG. 2.

Relative expression of hcnC in Pseudomonas sp. strain LBUM300 under soil conditions at day 0 (A), day 7 (B), and day 14 (C) postinoculation. Bars show the standard errors of the means. The x axes indicate the number (no) of bacterial cells inoculated at day 0 of the experiment. Values followed by a different letter were significantly different using Tukey's Studentized Range test (P < 0.05).

Absolute quantification of phlD and hcnC gene expression in soil.

For each gene, sampling date, and initial bacterial concentration, gene transcripts were quantified into absolute copy numbers or absolute copy numbers normalized to that of total RNA. phlD and hcnC gene transcripts were detected for at least 14 days. In general, transcripts of the phlD and hcnC genes for all five bacterial dilutions used were detected, but the phlD transcript copy numbers detected were lower than those detected for hcnC, regardless of initial bacterial concentration or sampling date. No amplification for nonspiked soil controls was detected.

The initial bacterial concentration had a significant impact on gene transcripts detected in soil samples. The overall effect of the initial bacterial inoculum concentration on the hcnC and phlD transcript copy numbers was highly significant (P ≤ 0.0001) at 0, 7, and 14 days postinoculation (Fig. 3). Overall, the absolute quantification method (normalized or not normalized to total RNA) had no significant effect on transcript copy numbers, and thus, results were not affected by the absolute quantification method (data not shown). No significant interactions were found between initial bacterial concentra- tions and the absolute quantification method used. The MANCOVA model's partial correlation values revealed a significant correlation between gene trends (R² = 0.293; P = 0.02), indicating that both genes followed similar transcriptional activity with regard to the initial bacterial concentration and sampling date.

FIG. 3.

Absolute expression of hcnC and phlD in Pseudomonas sp. strain LBUM300 under soil conditions, expressed as copy number/gram of soil at day 0 (A), day 7 (B), and day 14 (C) postinoculation. Bars show the standard errors of the means. The x axes indicate the number of bacterial cells inoculated at day 0 of the experiment. Values followed by a different letter were significantly different using Tukey's Studentized Range test (P < 0.05) and have been calculated for both hcnC (a, b, and c) and phlD (x, y, and z).

DISCUSSION

A universal protocol capable of high-efficiency extraction of mRNA from different soil types and conditions is yet to be developed (4). However, the use of a carefully optimized RNA extraction protocol, for example, the silica-based spin column method, has been proposed to alleviate some of the issues plaguing soil RNA extraction and provides more reproducible results than other commonly used approaches (19). For this reason, the commercially available, silica-based spin column method (MoBio RNA extraction kit) was used in our study. Apart from its ease of use, the resulting RNA yields and purity values were similar and sometimes higher than most of those reported in the literature (7, 9, 18, 28).

The few studies that have reported on microbial gene expression in soils have all relied on an absolute quantification approach (4, 21-23, 35). In theory, this quantification approach is unable to account for variations occurring during RNA extraction, which could result in a potentially inaccurate estimation of the target gene transcripts (10). The relative quantification approach, however, is able to address these limitations, but technical issues in developing both endo- and exogenous calibrators have prevented its successful implementation in soil studies to date (4). Since a detailed a priori knowledge of the expression patterns of endogenous calibrators (e.g., housekeeping genes) is required for relative quantification with reference to internal standards, researchers are now increasingly using exogenous spike-in RNA to normalize the relative quantification data (5, 8, 16, 31, 38). However, to the best of our knowledge, there is no report to date showing that such an approach in soil studies has been implemented.

With no previous published report, we first attempted to use the human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene transcripts as exogenous spike-in RNA in our experimental system. The choice and logic of using the GAPDH mRNA centered on the fact that it is highly characterized (26), available in copious amounts from commercial sources (e.g., Applied Biosystems), and most importantly, being of human origin, assumed to exhibit negligible cross-reactivity to coextracted soil components. Empirically, however, the qRT-PCR results using the GAPDH mRNA exogenous spike-in were inconsistent and very difficult to interpret, possibly due to higher cross-reactivity than that expected of the GAPDH primers/probes to coextracted nontarget sequences and other unexplainable factors (data not shown).

The recent development of a hypothetical nucleotide sequence for myIC (17), specifically for qPCR/qRT-PCR, provided a unique opportunity to practically test its suitability as an exogenous spike-in control under soil conditions. A preliminary experiment using the myIC primer/probes on the DNA/RNA extracted from our experimental test soil did not generate any qPCR or qRT-PCR signals, indicating its potential applicability in soil studies. We then proceeded to set up a controlled soil microcosm and quantified the temporal expression patterns of two bacterial gene transcripts, phlD and hcnC, using the myIC RNA as an external standard. The microcosm essentially consisted of fixed amounts of natural nonsterile field soil inoculated with defined numbers of the test bacterium Pseudomonas sp. strain LBUM300. In our experimental procedure, a defined amount (300 ng) of the myIC RNA was added exogenously during RNA extraction, specifically after the post-phenol step. Our initial attempts to spike any step upstream of the phenol treatment resulted in either no signal or a highly inconsistent qRT-PCR signal (data not shown). This could presumably be due to the degradation of the myIC RNA by the cell-disrupting steps or endogenous soil RNases, which we assume were sufficiently inhibited/eliminated postdeproteinization (phenol treatment).

Overall, both of the quantification methods, absolute and relative, showed a similar transcriptional trend for the phlD and hcnC gene transcripts, which is not surprising since both genes are subjected to similar regulation systems (6, 29). The use of a defined experimental system and a highly efficient RNA extraction procedure could also be a contributing factor toward the congruity of the final results generated using the two different approaches. The lack of significant difference in the absolute quantification data, normalized to total RNA (data not shown), also proved the robustness of the RNA extraction procedure. Interestingly, such normalization has been proposed to better control sample-to-sample variability during absolute quantification (10, 32). With both quantification methods, we observed that the phlD and hcnC transcripts diminished with respect to the initial inoculated bacterial concentration in soil, regardless of the sampling date. Factorial ANOVA confirmed the significant effect of initial bacterial concentration on the absolute copy number of gene transcripts detected for phlD and hcnC. In this experiment, using log dilutions of bacteria and assuming that all bacteria had equivalent levels of expression of the genes under study, increases or decreases in phlD or hcnC gene expression could be reliably detected. A posteriori analyses revealed no significant differences in the phlD and hcnC transcript copy numbers for some of the initial bacterial concentrations under study, especially those at the lowest concentrations. This suggests that transcripts present in very low numbers in soil could probably be detected but not be precisely quantified with the proposed approach. Also, regardless of the gene transcript under study, the initial concentration of the bacterium used, or the time, the number of transcripts detected was always lower than the number of bacterial cells originally inoculated. This could be best explained by either of the following: (i) not all cells at a particular cell concentration would transcribe the target gene(s) or (ii) extraction-related inefficiencies, which are typical in soils, could result in a significant loss of RNA, hence dampening the signal from the expected one.

The diminution in gene expression observed from days 7 to 14 could be attributed to the fact that gene expression of both phlD and hcnC is preponderant from late exponential phase to early stationary phase (13, 27). A recent study with Rhodococcus sp. strain RHA1 demonstrated that bacterial growth in sterile soil achieved stationary phase after 5 to 7 days (35). Gene expression quantification of phlD and hcnC was also performed at day 0, directly after soil inoculation, which is, however, more representative of gene expression in bacterial cultures than actual gene expression in soil.

In this study, although it can be assumed that the population of Pseudomonas sp. strain LBUM300 might have changed during the 14 days of the experiments, gene expression data are all related to the initial bacterial concentration at day 0 and not to the actual number of cells at harvesting periods. The reason for this is that metabolically active Pseudomonas sp. strain LBUM300 cells could not be reliably detected after inoculation because (i) culture-dependent approaches are not specific enough to quantify a specific strain under nonsterile soil conditions and (ii) culture-independent microbial quantification approaches might detect DNA from nonmetabolically active or residual dead cells, as harvesting periods were separated only by a few days. This is why correlation of the exact number of gene transcripts detected to the number of metabolically active cells was not possible under these conditions.

In summary, this study represents an important step in the development of a relative qRT-PCR protocol for quantitative analysis of bacterial gene transcripts in soil. The use of the myIC IAC generated specific and reproducible results. The noncommercial and nonproprietary usage of the myIC IAC is an added incentive for its wide-scale usage (15). Even though the use of the myIC IAC was successful in our controlled experimental setup, certain caveats need to be considered prior to its application in a specific soil experiment. First, researchers need to (i) empirically establish any cross-reactivity and complexation of the myIC RNA issues with their experimental test soil and (ii) finally determine the exact spike-in amounts based on the RNA extraction protocol in use and the intensity of the qRT-PCR fluorescence signal desired.