Proteins and fatty acids often exist in solutions containing biological matter that are treated with membranes. These proteins and fatty acids interact with each other as well as with the membranes thereby affecting the flux. Binding of fatty acids to proteins results in complexes that are much larger than fatty acid molecules. Exploitation of this size difference to remove difficult to separate fatty acids from aqueous solutions by ultrafiltration was investigated in this study. In addition, the fouling of membrane by the protein-fatty acid mixtures containing free dissolved fatty acids was studied using bovine albumin (BSA)-caprylic system. Binding of caprylic acid to native and pasteurized BSA was examined by diafiltering pre equilibrated fatty acid-BSA mixtures. The rate of mass transfer of fatty acid molecules through boundary film surrounding the protein molecules was estimated using a BSA solution as the adsorbent phase in an agitated column. A stirred cell fitted with a polyethersulfone membrane (30 kDa) was used for the diafiltrations. Accumulation of fatty acid in the BSA layers fouled on the membrane was also estimated. Binding studies indicate that a native BSA molecule (at pH 6.8) could bind 7 fatty acid molecules in specific binding cavities while approximately 44 molecules are bound onto the surface. When BSA was pasteurized the specific binding decreased from 7 to 2 indicating unfolding of the molecule. In addition, the total binding capacity decreased from 44 to 24 moles/BSA mole and the rate of mass transfer decreased from 4.5/min to 3.6/min, indicating heat induced aggregation of BSA. At alkaline pH levels fatty acid anion acts as an anionic surfactant stabilizing the molecular conformation of the protein and reducing fouling. When pH was lowered to 3, flux severely declined. Unusually large accumulation of fatty acid in the deposited protein layers (caprylic/BSA ~ 10,000 moles) occurred indicating capillary condensation of undissociated fatty acids in the protein layer. Agitated column studies showed that proteins could be used as an adsorbent to remove hard to separate dissolved fatty acids from aqueous solutions. The separated protein-fatty acid complex may be further processed to manufacture animal feed.

The multigene family of Ca2+-binding S100 proteins comprises 22 members that have various important intra- and extracellular roles. The three inflammation-associated members of this family???S100A8, S100A9 and S100A12 (collectively termed &quotcalgranulins&quot)???are constitutive neutrophil and monocyte proteins also expressed by macrophages within acute and chronic inflammatory lesions, but not in tissue macrophages. They are expressed in human/murine wounds and by appropriately activated macrophages, microvascular endothelial cells and keratinocytes in vitro. The &quot calgranulins&quot are implicated in leukocyte activation/deactivation, fatty acid transport, leukocyte/fibroblast chemotaxis, transmigration and adhesion, embryogenesis, wound healing, protection against oxidants and antibacterial defence. Chapter 3 of this thesis explores growth-factor- and cytokine-mediated regulation and expression of S100A8 and S100A9 in fibroblasts, and demonstrates spatio-temporal expression of S100A8 in rat dermal wounds. Fibroblasts are stromal resident cells with important regulatory immune-inflammatory functions in wound healing, tissue remodelling and fibrosis. Fibroblast migration, proliferation, differentiation and their synthetic repertoire are modulated by various factors including extracellular matrix components, growth factors, prostaglandins, reactive oxygen species and cytokines. Fibroblast growth factor-2 (FGF-2), interleukin-1?(IL-1? and platelet-derived growth factor (PDGF) are potent fibroblast mitogens; PDGF and transforming growth factor-? (TGF-? are fibroblast chemoattractants. FGF-2 and IL-1?promote fibroblast proliferation, whereas TGF-?promotes myofibroblast differentiation and collagen production. Lipopolysaccharide (LPS), interferon ?(IFN?, tumour-necrosis factor ? (TNF?, TGF-?and PDGF did not induce the S100A8 gene in fibroblasts whereas FGF-2 (25 ng/ml) maximally induced mRNA 12 hr. after stimulation and this declined over 36 hr. The FGF-2 response was strongly enhanced and prolonged by optimal levels of heparin (1-10 IU/ml), maximally at 18 hr. post-stimulation. FGF-2/heparin-induced responses depended on cell-cell contact in vitro. IL-1?(10 U/ml) alone, or in synergy with FGF-2/heparin strongly induced the gene in 3T3 and primary fibroblasts. Dexamethasone (10???6 M) enhanced LPS- and FGF-2/-IL-1?induced responses. S100A9 mRNA was not induced by any of these mediators. Induction of S100A8 in the absence of S100A9 was confirmed in primary fibroblast-like cells by real-time reverse-transcriptase polymerase chain-reaction. FGF-2-heparin- and IL-1?induced mRNA expression depended on de-novo protein synthesis and was partially mediated by the mitogenactivated protein kinase pathway of activation. Preliminary promoter deletion analyses indicated that FGF-2-responsive elements in the gene promoter were distinct from those responsive to IL-1? TGF-?(2 ng/ml) significantly suppressed gene induction mediated by FGF-2 ?heparin/LPS/dexamethasone, but not by IL-1? TGF-?may compromise mRNA stability. Protein levels in FGF-2-heparin-IL-1?stimulated fibroblasts correlated well with mRNA levels and expression was mainly cytoplasmic. Immunohistochemistry indicated S100A8 associated with keratinocytes, neutrophils, macrophage-like cells and some hair follicles in wounded rat skin. Rat wounds also contained numerous S100A8- positive fibroblast-like cells 2 and 4 days post-injury; numbers declined by 7 days. Upregulation of S100A8 by FGF-2/IL-1? down-regulation by TGF-? and time-dependent expression of S100A8 in wound fibroblasts suggest a role in fibroblast differentiation at sites of inflammation and repair. Intracellular fibroblast-derived S100A8 may also regulate intracellular redox equilibrium and antioxidant defence. Atherosclerosis is a progressive chronic disease with complex aetiology and pathogenesis. S100A1 and S100B are associated with dendritic cells and lymphocytes in experimental rodent and human atherosclerotic lesions. Monocytes and macrophages in plaques of ApoE???/??? mice express S100A9 but not S100A8. Myeloperoxidase and HOClmediated oxidative mechanisms are fundamental in the pathogenesis of atherosclerosis and S100A8 is exquisitely sensitive to HOCl oxidation which generates sulphinamide bonds, novel non-reducible cysteine-lysine covalent bonds. Chapter 4 of this thesis presents novel evidence that, in contrast to the murine ApoE???/??? model, the three human &quot calgranulins&quot were expressed in human atherosclerotic plaques, but not in normal arteries. High levels of S100A8, S100A9 and S100A12 were evident in macrophages and foam cells. Some neovessels were anti-S100A8-/anti-S100A9-immunoreactive; S100A9 staining was also evident on the extracellular matrix. Patterns of expression of S100A8, S100A9 and S100A12 were overlapping in serial sections, except that only smooth muscle cells were S100A12-positive. S100A8 and S100A9 mRNA were also expressed by macrophages, foam cells and endothelial cells, indicating gene up-regulation rather than passive protein uptake. Western blotting of plaque extracts revealed monomeric S100A8, S100A9 and S100A12 and larger complexes. Some were resistant to reduction, suggesting non-disulfide covalent cross-linking, possibly via sulphinamide bonds. Stable S100A8-S100A9 complexes were also detected after immunoaffinity purification. In an in-vitro system, molar ratios of HOCl of &gt1 generated stable complexes of S100A8 and S100A9 whereas ~800 and ~100-fold excess HOCl oxidises apolipoprotein B-100 and BSA, respectively. S100A8 and S100A9 protected low-density lipoprotein (LDL) against HOCl oxidation in a thiol-independent manner. Because HOCl-oxidised S100s did not contain epitopes recognised by an antibody used to detect HOCl-oxidised proteins in plaque, levels of oxidised proteins in plaque are likely to be significantly greater than described. S100A8 and S100A9 may protect LDL by functioning as HOCl-scavengers. However, chronic oxidative cross-linking of S100A8 and S100A9 with other proteins and extracellular matrix components may contribute to plaque pathogenesis. These studies support key roles for the &quot calgranulins&quot in chronic inflammation, wound healing and atherogenesis possibly by regulating cellular differentiation, activation and modulation of redox-dependent mechanisms.

The multigene family of Ca2+-binding S100 proteins comprises 22 members that have various important intra- and extracellular roles. The three inflammation-associated members of this family???S100A8, S100A9 and S100A12 (collectively termed &quotcalgranulins&quot)???are constitutive neutrophil and monocyte proteins also expressed by macrophages within acute and chronic inflammatory lesions, but not in tissue macrophages. They are expressed in human/murine wounds and by appropriately activated macrophages, microvascular endothelial cells and keratinocytes in vitro. The &quot calgranulins&quot are implicated in leukocyte activation/deactivation, fatty acid transport, leukocyte/fibroblast chemotaxis, transmigration and adhesion, embryogenesis, wound healing, protection against oxidants and antibacterial defence. Chapter 3 of this thesis explores growth-factor- and cytokine-mediated regulation and expression of S100A8 and S100A9 in fibroblasts, and demonstrates spatio-temporal expression of S100A8 in rat dermal wounds. Fibroblasts are stromal resident cells with important regulatory immune-inflammatory functions in wound healing, tissue remodelling and fibrosis. Fibroblast migration, proliferation, differentiation and their synthetic repertoire are modulated by various factors including extracellular matrix components, growth factors, prostaglandins, reactive oxygen species and cytokines. Fibroblast growth factor-2 (FGF-2), interleukin-1?(IL-1? and platelet-derived growth factor (PDGF) are potent fibroblast mitogens; PDGF and transforming growth factor-? (TGF-? are fibroblast chemoattractants. FGF-2 and IL-1?promote fibroblast proliferation, whereas TGF-?promotes myofibroblast differentiation and collagen production. Lipopolysaccharide (LPS), interferon ?(IFN?, tumour-necrosis factor ? (TNF?, TGF-?and PDGF did not induce the S100A8 gene in fibroblasts whereas FGF-2 (25 ng/ml) maximally induced mRNA 12 hr. after stimulation and this declined over 36 hr. The FGF-2 response was strongly enhanced and prolonged by optimal levels of heparin (1-10 IU/ml), maximally at 18 hr. post-stimulation. FGF-2/heparin-induced responses depended on cell-cell contact in vitro. IL-1?(10 U/ml) alone, or in synergy with FGF-2/heparin strongly induced the gene in 3T3 and primary fibroblasts. Dexamethasone (10???6 M) enhanced LPS- and FGF-2/-IL-1?induced responses. S100A9 mRNA was not induced by any of these mediators. Induction of S100A8 in the absence of S100A9 was confirmed in primary fibroblast-like cells by real-time reverse-transcriptase polymerase chain-reaction. FGF-2-heparin- and IL-1?induced mRNA expression depended on de-novo protein synthesis and was partially mediated by the mitogenactivated protein kinase pathway of activation. Preliminary promoter deletion analyses indicated that FGF-2-responsive elements in the gene promoter were distinct from those responsive to IL-1? TGF-?(2 ng/ml) significantly suppressed gene induction mediated by FGF-2 ?heparin/LPS/dexamethasone, but not by IL-1? TGF-?may compromise mRNA stability. Protein levels in FGF-2-heparin-IL-1?stimulated fibroblasts correlated well with mRNA levels and expression was mainly cytoplasmic. Immunohistochemistry indicated S100A8 associated with keratinocytes, neutrophils, macrophage-like cells and some hair follicles in wounded rat skin. Rat wounds also contained numerous S100A8- positive fibroblast-like cells 2 and 4 days post-injury; numbers declined by 7 days. Upregulation of S100A8 by FGF-2/IL-1? down-regulation by TGF-? and time-dependent expression of S100A8 in wound fibroblasts suggest a role in fibroblast differentiation at sites of inflammation and repair. Intracellular fibroblast-derived S100A8 may also regulate intracellular redox equilibrium and antioxidant defence. Atherosclerosis is a progressive chronic disease with complex aetiology and pathogenesis. S100A1 and S100B are associated with dendritic cells and lymphocytes in experimental rodent and human atherosclerotic lesions. Monocytes and macrophages in plaques of ApoE???/??? mice express S100A9 but not S100A8. Myeloperoxidase and HOClmediated oxidative mechanisms are fundamental in the pathogenesis of atherosclerosis and S100A8 is exquisitely sensitive to HOCl oxidation which generates sulphinamide bonds, novel non-reducible cysteine-lysine covalent bonds. Chapter 4 of this thesis presents novel evidence that, in contrast to the murine ApoE???/??? model, the three human &quot calgranulins&quot were expressed in human atherosclerotic plaques, but not in normal arteries. High levels of S100A8, S100A9 and S100A12 were evident in macrophages and foam cells. Some neovessels were anti-S100A8-/anti-S100A9-immunoreactive; S100A9 staining was also evident on the extracellular matrix. Patterns of expression of S100A8, S100A9 and S100A12 were overlapping in serial sections, except that only smooth muscle cells were S100A12-positive. S100A8 and S100A9 mRNA were also expressed by macrophages, foam cells and endothelial cells, indicating gene up-regulation rather than passive protein uptake. Western blotting of plaque extracts revealed monomeric S100A8, S100A9 and S100A12 and larger complexes. Some were resistant to reduction, suggesting non-disulfide covalent cross-linking, possibly via sulphinamide bonds. Stable S100A8-S100A9 complexes were also detected after immunoaffinity purification. In an in-vitro system, molar ratios of HOCl of &gt1 generated stable complexes of S100A8 and S100A9 whereas ~800 and ~100-fold excess HOCl oxidises apolipoprotein B-100 and BSA, respectively. S100A8 and S100A9 protected low-density lipoprotein (LDL) against HOCl oxidation in a thiol-independent manner. Because HOCl-oxidised S100s did not contain epitopes recognised by an antibody used to detect HOCl-oxidised proteins in plaque, levels of oxidised proteins in plaque are likely to be significantly greater than described. S100A8 and S100A9 may protect LDL by functioning as HOCl-scavengers. However, chronic oxidative cross-linking of S100A8 and S100A9 with other proteins and extracellular matrix components may contribute to plaque pathogenesis. These studies support key roles for the &quot calgranulins&quot in chronic inflammation, wound healing and atherogenesis possibly by regulating cellular differentiation, activation and modulation of redox-dependent mechanisms.

A series of compounds, 7-8 and 20-25, were tested as competitive inhibitors of 2,5-dihydroxyacetanilide epoxidase I (DHAE I) and DHAE II. A Hammett plot was constructed for each enzyme to determine the effect of electron density on inhibition. DHAE I gave a linear, highly correlated plot (r²=0.91) that signifies the importance of the amide oxygen in 1 on substrate binding. The plot for DHAE 11 is curved showing the greatest degree of inhibition with 7 suggesting steric factors within the active site control substrate binding. From these data, we conclude that each enzyme binds substrate in an opposite fashion and that this alone controls the stereochemistry of epoxide formation in 2 and 3. Alternative substrates, 26-29 and 33, were also synthesized and tested for product formation. All compounds, except 29, were accepted as alternative substrates, although the rates varied significantly. Surprisingly, 33 was accepted as an alternative substrate of DHAE II suggesting that the conformation of the amide bond in 33 is similar to the conformation required for catalytic activity in this enzyme. This information was then used to design ligands for affinity column separation of DHAE I and DHAE II from their protein mixtures. 35 and 36 were synthesized and attached to carbonyl di-imidazole activated agarose. Column I was tested three times with DHAE I enzyme preparations. The first attempt did not result in active enzyme being eluted from the column. The second attempt maintained the resin in the oxidized state. Protein was found to elute very quickly: no protein was found after fraction 4. The third attempt resulted in active enzyme in fractions 4-23. Column 2 was used twice for the attempted isolation of DHAE II from its protein mixture. The second attempt for column 2 mirrored the results for the third attempt with column 1. Neither column resulted in homogeneous enzyme by SDS-PAGE. / Graduation date: 2003

In proteomic studies, separate experimental protocols have been necessary to identify proteins, determine their function, and predict their three-dimensional structure. In this study, a function-based separation of proteins was conceived to fractionate proteins prior to enzymatic digestion. In the initial demonstration of this technique, a DNA substrate was used to separate the DNA-binding proteins from the rest of the proteins in a lysate in order to identify protein function and to simplify the complex mixture of proteins. A total of 232 putative DNA-binding proteins and over 540 proteins in all were identified from E. coli. Hypothetical or unknown proteins were found, some of which bind to DNA. As a part of this demonstration, changes in protein expression caused by different environmental conditions (aerobic and anaerobic atmospheres) were observed. In a second demonstration, aimed at determining the three-dimensional structure of the DNA binding proteins, binding sites were blocked with oligonucleotides, and the modified proteins were purified, enzymatically digested, and subjected to tandem mass spectrometry. The amino acids in the DNA-binding domains of three proteins were determined. In a final application of function-based separation, DNA-binding proteins were digested with trypsin and the resulting peptides were separated using HPLC and subsequently analyzed using MALDI TOF/TOF and ESI Q-TOF instruments to study the complementary nature of the two ionization techniques, taking into account the differences between the mass analyzers. Based on the analysis of a large data set containing hundreds of peptides and thousands of individual amino acids, some of the currently held notions regarding the ionization processes were confirmed. ESI tends to favor the analysis of hydrophobic amino acids and peptides while MALDI is disposed toward mainly basic and aromatic species. These tendencies in ionization account in large part for the complementary nature of the peptides and proteins identified by the ESI and MALDI instruments and make it necessary to employ both types of instruments to gain the most information out of a given sample in a proteomics study. / Graduation date: 2004

Graduation date: 1994

Les systèmes d’efflux tripartite de type Resistance, Nodulation and cell-Division (RND) sont essentiels dans le maintien de phénotypes de résistance multidrogues et contre les métaux lourds dans nombreuses bactéries Gram-négatives. Le transport de ces composés toxiques hors de la cellule est permis par l’assemblage d’une protéine de type antiporteur cation/proton (unité RND) insérée dans la membrane interne, connectée à une protéine insérée dans la membrane externe, pour former un canal de sorti qui traverse l’entièreté de l’enveloppe cellulaire. Le troisième composant du système, la protéine de type membrane fusion protein (MFP) qui est aussi appelée periplasmic adaptor protein (PAP), est requis pour permettre l’assemblage de tout ce complexe à trois composants. Cependant, les MFPs sont supposées jouer un rôle important et actif dans le mécanisme d’efflux du substrat. Pour mieux comprendre le rôle des MFPs au sein des systèmes d’efflux de type RND, nous avons étudié les protéines ZneB (précédemment appelée HmxB) et SilB, les composants périplasmiques des systèmes ZneCBA et SilABC responsables de la résistance aux métaux lourds chez Cupriavidus metallidurans CH34. Nous avons identifié la spécificité de liaison au substrat de ces protéines, montrant leur capacité à fixer le zinc (ZneB), ou le cuivre et l’argent (SilB). De plus, nous avons résolu la structure cristalline de ZneB à une résolution de 2.8 Å dans la forme apo- et avec un ion zinc fixé. La structure de ZneB possède une architecture générale composée de quatre domaines caractéristiques des MFPs, et la présence du site de coordination au zinc dans une région très flexible à l’interface des domaines β-barrel et membrane proximal. Les modifications structurales que la protéine subit lors de la fixation du zinc on été observée dans le cristal mais aussi en solution, ce qui suggère un rôle actif des MFPs dans le mécanisme d’efflux des métaux, vraisemblablement via la fixation et le relargage de l’ion à l’antiporteur. Les études de sélectivité de transport des antiporteurs ZneA et SilA montre que ces dernières et leurs protéines périplasmiques respectives ont des affinités similaires pour les métaux lourds. De plus, les études de transport ont apportés des arguments en faveur de l’hypothèse de capture cytoplasmique du substrat par l’antiporteur, tandis que la capacité des protéines périplasmiques à fixer les métaux lourds a apporté des arguments en faveur de l’hypothèse de capture périplasmique du substrat par l’antiporteur. Les deux modes de capture pourraient en réalité coexister ; cependant, le débat autour du compartiment cellulaire de capture du substrat par l’antiporteur est complexe et requiert de plus amples efforts afin d’être cerné. / Tripartite resistance nodulation cell division (RND)-based efflux complexes are paramount for multidrug and heavy metal resistance in numerous Gram-negative bacteria. The transport of these toxic compounds out of the cell is driven by the inner membrane proton/substrate antiporter (RND protein) connected to an outer membrane protein to form an exit duct that spans the entire cell envelope. The third component, a membrane fusion protein (MFP) also called periplasmic adaptor protein, is required for the assembly of this complex. However, MFPs are also proposed to play an important active role in substrate efflux. To better understand the role of MFPs in RND-driven efflux systems, we studied ZneB (formerly HmxB) and SilB, the MFP components of the ZneCAB and SilABC heavy metal RND-driven efflux complexes from Cupriavidus metallidurans CH34. We have identified the substrate binding specificity of the proteins, showing their ability to selectively bind zinc (ZneB), or copper and silver cations (SilB). Moreover, we have solved the crystal structure of the apo- and the metal-bound forms of ZneB to 2.8 Å resolution. The structure of ZneB displays a general architecture composed of four domains characteristic of MFPs, and it reveals the metal coordination site at the very flexible interface between the β-barrel and the membrane proximal domains. Structural modifications of the protein upon zinc binding were observed in both the crystal structure and in solution, suggesting an active role of MFPs in substrate efflux possibly through binding and release. The selectivity assays of the antiporter proteins ZneA and SilA demonstrated similar specificities in relation to their cognate MFPs toward heavy metal cations. Moreover, antiporter transport assays provide evidence for cytoplasmic substrate capture by this protein, whereas MFP substrate binding provides evidence for periplasmic substrate capture. Therefore, both modes of capture might co-exist; nevertheless, the substrate capture issue is a complex topic still needing consequent efforts to understand it.

Aromatic oligoamide foldamers possess a high potential for mimicking the secondary structures of biopolymers. These oligomers are efficiently designed, easy to synthesize, and allow one to reach a wide range of stable folded states.Thus far, we want to utilize stable states of these oligomers and a variety of groove binding agents combining together , and then study their influence on DNA.

This study investigates the influence of ionic strength on the binding constant (KDOC) between benzo(h)quinoline (BHQ) and LHA by using fluorescence quenching method. Being a basic polycyclic aromatic hydrocarbon, BHQ is the dominated solute as the solution¡¦s pH value is higher than BHQ¡¦s pKb. In contrast, BHQ+ is the major species as the solution¡¦s pH value is lower than BHQ¡¦s pKb. In a salty neutral or basic LHA solution, the cation will bind with the acidic functional groups of LHA, then the conformation of LHA would be coiled up to be small in size. Due to that, leading to the decrease of the corresponding BHQ¡¦s KDOC. Furthermore, the charge density of cation is an important factor in control of the variation of BHQ¡¦s KDOC. The lower charge density of cation is, the less BHQ¡¦s KDOC varied. Besides, SO4 2- may suppress the binding affinity between Na+ and the acidic function groups of LHA, so that lower variation of BHQ¡¦s KDOC was observed than that of Cl- in a Na+ contained LHA solution. In an acidic solution, cation will also bind with the acidic functional groups of LHA, leading to the decrease of the binding sites of BHQ+ on LHA and the corresponding BHQ+¡¦s KDOC. Besides, Mg2+ could provide more binding sites for the acidic functional groups of LHA than Na+, so that the variation of BHQ+¡¦s KDOC with Mg2+ addition is higher than that with Na+ addition.

Hox proteins are evolutionarily conserved transcription factors that control important developmental pathways in morphogenesis of the embryo. The Hoxa2 gene is expressed in the developing central nervous system in rhombomeres 2 to 7 and affects cellular differentiation. Few target genes of Hoxa2 protein have been identified so far and its mechanisms of regulating gene expression remain elusive. Previous work in our laboratory isolated Hoxa2 protein binding sequences from the E18 mouse spinal cord and hindbrain tissues using chromatin immunoprecipitation (ChIP). All isolated DNA fragments contain conserved GATG motifs. Sequence analysis revealed that one fragment belongs to the high temperature requirement factor A 3 (HtrA3) gene and another fragment belongs to the Dual specificity tyrosine kinase 4 (Dyrk4) gene. In this study, direct binding of Hoxa2 protein to the HtrA3 and Dyrk4 fragments was confirmed by electrophoretic mobility shift assays (EMSA). Site-directed mutagenesis and EMSA studies revealed that Hoxa2 protein binds to the multiple GATG motifs within these fragments. HtrA3 fragment also repressed luciferase gene expression in transient transfection and luciferase assays. Mutation of the DNA fragment showed that the repressive activity was affected by the GATG motifs, suggesting Hoxa2 protein regulated gene expression by binding to the GATG motif in the cis-regulatory element. In contrast to the inhibitory activity of Hoxa2 protein, a Hoxa2-VP16 fusion protein (Hoxa2 fused with an activation domain of a virion protein from herpes simplex virus) transactivates the luciferase expression by binding to GATG sites. RT-PCR and immunohistochemistry analysis revealed an upregulation of HtrA3 expression in Hoxa2-/- mice. This observation correlates with the inhibitory role of Hoxa2 protein acting upon the HtrA3 fragment in luciferase assays. Our data suggest that HtrA3 is a direct in vivo downstream target of Hoxa2 protein and support the activity regulation model in which Hox proteins selectively regulate target genes through occupation of multiple monomer binding sites.