Variation of the response to metal ions and nonsense-mediated mRNA decay across different Saccharomyces cerevisiae genetic backgrounds

Abstract

Regulation of mRNA steady-state levels is important in controlling gene expression particularly in response to environmental stimuli. This allows cells to rapidly respond to environment changes. The highly conserved nonsense-mediated mRNA decay (NMD) pathway was initially identified as a pathway that degrades aberrant mRNAs. NMD is now recognized as a pathway with additional functions including precisely regulating the expression of select natural mRNAs. Majority of these natural mRNAs encode fully functional proteins. Regulation of natural mRNAs by NMD is activated by NMD targeting features and environmental cues. Here, we show that Saccharomyces cerevisiae strains from three genetic backgrounds respond differentially to NMD depending on the environmental stimuli. We found that wild type and NMD mutant W303a, BY4741, and RM11–1a yeast strains respond similarly to copper in the environment but respond differentially to toxic cadmium. Furthermore, the PCA1 alleles encoding different mRNAs from W303a and RM11–1a strains are regulated similarly by NMD in response to the bio-metal copper but differentially in response to toxic cadmium.

1. INTRODUCTION

The highly conserved Nonsense-mediated mRNA decay (NMD) pathway triggers rapid degradation of mRNAs that prematurely terminate translation. This includes mRNAs containing premature termination codons (PTCs) and natural mRNAs. Majority of these natural mRNAs encode fully functional proteins that contain NMD-targeting features. These features include mRNAs subject to leaky ribosomal scanning (Guan et al., 2006; Welch & Jacobson, 1999), a translated upstream open reading frame or frames (Gaba et al., 2005; Guan et al., 2006), ribosomal frameshift signals (A.T. Belew et al., 2008, 2011), inefficiently spliced pre-mRNAs (He et al., 1993), and mRNAs with atypically long 3'-untranslated regions (3'-UTRs) (B.W. Kebaara & Atkin, 2009; Muhlrad & Parker, 1999). In addition to NMD triggering features, NMD-mediated degradation of mRNAs can respond to growth conditions. It has been shown that regulation of a number of transcripts by NMD depends on environmental conditions (Gaba et al., 2005; Johansson & Jacobson, 2010; M. Peccarelli et al., 2016, 2019). In Saccharomyces cerevisiae and other eukaryotes, three core NMD factors, Upf1p, Upf2p, and Upf3p, are required for the pathway to function.

NMD plays a role in bio-metal homeostatic mechanisms, including magnesium, copper, and iron homeostasis. Copper is a micronutrient that is essential for a number of cellular functions but is toxic in excess (Deliz-Aguirre et al., 2011; Hodgins-Davis et al., 2012). Copper serves as a cofactor in superoxide anion detoxification, iron metabolism, and mitochondrial oxidative phosphorylation. Although copper is required for normal cellular function, excessive amounts of copper or in its free form is extremely toxic to the cell. Thus, the copper concentration inside of the cell must remain around one free copper molecule per individual cell to avoid toxicity (Hodgins-Davis et al., 2012). In order to avoid cell death due to copper induced toxicity, cells have developed a number of mechanisms to maintain copper homeostasis. These mechanisms include copper compartmentalization and sequestration (Bird, 2015). Specific proteins involved in these mechanisms have been identified; however, the actual mechanisms involved in copper homeostasis are still under investigation, specifically the effect of copper on posttranscriptional regulation at the mRNA level.

Studies of gene expression in response to environmental stimuli frequently examine regulation of gene expression at the transcriptional level (i.e., increase or decrease of the level of transcription). However, the amount of mRNA present in a cell at any specific time is dependent on the mRNAs rate of synthesis and decay because mRNA levels can be increased by stabilization. This type of regulation would allow rapid response to changes in environmental stimuli. We have identified a set of mRNAs involved in copper homeostasis in S. cerevisiae that are regulated by NMD (M. Peccarelli et al., 2016). A number of these mRNAs are regulated by NMD in response to environmental copper levels. Thus, copper homeostasis in S. cerevisiae occurs at the level of mRNA stability.

These NMD regulated transcripts include mRNAs that encode proteins involved in copper transport, mitochondrial copper homeostasis (Murtha et al., 2018), and protection from metal toxicity. The two mRNAs found to be regulated by NMD and involved in protection from metal toxicity are CRS5 and PCA1 (M. Peccarelli et al., 2016). In yeast cells, the CRS5 (Copper-resistance-suppressor) and CUP1 genes encode the two copper binding metallothioneins. The major metallothionein, CUP1, is induced by the ACE1 transcription factor when yeast cells are exposed to elevated copper levels (Wang et al., 2013). Interestingly, CUP1 mRNA is not regulated by NMD, whereas CRS5 is an indirect NMD target (Murtha et al., 2018; M. Peccarelli et al., 2016). The decay rate of the CRS5 mRNA is comparable in wild-type and NMD mutant yeast cells grown in rich media. This was anticipated because CRS5 mRNA does not contain any identifiable NMD-targeting feature (M. Peccarelli et al., 2016).

The second mRNA regulated by NMD that encodes a protein involved in protection from metal toxicity is the PCA1 mRNA. PCA1 is an evolutionarily conserved P_1B-type cation-transporting ATPase (Rad et al., 1994). In addition to the PCA1 allele found in common laboratory strains, CAD2 an alias of PCA1 has also been characterized (Shiraishi et al., 2000). The protein encoded by CAD2 transports cadmium out of yeast cells (D.J. Adle et al., 2007). Cadmium (Cd) is a non-essential, toxic environmental contaminant that is classified as a human carcinogen (Wysocki & Tamas, 2010). A study on the evolution of metal resistance in natural yeast populations found that cadmium tolerance was controlled solely by the PCA1 locus (Chang & Leu, 2011). In addition, Pca1p has a role in copper and iron homeostasis. Copper increases expression of a FLAG tagged Pca1p in an ACE1 deletion background (D.J. Adle et al., 2007). Previous studies suggest that most laboratory yeast strains have 970R instead of 970G that alters the ATP-binding domain of the Pca1p. This version of Pca1p does not transport cadmium and mislocalizes to vesicle-like compartments (D.J. Adle et al., 2007). Copper tolerance mediated by Pca1p is understood to be through metal binding and sequestration rather than ATPase activity (D.J. Adle et al., 2007). PCA1 shares significant sequence similarity with ATP7B. ATP7B encodes a P-type ATPase that transports copper across membranes. In humans, a substitution in the ATP7B similar to the one found in PCA1 leads to Wilson's disease. Wilson's disease is characterized by excessive accumulation of copper in hepatic and neuronal tissues.

Using yeast strains from the W303 genetic background, we previously found that the PCA1 gene encodes two major transcripts in rich media. The longer PCA1 mRNA has an atypically long 3'-UTR of 650 nucleotides (nts) and may be subject to –1 programed ribosomal frameshifting (–1PRF). These are two known NMD targeting features in yeast (Peccarelli & Kebaara, 2014b). The longer PCA1 mRNA accumulated to higher levels in NMD mutants relative to wild-type yeast strains (M. Peccarelli et al., 2016). However, the decay rate of the PCA1 mRNAs was similar in wild-type and NMD mutant strains grown under normal growth conditions. Thus, under these conditions PCA1 mRNA is an indirect NMD target. This is despite PCA1 mRNA containing identifiable NMD targeting features suggesting that the mRNA maybe directly regulated by NMD under some conditions (M. Peccarelli et al., 2016). Direct NMD targets have altered decay rates in wild-type and NMD mutants.

Here, we examined the metal tolerance of three yeast strains and the connection regulation of PCA1 mRNAs by NMD has on metal tolerance. We used three genetically distinct yeast strains. The commonly utilized laboratory strains W303a, BY4741, and a natural yeast isolate RM11–1a that expresses the CAD2 gene (hereafter referred to as PCA1 970G; D.J. Adle et al., 2007). We found that wild-type W303a, BY4741, and RM11–1a are sensitive to toxic copper levels relative to the NMD mutants. As previously reported, RM11–1a wild-type are tolerant of cadmium (D.J. Adle et al., 2007), whereas both wild-type and NMD mutants W303a and BY4741 are sensitive to cadmium. BY4741 wild-type was more tolerant of cadmium relative to the NMD mutant. Deletion of PCA1 in BY4741 genetic background reversed the cadmium tolerance phenotype. Furthermore, the growth of W303a wild type was impaired under low copper conditions relative to the NMD mutant. Under low iron conditions, W303a, BY4741, and RM11–1a NMD mutants had impaired growth relative to the wild-type strains.

Analysis of PCA1 and PCA1 970G mRNA levels in wild-type and NMD mutants showed altered expression of the most highly expressed PCA1 and PCA1 970G mRNA isoforms under low copper conditions. In agreement with our previous study, PCA1 mRNAs are indirect NMD targets in rich media but are degraded at different rates depending on copper levels. Interestingly, when the yeast strains were grown in media containing cadmium, PCA1 and PCA1 970G mRNAs were differentially regulated by NMD. PCA1 mRNA from the W303a strain was regulated by NMD, whereas the PCA1 970G from RM11–1a which was highly induced under high cadmium was immune to NMD. Additionally, phylogenetic analysis of PCA1 genes from 165 laboratory, domesticated, clinical, and natural yeast strains revealed that 970R was the ancestral allele and 970G evolved from that ancestral lineage (Sardi et al., 2018).

2. RESULTS

2.1. Growth of W303a and RM11–1a wild-type and NMD mutants under iron and copper deplete conditions, elevated cadmium, and copper

We previously found that NMD mutants from a W303a genetic background are more tolerant of toxic copper levels above 400 μM relative to wild-type yeast strains (Deliz-Aguirre et al., 2011; Wang et al., 2013). To distinguish the metal tolerance of W303a and RM11–1a wild-type and NMD mutants, yeast strains were grown on complete minimal (CM), CM containing excessive amounts of copper (100 and 600 μM copper) and 100 μM cadmium. In addition, both strains were grown under copper and iron depleted conditions. For low copper, the strains were grown on media containing 100 μM Bathocuproinedisulfonic acid (BCS) and 100 μM Bathophenanthrolinedisulfonic acid (BPS) for low iron.

No difference in growth was observed between any of the yeast strains on CM and CM containing 100 μM copper after 3 days incubation at 30°C. All the strains grew equally well (Figure 1a). We found that W303a and RM11–1a NMD mutants are more tolerant of 600 μM copper relative to the wild-type strains (Figure 1a). Additionally, the W303a wild-type and NMD mutant had higher tolerance of 600 μM copper relative to RM11–1a wild-type and the NMD mutant. Furthermore, RM11–1a wild-type and NMD mutant expressing the PCA1 970G gene tolerated 100 μM cadmium, whereas W303a wild-type and NMD mutants had limited growth (Figure 1b). Because of the limited growth of the W303a wild-type and NMD mutant, there was no detectable difference in growth between the wild-type and NMD mutant. Furthermore, at 100 μM cadmium, there was no visible difference in growth between the RM11–1a wild-type and the NMD mutant (Figure 1b).

FIGURE 1.

W303a and RM11–1a NMD mutants are more tolerant of elevated copper levels; however, RM11–1a wild-type and NMD mutant tolerate cadmium relative to wild-type and NMD mutant W303a strains. UPF1 W303a (wild-type), upf1△ W303a (NMD mutant), UPF1 RM11–1a (wild-type), and upf1△ RM11–1a (NMD mutant) yeast cells were grown to mid-log phase in complete minimal media. Tenfold serial dilutions of the cells were spotted onto complete minimal medium and medium containing 100 μM copper or 600 μM copper (a), 100 μM cadmium (b), 100 μM Bathocuproinedisulfonic acid (BCS) (c), and 100 μM Bathophenanthrolinedisulfonic acid (BPS) (d) and incubated at 30°C for 3 days before photography. UPF1 W303a (wild-type) and upf1△ W303a (NMD mutant) were grown in complete minimal and 100 μM Bathophenanthrolinedisulfonic acid (BPS), low Fe conditions in liquid culture for 12 h (e)

W303a wild-type and NMD mutants grew well under copper depleted conditions (Figure 1c). The W303a NMD mutant was more tolerant of low copper conditions relative to the wild-type. Interestingly, RM11–1a wild-type and NMD mutant grew equally well and were both more tolerant of low copper conditions relative to W303a wild-type (Figure 1c). Under iron-depleted conditions, all four strains had limited growth although the W303a wild-type strain was more tolerant of low iron conditions relative to the NMD mutants both on plates and in liquid media after longer incubation (Figures 1d,e and 2d). It is important to note that the RM11–1a wild-type and NMD mutant strains tolerance to cadmium and enhanced growth under low copper conditions relative to W303a may not be fully attributable to the PCA1 970G gene because these yeast strains have different genetic backgrounds.

FIGURE 2.

Resistance of BY4741 and BY4741 pca1△ wild-type and NMD mutants to low copper and iron and elevated copper and cadmium. UPF1 W303a (wild-type), upf1△ W303a (NMD mutant), UPF1 BY4741 (wild-type), upf1△ BY4741 (NMD mutant), pca1△ UPF1 BY4741 (wild-type pca1△), and pca1△ upf1△ BY4741 (NMD mutant pca1△) yeast cells were grown to mid-log phase in complete minimal media. Tenfold serial dilutions of the cells were spotted onto complete minimal medium and medium containing 100 μM copper or 600 μM copper (a), 100 μM cadmium (b), 100 μM Bathocuproinedisulfonic acid (BCS) (c), and 100 μM Bathophenanthrolinedisulfonic acid (BPS) (d) and incubated at 30°C for 4 days before photographing

2.2. Growth of BY4741 wild-type and NMD mutants with PCA1 disrupted is affected under cadmium and low iron

To examine the effect PCA1 has on the growth of wild-type and NMD mutants on the growth conditions tested above, we utilized the BY4741 yeast strain that was used to generate the yeast deletion library. BY4741 is a derivative of S288C and contains the PCA1 allele found in W303a. We used BY4741 to avoid interference from other factors in the RM11–1a and W303a genetic backgrounds. We also used W303a as a control for growth under the same conditions tested in Figure 1.

No difference in growth was observed between W303a and BY4741 wild-type and NMD mutant strains after incubation for 4 days at 30°C on CM (Figure 2a). The plates were incubated for 4 days to better observe the growth differences under cadmium and low iron conditions. The same strains grown for 3 days are in Figure S1. BY4741 wild-type, NMD mutant and the PCA1 deletion strains grew equally well on media containing 100 μM copper (Figure 2a). We also found that BY4741 wild-type and NMD mutants with or without PCA1 are tolerant of 600 μM copper (Figure 2a). The tolerance of these strains to copper is the same as the W303a NMD mutant (Figure 2a). The BY4741 wild-type and NMD mutants grown on media containing 100 and 600 μM copper were brown compared to the W303a. This is likely due to copper sulfide (CuS) mineralization (Yu et al., 1996). This phenotypic observation suggests that there are other aspects of copper homeostasis that are different in BY4741.

Similar to the W303a wild-type and NMD mutant, BY4741 wild-type had limited growth under 100 μM cadmium (Figure 2b). Notably, under 100 μM, cadmium BY4741 NMD mutant had reduced growth relative to the wild-type strain. Interestingly, deletion of the PCA1 gene reversed this phenotype, BY4741 wild-type with PCA1 deleted was more sensitive to cadmium relative to the BY4741 NMD mutant lacking PCA1 (Figure 2b). These observations corroborate the observation that the PCA1 gene plays a role in cadmium tolerance. However, deletion of the gene in an NMD mutant background enhanced the growth of the yeast strains under cadmium suggesting that there are compensatory genes that are regulated by NMD and play a role in cadmium tolerance. Furthermore, under low copper BY4741 wild-type and NMD mutants grew at comparable levels as the W303a wild-type and NMD mutant (Figure 2c). Deletion of PCA1 in the BY4741 strain did not have any effect on the growth of the wild-type and NMD mutants in this condition (Figure 2c). Growth under low iron was impaired but appears to show that both W303a and BY4741 wild-type strains are more tolerant of low iron than their NMD mutants (Figure 2d). Deletion of PCA1 resulted in the wild-type and NMD mutant BY4741 growing at comparable rates. Altogether, these observations suggest that deletion of PCA1 in the BY4741 genetic background affects growth of wild-type and NMD mutants under cadmium and low iron conditions.

2.3. PCA1 and PCA1 970G mRNAs are differentially regulated by NMD depending on environmental conditions

To examine the molecular basis of the phenotypes observed, the role NMD plays in the expression of PCA1 and PCA1 970G under varying environmental conditions was investigated. We measured the steady-state accumulation levels of PCA1 and PCA1 970G mRNAs in varying amounts of copper and in cadmium. In agreement with our previous study, we found that the PCA1 gene encodes three mRNA isoforms with two major isoforms when yeast cells are grown in CM (Figure 3a). These three mRNA isoforms are specific to PCA1 because they are absent from BY4741 PCA1 deletion strain, and they vary in the length of the 3'-UTR. (Figure 3e, lane 3 pca1△). The short version of the PCA1 mRNA has a 3'-UTR of 200 nt (PCA1-short), whereas a longer more abundant mRNA has a 3'-UTR of 650 nt (PCA1-long; Figure 3a).

FIGURE 3.

Regulation of PCA1-long and PCA1970G-long mRNAs by NMD depends on copper levels. Schematic representations of PCA1 mRNA isoforms (a), the open triangle indicates the potential –1 programmed ribosomal frameshift (–1PRF) and the atypically long 3'-UTR. Representative mRNA steady-state accumulation levels (b–e). Steady-state accumulation levels were measured with total RNA from wild-type strains W303a (UPF1; Wente et al., 1992), RM11-a (UPF1; D.J. Adle et al., 2007), and NMD mutants (upf1△; B. Kebaara et al., 2003) and BKY111 (upf1△). The northern blots were probed with DNA specific to the PCA1 open reading frame, (ORF) indicated by the red line in (a). The major PCA1 970G-long mRNA isoform fold change in BKY111/RM11–1a (upf1△/UPF1) are shown to the right of the northern blots (b–e). The fold changes shown to the right of the northern blots are for the conditions listed above the specific northern blot. CUP1, CTR1, CYH2, and SCR1 were used as controls. CTR1 was used as a low copper control (e). CTR1 encodes a high affinity copper transporter of the plasma membrane. Low copper levels result in increased CTR1 expression. CUP1 was used as a control for high copper because CUP1 encodes a metallothionein that binds copper. The CUP1 gene is induced by the Ace1 transcription factor when cells are exposed to elevated copper levels. CUP1 mRNA expression increase in elevated copper conditions. CYH2 pre-mRNA was used as an NMD control because CYH2 pre-mRNA is degraded by NMD. SCR1 was used as a loading control for all northern blots. SCR1 is an RNA polymerase III transcript that is not regulated by NMD or sensitive to copper or cadmium levels

The PCA1 and PCA1 970G mRNAs accumulate to higher levels in NMD mutants when both strains were grown in CM media. The PCA1-long mRNA isoform was the most abundant in both yeast strains; therefore, we quantified and compared the levels of this mRNA isoform (Figure 3b). Interestingly, this mRNA isoform accumulated sevenfold higher in W303a relative to RM11–1a wild-type strain and fivefold higher in the W303a NMD mutant relative to the RM11–1a NMD mutant (Figure 3b). This is not due to a more efficient NMD pathway in the W303a strains because both W303a and RM11–1a yeast strains degrade the CYH2 pre-mRNA with comparable efficiency (Figure 3b, middle panel). The CYH2 pre-mRNA/mRNA ratio for wild-type W303a was 0.037 relative to 0.003 for the RM11–1a wild-type strains. Additionally, the CYH2 pre-mRNA accumulates to comparable levels in W303a and RM11–1a NMD mutants (0.89 and 0.66, respectively; Figure 3b, middle panel). Furthermore, both PCA1 and PCA1 970G mRNAs are regulated by NMD under normal growth conditions (Figure 3b and Table 1). The PCA1-long mRNA isoform accumulated 2.7 (±0.3) fold higher in the NMD mutant relative to the wild-type W303a strains (Figure 3b, lanes 1, 2). The PCA1 970G-long mRNAs accumulated 10.9 (±3.7) fold higher in the NMD mutant relative to wild-type RM11–1a (Figure 3b, lanes 3 and 4, and Table 1). These results indicate that both the PCA1-long transcripts found in most laboratory strains and the PCA1 970G-long found in most other yeast strains are regulated by NMD under normal growth conditions; however, they accumulate to different levels.

TABLE 1.

PCA1-long (W303a) and PCA1 970G-long (RM11-1a) relative mRNA accumulation levels in W303a and RM11-1a under different growth conditions

Growth media	W303 (upf1Δ/UPF1)	RM11-1 (upf1Δ/UPF1)
Complete minimal	2.71 (±0.3)	10.9 (±3.7)
100 uM copper	3.31 (±0.3)	8.0 (±3.2)
600 uM copper	2.15 (±1.2)	1.63 (±1.3)
100 uM cadmium	3.17 (±1.3)	1.0 (±0.3)
100 uM BCS*	1.73 (±0.5)	2.28 (±1.0)

Note: mRNA steady-state accumulation levels were done in triplicate (except W303 in 100 μM cadmium, which was done in duplicate) and reported as an average ± standard deviation (SD).

Abbreviation: BCS*, (Bathocuproinedisulfonic acid) Low copper.

Since Pca1p is involved in protection from metal toxicity we examined the extent to which regulation of PCA1-long and PCA1 970G-long mRNAs by NMD is responsive to changes in environmental copper and cadmium levels. Yeast strains with a functional and nonfunctional NMD pathway were grown under low copper, high copper (100 and 600 μM copper), and 100 μM cadmium.

The PCA1-long mRNA accumulated to higher levels in the NMD mutants under all conditions tested here in the W303a strain. Under 100 μM copper and cadmium, the PCA1-long mRNA accumulated approximately threefold higher in the NMD mutant relative to the wild-type strain (Figure 3c, lanes 1 and 2, and Table 1). Under low copper conditions, the PCA1-long mRNA accumulated 1.73 (±0.5) fold higher in the NMD mutant relative to the wild-type strain (Figure 3e, lanes 4 and 5). The steady-state accumulation levels of PCA1-long mRNA under low copper are significantly lower than the accumulations levels in CM (Figure 3d, lanes 1 and 2, and Table 1). These observations suggest that for the W303a strain, the regulation of the PCA1-long mRNA levels by NMD may be dependent on environmental copper levels.

The PCA1 970G-long mRNA transcripts expressed by the RM11–1a wild-type and NMD mutant are also responsive to NMD under CM, copper deplete, and copper supplemented conditions comparable to the W303a strain. However, the PCA1 970G-long mRNAs are immune to NMD when grown in media containing cadmium (Figure 4a lanes 1 and 2; Figure 4b lanes 3 and 4; and Table 1). The PCA1 970G-long mRNA accumulated 8.0 (±3.2) fold higher in the NMD mutant relative to the wild-type strain under 100 μM copper. These levels of accumulation are not significantly lower than CM but are significantly higher than any other conditions examined here. Furthermore, PCA1 970G-long accumulated 1.63 (±1.3) fold in the NMD mutant relative to the wild-type under 600 μM copper (Figure 3d, lanes 3 and 4). This level of accumulation is significantly lower than what was observed for 100 μM copper, but comparable to the W303a strains grown under identical conditions, suggesting that both yeast strains respond similarly to copper in the environment (Table 1).

FIGURE 4.

PCA1-long mRNA is sensitive to NMD in the presence of cadmium; however, PCA1 970G-long mRNAs escape NMD-mediated degradation under these conditions

PCA1-long and PCA1 970G-long mRNA steady-state accumulation levels (a and b). Steady-state accumulation levels were measured with total RNA from wild-type strains W303a (UPF1; Wente et al., 1992), RM11-a (UPF1; D.J. Adle et al., 2007), and NMD mutants (upf1△; B. Kebaara et al., 2003) and BKY111 (upf1△). The northern blots were probed with DNA specific to the PCA1, CUP1, CYH2, and SCR1. The fold changes (upf1△/UPF1) for the major PCA1-long (a) PCA1 970G-long and (b) mRNA isoforms are shown to the right of the northern blots. The fold changes shown to the right of the northern blots are for 100 μM cadmium. CUP1, CYH2, and SCR1 were used as controls as described in Figure 3

However, when W303a and RM11–1a wild-type and NMD mutant strains were grown on medium containing cadmium, they respond differently (Figure 4a,b). Notably, the expression of PCA1 970G-long mRNAs is induced and considerably higher in RM11–1a wild-type and NMD mutants when cells were grown in media containing 100 μM cadmium relative to CM media (Figure 4b). When detected on the same northern blot, PCA1 970G-long mRNAs are barely detectable in RM11–1a strains grown in CM media relative to those grown under high cadmium (Figure 4b, compare lanes 1 and 2 with lanes 3 and 4). The PCA1 970G-long mRNA accumulated 149 fold higher in RM11–1a wild-type strain grown under cadmium relative to CM and 27 fold higher in the RM11–1a NMD mutant grown under cadmium relative to CM media (Figure 4b). Additionally, these data show that in CM media, PCA1 970G-long mRNA is regulated by NMD but not under high cadmium. Under these growth conditions, the PCA1 970G-long mRNA does not accumulate to higher levels in the RM11–1a NMD mutant relative to the wild-type strain (Figure 4a, lanes 1 and 2, and Figure 4b, lanes 3 and 4). These results are distinct from PCA1-long mRNA encoded by W303a strains. Under the same conditions and on the same northern blot, the PCA1-long mRNA accumulates 3.17 (±1.3) fold higher in the NMD mutant relative to the W303a wild type strain (Figure 4a, lanes 3 and 4, and Table 1). In contrast, PCA1 970G-long mRNA accumulated 1.0 (±0.3) fold in the NMD mutant relative to the wild-type RM11–1a strain (Figure 4a, lanes 1 and 2; Figure 4b, lanes 3 and 4; and Table 1).

The lack of accumulation of the PCA1 970G-long mRNA in RM11–1a NMD mutants relative to the wild-type strain is not due to a non-functional or impaired NMD pathway under toxic cadmium. Under 100 μM cadmium, NMD is functional because the CYH2 pre-mRNA accumulates to higher levels in W303a and RM11–1a NMD mutants relative to wild-type yeast strains (Figure 4a,b, middle panels). Furthermore, the PCA1-long mRNA from the W303a accumulates to higher levels in NMD mutants relative to the wild-type strain. These observations suggest that PCA1-long and PCA1 970G-long are differentially regulated by NMD in the presence of cadmium. This differential regulation of PCA1 970G is specific to cadmium because similar escape from NMD was not observed when the yeast strains were grown under CM and variable copper concentrations.

CUP1 mRNA was used as a control for high copper conditions because it encodes for a metallothionein that binds copper. The CUP1 gene is induced by the copper activated Ace1 transcription factor when cells are exposed to elevated copper levels. CUP1 mRNA expression increase in elevated copper conditions and the mRNA is not an NMD target as stated previously. When W303a and RM11–1a wild-type and NMD mutant strains were grown under 100 μM cadmium, CUP1 mRNA accumulates to higher levels in W303a relative to RM11–1a strains (Figure 4a). Furthermore, the levels of CUP1 mRNA were lower in the RM11–1a strains grown under 100 μM cadmium relative to CM (Figure 4b).

2.4. PCA1-long and PCA1 970G-long mRNAs are indirectly regulated by NMD under normal growth conditions

The decay of PCA1-long and PCA1 970G-long mRNAs was measured under normal growth conditions to determine whether the mRNAs were directly or indirectly regulated by NMD. We previously measured the decay of the PCA1-long mRNA using W303a strains harboring the temperature sensitive allele of RNA polymerase II grown in rich media. The PCA1-long mRNA was expressed at higher levels in the NMD mutant strain as observed in the steady state accumulation northern (M. Peccarelli et al., 2016). Under rich media, the PCA1-long mRNA was indirectly regulated by NMD (M. Peccarelli et al., 2016).

To compare decay rates of PCA1-long to the PCA1 970G-long mRNA under identical growth conditions, we measured the decay of PCA1-long and PCA1 970G-long mRNAs using yeast strains grown in CM media using thiolutin to inhibit transcription (10 mg/ml in DMSO was added to a final concentration of 10 μg/ml). Under these conditions, both the PCA1-long and PCA1 970G-long mRNAs accumulated to levels lower than in rich media and were degraded by NMD with comparable decay rates (Figure 5a–d). The half-life of PCA1-long and PCA1 970G-long in the wild-type strains (UPF1) was biphasic with an initial decay rate of 12 min. The half-life of PCA1-long in the NMD mutants (upf1△) was 13 min relative to 15 min for the PCA1 970G long mRNA (Figure 5b,d). The biphasic decay of PCA1-long and PCA1 970G-long in the wild-type strains is not due to incomplete shut down of transcription by thiolutin. CYH2 pre-mRNA and mRNA controls of the northern blots demonstrate that the increase in PCA1 mRNA in W303a and RM11–1a is not due to ongoing transcription because expression of the CYH2 pre-mRNA is highest at time 0 followed by a steady decrease in mRNA abundance. This is consistent with a rapid and complete inhibition of transcription by thiolutin (Figure 5a,c, bottom panels). Furthermore, this initial decrease then increase in PCA1 mRNA levels after inhibition of transcription with thiolutin is not specific to thiolutin as a transcriptional inhibitor. We observed a similar decay trend when using W303a rpb1–1 yeast strains to measure CTR2 mRNA half-lives in wild-type and NMD mutants (Deliz-Aguirre et al., 2011). Altogether, these observations suggest that both PCA1-long mRNAs from W303a and PCA1 970G-long from RM11–1a are indirectly regulated by NMD under normal growth conditions, and additionally, they decay at comparable rates.

FIGURE 5.

Decay of PCA1-long and PCA1970G-long mRNAs in complete minimal media is indirect. Representative half-life Northern blots of the PCA1 mRNAs (a, b). Representative half-life northern blots for PCA1 970G mRNAs (c, d). The half-lives were measured with total RNA extracted wild-types W303a (UPF1; Wente et al., 1992), RM11-a (UPF1; D.J. Adle et al., 2007), and NMD mutants (upf1△; B. Kebaara et al., 2003) and BKY111 (upf1△)(a–d) grown in complete minimal media. Yeast strains were harvested over a 35 min time period after transcription was inhibited using thiolutin (a–d). The time points the cells were harvested are indicated above the Northern blots. The Northern blots were probed with radiolabeled DNA from the PCA1, CYH2, and SCR1 ORFs. Representative CYH2 pre-mRNA and mRNA half-life Northern blots for W303a (a, b) and RM11–1a (c, d) are shown. The half-lives were measured using SigmaPlot by calculating the time it takes for half of the original mRNA levels to degrade. The half-lives are shown to the right of each Northern blot

2.5. Regulation of PCA1-long mRNA by NMD under varying levels of copper is indirect

A previous study found that copper increases the expression of flag tagged Pca1p due to protein stabilization (D.J. Adle et al., 2007). We examined the extent to which varying copper levels in the environment affect the decay of the PCA1 mRNA by NMD. We measured the decay rate of the PCA1-long mRNA using W303a strains harboring the temperature sensitive allele of RNA polymerase II, because the decay of PCA1-long and PCA1 970G-long mRNAs in CM media was comparable and the response of the mRNAs to copper was similar. Additionally, we regularly use these strains to measure mRNA half-lives (Peccarelli & Kebaara, 2014a).

To determine the effect copper depletion has on the decay of PCA1-long mRNA, we measured the decay of the mRNA in wild-type and NMD mutants grown under low copper conditions. To confirm that the conditions were low copper, the northern blots were also probed with CTR1 (not shown). Under low copper conditions, the PCA1-long mRNA accumulated to higher levels in the NMD mutant relative to the wild-type and were indirectly regulated by the pathway (Figure 6a and Tables 1 and 2). Under these conditions, the half-life of the PCA1-long mRNA in the wild-type strain (UPF1) was 8.0 min relative to 10.3 min in the NMD mutant (Figure 6a and Table 2). In both the wild-type and NMD mutant, the PCA1-long mRNA decayed at comparable rates to CM media (Figures 5a,b and 6a).

FIGURE 6.

Regulation of PCA1-long mRNA by NMD under varying levels of copper is indirect. (a–c) Representative half-life Northern blots of the PCA1-long mRNAs under different environmental conditions. The half-lives were measured with total RNA extracted from wild-type strain AAY334 (UPF1 rpb1–1; B. Kebaara et al., 2003) and NMD mutant strain AAY335 (upf1△ rpb1–1; B. Kebaara et al., 2003) grown in media containing Bathocuproinedisulfonic acid (BCS), 100 and 600 μM copper. Yeast cells were harvested over a 35 min period after transcription was inhibited at time 0 and cells harvested at the times indicated above the Northern blots. The Northern blots were probed with radiolabeled DNA from the PCA1, CTR1, CUP1, CYH2, and SCR1 ORFs. The half-lives were measured using SigmaPlot by calculating the time it takes for half of the original mRNA levels to degrade. The half-lives are shown to the right of each Northern blot and are an average of at least three independent experiments

TABLE 2.

PCA1-long mRNA half-lives were measured in W303a wild-type (UPF1 rpb1-1) and NMD mutants (upf1Δ rpb1-1)

Growth media	UPF1	upf1Δ
YAPD	22.0 (± 7.4)	15.9 (± 3.3)
100 uM BCS*	8.0 (± 1.0)	10.3 (± 2.5)
100 uM Copper	14.0 (± 2.8)	17.7 (± 4.2)
600 uM Copper	11.7 (± 2.1)	6.7 (± 2.3)

Note: All yeast strains used were grown under the conditions indicated on the table. All mRNA half-lives were done in triplicate and reported as an average ± standard deviation (SD).

Abbreviation: BCS*, (Bathocuproinedisulfonic acid) Low copper.

Under copper replete conditions (100 μM copper), we found that the PCA1-long mRNA was stabilized in wild-type and NMD mutants relative to low copper (Figure 6b and Table 2). In these conditions, the half-life of PCA1-long mRNA in the wild-type strain (UPF1) was 14.0 mins relative to 17.7 min in the NMD mutant (Figure 6b and Table 2). Thus, the decay of PCA1-long mRNAs was comparable in the wild-type and NMD mutant. Surprisingly, increasing the concentration of copper to 600 μM did not further stabilize the PCA1-long mRNAs. Instead, the mRNAs were degraded at significantly faster rates, particularly in the NMD mutant (Figure 6c). Under 600 μM copper, the half-life of PCA1-long mRNA was 11.7 min in the wild-type strain (UPF1) relative to 6.7 min in the NMD mutant (Figure 6C and Table 2). The faster decay rates of the PCA1-long mRNA in NMD mutants under 600 μM copper correspond to the low steady-state accumulation levels of the mRNAs under these conditions (Table 1). These observations suggest that increasing copper concentration affects NMD-mediated decay and steady-state accumulation levels of the PCA1-long mRNAs. Additionally, we found that under all the conditions tested here PCA1-long mRNA are indirectly regulated by NMD, suggesting that there is an upstream factor regulated by NMD that regulates PCA1 mRNA expression.

2.6. Phylogenetic analysis of PCA1 genes

With the evidence of differential NMD regulation of PCA1 and PCA1 970G mRNAs in the presence of cadmium, we were interested in how the PCA1 gene evolved to gain insight into how these mechanisms of regulation emerged in yeast. We sequenced RM11–1a PCA1 970G and W303a PCA1 genes and conducted comparative genetic analysis of the PCA1 genes across 165 laboratory, domesticated, clinical and natural yeast strains (Sardi et al., 2018). Comparison of RM11–1a and W303a PCA1 sequences revealed a single nucleotide difference at the 970 codon associated with response to cadmium (RM11–1a–970G and W303a–970R). Analysis of all the PCA1 sequences revealed the 970G allele that facilitates tolerance to cadmium was the most common allele (Figure 7). Phylogenetic analysis revealed that 970R was the ancestral allele and while 970G evolved from that ancestral lineage (Figure 7). In addition, whether a yeast strain had the 970G or 970R alleles was not dependent on whether the strains are laboratory, domesticated, clinical, or natural strains.

FIGURE 7.

Maximum-likelihood tree of PCA1. Final ML Optimization Likelihood: –11952.815248. Only bootstrap values above 70 are shown. Black taxa are of undetermined origin

3. DISCUSSION

We found that the response of three genetically distinct yeast strains to bio-metals and toxic cadmium in the environment differs. Additionally, regulation of PCA1 and PCA1 970G mRNAs by NMD differs in the presence of cadmium but not excess copper. Moreover, lack of PCA1 affects growth of yeast strains in the presence of cadmium. Furthermore, wild-type and NMD mutant W303a, RM11–1a, and BY4741 yeast strains have comparable growth on CM and media containing 100 μM copper. Increasing the amount of copper to 600 μM results in NMD mutants being more tolerant of copper relative to wild type strains.

The RM11–1a strain is known to tolerate cadmium, whereas growth of both W303a and BY4741 is inhibited by cadmium. The wild-type and NMD mutant W303a strain appears to grow at comparable rates in the presence of cadmium, whereas BY4741 wild-type is more tolerant of cadmium than the NMD mutant, but deletion of PCA1 reverses this phenotype (Figure 2b). This observation demonstrates that PCA1 plays a role in cadmium tolerance but there are other compensatory factors that are regulated by NMD that allow the NMD mutant to be more tolerant of cadmium in the absence of PCA1. NMD is functional when yeast strains are grown on media containing cadmium as indicated by CYH2 pre-mRNA degradation. Furthermore, there was no discernible difference in growth between the RM11–1a wild-type and NMD mutants on media containing cadmium (Figure 1b). This observation suggests that the NMD pathway plays a role in homeostatic mechanisms of nutritional metals such as copper and iron but not to the same extent in non-essential, toxic environmental contaminants such as cadmium. This regulation appears to be yeast strain dependent. It could be that yeast strains found in diverse environments encounter environmental toxins that are not encountered by yeast strains maintained in laboratories and evolve to respond to their environment. Therefore, growth of RM11–1a in cadmium may reflect an adaptation that enable the strains to persist in their environment.

Regulation of PCA1 and other mRNAs involved in biometal homeostasis by NMD is dependent on the amount of metal ion in the environment. Conversely, the regulation of PCA1 970G by NMD is dependent on both copper and cadmium. This sort of regulation would allow yeast cells to respond rapidly to copper and cadmium levels at the mRNA level. It is critical for organisms to promptly respond to changes in their external environment because the environment is rapidly changing and cells must respond appropriately in order to survive.

Regulation of steady-state accumulations levels of PCA1 mRNA by NMD changes based on the amount of copper, specifically copper deplete conditions. Because the Pca1p present in most laboratory strains can confer copper tolerance in the ACE1 deletion background through binding and sequestering copper, we would anticipate that under low copper conditions PCA1 mRNA would be rapidly degraded by NMD and not accumulate to high levels. Because under low copper conditions, a reduced amount of Pca1p is necessary due to the deficiency of copper and would in fact be detrimental if expressed at high levels and sequesters the already scarce copper. In contrast, PCA1 mRNA was degraded at the slowest rate in wild-type W303a yeast strains grown in rich media and under 100 μM copper (Figure 6b), conditions where higher levels of Pca1p would be beneficial if required to sequester excess copper. Unexpectedly, the mRNA was degraded rapidly under 600 μM copper in NMD mutants. However, all three NMD mutant strains tested had higher tolerance to 600 μM copper relative to the wild-type strains (Figures 1a and 2a). Additional mRNAs regulated by NMD encode proteins that could play a role in the tolerance to copper at 600 μM copper. The tolerance of NMD mutants to 600 μM copper maybe attributable to other NMD regulated mRNAs that confer tolerance to copper like CTR2 and CRS5 (M. Peccarelli et al., 2014; M. Peccarelli et al., 2016).

Regulation of PCA1 and PCA1 970G mRNAs by NMD under some conditions is distinct. Steady-state accumulation levels of both PCA1 and PCA1 970G vary based on cadmium levels. This differential regulation of the mRNAs by NMD when yeast strains are grown in media containing cadmium is specific to cadmium because we did not observe differential regulation under high copper. This regulation is possibly due to a few factors: first, the induction of PCA1 970G by cadmium and, second, functionality of the proteins encoded by both mRNAs. As stated above, PCA1 encodes a protein that is impaired in cadmium detoxification due to a change in the ATPase domain. We expect that yeast strains encoding this allele of PCA1 are sensitive to cadmium and that the mRNA is regulated by NMD in the presence of cadmium. Furthermore, BY4741 which also encodes this PCA1 allele is also sensitive to cadmium. Conversely, PCA1 970G encodes a protein that can extrude cadmium, contributing to cadmium tolerance of yeast strains that contain this allele (Figure 1b). In growth media containing cadmium, the PCA1 970G mRNA is induced, expressed at higher levels and immune to NMD (Figure 4b). Thus, under conditions where more functional Pca1p is required, the PCA1 970G mRNA escapes NMD mediated regulation. Furthermore, this study and previous observations show that PCA1 970G expression is regulated by cadmium at both the mRNA and protein level (D.J. Adle & Lee, 2008). Multiple levels of regulation are beneficial and would allow precise and rapid response of yeast cells to cadmium levels in the environment because it is a toxic metal. The immunity to NMD-mediated regulation of PCA1 970G is physiologically significant to yeast cells because this correlates with RM11–1a tolerance to cadmium (Figure 1b). Furthermore, we have observed differential NMD-mediated regulation of mRNA involved in iron homeostasis based on iron availability (M. Peccarelli et al., 2019). The FRE2 mRNAs encode a ferric and cupric reductase and are regulated by NMD under normal growth conditions but not under low iron conditions (M. Peccarelli et al., 2019). Altogether, these observations suggest that NMD-mediated regulation of natural mRNAs that are responsive to environmental conditions depends on the following: (1) if the growth conditions are inducing or non-inducing and (2) the function of the encoded protein in the specific growth conditions. In future studies, we will investigate how PCA-970G mRNA escapes NMD-mediated degradation in the presence of cadmium and why the PCA1 is still sensitive to degradation.

The wide distribution of the 970G across the majority of the yeast strains in this study suggests the change leads to a widespread adaptation resulting in cadmium tolerance. The results of the phylogenetic analysis are consistent with this notion, as the most basal strains all carry the 970R and the majority of the later strains carry the 970G (bootstrap = 100). We confirmed the RM11–1a and W303a carry the 970G and 970R alleles, respectively, as previously observed (Figure 7). Given that both the W303a and RM11–1a demonstrate NMD-mediated regulation of mRNA involved in metal homeostasis and detoxification but only the RM11–1a carries the cadmium tolerance PCA1 allele, it is likely that this resistance allele encodes an mRNA that has evolved a way to counter NMD-mediated regulation when necessary, affording phenotypic plasticity. It is important to note that there are likely other genetic factors that contribute to cadmium tolerance, and the selective pressure for resistance may vary across different environments contributing to genetic variation at the PCA1 and other genes relevant to bio metal tolerance and toxic metal detoxification. There is evidence that in some environments, the cadmium resistance allele may incur a loss of fitness, given that a more recent strain (YJM269) does not carry the allele. Phylogenetic analysis reveals a high degree of PCA1 diversity among the strains that carry the 970G. Additional metal toxicity studies on some of the other strains representing the major clades and the YJM269 could provide additional confirmation of the role NMD plays in PCA1 expression and cadmium detoxification. Notably, most of the studies on NMD and regulation of gene expression at the mRNA level have been done on a small number of laboratory yeast strains. Thus, it is important to examine yeast strains from diverse background to understand how environmental conditions impact regulation of natural mRNAs by NMD.

4. MATERIALS AND METHODS

4.1. Yeast strains

Yeast strains used in this study are listed in Table 3. RM11–1a was kindly provided by Dr. Jaekwon Lee (Department of Biochemistry, University of Nebraska-Lincoln, NE). BKY111 was constructed from RM11–1a by a one-step gene disruption. BY4741 pca1△ upf1△ was also constructed from BY4741 pca1△ by a one-step gene disruption. A PCR fragment containing upf1-△2 (UPF1::URA3) was used to replace UPF1 in RM11–1a and BY4741 pca1△, respectively. The gene disruption was confirmed by PCR and by accumulation of CYH2 pre-mRNA on steady-state northern blots.

TABLE 3.

Saccharomyces cerevisiae strains used in this study

Yeast Strain	Genotype	Source
W303a	a, ade2–1 ura3–1 his3–11,15 trp1–1 leu2–3,112, can1–101	Wente et al. (1992)
AAY320	a, ade2–1 ura3–1 his3–11,15 trp1–1 leu2–3,112 can1–100 UPF1::URA3 (upf1-Δ2)	B. Kebaara et al. (2003)
AAY334	a, ADE2 ura3–1 or ura3–52 his3–52 his3–11,15, trp1–1 leu2–3,112 rpb1–1	B. Kebaara et al. (2003)
AAY335	a, ADE2 ura3–1 or ura3–52 his3–52 his3–11,15, trp1–1 leu2–3,112 rpb1–1, upf1-Δ2 (URA3)	B. Kebaara et al. (2003)
BY4741	a, his3D1, leu2D0, met15D0, ura3D0,	Transomics technologies (http://www.transomic.com/Home.aspx)
BY4741upf1Δ	a, his3D1, leu2D0, met15D0, ura3D0, URA3 (upf-Δ2)	Transomics technologies (http://www.transomic.com/Home.aspx)
BY4741 pca1Δ	a, his3Δ1 leu2Δ0 met15Δ0ura3Δ0 PCA1::KanMX4	Transomics technologies (http://www.transomic.com/Home.aspx)
BY4741 pca1Δ upf1Δ	a, his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 PCA1::KanMX4 URA3 (upf-Δ2)	This study
RM11–1a	a leu2Δ0 ura3Δ0 HO::KanMX6	D.J. Adle et al. (2007)
BKY111	a leu2Δ0 ura3Δ0 HO::KanMX6 UPF1::URA3 (upf1-Δ2)	This study

4.2. Growth under low copper and Iron

For growth under low copper conditions, yeast strains were grown in low copper complete minimal media. To attain low copper conditions, the media contained yeast nitrogen base without copper and iron (YNB-CuSO₄-FeCl₃) and 100 μM Bathocuproinedisulfonic acid (BCS; Sigma-Aldrich). To achieve low iron conditions, the media was prepared as described above for low copper replacing 100 μM BCS with 100 μM Bathophenanthrolinedisulfonic acid (BPS; Sigma-Aldrich). Glassware used in these experiments was soaked in 10% nitric acid overnight to remove trace amounts of copper and iron. All yeast cells used for low copper and iron were initially grown to saturation in complete minimal media then sub-cultured into copper or iron deficient media in acid washed glassware.

4.3. Growth under high copper

To grow W303a, BY4741, and RM11–1a wild type and NMD mutant strains under high copper conditions, the cells were grown in complete minimal media supplemented with 100 or 600 μM copper (high copper media). As with the low copper conditions, the yeast cells were first grown to saturation in complete minimal media then sub-cultured into media containing either 100 or 600 μM copper.

4.4. Growth in cadmium containing media

To culture W303a, BY4741, and RM11–1a wild type and NMD mutant strains under cadmium containing media, the yeast cells were grown in complete minimal media containing 100 μM cadmium. As with the low copper, low iron, and high copper conditions, the yeast cells were first grown to saturation in complete minimal media and then sub-cultured into media containing 100 μM cadmium.

4.5. RNA methods

Total S. cerevisiae RNA was used for all mRNA steady-state accumulations and half-life northerns. Yeast cells were grown in the different conditions described above. The cells were then harvested at mid-log phase as described in Peccarelli and Kebaara (2014). Total RNA was extracted from harvested cells using the hot phenol method. A total of 15 or 30 μg of total RNA was run on 1.0% agarose-formaldehyde gels for all steady-state and half-life northerns. The RNA was then transferred to GeneScreen Plus^® (PerkinElmer, Boston, MA) nylon membranes using the NorthernMax^™ Complete Northern Blotting kit (Thermo Fisher Scientific, Carlsbad, CA) transfer protocol. Northern blots were probed with oligolabeled DNA probes that were labeled with [α-³²P] dCTP using the RadPrime DNA Labeling System (Thermo Fisher Scientific, Carlsbad, CA). PCA1, CUP1, CTR1, CYH2, and SCR1 probes were generated by PCR. Northern blots were phosphorImaged^™ using a Typhoon Phosphorimager (GE Healthcare.)

4.6. Phylogenetic analysis

Phylogenetic analysis was performed using a maximum-likelihood approach. PCA1 sequences were provided by Sardi et al. (2018), and sequences were also generated for the yeast strains used in the present study. S. paradoxus was included as an outgroup to root the tree. Alignments of the DNA sequences were generated using MAFFT Version 7 (Folmer et al., 1994) and ragged ends trimmed in Mesquite 3.51 (Maddison, 2008). Phylogenetic relationships among the strains were inferred using RAxML. We utilized the GTRGAMA options with a thousand replications performed with the strategy of searching for the heuristically-best-scoring tree and generating bootstrap values for evaluation of support for the tree. Best scoring trees under ML with bootstrap values from RAxML were viewed and rooted under the outgroup criterion in FigTree (Rambaut, 2007).

4.7. Statistical analysis

Tests for significance were done using a two tailed t test. All experiments were done in triplicate unless otherwise stated. Significance was defined by p values determined from the t test. P < 0.05 was considered a significant difference in mRNA steady-state levels and half-life comparisons.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.