**2. Results**

#### *2.1. Lipidomics Analysis of the Effect of NO*• *on 15LOX-2/PEBP1 Peroxidation Activity*

To examine the effects of NO• on the production of the pro-ferroptotic signal, 15- HpETE-PE, by 15LOX-2/PEBP1, we incubated one of the lipoxygenase isoforms, 15LOX-2, and PEBP1 in the presence of an NO•-donor, propylamine (PAPA) NONOate [17]. This donor releases NO• at a constant rate with a decay half-time of 15 min. We found that NO• strongly suppressed the production of hydroperoxide of 1-SA-2-ETE-PE (HpETE-PE) (Figure 1a). Detailed analysis of redox lipidomics data led to the identification of a nitrosylated product, nitroso-ETE-PE (1-SA-2-ETE(-NO)-PE) generated by 15LOX-2/PEBP1 only in the presence of NO• (Figure 1b–d).

**Figure 1.** Effect of NO• on peroxidation of 1-SA-2-ETE-PE by the 15LOX-2/PEBP1 complex. (**a**) Kinetics of the production of 1-SA-2-HpETE-PE by 15LOX-2/PEBP1 in the presence and absence of NO• donor. Data are mean ± SD, *n* = 3, \*\*\* definition; \*\*\*\* *p* < 0.0001 vs. 15LOX-2/PEBP1 + NO, two way ANOVA, Sidak post-hoc analysis. (**b**) Mass spectrum showing the nitrosylated product formed upon 1-SA-2-ETE-PE incubation with 15LOX-2/PEBP1+PAPA NONOate (back). Nitroso-1- SA-2-ETE-PE (1-SA-2-ETE(-NO•)-PE) was not detected in the system without PAPA NONOate (front). Fragmentation analysis of 1-SA-2-HpETE-PE (**c**) and 1-SA-2-ETE(-NO)-PE (**d**) showing representative fragments. The insets at the top show the structures of 1-SA-2-HpETE-PE, 1-SA-2-ETE(-NO)-PE and their possible fragments. (**e**) Bar plot comparing 1-SA-2-ETE(-NO)-PE formation by 15LOX-2 and by 15LOX-2/PEBP1, both in the presence of NO•. Data are mean + SEM., *n* = 3, \*\*\* *p* = 0.0007, student's *t*-test. (**f**) Bar plot showing that the amount of 1-SA-2-HpETE-PE inhibited by NO• is significantly larger than the amount of 1-SA-2-ETE(-NO)-PE produced, as a function of reaction time. The inset bar plot shows the enlarged view of the lower plot, between the ordinate values 0.0<y< 0.04 (AU).

These observations sugges<sup>t</sup> that a direct reaction occurs between the ETE-PE carboncentered radical, an intermediate of the 15LOX-catalyzed dioxygenase cycle [18], and the NO• radical. No nitroxygenated products formed from the reaction of O2 with the reactive intermediates of NO• [19] could be found. Notably, the 15LOX-2/PEBP1 complex produced a greater amount of NO•-ETE-PE than 15LOX-2 alone (Figure 1e). Lack of appropriate standards of nitrosylated ETE-PE precluded the absolute quantification of nitrosylated ETE-PE and its comparison with 1-SA-2-HpETE-PE levels. However, upon normalization with the same internal standard, the relative amounts of the nitrosylated ETE-PE were about 50 folds lower than the inhibited amounts of 1-SA-2-HpETE-PE (Figure 1f).

Combined, these results sugges<sup>t</sup> that at least two mechanisms contribute to the observed suppression of ETE-PE peroxidation by NO• in the 15LOX-2/PEBP1 complex: (i) direct reaction of NO• with the ETE-PE carbon-centered intermediate, and (ii) potential presence or occupancy of the O2-binding site and channeling path by NO• or a competition between the two small molecules—O2 and NO•—interfering with the peroxidation of the oxidizable substrates. To explore the latter mechanism, we analyzed the changes in the conformational state and dynamics of 15LOX-2/PEBP1 invoked upon binding of the oxidizable substrate, as well as O2 and NO•.

#### *2.2. In silico Analysis of the Effect of NO*• *on the Structure, Dynamics, and Interactions of 15LOX-2/PEBP1*

We performed structure-based modeling and simulations to investigate the molecular basis of the experimentally observed reduced activity of 15LOX-2/PEBP1 in the presence of NO•. First, we explored how the accessibility of the catalytic site of 15LOX-2 to O2 and NO• molecules is altered upon complexation with PEBP1. Next, we investigated how the presence of NO• affects O2-binding and channeling to the catalytic pocket. We further analyzed the effect of oxidizable substrates (1-SA-2-ETE-PE, shortly called SAPE, vs. free ETE (AA)) on the accessibility of the 15LOX-2 catalytic site to O2 and NO•, in the presence and absence of PEBP1. Overall, we conducted 10 sets of molecular dynamics (MD) simulations in duplicate (or triplicate) runs of 150 nanoseconds (ns) under different conditions, summing up to total simulation duration of 3.6 microseconds in explicit water, as summarized in the Supplementary Table S1.

#### *2.3. Specific Porous Regions on 15LOX-2 Surface Enable Access of O2 and NO*• *to the Catalytic Cavity*

Lipoxygenases are non-heme iron-containing enzymes. The iron at the catalytic site in the central part of the catalytic domain of 15LOX-2 is coordinated by four highly conserved residues including three histidines: H373, H378 (on helix α12-14), H553 (on helix α27), and the C-terminal isoleucine, I676 (Supplementary Figure S1). Simulations revealed three porous regions on 15LOX-2 surface, denoted here as Entrances 1–3 (*E1–E3*), which could potentially allow for the entry of the radicals, O2 or NO• , to potentially access the catalytic cavity of the enzyme (Figure 2a): *E1*, near the loop Y154-P159 where Y154 and W158 side chains appear to be gating the entrance (Figure 2b and Supplementary Figure S1a,c); *E2*, at the 15LOX-2/PEBP1 binding interface (Figure 2a); and *E3*, near S573, P595, and A599 (Figure 2d). Although all three entrances could potentially provide access to the catalytic site [13], *E2* was largely obstructed by the oxidizable substrate (SAPE or AA in our simulations) and, therefore, inaccessible to O2 and NO•. In contrast, *E1* served as the main access point to a tunnel leading to the catalytic site, for both O2 and NO• (Figure 2a,b, orange trace), while *E3* provided an alternative path leading to the catalytic site (Figure 2a, blue trace and Figure 2d). Supplementary Figure S1 provides close-up views of *E1* and *E3* from different perspectives, to facilitate the visualization of their connection to the catalytic site; and Supplementary Figure S2 describes the secondary structure of 15LOX-2 [20].

The entrance *E1* was lined by Y154, N155, G157, W158, I421, I435, F438, and S439 (Figure 2b). It was transiently occupied by NO• or O2 at the initial stage of their interaction with the enzyme. After this first recognition event (at site *S0*), the small molecules moved deeper towards two attractive sites, binding sites *S1* and *S2*, that arrested them for extended durations (Figure 2c): *S1*, composed of N413, A416, R417, L374, and L379, was occupied by these ligands for more than 70 ns; and *S2*, containing I433, T431, V427, F365, and E369, for more than 35 ns, during 150 ns runs. Both sites made close contacts (atom-atom distance less than 3.5 Å) with NO• or O2 for sufficiently long durations (Figure 2e) meeting the criteria (see Methods) to qualify as binding sites.

**Figure 2.** Ligand (NO• and O2) binding and channeling to the catalytic site of 15LOX-2 and critical interactions mediating this process in the presence of SAPE. Results from MD simulations for 15LOX-2/PEBP1 in the presence of SAPE, five NO• and five O2 molecules are presented (Supplementary Table S1; last row). (**a**) 15LOX-2/PEBP1/SAPE complex stabilized by the end of a 150 ns MD run. 15LOX-2 and PEBP1 are displayed in white and cyan ribbon diagrams, respectively; and SAPE in cyan sticks (with atoms in CPK colors). Catalytic residues (H373, H378, H553, I676) are in green sticks, and the iron ion in pink sphere. Entrances 1 (red), 2 (inaccessible in the presence of PEBP1 and substrate), and 3 (yellow) are indicated by black/gray arrows. The series of orange dots represents the entry or diffusion path of O2 and NO• molecules through Entrance 1, and the blue dots those through Entrance 3. The region between helices α12 and α14 (K365-L380, UniProt ID: O15296), colored violet, contains two of the catalytic histidines. (**b**) A close-up view of Entrance 1, from a different angle. An O2 molecule (orange) approaches it around 37 ns and translocate to the catalytic site (orange dots in panels b and d) (**c**) A close-up view of the catalytic site with three NO• molecules occupying the recognition site *S0* (Entrance 1), and the binding sites *S1* (orange) and *S2* (blue). The cloud of small dots shows where and how long (more dots) NO• travelled or remained bound during simulations. NO• molecules are in blue-red spheres corresponding to their nitrogen and oxygen atoms, respectively. (**d**) A close-up view of the Entrance 3 providing access to a NO• molecule (in blue). Key residues lining the path are displayed in yellow. The series of orange dots represents the entry or diffusion path for O2 and NO• through Entrance 1 and the blue dots show the path through Entrance 3. (**e**) Illustration of the time evolution of contacts between 15LOX-2 residues (ordinate) and O2/NO• molecules (labeled NO\_805, OXY\_701 (shown in panel b) and OXY\_705; shown by the respective black, magenta, and blue traces) during a typical MD run. Red, yellow, orange, and blue arrows along the ordinate correspond to the Entrance 1 (*E1*), Entrance 3 (*E3*), Binding Sites *S1* and *S2*, respectively. All three molecules were arrested for extended durations at *S1* or *S2*, inserting to the catalytic site through Entrance 1 or 3.

The entrance *E3*, on the other hand, included (in addition to S573, P595, and A599), S430 and V603 near the surface (Figure 2d). These residues assist in directing the O2/NO• molecules from *E3* to the catalytic site. Entrances 1 and 2 are predominantly composed of hydrophobic residues: (i) the cluster I216, I604, F561, C564, and A565 that initially bound O2/NO• for ~40 ns; and (ii) the cluster L607, L610, L420, V426, and V427 that retained the molecules for ~11 ns. The third site that serves as a bridge between *E3* and catalytic site partially overlaps with *S2*: it contains the residues L246, E364, F365, H368, E369, and L570 that bound O2/NO• for ~40 ns. Thus, *S2* serves as an attractor for O2 or NO• binding entering through either *E1* or *E3*. Two highly conserved residues therein, F365 and E369, will be shown below to play an important role in binding or redirecting the O2 and NO• molecules.

A summary of all these residues involved in various roles, entry (*E1* and *E3*), binding to *S1* and *S2*, or insertion towards the catalytic site (mainly hydrophobic clusters), is given in Supplementary Table S2, along with corresponding secondary structures (based on [21]).

These simulations also indicated an additional pore that connected from the surface to the catalytic site through a short tunnel with a wide entrance. This path, available in both unbound- and PEBP1-bound-15LOX-2 (Supplementary Figure S3a,b, in green), was not selected by either NO• or O2. This tunnel included five residues highly conserved across LOX family members [22], L610, Q560, and the three catalytic residues, H373, H553, I676, along with E613 and S557. It is conceivable that this tunnel plays a role in 15LOX-2 dioxygenase activity, ye<sup>t</sup> to be explored.

#### *2.4. O2 and NO*• *Compete for the Catalytic Site*

Simulations with the substrate revealed an interesting effect which arises upon the complexation 15LOX-2 with PEBP1, namely a competition between O2 and NO• for a position near the C15 carbon in the arachidonoyl chain of SAPE within the catalytic site (Figure 3, Movie 1). It is known that 15LOX-2/PEBP1 complex converts SAPE substrate to its peroxidized form by facilitating the addition of a hydroperoxyl- group at the C15 position of the ETE sn-2 chain at the catalytic site. Thus, the presence of an O2 molecule close to C15 is necessary to initiate the process. Here, we observed that the positioning of O2 at C15 was disrupted by the interference of NO• radicals. Figure 3 illustrates such an observation. Panels a–i display a series of snapshots showing how NO• approaches the region originally occupied by an O2 molecule, to displace O2, and position itself near the C15 atom for extended durations (>50 ns), until the departure of O2 from the catalytic site, consistent with the experimentally observed ability of NO• to inhibit the peroxidation of SAPE.

The panel j in Figure 3 indicates the 15LOX-2 residues (ordinate) that make successive contacts with NO• and O2 molecules, starting from first encounters with O2 (left panel, 40 < t < 75 ns, shaded in green), followed by the arrival of NO• to the catalytic site and the competition between NO• and O2 for binding the catalytic pocket residues (both panels, 80 < t < 140 ns; shaded in orange (left) and yellow (right)), finally ending with the dislocation of O2, while NO• remained stabilized near C15 and Fe3+ ion. Among residues ligating (successively or competitively) both O2 and NO• during this trajectory, we note the highly conserved residues P365 and E369 in site S2, and the hydrophobic cluster V603, A606, L609, and L610, revealing that O2 and NO• channeled to the catalytic site through Entrance 3 in this case.

#### *2.5. Selected Residues Stabilize O2 and NO*• *Near 15LOX-2 Catalytic Site in a Substrate-Dependent Manner*

We further carried out a statistical analysis of four independent runs to identify the 15LOX-2 residues which made frequent and long-lived contacts with O2/NO• molecules (see Methods), among those located within 20 Å of the Fe3+ ion. Supplementary Figure S4a displays such residues. Panels b–d list the PEBP1-bound 15LOX-2, free 15LOX-2 and PEBP1 residues, respectively, that make contacts for extended durations with O2 and NO•, and the corresponding number of runs in which such contacts are observed. Clearly 15LOX-2 residues at Entrances 1 and 3 (*E1* and *E3*) are detected among them, along with sites *S1* and *S2* residues, as indicated by the bars between panels b and c. Of particular importance is the long helix α12-14 (F365-L379) comprising both *S1* (L374, L379) and *S2* (F365, E369) residues also noted [13] in the absence of the phospholipid substrate. Interestingly, our recent computational analysis of a dataset of 88 crystal structures resolved for lipoxygenase showed that this particular region acts as a strong effector of allosteric signals [22]. The highly conserved WxxAK motif (W353-K357) shared by other LOXs [22] also takes part in the same region. Comparison of panels b and c shows that certain residues (E168, L172-A179 (except for I174), and F219-P223) interact with O2 and NO• in the absence of PEBP1 but become inaccessible upon PEBP1 binding.

#### *2.6. The Change in 15LOX-2 Structure Upon PEBP1 Binding Renders the Catalytic Site Accessible to Both O2 and NO*• *Molecules*

We observed that in the presence of PEBP1, both O2 and NO• co-localize within the catalytic site (Figure 4a) where they compete for a position near the C15 atom of SAPE (Figure 3, Movie 1). Remarkably, in the absence of PEBP1 and at the small concentration of NO•, the NO• molecules were unable to access the catalytic site (Figure 4b, red oval). This behavior consistently reproduced in independent runs (see Supplementary Figure S5) was due to a conformational change (opening or exposure of a binding site) between the α2 helix and the T166-A179 region stabilized upon PEBP1 binding (black arrows Figure 4c). When this site was not exposed (in the absence of PEBP1), O2 and NO• molecules were attracted to the hydrophobic sn-1 (stearoyl) chain of SAPE and to the 15LOX-2 residues L172-A179. At a higher NO• concentration, however, this effect was partially suppressed and a few NO• molecules took a place near C15 atom of SAPE in the catalytic site (black dotted oval Supplementary Figure S6). Simulations repeated with AA bound to 15LOX-2 (Supplementary Figure S7) showed that NO• molecules were able to access the catalytic site in the absence of the sn-1 chain of SAPE that otherwise sequestered the NO• molecules.

These simulations therefore lead to the following conclusions: (i) PEBP1 increases the affinity of 15LOX-2 to bind both O2 and NO• molecules to the catalytic site; whereas in the absence of PEBP1 and without higher NO• concentration, only O2 molecules access the active site, while NO• molecules preferentially bind to the phospholipid tail of SAPE; (ii) competitive binding of NO• to the catalytic site in the presence of PEBP1 (Figure 4a) is expected to interfere with the peroxidation activity of 15LOX-2 resulting in lipid nitrosylation, as well as decreased conversion of SAPE to SAPE-OOH; and (iii) in absence of PEBP1, the NO• molecules co-localize at the exposed sn-1 stearoyl chain and do not compete with the O2 molecules at the catalytic site (Supplementary Figure S5b).

#### *2.7. The Precise Positioning of SAPE for Peroxidation by 15LOX-2 Is Assisted by α2 Residues N181, Y185, and G189, and by E3/S1 Residues A416 and A606*

Lipid peroxidation by 15LOX-2 requires O2 molecule to be positioned in proximity (<7.5 Å) of the carbon C15 in the arachidonoyl chain of SAPE, while the C13 in the arachidonoyl chain of SAPE would approach (by <7.5 Å) the iron ion at the catalytic site. We analyzed the MD trajectories to examine whether such poses were sampled. We identified a total of 121 and 538 poses, respectively, for PEBP1-bound and -unbound 15LOX-2/SAPE (Figure 5a) that satisfied this requirement. The 15LOX-2 residues, G189, A416 (at *S1*) and A606 (at *E3*), were distinguished by their high tendency (75% of MD snapshots) to coordinate SAPE, in both systems. Furthermore, α2-helix residues N181, Y185, and G189 exhibited a high propensity to coordinate the substrate, Y185 playing a major role in the presence of PEBP1; and N181 in the absence. Additionally, PEBP1 D144 and P186 contributed to stabilizing the SAPE.

**Figure 4.** Conformational change in 15LOX-2 induced upon PEBP1 binding allows NO• binding to the catalytic sites in the presence of SAPE. Panels **a** and **b** compare the binding patterns of O2/NO• to 15LOX-2 in the presence (**a**) and absence (**b**) of complexation with PEBP1. The positions of NO• (blue dots) and O2 (red dots) sampled during MD snapshots are displayed. These refer to contacts (within 3.5 Å) between O2/NO• and (**a**) SAPE-bound 15LOX-2/PEBP1 complex and (**b**) SAPE-bound 15LOX-2 (with O2/NO• molecules within 7 Å from any SAPE atom). Both O2 and NO• molecules sample the catalytic site in the presence of PEBP1 (panel a, black oval). In the absence of PEBP1 the catalytic site exclusively harbors the O2 molecules (red oval). Snapshots from another run (Supplementary Figure S5) illustrate the reproducibility of the results. (**c**) Structural change induced by PEBP1 binding. Structural alignment of 15LOX-2 structure (after 150 ns simulation) in PEBP1-bound (dark grey) and unbound (magenta) forms with SAPE substrate (spheres) and accumulation of NO• molecules near the loop L172-A179 (orange) on the surface, are shown. Black arrows show the direction of conformational change in the α2 helix and the T166-A179 loop, providing access to both NO• (blue dots) and O2 (red dots). Other NO• molecules attracted by the stearic acid sn-1 chain of SAPE are hidden for better visualization.

**Figure 5.** Substrate-binding residues of 15LOX-2 in the presence and absence of PEBP1. (**a**) Interfacial contacts (within 3.5 Å) between the substrate SAPE and PEBP1/15LOX-2 (left panel) and 15LOX-2 (right panel) observed in MD simulations. (**b**) 15LOX-2 residues exhibiting the highest probabilities of contacts with SAPE. The ordinate shows the probabilistic occurrence (the number of counts divided by the total number of selected frames that satisfy the contact requirement). Two PEBP1 residues, D144 and P186, also observed to make frequent contacts are also included (indicated by cyan stars). The fluctuations in the conformation of SAPE during simulations are indicated by yellow sticks.

#### *2.8. In Silico Saturation Mutagenesis Analysis Confirms the Critical Role of Selected Residues Amongst Those Identified to Mediate O2/NO*• *Entry and Translocation to the Catalytic Site*

The present study points to several residues implicated in initial entry, channeling, and binding of O2/NO• taking part in *E1* or *E3*, or associated sites *S1, S2*, and other clusters (of mostly hydrophobic residues) paving the way to the catalytic site (Table S2). As a test of the potential impact of substitutions at those sites, we performed an in silico saturation mutagenesis analysis, using *Rhapsody* [23]. This machine learning tool scans all possible 19 substitutions at all the *N* amino acid positions of the protein to predict the so-called pathogenicity probabilities, a measure of the expected impact of specific substitutions on the protein function, varying from zero (neutral) to 1 (most damaging or pathogenic). The predictions are based on the evaluation of the sequence-, structure- and dynamicsproperties of 15LOX family proteins in comparison to the features observed in more than 20,000 missense variants used as learning dataset [23,24].

The results for residue segments of interest are presented as a heatmap in Figure 6a, color-coded from blue (neutral) to red (deleterious). The wild-type (WT) residues are shown in white. The ribbon diagram in panel b is also colored by the expected 'pathogenicity' from blue to red, mainly using the average values plotted under the heat map for each residue. The three curves therein represent the residue-averaged pathogenicity values predicted by Rhapsody (red), EVmutation [25] (green), and PolyPhen-2 [26] (blue). EVmutation takes rigorous account of residue (co)evolutionary properties and has proven to provide highly accurate results, while Polyphen-2 is broadly used for evaluating the effect of

mutations due to its applicability in the absence of structural data or sufficiently large multiple-sequence-alignments. These curves provide a consolidated view of the sensitivity of 15LOX-2 amino acids to mutations, the peaks describing the sites that would be most resistant to substitutions in general.

**Figure 6.** In silico saturation mutagenesis results for human 15LOX-2. (**a**) Pathogenicity probabilities for all substitutions, plotted as a function of residue number (abscissa) for all possible amino acid substitutions (ordinate). The probabilities are represented by a heatmap color-coded from blue (neutral) to red (deleterious). Colored horizontal bars (red, yellow, orange, and blue) along the upper abscissa denote the sites linked to specific functions (see labels under the map, Figure 2 and Table S2). Those distinguished by highly deleterious response to substitution are labeled. Black dashed box highlights the conserved WxxAK motif. The curves in the panel under the map indicate the sensitivity of a given residue to any mutation, as predicted by Rhapsody (red curve), EVmutation (green), and PolyPhen-2 (blue). Entrance 2 (on α2 helix) is broadly neutral to substitutions and not included in the heatmap. (**b**) Ribbon diagram of 15LOX-2/SAPE color-coded by pathogenicity probabilities (if mutated). The regions shown in space-filling representation and labeled (pointed by the black/grey arrows) are the entrances E1–E3. Catalytic residues are displayed in green and SAPE as cyan-red-blue sticks. (**c**) Close-up view of the catalytic site with bound substrate coordinated by four residues, L374, L420, L610, and N413, whose substitutions would be highly damaging to function. Yellow arrow points to the α12-14 scaffolding helix.

As expected, substitutions at the catalytic residues H373, H378, H553, and I676, and at the WxxAK motif were deleterious, irrespective of the type of amino acid substitution. Of interest is, however, to see whether (or which of) the residues involved in mediating the interactions with the ligand (O2/NO•) or substrate (SAPE) are predicted to be critical to function. Our analysis in fact revealed that the residues labeled along the upper abscissa, written in boldface in Table S2, to be intolerant to mutations. These residues include P595

and A599 lining *E3*, W158-P159 and I421 at *E1*, and N413, R417, L421, and L374 at site *S1*. Among them we note that some play a dual role of coordinating the substrate SAPE too (L374, N413, L420, and L610; see Figure 6c). In contrast, the α2-helix residues A177- G199 (*E2* residues and binding interface for PEBP1) are found to be tolerant to mutations (Figure 6b), as well as the surface-exposed G189, even though it participates in 70% of the interactions with SAPE.

Residues located along the helix α12-14 (K350-Q391) are particularly sensitive to substitutions. Examination of the structure shows that this region, composed of a long helix (with disruptions at two positions, hence the labeling as α12, α13, and α14) spans the entire structure at the center, making contacts with both *E1* and *E3* residues and lining the catalytic pocket (Figure 6b, dotted black box; see also Supplementary Figure S1). Its scaffolding role and multiple contacts appear to be critical for maintaining the stability and functionality of the enzyme.

#### *2.9. Identification of Nitrosylated PE Species in Cells Treated with NO*• *Donors*

Encouraged by these results, we next examined whether the nitrosylated PE products are formed in cells in which ferroptosis is inhibited by NO•. We had previously shown that RAW 264.7 M2 macrophages are susceptible to ferroptosis when their phospholipid hydroperoxide-specific glutathione peroxidase is inhibited by RSL3 [17]. In these cells, ferroptosis induced by RSL3 is suppressed by two inhibitors, Ferrostatin-1 and DTPA NONOate. Ferrostatin-1 suppresses ferroptosis by inhibiting the 15LOX-2/PEBP1 complex and through radical trapping antioxidant action. DTPA NONOate is a donor of NO• [3]. We had earlier shown that rescue from ferroptosis by DTPA NONOate is associated with a reduction in HPETE-PE contents [17]. Both Fer-1 and DTPA NONOate rescued from RSL3 induced ferroptosis with approximately similar effectiveness (Figure 7a). We identified two nitrosylated PE products 1-SA-2-ETE(-NO)-PE and 1-OA-2-ETE(-NO)-PE with m/z values of 795.541 and 793.530, respectively (Figure 7b). The precursors for these nitrosylated lipids, 1-SA-ETE-PE and 1-OA-ETE-PE, are the two most abundant ETE-containing PE species in cells. There was about 10-fold excess of these nitosylated PE species in cells treated with DTPA NONOate compared to Fer-1 treated cells (Figure 7c). Though these amounts are very low and we do not have an appropriate standard to quantify these nitrosylated PE species, our findings point to the presence of NO• in the close proximity of carbon-centered radicals formed by lipoxygenase.

**Figure 7.** Formation of nitrosylated PE during the inhibition of ferroptosis by NO• (**a**) Bar graph showing the decrease in cell death upon addition of ferroptosis inhibitor Ferrostatin-1 and DTPA NONOate in RAW 264.7 M2 macrophages treated with RSL3. Data are mean ± SD, *n* = 3 and no statistical significance was observed. (**b**) Mass spectrum showing the presence of two nitrosylated PE species, 1-OA-2-ETE(-NO)-PE (top panel) and 1-SA-2-ETE(-NO)-PE (bottom panel) in RAW 264.7 M2 macrophages treated with RSL3+Fer-1 (bottom section) and RSL3+DTPA NONOate (top section). Inset shows the magnified spectrum. **(c)** Quantities of 1-OA-2-ETE(-NO)-PE and 1-SA-2-ETE(-NO)-PE in RAW 264.7 M2 macrophages treated with RSL3+Fer-1 and RSL3+DTPA NONOate. Data are mean ± SD, *n* = 3 for RSL3+Fer-1 cells and 6 for RSL3+DTPA NONOate cells. *p* values are calculated using two-way ANOVA followed by Sidak post-hoc test.
