INVESTIGATION OF FRESHWATER MUSSEL PHYSIOLOGY AND

REPRODUCTIVE BIOLOGY TO INFORM CONSERVATION

by

ANDREA KAY FRITTS

(Under the Direction of Robert B. Bringolf)

ABSTRACT

Filter feeding freshwater mussels (order Unionoida) fulfill an important ecological niche, but a suite of anthropogenic perturbations have made them the most imperiled faunal group in North America. Threats to mussels include habitat degradation, pollution, and alterations to natural flow regimes. The Apalachicola-

Chattahoochee-Flint River Basin is a watershed under increasing pressure from human populations and is also home to a diverse assemblage of aquatic organisms including five federally listed mussel species. The federal recovery plan for these species outlines specific research objectives that will better equip scientists and managers with the necessary tools to protect and restore populations of these species. This dissertation has focused on addressing some of these critical data gaps by investigating the ecological relevance of the sodium chloride glochidia viability test, conducting host determination trials for the Purple Bankclimber mussel (Elliptoideus sloatianus), and developing nonlethal methods for assessing the physiological response of mussels to various stress events. Freshwater mussels are characterized by a unique lifecycle in which the glochidia

larvae must attach to a vertebrate host to metamorphose into a juvenile mussel. This larval stage is used in toxicity testing to evaluate the effects of contaminants on freshwater mussels and for the derivation of water quality criteria. My results indicated that the viability of glochidia as measured by the sodium chloride test is an ecologically relevant measure of the health of glochidia. The discovery of Gulf Sturgeon (Acipenser oxyrinchus desotoi) as the primary hosts for Purple Bankclimber mussels has supplied important information for the preservation and management of wild mussel populations as well as providing the necessary data to initiate captive propagation. Changes in tissue glycogen and hemolymph chemistry parameters are potential biomarkers for monitoring stress in freshwater mussels. The factors of discharge, size, sex, and species were most commonly found to affect the biological responses in our models and we recommend that future research into the effects of drought and stress should include the use of alanine aminotransferase, aspartate aminotransferase, bicarbonate, and calcium. Combined, these data will allow scientists and managers alike to advance the conservation of these intriguing and ecologically important animals.

INDEX WORDS: Glochidia, Unionidae, Elliptoideus sloatianus, Apalachicola-

Chattahoochee-Flint River Basin, Biomarker

INVESTIGATION OF FRESHWATER MUSSEL PHYSIOLOGY AND

REPRODUCTIVE BIOLOGY TO INFORM CONSERVATION

by

ANDREA KAY FRITTS

B.S., University of Wisconsin-River Falls, 2007

M.S. Missouri State University, 2009

A Dissertation Submitted to the Graduate Faculty of The University of Georgia in Partial

Fulfillment of the Requirements for the Degree

DOCTOR OF PHILOSOPHY

ATHENS, GEORGIA

2013

© 2013

ANDREA KAY FRITTS

All Rights Reserved

INVESTIGATION OF FRESHWATER MUSSEL PHYSIOLOGY AND

REPRODUCTIVE BIOLOGY TO INFORM CONSERVATION

by

ANDREA KAY FRITTS

Major Professor: Robert B. Bringolf

Committee: James T. Peterson Mary C. Freeman C. Rhett Jackson

Electronic Version Approved:

Maureen Grasso Dean of the Graduate School The University of Georgia August 2013

DEDICATION

This dissertation would not have been possible without the unending support from

Mark Fritts: fellow biologist, terrific cook, and wonderful husband. Our hours together in the rivers of Georgia and our countless dinner conversations about research projects, experimental design, and delectable recipes have made the past four years not only memorable, but also incredibly fun! I look forward to our continued exploration of the natural world, with you and I side by side.

iv

ACKNOWLEDGEMENTS

When I first visited the University of Georgia and the lab of Dr. Robert Bringolf in January 2009, I realized very quickly that this was the place where I wanted to pursue my doctoral degree. Since the very first moment that I met Dr. Bringolf, he has been the type of advisor that every graduate student hopes for. Supportive and kind, yet one who challenges his students to jump out of their comfort zones and to grow as a scientist. His unending encouragement and passion for science has produced many wonderful experiences and memories that I will cherish as I leave Georgia and embark on new research adventures.

Special thanks go out to Bob Ratajczak, Colin Shea, and Jason Wisniewski. Bob was always willing to lend a hand in the field or lab and he made sure that we were equipped with all of the peculiar supplies that we needed. Colin and Jason taught me the mussels of the Flint River and graciously helped to collect mussels not only during the pleasant summer months but also during chilly winter sampling trips. Many other individuals have contributed to the successful completion of this dissertation research, including: Sandy Abbott, Brett Albanese, Chris Barnhart, Mike Bednarski, Shay Bush,

Ben Carswell, Greg Cope, Julie Creamer, Zack DeWolf, Justin Dycus, Dewayne Fox,

Mark Fritts, Pete Hazelton, Karen Herrington, Whitney Jacobs, Cecil Jennings, Jennifer

Johnson, Rachel Katz, Kristen Kellock, Teresa Newton, Doug Peterson, Sandy Pursifull,

Bernard Sietman, Channing St. Aubin, Jim Stoeckel, Brittany Trushel, Amos Tuck, Dan

Watrous, Deb Weiler, Naeem Willett, and Jim Williams. My committee members, Dr.

v

Mary Freeman, Dr. Jim Peterson, and Dr. Rhett Jackson have also been terrific sources of support and constructive input over the course of this project. Owens and Williams Fish

Farm were incredibly generous in donating a tremendous number of largemouth bass for an assortment of experiments.

Lastly, a big thank you to my family, who were the first to instill a love of nature within me. My parents, Bill and Anna Crownhart, my siblings, Maria, Amy, John, and

Rachel, and my sweet husband, Mark Fritts---you have all been incredible sources of support during my graduate school career.

vi

TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS ...... v

LIST OF TABLES ...... x

LIST OF FIGURES ...... xiii

CHAPTER

1 BACKGROUND AND CHAPTER ORGANIZATION ...... 1

INTRODUCTION ...... 1

REFERENCES ...... 10

2 ASSESSMENT OF TOXICITY TEST ENDPOINTS FOR THE

GLOCHIDIA LARVAE OF FRESHWATER MUSSELS ...... 16

ABSTRACT ...... 17

INTRODUCTION ...... 18

METHODS ...... 20

RESULTS ...... 26

DISCUSSION ...... 29

REFERENCES ...... 35

3 CRITICAL LINKAGE OF IMPERILED SPECIES: GULF STURGEON AS

HOST FOR THE PURPLE BANKCLIMBER MUSSELS...... 48

ABSTRACT ...... 49

INTRODUCTION ...... 51

vii

METHODS ...... 54

RESULTS ...... 58

DISCUSSION ...... 58

REFERENCES ...... 65

4 EVALUATION OF PHYSIOLOGICAL BIOMARKERS OF STRESS IN

FRESHWATER MUSSELS ...... 80

ABSTRACT ...... 81

INTRODUCTION ...... 83

METHODS ...... 87

RESULTS ...... 93

DISCUSSION ...... 95

REFERENCES ...... 103

5 NON-LETHAL ASSESSMENT OF FRESHWATER MUSSEL RESPONSE

TO CHANGES IN ENVIRONMENTAL FACTORS IN THE LOWER

FLINT RIVER BASIN, GEORGIA, U.S.A...... 120

ABSTRACT ...... 121

INTRODUCTION ...... 123

METHODS ...... 127

RESULTS ...... 133

DISCUSSION ...... 138

REFERENCES ...... 143

6 SYNTHESIS AND CONCLUSIONS ...... 170

INTRODUCTION ...... 170

viii

CHAPTER SYNTHESIS ...... 170

CONCLUSIONS...... 174

REFERENCES ...... 176

ix

LIST OF TABLES

Page

Table 2.1: Dates and locations of freshwater mussel collections ...... 39

Table 2.2: Results of one-way ANOVA comparing the change in glochidia viability or

metamorphosis over time for seven freshwater mussel species...... 40

Table 2.3: Results of one-way ANOVA comparing the change in glochidia viability or

metamorphosis for glochidia exposed to increasing concentrations of sodium

chloride (NaCl) or copper (Cu) ...... 42

Table 3.1: Fish species tested as potential hosts for Purple Bankclimbers ...... 76

Table 3.2: Mean (± 95% CI) total length (TL) and mass of fish species that facilitated

metamorphosis of Purple Bankclimber glochidia to juveniles ...... 79

Table 4.1: Water quality parameters (mean ± SE; n = 35 per temperature) measured

during thermal stress exposures of Elliptio crassidens, Villosa vibex, and Villosa

lienosa ...... 111

Table 4.2: Hemolymph and tissue parameter values for three mussel species (Elliptio

crassidens, Villosa vibex, and Villosa lienosa) exposed to elevated water

temperatures (25, 30 35°C) for 7 days ...... 112

Table 4.3: Results of ANOVA analysis for freshwater mussel hemolymph parameters that

responded to thermal stress ...... 114

x

Table 4.4: Cox proportional hazard regression analyses for effects of tissue and

hemolymph extraction on the long-term survival of Villosa vibex and Elliptio

crassidens ...... 116

Table 4.5: Hemolymph parameters measured from the anterior and posterior adductor

muscles of Elliptio crassidens (n = 10) ...... 117

Table 5.1: List of parameters included for candidate models for biomarkers response of

freshwater mussels in the lower Flint River Basin, southwestern Georgia ...... 149

Table 5.2: Akaike’s Information Criterion (AICc), number of parameters (K), ΔAICc, and

Akaike weights (wi) for the confidence set of models estimating hemolymph

alanine aminotransferase response to environmental parameters in the lower Flint

River Basin, Georgia ...... 150

Table 5.3: Akaike’s Information Criterion (AICc), number of parameters (K), ΔAICc, and

Akaike weights (wi) for the confidence set of models estimating hemolymph

aspartate aminotransferase response to environmental parameters in the lower

Flint River Basin, Georgia ...... 152

Table 5.4: Akaike’s Information Criterion (AICc), number of parameters (K), ΔAICc, and

Akaike weights (wi) for the confidence set of models estimating hemolymph

bicarbonate response to environmental parameters in the lower Flint River Basin,

Georgia ...... 154

Table 5.5: Akaike’s Information Criterion (AICc), number of parameters (K), ΔAICc, and

Akaike weights (wi) for the confidence set of models estimating hemolymph

calcium response to environmental parameters in the lower Flint River Basin,

Georgia ...... 155

xi

Table 5.6: Akaike’s Information Criterion (AICc), number of parameters (K), ΔAICc, and

Akaike weights (wi) for the confidence set of models estimating hemolymph

magnesium response to environmental parameters in the lower Flint River Basin,

Georgia ...... 156

Table 5.7: Akaike’s Information Criterion (AICc), number of parameters (K), ΔAICc, and

Akaike weights (wi) for the confidence set of models estimating tissue glycogen

response to environmental parameters in the lower Flint River Basin, Georgia

...... 157

xii

LIST OF FIGURES

Page

Figure 2.1: Comparison of percent viability (bars) and percent metamorphosis success

(lines; ± 95% CI) for glochidia following release from the female mussels of A)

Lampsilis siliquoidea (Jan), B) Lampsilis siliquoidea (Jun), C) Lampsilis cardium,

D) Lampsilis dolabraeformis, E) Amblema plicata, F) Villosa lienosa, and G)

Utterbackia imbecillis ...... 43

Figure 2.2: Comparison of percent viability (bars) and percent metamorphosis success

(lines; ± 95% CI) for glochidia of Ptychobranchus occidentalis: A) free glochidia

in reconstituted water, B) glochidia in conglutinates in reconstituted water, C) free

glochidia exposed to river water and sediment, and D) glochidia in conglutinates

exposed to river water and sediment ...... 45

Figure 2.3: Comparison of percent viability (bars) and percent metamorphosis success

(lines; ± 95% CI) for glochidia exposed to increasing concentrations of sodium

chloride (NaCl) or copper (Cu) for 24 h: A) Lampsilis cardium (Dec) NaCl

exposure, B) Lampsilis cardium (Jun) NaCl exposure, C) Lampsilis

dolabraeformis NaCl exposure, and D) Lampsilis dolabraeformis exposed to Cu

...... 46

Figure 2.4: Comparison of NaCl 24-h EC50 (± 95% confidence interval) for Lampsilis

cardium glochidia collected early in the brooding period (December) and late in

the brooding period (June) ...... 47

xiii

Figure 3.1: Apalachicola-Chattahoochee-Flint River (ACF) in Alabama, Florida, and

Georgia showing location of Jim Woodruff Dam ...... 72

Figure 3.2: Mean (± 95% CI) % metamorphosis of Purple Bankclimber glochidia on

sturgeons (Acipenser spp.) and darters (Percina spp.) ...... 73

Figure 3.3: Mean (± 95% CI) number of Purple Bankclimber juveniles produced/host fish

and number of juveniles produced/g host fish biomass during laboratory trials ...74

Figure 3.4: Mean (± 95% CI) median lethal concentrations (LC50) of NaCl for juvenile

Purple Bankclimber mussels produced from sturgeons (Acipenser spp.) and

darters (Percina spp.) ...... 75

Figure 4.1: Probability of survival (± SE) for Villosa vibex over 820 days after tissue and

hemolymph extraction ...... 118

Figure 4.2: Probability of survival (± SE) for Elliptio crassidens over 945 days after

tissue and hemolymph extraction ...... 119

Figure 5.1: Locations of five sampling sites within the lower Flint River Basin ...... 159

Figure 5.2: Relative weights of supporting models of six biological responses in

freshwater mussels of the lower Flint River Basin: A) alanine aminotransferase,

B) aspartate aminotransferase, C) bicarbonate, D) calcium, E) magnesium, and F)

tissue glycogen ...... 160

Figure 5.3: Parameter estimates (±SE) of the best supporting models of six biological

responses in freshwater mussels of the lower Flint River Basin: A) alanine

aminotransferase, B) aspartate aminotransferase, C) bicarbonate, D) calcium, E)

magnesium, and F) tissue glycogen ...... 164

xiv

Figure 5.4: Predicted response of alanine aminotransferase (ALT) to changes in discharge

at two different temperature regimes for three freshwater mussel species (Elliptio

crassidens, Villosa vibex, and Villosa lienosa) of average length during the spring

season in the Dougherty Plain physiographic province ...... 168

Figure 5.5: Predicted response of calcium to changes in discharge among three freshwater

mussel species: Elliptio crassidens, Villosa vibex, and Villosa lienosa ...... 169

xv

CHAPTER 1

BACKGROUND AND CHAPTER ORGANIZATION

INTRODUCTION

Freshwater mussels (order Unionoida) are organized into six families and more than 850 species worldwide (Graf and Cummings 2007). North America is an epicenter of unionid diversity with approximately 300 species currently recognized. The greatest diversity of species is found in the southeastern United States, specifically in the states of

Tennessee, Alabama, and Georgia (Parmalee and Bogan 1998, Williams et al. 2008).

The study of freshwater mussels has increased substantially in recent years (Haag

2012). This interest has largely been associated with the recognition that many mussel taxa have become extinct within recent memory and that many more taxa are currently threatened by extirpation and extinction in the wild (Ricciardi and Rasmussen 1999,

Lydeard et al. 2004, Strayer et al. 2004). Among the North American unionids, 72% of taxa are threatened with some risk for extinction (Williams et al. 1993). Anthropogenic threats to freshwater mussels include habitat degradation, pollution, and alterations to natural flow regimes associated with municipal and agricultural water withdrawals and the operations of hydroelectric facilities (Bogan 1993).

Many of these threats are of serious concern in the Flint River Basin in southwestern

Georgia, a watershed characterized by particularly high numbers of endemic unionids

(Williams et al. 2008). The Flint River is part of the Apalachicola-Chattahoochee-Flint

1

(ACF) River Basin and is highly impacted by agricultural water withdrawals. For the past 20 years, the ACF has been part of a tri-state water dispute between the states of

Georgia, Alabama, and Florida (Ruhl 2005). The Flint River Basin is home to a diverse assemblage of aquatic organisms, including five federally listed mussel species: shiny rayed pocketbook (Hamiota subangulata), oval pigtoe (Pleurobema pyriforme), Gulf moccasinshell (Medionidus penicillatus), purple bankclimber (Elliptoideus sloatianus), and the fat threeridge (Amblema neislerii) (Williams et al. 2008). These species were listed under the Endangered Species Act in 1998 because of their declining populations

(U.S. Fish and Wildlife Service 1998). The main factors cited in the decision to list these species included habitat loss, range restriction, population fragmentation and population size reduction (U.S. Fish and Wildlife Service 1998). The restricted distribution makes the remaining populations at risk from water quality degradations and the detrimental effects of genetic isolation. The recovery goal for these species calls for the restoration of viable populations throughout a substantial proportion of their historical distribution and the reduction or elimination of factors that may threaten the species’ long-term persistence (U.S. Fish and Wildlife Service 2003).

Strategic recovery plans for endangered taxa require a multi-faceted approach to aid the recovery of a given species or population. These approaches recognize the need for research advancing the knowledge of life history, the identification of suitable habitats and potential stressors. These areas of research are especially critical for advancing the conservation goals for the existing subpopulations of imperiled mussel species of the ACF.

2

The threats and challenges faced by freshwater mussel species in the Flint River

Basin are similar to those experienced by other imperiled mussels worldwide. Lessons gained from the conservation of Flint River basin species could advance the protection and restoration of other freshwater mussels in different watersheds. Therefore, the overarching goal of my dissertation research is to address these critical data gaps and advance our understanding of the information that is necessary to conserve of this unique group of animals.

Early life history of freshwater mussels

Members of the Order Unionoida have a unique reproductive cycle and are distinguished from all other bivalves by possessing parasitic larvae (Kat 1984, Graf and

Cummings 2007, Barnhart et al. 2008; Fritts et al. 2013). The larvae of Unionidae and

Margaritiferidae are called glochidia (s. glochidium). Female mussels brood their eggs within modified hollow gills (marsupia). Male mussels release sperm into the water column, which are filtered from the water by the female and fertilize the eggs. Glochidia develop within the eggs over a period of days to weeks. The larvae may be released quickly after finishing their development or may continue to be brooded in the female’s marsupia for up to several months (Kat 1984). However, parasitic glochidia generally do not develop further unless they attach to a vertebrate host. If the glochidia attach to the correct fish or amphibian species, they will be encapsulated on the gills or fins of their host and metamorphose to the juvenile stage. The hosts vary depending on mussel species, with some mussels able to metamorphose on a wide selection of fish species while others can apparently only metamorphose on a single host. Examples of hosts

3

include, but are not limited to: basses (Centrarchidae), catfishes (Ictaluridae), walleye

(Percidae), minnows (Cyprinidae), sturgeon (Acipenseridae) and gars (Lepisosteidae).

For many mussel species the fish hosts are unknown, or reported hosts are based on potentially erroneous identifications (Haag and Warren 2003). Accurate knowledge of host fish(es) is an integral component of a species protection plan because suitable host fish must be available for successful mussel recovery (U.S. Fish and Wildlife Service

2003). Without these data, researchers are limited in their understanding of an essential component of the early life history of imperiled species and the development of captive propagation programs may be hindered.

Host identification research is needed for the five federally listed species in the

Flint River. Initial host research was completed by O’Brien, Brim-Box and Williams

(1999, 2002), who identified marginal hosts for several of the listed mussels but failed to identify the primary hosts. Their data indicated some fish families that should be the focus of future research efforts, e.g., host fishes that should be examined for Elliptoideus sloatianus include Percidae (O'Brien and Williams 2002). The recovery plan for E. sloatianus specifically noted the need to test anadromous and other imperiled fishes (U.S.

Fish and Wildlife Service 2003). With the discovery of primary hosts for these listed species, it would be possible to begin captive propagation of these species and could provide managers with the necessary information to make more accurate decisions regarding the management of water resources in the Flint River Basin.

4

Viability test for freshwater mussel glochidia

Mussel researchers in the early 1900’s observed that the viability, the observed capacity of the mussel larva to close its valves, could be assessed by adding sodium chloride or a saturated saline solution to glochidia (Coker et al. 1921). Glochidia that close in response to the sodium chloride are considered viable and assumed able to infect a host fish. Those glochidia that do not close their valves are considered unviable and are assumed to be unable to infect a host and complete their lifecycle. This practice of assessing viability with the sodium chloride test has persisted to present studies. This technique is used not only for propagation and host work efforts, but also as the endpoint in toxicological experiments using mussel glochidia as test subjects (e.g. Bringolf et al.

2007, Wang et al. 2007, Cope et al. 2008). While the use of glochidia in toxicological trials is widely accepted, there is uncertainty about the ecological relevance of the viability measure as an indication of the true health and survival of mussel glochidia exposed to various contaminants. The recovery plan for the imperiled ACF mussel species calls for an investigation of the sensitivity of each freshwater mussel life stage to various contaminants (U.S. Fish and Wildlife Service 2003) and it is critical that any toxicological tests have ecologically relevant endpoints to accurately determine the effects of various contaminants on wild populations.

To date, there have been only a limited number of studies that have tested the ecological relevance of the sodium chloride test. However, there has been a rising concern among malacologists and toxicologists that the NaCl viability test is not representative of actual infectivity (i.e., ability to attach and metamorphose) of the glochidia. With the increasing interest in using glochidia for environmental toxicology

5

and the derivation of water quality criteria, it is necessary to assess whether the viability calculated with the sodium chloride test is indicative of the ability of glochidia to successfully attach and metamorphose on a host fish. More accurate knowledge of the ecological relevance of the glochidia viability test will help toxicologists more accurately determine the sub-lethal and lethal concentrations of various contaminants on larval mussels.

Non-lethal approach for assessing stress

The ability to monitor the health and stress levels of individuals within a population is an area that is currently lacking within freshwater mussel ecology.

Biomarkers show potential as a mechanism for evaluating the health of wild populations of mussels; they are developed by tracking specific biological/physiological processes and how those processes react to changes in the organism’s habitat (Gagne et al. 2002,

Blaise and Gagne 2009). Biomarkers of stress have been identified in commercially valued marine bivalves (Liu et al. 2004, Li et al. 2007), but information on freshwater bivalves is still scarce. Methods are needed for monitoring the relative health of mussel populations faced with extreme environmental variations and exposure to contaminants

(U.S. Fish and Wildlife Service 2003). Identification of useful biomarkers that can be sampled non-lethally is the primary challenge to creating an effective biomonitoring plan for endangered mussels (Gustafson et al. 2005a,b).

Prior research has indicated that decreases in glycogen, a primary energy storage molecule, may be a useful indicator of stress in freshwater mussels. Glycogen levels have been shown to decrease under stressful conditions such as starvation (Patterson et al.

6

1999) and zebra mussel infestation (Baker and Hornbach 2000). Amblema plicata and

Lampsilis radiata exhibited significant decreases in glycogen levels after three months of a zebra mussel infestation (Haag et al. 1993). Glucose is a primary source of energy in all organisms and glycogen is the primary energy storage molecule in freshwater mussels.

When glucose levels decline in the hemolymph, glycogen can be catabolized to supplement the glucose levels in the circulatory fluids (De Zwaan and Zandee 1972).

Therefore, the lower glycogen levels that result from stress events may indicate that glucose is being mobilized from glycogen reserves through catabolism.

Changes in hemolymph (mussel circulatory fluid) chemistry profiles have also shown potential as a tool for biomonitoring (Gustafson et al. 2005a), but have not yet been implemented for this use in freshwater mussel species. Traditional sampling of hemolymph from the ventricular chamber of the heart or from the pericardial sac has resulted in high mortality (100%) in freshwater mussels (Gustafson et al. 2005a). The efficacy and impacts of sampling hemolymph from the anterior adductor mussel have been investigated as an alternative collection locality. Gustafson et al. (2005a,b) extracted hemolymph from the anterior adductor muscle of 30 Elliptio complanata and monitored those individuals for 3 months to assess impacts on survival. There were no differences in survival between the control group and the treatment group. Additional animals were monitored to assess impacts of repeated sampling over a 5 month period.

Two months after the final sampling, there was no evidence of increased mortality in the cohort that was sampled repeatedly (Gustafson et al. 2005a). This technique has been utilized in the field to collect baseline data for Elliptio complanata populations in North

Carolina between the months of May-July (Gustafson et al. 2005b). The ability to take

7

small tissue biopsies and to collect hemolymph in situ allows researchers to monitor populations without the need to transport animals back to the laboratory, thus minimizing stress on the organisms.

Water temperatures in the southeastern United States have been documented to reach 34° C and higher (Dyar and Alhadeff 1997), and changes in climate patterns may result in even higher summer temperatures in the future. Pandolfo et al. (2010) demonstrated significant decreases in survival of three species of juvenile freshwater mussels at temperatures of 36°C and higher, which are environmentally relevant temperatures for the Southeast. It is imperative that we develop a greater understanding of how freshwater mussels will respond in the face of extreme temperatures and the secondary effects of drought conditions. The development of both in situ monitoring and ex situ laboratory experiments is needed for creating and refining biomonitoring programs of imperiled freshwater mussels.

Objectives and Chapter Organization

I addressed the following objectives in order to advance the conservation of freshwater mussels through the study of physiology and reproductive biology: 1) assess the ecological relevance of the sodium chloride viability test, 2) determine primary hosts for the federally listed purple bankclimber mussel (Elliptoideus sloatianus), and 3) establish biomarkers that can be used to non-lethally assess the health of freshwater mussel populations. This final broad objective was subdivided into a laboratory component, in which we attempted to establish useful biomarkers of stress in three species of freshwater mussels, and a second field component, in which we use

8

hierarchical modeling to evaluate the spatial and temporal response of our chosen biomarkers in natural riverine settings.

9

REFERENCES

Baker, S.M. and D.J. Hornbach. 2000. Physiological status and biochemical composition

of a natural population of unionid mussels (Amblema plicata) infested by zebra

mussels (Dreissena polymorpha). American Midland Naturalist 143:443-452.

Barnhart, M.C., W.R. Haag, and W.N. Roston. 2008. Adaptations to host infection and

larval parasitism in Unionoida. Journal of the North American Benthological

Society 27:370-394.

Blaise, C. and F. Gagne. 2009. Bioassays and Biomarkers, Two Pillars of Ecotoxicology:

Past, Present and Prospective Uses. Fresenius Environmental Bulletin 18:135-

139.

Bogan, A.E. 1993. Fresh-Water Bivalve Extinctions (Mollusca, Unionoida) - a Search for

Causes. American Zoologist 33:599-609.

Bringolf, R. B., W. G. Cope, M. C. Barnhart, S. Mosher, P. R. Lazaro, and D. Shea.

2007. Acute and chronic toxicity of pesticide formulations (atrazine, chlorpyrifos

and permethrin) to glochidia and juveniles of Lampsilis siliquoidea (Unionidae).

Environmental Toxicology and Chemistry 26:2101–2107.

Coker, R.E., A.F. Shira, H.W. Clark, and A.D. Howard. 1921. Natural history and

propagation of freshwater mussels. Bulletin of the U.S. Bureau of Fisheries

37:75–182.

Cope, W.G., R.B. Bringolf, D.B. Buchwalter, T.J. Newton, C.G. Ingersoll, N. Wang, T.

Augspurger, F.J. Dwyer, M.C. Barnhart, R.J. Neves, and E. Hammer. 2008.

Differential exposure, duration, and sensitivity of unionoidean bivalve life stages

10

to environmental contaminants. Journal of the North American Benthological

Society 27:451-462.

De Zwaan, A. and T.C.M. Wijsman. 1976. Anaerobic metabolism in Bivalvia (Mollusca).

Comparative Biochemistry and Physiology 54B: 313-324.

Dyar, T.R. and S.J. Alhadeff. 1997. Stream-temperature characteristics in Georgia. USGS

96-4203, Atlanta, GA.

Fritts, M.W., A.K. Fritts, S.A. Carleton, and R.B. Bringolf. 2013. Shifts in stable isotope

signatures confirm parasitic relationship of freshwater mussel glochidia attached

to host fish. Journal of Molluscan Studies 79:163-167.

Gagne, F., C. Blaise, I. Aoyama, R. Luo, C. Gagnon, Y. Couillard, P. Campbell, and M.

Salazar. 2002. Biomarker study of a municipal effluent dispersion plume in two

species of freshwater mussels. Environmental Toxicology 17:149-159.

Graf, D.L. and K.S. Cummings. 2007. Review of the systematics and global diversity of

freshwater mussel species (Bivalvia : Unionoida). Journal of Molluscan Studies

73:291-314.

Gustafson, L.L., M.K. Stoskopf, A.E. Bogan, W. Showers, T.J. Kwak, S. Hanlon, and

J.F. Levine. 2005a. Evaluation of a nonlethal technique for hemolymph collection

in Elliptio complanata, a freshwater bivalve (Mollusca : Unionidae). Diseases of

Aquatic Organisms 65:159-165.

Gustafson, L.L., M.K. Stoskopf, W. Showers, W.G. Cope, C. Eads, R. Linnehan, T.J.

Kwak, B. Andersen, and J.F. Levine. 2005b. Reference ranges for hemolymph

chemistries from Elliptio complanata of North Carolina. Diseases of Aquatic

Organisms 65:167-176.

11

Haag, W.R., D.J. Berg, D.W. Garton, and J.L. Farris. 1993. Reduced survival and fitness

in native bivalves in response to rouling by the introduced zebra mussel

(Dreissena polymorpha) in western Lake Erie. Canadian Journal of Fisheries and

Aquatic Sciences 50:13-19.

Haag, W.R. and M.L. Warren. 2003. Host fishes and infection strategies of freshwater

mussels in large Mobile Basin streams, USA. Journal of the North American

Benthological Society 22:78-91.

Haag, W.R. 2012. North American Freshwater Mussels: Natural History, Ecology, and

Conservation. Cambridge University Press, Cambridge, United Kingdom.

Kat, P.W. 1984. Parasitism and the Unionacea (Bivalvia). Biological Reviews of the

Cambridge Philosophical Society 59:189-207.

Lacoste, A., S.K. Malham, F. Gelebart, A. Cueff, and S.A. Poulet. 2002. Stress-induced

immune changes in the oyster Crassostrea gigas. Developmental and Comparative

Immunology 26:1-9.

Li, Y., J.G. Qin, C.A. Abbott, X.X. Li, and K. Benkendorff. 2007. Synergistic impacts of

heat shock and spawning on the physiology and immune health of Crassostrea

gigas: an explanation for summer mortality in Pacific oysters. American Journal

of Physiology-Regulatory Integrative and Comparative Physiology 293:R2353-

R2362.

Liu, S.L., X.L. Jiang, X.K. Hu, J. Gong, H. Hwang, and K. Mai. 2004. Effects of

temperature on non-specific immune parameters in two scallop species:

Argopecten irradians (Lamarck 1819) and Chlamys farreri (Jones & Preston

1904). Aquaculture Research 35:678-682.

12

Lydeard, C., R.H. Cowie, W.F. Ponder, A.E. Bogan, P. Bouchet, S.A. Clark, K.S.

Cummings, T.J. Frest, O. Gargominy, D.G. Herbert, R. Hershler, K.E. Perez, B.

Roth, M. Seddon, E.E. Strong, and F.G. Thompson. 2004. The global decline of

nonmarine mollusks. Bioscience 54:321-330.

O'Brien, C.A. and J.B. Box. 1999. Reproductive biology and juvenile recruitment of the

shinyrayed pocketbook, Lampsilis subangulata (Bivalvia : Unionidae) in the Gulf

coastal plain. American Midland Naturalist 142:129-140.

O'Brien, C.A. and J.A. Williams. 2002. Reproductive biology of four freshwater mussels

(Bivalvia : Unionidae) endemic to eastern Gulf Coastal Plain drainages of

Alabama, Florida, and Georgia. American Malacological Bulletin 17:147-158.

Pandolfo, T.J., W.G. Cope, and C. Arellano. 2010. Thermal Tolerance of Juvenile

Freshwater Mussels (Unionidae) under the Added Stress of Copper.

Environmental Toxicology and Chemistry 29:691-699.

Parmalee, P.W. and A.E. Bogan. 1998. The freshwater mussels of Tennessee. University

of Tennessee Press, Knoxville, TN.

Patterson, M.A., B.C. Parker, and R.J. Neves. 1999. Glycogen concentration in the

mantle tissue of freshwater mussels (Bivalvia : Unionidae) during starvation and

controlled feeding. American Malacological Bulletin 15:47-50.

Pipe, R.K., C. Porte, and D.R. Livingstone. 1993. Antioxidant Enzymes Associated with

the Blood-Cells and Hemolymph of the Mussel Mytilus-Edulis. Fish & Shellfish

Immunology 3:221-233.

Ricciardi, A. and J.B. Rasmussen. 1999. Extinction rates of North American freshwater

fauna. Conservation Biology 13:1220-1222.

13

Ruhl, J.B. 2005. Water wars, eastern style: divvying up the Apalachicola-Chattahoochee-

Flint River Basin. Journal of Contemporary Water Research and Education 131:

47-52.

Strayer, D.L., J.A. Downing, W.R. Haag, T.L. King, J.B. Layzer, T.J. Newton, and S.J.

Nichols. 2004. Changing perspectives on pearly mussels, North America's most

imperiled animals. Bioscience 54:429-439.

U.S. Fish and Wildlife Service. 1998. Endangered and Threatened Wildlife and Plants;

Determination of Endangered Status of Five Freshwater Mussels and Threatened

Status for Two Freshwater Mussels from the Eastern Gulf Slope Drainages of

Alabama, Florida, and Georgia. Federal Register 63(50): 12664-12687.

U.S. Fish and Wildlife Service. 2003. Recovery Plan for Endangered Fat Threeridge

(Amblema neislerii), Shinyrayed Pocketbook (Lampsilis subangulata), Gulf

Moccasinshell (Medionidus penicillatus), Ochlockonee Moccasinshell

(Medionidus simpsonianus), and Oval Pigtoe (Pleurobema pyriforme); and

Threatened Chipola Slabshell (Elliptio chipolaensis), and Purple Bankclimber

(Elliptoideus sloatianus). Atlanta, Georgia. 142 pp.

Wang, N., T. Augspurger, M.C. Barnhart, J.R. Bidwell, W.G. Cope, F.J. Dwyer, S. Geis,

I.E. Greer, C.G. Ingersoll, C.M. Kane, T.W. May, R.J. Neves, T.J. Newton, A.D.

Roberts, and D.W. Whites. 2007. Intra- and interlaboratory variability in acute

toxicity tests with glochidia and juveniles of freshwater mussels (Unionidae).

Environ Toxicol Chem 26:2029-2035.

14

Williams, J.D., A.E. Bogan, and J.T. Garner. 2008. Freshwater mussels of Alabama and

the Mobile Basin in Georgia, Mississippi and Tennessee. The University of

Alabama Press, Tuscaloosa, Alabama.

Williams, J.D., M.L. Warren, Jr., K.S. Cummings, J.L. Harris, and R.J. Neves. 1993.

Conservation status of freshwater mussels of the United States and Canada.

Fisheries 18:6-22.

Xue, Q.G. and T. Renault. 2000. Enzymatic activities in European flat oyster, Ostrea

edulis, and Pacific oyster, Crassostrea gigas, hemolymph. Journal of Invertebrate

Pathology 76:155-163.

15

CHAPTER 2

ASSESSMENT OF TOXICITY TEST ENDPOINTS FOR THE GLOCHIDIA

LARVAE OF FRESHWATER MUSSELS1

1 Fritts, A.K., M.C. Barnhart, M. Bradley, N. Liu, W.G. Cope, E. Hammer, and R.B.

Bringolf. Submitted to Environmental Toxicology and Chemistry, 8 June 2013

16

ABSTRACT

The objectives of our study were to determine if the viability of freshwater mussel larvae (glochidia) is an ecologically relevant endpoint for toxicity tests and to define the appropriate duration of those tests. We assessed 1) how viability (the closing response to sodium chloride) compares to infectivity (ability to attach to a host fish and successfully metamorphose to the juvenile stage), and 2) the decline of viability and infectivity over time after larvae were released from the female. Glochidia of seven mussel species were isolated from females, placed in water, and subsampled daily for 2-5 d. Viability, when ≥

90%, was generally a good predictor of infectivity; however, when viability was < 90%, infectivity was often disproportionately low, especially for glochidia collected near the end of the brooding period. Viability and infectivity declined more rapidly in natural water and sediment compared to reconstituted water. Following 24-h exposure to a toxicant (sodium chloride or copper), infectivity of the surviving viable glochidia did not differ among concentrations of toxicants. Our results indicate that the viability endpoint is a valid proxy for infectivity and is an ecologically relevant endpoint for standard toxicity tests with freshwater mussels for any test duration with control viability > 90%.

17

INTRODUCTION

Freshwater mussels (Order Unionoida) comprise six families and more than 850 species worldwide [1]. North America is an epicenter of unionid diversity with approximately 300 species currently recognized [2]. The study of freshwater mussels has increased markedly in recent years, spurred by the recognition that many mussel taxa have become extinct and many more are threatened by anthropogenic perturbations

[3,4,5]. The extinction or extirpation of species can have cascading, ecosystem-level effects because freshwater mussels provide essential ecosystem services such as water filtration, nutrient sequestration and cycling, and habitat for other aquatic organisms

[6,7,8].

The life cycle of freshwater mussels includes a parasitic stage in which the larvae

(glochidia) must attach to a host (usually a fish) to transform into the juvenile stage. To infect the appropriate host fish species, adult female mussels use a variety of strategies, including mantle lures, conglutinates, and broadcast of free glochidia into the water

[9,10]. In each case, the glochidia must be released from the female marsupial gills and from egg or conglutinate membranes before they are able to contact the host and attach.

Glochidial attachment occurs by clamping of the valves onto the host gills or skin.

Depending on the host infection strategy, glochidia may contact the host almost immediately after entering the water or they may remain free in the water for varying lengths of time before encountering a host.

Lefevre and Curtis [11] reported that glochidia closed in response to various salt solutions or to blood, and Coker et al. [12] recommended that the viability of glochidia could be assessed by observing the closing response. This approach remains popular

18

today for assessment of glochidia condition before propagation or host testing. Glochidia that are: a) initially open and b) able to close in response to NaCl are considered to be

“viable” and assumed to be able to infect a host fish [13]. The closing response has also been widely used as an endpoint in toxicological studies [14,15,16,17]. In this context, glochidia may either close in response to the toxicant or become unable to close in response to salt. Either condition prevents completion of the life cycle.

Few quantitative studies have investigated the correlation of the ability of glochidia to close in response to salt (viability) with the ability to attach to a host fish and metamorphose to the juvenile life stage (infectivity). Zimmerman and Neves [18] assessed effects of different holding temperatures (0°, 10°, 25°C) on glochidia viability and inoculated fish with glochidia from the 0° and 10°C treatments at 7 and 14 d post extraction from the females. Metamorphosis to the juvenile stage occurred, but the metamorphosis success (proportion of glochidia that metamorphosed to the juvenile stage) was not reported. Fisher and Dimock [19] held glochidia of Utterbackia imbecillis in reconstituted water and found that the ability of the glochidia to metamorphose in cell culture media declined more rapidly than the closing response to potassium chloride. To date, it appears that no other studies have quantitatively investigated the relationship between viability and metamorphosis success.

With the increasing interest in use of glochidia for toxicity tests and derivation of water quality criteria, it is necessary to assess whether the closing response is indicative of the more ecologically relevant ability of glochidia to successfully attach and metamorphose on a host fish. A related question is whether the duration of glochidia infectivity in laboratory conditions is similar to that in natural conditions. For example,

19

survival might be shortened by microorganisms present in natural water and sediment, but absent from reconstituted water used in laboratory tests. Survival might be lengthened in species whose glochidia are normally protected by conglutinates. Lastly, there is a need to evaluate the effect on metamorphosis of glochidial exposure to toxicants. Such data would allow scientists to assess if the decrease in reproductive potential is attributable to a smaller number of glochidia that are open and able to attach to a host and/or a decrease in metamorphosis success.

Our objectives for this study were to determine: 1) if the duration of glochidia viability (determined by a shell closing response to NaCl) is equivalent to the duration of infectivity--the ability to attach to a host fish and metamorphose successfully into the juvenile stage, 2) the protective effect of conglutinates, 3) the effect of natural water and sediment on glochidia viability and infectivity, and 4) the effect of exposure to a toxicant on metamorphosis success. These findings will provide insight into the ecological relevance of laboratory tests of glochidia viability, infectivity, and the appropriate duration for standard toxicity tests with glochidia.

METHODS

Test organisms

We used snorkeling, scuba diving, and tactile searches to collect gravid female mussels from four different states over the course of three years (Table 2.1). These seven mussel species use the three main infection strategies found in Unionidae: mantle lures

(Lampsilis siliquoidea, Villosa lienosa, Lampsilis cardium, Lampsilis dolabraeformis),

20

conglutinates (Ptychobranchus occidentalis), and glochidia broadcasters (Amblema plicata, Utterbackia imbecillis).

Duration of viability vs. duration of infectivity

We extracted glochidia from 3-8 female mussels per species by gently opening their valves and using a syringe and a gentle stream of water to flush the glochidia from the marsupial gills. Viability of glochidia from individual females was measured by uniformly suspending the glochidia in a known volume of water, removing ten 200-µl subsamples, and placing the subsamples as individual drops on a 10-cm Petri dish. Each subsample contained ~15 to 30 glochidia, for a total of approximately 150 to 300 glochidia tested for viability. A stereomicroscope at 10x magnification was used to count the open and closed glochidia in each of the ten subsamples. A drop of saturated NaCl solution was then added to each 200-µl subsample, and the number of glochidia that remained open 1 min after the addition of the saline solution were counted. Glochidia that were open initially and closed in response to NaCl were considered viable, those that were closed initially were considered functionally dead because they no longer had the ability to attach to a host fish, and those that remained open after the addition of NaCl were considered definitively dead. The viability of a batch of glochidia was calculated by subtracting the number of glochidia that did not close after the addition of NaCl from the number that were originally open, then dividing by the total number of glochidia

(initially open and closed).

Glochidia from female mussels (n= 3-8 with initial viability >90%) were pooled and randomly divided into replicate (n= 5) glass beakers with 2 L of aerated, moderately

21

hard reconstituted water [20] where they were maintained at 20°C with a 16L: 8D photoperiod. Water-chemistry variables were measured daily and ranged from 6.9 to 8.8 mg/L dissolved oxygen, pH: 6.9 to7.8, hardness: 75 to 85 mg CaCO3/L, and alkalinity: 55 to 65 mg CaCO3/L. At the beginning of the experiment and every 24 h thereafter, the viability of each replicate was calculated following the method described above.

Subsampling continued daily until mean viability for the five replicates was <10%. At each time point, a second subsample of glochidia from each replicate was tested for infectivity on known host fish (two fish per replicate) for a total of 10 fish per time point.

Host fish, largemouth bass (Micropterus salmoides, 8-18 cm) and rainbow darters

(Etheostoma caeruleum, 5-8 cm) were obtained from hatcheries, which ensured that the fish had not been previously exposed to glochidial infections. Host fish were exposed to

4000 viable glochidia per liter of water for 15 min; glochidia were kept in suspension with a large-bulb pipette and vigorous aeration.

Attachment and metamorphosis success were assessed in multi-unit recirculating aquarium systems (Aquatic Habitats, Apopka, Florida) as described by Dodd et al. [21].

Fish infested with glochidia were rinsed and placed individually into unit tanks (1-3 L) fitted with 153-μm filters to collect all sloughed glochidia and metamorphosed juveniles.

Juveniles were distinguished from glochidia by possessing a foot and valve movement.

Contents of the filters were examined under a stereomicroscope every 1-2 d for a minimum of 14 d and until no glochidia or juveniles were collected for at least three consecutive days. Percent metamorphosis was calculated by dividing the total number of juveniles by the total number of glochidia and juveniles recovered from each fish. The aquarium systems were maintained at 20-22°C. All water-chemistry variables were

22

measured daily throughout the study and ranged from 7.5 to 8.2 mg/L dissolved oxygen, pH: 6.8 to7.6, and total ammonia nitrogen <0.1 mg/L. All experimental activities were approved by the University of Georgia Institution Animal Care and Use Committee.

Effects of conglutinates and sediment

The durations of glochidia viability and infectivity in conglutinates and free in water were compared in P. occidentalis. These durations in reconstituted water and in natural water and sediment were also compared. Hosts were rainbow darter (Etheostoma caeruleum, 5-8 cm total length). Conglutinates were dissected from the marsupial gills

(demibranchs) of gravid female mussels using No. 0 insect pins. Conglutinates from six females were pooled, and each conglutinate was examined carefully under magnification to determine if it was intact. Only unbroken conglutinates were used for conglutinate exposures. Free glochidia were obtained by using pins to gently tear open the outer conglutinate membrane. The ruptured conglutinate was then gently drawn in and out of a

3-ml transfer pipette to dislodge the individual glochidia from the membrane.

Both free glochidia and intact conglutinates were exposed to reconstituted hard water or to river water with fine sediment. Reconstituted water was similar in formula to that of Smith et al. [22], but concentrations of all solutes were doubled. River water and wet sediment were collected from the James River (Greene County, Missouri).

Approximately 15 L of water and 250-ml of sediment were combined and filtered to 32

µm with Nitex filter cloth. Water and sediment were homogenized by stirring with a paint stirrer while being portioned in 200-ml aliquots to each of the test chambers (250- ml beakers). A layer of sediment about 1 mm deep settled in each beaker. Treatments

23

(three replicates per treatment x exposure time; 60 beakers total) were: 1) reconstituted water with free glochidia, 2) reconstituted water with conglutinates, 3) river water and sediment with free glochidia, and 4) river water and sediment with conglutinates. Water- chemistry variables were measured daily for both the reconstituted water and the river water. The pH averaged 7.9 for both water sources, hardness averaged 187 mg CaCO3/L for reconstituted water and 148 mg CaCO3/L for river water, temperature averaged

21.3°C for both water sources, and conductivity averaged 362 µS/cm for reconstituted water and 218 µS/cm for river water. Each beaker received 1000 glochidia or 3 conglutinates containing a similar total number of glochidia. At time zero, and then each

24-h interval for 96 h, viability and infectivity were assessed. In conglutinate treatments, the conglutinates were opened and the glochidia freed. Two darters were placed in each beaker for 15 min, while gently stirring the water with a pipette, to achieve infestation.

Methods for calculating viability and infectivity were similar to those previously described.

Effects of glochidial exposure to toxicants

We extracted glochidia from 3-8 female mussels per species by gently opening their valves and using a syringe and a gentle stream of water to flush the glochidia from the marsupial gills. Viability of glochidia from individual females was measured in the same manner as that previously described. According to the American Society for

Testing and Materials (ASTM) guidelines [23], glochidia from female mussels (n= 3-8 with initial viability >90%) were pooled and randomly divided into six concentrations of

NaCl (0, 0.5, 1.0, 1.5, 2.0, 2.5 ppt) or four concentrations of copper (0, 15, 30, 60 ppm)

24

dissolved in moderately hard reconstituted water [20]. Glochidia from each concentration were then split into replicate (n= 3) glass beakers where they were maintained at 20°C with a 16L: 8D photoperiod. After 24 h of exposure to the respective toxicants, the viability of each replicate was calculated following the method described previously. A second subsample of glochidia from each replicate was tested for infectivity on largemouth bass (two fish per replicate) for a total of six fish per concentration. The inoculation procedure follows the methods previously above. In each replicate, the inoculation suspension was adjusted to provide 4000 viable glochidia per liter. The aquarium systems were maintained at 18-22°C. All water-chemistry variables were measured daily throughout the study and ranged from 7.5 to 8.6 mg/L dissolved oxygen, pH: 6.8 to7.7, and total ammonia nitrogen <0.1 mg/L. Salinity values were verified daily in one replicate of each treatment with a handheld salinity meter (YSI 30,

Yellow Springs Instruments, Yellow Springs, OH) and were within 0.1 ppt of nominal

NaCl concentrations at all times. Water samples were collected and preserved by adjusting to pH <2 with HCl prior to Cu analysis by inductively coupled plasma mass spectrometry. All measured Cu concentrations (n=10) were 80-120% of nominal and duplicate samples (n=4) varied by <5%. All statistical analyses were based on nominal concentrations.

Statistical analysis

Data were tested for normality using the Kolmogorov-Smirnov test (SAS, version

9.3; SAS Institute, Cary, North Carolina). To achieve normality and equal variance, viability and metamorphosis data were transformed with the arcsine of the square root.

25

Transformed data were analyzed with a one-way ANOVA, followed by a post-hoc

Dunnett’s test to assess for differences between the control (i.e., 0 h in the aging exposures) and each time point or toxicant concentration. Attachment data (the number of glochidia attached per fish) were analyzed with a one-way ANOVA, followed by a post-hoc Tukey’s test. Level of significance for all tests was set at α = 0.05. A 24-h

EC50 was calculated for each 24-h glochidia viability test with the Trimmed Spearman-

Karber Method (ToxStat, WEST, Inc.). We considered EC50s significantly different based on non-overlapping 95% confidence intervals [15].

RESULTS

Duration of viability vs. duration of infectivity

The initial (0 h) glochidia viability met the current ASTM guideline of >90% [23] for all seven mussel species tested; however, the trial with L. siliquoidea collected in June had an initial viability of only 80%. Viability and metamorphosis remained higher for longer with L. siliquoidea glochidia collected in January (Fig. 2.1A) compared to those collected in June (Fig. 2.1B), which is at the end of the brooding season for many members of the tribe Lampsilini. At the 24-h time point, viability was <90% for L. cardium, U. imbecillis, and L. siliquoidea (June) (Fig. 2.1). According to the ASTM guideline for conducting laboratory toxicity tests with freshwater mussels [23], data from these time points would not be valid for toxicological testing purposes because the viability was <90%. Five species (L. siliquoidea (Jan), L. dolabraeformis, V. lienosa, A. plicata, and P. occidentalis) had >90% viability at 24 h (Fig. 2.1). For these species, metamorphosis success did not differ between 0 and 24 h (Fig. 2.1, Table 2.2). The

26

exception was P. occidentalis free glochidia, which had viability >90% through 72 h but had low (<5%) metamorphosis by 24 h (Fig. 2.2A).

For all but one species (L. dolabraeformis), metamorphosis success decreased rapidly after 24 h though viability remained near or above 60% for 3 d with four of the seven species. Overall, metamorphosis success ranged from 19 to 78% (mean 54%) at time zero for the species tested, typical of rates achieved in mussel propagation trials with known host fish-mussel pairs, at common temperatures, and with naïve test fish [21].

Total glochidia attachment did not differ among treatments for any of the mantle lure or conglutinate species tested (p = 0.08-0.64), but the two broadcasting species (U. imbecillis and A. plicata) had significantly greater glochidia attachment (p = <0.001) at the final inoculation time point (mean attachment ± 95% CI: U. imbecillis: 0 to 24 h =

353 ± 61, 48 h = 1306 ± 208; A. plicata: 0 to 48 h = 365 ± 44, 72 h = 745 ± 88).

Effects of conglutinates and sediment

Initial (0 h) glochidia viability of P. occidentalis was > 90% in all treatments. In reconstituted water, the viability of free glochidia remained >90% until 72 h. Viability of glochidia in conglutinates declined more slowly and was >90% for at least 96 h, at which the time at which the experiment was terminated (Fig. 2.2A, 2.2B). In unsterilized river water and sediment, viability of free glochidia decreased to 52% viable by 48 h, and viability of glochidia in conglutinates reached <90% at 72 h (Fig. 2.2C and 2.2D).

Metamorphosis success of free glochidia of P. occidentalis was high (45% ± 13) for glochidia placed on fish immediately after removal from the mussel, but decreased rapidly in both river water and reconstituted water to <5% for free glochidia aged 24 h or

27

more, even though viability remained >90% (Fig. 2.2, Table 2.2). In contrast, glochidia in conglutinates remained infective for at least 96 h. Total glochidia attachment did not differ among P. occidentalis conglutinate treatments nor free glochidia in reconstituted water (mean ± 95% CI = 38 ± 6; p = 0.11-0.61), but free glochidia exposed to river water and sediment had significantly lower attachment (p = 0.0113) at each time point after the initial inoculation (0 h = 45 ± 20; 24-96 h = 12 ± 3).

Effects of glochidial exposure to toxicants

In each toxicant test, viability decreased (ranged from 95% to 25%) as toxicant concentration increased (Fig. 2.3). However, for the remaining viable glochidia in each treatment, metamorphosis success was similar among treatments regardless of viability

(Table 2.3). Lampsilis cardium glochidia collected in December (early in the brooding period; Fig. 2.3A) maintained high viability longer than those of the same species collected in June (late in the brooding period; Fig. 2.3B) and also had consistently higher

(55-70%) metamorphosis rates than those collected in June (18-23%). Glochidia collected early in the brooding period (December) were significantly less sensitive (based on non-overlapping 95% confidence intervals) to NaCl than those collected late in the brooding period (June). The 24-h EC50 (± 95% CI) for December glochidia was 2.04 g/L ± 0.11 and for June glochidia was 0.80 g/L ± 0.16 (Fig. 2.4). Total glochidia attachment (mean ± 95% CI) did not differ among L. cardium or L. dolabraeformis trials

(p = 0.16-0.58).

28

DISCUSSION

Our results indicate that glochidia viability (as measured by the shell closing response to NaCl exposure) is indicative of the ability to metamorphose into the juvenile life stage, as long as control group viability was >90%. Age of glochidia also appears to be critical to glochidia health; viability of older glochidia (those from females collected late in the brooding season or from females held longer in the laboratory) declined more rapidly after removal from the female. The current ASTM guideline for conducting toxicity tests with glochidia of freshwater mussels [23] requires >90% viability for control groups at test termination (commonly 24 h) and our data support the continued use of this criterion. Additionally, glochidia that remained viable following exposure to a toxicant for 24 h were able to successfully metamorphose, thus the viability endpoint appears to have direct ecological relevance. Based on these data, we recommend that the current ASTM guideline for glochidia toxicity testing be retained for all mussel species.

We also recommend that glochidia from long term brooders to be used in toxicity tests should be collected well before the end of the brooding season because those collected late in the season appeared less healthy; our results that showed glochidia collected late in the brooding season were more sensitive to toxicants, had lower initial viability which declined quickly, and had poorer metamorphosis success. For species with previously defined duration of viability of ≥ 24 h, we recommend a 24 h toxicity test; however, for species with little or no data on duration of viability, we recommend that viability be assessed at an intermediate time point (e.g., 6, 12, 18h) during toxicity tests in the event that control viability is <90% by 24 h. Shipping time must also be considered, if glochidia cannot be obtained locally.

29

The duration of viability for free glochidia of P. occidentalis was shorter in river water and sediment than in reconstituted water, but the effect was much less pronounced for glochidia in conglutinates. We suspect that the relatively rapid decline of free glochidia in river water with sediment was the result of microorganisms that compromised glochidia survival and ability to attach to a host fish. In a preliminary test of this hypothesis, a group of free glochidia was tested in 0.4 µm filter-sterilized river water and they remained >80% viable at 72 h. This result indicated that microorganism removal positively affected viability. The subsequent infectivity of these glochidia was not tested, however. Whereas infectivity of free glochidia of P. occidentalis decreased rapidly in both treatments, infectivity of glochidia in conglutinates remained high for the duration of the study (96 h). The apparent protective effects of the conglutinate suggest that a 24-h toxicity test duration would be conservative for species that release glochidia in conglutinates. Conglutinates provide protection from some toxicants (e.g., ammonia, copper; M.J. Pillow, 2009, Master’s thesis, Missouri State University, Springfield, MO,

U.S.A), [24]) and thereby reduce toxicity (increase LC50, EC50 values).

Our results suggest that the conglutinate membranes of P. occidentalis provide protection from the external environment for at least 4 d. Even greater protection may be afforded by the marsupial gill of the brooding female, where glochidia of long-term brooders develop and survive for many months [10]. Once released from the protection of the conglutinate or the marsupia of the female mussel, the infectivity of free glochidia appears to decrease after 24 h, depending on the species. Our data indicate that metamorphosis success decreases before viability decreases; at 48 h post extraction, L. siliquoidea (January collection) and V. lienosa exhibited 80% viability, but the

30

corresponding metamorphosis success was reduced by 50% (Fig. 2.1A & 2.1F).

Importantly, the reduced infectivity was not a result of lower attachment to host fish, which remained consistent across inoculation days.

The copper and NaCl toxicant exposure results suggest that the decrease in glochidia reproductive potential after exposure to a toxicant is attributed to a smaller number of glochidia that are open and able to attach to a host, rather than a decrease in metamorphosis success. Because the inoculation suspensions were adjusted to a uniform number of viable glochidia, the consistent metamorphosis success supports the use of the viability test as a proxy for metamorphosis. These results were consistent with two mussel species and two toxicants with different modes of action.

This study provides insight into factors that reduce glochidia success. Both aging and water exposure outside of the female marsupia might reduce glochidia energy reserves, from maintenance metabolism during aging and from the increased metabolic demand of osmoregulation during water exposure. When the energy reserves were intact, metamorphosis success of attached glochidia was high, but as those reserves were depleted the ability to metamorphose was reduced. The experiments in which the glochidia were exposed to toxicants also support this hypothesis, indicating that reduced viability does not mean reduced infectivity of the glochidia that are still open and able to attach to a host fish.

For mussel species that are long-term brooders, the length of time that glochidia have aged within the female mussel also may affect glochidia function and condition.

Our data suggest that older glochidia may be in poorer physiological condition (e.g., less energy reserves) and are unable to remain viable for as long after the glochidia are

31

removed from the female. In our study, L. siliquoidea glochidia early in the brooding period (January) registered 80% viability 48-h after being removed from the female, whereas glochidia from the same species collected late in the brooding period (June) only had 20% viability at 48 h post extraction. Additionally, Lampsilis cardium glochidia tested late in the brooding period (June) also showed a rapid decline in viability compared to L. cardium tested early in the brooding period (December). Dodd et al. [25] reported that Lampsilis reeveiana glochidia aged 241-360 d in marsupia had significantly less metamorphosis success (67.5%) compared to glochidia aged 0-120 d (87.5%) and

121-240 d, (88.7%). Cope et al. [17] reported longer duration of viability (>90%) for some of the species used in the present study; however, timing of collection of the glochidia was not reported.

The implications for the results of this study are multifaceted. For the purposes of toxicological testing, our recommendation is to retain the current ASTM guideline for toxicity testing with glochidia and the use of the valve closure response to NaCl as an endpoint. Ideally, toxicity tests should be conducted with glochidia immediately following removal from the female mussel, thus shipping of brooding females rather than free glochidia may be considered. However, shipping of adult mussels is less practical and may create logistical and regulatory problems (e.g., interbasin transfer of species or threatened or endangered status), so shipping of free glochidia may be necessary. Our results indicate that this time frame may be suitable (i.e., viability and infectivity can remain >90% for 48 h), but we recommend that viability be assessed at one or more time points between 0 and 24 h (e.g., 6 or 12 h) in the event that viability of controls decreases to <90% by the 24 h time point of the toxicity test. This approach is especially important

32

when the duration of glochidia viability of a species is unknown. These recommendations are supported by Cope et al. [17], who showed that viability

(determined with the NaCl method) of glochidia from 20 species of mussels, three of which were tested in this study (L. cardium, L. siliquoidea, U. imbecillis), was suitable

(>90%) for 24 h or less. This time period corresponds with the ecologically relevant endpoint of infectivity determined in this study. The higher toxicant sensitivity, lower initial viability and lower metamorphosis success of glochidia collected late in the brooding season suggests they are not appropriate for use in toxicity tests. Use of glochidia in toxicity tests that are harvested from females late in the brooding period (of long-term brooders) may result in test outcomes that are substantially different than if glochidia were collected early in the brooding season. More research is required to better define the period when glochidia fitness begins to decline to the point that toxicant sensitivity is increased. Furthermore, implications for freshwater mussel host suitability research include the possibility of obtaining a false-negative result, if the glochidia have been outside of the female for more than 24 h. The decrease in metamorphosis success should be noted by those who conduct propagation efforts with imperiled mussels, and every effort should be made to place glochidia onto host fish as soon as practical after the glochidia are removed from the female mussel. Future propagation and culture research evaluating the infectivity of the glochidia life stage of freshwater mussels would benefit from accurate calculation and reporting of viability data.

In summary, we found that glochidia viability of >90% was indicative of infectivity (ability to metamorphose); however, after 24 h, infectivity often declined more rapidly than viability. Glochidia of long-term brooders collected late in the brooding

33

season lost infectivity more rapidly compared to those collected early in the brooding season. Glochidia in reconstituted water may exhibit a more prolonged duration of viability and infectivity than those exposed to the microbial community of a natural environment. Glochidia in conglutinates are protected and may remain viable and infective for more extended periods than free glochidia. Glochidia viability is an ecologically relevant endpoint in standard toxicity tests provided that control viability is high, and the current ASTM criteria for control animal viability (>90%) is therefore appropriate.

ACKNOWLEDGEMENTS

Funding was provided by the U.S. Environmental Protection Agency Region 5 Great

Lakes Restoration Initiative Program through Work Order no. 104 from the U.S.

Geological Survey. We thank C. Delos, L. Holst, L. Huff, D. Mount, C. Stephan, and B.

Thompson of the U.S. EPA for their valuable insight and assistance on project need and experimental design. We thank J. Creamer, M. Fritts, R. Ratajczak, A. Tuck, and J.

Wisniewski for assistance in the laboratory and field, and B. Sietman and T. Newton for providing mussels. Largemouth bass were donated by Owens and Williams Fish

Hatchery (Hawkinsville, Georgia) and rainbow darters were provided by Conservation

Fisheries (Knoxville, Tennessee). Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government. The views expressed in this article do not necessarily represent the views of the U.S. EPA.

34

REFERENCES

1. Graf DL, Cummings KS. 2007. Review of the systematics and global diversity of

freshwater mussel species (Bivalvia : Unionoida). Journal of Molluscan Studies

73:291-314.

2. Parmalee PW, Bogan AE. 1998. The Freshwater Mussels of Tennessee.

University of Tennessee Press, Knoxville, TN, USA.

3. Ricciardi A, Rasmussen JB. 1999. Extinction rates of North American freshwater

fauna. Conservation Biology 13:1220-1222.

4. Lydeard C, Cowie RH, Ponder WF, Bogan AE, Bouchet P, Clark SA, Cummings

KS, Frest TJ, Gargominy O, Herbert DG, Hershler R, Perez KE, Roth B, Seddon

M, Strong EE, Thompson FG. 2004. The global decline of nonmarine mollusks.

Bioscience 54:321-330.

5. Strayer DL, Downing JA, Haag WR, King TL, Layzer JB, Newton TJ, Nichols

SJ. 2004. Changing perspectives on pearly mussels, North America's most

imperiled animals. Bioscience 54:429-439.

6. Vaughn CC, Spooner DE. 2006. Unionid mussels influence macroinvertebrate

assemblage structure in streams. Journal of the North American Benthological

Society 25:691–700.

7. Strayer DL. 2008. Freshwater Mussel Ecology: A Multifactor Approach to

Distribution and Abundance. University of California Press, Berkeley, CA, USA.

8. Vaughn CC, Nichols SJ, Spooner DE. 2008. Community and foodweb ecology of

freshwater mussels. Journal of North American Benthological Society 27:409–

423.

35

9. Barnhart MC, Haag WR, Roston WN. 2008. Adaptations to host infection and

larval parasitism in Unionoida. Journal of the North American Benthological

Society 27:370-394.

10. Haag WR. 2012. North American Freshwater Mussels: Natural History, Ecology,

and Conservation. Cambridge University Press, Cambridge, United Kingdom.

11. Lefevre G, Curtis WC. 1912. Studies on the reproductive and artificial

propagation of fresh-water mussels. Bulletin of the Bureau of Fisheries 30:105-

201+12 plates.

12. Coker RE, Shira AF, Clark HW, Howard AD. 1921. Natural history and

propagation of freshwater mussels. Bulletin of the U.S. Bureau of Fisheries

37:75–182.

13. Ingersoll CG, Kernaghan NJ, Gross TS, Bishop CD, Wang N, Roberts A. 2007.

Laboratory toxicity testing with freshwater mussels. In Farris JL, Van Hassel JH,

eds, Freshwater Bivalve Ecotoxicology. CRC Press, Boca Raton, FL, USA, and

SETAC Press, Pensacola, FL, USA, pp 95-134.

14. Keller AE, Ruesller DS. 1997. The toxicity of malathion to unionid mussels:

relationship to expected environmental concentrations. Environmental Toxicology

and Chemisty 16(5):1028-1033.

15. Bringolf RB, Cope WG, Barnhart MC, Mosher S, Lazaro PR, Shea D. 2007.

Acute and chronic toxicity of pesticide formulations (atrazine, chlorpyrifos and

permethrin) to glochidia and juveniles of Lampsilis siliquoidea (Unionidae).

Environmental Toxicology and Chemistry 26:2101–2107.

36

16. Wang N, Ingersoll CG, Hardesty DK, Ivey CD, Kunz JL, May TW, Dwyer FJ,

Roberts AD, Augspurger T, Kane CM, Neves RJ, Barnhart MC. 2007. Acute

toxicity of copper, ammonia, and chlorine to glochidia and juveniles of freshwater

mussels (Unionidae). Environmental Toxicology and Chemistry 26:2036–2047.

17. Cope WG, Bringolf RB, Buchwalter DB, Newton TJ, Ingersoll CG, Wang N,

Augspurger T, Dwyer FJ, Barnhart MC, Neves RJ, Hammer E. 2008. Differential

exposure, duration, and sensitivity of unionoidean bivalve life stages to

environmental contaminants. Journal of the North American Benthological

Society 27:451-462.

18. Zimmerman LL, Neves RJ. 2002. Effects of temperature on duration of viability

for glochidia of freshwater mussels (Bivalvia: Unionidae). American

Malacological Bulletin 17:31-35.

19. Fisher GR, Dimock RV. 2000. Viability of glochidia of Utterbackia imbecillis

(Bivalvia: Unionidae) following their removal from the parent mussel.

Proceedings, Freshwater Mussel Conservation Society Symposium, Chattanooga,

TN, USA, March 18, 1999, pp 185-188.

20. American Society for Testing and Materials. 2006a. Standard guide for

conducting acute toxicity tests on test materials with fishes, macroinvertebrates,

and amphibians. E729-96. In Annual Book of ASTM Standards, Vol 11.06.

Philadelphia, PA, USA, pp 79–100.

21. Dodd BJ, Barnhart MC, Rogers-Lowery CL, Fobian TB, Dimock RV. 2005.

Cross resistance of largemouth bass to glochidia of unionid mussels. Journal of

Parasitology 91:1064-1072.

37

22. Smith ME, Lazorchak JM, Herrin LE, Brewer-Swartz S, Thoeny WT. 1997. A

reformulated, reconstituted water for testing the freshwater amphipod, Hyalella

azteca. Environmental Toxicology and Chemistry 16(6):1229-1233.

23. American Society for Testing and Materials. 2006b. Standard guide for

conducting laboratory toxicity tests with freshwater mussels E2455-06. In Annual

Book of ASTM Standards, Vol 11.06. Philadelphia, PA, USA, pp 1393–1444.

24. Gillis PL, Mitchell RJ, Schwalb AN, McNichols KA, Mackie GL, Wood CM,

Ackerman JD. 2008. Sensitivity of the glochidia (larvae) of freshwater mussels to

copper: assessing the effect of water hardness and dissolved organic carbon on the

sensitivity of endangered species. Aquatic Toxicology 88(2):137-145.

25. Dodd BJ, Barnhart MC, Rogers-Lowery CL, Fobian TB, Dimock RV. 2006.

Persistence of host response to glochidia larvae in Micropterous salmoides. Fish

and Shellfish Immunology 21:473–484.

38

Table 2.1. Dates and locations of freshwater mussel collections.

Species Water body County Date

Amblema plicata Sac River Cedar, MO Jul 2011

Cooleewahee Villosa lienosa Creek Baker, GA Oct 2010

Lampsilis cardium Mississippi River La Crosse, WI Jun 2011

Silver Fork of Lampsilis siliquoidea Perche Creek Boone, MO Jan 2011

Lampsilis siliquoidea Red Lake River Pennington, MN Jun 2012

Lampsilis dolabraeformis Oconee River Montgomery, GA Oct 2012

Utterbackia imbecillis Lake Oconee Morgan, GA Jul 2011

Ptychobranchus occidentalis St. Francis River Wayne, MO Feb 2011

39

Table 2.2. Results of one-way ANOVA comparing the change in glochidia viability or metamorphosis over time for seven freshwater mussel species. Degrees of freedom (df), test statistic (F value) and P values are presented. Level of significance was α = 0.05.

Species df F value P value

Viability

Amblema plicata 4, 20 45.23 <0.0001

Villosa lienosa 5, 24 59.83 <0.0001

Lampsilis cardium 3, 16 211.85 <0.0001

Lampsilis siliquoidea, January 5, 24 47.13 <0.0001

Lampsilis siliquoidea, June 2, 12 31.69 0.0005

Lampsilis dolabraeformis 4, 20 75.39 <0.0001

Utterbackia imbecillis 3, 16 286.73 <0.0001

Ptychobranchus occidentalis,

glochidia/water 4, 10 13.18 0.0005

Ptychobranchus occidentalis,

conglutinate/water 4, 10 0.83 0.5372

Ptychobranchus occidentalis,

glochidia/sediment 4, 10 35.17 <0.0001

Ptychobranchus occidentalis,

conglutinate/sediment 4, 10 4.00 0.0344

Metamorphosis

Amblema plicata 3, 16 7.75 0.0020

40

Villosa lienosa 3, 16 9.25 0.0009

Lampsilis cardium 2, 12 49.65 <0.0001

Lampsilis siliquoidea, January 4, 20 86.85 <0.0001

Lampsilis siliquoidea, June 1, 8 2.68 0.1404

Lampsilis dolabraeformis 3, 16 20.41 <0.0001

Utterbackia imbecillis 2, 12 105.09 <0.0001

Ptychobranchus occidentalis, glochidia/water 4, 10 32.69 <0.0001

Ptychobranchus occidentalis, conglutinate/water 4, 10 6.68 0.0070

Ptychobranchus occidentalis, glochidia/sediment 4, 10 6.77 0.0066

Ptychobranchus occidentalis, conglutinate/sediment 4, 10 1.37 0.3118

41

Table 2.3. Results of one-way ANOVA comparing the change in glochidia viability or metamorphosis for glochidia exposed to increasing concentrations of sodium chloride

(NaCl) or copper (Cu). Degrees of freedom (df), test statistic (F value) and P values are presented. Level of significance was α = 0.05.

Species df F value P value

Viability

Lampsilis cardium (Dec), NaCl 5, 12 45.65 <0.0001

Lampsilis cardium (Jun), NaCl 3, 8 165.67 <0.0001

Lampsilis dolabraeformis, NaCl 5, 12 92.31 <0.0001

Lampsilis dolabraeformis, Cu 3, 8 188.06 <0.0001

Metamorphosis

Lampsilis cardium (Dec), NaCl 5, 12 0.82 0.5595

Lampsilis cardium (Jun), NaCl 3, 8 2.67 0.1187

Lampsilis dolabraeformis, NaCl 5, 12 10.21 0.0005

Lampsilis dolabraeformis, Cu 3, 8 3.13 0.0875

42

Fig. 2.1. Comparison of percent viability (bars) and percent metamorphosis success

(lines; ± 95% CI) for glochidia following release from female mussels of A) Lampsilis siliquoidea (Jan), B) Lampsilis siliquoidea (Jun), C) Lampsilis cardium, D) Lampsilis dolabraeformis, E) Amblema plicata, F) Villosa lienosa, and* G) Utterbackia imbecillis. * Viability was determined with approximately 200 glochidia per replicate (n=5) and metamorphosis success was assessed on largemouth bass (n=10 fish, two per replicate).

Presence of a black asterisk (*) indicates a significant difference in viability, and a white asterisk indicates a significant difference in metamorphosis success.

43

44

Fig. 2.2. Comparison of percent viability (bars) and percent metamorphosis success

(lines; ± 95% CI) for glochidia of Ptychobranchus occidentalis: A) free glochidia in reconstituted water, B) glochidia in conglutinates in reconstituted water, C) free glochidia exposed to river water and sediment, and D) glochidia in conglutinates exposed to river water and sediment. Viability was determined with approximately 200 glochidia per replicate (n=3) and metamorphosis success was assessed on rainbow darter (n=6 fish, two per replicate). Presence of a black asterisk (*) indicates a significant difference in viability, and a white asterisk indicates a significant difference in metamorphosis success.

45

Fig. 2.3. Comparison of percent viability (bars) and percent metamorphosis success

(lines; ± 95% CI) for glochidia exposed to increasing concentrations of sodium chloride

(NaCl) or copper (Cu) for 24 h: A) Lampsilis cardium (Dec) NaCl exposure, B)

Lampsilis cardium (Jun) NaCl exposure, C) Lampsilis dolabraeformis NaCl exposure, and D) Lampsilis dolabraeformis exposed to Cu. Viability was determined with approximately 200 glochidia per replicate (n=3) and metamorphosis success was assessed on largemouth bass (n=6 fish, two per replicate). Presence of a black asterisk (*) indicates a significant difference in viability, and a white asterisk indicates a significant difference in metamorphosis success.

46

Fig. 2.4. Comparison of NaCl 24-h EC50 (± 95% confidence interval) for Lampsilis cardium glochidia collected early in the brooding period (December) and late in the brooding period (June). Asterisk indicates significant difference in EC50s (non- overlapping 95% confidence intervals).

47

CHAPTER 3

CRITICAL LINKAGE OF IMPERILED SPECIES:

GULF STURGEON AS HOST FOR PURPLE BANKCLIMBER MUSSELS2

2 Fritts, A.K., M.W. Fritts, D. Peterson, D. Fox, and R.B. Bringolf. 2012. Freshwater

Science. 31(4):1223-1232. Reprinted here with permission of the publisher.

48

ABSTRACT

One of the largest impediments to the conservation of freshwater mussels is the absence of host-fish data. Suitable hosts must be present in sufficient numbers and occur at the appropriate time for successful mussel recruitment. However, habitat degradation and fragmentation caused by dams and other anthropogenic alterations may reduce host availability. Host data are lacking for the federally threatened Purple Bankclimber mussel

(Elliptoideus sloatianus), which is endemic to the Apalachicola–Chattahoochee–Flint basin (ACF) in Alabama, Florida, and Georgia, and the Ochlockonee basin in Florida and

Georgia. We tested 29 fish species in 7 families as potential hosts for Purple

Bankclimbers and observed high metamorphosis success (79–89%) with 4 species of sturgeon: Gulf (Acipenser oxyrinchus desotoi), Atlantic (Acipenser oxyrinchus oxyrinchus), Lake (Acipenser fulvescens), and Shortnose (Acipenser brevirostrum).

Metamorphosis was less successful with Blackbanded Darters (Percina nigrofasciata) and Halloween Darters (Percina crypta) as hosts (34–36% metamorphosis), and the remainder of the fishes we tested were not suitable hosts. The federally threatened Gulf

Sturgeon is the only sturgeon species present in the ACF, but access of this migratory fish to most of the basin is blocked by Jim Woodruff Dam on the Apalachicola River. In the absence of sturgeon upstream of Jim Woodruff Dam, darters appear to have facilitated persistence of this mussel species, but at abundances far lower than historical conditions.

This relationship between the Purple Bankclimber and Gulf Sturgeon is the first description of a federally protected fish serving as a host for a federally protected mussel and is an archetypal example of the role of habitat fragmentation in the ecology of listed

49

species. Recovery of the Purple Bankclimber and other mussel species probably will require restoration of habitat connectivity for fish passage.

50

INTRODUCTION

Declines of freshwater communities have been attributed to a number of anthropogenic impacts, and extinction rates for the North American freshwater fauna are similar to those projected for critically imperiled tropical-rainforest communities

(Ricciardi et al. 1998, Ricciardi and Rasmussen 1999). Anthropogenic impacts to streams have led to the swift, well documented decline of many freshwater mussels

(Mollusca:Unionoida) that historically occurred at high densities throughout eastern

North America (Bogan 1993, Williams et al. 1993). Freshwater mussels provide essential ecosystem services, such as water filtration, nutrient sequestration, and nutrient cycling, and they provide habitat for other aquatic organisms. Consequently, declines of freshwater mussel populations may have cascading effects throughout aquatic communities (Strayer et al. 1994, Silverman et al. 1997, Vaughn and Spooner 2006,

Vaughn et al. 2008).

A unique and complex reproductive cycle makes mussels especially vulnerable to anthropogenic perturbation. In most species, parasitic larvae (glochidia) must attach to and become encysted on the gills or fins of fishes to complete metamorphosis to the juvenile stage. Suitable hosts vary widely among mussel species. Some mussels can metamorphose only on a single host species, whereas others can metamorphose on a variety of species (Barnhart et al. 2008). Glochidia that can metamorphose on >1 host species commonly have varying degrees of metamorphosis success on different host species. Identification of suitable hosts and the relative metamorphosis success on those species is essential for management of freshwater mussels. Resource managers must be able to determine whether suitable hosts are available in sufficient numbers at appropriate

51

times for reproduction, and to a large extent, captive propagation of mussels relies on knowledge of suitable hosts. Unfortunately, host data are incomplete or lacking for many mussels, including imperiled species.

The Apalachicola–Chattahoochee–Flint basin (ACF) in western Georgia, eastern

Alabama, and northwestern Florida is home to 32 unionid species, 8 of which are endemic to the basin and 6 of which are federally threatened or endangered (Brim Box and Williams 2000, Williams et al. 2008). Like many rivers, the ACF has been altered by numerous dams and intensive water withdrawal for municipal and agricultural use. Jim

Woodruff Lock and Dam, constructed by the US Army Corps of Engineers in 1952 on the Apalachicola River 1 km downstream of the confluence of the Flint and

Chattahoochee rivers (Fig. 3.1), blocks migration of several migratory fish species. For example, Gulf Sturgeon (Acipenser oxyrinchus desotoi) are restricted to the ~180-km reach of the Apalachicola River downstream of the dam and are unable to access 78% of their historical spawning habitat within the ACF (US Fish and Wildlife Service and Gulf

States Marine Fisheries Commission 1995). Despite ongoing recovery efforts below the dam, the ACF Gulf Sturgeon population has declined to only a few hundred individuals

(Zehfuss et al. 1999). Sturgeon and other migratory fishes (e.g., Striped Bass, herrings, salmon) provide valuable ecosystem services to freshwater systems (Limburg and

Waldman 2009, Merz and Moyle 2006). One of these services is their role as hosts for freshwater mussels. Furthermore, migratory fishes make lengthy and predictable movements that could serve as important dispersal mechanisms for mussels. The decline of migratory fishes has been implicated in the decline of a number of freshwater mussel species that specialize on migratory hosts, including the Ebonyshell (Reginaia

52

(Fusconaia) ebena), Elephant-ear (Elliptio crassidens), Black Sandshell (Ligumia recta), and Alewife Floater (Anodonta implicata) (Davenport and Warmuth 1965, Smith 1985,

Kelner and Sietman 2000, Haag 2012).

One of the federally protected mussel species in the ACF is the Purple

Bankclimber (Elliptoideus sloatianus). Purple Bankclimbers were once abundant in large streams of the ACF, but populations have declined over the past several decades and now remain above Jim Woodruff Dam only in the Flint River (Brim Box and Williams 2000,

Williams et al. 2008). The Flint River population is dominated by large individuals (~150 mm) and few juvenile mussels are found, a pattern suggesting that recruitment has been greatly reduced (J. Wisniewski, Georgia Department of Natural Resources, personal communication). Like other imperiled mussels, the decline of the Purple Bankclimber is broadly attributed to habitat fragmentation and degradation, especially dam construction and stream impoundment, but specific mechanisms causing these declines remain unknown. O’Brien and Williams (2002) identified 3 fish species that facilitated limited metamorphosis of juvenile Purple Bankclimbers: Eastern Mosquitofish (Gambusia holbrooki), Guppy (Poecilia reticulata), and Blackbanded Darter (Percina nigrofasciata).

Of these 3 species, only the Blackbanded Darter cooccurs with the Purple Bankclimber, but metamorphosis rates on this species were low, suggesting that other fish species may be more important as hosts. We examined host use of the Purple Bankclimber across a wide range of fish species, with emphasis on diadromous sturgeon, to assess the potential role of disruption of fish movement by Jim Woodruff Dam as a factor in the decline of this species.

53

METHODS

We collected adult Purple Bankclimbers from the Flint River in March 2010,

2011, and 2012 (Fig. 3.1). We used an oyster knife to open each mussel’s valves enough to observe the gills and identified gravid females as those individuals with swollen demibranchs. We placed gravid mussels (n = 5/y) in river water in aerated coolers, transported them to the Aquatic Science Laboratory (ASL) at the University of Georgia, and held them there until they released glochidia (<4 d). We returned adult mussels to their collection site within 21 d of glochidia release.

We conducted host-suitability trials using previously described methods (Neves et al. 1985) during March–May 2010, 2011, and 2012 at the ASL. We collected 29 fish species representing 7 families for host trials (Table 3.1). We collected most fishes from rivers throughout Georgia with a nylon seine and backpack electrofisher, but we obtained some species from federal, state, or private hatcheries. We captured wild juvenile

Atlantic Sturgeon (Acipenser oxyrinchus oxyrinchus) from the Altamaha River, Georgia, during May 2010 using 91-m-long anchored monofilament trammel nets with a 7.6-cm mesh (stretch measure) inner panel and 2 30.5-cm-mesh outer panels. We captured wild juvenile Gulf Sturgeon from the Choctawhatchee River, Florida, on 8 November 2011 using 60-m-long anchored monofilament gill nets with a single 10.2-cm-mesh panel. To avoid complications regarding acquired immunity of host fishes (Dodd et al. 2006), we focused collections of wild fish on rivers and streams where mussels were not present.

Regardless of source, we transported all fish in insulated coolers or hauling tanks with aerators to the laboratory where we held them in dechlorinated tap water until host trials.

54

Before beginning host trials, we tested a subsample of glochidia from each female mussel for viability with sodium chloride (NaCl) to stimulate valve closure (Zale and

Neves 1982). If >20% of the glochidia were unresponsive to NaCl or appeared to be weak (i.e., slow response), glochidia from that female were not used in trials. Only 7 of the 15 gravid female Purple Bankclimbers collected over the course of the study contained mature, healthy glochidia. We pooled viable glochidia from different females and counted them. We suspended the glochidia in a known volume of water and used a stereomicroscope to count the glochidia in ten 200-µL subsamples. We estimated the total number of viable glochidia by extrapolating from the subsample.

We inoculated potential fish hosts by immersion in a glochidia suspension of

4000 glochidia/L water. We placed the fish in the inoculation bath for 15 min and kept the glochidia in suspension with a large rubber-bulb pipette and vigorous aeration. After inoculation, we removed the fish from the solution, rinsed them with fresh water to remove unattached glochidia, and placed them in individual tanks for monitoring. We housed most fish in a modified recirculating aquaculture system (AHAB®; Aquatic

Habitats Inc., Apopka, Florida), with the outflow from each tank equipped with a 100-

µm-mesh filter cup to recover any glochidia or juveniles released from the fish. Atlantic,

Lake, and Shortnose sturgeons were too large to be housed in the AHAB unit, and we housed them individually in 80- or 100-L aquaria fitted with 2-mm-mesh false bottoms to prevent the consumption of glochidia or juveniles by the test fish. We housed Gulf

Sturgeon individually in 700-L tanks with false bottoms. The 1st day after inoculation and every 2nd day thereafter, we siphoned the aquarium and tank floors through 100-μm-mesh filters and changed the tank water to maintain water quality. We increased water velocity

55

in the AHAB tanks to flush any settled glochidia and juveniles into filter cups. We rinsed all samples were rinsed into Bogorov trays and counted glochidia or juveniles shed by the fish under a stereomicroscope.

We measured water temperature, dissolved O2 (DO), and pH in each fish holding system daily with a Hydrolab Quanta (Hach Hydromet, Loveland, Colorado). We monitored NH3 concentrations weekly with a LaMotte colorimeter (LaMotte Co.,

Chestertown, Maryland). Water temperature averaged 23 ± 1°C for all fishes held in the

AHAB systems, 21 ± 1°C for the Gulf and Shortnose sturgeons aquaria, and 18 ± 1°C for the Atlantic and Lake sturgeons aquaria. All other water-chemistry variables were maintained within suitable levels for aquatic organisms throughout the study (DO =7.5–

8.2 mg/L, pH = 6.8–7.6, total NH3 = <0.1 mg/L).

We removed a fish species from a trial after no glochidia or juveniles were recovered in 3 consecutive monitoring periods. We examined the fishes’ gills to ensure that no encysted glochidia remained. We approximated the number of glochidia that attached to each fish as the sum of sloughed glochidia and metamorphosed juveniles recovered by the end of each exposure trial. We calculated the % metamorphosis for each individual fish by dividing the number of juveniles by the sum of glochidia and juveniles recovered from that fish. We identified the number of days to peak metamorphosis (i.e., the day yielding the greatest number of juveniles) for each species. Percent metamorphosis was not normally distributed. Therefore, we tested the null hypotheses that metamorphosis success did not differ among fish species and genera with a Kruskal–

Wallis test (SAS, version 8.1; SAS Institute, Cary, North Carolina), followed by a post hoc Mann–Whitney U-test.

56

We used a NaCl challenge test to compare the relative health and vigor of juvenile mussels produced from fishes that facilitated metamorphosis during trials conducted in

2011. We followed established standardized guidelines for toxicity testing with early life stages of freshwater mussels (ASTM 2006). For the sturgeon group, we combined 80

Purple Bankclimber juveniles (age = 23–28 d) recovered from each of the 3 sturgeon species used in 2011 (Atlantic, Shortnose, and Lake; n = 240). For the darter group, we combined 120 Purple Bankclimber juveniles (age = 21–26 d) recovered from each of the

2 darter species (Blackbanded, Halloween; n = 240). Juveniles from each group were exposed to 1 of 5 NaCl concentrations and a control (0, 0.5, 1, 2, 4, 8 mg/L NaCl). We created test solutions by dissolving reagent-grade NaCl (Sigma-Aldrich, St. Louis,

Missouri) in dechlorinated tap water and renewed them (100%) after 48 h. We confirmed all NaCl concentrations with a refractometer at the start of the test and after renewal. We conducted exposures in 150-mL glass beakers with 100 mL of test solution. Each treatment concentration had 4 replicates with 10 mussels/replicate. We detected survival at 48 and 96 h by observing individual mussels for foot movement or valve closure in response to a stimulus (prodding with a probe). If we observed no response within 5 min, we considered the mussel dead. We removed any dead mussels observed at 48 h. We used the Trimmed Spearman–Karber Method (Hamilton et al. 1977) with TOXSTAT

(WEST, Laramie, Wyoming) software to determine the median lethal concentration

(LC50) of NaCl at 48 and 96 h. As a conservative comparison, we did not consider

LC50s as significantly different between groups if 95% confidence intervals (CI) for

LC50s overlapped.

57

RESULTS

Glochidia had high metamorphosis success on 4 of the 29 fish species tested

(Table 3.1) (% metamorphosis ± 95% CI, Gulf Sturgeon: 88 ± 3%, Atlantic Sturgeon: 89

± 7%, Lake Sturgeon: 79 ± 8%, Shortnose Sturgeon: 86 ± 6%. Two species of darters also produced juvenile Purple Bankclimbers but with lower metamorphosis success:

Blackbanded Darters (36% ± 14%) and Halloween Darters (P. crypta) (34% ± 18%) (Fig.

3.2). The mean number of juveniles produced/g fish was similar between darters and sturgeons, but the number of juveniles produced/fish was lower for darters (average for both species = 47) than for sturgeon, all of which produced >4000 juveniles (Fig. 3.3).

Peak day of metamorphosis varied by water temperature and species and occurred on day

16 for Gulf Sturgeon, day 21 for Atlantic Sturgeon, day 19 for Lake Sturgeon, day 14 for

Shortnose Sturgeon, and day 12 for darters (Table 3.2).

Based on overlapping 95% CIs, response to the NaCl challenge test was similar for Purple Bankclimber juveniles from sturgeon and darters (Fig. 3.4). The 48-h LC50 was 2.11 mg/L for juveniles from sturgeon and 2.41 mg/L for juveniles from darters. The

96-h LC50 for juveniles from sturgeon was 1.87 mg/Land 1.79 mg/L for juveniles from darters.

DISCUSSION

Our % metamorphosis data suggest that the federally threatened Gulf Sturgeon, the only sturgeon species present in the ACF (Boschung and Mayden 2004, Straight et al.

2009), probably is a primary host for the Purple Bankclimber. Furthermore, the period of glochidial brooding and release for the Purple Bankclimber (March–April; temperature

58

range 16–23°C; USGS, AKF, unpublished data) coincides with the historical timing of

Gulf Sturgeon spawning migrations in the ACF (Wooley and Crateau 1985, Fox et al.

2000, O’Brien and Williams 2002, Flowers et al. 2009), suggesting the potential for a close evolutionary linkage between these 2 species.

Darters appeared to be only marginally suitable hosts on the basis of their much lower metamorphosis success. It was necessary to hold sturgeon is aquaria separate from the AHAB units used to hold smaller fishes, such as darters. However, we were unable to maintain strictly equivalent temperatures among these holding facilities, and the temperature was 2°C warmer for darter trials than Gulf Sturgeon trials. Water temperature can strongly affect metamorphosis success in laboratory host trials (Roberts and Barnhart 1999). However, the temperatures of both darter and sturgeon trials were within the range experienced by Purple Bankclimbers during their brooding period (see

Methods). Furthermore, water temperatures in the Flint River during the brooding period show considerable daily fluctuation (AKF, unpublished data), and the small discrepancies in temperature among trials in our study probably had little effect on metamorphosis success.

Purple Bankclimber populations in the Flint River show only limited recruitment in recent years, and populations are dominated by older individuals (probably >50 y old) that probably recruited before installation of Jim Woodruff Dam (Jason Wisniewski,

Georgia Department of Natural Resources, unpublished data). Nevertheless, occasional observations of younger individuals (<40 mm) indicates that Purple Bankclimbers continue to reproduce in the absence of their primary host, probably because of their marginal ability to metamorphose on other fishes, particularly darters. Even though the

59

number of juvenile mussels produced/g host fish was similar among sturgeon and darters, a single adult Gulf Sturgeon (commonly >75 kg) can produce >200× more juveniles than an adult darter (~2 g). Therefore, the combination of lower % metamorphosis on darters and their smaller size results in far lower recruitment when darters are used as hosts than when sturgeon are used as hosts. However, total darter abundance probably is much higher than sturgeon abundance and may balance the inequity in juvenile production to some extent. Additional data and modeling approaches are required to understand better the relative contributions of juvenile mussels from darters and sturgeon in the wild. The livebearers (Poeciliidae), Eastern Mosquitofish, and Guppies, which were previously identified as hosts (O’Brien and Williams 2002), are unlikely to be important in the maintenance of Purple Bankclimber populations. Livebearers and topminnows

(Fundulidae) serve as marginal hosts for many mussel species, but these fishes are surface feeders and typically occur in shallow stream margins and backwaters not inhabited by mussels, so they probably rarely encounter glochidia in the wild (Haag and

Warren 1997).

Percent metamorphosis was higher on sturgeons than on darters, but the results of the NaCl challenge test suggest that juveniles produced from darters were of equivalent health to those produced from sturgeons. Additional studies (e.g., juvenile growth and survival) are needed to compare the fitness of juvenile mussels produced from different host species more completely. To our knowledge, no other investigators have attempted to compare fitness of juveniles produced from different host species.

From an evolutionary perspective, sturgeons have many desirable traits as mussel hosts: their benthic nature may increase interactions with gravid mussels or released

60

glochidia, their large size allows them to carry large numbers of glochidia, and their long spawning runs facilitate widespread mussel dispersal. Sturgeons have been identified as hosts for other freshwater mussel species in North America and elsewhere. The

Hickorynut (Obovaria olivaria), a Mississippi River basin species, metamorphoses primarily on Shovelnose Sturgeon (Scaphirhynchus platorynchus) and Lake Sturgeon

(Coker et al. 1921, Brady et al. 2004). A European species, Margaritifera auricularia, metamorphoses robustly on European Sea Sturgeon (Acipenser sturio), Siberian Sturgeon

(Acipenser baerii), and Adriatic Sturgeon (Acipenser naccarii) and marginally on River

Blenny (Salaria fluviatilis) (Araujo and Ramos 2000, Araujo et al. 2003, Lopez et al.

2007). Unfortunately, the worldwide decline of freshwater mussels has occurred simultaneously with a similar decline in sturgeons and other migratory fishes, making the long-term fate of mussels that are dependent on sturgeon uncertain at best.

Loss of a migratory host has several long-term ecological implications for mussels. For mussels that cannot use other hosts, recruitment ceases when their migratory hosts disappear and populations eventually become extirpated (Davenport and Warmuth

1965, Smith 1985, Kelner and Sietman 2000, Haag 2012). Other mussels species that have lost a migratory host have been able to persist by relying on a less mobile, but marginally suitable host (Araujo et al. 2003). However, as primarily sessile organisms, mussels depend on the parasitic relationship with their fish hosts for dispersal. Use of host species with small home ranges (e.g., darters, minnows, blennies) may limit dispersal and recolonization of mussels to new or restored habitats. Conversely, highly migratory species (e.g., sturgeons, Striped Bass [Morone saxatalis], river herrings [Alosa spp.]) could facilitate broader occupancy patterns for mussels. Lack of migratory hosts

61

and the subsequent decrease in dispersal can also compromise the genetic integrity of mussel populations (Kat 1984, Minns 1995, Schwalb et al. 2010).

Management actions outlined in the Federal Recovery Plan for the Purple

Bankclimber include captive propagation and population augmentation to help stabilize or reverse declines in the ACF (US Fish and Wildlife Service 2003). Our finding that

Purple Bankclimbers metamorphose at a high rate on a number of sturgeon species shows that mussel propagation efforts may not be dependent on use of the threatened Gulf

Sturgeon, but could be based on use of more readily obtained nonlisted sturgeon species.

However, further studies are needed to ensure that the fitness of juvenile mussels cultured in captivity on nonsympatric species is not impaired by domestication selection or reduction in genetic diversity (Jones et al. 2006).

Captive propagation can help stabilize Purple Bankclimber populations in the short term, but long-term sustainability of these populations is dependent on restoration of sturgeon access to the upper ACF by appropriate modifications to Jim Woodruff Dam.

Some research involving migratory species has been done at Jim Woodruff Dam.

Investigators assessed the efficacy of Alabama Shad (Alosa alabamae) passage through the lock (Young et al. 2012) and movement of tagged juvenile hatchery-reared Gulf

Sturgeon released above the dam (Weller 2002). Additional research is needed to evaluate structural or operational dam modifications that would allow sturgeon to navigate upstream through the reservoir to historical spawning areas where mussels occur. Equally important is the ability of sturgeon to pass downstream during outmigration (Hightower et al. 2002). Thus, dam modifications and passage designs for

62

sturgeon and other migratory fishes must optimize passage up- and downstream (US

Geological Survey 2008).

Recovery of imperiled mussel and fish populations will ultimately depend on ecosystem-wide habitat restoration for all life stages (Neves et al. 1997). The extirpation of migratory fishes from rivers worldwide has had cascading effects other than the diminished reproductive potential of freshwater mussels (Naiman et al. 2002, Agostinho et al. 2008). Our study highlights the interconnectedness and complexity of ecological relationships and their importance in conservation biology. Habitat fragmentation places the sturgeon population in the ACF at risk of extinction, and because of its dependence on sturgeon as hosts, the Purple Bankclimber also is at risk. Similar examples of species coextinctions, in which the fate of one species is directly tied to that of another, have been documented worldwide and they underscore the need for a holistic, ecosystem approach to conservation (Koh et al. 2004).

63

ACKNOWLEDGMENTS

We are grateful for the assistance of many individuals in the laboratory or field, including

M. Bednarski, J. Dycus, K. Herrington, C. Shea, D. Watrous, N. Willett, and J.

Wisniewski. Fish were generously donated by Owens and Williams Fish Farm. This research was completed under the Federal Fish and Wildlife Permit number TE10239A-0

(mussels), Florida Fish and Wildlife Conservation Commission Special Activity License

SAL-11-1346-SRP (Gulf Sturgeon holding), and Florida Fish and Wildlife Conservation

Commission Special Activity License (SAL-11-0891-SRP) (Gulf Sturgeon collection).

64

REFERENCES

Agostinho, A. A., F. M. Pelicice, and L. C. Gomes. 2008. Dams and the fish fauna of the

Neotropical region: impacts and management related to diversity and fisheries.

Brazilian Journal of Biology 68:1119–1132.

ASTM (American Society for Testing and Materials). 2006. Standard guide for

conducting laboratory toxicity tests with freshwater mussels. E2455-06. Pages

1393–1444 in Annual Book of ASTM Standards. Volume 11.06. American

Society for Testing and Materials Philadelphia, Pennsylvania.

Araujo, R., and M. A. Ramos. 2000. Status and conservation of the giant European

freshwater pearl mussel (Margaritifera auricularia) (Spengler, 1793) (Bivalvia:

Unionoidea). Biological Conservation 96:233–239.

Araujo, R., M. Quiros, and M. A. Ramos. 2003. Laboratory propagation and culture of

juveniles of the endangered freshwater mussel Margaritifera auricularia

(Spengler, 1793). Journal of Conchology 38:53–60.

Barnhart, M. C., W. R. Haag, and W. N. Roston. 2008. Adaptations to host infection and

larval parasitism in Unionoida. Journal of the North American Benthological

Society 27:370–394.

Bogan, A. E. 1993. Freshwater bivalve extinctions (Molluska: Unionoida): a search for

causes. American Zoologist 33:599–609.

Boschung, H.T., and R.L. Mayden. 2004. Fishes of Alabama. Smithsonian Books,

Washington, D.C.

65

Brady, T., M. Hove, C. Nelson, R. Gordon, D. Hornbach, and A. Kapuscinski. 2004.

Suitable host fish determined for hickorynut and pink heelsplitter. Ellipsaria 6:14–

15.

Brim Box, J., and J. D. Williams. 2000. Unionid mollusks of the Apalachicola Basin in

Alabama, Florida and Georgia. Alabama Museum of Natural History Bulletin

21:1–143.

Coker, R. E., A. F. Shira, H. W. Clark, and A. D. Howard. 1921. Natural history and

propagation of freshwater mussels. Bulletin of the U.S. Bureau of Fisheries

37:75–182.

Davenport, D., and M. Warmuth. 1965. Notes on the relationship between the freshwater

mussel Anodonta implicata Say and the alewife Pomolobus pseudoharengus

(Wilson). Limnology and Oceanography 10 (Supplement):R74–R78.

Dodd, B. J., M. C. Barnhart, C. L. Rogers-Lowery, T. B. Fabian, and R. V. Dimock.

2006. Persistence of host response to glochidia larvae in Micropterous salmoides.

Fish and Shellfish Immunology 21:473–484.

Flowers, H. J. , W. E. Pine, A. C. Dutterer, K. G. Johnson, J. W. Ziewitz, M. S. Allen,

and F. M. Parauka. 2009. Spawning site selection and potential implications of

modified flow regimes on viability of Gulf Sturgeon populations. Transactions of

the American Fisheries Society 138:1266–1284.

Fox, D. A., J. E. Hightower, and F. M. Parauka. 2000. Gulf sturgeon spawning migration

and habitat in the Choctawhatchee River system, Alabama–Florida. Transactions

of the American Fisheries Society 129:811–826.

66

Haag, W. R. 2012. North American freshwater mussels: natural history, ecology, and

conservation. Cambridge University Press, Cambridge, United Kingdom.

Haag, W. R., and M. L. Warren. 1997. Host fishes and reproductive biology of 6

freshwater mussels species from the Mobile Basin U.S.A. Journal of the North

American Benthological Society 16:576–585.

Hamilton, M. A., R. C. Russo, and R. V. Thurston. 1977. Trimmed Spearman–Karber

method for estimating median lethal concentrations in toxicity bioassays.

Environmental Science and Technology 11:714–719.

Hightower, J. E., K. P. Zehfuss, D. A. Fox, and F. M. Parauka. 2002. Summer habitat use

by Gulf sturgeon in the Choctawhatchee River, Florida. Journal of Applied

Ichthyology 18:595–600.

Jones, J. W., E. M. Hallerman, and R. J. Neves. 2006. Genetic management guidelines

for captive propagation of freshwater mussels (Unionoidea). Journal of Shellfish

Research 25:527–535.

Kat, P. W. 1984. Parasitism and the Unionacea (Bivalvia). Biological Reviews of the

Cambridge Philosophical Society 59:189–207.

Kelner, D. E., and B. E. Sietman. 2000. Relic populations of the ebony shell, Fusconaia

ebena (Bivalvia:Unionidae), in the upper Mississippi River drainage. Journal of

Freshwater Ecology 15:371–377.

Koh, L. P., R. R. Dunn, N. S. Sodhi, R. K. Colwell, H. C. Proctor, and V. S. Smith. 2004.

Species coextinctions and the biodiversity crisis. Science 304:1632–1634.

Limburg, K. E., and J. R. Waldman. 2009. Dramatic declines in North Atlantic

diadromous fishes. BioScience 59:955–965.

67

Lopez, M. A., C. R. Altaba, T. Rouault, and E. Gisbert. 2007. The European sturgeon

Acipenser sturio is a suitable host for the glochidia of the freshwater pearl mussel

Margaritifera auricularia. Journal of Molluscan Studies 73:207–209.

Merz, J. E., and P. B. Moyle. 2006. Salmon, wildlife and wine: marine-derived nutrients

in human-dominated ecosystems of central California. Ecological Applications

16:999–1009.

Minns, C. K. 1995. Allometry of home range size in lake and river fishes. Canadian

Journal of Fisheries and Aquatic Sciences 52:1499–1508.

Naiman, R. J., R. E. Bilby, D. E. Schindler, and J. M. Helfield. 2002. Pacific salmon,

nutrients, and the dynamics of freshwater and riparian ecosystems. Ecosystems

5:399–417.

Nelson, J. S., E. J. Crossman, H. Espinosa-Pérez, L. T. Findley, C. R. Gilbert, R. N. Lea,

and J. D. Williams. 2004. Common and scientific names of fishes from the United

States, Canada, and Mexico. 6th edition. American Fisheries Society Special

Publication 29. American Fisheries Society, Washington, DC.

Neves, R. J., A. E. Bogan, J. D. Williams, S. A. Ahlstedt, and P. W. Hartfield. 1997.

Status of aquatic mollusks in the southeastern United States: a downward spiral of

diversity. Pages 43–85 in G. W. Benz and D. E. Collins (editors). Aquatic fauna

in peril: the southeastern perspective. Special Publication No. 1, Southeast

Aquatic Research Institute. Lenz Design and Communications, Decatur, Georgia.

Neves, R. J., L. R. Weaver, and A. V. Zale. 1985. An evaluation of host fish suitability

for glochidia of Villosa vanuxemi and Villosa nebulosa (Pelecypoda, Unionidae).

American Midland Naturalist 113:13–19.

68

O'Brien, C. A., and J. A. Williams. 2002. Reproductive biology of four freshwater

mussels (Bivalvia: Unionidae) endemic to eastern Gulf Coastal Plain drainages of

Alabama, Florida, and Georgia. American Malacological Bulletin 17:147–158.

Ricciardi, A., R. J. Neves, and J. B. Rasmussen. 1998. Impending extinctions of North

American freshwater mussels (Unionoida) following the zebra mussel (Dreissena

polymorpha) invasion. Journal of Animal Ecology 67:613–619.

Ricciardi, A., and J. B. Rasmussen. 1999. Extinction rates of North American freshwater

fauna. Conservation Biology 13:1220–1222.

Roberts, A. D., and M. C. Barnhart. 1999. Effects of temperature, pH, and CO2 on

transformation of the glochidia of Anodonta suborbiculata on fish hosts in vitro.

Journal of the North American Benthological Society 18:477–487.

Schwalb, A. N., K. Cottenie, M. S. Poos, and J. D. Ackerman. 2010. Dispersal limitation

of unionid mussels and implications for their conservation. Freshwater Biology

56:1509–1518.

Silverman, H., S. J. Nichols, J. S. Cherry, E. Achberger, J. W. Lynn, and T. H. Dietz.

1997. Clearance rates of labortory-cultered bacteria by freshwater bivalves:

differences between lentic and lotic unionids. Canadian Journal of Zoology

75:1857–1866.

Smith, D. G. 1985. Recent range expansion of the freshwater mussel Anodonta implicata

and its relationship to clupeid fish restoration in the Connecticut River System.

Freshwater Invertebrate Biology 4:105–108.

69

Straight, C. A., B. Albanese, and B. J. Freeman. 2009. Fishes of Georgia. Georgia

Museum of Natural History, Athens, Georgia. (Available from:

http://fishesofgeorgia.uga.edu)

Strayer, D. L., D. C. Hunter, L. C. Smith, and C. Borg. 1994. Distribution, abundance,

and role of freshwater clams (Bivalvia: Unionidae) in the freshwater tidal Hudson

River. Freshwater Biology 31:239–248.

US Fish and Wildlife Service and the Gulf States Marine Fisheries Commission. 1995.

Gulf sturgeon recovery plan. US Fish and Wildlife Service, Atlanta, Georgia.

(Available from: http://www.nmfs.noaa.gov/pr/recovery/plans.htm)

US Fish and Wildlife Service. 2003. Recovery plan for endangered fat threeridge

(Amblema neislerii), shinyrayed pocketbook (Lampsilis subangulata), Gulf

moccasinshell (Medionidus penicillatus), Ochlockonee moccasinshell

(Medionidus simpsonianus), and oval pigtoe (Pleurobema pyriforme); and

threatened Chipola slabshell (Elliptio chipolaensis), and purple bankclimber

(Elliptoideus sloatianus). US Fish and Wildlife Service, Atlanta, Georgia.

(Available from: http://www.fws.gov/southeast/news/2003/7-mussel_recorvery-

plan.pdf)

US Geological Survey. 2008. White sturgeon passage at the Dalles Dam. Fact Sheet

2008-3081. US Geological Survey, Reston, Virginia.

Vaughn, C. C., S. J. Nichols, and D. E. Spooner. 2008. Community and foodweb ecology

of freshwater mussels. Journal of North American Benthological Society 27:409–

423.

70

Vaughn, C. C., and D. E. Spooner. 2006. Unionid mussels influence macroinvertebrate

assemblage structure in streams. Journal of the North American Benthological

Society 25:691–700.

Weller, R. R. 2002. Movements of hatchery-raised Gulf sturgeon in Lake Seminole.

Georgia Department of Natural Resources Wildlife Division report. Georgia

Department of Natural Resources, Albany, Georgia.

Williams, J. D., A. E. Bogan, and J. T. Garner. 2008. Freshwater mussels of Alabama and

the Mobile Basin in Georgia, Mississippi and Tennessee. University of Alabama

Press, Tuscaloosa, Alabama.

Williams, J. D., M. L. Warren, K. S. Cummings, J. L. Harris, and R. J. Neves. 1993.

Conservation status of freshwater mussels of the United States and Canada.

Fisheries 18(9):6–22.

Wooley, C. M., and E. J. Crateau. 1985. Movement, microhabitat, exploitation, and

management of Gulf of Mexico sturgeon, Apalachicola River, Florida. North

American Journal of Fisheries Management 5:590–605.

Young, S. P., T. R. Ingram, J. E. Tannehill, and J. J. Isely. 2012. Passage of Spawning

Alabama Shad at Jim Woodruff Lock and Dam, Apalachicola River, Florida.

Transactions of the American Fisheries Society 141:881–889.

Zale, A. V., and R. J. Neves. 1982. Identification of a host fish for Alasmidonta minor

(Mollusca: Unionidae). American Midland Naturalist 107:386–388.

Zehfuss, K. P., J. E. Hightower, and K. H. Pollock. 1999. Abundance of Gulf sturgeon in

the Apalachicola River, Florida. Transactions of the American Fisheries Society

128:130–143.

71

Fig. 3.1. Apalachicola–Chattahoochee–Flint River (ACF) in Alabama, Florida, and

Georgia showing location of Jim Woodruff Dam. Star indicates the location of mussel collections.

72

*** 100

80

60

40 % metamorphosis % 20

0 Gulf Shortnose Atlantic Lake Halloween Blackbanded

Sturgeons Darters

Fig. 3.2. Mean (± 95% CI) % metamorphosis of Purple Bankclimber glochidia on sturgeons (Acipenser spp.) and darters (Percina spp.). *** indicates significant difference between sturgeons and darters (Kruskal–Wallis, p < 0.0001). Bars under the same horizontal line are not significantly different.

73

Fig. 3.3. Mean (± 95% CI) number of Purple Bankclimber juveniles produced/host fish and number of juveniles produced/g host fish biomass during laboratory trials.

74

Fig. 3.4. Mean (± 95% CI) median lethal concentrations (LC50) of NaCl for juvenile

Purple Bankclimber mussels produced from sturgeons (Acipenser spp.) and darters

(Percina spp.).

75

Table 3.1. Fish species tested as potential hosts for Purple Bankclimbers. Hatchery reared species are denoted HR. all other species were field-collected. Nomenclature follows Nelson et al. (2004). Days to rejection indicates length of time until all glochidia had sloughed from nonhost fishes.

Family Species Common name n tested Days to rejection

Acipenseridae Acipenser brevirostrum Shortnose Sturgeon (HR) 4 –

Acipenser fulvescens Lake Sturgeon (HR) 2 –

Acipenser oxyrinchus desotoi Gulf Sturgeon 3 –

Acipenser oxyrinchus oxyrinchus Atlantic Sturgeon 4 –

Catostomidae Minytrema melanops Spotted Sucker 1 3

Hypentelium nigricans Northern Hogsucker 5 3

Moxostoma robustum Robust Redhorse (HR) 5 3

Centrarchidae Lepomis auritus Redbreast Sunfish 5 3

Lepomis cyanellus Green Sunfish 4 4

Lepomis macrochirus Bluegill 5 9

Lepomis microlophus Redear Sunfish 5 3

76

Lepomis punctatus Spotted Sunfish 3 9

Micropterus cataractae Shoal Bass 3 3

Micropterus salmoides Largemouth Bass 5 3

Cyprinidae Cyprinella venusta Blacktail Shiner 4 1

Nocomis leptocephalus Bluehead Chub 8 3

Notropis lutipinnis Yellowfin Shiner 4 5

Pimephales promelas Fathead Minnow (HR) 5 3

Semotilus atromaculatus Creek Chub 2 1

Ictaluridae Ameiurus brunneus Snail Bullhead 3 3

Ameiurus natalis Yellow Bullhead 3 3

Ictalurus punctatus Channel Catfish 3 3

Noturus leptacanthus Speckled madtom 1 4

Pylodictis olivaris Flathead Catfish 2 3

Moronidae Morone saxatilis Striped Bass 5 7

Percidae Etheostoma inscriptum Turquoise Darter 9 7

77

Etheostoma olmstedi Tessellated Darter 1 7

Percina crypta Halloween Darter 12 –

Percina nigrofasciata Blackbanded Darter 15 –

78

Table 3.2. Mean (± 95% CI) total length (TL) and mass of fish species that facilitated

metamorphosis of Purple Bankclimber glochidia to juveniles. Number of days to peak

juvenile metamorphosis (M) and mean temperature (±1°C) for the duration of the

experiment are shown.

Days to Temperature

Species n tested TL (mm) Mass (g) peak M (°C)

Gulf Sturgeon 3 687 ± 107 1416 ± 627 16 21

Shortnose Sturgeon 4 484 ± 35 551 ± 87 14 21

Lake Sturgeon 2 538 ± 5 581 ± 122 19 18

Atlantic Sturgeon 4 408 ± 42 228 ± 62 21 18

Blackbanded Darter 15 60 ± 7 1.8 ± 0.6 12 23

Halloween Darter 12 62 ± 6 1.9 ± 0.7 12 23

79

CHAPTER 4

EVALUATION OF PHYSIOLOGICAL BIOMARKERS OF STRESS

IN FRESHWATER MUSSELS3

3 Fritts, A.K., J.T. Peterson, P.D. Hazelton, and R.B. Bringolf. To be submitted to

Comparative Biochemistry and Physiology.

80

ABSTRACT

Biomarkers provide researchers an opportunity to evaluate the effects of emerging stressors on aquatic fauna. Freshwater mussels are an imperiled group that are highly susceptible to environmental alterations due to their diminished population sizes and primarily sessile behaviors, thus supporting the need to develop a non-lethal biomonitoring program to evaluate the health of the remaining populations. Our objectives were 1) to determine which freshwater mussel hemolymph biochemical parameters are consistently within detectable limits and how hemolymph and tissue glycogen respond to a thermal stress event, 2) evaluate the effects of tissue and/or hemolymph extraction on long-term growth and survival of small vs. large-bodied species, and 3) compare hemolymph parameters from anterior versus posterior adductor muscles within the same individual. Three species of mussels, Elliptio crassidens,

Villosa vibex, and Villosa lienosa, were exposed to elevated water temperatures in the lab

(25, 30, 35°C) for a period of seven days. Hemolymph was extracted from the adductor muscle sinus and tissue biopsies were taken from the foot tissue. Six of 16 hemolymph parameters were consistently within detectable limits and responded to the thermal stress event: alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase, bicarbonate, calcium, and magnesium. Glucose was detectable in some instances. The effects of hemolymph extraction and foot tissue biopsies were evaluated with Elliptio crassidens and V. vibex. Individuals of both species were randomly divided between four treatments: hemolymph extraction, tissue biopsy, combined tissue and hemolymph extraction, and control. Individuals were held for 820 to 945 days after the treatments were administered to evaluate the long term effects. Only a moderate increase

81

in risk of mortality was attributed to the treatments, supporting the use of these techniques in non-lethal biomonitoring programs. Hemolymph collected from anterior versus posterior adductor muscles in Elliptio crassidens was evaluated for differences between the collection locations. Expression of two enzymes (ALT and AST) was consistently elevated in the posterior adductor muscle; none of the remaining parameters varied between locations. These results provide useful data for developing a biomonitoring plan for imperiled freshwater mussels.

82

INTRODUCTION

Ecosystems around the globe have experienced substantial levels of modification from anthropogenic activities. Many of these changes have been at the expense of the native flora and fauna, which are often harmed by the myriad of new stressors that have been added to their habitats (Sherry 2003). In aquatic environments, negative effects may be caused by changes in water quality or water quantity (Haag and Warren 2008) and common stressors include sedimentation, habitat fragmentation, altered flow regime, chemical contaminants, thermal pollution and hypoxic zones, among others.

Researchers are actively trying to develop methods for evaluating the response of aquatic biota to these emerging stressors. The effects of stress can be divided into three levels: primary, secondary, and tertiary effects, with the severity of the stress event increasing with each level. Primary effects occur rapidly and include fluctuations in endocrine products such as hormones, secondary effects occur more slowly and include changes in blood chemistries, energy stores, or elevated heart rate, and tertiary effects are the most extreme (e.g. decreased growth, impaired disease resistance, mortality) and are also the most likely to be detected because they operate at the whole-organism or population level. However, once tertiary effects are observed it is often too late to take action to decrease the effect of the stress event. This is particularly true for imperiled species that already exist at diminished population levels. Stress events that alter molecular or physiological changes can also affect higher order responses such as growth, immune function, and reproduction. A tool that has emerged for evaluating stress events is the use of biomarkers, or the tracking of specific biological processes and how those processes react to changes in the organism’s habitat (Gagne et al. 2002, Blaise

83

and Gagne 2009). Shugart et al. (1992) defined a biomarker as, “a xenobiotically induced variation in cellular or biochemical components or processes, structures, or functions that is measurable in a biological system or samples.” These physiological responses to changing conditions may serve as early warning signs before acute mortality begins (Handy and Depledge 1999) or serve as a method of studying the effects of sublethal stress events (Sherry 2003).

The creation and calibration of useful biomarkers is a multi-step process. The first requirement is the determination of “normal” levels under specific conditions such as temperature, pH, and season (Handy and Depledge 1999). Determining the natural variability and acceptable variation among individuals within a population under specific conditions is imperative (Förlin et al. 1986); the potential influences of age and gender should also be investigated (Giesy and Graney, 1988). Developing non-lethal sampling techniques will also improve the usefulness of biomarkers, particularly when working with imperiled fauna (Sherry 2003). The next step is to establish reference ranges for the biomarkers in specific habitats and the rate at which the biomarker parameters can respond to changes. A biomarker for chronic stress must exhibit a consistent response to changes over time. Changes in noradrenaline and dopamine hormones are short term indicators of stress (e.g., mechanical disturbance) that respond very rapidly and also recover to initial levels in a short period of time (Lacoste et al. 2002). This rapid reaction time removes noradrenaline and dopamine from the list of potential biomarkers of chronic exposure to a stressor.

Physiological indicators of stress have been identified in marine bivalves (e.g. Liu et al. 2004, Li et al. 2007) due to the economic incentive of this fishery, but information

84

on freshwater bivalves is still scarce. Freshwater mussels (Unionoidea) comprise a diverse faunal group and function as key components of aquatic ecosystems. They provide numerous ecological services that benefit a broad range of aquatic species

(Vaughn et al. 2008). The past two decades have seen a substantial increase in research involving freshwater mussels, largely spurred by the alarming decline of this imperiled faunal group (see Haag 2012 for overview). Given the imperiled status of many freshwater mussel species, methods are needed for monitoring the health and status of the remaining populations facing environmental perturbations such as habitat fragmentation and exposure to toxicants. Development of useful biomarkers that can be evaluated non- lethally represents a substantial challenge to the creation of effective biomonitoring plans for endangered mussels.

Previous researchers have indicated the potential of two sources of samples for biomarker evaluation in freshwater mussels: somatic tissues and hemolymph. Somatic tissue is commonly used during assessment of various biomarkers, including contaminant accumulation and changes in glycogen (Haag et al. 1993, Waller et al. 1998). However, the majority of previous studies have sacrificed the mussels for tissue collection. In early attempts to develop a method for non-lethal tissue collection, Naimo et al. (1998) evaluated the effect of a 5-10 mg foot tissue biopsy on the survival of Amblema plicata.

This technique did not significantly alter survival over the course of 581 d for this large- bodied species. In another study, the biopsy of mantle tissue samples from Quadrula quadrula and Actinonaias ligamentina did not significantly affect survival for up to one year post sampling (Berg et al. 1995). These tissue extraction techniques may provide an effective method for non-lethal tissue collection for biomonitoring programs of

85

freshwater mussels. Hemolymph, the circulatory fluid of mussels, can be analyzed and evaluated in much the same way as vertebrate blood (McMahon and Bogan 2001).

Gustafson et al. (2005a) evaluated the effects of extracting hemolymph from two different locations, the cardiac chamber and the adductor muscle sinus, on the survival of

Elliptio complanata. Sampling hemolymph from the cardiac chamber resulted in high mortality (100%), but the extraction of hemolymph from the adductor muscle sinus presented a less destructive alternative collection location, with 90% survival three months after sampling (Gustafson et al. 2005a). There was no significant difference in survival between the control group and the those mussels having hemolymph extracted from the adductor sinus for a period of three months after a single sampling event and for seven months after being repeatedly sampled three times over the course of five months

(Gustafson et al. 2005a). While the work of previous researchers has provided an initial understanding of somatic tissues and hemolymph as sources of potential biomarkers and have assessed the mortality associated with the extraction of those tissues from large bodied mussel species (e.g. Q. quadrula, A. ligamentina, Berg et al. 1995; A. plicata,

Naimo et al. 1998; E. complanata, Gustafson et al. 2005a), no previous studies have attempted these tissue extraction methods on smaller mussel species that commonly represent the largest portion of mussel biomass present in smaller streams.

The Lower Flint River Basin (LFRB) in southwest Georgia, USA, is part of the

Apalachicola-Chattahoochee-Flint (ACF) River Basin, and is a watershed where the molluscan fauna is facing substantial amounts of habitat modification and alteration of natural flow regimes. Flows within the LFRB are heavily influenced by aquifer withdrawals for center pivot irrigation, which can lead to the drying of stream reaches

86

during the warmest summer months (Peterson et al. 2011, Rugel et al. 2012, Shea et al.

2013). Mussels stranded in these isolated pools are commonly subjected to hypoxic conditions and elevated temperatures (Golladay et al. 2004). The LFRB is home to five federally listed mussel species: Hamiota subangulata, Pleurobema pyriforme,

Medionidus penicillatus, Amblema neislerii, and Elliptoideus sloatianus. Given the imperiled nature of these endemic populations, there is an intense need to monitor the health of the freshwater mussel species in this basin. Therefore, our study addressed the following objectives: 1) determine which freshwater mussel hemolymph biochemical parameters are consistently within detectable limits, 2) examine if any of the hemolymph parameters or tissue glycogen respond in a predictable manner to a thermal stress event,

3) evaluate the effects of tissue and/or hemolymph extraction on long-term growth and survival, 4) compare the differences in survival between large and small-bodied species, and 5) compare hemolymph parameters from anterior versus posterior adductor muscles within the same individual.

METHODS

Thermal exposure

We tested three species of mussels for their response to thermal stress. Elliptio crassidens (Elephant Ear) were collected from the Flint River near Newton, GA on 13

July 2010. Villosa vibex (Southern Rainbow) and Villosa lienosa (Little Spectaclecase) were collected from Cooleewahee Creek near Newton, GA on 28 June 2010. These species were selected because they have stable, robust populations and represent two tribes within the family Unionidae; E. crassidens (Pleurobemini tribe) is a large-bodied

87

representative and V. vibex and V. lienosa (Lampsilini tribe) are small-bodied representatives (Williams et al. 2008). All species were placed in river water in aerated coolers and transported to the Aquatic Science Laboratory at the University of Georgia

(UGA). Mussels were acclimated in dechlorinated municipal water in the laboratory at

25° C for one week prior to the beginning of the thermal exposures. Thirty-eight animals of each species were used in the experiments. The average length (±SE) of the respective species was as follows: E. crassidens; 84.9 mm ± 0.7, V. vibex; 52.2 mm ± 0.9, and V. lienosa; 41.4 mm ± 0.9.

Three temperature treatments were used (25, 30, 35°C) with five replicates per temperature. Animals were held in 3-L acrylic tanks placed in a water bath to maintain a constant temperature among replicates. Water temperature was raised at a rate of

0.5°C*hr-1 until the target temperature was reached. All experimental chambers were aerated and mussels were not fed. Eight individuals per species were sampled for tissue and hemolymph on day 0 at 25°C as a control, and five individuals per species in all temperature treatments were sampled on days 3 and 7. Individuals were not repeatedly sampled; new individuals were used at each timepoint. Hemolymph was collected following the protocol of Gustafson et al. (2005a). In brief, hemolymph was extracted from the adductor muscle sinus with a 26 gauge needle and a 1-ml syringe. Hemolymph was extracted from the anterior adductor muscle of E. crassidens and the posterior adductor of V. vibex and V. lienosa. While the anterior adductor muscle is more easily accessible, the posterior adductor muscle appears to be larger (Cummings and Graf 2010) and holds a greater volume of hemolymph (A. Fritts, personal observation). It was necessary to sample the posterior adductor muscle of V. vibex and V. lienosa to obtain a

88

sufficient volume of hemolymph for biochemical analysis. The average volume of hemolymph (± SE) collected from E. crassidens was 0.85 ml ± 0.03, V. vibex: 0.45 ml ±

0.02, and V. lienosa: 0.43 ml ± 0.03. Tissue biopsies were extracted from the foot tissue of all three species with a 3-mm oval-cupped jaws biopsy forceps (Surgical Direct,

DeLand, Florida, USA). The average mass (± SE) of the tissue biopsies was 21 mg ± 0.5.

Hemolymph and tissue samples were placed in cryovials and stored in a cryogenic freezer (-80°C) until they were ready to be processed. Water quality parameters were measured daily (Table 4.1).

Mussel hemolymph samples were processed on a Hitachi 912 Blood Chemistry

Analyzer (Roche Diagnostics) at the Clinical Pathology Laboratory at UGA to evaluate the detection ability and thermal stress response for 16 different hemolymph biochemical parameters: alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (AlkPhos), bicarbonate, calcium, phosphorous, magnesium, creatine kinase, creatinine, total protein, albumin, sodium, potassium, chloride, amylase, and glucose.

Tissue samples were processed for glycogen content at the Aquatic Science

Laboratory at UGA following the method of Naimo et al. (1998) with a few modifications. Glassware was washed in an acid bath (10% HNO3) for eight hours. To create calibration standards, a stock solution of 2000 mg/L was made by dissolving 200 mg of powdered oyster glycogen (Sigma-Aldrich) in 100 mL deionized water. The stock solution was serially diluted to create standards with the following concentrations: 2000,

1000, 500, 250, 125 mg/L, plus a control (0 mg/L). As an internal standard, we used homogenized foot tissue from five Lampsilis cardium that was collected from Pool 8 of the Mississippi River in October 2011. We also created spiked standards by adding a

89

known quantity of internal standard tissue to each of the standard curve concentrations.

Each batch of samples was accompanied by three procedural blanks, three replicates of each standard curve concentration, three replicates of spiked standards, and three internal standard samples sampled in triplicate.

The glycogen assay procedure included a tissue digestion component followed by colorimetric analysis. Calibration standards were added to test tubes at a volume of 250

µL. Tissue samples were defrosted, then a 5-10 mg sample was weighed and placed into a test tube. A solution of 30% KOH was added to each test tube at a volume of 100 µL for tissue samples and 500 µL for standard curve samples. All samples were placed into a hot water bath at 100°C for 30 minutes. After heating, the samples were vortexed for 30 seconds and placed into an ice water bath for five minutes. A solution of 95% EtOH was then added to each test tube at a volume of 150 µL for tissues and 750 µL for standards.

Samples were vortexed for five seconds then placed into a hot water bath at 100°C for 20 minutes. After the tissue digestion was complete, samples were either processed immediately or stored in a cryogenic freezer (-80°C). The colorimetric analysis was begun by first diluting all samples to a common volume (7000 µL) with deionized water.

Samples were vortexed for 15 seconds, after which a 2 mL aliquot of the solution was placed into a new test tube. For the color development step, 100 µL of 80% phenol was added to the test tube, followed immediately by 5 mL of concentrated sulfuric acid. The test tube was vortexed for 10 seconds and allowed to stand at room temperature (20°C) for a minimum of 30 minutes to obtain maximum color development. Samples were processed by placing 300 µL of sample into a 96-well microplate, which was then read on a SpectraMax M2 plate reader at 490 nm and processed with SoftMax Pro software.

90

Final glycogen content was normalized for wet tissue mass and was reported as mg glycogen per g tissue.

The effects of elevated temperature and exposure time on the hemolymph parameters and tissue glycogen were analyzed with a two-way Analysis of Variance

(ANOVA) (SAS, version 9.3, SAS Institute, Cary, North Carolina), followed by a post- hoc Dunnett’s test. Values were deemed statistically significant at α ≤ 0.05.

Effects of hemolymph extraction and tissue biopsy

Two freshwater mussel species were used to evaluate the effects of tissue and hemolymph extraction on long-term survival and growth. A total of 100 E. crassidens, a large-bodied unionid, were collected on 6 Apr 2010 from the Flint River main stem near

Newton, GA. The average length (± SE) of individuals was 89 mm ± 0.7. Sixty-five V. vibex, a small-bodied unionid, were collected on 15 Oct 2010 from two tributaries of the

Flint River: Cooleewahee Creek near Newton, GA and Spring Creek near Colquitt, GA.

The average length (± SE) of individuals was 62 mm ± 0.6. Animals were placed into river water in aerated coolers and transported to the Aquatic Science Laboratory at UGA.

All individuals were uniquely identified with a plastic Hallprint shellfish tag (Hallprint,

Hindmarsh Valley, South Australia) that was affixed to the left valve with cyanoacrylate adhesive.

Individuals of both species were randomly divided into four treatment groups: 1) hemolymph extraction, 2) tissue biopsy, 3) tissue biopsy and hemolymph extraction, and

4) control (no tissue extraction). Individuals in the hemolymph extraction group and the tissue extraction group were sampled following the methods described above. The

91

average volume of hemolymph (± SE collected from E. crassidens was 0.90 ml ± 0.03 and V. vibex averaged 0.60 ml ± 0.03. The average mass (± SE) of the tissue biopsies for both species was 24 mg ± 0.9. Individuals in the combined tissue and hemolymph group were sampled with both of the aforementioned techniques. Animals in the control group were subjected to the same handling and holding conditions, but they were not sampled for tissue or hemolymph.

Additional hemolymph samples were collected from both the anterior and posterior adductor muscles of a separate group of E. crassidens (n=10). These samples were designed to test for differences in hemolymph parameters between adductor muscles within an individual. Villosa vibex were not included in this analysis; the anterior adductor muscle did not produce a sufficient volume of hemolymph (>0.25 ml) to be successfully processed with the blood chemistry analyzer.

Mussels were measured for length, height, and width at the beginning of the study. Following tissue extractions, animals were placed in floating plastic crates in an outdoor aerated pond at the Aquatic Science Laboratory at UGA. Survival was assessed weekly for the first two months, bi-monthly for the next 10 months, and monthly thereafter. E. crassidens were held and monitored for 945 days and V. vibex were monitored for 820 days. At the end of the experiment, all surviving individuals were again measured.

To test for changes in growth for V. vibex, growth data were found to be non- normal and efforts to achieve normality via transformation were unsuccessful, therefore growth was analyzed with a non-parametric Kruskal-Wallis test (SAS, version 9.3; SAS

Institute, Cary, North Carolina). E. crassidens were not included in this analysis because

92

the shells had eroded substantially during the holding period in the pond. Most of the E. crassidens individuals already exhibited some erosion near their umbo at the beginning of the study and this erosion increased substantially during the experiment. The elevated amounts of erosion likely occurred because the pond water was exceptionally soft (total hardness averaging 10-20 mg CaCO3/L). Effects on survival were evaluated for both species with a Cox proportional hazard regression (R, version 2.15.2; Statistical

Computing, Vienna, Austria).

RESULTS

Thermal exposure

Hemolymph parameters consistently within detectable limits for all three species included: alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (AlkPhos), bicarbonate, calcium, phosphorous, magnesium, and creatine kinase. Parameters consistently below detectable limits for all three species included: creatinine, total protein, albumin, sodium, potassium, chloride, and amylase. Glucose measurements were occasionally below detectable levels.

Exposure to elevated temperatures caused variable responses among hemolymph parameters and among species. The enzymes (ALT, AST, AlkPhos) exhibited substantial variation among individuals (Table 4.2). In general, ALT and AST levels were elevated in the 30°C treatments and values for AST in V. vibex were nearing significance among temperature treatments (p = 0.07; Table 4.3). Bicarbonate expression decreased on days

3 and 7 in the 30 and 35°C treatments for E. crassidens and V. vibex, but was elevated in the 35°C treatment for V. lienosa. Calcium levels were elevated in the 35°C treatment at

93

all timepoints for both E. crassidens and V. vibex. Magnesium decreased in the35°C treatments at all timepoint for both V. vibex and V. lienosa. Alkaline phosphatase was variable, but showed a general increase in the 35°C treatment for E. crassidens and V. vibex. Glucose was consistently elevated in the 35°C treatment at all timepoints for all three species. Glycogen content for E. crassidens decreased in the 35°C treatment on day

7, but there was a substantial amount of variability among all three species following exposure to elevated temperatures (Table 4.2 and 4.3).

Effects of hemolymph extraction and tissue biopsy

Shell volume of V. vibex was not adversely affected by hemolymph extraction and/or tissue biopsy (Kruskal Wallis, df = 3; F value = 1.044; p = 0.409). The average change in shell volume (as estimated by the cubic root of shell length x width x height;

Gustafson et al. 2005a) was 0.55 mm (range = 0.00 to 1.68 mm) for the control group,

0.47 mm (range = 0.38 to 0.56 mm) for the hemolymph extraction group, 1.29 mm (range

= 0.19 to 2.39 mm) for the tissue biopsy group, and 0.29 mm (range = 0.00 to 0.74 mm) for the combined tissue and hemolymph extraction group. Change in shell growth for E. crassidens was not quantified because of the excessive shell erosion that occurred during the course of the holding period.

Survival of V. vibex and E. crassidens differed greatly over the course of the holding period (Fig. 4.1 and 4.2). Survival of all treatment groups of E. crassidens was greater than 80% for the entirety of the monitoring period (945 d). Villosa vibex exhibited higher mortality in all groups, with survival ranging from 13-53% by the end of the holding period (820 d). The Cox proportional hazard regression did not indicate

94

significant differences in survival between any of the treatment groups and the control for

E. crassidens (p > 0.40) (Table 4.4). For V. vibex, there was a statistically significant increase in hazard risk for the treatment groups (p = 0.05-0.07; Table 4.4). Increased hazard for V. vibex was ~6 times higher for all treatment groups, with 13% survival of the hemolymph extraction and tissue biopsy treatments and 20% survival of the combined tissue and hemolymph extraction treatment, compared to 53% survival of the control group.

Adductor muscle collection site comparison

Of the six hemolymph parameters tested, two (both enzymes) were significantly different between the anterior and posterior adductor muscles of Elliptio crassidens (p =

< 0.035; Table 4.5). Aspartate aminotransferase (AST) and alanine aminotransferase

(ALT) consistently exhibited elevated levels (20 to 21% higher) in the posterior adductor muscle. The other four parameters (bicarbonate, calcium, magnesium, alkaline phosphatase) did not differ significantly between the two adductor muscles (p > 0.05).

DISCUSSION

Thermal exposure

Our results indicate that systematic changes in glycogen content and hemolymph biochemical parameters may be useful as biomarkers of thermal stress in freshwater mussels. In mussel hemolymph, two transaminase enzymes, ALT and AST, showed elevated levels in the 30°C treatment for both E. crassidens and V. vibex. Changes in the expression of transaminases, the enzymes that transfer amino groups during protein

95

production, have been noted as promising biomarkers in a variety of aquatic invertebrates. Increased expression of ALT and AST was documented in Pacific oysters

(Crassostrea gigas) as a result of elevated water temperature (Park et al. 2009) and elevated expression of these two enzymes was also documented with an increase in water temperature and salinity stress in ark shells (Scapharca broughtonii; An and Choi 2010).

Oysters (Crassostrea virginica) infected with the parasite Minchinia nelsoni also exhibited elevated levels of ALT and AST, particularly during the early onset stages of the infection (Douglass and Haskin 1976). As a result, increases in ALT and AST have been hypothesized as an indicator of tissue damage (Park et al. 2009). Elevation of these enzymes has been implicated as a mechanism for degrading foreign proteins, lipids, and carbohydrates (Xue and Renault 2000). Expression of these enzymes peaked at 25°C, which is just slightly below the upper thermal limits of these species (Lui et al. 2004). In the present study, ALT and AST peaked at 30°C, which is likely near the upper thermal tolerance for these species, based upon lethal values calculated for congeners (Pandolfo et al. 2010, Archambault 2012). The elevated enzyme levels indicate that these mussel species may be entering a sub-lethal stressed state at temperatures below those that result in mortality.

Alterations in glucose and glycogen levels have frequently been used as biomarkers of stress in other aquatic invertebrates and teleosts. Glucose is a primary source of energy in all organisms and glycogen is the primary energy storage molecule in freshwater mussels. When glucose levels decline in the hemolymph, glycogen can be catabolized to supplement the glucose levels in the circulatory fluids (De Zwaan and

Zandee 1972). Hypoxic conditions have led to an increase in hemolymph glucose in

96

Penaeus monodon (tiger shrimp; Hall and van Ham 2007) and Haliotis diversicolor supertexta (Tiawan abalone; Cheng et al. 2004), and exposure to pollutants resulted in elevated levels of blood glucose in fishes (Goss and Wood 1988). Tissue glycogen levels have been shown to decrease under stressful conditions such as starvation (Patterson et al.

1999) and zebra mussel infestation (Haag et al. 1993). Patterson et al. (1999) noted decreases in glycogen in as little as seven days of starvation; glycogen levels continued to drop over the course of the 30 d study. Amblema plicata and Lampsilis radiata exhibited significant decreases in glycogen levels after three months of a zebra mussel infestation

(Haag et al. 1993). In our study, all three species exhibited elevated levels of glucose in the highest temperature treatment (35°C), likely indicating that glucose is being mobilized from glycogen reserves through catabolism. We noted a decrease in glycogen content within E. crassidens at 35°C on day 7; while the decrease was not statistically significant, the general trend does support the linkage of glucose and glycogen.

Glycogen levels of both Villosa species were variable and did not demonstrate clear trends associated with the thermal stress. There were no differences in glycogen content between sexes for either Villosa species (mean = 71.9 mg/g for both sexes). We calculated glycogen content based on wet weight, but we feel that future work should be conducted with dried tissues to help decrease the variability among samples.

Additionally, our trials may not be been of sufficient length to elucidate a substantial decrease in tissue glycogen.

Mollusks, organisms that reside within calcareous shells, are uniquely dependent

2+ - upon maintaining optimal balances of ions, including Ca and HCO3 . Elliptio crassidens and V. vibex exhibited elevated levels of calcium in the 35°C treatment.

97

Another freshwater bivalve, Corbicula fluminea, exhibited an increase in hemolymph calcium under hypoxic conditions (Byrne et al. 1991). Both calcium and magnesium levels increased under hypoxic conditions in three other aquatic invertebrates: Saduria entomon (isopod; Hagerman and Szaniawska 1991), Leptograpsus variegatus (purple shore crab; Morris and Butler 1996), and Astacus astacus (crayfish; Nikinmaa et al.

1985). After an acute acid exposure, four species of unionids exhibited elevated levels of calcium in hemolymph collected from the anterior adductor muscle (Pynnönen 1990) and a subsequent study found that the increase in calcium was also mirrored by an increase in

2+ - bicarbonate (Pynnönen 1994). Concurrent increases in Ca and HCO3 offers substantial support that CaCO3 reserves from shell dissolution are being used to mitigate against hemolymph acidosis (Malley et al. 1988, Byrne and McMahon 1991). Acidosis (an increase in hemolymph CO2 concentration and decrease in hemolymph pH) may occur after exposure to acidic conditions, hypoxic conditions, or emersion in the air (De Zwaan and Wijsman 1976, Byrne and McMahon 1991, Pynnönen 1994, Michaelidis et al. 2005).

In the present study, the thermal stress exposure may have induced a slight degree of acidosis, indicated by the elevated expression of calcium. However, the response pattern of bicarbonate is less definitive.

The only previous studies that have evaluated freshwater mussel adductor muscle hemolymph parameters were reported by Gustafson et al. (2005a, b). In their studies, hemolymph was processed for eight parameters: magnesium, phosphorus, ammonia, protein, sodium, potassium, chloride and calcium. Our hemolymph profile included seven of the eight parameters used by the previous study; we did not include ammonia in our analysis because of the challenge of preserving ammoniated compounds in hemolymph

98

samples given the inherent volatility of those compounds (Shay Bush, Chief Medical

Technologist, University of Georgia, College of Veterinary Medicine, personal communication) and the travel distances (>300 km) between our collection sites and analytical facilities. Of the remaining seven parameters, only three (calcium, phosphorous, and magnesium) were consistently detectable in our study. Protein, sodium, potassium and chloride were always below detectable limits for the three species in our study.

When evaluating parameters as potential biomarkers, it is necessary to evaluate the response of the chosen parameters in a controlled environment, yet one that also represents realistic conditions that may be experienced in the field. Thermal stress is a highly pertinent condition that is occurring with increased frequency as our planet enters a period of substantial climate change. The temperatures used in our study are environmentally relevant for the southeastern United States (Dyar and Alhadeff 1997,

Mosner 2002); however, our exposures were aerated and thus the animals were not experiencing the hypoxic conditions that often accompany elevated water temperature in field conditions. In wadeable streams, high temperatures are also frequently associated with low flows, which may be another stressor for bivalves. Future studies should strive to more closely mimic natural conditions (e.g. oxygen limitation, low flow, and diel temperature fluctuations).

Effects of hemolymph extraction and tissue biopsy

The Cox proportional hazard analysis calculates a hazard risk, which is proportional to a baseline risk of the control group. There was higher survival among the

99

larger-bodied E. crassidens (>80% in all treatments) than the smaller-bodied V. vibex

(13-52%). The baseline risk of mortality was high in V. vibex, but there was also a significant increase in hazard risk associated with the three treatment groups. These results indicate that high mortality in V. vibex can likely be attributed to a combined influence of the treatment groups and also that the holding conditions within the pond were somewhat more inhospitable to V. vibex than E. crassidens. Naimo and Monroe

(1999) reported a decrease in tissue glycogen of A. plicata after 24 months of holding within a pond compared to an established population in the Mississippi River. These data indicate that pond conditions may not be favorable to riverine freshwater mussels, either in food supply/quality, temperature fluctuations, or lack of flow. In the present study, the statistics showed a significant effect of hemolymph extraction, tissue biopsy, and combined tissue and hemolymph extraction on survival of V. vibex. These treatments appear to add some risk of reduced survival to smaller bodied species.

Our results are consistent with other studies that have evaluated the effects of tissue and hemolymph extraction on the survival of large bodied species. Naimo et al.

(1998) found no difference in survival of A. plicata after foot tissue biopsy, consistent with Berg et al. (1995) who found no change in survival of Q. quadrula and A. ligamentina after mantle tissue biopsy. Hemolymph extraction did not affect the survival of E. complanata after a single sampling event or after being repeatedly sampled three times over the course of five months (Gustafson et al. 2005a).

100

Adductor muscle collection site variation

Our comparison of hemolymph parameters between anterior and posterior adductor muscles indicated that expression of enzymes may differ between the two locations, but the remainder of the parameters were consistent between the adductors

(Table 4.5). The labile nature of enzymes often results in high variability among individuals. While the anterior adductor muscle is more easily accessible in most species, this site may not provide sufficient volumes of hemolymph in smaller individuals. Therefore, we recommend reporting the specific adductor location from which hemolymph samples were obtained.

Conclusions

The creation of informative, specific biomarkers is a multistep process that necessitates evaluating the response of the chosen biomarkers under controlled conditions in the laboratory. The present study suggests that ALT, AST, bicarbonate, calcium, magnesium, glucose and tissue glycogen may respond to thermal stress in a predictable manner and thus hold promise for use in wild mussels. Future studies should be performed in the field to evaluate if these parameters are informative. We have determined that the hemolymph extraction and tissue biopsy techniques do not have a substantial negative effect on the long-term survival of large-bodied freshwater mussel species, but there does appear to be an increase in risk for small-bodied mussel species.

However, future studies are needed to assess the effects of tissue extraction in more natural and realistic environments. These tools may help to inform managers and policy

101

makers of the sub-lethal effects of low-flow and thermal stress events in habitats that are substantially affected by anthropogenic modifications.

ACKNOWLEDGEMENTS

Funding for this research was provided by the U.S. Geologic Survey. The authors thank

B. Carswell, Z. DeWolf, J. Dycus, M. Fritts, C. Shea, and J. Wisniewski for assistance in the field, and S. Bush of the University of Georgia, College of Veterinary Medicine for processing the hemolymph samples. The Georgia Cooperative Fish and Wildlife

Research Unit is jointly supported by the University of Georgia, Georgia Department of

Natural Resources, U.S. Geological Survey, U.S. Fish and Wildlife Service, and Wildlife

Management Institute. Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

102

REFERENCES

An, M.I. and C.Y. Choi. 2010. Activity of antioxidant enzymes and physiological

responses in the ark shell, Scapharca broughtonii, exposed to thermal and

osmotic stress: effects on hemolymph and biochemical parameters. Comparative

Biochemistry and Physiology B 155: 34-42.

Archambault, J.M. 2012. Thermal sensitivity of freshwater mussels: incorporating

benthic ecology into laboratory mesocosm experiments. MSc. Thesis. North

Carolina State University. 123 p.

Berg, D.J., W.R. Haag, S.I. Guttman, and J.B. Sickel. 1995. Mantle biopsy: a technique

for nondestructive tissue-sampling of freshwater mussels. Journal of the North

American Benthological Society 14: 577-581.

Blaise, C. and F. Gagne. 2009. Bioassays and Biomarkers, Two Pillars of Ecotoxicology:

Past, Present and Prospective Uses. Fresenius Environmental Bulletin 18:135-

139.

Byrne, R.A., B.N. Shipman, N.J. Smatresk, T.H. Dietz, and R.F. McMahon. 1991. Acid-

base balance during emergence in the freshwater bivalve Corbicula fluminea.

Physiological Ecology 64(3):748-766.

Byrne R.A. and R.F. McMahon. 1991. Acid-base and ionic regulation, during and

following emersion, in the freshwater bivalve, Anodonta grandis simpsoniana

(Bivalvia: Unionidae). Biological Bulletin 181: 289-297.

Cheng, W., I. Hsiao, C. Hsu, and J. Chen. 2004. Change in water temperature on the

immune response of Taiwan abalone Haliotis diversicolor supertexta and its

susceptibility to Vibrio parahaemolyticus.

103

Cummings, K.S. and D.L. Graf. 2010. Mollusca: Bivalvia in J.H. Thorp and A.P. Covich

(eds), Ecology and Classification of North American Freshwater Invertebrates,

third edition. Elsevier Publishing, Canada. 1021 p.

De Zwaan, A. and D.I. Zandee. 1972. The utilization of glycogen and accumulation of

some intermediates during anaerobiosis in Mytilus edulis L. Comparative

Biochemistry and Physiology 43B: 47-54.

De Zwaan, A. and T.C.M. Wijsman. 1976. Anaerobic metabolism in Bivalvia (Mollusca).

Comparative Biochemistry and Physiology 54B: 313-324.

Douglass, W.R. and H.H. Haskin. 1976. Oyster-MSX interactions: alterations in

hemolymph enzyme activities in Crassostrea virginica during the course of

Minchinia nelsoni disease development. Journal of Invertebrate Pathology 27:

317-323.

Dyar, T.R. and S.J. Alhadeff. 1997. Stream-temperature characteristics in Georgia. USGS

96-4203, Atlanta, GA.

Förlin, L., C. Haux, T. Anderson, P.E. Olsson, and Å. Larsson. 1986. Physiological

methods in fish toxicology: Laboratory and field studies. In S. Nilsson and S.

Homgren (eds) Fish Physiology: Recent Advances, pp. 158-169. London: Croom

Helm.

Gagne, F., C. Blaise, I. Aoyama, R. Luo, C. Gagnon, Y. Couillard, P. Campbell, and M.

Salazar. 2002. Biomarker study of a municipal effluent dispersion plume in two

species of freshwater mussels. Environmental Toxicology 17:149-159.

104

Giesy, J.P. and R.L. Graney. 1988. A review of selected clinical indicators of stress-

induced changes in aquatic organisms. In M.S. Evans (ed), Toxic Contaminants

and Ecosystem Health: a Great Lakes Focus. pp. 160-201. John Wiley & Sons.

Golladay, S.W., P.Gagnon, M. Kearns, J.M. Battle, and D. W. Hicks. 2004. Response of

freshwater mussel assemblages (Bivalvia: Unionidae) to a record drought in the

Gulf coastal plain of southwestern Georgia. Journal of the North American

Benthological Society 23(3): 494-506.

Goss, G.G. and C.M. Wood. 1988. The effects of acid and acid/aluminium expsoure on

circulating plasma cortisol levels and other blood parameters in the rainbow trout,

Salmo gairdneri. Journal of Fish Biology 32: 63-76.

Gustafson, L.L., M.K. Stoskopf, A.E. Bogan, W. Showers, T.J. Kwak, S. Hanlon, and

J.F. Levine. 2005a. Evaluation of a nonlethal technique for hemolymph collection

in Elliptio complanata, a freshwater bivalve (Mollusca : Unionidae). Diseases of

Aquatic Organisms 65:159-165.

Gustafson, L.L., M.K. Stoskopf, W. Showers, G. Cope, C. Eads, R. Linnehan, T.J. Kwak,

B. Andersen, and J.F. Levine. 2005b. Reference ranges for hemolymph

chemistries from Elliptio complanata of North Carolina. Diseases of Aquatic

Organisms 65:167-176.

Haag, W.R., D.J. Berg, D.W. Garton, and J.L. Farris. 1993. Reduced survival and fitness

in native bivalves in response to rouling by the introduced zebra mussel

(Dreissena polymorpha) in western Lake Erie. Canadian Journal of Fisheries and

Aquatic Sciences 50:13-19.

105

Haag, W.R., and M.L. Warren. 2008. Effects of severe drought on freshwater mussel

assemblages. Transactions of the American Fisheries Society 137:1165-1178.

Haag, W.R. 2012. North American Freshwater Mussels: Natural History, Ecology, and

Conservation. Cambridge University Press, New York, New York.

Hagerman, L. and A. Szaniawska. 1991. Ion regulation under anoxia in the brackish

water isopod Saduria (Mesidotea) entomon. Ophelia 33(2): 97-104.

Hall, M.R. and E.H. van Ham. 2007. The effects of different types of stress on blood

glucose in the giant tiger prawn Penaeus monodon. Journal of the World

Aquaculture Society 29(3): 290-299.

Handy, R.D. and M.H. Depledge. 1999. Physiological responses: their measurement and

use as environmental biomarkers in ecotoxicology. Ecotoxicology 8: 329-349.

Lacoste, A., S.K. Malham, F. Gelebart, A. Cueff, and S.A. Poulet. 2002. Stress-induced

immune changes in the oyster Crassostrea gigas. Developmental and Comparative

Immunology 26:1-9.

Li, Y., J.G. Qin, C.A. Abbott, X.X. Li, and K. Benkendorff. 2007. Synergistic impacts of

heat shock and spawning on the physiology and immune health of Crassostrea

gigas: an explanation for summer mortality in Pacific oysters. American Journal

of Physiology-Regulatory Integrative and Comparative Physiology 293:R2353-

R2362.

Liu, S.L., X.L. Jiang, X.K. Hu, J. Gong, H. Hwang, and K. Mai. 2004. Effects of

temperature on non-specific immune parameters in two scallop species:

Argopecten irradians (Lamarck 1819) and Chlamys farreri (Jones & Preston

1904). Aquaculture Research 35:678-682.

106

Malley, D.F., J.D. Huebner, and K. Donkersloot. 1988. Effects on ionic composition of

blood and tissues of Anodonta grandis grandis (Bivalvia) of an addition of

aluminum and acid to a lake. Archives of Environmental Contamination and

Toxicology 17: 479-491.

McMahon, R.F. and A.E. Bogan. 2001. Mollusca: Bivalvia. in J.H. Thorp and A.P.

Covich (eds) Ecology and Classification of North American Freshwater

Invertebrates, 2nd edition. Academic Press, San Diego, GA, p. 331-429.

Michaelidis, B., D. Haas, and M.K. Grieshaber. 2005. Extracellular and intracellular acid-

base status with regard to the energy metabolism in the oyster Crassostrea gigas

during exposure to air. Physiological and Biochemical Zoology 78(3): 373-383.

Morris, S. and S.L. Butler. 1996. Hemolymph respiratory gas, acid-base, and ion status of

the amphibious purple shore crab Leptograpsus variegatus (Fabricius) during

immersion and environmental hypoxia. Journal of Crustacean Biology 16(2): 253-

266.

Mosner, M.S. 2002. Stream-aquifer relations and the potentiometric surface of the upper

Floridan aquifer in the lower Apalachicola–Chattahoochee–Flint river basin in

parts of Georgia, Florida, and Alabama, 1999–2000. U.S. Geological Survey

Water-Resources Investigations Report 02-4244. 53 p.

Naimo, T.J., E.D. Damschen, R.G. Rada, and E.M. Monroe. 1998. Nonlethal evaluation

of the physiological health of unionid mussels: methods for biopsy and glycogen

analysis. Journal of the North American Benthological Society 17:121-128.

107

Naimo, T.J. and E.M. Monroe. 1999. Variation in glycogen concentrations within mantle

and foot tissue in Amblema plicata: implications for tissue biopsy sampling. Am

Malacol Bull. 15:51-56.

Nikinmaa, M.T. Jaervenpaeae, K. Westman, and A. Soivio. 1985. Effects of hypoxia and

acidification on the haemolymph pH value and ion concentrations in the

freshwater crayfish (Astacus astacus L.). Finnish Fisheries Research 5: 17-22.

Patterson, M.A., B.C. Parker, and R.J. Neves. 1999. Glycogen concentration in the

mantle tissue of freshwater mussels (Bivalvia : Unionidae) during starvation and

controlled feeding. American Malacological Bulletin 15:47-50.

Pandolfo, T.J., W.G. Cope, C. Arellano, R.B. Bringolf, M.C. Barnhart, and E. Hammer.

2010. Upper thermal tolerances of early life stages of freshwater mussels. Journal

of the North American Benthological Society 29:959-969.

Park, M.S., P.G. Jo, Y.K. Choi, K.W. An, and C.Y. Choi. 2009. Characterization and

mRNA expression of Mn-SOD and physiological responses to stresses in the

Pacific oyster Crassostrea gigas. Marine Biology Research 5(5): 451-461.

Peterson, J.T., J.M. Wisniewski, C.P. Shea, and C.R. Jackson. 2011. Estimation of mussel

population response to hydrologic alteration in a southeastern U.S. stream.

Environmental Management 48: 109-122.

Pynnönen, K. 1990. Physiological responses to severe acid stress in four species of

freshwater clams (Unionidae). Archives of Environmental Contamination and

Toxicology 19(4): 471-478.

108

Pynnönen, K. 1994. Hemolymph gases, acid-base status, and electrolyte concentration in

the freshwater clams Anodonta anatina and Unio tumidus during exposure to and

recovery from acidic conditions. Physiological Zoology 67(6):1544-1559.

R Core Team (2012). R: A language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0,

URL http://www.R-project.org/.

Rugel, K., C.R. Jackson, J.J. Romeis, S.W. Golladay, D.W. Hicks, and J.F. Dowd. 2012.

Effects of irrigation withdrawals on streamflows in a karst environment: lower

Flint River basin, Georgia, USA. Hydrological Processes 26: 523-534.

Shea, C.P., J.T. Peterson, M.J. Conroy, and J.M. Wisniewski. 2013. Evaluating the

influence of land use, drought and reach isolation on the occurance of freshwater

mussel species in the lower Flint River Basin, Georgia (U.S.A.). Freshwater

Biology 58: 382-395.

Sherry, J.P. 2003. The role of biomarkers in the health assessment of aquatic ecosystems.

Aquatic Ecosystems Health and Management 6(4): 423-440.

Shugart, L.R., J.F. McCarthy, and R.S. Halbrook. 1992. Biological markers of

environmental and ecological contamination: an overview. Risk Analysis 12:

353-360.

Vaughn, C.C., S.J. Nichols, and D.E. Spooner. 2008. Community and foodweb ecology

of freshwater mussels. Journal of North American Benthological Society 27:409–

423.

109

Waller, D.L., J.J. Rach, and J.A. Luoma. 1998. Acute toxicity and accumulation of the

piscicide 3-trifluoromethyl-4-nitrophenol (TFM) in freshwater mussels (Bivalvia:

Unionidae). Ecotoxicology 7: 113-121.

Williams, J.D., A.E. Bogan, and J.T. Garner. 2008. Freshwater mussels of Alabama and

the Mobile Basin in Georgia, Mississippi and Tennessee. The University of

Alabama Press, Tuscaloosa, Alabama.

Xue, Q.G. and T. Renault. 2000. Enzymatic activities in European flat oyster, Ostrea

edulis, and Pacific oyster, Crassostrea gigas, hemolymph. Journal of Invertebrate

Pathology 76:155-163.

110

Table 4.1. Water quality parameters (mean ± SE; n = 35 per temperature) measured during thermal stress exposures of Elliptio crassidens, Villosa vibex, and Villosa lienosa.

Target Actual Conductivity Dissolved temp (°C) temp (°C) (mS/cm) Oxygen (mg/L) pH E. crassidens

25 23.7 ± 0.08 0.13 ± 0.01 6.23 ± 0.10 7.21 ± 0.04

30 29.4 ± 0.04 0.14 ± 0.01 6.13 ± 0.03 7.33 ± 0.02

35 34.6 ± 0.03 0.14 ± 0.01 5.65 ± 0.02 7.43 ± 0.01

V. vibex and V. lienosa

25 24.9 ± 0.09 0.11 ± 0.01 8.92 ± 0.05 7.21 ± 0.03

30 29.9 ± 0.06 0.11 ± 0.01 7.89 ± 0.03 7.46 ± 0.01

35 34.1 ± 0.05 0.13 ± 0.01 7.37 ± 0.05 7.62 ± 0.02

111

Table 4.2. Hemolymph and tissue parameter values for three mussel species (Elliptio crassidens, Villosa vibex, and Villosa lienosa) exposed to elevated water temperatures (25, 30 35°C) for 7 days. Hemolymph parameters included alanine aminotransferase (ALT), aspartate aminotransferase (AST), glucose, calcium, magnesium, alkaline phosphatase (AlkPhos), and bicarbonate. Tissue parameters included glycogen. Values are means ± SE; n = 8 at day 0 and n = 5 per temperature per day for all species.

Trial Glucose Calcium Magnesium AlkPhos Bicarbonate Glycogen

Day Temp ALT (U/L) AST (U/L) (mg/dl) (mmol/L) (mg/dl) (U/L) (mmol/L) (mg/g)

E. crassidens

0 25 5.3 ± 0.92 10.4 ± 2.03 2.1 ± 0.30 18.0 ± 1.01 1.9 ± 0.14 8.8 ± 4.1 4.1 ± 0.35 81.8 ± 5.7

3 25 9.0 ± 2.05 15.8 ± 4.93 2.6 ± 0.24 16.7 ± 0.74 2.0 ± 0.14 11.0 ± 2.2 3.8 ± 0.20 74.2 ± 4.0

3 30 8.0 ± 2.07 19.0 ± 3.96 3.2 ± 0.58 15.7 ± 1.08 1.4 ± 0.15 22.8 ± 5.6 3.0 ± 0.32 73.0 ± 4.2

3 35 11.2 ± 1.69 22.6 ± 3.91 3.8 ± 0.58 17.6 ± 0.75 1.8 ± 0.12 19.6 ± 2.7 2.8 ± 0.20 75.2 ± 3.7

7 25 7.2 ± 1.39 14.4 ± 3.50 2.4 ± 0.40 17.0 ± 0.50 1.7 ± 0.06 20.6 ± 6.2 3.8 ± 0.20 75.0 ± 6.9

7 30 10.8 ± 2.87 21.0 ± 5.87 2.6 ± 0.40 15.8 ± 0.62 1.6 ± 0.19 16.8 ± 1.8 3.2 ± 0.20 76.1 ± 3.0

7 35 7.4 ± 1.60 12.2 ± 3.34 3.6 ± 0.24 20.4 ± 1.95 2.0 ± 0.22 18.4 ± 4.5 3.2 ± 0.20 67.4 ± 4.9

112

V. vibex

0 25 6.3 ± 1.18 14.3 ± 2.52 2.0 ±0.19 11.0 ± 0.57 2.7 ± 0.40 0.8 ± 0.09 4.9 ± 0.23 71.1 ± 2.9

3 25 9.6 ± 2.34 17.0 ± 3.71 1.6 ± 0.40 11.3 ± 0.71 1.6 ± 0.45 1.0 ± 0.27 4.0 ± 0.32 49.6 ± 8.5

3 30 11.3 ± 2.87 23.0 ± 4.32 1.3 ± 0.25 9.6 ± 1.12 1.5 ± 0.31 1.5 ± 0.84 3.8 ± 0.25 56.2 ± 7.9

3 35 6.5 ± 0.87 15.0 ± 1.91 2.5 ± 0.65 14.3 ± 0.43 0.9 ± 0.12 1.3 ± 0.60 4.0± 0.01 73.2 ± 5.5

7 25 2.5 ± 0.50 10.5 ± 1.71 2.3 ± 0.48 10.6 ± 0.90 1.9 ± 0.26 0.6 ± 0.13 4.5 ± 0.29 88.1 ± 12.8

7 30 6.2 ± 1.51 23.0 ± 6.38 2.3 ± 0.21 12.0 ± 1.07 1.8 ± 0.18 0.8 ± 0.11 3.7 ± 0.21 60.7 ± 6.7

7 35 3.9 ± 0.70 16.3 ± 2.50 3.3 ± 0.29 12.4 ± 0.57 1.5 ± 0.22 1.3 ± 0.18 4.0 ± 0.22 83.0 ± 6.4

V. lienosa

0 25 11.0 ± 1.65 21.1 ± 5.24 1.4 ± 0.20 12.5 ± 0.48 1.5 ± 0.13 2.3 ± 0.64 4.6 ± 0.30 71.2 ± 2.3

3 25 9.8 ± 1.77 21.2 ± 4.35 1.6 ± 0.24 12.8 ± 0.99 1.3 ± 0.16 3.8 ± 1.07 4.6 ± 0.40 70.6 ± 5.2

3 30 8.8 ± 2.24 20.0 ± 4.66 1.6 ± 0.40 13.1 ± 0.73 1.3 ± 0.11 6.2 ± 0.97 4.0 ± 0.32 74.6 ± 6.9

3 35 13.2 ± 3.01 24.2 ± 3.57 2.4 ± 0.51 13.0 ± 0.74 0.6 ± 0.14 5.5 ± 1.95 5.2 ± 0.49 70.2 ± 7.2

7 25 12.3 ± 3.16 26.5 ± 7.79 1.8 ± 0.40 13.2 ± 0.89 1.6 ± 0.25 2.5 ± 0.43 5.3 ± 0.42 69.6 ± 8.3

7 30 8.5 ± 1.20 20.0 ± 3.27 2.5 ± 0.34 12.8 ± 0.92 2.4 ± 0.26 2.0 ± 0.26 4.3 ± 0.21 83.1 ± 5.6

7 35 10.7 ± 1.12 22.2 ± 2.57 3.8 ± 0.31 12.4 ± 0.44 0.8 ± 0.15 3.3 ± 0.61 5.8 ± 0.40 80.2 ± 4.7

113

Table 4.3. Results of ANOVA analysis for freshwater mussel hemolymph parameters that responded to thermal stress. Values were deemed statistically significant at α ≤ 0.05; significant values are denoted in bold. A post-hoc Dunnett’s test was used to determine significant differences from the control (Day 0, 25°C). Asterisks (*) indicates a significant difference (α ≤ 0.05). Parameter abbreviations: alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (AlkPhos).

Day Temp

Parameter df F-value P-value Day 3 Day 7 df F-value P-value 30°C 35°C

Elliptio crassidens

ALT 2 2.70 0.0829 2 0.31 0.7381

Bicarbonate 2 5.54 0.0088 * * 2 4.57 0.0182 * *

Calcium 2 0.99 0.3820 2 4.53 0.0189

Magnesium 2 0.41 0.6695 2 3.18 0.0556 *

Glucose 2 3.43 0.0449 * 2 4.24 0.0235 *

Villosa vibex

AST 2 0.52 0.5977 2 2.86 0.0722

ALT 2 6.81 0.0035 2 2.33 0.1143

114

Bicarbonate 2 8.18 0.0014 * * 2 2.15 0.1332

Calcium 2 0.58 0.5671 2 4.51 0.0191 *

Magnesium 2 7.74 0.0019 * * 2 1.38 0.2654

Glucose 2 5.90 0.0067 2 6.00 0.0063 *

Glycogen 2 3.59 0.0378 2 3.52 0.0401

Villosa lienosa

Bicarbonate 2 2.15 0.1325 2 6.78 0.0034

Magnesium 2 5.37 0.0095 2 19.53 0.0001 *

AlkPhos 2 7.10 0.0027 * 2 0.92 0.4080

Glucose 2 7.87 0.0016 * 2 9.25 0.0006 *

115

Table 4.4. Cox proportional hazard regression analyses for effects of tissue and hemolymph extraction on the long-term survival of Villosa vibex and Elliptio crassidens.

This analysis compares the hazard risk of an action (i.e. tissue and hemolymph extraction) against the baseline risk (control). Values were deemed statistically significant at α ≤ 0.05. Exp (coef) = hazard ratio, SE (coef) = standard error, and upper

95% CI = the upper bound of a one-tailed 95% confidence interval, which represents the extent of the potential increase in risk for each action.

Treatment exp (coef) se (coef) p-value upper 95%

V. vibex

Hemolymph 2.433 0.471 0.059 6.124

Tissue 2.361 0.471 0.068 5.943

Tissue and hemolymph 2.443 0.455 0.050 5.959

E. crassidens

Hemolymph 1.458 0.764 0.622 6.515

Tissue 1.786 0.730 0.427 7.475

Tissue and hemolymph 1.422 0.764 0.645 6.354

116

Table 4.5. Hemolymph parameters measured from the anterior and posterior adductor muscles of Elliptio crassidens (n = 10). Paired t-tests were used to evaluate differences in hemolymph components from the two collection sites. Mean values (± SE) and the associated P-values are shown for six hemolymph parameters. Significant values are indicated in bold.

Hemolymph parameter Anterior Posterior P-value

Aspartate aminotransferase (U l-1) 20.1 ± 1.59 25.6 ± 2.08 0.0198

Alanine aminotransferase (U l-1) 8.3 ± 0.83 10.4 ± 1.05 0.0354

Bicarbonate (mmol l-1) 4.6 ± 0.22 4.7 ± 0.30 0.3434

Alkaline phosphatase (U l-1) 18.5 ± 3.71 14.4 ± 1.70 0.2752

Calcium (mg dl-1) 17.5 ±0.33 18.1 ±0.25 0.0869

Magnesium (mg dl-1) 1.5 ± 0.06 1.6 ± 0.07 0.0738

117

Fig. 4.1. Probability of survival (± SE) for Villosa vibex over 820 days after tissue and hemolymph extraction calculated using the Cox Proportional Hazard Regression. Bars are colored as follows: black= control, light= hemolymph, hatched= tissue, dark gray= tissue and hemolymph combined.

118

Fig. 4.2. Probability of survival (± SE) for Elliptio crassidens over 945 days after tissue and hemolymph extraction calculated using the Cox Proportional Hazard Regression.

Bars are colored as follows: black= control, light= hemolymph, hatched= tissue, dark gray= tissue and hemolymph combined.

119

CHAPTER 5

NON-LETHAL ASSESSMENT OF FRESHWATER MUSSEL RESPONSE TO

CHANGES IN ENVIRONMENTAL FACTORS IN THE LOWER FLINT RIVER

BASIN, GEORGIA, U.S.A.4

4 Fritts, A.K., J.T. Peterson, and R.B. Bringolf. To be submitted to Comparative

Biochemistry and Physiology.

120

ABSTRACT

The Southeastern U.S. is home to a diverse assemblage of freshwater mussels. Threats to this group include habitat degradation, pollution, and alterations to natural flow regimes.

Many of the aforementioned threats are of serious concern in the Flint River Basin in southwest Georgia, a system highly impacted by agricultural water usage. The Flint

Basin is home to a diverse assemblage of aquatic organisms, including five federally listed mussel species. Due to the imperiled status of these mussels, the development of effective nonlethal biomonitoring techniques is imperative. Changes in hemolymph chemistry profiles and tissue glycogen are potential biomarkers for non-lethally monitoring stress in freshwater mussels and to evaluate how these potential biomarkers respond to environmental changes, Villosa vibex, Villosa lienosa and Elliptio crassidens were sampled at five field sites in the Lower Flint River Basin over the course of two years. Hemolymph was extracted from the adductor muscle sinus and was analyzed with blood chemistry analyzer for a suite of five parameters: alanine aminotransferase, aspartate aminotransferase, bicarbonate, calcium, and magnesium. Tissue biopsies were extracted from the foot tissue of the mussels and processed for glycogen with a colorimetric assay. We used hierarchical linear models to evaluate the relationships between variation in the biomarkers and environmental factors. Our data indicate that the response of the hemolymph and tissue parameters was strongly related to stream discharge. The modeling results also illustrate that all hemolymph and tissue responses vary substantially between individuals of different size, sex, and species. Our study has provided a framework methodology for the emerging study of physiological biomarkers

121

of freshwater mussel populations and illustrates the value of hierarchical modeling techniques to account for the inherent complexity of aquatic ecosystems.

122

INTRODUCTION

Aquatic ecosystems are under increasing pressure from the demands of a vast spectrum of different stakeholders and changing climate patterns. Anthropogenic demands for the use of freshwater resources have hastened changes in both water quantity and quality in many rivers (Freeman et al. 2012). Changes in land use have exacerbated these depletions and degradations of freshwater resources. Agricultural water use, mostly for irrigation, constitutes ~80% of all freshwater consumption (Postel 1997), while the demands for municipal water withdrawals are also increasing with the growth of human populations. Changes in climate patterns will create even greater uncertainty about water availability (Araujo and Rahbek 2006). These changes, coupled with habitat alterations and fragmentation, create altered flow regimes in natural systems that can lead to disruptions in ecological functions (Postel 2000).

There is rising concern about how changes in freshwater ecosystems will impact native aquatic biota, particularly those species that are already threatened by previous perturbation (e.g. overharvest, habitat degradation). One faunal group that is of exceptional and growing concern is freshwater mussels (Williams et al. 1993). As primarily sessile organisms, freshwater mussels lack the ability to retreat to refuges when water levels drop or become inhospitable due to elevated temperatures or low levels of dissolved oxygen. The preservation of these imperiled species will require the development of a suite of non-lethal, early warning indicators to identify at-risk populations before species begin to exhibit reductions in growth and survival (Gustafson et al. 2005b). However, the specific biotic and abiotic mechanisms that drive population declines are poorly understood (Haag and Warren 2008, Haag and Williams 2013).

123

With the rising concern about water availability, a number of recent studies have focused on assessing the effects of drought on freshwater mussels. A study conducted in

Georgia documented decreased occupancy and detection of mussel populations with the occurrence of droughts (Shea et al. 2013). Haag and Warren (2008) recorded declines in mussel abundance in Alabama and Mississippi following a prolonged drought.

Interestingly, populations in stream reaches that ceased to flow but retained water exhibited similar decreases in mussel abundance when compared to stream reaches that were almost completely dewatered. This phenomenon indicates that mussel populations are at high risk not only from the loss of habitat from stream drying, but also from the secondary effects of drought, such as elevated water temperatures and decreases in dissolved oxygen (Haag and Warren 2008).

The lower Flint River basin (LFRB), which encompasses 13,952 km2 in southwest

Georgia, is located within the Fall Line Hills and Dougherty Plain districts of the Coastal

Plain physiographic province (Mosner 2002; Fig. 5.1). The Flint River joins the

Chattahoochee River at Lake Seminole to form the Apalachicola River, which flows to the Gulf of Mexico. This river system has been embroiled in a tri-state water dispute among the states of Georgia, Alabama, and Florida for more than 20 years (Ruhl 2005).

Water in this system is drawn upon heavily for use by municipalities, industry, and agriculture. The LFRB, in particular, supports a substantial amount of agricultural production. During the growing season from April-September, ~90% of the water consumption within this region is used for irrigation of row crops (Couch and McDowell

2006). The inception of center pivot irrigation in the 1970’s led to a >100% increase in water use between 1970-1976 (Pierce et al. 1984). The increase in irrigation has led to

124

substantial reductions in flows in the Flint River basin, as well as increasing the duration and severity of low flow events (Rugel et al. 2012). The region is typified by karst topography and is underlain by the Upper Floridian Aquifer (Hicks et al. 1987). A hydraulic connection exists between the surface waters in the LFRB and the Upper

Floridian Aquifer and withdrawals from the aquifer for center pivot irrigation leads to reduced flows in many of the springs and tributaries in the LFRB, particularly during the summer months (Torak and Painter 2006, Rugel et al. 2012).

The LRFB historically supported thirty species of freshwater mussels, but two of those species are presumed extirpated from the system (Clench and Turner 1956, Brim

Box and Williams 2000, Williams et al. 2008). This watershed is currently home to five federally listed mussel species: Hamiota subangulata, Pleurobema pyriforme,

Medionidus penicillatus, Amblema neislerii, and Elliptoideus sloatianus. Threats to this faunal group include pollution, habitat degradation/fragmentation, and altered flows in the form of droughts or reduction in flows caused by excessive irrigation withdrawals

(Golladay et al. 2004, Peterson et al. 2011, Haag 2012, Shea et al. 2013). Due to the imperiled nature of these organisms, there is a strong need for biomonitoring of the remaining mussel populations when faced with environmental changes in discharge, water temperature, and dissolved oxygen.

One biomonitoring tool that has emerged in recent years is the tracking of specific biological processes (biomarkers) and how those processes react to changes in the organism’s habitat (Gagne et al. 2002, Blaise and Gagne 2009). Development of specific and nonlethal biomarkers is the primary challenge to creating an effective biomonitoring plan for endangered mussels, and there is a growing body of literature that

125

is aimed at evaluating how various biotic and abiotic stressors influence the physiological response of mollusks. Prior research has indicated that decreases in tissue glycogen, a primary energy storage molecule, may be a useful indicator of stress in freshwater mussels. Glycogen levels have been shown to decrease under stressful conditions such as starvation (Patterson et al. 1999) and zebra mussel infestation (Haag et al. 1993, Baker and Hornbach 2000). Changes in hemolymph (mussel blood) biochemical profiles have also shown potential as a tool for biomonitoring (Gustafson et al. 2005a). This technique has been utilized to assess the effects of emersion in air (Byrne and McMahon 1991), acidification (Pynnonen 1994), anoxia (Dietz 1974), and thermal stress (Fritts 2013).

The identification of useful biomarkers requires that researchers establish reference values for healthy, unstressed populations (Gustafson et al. 2005b), understand how these parameters differ between species within the same taxonomic group, and evaluate how the parameters respond to environmental changes in a natural setting.

Therefore, our objectives were to 1) develop a biomonitoring program in the Flint River basin that utilized hemolymph and tissue biopsies collected over a broad spatial and temporal scale, 2) evaluate the response of our chosen biomarkers among three mussel species in a natural riverine setting, and 3) through the use of model selection determine which environmental parameters had the strongest influence on changes in the biomarkers in the Flint River basin.

126

METHODS

Site description

Our study was conducted in the lower Flint River Basin (LFRB) in southwest

Georgia. We focused our sampling efforts on five sites that were split between the Fall

Line Hills and the Dougherty Plain districts of the Coastal Plain physiographic province within the watershed (Fig. 5.1). Fall Line Hills sites included Lanahassee Creek upstream of the Hwy 280 crossing, 2 km ESE of Preston in Webster County and

Ichawaynochaway Creek upstream of the CR 51163 crossing, 11 km WNW of Dawson in

Terrell County. Sites within the Dougherty Plain included the Flint River above the State

Route 37 bridge near Newton in Baker County, Cooleewahee Creek downstream of the

Hwy 91 crossing near Newton in Baker County, and Spring Creek in Spring Creek Park,

Colquitt, Miller County. These five sites were each sampled on four to seven occasions from April 2010-June 2011. Discharge and other ancillary water quality parameters, including temperature, dissolved oxygen (DO), hardness, and alkalinity were measured during each sampling event.

Freshwater mussel tissue collection

Three species of mussels were sampled during this study. Villosa vibex and

Villosa lienosa were collected in the tributary sites, while Elliptio crassidens was limited to the mainstem Flint River site. A minimum of ten individuals per species per site were sampled during each field visit. Mussels were held in aerated river water while individuals were processed and samples (tissue and hemolymph) were collected. Each mussel was identified to species and measured with a calipers to the nearest mm (length x

127

height x width). Gravidity status was recorded, as well as the sex of the Villosa species, which were sexually dimorphic. Hemolymph was extracted from the adductor muscle sinus with a 26 gauge needle and a 1-ml syringe (Gustafson et al. 2005a,b, Fritts 2013).

Hemolymph was extracted from the anterior adductor muscle of E. crassidens and from the posterior adductor muscle of both Villosa species. After extraction, the hemolymph was placed into 2.0-mL cryovials and stored in liquid nitrogen for transport to the

Aquatic Science Laboratory at the University of Georgia (UGA). Tissue biopsies were extracted with a 3-mm oval-cupped jaws biopsy forceps (Surgical Direct, DeLand,

Florida, USA) from the foot tissue of all mussels. The tissue biopsies were also placed in

2.0-mL cryovials and stored in liquid nitrogen for transport to UGA. The forceps were thoroughly rinsed with clean water after each biopsy was extracted. After returning to the laboratory, all samples were transferred to a cryogenic freezer (-80°C) for storage until the samples were processed.

Mussel hemolymph samples were processed on a Hitachi 912 Blood Chemistry

Analyzer (Roche Diagnostics) at the UGA College of Veterinary Medicine Clinical

Pathology Laboratory for the following suite of parameters: alanine aminotransferase

(ALT), aspartate aminotransferase (AST), bicarbonate, calcium, and magnesium. This suite was selected based upon previous work by Gustafson et al. (2005a,b) and Fritts

(2013).

Tissue samples were processed for glycogen content at the Aquatic Science

Laboratory at UGA following the method described by Naimo et al. (1998) but modified for measurement on a 96-well plate. Briefly, glassware was washed in an acid bath (10%

HNO3) for eight hours. Calibration standards were created by dissolving 200 mg of

128

powdered oyster glycogen (Sigma-Aldrich) in 100 mL deionized water to make a stock solution of 2000 mg glycogen/L. The stock solution was serially diluted to create standards with the following concentrations: 2000, 1000, 500, 250, 125 mg glycogen/L, plus a control (0 mg/L). As an internal standard, we used homogenized foot tissue from five Lampsilis cardium that were collected from Pool 8 of the Mississippi River in

October 2011. We also created spiked standards by adding a known quantity of internal standard tissue to each of the standard curve concentrations. Each sequence of analyses was accompanied by three procedural blanks, three replicates of each standard curve concentration, three replicates of spiked standards, and three internal standard samples that were each measured in triplicate.

The glycogen assay procedure included a tissue digestion component followed by colorimetric analysis. Calibration standards were added to test tubes at a volume of 250

µL. Tissue samples were defrosted, then a 5-10 mg sample was weighed and placed into a 15 mL test tube. A solution of 30% KOH was added to each test tube at a volume of

100 µL for tissue samples and 500 µL for standard curve samples. All samples were placed into a hot water bath at 100°C for 30 minutes. After heating, the samples were vortexed for 30 seconds and placed into an ice water bath for five minutes. A solution of

95% EtOH was then added to each test tube at a volume of 150 µL for tissues and 750 µL for standards. Samples were vortexed for five seconds then placed into a hot water bath at 100°C for 20 minutes. Samples were either processed immediately after the tissue digestion was complete or stored in a cryogenic freezer (-80°C) until they could be processed.

129

The colorimetric analysis was begun by first diluting all samples to a common volume (7000 µL) with deionized water. Samples were vortexed for 15 seconds, after which a 2 mL aliquot of the solution was placed into a new test tube. For the color development step, 100 µL of 80% phenol was added to the test tube, followed immediately by 5 mL of concentrated sulfuric acid. The test tube was vortexed for 10 seconds and allowed to stand at room temperature (20°C) for a minimum of 30 minutes to obtain maximum color development. Samples were processed by placing 300 µL of sample into a 96-well microplate, which was then read on a SpectraMax M2 plate reader at 490 nm and processed with SoftMax Pro software. Final glycogen content was normalized for wet tissue mass and was reported as mg glycogen per g tissue.

Data analysis

The data were analyzed in R (R Core Team 2012) using linear models to assess the relations between the biological responses (changes in hemolymph parameters and tissue glycogen) and the other measured parameters: discharge, discharge2, physiographic region, DO, temperature, size, sex, season, and species. Three species were coded into the models as binary indicator variables (0,1); Villosa vibex was used as the baseline species. Three different durations of discharge were considered in the models: discharge at time of sampling, average daily discharge for 5 days prior to sampling, and average daily discharge for 15 days prior to sampling. Discharge data from the Flint River mainstem were obtained directly from the U.S. Geological Survey (USGS) gauging station at that site (2353000); discharge data for the four tributaries were calculated using published discharge models and data from surrounding USGS gauging stations (2356000,

130

2357000, 2353500, 2350900; McCargo and Peterson 2010, Dycus 2011). We normalized discharge by watershed area to control the effect of watershed area on stream discharge, since sites occurred in different sized streams. Quadratic terms for each discharge component were always included when discharge was in a candidate model because the biological responses of hemolymph and glycogen were assumed to be nonlinearly related to discharge. All continuous parameters (e.g. discharge, DO, temperature, size) were adjusted into standardized predictors, which were calculated by subtracting the mean and dividing by the standard deviation. Models were created to estimate the relation between changes in the aforementioned parameters on the responses (Table 5.1). All parameters were standardized with a mean of 0 and standard deviation of 1 to facilitate comparisons of parameter estimates for factors that were measured in very different scales. To avoid multicollinearity, relations between predictors were examined using Pearson correlations.

Parameters were not included together in models if r2 > 0.40.

A global model (discharge, discharge2, physiographic region, DO, temperature, size, sex, season, species, and the interaction between these parameters) was fit using linear regression and the residuals were analyzed using ANOVA to assess independence assumptions. The analysis revealed significant dependence among the samples collected at the same site (F5,418 = 15.50, p <0.0001) so hierarchical models were developed to account for the dependence (Byrk and Raudenbush, 1992, Royle and Dorazio 2008).

This technique accounts for dependence in the data by including random effects for lower level units, (i.e., hemolympth and tissue samples) nested within upper-level units (sites).

Different variance structures that included randomly varying intercepts and slopes and covariance between intercept and slopes were analyzed using Akaike’s Information

131

Criteria (AIC) (Akaike 1973) with small sample bias adjustment (AICc) (Huvich and Tsai

1989) of the global model to determine the optimal variance structure. The structure that produced the smallest AICc value was deemed optimal. After implementing a hierarchical linear model, the residuals of the global model were not autocorrelated and the data showed constancy of variance.

An information theoretic approach (Burnham and Anderson 2002) was used to assess the relationships between the model parameters (discharge, discharge2, DO, temperature, size, sex, season, species, and the interaction between these parameters) and the changes in hemolymph AST, ALT, bicarbonate, calcium, magnesium, and tissue glycogen. A global model and 102 additional candidate models were created for each hemolymph parameter and glycogen, each representing a hypothesized effect of local

(sampling site) and landscape-level features on the changes in hemolymph and glycogen

(Table 5.1). Combinations of predictors were created to identify the parameters that best explained the variability in hemolymph and glycogen responses within and among sampling sites. The relative fit of all models were assessed by calculating AICc. Akaike weights were calculated and used to examine the strength of evidence for each model; values range from 0 to 1 with the best fitting models having the largest values (Burnham and Anderson 2002). The number of parameters in each model included fixed and random effects and covariances. A confidence model set was developed to incorporate model selection uncertainty. Only those candidate models whose Akaike weights were within 10% of the largest weight were included in the confidence set of models. This is similar to the 1/8th rule proposed by Royall (1997). We also calculated Akaike importance weights for individual predictors within the confidence set of models to

132

evaluate the relative weight of evidence for the influence of biotic and abiotic factors on the expression of biomarkers. Precision of parameter estimates were assessed using 95% confidence intervals and if the confidence interval overlapped zero, the value was considered too imprecise to determine whether the direction of the effect was positive or negative.

A random effects ANOVA was used to partition variation in the response of hemolymph and glycogen among sites. This technique was used to calculate the intra- class correlation coefficient that describes the inherent variability within and among watersheds. In addition, we estimated the amount of variation that was explained by the candidate models by using the R2 coefficient of determination.

RESULTS

Modeling

Based on the AICc values, the best approximating variance structure for all models included an intercept and slope for discharge that varied randomly among sample sites. The model selection resulted in different model rankings for each biological response (i.e., variation in hemolymph parameters and tissue glycogen). However, discharge, size, sex, and species were included in the confidence set of models for all biological responses.

Alanine aminotransferase

The relative weights for the three durations of discharge, the main effects, and the interactions were calculated for the ALT candidate set of models (Fig. 5.2A). Discharge at the time of sampling had the highest importance weight (0.73) indicating that ALT is a

133

better indicator of near term (immediate) conditions rather than over moderate to long time intervals. The best approximating model for ALT included 21 parameters: season, region, size, sex, species, temperature, DO, discharge, discharge2, and eight 2-way interactions: temperature * size, temperature * sex, temperature * species (V. lienosa), temperature * species (E. crassidens), DO * size, DO * sex, DO * species (V. lienosa), and DO * species (E. crassidens; Table 5.2). This model was 1.2 times more likely than the next model that excluded the effects of region. The confidence set included 6 models

(Table 5.2). Parameter estimates generated from the best approximating model indicated that ALT levels were lower during the summer and that ALT expression increased with increasing temperature (Fig. 5.4), although this parameter estimate was marginally imprecise (Fig. 5.3A). The 95% confidence intervals overlapped zero for the remaining parameters, indicating that the parameter estimates were imprecise and it was not possible to discern the direction of the effect (i.e., positive or negative). The random effect indicated that the relation between discharge and ALT varied substantially among sites by 333%. The R2 for the best approximating ALT model was 31%.

Aspartate aminotransferase

The relative weights for the three durations of discharge, the main effects, and the interactions were calculated for the AST candidate set of models (Fig. 5.2B). The average discharge for 15 days prior to sampling had the highest importance weight (0.64) indicating that AST is a better indicator of long-term conditions rather than over short to moderate time intervals. The best approximating model for AST included 21 parameters: season, region, size, sex, species, temperature, DO, discharge, discharge2, and eight 2- way interactions: temperature * size, temperature * sex, temperature * species (V.

134

lienosa), temperature * species (E. crassidens), DO * size, DO * sex, DO * species (V. lienosa), and DO * species (E. crassidens; Table 5.3). This model was 2.1 times more likely than the next model that excluded the effects of region. The confidence set included 7 models (Table 5.3). Parameter estimates generated from the best approximating model indicated that AST expression increased with increasing temperature and decreased with increasing freshwater mussel size. AST levels appeared lower during the summer, although this parameter estimate was marginally imprecise

(Fig. 5.3B). The remaining parameters had 95% confidence intervals that overlapped zero, indicating that the parameter estimates were imprecise and it was not possible to discern the direction of the effect. The random effect indicated that the relation between discharge and AST varied substantially among sites by 567%. The R2 for the best approximating AST model was 28%.

Bicarbonate

The relative weights for the three durations of discharge, the main effects, and the interactions were calculated for the bicarbonate candidate set of models (Fig. 5.2C). The average discharge for 15 days prior to sampling had the highest importance weight (0.95) indicating that bicarbonate is a better indicator of long-term conditions rather than over short to moderate time intervals. The best approximating model for bicarbonate included

12 parameters: season, size, sex, species, temperature, DO, discharge, and one 2-way interaction: discharge * season (Table 5.4). This model was 2.2 times more likely than the next model that incorporated the effects of region and discharge2. The confidence set included 3 models (Table 5.4). Parameter estimates generated from the best approximating model indicated that bicarbonate levels were higher during the summer

135

and that bicarbonate expression decreased with increasing length, although this parameter estimate was marginally imprecise (Fig. 5.3C). Bicarbonate levels from E. crassidens were lower than levels in V. vibex and bicarbonate expression also decreased with increasing DO. The 95% confidence intervals overlapped zero for the remaining parameters, indicating that the parameter estimates were imprecise and it was not possible to discern the direction of the effect. The random effect indicated that the relation between discharge and bicarbonate varied substantially among sites by 11,078%.

The R2 for the best approximating bicarbonate model was 65%.

Calcium

The relative weights for the three durations of discharge, the main effects, and the interactions were calculated for the calcium candidate set of models (Fig. 5.2D). The average discharge for 15 days prior to sampling had the highest importance weight (0.96) indicating that calcium is a better indicator of long-term conditions rather than over short to moderate time intervals. The best approximating model for calcium included 13 parameters: size, sex, species, discharge, discharge2, and four 2-way interactions: discharge * size, discharge * sex, discharge * species (V. lienosa), and discharge * species (E. crassidens; Table 5.5). This model was 1.3 times more likely than the next model that incorporated the effects of region. The confidence set included 3 models

(Table 5.5). Calcium levels from both E. crassidens and V. lienosa were higher than levels in V. vibex (Fig. 5.5). Parameter estimates generated from the best approximating model indicated that calcium expression increased with increasing discharge, although this parameter estimate was marginally imprecise (Fig. 5.3D). The remaining parameters had 95% confidence intervals that overlapped zero, indicating that the parameter

136

estimates were imprecise and it was not possible to discern the direction of the effect. The random effect indicated that the relation between discharge and calcium varied among sites by 110%. The R2 for the best approximating calcium model was 32%.

Magnesium

The relative weights for the three durations of discharge, the main effects, and the interactions were calculated for the magnesium candidate set of models (Fig. 5.2E). The average discharge for 5 days prior to sampling had the highest importance weight (0.60) indicating that magnesium is a better indicator of moderate-term conditions. The best approximating model for magnesium included 9 parameters: region, size, sex, species, and discharge (Table 5.6). This model was 1.6 times more likely than the next model that incorporated the effects of an interaction between region and discharge. The confidence set included 7 models (Table 5.6). Parameter estimates generated from the best approximating model indicated that magnesium expression was strongly and positively associated with region; magnesium levels were elevated in the Fall Line Hills district.

Magnesium expression was higher in larger sized individuals and lower in females (Fig.

5.3E). The species and discharge parameters had 95% confidence intervals that overlapped zero, indicating that the estimates were imprecise and it was not possible to discern the direction of the effect. The random effect indicated that the relation between discharge and magnesium varied substantially among sites by 4507%. The R2 for the best approximating magnesium model was 60%.

Tissue glycogen

The relative weights for the three durations of discharge, the main effects, and the interactions were calculated for the glycogen candidate set of models (Fig. 5.2F). The

137

average discharge for 15 days prior to sampling had the highest importance weight (0.64) indicating that glycogen is a better indicator of long-term conditions rather than over short to moderate time intervals. The best approximating model for glycogen included 20 parameters: region, size, sex, species, temperature, DO, discharge, discharge2, and eight

2-way interactions: temperature * size, temperature * sex, temperature * species (V. lienosa), temperature * species (E. crassidens), DO * size, DO * sex, DO * species (V. lienosa), and DO * species (E. crassidens; Table 5.7). This model was 1.5 times more likely than the next model that incorporated the effects of season and excluded the effects of region. The confidence set included 7 models (Table 5.7). Parameter estimates generated from the best approximating model indicated that glycogen expression was strongly associated with region; glycogen levels were lower in the Fall Line Hills district

(Fig. 5.3F). The remaining parameters had 95% confidence intervals that overlapped zero, indicating that the parameter estimates were imprecise and it was not possible to discern the direction of the effect. The random effect indicated that the relation between discharge and glycogen varied among sites by 106%. The R2 for the best approximating glycogen model was 24%.

DISCUSSION

The inherent complexity of natural environments is difficult to characterize

(Wolfe 1996), thus necessitating the use of modeling to allow researchers to estimate the effects of various biotic and abiotic factors on observable phenomena (Burnam and

Anderson 2002). Our data indicate that the response of the hemolymph and tissue parameters was strongly related to discharge. The models also provided evidence

138

regarding which discharge duration was the best fit for each hemolymph and tissue parameter. Shifts in alanine aminotransferase (ALT) showed the largest relations with discharge at the time of sampling. Shifts in magnesium were most explained by the average discharge for five days prior to sampling. The remaining parameters (AST, bicarbonate, calcium, and glycogen) showed the strongest relations with average flows for 15 days prior to sampling. These data provide potential insights into the response time of these prospective biomarkers in a natural riverine setting.

The modeling results indicate that all hemolymph and tissue responses vary substantially among individuals of different size, sex, and species. The confidence set of models for all potential biomarkers had importance weights of 1.0 for size, sex, and species which indicates that future studies will need to continue to evaluate the different ways in which these biomarkers respond among different species/tribes, between males and females, and for individuals of different ages/sizes. Gustafson et al. (2005b) found calcium and bicarbonate to be positively correlated with shell length, while glucose,

ALT, and AST were negatively correlated with length. The current study also found

AST to be negatively related to length, but ALT, bicarbonate, and calcium did not show any precise relationship with length.

The two enzymes (ALT and AST) showed similar responses to biotic and abiotic changes. Both enzymes were lower in the summer, but showed increased expression with increasing water temperature. In other invertebrate and vertebrate species, elevated levels of these two enzymes have been used as an indicator of tissue damage (An and

Choi 2010). If freshwater mussels are becoming stressed by high water temperatures, it is conceivable that there would be increased expression of these enzymes during the stress

139

event. Expression of ALT was higher in V. lienosa than for V. vibex, possibly because these species were most exposed to significant changes in flows (i.e., dewatering) in the smaller tributary environments as opposed to mainstem sites. Additionally, AST expression was lower with an increase in the size of an individual, indicating that larger animals may be less susceptible to stress events than smaller individuals.

Bicarbonate and calcium are closely related in the respiration processes of freshwater mollusks (Byrne and McMahon 1991). If mussels close their valves to escape from stressful situations, such as desiccation, low DO, or elevated water temperatures, the lack of respiration may lead to acidosis of the circulatory fluids. A mechanism by which mollusks counteract this acidosis is through the use of bicarbonate and calcium ions to buffer the change in hemolymph pH (Pynnönen 1994). When mussels are exposed to a stressful event that disrupts respiration, it is likely that elevated levels of bicarbonate may be an indicator of that stress. Our best approximating model for bicarbonate indicated that bicarbonate levels were higher in the summer and lower with an increase in the size of the individual. Additionally, bicarbonate expression decreased with increasing levels of DO. These results likely indicate that smaller or younger individuals may be more susceptible to stress, particularly hypoxic events related to de-watering and high temperatures in southern streams.

Variation in magnesium levels were most closely tied to physiographic region, with individuals in the Fall Line Hills district exhibiting higher levels of magnesium than mussels in the Dougherty Plain. The Fall Line Hills district contains unique geographic features and is underlain by the Ocala Limestone (Hicks et al. 1987) and the differences in geomorphology are likely the source for contrasts in the levels of bioavailable

140

magnesium. Additionally, hemolymph magnesium levels were higher in larger individuals and lower for females compared to males.

The best approximating model for glycogen indicated that glycogen levels were lower in the Fall Line Hills district, indicating that food availability may be more limiting in that region. An alternative hypothesis would be that existence in the Fall Line Hills requires more energy use, thereby limiting the amount of nutrition/energy that can be stored as glycogen. Patterson et al. (1999) documented a decrease in tissue glycogen as a result of starvation. Additionally, our model indicated that glycogen content was predicted to be lower in larger individuals and to decrease as DO levels increased.

Among the best approximating models for the six different biological responses,

36-76% of the variability remains unexplained by the top model. The best approximating models for bicarbonate and magnesium account for a greater proportion of the variability, with only 36-40% remaining unexplained. The remaining responses (ALT, AST, calcium, and glycogen) have substantial amounts of variability that remain unexplained

(68-76%). The factors that might explain this remaining variability are unknown, but they may include microhabitat variables (e.g. sediment size, water velocity), exposure to previous stressors (e.g. low DO, elevated temperatures), or the duration of the stress event.

Growing evidence indicates that foot tissue biopsies and hemolymph extraction from the adductor muscle sinuses can be collected in a non-lethal manner (Naimo et al.

1998, Gustafson et al. 2005a,b, Fritts 2013), thereby opening opportunities to create biomonitoring plans for imperiled populations of freshwater mussels. Gustafson et al.

(2005a,b) were the first to make strides in advancing the use of non-lethal methods for

141

evaluating hemolymph parameters in freshwater mussels. Their collection of baseline data for Elliptio complanata in North Carolina was the first attempt to take this technique to the field. Our study represents the next steps in studying how these parameters respond to changes in natural systems. Despite our use of sensitive and nuanced modeling techniques, the inherent complexity of natural systems will continue to limit the capability to precisely model the effects of specific variables on the relative condition of freshwater mussels. The factors of discharge, size, sex, and species were most commonly found to affect the biological responses in our models and we recommend that future research into the effects of drought and stress should include the use of ALT, AST, bicarbonate, and calcium. Our collection of ancillary physical habitat data was limited by the scope of this research; however, future research must consider the effects of additional abiotic and biotic variables that could account for the variance that remains unexplained by our models. Nonetheless, this study has provided a framework methodology for the emerging study of physiological biomarkers of freshwater mussel populations and we strongly recommend that future researchers consider the use of similar modeling techniques to account for the inherent complexity of aquatic ecosystems.

ACKNOWLEDGEMENTS

Funding for this research was provided by the U.S. Geologic Survey. The authors thank

J. Dycus, M. Fritts, P. Hazelton, C. Shea, and J. Wisniewski for assistance in the lab and field. The Georgia Cooperative Fish and Wildlife Research Unit is jointly supported by the University of Georgia, Georgia Department of Natural Resources, U.S. Geological

142

Survey, U.S. Fish and Wildlife Service, and Wildlife Management Institute. Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

143

REFERENCES

Akaike, H. 1973. Information theory and an extension of the maximum likelihood

principle. In B.N. Petrov and F Casaki (eds), Second International Symposium on

Information Theory. Pp. 267-281. Akademiai Kiado, Budapest, Hungary.

An, M.I. and C.Y. Choi. 2010. Activity of antioxidant enzymes and physiological

responses in the ark shell, Scapharca broughtonii, exposed to thermal and

osmotic stress: effects on

hemolymph and biochemical parameters. Comp Biochem Phys B. 155:34-42.

Araujo, M.B. and C. Rahbek. 2006. How does climate change affect biodiversity?

Science 313: 1396-1397.

Baker, S.M. and D.J. Hornbach. 2000. Physiological status and biochemical composition

of a natural population of unionid mussels (Amblema plicata) infested by zebra

mussels (Dreissena polymorpha). American Midland Naturalist 143:443-452.

Blaise, C. and F. Gagne. 2009. Bioassays and Biomarkers, Two Pillars of Ecotoxicology:

Past, Present and Prospective Uses. Fresenius Environmental Bulletin 18:135-

139.

Brim Box, J. and J.D. Williams. 2000. Unionid mollusks of the Apalachicola Basin in

Alabama, Florida, and Georgia. Bulletin of the Alabama Natural History Survey

21: 1-143.

Burnham, K.P. and D.R. Anderson. 2002. Model Selection and Inference: An

Information -Theoretic Approach, 2nd edition. Springer-Verlag, New York.

Byrk, A.S. and S.W. Raudenbush. 1992. Hierarchical linear models: applications and data

analysis methods. Sage, Thousand Oaks, CA.

144

Byrne R.A. and R.F. McMahon. 1991. Acid-base and ionic regulation, during and

following emersion, in the freshwater bivalve, Anodonta grandis simpsoniana

(Bivalvia: Unionidae). Biological Bulletin 181: 289-297.

Clench, W.J. and R.D. Turner. 1956. Freshwater mollusks of Alabama, Georgia, and

Florida from the Escambia to the Suwannee River. Bulletin of the Florida State

Museum, Biological Sciences 1: 97-239.

Couch, C.A. and R.D. McDowell. 2006. Flint River Basin Regional Water Development

and Conservation Plan. Georgia Department of Natural Resources-Environmental

Protection Division: Atlanta, GA; 244 p.

Dietz, T.H. 1974. Body fluid composition and aerial oxygen consumption in the

freshwater mussel, Ligumia subrostrata (Say): effects of dehydration and anoxic

stress. Biological Bulletin 147:560-572.

Dycus, J.C. 2011. Masters Thesis. Athens, GA. University of Georgia.

Freeman, M.C., G.R. Buell, L.E. Hay, W.B. Hughes, R.B. Jacobson, J.W. Jones, S.A.

Jones, J.H. Lafontaine, K.R. Odom, J.T. Peterson, J.W. Riley, J.S. Schindler, C.P.

Shea, and J.D. Weaver. 2012. Linking river management to species conservation

using dynamic landscape-scale models. River Research and Applications DOI:

10.1002/rra.2575.

Fritts, A.K. 2013. Doctoral Thesis. Warnell School of Forestry & Natural Resources.

Athens, GA, University of Georgia. Ph.D.

Gagne, F., C. Blaise, I. Aoyama, R. Luo, C. Gagnon, Y. Couillard, P. Campbell, and M.

Salazar. 2002. Biomarker study of a municipal effluent dispersion plume in two

species of freshwater mussels. Environmental Toxicology 17:149-159.

145

Golladay, S.W., P.Gagnon, M. Kearns, J.M. Battle, and D. W. Hicks. 2004. Response of

freshwater mussel assemblages (Bivalvia: Unionidae) to a record drought in the

Gulf coastal plain of southwestern Georgia. Journal of the North American

Benthological Society 23(3): 494-506.

Gustafson, L.L., M.K. Stoskopf, A.E. Bogan, W. Showers, T.J. Kwak, S. Hanlon, and

J.F. Levine. 2005a. Evaluation of a nonlethal technique for hemolymph collection

in Elliptio complanata, a freshwater bivalve (Mollusca : Unionidae). Diseases of

Aquatic Organisms 65:159-165.

Gustafson, L.L., M.K. Stoskopf, W. Showers, G. Cope, C. Eads, R. Linnehan, T.J. Kwak,

B. Andersen, and J.F. Levine. 2005b. Reference ranges for hemolymph

chemistries from Elliptio complanata of North Carolina. Diseases of Aquatic

Organisms 65:167-176.

Haag, W.R., D.J. Berg, D.W. Garton, and J.L. Farris. 1993. Reduced survival and fitness

in native bivalves in response to rouling by the introduced zebra mussel

(Dreissena polymorpha) in western Lake Erie. Canadian Journal of Fisheries and

Aquatic Sciences 50:13-19.

Haag, W.R. and M.L. Warren. 2008. Effects of severe drought on freshwater mussel

assemblages. Transactions of the American Fisheries Society 137:1165-1178.

Haag, W.R. 2012. North American Freshwater Mussels: Natural History, Ecology, and

Conservation. Cambridge University Press, New York, New York.

Haag, W.R. and J.D. Williams. 2013. Biodiversity on the brink: an assessment of

conservation strategies for North American freshwater mussels. Hydrobiologia.

DOI: 10.1007/s10750-013-1524-7.

146

Hicks, D.W., H.E. Gill, and S.A. Longsworth. 1987. Hydrogeology, chemical quality,

and availability of ground water in the upper Floridan aquifer, Albany area,

Georgia. U.S. Geological Survey; Water-Resources Investigations Report 87-

4145; 31 p.

McCargo, J.W. and J.T. Peterson. 2010. An evaluation of the influence of seasonal base

flow and geomorphic stream characteristics on coastal plain stream fish

assemblages. Transactions of the American Fisheries Society 139: 29-48.

Mosner, M.S. 2002. Stream-aquifer relations and the potentiometric surface of the upper

Floridan aquifer in the lower Apalachicola–Chattahoochee–Flint river basin in

parts of Georgia, Florida, and Alabama, 1999–2000. U.S. Geological Survey

Water-Resources Investigations Report 02-4244. 53 p.

Naimo, T.J., E.D. Damschen, R.G. Rada, and E.M. Monroe. 1998. Nonlethal evaluation

of the physiological health of unionid mussels: methods for biopsy and glycogen

analysis. Journal of the North American Benthological Society 17:121-128.

Patterson, M.A., B.C. Parker, and R.J. Neves. 1999. Glycogen concentration in the

mantle tissue of freshwater mussels (Bivalvia : Unionidae) during starvation and

controlled feeding. American Malacological Bulletin 15:47-50.

Peterson, J.T., J.M. Wisniewski, C.P. Shea, and C.R. Jackson. 2011. Estimation of mussel

population response to hydrologic alteration in a southeastern U.S. stream.

Environmental Management 48: 109-122.

Pierce, R.R., N.L. Barber, and H.R. Stiles. 1984. Georgia Irrigation: 1970-1980: A

Decade of Growth. U.S. Geological Survey; Water Resources Investigations

Report 83-4177; Atlanta, GA; 180 p.

147

Postel, S.L. 1997. Last Oasis: Facing Water Scarcity. W.W. Norton: New York and

London.

Postel, S.L. 2000. Entering an era of water scarecity: the challenges ahead. Ecological

Applications 10: 941-948.

Pynnönen, K. 1994. Hemolymph gases, acid-base status, and electrolyte concentration in

the freshwater clams Anodonta anatina and Unio tumidus during exposure to and

recovery from acidic conditions. Physiological Zoology 67(6):1544-1559.

R Core Team (2012). R: A language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0,

URL http://www.R-project.org/.

Royall, R.M. 1997. Statistical Evidence: A Likelihood Paradigm. Chapman and Hall,

New York.

Royle, J.A. and R.M. Dorazio. 2008. Hierarchical Modeling and Inference in Ecology,

the Analysis of Data from Populations, Metapopulations, and Communities.

Academic Press, San Diego, CA.

Rugel, K., C.R. Jackson, J.J. Romeis, S.W. Golladay, D.W. Hicks, and J.F. Dowd. 2012.

Effects of irrigation withdrawals on streamflows in a karst environment: lower

Flint River basin, Georgia, USA. Hydrological Processes 26: 523-534.

Ruhl, J.B. 2005. Water wars, eastern style: divvying up the Apalachicola-Chattahoochee-

Flint River Basin. Journal of Contemporary Water Research and Education 131:

47-52.

Shea, C.P., J.T. Peterson, M.J. Conroy, and J.M. Wisniewski. 2013. Evaluating the

influence of land use, drought and reach isolation on the occurance of freshwater

148

mussel species in the lower Flint River Basin, Georgia (U.S.A.). Freshwater

Biology 58: 382-395.

Torak, L.J. and J.A. Painter. 2006. Geohydrology of the lower Apalachicola-

Chattahoochee-Flint River Basin, Southwestern Georgia, Northeastern Florida,

and Southeastern Alabama. U.S. Geological Survey, Report Number 2006-5070,

Atlanta, GA, 80 pp.

Williams, J. D., M. L. Warren, K. S. Cummings, J. L. Harris, and R. J. Neves. 1993.

Conservation status of freshwater mussels of the United States and Canada.

Fisheries 18(9):6–22.

Williams, J.D., A.E. Bogan, and J.T. Garner. 2008. Freshwater mussels of Alabama and

the Mobile Basin in Georgia, Mississippi and Tennessee. The University of

Alabama Press, Tuscaloosa, Alabama.

Wolfe, D.A. 1996. Insights into the utility of biomarkers or environmental impact

assessment and monitoring. Human and Ecological Risk Assessment 2: 245-250.

149

Table 5.1. List of parameters included for candidate models for biomarkers response of freshwater mussels in the lower Flint River

Basin, southwestern Georgia. Parameters are accompanied by the hypothesized effects. Hemolymph parameters include alanine aminotransferase (ALT), aspartate aminotransferase (AST), bicarbonate, calcium, and magnesium. Tissue parameters include glycogen content. Mussel species = Villosa vibex, Villosa lienosa, and Elliptio crassidens.

Parameter Interpretation/hypothesis Season All hemolymph and tissue parameters may vary by season. Physiographic region Calcium, magnesium, and bicarbonate may differ by physiographic region due to differences in geomorphology. Size All hemolymph and tissue parameters may vary by size of individual. Sex All hemolymph and tissue parameters may vary by sex. Species All hemolymph and tissue parameters may differ between species. Water temperature All hemolymph and tissue parameters may differ by water temperatures. Dissolved oxygen Enzymes (ALT & AST) may spike at low D.O. levels. Discharge (=discharge/watershed area) All hemolymph and tissue parameters may vary with different levels of discharge. Discharge * site Effects of discharge may vary by site for all hemolymph and tissue parameters. Effects of discharge may vary by size, sex, or species for all hemolymph and tissue Discharge * size, sex, species parameters. Dissolved oxygen * size, sex, species Effects of differing D.O. levels may vary by size, sex, or species for all hemolymph and tissue parameters. Water temperature * size, sex, Effects of temperature may vary by size, sex, or species for all hemolymph and tissue species parameters.

150

Table 5.2. Akaike’s Information Criterion (AICc), number of parameters (K), ΔAICc, and Akaike weights (wi) for the confidence set of models estimating hemolymph alanine aminotransferase response to environmental parameters in the lower Flint River Basin,

Georgia. Discharge is classified as follows: Day 0 = discharge at time of sampling, Day 5 = average discharge for five days prior to sampling, and Day 15 = average of discharge for 15 days prior to sampling. Mussel species = Villosa vibex, Villosa lienosa, and

Elliptio crassidens.

Candidate model Discharge AICc K ΔAICc wi (1 | site) + season + region + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 0 2896.73 21 0.00 0.330 (1 | site) + season + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 0 2897.09 20 0.36 0.276 (1 | site) + season + region + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen Day 0 2898.66 20 1.93 0.126

151

(1 | site) + season + region + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 15 2898.80 21 2.07 0.117 (1 | site) + season + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 15 2900.35 20 3.62 0.054 (1 | site) + season + region + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 5 2900.96 21 4.23 0.040

152

Table 5.3. Akaike’s Information Criterion (AICc), number of parameters (K), ΔAICc, and Akaike weights (wi) for the confidence set of models estimating hemolymph aspartate aminotransferase response to environmental parameters in the lower Flint River Basin,

Georgia. Discharge is classified as follows: Day 0 = discharge at time of sampling, Day 5 = average discharge for five days prior to sampling, and Day 15= average of discharge for 15 days prior. Species = Villosa vibex, Villosa lienosa, and Elliptio crassidens.

Candidate model Discharge AICc K ΔAICc wi (1 | site) + season + region + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 15 3300.44 21 0.00 0.388 (1 | site) + season + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 15 3301.92 20 1.48 0.185 (1 | site) + season + region + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 5 3302.91 21 2.47 0.113 (1 | site) + season + region + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 0 3303.35 21 2.91 0.091

153

(1 | site) + region + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 15 3303.99 20 3.55 0.066 (1 | site) + season + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 5 3304.78 20 4.34 0.044 (1 | site) + season + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 0 3304.84 20 4.40 0.043

154

Table 5.4. Akaike’s Information Criterion (AICc), number of parameters (K), ΔAICc, and Akaike weights (wi) for the confidence set of models estimating hemolymph bicarbonate response to environmental parameters in the lower Flint River Basin, Georgia. Discharge is classified as follows: Day 0 = discharge at time of sampling, Day 5 = average discharge for five days prior to sampling, and Day 15

= average of discharge for 15 days prior to sampling. Mussel species = Villosa vibex, Villosa lienosa, and Elliptio crassidens.

Candidate model Discharge AICc K ΔAICc wi (1 | site) + season + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + season * discharge Day 15 1341.90 12 0.00 0.518 (1 | site) + season + region + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + season * discharge + discharge^2 Day 15 1343.45 14 1.55 0.239 (1 | site) + season + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + season * discharge + discharge^2 Day 15 1343.86 13 1.96 0.194

155

Table 5.5. Akaike’s Information Criterion (AICc), number of parameters (K), ΔAICc, and Akaike weights (wi) for the confidence set of models estimating hemolymph calcium response to environmental parameters in the lower Flint River Basin, Georgia. Discharge is classified as follows: Day 0 = discharge at time of sampling, Day 5 = average discharge for five days prior to sampling, and Day 15 = average of discharge for 15 days prior to sampling. Mussel species = Villosa vibex, Villosa lienosa, and Elliptio crassidens.

Candidate model Discharge AICc K ΔAICc wi (1 | site) + size + sex + E.crassidens + V.lienosa + discharge + (discharge | site) + size * discharge + sex * discharge + E.crassidens * discharge + V.lienosa * discharge + discharge^2 Day 15 2004.64 13 0.00 0.484 (1 | site) + region + size + sex + E.crassidens + V.lienosa + discharge + (discharge | site) + size * discharge + sex * discharge + E.crassidens * discharge + V.lienosa * discharge + discharge^2 Day 15 2005.19 14 0.55 0.368 (1 | site) + season + region + size + sex + E.crassidens + V.lienosa + discharge + (discharge | site) + size * discharge + sex * discharge + E.crassidens * discharge + V.lienosa * discharge + discharge^2 Day 15 2007.56 15 2.92 0.112

156

Table 5.6. Akaike’s Information Criterion (AICc), number of parameters (K), ΔAICc, and Akaike weights (wi) for the confidence set of models estimating hemolymph magnesium response to environmental parameters in the lower Flint River Basin, Georgia. Discharge is classified as follows: Day 0 = discharge at time of sampling, Day 5 = average discharge for five days prior to sampling, and Day 15

= average of discharge for 15 days prior to sampling. Mussel species = Villosa vibex, Villosa lienosa, and Elliptio crassidens.

Candidate model Discharge AICc K ΔAICc wi (1 + discharge | site) + region + size + sex + E.crassidens + V.lienosa + discharge Day 5 547.00 9 0.00 0.305 (1 | site) + region + size + sex + E.crassidens + V.lienosa + discharge + (discharge | site) + region * discharge Day 5 547.90 10 0.90 0.194 (1 + discharge | site) + region + size + sex + E.crassidens + V.lienosa + discharge Day 0 548.69 9 1.69 0.131 (1 | site) + region + size + sex + E.crassidens + V.lienosa Day 0 549.28 8 2.28 0.097 (1 | site) + region + size + sex + E.crassidens + V.lienosa Day 5 549.28 8 2.28 0.097 (1 | site) + region + size + sex + E.crassidens + V.lienosa Day 15 549.28 8 2.28 0.097 (1 | site) + region + size + sex + E.crassidens + V.lienosa + discharge + (discharge | site) + region * discharge Day 0 550.34 10 3.34 0.057

157

Table 5.7. Akaike’s Information Criterion (AICc), number of parameters (K), ΔAICc, and Akaike weights (wi) for the confidence set of models estimating tissue glycogen response to environmental parameters in the lower Flint River Basin, Georgia. Discharge is classified as follows: Day 0 = discharge at time of sampling, Day 5 = average discharge for five days prior to sampling, and Day 15 = average of discharge for 15 days prior to sampling. Mussel species = Villosa vibex, Villosa lienosa, and Elliptio crassidens.

Candidate model Discharge AICc K ΔAICc wi (1 | site) + region + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 15 1365.41 20 0.00 0.317 (1 | site) + season + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 15 1366.25 20 0.84 0.208 (1 | site) + region + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 5 1366.51 20 1.10 0.183 (1 | site) + region + size + sex + E.crassidens + V.lienosa + discharge + (discharge | site) + size * discharge + sex * discharge + E.crassidens * discharge + V.lienosa * discharge + discharge^2 Day 15 1367.38 14 1.97 0.118

158

(1 | site) + season + region + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + discharge + (discharge | site) + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen + discharge^2 Day 5 1368.64 21 3.23 0.063 (1 | site) + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen Day 0 1369.39 17 3.98 0.043 (1 | site) + size + sex + E.crassidens + V.lienosa + temp + dissolved oxygen + size * temp + sex * temp + size * dissolved oxygen + sex * dissolved oxygen + E.crassidens * temp + V.lienosa * temp + E.crassidens * dissolved oxygen + V.lienosa * dissolved oxygen Day 5 1369.39 17 3.98 0.043

159

Fig. 5.1. Locations of five sampling sites within the lower Flint River Basin. Three freshwater mussel species (Elliptio crassidens, Villosa vibex, and Villosa lienosa) were collected. Each site was visited on four to seven occasions from May 2010 to June 2011.

160

Fig. 5.2. Akaike importance weights of parameters used to model six biological responses in freshwater mussels of the lower Flint River Basin: A) alanine aminotransferase, B) aspartate aminotransferase, C) bicarbonate, D) calcium, E) magnesium, and F) tissue glycogen. Data include the importance weights of three different discharge durations

(discharge at time of sampling, average discharge for five days prior, and average discharge for 15 days prior), main effects, and interactions. Freshwater mussel species include Elliptio crassidens, Villosa vibex, and Villosa lienosa.

161

Discharge Main effects Interactions

Discharge Main effects Interactions

162

Discharge Main effects Interactions

Discharge Main effects Interactions

163

Discharge Main effects Interactions

Discharge Main effects Interactions

164

Fig. 5.3. Parameter estimates (±SE) of the best supporting models of six biological responses in freshwater mussels of the lower Flint River Basin: A) alanine aminotransferase, B) aspartate aminotransferase, C) bicarbonate, D) calcium, E) magnesium, and F) tissue glycogen. Mussel species = Villosa vibex, Villosa lienosa, and

Elliptio crassidens. All continuous parameters (e.g. discharge, DO, temperature, size) were adjusted into standardized predictors, which were calculated by subtracting the mean and dividing by the standard deviation; the magnitudes of the estimates are directly comparable.

165

166

167

5.0 F

4.0

3.0

2.0

1.0

Parameter estimate 0.0

-1.0

-2.0

168

Fig. 5.4. Predicted response of alanine aminotransferase (ALT) to changes in discharge at two different temperature regimes for three freshwater mussel species (Elliptio crassidens, Villosa vibex, and Villosa lienosa) of average length during the spring season in the Dougherty Plain physiographic province.

169

Fig. 5.5. Predicted response of calcium to changes in discharge among three freshwater mussel species: Elliptio crassidens, Villosa vibex, and Villosa lienosa.

170

CHAPTER 6

SYNTHESIS AND CONCLUSIONS

INTRODUCTION

The decline of freshwater mussels is widespread and poses a challenge that managers and conservationists will continue to confront as growing human populations demand more of the earth’s aquatic resources. There have been great advances in our understanding of the causes of these declines, but clearly additional research is needed to improve our ability to preserve and restore mussel populations. The threats and challenges faced by freshwater mussel species in the Flint River Basin in southwest

Georgia are similar to those experienced by other imperiled mussels worldwide, and the lessons gained from this study of Flint River Basin species could advance the protection and restoration of other freshwater mussels in different watersheds around the world.

CHAPTER SYNTHESIS

I presented results in Chapter 2 that assessed the ecological relevance of the sodium chloride test for measuring the viability of freshwater mussel glochidia. These results help to support the use of the sodium chloride viability test when assessing the sensitivity of glochidia to various contaminants and contribute to our understanding of the effects of toxic substances on the health and survival of this earliest life stage of freshwater mussels. These results are also useful for increasing the accuracy of host-

171

determination trials. Accurate evaluation of the viability of glochidia will help researchers to avoid ‘false-negative’ results, which occur when researchers fail to identify a fish species that is a successful host because of the use of unhealthy glochidia. The realization that the sodium chloride challenge represents an ecologically relevant indication of the ability of glochidia to attach to their specific host and then transform into the juvenile stage provides previous and future researchers with some assurance that the methods that they use to test glochidial viability are trustworthy.

Our results indicated that glochidia viability (as measured by the shell closing response to saturated NaCl exposure) is indicative of the ability to metamorphose into the juvenile life stage, as long as control group viability was >90%. The age of glochidia also appeared to be critical to glochidia health. The viability of older glochidia (those from females collected late in the brooding season or from females held in the laboratory for longer periods of time) declined more rapidly after they were extracted from the female mussel. The current American Society for Testing and Materials (ASTM) guidelines for conducting toxicity tests with glochidia of freshwater mussels requires

>90% viability for control groups at test termination (commonly 24 h) and our results support the continued use of this criterion (ASTM 2006). Additionally, glochidia that remained viable following exposure to a toxicant for 24 h were able to successfully metamorphose, thus the viability endpoint appears to have direct ecological relevance.

Therefore, we recommended that the current ASTM guideline for glochidia toxicity testing be retained for all mussel species. We also recommended that glochidia from mussels with longer brooding periods should be collected well before the end of the brooding season when used for toxicity testing. Glochidia collected late in the brooding

172

season appeared less healthy and our results showed that older glochidia were more sensitive to toxicants, had lower initial viability which declined quickly, and had poorer metamorphosis success.

In Chapter 3, I highlighted the importance of host determination trials with the discovery of the federally protected Gulf Sturgeon serving as a host for the federally protected Purple Bankclimber mussel. While host determination studies may be in danger of becoming “passé” as new areas of research enter the arena of malacology, they still represent a substantial research need for the conservation of freshwater mussels (Haag and Williams 2013). Dependable host results are still lacking for a large number of mussel species; however, these efforts to document the complex relationship between

Gulf Sturgeon and Purple Bankclimber mussels have contributed new insights into the collective knowledge of how to preserve and restore these relatively large and charismatic species.

Accurate knowledge of host fish data not only provides information to managers for evaluating the health of wild host populations, but it also allows for the development of captive propagation programs. The U.S. Fish and Wildlife Service recovery plan for the federally listed species in the ACF calls for the stabilization of existing subpopulations and also the expansion of the species range through the addition of supplemental subpopulations---an action that would likely require the use of captive propagation (U.S. Fish and Wildlife Service 2003). Propagation may provide some immediate relief to the threat of extinction but the recovery of imperiled mussel and fish populations will ultimately depend on ecosystem-wide habitat restoration for all life stages (Neves et al. 1997). The extirpation of migratory fishes from rivers worldwide has

173

had cascading effects other than the diminished reproductive potential of freshwater mussels (Naiman et al. 2002, Agostinho et al. 2008). Our study highlighted the interconnectedness and complexity of ecological relationships and their importance in the conservation of mussels. Habitat fragmentation places the sturgeon population in the

ACF at risk of extinction, and because of its dependence on sturgeon as hosts, the Purple

Bankclimber also is at risk. Similar examples of species coextinctions, in which the fate of one species is directly tied to that of another, have been documented worldwide and they underscore the need for a holistic, ecosystem approach to conservation (Koh et al.

2004).

My fourth and fifth chapters presented results that contribute to our understanding of how mussels react to changes in their habitats. Drought conditions, and the secondary effects of drought (e.g., low flow, low dissolved oxygen, and elevated water temperatures in stagnant pools) are some of the stress events that aquatic organisms are likely to face with increasing frequency in the future as our planet enters a period of substantial climate change (Araujo and Rahbek 2006). The inherent complexity of natural environments makes it difficult to characterize the exact effect that these changes will have on mussel populations (Wolfe 1996), but the use of advanced multivariate and multi-scale modeling techniques can allow researchers to derive estimates of the effects of various biotic and abiotic factors on observable phenomena (Burnam and Anderson

2002).

The use of physiochemical biomarkers may provide future researchers with an effective method to monitor the health of mussel populations in their natural environments. The factors of discharge, size, sex, and species were most commonly

174

found to affect the biological responses in our models and we recommend that future research into the effects of drought and stress should include the use of alanine aminotransferase, aspartate aminotransferase, bicarbonate, and calcium. There is growing evidence that foot tissue biopsies and hemolymph extraction from the adductor muscle can be collected in a non-lethal manner (Naimo et al. 1998, Gustafson et al.

2005a,b, Fritts 2013), and opportunities to create biomonitoring plans for imperiled populations of freshwater mussels may be feasible in coming years. These studies have provided a framework methodology for the emerging study of physiological biomarkers of freshwater mussel populations and I strongly recommend that future researchers consider the use of similar modeling techniques to account for the inherent complexity of aquatic ecosystems.

CONCLUSIONS

The completion of these four research components has provided new information that could be used to advance the conservation of imperiled freshwater mussels.

Assessing the ecological relevance of the sodium chloride glochidia viability test will lend greater weight to studies involving freshwater mussel glochidia and will also improve the reliability of host determination trials. The discovery of a primary host for the threatened Purple Bankclimber mussel will allow for better management of this species and will enable the initiation of a captive propagation program. By establishing biomarkers that can be used for a successful biomonitoring program, our hope is that these data may be used to detect stressed populations prior to large scale mortality events.

This study has provided crucial information about conditions that cause stress in mussels

175

and these results can help lead to improved water resource allocation and overall management of rivers. For many of the imperiled mussel species, their continued existence is perilous. It is imperative that researchers act deliberately and systematically to increase the likelihood of preserving these fascinating animals that play a critical role in aquatic ecosystems.

176

REFERENCES

Abell, R.A. 2000. Freshwater Ecoregions of North America: A Conservation Assessment.

Island Press, Washington, D.C. USA.

Agostinho, A.A., F.M. Pelicice, and L.C. Gomes. 2008. Dams and the fish fauna of the

Neotropical region: impacts and management related to diversity and fisheries.

Brazilian Journal of Biology 68:1119–1132.

American Society for Testing and Materials. 2006. Standard guide for conducting

laboratory toxicity tests with freshwater mussels E2455-06. In Annual Book of

ASTM Standards, Vol 11.06. Philadelphia, PA, USA, pp 1393–1444.

Araujo, M.B. and C. Rahbek. 2006. How does climate change affect biodiversity?

Science 313: 1396-1397.

Burnham, K.P. and D.R. Anderson. 2002. Model Selection and Inference: An

Information -Theoretic Approach, 2nd edition. Springer-Verlag, New York.

Butler, R.S. 1993. Results of a status survey for eight freshwater mussels (Bivalvia:

Unionidae) endemic to eastern Gulf Slope drainages of the Apalachicolan Region

of southeast Alabama, southwest Georgia, and north Florida. Unpublished report,

U.S. Fish and Wildlife Service, Jacksonville, Florida. 41 pp.

Flowers, H.J. , W.E. Pine, A.C. Dutterer, K.G. Johnson, J.W. Ziewitz, M.S. Allen, and

F.M. Parauka. 2009. Spawning site selection and potential implications of

modified flow regimes on viability of Gulf Sturgeon populations. Transactions of

the American Fisheries Society 138:1266–1284.

177

Fritts, A.K., M.W. Fritts, D. Peterson, D. Fox, and R.B. Bringolf. 2012. Critical linkage

of imperiled species: Gulf sturgeon as hosts for purple bankclimber mussels.

Freshwater Science 31(4):1223-1232.

Fritts, A.K. 2013. Doctoral Thesis. Warnell School of Forestry & Natural Resources.

Athens, GA, University of Georgia. Ph.D.

Gustafson, L.L., M.K. Stoskopf, A.E. Bogan, W. Showers, T.J. Kwak, S. Hanlon, and

J.F. Levine. 2005a. Evaluation of a nonlethal technique for hemolymph collection

in Elliptio complanata, a freshwater bivalve (Mollusca : Unionidae). Diseases of

Aquatic Organisms 65:159-165.

Gustafson, L.L., M.K. Stoskopf, W. Showers, G. Cope, C. Eads, R. Linnehan, T.J. Kwak,

B. Andersen, and J.F. Levine. 2005b. Reference ranges for hemolymph

chemistries from Elliptio complanata of North Carolina. Diseases of Aquatic

Organisms 65:167-176.

Haag, W.R. and J.D. Williams. 2013. Biodiversity on the brink: an assessment of

conservation strategies for North American freshwater mussels. Hydrobiologia.

DOI: 10.1007/s10750-013-1524-7.

Kat, P.W. 1984. Parasitism and the Unionacea (Bivalvia). Biological Reviews of the

Cambridge Philosophical Society 59:189–207.

Koh, L.P., R.R. Dunn, N.S. Sodhi, R.K. Colwell, H.C. Proctor, and V.S. Smith. 2004.

Species coextinctions and the biodiversity crisis. Science 304:1632–1634.

Lydeard, C. and R.L. Mayden. 1995. A diverse and endangered aquatic ecosystem of the

Southeast United States. Conservation Biology 9:800-805.

178

Naiman, R.J., R.E. Bilby, D.E. Schindler, and J.M. Helfield. 2002. Pacific salmon,

nutrients, and the dynamics of freshwater and riparian ecosystems. Ecosystems

5:399–417.

Naimo, T.J., E.D. Damschen, R.G. Rada, and E.M. Monroe. 1998. Nonlethal evaluation

of the physiological health of unionid mussels: methods for biopsy and glycogen

analysis. Journal of the North American Benthological Society 17:121-128.

Neves, R.J., A.E. Bogan, J.D. Williams, S.A. Ahlstedt, and P.D. Hartfield. 1997. Status

of aquatic mollusks in the southeastern United States: a downward spiral of

diversity. Pages 43-85 in: G.W. Benz and D.E. Collins, eds. Aquatic fauna in

peril: the southeastern perspective. Special Publication 1, Southeast Aquatic

Research Institute, Lenz Design and Communications, Decatur, Georgia.

O'Brien, C.A., and J.A. Williams. 2002. Reproductive biology of four freshwater mussels

(Bivalvia: Unionidae) endemic to eastern Gulf Coastal Plain drainages of

Alabama, Florida, and Georgia. American Malacological Bulletin 17:147–158.

Schwalb, A.N., K. Cottenie, M.S. Poos, and J.D. Ackerman. 2010. Dispersal limitation of

unionid mussels and implications for their conservation. Freshwater Biology

56:1509–1518.

U.S. Fish and Wildlife Service. 1998. Endangered and Threatened Wildlife and Plants;

Determination of Endangered Status of Five Freshwater Mussels and Threatened

Status for Two Freshwater Mussels from the Eastern Gulf Slope Drainages of

Alabama, Florida, and Georgia. Federal Register 63(50): 12664-12687.

U.S. Fish and Wildlife Service. 2003. Recovery Plan for Endangered Fat Threeridge

(Amblema neislerii), Shinyrayed Pocketbook (Lampsilis subangulata), Gulf

179

Moccasinshell (Medionidus penicillatus), Ochlockonee Moccasinshell

(Medionidus simpsonianus), and Oval Pigtoe (Pleurobema pyriforme); and

Threatened Chipola Slabshell (Elliptio chipolaensis), and Purple Bankclimber

(Elliptoideus sloatianus). Atlanta, Georgia. 142 pp.

Wolfe, D.A. 1996. Insights into the utility of biomarkers or environmental impact

assessment and monitoring. Human and Ecological Risk Assessment 2: 245-250.

180