COMPARATIVE EVOLUTION AND DE VELOPMENT OF THE BUTTERFLY EYESPOT AND A SURVEY OF WING PATTERN GENE EXPRESSION ACROSS BUTTERFLY FAMILIES By MICHAEL W. PERRY A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2006
Copyright 2006 by Michael W. Perry
This document is dedicated to my parents, fo r instilling in me a l ove of biology and the natural world.
iv ACKNOWLEDGMENTS First and foremost, I thank my advisors. E ach has played a very special and critical role. I have worked for and with Dr. Thomas Emmel for over five years. He has provided numerous sometimes character-buildin g opportunities, for travel to foreign and exciting places for research and for meetings . Opportunities he has provided have played a key role in my decision to begin gra duate school in biology, and I cannot thank him enough for all he has done. I thank Dr. Charles Baer for overseeing my academic and research development, for acting as a soundingboard for ideas, for numerous fascinating discussions about aspects of biology in semina rs, lab meetings, and in discussion, and for at times cracking the whip when it was needed most. I thank my committee member Dr. Martin Cohn for an introduction to developmental biology, both academic and practical , and for his always valuable advice. I hope to someday make a go of it in evo-de vo, and he has given me a great introduction. I thank all the numerous faculty I met with, often at random, about ideas or concepts, especially Drs. Marta Wayne, Steve Phelps, David Reed, Ben Bolker, Rebecca Kimball, and Gustav Paulay. Faculty, students, and staff at the McGuire Center were of cons tant and consistent help. I thank Christine Eliaza r for making all things possible. I thank Dr. Jaret Daniels for all of his help in so many ways, from a dvice and help in reari ng larvae, to receiving shipments of plants from foreign countries, to always being there when I needed advice or expertise. I thank Dr. L ee Miller, and especially Dr. J ackie Miller, for sharing their
v vast knowledge. Thanks go to Dr. George Aust in Â– for being a great boss, and one of the hardest working people I have ever met. George is always a source of humor and knowledge. I thank Dr. Andrei Sourakov, alwa ys ready with help, useful advice, and incredible stories. I now alwa ys carry a bottle of vodka into the field with me in case I need to make friends in a hurry. A thank you goes to the entire Baer Lab Â– especially to Naomi Phillips and Andy Custer Â– for teachi ng me much of what I know about finding my way in the lab. I thank my fellow zool ogy graduate students, who are unfortunately too many to list Â– they are great. Many of you have helped me in important ways. Thanks go to my former roommates, especi ally to Chris and Mike Gmuer, for help with photoshop and all things computer related; for attempts at sta ndardizing a means of laser-cauterizing larvae with a professional la ser engraver/autocad system; and for putting up with years of larvae and bugs in the apartment. I thank Dr. Paul Brakefield and Dr. Antonia Montiero, for letting me visit their labs and for teaching me about butterfly wing patte rn. If it were not for my time learning technique in AntoniaÂ’s lab I would never ha ve been able to complete the research described here. I thank Dr. Da niel Janzen for assistance and advice in getting permits for collection and export in and fr om Guanacaste, Costa Rica, as well as about butterflies and navigation there. I thank Roger Blanco a nd Maria Marta Chevarria for permit assistance in the ACG. I thank Doctor J.D. Turner for samples, discussion about, and time collecting butterflies of the Riodinidae, and Dr. Jason Hall for discussion concerning the same. Thanks go to Dr. Nipam Patel and R on Parchem for discussion of techniques and of questions, and for two great visits to CA. I look forward to much future work with
vi them. I thank Ernesto Rodriguez and John Fazinni for making my time in Costa Rica possible and for hosting me. I could not have done it without them. I thank all who have sent me antibodies fo r this research, Drs. Sean Carroll, Rosa Barrio, and Nipam Patel. I thank the Deve lopmental Studies Hybridoma Bank of Iowa for the rest of the antibodies used. Finally, I thank all my friends and family, wherever they might be. Most importantly, I thank my Mom and Dad for suppor ting me in life and in my research. I thank them all.
vii TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF FIGURES.........................................................................................................viii ABSTRACT....................................................................................................................... ix CHAPTER 1 A BACKGROUND AND OVERVIEW OF BUTTERFLY WING PATTERN.........1 Overview....................................................................................................................... 1 Background and Overview of Butterfly Wing Pattern.................................................2 2 A COMPARISON OF HOMOLOGY AND PATTERNING MECHANISM Â– THE BUTTERFLY EYESPOT IN THE RIODINIDAE...........................................10 The Riodinid Eyespot.................................................................................................10 Results and Discussion...............................................................................................16 Larval Â“EyespotsÂ” and Some Initial Work.................................................................18 A Note on Phenotypic Complexity and th e Evolution of the Eyespot: an Ongoing Project....................................................................................................................21 3 A PHYLOGENETIC SURVEY OF PATTERNING GENE EXPRESSION............23 Patterning Gene Expression........................................................................................23 Expression of Sal in Swallo wtail Marginal Crescents................................................25 Possible Function of Sal Expressi on at the Future Wing Margin...............................28 Sal Expression Patterns Suggest a New Hypothesis...................................................30 Materials and Methods...............................................................................................37 Conclusion..................................................................................................................38 LIST OF REFERENCES...................................................................................................40 BIOGRAPHICAL SKETCH.............................................................................................44
viii LIST OF FIGURES Figure page 1. Examples of Â“eyespotÂ”-like patches pigment................................................................11 2. Examples of the riodinid eyespot compared to the satyrid eyespots.............................12 3. Mesosemia species.........................................................................................................15 4. Expression patterns in the Mesosemiini........................................................................17 5. Mesosemia sp. showing evidence that each white central spot may be an organizer....19 6. Close up of Euptychia rubrofasciata as an example of pattern when organizers are close to one another in the Satyridae........................................................................19 7. Neonympha areolata , as an example of a scattered or separated organizer..................19 8. Swallowtail larval ocelli: in this case, from Papilio palamedes ....................................20 9. Sal expression in Papilio pol yxenes, 5th instar larva.....................................................28 10. Swallowtail Sal expression..........................................................................................27 11. Growing trachea initially ove rshoot the future wing ma rgin; secondary branches grow in later to form the border lacunae..................................................................29 12. Diagram of the proposed mechanism..........................................................................31 13. Wing pattern of Cethosia biblus ..................................................................................33 14. Expression of rho-1 in Papilio cresphontes 5th instar larva.........................................36
ix Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science COMPARATIVE EVOLUTION AND DE VELOPMENT OF THE BUTTERFLY EYESPOT AND A SURVEY OF WING PATTERN GENE EXPRESSION ACROSS BUTTERFLY FAMILIES By Michael W. Perry August 2006 Chair: Charles F. Baer Cochair: Thomas C. Emmel Major Department: Zoology Recent growth in the field of evolutiona ry developmental biology, or evo-devo, has brought about a change in research emphasis from the sorting of phenotypic variation by selection, to the generation of that variation throu gh ontogeny. Research in this field has mainly focused on large evoluti onary changes that lead to sh ifts in animal body plan. Another fundamental aspect of evolution is the generation of morphological variants within a species. Selection can act on ly on these variants that are produced by development. An understanding of how deve lopmental mechanisms that lead to the production of variation arise, and change across groups and over time, is thus of fundamental importance to better understand evolution. One of the few examples in which we understand something of this process at multiple levels is in the butterfly eyespot. Studies of butterfly wing patterns have played an integral part in the development of evo-de vo as a field of research. Butterfly eyespots
x have clear adaptive value as a form of vi sual communication, and their development has been amenable to detailed characterization. Current research is aimed at further unraveling the genetic basis of eyespot deve lopment. We still know little about how often such developmental mechanisms have come to be used in pattern formation, or about how much these mechanisms change over time and across groups. Research described in Chapter 2 of this thesis attempte d to characterize differences or similarities in developmental mechanisms between two eyespot-containing groups . Data presented suggest that developmental mechanisms may not be homologous between the model butterfly eyespot and phenotypica lly-similar patterns in the butterfly family Riodinidae. A selective phylogenetic analysis of characte r evolution across the group may yet reveal if character complexity arose over time from a simpler spotting pattern. This test of the conservation of patterning mech anisms across taxa suggests th at the eyespot patterning mechanisms in the Nymphalidae arose independently, and perhaps only once. A survey of eyespot-patterning gene e xpression was conducted in an effort to determine what role these genes have played across multiple butterfly families. Data collected suggest that most wing patterns develop without the us e of the nymphalid eyespot patterning genes. One instance of e xpression was particularly interesting, and led me to propose new hypotheses about how the satyrid patterning mechanisms came to play a novel role in determining spatial expre ssion of pigment on the butterfly wing. It is possible that the eyespot patterning mechanism was directly co-opted from its ancestral function at the wing margin, after becoming coupled with pigmentation. This study is unique in the scale at which it addresses thes e questions, and it serves as an example of the direction studies of this system could begin to take.
1 CHAPTER 1 A BACKGROUND AND OVERVIEW OF BUTTERFLY WING PATTERN Overview Butterfly families are incredibly diverse, containing over 17,000 species, nearly all of which can be distinguished by examin ing differences in th e wing pattern (Nijhout 1991). These patterns often have known ad aptive value and function in both biotic interactions such as competition and pred ation (Vane-Wright and BopprÃ© 1993), and with abiotic variables in the natural environment, such as in thermoregulation (Kingsolver 1995). The wings of butterflies are their mo st prominent feature and their patterns are strongly shaped by natural and sexual selecti on. The selective forces at work are often identifiable, and they are critical variables in determining how evol ution interacts with developmental processes. Recent research has begun to dissect these interactions between environment, development, phenotype, and evolution in a model species, Bicyclus anynana , in the family Nymphalidae. Chapter 1 of this thesis contains a brief overview and review of what is known about the evolution and development of butterfly wing pattern, and provides context for the rese arch presented in Chapters 2 and 3. The butterfly wing pattern element that we know the most about is the butterfly eyespot. Recent research on this system ha s come to play a prominent role in the integration of evolutionary and developmenta l biology. Continuing research is aimed at further unraveling the genetic basis of eyespot development. Selection experiments have begun to show how evolution acts on that genetic basis over multiple generations. However, we know little about how often such developmental mechanisms have come to
2 be used in pattern formation, or about how conserved these mechanisms are. Work described in Chapter 2 of this thesis was aimed at testing the c onservation and homology of butterfly eyespot patterning mechanisms across butterfly families. It is important to know how these intera ctions relate to processes within and across larger taxonomic groups. An increasing amount of effort and resources have been devoted to characterizing genetic and deve lopmental processes in the model organism, but little is known about these processes outside the model species. We still know little about evolutionary and developmental mech anisms behind the formation of pattern elements, other than in the butterfly eyespot in B. anynana . We know little of how patterning genes that are now known to function in patterning the eyespot came to play a novel role in patterning pigm entation on the wing. Chapter 3 involved work aimed at surveying gene expression in a phylogenetic co ntext. The focus was on determining the phylogenetic origins of the use of the eyespot patterning genes in but terfly wing pattern. Data presented in Chapter 3 suggest a some what surprising hypothesis about the origins of these patterning genes in the development of wing pattern. Background and Overview of Butterfly Wing Pattern The patterns on butterfly wings are idea l for studying the interactions between evolution and development. A variety of factors make butte rfly wings nearly ideal for the study of the interactions between evolutionary and de velopmental processes that shape morphological variation, both within and among species. They represent the products of selection, and of ten have known adaptive value (for example, they may function in sexual selection, pred ator avoidance, or in intr a-specific recognition). The size and relative simplicity of butterfly wings make them attractive from a developmental perspective. For example, instead of threedimensional structures such as bristles or
3 appendages, the butterfly wing pattern is ma de up of spatially-arranged scales on a twodimensional surface. Each scale, which is a modified sensory bristle (Galant et al. 1998), is a single color, determined during devel opment by biochemical pathways that deposit pigment (exe. Koch et al. 1998). These patterns have proven amenable to a detailed developmental characterizati on, classically thro ugh surgical manipulations of pupal wings ( e.g. , Nijhout 1980, French an d Brakefield 1995), and more recently through the study of gene expression patterns ( e.g. , Brakefield et al. 1996, Keys et al. 1999, Brunetti et al. 2001, Reed and Serfas 2004). The pattern element we know the most a bout is the butterfly eyespot, a common pattern element composed of concentric rings of different colors. Research has been focused on the eyespots of th e tropical African butterfly Bicyclus anynana (discussed in Brakefield 1998). In this system, it has been shown that there is additive genetic variance for different characteristics of the eyespot such as size, color proportion, and shape (Monteiro et al. 1994, 1997a, 1997b). The inheri tance of many of these features is quantitative in nature, but the actual genes that contribute to much of the variation are unknown. One of the first genes to be quantita tively examined was Distal-less (Dll), and variation in this gene was shown to correlate with eyespot size (Beldade et al. 2002). Butterfly eyespot patterns are appropriat e for an integrated approach to the study of evolutionary developmental biology. The size and relative simplicity of butterfly wings allow them to be studied at multiple levels, from physiology and gene expression to population genetic an alysis (examples: Brakefield et al . 1998, Carroll et al. 1994, Holloway et al. 1993). The generation time of ma ny butterflies is ad equately short for selection experiments, and large populati on sizes can be maintained in laboratory
4 settings. Butterflies are larg e enough to be developmentally manipulable, for techniques such as hemolymph extraction, or for excisi on and transplantation of developing wing tissue, which can help reveal developmental pr ocesses that give rise to pattern (Nijhout 1991). Studies of the ecologi cal relevance of pattern, c oupled with its underlying developmental basis, allow for integrated st udies of morphological evolution. Selection acts to complete the integr ative feedback loop among genes, development, form, and function, and can be both strong and vari able, between and within species. A wide variety of experimental approach es have been used in these studies. Directional selection, combin ed with surgical manipulation, has been useful in differentiating the effects of genes and environment during development (Monteiro et al. 1994). In the last decade researchers have begun to identify the molecules and genes involved in pattern formation. Specific genes that play a ro le in development have been identified by removing wing primordia and staining with labeled antibodies against known pattering genes from D. melanogaster (Carroll et al. 1994). Researchers in the laboratory of Anthony Long, at the University of California at Irvine, are developing genomic resources such as a high density linka ge map composed of genes, which is then being used to identify quantitative trait loci (QTL) that contribute to eyespot size. Those products, in combination with large-scale seque ncing efforts, should begin to more fully unravel specific developmental pathways in the model species. There are few systems that can be studied at so many levels, using such a wide variety of experimental appr oaches to further our knowledge of morphological character evolution. The type of integr ated approach that is possibl e with this system of wing
5 pattern evolution has and will continue to grea tly contribute to the field of evolutionary developmental biology. Modularity and flexibility of pattern. It has been proposed th at the incredible diversification in butterfly wing patterns has been possible owing to developmental flexibility in the organization of pattern el ements (Nijhout 2001). In comparisons across species, it was observed that different types of element are independent, but that there are correlations across homologous elements in a series (Montiero et al. 1994, 1997a). This finding suggested that pattern elements in a series might make up a module or unit upon which evolution and development operate i ndependently from other modules, but in which traits tend to change in a definite manner (Beldade et al. 2002, Brakefield 2001). Elements from the same symmetry system can become developmentally Â“uncoupledÂ” and follow different evolutionary trajectories. One example is found in Precis coenia , where genetic correlation among eyes pots with different morphologies is lower than that observed among morphologi cally similar eyespots (Paulsen 1994). Independence could be achieved either by small gradual genetic changes or through major switches in developmental modularity that appreciably aff ect pattern in a subset of wing cells. This might occur through changes in gene regulation (Ca rroll et al. 2001). Furthermore, components within some of the more complex pattern elements show significant independence. One example is in B. anynana , where selection for size produces minor changes in color compositi on, and selection for changes in color composition has only minor effects on eyespot size (Monteiro et al. 1997a). In another example, also using B. anynana , artificial selection was us ed to quantify levels of developmental constraint on eyespot size when divergent eyespot size is selected for on
6 the same wing (Beldade et al. 2002). The system was shown to be remarkably flexible. This flexibility seems to be the result of compartmentalization of pattern elements in individual wing cells; it might result from the lack of physical communication between them and/or from the specific genetic composition of wing cells, which may regulate eyespot-forming genes (McMillan et al. 2002). To date, neither mechanism has experimental support. The nymphalid ground plan: from pattern to mechanism. Most of the more than 17,000 species of butterfly can be id entified by wing patt ern and coloration. Patterns can be variable within spec ies through genetic polymorphism, seasonal polyphenism, and sexual dimorphism, and most sp ecies have different (at least somewhat independent) patterns dorsally and ventrally (N ijhout 1991). Early in the last century, it was recognized that even radically different wing patterns are made up of relatively few shared pattern elements that are common acr oss groups. Two researchers independently proposed a similar system of wing pattern homologies, which were later expanded and popularized by the works of Nij hout (Nijhout 1991). This system came to be called the Â“nymphalid ground planÂ” and consisted of a sequence of bands and pattern elements (such as chevrons, bands, and eyespots) that belong to distinct symmetry systems, which are repeated in each wing cell. Each wi ng cell is a compartment outlined by the wing veins. Relatively large changes in the num ber, shape, location, si ze, and color of the individual elements were thought to explain the derivation of different wing patterns from this basic ground plan (Nijhout 1991). The ground plan was a framework in whic h homology could be identified across diverse groups in many species of Lepidoptera. It was used primarily as a means of
7 comparison, but began to provide a potential description of how the elements might be produced during development. The ground plan has been tested for actual pattern homology in very few species. Evidence agai nst this system of homology has been found in families and genera such as the swallowtails, Papilio, and longwings, Heliconius , which have patterns substantially differen t from the archetypal ground plan (Mallet 1991). At first, much literature was devoted to relatively extreme modification of the plan to accommodate these groups. Its lo w explanatory power for pattern evolution across these groups suggested that perhaps nove l patterning elements are involved, or that entirely different mechanisms are responsible for pattern. More recently, a variety of findings in the developmental biology of wing pattern has begun to provide evidence that a simple system of homology, combined w ith a single underlying process of pattern determination, might be overly simplistic. Perhaps partly because of these problems, the nymphalid ground plan has unfortunately seen little direct application by either those interested in pattern formation or by those interested in butterfly systema tics and taxonomy. Some of the general ideas involved were certainly valuab le, yet since that body of work, there have been few pattern development or evolution studies fo cused on any taxonomic level above that of closely related species. A 2001 study by Montei ro and Pierce developed a phylogeny of the Bicyclus genus, but unfortunately this phylogeny has not yet been used in any detailed analysis of pattern evolution across the very genus that we should know the most about (in terms of pattern evolution and developmen t). Current efforts remain focused on the goal of more fully understanding th e genetic basis of pattern in a single species. It is clear that it would be useful to learn something about the level of conservation of these
8 mechanisms in other taxonomic groups, bot h of the eyespot and of developmental mechanisms leading to production of other types of pattern elements. A variety of genes have been linked to pattern formation. The logical place to look for wing pattern genes homologies was in Drosophila . Researchers surveyed gene expression in butterfly larval and pupal wing primordia using i mmunological techniques and soon found that an entire suite of regul atory genes has been co-opted for use in pattern formation. Carroll and co -workers demonstrated that a Precis homologue of Distal-less , a Drosophila regulatory gene, is expressed in the cells that are fated to become the center of eyespots (Carroll et al. 1994). Homologues of conserved regulatory genes and even whole regulatory pathways are co-opted in butterfly wing pattern formation. It has since been suggested that developmental systems have largely evolved through changes in regulation of conser ved sets of patterning genes (Carroll et al . 2001). The long-range diffusible morphogens wingless and decapentaplegic are expressed in the forewing of Precis, in areas corresponding to future bands of the central symmetry system; these areas could be involve d in positioning eyespots (Carroll et al. 1994). Genes from the hedgehog signaling pathway are deployed in and around the eyespot foci in Precis (Keys et al . 1999). They may also play a functional role in eyespot development. These and others genes have unknown roles in pattern development. Many of these genes are transcri ption factors, and it has become important to test what functional role they play in determining phenotype. Researchers have just begun to identify genes that are involved in the wingpatterning process. Some of these genes ha ve been identified and the function of the homologues determined in greater detail in model systems such as Drosophila . In
9 Lepidoptera, we know little more than that th e function of these genes has been modified and co-opted into patterning pigmentation on the wing. Little is known about how they came to function in this newly-derived capac ity, or how they became coupled with genes that regulate pigment production. Even less is known about how evolution acts on these processes as a whole, and we are only beginni ng to figure out how selective forces act on the feedback network between genes, deve lopment, form, and function. Few of the putative morphogens involved in complex wi ng pattern element development have been identified and characterized. It seems likel y that at least several such morphogens are involved during different stages of developm ent in the wing, which first act to set up compartments, and then function to determin e shape and size of pattern elements in individual compartments. There is still mu ch that is unknown about even the eyespot, and even less well-studied are developmental mechanisms that determine all other pattern elements found on butterfly wings. Since little was known about what might be responsible for other types of pattern, I began my thesis work in this area by addressing questions of homology of the ey espot character itself, across different butterfly families.
10 CHAPTER 2 A COMPARISON OF HOM OLOGY AND PATTERNING MECHANISM â€“ THE BUTTERFLY EYESPOT IN THE RIODINIDAE. The Riodinid Eyespot The butterfly eyespot in wi ng pattern that has been prev iously studied is found in several species in the family Nymphalidae, and within the subgroup Satyridae. Almost all eyespots that share a complexity of patte rn including at least 3 concentric rings of color (white center, black ri ng, then yellow/gold ring) are found in this taxonomic group. Other potential attention-drawi ng patches of pigment, such as those found near tails in the Lycaenidae or in the Papilionidae (Figure 1), are sometimes referred to as â€œeyespotsâ€ but share much less direct phenotypic homology. Differences in placement on the wing, lack of concentric rings around an organizer, a nd other characters are a few ways in which they differ from the satyrid eyespot. Pa tterning gene expression homology was also examined in these families, and found to be unlikely (discussed in Chapter 3). One particular group, however, has an elaborate eyespot that bears striking resemblance to the satyrid eyespot (see Figure 1) . In at least some species, all three rings of color are present, and some have even mo re repeating concentric rings around what is potentially a focal organizer (evidence of an organizer is discussed below). These eyespot characters are found with in the butterfly family Riodinidae, and almost entirely within the â€œTribeâ€ Mesosemiini. This group of butterflies is relatively distinct and phylogenetically distant from the Satyridae, yet has a phenotypically similar character. Numerous intervening taxa have no eyespots, making it unlikely th at the most recent
11 common ancestor of the two groups had an eyes pot (for a phylogeny, see Wahlberg et al. 2006). A B Figure 1. Examples of Â“eyespotÂ”-like patches of colorful pigment in A) the Lycaenidae; Strymon acis and B) the Papilionidae; on the hindwing of Papilio polyxenes . These are thought to have a function sim ilar to the satyrid eyespot, in drawing the attention of predators away from the main body. These Â“eyespotsÂ” share much less visual similarity with eyes than do satyrid eyespots. Other characteristics of the riodinid eyespot make it seem likely that this eyespot evolved at least somewhat independently of the satyrid eyespot. Placement on the wing is in the central band of symmetry, in th e discal cell of the wing, as opposed to the marginal band of symmetry. Serial repetition of the element is much less common, although observed in some species in as many as five repeats, all within the central band of symmetry. A very few species also have a marginal eyespot, which may or may not be homologous in terms of devel opment to the discal eyespot and/or the Satyrid marginal eyespot. Notably, almost all of the eyespots in the Mesosemiini have at least two and usually three central white spots, instead of th e single centered white spot in most satyrid eyespots (Figure 2). There is evidence that each of these three spots may be an organizer (discussed below). It is somewhat unusual that the eyespot in this group is in the central
12 location; one assumption about th is type of pattern element is that it functions to not only startle predators, but that it al so might deflect predator attack away from the body Â– a task to which it seems less well suited when the eyespot is in center of the wing. A B C Figure 2. Examples of the riodin id eyespot compared to the satyrid eyespots: A) eyespots on various Mesosemia species. B) Opposing eyespots on a Mesosemia . B) some satyrid eyespots, on Bicyclus taenias . In an examination of patterning mechanis m and homology of the character, there were several specific questions to be addressed. In the model nymphalid system, it has been shown that a long-range diffusible mo rphogen (not yet characterized) acts as an
13 intercellular signal and is invol ved in pattern development. Specific transcription factors have been shown to respond to this signal in concentric rings around the source, possibly responding to a concentration gradient. Each transcription factor has been shown to be associated with (and may help control) th e expression of specific biochemical pigment pathways. In the uncharacterized system of interest in the Riodinidae, questions included: Is a long-range diffu sible morphogen involved? Is the transcription factor Distal-less involved? Are othe r transcription factors that ar e known to be involved in the nymphalid system operating, and do they have similar function? Th e first step was to obtain live material; the second was to determ ine if similar developmental mechanisms are involved. To obtain live specimens of butterflies in this group, it was necessary to travel to Central America. The Mesosemiini are found throughout the Neotropics, but life history and host plants for species in this gr oup are largely unknown. A large-scale ongoing effort to gather just such data for all speci es of non leaf-mining Lepidoptera is currently underway in the province of Guanacaste, in Co sta Rica, under the direction of researcher Dr. Daniel Janzen. Almost all known host r ecords for the group of interest were from this region. Appropriate collecting and export/import permits were obtained and a preliminary scouting trip in February of 2005 provided encouraging evidence that the Area de ConservaciÃ³n Guanacaste would pr ovide suitable collection sites. Many Riodinid species are myrmecophilous, and some are obligate myrmecophiles. Fortunately, Mesosemia have never been observed to a ssociate with ants, making rearing a less daunting endeavor than w ith other riodinid species.
14 Unfortunately, owing to low abundance of the target species in these areas (Volcanoes Cacao, Orosi, and Rincon de la Vieja were all scoute d by foot and using available local transportation and Â“lodgingÂ”), collecting during the first several weeks in the summer of 2005 yielded only a few males, one female, and three larvae. It became clear that it would be necessary to move the study area to another pa rt of Costa Rica that had a larger area at the appr opriate higher elevation, where it would be possible to find larger populations of Mesosemia . After obtaining modified permits though the appropriate channels, work was relocated to the San Luis research station near Monteverde in the Central Cordillera. Un fortunately, specific hos t plants were unknown for this area; they were likely plant species in the same genus ( Psychotria : Rubiaceae), of which there were many species present a nd on which finding larvae proved difficult (none was found). Adults, however, were found to be very abundant along streamsides in the cloud forest above the secondary growth ar eas. Perhaps only 3 females were seen for every 100 males of the species Mesosemia asa . Mesosemia grandis was nearly as abundant although no females were observed. Only a few Mesosemia carissima males were collected (Figure 3). For these obvious ly sexually dimorphic species, males perch or patrol sunlight patches al ong the stream edges from about 10 a.m. to as late as 4 p.m. All females collected were caught around midday, while either being pursued by males or wandering through undergrowth in cla ssic oviposition search behavior. All attempts at inducing oviposition from collected females failed. Attempts to induce oviposition were conducted in a vari ety of conditions and enclosures, ranging from small cups that encourage close contact with host plant cuttings, to relatively large (4Â’ by 4Â’ by 4Â’ hand-built flight enclosures), to actually en closing females with live host
15 plants using a large area of netting in the field in seemingl y suitable conditions. In the larger enclosures, predation (especially by an ts and spiders) was unfortunately a common occurrence. The adults proved difficult to force-feed sugar solution by hand. Longevity without predation was typically about 5-7 days. No eggs were ever collected, and attempts to remove eggs closest to oviposit ion were unsuccessful as these eggs failed to be viable. The original goal of keeping one of the more common species in captivity as a colony was discontinue d at this point. A B Figure 3. Mesosemia species. A) Mesosemia asa male and female B) Mesosemia grandis male After relocating back to Guanacaste a nd waiting out a hurricane that developed off the coast of Nicaragua and produced large qu antities of rain, attempts to find larvae at the Pitilla research station were finally successful. Multiple larvae were collected. Some imaginal wing discs were field-dissected and st ored in 100% methanol for the return trip; other larvae were shipped in Â“escape-proofÂ” c ontainers, according to pe rmit, to the lab in Florida, where they were fed on host plants that had been previously shipped back to Florida, also according to permit (shipped fr om San Jose without any soil or leaves, thoroughly washed and inspected). These larvae were dissected at th e appropriate stage,
16 during the 5th larval instar, and the wing discs were stained with the appropriate antibodies (see Materials and Methods). Results and Discussion There was no observed expression of the genes Distal-less ( Dll ) or Engrailed ( En ) in a pattern specific manner, despite clear expr ession of these genes in both internal and external positive controls (Figure 4). Expression was observed on the wing specific to the posterior compartment in the case of E n, and along the developing wing margin in the case of Dll, as was expected, but no expr ession was observed in the field of cells corresponding to the future location of the disc al cell in any of the samples stained. The stains seemed to have worked well, and th e age of the discs can be judged by the amount of tracheal invasion acro ss the disc. In the model satyri d species, regions corresponding to future eyespots express these genes throughout this time period (see Reed and Serfas 2003). This suggests that the riodinid eyes pot may not be developmentally homologous to the satyrid eyespot. That the antibody worked for these transcri ption factors in this particular group of butterflies suggests that it is unlikely the antibody simply failed to recognize the correct antigen when expressed in the pathway related to pigment production. The limited sample size was an unfortunate result of the difficulty of obtaining larvae at the correct stage. Regardless, these data provide evid ence that patterning mechanisms may be different in this unique group of butterflie s. Given that developmental patterning mechanisms are so often reused to pattern similar phenotype across large evolutionary time scales, this result was somewhat surprisi ng. It suggests that the butterfly eyespot, as described in the Satyridae, truly is an e volutionary novelty that evolved perhaps only once in that group. It then becomes more interesting to ask the same question about
17 homology of other types of pattern elements in different butterfly families, although to date little is known about developmental m echanisms responsible for producing these other patterns. Figure 4. Expression patterns in the Mesosemiin i. The dark region corresponds to slight tissue damage during staining. A) Leucochimona lagora adult. B) En expression. C) Dll expressi on. D) Light field image. Lacking the ability to keep at least one of these species in a colony in the lab or greenhouse, it is currently not possible to dete rmine what other genes might be involved in patterning the riodinid eyespot. Even without information on which patterning genes are responsible, we can infer at least a few things about patterning mechanism. Correlative evidence suggests the existence of a focal organizer at each of the white dots at the center of these eyespots. In several Mesosemia species, such as in M. laruhama
18 and M. melviana , some of the secondary eyespots have relatively widely separated centers, around which separate pool s of pigment seem to radiate (Figure 5). This is very similar to the effect of transplanting eyespots in Bicyclus in terms of the geometric shape produced when two organizers are close toge ther (Monteiro 2002). A similar effect is produced in some of the Bicyclus mutants in which shapes formed and centered around one organizer bleed into an area centered around a second focal organizer, such as in the Spotty mutant line. Many species in the Satyrida e exhibit pattern similar to that shown in Figure 5, when a secondary organizer adds a smaller amount of pigment onto a larger eyespot (example shown in Figure 6). This s uggests that each white spot is indeed an organizer that produces morphagen, with the en d result of pigment in concentric rings around the center. It is possibl e that asymmetry of the discal cell region or differences in pigment diffusion rates could make three or ganizers a necessity, in order to produce a symmetric, eyespot-like shape in the Riodi nidae. Selection may have acted on developmental mechanisms in this way, with the end result of multiple organizers. In Bicyclus , some mutant lines (such as Comet ) have been generated th at have lines of cells that act as organizers, instead of spots, and in many examples of natural variation in the Satyridae, organizers seem to be broken up into many separate groups of cells (Figure 7). Larval Â“EyespotsÂ” and Some Initial Work Almost entirely unexamined are pigment patterns laid down in concentric rings on larval and pupal cuticle (Figure 8). Thes e concentric rings of pigment on bodies of immature stages (with totally different struct ure and location) seem much less likely to share developmental patterning mechanisms with adult wings than do eyespots on wings of adults of different families. Initial work in the laboratory on larval eyespot patterns in
19 Figure 5. Mesosemia secondary eyespots showing evidence that each white central spot may be an organizer; note asymmetry of the eyespot especially near potential organizers (areas of white scales). Organizer location is not as clear-cut in the Riodinidae as it is in examples in the Satyridae. Figure 6. Close up of Euptychia rubrofasciata as an example of pattern when organizers are close to one anothe r in the Satyridae. A B Figure 7. Neonympha areolata . A) The less-focused organizers of Neonympha areolata , as an example of a scattered or separa ted organizer. B) a close up. Part A) photo courtesy of Jaret Daniels.
20 the Papilionidae suggests that these may also be determined by a mechanism involving a different suite of transcription factors. St ains of this region of developing cuticle at multiple developmental stages failed to produce evidence of gene expression of Dll , En , or Spalt ( Sal ). Cautery experiments to test for the presence of an organizer proved to be more difficult than cautery of the single ce ll layers of the pupal wing pad. Laser and physically cauterized larvae healed with in consistent effects on the pattern that was observed during the followi ng molt (results not shown). Many larvae and pupae of different species have spotting patterns, sometimes in the form of multiple concentric rings that seem phenotypically, and perhaps selectively, very sim ilar to those found on adult butterfly wings. Confir mation of any sort of select ive advantage provided by this type of pattern awaits expe rimental evidence. The stai ns done in this study were relatively preliminary, and difficulties in technique and in estimating the correct developmental timing could have produced the same negative results, so these data were inconclusive. Larval pigment patterning is a promising area for future research in development and evolution of patterning mechanisms. Figure 8. Swallowtail larval o celli: in this case, from Papilio palamedes . Photos courtesy of Jaret Daniels.
21 A Note on Phenotypic Complexity and th e Evolution of the Eyespot: an Ongoing Project An ongoing effort involves tracing what ar e potentially different stages of pattern evolution of the eyespot char acter onto a phylogeny of the Mesosemiini developed using molecular methods. Within the incredibly di verse genera of interest, there is a wide range of pattern elements in the same ge neral position on the wing. These range from simple spots, to something that seems very similar to the model Bicyclus eyespot, to eyespots with multiple extra concentric rings of color. A mechanism for eyespot evolution has been proposed in the literature, but never tested (Brunetti et al. 2001). The simpler, potentially Â“primitiveÂ” character state in the Riodinidae looks superficially like this hypothetical original eyes pot. Intermediates range up to the more complex state of the character, which is almost indistinguishable from the pattern element in Bicyclus anynana . Developing a molecular phylogeny would allow testing of this hypothesis. If the more complex character is more derived, an d if increases in the character complexity are monophyletic, this might suggest that e volution of the character occurred in this order. It will be unfortuna tely very difficult to follow this work with examination of molecular patterning mechanisms in species that exhibit different leve ls of complexity of eyespot character, due to the difficulty in inducing oviposition. This hypothesis may have some support in that it has been sugge sted by one of the top experts on riodinid systematics that this group may not be monophy letic in its currently accepted taxonomic order (Jason Hall, Smithsonian Institution, pers onal communication). Several genera that have more complex pattern elements may actua lly fall within the large group that also has more complex pattern; notably, the additi on of a yellow pigment ring around the black
22 area may be a monophyletic character after all. This hypothesis awaits testing upon generation of a phylogeny of the Riodinidae. This phylogeny has been partly completed, using selected members from the genera of interest. Almost all of the materi al required for such an analysis has been collected and deposited at the McGuire Center for Lepi doptera and Biodiversity, where much of this work has been conducted. A la rge number of species has been collected from across South and Central America within th e last ten years and is available for use. Cytochrome oxidase I was used, along with the nuclear genes Wingl ess and EF-1alpha; these genes show appropriate levels of diverg ence for a study at th is taxonomic level. There are many possible outcomes for how th e character could fall on a phylogeny. One possibility is that the Â“complexÂ” element will be monophyletic, or fall out near the tips of the tree, and that the less developed character will belong to species that diverged longer ago. If this is the case, we may have ev idence of how this pattern evolved in the Lepidoptera. Alternate scenarios could be hypothesized and then te sted by developing this phylogeny.
23 CHAPTER 3 A PHYLOGENETIC SURVEY OF PA TTERNING GENE EXPRESSION Given that the eyespot patterning genes ma y not be used to pattern other eyespotlike patterns, I began a survey to determin e just how far patterning mechanism homology might extend. Preliminary evidence from othe r wing pattern research sugested that the phylogenetic origins of these patterning gene s for their use in patterning pigmentation might have arisen earlier than just within the family Nymphalidae (Antnia Monteiro, Yale University, personal communication). It would be informative to know what other types of pattern element are patterned by genes in this pathway, and where in a phylogeny of the Lepidoptera this particular evolutionary nove lty first appeared. The key genes of interest were those that have been shown to be involved in organizer specification, and those that pattern the concen tric rings of pigment, especially the genes Distal-less , Spalt , Engrailed , and Notch . Patterning Gene Expression Efforts were made to stain at least seve ral sample species of each of the major butterfly families, to serve as examples. This work was focused especially on species that have patterns that are sometimes referre d to as â€œeyespotsâ€, such as the potentially attention-drawing patches of co lor located near tails on the hi ndwings of swallowtails and in the hairstreaks, or the somewhat concentr ic pattern of â€œleafspot sâ€ on cryptic pierid species. These patterns are phenotypically di fferent from the more traditional and more widely recognized satyrid eyespot; these patt erns are often less complex and do not have
24 as many concentric rings of color, if any. The species examined in cluded: in the family Pieridae, Eurema daira and Phoebis philea ; in the family Hesperidae, Urbanus proteus ; in the family Nymphalidae, Precis coenia , Morpho peleides , Caligo atreus , and Vanessa atalanta ; in the family Lycaenidae, Strymon melinus ; and in the family Papilionidae, Papilio cresphontes , Papilio glaucus , Papilio palamedes , Papilio polyxenes , and Battus polydamus . For most of these species, stains we re done on at least 23 different ages of larvae, at stages spanning the late 4th to late 5th larval instars. In some cases, multiple stains were attempted, as negative results are especially difficult to interpret, and most of the results were negative. With some especially interesting excep tions, discussed below, few of the stains produced recognizable examples of localization of gene expression that might be specific to any particular type of pattern. Excep tions mainly included already-characterized patterns in the nymphalid species that exhi bit eyespot patterns, or intervein line expression near the margin in many families; this type of pattern has been previously examined in Reed and Serfas (2003). This observed lack of expression could result for several reasons. Seve ral of these genes ( En and Sal ) are not normally expressed even in the eyespot until the pupal stage, and all of this work examined larval imaginal discs. It may be that these genes are active in patte rning pigment in the species later in wing development. Or, alternately, it may be that these genes only rarely play a role in patterning the wing in pigment pattern elements other than the eyespots. In order to narrow these alternatives, future work might examine species from these families at later developmental stages.
25 In only a few other examples have pattern ing genes been shown to be specific to other types of pattern. Th e two red forewing bands in Precis have been shown to correspond to expression of the gene Wingless ( Wg ), (Carroll et al. 1994). One other published work (Brunetti et al. 2001) attempted to examine gene expression in non-model species, although the species studied were still mainly nymphalid species (three of these), and only one species from outside that eyespo t-containing family. As in this survey, eyespot-like patterns in the Nymphalidae corre lated generally with eyespot patterning genes. The exception was a non-Â“eyespotÂ” (y et somewhat similar) type of pattern observed in Lycaeides melissa . In this species, expression was very similar to another type of expression described below. It would also be interesting to see how far eyespot-like pattern homology extends in terms of developmental mechanism. There are many satyrid and nymphalid species with eyespot-like patterns that arguably do not functio n selectively as an eyespot. These range from vaguely eyespot-like to simple spots, pe rhaps surrounded by half crescents, as in the Speyeria species. These other t ypes of pattern are relativel y close phylogenetically to species with eyespots patterned by known mechanisms. It should be possible to test these related yet varied patterns for ho mology in developmental mechanism. Overall, most of the stains were negativ e for gene expression in a recognizably pattern-specific manner. Negative results are notoriously difficult to interpret in studies of gene expression either, th rough immunohistochemistry or in situ hybridization. It is therefore the fewer positive result s that are more interesting. Expression of Sal in Swallowtail Marginal Crescents Expression of the patterning gene Sal correlates with future expression of the Â“marginal crescentÂ” pattern elements on swallo wtail wings. Sal seems in at least some
26 stains to be more localized and concentrated where the marginal crescents will be in the adult pattern of several species of swallowtail butterfly (Figure 9). This is one of the first descriptions of a potential patterning mechanis m for any wing pattern element that is part of the nymphalid ground plan, other than the ey espot. This type of expression, especially in a comparative context, is the best evid ence we have that any of the wing patterning genes play a role in patterni ng of pigmentation, even in the eyespot. Sal has been shown in this way to be tied to expr ession of black pigment in the Bicyclus eyespot (Brunetti et al. 2001). On swallowtail wings, it corresponds to regions that wi ll produce white or yellow patches of scales. In the most posteri or region of the forewi ng, this crescent is split around the intervein line both in Sal expr ession and in the adult pattern, providing some initial comparative evidence that the two are linked. A few swallowtail species lack marginal crescents, and these might al so provide a basis for comparison, although this would likely be difficult and dependent on a time series, for reasons discussed below. Ideally, once functional means of testing the role of pattern ing genes are developed, it will be possible to functionally demonstrate the effects of modification of Sal expression on the final pigment phenotype. A time series of Sal stains at different stages of development revealed that this expression might change over time. Sal s eems to be expressed along the future wing margin by at least the early 5th instar. In fact, in some stains, expression was observed along the entire future wing margin, with only slight increases in intensity where the marginal crescents will be specified (Figure 9) . Whether this varies over time, or whether it is an artifact specif ic to each individual stain, is unknown. It is possible, for example, that the line of expression c ondenses into crescents at a spec ific stage in development, but
27 this is as yet unclear. It will be necessary to complete a more extensive series of stains at more individual time steps to determine te mporally specific changes in expression. Figure 9. Sal expression in Papilio polyxenes , 5th instar larva. A) Note expression in the region where marginal crescents will be produced. B) Papilio polyxenes adult wing. C) a close-up showing nuclear localization of Sal expression. This correlation between marginal cres cents and Sal expression is consistent across multiple swallowtail species. This type of patterning gene expression was observed in Papilio polyxenes , P. palamedes , and P. cresphontes . In order to further an examination of patterning mech anism homology, the next st ep would be to test for marginal crescent homology in species in other families that also have marginal crescent patterns. This would be a test of devel opmental homology suggested originally by the nymphalid ground plan, for one of the major components of pattern. The expression of Sal along the entire future wing margin s uggests that there might perhaps be other functional roles for this gene
28 Possible Function of Sal Expression at the Future Wing Margin This earliest phylogenetic instance of Sa l expression in a manner related to pigment corresponds in time and space with several other potentially older and more A B C Figure 10. Swallowtail Sal expression. A) Papilio palamedes stain showing Sal expression along the entire future wing margin. B) The same sample/stain, light field photograph. The field of cells where the future wing margin will be is visually undifferentiated. C) A s econd exposure of the same sample/stain. conserved functions of this gene. To de velop hypotheses of th e potential function of patterning genes in butterflies, it may be useful to turn to a much better studied model Â– Drosophila . Functions of Sal on Drosophila wings may well be homologous to function on the developing butterfly wing. Several poten tial functions are part icularly obvious. In the Lepidoptera, trachea invade the wing disc over the course of the 5th instar. In my stains, the trachea often initially overshoot the wing margin itself, followed by
29 growth of secondary tracheal branches into the region of Sal expr ession along the margin; this forms the "border lacuna" (BL), (Figure 10). Growing trachea themselves may even be expressing Sal. In Drosophila , Sal expression is functionally necessary for tracheal growth in the dorsal trun k (Kuhnlein and Schuh 1996). More importantly, it regulates positioning of fly wing veins through activation of the gene rhomboid ( Rho ) (there is good evidence that this determines position of L2), and Rho mutant flies have trouble developing any distal wing veins at all (S urtevant et al. 1997, de Celis et al. 1996, Duncan 1935, Waddington, 1939). This suggests th at Sal expression along wing margins in Lepidoptera might play a role in gu iding tracheal/BL grow th, potentially via Rho activation. Figure 11. Growing trachea initially overs hoot the future wing margin; secondary branches grow in later to form the border lacunae. The primary candidate for possible function is in placing the wing veins that close off the wing, but another function may involve determining which ce lls undergo targeted cell death (apoptosis). Th e cells outside this line of BL at some point become fated to die
30 off during an ecdysone peak in the pupal stage. Ecdysteroid levels ar e directly involved in triggering apoptosis in th e cells outside this area (Koch et al. 2003, Fujiwara and Ogai 2001, Kodama et al. 1995). EGF receptor signali ng has recently been shown to protect smooth-cuticle cells from apoptosis during Drosophila ventral epidermis development (Urban et al 2004). Urban et al. fo cused especially on rho-1 expression ( Rho is also a prospective BL vein formation candidate in association with Sal and possibly Wg ). Rho itself is not ecdysone inducible, but inst ead causes the up-regulation of ecdysone receptors. There are numerous other less cl ear associations between Rho and Wg, and Rho and Dll that might also play a role in determining the wing ma rgin and cell apoptosis in this region that deserve further examination and comparison of butterflies with their function in Drosophila . Sal Expression Patterns Suggest a New Hypothesis The expression of Sal along the future wing margin was interesting for a variety of reasons. Initially, I was interested primarily in what functional roles it might play, expressed along the future wing margin at th at particular stage in development, as described above. With the addition of a fe w other key pieces of information and a few extra (but logical) steps, I developed a new hypothesis. This hypothesis involves how eyespot patterning genes may have originally come to play a ro le, both in producing pigment and in patterning the eyespot. 1. As summarized above, correlative evidence from function of Sal and related genes in developing Drosophila wings suggests several potenti al homologous functions at the butterfly wing margin th at seem plausible. 2. Given that Sal might play a functional role in determining and defining the future wing margin, it became clear that these f unctional roles would be much older and more highly conserved than any function in patterning pigmentation on the wing.
31 3. Furthermore, when comparing expression of the gene Dll at the growing imaginal disc margin (example Figure 4c) with Sa l expressed at the future wing margin (Figure 10), it is clear that in most taill ess species of butterflies and moths the two lines of expression are parall el and often not widely sepa rated. In the eyespot, Dll is upstream of Sal via a diffusible morphoge n; Sal is expressed in a concentration dependent manner in response to th is morphogen, in a ring surrounding the organizer. Dll expression has been shown to quantitatively vary with eyespot size, likely by increasing the amount of mor phogen produced. (See diagram in Figure 11.) The eyespot might actually be a subtle reversal of a very similar mechanism ancestrally found along the wing margin. A B Figure 12. Diagram of the proposed mechanism. A) Early in larval disc development, Dll is expressed along the disc margi n, shown in red. Diffusion of the uncharacterized morphogen is shown in gray scale. Placed on this gradient is a line of Sal expression that would come to determine the future wing margin, shown here in blue. The area outside the future wing margin degenerates and dies off during the pupal stage. B) Overvi ew of this circuit in the eyespot. Shown in red is Dll, blue is Sal, a nd yellow is an additional transcription factor, En. [F] is the concentration of the focal signal, in this case proposed to be the same signal as the morphogen depict ed in both A) and B) in greyscale. In both, Dll triggers production and di ffusion of morphogen, localizing Sal in a concentration-dependent manner. En is produced in response to the same morphogen, but in response to lower concentrations than Sal. 4. In this model, Dll expression at the wing margin would trigger production and diffusion of the same morphogen as in the ey espot, in turn trig gering Sal expression in a similar concentration-dependent manne r. This would produce pattern similar to that observed at the wing margin in Fi gures 8 or 9. Differences in rates of diffusion before, and/or in rates of cell di vision after establishment of this line of expression, would then help to shape the wing, and they would be involved in the differentiation of characters su ch as scalloped wing edges or tails. These characters in turn affect flight performance.
32 5. The diffusion of morphogen and estab lishment of Sal expression would of necessity occur early enough in developm ent that diffusion distances were not insurmountable. 6. The morphogen itself has not been char acterized, but it could be the gene Wingless ( Wg ), a known, long-range diffusible morphogen that happens to be associated with patterning wing margins in Drosophila . Wg is expressed in what seems to be just such a concentration gradient radiati ng out from the margin of the wing, and produced in association with Dll Â– see data presented in Kango-Singh et al . 2001. Evidence for and against this particular hypothesis is presented below. 7. If this circuit functions as suggested he re, at the wing margin, the only missing step would be for Sal to become coupled with pigmentation. Sal expressed along the wing margin provides preci se positional informati on. Any control of the transcription factor Sal over any of the biochemical machinery necessary for production pigment could provide the link ne cessary. This is perhaps the least clear link in this proposed series of steps, but once Sal became coupled with pigmentation, any misexpression of Dll a nywhere on the wing would lead to production of morphogen, then expression of Sal on that concentration gradient, and hence pigment in a concentric ring around the point of expression. This was the primitive eyespot, a selectable character. 8. An alternate sequence of events leading to the primitive eyespot might have been more gradual. Species such as Cethosia biblus (Figure 12), which are in the Nymphalidae, and phylogenetically close to species exhibiting the eyespot character, likely are similar in gene expression to all other studied nypmphalids (Reed and Serfas 2004), in that they shoul d have lines of expression of Dll on the midvein line. This could be an intermediate step between swallowtail marginal crescents and circularized pattern in a slig htly more basal pos ition, in the marginal band of symmetry. A sort of Â“pinching offÂ” over time of the line of expression would lead to just such a pattern. This type of expression is very similar to what is observed in the developing eyespot over time : a line of expression of Dll retreats toward the wing margin, leaving behind one spot of expression that becomes the eyespot organizer (Reed and Se rfas, 2004). If this line we re instead left intact, a pattern much like that seen in Cethosia would develop instead of an eyespot. This V-shaped pattern is even seen in some marginal crescents in some swallowtail species. This hypothesis is suggested by the Sal e xpression data (examples: Figures 8-9), and seems consistent with what is know n of wing patterning mechanisms in both Drosophila and in the Lepidoptera. Such a mech anism seems plausible and is the first proposed hypothesis describing how genes nor mally involved in appendage outgrowth might have taken on such a novel function, in pa tterning pigment in the butterfly eyespot.
33 Genes such as Dll are very highly conserved in function across a large range of multicellular life, yet in the eyespot they are used to pattern concentric rings of pigment produced by two-dimensional layers of cells. Figure 13. Wing pattern of Cethosia biblus . Note the v-shaped marginal pattern around the most distal sections of the intervein lines. Under the suggested hypothesis, Dll tr iggers diffusion of morphogen, in turn leading to expression of Sal. This mechanis m would allow selection to act on a range of variables to shape the wing itself. The amount of morphogen produced, rates of cell division following establishment of the line of Sal (and the cells of the future wing margin), the timing and length of expression Â– these are just a few ways selection might act to change the shape of the wing, th rough changes in the timing and level of expression of genes in the pathway. This is not the first time expression of th is type has been documented. In Brunetti et al. (2001), among others, Sal can be seen ex pressed along the future wing margin. However, the potential for function in defining the main wing area would not be nearly as noticeable in species without tails as it is in the swallo wtails, in data presented here. It is possible that En expression also falls on th e morphogen gradient, expressed in response
34 to even lower levels of mor phogen, just as it is in the Bicyclus eyespot. In Brunetti et al. (2001), this is the cas e at the margin of Lycaeides melissa . In my stains, En expression increased slightly in intensity in a region just interior of the line of Sal expression. Whether this is incidental or has some func tional role also has ye t to be determined. The gene Wg is expressed at the imaginal disc margin in a gradient extending away from the zone of Dll expression (K ango-Singh et al. 2001). Wg normally plays a role in defining the disc margin in Drosophila . The expression of Wg on the developing lepidopteran wing is in the correct position to be the potential mo rphogen linking Dll and Sal. There is evidence that Wg both (a) represses Sal during sens ory organ development in the thorax (de Celis et al., 1999) and (b) enhances Sal expression during tracheal development (Bradley and Andrew 2001, Llimargas 2000, Wappner et al. 1997, Chihara and Hiashi, 2000). These researchers do not addr ess one another in terms of whether Sal functions both as a repressor and an enhan cer in a stageor tissue-specific manner, although Sal has domains necessary for both re pression and enhancement. If Wg is upstream of Sal in tracheal growth, the sa me might be the case in butterfly wings, especially given that Wg normally has a role in wing margin determination. As a potential morphogen defining and shaping th e area between disc and wing margin, it deserves further examination as a potent ial morphogen in the eyespot. Potential associations and interactions between Wg and Dll in Drosophila wing margins are varied (Zecca 1996). A brief examination of butterfly Wg expression was described only briefly in Carroll et al. (1994); one figure showed an in situ hybridization for Wg. These stains showed the marginal expression of Wg, as well as expression in areas correlating with the red bands found on the wings of Precis , but no evidence of its expr ession in the eyespot.
35 Not published was information on what developmental stages were examined or how many attempts were made. Wg certainly de serves further examination as a potential candidate as the diffusible morphoge n that patterns the eyespot. An initial examination of rho-1 expre ssion revealed what may be ubiquitous expression in the larval disc (Figure 13). In this imag e, staining of nuclei in the peripodial membrane may obscure the actual wing tissue itself. In either case, expression of Rho is likely to be specific to certain developmental stages. The ecdysone peak that induces pupation precedes the ecdysone peak that causes cell death, suggesting that upregulation of Rho may occur between these two peaks and af ter the larval stage. If Rho was up-regulated before this point, it mi ght be expected that the first, larger ecdysone peak would induce cel l death earlier than appropr iate. Initial data is inconclusive, and rho-1 expressi on deserves further examination. Correlative evidence via staining of more re lated genes and at more developmental stages may begin to answer some of th ese questions. Detailed time steps during development might be useful to document ch anges in expression ove r time. Especiallyconclusive evidence may await means of tes ting gene function, which is not yet possible for specific patterning genes in the Lepidopt era. One of the most interesting, but difficult, components of this scenario of pa thway cooption involves how Sal, Dll, and En might have become coupled with pigmentation mechanisms. These genes must at some point have gained upstream control of genes n ecessary for the synthesis of pigment. That this occurred was already know n from their involvement in the eyespot. Work in this thesis suggests that perhaps this occurred in stages, but especially via a coupling of Sal and pigmentation at the wing margin. Sal and other patterning genes often provide very
36 precise positional information during developmen t, and it seems logical that some ability to link this information to adaptive traits w ould be selectively bene ficial. Whether this occurs through simple mutational modification of transcription fact ors or binding sites, through gene/regulatory region dup lication and divergence, or even through factors such as transposable elements causing rearrangeme nt of regulatory elements, a variety of possible mechanisms might be involved. Figure 14. Expression of rho-1 in Papilio cresphontes 5th instar larva. If Rho is involved in targe ting apoptosis to the area outside the wing margin via changes in the ecdysone signal-transduction pa thway, it is possible that it also plays a role in ecdysone-regulated pigmen tation. It has been shown th at ecdysone levels directly affect pattern in various examples of seas onal polyphenism, both in eyespot size in Bicyclus , and in seasonal pigmentation involved in thermoregulation. Rho seems like an excellent candidate for involvement in mo lecular signaling between ecdysone reception and an ultimate change in pigment phenotype.
37 Overall, the idea that Sal might have origin ally come to play a role in defining the wing margin, perhaps as a guide for trachea l growth to form the border lacuna, and perhaps to prevent apoptosis inside the area defined by the BL, is a plausible one. Sal certainly plays a role in wing tracheal development in Drosophila . It is also perhaps downstream from Wg, which plays a role in defini ng the wing margin in Drosophila . In more-derived lepidopteran taxa, Sal expr ession seems to have become tied to pigmentation. This is certainly the case in the eyespot, and now it is suggested that this may be the case in marginal crescents as well. Data on whether this is the link between function of these genes at the wing margin a nd their function in the eyespot would be an exciting area for future research. Materials and Methods Immunohistochemistry Butterfly larval di scs were dissected out and fixed for 30 minutes in 2% paraformaldehyde solu tion containing 0.1 M PIPES (pH 6.9), 1mM EGTA, 1% Triton X-100, and 2 mM MgSO4. To prevent nonspecific binding, discs were blocked for 2 hours in 50 mM Tris (p H 6.8), 150 mM NaCl, 0.5% NP40, and 1 mg/ml bovine serum albumin (BSA). Discs were th en incubated overnigh t in 50 mM Tris (pH6.8), 0.5% IGEPAL (NP40), 150 mM NaCl, and 1 mg/ml BSA (wash buffer) containing monoclonal mouse antibody 4F11 against En/Inv (1:5) (Patel et al . 1989) and either rabbit anti-Dll (1:100) (Panganiban et al. 1996) or rabbit antiSal (1:200) (Barrio et al. 1999). Discs were washed four times in wash buffer and incubated for 2 hours with wash buffer containing either anti-mouse or anti-goat Alexa Flour (Invitrogen) secondary antibodies. The discs were then washed four times and mounted in 70% glycerol. Images were collected on a Leica DMRE microscope with appropriate fluorescence emmiters/filters.
38 Conclusion There is much that is still unknown a bout developmental patterning mechanisms in the Lepidoptera. The adult wing pattern provides an easily observable and often exquisitely-detailed surface, describing the end result of molecular signaling mechanisms. These patterns are acted upon by various forms of selection or drift to produce nearly endless variation of shapes a nd colors. This variation, however, falls within the boundaries of what seem to be a se t of rules that have come to be known as the Â“nymphalid ground planÂ”, which can be used to describe homology of pattern types. This thesis describes some initial attempts to examine the mechanisms behind this pattern and set of rules in multiple species and families, in an effort to determine how much of the underlying mechanism is actually homologous, and how much it might vary. It is also an examination of a possibly unique example of evolutionary novelty through comparative methods. Though many difficulties ar ose, and though this is only an initial effort toward unraveling an extensive puzzle, data collected has already revealed a number of specific hypotheses that are worth ex amination and testing. The intricate patterns on butterfly wings serve as excellent examples of the result of the interaction between gene s, development, form, and function. These relatively complex patterns are extraordinar ily flexible in response to se lection. It has been said that butterfly wing patterns will become one of the first examples of morphological diversity to be successfully st udied at many levels of biological organization at once. Many pieces of the puzzle are still missing, and some of the most fasc inating include first a description of how patterning genes came to play a role in patterning the wing, and then how this set of genes and their interactions have changed over time and across multiple
39 families to produce the spectacular and dive rse array observable today. The work described here has provided some inroads in to how pattern develops and evolves across larger groups. This broader-scale comparativ e context may lead to different approaches to similar questions, or entirely new questions may arise. It is my hope that data and ideas provided here will prove useful in future research, bo th in my own efforts and by others.
40 LIST OF REFERENCES Barrio, R., de Celis, J.F., Bolshakov, S., Ka fatos, F.C. (1999). Identification of regulatory regions drivi ng the expression of the Drosophila spalt complex at different developmental stages. Dev. Biol. 215 , 33-47. Beldade, P., Brakefield, P.M., and Long, A.D. (2002a). Contribution of Distal-less to quantitative variation in bu tterfly eyespots. Nature 415 , 315. Beldade, P., Koops, K., and Brakefield, P.M. (2002b). Developmental constraints versus flexibility in morphologi cal evolution. Nature 416 , 844-847. Beldade, P. & Brakefield, P.M. (2003). Concerted evolution and developmental integration in modular butterfly wing patterns. Evol. and Devl. 5(2) , 169-179. Bradley, P. L. and Andrew, D. J. (2001).. ribbon encodes a novel BTB/POZ protein required for directed cell migration in Drosophila melanogaster. Development 128(15) , 3001-15. Brakefield, P. M. (2001). Structure of a char acter and the evolution of butterfly eyespot patterns. J. Exp. Zool. 291 , 93. Brakefield, P.M., Gates, J., Keys, D., Kesb eke, F., Wijngaarden, P.J., Monteiro, A., French, V., and Carroll, S.B. (1996). Deve lopment, plasticity and evolution of butterfly eyespot patterns. Nature 384 , 236. Brakefield, P.M. (1998). The evolution-deve lopment interface and advances with the eyespot patterns of Bicyclus butterflies. Heredity 80 , 265-272. Brakefield, P.M., Kesbeke, F. & Koch, P.B. (1998). The regulation of phenotypic plasticity of eyespots in the butterfly Bicyclus anynana . Am. Nat . 152 , 853-860. Brunetti, C.R., Selegue, S.E., Monteiro, A., French, V., Brakefield, P.M., Carroll, S.B. (2001). The generation and divers ification of butterfly eyes pot color patterns. Curr. Biol. 11 , 1578-1585. Carroll, S.B., Gates, J., Keys, D.N., Pa ddock, S.W., Panganiban, G.E., Selegue, J.E., Williams, J.A. (1994). Pattern formation a nd eyespot determination in butterfly wings. Science 265 , 109. Carroll, S. B., Greiner, J.K., Weatherbee, S.D. (2001). From DNA to Diversity: Molecular Genetics and the Ev olution of Animal Design , Blackwell Science. Maldon MA, USA.
41 Chihara, T. and Hayashi, S. (2000). C ontrol of tracheal tubulogenesis by Wingless signaling. Development 127 , 4433-4442. de Celis, J.F., Barrio, R., and Kafatos, F. C. (1996). A gene complex acting downstream of dpp in Drosophila wing morphogenesis. Nature 381(6581), 421-424. de Celis, J. F., Barrio, R. and Kafato s, F. C. (1999). Regulation of the spalt/spalt-related gene complex and its function during sens ory organ development in the Drosophila thorax. Development 126(12) , 2653-2662. Duncan, F.N. (1935). Veinlet, a first rank reces sive, located just to the right of roughoid in chromosome three of Drosophila melanogaster. Am. Nat . 69 , 94-96 French, V. and Brakefield, P. M. (1995). Eyespot development on butterfly wings: the focal signal. Dev. Biol . 168 , 112-123. Galant, R., Skeath, J.B., Paddock, S., Lewis, D.L., Carroll, S.B. (1998). Expression pattern of a butterfly achaete-scute homolog reveals the homology of butterfly wing scales and insect sens ory bristles. Curr. Biol. 8 , 807Â–813. Haruhiko, F., Shinya, O. (2001). Ecdysteroi d-induced programmed cell death and cell proliferation during pupal wing development of the silkworm, Bombyx mori , Dev. Genes and Evol. 211(3) , 118Â–123. Holloway, G.J., Brakefield, P.M., and Kofma n, S. (1993). The genetics of wing pattern elements in the polyphenic butterfly Bicyclus anynana . Heredity 70 , 179-186. Jiggins, C. D., Naisbit, R.E., Coe, R.L., and Mallet, J. (2001). Reproductive isolation caused by colour pattern mimicry. Nature 411 , 302Â–305. Joron, M. and Mallet, J. L. B. (1998). Divers ity in mimicry: paradox or paradigm? Trends Ecol. Evol . 13 , 461Â–466. Kango-Singh, M., Singh, A., and Gopinath an, K.P. (2001). The wings of Bombyx mori develop from larval discs exhibiting an early differentiated state: a preliminary report. J. Biosci, Indian Academy of Sciences. 26(2) , 167-177. Keys, D.N., Lewis, D.L., Selegue, J.E., Pearson, B.J., Goodrich, L.V., Johnson, R.L., Gates, J., Scott, M.P., and Carroll, S.B. (1999). Recruitment of a hedgehog regulatory circuit in butterfly eyespot evolution. Science 283 , 532Â–534. Kingsolver, J.G. (1995). Fitness consequences of seasonal polyphenism in Western white butterflies. Evolution 49 , 942Â–954. Koch, P.B., Keys, D.N., Rocheleau, T., Ar onstein, K., Blackburn, M., Carroll, S.B., ffrench-Constant, R.H. (1998). Regulat ion of dopa decarboxylase expression during colour pattern formation in w ild-type and melanic tiger swallowtail butterflies. Development 125 , 2303Â–2313.
42 Koch, B.P., Merk, R., Reinhardt, R., Weber, P. (2003). Localization of ecdysone receptor protein during colour pattern form ation in wings of the butterfly Precis coenia (Lepidoptera: Nymphalidae) and co-expr ession with Distal -less protein. Dev. Genes and Evol., 212(12) , 571Â–584. Kodama, R., Yohida, A., Mitsui, T. (1995). Pr ogrammed cell death at the periphery of the pupal wing of the butterfly, Pieris rapae , Dev. Genes and Evol. 204(7), 418. Kuhnlein R.P., and Schuh, R. (1996). Dual functi on of the region-specific homeotic gene spalt during Drosophila tracheal system development. Development 122(7) , 22152223. Llimargas, M. (2000). Wingless and its signa lling pathway have common and separable functions during tracheal development. Development 127 , 4407-4417. Mallet, J. (1991). Variations on a theme? Nature 354 , 368. Monteiro, A.F., Brakefield, P.M., & French, V. (1994). The evolutionary genetics and developmental basis of wing patte rn variation in the butterfly Bicyclus anynana . Evolution 48 , 1147-1157. Monteiro, A.F., Brakefield, P.M., & French, V. (1997a). Butterfly ey espots: the genetics and development of the color rings. Evolution 51 , 1207Â–1216. Monteiro, A.F., Brakefield, P.M., & Frenc h, V. (1997b). The genetics and development of an eyespot pattern in the butterfly Bicyclus anynana : response to selection for eyespot shape. Genetics 146 , 287-294. Monteiro, A. and Pierce, N. E. (2001). Phylogeny of Bicyclus (Lepidoptera: Nymphalidae) inferred from COI, COII, and EF-1 gene sequences. Mol. Phylog. Evol. 18 , 264Â–281. Nijhout, H.F. (1980). Pattern formation on le pidopteran wings: determination of an eyespot. Dev. Biol . 80 , 267Â–274. Nijhout, H.F. (1991) The Development and Evolution of Butterfly Wing Patterns , Smithsonian Institution Press. Washington D.C., USA. Nijhout, H.F. (2003). Polymorphic mimicry in Papilio dardanus : mosaic dominance, big effects, and origins. Evolution and Development 5 , 579-582. Panganiban, G., Sebring, A., Nagy, L., and Carroll, S. (1995). The development of crustacean limbs and the evolut ion of arthropods. Science 270 , 1363Â–1366. Patel, NH, Martin-Blanco, E., Coleman, K.G., Poole, S.J., Ellis, M.C., Kornberg, T.B., and Goodman, C.S. (1989). Expression of Engrailed proteins in arthropods, annelids, and chordates. Cell 58 , 955-968.
43 Paulsen, S. M. (1994). Quantitative genetics of butterfly wing color patterns. Dev. Genet. 15 , 79Â–91. Reed, R.D., & Serfas, M.S. (2004). Butterfly wing pattern evolution is associated with changes in a Notch/Distal-less temporal pattern formation process. Curr. Biol . 14 , 1159-1166. Vane-Wright, R.I. & BopprÃ©, M. (1993).Vis ual and chemical signa ling in butterflies functional and phylogenetic perspectives. Phil. Trans. R. Soc. Lond. B 340 , 197205. Sturtevant, M.A., Biehs, B., Marin, E., Bier, E. (1997) The spalt gene links the A/P compartment boundary to a linea r adult structure in the Drosophila wing. Development 124(1) , 21-32. Urban, S., Brown, G., Freeman, M. (2004). EGF receptor signaling protects smoothcuticle cells from apoptosis during Drosophila ventral epidermis development. Development 131(8), 1835-1845. Waddington, C.H. (1940). The genetic control of wing development in Drosophila . J. Genet . 41, 75-139. Wappner, P., Gabay, L. and Shilo, B. Z. (1997) . Interactions between the EGF receptor and DPP pathways establish distinct cell fates in the tracheal placodes. Development 124 , 4707-4716. Zecca, M., Basler, K. and Struhl, G. (1996). Direct and long-range action of a Wingless morphogen gradient. Cell 87 , 833-844.
44 BIOGRAPHICAL SKETCH I was born August 26th, 1982, in Washington, D.C. After living several places in Virginia, my parents and I relocated to Mo rgantown, West Virginia, where I completed the early stages of my education, K-8th grades. During this time, and throughout high school in South Florida, I trav eled extensively with my pare nts. Their love for natural areas, especially across the US and the Neot ropics, helped to develop in me a strong interest in biology from an early age. They are in fact biol ogists, and have over the years described themselves as working in fields su ch as aquatic ecology and fisheries biology. They currently hold positions in the National Park Service, administering to specific areas of research in Everglades National Pa rk. Their influence on my decision to pursue an academic career in biology has been seemingly incidental, as they know well the perils of the academic lifestyle. However, I like to think they are happy with my decision. I spent my undergraduate y ears in north-central Flor ida, while attending the University of Florida where I graduated in 2004 with a degree in both zoology and molecular and cell biology. In my conti nuing graduate program there I met amazing people, refocused my area of main interest , learned an immeasur able amount (too much, not too little), and overall had a wonderful time. My time there resulted in not only a wealth of experiences, but in receivi ng a National Science Foundation Graduate Fellowship. Next, I plan to begin graduate school in the upcoming fall semester (2006) at the University of California, Berkeley, in the laboratory of Nipam Patel.