The combined observations of the 16s rRNA data, predicted metagenome, and Biolog EcoPlate assays suggest that the soil bacterial community in our samples was in fact sensitive to the levels of APAP used in this study. Significant community differences were observed by 3 weeks of APAP treatment and significant differences in carbon metabolism were observed between treated and untreated samples collected at 7 weeks. Taken together, our results show that APAP was able to alter the soil bacterial communities and impact their functions.
APAP addition did not change individual community 𝛂-diversity levels as measured by the Shannon-Wiener index. The Shannon-Wiener diversity index remained relatively constant among treatments for the duration of the study. Although microbial communities can be sensitive to perturbations, plants can act to stabilize their associated soil microbial communities 58,59. Previous observations suggest that all microbes are not impacted equally by a given disturbance and some are more resistant to perturbations than others 39,40. Our results suggest that APAP was not able to completely displace many taxonomic groups of bacteria that were present in the soil. Instead, APAP caused certain bacterial groups to decrease in relative abundance and other groups to increase, thus allowing the community to maintain a similar level of 𝛂-diversity. Measured Shannon-wiener diversity would be less sensitive to community shifts in which the number of bacterial groups present remained the same. However, 𝛃-diversity, measured as Bray-Curtis distance matrices, would be more sensitive since it is affected by changes of relative abundance within and among different groups.
The observed significant differences of 𝛃-diversity between APAP treated and the initial sampling point suggests that the overall community structure (𝛽-diversity) was sensitive to the addition of APAP at concentrations found in RWW, especially after 7 weeks of exposure. Our observations are consistent with previous ones that indicate pharmaceutical products, including APAP, can impact microbial communities, and even hinder or disrupt key microbial functions 20–22,60. The APAP concentrations used in our study may appear low, but they represent levels typically found in RWW effluent 15,29,61. Our results suggest that the concentrations of APAP found in RWW can impact soil microbial communities, especially if they are repeatedly exposed to the CECs. These results are consistent with previous observations of microbes in agricultural soils that were observed to be sensitive to other pharmaceutical products present in treated wastewater 56,62. However, in these studies the resolution for detecting specific microbial community members was limited because they relied on non-sequencing-based approaches to characterize changes in the soil microbial community. In contrast, our study was able to detect specific shifts in the microbial community and identify specific bacterial groups that were impacted by APAP treatment by using Illumina sequencing based approaches.
The changes in the microbial community were most likely driven by changes in soil conditions caused by the addition of APAP. Many soil microbial communities are sensitive to soil conditions (such as pH, organic matter content, and mineral composition), and those caused by anthropogenic inputs such as pharmaceutical products 22,62. When the microbial community changes, the overall functions may change as well. Introduction of antibiotics like sulfamethoxazole into soil can cause decreases in the metabolism of a number of carbon substrates 56 or disrupt nitrogen cycling in bacterial communities 39,63. The Addition of APAP to soils has also been found to disrupt key aspects of nitrogen cycling although the concentrations of APAP (50 to 1,000 mg/L) used were greater than those found in RWW effluent 22,60. Our results from the Biolog assay showed altered microbial functions while using lower concentrations of APAP (5 ug/ L and 10 ug/L) that are in the range of those found in RWW effluent. Besides lower APAP concentration, our study distinguishes itself from previous ones in a few other ways. Unlike previous studies that focused on nitrogen cycling 39,63, our study examined a wide range of carbon metabolism pathways by using 31 ecologically relevant carbon sources (Supplementary Table 2). This is a very robust approach that can encompass nearly every member of the microbial community. The biolog approach allowed us to widen our scope beyond a specific set of community members such as anammox bacteria 60, or bacteria that contain amoA, napaA, or nifH genes for nitrification, denitrification, or nitrogen fixation respectively 39. Nitrogen cycling is critical for the soil microbiome and its associated environment, but some organisms that are not efficient nitrogen cyclers may be overlooked. Thus, by looking for variations in metabolic rates of various carbon sources we could screen for a wide variety of bacterial groups that were impacted by the addition of APAP to the soil. The study Liu et al., 201256did examine microbial community functions using biolog plates. However, their study did not simultaneously employ a method that would allow for the identification of specific microbial organisms that are shifting in abundance or that may be responsible for the changes in carbon metabolism they observed.
In this study, APAP treated samples had significantly higher rates of carbon metabolism in nearly every category measured (carbohydrate, amino acid, carboxylic acid, and polymer metabolism) compared to the control. We think this occurred because APAP is a carbon source for some organisms, therefore, its addition to the soil selects for microbes that use different sources of carbon more efficiently. Li et al., 201356 demonstrated that APAP is broken down in non-sterilized soil but not in sterilized soil. Indicating that soil microbes are able to break down APAP and use it as a carbon source. Metabolomics analyses revealed that when APAP is added to soil, the microbes were able to break it down to 8 identifiable intermediates, namely: 3-hydroxyacetaminophen, hydroquinone, 1, 4-benzoquinone, N-acetyl-p-benzoquinone imine, p-acetanisidide, 4-methoxyphenol, 2-hexenoic acid, and 1, 4-dimethoxybenzene 56. The intermediate 2-hexenoic acid is a carboxylic acid and it was found to be the most abundant metabolite in the soil after APAP treatment by Li and his colleagues (2013). This led us to conclude that treatment of soil with APAP increases carboxylic acid content in the soil, which in turn increases microbes that can use it as a carbon source. We were able to confirm this using our Biolog assay, which showed that soil treated with APAP had a significantly higher carboxylic acid metabolism compared to the control (ANOVA post hoc Tukey test, P< 0.01). Other carbon sources whose metabolism was found to be significantly higher in APAP treated soils were amino acids, carbohydrates, polymers and phenolics (Fig 6).
Furthermore, the Picrust analyses indicated that amino acid and carbohydrate metabolisms constituted the most prominently expressed genes within the soil communities from APAP treated soils, which was consistent with the Biolog assay results. However, the Picrust data did not detect differences in gene expression among APAP treatments that were used in the Biolog assay. This is not too surprising because Picrust predicted metagenomes rely on detecting the presences of different 16s rRNA genes from bacterial groups and it is based on previously established databases 55,64. Also many soil microbial communities can exhibit a high degree of genomic, and therefore metabolic, redundancy 65,66. It is possible that different soil microbes contained a high degree of overlapping genes. Therefore, Picrust predictions would be insensitive to changes in gene expression, if the total metagenome pool remains stable. Picrust may be only able to detect changes in genome expression among drastically different microbial communities with major differences in the constituting microbial groups. The Biolog plates, on the other hand, can detect changes in the community-level physiological profiles even when the microbial shifts are only in relative abundance changes of constituent organisms. Thus, the Biolog assay may be a better indicator of gene expression levels, especially for genes related to metabolic pathways.
The specific shifts in relative abundance of the soil microbial community members were very consistent with the observed changes in the microbial community function determined in the carbon substrate metabolism (i.e. BIOLOG assay). A. thermoflava and Cellvibrio sp. were two microbial groups that increased in relative abundance after APAP application. These groups were major contributors to community differences among the different soil communities and are capable of metabolizing a diverse set of carbon substrates, including glycosides 36–38. Glycosides are major breakdown products of APAP in the soil as a result of fungal 34,35 and plant detoxification activities 27. Another major breakdown product of APAP due to microbial activity is a carboxylic acid 30, which can be utilized by these two bacteria as well as some of the others that were observed to increase in abundance. In addition, A. thermoflava is also capable of degrading xanthine, a compound that caffeine is derived from and salicin which is a compound similar to aspirin 37. Both caffeine and salicin are also considered to be CECs, thus this particular group of soil bacteria may be positively selected for in soils contaminated with a variety of different CECs.
Cellvibrio genus was also a major driver in community structure differences as determined by the canonical correlation analysis (CCA). Cellvibrio is a genus of cellulolytic bacteria, capable of degrading plant cell walls. Members of Cellvibrio genus increased in abundance in the APAP treatment. Some Cellvibrios have been found to be able to utilize many different carbohydrates and can even utilize ⍶- and 𝛽- glycosides 36,38. When APAP contaminated water is used for irrigation, plants translocate APAP and detoxify it into a glutathionyl and a glycoside conjugate, which accumulate in the roots 27 and can get into the soil. Soil fungi have also been reported to break down APAP to a glycoside conjugate 34. The presence of glycosides in plant roots or in the soil probably led to the increase of glycoside metabolising organisms like Cellvibrio bacteria after APAP treatment in our study. This is a particularly interesting group because cellulolytic organisms can have major impacts to the soil community in general by degrading refractory cellulose, making the substrates available to other community members 67,68.
Enhydrobacter aerosaccus is another group of bacteria that increased in relative abundance in the APAP treated soils. This organism also seems to be positively selected for by the conditions generated by the addition of APAP as this organism is known to ferment many different carbohydrates, and can also utilize many amino acids for a carbon source 69. Metabolism of carbohydrates and amino acids was found to be significantly higher in APAP treated soils compared to the non-treated controls and both functional groups were found to be major drivers of the community differences as determined by the CCA.
The observed increases in metabolic rates of many carbon sources also suggest that there will be community members that do not tolerate the soil conditions caused by APAP addition, and will decline in relative abundance. Our results showed a decrease in the relative abundance of actinobacteria in APAP treated soil compared to the untreated control. 34 found that many strains of actinomyces (which is a genus of class Actinobacteria) were not able to metabolize APAP. This could explain the reduction in the relative abundance of Actinobacteria observed in our study. A large group of bacteria identified to the Xanthomonadaceae family also decreased in relative abundance in the presence of APAP. This particular bacteria family is very diverse and contains numerous plant beneficial or deleterious organisms 70. At this point it is unclear if this particular decrease in Xanthomonadaceae family was a result of its direct interactions with APAP, or if this was the result of indirect interactions, such as APAP benefiting a competitor. This warrants further investigation. One group of bacteria identified to the Gemmatiomadetes phylum decreased in abundance in APAP treated samples. Gemmatimonadetes bacteria are normally found in high abundance in a variety of different soil environments but, not much is known about the group as there are virtually no culturable representatives that can be studied directly at this point 71.