It was demonstrated that in solubilized mitochondria yeast Tim14/Pam18 and Tim16/Pam18 form a stable hetero-oligomeric complex with a reduced ability to stimulate the ATPase activity of mtHsp70, compared to Tim14/Pam18 alone. Additionally, the formation of the Tim14/Pam18-Tim16/Pam16 complex is essential for the correct function of both proteins in vivo [16]. Previous studies have also shown that the J-domains alone of Pam18/Tim14 and Pam16/Tim16 are able to form a complex [16, 20,26]. Therefore, in order to test the ability of human Tim14/Pam18 to interact with yeast Tim16/Pam16 and to form a complex, we cloned the soluble J domains of both proteins (a.a 24–116 and a.a 25–130, respectively) and co-expressed them in bacteria. Since only the yeast Tim16/Pam16s (a.a 25–130 of Tim16/Pam16) contains an octahistidine tag, the efficient purification of both proteins during all steps of isolation indicates that a complex is indeed formed between hTim14/Pam18s and yeast yTim16/Pam16s (figure 1). The proof of concept for this strategy, using the homologous yeast proteins, was published previously [26]. As shown in Figure 1, human Tim14/Pam18 was able to form a complex with yeast Tim16/Pam16. Using the same strategy, we were also able to demonstrate complex formation between yeast Tim14/Pam18 and human Tim16/Pam16 (not shown) and between human Tim14/Pam18 and human Tim16/Pam16 (not shown). We conclude that the putative human Tim14/Pam18 and Tim16/Pam16 do interact, in vitro, to form heter-oligomers and that they can also form heterologous complexes with their yeast counterparts.

Previous studies showed that yTim14/Pam18s and yTim16/Pam16s assemble into hetero-dimers in solution, [16,19,26]. We used cross-linking to examine the oligomeric state of the three complexes purified in this study (one homologous human hetero-dimer and two heterologous human/yeast hetero-dimers). The purified complexes were cross-linked using DSS and the cross-linking products were analyzed using SDS-PAGE. The results presented in figure 2 show clearly that upon exposure to the cross-linker, bands representing the monomeric proteins weaken while at the same time hetero-dimer cross-linking products appear, in all three complexes examined. A similar pattern, indicative of dimeric molecules, was observed previously for the homologous yeast complex [26]. The similarity in the cross-linking pattern of the three complexes examined in this study, together with the previously described cross-linking pattern for the yeast Tim14/Pam18-Tim16/Pam16 complex, supports the idea that the interaction between the yeast and human Tim14/Pam18-Tim16/Pam16 complex is evolutionarily conserved. This suggests that their function is also most likely conserved.

Yeast Tim14/Pam18 and Tim16/Pam16 are members of the J and J-like protein families, respectively. These proteins are characterized by a large domain composed of three α helices. As such, they exhibit distinct CD spectra [26]. Since hTim14/Pam18 and hTim16s/Pam16 are homologues of these yeast proteins, oligomers containing them should display a similar CD spectrum, assuming they fold into similar structures. To determine whether this is indeed the case, we carried out a CD analysis to determine the secondary structure of the purified constructs of hTim14/Pam18s (not shown), hTim16/Pam16s (figure 3), a complex of yTim14/Pam18s-hTim16/Pam16s (figure 3) and a complex of hTim14/Pam18s-hTim16/Pam16s (figure 3). A similar CD spectrum was obtained for all proteins tested, which was characterized by two minima at 222 nm and at 208 nm, typical of α-helical proteins (Figure 3). The results indicate that the human and yeast purified proteins are similarly folded and contain essentially α-helical structures.

We reported previously that yTim14/Pam18s and yTim16/Pam16s are only marginally stable proteins that undergo unfolding at very low temperatures (Tm values for the individual proteins of 16.5°C and 29°C, respectively) [26]. Upon mixing the purified proteins, or when both proteins are co-expressed in bacteria, yTim14/Pam18s and yTim16/Pam16s form a hetero-dimer that is thermally more stable (Tm of ~40° C) compared to the individual proteins. Consequently, we proposed that any dissociation of the yeast Tim14/Pam18-Tim16/Pam16 hetero-dimer complex in vivo would theoretically lead to denaturation of these essential import components. Therefore, it was speculated that the formation of a stable Tim14/Pam18-Tim16/Pam16 complex is favored in vivo and the regulation of their function on the translocation motor is exerted through conformational changes [26].

We determined unfolding midpoints (Tm) by monitoring changes in the secondary structure content, detected by CD spectroscopy at 222 nm. The unfolding midpoints of individual Tim/Pam proteins and their complexes are summarized in Table 1. The results indicate that while the three proteins yTim14/Pam18, yTim16/Pam16s and hTim16/Pam16s are essentially unstable proteins (Tm of 16.5, 29, 22.5° C, respectively) the human Tim14/Pam18s is significantly much more stable (Tm of ~45). Complex formation between the yeast proteins, yTim14/Pam18 and yTim16/Pam16s, and subsequent conformational changes lead to the stabilization of their complex [26]. Interestingly, when hTim16/Pam16s interacts with hTim14/Pam18s, the stability of the complex is similar to that of the latter protein. Thus, two different factors contribute to the stability of yeast and human Tim/Pam proteins. In the case of the yeast proteins, the individual proteins are significantly less stable than their complex. Thus, complex formation between them probably increases their folding which in its turn increases complex stability (41° C compared to 16.5° C and 29° C for the individual proteins). In the case of the human proteins, the Tm of their complex is very close to that of hTim14/Pam18s. In general, when we formed a complex that contained one of the human Tim/Pam proteins, the stability of the complex was not increased further than the stability of either of the individual proteins. We conclude that the thermal stability of human Tim14/Pam18s determines the stability of the full human complex.

As mentioned above, mtHsp70 serves as the core of the mitochondrial protein import motor [9,27,28]. This chaperone drives insertion of unfolded precursors into the matrix in an ATP-dependent manner. In vivo, mtHsp70 functions with the aid of several co-chaperones that regulate its ATPase activity. The first is the nucleotide exchange factor, Mge1, which enables ADP/ATP exchange and recycling of mtHsp70 [29–31]. The second co-chaperone is the yeast Tim14/Pam18, known to have a major role in stimulating the ATPase activity of mtHsp70 [11,14,15]. The third component, yeast Tim16/Pam16, antagonizes the function of Tim14/Pam18. Tim16/Pam16 specifically inhibits the Tim14/Pam18-induced ATPase stimulation of mtHsp70 [16,19]. We wanted to examine whether the human Tim/Pam homologues can affect the ATPase activity of yeast mtHsp70. An effect on ATP hydrolysis will indicate that not only the structure of the human Tim/Pam proteins is conserved, but also their function. To this end, we studied the ATPase activity of yeast mtHsp70 in the presence of the various co-chaperones (Figure 5).

In the absence of any additional components or in the presence of Mge1, mtHsp70 displays a low basal ATPase activity (1 turnover/min). Further addition of the purified yTim14/Pam18, containing a J-domain, significantly increases the rate of hydrolysis. Upon addition of the yTim16/Pam16s, there is an evident decrease in the ATPase activity [19,32], approximately 65% compared to the maximal hydrolysis activity, obtained for mtHsp70 in the presence of Mge1 and yTim14. These results demonstrate that the J domain of Tim14/Pam18 is less active in stimulating mtHsp70’s ATPase activity when in complex with the soluble domain of Tim16/Pam16. This is in agreement with previous reports [19,32]. A similar effect was observed using human Tim16/Pam16s, which inhibited the increase in the ATPase activity of mtHsp70 (obtained due to the presence of yTim14/Pam18) by almost 50%. The latter result indicates that hTim16/Pam16s can replace its yeast Tim16/Pam18s homologue, in vitro, and can act as a negative regulator of yTim14/Pam18. Based on the ATP hydrolysis experiments, we conclude that the function of Tim/Pam proteins is conserved between yeast and humans.

Human Tim14/Pam18 and human Tim16/Pam16 were isolated by PCR from a human cDNA library and cloned as individual proteins or complexes into a modified version of the bacterial expression vector pET21d, in which a TEV protease site was inserted between the histidine tag and the N-terminus of the protein. The proteins used in this study are detailed in Table 2. The purification of the constructs 1–2 in Table 2 was carried out as described previously for the soluble domains (constructs 3–4) of yeast Tim14s/Pam18s and Tim16s/Pam16s [26]. Constructs 5–7 were purified as described previously for construct 8 [26]. Constructs 9–10 were purified carrying an octa-histidine tag on Ni-agarose following manufacture's protocol. In constructs 1–8, the hisitidine tag was removed from the final purified protein by proteolysis with TEV protease. In constructs 9–10, the hisitidine tag was not removed from the final purified protein.

All CD measurements were performed with an Aviv CD spectrometer, as described previously [26].

Cross-linking of the proteins was carried out at room temperature in 20 mM of Na-HEPES (pH 7.4), containing 100 mM of KCl, with 1 mM DSS. A protein concentration of 0.5 mg/mL was used. The cross-linking reaction was stopped at different times by the addition of SDS-containing sample buffer and boiling for 5 min. The cross-linking products (20 μL) were analyzed by 16% SDS-PAGE [26].

mtHsp70 was purified as described in [34]. ATPase assays were carried out as described in [33]. The concentrations of proteins were determined with the Bicinchoninic Acid Protein Assay (Sigma; Cat. no.B9643) using BSA as a standard [33].