The nuclear localization of SWI/SNF proteins is subjected to oxygen regulation

Background Hypoxia is associated with many disease conditions in humans, such as cancer, stroke and traumatic injuries. Hypoxia elicits broad molecular and cellular changes in diverse eukaryotes. Our recent studies suggest that one likely mechanism mediating such broad changes is through changes in the cellular localization of important regulatory proteins. Particularly, we have found that over 120 nuclear proteins with important functions ranging from transcriptional regulation to RNA processing exhibit altered cellular locations under hypoxia. In this report, we describe further experiments to identify and evaluate the role of nuclear protein relocalization in mediating hypoxia responses in yeast. Results To identify regulatory proteins that play a causal role in mediating hypoxia responses, we characterized the time courses of relocalization of hypoxia-altered nuclear proteins in response to hypoxia and reoxygenation. We found that 17 nuclear proteins relocalized in a significantly shorter time period in response to both hypoxia and reoxygenation. Particularly, several components of the SWI/SNF complex were fast responders, and analysis of gene expression data show that many targets of the SWI/SNF proteins are oxygen regulated. Furthermore, confocal fluorescent live cell imaging showed that over 95% of hypoxia-altered SWI/SNF proteins accumulated in the cytosol in hypoxic cells, while over 95% of the proteins were nuclear in normoxic cells, as expected. Conclusions SWI/SNF proteins relocalize in response to hypoxia and reoxygenation in a quick manner, and their relocalization likely accounts for, in part or in whole, oxygen regulation of many SWI/SNF target genes.


Background
Living organisms ranging from yeast to mammals use oxygen to generate their cellular energy supply and to synthesize important biomolecules. Hence, they need to respond effectively to changes in oxygen levels in the environment, particularly to hypoxia [1,2]. In humans, hypoxia is responsible for death or damage by the ischemia accompanying heart attack, stroke, and traumatic injuries [3][4][5]. The molecular and cellular events induced by changes in oxygen levels are very broad in eukaryotes. For example, over 20% of yeast genes change their transcript levels in response to hypoxia [6]. In the human arterial endothelial cells, more than 8% of all genes alter their transcript levels by at least 1.5-fold in response to hypoxia [7]. In the human primary astrocytes, more than 5% of the genes alter their transcript levels by at least 2-fold in response to hypoxia [8]. Such broad changes in gene expression likely involve coordinated actions of multiple pathways and regulators.
Previous studies have identified several transcriptional regulators, including Mga1 and Rox1, that can mediate oxygen regulation of gene expression in the yeast Saccharomyces cerevisiae [9,10]. However, these regulators can account for the regulation of only a fraction of hypoxia-regulated genes [6]. Many other regulators are likely involved in mediating oxygen regulation. Recently, in an effort to systematically identify proteins that can mediate oxygen regulation and signaling, we performed a genomewide screen for proteins that exhibit altered cellular distribution patterns in response to hypoxia and reoxygenation [11]. We found that over 200 proteins alter their cellular locations in response to hypoxia. Particularly, under hypoxia, a good number (at least 121) of nuclear proteins do not localize to the nucleus, but accumulate in the cytosol.
In response to reoxygenation, they readily localize to the nucleus. Notably, many of these hypoxia-redistributed nuclear proteins are subunits of key regulatory complexes involved in chromatin remodeling (such as the SWI/SNF complex) [12][13][14], in transcriptional regulation (such as the SAGA complex) [15], and in splicing (such as the MRP complex) [16]. Hence, it is conceivable that some of these complexes can play a dominant role in mediating oxygen regulation of gene expression.
To further assess the roles of these regulators in mediating oxygen signaling and regulation, we examined the time course characteristics of the relocalization of these proteins in response to hypoxia and reoxygenation. We found a small group of nuclear proteins relocalized in a significantly shorter time period in response to both hypoxia and reoxygenation, when compared to other proteins. These proteins include three components of the SWI/SNF complex. Furthermore, using confocal fluorescent imaging of live cells, we quantitatively characterized the effect of hypoxia on the distribution of SWI/SNF proteins. We found that in live hypoxic cells, over 95% of Swi3, Snf5, Snf6, Snf11, Snf12 and Swp82 were in the cytosol, while over 95% of hypoxiaunaffected proteins, such as Swi2 and Taf14, were in the nucleus. These results suggest that hypoxia can significantly alter the composition and property of the SWI/SNF complex and mediate oxygen regulation of gene expression.

Results
Among the hypoxia-redistributed nuclear proteins we previously identified, some are likely involved in mediating oxygen signaling and regulation of gene expression. Particularly, proteins that change their locations in relatively shorter time periods are likely the regulators that initiate further downstream events in responses to hypoxia and reoxygenation. In other words, they are likely to be positioned in the upstream of the hierarchy of the molecular events elicited by hypoxia or reoxygenation, and are responsible for initiating downstream changes such as those in gene expression. We therefore decided to characterize the time course response of the hypoxia-redistributed nuclear proteins in response to hypoxia and reoxygenation.
First, we examined the time course characteristics of nuclear proteins in response to hypoxia. We found that all proteins became predominantly cytosolic after exposure to hypoxia for 12 hours; see Snf11 in Figure 1A for an example. One group of these proteins became predominantly cytosolic after only 6 hours or shorter times; see Swp82 in Figure 1A for an example. This group has 48 proteins (see Table 1). They include several transcriptional regulators and regulators of chromatin, DNA replication and repair, and RNA processing ( Figure 2). Notably, five components of the SWI/SNF complex relocalized in 6 hours (see Table 1 and Figure 2), suggesting that they may have a signaling role in initiating downstream events.
Second, we characterized the changes in protein distribution when cells grown under hypoxia were exposed to oxygen. We found that 76 hypoxia-redistributed nuclear proteins (see Table 2) recovered their nuclear locations in the majority of cells in one hour; see Swi3 in Figure 1B for an example. The rest of the proteins recovered their nuclear location in the majority of the cells in 2 or more hours; see Snf5 in Figure 1B for an example. Among these nuclear proteins, 17 of them responded to both hypoxia and reoxygenation in shorter times than the rest of the proteins (see Figure 3). Notably, 3 of these faster responding proteins are components  of the SWI/SNF complex ( Figure 3). These and previous results strongly suggest that the SWI/SNF proteins play regulatory roles in mediating oxygen regulation and hypoxia response. Given their roles in chromatin remodeling and transcriptional regulation [17,18], they are likely responsible for initiating certain changes in gene expression in response to changes in oxygen levels. Although 14 other proteins also responded to hypoxia and reoxygenation in shorter times, they are generally not components of one regulatory complex ( Figure 3). Therefore, we decided to further characterize the effect of hypoxia on the SWI/SNF proteins. First, we examined if changes in oxygen levels affect the protein levels of SWI/ SNF proteins. To this end, we detected and compared the levels of SWI/SNF proteins in hypoxic and normoxic cells. We used yeast strains expressing the SWI/SNF proteins with the TAP tag at the C-terminus from the natural chromosomal locations [19]. We found that the levels of all detected SWI/SNF proteins were not significantly affected by hypoxia ( Figure 4). The variations in the ratios of protein levels in hypoxic vs. normoxic cells were generally less than 30%, suggesting that hypoxia did not cause significant degradation of the Swi/Snf proteins during the time period when the proteins would be relocalized to the cytosol. These proteins include those whose cellular location was affected by hypoxia, such as Snf6, Swi3, Swp82 and Snf11 (see Figure 4). They also include all those SWI/SNF proteins whose localization was not affected by hypoxia. These results show that the levels of SWI/SNF proteins are not regulated by oxygen levels.
Therefore, the regulation of nuclear localization is likely the dominant mechanism mediating oxygen regulation of SWI/SNF proteins and the regulation of their targets. To further confirm the regulation of nuclear localization of the SWI/SNF proteins by oxygen, we quantitatively examined and compared their distribution in live hypoxic and normoxic cells, by using confocal fluorescent live cell imaging. As expected, for the SWI/SNF proteins whose localization was not affected by oxygen levels, over 95% of the proteins was present in the nucleus  Figure 2 Graphical representation of protein-protein interaction networks for the nuclear proteins that localized to the cytosol in response to hypoxia in a shorter time period. The information on the biochemical interactions and complex formation of the 48 faster responding nuclear proteins (listed in Table 1) was downloaded from the SGD database, and then imported to Cytoscape for network construction. The proteins are shown as round nodes in different colors based on their cellular functions. The GO terms for protein complexes or functional categorizations are indicated and are shown in square nodes. Nodes of the same sub-networks are colored similarly, and a key for the coloring of the nodes is shown. Lines represent an association of the protein to a particular complex or functional GO term. in both normoxic and hypoxic cells (see Figure 5). Figure 5A shows the distribution of Taf14 in air and under hypoxia, while Figure 5B shows the distribution of Swi2. For those proteins whose localization was affected by oxygen, over 95% of the proteins was present in the nucleus in air, whereas over 95% of the proteins was present in the cytosol under hypoxia (Figures 6 and 7). Figure 6A shows the distribution of Swi3 in normoxic cells, while Figure 6B shows the distribution of Swi3 in hypoxic cells. We also quantified the distribution of other hypoxia-relocalized SWI/SNF proteins (Figure 7). Figure 7A-E show the distribution of Snf5, Snf6, Snf11, Snf12 and Swp82 in hypoxic cells (The images for normoxic cells invariably showed nuclear localization, as expected and as shown in Figures 5   and 6, and are therefore omitted). Clearly, Swi2, Snf5, Snf6, Snf11, Snf12 and Swp82 proteins were transported to the nucleus in normoxic cells, but they accumulated in the cytosol in hypoxic cells.
To further ascertain the role of SWI/SNF proteins in oxygen regulation of gene expression, we determined if and how many oxygen-regulated genes are SWI/SNF protein targets as well. To this end, we used our previous microarray and computational work analyzing genes regulated by oxygen/ and Δ hap1 cells [6]. We also used the previously identified targets of 263 transcription factors [20]. Using these two sets of data and the R program, we identified those hypoxia-regulated genes that are targets of SWI/SNF proteins and calculated the p- values. Table 3 shows that in the wild type HAP1 cells, 95, 112, 67, 109, 9 and 19 target genes of Swi2, Swi3, Snf5, Snf6, Snf11 and Taf14, respectively, are oxygen regulated. In Δ hap1 cells, similar numbers of these SWI/SNF targets are hypoxia altered. These results strongly suggest that SWI/SNF proteins play a major role in mediating oxygen regulation and hypoxia responses. Furthermore, the changes in the relocalization of SWI/SNF proteins in response to hypoxia are completed between 6-12 hours or 1-2 generations; and the changes in the relocalization of SWI/SNF proteins in response to reoxygenation are completed in less than one generation. In contrast, the transcriptome response to hypoxia are completed after 5-6 generations; and the transcriptome response to reoxygenation are completed in 2 generations [21]. These results show that changes in SWI/SNF protein localization precede transcriptome responses. They therefore strongly suggest that oxygen regulation of SWI/SNF protein localization contribute to, at least in part, oxygen regulation of gene expression.

Discussion
The SWI/SNF complex is an ATP-dependent chromatin remodeling complex [22]. Its composition and function are conserved from yeast to humans [13]. In yeast, more than 10% of the genes are the targets of SWI/SNF proteins, although the targets of different SWI/SNF proteins are different [20]. Hypoxia and reoxygenation induce changes in gene expression in over 20% of yeast genes [6]. Such broad changes in gene expression involve the action of an array of regulators. Previous studies have shown that Mga2, Rox1, Hap1 and Mot3 are all involved in mediating oxygen regulation of several subsets of genes [6,21,23,24]. In this report, we show that several SWI/SNF proteins alter their subcellular localization readily in response to hypoxia or reoxygenation and that this change in subcellular localization likely contributes to oxygen regulation of SWI/SNF target genes. Because the targets of SWI/SNF proteins overlap but are not identical [20], it is likely that different SWI/SNF proteins act on different groups of genes and control their expression. Here, we show that six SWI/SNF proteins accumulate in the cytosol in hypoxic cells, and relocalize to the nucleus in response to reoxygenation. Furthermore, several of the SWI/SNF proteins respond to hypoxia or reoxygenation and relocalize in a relatively quick manner. The redistribution of the six SWI/SNF proteins in the cytosol should presumably change the composition, and thereby the function or selectivity of the SWI/SNF complexes in the nucleus. Hence, the relocalization can affect the target expression of not only these SWI/SNF proteins whose localization is altered by hypoxia, such as Swi3, but also those whose localization is not affected by hypoxia, such as Swi2 and Taf14  (Table 3). Very likely, under hypoxia, because Swi3 and other proteins are predominantly present in the cytosol, the nuclear SWI/SNF proteins, such as Swi2, are likely complexed with other proteins, and act as chromatin remodelers on different sets of target genes. This explains why many SWI/SNF target genes are altered by hypoxia/reoxygenation (Table 3). Previous studies showed that Swi2, Arp7 and Arp9 form a core subcomplex possessing the ATP-dependent remodeling activity [25], while Swi3 controls SWI/SNF assembly, ATPdependent H2A-H2B displacement, as well as recruitment to target genes [25,26]. Notably, the nuclear localization of Swi2, Arp7 and Arp9 is not affected by hypoxia, while Swi3 is affected. This supports the idea that the core complex can associate with other as yet unidentified proteins and form a different kind of SWI/ SNF complexes in the nucleus in hypoxic cells.
These results suggest a model for how oxygen may modulate SWI/SNF complex composition and function ( Figure 8). In normoxic cells, SWI/SNF components form complexes in the nucleus and remodel chromatin structure at their target genes. In hypoxic cells Swi3 and other five SWI/SNF proteins accumulate in the cytosol, leaving the Swi2-Arp7-Arp9 core subcomplex available to interact with other proteins. Consequently, a different kind of SWI/SNF complex containing the core complex and other proteins (A, B and C in Figure 8) can be formed and act to remodel chromatin and control gene expression in different sets of genes. In response to reoxygenation, Swi3 and other proteins can readily relocalize to the nucleus, forming the normoxic SWI/SNF complexes, and reestablish gene expression patterns under normoxic conditions. This model is also consistent with a recent study showing that Swi3 is a key regulator in controlling respiration genes [27]. The authors used a computational approach to analyze modules of genes with a common regulation that are affected by specific DNA polymorphisms. They integrated genotypic and expression data for individuals in a segregating population with complementary expression data of strains mutated in a variety of regulatory proteins, in order to identify regulatory-linkage modules. In so doing, they found that Swi3 is a dominant regulator in the control of respiratory gene expression [27]. The effect of swi3 deletion is stronger than that of known respiratory regulators, including Hap2/3/4/5, Mot3 and Rox1. This is in complete agreement with our results showing that hundreds of SWI/SNF targets are altered by hypoxia (Table 3), and supports our model ( studies showed that SWI/SNF proteins are important for oxygen regulation in mammalian cells [28,29]. It is likely that SWI/SNF proteins can respond to changes in oxygen levels and regulate gene expression in diverse eukaryotes.

Conclusions
Several SWI/SNF proteins, including Swi3, Snf6 and Swp82, respond to hypoxia or reoxygenation and alter their subcellular distribution in a relatively quick manner. This change in localization likely contributes to oxygen regulation of SWI/SNF target genes.

Yeast strains and antibodies
The yeast GFP clone collection of 4159 strains expressing GFP-tagged proteins [30] was purchased from Invitrogen Corp. The anti-TAP monoclonal antibody was purchased from Open Biosystems.

The creation of hypoxic growth conditions
Hypoxic (~10 ppb O 2 ) growth condition was created by using a hypoxia chamber (Coy Laboratory, Inc.) and by filling the chamber with a mixture of 5% H 2 and 95% N 2 in the presence of a palladium catalyst [31]. The oxygen level in the chamber was monitored by using the Model 10 gas analyzer (Coy Laboratory, Inc.). The precise level of oxygen was also estimated by using a CHEMetrics rhodazine oxygen detection kit (K-7511) with the minimum detection limit at 1 ppb, and a range of 0-20 ppb. The hypoxic state was further confirmed by measuring oxygen-controlled promoter activities, including UAS1/ CYC1, ANB1 and OLE1 [9,10,31].
Time course characterization of cellular localization of SWI/SNF proteins For a time course characterization of SWI/SNF protein relocalization in response to hypoxia or reoxygenation, we used a previously defined nuclear protein import assay in yeast [32][33][34]. Briefly, cells expressing GFPtagged proteins at various time points of hypoxia or reoxygenation treatment were collected, and images were acquired. At least 25 cells were counted at each time point, and three sets of cells were counted. A particular cell was counted as having the GFP-tagged protein in the nucleus if the nucleus was much brighter than the surrounding cytoplasm and a clear nuclear-cytoplasmic boundary was visible. Cells with excessive bright or weak fluorescence or with aberrant morphology were not scored.

Confocal fluorescent live cell imaging and quantitation
GFP-tagged strains were grown in synthetic complete media in air or in a hypoxia chamber. Cells were collected and subjected to confocal fluorescent imaging and quantitation. Image acquisition of live cells was performed by using a Perkin Elmer UltraView ERS Spinning Disc Confocal Microscope (Perkin Elmer, Waltham, MA) with a Zeiss 100x/1.4 Oil Immersion objective (Carl Zeiss, Thornwood, NY). High speed images were captured by using a Hamamatsu EMCCD C9100 digital camera (Hamamatsu Corporation, Bridgewater, NJ). Zstacks were recorded for the DAPI channel (EX 405nm) and the GFP channel (EX 488nm) by moving the objective turret with a UltraView z-focus drive (Perkin Elmer, Waltham, MA). Volocity 5.4.2 (Improvision, Perkin Elmer, MA, USA) was used for image acquisition. The 3D confocal images were analyzed, and statistical data were collected by using Imaris 7.4.0 (Bitplane, South Windsor, CT).

Preparation of yeast cell extracts and Western blotting
Yeast cells expressing various TAP-tagged proteins were grown to an optical density (OD 600 ) of approximately 0.8. Cells were harvested and resuspended in 3 packed cell volumes of buffer (20 mM Tris, 10 mM MgCl 2 , 1 mM EDTA, 10% glycerol, 1 mM dithiothreitol, 0.3 M NaCl, 1 mM phenylmethylsulfonyl fluoride, 1 mg of pepstatin per ml, 1 mg of leupeptin per ml). Cells were then permeabilized by agitation with 4 packed cell volumes of glass beads, and extracts were collected as described (See figure on previous page.) Figure 7 DAPI and merged confocal fluorescent images of cells expressing SWI/SNF proteins whose cellular location is affected by hypoxia. Cells expressing Snf5-GFP, Snf6-GFP, Snf11-GFP, Snf12-GFP and Swp82-GFP were grown in air or under hypoxia (Hyp), and the images were captured. Only the images of hypoxic cells are shown, because the normoxic cells all exhibit the same nuclear pattern as shown in Figures 5 and 6. The percentages of GFP fluorescence in the nucleus (N) or cytosol (C) were quantified and plotted here. The scale bar represents 1 μm.  Figure 8 A cartoon illustrating how oxygen may affect SWI/SNF composition and function. In normoxic cells, the components form the SWI/SNF complex in the nucleus, enabling it to remodel chromatin at the target genes. In hypoxic cells, Swi3 and five other components are retained in the cytosol, perhaps due to modifications of these components and/or interactions with unidentified factor(s) X. In the nucleus, Swi2 and other remaining components may interact with some other proteins (marked as A, B, and C), forming complexes with different composition and targeting different sets of genes.