Ssb shows low affinity interaction with classic Hsp70 substrate

In contrast to other Hsp70s, Ssb tolerates a number of alterations in its peptide-binding cleft without affecting its in vivo activity (PFUND et al., 2001; and data not shown). Several studies tried to monitor in vitro interaction of Ssb with peptide substrates, which is well established for canonical Hsp70 chaperones such as yeast Ssa or E. coli DnaK. However, so far no interaction of Ssb with peptides could be monitored (PFUND et al., 1998;

PFUND et al., 2001; GAUTSCHI et al., 2002; HUNDLEY et al., 2002). Apparently, there is a difference between Ssb and other Hsp70 chaperones with regard to substrate interaction. Two explanations are possible that are not mutually exclusive: either Ssb displays a very low affinity for peptide substrates and/or it possesses a different substrate-binding specificity compared to other Hsp70s.

To get insights into a potential difference in substrate binding of Ssb comparied to other Hsp70s, MD simulations of a crystallized DnaK-peptide complex (ZAHN et al., 2013) and a modeled Ssb1 variant binding to the same peptide (NRLLLTG) or to the classical Hsp70 substrate peptide APPY (PFUND et al., 2001) were performed. Analysis of protein-substrate binding revealed an overall very similar interaction pattern for both complexes (Fig. 24), suggesting that Ssb in general should be able bind to canonical Hsp70 substrate peptides.

Figure 24: Simulation of DnaK and Ssb show qualitatively identical substrate binding. A) DnaK-peptide complex after 750 ns MD simulation. SBD and peptide (NRLLLTG) are displayed in transparent green and magenta, respectively. The most stable protein-peptide hydrogen bonds (indicated by dotted black lines) are S427-L3 (89 %), S427-L5 (98 %) and M404-L4 (86 %) (with percentages in brackets representing stability during simulation). Residues forming hydrophobic contacts are depicted in stick representation involving I438, F426 and L5. B) Ssb1-peptide complex as depicted in A). The most stable protein-peptide hydrogen bonds are T433-L3 (82 %), T433-L5 (94 %) and Q410-L4 (55 %), residues forming hydrophobic contacts involve F444, F432 and L5. (Simulations performed by A. Jain).

To further address this issue, an experimental approach was developed in which the interaction of Ssb with several kinds of biotinylated peptides was monitored with size exclusion chromatography. The peptide APPY was used as a classic Hsp70 substrate for which Ssa possesses an apparent K_d of 5 $M (PFUND et al., 2001).

APPY was incubated either alone or in combination with Ssb1 and free peptide was separated from Ssb-bound one via gel filtration. After size exclusion, fractions were spotted from high to low molecular weight followed by the specific detection of either Ssb or peptide. Free APPY peptide elutes in fractions No. 29-34 (Fig. 25A top)

" " ./01230"456"

" (("

whereas Ssb runs faster in fractions 7-12 (Fig. 25A bottom). If incubated together, APPY shifts into earlier Ssb-containing fractions (7-12), indicating an interaction with Ssb (Fig. 25A, middle). A similar shift of APPY could be detected also for the ribosome-binding deficient Ssb1#601-13, indicating in vitro functionality of this mutant also with respect of substrate interaction (Fig. 25B).

Figure 25: Wild type and mutant Ssb1 interact with canonical Hsp70 substrate peptides in vitro. A) Analysis of Ssb1 and APPY peptide-interaction via gel filtration. Biotin-labeled APPY peptide was incubated either alone (top) or with Ssb1 (middle). Ssb1 alone served as a further control (bottom). Samples were applied to size exclusion chromatography to separate free peptides from Ssb-bound ones. Resulting elutions were fractionated and spotted from high to low molecular weight (fraction numbers are indicated in grey) onto nitrocellulose membranes followed by immunological detection of either APPY (via HRP-Streptactin; left) or Ssb1 (right). B) Analysis of APPY peptide binding by mutant Ssb1#601-13 protein as depicted in A).

To prove specificity of the APPY-Ssb1 interaction, Ssb was either pretreated with ATP to force the open conformation (Ssb_ATP) or with apyrase that hydrolyses Ssb-bound ATP to ADP/AMP, leading to lid closure (Ssb_ADP). Then the APPY peptide was added and in the case of Ssb_ATP a subsequent apyrase incubation step followed to stabilize the Hsp70-peptide complex. Analyses upon gel filtration showed that more APPY shifted to the Ssb-containing fractions (7-12) if it was added to Ssb in the ATP-open conformation (Fig. 26A).

Quantification of the shifted APPY signals showed that the peptide interaction was reduced to 71 % in the case of SsbADP (Fig. 26B). This proves that the nucleotide state of Ssb is critical for an efficient peptide binding and that the canonical Hsp70 ATPase cycle is involved in this interaction (see Fig. 11). Remarkably, the amount of APPY peptide that shifted into the 70 kDa fractions was substantially higher if the peptide was incubated with DnaK instead of Ssb (Fig. 26C), indicating a higher substrate affinity of DnaK. A similar in vitro Ssb binding of other peptides, characterized by the hydrophobic cluster of canonical Hsp70 substrates, could be observed (data not shown). Incubation with a non-Hsp70 substrate peptide P2-beta that is characterized by a hydrophobic

"

()"

Figure 26: Ssb interacts specifically with the APPY peptide, but binding is less efficient than the DnaK-APPY interaction. A) Analysis of Ssb1 and APPY peptide-interaction via gel filtration in dependency of the nucleotide state. Ssb1 was either pretreated with ATP (top) or apyrase (bottom) followed by incubation with biotinylated APPY peptide. Upon size exclusion chromatography the eluted fractions were spotted from high to low molecular weight (fraction numbers are indicated in grey) onto nitrocellulose membranes followed by the detection of either biotinylated APPY (via HRP-Streptactin; left) or Ssb1 (via Ponceau S staining; right). B) Quantification of shifted APPY peptide in dependency of the nucleotide state of Ssb1, as shown in A). Biotin and Ponceau S signals from fractions 1-12 were quantified with ImageJ and the interaction of Ssb1ATP with APPY was set as 1 (biotin / Ponceau S signal). Data are represented as mean of three independent approaches including experimental standard deviation (0 for SsbATP and 0.08 for SsbADP). C) Biotinylated APPY peptide was incubated either with Ssb or DnaK. Fractions after size exclusion were analyzed for the presence of peptide (via Streptactin; left) or the Hsp70 chaperone (via Ponceaus S; right). D) Ssb1 does not interact with the acidic peptide P2-beta.

Biotinylated P2-beta was incubated either alone or together with Ssb1. Fractions after size exclusion were analyzed for the presence of peptide (via Streptactin; left) or Ssb1 (via Ssb1 specific antibody; right).

" " ./01230"456"

" (*"

cluster flanked by negatively charged residues did not result in a shift of the peptide signal, neither to Ssb- (Fig. 26D) nor to DnaK-containing fractions (data not shown). Thus, it can be concluded that Ssb binds peptide substrates with specificity similar to other Hsp70s, favoring a hydrophobic cluster flanked by positively charged residues (RÜDIGER et al., 1997), but binds its substrates with lower affinity.

As shown above, the amount of shifted peptides in the case of Ssb seemed to be much lower than in the presence of DnaK (Fig. 26C), although in the experimental approach comparison of two different proteins was not suitable for an exact quantification. Therefore, fluorescence polarization measurements were performed to monitor anisotropy of IAANS-labeled model peptide sigma³²-Q132-Q144-C for which DnaK possesses a Kd of 78 nM (MAYER et al., 2000). In general, an increase in peptide anisotropy could be detected for both, wt and mutant Ssb1 protein, without major differences between the Ssb constructs (Fig. 27A). Peptide release kinetics of wt and mutant Ssb1 were monitored, showing again no substantial differences for all Ssb version (Fig. 27B) but a similar k_off as described for DnaK (0.001 s^-1; MAYER et al., 2000). This is important, since in the case of DnaK alterations in the lid increase the off-rate of peptides, as this region constitutes a physical lid-like barrier to substrate release (ZHU et al., 1996). If this had been detectable for the ribosome-binding deficient Ssb1#601-13 as well, one could have argued that the mutant is not hampered in ribosomal interaction but rather shows a higher degree of nascent polypeptide release. But, in vitro peptide binding and release are the same for wt and mutant Ssb1, indicating that mutations that hamper ribosome binding do not influence substrate detachment.

However, it could still be possible that wt Ssb1 more efficiently stimulates its intrinsic ATPase activity than the Ssb1#601-13 mutant, leading to a more stable peptide binding and thus possibly to a lower koff.

Figure 27: Wild type and mutant Ssb1 interact with sigma³² peptide in vitro and show a DnaK-like peptide release.

A) Peptide binding of wt and mutant Ssb1 was assessed using fluorescence anisotropy measurements. Recombinant Ssb1 and mutants thereof were incubated with IAANS-labeled sigma³²-Q132-Q144-C peptide followed by fluorescence polarization measurements. B) Peptide release kinetics of wt and mutant Ssb1 using IAANS-labeled peptide sigma³²-Q132-Q144-C.

(Performed by R. Kityk).

In general, in vitro binding of wt and mutant Ssb1 to canonical Hsp70 model peptides was detectable, although it was not possible to determine the exact K_d of any protein-peptide interaction since the binding curves did not reach saturation when titrating even up to 100 $M Ssb1 (data not shown). Therefore, it can finally be concluded that concerning substrate specificity wt and mutant Ssb behave like other Hsp70s but the affinity towards

"

(+"

substrates seems to be dramatically lower, at least in the absence of cofactors. This might be a mechanism that excludes stable binding of Ssb to substrates in the cytosol. One could speculate that a productive interaction between Ssb and nascent polypeptides is mediated via its multivalent ribosome binding of the CTD and SBD in combination with the cofactor RAC. This could position the putative substrate-binding pocket of Ssb in an optimal orientation to the ribosomal tunnel exit to guarantee a productive interaction with the nascent chain despite of its low affinity for peptide substrates.

Discussion & Outlook (A)

" (,"