BAYESIAN STATISTICAL METHODS IN GENE-ENVIRONMENT AND GENE-GENE INTERACTION STUDIES

Complex diseases such as cancer result from multiple genetic changes and environmental exposures. Due to the rapid development of genotyping and sequencing technologies, we are now able to more accurately assess causal effects of many genetic and environmental factors. Genome-wide association studies have been able to localize many causal genetic variants predisposing to certain diseases. However, these studies only explain a small portion of variations in the heritability of diseases. More advanced statistical models are urgently needed to identify and characterize some additional genetic and environmental factors and their interactions, which will enable us to better understand the causes of complex diseases. In the past decade, thanks to the increasing computational capabilities and novel statistical developments, Bayesian methods have been widely applied in the genetics/genomics researches and demonstrating superiority over some regular approaches in certain research areas. Gene-environment and gene-gene interaction studies are among the areas where Bayesian methods may fully exert its functionalities and advantages. This dissertation focuses on developing new Bayesian statistical methods for data analysis with complex gene-environment and gene-gene interactions, as well as extending some existing methods for gene-environment interactions to other related areas. It includes three sections: (1) Deriving the Bayesian variable selection framework for the hierarchical gene-environment and gene-gene interactions; (2) Developing the Bayesian Natural and Orthogonal Interaction (NOIA) models for gene-environment interactions; and (3) extending the applications of two Bayesian statistical methods which were developed for gene-environment interaction studies, to other related types of studies such as adaptive borrowing historical data. We propose a Bayesian hierarchical mixture model framework that allows us to investigate the genetic and environmental effects, gene by gene interactions (epistasis) and gene by environment interactions in the same model. It is well known that, in many practical situations, there exists a natural hierarchical structure between the main effects and interactions in the linear model. Here we propose a model that incorporates this hierarchical structure into the Bayesian mixture model, such that the irrelevant interaction effects can be removed more efficiently, resulting in more robust, parsimonious and powerful models. We evaluate both of the 'strong hierarchical' and 'weak hierarchical' models, which specify that both or one of the main effects between interacting factors must be present for the interactions to be included in the model. The extensive simulation results show that the proposed strong and weak hierarchical mixture models control the proportion of false positive discoveries and yield a powerful approach to identify the predisposing main effects and interactions in the studies with complex gene-environment and gene-gene interactions. We also compare these two models with the 'independent' model that does not impose this hierarchical constraint and observe their superior performances in most of the considered situations. The proposed models are implemented in the real data analysis of gene and environment interactions in the cases of lung cancer and cutaneous melanoma case-control studies. The Bayesian statistical models enjoy the properties of being allowed to incorporate useful prior information in the modeling process. Moreover, the Bayesian mixture model outperforms the multivariate logistic model in terms of the performances on the parameter estimation and variable selection in most cases. Our proposed models hold the hierarchical constraints, that further improve the Bayesian mixture model by reducing the proportion of false positive findings among the identified interactions and successfully identifying the reported associations. This is practically appealing for the study of investigating the causal factors from a moderate number of candidate genetic and environmental factors along with a relatively large number of interactions. The natural and orthogonal interaction (NOIA) models of genetic effects have previously been developed to provide an analysis framework, by which the estimates of effects for a quantitative trait are statistically orthogonal regardless of the existence of Hardy-Weinberg Equilibrium (HWE) within loci. Ma et al. (2012) recently developed a NOIA model for the gene-environment interaction studies and have shown the advantages of using the model for detecting the true main effects and interactions, compared with the usual functional model. In this project, we propose a novel Bayesian statistical model that combines the Bayesian hierarchical mixture model with the NOIA statistical model and the usual functional model. The proposed Bayesian NOIA model demonstrates more power at detecting the non-null effects with higher marginal posterior probabilities. Also, we review two Bayesian statistical models (Bayesian empirical shrinkage-type estimator and Bayesian model averaging), which were developed for the gene-environment interaction studies. Inspired by these Bayesian models, we develop two novel statistical methods that are able to handle the related problems such as borrowing data from historical studies. The proposed methods are analogous to the methods for the gene-environment interactions on behalf of the success on balancing the statistical efficiency and bias in a unified model. By extensive simulation studies, we compare the operating characteristics of the proposed models with the existing models including the hierarchical meta-analysis model. The results show that the proposed approaches adaptively borrow the historical data in a data-driven way. These novel models may have a broad range of statistical applications in both of genetic/genomic and clinical studies.

Characterization of the interaction between local cerebral metabolic rate for glucose and acid -base index in ischemic rat brain employing a double-isotope methodology

The association between increases in cerebral glucose metabolism and the development of acidosis is largely inferential, based on reports linking hyperglycemia with poor neurological outcome, lactate accumulation, and the severity of acidosis. We measured local cerebral metabolic rate for glucose (lCMRglc) and an index of brain pH--the acid-base index (ABI)--concurrently and characterized their interaction in a model of focal cerebral ischemia in rats in a double-label autoradiographic study, using ($\sp{14}$C) 2-deoxyglucose and ($\sp{14}$C) dimethyloxazolidinedione. Computer-assisted digitization and analysis permitted the simultaneous quantification of the two variables on a pixel-by-pixel basis in the same brain slices. Hemispheres ipsilateral to tamponade-induced middle cerebral occlusion showed areas of normal, depressed and elevated glucose metabolic rate (as defined by an interhemispheric asymmetry index) after two hours of ischemia. Regions of normal glucose metabolic rate showed normal ABI (pH $\pm$ SD = 6.97 $\pm$ 0.09), regions of depressed lCMRglc showed severe acidosis (6.69 $\pm$ 0.14), and regions of elevated lCMRglc showed moderate acidosis (6.88 $\pm$ 0.10), all significantly different at the.00125 level as shown by analysis of variance. Moderate acidosis in regions of increased lCMRglc suggests that anaerobic glycolysis causes excess protons to be generated by the uncoupling of ATP synthesis and hydrolysis. ^

Analysis of p53 function and regulation

p53 is a tumor suppressor gene that is the most frequent target inactivated in cancers. Overexpression of wild-type p53 in rat embryo fibroblasts suppresses foci formation by other cooperating oncogenes. Introduction of wild-type p53 into cells that lack p53 arrests them at the G1/S boundary and reverses the transformed phenotype of some cells. The function of p53 in normal cells is illustrated by the ability of p53 to arrest cells at G1 phase of the cell cycle upon exposure to DNA-damaging agents including UV-irradiation and biosynthesis inhibitors.^ Since the amino acid sequence of p53 suggested that it may function as a transcription factor, we used GAL4 fusion assays to test that possibility. We found that wild-type p53 could specifically activate transcription when anchored by the GAL4 DNA binding domain. Mutant p53s, which have lost the ability to suppress foci formation by other oncogenes, were not able to activate transcription in this assay. Thus, we established a direct correlation between the tumor suppression and transactivation functions of p53.^ Having learned that p53 was a transcriptional activator, we next sought targets of p53 activation. Because many transcription factors regulate their own expression, we tested whether p53 had this autoregulatory property. Transient expression of wild-type p53 in cells increased the levels of endogenous p53 mRNA. Cotransfection of p53 together with a reporter bearing the p53 promoter confirmed that wild-type p53 specifically activates its own promoter. Deletion analysis from both the 5$\sp\prime$ and 3$\sp\prime$ ends of the promoter minimized the region responsible for p53 autoregulation to 45 bp. Methylation interference identified nucleotides involved in protein-DNA interaction. Mutations within this protected site specifically eliminated the response of the promoter to p53. In addition, multiple copies of this element confer responsiveness to wild-type p53 expression. Thus, we identified a F53 responsive element within the p53 promoter.^ The presence of a consensus NF-$\kappa$B site in the p53 promoter suggested that NF-KB may regulate p53 expression. Gel-shift experiments showed that both the p50 homodimer and the p50/p65 heterodimer bind to the p53 promoter. In addition, the p65 subunit of NF-$\kappa$B activates the p53 promoter in transient transfection experiments. TNF $\alpha$, a natural NF-$\kappa$B inducer, also activates the p53 promoter. Both p65 activation and TNF $\alpha$ induction require an intact NF-$\kappa$B site in the p53 promoter. Since NF-$\kappa$B activation occurs as a response to stress and p53 arrests cells in G1/S, where DNA repair occurs, activation of p53 by NF-$\kappa$B could be a mechanism by which cells recover from stress.^ In conclusion, we provided the first data that wild-type p53 functions as a transcriptional activator, whereas mutant p53 cannot. The correlation between growth suppression and transcriptional activation by p53 implies a pathway of tumor suppression. We have analyzed upstream components of the pathway by the identification of both p53 and NF-$\kappa$B as regulators of the p53 promoter. ^

Inhibitor of cytochrome c messenger RNA -protein interaction in stimulated rat skeletal muscle

The purpose of this work was to examine the possible mechanisms for the regulation of cytochrome c gene expression in response to increased contractile activity in rat skeletal muscle. The working hypothesis was that increased contractile activity enhances cytochrome c gene expression through a cis-element. A 110% increase in cytochrome c mRNA concentration was observed in tibialis anterior (TA) muscle after 9 days of chronic stimulation. Similar difference (120%) exists between soleus (SO) muscle of higher contractile activity and white vastus lateralis (WV) muscle of lower contractile activity. These results suggest that the endogenous cytochrome c gene expression is regulated by contractile activity. Cytochrome c-reporter genes were injected into skeletal muscles to identify the cis-element that is responsible for the regulation. Although the data was inconclusive, part of it suggested the importance of the 3$\sp\prime$-untranslated region (3$\sp\prime$-UTR) in mediating the response to increased contractile activity.^ RNA gel mobility shift (GMSA) and ultraviolet (UV) cross-linking assays revealed specific RNA-protein interaction in a 50-nucleotide region of the 3$\sp\prime$-UTR in unstimulated TA muscle. Computer analysis predicted a stem-loop structure of 17 nucleotides, which provides a structural basis for RNA-protein interaction. These 17 nucleotides are 100% conserved among rat, mouse and human cytochrome c genes and their 13 pseudogenes, suggesting a functional role for this region. The RNA-protein interaction was significantly less in highly active SO muscle than in inactive WV muscle and was dramatically decreased in stimulated TA muscle due to a protein inhibitor(s) associated with ribosome. It is possible that cytochrome c mRNAs undergoing translation are subject to a compartmentalized regulatory influence.^ The conclusion from these results is that increases in contractile activity induce or activate a protein inhibitor(s) associated with ribosome in rat skeletal muscle. The inhibitor decreases RNA-protein interaction in the 3$\sp\prime$-UTR of cytochrome c mRNA, which may result in increased mRNA stability and/or translation. ^

A comparative analysis of hospital -based physician monitoring systems

The purpose of the multiple case-study was to determine how hospital subsystems (such as physician monitoring and credentialing; quality assurance; risk management; and peer review) were supporting the monitoring of physicians? Three large metropolitan hospitals in Texas were studied and designated as hospitals #1, #2, and #3. Realizing that hospital subsystems are a unique entity and part of a larger system, conclusions were made on the premises of a quality control system, in relation to the tools of government (particularly the Health Care Quality Improvement Act (HCQIA)), and in relation to itself as a tool of a hospital.^ Three major analytical assessments were performed. First, the subsystems were analyzed as to their "completeness"; secondly, the subsystems were analyzed for "performance"; and thirdly, the subsystems were analyzed in reference to the interaction of completeness and performance.^ The physician credentialing and monitoring and the peer review subsystems as quality control systems were most complete, efficient, and effective in hospitals #1 and #3. The HCQIA did not seem to be an influencing factor in the completeness of the subsystem in hospital #1. The quality assurance and risk management subsystem in hospital #2 was not representative of completeness and performance and the HCQIA was not an influencing factor in the completeness of the Q.A. or R.M. systems in any hospital. The efficiency (computerization) of the physician credentialing, quality assurance and peer review subsystems in hospitals #1 and #3 seemed to contribute to their effectiveness (system-wide effect).^ The results indicated that the more complete, effective, and efficient subsystems were characterized by (1) all defined activities being met, (2) the HCQIA being an influencing factor, (3) a decentralized administrative structure, (4) computerization an important element, and (5) staff was sophisticated in subsystem operations. However, other variables were identified which deserve further research as to their effect on completeness and performance of subsystems. They include (1) medical staff affiliations, (2) system funding levels, (3) the system's administrative structure, and (4) the physician staff "cultural" characteristics. Perhaps by understanding other influencing factors, health care administrators may plan subsystems that will be compatible with legislative requirements and administrative objectives. ^

The identification and regulation of a novel interaction occurring between $\beta$-catenin and the actin -bundling protein fascin

Catenins were first characterized as linking the cytoplasmic domains of cadherin cell-cell adhesion molecules to the cortical actin cytoskeleton. In addition to their essential role in modulating cadherin adhesion, catenins have more recently been indicated to participate in cell and developmental signaling pathways. $\beta$-catenin, for example, associates directly with receptor tyrosine kinases and transcription factors such as LEF-1/TCF, and tranduces developmental signals within the Wnt pathway. $\beta$-catenin also appear to a role in regulating cell proliferation via its interaction with the tumor supressor protein APC. I have employed the yeast two-hybrid method to reveal that fascin, a bundler of actin filaments, binds to $\beta$-catenin's central Armadillo-repeat domain. The $\beta$-catenin-fascin interaction exists in cell lines as well as in animal brain tissues as revealed by immunoprecipitation analysis, and substantiated in vitro with purified proteins. Fascin additionally binds to plakoglobin, which contains a more divergent Armadillo-repeat domain. Fascin and E-cadherin utilize a similar binding-site within $\beta$-catenin, such that they form mutually exclusive complexes with $\beta$-catenin. Fascin and $\beta$-catenin co-localize at cell-cell borders and dynamic cell leading edges of epithelial and endothelial cells. Total immunoprecipitable b-catein has several isoforms, only the hyperphosphorylated isoform 1 associated with fascin. An increased $\beta$-catenin-fascin interaction was observed in HGF stimulated cells, and in Xenopus embryos injected with src kinase RNAs. The increased $\beta$-catenin association with fascin is correlated with increased levels of $\beta$-catenin phosphorylation. $\beta$-catenin, but not fascin, can be readily phosphorylated on tyrosine in vivo following src injection of embryos, or in vitro following v-src addition to purified protein components. These observations suggest a role of $\beta$-catenin phosphorylation in regulating its interaction with fascin, and src kinase may be an important regulator of the $\beta$-catenin-fascin association in vivo. The $\beta$-catenin-fascin interaction represents a novel catenin complex, that may conceivably regulate actin cytoskeletal structures, cell adhesion, and cellular motility, perhaps in a coordinate manner with its functions in cadherin and APC complexes. ^

Analysis of VirB4, a membrane-associated ATPase subunit essential for T -complex transport in Agrobacterium tumefaciens

Agrobacterium tumefaciens is a plant pathogen with the unique ability to export oncogenic DNA-protein complexes (T-complexes) to susceptible plant cells and cause crown gall tumors. Delivery of the T-complexes across the bacterial membranes requires eleven VirB proteins and VirD4, which are postulated to form a transmembrane transporter. This thesis examines the subcellular localization and oligomeric structure of the 87-kDa VirB4 protein, which is one of three essential ATPases proposed to energize T-complex transport and/or assembly. Results of subcellular localization studies showed that VirB4 is tightly associated with the cytoplasmic membrane, suggesting that it is a membrane-spanning protein. The membrane topology of VirB4 was determined by using a nested deletion strategy to generate random fusions between virB4 and the periplasmically-active alkaline phosphatase, $\sp\prime phoA$. Analysis of PhoA and complementary $\beta$-galactosidase reporter fusions identified two putative periplasmically-exposed regions in VirB4. A periplasmic exposure of one of these regions was further confirmed by protease susceptibility assays using A. tumefaciens spheroplasts. To gain insight into the structure of the transporter, the topological configurations of other VirB proteins were also examined. Results from hydropathy analyses, subcellular localization, protease susceptibility, and PhoA reporter fusion studies support a model that all of the VirB proteins localize at one or both of the bacterial membranes. Immunoprecipitation and Co$\sp{2+}$ affinity chromatography studies demonstrated that native VirB4 (87-kDa) and a functional N-terminally tagged HIS-VirB4 derivative (89-kDa) interact and that the interaction is independent of other VirB proteins. A $\lambda$ cI repressor fusion assay supplied further evidence for VirB4 dimer formation. A VirB4 dimerization domain was localized to the N-terminal third of the protein, as judged by: (i) transdominance of an allele that codes for this region of VirB4; (ii) co-retention of a His-tagged N-terminal truncation derivative and native VirB4 on Co$\sp{2+}$ affinity columns; and (iii) dimer formation of the N-terminal third of VirB4 fused to the cI repressor protein. Taken together, these findings are consistent with a model that VirB4 is topologically configured as an integral cytoplasmic membrane protein with two periplasmic domains and that VirB4 assembles as homodimers via an N-terminal dimerization domain. Dimer formation is postulated to be essential for stabilization of VirB4 monomers during T-complex transporter assembly. ^

The cDNA cloning of murine HIP /L29 and the functional analysis of HIP/L29 domains

Human heparin/heparan sulfate interacting protein/L29 (HIP/L29) is a heparin/heparan sulfate (Hp/HS) binding protein found in many adult human tissues. Potential functions of this protein are promotion of embryo adhesion, modulation of blood coagulation, and control of cell growth. While these activities are diverse, the ability of human HIP/L29 to interact with Hp/HS at the cell surface may be a unifying mechanism of action since Hp/HS influences all of these processes. A murine ortholog has been identified that has 78.8% homology over the entire sequence and identity over the N-terminal 64 amino acids when compared to human HIP/L29. Northern, Western, and immunohistochemical analysis shows that murine HIP/L29 mRNA and protein are expressed in a tissue specific manner. Murine HIP/L29 is enriched in the membrane fraction of NmuMG cells where it is eluted with high salt, suggesting that it is a peripheral membrane protein. The ability of murine HIP/L29 to bind Hp is verified by studies using native and recombinant forms of murine HIP/L29. A synthetic peptide (HIP peptide-2) derived from the identical N-terminal region of HIP/L29 proteins was tested for the ability to bind Hp and support cell adhesion. This peptide was chosen because it conforms to a proposed consensus sequence for Hp/HS binding peptides. HIP peptide-2 binds Hp in a dose-dependent, saturable, and selective manner and supports Hp-dependent cell adhesion. However, a scrambled form of this peptide displayed similar activities indicating a lack of peptide sequence specificity required for activity. Lastly, an unbiased approach was used to identify sequences within human and mouse HIP/L29 proteins necessary for Hp/HS binding. A panel of recombinant proteins was made that collectively are deficient in every human HIP/L29 domain. The activities of these deletion mutants and recombinant murine HIP/L29 were compared to the activity of recombinant human HIP/L29 in a number of assays designed to look at differences in the ability to bind Hp/HS. These studies suggest that each domain within human HIP/L29 is important for binding to Hp/HS and divergences in the C-terminus of human and mouse HIP/L29 account for a decrease in murine HIP/L29 affinity for Hp/HS. It is apparent that multiple domains within human and mouse HIP/L29 contribute to the function of Hp/HS binding. The interaction of multiple HIP/L29 domains with Hp/HS will influence the biological activity of HIP/L29 proteins. ^

NMR structural analysis of cardiac troponin C: Monitoring conformational changes induced by binding calcium, troponin I, and calcium -sensitizing drugs

Contraction of cardiac muscle is regulated through the Ca2+ dependent protein-protein interactions of the troponin complex (Tn). The critical role cardiac troponin C (cTnC) plays as the Ca2+ receptor in this complex makes it an attractive target for positive inotropic compounds. In this study, the ten Met methyl groups in cTnC, [98% 13C ϵ]-Met cTnC, are used as structural markers to monitor conformational changes in cTnC and identify sites of interaction between cTnC and cardiac troponin I (cTnI) responsible for the Ca2+ dependent interactions. In addition the structural consequences that a number of Ca2+-sensitizing compounds have on free cTnC and the cTnC·cTnI complex were characterized. Using heteronuclear NMR experiments and monitoring chemical shift changes in the ten Met methyl 1H-13C correlations in 3Ca2+ cTnC when bound to cTnI revealed an anti-parallel arrangement for the two proteins such that the N-domain of cTnI interacts with the C-domain of cTnC. The large chemical shifts in Mets-81, -120, and -157 identified points of contact between the proteins that include the C-domain hydrophobic surface in cTnC and the A, B, and D helical interface located in the regulatory N-domain of cTnC. TnI association [cTnI(33–80), cTnI(86–211), or cTnI(33–211)] was found also to dramatically reduce flexibility in the D/E central linker of cTnC as monitored by line broadening in the Met 1H- 13C correlations of cTnC induced by a nitroxide spin label, MTSSL, covalently attached to cTnC at Cys 84. TnI association resulted in an extended cTnC that is unlike the compact structure observed for free cTnC. The Met 1H-13C correlations also allowed the binding characteristics of bepridil, TFP, levosimendan, and EMD 57033 to the apo, 2Ca2+, and Ca2+ saturated forms of cTnC to be determined. In addition, the location of drug binding on the 3Ca2+cTnC·cTnI complex was identified for bepridil and TFP. Use of a novel spin-labeled phenothiazine, and detection of isotope filtered NOEs, allowed identification of drug binding sites in the shallow hydrophobic cup in the C-terminal domain, and on two hydrophobic surfaces on N-regulatory domain in free 3Ca2+ cTnC. In contrast, only one N-domain drug binding site exists in 3Ca2+ cTnC·cTnI complex. The methyl groups of Met 45, 60 and 80, which are grouped in a hydrophobic patch near site II in cTnC, showed the greatest change upon titration with bepridil or TFP, suggesting that this is a critical site of drug binding in both free cTnC and when associated with cTnI. The strongest NOEs were seen for Met-60 and -80, which are located on helices C and D, respectively, of Ca2+ binding site II. These results support the conclusion that the small hydrophobic patch which includes Met-45, -60, and -80 constitutes a drug binding site, and that binding drugs to this site will lead to an increase in Ca2+ binding affinity of site II while preserving maximal cTnC activity. Thus, the subregion in cTnC makes a likely target against which to design new and selective Ca2+-sensitizing compounds. ^

Interactions and structural analysis of the zinc -binding motif in decorin, a small leucine rich proteoglycan in extracellular matrix

Decorin, a dermatan/chondroitin sulfate proteoglycan, is ubiquitously distributed in the extracellular matrix (ECM) of mammals. Decorin belongs to the small leucine rich proteoglycan (SLRP) family, a proteoglycan family characterized by a core protein dominated by Leucine Rich Repeat motifs. The decorin core protein appears to mediate the binding of decorin to ECM molecules, such as collagens and fibronectin. It is believed that the interactions of decorin with these ECM molecules contribute to the regulation of ECM assembly, cell adhesions, and cell proliferation. These basic biological processes play critical roles during embryonic development and wound healing and are altered in pathological conditions such as fibrosis and tumorgenesis. ^ In this dissertation, we discover that decorin core protein can bind to Zn2+ ions with high affinity. Zinc is an essential trace element in mammals. Zn2+ ions play a catalytic role in the activation of many enzymes and a structural role in the stabilization of protein conformation. By examining purified recombinant decorin and its core protein fragments for Zn2+ binding activity using Zn2+-chelating column chromatography and Zn2+-equilibrium dialysis approaches, we have located the Zn2+ binding domain to the N-terminal sequence of the decorin core protein. The decorin N-terminal domain appears to contain two Zn2+ binding sites with similar high binding affinity. The sequence of the decorin N-terminal domain does not resemble any other reported zinc-binding motifs and, therefore, represents a novel Zn 2+ binding motif. By investigating the influence of Zn2+ ions on decorin binding interactions, we found a novel Zn2+ dependent interaction with fibrinogen, the major plasma protein in blood clots. Furthermore, a recombinant peptide (MD4) consisting of a 41 amino acid sequence of mouse decorin N-terminal domain can prolong thrombin induced fibrinogen/fibrin clot formation. This suggests that in the presence of Zn2+ the decorin N-terminal domain has an anticoagulation activity. The changed Zn2+-binding activities of the truncated MD4 peptides and site-directed mutagenesis generated mutant peptides revealed that the functional MD4 peptide might contain both a structural zinc-binding site in the cysteine cluster region and a catalytic zinc site that could be created by the flanking sequences of the cysteine cluster region. A model of a loop-like structure for MD4 peptide is proposed. ^

Analysis of Agrobacterium tumefaciens VirB6 and VirB9 of the VirB /D4 type IV secretion system

Agrobacterium tumefaciens uses the VirB/D4 type IV secretion system (T4SS) to translocate oncogenic DNA (T-DNA) and protein substrates to plant cells. Independent of VirD4, the eleven VirB proteins are also essential for elaboration of a conjugative pilus termed the T pilus. The focus of this thesis is the characterization and analysis of two VirB proteins, VirB6 and VirB9, with respect to substrate translocation and T pilus biogenesis. Observed stabilizing effects of VirB6 on other VirB subunits and results of protein-protein interaction studies suggest that VirB6 mediates assembly of the secretion machine and T pilus through interactions with VirB7 and VirB9. Topology studies support a model for VirB6 as a polytopic membrane protein with a periplasmic N terminus, a large internal periplasmic loop, five transmembrane segments, and a cytoplasmic C terminus. Topology studies and Transfer DNA immunoprecipitation (TrIP) assays identified several important VirB6 functional domains: (i) the large internal periplasmic loop mediates interaction of VirB6 with the T-DNA, (ii) the membrane spanning region carboxyl-terminal to the large periplasmic loop mediates substrate transfer from VirB6 to VirB8, and (iii) the terminal regions of VirB6 are required for substrate transfer to VirB2 and VirB9. To analyze structure-function relationships of VirB9, the phenotypic consequences of dipeptide insertion mutations were characterized. Substrate discriminating mutations were shown to selectively export the oncogenic T-DNA and VirE2 to plant cells or a mobilizable IncQ plasmid to bacterial cells. Mutations affecting VirB9 interactions with VirB7 and VirB10 were localized to the C- and N- terminal regions respectively. Additionally, “uncoupling” mutations identified in VirB11 and VirB6 that block T pilus assembly, but not substrate transfer to recipient cells, were also identified in VirB9. These results in conjunction with computer analysis establish that VirB9, like VirB6, is also composed of distinct regions or domains that contribute in various ways to secretion channel activity and T pilus assembly. Lastly, in vivo immunofluorescent studies suggest that VirB9 localizes to the outer membrane and may play a role similar to that of secretion/ushers of types II and III secretion systems to facilitate substrate translocation across this final bacterial barrier. ^