Rebuttal to Correspondence on “Tracking Sources and Dissemination of Indicator Antibiotic Resistance Genes at a Watershed Scale”

Abstract

We thank Kirschner et al. for their recent commentary on potential opportunities for DNA loss during nucleic acid extraction within the context of monitoring environmental antibiotic resistance, using the example of our recent study. (1) In response, we submitted several relevant points. First, we appreciate the learning of the potential for bias introduced through the loss of DNA that can result from the sorption of nucleic acids to inorganic particles, affecting extraction efficiency. Given that this potential bias is not widely documented yet is relevant to potentially thousands of studies across the entire field of environmental DNA monitoring, this is an important opportunity to reconsider whether additional quality assurance and quality control (QA/QC) are needed in studies of environmental DNA. However, a critical evaluation of the need for this type of QA/QC should be placed within the broader context of the field. In particular, the field would benefit from a comprehensive summary of the state of knowledge on this topic and identifying research needs as this information is currently lacking. In ref (1), our team used appropriate controls as outlined by the Environmental Microbiology Minimum Information (EMMI) guidelines (2) and other best practices in the field, (3,4) such as those outlined in the recent Water Research Foundation report, “Standardizing Methods with QA/QC Standards for Investigating the Occurrence and Removal of Antibiotic Resistant Bacteria/Antibiotic Resistance Genes (ARB/ARGs) in Surface Water, Wastewater, and Recycled Water”. (5) The best practices used in our study included robust negative controls to evaluate contamination and the use of a dilution series to evaluate polymerase chain reaction (PCR) inhibition. However, we were not aware of the potential for DNA loss in this manner associated with sorption to inorganic particles, nor is that information available in protocols that are currently widely implemented in the field. Second, standardization and appropriate controls to ensure QA/QC are critical, but further research is needed to better understand what factors contribute to the impaired extraction efficiency and to identify appropriate controls. For example, Kirschner et al. propose using a defined target cell standard as a standard process control strain (DeTaCS), (6) but this practice has not been evaluated for use as a sample process control in monitoring environmental antimicrobial resistance (AMR). This is particularly important because while studies quantifying a gene originating from a single target organism may be appropriately evaluated using DeTaCS, using DeTaCS for evaluating the recovery of ARGs, which can originate from various target hosts within a single sample, introduces additional complexity. While the potential need for further QA/QC is clear, additional research is needed to determine the most effective approach for implementing these sample process controls within the context of environmental AMR research, where genetic targets may originate from a wide range of hosts in a single sample. Third, there are several reasons why the 16S rRNA gene concentrations measured in our study were low but plausible within the context of the study. Several sites highlighted as having lower 16S rRNA gene copy numbers (e.g., sites 2 and 7) were noted in the article to be strongly influenced by acid mine drainage (AMD). The harsh AMD environment is expected to have a microbial loading that is lower than that of pristine rivers or groundwater. (7−9) Significant seasonal variations in 16S rRNA genes measured in surface water have also been documented previously, with seasonally low concentrations comparable to those documented in our study. (10) Samples were collected year-round in our study, with water temperatures as low as 0.9 °C. Some of our sampling dates, such as February, April, and July 2021, coincided with severe snow and storm events, which could significantly impact microbial counts due to dilution and influence from snowmelt (i.e., near-freezing water temperature). Therefore, the low cell counts suggested by our 16S rRNA qPCR results are not “extremely unrealistic” but are plausible considering the appropriate context. Finally, we assert that the validity of our results and conclusions remains sound. The primary objective of our study was to assess the relative contributions of specific sources of fecal pollution to ARG loading throughout the studied watershed. While it is possible that extraction efficiency could have been impaired in some samples in this study, the most critical comparisons made in support of our key objective that form the basis for our conclusions were within-sample comparisons. Namely, comparisons between the abundance of microbial source tracking fecal markers and ARGs within samples were the primary point of comparison. In addition, in our study, we presented both absolute (ARG copies per milliliter) and relative (ARG copies per 16S rRNA gene copy) abundances of ARGs and found that absolute and relative abundances were highly correlated for all ARGs (Spearman’s ρ > 0.8, and p < 0.001). This approach of normalizing ARG abundances to quantities of the universal bacterial 16S rRNA gene has previously been used to correct for potential variations in DNA extraction efficiencies in complex environmental samples. (11) Therefore, even if our extraction efficiencies were impaired, the strong correlation between absolute and relative abundances observed in our study further suggests that our overall findings remain well supported by our collected data. While we welcome the opportunity to examine the need for additional QA/QC to improve collection and interpretation of data across the field as a whole, we do not believe there is a critical flaw in our published study that would compromise its findings. CRediT: Emily D. Garner conceptualization, writing - original draft; Mehedi Hasan Tarek conceptualization, writing - original draft; Jason A. Hubbart writing - review & editing. This article references 11 other publications. This article has not yet been cited by other publications.