Towards Autonomous Microcystin Detection: Investigating Methods for Automation

摘要

Due to increased anthropogenic activity, severe eutrophication is occurring in bodies of water around the world. Effects include decreased water quality, decreased value of surrounding land and recreational use (estimated loss in revenue of 0.67 and 3.96 U.S. billion dollars per year), and increased occurrence of toxin producing Harmful Algal Blooms (HABs). Microcystins are cyclic peptides made up of 7 amino acids and 800-1100 Daltons in size. They are one of the most predominantly produced of these toxins, and therefore was the focus of this study. Numerous structural variants of microcystin (referred to as congeners) exist, but microcystin-LR is one of the most common, having a World Health Organization (WHO) recommended limit of 1 microg/L in drinking water. In order to make informed public health decisions on potable and recreational water, an automated in situ instrument for detection of microcystin and its nucleic acids is needed. Very few detection systems have reached the market (i.e. Environmental Sample Processor, McLane Laboratories, USA), but all remain prohibitively costly and complex. Currently, research in many fields is directed towards developing a more cost effective automated in situ detection instrument that can collect and filter environmental samples, extract toxins and nucleic acids, and detect and quantify analytes, genes, and gene transcripts. In this study, a sample preparation method for on-filter collection, filtration, and dual extraction of microcystin and nucleic acids was developed during the summer of 2016 on environmental samples from two bodies of water, Lake Winnebago, WI and Veteran's Park Lagoon, Milwaukee, WI. Results were compared to a traditional laboratory bead beating method. Results showed that the median extraction ratios (quantified by mass spectrometry) obtained with on-filter method compared to bead beat method (comparative recovery) for microcystin congeners MC-LR, MC-YR, MC-RR, and MC-LA were 43% +/- 12%, 34% +/- 9%, 46% +/- 10% and 44% +/- 13%, respectively for Lake Winnebago. The median comparative recovery for MC-LR, MC-YR, and MC-RR was 51% +/- 9%, 49% +/- 12%, and 53% +/- 7%, respectively, for Veteran's Park Lagoon. Total RNA extraction by the on-filter result showed lower and more inconsistent ratios. Comparative recovery values for the Veteran's Park Lagoon ranged from 6% to 27% and 5% to 64% for Lake Winnebago. Further quantification with RT-qPCR is needed to evaluate extraction efficiency of the desired gene cluster (mcy). Methods that were evaluated for detection of microcystin included chemical derivatization (fluorescent derivatization) and optical signal amplification (direct and indirect hybridization schemes using DNA aptamers and oligonucleotide probes, nicking enzyme assisted fluorescent signal amplification (NEFSA)). Methods evaluated for detection of nucleic acids included optical signal amplification (direct and indirect hybridization, NEFSA, cascading amplification of nucleic acids (CANA)) and nucleic acid amplification (strand displacement amplification (SDA)). Of the techniques tested, SDA gave non-specific or no amplification, fluorescent derivatization was inconsistent, and all hybridization schemes resulted in non-specific binding. Preliminary results from NEFSA and CANA showed promise, but were inconsistent. Therefore, further optimization of reaction conditions is necessary to conclude if either could be viable options for use in an automated in situ detection system in combination with the on-filter sample preparation and extraction technique.