Is the lipochaperone activity of sHSP a key to the stress response encoded in its primary sequence?

Abstract

Several strategies have been put in place by organisms to adapt to their environment. One of these strategies is the production of stress proteins such as sHSPs, which have been widely described over the last 30 years for their role as molecular chaperones. Some sHSPs have, in addition, the particularity to exert a lipochaperone role by interacting with membrane lipids to maintain an optimal membrane fluidity. However, the mechanisms involved in this sHSP-lipid interaction remain poorly understood and described rather sporadically in the literature. This review gathers the information concerning the structure and function of these proteins available in the literature in order to highlight the mechanism involved in this interaction. In addition, analysis of primary sequence data of sHSPs available in database shows that sHSPs can interact with lipids via certain amino acid residues present on some β sheets of these proteins. These residues could have a key role in the structure and/or oligomerization dynamics of sHPSs, which is certainly essential for interaction with membrane lipids and consequently for maintaining optimal cell membrane fluidity.

Introduction

A weak modification of the internal cell environment in response to external stress conditions (e.g., heat, nutritional, oxidative, alcohol stresses) is enough to lead to protein misfolding and a subsequent loss of biological function (Booth 1985; Berger et al. 1996; Lim et al. 2000). To maintain protein homeostasis, several stress mechanisms are involved in cells, leading to the production of a group of stress proteins called heat shock proteins (HSP). These proteins act as molecular chaperones which bind and sequester denatured proteins (Ellis et al. 1989; Ellis et al. 1990; Hartl 1996; Beissinger and Buchner 1998; Tyedmers et al. 2010). Some of them are able to refold the denatured protein in an ATP-dependent manner. HSP are universal stress proteins found in almost all organisms including eukaryotes (plants, metazoan, yeast), bacteria, archaebacteria, and viruses (Wirth 2003; Ponomarenko et al. 2013; Carra et al. 2017). Besides being present in all kingdoms (Narberhaus 2002; Horwitz 2003; Sun and MacRae 2005; Haslbeck et al. 2005), these proteins have a high degree of homology to each other and are found ubiquitously in cells. They can be grouped into six main families according to their molecular mass, HSP 110/100, 90, 70, 60, 47, and less than 35 kDa (Wirth 2003; Bukau and Horwich 1998; Chang et al. 2007; Tang et al. 2007; Marcion et al. 2014). However, each family has high specificity concerning notably its mechanisms of action, intracellular localization, substrate specificity, and ATP requirement.

The stress proteins under 35 kDa, called small heat shock protein (sHSP), act mainly as ATP-independent chaperones (Baneyx and Mujacic 2004). For some of them, a lipochaperone activity, i.e., cell membrane protection activity, is described from the outset of stress conditions. However, this lipochaperone activity is still poorly described and understood (Mogk et al. 2019). As this activity is found in very different model organisms that diverged quite early in the evolution of sHSP, it is quite possible that this role is more common than currently described. Although highly heterogeneous in their primary sequence, the secondary structures and activities are well-conserved among sHSPs (Jong et al. 1998; Poulain et al. 2010). Thus, the question is to predict the potential molecular lipochaperone role of a sHSP on the basis of its primary sequence. Indeed, bioinformatics work on the primary sequences of 50 sHSPs has shown that certain protein motifs are better conserved in sHSPs known for their role as molecular lipochaperones.

sHSP, a key stress fighter

sHSP through the ages

Since the discovery of the first sHSP by Craig and Ingola in 1982 (Ingolia and Craig 1982), the number of identified sHSPs has exploded. These proteins are present in every kingdom of life in which they act primarily as molecular chaperons to control and protect the quality of the protein network (Obuchowski and Liberek 2020). The number of sHSP per organism may vary; however, higher organisms (plants, human, animals …) generally have a large number of sHSP (Haslbeck et al. 2005). For example, there are nineteen sHSPs in the plant Arabidopsis thaliana (Siddique et al. 2008), one in the lactic acid bacterium Oenococcus oeni (Guzzo et al 1997). Although sHSPs are widespread, small genomes such as Helicobacter pylori 26,695 (1.66 Mb), Mycoplasma pneumoniae M129 (0.81 Mb), or Lactococcus lactis subsp. lactis II1403 (2.37 Mb) are devoid of this type of protein (Narberhaus 2002; Han et al. 2008).

sHSPs derive from a common ancestor of the vertebrate lens protein (Mörner 1894; Ingolia and Craig 1982) and now form a superfamily of small heat stress proteins of about 20 kDa (Horwitz et al. 1993; Jakob et al. 1993; de Jong et al. 1993), characterized by the presence of a central domain named the α-crystallin domain (Boelens 2014; Carra et al. 2017; Caspers et al. 1995; Jong et al. 1998). This domain has been under significant evolutionary pressure, and certain amino acids are still conserved among sHSPs (De Maio et al. 2019). The analysis of the sHSPs evolutionary tree by Vogel and colleagues showed that terminal domain specialization occurred during evolution later, several times and in parallel with and independently of the α-crystallin domain (Vogel et al. 2004; Kriehuber et al. 2010). This is reflected in rather high sequence heterogeneity between sHSPs (Mogk et al. 2019) that may be linked to the acquirement of multiple activities (De Jong et al. 1993; Carra et al. 2017; Obuchowski and Liberek 2020; Waters et al. 1996; Kappe et al. 2002).

The first trace of sHSPs in the history is probably from cyanobacteria such as Synechococcus sp. and Prochlorococcus sp. that lived on Earth two billion years ago (Fu et al. 2006). Indeed, sHSPs, which share some structural characteristics of the bacterial class A sHSP (corresponding to proteins with characteristics similar to those of Escherichia coli IbpA and IbpB) (Münchbach et al. 1999; Fu et al. 2006) have been identified in cyanophages (Fu et al. 2006) able to infect them (Schopf 2006; Maaroufi and Tanguay 2013). Animal sHSPs therefore evolved from bacterial class A sHSP via the lateral transfer gene either by the endosymbiotic mitochondria or by bacterial pathogen infections. Using the same mechanism of evolution, plant sHSPs might have been evolved from plant bacterial sHSPs of class B, like those found in Rhizobia (Fu et al. 2006). Because sHSPs are the least conserved family of chaperone proteins, the hypothesis of very early evolutionary divergence is reasonable (Kappe et al. 2002). However, according to the literature, 2 hypotheses have been put forward to explain the evolution of sHSPs: one proposes divergence from a common ancestor and the other convergence from several ancestral genes (Fu et al. 2006).

Structure of sHSPs

Contrary to sequence homology, the structure of sHSPs is well conserved in all organisms (Poulain et al. 2010; Klevit 2020). The majority of sHSPs adopt an IgG-type sandwich structure and are organized in large spherical oligomeric assemblies ranging from 10 to 18 nm in diameter (Haslbeck et al. 2005; Augusteyn 2004), even if others may adopt smaller forms (monomers or dimers) (Weeks et al. 2014; Boelens 2020). Each protein monomer has a low molecular weight (12–35 kDa) (Ingolia and Craig 1982; Kim et al. 1998; van Montfort et al. 2001; Kriehuber et al. 2010) and is mainly composed of β strands (60 to 70 percent) and few or no α-helixes (Augusteyn 2004). All sHSPs are composed of three domains, the two less conserved termini domains playing an important role in sHSP oligomerization and activities (Carra et al. 2017) and the well-preserved α-crystallin domain (ACD) (van Montfort et al. 2001; Kriehuber et al. 2010; Junprung et al. 2021; Klevit 2020; Mainz et al. 2015; Carra et al. 2017).

The α-crystallin is composed of around 90 amino acid residues among which the well-conserved signature motif GVLTL (between the β8 and β9 sheets) (Lentze et al. 2004 and Lentze and Narberhaus 2004; Augusteyn 2004; Sun and MacRae 2005 Bozaykut et al. 2014; MacRae 2000; Sun et al. 2002). The structure of the ACD can be split into two types of monomer (Fig. 1): the “bacterial” (Kim et al. 1998) and the “metazoan” structures (Delbecq and Klevit 2013). In both the ACD (involved in the oligomerization process), these are composed of seven or eight anti-parallel β strands that form a β-sandwich (Tikhomirova et al. 2017): one with β4, β5, and β6/7 strands and the second with β2, β3, β8, and β9 strands (Kim et al. 1998; Haslbeck et al. 2005; Kriehuber et al. 2010; Poulain et al. 2010; Basha et al. 2012).

Fig. 1.

Predicted structure of model A metazoan sHSP (alphafold2 HSPB9 of H. sapiens) and B bacterial sHSP (alphafold2 sHSP of M. jannaschii). The numbers correspond to the sheets allowing interactions between the sHSP. The amino acid sequences SRLFQxFG and IXI correspond to highly conserved sequences in the N- and C-terminal domain in bacteria

The N-ter domain is variable in length and amino acid composition (Chen et al. 2010; Lelj-Garolla and Mauk 2011), with 90 residues on average from 40 in bacteria to 100 in animals (Kriehuber et al. 2010). Although they have low sequence identity, some motifs such as SRLFQxFG are highly conserved (Tikhomirova et al. 2017; Shatov et al. 2018). A hydrophobic groove formed by the association of the β2 and β7 sheets of the ACD of each monomer creates an open space into which the N-ter domain can engage to stabilize the dimer. This conformation is then said to be open because it facilitates the interaction of the sHSP with its substrate (Sudnitsyna et al. 2012; Tikhomirova et al. 2017, Boelens 2020).

The short variable C-ter domain (approximately 20 amino acids) contains the highly conserved I/L-x-I/L motif which acts as an “anchor” at the ACD and thus participates in oligomer formation (Saji et al. 2008; Poulain et al. 2010; Sudnitsyna et al. 2012; Delbecq and Klevit 2013; Hilton and Benesch 2012 and 2013).

Oligomerization of sHsp

As soon as they are produced, monomers are organized into dimers which are therefore the first level of the structural order (Delbecq and Klevit 2013). Common oligomerization mechanisms have been described between the sHPSs of different organisms according to the ACD organization, bacterial or metazoan. Either, there is an interaction between the β6 on one monomer with the β2 of the second monomer, which results in the so-called bacterial dimer (Fig. 2A) or the interaction between the β6 and β7 strands of two different monomers to produce the “metazoan”-type dimer (van Montfort et al. 2001; Sun and MacRae 2005; Delbecq and Klevit 2013; Hilton et al. 2013). Then sHSP dimers can interact to form a higher oligomeric structure. For most of sHSPs, the palindromic I/V-x-I/V motif in the C-ter domain on the first dimer interacts with the hydrophobic groove formed by β4 and β8 of the adjacent dimer ACD (Jong et al. 1993; van Montfort et al. 2001; Sun and MacRae 2005; Basha et al. 2012; Santhanagopalan et al. 2018; Mainz et al. 2015; Tikhomirova et al. 2017; Mogk et al. 2019) (Fig. 2B). The dimer assemblies will then interact with each other to form disks via the interaction between the still free C-ter domain and the β9 strand located at the end of the ACD domain, thus forming large oligomers (van Montfort et al. 2001) (Fig. 2C). However, for many mammalian, sHSP formation of large oligomers may be different, depending on N-terminal domain (Delbecq et al. 2015).

Fig. 2.

Oligomerization process of sHSP using the example of M. marinum sHSP (pdb 5ZUL) (Bhandari et al. 2019). A Dimerization: interaction between the β6 strand (green) of a monomer with the β2 strand of the adjacent monomer (blue). B Oligomerization: interaction between the C-ter domain of a dimer (red) with the hydrophobic groove formed by the β4 (turquoise) and β8 (orange) strands of an adjacent dimer. C Large oligomer formation: interaction between the free C-ter domain (pink) of the oligomer with the β9 (red) strand on another oligomer

These sHSPs are linked to the quality control of cellular proteins via the considerable dynamic dissociation and reassociation of oligomers around their substrates. The speed of these associations could be one of the main factors influencing chaperon activity (Fu et al. 2006 and 2012; Tikhomirova et al. 2017).

sHSPs, actors of stress response with multiple activities

The proper folding of proteins is essential for life (Klaips et al. 2017), but unfortunately stress conditions dramatically increase their level of unfolding (Tyedmers et al. 2010). Indeed, the modification of environmental conditions (i.e., acidification) may modify the physicochemical properties of amino acids, changing their interaction and consequently the final organization of the proteins. As the biological function of a protein is strongly linked to its tertiary structure, this denaturation will induce the loss of its function affecting the whole cellular and replicative metabolism (Berger et al. 1996; Lim et al. 2000). To survive, organisms have succeeded in adapting to their environment by developing various strategies for maintaining proteins. One of these strategies is the production of molecular chaperones such as sHSPs (Jakob et al. 1993; Boyle and Takemoto 1994; Singh et al. 1995; Wang and Spector 1995; Chang et al. 1996; Guzzo et al. 1997). Since the beginning of the 1990s, several studies have suggested the molecular chaperone roles played by eukaryotic and bacterial sHSPs. They constitute the first line of stress defense of the protein quality control system (Bukau et al. 2006; Chen et al. 2010; Haslbeck and Vierling 2015; Haslbeck et al. 2016; Mogk et al., 2019; Obuchowski and Liberek 2020). sHSPs confer resistance to several kinds of stresses such as heat, cold, salinity, ethanol, acid, and chemicals (Carra et al. 2017; Waters and Vierling 2020; Ma et al. 2021; Tian et al. 2012). sHSPs sequester denatured proteins in an ATP-independent manner (Jakob et al. 1993; Rodriguez Ospina et al. 2022) and thus could be considered more as “sequestrases” than “holdases” (Haslbeck and Vierling 2015; Haslbeck et al. 2016; Mogk et al. 2019; Reinle et al. 2022). To refold the misfolded protein, oligomeric sHSPs must collaborate with the Hsp70/100 complex (Veinger et al. 1998). The chaperone Hsp70 binds to the denatured sHSP-protein assemblies, allowing the recruitment of Hsp100 to extract and refold these proteins properly (Żwirowski et al. 2017). The same mechanisms have been observed both on eukaryotes such as Saccharomyces cerevisiae (Glover and Lindquist 1998; Cashikar et al. 2005), on bacteria like E. coli (Mogk et al. 2003), and cyanobacteria such as Synechocystis (Giese et al. 2005).

In addition to their role of molecular chaperone, the literature suggests a growing number of roles fulfilled by sHSPs. They have been described in some microorganisms as having an impact on virulence (Obuchowski and Liberek 2020), cyst and spore formation (Henriques et al. 1997; Cocotl-Yañez et al. 2014; Jee et al. 2018), dormancy (Yuan et al. 1997, Cunningham and Spreadbury 1998), cell cycle regulation (Kim et al. 2011), immune system modulation (Longo et al. 2020), and biofilm formation (Yeh et al. 1997; Zara et al. 2002; Diaz et al. 2006; Kuczynska-Wisnik et al. 2010; Pidot et al. 2010). Moreover, another role has been attributed to certain sHSPs corresponding to the capacity of maintaining cell membrane integrity. This new activity has given rise to the concept of sHSPs as lipochaperones (Maitre et al. 2014) and will be described more precisely in the next chapter.

Zoom on sHSP lipochaperone activity

The maintenance of the optimal state of fluidity of the cellular membrane is crucial to withstand stressful conditions. Organisms cannot synthetize the lipid bilayer de novo but can maintain a continuous process of lipid production and modification (Obuchowski and Liberek 2020). However, these two processes take time and are not sufficient at the beginning of stress to limit changes in the fluidity and/or structure of the membrane microdomains. These changes may act as a sensor perturbation, leading to the expression of sHSP that can exert their chaperon and eventually lipochaperone roles (Glatz et al. 2015, Vigh et al. 2007; Maitre et al. 2012). This second activity seems to be temporally favored, suggesting the importance to prioritize the regulation of membrane fluidity as demonstrated for Lo18 from O. oeni (Maitre et al. 2014). Currently, little is known about this lipochaperone activity; however, the few sHSPs described for it are present in model organisms representative of all kingdoms of life. Although they are known to interact with lipids (Table 1) (Coucheney et al. 2005; Balogi et al. 2005; Kim et al. 2011), the mechanism of interaction remains poorly understood. Among the knowledge, a distinction can be made between sHSPs described as interacting with the membrane and those that have a real role in maintaining membrane fluidity. Immunolabeling techniques allow to observe the interaction between sHSP 17.8 of Arabidopsis thaliana with the chloroplast membrane (Kim et al. 2011), human HSPB1 and HPB5 with the mitochondrial membrane system and lipid rafts (De Maio and Hightower 2020), HSPA from Synechococcus sp. with the thylakoid membrane (Nitta et al. 2005; Obuchowski and Liberek 2020), and HSP16.3 from Mycobacterium tuberculosis with positively charged lipid bilayers of M. tuberculosis (Sun and MacRae 2005; Zhang et al. 2005). Wilhelmy’s method showed that the interaction between the two sHSPs of Schizosaccharomyces pombe, HSP15.8 and HSP16, and different isolated lipids extracted from its membrane was sHsp-dose-dependent (Glatz et al. 2015).

Table 1.

sHSPs described for their lipochaperone activity in model organisms

	Organism	sHSP	References
Human	Homo sapiens	HSPB5	De Maio et al. (2019)
		HSPB1
Plant	Arabidopsis thaliana	HSP17.8	Balogi et al. (2008), Kim et al. (2011)
	Daucus carota	HSP17.7	Ahn and Zimmerman (2006), Balogi et al. (2008)
Yeats	Schizosaccharomyces pombe	Hsp15.8	Glatz et al. (2015)
		Hsp16
Bacteria	Lactiplantibacillus plantarum	Hsp18.55	Capozzi et al. (2011)
	Oenococcus oeni	Lo18	Coucheney et al. (2005), Maitre et al. (2012, 2014), Weidmann et al. (2010, 2017)
	Synechocystis sp.	Hsp17	Balogi et al. (2008), Tsvetkova et al. (2002), De Maio and Hightower (2020)
	Thermosynechococcus vulcanus	HspA	Roy et al. (1999)
	Synechococcus sp.	HspA	Nitta et al. (2005), Obuchowski and Liberek (2020)
	Mycobacterium tuberculosis	HSP16.3	Lini et al. (2008); Sun and MacRae (2005); Zhang et al. (2005)

Lipochaperone activity was observed in vivo and in vitro by fluorescence anisotropy measurements (reflecting the fluidity state of lipid bilayers) on Lo18 of O. oeni (Coucheney et al 2005; Maitre et al. 2012 and 2014; Weidmann et al. 2010 and 2017) and sHSP 18.55 from Lactiplantibacillus plantarum interacting with the charged membrane after ethanol stress and severe heat stress (> 48 °C) (Capozzi et al. 2011). Differential scanning calorimetry showed that sHSP17 from Synechocystis sp. can increase the physical order of the membrane to stabilize lipid membrane (Tsvetkova et al. 2002; De Maio and Hightower 2020) in response to heat stress (Balogi et al. 2008; Sun and MacRae 2005). Finally, electrolyte leakage analysis was performed to measure cell membrane stability at elevated temperature on sHSP17.7 of Daucus carota (Ahn and Zimmerman 2006).

Although the interaction mechanism is certainly lipid/sHSP dependent, other characteristics are also fundamental (Fig. 3):

Fig. 3.

Mechanism of interaction described between sHSP and the membrane. (1) Dissociation of the sHSP oligomer near the membrane allows lipid-protein interaction. (2) Interaction with the polar head of phospholipids according to their structure and state of fluidization (Tsvetkova et al. 2002; Maitre et al. 2012). (3) Electrostatic force, occurring during the disassembly process, can attract the protein to the membrane and enhance hydrophobic force. This electrostatic force may expose hydrophobic domains of sHSP leading to protein structure modification (Zhang et al. 2005) which can then bind to the membrane (Chen et al. 2003; Tsvetkova et al. 2002)

Association dissociation

The dissociation of the sHSP oligomer is a prerequisite for lipid binding. The dynamics of oligomerization (association-dissociation) seems to be very important for all the sHSPs described above. Indeed, the loss of this equilibrium between dissociation and association leads to the inability of sHSPs to interact with the membrane (Chen et al. 2003; Balogi et al. 2008; Maitre et al. 2014; Zeng et al. 2015). Balogi and colleagues suggest that the dissociation of the oligomer into the dimeric structural unit near the membrane will allow their uptake by the membrane (Balogi et al. 2008), certainly due to the exposure of certain protein domains initially hidden in the oligomer (Zeng et al. 2015). It is interesting to note that the dissociation of oligomers appears to occur, in vitro, in a temperature range of 40 to 50 °C. For example, this dissociation happens at 43 °C for HSP26 from S. cerevisiae and 40 °C for sHSP 16.9 from wheat (Chen et al. 2003), temperature range similar to that of membrane fluidization.

Specificity of sHsp-membrane interaction

The reversible interaction between sHSPs and membrane lipids is very specific and depends, on the one hand, on the sHSP, and on the other hand on the fluidity state of the membrane due to the degree of unsaturation of the fatty acids and the nature of the polar heads (Balogi et al. 2008; Csoboz et al. 2022). These interactions prevent the formation of hyperfluidic states by regulating the lipid polymorphism of the membrane during phospholipid disorganization (Glatz et al. 2015; Tsvetkova et al. 2002; Coucheney et al. 2005). Phospholipid composition seems to be a key determinant of sHSP-membrane interaction. Indeed, the lipid bilayer is not just a hydrophobic wall, but a highly dynamic structure with a natural movement of phospholipids, making the membrane a very fluid environment (De Maio et al. 2019). For example, Lo18 from O. oeni seems to interact better with a mix of saturated, unsaturated, and cyclic fatty acid than a highly rigid membrane with a high content of saturated fatty acids (Maitre et al. 2012 and 2014). HSP15.8 from S. pombe has a higher affinity for monolayers (prepared in the Langmuir balance system) but interacts in an equivalent manner for all the major phospholipid groups comprising polar heads such as inositol, ethanolamine, and glycerol. In contrast, the second sHSP of S. pombe, HSP16, is more specific since it has a greater affinity for glycerol but less so for inositol (Glatz et al. 2015). Moreover, infrared spectroscopy performed on HSP16.3 of M. tuberculosis indicates that interaction is mediated via the head group of phospholipids and the hydrophobic part of the protein (Chen et al. 2003), as was also demonstrated for Lo18 of O. oeni by measuring the anisotropy fluorescence of liposomes (Maitre et al. 2012).

Electrostatic and hydrophobic interactions

Electrostatic force is an important element in these interactions which may occur during the disassembly process. All the sHSPs described display a negative charge in neutral buffer solution and mainly interact with positively charged membranes (Capozzi et al. 2011, Glatz et al. 2015, Sun and MacRae 2005). However, there are also hydrophobic interactions between sHSPs and the lipid bilayer that should not be neglected. Indeed, from the protein side, electrostatic interactions generate a slight conformational change in sHSP, exposing hydrophobic zones (Chen et al. 2003; Tsvetkova et al. 2002). Then, the flexibility and dynamism of M. tuberculosis HSP16.3 increases in the presence of Mycobacterium membrane lipids (Zhang et al. 2005). From the membrane side, fluidization creates a hydrophobic docking surface on the cell that allows the binding of sHSP, as described for the human HSPB1 (Csoboz et al. 2022).

Analysis of primary structure: the beginning of understanding

Although the primary sequence of sHSPs is poorly conserved, the protein structure (Poulain et al. 2010) and the primordial activities of these proteins are rather homogeneous (Jong et al. 1998). This suggests that certain motifs remain highly conserved between sHSPs (Jong et al. 1998), such as those of the α-crystallin domain (Morimoto and Santoro 1998).

Previous studies have identified specific motifs containing key amino acids for the structure and/or chaperone activity of sHSP (van Montfort et al. 2001; Mao and Chang 2001; Lentze et al. 2003; Weidmann et al. 2010). Lentze and colleagues (2003) have developed a library of HspH proteins from B. japonicum modified by site-directed mutagenesis on highly conserved amino acids identified by multiple alignment of 12 sHSPs from different organisms (plants, animals, bacteria). The majority of the modified proteins were altered in their quaternary structure (total loss or alteration) and/or their chaperone activity. The two modified proteins G74E and F94A/D (mutated at glycine and phenylalanine respectively at position 74 and 94) are able to form only dimeric structures, resulting in a loss of chaperone activity. The same findings were reported for the wheat protein Hsp16.9 modified for the same residues (G83 and F106) (van Montfort et al. 2001). A reduction in oligomer size is observed for the modified HspH proteins L100A/A109S/L116A leading to a reduction or abolition of chaperone activity. The same mutation in Hsp16.3 of M. tuberculosis (L122) or in Lo18 of O. oeni (A123S) results in the formation of unstable or small complexes with low chaperone activity, respectively (Mao and Chang 2001; Weidmann et al. 2010). Finally, the loss of chaperone activity for the modified HspH protein G114A is observed despite the maintenance of quaternary structures suggesting an altered substrate interaction site.

All these works show that some residues, highly conserved on the sHSP primary sequence, are essential for establishing chaperone activity. In line with these studies, it seems very relevant to use a similar method to identify conserved motifs among sHSPs having the additional lipochaperone activity.

Twelve sequences of sHSPs (from human, plants, yeast, and bacteria) described as having lipochaperone activity (Table 1) were compared with 38 protein sequences of the sHSPs available in the literature, also representative of the different kingdoms (Fig. 4). However, in this second group, no lipochaperone activity is described, which does not exclude that some proteins may have it without being identified. Sequence alignment performed by ClustalW algorithms using Jalview 2.11.2.5 software between these two groups (described or not described for lipochaperone activity) shows the presence of four highly conserved amino acid motifs for the “lipochaperone activity” group (Fig. 4): “DKKET” on the β2, “ELPG” upstream of the β4 strand, the “LTISGKRE” on the β5, and “RSERSYGSFR” on the β7. Among these motifs, some amino acids are particularly conserved such as respectively glutamic acid (E), leucine (L), glycine (G), and phenylalanine (F) for each motif that are present in 77.3% of the cases for the sHSP of the lipochaperone group versus 50% for the other sHSPs (p value 0.042).

Fig. 4.

A sHSP primary sequence alignment by ClustalW algorithms using Jalview 2.11.2.5 software. Comparison of the alignments of 12 primary sHSP sequences described for their lipochaperone activity named “lipochaperone sHSP” with 38 primary sHSP sequences not described for this role named “Non lipochaperone sHSP.” Highly conserved (dark blue), conserved (blue), non-conserved (white), and unaligned (gray) residues are shown. Four domains, present on the β2 sheet (red arrow), before the β4 sheet (yellow arrow), β5 sheet (green arrow), and β7 sheet (blue arrow) respectively are more conserved on the lipochaperone sHSP. B Location of the four particularly conserved domains on a 3D predicted representation of an O. oeni sHSP Lo18 monomer

Almost no studies have investigated the primary structure for lipochaperone activity. Only the work on the Lo18 protein of O. oeni shows the importance of the second aromatic residue present on the last motif, “RSERSYGSFR” on β7 (Weidmann et al. 2010 and 2017; Maitre et al. 2012). Anisotropy measurements performed on the modified Lo18 protein Y107A (aromatic tyrosine replaced by alanine) showed its inability to prevent fluidization upon heat shock (Weidmann et al. 2010). Using a predicted model of Lo18 (obtained by trRosetta and confirmed by Alphafold 2), the aromatic ring of this tyrosine could act as a tenon between the two planes of the α-crystallin domain of Lo18, by interacting with adjacent amino acids present on different β sheets (Fig. 5A). The loss of the aromatic ring (transforming tyrosine into alanine) will change these interaction distances (Fig. 5A and B). In addition, it is interesting to note that three of them are present on highly conserved amino acid motifs described above (yellow and blue arrow, Fig. 4).

Fig. 5.

Amino acid interaction distances in O. oeni sHSPf Lo18. A Interaction between the aromatic ring of tyrosine at position 107 (red) and its adjacent amino acids (blue) within an average radius of 3.6 Å or B interaction of the alanine at position 107 from the tyrosine mutation (red) with adjacent amino acids (blue) within an average radius of 5.925 Å. The loss of the aromatic ring alters the interaction distances with other amino acids

Nevertheless, other sHSPs, not yet studied for lipochaperone activity, have shown that some amino acids of these motifs are important for their oligomeric structure. This is the case, for example, for HSP 16.5 from M. jannaschii modified for the second arginine present on the β7 sheet or for sHsP 16.9 from wheat described above (Jong et al. 1998; van Montfort et al. 2001).

It seems possible that these highly conserved amino acid motifs may play a role in the structure of sHSP, possibly allowing them to interact with the membrane. These results highlight the importance to study the lipochaperone activity in all sHSPs having these different motifs. Priority could be given to bacterial sHSPs, as they are more sensitive to point mutations, thanks to their structural rigidity (Studer et al. 2002; Lentze et al. 2003). They could be excellent candidates for in vitro and in vivo analysis of these four amino acid motifs.

Conclusion

To conclude, it appears that sHSPs can interact with lipids via certain amino acid residues present on the β2, inter β3/4, B5, and β7 strands. These residues could have an important role (i) in the structure and dynamics of oligomerization and (ii) for the interaction and maintenance of optimal cell membrane fluidity. To better understand the overall mechanisms allowing this sHSPs—cell membranes interaction, new experiments should focus on this activity and the highly conserved amino acid motifs described above.

sHSP deficiency may be involved in several human pathologies such as neurodegenerative diseases or familial exudative vitreoretinopathy. Indeed, this last pathology may be caused by the inability of human α-crystallin to interact with the retinal cell membrane suggesting a lipochaperone activity (Bakthisaran et al. 2015; Toth et al. 2015). Finally, the ability of sHSPs to protect membranes could be beneficial for probiotics or ferment production, especially for lactic acid bacteria. Indeed, this would allow them to resist production processes such as freeze-drying, which is known to induce a deleterious modification of the membrane fluidity of microorganisms (Arena et al. 2019). This has already been described for the Hsp18.55 of L. plantarum which induces cryotolerance by acting as an antifreeze protein (Arena et al. 2019).

It appears that understanding the lipochaperone activity of sHSPs could become an important area of research in the next decades.

Author contribution

Conceptualization, TB and SW; methodology, TB and SW; formal analysis, TB and SW; investigation, TB and SW; resources, TB and SW; writing—original draft preparation, TB; writing—review and editing, TB and SW; visualization, TB and SW; supervision, SW; project administration, SW; funding acquisition, SW. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by Conseil Régional de Bourgogne, the Université de Bourgogne grant number 2021Y-09559, and the ministère de l’Enseignement supérieur de la Recherche, grant number MESR 2020–04.

Declarations

Conflict of interest

The authors declare no competing interests.