|Home | About | Journals | Submit | Contact Us | Français|
The BTN1 gene product of the yeast Saccharomyces cerevisiae is 39% identical and 59% similar to human CLN3, which is associated with the neurodegenerative disorder Batten disease. Furthermore, btn1-Δ strains have an elevated activity of the plasma membrane H+-ATPase due to an abnormally high vacuolar acidity during the early phase of growth. Previously, DNA microarray analysis revealed that btn1-Δ strains compensate for the altered plasma membrane H+-ATPase activity and vacuolar pH by elevating the expression of the two genes HSP30 and BTN2. We now show that deletion of either HSP30 or BTN2 in either BTN1+ or btn1-Δ strains does not alter vacuolar pH but does lead to an increased activity of the vacuolar H+-ATPase. Deletion of BTN1, BTN2, or HSP30 does not alter cytosolic pH but diminishes pH buffering capacity and causes poor growth at low pH in a medium containing sorbic acid, a condition known to result in disturbed intracellular pH homeostasis. Btn2p was localized to the cytosol, suggesting a role in mediating pH homeostasis between the vacuole and plasma membrane H+-ATPase. Increased expression of HSP30 and BTN2 in btn1-Δ strains and diminished growth of btn1-Δ, hsp30-Δ, and btn2-Δ strains at low pH reinforce our view that altered pH homeostasis is the underlying cause of Batten disease.
Juvenile neuronal ceroid-lipofuscinoses, or Batten disease, is an autosomal progressive neurodegenerative disease in children, with an incidence as high as one in 12,500 live births and with about 440,000 carriers in the United States (1, 7). Diagnosis is often based on visual defects, behavioral changes, and seizures. Progression is characterized by a decline in mental abilities, increased severity of untreatable seizures, blindness, loss of motor skills, and premature death. The CLN3 gene, positionally cloned in 1995, was shown to be responsible for Batten disease, with most diseased individuals harboring a major deletion of the gene (11). However, the function of the CLN3 protein and the molecular basis for the disease remain elusive. Batten disease is characterized by the accumulation of autofluorescent hydrophobic material in the lysosomes of neurons and, to a lesser extent, in other cell types (9, 14). Furthermore, protein sequencing and immunological studies have revealed that subunit c of mitochondrial ATP synthase is the major component of the lysosomal storage material (8, 20). Although initially located in the mitochondria, mitochondrial ATP synthase subunit c accumulates in lysosomes of NCL cells, whereas the degradation of another mitochondrial inner membrane protein, cytochrome oxidase subunit IV, is unaffected, with no lysosomal accumulation (5, 15). CLN3 localizes in the lysosome (12, 13), suggesting that accumulation of mitochondrial ATP synthase subunit c in the lysosome is not due to defective mitochodria but rather to a defect within the lysosome.
Genes encoding predicted proteins with high sequence similarity to the Cln3 protein have been identified in mouse, dog, rabbit, Caenorhabditis elegans, and the yeast Saccharomyces cerevisiae (18, 28; GenBank accession no. U92812 and 249335 and Swiss-Prot accession no. O29611). We previously reported that the corresponding yeast gene, BTN1, encodes a nonessential protein that is 39% identical and 59% similar to human Cln3p (21). Deletion of BTN1 had no effect on mitochondrial function or the degradation of mitochondrial ATP synthase subunit c. We further demonstrated that yeast strains lacking Btn1p were resistant to d-(−)-threo-2-amino-1-[p-nitrophenyl]-1,3-propanediol (ANP) and that this phenotype was complemented by expression of human Cln3p, indicating that yeast Btn1p and human Cln3p have the same function (22). Furthermore, the degree of ANP resistance was correlated with point mutations identified in CLN3 that are associated with less severe forms of Batten disease, further establishing the functional equivalence of Btn1p and CLN3 (22).
The resistance of btn1-Δ yeast strains to ANP was caused by an apparent decrease in the pH of the growth media brought about by an elevated ability to acidify growth medium through an increased activity of the plasma membrane H+-ATPase (23). This increased activity is caused by a response to an imbalance in pH homeostasis within the cell, produced from an abnormally acidic vacuolar pH in btn1-Δ strains (23). DNA microarray analysis revealed that expression of only two genes, HSP30 and BTN2, is abnormally increased as btn1-Δ strains grow, apparently returning the plasma membrane H+-ATPase activity and vacuolar pH to normal. Furthermore, like CLN3, Btn1p has been localized to the lysosome, called the vacuole in yeast (4, 23). We report here that deletion of either HSP30 (hsp30-Δ) or BTN2 (btn2-Δ) does not result in ANP resistance, as vacuolar pH is unaffected by these mutations. However, the vacuolar H+-ATPase activity is nevertheless increased in hsp30-Δ or btn2-Δ, particularly btn2-Δ. Furthermore, btn1-Δ, hsp30-Δ, or btn2-Δ causes poor growth at low pH in the presence of sorbic acid. This new phenotype for the Batten disease yeast model, btn1-Δ, and for hsp30-Δ and btn2-Δ, whose gene products exhibit increased expression in btn1-Δ strains, provides more evidence that deletion of BTN1 affects the ability of a yeast cell to regulate intracellular pH.
The yeast strains used in this study are listed in Table Table1.1. ρ− derivatives were obtained by growing the ρ+ strains in the presence of ethidium bromide. Yeast strains were grown in either YPD (1% Bacto Yeast Extract, 2% Bacto Peptone, and 2% glucose) medium or YPG (1% Bacto Yeast Extract, 2% Bacto Peptone, and 2% glycerol) medium containing either ANP or sorbic acid at various concentrations. Growth in liquid media was measured with a Summerson photoelectric colorimeter (Klett Mfg. Co., Inc., New York, N.Y.). Deletion of BTN1 has been described previously (22). Deletion of HSP30 and BTN2 was performed by standard techniques. HSP30 was disrupted using the plasmid pSPHSP30 obtained from P. Piper (University College London). Briefly, a 1.1-kb HindIII fragment, containing the URA3 gene inserted in the HindIII site in the coding region of HSP30, was used for disruption of HSP30 by homologous recombination. BTN2 was disrupted with the plasmid pAB2197 containing a 1.1-kb HindIII fragment with the URA3 gene blunt ended and ligated in the NdeI site in the coding region of BTN2.
Plasma membranes and vacuoles were isolated and purified as previously described (23). Cell extracts were prepared, and plasma membranes were collected at the interface of a discontinuous sucrose density gradient, vacuoles were collected at the surfaces of discontinuous Ficoll gradients, and both were purified (23). Plasma membrane and vacuolar H+-ATPase activities were assessed by using pertinent inhibitors of mitochondrial, plasma membrane, or vacuolar H+-ATPase and by measuring Pi release as previously described (23). In addition, the protein levels of plasma membrane and vacuolar H+-ATPases were measured using a polyclonal antibody to Pma1p (a kind gift from Andre Goffeau, Universitae Catholique de Louvain, Louvain, Belgium) or a monoclonal antibody to Vph1p (Molecular Probes), respectively.
Vacuolar pH was measured using the fluorescent dye 6-carboxyflourescein (6-CFDA) in accordance with the method of Preston et al. (25), except that fluorescence was measured with a Fluorolog 2 spectrofluorimeter (Instruments S.A., Edison, N.J.). Labeling of cells with 5 μM 6-CFDA was performed in YPD medium which contained 50 mM citric acid adjusted to pH 3.0, and the cells were washed with YPD medium. Vacuolar pH was calculated from an in vivo calibration prepared by pretreating cells with ionophores to equilibrate the pH in vivo.
Cytosolic pH was measured using the fluorescent dye carboxyseminaphthorhodafluor-1 (C-SNARF-1) as described by Haworth et al. (10). Briefly, cells were harvested, washed twice in 30 mM MOPS (morpholine propanesulfonic acid), pH 7.0, and resuspended in the same solution to 2 × 107/ml. C-SNARF-1 (1 mg/ml) in dimethyl sulfoxide was loaded to a final concentration of 10 μM and incubated at room temperature for 10 min. Fluoresecence was determined with a Fluorolog 2 spectrofluorimeter with excitation at 534 nm and emission at 580 nm, and the intensities were compared to an in vivo calibration prepared as described previously (10).
Yeast strains were grown to either mid-logarithmic or stationary phase in YPD medium, washed in fresh YPD medium, and resuspended in YPD medium to an optical density at 600 nm of 0.4. Solutions of 1 M HCl or 1 M NaOH were carefully added until the yeast cell suspension had reached pH 3.0 or 10.0, respectively. This essentially gave a measure of tolerance, or the buffering capacity of yeast to the addition of external acid or alkali, for each strain.
A Met− btn2-Δ strain was transformed by standard techniques with a plasmid derived by ligating BTN2 in the EcoRI site of the pGFP-N-FUS plasmid (19) such that the green fluorescent protein (GFP) was fused to the N terminus of BTN2. GFP-Btn2p was visualized in yeast cells by fluorescence microscopy.
We previously showed that btn1-Δ elevated the expression of HSP30 and BTN2 (23). Because btn1-Δ caused ANP resistance through an elevated activity of the plasma membrane H+-ATPase and a decreased vacuolar pH in the early phase of growth, we investigated the effects of hsp30-Δ and btn2-Δ deletions on these properties in both BTN1+ and btn1-Δ strains. As shown in Fig. Fig.1,1, hsp30-Δ or btn2-Δ does not cause resistance to ANP. Furthermore, hsp30-Δ or btn2-Δ in btn1-Δ strains, and the presence of both hsp30-Δ and btn2-Δ in either BTN1+ or btn1-Δ strains, did not significantly alter the response to ANP (data not presented). The fact that hsp30-Δ and btn2-Δ do not cause ANP resistance similar to that caused by btn1-Δ indicates that despite being up-regulated in response to btn1-Δ, deletion of these genes does not necessarily produce the same phenotype. Therefore it is unlikely that hsp30-Δ or btn2-Δ would produce the same physiological response with regard to plasma membrane H+-ATPase activity and vacuolar pH as btn1-Δ.
We have previously reported that elevated expression of HSP30 and BTN2 in btn1-Δ strains appeared to inhibit the increased activity of plasma membrane H+-ATPase and decreased vacuolar pH during growth (23). The roles of BTN1, HSP30, and BTN2 in coordinating pH homeostasis were further investigated by measuring plasma membrane H+-ATPase activity, vacuolar H+-ATPase activity, and the vacuolar and cytosolic pH values of the various single-, double-, and triple-deletion strains after 6.5 (early exponential phase) and 25 h (late exponential phase) of growth (Table (Table2)2) (23). Clearly, the vacuolar pH at 6.5 h is decreased from 6.2 to 5.8 in btn1-Δ strains, irrespective of the presence or absence of HSP30 or BTN2. Similarly, the vacuolar pH returns to 6.2 at 25 h in btn1-Δ strains, independent of HSP30 or BTN2. Cytosolic pH is apparently unaffected by btn1-Δ, hsp30-Δ, or btn2-Δ, being pH 6.6 in all strains in both early and late exponential phases of growth (Table (Table2).2).
btn1-Δ causes a minor increase in vacuolar H+-ATPase activity in early growth (Table (Table2);2); however, the role this ATPase has in altering vacuolar pH in btn1-Δ strains is unclear. For any of the strains used in this study, vacuolar H+-ATPase activity was increased after 25 h of growth compared to 6.5 h. Furthermore, hsp30-Δ or btn2-Δ, but not btn1-Δ, clearly caused an increase in vacuolar H+-ATPase in cells grown for 25 h (Table (Table2).2). Using Western analysis of Vph1p, an integral membrane component of the vacuolar H+-ATPase, we have confirmed that increased activity of vacuolar H+-ATPase is not due to an increased amount of vacuolar H+-ATPase itself (data not presented). The increase in vacuolar H+-ATPase caused by the hsp30-Δ or btn2-Δ deletion may be related to the finding that expression of HSP30 and BTN2 is enhanced in btn1-Δ strains.
All btn1-Δ strains at the early phase of growth, 6.5 h, have elevated plasma membrane H+-ATPase activities (Table (Table2)2) that are reflected by resistance to ANP (Fig. (Fig.1).1). All strains show a characteristic increase of plasma membrane H+- ATPase activity through growth (2). Most significantly, all hsp30-Δ strains at the later phase of growth, 25 h, also have elevated plasma membrane H+-ATPase activities (Table (Table2).2). Piper et al. (24) previously reported that HSP30 diminishes plasma membrane H+-ATPase activity, implying that plasma membrane H+-ATPase activity is not inhibited in hsp30-Δ strains. The reason plasma membrane H+-ATPase activities are not significantly higher in hsp30-Δ strains at the early phase of growth may be the fact that HSP30 is not usually significantly expressed until later in the growth cycle (23).
Piper et al. (24) previously reported that hsp30-Δ strains grow poorly at pH 3.8 in the presence of low concentrations of sorbic acid. This finding was extended by examining the effects of hsp30-Δ and btn2-Δ deletions, as well as the btn1-Δ deletion, on the growth of strains on a medium containing 0.2 mM sorbic acid at pH 3.8. As shown in Fig. Fig.2A,2A, growth of the btn1-Δ, hsp30-Δ, and btn2-Δ strains was slightly diminished, and the growth of the double-deletion hsp30-Δ btn2-Δ strain was even further diminished. The diminished growth on sorbic acid medium was extended by examining the ρ− series of strains having btn1-Δ, hsp30-Δ, and btn2-Δ deletions (Table (Table1).1). Although the normal ρ− strain grows poorly on sorbic acid medium compared to the corresponding normal ρ+ strain, btn1-Δ, hsp30-Δ, or btn2-Δ causes a more severe growth defect compared to normal (Fig. (Fig.2B).2B). These growth defects are complemented by the introduction of BTN1, HSP30, or BTN2 plasmids in the corresponding btn1-Δ, hsp30-Δ, and btn2-Δ strains (results not presented). The diminished growth of the deletion strains on sorbic acid medium suggests a physiological link between the BTN1, HSP30, and BTN2 genes, possibly in maintaining the pH homeostasis of the cell. In addition, the fact that this growth defect is exacerbated in ρ− strains, which are therefore incapable of respiration, suggests that maintaining intracellular pH is linked to the cell's available energy. It is pertinent to cite our previous report that btn1-Δ does not alter mitochondrial function and that ρ+ BTN1+ and ρ+ btn1-Δ strains and ρ− BTN1+ and ρ− btn1-Δ strains have the same growth characteristics under normal growth conditions (21). Furthermore, if we compare the growth of ρ+ btn1-Δ, hsp30-Δ, and btn2-Δ strains and that of the corresponding ρ− btn1-Δ, hsp30-Δ, and btn2-Δ strains reported here, there is no defect in growth under normal growth conditions (data not presented).
To further confirm that btn1-Δ, hsp30-Δ, and btn2-Δ strains have compromised pH homeostasis, we simply measured the amount of acid or alkali that was required to substantially alter the pHs of the media in which these strains were suspended. In essence, we measured the buffering capacity for each strain. Each strain was resuspended to identical concentrations in YPD medium with a normal pH of approximately 6.5, and either acid or alkali was added to the suspension. As S. cerevisiae is able to manipulate the pH of its environment during growth, the rationale for this experiment was simply that compromised pH homeostasis would alter the amount of acid or alkali required to bring the medium pH to either 3, or 10, respectively. As indicated in Table Table3,3, deletion of btn1-Δ, hsp30-Δ, or btn2-Δ results in a decreased buffering capacity of the yeast to alkali so that less alkali is actually required to bring the yeast suspension up to pH 10.0 in mutant strains. The results shown are representative for btn1-Δ, hsp30-Δ, and btn2-Δ strains grown to either mid-logarithmic or stationary phase.
Hsp30p and Btn1p are localized in the plasma membrane and vacuole, respectively (4, 23, 24). The subcellular location of Btn2p was investigated with GFP fused to Btn2p at either the amino (GFP-Btn2p) or carboxyl (Btn2p-GFP) terminus. By expressing the fusion proteins and examining growth on sorbic acid medium, we determined that amino-terminal GFP-Btn2p is functional while the carboxyl-terminal Btn2p-GFP is nonfunctional. The lack of function of Btn2p-GFP suggests that the carboxyl region of the protein is especially important for function. The fluorescent pattern emitted by GFP-Btn2p established that Btn2p is localized in the cytosol (Fig. (Fig.3).3).
We have previously shown that btn1-Δ strains have a more acidic vacuole at the outset of growth, which is compensated for at the biochemical level through the elevated activity of the plasma membrane H+-ATPase and at the molecular level through increased expression of HSP30 and BTN2 (23). We have extended the study of our model for Batten disease to further understand this defect in pH homeostasis.
In this study, we demonstrate that deletion of either or both of the two genes HSP30 and BTN2, which have increased expression in btn1-Δ strains, did not alter the pH-dependent resistance to ANP in btn1-Δ strains nor did it result in resistance to ANP for BTN1+ strains. Furthermore, the btn1-Δ hsp30-Δ btn2-Δ strain, with all three genes deleted, is viable and shows no growth defect under normal conditions (results not presented). The ANP-resistant phenotype therefore serves as a marker for decreased vacuolar pH in early growth, as exhibited by btn1-Δ strains. We can predict that the HSP30 and BTN2 gene products most likely do not alter vacuolar pH, a conclusion experimentally verified by measuring the vacuolar pH. In fact, the characteristic low and normal vacuolar pH in btn1-Δ strains at early and later growth, respectively, is still apparent in the btn1-Δ hsp30-Δ strain, the btn1-Δ btn2-Δ strain, and the btn1-Δ hsp30-Δ btn2-Δ strain. Therefore, despite the deletion of the two genes that have elevated expression in btn1-Δ strains, vacuolar pH is still normalized during growth.
Increased vacuolar H+-ATPase activity was apparent in btn2-Δ strains compared to that in btn1-Δ strains. If we assume that an increase in vacuolar H+-ATPase activity results in a net gain of protons in the vacuole, it is puzzling that the pH of the vacuole itself is not more acidic than normal in btn2-Δ strains. However, there is no precedent for vacuolar pH correlating with the activity of vacuolar H+-ATPase. Previous studies revealed that btn1-Δ strains had a slightly elevated activity of vacuolar H+-ATPase in early growth which apparently returns to normal in later growth when expression of BTN2 is increased, whereas this study has revealed that btn2-Δ strains had an elevated activity of vacuolar H+-ATPase. Taken together, these data suggest that Btn2p may play a role in regulating activity of the vacuolar H+-ATPase. In another study, we found that in a cup5-Δ strain, which lacks a functional V-ATPase due to the absence of the CUP5-encoded subunit c, expression of BTN2 is elevated, again suggesting a role for Btn2p in regulating V-ATPase (S. Chattopadhyay and D. A. Pearce, unpublished data). Interestingly, the one constant in this study was that cytosolic pH was unaffected in btn1-Δ, btn2-Δ, and hsp30-Δ strains. This suggests that perhaps altered gene expression and modified vacuolar biochemistry contribute, at least in part, to maintaining a balanced cytosolic pH and that maintaining cytosolic and vacuolar pH is important. The fact that strains bearing a deletion of btn1-Δ, hsp30-Δ, or btn2-Δ have decreased buffering capacity, as determined by simply adding alkali to the strains, underscores our assertion that pH homeostasis is compromised. Clearly this study has not completely addressed how btn1-Δ, with btn2-Δ and hsp30-Δ mutations, balances vacuolar pH. One possibility is that S. cerevisiae has more than one redundant pathway that is activated to maintain intracellular pH homeostasis, which will be revealed by further gene expression studies in the appropriate btn1-Δ btn2-Δ hsp30-Δ genetic background. Another possibility requires that the role of the vacuolar H+-ATPase in the BTN pathway be scrutinized further. We have previously reported that Btn2p and human HOOK1 share 20% identity and 46% similarity over a 211-amino-acid span, although the two proteins do not appear to be functional orthologs, since HOOK1 does not complement btn2-Δ (results not shown). Nevertheless, such homology over a region of the protein implies similar, if not identical, functions. In Drosophilia, HOOK1 has been shown to be involved in endocytosis of the transmembrane ligand bride of sevenless (boss) into multivesicular bodies (MVBs) (16). Further studies have suggested that HOOK1 may indeed be a negative regulator of the fusion of these MVBs to the lysosome (27). If Btn2p has a related role in yeast, it is tempting to speculate that an increase of Btn2p in btn1-Δ strains may result in a decrease in the delivery of MVBs to the vacuole while btn2-Δ strains may have uncontrolled delivery of MVBs to the vacuole. As MVBs are likely to be acidified compartments, delivery of these MVBs may contribute to the underlying defect in pH homeostasis that is observed.
We have recently found that Btn2p physically interacts with known vesicular proteins, indicating a role in vesicular trafficking (Chattopadhyay and Pearce, unpublished). The key to defining Btn2p, which is up-regulated in btn1-Δ strains and, of course, absent in btn2-Δ strains, is to understand the function of Btn1p. We have previously speculated that Btn1p may be a transporter because, as a vacuolar transmembrane protein, a more acidic vacuolar pH results from its deletion, btn1-Δ. If Btn1p is a transporter, btn1-Δ causes an imbalance of transport, causing the vacuole to become more acidic. Btn2p expression is increased, which negatively regulates delivery of vesicles and their acidic contents to the vacuole. Therefore, there would be a decrease in delivery of acidic vesicles to the vacuole, which may decrease acid accumulation and result in balancing the pH. In this scenario, btn2-Δ would result in nonregulated delivery of acidic vesicles, and perhaps increased delivery of an acidic component, to the vacuole. This is at odds with the vacuolar pH, which is unaffected in btn2-Δ strains. However, there is increased activity of vacuolar H+-ATPase in btn2-Δ strains, which would suggest that an as-yet-unidentified acidic component was being transported out of the vacuole, which would be a candidate substrate for the Btn1p transporter. Alternatively, Btn2p may be a cytosolic protein that acts as a facilitator in the control of cytosolic pH.
Russell (26) established that the entry of sorbic acid into a yeast cell results in growth inhibition and that this inhibition is increased with medium acidification, essentially being proportional to the concentration of undissociated acid. Following entry into the cell in low-pH cultures, sorbic acid dissociates at the higher pH of the cytosol, leading to intracellular acidification, and previous reports noted that weak organic acids inhibit both fermentation and respiration in S. cerevisiae (6, 17). Clearly, growth at pH 3.8 in the presence of 0.2 mM sorbic acid places a homeostatic stress on the yeast cell, which is manifested as poor growth of btn1-Δ, btn2-Δ, or hsp30-Δ strains, implying a physiological relationship between Btn1p, Hsp30p, and Btn2p. Certainly sorbic acid will affect a yeast cell's ability to maintain a constant cytosolic pH through the action of Btn1p, Hsp30p, and Btn2p if any of these proteins is absent. Recent studies of hsp30-Δ strains indicated that Hsp30p acts as a stress-inducible regulator of plasma membrane H+-ATPase (24). The same authors showed that low-pH sorbic acid stress depletes cellular ATP but the depleted ATP is not reflected by a reduced metabolic activity. It was speculated that hsp30-Δ strains would be less efficient at reestablishing homeostasis within the cell due to an elevated plasma membrane H+-ATPase activity, which in turn would commit the cell to an even higher usage of ATP. Therefore, although we have not quantified ATP levels in this study, it is probable that btn1-Δ, btn2-Δ, and hsp30-Δ strains all exhibit similar energy-depleted growth inhibition in the presence of sorbic acid due to a reduced capacity to adapt to the pH stress. It is compelling that this phenotype is exacerbated in ρ− strains, which is highly suggestive of an energy requirement for the ability to grow normally in the presence of sorbic acid at low pH. No doubt, elevated activity of vacuolar H+-ATPase in btn1-Δ, btn2-Δ, and hsp30-Δ strains causes an energy drain on the cell due to the increased utilization of ATP. By uncovering this new phenotype for the study of the Batten disease yeast model, we should be able to refine our understanding of what may be causing perturbed pH homeostasis in btn1-Δ strains. Disturbed pH homeostasis in btn1-Δ strains may also result in an energy imbalance, as seen in hsp30-Δ strains. A combination of this altered pH homeostasis and energy imbalance may be a contributory factor in the neurodegeneration in Batten disease.
We thank Peter W. Piper (Department of Biochemistry and Molecular Biology, University College London, London, United Kingdom) for providing the plasmid pSPHSP30 and Andre Goffeau, Universitae Catholique de Louvain, Louvain, Belgium, for providing an antibody to Pma1p.
This work was supported by the National Institutes of Health grant R01 NS36610.