3. Results and Discussion

3.1. Performance of synthetic and plasmid DNA standards

For the six targeted genes which were tested with synthetic DNA and plasmid DNA standards (Table 1), both standards for qPCR quantifications yielded significantly (P < 0.001) linear calibration curves featuring a coefficient value (R2) of > 0.99 (Fig. 1), together with similar high R2 derived from the genes tested with synthetic DNA standards (Supplementary Fig. S1). The dilution series of both synthetic DNA and plasmid DNA standards exhibited smooth and exponential amplification curves (Fig. 2). Coefficients of variation of Cq values among the replicates for the standards in the range from 101 to 108 copies per μl were between 0.03% and 4.15%, which indicated good repeatability and reproducibility. Additionally, the slopes of the synthetic standards were similar to those of the plasmid standards, with only minor differences (Fig. 1). These, all together, provided compelling evidence that, similar to traditional plasmid standards, synthetic DNA standards of serial dilutions can be amplified effectively and produce high-quality and consistent standard curves. Comparable standard curves between synthetic DNA standards and traditional standards (PCR amplicons and plasmids of cloning) have been reported in previous studies targeting human mitochondrial gene (Conte et al., 2018), antibiotic resistance gene (Xu et al., 2019), human T-cell leukaemia virus type 1 (HTLV-1) (Bandeira et al., 2020) and HBV virus (Portilho et al., 2018). Yet, to the best of our knowledge, this is the first study to use synthetically designed and produced DNA fragments as qPCR standards targeting a broad range of genes involved in C and N cycling employed in microbial ecology.
Standard curves from both synthetic and plasmid standards showed high and similar amplification efficiency (E) values, confirming the reliability of synthetic gene fragments as qPCR standards (Fig. 1 and 2). PCR efficiency of the standard curves for 16S rRNA gene reached 0.84 for synthetic DNA and 0.96 for plasmid standards. It attained 0.98 and 0.90 for fungal ITS region for the synthetic and plasmid standard respectively. Efficiency values of the remaining four genes tested (mcr A, pmo A, nif H and nos Z), were all lower than 0.90, irrespective of plasmid or synthetic standards (Fig. 1). Ideally, an efficiency value over 0.90 is considered a well amplified standard and a qualified standard curve (Svec et al., 2015). However, often times due to the potential PCR inhibition, such as self-inhibition, polymerase and protein inhibition, primer specificity and contamination, E values can be as low as 0.70 (Luby et al., 2016; Xu et al., 2019). The slightly lower E value for synthetic 16S rRNA gene standard during PCR might be caused by a small peak (PCR byproduct) right before the main PCR product peak of 16S rRNA gene, especially the least diluted ones, implied by the melting curves (Supplementary Fig. S2A), which caused the differences in the standard curves from synthetic and plasmid standards. Lower E values of the genes mcr A, pmo A, nif H and nos Z from both synthetic and plasmid standards were likely also related to PCR inhibitions, in particular to the less diluted standards. Usually, 3.3 (a slope of – 3.3) cycles apart of the 10-fold dilutions were considered as an indicator of 100% PCR efficiency (Svec et al., 2015)). However, much higher Cq value differences between the dilution series were found for these four genes from both standards (Fig. 1), indicating an inhibition effect. Additionally, different instruments and volume for standard dilution also made huge differences in PCR efficiency, which turned out a larger volume transferred during dilution (10 µL) could increase the efficiency (Svec et al., 2015), while 2 µL was used in this study.
In spite of the similarity of qPCR standard curves between synthetic and plasmid standards, there were slight differences in the slopes and E values of standard curves between these two standards (Fig. 1). Formcr A, pmo A, nif H and nos Z, standard curves from synthetic standards were always steeper (higher absolute slopes) than those from plasmid standards, with Cq values of the least diluted standards from synthetic standards lower than those of the least diluted plasmid standards, even when the copy numbers of the least diluted standards from synthetic standards were lower than those from plasmid standards for mcr A and pmo A (Supplementary Table S4). This indicated a self-inhibition of the least diluted plasmid standards, which took more cycles (higher Cq values) to get fully amplified. For ITS region, standard curve from plasmid standard was slightly steeper than that from synthetic standard, which also indicated an effect of inhibition. In addition to self-inhibition, there might be also a conformation effect of non-linear plasmid standards, which could also impact PCR efficiency as well (Hou et al., 2010).

3.2. Microbial gene abundances in soils based on synthetic and plasmid DNA standards

In order to validate the reliability of our synthetic DNA standards, the abundances of the tested genes were quantified in eDNA extracts from soils by qPCR assays and the gene copy abundance calculated using standards from synthetic and plasmid DNA (Fig. 3). No amplification was observed in negative control reactions, confirming the absence of contamination during the reaction preparation steps. Overall, gene copies calculated with either standard curves (i.e. from synthetic or plasmid DNA) were not significantly different for all the genes studied across all soil samples except for few assays (marked with asterisks in Supplementary Fig. S3). These significant differences reflect the described differences of the standards curves thus that gene copy numbers calculated from synthetic standards were on average lower than those from plasmid standards (difference of 5.1±4.4% for mcr A, 15.4±1.2% for pmo A, 23.8±4.1% for nif H and 6.9±6.2% fornos Z) or higher for 16S rRNA gene (17.7±3.2%) and ITS region (41.6±16.4% higher). Gene copies varying within one log (10 times) were widely reported for qPCR quantifications of viruses with synthetic and plasmid standards, and such results have been considered good agreement of the two methods (Bandeira et al., 2020; Lima et al., 2017; Portilho et al., 2018; Tourinho et al., 2015).
Despite the described deviation, correlation of gene copies in the soil samples using the synthetic standard and copies by using plasmid standard, was significant with a squared coefficient (R2) of over 0.99 (Linear correlations, all significant P < 0.001) for all the six tested genes (Fig. 3), which showed highly identical results with both standards. Similarly, high R2 (0.83) based on Linear correlations were also observed in a study on human virus by qPCR when comparing synthetic DNA and plasmid DNA standards (Bandeira et al., 2020). Furthermore, when comparing the relative differences in gene copy numbers to the highest observed soil value within each gene, there were no any significant differences in the relative differences across all sites of all the six tested genes between synthetic and plasmid standards (Supplementary Fig. S4).
Therefore, considering the sensitivity and efficiency of qPCR, inhibition and anthropogenic interference (i.e. pipetting errors), differences in copy numbers within 50% variation in this study are very much acceptable, especially with gene concentrations reaching up to more than 1010 copies per dry gram soil. All taken together, our results demonstrated that the synthetic DNA standards are reliable for qPCR quantification of various taxonomic and functional genes in soils.