Stress pathway outputs are encoded by pH-dependent clustering of kinase components

Didan, Yuliia; Ghomlaghi, Milad; Nguyen, Lan K.; Ng, Dominic C. H.

doi:10.1038/s41467-024-50638-w

Download PDF

Article
Open access
Published: 05 August 2024

Stress pathway outputs are encoded by pH-dependent clustering of kinase components

Nature Communications volume 15, Article number: 6614 (2024) Cite this article

906 Accesses
1 Altmetric
Metrics details

Subjects

Abstract

Signal processing by intracellular kinases controls near all biological processes but how signal pathway functions evolve with changed cellular context is poorly understood. Functional specificity of c-Jun N-terminal Kinases (JNK) are partly encoded by signal strength. Here we reveal that intracellular pH (pHi) is a significant component of the JNK network and defines signal response to specific stimuli. We show pHi regulates JNK activity in response to cell stress, with the relationship between pHi and JNK activity dependent on specific stimuli and upstream kinases activated. Using the optogenetic clustering tag CRY2, we show that an increase in pHi promotes the light-induced phase transition of ASK1 to augment JNK activation. While increased pHi similarly promoted CRY2-tagged JNK2 to form light-induced condensates, this attenuated JNK activity. Mathematical modelling of feedback signalling incorporating pHi and differential contributions by ASK1 and JNK2 condensates was sufficient to delineate signal responses to specific stimuli. Taking pHi and ASK1/JNK2 signal contributions into consideration may delineate oncogenic versus tumour suppressive JNK functions and cancer cell drug responses.

Light-gated integrator for highlighting kinase activity in living cells

Article Open access 06 September 2024

Single-particle imaging of stress-promoters induction reveals the interplay between MAPK signaling, chromatin and transcription factors

Article Open access 23 June 2020

An intrinsic temporal order of c-JUN N-terminal phosphorylation regulates its activity by orchestrating co-factor recruitment

Article Open access 17 October 2022

Introduction

Signal transduction by intracellular kinases integrates and processes biochemical information generated by a multitude of cell-intrinsic and extrinsic stimuli. This is required to co-ordinate highly specific cell responses to their changing environments. The c-Jun N-terminal kinases (JNKs) are conserved members of the mitogen-activated protein kinase (MAPK) family, characterized by activation through a tiered phosphorylation cascade (MAP3K → MAP2K → MAPK) and serve as central nodes within these signal regulatory networks¹. The JNK pathway is activated by diverse sets of stimuli, including growth factors, cytokines, proteotoxic, environmental and metabolic stress². Following activation, JNK promotes a wide range of cellular responses, including cell fates that are often considered mutually exclusive, such as survival versus programmed death, and proliferation versus differentiation². In cancer, for example, both oncogenic and tumour suppressor roles for JNK signals are well recognized^3,4. The kinetics of JNK activation is a key mechanism that delineates between seemingly contradictory functions of the pathway. Sustained activation of JNK leads to cell cycle arrest and programmed death, while transient activation of the pathway is associated with survival and proliferation⁵. This raises the fundamental question of how JNK activation is changed in different cell contexts.

As the terminal kinases within a tiered cascade, JNK MAPKs receive signals from multiple upstream MAP3K activated by specific stimuli. For example, MLK3 signals primarily to JNK following activation by the inflammatory cytokine, TNFα⁶, while ASK1 is preferentially activated by oxidative stress⁷ and ASK3 responds to hyperosmolarity⁸. The signal transduction mediated by distinct MAP3Ks may result in differential engagement of feedback and feedforward loops, ultimately defining dynamic JNK signals and downstream cellular responses to stimulation. The organization of signal proteins within the cytosolic environment is also a critical determinant of how signals are propagated through pathways. An emerging concept is that signal networks may further evolve with intracellular pH (pHi) due to altered protonation state and regulation of higher-ordered signalling complexes^9,10.

Intracellular pH (pHi) is kept tightly regulated within a narrow range under physiological conditions to guard against excessive acidification or alkalinization of the cytosol that would be detrimental to protein function¹¹. However, it is now evident that small fluctuations in pHi regulate fundamental processes^12,13,14, and changes in pHi further reflect pathological cell states¹⁵. Notably, an increasing body of evidence supports a direct role of pHi in regulating signal transduction. pHi influences the stability of β-catenin and subsequent activation of the Wnt pathway¹⁰. Similarly, pHi-dependent conformational changes in Gα subunits are important determinants of G-protein coupled receptor-initiated signalling¹⁶. Fluctuations in pHi accompany the programmed death and proliferation of cancer cells, with JNK signalling being an important regulator of both processes. Given that phosphorylation networks regulate ion channels and transporters that determine cytosolic pH^17,18, pHi may represent embedded components of kinase regulatory networks that warrant more in-depth consideration. Moreover, the cross-regulation between protein and proton [H⁺] components within signalling networks is largely unexplored.

Here, we utilized live-cell biosensor measurements to reveal a tight association between JNK activity and changes in pHi in response to stress stimulation. We demonstrated that pHi was sufficient to regulate JNK signal output, and this was mediated through pH-dependent clustering and altered signal transduction of tiered kinases embedded within the JNK pathway. Interestingly, pH-regulated ASK1 at the MAP3K tier contributes differently to signal transduction compared to pH-dependent light-induced phase transitions of JNK2. Our mathematical model, which incorporates signal feedback on JNK activation, pHi, and the differential contribution by kinase condensates, accurately predicted JNK pathway responses to specific stimuli and explained how different JNK signal strengths are generated. Our study helps delineate the contextual functions of JNK signals, such as in drug-resistant cancers where JNK has been reported to serve dual functions.

Results

Cytosolic pH exerts influences stress-stimulated JNK signalling

An analysis of the Genomics of Drug Sensitivity in Cancer (GDSC) database¹⁹ indicated that cancer cell type sensitivity to JNK-targeting drugs was consistently negatively correlated with pHi (Supplementary Fig. S1A). There may be several reasons underlying this association, but our observations suggest a potential link between JNK signal transduction and pHi. This prompted us to investigate this more closely through live-cell measurements of pHi and JNK activity. Our studies utilized a fluorogenic dye (pHrodo) and a genetically encoded protein biosensor (ClopHensor) as two independent sensors of pHi in live cells in conjunction with a JNK activity reporter (JNK-KTR^20,21). pHrodo fluorescence is increased as pH is decreased, and ClopHensor is a pH-sensitive EGFP mutant that reports on elevated pH with an increased ratio of green to cyan fluorescence (Supplementary Fig. S1B)²². pHi measures were taken concurrently with the nuclear to cytoplasmic translocation of JNK-KTR to report on the JNK activation state^20,21 (Supplementary Fig. S1C). Our measures with pHrodo and ClopHensor (Supplementary Fig. S1D–H) were carefully calibrated against internal pH standards to enable direct comparison of pHi in different cell contexts and between sensors. As JNK activation is typically associated with stress responses, we first evaluated pHi and JNK activity responses to proinflammatory TNFα as a prototypical stress activator of the pathway²³. TNFα treatment of HEK293 cells led to JNK activation and modulation of pHi (Fig. 1A). Specifically, pHi decreased as JNK activity rapidly increased following TNFα treatment (Fig. 1B). After reaching peak JNK activation at around 25 mins, JNK activity decreased while pHi increased over time (Fig. 1B). Thus, JNK activity was inversely associated with changes in pHi following TNFα stimulation (Fig. 1C). In an expanded analysis of a panel of immortalized cell lines with measured pHi ranging between 6.9 to 7.6 (Supplementary Fig. S1E, H), we observed that the dynamics of JNK activation in response to TNFα varied markedly between different cell types (Fig. 1D). TNFα-stimulated JNK signals appeared to be enhanced in cells with comparatively acidic pHi such as MCF7 but dampened in cells with higher pHi such as U2OS (Fig. 1E and Supplementary Fig. S2A). Therefore, the strength of JNK signals generated by TNFα was negatively correlated with pHi both between different cell types and within the same cells over time.

**Fig. 1: Intracellular pH (pHi) modifies stress-stimulated activation of JNK MAPK.**

We next investigated whether pHi was similarly associated with JNK activity stimulated by hyperosmolarity, a physical stressor and robust activator of the JNK pathway²⁴. Surprisingly, we found that pHi fluctuated in close association with JNK activity in HEK293 cells exposed to hyperosmolarity (Fig. 1F). In response to osmotic stress, pHi increased alongside JNK activation, peaked, and then subsequently decreased as JNK signals were deactivated (Fig. 1G, H). Similarly, we observed a strong positive correlation between cell type-specific pHi and JNK signal strength stimulated by hyperosmolarity (Fig. 1I, J and Supplementary Fig. S2A). We further evaluated phosphorylation of JNK (pT183/pY185), indicative of activation state, by immunoblotting and confirmed stress-stimulated JNK activation in cell lines closely matched our live-cell KTR measurements (compare Fig. 1K, L with 1E and J). Compared to TNFα and osmotic stress, JNK activity stimulated by the ribotoxic stressor, anisomycin, was not as closely correlated to pHi in a time-course study, particularly during early time points as JNK activity increased (Supplementary Fig. S2B, C). In addition, we did not observe a clear correlation between cell type-dependent pHi with anisomycin-stimulated JNK activity (Supplementary Fig. S2D–F). Thus, stress-stimulated JNK signalling was closely associated with pHi, although the relationship between JNK activity and pHi appeared to be stimulus-dependent (Supplementary Fig. S2G).

To determine if a change in pHi was sufficient to alter JNK activity, we utilized pretreatment with chemicals widely used to modulate pH. As individual treatments may have effects independent of pH, we employed a diverse panel of pH modulators to both raise and lower pHi. We treated HEK293 cells, which had pHi in the middle of the physiological range and confirmed that NaHCO₃ and NH₄Cl treatment led to cytosolic alkalinization or increased pHi while S0859, DIDS, EIPA, Cariporide, and BafA resulted in acidification or reduced pHi (Fig. 1M). We then assessed the effect of chemically modulated pHi on JNK activity in response to our three candidate stress stimuli (Supplementary Fig. S3A–C). When JNK activity was plotted against pHi from all treatments, we found that acidification and alkalinization of the cytosol led to a predictable shift in JNK signalling capacity, which reinforced a strong inverse relationship between pHi and TNFα induced JNK activity and an opposite relationship for sorbitol (Fig. 1N). While a clear relationship was not evident when comparing cell types, we found that anisomycin-stimulated JNK activity was enhanced under more alkaline conditions similar to sorbitol (Fig. 1N). These studies reveal pHi influence over the strength of JNK activation following diverse stress stimulation, including ribotoxic stress.

We checked the effect of pH modulation on the phosphorylation of JNK and its downstream target, cJun, by immunoblotting. Chemicals that increased pHi, in particular NaHCO₃, augmented sorbitol- but not TNFα-stimulated JNK and cJun phosphorylation (Supplementary Fig. S4). Conversely, chemicals that decreased pHi (EIPA and DIDS) augmented TNFα but not sorbitol-stimulated JNK and c-Jun phosphorylation (Supplementary Fig. S4). A correlation analysis indicated pHi was positively associated with sorbitol-stimulated phosphorylation of JNK and cJun but negatively correlated with TNFα stimulation (Supplementary Fig. S4M, N) in good agreement with our imaging studies (Fig. 1N). The effect of NaHCO₃ on anisomycin-stimulated JNK phosphorylation appeared similar to sorbitol (Supplementary Fig. S4E, F), but unlike KTR findings, an association with pHi was not evident from immunoblots (Supplementary Fig. S4O). In general, the strength of associations (R-values) was not as high compared to KTR measures, possibly as immunoblots were limited to reporting on a single time point of stimulation. The more pronounced effects of NaHCO₃ on JNK phosphorylation may also be related to previously described effects of bicarbonate treatment on both extracellular and cytoplasmic pH²⁵. We also took this opportunity to evaluate p38^MAPK as a related stress-activated pathway. Our stress treatments stimulated a much reduced extent of p38^MAPK phosphorylation compared to JNK with weak to non-existent activation by TNFα (Supplementary Fig. S4). Thus, while our stress treatments activated intracellular pathways that lead to JNK activation, p38^MAPK responded selectively to hyperosmolarity and ribotoxic stress.

Cytosolic pH is regulated by stress-stimulated JNK signalling

Given previous reports that JNK may signal to protein targets that regulate pHi^26,27, we next looked at the consequence of targeting JNK on pHi. We utilized siRNA to deplete JNK1 or JNK2 individually or in combination (JNK1/2) and confirmed protein isoform-specific knockdown in HEK293 cells (Supplementary Fig. S5A). Firstly, we observed in control siRNA transfected cells that TNFα-stimulated JNK activity and pHi changes (Supplementary Fig. 2A) were not as tightly-associated compared to untransfected cells (Fig. 1B), which we attribute to liposome transfection. Nevertheless, TNFα-stimulated JNK activity (KTR) decreased with increasing amount of siRNA and was partially inhibited by JNK1 or JNK2 knockdown (Fig. 2A, B). Following JNK depletion, we observed pHi was elevated compared to control siRNA following TNFα stimulation, particularly at later time points and found similar effects regardless of the JNK isoform targeted (Fig. 2A, B). JNK activity was inversely associated with changes in pHi which tended to increase with reduced JNK activation from siRNA targeting (Fig. 2C). This negative association between JNK activity and pHi was similar to our findings with pHi modulation (Fig. 1). In response to hyperosmolarity, we found that siRNA depletion of JNK had the opposite effect on pHi (Fig. 2D–F). siRNA depletion of JNK1 or JNK2 led to reduced sorbitol-stimulated kinase activity and a decrease in pHi compared to control siRNA (Fig. 2D, E). In the context of hyperosmolarity, the extent of pHi decrease was closely associated with reductions in JNK activity (Fig. 2F). This was in agreement with the positive association between sorbitol-stimulated JNK activity and chemically modified pHi (Fig. 1). However, in response to anisomycin, the effect of JNK targeting on pHi mirrored TNFα-stimulation (Supplementary Fig. S5B–D) which was dissimilar to our findings when modifying pHi (Fig. 1). Finally, the effect of double depletion of JNK1 and JNK2 on stress-stimulated changes in pHi was not appreciably different to an isoform-specific knockdown (Supplementary Fig. S5E–G). Taken together, these studies reinforce a cross-talk between JNK signals and pHi. The lack of concordance between stress stimuli suggests signal context-dependent mechanisms at play that determine the specific nature of pHi signal interactions with the JNK pathway.

**Fig. 2: JNK depletion alters stress-stimulated change in pHi.**

Stress-stimulated clustering of JNK pathway kinases is regulated by pHi

As pH alters the charge states of proteins and their interactions within the crowded cytosol²⁸, we investigated the subcellular organization of JNK pathway signal components to gain insights into stress-responsive and pH-regulated signalling. We focused our attention on JNK2 and ASK1 as signal proteins in the JNK pathway reported to assemble higher ordered structures to regulate signalling^29,30. While ASK1 is implicated in transducing a broad range of stress signals, we found that pretreatment of HEK293 cells with an ASK1-specific inhibitor attenuated sorbitol and, to a lesser extent, anisomycin-stimulated JNK activity but did not change JNK activation in response to TNFα (Supplementary Fig. S6A, B). Similarly, ectopic expression of JNK2 enhanced JNK-KTR measured activity induced by all stress stimuli tested, while exogenous expression of ASK1 augmented sorbitol and anisomycin activation of JNK, but did not increase JNK activity in response to TNFα (Supplementary Fig. S6C). Moreover, ASK1 signals to p38^MAPK in addition to JNK, and we observed p38^MAPK phosphorylation in response to sorbitol and anisomycin but not TNFα (Supplementary Fig. S4). This indicates that ASK1 transduces stress signals to JNK in response to hyperosmolarity and ribotoxic stress but is less involved in TNFα-stimulated responses.

Our live tracking of mCherry-tagged proteins in HEK293 cells revealed that a proportion of JNK2 and ASK1 formed visible puncta or foci in the cytosol prior to stress (Fig. 3A). In response to each stress treatment tested, the number of JNK2 foci were reduced as JNK activity (KTR) increased and the number of JNK2 foci increased as JNK activity decreased following peak activation (Fig. 3B–D). Similar to JNK2, the number of ASK1 foci was inversely associated with the JNK activation state following TNFα stimulation (Fig. 3E, F). Interestingly, in response to anisomycin or sorbitol, ASK1 foci numbers increased and decreased in alignment with JNK activation and deactivation, respectively (Fig. 3G, H). Thus, while stress-induced assembly of JNK2 into cytosolic foci was negatively correlated with JNK activity, the formation of ASK1 puncta was positively correlated with JNK activity in hyperosmotic and ribotoxic stress contexts that signal through ASK1 (Fig. 3I). In comparison to JNK2 and ASK1, ectopically expressed mCherry-tagged JNK1 did not readily form foci (Supplementary Fig. S6D) and the puncta present were not closely associated with stress-stimulated JNK activity (Supplementary Fig. S6E–G).

**Fig. 3: Assembly of JNK2 and ASK1 cytoplasmic foci in response to stress and their association with pHi.**

We next compared the propensity for exogenously expressed mCherry-tagged JNK2 and ASK1 to form cytoplasmic foci in our panel of cell lines with differing pHi. The expression levels of JNK2 and ASK1, as determined by fluorescence intensity, were not a strong predictor of the propensity to form cytoplasmic foci in unstimulated cells (Supplementary Fig. S7). Instead, we observed increased numbers of JNK2 and ASK1 foci in unstimulated cells with comparatively more alkaline cytosols (e.g., U2OS), with the number of foci strongly correlated positively with pHi (Fig. 3J–M). In contrast, the number of JNK1 foci in various cell types was not associated with pHi (Supplementary Fig. S6H, J). Moreover, the formation of JNK1 foci was not strongly associated with pHi modulated by our panel of chemical pretreatments (Supplementary Fig. S6K–M). Thus, our next studies were focused on the pHi modulation of JNK2 and ASK1.

pHi modulates light-induced clustering of JNK2 and ASK1

To determine whether the formation of JNK2 and ASK1 foci altered signal transduction and was regulated by pHi, we utilized CRY2_clust³¹ fused to JNK2 and ASK1 followed by stable expression in HEK293 cells to enable blue light control of kinase clustering independent of stress stimulation. Blue light illumination induced clustering of CRY-JNK2 and CRY-ASK1 (Fig. 4A). CRY-JNK2 clusters were largely cytoplasmic while the majority of CRY-ASK1 clusters were atypically nuclear (Fig. 4A). Light-induced clustering of JNK2 and ASK1 was modestly attenuated or unaltered with pretreatment with pHi lowering compounds but potently enhanced with increased pHi due to NH₄Cl and NaHCO₃ (Fig. 4A–C and Supplementary Fig. S8). Light-induced JNK activity (KTR) in cells expressing CRY-JNK2 was enhanced as we decreased pHi with chemical pretreatments and attenuated as pHi was increased (Fig. 4D). In CRY-ASK1 cells, the impact of chemical pretreatment on light-stimulated JNK signal activity was not as clearly delineated between compounds that lowered versus increased pHi (Fig. 4E). Our studies indicate that kinase clustering was markedly enhanced at increased pHi, and this was strongly associated with attenuated JNK activity in cells expressing CRY-JNK2 (Fig. 4F–H). While CRY-ASK1 cluster assembly was enhanced at high pHi, this was moderately associated with enhanced light-induced JNK activity (Fig. 4I–K). The reduced correlation between CRY-ASK1 clusters and JNK activity (Fig. 4K), compared to CRY-JNK2 clusters (Fig. 4H), was due to contradictory effects of BafA and NaHCO₃ which augmented and dampened light-induced JNK activity (Fig. 4E) despite decreasing and promoting CRY-ASK1 clusters respectively (Fig. 4C). The reason for this is unclear but may be a result of competing effects of reduced condensation of endogenous JNK2 with BafA treatment that may promote JNK activation in CRY-ASK1 cells. In comparison, pHi effects on endogenous ASK1 were less of a factor in CRY-JNK2 cells as light directly stimulates JNK2. Taken together, our results indicate clustering of JNK2 appears to attenuate activity while ASK1 cluster assembly is associated with the activation of JNK. In control experiments, we repeated these studies with CRY2_clust variants of related MAP3Ks (CRY-MLK3 in Supplementary Fig. S9 and CRY-MEKK1 in Supplementary Fig. S10) but did not observe overt light-induced clustering of kinases nor an association between pHi and light-induced JNK activity in these stable lines (Supplementary Fig. S9, S10).

**Fig. 4: pHi regulates JNK2 and ASK1 subcellular organization and signal transduction.**

In seeking additional approaches to modulate pHi, we noted that the formation of dense monolayers also leads to increased pHi due to regulation by cell surface adhesive receptors³². We confirmed increased cytoplasmic and nuclear pH with increased seeding density of HEK293 cells (Supplementary Fig. S11A, B). The increased cell density led to enhanced clustering of CRY-JNK2 in response to light (Fig. 4L and Supplementary Fig. S11C). The light-induced formation of JNK2 clusters within dense monolayers was further promoted with NH₄Cl (Fig. 4L). In contrast, acidification with BafA attenuated CRY-JNK2 cluster formation (Fig. 4L and Supplementary Fig. S11D). As BafA disrupts lysosomal function³³ and NH₄Cl has indirect salt-driven effects¹¹ that may regulate kinase condensates³⁴ we also used nigericin, a carboxylic ionophore that promotes K⁺/H⁺ exchange³⁵, to alter cytosolic pH (pH_C) using extracellular media buffered at high (pHe 8.5) or neutral (pHe 7.0) pH. Using this approach we find CRY-JNK2 clusters were enhanced at high pH_C compared to slightly acidic pH_C defined by nigericin pretreatment (Fig. 4L and Supplementary Fig. S11D). Unlike JNK2, light induction of CRY-ASK1 clusters was not substantially increased at higher cell densities (Fig. 4M and Supplementary Fig. S11E) but was inhibited with BafA or nigericin at pHe 7.0 and augmented with NH₄Cl or nigericin at pHe 8.5 (Fig. 4M and Supplementary Fig. S11F). Thus, the increased pH from high-density cell cultures may not be sufficient to augment ASK1 foci formation compared to chemical alkalinization, which results in higher pH values (Supplementary Fig. S11A, B). As a control, the light-responsive oligomerization of the CRY2_clust domain alone was not enhanced by culturing cells at higher densities and was not substantially altered through chemical modulation of pHi (Supplementary Fig. S11G–I). A plot of cluster counts against measured pHi at various cell densities and chemical treatments indicates a strong positive association between pHi and the assembly of JNK2 and ASK1 clusters and confirms that CRY2_clust was not strongly correlated with pHi (Fig. 4N–P). Taken together with our control studies on related pathway components (Supplementary Figs. S6, S9 and S10), we find that JNK2 and ASK1 show a specific sensitivity to the cytosolic pH environment and enhanced assembly of intracellular foci under alkaline pHi. This was associated with attenuated or potentiated JNK activation with CRY-JNK2 and CRY-ASK1, respectively, indicating differential effects on signal flux through the pathway.

pH-regulated optogenetic phase transition of CRY-tagged JNK pathway kinases

The phase transition of kinases and self-assembly into protein condensates is sensitive to pH fluctuations and increasingly implicated in signal transduction³⁶. To determine if CRY-JNK2 and CRY-ASK1 foci demonstrated liquid-like properties, we performed live-cell tracking and observed light-induced CRY-JNK2 and CRY-ASK1 were highly dynamic with neighbouring clusters undergoing fusion upon contact (Fig. 5A). Light induction of CRY-JNK2 and CRY-ASK1 clusters were prevented with 1,6-hexanediol, (Fig. 5B, C) aliphatic alcohol that disrupts hydrophobic interactions involved in assembly of protein condensates in mammalian cells³⁷. In contrast, CRY2_clust clusters were not perturbed by 1,6-hexanediol (Fig. 5D) consistent with their formation of stable homo-oligomers³¹. In addition to increased numbers, CRY-JNK2 transitioned from elongated to spherical clusters with increased pHi, suggesting a pH-dependent change in their biophysical properties (compare pHi 7.0 with 7.5, Fig. 5E). Cytoplasmic CRY-JNK2 clusters at pHi 7.5 rapidly recovered 90% of maximum fluorescence intensity following photobleaching (Fig. 5E, F). In comparison, CRY-JNK2 clusters at pHi 7.0 did not recover fluorescence (Fig. 5E, F), supporting the notion that JNK2 underwent phase transitions at alkaline pHi. As CRY-ASK1 clusters formed predominantly in the nucleus, we compared fluorescence recovery with CRY2_clust located it in the nucleus as the appropriate control. CRY-ASK1 in the nucleus recovered 40% of maximum fluorescence following photobleaching (Fig. 5E, G). This indicates that CRY-ASK1, as with CRY-JNK2, assembled clusters that are dynamic and rapidly exchanged with their surrounding environment in the nucleoplasm. We note that a proportion of CRY-ASK1 remains immobile and does not fully recover fluorescence. This may be due to localization within the nucleus or the expanded domain organization of ASK1 compared to JNK2. As a control, CRY2_clust clusters, regardless of their localization in cytoplasm or nucleus, did not recover fluorescence following photobleaching (Fig. 5E–G). Thus, under alkaline pHi conditions, CRY-JNK2 and CRY-ASK1 appear to be assembled as protein condensates.

**Fig. 5: pHi regulates optogenetic assembly of JNK2 and ASK1 condensates.**

The formation of protein condensates is mediated by weak, multivalent interactions, commonly requiring low complexity or intrinsically disordered regions^28,37. SMART sequence analysis³⁸ identified regions of low complexity (aa 388–400 and 404–417) within the intrinsically disordered C-terminal tail of JNK2 (Fig. 5H), which we deleted to determine their contribution to the formation and dynamic properties of JNK2 condensates. The deletion of JNK2 aa 388–400, but not aa 404–417, reduced the light-induced formation of CRY-JNK2 clusters (Fig. 5I), while the deletion of either region was sufficient to blunt JNK2 fluorescence recovery following photobleaching studies (Fig. 5J). The SMART analysis identified a number of low-complexity motifs within the ASK1 sequence, and we selected three motifs (aa 178–202, 563–581 and 1291–1304) that were not located in structured regions for deletion (Fig. 5K). Deletions of each region individually were sufficient to reduce optogenetic stimulation of CRY-ASK1 clusters (Fig. 5L). However, fluorescence recovery of CRY-ASK1 clusters was specifically attenuated by Δ563–581 or Δ1291–1304 but not Δ178–202 (Fig. 5M). This suggests that these identified low-complexity regions are required for the formation of JNK2 and ASK1 condensates at permissive pHi. Moreover, blue-light stimulation of CRY-JNK2 triggered an increase in JNK activity with pHi at 7.0 but not when pHi was increased to 7.5 (Fig. 5N), consistent with the idea that JNK2 condensates are not conducive to signal transduction. Light-stimulated activation of CRY-JNK2 at pHi 7.5 was marginally improved with deletion of either Δ388–400 or Δ404–417 low-complexity motifs (Fig. 5O) and while C-terminal JNK2 truncation removing both motifs resulted in light-stimulated JNK activation similar to that observed at more neutral permissive pHi (Fig. 5N, O). Conversely, the deletion of Δ563–581 or Δ1291–1304 but not Δ178–202 in ASK1 reduced light-stimulated JNK activation (Fig. 5P) coinciding with reduced fluorescence recovery of CRY-ASK1 clusters (Fig. 5M). Thus, ASK1 low-complexity regions facilitate CRY tag-induced condensate formation at increased pHi to promote activation of JNK while cytosolic alkalinization promotes CRY-JNK2 condensates that are antagonistic to signalling.

Modelling of pH-regulated ASK1 and JNK2 predicts stress-context specific signal outputs

Our results suggest that the contrasting relationship between pHi and JNK activity with different stimuli may be explained by differential stress-activation of ASK1 and opposing effects of pHi on activities of kinases within the tiered JNK cascade. To examine this hypothesis, we employed a mathematical modelling approach to model the signal output from the pathway, incorporating pHi-regulated formation of JNK2 and ASK1 condensates and the differential ASK1 input caused by various stimuli (Fig. 6A). Our model builds upon the established tiered regulation of JNK by upstream MAP3Ks and negative feedback by DUSP1 phosphatases^39,40,41, which we further validated in our system using DUSP1 inhibitors, confirming that these resulted in enhanced stress-activation of JNK (Supplementary Fig. S12A). We adapted a rate equation (R7) previously utilized to model p38^MAPK oscillations³⁹ under the assumption that JNK induction of DUSP1 negative feedback is similarly conserved^40,41. We then incorporated in the model the pHi-assembled condensates that promote or attenuate the activity of ASK1 and JNK, respectively, and differential input by ASK1 in response to stimuli as indicated in our studies (Figs. 3, 4 and Supplementary Fig. S6). Our model also considered pHi as a dynamic signalling component that is increased by ASK1 and attenuated by JNK, based on the observed fluctuations of pHi in response to sorbitol and TNFα respectively (Fig. 1).

**Fig. 6: JNK cascade condensates and pHi predict signal output and functional consequences of the pathway in response to cell stress.**

The model, depicted in Fig. 6A, was formulated using ordinary differential equations (ODEs) and implemented in MATLAB. This effectively characterises the network interactions as a series of ODEs based on kinetic laws (see “Methods” for a detailed model description). We calibrated the model using the experimental data of JNK and pHi dynamics in response to TNFα, anisomycin and sorbitol (Fig. 6B). Simulation of stress treatments with the calibrated model recapitulated the experimentally observed changes in JNK activity, pHi and formation of ASK1 condensates (Figs. 6B,1, 3 and Supplementary Fig. S2). By validating our model using independent data, we found that it accurately predicted the effect of ASK1 inhibition on JNK activity stimulated by sorbitol but not TNFα (Fig. 6C). Using the model, we explored the mechanism regulating pHi dynamics in response to different stress stimuli. For this, we undertook a sensitivity analysis by systematically perturbing model parameters (see “Methods”) and simulating the effect on pHi dynamics in response to stress treatment. The results identified the strength of JNK-induced pHi acidification (Ki4) as the primary determinant of varying pHi dynamics after stress treatment. At low Ki4, pHi is increased by sorbitol treatment, but a higher Ki4 level switches this regulation from increasing pHi to a decrease (Fig. 6D).

Our model also accurately predicted the relationship between pHi and JNK activation and the strength under different stress types (Fig. 6E, F). Notably, simulations showed a negative correlation between pHi and JNK activity under TNFα treatment, but a positive correlation under sorbitol treatment. Next, we conducted a model sensitivity analysis to decipher the regulatory mechanisms affecting the pHi-JNK activity relationship (Fig. 6G and Supplementary Fig. S12B, C). The results indicate that the strength of JNK inhibition by JNK2con (ki3b) under sorbitol treatment, and the activation rate of JNK by ASK1 (kf3a) under TNFα stimulation are the key parameters in governing the pHi-JNK activity relationship (Supplementary Fig. S12B, C). For example, Fig. 6G depicts a shift from a positive correlation between pHi and JNK activity to a negative association, induced by increasing the ki3b parameter values (see Supplementary Fig. S12D for a similar analysis for kf3a). Together, our findings demonstrate how the dynamic interplay between ASK1 and JNK activities, alongside pHi-regulated kinase condensate formation, can explain the varied impact of pHi on stress-induced JNK pathway signalling.

We next hypothesized that knowledge of pHi and the kinases involved in the JNK network could help to predict the strength of the pathway’s signal output, which determines cellular outcomes. To test this idea, we turned to an analysis of cancer cell sensitivity to anti-tumour drugs in the GDSC database. Specifically, drugs like cisplatin, cetuximab, and fulvestrant, known to exert anti-tumour effects through JNK pathway activation, were considered^42,43. Although initial observations showed no clear correlation between drug sensitivity and pHi, taking into consideration ASK1 and JNK2 expression revealed reduced drug sensitivity at higher pHi among cancer cell types with low ASK1 expression relative to JNK2 (Fig. 6H). This is consistent with weak JNK signal outputs that are not conducive to programmed death induction. In contrast, cancer cells with high ASK1 levels relative to JNK2 showed a modest positive pHi correlation with drug sensitivity (Fig. 6H), suggesting that in this subgroup, elevated pHi enhances JNK activation leading to cell death. In contrast, correlation analyses pairing MLK3 or MEKK1, as MAP3Ks not sensitive to pHi (Supplementary Figs. S9, S10), with JNK2 poorly delineated cancer cell response to drugs (Supplementary Fig. S13). Thus, our findings suggest that pHi may help resolve the contextual function of JNK in cancer cell responses to anti-cancer therapy.

Discussion

The transduction of biochemical signals in cells relies on ions, nucleotides and phopsholipids⁴⁴. Our work indicates that cytosolic protons (H⁺) are intimately involved in signal transduction and should be considered an integral component of MAPK regulatory networks. While pHi is tightly regulated to guard against dramatic fluctuations that can damage proteins and trigger cell death, it is increasingly evident that localized physiological pHi changes have specific cell regulatory effects^12,13,14. Exocytosed protons participate in intercellular communication, including direct negative regulation of calcium channels during synaptic transmission⁴⁵, or positive signal feedback through stimulation of acid-sensing ion channels⁴⁶. In yeast, intracellular protons have been shown to act as second messengers in the glucose activation of protein kinase A⁹ and are coupled with GPCR activation for coincidence signal detection⁴⁷. Here, we demonstrated in mammalian cells that pHi changes are integrated within signal networks to determine the strength of JNK activation in response to specific stimuli. The JNK pathway has diverse, highly contextual functions that are dependent on how JNK signals are interpreted by cells². A pHi-regulated contextual JNK response may, therefore, serve to provide precise outcomes that are dependent on how signal inputs are generated.

We find that pHi altered signal transduction by modifying the biophysical properties and activity of specific kinases embedded within the JNK pathway. Optogenetic seeding of ASK1 and JNK2 condensates revealed that the relative alkalinization of the cytosolic environment promoted protein interactions. This was reminiscent of pHi-dependent conformational changes and oligomerization of Bcl-xL and Bax apoptotic proteins^48,49, mediated through altered protonation states of charged amino acids. Histidine residues are typically implicated as pH sensors due to the near-physiological pKa (~ 6.0) of their imidazole side group¹⁰. However, the pKa of other ionizable residues are known to shift dramatically towards physiological pH, and this is partly dependent on their position and topology within protein structures⁵⁰. Along these lines, the pH-dependent phase transition of a yeast prion protein, Sup35, is regulated by the protonation of a linear cluster of glutamic acid residues⁵¹. In this study, we identified regions in the C-terminal tail of JNK2 that were enriched with polar serine and negatively charged aspartate residues and were required for pH-dependent condensate formation upon optogenetic clustering. Similarly, low-complexity domains involved in CRY-ASK1 phase separation were comprised of positively charged lysines. This suggests that these domains respond to increased pHi to promote kinase interactions and potentially trigger condensate formation. The ectopic expression of JNK1 was not responsive to pHi. Importantly, the JNK1α1 spliceform tested lacks an extended C-terminal tail and the equivalent low complexity regions found in JNK2. This further indicates that these domains are involved in the pHi regulation of JNK proteins. However, we note that these sequences are reasonably conserved in the JNK1α2 spliceform with an extended C-terminal tail⁵². Therefore, it remains a possibility that the larger JNK1 spliceforms may undergo a pHi-regulated phase transition. We also cannot exclude the possibility that evolutionarily conserved histidines in JNK2 and ASK1 may further contribute to the observed kinase responses to pHi. The detailed biochemical mechanisms that underlie pH-dependent kinase responses require further characterization and may reveal new approaches to modulate signal transduction in cells.

In response to stress stimuli, pHi is known to regulate the phase transition of RNA-binding G3BP1, DEAD-Box helicase and poly(A)-binding proteins^53,54,55. However, the pH-dependent phase transitions of these proteins occur at pHi < 6.0, which is much lower than physiological pH. Under such conditions, phase separation represents an adaptive mechanism to sequester disrupted proteins into phase-separated granules to maintain protein function and cell survival⁵⁶. In contrast, CRY-JNK2 and CRY-ASK1 undertook light-induced phase transitions at relatively alkaline but physiological pHi between 7.5 and 8.0, which indicate mild changes in pHi function as signal regulators. In optogenetic studies, while light-stimulated assembly of CRY-JNK2 oligomers at relatively acidic pHi increased activity, the liquid-like phase transition at alkaline pHi inhibited signalling. So we speculate that CRY-JNK2 condensates are not conducive to trans-autoactivation²⁹ or are spatially separated from their downstream targets, including nuclear transcription factors⁵⁷. The phase transition of CRY-ASK1 has the opposing effect of enhancing downstream signalling to JNK. Thus, condensate formation may facilitate a quaternary state of ASK1 that enhances kinase activation⁵⁸ or concentrate ASK1 with its targets, perhaps by bringing together ASK1-binding signal scaffolds⁵⁹. This is consistent with phase-separated condensates acting as subcellular platforms that organize kinase signalling³⁶. We confirmed that JNK activation acts to regulate pHi in turn in a feedback loop with the effect of altering JNK levels on pHi also dependent on stress context. This is combined with opposing effects of phase separation on the activity of protein kinases organized within tiered cascades to create a complex signal feedback system where JNK activity is balanced with pHi to generate highly contextual signal outputs. Our mathematical modelling of the JNK pathway, which accounts for stimuli-induced alterations in pHi and consequent condensate formation of ASK1 and JNK2, was able to recapitulate experimental findings related to changes in JNK activity, pHi, and ASK1 condensate formation. Furthermore, it could predict JNK signal outputs with changed pHi under various stress conditions and identify potential regulatory interactions that explain how pHi influences JNK activity. However, it is likely that our current model may not fully capture how JNK, in turn, regulates pHi. Our model incorporates JNK-induced acidification of pHi. While this is consistent with some prior studies²⁷, JNK has also been reported to activate NHE1 to promote increases in pHi²⁶. Further characterization of JNK phosphorylation targets involved in pHi control would help refine our model and aid in the future delineation of the complex and diverse functions of JNK under different conditions.

Our studies suggest that the pHi relationship with ASK1/JNK2 may have a benefit in delineating signal regulation in cancer and predicting responses to anti-cancer therapy. JNK signals promote tumour growth and metastasis and mediate cell death in response to anti-cancer drugs⁴. Precise measures of pHi may delineate the role of JNK signalling in tumours, and an understanding of the positive or negative association between JNK signal capacity and pHi could be exploited therapeutically to prevent oncogenic signals while promoting JNK-killing of cancer cells in response to chemo- or radiotherapy. Future work will be required to advance these ideas and must take account of important considerations including the complex relationships between chemotherapeutic compounds and tumour pH. For example, prolonged cisplatin therapy causes pHi acidification in cancer cells, and fulvestrant antagonism of oestrogen receptors alters the expression of carbonic anhydrase family members that may modify pHi^60,61,62. Notably, the ability to maintain alkaline pHi delineates cancer cells that survive cisplatin treatment⁶⁰ consistent with the notion of reduced JNK activation under similar conditions.

More broadly, the hallmark inversion of the pH gradient in cancer cells suggests that inhibitors of proton pumps and exchangers may improve the chemosensitivity of tumours⁶³. However, trials have returned mixed results, with adverse patient outcomes reported when proton pump inhibitors were combined with kinase-targeting drugs or immune-checkpoint inhibitors^64,65. While this is in part due to drug-drug interactions⁶⁶, an improved understanding of the association between increased pHi and kinase signal regulation of cancer progression may help optimize pHi targeting as a therapeutic strategy. In addition to cancer, dysregulation of JNK signalling is linked to neurodegeneration, cardiac remodelling, metabolic and inflammatory disorders¹, where an improved understanding of JNK pathway interactions with pH may help decipher the contextual functions of the pathway.

Methods

Literature and database analysis

The Dataset for intracellular pHi measurements for Fig.6H and Supplementary Fig. S1A was generated from the previously published data and presented in Supplementary Table S1. Data for the efficiency of anticancer drugs (AS601245, JNK inhibitor VIII, JNK 9 L, cisplatin, cetuximab, and fulvestrant) as AUC of cell viability response to drug IC₅₀ plots were obtained from Genomics of Drug Sensitivity in Cancer (GDSC) database¹⁹ (www.cancerRxgene.org). Expression levels of JNK2, ASK1, MLK3, and MEKK1 in a panel of cell lines were obtained from the Human Protein Atlas⁶⁷ (www.proteinatlas.org).

DNA Constructs

Plasmids were constructed using standard restriction enzyme cloning. pPBbsr-JNKKTR-mCherry plasmid containing JNK activity reporter (JNK-KTR) was obtained from Addgene (#115493). mCherry in this plasmid was exchanged with iRFP713 (from Addgene #111510) fluorophore using NotI and SalI enzymes. ClopHensor plasmid was obtained from Addgene (#25938). JNK1α1, JNK2α2, ASK1, MLK3 and MEKK1 were amplified from pCMV-JNK1α1-Flag (a gift from M.A. Bogoyevitch), pCDNA3 Flag Jnk2a2 (Addgene #13755), pCMV6-Myc-MAP3K5 (ASK1, Origen #RC209913), pDONR223-MAP3K11 (MLK3, Addgene #23473), and pcDNA3MEKK1 (Addgene #12181) respectively. PCR amplified constructs were inserted into mCherry-CRY2clust (Addgene #105624), using BsrGI/HpaI sites to obtain mCherry tagged JNK2α2 and JNK1α1, and SgrAI/HpaI sites to obtain mCherry tagged ASK1 constructs. Amplified JNK2α2, ASK1, MLK3 and MEKK1 were inserted into mCherry-CRY2clust using BsrGI/BspEI, SmaI/HpaI, BamHI/MfeI, and BamHI/HpaI sites respectively to obtain CRY-JNK2, CRY-ASK1, CRY-MLK3 and CRY-MEKK1. Obtained constructs containing CMV promoter, mCherry, JNK2α2, ASK1, MLK3 or MEKK1, and CRY2_clust were then cloned into PiggyBac backbone (a gift from M. Jones) using SpeI/PmeI sites. Low complexity regions of JNK2α2 and ASK1 were deleted using Q5 Site-Directed Mutagenesis (NEB). Oligonucleotides used for cloning and site-directed mutagenesis are listed in Supplementary Table S2.

Cell Maintenance, transfection, and stable cell line generation

U87MG, MCF7, HEK293, SH-SY5Y, and U2OS cell lines were purchased from the American Type Culture Collection (ATCC). U87MG, MCF7, HEK293, and U2OS were cultured in DMEM (Gibco) supplemented with 10% FBS (Gibco) and 1% penicillin/streptomycin (Gibco) at 37 °C with 5% CO2 in a humidified incubator. SH-SY5Y was cultured in the same condition but with 15% FBS. For plasmids transfections, 200000 cells per well were seeded at the 8-well chamber slides (ibidi) and next day transfected with 0.5 µg plasmid and Lipofectamine (Invitrogen) at the ratio of 1:6. For siRNA-mediated JNK1 and JNK2 knockdown, cells were transfected with 1, 6 or 25 pmol ON-TARGET plus Human MAPK8 SMART pool siRNA (Dharmacon, L-003514-00-0020) and/or ON-TARGET plus Human MAPK9 SMART pool siRNA (Dharmacon, L-003505-00-0020) with 2 µl RNAiMAX (Thermo Fisher Scientific), according to the manufacturer’s protocol. As a control, cells were transfected with ON-TARGETplus Non-targeting Control Pool (Dharmacon, D-001810-01-20). Cells were used for imaging 24 h after transfection. For stable cell line generation, cells were seeded at the 12-well plate and transfected with plasmids (0.5 µg each) using Lipofectamine. For a generation of cells lines with CRY2_clust, CRY-JNK2, CRY-ASK1, CRY-MLK3, and CRY-MEKK1 cells were transfected with plasmids of interest and transposase (with the transposon to transposase ratio of 1:3) using Lipofectamine. 3 days after transfection, cells were treated with antibiotics according to resistance, 1 mg/ml G418 (Sigma-Aldrich) and/or 5 µg/ml blasticidin (Sigma-Aldrich). After 7 days of selection, cells were plated at the 96 well plates with a density of 1 cell per well. Obtained monoclonal cell populations with targeted proteins were cultured for 2 weeks prior to use in the experiments. The list of generated cell lines for this study is presented in Supplementary Table S3.

Stress and chemical treatments

8-well chamber slides were coated with fibronectin in PBS (2.5 µg/cm², Sigma-Aldrich) for 1 h. Cells were seeded at the coated chamber slides with a density of 200000 cells per well, and the next day, four hours before imaging, cells were stained with 1 µg/ml Hoechst 33342 for 30 min at 37 °C and washed 3 times for 10 min each with Live Cell Imaging Solution (Invitrogen). When applicable, cells were pre-treated with inhibitors (10 µM ASK inhibitor or 10 µM BCI (Sigma-Aldrich)) and then treated with 10 ng/ml TNFα (Sigma-Aldrich), 10 ng/ml anisomycin (Sigma-Aldrich) or 150 mM sorbitol (Sigma-Aldrich). For modulation of intracellular pH, cells were treated with 0.1 µM Bafilomycin A1 (Sigma-Aldrich), 50 mM NH₄Cl (Supelco) or 10 µM Nigericin (Invitrogen) 10 min before the imaging or treated with 2 µM EIPA (Sigma-Aldrich), 2 µM DIDS (Sapphire Bioscience), 100 µM S0859 (Sapphire Bioscience), 10 µM Cariporide (Sapphire Bioscience) or 25 mM NaHCO₃ (Sigma-Aldrich) 30 min prior imaging. Nigericin was added to cells by exchanging cell media with fresh media buffered at pH 7.0 or pH 8.5 and containing nigericin. For the duration of all imaging experiments, cells were kept at 37 °C and 5% CO2. Images of JNK-KTR translocation were obtained with a 633 nm laser, Leica DMi8 SP8. JNK activity was calculated as the ratio between intensities of fluorescence of JNK-KTR in the cytoplasm and in the nucleus (C/N ratio). Intensities and C/N ratio of JNK-KTR were measured using CellProfiler 4.0.7, data was then normalized and plotted in MATLAB R2021a. Normalized JNK-KTR represent values normalized against C/N ratios in cells prior to stress stimulation or light-illumination (i.e., relative change in KTR C/N values from the start of the experiment or 0 min). KTR area under curve (KTR AUC) values represent integrated JNK KTR signal values over the time duration of the experiment.

Immunoblotting

Following completion of cell treatments, protein extracts were prepared by lysis in RIPA buffer (50 mM Tris-HCl, pH 7.3, 150 mM NaCl, 0.1 mM EDTA, 1% (w/v) sodium deoxycholate, 1% (v/v) Triton X-100, 0.2% (w/v) NaF, and 100 μM Na₃VO₄) supplemented with protease inhibitors. After lysis, cell debris was removed by high-speed centrifugation (16,900 × g) at 4 ^oC. The concentration of protein in cleared lysates was determined by Bradford assay using serum albumin as a standard. Lysates were then diluted with Laemmli buffer before SDS-PAGE, transferred to PVDF membranes, blocked with 5% (w/v) skim milk, and immunoblotted with primary antibodies at the indicated dilutions (Supplementary Table S4). Proteins were detected using appropriate HRP-conjugated secondary antibodies with enhanced chemiluminescence. For reprobing, membranes were stripped using a mild buffer (200 mM Glycine, 0.1% w/v SDS, 1% v/v Tween 20, pH 2.2), washed with phosphate-buffered saline and Tris-buffered saline with 0.1% (v/v) Tween-20 before blocking and blotting with primary antibodies. Bands were imaged and quantified on either a BIO-RAD ChemiDoc MP or a LI-COR Odyssey Fc imaging system. Phosphorylated band densitometric values were normalized for changes in total protein levels, and stress-stimulated samples expressed as fold increase over appropriate controls without stress treatment.

Intracellular pH sensing

Intracellular pH in U87MG, MCF7, HEK293, SH-SY5Y, and U2OS cell lines was measured using the chemical dye pHrodo-Green AM (Invitrogen) and genetically encoded pH and chloride indicator ClopHensor. To measure intracellular pH with pHrodo, U87MG, MCF7, HEK293, SH-SY5Y, U2OS cells were seeded at the 8-well chamber slides with a density of 200000 cells per well and stained with 1 µg/ml Hoechst 33342 and pHrodo according to the manufacturer’s standard protocol (Invitrogen). At least four hours after staining, cells were imaged using a 488 nm laser, Leica DMi8 SP8. The intensity of pHrodo was used to calculate pH. Intracellular pH was also measured using the genetically encoded pH and chloride sensor ClopHensor. ClopHensor was transiently expressed in U87MG, MCF7, SH-SY5Y, and U2OS, and stably expressed in HEK293 cells. ClopHensor fluorescence was imaged with 458 nm (cyan emission) and 488 nm (green emission) laser, Leica DMi8 SP8. The ratio between fluorescent intensities at 488 nm and 458 nm was used to calculate pH. For pH quantification, calibration curves were built for all cell lines for pHrodo staining, and for ClopHensor expressed in HEK293 cells. To build calibration curves, intracellular pH was equilibrated to pH 4.5, 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, and 9.0. Cells were incubated in buffers of pH calibration buffer kit (Invitrogen), or additional pH buffers made by adjustments of pH with 1 mM NaOH (ChemSupply Australia), 10 μM of Valinomycin (Invitrogen) and 10 μM of Nigericin (Invitrogen) for 10 min before imaging. At least 200 cells were imaged for each pH value across 3-5 independent experiments for pHrodo, and at least 1000 cells per condition were imaged across 5-8 independent experiments for ClopHensor. pH for pHrodo stained cells was defined using linear regression analysis (MATLAB R2021a). The value for pHi for ClopHensor was calculated according to equation:$\,{\rm{pHi}}={\rm{pKa}}+\,\log \left(\frac{{\rm{RpH}}-{\rm{Ra}}}{{\rm{Rb}}-{\rm{RpH}}}\right)$, where pKa = 6.75, Ra = 5.8, Rb = 12, and RpH is a green-to-cyan ratio. Intensities of pHrodo and cyan-to-green ratio of ClopHensor were measured using CellProfiler 4.0.7, data was then plotted in MATLAB R2021a.

Optogenetic modulation of JNK2 and ASK1 clustering

HEK293 cells stably expressing CRY-JNK2, CRY-ASK1, CRY-MLK3, CRY-MEKK1, and CRY2_clust wear seeded at the 8-well chamber slides with a density 200000 cells per well (if other not specified). Mutated variants of CRY-JNK2 and CRY-ASK1 and controls were transiently expressed in HEK293 cells. To induce clustering of CRY-JNK2, CRY-ASK1, CRY-MLK3, CRY-MEKK1, and CRY2_clust, cells were illuminated with blue light (470 nm, 21 mW, Mightex) for the duration of the experiment which was up to 120 min. When applicable, prior to photoactivation, cells were treated with 0.1 µM Bafilomycin A1 (Sigma-Aldrich), 50 mM NH₄Cl (Supelco), 10 µM Nigericin (Invitrogen), 2 µM EIPA (Sigma-Aldrich), 2 µM DIDS (Sapphire Bioscience), 100 µM S0859 (Sapphire Bioscience), 10 µM Cariporide (Sapphire Bioscience), 25 mM NaHCO₃ (Sigma-Aldrich) or 1,6-hexanediol (Sigma-Aldrich) with a final concentration of 1% (v/v). Images of CRY-JNK2, CRY-ASK1, CRY-MLK3, CRY-MEKK1, and CRY2_clust clusters were acquired with a 561 nm laser, Leica DMi8 SP8. The number of clusters (counts) per cell was measured using CellProfiler 4.0.7, normalized, and plotted in MATLAB R2021a. Normalized counts represent relative change in cluster numbers from prior to stimulation (before light illumination) at the start of the experimental time course (0 min). The area under the curve (AUC) of normalized cluster counts are integrated values over the total duration of the experiment. Correlation analyses utilizing normalized cluster counts (Fig. 4F, I, N–P) were performed on a single time point where maximal cluster assembly was observed.

FRAP

FRAP assays were performed using Leica DMi8 SP8. Clusters were bleached with 80% intensity of 561 nm laser for 5 s, with the following image acquiring every 15 sec for 10 min. Intensities of fluorescence were extracted using ImageJ and fitted using MATLAB R2021a. The half-time of recovery (t_1/2) was calculated as the time required to recover half the maximum recovery intensity of fluorescence. Half of the maximum recovered fluorescence intensity was calculated as: I_1/2 = (I_sat + I₀)/2, where I_sat is the intensity of the bleached area after recovery at the maximum saturation point, and I₀ is the intensity of the bleached area immediately after bleaching.

Statistical & reproducibility

All statistical analyses were carried out with MATLAB R2021a. N values refer to independent experiments performed on new cells and over different days, with the exception of FRAP experiments, where N values indicated clusters/droplets across three independent experiments. In each experiment, cells or droplets were quantified in an automated fashion (CellProfiler 4.0.7), and averaged values were determined. Mean values are then derived from independent experimental repeats (n). Data were analysed using unpaired two-sided t tests. P-values < 0.05 were considered statistically discernible (*p < 0.05; **p < 0.001). No statistical method was used to predetermine the sample size. Cells were randomly allocated to each experiment but investigators were not blinded to allocation during experiments and outcome assessment.

Mathematical modelling

Model implementation. We developed a new kinetic, mathematical model of the JNK pathway and its regulation by pHi in response to stimulation by stress stimuli. The model schematic is given in Fig. 6A. The model was formulated using ordinary differential equations (ODEs) based on chemical kinetic laws using MATLAB (The MathWorks. Inc. 2020b) in conjunction with the IQM toolbox (https://iqmtools.intiquan.com/). The model was then coded and numerically solved to perform various simulations in MATLAB, utilising the SUNDIALS solver suite (https://computing.llnl.gov/projects/sundials).

Model fitting to experimental data. To ensure the model’s validity and predictive power, we fitted, i.e., calibrated, the model to experimental data. The data from Fig. 1B, G and Supplementary Fig. 6B were combined and utilized to fit the model. Model fitting involves computational estimation of the model kinetic parameters to minimize the discrepancy between model simulations and experimental data, quantified in the form of the below objective function J:

$$J(p)\,=\,\mathop{\sum }\limits_{j=1}^{M}\mathop{\sum }\limits_{i=1}^{N}\Bigg({\frac{{y}_{j,i}^{D}-{y}_{j}({t}_{i},p)}{D}}\Bigg)^{2}$$

Here, M denotes the number of experimental data sets used for fitting; N is the number of time points within each set; ${y}_{j}({t}_{i},\,p)$ indicates the simulation value of component j at the time point ${t}_{i}$ with parameter set $p$ and ${y}_{j,i}^{D}$ is the experimental data of component $j$ at time point ${t}_{i}$. D denotes the standard deviation of the experimental data.

We performed model fitting, specifically parameter estimation, using a genetic algorithm (GA) implemented with the Global Optimization Toolbox and the ‘ga’ function in MATLAB. The selection rules identify the individual solutions with the highest fitness values (i.e., elite solutions) from the current population. The elite count was set to 5% of the population size, and the crossover faction was set at 0.8. To determine the best-fitted parameter set, we conducted multiple GA runs with a population size of 400 and the number of generations set to 800. We provide the resulting best-fitted parameter sets in Supplementary Information, as well as in our GitHub folder below.

Sensitivity analysis

Model-based sensitivity analysis was employed to investigate the mechanism governing the correlation between intracellular pH (pHi) and JNK activity. To achieve this, we devised a metric that quantifies the relationship between pHi and JNK activity, defined as the slope of a simulated pHi-JNK dose-response curve. This curve was generated by varying pHi within a broad range (0.5 to 1.5), and the resulting JNK activity was determined using our model, as shown in Fig. 6F. We then applied this slope metric in the sensitivity analysis, where each model kinetic parameter was systematically altered within an extensive range (from 0.001 to 1000 times its nominal value), and the consequent effect on the slope metric was calculated. Finally, the parameters were ranked according to their impact, from the most significant to the least, as illustrated in Supplementary Fig. S12B C.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All data generated in this study are provided in the Supplementary Information and Source Data files. Source data are provided with this paper.

Code availability

For model reconstruction and reproducibility, a full set of the model reactions and their kinetic rates are given in Supplementary Table S5. They are also described in.txtbc and MATLAB files on Github [https://github.com/NguyenLab-IntegratedNetworkModeling/JNK_pHi] and published in Zenodo [https://doi.org/10.5281/zenodo.12633353].⁶⁸

References

Bogoyevitch, M. A., Ngoei, K. R., Zhao, T. T., Yeap, Y. Y. & Ng, D. C. c-Jun N-terminal kinase (JNK) signaling: recent advances and challenges. Biochim. Biophys. Acta Proteins Proteomics 1804, 463–475 (2010).
Article CAS Google Scholar
Davis, R. J. Signal transduction by the JNK group of MAP kinases. Cell 103, 239–252 (2000).
Article CAS PubMed Google Scholar
La Marca, J. E. & Richardson, H. E. Two-faced: roles of JNK signalling during tumourigenesis in the Drosophila model. Front. Cell Dev. Biol. 8, 42 (2020).
Article PubMed PubMed Central Google Scholar
Tournier, C. The 2 faces of JNK signaling in cancer. Genes Cancer 4, 397–400 (2013).
Article PubMed PubMed Central Google Scholar
Ventura, J.-J. et al. Chemical genetic analysis of the time course of signal transduction by JNK. Mol. Cell 21, 701–710 (2006).
Article CAS PubMed Google Scholar
Brancho, D. et al. Role of MLK3 in the regulation of mitogen-activated protein kinase signaling cascades. Mol. Cell. Biol. 25, 3670–3681 (2005).
Article CAS PubMed PubMed Central Google Scholar
Saitoh, M. et al. Mammalian thioredoxin is a direct inhibitor of apoptosis signal-regulating kinase (ASK) 1. EMBO J. 17, 2596–2606 (1998).
Article CAS PubMed PubMed Central Google Scholar
Watanabe, K., Umeda, T., Niwa, K., Naguro, I. & Ichijo, H. A PP6-ASK3 module coordinates the bidirectional cell volume regulation under osmotic stress. Cell Rep. 22, 2809–2817 (2018).
Article CAS PubMed Google Scholar
Dechant, R. et al. Cytosolic pH is a second messenger for glucose and regulates the PKA pathway through V‐ATPase. EMBO J. 29, 2515–2526 (2010).
Article CAS PubMed PubMed Central Google Scholar
White, K. A. et al. β-Catenin is a pH sensor with decreased stability at higher intracellular pH. J. Cell Biol. 217, 3965–3976 (2018).
Article CAS PubMed PubMed Central Google Scholar
Roos, A. & Boron, W. F. Intracellular pH. Physiol. Rev. 61, 296–434 (1981).
Article CAS PubMed Google Scholar
Spear, J. S. & White, K. A. Single-cell intracellular pH dynamics regulate the cell cycle by timing the G1 exit and G2 transition. J. Cell Sci. 136, jcs260458 (2023).
Article CAS PubMed PubMed Central Google Scholar
Srivastava, J. et al. Structural model and functional significance of pH-dependent talin–actin binding for focal adhesion remodeling. Proc. Natl. Acad. Sci. USA 105, 14436–14441 (2008).
Article ADS CAS PubMed PubMed Central Google Scholar
Ulmschneider, B. et al. Increased intracellular pH is necessary for adult epithelial and embryonic stem cell differentiation. J. Cell Biol. 215, 345–355 (2016).
Article CAS PubMed PubMed Central Google Scholar
Webb, B. A., Chimenti, M., Jacobson, M. P. & Barber, D. L. Dysregulated pH: a perfect storm for cancer progression. Nat. Rev. Cancer 11, 671–677 (2011).
Article CAS PubMed Google Scholar
Isom, D. G. et al. Protons as second messenger regulators of G protein signaling. Mol. Cell 51, 531–538 (2013).
Article CAS PubMed PubMed Central Google Scholar
Li, X., Augustine, A., Sun, D., Li, L. & Fliegel, L. Activation of the Na(+)/H(+) exchanger in isolated cardiomyocytes through beta-Raf dependent pathways. Role of Thr(653) of the cytosolic tail. J. Mol. Cell Cardiol. 99, 65–75 (2016).
Article CAS PubMed Google Scholar
Verkest, C., Diochot, S., Lingueglia, E. & Baron, A. C-Jun N-Terminal kinase post-translational Regulation of pain-related acid-sensing ion channels 1b and 3. J. Neurosci. 41, 8673–8685 (2021).
Article CAS PubMed PubMed Central Google Scholar
Yang, W. et al. Genomics of drug sensitivity in cancer (GDSC): a resource for therapeutic biomarker discovery in cancer cells. Nucleic Acids Res. 41, D955–D961 (2012).
Article PubMed PubMed Central Google Scholar
Miura, H., Kondo, Y., Matsuda, M. & Aoki, K. Cell-to-cell heterogeneity in p38-mediated cross-inhibition of JNK causes stochastic cell death. Cell Rep. 24, 2658–2668 (2018).
Article CAS PubMed Google Scholar
Regot, S., Hughey, J. J., Bajar, B. T., Carrasco, S. & Covert, M. W. High-sensitivity measurements of multiple kinase activities in live single cells. Cell 157, 1724–1734 (2014).
Article CAS PubMed PubMed Central Google Scholar
Arosio, D. et al. Simultaneous intracellular chloride and pH measurements using a GFP-based sensor. Nat. methods 7, 516–518 (2010).
Article CAS PubMed Google Scholar
Reinhard, C., Shamoon, B., Shyamala, V. & Williams, L. T. Tumor necrosis factor α-induced activation of c-jun N-terminal kinase is mediated by TRAF2. EMBO J. 16, 1080–1092 (1997).
Article CAS PubMed PubMed Central Google Scholar
Galcheva-Gargova, Z., Dérijard, B., Wu, I.-H. & Davis, R. J. An osmosensing signal transduction pathway in mammalian cells. Science 265, 806–808 (1994).
Article ADS CAS PubMed Google Scholar
Ritter, J., Doktor, H. & Benjamin, N. Paradoxical effect of bicarbonate on cytoplasmic pH. Lancet 335, 1243–1246 (1990).
Article CAS PubMed Google Scholar
Huc, L. et al. c-Jun NH2-Terminal kinase–related Na+/H+ exchanger isoform 1 activation controls hexokinase II expression in benzo (a) pyrene-induced apoptosis. Cancer Res. 67, 1696–1705 (2007).
Article CAS PubMed Google Scholar
Wang, X. et al. Carbonic anhydrase II regulates differentiation of ameloblasts via intracellular pH‐dependent JNK signaling pathway. J. Cell. Physiol. 225, 709–719 (2010).
Article CAS PubMed Google Scholar
Jin, X. et al. Effects of pH alterations on stress-and aging-induced protein phase separation. Cell. Mol. Life Sci. 79, 380 (2022).
Article CAS PubMed PubMed Central Google Scholar
Nitta, R. T., Chu, A. H. & Wong, A. J. Constitutive activity of JNK2α2 is dependent on a unique mechanism of MAPK activation. J. Biol. Chem. 283, 34935–34945 (2008).
Article CAS PubMed PubMed Central Google Scholar
Trevelyan, S. J. et al. Structure-based mechanism of preferential complex formation by apoptosis signal–regulating kinases. Sci. Signal. 13, eaay6318 (2020).
Article CAS PubMed Google Scholar
Park, H. et al. Optogenetic protein clustering through fluorescent protein tagging and extension of CRY2. Nat. Commun. 8, 30 (2017).
Article ADS PubMed PubMed Central Google Scholar
Galkina, S., Sud’Ina, G., Dergacheva, G. & Margolis, L. Regulation of intracellular pH by cell‐cell adhesive interactions. FEBS Lett. 374, 17–20 (1995).
Article CAS PubMed Google Scholar
Yoshimori, T., Yamamoto, A., Moriyama, Y., Futai, M. & Tashiro, Y. Bafilomycin A1, a specific inhibitor of vacuolar-type H (+)-ATPase, inhibits acidification and protein degradation in lysosomes of cultured cells. J. Biol. Chem. 266, 17707–17712 (1991).
Article CAS PubMed Google Scholar
Morishita, K., Watanabe, K., Naguro, I. & Ichijo, H. Sodium ion influx regulates liquidity of biomolecular condensates in hyperosmotic stress response. Cell Rep. 42, 112315 (2023).
Article CAS PubMed Google Scholar
Thomas, J. A., Buchsbaum, R. N., Zimniak, A. & Racker, E. Intracellular pH measurements in Ehrlich ascites tumor cells utilizing spectroscopic probes generated in situ. Biochemistry 18, 2210–2218 (1979).
Article CAS PubMed Google Scholar
López-Palacios, T. P. & Andersen, J. L. Kinase regulation by liquid–liquid phase separation. Trends Cell Biol. 33, 649–666 (2022).
Article PubMed PubMed Central Google Scholar
Alberti, S., Gladfelter, A. & Mittag, T. Considerations and challenges in studying liquid-liquid phase separation and biomolecular condensates. Cell 176, 419–434 (2019).
Article CAS PubMed PubMed Central Google Scholar
Schultz, J., Milpetz, F., Bork, P. & Ponting, C. P. SMART, a simple modular architecture research tool: identification of signaling domains. Proc. Natl. Acad. Sci. USA 95, 5857–5864 (1998).
Article ADS CAS PubMed PubMed Central Google Scholar
Tomida, T., Takekawa, M. & Saito, H. Oscillation of p38 activity controls efficient pro-inflammatory gene expression. Nat. Commun. 6, 8350 (2015).
Article ADS CAS PubMed Google Scholar
Kristiansen, M. et al. Mkp1 is a c-Jun target gene that antagonizes JNK-dependent apoptosis in sympathetic neurons. J. Neurosci. 30, 10820–10832 (2010).
Article CAS PubMed PubMed Central Google Scholar
Teng, C.-H., Huang, W.-N. & Meng, T.-C. Several dual specificity phosphatases coordinate to control the magnitude and duration of JNK activation in signaling response to oxidative stress. J. Biol. Chem. 282, 28395–28407 (2007).
Article CAS PubMed Google Scholar
Eke, I. et al. EGFR/JIP-4/JNK2 signaling attenuates cetuximab-mediated radiosensitization of squamous cell carcinoma cells. Cancer Res. 73, 297–306 (2013).
Article CAS PubMed Google Scholar
Zanke, B. W. et al. The stress-activated protein kinase pathway mediates cell death following injury induced by cis-platinum, UV irradiation or heat. Curr. Biol. 6, 606–613 (1996).
Article CAS PubMed Google Scholar
Newton, A. C., Bootman, M. D. & Scott, J. D. Second messengers. Cold Spring Harb. Perspect. Biol. 8, a005926 (2016).
Article PubMed PubMed Central Google Scholar
DeVries, S. H. Exocytosed protons feedback to suppress the Ca2+ current in mammalian cone photoreceptors. Neuron 32, 1107–1117 (2001).
Article CAS PubMed Google Scholar
Waldmann, R., Champigny, G., Bassilana, F., Heurteaux, C. & Lazdunski, M. A proton-gated cation channel involved in acid-sensing. Nature 386, 173–177 (1997).
Article ADS CAS PubMed Google Scholar
Kapolka, N. J. et al. Proton-gated coincidence detection is a common feature of GPCR signaling. Proc. Natl. Acad. Sci. USA 118, e2100171118 (2021).
Article CAS PubMed PubMed Central Google Scholar
Bhat, V. et al. Acidic pH promotes oligomerization and membrane insertion of the BclXL apoptotic repressor. Arch. Biochem. Biophys. 528, 32–44 (2012).
Article CAS PubMed PubMed Central Google Scholar
Cartron, P. F., Oliver, L., Mayat, E., Meflah, K. & Vallette, F. M. Impact of pH on Bax α conformation, oligomerisation and mitochondrial integration. FEBS Lett. 578, 41–46 (2004).
Article CAS PubMed Google Scholar
Isom, D. G. & Dohlman, H. G. Buried ionizable networks are an ancient hallmark of G protein-coupled receptor activation. Proc. Natl. Acad. Sci. USA 112, 5702–5707 (2015).
Article ADS CAS PubMed PubMed Central Google Scholar
Franzmann, T. M. et al. Phase separation of a yeast prion protein promotes cellular fitness. Science 359, eaao5654 (2018).
Article PubMed Google Scholar
Gupta, S. et al. Selective interaction of JNK protein kinase isoforms with transcription factors. EMBO J. 15, 2760–2770 (1996).
Article CAS PubMed PubMed Central Google Scholar
Guillén-Boixet, J. et al. RNA-induced conformational switching and clustering of G3BP drive stress granule assembly by condensation. Cell 181, 346–361. e317 (2020).
Article PubMed PubMed Central Google Scholar
Iserman, C. et al. Condensation of Ded1p promotes a translational switch from housekeeping to stress protein production. Cell 181, 818–831. e819 (2020).
Article CAS PubMed PubMed Central Google Scholar
Riback, J. A. et al. Stress-triggered phase separation is an adaptive, evolutionarily tuned response. Cell 168, 1028–1040. e1019 (2017).
Article CAS PubMed PubMed Central Google Scholar
Kroschwald, S. & Alberti, S. Gel or die: phase separation as a survival strategy. Cell 168, 947–948 (2017).
Article CAS PubMed Google Scholar
Maik-Rachline, G., Zehorai, E., Hanoch, T., Blenis, J. & Seger, R. The nuclear translocation of the kinases p38 and JNK promotes inflammation-induced cancer. Sci. Signal. 11, eaao3428 (2018).
Article PubMed Google Scholar
Liu, H., Nishitoh, H., Ichijo, H. & Kyriakis, J. M. Activation of apoptosis signal-regulating kinase 1 (ASK1) by tumor necrosis factor receptor-associated factor 2 requires prior dissociation of the ASK1 inhibitor thioredoxin. Mol. Cell. Biol. 20, 2198–2208 (2000).
Article CAS PubMed PubMed Central Google Scholar
Matsuura, H. et al. Phosphorylation-dependent scaffolding role of JSAP1/JIP3 in the ASK1-JNK signaling pathway: a new mode of regulation of the MAP kinase cascade. J. Biol. Chem. 277, 40703–40709 (2002).
Article CAS PubMed Google Scholar
Shirmanova, M. V. et al. Chemotherapy with cisplatin: insights into intracellular pH and metabolic landscape of cancer cells in vitro and in vivo. Sci. Rep. 7, 8911 (2017).
Article ADS PubMed PubMed Central Google Scholar
Rasmussen, J. K. & Boedtkjer, E. Carbonic anhydrase inhibitors modify intracellular pH transients and contractions of rat middle cerebral arteries during CO2/HCO3–fluctuations. J. Cereb. Blood Flow. Metab. 38, 492–505 (2018).
Article CAS PubMed Google Scholar
Barnett, D. H. et al. Estrogen receptor regulation of carbonic anhydrase XII through a distal enhancer in breast cancer. Cancer Res. 68, 3505–3515 (2008).
Article CAS PubMed Google Scholar
Koltai, T. Targeting the ph paradigm at the bedside: A practical approach. Int. J. Mol. Sci. 21, 9221 (2020).
Article PubMed PubMed Central Google Scholar
Chang, Y. et al. The Association between baseline Proton pump inhibitors, immune checkpoint inhibitors, and chemotherapy: A systematic review with network meta-analysis. Cancers 15, 284 (2022).
Article PubMed PubMed Central Google Scholar
Sharma, M. et al. The concomitant use of tyrosine kinase inhibitors and proton pump inhibitors: prevalence, predictors, and impact on survival and discontinuation of therapy in older adults with cancer. Cancer 125, 1155–1162 (2019).
Article CAS PubMed Google Scholar
van Leeuwen, R. W., van Gelder, T., Mathijssen, R. H. & Jansman, F. G. Drug–drug interactions with tyrosine-kinase inhibitors: a clinical perspective. Lancet Oncol. 15, e315–e326 (2014).
Article PubMed Google Scholar
Uhlén, M. et al. Tissue-based map of the human proteome. Science 347, 1260419 (2015).
Article PubMed Google Scholar
Didan, Y., Ghomlaghi, M., Nguyen, L. K., Ng,D. C. H. Stress pathway outputs are encoded by pH-dependent clustering of kinase components. zenodo https://doi.org/10.5281/zenodo.12633353 (2024).

Download references

Acknowledgements

We would like to thank the School of Biomedical Science Imaging and Analytical facility and Dr Shaun Walters for microscopy assistance. We thank Dr Sebastian Furness, Prof Walter Thomas, A/Prof Sean Millard and Dr Uda Ho for helpful discussion and for providing feedback on the manuscript. This project was funded by the National Health and Medical Research Foundation (GNT1162652), Australian Research Council (FT120100193) and Cancer Council Queensland (GNT1101931) grants awarded to A/Prof. Dominic Ng and a Research Fellowship from the Victorian Cancer Agency (MCRF18026) awarded to Lan K Nguyen.

Author information

Authors and Affiliations

School of Biomedical Science, Faculty of Medicine, University of Queensland; St Lucia, Brisbane, Australia
Yuliia Didan & Dominic C. H. Ng
Department of Biochemistry and Molecular Biology, School of Biomedical Sciences, Monash University, Clayton, Australia
Milad Ghomlaghi & Lan K. Nguyen
Biomedicine Discovery Institute, Monash University, Clayton, Australia
Milad Ghomlaghi & Lan K. Nguyen

Authors

Yuliia Didan
View author publications
You can also search for this author in PubMed Google Scholar
Milad Ghomlaghi
View author publications
You can also search for this author in PubMed Google Scholar
Lan K. Nguyen
View author publications
You can also search for this author in PubMed Google Scholar
Dominic C. H. Ng
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

Conceptualization: Y.D. and D.C.H.N. Methodology: Y.D., M.G., L.K.N., and D.C.H.N. Investigation: Y.D. and D.C.H.N. Visualization: Y.D. and M.G. Funding acquisition: D.C.H.N. Supervision: D.C.H.N Project administration: D.C.H.N. and L.K.N. Writing – original draft: D.C.H.N. Writing – review & editing: Y.D., M.G., and L.K.N.

Corresponding author

Correspondence to Dominic C. H. Ng.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Papa Kobina Van Dyck, and the other anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Didan, Y., Ghomlaghi, M., Nguyen, L.K. et al. Stress pathway outputs are encoded by pH-dependent clustering of kinase components. Nat Commun 15, 6614 (2024). https://doi.org/10.1038/s41467-024-50638-w

Download citation

Received: 20 August 2023
Accepted: 10 July 2024
Published: 05 August 2024
DOI: https://doi.org/10.1038/s41467-024-50638-w

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.