Redox-mediated upregulation ofhepatocyte iNOS transcriptionrequires coactivator PC4Carlos E. Marroquin, MD, Philip Y. Wai, MD, Paul C. Kuo, MD, MBA, and Hongtao Guo, MD, PhD,Durham, NC

Background. Redox-mediated upregulation of transcription of hepatocyte inducible nitric oxide synthase(iNOS) requires hepatocyte nuclear factor IV-alpha (HNF-4a). In this setting, PC4 is often isolated withHNF-4a in DNA-protein pull-down studies. Transcriptional coactivator PC4 facilitates activator-dependent transcription via interactions with basal transcriptional machinery that are independent ofPC4-DNA binding. We hypothesized that PC4 is a necessary component of HNF-4a--regulated redox-sensitive hepatocyte iNOS transcription.Methods. Murine CCL9.1 hepatocytes were stimulated with interleukin-1b (IL-1b; 1000 U/mL) in thepresence and absence of peroxide (H2O2; 50 nmol/L). Antisense and sense oligonucleotides to HNF-4aand PC4 were added selectively. Coimmunoprecipitation (Co-IP) studies determined the associationbetween HNF-4a and PC4. Transient transfection was performed with the use of a luciferase reporterconstruct containing the murine iNOS promoter (1.8 kb). Chromatin immunoprecipitation assaysdetermined in vivo binding of PC4 and HNF-4a to the iNOS promoter region.Results. Ablation of either HNF-4a or PC4 blunted the peroxide-mediated increase in the activation ofthe iNOS promoter. In IL-1b+H2O2 only, co-IP studies demonstrated the presence of an HNF-4a-PC4protein complex, and chromatin immunoprecipitation assays demonstrated that this complex binds to thegenomic iNOS promoter.Conclusions. Redox-mediated upregulation of hepatocyte iNOS transcription requires an HNF-4a-PC4transcriptional complex. (Surgery 2005;138:93-9.)

From the Department of Surgery, Duke University Medical Center, Durham, NC

CELLULAR REDOX MODULATORY SYSTEMS detoxify elec-trophiles and regulate key cellular functions suchas nucleotide synthesis, gene transcription andtranslation, post-translational protein modifica-tion, enzyme activation, cell-cycle regulation andsignal transduction.1-6 In this regard, we havepreviously demonstrated that hepatocyte expres-sion of inducible nitric oxide synthase (iNOS) andsynthesis of nitric oxide (NO) conveys protectiveantioxidant functions in models of sepsis, shock,and reperfusion injury.7,8 This hepatocellular re-

Presented at the 66th Annual Meeting of the Society ofUniversity Surgeons, Nashville, Tennessee, February 9-12, 2005.

Supported by National Institutes of Health grants AI44629,GM65113, and GM069331 (P.C.K.).

Reprint requests: Paul C. Kuo, MD, MBA, Department ofSurgery, 110 Bell Bldg, DUMC Box 3522, Durham, NC 27710.E-mail: kuo00004@mc.duke.edu.

0039-6060/$ - see front matter

� 2005 Mosby, Inc. All rights reserved.

doi:10.1016/j.surg.2005.03.014

dox regulatory system functions independently ofboth the oxidant species and the specific proin-flammatory cytokine. In interleukin (IL)-1b–treated rat hepatocytes, we have demonstrated thatiNOS gene transcription and promoter activity areincreased by oxidant stress mediated by peroxide,superoxide, or acetaminophen.9-13 Subsequently,in IL-1b–stimulated rat hepatocytes exposed tosuperoxide, we identified a redox-sensitive DR1cis-acting activator element (nt �1,327 to nt�1,315) in the iNOS promoter AGGTCAGGG-GACA. The corresponding transcription factorwas isolated by DNA affinity chromatography,sequenced, and identified to be hepatocyte nu-clear factor-4a (HNF-4a).9,14 In this setting, tran-scriptional coactivator PC4 is often isolated withHNF-4a in biotinylated DNA-protein pull-downstudies. PC4 facilitates activator-dependent tran-scription via interactions with basal transcriptionalmachinery that are independent of PC4-DNAbinding. We therefore hypothesized that PC4 is anecessary component of HNF-4a–regulated hepa-tocyte iNOS transcription.

SURGERY 93

SurgeryJuly 2005

94 Marroquin et al

MATERIAL AND METHODSMaterial. The mouse iNOS promoter (1745 bp;

GenBank accession no. L09126) was a gift fromProfessor Carl Nathan (Cornell University, Ithaca,NY) and subsequently was cloned into pGL3-basicluciferase reporter plasmid (Promega). CCL9.1mouse hepatocytes were maintained in Dulbeccomodified Eagle medium with 10% (v/v) heat-inactivated fetal calf serum, 100 U/mL penicillin,and 100 mg/mL streptomycin.

Induction of NO synthesis. IL-1b (1000 U/mL)was used in the absence of fetal calf serum toinduce NO synthesis. In selected instances, H2O2

(50 mmol/L) was added to induce oxidative stress.After incubation for 6 hours at 37�C in 5% CO2,the supernatants and cells were harvested forassays.

Transient transfection analysis. Hepatocyteswere transfected by the diethylaminoethyl(DEAE)-dextran technique.15,16 After the cellswere washed twice with media, 10 mg of plasmidDNA containing the mouse iNOS promoter con-struct was added per 107 cells to 1 mL of mediacontaining no serum but containing DEAE-dex-tran (250 mg/mL) and 50 mmol/L TRIS, pH 7.4,and prewarmed to 37�C. The suspension wasincubated at 37�C for 45 to 60 minutes, followedby a 1-minute shock with 10% (v/v) dimethylsulfoxide at room temperature (24�C). The cellswere washed, distributed to 100-mm plates, eachwith approx. 5 3 106 cells in 10 mL of completemedium, and incubated at 37�C in 5% CO2. Afterat least 24 hours, the medium was changed, andIL-1b (1000 U/mL), H2O2, or IL-1b+H2O2 wasadded. To control transfection efficiency betweengroups, we added 0.1 lg of pRL-TK to each well.Twenty-four hours after transfection, the cells wereharvested in 0.4 mL of reporter lysis buffer(Promega), and dual luciferase reporter assayswere performed by following the manufacturer’sprotocol. Forty microliters of lysate was used formeasurement in a luminometer (Turner DesignsTD-20/20). The sense and antisense oligonucleo-tides were designed according to GenBank se-quences to block the expression of HNF-4a(GenBank NM 008261; sense, 5#-GAGGGGCGAA-TGCGTCGCCA-3#; antisense, 5#-GCAGCTTGCTA-GATGGCT-3#) and PC4 (GenBank XM 372025;sense, 5#-AGGCAGTATTTAAAAAGCTC-3#; anti-sense, 5#-GCTCTCGAAGTCTCACCTGTC-3#).

Coimmunoprecipitation analysis. Cell culturemedium was removed and plates rinsed withphosphate-buffered saline (PBS) at room temper-ature. All the following steps were performed onice with the use of ice-cold buffers: 0.6 mL of

radioimmune precipitation buffer (1 3 PBS, 1%Nonidet P-40, 0.5% sodium deoxycholate, 0.1%sodium dodecylsulfate, 100 lg/mL phenylmethyl-sulfonyl fluoride (PMSF), and 60 lg/mL aproti-nin) was added to a 65-mm cell culture plate.Plates were scraped, and the cells lysed. Tenmicroliters of 10 mg/mL PMSF stock was added,followed by incubation for 30 to 60 minutes on ice.Whole-cell lysate was precleared by adding 0.25 lgof normal rat control IgG together with proteinA-agarose conjugate and incubated at 4�C for30 minutes. The beads were pelleted and thesupernatant incubated with primary antibody(polyclonal rabbit HNF-4a antibody; Santa CruzBiotechnology, Santa Cruz, Calif). Resuspendedprotein A-agarose was added and the tubes incu-bated at 4�C on a rocker platform overnight. Thepellet was collected by centrifugation at 1000g for5 minutes at 4�C and the supernatant discarded.The pellet was washed with radioimmune precip-itation buffer multiple times and resuspended inelectrophoresis sample buffer. Protein concentra-tion was determined by absorbance at 650 nm withthe use of protein assay reagent (Bio-Rad). Celllysate (50 lg/lane) were separated by 12% sodiumdodecylsulfate-polyacrylamide gel electrophoresis,and the products were electrotransferred topolyvinylidene difluoride membrane (AmershamBiosciences, Inc). The membrane was blockedwith 5% skim milk, PBS, 0.05% Tween for 1 hourat room temperature. After being washed 3,blocked membranes were incubated with goatPC4 polyclonal antibody (Santa Cruz Biotechnol-ogy) for 1 hour at room temperature, washedagain 3 times in PBS plus 0.05% Tween, and in-cubated with horseradish peroxidase–conjugatedsecondary antibody for 1 hour at room tem-perature. After an additional 3 washes, boundperoxidase activity was detected by the electro-chemiluminescence detection system (AmershamBiosciences, Inc.).

Nuclear extract preparation. Hepatocytes werewashed with PBS and harvested by scraping intocold PBS. The cell pellet obtained by centrifuga-tion was resuspended in buffer containing 10mmol/L HEPES, pH 7.9, 10 mmol/L KCl, 0.1mmol/L EDTA, 0.1 mmol/L EGTA, 1.0 mmol/Ldithiothreitol (DTT) and 0.5 mmol/L PMSF.Then, 10% Nonidet P-40 was added and vortexedbriefly, and the nuclei were pelleted by centri-fugation. The nuclear proteins were extractedwith buffer containing 20 mmol/L HEPES, pH7.9, 0.4 mmol/L NaCl, 1.0 mmol/L EDTA,1.0 mmol/L EGTA, 1.0 mmol/L DTT, and 1.0mmol/L PMSF. Insoluble material was removed by

SurgeryVolume 138, Number 1

Marroquin et al 95

Fig 1. A, Coimmunoprecipitation of HNF-4a and PC4.Coimmunoprecipitation experiments were performedwith the use of nuclear protein from cells treated with con-trol, IL-1b (1000 U/mL), catalase (100 mmol/L Heme),and/or H2O2, (50 nmol/L). Whole-cell lysate was pre-cleared and the supernatant incubated with primary anti-body (polyclonal rabbit HNF-4a antibody; Santa CruzBiotechnology). Protein concentration was determined,separated by 12% sodium dodecylsulfate-polyacrylamidegel electrophoresis, and the products were electrotrans-ferred to polyvinylidene difluoride membranes (Amer-sham Biosciences, Inc). Blocked membranes were thenincubated with goat PC4 polyclonal antibody (SantaCruz). After incubationwith horseradish peroxidase–con-jugated secondary antibody, bound peroxidase activitywere detected by the electrochemiluminescence detec-tion system (Amersham Biosciences, Inc). Blots represent3 experiments. B, PC4 protein expression. Immunoblotstudies were performed with nuclear protein from cells

centrifugation at 14000 rpm, and the supernatantcontaining the nuclear proteins was stored at�80�C until use.

Chromatin immunoprecipitation assay. Chroma-tin from hepatocytes was fixed and immuno-precipitated with the use of the chromatinimmunoprecipitation (ChIP) assay kit (UpstateBiotechnology, Inc) as recommended by the man-ufacturer. The purified chromatin was immuno-precipitated with the use of 10 lg of anti-PC4 oranti–HNF-4a or 5 lL of rabbit nonimmune serum.The input fraction corresponded to 0.1% and0.05% of the chromatin solution before immuno-precipitation. After DNA purification, the pres-ence of the selected DNA sequence was assessed bypolymerase chain reaction (PCR). The PCR pro-duct was 330 bp in length. The PCR program was94�C 3 4 minutes; followed by 94�C 3 45 seconds,55�C 3 45 seconds, and 72�C 3 45 seconds for atotal of 28 cycles; and then 72�C 3 7 minutes. PCRproducts were resolved in 10% acrylamide gels.The average size of the sonicated DNA fragmentssubjected to immunoprecipitation was 500 bp asdetermined by ethidium bromide gel electro-phoresis. The ChIP assay utilized PCR primersTCAATATTTCACTTTCATAA and TATATAGGAT-TATAATGTCC.

Statistical analysis. Data are presented as mean ±SEM of 3 or 4 experiments. Statistical analysis wasperformed with the Student t test; P values < .05were considered significant.

RESULTS

Coimmunoprecipitation studies. Murine CCL9.1hepatocytes were exposed to IL-1b to induce iNOSexpression. In selected instances, H2O2 was alsoadded. Nuclear protein was isolated and co-IPstudies performed to delineate potential associa-tion between HNF-4a and PC4 (Fig 1, A) Previousstudies suggest that HNF-4a and PC4 proteinsassociate in the presence of oxidative stress. Incontrol cells, and in cells stimulated with IL-1b andH2O2, there was no detectable PC4 protein. Incontrast, in the presence of IL-1b+H2O2, PC4 wasreadily detected. Subsequently, administration ofcatalase to inactivate peroxide in IL-1b+H2O2 cellsablated detectable PC4 in the co-IP assays. Immu-noblot analysis of nuclear HNF-4a in control cells,and in cells treated with IL-1b, H2O2, and IL-1b+H2O2 was also performed to normalize for

treated with control, IL-1b (1000 U/mL), catalase (100mmol/LHeme), and/orH2O2, (50nmol/L). Blots repre-sent 3 experiments.

SurgeryJuly 2005

96 Marroquin et al

HNF-4a expression. Equivalent amounts of HNF-4awere noted among the various treatment groups.These data suggest that a nuclear HNF-4a-PC4protein complex occurs exclusively in the presenceof both IL-1b– and H2O2-induced oxidative stress.

To determine whether PC4 might be inducedunder conditions of oxidative stress, we performedimmunoblot analysis of nuclear PC4 (Fig 1, B).PC4 protein expression was not altered underthese treatment conditions.

Transient transfection analysis. To determine apotential transcriptional role for the HNF-4a-PC4protein complex in redox-dependent transcrip-tional regulation of iNOS, transient transfectionanalysis was performed with the use of a luciferasereporter construct containing the murine iNOSpromoter (1.8kB; Fig 2). Again, murine CCL9.1hepatocytes were exposed to IL-1b to induce iNOSexpression. In selected instances, H2O2 and/or

Fig 2. Role of HNF-4a and PC4 in hepatocyte iNOS pro-moter activation. The mouse iNOS promoter (1745 bp;GenBank accession no. L09126) was cloned into pGL3-basic luciferase reporter plasmid (Promega). Hepato-cytes were transfected by the DEAE-dextran technique.After at least 24 hours, the medium was changed, andIL-1b (1000 U/mL), H2O2 or IL-1b+H2O2 was added.To control transfection efficiency between groups, 0.1lg of pRL-TK was added to each well. Twenty-four hoursafter transfection, the cells were harvested, and dual lu-ciferase reporter assays were performed by followingthe manufacturer’s protocol. Sense (S) and antisense(AS) oligonucleotides for both HNF-4a or PC4 were uti-lized to inhibit expression of these proteins. The histo-grams represent normalized luciferase activity. Thevalues are means ± SEM of 3 experiments. *P < .01 vsControl and H2O2; #P < .01 vs IL-1b and IL-1b+H2O2+Catalase.

catalase (100 mmol/L heme) was also added, asdetailed above. Sense (S) and antisense (AS)oligonucleotides for both HNF-4a or PC4 wereutilized to inhibit expression of these proteins. Inthis setting, treatment of wild-type cells with IL-1balone was associated with a >30-fold incease inluciferase activity, compared with unstimulatedControls or H2O2 alone (P < .01 vs Control andH2O2). In the presence of both IL-1b and H2O2,luciferase activity is increased by an additional 3-fold over that of IL-1b cells alone (P < .01 vs IL-1b alone). When catalase is added with IL-1b andperoxide, the additional luciferase activity noted inthe presence of this peroxide-mediated oxidativestress over IL-1b treatment alone is abolished. Thissame pattern holds for hepatocytes treated with S-HNF-4a and S-PC4. In contrast, in the presence ofAS-HNF-4a and/or AS-PC4, the incremental in-crease in luciferase activity in IL-1b+H2O2-treatedcells over that of IL-1b cells alone now is ablated.Despite this, the augmentation in luciferase activitynoted in IL-1b cells over unstimulated Controls ispreserved. These data suggest that both HNF-4aand PC4 are required for peroxide-mediatedupregulation of IL-1b–stimulated iNOS promoteractivity.

In vivo binding of an HNF-4a-PC4 complex. Toconfirm binding of an HNF-4a-PC4 complex to thedescribed region of the iNOS promoter in vivo,ChIP assays were performed with antibody toeither HNF-4a or PC4 (Fig 3, A). In unstimulatedControls, and IL-1b– and H2O2-treated cells, thereis no evidence for HNF-4a or PC4 binding to theregion of genomic DNA containing the iNOSpromoter. Only in the presence of IL-1b+H2O2

do HNF-4a and PC4 bind to the designated regionof DNA. No binding was noted in the presence ofrabbit immune serum or an irrelevant PCR primerset (data not shown). These results demonstratethat both HNF-4a and PC4 are bound to the sameregion of the iNOS promoter in the presence ofIL-1b+H2O2 stimulation. We then added AS-HNF-4a to these IL-1b+H2O2–treated cells. In this con-text in which HNF-4a expression is inhibited, noevidence exists for HNF-4a and PC4-DNA binding.To determine the extent of HNF-4a inhibition inthe presence of antisense oligonucleotides, immu-noblot analysis was also performed (Fig 3, B). Inthe presence of IL-1b+H2O2, AS-HNF-4a was asso-ciated with a 90% decrease in HNF-4a proteinexpression. Overall, these data suggest that PC4-DNA binding depends on the concomitant pres-ence of HNF-4a protein. PC4 is known to binddouble-stranded DNA (dsDNA) in a sequence-independent manner. Although PC4 possesses

SurgeryVolume 138, Number 1

Marroquin et al 97

both single-stranded DNA– and dsDNA-bindingactivities, which are important for transcriptionalrepression, only its dsDNA-binding activity appearsto correlate with the coactivator activity.17-19

DISCUSSION

Sepsis, shock, and reperfusion injury initiatehost responses characterized by proinflammatorycytokines and oxidative stress. This response mayultimately lead to organ dysfunction and multipleorgan failure, which remains a leading cause ofmorbidity and mortality in the critically ill surgicalpatient.20-24 Specifically, metabolic hepatic injuryafter sepsis or shock remains a poorly character-ized clinical problem. The high mortality rates inseptic patients with impaired hepatic amino aciduptake and protein synthesis underscore the sig-nificance of hepatic dysfunction in this setting.25-27

Multiple etiologies have been proposed, includingreactive oxygen species, alteration in microcircula-tory blood flow, neutrophil activation, and Kupffercell–induced hepatocyte dysfunction, but the rela-tive roles of these factors remain unknown. Simi-larly, cytoprotective mechanisms that act to inhibitor reverse sepsis- or shock-induced hepatic injuryare also poorly described. In this context, NO hasreceived a great deal of attention as a ubiquitous,multifunctional free radical produced duringshock and sepsis, which may function to eradicateinfection and limit tissue injury.28-30 The pervasive-ness of iNOS is emphasized by the finding that 33/33 human patients undergoing exploratory lapa-rotomy for trauma exhibited detectable iNOSmessenger RNA in the liver. (Timothy R. Billiar,MD, personal communication) However, the reg-ulatory pathways of hepatic NO production in themilieu of proinflammatory cytokines and reactiveoxygen species which characterize sepsis and shockremain poorly understood. 20-24,31-37

In this study, we used CCL9.1 murine hepato-cytes to demonstrate that (1) HNF-4a and PC4associate as a nuclear complex in IL-1b+H2O2–-

treated cells, (2) bothHNF-4a and PC4 are essentialelements in redox-mediated upregulation of IL-1b–stimulated iNOS transcription, (3) HNF-4a binds tothe iNOS promoter in vivo, and (4) PC4 binding tothe iNOS promoter in vivo requires HNF-4a. Intotal, these results indicate that redox-mediatedupregulation of hepatocyte iNOS transcription re-quires an HNF-4a-PC4 protein complex. As PC4does not bind to DNA and functions as a transcrip-tional coactivator, PC4must interact withHNF-4a tofulfill its coregulatory functions with the basaltranscriptional machinery.

Fig 3. A, Binding in vivo of HNF-4a and PC4 to iNOSpromoter. Chromatin was fixed and immunoprecipi-tated with the use of the ChIP assay kit as recommendedby the manufacturer (Upstate Biotechnology, Inc). Thepurified chromatin was immunoprecipitated with theuse of 10 lg of anti-PC4 or anti–HNF-4a, or 5 lL of rab-bit nonimmune serum. The input fraction correspondedto 0.1% and 0.05% of the chromatin solution before im-munoprecipitation. After DNA purification, presence ofthe selected DNA sequence was assessed by PCR. The gelrepresents 3 experiments. B, HNF-4a expression and an-tisense-HNF-4a. Immunoblot studies were performedwith nuclear protein from cells treated with IL-1b (1000 U/mL) and H2O2 (50 nmol/L) in the pres-ence or absence of antisense-HNF-4a (AS-HNF-4a). Blotsrepresent 3 experiments.

SurgeryJuly 2005

98 Marroquin et al

Transcription of messenger RNA--coding genesinvolves RNA polymerase II and 6 general tran-scription factors, which comprise the basal tran-scription machinery that recognizes the corepromoter elements and elicits the basal level oftranscription.38 Activated transcription requiresthe binding of activators to the regulatory DNAsequences typically present upstream of the corepromoter and their interactions with the generaltranscription machinery.39 In addition, activatedtranscription requires mediators or coactivators tomarkedly enhance responsiveness to activators.Coactivators may be grouped into 2 broad cate-gories according to the requirement of chromatinfor their action in biochemical assays. The coac-tivators that function on the templates withoutchromatin include the TATA box–binding protein-associated factors (TAFs) present in TFIID, positivecofactors (PCs), PC1, PC2, PC3, and PC4, derivedfrom the upstream factor stimulatory activity(USA) cofactor fraction, and metazoan multipro-tein complexes that are structurally related to theyeast mediator (TRAP/SMCC, ARC, DRIP, NAT,murine mediator, human mediator, CRSP, andPC2).40-42 The coactivators that require chromatintemplates for their functions include CBP/p300,PCAF and its related GCN5 proteins, and p160family proteins that display histone acetyltransfer-ase activities.43

PC4 is a 15-kDa polypeptide that serves as apotent coactivator in standard reconstitutedin vitro transcription systems. It mediates activa-tor-dependent transcription by RNA polymerase IIthrough interactions with the transcriptional acti-vator and basal transcription machinery. It issubject to in vivo phosphorylation events thatnegatively regulate its coactivator functions. Thevast majority (95%) of PC4 is phosphorylated andinactive in vivo. PC4 is proposed to promote theassembly of the preinitiation complex (PIC) inactivated transcription. However, given that tran-scription is a multistep process consisting of PICassembly, promoter opening, initiation, promoterescape, elongation, and reinitiation, steps otherthan PIC assembly are potential targets for regula-tion as well. Indeed, despite predominant effectsof activators---presumably in conjunction with co-activators---on PIC assembly, the effects on thesubsequent steps have also been demonstrated invarious systems, including promoter opening, pro-moter escape, elongation, and reinitiation.40-42

Interestingly, the 24 N-terminal residues ofHNF-4a (AF-1) constitute a critical structural ele-ment that has been demonstrated to bind toPC4.44 We have previously demonstrated that

HNF-4a binds with PC4 under conditions of IL-1b and superoxide stimulation.9 Cotransfection ofa mutant HNF-4a in which a critical PC4 bindingresidue has been substituted demonstrates abla-tion of redox-mediated iNOS promoter activation.It is unknown whether these stimulation condi-tions alter HNF-4 or PC4 to facilitate this interac-tion. However, because of the dependence of PC4activity on its phosphorylation status and the par-ticipation of various phosphatase activities in thecellular response to oxidative stress, it is temptingto speculate that PC4 may be a target. Thecoactivator activity of PC4 and its interaction withactivators, but not the single-stranded DNA–bind-ing activity, are lost on phosphorylation of theserine residues within its N-terminal region bycasein kinase II.44 Alternatively, IL-1b stimulationin the presence of oxidative stress may enhancebinding of PC4 to HNF-4a, expose or structurallyalter its DNA-binding domain, and enhance DNAbinding. These processes are currently the subjectof ongoing experiments in our laboratory.

REFERENCES

1. Powis G, Breihl M, Oblong J. Redox signalling and thecontrol of cell growth and death. Pharmacol Ther 1995;68:149-73.

2. Halliwell B. Antioxidant characterization. Biochem Phar-macol 1995;49:1341-8.

3. Harris ED. Regulation of antioxidant enzymes. FASEB J1992;6:2675-83.

4. Schulze-Osthoff K, Los M, Baeuerle PA. Redox signalling bytranscription factors NF-kB and AP-1 in lymphocytes.Biochem Pharmacol 1995;50:735-41.

5. Sen CK, Packer L. Antioxidant and redox regulation ofgene transcription. FASEB J 1996;10:709-20.

6. Jaiswal AK. Antioxidant response element. Biochem Phar-macol 1994;48:439-44.

7. Mayer B, Schrammel A, Klatt P, Koesling D, Schmidt K.Peroxynitrite induced accumulation of cyclic GMP inendothelial cells and stimulation of purified soluble gua-nylyl cyclase. J Biol Chem 1995;270:17355-60.

8. Sergent O, Griffon B, Morel I, Chevanne M, Dubos MP,Cillard P, Cillard J. Effect of nitric oxide on iron-mediatedoxidative stress in primary rat hepatocyte culture. Hepatol-ogy 1997;25:122-7.

9. Guo H, Cai CQ, Kuo PC. Hepatocyte nuclear factor-4alphamediates redox sensitivity of inducible nitric oxide synthasegene transcription. J Biol Chem 2002;277:5054-60.

10. Kuo PC, Abe K, Schroeder RA. Superoxide enhancesinterleukin 1beta-mediated transcription of the hepato-cyte-inducible nitric oxide synthase gene. Gastroenterology2000;118:608-18.

11. Kuo PC, Slivka A. Nitric oxide decreases oxidantmediated hepatocyte injury. Journal of Surg Res 1994;56:594-600.

12. Kuo PC, Abe KY. Interleukin-1 induced production of nitricoxide inhibits benzenetriol mediated oxidaitive injury in rathepatocytes. Gastroenterology 1995;109:206-16.

SurgeryVolume 138, Number 1

Marroquin et al 99

13. Punzalan C, Cai C, Schroeder RA, Kuo PC. Redox regula-tion of the rat hepatocyte iNOS promoter. Surgery 1999;126:450-5.

14. Guo H, Wei J, Inoue Y, Gonzalez FJ, Kuo PC. Serine/threonine phosphorylation regulates HNF-4alpha-depen-dent redox-mediated iNOS expression in hepatocytes. AmJ Physiol Cell Physiol 2003;284:C1090-9.

15. Lander HM. An essential role for free radicals and derivedspecies in signal transduction. FASEB J 1997;11:118-24.

16. Sambrook J, Fritsch EF, Maniatis T. Molecular cloning, alaboratory manual. New York: Cold Spring Harbor Press;1989.

17. Fukuda A, Nakadai T, Shimada M, et al. Transcriptionalcoactivator PC4 stimulates promoter escape and facilitatestranscriptional synergy by GAL4-VP16. Mol Cell Biol 2004;24:6525-35.

18. Banerjee S, Kumar BR, Kundu TK. General transcriptionalcoactivator PC4 activates p53 function. Mol Cell Biol 2004;24:2052-62.

19. Werten S, Stelzer G, Goppelt A, et al. Interaction of PC4with melted DNA inhibits transcription. EMBO J 1998;17:5103-11.

20. Abraham E, Bursten S, Shenkar R. Phosphatidic acidsignaling mediates lung cytokine expression and lunginflammatory injury after hemorrhage in mice. J ExperMed 1995;181:569-75.

21. DeLuca LG, Johnson DR, Whitley MZ. cAMP and tumornecrosis factor competitively regulate transcriptional activa-tion through and nuclear binding to the cAMP responsiveelement/activating transcription factor element of theendothelial leukocyte adhesion molecule promoter. J BiolChem 1994;269:19193-6.

22. Pizurki L, Bolla BS. cAMP modulates stress protein synthesisin human monocytes-macrophages. J Cell Physiol 1994;161:1669-77.

23. Liu PT, Symons AM, Howarth JA. Studies in surgical trauma:oxidative stress in ischemia-reperfusion of the rat liver. ClinSci 1994;86:453-60.

24. Schiller HJ, Reilly PM, Bulkley GB. Tissue perfusion incritical illnesses. Antioxidant therapy. Crit Care Med 1993;21:S92-102.

25. Rosenblatt S, Clowes G, George BC, Hirsch E, Lindberg B.Exchange of amino acids by muscle and liver in sepsis. ArchSurg 1983;118:167-75.

26. Sax HC, Talamini MA, Hasselgren PO, Rosenblum L, OgleCK. Increased synthesis of secreted hepatic proteins duringabdominal sepsis. J Surg Res 1988;44:109-16.

27. Clowes GHA, Hirsch E, George BC, Bigatello LM, MazuskiJE, Villee CA. Survival from sepsis: the significance ofaltered protein metabolism regulated by proteolysis induc-ing factor, the circulating cleavage factor of interleukin-1.Ann Surg 1985;202:446-58.

28. Nathan C. Nitric oxide as a secretory product of mamma-lian cells. FASEB J 1992;6:1157-75.

29. Kelly E, Morris SM, Billiar TR. Nitric oxide, sepsis, andarginine metabolism. JPEN 1995;19:234-8.

30. Liu S, Beckman JS, Ku DD. Peroxynitrite, a product ofsuperoxide and nitric oxide, produces coronary vasorelax-ation in dogs. J Pharmacol Exper Ther 1994;268:1114-21.

31. Kukreja RC, Kontos MC, Loesser KE. Oxidant stress in-creases heat shock protein 70 mRNA in isolated perfusedrat heart. Am J Physiol 1994;267:H2213-9.

32. Seekamp A, Warren JS, Remick DG. Requirements fortumor necrosis factor-a and interleukin-1 in limb ische-mia/reperfusion injury and associated lung injury. Am JPathol 1993;143:453-63.

33. Goris RJA, Boekhorst PA, Nuytink JKS. Multiple organfailure: generalized autodestructive inflammation? ArchSurg 1985;120:1109-15.

34. Moore EE, Moore FA, Franciose RJ. The postischemic gutserves as a priming bed for circulating neutrophils thatprovoke multiple organ failure. J Trauma 1994;37:881-7.

35. Liu P, McGuire GM, Fisher MA. Activation of Kupffer cellsand neutrophils for reactive oxygen formation is responsi-ble for endotoxin enhanced liver injury after hepaticischemia. Shock 1995;3:56-62.

36. Demling RH, LaLonde C, Daryani R. Relationship betweenthe lung and systemic response to endotoxin: comparisonof physiologic change and the degree of lipid peroxidation.Circulatory Shock 1991;30:97-106.

37. Machiedo GW, Powell RJ, Rush BF. The incidence ofdecreased red blood cell deformability in sepsis and theassociation with oxygen free radical damage and multiplesystem organ failure. Archives of Surgery 1989;124:1386-9.

38. Roeder RG. The role of general initiation factors in tran-scription by RNA polymerase II. Trends Biochem Sci 1996;21:327-35.

39. Lemon B, Tjian R. Orchestrated response: a symphony oftranscription factors for gene control. Genes Dev 2000;14:2551-69.

40. Kaiser K, Meisterernst M. The human general co-factors.Trends Biochem Sci 1996;21:342-5.

41. Malik S, Roeder RG. Transcriptional regulation throughMediator-like coactivators in yeast and metazoan cells.Trends Biochem Sci 2000;25:277-83.

42. Naar AM, Lemon BD, Tjian R. Transcriptional coactivatorcomplexes. Annu Rev Biochem 2001;70:475-501.

43. Brown CE, Lechner T, Howe L, Workman JL. The manyHATs of transcription coactivators. Trends Biochem Sci2000;25:15-9.

44. Green VJ, Kokkotou E, Ladias JA. Critical structural ele-ments and multitarget protein interactions of the transcrip-tional activator AF-1 of hepatocyte nuclear factor 4. J BiolChem 1998;273:29950-7.