Water Research X 1 (2018) 100009

Contents lists available at ScienceDirect

Water Research X
journal homepage: https://www.journals.elsevier.com/water-research-x

Interference of manganese removal by biologically-mediated

reductive release of manganese from MnOx(s) coated filtration media
Lindsay E. Swain a, William R. Knocke a, **, Joseph O. Falkinham III b, Amy Pruden a, *
a
Via Department of Civil and Environmental Engineering, Virginia Tech, Blacksburg, VA 24061, USA
b
Biological Sciences Department, Virginia Tech, Blacksburg, VA 24061, USA

a r t i c l e i n f o a b s t r a c t

Article history: Discontinuing application of pre-filter chlorine is a common water treatment plant practice to permit a
Received 29 May 2018 bioactive filtration process for the removal of soluble Mn. However, soluble Mn desorption has some-
Received in revised form times been observed following cessation of chlorine addition, where filter effluent Mn concentration
30 October 2018
exceeds the influent Mn concentration. In this paper it is hypothesized that Mn-reducing bacteria pre-
Accepted 4 November 2018
Available online 13 November 2018
sent in a biofilm on the filter media may be a factor in this Mn-release phenomenon. The primary
objective of this research was to assess the role of Mn-reducing microorganisms in the release of soluble
Mn from MnOx(s)-coated filter media following interruption of pre-filtration chlorination. Bench-scale
Keywords:
Drinking water treatment
filter column studies were inoculated with Shewanella oneidensis MR-1 to investigate the impacts of a
Manganese removal known Mn-reducing bacterium on release of soluble Mn from MnOx(s) coatings. In situ vial assays were
Media filtration developed to gain insight into the impacts of MnOx(s) age on bioavailability to Mn-reducing microor-
Manganese oxides ganisms and a quantitative polymerase chain reaction (qPCR) method was developed to quantify gene
Manganese-reducing bacteria copies of the mtrB gene, which is involved in Mn-reduction. Results demonstrated that microbially-
mediated Mn release was possible above a threshold equivalent of 2
102 S. oneidensis MR-1 CFU per
gram of MnOx(s) coated media and that those organisms contributed to Mn desorption and release.
Further, detectable mtrB gene copies were associated with observed Mn desorption. Lastly, MnOx(s) age
appeared to play a role in Mn reduction and subsequent release, where MnOx(s) solids of greater age
indicated lower bioavailability. These findings can help inform means of preventing soluble Mn release
from drinking water treatment plant filters.
© 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).

1. Introduction Microbial metal reduction is central to the geochemical cycling

of Fe, Mn and C in redox-stratified waters (Johnson, 2006). Yet, little
Mn is a challenging drinking water contaminant, owing to its is known about the microbial metabolic mechanisms by which the
widespread natural occurrence and multiple oxidation states (e.g., Mn reduction process takes place. In contrast to reduction of iron,
Mn(II), Mn(III), Mn(IV) in water). Elevated Mn concentrations the current state of understanding Mn-reduction is that it does not
(SMCL ¼ 0.05 mg/L in U.S.) can present a significant aesthetic provide energy to the cell (Lin et al., 2012). Under neutral pH
concern as the water travels through the distribution system and conditions, Mn oxides are highly insoluble and can be crystalline in
into a consumer's home plumbing. Upon entering the consumer's nature (Yang et al., 2013). For reduction to occur, microorganisms
home, water that contains oxidized Mn is often characterized by a are required to transfer electrons to Mn oxides that are external to
blackish-brownish color, which can lead to water discoloration the cell wall or membrane, since contact with a localized inner
complaints (Sly et al., 1990) and an undesirable metallic taste (Sain membrane electron transport chain is not possible (Lovley et al.,
et al., 2014). 2004). It is generally thought that Mn(IV) reduction occurs
directly at the outer membrane via single two-electron successive
transfers, resulting in Mn(II) as the final product (Thamdrup, 2000).
* Corresponding author. Via Department of Civil and Environmental Engineering, The first electron transfer forms soluble Mn(III) as a temporary
Virginia Tech, Blacksburg, VA 24061, USA.
intermediate before a final electron transfer and reduction to
** Corresponding author. Via Department of Civil and Environmental Engineering,
Virginia Tech, Blacksburg, VA 24061, USA.
Mn(II). The first electron transfer step increases the bioavailability
E-mail addresses: [email protected] (W.R. Knocke), [email protected] (A. Pruden). of Mn by reductive solubilization, while the second step is coupled

https://doi.org/10.1016/j.wroa.2018.100009
2589-9147/© 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
2 L.E. Swain et al. / Water Research X 1 (2018) 100009

to the production of inorganic carbon (Lin et al., 2012). methods applicable to this study. More detailed information may
Shewanella has been used as a model metal-reducing bacterium be found in Swain (2016).
due to its ability to utilize a wide variety of electron acceptors and
its capacity to transfer electrons to solid metal oxide compounds 2.1. Laboratory bench-scale filter column experimental and
(Osterman et al., 2008). The electron transport chain of S. oneidensis analytical procedures
is made up of several different interacting proteins, which allow the
utilization of extracellular electron acceptors for Mn reduction Bench-scale filter columns were set up to produce Mn break-
(Szeinbaum et al., 2014). The main outer membrane protein com- through curves as well as replicate Mn removal and desorption
plex is MtrCBA, an electron channel that allows their passage to trends observed in full-scale WTPs. Each laboratory bench-scale
extracellular metal hydroxide complexes (Burnes, 2000). MtrCBA is filter contained a filter media (sand or anthracite), received an
mainly associated with the reduction of Mn(IV). An outer mem- influent that contained soluble Mn and aluminum (Al, primarily as
brane b-barrel protein, MtrB, allows interaction and electron Al(OH)3(s)), and had free chlorine applied for a defined initial
transfer to extracellular metal hydroxide complexes and has been amount of time to promote formation of MnOx(s) coating on the
shown to be required for Mn(IV) and Mn(III) reduction (Lin et al., media and active site regeneration for the consistent uptake of
2012). soluble Mn. Particulate Al(OH)3(s) was added at a low concentra-
One of the more common methods of soluble Mn removal in tion (approximately 200 mg/L as Al) to mimic the presence of
water treatment facilities is through adsorption onto MnOx(s)- carryover floc from sedimentation basins in “real-world” surface
coated filtration media (Knocke et al., 1991; Tobiason et al., 2008). WTPs (Jones, 2012). During each study, free chlorine application
This process has been referred to as the natural greensand effect ceased after a set amount of time and the column influent and
and is a self-regenerating process that continues to remove addi- effluent were measured to evaluate the amount of soluble Mn
tional soluble Mn (Brandhuber et al., 2013). In a few water treat- passing through the column media.
ment plants (WTPs) where adsorption of soluble Mn to oxide- Glass columns were 24 inches in height and had a 7/16-inch
coated filtration media is the main method utilized for soluble inner diameter. Plastic tubing was employed to transfer feed solu-
Mn removal, Mn desorption and release from the media has been tions to each column and transport filtered effluent away from the
observed upon cessation of pre-filter free chlorine addition (Islam, column. Six inches of media depth was used in each column
2010). In certain cases, effluent Mn exceeded the influent Mn experiment.
concentration for extended periods of time (e.g., 1e2 weeks) after Influent feed solutions were prepared with deionized water in
chlorine application ceased (Gabelich et al., 2006). This desorption five-gallon increments and stored in larger plastic reservoirs. Each
phenomenon has specifically been documented at the Aquarion column had two influent feed solutions that were mixed just prior
Water Company (AWC) Stamford WTP (Stamford, CT) (Tobiason to entering the filter column. The primary feed solution contained
et al., 2008) and at the Henry J. Mills WTP (Riverside, CA) dissolved Mn, alkalinity, dissolved and particulate Al. The second
(Gabelich et al., 2006). Mn desorption from coated filter media has feed solution always contained alkalinity and contained free chlo-
also been observed in the laboratory using filtration media from the rine during the time period when an experiment required the
AWC Lantern Hill WTP (Stonington, CT) (Islam, 2010). Mn release presence of free chlorine in the filter-applied water. Once mixed
appeared to be due to biologically-mediated desorption from the together the resulting influent feed solution characteristics of the
anthracite filtration media. In addition, Islam (2010) indicated that water for each experiment were achieved.
Mn-reducing populations were only capable of growth on filtration Columns were backwashed as needed when excessive head loss
media and reduction of Mn in the absence of chlorine. Consistent occurred or at least once daily. Backwashing was performed by
with the above observations, Cerrato et al. (2010) were able to pumping 1 L of deionized water in an upflow direction through the
isolate Mn-reducing bacteria from the media of WTPs operating in column at a rate of 25 gpm/ft2. This flow rate corresponded to an
Virginia and North Carolina. Notably, some of these bacteria, mainly approximate 30% media expansion.
Bacillus, were capable of both oxidizing and reducing Mn. Further, There were two distinct sets of filter column experiments on
recent studies have revealed that a wide diversity of microbes two different MnOx(s)-coated filter medias to assess the potential
inhabit WTP media, even under chlorinated conditions (Chiao et al., role of Mn-reducing bacteria on Mn release. Those procedures are
2014). provided in the following text.
The overall goal of this research was to evaluate the potential for
biologically-mediated Mn-reduction from MnOx(s)-coated filtra- 2.1.1. Abiotic studies for assessing potential soluble Mn release from
tion media during drinking water treatment. The specific objectives media
were to: 1) characterize Mn release from columns with varying An initial set of experiments were designed to assess the po-
extent of MnOx(s) coating on the media under conditions of reduced tential for soluble Mn release from filter media under abiotic con-
microbial activity; 2) compare Mn-desorption from laboratory- ditions. These served as a “control” to determine whether soluble
scale filter columns containing porous media with and without Mn release would occur from MnOx(s)-coated media purely by
inoculation of S. oneidensis MR-1, a known Mn-reducing microor- chemical means (e.g., cessation of pre-filter chlorine addition).
ganism; 3) develop and apply a molecular assay for the quantifi- These experiments employed the use of new FilterSil sand (uni-
cation of the mtrB marker gene associated with biological Mn- formity coefficient 2.65; effective size 0.5 mm), which initially had
reduction; and 4) investigate the effects of MnOx(s) age on no MnOx(s) coating present. The media was autoclaved prior to the
bioavailability and Mn-reduction capacity. The goal was to generate start of MnOx(s) coating and laboratory experiments to minimize
information to inform improved operation of media filters at WTPs microbial activity. An initial MnOx(s) coating was generated on the
to optimize biological benefits of removing organic carbon via sand media through application of soluble Mn (200 mg/L) and free
biofiltration while minimizing potential concerns related to Mn chlorine (1e2 mg/L) for a defined period (either 5 or 15 days).
release. Extraction of the media coating by hydroxylamine sulfate addition
indicated an MnOx(s) coating level of 1.5 mg Mn/g media and
2. Experimental methods and materials 4.9 mg Mn/g media after 5 and 15 days of free chlorine application
respectively. This insured the presence of ample MnOx(s) coating
The following text describes the experimental and analytical for evaluating the potential for Mn reduction and release after free
 L.E. Swain et al. / Water Research X 1 (2018) 100009 3

chlorine addition ceased in these experiments. Additional influent potential led to Mn release from the media coatings.
water characteristics to the sand media were: pH e 7.0 to 7.3,
alkalinity - 2 meq/L, total Al e 200 mg/L, and TOC - <0.5 mg/L. 2.1.3. Analytical methods for filter column studies
Influent feed solutions were pumped into the column to yield a The pH of the influent feed solutions to the filter columns was
total hydraulic loading rate of 4 gpm/ft2. monitored and adjusted with concentrated hydrochloric acid (HCl)
Following the initial period (5 or 15 days) of MnOx(s) coating as needed. A HACH HQ40d pH meter was calibrated daily before
deposition on the sand, the free chlorine was eliminated from the measurements were taken. A HACH reagent set (DPD method) was
feed water. Column effluent was monitored for the presence of utilized for the measurement of free chlorine. Influent and effluent
soluble Mn to evaluate the potential for actual release of Mn from bench-scale column samples for Mn analysis were preserved in 2%
the media coating. trace metal grade nitric acid and were analyzed using inductively
coupled plasma mass spectroscopy (ICP-MS).
2.1.2. Experiments employing Mn-reducing bacteria to assess A description of the term “soluble Mn” is provided for clarity.
potential for Mn release from media coatings The influent solution to any media column contained only dis-
A second set of experiments sought to assess the potential role solved manganous (Mn(II)) ion. Likewise, the relatively brief (less
of Mn-reducing bacteria in promoting Mn release from MnOx(s)- than 2 min) contact time between the dissolved Mn(II) and free
coated media. These experiments employed anthracite coal media chlorine prior to application to the coated media was too short to
(uniformity coefficient 1.4; effective size 0.9e1.0 mm) obtained promote any significant Mn(II) oxidation and formation of partic-
from the Harwood Mills Waterworks WTP (Newport News, VA). ulate MnOx(s) in the water applied to the media (Brandhuber et al.,
The Harwoods Mill media was chosen for use due to its significant 2013). Preliminary testing with 0.45 mm membrane filtration of
(28e30 mg Mn/g media) MnOx(s) coating that was present. This column effluent samples showed no significant presence of par-
WTP employs free chlorine addition to the filter-applied water for ticulate, oxidized MnOx(s) being present, which is consistent with
soluble Mn control. practical experience when soluble Mn is the only source in the feed.
The influent characteristics for the anthracite media experiment As such the protocol for experimental testing did not involve 0.45
were as follows: pH e 6.3 to 6.6, alkalinity 2 meq/L, soluble Mn e mm filtration of the column effluent samples and assumed that Mn
50 mg/L, and total Al e 200 mg/L. The slightly lower pH condition present in the column effluent was in the soluble form.
employed (in relation to the abiotic study) was due to the fact that
the typical filter-applied pH at the Harwoods Mill WTP is between 2.2. Molecular detection of Mn-Reducing microorganisms
6.0 and 6.5. As such, it was desired to use a pH application similar to
what the media experienced in the WTP. A free chlorine concen- 2.2.1. Sample collection for molecular analyses
tration of 0.5e1.0 mg/L (again, comparable to Harwoods Mill WTP DNA was isolated from standard pure culture suspensions, lab-
practice) was present in both columns for the initial five-day period oratory samples, and environmental samples using the PowerSoil
of the test, after which chlorine was no longer fed to the column. kit from MoBio Laboratories, Inc. (Solon, OH) for use in quantitative
Whereas the abiotic study had no TOC added to the filter-applied polymerase chain reaction (qPCR). For extraction from anthracite
water, this experiment incorporated a small amount of sodium media, 0.25 g of each sample was measured and transferred to each
acetate (influent TOC ¼ 0.5 mg/L as C) to serve as a readily extraction kit vial. For liquid samples and standards, suspensions
degradable organic substrate. were mixed and 250 mL was added into the extraction kit vials.
One of the columns in this study was inoculated with Samples and standards were stored at 20  C until extraction.
S. oneidensis MR-1 (a well-characterized Mn-reducing bacterium)
while the other column had no microbes applied to the filter media. 2.2.2. S. oneidensis MR-1 qPCR
S. oneidensis MR-1 was obtained courtesy of Dr. Kenneth Nealson The S. oneidensis MR-1 16S rRNA gene RT-qPCR methods
from the University of Southern California. To prepare the described by Schicklberger et al. (2011) were modified for qPCR.
S. oneidensis MR-1 inoculum, a 10-mL sterile loop full of culture from SsoFast EvaGreen Supermix (Bio-Rad, Hercules, CA) was used, with
an R2A agar (BD Difco) plate was transferred to 100 mL of R2A broth a reaction volume of 10 mL. The mastermix contained 5 mL of SsoFast
(BD Difco). The flask was incubated for 48 h in a shaking water bath EvaGreen supermix, 0.8 mL of the forward and reverse primers,
at 30  C. S. oneidensis MR-1 was then used to inoculate a 1 L volume 2.4 mL molecular grade water and 1 mL DNA template. The DNA
of R2A broth and again incubated for 48 h in a shaking water bath at template was quantified using a Qubit Fluorometer (Bio-Rad) and
30  C. The entire volume of culture was further allowed to accli- diluted to a concentration where 5e15 ng/mL was added to each
mate to room temperature for an additional 24 h. The culture was well before use in qPCR. The optimal melting temperature was
centrifuged at 5,000 g for 20 min to concentrate the pellet. The determined by a temperature gradient and melt curve ranging from
pellet was suspended into a 200 mL volume of modified Mn- 57.0  C to 51.0  C, where 57.0  C was chosen as the optimal melting
reduction broth, which contained per liter of 10 mM HEPES buffer temperature for this protocol. A standard curve was developed
(pH 7.4): 0.2 g yeast extract and 2 g sodium acetate. The modified using DNA extracts from suspensions of S. oneidensis MR-1 culture
Mn-reduction broth, with concentrated S. oneidensis MR-1 inoc- of known colony-forming units per mL density to quantify this
ulum, was poured over the anthracite media in the column. The organism from laboratory samples and bench-scale filter column
culture was allowed to flow through the column for a short time studies.
and then was used to fill the column above the media level. The
column was left undisturbed for 24 h, after which it was lightly 2.2.3. mtrB gene detection
backwashed. Additional culture was added to the column and was Multiple sequence comparison by log-expectation “MUSCLE”
again subject to a short duration of flow. Culture filled the column alignment (Edgar, 2004) was chosen as the method to aid in design
above the media height and was left undisturbed for an additional of the mtrB primer set. From the NCBI GenBank website, 10 MtrB
24 h. After the period of column inoculation, influent flow was protein sequences were chosen for gene alignment based off the
introduced to the filter. targeted sequence description (Table S1). The final mtrB primers
These two columns were monitored for effluent Mn concen- obtained from the alignment is as follows: Forward 50 -CSTTCAAC-
trations before and after the cessation of pre-filter chlorine to VACATGGCCG-30 and Reverse 50 -SGAGATCTCSAGCAGGTC-3’.
evaluate whether the presence of the Mn-reducing bacteria For a 10-mL qPCR reaction volume, the mastermix contained 5 mL
4 L.E. Swain et al. / Water Research X 1 (2018) 100009

of SsoFast EvaGreen supermix, 0.8 mL of the forward and reverse sterile deionized water was added into the negative (uninoculated)
primers, 2.4 mL of molecular grade water and 1 mL of DNA template. control vials.
DNA template was added into the reaction wells at a concentration Percent transmittance of light at 540 nm was used as a
between 5 and 15 ng/mL. Annealing temperature ranges were tested comparative surrogate parameter for semi-quantitatively esti-
between 50.0  C and 58.0  C, where 50.0  C was chosen as optimal. mating the degree of Mn reduction occurring within the vial. As
To validate the assay, strains of Bacillus spp. that were verified not to MnOx(s) was reduced to soluble Mn, the percent transmittance of
have the capability to reduce Mn were analyzed as a the suspension in the vial increased. Percent light transmittance
phylogenetically-related negative control. This validation was measured frequently (a minimum of every 1e2 days) using a
confirmed non-detect levels of the mtrB gene for the negative spectrophotometer (Coleman Junior, Model 6A, Maywood, IL) that
control strains. was calibrated at the 540 nm wavelength before each use (Gerhardt
et al., 1981). Abiotically-reduced positive control vials were pre-
2.3. In-situ Mn-Reduction vial assay pared by adding 0.020 g of hydroxylamine sulfate suspended in
100-mL of sterile deionized water was added into a parallel set of
An in situ vial assay was developed to semi-quantitatively vials for each type of MnOx(s) sample. This allowed for an estimate
evaluate the bioavailability to S. oneidensis MR-1 of various aged of the percent transmittance of a sample wherein essentially 100%
MnOx(s) samples. Various modifications, detailed in the following of the MnOx(s) added was reduced to soluble Mn.
sections, were made from the protocol established by Nealson et al.,
(1991). A total of five MnOx(s) samples were examined using the vial
assay developed. Three of the MnOx(s) samples were prepared by 3. Results and discussion
adding potassium permanganate (KMnO4) at a stoichiometric
dosage to waters containing soluble Mn (obtained by dissolving 3.1. Abiotic filter media study - Mn sorption and release from
manganous chloride (MnCl2)). Once the reaction was complete, the MnOx(s)-Coated virgin sand
MnOx(s) formed was centrifuged at 10,000 g for 5 min to form a
pellet and remove excess water. The MnOx(s) samples were washed An initial experiment involved an abiotic study of the potential
by resuspension in nanopure water, and re-centrifuged to form a reduction of MnOx(s) media coatings. MnOx(s) coatings were
new pellet three times. These solids were precipitated approxi- deposited onto autoclaved sand by interaction between soluble Mn
mately one to seven months prior to use in the vial assay, and were and free chlorine on the media for an initial period of either 5 or 15
identified according to synthesis date (7-28-15, 11-18-15 and 1-11- days (the longer time led to greater MnOx(s) deposition on the
16). A fourth MnOx(s) sample (2008) was a purchased MnO2(s) sand). After the prescribed period had elapsed, the free chlorine
oxide from Sigma that had been stored in the laboratory for was removed from the filter-applied water and only soluble Mn
approximately 7 years. The fifth MnOx(s) sample (7-8-15) was loading continued.
artificially “aged” by drying the formed MnOx(s) at 103  C for 24 h. Results from this abiotic study are shown in Fig. 1. Time zero was
An attempt was made to inactivate any microbes possibly present set as the time chlorine addition ceased for each column. Soluble
on the MnOx(s) by exposure to an extremely strong free chlorine Mn breakthrough occurred more quickly when the media in the
environment. Chlorine inactivation of the MnOx(s) samples was columns had been coated for only five days. The rate of soluble Mn
completed by soaking the samples in a strong chlorine solution for removal is known to be directly associated with surface MnOx(s)
sufficient time to yield a Ct product of 10,000 mg/L-min. After the concentration, which impacts the amount of available active sites
desired Ct of 10,000 mg/L-min was reached, the MnOx(s) samples on a filter media (Knocke et al., 1991). Faster breakthrough from the
were washed three times using sterile deionized water. 5-day coated columns would be expected, as the MnOx(s) coating
A semisolid manganese reduction agar contained the following level was only 1.5 mg Mn/g media whereas 4.9 mg Mn/g media of
per liter of 10 mM HEPES buffer (pH 7.4): 0.2 g yeast extract, 2 g MnOx(s) coating was found on the 15-day coated media. Knocke
sodium acetate and 3 g agar (0.3%). The media was autoclaved on a et al. (1991) demonstrated that the number of MnOx(s) active sites
20-min liquid cycle and was allowed to cool before the addition of
MnOx(s). To determine the responsive range of the assay, percent
transmission at 540 nm was measured for lab-synthesized MnOx(s)
solids (1-11-16 and 2008 samples) over the range of 0e1.0 g/L.
Percent transmission versus MnOx(s) solids concentration fit a
polynomial curve with an R2 > 0.95. Based on this, a target initial
MnOx(s) sample addition of 0.7 or 0.35 g/L was selected in order to
maintain the assay within a measurable range over the duration of
the assay. As noted above, the commercial MnO2(s) solids and the
dried (at 103  C) MnOx(s) solids were much darker in color and thus
were added at the lower initial concentration (0.35 g/L) to achieve a
comparable initial transmittance measurement of 4e8% across the
vial assays. A volume of 10 mL of the Mn-reduction agar, with
MnOx(s) added, was placed into glass vials for inoculation with the
S. oneidensis MR-1 bacteria.
Vials were inoculated with fresh S. oneidensis MR-1 cultures
from R2A agar plates, which were incubated overnight at 30  C.
Cultures were removed from the agar plate with a sterile loop and
suspended in 1.5 mL of sterile deionized water and subject to serial
dilution. A 100-mL aliquot of the desired S. oneidensis MR-1 sus-
Fig. 1. Comparison of the percentage manganese measured in effluent of columns
pension was added into each vial and was vortexed. Each inoculum containing sand media coated with MnOx(s) for 5 days versus 15 days, relative to
was enumerated using standard heterotrophic pour plate methods influent feed concentration (0.2 mg/L Mn), after the discontinuation of free chlorine in
on R2A agar before the addition into vials. A volume of 100 mL of the feed.
 L.E. Swain et al. / Water Research X 1 (2018) 100009 5

positively correlate with the degree of MnOx(s) coating. Therefore, substrate that could have been coupled with the reduction of
the column with the 15-day MnOx(s) coating had increased MnOx(s) from the media surface by the Mn-reducing microorgan-
adsorption capacity on the media surface, which resulted in slower isms. This result compares well to Islam (2010), who were able to
Mn breakthrough. demonstrate a relationship between applied acetate concentration
Importantly, the concentration of Mn passing through the media to MnOx(s)-coated media and subsequent soluble Mn release from
depth never exceeded influent Mn concentrations for either the 5- that media.
day or 15-day coated columns. The breakthrough trends from this
column experiment can be explained by the loss of Mn uptake 3.2.1. S. oneidensis MR-1 quantification
capacity on the media sites due to Mn uptake becoming saturated, Relative abundances of S. oneidensis MR-1 occurring in the top,
given that regeneration with free chlorine was no longer taking middle and bottom two-inch sections of the inoculated and unin-
place. The cessation of free chlorine to filters in some full-scale WTP oculated columns were measured by qPCR after the completion of
(Gabelich et al., 2006) and laboratory studies (Islam, 2010) have column operation. Detection of S. oneidensis MR-1 was confirmed at
shown Mn desorption that results in effluent Mn concentrations all depths in the inoculated column at an average density across the
exceeding the influent for a period of several hours up to 1e2 three layers of 3.9 ± 0.89
102 CFU/mL. S. oneidensis MR-1 was not
weeks, apparently due to reduction of Mn from media coatings and detected at any depth in the uninoculated column. These results
release to the filter effluent. These abiotic column studies did not indicate that the process used to inoculate S. oneidensis MR-1 on the
yield situations where the column effluent Mn concentration anthracite media was successful and supports the hypothesis that
exceeded the influent Mn level of 0.2 mg/L. As such, operation of the presence of these Mn-reducing bacteria contributed to the
the columns under abiotic conditions did not yield evidence of Mn additional observed Mn-release. To the authors’ knowledge, no
reduction in the available MnOx(s) media coatings, suggesting a prior study has specifically confirmed the presence of S. oneidensis
lack of purely chemical mechanisms for promoting Mn reduction specifically on WTP media, though other species of Mn reducers
on the media. have been identified (Cerrato et al., 2010). In this specific study,
S. oneidensis MR-1 was selected as a well-known and highly char-
acterized model, with well-defined gene targets for monitoring,
3.2. Laboratory bench-scale filter study - Mn release from harwood
and thus served well in its purpose of representing conditions with
mills anthracite media with and without inoculum
confirmed Mn-reducers inoculated.
Relative abundances of mtrB genes in the top, middle and bot-
Data related to the passage of soluble Mn through the anthracite
tom two-inch sections of the inoculated and uninoculated columns
media filter depth for both the “Inoculated Column” (S. oneidensis
were also measured by qPCR. Results indicated that the
MR-1 inoculum) and “Uninoculated Column” are presented in
S. oneidensis MR-1 inoculated columns contained more mtrB gene
Fig. 2. During the initial five days when free chlorine was applied,
copies. The average Ct value for the S. oneidensis MR-1 inoculated
the Mn breakthrough in the inoculated column was consistently
columns was 27 ± 0.26 compared to uninoculated column, which
greater when compared to the uninoculated column. Likewise, af-
had an average Ct value of 38 ± 0.21 (where Ct values are inversely
ter free chlorine addition ceased, the release of soluble Mn was
proportional to target gene copy numbers). These findings further
much greater for the inoculated column. Experimental conditions
support the conclusion that there were far more microorganisms
(e.g., pH, applied Mn concentration, applied free chlorine concen-
that had the ability to reduce Mn in the inoculated column.
tration, chlorine demand across the media depth, time of chlorine
feed cessation, etc.) between both columns remained the same
3.2.2. Potential effects of aging on the bioavailability of Mn in
through the duration of the experiment. As such it is hypothesized
MnOx(s) coatings
that the difference in Mn breakthrough observed could be attrib-
Fig. 3 presents normalized transmittance values, equivalent to a
uted to the Mn reduction activity of the S. oneidensis MR-1
1 g/L MnOx(s) agar vial. Normalization of transmittance measure-
inoculum.
ments to that of a 1 g/L MnOx(s) agar vial enabled relative com-
Lovley and Phillips (1988) showed that S. oneidensis MR-1
parisons of Mn reduction capacities. Notably, the 2008 and 7-8-15
coupled the oxidation of electron donors (similar to the acetate
solids were characterized by the least extent of Mn reduction over
supplied in the feed to these two columns) to Mn reduction. In this
the duration of the study. All the other MnOx(s) solids evaluated in
specific experiment, acetate was supplied as a readily oxidized
this study were comparable in the observed extent of Mn reduction
and corresponding increase in sample percent transmittance.
Average S. oneidensis MR-1 cell counts were compared among
triplicate vials via qPCR over the course of the MnOx(s) incubations
(Fig. 4). An ANOVA test confirmed that mean cell concentrations
were highly similar across all the MnOx(s) agar vials throughout the
experiment (F < Fcrit; P < 0.05). Nonetheless, differences in Mn
reduction among the vials were clearly observed (Fig. 3), indicating
that variances in S. oneidensis MR-1 populations between vials were
not responsible for the differences. Further, for the 2008 and 7-8-15
MnOx(s) samples, even though the S. oneidensis MR-1 count
increased over time, Mn reduction rates were slower than observed
for the other samples. The 7-28-15, 11-18-15 and 1-11-16 MnOx(s)
samples had the same density of S. oneidensis MR-1, but could
reduce more MnOx(s). Because the 2008 and 7-8-15 vials had
similar S. oneidensis MR-1 counts, but lower extent of observed Mn
reduction, this pointed to a reduced bioavailability of the MnOx(s)
present in these two more “aged” samples.
Fig. 2. Percentage of Mn measured in effluent of “Inoculated” (S. oneidensis MR-1) and The ability of microbes to reduce freshly precipitated MnOx(s)
“Uninoculated” (sterile inoculum) Columns. when coupled with the oxidation of various organic compounds
6 L.E. Swain et al. / Water Research X 1 (2018) 100009

promotes soluble Mn removal (leading to MnOx(s) deposition on the

media) and results in a significant amount of MnOx(s) accumulation
over years. Tobiason et al. (2008) showed that MnOx(s) coatings on
anthracite media from surface WTPs addressing seasonal elevated
Mn concentration issues often had a “tree-ring-like” structures,
with many bands of high MnOx(s) placed throughout the coating.
Such deposition over many years would provide opportunity for
chemical aging and structural changes in the MnOx(s) present that
could decrease Mn bioavailability from the coatings. This would
correspond with the observation (Knocke, published data) that
significant residual MnOx(s) coating may be present on filter media
following an observed Mn release event from the filter. The pre-
sumption would be that only the more recently deposited MnOx(s)
coating was available for microbial reduction due to its less crys-
talline structure.

4. Conclusions

Bench-scale filter column studies served to demonstrate the

Fig. 3. Change in transmission at 540 nm as a measure of extent of Mn reduction that
potential of Mn-reducing microorganisms (specifically S. oneidensis
occurred with time in vials containing MnOx(s) samples and uniformly inoculated with
S. oneidensis MR-1. Each plot represents a distinct MnOx(s) solid sample, named for the MR-1) to contribute to the Mn-desorption phenomenon occasion-
date it was collected or generated, initially added at 0.70 g/L for 07/28/15, 01/11/16, ally seen in WTPs when pre-filtration free chlorine is removed.
and 11/18/15 samples and 0.35 g/L for 2008 and 07/08/15 sample to achieve an initial Specific conclusions that can be derived from this study include:
transmission of 4e8%. Values plotted are averages of triplicate vial incubations,
normalized to transmittance measured for 1 g/L of the corresponding MnOx(s).
 Inoculation of column media with the Mn-reducer S. oneidensis
MR-1 was associated with increased rates of Mn desorption.
 S. oneidensis MR-1 residing on the MnOx(s) surface was able to
remain viable and reduce Mn after sustained (up to five days)
contact with free chlorine.
 MnOx(s) age and crystalline structure could play an important
role in the bioavailability of Mn to Mn-reducing organisms.

Declaration of interests

☒ The authors declare that they have no known competing

financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
☐The authors declare the following financial interests/personal
relationships which may be considered as potential competing
interests:

Acknowledgements

The authors thank Dr. Kenneth Nealson from the University of

Fig. 4. Average S. oneidensis MR-1 counts (CFU/mL equivalents) of triplicate vials Southern California for providing the S. oneidensis MR-1 strain
incubating different MnOx(s) samples (indicated by the dates in the legend) measured used in this study and Dr. Dongjuan Dai for internal review and
via qPCR.
assistance with figures. Financial support was provided by the Via
Department of Civil and Environmental Engineering at Virginia
has been reported by a wide variety of authors, with one of the Tech, the Alfred P. Sloan Foundation Microbiology of the Built
seminal papers being published by Lovley and Phillips (1988). Environment Program, and the U.S. National Science Foundation
Aging has been shown to produce chemical and structural National Center for Earth and Environmental Nanotechnology at
changes in the forms of MnOx(s) present (Cui et al., 2010; Hinkle Virginia Tech (NSF NNCI - 1542100).
et al., 2016). Likewise, variations in MnOx(s) crystalline structure
affect Mn reactivity (Stone, 1987). Burdige et al. (1992) showed Appendix A. Supplementary data
that S. oneidensisMR-1 could reduce amorphous (more recently
precipitated) forms of MnOx(s) (d-MnO2(s)and birnessite) much Supplementary data to this article can be found online at
more readily than occurred when trying to reduce highly- https://doi.org/10.1016/j.wroa.2018.100009.
crystalline pyrolusite (MnO2(s)).
It is possible to speculate how MnOx(s) aging and potential References
reduced bioavailability to Mn reducing bacteria could be related to
observed Mn release occurring in WTPs with MnOx(s)-coated me- Brandhuber, P., Clark, S., Knocke, W., Tobiason, J., 2013. Guidance for the Treatment
dia. Typical anthracite or sand filtration media have a long service of Manganese. Water Research Foundation, Denver, CO.
Burdige, D.J., Dhakar, S.P., Nealson, K.H., 1992. Effects of manganese oxide miner-
life (often well over ten years) due to media durability (Edzwald, alogy on microbial and chemical manganese reduction. Geomicrobiol. J. 10 (1),
2011). In WTPs employing free chlorine, addition to the filter 27e48.
 L.E. Swain et al. / Water Research X 1 (2018) 100009 7

Burnes, B.S., 2000. Identification and Characterization of Genes Required for Mn(IV) organic carbon oxidation coupled to dissimilatory reduction of iron or man-
Reduction by Shewanella Putrefaciens. Doctoral Dissertation. Georgia Institute of ganese. Appl. Environ. Microbiol. 1472e1480.
Technology, Atlanta, GA. Lovley, D.H., Dawn, Nevin, K., 2004. Dissimilatory Fe(III) and Mn(IV) Reduction, vol.
Cerrato, J.M., Falkinham III, J.O., Dietrich, A.M., Knocke, W.R., McKinney, C.W., 49. Elsevier Science & Technology, England, pp. 219e286.
Pruden, A., 2010. Manganese-oxidizing and -reducing microorganisms isolated Nealson, K.H., Myers, C.R., Wimpee, B.B., 1991. Isolation and identification of
from biofilms in chlorinated drinking water systems. Water Res. 44 (13), manganese-reducing bacteria and estimates of microbial Mn(IV)-reducing po-
3935e3945. tential in the Black Sea. Deep-Sea Res. 38 (2), S907eS920.
Chiao, Tzu-Hsin, Clancy, Tara M., Pinto, Ameet, Xi, Chuanwu, Raskin, Lutgarde, 2014. Osterman, A.L.R., Jennifer, L., Beliaev, A.S., Zhulin, I.B., Auchtung, J.M., Driscoll, M.E.,
Differential resistance of drinking water bacterial populations to monochlor- Tiedje, J.M., Pinchuk, G., Romine, M.F., Serres, M.H., Rodionov, D.A.,
amine disinfection. Environ. Sci. Technol. 47 (7), 4038e4047. Saffarini, D.A., Spormann, A.M., Gardner, T.S., Rodrigues, J.L.M., Nealson, K.H.,
Cui, H., Liu, F., Feng, W., Tan, W., Wang, M.K., 2010. Aging promotes todorokite Fredrickson, J.K., 2008. Towards environmental systems biology of Shewanella.
formation from layered manganese oxide at near-surface conditions. J. Soils Nat. Rev. Microbiol. 6 (8), 592e603.
Sediments 10, 1540e1547. Sain, A., Griffin, A., Dietrich, A.M., 2014. Assessing taste and visual perception of
Edgar, R.C., 2004. MUSCLE: multiple sequence alignment with high accuracy and Mn(II) and Mn(IV). J. Am. Water Works Assoc. 106 (1). E32-E32.
high throughput. Nucleic Acids Res. 32 (5), 1792e1797. Schicklberger, M., Buecking, C., Schuetz, B., Heide, H., Gescher, J., 2011. Involvement
Edzwald, J., 2011. Water Quality and Treatment: a Handbook of Drinking Water. of the Shewanella oneidensis decaheme cytochrome MtrA in the periplasmic
American Water Works Association, American Society of Civil Engineers, stability of the beta-barrel protein MtrB. Appl. Environ. Microbiol. 77,
McGraw-Hill. ISBN: 9780071630115. 1520e1523.
Gabelich, C.G., Fredrick, W., Knocke, W.R., Lee, C.C., 2006. Sequential manganese Sly, L.I., Hodgekinson, M.C., Arunpairojana, V., 1990. Deposition of manganese in a
desorption and sequestration in anthracite coal and silica sand filter media. drinking water distribution system. Appl. Environ. Microbiol. 56 (3), 628e639.
American Water Works Association 98 (5), 116e127. Stone, A., 1987. Microbial metabolites and the reductive dissolution of manganese
Gerhardt, P., Murray, R., Costilow, R., Nester, E., Woods, W., Krieg, N., Phillips, G., oxides: oxalate and pyruvate. Geochem. Cosmochim. Acta 51 (4), 919e925.
1981. Manual of Methods for General Microbiology. American Society for Swain, Lindsay E., 2016. The effects of manganese-reducing bacteria on desorption
Microbiology, Washington, D.C. ISBN 0-914826-30-1. of manganese from MnOx(s) coated media. Retreived from Virginia Tech ETD
Hinkle, M.A.G., Flynn, E.D., Catalano, J.G., 2016. Geochem. Cosmochim. Acta 192, database. http://hdl.handle.net/10919/71643.
220e234. Szeinbaum, N., Burns, J.L., DiChristina, T.J., 2014. Electron transport and protein
Islam, A.A., 2010. Manganese Removal by Media Filtration: Release and Complex- secretion pathways involved in Mn(III) reduction by Shewanella oneidensis.
ation. Doctoral Dissertation. University of Massachusets, Amherst, MA. Environmental Microbiology Reports 6 (5), 490e500.
Johnson, K.S., 2006. Manganese redox chemistry revisited. Science 313 (5795), Thamdrup, B., 2000. Bacterial manganese and iron reduction in aquatic sediments.
1896e1897. Adv. Microb. Ecol. 16, 41e84.
Jones, A., 2012. The Role of Aluminum within MnOx(s)-coated Filtration Media in Tobiason, J., Knocke, W.R., Goodwill, J., Islam, A., Hargette, P., Bouchard, R.,
Drinking Water Treatment (Retrieved from the Virginia Tech ETD Database). Zuravnsky, L., 2008. Characterization and Performance of Filter Media for
Knocke, W.R.O., Suzanne, C., Hungate, Robert, 1991. Removal of soluble manganese Manganese Control. AWWA Research Foundation, London, U.K., p. 212. ISBN:
by oxide-coated filter media: sorption rate and removal mechanism issues. 1843392194.
American Water Works Association 83 (8), 64e69. Yang, W., Zhen, Z., Zhang, Z., Chen, H., Liu, J., Ali, M., Liu, F., Li, L., 2013. Population
Lin, H.S., Nadia, H., DiChristina, T.J., Taillefert, M., 2012. Microbial Mn(IV) reduction structure of manganese-oxidizing bacteria in stratified soils and properties of
requires an initial one-electron reductive solubilization step. Geochimica et manganese oxide aggregates under manganese-complex medium enrichment.
Cosmochimica 99, 179e192. PLoS One 8 (9). https://doi.org/10.1371/journal.pone.0073778.
Lovley, D.R., Phillips, E.J.P., 1988. Novel mode for microbial energy metabolism: