|Home | About | Journals | Submit | Contact Us | Français|
Fission yeast protein Sre1, the homolog of the mammalian sterol regulatory element binding protein (SREBP), is a hypoxic transcription factor required for sterol homeostasis and low oxygen growth. Nro1 regulates the stability of the N-terminal transcription factor domain of Sre1 (Sre1N) by inhibiting the action of the prolyl 4-hydroxylase-like Ofd1 in an oxygen-dependent manner. The crystal structure of Nro1 determined at 2.2 Å resolution shows an all-α-helical fold that can be divided into two domains: a small N-terminal domain and a larger C-terminal HEAT-repeat domain. Follow-up studies showed that Nro1 defines a new class of nuclear import adaptor that functions both in Ofd1 nuclear localization and in the oxygen-dependent inhibition of Ofd1 to control the hypoxic response.
Oxygen is essential for many biological processes. Therefore, cells have evolved different mechanisms to respond to fluctuating oxygen concentrations and hypoxic stress. In both mammals and yeast, transcription factors activate specific genes that allow growth and survival under low oxygen conditions (Emerling and Chandel, 2005; Kwast et al., 1998). In the fission yeast Schizosaccharomyces pombe, the mammalian SREBP homolog Sre1 is the principal activator of hypoxic gene expression (Hughes et al., 2005; Todd et al., 2006). Sre1, like mammalian SREBP, is an endoplasmic reticulum, membrane-bound transcription factor that controls cellular sterol homeostasis (Espenshade and Hughes, 2007; Goldstein et al., 2006). Under hypoxia, Sre1 is proteolytically cleaved and the N-terminal transcription factor domain (Sre1N) enters the nucleus and up-regulates genes required for sterol synthesis and hypoxic growth (Espenshade and Hughes, 2007; Hughes et al., 2005). Oxygen also controls the stability of Sre1N, such that Sre1N accumulates in the absence of oxygen, but is rapidly degraded in the presence of oxygen (Hughes and Espenshade, 2008).
It has been postulated that prolyl 4-hydroxylase, 2-oxoglutarate (OG)-Fe(II) dioxygenases (PHDs) can serve as oxygen sensors in cells (Chowdhury et al., 2008). In mammalian cells, proteins of the PHD family regulate the stability of the hypoxia-inducible factor α (HIF-α) subunit by hydroxylating two proline residues. This modification targets HIF-α for proteasomal degradation (Dann and Bruick, 2005; Kaelin and Ratcliffe, 2008; Mole et al., 2001; Semenza, 2007). Ofd1 is a PHD-like protein that has been shown to accelerate Sre1N degradation in the presence of oxygen (Hughes and Espenshade, 2008). Interestingly, the Ofd1 homolog in S. cerevisiae Tpa1 has been structurally proposed as a prolyl hydroxylase (Kim et al., 2010). Like Tpa1, Ofd1 consists of two domains: an N-terminal dioxygenase domain and a C-terminal degradation domain (CTDD). While the N-terminal dioxygenase domain may act as an oxygen sensor, the non-catalytic Ofd1CTDD is necessary and sufficient to target Sre1N for degradation (Hughes and Espenshade, 2008). Recently a negative regulator of Ofd1, Nro1, was identified and characterized as an enhancer of Sre1N stability (Lee et al., 2009). In absence of oxygen, Nro1 binds Ofd1 and inhibits Ofd1CTDD, leading to Sre1N accumulation. In the presence of oxygen, the interaction between Nro1 and Ofd1 is disrupted leading to the rapid degradation of Sre1N (Lee et al., 2009).
To gain insight into the mechanism for inhibition of Ofd1 by Nro1, we determined the structure of Nro1 to 2.2 Å resolution using x-ray diffraction methods. The crystal structure shows that Nro1 is an all-helical protein formed by α-helical repeats with a highly degenerate sequence motif, characteristic of members of the HEAT-repeat family of proteins (Andrade et al., 2001). Following from the structure determination, we found that Nro1 defines a new class of nuclear import adaptor that functions both in Ofd1 nuclear localization and in the oxygen-dependent inhibition of Ofd1 to control the hypoxic response.
Nro1 crystals belong to the space group P21 and contain two weakly associated molecules per asymmetric unit. The hydrodynamic behavior in size exclusion chromatography of Nro1 suggests a monomer as the physiological unit. The observed electron density in the Nro1 crystal corresponds to residues 12 to 34 and 62 to 392 of monomer A and residues 13 to 33 and 59 to 393 of monomer B.
The structure of Nro1 refined at 2.2 Å resolution (Table 1) exhibits an all α-helical fold containing fifteen helices (Fig. 1A and Figure S1A). An N-terminal α-helix (α0) lies against a large C-terminal domain (CTD) formed by the remaining α-helices (α1–α14). The CTD α-helices are grouped in six helical hairpins, with a connecting helix (α4) splitting the second hairpin (α3–α5) and a short C-terminal helix (α14). The six tandemly arrayed α-helical hairpins of the CTD fold as a solenoidal tape, shaped like an arched, twisted tape. The twist between the solenoid turns decreases abruptly after the third hairpin, flattening the solenoidal tape and dividing the CTD into two sub-domains of three repeats each (aa 60–223 or α1– α7 and aa 224–393 or α8– α14). Several salt-bridges, Glu-144 to Arg-181, Glu-145 to Arg-186, Glu-129 to Arg-233 and Arg-214 to Asp-256, are positioned to stabilize the greater twist of the first three repeats. The concave-side of the CTD arched solenoidal tape has an opening of 27 Å at its narrowest point and 42 Å at its widest point. The concave and convex sides of the arch-shaped CTD exhibit different charge-distributions (Fig. 1B, 1C). The convex side is slightly hydrophobic and has an even charge distribution (Fig. 1C). The concave side shows a concentration of negative charges on a groove between the fourth and sixth repeats that contains Glu-295, Asp-299, Asp-302, Asp-344, Asp-348 and Asp-383 (Fig. 1B).
In the crystal asymmetric unit, only about one-third of the N-terminal region is observed in the crystal electron density. The α0 helices are spanned by residues 12 to 34 of monomer A and residues 13 to 33 of monomer B. Both helices in the asymmetric unit attach to the convex side of a CTD, making few apolar contacts with the α0 helix C-terminus (Fig. 1C). No residue conservation among Nro1 homologs is associated with these contacts. The acidic residue-rich region, spanning residues 35 to 61, is predicted to be intrinsically unstructured by prediction algorithms (He et al., 2009) and forms an unobserved linker between α0 and the CTD (Fig. 1A). Since the linker that connects α0 to the rest of the fold is not observed, whether each α0 helix interacts with the CTD of the same molecule or an adjacent monomer is unknown. Indeed, the unobserved 26-residue linker could span the distance between the helix α0 C-terminus and the CTD of multiple asymmetric units in the unit cell. The unobserved linker and the low occupancy of α0 (50%, Fig. S1B), which is consistent with its poor interactions with the rest of Nro1, support the assignment of the involved region as a potentially flexible domain. Based on the above-described characteristics, Nro1 can be divided into two domains: a flexible N-terminal domain (NTD; residues 1–59) and a large CTD (residues 60–393) with a HEAT-repeat fold.
The overall fold of Nro1 resembles all-helical repeat proteins, like those formed by TPR-, ARM- and HEAT-repeats (Groves and Barford, 1999). These repeat proteins contain a tandem array of helical units stringed to form a solenoid. The basic repetitive unit of the Nro1 CTD is a single α-helical hairpin without a strong sequence motif, a common characteristic of HEAT-repeats (Andrade et al., 2001; Conti et al.). Members of this family are known to mediate protein-protein interactions and include proteins involved in cytoplasmic-nuclear transport, such as the Karyopherins Importin-β (Imp-β) with HEAT-repeats and Importin-α (Imp-α) with the evolutionary related ARM-repeats (Cook et al., 2007). The Nro1 NTD has a basic region (4RRPQGLRAAASLKKQQ19; Fig. 2A) that resembles a bipartite canonical nuclear localization signal (NLS: KRPX10–12KKKK) (Dingwall and Laskey, 1991; Dingwall and Laskey, 1998; Hahn et al., 2008). The six repeats of Nro1 CTD span 78 Å, roughly 75% percent the length of Imp-α which contains nine ARM-repeats (Fig. 1D) and 33% the length of an Imp-β molecule, which contains 18–20 HEAT-repeats (Fig. 1E). The structure of both CTD sub-domains is highly similar to that of a triad of HEAT-repeats from Imp-β (less than 2.0 Å rms deviation between 79 Cα-atoms, Fig. 1E).
Nro1 is a homolog of S. cerevisiae Yor051c that has been reported to be a nuclear pore associated protein (Rout et al., 2000). However, Nro1 does not localize to the nuclear rim (Lee et al., 2009). Rather, both Ofd1 and Nro1 localize to the nucleus (Hughes and Espenshade, 2008; Lee et al., 2009). Nro1 and S. cerevisiae Yor051c belong to a family of yeast proteins that share similar sequence profiles. Overall, there is 18% sequence identity between S. pombe Nro1 and S. cerevisiae Yor051c. The highest conservation is observed in the first CTD subdomain (α1–α7) (Fig. 2A). Minor conservation is observed in the second subdomain (α8–α14) of the CTD and is restricted only to hydrophobic residues between the hairpin helices. Amino acids conserved between these two orthologs are concentrated in the first CTD subdomain in a patch on the convex side of the CTD (Fig. 2C). A more extensive sequence alignment of yeast Nro1 orthologs reveals amino acid conservation in regions corresponding to the NTD (including α0) and CTD HEAT repeats 1, 2, 3, 5 and 6 (Fig. S2). This conservation highlights a region that potentially interacts with other proteins.
The structural similarity between Nro1 and karyopherins and its homology with the nuclear pore associated protein Yor051c suggested a potential role for Nro1 in nuclear protein import. To examine this possibility, we tested whether Nro1 is required for nuclear localization of its binding partner Ofd1. We performed indirect immunofluorescence in sre1N, sre1N ofd1Δ and sre1N nro1Δ cells using an antibody against endogenous Ofd1. In sre1N cells, Ofd1 was enriched in the nucleus with little cytosolic staining consistent with previous findings (Fig. 3A)(Lee et al., 2009). In sre1N nro1Δ cells, Ofd1 showed diffuse cytosolic staining indicating that Nro1 is required for Ofd1 nuclear localization. Staining for Ofd1 was specific as ofd1Δ cells showed no signal. Together, these results show that nuclear localization of Ofd1 is linked to Nro1.
To test whether additional proteins are involved in Nro1-dependent nuclear shuttling, we identified Nro1-binding proteins by affinity purification. Using an affinity purified anti-Nro1 antibody, we analyzed immunopurified Nro1 protein complexes from wild-type and nro1Δ cells by SDS-PAGE. Nro1-specific binding proteins were analyzed by mass spectrometry. A β-type Karyopherin, Kap123 (SPBC14F5.03c) also known as Imp-β4, was identified as a Nro1-binding protein (see Table S1) suggesting that this Imp-β may be involved in Nro1-dependent Ofd1 nuclear localization. To examine the role of Kap123 in the nuclear localization of Ofd1, we assayed Ofd1 localization in sre1N kap123Δ cells. Interestingly, nuclear localization of Ofd1 in kap123Δ cells was only partially abrogated compared to the complete lack of localization in nro1Δ cells (Fig. 3A, 3B). The residual Ofd1 nuclear localization may be due to the existence of additional Karyopherin-βs whose function overlap with Kap123. If Kap123 is required for Nro1-mediated Ofd1 nuclear import, we would expect to see a defect in Nro1 localization in kap123Δ cells. To test whether Kap123 is required for Nro1 nuclear transport, we examined the localization of Nro1 in sre1N, sre1N nro1Δ and sre1N kap123Δ cells. In sre1N cells, Nro1 was enriched in the nucleus with some cytosolic staining (Fig. 3C). The signal for Nro1 was specific as no signal was detected in sre1N nro1Δ cells. Indeed, Nro1 was partially mislocalized in sre1N kap123Δ cells indicating that Kap123 is also required for Nro1 nuclear localization. Nro1 localized normally in ofd1Δ cells (data not shown). Collectively, these data suggest that Kap123 interacts with Nro1 to shuttle the Nro1-Ofd1 complex into the nucleus.
Previous studies showed that Nro1 functions as a positive regulator of Sre1N by inhibiting the activity of Ofd1 in Sre1N degradation (Lee et al., 2009). In sre1N nro1Δ cells, Sre1N failed to accumulate under low oxygen due to constitutively active Ofd1. To determine whether Ofd1 nuclear localization affects Sre1N degradation, we engineered the sre1N NLS-ofd1 strain that expresses Ofd1 fused to a N-terminal nuclear localization signal (NLS-Ofd1). NLS-Ofd1 localized to the nucleus in nro1Δ cells (Fig. 4A), allowing us to assay the function of Nro1 in Sre1N degradation independent from Nro1’s function in Ofd1 nuclear localization. sre1N, sre1N ofd1Δ, sre1N nro1Δ and sre1N NLS-ofd1 nro1Δ cells were cultured in the absence of oxygen for increasing time and Sre1N levels were assayed by western blot (Fig. 4B). As previously reported, Sre1N accumulated in sre1N cells under anaerobic conditions (Fig. 4B, lanes 1–3)(Lee et al., 2009). In ofd1Δ cells, Sre1N accumulated in the presence and absence of oxygen (Fig. 4B, lanes 4–6). In nro1Δ cells with Ofd1 retained in the cytosol due to the loss of Nro1-mediated nuclear localization, Sre1N failed to accumulate in anaerobic conditions (Fig. 4B, lanes 7–9). Importantly, Sre1N also failed to accumulate in NLS-ofd1 nro1Δ cells in the absence of oxygen despite the restored Ofd1 nuclear localization (Fig. 4B, lanes 10–12). Taken together, these data support previous studies and demonstrate that Nro1 functions as a direct inhibitor of Ofd1 in Sre1N turnover.
HEAT-repeat proteins are intrinsically flexible and use different regions to interact with their binding proteins (Conti et al., 2006). To identify which region(s) of Nro1 interacts with Ofd1, we performed mutational analysis. First we targeted the acidic patch on the concave side of Nro1 formed by conserved residues (Fig. 1B and and2C).2C). Asp 295 was an exception and is not conserved (Fig. 2A). Two types of amino acid substitutions were chosen that satisfy different criteria: (1) substitutions that eliminate the negative charge (E ->Q, D ->N), and (2) substitutions that in addition remove hydrogen-bonding groups (E, D -> A). Single and multiple substitutions were introduced into Nro1 by site-directed mutagenesis. The effects of mutant Nro1 on Ofd1 binding, Ofd1 localization, and Sre1N accumulation were assayed (Fig. 5, Figure S3).
Table 2 summarizes the results. Single mutations that eliminated negative charge alone had no effect on Nro1 function (Table 2). Changes to the Nro1 CTD that removed H-bond forming groups weakened the interaction between Nro1 and Ofd1, but had little effect on Ofd1 localization or Sre1N regulation (Table 2 and Figure S3). Indeed, the Nro1(Y120H,D383A) mutant that includes a serendipitous mutation Y120H showed a dramatic decrease in Ofd1 binding (Fig. 5A, lanes 3–4), yet this resulted in only a minor effect on Sre1N regulation (Fig. 5B, compare lanes 5–6 to lanes 9–10). Interestingly, this Nro1 mutant also showed a partial defect in Ofd1 nuclear localization (Fig. 5C). The Y120H mutation is in a conserved region suggesting a role in protein-protein interaction (Fig. 2A and 2C). Consistent with this, the Y120H mutation enhances the negative effects of the D383A mutation on the Nro1-Ofd1 interaction (Table 2 and Fig. 5).
The effects of deleting the entire NTD or only its helix α0 --Nro1(Δ10–29)-- were evaluated using similar techniques. Coimmunoprecipitation experiments detected no binding between Ofd1 and these Nro1 mutants (Fig. 5A, lanes 3 and 6; Figure S3), indicating that the region containing most of α0 is required for Ofd1 binding. Furthermore, Sre1N protein failed to accumulate and Ofd1 was mislocalized to the cytosol in sre1N nro1Δ cells expressing Nro1(Δ10–29) (Fig. 5B, lanes 13–14 and Fig. 5C). Collectively, these studies showed that while mutation of residues on the acidic patch of the concave surface of the CTD weaken Ofd1 binding, only deletion of α0 or the NTD completely abrogates all known Nro1 functions.
Our mutational studies suggest that the NTD α0 is required for Ofd1 interaction, Ofd1 nuclear localization, and hypoxic Nro1-mediated inhibition of Ofd1. To evaluate the binding and better define the region of interaction between Nro1 and Ofd1, we carried out biophysical experiments using purified proteins including Ofd1CTDD, full-length Nro1 and two truncation mutants of Nro1: Nro1(Δ1–30) that removes α0 but retains the disordered linker, and the Nro1 CTD alone. Isothermal titration calorimetry (ITC) experiments on complex formation between Ofd1CTDD and either Nro1, Nro1(Δ1–30) or the CTD were performed. Analysis of the ITC data shows that Nro1 and Ofd1CTDD form a 1:1 complex with a dissociation constant (KD) of 4.0 ± 0.2 µM, a ΔH of −11.35 kcal/mol, and an unfavourable ΔS of −13.0 cal/(mol K) at 301 K (Fig. 6A). Heat of complex formation with Ofd1CTDD was not detected for either Nro1(Δ1–30) or the CTD at the maximal experimental reagent concentrations (up to 50 µM of Ofd1CTDD, data not shown).
To rule out the possibility of an entropically-driven complex formation in the case of the truncation mutants, size exclusion chromatography experiments were also performed. These experiments showed that while a near equimolar mixture Ofd1CTDD and intact Nro1 migrated as a heterodimer (Fig. 6B), mixtures containing Ofd1CTDD with either Nro1(Δ1–30) or the CTD migrated as single, monomeric species (data not shown). These in vitro data indicate that Nro1 and Ofd1CTDD form a heterodimer in solution and show that the first thirty residues of Nro1 interact directly with the Ofd1CTDD and are required for its binding.
To examine further the role of each Nro1 domain in Ofd1 binding, we performed yeast two-hybrid experiments with Ofd1CTDD and different domains of Nro1. Cells co-transformed with Ofd1CTDD and full-length Nro1, NTD or CTD plasmids were plated on non-selective or selective media. Cells expressing both Ofd1CTDD and either full-length Nro1 or NTD grew well on selective medium suggesting a strong interaction between the two proteins (Fig. 6C). However, cells containing the Nro1 CTD plasmid showed slow growth on selection medium, suggesting a weak interaction between the CTD alone and the Ofd1CTDD (Fig. 6C, lower right). Next, we tested whether the interaction between NTD and Ofd1CTDD is sufficient for Nro1-mediated Ofd1 inhibition in Sre1N degradation. Extracts from sre1N nro1Δ cells carrying an empty vector or a plasmid expressing GFP or GFP-NTD were analyzed by western blot using anti-Sre1N antibody. GFP-NTD, but not GFP alone, stabilized Sre1N (Fig. 6D, lanes 3–5), indicating that the NTD is sufficient to inhibit Ofd1 activity in Sre1N degradation. Collectively, these data are consistent with the hypothesis that the NTD is necessary and sufficient for Ofd1 binding and Nro1-mediated Ofd1 inhibition in Sre1N degradation.
Macromolecules require nuclear import receptors for transport across the nuclear membrane pore (Conti et al., 2006; Pemberton and Paschal, 2005). The Karyopherins Importin-α and -β mediate a well-studied nuclear import system in eukaryotic cells (Catimel et al., 2001; Cingolani et al., 1999; Conti et al., 1998). Imp-β interacts with nuclear localization signals in either cargo proteins or the adaptor Imp-α which in turn binds cargo (Goldfarb et al., 2004). Imp-β interacts with nucleoporins at the nuclear pore complex to enable the selective transport of macromolecules across the nuclear membrane (Cook et al., 2007). The adaptor protein Imp-α exhibits a two-domain architecture: an N-terminal Imp-β binding (IBB) domain connected by a linker to an ARM-repeat C-terminal domain. The IBB includes basic-residue rich regions that resemble a bipartite NLS that is recognized by Imp-β (Dingwall and Laskey, 1991)(Cingolani et al., 1999). To regulate protein shuttling, the concave side of the Imp-α C-terminal domain can recognize either the cargo protein or its own IBB in an auto-inhibitory fashion (Conti et al., 1998). This auto-inhibition explains the Imp-α switch between the cytosolic high-affinity for cargo and the nuclear low-affinity that results in directional transport (Fanara et al., 2000; Kobe, 1999; Kobe and Kemp, 1999).
Here, we report that like Imp-α Nro1 has a two-domain architecture with a flexible NTD and all α-helical repeat CTD. Functional studies reveal that Nro1 defines a new class of helical repeat, nuclear import adaptor. Several lines of evidence support this conclusion. First, Nro1 is required for nuclear localization of Ofd1 (Fig. 3). Second, Nro1 binds the Ofd1 cargo molecule through its NTD, not its C-terminal helical-repeat domain like Imp-α (Fig. 6). Third, the Nro1 NTD does not possess the auto-inhibitory property of the Imp-α IBB domain, since removal of the Nro1 NTD α0 (Δ10–29) blocks, rather than increases, cargo binding (Fig. 5)(Kobe, 1999). Fourth, both Nro1 and Ofd1 require the Imp-β family member Kap123 for nuclear import (Fig. 3), suggesting that Nro1 couples Ofd1 to a Karyopherin β for transport.
Nro1 functions as an adaptor and is structurally similar to both Imp-α, an ARM-repeat adaptor that imports NLS-containing cargo to the nucleus (Conti et al., 1998), and VPS35, a HEAT-repeat adaptor in the retromer cargo-recognition complex that sorts acid hydrolases to the lysosome (Hierro et al., 2007). These two proteins use their α-helical-repeat domains to recognize cargo, although in different manners. Imp-α recognizes only the cargo NLS peptide in a cleft on the concave side of its C-terminal domain (Catimel et al., 2001). VPS35 binds by wrapping itself around the cargo protein with most of the concave surface participating in the recognition (Hierro et al., 2007). Unlike these adaptor proteins, the Nro1 CTD is relatively less important for recognition of Ofd1 (Fig. 5 and Table 2).
In contrast, our data indicate a major role for the Nro1 NTD helix α0 in binding Ofd1 and a less important role for the concave surface of the Nro1 CTD (Figs. 5, ,66 and Table 2). In vitro experiments using Nro1, Nro1(Δ1–30), and the C-terminal degradation domain of Ofd1 indicate a direct interaction between the Ofd1CTDD and the first thirty residues of Nro1. Thermodynamic data obtained from Nro1-Ofd1CTDD complex formation showed an unfavourable entropic contribution to the binding (ΔS = −13 kcal mol−1 K−1), consistent with the notion that complex formation induces order in an otherwise disordered domain like the NTD. The removal of the Nro1 regions that include α0 abrogates Nro1 binding to Ofd1CTDD. Furthermore, we show by yeast two-hybrid that the NTD interacts with Ofd1CTDD, and GFP-NTD is sufficient to inhibit the effect of endogenous Ofd1 in Sre1N degradation (Figures 6C, D). Compared to large deletions in the NTD, select amino acid substitutions in the concave side of the CTD (Fig. 2) showed relatively minor negative effects on binding and nuclear localization of Ofd1 (Fig. 5 and Table 2). These mild effects may be explained by the small footprint of the selected substitutions within what could be a large area of recognition if Nro1 employs a VPS35-like mode of recognition.
Nro1 regulates the fission yeast hypoxic response by inhibiting Ofd1 function in Sre1N degradation. Nro1’s newly discovered role as a nuclear import adaptor raised the question of whether Nro1 simply regulates Sre1N by controlling the localization of Ofd1. For example, in nro1Δ cells Sre1N may be rapidly degraded due to the separation of Ofd1 from a hypothetical inhibitor localized in the nucleus. However, fusing a nuclear localization sequence to Ofd1 restored its nuclear localization, but had no impact on Sre1N stability (Fig. 4). These results indicate that the rapid degradation of Sre1N is not a consequence of Ofd1 localization, but rather due to the ability of Nro1 to bind and directly inhibit Ofd1. Thus, we propose that Nro1 has two different functions that require binding to Ofd1: (1) Nro1 is required for Ofd1 nuclear localization, and (2) Nro1 binding is required to inhibit Ofd1-mediated degradation of Sre1N.
Taken together these results suggest a mechanism for the Nro1-dependent regulation of Sre1N (Fig. 7). In the cytosol, Nro1 and Ofd1 form a heterodimer that is shuttled into the nucleus by a pathway that requires Kap123. Under low oxygen, the Nro1-Ofd1 heterodimer is more stable, which prevents Ofd1CTDD interaction with Sre1N (Lee et al., 2009). Under normoxic conditions, the interaction between Nro1 and Ofd1 is disrupted. Ofd1 is free to interact with Sre1N accelerating its turnover by the proteasome. One hypothesis for the oxygen-dependence of this mechanism is that in the presence of oxygen the enzymatic activity of the Ofd1 dioxygenase domain may modify Ofd1CTDD, Nro1, or both thereby interfering with the Nro1-Ofd1 heterodimer stability. The Ofd1 dioxygenase domain is presumably also active in the cytosol. However, nuclear localization of Ofd1 is not affected by inhibition of the Ofd1 dioxygenase domain (data not shown). Perhaps Kap123 further stabilizes the Nro1-Ofd1 complex, thus allowing nuclear import in the presence of oxygen (Fig. 7).
The work presented here shows that Nro1 defines a new class of nuclear import adaptor that is required for the nuclear transport of Ofd1. Understanding how Nro1 itself enters the nucleus, the role of Kap123 in this process, and what other cargo require Nro1 are important questions for future studies. Nro1 has a second function as a direct inhibitor of Ofd1CTDD in the regulation of hypoxic gene expression mediated by Sre1. Regulation of this hypoxic response triggered by oxygen is completely dependent on Nro1-Ofd1 complex dynamics. Elucidation of the molecular details of this interaction is the focus of current studies.
Yeast extract was obtained from Becton Dickinson and Co.; amino acids for medium from Sigma; Edinburgh minimal medium from MP Biomedical; SD base and dropout supplements from Clontech; horseradish peroxidase-conjugated affinity-purified donkey anti-rabbit and anti-mouse immunoglobulin G (IgG) from Jackson ImmunoResearch; oligonucleotides from Integrated DNA Technologies; dimethyloxalylglycine (DMOG) from Frontier Scientific, Inc.; and dithiobissuccinimidyl propionate (DSP) crosslinker from Pierce; SOURCE-Q resin, HisTrap™ HP, Superose 12 HR 10/30 and GL columns, and, the ÄKTApurifier™ instrument from GE Healthcare.
The S. pombe gene coding for Nro1 (SPCC4B3.07) was cloned into the BamHI/XhoI restriction sites of the pPROEX HTb (Gibco BRL) expression vector. The N-terminal 6xHis-tag Nro1 fusion protein was expressed in E. coli BL21(DE3) cells. Protein expression was induced with 0.6 mM IPTG in E. coli cultures and grown for 4 hours at 30°C. Cells were harvested and lysed in buffer containing 20 mM HEPES pH 8.0, 150 mM NaCl, 5 mM 2-mercaptoethanol (βME), 1 mM MgCl2, 10 µg/ml DNAse, 10 µg/ml RNase and 1 mM PMSF. The 6xHis-tagged Nro1 was purified by nickel affinity chromatography. The 6xHis-tag was removed by treatment with TEV protease. Subsequent anion exchange chromatography using SOURCE-Q resin and an additional step of nickel affinity chromatography produced 98% pure protein as judged by Coomassie Blue stained SDS-PAGE. Truncated constructs of Nro1, Nro1(Δ1–30) and the CTD, were cloned, expressed and purified using the same protocols as the unmodified protein.
The S. pombe gene coding for Ofd1CTDD (SPBC6B1.08c, residues 255 to 515) was cloned into the EcoRI/NotI restriction sites of the pPROEX HTb (Gibco BRL) expression vector. The N-terminal 6xHis-tag Ofd1 fusion protein was expressed in E. coli BL21(DE3) cells. Protein expression and purification was carried out as described above.
Proteins samples of 17 mg/ml Nro1 in 20 mM HEPES pH 7.5, 5 mM βME and 60 mM NaCl were used for crystallization. Initial screening of crystallization conditions of Nro1 were sought by vapour diffusion using a Mosquito® (TTP LabTech) crystallization robot using commercial crystallization buffers. Optimization of initial hits was performed in 24-well plates. Crystals suitable for data collection take between 1 and 2 weeks to grow in hanging drops containing 1 µl of 17 mg/ml Nro1 and 1 µl of well solution and equilibrated against 1 ml of reservoir containing 1.3 M of ammonium tartrate dibasic and 0.1 M Bis-Tris HCl pH 6.5, at 291 K.
For data collection single crystals of Nro1 were mounted in a loop and flash-frozen in liquid N2. Native data were collected at 100 K at the section 31 beamline of the Advanced Photon Source using radiation at a wavelength of 0.979 Å. Nro1 crystals were soaked in mother liquor solutions including 100 µM of samarium acetate (overnight) or with 100 µM di-μ-iodo-bis(ethylene-diamine)-di-Pt(II)-nitrate [PIP] (two days). Diffraction data of derivative crystals were collected in house using a FR-E SuperBright® x-ray generator equipped with VariMAX mirrors and a Saturn 944+ CCD detector (Rigaku Inc.). The native data were reduced with XDS (Kabsch, 1993) and merged using SCALA (CCP4, 1994). The derivative data were processed with the program package HKL2000 (HKL Research Inc.).
Phases were determined using Multiple Isomorphous Replacement including anomalous scattering (MIRAS) method. Experimental phases up to 4 Å resolution with a figure of merit of 0.47 were obtained with the program SOLVE using these two derivatives (Terwilliger and Berendzen, 1999). The non-crystallographic symmetry operator was identified in the experimental electron density and used to further improve the quality of the map by density averaging and phase extension to 2.2 Å with the program DM (CCP4, 1994). The atomic model was built in the improved density, using the program O (Jones et al., 1991) and refined using reflections to 2.2 Å with the program REFMAC (CCP4, 1994) (Table 1).
Both Nro1 and Ofd1CTDD were shown to be physiological monomers in solution by analytical size-exclusion chromatography using a Superose 12 HR 10/30 column. For the Nro1/Ofd1CTDD complex determination, approximately 1:1.2 ratio of Nro1/Ofd1CTDD were pre-mixed and incubated for 20 minutes on ice before loading on the Superose-12 GL column. Chromatography experiments were performed on a FPLC instrument in 20 mM HEPES pH 7.5, 150 mM NaCl, 5 mM βME with a flow rate of 0.5 ml/min.
All protein samples were dialyzed against 20 mM HEPES pH 7.5, 200 mM NaCl, 5 mM βME buffer and degassed at 25°C for 20 minutes before loading to the VP-isothermal titration calorimetry (MicroCal). 217 µM of Ofd1CTDD was loaded in the syringe and 13.8 µM of Nro1 in the cell. 10 µl of Ofd1CTDD were injected into the cell containing buffer, Nro1 or Nro1-deletion protein samples, ~25 times with a 300 seconds equilibration interval between injections. All the experiments were carried at 28°C. Data were analyzed with the program Origin (version 5.0) and fitted using a one-site model.
The procedure for indirect immunofluorescence was performed as previously described (Hagan and Hyams, 1988; Hughes and Espenshade, 2008; Lee et al., 2009). Cells were imaged after washing in PBS and mounting to cover slips on a bed of 2% agarose. Affinity purified anti-Nro1 antibody for immunofluorescence was prepared following a protocol from http://ygac.med.yale.edu/mtn/antibodycleanup.stm. Briefly, affinity purified anti-Nro1 antibody was incubated with 2×1010 spheroplasted nro1Δ cells for 8 hours to remove immunoglobulins with nonspecific interactions. Supernatant was collected and a 1:100 dilution was used for Nro1 staining. Cell images were taken using IVISION-Mac V.4.07 software. All images were normalized to control cells to minimize non-specific signal and then processed uniformly using Adobe Photoshop to optimize image contrast. Whole cell yeast extract preparation and western blot analysis using anti-Sre1 IgG and horseradish peroxidase-conjugated anti-rabbit IgG were performed as described previously (Hughes and Espenshade, 2008). Anti-Nro1 serum was used at 1:10,000 dilution.
Cells (5×107) were collected and washed with PBS. To crosslink protein complexes, cells were incubated with 2 mM DSP in PBS for 10 minutes after which DSP was quenched for 15 minutes by addition of 1 M Tris-HCl, pH 7.5 to a final concentration of 20 mM. Cells were lysed using glass beads (0.5 mm, Sigma) in 100 µl NP40 Lysis Buffer (50 mM Hepes, pH 7.4, 100 mM NaCl, 1.5 mM MgCl2, 1%(v/v) NP-40) plus protease inhibitors (Hughes et al., 2005). Insoluble material was removed by centrifugation at 2×104 g for 2 minutes, and the supernatant was subjected to immunoprecipitation in 1 ml NP40 Lysis Buffer for 2 hours using 5 µg of affinity purified anti-Ofd1 IgG and 30 µl Protein-A beads (Repligen). Beads were washed 3 times with NP40 Lysis Buffer, and coimmunopurified proteins were eluted by boiling in SDS-PAGE loading buffer (Hughes et al., 2005)
pCaMV-Nro1, encoding fission yeast nro1+ driven by the cauliflower mosaic virus (CaMV) promoter, was created using a nro1+ PCR fragment and inserted into the BamHI/PstI sites of pSLF101, which carries a leu2+ auxotrophic marker (Forsburg, 1993). The nro1+ point mutations were generated using the QuikChange multi site-directed mutagenesis kit (Stratagene) according to manufacturer’s instructions. Plasmids containing GFP and the NTD of Nro1 expressed from the thiamine-repressible nmt* promoter were generated by insertion of GFP and nro1+ PCR fragments into pREP41X (Forsburg, 1993). PCR products of full length, NTD or CTD of Nro1 were cloned into pGADT7, and the C-terminus of Ofd1 was cloned into pGBKT7 from the Matchmaker Gal4 two-hybrid system (Clontech).
S. pombe strains sre1N, sre1N ofd1Δ, and sre1N nro1Δ have been described previously (Lee et al., 2009). Strain NLS-ofd1 had a NLS sequence (MAPKKKRKV) in front of ofd1 ATG at endogenous locus. To engineer the NLS-ofd1 strain, endogenous ofd1 was first replaced with ura4+ by homologous recombination (Bahler et al., 1998). This strain was subsequently transformed with PCR products of NLS-ofd1 with ~600 bp flanking sequence of endogenous ofd1 ORF. Transformants were selected on minimal medium containing 5-FOA to counter-select for expression of the ura4 gene product (Boeke et al., 1987). Transformants were screened by PCR and confirmed by sequencing. Two-hybrid selection was performed following Matchmaker Gal4 Two-Hybrid System® user manual protocols (Clontech) using the yeast two-hybrid S. cerevisiae strain AH109 (MATα trp1-901 leu2-3,112 ura3-52 his3-200 gal4Δ gal80Δ LYS2GAL1UAS-GAL1TATA-HIS3 GAL2UAS-GAL2TATA-ADE2 URA3MEL1UAS-MELTATA-lacZ). S. pombe strains were grown to exponential phase at 30°C in yeast extract plus supplements (YES: 225 µg/ml each of histidine, leucine, adenine, lysine, and uracil) or in Edinburgh minimal medium where indicated, using standard techniques (Hughes et al., 2005). Anaerobic growth conditions were maintained using an In vivo2 400 workstation (Biotrace, Inc.) as described previously (Hughes et al., 2005; Todd et al., 2006).
We acknowledge the use of beamlines X4A, X4C and X6A at the Brookhaven National Laboratory and the use of the SGX Collaborative Access Team (SGX-CAT) beamline facilities at Sector 31 of the Advanced Photon Source at Argonne National Laboratory that was provided by Eli Lilly & Company who operates the facility, supported by the U. S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357. Thanks to the Taplin Biological Mass Spectrometry Facility (Harvard Medical School) for mass spectrometry data and to Tony Hunter (Salk Institute) for strains and plasmids. This work was supported by NIH grants: HL-077588 (PJE) and ARRA-1RO1NS061827 (LMA), in addition to an Established Investigator Award from the American Heart Association to PJE and American Heart Association Predoctoral Fellowships to C-YSL.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
The Nro1 coordinates and structure factors were deposited in the Protein Data Bank accession code 3MSV.
Supplemental information includes three figures and can be found with this article online.