University of Amsterdam: Institute of Biodiversity and Ecosystem Dynamics (IBED)
Department of Freshwater & Marine Ecology (FAME)
BSc Future Planet Studies (Biology major): bachelor’s thesis
16th of August, 2018
THE AGE OF BLUE-GREEN ALGAE
Killing of harmful cyanobacteria with hydrogen
peroxide under various nutrient limitations
Arian van Huis
Daily supervisor: Dr. ir. Giovanni Sandrini Supervisor: Dr. Petra Visser
This Bachelor’s thesis,
‘The Age of Blue-Green Algae: Killing of harmful cyanobacteria with hydrogen peroxide under various nutrient limitations’
by
Arian van Huis
is submitted in partial fulfillment of the requirements for the degree of:
Bachelor of Science in
Future Planet Studies
at the
University of Amsterdam
With special thanks to Giovanni Sandrini for his help and supportive feedback throughout the entire process. My gratitude goes to Tim Piel and Hongjie Qin for their guidance and making the lab- and fieldwork even better than it already was and to Petra Visser for her constructive comments on this
thesis.
Received:
………. Dr. ir. Giovanni Sandrini
Daily supervisor Date: ………. Dr. Petra Visser Supervisor Date: Submitted ………. Arian van Huis
TABLE OF CONTENTS
ABSTRACT p. 4 INTRODUCTION p. 5 MATERIALS & METHODS p. 8
RESULTS p. 11-20 Photosynthetic yield p. 11-12 Hydrogen peroxide p. 13-14 Biovolume Cell count KP mini-treatment p. 15-16 p. 17-18 p. 19-20 DISCUSSION p. 21-22 CONCLUSION p. 23 APPENDIX p. 24-29
Batch culture experiments over 24h Experimental control Hydrogen peroxide degradation data Hydrogen peroxide dilution protocol
p. 24 p. 25-27 p. 28 p. 29 REFERENCES p. 30-32
FIGURE 1 (cover): Massive toxic bloom of Microcystis aeruginosa in Lake Erie, US. Blooms in Lake Erie are mainly caused by nutrient leaching from surrounding farmland, but the increase of surface water temperatures and CO2 have amplified their intensity and length (NOAA, 2016; Kundewicz, 2008).
ABSTRACT
Anthropogenic climate change will cause rising water surface temperatures and elevated CO2 levels
worldwide, increasing the frequency and intensity of harmful cyanobacterial blooms in freshwater lakes, which are a great hazard to human health and account for billions of economic damage. Existing
methods to mitigate such blooms are either costly, impractical or exacerbate the issue. Hydrogen peroxide (H2O2) treatment of freshwater lakes has come up as a promising new method because of its
selectiveness for cyanobacteria and its quick degradation rate. However, little is known about the influence of nutrient limitations on the effectiveness of hydrogen peroxide on cyanobacteria. This research investigates the effectiveness of different concentrations of hydrogen peroxide on Microcystis
aeruginosa under (1) P-limitation, (2) High CO2 conditions and (3) Low CO2 conditions. The
photosynthetic yields, H2O2-degradation rate, biovolume and cell count of the cultures were measured.
P-limited cultures showed a mean -37% drop in photosynthetic yield after four hours, whilst High and Low CO2 cultures showed yield decreases of -66% and -77% respectively. High CO2 cultures degraded the
hydrogen peroxide at the fastest rate, while P-limited cultures did this much slower. At higher
concentrations, biovolumes and cell counts remained very constant over 24h for all nutrient limitations. P-limited cultures likely produced more protective enzymes in the chemostat compared to the other two groups, making their photosystems less sensitive to H2O2. Overall, these nutrient limitations will most
likely not influence the effectiveness of hydrogen peroxide treatment in any way that would affect treatment success.
INTRODUCTION
Cyanobacteria are among the most flexible organisms on the planet, being able to thrive in a wide variety of climates (from polar to
tropical) and conditions (eutrophic lakes to oligotrophic oceans) (Paerl & Huisman, 2009). However, some cyanobacterial species can form harmful cyanobacterial blooms in lakes (CyanoHAB’s), which produce toxins dangerous to humans and other organisms alike (Dawson, 1998; Paerl, Hall & Calandrino, 2011). CyanoHAB’s are responsible for over $2 billion loss in agricultural, recreational and drinking water resources in the US alone (Paerl, Hall & Calandrino, 2011).
Under anthropogenic climate change, these blooms are bound to increase in frequency, intensity and duration (Huisman et al., 2018). An increase in dissolved CO2 and bicarbonate due to rising CO2-levels
has already been shown to intensify and increase the duration of harmful cyanobacterial blooms (Sandrini et al., 2016). The same effects in relation to the rise of surface water temperatures and amount of solar irradiance have been well established: cyanobacteria become excellent competitors with other phototrophs at higher water temperatures (O’Neil et al., 2012; Paerl & Huisman, 2009; Robarts & Zohary, 1987) (Figure 2). However, eutrophication from human sources plays a dominant role in the increased occurrence of blooms. In fact, the excessive influx of key nutrients such as P and N has been identified as the single most important reason for the (increased) occurrence of CyanoHAB’s (Matthijs et al., 2012; Paerl, Hall & Calandrino, 2011; O’Neil et al., 2012). Thus, blooms originate due to the spatiotemporal convergence of three important factors: (1) an (over-)abundance of key nutrients such as P and N, (2) sufficiently warm surface water temperatures and (3) sufficient
sunlight over extended periods of time.
The crucial role of nutrient enrichment is the reason most policy from the 1970s onward has been focused on decreasing the input of P to combat blooms, such as measures drastically combating the use of fertilizer in the EU for instance, as well as the banning of
detergents and cleaning products containing P. This is because some CyanoHAB-species have the capacity to fix N2 when reactive N is depleted, thus it was reasoned
that limiting P would be more effective (Paerl, Hall & Calandrino, 2011). However, as has been observed in multiple cases, limiting P does reduce the dominance of N2-fixing cyanobacterial blooms, but non-N2 fixing
genera simply take their place, showcasing a typical ‘waterbed effect’ (Huisman et al, 2018).
Despite the type of cyanobacterium, another downside to reducing nutrients is that it is not a short-term
FIGURE 2: Growth rate as a function of temperature of three classes of eukaryotic phytoplankton and some cyanobacterial species that are frequent causes of CyanoHAB’s. Data has been obtained under conditions of light and nutrient saturation. A. form = Asterionella formosa, T. rot = Thalassiosira rotula, P. tric = Phaeodactylum tricornutum, H. triq = Heterocapsa triquetra, P. min = Prorocentrum minimum, C. furc = Ceratium furcoides, G. rad = Golenkinja radiata, Chlor. Sp. = Chlorella sp., S. cing = Staurastrum cingulum, A. ucr = Anabaena ucrainica, M. aer = Microcystis aeruginosa, A. flos = Aphanizomenon flos-aguae, C. sperm =
Cylindrospermopsis raciborskii . Figure and descriptive information obtained from Paerl, Hall & Calandrino (2011).
solution. This is because of a “legacy effect”, causing nutrients to be released from sediment long after their influx has seized (Matthijs et al., 2012). Thirdly, eutrophic lake systems have undergone a state change (from clear to turbid), which cannot simply be brought back to the initial desired state because of hysteresis (Weenink et al., 2015; Beisner, Haydon & Cuddington, 2003). Since lakes provide essential water for irrigation and drinking, it is of great importance to eradicate the blooms as quickly as possible (Matthijs et al., 2012; Paerl & Huisman, 2009). While nutrient input is the main driving force behind the unprecedented modern proliferation of CyanoHAB’s, reducing the input alone is not a short-term solution. For water bodies to remain usable in the short-term, a proper method to eliminate cyanobacteria must be devised.
In essence, there are three ways to combat cyanobacterial blooms: by prevention (nutrient reduction), control (artificial mixing, biomanipulation) and mitigation (algicides, surface scum removal) (Matthijs et
al., 2012). Unfortunately, many methods have severe drawbacks in that they are either costly,
impractical or exacerbate the issue. However, in recent years, a promising new method has appeared: hydrogen peroxide (H2O2) treatment.
Hydrogen peroxide is a ubiquitous compound used in a wide variety of applications, from cleaning to oxidizing and as an antimicrobial substance. Because it is produced as a by-product of photosynthesis, metabolic processes and the oxidation of dissolved organic carbon by UV-radiation, it happens to be one of the most abundant compounds in the world, found everywhere on the planet (Morris et al., 2011). In freshwater lakes, the background rate of hydrogen peroxide usually lies somewhere between one and thirty µg L-1 (Matthijs et al., 2012). The breakdown of hydrogen peroxide results in the production of
hydroxyl radicals (HO•). These deleterious compounds cause photoinhibition, inactivating PSII and severely damaging the cell (Barrington & Ghadouani, 2008).
The two main advantages of treatment with hydrogen peroxide are its selectiveness for cyanobacteria and its quick degradation. Already in 1986, Barroin & Feulliade showed that the cyanobacteria P.
rubescens experienced deleterious effects from a concentration of as little as 1.75 mg L-1 (Barroin &
Feulliade, 1986). The crux here is that cyanobacteria seem much more sensitive to hydrogen peroxide than other phytoplankton. In experiments conducted by Drábková, Admiraal, & Maršálek (2007), it was shown that Microcystis aeruginosa was about ten times more sensitive to hydrogen peroxide compared to the green algae P. subcapitata and the diatom N. seminulum. This selective elimination has also been demonstrated by later laboratory research, excluding toxicity for eukaryotic algae and most zooplankton below 5 mg L-1 (Weenink et al., 2015). Even the most sensitive fish species showed no observed effect at
concentrations below 50 mg L-1 (Gaikowski, Rach & Ramsay, 1999).
Further emphasizing its potential as a new mitigation strategy: hydrogen peroxide treatment has already been used successfully to treat an entire lake and relieve it from a Planktothrix rubescens bloom
(Matthijs et al., 2012). Using a concentration of 2.5 mg L-1 of H
2O2, the population of P. rubescens
collapsed by 90% over the course of three days, with cell counts declining from an initial concentration of 600*103 counts per mL, to near-zero. The decline in cell counts co-occurred with the decline of the
microcystin concentration. The hydrogen peroxide wasfully degraded after 2 days. The number of phytoplankton remained constant throughout the three months after the treatment. A different cyanobacterial species did invade again after a period of 7 weeks, underscoring that hydrogen peroxide treatment can only be observed as a mitigation strategy and not a permanent solution.
perform its task optimally: H2O2 degradation rates can vary due to the intensity of chemical and
biological processes, fueled by the presence of light and essential nutrients such as P and N (Matthijs et
al., 2012). Rapid degradation decreases the effectiveness of HP-treatment, meaning this increases the
uncertainty range on the H2O2-concentration required per individual case. Another issue is that, while
algae and diatoms might remain unharmed by these proposed concentrations, zooplankton appeared to be sensitive in the whole-lake experiment (Matthijs et al., 2012). To harm zooplankton whilst combatting cyanobacteria would be counterproductive, since zooplankton are natural predators of cyanobacteria and are valuable in maintaining a stable food web and thus, adequate fish populations. It is therefore of the utmost importance to possess the correct knowledge on exactly what dose of HP to apply in a lake treatment under different conditions of light and nutrient limitation.
The goal of this research is to determine in what way different nutrient conditions affect the
effectiveness of HP-treatment on CyanoHAB’s: High CO2, Low CO2 and Phosphorus-limitation (P-limited)
Microcystis aeruginosa is used as a model species in this study, it being one of the most notorious
common freshwater cyanobacteria (Barrington & Ghadouani, 2008). This species is capable of forming dense aggregates in freshwater lakes, having caused large proliferations of blooms in Lake Erie (near Chicago, USA) on a near-yearly basis, putting swimmers and local biota at risk (Rinta-Kanto et al., 2009; Dawson, 1998; Dittman & Wiegand, 2006). Understanding how different light and nutrient levels affect
Microcystis’ response to hydrogen peroxide will be a critical step in properly implementing and refining
MATERIALS & METHODS
The experimental set-up consisted of experiments with batch cultures of samples from chemostats in steady-state, performed in the IBED labs. The lab work was centered around the question in what way nutrient conditions influence the effectiveness of H2O2 on
M. aeruginosa, by investigating the responses to different
concentrations of hydrogen peroxide. In order to answer this question, four variables were measured in these batch culture experiments:
1. Photosynthetic yield (Φ), also known as chlorophyll fluorescence or quantum yield, of M. aeruginosa, is a measure of the photosynthetic vitality of a primary producer. It is calculated according to the following equation:
Φ =Fm − F0
Fm (Genty et al. , 1989) Where F0 is the minimum fluorescence and Fm the
maximum fluorescence. The yield of 0 mg L-1 is used as a
control value, which is the photosynthetic yield Yield is measured with a MINI-PAM II device (Walz, Effeltrich, Germany)
2. The degradation of hydrogen peroxide over time. This is done with the help of a nitrophenylboronic acid assay, performed by using a photospectrometer at the end of each batch experiment (Lu et al., 2012).
3. Biovolume (mm3 L-1), measured with CASY cell counting technology.
4. Cell concentration (cells mL-1), measured with CASY as well.
The cyanobacterium Microcystis aeruginosa PCC 7806 was cultivated in chemostats (see Figure 4) and samples were used in these batch culture experiments . This research made use of three distinct types of chemostat conditions: High CO2 (1450 ppm, standard BG11 medium without NaCO3); Low CO2 (150 ppm,
standard BG11 medium without NaCO3) and P-limited cultures (400 ppm, BG11 medium with 13,3x
lower P levels). An overview can be found in Table 1 below. BG11 medium was made according to standard protocol (Rippka et al., 1979) and chemostat light intensity was 50 µmol photons m-2 s-1 for all
three nutrient conditions.
Nutrient limitation Change in medium protocol (Rippka et al., 1979)
P-limited 13,3x lower P-concentration
High CO2 No NaCO3
Low CO2 No NaCO3
FIGURE 4: One of the four chemostats containing a P-limited culture of Microcystis aeruginosa PCC 7806.
Table 1: Changes in protocol during the preparation of BG-11 medium for three different nutrient limitations. Medium protocol was taken from Rippka et al. (1979).
Chemostats were monitored for approximately two months and sampled two times a week for
monitoring. As soon as they were in steady-state, samples from 2-4 replicate chemostats were taken for use in batch-culture experiments and diluted with medium in a large 3L flask to reach a biovolume of 100 mm3 L-1. The medium for P-limited cultures contained no P, while the medium for High CO
2 contained 10
mM of NaHCO3. The contents of this flask were distributed among twelve smaller flasks holding 200 mL
culture each. Different concentrations of hydrogen peroxide were added to the bottles later on. The twelve flasks were exposed to a light intensity of 15 µmol photons m-2 s-1 using white fluorescent tubes
(Philips Master TL-D 90 De Luxe 18 W/965; Philips Lighting, Eindhoven, The Netherlands) in combination with grey light filters. Usually, excess culture was poured and was removed with a pipet until all flasks contained exactly 200 mL of diluted culture. Cooling fingers were used to maintain a constant water temperature of 25 °C and air tubes bubbling compressed air were inserted to make sure hypoxia would not occur. Low CO2 cultures were bubbled with compressed air containing no CO2 and High CO2 cultures
were bubbled with normal compressed air.
t = 0 started when all the flasks were in place and is defined as the timepoint right before H2O2 was
added (t = 0.2)
Nutrient limitation Remarks
P-limited Added medium used for dilution contained no P. Low CO2 Cultures bubbled with air that contained no CO2.
High CO2 Added medium contained added 10 mM NaHCO
3,
bubbled with normal air.
Hydrogen peroxide was made with a 33%-stock (VWR International, Amsterdam, Netherlands) and dilutions were done according to protocol (refer to Appendix). After the measurements of t = 0 were finished, 1 mL of each stock was pipetted in its corresponding flasks so as to obtain the following desired concentrations (see FIGURE X): 0 mg L-1 (flask 1 & 2); 1 mg L-1 (3 & 4); 2 mg L-1 (5 & 6); 4 mg L-1 (7 & 8); 6
mg L-1 (9 & 10) and finally, 10 mg L-1 (flask 11 & 12). In addition, two flasks containing the same
BG11-medium used in the chemostats and two flasks containing pure Milli-Q were used as control groups for the monitoring of HP-degradation. 0 mg L-1 were the control groups for photosynthetic yield, biovolume
and cell count.
The batch culture experiment lasted for 24h, with several measuring points. Photosynthetic yield was measured at t = 0, t = 0.2, t = 0.5, t = 1 and every subsequent half hour until t = 4. After that, t = 24 was the only timepoint. Hydrogen peroxide concentration was measured from t = 0.2 onwards, with the same frequency as the yield (for the nitrophenylboronic acid assay protocol, refer to Lu et al., 2012). Biovolume and cell counts were measured twice for each flask at t = 0, t = 1.5, t = 3 and t = 24. Microscopy samples were taken at the same timepoints.
FIGURE 5: concentration in mg L-1 of hydrogen peroxide added to each of the batch culture flasks. Alongside these twelve flasks, two flasks with standard BG-11 medium and two with Milli-Q were used as controls for hydrogen peroxide degradation.
KP mini-treatment
An extra hydrogen peroxide experiment was done using a sample of lake water from the Kleine Klinkenbergerplas, KP for short (52°12'06.0"N 4°29'17.4"E, shown as the smaller of the two lakes in Figure 5 (right). KP is monitored for the occurrence of algal blooms and could be subject to treatment should such a bloom occur. On the 17th of May,
2018, a 10 L sample was taken from the center of the lake. Upon return to the laboratory, the sample was placed next on the windowsill and an air tube was inserted. Three days later, the 10 L sample was divided in 12 flasks (~800 mL each) which would be divided into four different types of categories: H2O2 concentration of 0 mg L-1 (flask 1-3); 2.5 mg
L-1 (4-6); 5 mg L-1 (7-9) and 2.5 mg L-1, containing 400 mL of
sample lake water and 400 mL of filtered sample lake water. Microscopy samples were taken at t = 0, t = 1.5, t = 3 and t = 24. The flasks were monitored for photosynthetic yield and hydrogen peroxide degradation, according to the same method as described above.
FIGURE 6: Aerial overview of the Grote (below) and Kleine Klinkenberger Plas (above), located in the municipality Oegstgeest in the province of South-Holland, The Netherlands (InTeylingen, 2018)
RESULTS
Graphs are shown from page 11 and onwards. Results over 24h for batch experiments is not shown in “Results”. For graphs, refer to 24h, refer to the Appendix.
Photosynthetic activity
In the first four hours, the photosynthetic yield of the High CO2 cultures showed an average decrease of
-37%. A concentratrion of 1 mg L-1 hydrogen peroxidedid not seem to have a negative effect on the
photosynthetic yield of M. aeruginosa (Figure 7A), in fact, the yield increased by +6,1% at t = 4, and ended with a +2% increase at t = 24. A concentration of 2 mg L-1 hydrogen peroxide(Figure 7B) seemed
much less effective than higher concentrations with a mere -39% drop at t = 4 compared to -100% for 4, 6 and 10 mg L-1 (Figure 7C, 7D and 7E respectively). At t = 24, 2 mg L-1 decreased further to -88%
compared to the control, while 4, 6 and 10 mg L-1 managed to increase by +5%, +2 and +8 percent,
respectively.
In Low CO2 cultures, a concentration of 1 mg L-1 hydrogen peroxide was also less effective than higher
concentrations, resulting in only a -16% decrease in yield at t = 4 and -42% at t = 24. Higher concentrations resulted in a decline in the yield: 2 mg L-1 hydrogen peroxide decreased by -89%
compared to the control at t = 4; 4, 6 and 10 mg L-1 hydrogen peroxide decreased with 95%, 92% and
-96% respectively. At t = 4, the yield in the cultures with 2 and 4 mg L-1 hydrogen peroxide decreased to 0
after 24h, while the yield in the cultures with 6 and 10 mg L-1 of hydrogen peroxide maintained 6% and
1% of the control yield after 24h, respectively.
At t = 4, the decrease in yield amongst P-limited cultures was lower for nearly every concentration compared to High and Low CO2 (except for 1 mg L-1, which produced a yield decrease of ~16% in both
P-limited and Low CO2). While High and Low CO2 showed dramatic yield collapse from 4 and 2 mg L-1
onward (respectively), higher concentrations produced nearly identical results. In contrast, P-limited cultured display a much more linear decrease under increasing HP-concentrations. At t = 4, one mg L-1
dropped by -16.8% compared to the control; 2 mg L-1 by -22%; 4 mg L-1 by -31%; 6 mg L-1 by -50% and 10
mg L-1 by -66,7%. At t = 24, however, the 1 mg L-1 cultures increased to +12% above the control; 2 mg L-1
recovered to +0.7% above the control; 6 mg L-1 recovered to a mere -6% below the control while only 10
mg L-1 decreased to a yield of 0.
All in all, phosphate limited M. aeruginosa cultures showed the smallest decrease in the photosynthetic yield: it dropped by a mean of -37% at t = 4, while High CO2 showed a mean decrease of -66%. Over the
same time period, Low CO2 showed a mean yield drop of -77%. At t = 4, the addition of 2 mg L-1 hydrogen
peroxide, closest to the concentration for field application, P-limited cultures dropped by -22,6 %; High CO2 by -39,7% and Low CO2 by -89%.
After 24h (refer to Appendix A), yield for P-limited cultures with a mean decrease of -29,4%, showing a recovery when compared to t = 4. High and Low CO2 cultures performed worse, with a median decrease
of -92,6% and -99,2%, respectively.For High CO2 and P-limited cultures, statistical analysis showed a
significantly negative correlation coefficient between yield decrease and increasing HP-concentration over four hours, while Low CO2 did not (p = 0.016 for High CO2; p = 0.005 for P-limited and p = 0.09 for
0 20 40 60 80 100 120 0 1 2 3 4 A ve rag e p h o to sy n th e tic yi e ld (Fv /Fm ) as % o f c o n tr o l Time (h)
10 mg L
-1Photosynthetic activity (4h)
0 20 40 60 80 100 120 0 1 2 3 4 A ve rag e p h o to sy n th e tic yi e ld (Fv /Fm ) as % o f c o n tr o l Time (h)4 mg L
-1 0 20 40 60 80 100 120 0 1 2 3 4 A ve rag e p h o to sy n th e tic yi e ld (Fv /Fm ) as % o f c o n tr o l Time (h)1 mg L
-1 0 1 2 3 4 Time (h)2 mg L
-1 Control High CO2 Low CO2 P-limited 0 1 2 3 4 Time (h)6 mg L
-1FIGURE 7A THROUGH 7E: photosynthetic vitality (Fv/Fm) of M. aeruginosa exposed to different concentrations of hydrogen peroxide: (A) 1 mg L-1, (B) 2 mg L-1, (C) 4 mg L-1, (D) 6 mg L-1, and (E) 10 mg L-1 H
2O2. The first four hours of the batch culture experiment are shown. “High CO2” indicates cells exposed to 1450 ppm CO2, “Low CO2”: indicates cells exposed to 150 ppm CO2, and “P-limitation” indicates cells exposed to P-limitation. Values were averaged for all duplicates (n= 12) and are displayed as a percentage of the control (0 mg L-1 of hydrogen peroxide), indicated by the dotted black line. The complete graphs with the 24h timepoint are shown in the Appendix.
7A)
B)
C)
D)
Hydrogen peroxide degradation
The average degradation of hydrogen peroxide is fastest in High CO2 cultures over the entire 24h, nearly
twice as fast as that of Low CO2 (see Table 1 below and refer to Appendix). While this relative difference
in degradation between High and Low CO2 remains the same over the course of the experiment, the
relative degradation rate of P-limited cultures does vary when comparing it to High CO2. Degradation of
hydrogen peroxide in P-limited cultures is 33% slower over 4 hours but 66% slower between 4 and 24h, when comparing them with High CO2 cultures (Table 1). Thus, with the exception of 1 and 2 mg L-1, the
concentration of hydrogen peroxide remained the highest overall in P-limited cultures at t = 24 with a remaining concentration of 0.9 mg L-1 (actual starting concentration [S]: 4.09 mg L-1); 3.3 ([S]: 6.25 mg L-1)
and 8.78 mg L-1 ([S]: 10.4 mg L-1).
For more data regarding the degradation of hydrogen peroxide, refer to Appendix.
Nutrient limitation Average slope of HP-degradation t = 1 – 4 t = 4 – 24 High CO2 -0.61 -0.13 Low CO2 -0.29 -0.07 P-limitation -0.41 +0.04
Table 1: average slope for the degradation of hydrogen peroxide in the three different nutrient limitations. Slopes are averaged between all concentrations and are shown for the first four hours (t = 1 – 4) and between four and 24 hours (t = 4 – 24).
0 1 2 3 4 5 6 7 8 9 10 11 12 0.2 0.5 1 1.5 2 2.5 3 3.5 4 24 [H 2 O2 ] (m g L -1) Time (h)
10 mg L
-1 0 1 2 3 4 5 6 7 8 0.2 0.5 1 1.5 2 2.5 3 3.5 4 24 Time (h)6 mg L
-1 0 1 2 3 4 5 0.2 0.5 1 1.5 2 2.5 3 3.5 4 24 [H2 O2 ] (m g L -1) Time (h)4 mg L
-1 0 0.2 0.4 0.6 0.8 1 1.2 0.2 0.5 1 1.5 2 2.5 3 3.5 4 24 [H2 O2 ] (m g L -1) Time (h)1 mg L
-1Hydrogen peroxide degradation (24h)
FIGURE 8A THROUGH 8E: degradation of various starting concentration of hydrogen peroxide over a period of 24h: (A) 1 mg L-1, (B) 2 mg L-1, (C) 4 mg L-1, (D) 6 mg L-1, and (E) 10 mg L-1 H
2O2. “High CO2” indicates cells exposed to 1450 ppm CO2, “Low CO2”: indicates cells exposed to 150 ppm CO2, and “limitation” indicates cells exposed to
P-8A)
B)
C)
D)
E)
0.0 0.5 1.0 1.5 2.0 2.5 0.2 0.5 1 1.5 2 2.5 3 3.5 4 24 Time (h)2 mg L
-1 High CO2 Low CO2 P-limited Starting concentrationBiovolume
Over a period of 24h, biovolumes showed the smallest changes for P-limited cultures, with σ = 6.88 mm3/L. For High CO
2, σ = 29.19 mm3/L and for Low CO2, σ = 12,6 mm3/L. Therefore, High CO2 seems to
have suffered the most when it comes to a drop in biovolume. A significant negative correlation was found between the decrease in biovolume as peroxide concentrations went up (p = 0.02 for High CO2, p =
0.009 for Low CO2 and p = 0.025 P-limited cultures (df = 4 and a = 0.05). Increasing the concentration
from 6 mg L-1 to 10 mg L-1 did not do much for the biovolume of any of the nutrient limitations, with only
a 2% difference between these two concentrations for P-limited cultures, 4% for High CO2 and 4% for Low CO2
At t = 24, High CO2 cultures of 1, 2 and 4 mg L-1 saw a change of +12%, +16% and -13% compared to the
control, respectively. At = 24, Low CO2 cultures with 1, 2 and 4 mg L-1 of peroxide saw a biovolume
difference of -12%, -37% and -48% compared to the control, respectively. At t = 24, P-limited cultures with 1, 2 mg L-1 saw a minor increase over the control with -0.4 and -0.3 percent respectively. Biovolume
0 1.5 3 24 Time (h)
2 mg L
-1 0 20 40 60 80 100 120 140 160 180 200 0 1.5 3 24 Bi o vo lu me ( mm 3/L ) Time (h)1 mg L
-1 0 20 40 60 80 100 120 140 160 180 200 0 1.5 3 24 B io vo lu m e ( m m 3/L) Time (h)10 mg L
-1 High CO2 Low CO2 P-limitation High CO2 control Low CO2 control P-limitation control 0 20 40 60 80 100 120 140 160 180 200 0 1.5 3 24 B io vo lu m e ( m m 3/L) Time (h)4 mg L
-1 0 1.5 3 24 Time (h)6 mg L
-1Biovolume (24h)
FIGURE 9A THROUGH 9E: development of biovolume of M. aeruginosa under different nutrient limitations over 24h, measured for different concentrations of hydrogen peroxide: (A) 1 mg L-1, (B) 2 mg L-1, (C) 4 mg L-1, (D) 6 mg L-1, and (E) 10 mg L-1 H
2O2. “High CO2” at 1450 ppm, “Low Co2” at 150 ppm. “P-limitation” indicates Microcystis cultivated
9A)
B)
C)
D)
Cell count
Cell counts remained very constant over 24h, except in High CO2 cultures which increased above the
control at 1 and 2 mg L-1, with an increase of +12% and +16% compared to the control; 4 mg L-1 saw a
decrease of -13%. Under Low CO2 conditions, 1 mg L-1 cell counts decreased by -3% compared to the
control; 2 mg L-1 by -4% and a concentration of 4 mg L-1 managed to decrease biovolume by -14%.
P-limited cultures cell counts changed with -3% compared to the control for 1 mg L-1; -8% for 2 mg L-1 and
no observed change for 4 mg L-1. Cell count in P-limited was thus the most constant with σ= 0.09*106
cells/mL for all concentrations over 24h. For High CO2, σ = 0.38*106 cells/mL (all concentrations, 24h) and
for Low CO2, σ = 2.9*106 cells/mL (all concentrations, 24h). Overall, a significant negative correlation
coefficient was found between the decrease in cell count and the increase in the concentration of hydrogen peroxide for Low CO2 cultures, but not for High CO2 or P-limited cultures. In P-limited cultures,
the correlation was even slightly positive with r = +0.04 (p = 0.065 for High CO2: p = 0.014 for Low CO2
Cell count (24h)
FIGURE 10A THROUGH 10E: cell count of M. aeruginosa over 24h under different nutrient limitations and different concentrations of hydrogen peroxide: (A) 1 mg L-1, (B) 2 mg L-1, (C) 4 mg L-1, (D) 6 mg L-1, and (E) 10 mg L-1 H
2O2. “High
10A)
B)
C)
D)
E)
0 2 4 6 8 10 12 14 16 18 20 0 1.5 3.5 24 Ce ll co u n t (*1 0 6) Time (h)1 mg L
-1 0 1.5 3.5 24 Time (h)2 mg L
-1 0 2 4 6 8 10 12 14 16 18 20 0 1.5 24 Ce ll co u n t (*1 0 6) Time (h)4 mg L
-1 0 1.5 24 Time (h)6 mg L
-1 0 2 4 6 8 10 12 14 16 18 20 0 1.5 24 Ce ll co u n t (*1 0 6) Time (h)10 mg L
-1 High CO2 Low CO2 P-limitation High CO2 control Low CO2 control P-limitation controlKP mini-treatment
The results of the KP mini-treatment showed the second biggest yield decrease: concentrations showed a mean drop of about 90% after four hours: 92% for 2.5 mg L-1; 90,9% for 5 mg L-1 and 90.66% for 2.5 mg
L-1 (50/50). Unfortunately, data on biovolume, species composition or cell count was unavailable at the
time of writing. However, during sampling at the lake, there was a substantial proliferation of (unidentified) cyanobacterial colonies present in the water, yet nowhere near the levels of a bloom. Therefore, it is safe to assume biovolume and cell count numbers were much lower than in the “normal” batch culture experiment.
0 1 2 3 4 5 6 7 8 0 5 10 15 20 [H 2 O2 ] (m g L -1) Time (h)
Degradation of hydrogen peroxide
0 mg/L 2.5 mg/L 5 mg/L 2.5 mg/L (50/50
KP mini-treatment (24h)
FIGURE 11A (above): photosynthetic vitality of a lake water sample taken from KP under different concentrations of hydrogen peroxide (indicated in the legend) during a full 24 hours. Values were averaged for all triplicates (n =12) FIGURE 11B (below): degradation of various hydrogen peroxide concentrations in lake water samples taken from KP, during a full 24 hours. Concentrations were determined with the nitrophenylboronic acid assay (Lu et al., 2012). Values were averaged for all triplicates (n =12)
A)
B)
0 20 40 60 80 100 120 0 1 2 3 4 A ve rag e c h lo ro p h yl l ac tiv ity (Fv /Fm ) as % o f co n tr o l Time (h)Photosynthetic yield as % of control (4h)
0 mg/L 2.5 mg/L 5 mg/L
DISCUSSION
Though the photo-inhibiting effect of hydrogen peroxide has been well established, there are some notable differences in the cyanobacterial response under different nutrient limitations and HP-concentrations.
P-limited cultures are most resistant to hydrogen peroxide
The fact that Low CO2 cultures are the most sensitive to hydrogen peroxide as shown by drops in
photosynthetic yield can be well observed in FIGURE 7B and FIGURE 7C for concentrations of 2 and 4 mg L-1 hydrogen peroxide, respectively. While this increased susceptibility could be explained by nutrient
starvation and oxidative stress, this answer is unsatisfactory, because P-limited cultures also suffer from a type of nutrient starvation (albeit a different type) and would thus also be more sensitive to
photoinhibition and environmental stressors (Ou, Wang & Cai, 2005). At the same time, the
photosynthetic yield of High CO2 cultures was lower on average than that of the P-limited cultures and
the latter could recover from exposure to 1, 2 and 4 mg L-1 hydrogen peroxide to yields above or at the
control level, while High CO2 cultures did not show any recovery. We are thus faced with a situation in
which the culture depleted of its essential nutrient is most resistant to oxidative stress from hydrogen peroxide, while supposedly vigorous, High CO2 Microcystis suffers greatly from the same compound.
Low CO2 cultures were the most sensitive to hydrogen peroxide probably because they have the least
resistance to further (oxidative) stress induced by hydrogen peroxide, since they are already suffering from CO2-starvation. Another reason might be a lower concentration of enzymes protecting against
oxidative stress. An increase in light exposure leads to increased production of enzymes involved in the protection against oxidative stress in M. aeruginosa (Hernando et al., 2016). Light intensity was the same in all conditions (15 µmol photons m-2 s-1), meaning there must have been a difference in biovolume.
Though chemostat data on Low CO2 was unavailable, it could be hypothesized that P-limited cultures had
the lowest biovolume out of the three limitations (394 mm3/L compared to 1000 mm3/L for High CO 2).
This is because M. aeruginosa is able to scavenge Ci very effectively at low concentrations thanks to their
Carbon Concentrating Mechanisms (Sandrini et al., 2015). Low CO2 cultures also have sufficient other key nutrients (P & N) at their disposal which promote growth. Hence, biovolume in Low CO2 cultures are
expected to have been slightly higher than those in P-limited cultures.
High CO2 cultures have the highest biovolume and would thus have suffered most from the effect of
self-shading (Garcia-Pichel, 1994), a common phenomenon in dense algal communities. Such self-shading
would have decreased the average light intensity per cell. Since biovolume of High CO2 was the highest,
this would have led to the lowest relative concentration of protective enzymes. In conclusion: this means that, relatively, P-limited cultures were exposed to the highest light intensity and consequently possibly produced more enzymes that offer protection against oxidative stress.
1 mg L-1 of hydrogen peroxide seemed to have a positive effect on the yield of High CO
2 compared to the
control. This could be because, at low concentrations, algae receive additional terminal electron acceptors from the breakdown of H2O2, enabling increase in primary production. This is also in line with
the observed increase in biovolume and cell count under High CO2 for lower HP-concentrations. At the
same time, we saw yield recovery between t = 4 and t = 24 for P-limited cultures at 1, 2 and 4 mg L-1
while biovolume did not increase and cell count remained very constant. This must mean that even affected cells in P-limited cultures somehow had the capacity to regenerate, even though hydrogen peroxide concentrations remained the highest after 24h out of all groups.
Nevertheless, it must be said that this deviation in yield or biovolume from the control groups cannot statistically be confirmed as neither significant nor insignificant, although it is clear that there exists a clear trend
High CO2 cultures degraded hydrogen peroxide at the fastest rate
High CO2 cultures were able to degrade hydrogen peroxide the quickest out of the three nutrient
limitations. When it comes to explaining the differences in rates of hydrogen peroxide degradation, the overall ‘health’ of M. aeruginosa might provide an answer. High CO2 cultures have all necessary nutrients
in abundance and are thus the fittest individuals out of all three conditions. This enables for quicker degradation of oxidative stressors such as hydrogen peroxide (possibly through peroxidase produced during the experiment), even using low HP-concentrations to their benefit in growth. In P-limited cultures, peroxidase could not have been involved because the results show that HP-concentrations degraded the slowest for P-limited cultures, whereas a higher concentration of catalase would mean a faster HP-degradation.
Biovolume and cell count remain constant under increasing concentrations
Whether degrading quickly or slowly, hydrogen peroxide concentrations above ~4 mg L-1 did not do
much to further decrease biovolume and cell count. In fact, cell count and biovolume only decreased significantly in High CO2 compared to the control, while for High CO2 and P-limited cultures, they
remained very constant. This might imply that the cells did not lyse, as was observed in the whole-lake treatment by Matthijs et al. (2012), although ammonium concentrations were not measured to conclusively prove this. Instead, it could be said that growth of the cells was simply inhibited and the cells were rendered innocuous, even recovering under P-limitation after 24h. The fact that biovolume and cell counts also stayed very constant when increasing the concentration from 6 to 10 mg L-1 suggests
that further increasing hydrogen peroxide concentration, within ecologically acceptable limits, might not necessarily be more effective.
KP mini-treatment further confirms effectiveness of hydrogen peroxide
Finally, the KP batch culture experiment demonstrated a severe decline in photosynthetic activity after 24, although not as severely as Low and High CO2 cultures. The presence of green algae in the lake water,
which contribute to the total yield but are resistant to HP, might explain this. Furthermore, the hydrogen peroxide was completely degraded for all groups after 24h, further supporting the notion by Weenink et
al. (2015) that hydrogen peroxide is likely not bound to linger long enough in a system to cause chronic
adverse effects.
Implications for lake treatment
Extrapolation of these results to the real environment would mean that under higher atmospheric CO2
conditions, blooms could be harder to treat. The difference would be scarcely noticeable though, seeing as High CO2 cultures were grown at 1450 ppm (>3.5 times the current atmospheric CO2 concentration)
showed a total collapse in photosynthetic yield after 24 hours, even at 2 mg L-1. Lakes low in P (with a
high N:P ratio) could prove to be an issue for effective treatment, seeing as cells under P-limitation were able to recover to levels above or near the control (as was the case up until 4 mg L-1). This would mean
desired treatment concentration of around 2.5 mg L-1 would only stimulate the primary production of
non N2-fixing genera, which would occur in such phosphorus-starved lakes. However, it’s unlikely such
blooms would contain cells with a high concentration of protective enzymes since blooms occur only when conditions are unusually favorable. The lack of any major decrease in biovolume and cell counts does warrant caution, since it could potentially mean that the cells do not lyse and are even able to recover. The same pattern for the drop in photosynthetic yield is observed in High CO2 and Low CO2: an
increase in hydrogen peroxide concentration causes yield to drop at a faster rate in the first four hours, but the same results are achieved after 24h regardless of concentration.
Combining these two findings means that, if conditions are favorable, the “golden standard” of 2.5 mg L-1
CONCLUSION
P-limited cultures demonstrate the highest overall resistance to hydrogen peroxide after 24h when it comes to photosynthetic yield, biovolume and cell count. In the first four hours, Low CO2 cultures were
most severely affected and while High CO2 performed better here on average, yet both groups dropped
to near-zero after 24h. The higher yield in P-limited cultures is most likely because of the increased production of protective enzymes due to low chemostat biovolume.
The degradation of hydrogen peroxide occurred at the fastest rate in high CO2 cultures because these
were the most vital, enabling the production H2O2-degrading enzymes during the experiment.
Degradation of hydrogen peroxide was considerably slower in Low CO2 cultures compared to High CO2
cultures, most likely because Low CO2 cells were too weak to deal with the oxidative stress, while
P-limited cultures seemed to tolerate hydrogen peroxide, barely breaking it down at higher concentrations while showing no change in biovolume or cell counts. This suggests that not peroxidase, but some other, non-H2O2-degrading enzyme played a vital role in protecting the algae against oxidative stress.
Lastly, under increasing concentrations of hydrogen peroxide, biovolume and cell count saw the biggest decrease for High CO2, while the two remained very constant for Low CO2 and P-limited cultures. An
increase from 6 to 10 mg L-1 barely affected biovolume or cell count for all nutrient limitations,
suggesting that hydrogen peroxide does not lyse nutrient-starved cells effectively and that an increase in hydrogen peroxide concentration will not help to lyse more cells.
Suggestions for further research
In order to further refine our understanding of the influence of hydrogen peroxide on cyanobacteria, it is worthwhile to investigate the effects of the limitation of important nutrients, such as manganese
(Gerloff & Skoog, 1957). In order to further refine the treatment of blue-green algae with hydrogen peroxide, we need a thorough understanding of the enzymatic pathway these organisms protect themselves from this toxic compound, seeing as there exist different enzymatic responses (or a seeming lack thereof) between cyanobacterial genera (Latifi, Ruiz & Zhang, 2009). Specifically, more research needs to be done on P-limited cultures and hydrogen peroxide to find out why these seem more resistant to hydrogen peroxide: is it because of protective enzymes or because of another, unknown factor? Additionally, since CyanoHAB’s can be formed by many cyanobacterial species, it would be wise to investigate if there is a difference between species in their response to hydrogen peroxide. In this way, it will be possible to further refine treatment, by adapting the required treatment concentration not only to nutrient conditions but also to the type of cyanobacteria. Lastly, an increase in hydrogen peroxide concentration did not lower biovolume and cell counts in any major way over the course of this experiment, it would be wise to investigate if cells truly lyse under the effect of hydrogen peroxide or are simply made innocuous.
A general remark of the results is that correlation coefficients were used to compare yield, biovolume, cell count to an increase in HP-concentration for statistically significant correlations between these variables. While correlations can provide valuable insight into the relationship between variables, the author must concede that the existence of a statistically significant correlation is only a part of the story. Unfortunately, statistical tests were impossible in this research because every concentration of hydrogen peroxide in the batch experiment were merely in duplo, which meant that the n for each timepoint was too small to perform meaningful statistical analysis. Additionally, the fact that some correlation
coefficients proved insignificant in some cases is also most likely due to the low n-value and not because of the experimental results themselves. Therefore, it would be a good course of action to replicate this research with more duplicates in order to draw statistically sound conclusions regarding the significance of declines in yield, cell count and biovolume.
APPENDIX A: BATCH CULTURE EXPERIMENTS RESULTS OVER 24H
APPENDIX
0 20 40 60 80 100 120 -1 4 9 14 19 24 Ph o to sy n th e tic yi e ld (Fv /Fm ) as % o f c o n tr o l Time (h)1 mg L
-1 0 5 10 15 20 Time (h)2 mg L
-1 Control High CO2 Low CO2 P-limited 0 20 40 60 80 100 120 0 5 10 15 20 Time (h)4 mg L
-1 0 5 10 15 20 Time (h)6 mg L
-1 0 20 40 60 80 100 120 0 4 8 12 16 20 24 Time (h)10 mg L
-1FIGURE A1 THROUGH A5: photosynthetic vitality (Fv/Fm) of M. aeruginosa exposed to different concentrations of hydrogen peroxide: (A) 1 mg L-1, (B) 2 mg L-1, (C) 4 mg L-1, (D) 6 mg L-1, and (E) 10 mg L-1 H
2O2. The whole 24h of the experiment are shown here. “High CO2” indicates cells exposed to 1450 ppm CO2, “Low CO2”: indicates cells exposed to 150 ppm CO2, and “P-limitation” indicates cells exposed to P-limitation. Values were averaged for all duplicates (n= 12) and are displayed as a percentage of the control (0 mg L-1 of hydrogen peroxide), indicated by the dotted black line. The complete graphs with the 24h timepoint are shown in the Appendix.
1)
2)
3)
4)
APPENDIX B: EXPERIMENTAL CONTROLS
B.1) Photosynthetic yield control batch experiments
B.2A) Hydrogen peroxide control Low CO2
0 0.1 0.2 0.3 0.4 0.5 0.6 0 5 10 15 20 A ve rag e p h o to sy n th e tic yi e ld (Fv /Fm ) Time (h)
Batch experiment PAM controls
High CO2 Low CO2 P-limited 0 2 4 6 8 10 12 0.2 0.5 1 1.5 2 2.5 3 3.5 4 24 [H 2 O2 ] (m g L -1) Time (h)
Low CO
2controls
BG11 a & b MQ a & bB.2B) Hydrogen peroxide control P-limited
B.2C) Hydrogen peroxide control High CO2
0 2 4 6 8 10 12 0.2 0.5 1 1.5 2 2.5 3 3.5 4 24 [H 2 O2 ] (m g L -1) Time (h)
P-limited controls
BG11 a & b MQ a & b 0 2 4 6 8 10 12 14 0.2 0.5 1 1.5 2 2.5 3 3.5 4 24 [H 2 O2 ] (m g L -1) Time (h)High CO
2controls
BG11 a & b MQ a & bFIGURE B1: average photosynthetic yield (Fv/Fm) of the control groups (0 mg L-1) of the three different nutrient limitations. Average duplicate values are shown
FIGURE B2A, B2B & B2C: concentration of hydrogen peroxide in BG11 medium and pure Mili-Q in Low CO2, P-limited and High CO2 cultures respectively. Concentrations were determined with the nitrophenylboronic acid assay (Lu et al., 2012). Average duplicate values are shown.
B.3A) KP HP
B.3B) KP photosynthetic yield control
FIGURE B3B: Photosynthetic yield (Fv/Fm) of KP lakewater control groups (0 mg L-1). Values were averaged for the triplicate.
FIGURE B3A: hydrogen peroxide concentration in BG11 medium and pure Mili-Q. Concentrations were determined with the nitrophenylboronic acid assay (Lu et al., 2012). Values were averaged for each duplicate. 0 1 2 3 4 5 6 7 8 0 5 10 15 20 [H2 O2 ] (m g L -1) Time (h)
KP HP controls data
BG11 a & b MQ a & b 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0 5 10 15 20KP yield control
KP ControlAPPENDIX C: HYDROGEN PEROXIDE DEGRADATION DATA
High CO2 Starting concentration (mg L-1) Actual starting concentration [S] (mg L-1) Slope Remaining H2O2 (mg L-1) t = Slope t = 4 t = 24 1 0.98 1 -1.15 <0.1 <0.1 2 2.12 2.5 -0.84 <0.1 <0.1 4 4.5 4 -0.53 2.45 <0.1 6 6.94 4 -0.67 4.66 <0.1 10 11.47 4 -1.07 7.16 0.51 Low CO2 Starting concentration (mg L-1) Actual starting concentration [S] (mg L-1) Slope Remaining H2O2 (mg L-1) t = Slope t = 4 t = 24 1 0.89 1 -0.71 <0.1 0.12 2 1.65 2.5 -0.32 0.7 0.18 4 3.35 4 -0.24 2.29 0.85 6 5.28 4 -0.43 3.99 1.27 10 8.84 4 -0.42 7.47 4.18 P-limited Starting concentration (mg L-1) Actual starting concentration [S] (mg L-1) Slope Remaining H2O2 (mg L-1) t = Slope t = 4 t = 24 1 1.04 1 -0.59 0.13 <0.1 2 2 2.5 -0.52 0.16 <0.1 4 4.09 4 -0.44 2.36 0.9 6 6.25 4 -0.50 4.18 3.3 10 10.37 4 -0.47 9.79 8.78KP batch culture experiment
Starting concentration (mg L-1) Actual starting concentration [S] (mg L-1) Slope Remaining H2O2 (mg L-1) t = Slope t = 4 t = 24 2.5 3.37 1 -0.31 2.24 <0.1 5 6.78 2.5 -0.53 4.79 <0.1 2.5 (50/50) 3.47 4 -0.20 2.67 <0.1
APPENDIX D: HYDROGEN PEROXIDE DILUTION PROTOCOL
H2O2 concentration(mg/l) solution Mili-Q
10000 33% H2O2 stock 1ml 32ml 5000 10000mg/l 5ml 5ml 4000 10000mg/l 4ml 6ml 2000 10000mg/l 2ml 8ml 1000 10000mg/l 1ml 9ml 500 5000mg/l 1ml 9ml 200 2000mg/l 1ml 9ml 100 1000mg/l 1ml 9ml 50 500mg/l 1ml 9ml 20 200mg/l 1ml 9ml 10 100mg/l 1ml 9ml 5 50mg/l 1ml 9ml 2 20mg/l 1ml 9ml 1 10mg/l 1ml 9ml 0.75 5mg/l 1.5ml 8.5ml 0.5 5mg/l 1ml 9ml 0.2 2mg/l 1ml 9ml 0 - 10 ml