INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at tfie upper left-hand comer and continuing from left to right in equal sections with small overlaps.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

Bell & Howell Information and Learning 300 North Zeeb Road, Ann Artx)r, Ml 48106-1346 USA UMI 800-521-0600

Towards a Molecular Analysis of Associative Learning in the Honey Bee,

Apis mellifera, via Massed Conditioning and Genetic Transformation

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in the Graduate School of The Ohio State University

By

Kellie Oline Robinson, 8.S.

*****

The Ohio State University

1999

Dissertation Committee: Dr. Brian Smith, Adviser Approved by

Dr. David Denlinger

Dr. John Oberdick Dr. Amanda Simcox Cdviser / y/

Dr. Harald Vaessin Molecular, Cellular,^d Developmental Biology Graduate Program UMI Number: 9951717

UMI Microform 9951717 Copyright 2000 by Bell & Howell Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code.

Bell & Howell Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, MI 48106-1346 ABSTRACT

Understanding conditions under which an organism correlates temporally related events is critical for establishing a link between behavioral mechanisms of associative learning and molecular mechanisms of acquisition and storage of this information. We examined the effect of massed-trial conditioning, which involves short (30 second) intervals between conditioning trials, on learning performance using the Proboscis Extension Conditioning (PEC) paradigm on worker honey bees. Specifically, we investigated the associative nature of the effect by incorporating several control procedures that have not been evaluated in a massed trials conditioning paradigm. A second goal was to define temporal events that occur during acquisition of information and consolidation as memory. Therefore, we tested recall at different post-conditioning intervals in order to investigate the temporal dynamics of memory consolidation. Acquisition of the Proboscis Extension Response (PER) is not dependent on the inter-trial interval (ITI) during forward- paired conditioning. The PER is the behavioral response of the subject (extension of the proboscis) after exposure to sucrose or conditioned stimuli. This response was used to score individual subjects during and after olfactory training utilizing the proboscis extension within various training presentations. Furthermore, the relatively higher response levels in forward-paired conditioning groups are dependent on a specific temporal and predictive relationship between the odor conditioned stimulus and sucrose reinforcement. Our data also indicate that massed conditioning with 30 second ITI produces robust associative recall over 24 hours A full understanding of the molecular basis of learning and memory' will depend upon the functional testing of implicated genes in transgenic honey bees. As a first step towards this goal, strategies to transform the genome of honey bees with foreign DNA plasmids were investigated Linearized plasmids were introduced with semen during instrumental insemination of virgin queen honey bees. Lar\'ae from a few such queens were subsequently identified as positive for the plasmid via PCR amplification of a plasmid fragment and fluorescent expression of green fluorescent protein encoded by the plasmid. These transgenic lines transmitted the transgene for two generations, demonstrating heritability. However, there was no evidence of

integration of the construct into the honey bee genome as determined by genomic

Southern analysis. For this reason, the use of a transposable element was explored. These experiments

tested the utility of a potential universal gene vector for transformation of the honey bee genome. The Tel transposon was originally isolated from C. elegans and is well

characterized. The data indicate sperm-mediated introduction of the Tel element via instrumental insemination results in PCR-positive progeny in the FO and the FI

generations. Ligation-mediated PCR and plasmid rescue from genomic DNA samples confirmed the persistence of the Tel element through the generations. However, recombination of the Tel element into the bee genome was not detected.

Therefore we consider it likely that the Tel element is maintained but persists as an

extrachromosomal element.

Ill Dedicated to my Mom. Bonnie Robinson, who would never let me give up and My Grandma, Emma Oline Jordan, who always has a good time with life.

IV ACKNOWLEDGMENTS

I wish to thank my advisor. Dr. Brian Smith, for guidance and infinite patience throughout this research. Brian you have always seen to it that everything that could be done to further my graduate studies was available. Thank you Sue Cobey for detailed instruction in honey bee inseminations. Dr. Holly Furgeson for pioneering the honey bee transformation project, and Dr. Seetha Bhagavan for discussions on

honey bee behavior. Thank you Dr. Geraldine Wright for help with the statistical analysis, scientific discussions, and friendship through the most trying of times.

Thank you to the Vaessin lab for their collaboration in the genetic transformation of honey bees. Dr. Harald Vaessin, thank you for patiently correcting all 101

graduate student mistakes and for continuous intellectual support. Thank you Dr. Kirsten Bremer, Dr. Julie Pinsonneault, Dr. Terrie Rife, Ling Li and Dr. Amian

Ahmed for scientific discussions, encouragement and continuous moral support

when it was most desperately needed.

Thanks to the Denlinger Lab for my start in scientific research and molecular

biology. Specifically thanks goes to Dr. David Denlinger for both serving on my

committee and advising me on numerous occasions. Dr. Karl Joplin, Dr. George Yocum, and Dr. Ron Flannagan for advice, friendship and support. Dr. John Oberdick, and Dr. Amanda Simcox, thank you for serving on my graduate committee, advising on my scientific career, reading and committing on my dissertation and manuscripts.

VI VITA

June 21, 1969 ...... Bom-Crystal City, Missouri

1989...... A.A., Palm Beach Community College

1992 B.S. Biology, Florida Atlantic University

1992-present...... Graduate Teaching and Research .Associate Molecular, Cellular and Developmental Biology,

The Ohio State University

PUBLICATIONS

FIELDS OF STUDY

Major Field; Molecular, Cellular, and Developmental Biology

emphasis on Apis mellifera

VII TABLE OF CONTENTS

Paae

Abstract...... ii Dedication...... iv

Acknowledgments ...... v

Vita...... vii List of Tables...... xi

List of Figures...... xii Chapters: 1. Introduction How is learning related to the formation of memory? ...... 1

Wliat are good model animals? Aplysia californien...... 4 Drosophila melanogasier...... 7

Why honey bees? ...... 10 Toward a molecular approach to studying learning & memory

in the honey bee ...... 12 Transformation in Apis mellifera...... 13

2. Massed Conditioning Produces Associative Learning in the Honey Bee,

Apis mellifera

Abstract...... 15 Introduction ...... 17

Materials and Methods ...... 21

viii Honey bee subjects...... 21 Proboscis extension conditioning (PEC) ...... 21 Treatment groups...... 22

Results...... 25

Effect of pairing condition on post-conditioning responses day 1...... 25

day 2 ...... 28 Effect if inter-trial intern als on post-conditioning responses of

massed and spaced forward-paired groups ...... 31 Discussion...... 34

3. Inheritance of Plasmid DNA via Sperm-mediated Introduction in the

Honey Bee, Apis mellifera

Abstract...... 38 Introduction ...... 40 Materials and Methods ...... 44

Honey bee maintenance ...... 44 Instrumental insemination of virgin honey bee queens...44

Honey bee queen rearing...... 46 Plasm ids ...... 47 Genomic DNA isolation ...... 48

PCR reactions ...... 48

Fluorescent microscopic imaging...... 49 Genomic Southern Blot ...... 49

LM-PCR...... 50 Plasmid rescue...... 51

IX Results...... 52

Molecular analysis of FO generation ...... 52 Expression of Green Fluorescent Protein in the

FO generation ...... 56 Genomic Southern Blot of the FO generation ...... 59

Molecular analysis of

FI Generation ...... 62

F2 generation ...... 66 Molecular analysis of FI and F2 Generations from summer 2 ...... 72

Ligation-mediated PCR and plasmid rescue ...... 76

Discussion...... 82 Expression and maintenance of plasmid DNA ...... 82 Inheritance and survivorship of transgenic progeny 83

Extrachromosomal elements vs. genomic integration

Of plasmid DNA ...... 85 Future experimental approach...... 85 Bibliography...... 87 LIST OF TABLES

Table Paae 1.1 Molecules involved in facilitation of synaptic response ...... 5

2.1 Treatment conditions ...... 24 3.1 Nomenclature of queen insemination and generations ...... 45

3.2 Insemination of Primary queens ...... 53

3.3 Insemination of Founder queens ...... 62

3.4 Insemination of FI queens ...... 66 3.5 Queen lineages from PO queens summer ...... 1 69 3.6 Queen lineages from PO queens summer 2...... 78

XI LIST OF FIGURES

Figure Page 2.1 Effect of pairing condition on post-conditioning responses day 1 ...... 26 2.2 Effect of pairing condition on post-conditioning responses day 2 ...... 29 2.4 Effect inter-trial intervals on post-conditioning responses ...... 32

3.1 PCR amplification of plasmid from FO genomic DNA ...... 55 3.2 Fluorescence of honey bee larvae ...... 58 3.3 Genomic Southen Blot of FO honey bee queens ...... 61 3.4 PCR amplification of phGFP-S65T plasmid from FI genomic DNA.. .65 3.5 PCR amplification ofphGFP-S65T plasmid from F2 genomic DNA...68

3.6 PCR amplification of Tel plasmid from FI genomic DNA ...... 75

XU CHAPTER 1

INTRODUCTION

How is learning related to the formation of memory? A fundamental challenge to any animal lies in recognizing important stimuli and responding in an appropriate manner. .An animal’s nervous system detects and discriminates a multitude of stimuli. Specific dedicated pathways in the nervous system allow an animal to make rapid reflexive responses to stimuli that have been stable over evolutionary time. For example, a male moth will fly upwind toward the source of female pheromone. Additionally more flexible pathways in the nervous system encode responses from an animal to stimuli that have been variable over evolutionary time. These pathways underlie an animal’s ability to learn about relationships among stimuli (Pavlovian conditioning) and about the consequences of its own actions (Instrumental conditioning). Our understanding of conditions during which an animal correlates events is critical for establishing a link between the behavioral mechanisms of learning and the molecular mechanisms for acquisition, storage and recall of this information.

Behavioral studies of learning and memory can be combined with an analysis of molecular events that take place during memory formation. Attempting to understand these processes requires a focus on well-defined “critical problems” and specific approaches to answer them utilizing the most appropriate animals (Miklos, 1993).

Learning mechanisms and the cellular processes for retaining information can be investigated in a variety of organisms utilizing a number of behavioral paradigms

(Rescorla, 1988). Learning and memory can be divided into a number of processes, which include nonassociative and associative learning. Nonassociative learning results in a change in the intensity of an animal’s response over a series of presentations of the same stimulus. After receiving mild stimulation there may be a short-term decrease in the intensity of the animal’s response, which is called habituation. Multiple presentations of the same stimulus will tdrm a long-term decrease in the intensity of the response. However, when an animal receives a strong noxious stimulus there is frequently a short-term increase in the intensity of its response, which is called sensitization. .As with habituation, multiple presentations of

a strong noxious stimulus will form a long-term increase in the response. This

nonassociative form of learning does not involve associative relationships with other

stimuli. Classical conditioning (Pavlovian) refers to the association between two stimuli.

During Pavlovian conditioning a meaningful unconditioned stimulus (e.g. food) is

temporally associated with a neutral conditioned stimulus (e.g. bell, odor, or color).

After formation of the association between the conditioned stimulus and the

unconditioned stimulus, the conditioned stimulus will invoke a conditioned response,

which may or may not be similar to that of the unconditioned stimulus.

One critical aspect in the investigation of molecular elements involved in learning

and memory requires the definition of specific behavioral and molecular approaches.

Learning involves a number of processes, and we chose to focus on associative

learning. Thus, we needed to develop a behavioral paradigm that was specific for the

formation of associative conditioning and which would be adaptable to the time frame

during which molecular events could be isolated and characterized. To determine if

our paradigm produces associative conditioning, as opposed to nonassociative conditioning, specific behavioral control groups were included in the initial study.

These control groups included groups with different temporal relationships between the conditioned and unconditioned stimulus, and it also included groups that experienced each stimulus presented alone or not at all. These control groups are critical to establish that the change in behavior is due to an associative affect. This

type of behavioral analysis had not yet been adequately done in either Apis mellifera

or Drosophila melanogasier for the specific time frame we used tor massed

conditioning (see below).

A second goal was to use this behavioral paradigm together with molecular

techniques to investigate molecular aspects of associative memory consolidation. One

or more nonassociative or associative conditioning trials will set into motion the

formation of memory that passes through different temporal phases. These memory

forms are largely defined by the molecular events that underlie them. Memory

formation is thus a multistage process that involves activation of parallel and serial

molecular pathways that are functionally distinct. Short-term memory (STM) will

produce transient molecular and behavioral changes and is consolidated through

intermediate forms of memories into a more persistent long-term memory (LTM).

STM formation is independent of protein synthesis (Menzel and Muller, 1996; Tully

et al., 1994; Wittstock et al., 1993; Wittstock and Menzel, 1994). Instead it depends

on second messenger-mediated covalent modification of previously synthesized

proteins to change the properties of nerve cells and synaptic connections (Bailey et al.,

1994; Bailey et al., 1996; Zhong and Wu, 1991). Long-term memories will produce

persistent changes in behavior requiring protein synthesis via transcription of

immediate early genes. These early genes activate transcription of late-effector genes

that will be necessary for LTM formation (Abel and Kandel, 1998; Alberini et al.,

1995; Alberini et al., 1994). One gene that has been found to be required for the

formation of LTM is CREB (cAMP-responsive element binding protein). This

transcription factor is phosphorylated by cAMP-dependent protein kinase (PBCA) and binds to promoter CRE sites (cAMP response elements) of genomic DNA activating the transcription of genes required for LTM. Blocking this transcriptional step, via cycloheximide or with CRE oligonucleotides, will block the formation of LTM.

A critical aspect is the time span under which these events unfold. The initial acquisition of information during a conditioning trial can be modified by later experience. A conditioning trial that activates molecular pathways that are involved in the formation of STM may overlap with molecular events involved in intermediate and/or long-term memory processes that were initiated during earlier trials. Therefore, in order to study molecular events involved in the formation of associative memories, we needed to condense the time span over which the animals were conditioned. This condensed time span would restrict the time of molecular expression to a narrow time window, which should allow us to much more clearly relate molecular events to specific types of memory formation. Therefore, by identifying the behavioral

paradigm as associative in nature and condensing the time span of the behavioral

conditioning, we were able to attempt a molecular analysis of events during learning

and memory.

fV/iat are good model animals? Aplysia californica In regard to memory phases, invertebrates are some of the most extensively studied

model animals. These include the mollusk Aplysia californica and the fruit fly

Drosophila melanogasier. In Aplysia californica the behavioral paradigm used to

study both habituation and sensitization is the gill withdraw reflex. During the gill

withdraw reflex the animal is stimulated with a small jet of water that causes reflexive

withdrawal of its gills. The neuronal circuit involved in mediating this gill withdraw

response includes a sensory neuron that synapses directly onto a motor neuron. The

withdrawal reflex habituates after repeated stimulation for a short time. Prolonged,

multiple stimulations will form long-term habituation leading to decrease in the

number of synaptic connections between the sensory neuron and the motor neuron. A stimulus that is strongly arousing will sensitize an animal and increase the response to subsequent stimulation. Multiple stimulations will form long-term sensitization leading to an increase in the number of synaptic connections between the sensory neuron and the motor neuron. Sensitization enhances synaptic transmission by increasing the release of neurotransmitter at the junction between the sensory and motor neurons. The number of and pattern of serotonin (5-HT) pulses induces different phases of PKA activation. This will induce different phases of synaptic facilitation of the synapses between the sensory neuron and the motor neuron of the gill withdraw reflex. Some of the other molecules involved in facilitation of synaptic response \n Aplysia with 5-HT are listed in Table 1.1.

Molecules Memory Stage

Synapsin I STM

Voltage dependent Ca‘" Channels STM

Adenylyl Cyclase STM

Ks (potassium channel) STM/LTM

CREB (cAMP response element binding protein) LTM

PKA (cA^IP-dependent protein kinase A) LTM

MAP kinase (microtuble associated protein kinase) LTM

Ap-uch (Aplysia-ubiquitin carboxy-terminal LTM

hydrolase)

Table 1.1 Summary of molecules involved in facilitation of synaptic response in

Aplysia with 5-HT. Three temporally and molecularly distinct forms of memory are induced by different exposures of serotonin (5-HT) to sensory neurons during the gill withdraw reflex. Stimulation of the sensory neuron with 5-HT results in the activation of adenylyl cyclase (AC), which will increase the concentration of cAMP within the neuron. This increase in cAMP concentration has several consequences, including the activation of cAMP-dependent protein kinase (PKA) via 3’, 5'-cyclic monophosphate

(cAMP). cAMP will bind to the regulatory (R) subunits of the PKA tetramer dissociating the catalytic (C) subunits. The PKA catalytic subunit will phosphorylate existing proteins, including a serotonin-sensitive potassium channel K(s). The K(s)

channel becomes inactivated, decreasing the amount of potassium that enters the cell

during an action potential. This prolonged, broadened action potential is capable of

increasing transmitter release at synapses. This potassium channel Ks is involved in

both the short-term facilitation (STF) and long-term facilitation (LTF) in Aplysia.

The PKA catalytic subunit then migrates into the nucleus while the PKA regulatory

subunit remains bound to the cAMP in the cytoplasm. The PKA regulatory subunit is

degraded by ubiquitin carboxy-terminal hydrolyase (Ap-uch), increasing the

activation time of the catalytic subunits. Therefore this ubiquitin-mediated proteolysis

is essential for formation of LTF (Chain et al., 1999), as this pathway produces a

persistently active PKA by proteolysis of the R subunits. Active PKA and protein

synthesis are both required for formation of LTF. PKA targets include molecular

processes that maintain the increased transmitter release from the sensory neuron, and

CREB to induce gene transcription resulting in LTF, and structural changes of the

neuron.

Other molecules involved in the formation of long-term sensitization in .-Iplysia

include CREB (Bartsch et al., 1998). CREB has three isoforms in .Aplysia CREB la,

CREB lb and CERBlc. These three isoforms have different patterns of cellular

localization, and different activation functions after 5-HT exposure. CREB la is

localized in the nucleus and after phosphorylation is necessary and sufficient for the formation of LTF with or without pulses of 5-HT. CREB lb is localized in the nucleus and dimerizes with CREB I a to repress CREB la activity. CREBlc is localized in the cytoplasm and enhances both STF and LTF.

Induction of LTF differs from STF in the concentration of 5-HT required and the number of pulses of 5-HT required. STF is induced by 1 to 4 pulses of serotonin,

requires covalent modification of existing proteins via PKA and PKC, and lasts up to

15 minutes. LTF is induced by over 5 pulses of serotonin, and requires both

transcription and translation. cAMP dependent gene expression via PKA

phosphorylated CREB and .ApC/EBP is required for the formation of LTF. LTF is

expressed after 24 hours. STF is not required for the tbrmation of LTF and does not

require the activity of ApC/EBP transcription factor (Emptage and Carew, 1993).

In Aplysia the gill withdraw reflex is also used to study associative conditioning.

The animal is stimulated with strong shock to the tail and reflexively withdraws its

gills. This tail shock is the unconditioned stimulus (US) and can be associated with a

weak shock to the mantle, the conditioned stimulus (CS). The reoccurrence of a weak

shock to the mantle will then result in the gill withdraw reflex for a short period of

time. Multiple stimulations will form long-term associative conditioning and increase

the number of synaptic connections between the sensory neuron and the motor

neuron.

Drosophila melanogaster Studies with Drosophila melanogaster have contributed to the understanding of the

molecular biology of learning and memory. Genetic screens have isolated over 24

mutants with disrupted olfactory learning including dunce, rutabaga, amnesiac,

radish, cabbage, turnip, latheo and linotte. These mutants reveal defects in an

olfactory shock avoidance paradigm because they all produce defects in olfactory

memory pathways. Analysis of these mutants has shown that the cAMP signaling

pathway is involved in Drosophila learning and memory. Other mutants involved in

olfactory learning also reveal large-scale changes in biochemical pathways and developmental deficits in structure including minibram, mushroom body miniature, mushroom body deranged, central body defect, ellipsoid body open, central complex deranged, central complex broad and no bridge.

Behavioral paradigms to test these mutants had to be developed in parallel to the molecular work. One paradigm that has been used successfully is the courtship- conditioning assay. This is an associative conditioning assay based on habituation of

male courtship. A male fly when placed with a virgin female fly will be stimulated

due to exposure to the courtship-stimulating pheromone. When this male fly is placed

with another virgin female fly there is no inhibition of courtship. Instead if the male

fly is first placed with a mated female fly there develops courtship inhibition due to

exposure to adversive cues. These male flies will habituate to the odor of mated

females and will not attempt to mate with other female flies (Griffith et al., 1993;

Griffith et al., 1994).

Another paradigm that has also been used to evaluate mutant lines is olfactory

shock avoidance learning (Quinn et al., 1974). During conditioning flies are placed

into a training chamber with an electric grid on the inside. The flies are then exposed

to odor and 12 pulses of 1.5 seconds of electric shock for 60 seconds. Massed

conditioning involves 10 of these training trials in quick succession, with only a few

seconds of rest in between trials. Spaced conditioning consists of 10 training trials

interspersed with 15 minutes of rest between each trial. Conditioned flies are tested in

a T-maze in which the conditioned odor is presented from one arm of the maze and a

novel odor from the other arm of the maze. The numbers of flies that move toward

each of these odors are counted after 2 minutes.

Memory formation in Drosophila involves 5 distinct temporal phases, including an

immediate learning phase (LRN), short-term memory (STM), middle-term memory

(MTM), amnesia resistant memory (ARM), and long-term memory (LTM). These

memories have different time courses, and their activation depends on the

conditioning procedure. LRN is the initial acquisition of the odor-shock association. STM is initially found after learning and will decay within 30 minutes. MTM is downstream of and dependent on STM formation such that if STM is blocked MTM fails to form. MTM is sensitive to cold-shock, but it is unaffected by treatment with drugs that inhibit protein synthesis. In contrast, ARM is insensitive to treatment with cold shock, and like MTM it is also unaffected by blockade of new protein synthesis.

ARM is formed during the first 2 hours post-conditioning and decays within 4 days.

LTM is produced by spaced trials and is dependent on synthesis of new protein. This

form of memory appears 24 hours post-conditioning and is not produced with massed

training trials. ARM and LTM are genetically and functionally independent forms of

long-lasting memory that exist in parallel for several days after spaced training.

These distinct forms of memory are dependent on several parameters of the

conditioning procedure. Presumably, these conditioning parameters affect the

activation of molecular pathways. Aspects of a conditioning procedure that are critical

in defining the type of memory formed include number of training trials, inter-trial

interval (ITI), and reinforcement strategy. For example, by comparing massed and

spaced training in Drosophila, Tully (Dubnau and Tully, 1998) showed that ITI is an

important component. If the ITI is too short, as in a massed training procedure, STM,

MTM and ARM will form but not LTM. Spaced trials with an ITI of 10 minutes

produced robust STM, MTM, and .ARM, and it also produced robust LTM.

Drosophila single gene mutants conditioned in an odor-avoidance paradigm

established a pattern of STM-LTM memory consolidation that can be genetically

disrupted. Acquisition of information can be blocked with the latheo and linotte

mutants (Boynton and Tully, 1992; Dura et al., 1995; Moreau-Fauvarque et al., 1998).

Initial memory is blocked when the cAMP concentration levels are disturbed. Two

mutants that are found in this pathway are dunce, which is a cAMP-specific

phosphodiesterase mutant (Dudai et al., 1976) and rutabaga, a Ca2+-calmodulin-

sensitive adenylate cyclase mutant (Davis, 1996; Levin et al., 1992). These mutants

have normal learning but have faster memory decay during STM (Xia et al., 1991). Medium-term memory (MTM) is deficient in amnesiac mutants (Hitier et al., 1998;

Quinn et al., 1979). The radish mutants when massed or spaced conditioned are lacking in anesthesia resistant memory and one day memory (Folkers et al., 1993).

However, radish mutants have normal LTM following space conditioning when tested 2-7 days post-conditioning. They have normal ARM after one day when treated with cycloheximide and space trained.

As in Aplysia, CREB affects LTM consolidation. In Drosophila CREB isoforms have opposing effects on LTM. The activation iso form dCREB2-a enhances performance by induction of LTM with one training trial. The dominant negative isoform dCREB2-h blocks PKA-responsive transcription and hence LTM (Yin et al.,

1994). The Drosophila dCREB2 gene has both a repressor and an activator form via alternative splicing (Yin et al., 1995; Yin et al., 1994). .Associative conditioning induces both forms. However, the repressor will inactivate faster then the activator.

During multiple spaced conditioning trials the activator form will accumulate and

produce LTM. During massed conditioning trials the activator does not accumulate

and LTM does not form.

Why honey bees? Social interactions of honey bees involve a wide variety of complex behaviors

including orientation and navigation during foraging flights, both chemical

(pheromones) and acoustic (piping/waggle dance) communication, nest building,

rearing of young, food storage, and defense of the colony. Honey bees employ

pheromones to organize colony activity, and they also learn about floral odors in order

to more efficiently locate nectar and pollen resources. Dedicated neuronal pathways

for pheromones result in specific and fairly stereotyped behavioral changes. In

contrast, floral odors that may alter information quality within the lifetime of a bee

have to be learned and stored in STM and LTM within a changing environmental

context.

10 Honey bees have a rich history as experimental organisms in neuroethological research (Hammer, 1997; Hammer and Menzel, 1995; Menzel and Muller, 1996).

Foragers require the ability to seek and find flowers in the field and navigate back to their colony with pollen and nectar loads. These traits are important in acquisition of resources, and we utilized these traits in our conditioning paradigms. The bee associates an odor (conditioned stimulus, CS) with sucrose (unconditioned stimulus.

US) when they are temporally related in such a way that the information provided by the odor predicts the sucrose reward.

The Proboscis Extension Response paradigm (PER) can be used to train honey

bees in a controlled setting. PER conditioning of restrained honey bees via association

of odor with sucrose reinforcement was developed by Kuwabara in 1957 (Menzel et

al., 1996). During conditioning, a sucrose solution touched to a bee’s antenna

stimulates chemoreceptors resulting in the bee extending its proboscis (mouthparts).

An odor that is appropriately paired with sucrose stimulation will cue oncoming

sucrose reward. After one to a few associations with sucrose, the odor alone will elicit

the proboscis extension response (PER) (Hammer and Menzel, 1995). The acquisition

of information correlating the novel odor stimuli with oncoming sucrose

reinforcement is retained effectively for the lifetime of a honey bee. During olfactory

conditioning of the honey bee the proboscis extension response (PER) is typically

used to score the conditioning of an odor (CS) to sucrose reward (US). General

temporal phases of learning with the honey bee have been investigated by utilizing

temperature and chemical disruption of neurological processes (Menzel and Muller,

1996).

Muller showed that associative memory produced in honey bees, like in Aplysia

and Drosophila, can be divided into at least three phases (Muller, 1994). A single

learning trial results in MTM with high recall for over eight hours post-conditioning.

MTM has both sensitive and resistant forms of memory to amnesia cooling. Honey

bees show some form of learning with massed associative conditioning trials after

11 injection with protein synthesis inhibitor cycloheximide (VVittstock et al., 1993;

Wittstock and Menzel, 1994). This agrees with Drosophila data that massed trials form STM, which is independent of protein synthesis.

Given the similarity of memory phases in Apis mellifera to Drosophila melanogaster, it would be useful to study post-conditioning recall after associative conditioning with massed and spaced ITIs. My study directly compared massed and spaced conditioning, and incorporated controls for associative conditioning, which had not yet been done for massed conditioning in honey bees. My first study was to determine how different inter-trial intervals affect post-conditioning recall. The associative conditioning paradigm used either 30 second massed ITl’s or 10 minute spaced ITl’s. We tested groups with extinction trials at either 30 minutes or 180 minutes, and then again after 24 hours post-conditioning. This procedure would allow us to further define the temporal dynamics of memory consolidation in the honey bee after massed-trials conditioning.

Toward a molecular approach to studying learning & memory in the honey bee The end objective was to investigate the molecular basis for learning and memory in the honey bee. The molecular technique we first used to attempt to identify gene expression due to associative conditioning was differential display. Unfortunately this arbitrary primed PCR technique has the disadvantage of generating false positives and was unsuccessful.

Thus we reevaluated our strategy and decided to focus on two different molecular approaches. One approach was to obtain a cDNA from Drosophila melanogaster that has a mRNA expression specific to the mushroom bodies (Han et al., 1996). We used this cDNA to screen our honey bee cDNA library to isolate the honey bee homologue.

This technique lead to the successful isolation of an Apis mellifera cDNA clone that

has sequence homology to aDrosophila melanogaster octopamine cDNA (Han et al.,

1998).

12 We also realized that it would be ultimately necessary to genetically transform the honey bee in order to augment and/or knockout targeted genes. This approach has been used successfully in Drosophila (see above) and the mouse (De Luca and

Giuditta, 1997; Miklos, 1993) to investigate the relationships between various genes involved in learning and memory formation.

Transformation in Apis mellifera During development a honey bee proceeds through programmed developmental

pathways that are the foundation for it’s future behavior (Winston, 1987). One

developmental determinate involves future reproductive status, as commitment to

worker, queen, or drone developmental trajectories takes place in embryonic through

early larval stages. Unfertilized eggs develop into drones, whereas fertilized eggs

develop into either queens or workers depending on external environmental cues. The

amount and quality of food fed to early instar female larvae will regulate

developmental differentiation of larva into either queens or workers.

Our goal was to alter gene expression during specific times and in specific spatial

areas of the organism. We first had to develop specific techniques to successfully

transform the honey bee genome. Some aspects of this technique already exist.

Instrumental insemination of honey bee queens is routinely performed (Cobey, 1997;

Harbo, 1985). Semen can be instmmentally collected from drones and used to

inseminate queens. Honey bee colony management has been well worked out, but

aspects of it had to be adapted for our transgenic studies.

We also needed to develop genetic tools to alter gene expression in a defined

manner. Specifically, we wanted to define techniques to introduce foreign DNA into

the honey bee genome. Transgenic transformation technology from other species (e.g.

injection) was not a suitable technique for the social honey bee. Worker bees will

discard any larvae that have been damaged or perturbed in any way. Furthermore, the

chorion of the honey bee is not as tough as the Drosophila chorion. Attempts to

13 directly inject eggs results in very high rates of lethality. Therefore, we attempted instrumental insemination of queens with sperm that carried linearized foreign DNA.

There is one report of successful use of sperm-mediated transformation in mice

(Lavitrano et al., 1989). However attempts by other labs to repeat this experiment with mice has not been successful (Bimstiel and Busslinger, 1989). Other species that have been successfully transformed utilizing the sperm-mediated technique are discussed in Chapter 2. To augment the possibility of genomic integration via instrumental insemination of foreign DNA, we tested a transposable element. Tel. for transposase-mediated transformation. We also used the green fluorescent protein as a marker to show that transgenic constructs can be expressed.

The various experiments discussed below have detailed the behavioral assays that show that massed forward-paired conditioning using the proboscis extension response

will produce associative conditioning. This condensed time span is appropriate for

future molecular studies of early events involved in the formation of associative

memory. The molecular work discussed below details the potential of utilizing

instrumental insemination coupled with various DNA constructs in transformation of

the honey bee. Furthermore, the requirements of successfully maintaining for long

periods of time manipulated flightroom honey bees are discussed.

14 CHAPTER 2

Massed Conditioniog Produces Associative Learning in the Honey Bee, Apis mellifera

ABSTRACT

Understanding conditions under which an organism correlates temporally related events is critical for establishing a link between behavioral mechanisms of associative learning and molecular mechanisms of acquisition and storage of this information.

We examined the effect of massed-trial conditioning, which involves short (30 second) intervals between conditioning trials, on learning performance using the

Proboscis Extension Conditioning paradigm on worker honey bees. Specifically, we investigated the associative nature of the effect by incorporating several control procedures that have not been evaluated in a massed trials conditioning paradigm. A second goal was to define temporal events that occur during acquisition of information and consolidation as memory. Therefore, we tested recall at different post-conditioning intervals in order to investigate the temporal dynamics of memory consolidation. Acquisition of the Proboscis Extension Response is not dependent on the inter-trial interval (ITI) during forward-paired conditioning. Furthermore, the relatively higher response levels in forward-paired conditioning groups are dependent on a specific temporal and predictive relationship between the odor conditioned

15 stimulus and sucrose reinforcement. Our data also indicate that massed conditioning with 30 second ITI produces robust associative recall over 24 hours.

16 INTRODUCTION

During associative learning an animal establishes a relationship between a neutral conditioned stimulus (CS) and a biologically important unconditioned stimulus (US)

(Rescorla, 1988). Under appropriate conditions this association is then consolidated into memory via a multistage process. These stages represent functionally distinct temporal phases that involve parallel and serial molecular pathways. Initial learning

(LRN) occurs rapidly and consolidates into short-term memory (STM). Short-term memory gives way to intermediate forms of memories before finally being consolidated into stable, more permanent forms of long-term memory (LTM). STM depends on second messenger-mediated covalent modification of existing proteins, which changes the properties of the nerve cells and their synaptic connections (Bailey et al., 1994; Bailey et al., 1996; Zhong and Wu, 1991). While STM does not require synthesis of new proteins (Tully et al., 1994; Wittstock et al., 1993; Wittstock and

Menzel, 1994), LTM is dependent on protein synthesis via activation of specific DNA transcription pathways.

Genetic pathways involved in learning and memory consolidation have been elucidated with Drosophila learning mutants (Tully et al., 1996). Mutation of the gene linotte will block initial learning. The behavioral effect of this mutation can be compared with wild type animals to reveal that mutant animals have reduced initial learning. Other genes from the cAMP pathway {dunce and rutabaga) do not block

initial learning. Instead, they block the formation of short-term memory from the

initial learning phase. STM consolidates into middle-term memory (MTM), which is

17 dependent on STM and can be dissected with specific genetic mutants {amnesiac and

DCO). Long-term memory (LTM) formation is dependent on the previous stage of

MTM and requires protein synthesis. Protein synthesis initiated via the transcription

factor CREB (cAMP-responsive element binding protein) is required for the

formation of LTM. However, the formation of a parallel form of memory, anesthesia

resistant memory (ARM), does not require protein synthesis or CREE activity.

A number of factors during behavioral conditioning can affect information

processing from LRN via STM through to LTM. Presumably, these conditioning

parameters can affect the activation of molecular pathways. One such factor, which

was the focus for the work discussed below, involves the inter-trial interval (ITl).

Inter-trial interval refers to one aspect of the timing over which stimulation is

presented to an animal. In an associative conditioning procedure ITI refers to the time

that elapses between learning trials. A learning trial encompasses the duration during

which the stimuli to be associated are presented. Most learning procedures involve an

optimal inter-trial interval for production of maximal long-term learning performance

(Rescorla. 1988). These inter-trial intervals can be given together during a short space

of time without a rest interval between each of the learning trials, which is referred to

as massed conditioning. Or they can be given over a longer space of time with a rest

interval between each of the learning trials, which is referred to as spaced

conditioning. Tully has shown for Drosophila that STM can be induced with either

massed or spaced conditioning procedures, involving either short or long ITIs,

respectively (Dubnau and Tully, 1998; Yin et al., 1994). However, LTM can be

induced with only spaced conditioning procedures. Spaced conditioning will induce

the formation of LTM via transcription of immediate early genes. These immediate

early genes will activate transcription of late-effector genes, which are necessary for

LTM formation (Abel and Kandel, 1998; Alberini et al., 1995; Alberini et al., 1994).

Honey bees have a rich history as experimental organisms in neuroethological

research into learning and memory formation (Hammer, 1997; Hammer and Menzel,

18 1995; Menzel and Muller, 1996). Foragers require the ability to seek and find flowers in the field and navigate back to their colony with pollen and nectar loads. During foraging honey bees associate the floral odor (conditioned stimulus, CS) with sucrose

(unconditioned stimulus, US) when the two stimuli are temporally related in such a way that the odor predicts the sucrose reward. These traits are important in acquisition of resources, and they are traits that can be utilized in a standardized conditioning paradigm under controlled conditions in the laboratory (Bitterman et al.. 1983;

Menzel, 1983).

Müller has shown that the honey bee has at least three phases of associative memory that are produced following forward pairing of odor and sucrose

reinforcement (Muller, 1994). A single learning trial results in STM and MTM with

high recall for over eight hours when tested post-conditioning. This form of MTM has

been divided into both amnesia-sensitive and amnesia-resistant forms by cooling

experiments of the honey bee brain. Honey bees continue to learn during associative

conditioning trials after injection with the protein synthesis inhibitor cycloheximide

(Wittstock et al.. 1993; Wittstock and Menzel, 1994). This agrees with Drosophila

data that massed trials form STM and MTM. which is independent of new protein

synthesis. Given the similarity of early and intermediate memory phases in Apis mellifera to

Drosophila melanogaster, it would be useful to more extensively evaluate post­

conditioning recall after conditioning with massed inter-trial intervals. Specifically,

massed conditioning in the honey bee has not been appropriately evaluated with the

proper pairing control groups to determine whether it produces predominantly

nonassociative or associative effects. Various pairing control groups are needed to be

able to conclude that a specific reinforcement strategy will produce associative

conditioning. Associative conditioning requires that the development of a conditioned

response be dependent on forward pairing of a conditioned stimulus with a salient

unconditioned stimulus. However, if any type of pairing, or even presentation of the

19 CS or US alone, produces an equivalent increase in the conditioned response it could indicate that the behavioral effect derives from nonassociative conditioning.

This behavioral analysis is critical in order to rigorously evaluate and relate associative conditioning to the molecular processes involved in learning and memory.

Our development in the honey bee of a well-controlled and compressed conditioning paradigm involving massed trials allows us not only to attribute the conditioning to a specific mechanism, but it also allows the behavioral work to be performed over a time scale that does not confound ongoing molecular events. Behavioral paradigms that require more than a few minutes risk confounding early molecular events, initiated by late behavioral trials, with intermediate-term molecular events that were initiated by early behavioral trials.

20 MATERIALS AND METHODS

Honey bee subjects Honey bee subjects were maintained at the Rothenbuhler Honey Bee Research

Laboratory at The Ohio State University. Foraging worker honey bees were captured on an outward flight in glass vials and chilled in an ice-water bath until they ceased movement. Subjects were individually secured into metal harnesses by tape along the top of the thorax, leaving the head, with the mandibles and antenna, free to respond to

stimulation. The subjects were fed I drop of 2.0 M sucrose with a Gilmont

micrometer syringe and then left two hours to acclimate to restraints.

Proboscis Extension Conditioning (PEC) The Proboscis Extension Response (PER), the extension of the proboscis after

exposure to sucrose or conditioned stimuli, was used to score individual subjects

during and after olfactory Proboscis Extension Conditioning (PEC). PEC uses the

proboscis extension within the training assay. This assay has been often used in

behavioral studies (Menzel, 1983); see also http;//iris.biosci.ohio-state.edu/honeybee).

During PER conditioning an individually harnessed subject was placed into the

training arena through which a continuous airflow was drawn past the subject and into

an exhaust hood. Thirty seconds later the subject was presented with conditioning

stimuli in a manner that depended on the specific treatment condition (see below), k

trial consisted of one of several procedures, all but one of which involved stimulation

of the subject with odor and/or sucrose. To prepare syringes used for odor delivery, 3

pi of pure odor was applied onto strips of filter paper, which were then placed into a

new 25 cc plastic syringe. Odor was manually delivered to the subject for 4 sec by

21 injecting it into the air stream being drawn across the subject’s head and antennae.

Hexanal was used for all conditioning and extinction trials in experiments outlined below. The sucrose-water solution (2.0 M) used as the US in all experiments was delivered from a Gilmont micrometer glass syringe. The sucrose drop ( 1 pi) was first touched to the subject’s antenna, which then elicited the PER. If the bee extended its proboscis it was fed off of the sucrose drop for the remainder of the conditioning trial.

Acquisition conditioning during the mass-training conditioning consisted of a total

of sixteen trials with a thirty second inter-trial interval (ITI). Mass trained subjects

were moved once into the training arena, then thirty seconds later were exposed to the

first conditioning trial. Successive trials were presented at every 30 seconds. The

training schedule was pseudorandomized such that all groups were trained in .ABBA

BAAB ABBA BAAB pattern, with each treatment group receiving a different

stimulus condition for trials A and B (Table 2,1) (Smith et al., 1991). The daily

sequence of conditioning groups was randomized.

Treatment Groups There were 6 different pairing conditions used to evaluate the associative nature of

massed conditioning (Table 2.1). These groups differed in their stimulus presentation

of the odor (CS) and/or the sucrose (US). The Forward-paired treatment group

received on ‘A’ trials both odor and sucrose together such that onset of odor preceded

the presentation of the sucrose. There was a 1 second overlap of the odor and the

sucrose presentation. This forward-paired group received no stimulus presentation on

‘B’ trials, referred to as ‘placement-only’ trials. The Backward-paired treatment

group also received on A trials both odor and sucrose presentation, but the order of

stimulation was reversed. The sucrose preceded the presentation of the odor with a 1

second overlap between them. This group also received placement-only during B

trials. The Unpaired treatment group received odor-only presentation during A trials,

and sucrose-only presentation during B trials. This group did not receive any

T) placement-only trials. These three groups, the forward-paired, backward-paired and unpaired, all receive the same number of CS and US presentations, but differed in how those stimuli are temporally related. Of the final three groups two received stimulus presentation on A trials and placement-only in B trials (Table 2.1). The

Odor-only and Sucrose-only groups receive one or the other stimulus during the trials, but they were never exposed to both. The Odor-only and Sucrose-only groups also received placement-only during the B trials. The Placement-only group received placement-only on both A and B trials, not receiving either odor or sucrose during the

conditioning.

Two different sets of the six treatment groups (Table 2.1) were conditioned in

parallel in order to evaluate the level of recall at different post-conditioning test

intervals. One set of six treatment groups was tested 30 minutes after the final

conditioning trial, while the second set of six groups was tested 180 minutes after the

final conditioning trial. The extinction test procedure involved exposure to the CS

(odor only) on 2 trials separated by a 2 minute ITl. All groups were additionally tested

in an identical manner 24 hours post-conditioning.

Another group included in our study of associative conditioning was a spaced

forward-paired conditioning group. This forward-paired group was given 16

pseudorandomized conditioning trials (ABBA BAAB .ABBA BAAB) with an inter­

trial interval of 10 minutes. Subjects in this group received both odor and sucrose

together such that onset of odor preceded the presentation of the sucrose. There was a

I second overlap of the odor and the sucrose presentation. This forward-paired group

received no stimulus presentation on ‘B’ trials, referred to as ‘placement-only’ trials.

Due to conditioning parameters, i.e., the ITI of 10 minutes, subjects in this group were

moved into the training arena, received their training trial, and then moved back to the

wait area until the next training trial.

23 Treatment Group Designations Trial A Trial B

Forward Paired F CS+/Sucrose Place

Backward Paired B Sucrose/CS+ Place

Unpaired U Sucrose CS+ Sucrose Only s Sucrose Place Odor Only 0 CS+ Place

Placement p Place Place

Table 2.1 Summary of treatment conditions to test for nonassociative and

associative conditioning effects.

24 RESULTS

Effect o f pairing condition on post-conditioning responses day 1 The effect of different pairing conditioning procedures (Table 2.1) on response levels after the completion of massed trials conditioning is shown in Fig. 2.1A and

2.IB. For the groups tested 30 minutes post-conditioning (Fig. 2.1 A), the si.x different conditioning groups differed in terms of the percentage of subjects in each group that

responded with proboscis extension 0, 1 or 2 times during the two extinction test

trials (.v2=54.9, df=10, n=119, p< 0.0001; F>B, Ü, S, O, P p<.01). Specifically,

subjects in the forward-paired group responded more often than did subjects in the

remaining five groups (see legend for post-hoc pairwise comparisons). Subjects in the

backward-paired and unpaired groups responded least often.

The same qualitative pattern of differences across the six different conditioning

groups was observed 180 minutes post-conditioning (Fig. 2.IB). As before, the six

different conditioning groups differed in terms of the percentage of subjects that

responded with proboscis extension 0, 1 or 2 times during the two extinction test

trials (A-=2l.O, df =10, n=106, p<0.02; F> B, U, S. 0. P p<0.05). Subjects in the

forward-paired group responded more often than did subjects in the remaining four

groups (see legend for post-hoc pairwise comparisons). Subjects in the backward-

paired and unpaired groups responded least often. Note that between the 30 and 180

minute test times there was an overall increase in response levels across all but the

forward-paired group, although the same qualitative pattern of within-group

differences held.

25 Figure 2.1 A.

100% c o ÎÂ 80% C Q) X 60% Ui (A 40% : Ô V) O a 20% O 0 % I B L B L s o S O

1

N umber Of Responses to Odor (max=2)

2.1 B.

100% c o ■(J5 80% c o 60% X til I M 40%

o 20% 13 O 0 % H i F B I S O P F B L S O P F B f S O P

Number Of Responses to Odor (max=2)

Figure 2.1 Effect of pairing condition on post-conditioning responses day 1

2 6 Figure 2.1. Effect of pairing condition (Table 2.1) on post-conditioning responses of massed trials groups on day 1. The figure shows the percentage of subjects that responded with proboscis extension 0, 1 or 2 times across the 2 extinction test trials. (A) Subjects were tested 30 minutes after completion of the massed trials procedure. In addition to the overall chi-squared comparison

(see text), we performed a post-hoc pairwise comparison between the forward- paired (F) group and each of the remaining 5 groups. The distribution of responses in the F group differed from B (A-=35.7. df = 2. p<0.000l), U

(A2=35.1, df =2, p <0.0001), S (A^-22.9, df =2, p <0.0001), 0 (.V^=14.3, df

=2, p<0.0008), and P (.T-=16.6, df =2, p< 0.0003). Sample sizes are as follows:

F=20; 8=19; U=20; S=20; 0=20; P=20. (B) Subjects were tested 180 minutes after the completion of the massed trials procedure. Post-hoc pairwise comparison was made between the forward-paired (F) group and each of the remaining 5 groups. The distribution of responses in the F group differed from

B (.^2=11.9, df =2, p<0.003), U (A^=14.3, df =2, p<0.0008), S (.V-=6.5, df = 2. p<0.04), 0 (.T-=4.9, df = 2, p<0.09), and P (.ï-=l.9, df =2, p<0.38). Sample sizes are as follows: F=20; B=20; U=20; S=20; 0=12; P=14. Effect of pairing condition on post-conditioning responses day 2 The effect of different pairing conditioning procedures (Table 2.1) on response levels on day 2, 24 hours after conditioning, is shown in Fig. 2.2A and 2.2B. Groups in Fig. 2A had been tested 30 minutes post-conditioning on day 1. The six different conditioning groups differed in terms of the percentage of subjects that responded with proboscis extension 0, 1 or 2 times during the two extinction test trials on day 2

(.V-=28.4, df=10, n=lGO, p<.0016; F> B, U, S. 0, P p<0.05). Subjects in the forward- paired group responded more often than did subjects in the remaining five groups (see

legend for post-hoc pairwise comparisons). As before, subjects in the backward-

paired and unpaired groups responded least often.

A similar qualitative pattern of differences across the six different 180 minutes

post-conditioning groups (Fig. 2.IB) was observed on day 2 (Fig. 2.2B). The six

different conditioning groups differed in terms of the percentage of subjects in each

group that responded with proboscis extension 0, 1 or 2 times during the two

extinction test trials (.V-=29.2, df=10, n=96, p<.0012; F> B, U, S p<0.01). Subjects in

the forward-paired group responded more often than did subjects in the backward-

paired and unpaired groups (see legend for post-hoc pairwise comparisons). In this

case, the difference between the forward-paired (F), odor-only (O) and placement-

only (P) groups was not significant.

28 Figure 2.2 A.

100%

80%

60%

w 40% N o .Q 20% I O v_ I CL 0 % I e B L s o p p B L s o p F B L S ü P

Number Of Responses to Odor (max=2)

2.2 B.

100% c o ■« c 0) X lU (A â o n o

B L S O P

Number Of Responses to Odor (max=2)

Figure 2.1 Effect of pairing condition on post-conditioning responses day 2

29 Figure 2.2. Effect of pairing condition (Table 2.1) on post-conditioning responses of massed trials groups on day 2. The figure shows the percentage of subjects that

responded with proboscis extension 0, 1 or 2 times across the 2 extinction test

trials. (A) Subjects that were initially tested 30 minutes after completion of the

massed trials procedure on day 1. We performed a post-hoc pairwise comparison

between the forward-paired (F) group and each of the remaining 5 groups. The

distribution of responses in the F group differed from B (.V-=20.9, df = 2,

p<0.0001), Ü (.V-=7.9. df = 2, p<0.02), S (.Y-=8.8. df =2. p<0.012), O (.V-=10.0, df

= 2, p<0.007 ), and P (.V-=7.5, df =2, p<0.023). Sample sizes are as follows: F=20;

8=18; U=20; S=18; 0=12; P=12. (B) Subjects that were initially tested 180

minutes after the completion of the massed trials procedure on day 1. Post-hoc

pairwise comparison was made between the forward-paired (F) group and each of

the remaining 5 groups. The distribution of responses in the F group differed from

B (W-=20.5, df =2. p<0.0001), U (X-=\AA, df =2, p<0.0008), S (A-=11.9. df = 2.

p<0.003), 0 (.V-=3.2, df = 2. p<0.2), and P (.V-=3.3, df = 2, p<0.2). Sample sizes

are as follows: F=19; B=17; U=19; S=20; 0=10; P=12.

30 Effect of inter-trial intervals on post-conditioning responses The effect of massed and spaced forward-paired conditioning procedures on post­ conditioning response levels is shown in Fig. 2.3A and 2.3B. These three different conditioning groups did not differ in terms of the percentage of subjects in each group that responded with proboscis extension 0, I or 2 times during the two extinction test trials on day I {X -= \.l, df=4, n=60, p<0.79 NS). A very high percentage (approx.

95%) of animals responded to the CS in all groups. In addition, the three groups also responded at equivalent levels 24 hours post-conditioning (Fig. 2.3B). As before, the three conditioning groups did not differ in terms of the percentage of subjects in each group that responded with proboscis extension 0. 1 or 2 times during the two extinction test trials on day 2 (.V-=3.9, df=4, n=58, p<0.41 NS)

31 Figure 2.3 A

100% c O ■55 80% c V X 60% LU (/) 40% N o .0 20% o 0 % l I-: ' v - M y X ' / l 30M 180M 30S 30 M 180M 30S 30M 180M 30S 0 I N umber Of Responses to O dor (max=2)

2.3 B

c 100% 0 1 80% Q) X 60% LU .2 40% U (A O 20% o ^ 0 % Ou 30M tSOM 30S 30M (80M 30S 30 M 180M 30S 0 1 ■>

Number Of Responses to Odor (max=2)

Figure 2.3 Effect of inter-trial intervals on post-conditioning responses

32 Figure 2.3. Effect of inter-trial intervals on post-conditioning responses (Table

2.1) of massed and spaced forward-paired conditioning trials groups. The figure

shows the percentage of subjects that responded with proboscis extension 0, 1

or 2 times across the 2 extinction test trials. (A) Subjects from forward-paired

massed (F30m) conditioned group and forward-paired spaced (Fs) conditioned

group were tested 30 minutes post-conditioning on day 1. Subjects from

forward-paired massed (FI80m) conditioned group were tested 180 minutes

post-conditioning on day 1. Sample sizes are as follows: F30m=20; F180m=20;

Fs=20. (B) Subjects from all three groups were tested again on day 2, 24 hours

after conditioning. Sample sizes are as follows: F30m-20; Fl80m=19;

F30s=19.

33 DISCUSSION

We have found that a massed forward-paired conditioning procedure will produce robust associative memory that lasts at least 24 hours. We base this claim on the higher post-conditioning response level of the forward-paired groups vs. other stimulus treatment groups. A change in an animal’s response level post-conditioning may be due to either nonassociative (habituation and sensitization) and/or associative

effects. Habituation will decrease an animal’s response level after repeated exposure

to a non-reinforced stimulus. Sensitization will increase an animal’s response after

stimulation with a salient stimulus. Unlike these nonassociative effects, associative

conditioning only occurs when an animal establishes relationships between stimuli.

Associative conditioning can produce either an increase or a decrease in an animal’s

response. Therefore, an increase or a decrease in an animal’s response level could be

due to either nonassociative and/or associative mechanisms. For our goal to elucidate

molecular events involved in the formation of associative memory it is critical to

establish to what extent each mechanism contributes to this behavioral response

following our conditioning procedure. Condensing the training time into a short

period by using a short inter-trial interval was important to be able to attempt a

molecular analysis. This avoids confounding the first few conditioning trials, and

hence the molecular signaling pathways that they would activate, with later

conditioning trials and the molecular activity they generate.

We use several control groups to establish the associative nature of massed forward-

paired conditioning with the honey bee proboscis extension response paradigm. To

34 determine that this conditioning produces an associative effect, we had to show that the change in response post-conditioning is dependent on both a specific temporal and predictive relationship between the odor (CS) and sucrose (US). Specifically, the change in response is found only when two conditions have been met. First the US is contingent on CS presentation, meaning that the US must be presented following the

CS. The animal’s response would decrease whenever the US occurs without prior presentation of the CS. Second, the two stimuli must be temporally contiguous, meaning that the US presentation must occur within a fi.xed time frame following presentation of the CS.

To establish the associative nature of massed forward-paired conditioning with the honey bee proboscis extension response paradigm, we included treatment groups in which contiguity and contingency between the CS and US were altered. Subjects in these groups experienced the same number of CS and US presentations, as did subjects in the forward-paired group. The backward-paired group disrupted contingency, in that the US preceded the CS, while the unpaired group disrupted contiguity between the CS and the US. Both of these treatment groups failed to respond at the level that was observed in the forward-paired group. This supports our claim that massed forward-paired conditioning is an associative effect.

To determine the level that nonassociative mechanisms may contribute to

responding in the forward-paired groups we included additional treatment groups that

experienced each stimulus only, the odor-only (CS), sucrose-only (US), and

placement-only (P). These groups tended to have a higher response level post­

conditioning in comparison to the backward-paired and unpaired groups. In general

they responded with lower rate of proboscis extension during the extinction trials vs.

the forward-paired group.

We performed a post-hoc pairwise comparison between the forward-paired group

and each of the remaining 5 treatment groups. On day 1, the distribution of responses

in the forward-paired group differed from all five treatment groups (backward-paired.

35 unpaired, sucrose-only, odor-only and placement-only) for both the 30 minute and the

180 minute groups. Note that between the 30 minute and 180 minute test times there was an overall increase in proboscis extension response levels across all but the forward-paired group, although the same qualitative pattern of within-group differences held. This general increase could have been due to an increase in motivation due to hunger over the 3 hours between the end of conditioning and testing.

We also tested the distribution of responses after 24 hours. From the groups tested after 30 minutes on day 1 the forward-paired group responses differed from all five treatment groups (backward-paired, unpaired, sucrose-only, odor-only and placement- only). For the groups tested after 180 minutes on day 1, the forward-paired group

responses differed from only three of the five groups - the backward-paired, unpaired,

and sucrose-only. In this case, the difference from the odor-only (0) and placement-

only (P) groups were not significant, which could reflect an increase in motivational

state in the latter two groups as neither group had received any sucrose during

conditioning. However, the lack of significant differences in these cases does not alter

the associative nature of the massed forward-paired conditioning procedure, because

the primary comparison is among groups that received the same number of stimulus

exposures, the backward-paired and the unpaired groups.

Some studies have reported a decrease in consolidation of memory after massed-

trials conditioning (Dubnau and Tully, 1998; Gerber et al., 1998). First, some kind of

interference of later trials may prevent animals from processing the information about

the relationship between the CS and the US from earlier trials. Second, the ITIs may

be so close together that they generate some level of backward inhibitory learning.

Our data over 24 hours fails to support either of these events. First, the massed 30

second ITI forward-paired groups (tested at 30 minutes, 180 minutes, and then 24

hours post-conditioning) are not significantly different from the spaced 10-minute ITI

forward-paired group (tested at 30 minutes and then 24 hours post-conditioning).

36 There were not any significant differences found in LTM between these three forward-paired groups. Second, we specifically included the backward-paired control groups to account for the possibility of the backward inhibitory learning. There are significant differences between the massed 30 second ITI forward-paired groups and the massed 30 second ITI backward-paired groups.

During the acquisition of information the intertrial interval is a critical parameter in conditioning studies (Rescorla, 1988). One aspect of our study was to test the effect of inter-trial interval in the PER conditioning paradigm by comparing massed conditioning to spaced conditioning. It is important to note that massed and spaced

forward-paired groups also differed in a few other parameters. The massed-

conditioned groups were placed into the training arena and were not moved between

the conditioning trials. While the spaced-trained groups were placed into the training

arena, given a conditioning trial, and moved back to a holding area until the next trial.

Yet the forward-paired conditioning groups used in our study did not differ in the

level of response across any test interval.

Finally the massed forward-paired conditioning paradigm is ethologically relevant

for the honey bee. The honey bee’s natural behavior is to visit many flowers during a

foraging flight and then return with the nectar and pollen load back to the colony.

Flower visits as such are within a few seconds to minutes of one another, which is

approximately equivalent to our massed conditioning paradigm. Thus the honey bee is

an ideal organism to mass condition in investigations of both molecular and

ethological aspects of learning and memory.

37 CHAPTER 3

Inheritance of Plasmid DNA via Sperm-mediated Introduction in the Honey Bee,Apis mellifera

ABSTRACT

A full understanding of the molecular basis of learning and memory will depend upon the functional testing of implicated genes in transgenic honey bees. As a first step towards this goal, strategies to transform the genome of honey bees with foreign

DNA plasmids were investigated. Linearized plasmids were introduced with semen during instrumental insemination of virgin queen honey bees. Larvae from a few such queens were subsequently identified as positive for the plasmid via PCR amplification of a plasmid fragment and fluorescent expression of green fluorescent protein encoded by the plasmid. These transgenic lines transmitted the transgene for two generations, demonstrating heritability. However, there was no evidence of integration of the construct into the honey bee genome as determined by genomic

Southern analysis.

38 For this reason, the use of a transposable element was explored. These experiments tested the utility of a potential universal gene vector for transformation of the honey bee genome. The Tel transposon was originally isolated from C elegcins and is well characterized. The data indicate sperm-mediated introduction of the Tel element via instrumental insemination results in PCR-positive progeny in the FO and the FI generations. Ligation-mediated PCR and plasmid rescue from genomic DNA isolates confirmed the persistence of the Tel element through the generations. However, recombination of the Tel element into the honey bee genome was not detected.

Therefore we consider it likely that the Tel element is maintained but persists as an

extrachromosomal element.

39 INTRODUCTION

Transgenic technology has become an important technique for exploration of the relationship between information processing on a molecular level and the natural behavioral repertoire of an organism. However, appropriate transformation techniques have not yet been fully developed for a wide array of organisms, including honey bees

(Handler and O'Brochta, 1991; Miklos, 1993). There are a few reasons that progress has been slow, particularly in animals such as the honey bee. First, transformation has been hampered by the inability to physically introduce molecular constructs into

honey bee eggs without the process causing lethality. Second, honey bees are highly

social. Worker bees closely monitor developing progeny (Rothenbuhler, 1979) and

any perturbations during development will result in the removal and destaiction of the

progeny. Thus the researcher has the possibly prohibitive option of artificially rearing

experimental progeny (Rothenbuhler, 1964). Finally, although there are

transformation vectors that are mobile in genomes of organisms outside of

Drosophila (Handler et al., 1998; Sarkar et al., 1997), none have been identified as a

universal transformation vector for invertebrates.

Nevertheless, there are compelling reasons to transform the genome of honey bees.

Honey bees have a rich history as an experimental organism for basic research in

social behavior (von Frisch, 1967), neuroethology (Fahrbach and Robinson, 1995),

and behavioral genetics (Page et al., 1998; Page and Robinson, 1991). They exhibit

complex social interactions, including, among other behaviors, dance communication

to indicate food resources (von Frisch, 1967), hormonally regulated behavioral castes

40 of worker bees (Robinson et aL, 1997), and well developed learning capabilities (Page et al., 1998). Behavioral genetic analyses have shown that behaviors such as pollen and nectar collection (Robinson and Page, 1989), colony guarding (Page and

Robinson, 1991), and colony hygienic behavior (Robinson and Page, 1988), all have genetic components that influence the behavior of individual worker bees

(Rothenbuhler, 1964). The honey bee is also economically important for agriculture.

The demands of increased crop production coupled with stresses of disease and pest infestation have increased the potential value of directly transforming the honey bee genome (Rothenbuhler, 1979).

The standard technique for introduction of plasmid DNA into an organism

involves microinjection. In mice the DNA is injected into an early stage pronucleus

(Palmiter and Brinster, 1986), in zebrafish the injection is into one- to two-cell

developmental stage (Fadool et al., 1998). or in fruit flies the injection is into eggs

(Rubin and Spradling, 1982). With the thought of introduction of manipulated

embryos into the in vivo rearing environment, we initially tried to transform eggs via

microinjection of plasmid DNA directly into eggs and liposome-mediated transfer.

Although we obtained some evidence of transient expression of the construct with

both techniques (H.J.F. unpubl.), we still were faced with the difficult task of rearing

manipulated honey bee embryos into adults given the high rejection rate that workers

show toward manipulated embryos.

For honey bees these techniques required manipulation of embryos in ways that

limited their réintroduction into the colony. This arises because social insects actively

feed and maintain embryos and developing larvae within the colony (Fahrbach and

Robinson, 1995; Page and Robinson, 1991). During close contact with eggs and

larvae workers detect perturbations in normal development. Injury or developmental

abnormalities frequently lead to destruction of a larva by nurse worker bees.

Therefore, use of microinjection would necessitate artificial {in vitro) rearing of

injected larvae into queens. Artificially rearing experimentally manipulated embryos

41 is possible (Rembold and Lackner, 1981; Shull and Dixon, 1986), but it is difficult to rear fully functional reproductive queens outside of a natural colony environment. As it would therefore be advantageous to develop a technique that utilizes in vivo rearing conditions, we tested the potential to transform honey bees by sperm-mediation.

We report here on the use of sperm for introduction of plasmid DNA (Francolini et al., 1993; Lauria and Gandolfi, 1993; Lavitrano et al., 1989; Lavitrano et al., 1997;

Lavitrano et al., 1992; Lavitrano et al., 1997; Shamila and Mathavan. 1998; Zani et al., 1995). Honey bees have a reproductive cycle that is accessible to experimental manipulation at many stages, and instrumental insemination is a standard procedure

(Cobey, 1997; Harbo, 1985; Laidlaw and Page, 1997). Instrumental insemination involves the artificial collection of drone semen followed by transferal of this semen into the oviducts of a virgin honey bee queen. The sperm then migrate into the spermatheca where it is stored until the queen uses it to fertilize her eggs. In our procedure linearized plasmid DNA is manually mixed with the collected semen and then introduced into the queen by instrumental insemination. We show that this technique of instrumental insemination allows for genetic transformation of the honey bee via semen-mediated transfer of linearized plasmid DNA. Use of sperm-mediation

for introduction of plasmid DNA takes advantage of the in vivo colony rearing conditions for transgenic progeny and it is not physically damaging to honey bee queens. This sperm-mediated transformation procedure has been controversial. It has been

used successfully to transform sea urchin blastulae (Arezzo, 1989), and rabbit eggs

(Brackett et al., 1971). Bombyx mori has also been transformed with injections of

DNA into the testis of larvae (Shamila and Mathavan, 1998). However sperm-

mediated transformation has been criticized for being difficult to replicate in mice

(Brinster et al., 1989).

As a universal transformation vector for arthropods has not yet been identified, we

needed to test possible vectors for transpositional activity in honey bees.

42 Transformation of D. melanogaster with P elements is a powerful genetic tool for manipulation of the genome (Rubin and Spradling, 1982; Spradling and Rubin, 1982).

Unfortunately P elements require host specific factors in the germline for activity. As they are host dependent for activity they are not suitable as a universal vector for arthropod transformation (O'Brochta and Atkinson, 1996). Overall there are lower levels of P element transposition within other Drosophila species but it is not successful outside ofDrosophila (Berg and Howe, 1989). Tc 1-like elements (TLE’s) were originally isolated in C. elegans (Liao et al., 1983) and the mechanism of transposition and insertion site sequence specificity have been extensively studied for the Tel element (Vos et al., 1996; Vos and Plasterk, 1994; Vos et al., 1993). The Tel element is able to transpose in species other than C. elegans (Gueiros-Filho and

Beverley, 1997; Luo et al., 1998) and is active in both soma and germline tissue. This element has been constructed into a two component transposon system composed of a nonautonomous transposon element and its transposase (Vos et al., 1996). One component is the Tel and the second is its Tel A transposase. We decided to utilize the Tel element in the hope that it would be effective in vivo to transform the honey bee genome.

In summary, we address two of the three issues broached above. First, we developed

a means to introduce DNA plasmids into the honey bee. Second, we developed the

techniques necessary for long-term maintenance of transgenic lines in a controlled

setting. Finally, although integration of plasmids nor transposition of the Tel element

into the honey bee genome was not detected, we consider the possibility that these

elements persist extrachromosomally.

43 MATERIALS AND METHODS

Honey Bee Maintenance Honey bee stocks were maintained at the Rothenbuhler Honey Bee Research

Laboratory at The Ohio State University. Camiolan queen honey bees used in the inseminations were reared in queen rearing colonies and maintained in individual wire mesh cages inside queenless colonies until instrumental insemination. Drones were available from a large number (ca. 200) of camiolan colonies. We isolated colonies containing putatively transgenic honey bee queens and their progeny in an indoor flightroom. This room consists of a metal shed with light from 3 natural sunlight fluorescent bulbs on a regime of 16 hours of light to 8 hours of dark. As an additional precaution there was a fme-mesh tent surrounding the colonies inside the shed. The colonies used to maintain the queens were small 2- or 3-frame colonies.

These flightroom colonies were fed a sucrose/Fumidil® solution, water, and pollen/sugar patties ad lib.

Instrumental Insemination o f Virgin Honey Bee Queens Semen was collected from 8 to 15 drones for instrumental insemination of each

virgin honey bee queen. We used 6 pi of freshly collected semen manually mixed

with a pipette tip with 2 pi linearized DNA plasmid at 400 ng/pl for instrumental

inseminations of primary queens (Table 3.1; PO). Queens of subsequent generations

(Table 3.1; FO and FI) were inseminated with 6 pi of untreated semen. Each queen

was anesthetized with CO 2 during the procedure and instrumentally inseminated in a

Schley instrument using a Harbo syringe apparatus with 8 pi of semen/DNA for PO

queens or untreated semen for FO and FI queens. To identify queens each was marked

44 with either a dot of paint or a plastic numbered queen tag on the thorax, and all queens were wing clipped immediately after instrumental insemination. Inseminated queens were maintained overnight in wire cages in a queenless honey bee colony.

Then 24 hours later all queens were given a second CO] treatment for five minutes to induce egg laying. Subsequentally, each queen was introduced into a small colony of young worker bees in the indoor flightroom.

Generation*

PO Primary Queen inseminated with semen mixed with DNA plasmid, PO queen

FO Founder PO progeny reared as a queen and inseminated with untreated semen, FO queen

FI FI FO progeny reared as a queen and inseminated with untreated semen, FI queen

F2 F2 F 1 progeny

Table 3.1. Nomenclature used to refer to generations of honey bee queens, their

progeny, and the next generation of queens throughout text.

45 Honey Bee Queen Rearing The offspring of PO queens were placed (grafted) into wax queen cell cups and put into a queenless colony of young nurse worker bees to be reared into virgin founder queen bees. Once capped these queen cells were isolated individually in a wire queen cage and placed into an incubator until emergence. After emergence these virgin queens were maintained in queenless colonies until instrumental insemination. This founder generation (FO) of putative transgenic queens would have acquired the plasmid DNA from the semen mixed with linear DNA plasmid used in the instrumental insemination of the PO queen. The following generations (FO, FI. and

F2) were collected at various developmental stages (eggs, larvae, and pupae) and tested for presence of the various plasmids via PCR amplification of a plasmid

fragment from genomic DNA isolates and/or fluorescent expression of green

fluorescent protein encoded by the plasmid.

Finally, some progeny were not assayed for plasmid but were instead reared into

honey bee queens. We refer to these queens via a lineage nomenclature (Table 3.1).

The first number refers to the wild type PO queen instrumentally inseminated with

linear DNA plasmid and semen. The second number refers to PO progeny, which

comprises the FO generation. Some of these progeny were reared into FO queens. The

third number refers to FI progeny of these FO queens. Note that FI and F2 progeny

that test positive for plasmid with PCR amplification of a plasmid fragment from

genomic DNA isolates would have inherited the construct from the queen, as these

queens were instrumentally inseminated with untreated semen (not mixed with

plasmid DNA).

46 Plasmids In our transformation protocol we utilized different plasmid DNA combinations for instrumental insemination of wild type PO honey bee queens. First, the phGFP-S65T plasmid (Clontech) was linearized with BamHI, for queen inseminations. Second, we used a phGFP-S65T plasmid that included the Drosophila hsp70 inducible promoter that had been subcloned into the S a d and H indlll sites. We refer to this construct as phspGFP. This plasmid was also linearized with Bam H l for queen inseminations.

Third, we used the Tel element and its transposase Tel A, for queen inseminations

(pRP466/pRP470) (Vos et al., 1996; Vos et al., 1993). Fourth, we subcloned the

Drosophila hsp70 promoter and the GFP marker in the Sacl-BaniHI fragment of

phspGFP plasmid into the A pal site within the Tel element. It is now referred to as

p466-hspGFP. We used this p466-hspGFP with the Tel A transposase, pRP470. Fifth,

we cloned the Drosophila hsp70 promoter into the pCR^II vector (Invitrogen®)

plasmid to drive the TclA transposase of pRP470. It is now referred to as pCRhsp-

470. We used the p466-hspGFP with this pCRhsp-470.

.All of these plasmids were linerized, precipitated, and resuspended in bee saline

solution (Harbo, 1985) to a concentration of 300 to 400 ng/pl. For the instrumental

insemination of wild type primary honey bee queens (PO) we used 2 pi of the

linearized plasmid mixed with 6 pi of honey bee drone semen. The honey bee Tris-

buffered saline that was used for mixing and storing semen is as follows: 100 ml.

distilled water, l.ll g. sodium chloride, O.IO g. glucose, 0.01 g. L - lysine. 0.01 g. L -

arginine, 0.01 g. L - glutamic acid, 0.329 g. Trizma HCL, and 0.329 g. Trizma base.

The pH was adjusted to pH 8.6, followed by bacteriological filtering with a pore size

of 0.2 um and autoclaving. After cooling the solution to room temperature the

antibiotic (0.25 g dihydrostreptomycin) was added.

47 Genomic DNA Isolation Genomic DNA samples were isolated from adult honey bee queens, workers and late stage pupae by fast freezing in liquid nitrogen and grinding with a mortar and pestle. The mortar and pestle was washed in bleach and autoclaved between each sample isolation. These samples were incubated with genomic DNA extraction solution of 50 mM Tris, 10 mM EOTA. and .2% SDS with proteinase K (. 1 mg/ml) at

65° for 2-5 hours. This crude extract was then phenol extracted, phenol/chloroform extracted, and chloroform extracted. The DNA was precipitated with 1/ 10th volume of 3M NaAc. and 21/2 volume EtOH, centrifuged and the pellet washed with 75%

EtOH. The genomic DNA sample was resuspended in TE and RNAse. Honey bee eggs, and larvae were ground in an eppendorf tube using a small pellet pestle and the genomic DNA was extracted with QIAGEN'® QLAamp Tissue kit via their procedures

for the spin column.

PCR Reactions Genomic DNA samples were isolated from various developmental stages of

progeny (FO, FI, or F2) to determine if these progeny were positive for containing

plasmid DNA via PCR amplification of a plasmid fragment. PCR amplifications with

phGFP-S65T specific primers (GFP-1210) 5’ AAA GGC ATT CCA CCA CTG CTC

3’ and (GFP-1760) 5’ GGA GGG CAT CGA CTT CAA GGA G 3’ and would

amplify a fragment of 550 bp in construct positive samples. These primers amplify

within the GFP region and were used with genomic DNA isolates from progeny that

putatively contain either phGFP-S65T or phspGFP plasmids and genomic DNA from

control animals. PCR primers specific for the Tel element within pRP466 are

(pRP466-227) 5’ ACG CAC TCT GTT TGT TGC ACT G 3’ and (pRP466-680) 5’

GGT TTC TTG ACT GGC TTT CGT C 3’ and would amplify a fragment of 453 bp

in construct positive samples. PCR conditions, using Ventr (exo-) DNA polymerase

from New England Biolabs, Inc., for the first roimd of amplification are: 5 minutes at

48 94° for one cycle, 94°/30 seconds, 45°/2 minutes, and 72°/l minute for 30 cycles. A second roimd of PCR amplification, with 1/100 to 1/500 dilution of the first PCR reactions, conditions are: 2 minutes at 94° for one cycle, 94°/30 seconds, 50°/1 minutes, and 72°/1 minute for 30 cycles. Negative control wild type genomic bee

DNA and positive control of plasmid DNA was run concurrently with offspring genomic DNA samples for each PCR reaction. All PCR reactions were made with aerosol resistant pipette tips. The PCR reactions were Southern blotted (Maniatis et al., 1982) and probed with a random primed GFP or pRP466 plasmid probe to verify that these PCR amplification bands indeed represent the expected sequence of plasmid.

Fluorescent Microscopic Imaging Putative transgenic larvae that may contain plasmids expressing the Green

Fluorescent Protein were collected during different stages of development and analyzed for fluorescent expression with a Zeiss microscope at 450 nm wavelengths.

We visualized GFP fluorescent expression in larvae because the background

auto fluorescence of egg yolk obscured our ability to detect GFP fluorescent

expression within honey bee eggs.

Genomic Southern Blot A genomic Southern blot was used to determine if plasmids had integrated into the

genome of the honey bee. Genomic DNA isolates were digested with the restriction

enzyme Xba I overnight in a 37° hot block. The digested genomic DNA samples were

phenol/chloroform extracted, and EtOH precipitated. After the DNA was resuspended

in water, a Southern hybredization was performed according to standard protocols

(Maniatis et al., 1982). The digested genomic DNA was seperated in a 1% agarose gel

(TAB), denatured in 1.0 M NaCl and 0.5 M NaOH, neutralized in 3 M NaCL and 0.5

M Tris-HCl (pH 7.5), and blotted via capillary transfer using 20X SSC onto a

nitrocellulose membrane. This membrane was then baked and prehybredized at 42°.

49 The prehybredization solution was 50% Form amide, 45% 2.5X prehybridization solution, 1% dry milk, 1 ml ssDNA and 5 ml of 20% SDS with the 2.5X prehybridization solution ( 0.25 M PIPES pH 6.8, 2.5 M NaCl, 0.025% Sarkosyl,

0.25% Ficoll, 0.25% PVP-360, and 0.25% BSA).

The random primed plasmid DNA probe was added to the prehybredization solution and the membrane was hybridized overnight at 42°. Washes were IX 2XSSC for 1 minute, followed by 15 minute washes at room tempreture with 2X SC and

0.1% SDS; 0.5X SSC and 0.1% SDS; O.IX SSC and 0.1% SDS; and final wash at 50°

O.IX SSC and 1% SDS. Exposure of the genomic Southern blot was with a

phospho imager.

Ligation-mediated PCR A ligation-mediated PCR (LM-PCR) strategy was used to determine if the founder

queens progeny (FI) had integration of the Tel element within the genome of the

honey bee. Genomic DNA was digested with Hind // (or Hinc II) restriction enzyme.

This restriction enzyme digests internally in the plasmid pRP466. The genomic DNA

isolate was digested, phenol/chloroform extracted, and precipitated. After the DNA

was resuspended in water it was ligated to pre-annealed vector adapters (Eggert et al..

1998). If the plasmid had integrated into the genome, then PCR amplification between

primers specific to vector adapters ligated to genomic DNA and primers specific to

plasmid sequences would amplify the regions of genomic DNA between them.

Sequence for the vector adapters and the vectorette primer was (VA-T) 5’ AAG GAG

AGG ACG CTG TCT GTC GAA GGT AAG GAA CGG ACG AGA GAA GGG

AGA G 3’ and (VA-B) 5’ CTC TCC CTT CTC GAA TCG TAA CCG TTC GTA

CGA GAA TCG CTG TCC TCT CCT T 3’ and (VA-P 224) 5’ CGA ATC GTA

ACC GTT CGT ACG AGA ATC GCT 3’. The Tel Im-pcr primer for the first round

of amplification was 5' CCA AAC AAA TCC AGT GCA ACA AAC AGA GTG CG

3'. Conditions for PCR cycling conditions with Expandfrom Boehringer

Mannheim was: 5 minutes at 94° followed by 10 cycles: 92°/10 seconds, 60°'30

50 seconds, and 68°/3 minutes and then 15 cycles as above except the elongation is 68°/3 minutes + 20 seconds/cycle. The nested PCR reaction primer was 5' GCG TAA CAT

TTC GCT TTA TGC ACA CGG 3'. PCR cycling conditions was 10 cycles: 92°/10 seconds, 65°/30 seconds, and 68°/3 minutes, and then 15 cycles as above except the elongation is 68°/3 minutes + 20 seconds/cycle.

Plasmid Rescue Genomic DNA isolates were digested with the restriction enzyme Xba I overnight

in a 37° hot block. The digested genomic DNA sample was phenol/chlorotbrm

extracted, precipitated, and resuspended in water. Ligation of the genomic DNA

sample was done with T4 DNA ligase overnight at 14°. Chemically competent Xll-

blue bacteria were transformed with the ligation reaction. Transformation of the

chemically competent bacteria with ligated DNA was as follows. Place reaction in

ice-water for 30 minutes, followed by a heat block at 37° for 2 minutes, and back into

ice-water for 5 minutes. Add 500 pi of LB liquid media and put into a heat block at

37° for 45 minutes. Plate out 200 pi of bacteria on LB ampicillin plates. Grow

bacteria overnight at 37° in a bacteria culture chamber. Bacterial colonies were grown

overnight at 37° in LB-ampicillin liquid media. Plasmids were isolated and analyzed

on agarose gel by electrophoresis. These gels were Southern blotted and then probed

with random primed pRP466.

51 RESULTS

Molecular Analysis of Founder (FO) Generation We obtained evidence of heritability using our transformation protocol with 5 different plasmid DNA combinations over two summers (Tables 3.5 and 3.6). During summer 1 a total of 42 virgin wild type PO queens were inseminated with semen mixed with linear DNA plasmids. Four of the 5 plasmid combinations were used in this set of inseminations (Table 3.2). These PO queens were then introduced into small colonies in the indoor flight room, and 24 of those queens produced eggs and/or larvae. The remaining 18 PO queens either were rejected by the colonies during introduction or did not produce FO progeny. We tested samples of FO progeny from

22 of 24 egg laying PO queens via PCR amplification of a plasmid fragment from genomic DNA isolates. We obtained PCR-positive FO samples from 8 PO queens. The remaining 14 PO queens failed to generate PCR-positive FO progeny over several samplings taken from their brood frames. .\11 four of the plasmid combinations used in inseminations of PO queens were represented in this PCR-positive subset of FO progeny. Finally, we were able to rear a total of 18 Founder (FO) queens from progeny of 3 of the 8 PO queens that had produced PCR-positive FO progeny samples.

PCR amplification of a plasmid fragment from genomic DNA isolates of FO generation is displayed in Fig. 3.1. The 6 PO queens were instrumentally inseminated with either phGFP-S65T or phspGFP plasmids (Table 3.5). Single FO larv'a were

tested per PCR reaction from PO queens #4, #6, while groups of 3 to 5 FO larvae were

tested per PCR reaction from PO queens #1, #2, #3, and #5. In this example only 2 PO

queens #1 and #4 had PCR-positive FO samples. FO progeny from PO queen #1 in lane

52 19 tested positive, while another within cohort FO sample in lane 18 tested negative.

Furthermore, FO larvae from PO queen #4 were tested both for GFP expression and in the PCR assay. PCR amplifications of a plasmid fragment from genomic DNA isolates from FO progeny samples from PO queens listed in Tables 3.2 and 3.5 but not

shown in Fig. 3.1 showed similar positive and negative expression patterns.

DNA PO Q ueens PO Q ueens PC R of PCR FO* C onstruct Inseminated Laying FO progeny +/- Q ueens pGFP-S65T 11 6 5 2/3 14/0

phspGFP 20 8 8 2/6 0/0

pRP466/ 8 8 7 3/4 0/3/0 pRP470

p466-hspGFP 3 3 2 1/1 1 /pCRhsp-470

Total 42 25 22** 8/14 18

Table 3.2. Production of Founder queens (FO) after instrumental insemination of

wild type Primary queens (PO) with linear plasmid DNA mixed with

semen during summer 1.

* Number of FO queens successfully reared into adult queens. These were from PO

queens with FO progeny that tested PCR-positive for plasmid. PCR amplification

was with genomic DNA isolates of the FO generation.

**The lineages from these 22 PO queens that had FO progeny PCR tested for

plasmid are shown in detail in Table 3.5.

53 Fig. 3.1. PCR amplification of a plasmid fragment from genomic DNA isolates of the

Founder generation (FO). During summer 1 of a total 22 PO queens that had FO

generation progeny PCR tested for presence of plasmid. 8 PO queens had FO

progeny that tested PCR positive. In this example 2 PO queens were inseminated

with pGFP-S65T plasmid #1 and #3, while another 4 PO queens #4, 45, 46, and 47

were inseminated with phspGFP plasmid. PO queens 43 and 45 had PCR

amplification on genomic samples of 3 to 5 FO larvae per lane. PO queens 41, 44,

46, and 47 had PCR amplification on genomic samples of individual FO larva per

lane. Positive FO samples were identified from PO queens 44 (lane 21) and 41

(lanes 19). Additionally, lanes 21 and 22 correspond to fluorescent GFP positive

(C) and negative (B) larvae seen in Fig. 3.2.

Wild type honey bee larvae (W) in lanes 2, 9, and 10; wild type PO queens 41, 3, 4, 5.

6, and 7 had PCR amplification of a plasmid fragment from FO offspring genomic

DNA isolates. 44 progeny: lanes 1, 3, 4, 7, 12, 21, 22, 23, and 24; 45 progeny; lane

5 and 6; 41 progeny: lane 8, 15, 17, 18, and 19; 43 progeny: lane 11. and 16; 46

progeny: lane 13, and 14; 47 progeny: lane 25-37; M ///m////marker, - wild type

bee genomic DNA, and + plasmid DNA.

54 Queen # 4W445541WVV 34771 3 I 1 I lane # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 M

Queen # 4 4446666666666666 lane# 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 M - +

Figure 3.1 PCR Amplification of GFP from FO Larvae

33 Expression of Green Fluorescent Protein in the FO and F2 generations We viewed using fluorescence microscopy a wild type larva from a queen inseminated with semen only (Fig. 3.2A), FO larvae from a PO queen (Fig. 3.2B and

C), and F2 larvae from a FI queen (Fig. 3.2D and E). These larvae were tested for

fluorescent expression of the marker green fluorescent protein (GFP). Some

autofluorescence was evident in wild type larvae (A), but it was always restricted to

the gut. The fluorescence pattern of FO larvae indicates expression of the GFP

construct in some but not in all tissues. Furthermore, as we would expect from the

PCR data. GFP expression was not evident in all larvae. Fig. 3.2B shows an example

of an FO larva in which there was no detectable GFP expression. The PCR result from

this larva was also negative (Fig. 3.1, lane 22). In contrast. Fig. 3.2C shows an FO

larva from the same cohort with the fluorescent expression pattem evident in the

anterior dorsal region. The PCR result for this larva was positive (Fig. 3.1. lane 21).

We also viewed F2 larvae using fluorescence microscopy (Fig. 3.50 and E). Fig.

3.2D shows an example of an F2 larva in which there was no detectable GFP

expression. The PCR result from this larva was also negative (Fig. 3.5B, lane 4). In

contrast. Fig. 3.2E shows an F2 larva from the same cohort with the fluorescent

expression pattem evident in the anterior region. The PCR result for this larva was

positive (Fig. 3.5B, lane 1).

56 Figure 3.2. Fluorescence imaging of a wild type larva, FO honey bee larvae from a

wild type Primary queen (PO) inseminated with the phspGFP construct, and F2

honey bee larvae from a FI queen that carried the pGFP-S65T construct. Larvae

were collected at day 6 and fluoresced in a Zeiss microscope at 450 nm

wavelength. The wild type larva is from a queen inseminated with semen only in

A. FO larvae collected from the same Primary queen #4 inseminated with phspGFP

are in B and C. The fluorescence of the GFP can be seen in the anterior dorsal

region of one FO larva C, while another within cohort FO larva B fails to exhibit

evidence of GFP expression. F2 larvae collected from the same FI queen #13.1.2

carrying the pGFP-S65T are in D and E. The fluorescence of the GFP is seen in the

extreme anterior region of one F2 larva E, while another within cohort larva fails to

exhibit evidence of GFP expression D.

57 Wild Type Larvae FO Larvae F2 Larvae

Figure 3.2 Fluorescence Imaging of Honey Bee Larvae

5 8 Genomic Southern Blot from the FO generation Genomic Southern blot from genomic DNA isolates of Founder generation honey bee queens (FO). In this example, 4 FO honey bee queens had been inseminated with seman only and set into flightroom colonies. One of these FO queens (#32.1) was from a PO queens inseminated with both p466hspGFP/pRP470, while the remaining 3 FO queens (#27.7, #26.14 and #28.6) were from PO queens inseminated with both pRP466/pRP470. Genomic DNA isolates from these FO queens were restriction digested with X hul. and Southern blotted. .4. sample positive for the plasmid DNA was identified from the genomic isolate of FO queens

#26.14 (lane 4). The sizes of the bands are 9 kb and 6 kb, the size of linearized plasmid DNA. FO queen #32.1 (lane 1) was from a PO queen that had been inseminated with p466hspGFP/pRP470 while FO queens #27.7, #26.14 and #28.6

(lanes 2, 3, and 4) were from PO queens inseminated withpRP466/pRP470.

59 Figure 3.3. Genomic Southern blot from genomic DNA isolates of Founder

generation honey bee queens (FO). In this example, 4 FO honey bee queens had

been inseminated with seman only and set into flightroom colonies. One of these

FO queens (#32.1) was from a PO queens inseminated with both

p466hspGFP/pRP470, while the remaining 3 FO queens (#27.7, #26.14 and #28.6)

were from PO queens inseminated with both pRP466/pRP470. Genomic DNA

isolates from these FO queens were restriction digested with Xba I, and Southern

blotted. A sample positive for the plasmid DNA was identified from the genomic

isolate of FO queens #26.14 (lane 4). The sizes of the bands are 9 kb and 6 kb, the

size of linearized plasmid DNA.

Genomic Southern blot of plasmid DNA from genomic DNA isolates of FO honey bee

queens. FO queen #32.1 (lane 1) was from a PO queen that had been inseminated

with p466hspGFP/pRP470 while FO queens #27.7, #26.14 and #28.6 (lanes 2, 3,

and 4) were from PO queens inseminated with pRP466/pRP470.

60 Queen # 32.1 27.7 26.14 28.6 lane # 1 2 3 4

W-

Figure 3.3 Genomic Southern Blot of FO Queens

6 1 Molecular Analysis of FI Generation During summer I we were able to rear a total of 18 FO putative transgenic queens from PC queens #13, #19, and #22 (Tables 3.3 and 3.5). These 18 FO queens were all inseminated with semen without plasmid DNA, and 14 of them produced eggs and/or larvae in flightroom colonies. We tested samples from 10 FO queens via PCR amplification of genomic DNA isolated from their FI progeny of eggs/larvae (Table

3.3 and 3.5). We obtained PCR-positive results from FI progeny of 3 FO queens.

These PCR-positive samples would be due to maternal inheritance of construct from the FO queen. The remaining 7 FO queens failed to generate FI progeny that were

PCR-positive. In addition, it is perhaps noteworthy that 2 PO queens with FO progeny that tested PCR positive also had FO queen grafts that were completely destroyed by nurse bees during queen rearing.

DNA FO* Queens FO Queens PCR of FI PCR FI Construct Inseminated Laying Progeny +/- Queens pGFP-S65T 14 9 6 2/4 7/0

pRP466 /pRP470 3 3 3 12 0

p466-hspGFP /pCRhsp-470 1 1 1 0/1 -

Total 18 13 10 3/7 7

Table 3.3. Production of FI queens after insemination of Founder (FO) queens with untreated semen during summer 1.

* Numbers in this column represent FO queens shown in the last column in Table 3.2.

62 An example of one of these FO queens with FI progeny that tested PCR-positive is shown in Fig. 3.4A. FO queen #13.1 was reared from PO queen #13, which had been inseminated with phGFP-S65T plasmid. FI progeny, from FO queen #13.1, were tested with PCR amplification of pGFP-S65T from genomic samples containing 3-5 eggs and/or larvae per genomic isolate. PCR amplification revealed that one sample

(Fig. 3.4A, lane 1) tested positive while the other 2 within-cohort FI samples tested negative (Fig. 3.4A, lanes 2 and 3). PCR amplification of pGFP-S65T from FI generation genomic samples containing 1 egg per genomic isolate also tested positive for the presence of plasmid (Fig. 3.4B, lanes 3 and 10) while other within-cohort FI samples tested negative. Indicating that FO queen #13.1 is a transgenic founder honey bee queen.

63 Figure 3.4. PCR amplification of a plasmid fragment from genomic DNA isolates of

the FI generation. During summer 1 of a total 10 FO queens that had FI

generation progeny PCR tested for presence of plasmid, 3 FO queens had FO

progeny that tested PCR positive. In this example Founder queen #13.1 was

inseminated with untreated semen and FI generation eggs and larvae were tested

for inheritance of the pGFP-S65T plasmid via PCR amplification on genomic

DNA samples.

(A). Lanes 1-3 contains mixed stages of FI progeny approximately 5 eggs and/or

larvae. One sample tested positive in lane 1.

(B). Lanes 1-18 each represents a single FI egg collected from FO queen #13.1, two

samples tested positive, lanes 3 and 10. In both figures .VI H indlll marker, - wild

type bee genomic DNA, and + plasmid DNA.

64 3.4 A.

Queen # lane #

3.4 B.

Queen # lan e# 1 2 3 4 5 6 7 8 M 9 10 11 12 13 14 15 16 17 18 - +

Figure 3.4 PCR Amplification of GFP from FI Progeny

6 5 Molecular Analysis o f F2 Generation Seven FI queens were reared from progeny of FO queen #13.1 (Tables 3.4 and 3.5) and were inseminated with untreated semen. We were able to test 6 of 7 FI queens with PCR amplification of a plasmid fragment from genomic DNA samples isolated from their F2 generation progeny. We used F2 generation genomic DNA isolated from approximately 30 eggs per sample for each PCR reaction. Three of the FI queens #13.1.1, #13.1.2, and #13.1.6 had F2 progeny that tested PCR-positive for phGFP-S65T plasmid (Fig. 3.5A). Furthemiore, one of these FI queens #13.1.2 still produced F2 progeny that tested PCR-positive 6 months after the original samples were collected (Fig. 3.SB). Additionally, F2 larvae from FI queen #13.1.2 shown in

Fig. 3.5B lanes 1 and 4, were tested under fluorescence for expression of the GFP marker. The F2 larva from lane 4 in Fig. 3.2D was negative for GFP fluorescence, whereas the F2 larva from lane 1 in Fig. 3.2E was positive for GFP fluorescence. We also viewed a wild type larva from a queen inseminated with untreated semen (Fig.

3.2A).

DNA FI* Queens FI Queens PCR of F2 PCR** Construct Inseminated Laying Progeny______+/- pGFP-S65T ' 7 6 6 ' 4/2'

Total 7 6 6 4/2

Table 3.4. Production of F2 progeny after insemination of FI queens with untreated semen during summer 1.

* Numbers in this column represent FI queens shown in the last column in Table 3.3. ** These larvae were not reared into adult reproductive honey bee queens.

6 6 Figure 3.5. PCR amplification of pGFP-S65T plasmid from F2 genomic DNA

samples from the transgenic line #13. FI queens were inseminated with semen

only and F2 generation eggs were tested for inheritance of the pGFP-S65T

plasmid via PCR amplification of genomic DNA isolated from approximately 30

eggs/reaction.

(A). FI queens are #13.1.1, #13.1.2, and #13.1.6. #13.1.6 progeny lanes 1 and 2;

#13.1.1 progeny in lanes 5 and 6; and #13.1.2 progeny in lanes 7-9. Primary queen

P0#14 inseminated with linear pGFP-S65T construct mixed with semen was

tested with PCR amplification of FO genomic DNA from eggs in lanes 3 and 4.

(B). FI queen #13.1.2 with F2 progeny reared as queen larvae and tested with PCR

amplification of genomic DNA in lanes 1-4 with 1 F2 larva/lane. VI H indlll

marker, -wild type bee genomic DNA. - wild type bee genomic DNA, and -

plasmid DNA.

67 3.5 A. Queen # 13.1.6 14 13.1.1 13.1.2 lane # 123456789 - + M

3.5 B. Queen # 1 3 X 2 . lane # 1 2 3 4 \1

Figure 3.5 PCR Amplification of GFP from F2 Eggs

68 Queen Generations

PO DNA PCR of FO FO PCR of FI F1 PCR of F2 Queen Inseminated Generation Queen Generation Queen Generation + 1- +/- 1 pGFP-S65T - r ' ' ^

2 pGFP-S65T A

3 pGFP-S65T A

4 phspGFP - r

5 phspGFP A

6 phspGFP _ A

7 phspGFP A

8 phspGFP A

9 phspGFP A

10 pRP466/ _ A pRP470 11 pRP466/ A pRP470 12 pRP466/ pRP470

Table 3.5. Details of the honey bee queen lineages produced from 22 PO queens

(continued).

69 Table 3.5 (continued)

13 pGFP-S65T 13.1 + b 13.I.I

13.1.2

13.1.3

13.1.4 _ E

13.1.5

13.1.6 _ E 13.2

13.3 _i)

13.4 . K

13.5 E

13.6 E 14 pGFP-S65T

15 phspGFP

16 phspGFP

Table 3.5 (continued)

70 Table 3.5 (continued)

17 pRP466/ pRP470

18 pRP466/ pRP470

19 pRP466/ 19.1 pRP470

19.2

19.3

20 pRP466/ pRP470

21 p466- hspGFP/ pCRhsp 470 p466- 22.1 hspGFP/ pCRhsp 470

Table 3.5. Details of the honey bee queen lineages produced from 22 PO queens.

' Either environment was not conductive to gratling progeny and rearing honey bee queens, or the PO queens died before we could graft. ® Graft was successful and we reared adult virgin honey bee queens. Grafts from these honey bee queens were removed by worker bees or died during queen development such that virgin honey bee queens could not be reared into adults. ° These honey bee queens were removed from the colonies and sacrificed. ^ These honey bee queens died within the colony before progeny could be grafted and reared into adult honey bee queens. ^ Collected the grafted honey bee queens before emergence for fluorescence testing and PCR analysis

71 Molecular Analysis of FO and FI generations during summer 2 We obtained evidence of heritability using our transformation protocol with 3 different plasmid DNA combinations (Table 3.6). During summer 2 a total of 18 virgin wild type PO queens were inseminated with semen mixed with linearize plasmids and introduced into small colonies in the indoor flightroom (Table 3.6). Of these, 6 PO queens were rejected by the colonics during introduction and killed by the workers. The surviving 12 PO queens produced FO progeny. We attempted to rear 250 of these FO progeny into queens. We inseminated a total of 70 Founder (FO) queens with untreated semen and introduced them into colonies in the flightroom. Of the 70

FO queens inseminated with untreated semen 40 of these queens produced eggs and/or larvae in flightroom colonies. We tested samples of FI eggs via PCR amplification of a plasmid fragment from genomic DNA isolates for the persistence of the Tel element. We obtained PCR-positive results from FI progeny of 8 FO queens. The remaining 32 FO queens failed to generate FI progeny that were PCR-positive. Of the

8 FO queens that produced FI samples that were PCR-positive 6 also had other FI egg isolates that tested PCR-negative for presence of the Tc 1 element.

Examples of FO queens with FI progeny that tested PCR-positive are shown in Fig.

3.6. In Fig. 3.6 there are two plasmid combinations from the subset of 16 FO queens.

These FO queens had genomic DNA from their eggs isolated and PCR amplification for the Tel element done. Plasmids pRP466/pRP470 would be inherited from FO queens #26.1, #28.1, #28.2, #28.3, #28.5, #27.1, #27.4, #26.7, #26.9, #26.10, and

#26.6. Plasmids p466-hspGFP/pCRhsp-470 would be inherited from FO queens

#39.2, #40.3, #40.5, #40.6, and #40.4. FI progeny of FO queens #26.1, #27.1, #39.2,

#26.10 and #40.4 were scored as PCR-positive (Fig. 3.6, lanes 2, 7, 9, 12, and 17).

Of the 70 FO queens that were instrumentally inseminated with untreated semen and

introduced into the indoor flightroom, 53 FO queens that had their genomic DNA

isolated and tested by PCR amplification for the Tc 1 element. The other 17 FO queens

72 died or were removed by worker bees. From the 53 FO queens collected and genomic

DNA isolated, 10 FO queens tested PCR-positive for the Tel element. The other 43

FO queens tested PCR-negative. There were a total of 40 FO queens that produced FI eggs that were collected from the flightroom colonies. Of these 8 FO queens had eggs that tested PCR-positive, the other 32 FO queens produced FI eggs that tested PCR- negative.

From the 10 FO queens that tested PCR-positive, 2 FO queens did not produce any

FI progeny, 2 FO queens produced FI eggs that were PCR-positive, and the other 6

FO queens produced FI eggs that were PCR-negative. There were another 4 FO queens that produced FI progeny collected early that tested PCR-positive (Fig. 3.6) and produced FI samples later that were PCR-negative. These FO queens were isolated and tested PCR-negative for the Tel plasmids.

73 Figure 3.6. PCR Amplification of Tel plasmid from genomic DNA isolated from FI

generation eggs. In this example eggs from 16 FO queens were collected and

genomic DNA isolated. These isolates were analyzed for the presence of the Tel

element by PCR amplification. Each PCR reaction had -50 eggs/reaction. The

PCR reactions were run on an agarose gel and hybridized to radiolabeled pRP466

plasmid. PCR-positive progeny are in lanes 2, 7, 9, 12, and 17.

Lanes 1 and 20 is H indlll marker. Lanes 2, 3, 4, 5. 6, 7. S, 10, 11, 12, and 13 are FI

eggs from FO queens that would have inherited the pRP466/pRP470 plasmid.

Lanes 9, 14, 15, 16, and 17 are FI eggs from FO queens that would have inherited

the p466-hspGFP/pCRhsp-470. Lane 18 is wild type bee genomic DNA. Lane 19 +

is pRP466 plasmid DNA.

74 Queen# M | g g g g g g g | H ^ „ lane# 1 2 3 4 5 6 7 8 9 10 U 12 13 14 15 16 17 18 19 20

Figure 3.6 PCR Amplification of Tel from FI Eggs

7 5 Ligation-mediated PCR and Plasmid Rescue o f FO and FI generations during summer 2 To determine if the transposons were integrated into the genomic DNA, we attempted ligation mediated PCR (LM-PCR) analysis of genomic DNA (Riley et al.,

1990; Roberts et al., 1992). Transposase-mediated transposition would incorporate the Tel element, without plasmid sequences, into the honey bee genome. Ligation mediated PCR allows for the isolation of unknown genomic DNA sequences adjacent to known sequences. Genomic DNA is digested with restriction enzymes and ligated vectorette adapters. We attempted LM-PCR with samples of FI eggs from FO queens

those with the pRP466 Tel element #26.1 (Fig. 3.6, lane 2), #26.10 (Fig. 3.6, lane

12), #27.1 (Fig. 3.6, lane 7), #27.8 (Fig. 3.6, lane 4), #27.9 (Fig. 3.6. lane 5), or the

p466-hspGFP/pCRhsp-470 element and #39.2 (Fig. 3.6, lane 9), and #40.4 (Fig. 3.6,

lane 17). We had LM-PCR bands that we cloned from #26.1. #27.1. and #27.9. The

other samples did not give LM-PCR bands. These clones were Southern blotted and

hybridized to the pRP466 plasmid. LM-PCR clones were sequenced from FO queens

#26.1. and #27.1 and show concatamers of the Tel element (data not shown). This

does not disprove genomic integration, but may indicate the plasmid has been

inherited extrachromosomally.

.An alternative technique to isolate plasmid DNA integrated into genomic DNA is

plasmid rescue. If transposase-mediated excision and integration of the Tel element

occurs then the internalized Tel element and adjacent genomic DNA will not be

isolated. The Tel element itself does not have the ampicillin resistance or the

bacterial origin of replication, these sequences are lost during transposase-mediated

excision and integration. Transopsase-independent elements would be isolated, those

that undergo genomic integration due to cellular repair enzymes, or incomplete

excision and integration of the Tel element with plasmid sequences. Depending on

the size, extrachromosomal arrays might be isolated by plasmid rescue.

Genomic DNA was digested, self-ligated, and transformed into chemically

competent XLl blue bacteria. We attempted plasmid rescue with samples of FI egg

76 isolates from FO queens #26.1, #26.10, #27.1, #27.8, #27.9. #39.2, #40.4 and FO queen #36.1. Plasmids were isolated from FI isolates from FO queens #26.10, #27.1,

#27.9, #40.4 and FO queen #36.1. These plasmids were Southern blotted and hybridized with radiolabeled pRP466 (data not shown). PCR amplification with Tel specific primers was Southern blotted and some of the plasmids produced bands that hybridized to the pRP466 (data not shown).

77 Queen Generations

PO DNA # F 0 FO PCR of PCR of Queen Inseminated Q ueens Q ueen’s FO Q ueens FI eggs G rafted Inseminated +/- +/- 23 pRP466/ X pRP470 24 pRP466/ X pRP470 25 pRP466/ X pRP470 26 pRP466/ 39 26.1 X pRP470 26.2 X X

26.3 X X

26.4 - -

26.5 X X

26.6 --

26.7 - -

26.8 - X

26.9 - -

26.10* -

26.11 - X

26.12 X X

26.13

Table 3.6 Queens Insemination and Generation during gene transfer and analysis with PCR amplification of a plasmid fragment from genomic DNA isolates.

78 Table 3.6 (continued)

26.14

26.15

27 pRP466/ 16 27.1 A . Q pRP470 T 7 1 X X 27.3 - X 27.4 27.5 X X 27.6 - X 27.7 27.8 X 27.9* - "

28 pRP466/ 16 28.1 pRP470 28.2 28.3 28.4 - X 28.5 28.6 28.7 28.8 28.9 X X 28.10 28.11 28.12 - X 29 p466hspGFP/ X pRP470 30 p466hspGFP/ X pRP470 31 p466hspGFP/ X pRP470 32 p466hspGFP/ 13 32.1 pRJP470 32.2 - X Table 3.6 (continued)

79 Table 3.6 (continued)

33 p466hspGFP/ 7 33.1 pRP470 33.2 - X 33.3 33.4 - X 33.5 - X 34 p466hspGFP/ 13 34.1 X X pRP470 34.2 - X 34.3 X 35 p466hspGFP/ 13 0 pRP470 36 p466hspGFP/ 13 36.1* - pRP470 36.2 36.3 X 36.4 - X 37 p466hspGFP/ 13 37.1 X X pRP470 37.2 37.3 X X 37.4 38 p466hspGFP/ 22 38.1 - X pCRhsp470 38.2 38.3 - X 39 p466hspGFP/ 31 39.1 X X pCRhsp470 39.2 - 39.3 39.4 39.5 - X 39.6 - X 39.7 + X Table 3.6 (continued)

80 Table 3.6 (continued)

40 p466hspGFP/ 9 40.1 X X pCRhsp470 40.2 X X 40.3 40.4 A . B 40.5 40.6

* These Founder queens had early collections of FI genomic DNA from eggs test PCR-positive for the Tel element. However later FI egg samples tested PCR- negative. When these queens were then collected and their genomic DNA tested via PCR. 2 of the 3 FO queens tested PCR-negative. There was 7 lines (#26.1, #26.10, #27.1, #27.8, #27.9, #39.2, and #40.4) that were tested with ligation-mediated PCR of genomic DNA from FI generation eggs for the integration of the Tel element. Of these 2 (#26.1, and #27.1) gave PCR amplifications that were cloned and sequenced. The others did not give LM-PCR results. ® These 7 lines (#26.1, #26.10, #27.1, #27.8, #27.9, #39.2, and #40.4) and 1 FQ #36.1 were tested with plasmid rescue for the integration of the Tel element. Of these 4 (#26.10, #27.1, #27.9, and #40.4) and the FQ #36.1 had plasmids that were retrieved from the genomic DNA and hybridized to pRP466.

81 DISCUSSION

Expression and Maintenance of Plasmid DNA in Transgenic Honey Bees This work demonstrates that plasmid DNA can be introduced to the honey bee genome by mixing linear plasmid DNA with semen used to instrumentally inseminate wild type virgin honey bee queens. We were able to detect the plasmid in positive progeny across three generations via two protocols. First, some larvae expressed the expected green fluorescent protein from DNA plasmids containing the GFP marker introduced during insemination of primary queens. GFP expression could eventually serve as a dominant visual marker to nondestructively identify potentially positive larvae to be reared into adult reproductive honey bee queens (Long et al.. 1997).

Second, the strongest argument for successful transformation derives from detection by PCR amplification of the plasmid DNA from animals across several generations.

This work shows that the introduction of plasmid DNA. identification of progeny positive for this plasmid DNA, and maintenance of transgenic honey bees are all possible. These steps are necessary for a long-term goal of manipulating the honeybee genome to investigate areas of honey bee biology, including caste specific developmental regulation, kin recognition, learning and improving agricultural management of this species. But the broader implication is that this technique may also be viable for a number of insect species.

82 Our success with this technique for honey bee transformation may be due to differences in the sperm utilization and fertilization process between mammals and insects. One reason that a sperm-mediated approach may work in insects involves the means through which sperm gain access to the egg. Specifically, honey bee queens mate with up to 20 drones and store viable sperm for several years in her spermatheca.

During fertilization one to several sperm enter the egg via the micropyle (Kerkut and

Gillbert, 1985). Honey bee sperm will bind DNA along the entire length of the sperm, and DNAase treatment removed this DNA (Atkinson et al., 1991). Thus it seems that

insect sperm do not internalize DNA, instead it is likely that sperm carry the DNA

externally as the sperm enter the egg through the micropyle. External foreign DNA

would degrade over time as the sperm is stored in the queen’s spermatheca. Indeed we

observed that the FO progeny of PO queens, which had been inseminated with semen

mixed with plasmid DNA, tested positive for the construct only within 2-3 weeks

after instrumental insemination. We rarely found positive FO progeny after that, and

we now routinely do not maintain PO queens for more than 4 weeks. FO or later

generation queens can be maintained for much longer, over a year or more, and still

produce positive progeny. This is consistent with cellularization of the construct in

those FO or later generation queens. This inheritance through generations is consistent

with the claim of transformation of the honey bee.

Inheritance and Survivorship of Transgenic Progeny We have made several observations from our data that impinge on the long term

success of these techniques in creating and maintaining transgenic honey bees. One

key point is the inheritance from FO queen #13.1 (Fig. 3.4B; Table 3.5) shows the

plasmid can be detected in approximately 1 out of 10 eggs. Therefore, this FO queen

would most likely have been a mosaic founder queen. The plasmid may have been

integrated in only a ft-action of its germ line cells resulting in nonMendelian

inheritance of plasmid in the progeny tested. This nonMendelian inheritance of

83 plasmid from a mosaic founder is not unusual especially if the plasmid is not integrated into the genome of the organism but is instead maintained within cells as large extrachromosomal arrays.

Additional observations indicate the introduction of plasmid DNA can be detrimental to the survivorship of progeny. In several cases we specifically tested eggs because we failed to detect the construct in later stages of worker-developed progeny.

We also observed that queens #13.1 (FO) and #13.1.1 (FI) produced 'spotty’ brood patterns, which is an intermixing of different ages of progeny on a frame of brood.

This pattern contrasts with the normally solid, unbroken brood patterns produced by wild-type queens in our flightroom. This spotty brood pattern would be consistent with an effect of the plasmid on larval survivorship. Under normal conditions queens lay eggs in clusters on a comb so that the brood in any given patch will be of a similar age. This results in an unbroken pattern in the distribution of brood (eggs, larvae and pupae) over the face of a comb. A spotty brood pattern indicates the progeny has either died or has been removed by worker bees. To further analyze this phenomenon and to continue this transgenic line, we reared very young larvae in queen rearing cell builders.

From the transgenic line of FO queen #13.1 we were able to rear adult FI queens.

Under queen rearing conditions, we were able to rear the progeny of FI queen #13.1.1 to late stage F2 larvae that were PCR-positive for the plasmid. Thus queen rearing appears to at least partially overcome lethal effects within this transgenic line.

However, only a fraction of the grafted larvae from this transgenic line were reared

into viable reproductive adult queen honey bees in the queen rearing colonies by worker nurse bees. These same queen rearing colonies simultaneously successfully

reared almost all the control wild type progeny into queens.

84 Extrachromosomal Elements vs. Genomic Integration of Plasmid DNA The long time period between insemination of the PO queens and the ultimate

PCR-positive inheritance of plasmid in the F2 larva after 2 generations indicates that the sperm-mediation technique is successful for transformation. However, it does not mean that the construct is integrated. To determine if the constructs were internalized and maintained as extrachromosomal arrays, or integrated into the honey bee genome via DNA repair mechanisms, we attempted both ligation-mediated PCR analysis of genomic DNA, plasmid rescue (Riley et al., 1990; Roberts et al., 1992), and genomic

Southern blots. Ligation-mediated PCR is a technique that allows for the isolation of unknown genomic DNA sequences adjacent to known plasmid sequences after genomic DNA is restriction digested and ligated to known vectorette adapters.

Plasmid rescue is a technique that allows for the isolation of unknown genomic DNA sequences adjacent to integrated plasmid sequences containing both a bacterial origin of replication and antibiotic resistance sequences. Our ligation mediated-PCR (LM-

PCR) sequence data show concatamers of the inseminated DNA plasmid (data not shown). This is not surprising as linear DNA in pronuclear injections of mice embryos is found as concatamers (Bishop and Smith, 1989). The plasmid rescue (data not shown) and genomic Southern blot (Fig. 3.3) show presence of plasmid DNA. All of these data neither conclusively proves nor disproves genomic integration but, coupled with the nonMendelian inheritance PCR data, it may mean that our plasmid is inherited extrachromosomally. Large extrachromosomal elements have been found in other transformed species (Mello et al., 1991; Nikolaev et al., 1993).

A Future Experimental Approach The most utilized technique with insect genomic transformation occurs via

injection of Drosophila embryos with P-elements (Rubin and Spradling, 1982;

Spradling and Rubin, 1982). This is not routinely possible with non-drosophilid

insects due to inhibition of the P-element from activity in other species and the lack of

85 an alternative universal transposable element (Ashbumer et al., 1998). Yet several transposable elements show promise in species other than Drosophila. These include the Mediterranean fhiitfly Ceratitis capitata transformed with Minos (Loukeris et al.,

1995), mosquitoAedes aegypti with germ-line transformation both with Hermes

(Jasinskiene et al., 1998) and mariner (Coates et al., 1998).

Integration of foreign DNA into the honey bee genome may eventually be enhanced via use of one or more of these transposable elements (Handler et al.. 1998;

Rezsohazy et al., 1997). Having already developed the techniques to introduce plasmid DNA, select positive larvae with green florescent marker, and maintain the developed transgenic lines, we propose utilizing this sperm-mediated plasmid DNA introduction via instrumental insemination approach to screen a variety of transposable elements from other species as a logical next step.

8 6 BIBLOGRAPHY

Abel. T., and Kandel, E. (1998). Positive and negative regulatory mechanisms that mediate long-term memory storage. Brain Res Brain Res Rev 26, 360-78.

Alherini, C. M., Ghirardi, M., Huang, Y. Y.. Nguyen, P. V., and Kandel. E. R. (1995). A molecular switch for the consolidation of long-term memory; c.\MP- inducihle gene expression. Ann N Y Acad Sci 758, 261-86.

Alherini, C. M.. Ghirardi, M., Metz, R., and Kandel, E. R. (1994). C/EBP is an immediate-early gene required for the consolidation of long-term facilitation in Apiysia. Cell 76, 1099-114.

Arezzo, F. (1989). Sea urchin sperm as a vector of foreign genetic information. Cell Biology International Reports 15, 391-404.

Ashhumer, M., Hoy, M. A., and Peloquin, J. J. (1998). Prospects for the genetic transformation of arthropods. Insect Mol Biol 7, 201-13.

Atkinson, P. W., Hines, E. R., Beaton, S., Matthaei, K. I., Reed, K. C., and Bradley, M. P. (1991). Association of exogenous DNA with cattle and insect spermatozoa in vitro. Mol Reprod Dev 29, 1-5.

87 Bailey, C. H., Alberini, C., Ghirardi, M., and Kandel, E. R. (1994). Molecular and structural changes underlying long-term memory storage in Apiysia. Adv Second Messenger Phosphoprotein Res 29, 529-44.

Bailey, C. H., Bartsch, D., and Kandel, E. R. (1996). Toward a molecular definition of long-term memory storage. Proc Natl Acad Sci U S A 93, 13445-52.

Bartsch, D., Casadio, A., Karl, K. A., Serodio, P., and Kandel, E. R. (1998). CREBl encodes a nuclear activator, a repressor, and a cytoplasmic modulator that tbrm a regulatory unit critical for long-term facilitation. Cell 95, 211-23.

Berg, D. E., and Howe, M. M. (1989). Mobile DNA (W ashington, D.C.; .American Society for Microbiology).

Bimstiel, M. L., and Busslinger. M. (1989). Dangerous liaisons; spermatozoa as natural vectors for foreign DNA? Cell 57, 701-2.

Bishop, J. 0., and Smith, P. (1989). Mechanism of chromosomal integration of microinjected DNA. Mol Biol Med 6, 283-98.

Bitterman, M. E., Menzel, R., Fietz, A., and Schafer, S. (1983). Classical conditioning of proboscis extension in honeybees (.Apis mellifera). J Comp Psychol 97, 107-19.

Boynton, S., and Tully, T. (1992). kitheo, a new gene involved in associative learning and memory in Drosophila melanogaster, identified from P element mutagenesis [published erratum appears in Genetics 1992 Sep;132( l):289]. Genetics /i/,655-72.

Brackett, B. G., Killen, D. E., and Peace, M. D. (1971). Cleavage of rabbit ova inseminated in vitro after removal of follicular cells and zonae pellucidae. Fertil Steril 22, 816-28.

88 Brinster, R. L., Sandgren, E. P., Behringer, R. R., and Paimiter, R. D. (1989). No simple solution for making transgenic mice [letter; comment]. Cell 59, 239-41.

Chain, D. G., Casadio, A., Schacher, S., Hegde, A. N., Valbrun, M., Yamamoto, N., Goldberg, A. L., Bartsch, D., Kandel, E. R., and Schwartz, J. H. (1999). Mechanisms for generating the autonomous cAMP-dependent protein kinase required for long­ term facilitation in Apiysia. Neuron 22, 147-56.

Coates, C. J., Jasinskiene, N., Miyashiro, L„ and James, A. .\. (1998). Mariner transposition and transformation of the yellow fever mosquito, .Aedes aegypti. Proc Natl Acad Sci U S A 95, 3748-51.

Cobey, S. (1997). Instrumental insemination and honey bee breeding: short course.

Davis, R. L. (1996). Physiology and biochemistry of Drosophila learning mutants. Physiol Rev 76, 299-317.

De Luca, .4., and Giuditta, A. (1997). Role of a transcription factor (CREB) in memory processes. Riv Biol 90, 371-84.

Dubnau, J., and Tully, T. (1998). Gene discovery in Drosophila: new insights for learning and memory. Annu Rev Neurosci 21, 407-44.

Dudai, Y., Jan, Y. N., Byers, D., Quinn, W. G., and Benzer, S. (1976). dunce, a mutant o îDrosophila deficient in learning. Proc Natl Acad Sci U S A 73, 1684-8.

Dura, J. M., Taillebourg, E., and Preat, T. (1995). The Drosophila learning and memory gene linotte encodes a putative receptor tyrosine kinase homologous to the human RYK gene product. FEBS Lett 370, 250-4.

Eggert, H., Bergemann, K., and Saumweber, H. (1998). Molecular screening for P- element insertions in a large genomic region of Drosophila melanogaster using polymerase chain reaction mediated by the vectorette. Genetics 149, 1427-34.

89 Emptage, N. J., and Carew, T. J. (1993). Long-term synaptic facilitation in the absence of short-term facilitation in Apiysia neurons. Science 262, 253-6.

Fadool, J. M., Hartl, D. L., and Dowling, J. E. (1998). Transposition of the m anner element from Drosophila maiiritiana in zebrafish. Proc Natl Acad Sci U S A 95, 5182-6.

Fahrbach, S. E., and Robinson, G. E. (1995). Behavioral development in the honey bee: Toward the study of learning under natural conditions. Learning and Memory 2, 199-224.

Folkers, E., Drain. P.. and Quinn, W. G. (1993). Radish, a Drosophila mutant deficient in consolidated memory. Proc Natl Acad Sci U S .A. 90, 8123-7.

Francolini, M., Lavitrano, M., Lamia, C. L., French. D.. Frati, L.. Cotelli. F., and Spadafora. C. (1993). Evidence for nuclear internalization of exogenous DNA into mammalian sperm cells. Mol Reprod Dev 34, 133-9.

Gerber, B., Wustenberg, D., Schütz, A., and Menzel, R. (1998). Temporal determinants of olfactory long-term retention in honeybee classical conditioning; nonmonotonous effects of the training trial interval. Neurobiology of Learning and Memory <59, 71-78.

Griffith, L. C., Verselis, L. M., Aitken, K. M., Kyriacou, C. P., Danho, W., and Greenspan, R. J. (1993). Inhibition of calcium/calmodulin-dependent protein kinase in Drosophila disrupts behavioral plasticity. Neuron 10, 501-9.

Griffith, L. C., Wang, J., Zhong, Y., Wu, C. F., and Greenspan, R. J. (1994). Calcium/calmodulin-dependent protein kinase II and potassium channel subunit eag similarly affect plasticity in Drosophila. Proc Natl Acad Sci U S A 91, 10044-8.

9 0 Gueiros-Filho, F. J., and Beverley, S. M. (1997). Trans-kingdom transposition of the Drosophila element m anner within the protozoan Leishmania [see comments] [published erratum appears in Science 1997 .A.ug 8;277(5327);753]. Science 276, 1716-9.

Hammer, M. (1997). The neural basis of associative reward learning in honeybees. Trends Neurosci 20, 245-52.

Hammer, M„ and Menzel, R. (1995). Learning and mcmor\- in the honeybee. J Neurosei 15, 1617-30.

Han, K. A., Millar. N. S., and Davis, R. L. (1998). A novel octopamine receptor with preferential expression in Drosophila mushroom bodies. J Neurosci 18, 3650-8.

Han, K. .A., Millar, N. S., Grotewiel, M. S., and Davis, R. L. (1996). DAMB, a novel dopamine receptor expressed specifically in Drosophila mushroom bodies. Neuron 16, 1127-35.

Handler, A. M., McCombs, S. D., Fraser, M. J., and Saul, S. H. (1998). The lepidopteran transposon vector, piggyBac, mediates germ-line transformation in the Mediterranean fruit fly. Proc Natl Acad Sci U S .\ 95, 7520-5.

Handler, A. M., and O'Brochta, D. A. (1991). Prospects for gene transformation in insects. Annu Rev Entomol 36, 159-83.

Harbo, J. R. (1985). Instrumental insemination of queen bees. American Bee Journal 127, 282-287.

Hitler, R., Heisenberg, M., and Preat, T. (1998). Abnormal mushroom body plasticity in the Drosophila memory mutant amnesiac. Neuroreport 9, 2717-9.

91 Jasinskiene, N., Coates, C. J., Benedict, M. Q., Gomel, A. J., Rafferty, C. S., James, A. A-, and Collins, F. H. (1998). Stable transformation of the yellow fever mosquito, Aedes aegypti, with the Hermes element from the housefly. Proc Natl Acad Sci U S A 95, 3743-7.

Kerkut, G. A., and Gillbert, L. 1. (1985). Comprehensive insect physiology, biochemistry, and pharmacology (Oxford, New York: Fergamon Press).

Laidlaw, H. H., and Page, R. E. (1997). Queen rearing and bee breeding, 1st Edition (Cheshire, Conn.: Wicwas Press).

Lauria, A., and Gandolfi, F. (1993). Recent advances in sperm cell mediated gene transfer. Mol Reprod Dev 36, 255-7.

Lavitrano, M., Camaioni, A., Fazio, V. M., Dolci, S., Farace, M. G., and Spadafora, C. (1989). Sperm cells as vectors for introducing foreign DNA into eggs: genetic transformation of mice [see comments]. Cell 57, 717-23.

Lavitrano, M., Fomi, VL, Varzi. V., Pucci, L., Bacci, M. L., Di Stefano, C., Fioretti, D., Zoraqi, G., Moioli, B.. Rossi, M., Lazzereschi, D., Stoppacciaro, .\., Seren, E., Alfani, D., Cortesini, R., and Frati, L. (1997). Sperm-mediated gene transfer: production of pigs transgenic for a human regulator of complement activation. Transplant Proc 29, 3508-9.

Lavitrano, M., French, D., Zani, M., Frati, L., and Spadafora, C. (1992). The interaction between exogenous DNA and sperm cells. Mol Reprod Dev 31, 161-9.

Lavitrano, M., Maione, B., Forte, E., Francolini, M., Sperandio, S., Testi, R., and Spadafora, C. (1997). The interaction of sperm cells with exogenous DNA: a role of CD4 and major histocompatibility complex class II molecules. Exp Cell Res 233, 56- 62.

92 Levin, L. R., Han, P. L., Hwang, P. M., Feinstein, P. G., Davis, R. L., and Reed, R. R. (1992). The Drosophila learning and memory gene rutabaga encodes a Ca2+/CalmoduIin-responsive adenylyl cyclase. Cell 6 8 ,479-89.

Liao, L. W., Rosenzweig, B., and Hirsh, D. (1983). Analysis of a transposable element in Caenorhabditis elegans. Proc Natl Acad Sci U S A 50, 3585-9.

Long, Q., Meng. A., Wang, H., lessen, J. R., Farrell, M. J., and Lin, S. (1997). GATA-1 expression pattern can be recapitulated in living transgenic zebrafish using GFP reporter gene. Development 124, 4105-4111.

Loukeris, T. G., Livadaras, I., Area, S., Zabalou, S., and Savakis, C. (1995). Gene transfer into the medfly, Ceratitis capitata, with a Drosophila hydei transposable element [see comments]. Science 270, 2002-5.

Luo. G., Ivies, Z., Izsvak, Z., and Bradley, A. (1998). Chromosomal transposition of a T el/mariner-like element in mouse embryonic stem cells. Proc Natl .Acad Sci U S A 95, 10769-73.

Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982). Molecular Cloning: a laboratory manual (Cold Spring Harbor, N. Y.: Cold Spring Harbor Laboratory).

Mello, C. C., Kramer, J. M., Stinchcomb, D.. and Ambros, V. (1991). Efficient gene transfer in C .elegans: extrachromosomal maintenance and integration of transforming sequences. Embo i 10, 3959-70.

Menzel, R. (1983). Neurobiology of learning and memory; the honeybee as a model system. Naturwissenschaften 70, 504-11.

Menzel, R., Hammer, M., Muller, U., and Rosenboom, H. (1996). Behavioral, neural and cellular components underlying olfactory learning in the honeybee. J Physiol Paris 90, 395-8.

93 Menzel, R., and Muller, U. (1996). Learning and memory in honeybees: from behavior to neural substrates. Annu Rev Neurosci 79, 379-404.

Miklos, G. L. (1993). Molecules and cognition: the latterday lessons of levels, language, and lac. Evolutionary overview of brain structure and function in some vertebrates and invertebrates. J Neurobiol 24, 842-90.

Moreau-Fauvarque, C., Taillebourg, E., Boissoneau, E.. Mesnard. J., and Dura, J. M. (1998). The receptor tyrosine kinase gene linoiie is required for neuronal pathway selection in the Drosophila mushroom bodies. Mech Dev 78, 47-61.

Muller, U. (1994). Ca2+/calmodulin-dependent nitric oxide synthase in .Apis mellifera and Drosophila melanogaster. Eur J Neurosci6, 1362-70.

Nikolaev, A. I., Tchkonia, T. T., Kafiani-Eristavi, C. A., and Tarantul. V. Z. (1993). Preferential extrachromosomal localization of exogenous DNA in transgenic silkworm Bombv.x mori L. Mol Gen Genet 236, 326-30.

O’Brochta, D. .K., and Atkinson, P. W, (1996). Transposable elements and gene transformation in non-drosophilid insects. Insect Biochem Mol Biol 26, 739-53.

Page, R. E., Jr., Erber, J., and Fondrk, M. K. (1998). The effect of genotype on response thresholds to sucrose and foraging behavior of honey bees (.Apis mellifera L.). J Comp Physiol [A] 182,489-500.

Page, R. E., and Robinson, G. E. (1991). The genetics of division of labour in honey bee colonies. Advances Insect Physiology 23, 117-171.

Paimiter, R. D., and Brinster, R. L. (1986). Germ-line transformation of mice. Annu Rev Genet 2 0 , 465-99.

Quinn, W. G., Harris, W. A., and Benzer, S. (1974). Conditioned behavior in Drosophila melanogaster. Proc Natl Acad Sci U S A 77, 708-12.

94 Quinn, W. G., Sziber, P. P., and Booker, R. (1979). The Drosophila memory mutant amnesiac. Nature 277, 212-4.

Rembold, H., and Lackner, B. (1981). Rearing of honeybee larvae in vitro: effect of yeast extract on queen differentiation. Journal of Apicultural Research 20, 165-171.

Rescorla, R. A. (1988). Behavioral studies of Pavlovian conditioning. .Annu Rev Neurosci //, 329-52.

Rezsohazy, R., van Luenen, H. G., Durbin, R. M., and Plasterk, R. H. (1997). Tc7, a T el-hitch hiking transposon in Caenorhabditis elegans. Nucleic Acids Res 25, 4048- 54.

Riley, J., Butler, R., Ogilvie, D., Finniear, R., Jenner, D., Powell. S., .Anand, R.. Smith, J. C.. and Markham, A. F. (1990). A novel, rapid method for the isolation of terminal sequences from yeast artificial chromosome (YAC) clones. Nucleic Acids Res 2887-90.

Roberts, R. G., Bobrow, M., and Bentley, D. R. (1992). Point mutations in the dystrophin gene. Proc Natl Acad Sci U S A 89. 2331-5.

Robinson, G. E., Fahrbach, S. E., and Winston, M. L. (1997). Insect societies and the molecular biology of social behavior. Bioessays /P. 1099-108.

Robinson, G. E., and Page, R. E. (1988). Genetic determination of guarding and undertaking in honey bee colonies. Nature 333, 356-358.

Robinson, G. E., and Page, R. E. (1989). Genetic determination of nectar foraging, pollen foraging, and nest-site scouting in honey bee colonies. Behavioral Ecology Sociobiology24, 317-323.

95 Rothenbuhler, W. C. (1964). Behavioral genetics of nest cleaning in honey bees, IV. Responses of F2 and backcross generations to disease-killed brood. American Zoology4, 111-123.

Rothenbuhler, W. C. (1979). Semidomesticated insects: Honeybee breeding. In Genetics in Relation to Insect management, M. A. Hoy and J. J. McKelvey, eds. (New York: Rockefeller Found), pp. 84-92.

Rubin, G. M., and Spradling, A. C. (1982). Genetic transformation of Drosophila with transposable element vectors. Science 218, 348-53.

Sarkar, A., Coates, C. J., Whyard, S., Willhoeft, U., Atkinson, P. W., and O'Brochta, D. A. (1997). The Hermes element from Miisca domestica can transpose in four families of cyclorrhaphan flies. Genetica 99, 15-29.

Shamila, Y., and Mathavan, S. (1998). Sperm-mediated gene transfer in the silkworm Bombyx mori. Arch Insect Biochem Physiol 37, 168-77.

Shull, R. W., and Dixon, S. E. (1986). An artificial diet for laboratory rearing of honeybees. Journal of Apicultural Research 25, 35-43.

Smith, B. H., Abramson, C. 1., and Tobin, T. R. (1991). Conditional withholding of proboscis extension in honeybees (Apis mellifera) during discriminative punishment. J Comp Psychol 105, 345-56.

Spradling, A. C., and Rubin, G. M. (1982). Transposition of cloned P elements into Drosophila germ line chromosomes. Science 218, 341-7.

Tully, T., Bolwig, G., Christensen, J., Connolly, J., DeZazzo, J., Dubnau, J., Jones. C., Pinto, S., Regulski, M., Svedberg, F., and Velinzon, K. (1996). Genetic dissection of memory in Drosophila. PhysiolJ Paris 90, 383.

9 6 Tully, T., Cambiazo, V., and tCruse, L. (1994). Memory through metamorphosis in normal and mutant Drosophila. J Neurosci 14, 68-74.

Tully, T., Preat, T., Boynton, S. C., and Del Vecchio, M. (1994). Genetic dissection of consolidated memory in Drosophila. Cell 79, 35-47.

von Frisch, K. (1967). The orientation and dance language of the honey bee colonies (Cambridge, MA: Harvard University Press).

Vos, J. C., De Baere, I., and Plasterk, R. H. (1996). Transposase is the only nematode protein required for in vitro transposition ofTcl. Genes Dev 10, 755-61.

Vos, J. C., and Plasterk, R. H. (1994). Tel transposase of Caenorhabditis elegans is an endonuclease with a bipartite DNA binding domain. Embo 113, 6125-32.

Vos, J. C., van Luenen, H. G., and Plasterk, R. H. (1993). Characterization of the Caenorhabditis elegans Tel transposase in vivo and in vitro. Genes Dev 7, 1244-53.

Winston, M. L. (1987). The biology of the honey bee (Cambridge, MA: Harvard University Press).

Wittstock, S., Kaatz, H. H., and Menzel, R. (1993). Inhibition of brain protein synthesis by cycloheximide does not affect formation of long-term memory in honeybees after olfactory conditioning. J Neurosci 13, 1379-86.

Wittstock, S., and Menzel, R. (1994). Color learning and memory in honey bees are not affected by protein synthesis inhibition. Behav Neural Biol 62, 224-9.

Xia, Z. G., Refsdal, C. D., Merchant, K. M., Dorsa, D. M., and Storm, D. R. (1991). Distribution of mRNA for the calmodulin-sensitive adenylate cyclase in rat brain: expression in areas associated with learning and memory. Neuron 6, 431-43.

97 Yin, J. C., Del Vecchio, M., Zhou, H., and Tully, T. (1995). CREB as a memory modulator; induced expression of a dCREB2 activator iso form enhances long-term memory in Drosophila. Cell 81, 107-15.

Yin, J. C., Wallach, J. S., Del Vecchio, M., Wilder, E. L., Zhou, H.. Quinn, W. G., and Tully, T. (1994). Induction of a dominant negative CREB transgene specifically blocks long-term memory in Drosophila. Cell 79, 49-58.

Zani, M., Lavitrano, M„ French, D., Lulli, V., Maione. B., Sperandio, S., and Spadafora, C. (1995). The mechanism of binding of exogenous DNA to sperm cells: factors controlling the DNA uptake. Exp Cell Res 2 17. 57-64.

Zhong, Y.. and Wu, C. P. (1991). Altered synaptic plasticity in Drosophila memory mutants with a defective cyclic AMP cascade. Science 251. 198-201.

98