1 INTRODUCTION 1.1 Inspiration from Winogradsky

Sergei Winogradsky (1856–1953) was a Russian microbiologist who revolutionized the field in the late 19^th and early 20^th centuries and perhaps invented microbial ecology (Dworkin and Gutnick, 2012). His simple techniques and diligent observations altered our understanding of the nutritional metabolisms and energy sources of microorganisms, especially photoautotrophy, chemolithotropy, and anerobic respiration (e.g., Stanier, 1951; Waksman, 1953; Dworkin and Gutnick, 2012). He also dispelled several misconceptions, including settling a scientific debate regarding pleomorphism versus monomorphism in the 1880' s (Waksman, 1953). This contention stemmed from certain species of sulfur-oxidizing bacteria mutating under sterile conditions, leading many to think, incorrectly, that simple bacteria could have multiple forms (Wainwright, 1997; Dworkin and Gutnick, 2012). To explore this, Winogradsky designed his namesake columns. By placing water, mud, and nutrients into a closed container, he could watch the absorption and effect of H₂S on the sulfur-oxidizing bacteria communities inside. He observed morphologies that were described in natural conditions, ending the debate (Winogradsky, 1888).

Much of modern microbial ecology is focused on microbial communities (e.g., Almeida-Dalmet et al., 2015; Lindsay et al., 2020) and is distinct in approach to traditional isolate-and-characterize techniques. Researchers in this field are developing an understanding of shared metabolites (Dillard et al., 2021), and community roles for individual species as opposed to how a strain behaves when separated from its environment. It is interesting that Winogradsky recognized this perspective more than a century ago, and his columns were exactly that, a study in polyculture, where communities could be analyzed while working together in compartments that mimicked the natural environment (Winogradsky, 1888).

1.2 Extremophile microbial communities of Great Salt Lake

We applied Winogradsky' s insight to a large extreme ecosystem, the hypersaline Great Salt Lake (GSL), where we had been studying microbial diversity for some time (Baxter and Zalar, 2019). GSL is a high elevation lake in the western US, bisected by a railroad causeway into two sections, the north and south arms (Fig. 1) (Baxter, 2018). The north arm is visibly pink due to aerobic, carotenoid-pigmented haloarchaea (Figs. 1 & 2a), contrasting with the greenish south arm eukaryotic microalgae and cyanobacteria (Figs. 1 & 2b). The north arm has a salinity between 25%–34% over seasons (Almeida-Dalmet and Baxter, 2020). The haloarchaea, which dominate this system, have many superpowers that allow them to thrive in not only high salinity, but also high ultraviolet radiation flux and cycles of desiccation (Bayles et al., 2020). Haloarchaea employ carotenoid pigments as a means for photoprotection (Jones and Baxter, 2017). The saturated salt concentrations of the GSL north arm limit bacteria in the oxygenated water column, but a few species (e.g., Salinibacter spp.) have been detected (Baxter and Zalar, 2019; Almeida-Dalmet and Baxter, 2020). We recognize the pink coloration is a visual feature of the GSL north arm, observed not only by microbiologists (Fig. 2a), but also by astronauts (NASA, 2019), kayakers (Larsen, 2021), artists (Smithson, 1996; Dia Art Foundation, 2022), and members of the public (e.g., The Salt Project, 2021).

Although the north arm brine is low in dissolved oxygen due to the high concentration of salt interacting with water molecules, there is a truly anoxic sediment layer that has received very little scientific attention (e.g., Boyd et al., 2017). One study previously demonstrated biogenic methanogenesis in north arm sediment microcosms (Baxter et al., 2005). Sulfate reducers have been discovered (Brandt and Ingvorsen, 1997; Brandt et al., 1999; Brandt et al., 2001; Ingvorsen and Brandt, 2002), and the rate of sulfate reduction was measured showing decreased activity in the north arm, correlated with increased salinity (Kjeldsen et al., 2007).

The south arm aerobic microbiota is more diverse (Meuser et al., 2013), with a larger abundance of eukaryotic microalgae and cyanobacteria that serve as food for the brine shrimp and fly larvae that inhabit the water (Figs. 1 & 2b) (Belovsky et al., 2011; Barrett and Belovsky, 2020; Brown et al., 2022). Microbial activities of the anaerobic compartments such as sediment and stratified deep brine layers in this region south of the causeway have been explored due to the concerning presence of deposited heavy metals (Domagalski et al., 1990; Wurtsbaugh, 2007; Naftz et al., 2008; Saxton et al., 2013; Boyd et al., 2017). Also, anaerobic metabolisms such as methanogenesis (Phelps and Zeikus, 1980; Zeikus et al., 1983; Lupton et al., 1984) and sulfate reduction (Brandt et al., 2001; Ingvorsen and Brandt, 2002; Ionescu et al., 2007; Boyd et al., 2017) have been studied in these anoxic compartments of the south arm.

1.3 Great Salt Lake Winogradsky columns: exhibits and analyses

Informal education venues typically go beyond the traditional pedagogical approaches of a classroom, and museums who engage visitors through dynamic and interactive displays have enormous potential to enhance learning (Mujtaba et al., 2018). As we collaborated to create a "Great Salt Lake Gallery" at the Natural History Museum of Utah (NHMU) (Natural History Museum of Utah, 2022), we focused the exhibits largely on the visible food web, which includes the south arm invertebrates and the ten million birds who eat them (Baxter and Butler, 2020). However, we felt the GSL hypersaline north arm, though remote and less accessible to tourists, should be represented. We considered designs that would underscore the significance of the microbiology in this extreme ecosystem and settled on creating a large Winogradsky column that would achieve this aim. This interactive would be a long-term part of the exhibit with the goal of educating museum goers about the extremophile community and would incorporate visual cues that people would be primed to recognize in the natural environment. The NHMU GSL north arm Winogradsky column (Fig. 3) features a bright reddish hue, which is impactful to visitors who learn about the haloarchaea community in that part of the lake. After incubation for several years, we designed experiments to explore the current microbial community of the column, especially to note which genera were enriched in these museum conditions.

Recently, we incorporated our lessons learned at a natural history museum to apply them in an art museum environment. The Utah Museum of Fine Arts (UMFA) planned a temporary exhibit, Confluence (Utah Museum of Fine Arts, 2021), which would highlight issues of water in Utah. The curators wanted to feature GSL as the low point of the watershed of this terminal basin was reached. Whereas the mission of NHMU and UMFA institutions may differ, our goal was similar, to educate about the microorganisms in the environment while paying attention to a visual aesthetic that would excite visitors. In addition, macrofauna in the UMFA columns created an in-person animal experience that served to convey educational information and forge an empathic connection that impacts learning and action (Moss et al., 2017). Here, we present data not only on design and construction, but also on GSL column biology and insights on how the microbial communities on exhibit may align or differ from the natural phenomena.

2 MATERIAL AND METHOD 2.1 Natural History Museum of Utah column design and exhibit implementation

NHMU opened a new facility in 2011, and the Winogradsky column was installed at that time. During the design phase and engineering, careful considerations were made to future maintenance and support of a biological sample. The column is a plastic tube inside an outer tube, which creates a hollow cavity on the inside, and only the outer tube is filled (Fig. 3a). It measures 137 cm in height and 30.5 cm in diameter and is glued to a plastic circle which serves as a base that was mounted securely to the floor. The column is surrounded by a counter with explanatory text (Fig. 3b). It was designed to be refillable so that the brine could be replaced in the future and to allow water to be added in case of evaporation. Water and sediment samples from Rozel Point at the GSL north arm were collected by museum curators and used to fill the installed column. This original column did not have a light installed and was at ambient temperature. The brine remained pale in color and the column did not enrich pigmented halophiles as desired.

In 2014, we adjusted the column design to add light and heat. Just above the sediment line we installed a 10-cm halogen bulged reflector spotlight. Inside the empty center tube, we placed a strip of 100-W lights on a metal support, which heated the tube to a consistent temperature of 37.8 ℃. We collected new samples from another GSL north arm site, the Little Valley Boat Harbor at Promontory Point, just south of Rozel Point (41.255 7°N, 112.499 1°W). To the sediment layers, we added a mixture of nutrients (including sawdust, chalk, dry milk, rice, parmesan cheese, oatmeal, ashes, shredded newspaper, eggs, gypsum (CaSO₄·2H₂O), epsom salt (MgSO₄·7H₂O), Miracle-Gro^® fertilizer (28% nitrogen, 8% phosphorus, and 16% potassium), and TUMS® antacid tablets (CaCl₂)) in quantities suggested by resources for building Winogradsky columns (Bordenstein, 2022; Genetic Science Learning Center, University of Utah, 2022). This time as the column was filled, the oolitic sand sediment was enriched and deposited in the bottom of the column in four layers, so that the nutrients were dispersed. The brine was layered on top. Within a few weeks, the NHMU Winogradsky column began to redden, and slowly the tube darkened to a deep red (Fig. 3b). Halite crystals formed at the sediment interface. Currently, our maintenance involves only replacing the evaporated water from the top access.

2.2 DNA sequencing analysis of the Natural History Museum of Utah column

Samples were collected from the NHMU column after seven years of ambient temperature growth in the exhibit. A total of 50 mL of the GSL water above the sediment was diluted 1:10 with deionized water to reduce the viscosity of the brine and then filtered through a 0.45-µm Pall MicroFunnel Filter (Show Low, AZ, USA). The DNA was extracted from the GN-6 Membrane of the filter with the DNeasy PowerWater Kit (Hilden, Germany). The extraction resulted in ~10-µg DNA from the microbial community inhabiting the column.

The DNA sample from the NHMU column was sent to MRDNA (Shallowater, TX, USA) to amplify target genes and sequence the PCR products. We amplified the 16S rRNA genes of bacteria and archaea using the 515 forward and reverse primers for the V4 variable region (Walters et al., 2016), the 18S rRNA genes of eukaryotes using EUK1391 forward and reverse primers for the hypervariable region (Rodriguez-Sanchez et al., 2019), and the Internal Transcribed Spacer (ITS) gene of fungi using ITS forward and reverse primers (Walters et al., 2016). Each reaction was subjected to PCR using the HotStarTaq Plus Master Mix Kit (Qiagen, USA) under the following conditions: 95 ℃ for 5 min, followed by 30–35 cycles of 95 ℃ for 30 s, 53 ℃ for 40 s and 72 ℃ for 1 min, after which a final elongation step at 72 ℃ for 10 min was performed. After amplification, samples were verified on a gel and multiplexed using unique dual indices, then pooled together in equal proportions based on their molecular weight and DNA concentrations. Pooled samples were purified using calibrated Ampure XP beads. The pooled and purified PCR products were used to prepare an Illumina DNA library. Sequencing was performed at MR DNA (www.mrdnalab.com, Shallowater, TX, USA) on MiSeq following the manufacturer' s guidelines. Sequence data were processed using MR DNA analysis as follows. The Q25 sequence data were processed using the MRDNA ribosomal and functional gene analysis pipeline. Sequences were depleted of primers, short sequences < 150 bp and those with ambiguous base calls were removed. The data was quality filtered using a maximum expected error threshold of 1.0 and dereplicated. The unique sequences were denoised, including removing point errors and chimera, resulting in "zOTUs". Final zOTUs were taxonomically classified using BLASTn against a curated database derived from NCBI (www.ncbi.nlm.nih.gov) and compiled into each taxonomic level by abundance. In the analysis, bacteria and archaea were binned at the level of genus, and eukaryotes were binned at level of order. We also sequenced the 16S rRNA partial gene for several species that we cultivated and isolated in the lab, using procedures previously described (Kemp et al., 2018).

2.3 Microscopy of the Natural History Museum brine

A sample of brine from the water column of the NHMU column was examined by microscopy to look at the community morphologies. After a simple staining procedure (Dyall-Smith, 2018), we examined the slides with a brightfield microscopy (Olympus BX51, Breinigsville, PA, USA). Micrographs were captured at 1 000× magnification under oil immersion with applied scale bars of 10 µm.

2.4 Pigment analysis of the Natural History Museum of Utah column

To examine the pigment composition, we performed an extraction on a 2-mL sample from the NHMU column (Oren, 2011). The aliquot was centrifuged to pellet the cells. The supernatant was removed and replaced with 2 mL of a 1:1 methanol-acetone solution then incubated in a dark room for one hour. The extract was cleared via centrifugation and subjected to spectroscopy with a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). We recorded the absorbance peaks of the supernatant from the ultraviolet to the infrared light range.

2.5 Utah Museum of Fine Arts column design, sampling, and exhibit implementation

When invited to participate in the UMFA exhibit on the local watershed, Confluence, we planned to demonstrate the salinity gradient of the GSL ecosystem. We chose three sites based on their unique properties and organisms. To this end, we designed three clear extruded acrylic cylinders (Delvies Plastics, Salt Lake City, UT, USA), with closed bottoms, for the columns (15.2 cm in diameter, 0.6 m in height) (Fig. 4a). Considering what we learned about the importance of lighting from the NHMU column design, we included full spectrum lighting on the top of each column in the form of 3 high wattage, full-spectrum lamps (GE BR30 9-watt LED; General Electric, Boston, MA, USA) that provided enough light to illuminate the entire length of the column. Bulbs were installed in wired light sockets with switches (Cable Matters, Southborough, MA, USA).

To house and secure the bulbs in place, as well as provide additional support for the column, we modeled and 3-D printed structural caps that would simultaneously anchor the column to the display base and provide an attachment for the lightbulbs. The models were created using TInkercad software (Autodesk, San Francisco, CA, USA) (Fig. 4b) and printed using fused deposition modeling (FDM) in polylactic acid (PLA) filament on Lulzbot Taz 6 3D printers (Lulzbot, Fargo, ND, USA). The top of the cap was designed to be removable in case of bulb failure, without needing to remove the entire mounting structure from the display base. The rear of each cap featured a structural element that accepted an acrylic rod that traveled the length of each cylinder, holding them structurally to the display base and helping prevent detachment of the light or loss of material if the column were bumped by museum visitors. Threaded inserts allowed the rods to be removed and replaced if needed and minimized visual distractions behind the cylinders. Cables were run down the acrylic rod and into the display base and connected to a programmable timer (BN-LINK, Santa Fe Springs, CA, USA) underneath the display base that powered all three lights on and off to recreate diurnal cycles.

After column construction, each was filled using sediment and water from three distinct sampling sites: Farmington Bay Wildlife Management Area, Antelope Island, and Rozel Point (Fig. 5). These sites varied in salinity, shore vegetation, nutrients, and microbial communities. Water and soil samples (approximately a 15-L total volume) were collected in large buckets and returned to the lab where the column assembly took place. The sediment was collected from the top 10–15 cm at the lake sampling sites. At the hypersaline Rozel Point location, halite crystals were also collected from the surface of the oolitic sand beneath the water.

Before we installed the exhibit and prior to the addition of the lamp mounts, the columns were incubated at the ambient temperature of the laboratory and placed on a windowsill to provide sunlight for approximately 3 months. We monitored the columns and noted changes in appearance. The Farmington Bay column community was not adjusted, but we did make changes to the other two to enhance both the aesthetics and the science message that was to be conveyed by the exhibit. To the Rozel Point column, we added ~50 mL of NHMU column brine via syringe to the center of the column in an attempt to redden the column (visible in Fig. 4a). To the Antelope Island column, with a 100-µL lab spatula scoop, we added brine shrimp Artemia franciscana cysts, which would hatch and form adults that are prevalent during most of the year at this GSL site but were missing from our column due to the sampling season. We also added some supplemental yeast to feed the shrimp. A packet of dried Saccharomyces cerevisiae (Fleischmann' s Active Dry Yeast) was added to 50-mL sterile water. After a 30-min incubation, 100 µL of yeast solution was added to this column.

We installed the columns, in the context of the larger exhibit on the watershed (Fig. 4c–d), with the help of UMFA exhibit design staff. The surrounding interpretive information informed visitors of the sampling locations and the salinity gradient. The columns were accompanied by a video monitor slide show of the biology of GSL coupled with recorded lake soundscapes.

3 RESULT 3.1 Microbial community of the natural history museum Winogradsky column

In the design of the Great Salt Lake Gallery at the NHMU, we wanted to show the pink color of the north arm, bringing this unusual facet of the lake ecosystem into the exhibit design. We landed on the idea of a large Winogradsky column, constructed, and maintained as described in the materials and methods. The column deepened in pigmentation over a seven-year incubation (Fig. 6), especially in response to the addition of nutrients and broad-spectrum lights. The nutrients may have enriched certain microbiota and the lights increased the temperature of the brine, perhaps stimulating growth. We became curious about what taxa we were enriching with these artificial conditions.

First, we sampled the NHMU column and surveyed the community with microscopy (Fig. 7). In a background of irregular and coccoid morphologies, we noted prevalent rods and curved rods 2–4 µm long and about 0.25 nm wide (Fig. 7). On all slides examined, we saw clumping of these rods (Fig. 7; right panel).

In addition, we inoculated haloarchaea media with the column brine and isolated strains under aerobic conditions. This resulted in the cultivation of two strains of Halorhabdus and one species of Halovenus, which we characterized with 16S rRNA gene amplification. In many other studies culturing from GSL brine or minerals, our lab had never isolated cultivars from these genera (e.g., Kemp et al., 2018). We did not cultivate any bacterial species, but we also did not use other media recipes, beyond those for haloarchaea, nor did we attempt anaerobic culturing techniques.

Of course, this cultivation-dependent approach is limiting, and we wanted to envision the entire column microbial consortia. To this end, we extracted DNA from a sample of NHMU column brine (not sediment) and used 16S rRNA gene primers to probe the bacteria and archaea taxa (Fig. 8a–b). The total number of reads for operational taxonomic units (OTUs) were 19 640 for archaea and 12 189 for bacteria. Since we used separate primer sets specific to either bacteria or archaea, we cannot compare abundance between these two categories of prokaryotes, but we can reflect proportions within each. More than 99.5% of bacterial species in the column are in the class of Gammaproteobacteria, which has a genera richness second only to Firmicutes (Williams et al., 2010). The following Gammaproteobacteria genera are represented in these percentages of the entire bacterial population: Halorhodospira (69%), Halomonas (14%), and Idiomarina (11%) (Fig. 8a). Any taxa represented at an abundance lower than 5% is not counted in the major community members and shown as "other" in the figure (Genera: Bacillus, Balneola, Halanaerobium, Halobacillusm, Paraliobacillus, Rhodovibrio, Thalassobacillus). The largest proportion of haloarchaea identified fall into seven genera with the following representation among all archaea in the column: Halorubrum (56%), Natronomonas (9.2%), Halovenus (8.8%), Halorhabdus (6.9%), Salinarchaeum (6.3%), Haloquadratum (5.5%), and Halorientalis (5.1%) (Fig. 8b). "Other" is represented by one genus, Archaeoglobus, which is less than 2% of the population.

To assess eukaryotes in the DNA extracted from the NHMU column brine, we amplified with primers for 18S rRNA genes as well as ITS spacers to make sure our results were inclusive of fungi (Anderson et al., 2003). Amplicons from the 18S rRNA gene primers indicated no evidence of eukaryotic phytoplankton (algae) despite the abundance of Dunaliella salina (Fig. 2, right panel) and Tetracystis species previously demonstrated in the GSL north arm (Baxter and Zalar, 2019). This experiment revealed two fungal orders (data not shown), but since the ITS primer amplification captured all fungi; we used this data set instead to avoid redundancy and to more accurately express abundance (Fig. 8c). The majority of the fungal taxa indicated by the amplicons were in the phylum Ascomycota. The classes observed under this phylum were Dothideomycetes (orders: Pleosporales, Mycosphaerellales, and Cladosporiales) and Sordarimycetes (orders: Microascales, Hypocreales, Coniochaetales, Glomerellales, and Melanosporales). Under the phylum, Basidiomycota, only the class Cystobasidomycetes (order: Cystobasidiales) was noted. The "other" category with fungi represented diverse orders (Cantharellales, Corticiales, Eurotiales, Hymenochaetales, Malasseziales, Saccharomycetales, and Sporidiobolales).

3.2 Pigment spectroscopy

Pigment analyses can augment and support taxonomic data (Oren, 2011), and there are a large variety of carotenoids in halophilic bacteria and archaea. Halophilic anoxygenic phototrophs, which are represented in the microbial population of the NHMU column brine, have other pigments such as bacteriochlorophylls. Since the color of the column was the feature for which we were selecting, we wanted to determine the pigment composition. We extracted pigments from the NHMU column brine and analyzed the extract by UV-visible spectroscopy. Absorbance peaks were observed at 600 nm and 760 nm (data not shown), which corresponds with results anticipated for bacteriochlorophyll a found in (Alpha-, Beta- and) Gammaproteobacteria such as the identified Halorhodospira species (Oelze, 1985; Oren, 2011). We also see an indication of absorbance at the characteristic visible light peaks (namely 470 nm, 496 nm, and 530 nm) of bacteriorubrin-enriched haloarchaea (Kelly et al., 1970; Jones and Baxter, 2017), such as the Halorubrum identified in the column community. However, the strongest signal was the 760-nm peak. This indicated an abundance of bacteriochlorophyll a that is contributing to the coloration, and it supports the genetic data suggesting Halorhodospira are abundant in the NHMU mature column community.

3.3 Microbial communities along a salinity gradient: Utah Museum of Fine Arts Winogradsky column collection

We were invited to participate in Confluence (Utah Museum of Fine Arts, 2021), an exhibit highlighting the northern Utah watershed, and we wanted to use Winogradsky columns to bring the biology into the art museum. Applying lessons that we learned studying the NHMU column, we designed a trio of smaller columns (Fig. 4d) to represent various sites around GSL with varying salinity: Farmington Bay Wildlife Management Area, the north tip of Antelope Island, and Rozel Point in the north arm of the lake (Table 1; Fig. 5). The field sampling occurred in the autumn, and at that time, the columns were constructed as described in Materials and Methods. The salinity was measured in the field with a refractometer and ranged from 28 to 320 (Table 1). The columns were incubated in the lab on a windowsill for three months before installation, and we altered them as described below and recorded observations during that time (Table 1). Because lighting had proved so important in the NHMU column exhibit, we were concerned about the lack of natural sunlight in the art museum. We thus designed lighting mounts for the top of each column, which put a full spectrum light at the top as a part of the lid (Fig. 4c).

3.3.1 Farmington Bay column

Beginning construction in 1935, US federal agencies installed dams upstream of the GSL basin to create the Farmington Bay Wildlife Management Area with the goal of producing bird habitat (Hoven, 2012). This created a brackish bay to the east of Antelope Island in GSL. Due to the lower salinity, relative to the rest of the lake (and relative to ocean salinity of about 35), this area possesses a unique microbiology, and is prone to cyanobacterial blooms (Marcarelli et al., 2006). At Farmington Bay, large birds are frequently observed eating the fish that thrive here such as carp, and the banks are covered with Phragmites and other grasses (Kettenring et al., 2020). Therefore, this ecosystem would represent a very different environment than the sites with higher salinities.

The Farmington Bay column contained soils, water, and detritus from the pond we sampled (Fig. 5). A few days following lab incubation, we noted that bladder snails (family Physidae) were emerging from the sediment (Fig. 9) as well as tiny crustaceans, "scuds" (order Amphipoda) (Schoch et al., 2020). This colony of invertebrates grew and multiplied until the column was placed in the UMFA exhibit. Within two weeks, the sides of the column and the soil-water interface were coated with green biofilms, and the water column contained visible phytoplankton. After installation, under the glow and warmth of the broad-spectrum light, these photosynthesizers continued to thrive. Their invertebrate companions experienced a population shift; we noted a decrease in snails and an increase in scuds. This observation provided insight to another use of Winogradsky columns in monitoring macrofauna populations over time.

3.3.2 Antelope Island column

We sampled Antelope Island on the north tip of the island (Fig. 5) where Baxter and others have previously studied the geobiology of microbialites, calcium carbonate structures which were formed by the action of microbial mats (e.g., Lindsay et al., 2017). Over years of sampling this site, the salinity has varied from 80 to 130, vacillating locally as wind events can assist inputs from a freshwater source nearby. We did not include mats or microbialite samples in our column, only water and oolitic sand, but we expected the sample to be rich in phytoplankton, and the water at the site was bright green (Fig. 5). In fact, we observed a green biofilm forming on the top of the water column in the first week of windowsill incubation and a microbial mat forming on the surface of the sediment (Fig. 10a). After a few weeks, the column formed visible layers that differed in color, possibly a halocline related to the lack of mixing, indicating stratification of brine by density.

Since brine shrimp Artemia franciscana are an inhabitant of this part of the lake (Marden et al., 2020), and because we thought these would excite museum visitors, we added both adults and encysted embryos from an aquaculture shrimp tank in the lab after the column had been incubated on the windowsill for approximately three weeks. They died before the installation and had to be replenished. With the second Artemia inoculation, we added some Saccharomyces cerevisiae for shrimp food. After installation in the museum, the Artemia population thrived. We observed these filter feeders grazing on the green phytoplankton film at the air interface (Fig. 10c) and from the mat growing on top of the sediment (Fig. 10d).

3.3.3 Rozel Point column

At the time of sampling, the north arm GSL brine at the site was at salt-saturation and had a rosy pink color (Table 1 & Fig. 5). In the Rozel Point column, the water was only tinged pink and did not convey the visual impression of this part of the lake. To accentuate the pink pigmentation, we added a 50-mL spike of fluid from the NHMU column described above (Fig. 6) with a sterile syringe to the top of the column. Within a week of incubation on a windowsill in the laboratory at ambient temperature, the vibrant pink microorganisms had migrated toward a more central location in the column (Fig. 11a). Before installation, this part of the microbial community floated as a 6 cm pink band of color about 6 cm below the top of the water column and 9 cm above the sediment. After installation, phytoplankton began to accumulate and form green mats, which have not been reported in the north arm of the lake, near the air interface and beneath the lamp (Fig. 11b–c). Also, the purplish-pink coloration moved slowly toward the bottom of the brine and took on an orangish-pink hue, suggesting a shift in carotenoid biosynthesis pathways or a shift in the most prominent members of the microbial community (Fig. 11d). As these UMFA columns are still on exhibit, we have not done any analysis of the consortia.

4 DISCUSSION

The grand arc of microbiology illustrates a balance between the values of purity and diversity (Grote, 2018). The last century of studies of microorganisms was propelled by isolation and characterization of individual species, which gave us insight into various lifestyle strategies of microorganisms, including pathogens. Modern genomics and microbiome studies, however, underscore the significance of microbial communities. It is thus satisfying to see that the pioneer, Sergei Winogradsky, so long ago, developed a column technique that implemented polyculture instead of monoculture and preserved the natural microbial consortia in the laboratory setting. Winogradsky columns allow for the exchange of metabolites between species and may more closely mimic these processes as they occur in nature; they include both a sediment compartment and a water column. Community structure and species abundance are more likely to be impacted by intrinsic components rather than extrinsic (Daniel, 2005; Bowen et al., 2009). For this reason, co-culturing is an important technique, even in applied science (Peng et al., 2016).

To better understand this phenomenon of shifts in abundance of community members over time, we analyzed pigments in the NHMU column, and we constructed a profile of the community members with DNA sequencing and analysis. Although we did not do this at time zero, when we set the column up, in this study, we assumed that the original brine and sediment consortia reflected that of the natural environment from which it was removed (Almeida-Dalmet and Baxter, 2020). Therefore, our analysis considered how the mature column is different from the natural system. We do not include any such analysis of the UMFA columns as they are much younger and are still on exhibit.

The goal of building the NHMU Winogradsky column was to put on display a thriving haloarchaeal community, but the data showed a limited number of these species evident in our analysis. Also, the haloarchaea taxa abundance measured in the mature NHMU column are in contrast with that of the natural environment (Bayles et al., 2020). One caveat, in filtering the dense brine from the column, we had to dilute the sample, which could have lysed some haloarchaea if the molarity dropped below 1.5 mol/L (Schneegurt, 2012), therefore, they may be underrepresented. Our cultivation of haloarchaea from column brine resulted in genera not seen before in prior GSL work; this was the first signal that the community was distinct from the lake. The 16S rRNA gene profile analysis expanded our view of the consortia. The largest proportion of haloarchaea identified in the NHMU column community fall into seven genera that are commonly identified in other hypersaline environments (e.g., Xin et al., 2000; Oren, 2002; Cui et al., 2006, 2007; Oh et al., 2010), in order of abundance: Halorubrum, Natronomonas, Halovenus, Halorhabdus, Saliniarchaeum, Haloquadratum, and Haloorientalis (Fig. 8b). During a three-year temporal study of north arm microbial communities, Halorubrum species were far less abundant than represented in this column; the highest value measured approached 25% during one summer sampling (Almeida-Dalmet et al., 2015) compared with 56% of the NHMU haloarchaea. Species in several genera were overrepresented in the NHMU column archeal profile, but not detected at all in the lake analysis: Natronomonas, Halovenus, Saliniarchaeum, and Haloorientalis. These data suggest that the selection pressures of the column conditions, for example depleted nutrients and oxygen, amplified microorganisms that were previously underrepresented, and they became dominant members of the mature column consortia. Although haloarchaea have a host of carotenoids associated with their membranes (Jones and Baxter, 2017), our pigment analysis did not see significant absorbance at these wavelengths, suggesting the pink coloration was not due to haloarchaea alone.

The dominant bacterial genera discovered with our phylogenetic analysis of the NHMU column included Halomonas (Fendrich, 1988; Sorokin and Tindall, 2006) and Idiomarina (Choi and Cho, 2005) (Fig. 8a). These species are common aerobic bacteria in the brine of GSL (Almeida-Dalmet et al., 2015) and other salt lakes, such as Lake Urmia (Vahed et al., 2011). However, considering past studies on bacteria in the north arm of GSL (reviewed in Baxter and Zalar, 2019 and Almeida-Dalmet and Baxter, 2020), the most interesting finding in the column analysis was the overabundance of Gammaproteobacteria, especially the genus Halorhodospira, which was represented at 69% of the bacterial population (Fig. 8a). When contemplating how the NHMU column differs from the actual GSL ecosystem, the dominance of this genus should be noted as it was not detected in prior north arm sampling that measured abundance or species over time (Almeida-Dalmet et al., 2015; Almeida-Dalmet and Baxter, 2020). Importantly, this prior sampling of brine assessed the aerobic compartment of GSL. Dissolved oxygen is low in this hypersaline surface water, but oxygen is replenished by wave action (Belovsky et al., 2011). The NHMU column remained capped and stationary without introduced oxygen or disturbance and thus, it likely reached the threshold that would enrich anaerobic species and not support obligate aerobes. Halorhodospira are anaerobic purple sulfur bacteria that are proficient at osmophily (Fendrich, 1988; Imhoff, 2001) and at low-light-energy photosynthesis (Frigaard and Dahl, 2008; Kimura et al., 2021). These bacteria do photolithotrophic growth and use sulfide or elemental sulfur as electron donors (e.g., Imhoff and Trüper, 1981; Then and Trüper, 1983). They are found as reddish mats in the sediment of hypersaline systems (Hirschler-Réa et al., 2003). In GSL, this genus was previously detected in one study that combined multiple water samples across the north and south arms of the lake to produce a clone library (Tazi et al., 2014). Perhaps in some stagnated areas that the authors sampled, Halorhodospira was flourishing. When others examined lake sediments (not water) across a salinity gradient, the north arm samples contained 30% Gammaproteobacteria (Boyd et al., 2017). The cells we see in the micrographs (Fig. 7) are consistent in size and shape with Halorhodospira images published by others (Hirschler-Réa et al., 2003). All of these results taken together suggest that in the NHMU column, the Halorhodospira were in the collected sediment originally, and they migrated into the brine as it became anaerobic over time. This model of column maturation is confirmed by the lack of detected aerobic phytoplankton, as no 18S rRNA gene signal resulted for eukaryotic phototrophs. This is consistent with observations of others that suggest a column microbial community is structured from founder effect followed by enrichment based on anaerobic conditions, thus forcing the increased abundance of phyla such as Proteobacteria (Rundell et al., 2014).

Could the enrichment of pink coloration be due to the overabundance of Halorhodospira? In addition to their characteristic sulfur metabolism, this genus is also known to be obligately photosynthetic, utilizing bacteriochlorophyll pigments (Imhoff, 2001; Oren, 2011). Our pigment analysis did indicate the presence of bacteriochlorophyll a, which supports the overrepresentation of Halorhodospira. In addition, in salt-saturated environments such as salterns or the GSL north arm, the dry weight of pigments in bacteria exceeds those in archaea by a factor of three (Khanafari et al., 2010). Since the anaerobic Halorhodospira functions in the oxidation of sulfide, examining sulfate metabolism in the column would be an interesting experiment to re-imagine for the future, but we suggest it be accomplished under anaerobic conditions.

Before the DNA sequencing data was available for the NHMU column, we built the salinity gradient columns for the other museum. In an effort to speed up the pink pigmentation of the Rozel Point UMFA column, we inoculated it with pink brine from the NHMU column. Since the inoculum was pigmented, we visually tracked its migration away from the water-air interface and into a deeper part of the brine (Figs. 4a & 11a). This observation, that the vivid pink community sought a location with less oxygen, is consistent with an anaerobic consortium being responsible for the coloration. The UMFA Rozel Point column, despite being spiked with an NHMU brine sample, has not yet achieved the vibrancy of color (Fig. 11b) seen in the NHMU column (Fig. 6), but if we maintain this column over years, we suggest it may mature in similar ways.

In our community analysis, we wanted to include fungi as it is becoming clear that these eukaryotes are a part of the north arm community, even if we do not yet understand their role. A recent study cultured and identified 32 fungal species from Rozel Point water and oolitic sediment (Baxter and Zalar, 2019). Our DNA sequencing analysis of the NHMU column indicated a broad array of fungi, from two distinct phyla and three different classes (Fig. 8c). In the context of the shift to anaerobic conditions, the fungal community remains diverse. This suggests an ecological role, not only in the column, but in the lake itself in micro-niches that shift to anaerobic conditions. Also, we echo the pleas of mycologists to close the "fungi gap" (e.g., Zhang et al., 2015); researchers of hypersaline communities like GSL should include fungal specific primers in when amplifying DNA for sequencing less they miss this segment of the microbial consortia.

Winogradsky columns installed in public exhibits serve as models of natural phenomena (McGenity et al., 2020), which through visual representations communicate scientific ideas to visitors. Models play important roles in engaging science learners, encouraging observation, inquiry, prediction, and other science skills (Oh and Oh, 2011). However, models are limiting, and another level of learning is achieved when one begins to dissect how the model is distinct from the phenomenon. Our work presented here suggests that the sealed NHMU column may have reached a point where the microbial community was no longer representative, and yet it still served the purpose of characterizing the ecosystem' s pink water and thus eliciting inquiry and insight from visitors.

The visual language of science can communicate concepts in impactful, artistic ways. A public display of Winogradsky columns in Barcelona engaged visitors to the exhibit in the experimentation with nutrients that shifted communities under anaerobic versus aerobic conditions via two separate recipes, including sulfate/sulfide compounds to select for anaerobic communities (Urmeneta and Duró, 2011). The results were strikingly different visuals and viewers were inspired to ask why. When putting science on display in public exhibits, the aesthetics are important. This too calls to Winogradsky, who left science to pursue two years in his own purely aesthetic endeavors as an accomplished pianist, but missed "true inquiries, " and returned to invent his columns (Waksman, 1953). Indeed, Winogradsky columns may be seen as living works of art (Wightman, 2022) that can entice visitors to gaze more deeply, inspiring the inquiry that Winogradsky sought and using aesthetic as a tool for engagement to demonstrate the microbial splendor of extreme environments.

5 CONCLUSION

Winogradsky columns are unique representations of microbial consortia and a fascinating way to observe changes that may occur over time. In exhibiting the microbiota of the extreme environment of the salt-saturated north arm of GSL, we sought to display the robust pigmentation of haloarchaea. Instead, our first column at the NHMU taught us much about shifting microbial communities. The lack of mixing, nutrient addition, and steady shift toward anaerobic conditions selected instead for Gammaproteobacteria, especially the genus Halorhodospira, which are also pigmented. Although these conditions restricted eukaryotic phytoplankton, fungi were found in rich abundance. The purpose of our second exhibit at the UMFA was to show the diversity of life across a salinity gradient. The analysis of these columns again demonstrated the artificial museum conditions, but also this experiment gave us the opportunity to consider macrofauna in the system. Such analyses of Winogradsky columns provide the museum visitor, as well as the scientists, an opportunity to observe an experiment in progress and to reflect on the natural ecosystem under study.

6 DATA AVAILABILITY STATEMENT

All data are available upon request and are maintained and stored through Great Salt Lake Institute at Westminster College. DNA sequence information from the NHMU column biosample is available in the NCBI GenBank Sequence Read Archive (SRA) database, "GSL_NHMU_ Winogradsky_Community" under the BioProject accession number: PRJNA847570.

7 ACKNOWLEDGMENT

The authors of this manuscript include artists, museum exhibit designers, and scientists, and we share a mutual passion for interdisciplinary approaches to a study such as this one. We are indebted to those who worked alongside us in lake sampling, especially the students at Great Salt Lake Institute. At UMFA, we are indebted to the Confluence team: Annie Burbidge REAM, Sarah PALMER, Virginia CATHERALL, and June MACDONALD. Dr. Aharon OREN served as an informal advisor on this project, and we are particularly grateful for his insights on bacteriochlorophylls.