The Pennsylvania State University

The Graduate School

Department of Biology

PHYLOGEOGRAPHY OF DEEP-SEA VESTIMENTIFERANS AND A

POPULATION GENETICS STUDY OF TWO SPECIES, LAMELLIBRACHIA

LUYMESI AND SEEPIOPHILA JONESI, FROM THE GULF OF MEXICO

A Thesis in

Biology

by

Erin R. McMullin

© 2003 Erin R. McMullin

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

August 2003

The thesis of Erin McMullin has been reviewed and approved* by the following:

Charles R. Fisher Professor of Biology Thesis Co-Adviser Co-Chair of Committee

Stephen W. Schaeffer Associate Professor of Biology Thesis Co-Adviser Co-Chair of Committee

Andrew Clark Professor of Biology

Lee Kump Professor of Geosciences

Kimberlyn Nelson Forensic Examiner Mitotyping Technologies Special Signatory

Douglas Cavener Professor of Biology Head of the Biology Department

* Signatures are on file at the Graduate School iii

ABSTRACT

First discovered in 1977 on the Galapagos Rift, vestimentiferans are a group of deep-sea annelids found in a variety of environments worldwide. Vestimentiferan communities are isolated pockets of high biomass in the otherwise nutrient-poor deep-sea. Chemosynthesis, not photosynthesis, is the underlying energy source for vestimentiferans, which entirely lack a digestive tract and rely on internal sulfide-oxidizing symbionts for fixed carbon. Symbionts appear to be acquired by the motile vestimentiferan larvae before they settle and become sessile adults. The ability of both these organisms and their symbionts to disperse across the sometimes considerable distances between sulfidic environments has been a topic of study for over a decade. This study addresses the question of dispersal by two different methods, with a biogeographical approach of all vestimentiferan species, and through the population genetic analysis of two species within the Gulf of Mexico. Species distributions can reveal barriers to gene flow, and the range over which an organism maintains a single species reflects the ability of individuals in that species to interbreed between populations. In the case of the vestimentiferan symbionts, which do not interbreed, symbiont strain distribution reflects the presence of a particular bacterial strain near the vestimentiferan larvae before symbiont acquisition. Vestimentiferan biogeography reveals that many species have very large species ranges, indicating the exchange of at least one migrant per generation between sites as distant as 6000km. However, different vestimentiferan species are found between sites that are geographically close but that are at very different depths, suggesting depth may be a barrier to either dispersal or survival of larvae between sites. The same vestimentiferan symbiont strains are found worldwide, with different host species in sites as distant as the Florida Escarpment and the coast of Oregon containing the same symbiont. However, as with the vestimentiferan hosts, evidence suggests that depth may affect the presence of symbiont strains. Symbiont strains may in fact be ubiquitously distributed throughout the world’s oceans but only within a certain depth range, and the particular strain a vestimentiferan host contains may depend predominantly on the depth at which that host lives. Biogeography serves to define species limits, but does not directly measure gene flow within a species. Microsatellite markers were developed for use in a population genetics study of Lamellibrachia luymesi and Seepiophila jonesi of the Louisiana Slope of the Gulf of Mexico. iv

Microsatellite markers are regions of repetitive DNA that tend to be highly polymorphic within a species. Five polymorphic microsatellite loci were isolated from L. luymesi and seven from S. jonesi; these loci had between 7 and 50 alleles. Samples were collected from aggregations of tubeworms from nine hydrocarbon seep sites from the Louisiana Slope, and screened with the microsatellite loci to reveal departures from Hardy Weinberg Equilibrium. No obvious population structure was found between aggregations within a sample site in either species. L. luymesi showed some evidence for isolation by distance across the 480km from which it was sampled, but S. jonesi showed no evidence for isolation by distance over 580km. Both L. luymesi and S. jonesi, however, showed a generalized excess of homozygous individuals which did not reflect underlying population substructure. These data suggest that both species have high geneflow between aggregations and sample sites, but that L. luymesi dispersal is somewhat more limited than S. jonesi. The generalized homozygote excess observed reflects a departure from HWE the cause of which is not clearly identified in this study. Cohorts may exist within an aggregation that were not revealed with this sampling plan, leading to mating between related individuals. A second hypothesis to explain the observed homozygote excess may be strong selection against heterozygous larvae, as is suggested in some marine bivalves. Unlike most of the deep sea, vestimentiferan communities are characterized by very high biomass. These communities, however, depend on resources that are patchily distributed across the globe. Biogeographic data suggest that both vent and seep vestimentiferans have strong dispersal capabilities, and the population genetic analyses presented here support the same conclusion. However, vestimentiferan dispersal is affected by at least one physical barrier, depth, and may also be affected by water currents. The polymorphic markers isolated in this study are a tool that can be used to address the presence of physical barriers between vestimentiferan communities, and can also be used to reveal non random mating within a species, as in the Gulf of Mexico study. v

TABLE OF CONTENTS

List of Figures vii List of Tables viii Preface ix Acknowledgements x

Chapter 1: General background and purpose………………………………………. 1

Chapter 2: Metazoans in extreme environments: Adaptations of hydrothermal vent and hydrocarbon seep fauna……………………………………………………. 12

Introduction…………………………………………………………………... 12

Temperature………………………………………………………………….. 12

Hypoxia/Anoxia……………………………………………………………… 12

Toxicity………………………………………………………………………. 12 Sulfide………………………………………………………………… 12 Metals………………………………………………………………… 12

Summary and Conclusions…………………………………………………… 12

References……………………………………………………………………. 12

Chapter 3: Phylogeny and biogeography of deep sea vestimentiferan tubeworms 13 and their bacterial symbionts…………………………………………………………

Introduction…………………………………………………………………... 13

Materials and Methods……………………………………………………….. 13

Results and Discussion………………………………………………………. 13 Vestimentiferan hosts………………………………………………… 13 Vestimentiferan symbionts…………………………………………… 13 Conclusion…………………………………………………………………… 13

References…………………………………………………………………… 13

vi

Chapter 4: Twelve microsatellites for two deep sea polychaete tubeworm species, Lamellibrachia luymesi and Seepiophila jonesi, from the Gulf of Mexico……………. 14

Text……………………………………………………………………………... 16

References………………………………………………………………………. 19

Tables and Figures……………………………………………………………… 20

Chapter 5: Genetic diversity and population structure of two deep sea tubeworms, Lamellibrachia luymesi and Seepiophila jonesi, from the hydrocarbon seeps of the Gulf of Mexico…………………………………………………………………………. 22

Introduction……………………………………………………………………... 23

Materials and Methods…………………………………………………………. 25 Sampling………………………………………………………………… 25 Molecular Analyses……………………………………………………... 26 Descriptive Statistics……………………………………………………. 26

Results………………………………………………………………………...... 29

Discussion………………………………………………………………………. 33 Population Structure…………………………………………………...... 33 A Generalized Heterozygote Deficiency………………………………... 35

Conclusions……………………………………………………………………... 37 References………………………………………………………………………. 39

Tables and Figures……………………………………………………………… 43

Appendix A: Details on the isolated microsatellite loci from Chapter 4……………… 49

Tables and Figures……………………………………………………………… 50

Appendix B: Details on the hydrocarbon seep sites sampled in Chapter 5…………… 56

Tables and Figures……………………………………………………………… 57

vii

LIST OF FIGURES

Figures shown in Chapter 2: Figure 1. Hydrothermal vents manifest on the sea floor as chimney structures and diffuse flow fields…………………………………...... 12 Figure 2. Temperature, oxygen, and sulfide concentrations measured in situ at the Galapagos Rift…………………………………………...... 12 Figure 3. Mechanisms allowing vent tubeworms to tolerate and exploit their sulfide-rich environments………………………………………………….. 12

Figures shown in Chapter 3: Figure 1. Neighbor joining tree showing molecular evolutionary relationships among vestimentiferan cytochrome oxidase I sequences…………………………………………………………...... 13 Figure 2. Worldwide distribution of vestimentiferans labeled by species when known………………………………………………………………………. 13 Figure 3. An unrooted neighbor joining tree of the vestimentiferan COI sequences used in the relative rates test……………………………………. 13 Figure 4. Neighbor joining tree showing molecular evolutionary relationships among vestimentiferan symbiont rDNA 16S sequences…………………………………………………………...... 13 Figure 5. Distribution of vestimentiferan host species and symbiont strains in paired samples……………………………………………………………… 13

Figures shown in Chapter 5: Figure 5.1. A map of sample locations on the Louisiana Slope of the Gulf of Mexico……………………………………………………………………... 43 Figure 5.2. Null allele frequencies estimated by two methods graphed against FIT values for each locus…………………………………...... 45 Figure 5.3. A neighbor joining tree for L. luymesi built with pairwise Nei’s Da genetic distances…………………………………………………………… 46 Figure 5.4. : A neighbor joining tree for S. jonesi built with pairwise Nei’s Da genetic distances…………………………………………………………… 47 Figure 5.5. A test for isolation by distance comparing geographic distance to genetic distance between pairs of samples. a) comparisons of L. luymesi geographic and genetic distances, b) comparisons of S. jonesi geographic and genetic distances……………………………………………………….. 48

Appendix A: Figure A.1. Frequency distributions of five variable microsatellite loci for L. luymesi……………………………………………………………………... 51 Figure A.2. Frequency distributions of seven variable microsatellite loci for S. jonesi……………………………………………………………………….. 52

viii

LIST OF TABLES

Tables shown in Chapter 3: Table 3.1. GenBank Accession numbers of sequences used in vestimentiferan host and symbiont phylogenetic analyses………………………………….. 13 Table 3.2. Vestimentiferan collections worldwide, including site depth, habitat type and vestimentiferan species present………...... 13

Tables shown in Chapter 4: Table 4.1. Characteristics of 12 vestimentiferan tubeworm microsatellite DNA loci. Included are locus designation, primer sequence, GenBank accession numbers, repeat motif, range of PCR products in base pairs, and number of alleles observed…………………………………………………………….. 21 Table 4.2. Cross amplification of isolated microsatellite loci. All twelve loci were tested in seven different vestimentiferan species at 52°C annealing temperature………………………………………………………………… 22

Tables shown in Chapter 5: Table 5.1: A list of the site names for each sample site used in this study. Sites are listed from west to east…………………………………………...... 44

Table 5.2: Pairwise geographic distances between sites, in km……...... 44

Tables shown in Appendix A: Table A.1: Information on all repetitive loci isolated from L. luymesi, including locus designation, primer sequences (if applicable), GenBank Accession number, annealing temperature, type of repeat, status of locus for use as a population marker, and allele size range…………………………………… 53 Table A.2: Information on all repetitive loci isolated from S. jonesi, including locus designation, primer sequences (if applicable), GenBank Accession number, annealing temperature, type of repeat, status of locus for use as a population marker, and allele size range…………………………………... 54

Tables shown in Appendix B: Table B.1: Support information on samples sites used in Chapter 5. Sample sites are listed by site name, with collection designations from Chapter 5 of this thesis……………………...... 57

ix

PREFACE

Contributions to the research and writing of multi-author chapters:

Chapter 2: Derk Bergquist wrote the sections on adaptations to sulfide and metal toxicity and collaborated on the Conclusions.

Chapter 3: Stephane Hourdez provided information on the morphological characters used in the vestimentiferan taxonomy, and ran the TREEMAP analysis of coevolution between hosts and symbionts. Chuck Fisher was invaluable in reviewing, editing, and structuring the text of this paper. All new molecular data were generated in Steve Schaeffer’s lab, and the work was funded by both Chuck Fisher and Steve Schaeffer.

Chapter 4: (possible change in authorship) John Wood and Sarah Hamm Marrus assisted in the lab with DNA isolations and early microsatellite development

Chapter 5: Both Chuck Fisher and Steve Schaeffer were involved in the overall project design and funding. Chuck Fisher provided the means for sample collections, and Steve Schaeffer provided the laboratory facilities used to generate all data shown in this chapter, as well as consultation on data analyses.

x

ACKNOWLEDGEMENTS

Professional Acknowledgements:

I would like to offer my sincere thanks to my co-advisors, Chuck Fisher and Steve Schaeffer, who together have given me the opportunity to study these unique organisms found in such interesting environments. I thank Dr. Fisher for sharing his expertise and knowledge of these complex ecosystems, and for providing me with access to these sites in the form of research expeditions and financial support. I also thank Chuck for his ability, in both the design of research and the discussion of results, to focus on the most biologically interesting questions, and for his sometimes merciless but always effective editing of manuscripts. I thank Steve Schaeffer for being my mentor in the field of population genetics, and for providing insight and direction in both identifying what questions to ask of the data, and how to approach these questions with population genetic analyses. I thank Steve also for unrestricted use of his laboratory equipment (BIG thanks for that!), and for always being both financially and emotionally supportive during this at times difficult project. Kimberlyn Nelson was integral in the initial design of this study, and I thank her for laying out the overall sampling scheme which underlies Chapters 4 & 5. I also thank her deeply for her time and advice during the early years of my Ph.D., and for teaching me the basic molecular biology tools I continued to use throughout my Ph.D. I would also like to thank Andrew Clark for participating in my Ph.D. committee and for providing much appreciated guidance and comments on my population genetic analyses, and I would like to thank Lee Kump for being the self-appointed committee referee to ensure “fair play”. Many people were involved in collecting the samples used in this thesis. I thank Kim Nelson most of all for collecting the core of the population genetics samples from the Louisiana Slope. I also thank Erik Cordes, John Freytag, Stephane Hourdez and Chuck Fisher for collecting additional samples from the Louisiana Slope. I thank Heiko Sahling for sending me samples from Costa Rica and from Java, and Eve Southward for suggesting it to him. I thank Jim Barry and Paul Yancey for providing samples from the west coast of North America. I thank John Wood, Sarah Hamm, Christi Ludwig, and Therese Waltz for their time and effort in the lab, and Stephane Hourdez for advice and help on molecular phylogenetics (and Dr. Nei’s homework assignments!). Special thanks to the Harbor Branch Oceanographic Institution and to the pilots and crew of the R/V Seward Johnson and the DSRV Johnson Sea Link, without whom none of this work would have been possible.

Personal Acknowledgements:

For sharing long hours at sea together in good spirit, I thank Derk, Ish, KT, John, Stephane, Jason, and Melanie. For making days in a research lab fun, I thank Sue, Miro, and Therese, and all of the Fisher and Schaeffer labs. For sharing long hours at the Skellar drinking good spirits I thank Hany, Ian, Sue, Erik, Ruud, Breea, Joel, and the rest of the crew. And of course special thanks to my roommates, Gioia, Renee, Elene, and Sharmishtha, and assorted cats for making graduate school fun. Yes, it can be done. Finally, thanks to Rob for putting up with the crazy work schedule of a graduate student, and the irritability associated with writing a thesis. And for bringing me flowers ☺ 1

CHAPTER 1

General Background and Purpose 2

Chapter 1

General Background and Purpose

Vestimentiferan tubeworms are a group of sessile marine polychaete annelids of the family

Siboglinidae (McHugh, 1997, Rouse, 2001, Halanych et al., 2001) found in regions of hydrothermal venting and seepage of the reduced chemical sulfide (Corliss et al., 1979, Paull et al., 1984,

Kennicutt et al., 1985). Vestimentiferans completely lack a digestive tract and rely on internal sulfide oxidizing symbionts for fixed carbon (Cavanaugh et al., 1981, Childress et al., 1991). The species that comprise the vestimentifera live at depths of between 300m and 3300m, with one shallow water exception found at 82m depth (Tunnicliffe et al., 1998, Sibuet & Olu, 1998,

Hashimoto et al., 1993). Biogeographic and population studies of these deep sea tubeworms are limited by the technological difficulty involved in sampling and observing deep sea communities.

Molecular genetic techniques which make use of frozen or preserved tissue samples therefore provide useful tools for exploring relationships both between and within species in these difficult- to-sample organisms. This study uses genetic approaches first to explore the biogeography of vestimentiferans and their symbionts, and second to explore the genetic diversity, population substructure, and gene flow within two species of vestimentiferans found at the hydrocarbons seeps of the Louisiana Slope in the Gulf of Mexico.

Vestimentiferan tubeworms are found worldwide in deep sea environments in which both hydrogen sulfide and oxygen co-occur (Childress & Fisher, 1992, Sibuet & Olu, 1998).

Vestimentiferan symbionts require sulfide as an electron source for chemoautotrophy, oxidizing sulfide to sulfate in order to produce the energy required for carbon fixation (Childress et al., 1991).

The symbionts, therefore, require both sulfide and oxygen, and the vestimentiferan host requires oxygen for cellular respiration. Because oxygen and sulfide spontaneously react with each other 3 (Johnson et al., 1986), the two molecules only co-occur in areas where their sources are either spatially or temporally separated. Examples of such environments include hydrothermal vents, hydrocarbon seeps, accretionary prisms, whale carcasses and shipwrecks (Johnson et al., 1986,

MacDonald et al., 1990, Smith et al., 1989, Dando et al., 1992). Sulfide, however, is toxic to the respiratory pathway of all metazoans, poisoning the electron transport chain at cytochrome c oxidase (Grieshaber & Völkel, 1998, Somero et al., 1989). Vent and seep environments provide additional challenges for vestimentiferans, including the extreme pressure associated with depth, the presence of additional toxic chemicals in vent or seep fluid, and for vent organisms, extremes of temperature over very short distances and time scales. Chapter 2 provides an introduction to the geology and geochemistry of the vent and seep environments, discusses the implications of these extreme environments for metazoan physiology, and introduces adaptations of organisms such as vestimentiferans to such environments.

Reports of hydrothermal vent tubeworms are limited to the Pacific Ocean, with four species described from the East Pacific and Northeast Pacific spreading centers (Tunnicliffe et al., 1998), and two other vent taxa found in the western Pacific near Japan (Kojima et al., 2002, 2003). No vestimentiferans have been found at the hydrothermal vents along the much-studied Mid Atlantic

Ridge, and no vestimentiferans have yet been found at venting sites in the Indian Ocean. Cold seep vestimentiferan communities are more globally distributed than the hydrothermal vent vestimentiferans, and have been found in both the Atlantic and Pacific Oceans, as well as in the

Mediterranean and on the west coast of Java (Sibuet & Olu, 1998, McMullin, 2003). Based on their frequency of occurrence in explored seep sites, seep vestimentiferans are likely to be present in the less-explored Indian Ocean as well. Though vent and seep vestimentiferans have been reported from many locations, many have not yet been identified to the species or even genus level. 4 A large amount of the vestimentiferan body is occupied by the trophosome, a specialized tissue containing the intracellular gamma proteobacterial symbionts (Powell & Somero, 1986).

Vestimentiferans appear to have only one species of bacterium present within an animal (Distel et al., 1988), and evidence supports the acquisition of symbionts from the environment by vestimentiferan larvae (Cary et al., 1989, 1993). Therefore both vestimentiferan larval dispersal and symbiont dispersal define the biogeography of these organisms. Previous work with symbiont phylogenetics was done through comparisons of the sequence of the 16S ribosomal RNA gene (16S rRNA). These analyses revealed little of vestimentiferan symbiont biogeography other than the existence of a deep division between the symbionts of vent and seep vestimentiferans and the presence of three clusters of symbionts within the seep vestimentiferans (Feldman et al., 1997,

Nelson & Fisher, 2000).

Chapter 3 focuses on the biogeography of both the vestimentiferan hosts and their symbionts using sequence comparisons of either the host mitochondrial cyotochrome oxidase I gene or the symbiont 16S ribosomal RNA gene. The chapter first provides a comprehensive overview of the current knowledge on vestimentiferan host and symbiont biogeography. Second, new data are presented for the seep vestimentiferan hosts to explore species distribution and to define the species present on the Upper Louisiana Slope of the Gulf of Mexico, which are the subject of the population studies presented in Chapter 5. Additionally, these data combined with already published data are used to explore evolutionary rates among the major groups of vestimentiferans. Third, a similar study of evolutionary rate is conducted for the vestimentiferan symbionts, and new data are provided for seep symbionts which suggest that seep symbiont strains have a global occurrence that may be limited by depth and not geography.

Study of the reproduction and dispersal of vestimentiferan species is seriously affected by the remoteness of these communities and by the fact that sites are visited for just a few hours at a 5 time. Additionally, due to weather, funding, and scheduling issues, sites are often visited at the same time each year, or visits are separated by spans of well over a year (Tyler & Young, 1999).

Previous work has concentrated on the reproduction and dispersal in the tubeworm species of the

East Pacific hydrothermal vents. Allozyme and DNA sequence data have revealed little population differentiation among these species over large distances (3000+km), indicating either long distance dispersal or the presence of many vent “stepping stones” along the axis (Black et al., 1994,

Tunnicliffe et al., 1998, Hurtado, 2002). Seep tubeworms occur in a very different environment from hydrothermal vents, and appear to be different from vent vestimentiferans in such physiological and ecological traits as metabolic rate, lifespan, fecundity, and generation time

(Freytag et al., 2001, Bergquist et al., 2000, 2002). Seep vestimentiferans, like vent vestimentiferans, appear to have very large species ranges but research has thus far been limited to gene and promoter sequence comparisons (reviewed in McMullin, 2003, Kojima et al., 2003).

Chapters 4 presents the molecular genetic tools used in Chapter 5 to address the question of dispersal and genetic diversity of two species of deep sea hydrocarbon seep vestimentiferans tube worms, Seepiophila jonesi and Lamellibrachia cf. luymesi, from the Louisiana Slope of the Gulf of

Mexico (GoM). Hydrocarbon seepage on the Louisiana Slope (LS) supports large communities of these co-occurring cold seep vestimentiferan species, which are found in regions of active oil seepage (MacDonald et al., 1989, 1990). Vestimentiferans have separate sexes and most evidence supports reproduction by spawning of gametes into the water column (Young et al., 1996, Tyler &

Young, 1999). The motile larval stage settles on, and remains attached to, exposed carbonate rock formed in areas of active oil seepage (Fisher et al., 1997). Settlement of both species of larvae occurs in a limited time window (~20 years) after the formation of the exposed carbonate rock and before the rock becomes buried in sediment (Bergquist et al., 2002). Because settlement occurs on patchy exposed surfaces, vestimentiferan communities on the Louisiana Slope are characterized by 6 large bush-like aggregations of similarly-aged individuals, which can persist in excess of 200 years

(Bergquist et al., 2000, 2002).

The seep communities of the LS are a significant source of organic matter in the deep sea, and may be a resource for non-seep deep sea fauna (MacAvoy et al., 2002). The location of hydrocarbon seep communities and oil company interests often overlap, and seep communities can be seriously damaged during the building and maintenance of oil rigs. Genetically diverse and interbreeding populations that span wider geographical ranges are theoretically more able to withstand damage and loss of individuals within a single site than highly fragmented and isolated communities (Wright, 1977, 1978). Microsatellite markers were developed (Chapter 4) to address questions of genetic diversity and geneflow among sites. Microsatellite DNA is repetitive coding or noncoding DNA with a higher mutation rate than most genomic DNA (Weber JL, 1993, Schug et al., 1998). This high mutation rate generally results in the maintenance of a large number of alleles per locus within populations of a species. This variation within a species can be used to reveal reproductive characteristics that violate the assumptions of the Hardy-Weinberg Equilibrium

(HWE), such as the existence of nonrandom mating or underlying population structure (Wright,

1969). Microsatellites are found at varying levels throughout eukaryotic genomes (Tautz & Renz,

1984) and the frequency of microsatellite DNA within a particular organism varies with taxonomic group (Zane et al., 2002). Details on the development of these markers are presented in Chapter 4 with in depth information on the eleven polymorphic loci used in further population studies.

Additional information for these 11 loci and other repetitive sequences isolated but not used in further studies are provided in Appendix A.

The microsatellite primers isolated in Chapter 4 were used in a population study of L. luymesi and S. jonesi from nine sites on the LS separated by between 2 to 580km (Chapter 5).

Individuals were sampled from between one and four aggregations within eight of the sites. The ninth site, Bush Hill, was the subject of intensive sampling, with individuals of each species taken 7 from 10 or more aggregations. The sampling plan in Bush Hill, based on evidence that tube length is correlated with tubeworm age (Bergquist et al., 2000), included at least two small ‘juvenile’ aggregations, two large ‘adult’ aggregations, and two ‘senescent’ aggregations covered with epifaunal growth. Genomic DNA was isolated from 235 L. luymesi and 165 S. jonesi, and individuals were PCR screened with primers to five microsatellite loci in L. luymesi and seven in S. jonesi. Because mutations in microsatellite DNA generally result in the loss or gain of a repeat unit, microsatellite alleles are distinguishable by size of PCR product. In Chapter 5, numbers of alleles at each loci are used to explore overall genetic diversity. Allele frequencies are compared within individuals, within aggregations, within sites, and between sites to reveal departures from HWE predicted values at each possible level of population structure. Allele frequencies are also compared between each pair of collections to reveal clustering of collections within or between sites, and to explore the correlation between the genetic distance and geographic distance.

8 References:

Bergquist, D. C., Urcuyo, I. A. and Fisher, C. R. (2002). Establishment and persistence of seep vestimentiferan aggregations on the upper Louisiana slope of the Gulf of Mexico. Marine Ecology-Progress Series 241: 89-98.

Bergquist, D. C., Williams, F. M. and Fisher, C. R. (2000). Longevity record for deep-sea invertebrate. Nature 403: 499-500.

Black, M. B., Lutz, R. A. and Vrijenhoek, R. C. (1994). Gene flow among vestimentiferan tube worm (Riftia pachyptila) populations from hydrothermal vents of the Eastern Pacific. Marine Biology 120: 33-39.

Cary, S. C., Felbeck, H. and Holland, N. D. (1989). Observations on the reproductive biology of the hydrothermal vent tube worm, Riftia pachyptila. Mar. Ecol. Prog. Ser. 52: 89-94.

Cary, S. C., Warren, W., Anderson, E. and Giovannoni, S. J. (1993). Identification and localization of bacterial endosymbionts in hydrothermal vent taxa with symbiont-specific polymerase chain reaction amplification and in situ hybridization techniques. Molecular Marine Biology and Biotechnology 2: 51-62.

Cavanaugh, C. M., Gardiner, S. L., Jones, M. L., Jannasch, H. W. and Waterbury, J. B. (1981). Prokaryotic cells in the hydrothermal vent tube worm Riftia pachyptila: Possible chemoautotrophic symbionts. Science 213: 340-342.

Childress, J. J. and Fisher, C. R. (1992). The biology of hydrothermal vent animals: physiology, biochemistry and autotrophic symbioses. Oceanogr. Mar. Biol. Annu. Rev. 30: 337-441.

Childress, J. J., Fisher, C. R., Favuzzi, J. A., Kochevar, R., Sanders, N. K. and Alayse, A. M. (1991). Sulfide-driven autotrophic balance in the bacterial symbiont-containing hydrothermal vent tubeworm, Riftia pachyptila Jones. Biol. Bull. 180: 135–153.

Corliss, J. B., Dymond, J., Gordon, L. I., Edmond, J. M., Herzen, R. P. V., Ballard, R. D., Green, K., Williams, D., Bainbridge, A., Crane, K. and Andel, T. H. v. (1979). Submarine thermal springs on the Galapagos Rift. Science 203: 1073-1083.

Dando, P. R., Southward, A. J., Southward, E. C., Dixon, D. R., Crawford, A. and Crawford, M. (1992). Shipwrecked tube worms. Nature 356: 667.

Distel, D. L., Lane, D. J., Olsen, G. J., Giovannoni, S. J., Pace, B., Pace, N. R., Stahl, D. A. and Felbeck, H. (1988). Sulfur-oxidizing bacterial endosymbionts: analysis of phylogeny and specificity by 16 S rRNA sequences. J. Bacteriol. 170: 2506-10.

Feldman, R. A., Black, M. B., Cary, C. S., Lutz, R. A. and Vrijenhoek, R. C. (1997). Molecular phylogenetics of bacterial endosymbionts and their vestimentiferan hosts. Molecular Marine Biology and Biotechnology 6: 268-277.

Fisher, C. R., Urcuyo, I. A., Simpkins, M. A. and Nix, E. (1997). Life in the slow lane: Growth and longevity of cold-seep vestimentiferans. Marine Ecology 18: 83-94. 9

Freytag, J. K., Girguis, P. R., Bergquist, D. C., Andras, J. P., Childress, J. J. and Fisher, C. R. (2001). A paradox resolved: Sulfide acquisition by roots of seep tubeworms sustains net chemoautotrophy. Proceedings of the National Academy of Sciences of the United States of America 98: 13408-13413.

Grieshaber, M. K. and Völkel, S. (1998). Animal adaptations for tolerance and exploitation of poisonous sulfide. Annu. Rev. Physiol. 60: 33-53.

Halanych, K. M., Feldman, R. A. and Vrijenhoek, R. C. (2001). Molecular evidence that Sclerolinum brattstromi is closely related to vestimentiferans, not to frenulate pogonophorans (Siboglinidae, Annelida). Biological Bulletin 201: 65-75.

Hashimoto, J., Miura, T., Fujikura, K. and Ossaka, J. (1993). Discovery of Vestimentiferan Tube- Worms in the Euphotic Zone. Zoological Science 10: 1063-1067.

Hurtado, L. A. (2002). Evolution and biogeography of hydrothermal vent organisms in the eastern Pacific Ocean. New Brunswick, NJ: Rutgers University.

Johnson, K. S., Beehler, C. L., Sakamoto-Arnold, C. M. and Childress, J. J. (1986). In situ measurements of chemical distributions in a deep-sea hydrothermal vent field. Science 231: 1139-1141.

Kennicutt, M. C., II, Brooks, J. M., Bidigare, R. R., Fay, R. R., Wade, T. L. and McDonald, T. J. (1985). Vent-type taxa in a hydrocarbon seep region on the Louisiana Slope. Nature 317: 351-353.

Kojima, S., Ohta, S., Yamamoto, T., Miura, T., Fujiwara, Y., Fujikura, K. and Hashimoto, J. (2002). Molecular taxonomy of vestimentiferans of the western Pacific and their phylogenetic relationship to species of the eastern Pacific: II. Families Escarpiidae and Arcovestiidae. Marine Biology 141: 47-55.

Kojima, S., Ohta, S., Yamamoto, T., Yamaguchi, T., Miura, T., Fujiwara, Y., Fujikura, K. and Hashimoto, J. (2003). Molecular taxonomy of vestimentiferans of the western Pacific, and their phylogenetic relationship to species of the eastern Pacific III. Alaysia-like vestimentiferans and relationships among families. Marine Biology 142: 625-635.

MacAvoy, S. E., Carney, R. S., Fisher, C. R. and Macko, S. A. (2002). Use of chemosynthetic biomass by large, mobile, benthic predators in the Gulf of Mexico. Marine Ecology- Progress Series 225: 65-78.

MacDonald, I. R., Boland, G. S., Baker, J. S., Brooks, J. M., Kennicutt, M. C., II and Bidigare, R. R. (1989). Gulf of Mexico hydrocarbon seep communities. II. Spatial distribution of seep organisms and hydrocarbons at Bush Hill. Mar. Biol. 101: 235-247.

MacDonald, I. R., Guinasso, N. L., Reilly, J. F., Brooks, J. M., Callender, W. R. and Gabrielle, S. G. (1990). Gulf of Mexico hydrocarbon seep communities: VI. patterns in community structure and habitat. Geo-Mar. Lett. 10: 244-252.

10 McHugh, D. (1997). Molecular evidence that echiurans and pogonophorans are derived annelids. Proceedings of the National Academy of Sciences of the United States of America 94: 8006- 8009.

McMullin, E. R., Hourdez, S., Schaeffer, S.W., and Fisher, C.R. (2003). Phylogeny and biogeography of deep sea vestimentiferan tubworms and their bacterial symbionts. Symbiosis 34: 1-41.

Nelson, K. and Fisher, C. R. (2000). Absence of cospeciation in deep-sea vestimentiferan tube worms and their bacterial endosymbionts. Symbiosis 28: 1-15.

Paull, C. K., Hecker, B., Commeau, R., Freeman-Lynde, R. P., Neumann, C., Corso, W. P., Golubic, S., Hook, J. E., Sikes, E. and Curray, J. (1984). Biological communities at the Florida escarpment resemble hydrothermal vent taxa. Science 226: 965-967.

Powell, M. A. and Somero, G. N. (1986). Adaptations to sulfide by hydrothermal vent animals: sites and mechanisms of detoxification and metabolism. Biol. Bull. 171: 274-290.

Rouse, G. W. (2001). A cladistic analysis of Siboglinidae Caullery, 1914 (Polychaeta, Annelida): formerly the phyla Pogonophora and Vestimentifera. Zoological Journal of the Linnean Society 132: 55-80.

Schug, M. D., Hutter, C. M., Noor, M. A. and Aquadro, C. F. (1998). Mutation and evolution of microsatellites in Drosophila melanogaster. Genetica 102: 359-367.

Sibuet, M. and Olu, K. (1998). Biogeography, biodiversity and fluid dependence of deep-sea cold- seep communities at active and passive margins. Deep-Sea Research Part Ii-Topical Studies in Oceanography 45: 517-+.

Smith, C. R., Kukert, H., Wheatcroft, R. A., Jumars, P. A. and Deming, J. W. (1989). Vent fauna on whale remains. Nature 341: 27-28.

Somero, G. N., Anderson, A. E. and Childress, J. J. (1989). Transport, metabolism and detoxification of hydrogen sulfide in animals from sulfide-rich marine environments. Rev. Aquat. Sci. 1: 591-614.

Tautz, D. and Renz, M. (1984). Simple sequences are ubiquitous repetitive components of eukaryotic genomes. Nucleic Acids Research 12: 4127-38.

Tunnicliffe, V., McArthur, A. and McHugh, D. (1998). A biogeographical perspective of the deep- sea hydrothermal vent fauna. Advances in Marine Biology 34: 353-442.

Tyler, P. A. and Young, C. M. (1999). Reproduction and dispersal at vents and cold seeps. J Mar Biol Assn Uk 79: 193-208.

Weber JL, W. C. (1993). Mutation of human short tandem repeats. Human Molecular Genetics 2: 1123-8.

11 Wright, S. (1969). Evolution and the Genetics of Populations: Theory of Gene Frequencies, vol. 2. Chicago: University of Chicago Press.

Wright, S. (1977). Evolution and the Genetics of Populations: Experimental Results and Evolutionary Deductions, vol. 3, pp. 614. Chicago: The University of Chicago Press.

Wright, S. (1978). Evolution and the Genetics of Populations: Variability within and among natural populations, vol. 4, pp. 565. Chicago: The University of Chicago Press.

Young, C. M., Vazquez, E., Metaxas, A. and Tyler, P. A. (1996). Embryology of vestimentiferan tube worms from deep-sea methane/sulphide seeps. Nature 381: 514-516.

Zane, L., Bargelloni, L. and Paternello, T. (2002). Strategies for microsatellite isolation: a review. Molecular Ecology 11: 1-16. 12

CHAPTER 2

Metazoans in Extreme Environments: Adaptations of Hydrothermal Vent and Hydrocarbon Seep Fauna

Published in: Gravitational and Space Biology Bulletin Volume 13, Issue 2, June 2000 Pages 13-23

Metazoans in Extreme Environments: Adaptations of Hydrothermal Vent and Hydrocarbon Seep Fauna Erin R. McMullin, Derk C. Bergquist and Charles R. Fisher* Department of Biology, 208 Mueller Lab, Pennsylvania State University, University Park PA

ABSTRACT (reviewed in Somero, 1998). High pressure can alter enzyme Some of the most extreme environments where animals kinetics and efficiency, change protein structure (reviewed in survive are associated with active vents and seeps in the deep sea. Gibbs, 1997). The sheer distance of much of the ocean floor from a In addition to the extreme pressure, low temperatures, and lack of source of photosynthetic production makes it largely a nutrient- light that characterize the deep sea in general, a variety of other poor environment, where any carbon consumed by an organism factors that are hostile to most animals prevail in these must be strictly conserved (Sibuet, 1992; Gage and Tyler, 1996). A environments. Hydrothermal vent regions show extremes in diverse array of metazoans have responded to this combination of temperature, areas of very low oxygen, and the presence of toxic environmental extremes by developing specialized proteins and hydrogen sulfide and heavy metals. Hydrocarbon seeps, though lipids, capable of functioning under temperatures and pressures much cooler than vents, also have regions of very low oxygen and high hydrogen sulfide, as well as other potentially harmful that would hinder shallower-dwelling species, and depressed substances such as crude oil and supersaturated brine. Specially metabolic rates and activity that conserve a limited energy supply adapted animals not only tolerate these conditions, they often (reviewed in Gibbs, 1997 and Somero, 1998). thrive under them. In most cases this tolerance is due to a Hydrothermal vent environments have the same high combination of physiological and behavioral adaptations that allow pressure and continuous darkness characteristic of the deep sea in animals to avoid the extremes of their habitats and yet benefit from general, but they differ in nutrient supply, temperature variation, the chemoautotrophic production characteristic of these oxygen concentration, pH, and levels of such potentially toxic environments. chemicals as sulfide and heavy metals (Childress and Fisher, 1992). In active venting regions, entrained seawater is superheated by

deep, subsurface, hot basalt, converting geothermal energy into the INTRODUCTION chemical energy of hot, highly reduced hydrothermal fluid Metazoans are multicellular animals and, as such, are more (Jannasch, 1989). This superheated fluid rises through a system of complex in their body plans and often have stricter physiological interconnected subterranean cracks and fissures in the newly requirements than unicellular animals or bacteria. Individual cells in formed sea floor, resulting in multiple flow sites and flow these organisms perform specialized tasks, and cells of similar manifestations at the surface. In areas where the effluent reaches function are typically organized into layers or compartments. In temperatures up to 400°C and actively mixes with the cold oxygen- most cases, this means that materials cannot simply diffuse from bearing ambient seawater, chimneys or mounds may form when the environment to all cells. Therefore, systems must be in place to chemicals rapidly precipitate out of solution. Hydrothermal fluids transport materials throughout the body, to allow communication may also mix with cooler water as it rises through the basalt, between different compartments, and to regulate the internal emerging as diffuse flow fields at temperatures ranging from environment in which these processes occur. Strictly speaking, all slightly above ambient to about 100°C (Delaney et al., 1992) known metazoans are heterotrophic and must depend on the (Figure 1). oxidation of autotrophically produced organic carbon compounds Hydrocarbon seeps in the Gulf of Mexico provide a chemical for energy. This metabolic process can occur in the absence of environment similar to that of hydrothermal vents, though differing oxygen (anaerobically). However, in the presence of oxygen significantly in temperature, sulfide concentrations, and types of (aerobically), the amount of energy harnessed per unit of fixed toxins. Sulfide in seep areas is produced when organic carbon carbon oxidized increases dramatically. As metazoans often have (crude oil and natural gas) migrates to the sea floor from deep relatively large body sizes and high energy demands, they require reservoirs. Sediment layers at the sea floor diffuse and retain the not only ample supplies of fixed carbon and oxygen but also seeping oil, and sulfate reducers in the upper meters of sediment environmental conditions under which their complex biochemical produce hydrogen sulfide. The upward migration of fluids at cold pathways may function. An extreme environment for a metazoan, seeps is to some extent thermally driven (Kennicutt et al., 1992). then, may be one in which fixed carbon or oxygen is scarce or Unlike the process in vents, however, this process is more passive, absent, or where environmental conditions—such as temperature, and does not result in vigorous mixing of seep fluid and ambient pressure, or the presence of toxins—hinder or prevent bottom water. Seep fluids are normally at ambient water physiological processes. The deep sea is a relatively inhospitable environment for temperatures by the time they reach the sediment/water interface. metazoans: ambient temperature is constantly low (~2o C), Sulfide levels found in emitted seep fluid are generally much lower pressure is high, light is absent, and organic carbon is scarce. Low and more stable temporally and spatially, than levels found in temperatures may slow or impede many biochemical reactions and vents, (MacDonald et al., 1989; Julian et al., 1999). However, decrease the fluidity of lipids, a factor of primary importance to interstitial sulfide concentrations can be comparable to those found cell membrane function at hydrothermal vents (Kennicutt et al, 1989). Unlike

*Correspondence to: Charles R. Fisher: e-mail: [email protected]

Gravitational and Space Biology Bulletin 13(2), June 13 METAZOANS IN EXTREME ENVIRONMENTS

Figure 1. Hydrothermal Vents Manifest on the Sea Floor as Chimney Structures and Diffuse Flow Fields. Superheated seawater carries highly reduced compounds to the surface, where they rapidly precipitate out of solution or fuel extensive biological communities based on chemoautotrophic production. (Modified from Jannash, 1989) the fauna in vent regions, fauna in seeps experience temperatures concentrations for significant periods of time (Millero, 1986; that do not vary appreciably from ambient deep-sea temperatures Johnson et al., 1986). The needs of the symbiont or (~8°C). Methane and crude oil may also bubble through overlying free-living bacteria for sulfide and oxygen, and of the animal for sediments at these sites, and dense anoxic brines can form pools on additional oxygen, limit vent fauna to areas where hot, sulfide- the seafloor (MacDonald, 1990; MacDonald, 1998). bearing vent water and cold, oxygen-bearing ambient water actively The abundant reduced chemicals found in both hydrothermal mix (Childress and Fisher, 1992). Animals in such regions therefore vent and hydrocarbon seep fluids, hydrogen sulfide in particular, experience rapid shifts in temperature that coincide with changes in can be used by chemosynthetic prokaryotes as an energy source oxygen and sulfide concentration. Seep fauna are also limited by for carbon fixation. For metazoan life, this offers the potential for a access to sulfide and, as sulfide is present in areas of active large food source more or less independent from the overlying seepage, are exposed to potentially toxic levels of hydrocarbons. photic zone (Jannasch, 1989). Most vent and many seep organisms The extraordinarily high level of chemosynthetic-based rely entirely on this chemosynthetic primary production by primary production that is supported by venting and seepage is actively feeding on free-living bacteria or by forming symbiotic one of the outstanding differences between these regions and the relationships with chemosynthetic bacteria for the bulk of their average deep-sea environment. Metazoans are found at high food supply (Childress and Fisher, 1992; Kennicutt and Burke, densities in vent and seep regions, but they are of low diversity and 1995; Fisher, 1996). The sulfide-containing hydrothermal vent fluid have a high degree of endemism (Tunnicliffe et al., 1998; Sibuet and on which primary production relies is hot, extremely deficient in Olu, 1998). Organisms of taxonomic and functional similarity are oxygen, and laden with toxic chemicals. Moreover, sulfide and found at both seeps and vents. It appears that vents and seeps are oxygen react spontaneously and do not coexist in significant both areas of significant resources and of extreme environmental

14 Gravitational and Space Biology Bulletin 13(2), June 2000

METAZOANS IN EXTREME ENVIRONMENTS demands that are largely exploited by a limited group of animals (Chevaldonne et al., 1992). Temperatures of 20-80°C have been with specialized physiological and biochemical adaptations measured on surfaces colonized by P. sulfincola and P. palmiformis (Childress and Fisher, 1992; Fisher, 1996). (Juniper et al., 1992), and temperature measurements taken in alvinellid tubes have recorded temperatures of 68°C, with spikes as TEMPERATURE high as 80°C (Desbruyères et al., 1998). These recordings have led Temperature variation is one of the striking characteristics of some authors to conclude that alvinellids regularly experience such hydrothermal vent environments. Water temperature can range high temperatures, and that alvinellid tubes may open to vent fluid from 2°C to 400°C within a centimeter, and animals may have at the back, allowing warm vent water to flow outward over the occasional brief contact with 100°C+ water (Chevaldonne et al., animals (Cary et al., 1998). 1992; Delaney et al., 1992; Cary et al., 1998). In areas of diffuse Measurements of temperature effects on alvinellid proteins, flow, a vent animal may experience water temperatures of 2, 20 and however, indicate that these animals cannot survive body 40+°C in rapid succession, or even simultaneously over the length temperatures over 50°C for extended periods. For example, both of its body (Johnson et al., 1988; Cary et al., 1998). Additionally, Alvinella pompejana and A. caudata have hemoglobin that is at least one life stage of these organisms must be able to withstand unstable at 50°C, with highest oxygen binding affinities occurring at extended periods of cold (~2°C) during dispersal. roughly 15° and 25°C, respectively (Toulmond et al., 1990). More Many biological structures, such as enzymes and lipid striking are the melting temperatures of alvinellid collagens, which bilayer membranes, depend on a particular degree of molecular in A. pompejana are 40°C for cuticle collagen and 46°C for instability or fluidity, which is directly affected by temperature. interstitial collagen (Gaill et al., 1991). Based on these Increasing temperatures can increase reaction rates and affect measurements, these animals likely live at temperatures averaging reaction equilibria through higher kinetic energy, and high 30 to 35°C rather than at the higher temperatures proposed by temperatures can cause protein denaturation, resulting in complete others (Childress and Fisher, 1992; Fisher, 1998). The conflict and often irretrievable loss of function (Hochachka and Somero, between observed temperature probe readings (Chevaldonne et al., 1984). Adaptation to the deep sea requires more “fluid” proteins 1992; Cary et al., 1998) and the in vitro physiological limits and lipids to compensate for the stabilizing forces of high pressure probably reflects both the steep temperature gradient that exists and low temperature (Hochachka and Somero, 1984). High within alvinellid tubes and the difference between a temperature temperature, on the other hand, tends to destabilize molecules, and probe reading and actual body temperature. Even if these animals selection in such environments is for more stable forms. At vent do not experience body temperatures of 60+°C, they live at regions, however, temperature can vary by almost 100°C within an temperatures much higher than ambient and may even take animal’s habitat (Cary et al., 1998; Desbruyères et al., 1998). How, advantage of the large temperature gradient present over the lengths then, do vent organisms maintain proper function in such extremes? of their bodies—from 22°C near the gills to 60+°C at the body As with all metazoan challenges, this threat can be met trunk (Cary et al., 1998) (see discussion of alvinellid hemoglobin). through morphological, physiological, and behavioral adaptations. Animal survival in extreme temperatures depends as much on Biochemically, changes that favorably affect reaction rate equilibria behavior as on physical adaptations. Different species in vent and molecular stability become fixed in a population. A number of fields have different environmental requirements, and they will molecules show a particularly strong correlation between functional either settle in or migrate to regions that meet their particular needs. and denaturation temperatures and the temperature of an Within specific habitats, animals may also have behavioral organism’s environment (Hochachka and Somero, 1984). Molecules mechanisms for modifying environmental conditions to their of particular interest are enzymes, collagen, and lipids. Enzymes benefit. Paralvinella sulfonicola colonizes young sulfide chimneys quickly lose function above and below an optimal temperature. The in the early stages of sulfide mineralization, where both subunits of the structural polymer collagen, the most abundant temperature and sulfide levels have begun to decrease (Juniper and animal protein, often show a melting temperature (Tm) close to the Martineau, 1995). Additionally, P. sulfonicola has been found to upper lethal limit for an animal, though the polymer itself is have a 2-mm-thick layer of FeS2 below its tubes, which forms a somewhat more stable. Lipids of cell membrane bilayers must be barrier between hot vent water and cold sea water—a barrier that both fluid and structurally coherent to form a functional membrane, may be formed by the activity of the animal itself (Juniper and a characteristic that is also is very sensitive to temperature change Martineau, 1995). Alvinellids, in general, may also actively cool (Hochachka and Somero, 1984). their tubes by pumping in ambient seawater (Chevaldonne et al., Many hydrothermal vent animals must have a wide 1991). temperature tolerance (especially compared to ambient deep-sea Though the vent animals described here likely withstand fauna). Of particular interest are such chimney-dwelling body temperatures that rival those of the most thermotolerant polychaetes as Alvinella pomejana, A. caudata, and Paralvinella metazoans, the best-documented thermotolerant animals are found sulfincola, which live on newly formed vent walls very near the in other, more easily studied environments. Desert-adapted bees super-hot vent fluid. Photographic data from one such site show an fly with a sustained internal temperature of 46°C (Willmer and alvinellid crawling over a temperature probe that is reading 105°C Stone, 1997), and desert ants can survive even hotter temperatures.

Gravitational and Space Biology Bulletin 13(2), June 2000 15 METAZOANS IN EXTREME ENVIRONMENTS

The Australian ant Melophorus bagoti actively forages at soil do not normally co-occur (Millero, 1986). temperatures above 70°C and has a critical thermal maximum of Vent and seep animals show a variety of behavioral and 56.7°C, surviving for one hour at 54°C (Christian and Morton, physiological adaptations to low and variable oxygen tensions. 1992). Unlike desert insects, however, vent organisms live in a Behavioral mechanisms that help to maintain an aerobic metabolism in these environments include: world of rapidly shifting temperature extremes: within minutes, water temperatures vary in tens of degrees at a given location and · spatially spanning an oxic/anoxic transition zone; in hundreds of degrees over centimeters. Vent organisms must · temporally spanning the transition zone by moving therefore be adapted to extreme temperature variation as well as to from oxic to anoxic water; extreme temperatures. · acquiring oxygen by pumping oxic water into an anoxic burrow or tube.

HYPOXIA/ANOXIA Alvinellids build tubes that protrude from chimney walls, allowing Metazoan metabolic energy is produced via the release and their gills access to oxygen-bearing ambient water. Mobile transfer of electrons from reduced-carbon electron donor molecules predators, such as Bythograea thermydron, may move between to more oxidized electron acceptors. In aerobic metabolism, the pools of oxic and anoxic water. Hesiocaeca methanicola, a electron acceptor is oxygen; in metazoan anaerobic metabolism, the polychaete that can live on buried anoxic methane hydrates in the electron acceptor is an organic molecule such as lactate or fumarate. cold seeps of the Gulf of Mexico, appears to increase circulation of Aerobic respiration yields 36 ATP/mol glucose oxidized, while oxygenated water in its habitat using its parapodia (Fisher et al., anaerobic respiration yields 2-8 ATP/mol glucose, depending on 2000), as might the alvinellid polychaetes that live in tubes on the pathway and electron acceptor utilized (Fenchel and Finlay, hydrothermal chimneys (Desbruyères, 1998). 1995). Many organisms found in hypoxic environments are able to Metazoans rely primarily on the efficient, high-energy maintain aerobic respiration and normal metabolic rate even at very output of aerobic respiration to maintain a normal level of low oxygen tensions (oxyregulation of respiration). A number of metabolism. Metazoans forced to use inefficient fermentation for physiological adaptations are seen in oxyregulators, including large energy production over extended periods need a large and surface areas for gas exchange, short diffusion distances from constantly replenished food source to maintain normal bodily external surface to blood spaces, a well-developed function and reproduction, a situation not common in nature, although endoparasites are a possible exception (Bryant, 1991). Oxygen is also involved in a number of critical biosynthetic pathways, and no metazoan has yet been conclusively documented to complete its entire lifecycle without its presence, though obligate anaerobe eukaryotic protozoa do exist (Fenchel and Finlay, 1995). Some nematode and oligochaete meiofauna, which occur deep in the sulphidic zone of sediments, may prove to be capable of growth and reproduction in the absence of oxygen (these environments have not yet been established as completely anaerobic) (Fenchel and Finlay, 1995). Yet numerous metazoans from diverse groups and environments (e.g., nematodes, annelids, amphipods, and goldfish) can withstand extended periods of anoxia (Bryant, 1991; Panis et al., 1996; Hagerman et al., 1997). Vent and seep fluids are highly reduced and contain significant levels of sulfide. Oxygen in vent habitats varies inversely with temperature, and organisms in areas of actively mixing hydrothermal and ambient water may experience rapid fluctuations in both (Childress and Fisher, 1992) (Figure 2). At hydrocarbon seeps, oxygen concentration decreases with proximity to the substrate and with increased depth in the sediment (Kennicutt et al., 1989). Seep organisms colonize both brine pools Figure 2. Temperature, Oxygen, and Sulfide and methane hydrates, which are habitats of particularly low Concentrations Measured in Situ at the Galapagos Rift. The oxygen (MacDonald, 1990; Fisher et al., 2000). Metazoans with sampling probe was moved from ambient bottom water to within a chemoautotrophic sulfide-oxidizing symbionts at vent and seep vent mussel bed and then back, a distance of about 50 cm. Oxygen sites require oxygen for aerobic respiration, as well as both oxygen and sulfide react spontaneously, and do not coexist for significant and sulfide to power chemoautotrophic carbon fixation. This dual amounts of time. In areas where hot, sulfide-rich vent water mixes requirement can prove a challenge for metazoans that are dependent with cold, oxygen-rich ambient water, oxygen concentration upon chemoautotrophic symbionts, because they will require decreases as water temperature and sulfide increase. (Johnson et substantial amounts of both oxygen and sulfide, two chemicals that al., 1986)

16 Gravitational and Space Biology Bulletin 13(2), June 2000

METAZOANS IN EXTREME ENVIRONMENTS circulatory system, and the presence of respiratory pigments where roughly two-thirds of the oxygen consumption occurs (Weber, 1978; Bryant, 1991). Some vent fauna, such as B. (reviewed in Childress and Fisher, 1992). The utility of the thermydron, have been shown to oxyregulate to very low levels of Bohr effect for oxygen offloading to the trophosome was environmental oxygen (Mickel and Childress, 1982). A seep- questioned by Childress and Fisher (1992) because CO2 endemic orbiniid and the methane hydrate polychaete H. consumption by the symbionts in the trophosome would methanicola can also oxyregulate to very low oxygen tensions overshadow tissue CO2 production. However, Goffredi et al. (Fisher et al., 2000; Hourdez et al., personal communication). (1999) have recently demonstrated a net H+ ion production One common adaptation to improve oxygen exchange is the by autotrophic Riftia, indicating H+ ion production in the development of large gills or specialized surfaces for gas exchange, trophosome, which could explain the utility of the Bohr such as the plume of R. pachyptila (Jones, 1981, 1988; Arp et al., effect in this symbiosis. 1985) and the hypertrophied gills of Alvinellids (Jones, 1981; In the absence of oxygen, some organisms can use anaerobic Jouin and Gaill, 1990). Decreased diffusion distances between gas metabolism for extended periods. Indeed, sulfide may poison exchange surfaces and the blood supply, another common aerobic respiration at the electron transport chain, forcing the adaptation that facilitates uptake of dissolved gases, has been organism to rely on anaerobic metabolism even in the presence of documented in both alvinellids (Jouin and Gaill, 1990) and the seep oxygen (Bryant, 1991). In general, annelids and molluscs are able to orbiniid (Hourdez et al., 2000). Finally, well-developed and highly use more efficient mitochondrial pathways of fermentation vascularized circulatory systems have been found in alvinellids, (Bryant, 1991; Fenchel and Finlay, 1995; Tielens and Van vestimentiferans, and a seep orbiniid (Jones, 1981, 1988; Jouin et Hellemond, 1998), whereas no crustacean has been found that can al., 1996; Hourdez et al., 2000). Respiratory pigments, such as use a pathway beyond glycolysis, which yields only two to three hemoglobin or hemocyanin, with high affinities and ATP/glucose (reviewed in: DeZwaan and Putzer, 1985; and capacities for oxygen, are a particularly useful adaptation for Bryant, 1991). Again, only selected vent and seep species an organism experiencing low and/or variable oxygen have been tested for anaerobic tolerance. The vent crab tensions. Although the respiratory pigments of most vent Bythograea thermydron can survive only about 12 hours in and seep fauna have not been characterized, both Riftia the absence of oxygen, and the glycolytic endproduct pachyptila and Alvinella spp. contain circulating lactate is accumulated during this time (Mickel and hemoglobins with very high oxygen affinities (Terwilliger et Childress, 1982). However, Riftia tolerates anoxia up to 60 al., 1980; Arp and Childress, 1981; Terwilliger and Terwilliger, hours and accumulates succinate when kept under anaerobic 1984; Toulmond et al., 1990), which allows them to take up conditions, indicating the use of a modified citric acid cycle oxygen from very low concentrations and accumulate it to for fermentation (Arndt et al., 1998). Similarly, both the help withstand short periods of anoxia. B. thermydron hydrate worm Hesiocaeca methanicola and a seep orbiniid hemocyanin affinity is increased by the presence of can survive four to five days in the absence of oxygen thiosulfate (a sulfide detoxification product) and lactate (a (Fisher et al., 2000; Hourdez et al., 2000). byproduct of anaerobic metabolism) (Sanders and Childress, Thus we see the same pattern with respect to oxygen 1992). Riftia hemoglobins (R. pachyptila has three different that we saw with temperature. The vent and seep fauna that Hb's) bind sulfide, as well as oxygen, with high affinity and have been investigated are not significantly more tolerant of at high capacity, allowing simultaneous transport of both anoxia or high temperature than the best-adapted fauna from gasses to the symbionts in their internal trophosome while other environments, but are certainly very well adapted for preventing the reaction of sulfide and oxygen in the blood extremes in these parameters. (Arp and Childress, 1983; Childress et al., 1984). Alvinellid hemoglobin oxygen affinity is reduced by low pH (normal TOXICITY Bohr effect) and high temperature (Toulmond et al., 1990), Potentially toxic chemicals abound in hydrothermal vent and which may facilitate the uptake and delivery of oxygen from cold seep environments (Corliss et al., 1979; Johnson et al., 1986; the plume, normally extended outside of the tube in cooler McDonald, 1990; Nix et al., 1995). Of these, sulfide is perhaps the waters, to the body, which is often bathed in highly reduced most abundant and well studied, and its consequences for biological high-temperature fluids (Desbruyères et al, 1998; Cary et al., systems have been well documented in many other reducing 1998). environments, including mud flats, mangrove swamps, and sewage Riftia hemoglobins have such a high affinity for outfalls (see reviews in Somero et al., 1989 and Grieshaber and oxygen that study of the binding properties is difficult, Volkel, 1998). Heavy metals may also occur in extremely high which has led to some variation among the results of concentrations at hydrothermal vents, where they precipitate out different investigators. Overall, the data suggest that oxygen of solution to form chimneys and sometimes coat tubeworm tubes binding by Riftia hemoglobins shows a moderate normal and mollusk shells. In spite of our limited knowledge regarding Bohr and temperature effect that may assist offloading to the specific adaptations to these toxins in vent and seep organisms, more posterior animal tissues, particularly in the trophosome, what we know of their physiologies and those of their shallower-

Gravitational and Space Biology Bulletin 13(2), June 2000 17 METAZOANS IN EXTREME ENVIRONMENTS dwelling relatives should allow us to investigate some potential are exposed to sulfide in the surrounding water. mechanisms for detoxifying these substances. Likewise, oxidation of sulfide to more benign sulfur compounds, most commonly thiosulfate, may occur within the animal by a variety of means, including sulfide-oxidase enzymes Sulfide and mitochondrial oxidation. The vent crab Bythograea thermydron maintains aerobic metabolism by steadily increasing its rate of Sulfide is a toxin that, in just micromolar amounts, is capable oxygen consumption up to environmental sulfide concentrations of of impairing biological processes necessary to metazoan function. about 800mM and apparently detoxifying sulfide via a sulfide- Its most important physiological effect may be to severely inhibit oxidase (Vetter et al., 1987; Childress and Fisher, 1992). Sulfide- aerobic respiration by interfering with cellular respiration and blood oxidizing activity has also been found in tissues from several other oxygen transport (reviewed in: Somero et al., 1989; Vismann, 1991; vent species, including the crab Munidopsis subsquamosa, the Grieshaber and Völkel, 1998). In the mitochondria, sulfide may shrimp Alvinocaris lusca, Riftia pachyptila, and Calyptogena poison the respiratory enzyme cytochrome c oxidase, thus magnifica (Vetter et al., 1987; Powell and Somero, 1986b), as well inhibiting ATP production by the electron transport chain. Sulfide as a host of species living in non-vent, reducing habitats (Lee et al., may also bind to the hemoglobin molecule in blood, reducing its 1996; Grieshaber and Völkel, 1998). Oxidation of sulfide may also capacity to carry oxygen and, in high concentrations, rendering it be coupled to energy production directly in the mitochondria of nonfunctional. In addition, a recent study found that sulfide is some animals or indirectly by providing reduced sulfur capable of inhibiting muscular contraction independent of its intermediates to symbionts. Solemya reidi, a clam that inhabits effects on aerobic metabolism (Julian et al., 1998). areas organically enriched by sewage and paper mill effluent, links To avoid these toxic effects, an organism has several practical sulfide oxidation to ATP production in its mitochondria at low-to- options: avoid sulfide, switch to anaerobic metabolism, exclude moderate sulfide concentrations, but this ability becomes inhibited sulfide from sensitive tissues, or oxidize sulfide to more benign at high concentrations (Powell and Somero, 1986a). Similarly, the forms. Most inhabitants of vent and seep environments do not intertidal lugworm Arenicola marina possesses the ability to realistically have the option of avoiding sulfide altogether. Species oxidize sulfide in its mitochondria even at very high sulfide containing sulfide-oxidizing bacteria must supply this chemical to concentrations (Volkel and Grieshaber, 1996). It is not known their symbionts, thus requiring them to inhabit areas where sulfide whether sulfide oxidation by animal tissues of the vent mussel is abundant. Nonsymbiotic endemic heterotrophs not only must Bathymodiolus thermophilus is linked directly to ATP production, forage in areas where at least brief exposure is likely, but some but the resultant thiosulfate is supplied to its bacterial must also consume symbiotic or free-living sulfide oxidizers that endosymbionts, where it is further oxidized to fuel often contain high levels of sulfide (Somero et al., 1989). To chemoautotrophic carbon fixation (Powell and Somero, 1986a; prevent poisoning of the electron transport chain in the presence of Fisher et al., 1987; Nelson and Fisher, 1995). In this way, sulfide high concentrations of sulfide, many invertebrates temporarily oxidation in the animal’s tissues is indirectly linked to energy switch from aerobic to anaerobic metabolism. (Grieshaber and production. As research continues at vents and seeps, sulfide Völkel, 1998). As discussed above, several vent and seep species oxidation mediated by mitochondria may prove to be a common have considerable anaerobic capacity; whether this occurs in the method of detoxification. presence of sulfide has not been directly tested. Exclusion from Metazoans hosting sulfide-oxidizing bacterial symbionts sensitive tissues and oxidation within the body are the two best- must not only tolerate this potential toxin but must also acquire documented strategies to prevent sulfide poisoning among vent and both sulfide and oxygen from the environment and transport them seep animals, and symbiont-containing species often use them in to the symbionts. In many cases, these animals employ specialized conjunction. blood proteins that bind sulfide reversibly to prevent inhibition of Exclusion of sulfide from tissues may involve physical, oxygen transport, poisoning of cytochrome c oxidase in the biological, or chemical barriers around or within an animal. Thick animals' tissues, and spontaneous reaction of sulfide with oxygen. tubes or cuticles may reduce or prevent exposure of some external In Riftia pachyptila, two different extracellular hemoglobins in the tissues to sulfide, and epibiotic bacteria and abundant metal ions vascular blood and one in the coelomic fluid bind sulfide and may oxidize sulfide before it makes contact with external tissues. oxygen simultaneously and reversibly, with high affinity (Arp et The Pompeii worm, Alvinella pompejana, resides on active al., 1985). Cysteine residues and disulfide groups on these chimney structures where sulfide is abundant, within a secreted hemoglobins apparently provide the sulfide-binding mechanism proteinaceous tube that it shares with epibiotic bacteria (Zal et al., 1998). Their presence as well on the extracellular (Desbruyères et al., 1998). Although specific data is lacking, the hemoglobin of A. pompejana indicates that they may be a common tube is thought to provide a regulated environment (potentially adaptation to sulfide-rich vent habitats (Zal et al., 1997; Zal et al., lower in sulfide than the immediately surrounding vent fluid), and 1998). These hemoglobins also bind sulfide with a high enough the bacteria to supply both nutrition and a means of sulfide affinity to prevent the sulfide poisoning of cytochrome c oxidase detoxification for A. pompejana (Desbruyères et al., 1998). Several (Powell and Somero, 1983). Deep-sea clams in the family other worms living in direct contact with vent fluid, including the Vesicomyidae are also capable of binding both vestimentiferans, also secrete tubes and only expose portions of their bodies directly to sulfidic fluids. These structures may, in part, serve to limit what tissues and how much tissue surface area

18 Gravitational and Space Biology Bulletin 13(2), June 2000

METAZOANS IN EXTREME ENVIRONMENTS

extracellular sulfide binding factor that binds with high affinity (Arp et al., 1984; Childress et al., 1991). Sulfide binding can only protect sensitive tissues when a sink for the bound sulfide is available to remove sulfide and maintain free binding sites. In these species, internal chemoautotrophic microbial symbionts oxidize sulfide, acting as an internal sink while providing a source of fixed carbon for the host (Figure 3). All vestimentiferans and vesicomyid clams appear to utilize a similar system to tolerate and exploit the sulfide-rich environments in which they live. This system is characterized by the exclusion of sulfide from sensitive tissues via high-affinity binding to blood components, transport to symbionts via the blood, and the oxidation of sulfide by intracellular chemoautotrophic bacteria.

Metals

Due to the interactions between circulating crustal water and hot basalts, dissolved heavy metals are particularly abundant in hydrothermal vent systems. Metals, in general, may interfere with a wide array of biological processes, including respiration, muscular function, osmoregulation, reproduction, development, and protein utilization (Luoma and Carter, 1991). Metals can also cause morphological abnormalities, histopathological problems, and instability of genetic material (Luoma and Carter, 1991). Metazoans typically detoxify absorbed or ingested metals by using metal-binding proteins (metallothioneins) and forming subcellular inclusions. These mechanisms often act jointly to consolidate and enclose excess metals, which then accumulate within tissues and/or skeletal structures over time (Beeby, 1991; Luoma and Carter, 1991). The few investigations into potential metal detoxi-fication in vent and seep fauna indicate that strategies used by these animals are not very different from those studied elsewhere. Polychaetes of the genus Paralvinella secrete mucus, rich in metallothionien-like proteins, that removes inorganic material from the epidermis and may also remove metals from the immediate external environment (Juniper et al., 1986). Metallothionien-like proteins and metal-rich inclusions have been found within the tissues of R. pachyptila, A. pompejana, C. magnifica, and a Bathymodiolid mussel from the Mid-Atlantic Ridge, and elevated levels of metals have been found within the shells of C. magnifica and B. thermophilus (reviewed in Childress

and Fisher, 1992; Geret et al., 1998). What may separate vent and Figure 3. Mechanisms Allowing Vent Tubeworms to seep species from shallower marine taxa is not the detoxification Tolerate and Exploit Their Sulfide-rich Environments. (A) mechanism, per se, but rather the ability of the mechanism to Using its plume, the only part of its body that protrudes from the function effectively at high metal concentrations and over long open end of the chitinous tube it inhabits, Riftia pachyptila acquires exposure times. sulfide and oxygen from areas where vent fluids actively mix with ambient sea water. (B) Specialized hemoglobins in the blood of R. pachyptila simultaneously and reversibly bind both sulfide and SUMMARY AND CONCLUSIONS oxygen, and (C) transport them to internally housed, sulfur- Metazoans colonizing vent and seep habitats must tolerate oxidizing, chemoautotrophic bacteria that act as a sink for the not only the already extreme characteristics of the deep sea but also potentially poisonous sulfide. (Modified from Arp et al., 1985) a wide range of additional conditions that result from the complex

geologic and microbiological processes driving these environments. Ironically, high concentrations of hydrogen sulfide, one of the oxygen and sulfide simultaneously. However, unlike R. pachyptila, primary characteristics that should make these environments they utilize two different binding molecules: (1) an intracellular inhospitable to metazoan life, also drives biological production to hemoglobin that binds oxygen with moderate affinity, and (2) an

Gravitational and Space Biology Bulletin 13(2), June 2000 19 METAZOANS IN EXTREME ENVIRONMENTS levels far exceeding those of the surrounding deep sea. To exploit Acknowledgements the energetic abundance of vents and seeps, metazoans must The authors are supported by the National Science tolerate not only the sulfide but also a whole suite of factors Foundation, the Mineral Management Service, and the NOAA intrinsically correlated with its presence. At vents, the presence of National Undersea Research program. sulfide corresponds directly to the high temperatures, absence of oxygen, and presence of heavy metals characteristic of the effluent REFERENCES waters. At seeps, as at vents, the presence of sulfide corresponds to the absence of oxygen; but, unlike the situation at vents, it also Arndt, C., Schiedek, D., and Felbeck, H. 1998. Metabolic corresponds to the presence of potentially toxic hydrocarbons. responses of the hydrothermal vent tube worm Riftia pachyptila to Although not well studied in seep fauna, crude oil and its individual severe hypoxia. Marine Ecology Progress Series 174: 151-158. chemical components display fouling and mutagenic effects on a host of biological functions, including feeding, respiration, Arp, A.J. and Childress, J.J., 1981. Blood function in the excretion, reproduction, development, and chemoreception (Bayne hydrothermal vent vestimentiferan tube worm. Science 213:342- et al., 1982; Suchanek, 1993). 344. To cope with the numerous, potentially interacting extremes of these environments, the denizens of vents and seeps employ and Arp, A.J. and Childress, J.J. 1983. Sulfide binding properties of the combine a vast array of morphological, physiological, and blood of the hydrothermal vent tube worm Riftia pachyptila. behavioral adaptations. By inhabiting active chimney structures at Science 219:295-297. hydrothermal vents, A. pompejana and P. sulfinicola experience extremes in temperature, hypoxia, sulfide, and heavy metals. They Arp, A.J., Childress, J.J. and Fisher, C.R. 1984. Metabolic and tolerate such varied extremes by modifying their blood gas transport characteristics of the hydrothermal vent microenvironments, interacting with other organisms, and relying bivalve, Calyptogena magnifica. Physiological Zoology 57:648-662. on specialized biochemical adaptations. Combined, such adaptations allow these species to exploit an environment where Arp, A.J., Childress, J.J. and Fisher, C.R. 1985. Blood gas very few other metazoans could survive. transport in Riftia pachyptila. Biological Society of Washington Much of what we know regarding metazoan adaptations to Bulletin 6:289-300. these deep-sea reducing environments comes from investigations at hydrothermal vents. Therefore, research into other extreme deep- Bayne, B.L., Widdows, J., Moore, M.N., Salkeld, P., Worrall, sea environments is sure to expose previously unknown and novel C.M. and Donkin, P. 1982. Some ecological consequences of the adaptations. At cold seep sites on the upper Louisiana slope in the physiological and biochemical effects of petroleum compounds on Gulf of Mexico, for example, methane hydrates and brine pools, marine molluscs. Philosophical Transactions of the Royal Academy like chimney structures at hydrothermal vents, present of London, Series B 297:219-239. combinations of extreme environmental characteristics, yet they support surprisingly dense metazoan communities. Although Beeby, A. 1991. Toxic metal uptake and essential metal regulation methane hydrates are characterized by the presence of sulfide, in terrestrial invertebrates. in Metal Ecotoxicology: Concepts and methane, and crude oil, and a corresponding absence of oxygen, the Applications. Chelsea, FL: Lewis Publishers, pp. 65-89. polychaete H. methanicola flourishes on their surfaces (Fisher et al., 2000). Similarly, the brine pools in this region possess salt Bryant, C. (ed.). 1991. Metazoan Life without Oxygen. Chapman concentrations several times higher than normal seawater, almost and Hall; Suffolk, Great Britain. no oxygen, and abundant crude oil; nonetheless, massive communities of methanotrophic mussels sometimes live on their Cary, S.C., Shank, T. and Stein, J. 1998. Worms bask in extreme surfaces (MacDonald et al., 1990). temperatures. Nature. 391:545-546. The low diversity of species found at hydrothermal vents and cold seeps, as compared to other deep-sea environments, clearly Chevaldonne, P., Desbruyères, D. and Childress, J.J. 1992. …and suggests that few organisms have been able to successfully some even hotter. Nature 359:593-594. establish and reproduce under the extreme conditions these habitats present. Yet the high density, high standing-stock biomass, and Chevaldonne, P., Desbruyères, D. and Le Haitre, M. 1991. Time- high levels of production in the communities that do persist in series of temperature from three deep-sea hydrothermal vent sites. these environments indicate a significant advantage to those Deep-Sea Research 38 (11):1417-1430. organisms capable of tolerating these conditions. By tolerating extreme environments, these select few animals take advantage of Childress, J.J., Arp, A.J. and Fisher, C.R. 1984. Metabolic and an abundant yet toxic energy source, which allows them to flourish blood characteristics of the hydrothermal vent tube worm Riftia in the very nutrient-poor deep sea. pachyptila. Marine Biology 83:109-124.

Childress, J.J. and Fisher, C.R. 1992. The biology of hydrothermal vent animals: physiology, biochemistry, and autotrophic

20 Gravitational and Space Biology Bulletin 13(2), June 2000

METAZOANS IN EXTREME ENVIRONMENTS symbioses. Oceanography and Marine Biology Annual Review Fisher, C.R., MacDonald, I.R., Sassen, R., Young, C.M., Macko, 30:337-441. S.A., Hourdez, S., Carney, R.S., Joye, S. and McMullin, E. 2000. Methane ice worms: Hesiocaeca methanicola colonizing fossil fuel Childress, J.J., Fisher, C.R., Favuzzi, J.A. and Sanders, N.K. 1991. reserves. Naturwissenschaften 87:184-187. Sulfide and carbon dioxide uptake by the hydrothermal vent clam Calyptogena magnifica and its chemoautotrophic symbionts. Gage, J.D. and Tyler, P.A. 1996. Deep-Sea Biology: A Natural Physiological Zoology 64:1444-1470. History of Organisms at the Deep-Sea Floor. New York: Cambridge University Press. Christian, K.A., and Morton, S.R. 1992. Extreme thermophila in a central Australian ant, Melophorus bagoti. Physiological Zoology Gaill, F., Weidemann, H., Mann, K., Kuhn, K., Timpl, R. and 65 (5):885-905. Engel, J. 1991. Molecular character of cuticle and interstitial collagens from worms collected at deep sea hydrothermal vents. Corliss, J.B., Dymond, J., Gordon, L.I., Edmond, J.M., von Journal of Molecular Biology 221:209-223. Herzen, R.P., Ballard, R.D., Green, K., Williams, D., Bainbridge, A., Crane, K. and van Andel, T.H. 1979. Submarine thermal Geret, F., Rousse, N., Riso, R., Sarradin, P.M. and Cosson, R.P. springs on the Galapagos Rift. Science 203:1073-1083. 1998. Metal compartmentalization and metallothione isoforms in mussels from the Mid-Atlantic Ridge; preliminary approach to the Delaney, J.R., Robigou, V., McDuff, R.E. and Tivey M.K. 1992. fluid-organism relationship. Cahiers de Biologie Marine 39:291- Geology of a vigorous hydrothermal system on the endeavof 293. segment, Juan de Fuca Ridge. Journal of Geophysical Research- Solid Earth 97(B13):19663-19682. Gibbs, A.G. 1997. Biochemistry at depth. In: Fish Physiology; Vol. 16, Deep-Sea Fishes. (Randall, D.J. and Farrell, A.P., Eds.) New Desbruyères, D., Chevaldonnè, P., Alayse, A.-M., Jollivet, D., York: Academic Press, pp. 239-277. Lallier, F.H., Jouin-Toulmond, C , Zal, F., Sarradin, P.-M., Cosson, R., Caprais, J.-C., Arndt, C., O'Brien, J., Guezennec, J., Hourdez, Goffredi, S.K., Childress, J.J., Lallier, F.H. and Desaulniers, S., Riso, R., Gaill, F., Laubier, L. and Toulmond, A. 1998. Biology N.T. 1999. The ionic composition of the hydrothermal vent tube 2- and ecology of the "Pompeii worm" (Alvinella pompejana worm Riftia pachyptila: Evidence for the elimination of SO4 Desbruyères and Laubier), a normal dweller of an extreme deep-sea and H- and for Cl-/HCO3- shift. Physiological and Biochemical environment: A synthesis of current knowledge and recent Zoology 72(3):296-306. developments. Deep-Sea Research II 45:383-422. Grieshaber, M.K. and Volkel. S. 1998. Animal adaptations for DeZwaan, A. and Putzer, V. 1985. Metabolic adaptations of tolerance and exploitation of poisonous sulfide. Annual Review of intertidal invertebrates to environmental hypoxia (a comparison of Physiology 60:33-53. environmental anoxia to exercise anoxia). In: Physiological Adaptations of Marine Animals, vol. 39, Symposia of the Society Hagerman, L., Sandberg, E. and Vismann, B. 1997. Oxygen-binding of Experimental Biology (Laverck, M.S., Ed.). Cambridge UK: properties of haemolymph from the benthic amphipod Company of Experimental Biologists Ltd., University of Monoporeia affinis from the Baltic. Marine Biology 130:209-212. Cambridge, pp. 33-62. Hochachka, P.W. and Somero. G.N. 1984. Biochemical Adaptation. Fenchel, T and Finlay, B.J. 1995. Ecology and Evolution in Anoxic Princeton NJ: Princeton Academic Press, pp. 355-495. Worlds. Oxford University Press; New York. Hourdez, S., Frederick, L.A., Schernecke, A. and Fisher, C.R. 2000. Fisher, C.R. 1996. Ecophysiology of primary production at deep- Functional respiratory anatomy of a deep-sea orbiniid polychaete sea vents and seeps. In: Deep-sea and Extreme Shallow-water from the Brine Pool NR-1 in the Gulf of Mexico. Invertebrate Habitats: Affinities and Adaptations. (Uiblein, F., Ott, J. and Biology. In press. Stachowtisch, M., Eds.). Biosystematics and Ecology Series 11:313-336. Jannasch, H.W. 1989 Chemosynthetically sustained ecosystems in the deep sea. In: Hydrothermal Processes at Sea Floor Spreading Fisher, C.R. 1998. Temperature and sulphide tolerance of Centers (Schlegel, H.G. and Bowien, B., Eds.). New York: Plenum, hydrothermal vent fauna. Cahiers de Biologie Marine 39 (3-4):283- pp. 677-709. 286. Johnson, K.S., Beehler, C.L., Sakamoto-Arnold, C.M. and Childress, J.J. 1986. In situ measurements of chemical distributions Fisher, C.R., Childress, J.J., Oremland, R.S. and Bidigare, R.R. in a deep-sea hydrothermal vent field. Deep-Sea Research 35:1711- 1987. The importance of methane and thiosulfate in the metabolism 1722. of the bacterial symbionts of two deep-sea mussels. Marine Biology 96:59-71. Johnson, K.S., Childress, J.J., Hessler, R.R., Sakamoto-Arnold, C.M and Beehler, C.L. 1988. Chemical and biological interactions

Gravitational and Space Biology Bulletin 13(2), June 2000 21 METAZOANS IN EXTREME ENVIRONMENTS in the Rose Garden hydrothermal vent field. Deep-Sea Research of oxygen consumption rate and cytochrome reduction in the gill of 35:1724-44. the estuarine mussel Geokensia demissa. Biological Bulletin 191:421-430. Jones, M.L. 1981. Riftia pachyptila, new genus, new species, the vestimentiferan tubeworm from the Galapagos Rift geothermal Luoma, S.M. and Carter, J.L. 1991. Effects of trace metals on vents. Proceedings of the Biological Society of Washington aquatic benthos. In: Metal Ecotoxicology: Concepts and 93:1295-1313. Applications. Chelsea FL: Lewis Publishers, pp. 261-300.

Jones, M.L. 1988. The Vestimentifera: their biology, systematic MacDonald, I.R. 1998. Habitat formation at Gulf of Mexico and evolutionary patterns. Oceanologica Acta 8:69-82. hydrocarbon seeps. Cahiers de Biologie Marine 39:337-340.

Jouin, C. and Gaill, F. 1990. Gills of hydrothermal vent annelids: MacDonald, I.R., Boland, G.S., Baker, J.S., Gbooks, J.M., structure, ultrastructure, and functional implications in two Kennicutt, M.C., II, and Bidigare, R.R. 1989. Gulf of Mexico alvinellid species. Progress in Oceanography 24:59-69. chemosynthetic communities II: spatial distribution of seep organisms and hydrocarbons at Bush Hill. Marine Biology 101:235- Jouin-Toulmond, C., Augustin, D., Desbruyères, D. and 247. Toulmond, A. 1996. The gas transfer system in alvinellids. Cahiers de Biologie Marine 37:135-151. MacDonald, I.R., Reilly J.F., II, Guinasso, N.L., Jr., Brooks, J.M., Carney, R.S., Bryant, W.A. and Bright, T.J. 1990. Chemosynthetic Julian, D., Dalia, W.E. and Arp, A.J. 1998. Neuromuscular mussels at a brine-filled pockmark in the northern Gulf of Mexico. sensitivity to hydrogen sulfide in the marine invertebrate Urechis Science 248:1096-1099. caupo. Journal of Experimental Biology 201:1393-1403. McDonald, S.J. 1990. Benzo[a]pyrene metabolism in a deep-sea Julian, D., Gaill, F., Wood, E., Arp, A. and Fisher, C.R. 1999. mussel associated with natural petroleum seeps in the Gulf of Roots as a site of hydrogen sulfide uptake in the hydrocarbon Mexico [dissertation]. College Station TX: Texas A&M seep vestimentiferan Lamellibrachia sp. Journal of Experimental University. Biology 202:2245-2257. Mickel, T.J. and Childress, J.J. 1982. Effects of temperature, Juniper, S.K., Jonasson, I.R., Tunnicliffe, V. and Southward, A. pressure, and oxygen concentration on the oxygen consumption 1992. Influence of a tube-building polychaete on hydrothermal rate of the hydrothermal vent crab Bythograea therydron chimney mineralization. Geology 20:895-898. (Brachyura). Physiological Zoology 55:199-207.

Juniper, S.K. and Martineau, P. 1995. Alvinellids and sulfides at Millero, F.J. 1986. The thermodynamics and kinetics of the hydrothermal vents of the eastern Pacific: A Review. American hydrogen sulfide system in natural waters. Marine Chemistry Zoology 35:174-185. 18:121-147.

Juniper, S.K., Thompson, A.J. and Calvert, S.E. 1986. Nelson, D.C. and Fisher, C.R. 1995. Chemoautotrophic and Accumulation of minerals and trace elements in biogenic mucus at methanotrophic endosymbiotic bacteria at deep-sea vents and hydrothermal vents. Deep-Sea Research 33:339-347. seeps. In: The Microbiology of Deep-Sea Hydrothermal Vents. Ed. D.M. Karl. New York: CRC Press, pp.25-167. Kennicutt, M.C., II, Brooks, J.M. and Burke, R.A., Jr. 1989. Hydrocarbon seepage, gas hydrates, and authigenic carbonate in the Nix, E.R., Fisher, C.R., Vodenichar, J. and Scott, K.M. northwestern Gulf of Mexico. Offshore Technology Conference Physiological ecology of a mussel with methanotrophic 5952:649-654. endosymbionts at three hydrocarbon seep sites in the Gulf of Mexico. Marine Biology 122:605-617. Kennicutt, M.C., II, Burke, R.A., Jr. 1995. Stable isotopes: clues to biological cycling of elements at hydrothermal vents. In: Panis, L.I., Goddeeris, B. and Verheyen, R. 1996. On the Microbiology of Deep-Sea Hydrothermal Vents (Karl, D.M., Ed.) relationship between vertical microdistribution and adaptations to Boca Raton FL: CRC Press, pp. 275-287. oxygen stress in littoral Chironomidae. Hydrobiologia 318:61-67.

Kennicutt, M.C., II, McDonald, R.J., Comet, P.A., Denoux, G.J. Powell, M.A. and Somero, G.N. 1983. Blood components prevent and Brooks J.M. 1992. The origins of petroleum in the northern sulfide poisoning of respiration of the hydrothermal vent tube Gulf of Mexico. Geochimica et Cosmochimica Acta 56:1259- worm Riftia pachyptila. Science 219:297-299. 1280. Powell, M.A. and Somero, G.N. 1986a. Hydrogen sulfide oxidation Lee, W.L., Kraus, D.W. and Doeller, J.E. 1996. Sulfide stimulation is coupled to oxidative phosphorylation in mitochondria of

22 Gravitational and Space Biology Bulletin 13(2), June 2000

METAZOANS IN EXTREME ENVIRONMENTS

Solemya reidi. Science 233:563-566. thermydron and other decapod crustaceans. Physiological Zoology 60:121-137. Powell, M.A. and Somero, G.N. 1986b. Adaptations to sulfide by hydrothermal vent animals: sites and mechanisms of Vismann, B. 1991. Sulfide tolerance: physiological mechanisms and detoxification and metabolism. Biological Bulletin 171:274-290. ecological implications. Ophelia 34:1-27.

Sanders, N.K. and Childress, J.J. 1992. Specific effects of Volkel, S. and Grieshaber, M.K. 1996. Mitochondrial sulfide thiosulfate and L-lactate on hemocyanin-O2 affinity in a oxidation in Arenicola marina. Evidence for alternative electron brachyuran hydrothermal vent crab . Marine Biology 113 (2):175- pathways. European Journal of Biochemistry 235:231-237. 180. Weber, R.E. 1978. Respiratory Pigments. In Physiology of Annelids Sibuet, M. 1992. Deep-sea benthic ecosystems: dynamic aspects. (Mill, P.J., Ed.). London: Academic Press, pp. 393-446. In: Use and Misuse of the Seafloor (Hsu, K.J. and Thiede, J., Eds). John Wiley & Sons, pp. 319-334. Willmer, P. and Stone, G. 1997. Temperature and water relations in desert bees. Journal of Thermal Biology 22(6):453-465. Sibuet, M. and Olu, K. 1998. Biogeography, biodiversity and fluid dependence of deep-sea cold-seep communities at active Zal, F., Green, B.N., Lallier, F.H. and Toulmond, A. 1997. and passive margins. Deep Sea Research II 45(1-3):517-540. Investigation by electrospray ionization mass spectroscopy of the extracellular hemoglobin from the polychaete annelid Alvinella Somero, G.N. 1998. Adaptation to cold and depth: contrasts pompejana: an unusual hexagonal bilayer hemoglobin. Biochemistry between polar and deep-sea animals. In Cold Ocean Physiology 36:11777-11786. (Portner, H.O. and Playle, R.C., Eds.). Cambridge, UK: Cambridge University Press, pp. 33-57. Zal, F., Leize, E., Lallier, F.H., Toulmond, A., Van Dorsselaer, A. and Childress, J.J. 1998. S-sulfohemoglobin and disulfide exchange: Somero, G.N., Childress, J.J. and Anderson, A.E. 1989. Transport, the mechanisms of sulfide binding by Riftia pachyptila hemoglobins. metabolism, and detoxification of hydrogen sulfide in animals from Proceedings of the National Academy of Sciences 95:8997-9002. sulfide-rich marine environments. Critical Review in Aquatic Science 1:591-614.

Suchanek, T.H. 1993. Oil impacts on marine invertebrate populations and communities. American Zoology 33:510-523.

Terwilliger, N.B. and Terwilliger, R.C. 1984. Hemoglobin from the “Pompeii worm,” Alvinella pompejana, an annelid from a deep sea hot hydrothermal vent environment. Marine Biology Letters 5:191- 201.

Terwilliger, R.C., Terwilliger, N.B. and Schabtach, E. 1980. The structure of hemoglobin from an unusual deep-sea worm (vestimentifera). Comparative Biochemistry and Physiology 65B:531-535.

Tielens, A.G.M. and Van Hellemond, J.J. 1998. The electron transport chain in anaerobically functioning eukaryotes. Biochimica et Biophysica Acta-Bioenergetics 1365 (1-2):71-78.

Toulmond, A., Slitine, F., Frescheville, J. and Jouin, C. 1990. Extracellular hemoglobins of hydrothermal vent annelids: structural and functional characteristics in three alvinellid species. Biological Bulletin 179:366-373.

Tunnicliffe, V., McArthur, A.G. and McHugh, D. 1998. A biogeographical perspective of the deep-sea hydrothermal vent fauna. Advances in Marine Biology 34:353-442.

Vetter, R.D., Wells, M.E., Kurtsman, A.L. and Somero, G.N. 1987. Sulfide detoxification by the hydrothermal vent crab Bythograea

Gravitational and Space Biology Bulletin 13(2), June 2000 23

24 Gravitational and Space Biology Bulletin 13(2), June 2000 13

CHAPTER 3

Phylogeny and Biogeography of Deep Sea Vestimentiferan Tubeworms and Their Bacterial Symbionts

Published in: Symbiosis Volume 34, March 2003 Pages 1-41

Symbiosis, 34 (2003) 1–41 1 Balaban, Philadelphia/Rehovot

Review article Phylogeny and Biogeography of Deep Sea Vestimentiferan Tubeworms and Their Bacterial Symbionts

ERIN R. MCMULLIN*, STÉPHANE HOURDEZ, STEPHEN W. SCHAEFFER, and CHARLES R. FISHER The Pennsylvania State University, Department of Biology, 208 Mueller Laboratory, University Park, PA 16802, USA, Tel. +1-814-863-8360, Fax. +1-814-865-9131, Email. [email protected]

Received June 27, 2002; Accepted October 24, 2002

Abstract The present study combines previously published morphological descriptions and molecular-based characterizations of vestimentiferans and their symbionts with new molecular data to summarize and extend the understanding of vestimentiferan host and symbiont phylogeny and biogeography. Host cytochrome oxidase I (COI) and symbiont 16S ribosomal gene (16S) DNA sequences were used to explore evolutionary relationships among the vestimentiferans and their symbionts. Lamellibrachids of the northern Gulf of Mexico (GOM) are identified as a single species, Lamellibrachia cf luymesi, and new data and analyses are presented for Lamellibrachia barhami, Paraescarpia echinospica, and Arcovestia ivanovi. In general, both vestimentiferan hosts and symbionts have very large species ranges that are interrupted by depth. No evidence for cospeciation was found between vestimentiferans and their symbionts, supporting an environmental acquisition of the symbionts. Symbiont acquisition depends on host type (vent or seep), depth of site, and possibly host species. A test of evolutionary rate showed that vent

*The author to whom correspondence should be sent.

0334-5114/2003/$05.50 ©2003 Balaban 2 E.R. MCMULLIN ET AL.

vestimentiferans had a significantly faster COI sequence evolution than lamellibrachids, and symbionts from vent and seep vestimentiferans from deep water sites had a significantly slower rate of evolution than those from mid-depth and shallow sites.

Keywords: Vestimentiferan, symbiont, phylogenetics, biogeography, COI, 16S

1. Introduction

Vestimentiferan tube worms were first described by Webb in 1969 when a deep sea trawl recovered large worms in thick tubes from deep waters off the coast of California. These large animals bore a resemblance to a previously described group, the Pogonophora. Though first described in 1969, vestimentiferans were found only twice before 1977 (Webb, 1969; van der Land and Norrevang, 1975) in trawl and dredge collections, their habitats remaining unknown. In the late 1970s extremely high concentrations of organisms, including the 'giant tube worm' Riftia pachyptila, were found clustered around the hydrothermal vents of the Galapagos Rift spreading center (Lonsdale, 1977; Corliss et al., 1979). Riftia pachyptila were similar to the vestimentiferans originally described by Webb (1969) and van der Land (1975), but were much larger than those previously seen. In the early and mid 1980s aggregations of vestimentiferan tube worms were found at a variety of 'cold seep' sites worldwide, including the Oregon subduction zone (Suess et al., 1985), on the Florida Escarpment (Paull et al., 1984), off the coast of San Diego (Jones, 1985) and at the hydrocarbon seeps in the northern Gulf of Mexico (Kennicutt et al., 1985). The energy source for most deep sea life is organic matter generated by photosynthesis in the upper water column, very little of which reaches the bottom of the ocean. The high biomass at hydrothermal vent and cold seep sites are a direct result of the unique chemistry of these environments. The reduced chemicals present in vent and seep fluid, specifically sulfide and methane, are an energy source for chemoautotrophic bacteria, providing a rich local food source for deep sea organisms able to live in these extreme environments (Fisher, 1996). Though chemically similar, vents and seep environments differ in the geological setting that leads to the production of reduced chemicals. At hydrothermal vents seawater interacts with hot, newly formed rock several kilometers deep in the crust. As a result of interactions with hot basalts, the superheated water exiting the seafloor is devoid of oxygen and carries high concentrations of reduced metals and chemicals like hydrogen sulfide. Because of their tectonic nature, vent habitats have extreme VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 3 temporal and spatial fluctuations of temperature (2–60°C) and chemistry, and are ephemeral, normally lasting only years or decades (Tunnicliffe et al., 1998). The reduced chemicals present at cold seeps are a result of thermogenic or biogenic degradation of organic matter. Seeps can be associated with hydrocarbon reservoirs, subduction zone accretionary prisms, landslides, slumps, and other geologic features that create compressed sediments (MacDonald et al., 1990). Unlike vent environments, seep habitats are generally near ambient temperature, are highly sedimented, and may be stable for centuries to millennia (Fisher et al., 1997). Though the relationships of species and genera within the vestimentiferans are fairly well understood, the higher order classification of the vestimentiferans has been a continued source of debate. Alternately considered a phylum, class, or order of their own (Jones, 1985; Mañé-Garzón and Montero, 1985; Webb, 1969; van der Land and Nørrevang, 1977), new molecular data (Williams et al., 1993; Kojima et al., 1993; Winnepenninckx et al., 1995; McHugh, 1997; Halanych et al., 2001) and a reevaluation of morphological data (Rouse, 2001) support the placement of vestimentiferans in the polychaete family Siboglinidae. Siboglinidae is therefore comprised of the vestimentiferans, the pogonopharans and Sclerolinum brattstromi, a species previously thought to be a basal pogonophoran but recently found to be more similar to the vestimentiferans (Rouse, 2001; Halanych et al., 2001). This classification of vestimentiferans places them as a clade within the Family Siboglinidae, within the Class Polychaeta, under Phylum Annelida (Black et al., 1997; Kojima et al., 1997; McHugh, 1997, 2000; Southward, 1999; Bartolomaeus, 1999; Halanych et al., 1998, 2001; Rouse and Fauchald, 1997; Rouse, 2001). The vestimentiferans described to date fall into seven higher order taxa (previously families) and ten genera. The seven higher order taxa will be referred to here by a contraction of their now outdated family names (e.g. Lamellibrachiidae becomes lamellibrachid). Four species from three different taxa, the ridgeids, tevnids, and riftids, are known from hydrothermal vents on the east Pacific and northeast Pacific spreading centers (Jones, 1981, 1985), and two additional monospecific taxa, the arcovestids and alaysids, are known from hydrothermal vent sites in the Lau back arc basin (Southward, 1991; Southward and Galkin, 1997). Alaysid-like vestimentiferans have also been reported from diffuse flow sites near Japan (Kojima, 2002; Kojima et al., 2002). The two groupings of cold seep vestimentiferans, the escarpids and lamellibrachids (Jones, 1985), are generally more widely distributed than the hydrothermal vent taxa, and live in sedimented habitats of near-ambient temperatures with more diffuse sulfide seepage. Escarpids, containing four described species, have been found off the coast of Japan, off the North American Pacific coast, in Guaymas Basin, in the Gulf of Mexico, and off the 4 E.R. MCMULLIN ET AL. coast of Barbados. Members of the lamellibrachids, with five described species, are more widespread than the escarpids, and have been collected from both sides of the Pacific Ocean (near Japan, in Lau Basin, off the North American Pacific coast) and of the Atlantic (in the Gulf of Mexico, in the Mediterranean, and off the coasts of Barbados, Uruguay, Portugal, and West Africa). Seep vestimentiferan communities have been reported at depths ranging from 82 to 3,300 m (Hashimoto et al., 1993; Paull et al., 1984). Vestimentiferan tube worms, like their pogonophoran relatives, lack a digestive tract, and rely directly on primary production by sulfide oxidizing eubacterial symbionts (reviewed in Nelson and Fisher, 1995). A large amount of the vestimentiferan body is occupied by the trophosome, a specialized tissue containing large numbers (1011 bacteria/gram) of intracellular sulfide-oxidizing bacteria (gamma proteobacteria) (Powell and Somero, 1986). Researchers report the presence of a single species of bacterium within an individual (Distel et al., 1988), although one report suggests the presence of a second bacterial type, an epsilon proteobacterium, in the trophosome of a cold seep vestimentiferan (Naganuma et al., 1997a, 1997b). Highly integrated symbioses such as these involve a well-coordinated interweaving of responses between a host and symbiont pair that evolves over many generations. Vertical transmission of symbionts in the egg, such as that seen in some vent and seep bivalves (Peek et al., 1997), provides offspring with the specific strain of bacteria that was in the parent, guaranteeing the presence of a successful symbiont in the next generation. On the other hand, hosts like the lucinid bivalve Codakia orbicularis (Gros et al., 1996) acquire symbionts from the environment with each generation, a process known as environmental or horizontal symbiont transmission. The highly integrated and obligate nature of the symbiosis between vestimentiferans and their endosymbionts suggests that evolution would favor a mechanism by which offspring are guaranteed the presence of the symbiont so critical for vestimentiferan life. Despite the apparent benefit of a direct transmission of vestimentiferan endosymbionts, evidence does not support this mode of symbiont transmission between generations. No bacteria have been found in either vestimentiferan sperm or eggs (Cavanaugh et al., 1981), and molecular detection by bacterial-specific PCR and in situ hybridizations with gonadal tissue and freshly released sperm and eggs have both failed (Cary et al., 1993). Additionally, a functional form of the flagellin gene, which is involved with motility in free living bacteria, has been isolated from the endosymbiont of Riftia pachyptila. The presence of this gene is interpreted as evidence for a motile stage within the endosymbiont's lifecycle (Millikan et al., 1999). Vertical transmission of symbionts links host and symbiont DNA between generations, resulting in phylogenies with a degree of congruent evolution, where host speciation events are reflected in the phylogenetic gene VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 5 tree of the symbionts (Cary, 1993, 1994; Distel et al., 1994; Durand et al., 1996; Funk et al., 2000). No such congruence has been seen in the molecular phylogenetic trees of symbionts and their vestimentiferan hosts (Feldman et al., 1997; Di Meo et al., 2000; Nelson and Fisher, 2000). Indeed, three different vent vestimentiferan species share one symbiont type (Laue and Nelson, 1997; Nelson and Fisher, 2000) and multiple species of seep vestimentiferan share a second (Feldman et al., 1997; Nelson and Fisher, 2000), while different symbionts are seen in the same host species collected from different geographic regions (Nelson and Fisher, 2000). The present study combines previously published morphological descriptions and molecular-based characterizations of vestimentiferans and their symbionts with new molecular data to summarize and extend the understanding of vestimentiferan host and symbiont phylogeny and biogeography.

2. Materials and Methods

The phylogenies of vestimentiferan hosts presented here were generated based on the mitochondrial cytochrome oxidase I gene (COI), whose gene product is involved in aerobic respiration, a critical metabolic pathway. Twelve new vestimentiferan COI sequences were generated in this study, and were aligned with those vestimentiferan COI sequences previously available in GenBank, using the pogonophoran Galathealinum brachiosum (GenBank #AF178679) COI sequence as an outgroup. The GenBank accession numbers and collection locations for the sequences used in this study are listed in Table 1. The observed variation in vestimentiferan COI sequences generated no amino acid changes in the vestimentiferan COI protein sequence. These silent DNA substitutions should be selectively neutral; changes at these sites are expected to accumulate between two reproductively isolated species at a rate dependent on the mutation rate, the generation time, and the population size of each species (Kimura, 1983). Symbiont phylogeny was investigated using the sequence of the small subunit ribosomal RNA gene (16S), a gene whose product is an RNA molecule involved in the machinery of protein production. Though the shape of the 16S rRNA gene product is important to ribosome function, portions of the 16S sequence are not as constrained as the COI gene. The eight new seep symbiont 16S sequences generated for the current study were aligned with data presented in Feldman (1997) and Nelson and Fisher (2000), using the 16S sequence of the sulfide oxidizing endosymbiont of the bivalve Thyasira flexuosa (Genbank #L01575) as an outgroup. GenBank accession numbers for all symbiont sequences used in this study are shown in Table 1. Several sequences available in GenBank (e.g. U77479 in Feldman (1997)) and four sequences from DiMeo et al. (2001) were

VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 11 excluded from the analysis because the host species was unclear, or the sequences differed markedly from all others analyzed. The DiMeo symbiont sequences may represent entirely new groups of vent and seep symbionts, but the absence of additional sequences to support the observations make these data worrisome. To obtain the new COI and 16S rRNA sequences, genomic DNA was isolated from paired vestimentum (symbiont free) and trophosome (symbiont containing) tissue samples. The paired samples were frozen immediately after collection in liquid nitrogen, and were stored at –80°C. A small piece of either vestimentum or trophosome was digested overnight in a proteinase K digestion buffer, and DNA extracted by the standard phenol/chloroform technique (Ausubel et al., 1989). The density of bacterial cells in the trophosome ensured the isolation of predominantly symbiont genomic DNA, while the symbiont-free vestimentum generated pure host genomic DNA. Amplification of the COI and 16S was done directly from the isolated genomic DNA. A 1250 base pair fragment of the COI gene was amplified from vestimentum genomic DNA using primers based on COI regions conserved in invertebrates (Nelson and Fisher, 2000). 1550 base pairs of the 16S rRNA gene were amplified from the trophosome genomic DNA using universal bacterial primers published by Lane (1991). A single PCR band was generated with each primer set, COI or 16S. The two outer COI primers and four additional inner COI primers were used to sequence the COI fragment for a total of three overlapping sequence fragments on each strand. The COI sequencing primers used are as follows: fw2: GC(CT)GG(AG)ACAGGATGAAC(AT)GT, fw3: TTCTTTGA(TC)CC(TC)GCAGGAGG, rv2: GC(GA)AAAT(GA)GCTA(GA)ATCAATGCATGG, rv3: AC(AT)GTTCATCCTGT(CT)CC(AG)GC. The 1550 base pair16S PCR product was sequenced with the two original 16S PCR primers as well as six additional internal 16S primers based on conserved eubacterial regions, generating four overlapping fragments on each strand (Lane, 1991; Weisburg et al., 1991). Nucleotide sequences generated from these primers were unambiguous, indicating the presence of a single PCR product, COI or 16S, amplified from the genomic DNA. The COI and 16S sequences obtained by this method were similar to those previously published for vestimentiferans and their symbionts. Sequences were replicable, with the same animal consistently generated the same COI and 16S sequence. Sequences for the two genes were generated using chain terminating fluorescently labeled nucleotides (Beckman DTCS Quick Start Kit), and cycle sequenced according to the manufacturer's protocol. Products of this cycle sequencing reaction were run on a Beckman CEQ2000 automated sequencer (Beckman, Palo Alto, CA). The sequence fragments for each gene were assembled in Seqman (Lasergene, Madison, WI) and were edited by eye, 12 E.R. MCMULLIN ET AL. generating a 1200 bp contig of the COI gene and a 1550 bp contig of the 16S rRNA gene. Newly generated COI or 16S sequences were aligned (Lasergene, Madison, WI) with those available from GenBank (Table 1) using Clustal (Thompson et al., 1994) and were then adjusted by eye. The COI sequences used in this analysis overlapped for 600 to 1090 base pairs, while the 16S sequences overlapped for 1300 base pairs. Phylogenetic analyses were performed with MEGA 2.1 (Kumar et al., 2001). Pairwise genetic distances among the COI or 16S sequences were calculated using the proportion of differences between samples, and the Kimura two parameter correction (Kimura, 1980) was applied to correct for multiple substitutions. Gene trees were generated using both the neighbor joining (NJ) (Saitou and Nei, 1987) and minimum evolution (ME) methods (Rzhetsky and Nei, 1993). A maximum likelihood (ML) COI tree was also generated using PHYLIP (version 3.6a2, Felsenstein, 2001). The significance of the branching order in the COI and 16S neighbor joining and minimum evolution trees was evaluated by the bootstrap analysis of 1000 computer- generated trees. The substitution rates among both host and symbiont lineages were compared using two methods. First, Tajima's relative rate test (Tajima, 1993) was used within the MEGA package to compare substitution rates in pairs of sequences versus an outgroup (Galathealinum brachiosum or the endosymbiont of Thyasira flexuosa). A second program, RRTree (Robinson et al., 1998; Robinson- Rechavi and Huchon, 2000) was used to compare substitution rates between lineages, rather than individual sequences. RRTree calculates the probability of the substitution rates of two lineages being identical by comparing the mean substitution rate of each lineage to that of an outgroup. For RRTree a reduced data set of vestimentiferan COI sequences was produced using a 589 base pair fragment from one individual of each vent and seep species, excepting Arcovestia ivanovi, which did not overlap in this region. A test of coevolution, TREEMAP 1.0 (Page, 1995), was performed on the 15 vestimentiferans and symbionts for which paired host and symbiont sequences were available. TREEMAP evaluates the congruence of the phylogenies for host and symbiont pairs. The number of predicted cospeciation events between the 15 pairs was inferred and compared to the distribution obtained from a set of 1,000 randomly permuted tree topologies.

3. Results and Discussion

Vestimentiferan hosts

Fig. 1 shows the neighbor joining tree of available COI sequences for vestimentiferan tube worms. Data presented here include published sequences

VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 15 from previous studies (Black et al., 1997; Miura et al., 1997; Feldman et al., 1998; Kojima et al., 1997, 2002), as well as new data (marked with an asterisk). Though the family names Escarpiidae and Lamellibrachiidae are no longer applicable under the new classification of vestimentiferans as a clade within the Family Siboglonidae (Halanych et al., 2001; Rouse, 2001), the former families maintain two distinct and highly supported (100% bootstrap values) groups in all (NJ, ME, and ML) COI phylogenetic trees. The ML tree differed from the NJ and ME trees only in its deep branches, placing the vent vestimentiferans basal to the lamellibrachids and escarpids. The five species of vent vestimentiferans compared here form a third group in the COI tree with lower bootstrap support (85%). Previously published data were combined with the molecular identifications presented here to generate a map of worldwide vestimentiferan species distribution (Fig. 2). Table 2 is a detailed table of the collection locations, depth, and habitat type of vestimentiferan collections. A test of evolutionary rate constancy, the RRTree relative rate test, found a significant difference (p=0.028) in substitution rates between lamellibrachids (0.217 substitutions/site) and the vent species (0.252 substitutions/site). RRTree substitution rates are the number of substitution per site between a particular lineage and an outgroup. Tajima's relative rate test revealed a significant difference in evolutionary rate only between Lamellibrachia satsuma and Riftia pachyptila (P=0.02). Because R. pachyptila and L. satsuma have the fastest and slowest substitution rates they may affect RRTree comparisons between their respective groups. An RRTree test comparing evolutionary rates between vent vestimentiferans and lamellibrachids with R. pachyptila and L. satsuma removed still showed a significant difference. No significant difference was found between the escarpid group (0.243 substitutions/site) and either the vent species or lamellibrachids. However, the new escarpid species from 300 m Nankai showed a significantly higher substitution rate than both the new escarpid from the Louisiana Slope and the lamellibrachid group at (p<0.05). From these data lamellibrachids appear to have the slowest evolutionary rate of the three vestimentiferan groups, and vent species the fastest (Fig. 3). The escarpid substitution rate is somewhat less than is seen in the vent species, but substitution rates vary significantly within the group.

The escarpids

Members of the escarpids appear basal to both the lamellibrachids and the vent species in Fig. 1, though this branching pattern is not supported by strong bootstrap values (70%). Five distinct branches are seen within the escarpids. COI sequence data are available for at least one individual of each of the four described escarpid species. Sequences of two described escarpids, Paraescarpia

VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 17

California (type specimen, Jones, 1985), in both seep and vent sites of the Guaymas Basin (Black et al., 1997), as well as at a whalefall off Santa Catalina Island, California (Feldman et al., 1998). The two species of escarpids look very similar but differ by some key morphological characters, including the obturacular process, the ventral ridge (present in E. spicata only), the anterior margin of the vestimentum, and the number of paired branchial lamellae (Jones, 1985). As previously reported, E. laminata and E. spicata have extremely similar COI sequences (Black et al., 1997; Feldman et al., 1997; Nelson and Fisher, 2000), and form a single cluster with 100% bootstrap values (Fig. 1). Three new sequences of E. laminata COI, two from Alaminos Canyon and one from the Florida Escarpment, were used here to further explore the diversity of this gene in the two species. Of the seven sequences used (each 1050 bp in length), the variation observed within each species (0.1% to 0.3%) was similar to that seen between the two species (0.2% to 0.4%). In comparison, the genetic distance between the next most similar COI sequences, Lamellibrachia cf luymesi and Lamellibrachia columna, is an order of magnitude greater (3.2% to 4.8%). The connection between the Gulf of Mexico and the Pacific Ocean was broken by the formation of the Panama Isthmus roughly 3 million years ago, with deep water connections disappearing as many as 10 million years ago (Knowlton et al., 1993). Both E. laminata and E. spicata are currently found in very deep water (2000 to 3300 m) and were likely deep water species at the time of the Isthmus formation. In general, the estimated accumulation of mitochondrial mutations for many species is between 1% and 2% every 1 million years (Avise, 1991), and shallow water urchin species divided by the closing of the Panama Isthmus show COI divergence of 1.6% to 3.4% per million years (Bermingham and Lessios, 1993; McCartney et al., 2000). If E. laminata and E. spicata COI genes are accumulating mutations at 1% per million years, an expected difference of 3% in COI sequences is a conservative estimate of expected divergence. The observed difference between E. laminata and E. spicata COI sequences, however, is no more than 0.4%. Recent work suggests that the evolutionary rate of the COI gene within vent annelids may in fact be nearer to 0.2% per million years (Chevaldonné et al., 2001), generating an expected difference of 0.6% to 2.0% for E. laminata and E. spicata COI sequences, a divergence somewhat greater than the observed value. A test of evolutionary rate variation among the different vestimentiferan species did not show an unusually slow evolutionary rate for these two species compared to other vestimentiferans. The low divergence between E. laminata and E. spicata appears to be the result of a generally slow evolutionary rate in vestimentiferan COI genes rather than of a particularly slow rate in these two species. 18 E.R. MCMULLIN ET AL.

Seepiophila jonesi Seepiophila jonesi, a recently described species (Gardiner et al., 2001), is found at the hydrocarbon seep sites at relatively shallow depths (550–650 m) on the Louisiana Slope (MacDonald et al., 1989; Brooks et al., 1990). Previous molecular data (Gardiner et al., 2001) supported the existence of a single escarpid species within four Louisiana Slope sites which spanned 100 km. New sequences from animals collected 350 km to the east and 100 km to the west of the original sites were used in this study. COI sequences from all but one of the escarpids from the northern Louisiana Slope cluster in a single group, supported by 100% bootstrap values, with low genetic variation (0.03%) among the sequences. These data support the existence of a single dominant escarpid species inhabiting seep sites in the northern Gulf of Mexico as much as 580 km apart and 540 m to 650 m in depth.

Paraescarpia echinospica COI sequence from an escarpid from a 1200 m deep seep site on the Nankai Trough and a second individual from 1400 m deep a vent site on the Okinawa Trough (Kojima et al., 1997) form a single highly supported cluster with the COI sequence from an individual identified as Paraescarpia echinospica from a 1650 m deep vent site off Lihir Island (Southward et al., 2002, sequence provided by Ken Halanych). The COI sequences among these three individuals show very little divergence (0.4%). This same morphotype of escarpid has also been found near Japan in the Ryuko Canyon (1100 m), a seep site on the Omaezaki Spur (1200 m), and in a seep area off Kikaijima Island (1400 m) (Kojima, 2002). This species has also been reported from a 1500 m site in the Java Trench (Southward et al., 2002), as well as from a second site within the Manus Basin and from near Papua New Guinea (Kojima et al., 2002). Given the high degree of similarity in sequence between the samples reported here and in Kojima et al. (2000), and assuming based on morphological analysis that the individual from the Java Trench is also P. echinospica, this appears to be a single species of escarpid from intermediate depths (1200–1650 m) that spans at least the 8000 km from the Nankai Trough to the Java Trench. Excluding the Java Trench individual, for which no sequence data are available, the sequence data from the remaining samples support the existence of a single species that spans 4500 km.

Other escarpid species Kojima (Kojima et al., 1997, 2002) reported the COI sequence for a second morphotype of escarpid collected from a shallow site (300 m) in the Nankai Trough. The molecular phylogeny of these sequences shows a very distinct division between this second escarpid morphotype and the P. echinospica VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 19 individuals collected at deeper sites in the Nankai Trough (1200 m) and the Iheya Ridge (1400 m) (Kojima et al., 1997). Given the different morphology and the COI sequence divergence (8.5%) of this morphotype from P. echinospica, the individuals from the 300 m site in the Nankai Trough appear to be a new and undescribed species. This as yet undescribed shallow water escarpid has only been reported at one site, with a higher genetic diversity among those individuals than is seen in other vestimentiferans (Kojima et al., 2002). A second morphologically and genetically distinct escarpid has been collected, though rarely, from the Louisiana Slope sites in the Gulf of Mexico. Only one sample of this morphotype was preserved for genetic analysis, while three others were fixed for morphologic description. The COI sequence from this individual is significantly different from both Seepiophila jonesi (7.8%) of the Louisiana Slope (600 m) sites, and Escarpia laminata (6.8%), which is found at the Florida Escarpment (3300 m) and Alaminos Canyon (2200 m). The COI neighbor joining tree places this new GOM escarpid well outside of the S. jonesi cluster and with the E. laminata/E. spicata group at very low bootstrap values (<70%). The genetic and morphological differences between this morphotype and previously reported escarpids support it as a new species. This rare escarpid may be a representative of an intermediate depth escarpid species, which occurs only rarely at the shallower Louisiana Slope sites. Another new escarpid species was reported from collections near Barbados, but again no sequence data are available (Olu et al., 1997; Sibuet and Olu, 1998).

The lamellibrachids

Fig. 1 shows a possible evolutionary association between the lamellibrachids and the vent species, though this is supported with only intermediate bootstrap values (70%). A similar branching pattern was reported by Nelson (2000). Within the lamellibrachids five species have been named and described: Lamellibrachia luymesi (van der Land), L. victori (Mane- Garzon), L. columna (Southward), L. barhami (Webb), and L. satsuma (Miura), although no sequence data are available for either L. luymesi or L. victori. The lamellibrachid group in Fig. 1 contains six branches, three of which contain a sequence from one of the three described lamellibrachid species. A fourth highly supported (100% bootstrap value) cluster of sequences is comprised of sequences from individuals from the shallow water seep sites in the Gulf of Mexico, samples which bear a strong morphological resemblance to L. luymesi. Two additional sequences, the first from Manus Basin and the second from Kuroshima Knoll, did not cluster with any previously known species and were 20 E.R. MCMULLIN ET AL. different enough from each other to suggest they are separate species, though together they form a highly supported (97% bootstrap value) cluster within the lamellibrachids.

Lamellibrachia barhami Lamellibrachia barhami was first described by Webb (1969) (emended by Jones, 1985) from the coast of California near San Diego. It has since been reported from a variety of sites along the west coast of the North America, including the Monterey Canyon (Barry et al., 1996), the Oregon Slope (Suess et al., 1985), and from a low flow vent site in Middle Valley (Williams et al., 1993). The COI sequence has been published (Black et al., 1997) for L. barhami from the Oregon Slope (2100 m) and from Middle Valley (2400 m). Additional COI sequences were generated in this study from lamellibrachids sampled from 1300 m depth on the Vancouver Island Margin, from 1800 m and 2200 m depths off the Oregon Slope, and from 1000 m depth in Monterey Canyon. Samples of lamellibrachids from all sites on the North American Pacific coast fall within a single cluster, which includes individuals identified as Lamellibrachia barhami, supported by 100% bootstrap values. Only two nucleotide substitutions are seen in 1050 bp of COI sequence among the samples, and both substitutions are seen in individuals from the northern Vancouver Margin site as well as the more southern Monterey Canyon site. Additionally, a newly obtained sample from 1400 m depth off the coast of Costa Rica is identical to the L. barhami COI sequences presented here. Overall, samples within the highly supported Lamellibrachia barhami cluster (100% bootstrap values) show a genetic distance of 0.3% to 0.5%, similar to the genetic distance seen within Seepiophila jonesi, Lamellibrachia cf luymesi, and Lamellibrachia columna (0.1% to 0.5%). This single species of lamellibrachid is found in multiple seep and low activity vent sites spanning 6000 km, from the Vancouver Island Margin to the coast of Costa Rica, at depths ranging from 900 m (Barry et al., 1996) to 2400 m (Tunnicliffe, 1991).

Lamellibrachia luymesi Lamellibrachia luymesi was described in 1975 from a 500 m deep site near Guyana (van der Land and Norrevang, 1975). No type sample of L. luymesi is available for morphological comparison to other lamellibrachid species; however a lamellibrachid found on the Louisiana Slope of the Gulf of Mexico is morphologically very similar to the description of L. luymesi. The morphology of the single specimen of L. luymesi as described by van der Land and Norrevang (1975) falls within the morphological variation exhibited by the Louisiana Slope species. The Louisiana Slope species will therefore be referred to as Lamellibrachia cf luymesi. A redescription of L. luymesi, based on material VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 21 collected from sites on the Louisiana Slope, is in progress (Gardiner and Hourdez, submitted). This shallow water lamellibrachid co-occurs with Seepiophila jonesi, forming bush-like aggregations composed of both species (MacDonald et al., 1990). Lamellibrachia cf luymesi were sampled from sites spanning 480 km, and depths of 550 m to 650 m. COI sequences from all samples were very similar, with a genetic distance of 0.2% to 0.4% within the group, and are supported as a single phylogenetic group with 100% bootstrap values. Similar to L. barhami on the west coast of North America, L. luymesi appears to maintain a single species in a range exceeding 4000 km, between the Guyana shelf and the northern Gulf of Mexico. A second species of Lamellibrachia, L. victori, has been reported from the coast of Uruguay (Mañé-Garzón and Montero, 1985). The description of L. victori, trawled from approximately 300 m depth, is also similar to that of L. luymesi. The two species mainly differ in the number of sheath lamellae: 6 pairs for L. luymesi and 7 for L. victori (Mañé-Garzón and Montero, 1985) in samples of one and possibly two individuals, respectively. The number of sheath lamellae in L. cf luymesi of the Louisiana Slope typically varies from 4–8 pairs (Gardiner, submitted). Sheath lamellae appear to be a variable trait; Southward (1991) documented 8–16 pairs of sheath lamellae for L. columna. The three species are also found at similar depths, which may be a strong factor in defining species range. Without additional samples of L. victori available for morphological studies and genetics, however, the identity of this species with regard to those in the Gulf of Mexico and Caribbean cannot be determined. If this animal is indeed the same species as the Louisiana Slope lamellibrachid, then L. cf luymesi has a species range of 8000 km, comparable to that seen in Paraescarpia echinospica.

Lamellibrachia satsuma Lamellibrachia satsuma is one of two species of Lamellibrachia described from vent and seep sites around Japan. L. satsuma, described by Miura et al. (1997) from 82–110 m depth vent sites in Kagoshima Bay, is the shallowest species of vestimentiferan known (Hashimoto et al., 1993). Two different morphologies of vestimentiferan were found at a slightly deeper Nankai Trough site (Kinsu-no-se, 300 m). The COI sequence of the first of these morphotypes matched that of L. satsuma from Kagoshima bay, while the COI sequence of the second morphotype matched the sequence for L. columna (Black et al., 1997) from Lau Basin (Kojima et al., 1997). A third vestimentiferan, a lamellibrachid (Black et al., 1997) from a 433 m deep vent community on Nikko Seamount, also matches that from L. satsuma (Kojima et al., 2001). Kojima (2001) reports a very low genetic diversity among the COI sequences from L. satsuma of Kagoshima Bay, and argues that this low diversity may be the 22 E.R. MCMULLIN ET AL. result of a relatively recent colonization by a small founding population. The co-occurrence of L. columna and L. satsuma at the 300 m deep Kinsu-no-se site also suggests that perhaps this is a transition depth for the two species, with L. satsuma becoming dominant at shallower depths and L. columna dominant at deeper depths.

Lamellibrachia columna Lamellibrachia columna was first described by Southward (1991) from the Lau Basin, and the COI sequence of a Lau Basin L. columna was published by Black (1997). Kojima (2001) has shown that samples of lamellibrachids from eight different sites near Japan, both vents and seeps, are very similar to the published sequence of L. columna, a grouping which is supported in the current study with high (100%) bootstrap values. Though COI sequences from these eight sites are very similar (0.1% to 0.5% genetic distance), a marked division appears between individuals from shallow and deep collections. No overlap of COI haplotypes was seen between 45 shallow (300–1400 m) and 22 deep water (2000–3270 m) individuals sampled within or near the Nankai Trough. Of 14 animals sampled from 680–1400 m in the Okinawa Trough and Kuroshima Knoll, two had unique COI haplotypes and twelve had a haplotype seen in shallow Nankai Trough lamellibrachids; no deepwater haplotypes were seen within these collections (Kojima et al., 2001). The division of COI haplotype by depth indicates that L. columna from these sites may in fact be genetically isolated from each other. The pairwise genetic distance between the three distinct clusters within the L. columna group have a genetic distance of 0.5% to 1.1% (Kojima et al., 2001), larger than the distance between E. laminata and E. spicata, but less than that seen within the vent species Ridgeia piscesae. If these three clusters do indeed represent a single species, albeit subdivided, then Lamellibrachia columna is found over a distance of 8000 km and in a depth range of 300 to 3270 m. Kojima's data (2001), however, raise the possibility that these are three separate sister species with very low levels of COI divergence.

Possible new species of Lamellibrachia Kojima (2001) reports COI sequences for two additional vestimentiferans which, though they cluster within the lamellibrachids, are sufficiently different from all other known sequences to suggest that they are both from new species. The first new lamellibrachid sequence is from a sample collected from a 1650 m deep seep site in the Manus Basin, a region where Arcovestia ivanovi (Southward and Galkin, 1997), an escarpid-like, a ridgeid-like, and an alaysid-like species (Hashimoto et al., 1999; Southward et al., 2002) were VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 23 previously reported. A second unique sequence is from an individual sampled from a vent site (680 m) on the Kuroshima Knoll where L. columna is also found. While only one sequence is available from each of these new vestimentiferans (Kojima et al., 2001), the divergence between each sequence and other lamellibrachid sequences indicates the two individuals represent two new lamellibrachid species. Additional endemic species of Lamellibrachia may also exist in the Manus Basin and the seep area off the New Guinea Island (Kojima, 2002). A vestimentiferan collected from a shipwreck 1160 m deep off the coast of Portugal (Dando et al., 1992) was identified by tube morphology and 28S sequence as a Lamellibrachia sp. Tissue from the sample was too degraded for morphological identification, and no COI sequence was obtained, limiting comparisons to other known lamellibrachid species. The 28S data, however, did differentiate this sample from samples from the Gulf of Mexico, from Canadian samples (most likely L. barhami) (Williams et al., 1993), and from the Japanese L. satsuma (Brown et al., 1999). Because this sample is from a very different geographic region from the only remaining described species, L. columna of the southeast Pacific, it is likely to be another new species of lamellibrachid. Lamellibrachids have been collected from the Alaminos Canyon (2200 m) in the Gulf of Mexico (Brooks et al., 1990), but unfortunately no samples were available for molecular study. Given the division of species based on depth seen in sample sites near Japan, these deep water Gulf of Mexico lamellibrachids are likely a separate species from those found on the shallower Louisiana Slope (Brooks et al., 1990). Sibuet (1998) also reports a species of Lamellibrachia collected from below 1000 m off the coast of Barbados, again with no morphological species identification or samples for genetics. Additional reports of lamellibrachids collected from the Mediterranean (Olu-LeRoy et al., 2001b) and from the west coast of Africa (Olu-LeRoy et al., 2001a) are intriguing, but no species descriptions or genetic data have yet been published.

The vent vestimentiferans

Four vestimentiferan species, Riftia pachyptila, Tevnia jerichonana, Oasisia alvinae, and Ridgeia piscesae were described by Jones (1985) from venting sites in the east and northeast Pacific. Two additional vent species, Arcovestia ivanovi and Alaysia spiralis, were subsequently described from hydrothermal vent sites at the Lau back arc basin (Southward, 1991; Southward and Galkin, 1997). COI sequences from the four east Pacific vestimentiferans and newly published COI sequence from Arcovestia ivanovi 24 E.R. MCMULLIN ET AL.

(Kojima et al., 2002) form a single moderately supported vent clade (85%) (Fig. 1). This vent clade in turn pairs with the lamellibrachids, though at a low bootstrap value (70%). Deep branching patterns among the three vestimentiferan clades are not clearly resolved by COI sequences, as previous studies with this gene have placed vent vestimentiferans either outside of an escarpid/lamellibrachid seep clade (Nelson and Fisher, 2000) or within an escarpid/vent clade to the exclusion of the lamellibrachids (Black et al., 1997; Feldman et al., 1997). A similar phylogenetic tree using 28S sequence also failed to resolve the issue, grouping three vent genera, Tevnia, Ridgeia, and Riftia, in a cluster with Escarpia at low bootstrap values (63%) (Williams et al., 1993). Within the vent clade, Oasisia and Ridgeia form a group with low bootstrap support and Riftia, Tevnia, and Arcovestia form a second group. Oasisia/Ridgeia and Riftia/Tevnia pairings have been reported by a number of researchers (Black et al., 1997; Feldman et al., 1997; Nelson and Fisher, 2000). The new Arcovestia ivanovi COI sequence is most similar to that from Tevnia jerichonana, and the two sequences form a moderately supported (70%) cluster.

Riftia pachyptila Riftia pachyptila is a dominant organism at venting regions throughout the east Pacific and is found at Guaymas Basin and Rift, on the Galapagos Rise, and from 32°S to 21°N on the East Pacific Rise (Tunnicliffe et al., 1998). Like R. piscesae, R. pachyptila shows a degree of morphological placticity not reflected in molecular data (Black et al., 1994). Based on genetic analysis, Riftia pachyptila appears to maintain a single species over thousands of kilometers (Black et al., 1994; Tunnicliffe et al., 1998), though the Riftia at 32°S EPR appear be reproductively isolated from individuals from the northern sites (Hurtado, 2002). R. pachyptila often co-occurs on the EPR with two other vestimentiferan species, Tevnia jerichonana and Oasisia alvinae.

Tevnia jerichonana and Oasisia alvinae Tevnia jerichonana and Oasisia alvinae co-occur with the larger R. pachyptila at most east Pacific hydrothermal sites. The three species overlap on the East Pacific Rise from 32°S to 21°N, though T. jerichonana is absent from the northernmost EPR site, and only R. pachyptila is present on the Galapagos Rise (Segonzac et al., 1997; Tunnicliffe et al., 1998; Hurtado, 2002). R. pachyptila and T. jerichonana dominate different successional stages of the short-lived EPR communities (Lutz et al., 1994; Shank et al., 1998; Mullineaux et al., 2000), with T. jerichonana the first to colonize newly opened vent sites, only to be subsequently replaced by R. pachyptila. O. alvinae, found only sporadically and hard to distinguish from T. jerichonana (Lutz et al., 1994; Shank et al., 1998; Mullineaux et al., 2000), may occupy a different ecological VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 25 niche from the two other EPR species (pers. obs. C.R.F.). As seen in R. pachyptila, T. jerichonana and O. alvinae from 32°S EPR have limited gene flow with the more northern EPR populations. O. alvinae at this site show a particularly high genetic divergence from the northern Oasisia, and may in fact be a second Oasisia species (Hurtado et al., 2002).

Ridgeia piscesae Hydrothermal vent vestimentiferans found in the northeast Pacific were initially described as two different species, Ridgeia phaeophiale and Ridgeia piscesae (Jones, 1985). As many as six additional species were suspected based on differences in tube and body morphology among samples (Jones, 1985; Tunnicliffe, 1991). Ridgeia are found in areas of high and diffuse flow, with different morphotypes found in these different flow regimes (Tunnicliffe, 1991). For example, a long and thin Ridgeia morphotype is found on basaltic substrate in diffuse, low flow environments where temperatures are barely above ambient (2°C) and sulfide is often undetectable, while a short and thick morphotype is found on sulfide structures in higher flow environments (sulfide ~200 µM, temperature ~30°C) (Sarrazin et al., 1997; Urcuyo et al., 1998). Genetic studies, however, failed to find any evidence for reproductive isolation among the various Ridgeia morphotypes (Black, 1991; Southward et al., 1995; Black et al., 1998). Ridgeia phaeophiale and R. piscesae have subsequently been redescribed (Southward et al., 1995) as a single species, R. piscesae, with a number of highly variable morphological features. The distinct phenotypes of R. piscesae may be induced by local environmental conditions (Southward et al., 1995; Tunnicliffe et al., 1998). Ongoing studies are investigating physiological and gene expression differences that may explain how this highly polymorphic tubeworm species can inhabit such a broad range of microhabitats (Carney et al., 2002; Flores et al., 2002).

Arcovestia ivanovi and Alaysia spiralis Arcovestids and alaysids were originally reported from sedimented hydrothermal vent sites at the Lau back arc basin (Southward, 1991; Southward and Galkin, 1997). More recent collection sites of these two vent species often overlap collections of escarpids and lamellibrachids, classically considered the 'seep' vestimentiferans. Arcovestids have been collected from three vent sites within Manus Basin (Hashimoto et al., 1999), a vent site in the North Fiji Basin (Southward et al., 2002) and from a vent site near Papua New Guinea (Kojima et al., 2002). Genetic analysis of the COI sequence of samples from two sites within the Manus Basin and a third near Papua New Guinea show no genetic differentiation (Kojima et al., 2002) suggesting that samples from these three sites comprise a single species, Arcovestia ivanovi. These 26 E.R. MCMULLIN ET AL. newly released COI data (Kojima et al., 2002), when added to the aligned sequences used here, place Arcovestia ivanovi firmly within the 'vent' vestimentiferans, with Tevnia jerichonana as its closest relative. Though no sequence data have yet been published for Alaysia spiralis, reports and collections of this second western Pacific vent vestimentiferan have increased. A possible new Alaysia species was reported from Iheya Knoll as a dominant species. In 2000 a tentative new species of Alaysia was collected from a new vent field at 1350 m on the Daiyon Yonaguni Knoll (south Okinawa Trough) (Kojima, 2002). Kojima et al. (2002) report that molecular analysis is being done on an alaysid vestimentiferan sampled from a 2000 m deep seep site near Sagami Bay (the Yukie Ridge), and that preliminary data show that this alaysid species is closely related to Arcovestia ivanovi (Kojima et al., 2002). This close relationship would place Alaysia as well as Arcovestia within the existing vent vestimentiferan clade, supporting a common origin for all species of vent vestimentiferans.

Vestimentiferan symbionts Previous work has shown that vestimentiferan endosymbiont phylogenies do not parallel the phylogenies of their host species. For example, Escarpia laminata and a novel species of lamellibrachid from Alaminos Canyon have indistinguishable symbionts, while E. laminata from the Florida Escarpment harbors a different symbiont (Nelson and Fisher, 2000). These data, as well as the apparent lack of symbionts in vestimentiferan eggs and early larva (Cavanaugh et al., 1981; Cary et al., 1989b, 1993), support a horizontal or environmental transmission of symbionts from generation to generation. A number of factors affect symbiont acquisition in species with horizontal transmission, the most fundamental of which is that such species can only acquire a bacterial symbiont if it is present in the environment of the host. Because vestimentiferan larvae must interact directly with potential symbionts, motile bacteria found in the immediate vicinity of the larvae are the prime candidates for initiating a symbiosis. The physical presence of a potential symbiont is not enough to form a symbiosis; the host and symbiont must communicate in a specific way to initiate the process, and the host must be able to provide the appropriate conditions for bacterial growth while screening out unwanted bacteria (Smith and Douglas, 1987; Hirsch and McFall-Ngai, 2000). Varying degrees of specificity are seen in different host/symbiont associations. Some host/symbiont interactions are extremely specific: the host will only acquire a specific bacterial strain even if it is at low densities in the environment (e.g. the squid Euprymna scolopes (Hirsch and McFall-Ngai, 2000)). Other interactions are less specific: hosts will acquire any one of a variety of symbionts, depending on which strains of bacteria are available in VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 27 the environment (e.g. Ectomycorrhizae (Smith and Douglas, 1987)). New data presented here further explore the interplay of geography, host species, and the physical characteristics of a site in defining which strain or species of endosymbiont is found in a particular vestimentiferan host. The deepest divergence in the vestimentiferan endosymbiont 16S rRNA tree (Fig. 4) is between endosymbionts sampled from vent species and those from seep species. This division of vent and seep symbionts has been previously reported (Feldman et al., 1997; Nelson and Fisher, 2000; Di Meo et al., 2000), and is clearly illustrated by the two species L. barhami and R. piscesae, which occur in close proximity in the venting regions of Middle Valley but which contain very different symbionts (Nelson and Fisher, 2000). The 16S sequences of endosymbionts of the four east Pacific vent vestimentiferans form a single highly supported group, and have an average of 4.6% difference in 16S sequence from endosymbionts of seep vestimentiferans. Nelson and Fisher (2000) report a similar value, and argue that this level of difference between 16S sequences reflects an old divergence between these two symbiont groups (108 to 215 mya), perhaps indicating that seep and vent symbionts are different bacterial species. Identical 16S sequences were found in symbionts from three different host species at 9°N on the East Pacific Rise, and very slightly divergent symbionts were found in Ridgeia piscesae from the NEP spreading centers (Nelson and Fisher, 2000). This slight level of divergence is also supported by restriction fragment length polymorphism data (Laue and Nelson, 1997) and suggests that the symbionts of Ridgeia piscesae may be of a different population or strain from those found at 9°N on the EPR (Nelson and Fisher, 2000). The 16S sequences of endosymbionts from seep vestimentiferans form three distinct clusters, each supported with 100% bootstrap values. This result differs from the findings of Feldman (1997), who found that the symbionts from four sedimented seep host samples comprise a single bacterial species. The data are in agreement with Nelson (2000), who also found that the seep symbiont clade was composed of three separate clusters. The most basal cluster (Group 1) of symbionts is composed of samples from Lamellibrachia barhami from the Oregon Slope and from Middle Valley, E. laminata from the Florida Escarpment, and L. columna from Fiji-Lau (Fig. 5). The 16S sequences of symbionts within Group 1 show very low divergence (0.1% or 1.3 nucleotides), and differ by 2% (25.6 nucleotide differences) from the two other seep symbionts groups. The remaining two clusters of seep symbionts in turn form a highly (99%) supported subgroup. The first of these (Group 2) includes the symbionts of Lamellibrachia sp. and E. laminata collected from Alaminos Canyon, L. barhami from the Vancouver Island Margin and Monterey Canyon, and E. spicata from a whalefall near Santa Catalina California (Fig. 5). As reported

30 E.R. MCMULLIN ET AL. in Nelson and Fisher (2000) symbionts of two species, E. laminata and a Lamellibrachia sp. from Alaminos Canyon, have only one nucleotide replacement among them. These symbionts form a single branch with high bootstrap values, and are 0.2% (2.6 nucleotides) different from the other symbionts within Group 2, data which suggest that the symbionts of Alaminos Canyon are a separate population of bacteria. Group 3 symbionts are from the three species of the Louisiana Slope (GOM), Seepiophila jonesi, Lamellibrachia cf luymesi, and a new escarpid species (Fig. 5). Symbionts from Group 2 and Group 3 have a 1% (or 12.8 bp) difference in their 16S sequences, while within each group individuals are between 0.2% and 0.3% different. Previous authors (Feldman et al., 1997; Di Meo et al., 2000; Nelson and Fisher, 2000) have used the lack of congruence between host and symbiont phylogenies as evidence for environmental transmission of symbionts in the vestimentiferans. In the present study, a test for coevolution (TREEMAP) between the host COI tree and the symbiont 16S tree found five cospeciation events between the two phylogenetic trees. This result, however, is not significant when compared to a normal distribution of cospeciation events generated with 1000 random tree topologies. Vestimentiferans appear to acquire the locally available symbiont strain independent of host species, as illustrated by the collections from Alaminos Canyon, the Louisiana Slope, and EPR 9°N, which each have only one discernable symbiont but multiple host species (Fig. 5). Vestimentiferans do not, however, take up bacterial symbionts in an indiscriminate fashion. Only four strains or species of bacteria have been found in vestimentiferans (with the possible exception of the Guaymas Basin samples not included in this analysis), yet countless other bacteria are available in the water surrounding larval vestimentiferans. In particular, sulfide oxidizing gamma proteobacterial symbionts from local vent and seep bivalve species may be available to the vestimentiferan larvae, yet these particular symbionts are never found in a vestimentiferan host. Vestimentiferans are clearly discriminating among an array of available bacteria, acquiring only bacteria with a narrow range as symbionts. An even finer level of host/symbiont specificity is apparent in the animals collected from Middle Valley in the northeast Pacific. At this site Lamellibrachia barhami and Ridgeia pisceseae, collected just meters apart, harbor different species of symbionts. Because the hosts are in such close proximity, both symbiont types are presumably available to both vestimentiferan species, yet the vent species R. piscesae contains the classic 'vent' symbiont while the L. barhami symbiont falls into the seep symbiont cluster. Discrimination therefore occurs between very similar 'vent' and 'seep' symbionts species, and the appropriate association between vestimentiferans and their symbionts results even when numerous other bacterial species are available. VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 31

The symbiont that a vestimentiferan acquires from its environment is therefore a product of three filtering events. First, the bacteria must be available in the host's environment; second, the bacteria must be one of a limited set of bacteria which can be vestimentiferan symbionts; and third, the potential bacterial symbiont must be of the same environmental type, vent or seep, as its host. Questions of symbiont distribution can therefore be answered with a data set such as the one presented here. Within the two groups of vestimentiferans, vent and seep, symbiont acquisition appears to be defined by the availability of the free living symbiont. Mapping the locations and site characteristics of vent and/or seep host collections may reveal boundaries, geographic, physical, or chemical, that define symbiont distribution. Little information can be obtained from the vent symbionts because of their low 16S rRNA genetic variation. The three different strains of seep symbionts, however, are informative for such a comparison. Symbiont distribution does not appear to be strongly affected by geography (Fig. 5). The 16S sequences of symbionts of L. barhami from the Vancouver Island Margin and from Middle Valley, only 135 km apart, are 1.8% different from each other, whereas sequences of symbionts from three different host species from Fiji-Lau (L. columna), Middle Valley (L. barhami), and the Florida Escarpment (E. laminata), separated by as much as 11,000 km, are only 0.2% different. Symbionts from Group 1 and Group 2 are found in both the Atlantic and the Pacific, and representatives of all three groups are found within the Gulf of Mexico, separated at most by 1000 km. These data suggest that symbionts of Group 1 and Group 2 are available globally, and that some factor other than distance between hosts controls the presence of each strain. One possible environmental feature that could limit the range of a free living symbiont is depth. Group 1 symbionts were collected from generally deeper sample sites (1800 to 3300 m), while Group 2 is comprised of samples from intermediate depths (900m to 2200 m), and Group 3 from shallower sites (550 to 650 m) (Fig. 4). The collections depths of Group 1 and 2 show some overlap; these two strains may not have a clear depth boundary between them. Additionally, the deeper Vent and Group 1 symbiont collections, with mean substitution rates of 0.061 and 0.062 nucleotides/site respectively, show a significantly slower evolutionary rate (p<0.001) than the intermediate Group 2 (0.085 nucleotides/site) and shallow Group 3 (0.087 nucleotides/site) collections. Though the discussion above relies on the assumption that, on a local level, symbiont strain is defined first by collection site, new data presented here (Fig. 4) also suggest fine scale division of symbiont strain between vestimentiferan hosts within a collection site. The GOM symbiont sequences within Group 3 appear to be divided by host species. The three S. jonesi symbiont sequences form a single cluster (71% bootstrap), while the three sequences from L. cf 32 E.R. MCMULLIN ET AL. luymesi form a second cluster, with low bootstrap support (61%). The symbiont from the second GOM escarpid species falls basal to these two clusters in its own well supported branch (87% bootstrap). While Group 1 and Group 2 Alaminos Canyon sequences have very little 16S rRNA variation, and the remaining Group 2 symbionts are from a single host species, Group 3 symbionts may have enough genetic variation in their 16S sequences for fine scale divisions to be seen. Preliminary data from RFLP analysis of 16S variation within the Louisiana Slope tubeworms shows a statistically significant tendency for one symbiont strain to be found within S. jonesi and a second symbiont strain within L. cf luymesi. The above observations lead to a number of predictive hypotheses that may be tested with known vestimentiferan collection sites, as well as with new collections in the future. First, we predict that deep water symbionts, both vent and seep, will show a lower diversity and slower evolutionary rate than intermediate and shallow water symbionts. Second, all samples of vent vestimentiferans will harbor symbionts of the 'vent' symbiont group. This prediction includes the alaysids and arcovestids of Lau Basin and the waters near Japan; a large species range in vent symbionts would not be unprecedented as similar ranges are seen in both Group 1 and Group 2 symbionts. Third, seep vestimentiferan symbiont type will depend on the depth of collection: all seep vestimentiferans species collected from >2000 m will harbor Group 1 seep symbionts, those from intermediate depths (1000–2000 m) will harbor Group 2 symbionts, and shallow water vestimentiferans (<1000 m) will harbor Group 3 symbionts. The predictions for shallow water symbionts are based at present on a limited amount of data, and while Group 3 symbionts may be found worldwide in shallow water seep sites, they may also be found only locally in the northern Gulf of Mexico. Samples of L. luymesi from the shallow water site near Guyana would be particularly useful in answering this question. Finally, different host species from the same sample site may show a tendency toward having one or another symbiont strain, though not every individual of a given host species will have the preferred symbiont. A number of collection sites could be particularly useful in addressing the above questions. It would be interesting to see if symbionts from the two west Pacific vent vestimentiferans, Alaysia spiralis and Arcovestia ivanovi, are the same symbiont strain as the vent vestimentiferans of the eastern Pacific. Sites with co-occurring vent and seep species, such as in Guaymas Basin and in the waters near Japan, could be used to further test the specificity of these vestimentiferans for their respective symbiont types. Vestimentiferan communities near Japan include multiple vent and seep vestimentiferan species which span a wide range of depths and include seep, vent, and whalefall sites. These sites provide an exciting and extensive opportunity for studies concerning VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 33 symbiont strain and evolutionary rate as a function of host species, depth, or other physical characteristics of a site.

4. Conclusion

While symbiont evolutionary rates appear to decrease with depths, vent vestimentiferans have the fastest COI evolutionary rate, while the slowest rate is in the shallow L. satsuma. Evolutionary rate is the accumulation of mutations in a gene per generation, and is therefore influenced by such factors as the mechanics of DNA replication and repair, and the generation time of the species (Hartl and Clark, 1997). The generation time effect may explain why the relatively short-lived vent species have a faster COI sequence evolution than the very long-lived lamellibrachids (Bergquist et al., 2000), but does not explain why the escarpids have an evolutionary rate more similar to vent vestimentiferans when evidence suggests that escarpids live as long as, if not longer than, the lamellibrachids (Bergquist et al., 2002). In fact, the two seep vestimentiferan with the slowest evolutionary rate and lowest genetic diversity (L. satsuma), and the fastest evolutionary rate and highest genetic diversity (the escarpid), are from the same 300 m deep site of the Nankai Trough. Though the low diversity in L. satsuma of the Nankai Trough may be, as Kojima et al. (2002b) argue, the result of a recent colonization event by a small founding population and/or a stressful environment, the same low diversity is not seen in the escarpid. Most likely the differences in evolutionary rate and genetic diversity between these two species are the result of a combination of different physiologies, life history traits, and population histories. The slower COI evolution rate in lamellibrachids may explain why COI-based trees show L. columna from the Nankai Trough as a single species while detailed genetic studies reveal two genetically isolated communities. Lamellibrachid species with large geographical and/or depth ranges (L. barhami and L. luymesi) should therefore be subjected to more detailed genetic studies to reveal genetic isolation undetected by COI-based phylogenetics. We have combined information from the current literature on vestimentiferan biogeography and genetics with new genetic data to present a more complete description of vestimentiferan host and symbiont occurrence worldwide. While vent vestimentiferans are notably absent from the Atlantic and the recently explored Indian Ocean vents, seep vestimentiferans are found in a variety of sulfidic environments in the Atlantic, the Pacific, and the Mediterranean, over a wide range of depths (82 m to 3300 m). Often multiple species are found at the same site, with seep and vent species at times occurring meters apart, and multiple vent or seep species inhabiting the same site or even aggregation. In general, vestimentiferans (eg. L. barhami, L. columna and R. 34 E.R. MCMULLIN ET AL. pachyptila) have very large species ranges that are interrupted by changes in depth (eg. the four species of the Nankai Trough). Symbionts also have very large ranges, with nearly identical 16S sequence found in hosts separated by thousands of kilometers. Vent symbionts appear specific to vent vestimentiferan hosts, while three different symbiont strains inhabit seep vestimentiferans. Site depth appears to be a factor in defining which of these three strains is found in a particular seep host, and different hosts may show a preference for variants within a symbiont strain. Depth may also be directly influencing seep and vent symbiont diversity and evolutionary rate. A number of known sample sites, particularly in the western Pacific, offer an exciting opportunity for genetic studies of host and symbiont biogeography similar to those presented here, which would be greatly advanced by similar analyses of samples from the eastern Atlantic. It will be interesting to see if the trends in host and symbiont phylogenetics, biogeography and evolutionary rate shown in this study are supported in future samples.

Acknowledgments

This work was supported by the NOAA National Undersea Research Program at the University of North Carolina, Wilmington, Harbor Branch Oceanographic Institution, and the Minerals Management Service, Gulf of Mexico Regional OCS Office through the contract number 1435-10-96-CT30813. We thank the captain and crew of the RV Edwin Link and the Canadian Coast Guard Ship John P. Tulley, as well as the submersible crew and pilots of the Johnson Sea Link and of ROPOS. We thank Paul Yancey (NSF grant number IBN-9407205 to Joseph Siebenaller, Louisiana State University), Jim Barry and the Monterey Bay Aquarium Research Institute, Jason Flores, Susan Carney, and Kim Juniper (Natural Sciences and Engineering Research Council of Canada, Collaborative Research Opportunities grant), and Heiko Sahling (Sonderforschungsbereich 574 "Volatiles and fluids in subduction zones", Kiel University) for collecting and providing samples, and Stephen Gardiner for his input on vestimentiferan morphological phylogenetics.

REFERENCES

Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Smith, J.A., and Struhl, K. 1989. Current Protocols in Molecular Biology. John Wiley & Sons, New York. Avise, J.C. 1991. Ten unorthodox perspectives on evolution prompted by comparative population genetic findings on mitochondrial DNA. Annual Review of Genetics 25: 45–69. Barry, J.P., Greene, H.G., Orange, D.L., Baxter, C.H., Robison, B.H., Kochevar, R.E., Nybakken, J.W., Reed, D.L., and McHugh, C.M. 1996. Biologic and geologic VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 35

characteristics of cold seeps in Monterey bay, California. Deep-Sea Research Part I- Topical Studies in Oceanography 43: 1739. Bartolomaeus, T. 1999. Structure, function and development of segmental organs in Annelida. Hydrobiologia 402: 21–37. Bergquist, D.C., Williams, F.M., and Fisher, C.R. 2000. Longevity record for deep-sea invertebrate. Nature 403: 499–500. Bergquist, D.C., Urcuyo, I.A., and Fisher, C.R. 2002. Establishment and persistence of seep vestimentiferan aggregations from the upper Louisiana slope of the Gulf of Mexico. Marine Ecology Progress Series 41: 89–98. Bermingham, E. and Lessios, H.A. 1993. Rate variation of protein and mitochondrial-DNA evolution as revealed by sea-urchins separated by the Isthmus of Panama. Proceedings of the National Academy of Sciences of the USA 90: 2734–2738. Black, M.B. 1991. Genetic (allozyme) variation in Vestimentifera (Ridgeia spp.) from hydrothermal vents of the Juan de Fuca Ridge (Northeast Pacific Ocean). Thesis, University of Victoria. Black, M.B., Lutz, R.A., and Vrijenhoek, R.C. 1994. Gene flow among vestimentiferan tube worm (Riftia pachyptila) populations from hydrothermal vents of the Eastern Pacific. Marine Biology 120: 33–39. Black, M.B., Halanych, K.M., Maas, P.A.Y., Hoeh, W.R., Hashimoto, J., Desbruyeres, D., Lutz, R.A., and Vrijenhoek, R.C. 1997. Molecular systematics of vestimentiferan from hydrothermal vents and cold-water seeps. Marine Biology 130: 141–149. Black, M.B., Trivedi, A., Maas, P.A.Y., Lutz, R.A., and Vrijenhoek, R.C. 1998. Population genetics and biogeography of vestimentiferan tube worms. Deep-Sea Research Part I- Topical Studies in Oceanography 45: 365–382. Brooks, J.M., Wiesenburg, D.A., Roberts, H., Carney, R.S., MacDonald, I.R., Fisher, C.R., Guinasso, N.L.J., Sager, W.W., McDonald, S.J., Burke, R. A.J., Aharon, P., and Bright, T.J. 1990. Salt, seeps and symbiosis in the Gulf of Mexico. EOS 71: 1772–1773. Brown, S., Rouse, G., Hutchings, P.A., and Colgan, D.J. 1999. Assessing the usefulness of histone H3, U2, snRNA and 28S rDNA in analyses of polychaete relationships. Australian Journal of Zoology 47: 499–515. Carney, S.L., Peoples, J.R., Fisher, C.R., and Schaeffer, S.W. 2002. AFLP analyses of genomic DNA reveal no differentiation between two phenotypes of the vestimentiferan tubeworm, Ridgeia piscesae. Cahiers de Biologie Marine 43: 363–366. Cary, S.C., Felbeck, H., and Holland, N.D. 1989a. Observations on the reproductive biology of the hydrothermal vent tube worm, Riftia pachyptila. Marine Ecology Progress Series 52: 89–94. Cary, S.C., Vetter, R.D., and Felbeck, H. 1989b. Habitat characterization and nutritional strategies of the endosymbiont-bearing bivalve Lucinoma aequizonata. Marine Ecology Progress Series 55: 31–45. Cary, S.C., Warren, W., Anderson, E., and Giovannoni, S.J. 1993. Identification and localization of bacterial endosymbionts in hydrothermal vent taxa with symbiont- specific polymerase chain reaction amplification and in situ hybridization techniques. Molecular Marine Biology and Biotechnology 2: 51–62. 36 E.R. MCMULLIN ET AL.

Cary, S.C. and Giovannoni, S.J. 1993. Transovarial inheritance of endosymbiotic bacteria in clams inhabiting deep-sea hydrothermal vents and cold seeps. Proceedings of the National Academy of Sciences of the USA 90: 5695–5699. Cary, S.C. 1994. Vertical transmission of a chemoautotrophic symbiont in the protobranch bivalve, Solemya reidi. Molecular Marine Biology and Biotechnology 3: 121–130. Cavanaugh, C.M., Gardiner, S.L., Jones, M.L., Jannasch, H.W., and Waterbury, J.B. 1981. Prokaryotic cells in the hydrothermal vent tube worm Riftia pachyptila: Possible chemoautotrophic symbionts. Science 213: 340–342. Chevaldonné, P., Jollivet, D., Desbruyères, D., Lutz, R.A., and Vrijenhoek, R.C. 2001. Sibling species of the eastern Pacific hydrothermal vent worms (Ampharetidae, Alvinellidae, Vestimentifera) provide new mitochondrial COI clock calibration. In: The 2nd International Symposium on Deep-Sea Hydrothermal Vent Biology. Brest, France. Corliss, J.B., Dymond, J., Gordon, L.I., Edmond, J.M., Herzen, R.P.V., Ballard, R.D., Green, K., Williams, D., Bainbridge, A., Crane, K., and van Andel, T.H. 1979. Submarine thermal springs on the Galapagos Rift. Science 203: 1073–1083. Dando, P.R., Southward, A.J., Southward, E.C., Dixon, D.R., Crawford, A., and Crawford, M. 1992. Shipwrecked tube worms. Nature 356: 667. Di Meo, C.A., Wilbur, A.E., Holben, W.E., Feldman, R.A., Vrijenhoek, R.C., and Cary, S.C. 2000. Genetic variation among endosymbionts of widely distributed vestimentiferan tubeworms. Applied Environmental Microbiology 66: 651–658. Distel, D.L., Lane, D.J., Olsen, G.J., Giovannoni, S.J., Pace, B., Pace, N.R., Stahl, D.A., and Felbeck, H. 1988. Sulfur-oxidizing bacterial endosymbionts: analysis of phylogeny and specificity by 16S rRNA sequences. Journal of Bacteriology 170: 2506–10. Distel, D.L., Felbeck, H., and Cavanaugh, C.M. 1994. Evidence for phylogenetic congruence among sulfur-oxidizing chemoautotrophic bacterial endosymbionts and their bivalve hosts. Journal of Molecular Evolution 38: 533–542. Durand, P., Gros, O., Frenkiel, L., and Prieur, D. 1996. Phylogenetic characterization of sulfur-oxidizing bacterial endosymbionts in three tropical Lucinidae by 16S rDNA sequence analysis. Molecular Marine Biology and Biotechnology 5: 37–42. Feldman, R.A., Black, M.B., Cary, C.S., Lutz, R.A., and Vrijenhoek, R.C. 1997. Molecular phylogenetics of bacterial endosymbionts and their vestimentiferan hosts. Molecular Marine Biology and Biotechnology 6: 268–277. Feldman, R.A., Shank, T.M., Black, M.B., Baco, A.R., Smith, C.R. and Vrijenhoek, R.C. 1998. Vestimentiferan on a whale fall. Biological Bulletin 194: 116–119. Felsenstein, J. 2001. PHYLIP: University of Washington. http://evolution.genetics.washington.edu/phylip.html. Fisher, C.R., Urcuyo, I.A., Simpkins, M.A., and Nix, E. 1997. Life in the slow lane: Growth and longevity of cold-seep vestimentiferans. Marine Ecology 18: 83–94. Flores, J.F., Green, B.N., Freytag, J.K., Hourdez, S., and Fisher, C.R. 2001. Structural and functional plasticity of the extracellular hemoglobins from the polymorphic vestimentiferan tubeworm, Ridgeia piscesae. In: The 2nd International Symposium on Deep-Sea Hydrothermal Vent Biology. Brest, France. Funk, D.J., Helbling, L., Wernegreen, J.J., and Moran, N.A. 2000. Intraspecific phylogenetic congruence among multiple symbiont genomes. Proceedings of the Royal Society of London Series B-Biological Sciences 267: 2517–2521. VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 37

Gardiner, S.L., McMullin, E., and Fisher, C.R. 2001. Seepiophila jonesi, a new genus and species of vestimentiferan tube worm (Annelida: Pognophora) from hydrocarbon seep communities in the Gulf of Mexico. Proceedings of the Biological Society of Washington 114: 694–707. Gros, O., Darrasse, A., Durand, P., Frenkiel, L., and Moueza, M. 1996. Environmental transmission of a sulfur-oxidizing bacterial gill endosymbiont in the tropical lucinid bivalve Codakia orbicularis. Applied Environmental Microbiology 62: 2324–2330. Halanych, K.M., Lutz, R.A., and Vrijenhoek, R.C. 1998. Evolutionary origins and age of vestimentiferan tube-worms. Cahiers de Biologie Marine 39: 355–358. Halanych, K.M., Feldman, R.A., and Vrijenhoek, R.C. 2001. Molecular evidence that Sclerolinum brattstromi is closely related to vestimentiferans, not to frenulate pogonophorans (Siboglinidae, Annelida). Biological Bulletin 201: 65–75. Hartl, D.L. and Clark, A.G. 1997. Principles of Population Genetics, 3rd edition, Sinauer Associates, Inc., Sunderland, Massachusetts. Hashimoto, J., Miura, T., Fujikura, K., and Ossaka, J. 1993. Discovery of vestimentiferan tube-worms in the euphotic zone. Zoological Science 10: 1063–1067. Hashimoto, J., Ohta, S., Fiala-Medioni, A. et al. 1999. Hydrothermal vent communities in the Manus Basin, Papua New Guinea: results of the BIOACCESS Cruises '96 and '98. InterRidge News 8: 12–18. Hirsch, A.M. and McFall-Ngai, M.J. 2000. Fundamental concepts in symbiotic interactions: Light and dark, day and night, squid and legume. Journal of Plant Growth Regulation 19: 113–130. Hurtado, L.A. 2002. Thesis: Evolution and biogeography of hydrothermal vent organisms in the eastern Pacific Ocean. Ph.D. Thesis, Rutgers University. Hurtado, L.A., Mateos, M., Lutz, R.A., and Vrijenhoek, R.C. 2002. Molecular evidence for multiple species of Oasisia (Annelida: Siboglinidae) at the eastern Pacific hydrothermal vents. Cahiers de Biologie Marine 43: 377–380. Jones, M.L. 1981. Riftia pachyptila, new genus, new species, the vestimentiferan tubeworm from the Galapagos Rift geothermal vents. Proceedings of the Biological Society of Washington 93: 1295–1313. Jones, M.L. 1985. On the vestimentifera, new phylum: six new species, and other taxa, from hydrothermal vents and elsewere. Bulletin of the Biological Society of Washington 6: 117–158. Kennicutt, M.C. II, Brooks, J.M., Bidigare, R.R., Fay, R.R., Wade, T.L., and McDonald, T.J. 1985. Vent-type taxa in a hydrocarbon seep region on the Louisiana Slope. Nature 317: 351–353. Kimura, M. 1980. A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. Journal of Molecular Evolution 16: 111–120. Kimura, M. 1983. The Neutral Theory of Molecular Evolution. Cambridge University Press, New York. Knowlton, N., Weigt, L.A., Solorzano, L.A., Mills, D.E.K. and Bermingham, E. 1993. Divergence in proteins, mitochondrial DNA, and reproductive compatability across the Isthmus of Panama. Science 260: 1629–1632. 38 E.R. MCMULLIN ET AL.

Kojima, S., Hashimoto, T., Hasegawa, M., Murata, S., Ohta, S., Seki, H., and Okada, N. 1993. Close phylogenetic relationship between vestimentifera (tube worms) and Annelida revealed by the amino-acid-sequence of elongation Factor-1-Alpha. Journal of Molecular Evolution 37: 66–70. Kojima, S., Segawa, R., Hashimoto, J., and Ohta, S. 1997. Molecular phylogeny of vestimentiferans collected around Japan, revealed by the nucleotide sequences of mitochondrial DNA. Marine Biology 127: 507–513. Kojma, S., Ohta, S., Miura, T., Fujiwara, Y., and Hashimoto, J. 2000. Molecular phylogenetic study of chemosynthesis-based communities in the Manus Basin. JAMSTEC Deep Sea Research 16: 7–13. Kojima, S., Ohta, S., Yamamoto, T., Miura, T., Fujiwara, Y., and Hashimoto, J. 2001. Molecular taxonomy of vestimentiferans of the western Pacific and their phylogenetic relationship to species of the eastern Pacific. I. Family Lamellibrachiidae. Marine Biology 139: 211–219. Kojima, S. 2002a. Deep-sea chemoautosynthesis-based communities in the Northwestern Pacific. Journal of Oceanography 58: 343–363. Kojima, S., Ohta, S., Yamamoto, T., Miura, T., Fujiwara, Y., Fujikura, K., and Hashimoto, J. 2002b. Molecular taxonomy of vestimentiferans of the western Pacific and their phylogenetic relationship to species of the eastern Pacific: II. Families Escarpiidae and Arcovestiidae. Marine Biology 141: 57–64. Kumar, S., Tamura, K., Jakobsen, I.B., and Nei, M. 2001. MEGA2: Molecular Evolutionary Genetics Analysis software. Bioinformatics 17: 1244–1245. Land, J. van der and Nørrevang, A. 1975. The systematic position of Lamellibrachia (Annelida, Vestimentifera). In: The Phylogeny and Systematic Position of Pogonophora. A. Nørrevang, ed., pp. 86–101. Zeitschrift für Zoologische Systematik und Evolutionsforschung. Sonderheft. Land, J. van der and Nørrevang, A. 1977. Structure and relationships of Lamellibrachia (Annelida, Vestimentifera). Kongelige Danske Videnskaps Selskaps Skrifter 21: 1–102. Lane, D.J. 1991. 16S/23S rRNA Sequencing. In: Nucleic Acid Techniques in Bacterial Systematics. E. Stackebrandt and M. Goodefellow, eds. John Wiley and Sons, pp. 115–175. Laue, B.E. and Nelson, D.C. 1997. Sulfur-oxidizing symbionts have not co-evolved with their hydrothermal vent tube worm hosts: An RFLP analysis. Molecular Marine Biology and Biotechnology 6: 180–188. Lonsdale, P. 1977. Clustering of suspension-feeding macrobenthos near abyssal hydrothermal vents at aceanic spreading centers. Deep-Sea Research 24: 857–863. Lutz, R.A., Shank, T.M., Fornari, D.J., Haymon, R.M., Lilley, M.D., Von Damm, K.L., and Desbruyères, D. 1994. Rapid growth at deep-sea vents. Nature 371: 663–664. MacDonald, I.R., Boland, G.S., Baker, J.S., Brooks, J.M., Kennicutt, M.C., II, and Bidigare, R.R. 1989. Gulf of Mexico hydrocarbon seep communities. II. Spatial distribution of seep organisms and hydrocarbons at Bush Hill. Marine Biology 101: 235–247. MacDonald, I.R., Guinasso, N.L., Reilly, J.F., Brooks, J.M., Callender, W.R., and Gabrielle, S.G. 1990. Gulf of Mexico hydrocarbon seep communities: VI. patterns in community structure and habitat. Geo-Marine Letters 10: 244–252. VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 39

Mañé-Garzón, F. and Montero, R. 1985. Sobre una nueva forma de verme tubicola – Lamellibrachia victori n. sp. (Vestimentifera) – Proposicion de un nuevo phylum: Mesoneurophora. Revista de Biologia del Uruguay 8: 1–28. McCartney, M.A., Keller, G., and Lessios, H.A. 2000. Dispersal barriers in tropical oceans and speciation in Atlantic and eastern Pacific sea urchins of the genus Echinometra. Molecular Ecology 9: 1391–1400. McHugh, D. 1997. Molecular evidence that echiurans and pogonophorans are derived annelids. Proceedings of the National Academy of Sciences of the USA 94: 8006–8009. McHugh, D. 2000. Molecular phylogeny of the Annelida. Canadian Journal of Zoology-Revue Canadienne De Zoologie 78: 1873–1884. Millikan, D.S., Felbeck, H., and Stein, J.L. 1999. Identification and characterization of a flagellin gene from the endosymbiont of the hydrothermal vent tubeworm Riftia pachyptila. Applied Environmental Microbiology 65: 3129–3133. Miura, T., Tsukahara, J., and Hashimoto, J. 1997. Lamellibrachia satsuma, a new species of vestimentiferan worms (Annelida: Pogonophora) from a shallow hydrothermal vent in Kagoshima Bay, Japan. Proceedings of the Biological Society of Washington 110: 447–456. Mullineaux, L.S., Fisher, C.R., Peterson, C.H., and Schaeffer, S.W. 2000. Tubeworm succession at hydrothermal vents: use of biogenic cues to reduce habitat selection error? Oecologia 123: 275–284. Naganuma, T., Kato, C., Hirayama, H., Moriyama, N., Hashimoto, J., and Horikoshi, K. 1997a. Intracellular occurrence of e-proteobacterial 16S rDNA sequences in the vestimentiferan trophosome. Journal of Oceanography 53: 193–197. Naganuma, T., Naka, J., Okayama, Y., Minami, A., and Horikoshi, K. 1997b. Morphological diversity of the microbial population in a vestimentiferan tubeworm. Journal of Marine Biotechnology 5: 119–123. Nelson, D.C. and Fisher, C.R. 1995. Chemoautotrophic and methanotrophic endosymbiotic bacteria at deep-sea vents and seeps. In: Microbiology of Deep-Sea Hydrothermal Vents. D.M. Karl, ed., CRC Press Inc, Boca Raton, Fl, pp. 125–167. Nelson, K. and Fisher, C.R. 2000. Absence of cospeciation in deep-sea vestimentiferan tube worms and their bacterial endosymbionts. Symbiosis 28: 1–15. Olu, K., Lance, S., Sibuet, M., Henry, P., Fiala-Medioni, A., and Dinet, A. 1997. Cold seep communities as indicators of fluid expulsion patterns through mud volcanoes seaward of the Barbados accretionary prism. Deep Sea Research Part I Oceanographic Research 44: 811. Olu-LeRoy, K., Rigaud, V., Fifis, A., Fabri, M.C., Cochonat, P., Ondréas, H., and Sibuet, M. 2001a. Spatial distribution of chemosynthetic fauna from video records and mosaic analysis at a new cold seep site in the Gulf of Guinea. In: The 2nd International Symposium on Deep-Sea Hydrothermal Vent Biology. Brest, France. Olu-LeRoy, K., Sibuet, M., Levitre, G., Gofas, S., and Fiala-Médioni, A. 2001b. Cold seep communities in the deep Mediterranean Sea (south of Crete and Turkey): Faunistic composition and spatial distribution. In: The 2nd International Symposium on Deep-Sea Hydrothermal Vent Biology. Brest, France. Page, R. 1995. TREEMAP: University of Glasgow. http://taxonomy.zoology.gla.ac.uk/rod/treemap.html. 40 E.R. MCMULLIN ET AL.

Paull, C.K., Hecker, B., Commeau, R., Freeman-Lynde, R.P., Neumann, C., Corso, W.P., Golubic, S., Hook, J.E., Sikes, E., and Curray, J. 1984. Biological communities at the Florida escarpment resemble hydrothermal vent taxa. Science 226: 965–967. Peek, A., Gustafson, R., Lutz, R., and Vrijenhoek, R. 1997. Evolutionary relationships of deep-sea hydrothermal vent and cold-water seep clams (Bivalvia: Vesicomyidae): Results from the mitochondrial cytochrome oxidase subunit I. Marine Biology 130: 151–161. Powell, M.A. and Somero, G.N. 1986. Adaptations to sulfide by hydrothermal vent animals: sites and mechanisms of detoxification and metabolism. Biological Bulletin 171: 274–290. Robinson, M., Gouy, M., Gautier, C., and Mouchiroud, D. 1998. Sensitivity of the relative- rate test to taxonomic sampling. Molecular Biology and Evolution 15: 1091–1098. Robinson-Rechavi, M. and Huchon, D. 2000. RRTree: Relative-Rate Tests between groups of sequences on a phylogenetic tree. Bioinformatics 16: 296–297. Rouse, G.W. and Fauchald, K. 1997. Cladistics and polychaetes. Zoologica Scripta 26: 139–204. Rouse, G.W. 2001. A cladistic analysis of Siboglinidae Caullery, 1914 (Polychaeta, Annelida): formerly the phyla Pogonophora and Vestimentifera. Zoological Journal of the Linnean Society 132: 55–80. Rzhetsky, A. and Nei, M. 1993. Theoretical foundation of the minimum-evolution method of phylogenetic inference. Molecular Biology and Evolution 10: 1073–1095. Saitou, N. and Nei, M. 1987. The neighbor-joining method: A new method for reconstructing phylogenetic trees. Molecular Biology and Evolution 4: 406–425. Sarrazin, J., Robigou, V., Juniper, S.K., and Delaney, J.R. 1997. Biological and geological dynamics over four years on a high-temperature sulfide structure at the Juan de Fuca Ridge hydrothermal observatory. Marine Ecology Progress Series 153: 5–24. Segonzac, M., Hekinian, R., Auzende, J.M., and Francheteau, J. 1997. Recently discovered animal communities on the South East Pacific Rise. Cahiers de Biologie Marine 38: 140–141. Shank, T.M., Fornari, D.J., Von Damm, K.L., Lilley, M.D., Haymon, R.M., and Lutz, R.A. 1998. Temporal and spatial patterns of biological community development at nascent deep-sea hydrothermal vents (9°50'N, East Pacific Rise). Deep-Sea Research II 45: 465–515. Sibuet, M. and Olu, K. 1998. Biogeography, biodiversity and fluid dependence of deep-sea cold-seep communities at active and passive margins. Deep-Sea Research Part I-Topical Studies in Oceanography 45: 517. Smith, D.C. and Douglas, A.E. 1987. The Biology of Symbiosis. Edward Arnold, Baltimore. Southward, E.C. 1991. 3 new species of Pogonophora, including 2 vestimentiferans, from hydrothermal sites in the Lau Back-Arc Basin (Southwest Pacific-Ocean). Journal of Natural History 25: 859–881. Southward, E.C., Tunnicliffe, V., and Black, M. 1995. Revision of the species of Ridgeia from northeast pacific hydrothermal vents, with a redescription of Ridgeia piscesae Jones (Pogonophora, Obturata Equals Vestimentifera). Canadian Journal of Zoology- Revue Canadienne De Zoologie 73: 282–295. VESTIMENTIFERAN TUBEWORM/SYMBIONT PHYLOGEOGRAPHY 41

Southward, E.C. and Galkin, S.V. 1997. A new vestimentiferan (Pogonophora: Obturata) from hydrothermal vent fields in the Manus Back-arc Basin (Bismarck Sea, Papua New Guinea, Southwest Pacific Ocean). Journal of Natural History 31: 43–55. Southward, E.C. 1999. Development of Perviata and Vestimentifera (Pogonophora). Hydrobiologia 402: 185–202. Southward, E.C., Schulze, A., and Tunnicliffe, V. 2002. Vestimentifera (Pogonophora) in the Pacific and Indian Oceans: a new genus from Lihir Island (Papua New Guinea) and the Java Trench, with the first report of Arcovestia ivanovi from the North Fiji Basin. Journal of Natural History 36: 1179–1197. Suess, E., Carson, B., Ritger, S.D., Moore, J.C., Jones, M.J., Kulm, L.D., and Cochrane, G.R. 1985. Biological communities at vent sites along the subduction zone off Oregon. Bulletin of the Biological Society of Washington 6: 475–484. Tajima, F. 1993. Unbiased estimation of evolutionary distance between nucleotide sequences. Molecular Biology and Evolution 10: 677–688. Thompson, J.D., Higgins, D.G., and Gibson, T.J. 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position specific gap penalties and weight matrix choice. Nucleic Acids Research 22: 4673–4680. Tunnicliffe, V. 1991. The biology of hydrothermal vents – ecology and evolution. Oceanography and Marine Biology 29: 319–407. Tunnicliffe, V., McArthur, A., and McHugh, D. 1998. A biogeographical perspective of the deep-sea hydrothermal vent fauna. Advances in Marine Biology 34: 353–442. Urcuyo, I.A., Massoth, G.J., MacDonald, I.R., and Fisher, C.R. 1998. In situ growth of the vestimentiferan Ridgeia piscesae living in highly diffuse flow environments in the main Endeavor Segment of the Juan de Fuca Ridge. Cahiers de Biologie Marine 39: 267–270. Webb, M. 1969. Lamellibrachia barhami, gen. nov., sp. nov. (Pogonophora), from the northeast Pacific. Bulletin of Marine Science 19: 18–47. Weisburg, W.G., Barns, S.M., Pelletier, D.A., and Lane, D.J. 1991. 16S ribosomal DNA amplification for phylogenetic study. Journal of Bacteriology 173: 697–703. Williams, N.A., Dixon, D.R., Southward, E.C., and Holland, P.W.H. 1993. Molecular evolution and diversification of the vestimentiferan tube worms. Journal of Marine Biological Association of the UK 73: 437–452. Winnepenninckx, B., Backeljau, T., and Dewachter, R. 1995. Phylogeny of protostome worms derived from 18S rRNA sequences. Molecular Biology and Evolution 12: 641–649. 14

CHAPTER 4

Twelve microsatellites for two deep sea polychaete tubeworm species, Lamellibrachia luymesi and Seepiophila jonesi, from the Gulf of Mexico

Accepted with revisions: Molecular Ecology Notes Resubmitted May 1, 2003 15

Twelve microsatellites for two deep sea polychaete tubeworm species, Lamellibrachia luymesi and Seepiophila jonesi, from the Gulf of Mexico.

E.R. McMullin, J. Wood, andS. H. Marrus

Abstract: The vestimentiferan tube worms Lamellibrachia luymesi and Seepiophila jonesi are found at hydrocarbon seeps in the Gulf of Mexico (GoM). Primers for polymorphic microsatellite loci were developed from genomic libraries of L. luymesi (5 loci) and from S. jonesi tissue (8 loci), and were used to screen individuals collected from nine northern GoM hydrocarbon seep sites. Loci had from four to more than fifty alleles, with high expected levels of heterozygosity. Cross-species amplification, tested on seven vestimentiferan species including both hydrothermal vent and cold seep species, was generally strong in similar species, but weak in more genetically distant species. 16

Text: Deep sea vestimentiferan tubeworms lack a digestive tract and rely on internal sulfide- oxidizing symbionts for the bulk of their energy needs. These polychaete annelids of the family Siboglinidae (Rouse, 2001, Halanych et al., 2001) are known from depths of up to 3500m in areas of sulfide seepage or venting (McMullin, 2003). Vestimentiferans fall into three distinct groups, the hydrothermal vent vestimentiferans and the cold seep escarpids and lamellibrachids (McMullin, 2003). Lamellibrachia luymesi and Seepiophila jonesi are found at hydrocarbon seeps in the northern Gulf of Mexico (GoM). These co-occurring sedentary deep sea annelids are sexually dimorphic, forming a motile larval stage from gametes released into the water column (Young et al., 1996). Larvae appear to settle on exposed carbonate rock formed in areas of high levels of hydrocarbon seepage (Behrens, 1988). Once settlement occurs, the sedentary adult vestimentiferans live 200 or more years (Bergquist et al., 2000). Partial genomic libraries were created using DNA isolated from the plume region of a single L. luymesi and a single S. jonesi. A CAA-enriched library was attempted first (Armour et al., 1994), but resulted in the isolation of very repetitive DNA that generated few effective primers. Genomic DNA was digested with Sau3A1 (New England Biolabs) and fragments of 200-1000bp were isolated, purified (QIAGEN) and ligated into a BamH1-digested pBluescriptSK+ vector (Stratagene). E. coli (Epicurian Coli Supercompetent cells, Stratagene) were transformed as per manufacturer’s instructions. Insert-containing clones were screened with four repetitive biotinylated probes, GA, GAAA, and GAT/CA (Ausubel et al., 1989) and probed filters were developed with a Biotin Luminescent Detection Kit (Boehringer-Mannheim). Positive clones were amplified with pBluescript-specific primers and clones containing inserts >150bp were sequenced (Beckman CEQ 2000XL) with the same primers using CEQ DTCS Quickstart cycle sequencing kit (Beckman). Primers were designed to the flanking regions of microsatellite DNA using Primer Designer (Scientific & Educational Software). Of 46 sequenced L. luymesi clones, 21 contained inserts with repetitive DNA. Three microsatellite sequences were found twice, one (L1-2E) containing two separate microsatellite regions, and two (L2-2D and L2-3B) containing both microsatellite and minisatellite regions. The remaining 15 contained a single unique repeat sequence. Primers were designed for 15 repeat regions, including specific primers flanking each repeat in L1-2E and L2-2D. Five loci generated monomorphic products, and of the ten remaining polymorphic loci five (including L1-2E#1) were unscorable due to nonspecific amplification or stuttering. Five loci (Table 1) generated repeatable 17 and scorable PCR products, including both L2-2D repeats. Twenty five of the 74 sequenced S. jonesi clones contained repetitive DNA. No identical sequences were found, but eight clones contained similar inserts; for example, three clones (E2-7G, ECL2-8B, and E1-8E) were only 1-2% different (see Table 1 for accession numbers). A single insert (E1-10A) contained two repeat regions. Primers were designed for 19 of the repeat regions, including primers flanking each repeat in E1-10A. Three loci generated monomorphic products, one locus failed to amplify, two generated products with alleles outside of the scorable size range (>600bp), and five loci were unscorable due to stuttering within the PCR product. A final locus (E3-10A) was monomorphic in 100 individuals except for three rare alleles. Seven loci (Table 1) generated repeatable and scorable PCR products, including one of the two repetitive regions in E1-10A. Of the similar sequences mentioned above, only E1-10A was among the final seven loci. Twelve variable primers (Table 1) were used to screen L. luymesi and S. jonesi from five collection areas in the northern GoM, spanning a total of 580km (530m to 640m in depth). Collection areas were spaced roughly 140km apart, the centermost of which included five samples sites located between 2km and 57km of each other. The remaining four collection areas included one to three sample sites separated by less than 100m. Tissue samples collected from the individuals were immediately frozen in liquid nitrogen, and were transferred to a -80°C freezer upon return to the lab. Genomic DNA was isolated from these samples by standard phenol/chloroform extraction procedures (Ausubel et al., 1989). PCR amplification was carried out in a 10µL reaction volume using a Perkin Elmer thermocycler. Final amplification conditions consisted of 20ng unlabeled reverse primer, 20ng Beckman fluorescently labeled forward primer,

50µM of each dNTP, 0.25 units of Taq DNA polymerase and buffer with 1.5mM MgCl2 (Gene Choice). The thermal profile for PCR amplification was 95°C for 5 min followed by 30 cycles of 95°C for 1 min, a primer specific annealing temperature (Table 1) for 1 minute, and 72°C for 1 minute, with a final extension of 72°C for 5 minutes. Allele sizes were determined by separation of the PCR products on a Beckman CEQ2000XL capillary sequencer, and fragment length was assigned with the Fragment Analysis software package, using an internal 600bp size standard (Beckman). ARLEQUIN (Schneider, 2000) was used to calculate heterozygosity at each locus, to test for Hardy Weinberg equilibrium of the genotypic frequencies (p<0.01) and to test for linkage disequilibrium between loci (p<0.01). Table 1 also shows the number and size of observed alleles and observed and expected heterozygosity for each locus. Allele numbers ranged from 4 to over 50, and the average observed heterozygosity for each species, excluding results from E3-10A, was 0.78 18 in L. luymesi and 0.70 in S. jonesi. The observed heterozygosity was significantly lower than expected in nine of the twelve loci amplified. Cross-species amplification for the twelve loci was tested (52°C annealing temp) in six species of vestimentiferans other than the original species (Table 2). PCR primers were scored by ability to generate polymorphic product at a 52°C annealing temperature. Successful amplification of microsatellite loci generally occurred in species similar to the source species, but was less successful across deeper evolutionary distances. Four of the five L. luymesi primers amplified L. barhami DNA, and between four and seven of the eight S. jonesi loci successfully amplified other escarpid species. Only two escarpid loci generated strong products in a lamellibrachid species, and no escarpid or lamellibrachid locus generated a strong product in the hydrothermal vent species Ridgeia piscesae. Twelve of the polymorphic loci isolated here (excluding E3-10A) are currently being used in a study of population structure and migration among populations of L. luymesi and S. jonesi in the northern Gulf of Mexico. In addition, the cross-amplification of these markers should make them useful in population studies of other seep vestimentiferan species worldwide.

Acknowledgements:

Thanks to Christi Ludwig for DNA extractions and preliminary primer development, and to K. Nelson, E. Cordes, S. Hourdez, J. Freytag and C. Fisher for sample collections. We thank the captain and crew of the RV Edwin Link and the submersible crew and pilots of the Johnson Sea Link. This work was supported by grants to C. Fisher and S. Schaeffer from the NOAA National Undersea Research Program at the University of North Carolina, Wilmington, the Minerals Management Service, Gulf of Mexico Regional OCS Office through contract number 1435-10-96- CT30813 and NSF OCE 0117151.

Figures:

Table 1 Characteristics of 12 vestimentiferan tubeworm microsatellite DNA loci. Included are locus designation, primer sequence, GenBank accession numbers, repeat motif, range of PCR products in base pairs, and number of alleles observed. Heterozygosities are given for the total population as well as for those samples sites where n>20 in at least one species (top number=HO, 19 bottom=HE, asterisks indicate a significant difference between HO and HE, p<0.01). Pairs of loci which failed a linkage equilibrium test (p<0.01) are outlined in black. Other sequences isolated for this study are available in Genbank (AY263742 -AY263780), and additional primer sequences are available upon request from the author.

Table 2 Cross amplification of isolated microsatellite loci. All twelve loci were tested in seven different vestimentiferan species at 52°C annealing temperature. Symbols indicate variable products (+++), possible monomorphic product (++), product that may improve with PCR optimization (+), or little or no discernable product (-). N = number of animals screened.

References: Armour, J. A. et al. (1994). Isolation of human simple repeat loci by hybridization selection. Human Molecular Genetics 3(4), 599-565.

Ausubel, F. et al. (1989). Current Protocols in Molecular Biology, John Wiley & Sons, New York.

Behrens, E. W. (1988). Geology of a continental slope oil seep, Northern Gulf of Mexico. The American Association of Petroleum Geologists Bulletin 72(2), 105-114.

Bergquist, D. C., Williams, F. M. & Fisher, C. R. (2000). Longevity record for deep-sea invertebrate. Nature 403(6769), 499-500.

Halanych, K. M., Feldman, R. A. & Vrijenhoek, R. C. (2001). Molecular evidence that Sclerolinum brattstromi is closely related to vestimentiferans, not to frenulate pogonophorans (Siboglinidae, Annelida). Biological Bulletin 201(1), 65-75.

McMullin. et al. (2003). Phylogeny and biogeography of deep sea vestimentiferan tubeworms and their bacterial symbionts. Symbiosis 34(1), 1-41.

Rouse, G. W. (2001). A cladistic analysis of Siboglinidae Caullery, 1914 (Polychaeta, Annelida): formerly the phyla Pogonophora and Vestimentifera. Zoological Journal of the Linnean Society 132(1), 55-80.

Schneider, S., Roessli, D., Excoffier, L. (2000). Arlequin: A software for population genetic data. 2.0 edit. Genetics and Biometry Laboratory, University of Geneva, Switzerland.

Young, C. M. et al. (1996). Embryology of vestimentiferan tube worms from deep-sea methane/sulfide seeps. Nature 381, 514.

20 Table 4.1 T Repeat motif in Size # of Primer sequences (5'-3') GenBank Sample locations °C clone range alleles Total VK GC2 BP BH GC3 Lamellibrachia luymesi n=235 n=25 n=34 n=23 n=93 n=22

L1- Fw: GGC AAT TGT TGA GGA CGT GT 339- 0.71* 0.80 0.70* 0.68 0.67* 0.76 AY263756 62 (CAC/AG)12 12 2E#2 369 Rv: GGA AGT GAA CCA ATG CTC TG 0.77 0.81 0.80 0.74 0.74 0.77

L2-2D Fw: ATA AGA TGC GAC TTC GAT GC 245- 0.78* 0.72 0.79 0.78 0.77* 0.90 AY263755 62 (CA)13/(CACG)3 22 inner 287 Rv: CTC TAC ATG AAC AAG TTT GC 0.90 0.91 0.89 0.92 0.91 0.93

L2-2D Fw: GTA AAC CGA TAT ACC CAC CT (CA)13/(CACG)3, 342- 0.81* 0.72* 0.50 0.57 0.44* 0.63 AY263755 62 20 all (ATCACGTATT)7 396 Rv: TGT CTC TTC TTT ACA GCT GC 0.91 0.93 0.55 0.67 0.64 0.85

Fw: GGC AAT AAT CCG CTA GAT GG 219- 0.76* 0.77 0.82 0.78 0.77* 0.71 L4-1E AY263750 62 (CA)10/ (TAAA)3 20 257 Rv: CGT AGC ACG TAT CGT TCT CA 0.91 0.93 0.91 0.92 0.90 0.88

Fw: TCT GAG CGG TGA ACT GTA TC (CAAA)12/ 320- 0.87* 0.84* 0.82 0.96 0.87* 0.73* L4-5E AY263752 62 30 (CAGA)10 449 Rv: GAT AAG TTC TCG TCG CTC GT 0.93 0.90 0.93 0.93 0.94 0.94 Seepiophila jonesi n=168 n=12 n=8 n=17 n=74 n=21

E1- Fw: GTC TTC TGA TGT CAT GGT CC 128- 0.85 0.67 1.00 0.82 0.81 0.85 AY263764 60 (CACG)9/(CA)13 23 10A#2 234 Rv: CGA ACA CCT CGA CAA TCA AC 0.87 0.89 0.82 0.89 0.85 0.87

Fw: CAG TGT CAT ACA GCG GGC TT 127- 0.47* 0.58 0.63 0.41 0.49* 0.33 E2-2A AY263772 55 (TGCGGGCG)4 7 179 Rv: ATC CGC CTC TGC CTA CGT TT 0.63 0.70 0.77 0.63 0.64 0.51

Fw: GAT CTC TTC TGG CAG GCG TT 128- 0.88 0.92 0.88 0.88 0.93 0.90 E3-2D AY263777 60 (CGTG)9 24 195 Rv: CTC CAT TGC ATC TAC GGG TC 0.87 0.87 0.87 0.91 0.89 0.87

Fw: AGGTCAGAGGCATTGCCATA 305- 0.04 n/a n/a n/a n/a n/a E3-10A AY263780 60 (TGTC)8 4 329 Rv: ACAACGGACAGGTCTGCATT 0.04

Fw: TCA TTG CTC TCC TGG TTT AG 271- 0.80* 1.00 0.88 0.82 0.77 0.71 E4-7G AY263779 60 (GT)19 20 316 Rv: GCC TGG TTT CTG ATG ACT TA 0.83 0.90 0.94 0.88 0.82 0.75

ECL1- Fw: TAA CAC TGA CCA CGT CAC TG 260- 0.84* 0.75 0.86 0.82 0.76* 0.90 AY263761 60 (GT)22 40 5B 395 Rv: GTC GCG TAG ACC TAG CAA TA 0.95 0.96 0.97 0.95 0.95 0.94

ECL2- Fw: GGA GCC TCC TTG ACT TTA CA 194- 0.27* 0.08 0.23 0.53 0.24 0.38 AY265354 55 (CTGT)7 5 12C 224 Rv: CAA AGC TCA TCG CAC ACT TG 0.33 0.31 0.53 0.43 0.32 0.40

ECL3- Fw: GCG TTG CTA ACT GCC AAG TG 323- 0.76 0.83 0.88 0.63* 0.75 0.75 AY263763 57 (TGCG)9 54 8D 563 Rv: ATC TTG ACC AGT CGC TGA CC 0.93 0.96 0.95 0.97 0.97 0.95 21 Table 4.2 Lamellibrachids Escarpids Vent species

Lamellibrachia Lamellibrachia Seepiophila undescribed Escarpia Paraescarpia Ridgeia Locus luymesi barhami jonesi escarpid laminata echinospica piscesae

n=235 n=4 n=165 n= 4 n=3 n=1 n=2

L1-2E#2 + + + + + + - + + -

L2-2D, inner + + + + + + + + + -

L2-2D, all + + + + - - - + -

L4-1E + + + + + + - - - + -

L4-5E + + + + + + + - - + -

L5-1B + + + + + + - + + -

E1-10A#2 + + - + + + + + + + + + + + +

E2-2A + - + + + + + - - -

E3-10A + - + + + + + - + -

E3-2D + + + + + + + + + + + + + + + + +

E4-7G + - + + + + + + + + + + + -

ECL1-5B + - + + + - + + + - +

ECL2-12C - - + + + + + + + + + + -

ECL3-8D + - + + + + + + + + - +

22

CHAPTER 5

Genetic diversity and population structure of two deep sea tubeworms, Lamellibrachia luymesi and Seepiophila jonesi, from the hydrocarbon seeps of the Gulf of Mexico

For submission to: Molecular Ecology 23

Genetic diversity and population structure of two deep sea tubeworms, Lamellibrachia luymesi and Seepiophila jonesi, from the hydrocarbon seeps of the Gulf of Mexico.

Erin R. McMullin, S.W. Schaeffer, and C.R. Fisher

Vestimentiferan tubeworms are a group of sessile deep sea marine polychaete annelid (family Siboglinidae) that as adults completely lack a digestive tract and rely on internal sulfide oxidizing symbionts for fixed carbon (Cavanaugh et al., 1981; Childress et al. 1991). Vestimentiferans therefore inhabit regions where hydrogen sulfide is expressed on the seafloor, such as the hydrothermal vents associated with ocean spreading centers and cold seeps associated with geologic features such as hydrocarbon reservoirs. (Tunnicliffe et al., 1998; Sibuet & Olu, 1998). The species that compose this group of siboglinids live at depths between 82m and 3300m (reviewed in McMullin, 2003). Cold seep vestimentiferan communities are more globally distributed than the hydrothermal vent tubeworms (Tunnicliffe et al., 1998; Sibuet & Olu, 1998; McMullin, 2003), and are more opportunistic in habitat, with members of the group found in habitats as different as hydrocarbon seeps (MacDonald et al., 1990), decaying whalefalls (Smith et al., 1989), and even a shipwrecked coffee shipment (Dando et al., 1992). Vestimentiferans have lecithotrophic trochophore larvae (Young et al., 1996); dispersal of these larvae depends both on the duration and position of the larvae in the water column and on the currents they experience. Measurements of metabolic rate and energy stores in larvae of the vent species Riftia pachyptila suggest a larval lifespan of less than 40 days (Marsh et al., 2001). Allozyme and genetic data have revealed little population differentiation among Riftia pachyptila across 4000km of the East Pacific Rise spreading center (Bucklin, 1988; Jollivet, 1996; Vrijenhoek, 1997). Hydrothermal vent sites are typically linearly arrayed along spreading center ridge axes, with vent plumes that rise upward through the water column, and near-bottom currents that flow along the axis of the rift valley. Propagule dispersal modeling of vent organisms based on this water current regime cannot completely explain the amount of genetic homogeneity observed in vent vestimentiferans (Chevaldonné et al., 1997; Kim & Mullineaux, 1998). Modeling larval dispersal between seep sites is more complicated than between vent sites, where water currents are strongly affected by the geology and topography of the spreading centers. Seep sites are formed in a variety of geologic and biologic settings, are not necessarily linearly 24 arrayed, and are subject to local bottom water currents. Additionally, hydrothermal vents are ephemeral environments, and hydrothermal vent vestimentiferan communities survive on the order of years to decades (Hessler et al., 1988; Sarrazin et al., 1997). Cold seeps, on the other hand, are stable with estimated population persistence on the order of decades to centuries (Aharon, 1994). While vent vestimentiferans are among the fastest growing invertebrates on the planet, seep vestimentiferans are among the slowest (Bergquist et al., 2000). These differences in habitat and life history traits could result in very different patterns of geneflow and dispersal in vent and seep vestimentiferans. Seep vestimentiferans, like vent vestimentiferans, have very large species ranges, but research has thus far been limited to comparisons between gene and promoter region sequences (reviewed in McMullin, 2003; Kojima et al., 2003). To our knowledge only one population genetic study, which showed an isolation of L. columna populations by depth, has addressed the question of geneflow and dispersal within a cold seep species (Kojima et al., 2003). Hydrocarbon seepage on the Louisiana Slope (LS) in the northern Gulf of Mexico (GoM) supports large communities of the co-occurring cold seep vestimentiferan species Lamellibrachia luymesi and Seepiophila jonesi(Bergquist et al., 2000; Bergquist et al., 2002). The species found on the LS are unique from those found at Alaminos Canyon (2200m) in the western GoM (Brooks et al., 1990) and on the Florida Escarpment (3300m) in the eastern GoM (Paull et al., 1984), indicating an historical lack of gene flow between these populations. Species on the shallower LS (~600m), however, are constant across a large geographic region (Brooks et al., 1990), and are therefore good subjects for studies of geneflow and dispersal. L. luymesi and S. jonesi have separate sexes and reproduce by spawning gametes into the water column (Young et al., 1996). The larvae of the two species of seep vestimentiferans, Lamellibrachia luymesi and Seepiophila jonesi, survived in culture for approximately three weeks (Young et al., 1996). This motile larval stage settles on exposed carbonate rock formed in areas of active oil seepage (Fisher et al., 1997). The periodicity of spawning and the possible settlement cues are not known for these or other vestimentiferan species. After settlement, the worm remains attached to this rock, its body encased in a chitinous tube secreted by glands in the body wall of the animal. Seep vestimentiferans inhabit the entire length of their tube and are not capable of surviving if removed from it. Settlement of both species of larvae occurs in a limited time window after the formation of the exposed carbonate rock and before the rock becomes buried in sediment. The estimate of this settlement window ranges between 15 and 40 years (Bergquist et al., 2002), with the sessile adults living in excess of 200 years (Bergquist et al., 2000). LS species can be reproductively active before the end of this settlement window, and may remain reproductively 25 capable for their entire lifespan (E.C. Cordes, pers. comm.). Because settlement occurs on patchy exposed surfaces, vestimentiferan communities on the LS are characterized by large bush-like aggregations composed of hundreds or even thousands of similarly-aged individuals. A single LS hydrocarbon seep “site” is composed of many aggregations of varying ages, and sites themselves are patchily distributed across the LS. In this study we used polymorphic microsatellite loci to assess the structure of genetic variability in L. luymesi (five loci) and S. jonesi (seven loci) (McMullin, submitted) over a range of geographic distances. Multiple tubeworm aggregations were sampled from nine hydrocarbon seep sites spanning 580km across the Louisiana Slope (LS), and tests were conducted to investigate the population substructure among these sites.

Materials and Methods

Sampling

Samples were collected over multiple years, from 1997 to 2002, with either the Johnson SeaLink I or II and associated support ship (operated out of the Harbor Branch Oceanographic Institute). Tissue was collected from individual vestimentiferans from nine seep sites on the Louisiana Slope of the Gulf of Mexico (Figure 5.1). Seep sites were at 585 ± 55 m depth, and were between 2 km and 580 km apart (Table 5.1, Table 5.2). Within each seep site collections were made from between one and twelve tubeworm aggregations, and are indicated as in “BH-1”, “BH- 2”, etc. All collections were made from different aggregations except for GC2-3 and GC2-4, which were taken from the center and from one edge of the same aggregation. Individuals sampled were between 10 cm and 2 m in length above the point of attachment to the substrate. Four to 18 individuals were collected from each aggregation, and a total of 235 L. luymesi and 165 S. jonesi were analyzed for this study. The total number of samples taken from each species in each seep site is shown in Figure 5.1. To ensure the isolation of vestimentiferan and not bacterial endosymbionts DNA, tissue samples for DNA isolation were taken from the symbiont –free vestimentiferan plume. Tissue samples were immediately frozen in liquid nitrogen and were transferred to -80°C upon return to the laboratory.

26 Molecular analyses

Total DNA was extracted from approximately 1g of tissue by the standard phenol/chloroform extraction method (Ausubel et al., 1989). DNA variation was assayed using the five L. luymesi and seven S. jonesi microsatellite loci isolated from these species (McMullin, submitted). One primer from each locus set was fluorescently labeled (Beckman dye), and amplification reactions were carried out as described in (McMullin, submitted). One µL of PCR product was mixed with 35µL of formamide-containing sample loading solution (SLS, Beckman) and 0.25 µL of fluorescently labeled size standard (600bp, Beckman). Products were separated on a Beckman CEQ2000XL capillary sequencer. Locus allele size was assigned with the Fragment Analysis software package (Beckman) using the internal size standard. The exact composition and number of repeat units in the microsatellites of these alleles is unknown due to difficulties in sequencing through the repetitive region and to the high number of imperfect and minisatellite repeats found in or near the microsatellite repeat regions. Those individuals in which more than half the loci failed to amplify (<5 in each species) were removed from further calculations.

Descriptive statistics

Genetic Diversity, HWE and Linkage Disequilibrium: Populations were assessed for gene diversity (Nei, 1987), Hardy-Weinberg equilibrium and linkage equilibrium of loci using ARLEQUIN (Schneider, 2000). The observed number of heterozygotes in the entire dataset was compared first to the HWE expected heterozyosities calculated assuming all data was from one population and then to the HWE expected heterozyogosity calculated within each collection. The difference between the observed and expected number of heterozygous individuals was tested with a χ2 test. We tested for significant non-random association of loci using a likelihood-ratio test that compares the likelihood of the distribution of the data assuming linkage equilibrium (based on haplotype frequencies calculated as the product of the allele frequencies), to that assuming no linkage equilibrium (based on haplotype frequencies estimated with the expectation maximization [EM] algorithm). The test assumes Hardy-Weinberg proportions of genotypes, and a failure of the linkage equilibrium test could also result from a departure from HWE. Because independence of loci may be influenced by underlying population structure, linkage disequilibrium was also tested within each seep site and within each collection. 27

Null alleles: Null alleles fail to amplify due to alteration in one or both of the PCR oligonucleotide annealing sites, leading to the misidentification of null heterozygotes as homozygotes. An additional problem with microsatellite loci is the possibility of having “called” more alleles than actually exist due to stuttering in the product or artifacts of amplification, which would inflate estimates of expected heterozygosity, and result in an apparent heterozygote deficiency similar to that seen in the presence of null alleles.

The frequency of null alleles at each microsatellite locus was estimated assuming those individuals that failed to amplify at a given locus were null homozygotes (FA-null). Null allele frequencies were also estimated for each locus with GENEPOP (GP-null) (Raymond & Rousset, 2000) which applies an EM algorithm and a likelihood test to estimate the frequency of an assumed null allele to best explain the observed data. This estimate of null alleles assumes HWE and therefore may be influenced by underlying population structure. The distribution of null allele frequencies under each model was compared across loci and across inbreeding coefficients under the expectation that null alleles are uncommon, and that any locus with a null allele (or too many alleles) would have a higher than average estimated null allele frequency. Null frequencies were compared to inbreeding coefficients to compensate for the assumption made in GENEPOP of HWE.

Population Structure: Wright’s F-statistics were used to examine the structure of the genetic variability among samples (Wright, 1951) calculated in ARLEQUIN (Schneider, 2000). Nei’s standard genetic distance (D) (Nei, 1972) and Nei’s net divergence (Da) (Nei et al., 1983) were also calculated using ARLEQUIN and POPTREE. Due to the number of imperfect repeats and the difficulty in assigning repeat numbers to microsatellite-containing alleles, only statistics assuming the infinite-alleles model (IAM), and not the stepwise-model of microsatellite mutation (SMM), were estimated. The estimation of population differentiation from genetic distances based on the IAM model of mutation are fairly consistent with those which assume a SMM when migration rate is much greater than mutation rate (Balloux & Lugon-Moulin, 2002), which appears to be true in these data. Wright’s F statistics were estimated in ARLEQUIN with an Analysis of Molecular Variance

(AMOVA) approach using variance of gene frequencies. AMOVA-based calculation of FST was accomplished using a hierarchical partitioning of the total variance into covariance components due 28 to intra-individual differences, inter-individual differences, and/or inter-population differences.

These covariance components were then used to estimate Wright’s fixation index (FST). Fixation indices were calculated with this method for departure from HWE at four levels: 1) haplotypes within individuals, 2) individuals within collections, 3) collections within seep sites, and 4) seep sites geographic regions. Significant differences of FST values from zero were tested by comparing observed values to a null distribution estimated by randomly permuting the haplotypes, individuals, or populations, among individuals, populations, or groups of populations. Pairwise FST values were calculated between collections, between seep sites, and between geographic regions. Again the significance of the resulting FST values was tested against a null distribution obtained by randomly permuting haplotypes between the two populations. The probability for these tests is the proportion of permutations leading to an FST value larger than or equal to the observed one.

Pairwise D and Da values were used to construct Neighbor Joining (NJ) trees (Saito & Nei, 1987) for each species in all collections larger than 5 individuals. Robustness of the branching order of the NJ trees was tested by bootstrapping (10,000 replications) using POPTREE (Takezaki, 1998). Because L. luymesi showed signs of underlying population structure, a NJ tree was also constructed for pairwise Da values between L. luymesi seep sites.

Individual-based analyses: The relationship among the 165 S. jonesi or 235 L. luymesi was also explored by comparing the multilocus genotypes of each individual. Shared allele distances (DAS) were estimated with the Microsatellite Toolkit plug-in for EXCEL (Park, 2001). This distance matrix was then used to construct a NJ tree (Saito & Nei, 1987) using MEGA2 (Kumar et al., 2001) to test the hypothesis of no population structure between individuals from different collections. Underlying substructure was also explored with STRUCTURE 2.0 under the population admixture model (Pritchard et al., 2000), which assigns individuals to populations assuming that loci are at HWE within each population. STRUCTURE 2.0 estimates the population of origin of an individual and the allele frequencies in all populations from the observed genotypes of individuals using a Markov Chain Monte Carlo, and then calculates a membership coefficient (Q) for each individual to a particular population. Individuals will have a high Q value for any populations for which they show a strong affinity and low Q values for all other populations. STRUCTURE 2.0 was also used to estimate of number of populations from all L. luymesi and all S. jonesi microsatellite data. This is done by calculating the 29 probability (under the assumption of unlinked loci in HWE within each population) of the observed genotype data under a range of population sizes.

Gene Flow: Estimates of overall gene flow were calculated using GENEPOP in which the estimate of Nm is based on the average frequency of “rare” alleles found in only one population (Slatkin, 1985; Barton & Slatkin, 1986).

Isolation by distance: Isolation-by-distance was explored by comparing geographical and genetic distances among all collections. The log of geographic distance was plotted against Slatkin’s linearized FST, FST/(1-

FST) (Slatkin, 1995). The relationship between genetic distance and geographical distance was further examined with GENEPOP, which estimates the Spearman Rank correlation coefficient between FST/(1-FST) and the distance in km. Using this program, the significance of the correlation coefficient was tested with a Mantel test (50,000 permutations).

Results

General:

All loci used in this study show high variability, with numbers of alleles ranging from 7 to 50 (McMullin, submitted), and an average of 20 alleles (18.5 in L. luymesi and 22 in S. jonesi) per locus. Observed heterozygosity was also high across all loci, ranging from 0.71 to 0.90 in L. luymesi, and from 0.27 (5 alleles) to 0.84 in S.jonesi. The average gene diversity per locus in each collection ranged from 0.80 to 0.91 in L. luymesi and from 0.70 to 0.87 in S. jonesi.

HWE heterozygosity:

We used a chi squared goddness of fit test to determine if the observed heterozygosities at each locus departed from HWE expectations. When all samples from each species were assumed to be from a single interbreeding populations, all five L. luymesi loci and five of seven S. jonesi loci had a significant (p<.001) deficiency of heterozygotes. A chi squared goodness of fit test across all loci rejected the null hypothesis of HWE (p<0.01) for each species. Within each collection, 30 observed heterozygosities were consistently lower than expected for both species, but were often not significant. L. luymesi averaged one out of five loci significantly different from expected in individual collections, though all five loci were significantly different from expected in the collection with the largest sample size (BH-3, n=18). In S. jonesi, 20 of the 23 collections had no loci significantly different from expected heterozyogosity, and three had only one loci significantly different (p<0.01).

Linkage equilibrium:

Significant linkage disequilibrium was found between two L. luymesi loci, L2-2Dinner and L2-2Dall, in half of the eight seep sites (p<0.005). As reported by (McMullin, submitted) these loci are in fact within 200 bp of each other. Because L2-2Dinner and L2-2Dall are not independent loci, L2-2Dall was excluded from subsequent calculations. Significant departure from linkage equilibrium was also found within the Bush Hill samples (BH, N=93) between the L. luymesi loci L1-2E and L2-2Dinner, and between L1-2E and L4-1E. The apparent linkage disequilibrium between these loci could be due to non-random mating and not to physical proximity of the loci, therefore only L2-2Dall was removed from the dataset. No linkage disequilibrium (p<0.01) was found in the S. jonesi loci in any of the nine seep sites.

Null alleles: The observed heterozygote deficiency in both L. luymesi and S. jonesi could be the result of null alleles present in one or more loci or of underlying population structure. We examined whether null alleles were responsible for the heterozygote deficiency by plotting two different null allele estimates (FA-null and GP-null) for each locus versus an inbreeding coefficient, FIT.

Figure 5.2 shows both estimates of null allele frequencies vs. overall FIT values assuming all data are from one population. The presence of null alleles in a locus should result in a higher estimate of FIT, therefore if failure to amplify is the result of null alleles, those loci with higher FA- null values should also show higher FIT values. FA-null estimates were highly variable, ranging from 0.0 to .20, and showed no positive correlation with FIT values. The lack of correlation between

FA-null estimates and FIT values indicates that failure to amplify at a given locus is not due to the presence of null alleles. The GENEPOP null allele estimate (GP-null) assumes no population structure and therefore reflects both population structure and the presence of null alleles. GP-null estimates should be of 31 similar value if they predominantly reflect population structure. If null alleles are uncommon, the ratio of null allele frequency to FIT should be roughly constant across all loci (a positive correlation between the two numbers), and those loci with unusually high GP-null estimates may suffer from null alleles. GP-null estimates (Figure 5.2, triangles) were very similar in value, with more than half the estimates near 0.66. Two points, which correspond to the S. jonesi loci ECL3-8D and E2- 2A, were higher than the other nine values. Locus ECL3-8D had the largest number of alleles (50+) and may suffer from an inflated number of alleles called. Further analyses with S. jonesi were conducted both with and without these two loci, with no significant differences in the results.

Genetic Structure: Calculations of Wright’s F statistics were conducted to explore possible population structure. The most significant departure from HWE predicted values was in FIS

(haplotypes/individuals), which were similar in both L. luymesi (FIS=0.088, P<0.001) and S. jonesi

(FIS=0.104, P<0.001), with FIS explaining 90% of the genetic variation in both species. FIT (haplotype/total population) was also significantly positive in both species (0.085 in L.luymesi and

0.104 in S. jonesi) but FST (individuals/collections) was not significantly different from zero in either species. Wright’s F statistics were calculated a second time excluding the level of the individual to reveal more subtle population structure. L. luymesi showed an FST (seep sites/total population) of 0.008 and an FCT (geographic location/total population) of 0.009, both highly significant at p<0.0001. These values, however, only explain 1% of the variation seen among L. luymesi samples. FST or FCT was not significantly different from zero for S. jonesi.

Pairwise FST values were calculated between each collection, between each seep site, and between each geographic region for each species. Pairwise FSTs between L. luymesi collections ranged from -0.033 to 0.076, with significant difference (Bonferroni correction p<0.00015) between

GC3-3 and GC3-2, BH-3, BH-5, VK-2; and between VK-2 and GB4-1. Pairwise FSTs between L. luymesi seep sites ranged from -0.012 and 0.041, with significant differences (Bonferroni correction p<.0018) between VK and GC2, BH, GB4; between GC3 and GC2, BH, GB4; and between MC and GB4. When L. luymesi seep sites were combined by geographic region (East, Central, and West), significant differences were found between East and both Central (FST=0.017) and West

(FST=0.014), but not between Central and West (FST=-0.004). In S. jonesi FSTs between collections ranged from -0.035 to 0.097, and only FSTs between GC3-3 and GC2-1/2 (individuals from the neighboring aggregations GC2-1 and GC2-2), MC were significantly different from zero

(Bonferroni correction p<0.0003). Pairwise FST values between S.jonesi seep sites ranged from - 32 0.013 to 0.022, with a significant difference (Bonferroni correction p<0.0015) only between GC3 and MC. No significant difference was found in FST comparisons between geographic regions for this species.

Population structure was further explored with Nei’s pairwise Da values. Pairs of populations with Da values significantly different from zero in L.luymesi included GC3-3 and GC3-

2, VK-2, BH-3, BH-1; and BH-1 and BP-1. S. jonesi pairs with significant Da values were GC3-3 and GC2-1/2, MC. NJ trees of pairwise Nei’s D and Da values (10,000 bootstrap replications) were very similar in branching order and bootstrap values, therefore only the Da trees are shown here. NJ trees in both species reveal little population structure, with low bootstrap support (<75%) for all nodes. L. luymesi(Figure 5.3), with 6 nodes supported by bootstrap values of 50% or greater, generally shows slightly higher bootstrap values than S. jonesi (Figure 5.4), which only has two nodes supported with bootstrap values greater than 50%. Overall both trees show a lack of population structure, with no resolution of deeper branches and all sites appearing to cluster at random with each other regardless of geographic location. A neighbor joining tree of pairwise

Nei’s Da values between L.luymesi seep sites showed no branching order supported with bootstrap values higher than 50%.

Individual-based analyses: A neighbor joining tree constructed from shared allelic differences between individuals in both species generated a star-shaped tree with very short deep branches and very long shallow branches. Both trees showed little grouping of individuals by collection or sample location, though the tree for L.luymesi showed a slight tendency toward individuals from the same site falling into the same cluster. No underlying population structure was revealed by analyses in STRUCTURE 2.0. When the number of populations (k) was assumed to be equal to the number of seep sites, most individuals showed no clear affinity to any one of the k populations. In both species the probability of a given number of populations failed to converge on a single or small range of numbers of populations.

Gene Flow: Estimates of overall migration between sites using rare alleles were between 3 and 4 migrants per generation in both species (L. luymesi =3.8 migrants, S. jonesi = 3.0 migrants).

33 Isolation by distance: Correlation between geographic and genetic distances was explored first by plotting the log of geographic distance (with 0 distances rounded up to 1km) against Slatkin’s linearized FST (Figure

5.5). In general for both species, pairwise FST values vary widely even within seep sites (represented in the scatterplot by values on the Y-axis), and similar variation is seen for pairwise

FST values over larger distances. A trendline plotted through the pairwise S. jonesi FST values (Figure 5.5a) shows no linear trend (slope =2E-05, r2<0.0001) for increasing genetic distance with increasing distance, and the Spearman Rank correlation coefficient between genetic and geographic distance was, at 0.024, not significantly different from zero. A similar trendline through the pairwise L. luymesi FST values (Figure 5.5b) shows a very slight positive slope (slope =0.0018, r2=0.007). This slightly positive relationship between genetic and geographic distance is further supported by a low but significant Spearman Rank correlation coefficient (0.386, p<0.002).

Discussion

In both Lamellibrachia luymesi and Seepiophila jonesi the molecular genetic markers reveal an overall deficit in heterozygosity that does not appear to be the result of strong underlying population substructure. Very little of the genetic variability is due to geographic structure of either species, though L. luymesi shows a small amount of isolation by distance within the study area

(476km). Additionally, the high FIS levels are not easily explained by the presence of null alleles because similar values were obtained with and without suspect loci in the dataset. Both L. luymesi and S. jonesi show high levels of genetic variation at all microsatellite loci observed in this study, suggesting that neither species is under extreme selective pressure or has suffered a recent bottleneck event.

Population structure Aggregations of L. luymesi and S. jonesi show high levels of genetic patchiness between sites, within sites, and even within aggregations. Pairwise FST values between collections within a single seep site, which range from -0.03 to 0.11, match or exceed those between sites geographically distant from each other (Figure 5.5). Even animals sampled within a single aggregation show enough genetic difference between individuals that the two collections, GC2-3 and GC2-4, do not cluster together on the Da NJ tree (Figure 5.3). Given the number of alleles in 34 the microsatellite loci used in this study, and that each seep site is potentially made up of well over 1000 individuals, the apparent genetic differences between collections could simply be the result of sampling error. Alternatively, these differences may reflect an underlying population structure that is not statistically significant due to the relatively small sample sizes analyzed here. No evidence for population substructure was seen in S. jonesi. Comparisons of S. jonesi collections, seep sites and geographic regions failed to show FST values significantly different from zero, and a test for isolation by distance showed no correlation between genetic distance and geographic distance. Additionally, the Da based NJ tree showed no clustering of collections (Figure 5.4), and neither individual-based comparisons revealed any latent structure. All analyses conducted here suggest that S. jonesi from at least eight of the nine seep sites spanning 580 km have no population structure by collection location or geographic region. The only significant difference in S. jonesi was found between collections from GC3 and two other sites, GC2 and MC, indicating that S. jonesi in GC3 are somewhat genetically isolated from the other seep sites in the LS. L. luymesi shows evidence for low levels of population structure and isolation by distance among the eight sites from which it was sampled (476 km). Though the low FST values explain only a very small percentage of the overall genetic variation, the values are consistent (between sites = 0.008, between regions = 0.009) and significant. The neighbor joining tree for L. luymesi (Figure 5.3) is similar to S. jonesi in that no branch nodes are supported by strong bootstrap values. However, more nodes are supported by values over 50 than in the S. jonesi tree. Additionally, pairwise FSTs between collections suggest isolation by distance, with an eastern collection,VK-2, significantly different from GB4-1, the westernmost L. luymesi collection. Pairwise FSTs between seep sites show this same trend more strongly, probably due to the larger samples sizes. VK (eastern) was significantly different from GC2 and BH (central), and both VK and MC (eastern) were significantly different from GB4 (western). Comparisons of the three geographic regions revealed significant differences between East sites and both Central and West sites, supported by somewhat higher FST values (0.017 and 0.014, respectively). Finally, a test of isolation by distance revealed a small (0.386) but significant correlation between genetic and geographic distance. In addition to a low level of isolation by distance, L. luymesi show the same tendency for genetic isolation of seep site GC3 as seen in S. jonesi. Collection GC3-3 was significantly different from four collections, including a second collection from the same seep site (BH-3, BH-5, VK-2 and GC3-2), and the combined L. luymesi from seep site GC3 were significantly different from those of GC2, BH, and GB4. 35 These data indicate that L. luymesi dispersal may be more limited than S. jonesi, an interesting result for two species which almost always co-occur in the same aggregations. The result is also interesting because L. luymesi tend to dominate the top of a bush-like aggregation, while S. jonesi are most often found at the base of the aggregation near the sediment-water interface (Bergquist et al., 2002). Current data suggests that both species have slightly buoyant larvae that rise in the water column, persisting for possibly three weeks before settlement (Young et al., 1996). A larva will be subjected to different current patterns depending on the depth to which it rises. The observed difference in gene flow among sites in L. luymesi and S. jonesi may very well reflect differences in either larval longevity or buoyancy. Little is known of the deep currents in the Gulf of Mexico. A strong loop current enters the southern GoM and exits south of Florida. This current can generate eddies along its edges that move in the opposite direction, from east to west. Preliminary data on deep sea currents suggest, however, that the northern Gulf of Mexico (500m to 1000m) experiences an average easterly current with a mean velocity of 10cm/s (Oceanexplorer, online). A calculation from this average current speed results in an estimated dispersal of only roughly 26 km in the three week larval lifespan. Loop current eddies may interrupt this average velocity, moving water across both depth and distance for transient higher velocities, which could possibly explain the lack of population structure and isolation by distance among these site. Alternatively, undiscovered seep sites may exist between those sampled in this study, allowing for genetic exchange across shorter distances. Given the overall difference in population structure between S. jonesi and L. luymesi, it is surprising that both species show an interruption in gene flow with GC3. GC3, located only 35 km from BH, is a centrally located site which does not differ in depth or latitude from the other sites from which S. jonesi and L. luymesi were collected. GC3 may experience some other physical barrier to dispersal, such as interrupted water currents between it and other sites. Whatever barrier does exist must affect the different dispersal patterns of both L. luymesi and S. jonesi. Interestingly, faunal assemblages in tube worm aggregations collected from these sites are slightly different from those found at other sites sampled (Erik Cordes, pers. comm.). Genetic and ecologic data both point toward isolation of this site; further research on the physical characteristics could reveal interesting differences between it and other LS sites.

A general excess of homozygotes A striking result of this study is the general deficiency of heterozygotes. Individuals of both species are more often homozygous for one of the microsatellite loci tested than is expected under 36

HWE. Both species show FIS values near 0.09, an order of magnitude higher than the FST values supporting population substructure in L. luymesi. A deficiency in heterozygotes can be explained by a number of reasons. The existence of substructure within a defined ‘population’ can generate an apparent heterozyote deficiency by violating the Hardy Weinberg assumption of random mating within the population. Other violations of the assumption of random mating, such as mating of related individuals, assortative mating, and self fertilization, will also cause a decrease in heterozygosity. The presence of null alleles or other issues in the scoring alleles can also cause an apparent loss of heterozygosity. Population structure is not strongly supported by the data presented here. Evidence supports a general absence of population structure within a collection, within a site, and within a geographic region. Vestimentiferans have separate sexes, and no support exists for the presence of self fertilization in this group. Additionally, because they reproduce by spawning gametes into the water column, adult vestimentiferans have no opportunity for assortative mating. Assortative mating could occur between gametes, but seems highly unlikely due to the probable low density of gametes in the water column. Finally, null alleles do not appear to be a significant problem in this dataset. Comparisons of null-allele frequencies (FA-null and GP-null) and inbreeding coefficients across the loci used in this study do not support a strong effect of null alleles. When the two loci with higher GP-null allele frequencies were removed from the analyses FIS values did not change appreciably. Observed FIS values are therefore most likely not a product of null alleles. The source of the departure from HWE that generates the observed heterozygote deficiency is not clearly defined by the data presented here. We will suggest two possible explanations for the observed trends that can be tested with further sampling. First, even though no population substructure was revealed by these analyses, substructure may in fact exist within an aggregation. Tube worm aggregations on the LS often have in excess of 1,000 individuals (Bergquist et al., 2002), and sibling cohorts may not be revealed by sample sizes of 6 to 18 individuals using highly polymorphic loci such as microsatellites, as used in this study. If an aggregation is composed of multiple cohorts of siblings, these siblings would only show excess homozygosity if their parents had a higher than expected chance of sharing an allele at a given loci. Because these animals reproduce by spawning gametes into the water column, a larva can potentially be the product of any male and female tubeworm that spawn at the same time, which suggests that the production of inbred larvae is unlikely. However, if an aggregation is made up of multiple sibling groups, any individual will be related to a certain fraction of others within the aggregation. The chance of producing larvae is likely greatly increased if two neighboring tubeworms happen to release 37 gametes at the same time. Based on these assumptions, a significant fraction of larvae produced may in fact be the offspring of related parents. Even a small increase in the chance of mating between related individuals could over time result in a deficiency in heterozygotes. However, inbreeding should increase the occurrence of individuals homozygous at multiple loci, and will make loci move through the population in a nonrandom fashion, which can be revealed by a linkage disequilibrium test. The data presented here show little evidence for linkage disequilibrium between loci in either species. L. luymesi shows some linkage disequilibrium between two loci at BH, the site with the largest sample size, but at no other site. S. jonesi loci do not appear to be in linkage disequilibrium at any seep site. It is possible that linkage disequilibrium does in fact exist in these data, but that the samples sizes and robustness of the test are not sufficient to reveal it. Selection is a second possible explanation for the observed heterozygote deficiency. A similar general deficiency in heterozygotes has been seen in marine bivalves (Zouros & Foltz, 1984). The observed deficiency in heterozygotes has not been successfully explained by the presence of null alleles or population structure (Raymond et al., 1997). One possible explanation for this deficiency is differential selection in sessile adults and planktonic larvae. Heterozygous marine bivalves often demonstrate superior growth capability to homozygous bivalves, and evidence suggests that heterozygous bivalves are more reproductively successful than inbred ones (Launey & Hedgecock, 2001). However, those fitness traits that are advantageous in the adult bivalve may be selectively disadvantageous in planktonic larvae, because heterozygous larvae may have higher metabolic demands and may not survive long dispersal periods as well as inbred larvae (Haag & Garton, 1995). Vestimentiferan larvae are the only means for vestimentiferan dispersal. If heterozygosity affects vestimentiferan metabolism as it does in marine bivalves, fewer heterozygous larvae may survive through dispersal to settlement and growth into adult tubeworms, resulting in a generalized heterozygote deficiency in the adult animals sampled.

Conclusions:

This study represents the first detailed analysis of population structure within seep vestimentiferan species. Given the differences between vent and seep vestimentiferan metabolism and longevity, and between seep and vent habitat characteristics and water currents, seep vestimentiferans could be expected to show very different gene flow patterns than vent vestimentiferans. However, like their vent vestimentiferan cousins, L. luymesi and S. jonesi maintain high levels of geneflow across large distance. L. luymesi and S. jonesi, however, show 38 different degrees of population structure and isolation by distance. The two species co-occur at the same seep sites and in the same aggregations, and therefore likely face the same factors effecting larval dispersal. The observed differences in population structure may reflect differences in the larval biology of the two species, possibly in larval buoyancy or the duration of the larvae in the water column. A consistent feature of both species is the isolation of GC3 from all other sites, suggesting a barrier exists between GC3 and the other sites that affects both L. luymesi and S. jonesi larvae. The generalized heterozygote deficiency observed in these data cannot be satisfactorily explained by the presence of null alleles. Two possible mechanisms for the observed data are: 1) the presence of population structure within aggregations that leads to mating between related individuals, and 2) a selective advantage for homozyogous individuals in the larval stage as is suggested for marine bivalves. Population structure within aggregations can be addressed by a similar study to that presented here by increasing the sample size of individuals from a single aggregation or by sampling very recently settled vestimentiferans. Selective advantage of homozygous larvae could be tested by comparing the heterozygosity of vestimentiferan larvae to that of settled adult vestimentiferans. Studies of the reproduction and dispersal of the vestimentiferans are often indirect due to the technological difficulty in sampling, culturing, and observing deep sea communities. Many aspects of the reproductive cycle, such as the periodicity of spawning, the duration and position of the larvae in the water column, and the existence of settlement cues, are difficult to address and are only partially understood. Results of population genetic data such as presented here can be used to design effective sampling strategies that make more efficient use of the time and resources available for studies of deep sea biology.

Figures:

Figure 5.1: A map of sample locations on the Louisiana Slope of the Gulf of Mexico. Sites range from 2km to 580km apart. Three sample sites are included in one label: BP, GC2, and TA, which are within 5km of each other.

Table 5.1: A list of the site names for each sample site used in this study. Sites are listed from west to east. The table also shows the depth of each site and the geographic region of each site used in pairwise population comparisons. The “L” and “S” columns indicate the number of collections and the total number of animals analyzed for each seep site as “#collection, # animals”.

Table 5.2: Pairwise geographic distances between seep sites, in km.

Figure 5.2: Null allele frequencies estimated by two methods graphed against FIT values for each 39 locus. Diamonds represent null allele frequencies calculated from the frequency of individuals that failed to amplify (FA-null, squares) and frequencies estimated in GENEPOP (GP- null, triangles).

Figure 5.3: A neighbor joining tree for L. luymesi built with pairwise Nei’s Da genetic distances. All bootstrap values are shown (from 10,000 replications)

Figure 5.4: A neighbor joining tree for S. jonesi built with pairwise Nei’s Da genetic distances. All bootstrap values are shown (from 10,000 replications)

Figure 5.5: A test for isolation by distance comparing geographic distance to genetic distance between pairs of samples. a) comparisons of L. luymesi geographic and genetic distances, b) comparisons of S. jonesi geographic and genetic distances.

References:

Aharon, P. (1994). Geology and biology of modern and ancient submarine hydrocarbon seeps and vents: An Introduction. Geo-Marine Letters 14, 69-73.

Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A. & Struhl, K. (1989). Current Protocols in Molecular Biology, John Wiley & Sons, New York.

Balloux, F. & Lugon-Moulin, N. (2002). The estimate of population differentiation with microsatellite markers. Molecular Ecology 11, 155-65.

Barton, N. H. & Slatkin, M. (1986). A quasi-equilibrium theory of the distribution of rare alleles in a subdivided population. Heredity 56(Pt 3), 409-415.

Bergquist, D. C., Urcuyo, I. A. & Fisher, C. R. (2002). Establishment and persistence of seep vestimentiferan aggregations on the upper Louisiana slope of the Gulf of Mexico. Marine Ecology-Progress Series 241, 89-98.

Bergquist, D. C., Williams, F. M. & Fisher, C. R. (2000). Longevity record for deep-sea invertebrate. Nature 403(6769), 499-500.

Brooks, J. M., Wiesenburg, D. A., Roberts, H., Carney, R. S., MacDonald, I. R., Fisher, C. R., Guinasso, N. L. J., Sager, W. W., McDonald, S. J., Burke, R. A. J., Aharon, P. & Bright, T. J. (1990). Salt, seeps and symbiosis in the Gulf of Mexico. EOS 71(45), 1772-1773.

Bucklin, A. (1988). Allozymic variability of Riftia pachyptila populaitons from the Galapogos Rift and 21°N hydrothermal vents. Deep-Sea Res. 35, 1759-1768.

Chevaldonné, P., Jollivet, D., Vangriesheim, A. & Desbruyères, D. (1997). Hydrothermal-vent alvinellid polychaete dispersal in the eastern Pacific .1. Influence of vent site distribution, bottom currents, and biological patterns. Limnol Oceanogr 42(1), 67-80.

Dando, P. R., Southward, A. J., Southward, E. C., Dixon, D. R., Crawford, A. & Crawford, M. (1992). Shipwrecked tube worms. Nature 356, 667.

40 Fisher, C. R., Urcuyo, I. A., Simpkins, M. A. & Nix, E. (1997). Life in the slow lane: Growth and longevity of cold-seep vestimentiferans. Marine Ecology 18, 83-94.

Haag, W. R. & Garton, D. W. (1995). Variation in genotype frequencies during the life history of the bivalve, Dreissena polymorpha. Evolution 49(6), 1284-1288.

Hessler, R. R., Smithey, W. M., Boudrias, M. A., Keller, C. H., Lutz, R. A. & Childress, J. J. (1988). Temporal change in megafauna at the Rose Garden hydrothermal vent. Deep-Sea Res. 35(10/11), 1681-1710. http://oceanexplorer.noaa.gov/explorations/02mexico/background/currents/currents.html Jollivet, D. (1996). Specific and genetic diversity at deep-sea hydrothermal vents: An overview. Biodivers Conserv 5(12), 1619-1653.

Kim, S. L. & Mullineaux, L. S. (1998). Distribution and near-bottom transport of larvae and other plankton at hydrothermal vents. Deep-Sea Research II 45, 423-440.

Kojima, S., Ohta, S., Yamamoto, T., Yamaguchi, T., Miura, T., Fujiwara, Y., Fujikura, K. & Hashimoto, J. (2003). Molecular taxonomy of vestimentiferans of the western Pacific, and their phylogenetic relationship to species of the eastern Pacific III. Alaysia-like vestimentiferans and relationships among families. Marine Biology 142(4), 625-635.

Kumar, S., Tamura, K., Jakobsen, I. B. & Nei, M. (2001). MEGA2: Molecular Evolutionary Genetics Analysis software. Bioinformatics 17, 1244-1245.

Launey, S. & Hedgecock, D. (2001). High genetic load in the pacific oyster Crassostrea gigas. Genetics 159, 255-265.

MacDonald, I. R., Guinasso, N. L., Reilly, J. F., Brooks, J. M., Callender, W. R. & Gabrielle, S. G. (1990). Gulf of Mexico hydrocarbon seep communities: VI. patterns in community structure and habitat. Geo-Mar. Lett. 10, 244-252.

Marsh, A. G., Mullineaux, L. S., Young, C. M. & Manahan, D. T. (2001). Larval dispersal potential of the tubeworm Riftia pachyptila at deep-sea hydrothermal vents. Nature 411(6833), 77-80.

McMullin, E. R. (submitted). Twelve microsatellites for two deep sea polychaete tubeworm species, Lamellibrachia luymesi and Seepiophila jonesi, from the Gulf of Mexico. Molecular Ecology Notes.

McMullin, E. R., Hourdez, S., Schaeffer, S.W., and Fisher, C.R. (2003). Phylogeny and biogeography of deep sea vestimentiferan tubworms and their bacterial symbionts. Symbiosis 34(1), 1-41.

Nei, M. (1972). Genetic distance between populations. 106(949), 283-292.

Nei, M. (1987). Molecular Evolutionary Genetics, Columbia University Press, New York.

Nei, M., Tajima, F. & Tateno, Y. (1983). Accuracy of estimated phylogenetic trees from molecular data. II. Gene frequency data. Journal of Molecular Evolution 19(2), 153-70. 41

Park, S. D. E. (2001). Trypanotolerance in West African cattle and the population genetic effects of selection. Ph.D. thesis, University of Dublin.

Paull, C. K., Hecker, B., Commeau, R., Freeman-Lynde, R. P., Neumann, C., Corso, W. P., Golubic, S., Hook, J. E., Sikes, E. & Curray, J. (1984). Biological communities at the Florida escarpment resemble hydrothermal vent taxa. Science 226(23 nov 1984), 965-967.

Pritchard, J. K., Stephens, M. & Donnelly, P. (2000). Inference of population structure using multilocus genotype data. Genetics 155(2), 945-959.

Raymond, M. & Rousset, F. (2000). GENEPOP On The Web 3.3 edit. http://wbiomed.curtin.edu.au/genepop/.

Raymond, M., Vaanto, R. L., Thomas, F., Rousset, F., deMeeus, T. & Renaud, F. (1997). Heterozygote deficiency in the mussel Mytilus edulis species complex revisited. Marine Ecology-Progress Series 156, 225-237.

Saito, N. & Nei, M. (1987). The neighbor-joining method: a new method for reconstructing phylogenetic trees. Molecular Biology and Evolution 4, 406-425.

Sarrazin, J., Robigou, V., Juniper, S. K. & Delaney, J. R. (1997). Biological and geological dynamics over four years on a high-temperature sulfide structure at the Juan de Fuca Ridge hydrothermal observatory. Marine Ecology Progress Series 153, 5-24.

Schneider, S., Roessli, D., Excoffier, L. (2000). Arlequin: A software for population genetic data. 2.0 edit. Genetics and Biometry Laboratory, University of Geneva, Switzerland. Sibuet, M. & Olu, K. (1998). Biogeography, biodiversity and fluid dependence of deep-sea cold- seep communities at active and passive margins. Deep-Sea Research Part Ii-Topical Studies in Oceanography 45(1-3), 517-+.

Slatkin, M. (1985). Gene flow in natural populations. Annual Review of Ecology and Systematics 16, 393-430.

Slatkin, M. (1995). A measure of population subdivision based on microsatellite allele frequencies. Genetics 139, 457-462.

Smith, C. R., Kukert, H., Wheatcroft, R. A., Jumars, P. A. & Deming, J. W. (1989). Vent fauna on whale remains. Nature 341, 27-28.

Takezaki, N. (1998). POPTREE.

Tunnicliffe, V., McArthur, A. & McHugh, D. (1998). A biogeographical perspective of the deep-sea hydrothermal vent fauna. Advances in Marine Biology 34(3-4), 353-442.

Vrijenhoek, R. C. (1997). Gene flow and genetic diversity in naturally fragmented metapopulations of deep-sea hydrothermal vent animals. Journal of Heredity 88, 285-293.

Wright, S. (1951). The genetical structure of populations. Annals of Eugenics 15, 323-354. 42

Young, C. M., Vazquez, E., Metaxas, A. & Tyler, P. A. (1996). Embryology of vestimentiferan tube worms from deep-sea methane/sulphide seeps. Nature 381(6582), 514-516.

Zouros, E. & Foltz, D. W. (1984). Possible explanations of heterozygote deficiency in bivalve molluscs. Malacologia 25, 583-591.

43

44

Table 5.1

sample Site depth L.luymesi S. jonesi geographic site abbreviation (m) collections collections location

GB535 GB5 600 0 1, 10 Western GB425 GB4 600 2, 18 1, 9

GC354 GC3 540 2, 22 1, 13

Bush Hill BH 560 10, 93 8, 74

TAMU-17 Tamu 540 1, 9 1, 7 Central Brine Pool BP 640 2, 23 2, 17

GC234 GC2 540 4, 34 1, 8 Mississippi Canyon MC 625 1, 11 1, 10 Eastern Viosca Knoll VK 550 3, 25 3, 20

Table 5.2

MC GC2 BP TA BH GC3 GB4 GB5 Vk 205 354 356 357 374 411 476 578 MC 156 159 161 181 214 284 388 GC2 4 6 28 57 128 233 BP 2 26 56 126 230 TA 22 54 124 228 BH 35 104 207 GC3 70 176 GB4 106

45

Figure 5.2

46

47

48

49

Appendix A 50

Figures:

Figure A.1: Number of alleles observed in five variable microsatellite loci for L. luymesi from the LS of the GoM. Number of alleles (Y-axis) is graphed vs. the size of the PCR product (X-axis). Two loci were linked, a minisatellite and a microsatellite region. One primer pair (L2-2Dall) amplified both repetitive regions and a second (L2-2Dinner) amplified only the microsatellite repeat. A figure is shown for L2-2Dall-L2-2Dinner, which shows the variation in the minisatellite repeat.

Figure A.2: Number of alleles observed in seven variable microsatellite loci for S. jonesi from the LS of the GoM. Number of alleles (Y-axis) is graphed vs. the size of the PCR product (X-axis).

Table A.1: Information on all repetitive loci isolated from L. luymesi, including locus designation, primer sequences (if applicable), GenBank Accession number, annealing temperature, type of repeat, status of locus for use as a population marker, and allele size range.

Table A.2: Information on all repetitive loci isolated from S. jonesi, including locus designation, primer sequences (if applicable), GenBank Accession number, annealing temperature, type of repeat, status of locus for use as a population marker, and allele size range (continued on 2nd page).

51

Figure A.1

52 Figure A.2

53 Table A.1 GenBank Annealing Repeat motif in Size Locus Primer sequences (5'-3') Accession Status temp. (°C) clone range no.

L1- TGTTGAGGACGTGTTACAAG overlapping AY263756 62 (GT)30 200 2E#1 size classes ACGTGCATGTGTTTACTAGC

L1- GGCAATTGTTGAGGACGTGT 339- AY263756 62 (CAC/AG)12 scored 2E#2 369 GGAAGTGAACCAATGCTCTG

(TACAG), repetitive L1-2F N/A AY263743 N/A N/A N/A flanking regions

GTAAACCGATATACCCACCT (CA)15/(ATCAT/CGT 342- L2-2D AY263755 62 scored ATT)5 396 TGTCTCTTCTTTACAGCTGC

GTAAGCTGTAATCTCCGTCG L2-3A AY263744 55 minisatellite monomorphic 400 ACATCGACAGCGTAAGTTCT

GATCCAGTATACGCCAGGCA L3-7B AY263747 55 (GT)30 monomorphic 400 GCGGCTGTGTTCAACTTGTG

GGTTGTCTGTGGAACATGGA single nucleotide L3-7H AY263748 60 monomorphic 250 repeat CAGCAGAATGAGGTGTGGAT

CTGCAGCAAAGGGAAGGACA repetitive but no L3-11A AY263746 60 monmorphic 360 microsatellite AGATGACGTTATCGGCGTGG

GGCAATAATCCGCTAGATGG 219- L4-1E AY263750 62 (CA)10/ (TAAA)3 scored 257 CGTAGCACGTATCGTTCTCA

TCGGCTCGACACTCCATGTA too many L4-2B AY263751 60 (CACG)12/(CA)8 450 bands GCGAAGGTCAGTGGTTAAGG

TCTGAGCGGTGAACTGTATC 320- L4-5E AY263752 62 (CAAA)12/(CAGA)10 scored 449 GATAAGTTCTCGTCGCTCGT

L4-8A, ACTGGGCCATAACGTGTTTC doesn't amplify AY263745 62 (CAA)18 340 L2-3B well -smear GCGTGGGACAAACAACTCTA

GGGTTGAGAATAACGAAAGC smeary L4-7B AY263753 60 (TGTC)8 180 product TCAGCAAGGAGTCGTTAGTA repetitive flanking L4-10C N/A AY263749 N/A N/A N/A regions

CAGTTGACGGCATCACCAGT overlapping L5-1B AY263754 62 (GA)20 400 size classes ACGCTGGAAGTGCACGATGT

ATAAGATGCGACTTCGATGC 245- L5-2H AY265354 62 (CA)13 / (CACG)3 scored 287 CTCTACATGAACAAGTTTGC

L5-8G N/A AY263742 N/A minisatellite N/A N/A

54 Table A.2 GenBank Annealing Size Primer sequences (5'-3') Accessio Repeat motif in clone Status Locus temp. (°C) range n no.

CGTACAATGCAGGTGAACGA too big to 350, E1-5F AY263765 55 (CACACG)10/(CA)13 score 700 TGCAATCCCGATGTGGTACG

(CGT/GT)35 no flanking E1-7H N/A AY263767 N/A N/A N/A region

CTTGGCTAGGGTAGATGTGA E1-8A AY263768 54 (GCGT(T))10 monomorphic 220 CGTACACAACTCCCGCAACT

G/A repetitive region E1-8E N/A AY263769 N/A N/A N/A through cut site

GATCGGGTAACCGCGATAGT E1-9E AY263770 64 (GT)37 smeary 350 GCCACGTCTCTTCTCGAAGT

AGAAGACCGCGTCTCGATAG E1- overlapping AY263764 60 (CT)21 400 10A#1 size classes GGAATCTCGCCATGTATCGT

GTCTTCTGATGTCATGGTCC E1- 128- AY263764 60 (CACG)9/(CA)13 scored 10A#2 234 CGAACACCTCGACAATCAAC

CAGTGTCATACAGCGGGCTT 127- E2-2A AY263772 55 (TGCGGGCG)4 scored 179 ATCCGCCTCTGCCTACGTTT

TACACGCGAACAGACTAGAG E2-7E AY263773 55 (CA)34 stutters 400 TATACTACCCTGTGACCACC

C/T repetitive region E2-7G N/A AY263774 N/A N/A N/A through cut site

GTACAATCTCACCTGCGTGC E2-9E AY263775 55 (GTT)8 monomorphic 350 GCATGTCGCTCCGTATGTAG

(GG/AAT)17, GC poor E2-10B N/A AY263771 N/A N/A N/A flanking region

E3-1E N/A AY263776 N/A CA repeat N/A N/A

55 Table A.2 cont. GenBank Annealing Repeat motif in Size Locus Primer sequences (5'-3') Accession Status temp. (°C) clone range no.

GATCTCTTCTGGCAGGCGTT 128- E3-2D AY263777 60 (CGTG)9 scored 195 CTCCATTGCATCTACGGGTC

GTACGGTGCAACCTATGTAT variable, too E3-9G AY263778 55 (CA)25 600 large GACAATGGCGTTTACATTCG

AGGTCAGAGGCATTGCCATA E3-10A AY263780 55 (TGTC)8 scored 305 ACAACGGACAGGTCTGCATT

TCATTGCTCTCCTGGTTTAG 271- E4-7G AY263779 60 (GT)19 scored 316 GCCTGGTTTCTGATGACTTA

GATCGCTCCACTACTCACACAA overlapping E4-10D AY263759 55 (GT)20 size classes CTGGACAAGGATGAGAGCAACA

TAACACTGACCACGTCACTG ECL1- 260- AY263761 60 (GT)22 scored 5B 395 GTCGCGTAGACCTAGCAATA

ECL1- N/A AY263757 N/A minisatellite N/A N/A 7D

ECL2- G/A repetitive region N/A AY263758 N/A N/A N/A 8B through cut site

GGAGCCTCCTTGACTTTACA ECL2- 194- 55 (CTGT)7 scored 12C 224 CAAAGCTCATCGCACACTTG

GTACAATCTCACCTGCGTGC ECL3- AY263760 55 (GTT)8 smeary 230 4C GCATGTCGCTCCGTATGTAG

ECL3- repetitive but no N/A AY263762 N/A N/A N/A 7E microsatellite

GCGTTGCTAACTGCCAAGTG ECL3- 323- AY263763 57 (TGCG)9 scored 8D 563 ATCTTGACCAGTCGCTGACC

56

Appendix B

57 Figure:

Table B.1: Support information on samples sites used in Chapter 5. Sample sites are listed by site name, with collection designations from Chapter 5 of this thesis, as well as the collection label for lab notes and frozen samples, and the year of the collection. Information is also provided for identification of these same samples in three chapters of D.C. Bergquist’s Ph.D thesis (2001). A rough estimate is also provided where known for the size of tubeworms in each collection (from dive notes).

58 Table B.1

Collection Collection Bergquist Thesis 2001 Rough Sample Year name for name in my size of Site Collected publication notes Chap 3 Chap 4 Chap 5 animals

GB535 GB5 GB54241 2000 ------<70cm

GB4-1 GB4041 1998 ------N/A GB425 GB4-2 GB4053 1998 ------N/A

GC3-4 GC33347 2002 ------>50cm GC354 GC3-2 GC34237 2000 ------<70cm GC3-3 GC34238 2000 ------>50cm

BH-1 BH2853AT5 1997 BH-4 ---- BH-5 >70cm BH-2 BH2857TT6 1997 BH-6 ---- BH-7 >70cm BH-3 BH2865 1997 ------<70cm BH-11 GC2868 1997 ------>70cm BH-4 BH2873 1997 ------<70cm Bush BH-5 BH2874JT5 1997 BH-5? ------N/A Hill BH-12 BH2874AT5 1997 BH-5? ---- BH-6 <50cm BH-6 BH2892JT1 1997 ------<20cm BH-7 BH2892JT2 1997 ------<70cm BH-8 BH4028JT1 1998 BH-3 BH-J1 BH-1 <20cm BH-9 BH4035AT1 1998 BH-7 BH-A1 BH-3 N/A BH-10 BH4046JT2 1998 BH-1 BHJ2 BH-2 <50cm

TAMU TA TAMU 1998 ------<20cm

Brine BP-1 BP2850 1997 ------<50cm Pool BP-2 BP4384 1998 ------N/A

GC2-1 GC2866NM1 1997 GC-6 ------<20cm GC2-2 GC2866NM2 1997 GC-7 ------<20cm GC234 GC2-3 GC4033center 1998 GC-3 GC-A1 GC-3 N/A GC2-4 GC4033edge 1998 ------N/A

MS MC MC3339 2002 ------N/A Canyon

VK-2 VK3337 2002 ------N/A Viosca VK-4 VK3355 2002 ------<70cm Knoll VK-1 VK4301 2001 ------N/A VK-3 VK4303 2001 ------N/A

Erin McMullin Curiculum Vitae

Education: Pennsylvania State University: PhD in Biology, August 2003 Thesis: Molecular evolution and dispersal pathways of vestimentiferan tube worms from the hydrocarbon seeps in the northern Gulf of Mexico. Advisors: Dr. Charles Fisher and Dr. Stephen Schaeffer Oberlin College: B.A. in Biology, 1993

Publications: In prep. E.R. McMullin, S.W. Schaeffer, C.R. Fisher. “Molecular population structure of two vestimentiferan tubeworms, Lamellibrachia luymesi and Seepiophila jonesi, from the hydrocarbon seep communities in the Gulf of Mexico.” In preparation for Molecular Ecology. submitted E.R. McMullin. “Variable microsatellite markers for two hydrocarbon seep vestimentiferan tubeworm species, Lamellibrachia luymesi and Seepiophila jonesi, from the northern Gulf of Mexico.” for Molecular Ecology Notes. 2003 E.R. McMullin, S. Hourdez, S.W. Schaeffer, C.R. Fisher. In press. “Phylogenetics and biogeography of deep sea vestimentiferan tubeworms and their bacterial endosymbionts” Symbiosis, 34 (1): 1-41. 2001 S.L. Gardiner, E.R. McMullin, and C.R. Fisher. “Seepiophila jonesi: a new genus and species of vestimentiferan tube worm (Annelida: Pogonophora) from hydrocarbon seep communities in the Gulf of Mexico.” Proceedings of the Biological Society of Washington. 114: 694-707. 2000 C.R. Fisher, I.R MacDonald, R. Sassen, C.M. Young, S.A. Macko, S. Hourdez, R.S. Carney, S. Joye, and E. McMullin “Methane ice worms: Hesiocaeca methanicola colonizing fossil fuel reserves.” Naturwissenschaften, 87(4):184-187. 2000 E.R. McMullin, D.C. Bergquist, and C.R. Fisher. “Metazoans in extreme environments: Adaptations of hydrothermal vent and hydrocarbon seep fauna.” Gravitational and Space Biology Bulletin. 13(2):13-23. 1996 C.D. Woodworth, J. Chung, E. McMullin, G.D. Plowman, S. Simpson, and M. Iglesias. “Transforming growth factor beta 1 supports autonomous growth of human papillomavirus-immortalized cervical keratinocytes under conditions promoting squamous differentiation.” Cell Growth and Differentiation, 7:(6)811-820. 1995 C.D. Woodworth, E. McMullin, M. Iglesias, and G.D. Plowman, “Interleukin-1 and Tumor Necrosis Factor-a stimulate autocrine amphiregulin expression and proliferation of human papillomavirus- immortalized and carcinoma-derived cervical epithelial cells.” PNAS, 92:(7)2840-2844.

Meetings, Posters and Presentations: Mar. 2003: McMullin, E.R. Presentation: “Migration and genetic structure of two deep sea vestimentiferans from the Gulf of Mexico revealed using variable microsatellite loci.” IMEG (Institute of Molecular Evolutionary Genetics) seminar, Penn State University, Pennsylvania. Oct. 2001: McMullin, E.R., J.Wood, T.W. Waltz, S.W. Schaeffer, and C.R. Fisher. Presentation: “Estimated migration rates of two Gulf of Mexico hydrocarbon seep tube worm species using microsatellite markers.” 2nd International Symposium on Deep-sea Hydrothermal Vent Biology, October 8-12, 2001,Brest, France. May 2001: McMullin, E.R., J. Wood, S.W. Schaeffer, and C.R. Fisher. Poster: “Cold seep vestimentiferans: molecular phylogeny and population genetics” LARVE meeting, Ogden, Utah. Mar. 2001: McMullin, E.R., J. Wood, S.W. Schaeffer, C.R. Fisher. Poster: “Larval dispersal in deepsea tubeworms from Gulf of Mexico oil seeps” Penn State Graduate Research Exhibition, University Park Pennsylvania. Jan. 2001:McMullin, E.R. Presentation: “Molecular Phylogeny and population genetics of Gulf of Mexico deep sea tube worms and their symbionts.” IMEG (Institute of Molecular Evolutionary Genetics) seminar, Penn State University, Pennsylvania. Oct. 2000: Invited talk: “Life in Extreme Environments” New Jersey Spacegrant, NJ Teachers Symposium, Somerset, NJ.