Porphyrobacter mercurialis sp. nov., isolated from a stadium seat and emended description of the genus Porphyrobacter

David A. Coil^1, Jennifer C. Flanagan^1,, Andrew Stump ^1,, Alexandra Alexiev^1, Jenna M. Lang^1, Jonathan A. Eisen^1,2,#,

1 University of California Davis Genome Center, Davis, CA, USA.

2 University of California Davis Department of Evolution and Ecology, Department of Medical Microbiology and Immunology, Davis, California, USA

# Corresponding author: jaeisen@ucdavis.edu



A novel, Gram-negative, non-spore-forming, pleomorphic yellow-orange bacterial strain was isolated from a stadium seat. Strain CoronadoT falls within the Erythrobacteraceae family and the genus Porphyrobacter based on 16S rRNA phylogenetic analysis. This strain has Q-10 as the predominant respiratory lipoquinone, as do other members of the family. The fatty acid profile of this strain is similar to other Porphyrobacter, however CoronadoT contains predominately C18:1ω7cis and C16:0, a high percentage of the latter not being observed in any other Erythrobacteraceae. This strain is catalase-positive and oxidase-negative, can grow from 4-28 °C, at NaCl concentrations 0.1-1.5%, and at pH 6.0-8.0. On the basis of phenotypic and phylogenetic data presented in this study, strain CoronadoT represents a novel species in the Porphyrobacter genus for which the name Porphyrobacter mercurialis sp. nov. is proposed; the type strain is CoronadoT (=DSMZ 29971, =LMG 28700).


In this study, strain CoronadoT was isolated from a stadium seat at Niedermeyer Field, Coronado High School in Coronado, California, USA as part of a nationwide Citizen Science project (Project MERCCURI - www.spacemicrobes.org.) One goal of Project MERCCURI was to collect bacterial isolates to be used for an experiment aboard the International Space Station (ISS). The 16S rRNA gene sequenced from this particular isolate was at least 99% identical to rRNA genes from a few uncultured organisms. Uncultured isolates with high identity (=>99%) were found in samples from deep ocean sediment and the human skin microbiome. However, the highest identity to a cultured organism (Porphyrobacter donghaensis) (Yoon 2004) was only 95.5% (as determined by BLAST, (Altschul 1990)). Given the low identity to characterized species, a more detailed study of this isolate was undertaken.

A phylogenetic analysis of the Alphaproteobacteria class led in 2005 to the creation of a new family, Erythrobacteraceae, to house the genera Erythrobacter, Porphyrobacter and Erythromicrobium (Lee 2005). These genera were later joined by Altererythrobacter (Kwon 2007) and Croceicoccus (Xu 2009), the latter work also emended the description of the family. Members of the Erythrobacteraceae family are Gram-negative, aerobic bacteria that contain carotenoids, usually appearing pink, orange or yellow. They do not form spores, are chemo-organotrophic, and are most often associated with aquatic environments. The Porphyrobacter genus was established in 1993 with the description of Porphyrobacter neustonensis, isolated from freshwater (Fuerst 1993). All subsequent Porphyrobacter species have also been isolated from aquatic sources (hot springs, seawater, and swimming pools) including P. tepidarius (Hanada 1997), P. sanguineus (Hiraishi 2002), P. cryptus (Rainey 2003), P. donghaensis (Yoon 2004), P. dokdonensis (Yoon 2006) and P. colymbi (Furuhata 2013).

Phylogenetic and biochemical characteristics presented here show that our isolate is clearly distinct from other members within the Erythrobacteraceae family and is most closely related to the genus Porphyrobacter. However, a major taxonomic revision of this family is most likely required, as has been suggested by others (e.g. (Rainey 2003) (Huang 2015)). Here we describe the genotypic, morphologic, and biochemical characteristics of strain CoronadoT, based on which we propose the name of Porphyrobacter mercurialis sp. nov.


Cells were initially grown on plates containing either Reasoner's 2A agar (R2A), or lysogeny broth agar (LB). LB was made with 10 g tryptone, 10 g NaCl, and 5 g yeast extract per liter. A clear preference for growth on LB was observed and so was used for all subsequent experiments. Salt tolerance was measured by growth in liquid media (25 °C) from 0% to 20% w/v NaCl. pH tolerance was measured by growth in liquid media (25 °C) from pH 3.4 to pH 8.0. Temperature preference was measured by growth in liquid culture across the range 4 °C to 30 °C. Growth in microgravity (OD600) was measured aboard the International Space Station (ISS).

Cell morphology, motility, and presence/absence of flagella were examined by light microscopy (Zeiss Axio Lab.A1) and transmission electron microscopy (TEM). Cell cultures in either exponential or stationary phase were prepared for TEM by the UC Davis Electron Microscopy lab as follows. 400 mesh copper grids with formvar/carbon support film (Ted Pella, Inc., Redding, Ca.) were placed on dental wax. A 10 μl drop of fixed or unfixed sample was placed onto the grid and left in a dust-free environment for 10 minutes. Then excess was wicked off with filter paper. A 10 μl drop of 1% PTA pH 5.8 (phosphotungstic acid) or 1% ammonium molybdate in double-distilled water was added to the grid and wicked off immediately. Grids were allowed to air-dry completely before viewing in a Philips CM120 (FEI/Philips Inc, Hillsborough, Or.) electron microscope at 80KV.

Oxidase activity was measured using a solution of tetramethyl-p-phenylenediamine and catalase activity was measured by the addition of hydrogen peroxide to plated cells. The hydrolysis of starch and casein were measured by standard plate methods (beef agar with soluble starch and iodine staining, and milk agar with a pancreatic digest of casein respectively). Carbon source oxidation was assayed using the Phenotypic MicroArray (TM) services offered by Biolog, Inc. using their standard procedures for gram-negative bacteria as follows. Colonies were grown on blood agar at room temperature and suspended in IF-0a inoculating fluid (Biolog) to a density of 42% transmittance. The cell suspension was diluted 1:6 in IF-0a plus 1x Dye H (Biolog) and a carbon source utilization MicroPlate (PM1; Biolog) was inoculated with 100 μl per well. The PM1 microplate was incubated at 23 °C and read by the OmniLog instrument every 15 minutes for 96 hours. Duplicate sets of OmniLog data were converted to average read value and a threshold of 78 was required in both replicates for a positive call.

Respiratory quinones, polar lips, and fatty acids

Cells were grown to post exponential phase (~72 hours) in 1 L of LB at 23 °C for large-scale biomass production, then centrifuged to pellet cells and lyopholized (VirTis Freezemobile). Analysis of respiratory quinones/polar lipids and fatty acids were carried out by the Identification Service, DSMZ, Braunschweig. Germany. Protocol details and references can be found at the following DSMZ pages; https://www.dsmz.de/services/services-microorganisms/identification/analysis-of-polar-lipids.html, https://www.dsmz.de/services/services-microorganisms/identification/analysis-of-respiratory-quinones.html, https://www.dsmz.de/services/services-microorganisms/identification/analysis-of-cellular-fatty-acids.html.

16S rDNA, genome sequencing, and phylogenetic analysis

Genomic DNA was extracted using a Wizard Genomic DNA Purification Kit (Promega). A nearly full-length 16S rRNA gene sequence was amplified using the 27F (5'-AGAGTTTGATCMTGGCTCAG-3') and 1391R (5'-GACGGGCGGTGTGTRCA-3') "universal" primers. Sanger sequencing was provided by the College of Biological Science UC-DNA Sequencing Facility (UC Davis). This DNA was also used for Illumina sequencing of the draft genome as described elsewhere (Coil and Eisen, in press). The genome sequence was annotated using the RAST server (Aziz 2008) (Overbeek 2013).

The 1482bp 16S rDNA sequence was obtained from the genome assembly in RAST (GenBank: KP122961), and uploaded to the Ribosomal Database Project (RDP) (Cole 2013). Since the RDP database is incomplete with respect to the Erythrobacteraceae family, additional type strain sequences were obtained from NCBI to ensure that every member of the family with official standing in nomenclature (http://www.bacterio.net/) was present in the alignment downloaded from RDP. Because the taxon names exported with this alignment contained special characters that were not compatible with phylogenetic reconstruction software, a custom script was used to remove or replace those characters with underscores. A description of and link to this script can be found in (Dunitz 2015). The alignment was manually examined using MView (http://www.ebi.ac.uk/Tools/msa/mview/t), the secondary structure was generated using the RNAfold Web Server (http://rna.tbi.univie.ac.at/cgi-bin/RNAfold.cgi) and visualized with Forna (http://nibiru.tbi.univie.ac.at/forna/). This alignment was used to generate phylogenetic trees using a variety of methods including maximum likelihood (RAxML, (Stamatakis 2014)), Bayesian (MrBayes, (Ronquist 2003) (Huelsenbeck 2001)), neighbor-joining (MEGA6, (Tamura 2013), NINJA, http://nimbletwist.com/software/ninja/), and maximum parsimony (MEGA6). Dendroscope 3 (Huson 2012) and FigTree (http://tree.bio.ed.ac.uk/software/figtree/) were used to view and edit the phylogenetic trees.