Mixed‐class J‐domain protein scaffolds promote expanded aggregate handling and multivalent Hsp70 engagement during functional disaggregase assembly

Abstract

Protein aggregates threaten cellular proteostasis and are linked to aging and disease. In metazoa, aggregate resolution relies on Hsp70‐J‐domain protein (JDP)‐based disaggregases. Previous studies showed human class A and class B JDP assemblies enhance Hsp70‐mediated disaggregation, but the underlying mechanism has remained unclear. Using J‐domain mutants that impair Hsp70 binding while preserving mixed‐class JDP interaction, we show that synergistic disaggregation is lost when either JDP partner cannot engage Hsp70. Size‐resolved disaggregation assays further reveal that mixed‐class JDP assemblies influence the processing of distinct luciferase aggregate populations, including aggregate species inefficiently handled by either JDP alone. Our findings support a model in which mixed‐class JDP assemblies enhance Hsp70 disaggregation through expanded aggregate‐processing capacity and multivalent Hsp70 recruitment by both JDP partners.

Protein aggregates threaten proteostasis and cell health. In human cells, Hsp70–J‐domain protein‐based disaggregases remove aggregates, but how they assemble remains unclear. Our biochemical findings show that DNAJA2‐ and DNAJB1‐containing disaggregase scaffolds enhance luciferase aggregate targeting, and that Hsp70 recruitment by both J‐domain proteins within these scaffolds is crucial for building high‐capacity disaggregases.

Abbreviations

CTD, C‐terminal domain

EEVD, Glutamic acid‐Glutamic acid‐Valine‐Aspartic acid

FRET, Förster resonance energy transfer

HPD, Histidine‐Proline‐Aspartic acid

JDP, J‐domain protein

NEF, nucleotide exchange factor

PQC, protein quality control

QPN, Glutamine‐Proline‐Asparagine

RRR, D6R, E61R, E64R in DNAJA2, and D4R, E69R, E70R in DNAJB1

SEC, size‐exclusion chromatography

sHSP, small heat shock protein

Proteins must fold into precise three‐dimensional structures to perform their biological functions. Errors in protein biogenesis, destabilizing mutations, proteotoxic stress, and age‐related decline in protein quality control (PQC) pathways can disrupt this process, leading to the accumulation of misfolded proteins that form toxic aggregates. Such aggregates impair cellular homeostasis and can ultimately trigger cell death [1, 2]. To counteract protein aggregation, cells deploy protein disaggregases—specialized chaperone machines that solubilize aggregates by extracting trapped polypeptides for refolding or degradation [1]. Protein disaggregation activity was first identified in bacteria and fungi, where Hsp100‐type disaggregases—hexameric AAA+ ATPases—drive the disassembly of protein aggregates and confer thermotolerance [3, 4, 5]. Extensive biochemical and structural studies have revealed how these machines extract polypeptides from aggregates through ATP‐driven translocation mechanisms [6, 7, 8, 9]. However, Hsp100‐type disaggregases are not conserved in the cytosol or nucleus of human cells [10]. Instead, biochemical studies show that human cells rely on disaggregation systems centered on the Hsp70 chaperone machinery [1].

Hsp70 disaggregase systems are formed through the coordinated action of JDPs and Hsp110‐type nucleotide exchange factors [11, 12, 13]. Among these systems, particularly efficient Hsp70 disaggregases arise when canonical class A and class B JDPs assemble (referred to as mixed‐class JDP assemblies) within the disaggregase machine [12]. These JDP assemblies synergistically enhance the disaggregation of amorphous protein aggregates, including the widely used model substrate heat‐aggregated firefly luciferase [12, 14, 15, 16]. Apart from human JDPs, reconstituted systems containing class A and B JDPs from yeast and nematodes also show increased activity [14, 17, 18, 19], suggesting that cooperative interactions between JDPs from these two classes represent a conserved strategy for potentiating Hsp70‐mediated aggregate disassembly.

Despite these observations, the mechanistic basis underlying the architecture and assembly steps of these potent mixed‐class JDP‐containing disaggregases remains poorly understood. Progress in defining their molecular organization has been hindered in part by the structural heterogeneity of these systems, which are thought to form dynamic oligomeric assemblies on aggregate surfaces. Currently, the primary mechanistic insight is that class A and class B JDPs can form transient mixed‐class complexes mediated by electrostatic interactions between the J‐domain of one JDP and the C‐terminal domain (CTD) of the partner. This interaction has been demonstrated for the human class A JDP DNAJA2 and the class B JDP DNAJB1, providing a molecular basis for mixed‐class JDP scaffold formation [12].

One model proposes that such JDP scaffolds enhance disaggregase activity by recruiting multiple Hsp70 molecules to aggregate surfaces, thereby facilitating the rapid formation of higher‐order Hsp70 assemblies capable of extracting polypeptides from aggregates [10, 20]. However, direct experimental evidence for such cooperative or multivalent Hsp70 recruitment during mixed‐class disaggregase assembly has been lacking. In particular, it has remained unclear whether both JDP partners within mixed‐class complexes must actively recruit Hsp70, or whether engagement of Hsp70 by a single JDP component is sufficient once the mixed‐class scaffold has formed.

Here, we address this question using DNAJA2 and DNAJB1 mutants in which the conserved His‐Pro‐Asp (HPD) motif within the J‐domain is replaced with Gln‐Pro‐Asn (QPN). The HPD motif is required for productive JDP‐mediated stimulation of Hsp70's ATPase activity [21, 22]. We selected the QPN substitution because it disrupts this conserved motif while preserving the central proline residue in the HPD loop, allowing us to test whether Hsp70 recruitment can be functionally uncoupled from mixed‐class JDP scaffold formation. Using complementary biochemical approaches, we show that efficient disaggregation of heat‐induced luciferase aggregates requires both DNAJA2 and DNAJB1 to remain competent for productive Hsp70 recruitment. We further show that mixed‐class JDP interactions influence the processing of distinct luciferase aggregate populations. Together, these findings reveal that enhanced aggregate‐processing capacity and multivalent Hsp70 recruitment are distinct yet coupled functions of mixed‐class JDP scaffolds that drive high‐capacity disaggregation.

Materials and methods

Protein expression and purification

Human DNAJA2 and DNAJB1 generated and expressed from vector pCA528/pSUMO with a 6 × His–Smt3 tag as previously described [12, 14]. DNAJA2 variant for N‐terminal ReAsH labeling was generated by incorporating the Cys‐Cys‐Pro‐Gly‐Cys‐Cys tag (5′‐TGTTGTCCAGGGTGCTGC‐3′) after N‐terminal methionine. G278C mutation was generated in DNAJB1 for labeling C‐terminal domain (CTD) with Alexa 488. HPD motifs were mutated to QPN by changing H36Q + D38N and H32Q + D34N in DNAJA2 and DNAJB1, respectively, using standard PCR mutagenesis techniques and verified by sequencing. Similarly, mixed‐class JDP scaffold forming mutants were created by mutating residues D6R, E61R, E64R (RRR) in DNAJA2, and D4R, E69R, E70R (RRR) in DNAJB1 as described in Nillegoda et al., [12]. Purification of J‐proteins and their mutant variants was performed as previously described [12]. In brief, BL21(DE3)/pRARE E. coli strains carrying the corresponding expression vectors were induced for protein expression with 0.5 mm isopropyl‐1‐thio‐ d ‐galactopyranoside (IPTG; Sigma‐Aldrich Chemie GmbH, Taufkirchen, Germany) for 3 h at 30 °C. For the lysis of cells expressing human J‐domain proteins, a buffer containing 50 mm HEPES‐KOH, pH 7.5, 750 mm KCl, 5 mm MgCl₂, protease inhibitor cocktail (Roche Diagnostics GmbH, Mannheim, Germany), 2 mm phenylmethylsulphonyl fluoride and 10% glycerol was used. After centrifugation at 30 000 g (30 min, 4 °C), the resulting supernatants were applied to a Ni‐NTA/Ni‐IDA matrix (Macherey‐Nagel; Ni‐NTA, Pierce) and incubated for 60 min at 4 °C. Subsequent washing steps were performed with high‐salt buffers (50 mm HEPES‐KOH, pH 7.5, 750 and 500 mm KCl, 5 mm MgCl₂ and 10% glycerol). Protein elution was performed with 300 mm imidazole in the corresponding low‐salt buffers. Dialysis was performed overnight at 4 °C in the presence of 4 μg His‐tagged Ulp1 per mg substrate protein for proteolytic cleavage of the 6 × His–Smt3 tag. The 6 × His–Smt3 tag and His–Ulp1 were removed by incubating the dialyzed proteins in Ni‐NTA/Ni‐IDA matrix for 60 min at 4 °C. The targeted proteins were further purified using Superdex 200 column. Firefly luciferase, human HSPA8 and human HSPH2 were purified as previously described [11].

Fluorescent labeling of DNAJA2 and DNAJB1

DNAJA2 with tetracysteine motif CCPGCC and DNAJB1 G278C mutant were reduced with 50‐fold molar excess of TCEP for 30 min at room temperature and incubated with ReAsH (gift from A. Krezel, PhD) at a 1 : 1.5 protein/label ratio for 4 h at 4 °C, or 20‐fold excess of Alexa‐Fluor‐488‐maleimide for 2 h at room temperature, respectively. Progress of labeling reaction for the biarsenical dye was monitored in a spectrophotometer using A _280nm (DNAJA2 ε _280nm = 24 000 M⁻¹·cm⁻¹; DNAJB1 ε _280nm = 19 035 M⁻¹·cm⁻¹) for protein concentration and absorption at A _{590 nm} for ReAsH (ε _590nm = 68 000 M⁻¹·cm⁻¹) and A _494nm for Alexa Fluor 488 (ε _494nm = 71 000 M⁻¹·cm⁻¹). The excess of unbound dye was removed using a self‐packed Sephadex G‐25 column (GE Healthcare GmbH, Düsseldorf, Germany), and the activity of labeled proteins was confirmed by luciferase refolding assay.

FRET measurements

As previously described [12], FRET emission spectra were recorded on a Jasco FP6500 spectrofluorimeter between 510 and 650 nm, at an excitation wavelength of 488 nm for Alexa Fluor 488 (donor fluorophores). Quenching of donor fluorescence at 519 nm for Alexa Fluor 488 and an increase in acceptor emission at 608 nm for ReAsH were quantified. Acceptor fluorescence measurement was refined by subtracting the fluorescence from donor‐labeled J‐domain protein to minimize background. The Förster radius of the Alexa‐Fluor‐488–ReAsH FRET pair was calculated to 62 Å [23]. For FRET experiments, J‐domain proteins labeled with donor and acceptor fluorophores were mixed at 0.1 μm (donor) and 1 μm (acceptor) in a buffer containing 25 mm HEPES, pH 7.5, 50 mm KCl, and 5 mm MgCl₂, and allowed to equilibrate for 15 min at 30 °C before measuring the steady‐state fluorescence. FRET efficiencies were calculated based on the donor fluorescence quenching and presented as a percentage of donor fluorescence in the absence of acceptor.

Luciferase refolding‐only and disaggregation‐refolding assays

Luciferase refolding‐only and disaggregation‐refolding assays were performed as previously described [12]. For refolding‐only assays, 20 nm luciferase was incubated with 2 μm HSPA8, 100 nm HSPH2, 500 nm J‐domain protein, and 100 nm Hsp26 (or otherwise indicated) at 42 °C for 10 min in HKM buffer (50 mm HEPES‐KOH, pH 7.5, 50 mm KCl, 5 mm MgCl₂, 2 mm dithiothreitol, 2 mm ATP, and 10 μm BSA) to generate thermally denatured luciferase. Refolding was initiated by shifting the reaction to 30 °C and adding an ATP‐regenerating system (3 mm phosphoenolpyruvate and 20 ng μL⁻¹ pyruvate kinase). Luciferase activity was measured at the indicated time points using a Lumat LB 9507 luminometer (Berthold Technologies GmbH & Co. KG, Bad Wildbad, Germany) by transferring 1 μL of sample to 100 μL of assay buffer (25 mm glycylglycine, pH 7.4, 5 mm ATP, 100 mm KCl, and 15 mm MgCl₂) supplemented with 100 μL of 0.25 mm luciferin. Luciferase aggregates used for disaggregation‐refolding reactions were generated as previously described [12]. Briefly, 25 nm luciferase was incubated with a fivefold molar excess of Hsp26 at 45 °C for 15 min to generate aggregates. Disaggregation‐refolding reactions were initiated by adding the indicated chaperone mixture cocktails (e.g., 2 μm HSPA8 (or 15 μm), 100 nm HSPH2 (or 750 nm), and 500 nm J‐domain protein) to preformed luciferase aggregates followed by incubation at 30 °C. Representative graphs (n = 1) are shown for reactions containing higher HSPA8/HSPH2 concentrations.

SEC and aggregate profiling

As previously described [12], tritium (³H) labeling of firefly luciferase was performed with N‐succinimidyl‐[2,3‐³H] propionate (Hartmann Analytic) according to manufacturer's guidelines. Unincorporated N‐succinimidyl‐[2,3‐³H] propionate was removed using dialysis with HKM buffer [12] at 4 °C, overnight. ³H‐labeled luciferase was aggregated as described in the disaggregation‐refolding assays. Luciferase reactivation was performed by adding specified chaperone cocktails and incubating at 30 °C. Reactions were quenched with apyrase (0.8 μg·μL⁻¹) at 40 min and placed on ice. Aggregated luciferase species were separated using an ÄKTA purifier system with a Superose 6 Tricorn column (10/300 GL, GE Healthcare Life Sciences Europe GmbH, Freiburg, Germany). Running buffer contained 50 mm HEPES‐KOH, pH 7.5, 150 mm KCl, 5 mm MgCl₂, 2 mm DTT, and 10% glycerol. The ³H signal in each elution fraction was measured with liquid scintillation counting and normalized to the total counts of the corresponding SEC run after background subtraction and presented as a percentage of the total counts (F1–F4). A SEC standard (BIO‐RAD Laboratories GmbH, Feldkirchen, Germany) was used to determine the size of the elution peaks. Void volume contains any complexes ≥ 5000 kDa. Notably, ∼ 40–50% of the input material was lost during SEC as a result of nonspecific binding to filters and column matrix. These experiments were conducted as an extension of the work reported in Nillegoda et al., and Nillegoda et al., [12, 14].

Results

Mixed‐class JDP complex formation is not affected by HPD‐to‐QPN mutation in the J‐domain

J‐domains are ~ 70‐amino acid subdomains that adopt a conserved four‐helix bundle architecture (helices I–IV) arranged in a hairpin‐like structure [24]. Helices II and III form an antiparallel coiled‐coil stabilized by interchain hydrophobic interactions, whereas helices I and IV cap the structure at the N‐ and C‐termini, respectively. A hallmark of functional J‐domains is the conserved HPD motif located in the loop between helices II and III (Fig. 1A). This motif is essential for productive JDP‐mediated stimulation of Hsp70's ATPase activity, as shown by HPD‐disrupting mutations that abolish JDP‐Hsp70 functional coupling [21, 22]. In addition to mediating Hsp70 interaction, the J‐domains of human DNAJA2 and DNAJB1 also contribute to formation of mixed‐class JDP scaffolds (Fig. 1B) that initiate assembly of high‐capacity Hsp70 disaggregases [12, 14]. To test whether disruption of the HPD motif affects mixed‐class JDP scaffold formation, we introduced an HPD‐to‐QPN substitution into DNAJA2 (H36Q, D38N; DNAJA2^QPN). This mutation was chosen to impair productive Hsp70 recruitment/stimulation while retaining the conserved proline residue in the HPD loop, thereby minimizing the risk that loss of disaggregation activity would simply reflect broader disruption of J‐domain architecture. We then examined whether DNAJA2^QPN retained the ability to interact with wild‐type DNAJB1 using a previously established Förster resonance energy transfer (FRET) assay that monitors proximity between the DNAJA2 J‐domain region and the DNAJB1 CTD in vitro [12, 14]. For these experiments, a tetracysteine motif (CCPGCC, see Methods) was inserted immediately upstream of the DNAJA2 J‐domain and labeled with ReAsH (acceptor fluorophore), whereas DNAJB1 carrying a G278C substitution in its CTD was labeled with Alexa Fluor 488 (donor fluorophore). FRET efficiency measurements showed that both wild‐type DNAJA2 and DNAJA2^QPN mutant interacted with DNAJB1 to a similar extent under the conditions tested (Fig. 1C). Controls for fluorophore labeling positions are described in Nillegoda et al., [12]. Because FRET is a proximity‐based assay, these data do not define absolute binding affinity or stoichiometry. Rather, they show that the HPD‐to‐QPN substitution does not abolish detectable mixed‐class JDP scaffold formation through the previously defined DNAJA2–DNAJB1 interaction interfaces. This result provided the basis for testing whether scaffold formation alone is sufficient for synergistic disaggregation when productive Hsp70 recruitment by one JDP partner is impaired.

Fig. 1.

(A) Ribbon representation of the four‐helix bundle (helices I–IV) of the J‐domains of human DNAJA2 (green) and DNAJB1 (blue) derived from the alphafold prediction (AF‐P25685‐F1‐model_v6 and AF‐O60884‐F1‐model_v6). The conserved His‐Pro‐Asp (HPD) motif, depicted in ball‐and‐stick representation, is located between helices II and III. (B) Mixed‐class J‐domain protein complexes form between DNAJA2 and DNAJB1 via electrostatic interactions between the J‐domain (JD) and the C‐terminal domain (CTD) (represented by dotted lines, see Nillegoda et al., [12]). Dimerization domain labeled as D. The cartoon depicts only protomers for simplicity; adapted from Nillegoda et al., [12]. (C) FRET analysis of interactions between DNAJB1 and wild‐type DNAJA2 or the mutant DNAJA2^QPN (H36Q, D38N). Bars represent donor quenching efficiency reflecting JD‐to‐CTD interactions. Cartoons below illustrate the positions of the fluorophores (orange‐yellow stars) on J‐domain protein protomer pairs. The N‐terminus of the J‐domain of DNAJA2 was labeled with the acceptor fluorophore ReAsH, whereas the C‐terminal domain (CTD) of DNAJB1^G278C was labeled with the donor fluorophore Alexa Fluor 488 at amino acid residue position 278 (mean ± SEM from; n = 3). Controls for fluorophore labeling positions are described in Nillegoda et al., [12].

HPD‐to‐QPN mutation in DNAJA2 abolishes synergistic protein disaggregation with DNAJB1

Because mixed‐class JDP complex formation remained detectable, we next asked whether DNAJA2^QPN could still cooperate with wild‐type DNAJB1 to enhance protein aggregate disaggregation in vitro. Amorphous aggregates were generated by heat denaturation of firefly luciferase in the presence of a small heat shock protein (sHSP; see Materials and methods). sHSPs co‐aggregate with misfolded proteins and modify aggregate morphology, size, and hydrodynamic properties thereby rendering aggregates more amenable to disassembly [11, 12, 15]. Incubation of these preformed aggregates with human HSPA8 (Hsc70), DNAJA2, DNAJB1, and the Hsp110‐type nucleotide exchange factor (NEF) HSPH2 resulted in strong synergistic luciferase reactivation after disaggregation (magenta filled circle) compared with reactions containing HSPA8, HSPH2, and a single JDP class (green filled circle, DNAJA2 only; blue filled circle, DNAJB1 only; Fig. 2A). This observation is consistent with previous reports [12, 14]. In contrast, the reaction containing HSPA8, DNAJA2^QPN, DNAJB1, and HSPH2 showed (orange, open circle) a luciferase reactivation level comparable to that observed in HSPA8, DNAJB1, and HSPH2 containing reaction (Fig. 2A, left). Thus, although DNAJA2^QPN retained detectable mixed‐class scaffold formation with DNAJB1, the resulting assembly did not support the synergistic increase in disaggregation and refolding activity normally observed when both wild‐type JDPs are present. Reciprocal experiments using DNAJB1^QPN (H32Q, D34N) produced a similar loss of synergistic disaggregation activity (Fig. 2A, right). Together, these data indicate that efficient disaggregase function requires both DNAJA2 and DNAJB1 to remain competent for productive Hsp70 recruitment. Mixed‐class JDP complexes therefore appear to assemble multivalent disaggregase configurations in which both JDP partners contribute to the Hsp70 recruitment required for high‐capacity disaggregation.

Fig. 2.

(A) Representative graphs showing reactivation of preformed luciferase aggregates (disaggregation‐refolding) by the human HSPA8 (Hsc70)‐HSPH2 (Hsp110/Apg2) system in the presence of the J‐domain proteins DNAJA2 (green), DNAJB1 (blue), or both DNAJA2 and DNAJB1 (magenta), as well as reactions containing the corresponding QPN mutants. Titration of DNAJA2^QPN (H36Q, D38N) mutant against wild‐type DNAJB1 at 1 : 1 and 3 : 1 indicated in orange and red colored curves, respectively. Titration of DNAJB1^QPN (H32Q, D34N) mutant against wild‐type DNAJA2 at 1 : 1 and 3 : 1 indicated in organge and red colored curves, respectively. Reactions performed in the absence of J‐domain proteins are indicated by black circles (each replicate shown; n = 3). The sequence of heat treatment and chaperone addition is indicated in the scheme above. (B) As in (A), representative graphs from an experiment showing reactivation of preformed luciferase aggregates by the human HSPA8‐HSPH2‐class A/B J‐domain protein‐containing disaggregase systems using higher concentrations of HSPA8 (15 μm) and HSPH2 (750 nm).

Increasing the ratio of DNAJA2^QPN to DNAJB1 from 1 : 1 to 3 : 1 further reduced both the rate and final yield of luciferase reactivation (Fig. 2A, left). Similar inhibitory effects were observed in the reciprocal reactions containing DNAJB1^QPN and DNAJA2 (Fig. 2A, right), and when QPN mutants were mixed with their corresponding wild‐type proteins (Fig. S1A). We therefore describe this effect as a dose‐dependent inhibition consistent with increased incorporation of Hsp70‐recruitment‐defective JDPs into otherwise functional disaggregase assemblies. Importantly, mixing wild‐type and QPN‐mutant JDPs at these ratios did not substantially affect the ability of the wild‐type JDP to support refolding of denatured monomeric luciferase (Fig. S1), where JDP‐JDP complex formation is dispensable [12, 14].

One possible explanation for this inhibitory effect is that JDPs carrying the HPD mutation might retain weak Hsp70 binding either through their J‐domains [21] or through additional regions outside the J‐domain [25], such as the C‐terminally located EEVD motif of DNAJB1 [26], thereby forming nonproductive JDP‐Hsp70 complexes. To test whether reduced disaggregation resulted from sequestration of limiting Hsp70, we repeated the reactions using 7.5‐fold higher concentration of HSPA8 while maintaining the HSPA8:HSPH2 ratio. The inhibitory effect persisted, although to a lesser extent (Fig. 2B and Fig. S1B), indicating that reduced disaggregation cannot be explained primarily by simple Hsp70 sequestration. Another possibility is that QPN mutants compete with wild‐type JDPs for substrate binding. Though this may be possible at very high mutant JDP ratios, this is unlikely to be the main explanation under the conditions used because the inhibitory effect was minimal or absent in luciferase refolding‐only reactions performed at the same JDP concentrations (Fig. S2). The data are therefore most consistent with a model in which Hsp70‐recruitment‐defective JDPs can enter mixed‐class assemblies but produce non‐productive disaggregase configurations in which insufficient multivalent Hsp70 recruitment prevents efficient aggregate solubilization.

DNAJA2^QPN disrupts the aggregate targeting spectrum of DNAJB1

To further understand how mixed‐class JDP assemblies contribute to efficient disaggregase formation, we next examined how these systems process distinct aggregate populations generated from the luciferase substrate. We used a previously established size‐exclusion chromatography (SEC)‐based assay that monitors disaggregation of ³H‐labeled luciferase aggregate species generated by heat denaturation (Fig. 3A) [12, 14]. In this assay, aggregates elute in fractions F1–F3 according to apparent size and hydrodynamic properties, whereas solubilized luciferase monomers appear in fraction F4 (Fig. 3B) [12, 14]. This approach allowed us to examine how the different JDP configurations‐containing disaggregase systems, including those containing QPN mutants, affect distinct luciferase aggregate populations, rather than relying only on assays measuring total luciferase reactivation.

Fig. 3.

(A) Flowchart illustrating the key steps of the SEC‐based luciferase disaggregation assay. (B) Quantification of ³H‐luciferase aggregate populations across SEC fractions (F1–F3) after 40 min of disaggregation and refolding by distinct configurations of the human Hsp70 disaggregase systems consisting of HSPA8, HSPH2, DNAJA2 and/or DNAJB1 (mean ± SEM; each replicate shown, n = 3 (two‐tailed Student's t‐test; *P < 0.05). Reactions containing DNAJA2^QPN or DNAJB1^QPN alone, n = 2. Reactions containing the corresponding QPN mutants are indicated. Elution fractions are labeled F1–F4 (red lines); F4 corresponds to disaggregated luciferase monomers (~ 63 kDa). Loss of ³H‐luciferase from aggregate fractions (F1–F3) occurs concomitantly with accumulation of monomeric luciferase in F4. Additional controls for SEC‐based disaggregation assay described in Nillegoda et al., [12].

Consistent with previous findings [12, 14], Hsp70 disaggregase system reconstituted with DNAJA2 alone (Fig. 3B; filled green bars) preferentially solubilized smaller aggregates eluted in the F3 fraction (~ 200–700 kDa) within 40 min of reaction time, but showed no activity toward the medium and large size aggregate species eluted in F2 fraction (~ 700–4000 kDa) and F1 fraction (≥ 5000 kDa), respectively. Conversely, reactions containing DNAJB1 alone preferentially targeted large aggregates in F1 fraction but did not efficiently solubilize F2 or F3 aggregates within the same reaction time (filled blue bars). Thus, under these conditions, neither single‐class JDP system efficiently processed the F2 aggregate population. In contrast, the reactions supplied with both DNAJA2 and DNAJB1 depleted aggregate populations across F1–F3 and produced a corresponding ~ 4‐fold increase in monomeric luciferase in F4 fraction (filled magenta bars). Importantly, aggregates in the F2 fraction were efficiently processed within 40 min only when both JDPs were present, indicating that DNAJA2 and DNAJB1 act together to promote processing of this aggregate population (Fig. 3B).

As expected, reactions containing DNAJA2^QPN or DNAJB1^QPN alone did not support Hsp70‐dependent disaggregation, and the luciferase aggregate profile remained unchanged (Fig. 3B; open green and blue bars). However, when DNAJA2^QPN was combined with wild‐type DNAJB1, we observed not only the solubilization of the large F1 aggregate population, which can be supported by DNAJB1 alone, but also a modest but significant reduction in the F2 fraction which was unexpected (Fig. 3B; open red bars). This suggests that DNAJA2 can influence the aggregate‐processing behavior of DNAJB1 through mixed‐class JDP interaction even when DNAJA2 is defective in productive Hsp70 recruitment. In the reciprocal reaction containing wild‐type DNAJA2 and DNAJB1^QPN, no comparable reduction of the aggregates in F1 or F2 fractions was observed; only the F3 aggregate population, which can also be targeted by DNAJA2 alone (Fig. 3B; open orange bars).

To test whether the altered aggregate‐handling profile observed in the DNAJA2^QPN and DNAJB1‐containing reaction depends on mixed‐class JDP interaction, we used previously established JDP^{class A}–JDP^{class B} interaction‐disrupting charge reversion (RRR) mutants (DNAJA2^{D6R,E61R,E64R}/DNAJA2^RRR and DNAJB1^{D4R,E69R,E70R}/DNAJB1^RRR) described in Nillegoda et al., [12]. These RRR mutations disrupt the electrostatic interface driving mixed‐class scaffold formation between DNAJA2–DNAJB1 while preserving the intrinsic Hsp70‐recruitment capacity of the individual JDPs. Prior FRET analysis showed that DNAJA2^RRR mixed with wild‐type DNAJB1 retains partial interaction (~ 50% donor quenching signal from mixed‐class JDP interaction), likely mediated through the wild‐type J‐domain of DNAJB1, whereas the DNAJA2^RRR–DNAJB1^RRR combination reduced % donor quenching signal by ~ 80% [12]. The paired RRR mutant condition, which shows significant loss of interaction between DNAJA2 and DNAJB1, therefore provided an appropriate test of whether the expanded luciferase aggregate‐handling profile depends on mixed‐class JDP scaffold formation.

SEC‐based luciferase disaggregation analyses showed that DNAJA2^RRR alone preferentially processed the aggregates in F3 fraction (open green bars), whereas DNAJB1^RRR alone preferentially processed the aggregates in F1 fraction (open blue bars), broadly recapitulating the aggregate‐processing profiles of the corresponding wild‐type single‐JDP reactions (Fig. S3). When DNAJA2^RRR and DNAJB1^RRR were combined, however, we did not observe significant depletion of the F2 aggregate fraction (Fig. S3; open magenta bars). Thus, disrupting mixed‐class JDP interaction abolished the expanded aggregate‐handling profile normally observed in disaggregase systems with DNAJA2–DNAJB1 assemblies. Together, the results from these proof‐of‐principle experiments indicate that mixed‐class JDP assemblies could contribute to aggregate targeting or handling, whereas productive Hsp70 recruitment by both JDP partners in such assemblies is required for efficient aggregate solubilization. The data therefore support a model in which mixed‐class JDP scaffold formation and productive Hsp70 recruitment represent separable but coupled steps during assembly of a high‐capacity Hsp70 disaggregase, in vitro.

Discussion

Unlike the structurally well‐defined hexameric Hsp100 disaggregases present in bacteria and fungi [8, 9], the architecture of metazoan Hsp70‐based disaggregases remains poorly understood. These disaggregase systems are thought to assemble as dynamic and heterogeneous machines on aggregates that vary in size, density, composition, and type. This complexity has made it difficult to assign fixed affinities, stoichiometries, or aggregate‐bound assembly states. In the present study, we therefore used a functional biochemical strategy to define a specific assembly requirement within an Hsp70 disaggregase formed by mixed‐class JDP scaffolds.

Previous work showed that canonical class A and class B JDPs can cooperate to enhance Hsp70‐mediated disaggregation of amorphous aggregates [12, 14, 15, 16, 18, 19]. These JDPs can form mixed‐class complexes through interactions between the J‐domain of one partner and the CTD of the other [12, 14], providing a plausible scaffold for efficient Hsp70 recruitment when forming disaggregases on aggregate surfaces [1, 20, 27]. The unresolved question addressed here was whether formation of this mixed‐class scaffold is sufficient to enhance disaggregation, or whether both JDP partners must also remain competent for productive Hsp70 recruitment. Our results support the latter model. HPD‐to‐QPN mutation of either DNAJA2 or DNAJB1 impaired synergistic luciferase disaggregation, even though DNAJA2^QPN retained detectable mixed‐class interaction with DNAJB1 in the FRET assay. Thus, mixed‐class JDP scaffold formation and productive Hsp70 recruitment can be functionally separated. Scaffold formation alone is not sufficient to generate a high‐capacity disaggregase; both JDP partners must contribute to productive Hsp70 engagement.

This conclusion is consistent with models in which mixed‐class JDP assemblies increase the local density of Hsp70 molecules at aggregate surfaces. Once a mixed‐class JDP scaffold forms on an aggregate, exposed J‐domains from both JDP partners could recruit Hsp70 molecules independently, promoting multivalent Hsp70 engagement at the same aggregate region. Clustering of Hsp70 molecules has been proposed to promote polypeptide extraction through mechanisms such as entropic pulling, in which multiple chaperones bound to a substrate generate steric and entropic forces that favor release from aggregates [28, 29]. Disrupting Hsp70 recruitment by either JDP partner would be expected to weaken this multivalent recruitment mechanism, consistent with the loss of synergistic disaggregation observed here.

The SEC‐based assays further suggest that mixed‐class JDP assemblies influence aggregate handling in a manner that is not explained solely by Hsp70 recruitment. In this setup, DNAJA2 alone‐containing systems preferentially processed smaller luciferase aggregates, whereas DNAJB1 alone‐containing systems preferentially processed larger aggregates. The combination of DNAJA2 and DNAJB1 enabled efficient processing across a broader range of heat‐induced luciferase aggregate populations (based on size/density), including the F2 population that was inefficiently handled by either JDP alone.

The experiments conducted with QPN and RRR assembly‐defective mutants, together, refine this interpretation. DNAJA2^QPN, although defective in productive Hsp70 recruitment, partially altered the aggregate‐processing behavior of DNAJB1, suggesting that mixed‐class JDP interaction can influence aggregate handling independently of Hsp70 recruitment. However, this interaction was not sufficient for efficient solubilization in the absence of productive Hsp70 recruitment by both partners. Conversely, the RRR mutants that cannot effectively form JDP assemblies abolished the expanded aggregate‐handling profile, supporting the conclusion that mixed‐class JDP scaffold formation contributes to aggregate targeting, while productive Hsp70 recruitment by both JDP partners is required to promote efficient disaggregation.

Several limitations should be noted. The FRET assay used here reports proximity between labeled DNAJA2 and DNAJB1 and should not be interpreted as an absolute affinity measurement. The present experiments also do not define a fixed stoichiometry for mixed‐class JDP assemblies in solution or on aggregate surfaces. Such measurements remain challenging because the assemblies are transient, oligomeric, and likely continuously remodeled on structurally heterogeneous aggregate substrates. These unresolved parameters are important for future structural and quantitative biophysical studies, but they are not required for the central conclusion of the present work.

Together, our findings support a model in which mixed‐class JDP assemblies formed by human DNAJA2 and DNAJB1 promote Hsp70 disaggregation through two functions. First, mixed‐class JDP scaffold formation influences the handling of distinct luciferase aggregate populations. Second, productive Hsp70 recruitment by both JDP partners enables multivalent Hsp70 engagement and assembly of an efficient disaggregase. These findings identify a previously unresolved step in Hsp70 disaggregase assembly: mixed‐class JDP scaffold formation alone is insufficient for robust aggregate solubilization and must be accompanied by productive, multivalent Hsp70 engagement by both JDP partners, in vitro.

Author contributions

NBN conceived and supervised the study; NBN and AS designed experiments; NBN and AS performed experiments; NBN and AS analyzed data; NBN wrote the manuscript; NBN and AS made manuscript revisions.

Supporting information

Fig. S1. Representative graphs showing reactivation of preformed luciferase aggregates (disaggregation‐refolding) by the human HSPA8 (Hsc70)‐HSPH2 (Hsp110/Apg2) system in the presence of the J‐domain proteins DNAJA2 (green), DNAJB1 (blue), or both DNAJA2 and DNAJB1.

Fig. S2. Representative graphs showing refolding of misfolded monomeric luciferase by the human HSPA8 (Hsc70)‐HSPH2 (Hsp110/Apg2) system in the presence of the J‐domain proteins DNAJA2 (green), DNAJB1 (blue), or both DNAJA2 and DNAJB1 (magenta), as well as reactions containing the corresponding QPN mutants.

Fig. S3. Quantification of 3H‐luciferase aggregate populations across SEC fractions (F1–F3) after 40 min of disaggregation and refolding by distinct configurations of the human Hsp70 disaggregase systems consisting of HSPA8, HSPH2 and JDP scaffold forming defective mutants DNAJA2RRR (D6R, E61R, E64R) and/or DNAJB1RRR (D4R, E69R, E70R).

Data accessibility

Data available on request from the authors.