Indoleamine dioxygenase and tryptophan dioxygenase activities are regulated through control of cell heme allocation by nitric oxide

Indoleamine-2, 3-dioxygenase (IDO1) and Tryptophan-2, 3-dioxygenase (TDO) catalyze the conversion of L-tryptophan to N-formyl-kynurenine and thus play primary roles in metabolism, inflammation, and tumor immune surveillance. Because their activities depend on their heme contents, which vary in biological settings and go up or down in a dynamic manner, we studied how their heme levels may be impacted by nitric oxide (NO) in mammalian cells. We utilized cells expressing TDO or IDO1 either naturally or via transfection and determined their activities, heme contents, and expression levels as a function of NO exposure. We found NO has a bimodal effect: a narrow range of low NO exposure promoted cells to allocate heme into the heme-free TDO and IDO1 populations and consequently boosted their heme contents and activities 4- to 6-fold, while beyond this range the NO exposure transitioned to have a negative impact on their heme contents and activities. NO did not alter dioxygenase protein expression levels, and its bimodal impact was observed when NO was released by a chemical donor or was generated naturally by immune-stimulated macrophage cells. NO-driven heme allocations to IDO1 and TDO required participation of a GAPDH–heme complex and for IDO1 required chaperone Hsp90 activity. Thus, cells can up- or downregulate their IDO1 and TDO activities through a bimodal control of heme allocation by NO. This mechanism has important biomedical implications and helps explain why the IDO1 and TDO activities in animals go up and down in response to immune stimulation.

Indoleamine-2, 3-dioxygenase (EC 1.13.11.52; IDO1) and Tryptophan-2, 3-dioxygenase (EC 1.13.11.11; TDO) are heme proteins that catalyze the conversion of tryptophan (Trp) to Nformyl-kynurenine (1,2). The activities of these enzymes directly depend on the content of their ferrous heme, which binds dioxygen to enable its insertion into the indole ring of Trp (3,4). TDO is selective for L-Trp as a substrate, while IDO1 accepts a broader range of indole-containing molecules (5,6). Although IDO1 has a monomeric structure and TDO is tetrameric (7), their crystal structures indicate there are similarities in their heme binding environments and substrate-binding sites (8), consistent with their having similar catalytic mechanisms (9,10). TDO expression and activity levels are highest in the liver and upregulated by glucocorticoids and L-Trp (11,12). In comparison, IDO1 is constitutively expressed in tissues such as the lung and its expression can be broadly induced by stimuli associated with immune activation and inflammation, including interferon-γ, lipopolysaccharide (LPS), and tumor necrosis factor (13)(14)(15). Increased IDO1 or TDO activities cause Trp depletion and increased production of Kynurenine (Kyn) and its metabolites, which are bioactive (16)(17)(18)(19). In this way, increased IDO1 activity helps downregulate several inflammatory diseases (20,21), but on the other hand, it is also associated with neurologic disorders (22,23), and in cancer cells or in host dendritic cells is associated with suppression of effector T-cell responses toward tumors (24,25). Thus, careful control of the IDO1 and TDO activities is needed and they are targets for pharmacologic intervention (26)(27)(28).
We recently reported that maturation of functional IDO1 and TDO requires a GAPDH-dependent heme delivery and in the case of IDO1 also requires heat shock protein 90 (Hsp90) to drive its heme insertion (29). Moreover, we found that in cells under normal growth conditions both IDO1 and TDO were about 40% heme saturated (29), consistent with earlier reports showing that 30 to 60% of rodent liver TDO normally exists in a heme-free state in healthy animals (30)(31)(32) and that IDO1 is predominantly expressed in its heme-free form in cells (33,34). Older studies also showed that the heme saturation levels of TDO and IDO1 could be dynamically altered: the heme level of TDO could be increased by giving the animals Trp (35,36) or by boosting their heme biosynthesis (31,32). Moreover, these changes could be caused by immune stimulation: rats injected with the immune stimulant Escherichia coli bacterial LPS displayed a temporal increase in liver TDO heme content and activity, reaching a maximum 4 to 6 h post injection and then returning back to or falling below the original baseline after 10 h (37,38). A similar bimodal effect was observed for cellular IDO1 activity in response to various immunologic stimuli (39,40). Although such dynamic regulation of TDO and IDO1 activities is likely to be biomedically relevant, how it occurs upon immune activation is currently unknown.
In researching this topic we noticed that the temporal changes in the rat liver TDO heme content and activity following LPS injection were quite similar to the induction of nitric oxide (NO) synthesis activity that is caused by a similar LPS injection in mice (41). We therefore investigated a possible role for NO in driving these immune-related changes in the activities and heme contents of the two dioxygenases. NO at relatively high levels is already known to block heme insertion into several heme proteins including inducible NO synthase (NOS) (42), endothelial NOS, neuronal NOS, two cytochrome P450 enzymes (CYP 3A4 and 2D6), catalase, and hemoglobin (43). Conversely, lower levels of NO were shown to cause cells to insert heme into the heme-free form of soluble guanylyl cyclase β subunit (sGCβ) and thus trigger assembly of the sGCαβ heterodimer (44), which is the functional NO sensor. In our present study we investigated mammalian cells in culture that either naturally express TDO or IDO1 or did so upon transfection and we utilized NO generated either by a well-characterized NO donor compound or by macrophage cells induced to express NOS. We determined IDO1 and TDO activities and heme contents by measuring Kyn production and radiolabel heme incorporation, respectively, and we also investigated possible roles for cellular GAPDH and Hsp90 in the NO-directed processes. Our findings establish that TDO and IDO1 activities in cells are beholden to NO-driven effects on their heme contents. This helps to explain how immune stimulation regulates their activities and reveals a new mechanism by which IDO1 and TDO function may be up-or downregulated in health and disease.

Results
NO has a concentration-dependent, bimodal effect on IDO1 and TDO dioxygenase activities We first examined how varying NO concentration would impact the activities of TDO and IDO1 in cells. Human cell lines that either constitutively express human TDO (HepG2) or did so upon transfection (HEK293T) were cultured for 12 h in the presence of varying concentrations of the slow-release NO donor 2,2ʹ-(hydroxynitrosohydrazino)bis-ethanamine (NOC-18), whose half-life is reported to be 21 h at pH 7.4 and 37 C (https://www.dojindo.eu.com/store/p/237-NOC-18. aspx) and was calculated to have a half-life of 16 h based on our measurements of NO release from NOC-18 under our particular cell culture conditions (Fig. S1). HEK293T cells that had been transfected to express human IDO1 underwent identical treatment. Dioxygenase activities were determined from the product N-formyl-kynurenine that accumulated in the culture fluid (measured after hydrolyzing it to Kyn) over the 12-h period. Figure 1, A-C shows that cells exposed to a low range of NOC-18 concentrations (0.1-5 μM) increased their TDO and IDO1 activities between 4-and 6-fold in a concentration-dependent manner. The effect then steadily ebbed such that cells receiving 25 μM NOC-18 displayed nearbasal levels of activity while those receiving NOC-18 at 50, 75, or 100 μM had activities below the basal levels. In the empty vector-transfected HEK293T cells (Fig. 1D) there was no Kyn formation in the entire range of NOC-18 treatments. Western blot analysis showed that the changes in activity were not due to changes in the levels of TDO or IDO1 protein expression in the cells, which remained similar across the range of NOC-18 exposures (Figs. S2-S4). Thus, IDO1 and TDO when expressed in cells underwent a concentration-dependent bimodal change in their activities in response to NO exposure that was independent of the cell identity or the manner of enzyme expression (natural versus transient transfection).
The NO effect on dioxygenase activities correlates with a similar effect on their heme contents We next examined if the changes in IDO1 and TDO activities were related to changes in their heme contents. Glycine auxotroph Chinese hamster ovary cells (GlyA-CHO) were transfected to express human TDO-FLAG and IDO1-FLAG proteins in 14 C-Glycine-containing medium for 48 h, after which protein expression was stopped with cycloheximide (Chx). Cells were then treated with NOC-18 at different concentrations for 12 h, and the effects on the IDO1 and TDO activities and 14 C-heme contents were determined by measuring Kyn production and by immunoprecipitation with anti-FLAG antibody and 14 C scintillation counting, respectively. Controls included cells not transfected to express the dioxygenases, and transfected cells whose heme biosynthesis was intentionally blocked by inclusion of succinyl acetone (29,43), which enabled us to calculate what percentage of 14 C counts in the IDO1 or TDO pull-downs were due to their 14 Cheme incorporation versus 14 C-Gly incorporation into the protein chains. We observed that the counts due to 14 C-Gly in the proteins were similar across all the different culture conditions (Fig. S14) and these were subtracted in each case from the total sample counts in the pull-downs to obtain the 14 Cheme specific counts. Figure 2, A and B show that NOC-18 treatment of the GlyA-CHO cells had a concentration-dependent and bimodal impact on the IDO1 and TDO activities that were essentially identical to what we had observed when these enzymes were expressed in the HepG2 and HEK293T cells. Figure 2, C and D show that the change in activities correlated directly with changes in the level of 14 C-heme bound within IDO1 or TDO. Their 14 C- Figure 2. NO regulates the activities of TDO and IDO1 by regulating their heme contents in cells. GlyA-CHO cells were transfected to express TDO-FLAG and IDO1-FLAG in 14 C-Gly-containing media for 48 h after which protein expression was stopped with Chx and NOC-18 was added at the indicated concentrations and culture resumed for an additional 12 h, followed by assay of Kyn accumulation and Ab pull-down of FLAG-labeled TDO or IDO1 to measure their 14 C-heme contents. A and B, IDO1 and TDO activities. C and D, IDO1 and TDO 14 C-heme counts. Data are the mean ± SD; n = 3 experiments. ***p < 0.001, one-way ANOVA.
heme contents rose when cells were given 0.1 to 5 μM NOC-18 to reach an approximate 3-fold increase in 14 C-heme and then steadily fell when given higher NOC-18 concentrations such that IDO1 and TDO contained only original levels of 14 Cheme in cells incubated with 25 μM NOC-18 and contained subbasal levels in cells that had been incubated with 50 μM NOC-18 or above. TDO or IDO1 protein expression in the cells remained similar across the range of NOC-18 exposures (Figs. S5 and S6). Thus, NO-driven changes in IDO1 and TDO heme contents correlated with and could explain the NOdriven changes in their catalytic activities.
To estimate the extent to which NOC-18 boosted the IDO1 and TDO heme incorporation, we compared the extent to which heme addition would boost IDO1 and TDO activities in supernatants prepared from cells that had or had not been cultured for 6 h with 5 μM NOC-18. Heme addition in this circumstance is known to fully saturate the dioxygenases and has been used as a way to estimate their original levels of heme saturation (32,33,38). The activities that we obtained are compiled in Table S1. For the supernatants of cells cultured in the absence of NOC-18, addition of heme at 3 or 6 μM similarly caused 5-and 3-fold increases in the IDO1 or TDO activities, respectively, consistent with the cells expressing IDO1 and TDO predominantly in their heme-free forms under normal culture conditions. In contrast, supernatants from the NOC-18 treated cells displayed the same high IDO1 and TDO activities even in the absence of added heme, and heme addition to these supernatants caused no further increase in their activities. These results indicate that the 5 μM NOC-18 treatment caused the cells to completely saturate their apo-IDO1 and apo-TDO subpopulations with cell-generated heme.

NO generated by immune-stimulated cells alters TDO and IDO1 activities and heme contents
To test if NO naturally generated by immune-stimulated cells would have a similar effect as the NO released from NOC-18, we performed a coculture experiment (Fig. 3A) where RAW264.7 macrophage cells that had been cultured in permeable transwell inserts and either not stimulated or activated by E. coli LPS to induce expression of inducible NOS and NO production were placed above a monolayer of GlyA-CHO cells that expressed either TDO-FLAG or IDO1-FLAG that had 14 C-heme incorporated due to their prior culture with 14 Cglycine. The transwell inserts contained two different quantities of activated RAW264.7 cells (25% and 100% confluent) to create a lower and higher NO production condition, respectively. The cocultures were analyzed at 0 h or after 6 h had elapsed. Figure 3B shows that nitrite, a NO-derived product, was generated by the activated RAW264.7 cells and accumulated in the coculture fluid after 6 h at levels that corresponded to the two different quantities of activated cells that were present in the inserts. Figure 3C shows that the GlyA-CHO Figure 3. NO released by immune-activated macrophage cells regulates the heme level of TDO expressed in underlying cells. A, transwell inserts containing RAW264.7 macrophage cells at 0, 25, and 100% confluency that had or had not been activated by bacterial LPS to express NOS and generate NO were placed into culture wells that contained a monolayer of GlyA-CHO cells expressing TDO-FLAG that contained 14 C-heme incorporated due to prior culture of the cells with 14 C-Gly. After 6 h of coculture nitrite (B) and Kyn (C) produced in the cell cultures were measured, along with the 14 C heme counts in TDO-FLAG pulled down after 0 h and 6 h of coculture (D). Data are the mean ± SD of three experiments. ***p < 0.001, **p < 0.01, one-way ANOVA. cells that were cocultured for 6 h with the lesser number of activated RAW264.7 cells displayed an increase in their TDO activity, while those cocultured with the higher quantity of activated cells displayed a decrease in TDO activity. These activity changes correlated with a gain or loss in TDO 14 Cheme level relative to what was initially present in TDO in the cell cultures analyzed at t = 0 h (Fig. 3D). In replica experiments using transwell inserts containing nonactivated RAW264.7 macrophages we observed no change in the cell TDO activity or heme content after the 6 h coculture (Fig. 3, C and D). Nearly identical results were obtained in coculture experiments where expression of IDO1 substituted for TDO in the GlyA-CHO cells (Fig. 4, A-C). The expression levels of TDO or IDO1 were similar across all cultures and conditions (Fig. S7). Thus, NO released from immune-stimulated macrophage cells caused a concentration-dependent bimodal change in cell TDO and IDO1 activities and heme contents that mirrored what we observed for cells exposed to low or high levels of NO released by the donor NOC-18.

Time course of the NO-driven changes in dioxygenase activities and heme contents
To investigate the kinetics of the NO-driven changes we expressed TDO or IDO1 in 14 C-Gly-treated GlyA-CHO cells and then exposed them to media alone or either to a lower (5 μM) or higher (100 μM) concentration of NOC-18. Figure 5 shows the buildup of product Kyn in the cultures versus time under the three conditions for cells expressing IDO1 (panel A) or TDO (panel B). Cells given 5 μM NOC-18 showed greater Kyn buildup from IDO1 or TDO than did the control cultures at all time points, which resulted in a seven or three times greater Kyn concentration accumulated by 12 h, respectively. In contrast, cells given 100 μM NOC-18 showed less Kyn buildup from IDO1 or TDO than did the control cells, which resulted in a four or five times less Kyn concentration accumulated by 12 h. Figure 5, C and D report the corresponding levels of 14 C-heme in IDO1 or TDO after they were immunoprecipitated at each time point for the three different culture conditions. In the control cultures the IDO1 and TDO maintained constant levels of 14 C-heme throughout the time course. Cells given 5 μM NOC-18 stimulated incorporation of 14 C-heme into IDO1 and TDO that could be detected even after the first hour of exposure and then continued to increase until 6 h of exposure, after which it remained at a constant level that was 2 to 3 times above the original 14 C-heme contents of IDO1 and TDO at time = 0 and that was present in IDO1 or TDO in the control cell cultures. In contrast, in cells given 100 μM NOC-18 there was no gain in 14 C-heme content in IDO1 or TDO at any time point and instead their 14 C-heme Figure 4. NO released by immune-activated macrophage cells regulates the heme level of IDO1 expressed in underlying cells. Transwell inserts containing RAW264.7 macrophage cells at 25 and 100% confluency that had or had not been activated by bacterial LPS to express NOS and generate NO were placed into culture wells that contained a monolayer of GlyA-CHO cells expressing IDO1-FLAG that contained 14 C-heme incorporated due to prior culture of the cells with 14 C-Gly. After 6 h of coculture the nitrite (A) and Kyn (B) produced in the cell cultures were measured, along with the 14 C heme counts in TDO-FLAG pulled down after 0 h and 6 h of coculture (C). Data are the mean ± SD of three experiments. ***p < 0.001, **p < 0.01, one-way ANOVA. levels steadily decreased relative to the control cultures, even by the first hour of exposure. Western blot analysis indicated that none of the changes in activity or heme content could be attributed to changes in IDO1 or TDO protein expression in the cells, which remained similar across the time points and different culture conditions (Figs. S8 and S9). Thus, the low NO exposure caused cells to begin steadily incorporating heme into their apo-IDO1 or apo-TDO populations within an hour, whereas high NO exposure never stimulated any heme incorporation and instead caused the dioxygenases to begin losing their heme.

NO stimulates cell heme delivery to IDO1 and TDO through a GAPDH-dependent mechanism
During their normal maturation process, cell heme delivery to apo-IDO1 and apo-TDO requires the formation and participation of a GAPDH-heme complex (29). We therefore investigated if the NO-driven heme acquisitions also involved GAPDH by employing our established siRNA knockdown and rescue strategy (29). GlyA-CHO cells that first underwent targeted siRNA knockdown of GAPDH expression or treatment with scrambled siRNA were given 14 C-Gly and were then transiently transfected to express IDO1 or TDO alone or in combination with siRNAresistant forms of wildtype hemagglutinin (HA)-GAPDH or the heme-binding defective HA-GAPDH-H53A variant.
Cells were then given 0, 5, or 100 μM NOC-18, and the activities and heme contents of their IDO1 and TDO were determined after 12 h.
The targeted knockdown of GAPDH for 48 h lowered its cell expression by approximately 80% relative to controls, consistent with our previous results (29); the expression of either HA-GAPDH protein in the knockdown cells restored their total GAPDH expression level to a normal value; and the cell IDO1 and TDO expression levels were unaffected by the different treatments (Figs. S10 and S11). Figure 6, A and B show that GAPDH knockdown severely diminished Kyn accumulation in the cell cultures expressing IDO1 or TDO, as we reported previously (29), and it also severely diminished the gain in Kyn production that otherwise occurred in response to 5 μM NOC-18. Under both circumstances (without or with 5 μM NOC-18) the expression of the wildtype HA-tagged GAPDH in the knockdown cells restored TDO and IDO1 activities to normal as judged from the levels of Kyn buildup, whereas expression of the GAPDH heme binding mutant (HAtagged H53A GAPDH) in the knockdown cells did not. In cells undergoing higher NO exposure from 100 μM NOC-18, there was no increase in IDO1 or TDO activities under any circumstance and instead their activities were inhibited relative to the control cells. Figure 6, C and D reveal that the GAPDH knockdown and rescue procedures impacted the heme contents of IDO1 and TDO in a manner that matched the impact on their catalytic activities, in both the control cells and in the cells that were treated with 5 μM NOC-18. We conclude that the cell heme allocation into apo-IDO1 and apo-TDO that was driven by 5 μM NOC-18 and their consequent gain in activity is dependent on the cell GAPDH expression level and on the ability of GAPDH to bind intracellular heme.

NO-stimulated heme insertion into IDO1 requires Hsp90
Because Hsp90 is known to drive heme insertion into apo-IDO1 but not into apo-TDO during their normal maturation process (29), we investigated if Hsp90 was also needed for NOdriven heme allocation. We used the specific inhibitor radicicol to test how blocking the Hsp90 ATPase function in GlyA-CHO cells would impact the NO-driven change in IDO1 and TDO activities and heme contents. Figure 7, A and B show that a 6-h preincubation with radicicol completely prevented 5 μM NOC-18 from stimulating cells to increase the activity and heme content of IDO1, while in contrast it had no ability to block 5 μM NOC-18 from increasing cell TDO activity and heme content (Fig. 7, C and D). Western blot analyses showed that radicicol treatment did not change the expression levels of IDO1 or TDO under any circumstance, which remained constant throughout the experiment (Figs. S12 and S13). These results indicate that functional Hsp90 is needed for the NO-stimulated heme allocation into apo-IDO1 but not into apo-TDO.

Discussion
We found that NO can positively or negatively regulate IDO1 and TDO activities in cells through its influence on cell heme allocation. Cells exposed to a low-level of NO generation were stimulated to increase the heme contents and catalytic activities of their IDO1 and TDO. Beyond this range, NO steadily lost its positive effect and ultimately caused a loss of the existing heme that was bound within either dioxygenase enzyme. This bimodal effect was observed when NO was released by a chemical NO donor or was released naturally by immune-activated macrophage cells. It was also independent of what cell type expressed IDO1 or TDO and whether the cells naturally expressed the dioxygenases or did so by transfection. Thus, we conclude that NO has a broad ability to both positively or negatively impact cell IDO1 and TDO activities through it regulating cell heme allocation to or from the dioxygenases.
Our experiments utilized cells that were cultured under standard conditions and thus contained their normal or resting levels of heme. In this circumstance both IDO1 and TDO are known to be only partially heme-saturated (29), which allowed us to find that low NO exposures promote heme allocation into the apo-IDO1 and apo-TDO subpopulations that are naturally present in the cells. We saw that NO-driven heme allocation into apo-IDO1 or apo-TDO began within the first hour of the low NO exposure and continued in a steady manner for 6 h until the TDO and IDO1 heme levels had increased by about 3-to 4-fold. The overall effect on their activities was significant and led to a 7-fold increase in total Kyn production and corresponded with both IDO1 and TDO becoming fully heme replete. Figure 6. NO-driven cell heme allocation to apo-TDO and apo-IDO1 is GAPDH dependent. GlyA-CHO that underwent siRNA knockdown of GAPDH expression or treatment with scrambled siRNA were given 14 C-Gly and then transiently transfected to express IDO1 or TDO alone or in combination with siRNA-resistant forms of wildtype HA-GAPDH or the heme-binding defective HA-GAPDH-H53A variant. Cells were then given 0, 5, or 100 μM NOC-18 and the Kyn production (A and B) and 14 C-heme contents of IDO1 and TDO (C and D) were determined after 12 h. Circle, control siRNA; square, GAPDH siRNA; triangle, GAPDH siRNA plus HA-GAPDH; inverted triangle, GAPDH siRNA plus H53A HA-GAPDH. Data are the mean ± SD; n = 3 experiments. ***p < 0.001, ns = not significant, one-way ANOVA.
It is remarkable how low an NO exposure triggered the cells to allocate heme into apo-TDO and apo-IDO1. NOC-18 has the slowest NO release rate of any NO donor that is commercially available. Despite this, the NO generated by the lowest NOC-18 concentration that we utilized (0.1 μM) still caused a measurable gain in the IDO1 and TDO heme contents over a 12-h period, and exposing the cells to only 5 μM NOC-18 promoted an optimal heme allocation, with an increase in the IDO1 and TDO heme contents being observable even within the first hour exposure. The NO release from NOC-18 at these concentrations is extremely low, 0.13 and 6.6 nM/min, respectively, for 0.1 μM and 5 μM NOC-18 based on measures we made under our cell culture conditions. In general, NOC-18 has been reported to create solution NO concentrations that are about 1000 times less than its initial concentration (45). This suggests that NO concentrations of only 100 pM to 5 nM were being reached in our cell cultures yet still could stimulate cells to allocate heme into apo-IDO1 and apo-TDO. Such concentrations of NO can enable metalnitrosyl formation, for example, NO binding to the heme in sGC to activate its cGMP synthesis, and also can drive cells to allocate heme into apo-sGC β (44, 46), but are somewhat lower than the NO concentrations required to stimulate cells to phosphorylate extracellular signal-regulated kinase (ERK) or Akt, and are 20 to 80 times lower than the NO concentrations required to promote hypoxia-inducible factor 1 α stabilization or P53 phosphorylation in cells (47). Thus, stimulating cell heme allocation may be one of the more sensitive NO effects in biology. It is equally remarkable how narrow the window of NOC-18 concentration was that could promote cell heme allocation to apo-IDO1 and apo-TDO. Its effectiveness peaked at only 5 μM NOC-18 and then decreased fairly sharply in all four cell types that we examined, such that the positive effect was completely lost by 50 μM NOC-18, which is a concentration that still generates a relatively low NO exposure. Indeed, because NOC-18 has almost never been used below 50 μM in the literature and instead is typically used at concentrations of 100 to 2000 μM, we suspect that the impact of very low NO on other heme proteins and in biological systems in general has been understudied and to a large extent is still overlooked.
Regarding mechanism, NO exposure is known to increase the level of exchangeable heme inside cells as detected by an intracellular heme sensor (48). This effect occurred over tens of minutes following NO exposure, which is consistent with the kinetics for the NO-driven heme allocations to apo-TDO and apo-IDO1 that we observed here. NO is also known to speed heme intake and utilization within mammalian cells (49), and NO can trigger heme transfer between purified proteins (50). This suggests that NO may improve heme availability and its mobility inside the cell and deserves further study. We also found that the NO-driven heme allocations to apo-IDO1 and apo-TDO relied on formation of a GAPDHheme complex in the cells, and for IDO1 also relied on Figure 7. Hsp90 is needed for NO-driven heme allocation to apo-IDO1 but not to apo-TDO. GlyA-CHO cells were given 14 C-Gly and transfected to express IDO1 or TDO, then given 10 μM radicicol for a further 6 h. NOC-18 at 0, 5, or 100 μM was added to the cells in the continued presence of radicicol (indicated in red, green, and blue respectively, time = 0) and cells were harvested at the indicated times. A and B, IDO1 Kyn production and 14 C-heme contents. C and D, TDO Kyn production and 14 C-heme contents. Data are the mean ± SD; n = 3 experiments. ***p < 0.001, ns = not significant, one-way ANOVA.
Hsp90 function, which means that the NO acts through a mechanism that involves the same machinery that cells normally use to deliver heme to these dioxygenases (29) and to several other hemeproteins during their maturation (51)(52)(53)(54). Within this context NO could act in several ways. The initial heme provision to apo-IDO1 and apo-TDO that occurs within the first hour of low NO exposure likely results from a relatively direct redistribution of the existing cellular heme, for example, by NO possibly promoting heme transfer from preexisting GAPDH-heme complexes in the cell or by it increasing heme loading onto GAPDH. Beyond this time point the possible mechanisms of action can expand and conceivably include NO-induced changes in cell heme biosynthesis, protein expression, or NO-based posttranslational protein modifications. This warrants further study. We also note that the heme allocations to apo-IDO1 and apo-TDO induced by low NO exposure did not reverse after reaching a maximal level after 6 h exposure. This suggests there is a range of low NO exposure where negative impacts of NO on heme allocation never manifest or do so in ways that can be compensated for by protective processes that may be operating constitutively within the cells.
We observed that higher NO exposure negatively impacted TDO and IDO1 by inhibiting their activities, consistent with previous reports showing NO inhibits IDO1 activity at higher concentrations (39). Curiously, the higher NO exposure did not trigger any initial heme allocation into apo-IDO1 and apo-TDO and instead only caused the heme-containing IDO1 and TDO subpopulations that were present in the cells to steadily lose their heme. There was no loss of 14 C-heme from these enzymes if cell cultures did not receive the NO, and because the experiments were done in the presence of Chx, the results were not complicated by any new IDO1 or TDO protein synthesis. Moreover, Western analysis showed that there was no change in the TDO or IDO1 protein expression levels over the time course of high NO exposure, indicating that the loss of 14 C-heme counts was not due to the IDO1 or TDO degradation. This leads us to conclude that the higher NO exposure caused cells to "allocate" heme away from the subpopulation of heme-replete IDO1 and TDO. This is reminiscent of reports showing that exposure to higher NO levels prevented cells from allocating heme into globins, NOS enzymes, cytochrome P450s, and catalase (43) and could even result in heme loss from cytochrome P450s (55). However, it contrasts with recent results we obtained for the hemeprotein sGCβ, where we found that NO exposure at both a low and high level caused cells to allocate heme into apo-sGCβ (44). This apparent discrepancy implies that higher NO levels may impact cell heme allocations differently depending on the identity of the recipient hemeprotein.
How NO blocks cell heme allocation or causes heme loss from proteins is still mostly unclear. We know that it blocks cell heme allocation to NOS by causing buildup of a particular posttranslational modification in GAPDH, namely, Cysnitrosated GAPDH (SNO-GAPDH) (56). Whether this is a common mechanism whereby NO blocks cell heme allocations to other hemeproteins whose heme deliveries are also GAPDH dependent (29,53,54,56) is an intriguing possibility. Because cells also express enzymes that denitrosate SNO-GAPDH and recover its function (57), this activity in cells could conceivably diminish or delay the buildup of SNO-GAPDH and thus allow for a window of NO levels to have a "beneficial" effect on GAPDH-dependent heme allocations. In any case, it seems likely that the mechanisms by which low NO promotes cell heme allocations differ from the mechanisms by which high NO blocks heme allocations or causes heme loss in heme proteins. Together, these mechanisms likely combine to allow NO to up-and downregulate IDO1 and TDO activities in cells and can now be further investigated.
NO-driven cell heme allocation into TDO and IDO1 may finally explain the classical observation that immune activation in rats caused a temporary increase in their liver TDO activity and heme content, as reported decades ago (11,12,30,36), and may also explain how immune stimulation or NO exposure was found to change IDO1 activity in cells and animals (39,58). Indeed, it is remarkable how the time course of NO-driven 14 C-heme incorporation into apo-TDO or apo-IDO1 as we found here (Fig. 5) correlates with the reported gain in liver TDO activity and heme content in rats following an LPS injection (13), which in turn correlates with the reported increase in the blood levels of NOS-derived nitrate (an NO breakdown product) in mice over time following an LPS injection (41). Our finding that the NO generated by lower numbers of LPS-stimulated macrophages could increase the activities and heme contents of TDO and IDO1 in neighboring cells recapitulates these early animal studies and overall suggests a mechanism whereby the LPS injections, by inducing NOS expression and an increased NO production in the animals, can initially stimulate heme allocation into their apo-TDO (or apo-IDO1) populations, thus boosting their overall activity for Trp metabolism. In short, our current findings potentially explain how immune stimulation and the resultant biological NO generation can up-or downregulate Trp metabolism in mammals via an ability to up or downregulate the heme contents of IDO1 and TDO.
Our findings have other biomedical implications. For example, they may help explain why low-level NO production has often been found to be procancerous (59,60). Given that IDO1 and TDO are often expressed in malignant tumors and cause unwanted suppression of the immune system via their production of Kyn (27,61), we speculate that a low level of NO might actually be boosting the IDO1 and TDO heme contents, thereby increasing Kyn production and the resultant immune suppression that helps tumor cells escape from immune surveillance. Likewise, a high level of NO via high iNOS expression in tumors is associated with limiting or reversing cancer growth (62). Because IDO1 and TDO both lost their heme when cells were given a higher NO exposure, it is possible that inactivating IDO1 and TDO in this way through heme loss could diminish Kyn production and limit immune suppression, thereby enabling the T-cell differentiation that is needed to achieve tumor surveillance. In autoimmune diseases like severe asthma, high levels of NO are often produced in the airway due to an increased iNOS expression (63). Although IDO1 protein expression is typically upregulated in asthma, the higher NO levels could cause a heme deficit that limits IDO1 heme content and Kyn production, thus diminishing immune suppression and enabling the lung inflammation that is a characteristic of severe asthma (64). Indeed, IDO1 expressed in airways is reported to have lower-than-normal activity in both adult and pediatric patients with severe asthma (65), and in a mouse asthma model, active IDO1 protected the animals from developing severe airway inflammation (58). Taken together, this suggests that the bimodal impact of NO on TDO and IDO1 heme contents and activities as described here is likely to be medically important and may present new opportunities to preserve, boost, or limit Kyn production in health and disease.

Summary
There is a dynamic and bimodal regulation of heme levels in IDO1 and TDO by cells in response to different NO levels. The bimodal impact of NO on cell heme allocation could explain prior observations where inflammation and resultant NO generation altered the activities of IDO1 or TDO. Because several other hemeproteins naturally exist in a partially hemesaturated state, and NO also promotes cell heme allocation to sGC β (44), we speculate that NO may broadly impact the biological functions of hemeproteins in this manner.

Experimental procedures
Materials [ 14 C]-glycine (0.5 mCi) was purchased from ICN Biomedicals. All other reagents used were purchased from Sigma unless otherwise mentioned.

Growth of HepG2 cells
HepG2 cells (ATCC # HB-8065) naturally expressing TDO were cultured in tissue culture treated plates containing Eagle's minimum essential medium (ATCC # 30-2003) containing 10% fetal bovine serum (Gibco) until 90% confluent after which Chx (Sigma # C7698) was used to treat cells at 5 μg/ml for 12 h to inhibit further protein synthesis. The expression of TDO was checked using anti-TDO antibody (Proteintech # 15880-1-AP). Cells were then utilized for experiments involving treatment with NOC-18 NO donor for indicated doses and time points.

Treatment of cells with radicicol
Hsp90 inhibitor radicicol (Sigma # R2146) was used to pretreat cells for 6 h before treatment with NOC-18 at a final concentration of 10 μM. After pretreatment, cells were maintained in the presence of 10 μM radicicol during NOC-18 treatment for the indicated time points and doses. 14 C-labeled heme production in cells and measuring radiolabeled heme counts 14 C glycine uptake by GlyA-CHO cells generated 14 C-labeled heme which was incorporated into FLAG-IDO1/ TDO. We maintained a negative control where GlyA-CHO cells were treated with heme synthesis inhibitor succinyl acetone (Sigma # D1415) at 400 μM for 72 h post transfection of FLAG-IDO1/TDO plasmids along with 2 μCi/ml of [ 14 C] glycine. In this circumstance the heme proteins are expressed with 14 C glycine incorporated into the polypeptide but do not have any heme incorporated into them. The 14 C counts of the FLAG-IDO1/TDO from these negative controls were subtracted from the experimental samples. Chx was used to treat cells at 5 μg/ml for 12 h to inhibit further protein synthesis. The method for measuring 14 C heme counts using a scintillation counter is described in (66).

Transfection of siRNA and gene silencing
Cell GAPDH protein expression was reduced using siRNA against human GAPDH mRNA. Commercially available siRNA against human GAPDH (# D-001830-01-05) and scrambled siRNA (# D-001810-10-05) were purchased from Dharmacon and used at a final concentration of 100 nM in cultures of mammalian cells of low passage number along with Lipofectamine 2000. The siRNA-treated cells were cultured for 72 h before they received transfections with protein expression plasmids as described above.

IDO1 and TDO activity assay from cell culture medium
The enzyme activity of IDO1 and TDO was measured using a colorimetric assay. The cells after treatment with Chx were treated with IDO1/TDO substrate L-Trp at 2 mM in phenol red-free DMEM/Eagle's minimum essential medium containing the appropriate type of serum (normal or heme depleted) for indicated time to allow for substrate utilization and Kyn product formation. The medium was collected and deproteinized by adding an equal volume of 3% trichloroacetic acid and incubated at 50 C for 30 min. The tubes were centrifuged at 9000g for 10 min at room temperature to precipitate proteins from the medium. Equal volumes of deproteinized sample were mixed with a freshly made 20 mg/ ml solution of p-dimethyl-amino-benzaldehyde (Ehrlich's reagent; Sigma # 109762) in glacial acetic acid at room temperature to allow for the formation of a yellow-colored product. The end point absorbance of this product was measured at 492 nm (Molecular Devices) to determine the concentration of Kyn in the medium. A standard curve was obtained using commercial Kyn (Sigma # K8625) dissolved in 0.5 N hydrochloric acid in various concentrations using the exact same method.

IDO1 and TDO activity assay in cell supernatants
The activities of IDO1 and TDO in cell supernatants were measured as described in (67) with modifications. Briefly, 500 μg of cell supernatant protein was added to a 1-ml reaction containing 50 mM potassium phosphate buffer (pH 6.5), 20 mM ascorbate, 10 μM methylene blue, 100 μg/ml catalase, and 1 mM L-Trp and incubated at 37 C for 30 min. The reactions were stopped by adding 1 ml of 30% trichloroacetic acid to denature the proteins. The next steps were followed as per the previously described colorimetric detection method of Kyn using Ehrlich's reagent. In some cases, the cell supernatants were given heme prior to assay. Supernatants (1 mg protein) were made 3 or 6 μM in heme, incubated at room temperature for 1 h, passed through a desalting column (Sephadex G-25 resin), and then assayed for activity as described above.

Immunoprecipitation and Western blot
GlyA-CHO cells were lysed using 50 mM Tris-HCl pH 7.4 buffer with 0.1% Triton X-100, 5 mM Na-molybdate, and EDTA-free protease inhibitor cocktail (Roche). Protein concentration was measured using the Bradford method (Bio-Rad # 500-0006). Immunoprecipitation pull-downs were performed using 1 mg of whole cell extracts with anti-FLAG antibody (Sigma # F1804). Protein G agarose beads (Millipore # 16-201) were used to pull down the antibody-protein complex. The beads were washed well with lysis buffer, the 1.5-ml tubes were inserted into 5-ml scintillation vials and 4 ml of scintillation fluid (Liquiscint, National Diagnostics # LS-121) was added to each vial. The method for measuring 14 C heme counts using a scintillation counter is described in (66).

Transwell coculture experiment
We used a six-well plate format (Corning # CLS3450) for our transwell coculture experiments. We seeded RAW264.7 (ATCC # TIB-71) cells in the upper chamber. These cells at a final confluency reached 25% and 100% when they were activated with LPS at 1 μg/ml (Sigma # L2630) for 6 h. The bottom chamber had GlyA-CHO cells expressing FLAG-IDO1/TDO in 14 C glycine-containing medium. After initial activation of the RAW264.7 with LPS for 6 h, the upper chamber baskets were introduced to the GlyA-CHO cells along with fresh medium in the bottom chamber for 6 h. The medium was phenol red free and was used to measure the activities of IDO1/TDO as described previously. Also, nitrite accumulation in this medium was measured using the Griess reagent system (Promega # G2930). The lysates of the RAW264.7 cells were used to check iNOS expression by Western blot. The lysates of the GlyA-CHO cells were used in immunoprecipitation to measure 14 C heme counts and to determine FLAG-IDO1/ TDO expressions by Western blot.

NO release from NOC-18
The NO-mediated conversion of oxyhemoglobin to methemoglobin was used to determine the rate of NO release from NOC-18 at 37 C. Various concentrations of NOC-18 were added to cuvettes that contained phenol red-free DMEM, 10% fetal bovine serum, 2 mM L-Trp, and 10 μM oxy-hemoglobin. The absorbance gain at 401 nm was recorded per minute over a 3-h period for each concentration of NOC-18. The rate of NO release was calculated using the difference extinction coefficient of 38 mM −1 cm −1 (43).

Statistical analyses
All experiments were done in three independent trials, with three replicates per trial. The results are presented as the mean of the three trial values ± standard deviation. The statistical test used to measure significance (p-values) was one-way ANOVA in the software GraphPad Prism (v9).

Data availability
All data are contained within the article.
Supporting information-This article contains supporting information.