The bovine papillomavirus type 1 E2 protein is a multifunctional early viral protein with roles in all phases of the cell cycle. E2 is required during G1 as a transcription factor, in S phase to initiate viral replication and during mitosis to tether the viral genome to dividing DNA. The viral genome contains 17 E2 binding sites, the majority of which are concentrated in the long control region (LCR), a regulatory region that is upstream of the viral coding sequence. The role of these binding sites has been explored in vitro using small plasmids and E1 and E2 proteins expressed in bacteria and insect cells. In this study we attempt to examine the placement of E2 on its binding sites during all phases of the cell cycle and in the context of a stably replicating viral system. As part of the examination of the role of E2 during mitosis, we have also examined the role of the cohesin protein Scc1 in viral tethering. Two groups have published disparate reports identifying the cellular protein that binds to the transactivation domain of E2 to stably maintain viral genomes during cell division. Our group has published that it is the DNA helicase ChlR1 that is required for viral tethering, while it has been reported that it is the bromodomain protein Brd4 that is responsible. In this study we contribute to a report that shows that the cellular protein Scc1 binds to the viral genome through a ChlR1 independent mechanism. The cohesin protein binds to BPV-1 E2 at intermittent stages of the cell cycle and may be a factor in viral genome tethering. This interaction may also be important for regulating viral transcription.

Cytokinesis is the cytoplasmic division of one cell into two independent daughter cells. Precise regulation of cytokinesis during cell cycle is essential for healthy and rapid multiplication of any organism. Schizosaccharomyces pombe has emerged as an excellent model system to study eukaryotic cell division regulation. This rod shaped organism grows by bipolar elongation in interphase when its actin cytoskeleton is concentrated at the cell ends (poles). However, growth stops in mitosis and the actin cytoskeleton is rearranged to facilitate assembly of the contractile actomyosin ring at the cell middle. Although several studies have focused on the separate processes of growth and division, it was unclear how cells regulate the cytoskeletal remodeling during the transition between the different stages of the cell cycle. The Septation Initiation Network (SIN) is a signaling cascade essential for fission yeast cytokinesis (Balasubramanian et al., 1998; Mishra et al., 2004) and the MOR (morphogenesis) signaling pathway is essential for interphase bipolar growth (Kanai et al, 2005). Interestingly, inactivation of the SIN not only failed to maintain the cytokinetic apparatus at the cell middle but also caused the redistribution of the cytoskeletal elements like actin to the cell ends that led to bipolar cell elongation similar to cells in interphase (Mishra et al., 2004). These results suggested that SIN signaling inhibits interphase bipolar growth, but it was not clear if the SIN had a direct role in inhibition of interphase growth during mitosis and this question was the major focus of this thesis. The results presented in Chapter II show a novel cross-pathway interaction between the SIN and the MOR in the fission yeast. Our results in Chapter III suggest that some of the MOR pathway components might be important for coordination between nuclear and cytoplasmic divisions in mitosis, revealing novel roles of the pathway. In a separate study (Chapter IV) we sought to identify additional regulators of the SIN and cytokinesis through a suppressor screen and found that the nucleolar rDNA transcription machinery inhibits cytokinesis in fission yeast.

Bile acid metabolism is integrally involved in cholesterol homeostasis in mammals because it is the major means by which cholesterol is eliminated from the body. We have undertaken an effort to study the molecular mechanisms underlying the regulation of bile acid metabolism by isolating and characterizing the cDNA and gene for an enzyme that hydroxylates lithocholic acid (LCA) at position 6β, lithocholic acid 6β-hydroxylase; the first bile acid-induced gene reported. LCA is a very hydrophobic, toxic bile acid formed from chenodeoxycholic acid in the gut lumen upon reduction of the 7α-hydroxy group by microbial enzymes. The proper elimination of LCA is essential for maintenance of the bile acid pool and for prevention of cholestasis which results from LCA precipitating in the cannaculi of the liver when its concentration is high. The LCA 6β-hydroxylase cDNA was isolated by differential hybridization of hamster liver libraries prepared from animals fed either a cholic acid enriched diet or a cholestipol-rich chow and was named CYP3A10 based on its homology with other cytochrome P450s (P450) in family 3A. We found that CYP3A10 was essentially expressed only in males. A statistical analysis of RNA from young males fed with cholic acid and normal chow showed that the cholic acid induction was about 50% at the RNA level. We determined the biological nature of the protein encoded by CYP3A10 by expression of the cDNA in COS cells. Microsomes prepared from transfected cells were assayed with LCA as a substrate and found to hydroxylate LCA predominantly at position 6β. We examined whether CYP3A10 could hydroxylate other steroid compounds by assays with testosterone, progesterone and androstenedione and found that, although 6β-hydroxylase (as well as others) activity was observed with all three, LCA was the preferred substrate based on kinetic analysis. A developmental time course of CYP3A10 expression in males showed little expression before puberty, a striking induction of expression at puberty and a fourfold induction thereafter through adulthood. We then examined the male-specific expression of CYP3A10 in hamster liver. We disrupted the pattern of GH secretion in male hamsters by hypophysectomy, neonatal glutamate treatment and by continuous infusion of GH via osmotic minipumps (to mimic the female pattern of GH secretion) and found no significant effect on CYP3A10 expression. Conversely, in females, hypophysectomy and neonatal glutamate treatment significantly induced CYP3A10 expression 5- to 10-fold. Additionally, when females treated neonatally with glutamate were injected twice daily with GH as adults (to mimic the male pattern of GH secretion), the levels of CYP3A10 expression were not significantly different from those of normal males. These results led us to conclude that the pattern of GH secretion in males does not control the male-specific expression of CYP3A10 but that in females expression can be induced by altering the tonic secretion of GH. No significant effect on CYP3A10 expression was observed by castration of adult males, indicating that circulating androgens were not required for expression. We found that gonadal hormones (e.g. estrogen and progesterone) do not have a suppressive effect on CYP3A10 expression in females since ovariectomy did not induce expression. Many genes are "imprinted" neonatally by exposure to a given effector for developmental-, tissue- or sexually regulated expression. We investigated whether neonatal androgen exposure was required for male-specific expression of CYP3A10 by castrating hamsters neonatally and determining the level of CYP3A10 expression in adulthood. Our results indicate that androgens are required neonatally for CYP3A10 expression since no expression was observed in neonatally castrated hamsters. We were unable to induce expression in neonatally castrated hamsters by either GH or testosterone injections. These results suggest several notable points 1) that CYP3A10 expression is programmed neonatally by androgen exposure; 2) that androgens exert their effect directly on the liver and not via the hypothalamus; 3) that neither testosterone nor GH can restore CYP3A10 expression when males have not been exposed to androgens neonatally; and 4) that in experimental conditions, females can be induced to express CYP3A10, which indicates that there are two modes for regulating expression: by "imprinting" in males and by GH and testosterone in females. We are now studying the molecular mechanisms involved in the bile acid-mediated induction and the male-specific expression of CYP3A10. We have cloned approximately 8 kb of 5' flanking DNA from a hamster genomic library and sequenced about 1 kb of proximal DNA. Primer extension and S1 digestion analyses indicate that the mRNA for CYP3A10 has multiple transcription initiation sites clustered about 90 bp from the initiator methionine codon. We have also prepared CYP3A10 promoter/lacZ chimeric constructs to begin delineating the cis-acting elements controlling CYP3A10 expression and regulation. We used H2.35 cells as recipients because they are a mouse hepatocyte cell line that has been transformed with a temperature sensitive SV40. These cells can be grown at the permissive temperature and can be induced to behave like liver cells, the differentiated condition, by switching to a nonpermissive temperature. We have found that the construct with 1 kb of proximal CYP3A10 5' flanking DNA was able to express the reporter gene at higher levels under differentiated conditions, which were consistent with higher expression of an albumin promoter/lacZconstruct, upon switching the cells to the more liver phenotype. The system characterized and described here is ideally suited for dissecting the molecular details governing bile acid-mediated regulation and sexually dimorphic expression of liver genes. Very little is known about both these very important biological phenomena. Much could be learned about transcriptional regulation of liver genes by investigating the cis-elements and trans-acting factors mediating regulation of CYP3A10 expression.

During embryonic development of the nematode Caenorhabditis elegans, cell fates are specified by asymmetric segregation of cell fate determinants and via cell-cell signaling events. Specification of the eight-cell stage blastomere E, the endoderm progenitor cell, requires both cell signaling and asymmetric cell division. At the four-cell stage, a polarity-inducing signal from the P2 cell is required for the EMS cell to divide asymmetrically to produce an anterior daughter MS, and posterior daughter E. In the absence of signal, the EMS cell divides symmetrically to produce two daughters that adopt the MS fate. This thesis describes the identification and analyses of seven genes required to tranduce this polarity-inducing signal and specify endoderm formation. The mom-1, mom-2, mom-5, apr-1, and wrm-1 genes are homologous to components of the Wnt/Wingless signal transduction pathway, and the mom-4, and lit-1 genes are related to components of the mitogen-activated protein kinase pathway. Biochemical analysis of these signaling molecules reveal a novel convergence of these pathways at the level of the LIT-1 and WRM-1 proteins, which appear to function as a kinase complex and are required for the downregulation of POP-1. Together these genes constitute components of a complex genetic pathway required for specification of the E cell fate.

The proper maintenance of the pathways governing cell growth is critical to ensure cell survival and DNA fidelity. Much of our understanding of how the cell cycle is regulated comes from studies examining the relationship between DNA viruses and the mechanisms of cell proliferation control. There are numerous examples demonstrating that viruses can alter the host cell environment to their advantage. In particular, the small DNA tumor viruses, which include adenovirus, simian-virus 40 (SV-40), and human papillomavirus (HPV), can modulate the host cell cycle to facilitate viral DNA replication. Due to the fact that these viruses infect quiescent, non-cycling cells and lack the necessary enzymes and resources to replicate their DNA (e.g. DNA polymerase), the small DNA tumor viruses must activate the host cell replication machinery in order to expedite viral DNA replication. The capacity of these viruses to perturb normal cell proliferation control is dependent upon their oncogene products, which target p53 and members of the Retinoblastoma (RB) family of proteins and inactivate their respective functions. By targeting these key cell cycle regulatory proteins, the small DNA tumor viruses induce the infected host cells to enter S-phase and activate the components involved with host cell DNA synthesis thereby generating an environment that is conducive to viral DNA replication. In contrast, the larger, nuclear-replicating DNA viruses such as those from the family Herpesviridae, do not share the same stringent requirement as the small DNA viruses to induce the infected host cell to enter S-phase. The herpesviruses encode many of the components to stimulate nucleotide biosynthesis and the necessary factors to facilitate virus DNA replication including a viral DNA polymerase and other accessory factors. Additionally, many herpesviruses encode gene products that arrest the host cell cycle, in most instances, prior to the G1/S transition point. Inducing cells to growth arrest appears to be a prerequisite for the replication of most herpesviruses. However, in addition to encoding factors that inhibit the cell cycle, many herpesviruses encode proteins that can promote cell cycle progression in a manner similar to the small DNA tumor virus oncoproteins. By targeting members of the RB family and p53 protein, the herpesvirus proteins induce S-phase and activate S-phase associated factors that playa role in DNA replication. In this manner, the herpesviruses may promote an environment that is favorable for DNA replication. Consistent with the other herpesviruses, human cytomegalovirus (HCMV)induces human fibroblasts to growth arrest. However, in other cell types, virus infection causes cells to enter S-phase. In addition, HCMV replication requires several cellular factors that are present only during S-phase. Furthermore, HCMV induces the activation of S-phase-associated events as well as the increased expression of numerous S-phase genes following infection. HCMV encodes two immediate early (IE) gene products, IE1-72 and IE2-86, which can interact with members of the RB family of proteins. Additionally, the IE2-86 protein can bind to and inhibit p53 protein function. Given the functional resemblance between the HCMV IE proteins and the oncoproteins of the small DNA tumor viruses, we hypothesized that expression of the HCMV IE proteins could modulate cell cycle control. Specifically, we determined that expression of either IE1-72 or IE2-86 can induce quiescent cells to enter S-phase and delay cell cycle exit following serum withdrawal. Moreover, IE2-86 mediates this effect in the presence or absence of p53, whereas IE1-72 fails to do so in p53-expressing cells. Furthermore, both IE1-72 and IE2-86 induce p53 protein accumulation that is nuclear localized. Because IE1-72 fails to promote S-phase entry in cells expressing p53 and induces p53 protein levels, the mechanism by which IE1-72 alters p53 levels was examined. IE1-72 elevates p53 protein levels by inducing both p19ARF protein and an ATM-dependent phosphorylation of p53 at Ser15. IE1-72 also promotes p53 nuclear accumulation by abrogating p53 nuclear shuttling. As consequence of this IE1-72-mediated increase in p53 levels, p21 protein is induced leading to a p21-dependent growth arrest in cells expressing IE1-72. These findings demonstrate that the HCMV IE proteins can alter cell proliferation control and provide further support to the notion that HCMV, through the expression of its IE proteins, induces S-phase and factors associated with S-phase while blocking cell DNA synthesis, to possibly generate an environment that is suitable for viral DNA replication.

This dissertation describes the purification and partial characterization of CBP from the epidermis of the monarch butterfly larvae (Danaus plexippus). A yellow protein-carotenoid complex was extracted from the yellow pigmented epidermal tissue from monarch butterfly larvae by homogenization. Additional steps in the purification process included differential precipitation with ammonium sulfate, cation and anion chromatography, and lastly size exclusion chromatography. Polyacrylamide gel electrophoresis demonstrates that a single protein was isolated (M-LBP) having a ~60 kDa molecular weight, the value has subsequently been confirmed by HR-tandem MS. Lutein is the sole carotenoid bound by M-LBP with a stoichiometry of the binding of 2: 1. Immunohistochemistry results show that M-LBP has no cross-reactivity to antibodies for silk worm CBP (Bombix mori) but does have cross-reactivity with antibodies for horn worm epidermal CBP (Agrius convolvuli). Binding affinities were determined using surface plasmon resonance for the carotenoids lutein (KD = 18.6 ± 0.7), R,R-zeaxanthin (KD = 990 ± 60), R,S-zeaxanthin (KD = 60 ± 2). Tryptophyphan fluorescence lifetimes were determined for the apoprotein and compared to those of the native M-LBP. Tryptophan fluorescence lifetimes were found to be 3.9 ns and 3.0 ns, respectively for these two forms of the protein, indicating that upon dissociation of the carotenoid from the protein the tryptophan fluorophore adopts a position where it has less interaction with the polar surface environment.

One of the main components underlying drug addiction is the emergence of tolerance. Although its development is a complex issue, and is believed to have both psychological and physiological connotations, it is clear that some physiological change must occur that would enable an organism to withstand drug concentrations lethal to a naïve system. The purpose of this thesis was to identify and study a physiological mechanism, whose characteristics were altered due to chronic exposure to ethanol. Vasopressin (AVP), whose primary function is to control water balance, release from the neurohypophysis is suppressed by an acute ethanol challenge. Therefore, I hypothesized; 1) that chronic ethanol exposure would reduce the normal suppression of AVP release during an acute ethanol challenge and 2) that the ion channels that are acutely sensitive to ethanol, involved in the control of AVP release, would exhibit a change in their ethanol sensitivity and characteristics. To study the hypothesis, I utilized the neurohypophysis from rats chronically exposed to ethanol and yoked controls to determine whether chronic exposure would modify the acute ethanol sensitivity of the neurohypophysial vasopressin release mechanism. I examined whether the long-term ethanol exposure affected the suppression of vasopressin release from either or both the intact neurohypophysis and the isolated neurohypophysial terminals. In addition, I investigated how chronic exposure affected two types of potassium channels, the ethanol sensitive large conductance Ca+2-activated (BK) channel and the fast inactivating (IA) channel known to be insensitive to physiologically relevant concentrations of ethanol. I was able to establish that chronic ethanol exposure reduced the suppression of vasopressin release by an acute ethanol challenge from both the intact neurohypophysis and the isolated neurohypophysial terminals. In addition, I discovered that oxytocin release was affected similarly. I concluded from this data that chronic exposure to ethanol affected a general mechanism, which controlled hormone release from the neurohypophysis, and that this mechanism could be isolated to the neurohypophysial terminals. I also used electrophysiological techniques to study ion channel characteristics of both the BK and IA potassium channels. I found that in naïve rats, BK channels were potentiated and IA channels insensitive to physiological relevant concentrations of ethanol. But in chronic ethanol-exposed rats the BK channels exhibited a reduced sensitivity to ethanol while IA channels were inhibited. In addition, the current density of the BK channel was significantly reduced. These results show that at least one characteristic of each potassium channel has been modified. This suggests that chronic exposure can not only modify the ethanol sensitivity of ion channels known to be ethanol-sensitive, but also those believed to be relatively insensitive. Therefore, since modifications in these channels have previously been shown to alter the duration and frequency of action potentials, I conclude that these ethanol-induced modifications play a role in the modified hormone release patterns observed in the chronically exposed rats.

Chromatin remodeling is crucial to transcriptional regulation in vivo and a number of protein complexes capable of altering genomic architecture in the budding yeast Saccaromyces cerevisiaehave been identified. Among these, the SWI/SNF complex, a 2 MDa, eleven subunit protein assembly, has been the most extensively characterized. The SWI/SNF complex is required for the proper expression of a number of genes in yeast, although it is completely dispensable for the expression of others. Likewise, some, but not all, transcriptional activator proteins require SWI/SNF activity in order to function in vivo. The goal of this thesis work was to identify those components of the transcription process which dictate this dependence on SWI/SNF activity. Using the well characterized UASGALsystem, we have determined that one of these components is the nucleosome state of activator binding sites within a promoter. We find that while SWI/SNF activity is not required for the GAL4 activator to bind to and activate transcription from nucleosome-free binding sites, the complex is required for GAL4 to bind and function at low affinity, nucleosomal binding sites in vivo. The SWI/SNF -dependence of these nucleosomal binding sites can be overcome by 1) replacing the low affinity sites with higher affinity, consensus GAL4 binding sequences, or 2) placing the low affinity sites into a nucleosme-free region. These results provide the first in vivo evidence that the SWI/SNF complex can regulate gene expression by modulating the DNA binding of a transcriptional activator protein. To determine whether specific components of the GAL4 protein are necessary in order for the SWI/SNF complex to modulate binding to nucleosomal sites in our model system, we tested the SWI/SNF-dependent DNA binding of various derivative GAL4 proteins. We find that a functional activation domain is not required for SWI/SNF to modulate GAL4 binding in vivo. Interestingly, like the full length protein, GAL4 derivatives in which the activation domain has been mutated are able to partially occupy nucleosomal sites in the absence of SWI/SNF (binding in the absence of SWI/SNF is at least forty percent lower than in the presence of SWI/SNF), indicating the activation domain is also not required for SWI/SNF-independent DNA binding. These results support a model in which the SWI/SNF-dependence of a gene reflects the nucleosomal context of its important regulatory sequences, e.g. binding sites for transcriptional regulatory proteins. Although nucleosomal promoter regions have been correlated with SWI/SNF-dependence in the past, there has of yet been no gene at which nucleosome location has correlated with a specific genetic function. In the final part of this thesis work, we initiated a search for an endogenous SWI/SNF-dependent gene for which the nucleosome state of activator binding sites could be determined.

Transcriptional regulation of cell cycle controlled genes is fundamental to cell division in eukaryotes and a broad spectrum of physiological processes directly related to cell proliferation. Expression of the cell cycle dependent human H4, H3 and H1 histone genes is coordinately regulated at both the transcriptional and posttranscriptional levels. We have systematically analyzed the protein/DNA interactions of the immediate 5'regions of three prototypical cell cycle controlled histone genes, designated H4-F0108, H3-ST519 and H1-FNC16, to define components of the cellular mechanisms mediating transcriptional regulation. Multiple biochemically distinct protein/DNA interactions were characterized for each of these genes, and the binding sites of several promoter-specific nuclear DNA binding activities were delineated at single nucleotide resolution using a variety of techniques. These findings were integrated with results obtained by others and revealed that the in vitro factor binding sites in H4, H3 and H1 histone promoters coincide with genomic protein/DNA interaction sites defined in vivofor the H4-F0108 and H3-STS19 genes, and with evolutionarily conserved cis-acting sequences shown to affect the efficiency of histone gene transcription. Specifically, we have defined binding sites for Sp1, ATF, CP1/NF-Y, HiNF-D, HiNF-M, HiNF-P and HMG-I related factors. Based on sequence-similarities and cross-competition experiments, we postulate that most of these protein/DNA interaction elements are associated with more than one class of histone genes. Thus, the protein/DNA interactions characterized in this study may represent components of a cellular mechanism that couples transcription rates of the various histone gene classes. Regulation of the protein/DNA interactions involved in transcriptional control of these H4, H3 and H1 histone genes was investigated in a spectrum of cell types using several distinct in vitro cell culture models for the onset of differentiation and quiescence, as well as cell cycle progression. Moreover, we studied control of histone gene associated DNA binding activities during hepatic development from fetus to adult in transgenic mice reflecting the onset of differentiation and quiescence in vivo. We show that the H4 histone promoter protein/DNA interaction mediated by factor HiNF-D is selectivelymodulated, and directly at the level of DNA binding activity, during the entry into, progress through and exit from the cell cycle in normal diploid cells, as well as during hepatic development. The regulation of this protein/DNA interaction occurs in parallel with analogous interactions occurring in H3 and H1 histone genes. Moreover, these proliferation-specific protein/DNA interactions are collectively deregulated during the cell cycle in four distinct cell types displaying properties of the transformed phenotype. Hence, the cellular competency to coordinately transcribe distinct classes of histone genes during the cell cycle may be mediated by the intricate interplay of constitutively expressed general transcription factors and temporally regulated, cell growth controlled nuclear factors interacting specifically with cell cycle dependent histone genes. Finally, we show that HiNF-D is represented by two electrophoretically distinct species. The ratio of these forms of HiNF-D fluctuates dramatically during the cell cycle of normal diploid cells, but remains relatively constant in tumor cells. Total HiNF-D binding activity embodied by both HiNF-D species is negatively influenced in vitro by incubation with exogenous phosphatase activity. These observations provide a first indication for the hypothesis that HiNF-D may exist in distinct post-translationally modified forms that are subject to a stringent cell growth control mechanism involving protein kinases and phosphatases. Such a cellular post-translational modification mechanism, which directly impinges on (or activates) the DNA binding activity of a key factor controlling histone genes, would provide a highly efficient means by which to influence the rate of transcription in rapid response to intra-cellular requirements for histone mRNA and extra-cellular cues signalling the onset and cessation of cell proliferation.

Cotranslational translocation is initiated by targeting of a ribosome-bound nascent polypeptide chain (RNC) to the endoplasmic reticulum (ER) membrane. The targeting reaction is coordinated by the signal recognition particle (SRP) through its interaction with the RNC and the membrane-bound SRP receptor (SR). A vacant translocon is a prerequisite for the subsequent nascent chain release from SRP-SR-RNC complex. It has been proposed that the protease-accessible cytosolic domains of the Sec61p complex play an important role in posttargeting steps by providing the binding site for the ribosome or interacting with the SR to initiate the signal sequence releasing. In this study, we have investigated the detailed mechanism that allows transfer of the ribosome-nascent chain (RNC) from the SRP-SR complex to the translocon using yeast S. cerevisiaeas the model system. Point mutations in cytoplasmic loops six (L6) and eight (L8) of yeast Sec61p cause reductions in growth rates and defects in translocation of nascent polypeptides that utilize the cotranslational translocation pathway. Sec61 heterotrimers isolated from the L8 sec61 mutants have a greatly reduced affinity for 80S ribosomes. Cytoplasmic accumulation of protein precursors demonstrates that the initial contact between the large ribosomal subunit and the Sec61 complex is important for efficient insertion of a nascent polypeptide into the translocation pore. In contrast, point mutations in L6 of Sec61p inhibit cotranslational translocation without significantly reducing the ribosome binding activity, indicating that the L6 and L8 sec61mutants impact different steps in the cotranslational translocation pathway. An interaction between the signal recognition particle receptor (SR) and the Sec61 complex has been proposed to facilitate transfer of the ribosome-nascent chain (RNC) complex to an unoccupied translocon. The slow growth and cotranslational translocation defects caused by deletion of the transmembrane span of yeast SRβ (srp102pΔTMD) are exaggerated upon disruption of the SSH1 gene, which encodes the pore subunit of a cotranslational translocation channel. Disruption of the SBH2 gene, which encodes the β-subunit of the Ssh1p complex, likewise causes a synthetic growth defect when combined with srp102pΔTMD. The in vivo kinetics of translocon gating by RNCs were slow and inefficient in the ssh1Δ srp102pΔTMD mutant. A critical role for translocon β-subunits in SR recognition is supported by the observation that deletion of both translocon β-subunits causes a block in the cotranslational targeting pathway that resembles elimination of either subunit of the SR, and could be partially suppressed by expression of carboxy-terminal Sbh2p fragments.