CHARACTERIZATION of STOMATIN SUPPRESSORS Ssu-1 and Ssu-2

Home , Alcohol sulfotransferase, Flurothyl, Malignant hyperthermia, Stomatin

CHARACTERIZATION OF STOMATIN SUPPRESSORS ssu-1 AND ssu-2

BRYAN THOMAS CARROLL

Submitted in partial fulfillment of the requirements

For the degree of Doctor of Philosophy

Dissertation Adviser: Dr. Philip Morgan

Department of Genetics

CASE WESTERN RESERVE UNIVERSITY

August, 2005

CASE WESTERN RESERVE UNIVERSITY

SCHOOL OF GRADUATE STUDIES

We hereby approve the dissertation of

______Bryan Thomas Carroll ______

candidate for the Ph.D. degree *.

(signed)_____Ron Conlon______(chair of the committee) ______Philip Morgan______

______Helen Salz______

______Robert Peterson______

______George Dubyak______

______

(date) ______August 2005______

*We also certify that written approval has been obtained for any proprietary material contained therein.

To Gina and my family,

Thank you.

iii Table of Contents

LIST OF FIGURES ...... 3 GLOSSARY OF C. ELEGANS GENE NAMES USED IN THIS PAPER ...... 6 NOMENCLATURE IN C. ELEGANS: ...... 6 AN ALPHABETICAL LISTING OF THE GENES USED IN THIS PAPER: ...... 6 ABSTRACT...... 9 I INTRODUCTION ...... 11 A) A BRIEF HISTORY OF THE STUDY OF VOLATILE ANESTHETICS ...... 11 1) The significance of volatile anesthetic research...... 11 2) The Meyer-Overton rule and the debate over the functional target of volatile anesthetics...... 11 B) AN INTRODUCTION TO THE GENETIC STUDY OF VOLATILE ANESTHETICS ...... 13 1) The model system Caenorhabditis elegans...... 13 C) A REVIEW OF THE GENETIC STUDIES OF VOLATILE ANESTHETICS...... 15 1) in vitro targets of volatile anesthetic...... 15 2) Genetic studies of sensitivity to volatile anesthetics in yeast...... 17 4) Genetic studies of sensitivity to volatile anesthetics in C. elegans as measured by other endpoints21 5) Genetic studies of sensitivity to volatile anesthetics in Drosophila...... 24 6) Genetic studies of sensitivity to volatile anesthetics in mice...... 26 7) A possibly conserved target associated with malignant hyperthermia ...... 27 D) MODIFICATION OF VOLATILE ANESTHETIC SENSITIVITY BY STOMATIN-LIKE PROTEINS ...... 28 1) The significance of stomatin-like proteins to the study of volatile anesthetics ...... 28 2) A review of stomatin-like proteins ...... 30 II GENETIC SUPPRESSORS OF THE STOMATIN UNC-1 ...... 38 A) THE RATIONALE FOR USING GENETIC SUPPRESSORS TO STUDY UNC-1 ...... 38 B) A DETAILED PHENOTYPIC DESCRIPTION OF THE GENETIC SUPPRESSOR SSU-1...... 41 C) A DETAILED PHENOTYPIC DESCRIPTION OF THE GENETIC SUPPRESSOR SSU-2...... 43 III SULFONATION OF A SIGNAL FROM SENSORY NEURONS MODULATES STOMATIN- DEPENDENT COORDINATION IN C. ELEGANS...... 46 ABSTRACT ...... 47 INTRODUCTION ...... 48 RESULTS ...... 49 Isolation of the genetic suppressor ssu-1...... 49 Identification of ssu-1 ...... 50 The similarity between SSU-1 and cytosolic sulfotransferases...... 52 SSU-1 in vitro sulfotransferase activity...... 54 SSU-1 expression in the ASJ amphid neurons...... 57 The shared dauer phenotype of ssu-1 and the ASJ amphid neurons ...... 60 Genetic characterization of the SSU-1 substrate...... 63 CONCLUSIONS ...... 64 MATERIALS AND METHODS ...... 70 ACKNOWLEDGEMENTS ...... 73 SUPPLEMENTAL DATA...... 74 IV SSU-2: LIMITED EXPRESSION OF SUPPRESSOR CONFINES SITE OF STOMATIN FUNCTION ...... 75 ABSTRACT ...... 75 INTRODUCTION ...... 76 RESULTS ...... 76

1 Isolation of the genetic suppressor ssu-2...... 76 Identification of ssu-2 ...... 78 The similarity between SSU-2 and fls485-like proteins ...... 83 The full-length ssu-2 gene might include the coding region of the neighboring gene Y67A10A.7. ..84 ssu-2 promoter GFP expression in the DVA and RMD neurons...... 86 ssu-2 expression of UNC-1::GFP fusion protein...... 91 CONCLUSIONS ...... 93 FUTURE DIRECTIONS ...... 97 MATERIALS AND METHODS ...... 99 ACKNOWLEDGEMENTS ...... 102 V SUMMARY...... 103 A) HIGHLIGHTS OF SECTIONS I-IV...... 103 B) MODEL WITH THE COMBINED FINDINGS ...... 106 1) Given the expression patterns of ssu-1 and ssu-2, the likelihood of autonomous vs. nonautonomous function ...... 106 2) Given the signaling function of SSU-1, a model of SSU-2 function that would place SSU-2 as a receiver of the SSU-1 signal...... 110 C) RELATING THE FINDINGS OF SSU-1 AND SSU-2 TO THE UNDERSTANDING OF THE FUNCTIONS OF STOMATINS AND VOLATILE ANESTHETICS IN OTHER SPECIES...... 112 D) FUTURE DIRECTIONS...... 113 1) Exploring the roles of the ASJ, RMD, and DVA neurons in affecting stomatin function and modulating volatile anesthetic sensitivity ...... 113 2) Mapping of the sites of action for genes involved in modulating volatile anesthetic sensitivity .116 3) Further biochemical characterization of SSU-1...... 117 4) Supporting the model of ssu-1 signaling through G-proteins ...... 118 5) A saturated screen for suppressors of unc-1 using RNAi...... 118 APPENDIX I: PROTEIN SEQUENCE ALIGNMENTS ...... 120 BIBLIOGRAPHY ...... 124

2 List of Figures

Figure 1. Genetic determinants modify volatile anesthetic sensitivity in C. elegans...... 19 Figure 2. Immobility is a relevant endpoint to study volatile anesthetic sensitivity in C. elegans ...... 23 Figure 3. The three stomatin-like proteins associated with neuronal phenotypes in C. elegans ...... 30 Figure 4. Neuronal stomatin-like proteins affect cellular distribution of other neuronal stomatin-like proteins...... 36 Figure 5. The twitch assay measurement of the movement phenotypes of the genetic suppressors ssu-1 and ssu-2 ...... 40 Figure 6. ssu-1 modifies volatile anesthetic sensitivity...... 42 Figure 7. ssu-2 modifies volatile anesthetic sensitivity...... 44 Figure 8. ssu-1(fc73) has a mutation in the predicted cytosolic sulfotransferase gene Y113G7A.11...... 51 Figure 9. SSU-1 is similar to the cytosolic alcohol sulfotransferase enzyme family ...... 53 Figure 10. Bacterially expressed isoforms of SSU-1 demonstrate sulfotransferase activity in vitro...... 55 Figure 11. The bacterially expressed full length isoform of SSU-1 demonstrates in vitro sulfotransferase activity against sterols...... 56 Figure 12. SSU-1 is expressed in only two cells, the ASJ amphid neurons ...... 58 Figure 13. Fusion with UNC-1 changes GFP distribution...... 60 Figure 14. The physical and functional connections of ASJ amphid neurons with other neurons...... 61 Figure 15. Model: The SSU-1 sulfonation of a signaling molecule in the ASJ neurons remotely affects the site of UNC-1 deficiency...... 66 Figure 16. Steroid hormone regulation of the UNC-8 homolog ENaCs in humans...... 68 Figure 17. Model: The SSU-1 sulfonated signaling molecule regulates the transcription of activators of the UNC-8 sodium channels...... 69 Figure 18. ssu-2 modifies volatile anesthetic sensitivity ...... 77 Figure 19. ssu-2 was mapped onto chromosome IV...... 79

3 Figure 20. RT-PCR amplification of multiple ssu-2(fc71) spliced forms ...... 80 Figure 21. Rescuing ssu-2 constructs with and without the neighboring predicted gene Y67A10A.7...... 82 Figure 22. SSU-2 is a member of the fls485 protein family...... 84 Figure 23. Anti-SSU-2 antibody recognizes a 90kd band in fc71 total worm protein but not N2 protein...... 85 Figure 24. ssu-2 promoter::GFP expression in seven neurons ...... 87 Figure 25. The ssu-2 promoter GFP expression fills the full length of the DVA neuron. 88 Figure 26. The multiple neuronal connections of RMD neurons with the motoneurons of the head...... 89 Figure 27. The RMD neurons coordinate head movements in response to stimuli...... 90 Figure 28. The primary circuit of locomotion...... 90 Figure 29. The DVA makes many synaptic or gap junction connections with the primary circuit of locomotion...... 91 Figure 30. Expression of UNC-1::GFP in the RMD and DVA neurons fails to rescue the Unc-1 phenotype...... 92 Figure 31. Model: ssu-2 suppresses unc-1 locomotion defect through the DVA neuron to directly affect the coordination of the primary circuit of locomotion...... 96 Figure 32. Model: Model: UNC-1 and its suppressors SSU-1 and SSU-2 function autonomously...... 107 Figure 33. Model: UNC-24 also functions autonomously from SSU-1 and SSU-2...... 109 Figure 34. Model: A HSP40 predicted function of SSU-2 gives SSU-2 a role in modifying SSU-1 signaling...... 111

4 Acknowledgements

This work was supported in part by the Department of Anesthesiology of University

Hospitals of Cleveland and by the Department of Genetics of Case Western Reserve

University. Bryan Carroll was supported in part by NIH T32 GM07250 and the Case

Medical Scientist Training Program. Additional funding came from National Institutes of

Health (NIH) Grant GM-45402

I would like to thank the C. elegans community for creating and maintaining a great

collection of resources. I would like to thank Zeynep Altun for assistance with cell

identification, and Sara Olson for assistance with being a sulfotransferase-studying

graduate student.

I would like to thank George Dubyak and his laboratory for providing materials, technical

assistance, and emotional support for the study of the ssu-1 sulfotransferase.

I would like to thank the S+M lab past and present.

I appreciate everything that Philip Morgan and Margaret Sedensky have done for me far

more than they know. I am grateful for their mentorship during the last five years.

I am grateful to Kenneth Thimann, Barbara Weber, and Fergus Couch for giving me the

foundation of my molecular biology training.

5 Glossary of C. elegans gene names used in this paper

Nomenclature in C. elegans:

In C. elegans literature, gene names are generally recognized as three letter names

followed by a dash and a number, all italicized. Different mutations can occur in the same gene and these are given allele names. Allele names are one or two letters followed

by a number with no dash. Genes can be written with a parenthetical note to express loss

of function (lf), semidominant (sd), or null (null). The protein made from a gene is

capitalized and not italicized. The phenotype associated with a mutation in a gene is

referred to by capitalizing the first letter of the gene name with no italics. The wild type

strain of worm used in these experiments is Bristol, abbreviated N2.

An alphabetical listing of the genes used in this paper: cat-1/cat-2/cat-4 – Genes involved in catecholamine biosynthesis and processing. daf-5/daf-9/daf-12/daf-16/daf-22 – Genes associated with modifying the formation of

dauer larvae.

dpy-4/dpy-20 – Marker genes that increase the thickness of the worm. Used in the

genetic mapping of ssu-2. e580 – An allele of unc-1. e114 – An allele of unc-1. eDf28 – An allele of unc-24. fc71 – An allele of ssu-2. fc73 – An allele of ssu-1. fc83 – A mutation that increases the formation of dauer larvae when paired with unc-24.

6 gas-1 – A component of mitochondrial complex I that modifies volatile anesthetic

sensitivity as measured by immobility.

hs-1 – An allele of unc-1.

mec-2 – A neuronal stomatin-like protein that modifies mechanosensation.

mec-4/mec-10 – Two amiloride-sensitive DEG/ENaC cation-selective channel subunits

that modify mechanosensation.

moc-1(ok366) – The source of fc83. npc-1/npc-2 – Two genes that code for large transmembrane glycoproteins that are predicted to function in intracellular cholesterol and glycolipid trafficking. n494 – An allele of unc-1. ozDf1/ozDf2 – Two large genomic deletions of chromosome V. Both deletions remove

ssu-1.

rol-6/rol-9 – Two genes that when mutated cause a cuticle defect that results in a

spiraling pattern of locomotion. Both have been used in the genetic mapping and

transgenic characterization of ssu-1, ssu-2, and unc-1. ssu-1 – A cytosolic sulfotransferase that suppresses unc-1 and unc-24.

ssu-2 – A novel gene that suppresses unc-1 and unc-24.

tm1117 – An allele of ssu-1.

tm1633 – An allele of ssu-2.

tra-3 – A gene used as a marker in the genetic mapping of ssu-2.

unc-1 – A neuronal stomatin-like protein that modifies volatile anesthetic sensitivity as

measured by immobility.

7 unc-7 – A gap junction innexin that modifies volatile anesthetic sensitivity as measured by immobility. unc-8 – An amiloride-sensitive DEG/ENaC cation-selective channel subunit that modifies volatile anesthetic sensitivity as measured by immobility. unc-9 – A gap junction innexin that modifies volatile anesthetic sensitivity as measured by immobility. unc-24 – A neuronal stomatin-like protein that modifies volatile anesthetic sensitivity as measured by immobility. unc-51 – An uncoordinated mutant used as a marker for the genetic mapping of ssu-1. unc-79 – A novel gene that modifies volatile anesthetic sensitivity as measured by immobility. unc-80 – An uncloned gene that modifies volatile anesthetic sensitivity as measured by immobility.

Abstract

Our laboratory is primarily interested in the mechanism of action of volatile anesthetics. We have chosen to use the model system Caenorhabditis elegans to study the genetic determinants of volatile anesthetic sensitivity. Previous work has shown that the function of stomatin-like proteins is one determinant controlling volatile anesthetic response. We isolated two genetic suppressors of the volatile anesthetic-sensitive phenotype resulting from the deficiency of neuronal stomatin-like protein. This work describes the identification and characterization of these two genetic suppressors, ssu-1 and ssu-2. ssu-1 is the only representation in the C. elegans genome of a type of enzyme involved in sending signals between cells through the circulation. While there are several types of sulfotransferase in C. elegans, ssu-1 is the only cytosolic alcohol sulfotransferase. The cytosolic alcohol sulfotransferase affects the action of signaling molecules by attaching a sulfonate to an alcohol group. Our characterization of this sulfotransferase and its relationship to stomatin-like proteins suggests a common role to the conserved structure of stomatin-like proteins. The second genetic suppressor, ssu-2, is a novel gene that is expressed in seven neurons. The location and neuronal connections of these neurons suggest that they are directly involved in coordinating locomotion, a second phenotype that is altered by stomatin deficiency. Since ssu-2 affects both locomotion and sensitivity to volatile anesthetics, this expression pattern suggests that the same neurons are involved in coordinating locomotion and modifying sensitivity to volatile anesthetics. The functional characterization of ssu-1 and the

9 localization of the expression of ssu-2 offer complementary perspectives into the function of neuronal stomatin-like proteins.

10 I Introduction

A) A brief history of the study of volatile anesthetics

1) The significance of volatile anesthetic research.

On Oct. 16, 1846, William T.G. Morton, a Boston dentist, demonstrated the use of ether during surgery. Volatile anesthetics liberated patients from the trauma of experiencing their own surgery. The effects of volatile anesthetics are reversible, with no long term effects on neurological function. The efficacy of volatile anesthetics combined with their speed of induction and recovery transformed the practice of surgery. As beneficial as volatile anesthetics are, however, they are not perfect. There is a high risk of mortality when working with these drugs with their LC50 (the concentration lethal to

50% of the population) so close to their EC50 (the concentration effective for 50% of the

population). In addition, potentially fatal malignant hyperthermia occurs in 1 in 15,000

anesthetic administrations in children and 1 in 50,000 to 100,000 in adults. Half the

patients who develop the syndrome have had previous anesthesia without recognized malignant hyperthermia (Nelson and Flewellen 1983). A better understanding of volatile

anesthetic function will reveal both the mechanism of a great medical mystery and

advance the progress of anesthetic safety.

2) The Meyer-Overton rule and the debate over the functional target of volatile

anesthetics

General anesthesia can be induced by a wide variety of structurally dissimilar molecules. The target shared by volatile anesthetics has been debated for over one hundred years. There is even disagreement over what class of molecules, proteins or

11 lipids, is targeted by volatile anesthetics. The first piece of data supports lipids as the

target of volatile anesthetics. It is based on an association that has been observed in all

species tested that volatile anesthetics follow the relationship defined by the Meyer-

Overton rule (Meyer 1899; Overton 1901; Tanifuji, Eger et al. 1977). The Meyer-

Overton rule describes the correlation between potency and lipid solubility of a volatile

anesthetic. The Meyer-Overton rule predicts that the greater the lipid solubility of an anesthetic, the greater the potency of the drug. The Meyer-Overton rule reveals two important features of volatile anesthetics. First, the Meyer-Overton rule’s applicability to all species suggest that there is a conserved target to volatile anesthetic that has been conserved throughout evolution. This conservation of effect is part of our justification for using a genetic model to study the mechanism of volatile anesthetics.

Secondly, the Meyer-Overton rule suggests that the mechanism of volatile anesthetics involves a lipid-dependent action.

In line with focus on lipids by the Meyer-Overton rule, some have postulated that volatile anesthetics confer their effects by interacting directly with lipids to modify the chemical environment of the biological membrane. Ueda et al claimed that the

“primary effect of anesthetics on membranes is to weaken the lipid-water interaction forces” (Ueda, Hirakawa et al. 1986). The disruption of the lipid-water interaction forces is thought to then interfere non-specifically with the normal functions of membrane machinery.

There are opposing views in the field that hold proteins as the direct targets of volatile anesthetics. Franks and Lieb proposed that volatile anesthetics can mimic ligands and interfere with protein signaling (Franks and Lieb 1981; Franks and Lieb

12 1984; Franks and Lieb 1994). The stereoisomers of volatile anesthetics have been compared to test whether the site of such actions is purely lipid in nature or may contain a protein target. In both C. elegans and mice, stereoisomers of volatile anesthetics demonstrated differing potencies (Harris, Moody et al. 1992; Sedensky, Cascorbi et al.

1994; Morgan, Usiak et al. 1996). The different potencies of stereoisomers would support that the site targeted by volatile anesthetic is able to differentiate steric structures.

Proteins, more so than lipids, would possess such a recognition site.

159 years after the first use of volatile anesthetics, there is evidence supporting both lipids and proteins as the relevant targets of volatile anesthetics. To study the mechanism of action of volatile anesthetics, we are employing a genetic model. Variations in genes will be able to modify the targets of volatile anesthetics whether they are protein or lipid.

By identifying and characterizing genes that alter sensitivity, the target of volatile anesthetics may be revealed.

B) An introduction to the genetic study of volatile anesthetics

1) The model system Caenorhabditis elegans

We have chosen to use the model system Caenorhabditis elegans to study the genetic determinants of anesthetic sensitivity. Of the established metazoan genetic model systems, C. elegans has the simplest nervous system. 302 of the 959 somatic cells of a hermaphrodite C. elegans are neurons (White, Southgate et al. 1986). With support cells, the nervous system of the hermaphrodite includes 37% of the somatic cells. Electron microscopic dissection was used to produce a map of all of the connections of these neurons (White, Southgate et al. 1986). 5000 chemical synapses, 2000 neuromuscular

13 junctions, and 600 gap junctions connect the nervous system. The simplicity of the C.

elegans nervous system will facilitate the identification of the mechanism of action of

volatile anesthetics.

Developing a new model organism to explore the genetics of behavior and the

nervous system was the aim of Sidney Brenner when he began working with the

nematode Caenorhabditis elegans in 1963. To identify the genes involved in the C.

elegans nervous system, Brenner first isolated mutants with uncoordinated locomotion.

This uncoordinated class of mutants was named unc. Validating Brenner’s goals for C.

elegans, a great degree of conservation within this family of unc genes has been revealed between the C. elegans genes and genes associated with the human nervous system.

Most receptors, neurotransmitters, and synaptic vesicle proteins are highly conserved between C. elegans and humans (Hosono, Sassa et al. 1987; Ogawa, Harada et al. 1998;

Richmond and Jorgensen 1999). Also conserved between C. elegans and humans is a

sensitivity to volatile anesthetics. In humans, volatile anesthetics are used for the

induction of unconsciousness, immobility, amnesia, and analgesia. Due to the difficulties

in measuring consciousness, pain, and memory in C. elegans, immobility is used to quantify the effect of volatile anesthetics. In both humans and C. elegans, lethality of volatile anesthetics is 2-3 times the concentration needed for immobility. The conservation of genes of the nervous system and of sensitivity to volatile anesthetics between C. elegans and humans is the foundation of our work. The identification and

characterization of C. elegans genes that modify volatile anesthetics will be relevant to

the understanding of the mechanism of action of volatile anesthetics in humans.

14 To give context to our studies and the known C. elegans genetic modifiers of volatile

anesthetics, I will briefly review the previous genetic studies of volatile anesthetics from

in vitro studies of individual genes to mice. This review represents the current

understanding of volatile anesthetics and will not produce a coherent picture of the

mechanism of volatile anesthetics. Divergent themes are intentionally introduced to best represent the numerous pieces of data that lay outside all current models of volatile anesthetic action. The goal of this review is to convey the complexity of the mechanism of volatile anesthetics. Volatile anesthetics do not appear to act through a mechanism of a simple ligand and a receptor working together as a key and a lock. The future understanding of the function of volatile anesthetics will likely permit a model which includes all of the experimentally observed effects of volatile anesthetics. In the absence of unifying model with which to organize this data, the genetic studies of volatile anesthetics are presented by organism.

C) A review of the genetic studies of volatile anesthetics

1) in vitro targets of volatile anesthetic

The identity of the target of volatile anesthetics has been pursued by in vitro methods resulting in a variety of candidate targets. Expressing candidate genes such as acetylcholine receptors and G protein-coupled receptors in Xenopus has found many in vitro effects that fail to correlate with significant effect on the total organism (Honemann,

Nietgen et al. 1998; Nietgen, Honemann et al. 1998; Nietgen, Honemann et al. 1998).

Mammalian electrophysiological studies have implicated multiple classes of channels including potassium, sodium, and chloride channels (Jones, Brooks et al. 1992;

15 Ratnakumari and Hemmings 1998; Sirois, Lei et al. 2000) as targets of volatile anesthetic

function. While successful in identifying molecules with altered function in the presence

of volatile anesthetics, these studies have not yielded targets that are clearly linked with

the organismal responses to anesthetic. The presence or absence of some of these

candidate channels and receptors has proven to have no effect on the response of the

organism to volatile anesthetics.

These studies demonstrate that volatile anesthetics have the ability to affect the function of many targets in vitro. The absence of altered volatile anesthetic sensitivity in whole animals with single knockouts of these targets suggests three possibilities. First,

studies of the in vitro effects of volatile anesthetics have no relevance to the mechanism

of volatile anesthetics in whole animals. It is unlikely that the observed in vitro actions

of volatile anesthetics have no relevance, but the meaning of these findings is unclear.

Second, the effect of volatile anesthetics on whole animals targets multiple sites such that

the knockout of a single channel is insufficient to change the cumulative actions on the

other targeted sites. Even if this is true, the value of in vitro studies for identifying

proteins that respond to volatile anesthetics is limited. The absence of effect in the whole

animal of some of the in vitro identified targets of volatile anesthetics is a strike against the significance of the in vitro identified proteins. Third, the in vitro identified targets are

not relevant for understanding the response of the animal but the mechanism by which

volatile anesthetics affect targets in vitro is the same mechanism that acts on the relevant target in the animal. The third possibility is the most likely and indicates the perspective

with which the in vitro studies should be viewed. The in vitro studies are not perfectly

clean combinations of candidate channels and volatile anesthetics. In the Xenopus

16 oocyte, for instance, the artificially expressed channel is surrounded by the cellular machinery of the oocyte. The in vitro identified targets most likely share a common association within the oocyte such as a specific protein processing complex or a similar membrane microenvironment. In short, the in vitro studies support that the mechanism of volatile anesthetic response is complex and not similar to a key and lock model.

Genetic studies of volatile anesthetics in model organisms have the advantage over in vitro studies by starting with a physiological significance of identified targets. One limitation of genetic studies is the selection of endpoints. The potential implications of different end points are discussed in the following discussion of volatile anesthetics studied in genetic systems.

2) Genetic studies of sensitivity to volatile anesthetics in yeast

Using growth inhibition as an endpoint, yeast were screened for resistance to volatile anesthetics (Keil, Wolfe et al. 1996). Five volatile anesthetics, isoflurane, halothane, enflurane, sevoflurane, and methoxyflurane, were tested. Potency as a growth inhibitor of yeast correlates with lipophilicity as predicted by the Meyer-Overton rule. The growth inhibition is both dose-dependent and reversible. Several genes which when mutated confer resistance to the growth inhibition of volatile anesthetics were identified. The first two isolated genes are associated with ubiquitin metabolism. BUL1/ZZZ1 has been shown to bind ubiquitin ligase by yeast two hybrid analysis (Yashiroda, Oguchi et al.

1996). The other gene, DOA1/UFD3/ZZZ4, is involved in ubiquitin-dependent protein degradation. Additional sensitivity modifying genes have linked anesthetic response with

17 transcription regulation in response to nutritional availability of amino acids

(Palmer, Wolfe et al. 2002; Palmer, Shoemaker et al. 2005).

The genes identified in yeast introduce the mechanisms of protein processing as

determinants of volatile anesthetic response. The association of protein processing with

volatile anesthetics may be conserved in C. elegans. Introduced in the next paragraph and discussed in depth in section D, the volatile anesthetic sensitivity modifier unc-1 belongs to a family of genes associated with protein processing. Similar to the in vitro studies, the studies in yeast support a complex mechanism of volatile anesthetic response.

3) Genetic studies of sensitivity to volatile anesthetics in C. elegans as measured by

immobility

Immobility is the endpoint of volatile anesthetic response that is used in this work.

Immobility in C. elegans is used as a behavior analogous to insensibility to a surgical stimulus in mammals (Morgan, Sedensky et al. 1990; Kayser, Rajaram et al. 1998). The merits of the endpoint of immobility relative to other endpoints in C. elegans will be discussed in the next section. Mutants sensitive to volatile anesthetics as measured by immobility can be divided into three classes. The first group is sensitive to thiomethoxyflurane, methoxyflurane, chloroform, halothane and diethylether, but does not change sensitivity to enflurane, isoflurane, fluroxene, or flurothyl. unc-79 and unc-

80 are included in this group (Figure 1). unc-80 and unc-79 genes have yet to be

characterized. The second group is able to suppress the sensitivity produced by unc-79.

unc-1, unc-8, unc-24, unc-7, and unc-9 belong to this group. unc-8 is an amiloride

sensitive sodium channel (ENaC) and unc-1 and unc-24 are stomatin-like gene

18 homologues. unc-7 and unc-9 are innexins, components of gap junctions. The final class

is sensitive to all anesthetics. gas-1 is a representative of this group and codes for a

mitochondrial protein.

Figure 1. Genetic determinants modify volatile anesthetic sensitivity in C. elegans The histogram shows relative volatile anesthetics sensitivities of unc-79 and unc-79;unc- 1 relative to N2. For each of the nine different volatile anesthetics, the concentration needed to immobilize 50% of the animals (EC50) was determined for the three different genotypes N2, unc-79, and unc-79;unc-1. The unc-79 mutation increases sensitivity to five of the nine volatile anesthetics relative to N2. Four anesthetics are equally potent in N2 relative to unc-79. The genetic suppressor unc-1 restores normal sensitivity to all volatile anesthetics tested except for diethyl ether. The different sensitivities of animals with these mutations divide the volatile anesthetics into three groups: anesthetics with responses in unc-79 that are more potent than N2, not changed relative to the response of N2, or not suppressed by unc-1.

19 unc-1, unc-8, and unc-24 are a focus of this paper and will be discussed in depth

later. The most interesting features of the other genes associated with the endpoint of

immobility are discussed next.

unc-79 –unc-79 modifies sensitivity to volatile anesthetics in a manner that deviates from the Meyer-Overton rule. Independent of the predicted potency of volatile anesthetics, unc-79 increases sensitivity to some volatile anesthetics more than others. If lipid solubility was the sole determinant in potency, unc-79 would be expected to alter sensitivity to all volatile anesthetics similarly. Instead, the anesthetic sensitivity of unc-

79 divides volatile anesthetics into three groups: anesthetics with responses in unc-79 that are more potent than N2, not changed relative to the response of N2, or not suppressed by unc-1 (Figure 1). The ability to alter sensitivity to some anesthetics but not others suggests that unc-79 functions in a specific fashion to alter sensitivity.

Characterizing the function of unc-79 should be very informative about the mechanism of volatile anesthetics. Since unc-1 and unc-24 are modifiers of unc-79, the characterization of unc-1 and unc-24 will contribute to the characterization of unc-79.

unc-7, unc-9 – Initially, three genes were identified as suppressors of unc-79

(Morgan, Sedensky et al. 1990). One of the genes was unc-1. The other two genes were unc-7 and unc-9. These latter two genes are innexins, the invertebrate equivalent of the gap junction proteins connexins. Unlike unc-1, no allele of either innexin has been found

to increase sensitivity to volatile anesthetics. This difference between unc-1 and both

unc-7 and unc-9 is the main reason that we have chosen to focus on unc-1. The similar

phenotype of suppressing unc-79 suggests a shared function between unc-1, unc-24, unc-

20 7, and unc-9. The characterization of unc-1 and unc-24 may elucidate this shared function with innexins.

gas-1 – While gas-1 does not genetically interact with unc-1 and unc-24, it is important to introduce because it validates our studies of volatile anesthetics in C. elegans and the endpoint of immobility. gas-1 modifies the sensitivity to all volatile anesthetics. A component of mitochondrial complex I, the exact function of gas-1 is not known. Our laboratory has taken the association between mitochondrial complex I dysfunction and volatile anesthetic sensitivity into humans by showing increased sensitivity to volatile anesthetics in individuals with mitochondrial complex I myopathies

(Morgan, Hoppel et al. 2002).

4) Genetic studies of sensitivity to volatile anesthetics in C. elegans as measured by other endpoints

In addition to the endpoint of immobility used in our laboratory, C. elegans screens for volatile anesthetic sensitivity have used other behavioral endpoints which more closely approximate the absolute concentration of anesthetic used in clinical practice.

Called “clinically significant” by the authors of these studies, these concentrations induce only small changes to C. elegans behavior (van Swinderen, Shook et al. 1997). To work with these concentrations, subtle endpoints of volatile anesthetic effects have been studied. The defecation cycle rate and the “dispersal” assay, which measures C. elegans velocity of locomotion, are two of the anesthetic assays. The endpoints are problematic in that they require observation of individual worms for extended periods, limiting the

21 number of worms that can be screened. Screening randomly generated mutants would be

a Herculean task, and has never been attempted using these more subtle endpoints.

While these endpoints quantify some effect of volatile anesthetics, the “clinically

significant” concentrations may not be as significant as maintained. The “clinically

significant” title does not address that C. elegans have an increased resistance to the

lethal effects of volatile anesthetics. Relative to itself, C. elegans mirrors the human

pattern of LC50 being two to three times EC50 when EC50 is determined by immobility

(Figure 2). In the figure, MAC is the minimum alveolar concentration. MAC is the

concentration of anesthetic needed to eliminate movement among 50% of patients

challenged by a standardized skin incision. When considered relative to the LC50, the concentrations of volatile anesthetic used by van Swinderen et al. mirror the concentrations that produce a myriad of subtle effects in humans.

Figure 2. Immobility is a relevant endpoint to study volatile anesthetic sensitivity in C. elegans The schematic depicts the affects of volatile anesthetic relative to dose. Increasing concentrations are represented left to right. In humans, the mean alveolar concentration, MAC, is the amount of volatile anesthetic required to induce surgical anesthesia. Lethality in humans occurs at 2-3 times MAC. Concentrations below MAC are associated with subtle affect on mental performance and smell. Relative to C. elegans lethality in volatile anesthetics, immobility is the most relevant endpoint to study the conserved action of volatile anesthetics. Other endpoints in C. elegans are induced by relative concentrations of anesthetic that are equivalent to the relative concentration that induce the subtle affects in humans.

By screening existing mutants with behavioral endpoints other than immobility, van

Swinderen et al. has identified unc-64, a syntaxin, and goa-1, the alpha subunit of Go, as modifiers of volatile anesthetic sensitivity (van Swinderen, Saifee et al. 1999; van

Swinderen, Metz et al. 2001). There have been no published correlations between the genes involved in immobility and the genes associated with other endpoints of anesthetic

23 response. Discussed later, the work of this thesis will suggest some shared features between the different C. elegans genes and some possible experiments to unite the field.

5) Genetic studies of sensitivity to volatile anesthetics in Drosophila

The first genetic studies of sensitivity to volatile anesthetics in a model organism were done in Drosophila. Ether was administered to Drosophila in an attempt to study the genetic control of behavior (Kaplan and Trout 1969). While not pursuing volatile anesthetic mutants, two of the mutants characterized by Kaplan remain a focus of volatile anesthetic studies today. After 20 seconds in the etherizer, wild type flies lay still while the affect on the behavior of the four mutants is to roll around excitedly, shaking their legs. Ether a go-go (eag) and Shaker (Sh) are two of the four genes identified by Kaplan et al. Both genes code for voltage sensitive potassium channels and have been found to modify response to other volatile anesthetics than ether. Shaker mutations also modify the volatile anesthetic effect on pain sensation. After exposure to volatile anesthetics, responses to the noxious stimulus of heat is diminished significantly in two-day-old adult

Sh mutants (Tinklenberg, Segal et al. 1991).

Studies of ShakerB K+ channel function expressed in Xenopus oocytes demonstrated that volatile anesthetics affect conductance mainly due to changes in the transitions in and out of the open state. Volatile anesthetics differentially affected the kinetics of the return of the gating charge in a mutant non-inactivating channel. CHCl3 and isoflurane had opposite effects while halothane had no effect (Correa 1998).

Similar to studies of immobility in C. elegans, work with Drosophila has linked a sodium channel with sensitivity to volatile anesthetics (Tanaka and Gamo 2001). Like

24 unc-79, other mutations in flies have also been shown to differentiate anesthetics independent of the Meyer-Overton predicted potency (Campbell and Nash 1994). These similarities suggest that the pathways being characterized in C. elegans and Drosophila may eventually be tied together.

Additional Drosophila genes are being genetically mapped and identified. Two ABC transporters that were initially identified by changes in the body color of mutants were shown to change responses to enflurane and halothane as measured by the behavioral test of climbing ability (Campbell and Nash 2001). The halothane-resistant (har) mutants were isolated in 1990 using halothane delivered via the “inebriometer” (Krishnan and

Nash 1990). The inebriometer is a device that uses a series of baffles to separate flies based upon their coordination of flight. The cation channel narrow abdomen (na) is the first har strain to be identified (Nash, Scott et al. 2002).

Patch-clamp comparisons of wild-type (Ore-R) and halothane resistant (har) mutants of Drosophila melanogaster revealed glutamatergic nerves of the larval neuromuscular junction to be immune to halothane suppression of the amplitude of nerve-evoked excitatory junctional currents. The miniature excitatory junctional currents were not reduced in the wild-type worms suggesting halothane’s effect is independent of vesicle packaging (Nishikawa and Kidokoro 1999). Miniature excitatory junctional currents are the invertebrate homolog to miniature end-plate potentials. Recorded spontaneous potentials in the absence of stimulation, miniature excitatory junctional currents measure the postsynaptic potential resulting from a single synaptic vesicle of neurotransmitter.

In addition to neuromuscular junction patch clamping, extracellular recording of light-evoked mass potentials from the surface of the eye has recently been shown to be

25 sensitive to volatile anesthetics (Rajaram and Nash 2004). This technique offers another

means of screening for and characterizing mutants that modify volatile anesthetic

response.

Unlike yeast and C. elegans, all of the genes currently associated with modifying

volatile anesthetic sensitivity in Drosophila are channels. This predominance of channels

may be suggestive of the mechanism of action of volatile anesthetics is to target ion flux.

6) Genetic studies of sensitivity to volatile anesthetics in mice

A genetic screen for mutants sensitive or resistant to volatile anesthetics has not been

completed in mice. The study of volatile anesthetics and mice has centered on the

gamma-aminobutyric acid type A receptor. Electrophysiological studies helped establish

the GABAA receptor as a prime anesthetic target (Franks and Lieb 1994). Also

supporting a role of GABAA receptor as a target for volatile anesthetics, mice with altered

sensitivities to diazepam, which is known to target GABAA receptors, were found to be

sensitive to volatile anesthetics (McCrae, Gallaher et al. 1993). The sensitivity of the

mice to both diazepam and volatile anesthetics suggests both drugs might share the same

target. Since diazepam acts in part through the GABAA receptor, the sensitivity of these

mice suggest volatile anesthetics might act in part through the GABAA receptor. The

data is incomplete as no mutations were reported in the GABAA receptor of these mice.

The GABAA receptor and volatile anesthetics were further studied in mice lacking a

functional alpha6 subunit of the GABAA receptor. Homozygous mice are viable and

fertile and have grossly normal cerebellar cytoarchtecture. Mice were exposed to halothane and enflurane and challenged with the “carousel”, a device that measures

26 balance and motor coordination. There were no differences between the different

genotypes (Homanics, Ferguson et al. 1997). Subtle mutations in GABAA receptor which modify other drug responses, fail to significantly affect volatile anesthetic response

(Jurd, Arras et al. 2003).

Only two mutants have been shown to affect volatile anesthetic response in mice.

Mutations in the glycine receptor have been shown to both increase and decrease sensitivity to enflurane and halothane as measured by several behavioral endpoints

(Quinlan, Ferguson et al. 2002). The second gene is stomatin, the focus of this paper.

Unpublished data in our laboratory has demonstrated that stomatin deficient mice have modified sensitivity to volatile anesthetics as measured by tail-clamp assay, which measures the response to a non-damaging pain stimulus. The study of volatile anesthetics would be greatly served by a true genetic screen for altered anesthetic response in mice.

7) A possibly conserved target associated with malignant hyperthermia

Insights into the function of volatile anesthetics might be found in the mechanism of malignant hyperthermia. Malignant hyperthermia is a state of elevated skeletal muscle metabolism that may occur during general anesthesia in genetically pre-disposed individuals. Elevated temperature, rhabdomyolysis, and cardiovascular instability rapidly develop in response to certain volatile anesthetics. Malignant hyperthermia is introduced within a discussion of genetic studies of volatile anesthetics because the mechanism of action of volatile anesthetics on muscles to produce malignant hyperthermia might be conserved with the mechanism of action of volatile anesthetics on nerves to produce anesthesia. Therefore, the genes associated with malignant hyperthermia might be

27 relevant to the understanding of the mechanism of action of volatile anesthetics on nerves

to produce anesthesia. The physiology of malignant hyperthermia is well understood.

Malignant hyperthermia results from altered control of sarcoplasmic reticulum

Ca2+ release. Mutations have been identified in malignant hyperthermia-susceptible

individuals in two key proteins of excitation-contraction coupling, the Ca2+ release

channel of the SR, ryanodine receptor type 1 and the alpha1-subunit of the

dihydropyridine receptor (L-type Ca2+ channel) (Melzer and Dietze 2001). Dantrolene

sodium is the primary specific therapeutic agent. Dantrolene is used for chronic

spasticity and its effectiveness in malignant hyperthermia appears to be related to its

action on skeletal muscle where it re-associates the excitation-contraction coupling,

probably by interfering with the release of Ca2+ from the sarcoplasmic reticulum (Nelson and Flewellen 1983). Similar to the genetic determinants of volatile anesthetic sensitivity

in C. elegans, Drosophila, and mice, the genes associated with malignant hyperthermia affect ion flux across membranes. The mechanism of malignant hyperthermia is

something to keep in mind when considering the mechanism of action of volatile

anesthetics.

D) Modification of volatile anesthetic sensitivity by stomatin-like proteins

1) The significance of stomatin-like proteins to the study of volatile anesthetics

In 1990, this laboratory linked mutations in unc-1 with C. elegans sensitivity to volatile anesthetics as measured by immobility (Morgan, Sedensky et al. 1990). unc-

1(null) was first identified as a suppressor of the sensitivity to volatile anesthetics generated by unc-79. Additional alleles of unc-1 were found to have the opposite

28 phenotype of increasing sensitivity. The ability to both increase and decrease

sensitivity suggests a pivotal role for unc-1 in the mechanism of volatile anesthetics.

As mentioned above, this laboratory has studied the association of stomatins and volatile anesthetic sensitivity in mammals. Mice deficient of stomatin were shown to have altered volatile anesthetic sensitivity. Similar to unc-1(null) worms, the stomatin deficient mice are more sensitive to ether. This observation in mice validates the relevance of the C. elegans model of volatile anesthetic response as measured by

immobility.

The first of the 131 isolated unc strains, unc-1 codes for a membrane bound protein

with homology to the stomatin-like protein family of proteins (Rajaram, Sedensky et al.

1998). There are ten stomatin-like proteins in C. elegans. Three stomatin-like genes

have been associated with neuronal phenotypes (Figure 3) (Barnes, Jin et al. 1996;

Huang, Gu et al. 1995). The locomotion defect of unc-1(null) appears to be

hyperspasticity with random muscle contractions distorting the shape of the worm and

prohibiting the normal sinusoidal locomotion. Loss of function alleles of unc-1, e580,

fc53 and e114, are able to suppress the increased sensitivity of unc-79 and unc-80. A

dominant allele of unc-1, n494, results in worms being immobilized by one third the concentration of volatile anesthetic needed to immobilize wild type worms. Worms carrying both e580 and unc-79 or unc-80 require an equivalent amount of volatile anesthetic as wild type worms to be immobilized. A temperature sensitive allele of unc-

1, hs1, can induce or reverse the unc-1 phenotype roughly eight hours after being moved into the restrictive or permissive temperatures, respectively. All alleles of unc-1 are associated with an increased sensitivity to ether (Rajaram, Sedensky et al. 1998). This

29 link to loss of coordination and altered sensitivity to volatile anesthetics suggests a significant role for unc-1 in neuron function.

Figure 3. The three stomatin-like proteins associated with neuronal phenotypes in C. elegans UNC-1, UNC-24, and MEC-2 share a 160aa similarity with stomatin-like proteins. The amino and carboxy termini of stomatin-like proteins differ between the different proteins. The carboxy terminus of UNC-24 shares similarity with a SCP2 sterol binding domain (blue). Of the ten predicted stomatin-like proteins in C. elegans, these three proteins are the only C. elegans stomatin-like proteins that have been associated with neuronal phenotypes.

2) A review of stomatin-like proteins

The stomatin-like protein family gets it name from the human anemia stomatocytosis.

The fragile red blood cells in this disorder often resemble stoma, στοµα, the Greek word for lips. Patients with stomatocytosis are deficient in erythrocyte surface protein band 7.2, stomatin (Lande, Thiemann et al. 1982; Stewart, Hepworth-Jones et al. 1992).

Despite the absence of the stomatin protein in these individuals, no mutations have been

30 found in the stomatin gene. In addition, a mouse knockout for stomatin does not have red

blood cell abnormalities (Zhu, Paszty et al. 1999). This suggests that there is a second

unknown gene that needs to be mutated to cause the loss of stomatin from red blood cells and to cause stomatocytosis. In addition, it suggests that the mere absence of stomatin from the red blood cell membrane is not sufficient to cause the red blood cell abnormalities of stomatocytosis. Recently, a variant of stomatocytosis with a neurological component was described (Fricke, Jarvis et al. 2004). In addition to the red blood cell abnormalities and a lack of stomatin protein in red blood cells, the individuals experienced mental retardation and seizures.

A second human stomatin-like protein, podocin, is mutated in autosomal recessive steroid-resistant nephrotic syndrome (Boute, Gribouval et al. 2000). The normal location of podocin is in the podocytes of the glomerulus facing the slit diaphragm

(Roselli, Gribouval et al. 2002). This site in the glomerulus is a filter that serves as the portal between the blood and the tubular fluid of the nephron. The pathogenic mechanisms of podocin-associated nephropathy and stomatocytosis remain uncertain.

The molecular mechanism of stomatin-like proteins is not well understood. Studies in model organisms offer some insights into the molecular function of stomatin-like proteins. The common feature of all stomatin-like proteins across all phyla is a central domain of 150 conserved amino acids (NCBI conserved domain smart00244) (Marchler-

Bauer, Anderson et al. 2005). Every species has multiple stomatin-like proteins. Unique

carboxy and amino ends differentiate different stomatin-like proteins. In bacteria, there are two stomatin-like proteins called HflK and HflC (Kihara, Akiyama et al. 1997).

Originally isolated as a single locus named high frequency lysogenization A (HflA) for

31 the ability to modify the lytic/lysogenization decision of bacteriophage, HflKC form a large complex with an AAA protease named FtsH, also known as HflB. Together, the stomatin-like proteins HflKC and protease FtsH regulate the stability and localization of native transcription factors involved in stress response (Kihara, Akiyama et al.

1998). The stress responding function of the stomatin-like proteins HflKC is used by the bacteriophage. The bacteriophage’s decision to integrate into the bacterial genome to wait for better times or to replicate when conditions are good is decided by the stomatin- like proteins HflKC. The transcription factor CII regulates transcription to activate either the lytic or lysogenization pathways. The stability and availability of CII is regulated by stomatin-like proteins HflKC and protease FtsH. This ancestral role for stomatin-like proteins is suggestive of both a chaperone function and a possible stress response role for stomatin-like proteins in multicellular organisms.

In yeast, the two stomatin-like proteins are called prohibitins, Phb1 and Phb2 (Berger and Yaffe 1998). Like bacterial stomatin-like proteins, prohibitins also form large molecular weight complexes that are associated with AAA proteases (Steglich, Neupert et al. 1999). Also like bacteria, Phb2 has been shown to also regulate transcription through interactions with transcription factors (Sun, Liu et al. 2004). The prohibitins have also been associated with chaperoning newly made protein complexes into the mitochondria (Coates, Jamieson et al. 1997). Part of the mitochondrial chaperone role is the protection of newly synthesized peptides against proteolysis before the peptides can be assembled into active mitochondrial complexes (Nijtmans, de Jong et al. 2000). It is interesting that both the volatile anesthetic studies in yeast and the stomatin-like protein function in yeast share the common mechanism of protein processing. The association of

32 protein processing with both of these yeast traits is suggestive of a protein processing function for neuronal stomatin-like proteins in the modulation of volatile anesthetic response. The conserved association of stomatin-like proteins with protein processing between bacteria and yeast suggests that this relationship should carry forward into metazoans. While no published work has reported a protein processing function for neuronal stomatin-like proteins, our data presented here does not exclude this possibility.

The associations with stomatin-like proteins that are most clearly relevant to the study of unc-1 were first identified in worm studies of mechanosensation. Stomatin-like proteins were found to act in neurons to regulate the function of sodium channels.

Neuronal stomatin-like gene mec-2 and the sodium channels mec-4 and mec-10 were shown to genetically interact in mechanosensation (Huang, Gu et al. 1995).

Mechanosensation in C. elegans is measured by observing the response to touch with an eyelash. The regulation of the sodium channels MEC-4 and MEC-10 by the stomatin- like protein MEC-2 has been observed with Xenopus oocyte injection studies (Goodman,

Ernstrom et al. 2002). These studies involve patch-clamp electrophysiology recordings of MEC-4 and/or MEC-10 channels being artificially expressed in the Xenopus oocyte.

Coexpression with the stomatin-like protein MEC-2 greatly enhances cation transport through the channels. In addition, the stomatin-like protein MEC-2 and the sodium channels MEC-4 and MEC-10 have been shown to bind together in co- immunoprecipitation studies and to colocalize within the six mechanosensory neurons

(Zhang, Arnadottir et al. 2004). The stomatin-like protein UNC-1 has been shown to interact with the ENaC ortholog, UNC-8 (Rajaram, Spangler et al. 1999). A third neuronal stomatin-like protein, UNC-24, interacts with both of the neuronal stomatin-like

33 proteins and their respective phenotypes (Zhang, Arnadottir et al. 2004; Sedensky,

Siefker et al. 2001). In addition to genetically interacting with the volatile anesthetic

modifying sodium channel UNC-8, the stomatin-like protein UNC-24 has been shown to

colocalize with MEC-4 and MEC-10 (Zhang, Arnadottir et al. 2004).

The interaction of stomatins and ENaCs has also been identified in mammals.

Mammalian stomatin and ENaCs coexpressed in neurons of the mouse dorsal root ganglion (Mannsfeldt, Carroll et al. 1999). Interactions between stomatin-like proteins

and epithelial sodium channels have been described in both C. elegans and rats (Fricke,

Lints et al. 2000; Sedensky, Siefker et al. 2001). Stomatin has been shown to modulate functioning of the ENaCs in tissue cultured mammalian cells expressing stomatin and a subset of ENaCs called acid sensitive channels, ASICs (Price, Thompson et al. 2004).

These mammalian studies support the relevance of the relationship between ENaCs and

neuronal stomatin-like proteins.

The final feature of stomatin-like proteins is an association with dynamic vesicles.

Two types of human blood cells mobilize stomatin-rich vesicles in response to stimuli.

Red blood cells shed stomatin-rich microvesicles in response to Ca2+ (Salzer,

Hinterdorfer et al. 2002), and the activation of platelets mobilizes stomatin-rich alpha

granules to fuse with the plasma membrane (Mairhofer, Steiner et al. 2002). Supporting

the importance of stomatin in vesicle dynamics, red blood cells from individuals with

severe stomatocytosis were found to lack ATP-dependent endocytic vesiculation (Turner,

Jarvis et al. 2003). The ease of purifying intact tissue makes mammalian blood tissue the

best system in which to look for stomatin-like protein associations with dynamic vesicles.

The lack of similar observations in other models and systems is most likely due to

34 experimental challenges and not due to lack of conserved function. Though the dynamic changes in distribution of stomatin-like proteins have eluded study in other organisms, the importance of cellular distribution has been shown to be important for the function of stomatin-like proteins in other organisms. While there is no visual demonstration of dynamic changes in stomatin-like protein distribution in bacteria and yeast, the bacterial- and yeast-associated functions of protein processing have many steps that are dependent on trafficking. In C. elegans, the cellular distribution of unc-1 is dramatically affected by mutations in unc-24 (Figure 4) (Sedensky, Siefker et al. 2001). Antibody staining of

UNC-1 in unc-24 null animals reveals faint staining next to the nucleus instead of the bright wild type staining along the neuronal processes. This data suggests C. elegans neuronal stomatin-like proteins participate in regulating cellular distribution similar to stomatin-like proteins in mammalian blood tissues. In short, protein trafficking is a shared function of several stomatin-like proteins and is likely to be involved in the function of neuronal stomatin-like proteins.

Figure 4. Neuronal stomatin-like proteins affect cellular distribution of other neuronal stomatin-like proteins Two fluorescent images of the midsection of N2 and unc-24(null) worms show different anti-UNC-1 antibody staining patterns. The cartoon of the worm highlights the imaged section of the worm. The spermatheca (Sp) can be seen as a blue oval in the cartoon. UNC-1 antibody staining is red and DAPI nuclear staining is blue. The staining of UNC- 1 protein is greatly diminished in the unc-24(null) animals. The increased exposure to capture the staining pattern in the unc-24(null) animals can be seen in the greatly increased amount the background staining in the image. In the N2 animal, the UNC-1 staining is seen as numerous, bright, punctate dots along the dorsal and ventral nerve cords (DC and VNC). In the unc-24(null) animal, the UNC-1 staining is limited to small, paired structures juxtaposed to the nuclei (N). This data shows that neuronal stomatin- like proteins can affect the cellular distribution of proteins.

To review, stomatin-like proteins are associated with at least two human syndromes, an anemia with neurological variant and a nephrotic syndrome. A chaperone function associated with proteolysis and substrate delivery/availability is conserved in bacteria and yeast. Neuronal stomatin-like proteins have been shown to regulate sodium channels in

36 both worms and mammals. Finally, mammalian stomatin has been associated with dynamic vesicles. It is unclear whether stomatin-like proteins share a unifying function in different species or if they simply share a common membrane binding region with widely varying functions. The understanding of stomatin-like proteins needs to be improved to better understand the role of stomatins in human disease and in our pursuit of the mechanism of volatile anesthetics. Since the function of neuronal stomatin-like proteins is unclear, we reasoned that identification of mutations which suppress loss of stomatin function might identify the role of stomatin-like proteins.

37 II Genetic suppressors of the stomatin unc-1

A) The rationale for using genetic suppressors to study unc-1

Genetic suppressors of the locomotion defect of unc-1(null) were generated to identify genes with known functions that might elucidate the function of neuronal stomatin-like proteins. A genetic suppressor is a mutated gene that is able to reverse the phenotype caused by mutations in a second gene of interest. The Unc-1 locomotion phenotype was used in the screen instead of volatile anesthetic sensitivity phenotype to expedite the screen. The isolated genetic suppressors have since been shown to also suppress the Unc-1 volatile anesthetic sensitivity phenotype. A genetic suppressor of a null mutation of unc-1 by its nature selects for indirect interactions with unc-1. It is hoped that the genetic suppressor will target the pathway disrupted by the absence of unc-

1, and that the function of the genetic suppressor will indirectly characterize the function of unc-1.

Genetic suppressors of unc-1 worms were created by EMS mutagenesis. F2 worms were screened for recessive suppressors of the unc-1 locomotion defect. The improved locomotion is easily seen as two successful waves propelling the worm forward or backwards. The unc-1(null) rarely completes a full sinusoidal wave. Out of 20,000 animals screened, four genetic suppressors were identified. The identified mutants were given the allele name of our lab, Forest City (fc). Of the four genetic suppressers, two of them, fc70 and fc72, are alleles of unc-8. The other two genetic suppressors, fc71 and fc73, are not allelic with any of the identified genetic suppressors. With only one allele identified for two of the three genetic suppressors, the screen is not considered saturated.

38 Suppressor of Stomatin Uncoordination (ssu) is the gene class for the new genes identified by this screen.

The two genetic suppressor genes, ssu-1(fc73) and ssu-2(fc71), have been cloned and characterized. The phenotypes of each gene are discussed below. The molecular characterization of each gene will be discussed in sections III and IV. Pictorial representation of the locomotion phenotypes of these worms is problematic. Watching worms under a microscope is the best way to appreciate their differences in locomotion.

Crosses and microinjection rescue studies were scored blinded by more than one observer to avoid bias in scoring the locomotion phenotypes. Quantification of the locomotion phenotype was attempted using a twitch assay in which the movement of the animal in buffer is observed for two minutes. A “twitch” is recorded when the animal moves its head and tail both across and back the center line of its body. Wild type worms complete this motion at a very rapid rate (greater than 200 twitches in two minutes). Both unc-1 and unc-24 animals move so poorly that they only score the occasional twitch when a convulsion bends the body in a way that meets the definition of a twitch (less than 10 twitches in two minutes). While able to demonstrate the improved locomotion of the genetic suppressors ssu-1(fc73) and ssu-2(fc71), the twitch assay fails to accurately represent the improvement because of variability in attempts to move by the mutant worms. While animals of the same genotype twitch at a similar pace when moving, some animals move for the entire two minutes while others would only attempt to move for less than 30 seconds. This tendency to stop moving in buffer is not consistent with the locomotion of the mutants on agar where they do not rest for extended periods. Because the tendency to stop was a feature specific to movement in buffer, the twitch assay was

39 not performed exhaustively and only two worms from each genetic background were

observed (Figure 5). Nonetheless, this twitch assay does pictorially represent the improved locomotion of both unc-1(null) and unc-24(null) when in combination with

ssu-1(fc73) and ssu-2(fc71). Additionally, visual observation of the movement on agar concurs with the twitch assay’s representation of greater genetic suppression of the

locomotion defect of unc-1(null) relative to unc-24(null).

Figure 5. The twitch assay measurement of the movement phenotypes of the genetic suppressors ssu-1 and ssu-2 A twitch assay was used to quantify the movement phenotypes of the genetic suppressors ssu-1 and ssu-2. Individual worms were placed into buffer and their movement was observed for two minutes. A "twitch" was recorded when the head and tail moved both across and back the midline of the animal. Error bars show the standard deviation from the mean. This assay is less than ideal because of a tendency of the observed worms to not attempt to move for extended periods. The lack of attempts at movement is not a feature of these animals when moving on agar. Because of this assay-specific problem, only two worms per genotype were observed. The numbers do grossly represent the compromised locomotion phenotypes of unc-1(null) and unc-24(null) animals and the improved movement when in combination with the genetic suppressors ssu-1 and ssu-2.

40 B) A detailed phenotypic description of the genetic suppressor ssu-1

ssu-1(fc73) is a recessive suppressor of the locomotion phenotypes of both stomatin- like genes unc-1(null) and unc-24(null). ssu-1(fc73) also suppresses the volatile anesthetic resistance of unc-1(null) (Figure 6). The mechanosensation phenotype of the stomatin-like gene mec-2(null) is not suppressed by ssu-1(fc73). No phenotype has been observed in worms carrying only ssu-1(fc73). Only in the presence of unc-1 or unc-24 is the action of ssu-1(fc73) observed. ssu-1(fc73) was created and is maintained in hermaphrodites. It was only during attempts to mate ssu-1(fc73) that it was observed that ssu-1(fc73) fails to suppress unc-1 locomotion in male worms. Due to the challenges of maintaining male cultures of severely uncoordinated worms (which fail to mate), measuring the anesthetic sensitivity of male ssu-1(fc73);unc-1 has not been attempted. It is a more complete suppressor of uncoordination than ssu-2 and it does not add any distinctive movements.

Figure 6. ssu-1 modifies volatile anesthetic sensitivity.

The bars depict EC50 in halothane as measured by immobility for several worm strains. Error bars show the standard deviation from the mean EC50. Wild type C. elegans are immobilized by 3.2% halothane. The sensitivity created by mutations in unc-79 can be suppressed by unc-1. The triple mutant unc-79;unc-1;ssu-1 has increased sensitivity as a result of ssu-1 suppressing the volatile anesthetic resistance of unc-1. Individually, unc- 1(null) and ssu-1(fc73) do not modify anesthetic sensitivity. Modifying volatile anesthetic sensitivity is one of three neuronal stomatin phenotypes that can be suppressed by ssu-1.

Dauer is a larval stage in C. elegans that forms in response to adverse environmental conditions. An increased incidence of dauer larvae was found in a double mutant unc-

24(eDf28) and moc-1(ok366), a chemically mutated strain created by the C. elegans Gene

Knockout Consortium. The phenotype was not linked to the moc-1(ok366) mutation, but to an additional mutant in the background of moc-1(ok366). The additional mutation was

42 given the allele name fc83 and was mapped to chromosome III. unc-24, but not unc-1,

interacts with fc83 to increase dauer formation to greater than 95% of the offspring at

20ºC. The unc-24(eDf28);fc83 animals that enter dauer maintain dauer for variable

durations from less than 24hrs to indefinitely. Two unc-24(eDf28);fc83 animals were

observed to escape from dauer after six months. ssu-1(fc73) reverses the dauer

phenotype of unc-24(eDf28);fc83 so that less than 5% of offspring form dauer. To

maintain unc-24(eDf28);fc83, adult worms but not dauer larvae are moved to fresh plates.

Adult unc-24(eDf28);fc83 have roughly wild type fecundity. The second suppressor of

unc-1, ssu-2, does not suppress the dauer constitutive phenotype of unc-24(eDf28);fc83.

This phenotype is further discussed in section III.

C) A detailed phenotypic description of the genetic suppressor ssu-2

ssu-2(fc71) is a semi-dominant suppressor of unc-1. ssu-2(fc71);unc-1 suppresses

unc-1 uncoordination much better than +/ssu-2(fc71);unc-1. Both +/ssu-2(fc71);unc-1

and unc-1 maintain a contorted body posture that makes visual separation impossible.

They are easily distinguished by comparing successful locomotion over a 24 hour period.

Successful locomotion is clearly recorded in the tracks in the bacterial lawn of the normal

growth conditions. The heterozygote +/ssu-2(fc71);unc-1 enhances locomotion

sufficiently that there are multiple tracks crisscrossing the bacterial lawn. By itself, unc-1

will not significantly move from the origin in 24 hours. ssu-2(fc71) suppresses the

locomotion phenotypes of both stomatin-like genes unc-1(null) and unc-24(null). ssu-

2(fc71) also suppresses the volatile anesthetic resistance of unc-1(null) (Figure 7). The mechanosensation phenotype of the stomatin-like gene mec-2(null) is not suppressed by

43 ssu-2(fc71). Similar to ssu-1(fc73), ssu-2(fc71) fails to suppress the Unc-1 locomotion

phenotype in males.

Figure 7. ssu-2 modifies volatile anesthetic sensitivity. The bars depict EC50 in halothane as measured by immobility for several worm strains. Error bars show the standard deviation from the mean EC50. Wild type C. elegans are immobilized by 3.2% halothane. The sensitivity created by mutations in unc-79 can be suppressed by unc-1. The triple mutant unc-79;unc-1;ssu-2 has increased sensitivity as a result of ssu-2 suppressing the volatile anesthetic resistance of unc-1. Individually, unc- 1(0) and ssu-2(fc71) do not modify anesthetic sensitivity. Modifying volatile anesthetic sensitivity is one of two neuronal stomatin phenotypes that can be suppressed by ssu-2.

ssu-2(fc71) also adds two distinctive features of locomotion when combined with unc-1. First, ssu-2(fc71);unc-1 display a mild coiler phenotype. This pattern of movement looks like the worm is winding up into a circle. The motion starts from either

44 end and pulls the body into the expanding circle. This motion appears to be the complete

dominance of an entire side of muscles. It appears that the coordination of alternating

contraction and relaxation of the ventral and dorsal muscle is momentarily lost, and only

the ventral side gets the successive commands to contract while the dorsal side remains

flaccid. We say that ssu-2(fc71) is a mild coiler because this movement does not happen

very frequently, and when it does, the worms recover within a second or two. The

second feature of ssu-2(fc71) is piling of the bacterial lawn. This pattern can be observed

by eye. The bacterial lawn is pushed into hills and furrows to result in an evenly spaced

grid of 1mm2 piles. It is the only worm that I can score without a microscope.

The second allele ssu-2(tm1633) is recessive. ssu-2(fc71) over ssu-2(tm1633) is able to suppress unc-1(null). Additionally, ssu-2(tm1633);unc-1 is also suppressed for locomotion. ssu-2(tm1633) has a stronger coiler phenotype than fc71. The majority of ssu-2(tm1633);unc-1 worms are stuck in coils. The coiling phenotype of tm1633 is

severe enough to make working with the semidomant fc71 preferable in microinjection

rescue studies. ssu-2(tm1633) does not form hills of bacteria as does ssu-2(fc71). ssu-

2(tm1633) only recently arrived in the lab and this allele has not been thoroughly

characterized. The interactions of ssu-2(tm1633) with unc-24 and mec-2 have not been

explored. The effects on anesthetic sensitivity of ssu-2(tm1633) have not been assayed.

III Sulfonation of a Signal from Sensory Neurons Modulates Stomatin-Dependent

Coordination in C. elegans

Bryan Carroll, Margaret Sedensky, Philip Morgan

Department of Genetics, Case Western Reserve University

and Department of Anesthesiology, University Hospitals of Cleveland

Reference: Paper Submitted to Nature Neuroscience

46 Abstract

Stomatins form a large family of membrane bound proteins conserved from bacteria through humans (Nijtmans, Artal et al. 2002; Stewart and Fricke 2003). In humans, defects in stomatin-like proteins are associated with hematopoetic, nephrotic, and neurological disorders (Lande, Thiemann et al. 1982; Boute, Gribouval et al. 2000;

Fricke, Jarvis et al. 2004). Mutations in neuronal stomatin-like proteins in C. elegans

alter sensitivity to volatile anesthetics (Rajaram, Sedensky et al. 1998). While stomatin-

like proteins in bacteria and yeast have well-studied chaperone relationships with AAA-

proteases (Kihara, Akiyama et al. 1997; Steglich, Neupert et al. 1999), little is known

about the function of stomatin-like proteins in metazoans. We have explored the function

of stomatin in the nematode C. elegans through genetic studies of genes which interact

with neuronal stomatin genes, unc-1 and unc-24. A suppressor of unc-1 and unc-24,

named ssu-1(fc73) (for suppressor of stomatin uncoordination), suppresses three

phenotypes of neuronal stomatin deficiency: volatile anesthetic sensitivity, uncoordinated

locomotion, and a constitutive alternative developmental phenotype (known as dauer).

Here we provide the first characterization of ssu-1, predicted to be the only C. elegans

cytosolic sulfotransferase. In vitro enzyme analysis of bacterially expressed SSU-1

demonstrates sulfotransferase activity. While unc-1 is expressed in the majority of

neurons, expression of SSU-1 protein in only the two ASJ amphid interneurons is

sufficient to restore the wild type phenotype. The predicted small molecule sulfonation

activity of ssu-1 is consistent with distinct sites of unc-1 and ssu-1 function. The data

support the role of the cytosolic sulfotransferase in C. elegans as a systemic modifier of

neuronal phenotypes.

47 Introduction

The common feature of all stomatin-like proteins across all phyla is a central domain of

150 conserved amino acids (NCBI conserved domain smart00244) (Marchler-Bauer,

Anderson et al. 2005). Unique carboxy and amino ends differentiate individual stomatin-

like proteins. It is unclear whether stomatin-like proteins share a unified function in different species or if they simply share a common membrane binding region with widely varying functions. The current understanding of neuronal stomatin-like protein function in C. elegans is limited to an association between stomatin-like proteins and epithelial sodium channels, ENaCs (Sedensky, Siefker et al. 2004; Zhang, Arnadottir et al. 2004).

The stomatin-like gene mec-2 has been shown to modulate functioning of the ENaCs involved in mechanosensation in C. elegans (Zhang, Arnadottir et al. 2004). In addition, the stomatin-like protein UNC-1 has been shown to interact with the ENaC ortholog,

UNC-8 (Rajaram, Spangler et al. 1999). A similar relationship has been observed in tissue cultured mammalian cells expressing stomatin and a subset of ENaCs called acid sensitive channels, ASICs (Price, Thompson et al. 2004). In line with a potential cation regulatory role for stomatin-like proteins, abnormal red blood cell electrolyte concentrations are seen in stomatocytosis, an anemia associated with the loss of stomatin protein from red blood cells (Lande, Thiemann et al. 1982). However, the role of stomatin in the normal functioning of sodium channels is unknown. Since the function of stomatin-like proteins is unclear, we reasoned that identification of mutations which suppress loss of stomatin function might identify the role of stomatin-like proteins.

48 Of the ten stomatin-like proteins in C. elegans, we study the neuronal stomatin-like

proteins UNC-1 and UNC-24 because of their ability to modify sensitivity to volatile anesthetics as measured by immobility (Sedensky, Siefker et al. 2001). Immobility in C.

elegans has been used as a behavior analogous to insensibility to a surgical stimulus in

mammals (Morgan, Sedensky et al. 1990; Kayser, Rajaram et al. 1998). UNC-1 and

UNC-24 are also required for normal neuronal coordination of locomotion. By studying

genes that affect these behaviors, we hope to both better understand volatile anesthetic

action and also to understand conserved neuronal mechanisms.

Results

Isolation of the genetic suppressor ssu-1

To study the function of unc-1, we created a genetic suppressor of unc-1. Different

alleles of the pan-neuronal stomatin unc-1 (Rajaram, Spangler et al. 1999) can increase or

decrease volatile anesthetic sensitivity (Rajaram, Sedensky et al. 1998) and disrupt

coordination of locomotion (Brenner 1974; Park and Horvitz 1986). For our genetic

suppressor studies, the null unc-1(e580) was chosen to avoid selection of additional

alleles of unc-1. unc-1(e580) suppresses the volatile anesthetic sensitivity of a second

gene, unc-79 (Figure 6). We mutagenized unc-1(e580) and selected for genetic

suppressors of the Unc-1 uncoordinated locomotion phenotype. Isolated as suppressors

of the locomotion defect of unc-1, the gene class was given the name Suppressor of

Stomatin Uncoordination (ssu). One of the isolates, ssu-1(fc73), suppresses the volatile

anesthetic sensitivity and locomotion defect of unc-1(e580) and the locomotion defect of

another neuronal stomatin, unc-24(eDf28) (also a null mutation) (Figure 6). ssu-1(fc73)

49 is not a universal suppressor of stomatins; it fails to suppress the mechanosensation

defect of mec-2(e75). Curiously, ssu-1(fc73) also fails to suppress the unc-1 locomotion

defect in male worms.

Identification of ssu-1

ssu-1 was mapped on the end of chromosome V. Sequencing identified a mutation in

fc73 that codes for a premature termination at amino acid 485 in the predicted gene

Y113G7A.11 (Figure 8). RT-PCR of wild type cDNA identified two spliced forms of

ssu-1: a 1.2kb fragment including the five predicted exons, and a shorter 1kb spliced

form that skips the second exon without disrupting the reading frame. For biological

confirmation that Y113G7A.11 is ssu-1, rescuing constructs were designed for

microinjection into worms. From the identified transcripts, two constructs were created

to encode the long and short spliced forms of Y113G7A.11 (Figure 8). To provide the

native promoter, the constructs included the 500 bp of genomic DNA 5’ of the start of the

first exon. The 500 bp promoter fragment of genomic DNA includes three predicted

potential promoters. Each of these constructs was separately microinjected into ssu-

1(fc73);unc-1(e580). Scored by coordinated movement, only the longer cDNA construct

eliminated the suppression of unc-1(null) by ssu-1(fc73). The shorter spliced form did

not rescue the ssu-1 phenotype. The longer spliced form, but not the shorter form, also

rescued the change in anesthetic sensitivity associated with ssu-1 (data not shown).

Figure 8. ssu-1(fc73) has a mutation in the predicted cytosolic sulfotransferase gene Y113G7A.11 In the schematic, the top bar represents the relative positions of the five exons (blue) of the predicted gene Y113G7A.11. Rescuing constructs are depicted below. a) The positions of two allelic variations of ssu-1 in the predicted gene Y113G7A.11 on chromosome V. The 189bp deletion of tm1117 removes 23bp of the first exon and 166bp 5' of the first exon. The variation in fc73 is at 852bp, TGG -> TGA, which creates a stop codon at 284aa. b) The red bar designates the 60% of SSU-1 that shares similarity with cytosolic alcohol sulfotransferases. Rescuing constructs combined 500bp genomic DNA 5' of the first exon with the two cDNA isoforms. GFP fusions were designed to remove the native ssu-1 stop codon.

A second ssu-1 allele was created by the National Bioresource Project for the

Experimental Animal “Nematode C. elegans” with a screen for genomic deletions. The

189bp deletion in ssu-1(tm1117) removes two of the three predicted promoters (Figure

8), yet worms carrying the ssu-1(tm1117) allele fail to suppress unc-1 as a homozygote or

51 in trans to ssu-1(fc73). Larger deletions ozDf1 and ozDf2, which remove all of ssu-1 and

several neighboring genes, in trans to ssu-1(fc73) suppress unc-1. ssu-1(tm1117) has no observed phenotype, and is able to produce ssu-1 mRNA. This suggests the ssu-

1(tm1117) deletion does not affect the activity of the SSU-1 protein and that the first promoter is sufficient for ssu-1 expression.

The similarity between SSU-1 and cytosolic sulfotransferases

The SSU-1 protein is similar to the cytosolic alcohol sulfotransferase family (Figure 9).

Alcohol sulfotransferase enzymes transfer a sulfonate from 3’-phosphoadenosine-5’- phosphosulfate (PAPS) onto an alcohol group of a variety of compounds (Strott 2002).

Alcohol sulfotransferases target a different acceptor group than the amino sulfotransferases which include several heparin sulfotransferases in C. elegans. Unlike protein-tyrosine alcohol sulfotransferases, cytosolic sulfotransferases do not target proteins. Sulfonation by cytosolic sulfotransferases is involved in clearance of toxins and metabolizing small molecules including steroid hormones, catecholamines, and drugs

(Strott 2002). The addition of sulfonate can alter the effect of a signaling molecule on its intended target. Other species possess multiple cytosolic sulfotransferases which are classified according to substrate preference. BLAST analysis of the completed C. elegans genome suggests that ssu-1 is the only predicted C. elegans representative of the

cytosolic type of enzyme. SSU-1 averages 31% identity and 51% similarity over 60% of its length to mammalian cytosolic alcohol sulfotransferases (Figure 9) (Protein alignment in Appendix I). The ssu-1 predicted protein sequence comparisons do not favor any of the substrate subdivisions of cytosolic alcohol sulfotransferases. This suggests that

52 mammalian genomic expansion of cytosolic sulfotransferase genes occurred after the

evolutionary split between mammals and nematodes.

Figure 9. SSU-1 is similar to the cytosolic alcohol sulfotransferase enzyme family a) The active sites of the human sulfotransferase SULT2A3 are shown as they relate to the region of cytosolic sulfotransferase similarity in the full length and short spliced forms of SSU-1. The early truncation of ssu-1(fc73) falls outside of the most terminal predicted active site. The short spliced form lacks one of the three substrate binding sites but retains that predicted PAPS binding site. b) There are several types of sulfotransferase in C. elegans. SSU-1 is the only cytosolic sulfotransferase. C. elegans has at least one tyrosylprotein sulfotransferase and several heparin sulfotransferases which all function within the golgi. c) The reaction catalyzed by sulfotransferases transfers the sulfate from phosphoadenosinephosphosulfate onto the alcohol group of the substrate.

SSU-1 in vitro sulfotransferase activity

To confirm sulfotransferase activity by ssu-1, we produced bacterially expressed

proteins for the full-length spliced form, short-length spliced form, and fc73 mutant forms of ssu-1. Thin layer chromatography was employed to visualize the transfer of S35 sulfonate from the phosphoadenosinephosphosulfate donor onto potential substrates. 1- hydroxypyrene is a universal sulfonate acceptor of sulfotransferases (Ma, Shou et al.

2003). Since mammalian sulfotransferases can change substrate specificity with alternative splicing of the first exon (Javitt, Lee et al. 2001), we predicted that the non- rescuing short splice form would be active as a sulfotransferase, but would target substrates irrelevant to the locomotion phenotype. Enzymatic activity of all three bacterially expressed isoforms was demonstrated by the radiolabelling of 1- hydroxypyrene (Figure 10). The full-length spliced form is much more active in these in vitro conditions against this specific substrate. The full-length spliced form also has greater affinity for sterols (Figure 11). The presence of activity of the fc73 mutant form may partially explain why this mutation of the only cytosolic sulfotransferase in C.

elegans is not catastrophic for the animal. The difference in activities could be a property

of substrate preferences of the different isoforms or a result of differences in enzyme

stability in these in vitro conditions. Further biochemical dissection of the three isoforms

offers the opportunity to connect separate ssu-1 substrates with different phenotypes. It is

possible that this biochemical analysis will facilitate the first identification of a C.

elegans steroid hormone.

Figure 10. Bacterially expressed isoforms of SSU-1 demonstrate sulfotransferase activity in vitro a) Coomassie blue staining of purified bacterially expressed SSU-1::GST fusion protein. b) The SSU-1 enzyme catalyzes the transfer of sulfonate from phosphoadenosinephosphosulfate onto hydroxypyrene. c) Organic phase TLC assay of [35S]PAPS reaction with 1-hydroxypyrene. Completed reactions are spotted along the origin of a silica thin layer chromatography plate. The plates are developed in an organic solvent that mobilizes the lipophilic pyrene. The hydrophilic PAPS does not move from the origin. The positions of the radiolabeled pyrene and PAPS are marked on the right of the image. The middle band that is present in all lanes is also present in reactions run with no substrate. The three reactions per SSU-1 isoform have increasing amounts of enzyme going left to right. The decrease in the third full-length isoform reaction is most likely due to glycerol inhibition of the reaction. The activity of the fc73 isoform is very faint.

Figure 11. The bacterially expressed full length isoform of SSU-1 demonstrates in vitro sulfotransferase activity against sterols. a) The chemical structure of the substrates used to assay in vitro activity of SSU-1. Cholesterol and lophenol, which contains an additional methyl group, are both essential for normal worm developement. Ecdysone, lanosterol, and 25-hydroxycholesterol were chosen as substrates to test the substrate affinities of SSU-1. 1-Hydroxypyrene is the control substrate. b) Organic phase TLC assay of [35S]PAPS reaction with 1- hydroxypyrene and a cocktail of the five sterols listed in (a). The positions of the radiolabeled pyrene and PAPS are marked on the right of the image. The middle band that is present in all lanes is also present in reactions run with no substrate (data not shown). Six of the reactions contained 1-hydroxypyrene as the substrate, three reactions contained the cocktail of the five sterols. Duplicate reactions were run for each of the three SSU-1 isoforms with 1-hydroxypyrene. Single reactions were run for each of the three SSU-1 isoforms with the cocktail of five sterols. Again, all three isoform demonstrate activity against 1-hydroxypyrene. The banding pattern in the lane with the reaction containing the cocktail of five sterols and the full length isoform is different from the lanes with sulfonated 1-hydroxypyrene. This result suggests that full length SSU-1 sulfonates a substrate within the cocktail of five sterols.

56 SSU-1 expression in the ASJ amphid neurons

We sought to localize ssu-1 expression to better understand the role of ssu-1. Green

fluorescent protein (GFP) was fused with both spliced forms of ssu-1 and the predicted

ssu-1 promoter (Figure 8). Consistent with the previously described microinjections of the non-GFP constructs, only the full-length ssu-1::GFP fusion rescued the suppression of the locomotion defect of unc-1(null) by ssu-1(fc73). However, both constructs produced

strong fluorescence in only two head neurons. Using the amphid stain DiI, these two

head neurons were identified as the ASJ amphid neurons (Figure 12). The ASJ amphid

neurons are not included in the primary circuit of locomotion which is composed of

motoneurons and the AVA, AVB, PVC, and AVD/E interneurons. Despite the extensive

mapping of the neuronal wiring in C. elegans, no synapses have been observed that connect ASJ amphid neurons with the primary circuit of locomotion (White, Southgate et

al. 1986). Therefore, the ASJ amphid neurons are not likely to affect locomotion through

a direct connection with neurons or muscles. It is most likely that the ssu-1 expression in

the ASJ amphid neurons affects locomotion through systemic signaling to the primary

circuit of locomotion.

Figure 12. SSU-1 is expressed in only two cells, the ASJ amphid neurons a) The GFP fluorescence pattern in animals expressing SSU-1::GFP driven by 500bp promoter. On agar, these transgenic animals display rescue of the Ssu-1 locomotion phenotype. In this plane of focus, two cell bodies (white arrow) reside posterior to additional staining of the processes of these two neurons. The pair of ASJ amphid neurons sends processes to the nerve ring and to the nose. b) DiI staining (red) of the six pairs of amphid neurons in the head. In this single plane of focus, six cell bodies are shown stained with DiI. One of these neurons (yellow arrows) also expresses GFP. The pair of GFP expressing neurons in the SSU-1::GFP animals co-stains with one pair of the DiI staining neurons. The relation of this pair to the other five pairs matches the mapped positions of ASJ amphids.

SSU-1::GFP expression in the ASJ amphid neurons is much more restricted than the

broad neuronal expression of unc-1 (Figure 13). We hypothesized that unc-1 does not act

to affect locomotion from the ASJ amphid neurons. We attempted to rescue unc-1 with a

UNC-1::GFP fusion protein driven by the ssu-1 promoter fragment. The GFP fusion was

employed to confirm the presence of the fusion protein within the ASJ amphid neurons.

58 As expected, the ssu-1 promoter UNC-1::GFP construct failed to rescue any of 450 F1 worms carrying the coinjected marker roller phenotype. However, the ssu-1 promoter- driven expression of the UNC-1::GFP in the ASJ could not be supported by visualization of GFP fluorescence. The control unc-1 promoter driven UNC-1::GFP fusion protein did rescue the Unc-1 phenotype and display expression characteristic of UNC-1 (Figure 13).

The cytoplasmic expression of SSU-1::GFP is much easier to see than the subcellular, punctate pattern of UNC-1::GFP. The absence of observable fluorescence is probably due to the specific localization of UNC-1::GFP to small compartments within the two

ASJ amphid neurons. Although not definitively proven, the experiment suggests that the activity of ssu-1 in ASJ amphid neurons suppresses the defect of unc-1 in remote neurons. This is consistent with the SSU-1 sulfonating a signaling molecule. The separation of the site of action of unc-1 and ssu-1 would be consistent with SSU-1 sulfonating a signaling molecule.

Figure 13. Fusion with UNC-1 changes GFP distribution Confocal images of epifluorescence of unc-1 promoter-driven GFP expression in the head. The arrow marks the nerve rings of each worm. The brackets mark the autofluorescence of the gut. a) The unc-1 promoter expression of GFP in the cytosol of many neurons. b) The unc-1 promoter expression of a UNC-1::GFP fusion. The GFP is no longer seen in the cell bodies, and the brightest fluorescence is in the nerve ring (arrow), the main convergence of nerve processes in the head. The UNC-1::GFP fusion construct rescues the Unc-1 phenotype and produces much less fluorescence per cell than the free GFP expression in (a). The autofluorescence of the gut (brackets) is overexposed in B to capture the reduced emissions of the UNC-1::GFP fusion.

The shared dauer phenotype of ssu-1 and the ASJ amphid neurons

The expression of SSU-1 in the ASJ amphid neurons is remarkable because both ssu-1

and ASJ amphid neurons modify dauer formation. Dauer is a larval stage in C. elegans

that forms in response to adverse environmental conditions. In laser ablation studies in

60 mutant worms with altered dauer formation, ASJ amphid neurons have been associated

with promoting the maintenance of the dauer larvae (Figure 14) (Schackwitz, Inoue et al.

1996). The ASJ amphid neurons have only three physical connections with other cells

(White, Southgate et al. 1986). Despite the limited number of physical connections with

other neurons, the ASJ amphid neurons uniquely function with other amphids involved in

dauer larvae formation (Bargmann and Horvitz 1991; Schackwitz, Inoue et al. 1996).

Figure 14. The physical and functional connections of ASJ amphid neurons with other neurons This schematic shows the connections of the ASJ amphid neurons (green triangle) relative to the rest of the amphid neurons. The ASJ amphid neurons receive input from the AIM amphid neurons and send outputs to the ASK and the PVQ neurons. Also highlighted in this picture are the other amphid neurons that affect dauer formation when laser ablated. The ASK and ADL amphid neurons' (red triangles) effect on dauer larvae formation is dependent on the simultaneous ablation of the ASJ amphid neurons.

61 Ablation of the ADF, ASG, and ASI amphid neurons (blue circles) results in transient dauer larvae formation in good environmental conditions. ASJ ablation perpetuates the dauer state in animals with ablation of ADF, ASG, and ASI. Despite a limited number of physical connections with other neurons, the ASJ amphid neuron uniquely functions with the other amphids involved in dauer larvae formation (picture adapted from White, Southgate et al. 1986).

The role of ASJ amphid neurons in dauer larvae formation and the expression of SSU-1 in ASJ amphid neurons fit well with the association of ssu-1 and dauer larvae formation.

ssu-1 is associated with dauer formation as a suppressor of a neuronal stomatin-like

protein dauer formation phenotype. An increased incidence of dauer larvae to over 95%

of offspring in normal growth conditions at 20˚C was found in a double mutant unc-

24(eDf28) and moc-1(ok366), a chemically mutated strain created by the C. elegans Gene

Knockout Consortium. The phenotype was not linked to the moc-1(ok366) mutation, but

to an additional mutation in the background of moc-1(ok366). The additional mutation

has been given the allele name fc83 and has been mapped to chromosome III. unc-24, but not unc-1, interacts with fc83 to increase dauer formation at 20ºC. ssu-1(fc73) reverses the dauer phenotype of unc-24(eDf28); fc83 so that less than 5% of offspring abnormally form dauer larvae. The suppression of a coordination-modifying stomatin- like protein deficiency by mutant ssu-1 expression in the ASJ amphids suggests endocrine signaling is allowing the two cell populations to interact. The dauer phenotype also supports endocrine classification of the ASJ amphid neurons and the ssu-1 expressed within ASJ neurons. The cytosolic sulfotransferase function of ssu-1 fits well with an endocrine model of action.

62 Genetic characterization of the SSU-1 substrate

With endocrine signaling molecules as the likely substrates for sulfonation, the phenotypes and genetic interactions of ssu-1 give some clues to the substrate preferences of the ssu-1 enzyme. Catecholamine deficient mutants, cat-1(e1111)X, cat-2(e1112)II,

and cat-4(e1141)V, failed to dramatically alter the unc-1, unc-24 and ssu-1 locomotion

phenotypes (data not shown). These crosses argue against catecholamines as being the

substrate of ssu-1 that is relevant to the phenotypes of interest. The majority of identified

mammalian sulfotransferases prefer sterol substrates (Strott 1996). Suggestive of a sterol

target, unc-24 has a sterol binding domain attached to the amino end of the central

stomatin-like domain.

Sterols are also hypothesized to be involved in dauer larvae formation (Sym, Basson et

al. 2000; Gerisch, Weitzel et al. 2001). Current models for the dauer decision-making

pathway have several signaling cascades converge at daf-12, a predicted nuclear hormone receptor (Matyash, Entchev et al. 2004). To further determine the role of ssu-1, we created triple mutants of dauer defective mutations with fc83;unc-24 to determine the position of fc83;unc-24 within the dauer pathway. The dauer defective daf-12(m20),

similar to ssu-1, suppresses the Fc83;Unc-24 dauer phenotype (supplemental data). daf-

12(m20) is a null allele that is indifferent to the predicted steroid hormone which

promotes dauer formation. The epistatic relationship of daf-12 and fc83;unc-24 is

suggestive of a shared signaling sterol. Also supporting sterol involvement in fc83:unc-

24 dauer formation, the Fc83;Unc-24 dauer phenotype has a transient duration of a few

hours to greater than six months before escape to fertile adult. This transient dauer

63 phenotype has only been observed in dauer modifying genes that are predicted to act on sterols: daf-9 (Gerisch, Weitzel et al. 2001) and npc-1;npc-2 (Sym, Basson et al. 2000).

daf-9 is a predicted sterol modifying p450 that is placed directly upstream of daf-12 in

some epistatic models (Gerisch, Weitzel et al. 2001). Since daf-9 promotes dauer

formation, we tested whether unc-24 alone would increase dauer formation or if ssu-

1(fc73) decreased dauer formation. The transient dauer phenotype of daf-9 does not

dramatically increase nor decrease when paired with unc-24 or ssu-1(fc73) respectively.

The lack of interaction suggests that ssu-1 acts upstream or in parallel to daf-9. Finally,

daf-16 is genetically positioned upstream of daf-9. fc83;unc-24 is epistatic to daf-16.

The narrow epistatic placement of fc83;unc-24 downstream of daf-16, and fc73 upstream

of daf-9 suggests a direct role in modifying the steroid hormone ligand of the nuclear

hormone receptor daf-12.

Conclusions

This is the first phenotypic and biochemical characterization of the only cytosolic

sulfotransferase in C. elegans. That the single representation of cytosolic sulfotransferases in C. elegans modifies neuronal phenotypes including sensitivity to volatile anesthetics suggests relevance of ssu-1 to understanding human cytosolic sulfotransferases. The phenotypic differentiation between the two in vitro active spliced forms of ssu-1 indicates the two spliced forms have different substrate affinities. The difference in substrate affinity between spliced forms should assist in the identification of the first steroid hormone in C. elegans. To date, no steroid hormones have been identified despite over 200 predicted nuclear hormone receptors in the genome of C.

64 elegans. The substrate specificity of SSU-1 combined with radiolabeled sulfonate may allow for further phenotypic classification of sterol species in C. elegans. The expression of SSU-1 in the ASJ amphid neurons confirms an endocrine role for these cells. The expression of SSU-1 only in the ASJ amphid neurons supports the model that part of the function of ASJ neurons is to produce sulfonated signals, presumably sterols, to affect other cells. This association of ssu-1 and ASJ cells is the first molecular explanation for an ASJ effect on dauer larvae formation.

The suppression of the volatile anesthetic sensitivity phenotype of unc-1 by the mutation of ssu-1 extends the known interactions between dauer larvae formation and volatile anesthetic sensitivity. N2 dauer larvae are resistant to volatile anesthetics relative to adult N2 worms as measured by immobility (data not shown). Adult N2 worms exposed to conditions that encourage dauer larvae formation have altered sensitivity to volatile anesthetics as measured by behaviors dependent on coordinated locomotion (van

Swinderen, Metz et al. 2002). The SSU-1 sulfonated sterol from the ASJ amphid neuron that affects dauer formation is likely to be the same sulfonated signal that affects sensitivity to volatile anesthetics. Clearly, the dauer formation pathway plays a part in determining sensitivity to volatile anesthetics in C. elegans.

The ASJ-produced, SSU-1-sulfonated signal most likely acts on the primary circuit of locomotion to affect UNC-1 deficiency. Artificial expression of the stomatin-like protein

UNC-24 within the primary circuit of locomotion rescues unc-24(null) (David Miller, personal communication). Since unc-24(null) phenocopies the locomotion defect unc-

65 1(null) and unc-24(null) affects the cellular distribution of UNC-1 protein, we make the

conclusion that UNC-1 is likely to be functioning within the same cells as UNC-24. We

hypothesize that the mutation in ssu-1 alters signaling from the ASJ amphid neurons to

affect the deficiency of UNC-1 in the primary circuit of locomotion (Figure 15).

Figure 15. Model: The SSU-1 sulfonation of a signaling molecule in the ASJ neurons remotely affects the site of UNC-1 deficiency. The cartoons of the worm have two dotted circles in the head to represent the anterior and posterior pharyngeal bulbs. The cartoons include the site of SSU-1 function (green ovals) and the likely site of UNC-1 function (blue circles). a) The loss of UNC-1 likely affects locomotion through the primary circuit of locomotion (blue circles). The primary circuit of locomotion consists of motoneurons and interneurons concentrated along the ventral nerve cord of the worm. b) The mutation of SSU-1 alters the signaling of the ASJ amphid neurons (green ovals) in the head to remotely affect the primary circuit of locomotion and suppress the Unc-1 phenotype.

66 An epithelial sodium channel is a likely shared target of UNC-1 and SSU-1. A genetic

suppressor of a null mutation of unc-1 by its nature selects for indirect interactions with

UNC-1. The suppression of unc-1 by ssu-1 suggests that the normal function of stomatin

is to regulate the action of the substrate of SSU-1. We hypothesize that in the absence of

stomatins, levels of the substrate of SSU-1 need to be altered to restore the normal

function of the shared target of UNC-1 and the substrate of SSU-1. The well-

characterized genetic and protein interactions between unc-1 and the epithelial sodium

channel unc-8 (Rajaram, Spangler et al. 1999; Sedensky, Siefker et al. 2004) strongly

indicate ENaCs as the shared target of unc-1 and ssu-1. We propose that the substrate of

SSU-1 acts through the ENaC UNC-8. There is a strong precedent for steroid hormone

regulation of the UNC-8 homolog ENaCs in humans (Figure 16). Human variations in

ENaCs result in the pseudoaldosteronism hypertension of Liddle’s syndrome (Shimkets,

Warnock et al. 1994) and pseudohypoaldosteronism type I (Chang, Grunder et al. 1996).

Both the constitutive activity of the mutant ENaCs in Liddle’s syndrome and the loss of function mutations in pseudohypoaldosteronism type I render individuals insensitive to aldosterone. Aldosterone is proposed to affect ENaC channels through transcription regulation of G-proteins (Staruschenko, Pochynyuk et al. 2005). We propose that C.

elegans neuronal ENaC-like channels, including UNC-8, are regulated by sulfonated

mineralocorticoids (Figure 17). In this model, loss of UNC-1 function results in ectopic

expression of UNC-8. Supporting this model that UNC-1 normally negatively regulates

UNC-8, unc-8(lf) partially suppresses unc-1(lf). We predict the fc73 mutant form of

SSU-1 to be less efficient at producing a signaling molecule that increases the activity of

the UNC-8 sodium channel. Loss of this activating ligand decreases the ectopic UNC-8

67 expression and restores normal neuronal function. Similar to the aldosterone insensitive

mutations in human ENaCs, we have found ssu-1(fc73) does not modify the phenotypes

of constitutively active nor loss of function alleles of unc-8.

Figure 16. Steroid hormone regulation of the UNC-8 homolog ENaCs in humans The diagram depicts three epithelial cells from the nephron in the human kidney. a) Normal function of these epithelial cells depends on the proper function of the UNC-8 homolog epithelial sodium channels (ENaCs). These ENaCs are normally responsive to the steroid hormone aldosterone, which affects the transcription of G-proteins to regulate channel activity. b) Liddle’s syndrome, also known as pseudoaldosteronism, results from the constitutive activity of a mutated form of the ENaC. With the ENaC always open, individuals with this mutation retain sodium and develop high blood pressure; mimicking the effects of excess aldosterone. c) Pseudohypoaldosteronism type I is the result of loss of function mutations in the ENaCs. As opposed to constitutive channel activity in Liddle's syndrome, the loss of channel activity produces salt wasting and mimics the absence of aldosterone. We have based our model of SSU-1 activity on this interaction between steroid hormones and sodium channels in humans.

Figure 17. Model: The SSU-1 sulfonated signaling molecule regulates the transcription of activators of the UNC-8 sodium channels a) We propose that UNC-1 acts directly to regulate UNC-8 and SSU-1 acts through sulfonation of a signaling molecule to regulate UNC-8. b) Loss of the UNC-1 permits greater availability of UNC-8 to G-proteins and results in increased activity of the UNC-8 sodium channel. c) The mutation of SSU-1 results in the loss of a signaling molecule. This results in decreased transcription of activating G-proteins and decreased activity of the UNC-8 channel.

This work has demonstrated that the only predicted cytosolic sulfotransferase in C. elegans is a functional sulfotransferase which likely modifies endocrine signaling such that neuronal stomatin-associated phenotypes are markedly suppressed. The expression

of SSU-1 in the ASJ neurons refines the understanding of the function of these endocrine

cells. The relationship of unc-1, unc-24 and ssu-1 is the first association of neuronal stomatins sharing regulatory roles with an endocrine signaling cascade.

69 Materials and methods

Strains and General Methods:

Basic genetic techniques and those for culturing C. elegans along with genetic and physical mapping were as described by Brenner (Brenner 1974) and Williams (Williams,

Schrank et al. 1992). Molecular biology techniques were done by standard procedures except as noted. All but two nematode strains were obtained from the Caenorhabditis

Genetics Center in Minneapolis, MN. We used the standard C. elegans EMS mutagenesis protocol (Brenner 1974) to create ssu-1(fc73). ssu-1(tm1117) was kindly

provided to us by Dr. Mitani and the National Bioresource Project for the Experimental

Animal “Nematode C. elegans”. Nematodes were anesthetized as described previously

(Morgan, Sedensky et al. 1990), and EC50’s with standard errors were calculated as described by Waud (Waud 1972). Nematodes were cultured and all anesthetic experiments were performed at 20˚C.

Crosses and Mapping:

Preliminary SNP analysis crossed fc73;unc-1 into the polymorphic strain CB4856. These crosses placed fc73 at the end of chromosome V. For more refined mapping, fc73 was separately paired with neighboring marker genes unc-51(e369) and rol-9(sc148). The doubles with unc-1 were then crossed into CB4856 and recombinants that separated fc73

from the marker genes were collected (22 for unc-51, and 4 for rol-9). Selecting for the

marker genes excluded CB4856 polymorphisms at the position of the marker genes, but selected for CB4856 polymorphisms at the site of fc73. The recombination rates with the known positions of unc-51 and the SNPs predicted that fc73 would lie around the 20080

70 kb on chromosome V. Genes in the proximity of this site were sequenced and a variation

was found in Y113G7A.11, at 20120 kb on chromosome V.

Microinjection:

cDNA for the three isoforms were cloned into pCR2.1 vector (Invitrogen, Carlsbad, CA).

GFP constructs used the pEGFP-1 vector (BD Biosciences, Mountain View, CA).

Construct DNA was prepared with the Plasmid Miniprep kit (Qiagen, Chatsworth, CA).

Mutant rescue was done as described by Mello et al. (Mello, Kramer et al. 1991).

Briefly, this technique involves injecting wild-type DNA into the gonad of mutant animals. The injected DNA is taken up by the developing oocytes and can form free linear arrays. ‘‘Rescued’’ is defined by the reversal of the phenotype of the gene of interest. Rescued progeny of unc-1(null) parents will move normally. Rescued progeny of ssu-1;unc-1(null) parents have uncoordinated locomotion. In general, the test DNA was injected at 10 µg/ml; a marker for successful microinjection, rol-6 DNA [pRF4, a plasmid containing the dominant rol-6(su1006) mutation] was coinjected at 100 µg/ml.

F1 Rollers were picked to establish stable lines.

Microscopy:

Pictures of GFP transgenic animals were taken with a Zeiss Axioplan with confocal microscope (model 600; Bio-Rad). Pictures of a single focal plane of animals were taken with a Zeiss Axiophot microscope equipped for fluorescence, using a Zeiss Axiocam digital camera.

71 DiI:

Amphid cells were stained using the standard C. elegans DiI stain by Michael Koelle,

from Beth Sawin. Briefly, a stock solution of 2mg DiI (Invitrogen, Carlsbad, CA) per 1

ml dimethyl formamide was diluted 1:4000. Several worms were incubated in the diluted

solution at room temperature for 1 hour. Worms were then transferred to a normal

growth medium plate to destain for 1 hour. Stained worms were immobilized on agar

pads with 1mM levamisole for fluorescent microscopy.

Bacterially expressed protein purification, in vitro reactions, and TLC:

cDNA for the three isoforms were cloned into pCR2.1 vector (Invitrogen, Carlsbad, CA).

cDNA was inserted with a fusion at the amino start of the SSU-1 protein into pGEX-4T-2

protein expression vector (Amersham Biosciences, Piscataway, NJ) with BamHI/EcoRI –

using BamHI that was incorporated in frame into the 5’ primer and EcoRI of pCR2.1

vector. Constructs were transformed bl21 pLysS competent cells. To avoid inclusion

bodies, the bacteria were grown at room temperature.

Protein was purified using the Novagen Bugbuster GST·Bind Purification kit (Novagen,

Darmstadt, Germany) per manufacturer’s directions. Eluate was concentrated using a

Millipore Microcon YM-50 Ultrafilter spin tube (Millipore, Billerica, MA). 4ul of the concentrate was run in Figure 3. The sample was mixed to 50% glycerol and frozen at -

20˚C. The GST tag was not removed.

72 50µl reactions contained 100 mM pH 7.0 Potassium Phosphate, 1mM MgCl2, 10mM

DTT, 1mg/ml BSA, .2mM 2-hydroxypropyl-β-cyclodextrin, 1µM [35S]PAPS, and 6ul enzyme glycerol mix. Reactions were run at a nematode physiological temperature of

20˚C for 24 hours to maximize formation of product.

5µl of the reaction mix was mixed with 10µl EtOH saturated with 1-hydroxypyrene as a visible marker. This mix was spotted on a 250µm KG silica gel 60 Å TLC plate and developed for one hour with a solvent of chloroform/methanol/acetone/acetic acid/H2O –

8:2:4:2:1.

Plates were then either sprayed with Perkin Elmer En3hance and placed on film, or lanes

were scraped into Biosafe scintillation fluid and the samples were read on a scintillation

counter.

Acknowledgements

John Au contributed excellent technical assistance to this work.

We would like to thank Zeynep Altun for confirmation of the ASJ cell identification.

This work was funded in part by the Department of Anesthesiology of University

Hospitals of Cleveland and by the Department of Genetics of Case Western Reserve

University.

Bryan Carroll was supported in part by NIH T32 GM07250 and the Case Medical

Scientist Training Program.

Additional funding came from National Institutes of Health (NIH) Grant GM-45402

73 We would like to thank Bernhard Kayser and Rish Pai for their critical reading of this manuscript.

Supplemental data

Dauer crosses

The propensity to form dauer larvae was observed in the following animals. The phenotype was defined as “Dauer” if greater than 95% of the animals form dauer larvae in normal growth conditions at 20˚C. The phenotype was defined as “Not Dauer” if less than 5% of the animals form dauer larvae in the same conditions.

Genotype Phenotype

daf-5(e1386);fc83;unc-24 Dauer

daf-16(m27);fc83;unc-24 Dauer

daf-22(m130);fc83;unc-24 Dauer

fc83;unc-24; daf-12(m20) Not Dauer

fc83;unc-24;fc73 Not Dauer

74 IV ssu-2 : Limited Expression of Suppressor Confines Site of Stomatin Function

Abstract

The mechanism of volatile anesthetics remains unknown. In C. elegans, the gene unc-1 affects anesthetic sensitivity and normal locomotion. The neuronal-expressed

UNC-1 belongs to a large family of membrane-bound chaperones, called stomatin-like proteins. To better understand the function of unc-1 and the mechanism of volatile anesthetics, we created ssu-2, a genetic suppressor of the unc-1 null anesthetic sensitivity and uncoordinated locomotion phenotypes. The SSU-2 protein was found to have homology with the HSP40 chaperone family. The ssu-2 promoter drives GFP expression

in seven specific neurons. The limited expression of the ssu-2 promoter may delineate the site of action of volatile anesthetics as measured by immobility.

75 Introduction

Volatile anesthetics revolutionized surgery nearly 160 years ago. Despite the

continued dependence on volatile anesthetics, the mechanism of action remains a

mystery. Current volatile anesthetics are lethal at twice the doses needed to produce the four requirements of general anesthesia: unconsciousness, immobility, amnesia, and analgesia. A better understanding of the action of volatile anesthetics might enable the

development of safer anesthetics and techniques. The C. elegans neuronal stomatin-like proteins UNC-1 and UNC-24 modify sensitivity to volatile anesthetics as measured by immobility (Sedensky, Siefker et al. 2001). Immobility in C. elegans has been used as a behavior analogous to insensibility to a surgical stimulus in mammals (Morgan, Sedensky et al. 1990; Kayser, Rajaram et al. 1998). unc-1(null) was first identified as a suppressor of the sensitivity to volatile anesthetics generated by unc-79 (Rajaram, Sedensky et al.

1998). Additional alleles of unc-1 were found to have the opposite phenotype of

increasing sensitivity. The ability to both increase and decrease sensitivity suggests a

pivotal role for unc-1 in the mechanism of volatile anesthetics. By studying genes that

affect behavior in volatile anesthetics, we hope to both better understand volatile

anesthetic action and also to understand conserved neuronal mechanisms.

Results

Isolation of the genetic suppressor ssu-2

To study the function of unc-1, we created a genetic suppressor of unc-1. For our genetic suppressor studies, the null unc-1(e580) was chosen to avoid selection of additional alleles of unc-1. We mutagenized unc-1(e580) and selected for genetic

76 suppressors of the Unc-1 uncoordinated locomotion phenotype. Isolated as suppressors

of the locomotion defect of unc-1, the gene class was given the name Suppressor of

Stomatin Uncoordination (ssu). One of the isolates, ssu-2(fc71), is a semidominant

suppressor of the volatile anesthetic sensitivity and locomotion defect of unc-1(e580) and

the locomotion defect of another neuronal stomatin, unc-24(eDf28) (Figure 18). ssu-

2(fc71) is not a universal suppressor of stomatins; it fails to suppress the mechanosensation defect of mec-2(e75).

Figure 18. ssu-2 modifies volatile anesthetic sensitivity

The bars depict EC50 in halothane as measured by immobility for several worm strains. Error bars show the standard deviation from the mean EC50. Wild type C. elegans are immobilized by 3.2% halothane. The sensitivity created by mutations in unc-79 can be suppressed by unc-1. The triple mutant unc-79;unc-1;ssu-2 has increased sensitivity as a result of ssu-2 suppressing the volatile anesthetic resistance of unc-1. Individually, unc- 1(null) and ssu-2(fc71) do not modify anesthetic sensitivity. Modifying volatile

77 anesthetic sensitivity is one of two neuronal stomatin phenotypes that can be suppressed by ssu-2.

Identification of ssu-2

ssu-2 was mapped adjacent to tra-3 on chromosome IV. Sequencing genes in the candidate region identified a splice site mutation in the predicted gene Y67A10A.6. The mutation is an AG→AA change at the splice acceptor site of the seventh exon (Figure

19). The mutation was initially identified within a retained intron in a RT-PCR fragment

amplified from fc71 cDNA. Additional RT-PCR across the mutation shows numerous

products of differing lengths in fc71 cDNA compared to the single fragment seen in N2

cDNA (Figure 20). In the figure, primers amplified products that combine neighboring predicted genes Y67A10A.6 and Y67A10A.7 from both N2 and ssu-2(fc71) cDNA. The significance of this primer design will be discussed later. This data supports a deleterious

effect on the normal processing of ssu-2. The transcript with the retained intron has a

frame shift of the reading frame in the seventh exon that results in an early truncation.

The truncated protein contains 338 amino acids instead of the 447 amino acids of the total Y67A10A.6 predicted gene.

Figure 19. ssu-2 was mapped onto chromosome IV a) The top bar represents the relative position of Y67A10A.6 on chromosome IV. The second bar represents the predicted length of Y67A10A.6 DNA. The blue bars represent exons. The third bar shows the sequence around the ssu-2(fc71) mutation. Brackets designate exon/intron boundaries. Sequencing genes in the candidate region identified splice site mutation in the predicted gene Y67A10A.6. The mutation is an AG→AA change at the splice acceptor site of the seventh exon. The retained intron produces a frame shift and an early truncation. b) A second ssu-2 allele, tm1633, contains a 515bp deletion that removes the third exon. The loss of the third exon results in a frame shift and an early truncation.

Figure 20. RT-PCR amplification of multiple ssu-2(fc71) spliced forms The schematic shows the neighboring genes: Y67A10A.6 and Y67A10A.7. The RT- PCR primers were designed to include the fc71 mutation site and both predicted genes. The picture of the agarose gel shows the products of the RT-PCR visualized with ethidium bromide. The N2 reaction has a single band (arrow). Triplicate reactions with ssu-2(fc71) reveal many bands of various sizes. Sequencing of several of these bands revealed them all to be derived from these two predicted genes. Some of the ssu-2(fc71) fragments have open reading frames that are predicted to produce a protein product that combines the neighboring predicted proteins. The single N2 fragment does not have an open reading frame that is predicted to produce a protein product that combines the neighboring predicted proteins. This RT-PCR supports the combination of the neighboring predicted genes into a single gene. The multiple products from the ssu- 2(fc71) cDNA also support the adverse effect of the splice site mutation in ssu-2(fc71).

Biological confirmation that ssu-2 is encoded by the predicted gene Y67A10A.6 was provided by complementation tests between ssu-2(fc71) and tm1633, a knockout of

Y67A10A.6 that was created by the National Bioresource Project for the Experimental

Animal “Nematode C. elegans” with a screen for genomic deletions. tm1633 in trans to ssu-2(fc71) suppresses the Unc-1 locomotion phenotype. ssu-2(tm1633);unc-1 also

80 suppresses the Unc-1 locomotion phenotype. In addition to suppressing unc-1, both ssu-2

alleles share a coiler phenotype when paired with unc-1. This distinctive motor

dysfunction appears as the dominance of the muscles on ventral side over the dorsal side.

The result is a worm that moves normally between episodes of tight coiling. ssu-

2(tm1633) contains a 515bp deletion that removes the third exon. RT-PCR has not yet

been performed on ssu-2(tm1633) cDNA to delineate the effect of the deletion on the

mRNA transcript. A transcript that joins the second and fourth exons would result in a frame shift and an early truncation at 178 amino acids. Alternative splicing of exons would need to pair the second and seventh exons to return to the original reading frame.

Ideally, the predicted gene can be inserted into a construct that allows biological confirmation with microinjection transgenic experiments. ssu-2 has not yet been rescued by microinjection. Assembly of the wild type construct was impeded by a repeated element of DNA that hindered primer design. The 31bp repeated section of DNA differs by only 2bp and sits 4bp from the end of the predicted coding region. Attempts to amplify the wild type cDNA produced fragments including the first repeated section but not the end of the gene. A full-length representation of the wild type coding region was not created. Genomic fragments including the entire 7.4kb gene did not amplify. The best construct that could be created combines the cDNA including the first 6 exons and a

1kb genomic fragment including the last three exons (Figure 21). This construct is completed, but has yet to be injected into ssu-2(fc71);unc-1.

Figure 21. Rescuing ssu-2 constructs with and without the neighboring predicted gene Y67A10A.7. The relative positions of neighboring predicted genes Y67A10A.6 and Y67A10A.7. Exons are depicted by the blue lines and introns are the black lines between exons. Two rescuing constructs have been created with the genomic and cDNA components. A) The first construct combines the ssu-2 promoter with a .8kb section of cDNA containing the first six exons of Y67A10A.6 and a 1kb section of genomic DNA containing the last three predicted exons of Y67A10A.6. B) The second construct adds Y67A10A.7 to the first construct using 1kb of genomic DNA 5' of the Y67A10A.7 start and 1kb of cDNA containing the entire predicted Y67A10A.7 gene.

The semidominant phenotype of ssu-2(fc71) could mean that the mutation is a dominant negative suppressor of unc-1. If ssu-2(fc71) was a dominant negative suppressor of unc-1, a rescuing construct with the mutant form of the protein should phenocopy the Ssu-2 phenotype of suppressing the Unc-1 locomotion defect. Since the ssu-2(fc71) mutation is before the repeated section of DNA that hindered the full-length wild type construct, an RT-PCR fragment could be created that encoded the full length of the mutant protein. The construct with a 1kb genomic fragment including the predicted

82 promoter and the ssu-2(fc71) cDNA including the retained intron failed to change the

phenotype of injected unc-1 worms.

The similarity between SSU-2 and fls485-like proteins

SSU-2 as defined by the predicted protein of Y67A10A.6 has the greatest amount of similarity with the fls485 protein family (Figure 22). SSU-2 has 49% similarity and 29% identity to the human fls485, the protein for which the family is named (Protein alignment in Appendix I). The similarity extends over 360aa, 80% of the length of SSU-

2. The human fls485 was described in an RNA microarray study of uveal melanoma in humans (Tschentscher, Husing et al. 2003). The human fls485 was found to have decreased expression in tumors. In addition, fls485 is located within a uveal melanoma susceptibility locus at 3p25. The remainder of the fls485-like proteins is uncharacterized.

No C. elegans predicted protein has similarity over the length of the fls485 similarity. A smaller domain within the fls485-like section shares similarity with part of the dnaJ domain of the HSP40 heat shock protein. This 100aa region in HSP40 includes a zinc finger domain that is predicted to bind to the hydrophobic portions of substrates. Many proteins share similarity with this 100aa between 240aa and 340aa. A Psi/Phi Blast search of this region yields more than 50 hits. Since the similarity with this 100aa section is only part of SSU-2 and part of HSP40, it is unclear how much functional similarity is

shared between the proteins. The similarity is too limited to make any strong predictions

of the function of SSU-2 based on homology. In addition to the predicted protein

Y67A10A.6, it is possible that the full length of SSU-2 might include the neighboring

predicted protein Y67A10A.7.

Figure 22. SSU-2 is a member of the fls485 protein family SSU-2 has 49% similarity and 29% identity to the human fls485, the protein for which the family is named. The similarity extends over 360aa, 80% of the length of SSU-2. A smaller domain within the fls485-like section shares similarity with part of the dnaJ domain of the HSP40 heat shock protein. This 100aa region in HSP40 includes a zinc finger domain that is predicted to bind to the hydrophobic portions of substrates. The similarity between these 100aa sections is only part of SSU-2 and part of HSP40, it is unclear how much functional similarity is shared between the proteins.

The full-length ssu-2 gene might include the coding region of the neighboring gene

Y67A10A.7.

The predicted Y67A10A.6 protein may be shorter than the actual length of SSU-2.

An antibody against SSU-2 was created. The first Western blot comparing N2 and ssu-

2(fc71) revealed a 90kd protein in the ssu-2(fc71) lanes (Figure 23). Unfortunately, this western was not reproducible. After the first blot, subsequent isolations of worm protein yielded only non-specific banding patterns. Two pieces of data support the legitimacy of the initial observation. First, in vitro rabbit reticulocyte expressed SSU-2 protein is reactive with the anti-SSU-2 antibody at the predicted size of the protein. Secondly, RT-

PCR with a forward primer in Y67A10A.6 and a reverse primer in neighboring predicted geneY67A10A.7 amplifies products in both N2 and ssu-2(fc71) (Figure 20). Sequencing of several of these RT-PCR bands revealed them all to be derived from these two predicted genes. Some of the ssu-2(fc71) fragments have open reading frames that are predicted to produce a protein product that combines the neighboring predicted proteins.

84 The single N2 fragment does not have an open reading frame that is predicted to produce

a protein product that combines the neighboring predicted proteins. While none of this

data can stand by itself, it is enough to force additional experiments to test the association

of the two neighboring genes. These experiments will be discussed later within future

directions.

Figure 23. Anti-SSU-2 antibody recognizes a 90kd band in fc71 total worm protein but not N2 protein. Total worm protein from two cultures of fc71 and one culture of wild type N2 were separated on a denaturing protein gel. Blots were incubated with either anti-SSU-2 or preimmune sera. A 90kd band (arrow) appears in both mutant protein preparations, but not in the N2 protein. This 90kd band could not be repeated in subsequent harvests.

85 ssu-2 promoter GFP expression in the DVA and RMD neurons.

We sought to localize ssu-2 expression to better understand the role of ssu-2. Ideally,

a construct fusing the SSU-2 protein with green fluorescent protein driven by the ssu-2

would be microinjected into worms. The protein fusion would permit screen for rescue of the ssu-2(fc71). A transgenic worm with a rescued ssu-2 phenotype would confirm the expression pattern as the minimal expression of a gene required for rescue. The lack of a ssu-2 rescuing construct prevents the creation of a SSU-2::GFP fusion transgenic. As the best alternative, green fluorescent protein was fused with 1kb of genomic DNA from 5’ of the ssu-2 start site. This 1kb section contains one predicted promoter. The ssu-2

promoter::GFP constructs were co-injected with the marker construct carrying the

dominant rol-6 gene. F1 worms expressing the marker Rol-6 phenotype were selected.

Clones that maintain the Rol-6 phenotype through several generations typically carry any

additional constructs that were co-injected with the marker rol-6 construct. Twenty stable lines were observed for GFP fluorescence. Twelve of these lines visibly express the GFP. The twelve lines displayed differing extents of the same pattern. The maximal expression includes seven neurons, three pairs in the head and a single neuron in the tail.

The three pairs in the head are RMD interneurons, and the single neuron in the tail is the

DVA neuron (Figure 24). The most consistent expression is in the DVA neuron (Figure

25). All of the 12 lines brightly expressed GFP within the DVA neuron. In some lines, there was only DVA expression with no GFP expression in the head. Never was expression in the head seen in the absence of DVA expression.

Figure 24. ssu-2 promoter::GFP expression in seven neurons The ssu-2 promoter drives expression of GFP in DVA tail neuron and 3 pairs of RMD head neurons. a) Head and tail fluorescence can be seen in a single animal. One of the six RMD neurons can be seen in the head (white arrow). The DVA neuron can be seen in the tail (red arrow). b) GFP fluorescence in the DVA cell body (red arrow). The long neuronal process of the DVA neuron can be seen tracking down the ventral cord on the left side of the worm. c) Three planes of focus through the head show several RMD neurons (white arrows).

Figure 25. The ssu-2 promoter GFP expression fills the full length of the DVA neuron a) The GFP fluorescence pattern in transgenic animals carrying ssu-2 promoter::GFP. Several single planes of focus are compiled to track the process of the DVA neuron from the cell body (white arrow) in the tail to the nerve ring in the head. The dotted circles mark the posterior and anterior pharyngeal bulbs in the head. b) The cartoon shows the process of the DVA neuron extending from the cell body (black arrow) in the tail to the nerve ring in the head (cartoon from WormAtlas).

Both the DVA and the RMD neurons integrate sensory inputs and coordinate the motor response. The RMD neurons coordinate head movements involved in foraging and withdrawal from touch (White, Southgate et al. 1986; Sakata and Shingai 2004) (Figure

26) (Figure 27). Connecting with six of the ten classes of neurons within the primary circuit of locomotion (White, Southgate et al. 1986) (Figure 28), the DVA neuron is well situated to affect locomotion (Figure 29). The DVA neuron modulates the touch withdrawal response (Wicks and Rankin 1995). This phenotype combines the sensory input of mechanosensation with the execution of locomotion. Wicks and Rankin

88 demonstrated that laser ablation of the DVA neuron results in diminished responses to

touch. The DVA is predicted to function just next to the primary circuit of locomotion as

an integrator of multiple inputs.

Figure 26. The multiple neuronal connections of RMD neurons with the motoneurons of the head. The green circle highlights the central position of the RMD neurons within the circuitry of the motoneurons of the head. These connections were mapped using electron microscopic dissection to identify synapses and gap junctions. A behavior has been associated with the interactions between RMD neurons, the OLQ (blue rectangle), and IL1 (blue triangle). Depicted more clearly in the next figure, these neurons coordinate head motions involved in foraging and moving away from touch (picture from White, Southgate et al. 1986).

Figure 27. The RMD neurons coordinate head movements in response to stimuli. The RMD neurons function down stream of the OLQ and IL1 neurons to coordinate head motions involved in foraging and moving away from touch (picture adapted from Riddle, Blumenthal et al. 1997).

Figure 28. The primary circuit of locomotion.

90 Motoneurons (triangles) and interneurons (rectangles) coordinate the contraction and relaxation of muscles for the purpose of locomotion (picture from Riddle, Blumenthal et al. 1997).

Figure 29. The DVA makes many synaptic or gap junction connections with the primary circuit of locomotion. The DVA neuron (hexagon) is well positioned to directly act on the coordination of locomotion. Arrows depict synapses and bars depict gap junctions identified by electron microscopy (White, Southgate et al. 1986) (Picture was adapted from Riddle, Blumenthal et al. 1997).

ssu-2 expression of UNC-1::GFP fusion protein

Since ssu-2 suppresses unc-1 deficiency, it is possible that SSU-2 and UNC-1 affect the locomotion phenotype by being expressed in the same cells. We tested whether

UNC-1 expression within the seven neurons associated with the ssu-2 promoter is sufficient to rescue. This experiment also has the potential to partially validate the significance of the ssu-2 promoter expression in the absence of a rescuing SSU-2::GFP fusion. Transgenic animals were identified by the presence of the rolling phenotype of

91 the co-injected rol-6 marker construct. While none of ten ssu-2 promoter::UNC-1::GFP

transgenic clones rescue the Unc-1 phenotype, one of the clones did express GFP. The

GFP pattern in these worms was a combination of ssu-2 promoter expression limited to

the DVA neurons with the subcellular punctate pattern that is characteristic of UNC-

1::GFP fusions (Figure 30). The positive control of the unc-1 promoter::UNC-1::GFP successfully rescued the Unc-1 phenotype and displayed the classic unc-1 pan-neuronal, punctate pattern (Figure 13).

Figure 30. Expression of UNC-1::GFP in the RMD and DVA neurons fails to rescue the Unc-1 phenotype Two epifluorescence photos of worms expressing UNC-1::GFP that failed to rescue the Unc-1 phenotype. a) The head of ssu-2 promoter::UNC-1::GFP animal has several small dots of GFP fluorescence (arrows). The two focal planes show all of the GFP fluorescence that was observed in the head of this animal. This GFP pattern is in the

92 same region as the RMD neurons. b) The tail of a ssu-2 promoter::UNC-1::GFP animal also has GFP fluorescence (arrow). This pattern is the expected combination of the punctate pattern that results from UNC-1 fusion with GFP and the ssu-2 expression in only seven neurons.

Conclusions

The identification of ssu-2 has expanded the known interactions with UNC-1 and neuronal stomatin-like proteins. The complementation tests of two alleles of Y67A10A.6 biologically confirm the relevance of the mutations. The altered splicing in ssu-2(fc71) confirms the mutation’s adverse affect on splicing. The SSU-2 protein is unlike any protein that has been previously associated with stomatin-like proteins. Further characterization of ssu-2 and its interaction with unc-1 mutations will provide unique insights into the function of neuronal stomatin-like proteins and potentially the mechanism of volatile anesthetics.

The expression pattern of the ssu-2 promoter supports the endpoint of immobility for the study of the mechanism of action of volatile anesthetics. Of the four features of surgical anesthesia, immobility is arguably the feature most likely to be affected by the peripheral nervous system. Consciousness, amnesia, and to a lesser extent analgesia are predominately executed by the central nervous system. There was the risk that in screening for immobility we might identify a mechanism of action of volatile anesthetics affecting only the peripheral nervous system. The expression pattern of the ssu-2

promoter suggests that our screen is targeting a mechanism that could be defined as

central nervous system. Both the DVA and RMD neurons are positioned at an interesting

intersection of sensory inputs and motor outputs. With the simplicity of the worm

nervous system, the interneurons that modulate the motor response to sensory stimuli are

93 the most similar class of C. elegans neurons to the human central nervous system. The

connections and associated functions of the DVA and RMD neurons indicate that they

might be the worm equivalent of a central nervous system. The DVA and RMD neurons

may perform the same cellular function that volatile anesthetics target in the cells

affected in all of the human features of volatile anesthetic response. The roles of the

DVA and RMD neurons are suggestive that our screen for volatile anesthetic sensitivity

with the endpoint of immobility may offer insights into the other three features of

surgical anesthesia: unconsciousness, amnesia, and analgesia.

At the cellular level, my hypothesis is that ssu-2 changes the level of activity of the

DVA neuron which coordinates the UNC-1-dependent DA and VA motoneurons (Figure

31). In the ssu-2 promoter::GFP transgenic worms, the GFP expression is brighter and

more consistently present in the DVA neuron than the RMD neurons. The strength of the

DVA neuron GFP expression and the neuronal connections of the DVA neuron are suggestive that the DVA neuron is the site of SSU-2 action on locomotion. The predicted

chaperone function of SSU-2 is most likely to act in close proximity to the site of UNC-1

action. The expression pattern of the ssu-2 promoter in seven neurons greatly reduces the

candidate neurons for UNC-1 function. UNC-1 is likely expressed in the same neuron as

SSU-2 or in a neuron that is close to SSU-2 expressing cells. The following data argues

against co-expression of UNC-1 and SSU-2. The ssu-2 promoter expression of the UNC-

1::GFP fusion did not rescue the Unc-1 phenotype. We believe that UNC-1 is expressed in the VA motoneurons based on the following data. David Miller has found the

stomatin-like protein UNC-24 to rescue when expressed exclusively in the VA motoneurons (personal communication). unc-24(null) affects the distribution of UNC-1

94 protein (Sedensky, Siefker et al. 2004). Since unc-24(null) phenocopies unc-1(null), we make the conclusion that the Unc-24 phenotype is actually the result of loss the normal distribution of UNC-1. With this assumption, David Miller’s unc-24 data can be directly applied to unc-1. The VA motoneurons direct backwards locomotion. In my model

(Figure 31), loss of UNC-1 results in hyperactivity of the VA motoneurons resulting in the spastic body posturing characteristic of unc-1(null). The second mutation in ssu-2 results in decreased signaling from the DVA neuron to the hyperactive VA motoneurons resulting in decreased spasticity and normal locomotion. As a side effect of disrupted

DVA coordination of motoneurons, the coiler phenotype is seen in ssu-2;unc-1 worms.

Further characterization of the interactions between DVA and VA neurons will elucidate the mechanism of the locomotion deficit and anesthetic sensitivity of unc-1 and ssu-2.

Figure 31. Model: ssu-2 suppresses unc-1 locomotion defect through the DVA neuron to directly affect the coordination of the primary circuit of locomotion. The neuronal connections between the ssu-2 expressing DVA and the primary circuit of locomotion are suggestive of a direct affect by ssu-2 on coordination of locomotion. a) The Unc-1 phenotype is the result of hyperactive VA motoneurons resulting in uncoordinated, ectopic entry into backward locomotion. b) Through the DVA neuron, ssu-2 affects the primary circuit of locomotion to suppress the abnormal signaling (picture adapted from Riddle, Blumenthal et al. 1997).

The protein similarities with Y67A10A.6 and the possible expansion of SSU-2 to include Y67A10A.7 are very interesting. The similarity with the uncharacterized fls485 proteins gives further characterization of the SSU-2 protein the significance of defining a novel protein family. Our further characterization of SSU-2 might be applicable to the

96 mammalian homologues. The broader similarity with the dnaJ domain of HSP-40

chaperones points to a direction for additional study.

Future Directions

Confirmation of the full length of SSU-2 would allow for the writing of a descriptive paper about ssu-2. Three experiments that will address the full length of SSU-2 will be completed in the coming months.

First, a knockout of Y67A10A.7 is being made by the National Bioresource Project for the Experimental Animal “Nematode C. elegans”. Complementation test between the knockout of Y67A10A.7 and the alleles of ssu-2 could support the inclusion of

Y67A10A.7 within the defined length of SSU-2.

Second, RT-PCR across Y67A10A.6 and Y67A10A.7 from N2 and ssu-2(fc71)

cDNA successfully yields products. Investigating the RT-PCR products of the second

ssu-2 allele, tm1633, might rule out the combination of the neighboring predicted

proteins. The amplification of Y67A10A.7 specific products combined with the absence

of products across the two genes would support the neighboring genes functioning

separately. A Northern to compare transcripts would be ideal, but efforts have failed to

produce ssu-2 banding patterns in N2 and ssu-2(fc71) RNA Northerns. The expression of

the ssu-2 promoter in seven neurons is suggestive that the Northerns have failed because

there is too little ssu-2 mRNA to detect.

Finally, microinjection of a rescue construct is the ideal experiment to delineate the

full extent of the SSU-2 protein. With the possibility that ssu-2 may be represented by a

19kb gene, a complicated rescuing construct was built (Figure 21). Briefly, cDNA

97 fragments of Y67A10A.6 and Y67A10A.7 were combined with 2 kb of genomic DNA

that likely holds the information needed to combine the fragments into a single mRNA

transcript. Microinjecting both the previously described Y67A10A.6-specific construct

and this combined construct should rescue the Ssu-2 phenotype of ssu-2(fc71);unc-1.

Successful rescue with one of these constructs will also allow for microinjection of the

SSU-2::GFP fusion. The rescue with SSU-2::GFP fusion will permit validation of the neuronal expression of SSU-2. The subcellular distribution of the SSU-2::GFP fusion will also help elucidate the function of SSU-2. It will be interesting to see if SSU-2::GFP localizes in a punctate pattern within the neuronal processes or if it stays close to the nucleus. The SSU-2::GFP fusion should be very revealing of the nature of the interaction between SSU-2 and UNC-1.

A longer term goal in the study of ssu-2 is to pursue the molecular function of the

SSU-2 protein. The protein similarity with the heat shock protein domain suggests that

SSU-2 functions as a stress-responding chaperone. Temperature, ionic stress, oxidative stress, and nutritional restrictions are easily modified conditions that assay the function stress responding genes in C. elegans. Comparing the ssu-2 with N2 in these various conditions might support the idea that SSU-2 is a heat shock protein.

A second term goal in the study of ssu-2 is to pursue the role of the DVA neuron.

The DVA can be laser ablated (Wicks and Rankin 1995). Ablation of the DVA neuron within our mutants will confirm the significance of the DVA neuron. Since the published account of DVA ablation did not mention an Unc-1-like phenotype, ablation will be most

98 informative in establishing whether DVA neuron activity is increased or decreased in the suppression of unc-1(null) by ssu-2(fc71).

This work has biologically confirmed that Y67A10A.6 is ssu-2. The expression of the ssu-2 promoter is limited to the DVA neuron and the 3 pairs of RMD neurons in the head. The homology of SSU-2 predicts the function of a HSP40-like chaperone.

Materials and methods

Strains and General Methods:

Basic genetic techniques and those for culturing C. elegans along with genetic and physical mapping were as described by Brenner (Brenner 1974) and Williams (Williams,

Schrank et al. 1992). Molecular biology techniques were done by standard procedures except as noted. All but two nematode strains were obtained from the Caenorhabditis

Genetics Center in Minneapolis, MN. We used the standard C. elegans EMS mutagenesis protocol (Brenner 1974) to create ssu-2(fc71). ssu-2(tm1633) was kindly provided to us by Dr. Mitani and the National Bioresource Project for the Experimental

Animal “Nematode C. elegans”. Nematodes were anesthetized as described previously

(Morgan, Sedensky et al. 1990), and EC50’s with standard errors were calculated as described by Waud (Waud 1972). Nematodes were cultured and all anesthetic experiments were performed at 20˚C.

99 Crosses and Mapping:

Preliminary SNP analysis crossed fc71;unc-1 into the polymorphic strain CB4856. These crosses placed fc71 on chromosome IV.

For more refined mapping, fc71 was separately paired with neighboring marker genes dpy-20, dpy-4, and tra-3. The doubles with unc-1 were then crossed into CB4856 and recombinants that separated fc71 from the marker genes were collected (5 for dpy-20, 39

for dpy-4, and 3 for tra-3). Selecting for the marker genes excluded CB4856

polymorphisms at the position of the marker genes, but selected for CB4856

polymorphisms at the site of fc71. The recombination rates with the known positions of

the SNPs predicted that fc71 would lie around the 14420 kb on Chromosome IV. Genes

in the proximity of this site were sequenced and a variation was found in the cDNA of

predicted gene Y67A10A.6.

Microinjection:

Genomic and cDNA PCR/RT-PCR fragment was cloned into pCR2.1 vector (Invitrogen,

Carlsbad, CA). GFP constructs used the pEGFP-1 vector (BD Biosciences, Mountain

View, CA). Construct DNA was prepared with the Plasmid Miniprep kit (Qiagen,

Chatsworth, CA). Mutant rescue was done as described by Mello et al. (Mello, Kramer et al. 1991). Briefly, this technique involves injecting wild-type DNA into the gonad of mutant animals. The injected DNA is taken up by the developing oocytes and can form free linear arrays. ‘‘Rescued’’ is defined by the reversal of the phenotype of the gene of

interest. Rescued progeny of unc-1(null) parents will move normally. Rescued progeny of ssu-2;unc-1(null) parents have uncoordinated locomotion. In general, the test DNA

100 was injected at 10 µg/ml; a marker for successful microinjection, rol-6 DNA [pRF4, a

plasmid containing the dominant rol-6(su1006) mutation] was coinjected at 100 µg/ml.

F1 Rollers were picked to establish stable lines.

Microscopy:

Pictures of GFP transgenic animals were taken with a Zeiss Axioplan with confocal

microscope (model 600; Bio-Rad). Pictures of a single focal plane of animals were taken with a Zeiss Axiophot microscope equipped for fluorescence, using a Zeiss Axiocam

digital camera.

RT-PCR:

Mixed stage RNA was prepared from C. elegans by standard methods (Stuart Kim

Laboratory). Briefly, worms that had recently cleared all of the bacteria from a single

15cm plate were washed once in SB. Pelleted worms were re-suspended in lysis buffer

and freeze thawed 6 times in liquid nitrogen. The lysate was then purified on RNA

purification columns (Invitrogen, Carlsbad, CA) per the manufacturer’s instructions.

RNA was reversed transcribed with random hexamers and AMV reverse transcriptase

(Roche, Indianapolis, IN). cDNA was PCR amplified with standard methods.

Antibodies:

Two rabbits were inoculated four times with the SSU-2 peptide YILESFTEARSTSEAT

(Sigma-Genosys). The final bleed was used for Westerns.

101 Western blots:

Total protein samples were prepared by sonicating 0.5 g worm pellet in ice-cold lysis buffer (150mM NaCl, 3.5mM MgCl2, HEPES 10mM, pH 7.4, 0.3% Triton X-100) with protease inhibitor cocktail (Calbiochem, Set III) followed by centrifugation at 1000xg for

10 minutes at 4˚C. as described by Duerr et al. (Duerr, Frisby et al. 1999). The supernatant was collected, diluted 1:1 in 2x SDS Sample Buffer and boiled 3 minutes.

Protein samples were run on 4-20% gradient SDS-polyacrylamide gels. The concentrations of protein samples were determined using the BCA protein assay kit

(Pierce, Rockford, IL). The resulting concentrations were used to ensure that 140ug of the resulting protein were loaded in each lane. The anti-SSU-2 antibody and preimmune serum were used at 1::50k dilutions.

Acknowledgements

We would like to thank Dr. Zeynep Altun for identification of the DVA and RMD neurons.

This work was funded in part by the Department of Anesthesiology of University

Hospitals of Cleveland and by the Department of Genetics of Case Western Reserve

University.

Bryan Carroll was supported in part by NIH T32 GM07250 and the Case Medical

Scientist Training Program.

Additional funding came from National Institutes of Health (NIH) Grant GM-45402

102 V Summary

A) Highlights of sections I-IV

Genetic suppressors to the locomotion defect of unc-1(null) were generated to identify genes with known functions that might elucidate the function of neuronal stomatin-like proteins. To this end, the characterization of ssu-1 and ssu-2 were successful. The known function of the SSU-1 enzyme and the known function of the

SSU-2 expressing DVA neuron have clarified the function of unc-1.

To appreciate the position of the new information from the characterization of ssu-1 and ssu-2, I will briefly review the highlights of the paper.

Structurally dissimilar volatile anesthetics are bound by only one correlation, Meyer-

Overton rule (Meyer 1899; Overton 1901; Tanifuji, Eger et al. 1977). This rule describes the correlation between potency and lipid solubility of a volatile anesthetic. The Meyer-

Overton rule began a 100 year old debate of what is the relevant target for volatile anesthetics. The Meyer-Overton rule favors lipids as the relevant target, while the opposing view holds proteins as the central targets of volatile anesthetics. The endurance of this debate is a likely indicator that the target of volatile anesthetics is an interface of lipids and proteins.

Genetic studies have created a diverse collection of volatile anesthetic modifiers. In yeast, ubiquitin-associated proteolysis and the stress response to nutritional depletion have been implicated in altering volatile anesthetic response (Wolfe, Reiner et al. 1999;

Palmer, Shoemaker et al. 2005). Low dose volatile anesthetic sensitivity in C. elegans has identified two genes involved in vesicle fusion (Hawasli, Saifee et al. 2004). Higher

103 doses of volatile anesthetics were used to characterize the mutants in this paper. In

addition to the stomatin-like genes and their suppressors, genes with known functions

that modify high doses of volatile anesthetics include gap junction proteins and a

mitochondrial complex I component (Boswell, Morgan et al. 1990; Kayser, Morgan et al.

1999). In Drosophila, voltage sensitive potassium channels, a sodium channel, and two

ABC transporters alter volatile anesthetic sensitivity (Kaplan and Trout 1969; Campbell and Nash 2001; Tanaka and Gamo 2001). In mice, a glycine receptor and stomatin have both been shown to affect volatile anesthetic sensitivity (Quinlan, Ferguson et al. 2002)

(the stomatin association is unpublished data). Lastly, mitochondrial myopathies in humans have been shown to be associated with increased sensitivity to volatile anesthetics (Morgan, Hoppel et al. 2002).

The ability to both increase and decrease sensitivity suggests a pivotal role for unc-1 in

the mechanism of volatile anesthetics. unc-1(null) is suppressed by mutations in unc-8, ssu-1, and ssu-2, suggesting that these genes function in a common pathway.

Similar to the genetic modulators of volatile anesthetics, the collection of associations with stomatin-like proteins is diverse. To review, stomatin-like proteins are associated with two human syndromes: an anemia with neurological variant and a nephrotic syndrome. A chaperone function associated with proteolysis and substrate delivery/availability is conserved in bacteria and yeast. Neuronal stomatin-like proteins have been shown to regulate sodium channels in both worms and mammals. Finally, mammalian stomatin has been associated with dynamic vesicles. Since the function of neuronal stomatin-like proteins is unclear, we reasoned that identification of mutations which suppress loss of stomatin function might identify the role of stomatin-like proteins.

104

ssu-1 was mapped to the predicted gene Y113G7A.11 on the end of chromosome V.

The SSU-1 protein is similar to the cytosolic alcohol sulfotransferase family. Bacterially expressed SSU-1 isoforms demonstrate in vitro sulfotransferase activity. Rescuing SSU-

1::GFP expression is seen in the two ASJ amphid neurons. This pair of neurons in the head is not in the primary circuit of locomotion. UNC-1 is unlikely to affect locomotion from within the ASJ amphid neurons. The VA motoneurons are the likely site of UNC-1 action on locomotion. SSU-1 likely affects the VA motoneurons from the ASJ amphid neurons by sulfating signaling molecules. The ASJ amphid neurons and ssu-1(fc73) share a common association with modifying the formation dauer larvae. Because of the dominant role of endocrine signaling in the dauer larvae formation, this commonality strengthens the argument that SSU-1 signaling can be classified as endocrine.

Our model of interaction between ssu-1 and unc-1 deficiency nominates the ENaC unc-8 as the shared target of unc-1 and ssu-1. The model predicts that ssu-1 can change transcriptional regulation of G-protein expression to induce sodium channels that are normally sheltered from G-proteins by UNC-1.

ssu-2 was mapped to the predicted gene Y67A10A.6 on chromosome IV. RT-PCR of ssu-2(fc71) cDNA confirms the deleterious effect of the ssu-2(fc71) mutation on the normal processing of ssu-2. Complementation tests between two mutations in

Y67A10A.6 provided biological confirmation that ssu-2 is encoded by the predicted gene

Y67A10A.6. SSU-2 shares the greatest amount of similarity with the uncharacterized fls485 protein family. A smaller domain within the fls485-like section shares similarity

105 with the dnaJ domain of the HSP40 heat shock protein. Our data suggests that the full- length SSU-2 protein includes the neighboring predicted protein Y67A10A.7. The full length of SSU-2 is still being pursued. The ssu-2 promoter drives GFP expression in the

DVA neuron and the three pairs of RMD neurons. Both the DVA and the RMD neurons integrate sensory inputs and coordinate the motor response. The UNC-1::GFP fusion protein did not rescue the Unc-1 phenotype despite GFP expression within a ssu-2 promoter expression pattern.

Our model of interaction between ssu-2 and unc-1 deficiency focuses on the known function of the DVA neuron to coordinate locomotion. The model predicts that the ssu-2 altered DVA neuron can suppress the hyperactivity of the unc-1 altered VA motoneuron.

B) Model with the combined findings

1) Given the expression patterns of ssu-1 and ssu-2, the likelihood of autonomous vs. non-autonomous function

Combined, the distinct expression patterns of ssu-1 and ssu-2 strengthen the argument for autonomous expression of UNC-1, SSU-1, and SSU-2 (Figure 32). The promoters of both suppressors provide very informative GFP expression patterns. The site where unc-

1 deficiency affects locomotion cannot be both the expression site of SSU-1 and SSU-2.

The expression sites of SSU-1 and SSU-2 are not functionally related. The ASJ neuron does not have any synapses or gap junctions that would suggest a direct role in coordinating motoneurons. Additionally, the ASJ amphid neurons are not directly or closely connected with the DVA or RMD neurons. Compared to the predicted function

106 of SSU-2, the signaling molecule sulfation of SSU-1 is the most likely to act remotely.

The predicted chaperone function of SSU-2 is more likely to act locally within the DVA

or RMD neurons.

Figure 32. Model: Model: UNC-1 and its suppressors SSU-1 and SSU-2 function autonomously. A depiction of whole worms with the head marked by two dotted circles representing the anterior and posterior pharyngeal bulbs shows the likely sites of action of UNC-1, SSU-1, and SSU-2. UNC-1 (blue) is modeled to affect locomotion from within the VA motoneurons. SSU-1 (green) is modeled to function in the ASJ amphid neurons between the two pharyngeal bulbs. SSU-2 (red) is modeled to function with the DVA tail neuron. a) Loss of UNC-1 protein from the VA motoneurons results in hyperactive management of backwards locomotion. b) Systemic circulation of the SSU-1 sulfonated signaling molecule allows SSU-1 to affect locomotion indirectly from the ASJ amphid neurons. c) A direct connection between the DVA neurons and the primary circuit of locomotion allows ssu-2 to affect locomotion and suppress loss of UNC-1.

107 The interactions of UNC-24, SSU-1, and SSU-2 further support the cell-

nonautonomous function for each protein (Figure 33). The locomotion, volatile

anesthetic sensitivity, dauer and mechanosensation phenotypes of unc-24 are likely the

result of UNC-24 deficiency in at least three different neurons. unc-24 affects

mechanosensation from within the six mechanosensory neurons (Zhang, Arnadottir et al.

2004). unc-24 likely affects locomotion and volatile anesthetic sensitivity from a neuron

within the primary circuit of locomotion. unc-24 is not likely to affect dauer formation

from the same site as it affects locomotion. The differing activities of ssu-1 and ssu-2 support this distribution of unc-24 function. While both ssu-1 and ssu-2 suppress the

Unc-24 locomotion phenotype, only ssu-1 suppresses the unc-24 dauer phenotype. The

lack of ssu-2 suppression of the Unc-24 dauer phenotype supports a model in which SSU-

2 acts locally within the DVA or RMD neurons. Additionally, the lack of ssu-2

suppression of the Unc-24 dauer phenotype supports unc-24 affecting dauer formation

from a cell outside of the primary circuit of locomotion. Suppression of the Unc-24

dauer phenotype by ssu-1 supports the ability of SSU-1 to act on remote cells.

108

Figure 33. Model: UNC-24 also functions autonomously from SSU-1 and SSU-2. A depiction of whole worms with the head marked by two dotted circles representing the anterior and posterior pharyngeal bulbs shows the likely sites of action of UNC-24, SSU- 1, and SSU-2. UNC-24 (blue) is represented with additional blue cells to represent likely sites of function associated with the additional Unc-24 phenotypes. The string of twelve blue dots on the right side of the worm depicts the VA motoneurons. unc-24 is modeled to affect locomotion from within the VA motoneurons. SSU-1 (green) is modeled to function in the ASJ amphid neurons between the two pharyngeal bulbs. SSU-2 (red) is modeled to function with the DVA tail neuron. a) As in the previous figure, the VA motoneurons are predicted to be hyperactive in the absence of neuronal stomatin-like protein UNC-24. b) The systemic circulation of the SSU-1 sulfonated signaling molecule is able to suppress both the locomotion defect of the VA motoneurons and the aberrant dauer larvae formation of the UNC-24 dependent cells in the head. c) The direct connection between the DVA neuron and the primary circuit of locomotion allows ssu-2 to affect the locomotion deficit of unc-24 but the other unc-24 phenotypes are not affected.

The Ssu-2 idiosyncratic coiler phenotype and bacterial piling phenotypes are the final pieces of data supporting a local function for SSU-2 and a remote function for SSU-1.

109 Both of these Ssu-2 phenotypes can be modeled as deficits of DVA or RMD neurons.

The DVA neuron shares the coordination of locomotion with additional neurons that all cross talk. The abnormal activity of the ssu-2 DVA neuron would act locally to imbalance the normal feedback loops within the circuit coordinating locomotion. A subtle imbalance that offset the timing of the ventral/dorsal contractions could be projected to produce a peculiar phenotype similar to the Ssu-2 coiler phenotype. Acting systemically, the ssu-1 signaling would more likely to affect all parts of the circuit equally and be less likely to produce a left/right or ventral/dorsal imbalance. The bacterial piling phenotype was mentioned in the phenotypic description of ssu-2 (section

IIB). The involvement of RMD neurons in coordinating foraging motions of the head might explain this Ssu-2 phenotype. The fact that ssu-2 worms have two phenotypes that can be explained by the two types of neurons with ssu-2 promoter expression of GFP supports a local function of SSU-2 within proximity of the DVA and RMD neurons.

2) Given the signaling function of SSU-1, a model of SSU-2 function that would place SSU-2 as a receiver of the SSU-1 signal

The proposed signaling function of the sulfotransferase ssu-1 may indicate the role of ssu-2 in the suppression of stomatin-like gene unc-1. This hypothesis is driven more by a desire for a common theme than by the data. HSP40 chaperones regulate the ligand binding and nuclear import of steroid hormone receptors (Murphy, Morishima et al.

2003). It is possible that the dnaJ HSP40-like domain of SSU-2 confers the function of regulating the steroid hormone receptor of the molecule sulfonated by SSU-1. In this model, the affects of ssu-1 and ssu-2 would converge at the regulation of the same

110 transcription (Figure 34). Similar to ssu-1, ssu-2 would affect the loss of unc-1 regulation

of the sodium channel unc-8 by altering the transcription of G-proteins.

Figure 34. Model: A HSP40 predicted function of SSU-2 gives SSU-2 a role in modifying SSU-1 signaling. HSP40-like proteins regulate nuclear hormone receptors by affecting availability to binding with ligand and controlling access into the nucleus. It is possible that SSU-1 and SSU-2 both affect the same transcriptional regulation. This model needs for the SSU-1 sulfonated signaling molecule to act on the DVA neurons. A) The loss of UNC-1 results in ectopic activity of the UNC-8 sodium channel. The increased channel activity makes the VA motoneurons hyperactive and backwards motion in initiated inappropriately. B) Loss of the SSU-1 sulfonated signaling molecule results in an inhibitory signal being relayed through neuronal connections onto the VA motoneurons. The altered signaling from the DVA abolishes the aberrant activity of the VA motoneurons. The nuclear hormone receptor, stabilized by SSU-2, remains in the cytoplasm. C) Mutation of SSU- 2 destabilizes the nuclear hormone receptor and alters DVA signaling. The SSU-1 sulfonated signal is blocked by having no nuclear hormone receptor on which to act.

111 C) Relating the findings of ssu-1 and ssu-2 to the understanding of the functions of

stomatins and volatile anesthetics in other species

The remote suppression of neuronal stomatin-like protein UNC-1 deficiency by both

ssu-1 and ssu-2 adds to the complicated picture of stomatin function. Genetic

suppressors of unc-1(null) were predicted to yield indirect associations between the

function of the genetic suppressers and unc-1. While bacterial stomatin-like proteins and

human blood cell stomatin-like proteins have been shown to respond to soluble signals

(Kihara, Akiyama et al. 1998; Salzer, Hinterdorfer et al. 2002), the suppression of unc-

1(null) by ssu-1(fc73) is the first association of soluble signals with neuronal stomatin

function. Additionally, the likely direct cell to cell signaling by ssu-2 to affect unc-1

deficiency is a first for neuronal stomatin-like proteins.

The remote suppression of unc-1 by both ssu-1 and ssu-2 does not exclude any of the

functions associated with stomatin-like proteins in other species. Our model of UNC-1

regulating access of G-proteins to UNC-8 is partially based upon the role of stomatin-like

proteins in bacteria and yeast where the stomatin-like proteins HlfKC and prohibitin regulate access of proteases to substrates. Based on the bacterial and yeast functions of stomatin-like proteins, the identification of additional genetic suppressors of unc-1 would be expected to find a protease that directly interacts with UNC-1. By including UNC-8, our model also incorporates the association of stomatin-like proteins with ENaCs. The association of stomatin-like proteins with dynamic vesicles was neither supported nor excluded for UNC-1 by the characterization of ssu-1 and ssu-2. Since the suppressors suppress unc-1(null), support for the role of dynamic vesicles or cellular distribution will most likely be seen in antibody studies of UNC-8. Finally, it is too early in the

112 characterization of the relationship of unc-1 with ssu-1 and ssu-2 to extrapolate our

findings to the human conditions of stomatocytosis and nephrotic syndrome.

The 160 year old mystery of volatile anesthetics was assisted but not solved by the

characterizations of ssu-1 and ssu-2. Most significantly, the dauer-modifying phenotype

of ssu-1 and the potential to be acting through G-protein signaling offer the strongest link to date between the pathways of volatile anesthetic modifying genes in C. elegans identified by the endpoints of immobility and the endpoints of behavior endpoints affected by smaller concentrations of volatile anesthetics. Both dauer larvae formation and G-protein signaling have been associated with the other endpoints of volatile anesthetic sensitivity (van Swinderen, Metz et al. 2001; van Swinderen, Metz et al. 2002).

The possible overlap of these two collections of volatile anesthetic modifying genes would greatly empower the study of volatile anesthetics using C. elegans.

D) Future directions

1) Exploring the roles of the ASJ, RMD, and DVA neurons in affecting stomatin function and modulating volatile anesthetic sensitivity

Our models for the action of ssu-1 and ssu-2 predict a loss of activity in the ASJ amphid neurons and an increased negative signaling from the RMD and DVA neurons.

The limited expression of ssu-1 and ssu-2 allows for two conceptually simple methods of determining whether the gain or loss of activity from these neurons is involved in the

Ssu-1 and Ssu-2 phenotypes.

The first method is laser ablation, which allows for the determination of the effect of a complete loss of function of the individual neurons. The ablation studies are only

113 practical for studying the locomotion phenotype and not the anesthetic sensitivity

phenotype. The limited number of worms that can be ablated precludes accurate

determination of EC50 because this number is determined by the study of populations.

Attempts were made to initiate these experiments, but I could not find a lab with an

operational laser ablation system. For the study of ssu-1, laser ablation of the ASJ amphid neurons in unc-1(null) and ssu-1(fc73);unc-1(null) animals should reveal whether

ssu-1(fc73) produces a gain or loss of function. From our model, we would expect

ablation of the ASJ amphid neurons to result in suppression of the Unc-1 phenotype in

animals of either genotype. This result would support the model of the loss of an ASJ-

produced, SSU-1-sulfonated signaling molecule affecting the suppression of UNC-1

deficiency in remote cells. There was no mention of uncoordination in the previous

studies in which the ASJ amphid neurons were ablated in N2 animals for the study of

dauer (Bargmann and Horvitz 1991). It is possible, that in studying dauer, the authors did

not notice or record locomotion phenotypes. It is unlikely that a locomotion phenotype

as severe as that of unc-1(null) was not observed if it was present. We expect the

ablation of ASJ amphid neurons in N2 to have no effect on locomotion. The ablation of

ASJ amphid neurons in N2 could be repeated to confirm the lack of effect. For the study

of ssu-2, individual and combined laser ablations of the RMD and DVA neurons in N2,

unc-1(null), and ssu-2(fc71);unc-1(null) animals should characterize the involvement of

these neurons in locomotion deficit of unc-1(null). From our model, we expect increased

inhibition from the DVA neuron as a result of ssu-2(fc71). Therefore, we predict a gain of function in the DVA neuron and laser ablation of the DVA neuron in ssu-2(fc71);unc-

1(null) animals should restore the Unc-1 locomotion phenotype. Again, laser ablation of

114 the RMD and DVA neurons in N2 was not reported to result in severe uncoordination

(Wicks and Rankin 1995; Sakata and Shingai 2004). While the ablation of N2 should be

repeated to confirm previous studies, the ablation is expected to be most informative in

clarifying the role of these neurons as suppressors.

The second method to determine whether there is gain or loss of activity from SSU-1

and SSU-2 expressing neurons is the targeted expression of the tetanus toxin light chain

to specifically eliminate synaptic transmission (Sweeney, Broadie et al. 1995). Using the

ssu-1 and ssu-2 promoters to express the tetanus toxin light chain in the ASJ, RMD, and

DVA neurons would support the role of direct signaling in suppression of unc-1(null).

Unlike laser ablation, these studies would not be limited in the number of altered animals

and all of the relevant phenotypes could be studied. While the potential data is very exciting, there are currently two challenges: experiments expressing tetanus toxin light chain have not yet been performed in C. elegans and I do not have a suggestion for the positive control for confirming expression of the tetanus toxin light chain specifically in the desired cells. If the technical hurdles could be worked out, the targeted expression of tetanus toxin light chain would allow for the testing of our models. For the study of ssu-

1, expression of tetanus toxin light chain in the ASJ amphid neurons is expected to have not effect on the endocrine signaling. Therefore, expression of tetanus toxin light chain in the ASJ amphid neurons would not be expected to affect unc-1(null) nor ssu-

1(fc73);unc-1(null). Any effect on locomotion or volatile anesthetic sensitivity by the

ASJ expression of tetanus toxin light chain would require a new model. For the study of ssu-2, the expression of tetanus toxin light chain in the RMD and DVA neurons is expected to have an effect on the direct communications between the ssu-2(fc71) altered

115 DVA neuron and the primary circuit of locomotion. ssu-2(fc71);unc-1(null) animals

expressing tetanus toxin light chain in the RMD and DVA neurons are expected to

phenocopy Unc-1. Alternatively, if ssu-2(fc71) acts by decreasing DVA signaling, unc-

1(null) animals expressing tetanus toxin light chain in the RMD and DVA neurons would

be expected to suppress the Unc-1 locomotion defect.

2) Mapping of the sites of action for genes involved in modulating volatile anesthetic

sensitivity

Determining the minimal expression for rescue of phenotype for the genes associated

with unc-1 and volatile anesthetic sensitivity would greatly advance the understanding of

this field. This task has not been completed before because screening the previous

number of candidate neurons would have been very laborious. The data presented here

greatly reduce the number of constructs to make and the number of worms to screen for

changes in phenotypes. The laboratory currently has two promoters to work with. One is

the ssu-2 promoter, and the other targets VA motoneurons. The promoters will be tested with rescuing GFP fusions with UNC-1, UNC-8, and UNC-24. If our laboratory can validate David Miller’s results with the VA motoneuron expression of UNC-24, then constructs could be made to test all genes associated with modifying volatile anesthetics in C. elegans as measured by immobility.

The characterization of ssu-1 suggests an association between the C. elegans genes

that modify behavior in low doses of volatile anesthetic and the genes associated with the

concentrations used in our studies. Minimal expression of the genes from these other

studies in the same neurons that are required for the function of the genes in our pathway

116 would support a common mechanism between the two pathways. Demonstrating a link

between the genes affecting immobility in volatile anesthetics and the genes identified by

phenotypes in low concentrations of volatile anesthetic would be an important link within

the field.

3) Further biochemical characterization of SSU-1

The biochemical characterization of SSU-1 needs to be completed. Having

demonstrated that the bacterially expressed SSU-1 isoforms transfer sulfonate in vitro,

complete characterization of substrate specificity of the only cytosolic sulfotransferase in

C. elegans now requires running many candidate substrates through the reaction.

Short of a complete characterization of the enzyme, the activity of SSU-1 towards sterols should be pursued. We have preliminary data that demonstrates SSU-1

sulfotransferase activity against a panel of five sterols. These five sterols should be assayed individually to both confirm SSU-1 activity against sterols and to characterize the substrate preferences of SSU-1. Knowing the sterol preference of SSU-1 would allow predictions of the substrate relevant to the Ssu-1 phenotype.

A better prediction of the substrate relevant to the Ssu-1 phenotype would allow for targeted modification of the sterol composition in the worm culture. Total cholesterol deprivation did not produce any phenotypic changes in any of our mutant worms. It is possible that the small amount of sterol needed for signaling is not depleted before the third reproduction cycle in which the worms growth arrest from cholesterol depletion.

Knowing the sterol preference of SSU-1 would allow for pharmaceutical challenge

117 against the preferred class of sterol. This more subtle approach might detect sterol

involvement in the Ssu-1 and Unc-1 phenotypes.

4) Supporting the model of ssu-1 signaling through G-proteins

Animals with genetic mutations of G-proteins are available to test our model that the

ASJ-expressed, SSU-1-sulfonated signaling molecule regulates the transcription of a G- protein to affect UNC-1 regulated ion flux. Many if not all of the in silica identified C. elegans G-proteins have available animals with null alleles. Promoter expression studies for each of these G-proteins indicate which proteins are expressed in neurons. The five to ten G-proteins with null alleles and neuronal expression could be crossed into unc-1(null) to test our model that loss of the relevant G-protein phenocopies ssu-1(fc73). The first G- protein to test is goa-1 because goa-1 has been shown to modify volatile anesthetic sensitivity in C. elegans as measured by other behavioral endpoints than immobility (van

Swinderen, Metz et al. 2001). An interaction between goa-1 and unc-1 would be the first confirmed link between the genetic pathways identified by the different endpoint of volatile anesthetic response.

5) A saturated screen for suppressors of unc-1 using RNAi

The study of unc-1 needs a saturated screen for suppressors. RNAi is currently the ideal screen for mutants. Unfortunately, the method has not yet been optimized for targeting neuronal genes. Our attempts to affect unc-1, unc-24, ssu-1 and ssu-2 with

RNAi did not produce any change in phenotype. The worm community is working on creating the techniques to allow for better RNAi results against neuronal genes. A

118 saturated screen for suppressors of unc-1 should happen when the techniques are

improved.

Additional alleles of ssu-1 and ssu-2 should be created. The previous attempts at

creating new alleles led to the discovery that ssu-1 and ssu-2 do not suppress the Unc-1 phenotype in male worms. If new alleles could be isolated, the epistatic relationship

between ssu-1 and ssu-2 could be established. If ssu-2 does function on the receiving end of the ssu-1 signal, ssu-2 mutations interfering with the steroid hormone receptor should be epistatic to all alleles of ssu-1. If ssu-2 proves to have a novel function, the additional alleles will assist in the characterization of the protein.

Thank you for reading about this work.

119 Appendix I: Protein sequence alignments

The sequence alignment of the SSU-1 protein with the sulfotransferase domain

The font has been changed to Courier New to allow for the proper character spacing within the alignment.

SSU-1 CD-Length = 412 residues, 58.7% aligned

SULF gnl|CDD|25588 pfam00685 Sulfotransferase domain. CD-Length = 263 residues, 100.0% aligned Score = 195 bits (497), Expect = 8e-51

Red is Identity. Blue is Similarity.

SSU-1: 86 ETDVVIATYPKCGTTWLQHITSQLIKGHDYKAGKGNEL---CVQSPMIER------132 1 PDDVLIVTYPKSGTTWLQELLSLIPNRGDFEKSEEPLLFNARDRSPFLEWYDMFFPDAYE 60 SULF:

SSU-1: 133 ------MGAAFADNIKGPRVLKTHFHHYNIPK---YPDTKYIYCVRNPKDC 174 61 PQLAFEKCRSYFDVELKPVRLAALPSPRLLKTHLPLHLLPKSLLDPNAKIIYLVRNPKDV 120 SULF:

SSU-1: 175 LTSYFHHNRNFKIYNWANGTWDVFLDLFASGQLAFGDYFEHLLSWLPCLKDDNVLFLKYE 234 121 AVSRYHFKRALKDLKAPGTPFEEFFDLFLAGKVVCGSYFDHVLGWLKLRLPGQVLFLRYE 180 SULF:

SSU-1: 235 DMFQDLENAVYKIGQFLGGEAAHRVENPEILREIVDNSTIDAMKKDQKRWFPESQLHKVE 294 181 DLKRDPAGEIKRIAEFLGLP------LEELESILKHLSFENMKGNPCVNYSELS----S 229 SULF:

SSU-1: 295 FIRKGGSRDWKNYFTREQSDRIDSIFAAKFAGTP 328 230 FFRKGLVGDWKNYFTPEQIEKLDEVYREKLADLG 263 SULF:

120 The sequence alignment of the SSU-2 protein with fls485-like proteins.

SSU-2 is aligned with the five best aligned proteins (greater than Score = 90, Expect = 1e-17). To maintain the formatting of the protein alignment, each protein has been given an abbreviated name.

BRIG = LOCUS CAE73996 377 aa linear INV 24-NOV-2003 DEFINITION Hypothetical protein CBG21630 [Caenorhabditis briggsae]. BLAST Score = 725 bits(1871),Expect = 0.0, Identities = 347/377 (92%), Positives = 364/377 (96%), Gaps = 0/377 (0%)

XENO = LOCUS AAH80956 389 aa linear VRT 24-NOV-2004 DEFINITION LOC493204 protein [Xenopus tropicalis]. BLAST Score = 106 bits(264),Expect = 2e-21, Identities = 83/339 (24%), Positives = 137/339 (40%), Gaps = 73/339 (21%)

DANIO = LOCUS XP_684280 406 aa linear VRT 30-JUN-2005 DEFINITION PREDICTED: hypothetical protein XP_679188 [Danio rerio]. BLAST Score = 104 bits(260),Expect = 6e-21, Identities = 91/368 (24%), Positives = 145/368 (39%), Gaps = 72/368 (19%)

RAT = LOCUS XP_575636 440 aa linear ROD 15-APR-2005 DEFINITION PREDICTED: similar to Fls485 protein [Rattus norvegicus]. BLAST Score = 94.0 bits(232),Expect = 1e-17, Identities = 45/155 (29%), Positives = 83/155 (53%), Gaps = 4/155 (2%)

HOMO = LOCUS NP_057015 353 aa linear PRI 02-MAR-2005 DEFINITION fls485 [Homo sapiens]. BLAST Score = 93.2 bits(230),Expect = 2e-17, Identities = 44/153 (28%), Positives = 83/153 (54%), Gaps = 2/153 (1%)

Families of amino acids: Basic side chains K, R, H Acidic side chains D, E Uncharged polar side chains N, Q, S, T, Y Nonpolar side chains A, V, L, I, P, F, M, W, C

121 CLUSTAL W (1.82) multiple sequence alignment

SSU-2--MTDVFNSLLPPDEDLEKHKKFSTSASSKMQYFDDIIPFGEFECREALEKEVKRHRYWK 58 BRIG ------TATSKMHYFDDIIPFGEFECREALEREVKRHRYWK 35 XENO ------SPYSNNPGSSQAAPPYQSVGRDSPTVLLLGAGAQDMMVP------T 40 DANIOMYHEIVTLHCCFLTLKTKVDSSLATMDKTHLLSDQDGASGPGRSYGTCTDPNTEANDGAT 60 RAT ------MDICSHGHHSPGLELLLSIGKHQKTVLPPQGTVLTPVSAEIKCLQLSR 48 HOMO ------MP------SPVG------LLR 9

SSU-2SKTVKKMNFDRIEPS------73 BRIG NKTVRKMNFDRIETS------50 XENO VSPSDMFDPIPGYEG------LTNDNSER------63 DANIOAPPADMMPVVPGYEN------LGPN------79 RAT QSPVVVHQFSVFLRDLKPALGWYMENQDDLVLGLMSDREGFEFWLGTIAGCMADLSFEAE 108 HOMO ALPLPWPQFLACT------LRRLAG------28 .

SSU-2------TSIHYILESFTEARSTSEATEAANFAAMSEAS 105 BRIG ------TSIQYILESFTEARSTSEANEAANFATMSEAS 82 XENO ------FVPPPLSAMIPGPVQPAPVDPEWIIPCITKDA 95 DANIO------VIPPSHFGSSQPQAPPRAPERRFDIPAISEEL 111 RAT SPVMPPDELLEGLPSFDWLLQGRGRQVFFPPLEALGRSQEPTCWSSVLGHSRVPVVTEEV 168 HOMO ------PRESTGPSQKPPPLCSVP--CRVPAMTEEV 56 .. :::

SSU-2CSLTGGSGALSPWDFEVMPGHLFVDQVRVFEMPGSSQINPCSACNSDGTIHCFHCRGYGT 165 BRIG CSLSGGGGALSPWDFEVLPNQLFVDQVRVFEMPGSSQINPCSACNSEGTIHCFHCRGYGT 142 XENO AKEAFVEYAIKDCCYGTLP----AQEMRFREFEPLNVYFYRLETFTESRTIERRTKPF-D 150 DANIOAQEAFTEYVSSKCCYSSKP----VKEMVFTDLQSLNTYRYRLETFTESRTTEWDSEPY-N 166 RAT AREALLSFVNSQCCYSSAA----AGDFIIQELRQQTLCRYRLETFSESRVSEWTFQPV-T 223 HOMO AREALLSFVDSKCCYSSTV----AGDLVIQELKRQTLCRYRLETFSESRISEWTFQPF-T 111 . : . . : . :. . :: . ::. .

SSU-2DKCSFCRGTGMKSGVAHPAVYTHPMIATFPHADLSRGYASSSTAMVRHGGSGSGSNSYGI 225 BRIG DKCSFCRGTGMKSGVAHPAVYTHPMIATFPHADLSRGYASSSTAMVRHGGSGSGGNTYGI 202 XENO GHTVDSRVYGTPPQPWDIPVPYLALFKNEEKKIPIPGTSSLKTCPQCIGVGKIFCTKCTG 210 DANIOGQVVDG--FGVAPEPWSIPVPVPSLFKNCKKSVRVPHTSTVKGCHSCLNLGRSACRKCVN 224 RAT NHSVDGPQRGTSPRLWDMKVQVPPMFQEDTRKFQVPHSSLVKECHKCHGRGRYKCSGCHG 283 HOMO NHSVDGPQRGASPRLWDIKVQGPPMFQEDTRKFQVPHSSLVKECHKCHGRGRYKCSGCHG 171 .: * . * .:: : : . . . .

SSU-2GTPMHFMSKTGVPPPGIGTHDLCYMCHGRGIKECHHCKGGGKKPCTSCSGTGSVRNYTRI 285 BRIG GTPMHFMSKTGVPPPGIGTHDLCYMCHGRGIKECHHCKGGGKKPCTSCSGTGSVRNYTRI 262 XENO TGWVKCGSCLGTGRRQGGDQ--CYSCSIYGTKSCGSCS-KGKLNCDGCSGTGKIVNFIQL 267 DANIOSGRTRCGHCCGTGWISSSERRRCGLCSGSGMVRCHSCGGVGSITCKTCKGHGKLLCFIKL 284 RAT AGMV------TKQCSTCSGRGNKTCATCKGERKLEHFVQL 317 HOMO AGTVRCPSCCGAKRKAKQSRR-CQLCAGSGRRRCSTCSGRGNKTCATCKGEKKLLHFIQL 230 * * *. * *.* .: : ::

122

SSU-2KVYFKNEKSEHYTES--EIPEKLLFQAEGKRIFEEEQDYIIPISQYQEEDVNKMSKLFCA 343 BRIG KVYFKNEKSEHYTEC--EIPEKLLFQAEGKRIFEEEQEYIIPISKYPQEDVNNMSKLFCA 320 XENO AVTWRNNIFEFVADHNSDFPSNRIRKVTGVTLYTDEQDLVSPLVTFPKQSINQASQDGQK 327 DANIOIVTWKNNIHVSVIDKGSGFPLKLLNQITGEKLLTDMAPMVYPVVSFPDHSVNAASESAVR 344 RAT VIVWKNSLFEFVSPHHLHCPGELLTKARGENLFKDENAMVYPIVDFPLKDISLASKRGIE 377 HOMO VIMWKNSLFEFVSEHRLNCPRELLAKAKGENLFKDENSVVYPIVDFPLRDISLASQRGIA 290 : ::*. * : : : * : : : *: : ..:. *:

SSU-2QHLQKCMGVCRVIRQRHYMNAIPISKVHFSLGGEKGIFYVYGTQKLCYFPNFPSKCLADR 403 BRIG QHLQKCMGVCRVIRQRHYMNAIPISKVHFTVGNEKGVFYVYGTQRLCYFPNFPSKCV--- 377 XENO EHHAKYSSSSRILRQRQTIELLPLTKVHYTWEGSPYSYFVYGRENKVYTKNYPEK-CWFC 386 DANIOEHQAQFCTTCRILQQRQTIELIPITQVYYAWKEKTHTFYVYGTEHKVYAKDYPAT-CCCC 403 RAT EHSAMLSSRARILQQRQTIELIPITEVHYWYQGKTSVYYIYGTDHQVYVADYPERYCCGC 437 HOMO EHSAALASRARVLQQRQTIELIPLTEVHYWYQGKTYVYYIYGTDHQVYAVDYPERYCCGC 350 :* .*:::**: :: :*:::*:: . :::** :. * ::*

SSU-2NYAHLQLKKSFCFSINLYSNKYLTAQFSTIKPSHAISRISHQNV 447 BRIG ------XENO NLL------389 DANIOSIL------406 RAT TIL------440 HOMO TIV------353

123 Bibliography

Bargmann, C. I. and H. R. Horvitz (1991). "Control of larval development by chemosensory neurons in Caenorhabditis elegans." Science 251(4998): 1243-6. Barnes, T. M., Y. Jin, et al. (1996). "The Caenorhabditis elegans behavioral gene unc-24 encodes a novel bipartite protein similar to both erythrocyte band 7.2 (stomatin) and nonspecific lipid transfer protein." J Neurochem 67(1): 46-57. Berger, K. H. and M. P. Yaffe (1998). "Prohibitin family members interact genetically with mitochondrial inheritance components in Saccharomyces cerevisiae." Mol Cell Biol 18(7): 4043-52. Boswell, M. V., P. G. Morgan, et al. (1990). "Interaction of GABA and volatile anesthetics in the nematode Caenorhabditis elegans." Faseb J 4(8): 2506-10. Boute, N., O. Gribouval, et al. (2000). "NPHS2, encoding the glomerular protein podocin, is mutated in autosomal recessive steroid-resistant nephrotic syndrome." Nat Genet 24(4): 349-54. Brenner, S. (1974). "The genetics of Caenorhabditis elegans." Genetics 77(1): 71-94. Campbell, D. B. and H. A. Nash (1994). "Use of Drosophila mutants to distinguish among volatile general anesthetics." Proc Natl Acad Sci U S A 91(6): 2135-9. Campbell, J. L. and H. A. Nash (2001). "Volatile general anesthetics reveal a neurobiological role for the white and brown genes of Drosophila melanogaster." J Neurobiol 49(4): 339-49. Chang, S. S., S. Grunder, et al. (1996). "Mutations in subunits of the epithelial sodium channel cause salt wasting with hyperkalaemic acidosis, pseudohypoaldosteronism type 1." Nat Genet 12(3): 248-53. Coates, P. J., D. J. Jamieson, et al. (1997). "The prohibitin family of mitochondrial proteins regulate replicative lifespan." Curr Biol 7(8): 607-10. Correa, A. M. (1998). "Gating kinetics of Shaker K+ channels are differentially modified by general anesthetics." Am J Physiol 275(4 Pt 1): C1009-21. Duerr, J. S., D. L. Frisby, et al. (1999). "The cat-1 gene of Caenorhabditis elegans encodes a vesicular monoamine transporter required for specific monoamine- dependent behaviors." J Neurosci 19(1): 72-84. Franks, N. P. and W. R. Lieb (1981). "Is membrane expansion relevant to anaesthesia?" Nature 292(5820): 248-51. Franks, N. P. and W. R. Lieb (1984). "Do general anaesthetics act by competitive binding to specific receptors?" Nature 310(5978): 599-601. Franks, N. P. and W. R. Lieb (1994). "Molecular and cellular mechanisms of general anaesthesia." Nature 367(6464): 607-14. Fricke, B., H. G. Jarvis, et al. (2004). "Four new cases of stomatin-deficient hereditary stomatocytosis syndrome: association of the stomatin-deficient cryohydrocytosis variant with neurological dysfunction." Br J Haematol 125(6): 796-803. Fricke, B., R. Lints, et al. (2000). "Epithelial Na+ channels and stomatin are expressed in rat trigeminal mechanosensory neurons." Cell Tissue Res 299(3): 327-34. Gerisch, B., C. Weitzel, et al. (2001). "A hormonal signaling pathway influencing C. elegans metabolism, reproductive development, and life span." Dev Cell 1(6): 841-51.

124 Goodman, M. B., G. G. Ernstrom, et al. (2002). "MEC-2 regulates C. elegans DEG/ENaC channels needed for mechanosensation." Nature 415(6875): 1039-42. Harris, B., E. Moody, et al. (1992). "Isoflurane anesthesia is stereoselective." Eur J Pharmacol 217(2-3): 215-6. Hawasli, A. H., O. Saifee, et al. (2004). "Resistance to volatile anesthetics by mutations enhancing excitatory neurotransmitter release in Caenorhabditis elegans." Genetics 168(2): 831-43. Homanics, G. E., C. Ferguson, et al. (1997). "Gene knockout of the alpha6 subunit of the gamma-aminobutyric acid type A receptor: lack of effect on responses to ethanol, pentobarbital, and general anesthetics." Mol Pharmacol 51(4): 588-96. Honemann, C. W., G. W. Nietgen, et al. (1998). "Influence of volatile anesthetics on thromboxane A2 signaling." Anesthesiology 88(2): 440-51. Hosono, R., T. Sassa, et al. (1987). "Mutations affecting acetylcholine levels in the nematode Caenorhabditis elegans." J Neurochem 49(6): 1820-3. Huang, M., G. Gu, et al. (1995). "A stomatin-like protein necessary for mechanosensation in C. elegans." Nature 378(6554): 292-5. Javitt, N. B., Y. C. Lee, et al. (2001). "Cholesterol and hydroxycholesterol sulfotransferases: identification, distinction from dehydroepiandrosterone sulfotransferase, and differential tissue expression." Endocrinology 142(7): 2978-84. Jones, M. V., P. A. Brooks, et al. (1992). "Enhancement of gamma-aminobutyric acid- activated Cl- currents in cultured rat hippocampal neurones by three volatile anaesthetics." J Physiol 449: 279-93. Jurd, R., M. Arras, et al. (2003). "General anesthetic actions in vivo strongly attenuated by a point mutation in the GABA(A) receptor beta3 subunit." Faseb J 17(2): 250-2. Kaplan, W. D. and W. E. Trout, 3rd (1969). "The behavior of four neurological mutants of Drosophila." Genetics 61(2): 399-409. Kayser, B., S. Rajaram, et al. (1998). "Control of anesthetic response in C. elegans." Toxicol Lett 100-101: 339-46. Kayser, E. B., P. G. Morgan, et al. (1999). "GAS-1: a mitochondrial protein controls sensitivity to volatile anesthetics in the nematode Caenorhabditis elegans." Anesthesiology 90(2): 545-54. Keil, R. L., D. Wolfe, et al. (1996). "Molecular genetic analysis of volatile-anesthetic action." Mol Cell Biol 16(7): 3446-53. Kihara, A., Y. Akiyama, et al. (1997). "Host regulation of lysogenic decision in bacteriophage lambda: transmembrane modulation of FtsH (HflB), the cII degrading protease, by HflKC (HflA)." Proc Natl Acad Sci U S A 94(11): 5544-9. Kihara, A., Y. Akiyama, et al. (1998). "Different pathways for protein degradation by the FtsH/HflKC membrane-embedded protease complex: an implication from the interference by a mutant form of a new substrate protein, YccA." J Mol Biol 279(1): 175-88. Krishnan, K. S. and H. A. Nash (1990). "A genetic study of the anesthetic response: mutants of Drosophila melanogaster altered in sensitivity to halothane." Proc Natl Acad Sci U S A 87(21): 8632-6.

125 Lande, W. M., P. V. Thiemann, et al. (1982). "Missing band 7 membrane protein in two patients with high Na, low K erythrocytes." J Clin Invest 70(6): 1273-80. Ma, B., M. Shou, et al. (2003). "Solvent effect on cDNA-expressed human sulfotransferase (SULT) activities in vitro." Drug Metab Dispos 31(11): 1300-5. Mairhofer, M., M. Steiner, et al. (2002). "Stomatin is a major lipid-raft component of platelet alpha granules." Blood 100(3): 897-904. Mannsfeldt, A. G., P. Carroll, et al. (1999). "Stomatin, a MEC-2 like protein, is expressed by mammalian sensory neurons." Mol Cell Neurosci 13(6): 391-404. Marchler-Bauer, A., J. B. Anderson, et al. (2005). "CDD: a Conserved Domain Database for protein classification." Nucleic Acids Res 33(Database issue): D192-6. Matyash, V., E. V. Entchev, et al. (2004). "Sterol-Derived Hormone(s) Controls Entry into Diapause in Caenorhabditis elegans by Consecutive Activation of DAF-12 and DAF-16." PLoS Biol 2(10): e280. McCrae, A. F., E. J. Gallaher, et al. (1993). "Volatile anesthetic requirements differ in mice selectively bred for sensitivity or resistance to diazepam: implications for the site of anesthesia." Anesth Analg 76(6): 1313-7. Mello, C. C., J. M. Kramer, et al. (1991). "Efficient gene transfer in C.elegans: extrachromosomal maintenance and integration of transforming sequences." Embo J 10(12): 3959-70. Melzer, W. and B. Dietze (2001). "Malignant hyperthermia and excitation-contraction coupling." Acta Physiol Scand 171(3): 367-78. Meyer, H. (1899). "Zur Theorie der alkoholnarkose. I. Mitt. Welche eigenschaft der anasthetika bedingt ihre narkotische wirkung?" Arche expe Pathol Pharmak 42: 109-119. Morgan, P. G., C. L. Hoppel, et al. (2002). "Mitochondrial defects and anesthetic sensitivity." Anesthesiology 96(5): 1268-70. Morgan, P. G., M. Sedensky, et al. (1990). "Multiple sites of action of volatile anesthetics in Caenorhabditis elegans." Proc Natl Acad Sci U S A 87(8): 2965-9. Morgan, P. G., M. F. Usiak, et al. (1996). "Genetic differences affecting the potency of stereoisomers of isoflurane." Anesthesiology 85(2): 385-92. Murphy, P. J., Y. Morishima, et al. (2003). "Visualization and mechanism of assembly of a glucocorticoid receptor.Hsp70 complex that is primed for subsequent Hsp90-dependent opening of the steroid binding cleft." J Biol Chem 278(37): 34764-73. Nash, H. A., R. L. Scott, et al. (2002). "An unusual cation channel mediates photic control of locomotion in Drosophila." Curr Biol 12(24): 2152-8. Nelson, T. E. and E. H. Flewellen (1983). "Current concepts. The malignant hyperthermia syndrome." N Engl J Med 309(7): 416-8. Nietgen, G. W., C. W. Honemann, et al. (1998). "Volatile anaesthetics have differential effects on recombinant m1 and m3 muscarinic acetylcholine receptor function." Br J Anaesth 81(4): 569-77. Nietgen, G. W., C. W. Honemann, et al. (1998). "Influence of anesthetics on endogenous and recombinantly expressed G protein-coupled receptors in the Xenopus oocyte." Toxicol Lett 100-101: 319-27.

126 Nijtmans, L. G., S. M. Artal, et al. (2002). "The mitochondrial PHB complex: roles in mitochondrial respiratory complex assembly, ageing and degenerative disease." Cell Mol Life Sci 59(1): 143-55. Nijtmans, L. G., L. de Jong, et al. (2000). "Prohibitins act as a membrane-bound chaperone for the stabilization of mitochondrial proteins." Embo J 19(11): 2444- 51. Nishikawa, K. and Y. Kidokoro (1999). "Halothane presynaptically depresses synaptic transmission in wild-type Drosophila larvae but not in halothane-resistant (har) mutants." Anesthesiology 90(6): 1691-7. Ogawa, H., S. Harada, et al. (1998). "Functional properties of the unc-64 gene encoding a Caenorhabditis elegans syntaxin." J Biol Chem 273(4): 2192-8. Overton, E. (1901). "Me too Studien uber die narkose." Jena, Verlag von Guston Fischer. Palmer, L. K., J. L. Shoemaker, et al. (2005). "Inhibition of Translation Initiation by Volatile Anesthetics Involves Nutrient-sensitive GCN-independent and - dependent Processes in Yeast." Mol Biol Cell. Palmer, L. K., D. Wolfe, et al. (2002). "Volatile anesthetics affect nutrient availability in yeast." Genetics 161(2): 563-74. Park, E. C. and H. R. Horvitz (1986). "Mutations with dominant effects on the behavior and morphology of the nematode Caenorhabditis elegans." Genetics 113(4): 821- 52. Price, M. P., R. J. Thompson, et al. (2004). "Stomatin modulates gating of ASIC channels." J Biol Chem. Quinlan, J. J., C. Ferguson, et al. (2002). "Mice with glycine receptor subunit mutations are both sensitive and resistant to volatile anesthetics." Anesth Analg 95(3): 578- 82, table of contents. Rajaram, S. and H. A. Nash (2004). "A specific alteration in the electroretinogram of Drosophila melanogaster is induced by halothane and other volatile general anesthetics." Anesth Analg 98(6): 1705-11, table of contents. Rajaram, S., M. M. Sedensky, et al. (1998). "Unc-1: a stomatin homologue controls sensitivity to volatile anesthetics in Caenorhabditis elegans." Proc Natl Acad Sci U S A 95(15): 8761-6. Rajaram, S., T. L. Spangler, et al. (1999). "A stomatin and a degenerin interact to control anesthetic sensitivity in Caenorhabditis elegans." Genetics 153(4): 1673- 82. Ratnakumari, L. and H. C. Hemmings, Jr. (1998). "Inhibition of presynaptic sodium channels by halothane." Anesthesiology 88(4): 1043-54. Richmond, J. E. and E. M. Jorgensen (1999). "One GABA and two acetylcholine receptors function at the C. elegans neuromuscular junction." Nat Neurosci 2(9): 791-7. Roselli, S., O. Gribouval, et al. (2002). "Podocin localizes in the kidney to the slit diaphragm area." Am J Pathol 160(1): 131-9. Sakata, K. and R. Shingai (2004). "Neural network model to generate head swing in locomotion of Caenorhabditis elegans." Network 15(3): 199-216.

127 Salzer, U., P. Hinterdorfer, et al. (2002). "Ca(++)-dependent vesicle release from erythrocytes involves stomatin-specific lipid rafts, synexin (annexin VII), and sorcin." Blood 99(7): 2569-77. Schackwitz, W. S., T. Inoue, et al. (1996). "Chemosensory neurons function in parallel to mediate a pheromone response in C. elegans." Neuron 17(4): 719-28. Sedensky, M. M., H. F. Cascorbi, et al. (1994). "Genetic differences affecting the potency of stereoisomers of halothane." Proc Natl Acad Sci U S A 91(21): 10054- 8. Sedensky, M. M., J. M. Siefker, et al. (2004). "A stomatin and a degenerin interact in lipid rafts of the nervous system of Caenorhabditis elegans." Am J Physiol Cell Physiol 287(2): C468-74. Sedensky, M. M., J. M. Siefker, et al. (2001). "Model organisms: new insights into ion channel and transporter function. Stomatin homologues interact in Caenorhabditis elegans." Am J Physiol Cell Physiol 280(5): C1340-8. Shimkets, R. A., D. G. Warnock, et al. (1994). "Liddle's syndrome: heritable human hypertension caused by mutations in the beta subunit of the epithelial sodium channel." Cell 79(3): 407-14. Sirois, J. E., Q. Lei, et al. (2000). "The TASK-1 two-pore domain K+ channel is a molecular substrate for neuronal effects of inhalation anesthetics." J Neurosci 20(17): 6347-54. Staruschenko, A., O. M. Pochynyuk, et al. (2005). "Ras couples phosphoinositide 3-OH kinase to the epithelial Na(+) channel." Biochim Biophys Acta 1669(2): 108-15. Steglich, G., W. Neupert, et al. (1999). "Prohibitins regulate membrane protein degradation by the m-AAA protease in mitochondria." Mol Cell Biol 19(5): 3435-42. Stewart, G. W. and B. Fricke (2003). "The curious genomic path from leaky red cell to nephrotic kidney." Nephron Physiol 93(2): p29-33. Stewart, G. W., B. E. Hepworth-Jones, et al. (1992). "Isolation of cDNA coding for an ubiquitous membrane protein deficient in high Na+, low K+ stomatocytic erythrocytes." Blood 79(6): 1593-601. Strott, C. A. (1996). "Steroid sulfotransferases." Endocr Rev 17(6): 670-97. Strott, C. A. (2002). "Sulfonation and molecular action." Endocr Rev 23(5): 703-32. Sun, L., L. Liu, et al. (2004). "Akt binds prohibitin 2 and relieves its repression of MyoD and muscle differentiation." J Cell Sci 117(Pt 14): 3021-9. Sweeney, S. T., K. Broadie, et al. (1995). "Targeted expression of tetanus toxin light chain in Drosophila specifically eliminates synaptic transmission and causes behavioral defects." Neuron 14(2): 341-51. Sym, M., M. Basson, et al. (2000). "A model for niemann-pick type C disease in the nematode Caenorhabditis elegans." Curr Biol 10(9): 527-30. Tanaka, Y. and S. Gamo (2001). "Sensitivity to diethylether anesthesia of fruit flies primarily depends on the genotypes of the sodium channel gene rather than the states of the membranes and the mechanisms might be different from heat- induced paralysis." Colloids Surf B Biointerfaces 22(1): 39-53. Tanifuji, Y., E. I. Eger, 2nd, et al. (1977). "Some characteristics of an exceptionally potent inhaled anesthetic: thiomethoxyflurane." Anesth Analg 56(3): 387-90.

128 Tinklenberg, J. A., I. S. Segal, et al. (1991). "Analysis of anesthetic action on the potassium channels of the Shaker mutant of Drosophila." Ann N Y Acad Sci 625: 532-9. Tschentscher, F., J. Husing, et al. (2003). "Tumor classification based on gene expression profiling shows that uveal melanomas with and without monosomy 3 represent two distinct entities." Cancer Res 63(10): 2578-84. Turner, E. J., H. G. Jarvis, et al. (2003). "ATP-dependent vesiculation in red cell membranes from different hereditary stomatocytosis variants." Br J Haematol 120(5): 894-902. Ueda, I., M. Hirakawa, et al. (1986). "Do anesthetics fluidize membranes?" Anesthesiology 64(1): 67-71. van Swinderen, B., L. B. Metz, et al. (2002). "A Caenorhabditis elegans pheromone antagonizes volatile anesthetic action through a go-coupled pathway." Genetics 161(1): 109-19. van Swinderen, B., L. B. Metz, et al. (2001). "Goalpha regulates volatile anesthetic action in Caenorhabditis elegans." Genetics 158(2): 643-55. van Swinderen, B., O. Saifee, et al. (1999). "A neomorphic syntaxin mutation blocks volatile-anesthetic action in Caenorhabditis elegans." Proc Natl Acad Sci U S A 96(5): 2479-84. van Swinderen, B., D. R. Shook, et al. (1997). "Quantitative trait loci controlling halothane sensitivity in Caenorhabditis elegans." Proc Natl Acad Sci U S A 94(15): 8232-7. Waud, D. R. (1972). "On biological assays involving quantal responses." J Pharmacol Exp Ther 183(3): 577-607. White, J. G., E. Southgate, et al. (1986). "The Structure of the Nervous System of the Nematode Caenorhabditis elegans." Philosophical Transactions of the Royal Society of London 314(1165): 1-340. Wicks, S. R. and C. H. Rankin (1995). "Integration of mechanosensory stimuli in Caenorhabditis elegans." J Neurosci 15(3 Pt 2): 2434-44. Williams, B. D., B. Schrank, et al. (1992). "A genetic mapping system in Caenorhabditis elegans based on polymorphic sequence-tagged sites." Genetics 131(3): 609-24. Wolfe, D., T. Reiner, et al. (1999). "Ubiquitin metabolism affects cellular response to volatile anesthetics in yeast." Mol Cell Biol 19(12): 8254-62. Yashiroda, H., T. Oguchi, et al. (1996). "Bul1, a new protein that binds to the Rsp5 ubiquitin ligase in Saccharomyces cerevisiae." Mol Cell Biol 16(7): 3255-63. Zhang, S., J. Arnadottir, et al. (2004). "MEC-2 Is Recruited to the Putative Mechanosensory Complex in C. elegans Touch Receptor Neurons through Its Stomatin-like Domain." Curr Biol 14(21): 1888-96. Zhu, Y., C. Paszty, et al. (1999). "Stomatocytosis is absent in "stomatin"-deficient murine red blood cells." Blood 93(7): 2404-10.

129