Bioreduction of Cr(VI) using a propane-based membrane biofilm reactor

The strong physiological toxicity of Cr(VI) makes it widely concerned in wastewater treatment. At present, the simplest and harmless method for treating Cr(VI) is known to be biologically reducing it to Cr(III), making it precipitate as Cr(OH)3(s), and then removing Cr(III) by solid separation technology. Studies have shown that Cr(VI) reduction bacteria can use CH4 and H2 as electron donors to reduce Cr(VI). Based on this, in this study, C3H8 was used as the only electron donor to investigate the potential of C3H8 matrix membrane bioreactor in the Cr(VI) wastewater treatment. The experiment was divided into three stages, each of which run stably for at least 30 days, and the whole process run for 120 days in total. The experiment is divided into three stages, each stage runs stably for at least 30 days, for a total of 120 days. With the increase of the Cr(VI) load, the removal rate gradually decreased. In stage 3, when Cr(VI) concentration was 2.0 mg·L−1, the removal rate was reduced from 90% in the first stage to 75%. According to X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) analysis, it is known that Cr(III) is the main product during this process and it is adsorbed on the biofilm as Cr(OH)3 precipitate. During the experiment, the amount of extracellular polymeric substance (EPS) produced by microorganisms increased initially and then decreased, and the amount of polysaccharides (PS) was always more than protein (PN). By analyzing the microbial community structure after inoculating sludge and adding Cr(VI), Nocardia and Rhodococcus dominate the biofilm samples. Chromate reductase, cytochrome c, nitrate reductase, and other functional genes related to chromate reductase increased gradually during the experiment.


Introduction
Chromium and its compounds are widely used in metallurgy, tanning, petroleum refining, textile, dye manufacturing, and other industries due to its special properties. Improper discharge of wastewater from these industries can cause serious chromium pollution (Barnhart 1997;Gu et al. 2013). Chromium mainly exists in the form of Cr(VI) and Cr(III) in water, both of which are carcinogenic (Dayan and Paine 2001). Among them, Cr(III) forms a precipitate in the form of hydroxide under neutral and alkaline conditions, which can be easily removed from water by solid separation technology. While Cr(VI) has high solubility, bioavailability, and mobility, and usually exists in the form of anion. Therefore, the toxicity of Cr(VI) is more than 100 times higher than that of Cr(III). The US Environmental Protection Agency (EPA) has identified that Cr (VI) is one of the 17 chemical substances that pose the greatest threat to humans (Marsh and Mcinerney 2001). The US Environmental Protection Agency (The US EPA) defines the maximum contaminant limit (MCL) standard for Cr contained in drinking water as 100 μg-Cr·L −1 . (US EPA, National Primary Drinking Water Standard 2003).
Traditional water treatment methods have poor chromate removal effects, are not suitable for treating low-concentration chromium-containing wastewater, and are prone to cause secondary pollution of the environment (Gui et al. 2009;Poopal and Laxman 2009). The frequently used processes to treat wastewater containing Cr ions from industrial Responsible Editor: Ioannis A. Katsoyiannis waste and waste-water include physical (adsorption, filtration, and flotation), chemical (coagulation, oxidation, reduction, electrolysis, solar photolysis, and photochemical), and biological (aerobic, anaerobic degradation). These methods have their limitations such as costs are high, and subsequent treatment and disposal are often required. In contrast, biological reduction to remove Cr(VI) from groundwater is a more suitable and effective method (Derek et al. 1997;Zahoor and Rehman 2009;Al Hasin et al. 2010;Ruggaber and Talley 2006). Cr (VI) is reduced to Cr(III) under the action of microorganisms, and Cr(III) exists in the form of Cr(OH) 3 (s), which is then removed by solid separation.
Membrane bioreactor (MBfR) is a novel and efficient technology to transfer substrates to microorganisms. Hollow fiber membrane surface micropores are used to supply donor substrate, which can not only increase the microbial attachment area, reduce the biological deposition, but also improve the mass transfer and utilization efficiency of substrates . The commonly used donor substrates are H 2 and CH 4 , and it has been proven that MBfR with H 2 and CH 4 as substrates can be used for bioreduction of inorganic anions (NO 3 − , NO 2 − , ClO 4 − , etc.) and heavy metal pollutants (Cr(VI), TeO 3 2− , etc.) (Chung et al. 2006;Luo et al. 2015;Nerenberg et al. 2008;Rittmann et al. 2011;Ziv-El and Rittmann 2009;Lai et al. 2016;Shi et al. 2019). Propane (C 3 H 8 ) is a by-product of the petrochemical industry, with abundant sources and low price; and C 3 H 8 is also one of the important components of VOCs. If it is not treated, it will cause serious harm to the environment. In this study, we attempted to use C 3 H 8 as an electron donor to remove Cr(VI), which can achieve the purpose of "using waste to treat waste" to a certain extent. Based on this, this study investigated the feasibility of Cr(VI) bioreduction in the MBfR reaction system using C 3 H 8 as the only electron donor, and on this basis, the reduction performance of Cr(VI) with different loading was investigated. Scanning electron microscope (SEM) and energy dispersive X-ray analysis (EDS) were used to analyze the cell morphology of chromate-reducing bacteria at different stages, and X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) products were used to analyze Cr(VI) reduction products.
Extracellular polymeric substance (EPS) is a complex polymer matrix polymer that exists in pure bacteria, activated sludge, granular sludge, and biofilm systems. It is produced by microorganisms in response to environmental stress (Sheng et al. 2008;Kantar et al. 2011;Liu et al. 2001;Guibaud et al. 2005). The interaction between EPS secreted by cells on biofilm and microbial reduction of Cr(VI) was investigated by analyzing the amount of EPS produced during the experiment. In order to investigate the mechanism of Cr(VI) reduction by microorganisms in the MBfR reaction system with C 3 H 8 as substrate, we identified the microbial community structure during Cr(VI) reduction by highthroughput Illumina sequencing of 16S rRNA and predicted the changes of functional genes by PICRUST technology.

Experimental device
The MBfR reactor used in the experiment is shown in Fig. 1. The MBfR reaction system was consisted of a membrane module and a circulation loop. Its core part was a composite gas transmission fiber (280 μm in OD, 180 μm in ID, which was produced by Mitsubishi Rayon). One end of the fibers was open and the other was sealed as a dead end in order to well deliver C 3 H 8 gas. The fiber membrane module was composed of 70 fiber membrane filaments. Ensure that the supply of C 3 H 8 was sufficient during the experiment. The total volume of the MBfR reaction system was 175 mL and the total surface area of the membrane filament was 197 cm 2 . The circulation loop was circulated by a peristaltic pump with a circulation flow rate of 40 rpm·min −1 to ensure complete mixing of the solution in the reaction system.

Start-up and operation of the reactor
The experimental inoculation-activated sludge was taken from Qingdao Nibuwan Wastewater Treatment Plant. After the sludge was inoculated, it was cultured with 30 mg·L −1 NO 3 − for 30 days to make the microorganisms evenly attached to the membrane filaments, and the experiment was started after the successful hanging of the membrane. The experiment was divided into 3 stages, the influent Cr(VI) concentrations were 0.5, 1.0, and 2.0 mg·L −1 , respectively, with the same influent flow rate (0.3 mL·min −1 ) in all three stages and the HRT was 12 h. Each stage run for at least 20 days to make the Cr(VI) removal rate reach a steady state. The experiment used the prepared inorganic medium to simulate Cr(VI) wastewater. A certain amount of potassium chromate was added according to the needs of the experiment, and the pH value of the simulated wastewater was adjusted to 7.5 ± 0.5. The specific inorganic culture medium and trace element formula are shown in Table 1.

Analytical method
Ten-milliliter samples were taken from the inlet and outlet water of the MBfR reactor once in 2 days immediately through a 0.22 μm membrane to remove the bacteria contained in the samples. Total chromium concentration was determined using atomic absorption spectrometry (contrAA 700, Germany) and Cr(VI) concentration was determined using diphenylcarbohydrazide spectrophotometry (method 3500-cr D, APHA1998). EPS was extracted from the inoculated microorganisms and different stages of the experiment using thermal extraction method. The protein (PN) content was determined using Coomassie bright blue method, and the polysaccharide (PS) content was determined using anthrone colorimetric method. Influent and outlet pH were measured by pH meter (PHS-25, Shanghai).

Flux calculation
We calculated the surface load of Cr(VI) according to formula (1) and the removal flux of Cr(VI) according to formula (2).
In the formula, S 0 and S represent the concentration of Cr(VI) influent and effluent, in g·L −1 ; Q represents the influent flow rate in the MBfR reaction system, in L·day −1 ; A represents the surface area of the central control fiber membrane, in m 2 .

Microbial morphology analysis and precipitation characterization
When each stage in the reaction system reached a stable state, the biofilm samples in the reactor were collected for SEM and EDS. The sample pretreatment method was referred to the methods and steps described by Wu et al. (2010). Five-milliliter of biofilm sample was washed once with deionized water, centrifuged, and removed the supernatant; soaked in 2.5% glutaraldehyde solution for 4 h, rinsed in PBS buffer solution and centrifuged; dehydrated twice with 10%, 30%, 50%, 70%, 90%, and 100% ethanol solution for 15 min each time; replaced with ethanol:isoamyl acetate 1:1 solution and pure isoamyl acetate for 15 min each time. The replaced samples were placed in a desiccator to dry for 8 h; the dried samples were coated and observed in a Schematic of the reaction device for this study scanning electron microscope, and the element composition of the biofilm sample was determined and analyzed with an EDS detector. Biofilm samples were at the end of the experiment and used for XRD, XPS, and Fourier infrared transform spectroscopy (FTIR) to analyze the valence state and existence of Cr in the precipitate within the reaction system. Three microbial samples were taken from the inside of the reactor, one of which was soaked in phosphate buffer solution (PBS, PH = 7.0) for 30 min, lyophilized for 48 h, and ground into powder for XRD analysis; the second sample was taken out and washed with ionized water and centrifuged for 15 min. The supernatant was removed and dried in an oven at 80 °C for XPS analysis. The third sample was taken out and freeze dried directly for FTIR analysis.

High-throughput sequencing analysis and functional gene prediction analysis
At the end of each stage of the experiment, samples of the biofilm were taken for high-throughput sequencing analysis. Primers 515F (5′-GTG CCA GCMGCC GCG G-3′) and 907R (5′-CCG TCA ATTCMTTT RAG TTT-3′) were used for PCR amplification of the conservative region of the bacterial 16S rRNA V4-V5 gene and the AxyPrep gel extraction kit (Axygen Biosciences, Union City, CA, USA) was used to purify PCR products. The purified amplicon was sent to Majorbio Biotechnology Company (Shanghai, China) for highthroughput sequencing. High-throughput sequencing was performed using the Illumina MiSeq processing platform, and the data were analyzed by QIIME (version 1.9.1). We used the UPARSE tool (version 7.0.1090) to screen out the 16S rRNA with ≥ 97% similarity in the clustering of operating units (OTUs), and the RDP Classifier tool (version 2.11) was used to compare the silva (SSU115) 16S rRNA database with sequences represented by 70% confidence intervals to analyze the genetic relationship of the phylogeny. PCoA analysis was performed on the inoculation source and biofilm samples at each stage to study the similarities and differences of the sample community composition. The alpha diversity analysis was performed using the Mothur tool (version 1.30.2) to obtain the distribution of species at different taxonomic levels for each sample based on the taxonomic analysis.
The PICRUST 2 tool was used to predict the functional gene abundance in inoculum source and biofilm samples at each stage based on the latest Kyoto Encyclopedia of Genes and Genomes (KEGG) database. Figure 2 shows the concentration of influent Cr(VI), effluent Cr(VI), and effluent soluble total chromium in the three stages of the C 3 H 8 -MBfR reaction system. Figure 3 shows the Cr(VI) removal rate, Cr(VI) flux, and its surface loading in the three stages of the C 3 H 8 -MBfR reaction system. According to the concentration of Cr(VI) in the effluent and total soluble chromium in the effluent, the content of soluble Cr(III) in the effluent can be neglected, which can be taken as a proof that almost all the Cr(VI) removed within this reaction system is converted to Cr (III) precipitation.

C 3 H 8 -MBfR processing Cr (VI) performance overview
According to Fig. 3, it can be seen that the removal rate of Cr(VI) gradually decreased as the Cr(VI) load increased, and the removal rate of Cr(VI) stabilized at about 95% at the influent Cr(VI) concentration of 0.5 mg-L −1 in the first stage. The increase of Cr(VI) concentration to 1.0 mg-L −1 led to an obvious decline of Cr(VI) removal efficiency to 80%. Further increasing Cr(VI) concentration to 2.0 mg-L −1 led to a fall of Cr(VI) removal efficiency to 75% in the third stage.

The effect of Cr(VI) on EPS
We measured the amount of EPS produced by microorganisms in the inoculated sludge and at different stages of the experiment, and the results are shown in Table 2.
According to Table 2, the amounts of PN and PS in EPS secreted by microorganisms without the addition of Cr(VI) were 1.13 mg/g·vss and 1.35 mg/g·vss, respectively. After adding Cr(VI), at the end of the first phase of the experiment, the productions of PN and PS were 1.51 mg/g·vss and 2.24 mg/g·vss, respectively, with an increase in both PN and PS compared to the inoculated sludge. At the end of the second phase, the amount of PN increased to 7.49 mg/g·vss, the amount of PS increased to 9.43 mg/g·vss, and the increase in PN and PS increased. While at the end of the third stage, the amount of PN decreased to 2.88 mg/g·vss and the amount of PS decreased to 3.48 mg/g·vss. The amount of EPS secreted by microorganisms increased with the increase of Cr(VI) load in stage 1 and stage 2. In stage 3, when the Cr(VI) inflow is 0.5 mg·L −1 , the amount of EPS decreased by 60% compared to stage 2. We speculate that the EPS secreted by microorganisms has a certain effect on the reduction of Cr(VI), but the high load of Cr(VI) will also inhibit the secretion of EPS by microorganisms. High concentrations of Cr(VI) might also present irreversible inhibition on the production of EPS. Liu et al. also obtained similar results ). Figure 4 shows the changes in the cell morphology of chromate-reducing bacteria at each stage in the reaction system and the signals of each element contained in the biofilm sediment at each stage.

Characterization of cell morphology and sedimentation of chromate reducing bacteria
The results of SEM analysis show that when Cr(VI) is not added, most of the microorganisms from the inoculation source had a smooth surface and mostly rod shaped (Fig. 4A1). In the first and second stages after the addition of Cr(VI), obvious irregular wrinkles appeared on the cell surface, and the first stage was particularly obvious (Fig. 4B1 and 4C1). Based on EDS analysis, it was found that Cr signals had appeared in the biofilm deposits in the first stage, and obvious Cr signals were detected in the second and third stages (Fig. 4B2, C2, and D2). It proved that the irregular folds on the cell surface were Cr deposits. In the third stage of the experiment, the cell surface returned to smoothness (Fig. 4D1), indicating that the chromate-reducing bacteria gradually adapted to the Cr(VI) environment as the experiment progressed. Studies have shown that chromate has strong oxidizing ability and can change cell morphology directly or through the  production of reactive oxygen species (ROS) (Lai et al. 2016). In many studies, it has been reported that the cell surface of Cr(VI)-reducing bacteria appears deformed and wrinkled (Lai et al. 2016;Pei et al. 2009). After the end of the third stage, we analyzed the precipitates within the reaction system by XPS. The results are shown in Fig. 5. We used chromium chloride hexahydrate (CrCl 3 ·6H 2 O) and potassium dichromate (K 2 Cr 2 O 7 ) as Cr( III) and Cr(VI) standard materials, respectively. The XPS results showed that the sample patterns were consistent with the Cr(III) standard material pattern, which could prove that the Cr in the biofilm sediment in the reaction system was trivalent. Also, we analyzed the precipitates by XRD, and the results are shown in Fig. 6. The XRD results also showed the presence of Cr(OH) 3 ·3H 2 O in the precipitates, which is consistent with the results of Min Long et al. (2017). Therefore, through XPS and XRD analysis, we can conclude the presence of Cr(III) precipitates on the biofilm within the reaction system, and that Cr(III) was present in the form of Cr(OH) 3 ·3H 2 O.
The FTIR results in Fig. 7 also further prove that Cr(III) is adsorbed on the biofilm. Figure 7 compares the FTIR spectra of the biofilm before and after the addition of Cr(VI). The results show that the addition of Cr caused a significant change in the absorption peak. It can be inferred that the hydroxyl group, carboxyl group, nitro group, and sulfonic acid group of the biofilm are related to the combination of Cr by the change of the absorption peak. The biofilm sample showed a broad, stretched, and strong peak at 3413.9 cm −1 , which is characteristic of -OH stretching vibration. The adsorption band of the sample at 2925.9 cm −1 is characteristic of -CH tensile vibration. The adsorption peaks at 1661.5 cm −1 and 1453.2 cm −1 reflect the stretching vibration of C = C and the asymmetric stretching vibration of C-O-C, which are related to the obvious deviation in the adsorption process. A new adsorption band appears at 1738.8 cm −1 after the deposition of Cr on the biofilm. This is a characteristic of -COOH stretching. Studies have shown that negatively charged chemical reaction functional groups such as -OH and -COOH can bind to highly valent metals and can form a barrier that impede the penetration of heavy metals into cells, thereby reducing the toxicity of heavy metals to cells (Joshi and Juwarkar 2009;Kang et al. 2014;Wang et al. 2014;Zhu et al. 2012).

Microbial community structure analysis
We performed high-throughput sequencing analysis on the inoculated sludge and the sludge samples at each stages of the experiment to obtain the original sequences of the four samples. The original sequences were subjected to OTU cluster analysis after quality control according to 97% similarity. The numbers of valid sequences and OTUs are shown in Table 3.
As shown in Table 3, at a certain depth of sequencing, the effective sequence number and OTU number of different samples varied among different samples. In general, the highest numbers if OUTs and more effective sequences were found in the inoculated samples, and the OTU number gradually decreased with the increase of Cr(VI) concentration in the influent water, indicating that high concentration of Cr(VI) would reduce the abundance of the community.
Alpha diversity analysis was performed on the samples to obtain the alpha diversity indices of different samples in Table 4. Among them, the Sobs, Chao, and Ace indexes reflect the abundance of the community and are related to the actual number of species in the microbial community. The larger the value, the richer the community. The Shannon and Simpson indexes reflect the community diversity, and the coverage index reflects community coverage. According to Table 4, the Sobs, Chao, and Ace index values of the inoculum were significantly higher than those of each stage of the experiment, indicating that the addition of Cr(VI) had a great impact on the abundance of the community. The Shannon and Simpson indexes showed different degrees of fluctuation throughout the reactor operation stage, indicating that different Cr(VI) concentrations had a significant impact on the composition of the microbial community. This result is consistent with the research conclusion of Aquino et al. (2004). Figure 8 shows the composition of the microbial community at the phylum level in inoculum source and at each stage. At the phylum classification level, Actinobacteria, Alphaproteobacteria, and Gammaproteobacteria dominated. Among them, the abundance of Actinobacteria and Alphaproteobacteria in the inoculated sludge was 10.84% and 5.67%, respectively. With the addition of Cr(VI), the abundance of Actinobacteria and Alphaproteobacteria in the first stage increased to 45.97% and 16.97%, respectively, occupying a dominant position. The abundance of Alphaproteobacteria gradually increased with the increase of the influent Cr(VI) concentration and dominated in both stage 2 and stage 3, while the abundance of Actinobacteria gradually decreased to 15.68% in stage 3, which was consistent with the trend of gradually decreasing of the Cr(VI) removal rate. Figure 9 shows the composition of the microbial community at the genus level in inoculum source and at each Fig. 4 (Left column) SEMs of the MBfR biofilms and Cr(III) precipitates for the inoculum (A1) and stages 1 (B1), 2 (C1), and 3 (D1). Samples were taken at day 25, day 70, and day 110 for stages 1, 2, and 3, respectively. (Right column) Representative EDS for the inoculum (A2) and stages 1 (B2), 2 (C2), and 5 (D2) ◂ stage. At the genus classification level, the abundance of Nocardia, Rhodococcus, and Mycobacterium belonging to Actinobacteria in the inoculated sludge is zero, and the abundance of both increased as the experiment progressed. The genera Nocardia, Rhodococcus, and Mycobacterium have been shown to have the ability to use gaseous normal alkanes as the sole carbon source (Hamamura et al. 2001;William et al. 1994). N. R. Woods (1989) found that a strain of Rhodococcus rhodochrous isolated from the soil was able to use propane as the sole carbon source. Some studies have found multiple strains of propane oxidizing bacteria in Rhodococcus and Mycobacterium, which can use propane as a substrate to biodegrade trichloroethylene, methyl tertbutyl ether, dioxane, and other organic pollutants (Wackett et al. 1989;Tupa et al. 2018;Deng et al. 2018). Studies have reported that Nocardia has the ability to remove Cr(VI) (Zhang et al. 2019). Dimitroula et al. (2015) have reported that a strain of acidophilus was able to completely reduce  (Dimitroula et al. 2015). Previous studies have also reported that strains belonging to the genus Rhodococcus were able to reduce Cr(VI) (Sun et al. 2011;Revelo Romo et al. 2019;Kuyukina et al. 2017;Patra et al. 2010). Soumya Banerjee (2017) has reported that R. erythropolis removed Cr(VI) through bioconcentration (Banerjee et al. 2017). Nocardia and Rhodococcus had the highest abundance at stage 1, and then the abundance decreased with the increase of the influent Cr(VI) concentration, which was consistent with the gradually decreasing trend of the Cr(VI) removal rate. Table 5 shows the predicted abundance of some functional genes of metagenomics on the biofilm in the inoculation source and the various stages of the experiment based on the PICRUST 2 function prediction analysis. Among them, the abundance of ABC transport system at all stages of the experiment was higher than that of the inoculation source, and its abundance was the highest in stage 1. The abundance of ABC transporters at all stages of the experiment was higher than that of the inoculation source. ABC transporters can transport heavy metals through the cell membrane and excrete them from the cells, which are part of the cell defense system (Torre et al. 2012). The abundance of chromate reductase and chromate transporter gradually increased with the increase of the influent Cr(VI) concentration in the influent water and reached the highest in stage 3. Cytochrome c belongs to chromate reductase and has Cr(VI) reduction ability, which reduces Cr(VI) to Cr(III) via Cr(V) intermediates. Its abundance increased with the concentration of Cr(VI) influent concentration (Shi et al. 1990;Ackerley et al. 2004). NADH, as an electron donor for reduction of intracellular Cr(VI), had an abundance 2 times that of the inoculation source at each stage, and its abundance in stage 1 with the highest Cr(VI) removal rate was 3 times that of the inoculation source (Thatoi et al. 2014). Flavin reductase and nitrate reductase are both soluble chromate reductase, which are extracellular enzymes, and their abundance gradually increased with the progress of the experiment (Cheung et al.    2017). The abundances of cytochrome c551 and thioredoxin reductase, which are also related to chromate reduction, increased in stage 1, and then decreased with the increase of Cr(VI) influent concentration (Li, L. Z. P. L. X. 2013; Zhong et al. 2017). Cysteine is a non-enzymatic chromate reducing agent, and its abundance in the reaction system is negligible (Poljsak et al. 2010). Compared with the vaccination source, some genes related to the oxidation of propane were enriched (aldehyde dehydrogenase, propanal dehydrogenase, acetone monooxygenase, propane monooxygenase coupling protein, NAD + -dependent secondary alcohol dehydrogenase Adh1), and there were four main pathways for propane oxidation, namely, monooxygenase-mediated terminal oxidation of propan-1-ol, monooxygenase-mediated sub-terminal oxidation of propan-2-ol, oxidized by the oxygenase mechanism to produce prop-1-ol, and prop-The 2-alcohol mixture produces propane-1,2-diol by oxidizing terminal and subterminal carbon atoms by dioxygenase (Hamamura 2001). In this study, based on the measured abundance of functional genes, it can be inferred that during the experiment, under the action of microorganisms, propane was oxidized via the monooxygenase-mediated pathway via propane-1-ol or prop-2-ol terminal or sub-terminal and produced metabolic

Research on Cr(VI) reduction mechanism
According to the analysis of microbial community diversity and PICRUST 2 function prediction analysis, the reduction mechanism of Cr(VI) in the reaction system is speculated (Fig. 10).
The reduction of Cr(VI) includes enzymatic reduction and non-enzymatic reduction (Derek et al. 2009;Losi et al. 1994;Fendorf et al. 1996). Cysteine is a non-enzymatic chromate-reducing agent. According to the prediction of the abundance of functional genes, the abundance of this gene in the reflection system is negligible. Therefore, in this study, the reduction of Cr(VI) is mainly through an enzymatic process.
Enzyme-mediated reduction of Cr(VI) can be divided into two parts, the reduction of Cr(VI) in the cell membrane and the reduction of Cr(VI) outside the cell membrane. Cr(VI) in the cell is reduced to Cr(III) under the action of chromate reductase and cytochrome c. This process may release reactive oxygen species (ROS), and the generation of ROS will affect cell viability and Cr(VI) reduction (Thatoi et al. 2014). The extracellular Cr(VI) is reduced to Cr(III) under the action of flavin reductase and nitrate reductase; the reduction product Cr(III) combines with the functional groups on the cell surface (Ngwenya et al. 2011).

Conclusion
In this study, we concluded that with the continuous increase of influent Cr(VI) load, the removal rate of Cr(VI) gradually decreases. When the influent Cr(VI) concentration is 2.0 mg·L −1 , Cr(VI), the removal rate is only 75%. Through the effluent soluble total chromium concentration, Cr(VI) concentration, X-ray diffraction (XRD), and X-ray photoelectron spectroscopy (XPS) analysis, we proved that all the reduced Cr(VI) was converted to Cr(III) and reduced. The product Cr(III) is in the form of Cr(OH) 3 ·3H 2 O. By measuring the EPS secreted by the microorganisms during the reaction, it is inferred that the EPS secreted by the microorganisms has a certain effect on the reduction of Cr(VI), but the high load of Cr(VI) will also inhibit the secretion of EPS by the microorganisms. High-throughput sequencing analysis showed that Nocardia and Rhodococcus were the dominant bacterial genera. The abundance of the two bacterial genera in the inoculation source was 0. After Cr(VI) was added, the abundance was enriched. The latter abundance increased with Cr(VI) decrease in removal rate. Through PICRUST 2 function prediction analysis, it can be seen that the abundance of some functional genes related to Cr(VI) reduction increases with the increase of Cr(VI) concentration (such as chromate reductase, cytochrome c, NADH, flavin reductase, and nitrate reductase); the abundance of some functional genes decreases with the increase of Cr(VI) concentration (such as cytochrome c551 and thioredoxin reductase). At the same time, functional genes related to propane oxidation are also enriched.
Author contribution All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by Chunshuang Liu, Luyao Zhang, and Haitong Yu. The first draft of the manuscript was written by Luyao Zhang and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Data availability All data generated or analyzed during this study are included in this published article (and its supplementary information files).

Declarations
Ethics approval and consent to participate Not applicable.

Competing interests
The authors declare no competing interests.