NO dioxygenase activity in hemoglobins is ubiquitous in vitro, but limited by reduction in vivo

Abstract

Genomics has produced hundreds of new hemoglobin sequences with examples in nearly every living organism. Structural and biochemical characterizations of many recombinant proteins reveal reactions, like oxygen binding and NO dioxygenation, that appear general to the hemoglobin superfamily regardless of whether they are related to physiological function. Despite considerable attention to "hexacoordinate" hemoglobins, which are found in nearly every plant and animal, no clear physiological role(s) has been assigned to them in any species. One popular and relevant hypothesis for their function is protection against NO. Here we have tested a comprehensive representation of hexacoordinate hemoglobins from plants (rice hemoglobin), animals (neuroglobin and cytoglobin), and bacteria (Synechocystis hemoglobin) for their abilities to scavenge NO compared to myoglobin. Our experiments include in vitro comparisons of NO dioxygenation, ferric NO binding, NO-induced reduction, NO scavenging with an artificial reduction system, and the ability to substitute for a known NO scavenger (flavohemoglobin) in E. coli. We conclude that none of these tests reveal any distinguishing predisposition toward a role in NO scavenging for the hxHbs, but that any hemoglobin could likely serve this role in the presence of a mechanism for heme iron re-reduction. Hence, future research to test the role of Hbs in NO scavenging would benefit more from the identification of cognate reductases than from in vitro analysis of NO and O(2) binding.

Figure 1. NO reactions with hemoglobins.

1, NO dioxygenase activity; 2, NO binding to the ferric Hbs; 3, NO-induced heme iron reduction; 4, O₂ nitroxylase activity; 5, reduction reaction of ferric Hbs (4).

Figure 2. NO dioxygenation and stoichiometric NO consumption.

Plots of k_obs,NOD vs [NO] for the NOD reaction following rapid mixing.

Figure 3. NO binding to ferric Mb and hxHbs.

Plots of k_obs versus [NO] for Mb and each hxHb. Time courses giving rise to these values were measured at different [NO] ranging from 50 to 1700 µM (after mixing) and were fitted to a single exponential to extract the observed rate constants (k_obs). A linear fit of these data provides the observed ferric NO binding association rate constant (k_{obs, NO(Fe3+)}).

Figure 4. Absorbance spectra associated with NO-induced reduction.

Panels (A–E) show the Hb(3+) (thick solid line), Hb(3+)-NO (dotted line), Hb(2+)-NO (thin dotted line) absorbance spectra, and the spectrum of the sample 30 minutes after mixing NO with ferric protein (dashed line). Ferric and ferrous oxidation states are indicated by 3+ and 2+, respectively. F. Percentage of Hb(2+)-NO after 30 minutes of reduction.

Figure 5. Enzymatic reduction by ferredoxin-NADP reductase and catalytic NO consumption

. A. Absorption spectra associated with the reduction of Ngb (10 µM) by 1 µM ferredoxin-NADP reductase in the presence of CO (1 mM). B. Change in absorbance at the Soret peak (CO-bound) associated with the reduction is plotted vs time for each Hb (10 µM). The data are normalized to the absorbance change expected for complete reduction of the Hb in question. C. Enzymatic reduction of Mb and hxHbs by ferredoxin-NADP reductase. V/[E_T] calculated at different Hb concentration is plotted vs [Hb]. The fit to the Michaelis-Menten equation gives Km and Vmax for reduction of each protein. D. Consumption of 40 µM NO by 10 µM oxyHb as measured by an NO electrode, in the presence of ferredoxin-NADP reductase (1 µM). For each protein, after addition of NO, the signal drops by ∼10 µM corresponding to stoichiometric NOD. [NO] then decreases linearly indicating catalytic NO destruction. Rates of consumption were calculated from the linear phase of catalytic NO removal.

Figure 6. Protection of the E. coli hmp mutant from NO by Mb and hxHbs.

A. The growth (OD₆₀₀) of wild type (AB1157) and the hmp mutant (AB1000) E coli strains carrying Mb and hxHbs under control of the flavoHb promoter were measured after 14h in the absence and presence of 3 mM GSNO. Photolyzed GSNO was also used as control. Only flavoHb (hmp) and SynHb are able to rescue the phenotype observed for the wild type strain. The data are an average of the duplication of at least three independent experiments. B. Hemoglobin expression levels driven by the hmp promoter. The values reported are the averaged maximum absorbance of the CO-difference spectra peak. Data are an average of two independent experiments. C. NO consumption by different E. coli strains. Compared to flavoHb, the consumption by other Hbs is significantly slower. Data are an average of two replications of six independent experiments. D. Growth curves of AG1000 strains transformed with pANX and expressing flavoHb and SynHb (average of two independent experiments). In the absence of GSNO, no differences are observed between strains. In the presence of GSNO, there is no growth in the strain transformed with pANX alone. With flavoHb, the growth is similar to the one without GSNO. In presence of SynHb, growth occurs but is significantly retarded.