A kinase cascade on the yeast lysosomal vacuole regulates its membrane dynamics: conserved kinase Env7 is phosphorylated by casein kinase Yck3

Abstract

Membrane fusion/fission is a highly dynamic and conserved process that responds to intra- and extracellular signals. Whereas the molecular machineries involved in membrane fusion/fission have been dissected, regulation of membrane dynamics remains poorly understood. The lysosomal vacuole of budding yeast (Saccharomyces cerevisiae) has served as a seminal model in studies of membrane dynamics. We have previously established that yeast ENV7 encodes an ortholog of STK16-related kinases that localizes to the vacuolar membrane and downregulates vacuolar membrane fusion. Additionally, we have previously reported that Env7 phosphorylation in vivo depends on YCK3, a gene that encodes a vacuolar membrane casein kinase I (CKI) homolog that nonredundantly functions in fusion regulation. Here, we report that Env7 physically interacts with and is directly phosphorylated by Yck3. We also establish that Env7 vacuole fusion/fission regulation and vacuolar localization are mediated through its Yck3-dependent phosphorylation. Through extensive site-directed mutagenesis, we map phosphorylation to the Env7 C terminus and confirm that Ser-331 is a primary and preferred phosphorylation site. Phospho-deficient Env7 mutants were defective in negative regulation of membrane fusion, increasing the number of prominent vacuoles, whereas a phosphomimetic substitution at Ser-331 increased the number of fragmented vacuoles. Bioinformatics approaches confirmed that Env7 Ser-331 is within a motif that is highly conserved in STK16-related kinases and that it also anchors an SXXS CKI phosphorylation motif (₃₂₈SRFS₃₃₁). This study represents the first report on the regulatory mechanism of an STK16-related kinase. It also points to regulation of vacuolar membrane dynamics via a novel Yck3–Env7 kinase cascade.

In eukaryotic cells, organelles modulate their shape, size, copy number, and contents through highly regulated fusion and fission events (reviewed in references 1–4). Organellar fusion/fission equilibria drastically shift during biogenesis and inheritance as well as in response to changes in the intracellular and extracellular milieu of the cell. A model system in seminal studies of organelle biogenesis and membrane fusion/fission has been the lysosomal vacuole of Baker's yeast Saccharomyces cerevisiae (reviewed in references 5–7). Yeast vacuoles are functionally analogous to mammalian lysosomes; both are highly dynamic organelles responsible for maintaining cellular homeostasis (reviewed in references 8–10). Both organelles contain resident acidic hydrolases through which they carry out degradation and recycling of cargo, receptor downregulation, stress survival functions, and autophagy (reviewed in references 11 and 12). Whereas defects in lysosomal function and trafficking have been associated with diseases for many years, most recently, defects specifically in lysosomal fusion dynamics have been implicated as an underlying mechanism in lysosomal storage disease and Alzheimer's disease pathologies as well as cellular aging (reviewed in references 3, 13–16). Yeast lysosomal vacuoles have served as a productive model for fusion studies, as they are prominent dynamic landmarks that constitute more than 25% of the cell volume and undergo controlled fusion and fission in response to external and internal stimuli (reviewed in references 17–20). Vacuoles fragment under hyperosmotic stress as an adaptive response to maintain osmotic balance (21–24). Vacuole fusion/fission equilibrium is also regulated during cell cycle progression and autophagy (reviewed in references 25–27). Through in vivo vacuolar morphology studies and in vitro homotypic vacuole fusion assays, many components of vacuole fusion have been identified, and fusion machinery has been molecularly dissected into priming, tethering, docking, and bilayer mixing stages (reviewed in references 28–32). Whereas the machinery of vacuolar membrane fusion has been well dissected, the regulation of membrane fusion/fission dynamics remains poorly understood. Yeast vacuolar casein kinase 3 (Yck3) was the first vacuolar protein kinase shown to be implicated in the regulation of homotypic fusion (33). Yck3 inhibits vacuolar membrane fusion by phosphorylation of at least two proteins involved in vacuolar fusion, Vps41, which is a component of homotypic fusion and vacuole protein sorting (HOPS) complex involved in membrane tethering (33–38), and Vam3, a vacuolar Q-SNARE essential for homotypic fusion (21, 39). Yck3-dependent phosphorylation of the HOPS complex has also been implicated in conferring guanine nucleotide specificity of a vacuolar fusion Rab G-protein, Ypt7p (40, 41). Yck3 has also been found necessary to phosphorylate Mon1p, a component of autophagy-related cytoplasm to vacuole targeting pathway that is released from the vacuolar membrane during membrane fusion (42). These findings support the complex modulation of vacuolar fusion/fission equilibrium directly and indirectly through phosphorylation events.

We uncovered ENV7 in a genomic screen (43) and have identified its protein product, Env7, as a second protein kinase involved in vacuolar membrane dynamics (44). Env7 is a palmitoylated vacuolar membrane protein kinase with a function similar yet nonredundant to that of Yck3 in negative regulation of membrane fusion; it has a human ortholog and belongs to the underdefined family of STK16-related kinases (44, 45). We have also shown that native Env7 is phosphorylated in cells in a YCK3-dependent manner (46). Furthermore, we reported strong negative genetic interactions between ENV7 and YCK3 as the double deletion mutant exhibits severely perturbed cell fitness, budding, and vacuolar morphology (46). These results indicate that Yck3- and Env7-dependent vacuolar membrane flux is essential to normal cell physiology. In this study, we investigated whether Env7 is a direct substrate of Yck3, as both are physically localized to the vacuolar membrane (44, 47). We show that Env7 phosphorylation levels are dependent on Yck3 levels, that Yck3 and Env7 can physically interact, and that Yck3 can directly phosphorylate Env7. We also map the Yck3-dependent phosphorylation to the C terminus of Env7 and to a casein kinase I substrate phosphorylation motif, where S331 is the primary and preferred phosphorylation site. Lastly, we establish that the phosphorylation state of Env7 affects its localization and cellular function in regulating vacuolar membrane dynamics.

Results

Phosphorylation levels of native Env7 are dependent on Yck3 levels

We have previously identified that Env7 is a conserved vacuolar membrane protein kinase with a function similar but nonredundant to Yck3 in negative regulation of membrane fusion (44). We have also shown that native Env7 is not phosphorylated in yck3Δ cells (44, 46). In vivo, Env7 phosphorylation can be assessed based on Env7 phosphatase-sensitive upshift on a low-percentage SDS gel (8–10%), as reported by us previously (46) and presented here (Fig. 1A). Based on our gel mobility upshift assay, WT Env7 phosphorylation is at 40–55% depending on various conditions, including growth phase and media. We first attempted to confirm if phosphorylation levels of native Env7 are dependent on Yck3 levels. We transformed env7Δ and yck3Δ cells with CEN pRS316-ENV7 and CEN pRS316-YCK3, respectively. These plasmids express Env7 or Yck3 at native levels. We then analyzed vacuole-enriched membrane fractions (P13) for Env7 phosphorylation by Western blotting using anti-Env7 antibodies. Our data confirm that exogenously expressed Env7 is phosphorylated in the presence of native Yck3 in env7Δ cells (Fig. 1B, left), and that exogenously expressed Yck3 restores phosphorylation of native Env7 in yck3Δ cells (Fig. 1B, right). To explore if Env7 phosphorylation is Yck3 dose dependent, we transformed yck3Δ yeast cells with plasmids that overexpress either N- or C-terminally HA-tagged Yck3 from a constitutive PGK promoter and analyzed P13 fractions for phosphorylation of Env7. Overexpressed Yck3 resulted in increased phosphorylation of native Env7 independent of the HA-tag orientation (Fig. 1C). Similar effects were also observed with vacuole fractions isolated from these cells by Ficoll density gradient floatation (S. P. Manandhar, unpublished observations). These results confirm that phosphorylation of Env7 is Yck3 dose dependent. N-terminally HA-tagged Yck3 levels were consistently 5- to 10-fold less than those of WT or C-terminally tagged constructs, as seen in Fig. 1C. We have not yet explored the cause or the nature of decreased protein levels but proceeded to use C-terminally tagged Yck3 in subsequent studies.

Figure 1.

Phosphorylation levels of native Env7 are dependent on endogenous and exogenous Yck3 levels. A, upshifted Env7 migration is phosphatase sensitive. Vacuolar enriched P13 membrane fraction was isolated from env7Δ cells expressing Env7 from exogenous plasmid and was treated without and with phosphatase (PPase) and analyzed by Western blotting using polyclonal antibodies against Env7. B, exogenously expressed Yck3 phosphorylates Env7. P13 membrane fractions from indicated strains expressing native levels of Env7 or Yck3 from plasmids were analyzed by Western blotting, as described for panel A. Blots were stripped and reprobed for ALP as a loading control (lower) with an anti-ALP mAb (*, a partially degraded product of ALP). C, Env7 phosphorylation levels are dependent on Yck3 levels. P13 membrane fractions from indicated strains expressing either chromosomal Yck3 (WT) or exogenously overexpressed HA-tagged Yck3 were analyzed, as described for panel A. The intensity of upshifted phosphorylated and unphosphorylated bands of Env7 in each lane was quantified densitometrically using ImageJ (NIH) and expressed as percent upshift as described in Experimental procedures. Spliced lanes are indicated by a solid line.

Yck3 interacts with and directly phosphorylates Env7

We next tested whether Env7 is a direct substrate for Yck3. We first queried whether the two vacuole membrane kinases can physically interact by using a pulldown approach. For this inquiry, a triple cysteine-to-serine Env7 mutant (Env7-C13-15S-HA), which we have reported to be both palmitoylation and phosphorylation deficient, was used as a negative control (44). A small pool of Env7-C13-15S-HA remains associated with P13 membranes, and we hypothesized that such a membrane-associated pool would be defective in interactions with Yck3. Vacuolar membrane-enriched P13 fractions prepared from cells overexpressing WT Env7-HA or Env7-C13-15S-HA were incubated with bacterially expressed and purified Yck3-His, followed by centrifugation. The membrane pellet was analyzed by Western blotting using a monoclonal anti-HA antibody for Env7-HA and anti-His antibody for Yck3-His. Our results show that WT Env7-HA can pull down bacterially expressed and purified Yck3-His, consistent with a direct physical interaction(s) between the two proteins (Fig. 2A).

Figure 2.

Yck3 interacts with and directly phosphorylates Env7. A, pulldown of Yck3-His by P13 membranes expressing WT Env7-HA. P13 membrane fractions from yeast cells expressing WT Env7-HA or phosphorylation-deficient Env7-C13-15S-HA were incubated with bacterially expressed and purified Yck3-His (1 μm) for 2 h and centrifuged to pellet membranes. Five-fold excess mutant P13 fractions were utilized in incubations to input equivalent amounts of WT and mutant Env7. Pellets were analyzed by Western blotting using monoclonal antibodies against HA and His tags for Env7-HA and Yck3-His, respectively. Blots were stripped and reprobed for ALP as a loading control (lower). B, bacterially expressed and purified Yck3-His phosphorylates yeast Env7 in vitro. P13 membrane fractions from WT and yck3Δ cells were incubated with or without bacterially expressed and purified Yck3-His (1 μm) for 1 h and analyzed by Western blotting using anti-Env7 polyclonal antibodies. Spliced lanes are indicated by a solid line. C, bacterially expressed and purified Yck3-His phosphorylates bacterially expressed and purified Env7-His in vitro. Bacterially expressed and purified Env7-His alone (first two lanes), Yck3-His alone (second two lanes), and Env7-His+Yck3-His with either untreated (third two lanes) or heat-killed (HK, fourth two lanes) Yck3-His were incubated at 30 °C for 90 min in the absence or presence of the ATP regeneration system. All Yck3-His reactions were stopped by addition of an equal volume of 2× Laemmli sample buffer. Samples were analyzed by Western blotting using phospho-(Ser/Thr)-specific polyclonal antibodies (bottom). Blots were stripped and reprobed for protein level using an anti-His mAb to detect Yck3-His (top) and Env7-His (middle). Percent phosphorylation was determined as described for Fig. 1. D, on vacuolar enriched membranes, native Yck3 phosphorylates native Env7 in cis and trans. P13 membrane fractions from indicated yeast strains were incubated at 30 °C for 90 min in specified combinations in the absence or presence of the ATP regeneration system and analyzed by Western blotting using anti-Env7 polyclonal antibodies. Env7 phosphorylation levels were quantified densitometrically as described for Fig. 1. Trans upshift (indicating phosphorylation) is shown by an arrow. Spliced lanes are indicated by a solid line.

We then asked whether Env7 is a substrate for Yck3 in vitro. For this, P13 fractions isolated from both WT and yck3Δ cells were tested for ATP-dependent upshift of native Env7 in the presence and absence of bacterially expressed and purified His-tagged Yck3, and Env7 was detected by Western blotting using anti-Env7 polyclonal antibodies (Fig. 2B). Excess of bacterially purified Yck3-His led to extensive upshifted Env7 species, indicative of hyperphosphorylation of Env7 in both WT and yck3Δ cells (Fig. 2B). Without the addition of purified Yck3-His, native levels of Env7 phosphorylation were seen in the WT background where native Yck3 is present, and no Env7 upshift was observed in yck3Δ samples. These data confirm that bacterially expressed and purified Yck3-His is functionally active and can lead to phosphorylation of membrane-associated Env7. They are also consistent with the Yck3 dose-dependent phosphorylation of Env7 seen in Fig. 1C.

We next tested whether bacterially expressed and purified recombinant Yck3-His directly phosphorylates bacterially expressed and purified Env7-His. Yck3 is a serine/threonine-protein kinase. Therefore, we used phospho-(Ser/Thr)-specific polyclonal antibodies commonly used for the detection of Ser/Thr phosphorylated proteins (48, 49). Briefly, we incubated bacterially expressed and purified Env7-His and Yck3-His individually or together in the absence or presence of an ATP-regenerating system, as described previously (34, 44), and analyzed samples by Western blotting using phospho-(Ser/Thr)-specific antibodies for the detection of phosphorylated Env7-His and anti-His mAb for detection of proteins. As shown in Fig. 2C, bacterially expressed and purified Env7-His is phosphorylated by bacterially expressed and purified Yck3-His. Heat-killed Yck3-His did not lead to detectable phosphorylation of Env7-His. These data indicate that Yck3-His directly phosphorylates Env7-His in vitro.

We also tested if Yck3 from the env7Δ vacuolar enriched membrane fraction can phosphorylate Env7 from the yck3Δ vacuolar enriched membrane fraction in trans interactions where both proteins are expressed at native levels. Membrane proteins have been known to interact in both cis and trans during membrane and vesicle fusion (50, 51). For this, we incubated P13 fractions isolated from WT, env7Δ, and yck3Δ backgrounds in various combinations (cis or trans) in the absence or presence of an ATP regenerating system and analyzed them for Env7 upshift. We observed a small but consistently reproducible percentage (∼13%) of Env7 upshift in trans when membranes isolated from env7Δ and yck3Δ backgrounds were used (Fig. 2D).

Env7 phosphorylation is at its C-terminal region, where a casein kinase I substrate motif resides

In a series of experiments (Fig. 2, B–D), we confirmed that Yck3 directly phosphorylates Env7. Now the question arises, which Env7 amino acid residues are phosphorylated? Yck3 is orthologous to casein kinase I and is a serine/threonine kinase that phosphorylates several other vacuolar membrane proteins, including Vam3, Vps41, and Mon1p (33–35, 42). We searched for possible sites of Env7 phosphorylation by Yck3. Most protein kinases are activated and regulated by phosphorylation events in their activation loop and in a GHI subdomain C terminus of the loop (52). Bioinformatics analysis and structural modeling (iTASSER software) of the 365-amino-acid Env7, using its orthologous STK16 crystal structure (53), predict an atypical activation loop at amino acid residues 233–269 and a GHI subdomain at amino acid residues 298–352. These findings directed us to C-terminal half of Env7 for regulatory modifications. In partial degradation products of Env7-HA, phosphorylated (upshifted) species are detected in various C-terminal fragments, including an ≈20-kDa C-terminal fragment of WT Env7-HA, but not in those of phosphorylation-deficient Env7-C13-15S-HA (Fig. 3A). This yielded valuable information to predict that the phosphorylation site is located near the C terminus of Env7. To test this, we created a truncated version of Env7 by introducing a stop codon that led to a truncated Env7 (1–304 aa). P13 fractions were isolated and analyzed for ATP-dependent upshift by Western blotting using anti-Env7 antibodies (Fig. 3B). C-terminally truncated Env7, shorter than full-length Env7 by ∼10 kDa, showed complete loss of upshift phosphorylation compared with full-length Env7. This confirms our prediction and maps Env7-phosphorylated residues to its C-terminal aa 305–365.

Figure 3.

Env7 phosphorylation is at its C-terminal region, where a putative casein kinase I phosphorylation motif resides. A, C-terminal fragments of Env7 remain phosphorylated. WT Env7-HA and phosphorylation-deficient Env7-C13-15S-HA cell lysates were analyzed by Western blotting in higher percent SDS gel to assess full-length and N-terminal degradation products using anti-HA mAb. Upshifted Env7 species, an indicator of phosphorylation, is marked for the smallest fragment by an arrow. B, C-terminally truncated Env7 is not phosphorylated. A truncated Env7 missing 59 amino acids from its C terminus was engineered by site-directed mutagenesis. P13 membrane fractions from the indicated yeast strains were analyzed by Western blotting using anti-Env7 polyclonal antibodies. C, conserved serines and a putative casein kinase phosphorylation motif reside in the C terminus of Env7. Eight specified phosphorylation prediction tools were used to predict potential phosphorylation sites of Env7. Percent conservation is based on MSA of 114 fungal orthologs of Env7p. Sites denoted with a ● are potentially phosphorylated by a member of the casein kinase family I or II.

To identify the phosphorylation sites on Env7, we searched several databases for Yck3 phosphorylation sites using bioinformatics tools. Within the activation loop, there are six potential phosphorylation sites, of which S236, T262, and Y265 were predicted as phosphorylation sites by at least 2 of 8 phosphorylation prediction programs used (NetPhos Yeast, PhosphoSitePlus, and Predikin). However, none of these residues are >40% conserved, and they reside N-terminal to our mapped Env7 region for phosphorylation sites; hence, they were not pursued. Several other putative phosphorylation sites were identified within the mapped C terminus, including S323, S328, S331, and S352. Among those, S331 is the only serine residue that is >90% conserved and is predicted to be phosphorylated by 5 of 8 programs (the above 3 programs, plus DISPHOS and PHOSIDA) (Fig. 3C). Interestingly, S₃₂₈-XX-S₃₃₁ is a predicted substrate motif for casein kinase I (CKI), whose gene is orthologous to YCK3 (54).

Env7 S331 is a primary phosphorylation site

The schematic of Env7 protein in Fig. 4A highlights two clusters of conserved Ser/Thr residues within the C-terminal stretch that is absent in the truncated and nonphosphorylated Env7 (depicted in blue). One is the highly conserved cluster of serine residues encompassing the CKI phosphorylation motif (yellow), and the other is a more C-terminal threonine and serine residue cluster outside the CKI motif (gray shaded). Based on the bioinformatics details described in Fig. 3C and the S₃₂₈XXS₃₃₁ CKI phosphorylation motif present in the first cluster, we predicted that Env7 S331 is a primary phosphorylation site and that its phosphorylation is YCK3 dependent. We used the gray-shaded Ser/Thr cluster (T351-S353) as a non-CKI motif control within the Env7 C terminus. Ser/Thr codons were mutated to nonphosphorylatable alanine codons by PCR-based site-directed mutagenesis. Initially, four Env7 mutants were generated and assayed, mutant at all six Ser/Thr codons (S323, S328, S331, T351, T352, and S353) spanning both clusters (sextuple mutant Env7-6×); mutant at three Ser codons in/near the CKI-motif (S323, S328, and S331); mutant at three distal Ser/Thr codons (T351, T352, and S353); and mutant at S331 codon only (Env7-S331A). Plasmids expressing WT or mutant Env7 at native levels from the ENV7 native promoter were transformed into env7Δ cells, and P13 membrane fractions were isolated and analyzed for upshift (Fig. 4B). Env7-6× showed complete loss of upshift, mimicking what is observed in yck3Δ cells in this and previous figures, as well as with C-terminally truncated Env7 in env7Δ cells. This result confirms that phosphorylation leading to upshift is indeed at one or more of the targeted six amino acid residues. Whereas the phosphorylation state of the Env7 nonmotif triple mutant (T351A-T352A-S353A) consistently exhibited little or no change from that of the WT based on persistent upshift in Western blots, phosphorylation of Env7 mutant at the CK-1 motif (S323, 328, 331A) was consistently abolished. This result narrowed the Yck3-dependent phosphorylation site to Env7 residues S323, S328, and/or S331. However, the Env7-S331A single mutant consistently showed WT Env7 levels of upshift, suggesting either more than one phosphorylation site within the CKI motif cluster or utilization of S323 and/or S328 as an alternative phosphorylation site(s) in the absence of S331.

Figure 4.

Env7 S331 is a primary Yck3-dependent phosphorylation site. A, mutagenized Ser/Thr residues at Env7 C terminus and their significance. Truncated version of Env7 with 304 amino acids (blue) is not phosphorylated. Env7 C-terminal Ser/Thr cluster (yellow) includes putative casein kinase I phosphorylation motif S₃₂₈XXS₃₃₁ (boldface letters, darker yellow). Serine residues within this cluster were mutagenized to alanine by site-directed mutagenesis. Another Ser/Thr cluster (gray) located C terminal to the phosphorylation motif was also mutagenized to alanine as a control. B, phosphorylation status of single, triple, and sextuple mutants of Env7 expressed at native levels. P13 membrane fractions from env7Δ or yck3Δ cells expressing native levels of WT or indicated Ser/Thr mutants of Env7 were prepared and analyzed as described for Fig. 1. Upshift is indicative of phosphorylation. Blots were stripped and reprobed for ALP as a loading control (lower). C, add-back of S323 and S331 to sextuple Env7-6× mutant restores its Yck3-dependent upshift/phosphorylation. P13 membrane fractions from env7Δ and yck3Δ cells expressing the indicated Env7 species from a plasmid were analyzed as described for Fig. 1. D, the double (S323A/S331A) mutant is phosphorylated. P13 membrane fractions expressing native levels of WT and indicated Ser/Thr mutants of Env7 were prepared and analyzed by Western blotting as described for Fig. 1. E, add-back of S323 and S331 to overexpressed sextuple Env7-6×-HA mutant restores its upshift/phosphorylation. P13 membrane fractions from env7Δ cells overexpressing indicated Env7-HA species were analyzed by Western blotting, as described for Fig. 1. F, S323 and S331 add-backs to sextuple mutant Env7-6×-HA undergo some Yck3-independent upshift/phosphorylation upon overexpression. P13 membrane fractions from yck3Δ cells overexpressing specified Env7-HA species were analyzed as described for Fig. 1.

To further dissect the involvement of S323, S328, and S331 in Env7 phosphorylation in the absence of alternative neighboring serines, we took an add-back approach. Each of the CKI motif cluster alanine codons (S→A) of the sextuple mutant were individually mutated back to the original serine codons (A→S). The add-back mutant plasmids were transformed into env7Δ and yck3Δ cells, and P13 membrane fractions were isolated from these cells and tested for upshift phosphorylation of native levels of Env7 species using anti-Env7 antibodies (Fig. 4C). Env7-5×-A331S was the only add-back that consistently fully restored upshift of sextuple mutant Env7 to or above WT levels in the env7Δ background. Env7-5×-A323S consistently resulted in partially restored upshift, whereas Env7-5×-A328S resulted in minimal restoration of upshift. Importantly, no upshifted Env7 species could be detected in the yck3Δ background with the same add-back mutants (Fig. 4C, right). Thus, observed upshift indicative of phosphorylation is strictly Yck3 dependent. These results are consistent with Env7 S331 being the primary Yck3-dependent phosphorylation site.

To examine if S323 and S331 are the only residues associated with phosphorylation of Env7, we created a double (S323A/S331A) mutant and assessed its phosphorylation state when expressed at native levels. The double mutant showed upshift phosphorylation at WT levels (Fig. 4D). It is difficult to fully interpret this result, as both the triple mutant at 323/328/331 and the A328S add-back mutant remained consistently defective in detected phosphorylation. The double mutant was reconfirmed by sequencing before and after phosphorylation assays. It may be that in the absence of serines at aa 323 and 33, alternative S328 detectible phosphorylation can only occur in the presence of a WT stretch of Ser/Thr residues at aa 351–353. Alternatively, S328 may be facilitating phosphorylation at the aa 351–353 residue(s) in the double mutant.

We also created mutants in Env7-HA-overexpressing plasmid (under PGK promoter control) and assayed them for upshift as before (Fig. 4E). Overexpressed sextuple mutant Env7-6×-HA showed a lack of detectable upshift similar to that of its native-level counterpart. Overexpressed Env7-5×-A323S-HA, Env7-5×-A328S-HA, and Env7-5×-A331S-HA partially, minimally, and fully restored upshift, respectively, consistent with what was seen with their native-level expression. However, upshift levels of A323S and A331S were consistently more than 20% higher than those observed at native levels. Overexpression of those two mutants in the yck3Δ background showed a fraction of detected upshift to be Yck3 independent (Fig. 4F). Thus, based on upshift levels as indicative of phosphorylation levels, although native levels of Env7 are phosphorylated at S331 and S323 in a Yck3-dependent manner, overexpressed Env7 undergoes some Yck3-independent phosphorylation at those residues. This may be because of detectable autophosphorylation and/or phosphorylation by other kinases when Env7 is overexpressed.

Env7 phosphorylation is required for proper folding and membrane localization

In a series of experiments, we were unable to extract overexpressed phosphorylation-deficient mutants from P13 membranes. We have previously shown that a palmitoylation-deficient Env7 mutant forms detergent-resistant aggregates and does not localize to the vacuole (45). Therefore, we hypothesized that the phosphorylation-deficient mutants found in P13 are misfolded, detergent-insoluble aggregates. To better understand the physical state and localization of these mutants, P13 fractions from WT, sextuple, and add-back mutants Env7-5×-A323S-HA, Env7-5×-A328S-HA, Env7-5×-A331S-HA, and Env7-5×-A353S-HA were solubilized in 1% Triton X-100 and centrifuged to separate the soluble (supernatant, S) and insoluble (residual pellet, RP) fractions. Fractions were then analyzed by Western blotting using an anti-HA antibody (Fig. 5A). Whereas WT Env7 and add-backs Env7-5×-A323S-HA and Env7-5×-A331S-HA were solubilized and transferred to the soluble fraction, phosphorylation deficient sextuple and add-back mutants Env7-5×-A328S-HA and Env7-5×-A353S-HA were mostly insoluble and remained in the residual pellet (Fig. 5A, upper). Alkaline phosphatase (ALP), a vacuolar marker, was solubilized in all samples, suggesting that the vacuolar integrity of these mutants remains intact (Fig. 5A, lower). Increasing the concentration of Triton X-100 to 2% did not improve solubility of these mutants (data not shown). To further examine if Yck3-dependent phosphorylation is a factor in proper folding of WT Env7-HA, we examined the localization patterns of both WT and sextuple Env7-HA in env7Δ and yck3Δ backgrounds as described above. The behavior/membrane association of the sextuple mutant is the same in both backgrounds; it is mostly aggregated in residual pellet (RP). Behavior of WT Env7-HA, however, is different in the yck3Δ strain (where it cannot be phosphorylated by Yck3) and forms a higher percentage of RP aggregates (35%) than that detected in the WT background (6%) (Fig. 5B). This is consistent with Yck3-dependent phosphorylation having a role in correct membrane association/folding of Env7. These results are based on overexpressed Env7, and WT overexpressed Env7 undergoes some Yck3-independent phosphorylation (Fig. 4F).

Figure 5.

Overexpressed phosphorylation-deficient Env7-HA mutant proteins mislocalize to detergent-insoluble fraction. A, P13 membrane fractions from env7Δ cells overexpressing WT and indicated mutant Env7-HA were solubilized in lysis buffer containing 1% Triton X-100 and centrifuged to obtain the soluble fraction (S) and residual pellet (RP) as described in Experimental procedures. All fractions, including the original P13 fractions (P13), were analyzed by Western blotting using anti-HA antibody (top). Blots were stripped and reprobed for ALP as a loading control (lower). B, P13 membrane fractions from env7Δ and yck3Δ cells overexpressing WT and sextuple mutant Env7-HA were analyzed by Western blotting using anti-HA antibody (top) and anti-ALP (lower) as described for panel A. C, equivalent amounts of 0–4% Ficoll floatation vacuolar fraction were resuspended in 2× Laemmli sample buffer and analyzed by Western blotting using anti-HA antibody (top). Blots were stripped and reprobed for ALP as a loading control (lower). A graphical presentation of densitometric quantification showing the ratio of Env7-HA to ALP is included. Spliced lanes are indicated by a solid line.

To confirm if aggregating mutants are indeed absent from vacuoles, we isolated vacuoles from cells overexpressing WT Env7-HA, Env7-6×-HA, and add-backs Env7-5×-A328S-HA and Env7-5×-A331S-HA using Ficoll gradient floatation as described previously (44) and analyzed the vacuolar fractions by Western blotting using anti-HA mAb. Our results demonstrate that WT Env7 and Env7-5×-A331S-HA mutant are mostly localized to the vacuolar fraction, whereas the phosphorylation-deficient Env7-6×-HA and Env7-5×-A328S-HA are not (Fig. 5C). Subsequent densitometric and quantitative analysis of the blots show a significant decrease in the vacuolar levels of these two mutants compared with those of the WT Env7 and Env7-A331S-HA add-back mutant (shown in graph). Thus, Env7 species that show phosphorylation are the ones that do not form insoluble aggregates and correctly localize to vacuoles. These data strongly support that phosphorylation is essential for Env7 proper folding and localization to vacuolar membranes.

Env7 phosphorylation at its C terminus affects its in vivo regulatory function in vacuolar membrane dynamics

We have previously shown that Env7 negatively regulates vacuolar fusion during hyperosmotic stress and budding, and that overexpression of Env7 leads to an increase in multilobed vacuoles (44, 46). We hypothesized that the Env7 phosphorylation state regulates this cellular function. To test this hypothesis, env7Δ cells overexpressing WT, truncated, or phosphorylation site mutant Env7-HA species were analyzed for vacuolar morphology using live-cell confocal microscopy (Fig. 6A). Consistent with our hypothesis, overexpressed C-terminally truncated Env7-HA, which showed loss of phosphorylation by gel upshift assay (Fig. 3B), resulted in a decrease in multilobed vacuoles relative to WT Env7 overexpression, indicating a defect in vacuolar fusion/fission equilibrium. This vacuolar morphology defect was also observed with overexpressed sextuple mutant Env7-6×-HA, which also had shown complete loss of phosphorylation (Fig. 4E). Individual add-back mutants Env7-5×-A328S and Env7-5×-A353S, which did not significantly restore phosphorylation of sextuple mutants (data not shown), also did not lead to WT levels of multilobed vacuoles. However, individual add-back mutants Env7-5×-A323S and Env7-5×-A331S, which partially or entirely restored the phosphorylation state of the sextuple mutant, respectively (Fig. 4E), led to multilobed vacuole levels that were similar to that seen with WT Env7 overexpression. These results implicate S323 and S331 phosphorylation in the fusion/fission regulatory role of overexpressed Env7 and show that Env7 S323 or S331 phosphorylation alone is sufficient for maintaining vacuolar fusion/fission dynamics when Env7 is overexpressed. These results are also consistent with correct localization and vacuolar membrane association of overexpressed Env7-5×-A323S-HA and Env7-5×-A331S-HA add-back mutants shown in Fig. 5.

Figure 6.

Env7 phosphorylation affects its cellular function in regulating vacuolar fusion/fission dynamics. Specified strains expressing WT, truncated, or phosphorylation site mutant Env7 either from an overexpressed plasmid promoter (A) or from an endogenous plasmid promoter (B–E) were grown to mid-log phase in SM-URA. For confocal microscopy, live cells were stained with vital dye FM4-64 to mark the vacuole. A, phosphorylation-deficient mutants of overexpressed Env7-HA are defective in regulation of vacuolar fusion/fission dynamics. The lower panels show the quantification of vacuolar morphology represented in the corresponding upper panels. Averages from at least four different experiments were compared using a chi-squared test. *, p < 0.05. Error bars signify standard deviation. B, Env7 phosphorylation or phosphomimetic change at S331 is essential for its regulatory function in vacuolar fusion/fission dynamics upon native level expression. Quantification and statistical analysis of vacuolar morphology are presented in the lower panel as described for panel A. C, phosphomimetic mutant Env7 expressed at native levels. Serine 331 (S331) in WT Env7 and alanine 331 (A331) in the sextuple Env7 mutant were replaced with aspartic acid (331D) and exogenously expressed from native promoter. P13 membrane fractions were isolated from the indicated background and analyzed by Western blotting using anti-Env7 polyclonal antibodies. Blots were stripped and reprobed for ALP as a loading control (lower). D, Yck3-dependent phosphorylation at S331 is necessary for Env7 regulatory function in vacuolar fusion/fission dynamics. env7Δ yck3Δ cells expressing control plasmid (no Env7), WT, or phosphorylation site mutants of Env7 at native levels were studied as described for panel A. Quantification and statistical analysis of vacuolar morphology are presented in the lower panel as described for panel A. E, GFP-tagged Env7 phosphomimetic mutant localizes to vacuoles, whereas phosphorylation-deficient mutant does not. Live env7Δ cells expressing native levels of GFP-tagged WT, phosphomimetic, or phosphorylation-deficient Env7-6× mutant from an endogenous plasmid promoter were subjected to confocal microscopy to visualize Env7-GFP localization as well as FM4-64-stained vacuoles. Quantification and statistical analysis of GFP-tagged Env7 localization are presented in the lower panel. Averages from at least two different experiments were compared using a chi-squared test. *, p < 0.05 compared with WT. Error bars signify standard deviations.

We next analyzed vacuolar morphology upon native-level expression of WT, sextuple, and the two add-back mutants that resulted in WT vacuolar morphology when overexpressed, Env7-5×-A323S and Env7-5×-A331S (Fig. 6B). Interestingly, native levels of only Env7-5×-A331S mutant led to WT patterns of vacuolar morphology with >25% multilobed vacuoles. Env7-5×-S323A, as well as Env7-6×, led to a statistically significant decrease in multilobed vacuoles relative to the WT when expressed at native levels. These results correlate with our findings on the phosphorylation state of those mutants, where the sextuple mutant expressed at native levels showed complete loss of Env7 phosphorylation and Env7-5×-A331S had a stronger restoration of phosphorylation than Env7-5×-A323S (Fig. 4C). As stated in a previous section, overexpression of the same two add-back mutants led to increased restoration of phosphorylation relative to their native expression levels; both mutants were also biologically active in cells upon overexpression. Thus, levels of phosphorylation of Env7 correlate with its cellular functionality, further implicating the importance of phosphorylation in the cellular fusion/fission regulatory function of Env7.

Our biochemical results implicate S331 as the primary phosphorylation site for Env7. Our microscopic studies show that only the add-back at A331S is sufficient for the WT profile of vacuolar fusion/fission dynamics when expressed at native levels. Lastly, S331 is the hypothesized phosphorylated anchor of the CKI substrate motif (S₃₂₈XXS₃₃₁). Therefore, we also investigated the cellular function of a phosphomimetic mutant at S331. We used site-directed mutagenesis again to create the phosphomimetic mutant Env7-S331D by replacing S331 with negatively charged aspartic acid to mimic phosphorylation. Additionally, A331D phosphomimetic mutation was introduced in the sextuple Env7, rendering Env7-5×-A331D. Recombinant plasmids expressing Env7 species at native levels were transformed into WT, env7Δ, yck3Δ, and env7Δ yck3Δ double deletion strains. P13 fractions were isolated, and Env7 species were analyzed by Western blotting using anti-Env7 polyclonal antibodies (Fig. 6C). In these experiments, WT chromosomal Env7 is present in WT and yck3Δ yeast strains and is, presumably, the phosphorylated portion of detected Env7 in the WT background. As expected, no Env7 phosphorylation upshift is detected in yck3Δ or env7Δ yck3Δ backgrounds, since Yck3 is not present. Importantly, no phosphorylated Env7-S331D or Env7-5×-A331D is detected in the env7Δ background. Thus, the presence of a phosphomimic at Env7 amino acid residue 331 abrogates secondary phosphorylation events seen with Env7-S331A mutant in the env7Δ background in our earlier experiments. This further confirms that Env7 S331 is the preferred site for Yck3-mediated phosphorylation. Consistent with Yck3-independent phosphorylation seen upon overexpression of Env7, overexpressed phosphomimetic mutant Env7-S331D-HA was phosphorylated in the yck3Δ background (data not presented). Phosphomimetic mutants were included in experiments directed at probing the role of Env7 phosphorylation in its cellular function. Phosphomimetic Env7-S331D, both as a single mutant as well as in the sextuple background, led to WT vacuolar morphology profiles under native-level expression (Fig. 6B). Thus, a phosphomimic and, hence, phosphorylation at Env7 S331 is sufficient to direct its vacuolar fusion/fission regulatory function in cells.

As our results implicate Env7 S331 as the phosphorylation site by Yck3, we further examined the cellular function of our various Env7 phosphorylation site mutants in the absence of Yck3. As our biochemical studies indicated Yck3-independent Env7 phosphorylation upon overexpression, we focused our efforts on native-level expression of Env7. The double mutant env7Δ yck3Δ has been described previously (46). env7Δ yck3Δ cells were transformed with plasmids that express WT and phosphorylation site mutant Env7 species at native levels, and transformed cells were analyzed microscopically for vacuolar morphology (Fig. 6D). Consistent with our previous report (46), the double mutant cells were enlarged and exhibited more than 40% multilobed vacuoles. Native expression of WT Env7 in the double mutant did not significantly increase the percentage of multilobed vacuoles. This is consistent with our results thus far, as without Yck3 as the upstream kinase, Env7 is not expected to be phosphorylated and is expected to be defective in its negative regulation of vacuolar fusion. As expected, native-level expression of add-back mutant Env7-5×-A331S, which exhibited strong Yck3-dependent phosphorylation in our biochemical assays (Fig. 4C) and restored cellular function in microscopic assays in presence of Yck3 (Fig. 6B), did not alter the vacuolar morphology phenotype of the double mutant. Env7-6× and Env7-5×-A323S mutants exhibited the same lack of cellular function as they did in the presence of Yck3 seen in Fig. 6B. Phosphomimetic single mutant Env7-S331D, however, significantly increased multilobed vacuoles by more than 10%, consistent with a biologically functional Env7. Together, these results confirm that phosphorylation at S331 is Yck3 dependent and is essential for the cellular function of Env7 in negative regulation of vacuolar fusion. The phosphomimetic mutant in the sextuple background did not lead to a vacuolar morphology profile that was significantly different from that of the control. Although we cannot readily explain this result, it would be consistent with an essential role at one or more of the remaining five targeted Ser/Thr residues in directing Env7 function in cells, especially in the absence of Yck3. Our results demonstrate that Yck3-dependent phosphorylation at S331 is essential for Env7's role as a negative regulator of vacuolar membrane fusion.

Lastly, we assessed localization of the phosphomimetic mutant by live-cell confocal microscopy of GFP-tagged Env7 species. Our biochemical studies established that the Env7 S331 residue, presumably through its phosphorylation, was essential for WT patterns of vacuolar membrane localization and absence of misfolding/aggregation (Fig. 5). In those assays, Env7-6× aggregated into detergent-insoluble precipitants and was not associated with the vacuolar fraction. We constructed plasmids expressing native levels of GFP-tagged WT Env7, phosphomimic Env7-S331D, and Env7-6× sextuple mutant and examined Env7 localization by live-cell confocal microscopy (Fig. 6E). Both WT Env7 and Env7-S331D showed similar levels of >55% localization to the vacuolar membrane, whereas 90% of Env7-6×-GFP was cytoplasmic. Taken together, these results indicate that Env7 phosphorylation is essential to its correct vacuolar localization as well as vacuolar membrane fusion/fission regulatory function.

Discussion

We previously identified Env7 as a conserved STK16-related yeast vacuolar membrane kinase implicated in negative regulation of vacuolar fusion; we also previously established that Env7 phosphorylation depends on the presence of another vacuolar membrane kinase gene involved in negative regulation of vacuolar fusion, the casein kinase I (CKI) homolog, YCK3 (33, 46). We establish here that Env7 is a substrate of Yck3 and is phosphorylated at a conserved consensus sequence. We also show that Yck3-dependent phosphorylation of Env7 is essential for its in vivo regulation of vacuole fusion dynamics as well as its correct localization.

Env7 Ser331 is a main phosphorylation target of Yck3 and anchors a conserved casein kinase substrate phosphorylation motif

We show that Yck3 directly phosphorylates Env7 and map phosphorylation to the C-terminal tail of Env7 (305–364 aa). Within that region, Env7 structural modeling using the orthologous human STK16 crystal structure (53) predicts a kinase GHI subdomain where many kinases are phosphorylated for regulation. The predicted Env7 GHI subdomain contains a CKI substrate motif, SXXS, at S₃₂₈-XX-S₃₃₁, where the C-terminal serine is predicted to be phosphorylated by CKI (55). Mutagenesis of serine and threonine codons in and around this motif at ₃₂₃-SFPRNSRFS-₃₃₁ and/or at a more distal cluster of Ser/Thr at ₃₅₁-TTS-₃₅₃, as well as specific singular add-back or phosphomimetic mutants, confirmed that Ser331 is a primary and preferred Yck3-dependent phosphorylation site.

The CKI substrate phosphorylation motif and the predicted phosphorylation site at Ser331 are highly conserved across STK16-related orthologues of Env7, as shown in Fig. 7. This observation suggests that phosphorylation by casein kinase is involved in regulating STK16-related kinases. Most interestingly, specific residues both N-terminal and C-terminal to the motif are highly invariant (Fig. 7, magnified inset). Ser323 is more invariant than Ser328 of the putative S₃₂₈-XX-S₃₃₁ tetra motif. This is consistent with our results implicating Ser323, but not Ser328, as the preferred alternative phosphorylation site to Ser331. The observed continued phosphorylation of the Env7 double mutant at Ser323 and Ser331, however, is consistent with a functional role for Ser328 in Env7 phosphorylation. One possibility is that Ser328 presence and/or its phosphorylation is necessary for priming of efficient phosphorylation at Ser331 and, in its absence, at other Ser/Thr residues in the Env7 C terminus.

Figure 7.

Multiple-sequence alignment of Env7 and its orthologues. MAFTA MSA of Env7 and a representative set of protein sequences from orthologous genes to ENV7. A conserved C-terminal region encompassing a proposed CK1 substrate phosphorylation motif is magnified as an inset. Species are listed in their evolutionary order. Amino acid residues are shaded based on their similarities as indicated.

Additionally, Env7 Pro325 and Leu333 are conserved to 100% similarity in the species queried, as is Env7 Ser331. Lastly, there is an invariant hydrophobic residue just before the conserved Pro325. Hence, the CKI phosphorylation consensus motif may be more extensive in STK16-related kinases than the SXXS motif proposed in reference 54. We propose a more extensive motif of SΘPXXXXXSXXL (Θ = hydrophobic residue) corresponding to Env7 ₃₂₃SFPxxSxxS₃₃₁xxL₃₃₄, where the underlined residues are the most conserved elements and the conserved serine flanked by invariant proline and leucine is the preferred phosphorylation site by CKI and its homologs in STK16-related kinases. We will be probing the significance of these motif components in the near future.

Yck3-dependent Env7 phosphorylation is essential for its proper localization and cellular function

C-terminally truncated Env7 and Env7-6× are phosphorylation deficient and defective in regulating vacuolar fusion/fission dynamics in vivo. Upon native-level expression in cells, only the single add-back mutant at Env7 A331S was able to fully restore phosphorylation state of sextuple mutant Env7-6× and lead to an increase in multilobed vacuoles to WT levels; both phosphorylation and cellular function were Yck3 dependent. The Env7 A323S add-back mutant partially restored phosphorylation of Env7-6× and did not restore cellular function to the sextuple mutant upon native-level expression. Env7 add-back mutants at A328S and A353S restored neither significant phosphorylation to Env7-6× nor cellular function. Lastly, phosphomimetic Env7 mutant at S331 led to WT levels of fusion/fission equilibria in the absence of any detectable additional phosphorylation. Taken together, these results directly implicate S331 Yck3-dependent phosphorylation as essential for the fusion/fission regulatory role of Env7 in cells.

Our current study also shows that Yck3-dependent Env7 phosphorylation is essential for its proper folding and localization. In biochemical assays, overexpressed phosphorylation-deficient mutants form detergent-resistant aggregates that are not localized to purified vacuoles. In live-cell imaging studies, the native level of phosphorylation-deficient sextuple mutant mislocalizes to the cytoplasm and is not associated with vacuolar membranes, whereas phosphomimetic Env7 shows WT patterns of vacuolar membrane localization. Phosphorylation has been known to modify protein conformation, catalytic properties, and interactions with partners, thereby mediating the effects of signaling pathways on cells (56, 57). It is possible that phosphorylation-deficient mutants are not correctly folded, form aggregates in the cytoplasm that cannot associate with membranes, and, as such, cannot direct vacuolar fusion/fission regulation.

A vacuolar membrane kinase cascade

Yck3 and Env7 are vacuolar membrane kinases involved in the negative regulation of vacuolar fusion (33, 44). Yck3 has several previously established downstream nodes, including Vps41, Vam3, and Mon1p (44–46, 58). Vps41 is a component of the homotypic fusion and vacuole protein sorting (HOPS) complex, involved in membrane tethering (33–36), and Vam3 is a vacuolar Q-SNARE essential for homotypic fusion (21, 39). Both are directly phosphorylated by Yck3, and their phosphorylation directly inhibits vacuolar membrane fusion. Mon1p, a component of the autophagy-related cytoplasm-to-vacuole targeting pathway, is released from vacuolar membranes during cytoplasm-to-vacuole targeting fusion events upon Yck3-dependent phosphorylation (42). This study adds Env7 to the known Yck3 substrates. Interestingly, Env7 is the only kinase identified to date as a Yck3 substrate. As such, it implicates a kinase cascade existing on the vacuolar membrane and involved in regulating vacuolar membrane dynamics. Although Env7 also negatively regulates vacuolar fusion, it does not phosphorylate the Yck3 substrate Vps41 or suppress YCK3 deletion when overexpressed (44, 46). We have not yet examined the role of Env7 in phosphorylation of the two other known substrates of Yck3. Based on the results at hand, Env7 may represent a distinct node with its own unidentified set of effectors. We are currently exploring both targeted candidates as well as proteomic approaches to identify Env7 substrates.

Lastly, a vacuolar kinase cascade regulating vacuolar membrane dynamics may include additional kinases as well as a combination of both upregulation and downregulation mediated by phosphorylation. Overexpressed Env7 undergoes Yck3-independent phosphorylation in cells, as we report here. We also detect hyperphosphorylation of Env7 in the absence of the kinase Tor1 (in tor1Δ cells) and following treatment with the Tor complex 1 (TORC1) inhibitor rapamycin. TORC1, a master regulator of cell growth and cell division, resides on vacuolar membranes; it also regulates autophagy, which is another vacuolar event accompanied by significant changes in vacuolar fusion/fission dynamics (58) (reviewed in references 55 and 59). Furthermore, TORC1 has been implicated in regulating vacuolar fragmentation (reviewed in reference 1). It is intriguing to speculate on a TORC1-modulated Yck3-Env7 kinase cascade at the vacuole membrane orchestrating intricate membrane dynamics during cell cycle progression and key stress survival responses.

Experimental procedures

Materials and media

Phusion and Taq DNA polymerases, all restriction enzymes, and Lambda protein phosphatase (Lambda PP) were purchased from New England BioLabs, Inc. (Beverly, MA). A quick-change mutation kit was supplied from Agilent Technologies (Santa Clara, CA). Oligonucleotides were ordered from Operon (Alameda, CA). Analytical-grade Tris, BSA, protease inhibitor cocktails, leupeptin, aprotinin, and other chemicals were purchased from Sigma (St. Louis, MO). All growth media were from Difco.

Yeast strains and growth media

Yeast strains used in this study are presented in Table 1. These are BY4742 (WT, MATα his3Δ1 leu2Δ0 lys2Δ0 MET15Δ0 ura3Δ0), BY4742/env7Δ, and BY4742/yck3Δ (MATa his3Δ1 leu2Δ0 lys2Δ0 MET15Δ0 ura3Δ0 env7Δ::KanMX4) (gifts from Greg Payne, UCLA, Los Angeles, CA). The env7Δ yck3Δ double deletion mutant used in this study was created by deleting YCK3 in the env7Δ background and has been described before (44). Cells were routinely grown in rich medium (yeast extract-peptone-dextrose [YPD]; 1% yeast extract, 2% peptone, and 2% glucose) or synthetic minimal dropout medium (SMD; 0.67% yeast nitrogen base, 2% glucose, and specific amino acids) at 30 °C unless otherwise stated.

Table 1.

Yeast strains used in this study

Strain	Genotype	Reference
BY4742 WT	MAT-α his3Δl leu2Δ0 lys2Δ0 ura3Δ0	63
BY4742 env7Δ	MAT-α his3Δl leu2Δ0 lys2Δ0 ura3Δ0 env7::KanMX4	63
BY4742 yck3Δ	MAT-α his3Δl leu2Δ0 lys2Δ0 ura3Δ0 yck3Δ::KanMX4	63
BY4742 env7Δ/yck3Δ	MAT-α his3Δl leu2Δ0 lys2Δ0 ura3Δ0 env7Δ::KanMX4 yck3Δ::LEU2	46

Antibodies

Anti-Env7 rabbit polyclonal antibodies were generated as described previously (44). Anti-hemagglutinin epitope (HA) mAb, anti-6xHis mAb, rabbit anti-HA antibody conjugated to Sepharose beads, HRP-conjugated secondary antibodies against mouse IgG, and phospho-(Ser/Thr)-specific rabbit polyclonal antibodies were purchased from Cell Signaling Technology (Danvers, MA). Anti-alkaline phosphatase mAb was supplied by Abcam (Cambridge, MA).

Site-directed mutagenesis and recombinant DNA approaches for the construction of Env7 phosphorylation mutants

Plasmids harboring single, triple, and sextuple putative phosphorylation site mutants and their add-backs to the WT codon in sextuple mutant Env7-6× or Env7-6×-HA were constructed by using a quick-change lightning site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA) and two-PCR-based site-directed mutagenesis (44). Serine codons were replaced with alanine codons at residues Ser³²³, Ser³³¹, Ser^323-328-331, and Thr^351-352S³⁵³ and the sextuple mutant (Ser^323-328-331; T^351-352S³⁵³). Phosphomimetic mutants were created by replacing serine or alanine at amino acid position 331 with aspartic acid (D) in WT Env7 and sextuple Env7-6×, respectively. A truncated Env7 was created by introducing a stop codon at a designated position (1–304 aa). Mutants were constructed for expression from native or overexpression promoters from plasmids. All plasmids used in this study are listed in Table 2, and the primers used to construct the plasmids are listed in Table 3. pRS316:ENV7 mutants were created either by site-directed mutagenesis or by subcloning from pSMG17 constructs. All plasmid constructs and mutagenesis were confirmed by DNA sequencing (Macrogen, Seoul, Korea).

Table 2.

Plasmids used in this study^a

Plasmid	Mutation introduced	Genotype	Reference or source
pMSG370	WT	CEN URA3 pRS316-ENV7	43
pMSG470	WT	2 μm URA3 pRS426-ENV7	43
pSMG44	WT	CEN URA3 pRS316-YCK3	This study
pSMG45	WT	2 μm URA3 PPGK-YCK3-3×HA	This study
pSMG46	WT	2 μm URA3 PPGK-3×HA-YCK3	This study
pSMG47	WT	pET24d(+)-YCK3-6×His	This study
pSMG34	WT	pET24d(+)-ENV7-6×His	44
pSMG17	WT	2 μm URA3 PPGK-ENV7-3×HA	44
pSMG19	WT	CEN URA3 PGAL1- ENV7-GFP	44
pSMG30	Env7-C13-15S-HA	2 μm URA3 PPGK-ENV7-C13-15S-3×HA	44
pSMG48	Env7-(1-304)-HA (truncated)	2 μm URA3 PPGK-ENV7-(1-304)-3×HA	This study
pSMG49	Env7-S331A-HA	2 μm URA3 PPGK-ENV7- S331A-3×HA	This study
pSMG50	Env7-S323A-S328A-S331A-HA	2 μm URA3 PPGK-ENV7-S323A-S328A-S331A-3×HA	This study
pSMG51	Env7-T351A-T352A-S353A-HA	2 μm URA3 PPGK-ENV7-T351A-T352A-S353A-3×HA	This study
pSMG52	Env7-6×-HA (sextuple mutant, overexpressed)	2 μm URA3 PPGK-ENV7-S323A-S328A-S331A-T351A-T352A-S353A-3×HA	This study
pSMG53	Env7-5×-A323S-HA (add-back)	2 μm URA3 PPGK-ENV7-A323S-S328-AS331A-T351A-T352A-S353A-3×HA	This study
pSMG54	Env7-5×-A328S-HA (add-back)	2 μm URA3 PPGK-ENV7-S323A-A328S-S331A-T351A-T352A-S353A-3×HA	This study
pSMG55	Env7-5×-A331S-HA (add-back)	2 μm URA3 PPGK-ENV7-S323A-S328A-A331S-T351A-T353A-353A-3×HA	This study
pSMG56	Env7-5×-A353S-HA (add-back)	2 μm URA3 PPGK-ENV7-S323A-S328A-A331S-T351A-T352A-353S-3×HA	This study
PSMG57	Env7-6× (sextuple mutant, native)	CEN URA3 pRS316-ENV7-A323S-S328A-S331A-T351A-T352A-S353A	This study
pSMG58	Env7-5×-A323S (add-back)	CEN URA3 pRS316-ENV7-A323S-S328A-S331A-T351A-T352A-S353A	This study
pSMG59	Env7-5×-A328S (add-back)	CEN URA3 pRS316-ENV7-S323A-A328S-S331A-T351A-T352A-S353A	This study
pSMG60	Env7-5×-A331S (add-back)	CEN URA3 pRS316-ENV7-S323A-S328A-A331S-T351A-T352A-S353A	This study
pSMG61	Env7-S331d-HA (overexpressed)	2 μm URA3 PPGK-ENV7-S331d-HA	This study
pSMG62	Env7-5×-A331D-HA (overexpressed)	2 μm URA3 PPGK-ENV7-S323A-S328A-A331D-T351A-T352A-S353A-3×HA	This study
pSMG63	Env7-S331D (native)	CEN URA3 pRS316-ENV7-S331D	This study
pSMG64	Env7-5×-A331D (native)	CEN URA3 pRS316-ENV7-A323S-S328A-A331D-T351A-T352A-S353A	This study
pSMG65	Env7-GFP (WT, native)	CEN URA3 pRS316-ENV7-GFP	This study
pSMG66	Env7-S331D-GFP (native)	CEN URA3 pRS316-ENV7-S331D-GFP	This study
pSMG67	Env7-5×-A331D-GFP (sextuple mutant, native)	CEN URA3 pRS316-ENV7-S323A-A328S-A331D-T351A-T352A-S353A-GFP	This study
pSMG68	Env7-S323A-S331A-HA	2 μm URA3 PPGK-ENV7-S323A-S331A-3×HA	This study

^aEnv7 amino acid change(s) introduced is shown in boldface. Overexpressed HA-tagged Env7 and Yck3 were expressed from yeast 2 μm URA3 P_PGK plasmids. Native levels of Env7, Yck3, and GFP-tagged Env7 were expressed from their native promoters in yeast CEN URA3 pRS316 plasmids.

Table 3.

Oligonucleotides used to construct plasmids in this study^a

Oligonucleotide and mutation	Sequence (5'→3')	Source
YCK3 FP	CGCGGTGGCGGCCGCTCTAGAACTAGTGGATCCCCCGGGCTGCAGtctgttccgtggcaaatgtttctcc	This study
YCK3 RP	CAAAAGCTGGGTACCGGGCCCCCCCTCGAGGTCGACGGTATCGATttacatattattttggatattcttaaaaaaatgc	This study
YCK3-HAc FP	CTACTTTTTACAACAAATCTAGAATTCCTGCAGCCCGGGGGATCCatgtcccaacgatcttcacaacac	This study
YCK3-HAc RP	CCCGCATAGTCAGGAACATCGTATGGGTAAAAGATGCGGCCCAGATCacagcaacaaaaacagcaacaac-3'	This study
HAn-YCK3 FP	CGCTGCTCAGTGCGGCCCAGATCTGTCCCAACGATCTTCACAACACATTGTAGG	This study
HAn-YCK3 RP	TAGTACATAGGAAATTTTAAACGGTATACAAGTACGTAAAAAAGGTCAACAGCAACAAAAACAGCAACAAC	This study
Truncated Env7 (1-304) FMP	gttcggtatctccccctttgagcgataagagcagatacatggagcttctttaacc	This study
Truncated Env7 (1–304) RMP	ggttaaagaagctccatgtatctgctcttatcgctcaaagggggagataccgaac	This study
Env7-S331A FMP	gagaaattccagatttgctgaggggcttttgag	This study
Env7-S331A RMP	ctcaaaagcccctcagcaaatctggaatttctc	This study
Env7-S323A-S328A-S331A FMP	cctacgctataaacactggtaagtacgctttcccgagaaatgccagatttgctgaggggcttttgagtgt	This study
Env7- S323A-S328A-S331A RMP	acactcaaaagcccctcagcaaatctggcatttctcgggaaagcgtacttaccagtgtttatagcgtagg	This study
Env7-T351A-T352A-S353A FMP	gcattcaagtggatcctatacaaaggcctgctgccgcccaattattaaatcttttacaagatttagac	This study
Env7- T351A-T352A-S353A RMP	gcttaaatcttgtaaaagatttaataattgggcggcagcaggcctttgtataggatccacttgaatgc	This study
Env7-5×-A323S FMP	cctacgctataaacactggtaagtacagtttcccgagaaatgccagattttctgagg	This study
Env7-5×-A323S RMP	cctcagcaaatctggcatttctcgggaaactgtacttaccagtgtttatagcgtagg	This study
Env7-5×-A328S FMP	ggtaagtacgctttcccgagaaattccagatttgctgaggggcttttgag	This study
Env7-5×-A328S RMP	ctcaaaagcccctcagcaaatctggaatttctcgggaaagcgtacttacc	This study
Env7-5×-A331S FMP	cgctttcccgagaaatgccagattttctgaggggcttttgagtgtaatcaag	This study
Env7-5×-A331S RMP	cttgattacactcaaaagcccctcagaaaatctggcatttctcgggaaagcg	This study
Env7-5×-A353S FMP	ggatcctatacaaaggcctgctgccagccaattattaaatcttttacaag	This study
Env7-5×-A353S RMP	cttgtaaaagatttaataattggctggcagcaggcctttgtataggatcc	This study
Env7-S331D FMP	cctacgctataaacactggtaagtacagtttcccgagaaattccagatttgatgaggggcttttgagtg	This study
Env7-S331D RMP	cactcaaaagcccctcatcaaatctggaatttctcgggaaactgtacttaccagtgtttatagcgtagg	This study
Env7-5×-A331D FMP	cctacgctataaacactggtaagtacgctttcccgagaaatgccagatttgatgaggggcttttgagtgt	This study
Env7-5×-A331D RMP	acactcaaaagcccctcagcaaatctggcatttctcgggaaagcgtacttaccagtgtttatagcgtagg	This study
Env7-S323A-S331A FMP	cctacgctataaacactggtaagtacgctttcccgagaaattccagatttgctgaggggcttttgagtgt	This study
Env7-S323A-S331A RMP	acactcaaaagcccctcagcaaatctggaatttctcgggaaagcgtacttaccagtgtttatagcgtagg	This study

^a Vector-specific sequences are capitalized, and ENV7-specific sequences are in lowercase. Mutations are indicated in boldface, and affected codons are underlined. FP, forward primer; RP, reverse primer; FMP, forward mutant primer; RMP, reverse mutant primer.

Subcellular fractionation and Western blot analysis

Subcellular fractionation of yeast cells was performed as described previously (44–46). Briefly, cells were grown in SM-URA to late log phase (A₆₀₀ = 1.0), spheroplasted with 40 units of zymolyase, lysed with 4 µg/ml DEAE-dextran in the presence of protease and phosphatase inhibitors, and centrifuged to yield P13, S100, and P100 fractions. Samples were analyzed by Western blotting using anti-Env7 polyclonal and anti-HA monoclonal antibodies. Blots were stripped and reprobed for vacuolar membrane alkaline phosphatase (ALP) as a loading control with an anti-ALP mAb, as described previously (44).

Ficoll gradient isolation of vacuoles

Vacuoles were isolated using four-tiered Ficoll gradient floatation as described previously (44, 60). The vacuolar layer (0–4%) was analyzed by Western blotting using anti-HA and anti-ALP (loading control) monoclonal antibodies, as described above.

Protein purification

Purification of His-tagged Env7 and Yck3 from E. coli cells was performed using Ni-NTA column chromatography as described previously (44).

Mobility shift assays for in vivo Env7 phosphorylation

Env7 phosphorylation in vivo was assessed based on mobility upshift of phosphorylated Env7 in 8–10% SDS gels. P13 membrane fractions were incubated without or with an ATP regeneration system (5 mm ATP, 1 mg/ml creatine kinase, 400 mm creatine phosphate, and 200 mm sorbitol in 10 mm Pipes buffer, pH 6.8) as described previously (44, 46) and analyzed by low-percentage (8–10%) SDS-PAGE followed by Western blotting using either anti-Env7 polyclonal or anti-HA monoclonal antibodies, used to minimize potential ATP depletion. For upshift phosphorylation, bands of each protein species were densitometrically quantified using ImageJ version 1.46 (National Institutes of Health). The extent of phosphorylation (upshifted species) was calculated as a percentage of the total (upshifted + nonupshifted species) for each protein and expressed as % upshift. In another set of experiments, a P13 fraction of exogenously expressed WT Env7 was treated with phosphatase (100 units) and analyzed by Western blotting, as previously described (46).

Kinase assays

Direct phosphorylation of bacterially expressed and purified Env7-His by bacterially expressed and purified Yck3-His was performed in vitro by incubating the recombinant proteins in the absence and presence of an ATP regeneration system at 30 °C for 90 min. The reaction was stopped by adding an equal volume of 2× Laemmli sample buffer. Phosphorylated proteins were analyzed by Western blotting using Ser/Thr-specific polyclonal antibodies (1:1000). Blots were then stripped and reprobed with His-tag mAb to determine total protein. The phosphorylation level of Env7 in each combination was then densitometrically quantified and expressed as percent phosphorylated, as described above.

Phosphatase assay

Phosphatase assay was performed according to the standard protocol provided by the company (NEB Biolabs), with some modifications. Briefly, 100 µl of P13 membrane fractions was resuspended in lysis buffer that includes 1× NEB buffer and 1 mm MnCl₂. The reaction mixture was incubated without or with Lambda protein phosphatase (100 units) at 30 °C for 1 h. The reaction mixture was then centrifuged at 13,000 × g for 15 min to pellet the membranes and analyzed by Western blotting using anti-Env7 polyclonal antibodies as described previously (46). The upshift phosphorylation was densitometrically quantified as described above.

Detergent solubilization of P13 membrane fractions

To assess the detergent solubility of specific Env7-HA mutants, P13 membrane fractions were treated with 1–2% Triton X-100 for 1 h at 4 °C with shaking and subjected to centrifugation to yield the soluble fraction (supernatant, S) and residual membranes (pellet). The pellet was resuspended in lysis buffer and adjusted to its original volume. Equal volumes of original P13, residual pellet (RP), and supernatant were analyzed by Western blotting using anti-HA mAb as described above. For percent Env7-HA localized, band intensity of RP and S for each protein was densitometrically quantified using ImageJ version 1.46 (National Institutes of Health). The extent of localization was calculated as a percentage of the total (RP + S) for each protein species and expressed as percent localized.

Phosphorylation site prediction

Eight different programs were used, as specified in Fig. 3, to find amino acid residues in Env7 that possess a good potential for phosphorylation. Default settings were chosen for each prediction method, and S. cerevisiae was specified as the target organism. If a putative phosphorylation site was only predicted by a single program, it was precluded from further analysis. The resulting proposed phosphorylation sites were fashioned into a table and visually manipulated with Adobe Photoshop CS5 (61).

Microscopy

Vacuolar morphology and Env7-GFP localization were analyzed by confocal microscopy. For vacuolar morphology studies, env7Δ, yck3Δ, or env7Δ yck3Δ cells expressing either WT or mutant Env7 from an endogenous (untagged Env7) or constitutive (Env7-HA) plasmid promoter were grown to an optical density at 600 nm of 0.8–1.0 in SM-URA. Cells were then transferred to YPD medium and stained with the vital vacuolar dye FM4-64, as described previously (44). Live cells were viewed by confocal microscopy at 1,800× magnification and scored for prominent (1–3 vacuoles/cell) versus multilobed (>3 vacuoles/cell) vacuoles. For localization studies of GFP-tagged Env7, env7Δ cells transformed with plasmids expressing either GFP-tagged WT or mutant Env7 from an endogenous promoter were grown to an optical density at 600 nm of 0.8 in SM-URA, stained with FM4-64, and viewed by confocal microscopy as described previously (62). Images were equally enlarged to ×4,000 and were analyzed and processed with Photoshop CS5.

Protein determination

Protein concentrations were determined by the Bradford method as described previously (44–46). P13 membrane fractions or purified proteins were accordingly diluted and calculated from a linear standard curve of BSA using Microsoft Excel.

Statistical analyses

For microscopy analyses, 150–200 cells from multiple independent experiments (as specified in the Fig. 6 legend) were blind-scored from random fields, and their mean values and standard deviations were calculated using a standard statistical tool (Excel). p values were calculated using a chi-squared test. p values of <0.05 were considered statistically significant.

Data availability

All presented data are contained within the article.