Comparative structural and functional analysis of the glycine‐rich regions of Class A and B J‐domain protein cochaperones of Hsp70

J‐domain proteins are critical Hsp70 co‐chaperones. A and B types have a poorly understood glycine‐rich region (Grich) adjacent to their N‐terminal J‐domain (Jdom). We analyzed the ability of Jdom/Grich segments of yeast Class B Sis1 and a suppressor variant of Class A, Ydj1, to rescue the inviability of sis1‐∆. In each, we identified a cluster of Grich residues required for rescue. Both contain conserved hydrophobic and acidic residues and are predicted to form helices. While, as expected, the Sis1 segment docks on its J‐domain, that of Ydj1 does not. However, data suggest both interact with Hsp70. We speculate that the Grich–Hsp70 interaction of Classes A and B J‐domain proteins can fine tune the activity of Hsp70, thus being particularly important for the function of Class B.

Hsp70 chaperone systems play critical roles in many cellular functions under both normal and stress conditionscyclically binding and releasing protein substrates in a nucleotide-dependent manner [1-3].J-domain proteins (often called JDPs or J-proteins) are essential Hsp70 co-chaperones [4,5].All perform the critical function of stimulating Hsp70's ATPase activity, converting its bound ATP to ADP, thereby driving conformational changes that stabilize interaction with substrate.This co-chaperone role is carried out by the J-domain (J dom ), which is comprised of four helices, two of which (II and III) interact with Hsp70.Many JDPs also bind substrate, orchestrating their delivery to Hsp70.
Although a single cellular compartment can have 12 or more different JDPs, Classes A and B are typically the most abundant [6].They are also among the most complex [7].Members of both classes have an N-terminal J-domain and an adjacent glycine-rich region (G rich )often called the glycine/phenylalanine or G/F region.The G rich is followed by the substrate-binding domain, two structurally similar b-barrels, often called C-terminal domain (CTD) I and II, and, at the C-terminus, a dimerization domain.However, Classes A and B b-barrel JDPs differ, both structurally and functionally.Class A members have a zinc-binding domain protruding from the more N-terminal barrel, while the most abundant cytosolic Class B members have a binding site for the C-terminal EEVD found in cytosolic Hsp70s [8,9].Clearly, Class B JDPs can carry out functions that Class A does not, such as disaggregation of amyloid fibrils [10][11][12].However, open questions remain about the basis of such functional differences.
Saccharomyces cerevisiae is a particularly advantageous model system to dissect functional differences between Classes A and B JDPs, as Class B Sis1 is essential, while Class A Ydj1, although more abundant, is not [13,14].Moreover, a 121 residue N-terminal fragment containing only the 67 residues J dom and a segment of the G rich (Sis1 121 ) allows growth of cells having a deletion of SIS1 (sis1-Δ) [15].Although Ydj1 cannot normally substitute for Sis1, that is, support growth of sis1-Δ cells, a single residue change at the junction of the J dom and G rich of Ydj1 enables growth, whether in full-length Ydj1 or a 109 residue N-terminal fragment containing only the J dom and G rich (Ydj1 109 G70N) [16].This substitution destabilizes the junction but appears to have no effect on helices II/III of the J dom or distal G rich sequences.
Residues in both the J dom and G rich of Sis1 121 are critical for Sis1 121 function [17].Importance of the J dom is unsurprising, as it has long been known to play a fundamental role in all JDPs [18].G rich segments are much less studied.However, recent reports have pointed to unanticipated complexity in this region.The extreme proximal region of the G rich of both Sis1 and its human ortholog DNAJB1 forms a short helical region that folds back and interacts with adjacent Helix IV of the J dom [9,10].In the case of DNAJB1, a distal helical portion of the G rich (called Helix V) loops back and interacts with the face of Helices II and III that interacts with Hsp70, thus inhibiting J dom binding to Hsp70 [10].In the context of full-length protein, this inhibitory intramolecular interaction is released upon DNAJB1 binding to Hsp70's EEVD.
Results to date point to functional importance of Sis1's G rich , but a thorough understanding of how it functions has remained elusive.To better understand G rich function, we carried out a combination of in vivo genetic and in vitro nuclear magnetic resonance (NMR) comparisons of J dom /G rich fragments of Sis1 and Ydj1, exploiting the ability of Ydj1 109 G70N to support growth of cells lacking Sis1.
Escherichia coli strains DH5a and Rosetta 2 (DE3) pLysS were used for plasmid construction and protein purification, respectively, and cultured using Luria broth or M9 minimal media with 15 NH 4 Cl and 13 C-glucose added for 15 N or 15 N 13 C sample labeling (Cambridge Isotope Laboratories, Inc., Andover, MA, USA) [22], along with the addition of necessary antibiotics for plasmid maintenance.Plasmids are listed in Table S1.All site-directed mutations were made using the QuikChange protocol (Stratagene, La Jolla, CA, USA) and verified by sequencing; and restriction enzymes were obtained from New England BioLabs, Ipswich, MA, USA.

Protein purification
Escherichia coli strains were grown at 37 °C, induced at OD 600nm 0.6 by addition of 0.5 mM isopropyl ß-D-1-thiogalactopyranoside (IPTG), and incubated overnight at 18 °C.After harvesting, cells were lysed using a T2 cell disrupter (Constant Systems, Ltd, Horthants, UK).Slide-A--Lyzer dialysis cassettes (Thermo Scientific, Rockford, IL, USA) were used for dialysis and Amicon Ultra Centrifugal Filter units (Merck Millipore Ltd., Burlington, MA, USA) for sample concentration.Final protein concentrations were determined using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific) and purity was assessed using SDS/PAGE followed by Coomassie blue staining.
To purify Ssa1T201A, cells expressing the His 6 -SUMO-tagged version were resuspended in buffer A (25 mM HEPES pH 7.5, 150 mM NaCl, 30 mM KCl, 10 mM Mg (C 2 H 3 O 2 ) 2 , 25 mM imidazole, 1 mM DTT, and 5% (v/v) glycerol) with protease inhibitor added.After lysis and clarification spin, supernatant was applied to HisPur NiNTA resin (Thermo Fisher) and protein was purified as recommended by the manufacturer including a wash step using buffer A containing 750 mM NaCl and elution with buffer A containing 300 mM imidazole.Peak fractions were pooled and treated with His 6 -tagged Ulp1 protease overnight at 4 °C.The resulting mixture was incubated with 3.5 mL ATP-agarose beads (Sigma Chemical Co.) in batch for 2 h at 4 °C.After transferring to the column, the beads were washed sequentially with 35 mL of buffer B [25 mM HEPES pH 7.5, 100 mM NaCl, 30 mM KCl, 10 mM Mg (C 2 H 3 O 2 ) 2 , 1 mM DTT, and 5% (v/v) glycerol], 35 mL of buffer B containing 650 mM NaCl, followed with 20 mL of buffer B. Ssa1T201A was eluted with buffer B containing 5 mM ATP. Peak fractions were pooled and concentrated using filter unit with 50 kDa cutoff and passed over a PD10 column (GE Healthcare, Chicago, IL, USA) equilibrated with the final storage buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 30 mM KCl, and 1 mM DTT).Final sample was concentrated and stored at 4 °C with addition of protease inhibitor and NaN 3 (0.05%, w/v).

Analysis of Helix V conservation and length
To analyze sequence conservation of G rich segments homologous to helices V of Sis1 and Ydj1, we assembled a data set of Sis1/DNAJB1 and Ydj1/DNAJA1 orthologs from fungi and metazoan.Because both Sis1 and Ydj1 have multiple orthologs in Metazoa, and to ensure a fair comparison, we used OMA groups [24], which are groups of sequences related exclusively by speciation events, including at most one sequence per species per group.A set of Sis1, DNAJB1, Ydj1, and DNAJA1 protein sequences were gathered from OMA database (OMAGroup: 681110, 683249, 681430, 681954) [24].DnaJ and CbpA sequences from E. coli and Scj1 and DNAJB11 sequences from fungi and metazoans, respectively, were added as outgroups.Sequences were then aligned using CLUSTAL OMEGA v1.2.2 with default parameters [25].Protein phylogeny was calculated with 1000 maximum-likelihood (ML) searches using iQtree with 1000 rapid bootstrap replicates [26], under LG model of amino acid substitution with FreeRate model of rate heterogeneity using eight discrete rate categories (LG + R8) [27], which was determined as the bestfit model by Bayesian Akaike information criterion [28].For sequence conservation analysis, Sis1 and Ydj1 orthologs were grouped based on their clustering on the tree (Fig. S4) and realigned, while sequences that did not cluster with either Sis1 or Ydj1 were removed from the analysis.Sequence conservation was calculated separately for Sis1/DNAJB1 and Ydj1/D-NAJA1 orthologs.Alignment logos were generated using the WebLogo server [29].
To analyze the length of Helix V, we used AlphaFold structural models from the AlphaFold Protein Structure Database [30].Available AlphaFold structural models were found, based on UniProt ID, for each analyzed sequence from the data set of Ydj1/DNAJA1 and Sis1/DNAJB1 orthologs.Next, secondary structure was inferred for each residue of each structure using DSSP software [31].The position of Helix V in each analyzed structure was found based on the position of the conserved sequence block in the multiple-sequence alignments of Ydj1/DNAJA1 and Sis1/DNAJB1 orthologs.A five amino acid sequence was added, both upstream and downstream, to each conserved block to avoid missing any helical residues within the region containing Helix V.The number of residues forming the longest helix within this region was calculated using a custom Python script.Statistics were calculated and plotted with Seaborn [32].

NMR data collection, processing, and analysis
NMR spectra were collected using Varian VNMRS and Bruker Avance III spectrometers operating at 600, 800, or 900 MHz, and equipped with cryogenic triple-resonance probes.Protein samples were labeled uniformly with 15 N or 15 N and 13 C. NMRPipe [33] software was used to process NMR spectra and analyzed via NMRFAM-Sparky [34] software.
For Sis1 121 backbone resonance assignments, data were acquired from the following two-dimensional (2D) and three-dimensional (3D) NMR experiments collected at 25 °C: 2D 1 H, 15 N HSQC, 3D HNCO, 3D HNCACB, 3D CBCA(CO)NH, 3D HNCA, and 3D HN(CO)CA.All 3D spectra were recorded using nonuniform sampling (NUS) with a sampling rate of 50%.The software package SMILE [35], available in NMRPipe [33], was used to reconstruct spectra from the NUS data.The assignments were determined using previously described approach [16], utilizing automatic APES and PINE algorithms followed by manual verification via PINE-SPARKY and deposited in Biological Magnetic Resonance Bank (BMRB ID: 52234).To resolve the overlap of resonances in the G rich , an additional pair of high-resolution 3D HNCA and 3D HN(CO) CA spectra was used as described previously [16].Apart from the first methionine, this approach allowed for unambiguous assignment of all 112 nonproline Sis1 121 residues for Sis1 121 B and 77 residues for Sis1 121 U .The relative fractions of Sis1 121 B and Sis1 121 U were determined based on the volumes of signals corresponding to T39 and K32 residues and measured using NMRFAM-SPARKY [34] software.Previously determined assignments of all 103 nonproline residues of Ydj1 109 (besides D36) (BMRB ID: 52286) and 88 residues of Ydj1 109 G70N (BMRB ID: 52287) were used [15].Assignments for 74 of 75 nonproline residues of Sis1 81 were available (BMRB ID: 30457).Assignments of 71 of 75 resides of Ydj1 79 were determined using available assignments for Ydj1 J dom (BMRB ID: 30293) and Ydj1 109 [15] as reference.Secondary structure prediction based on chemical shifts for Sis1 121 and Ydj1 109 was carried out using TALOS-N [36].
The chemical shift perturbation (CSP) analysis was performed by comparing the positions of signals in 2D 1 H 15 N HSQC spectra of indicated proteins, recorded at 25 °C, or 10 °C if compared with Ydj1 109 G70N.Chemical shift differences were determined in the 1 H (d H ) and 15 N (d N ) dimensions, and were converted to a consensus chemical shift difference ( D d HN ), as shown in the bar graphs, using the equation: To determine backbone dynamics, steady-state heteronuclear 1 H- 15 N NOEs (nOe) values were calculated from the ratio of signal intensities between two spectra recorded with and without 3 s proton saturation, using a standard 1 H, 15 N HSQC-based pulse sequence.Spectra were collected at 25 °C for Sis1 121 and at 10 °C for Ydj1 109 WT and Ydj1 109 G70N.
In experiments involving Hsp70, 1 H-15 N HSQC spectra were collected for 15 N-labeled Ydj1 and Sis1 variants individually or in the presence of equimolar concentration (300 lM) of nonlabeled Ssa1 T201A variant.Experiments were conducted in buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 5 mM DTT, 2 mM ATP, 2 mM Mg(C 2 H 3 O 2 ) 2 , 0.02% NaN 3 , and 7% 2 H 2 O) at 10 °C and 15 °C for Ydj1 and Sis1 variants, respectively.The ratio of signal intensity (I/I 0 ) for each residue, where I and I 0 are signal intensities for Hsp70-bound and -free JDP, was plotted in form of the bar graphs.

Miscellaneous
All chemicals used were purchased from Sigma-Aldrich (St. Louis, MO) unless noted otherwise.

Results
Residues in G rich of Ydj1 109 G70N are required for rescue of sis1-Δ We first asked whether the G rich of Ydj1 109 G70N is required for its ability to substitute for Sis1.We constructed a set of 8 Ydj1 109 G70N variants having consecutive deletions of 4-6 residues across its G rich (from residues 71 to 109).Two, 92-95Δ and 96-99Δ, did not support growth of sis1-Δ cells; the others, both the four more proximal and the two more distal, had little or no effect on sis1-Δ growth compared to those expressing Ydj1 109 G70N with an intact G rich .To dissect this functionally important part of the G rich , we individually substituted residues between positions 88 and 101.Two variants (alanine substitution of phenylalanine at position 95 or 98) were unable to support growth of sis1-Δ (Fig. 1, Fig. S1).Slow growth, compared to that resulting from expressing Ydj1 109 G70N was observed when two other hydrophobic residues, I94 and F99, were substituted.In addition, alteration of D92 or D93, the only negatively charged residues in this region, severely compromised Ydj1 109 G70N function.Thus, both hydrophobic and negatively charged G rich residues are important for the ability of Ydj1 109 G70N to substitute for Sis1.As Sis1 121 also requires distal G rich sequences to support sis1-Δ growth [17], these results raise the question of whether the Class A Ydj1 G rich and the Class B Sis1 G rich act in a similar manner.As described in the sections below, we addressed this issue by comparing their structural properties.

The G rich of Sis1 docks on helices II and III of its J-domain
Using solution NMR, we first asked if the G rich of Sis1 folds back and binds to its J dom , as is the case for mammalian DNAJB1 [10].Because such an interaction would affect the position of signals (i.e., chemical shift) in an 1 H- 15 N HSQC spectrum, we compared signal positions of J dom residues in spectra of Sis1 121 with those of Sis1 81 , which contains the entire J dom , but lacks most of the G rich .Significant differences were observed for a cluster of J dom residues in Helices II and III (Fig. 2A, upper panel) consistent with G rich docking on the J dom , forming a "bound" conformation (Sis1 121

B
).These results are also consistent with the recent Alpha-Fold model of Sis1 [37,38], which predicts that a distal 12 residue segment of the Sis1 G rich (residues 108-119) forms a helix that folds back upon Helices II and III of the J dom (Fig. 2B).In analogy to DNAJB1 nomenclature [10], we call this Helix V of Sis1 throughout.
In addition to the Sis1 121 B signals, we identified a set of minor signals 1 H- 15 N HSQC corresponding to approximately 10% of the molecule population in the sample (Fig. 2C; Fig. S2).We found no significant differences in position between these minor signals and those of the J dom of Sis1 81 (Fig. 2A, lower panel and C).Furthermore, the main differences in positions between these minor signals and those of Sis1 121 B were clustered in Helices II and III of the J dom and Helix V (Fig. S2).These results are consistent with the minor signals representing "unbound" Sis1 121 having no intramolecular J dom -G rich interaction (called Sis1 121 U throughout).Together our results indicate that Sis1 121 exists in two conformations-one in which Helix V of the G rich interacts with the J dom , as observed for DNAJB1 [10], and one having no intramolecular interaction between the J dom and the G rich .

The G rich of Ydj1 does not dock on Helices II and III of its J dom
We carried out an analogous analysis of Ydj1 J dom / G rich -comparing 1 H- 15 N HSQC spectra of the 109 residue fragment to a 79 residue fragment lacking the majority of the G rich (Ydj1 79 ).No significant differences in the position of Helix II or III J dom residue signals were observed (Fig. 3A).We also tested Ydj1 109 G70N.As expected, based on our previous analysis [16], the Ydj1 109 G70N and Ydj1 79 comparison showed differences in/near Helices I and IV due to their close proximity to the G70N substitution (Fig. S3).However, like the Ydj1 109 comparison to Ydj1 79 , no differences were observed for residues in Helices II and III.These results suggest that Ydj1's G rich does not fold back on Helices II and III of its J-domain in a manner similar to Sis1's G rich .
However, our genetic results (Fig. 1) demonstrate that residues in the distal G rich region are important for Ydj1 109 G70N's ability to support growth of cells lacking Sis1.We therefore decided to investigate further the structural character of Sis1 and Ydj1 J dom /G rich fragments.To gain insight into each residue's dynamics, we recorded NMR heteronuclear 1 H- 15 N NOE (nOe) experiments.As expected, both J dom were found to be rigid, while significant dynamics were observed within both G rich .The Helix V region of the Sis1 G rich had a level of rigidity comparable to the J dom (Fig. 3B), as expected based on its helical character and interaction with the J-domain.
Unexpectedly, Ydj1's distal region between residues ~90 and 100 also displayed rigidity, although not to the same extent as that observed for Sis1.TALOS-N secondary structure prediction based on NMR chemical shifts suggests a short helical region (residues 94-97) within this segment that exhibits rigidity (Fig. S4).Together these experimental results are consistent with the formation of a short, likely quite transient, helix.
The AlphaFold [37,38] model of Ydj1 predicts a helix within that segment (residues 92-99) that does not interact with its J dom (Fig. 3C).By analogy, we call this region Helix V throughout (Figs 3 and 4).
Conserved helices of Sis1 121 and Ydj1 109 G rich are critical for growth of sis1-Δ To better compare the two G rich , we decided to carry out a more systematic substitution analysis of Sis1's G rich as shown in Fig. 1 for Ydj1.Two important residues of Sis1 121 's G rich were identified previously-N108 and D110 [17], but these were from a nonsaturating screen employing random mutagenesis.Therefore, we changed, individually, each residue between G102 and G121 to alanine in Sis1 121 and tested the ability of each variant to sustain growth of sis1-Δ (Fig. 4A; Fig. S5).In addition to the expected growth defects caused by substitutions in the 108-110 regions, deleterious effects were also observed for more distal changes-notably, upon substitution of phenylalanines.Sis1 121 F115A did not support growth, while F118A and F119A cells were temperature sensitive.Comparing the Ydj1 and Sis1 segments of the G rich regions of Ydj1 109 G70N and Sis1 121 important for supporting growth of sis1-Δ cells, we noted similarities.Both are predicted to form helices, with negatively charged amino acids followed by hydrophobic amino acids, mainly phenylalanines (Fig. 4B).However, Sis1 Helix V is longer (12 residues compared to 8 residues for Ydj1) with important phenylalanines more distant from negatively charged residues.
Because of the unexpected presence of a functionally important segment of Ydj1 G rich with residues of similar character to those in Sis1, we asked if these segments were conserved.We assembled and analyzed two datasets that included sequences from Fungi and Metazoa-one having 442 Ydj1/DNAJA1 orthologs and one having 411 Sis1/DNAJB1 orthologs (Fig. S6; Table S2).We then generated sequence profiles, which revealed conservation of acidic, followed by four hydrophobic residues in both-with an expanded distance between them in Sis1/DNAJB1 compared to Ydj1/DNAJA1 orthologs, as seen in S. cerevisiae Sis1 and Ydj1 (Fig. 4B).Furthermore, analysis of available AlphaFold structural predictions indicate that the conserved sequence forms a helix, with a median two residue longer helix length for Sis1/DNAJB1 compared to Ydj1/DNAJA1 (Fig. S7).Together our data support the idea that Helix V of the G rich region of both Ydj1 and Sis1 is important for the ability of their J dom /G rich fragments to support growth of sis1-Δ cells and is conserved in Ydj1 and Sis1 orthologs.
Evidence of interaction of Sis1 121 and Ydj1 109 G rich with Hsp70 Ydj1 contains residues in the G rich segment important for function, even though the J dom /G rich Ydj1 fragment shows no intramolecular interaction with Helices II and III of the J dom in our analyses.This implies that the G rich interacts intermolecularly with another entity in the cell.Hsp70 is an obvious candidate interactor.During productive cycles of substrate protein-binding Hsp70 is very close to the G rich because of its proximity to the J dom .To begin to test this idea, we compared the 1 H-15 N HSQC spectra of Ydj1 109 in the presence and absence of Hsp70, since interaction results in significant broadening, resulting in an apparent reduction in signal intensity.In addition to the expected dramatic reduction for Ydj1 109 J dom residues, a significant depression in signal intensity was observed for residues in the more distal region of the G rich that encompasses Helix V (Fig. 5A).Similar results were obtained when Ydj1 109 G70N was tested.
When an analogous experiment was carried out for Sis1 121 , signals from residues in the distal G rich that encompass Helix V were less in the presence of Hsp70 (Fig. 5B).This was also the case for J dom residues.However, the suppression of J dom signals was not as severe as that observed for Ydj1.Reasoning that the J dom -G rich intramolecular interaction diminished the availability of the J dom for Hsp70 interaction, we tested a shorter J dom -containing fragment that lacked most of the G rich (Sis1 81 ).As expected, the J dom signals of Sis1 81 dramatically broadened upon Hsp70 addition.Although these results do not conclusively demonstrate a direct interaction of Helix V with Hsp70, they are consistent with such an interaction, in addition to that of the J dom of both Ydj1 and Sis1.

Discussion
The most straightforward hypothesis consistent with our genetic and biophysical data is that Helix V of the G rich engages in two functionally important interactions.One is the previously identified inhibitory action caused by intramolecular interaction with the J dom in Class B JDPs.The second, uncovered here through the analysis of the G rich of Ydj1 for which no J dom intramolecular interaction is detected, is intermolecular.This second interaction-likely with Hsp70-occurs in both Class A and Class B JDPs.
The intramolecular interaction of the G rich of Sis1 with its J dom corresponds to that observed for its mammalian ortholog DNAJB1 [10]-folding back onto the surfaces of Helices II and III that are required for Hsp70 interaction.This J dom -Helix V interaction may be less robust for yeast Sis1 than for human DNAJB1, at least for the truncated variant, as we also observed an apparent undocked conformation in our NMR analysis, albeit as a minor population.Perhaps it is the less robust character of the intramolecular interaction that allowed detection of Hsp70 interaction with both the J-domain and Helix V of the Sis1 121 fragment-as the sites of interaction are inaccessible in the bound conformation.
A functional role for this intramolecular interaction in the minimal system exploited here having only a segment of the G rich and the J dom is questionable.But in the much more architecturally complex full-length protein, this J dom -G rich interaction has been reported to serve the role of inhibiting J dom binding to Hsp70 until the inhibition is released via effects of CTD binding to the C-terminal EEVD of Hsp70 [10].But another Class B JDP having a structurally different C-terminal substrate-binding domain that does not bind the Hsp70 EEVD (DNAJB6) has a similar intramolecular interaction [39], suggesting that the J dom -G rich interaction may serve regulatory functions more generally.
Our hypothesis that Helix V is responsible for both intramolecular J dom and intermolecular Hsp70 interactions makes it tempting to speculate that the ability of G rich regions to interact with Hsp70 arose first, and later in evolution mutations were selected for that allowed intramolecular interaction.Interestingly, a repeated motif of aspartic acid, isoleucine and phenylalanine in the G rich region of DnaJ, the E. coli Class A ortholog of Ydj1, (called the "DIF" repeat) has been reported to be important for certain functions, including growth at cold-and warm-temperature extremes [40].In previous studies, no interaction of this repeat with DnaJ's J dom was found, but it did exhibit decreased flexibility [41] and has been predicted to form a helix [37,38].It is possible that insertion occurred in Class B JDPs that maintained the relative positioning of residues in Helix V but facilitated interaction with the J dom .
Further analysis is required to determine whether such interactions can occur simultaneously or whether they are mutually exclusive, as well as the nature and function of the Hsp70 interaction.Because Sis1 is known to have unique functional ability not normally carried out by Ydj1 [16,42], it is likely that such an interaction modulates substrate capture by fine-tuning Hsp70's conformational transition and/or substrate binding, but such ideas remain speculation.Alternatively, it is possible that in both bacterial and eukaryotic Class A JDPs, Helix V forms only transiently and weakly interacts intramolecularly, but thus far such an interaction has not been detected experimentally.
Regardless, the results presented here, exploiting the G70N alteration in Ydj1 and those from E. coli DnaJ indicate that sequences in this region are functionally important in Class A JDPs.Indeed, many questions pertaining to Classes A and B JDP function remain.Pertinent to this report is the junction between the J dom and G rich .Both molecular modeling [37,38] and structural studies [16] suggest that a proximal segment of the G rich region of both Sis1 and Ydj1 forms a structural junction that includes J dom residues of Helices I and IV, as well as the first few residues of G rich .NMR results, both nOe presented here and CSP reported previously [16], indicate that the G70 alteration that allows Ydj1 109 to substitute for Sis1 destabilizes this region.Likely such an alteration increases the flexibility or reach of the G rich .This is interesting in light of the fact that the spacing between the end of the J dom and Helix V is typically longer in Sis1 than in Ydj1 orthologs.It is also important to note that, while we have exploited the ability of minimal Ydj1 109 G70N and Sis1 121 fragments to support growth of sis1-Δ cells, neither do so, as well as full-length Sis1 [16]-and Ydj1 109 G70N less well than the Sis1 121 .Overall, the flexibility of the G rich likely plays a pivotal role in fine-tuning Hsp70 action by modulating both intramolecular and intermolecular interactions of Helix V.

Fig. 1 .
Fig. 1.G rich of Ydj1 109 G70N is required for rescue of inviability of sis1-Δ.Analysis of ability of variants having short deletions (top) or single-amino-acid substitutions (bottom) within the G rich of Ydj1 109 G70N to support growth of sis1-Δ.Ten-fold serial dilutions of sis1-Δ cells expressing Ydj1 109 G70N with no additional alteration (wt) or indicated change in the G rich or, as a control, no Ydj1 109 G70N (-) were plated and grown for 2 days at the indicated temperature.(middle) Diagram of Ydj1 109 G70N, with the red arrow indicating position of G70N substitution: J dom (blue) and G rich (green).Deletions tested across the G rich are indicated in red above the diagram; single residues between F88 and A101 are shown below the diagram, those substituted in red, with changes indicated beneath.

Fig. 3 .
Fig. 3.No detection of Ydj1 G rich intramolecular interaction with J dom Helices II and III.(A) Comparison of 1 H-15 N HSQC spectra of fragments of Class A JDP Ydj1.(top) Diagram, analogous to Fig. 2A, of Class A JDP Ydj1: same domains as in Sis1, plus zinc-binding domain (ZnBD), gold.(middle) Enlarged 1-79 segments of Ydj1 (Ydj1 79 ) showing secondary structure arrangement-four a-helices of J dom (HI-HIV) plus short helical extension into G rich region (green).(bottom) Change in signal position of each residue of Ydj1 79 compared to Ydj1 109 , calculated as the chemical shift perturbations (CSP) in ppm with smallest (≤ 0.1) indicated in gray.(B) NMR heteronuclear nOe analysis illustrating dynamics of each residue.(top) Ydj1 109 and Ydj1 109 G70N, closed and open circles, respectively.(bottom) Sis1 121 B .Cartoon indicates that higher and lower nOe values are associated with a more rigid or more flexible structure, respectively.Error bars calculated as standard deviation for each signal are included but in most cases obscured by data points.(C) Ydj1 109 structure from AlphaFold full-length model (ID: AF-P25491-F1).Note, lack of intramolecular interaction between J dom (light blue) and G rich (green), but presence of a helix (HV).

Fig. 4 .
Fig. 4. Conservation of G rich Helix V in Class A and Class B JDPs. (A) Analysis of growth of sis1-Δ cells expressing Sis1 121 variants having substitution within the distal portion of the G rich between G102 and G121.Positions changed to alanine, with exception of A105 and A111 changed to glycine.Controls: no Sis1 121 mutation (wt); no Sis1 sequences in the plasmid (-).Cells were plated (two plates for each temperature) and incubated for 2 days at the indicated temperature.(B) Helix V (HV) of Sis1 and Ydj1 G rich and its conservation.Sequences of G rich segments encompassing HV of Ydj1 and Sis1 are shown, with phenotypic effect of individual substitutions summarized by circles whose shading is based on results in Fig. 1 (Ydj1) and 4A (Sis1): lethality (black), temperature sensitive (gray), no effect (white), or if not tested (dotted-line).HV conservation presented as sequence logo generated based on the sequence alignments of 442 Ydj1 orthologs (Class A) and 411 Sis1 orthologs (Class B) from Fungi and Metazoa.

1474FEBS
Letters 598 (2024) 1465-1477 ª 2024 The Authors.FEBS Letters published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies.