Heme delivery into soluble guanylyl cyclase requires a heme redox change and is regulated by NO and Hsp90 by distinct mechanisms

Abstract

Nitric oxide (NO) signaling often relies on it activating cGMP production by the heterodimeric enzyme soluble guanylyl cyclase (sGC). To mature to function, an sGCβ subunit must first incorporate heme and then form a heterodimer with a partner α subunit. Our previous studies in cells showed that glyceraldehyde 3-phosphate dehydrogenase (GAPDH) supplies heme to the apo-sGCβ subunit, which is complexed with the cell chaperone Hsp90. Through its ATP hydrolysis, Hsp90 then promotes heme insertion into apo-sGCβ and consequent formation of a functional heterodimer. NO at physiologic levels somehow stimulates cell heme allocation into apo-sGCβ by this process. To gain insight, we utilized purified apo-sGCβ and GAPDH reporter proteins whose heme contents can be followed by fluorescence and determined the impact of Hsp90 and NO on heme transfer between them. Results show that heme transfer out of GAPDH and into apo-sGCβ is tightly coupled in all circumstances and is limited by the ability of the apo-sGCβ to incorporate the heme, which in turn relies on a ferric to ferrous heme transition taking place inside the sGCβ. Hsp90 can influence the heme transfer kinetics in a negative or positive manner through its conformational effects on apo-sGCβ, while NO speeds heme transfer by binding to the heme iron and thus speeding heme dissociation from GAPDH. Our findings provide new mechanistic understanding of sGC maturation and how Hsp90 and NO combine to dynamically regulate heme incorporation for sGC heterodimer formation and consequent cGMP production in biological settings.

Keywords: GAPDH; guanylyl cyclase; heat shock protein 90; heme; heme transfer; nitric oxide.

Figure 1

Monitoring heme binding in TC-sGCβ and TC-GAPDH. The fluorescence intensity of either FlAsH-labeled protein is inversely proportional to its heme content, allowing changes in their heme contents to be followed versus time.

Figure 2

FlAsH fluorescence traces indicating the kinetics of ferric or ferrous heme transfer from GAPDH and into apo-sGCβ. Reactions contained 1 μM each of the FlAsH-labeled and unlabeled partner proteins and were initiated by adding the unlabeled partner protein. Representative traces are the mean ± SD of three replicate reactions and show the gain or loss in fluorescence intensity versus time that indicate the loss of ferric heme (A) or ferrous heme (B) from FlAsH-TC-GAPDH or their incorporation into FlAsH-TC-apo-sGCβ during the reactions, respectively. Red lines are the fitted curves.

Figure 3

The impact of Hsp90 and its ATPase activity on heme transfer from GAPDH to apo-sGCβ. Reactions contained 1 μM each of unlabeled or FlAsH-labeled versions of the TC-GAPDH-heme complex or TC-apo-sGCβ either alone or in complex with Hsp90. Reactions were initiated by adding the unlabeled partner protein. In some cases, ATP was present or a variant of Hsp90 (D88N) or TC-apo-sGCβ (HD) were used as indicated. (A and C) contain representative fluorescence traces (mean ± SD of three replicates) that indicate (A) the kinetics and extent of heme gain into the indicated FlAsH-labeled TC-apo-sGCβ proteins or (C) heme loss from FlAsH-TC-GAPDH in companion reactions. The relative rates of heme gain and heme loss under the various reaction conditions are compared in (B and D), respectively. Data points in (B and D) are derived from three independent trials, and the changes observed in each trial were normalized by assigning a value of 100 to the mean rate (from n = 3 replicates) obtained for heme transfer in the simple two-component reaction run in each trial. Inserts show the data using a smaller y-axis scale. ∗p ≤ 0.05, significantly different as indicated. HD, heme binding deficient; ns, nonsignificant.

Figure 4

Effect of NO on GAPDH heme transfer into FlAsH-TC-apo-sGCβ. Reactions contained 1 μM FlAsH-labeled TC-apo-sGCβ alone or in complex with Hsp90 and were initiated by adding the GAPDH–heme complex to give a final concentration of 1 μM. In some reactions, NOC18, ATP, or and/or radicicol were present or a variant of Hsp90 (D88N) was used as indicated. A and C, representative fluorescence traces versus time (mean ± SD of three replicates) indicating the kinetics and extent of heme gain into FlAsH-TC-apo-sGCβ under the different reaction conditions. B, the rates of heme incorporation for each reaction condition as derived by fitting the traces in replica reactions to a single exponential equation. ∗p ≤ 0.05, significantly different as indicated.

Figure 5

Effect of NO on heme transfer from FlAsH-TC-GAPDH to apo-sGCβ. Reactions contained 1 μM of FlAsH-TC-GAPDH-heme complex and were initiated by adding apo-sGCβ either alone or in complex with Hsp90. In some reactions, NOC18 and/or ATP was present as indicated. A, representative fluorescence traces (mean ± SD of three replicates) indicating the kinetics and extent of heme loss from FlAsH-TC-GAPDH under the different reaction conditions. B, the relative rates of heme incorporation for each reaction condition as derived by fitting the traces in replica reactions to a single exponential equation. ∗p ≤ 0.05, significantly different as indicated.

Figure 6

UV-visible difference spectra indicating ferric heme becomes reduced to ferrous during its transfer from GAPDH into apo-sGCβ. Reactions contained 2.5 μM apo-sGCβ either alone (A and C) or with Hsp90 and ATP (B and D) and were initiated by adding the GAPDH-ferric heme complex (final concentration 2.5 μM GAPDH tetramer containing 1.2 μM heme) and then were run for 30 min in the absence (A and B) or presence (C and D) of 100 μM NOC18 at RT. After this 30 min reaction, the samples received ODQ (10 μM), and the reactions were run for a further 30 min. Traces shown are the difference spectra calculated by subtracting the UV-visible spectra recorded at reaction time = 0 from the spectra recorded for the 30 min reaction samples both before (black traces) and after (red traces) the additional 30 min incubation with ODQ. Traces shown are representative of three independent trials.

Figure 7

Having ODQ present during GAPDH heme transfer to apo-sGCβ mutes the spectroscopic changes associated with heme reduction. Reaction solutions that contained 2.5 μM apo-sGCβ either alone (A and B) or with Hsp90 and ATP (C and D), without (A and C) or with (B and D) 100 μM NOC18 and with or without 10 μM ODQ as indicated, had GAPDH-ferric heme complex added (final concentration 2.5 μM GAPDH tetramer containing 1.2 μM heme) to initiate the heme transfer reactions which were then run for 30 min while monitoring UV-visible absorbance at 427 nm (A and C) or 390 nm (B and D). Absorbance traces shown are the mean ± SD of three replicates and representative of two independent trials.

Figure 8

Having ODQ present diminishes the transfer of ferric heme from GAPDH into FlAsH-TC-apo-sGCβ. Reaction solutions that contained 1 μM FlAsH-TC-apo-sGCβ alone (two component) or with Hsp90 and ATP (3 component), with or without 100 μM NOC18 and with or without 10 μM ODQ, had either a GAPDH-ferric heme complex (A–D) or a GAPDH ferrous heme-NO complex (PanelsE and F) added to give 1 μM final concentration to start the transfer reactions. The fluorescence emission was then monitored at RT for 30 min. The traces shown (mean ± SD, three replicates) indicate the kinetics and extent of heme incorporation into FlAsH-TC-apo-sGCβ in the reactions and are representative of two independent trials.

Figure 9

Effect of CO on the transfer of GAPDH ferrous versus ferric heme into FlAsH-TC-apo-sGCβ. Reactions were run either in air or under an N₂ atmosphere and were initiated by adding a GAPDH-ferric or GAPDH-ferrous heme complex, respectively, to a FlAsH-labeled TC-apo-sGCβ that was in complex with Hsp90 in the presence of ATP. Some reactions also contained 100 μM NOC18 and/or 500 μM CORM-A (CO donor) as indicated. A and B, representative fluorescence traces (mean ± SD of three replicates) indicating the kinetics and extent of ferric or ferrous heme transfer into FlAsH-labeled TC-apo-sGCβ under the indicated reaction conditions. Results are representative of two or three independent trials.

Figure 10

NO increases the rate of heme dissociation from GAPDH via ligation to the heme iron. Reactions were initiated by mixing either a FlAsH-TC-GAPDH ferric heme complex or a FlAsH-TC-GAPDH ferrous heme-NO complex, with a 30-fold molar excess of GAPDH. In some cases, 100 μM NOC18 or 3 μM miconazole (a ferric iron heme ligand) were also present in the reactions as indicated. A and B, representative fluorescence traces (mean ± SD, three replicates) indicating the kinetices of heme dissociation from FlAsH-labeled TC-GAPDH under indicated conditions. The fluorescence traces were fit to a single exponential equation (green line) to obtain rates and are representative of two or three independent trials.

Figure 11

Ferric heme transfer from GAPDH into apo-sGCβ is regulated by a heme redox transition, NO, and Hsp90/ATP via distinct mechanisms. The transfer of ferric heme from GAPDH into apo-sGCβ is tightly coupled regarding its rate and extent, relies on protein–protein contact, and results in reduction of the ferric heme to ferrous inside apo-sGCβ. A, heme transfer occurs to a minor extent in the absence of heme reduction. B, the heme redox transition greatly drives the equilibrium toward the right and increases the observed extent and rate of the heme transfer. In this circumstance, NO and Hsp90 can speed heme transfer by two different mechanisms: NO acts by binding to the ferric heme iron in GAPDH to increase its rate of release, while Hsp90 speeds heme intake into the apo-sGCβ by causing protein conformational changes that are driven by its ATP hydrolysis. In all cases, the extent and rate of heme transfer from GAPDH is determined by the ability of the apo-sGCβ to accept the heme.