Introduction

Hsp70s are highly conserved molecular chaperones with wide-ranging roles in cellular homeostasis. At the core of their function is an ATP-driven conformational cycle: Hsp70s alternate between an open ATP-bound state and a closed ADP-bound state, enabling them to bind and release client proteins in a nucleotide-dependent manner¹. Through this mechanism, Hsp70s participate in diverse biological processes, including protein homeostasis¹, metabolic regulation², the DNA damage response³, and cell cycle control^4,5. Beyond their central roles in cell biology Hsp70s have been implicated in the pathology of numerous diseases. For example, Hsp70s can inhibit amyloid-beta aggregation in vitro and reduce its accumulation in mouse neurons, suggesting protective roles in neurodegenerative diseases such as Alzheimer’s disease^6,7. Hsp70s are also overexpressed in various cancers and influence tumor cell survival and proliferation ⁸. Thus, understanding how Hsp70 activity is regulated has broad implications for both basic biology and therapeutic development.

Hsp70 function is tightly controlled by co-chaperones and post-translational modifications (PTMs). Cochaperones — namely J proteins and nucleotide exchange factors — catalyze nucleotide hydrolysis and exchange, facilitate and specify client engagement, and determine client fate ¹. In recent years, PTMs have emerged as an important layer of regulation. A computational analysis of the budding yeast Hsp70, Ssa1, identified two conserved ‘hotspots’ in the nucleotide binding domain (NBD) and substrate binding domain (SBD), that are likely regions of PTM-mediated regulation⁹. Experimental work has validated some of the PTMs at these site. For example, T36 in the NBD of Ssa1 is phosphorylated both in response to nutrient stress and exposure to mating pheromone (α-factor), impairing chaperone activity and leading to cell cycle arrest⁴. In mammalian cells, the endoplasmic reticulum (ER) resident Hsp70, BiP, is AMPylated at T518, locking it in an ‘open’ conformation and tuning its activity to the unfolded protein burden in the ER ^10,11. Our lab previously showed that the Legionella pnuemophila (L.p.) kinase LegK4 phosphorylates cytosolic human Hsp70 (Hsc70; HSPA8) at T495 during infection, increasing its association with polysomes and globally reducing protein synthesis¹². Notably, Hsc70(T495) is structurally analogous to BiP(T518), indicating that both AMPylation and phosphorylation at this site have regulatory consequences. This finding led us to ask whether phosphorylation at T495 occurs outside of L.p. infection, and if it serves endogenous regulatory functions. Mining phosphoproteomics datasets revealed that the homologous residue in S. cerevisiae Ssa1, T492, is phosphorylated during DNA alkylation damage, and when mitotic exit is prevented^13,14. However, the functional consequences of this modification have not been studied. In this work, we show that Hsp70 is phosphorylated at T495 (pHsp70) in human cells during base excision repair (BER) of DNA damage. We find that a phosphomimetic mutant T495E allosterically impairs ATP hydrolysis in vitro and stabilizes an open-like conformation while still permitting substrate engagement. In yeast, the analogous mutation T492E causes a growth defect, leads to accumulation of cells in G1, and delays cell cycle re-entry after alkylation damage. Likewise, the phosphonull mutation (T492A) in yeast lacking a compensatory Hsp70 homolog also causes a growth defect and cell cycle dysregulation after alkylation damage. Together, these data suggest that the dynamic phosphorylation of Hsp70 at this conserved site acts as a regulatory switch to tune chaperone activity in coordination with DNA repair and cell cycle progression.

Results

Phosphomimetic Hsc70 (T495E) is locked in an open-like conformation

Previous work showed that the Legionella pneumophila kinase LegK4 phosphorylates Hsc70 at a conserved threonine (T495) in its substrate SBD (Figure 1a), reducing its chaperone activity¹². However, in vitro phosphorylation by LegK4 only yields phosphorylation of ∼53% of the Hsc70, complicating bulk biochemical analyses ¹². To overcome this limitation, we generated a phosphomimetic mutant Hsc70 (T495E). To determine whether this mutant mimics phosphorylated Hsc70, we assessed the ATPase activity of the purified protein using a malachite green assay. As Hsc70 has a low intrinsic rate of ATPase activity, we titrated in the co-chaperone DnaJA2 to stimulate hydrolysis^12,15. T495E exhibited significantly reduced J-protein stimulated ATPase activity compared to wild-type Hsc70 (WT) (Figure 1b), consistent with prior observations of the LegK4-phosphorylated protein¹². To investigate if the reduced ATPase activity is due to impaired nucleotide binding, we performed a fluorescence polarization (FP) assay using fluorescently-labeled ATP-FAM. ATP binding was not diminished in T495E (Figure 1c), suggesting that T495E allosterically inhibits J protein stimulated ATPase activity without preventing nucleotide binding.

Figure 1.

Hsp70 chaperone activity is driven by cycling through nucleotide-dependent conformations¹. In the ATP-bound state, Hsp70s adopt an ‘open’ conformation, and ATP hydrolysis to ADP induces large-scale structural rearrangements to a ‘closed’ state¹. As T495E poorly hydrolyzes ATP, we reasoned that this mutation may lock the protein into either the open or closed conformation. Indeed, AMPylation of BiP at the analogous residue is known to lock the protein in an open state¹⁶. To assess the conformation of T495E Hsc70, we performed partial proteolysis using trypsin, which produces distinct cleavage patterns depending on the protein’s conformation^17,18. As expected, WT Hsc70 displayed nucleotide-dependent banding: protection of band 2 and loss of band 3 in the presence of ADP compared to ATP. In contrast, T495E exhibited an ATP-like banding pattern regardless of nucleotide, consistent with a locked open conformation (Figure 1d). While this agrees with the studies on BiP AMPylation, our previous work showed that phosphorylation coincides with increased occupancy of Hsc70 on polysomes, suggesting continued substrate engagement¹². We therefore asked whether T495E can still bind client proteins. An ELISA with tau, a known Hsc70 substrate¹⁹, showed that T495E retains the ability to bind tau despite the conformational lock (Figure 1e). These results suggest that T495E locks Hsc70 in a pseudo-open conformation: structurally similar to the ATP-bound open state in terms of protease sensitivity and domain exposure, yet still capable of substrate engagement typically associated with the closed state.

Base excision repair leads to Hsp70(T495) phosphorylation

Pathogens often mimic host proteins and hijack cellular pathways to create a favorable niche for replication. Thus, these pathogen-driven modifications frequently reveal biologically significant regulatory nodes within host systems²⁰. Motivated by this concept, we asked whether the phosphoregulation of Hsp70 employed by Legionella might occur endogenously in mammalian cells. A phospho-proteomics study in S. cerevisiae identified phosphorylation at the analogous residue, Ssa1(T492), after treatment with the alkylating agent methyl methanesulfonate (MMS)¹³. Inspired by this finding, we analyzed human cells treated with MMS and observed that this phosphorylation is conserved(Figure 2a).

Figure 2.

This finding suggested that phosphorylation of Hsp70 might be involved in the response to DNA damage. MMS is primarily used as a DNA alkylating agent, though its chemical activity is not limited to DNA targets²¹. Thus, we sought to determine whether the response to MMS-induced DNA alkylation leads to Hsp70 phosphorylation. Aberrant DNA alkylation is predominantly repaired through the base excision repair (BER) pathway²². In BER, DNA glycosylases recognize and excise damaged bases, producing an abasic site (AP site). MMS-induced lesions are processed by N-methylpurine DNA glycosylase (MPG)^22,23 (Figure 2b). The backbone at the AP site is then cleaved by the endonuclease APE1, and the resultant single stranded break is repaired by a variety of proteins, including the DNA polymerase Polβ²² (Figure 2b). Interestingly, the predominant DNA adduct generated by MMS, 7-methylguanine, is not intrinsically cytotoxic, but spontaneously depurinates to a toxic and mutagenic intermediate and so requires rapid repair^24,25. The repair intermediates generated by the BER pathway, however, are cytotoxic²⁴. Because Polβ activity is rate limiting, overexpression of the glycosylase that initiates BER can lead to an accumulation of these cytotoxic intermediates^22,26. To determine whether increased BER activity affects Hsp70 phosphorylation, we treated cells overexpressing MPG with MMS. MPG overexpression modestly increased MMS-induced pHsp70 levels (Figure 2c). Conversely, inhibition of the subsequent BER step strongly reduced the levels of MMS-induced pHsp70. Pharmacological inhibition of APE1 led to a dose dependent decrease in Hsp70 phosphorylation (Figure 2d). Similarly, treatment with methoxyamine (Mx), which covalently binds AP sites and impairs APE1 cleavage and Polβ activity²⁷, caused a striking reduction in pHsp70 levels (Figure 2b,e). Surprisingly, overexpression of APE1 did not enhance pHsp70 (Figure S1a), suggesting that APE1 activity is necessary but not rate-limiting in this context. Additionally, overexpression of Polβ did not reduce pHsp70 upon MMS treatment (Figure S1b).These data confirm that the DNA alkylation by MMS is responsible for Hsp70 phosphorylation in human cells, and the results specifically indicate that BER intermediates may trigger this response.

To test the specificity of Hsp70 phosphorylation in response to DNA damage, we treated cells with a panel of genotoxic compounds that activate distinct branches of the DNA damage response (DDR), including bleomycin (induces double stranded breaks), camptothecin (a topoisomerase I inhibitor), hydroxyurea (depletes the dNTP pool), MMS, and arsenite (generates reactive oxygen species). While activation of DDR kinases (pDNA-PKcs(S2056), pATM(S1981) pChk2(T68), pChk2(S516)), and an increase of the DDR marker γH2AX confirm DNA damage and response induction in the drug treatment conditions, pHsp70 was only observed following high-dose MMS treatment, and exposure to sodium arsenite (Figure 2f). Arsenite causes oxidative DNA damage that can be repaired through BER^22,28. Surprisingly, arsenite has also been shown to both inhibit the activity of enzymes necessary for this pathway, including OGG (the oxidative glycosylase)²⁹, and decrease the levels of others (e.g. APE1 and Polβ) ³⁰. Despite this, we found that both treatment with Mx and inhibition of APE1 were still sufficient to prevent Hsp70 phosphorylation upon arsenite treatment (Figure S1c,d). These findings suggest that Hsp70(T495) phosphorylation is selectively triggered by repair intermediates generated during BER.

Phosphorylation of Hsp70 requires the DDR kinases ATM, DNA-PKcs, Chk2, and CK1

Though BER can function as a self-contained repair pathway, several DDR kinases are known to facilitate its activity^31,32. To determine if DDR signaling contributes to MMS-induced Hsp70 phosphorylation, we performed siRNA-mediated knockdowns of the three master kinases of the DDR: ataxia-telangiectasia mutated (ATM), ataxia-telangiectasia and Rad3-related (ATR), and DNA-dependent protein kinase catalytic subunit (DNA-PKcs). Among these, only knockdown of DNA-PKcs prevented Hsp70 phosphorylation (Figure 3a, Supp Figure 2a,b). Interestingly, while ATM knockdown had no effect, pharmacological inhibition of ATM decreased Hsp70 phosphorylation (Figure 3b). To rule out off-target effects, we tested two independent ATM inhibitors and observed a similar reduction in pHsp70 with both (Figure 3b). This discrepancy is consistent with prior reports showing that genetic and pharmacological inhibition of ATM can yield diverging results^33,34. Pharmacological inhibition of DNA-PKcs confirmed the DNA-PKcs knockdown experiment (Figure 3c). Both DNA-PKcs and ATM phosphorylate the transducer kinase Chk2, activating it and promoting propagation of the DDR³⁵. Inhibition of Chk2 also prevented Hsp70 phosphorylation (Figure 3c). Given that casein kinase 1 (CK1) has been previously reported to act in concert with Chk2 in the DDR³⁶, we tested its involvement and found that CK1 inhibition also suppressed Hsp70 phosphorylation (Figure 3d, Figure S2c). Inhibition of these kinases also prevented Hsp70 phosphorylation upon arsenite treatment (Figure S2d).

Figure 3.

Having identified a set of kinases required for Hsp70 phosphorylation, we noted that the delayed timing of this event is striking. Canonical DDR signaling is initiated within minutes of damage induction^31,37, yet pHsp70 only emerges after prolonged MMS exposure (Fig 3d). Indeed, we observed robust activation of DNA-PKcs, ATM and Chk2 well before Hsp70 phosphorylation (Figure 3d). These observations raise the possibility that either prolonged MMS exposure or a secondary cellular response is required to trigger Hsp70 phosphorylation.

Hsp70 phosphorylation occurs after M phase onset

To better understand the timing of Hsp70 phosphorylation, we treated cells with MMS for varying durations (1 to 5 hours), followed by recovery in fresh media up to a total of 5 hours. In previous experiments, we observed robust Hsp70 phosphorylation only after 5 hours of continuous MMS treatment, with minimal signal at earlier timepoints (see Fig. 3d). However, in our pulse-chase assay, 2 hour MMS treatment, when followed by a 3 hour MMS-free chase, resulted in pHsp70 accumulation (Fig. 4a). This result indicates that prolonged MMS exposure alone does not fully explain the lag time between damage initiation and the appearance of pHsp70, suggesting that secondary cellular events or signaling contribute to Hsp70 phosphorylation.

Figure 4.

Damage from MMS is thought to occur when replication forks collide with DNA repair intermediates in S phase³⁸. This hypothesis, in combination with our observations that pHsp70 accumulation lags behind initial DDR activation, suggested that cell cycle stage might be a key determinant in Hsp70 phosphorylation. In asynchronous populations, Hsp70 phosphorylation might only appear when enough cells reach a permissive phase of the cell cycle. If this were true, synchronizing cells at the beginning of S phase should potentiate MMS-induced Hsp70 phosphorylation. To test this idea, we synchronized cells at the beginning of S phase using a double thymidine block, then released them into MMS-containing or untreated media. Unexpectedly, synchronization at G1/S does not increase Hsp70 phosphorylation (Figure 4b). Even more surprising, MMS-treated cells entered mitosis at an increased rate compared to controls, as evidenced by the accumulation of the mitotic marker phospho-Histone H3 (S10) (pH3), reduced levels of S phase markers such as the replication licensing factor CDT1 and thymidine kinase, and pCdk1 (Y15), the inhibited cyclin dependent kinase that must be de-phosphorylated for M phase entry (Figure 4b). These data suggest that MMS drives cells into mitosis, and that Hsp70 phosphorylation occurs after this transition. Indeed, blocking mitotic entry by treatment with a CDK1 inhibitor (Ro3306) prevented both MMS and arsenite induced Hsp70 phosphorylation (Figure 4c, Figure S3a). Releasing the CDK1 block upon MMS or arsenite treatment restored Hsp70 phosphorylation (Figure 4c, Figure S3a), suggesting this response does not require passage through S phase. While passage through mitosis upon DNA damage is necessary for Hsp70 phosphorylation, the presence of pHsp70 in a distinct nuclear fraction indicates that it persists after nuclear envelope reformation (Figure 4d). Indeed, a 2 hour pulse chase of MMS revealed that pHsp70 accumulates even after pH3 disappears, indicating that phosphorylation is maintained after mitotic exit (Figure 4e). Notably, we do not see an increase of S phase markers such as thymidine kinase, nor changes in cyclin A or cyclin B levels, suggesting that the cell cycle progression may be dysregulated following MMS treatment.

Phosphoregulation of Ssa1 at T492 is important for G1/S transition in yeast

To investigate the functional significance of Hsp70 phosphorylation at T495, we turned to S. cerevisiae due to its genetic tractability. S. cerevisiae has four cytosolic Hsp70s, Ssa1-4. Of these, Ssa1 and Ssa2 are constitutively expressed, whereas Ssa3 and Ssa4 are stress-inducible³⁹. Phosphoproteomics studies revealed that Ssa1 can be phosphorylated at threonine 492, which corresponds to human Hsp70 T495^13,14. To probe the role of this modification, we created phosphomimetic (T492E, TE) and phosphonull (T492A, TA) point mutations at the endogenous SSA1 locus. Given the known redundancy between Ssa1 and Ssa2⁴⁰, we disrupted SSA2 by inserting an antimicrobial resistance gene into its coding sequence to better assess the functional consequences of Ssa1 phospho-variants. Strikingly, even in the SSA2 background, we observed a growth defect in TE yeast (Figure 5a,b). This is not likely explained by hypomorphism, as neither the TA mutant nor the WT;ssa2Δ strains show such a defect. Rather, phosphomimetic Ssa1 seems to act in a dominant manner. In the ssa2Δ background, strains expressing either phosphomimetic or phosphonull Ssa1 grew more slowly than the wild-type (Figure 5 a,b).

Figure 5.

Phosphorylation of Ssa1 at T492 was previously reported to occur in a cell cycle dependent manner, specifically when yeast are prevented from exiting mitosis through the expression of a non-degradable cyclin B¹⁴. This information, in conjunction with the cell cycle dependence of pHsp70 in human cells and the growth defects we observed in our yeast phosphomutants, led us to examine the impact of these mutations on cell cycle progression. With SSA2 present, the Ssa1 WT and phosphonull (TA) strains showed similar cell cycle distributions. In contrast, the phosphomimetic mutant (TE) showed an increase in cells with 1N DNA content, consistent with a G1 stall (Figure 5c). In the ssa2Δ background for both Ssa1 WT and TA, we see a slight increase in the G1 population compared to G2, and the difference is even more pronounced in T492E yeast (Figure 5c). T492A yeast showed increased cell size in G1 in the ssa2Δ background, a phenotype previously linked to G1/S stalling, perhaps due to decreased Cln3 stability⁴ (Figure S4a). Intriguingly, both Δssa2 point mutants exhibited a large cell size in G2/M (Figure S4a) — a finding that warrants additional investigation. Together, these data suggest that dynamic phospho-regulation of Ssa1 at T492 is critical for proper cell cycle progression.

We were curious if this modification served a similar cell cycle regulatory function during MMS treatment. To assess the overall viability of the point mutants in response to alkylation damage we performed a spot test in the presence and absence of MMS. In the ssa2Δ background, both phosphonull and phosphomimetic Ssa1 led to impaired growth on MMS-containing media, though growth in the in the SSA2 background was only mildly impacted (Figure 5d). To examine how these mutations influence cell cycle progression in response to MMS, we treated the strains with 0.05% MMS for three hours as previously shown to cause Ssa1 phosphorylation at T492¹³, and released them into fresh YPAD for the indicated times (Figure 5e). Ssa1 WT strains with and without SSA2 behave similarly: immediately after the MMS treatment the cells display a bimodal distribution around G1 and early S. Upon MMS release, the G1 peak gradually disappeared, accompanied by synchronized progression through S phase, with DNA content slowly increasing over time. In the SSA2 background, the T492A was nearly indistinguishable from the WT strains. However, in the ssa2Δ background, T492A no longer exhibited the bimodal distribution of cells near 1N immediately after MMS treatment. Rather, there was a single peak. Upon release, the cells failed to undergo coordinated S phase progression. Instead, DNA content increased in an asynchronous manner, suggesting impaired checkpoint regulation. The T492E mutants also showed distinct phenotypes. In the SSA2 background, T492E cells lacked the bimodal distribution around 1N immediately after MMS treatment, and instead exhibited a more pronounced G1 arrest. Upon release, these cells progressed through S phase more slowly than WT, with a prominent left shoulder, indicating delayed DNA replication in a subset of cells. In the ssa2Δ background, the phenotype was even more prominent. A strong 1N peak persisted after MMS release, with minimal evidence of bulk S phase progression. Instead, we observed a gradual increase in DNA content across the population, reminiscent of the T492A mutant in ssa2Δ, but with a key difference: while T492A;ssa2Δ cells showed more rapid entry to and passage through S phase, T492E;ssa2Δ cells retained a distinct 1N peak and slower S phase progression. These data indicate that both phosphonull and phosphomimetic mutations disrupt the coordination of cell cycle re-entry following genotoxic stress. Loss of phosphorylation appears to weaken G1 arrest and promote uncontrolled S-phase entry, while constitutive phosphorylation reinforces arrest but impairs orderly progression through S phase. Eliminating the functional redundancy of Ssa2 in the T492A;ssa2Δ and T492E;ssa2Δ strains reveals the importance of dynamic Ssa1 phospho-regulation in mounting an effective DNA damage response and ensuring proper cell cycle transitions, as both phosphonull and phosphomimetic Ssa1 variants result in cell cycle dysregulation and altered growth in the absence of Ssa2.

Discussion

Our work demonstrates that Hsp70s are phosphorylated at T495 as part of a conserved response to DNA damage, and that dynamic phosphoregulation of this residue exerts control over the cell cycle. While prior phosphoproteomics studies in yeast detected phosphorylation of the analogous residue in Ssa1 (T492) following MMS exposure¹³ or mitotic exit block¹⁴, our findings establish that this modification is a regulated, damage-responsive event that is conserved in human cells. We show that phosphorylation is induced by two chemically distinct DNA-damaging agents, MMS and sodium arsenite, both of which generate BER substrates. Critically, inhibition of the BER endonuclease APE1 or chemical blockade of abasic site recognition and repair abolishes pHsp70 formation. Of note, high doses and long treatment times are required to elicit Hsp70 phosphorylation with both compounds, which raises questions as to the precise driver of this response. BER is a streamlined pathway: it is thought to follow a ‘baton-passing’ model, with each enzyme passing off intermediates to the next in the cascade⁴¹. However, BER intermediates are cytotoxic. APE1-mediated cleavage of the DNA backbone produces single-strand breaks that, if encountered by replication forks, can be converted into double-stranded breaks (DSBs)³⁸. The requirement for APE1 activity, along with the involvement of the DSB repair kinases ATM and DNA-PKcs, support a model in which BER-induced DSBs lead to Hsp70 phosphorylation.

While it is clear that Hsp70 is phosphorylated in response to DNA damage, the lag time between insult and pHsp70 suggests a complex signaling pathway. Phosphorylation does not arise simply during damage exposure; instead, it requires both prolonged insult and progression through mitosis. Despite the longstanding view that MMS-induced damage primarily arises during S phase, we find that synchronizing cells at G1/S does not enhance Hsp70 phosphorylation. This paradox may be resolved by emerging evidence that BER-induced DSBs can occur in non-replicating cells⁴². Considering that the DDR is considerably rewired during mitosis, and though BER occurs, repair of lesions such as DSBs is forestalled until mitotic exit^43–45, it is possible that pHsp70 arises in response to high lesion burden in M phase to prevent cell cycle entry into the vulnerable S phase (Figure 6a).

Figure 6.

Our genetic studies in yeast directly link phosphoregulation of Ssa1 at T492 to cell cycle progression. Both phosphonull (T492A) and phosphomimetic (T492E) mutants cause growth defects in the absence of the compensatory isoform Ssa2, underscoring the importance of dynamic, reversible phosphorylation at this site. Cell cycle profiling reveals that T492E yeast in both SSA2 and ssa2Δ backgrounds accumulate in G1 and fail to efficiently enter S phase following MMS exposure, while T492A strains show delayed recovery in the absence of Ssa2. Further, in the absence of MMS insult, the Ssa1 T492E mutant strain exhibits altered growth and cell cycle progression even in the SSA2 background, suggesting a dominant effect of the phosphomimetic variant. The dominant phenotype of T492E is consistent with biochemical data suggesting persistent, aberrant substrate engagement, akin to the effects of mutations that trap Hsp70 in a closed conformation⁴⁶. Together, these results support a model in which pHsp70 serves as a reversible molecular brake, preventing premature S-phase entry under conditions of BER stress. Interestingly, this mechanism parallels, but is distinct from the previously described phospho-Ssa1 (T36) mediated regulation of the cell cycle. Phosphorylation at T36/38 also enforces G1/S arrest in yeast. However, the residue is in the NBD, and phosphorylation decreases nucleotide binding. Furthermore, the activity is ER-localized, where it slows the accumulation of Cdk1-Cln3 by sequestering and promoting the degradation of Cln3⁴. In contrast, pT492/495 occurs in the SBD, modulates ATP hydrolysis without preventing nucleotide binding, and occurs in the nucleus. Given the established role of phosphorylation in directing Hsp70-client fate ^47,48, it is plausible that pT492/T495 controls the stability or activity of nuclear substrates critical for G1/S transition. Our data support a model wherein Hsp70 engagement with both a client and some accessory machinery (e.g. a cochaperone) are controlled by phosphorylation to regulate S phase entry (Figure 6b). For instance, if a specific Hsp70 client must be degraded to exit G1, phosphorylation may promote engagement with said client and proper degradation machinery, and dephosphorylation could allow client release. In this case, the phosphomimetic Hsp70 would be unable to disengage, and therefore cause the G1/S stall we report. Conversely, the phosphonull Hsp70 would be unable to establish proper interactions with the client and/or accessory machinery, thereby preventing efficient degradation and leading to the observed dysregulation of S phase re-entry. While this model explains our experimental observations, other possibilities remain, and future work is required to further establish mechanistic details.

Collectively, our findings reveal a conserved phospho-switch on Hsp70 that links DNA damage sensing during BER to cell cycle control, adding an unanticipated regulatory layer to the integration of repair and checkpoint pathways. Strikingly, this pathway came to light because it is targeted by a pathogenic effector kinase, once again illustrating how pathogens can hijack, and thereby illuminate, fundamental aspects of cell biology. Such examples underscore the enduring value of pathogens as powerful tools to uncover deeply conserved regulatory mechanisms. This work opens the door to identifying the upstream kinase(s) and downstream client(s) that mediate this effect, and to determining whether T495 phosphorylation serves as a general mechanism for coordinating repair with proliferation in diverse physiological contexts. By coupling BER to a reversible molecular brake on S phase entry, phosphorylation of Hsp70 at T495 emerges as a conserved checkpoint signal that safeguards genome integrity under conditions of repair-induced stress.

Resources

Plasmids

RNA Oligos

DNA Oligos

Antibodies

Yeast Strains

Software

Methods

Cell Culture

HEK293T and HeLa lines were obtained from ATCC. All cell lines were regularly tested for mycoplasma contamination. Cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM, Gibco) supplemented with 10% fetal bovine serum (FBS, VWR) in a humidified incubator at 37°C, 5% CO₂. Cells were treated with 10 mM methyl methanesulfonate (MMS) (Fisher AC156890050) for 5 h unless otherwise indicated. Cells were treated with methoxyamine hydrochloride (Sigma Aldrich 226904), APE1 compound III (EMD Millipore 262017), bleomycin (sulfate) (Thomas Scientific C830H18), camptothecin (Selleck S1288), hydroxyurea (Sigma-Aldrich H8627), sodium arsenite (Fisher Scientific S88733), KU-60019 (MedChem Express HY-12061), AZD1390 (MedChem Express 2089288-03-7), CCT241533 (MedChem Express HY-14715B), PF-670462 (Sigma-Aldrich SML0795), thymidine (Fisher 501882638), Ro-3306 (MedChem Express HY-12529), and AZD7648 (MedChem Express 2230820-11-6) as described in the figure legends.

siRNA and Plasmid Transfection of Mammalian Cells

Cells were transfected with high quality maxiprepped plasmid DNA (Sigma Genelute HP kit), custom synthesized siRNA duplexes (Sigma-Aldrich VC30002), or Mission® siRNA Universal Negative Control #1 (Sigma Aldrich SIC001) using the jetPRIME transfection system (VWR 89129-926). For plasmid transfection, cells were grown to 70% confluence and transfected overnight with total µg of DNA added as recommended by supplier for the culture vessel. For siRNA transfection, knockdowns were performed per manufacturer suggestions in HeLa cells. Cells were transfected for 24 h after which the media was replaced with fresh media. Cells were allowed to grow for an additional 48 h before any additional treatment and harvesting.

Whole Cell Lysate Preparation and Western Blotting

Once experimental manipulations were complete, media was aspirated and cells were washed 3X with ice cold 1X TBS on ice. Cells were gently scraped into 1 mL ice cold 1X TBS and transferred to 1.5 mL microcentrifuge tubes. Cells were pelleted at 3000xg for 5 min at 4°C, and the supernatant was aspirated. Pellets were stored at −80°C until lysis.

1x RIPA lysis buffer (Cell Signaling Technology 9806S) supplemented with 1x Roche cOmplete protease inhibitor cocktail (Roche 11697498001), 1x Roche PhosStop (Roche 04906837001), and 1mM PMSF was added to the frozen pellets which were then agitated for 30 min at 4 °C then centrifuged at 16,000 xg for 15 min at 4 °C. Cleared lysates were transferred to a new tube, and protein concentrations were quantified using the Pierce BCA Assay Kit (Thermo Scientific 23227). 30-100 µg total protein was denatured in 1x SDS sample buffer with 10mM dithiothreitol (DTT) for 8 min at 95 °C. Lysates were loaded onto polyacrylamide gels (8, 10% or Bio-Rad Mini-PROTEAN TGX Precast Protein Gels, 4-20%) (BioRad) and separated by SDS-PAGE. Proteins were transferred to methanol-activated 0.2 µm pore PVDF membranes overnight at 4°C in 1X CAPS, 10% methanol. Membranes were blocked in either 5% bovine serum albumin (BSA) (VWR) for phopsho-specific antibodies, or 5% non-fat dry milk (BioRad) in TBS-T (0.1% Tween-20 in TBS) for 1 h at RT. Primary antibodies were diluted in 5% BSA, 0.02% NaN₃ in TBS-T. Membranes were incubated with primary antibody solutions overnight at 4°C. Membranes were washed 3X in TBS-T for 10 min each wash, then incubated with 1:5000 secondary antibody in blocking buffer for 1 h at RT. Membranes were washed 3X in TBS-T for 10 min each wash, then developed for 1 min in Amersham ECL Western Blotting Detection Reagent (Cytiva RPN2209) or Immobilon Crescendo (Millipore WBLUR0500) and imaged on a ChemiDoc Imaging System (BioRad).

Nuclear and Cytoplasmic Extraction

Nuclear and cytoplasmic extraction was performed at 4 °C using the NE-PER Nuclear and Cytoplasmic Reagents (Fisher Scientific PI78835, Thermo Scientific 78833) following the manufacturer’s protocol. Protein quantification and Western blotting was performed as described above.

Protein Purification

Hsc70 and point mutants were purified as described previously⁵⁰. In brief, plasmids containing His-tag proteins were transformed into Rosetta(DE3) E. coli. Bacteria were grown in 2 L Terrific Broth (TB) to OD₆₀₀=0.6. Bacterial cultures were cooled to 20 °C, and protein expression was induced with 200µM IPTG at 20 °C overnight. The cells were harvested by centrifugation (7500 r.p.m. 10 min, JLA 8.1 rotor) and then resuspended in 20 mL His-Binding Buffer (50 mM Tris, 10 mM imidazole, 500 mM NaCl, pH 8) + 2 tablets EDTA-free Roche cOmplete Protease Inhibitor per liter culture using a Dounce homogenizer. Cells were then lysed by sonication at Amp 35% for 5 min, 30 sec on, 30 sec off, on ice at 4 °C. Lysate was separated by centrifugation (18000 r.p.m. 30 min, JA-20 Beckman rotor) and incubated with 10 mL/L pre-equilibrated Ni-NTA resin (EMD Millipore 70666-5)/liter of culture at 4 °C for 1 h. Bound protein was first washed with 200 mL His-binding buffer, then 100 mL His-Washing buffer (50 mM Tris, 30 mM imidazole, 300 mM NaCl, pH 8), and then eluted with His-Elution buffer (50 mM Tris, 300 mM imidazole, 300 mM NaCl, pH 8). This His-tag was then removed by the spiking in 5 mM β-mercaptoethanol and 600 µg TEV protease to the sample and dialyzing overnight into Buffer A (25 mM HEPES, 5 mM MgCl₂, 10 mM KCl, pH 7.5) in 10 KDa MWCO snakeskin dialysis tubing (Thermo Fisher 68100). The protein was further purified by an ATP-agarose column using previously established protocols⁵⁰. Purified DnaJA2 and tau were acquired from J.E.G.

ATPase Assays with Malachite Green

The ATPase activity of Hsc70 and Hsc70(T495E) was performed with malachite green (MG) (Sigma Aldrich) as described previously⁵¹. Briefly, in a clear 96-well plate, Hsc70 or Hsc70(T495E) were incubated with increasing concentrations of human DnaJA2 (DJA2) in 25 μL total volume. Reactions were also performed with DnaJA2 in the absence of Hsc70 for background subtraction. The assay buffer was 100 mM Tris at pH 7.4, 20 mM KCl, 6 mM MgCl₂, and 0.01% Triton. The reaction was initiated by the addition of ATP at a final concentration of 1 mM and incubated at 37 °C for 1 h. After incubation, 80 μL of MG reagent was added, followed by 10 μL of saturated sodium citrate to quench the reaction. Absorbance was measured at 620 nm on a SpectraMax M5 plate reader (Molecular Devices). ATP hydrolysis rates were calculated by comparison to a phosphate standard. Displayed curves are a combination of 6 replicates.

Fluorescence Polarization (FP) of ATP-FAM

Nucleotide of Hsc70 and Hsc70(T495E) was assessed with a fluorescence polarization (FP) assay using labeled ATP. First, apo-Hsc70 was generated by subsequent 6-12h dialyses in the following buffers: buffer 1 (25 mM HEPES, 100 mM NaCl, 5 mM EDTA (pH 7.5)), buffer 2 (25 mM HEPES, 100 mM NaCl, 1 mM EDTA (pH 7.5)), buffer 3 (25 mM HEPES, 5 mM MgCl₂, 10 mM KCl (pH 7.5)). Following this, 40 kDa MWCO Zeba Buffer Exchange columns (Thermo Fisher A57760) were used to exchange the buffer to FP assay buffer 100 mM Tris, 20 mM KCl, 6 mM MgCl₂, pH 7.4) before protein quantification. The assay was performed in 384-well black round-bottom low-volume plates (Corning 4511). 40 nM ATP-FAM (Jena Bioscience nu-805-5fm) was added to each well for a final concentration of 20 nM in 20 µL. Hsc70 and Hsc70(T495E) were added for a starting concentration of 20 µM in 20 µL, and then serially diluted with FP assay buffer (2-fold 13 times into the ATP-FAM. Plates were covered and incubated for 30 min at room temperature. Fluorescence polarization was read on a SpectraMax M5 plate reader (excitation, 485 nm; emission, 535 nm)

Partial Proteolysis

The partial proteolysis protocol to identify Hsp70 conformations was modified from previous work¹⁷. Hsc70 and Hsc70(T495E) was buffer exchanged to proteolysis buffer (40 mM HEPES, 20 mM NaCl, 8 mM MgCl₂, 20 mM KCl, 0.3 mM EDTA, pH 8) using Zeba 40 kDa MWCO buffer exchange columns and diluted to 3 µM in this buffer. ATP or ADP was added to 1 mM and incubated for 30 min at room temperature. 1.5 µM trypsin (Sigma EC 3.4.21.4) was added for 2 h at room temperature. The reaction was quenched by boiling in SDS loading buffer (125 mM Tris-HCl pH 6.8, 5% SDS, 10% β-mercaptoethanol, 20% glycerol, 0.025% bromophenol blue), run on a precast SDS-PAGE gel (Bio-Rad) and stained by Coomassie.

Hsc70-tau binding ELISA

Hsc70-tau ELISA was modified from a previously published protocol⁵². It was performed in a Fisherbrand, Flat bottom 96-well plates, clear, PS (cat no 12565501). 1µM Hsc70 or Hsc70(T495E) (diluted in dialysis buffer 3 from FP assay) was added to wells along with 1 mM ATP and incubated overnight at 37°C. Protein was discarded from wells followed by 3x 3min washes on rocker with PBS-T, discarding the PBS-T and vigorously blotting the inverted plate on a paper towel between each wash. In triplicate, add 30 uL of tau to each well, spanning a 12-dose 3-fold concentration gradient (from 70 µM). Plates were covered and incubated at room temperature for 3 h. Solutions were removed and wells washed 3x with PBS-T as described above. 100 µL of blocking solution (5% non-fat milk in TBS-T) was added to all wells and incubated at room temperature for 5 min without rocking. Solution was removed without washing. 50 uL of primary antibody was added to all wells (1:2000 mouse anti-tau clone D-8 from SantaCruz Biotechnologies in blocking solution). The plate was incubated for 1 h at room temperature without shaking. Solution was removed and wells washed 3x with PBS-T. 50 µL of secondary antibody (1:2000 goat anti-mouse (Jackson 115-035-146)in blocking solution) was added to the wells. The plate was incubated for 1h at room temperature and then washed 3x with PBS-T. 50 µL of TMB substrate was added to the wells (ThermoFisher 34028) and incubated for 15 min at room temperature in the dark. 50 µL 1M HCl was added and the plate read at 450 nm on SpectraMax M5 plate reader.

Yeast Strain Generation

W303α yeast and W303α yeast with SSA1 NAT were gifted to us by David Morgan. Oligos for yeast point mutations were generated by PCR amplification of NAT-SSA1with the relevant codon substituted in the primer sequence. Oligos for SSA2 interruption were generated by PCR amplification of KanMX from pYM13 with flanking sequences to SSA2. gDNA was extracted from yeast by first suspending yeast colonies from a YPAD plate in 100 µL LiOAc/SDS buffer (0.2 M LiOAc, 1% SDS). Cells were incubated at 70 °C for 10 min, and then 300 µL 96-100% ethanol was added. Sample was then mixed by vortexing. Sample was the spun at 15,000 xg for 3 min, supernatant removed, and pellet dissolved in 100 µL nuclease free water. The sample was then spun down at maximum speed for >15 sec, and supernatant was transferred to a new tube. 1 µL of supernatant was used for subsequent PCRs.

For yeast transformations, competent cells were generated by first diluting an overnight culture 1:20 in YPAD and grown on a rotary shaker at 30 °C to mid-log. The yeast were spun at 3000 xg for 1min. While spinning the yeast, 10 mg/mL salmon sperm DNA (Sigma-Aldrich D9156) was boiled for 5 min and then put on ice. The supernatant was removed from the yeast pellet which was then washed 1x in water and re-pelleted. The water was remove and the yeast was resuspended in LiOAc/TE solution (10 mM Tris-HCl pH 8, 1 mM EDTA, 0.1 M lithium acetate (Sigma-Aldrich L4158)). This was then pelleted, the supernatant aspirated, and yeast were resuspended in 200 uL LiOAc/TE solution to generate competent cells. For each transformation, 50uL of these competent cells were mixed with 1µg of DNA along with 10 µL of boiled salmon sperm DNA and 500 µL of PEG/LiOAc/TE solution (10 mM Tris-HCl pH 8, 1 mM EDTA, 0.1 M lithium acetate, 40% PEG 3350). This mix was incubated at 30 °C with shaking at 550 r,p,m, on a ThermoMixer F1.5 (Eppendorf), then spun for 1 min at 3000 xg. The supernatant was removed and the yeast resuspended in sterile water and plated on YPAD and grown overnight at 30 °C. The following day, these plates were replica plated onto plates with the appropriate selection marker.

Yeast Growth

For all yeast experiments, yeast were first grown overnight in YPAD on a rotary shaker at 30 °C. The following morning, yeast were diluted to OD₆₀₀ =0.3 and grown at 30 °C to midlog (OD₆₀₀ = ∼0.6). Strains were then concentrated (by centrifugation at 3000xg) to 2e7 cells/mL in 1.5 mL microcentrifuge tubes.

Growth Curve

The above yeast were diluted 1:30 into a clear flat bottom 96 well plate (Costar 3370) and grown for 25 h with continuous orbital shaking at 30 °C on a Cytation 5 Imaging Reader (BioTek) with OD₆₀₀ collected every 20 min.

Yeast Spot Test

The day prior to the experiment, Nunc Rectangular Dishes (Thermo 267060) were made with either YPAD or YPAD + 0.0095% MMS. The above yeast were serially diluted 10-fold 5 times. 5 µL of the dilutions were spotted on the YPAD or YPAD + MMS plates. Plates were grown for three days at 30 °C and imaged daily on a ChemiDoc Imaging System (BioRad).

Yeast Cell Cycle Analysis

This protocol was modified from previous work^53,54 For cell cycle analysis of yeast at mid-log, we took 500 µL of the 2e7 cell/mL yeast solution, spun it at 14,000 xg for 30 sec, removed the supernatant, washed 1x in water and re-pelleted. The water was then decanted, and the yeast were vortexed to resuspend them in the water remaining in the tube. 95% ethanol prechilled to - 20 °C was added to the resuspended yeast which were then vortexed and placed on ice. For MMS treatment, yeast were grown to mid-log as previously described. 0.05% MMS (Fisher Scientific AC156890050) was then added to the yeast for 3 h¹³, rotating at 30 °C. Samples were harvested as described for ‘no recovery’ conditions, and the remaining yeast was transferred to 15 mL centrifuge tubes (Fisher Scientific 12-565-268),spun at 300 xg. The pellet was washed 1x with sterile water, spun down again, and then resuspended in the starting volume of YPAD. The caps were partially unscrewed and the yeast was returned to the rotary at 30 °C with samples harvested as detailed above at the indicated time points. Once all samples were resuspended in ethanol the yeast were stored at −20 °C for at least overnight. The yeast were then resuspended by vortexing and spun at 14000 xg for 1 min. The ethanol was removed, the yeast resuspended in water and then pelleted again. The water was removed and the pellet was resuspended in 1mL sodium citrate buffer (50 mM sodium citrate (C8532), pH to 7.5 with citric acid (BP399)). 8 µL of 10 mg/mL RNAase A (Thermo EN0531) was then added to the samples and incubated for 2 h at 37 °C. Following this, 10 µL of 20 mg/mL Proteinase K (NEB P8107S) was added to the samples and incubated for 1 h at 50 °C. The samples were then incubated at 4 °C overnight. The samples were spun down at 14000 xg for 1 min and the pellets resuspended in 500 µL sodium citrate buffer + 2.5 µM SYTOX Green (Thermo Fisher S7020), or sodium citrate buffer alone for the no-stain control. Samples were placed on ice and sonicated with a probe sonicator (QSonica) at 30% amplitude for 10 sec. The samples were then stored in the dark at 4 °C until flow cytometry analysis.

All yeast flow cytometry experiments were performed on a LSRII SORP (BD Biosciences) using a 488 nm laser, and the software analysis done on FlowJo (10.10.0). Cells were first selected by graphing FSC-A by SSC-A. Singlets were then gated by graphing FSC-A by FSC-H and then analyzed.

Data Reproducibility and Statistical Analysis

All statistics were performed in Prism as described in the relevant figure legends. Figures are either mean + SD, or representative of the number of replicates described in the figure legend.

Data availability

We will make the data freely and widely available as needed

Acknowledgements

We thank Dr. Adriana Steinbach and Dr. Michael Metrick for critical evaluation and discussions of the data. We thank Dr. Henry Ng and Dr. David O. Morgan for yeast strains and technical advice. We thank Dr. Oleta Johnson, Dr. Cory Nadel, and Dr. Emma Carroll from the J.E.G. lab for technical assistance and advice. We acknowledge Vinh Nguyen for his technical support and the PFCC (RRID:SCR_018206) for assistance generating Flow Cytometry data. Research reported here was supported in part by the DRC Center Grant NIH P30 DK063720. T.M. acknowledges support from the MPHD T32 training grant. J.E.G. acknowledges support from R01NS059690. S.M. acknowledges financial support from the National Institutes of Health (grant nos. R01GM140440 and R01GM144378), the Pew Charitable Trust (grant no. A129837), a Bowes Biomedical Investigator award, and a gift fund from the Chan-Zuckerberg Biohub.

Additional information

Author contributions

Conceptualization, T.M., S.M.; Experiment design, T.M., J.E.G., S.M.; Experimentation, T.M., A.W. K.B. M.C..; Formal Analysis, T.M..; Writing, T.M., S.M.; Review & Editing T.M, A.W., J.E.G., S.M.; Funding Acquisition: J.E.G., S.M.

Funding

HHS | NIH | National Institute of General Medical Sciences (NIGMS) (R01GM140440)

• Shaeri Mukherjee

HHS | NIH | National Institute of General Medical Sciences (NIGMS) (R01GM144378)

• Shaeri Mukherjee

HHS | NIH | National Institute of Neurological Disorders and Stroke (NINDS) (R01NS059690)

• Jason E Gestwicki

Additional files

Figure S1

Figure S2

Figure S3

Figure S4

Supplementary Figure Legend