Why it's better to be blue than greenInvestigating the environmental drivers of cyanobacterial harmful algal bloom growth and

toxicity in western Lake Erie

Timothy Davis

Co-authors & Funding• Greg Dick, Rose Cory, Michelle Berry, Vincent Denef, Melissa Duhaime,

Derek Smith and Kevin Meyer – University of Michigan

• Duane Gossiaux, Steve Ruberg – NOAA-GLERL

• Greg Doucette, Tina Mikulski, Richard Stumpf, Timothy Wynne – NOAA-NCCOS

• Tom Johengen, Alicia Ritzenthaler, Ashley Burtner & Danna Palladino –CILER

• George Bullerjahn and Robert Michael McKay, Mark Rozmarynowycz –Bowling Green State University

• Justin Chaffin – OSU Stone Lab

• Susan Watson– Environment Canada

• Funding: EPA-Great Lakes Restoration Initiative, Great Lakes Nutrient Initiative, NOAA, University of Michigan Water Center

"If you are interested in stories withhappy endings, you would be betteroff listening to some otherpresentation. In this talk, not only isthere no happy ending, there is nohappy beginning and very fewhappy things in the middle."

Adopted from: Lemony Snicket’s ‘A Series of Unfortunate Events: The Bad Beginning’

Disclaimer

Harke et al., 2016; Harmful Algae

HABs have expanded globally in the past few decades

Why have HABs proliferated?

• Anthropogenic nutrient loading• N vs P

• Loss of key food web predators

• Invasive species may alter ecosystems that give HAB species an advantage

• Global climate change producing a wider range for some HABs.

• More comprehensive monitoring and reporting

HABs and human healthNortheast, USA (Maine – NY) Southeast, USA (Florida - Texas)

Pseudo-nitzschia seriata -(Cleve) H. Peragallo (400 x)

Karenia brevis (formally known as Gymnodinium breve)

Domoic Acid

http://www.regione.emilia-romagna.it/crm/images/im_Pseudo_nitzschia_seriata.jpghttp://en.wikipedia.org/wiki/Image:Domoic-acid.png

HABs and human healthWest coast, USA (California - Washington) Inland, coastal and Great Lakes

Pseudo-nitzschia seriata -(Cleve) H. Peragallo (400 x)

Domoic Acid

Anabaena /Dolichospermum

Microcystin Anatoxin-a

Planktothrix

http://www.regione.emilia-romagna.it/crm/images/im_Pseudo_nitzschia_seriata.jpghttp://en.wikipedia.org/wiki/Image:Domoic-acid.png

HABs have inspired horror movies

Cyanobacterial blooms are a global phenomenon

Lake Taihu, ChinaCredit: T. Davis

Matilda Bay, Western Australia

Credit: D. Sarson

Lake Erie, USACredit: NOAA

Lake Victoria, E. AfricaCredit: H. Paerl

Han River, SeoulCredit: HanKyoReh newspaper

Lake Okeechobee, USACredit: USGS

Paerl and Otten 2013, Science

The ecology of cyanobacterial blooms is complex

Cyanotoxins impact all levels of the food web, including humans

• The World Health Organization (WHO) has set a 1µg per liter concentration for microcystin as a guideline level for safe drinking water.

• Cyanotoxins can cause serious neurological and gastrointestinal disorders and have been associated with gastrointestinal cancers.

• On occasion, these compounds have lead to death in humans (i.e. Brazilian hemodialysis patients).

• Cyanotoxins have been linked to deaths in waterfowl, cattle, dogs, and marine mammals.

Photo courtesy of The China Digital Times

Photo courtesy of Jef Huisman et al.Miller et al., 2010

Toledo water crisis of 2014

Lake Okeechobee State of Emergency 2016Photo: miamiherald.com

Photo: surfline.com

Utah Lake algal bloom 2016

‘Utah Poison Control says it has taken hundreds of calls about the bloom, about 130 of which involved people that likely came in contact with the water reporting vomiting, diarrhea, headache and rashes.’

Maumee River algal bloom 2017

Photo: ABC 13 News

Wuxi water crisis of 2007

Kisumu Bay, Lake Victoria, Kenya, 2004

Overarching research statement: Understanding the drivers of

bloom ecology will aid in enhancing predictive models that forecast bloom size, location AND

toxicity

HABs occur throughout the Great Lakes basin

NOAA studies HABs throughout the Great Lakes

Lake St. Clair bloom during late August, 2013

Image: MODIS

Spatial sampling of Lake St. Clair, 23 August, 2013

Samples collected (17 sites)• Sterivex filter (DNA)• Cell enumeration (Lugols)• Dissolved and total nutrients• Total microcystins (ELISA)

Genetic analysis• Targeted the mcyA region of the MC

gene operon• All samples yielded PCR products for

mcyA• Samples were sequenced at

GENEWIZ, NJ

0%

20%

40%

60%

80%

100%M

B1

142

140

Tham

es 1

Tham

es 2

139

138

136

RC

MR

IV

135 8

134

803 FI

804

1159

CC

GB

Dinophyta

Cryptophyta

Chrysophyta

Bacillariophyta

Euglenophyta

Chlorophyta

Cyanophyta

% p

hyto

plan

kton

bio

mas

s

Site

Cyanobacteria dominated the phytoplankton community at most sites

Davis et al., 2014; PLOS ONE

Microcystis spp. dominated the cyanobacterial community

Davis et al. 2014; PLOS ONE

MC producers were homologous throughout LSC and similar to strains in Lake Erie and Lake Ontario

Microcystis spp.Anabaena spp.Planktothrix spp.

Davis et al., 2014; PLOS ONE

• Maximum-likelihood tree was generated using the Jones-Taylor-Thornton (JTT) algorithm• Bootstrap values were obtained for 1,000 replicates

Bay of Quinte, Lake OntarioHamilton Harbour, Lake OntarioLake St. ClairNorthshore Lake Erie

Partnerships between Federal, State, academic and citizen scientists help to monitor western Lake Erie

Weekly sampling reveals important trends that highlight potential environmental drivers and inform future experiments

• Toxicity changes throughout the bloom

• Relationship between nitrogen and toxicity

• Microcystis blooms occur even when phosphorus concentrations are low

• Other factors beyond nutrients may be important in driving bloom structure and function

Toxi

city

H2O

2(m

g L-

1 )(μ

g L-1

)(μ

g L-1

)(n

M)

(μg

L-1)

ic

ys

tis

2014 Toledo Water Intake (Station 12)

Weekly sampling reveals important trends that highlight potential environmental drivers and inform future experiments

Trends are consistent between years

From: Gobler, Burkholder, Davis, et al., 2016, Harmful Algae

Blooms are patchy!

Sensors:. YSI EXO sondes

. Chlorophyll, Phycocyanin, Turbidity, CDOM

. Wetlabs Cycle -SRP

Western Lake Erie Real-time HABs Monitoring

= Real-time observation station in 2016

. SeaBird SUNA nitrate sensor

Continuous nutrient monitoring tracks field collections well Continuous sensor work being conducted by Tom Johengen, Steve Ruberg, Danna Palladino, Heidi Purcell, Steve Constant, Ron Muzzi and Russ Miler

Continuous pigment monitoring yielded more variability

• Overall trend is seen but much greater variation between grab and continuous data than the nutrient sensors

• ‘Packaging effects’ of colonies interferes with the ability to accurately detect fluorescence signal

• Buoy setup instruments record data about 2m below the surface whereas Niskinsamples are taken at 1m

• Highlights the need to be careful when integrating datasets

Continuous sensor work being conducted by Tom Johengen, Steve Ruberg, Danna Palladino, Heidi Purcell, Steve Constant, Ron Muzzi and Russ Miler

Viking Mission(s) to MarsFirst time a biological probe was used to search for life outside of this planet

Autonomous near real-time toxin detection for Lake Erie• Purchased with GLRI funds post-Toledo Water crisis• Collaboration between NOAA, MBARI, WHOI, OSU• Truly emerging technology

• Fewer than 20 worldwide• ESPniagara is the first in freshwater • Fine-scale MC observations are critical for toxicity models

Stumpf, Davis et al., 2016; Harmful Algae

Deployment of ESPniagara

• 5’6” tall• 5ft x 5ft footprint• < PSI pressure on

footprint• Deck weight ~ 1800 lbs• In-water weight ~ 750lbs

June July Sept/Oct

Test of communication system and general operations

Science testing at Ohio State University Stone Laboratory on South Bass/Gibraltar Islands

Full Mission Deployment• Microcystin sampling every other day• Archived samples for future DNA sequencing

Pressure housing for ESP − Rated to 48 m depth

Sampling manifold

3-way sample valve

Bolted lifting assembly

ESP battery assemblies(and opposite)

Tagline attachment points

ESP MC assay design

• Using mAb against ADDA moiety and MC-BSA conjugate – for comparison with Abraxis ELISA

• Compatible with extraction solvent

• Good range of detection on benchtop “mimic”

• Good overall intensity values

0.2 ng/mL MCLR 2.0 ng/mL MCLR 20.0 ng/mL MCLR 200.0 ng/mL MCLR

Work being conducted by Greg Doucette and Tina Mikulski

Detection and quantification limits during the September 2016 deployment

Sample volume filtered (ml)

Lower limit of detection*

(LLOD)

Lower limit of Quantification*

(LLOQ)

Upper limit of Quantification*

(ULOQ)1000 0.01 0.02 0.15500 0.03 0.04 0.31200 0.07 0.09 0.77150 0.09 0.12 1.03100 0.13 0.19 1.5475 0.18 0.25 2.0550 0.27 0.37 3.0830 0.44 0.62 5.1325 0.53 0.75 6.16

All limits in ug/L*using 15sec autoexposure image

Work being conducted by Greg Doucette and Tina Mikulski

Multiplexing toxins on ESP

Rabbit IgG (6-10)

STX-OVA (11-21)

Mouse IgG (1-5)

DA-OVA (22-32)

NeoSTX-OVA (33-43)

Spot currently used

Spot currently empty

New spots for 7x7 array

1

2

3

4

5

6

7

8

9 10

11

14

15

17

1920

21

12

13

18

16

22 23

25 24

26 27

285

29

30

31 32

33

39

42

35

34

36

3738

43

4041

Work being conducted by Greg Doucette and Tina Mikulski

2G ESP

3G ESP

MBio Waveguide‘3.1G’ ESP

Evolving and Emerging Ecogenomic Sensor Technologies

Spreeta SPR chips

Analysis of Microcystis subpopulations:

% toxic Microcystis = proportion of Microcystis cells containing the genetics to produce microcystins (mcyD or E / 16S x 100)

Determined by (1) qPCR and (2) metagenomics

Microcystis (16S rDNA)

Potentially-toxic Microcystis

(mcy-containing)

Non-toxic Microcystis

(16S without mcy)

NIES44(non-toxic)

NIES1050(toxic)

Other cyanobacteria

Chart1

70

30

Sheet1

100

60

70

30

Sheet1

Sheet2

Sheet3

Chart1

40

60

Sheet1

40

60

Sheet1

Sheet2

Sheet3

What is omics?

Dick and Lam, Elements Magazine, in review

relative abundance (%)

(Illumina MiSeq16S v4 region)

CyanobacterialOTUs (%)

0

25

50Bacterial community

Microcystinconcentration (ug/L)

Phycocyaninconcentration (ug/L)

Shift in toxicity not due to changes in cyanoHAB genera

(toxin)

(total cyanobacterialcommunity)

Community shifts correlate with (1) seasonal parameters (e.g., temp) (2) the bloom (pH)

pre-bloomearly bloommid-bloomlate-bloompost-bloom

post-bloom

late bloom

early bloom

pre-bloom

Berry, Davis et al., in press, Environmental Microbiology

Microcystis has ~1900 core genes and a diverse flexible genome

• Within an individual strain 40-50% are core genes

• Prochlorococcus 40-67% (Kettler et al., 2007 PLoSGenetics)

• Anabaena 40% (Simm et al., 2015 Frontiers in Microbiology)

• Of all genes identified• 11%-14% of identified genes

are highly represented (core genome)

• 62% of genes are rare (only found in 1-2 strains)

Meyer, Davis et al., subm.

COG distribution of core and flexible genes in Lake Erie isolates show an enrichment of defense-

oriented genes in the flexible genome

Meyer, Davis et al., subm.

Many of the unique genes found in new Microcystis isolates involve genome flexibility

Meyer, Davis et al., subm.

• Most unique genes are currently hypothetical• LE3: 42%• LSC13-02: 43%• LE013-01: 43%

Unique genes are present and expressed in a naturally occurring bloom

Lake Erie Metagenomes

Lake Erie Metatranscriptomes

relative abundance (%)

(Illumina MiSeq16S v4 region)

CyanobacterialOTUs (%)

0

25

50Bacterial community

Particulate Microcystins

(ug/L)

Phycocyaninconcentration

(ug/L)

Toxicity decreases as the bloom progresses

Toxic strains decline with lower nitrate concentrationsN

itrate conc. (mg L -1)

Part

icul

ate

MCs

(μg

MC-

LR e

quiv.

L-1

)

SRP

conc

. (μg

L-1

)

Nitrate = 0.14 ± 0.03 mg L-1Ammonium = 2.3 ± 2.7 µg L-1SRP = 10.5 ± 5.2 µg L-1

% Toxic Microcystis>100µm = 4 ± 3%

53-100µm = 29 ± 26%3-53µm = 5.7 ± 10.1%

Microcystins = 0.7 ± 0.6 µg L-1

(Average ± SD of 4 sampling dates)

Nitrate = 0.41 ± 0.47 mg L-1Ammonium = 2.8 ± 2.6 µg L-1SRP = 1.9 ± 1.4 µg L-1

% Toxic Microcystis>100µm = 63 ± 31%

53-100µm = 48 ± 31%3-53µm = 5.6 ± 5.2%

Microcystins = 4.0 ± 2.6 µg L-1

(Average ± SD of 5 sampling dates)

Rela

tive

expr

essio

n (m

cyD:

RPO

C1)

Bars = Mean ± SE; A and B = significantly up-regulated (p < 0.05)

A

B B

BB

A

Microcystin synthesis genes up-regulated within 4 hours of exposure to increased N

A

B

Chaffin, Davis, et al, in prep.

Nitrogen constrains growth and toxicity of Planktothrix in Sandusky Bay

Davis et al., 2015 ES&T

Control

+PO4 only

+NO

3 only

+NH

4 only

NH

4 + PO4

NO

3 + PO4

+Urea only

Urea + PO

4

After 48 hours of incubation

Proteobacteria and Cyanobacteria dominate but cyanos are doing the work

Eco-transcriptomic surveys of LE CHABs

Harke, Davis et al., 2016; ES&T

0.0

0.5

1.0

1.5

2.0

LET7 LET6 LET5 LET4 LET3 LET2 LET1

DIP

[µM

]

020406080

100120

LET7 LET6 LET5 LET4 LET3 LET2 LET1Com

mun

ity c

ompo

sitio

n(%

tran

scrip

ts)

Station

Microcystis Anabaena Planktothrix

Flavobacterium Pseudomonas Other prokaryotes

Niche differentiation among cyanobacterial populations

Harke, Davis et al., 2016; ES&T

0.0

0.5

1.0

1.5

2.0

LET7 LET6 LET5 LET4 LET3 LET2 LET1

DIP

[µM

]

012345678

LET6 LET5 LET4 LET3 LET2 LET1

Log 2

Fold

Cha

nge

pstA pstB pstB2 pstS pstS sphX

Microcystis is an excellent P scavenger

Harke, Davis et al., 2016; ES&T

Solu

ble

Rea

ctiv

e P

conc

entr

atio

ns

From: Gobler, Burkholder, Davis, et al., 2016, Harmful Algae

0 0.5 1.0

Microcystins may protect cyanobacteria from oxidative damage….

H2O2 (μM)

Pigment-inferred growth and particulate toxicity not impacted by ROS

August 19, 2014

Phyc

ocya

nin

(μg

L-1)

•Control•+Ni+P (20 & 2 μM)•+ROS (H2O2; 1μM)•+Ni+P + H2O2

Ni = NH4NO3P = KH2PO4x3

48 hours @

ambient light and temperature

2L bottles

x3

Part

icul

ate

Mic

rocy

stin

s(μ

g M

C-LR

equ

ival

ents

L-1

)

Will increases in H2O2 impact Microcystis growth or toxicity

+H2O2 treatments = 1081 ± 439 nMControl treatments = 364 ± 158 nM

(Average ± SD of 18 measurements)

MS = Methanol soluble MCs (Free) MI = Methanol insoluble MCs (protein-bound)

Catalase is the main sink of H2O2 in marine waters

2H2O2 → O2 + H2O

The catalase test (Jesse Fenno)

Moffett and Zafiriou, 1990 Limnology and Oceanography

Colony-associated catalase gene expression shows similar trends to microcystins concentrations

Microcystins (ug/L)Re

lativ

e Ex

pres

sion

0

2

4

6

8

10

0.00

0.10

0.20

0.30Catalase katG Cytochrome c Peroxidase Microcystin

n = 19 n = 6 n = 16 n = 13 n = 45 n = 36

n = number of genomes queried

Microcystis and many other cyanos lack catalase genesExternal

Internal

(H2O2) (H2O2 or ROOH)

(H2O2) (H2O2 or ROOH)

(H2O2 or ROOH)

(H2O2 or ROOH)

(H2O2 or ROOH)

(O2-)

Isolation of bloom-associated heterotrophic bacteria from Lake Erie

not tested

not tested

not tested

not tested

not tested

not tested

compare to LE 16S

Over half of all isolates produce catalases

Isolate

Code

Class

Family

Genus

Boot-

strap

Seq-

matchb

OTUc

OTU

%ID

Cata-lase

LE-L6

Gammaproteobacteria

Pseudomonadaceae

Pseudomonas

100

-

00127

97.6

-

LE-E3

Gammaproteobacteria

Pseudomonadaceae

Pseudomonas a

-

0.987

00122

99.2

+

LE-E6

Gammaproteobacteria

Pseudomonadaceae

Pseudomonas a

-

0.989

00127

98.0

-

LE-L1

Gammaproteobacteria

Enterobacteriaceae

Citrobacter a

-

0.994

00469

100

+

PE-6

Gammaproteobacteria

Xanthomonadaceae

Stenotrophomonas

100

-

00985

99.2

+

LE-E1

Bacilli

Bacillaceae

Bacillus

98

-

00683

96.1

+

LE-L2

Bacilli

Bacillaceae

Bacillus

100

-

00276

95.7

+

LE-L3

Bacilli

Bacillaceae

Bacillus

100

-

00683

94.5

+

LE-L4

Bacilli

Bacillaceae

Bacillus

98

-

00683

94.5

-

LE-L5

Bacilli

Bacillaceae

Bacillus

100

-

00683

96.1

-

PE-1

Bacilli

Bacillaceae

Bacillus

97

-

00276

94.9

-

PE-5

Bacilli

Bacillaceae

Bacillus a

-

0.988

00276

94.5

NT

LE-E4

Actinobacteria

Micrococcaceae

Micrococcus

100

-

00036

92.5

-

PE-3

Actinobacteria

Micrococcaceae

Kocuria

100

-

00385

92.1

+

PE-4

Actinobacteria

Micrococcaceae

Kocuria

100

-

00385

92.1

+

PE-2

Alphaproteobacteria

Rhizobiaceae

Rhizobium

96

-

00989

92.5

ME-1

Alphaproteobacteria

Brucellaceae

Ochrobactrum

100

-

00106

92.1

NT

ME-2

Alphaproteobacteria

Brucellaceae

Ochrobactrum

100

-

00106

92.1

NT

ME-3

Alphaproteobacteria

Brucellaceae

Ochrobactrum

100

-

00106

92.1

NT

ME-4

Alphaproteobacteria

Brucellaceae

Ochrobactrum

100

-

00106

92.1

NT

ME-5

Alphaproteobacteria

Brucellaceae

Ochrobactrum

100

-

00106

92.1

NT

Cultured bacteria are abundant during Lake Erie cyanoHABs, rapidly decompose H2O2

Abundance of OTU00127 (isolates LE-E6 and LE-L6) in Lake Erie.

Estimated rates of H2O2 decay in lab cultures

Heterotrophic bacteria protect Microcystis from external H2O2

A DAPI-stained Microcystis colony, courtesy of Melissa Duhaime

Microcystis cells

associated bacteriaH2O2

H2O + O2

Does Lake Erie represent everywhere, always?

Green Bay, Lake Michigan

Lake Winnipeg, Mantoba

Lake Okeechobee, FL Saginaw Bay, Lake Huron

Dianshan Lake, reservoir for Shanghai (Pop. 24+ million)

Questions?

Place Photo Here

Slide Number 1Co-authors & FundingSlide Number 3HABs have expanded globally in the past few decadesWhy have HABs proliferated?HABs and human healthHABs and human healthHABs have inspired horror moviesCyanobacterial blooms are a global phenomenonThe ecology of cyanobacterial blooms is complexCyanotoxins impact all levels of the food web, including humans Slide Number 12Slide Number 13Slide Number 14Slide Number 15Slide Number 16Slide Number 17Kisumu Bay, Lake Victoria, Kenya, 2004Overarching research statement: Understanding the drivers of bloom ecology will aid in enhancing predictive models that forecast bloom size, location AND toxicitySlide Number 20Slide Number 21Slide Number 22Slide Number 23Slide Number 24Slide Number 25Slide Number 26Slide Number 27Slide Number 28Slide Number 29Trends are consistent between yearsSlide Number 31Slide Number 32Slide Number 33Slide Number 34Viking Mission(s) to Mars��First time a biological probe was used to search for life outside of this planet��Autonomous near real-time toxin detection for Lake ErieSlide Number 37Slide Number 38Detection and quantification limits during the September 2016 deploymentMultiplexing toxins on ESPSlide Number 41Analysis of Microcystis subpopulations: �Slide Number 43Slide Number 44Slide Number 45Microcystis has ~1900 core genes and a diverse flexible genomeCOG distribution of core and flexible genes in Lake Erie isolates show an enrichment of defense-oriented genes in the flexible genomeMany of the unique genes found in new Microcystis isolates involve genome flexibilityUnique genes are present and expressed in a naturally occurring bloomSlide Number 50Toxic strains decline with lower nitrate concentrationsSlide Number 52Slide Number 53Proteobacteria and Cyanobacteria dominate but cyanos are doing the workEco-transcriptomic surveys of LE CHABsSlide Number 56Slide Number 57Slide Number 58Slide Number 59Pigment-inferred growth and particulate toxicity not impacted by ROSSlide Number 61Slide Number 62Slide Number 63Slide Number 64Isolation of bloom-associated heterotrophic bacteria from Lake Erie�Cultured bacteria are abundant during Lake Erie cyanoHABs, rapidly decompose H2O2 Slide Number 67Does Lake Erie represent everywhere, always?Dianshan Lake, reservoir for Shanghai (Pop. 24+ million) Slide Number 70