Development of Microbial Mercury

Detoxification Processes Using

Mercury-Hyperresistant Strain of
Pseudomonas aeruginosa PU21
Jo-Shu Chang, Wen-Shing Law

Department of Chemical Engineering, Feng Chia University, Taichung,

Taiwan, Republic of China; telephone: 886-4-4517250, ext. 3651; fax:
886-4-4510890; e-mail: changjs@fcu.edu.tw
Received 7 February 1997; accepted 24 July 1997

Abstract: A mercury-hyperresistant strain of Pseudomonas aeruginosa PU21 harboring plasmid Rip64 was utilized to develop bioprocesses able to detoxify and recover soluble mercuric ions in aquatic systems. The kinetics of mercury detoxification was investigated to
determine the parameters needed for the design of the
bioprocesses. Batch, fed-batch, and continuous bioreactors were utilized to evaluate the efficiency and feasibility
of each mode of operation. The results showed that the
specific mercury detoxification rate was dependent on
cell growth phases, as well as the initial mercury concentrations. Cells at the lag growth phase exhibited the best
specific detoxification rate of approximately 1.1 106 g
Hg/cell/h, and the rate was optimal at an initial mercury
concentration of 8 mg/L. In batch operations with initial
mercuric ions ranging from 2 to 10 mg/L, the mercuric
ions added were rapidly volatilized from the media in
less than 23 h. With periodic feeding of 3 or 5 mg Hg/L
at fixed time intervals, the fed-batch processes had mercury removal efficiencies of 2.9 and 3.3 mg Hg/h/L, respectively. For continuous operations, the effluent cell
concentration (Xe) was essentially invariant at 527 and
523 mg/L with the dilution rates (D) of 0.18 and 0.325 h1,
respectively. The increase in mercury feeding concentrations (Hgf) from 1.0 to 6.15 mg Hg2+/L did not affect the
steady-state cell concentration (Xe) but forced the effluent mercury concentration (Hge) to increase. The decrease in the dilution rate, however, resulted in lower
Hge values. It was also found that sequential mercury
vapor absorption columns recovered over 80% of the Hg
released from the bioreactor while the residual mercury
vapor was subsequently immobilized by an activated
carbon trap in the down stream of the absorption column. 1998 John Wiley & Sons, Inc. Biotechnol Bioeng 57:
462470, 1998.

Keywords: mercury detoxification: Pseudomonas aeruginosa; bioreactor design

INTRODUCTION
Mercury is known as one of the most toxic heavy metals on
earth. Mercury and mercurial compounds have been used in

Correspondence to: J.-S. Chang

Contract grant sponsor: National Sciences Council of the ROC
Contract grant number: NSC-85-2214-E-035-004

1998 John Wiley & Sons, Inc.

a variety of industries, causing severe mercury pollution in

aquatic systems and soils (DItri, 1972; Noyes et al., 1976;
Nriagu, 1979). Due to their high mobility, the polluted mercurial compounds disperse widely into the environment via
physical, chemical, and biological pathways during which
time they are potentially concentrated hundreds of times
through the food chain (DItri, 1972; Nriagu, 1979; Saouter
et al., 1994). Conventional mercury removal processes
mainly utilize physical and chemical approaches that involve either trapping and collecting of mercury from contaminated sites or chemical precipitation of mercury compounds so that they can be easily separated (Jones, 1971;
Noyes et al., 1976). The drawback of physical methods has
been the requirement for additional treatments, while
chemical methods often leave hazardous by-products or residual sludges. Therefore, it is necessary to search for alternative approaches, such as biological methods, for a more
natural and efficient cleanup of mercury waste at a relatively low cost.
It is well known that mercury-resistant microorganisms
can resist mercury due to their ability to volatilize soluble
forms of mercury from the environment via a sequence of
enzymatic reactions, which are recognized as mercury detoxification (Silver et al., 1989; Summers and Silver, 1972,
1978; Wood and Wang, 1983). The genetic basis of mercury
resistance was found to be encoded in mer operons located
on either plasmids or transposable elements (Foster, 1987;
OHalloran, 1993; Summers, 1986). With the aid of organomercurial lyase originated from the merB gene, mercuryresistant microorganisms are able to cleave the CHg bonds
of organomercurial compounds; the resulting mercuric ions
are enzymatically reduced to less toxic and more volatile
metallic mercury (Hg) by the merA product mercuric reductase (Belliveau and Trevors, 1989; Foster, 1987; Misra,
1992; Robinson and Tuovinen, 1984; Summers, 1986). The
merP and merT genes in mer operons also express cysteinerich proteins located on the periplasmid space and inner
membrane, respectively, for the specific delivery of ambient
mercuric ions toward mercuric reductase located in the cytoplasm where Hg2+ is reduced to volatile Hg (Lund and

CCC 0006-3592/98/040462-09

Brown, 1987; Misra, 1992). The constitutive mer operon is

induced by the subtoxic level of mercuric ions (Clark et al.,
1977; OHalloran, 1993). The molecular genetics of the mer
operon and the biochemistry of mercury detoxification have
been intensively studied for years (Belliveau and Trevors,
1989; Foster, 1987; Misra, 1992; Robinson and Tuovinen,
1984; Summers, 1986). However, far less effort has been
devoted to the understanding of mercury detoxification kinetics or the development of microbial mercury detoxification strategies (Chang, 1993; Chang and Hong, 1995; Philippidis et al., 1990, 1991). Recently Saouter et al. (1994)
reported their preliminary investigation of using Hg2+ reducing strains to decontaminate a polluted freshwater pond
in Tennessee. Other authors (Ghosh et al., 1996a,b) demonstrated their studies on volatilization of mercury using
resting or immobilized cell systems. Direct utilization of
mercuric reductase to remediate mercury was also attempted with immobilization of the enzyme by activated
supports (Anspach et al., 1994; Uo et al., 1992). Our previous work (Chang, 1993; Chang and Hong, 1995) showed
that mercury-hyperresistant Pseudomonas aeruginosa PU21
strain, which contains plasmid Rip64 that encodes for the
mer operon, was able to remove mercuric ions effectively
from the contaminated water. This observation motivated us
to look for the possibility of utilizing the microorganism to
develop mercury detoxifying processes.
This research began with the investigation of the effect of
mercury concentrations on the transient growth of the microorganism and the dependence of detoxification kinetics
on the bacterial growth phases and on the substrate (mercury) concentrations. After the required kinetic parameters

were obtained, batch, fed-batch, and continuous bioreactors

were utilized to demonstrate operation strategies applicable
to the practical mercury treatment processes. A tentative
mercury vapor recovery device was also designed to prevent
the resulting product of mercury detoxification, Hg, from
being released into the atmosphere. Experimental results
obtained from this study were evaluated to justify the feasibility of the bioprocess for mercury remediation.
MATERIALS AND METHODS
Microorganism and Cultivation
P. aeruginosa PU21, an auxotrophic derivative of PAO1,
was isolated from clinical sewage by Jacoby (1986). The
strain harbors a 142.5 kb Rip64 plasmid, which encodes for
the narrow spectrum mercury resistance in addition to resistance to various antibiotics (Jacoby, 1986). Mercury hyperresistance of the PU21 strain was generated by serial
adaptation of the cells in liquid media containing elevated
mercury concentrations up to 150 mg/L. Detailed procedures for the selection of the mercury-hyperresistant PU21
strain were described elsewhere (Chang, 1993). The hyperresistant strain was cultivated aerobically at 37C with either LuriaBertani (LB) broth (Difco) or Pseudomonas
minimal medium (PMM) (Miller and Ku, 1978) supplemented with 0.41 mg/mL of leucine, isoleucine, and valine.
The stock bacterial culture was prepared on LB medium
containing 25 mg Hg2+/L or PMM amended with 2 mg
Hg2+/L to keep the constitutive mer operon induced.

Figure 1. Schematic description of a mercury detoxification bioreactor and the mercury recovery device.

CHANG AND LAW: MICROBIAL Hg DETOXIFICATION PROCESSES

463

Mercury Measurement

Bioreactor Operations

Total mercury and mercuric ion concentrations in the

growth medium were measured with a mercury analyzer
(Buck Scientific, model 400A) applying cold vapor atomic
absorption spectrometry. In general, samples were treated
with 2 mL HNO4 (30.3%), 2 mL H2SO4 (48.3%), and 5 mL
KMnO4 (5%) at 95C for 1 h to oxidize all the mercury
present to the ionic form (Hg2+) before measurement. The
excess KMnO4 was neutralized with hydroxylamine hydrochloride (NH4OH z HCl). Mercuric ions were then reduced
to metallic mercury (Hg) by adding 1 mL of 20% stannous
chloride (SnCl2). For the measurement of mercuric ions, the
oxidation steps were omitted (Chang et al., 1993). The mercuric compound used in this research was HgCl2 obtained
from Sigma (99.6% pure).

The batch, fed-batch, and chemostat operations were conducted with a 5-L fermentor (Eyela Jar Fermentor MDF)
equipped with devices that measure and control temperature, dissolved oxygen level (DO), pH, and agitation speed.
Figure 1 is the schematic description of the bioreactor and
mercury recovery device. The cultivation temperature and
agitation rate were controlled at 37C and 450 rpm, respectively. Because our previous work (Chang et al., 1993)
showed that the components of LB medium interact with
mercuric ions to form complexes that could not be reduced
by the P. aeruginosa PU21 strain, the growth medium used
in the bioreactor experiments was PMM, which would not
form nonreducible complexes with mercuric ions if the concentration of Hg2+ was lower than 20 mg/L (Chang, 1993).

Figure 2. Mercury profiles during transient growth of Pseudomonas aeruginosa PU21 as (a) 2 mg Hg2+/L was added at the cultivation time of 0, 4.7,
7, 8.6, and 12.6 h; (b) 5 mg Hg2+/L was added at the cultivation time of 0, 5.2, and 8.2 h; (c) 8 mg Hg2+/L was added at the cultivation time of 0, 11, and
21 h; and (d) 10 mg Hg2+/L was added at the cultivation time of 0, 13, and 43 h. (m) Cell growth and (s) residual mercury concentration.

464

BIOTECHNOLOGY AND BIOENGINEERING, VOL. 57, NO. 4, FEBRUARY 20, 1998

by batch detoxification experiments. The concentration of

mercury in each HNO3 column was analyzed after the residual mercury inside the bioreactor reached essentially
zero. The ratio between the amount of mercury trapped by
the acid columns and the amount of mercury added into the
bioreactor was defined as the efficiency of mercury recovery.
RESULTS AND DISCUSSION
Mercury Profiles in Course of Cell Cultivation

Figure 3. Dependence of specific mercury detoxification rates on mercury concentrations at different cell growth phases: (d) lag phase, (j)
exponential phase, and (m) stationary phase.

The air flow rate was 1 vvm (1 vol air/vol medium/min) for
continuous operations. In fed-batch operation, solutions
with mercuric ion concentrations of 3 and 5 mg/L were fed
into the reactor at fixed time intervals (1 or 1.5 h). In chemostat operations, the performance was examined for two
dilution rates (0.18 and 0.325 h1) combined with four mercury feeding concentrations (1.0, 2.0, 4.0, and 6.15 mg/L).
It should be noted that the maximum specific growth rate of
the PU21 strain with PMM was approximately 0.56 h1, so
the highest dilution rate attempted in this study was 0.325
h1 to ensure a stable chemostat operation. In all the experiments, the optical density (at 550 nm) and mercury concentration were monitored as a function of time.
Recovery of Mercury Vapor
The mercury vapor collection device was the combination
of gas and liquid absorption columns using HNO3 as the
oxidant and the activated carbon adsorption trap (Fig. 1).
Connected to the gas exit of the bioreactor were three glass
columns (36 cm long, 6.5 cm in diameter) in series. Each
column was packed with small glass tubes (2.5 cm long,
outer diameter 0.8 cm, inner diameter 0.65 cm), resulting in
a porosity of 80%. The columns were filled with 30% concentrated nitric acids to dissolve mercury vapor (Hg) released from the bioreactor. All the glassware used in the
mercury recovery operations was treated with mercury-free
concentrated HNO3 solution prior to each experiment to
avoid the possible adsorption of mercury on the glass surface. The gas effluent of the third acid absorption column
was connected to an activated carbon packed cylindrical
vessel with a bed height of 25 cm and a diameter of 2.6 cm
(Fig. 1). The mercury recovery efficiency was determined

P. aeruginosa PU21 (Rip64) was grown with PMM that

initially contained 2, 5, 8, and 10 mg Hg2+/L. In each run,
the same mercury concentrations were added into the culture at the exponential growth phase, as well as at the stationary phase. The mercury concentration in the culture was
recorded as a function of time after each mercury addition.
The results of these experiments are illustrated in Figure 2.
It was found that most of the mercury introduced into the
cell culture disappeared within 23 h, reflecting rapid mercury detoxification by the cells of P. aeruginosa PU21. The
time interval required to achieve a complete detoxification
tended to extend as the mercury concentration introduced to
the culture was increased (Fig. 2). The experimental data
also showed that cell death occurred at the beginning of
each mercury addition; and the extent of cell death appeared
to become more significant during the early stage (lag
phase) of the transient growth, which may be attributed to
lower cell concentration in the lag-phase culture. Addition
of 28 mg/L of Hg2+ into the exponential growth phase
culture resulted in a slight decrease in cell concentration; the
decrease was approximately 0.9, 3.5, and 1.5% for the mercury dosage (Hgd) of 2, 5, and 8 mg Hg/L, respectively (Fig.
2ac). In contrast, as Hgd was raised to 10 mg/L, a nearly
42% decrease in cell concentration was observed after the
second mercury addition (Fig. 2d), indicating that 10 mg/L
of mercury caused severe growth inhibition to the exponential growth phase cells whereas the toxic effect was much
less significant as Hgd was below 8 mg/L. Figure 2 also
demonstrates that addition of mercury at the stationary
phase did not lead to considerable cell death. The results
seem reasonable because the cell concentration at the stationary phase was very high, forcing an instant removal of
mercury from the culture to minimize the cell death.
Detoxification Kinetics
Specific mercury detoxification rates (RHg) were estimated
from experimental data shown in Figure 2ad. Being determined in the course of the transient growth, the RHg value is
actually an in situ measurement of the specific mercury
removal activity for cells at different growth stages. The
equation used to calculate RHg is
1 D@Hg2+#
,
?
RHg =
Xavg
Dt

CHANG AND LAW: MICROBIAL Hg DETOXIFICATION PROCESSES

465

Dependence of RHg on Bacterial Growth Phase

Figure 4. Cell growth and mercury profiles in the fed-batch operation as

3 mg Hg2+/L was added every 1 h: (m) cell growth and (s) residual
mercury concentration.

where D[Hg2+] is the difference in mercury concentration

during the time interval Dt and Xavg is the average cell
concentration over the time interval. Figure 3 demonstrates
the effect of mercury concentration on specific mercury
detoxification rates. As shown in Figure 3, for all growth
phases the specific mercury detoxification rate reached a
maximum when the initial mercury concentration in the
culture was nearly 8 mg/L, similar to the optimal mercury
concentration identified by Philippidis et al. (1990, 1991) in
their studies on mercury detoxification kinetics with a recombinant Escherichia coli strain. The highest RHg obtained
from our batch experiments was approximately 1.11 106
mg Hg/h/cell, which was measured during the lag phase
with the initial mercury concentration of 8 mg/L.

Figure 3 showed that cells at the lag phase exhibited the

highest specific mercury detoxification rates, followed by
those at the exponential phase; at the stationary phase the
cells had the lowest RHg. This trend may be correlated with
the physiological state of cells at different growth phases. At
the lag phase cell density was low, and thus their mercury to
cell ratio (rHg/x) was higher than that of other growth phases.
To survive the toxic environment, lag-phase cells may tend
to dedicate most of their resources to reducing the mercury
concentrations only and thereafter start growing. Therefore,
with RHg representing the average detoxification performance of each cell in the population, the lag-phase cells
possibly achieved better RHg than cells at any other phases.
As the exponential growth phase is reached, the rHg/x ratio
reduces significantly due to higher cell concentrations.
Thus, according to the reaction kinetics, the RHg would be
lower as reactant (Hg) concentration dropped. Besides,
rapid cell division occurring during the exponential phase
may cause uneven distribution of mer operon containing
plasmids in the succeeding generations, and thus some low
copy number cells may be produced. As a result, the average mercury detoxification activity of the whole cell population may be reduced due to the presence of those low copy
number cells. Therefore, the specific mercury detoxification
rate may not be optimized at the exponential growth phase.
It also seems reasonable to observe that the lowest RHg
occurred at the stationary phase, because the stationaryphase culture has the lowest rHg/x ratio. Besides, during the
stationary phase the cellular metabolic energy and NADPH
production may be limited, so the energy driven cellular
mercury transport process (Misra, 1992) may be retarded
and the supply of the cofactor (reducing power) for the
enzymatic reduction of mercuric ions may be insufficient.
Any of these cases could result in a decrease in RHg.

Table I. Comparison of Hg2+ conversion and mercury detoxification efficiency between fed-batch
and chemostat operations.

Feeding strategy

Mercury in feed (mg/L)

Hg2+
conversiona
(%)

Every 1.0 h
Every 1.5 h
D 4 0.18 h1

3.0
5.0
1.0
2.0
4.0
6.5
1.0
2.0
4.0
6.5

100.0
100.0
99.91
99.75
99.15
98.21
99.66
98.40
97.90
97.07

Operation conditions
Mode of
operation
Fed batch
Chemostat

D 4 0.325 h1

Mercury detoxification
efficiencyb (mg/L/h)
2.90
3.30
0.18
0.36
0.71
1.09
0.32
0.64
1.27
1.94

Hg2+ conversion 4 the amount of Hg2+ reduced/the amount of Hg2+ added into bioreactor.
Mercury detoxification efficiency 4 the amount of Hg2+ reduced at the end of the operation/(total
volume treated) z (time of operation).
a

466

BIOTECHNOLOGY AND BIOENGINEERING, VOL. 57, NO. 4, FEBRUARY 20, 1998

Fed-Batch Mercury Detoxification Operations

Figure 2a and 2b show that batch growth of P. aeruginosa
PU21 with PMM containing 2 and 5 mg Hg/L exhibited no
considerable lag phase and more than a 90% reduction of
mercury concentration within 12 h. Fed-batch operations
were therefore designed by pulse feeding of 3 and 5 mg
Hg/L every 1 and 1.5 h, respectively, until the cell growth
was inhibited. This mercury feeding strategy eliminates the
toxic effect of mercury on cell growth and thus should allow
cells to detoxify mercury more efficiently. Figure 4 demonstrates the transient behavior of cell growth and the residual mercury profile with pulse feeding of 3 mg Hg/L
every 1 h. It was observed that the mercury feeding did not
cause considerable cell death, and after a short 2-h lag

phase, cells started to grow exponentially. It also shows that

most of the mercury concentration in the fermentor was
volatilized at the end of each hour during the first 12 h of
cell cultivation, while the extent of detoxification declined
after operation for 12 h (Fig. 4). The overall detoxification
efficiency was 2.9 mg Hg/L/h (Table I).
The results of 5 mg Hg/L feeding (Dt 4 1.5 h) with
different inoculum sizes are presented in Figure 5ac. With
the inoculum size of 5.6 108 cells/mL (Fig. 5a), the cell
concentration remained almost constant throughout the 19-h
operation, suggesting that under the operation conditions
the cell death rate and cell growth rate were essentially
equal. It was found that the added mercuric ions were detoxified almost completely within each 1.5-h feeding inter-

Figure 5. Cell growth and mercury profiles in the fed-batch operation as 5 mg Hg2+/L was added every 1.5 h: (a) inoculum size 4 5.6 108 cells/mL,
(b) inoculum size 4 1.49 109 cells/mL, and (c) inoculum size 4 1.4 108 cells/mL. (m) Cell growth and (s) residual mercury concentration.

CHANG AND LAW: MICROBIAL Hg DETOXIFICATION PROCESSES

467

these results seemed to imply that inoculum size must exceed certain critical values to ensure a successful fed-batch
operation. The critical cell inoculum may be required to
provide sufficient initial mercury detoxification activity, allowing a rapid drop of soluble mercury concentration to a
nontoxic level. It is also possible that the critical amount of
inoculum reflects the number of cells necessary to contain a
significant segment of the population with mercury hyperresistance.
Continuous Mercury Detoxification Operations
Figure 6a and 6b demonstrate typical profiles of cell and
mercury concentrations in chemostat operations with the
dilution rates (D) of 0.18 and 0.325 h1. After the continuous culture reached the steady state, the cell (Xe) and mercury (Hge) concentrations in the effluent of the fermentor
remained invariant for over 50 h (80 generations). For continuous culture operated with D 4 0.18 and 0.325 h1, the
effluent cell mass concentrations (Xe) were approximately
527 and 523 mg/L, respectively, and the Xe appeared to be
independent of the feeding mercury concentrations (Hgf)
used in this study. The dependence of mercury concentrations in the effluent (Hge) on those in the feed (Hgf) is
presented in Figure 7. It shows that for both dilution rates,
Hge appeared to increase as Hgf increased from 1.0 to 6.15
mg/L, while Hge values were lower for the lower dilution
rate. For the runs with mercury feeding of 1 mg/L, the Hge
value was nearly undetectable. As the steady-state cell concentration (Xe) did not change with Hgf and the specific
detoxification rate (RHg) was assumed to be uniform in the
chemostat culture, increases in feed mercury should also

Figure 6. Cell concentration and mercury profiles in the chemostat operation: (a) D 4 0.18 h1, Hgf 4 6.15 mg/L; and (b) D 4 0.325 h1, Hgf
4 6.15 mg/L. () Hgf, (d) cell concentration, and (.) residual mercury
concentration.

val for the first 8 h of cultivation, after which the residual

mercury concentration rose slightly (Fig. 5a). With the operation procedures, the overall detoxification efficiency was
3.3 mg/L/h, which is about 15% higher than that with a 3
mg Hg/L feeding (Table I). As the inoculum size was increased to 1.49 109 cells/mL (Fig. 5b), cell death was not
observed; and cell concentration increased from the time of
inoculation, reaching saturation after an 18-h cultivation. In
contrast, as the inoculum size was decreased to lower than
1.4 108 cell/mL, severe cell death was observed and the
detoxification activity declined rapidly (Fig. 5c). Therefore,

468

Figure 7. Dependence of effluent mercury concentration (Hge) on feeding mercury concentration in chemostat operations: (d) D 4 0.18 h1 and
(j) D 4 0.325 h1.

BIOTECHNOLOGY AND BIOENGINEERING, VOL. 57, NO. 4, FEBRUARY 20, 1998

Table II. Mercury recovery efficiency of successive nitric acid columns in batch detoxification
operations with various initial mercury concentrations.
% Hg trapped in each columna
Initial mercury
concentration (mg/L)
2.0
5.0
8.0
10.0

First column

Second column

Third column

Total Hg recovery
efficiencyb (%)

83.2
76.6
67.8
60.7

13.1
18.8
23.3
26.6

3.70
4.60
8.90
13.8

80.6
82.2
80.0
88.3

a
% Hg trapped in each column 4 the amount of Hg trapped in the column of interest/total
amount of Hg trapped in three columns.
b
Total Hg recovery efficiency 4 total amount of Hg recovered in three columns/total amount of
Hg2+ added into the reactor.

lead to an increase of Hge. Therefore, the effect of Hgf on

Hge shown in Figure 7 seems reasonable. Also, because the
mean residence time for D 4 0.18 h1 is nearly double of
that for D 4 0.325 h1, one can expect to see lower Hge
values for D 4 0.18 h1 than for 0.325 h1. The conversions
of Hg2+ to Hg for the chemostat operations carried out in
this study were always over 97%. Table I summarizes the
mercury detoxification efficiencies of fed-batch and continuous operations. As the Hgf was increased from 1.0 to 6.5
mg/L, the mercury detoxification efficiency for chemostat
operations tended to increase, whereas the Hg2+ conversion
dropped slightly. It can also be found in Table I that the
continuous cultures exhibited lower mercury detoxification
efficiencies than those of fed-batch runs. The fed-batch detoxification processes were more cost effective than chemostat operations, which can, however, handle a larger quantity of mercury contaminated water than fed-batch processes. Further evaluations may be required to justify which
mode of operation would be more effective, because in this
study the fed-batch and continuous runs may not have been
operated with their optimal strategies.
Mercury Recovery
The end product of the mercury detoxification process was
metallic mercury (Hg), which was readily volatile under
the operation conditions and was subsequently released
through the gas exit of the bioreactor. The Hg in the gas
effluent was stabilized with concentrated HNO3 in a series
of gas absorption columns (Fig. 1). The mercury recovery
efficiency was tested in batch operations with initial mercury concentrations of 2, 5, 8, and 10 mg/L. The experimental results, listed in Table II, show that most of the
vaporized mercury was absorbed in the first column. As the
initial concentration in the batch culture was decreased from
10 to 2 mg Hg2+/L, the relative amount of Hg trapped in
the first column increased from 61 to 83%. In contrast to the
first column, the efficiency of mercury vapor absorption in
the second and third columns decreased sharply, and the
absorption percentage tended to increase as initial mercury
concentrations were raised. As indicated in Table II, the
total mercury recovery efficiency by the nitric acid columns

was in the range of 8090% for batch operations starting

with 210 mg/L of mercuric ions. The residual mercury
vapor released from the acid columns was passed through a
column containing activated carbon. It was also found that
the mercury concentration in the exiting stream of the activated carbon column was below the detection limit (approximately 0.2 mg/L) of the mercury analyzer used in this
study.
CONCLUSIONS
P. aeruginosa PU21 (Rip64) is able to detoxify mercury
rapidly during its transient growth. The reduction of mercury with growing cells exhibited a maximal rate at the
substrate concentration of 8 mg Hg/L. The specific mercury
detoxification rate was also correlated to the bacterial
growth phases, with cells at the lag phase and exponential
phase having higher specific detoxification rates. The highest rate obtained from our experiments so far was approximately 1.11 106 mg Hg/cell/h. The fed-batch operations
with stepwise feeding of mercury to the bioreactor appeared
to exhibit excellent mercury detoxification efficiencies of
2.9 and 3.3 mg/L/h for 2 and 5 mg Hg/L feedings, respectively. Showing slightly lower detoxification efficiencies
than those obtained from fed-batch mode, chemostat operations were able to continuously reduce the feeding mercury
concentrations from 1.0 to 6.15 mg/L to extremely small
amounts, and the steady-state operation could be maintained
for a prolonged period (nearly 55 h). However, optimal
operation conditions still need to be identified to develop
more effective detoxification strategies. Metallic mercury
resulting from the detoxification process was recovered
from the gaseous stream by acid columns and by an activated carbon trap. The recovery efficiency in the acid columns was over 80%. This research is a preliminary step
toward the ultimate goal of developing practical mercury
detoxification bioprocesses for the treatment of mercury
wastes.
The authors gratefully acknowledge financial support from the
National Sciences Council of the Republic of China under Grant
NSC-85-2214-E-035-004. The authors also thank Dr. Juan Hong
for much valuable advice.

CHANG AND LAW: MICROBIAL Hg DETOXIFICATION PROCESSES

469

References
Anspach, F. B., Huckel, M., Brunke, M., Schutte, H., Deckwer, W.-D.
1994. Immobilization of mercuric reductase from a Pseudomonas
putida strain on different activated carriers. Appl. Biochem. Biotechnol. 44: 135150.
Belliveau, B. H., Trevors, J. T. 1989. Mercury resistance and detoxification
in bacteria. Appl. Organomet. Chem. 3: 283294.
Brown, N. L., Lund, P. A., Bhriain, N. N. 1989. Mercury resistance in
bacteria, pp. 175195. In: D. A. Hopwood and K. F. Chater (eds.),
Genetics of bacterial diversity. Academic Press, New York.
Chang, J.-S. 1993. Kinetic modelling and optimization of bioreactor operation for mercury detoxification in Pseudomonas aeruginosa PU21
(Rip64). Ph.D. thesis, University of California, Irvine.
Chang, J.-S., Hong, J. 1995. Estimation of kinetics of mercury detoxification from low-inoculum batch cultures of Pseudomonas aeruginosa
PU21 (Rip64). J. Biotechnol. 42: 8590.
Chang, J.-S., Hong, J., Ogunseitan, O. A., Olson, B. H. 1993. Interaction of
mercuric ions with the bacterial growth medium and its effects on
enzymatic reduction of mercury. Biotechnol. Progr. 9: 526532.
Clark, D. L., Weiss, A. A., Silver, S. 1977. Mercury and organomercurial
resistances determined by plasmid in Pseudomonas. J. Bacteriol. 132:
186196.
DItri, F. M. (ed.) 1972. The environmental mercury problem. CRC Press,
Cleveland, OH.
Foster, T. J. 1987. The genetics and biochemistry of mercury resistance.
CRC Crit. Rev. Microbiol. 15: 117140.
Ghosh, S., Sadhukhan, P. C., Chaudhuri, J., Ghosh, D. K., Mandal, A.
1996a. Volatilization of mercury by immobilized mercury-resistant
bacterial cells. J. Appl. Bacteriol. 81: 104108.
Ghosh, S., Sadhukhan, P. C., Ghosh, D. K., Chaudhuri, J., Mandal, A.
1996b. Volatilization of mercury by resting mercury-resistant bacterial
cells. Bull. Environ. Contam. Toxicol. 56: 259264.
Jacoby, G. A. 1986. Resistance plasmids in Pseudomonas aeruginosa, pp.
497514. In: I. C. Gunsalus, J. R. Sokatch, and L. N. Ornston (eds.),
The bacteria, Vol. X. Academic Press, Orlando, FL.
Jones, H. R. (ed.) 1971. Mercury pollution control. Noyes Data Corp., Park
Ridge, NJ.
Lund, P. A., Brown, N. L. 1987. Role of the merT and merP gene products

470

of transposon Tn501 in the induction and expression of resistance to

mercuric ions. Gene 52: 207214.
Miller, R. V., Ku, C. C. 1978. Characterization of Pseudomonas aeruginosa mutants deficient in establishment of lysogeny. J. Bacteriol. 134:
875883.
Misra, T. K. 1992. Bacterial resistances to inorganic mercury salts and
organomercurials. Plasmid 27: 416.
Noyes, O. R., Hamdy, M. K., Muse, L. A. J. 1976. Control of mercury
pollution. Toxicol. Environ. Health 1: 409420.
Nriagu, J. O. (ed.) 1979. Biogeochemistry of mercury in the environment.
Elsevier/North-Holland Biomedical Press, New York.
OHalloran, T. V. 1993. Transition metals in control of gene expression.
Science 261: 715725.
Philippidis, G. P., Schottel, J. L., Hu, W. S. 1990. Kinetics of mercuric
reduction in intact and permeabilized Escherichia coli cells. Enzyme
Microb. Technol. 12: 854859.
Philippidis, G. P., Schottel, J. L., Hu, W.-S. 1991. A model for mercuric
ion reduction in recombinant E. coli. Biotechnol. Bioeng. 37: 4754.
Robinson, J. B., Tuovinen, O. H. 1984. Mechanisms of microbial resistance and detoxification of mercury and organomercury compounds:
Physiological, biochemical, and genetic analysis. Microbiol. Rev. 48:
95124.
Saouter, E., Turner, R., Barkay, T. 1994. Microbial reduction of ionic
mercury for the removal of mercury from contaminated environments.
Ann. NY Acad. Sci. 721: 423427.
Silver, S., Misra, T. K., Laddaga, R. A. 1989. Bacterial resistance to toxic
heavy metals, pp. 121139. In: T. J. Beveridge (eds.), Metal ions and
bacteria. Wiley, New York.
Summers, A. O. 1986. Organization, expression, and evolution of genes for
mercury resistance. Ann. Rev. Microbiol. 40: 607634.
Summers, A. O., Silver, S. J. 1972. Mercury resistance in a plasmidbearing strain of E. coli. J. Bacteriol. 112: 12281236.
Summers, A. O., Silver, S. 1978. Microbial transformations of metals.
Ann. Rev. Microbiol. 32: 637672.
Wood, J. M., Wang, H. K. 1983. Microbial resistance to heavy metals.
Environ. Sci. Technol. 17: 582A590A.
Uo, M., Numata, M., Suzuki, M., Tamiya, E., Karube, I., Makishima, A.
1992. Preparation and properties of immobilized mercuric reductase in
porous glass carriers. J. Ceram. Soc. Jpn. 100: 430433.

BIOTECHNOLOGY AND BIOENGINEERING, VOL. 57, NO. 4, FEBRUARY 20, 1998