1. Biochemistry and Chemical Biology
  2. Evolutionary Biology

Evolution towards simplicity in bacterial small heat shock protein system

  1. Piotr Karaś
  2. Klaudia Kochanowicz
  3. Marcin Pitek
  4. Przemyslaw Domanski
  5. Igor Obuchowski
  6. Barlomiej Tomiczek  Is a corresponding author
  7. Krzysztof Liberek  Is a corresponding author
  1. Intercollegiate Faculty of Biotechnology UG-MUG, University of Gdansk, Poland
https://doi.org/10.7554/eLife.89813.3
Share this article
Cite this article
  1. Piotr Karaś
  2. Klaudia Kochanowicz
  3. Marcin Pitek
  4. Przemyslaw Domanski
  5. Igor Obuchowski
  6. Barlomiej Tomiczek
  7. Krzysztof Liberek
(2023)
Evolution towards simplicity in bacterial small heat shock protein system
eLife 12:RP89813.
https://doi.org/10.7554/eLife.89813.3
  2. Download BibTeX
  3. Download .RIS

Abstract

Evolution can tinker with multi-protein machines and replace them with simpler single-protein systems performing equivalent functions in an equally efficient manner. It is unclear how, on a molecular level, such simplification can arise. With ancestral reconstruction and biochemical analysis, we have traced the evolution of bacterial small heat shock proteins (sHsp), which help to refold proteins from aggregates using either two proteins with different functions (IbpA and IbpB) or a secondarily single sHsp that performs both functions in an equally efficient way. Secondarily single sHsp evolved from IbpA, an ancestor specialized in strong substrate binding. Evolution of an intermolecular binding site drove the alteration of substrate binding properties, as well as the formation of higher-order oligomers. Upon two mutations in the α-crystallin domain, secondarily single sHsp interacts with aggregated substrates less tightly. Paradoxically, less efficient binding positively influences the ability of sHsp to stimulate substrate refolding, since the dissociation of sHps from aggregates is required to initiate Hsp70-Hsp100-dependent substrate refolding. After the loss of a partner, IbpA took over its role in facilitating the sHsp dissociation from an aggregate by weakening the interaction with the substrate, which became beneficial for the refolding process. We show that the same two amino acids introduced in modern-day systems define whether the IbpA acts as a single sHsp or obligatorily cooperates with an IbpB partner. Our discoveries illuminate how one sequence has evolved to encode functions previously performed by two distinct proteins.

eLife assessment

This valuable study advances our understanding of the evolution of protein complexes and their functions. Through convincing experimental and computational methodologies, the authors show that the specialization of protein function following gene duplication can be reversible. The work will be of interest to investigators working in biochemical evolution and those working on heat shock proteins.

https://doi.org/10.7554/eLife.89813.3.sa0

Introduction

Gene birth and loss is a hallmark of protein family evolution, however, molecular determinants and genetic mechanisms enabling that process are not well understood (Fernández and Gabaldón, 2020; Worth et al., 2009). Gene loss and differential retention of paralogues reshapes the divergence of organisms both in Animalia, Archaea, and Bacteria (Fernández and Gabaldón, 2020; Iranzo et al., 2019; Puigbò et al., 2014). Examples of gene loss often involve adaptive changes in response to changing environmental niches, like loss of genes encoding olfactory receptors in primates (Demuth and Hahn, 2009) or differential retention of paralogous genes encoding venom toxins in different rattlesnake lineages (Dowell et al., 2016). In prokaryote genomes, gene loss is one of the main evolutionary processes accelerating sequence divergence leading to functional innovations (Puigbò et al., 2014). In complex protein systems, execution of a cellular function can be shared between several proteins. In bacteria, the cost of maintaining additional gene copy is very high and maintaining a low gene count is important for keeping the replication energy costs low (Kempes et al., 2017; Lever et al., 2015; Lynch and Marinov, 2015). Still, it remains unclear how a multi-protein system can undergo simplification. Here, we asked what are the molecular events that enabled the gene loss, and how one of the biochemical functions has been taken over by the other protein. We investigated these questions using sHsp system, which underwent simplification within Enterobacterales (which include common bacteria species like Escherichia coli, Salmonella enterica and Erwinia amylovora) as a model (Figure 1).

Figure 1
Download asset Open asset
small heat shock protein (sHsp) systems in Enterobacteriacrae and Erwiniaceae.

Left - schematic phylogeny of sHsps in Enterobacterales. Gene duplication resulting in IbpA + IbpB two-protein system is marked with a star, while the loss of ibpB gene in Erwiniaceae clade is marked with a cross. AncA0 - reconstructed last common ancestor of IbpA from Erwiniaceae and Enterobacteriaceae, expressed as a part of two-protein system. AncA1 - reconstructed last common ancestor of secondarily single IbpA from Erwiniaceae. Right - representative extant sHsps’ ability to stimulate luciferase refolding. sHsps were present during the luciferase thermal denaturation step. Refolding of denatured luciferase was performed by the Hsp70-Hsp100 chaperone system. Activity of luciferase was measured after 1 hr refolding at 25 °C and shown as an average of 3 repeats. Error bars represent standard deviation.

Figure 1—source data 1

Spreadsheet containing raw data for the luciferase refolding graphs shown in Figure 1.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig1-data1-v1.xlsx

sHsps are a family of ATP-independent molecular chaperones present in all living organisms with various copy numbers (ten representatives in human) (Haslbeck and Vierling, 2015). They bind misfolded proteins and sequester them into refolding – prone assemblies, preventing uncontrolled aggregation and helping to maintain proteostasis at stress conditions. sHsp is composed of a highly conserved α–crystallin domain (ACD), in a form of so-called β–sandwich, flanked by less conserved, unstructured N- and C-terminal regions (Haslbeck and Vierling, 2015; Haslbeck et al., 2019; Reinle et al., 2022). Their smallest functional unit is usually a dimer, formed by the interaction between ACDs of two neighboring sHsps. Stable sHsp dimers in turn tend to form variable and dynamic higher-order oligomers, stabilized by N- and C-terminal region interactions. Particularly the interaction between the IXI motif, highly conserved in sHsps C-termini, and the cleft formed by β4 and β8 strands of ACD is critical for oligomer formation (Haslbeck and Vierling, 2015; Kennaway et al., 2005; Mani et al., 2016; Simmons and Ochoterena, 2000). Oligomers of bacterial sHsps reversibly dissociate into smaller forms when the temperature increases. It is considered their activation mechanism, probably uncovering substrate interaction sites. The mechanism of sHsps’ interaction with misfolded substrates is not yet fully understood, but both N-termini and β4–β8 cleft region have been found to play a role in this process (Basha et al., 2006; Fuchs et al., 2009; Jaya et al., 2009; Lee et al., 1997; Reinle et al., 2022).

In most Enterobacterales a two-protein sHsps system exists, consisting of IbpA and IbpB proteins (Mogk et al., 2003; Obuchowski et al., 2019). IbpA and IbpB have originated via duplication, form a heterodimer partnership, and are functionally divergent from one another (Obuchowski et al., 2019; Piróg et al., 2021; Figure 1A and B). IbpA is specialized in tight substrate binding (sequestrase activity), while IbpB is required for the dissociation of both sHps from the aggregates, a step necessary to initiate Hsp70-Hsp100 dependent substrate disaggregation and refolding (Obuchowski et al., 2021; Ratajczak et al., 2009). In a subset of Enterobacterales (Erwiniaceae), as a result of ibpB gene loss, the secondarily single-protein sHsp (IbpA) system has emerged (Figure 1). The term ‘secondarily single’ is used in order to distinguish it from single-protein IbpA from clades in which the duplication did not occur (for example Vibrionaceae; Obuchowski et al., 2019). How did IbpA evolve to become independent of its partner? In this study, using ancestral reconstruction, we identify mutations, which allowed the secondarily single IbpA to be fully functional without its partner in substrate sequestration and handover to Hsp70-Hsp100-mediated disaggregation and refolding.

Results

New activity of Erwiniaceae IbpA has evolved in parallel to IbpB gene loss

To better understand the evolution of sHsps after gene loss, we reconstructed the IbpA ancestors from before and after the loss of its IbpB partner. This technique uses multiple sequence alignments of modern-day proteins from different species to infer amino acid sequences of its common ancestors (Ashkenazy et al., 2012; Pupko et al., 2002) and is widely used to investigate various evolutionary questions (Gaucher et al., 2008; Longo et al., 2020; Thomson et al., 2005; Thornton et al., 2003). We created a multiple sequence alignment of 77 IbpA sequences from Enterobacterales (Supplementary file 1), from which we inferred the phylogeny of IbpA using the maximum likelihood method (Figure 2, Supplementary file 2 – phylogenetic tree in Newick format). From that, we inferred ancestral sequences, which have the highest probability of producing the modern-day sequences using the empirical Bayes method (Ashkenazy et al., 2012; Cohen and Pupko, 2011; Cohen et al., 2008; Pupko et al., 2002; Simmons and Ochoterena, 2000). Next, we resurrected (i.e. expressed and purified) the last ancestor of IbpA present before (AncA0) and after (AncA1) the differential gene loss (Figure 2, Supplementary file 3).

Figure 2
Download asset Open asset
IbpA phylogeny in Enterobacterales.

Phylogeny was reconstructed from 77 IbpA orthologs from Enterobacterales using the Maximum Likelihood algorithm with JTT + R3 substitution model. AncA0 - node representing the last common ancestor of IbpA from Erwiniaceae and Enterobacteriaceae. AncA1 - node representing the last common ancestor of IbpA from Erwiniaceae. Bootstrap support is noted for the major nodes. Extant IbpAs from E. coli and E. amylovora are marked with a red frame. Scale bar - substitutions per position.

Similarly, to modern-day IbpA proteins both AncA0 and AncA1 were fully folded, and reversibly deoligomerized into smaller species under elevated temperature (Figure 3—figure supplement 1). Moreover, both ancestral proteins were able to sequester aggregating firefly luciferase in sHsp-substrate assemblies. AncA0 exhibited sequestrase activity on the level comparable to IbpA from Escherichia coli (IbpAE. coli). AncA1 was moderately efficient in this process and IbpA from Erwinia amylovora (IbpAE.amyl) was the least efficient sequestrase (Figure 3A). The differences in sequestrase activity were especially pronounced at lower sHsp concentrations. Next, we tested their ability to bind protein aggregates in real time (Figure 3B). Ancestral proteins’ interaction with the aggregated substrates was stronger than in the case of extant E. amylovora IbpA, but weaker than in the case of extant E. coli IbpA (Figure 3B).

Figure 3 with 3 supplements see all
Download asset Open asset
Functional changes during the evolution of secondarily single small heat shock protein (sHsp) in Erwiniaceae.

(A) Sequestrase activity of extant and ancestral sHsps. Luciferase was heat denatured in the presence of different concentrations of sHsps and size of formed sHsps-substrate assemblies was measured by DLS. For every measurement, ten subsequent series of ten 10-s runs were averaged and particle size distribution was calculated by fitting to 70 size bins between 0.4 and 10,000 nm. Results are presented as an average hydrodynamic diameter of measured particles weighted by intensity (Z-average), calculated with Malvern Zetasizer Software 7.13. Error bars represent standard deviation obtained from the distribution. (B) Binding of extant and ancestral sHsps to heat-aggregated E. coli proteins. E. coli proteins were heat aggregated and immobilized on a Biolayer interferometry (BLI) sensor. sHsps were heat-activated before the binding step. (C) Extant and ancestral sHsps’ ability to stimulate luciferase refolding. Experiment was performed at 25 °C. Luciferase activity at each timepoint was shown as an average of 3 repeats. Error bars represent standard deviation.

Figure 3—source data 1

Spreadsheet containing raw data for the sequestrase activity graph shown in Figure 3A.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig3-data1-v1.xlsx
Figure 3—source data 2

Spreadsheet containing raw data for the binding and dissociation curve shown in Figure 3B.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig3-data2-v1.xlsx
Figure 3—source data 3

Spreadsheet containing raw data for the luciferase refolding graph shown in Figure 3C.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig3-data3-v1.xlsx

Finally, we asked how the modification of the substrate aggregation process by reconstructed proteins influences subsequent substrate refolding by the Hsp100 and Hsp70 chaperones. AncA1 stimulated luciferase refolding, however, its effectiveness was around half of both analyzed extant sHsp systems (single IbpA from E. amylovora or IbpA + IbpB system from E. coli), similar to extant IbpA form E. coli without its IbpB partner. AncA0, in contrast, inhibited luciferase refolding in comparison to control (no sHsps at substrate aggregation step) (Figure 3C).

To test the robustness of our observations, we repeated the analysis for the alternative ancestors, which have the second highest probability to produce the modern-day sequences (AltAll) (Figure 3—figure supplement 2, supplementary file 3 – posterior probabilities) (Eick et al., 2017). Both AltAll variants behaved similarly to most likely (ML) variants in reversible deoligomerization, sequestrase activity and stimulation of substrate refolding assays (Figure 3—figure supplement 3A-D). However, the last property (the influence on refolding) required higher Hsp70 system concentration to observe AltAll variants activity (Figure 3—figure supplement 3C, D). Together, these data show that reconstructed ML and AltAll ancestors are functional. What is of particular interest, these data clearly point out that the ability of reconstructed sHsps to stimulate Hsp70-Hsp100-dependent substrate refolding arose between A0 and A1 nodes.

We performed a molecular evolution analysis to test for positive selection across IbpA phylogeny using both branch models and branch-site models in codeml (Jeffares et al., 2015; Yang, 1998; Yang, 2007; Yang and Nielsen, 2002). The analysis shows a significantly increased ratio of nonsynonymous to synonymous substitutions, after the gene loss, at the branch leading to A1 with both tests. This result indicates that the new IbpA functionality likely arose due to an episode of positive selection rather than genetic drift (Figure 4A, Supplementary file 4 A,B). The result of the branch-site test indicates possible positive selection acting on all sites substituted at the branch leading to A1 with pp >0.5, therefore, we aimed to identify the minimum number of mutations that are responsible for a change in functional properties of IbpA.

Figure 4 with 2 supplements see all
Download asset Open asset
Substitutions at positions 66 and 109 that occurred between nodes A0 and A1 are crucial for ancestral small heat shock proteins (sHsps) to work as a single protein.

Luciferase refolding assay was performed as in Figure 1. Activity of luciferase was measured after 1 hr refolding at 25 °C and shown as an average of 3 repeats (6 in the case of no sHsp and proteins AncA0­ and AncA1 in panel C). Error bars represent standard deviation. (A) Schematic phylogeny of Enterobacterales IbpA showing increased ratio of nonsynonymous to synonymous substitutions (ω) on the branch between nodes AncA0 and AncA1. Loss of cooperating IbpB is marked on a tree. Value of the Likelihood Ratio Test (LRT) is given for the selection model. (B) Identification of substitutions necessary for AncA0 to obtain AncA1-like activity in luciferase disaggregation; seven candidate mutations were introduced into AncA0 (AncA0 + 7); subsequently, in series of six mutants, each of the candidate positions was reversed to a more ancestral state (AncA0 + 6* variants) (C) Effect of substitutions at positions 66 and 109 on the ability of AncA0 and AncA1 to stimulate luciferase refolding.

Figure 4—source data 1

Spreadsheet containing raw data for the luciferase refolding graph shown in Figure 4B.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig4-data1-v1.xlsx
Figure 4—source data 2

Spreadsheet containing raw data for the luciferase refolding graph shown in Figure 4C.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig4-data2-v1.xlsx

Identification of residues defining ancestral sHsps activities

In order to identify amino acids responsible for the observed new functionality of AncA1, we compared the sequences of two ancestral proteins, selecting 7 out of 10 substitutions as probable candidates. Three substitutions were removed from the analysis based on the low conservation of these positions in extant proteins (Figure 4—figure supplement 1). The remaining seven were introduced into AncA0. Resulting AncA0 + 7 protein stimulated Hsp70-Hsp100-dependent luciferase refolding at the level comparable to AncA1 (Figure 4B). To further specify key mutations, we prepared seven additional variants. In each variant, a different position in AncA0 + 7 was reversed to a more ancestral state. The substantial decrease in luciferase refolding stimulation was observed for positions 66 and 109 (Figure 4B). Next, each of these substitutions on its own (Q66H or G109D) was separately introduced into AncA0. This was not sufficient to increase AncA0 ability to stimulate luciferase refolding. However, when both substitutions were introduced simultaneously, the resulting sHsp exhibited activity similar to AncA1 (Figure 4C). What is more, when in AncA1 these two positions were reversed to AncA0-like state, the resulting sHsp lost the ability to stimulate luciferase refolding (Figure 4C). All analyzed proteins, namely AncA0 + 7, AncA0 Q66H G109D and AncA1 H66Q D109G, possess biochemical properties characteristic for sHsps, exhibiting reversible thermal deoligomerization and sequestrase activity (Figure 4—figure supplement 2 A,B). All these results show that substitutions Q66H and G109D are both sufficient and necessary for the increase in activity observed for ancestral sHsps between A0 and A1 nodes.

Identified substitutions influence α-crystallin domain properties

Substitutions Q66H and G109D, responsible for gaining single sHsp activity, are located in the ACD within β4 and β8 strands, which form a cleft responsible for the interaction with unstructured C-terminal peptide of the neighboring sHsp dimer (Figure 5A). In order to identify possible structural underpinnings of the single sHsp activity we have predicted structures of AncA0 and AncA0 Q66H G109D ACD dimers in complex with C-terminal peptide using AlphaFold2 and in silico mutagenesis and subjected them to 0.5 µs equilibrium molecular dynamics (MD) simulations. Analysis of the C-terminal peptide interface contact probabilities in MD trajectories showed that both substituted residues contact the C-terminal peptide, although the overall contact pattern remain similar upon their introduction (Figure 5—figure supplement 1) and no major differences in the overall ACD domains structure were observed (Figure 5—figure supplement 2). To explore the possibility that identified substitutions affect the strength of this interaction, we analyzed the binding of purified ACDs of AncA0 and AncA0 Q66H G109D to C-terminal peptide using biolayer interferometry. Titrations of immobilized C-terminal peptides by different ACDs (Figure 5B, Figure 5—figure supplement 3) allowed us to determine the dissociation constants. These two substitutions increased the K0.5 of ACD binding to the C-terminal peptide from 4.3 µM to 7.1 µM at the same time increasing the Hill coefficient of the interaction from 2.3 to 3.7, indicating a modest decrease in affinity, accompanied by an increase in binding cooperativity.

Figure 5 with 5 supplements see all
Download asset Open asset
Substitutions at positions 66 and 109 decreased the affinity of AncA0 ACD to C-terminal peptide and aggregated substrate.

(A) Structural model of complex formed by AncA0 Q66H G109D ɑ-crystallin domain dimer (purple and lilac) and AncA0 C-terminal peptide (orange). (B) Effect of Q66H G109D substitutions (green) on AncA0 (purple) ACD’s affinity to the C-terminal peptide assayed by Biolayer interferometry (BLI). Biolayer thickness at the end of the association step was used to calculate the fraction of bound peptide. Filled circles represent means of triplicate measurements, individual data points are shown as hollow circles and were fitted to the cooperative binding model (Hill equation). Values of fitted binding affinities [K0.5] (AncA0 4.3 ± 0.2 µM, AncA0 Q66H G109D 7.1 ± 0.2 µM) and Hill coefficients [n] (AncA0 2.3 ± 0.17, AncA0 Q66H G109D 3.7 ± 0.34) are indicated on the plot. (C) Effect of Q66H G109D substitutions on AncA0 ACD’s affinity to aggregated E. coli proteins bound to BLI sensor. Analysis was performed as in Figure 3A.

Figure 5—source data 1

PDB file of the structural model of complex formed by AncA0 Q66H G109D ɑ-crystallin domain dimer and AncA0 C-terminal peptide shown in Figure 5A and Figure 5—figure supplement 2.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig5-data1-v1.zip
Figure 5—source data 2

Spreadsheet containing raw values of biolayer thickness at the end of the association step (for binding to the C-terminal peptide and nonspecific binding to His6-SUMO), used for fitting the curve shown in Figure 5B.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig5-data2-v1.xlsx
Figure 5—source data 3

Spreadsheets containing raw data for binding and dissociation curves used to obtain biolayer thickness at the end of association curve values shown in Figure 5—source data 2.

Different repeats as well as curves representing nonspecific binding of α–crystallin domains (ACDs) to His6-SUMO are in separate sheets. Repeat 1 was used as a representative binding curves shown in Figure 5—figure supplement 3.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig5-data3-v1.xlsx
Figure 5—source data 4

Spreadsheets containing raw data for binding and dissociation curves shown in Figure 5C.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig5-data4-v1.xlsx

As this interaction is known to play a crucial role in the formation of sHsp oligomers (Fu et al., 2005; Mani et al., 2016; Simmons and Ochoterena, 2000), we used dynamic light scattering to investigate how Q66H and G109D substitutions influence the size of oligomers formed by AncA0 at different temperatures. In agreement with decreased affinity between ACD and the C-–terminal peptide, we have shown that these substitutions slightly decrease the oligomer size and facilitate AncA0 deoligomerization (Figure 5—figure supplement 4 A,B).

β4-β8 cleft in certain sHsps, in addition to its interaction with C-terminal peptide, was also shown to participate in interactions with substrates and partner proteins (Fuchs et al., 2009; Jaya et al., 2009; Lee et al., 1997; Reinle et al., 2022). Therefore, we tested whether the ACD of AncA0 binds protein aggregates and whether this interaction is influenced by Q66H G109D substitutions. We observed that AncA0 ACD efficiently binds to either aggregated E. coli lysate or aggregated luciferase, and this binding was weakened by analyzed substitutions (Figure 5C, Figure 5—figure supplement 5). This suggests that ACD of bacterial sHsps interacts with the substrate, most likely through the β4-β8 cleft.

These results allow us to conclude that substitutions Q66H and G109D in AncA0 substantially increased the sHsp ability to stimulate Hsp70-Hsp100-dependent substrate refolding by weakening the interaction of β4-β8 cleft with both the C-terminal peptide and the aggregated substrates. Despite its ability to bind aggregated substrates in biolayer interferometry assay, analyzed ACDs do not exhibit sequestrase activity and were unable to positively influence substrate refolding by the Hsp70-Hsp100 system (Figure 5—figure supplement 5B, C).

Identified substitutions define the mode of action of extant sHsps

As more ancestral, AncA0-like state in positions 66 and 109 is conserved in IbpA of E. coli while more modern, AncA1-like state is conserved in IbpA of E. amylovora, we decided to ask whether this difference is sufficient to explain functional differences between the two extant proteins. Therefore, we introduced AncA1-like substitutions into IbpAE.coli and AncA0-like substitutions into IbpAE.amyl. Resulting IbpAE.coli Q66H G109D, in comparison to wild-type IbpAE.coli, exhibited an increased ability to stimulate Hsp70-Hsp100-dependent luciferase refolding, as well as a decreased ability to bind aggregated substrates becoming more similar to modern E. amylovora IbpA. At the same time, IbpAE.amyl H67Q D110G significantly less efficiently stimulated luciferase refolding, while exhibiting increased ability to bind aggregated substrates in comparison to wild-type IbpAE.amyl (Figure 6A–C). Still, both new IbpA variants exhibited properties characteristic of sHsps, namely reversible thermal deoligomerization and sequestrase activity (Figure 6—figure supplement 1A, B).

Figure 6 with 1 supplement see all
Download asset Open asset
Differences at positions 66 and 109 determine functional differences between extant IbpA proteins from E. coli and E. amylovora.

(A) Effect of substitutions at position 66 and 109 (and homologous) on the ability of IbpA from E. amylovora and E. coli to stimulate luciferase refolding. Assay was performed as in Figure 1B. Activity of luciferase was measured after 1 h refolding at 25 °C and shown as an average of 3 repeats. Error bars represent standard deviation. (B, C) Effect of substitutions at analyzed positions on binding of IbpA from E. coli (B) and E. amylovora (C) to heat - aggregated E. coli proteins. Assay was performed as in 3 A. (D–H) Effect of substitutions at analyzed positions on inhibition of Hsp70 system binding to aggregates by extant small heat shock proteins (sHsps) (D) Experimental scheme. (E–H) Aggregate-bound sHsps differently inhibit Hsp70 binding. Biolayer interferometry (BLI) sensor with immobilized aggregated luciferase and aggregate bound sHsps was incubated with Hsp70 or buffer (spontaneous dissociation curve).Dark gray traces (dashed line) represent Hsp70 binding to immobilized aggregates in the absence of sHsps. Results are presented as an average of 5 (in the case of IbpA E. coli in panel G), 4 (in the case of IbpA E. coli Q66H G109d with Hsp70 in panel H), 2 (in the case of IbpA E. coli Q66H G109d without Hsp70 in panel H) or 3 repeats (in the case of the remaining curves). Error bars (shades) represent standard deviation.

Figure 6—source data 1

Spreadsheet containing raw data for the luciferase refolding graph shown in Figure 6A.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig6-data1-v1.xlsx
Figure 6—source data 2

Spreadsheet containing raw data for binding and dissociation curve shown in Figure 6B.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig6-data2-v1.xlsx
Figure 6—source data 3

Spreadsheet containing raw data for binding and dissociation curve shown in Figure 6C.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig6-data3-v1.xlsx
Figure 6—source data 4

Spreadsheet containing raw data for binding and dissociation curves shown in Figure 6E.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig6-data4-v1.xlsx
Figure 6—source data 5

Spreadsheet containing raw data for binding and dissociation curves shown in Figure 6F.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig6-data5-v1.xlsx
Figure 6—source data 6

Spreadsheet containing raw data for binding and dissociation curves shown in Figure 6G.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig6-data6-v1.xlsx
Figure 6—source data 7

Spreadsheet containing raw data for binding and dissociation curves shown in Figure 6H.

https://cdn.elifesciences.org/articles/89813/elife-89813-fig6-data7-v1.xlsx

Above results suggest that tight sHsp binding to aggregates negatively affects the subsequent Hsp70-Hsp100-dependent substrate refolding process. It is initiated by binding of the Hsp70 system (DnaK and cochaperones DnaJ and GrpE) to aggregates that require sHsps to be outcompeted from aggregates (Żwirowski et al., 2017). To gain insight into the competition between sHsps and Hsp70 we modified the biolayer interferometry experiments and introduced the sensor with sHsps bound to luciferase aggregates into a buffer containing Hsp70 system (Figure 6D). Although biolayer interferometry cannot distinguish between proteins bound to the sensor, we took advantage of the differences in the thickness of the protein layers specific for sHsp or Hsp70 binding and also in the binding kinetics. The analysis of the Hsp70 binding to the aggregates covered with sHsps clearly shows that the presence of IbpAE.amyl or IbpAE. coli Q66H G109D on aggregates only weakly inhibits Hsp70 binding (Figure 6E and H). In contrast, the inhibition is much more pronounced when IbpAE.amyl H67Q D110G or IbpAE. coli are present on aggregates (Figure 6F and G).

All the above experiments indicate that two specific amino acids in positions 66 and 109 in ACD of IbpA proteins define the mode of IbpA activity. Glutamine 66 and glycine 109 are characteristic of IbpA proteins which bind tightly to substrates and thus are not easily outcompeted from the aggregates by Hsp70s. Such IbpAs require IbpB partner cooperation to function properly. Substitutions at these positions to histidine (position 66) and aspartic acid (position 109) allowed for the emergence of a single sHsp which binds to aggregating substrate less tightly and can be outcompeted from the aggregates by Hsp70s in the absence of IbpB.

Discussion

In this study, we traced, at the molecular level, how the two-protein sHsp system, a part of the cellular protein refolding machinery, underwent simplification in a way that its biochemical functions are performed by a single protein. In most Enterobacterales two sHsps (IbpA and IbpB) drive the sequestration of misfolded proteins into the reactivation-prone assemblies (Obuchowski et al., 2019). Together, IbpA and IbpB form a functional heterodimer, in which IbpA specializes in substrate binding, preventing the substrates from creating large aggregates (sequestrase activity), while IbpB promotes sHsps dissociation from the aggregates required for subsequent Hsp70-Hsp100-dependent substrate refolding (Obuchowski et al., 2019; Piróg et al., 2021). We showed that, in parallel to the ibpB gene loss in Erwiniaceae, new functions of IbpA have emerged, i.e., a lower substrate sequestrase activity, which correlates with the efficient substrate refolding. We have identified two amino acid substitutions (Q66H and G109D) responsible for this new IbpA functionality. Selection analysis shows that these two substitutions were likely driven by positive selective pressure. This indicates that this change possibly had an adaptive character in the ancestral background. It is important to note, however, that models used for the analysis do not account for the variation of synonymous substitution rate and multinucleotide substitution events, which in some cases might lead to false positive results (Lucaci et al., 2023). Because of that, the alternative hypothesis that substitutions Q66H and G109D occurred in the common ancestor of Erwiniaceae due to genetic drift, enabling the subsequent loss of IbpB gene, cannot be fully discounted. Functional differences observed between modern-day sHsps from E. coli and E. amylovora are at least partially defined by the presence of specific amino acids in these two positions and can be diminished by their swapping between extant proteins. Their occurrence in the last common ancestor of Erwiniaceae IbpA resulted in a decreased affinity of ACD’s β4-β8 cleft to aggregated substrates as well as to the C-termini of the other sHsps. These interactions might be of particular importance for the stabilization of sHsps on a surface of sequestered aggregated substrates leading to the formation of so-called protective shell preventing further uncontrolled aggregation. Apparent role of C-termini and ACD interaction is in agreement with earlier studies, showing that the addition of free C-terminal peptide causes E. coli IbpA and IbpB dissociation from the outer shell of sHsp - substrate complex (Żwirowski et al., 2017). It was also shown that, in case of Hsp 16.6 from cyanobacterium Synechocystis, substitutions that slightly weakened the interaction between ACD and C-termini lead to increased stimulation of luciferase refolding. However, abolishing this interaction resulted in a non-functional protein, most likely due to the loss of the sequestrase activity (Giese and Vierling, 2002). Destabilization of the protective shell of secondarily single IbpA by weakening these interactions had an effect functionally analogous to a role of IbpB in the two-protein system, facilitating IbpA dissociation from the substrate (Obuchowski et al., 2019). Two identified substitutions also weaken the ACD interaction with aggregated substrates which is an additional factor shifting the sHsps balance towards dissociation, a step necessary to initiate Hsp70-Hsp100 dependent disaggregation and refolding.

Our results show how ACD substitutions can fine-tune sHsp system by exerting pleiotropic effects on ACD-C-terminal peptide and ACD-substrate interactions. These relatively small changes strongly influence the effectiveness of sHsp functioning in complex process of aggregated protein rescue by molecular chaperones. It might be particularly important in the case of a conserved interaction, like the one between C-terminal peptide and ACD, when excessive changes of affinity may be detrimental to the overall protein function (Giese and Vierling, 2002). Our approach enabled us to find functional residues in sHsp system, which would not have been possible by using conventional mutagenesis and highlights the importance of using the vertical approach in biochemical studies. This study closely follows the evolutionary process, in which mutations in one of the two cooperating proteins tinker it to a point where it becomes independent of its partner, enabling the simplification of a more complex system by partner loss while maintaining its overall function.

Following, with molecular precision, the genetic events associated with the gene loss allowed us to answer several questions about protein family evolution. The first question concerns how the lost function is incorporated into a remaining partner protein. The results of our experiments indicate that even though the primary function of a partner protein is maintained, it is altered so that it does not interfere with the newly incorporated one. In our case substrate binding properties, as well as the formation of higher-order oligomers were altered in a way to keep the sequestrase function maintained but allow for more efficient stimulation of Hsp70-Hsp100-dependent substrate refolding. The second question concerns the context-dependence of mutations in protein evolution. We successfully transplanted the mutations that appeared in E. amylovora into E. coli IbpA ortholog, artificially creating an efficient single protein system from a protein that normally needs a partner for efficient substrate refolding. In contrast to other studies (Mogk et al., 2003) the context of E. coli IbpA protein did not influence the ability of IbpAE.coli Q66H G109D to work without IbpB partner. This brings the question about the accessibility of adaptive solutions. In case of E. coli IbpA, there was no need for adaptive changes because the gene loss did not occur in this clade, suggesting that the loss of a protein can push another one towards an adaptation, which leads to finding efficient molecular innovations.

Materials and methods

Key resources table
Reagent type (species) or resourceDesignationSource or referenceIdentifiersAdditional information
Software, algorithmClustal OmegaClustal OmegaRRID:SCR_001591European Bioinformatics Institute
Software, algorithmiq-treeIQ-TREE
Software, algorithmFastMLFastMLRRID:SCR_016092Tel Aviv University
Software, algorithmcodemlcodemlRRID:SCR_014932Part of PAML software package
Software, algorithmGromacs 2019.2GROMACSRRID:SCR_014565
Software, algorithmAlphaFold-MultimerAlphaFold-MultimerColabFold implementation
Software, algorithmPyMolPyMolRRID:SCR_000305
Recombinant DNA reagentpET3a – AncA0 (plasmid)This paperpET3a plasmid carrying AncA0 gene was ordered from Genscript on the basis of reconstructed amino acid sequence. For the sequence of synthesized gene see Supplementary file 5
Recombinant DNA reagentpET3a – AncA1
(plasmid)		This paperpET3a plasmid carrying AncA1 gene was ordered from Genscript on the basis of reconstructed amino acid sequence. For the sequence of synthesized gene see Supplementary file 5
Recombinant DNA reagentpET3a – AncA0alt_all
(plasmid)		This paperpET3a plasmid carrying AncA0alt_all gene was ordered from Genscript on the basis of reconstructed amino acid sequence For the sequence of synthesized gene see Supplementary file 5
Recombinant DNA reagentpET3a – AncA1alt_all
(plasmid)		This paperpET3a plasmid carrying AncA1alt_all gene was ordered from Genscript on the basis of reconstructed amino acid sequence. For the sequence of synthesized gene see Supplementary file 5
Recombinant DNA reagentpET3a – AncA0 + 7
(plasmid)		This paperpET3a plasmid carrying AncA0 + 7 gene was ordered from Genscript on the basis of reconstructed amino acid sequence. For the sequence of synthesized gene see Supplementary file 5
Recombinant DNA reagentpET3a – IbpAEa
(plasmid)		This paperpET3a plasmid carrying gene encoding Erwinia amylovora IbpA protein was ordered from Genscript on the basis of amino acid sequence obtained from PDB database (accession number E5BAR7). For the sequence of synthesized gene see Supplementary file 5
Recombinant DNA reagentpET28a-AncA0 ACD (plasmid)This paperpET28a plasmid carrying AncA0 ACD gene was ordered from Genscript on the basis of reconstructed amino acid sequence. For the sequence of synthesized gene see Supplementary file 5
Recombinant DNA reagentpET28a-AncA0 Q66H H109D ACD (plasmid)This paperpET28a plasmid carrying AncA0 ACD Q66H G109D gene was ordered from Genscript on the basis of reconstructed amino acid sequence. For the sequence of synthesized gene see Supplementary file 5
Recombinant DNA reagentChampion pET SUMO Expression System (plasmid)Thermo Fisher ScientificCat. # K30001
Recombinant DNA reagentpET28a – His6-SUMO-C-peptide (plasmid)This paperpET28a plasmid carrying His6-SUMO-C-peptide gene was ordered from Genscript on the basis of reconstructed amino acid sequence of AncA0 C-terminal peptide. For the sequence of synthesized gene see Supplementary file 5
Peptide, recombinant proteinOuantiLum Recombinant LuciferasePromegaCat. # E1701
Peptide, recombinant proteinCreatin KinaseRocheCat. # 10127566001
Peptide, recombinant proteinDnaKDOI:10.1016 /j.jmb.2008.12.009DnaK protein from E. coli
Peptide, recombinant proteinDnaJDOI:10.1016 /j.jmb.2008.12.009DnaJ protein from E. coli
Peptide, recombinant proteinClpBDOI:10.1016 /j.jmb.2008.12.009ClpB protein from E. coli
Peptide, recombinant proteinIbpAEcDOI:10.1016 /j.jmb.2008.12.009IbpA protein from E. coli
Peptide, recombinant proteinIbpBEcDOI: 10.1016 /j.jmb.2021.167054IbpB protein from E. coli
Peptide, recombinant proteinGrpEDOI: j.jmb.2008.12.009GrpE protein from E. coli
Peptide, recombinant proteinHis-tagged luciferaseDOI: 10.1371/journal.pgen.1008479
Commercial assay or kitLuciferase Assay SystemPromegaCat. # E151A, E152A
Strain, strain background (Escherichia coli)E. coli BL21 (DE3)therLaboratory collection

Reconstruction of IbpA phylogeny

Request a detailed protocol

Amino acid sequences of 77 IbpA orthologs from Enterobacterales were obtained from NCBI and UniProt databases and aligned using Clustal Omega (Sievers et al., 2011). Alignment was trimmed manually. JTT + R3 was identified as the best-fit model by iq-tree, using Bayesian Information Criterion, and was used in the analysis (Kalyaanamoorthy et al., 2017; Nguyen et al., 2015). The phylogenetic tree was inferred using iq-tree on the basis of 328 iterations of ML search with 100 rapid bootstraps replicates (Nguyen et al., 2015).

Reconstruction of ancestral IbpA amino acid sequences

Request a detailed protocol

Ancestral sequence reconstruction was performed on the basis of multiple sequence alignment of 77 amino acid sequences of IbpA orthologs from Erwiniaceae and Enterobacreriaceae as well as a phylogenetic tree of those orthologs (see above). Marginal reconstruction of ancestral sequences was performed with FastML program based on ML algorithm and Bayesian approach using JTT substitution matrix with gamma parameter (Ashkenazy et al., 2012; Cohen and Pupko, 2011; Cohen et al., 2008; Jones et al., 1992; Pupko et al., 2002; Simmons and Ochoterena, 2000).

Alternative ancestral sequences for AncA0 and AncA1 proteins were obtained by substituting the most likely amino acid on every uncertain position (defined as a position with more than one amino acid with posterior probability ≥0.2) with the amino acid with the second highest posterior probability (Eick et al., 2017).

Analysis of natural selection

Request a detailed protocol

Analysis of natural selection was performed using codeml. First, Pal2Nal was used to obtain codon alignment based on the multiple sequence alignment of amino acid sequences of IbpA orthologs from Enterobacterales (see above) as well as corresponding nucleotide sequences obtained from the NCBI database. Resulting codon alignment was then trimmed manually and used together with the phylogenetic tree obtained earlier (see above) for the selection analysis.

For branch model analysis, models M0 (null hypothesis) and two-ratio (with either AncA0-AncA1 branch or Erwiniaceae clade as foreground) were used. For branch-site model analysis, models A null (null hypothesis) and A were used, with foreground branches selected as above. Statistical significance of different models was estimated with Likelihood Ratio Test (LRT) (Jeffares et al., 2015; Yang, 1998; Yang, 2007; Yang and Nielsen, 2002).

Protein purification

Purification of IbpA proteins

Request a detailed protocol

pET3a plasmids containing AncA0, AncA1, AncA0 + 7, AncA0 alt_all, AncA1alt_all, and IbpAEa genes were ordered from GeneScript (Supplementary file 5). Point mutations were introduced using site-directed mutagenesis and confirmed by sequencing. Proteins were overproduced in E. coli BL21(DE3). Cells were then lysed by sonication in Qsonica sonicator (13% amplitude, 2 min 30 s process time, 15 s pulse-ON time, 45 s pulse-OFF time) in lysis buffer L1 (50 mM Tris pH 7.5, 50 mM NaCl, 5 mM EDTA, 10% glycerol, 5 mM β-mercaptoethanol). Insoluble fraction containing proteins of interest was separated by centrifugation (75,000 x g, 30 min, 4 °C) and resolubilized in buffer A (40 mM Tris pH 7.5, 50 mM NaCl, 10% glycerol, 5 mM β-mercaptoethanol, 6 M urea) and then centrifuged (75,000 x g, 30 min, 4 °C). Supernatant was loaded on Q-Sepharose chromatography column equilibrated with buffer A and eluted in 50 mM–500 mM NaCl gradient. Fractions containing proteins of interest were then dialyzed to buffer B (40 mM Tris pH 8.5, 50 mM NaCl, 10% glycerol, 5 mM β-mercaptoethanol) and loaded on Q-Sepharose chromatography column equilibrated with buffer B. Flow-through fraction was collected and dialyzed to buffer C (50 mM Tris pH 7.5, 150 mM KCl, 5% (v/v) glycerol, 5 mM β-mercaptoethanol).

Purification of ACD domains

Request a detailed protocol

ACDs of IbpAEc, AncA0, and AncA0 Q66H G109D were purified as described previously (Piróg et al., 2021) and as a final step dialyzed to buffer G (50 mM Tris pH 7.5, 150 mM KCl, 5 mM β-mercaptoethanol).

Purification of His6-SUMO and His6-SUMO-C-terminal peptide of AncA0 construct

Request a detailed protocol

His6-SUMO was purified using the Champion pET SUMO Expression System. pET28a plasmid containing gene encoding His6-SUMO fused with C-terminal peptide of AncA0 (PEAMKPPRIEIN) was ordered from GeneScript (Supplementary file 5). Proteins were overproduced in E. coli BL21(DE3). Cells were then lysed by sonication in Qsonica sonicator (20% amplitude, 2 min process time, 5 s pulse-ON time, 10 s pulse-OFF time) in lysis buffer L2 (40 mM Tris pH 7.5, 100 mM NaCl, 10% glycerol, 10 mM imidazole, 2 mM β-mercaptoethanol). Insoluble fractions were separated by centrifugation for 30 min at 70,000 x g and supernatants, containing proteins of interest, were incubated for 1 hr with Ni-NTA resin equilibrated with buffer L2. Resins were then washed with the buffer D (40 mM Tris pH 7.5, 100 mM NaCl, 10% glycerol, 40 mM imidazole, 2 mM β-mercaptoethanol) Proteins of interest were eluted from the columns with the buffer E (40 mM Tris pH 7.5, 100 mM NaCl, 10% glycerol, 400 mM imidazole, 2 mM β-mercaptoethanol) and then dialyzed to buffer C (as above).

DnaK, DnaJ, GrpE, ClpB, and IbpAEc proteins were purified as described previously (Ratajczak et al., 2009). IbpBEc protein was purified as described previously (Piróg et al., 2021). His-tagged luciferase used for BLI measurements was purified as described previously (Obuchowski et al., 2019).

Purity of purified proteins was assessed with SDS-PAGE electrophoresis with Coomassie Blue staining. Protein concentrations were measured using the Bradford reaction, with Bovine Serum Albumin as a standard. In the case of His6-SUMO fused with C-terminal peptide of AncA0, concentration was measured with SDS-PAGE electrophoresis with Coomassie Blue staining coupled with densitometric analysis with Bovine Serum Albumin used as a standard.

OuantiLum Recombinant Luciferase was purchased from Promega. Creatin Kinase from rabbit muscle was purchased from Roche.

Luciferase refolding assay

Request a detailed protocol

1,5 μM recombinant firefly luciferase in buffer F (50 mM Tris pH 7.5, 150 mM KCl, 20 mM MgCl2, 2.5 mM DTT) was denatured by incubation for 10 min at 44 °C alone or in the presence of 10 μM sHsps (3 uM IbpAEc + 7 μM IbpBEc in the case of two-protein system from E. coli). Denatured luciferase was then incubated at 25 °C with Hsp70 system (1 μM DnaK, 0.4 μM DnaJ, and 0.3 μM GrpE), 2 μM ClpB, and ATP regeneration system (5 mM ATP, 0.1 mg/ml creatine kinase and 18 mM creatine phosphate). For experiment presented in Figure 3—figure supplement 3D, higher concentration of the Hsp70 system was used (2 μM DnaK, 0.8 μM DnaJ, and 0.6 μM GrpE). At different timepoints luciferase activity was measured with GLOMAX 20/20 luminometer, using the Luciferase Assay System from Promega. Results are presented as averages of at least three independent repeats ± standard deviation.

DLS measurements

Request a detailed protocol

Dynamic Light Scattering measurements were performed using Malvern Instruments ZetaSizer Nano S instrument, at 40 μl sample volume, scattering angle of 173° and wavelength of 633 nm. For every measurement, a minimum of 10 subsequent series of ten 10 s runs were averaged and particle size distribution was calculated by fitting to 70 size bins between 0.4 and 10,000 nm, as previously described (Żwirowski et al., 2017).

For reversible deoligomerization assay, size of oligomers formed by 10 μM sHsps in buffer F (50 mM Tris pH7.5, 150 mM KCl, 20 mM MgCl2, 2.5 mM DTT) were measured by DLS. First measurement was performed at 25 °C and then the sample was heated to 44 °C, cooled to 25 °C, heated to 44 °C and cooled to 25 °C, with measurements performed after each change in temperature. Results are presented as size distribution by volume.

For measuring the influence of substitutions on oligomer formation, the size of oligomers formed by either 10 µM AncA0 or 10 µM AncA0 Q66H G109D in buffer F (as above) was measured by DLS at 25, 27, 29, 31, 33, 35, 37, 39, 41, 43, and 45°C. Results were presented as size distribution by intensity (for temperatures 25 °C, 35°C, and 45°C) or as an average hydrodynamic diameter corresponding to the maximum of a dominant peak of size distribution by volume plotted against temperature ± standard deviation.

For assembly formation (sequestrase) assay, 1.5 μM firefly luciferase in buffer F was denatured alone or in the presence of different sHsp concentrations by incubation at 44 °C for 10 min. Size of obtained luciferase aggregates was then measured by DLS at 25 °C. Results are presented as an average hydrodynamic diameter of measured particles weighted by intensity (Z-average) ± standard deviation.

Biolayer interferometry (BLI) measurements

Request a detailed protocol

sHsps interactions with aggregated luciferase or aggregated E. coli lysate were measured using the Octet K2 system. Anchoring layer of his-tagged luciferase was attached to Octet NTA Biosensors by 5 min incubation in 0.6 mg/ml his-tagged luciferase in denaturing conditions in buffer UF (50 mM Tris pH 7.5, 4.5 M urea, 150 mM KCl, 20 mM MgCl2, 2.5 mM DTT) at 25 °C with 350 rpm shaking. Sensors were then incubated for 5 min as above in buffer H (50 mM Tris pH 7.5, 150 mM KCl, 20 mM MgCl2, 5 mM β-mercaptoethanol) to remove urea and unbound luciferase. The protein aggregate was then formed on the sensor by incubation for 10 min in 0,5 mg/ml His-tagged luciferase or 0.2 mg/ml E. coli lysate in buffer H at 44 °C (in case of the luciferase) or 55 °C (in case of the lysate). Sensors were then again incubated in buffer H for 5 min at 25 °C with 350 rpm shaking to remove excess protein. Sensors with attached aggregate were placed in the Octet system in H buffer for 60 s baseline measurement and then placed for 1 hr in 5 μM sHsp solution in H buffer to measure sHsps association. Sensors were then moved for 1 hr to buffer H to measure protein dissociation. Measurements were performed with 1000 rpm shaking at 44 °C (in case of full-length sHsps) or at 25 °C (in case of ACDs). Full-length proteins were preincubated at 44 °C for 10 min before measurement.

sHsps displacement by Hsp70 system was measured using ForteBio BLItz. Sensors with attached aggregates were prepared as described above, with buffer F (50 mM Tris pH 7.5, 150 mM KCl, 20 mM MgCl2, 2.5 mM DTT) instead of buffer H. Baseline biolayer was measured for 60 s. Sensors were then placed in 5 µM sHsp solution in buffer F, previously preincubated for 10 min at 44 °C. sHsps association was measured for 10 min. Sensors were then moved to either buffer F or Hsp70 system in buffer F (0.7 µM DnaK, 0.28 µM DnaJ, 0.21 µM GrpE, 5 mM ATP, 0.1 mg/ml creatine kinase, 18 mM creatine phosphate). Hsp70 system binding and sHsps dissociation were measured for 1 hr. Measurements were performed at room temperature with 2000 rpm shaking.

ACD interactions with C-terminal peptide were measured using Octet K2 system. Octet NTA Biosensors were placed in buffer C (50 mM Tris pH 7.5, 150 mM KCl, 5% glycerol, 5 mM β-mercaptoethanol) and baseline signal was measured for 60 s. Sensors were then placed in 2.5 µM His6-SUMO-C-peptide solution in buffer C and incubated for 15 min. Surplus His6-SUMO-C-peptide was then removed by incubation in G buffer for 15 min. Sensors were then moved to ACD solution and association was measured for 48 min. After that, ACD dissociation was measured in buffer G for 10 min. Measurements were performed with 1000 rpm shaking at 25 °C. Measurements were performed for different ACD concentrations in triplicates. Biolayer thickness at the end of association stage was corrected for the nonspecific binding using control substituting His6-SUMO for His6-SUMO-C-peptide and converted to fraction bound by min-max scaling between 0 and 1 using minimal and maximal triplicate averages of biolayer thickness. To determine dissociation constant fraction bound as a function of ACD concentration was fitted to the Hill equation using SciPy implementation (Virtanen et al., 2020) of dogbox algorithm:

fb = XnK0.5n+Xn

Where [X] is the total ACD concentration, K0.5-ACD concentration required to reach half-maximum binding at equilibrium, n-Hill coefficient, fb-fraction bound. Standard deviations of fitted K0.5 and n values were derived from the diagonal of the optimized parameters covariance matrix.

Molecular dynamics (MD)

Request a detailed protocol

All MD simulations were performed using Gromacs 2019.2 (Van Der Spoel et al., 2005) and CHARMM36-jul2021 as a force field (Huang et al., 2017). Simulations were performed in the isothermal-isobaric (NPT) ensemble, where the temperature was kept at 310 K using v-rescale thermostat (Bussi et al., 2007) with a time constant of 0.1 ps and the pressure was held at 1 bar using Parinello-Rahman barostat (Parrinello and Rahman, 1981). Lennard-Jones potential with a cut-off of 1.0 nm was used to describe Van der Waals interactions. Computation of long-range electrostatic interactions was performed using the particle mesh Ewald (PME) method (Essmann et al., 1995) with a Fourier grid spacing of 0.12 nm and a real space cutoff of 1.0 nm. Bonds between hydrogen and protein heavy atoms were constrained by P-LINCS (Hess et al., 1997), and water molecules geometry was constrained by SETTLE (Miyamoto and Kollman, 1992). Integration of equations of motion was performed by leap-frog algorithm (Van Gunsteren and Berendsen, 1988) with a time step of 2 fs. Periodic boundary conditions were applied in all dimensions.

The initial conformation of AncA0 ACD-C-terminal peptide complex was predicted by ColabFold implementation (Mirdita et al., 2022) of AlphaFold-Multimer (Evans et al., 2022) Substitutions Q66H and G109D were introduced to the complex using PyMol Mutagenesis Wizard (http://www.pymol.org). Complexes were then placed in rhombic dodecahedral boxes measuring 10.4 nm in all dimensions and solvated by CHARMM-modified TIP3P water model (Jorgensen et al., 1983). Concentrations of sodium and chloride ions were adjusted to 0.15 M and net zero charge of the system. Each system was subjected to a 3-step energy minimization protocol, where during the first step protein conformation was constrained, during the second step constraint was reduced to protein backbone only and during the third step positions of all protein heavy atoms were restrained with a force constant of 1000 kJ*mol–1*nm–1. Minimized systems were equilibrated for 10 ns while positions of protein backbone atoms were kept constrained, equilibration was then continued without constraints for further 500 ns. The first 100 ns of equilibration was discarded, and the rest was used for ACD-C-terminal peptide contact determination using GetContacts tool (https://getcontacts.github.io/; Fonseca and Ma, 2018). Contact between interfacial residues was defined as any of the following types of interaction: hydrogen bond, ionic, π-stacking, π-cation, or van der Waals between purely hydrophobic residues. The default GetContacts interaction criteria were used for all interaction types except for hydrogen bond detection, where a more stringent 30° cutoff for hydrogen-donor-acceptor angle was used. Contact heatmaps were prepared using seaborn (Waskom, 2021).

The representative conformations of ACD-C-terminal peptide complexes were chosen by clustering the last 400 ns of equilibrium MD trajectories (frames spaced every 0.5 ns) using ‘gmx cluster’ tool and Jarvis-Patrick clustering method (Jarvis and Patrick, 1973). RMSD cutoff of 0.2 nm, was used for the Jarvis-Patrick algorithm and conformations possessing at least three neighbors in common out of 15 closest conformations were assigned to the same cluster. RMSD calculation was based on coordinates of heavy backbone atoms of the C-terminal peptide interacting ACD monomer without dimerization loop (residues 40–74 and 95–126) and stably interacting region of the C-terminal peptide (residues 132–137). In the case of both simulated complexes the biggest cluster contained more than 90% of simulation frames (92.3% for AncA0 ACD complex and 93.3% for AncA0 Q66H G109D complex) and its middle frame was chosen as representative conformation. Visual Molecular Dynamics (VMD) (Humphrey et al., 1996) and Blender (https://www.blender.org/) were used for structure visualization.Supplementary file 1Supplementary file 3.

Data availability

All data generated or analysed during this study are included in the manuscript and supporting files. Source data files have been provided for Figs 1, 3, 4, 5, 6 and Figure supplements. Multiple sequence alignment of Enterobacterales IbpA orthologs used for phylogenetic analysis and ancestral reconstruction in fasta format is attached as Supplementary file 1. Posterior probability statistics for ancestral sequence reconstruction of AncA0 and AncA1 nodes are attached as Supplementary file 3. Raw data generated in this study are attached as source data files.

References

    1. Cohen O
    2. Rubinstein ND
    3. Stern A
    4. Gophna U
    5. Pupko T
    (2008) A likelihood framework to analyse phyletic patterns
    Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 363:3903–3911.
    https://doi.org/10.1098/rstb.2008.0177

Peer review

Reviewer #1 (Public Review):

The work in this paper is in general done carefully. Reconstructions are done appropriately and the effects of statistical uncertainty are quantified properly. I was glad to see that the tree and alignment are now deposited.

The paper identifies which mutations are crucial for the functional differences between the ancestors tested. This is done quite carefully - the authors even show that the same substitutions also work in extant proteins.

These substitutions very slightly lower the affinity and increase the cooperativity of the C-terminal peptide binding to the alpha crystallin domain - a key oligomeric interaction. These relatively minor changes nevertheless apparently affect the subunit exchange behaviour and oligomerization of the sHSP.

Lastly, the authors use likelihood methods to test for signatures of selection. This reviewer is not a fan of these methods, as they are easily misled by common biological processes (see PMID 37395787 for a recent critique). The paper is relatively careful in the interpretation of this test though, and I think the importance of the other findings does not depend on the action of selection along this branch.

https://doi.org/10.7554/eLife.89813.3.sa1

Reviewer #2 (Public Review):

This was an interesting study, and I enjoyed seeing different experimental approaches used to compare the properties of the different native proteins, the ancestral reconstructions, and the other mutants. In the original manuscript, I felt that the authors had over-simplified their explanations, as the differences between the ancestral proteins, and the changes induced by the two mutations, only partially explain the differences between IbpA proteins from the two different species. However, with their revised version, I think the presentation and discussion of their results are much better. Overall, I think this represents a valuable contribution to the field, providing convincing mechanistic evidence as to how these small heat shock proteins have evolved.

https://doi.org/10.7554/eLife.89813.3.sa2

Author response

The following is the authors’ response to the original reviews.

Reviewer #1 (Recommendations For The Authors):

My main request is to show the phylogeny in the main text, so the reader knows what nodes are being compared.

Full phylogeny was added to the main text as Fig. 2. Additionally, phylogenetic tree in Newick format is presented as a Supplementary file 2.

I also suggest the authors check their figure legends carefully. At least in figure one, I think there is some mix-up with the letter labelling of the panels.

Our mistake. Figure legend was corrected. In this version of the manuscript Figure 1 was split into Fig. 1 and Fig. 3. Corrected version is presented in the legend to Fig. 3.

And lastly, I urge the authors to deposit the tree, alignment, and reconstructed sequences in a public repository.

Alignment in fasta format and phylogenetic tree in Newick format were added as supplementary files to the publication (supplementary file 1 and supplementary file 2, respectively). Reconstructed sequences(both Most likely and AltAll variants) were shown as a figure supplement (Figure 3 – figure supplement2). Posterior probabilities for all positions of the reconstructed sequences were added as a supplementary file (supplementary file 3).

Reviewer #2 (Recommendations For The Authors):

-I find the term "secondarily single sHsp" to be a little confusing, especially because it is often used in relation to IbpA/B, but it is just IbpA in another species. I think it would be more clear for the reader to consistently refer to it as Erwiniaceae IbpA vs Escherichia IbpA, or something similar.

In the introduction we clarified (page 4 lines 11-13) that the term “secondarily single” IbpA refers to IbpA that lacks partner protein as a result of ibpB gene loss. This is in opposition to “single-protein” IbpA from a clade in which gene duplication leading to creation of two – protein sHsp system did not occur (like Vibrionaceae or Aeromonadaceae) - see Obuchowski et al., 2019.

-Figure 1B. The labels are not defined. What is L? A and B refer to IbpA and IbpB but this should be made more clear to the reader. Why is this panel only referred to in the Introduction and not the Results? Why is there a second panel for E.amy, rather than including it in the same panel, as for other experiments? What are the error bars? (That goes for every error bar in the paper, none are defined).

Labels in Fig.1B were corrected; “L” was used in reference to “luciferase alone” and it has been corrected for consistency to “no sHsp”. The sHsps activity measurements (obtained in the same experiment) were split into two separate panels as a correspondence to the two branches of the simplified tree in Fig. 1. The figure was modified to make it clearer and avoid confusion. Definitions of error bars were added to this and other figures.

-"AncA0 exhibited sequestrase activity on the level comparable to IbpA from Escherichia coli(IbpA E. coli). AncA1 was moderately efficient in this process and IbpA from Erwinia amylovora(IbpAE.amyl) was the least efficient sequestrase (Fig. 1D)." - First, this should be referring to Fig. 1C. Second, the text doesn't quite match the panel. A0 appears to have the strongest sequestrase activity over most concentrations. Can the authors comment on in what concentration range these differences are most meaningful?

Figure legend was corrected. Descriptions of panels C and D were fixed. Now these data are presented in panels A and B of a new Fig. 3. In our opinion differences in sequestration are most meaningful at lower sHsp concentrations (in this case lower than 5 µM), as with high enough sHsp concentration even less effective sequestrases seem to be able to effectively sequester aggregated proteins. Comment about it was added to the main text (page 5, line 6)

-"Ancestral proteins' interaction with the aggregated substrates was stronger than in the case of extant E. amylovora IbpA, but weaker than in the case of extant E. coli IbpA (Fig. 1C)." - Is this referring to Fig.1C, or to the unlabelled panel on the bottom right panel of Fig 1 (that is not referred to in the legend)? Can the authors comment on why they think the 2 ancestral proteins are much more similar to each other than they are to either of the native IbpAs?

Due to our mistake descriptions of panels C and D were switched.

Figure 1 was rearranged and split into Figures 1 and 3. Former figure S1 (full phylogeny) was inserted into the main text, as Fig. 2, per request of reviewer #1. Former panel 1D (now 3B) was rearranged, as graph was not apparent to be a part of that panel and looked as if it was unlabeled.

The fact that the two ancestral proteins are more similar to each other than to the extant E. coli and E. amylovora proteins in their interaction with model substrate might be caused by higher sequence identity between the two ancestral proteins than between ancestral and extant proteins (10 amino acid differences between AncA0 and AncA1 compared to 20 differences between AncA1 and IbpA from E. amylovora or 11 differences between AncA0 and IbpA from E. coli). One also has to remember that this property is only one aspect of sHsp activity – proteins AncA0 and AncA1 are much less similar to each other if other activities such as sequestrase activity are considered. Substrate affinity and sequestrase activity are connected to each other, but there isn’t a strict correlation, as can be seen in the case of free ACD domains, which strongly bind aggregated substrate while effectively lacking sequestrase activity (fig. 5 A, fig. 5 – figure supplement 4 A,B).

-Figure 1E should have E. coli IbpA and IbpB, by themselves, included for comparison. Strangely, it seems, by comparison to Fig 1B, that the "inhibitory" activity of A0 is not present in the E. coli protein, and the authors should comment on this. Similarly, A1 disaggregation looks like it might not be significantly different than the E. coli protein. Can the authors comment on why disaggregation might be so low in A1 compared to E.amy?

E. coli IbpA alone was added to Fig. 1E (Fig. 3C in the new version) as suggested.

AncA1 indeed exhibits similar activity to extant IbpA from E. coli, which, at the conditions of the experiment, does not possess inhibitory effect observed for AncA0. This suggests that:

-There was an additional increase in ability to stimulate luciferase disaggregation between AncA1 and extant IbpA from E. amylovora

-There was also an increase of ability to stimulate luciferase refolding between AncA0 and extant E. coli IbpA, albeit to a significantly lesser degree than in the Erwiniaceae branch.

It is quite likely that after separation of Erwiniaceae and Enterobacteriaceae sHsp systems, they underwent further optimization through evolution. This might have led to observed higher effectiveness of modern IbpAs from both clades in refolding stimulation in comparison to the reconstructed ancestral proteins.

Despite the above, effects of substitutions on positions 66 and 109 on activities of the extant E. coli and E. amylovora proteins suggests that the two identified positions still play key role in differentiating extant IbpAs from Erwiniaceae and Enterobacteriaceae.

Nevertheless, additional mutations that lead to increased ability to stimulate luciferase reactivation must have occurred in both Erwiniaceae and Enterobacteriaceae branches of the phylogeny during evolution. These substitutions would be a worthwhile subject of further study.

-Fig 1D - lizate should be lysate.

The typo was corrected.

-What is the bottom right panel in Fig 1? It doesn't seem to be referred to in the legend.

This panel was intendent to be the part of figure 1D, but it was not clearly visible. This figure was rearranged to make it clearer. Now these data are presented as Fig. 3B.

-Sequences are provided for the ancestral proteins, but I don't see them anywhere for the alternative ancestral proteins. How similar are the Anc proteins to the AltAlls? If they are very similar, this may not tell us anything about "robustness".

Sequences of alternative proteins are added as a figure supplement (Fig. 3 - figure supplement 2). Full sequences of ML and alternative ancestors with posterior probabilities for each reconstructed position are presented in supplementary file 3

The testing of the robustness to statistical uncertainty was intended to test to what extent properties of reconstructed ancestral proteins could be influenced by uncertainty present in a given reconstruction due to probabilistic nature of the process. Relatively high similarity between ML and AltAll sequences would indicate low uncertainty of the reconstruction (most likely due to high conservation during evolution). In such a case similar properties of AltAll and ML proteins would simply indicate that they are robust to the level of uncertainty present in a given reconstruction (which may be low). It would not tell us much about “general” robustness to mutations, but it was not relevant to research questions considered.

-If the functional gain by IbpA comes down to only two amino acid substitutions, I'm not convinced this would be meaningfully reflected in any tests of positive selection.

After considering Reviewer #1’s comments about limitations of models used for selection analysis we added acknowledgment in the discussion (page 9, line 9 - 13) that results indicating positive selection in our dataset should not be considered conclusive (see answer to Reviewer #1’s public review below).

-The full MSA should be provided as supplemental material.

The full MSA in fasta format is presented in the supplementary file 1.

-For the aggregate binding panels in Figs 3 and 4, it would be helpful to show the native and ancestral proteins for comparison. I know this is a bit redundant, as they're present in Fig 1, but I find it hard to judge the scale of change. This is especially important because A0 and A1 are very similar in Fig 1, so I want to see what kind of difference the 2 mutations make.

Data presented in Fig. 3C (Fig. 5C in the new version) refer to the binding of α-crystallin domains (A0ACD and A0ACD Q66H G109D) and not full length sHsps to E. coli proteins aggregated on a BLI sensor. Our intention was to show the influence of the two crucial substitutions (Q66H G109D) on the properties of A0 ancestral α-crystallin domain.

Figure 4 (Fig. 6 in the new version) represent the effects of the substitutions on the identified positions 66 and 109 on the properties of extant IbpA orthologs from E. coli and E. amylovora, showing that these two positions play a key role in differentiating properties of those extant proteins. Changes in binding to aggregated substrate caused by those substitutions, as shown in Figure 6 B,C (new version), are indeed larger than observed between AncA0 and AncA1, as shown in Fig. 3B (new version).

One has to remember, however, that the experiment shown in Fig.3 (new version) shows the effects of all 10 amino acid changes between the nodes A0 and A1 and not only the two analyzed substitutions, as was the case in experiment shown in Fig. 6 B,C (new version). Moreover, due to relatively large number of differences between ancestral and extant sequences (11 differences between AncA0 and E. coli IbpA, 20 differences between AncA1 and E. amylovora IbpA), substitutions in the two experiments are introduced into different sequence context.

Because of the above, we believe that direct comparison of the results obtained for ancestral proteins with the results obtained for substitutions introduced into extant proteins would not meaningfully contribute to answering the question of the role of analyzed substitution in the context of extant proteins, while decreasing clarity of presented information.

-Some of the luciferase plots show a time course, but others just show a single %. What is the time point used for the single % plots?

Information was added to appropriate figure legends that for experiments showing a single timepoint the luciferase activity was measured after 1h of refolding.

Reviewer #3 (Recommendations For The Authors):

1. In the Introduction, it would be beneficial to explore additional instances where this evolutionary simplification process has been observed in nature. Investigating the prevalence of this phenomenon and identifying other multi-protein systems that have undergone simplification could enhance the understanding of its significance and implications.

The section of the introduction concerning gene loss and differential paralog retention was expanded with additional examples of gene loss that is considered adaptive (page 3 lines 1 - 12).

1. I am intrigued by the reasons why certain organisms continue to maintain a two-protein system despite the viability of a single-protein system. This aspect is particularly relevant for bacteria, considering the fitness cost associated with maintaining extra gene copies. Do you have any hypotheses or theories that may shed light on this intriguing observation?

Refolding of proteins from aggregates requires the functional cooperation of sHsps and chaperones from Hsp70 system and Hsp100 disaggregase. In two protein sHsps system one sHsp (IbpA) is specialized in substrate binding, while the second one (IbpB) possesses low substrate binding potential and enhances sHps dissociation from substrates (Obuchowski et al, 2019). Thus, the presence of IbpB reduces the amount of chaperones from Hsp70 system required to outcompete sHsps from aggregated substrates to initiate refolding process. The cost associated with maintaining extra sHsp gene copy (ibpB) in bacteria might be compensated by lower requirement for Hsp70 chaperones for efficient and fast protein refolding following stress conditions.

In this study we have demonstrated how such a system could have been simplified to a single – protein system capable of efficient substrate sequestration as well as stimulation of reactivation. This indeed leads to the question why such single – protein system isn’t more prevalent in Enterobacterales.

One possibility may be that there are very specific requirements for efficient reactivation by a single – protein sHsp system. We have shown that new, more efficient IbpA functionality observed in Erwiniaceae required at least two separate mutations. It is possible, that such combinations of two substitutions simply did not occur in Enterobacteriaceae clade, in which IbpA still required partner protein for efficient reactivation stimulation.

One must also remember that experiments performed in this study were performed in vitro in a specific set of conditions, which most likely does not represent whole spectrum of challenges faced by different bacteria. It is possible that two – protein system has some other additional adaptive effects, counterbalancing the additional cost of gene maintenance. It was for example recently shown (Miwa & Taguchi, PNAS, 120 (32) e2304841120) that bacterial sHsps play an important role in regulation of stress response. Two – protein system could potentially allow for more complex regulation.

1. Incorporating X-ray crystallization as an additional technique in the methodology would offer detailed molecular insights into the effects of Q66H and G109D substitutions on ACD-C-terminal peptide and ACD-substrate interactions. The inclusion of such data would further strengthen the results section and provide robust support for your findings. Since the x-ray data might be difficult to collect, the authors might think to get alphafold model or some rosetta score for the model to discuss the finding further.

In response to reviewer comment we added the comparison of the structural models of AncA0 and AncA0 Q66H G109D ACD dimers complexed with the C-terminal peptides, representing middle structures of largest clusters obtained from equilibrium molecular dynamics simulation trajectories based on the AlphaFold2 prediction and in silico mutagenesis (Fig. 5 – figure supplement 2). Model comparison as well as C-terminal peptide – ACD contact analysis did not reveal any major changes in mode of peptide binding or α-crystallin domain conformation, although we do acknowledge that simulation timescale limits the conformational sampling.

Reviewer #1 (Public Review):

The work in this paper is in general done carefully. Reconstructions are done appropriately and the effects of statistical uncertainty are quantified properly. My only slight complaint is that I couldn't find statistics about posterior probabilities anywhere and that the sequences and trees do not seem to be deposited.

Posterior probabilities for all positions of reconstructed proteins were added as a supplementary file 3. MSA of all sequences used for ancestral reconstruction as well as phylogenetic tree in Newick format were added as supplementary files 1 and 2, respectively.

I would also have preferred to have the actual phylogeny in the main text. This is a crucial piece of data that the reader needs to see to understand what exactly is being reconstructed.

Full phylogeny was added to the main text as Fig. 2.

The paper identifies which mutations are crucial for the functional differences between the ancestors tested. This is done quite carefully - the authors even show that the same substitutions also work in extant proteins. My only slight concern was the authors' explanation of what these substitutions do. They show that these substitutions lower the affinity of the C-terminal peptide to the alpha-crystallin domain - a key oligomeric interaction. But the difference is very small - from 4.5 to 7 uM. That seems so small that I find it a bit implausible that this effect alone explains the differences in hydrodynamic radius shown in Figure S8. From my visual inspection, it seems that there is also a noticeable change in the cooperativity of the binding interaction. The binding model the authors use is a fairly simple logarithmic curve that doesn't appear to consider the number of binding sites or potential cooperativity. I think this would have been nice to see here.

The binding model we used is equivalent to the Hill equation as it accounts for the variable slope of sigmoid function by inclusion of input scaling factor k, which is equivalent to the hill coefficient. Simple one site binding model and two site binding model were also considered but provided worse fits to the data than model including binding cooperativity. Not providing values of fitted parameter k was our mistake, and it was corrected (Fig. 5. with a legend). Additionally, output scaling parameter L is not necessary as fraction bound takes values from 0 to 1, therefore we have fitted the curves again without this parameter. The new values of fitted parameters are very similar to the previous ones. To make text more accessible to the reader, we have used a conventional form of Hill equation. Indeed, AncA0 Q66H G109D ACD displays higher binding cooperativity than more ancestral AncA0 ACD (hill coefficient 2.3 for AncA0 vs 3.7 for AncA0 Q66H G109D). Fitted values of Hill coefficients are higher than one can expect for 2-site ACD dimer, which is probably caused by an experimental setup of BLI, where C-terminal peptide is immobilized on the sensor and ACD is present in solution as bivalent analyte leading to emergence of avidity effects. Both cooperativity and avidity are reflected in the value of Hill coefficient, however as ligand density on the sensor is the same in all experiments only change in ACD binding cooperativity can account for observed difference in the value of Hill coefficients. Difference in the C-terminal peptide binding cooperativity may influence the process of sHsp oligomerization and assembly formation despite similar binding affinity, especially if avidity of multiple binding sites within oligomer is considered.

In addition, we changed the legend to Figure S8 (now called Fig. 5 – figure supplement 4A ) to clarify the fact that the differences in average hydrodynamic radius are in fact ferly small. To highlight the observation that there are two populations of particles in AncA0 and AncA0 Q66H G109D measured at 25, 35 and 45 °C with different hydrodynamic diameters, we used % of intensity in DLS measurement. It allows us to show the change in the hydrodynamic diameter distribution that is relatively small. We recognize it was not properly explained in the article and added a clarification in figure description.

Lastly, the authors use likelihood methods to test for signatures of selection. This reviewer is not a fan of these methods, as they are easily misled by common biological processes (see PMID 37395787 for a recent critique). Perhaps these pitfalls could simply be acknowledged, as I don't think the selection analysis is very important to the impact of the work.

We thank the reviewer for pointing to the recent research about limitations of methods used in our work in selection analysis. As per recommendation we added acknowledgment of limitations of methods used to discussion (page 9, line 9 - 13), modifying wording of our conclusions to deemphasize significance of selection analysis results.

https://doi.org/10.7554/eLife.89813.3.sa3

Article and author information

Author details

  1. Piotr Karaś

    Intercollegiate Faculty of Biotechnology UG-MUG, University of Gdansk, Gdańsk, Poland
    Contribution
    Conceptualization, Data curation, Formal analysis, Investigation, Visualization, Methodology, Writing - original draft, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-5270-7938

  2. Klaudia Kochanowicz

    Intercollegiate Faculty of Biotechnology UG-MUG, University of Gdansk, Gdańsk, Poland
    Contribution
    Investigation, Methodology
    Competing interests
    No competing interests declared

  3. Marcin Pitek

    Intercollegiate Faculty of Biotechnology UG-MUG, University of Gdansk, Gdańsk, Poland
    Contribution
    Investigation, Visualization, Methodology, Writing - original draft, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-1300-4364

  4. Przemyslaw Domanski

    Intercollegiate Faculty of Biotechnology UG-MUG, University of Gdansk, Gdańsk, Poland
    Contribution
    Investigation, Visualization
    Competing interests
    No competing interests declared

  5. Igor Obuchowski

    Intercollegiate Faculty of Biotechnology UG-MUG, University of Gdansk, Gdańsk, Poland
    Contribution
    Methodology, Writing - review and editing
    Competing interests
    No competing interests declared

  6. Barlomiej Tomiczek

    Intercollegiate Faculty of Biotechnology UG-MUG, University of Gdansk, Gdańsk, Poland
    Contribution
    Conceptualization, Formal analysis, Supervision, Validation, Visualization, Writing - original draft, Project administration, Writing - review and editing
    For correspondence
    bartlomiej.tomiczek@biotech.ug.edu.pl
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-9295-663X

  7. Krzysztof Liberek

    Intercollegiate Faculty of Biotechnology UG-MUG, University of Gdansk, Gdańsk, Poland
    Contribution
    Conceptualization, Supervision, Funding acquisition, Validation, Writing - original draft, Project administration, Writing - review and editing
    For correspondence
    krzysztof.liberek@ug.edu.pl
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-7532-9279

Funding

Narodowe Centrum Nauki (OPUS 17 2019/33/B/NZ1/00352)

  • Piotr Karaś
  • Klaudia Kochanowicz
  • Marcin Pitek
  • Igor Obuchowski
  • Krzysztof Liberek

Narodowe Centrum Nauki (OPUS 21 2021/41/B/NZ8/02835)

  • Przemyslaw Domanski
  • Barlomiej Tomiczek

Infrastruktura PL-Grid (PLG/2023/016546)

  • Barlomiej Tomiczek

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

This work was supported by a grant from the Polish National Science Centre (OPUS 17 2019/33/B/NZ1/00352). The work of B.T. and P.D. was supported by a National Science Center grant (OPUS 21 2021/41/B/NZ8/02835)(to B.T.). We gratefully acknowledge Poland’s high-performance computing infrastructure PLGrid (HPC Centers: ACK Cyfronet AGH) for providing computer facilities and support within computational grant no. PLG/2023/016546. We thank Prof. Max Telford, Prof. Jaroslaw Marszalek, and Dr. Agnieszka Kłosowska for their helpful discussions.

Senior and Reviewing Editor

  1. Christian R Landry, Université Laval, Canada

Version history

  1. Preprint posted: May 30, 2023 (view preprint)
  2. Sent for peer review: June 22, 2023
  3. Preprint posted: August 21, 2023 (view preprint)
  4. Preprint posted: November 23, 2023 (view preprint)
  5. Version of Record published: December 8, 2023 (version 1)

Cite all versions

You can cite all versions using the DOI https://doi.org/10.7554/eLife.89813. This DOI represents all versions, and will always resolve to the latest one.

Copyright

© 2023, Karaś et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Piotr Karaś
  2. Klaudia Kochanowicz
  3. Marcin Pitek
  4. Przemyslaw Domanski
  5. Igor Obuchowski
  6. Barlomiej Tomiczek
  7. Krzysztof Liberek
(2023)
Evolution towards simplicity in bacterial small heat shock protein system
eLife 12:RP89813.
https://doi.org/10.7554/eLife.89813.3

Categories and tags

Research organism

Further reading

    1. Biochemistry and Chemical Biology
    2. Computational and Systems Biology

    Death by a thousand cuts through kinase inhibitor combinations that maximize selectivity and enable rational multitargeting

    Ian R Outhwaite, Sukrit Singh ... Markus A Seeliger
    Research Article

    Kinase inhibitors are successful therapeutics in the treatment of cancers and autoimmune diseases and are useful tools in biomedical research. However, the high sequence and structural conservation of the catalytic kinase domain complicates the development of selective kinase inhibitors. Inhibition of off-target kinases makes it difficult to study the mechanism of inhibitors in biological systems. Current efforts focus on the development of inhibitors with improved selectivity. Here, we present an alternative solution to this problem by combining inhibitors with divergent off-target effects. We develop a multicompound-multitarget scoring (MMS) method that combines inhibitors to maximize target inhibition and to minimize off-target inhibition. Additionally, this framework enables optimization of inhibitor combinations for multiple on-targets. Using MMS with published kinase inhibitor datasets we determine potent inhibitor combinations for target kinases with better selectivity than the most selective single inhibitor and validate the predicted effect and selectivity of inhibitor combinations using in vitro and in cellulo techniques. MMS greatly enhances selectivity in rational multitargeting applications. The MMS framework is generalizable to other non-kinase biological targets where compound selectivity is a challenge and diverse compound libraries are available.

    1. Biochemistry and Chemical Biology
    2. Microbiology and Infectious Disease

    Indole produced during dysbiosis mediates host–microorganism chemical communication

    Rui-Qiu Yang, Yong-Hong Chen ... Cheng-Gang Zou
    Research Article Updated

    An imbalance of the gut microbiota, termed dysbiosis, has a substantial impact on host physiology. However, the mechanism by which host deals with gut dysbiosis to maintain fitness remains largely unknown. In Caenorhabditis elegans, Escherichia coli, which is its bacterial diet, proliferates in its intestinal lumen during aging. Here, we demonstrate that progressive intestinal proliferation of E. coli activates the transcription factor DAF-16, which is required for maintenance of longevity and organismal fitness in worms with age. DAF-16 up-regulates two lysozymes lys-7 and lys-8, thus limiting the bacterial accumulation in the gut of worms during aging. During dysbiosis, the levels of indole produced by E. coli are increased in worms. Indole is involved in the activation of DAF-16 by TRPA-1 in neurons of worms. Our finding demonstrates that indole functions as a microbial signal of gut dysbiosis to promote fitness of the host.

    1. Biochemistry and Chemical Biology
    2. Structural Biology and Molecular Biophysics

    The neuronal calcium sensor NCS-1 regulates the phosphorylation state and activity of the Ga chaperone and GEF Ric-8A

    Daniel Muñoz-Reyes, Levi J McClelland ... Maria Jose Sanchez-Barrena
    Research Article

    The Neuronal Calcium Sensor 1, an EF-hand Ca2+ binding protein, and Ric-8A coregulate synapse number and probability of neurotransmitter release. Recently, the structures of Ric-8A bound to Ga have revealed how Ric-8A phosphorylation promotes Ga recognition and activity as a chaperone and guanine nucleotide exchange factor. However, the molecular mechanism by which NCS-1 regulates Ric-8A activity and its interaction with Ga subunits is not well understood. Given the interest in the NCS-1/Ric-8A complex as a therapeutic target in nervous system disorders, it is necessary to shed light on this molecular mechanism of action at atomic level. We have reconstituted NCS-1/Ric-8A complexes to conduct a multimodal approach and determine the sequence of Ca2+ signals and phosphorylation events that promote the interaction of Ric-8A with Ga. Our data show that the binding of NCS-1 and Ga to Ric-8A are mutually exclusive. Importantly, NCS-1 induces a structural rearrangement in Ric-8A that traps the protein in a conformational state that is inaccessible to Casein Kinase II-mediated phosphorylation, demonstrating one aspect of its negative regulation of Ric-8A-mediated G-protein signaling. Functional experiments indicate a loss of Ric-8A GEF activity towards Ga when complexed with NCS-1, and restoration of nucleotide exchange activity upon increasing Ca2+ concentration. Finally, the high-resolution crystallographic data reported here define the NCS-1/Ric-8A interface and will allow the development of therapeutic synapse function regulators with improved activity and selectivity.