Sphingopyxis Species Isolated from Sand Filter Biofilm at an Australian Drinking Water Treatment Works

ABSTRACT Three strains isolated by geosmin enrichment from a sand filter in an Australian drinking water treatment works were genome sequenced to identify their taxonomic placement, and a bench-scale batch experiment confirmed their geosmin-degrading capability. Using the average nucleotide identity based on the MUMmer algorithm (ANIm), pairwise digital DNA-DNA hybridization (dDDH), and phylogenomic analyses, the strains were identified as Sphingopyxis species.

Sphingopyxis strains Geo24, Geo25, and Geo48 were originally isolated in the lab by geosmin enrichment from a sand filter from an Australian drinking water treatment works (13,14). Geo24 and Geo25 were identified as part of a bacterial consortium able to biodegrade geosmin as the sole carbon source (13), and Geo48 was later identified as an isolate capable of geosmin biodegradation (14). The isolates were stored long term at 280°C, shipped on charcoal transport swabs, streaked onto tryptic soy agar (TSA), and grown for 48 h at 30°C. All strains were streaked three times to ensure the purity of individual colonies.
For genome sequencing, the strains were grown in 5 mL tryptic soy broth (TSB) at 30°C for 48 h, and total DNA was extracted using the FastDNA spin kit for soil (MP Biomedicals). Sequencing was performed on a NovaSeq 6000 SP instrument using a NEBNext Ultra II DNA library prep kit. Between 7 and 9 million (150-bp) read pairs were generated for each genome. The read quality was checked using FastQC v0.11.9, trimming and adapter removal was performed using Fastp v0.20.1, and the genomes were assembled using Unicycler v0.4.7. The genome sizes and other metrics are as follows: Geo24 has a size of 3.86 Mbp, 23 contigs, an N 50 value of 665,169 bp, and a GC content of 65.3%; Geo25 comprises 3.87 Mbp, 23 contigs, an N 50 value of 672,543 bp, and a GC content of 65.3%; and Geo48 has a size of 3.96 Mbp, 25 contigs, an N 50 value of 978,883 bp, and a GC content of 65.2%.
Phylogenetic analysis of Sphingopyxis strains Geo24, Geo25, and Geo48 was performed against all available Sphingopyxis type strain genomes, downloaded using the NCBI genome download tool v0.2.10 and annotated using Prokka v1.14.6. The average nucleotide identity Microbiology Resource Announcements (ANI) was calculated using PyANI v0.2.12 (Fig. 1A), and a phylogenomic tree was constructed using OrthoFinder v2.5.4 (15,16) and RAxML-NG v1.1 (17) (Fig. 1B). All three strains showed an ANI of ,94% to the most closely related type strains and displayed phylogenomic distinction, indicating them as potentially novel species (Fig. 1B). Pairwise digital DNA-DNA hybridization (dDDH) values were determined using the Type Strain Genome Server (TYGS) with the most closely related type strains (18); the results indicated that all three Sphingopyxis strains represent different species (dDDH, ,70%) than those previously described (Table 1). Geosmin biodegradation capacity was confirmed in a microcosm batch experiment by analyzing the geosmin concentration over 7 days. Microcosms comprised of 10 mL bacteria, diluted to 0.1 optical density at 600 nm (OD 600 ), in basal salts medium (BSM) (19) in vials with 20 mL headspace. Geosmin losses by volatilization were controlled with microcosms of 10 mL BSM. Geosmin was added at 100 ng/L to each microcosm and measured at 0, 4, and 7 days in triplicate using solid-phase microextraction and gas chromatography mass spectrometry (GCMS) analysis. Significant removal of geosmin was observed for all inoculated microcosms compared to the control (Fig. 1C), demonstrating that all three Sphingopyxis strains can degrade geosmin.
Data availability. The genome sequences and raw reads have been deposited in the European Nucleotide Archive (ENA) under the project/study number PRJEB60073. The accession numbers for the genome sequences are ERS14837053, ERS14837054, and ERS14837055 for Geo24, Geo25, and Geo48, respectively.

ACKNOWLEDGMENTS
This study was supported by funding from KESS2 East and Dŵr Cymru Welsh Water. All analyses were conducted using Cloud Infrastructure for Big Data Microbial Bioinformatics (CLIMB), funded by the Medical Research Council (MR/L015080/1). We thank South Australian Water (SA Water) for providing the bacterial strains investigated in this paper, and we acknowledge L. Ho, G. Newcombe, A. Keegan, and W. Aunkofer for their contributions to the initial isolation of the strains. nucleotide identity, as specified by the color bar key. Novel species are shown in bold against Sphingopyxis type strains. (B) Phylogenetic tree created using OrthoFinder v2.5.4 and RAxML-NG v1.1 to show the relationship between novel Sphingopyxis species and available type strain species, rooted with the type strain from the sister genus Novosphingopyxis. The maximum likelihood method was used with the LG model and G4 distribution, with bootstrap support (100 replicates) shown next to each node. Novel Sphingopyxis species are shown in bold. (C) Geosmin concentrations for microcosms in batch experiment comparing geosmin removal of novel Sphingopyxis species over 7 days. Bacteria (1 mL) grown in TSB, washed, and controlled to an OD 600 of 1 were added to 9 mL BSM and spiked with 100 ngL 21 geosmin. Blank control microcosms with 10 mL BSM, spiked with 100 ngL 21 geosmin and with no inoculum, were run simultaneously. Geosmin concentration analysis was performed using SPME with GCMS analysis. Asterisks indicate statistical significance from the blank control at each time point, determined using the Mann-Whitney test. a dDDH values were calculated using TYGS and according to the Sphingopyxis phylogenetic tree (Fig. 1A).