Glutathione Reductase Promotes Fungal Clearance and Suppresses Inflammation during Systemic Candida albicans Infection in Mice.

Glutathione reductase (Gsr) catalyzes the reduction of glutathione disulfide to glutathione, which plays an important role in redox regulation. We have previously shown that Gsr facilitates neutrophil bactericidal activities and is pivotal for host defense against bacterial pathogens. However, it is unclear whether Gsr is required for immune defense against fungal pathogens. It is also unclear whether Gsr plays a role in immunological functions outside of neutrophils during immune defense. In this study, we report that Gsr-/- mice exhibited markedly increased susceptibility to Candida albicans challenge. Upon C. albicans infection, Gsr-/- mice exhibited dramatically increased fungal burden in the kidneys, cytokine and chemokine storm, striking neutrophil infiltration, histological abnormalities in both the kidneys and heart, and substantially elevated mortality. Large fungal foci surrounded by massive numbers of neutrophils were detected outside of the glomeruli in the kidneys of Gsr -/- mice but were not found in wild-type mice. Examination of the neutrophils and macrophages of Gsr-/- mice revealed several defects. Gsr -/- neutrophils exhibited compromised phagocytosis, attenuated respiratory burst, and impaired fungicidal activity in vitro. Moreover, upon C. albicans stimulation, Gsr -/- macrophages produced increased levels of inflammatory cytokines and exhibited elevated p38 and JNK activities, at least in part, because of lower MAPK phosphatase (Mkp)-1 activity and greater Syk activity. Thus, Gsr-mediated redox regulation is crucial for fungal clearance by neutrophils and the proper control of the inflammatory response by macrophages during host defense against fungal challenge.

GSR is not essential for the maintenance of antioxidant defenses in mouse cochlea: Possible role of the thioredoxin system as a functional backup for GSR.

Glutathione reductase (GSR), a key member of the glutathione antioxidant defense system, converts oxidized glutathione (GSSG) to reduced glutathione (GSH) and maintains the intracellular glutathione redox state to protect the cells from oxidative damage. Previous reports have shown that Gsr deficiency results in defects in host defense against bacterial infection, while diquat induces renal injury in Gsr hypomorphic mice. In flies, overexpression of GSR extended lifespan under hyperoxia. In the current study, we investigated the roles of GSR in cochlear antioxidant defense using Gsr homozygous knockout mice that were backcrossed onto the CBA/CaJ mouse strain, a normal-hearing strain that does not carry a specific Cdh23 mutation that causes progressive hair cell degeneration and early onset of hearing loss. Gsr-/- mice displayed a significant decrease in GSR activity and GSH/GSSG ratios in the cytosol of the inner ears. However, Gsr deficiency did not affect ABR (auditory brainstem response) hearing thresholds, wave I amplitudes or wave I latencies in young mice. No histological abnormalities were observed in the cochlea of Gsr-/- mice. Furthermore, there were no differences in the activities of cytosolic glutathione-related enzymes, including glutathione peroxidase and glutamate-cysteine ligase, or the levels of oxidative damage markers in the inner ears between WT and Gsr-/- mice. In contrast, Gsr deficiency resulted in increased activities of cytosolic thioredoxin and thioredoxin reductase in the inner ears. Therefore, under normal physiological conditions, GSR is not essential for the maintenance of antioxidant defenses in mouse cochlea. Given that the thioredoxin system is known to reduce GSSG to GSH in multiple species, our findings suggest that the thioredoxin system can support GSSG reduction in the mouse peripheral auditory system.

Extracellular superoxide provokes glutathione efflux from Escherichia coli cells.

The aim of the study was to elucidate a possible relationship between transmembrane cycling of glutathione and changes in levels of external superoxide. Exposure of growing Escherichia coli to exogenous reactive oxygen species (ROS) generated by xanthine and xanthine oxidase (XO) stimulates reversible glutathione (GSH) efflux from the cells that is considerably lowered under phosphate starvation. This GSH efflux is prevented by exogenous SOD, partially inhibited by catalase, and is not dependent on the GSH exporter CydDC. The γ-glutamyl transpeptidase (GGT) deficiency completely prevents a return of GSH to the cytoplasm. In contrast to wild-type E. coli, mutants devoid of GGT and glutathione reductase (GOR) show enhanced accumulation of oxidized glutathione in the medium after exposure to xanthine and XO. Under these conditions, sodC, ggt and especially gshA mutants reveal more intensive and prolonged inhibition of growth than wild-type cells. Treatment with XO does not influence E. coli viability, but somewhat increases the number of cells with lost membrane potential. In summary, data obtained here indicate that transmembrane cycling of GSH may be involved in E. coli protection against extracellular ROS and may promote rapid growth recovery.

S-nitrosoglutathione reductase-dependent PPARγ denitrosylation participates in MSC-derived adipogenesis and osteogenesis.

Bone marrow-derived mesenchymal stem cells (MSCs) are a common precursor of both adipocytes and osteoblasts. While it is appreciated that PPARγ regulates the balance between adipogenesis and osteogenesis, the roles of additional regulators of this process remain controversial. Here, we show that MSCs isolated from mice lacking S-nitrosoglutathione reductase, a denitrosylase that regulates protein S-nitrosylation, exhibited decreased adipogenesis and increased osteoblastogenesis compared with WT MSCs. Consistent with this cellular phenotype, S-nitrosoglutathione reductase-deficient mice were smaller, with reduced fat mass and increased bone formation that was accompanied by elevated bone resorption. WT and S-nitrosoglutathione reductase-deficient MSCs exhibited equivalent PPARγ expression; however, S-nitrosylation of PPARγ was elevated in S-nitrosoglutathione reductase-deficient MSCs, diminishing binding to its downstream target fatty acid-binding protein 4 (FABP4). We further identified Cys 139 of PPARγ as an S-nitrosylation site and demonstrated that S-nitrosylation of PPARγ inhibits its transcriptional activity, suggesting a feedback regulation of PPARγ transcriptional activity by NO-mediated S-nitrosylation. Together, these results reveal that S-nitrosoglutathione reductase-dependent modification of PPARγ alters the balance between adipocyte and osteoblast differentiation and provides checkpoint regulation of the lineage bifurcation of these 2 lineages. Moreover, these findings provide pathophysiological and therapeutic insights regarding MSC participation in adipogenesis and osteogenesis.

S-nitrosoglutathione reductase deficiency-induced S-nitrosylation results in neuromuscular dysfunction.

AIMS: Nitric oxide (NO) production is implicated in muscle contraction, growth and atrophy, and in the onset of neuropathy. However, many aspects of the mechanism of action of NO are not yet clarified, mainly regarding its role in muscle wasting. Notably, whether NO production-associated neuromuscular atrophy depends on tyrosine nitration or S-nitrosothiols (SNOs) formation is still a matter of debate. Here, we aim at assessing this issue by characterizing the neuromuscular phenotype of S-nitrosoglutathione reductase-null (GSNOR-KO) mice that maintain the capability to produce NO, but are unable to reduce SNOs. RESULTS: We demonstrate that, without any sign of protein nitration, young GSNOR-KO mice show neuromuscular atrophy due to loss of muscle mass, reduced fiber size, and neuropathic behavior. In particular, GSNOR-KO mice show a significant decrease in nerve axon number, with the myelin sheath appearing disorganized and reduced, leading to a dramatic development of a neuropathic phenotype. Mitochondria appear fragmented and depolarized in GSNOR-KO myofibers and myotubes, conditions that are reverted by N-acetylcysteine treatment. Nevertheless, although atrogene transcription is induced, and bulk autophagy activated, no removal of damaged mitochondria is observed. These events, alongside basal increase of apoptotic markers, contribute to persistence of a neuropathic and myopathic state. INNOVATION: Our study provides the first evidence that GSNOR deficiency, which affects exclusively SNOs reduction without altering nitrotyrosine levels, results in a clinically relevant neuromuscular phenotype. CONCLUSION: These findings provide novel insights into the involvement of GSNOR and S-nitrosylation in neuromuscular atrophy and neuropathic pain that are associated with pathological states; for example, diabetes and cancer.

Characterizing roles for the glutathione reductase, thioredoxin reductase and thioredoxin peroxidase-encoding genes of Magnaporthe oryzae during rice blast disease.

Understanding how pathogenic fungi adapt to host plant cells is of major concern to securing global food production. The hemibiotrophic rice blast fungus Magnaporthe oryzae, cause of the most serious disease of cultivated rice, colonizes leaf cells asymptomatically as a biotroph for 4-5 days in susceptible rice cultivars before entering its destructive necrotrophic phase. During the biotrophic growth stage, M. oryzae remains undetected in the plant while acquiring nutrients and growing cell-to-cell. Which fungal processes facilitate in planta growth and development are still being elucidated. Here, we used gene functional analysis to show how components of the NADPH-requiring glutathione and thioredoxin antioxidation systems of M. oryzae contribute to disease. Loss of glutathione reductase, thioredoxin reductase and thioredoxin peroxidase-encoding genes resulted in strains severely attenuated in their ability to grow in rice cells and that failed to produce spreading necrotic lesions on the leaf surface. Glutathione reductase, but not thioredoxin reductase or thioredoxin peroxidase, was shown to be required for neutralizing plant generated reactive oxygen species (ROS). The thioredoxin proteins, but not glutathione reductase, were shown to contribute to cell-wall integrity. Furthermore, glutathione and thioredoxin gene expression, under axenic growth conditions, was dependent on both the presence of glucose and the M. oryzae sugar/ NADPH sensor Tps1, thereby suggesting how glucose availability, NADPH production and antioxidation might be connected. Taken together, this work identifies components of the fungal glutathione and thioredoxin antioxidation systems as determinants of rice blast disease that act to facilitate biotrophic colonization of host cells by M. oryzae.

The glutathione reductase GSR-1 determines stress tolerance and longevity in Caenorhabditis elegans.

Glutathione (GSH) and GSH-dependent enzymes play a key role in cellular detoxification processes that enable organism to cope with various internal and environmental stressors. However, it is often not clear, which components of the complex GSH-metabolism are required for tolerance towards a certain stressor. To address this question, a small scale RNAi-screen was carried out in Caenorhabditis elegans where GSH-related genes were systematically knocked down and worms were subsequently analysed for their survival rate under sub-lethal concentrations of arsenite and the redox cycler juglone. While the knockdown of γ-glutamylcysteine synthetase led to a diminished survival rate under arsenite stress conditions, GSR-1 (glutathione reductase) was shown to be essential for survival under juglone stress conditions. gsr-1 is the sole GSR encoding gene found in C. elegans. Knockdown of GSR-1 hardly affected total glutathione levels nor reduced glutathione/glutathione disulphide (GSH/GSSG) ratio under normal laboratory conditions. Nevertheless, when GSSG recycling was impaired by gsr-1(RNAi), GSH synthesis was induced, but not vice versa. Moreover, the impact of GSSG recycling was potentiated under oxidative stress conditions, explaining the enormous effect gsr-1(RNAi) knockdown had on juglone tolerance. Accordingly, overexpression of GSR-1 was capable of increasing stress tolerance. Furthermore, expression levels of SKN-1-regulated GSR-1 also affected life span of C. elegans, emphasising the crucial role the GSH redox state plays in both processes.

Dynamic denitrosylation via S-nitrosoglutathione reductase regulates cardiovascular function.

Although protein S-nitrosylation is increasingly recognized as mediating nitric oxide (NO) signaling, roles for protein denitrosylation in physiology remain unknown. Here, we show that S-nitrosoglutathione reductase (GSNOR), an enzyme that governs levels of S-nitrosylation by promoting protein denitrosylation, regulates both peripheral vascular tone and ß-adrenergic agonist-stimulated cardiac contractility, previously ascribed exclusively to NO/cGMP. GSNOR-deficient mice exhibited reduced peripheral vascular tone and depressed ß-adrenergic inotropic responses that were associated with impaired ß-agonist-induced denitrosylation of cardiac ryanodine receptor 2 (RyR2), resulting in calcium leak. These results indicate that systemic hemodynamic responses (vascular tone and cardiac contractility), both under basal conditions and after adrenergic activation, are regulated through concerted actions of NO synthase/GSNOR and that aberrant denitrosylation impairs cardiovascular function. Our findings support the notion that dynamic S-nitrosylation/denitrosylation reactions are essential in cardiovascular regulation.

Glutathione reductase facilitates host defense by sustaining phagocytic oxidative burst and promoting the development of neutrophil extracellular traps.

Glutathione reductase (Gsr) catalyzes the reduction of glutathione disulfide to glutathione, which plays an important role in the bactericidal function of phagocytes. Because Gsr has been implicated in the oxidative burst in human neutrophils and is abundantly expressed in the lymphoid system, we hypothesized that Gsr-deficient mice would exhibit marked defects during the immune response against bacterial challenge. We report in this study that Gsr-null mice exhibited enhanced susceptibility to Escherichia coli challenge, indicated by dramatically increased bacterial burden, cytokine storm, striking histological abnormalities, and substantially elevated mortality. Additionally, Gsr-null mice exhibited elevated sensitivity to Staphylococcus aureus. Examination of the bactericidal functions of the neutrophils from Gsr-deficient mice in vitro revealed impaired phagocytosis and defective bacterial killing activities. Although Gsr catalyzes the regeneration of glutathione, a major cellular antioxidant, Gsr-deficient neutrophils paradoxically produced far less reactive oxygen species upon activation both ex vivo and in vivo. Unlike wild-type neutrophils that exhibited a sustained oxidative burst upon stimulation with phorbol ester and fMLP, Gsr-deficient neutrophils displayed a very transient oxidative burst that abruptly ceased shortly after stimulation. Likewise, Gsr-deficient neutrophils also exhibited an attenuated oxidative burst upon encountering E. coli. Biochemical analysis revealed that the hexose monophosphate shunt was compromised in Gsr-deficient neutrophils. Moreover, Gsr-deficient neutrophils displayed a marked impairment in the formation of neutrophil extracellular traps, a bactericidal mechanism that operates after neutrophil death. Thus, Gsr-mediated redox regulation is crucial for bacterial clearance during host defense against massive bacterial challenge.

Inherited glutathione reductase deficiency and Plasmodium falciparum malaria--a case study.

In Plasmodium falciparum-infected red blood cells (RBCs), the flavoenzyme glutathione reductase (GR) regenerates reduced glutathione, which is essential for antioxidant defense. GR utilizes NADPH produced in the pentose phosphate shunt by glucose-6-phosphate dehydrogenase (G6PD). Thus, conditions affecting host G6PD or GR induce increased sensitivity to oxidants. Hereditary G6PD deficiency is frequent in malaria endemic areas and provides protection against severe malaria. Furthermore, GR deficiency resulting from insufficient saturation of the enzyme with its prosthetic group FAD is common. Based on these naturally occurring phenomena, GR of malaria parasites and their host cells represent attractive antimalarial drug targets. Recently we were given the opportunity to examine invasion, growth, and drug sensitivity of three P. falciparum strains (3D7, K1, and Palo Alto) in the RBCs from three homozygous individuals with total GR deficiency resulting from mutations in the apoprotein. Invasion or growth in the GR-deficient RBCs was not impaired for any of the parasite strains tested. Drug sensitivity to chloroquine, artemisinin, and methylene blue was comparable to parasites grown in GR-sufficient RBCs and sensitivity towards paraquat and sodium nitroprusside was only slightly enhanced. In contrast, membrane deposition of hemichromes as well as the opsonizing complement C3b fragments and phagocytosis were strongly increased in ring-infected RBCs of the GR-deficient individuals compared to ring-infected normal RBCs. Also, in one of the individuals, membrane-bound autologous IgGs were significantly enhanced. Thus, based on our in vitro data, GR deficiency and drug-induced GR inhibition may protect from malaria by inducing enhanced ring stage phagocytosis rather than by impairing parasite growth directly.

Endogenous S-nitrosothiols protect against myocardial injury.

Despite substantial evidence that nitric oxide (NO) and/or endogenous S-nitrosothiols (SNOs) exert protective effects in a variety of cardiovascular diseases, the molecular details are largely unknown. Here we show that following left coronary artery ligation, mice with a targeted deletion of the S-nitrosoglutathione reductase gene (GSNOR(-/-)) have reduced myocardial infarct size, preserved ventricular systolic and diastolic function, and maintained tissue oxygenation. These profound physiological effects are associated with increases in myocardial capillary density and S-nitrosylation of the transcription factor hypoxia inducible factor-1alpha (HIF-1alpha) under normoxic conditions. We further show that S-nitrosylated HIF-1alpha binds to the vascular endothelial growth factor (VEGF) gene, thus identifying a role for GSNO in angiogenesis and myocardial protection. These results suggest innovative approaches to modulate angiogenesis and preserve cardiac function.

A genetic analysis of nitrosative stress.

Nitrosative stress is induced by pathophysiological levels of nitric oxide (NO) and S-nitrosothiols (e.g., S-nitrosoglutathione, GSNO) and arises, at least in significant part, from the nitrosylation of critical protein Cys thiols (S-nitrosylation) and metallocofactors. However, the mechanisms by which NO and GSNO mediate nitrosative stress are not well understood. Using yeast Saccharomyces cerevisiae strains lacking NO- and/or GSNO-consuming enzymes (flavohemoglobin and GSNO reductase, respectively), we measured the individual and combined effects of NO and GSNO on both cell growth and the formation of protein-bound NO species. Our results suggest an intracellular equilibrium between NO and GSNO, dependent in part on cell-catalyzed release of NO from GSNO (i.e., "SNO-lyase" activity). However, whereas NO induces multiple types of protein-based modifications, levels of which correlate with inhibition of cell growth, GSNO mainly affects protein S-nitrosylation, and the relationship between S-nitrosylation and nitrosative stress is more complex. These data support the idea of multiple classes of protein-SNO, likely reflected in divergent routes of synthesis and degradation. Indeed, a significant fraction of protein S-nitrosylation by NO occurs in the absence of O(2), which is commonly assumed to drive this reaction but instead is apparently dependent in substantial part upon protein-bound transition metals. Additionally, our findings suggest that nitrosative stress is mediated principally via the S-nitrosylation of a subset of protein targets, which include protein SNOs that are stable to cellular glutathione (and thus are not metabolized by GSNO reductase). Collectively, these results provide new evidence for the mechanisms through which NO and GSNO mediate nitrosative stress as well as the cellular pathways of protein S-nitrosylation and denitrosylation involving metalloproteins, SNO lyase(s) and GSNO reductase.

Evidence for an additional disulfide reduction pathway in Escherichia coli.

An Escherichia coli cell-free protein synthesis cell extract has been created that lacks all known cytoplasmic disulfide reduction pathways but still retains significant reductase activity. Oxidized glutathione was partially stabilized by deleting the gene for glutathione reductase. To avoid previously reported AhpC mutations, thioredoxin reductase was only removed after extract preparation. The trxB gene was extended to encode a hemagglutinin tag so that TrxB could be removed by affinity adsorption. However, significant glutathione reductase activity remained. The unknown glutathione reductase pathway is disabled by iodoacetamide, is inhibited by NADH, and appears to use NADPH as an electron source.

Molecular basis of glutathione reductase deficiency in human blood cells.

Hereditary glutathione reductase (GR) deficiency was found in only 2 cases when testing more than 15 000 blood samples. We have investigated the blood cells of 2 patients (1a and 1b) in a previously described family suffering from favism and cataract and of a novel patient (2) presenting with severe neonatal jaundice. Red blood cells and leukocytes of the patients in family 1 did not contain any GR activity, and the GR protein was undetectable by Western blotting. Owing to a 2246-bp deletion in the patients' DNA, translated GR is expected to lack almost the complete dimerization domain, which results in unstable and inactive enzyme. The red blood cells from patient 2 did not exhibit GR activity either, but the patient's leukocytes contained some residual activity that correlated with a weak protein expression. Patient 2 was found to be a compound heterozygote, with a premature stop codon on one allele and a substitution of glycine 330, a highly conserved residue in the superfamily of NAD(P)H-dependent disulfide reductases, into alanine on the other allele. Studies on recombinant GR G330A revealed a drastically impaired thermostability of the protein. This is the first identification of mutations in the GR gene causing clinical GR deficiency.

Diquat induces renal proximal tubule injury in glutathione reductase-deficient mice.

Reactive oxygen species (ROS) have been associated with many human diseases, and glutathione (GSH)-dependent processes are pivotal in limiting tissue damage. To test the hypothesis that Gr1(a1Neu) (Neu) mice, which do not express glutathione reductase (GR), would be more susceptible than are wild-type mice to ROS-mediated injury, we studied the effects of diquat, a redox cycling toxicant. Neu mice exhibited modest, dose- and time-dependent elevations in plasma alanine aminotransferase (ALT) activities, 126+/-36 U/l at 2 h after 5 micromol/kg of diquat, but no ALT elevations were observed in diquat-treated C3H/HeN mice for up to 6 h after 50 micromol/kg of diquat. Histology indicated little or no hepatic necrosis in diquat-treated mice of either strain, but substantial renal injury was observed in diquat-treated Neu mice, characterized by brush border sloughing in the proximal tubules by 1 h and tubular necrosis by 2 h after doses of 7.5 micromol/kg. Decreases in renal GSH levels were observed in the Neu mice by 2 h post dose (3.4+/-0.4 vs 0.2+/-0.0 micromol/g tissue at 0 and 50 micromol/kg, respectively), and increases in renal GSSG levels were observed in the Neu mice as early as 0.5 h after 7.5 micromol/kg (105.5+/-44.1 vs 27.9+/-4.8 nmol/g tissue). Blood urea nitrogen levels were elevated by 2 h in Neu mice after doses of 7.5 micromol/kg (Neu vs C3H, 32.8+/-4.1 vs 17.9+/-0.3 mg/dl). Diquat-induced renal injury in the GR-deficient Neu mice offers a useful model for studies of ROS-induced renal necrosis and of the contributions of GR in defense against oxidant-mediated injuries in vivo.

Glutathione reductase activity deficiency in homozygous Gr1a1Neu mice does not cause haemolytic anaemia.

A glutathione reductase (GR) mutant with approximately 50% residual enzyme activity in blood compared with wild-type was detected amongst offspring of isopropyl methanesulphonate-treated male mice. Homozygous mutants with only 2% residual enzyme activity were recovered in progeny of inter se matings of heterozygotes. Results of linkage studies indicate a mutation at the Gr1 structural locus on chromosome 8. The loss of GR activity was evident both in blood and in other tissue extracts. Erythrocyte and organo-somatic indices did not show differences between wild-types and homozygous mutants, indicating no association between the GR deficiency and haemolytic anaemia in this potential animal model.

Glutathione reductase deficiency in Saudi Arabia.

Glutathione reductase (GR) is a ubiquitous enzyme required for the conversion of oxidized glutathione (GSSG) to reduced glutathione (GSH) concomitantly oxidizing reduced nicotinamide adenine dinucleotide phosphate (NADPH) in a reaction essential for the stability and integrity of red cells. Mutations in the GR gene and nutritional deficiency of riboflavin, a co-factor required for the normal functioning of GR, can cause GR deficiency. We conducted a study on 1691 Saudi individuals to determine the overall frequency of GR deficiency and to identify whether the deficiency results from genetic or acquired causes or both. The activity of GR was measured in freshly prepared red cell haemolysate in the presence and absence of flavin adenine dinucleotide (FAD) and the activity coefficient (AC) was determined. Samples with low GR activity (> 2.0 IU/g haemoglobin) both in the presence and absence of FAD and an AC between 0.9 and 1.2 were considered GR-deficient. Samples with AC > or = 1.3 were considered riboflavin-deficient. The overall frequency of partial GR deficiency was 24.5% and 20.3% in males and females respectively. In addition, 17.8% of males and 22.4% of females suffered from GR deficiency due to riboflavin deficiency. This could be easily corrected by dietary supplementation with riboflavin. No cases of severe GR deficiency were identified.

[Hemolytic anemia due to the dysfunction of the protection against oxidative attack].

The most important products of the hexose monophosphate pathway is reduced nicotinamide adenine dinucleotide phosphate (NADPH). Reduced glutathione (GSH) maintained by the reduction of oxidized glutathione (GSSG) using NADPH as a cofactor, is a major reducing agent in the red cell and the ultimate source of protection against oxidative attack. In the syndromes associated with dysfunction of the hexose monophosphate pathway and glutathione synthesis and metabolism, oxidative denaturation of hemoglobin is the major contributor to the hemolytic process. Glucose-6-phosphate dehydrogenase (G6PD) plays a key role in the generation of NADPH. G6PD deficiency is the most common metabolic disorder, and it is associated with chronic hemolytic anemia and/or drug- or infection-induced acute hemolytic attack. It is estimated that 400 million people are affected worldwide. The mutations responsible for 101 variants have been determined. Some of them have polymorphic frequencies in different populations. Most variants are produced by one or two nucleotide substitutions. Molecular studies have disclosed that most of the class 1 G6PD variants associated with chronic hemolysis have the mutations surrounding the site of dimer formation.