Culture of attached and suspended Rhodopseudomonas faecalis in the presence of decomposing fish feed

Abstract An approach to culturing attached and suspended forms of Rhodopseudomonas faecalis by using compound fish feed with tap water in transparent containers is reported in this study. The ratio of fish feed to tap water was 14.3–50.8 g/L, and no other inoculum or substances were added during the culture process. When the ratio of fish feed to tap water was 14.3 g/L, the highest total nitrogen, total phosphorus, and total dissolved carbon content recorded in the water in the containers were approximately 730 mg/L, 356 mg/L, and 1,620 mg/L, respectively, during the process of feed decay. Comamonas, Rhodopseudomonas, and Clostridium successively dominated during the culture process. Rhodopseudomonas was the most common dominant genus in both the attached and suspended forms when the water was dark red, and the relative operational taxonomic unit abundance reached 80‒89% and 24.8%, respectively. The dominant species was R. faecalis. The maximum thickness of attached bacteria and the biomass of attached Rhodopseudomonas reached up to 0.56 mm and 7.5 mg/cm2, respectively. This study provides a method for the mass culture of Rhodopseudomonas by using the fermentation of aquatic compound fish feed.

treatment of organic wastewaters from the sugar, food processing, chicken abattoir, dairy, and fermented starch industries. Some reports have indicated the dominance of PNSB in both the suspended and attached forms. Rhodopseudomonas species are found in various environments including limnetic zones, marine habitats, and sewage sludge, which are particularly abundant in eutrophic environments (Wei, Okunishi, Yoshikawa, Kamei, & Maeda, 2016;Zhang, Yang, Huang, Zhang, & Liu, 2002). Mixed communities of anoxygenic PPB and Chlorella-type green algae have been reported to dominate leachate sewage ponds, which are typical hypereutrophic red water ponds (Chandaravithoon, Nakphet, & Ritchie, 2018). Some pink microbial mats have been found to naturally occur in a swine wastewater ditch, in which the microbial mats were dominated by the genera Rhodobacter and Rhodopseudomonas, and the Rhodopseudomonas species was more dominant than the Rhodobacter species (Okubo, Futamata, & Hiraishi, 2006). A continuous photo-anaerobic membrane reactor started with raw domestic wastewater has been found to change to purple with PPB dominance during culture, and Rhodocyclus or Rhodopseuodomonas and Rhodobacter or Rhodobacter dominance have been found to occur during different culture stages (Hülsen et al., 2016).
Before the experiment reported in this paper, we attempted to obtain organic and inorganic nutrients from fish feed decomposition to study Microcystis blooms. In 2014, fish feed and tap water were placed in tanks without aeration. After slow degradation during the first few weeks, the fish feed remained relatively stable, as the dissolved oxygen in the water was consumed, and the feed remained unchanged in an almost anaerobic environment. To promote the decomposition of fish feed, we used aerated tanks starting from January 2015. During warm periods in 2015, we observed some interesting phenomena in the tanks. Some dark red material attached to the inner walls of the tanks, and the water in the tanks also gradually became dark red. Analysis of the dark red material and the dark red water revealed that the color was due to the dominance of Rhodopseudomonas. On the basis of these observations, we carried out further experiments to test and verify the ability of Rhodopseudomonas to be cultured in the presence of fish feed after decomposition produces sufficient nutrients. Therefore, the aim of the current study was to determine (a) whether the dark red microbial growth in the presence of decomposing fish feed was a random occurrence; (b) whether the culture method could be replicated; and (c) if the phenomenon was a regular occurrence, then what microbial changes occur during the cultivation process, and what factors stimulate these changes.
We reasoned that if the results of this study indicated that both attached and suspended dark red microbial growth could be achieved by using compound fish feed decomposition in transparent containers, then this research might provide a technical reference for the use of organic waste such as livestock manure, concentrated cyanobacterial blooms, and food waste. These waste products could be converted to high densities of Rhodopseudomonas, and the bacteria could then be used in aquaculture water quality control, animal feed additives, or hydrogen production. Eight transparent containers (TK1-TK8) were filled with tap water and fish feed for the culture process; the details are provided in Table 1. The ratio of fish feed to tap water was 14.3-50.8 g/L.

| Experimental setup and management
During the incubation, no other substances or strains were added.
The compound fish feed was composed of high-quality fish meal, soybean meal, animal protein, yeast powder, flour, calcium dihydrogen phosphate, trace elements, vitamins, and amino acids; the content of crude protein, calcium dihydrogen phosphate, and trace elements was 35.0%, 1.0%, and 0.3%, respectively.
The process in two of the eight tanks (TK1 and TK2, 60 cm × 35 cm × 40 cm) is described below. On 1 December 2015 (winter in Shanghai, China), 1.0 kg of compound fish feed was added to TK1 and TK2. Tap water was then added to the tanks, to a total volume of 70 L.
Tanks TK1 and TK2 were heated to 33-35°C with a 300 W heating rod from the start of the experimental period until day 64 (3 February 2016). During the culture process, the heating rod became damaged by the growth of fouling organisms. After the old heating rod stopped functioning properly, it was replaced with a new heating rod. From day 64 onward, heating was stopped, and the culturing process continued at ambient temperature.
Tanks TK1 and TK2 were continuously aerated with an air stone, which was set at the bottom of each tank. The aeration strength was adjusted to avoid any foam overflow in the tank. Aeration provides air, particularly oxygen, which is required for fish feed decomposition.
During the culturing process, tap water was replenished approximately every 3 days to restore the total volume in the tank to 70 L.
The culturing process in TK3-TK8 was very similar to that for TK1 and TK2, and was performed in the same greenhouse under the same ambient temperature. Further details are presented in Table 1.

| Water quality sampling and analysis
During the culturing process, water samples were collected for determination of total nitrogen (TN), total phosphorus (TP), and total dissolved carbon (TOC) from TK1 and TK2 between 09:00 and 10:00 approximately every 15-20 days. Measurements of TN and TP were performed according to the methods of Gross and Boyd (Gross & Boyd, 1998), and TOC content was determined with a TOC analyzer (MultiN/C UV, Analytik Jena Inc.). The pH values, dissolved oxygen (DO), and salinity were measured in situ in the middle of the tanks with a handheld multiparameter meter (Professional Plus, YSI).

| Microbial sample collection
During the experiment, similar phenomena occurred in tanks TK1 and TK2. However, the water color in TK1 changed faster than that in TK2; therefore, microbial samples were collected in tank TK1.
Suspended bacterial samples were collected in TK1 on days 10, 34, 70, and 142, and the attached bacteria were sampled on days 55 and 70 when the inner wall was dark red or red. The water color in TK1 was dark brown with a reddish tinge on day 70 and became dark brown on day 142. A 0.22-μm filter (Millipore) was used to collect the bacterial cells.

| Illumina MiSeq sequencing
Purified amplicons were pooled in equimolar amounts and subjected to paired-end sequenced (2 × 300) on an Illumina MiSeq platform (Illumina) according to the standard protocols by Majorbio Bio-Pharm Technology Co. Ltd.

| Processing of sequencing data
Raw fastq files were demultiplexed, quality-filtered with trimmomatic, and merged with FLASH with the following criteria: (a) The reads were truncated at any site receiving an average quality score <20 over a 50 bp sliding window. (b) Primers were exactly matched, allowing 2 nucleotide mismatches, and reads containing ambiguous bases were removed. (c) Sequences whose overlap was longer than 10 bp were merged according to their overlap sequence.
According to cluster information, alpha diversity parameters referring to species richness (Chao), community diversity (Shannon index), and sequencing depth (Good's coverage) of the bacterial community were analyzed from 16S rRNA gene amplicon sequence data with MOTHUR software (http://www.mothur.org).

| Microbial biomass analysis
Microbial biomass analysis was conducted on the microbial samples collected in tank TK1, including the suspended samples from day 70 and attached samples from days 55 and 70. The microbial biomass was expressed as wet weight and dry-cell weight. Wet weight was changed to dry-cell weight by drying in an oven at 100°C for 12 hr (Kim & Lee, 2000).
For suspended cell biomass analysis, the conversion to wet weight biomass was based on the assumption that 1 mm 3 of volume was equivalent to 1 mg of wet weight biomass. Bacterial volumes were calculated from cell density and cell size measurements.
At least 50 bacterial units were measured to obtain the mean cell size. The number of dark red bacteria was calculated in 0.1 ml counting chambers with two replicate aliquots, which were observed at ×1,000 magnification by light microscopy with oil immersion objective and analyzed through phytoplankton biomass counting methods (Wang, Qin, Gao, & Paerl, 2010).
In attached cell biomass analysis, a specific area of the attached bacterial layer was scraped down to obtain the wet weight. For this process, a thick area with a smooth surface was selected. Care was taken to ensure that as few crystallized salt grains as possible were mixed during this operation.

| Inorganic salt crystal collection and analysis
During the culture, some clear and colorless crystals appeared both on the inner walls of the culture tanks and the heating rods. The inorganic salt that had crystallized on the heating rod in TK1 was collected on 27 January 2016 to perform elemental analysis through an Axios X-ray fluorescence spectrometer (PANalytical B.V.). The thickness of the attached bacteria was measured with Digimatic calipers (Mitutoyo) after some water was discharged from the tanks.

| Observations during the culture process
The water color changed as the fish feed decomposed. The variations in water color in each experimental tank were similar. As the fish feed decayed, the water gradually became pale yellow and odorous; then, the water gradually became pale gray before becoming brown or black brown. Subsequently, the water became reddish brown or even dark red, and the inner walls of the containers became dark red before the water inside the tank became dark red. After becoming dark red, the water gradually became dark brown and then light brown.
Similar results were achieved in all containers, with dark red material becoming attached to the inner walls and the water then turning dark red. When the inner walls were dark red, the thickest attached layer of the inner walls was approximately 0.5-1.0 mm in containers TK1-TK8. Figure 1 shows photographs of containers TK1-TK8 with the attached dark red bacteria on their inner walls. Rhodopseudomonas faecalis was the main Rhodopseudomonas species in both the suspended and attached samples.

| Microorganisms
The dark red bacteria were dominated by rod-shaped cells. The dark red cells were 0.5-0.8 μm in width and 1.0-2.5 μm in length.
Microbial biomass analysis indicated that the cell density of the suspended dark red bacteria on day 70 was 1.983 ×1.0 12 , with a dry-cell weight of 0.056 g/L, and the dry-cell weight of the thickest layer on days 55 and 70 was 6.86 and 7.50 mg/cm 2 , respectively (Table 3).

| Water quality in TK1 and TK2
As seen in Figure 4, the pH value on day 10 was approximately 3.7 and increased to approximately 7 on day 40, 8 on day 80, then subsequently increased to approximately 8.5. The DO was nearly 0 mg/L from the start of the culture period until day 83, after which it increased to 0.14-0.6 mg/L. The salinity was approximately 2.5 % on day 10, increased to 3.0‰ on day 42, and subsequently decreased.
The TN, TP, and TOC levels in TK1 and TK2 (Table 4) were highest on day 57, and those in TK1 were lower than those in TK2. The highest TN, TP, and TOC values in TK2 (on day 57) were approximately 730 mg/L, 356 mg/L, and 1,620 mg/L, respectively.
Some inorganic salt was crystallized on the inner walls and the heating rod of TK1 ( Figure 5). X-ray fluorescence spectrometry analysis indicated that the main elements present in the salt were O, P, Mg, C, and N (Table 5). However, the light element hydrogen cannot be measured by this method, and substantial amounts of compounded hydrogen were likely to be present in the salt crystals.
The crystallized inorganic salt was able to partially dissolve in water, and some ammonium and phosphate radicals were measured in the dissolved solution, thus indicating that the salt may have contained more than one compound. On the basis of some calculations using the elemental composition of the crystallized inorganic salt (  Figure 1 shows that containers TK1-TK8 with different fish feed  . C. denitrificans is a Gram-negative, oxidase-, and catalase-positive bacterium with efficient denitrifying ability, which can reduce nitrate to nitrogen gas (Gumaelius et al., 2001). C. denitrificans 123 is found in wastewater treatment systems, and it has been reported to form pure culture biofilms in four different culture media (Andersson et al., 2008). Pseudoxanthomonas is a Gram-negative, aerobic, heterotrophic bacterium, which can degrade cellulose and lignocellulose (Kumar et al., 2015). Pseudomonas is ubiquitous in soil and water, and Fusibacter is a minor genus within the Firmicutes phylum, which mainly comprises four Gram-positive species with valid published names (Ben et al., 2012). Fusibacter was first recorded with the discovery of the thiosulfate-reducing bacterium Fusibacter paucivorans (Ravot et al., 1999). One of the species, Fusibacter tunisiensis, has been used to treat olive-mill wastewater (Ben et al., 2012).  Oxygen is an important factor affecting the occurrence of PPB.

Rhodopseudomonas faecalis dominance in both the
Rhodopseudomonas palustris has been reported to be much more tolerant of oxygen than previously believed (Chandaravithoon et al., 2018). Some PNSB species can oxidize under light, microaerobic, or aerobic conditions (Garrity, Boone, & Castenholz, 2001). In the current experiment, the culture tanks were continuously aerated with air stones to improve the feed degradation, although the measured DO was nearly 0 mg/L during the initial 90 days (Figure 4). The aeration process can cause part of the feed material to become suspended in the water column, thus enabling the feed material to come into contact with the aerated oxygen. The dominance of denitrifying C. denitrificans on both days 10 and 34 (Figure 3) showed that some oxygen was already present in the system. With the degradation of the organic matter from the fish feed, the process of organic decay consumed less oxygen, resulting in a gradual increase in DO levels in the water (Figure 4).
Water temperature can affect the growth of Rhodopseudomonas.
In this experiment, the time for attached Rhodopseudomonas to be- Notes: The weight unit "g/L" is for suspended bacteria, and "mg/cm 2 " is for attached bacteria. and the dominance of Rhodopseudomonas occurred more quickly than that in other tanks (Table 1). This result indicates that suitable high temperatures are conducive to the rapid growth of the bacteria.
Light is also an important factor affecting the PPB community.
PNSB are photoheterotrophs capable of photoautotrophic nutrition and can be grown effectively in the dark (Madigan & Jung, 2009 found that some strains of Rhodopsedumonas such as R. faecalis (Zhang et al., 2002) and R. palustris-17 (Dönmez, Öztürk, & Çakmakçi, 1999) are unable to grow in oxygen-rich, illuminated environments.
In the current experiment, the attached Rhodopseudomonas grew on the inner wall of the container, where the frame of the greenhouse decreased the light intensity by approximately 50%, demonstrating that the attached bacteria were photoheterotrophic. The containers in this experiment were transparent, and the provision of light is an important factor in achieving growth of attached Rhodopseudomonas.
Previous investigations have shown that substrate loading is a key parameter in determining biofilm structure and function (Wijeyekoon, Mino, Satoh, & Matsuo, 2004). Peyton (1996) has found that Pseudomonas aeruginosa biofilm thickness increases with the substrate loading rate. Substrate and hydrodynamic conditions significantly influence the structure, density, thickness, and substrate conversion rate of the resulting biofilm (Wäsche, Horn, & Hempel, 2002). R. faecalis strain A isolated from swine sewage wastewater has been found to grow in twice diluted pig farm effluent of 4,575 mg/L TOC and 693 mg/L ammonia nitrogen, in which it enters the logarithmic growth phase after 48 hr of incubation (Wei et al., 2016). Rhodopseudomonas faecalis PA2 has also been cultivated in sterilized domestic wastewater of approximately 176 mg/L total Kjeldahl nitrogen and 121 mg/L total phosphate, with inoculation of 10% in different light intensities and agitation speeds (Saejung & Ampornpat, 2017). The bacteria grow rapidly within 10 days (Saejung & Ampornpat, 2017) but do not form a biofilm on the inner walls of the containers during that time.
Research (Hülsen et al., 2016) similar to the current study has been carried out, in which domestic wastewater was collected and used as reactor feed, and a minimum of 200 mg/L of ethanol was supplied, with no inoculum addition. In that experiment (Hülsen et al., 2016), the reactor biomass color became purple within 22 days, and Rhodocyclus was replaced with Rhodopseuodomonas and Rhodobacter from day 115. The Rhodopseuodomonas abundance was 60% at day 193 (Hülsen et al., 2016), and during the culture process, the concentration of total Kjeldahl nitrogen and total phosphorus in the wastewater was 60 ± 11-63 ± 8 mg/L and 8.6 ± 1.8-8.7 ± 1.6 mg/L, respectively (Hülsen et al., 2016). However, there was no clear biofilm formation on the reactor walls. In the current study, the highest levels of TN, TP, and TOC in TK1 and TK2 were approximately 730 mg/L, 356 mg/L, and 1,620 mg/L, respectively (Table 4), values much higher than those reported by (Hülsen et al., 2016). Wijeyekoon et al. (2004) (Fernández, Díaz, Amils, & Sanz, 2008). Biofilms with high cell densities can be formed in batch biogas reactors, both at high and low organic loading rates, with anaerobic digestion to digest organic waste (Langer, Schropp, Bengelsdorf, Othman, & Kazda, 2014). However, the dominant bacterial species were not reported in Langer et al. (2014). The natural attachment methods occurring during the formation of biofilm promote maximum cell viability and biochemical activity after attachment (Robledo-Ortíz et al., 2010).
Improving surface attachment for bacteria is considered to be a reasonable approach for photofermentation H 2 production with R. faecalis RLD-53 . Increasing the biomass of Rhodopseudomonas has been the goal of many studies, because biomass is very important in the application of Rhodopseudomonas, including the production of carotenoids, animal feed additives, or hydrogen. In this experiment, the cell density and biomass of suspended Rhodopseudomonas on day 70 (Table 3) were lower than that reported in several previous studies (Kim & Lee, 2000;Saejung & Ampornpat, 2017). Kim and Lee (2000) have reported that the maximum number of viable cells is 2.65 g/L of dry-cell weight in batch fermentation with a modified MYC medium. Maximum biomass production can be as high as 33.9 g/L in domestic wastewater (Saejung & Ampornpat, 2017). However, the thickness of the attached bacterial layer in the current study was greater than those studies Liao, Zhong, Zhu, Huang, & Chen, 2015;Wang, Liao, Wang, Zhu, & Li, 2011), although the culture process was longer in the current study. Wang et al. (2011) have reported that the dry weight and thickness of Rhodopseudomonas biofilms are approximately 1.0 mg/ cm 2 and 0.02 mm, respectively. The highest value for the dry weight of the biofilms has been reported to be approximately 0.8 mg/cm 2 (Liao et al., , 2015, and the thickness of the biofilms has been reported to be as high as 0.18 mm in a synthetic medium (Liao et al., 2015). Therefore, the methods used in the current study produced an increased biomass of attached Rhodopseudomonas.
Photosynthetic bacteria are highly metabolically diverse.
Rhodopseudomonas can use a wide range of compounds such as sugar, low fatty acids, organic acids, and mixed alcohols as carbon sources (Basak, Jana, & Das, 2016), and some Rhodopseudomonas spp. are able to grow anaerobically in the light or aerobically in the dark with many carbon sources and electron donors (Zhang et al., 2002). Rhodopseudomonas palustris strain RP2 is capable of mineralizing diesel range hydrocarbons in anaerobic environments (Hülsen et al., 2016). Although the composition of the materials from the degradation of compound fish feed was not analyzed in this study, the pH values at the early stage of degradation may be as low as 3.7, and those during the initial 40 days were always less than 7.0 ( Figure 4), thus indicating that some acids were produced during the degradation of fish feed. Similarly, the presence of oxygen allows for the proliferation of acetic acid bacteria to produce high concentrations of acetic acid during water kefir fermentation (Hülsen et al., 2016), whereas high concentrations of short-chain fatty acids promote the dominance of PNSB (Hiraishi, 1991;Okubo et al., 2005).
The bioavailability of lower fatty acid concentrations in wastewater is a key factor allowing the formation of visible microbial mats with dominance of Rhodobacter and Rhodopseudomonas: Members of the genus Rhodopseudomonas have a wider spectrum of lower fatty acids utilization than those of the genus Rhodobacter (Okubo et al., 2006).
The composition of the synthetic culture medium for Rhodopseuodomonas (Weaver et al., 1975) showed that metal cations and trace elements, as well as nitrogen, phosphorus, and carbon sources, are necessary factors for Rhodopseuodomonas growth. In this experiment, some crystallized salt appeared, and ammonium and phosphate radicals were measured in the dissolved solution of the crystallized salt, thus indicating that the water from the decomposition of fish feed was high in dissolved nitrogen and phosphorus nutrients. In addition, the TN content, TP content, and TOC content in the water during the culture period were high (Table 4). Moreover, the elemental composition of the crystallized salt (Table 5) indicated the presence of many metal cations in the water, supplying metal cations and some trace elements. The compound fish feed also contained trace elements. Thus, the metal cations and trace elements, in addition to the high content of nitrogen, phosphorus, and carbon, in the eutrophic water (Table 4) facilitated the fast growth of attached and suspended Rhodopseuodomonas. In particular, the attachment of salt to the inner tank wall is an important factor promoting the formation of a thick bacterial layer, by providing a persistent and stable substrate for mass growth.
There was an increase in pH and DO levels in TK1 and TK2, with a decrease in salinity levels ( Figure 4). Some crystallized inorganic salt formed during the culture process ( Figure 5), which was likely com-  Borgerding (1972) reported that struvite solubility could decrease from around 3,000 mg/L to less than 100 mg/L with an increase in pH from 5 to 7.5, while Buchanan, Mote, and Robinson (1994) identified pH 9 as the minimum pH required for struvite solubility. As the pH value increased to about 8.5 in this experiment (Figure 4), the struvite solubility likely decreased.
Thus, the salinity decreased gradually after 40 days.

| CON CLUS ION
In conclusion, this study provides details of a method for the mass culture of Rhodopseudomonas in the presence of compound fish feed. This method involves placing compound fish feed and tap water into aerated transparent containers. The ratio of fish feed to tap water was 14.3-50.8 g/L. No other inoculum enrichment or substances were needed during the culture process. As the fish feed decomposed, high nutrient concentrations, including high levels of nitrogen, phosphorus, carbon, and metal cations, were achieved, thus allowing both attached and suspended Rhodopseudomonas to become dominant from days 45 to 120 of the experiment. The biomass of attached Rhodopseudomonas attained in the current study is higher than that reported in many other studies. This method can facilitate large-scale cultivation of Rhodopseudomonas spp. by using aquatic fish feed fermentation.

CO N FLI C T O F I NTE R E S T S
None declared.

E TH I C S S TATEM ENT
None required.

DATA AVA I L A B I L I T Y S TAT E M E N T
The sequence data have been deposited in the Sequence Read Archive (SRA) under accession number SRP190363.