The microbiology of table salts Seana P. Thompson Cooperative Institute for Research in Environmental Sciences and Department of Molecular, Cellular, and Developmental Biology, University of Colorado, Boulder, Colorado, USA Honors Thesis Committee: Noah Fierera, Abby Hickcoxb, Ravinder Singhc, Alison Vigersd a Department of Ecology and Evolutionary Biology, Cooperative Institute for Research in Environmental Sciences, University of Colorado, Boulder, Colorado, USA b Arts and Sciences Honors Program, University of Colorado, Boulder, Colorado, USA c Department of Molecular, Cellular, Developmental Biology, University of Colorado, Boulder, Colorado, USA d Thesis Advisor. Department of Molecular, Cellular, Developmental Biology, Department of Psychology and Neuroscience, University of Colorado, Boulder, Colorado, USA Defense Date: April 3rd, 2020 Abstract The dominant bacteria and archaea on table salts from around the world were identified and how these halophilic families varied as a function of salt chemical makeup was investigated. Through this study, the dominant taxa found on table salts were reported to be: Haloferacaceae, Pseudomonadaceae, Halomicrobiaceae, Halobacteriaceae, Nanohaloarchaeaceae, Rhodothermaceae, Burkholderiaceae, and Rhizobiaceae Marinobacteraceae, Alteromonadaceae, Rhodobacteraceae, Micrococcaceae, Moracellaceae, Staphylococcaceae, Oligoflexaceae, Halomonadaceae, and Flavobacteriaceae. The relative abundance of these microorganisms across samples was variable and related to the chemical composition of the table salts, including concentrations of iron (Fe), potassium (K), sodium (Na), magnesium (Mg), and calcium (Ca). The microbial families clustered into three general clusters. The three clusters were defined by Fe/K abundance, Na abundance, and Ca/Mg abundance. The Fe/K cluster showed the smallest range of microbial diversity and hosted a community dominated by Haloferacaceae. The Na cluster also hosted a dominant community of Haloferacaceae. This cluster also had a greater range of other microbes including Pseudomonadaceae, Halomicrobiaceae, Halobacteriaceae, Nanohaloarchaeaceae, Rhodothermaceae, Burkholderiaceae, and Rhizobiaceae. The Ca/Mg cluster showed the greatest diversity of microorganisms including Haloferacaceae, Pseudomonadaceae, Halomicrobiaceae, Halobacteriaceae, Nanohaloarchaeaceae, Rhodothermaceae, Burkholderiaceae, and Rhizobiaceae Marinobacteraceae, Alteromonadaceae, Rhodobacteraceae, Micrococcaceae, Moracellaceae, Staphylococcaceae, Oligoflexaceae, Halomonadaceae, and Flavobacteriaceae. These finding indicate a potential correlation between certain elements present in table salts and the microorganisms that reside there. Diverse microbial families thrive on table salts of varying composition, emphasizing the unique microbiology of a widely used food item. Introduction Environments with high concentrations of salt are considered extreme. These places foster microbial families that are unique from more hospitable environments because microorganisms must have a high salt tolerance to survive. Even something as small as a grain of salt can host a microbiome that is distinct from other environments. Microbes that thrive in salty conditions are called halophiles, and they are found across all three domains of life. Within the Bacteria domain of life, halophiles are found within the Fig. 1 The set of 40 table salt samples received from the Robert Dunn Lab at North Carolina State University. Cyanobacteria, Proteobacteria, Firmicutes, Photo taken by Lauren Nichols from the Robert Dunn Lab. Actinobacteria, Spirochaetes, and Bacteroidetes phyla [7]. Within the Archaea domain, halophilic microorganisms are found in the class Halobacteria and in Methods the order Methanococci [7]. Halophiles and non-halophilic microbes are often found A set of 40 table salt samples was collected within the same lineages. However, certain by the Robert Dunn Lab at North Carolina cohorts are made up entirely of halophilic State University from various salt species. These halophile-specific groups are companies across the world (Fig. 2). This set the order Halobacteriales of family was sent to the Fierer Lab and were stored in Halobacteriaceae (Euryarchaeota), the the -20C freezer to arrest microbial growth. order Halanaerobiales (Firmicutes) of The following procedures were followed for bacteria, the family Halomonadaceae the 40 samples examined in this study. (Gammaproteobacteria). Table salt is known generally by the chemical formula NaCl. However, there are Filtering Protocol other elements present in salt and each salt has a unique elemental makeup. With this in 0.25 grams of each salt was weighed and put mind, and the knowledge that different in individual 50 mL conical tubes. 50 mL elemental abundances can host unique microbial families, elemental composition of sterile water was measured into each tube. Tubes containing salt and sterile water were environments could be used to explain vortexed until the salt was fully dissolved. variation in microbial families. Specifically, The contents of each tube were poured over the elemental composition of different table a filter and filtered via vacuum filtration. salts could explain some of the apparent Another 50 mL of sterile water was poured variation in microorganisms across different salts. Fig. 2 The metadata that was received with the collection of salt samples. Not all information was available for each sample. An Invitrogen SequelPrep Normalization over the same filter and filtered via vacuum Plate (96) Kit was used to normalize the filtration. The filter was then removed with amplified PCR products. tweezers and put in individually labeled The microbial dataset was produced whirlpak bags and stored in a -20C freezer. using Illumina DNA sequencing, in which When original samples contained enough DNA is amplified, sequenced, and then salt to run the protocol again, duplicates analyzed. More information on sequencing were taken. can be found in the references [1]. Note: Filter blanks, extraction blanks, DNA Extractions and PCR blanks were all done as controls to compare to microorganisms found in all Salt sample filters were cut into pieces on samples. sterile petri dishes with tweezers and razor blades to make them the correct size to fit Salt Samples Chemical Analysis into DNA extraction single tubes. After the filters were cut and placed into DNA Each salt sample was chemically analyzed extraction single tubes, the DNeasy for presence and abundance of iron, PowerSoil Kits (Qiagen, DE) instructions potassium, magnesium, calcium, and sodium were followed. using inductively coupled plasma mass spectrometry. Mantel tests were run using PCR Bray-Curtis transformation to test for correlations between chemical elements. 16S rRNA gene PCR was performed on the extracted salt samples to amplify the DNA. Excel and R Studio Analysis Gel electrophoresis was done to verify the success of the PCR and DNA extractions Raw data was processed and filtered using using agarose gel and a DNA ladder for standard methods via DADA2 [1]. Blanks comparison. The gel was made with 135 mL were found to have very low community of Syber Safe and 2.5 grams of agarose. It abundance compared to the majority of the was run at 140 V for 45 minutes. samples. Thus, a cutoff of 8,000 total ASV reads was used to remove the blanks from downstream microbial analysis. Chloroplast, check if chemical elements were driving mitochondria, eukaryote, unassigned, and microbial community difference by NAs were filtered out and data was rarefied comparing individual matrices of chemistry to a depth of 8,000. Taxonomic relative and microbial data. PCoA was made to abundances were summarized. Using the extract ordination score and check metadata that came with the salt samples, correlation with chemical variables. PCoA dissimilarity matrices and ordination plots with Hellinger transformation was graphed were used to try to find clustering of the using Bray-Curtis Matrix and different samples. Ordinations were done with the metadata factors to find possible explanation metadata “Color”, “Type”, “Salt”, and for clusters. Chemistry data vectors were “Continent”. Microbial taxonomy was added to the microbial clustering graph. plotted in stacked bar plots at the order and Heatmaps for chemical abundance in family levels. Duplicates were dropped since samples and heatmaps for dominant duplicates were not able to be done for all microbial taxa in samples were created using samples. A taxa dissimilarity matrix was the ggplot2 package [2] and using Bray- calculated using the Bray-Curtis square root Curtis Square Root and Hellinger transformation. A mantel test was used to transformations. Fig. 3 The top 50 ASV reads from the microbial sequencing data. A majority of the top ASVs are salt- specific microbes. Fig. 4 Stacked bar plot showing the top 19 microbes found in the salt samples at the order level. “Na” is an unknown order of the Millicutes class of microbes. Results range of variation in families. There was a total range of 4,709 unique ASVs. Sequencing shows dominant microbial Ordinations done with the metadata “Color”, taxa found to be salt-specific microbes “Type”, “Salt”, and “Continent” were not valuable for explaining the clustering of Sequencing of the DNA extracted from the microbial samples. Without different salt samples gave us taxonomy (taxa) metadata, only stacked bar plots comparing information about the microorganisms found the abundances of microbial families within on these samples. The top