Construction of Recombinant Magnetospirillum Strains for Nitrate Removal from Wastewater Based on Magnetic Adsorption

Abstract

Nitrate ion (NO₃⁻) in wastewater is a major cause of pollution in aquatic environments worldwide. Magnetospirillum gryphiswaldense (MSR-1) has a complete dissimilatory denitrification pathway, converts NO₃⁻ in water into nitrogen (N₂) and simultaneously removes ammonium ions (NH₄⁺). We investigated and confirmed direct effects of regulatory protein factors Mg2046 and MgFnr on MSR-1 denitrification pathway by EMSAs and ChIP-qPCR assays. Corresponding mutant strains were constructed. Denitrification efficiency in synthetic wastewater medium during a 12-h cell growth period was significantly higher for mutant strain Δmgfnr (0.456 mmol·L⁻¹·h⁻¹) than for wild-type (0.362 mmol·L⁻¹·h⁻¹). Presence of magnetic particles (magnetosomes) in MSR-1 greatly facilitates collection and isolation of bacterial cells (and activated sludge) by addition of a magnetic field. The easy separation of magnetotactic bacteria, such as MSR-1 and Δmgfnr, from wastewater using magnetic fields is a unique feature that makes them promising candidates for practical application in wastewater treatment and sludge pretreatment.

1. Introduction

Nitrate (NO₃⁻) pollution in water and wastewater is a major environmental problem worldwide. The primary sources of NO₃⁻ contamination are anthropogenic activities such as excessive fertilization in agriculture [1]. In numerous lakes and rivers in Europe and the U.S., NO₃⁻ concentration has doubled in less than a decade [2]. In certain coastal zones (e.g., Gulf of Mexico, Yellow Sea, Baltic Sea, Chesapeake Bay) increasing NO₃⁻ levels cause severe algal blooms that lead to “dead zones”. NO₃⁻ can infiltrate drinking water sources, posing a threat to human health. A 2020 report from the Ireland Environmental Protection Agency [3] concludes that “urgent” action is needed to reduce water NO₃⁻ levels.

Among various approaches for NO₃⁻ removal, biological denitrification is more cost-effective than physicochemical methods, such as adsorption, ion exchange, and reverse osmosis [4,5]. Biological denitrification has been applied widely for treatment of municipal, industrial, and agriculture wastewater and NO₃⁻-contaminated groundwater. During this process, NO₃⁻ functions as an electron acceptor and is converted to harmless nitrogen gas by denitrifying microorganisms [6,7,8,9,10]. An electron donor is necessary as a source of electrons and energy during denitrification processes [11].

Denitrification is classified as heterotrophic or autotrophic, depending on the electron donor. In heterotrophic denitrification, the electron donor is a low-molecular-weight (e.g., acetate, methanol, glucose, benzene, methane) or high-molecular-weight (e.g., cellulose, polylactic acid, polycaprolactone) organic compound [11]. In autotrophic denitrification, the electron donor is an inorganic compound such as hydrogen gas (H₂), reduced sulfur compound (e.g., sulfide, elemental sulfur, thiosulfate), ferrous iron (Fe²⁺), iron sulfide (e.g., FeS, Fe_1−xS, FeS₂), arsenite (As^III), or manganese (Mn^II) [12]. The nitrate removal rate is generally higher for heterotrophic denitrification than for autotrophic denitrification because heterotrophic bacteria have a faster growth rate [4].

Many studies during the past decade have focused on denitrification processes involving Fe²⁺ as an electron donor. Most Fe^II-oxidizing denitrifiers isolated to date are heterotrophic and can grow only in the presence of organic carbon [13,14]. Effective follow-up treatments for excess sludge from heterotrophic denitrification are needed [15]. Following long-term operations, accumulated Fe²⁺ may form an iron crust that inhibits subsequent activity of denitrifiers.

Magnetotactic bacteria (MTB) produce intracellular magnetic nanoparticles composed of magnetite (Fe₃O₄) or greigite (Fe₃S₄) and enclosed by a lipid bilayer, termed “magnetosomes” [16]. These nanosized organelles enable MTB to orient and migrate along magnetic lines of force. MTB have been applied in cancer therapy and in heavy metal biosorption from wastewater or polluted water [17,18]. They are useful for nitrate removal from wastewater because MTB biomass is easily removed from sewage treatment systems by magnetic separation. Li et al. [19] described a complete denitrification pathway, including functional genes for reduction of nitrate (nap), nitrite (nir), nitric oxide (nor), and nitrous oxide (nos), in the extensively studied MTB strain Magnetospirillum gryphiswaldense MSR-1. The present study is focused on regulation of the denitrification pathway in MSR-1, and construction of a recombinant MTB strain with enhanced nitrate removal capability.

2. Materials and Methods

2.1. Bacterial Strains and Culture Conditions

Bacterial strains and plasmids used in this study are listed in

Table 1. Primers used in this study.

Primer	Sequence (5′-3′)	Description
Qrpoc-F	ATCTGGTCTACCGCCATTG	qRT-PCR for rpoc gene
Qrpoc-R	CCTTGCCGAACGAAATACC	qRT-PCR for rpoc gene
QmamA-F	GCCTATCCGTGGCGAAGAA	qRT-PCR for mamA gene
QmamA-R	TCGGCATCGTAAACCTGCT	qRT-PCR for mamA gene
QmamB-F	AGGTCGTGTGGTGGGCAT	qRT-PCR for mamB gene
QmamB-R	CGCTCATCCGCAGGCTTA	qRT-PCR for mamB gene
Qmms6-F	GGTTGGCGTTGGGAAGGT	qRT-PCR for mms6 gene
Qmms6-R	CATCGCTCTGTGCCGCTT	qRT-PCR for mms6 gene
QmmsF-F	TCGGGACGACGAGTTTGTC	qRT-PCR for mmsF gene
QmmsF-R	GGAACACCACGGAGACCAA	qRT-PCR for mmsF gene
QnapF-F	TGATGTCGCACAGCCTTAG	qRT-PCR for napF gene
QnapF-R	TGATGTCGCACAGCCTTAG	qRT-PCR for napF gene
QnirT-F	CCATTCACTACACCAACCGTTC	qRT-PCR for nirT gene
QnirT-R	ATGGCAGTTGCGGCATTC	qRT-PCR for nirT gene
QnorC-F	CGGTGTTCGTTGCCTTGA	qRT-PCR for norC gene
QnorC-R	CAGACATTGCCCAGTTCCG	qRT-PCR for norC gene
QnosZ-F	TCGCCACGGTGTCCTTT	qRT-PCR for nosZ gene
QnosZ-R	ATCACCTGACCGCTTTGGC	qRT-PCR for nosZ gene
Qmg2046-F	GCTCCATACCCAATGACGC	qRT-PCR for mg2046 gene
Qmg2046-R	TGTCCACATCCTCGCCC	qRT-PCR for mg2046 gene
Qmgfnr-F	GAGTTGAACCACGACGAAATCA	qRT-PCR for mgfnr gene
Qmgfnr-R	CGAACATCTCGCCCGAAA	qRT-PCR for mgfnr gene
mgfnrSF-EcoRI	CGGAATTCACCCTGACCGTGGGCAAGCCGGAA	Amplification upstream of mgfnr
mgfnrSR-SacI	CGAGCTCACCTTGTGATCGTCGTAATCC	Amplification upstream of mgfnr
mgfnrR-SacI	GCTGTTGTTCTTCCTGCT	Confirmation of mgfnr mutant
mgfnrF-SacI	TCCACCGAAATGAAACCG	Confirmation of mgfnr mutant
mgfnrXF-SacI	CGAGCTCTACCCAGTTGAAGCGTGAAG	Amplification downstream of mgfnr
mgfnrXR-XbaI	GCTCTAGAGAAATCGGAAAACAGCCCCA	Amplification downstream of mgfnr
Pmg2046-F	GGAATTC ATGACGACGATGATCCA	Expression of mg2046 protein
Pmg2046-R	CCTCGAG TTAAACGTTCTCCCATC	Expression of mg2046 protein
Pmgfnr-F	CCGGAATTCGTGATCCCCATGCCGCC	Expression of mgfnr protein
Pmgfnr-R	CCGCTCGAGCTAATGCGCCCCGCCGC	Expression of mgfnr protein
Erpoc-F	TGAAGGAAGCCAAGGACCT	EMSA for rpoc gene promoter
Erpoc-R	CGAGGGACGG GTCAAATCCC	EMSA for rpoc gene promoter
EnapF-F	CGGCGGTCAAGAAGATGA	EMSA for nap operon
EnapF-R	GAGTGCGCCCGAACAAGG	EMSA for nap operon
EnirT-F	AAGCAGCAGGGCGTTCCT	EMSA for nir operon
EnirT-R	AGTATTTTCATTTTGGACA	EMSA for nir operon
EnorC-F	AACCGATCTCATCGGCGAA	EMSA for nor operon
EnorC-R	TACCGCCATAAAAGATATT	EMSA for nor operon
EnosZ-F	AGACGTCGGGGCAGAAGGT	EMSA for nos operon
EnosZ-R	AGCGCGCCAAAGGACACCGT	EMSA for nos operon
Emg2046-F	TGTTGGGCGAAATCCTCGT	EMSA for mg2046 gene promoter
Emg2046-R	GGCCAGGATG TCCACATCCT	EMSA for mg2046 gene promoter
Emgfnr-F	TGGCTGAAATCTGCGAGGTT	EMSA for mgfnr gene promoter
Emgfnr-R	TCACGGGCGA CCTTGTGAT	EMSA for mgfnr gene promoter

. MSR-1 was autoclaved and cultured at 30 °C in sodium lactate medium (SLM), which contained (per L) 2.25 g of sodium lactate solution (55–65%), 0.4 g of NH₄Cl, 0.1 g yeast extract, 0.5 g of K₂HPO₄, 0.1 g of MgSO₄, and 0.5 mL of trace element mixture; ferric citrate was added to obtain a final concentration of 60 µM. To make nitrate medium or ammonium/nitrate medium, 0.4 g of NH₄Cl in the above recipe was replaced respectively with 1 g of NaNO₃ or 0.2 g of NH₄Cl + 0.5 g NaNO₃. Synthetic wastewater medium was made as described previously [20] and sterilized at 121 °C for 30 min. MSR-1 cultures were incubated on a shaker (100 rpm, 30 °C) with MSR-1 antibiotics: nalidixic acid (Nx) 5 µg·mL⁻¹, ampicillin 5 µg·mL⁻¹, kanamycin sulfate (Km) 5 µg·mL⁻¹, and gentamicin (Gm) 5 µg·mL⁻¹. Escherichia coli cultures were incubated (shaking: 200 rpm, 37 °C) with E. coli antibiotics: ampicillin 100 µg·mL⁻¹, Km 100 µg·mL⁻¹, and Gm 20 µg·mL⁻¹.

2.2. Cell Growth and Magnetic Response

Optical density at a wavelength of 565 nm (OD₅₆₅) was measured every 2 h using a UV-visible spectrophotometer (model UNICO2100; Unico Instrument Co.; Shanghai, China). Magnetic response (Cmag) was determined as described previously [21]. Residual iron concentration was measured by ferrozine method in broth supernatant aspirated after centrifugation [22].

2.3. Construction of mgfnr Mutant Strain Δmgfnr

To construct mgfnr deletion mutant, an upstream fragment (1054 bp) was amplified using primer pair mgfnrS-F/mgfnrS-R (Table 1), and a downstream fragment (1207 bp) was amplified using primer pair mgfnrX-F/mgfnrX-R. These fragments were ligated into pMD18-T simple vector (TaKaRa Biotechnology; Dalian, China) for sequencing. Amplified upstream fragment was digested with EcoRI and SacI, and amplified downstream fragment was digested with SacI and XbaI. Gm cassette was digested from pUC-Gm vector with SacI. The above fragments were fused together into XbaI and BamHI sites of pUX19 vector to yield pUX-mgfnr. pUX- mgfnr was transformed into wild-type (WT).

MSR-1 by biparental conjugation using E. coli S17-1 as donor strain (

Table 2. Strains used in this study.

Strain or Plasmid	Description	Source or Reference
E. coli DH5α	endA1 hsdR17 [r-m+] supE44 thi-1 recA1 gyrA [NalR] relA relA1 Δ[lacZYA-argF] U169 deoR [Ø80Δ (LacZ) M15]	Novagen
E. coli DH5α-mgfnrS	DH5α containing pMD18-T-mgfnrS, Ampr	This study
E. coli DH5α-mgfnrX	DH5α containing pMD18-T-mgfnrX, Ampr	This study
E. coli S17-1	Thi endA recA hsdR with RP4-2-Tc::Mu-Km::Tn7 integrated in chromosome; Smr, Tra	Novagen
E. coli S17-1-Δmgfnr	S17-1 containing pUX-19-mgfnrS-Gm-mgfnrX, Kmr, Gmr	This study
E. coli BL21 (DE3)	F- ompT hdSB (rB− mB−) gal dcm (DE3), general purpose expression host	Novagen
E. coli BL21-Pmg2046	BL21 containing pET28a (+)-Pmg2046, Kmr	This study
E. coli BL21-Pmgfnr	BL21 containing pET28a (+)-Pmgfnr, Kmr	This study
MSR-1 WT	WT M. gryphiswaldense, Nxr	DSM 6361
MSR-1 Δmg2046	mg2046-deficient mutant, Nxr, Gmr	This study
MSR-1 Δmgfnr	mg2046-deficient mutant, Nxr, Gmr	This study

). Colonies were screened and selected using Gm and Nx. Successful knockout of mgfnr, without deletion of magnetosome island-related genes, was confirmed by PCR.

2.4. Transmission Electron Microscopic (TEM) Observation

Bacterial strains with 60 µM of ferric citrate were placed in SLM and incubated for 24 h at 30 °C. Cells were collected by centrifugation, washed twice with ddH₂O, resuspended, dropped onto a copper net, air-dried, and observed by TEM (model JEM-1230; JEOL; Tokyo, Japan). Number and size of magnetosomes were measured using ImageJ software program (NIH; Bethesda, MD, USA).

2.5. Quantitative Real-Time Reverse Transcription PCR (qRT-PCR)

WT and Δmg2046/Δmgfnr strains were cultured for periods of 6, 12, 18, and 24 h. RNA was extracted with TRIzol reagent (Tiangen; Beijing, China), and DNA was digested with DNase I (Takara; Shiga, Japan). RNA was reverse transcribed into cDNA using Moloney murine leukemia virus (M-MLV) reverse transcriptase (Promega; Madison, WI, USA). qRT-PCR was performed as per manufacturer’s instructions, using an optical circulator 480 RT-PCR system and 480 SYBR Green I Master Kit (Roche; Mannheim, Germany). Reaction volume (20 µL total) contained 50 ng of template cDNA, 10 µL of SYBR Green I mix, and 0.5 mM of the particular primer. The housekeeping gene rpoc was used as the internal reference. Calculations were made using the 2^-ΔΔCt method, with triplicate samples.

2.6. Electrophoretic Mobility Shift Assays (EMSAs)

EMSAs were performed using digoxigenin gel shift kit (DIG) (2nd generation; Roche). Target DNA sequence was amplified and DIG-labeled in 20-µL reaction volume consisting of 0.4 nM of DIG-labeled probe, various concentrations of His₆-Mgfnr/His₆-Mg2046, 4 µL of bonding buffer, 1 µL of poly[d(I-C)], and ddH₂O. Specificity of Rok7B7/probe interaction was confirmed by adding 300x unlabeled rpoc (nonspecific probe) and specific probe to the reaction system. The sample was incubated for 30 min at 25 °C, PAGE was performed to separate protein-bound DNA from free DNA, gel transferred to positively-charged nylon membrane (Roche), and chemiluminescence signal detected using an imaging system.

2.7. Iron Absorption Capability

Iron concentration in medium was assayed using ferrozine [3-(pyridyl-2-yl)-5,6-bis(4-sulfophenyl)-1,2,4-triazine disodium salt]. Standard curves for the three strains were constructed based on various concentrations (0, 20, 40, 60, 80, 100 µM) of ferric citrate. One hundred microliters of ferric citrate was added with hydroxylammonium chloride (250 µL; 10% [w/v]), incubated for 5 min at room temperature, added with 20 µL of ferrozine (25 g·L⁻¹) and 630 µL of ddH₂O, and incubated for another 15 min. OD₅₆₅ was measured by spectrophotometry. At 6-h intervals, supernatants were obtained by centrifuging 1 mL of cell suspension (relative centrifugal force 8000; 3 min). Iron concentrations in supernatants were measured using the standard curves.

2.8. Protein Expression Vector Construction and Anaerobic Purification

Expression plasmids pET-28a-mg2046 and pET-28a-mgfnr were constructed by PCR amplification using MSR-1 genome as template. mg2046 (723 bp) and mgfnr (762 bp) gene fragments (upstream and downstream; containing EcoR I and Xho I restriction site sequences, respectively) were amplified, ligated into pMD18-T simple vector, and transformed into E. coli DH5α. Correctly sequenced strains were selected, strains containing pET-28a(+) plasmids were cultured, and plasmids were extracted and double digested with EcoR I and Xho I to recover target fragments and linear pET-28a(+) plasmids. Target fragments were combined with pET-28a(+), ligated, and transformed into E. coli BL21.

Constructed strains as above were cultured at 37 °C in 200 mL of LB broth added with 50 µg·mL⁻¹ of Km until OD₆₀₀ reached 0.6. Cells were placed in an ice bath for 18 min, and added with isopropyl β-D-1-thiogalactopyranoside (IPTG) to achieve a final concentration 1 mM of IPTG. mg2046 was incubated at 150 rpm, at 16 °C, and mgfnr was incubated at 200 rpm and induced for 1 h at 37 °C. Methionine and ferric citrate were added to respective final concentrations of 25 and 200 µM, and incubation was continued. Cells were collected, disrupted in anaerobic environment by low-temperature centrifugation, and protein-containing supernatant was placed in nickel-nitrilotriacetic acid-agarose (Ni-NTA) column (GE Healthcare; Danderyd, Sweden), repeatedly vacuumed and exposed to nitrogen gas, transferred to anaerobic glovebox cabinet (MBRAUN LABstar; München, Germany), and eluted with buffers containing 20 mM Tris (pH 8.0), 300 mM NaCl, and imidazole at various concentrations. Purified proteins were stored in liquid nitrogen.

2.9. Chromatin Immunoprecipitation-Quantitative PCR (ChIP-qPCR)

Following anaerobic purification as above, proteins were sent out (BGI Genomics; Beijing, China) in order to prepare polyclonal antibodies for ChIP-qPCR. Purified Mg2046 and Mgfnr proteins were injected into rabbits to generate respective polyclonal antibodies. qPCR was performed using primer pairs listed in Table 1, and relative levels of Mg2046 and Mgfnr-precipitated DNA were determined by comparison with IgG control.

2.10. Total Nitrogen Content

Tested bacterial cells were cultured and harvested as in preceding sections. Total nitrogen content of cells was determined using a kit (cat # LH-NT-100; Lianhua Technology; Beijing, China) as per the manufacturer’s instructions.

3. Results

3.1. Combined Effects of Dissimilatory Denitrification Pathway Genes Determined by EMSAs and ChIP-qPCR

Denitrifying effects of microorganisms in sewage are closely correlated with degree of denitrification gene expression. Two protein factors (Mg2046, MgFnr) were previously implicated in control of MSR-1 denitrification pathway [23]. We examined the capability of these factors to directly regulate specific denitrification genes, including nap, nir, nor, and nos.

His₆-labeled Mg2046 and Mgfnr proteins were expressed in E. coli BL21 and confirmed by western blotting (Figure S1). The capability of Mg2046 and Mgfnr to directly regulate genes involved in the dissimilatory denitrification pathway (NO₃⁻→ NO₂⁻→ NO → N₂O → N₂) was determined by EMSAs using anaerobically-purified His₆-Mgfnr and His₆-Mg2046. Both these tagged proteins combined with upstream sequences of napF, nirT, nosZ, and mg2046 (Figure 1A,B), and His₆-Mg2046 combined with norC (Figure 1A), indicating the capability of both Mg2046 and Mgfnr to directly regulate dissimilatory denitrification genes. Mgfnr directly regulated mg2046, whereas no direct regulator of mgfnr by Mg2046 was found.

Combination of Mgfnr and Mg2046 with dissimilatory denitrification gene promoters in vivo was further examined by ChIP-qPCR using WT, Δmgfnr (constructed in this study; confirmed by PCR) (Figure S2), and Δmg2046 (constructed by Wang et al. [24]). WT strain was used as negative control, qRT-PCR analysis of product DNA showed that no enrichment of Mg2046 on mgfnr was detected; enrichment levels of Mgfnr and Mg2046 on napF, nirT, norC, nosZ and mg2046 were at least 1-fold higher in Δmg2046/Δmgfnr than that in control in all samples immunoprecipitated (Figure 1C). These findings indicate that Mg2046 regulates dissimilatory denitrification genes and itself, but does not directly regulate mgfnr, whereas Mgfnr regulates the dissimilatory denitrification genes and also mg2046.

3.2. Transcription Levels of Dissimilatory Denitrification Pathway Genes

Variations of dissimilatory denitrification pathway in Δmg2046 and Δmgfnr were evaluated by qRT-PCR analysis of key genes (napF, nirT, norC, nosZ). Elimination of nap and nir impairs magnetite biomineralization, resulting in fewer, smaller, and more irregular crystals during the denitrification process. Microaerobic respiration may interfere with magnetosome synthesis by disrupting proper oxide balance [19,25]. Δmg2046, relative to WT, did not show significant difference of nirT or norC transcription levels at 6 or 12 h, nor of napF or nosZ levels at 6 h. Levels of napF, nirT, norC, and nosZ decreased significantly in Δmg2046 relative to WT during 18–24 h but increased significantly in Δmgfnr (Figure 2). Levels of dissimilatory denitrification genes were relatively low from 6–12 h. Consumption of oxygen in medium from 18–24 h resulted in increased levels of dissolved oxygen and of dissimilatory denitrification genes. Gene transcription was significantly higher in WT than in Δmg2046, indicating a positive regulatory effect of mg2046 on denitrification pathway as oxygen content declined. Transcription levels of mgfnr genes did not differ significantly between Δmg2046 and WT, whereas levels of mg2046 genes were significantly lower (Figure 2E,F), indicating a unidirectional positive regulatory effect of mgfnr on mg2046.

Energy-saving wastewater treatment plants have received increasing research attention [26]. Microaerobic wastewater treatment has the potential to reduce energy consumption and improve operational efficiency [27,28]. Required DO concentration in microaerobic reactors is as low as 0.3–1.0 mg·L⁻¹ and can save aeration energy [29]. mgfnr genes had a negative regulatory effect on the pathway during all growth periods, suggesting that Δmgfnr mutation could potentially increase nitrate utilization efficiency. We utilized a microaerophilic MTB strain, MSR-1, to remove nitrate from synthetic wastewater medium. Δmgfnr is a promising candidate for industrial nitrate removal from wastewater.

3.3. Phenotypic Analysis of WT and Δmg2046/Δmgfnr

Large quantities of activated sludge (an umbrella term for microbial communities and the organic and inorganic substances on which they depend [30]) are typically produced during sewage treatment. During denitrification processes, sludge is distributed throughout the sewage system and needs to be collected for treatment. Ease of solid/liquid separation is directly related to treatment efficiency of activated sludge. The activated sludge with MTB can be quickly separated from the processing wastewater under an added magnetic field, which presents a unique advantage in sewage treatment. An important consideration in studying the denitrification power of mutant strains is that their magnetosome synthesis capability be fairly unaffected in order to facilitate application of magnetic adsorption.

WT and Δmgfnr MSR-1 strains were observed by TEM, and magnetosome numbers and particle sizes were quantified. In WT, magnetosomes were densely arranged and had uniform particle size. In Δmg2046, magnetosomes were less dense and had shorter chains. Δmgfnr showed greater synthesis of irregular magnetosomes, including some with small size (Figure 3A). Relative to WT, magnetosome number and particle size were significantly smaller for Δmg2046, but similar for Δmgfnr (Figure 3B).

MTB were cultured in a thermostatic shaker. For construction of the Cmag curve, vertical and horizontal magnetic fields were added by modifying the spectrophotometer, measuring absorbance, and calculating Cmag values to reflect differences in number and maturity of magnetosomes. OD₅₆₅ growth curves of Δmgfnr/Δmg2046 were similar to that of WT (Figure 3C). Comparison of Cmag curves showed that magnetosome synthesis in Δmg2046 began at 12 h and that Cmag value was significantly lower than that of WT, indicating slower magnetosome synthesis and reduced magnetic response. The magnetic response curve of WT was similar to that of Δmgfnr, indicating that absence of mgfnr did not strongly affect magnetosome synthesis (Figure 3D). Iron absorption rates were calculated at 6-h intervals, based on measurement of residual iron content in cells. Rates for Δmg2046/Δmgfnr were lower than for WT over various time periods. Rates for Δmgfnr were higher than those for Δmg2046 during the 0–6 h and 18–24 h periods, consistent with the higher Cmag value of Δmgfnr (Figure 3E).

3.4. Transcription of Genes Involved in Magnetosome Synthesis

The phenotypic analyses clearly showed that magnetosome synthesis was altered by mutations of mg2046 and mgfnr. Changes in gene transcription levels of magnetosome islands (MAI) were evaluated by examining key MAI genes mamA, mamB, mamF, and mamY. The mamAB gene cluster plays a key role in magnetosome synthesis [31]. Transcription levels of the four genes during various periods were significantly higher for WT than for Δmg2046, consistent with the reduced magnetosome synthesis and magnetic response of Δmg2046 during cell growth (Figure 4A,B). Transcription levels of the four genes during the most crucial magnetosome synthesis period (6–12 h) did not differ significantly between Δmgfnr and WT. During the 18–24 h period, mamA, mamB, and mamY levels were significantly lower for Δmgfnr than for WT (Figure 4C,D). These findings suggest that magnetosome synthesis in Δmg2046 was blocked, and abnormal crystals were generated during early cell growth stages because of reduced expression of the four genes. Gene transcription levels for Δmgfnr during the crucial 6–12 h period did not differ notably from those for WT, indicating that this mutation had no major effect on magnetosome synthesis and will not affect magnetic adsorption sedimentation. In contrast, normal functioning of mg2046 plays an important role in magnetosome synthesis through direct or indirect positive regulatory effects.

3.5. Comparative Utilization of Various Nitrogen Sources by WT and Δmg2046/Δmgfnr

Absorption and utilization capabilities of WT and Δmg2046/Δmgfnr for various nitrogen sources (ammonium [NH₄⁺], nitrate [NO₃⁻], ammonium/nitrate [NH₄⁺/NO₃⁻]) were evaluated. In NH₄⁺ medium, total nitrogen removal rates after 12 h culture for WT, Δmg2046, and Δmgfnr were respectively 0.363, 0.178, and 0.393 mmol·L⁻¹·h⁻¹, and total nitrogen content in medium did not differ notably for WT vs. Δmgfnr (Figure 5A,B). Total nitrogen content in medium, relative to WT, was significantly higher for Δmg2046. These findings suggest that NH₄⁺ absorption and utilization capabilities were similar for WT and Δmgfnr but lower for Δmg2046.

Total nitrogen removal rates after 12 h culture for WT and Δmg2046/Δmgfnr were respectively 0.302, 0.234, and 0.412 mmol·L⁻¹·h⁻¹ in NO₃⁻ medium, and 0.454, 0.304, and 0.505 mmol·L⁻¹·h⁻¹ in NH₄⁺/NO₃⁻ medium (Figure 5C–F). Total nitrogen content in medium, relative to WT, was significantly higher for Δmg2046 but lower for Δmgfnr, and nitrogen removal rates were respectively 69.20%, 65.09%, and 80.50%. Total nitrogen content in bacteria, relative to WT, was significantly lower for Δmg2046 but higher for Δmgfnr. These findings indicate that NO₃⁻ and NH₄⁺/NO₃⁻ absorption and utilization capabilities of Δmgfnr, relative to WT, are significantly higher, and it has superior pollution treatment capacity for major nitrogen-source nitrates in the environment. Δmgfnr can use NO₃⁻ and NH₄⁺ simultaneously as nitrogen sources and adapt to a variety of sewage environments.

3.6. Comparative Utilization of Nitrogen Sources from Synthetic Wastewater by WT and Δmg2046/Δmgfnr

The three bacterial strains were cultured in synthetic wastewater to evaluate feasibility of using MTB for treatment of industrial wastewater from the stainless steel industry. Samples were taken at 0, 12, and 24 h for determination of total nitrogen content in medium and in bacteria. Total nitrogen levels did not differ notably at 0 h. Total nitrogen removal rates in medium for WT, Δmg2046, and Δmgfnr were respectively 0.362, 0.302, and 0.456 mmol·L⁻¹·h⁻¹ at 0–12 h, and 0.210, 0.205, and 0.155 mmol·L⁻¹·h⁻¹ at 12–24 h (Figure 6A). Total nitrogen content in bacteria generally increased with culture time, and relative to WT was higher for Δmgfnr and lower for Δmg2046 (Figure 6B). Thus, denitrification efficiency in synthetic wastewater declined for Δmg2046 but increased for Δmgfnr (0.456 mmol·L⁻¹·h⁻¹ during 0–12 h; 0.155 mmol·L⁻¹·h⁻¹ during 12–24 h). Total nitrogen removal rates for Δmgfnr were 64.09% at 12 h and 85.83% at 24 h. Δmgfnr is potentially useful for nitrate processing in synthetic industrial wastewater from the stainless steel industry, with a variety of applications.

Feasibility of magnetic adsorption by the three MTB strains was evaluated by placing cultures above a strong magnet and measuring OD₅₆₅; cultures without a magnet were used as controls. Sedimentation rates of the three control groups were similar (OD₅₆₅ decreased from 1.0 at 0 h to 0.3 at 148 h); the rates declined rapidly from 12 to 96 h but showed minimal change from 96 to148 h (Figure 6C). Magnetic adsorption rates were significantly higher for experimental groups than control groups (Figure 6D). OD₅₆₅ for experimental groups declined from 1.0 at 0 h to 0.15–0.3 at 24 h, whereas OD₅₆₅ for control groups was ~0.7 at 24 h. OD₅₆₅ for WT and Δmgfnr was near 0 at 72 h, whereas that for Δmg2046 approached 0 at 96 h, presumably because magnetosome synthesis was blocked. These findings suggest that magnetic adsorption enhances sedimentation rates for MTB, reduces energy consumption required for separation of activated sludge, and has a wide variety of potential applications.

The traditional activated sludge system for wastewater treatment has always been the object of a lot of research [32,33]. Wastewater treatments using current activated sludge process (ASP) inevitably produce waste biomass. The water treatment industry has developed rapidly during the past decade, with associated increase in amount of activated sludge produced by sewage processing [34]. Various pretreatment methods for active sludge (including chemical, mechanical, biological, and thermal hydrolysis methods) have been developed [35,36]. For all of these, improved separation of liquid and solid components will result in greater energy efficiency, better sludge dewater ability, and reduced sludge viscosity, making the method more feasible and practicable [37]. The fact that MTB can be easily separated from wastewater using magnetic fields [18] is a unique feature that makes Δmgfnr promising candidates for practical application in sludge pretreatment.

4. Conclusions

Proteins Mg2046 and Mgfnr were shown to directly regulate genes involved in the dissimilatory denitrification pathway. M. gryphiswaldense MSR-1 deletion mutant strains Δmg2046 and Δmgfnr were constructed, and their functions and relationships were evaluated based on phenotypic and genetic characteristics. mgfnr mutation did not notably affect magnetosome synthesis, and transcription levels of dissimilatory denitrification genes for Δmgfnr were higher than those for WT. Total nitrogen removal rates in NH₄⁺, NO₃⁻, NH₄⁺/NO₃⁻, and synthetic wastewater medium were significantly higher for Δmgfnr than for WT or Δmg2046. Magnetic adsorption was shown experimentally to significantly increase the sedimentation rate of activated sludge for MTB.