Human Molecular Genetics, 2002, Vol. 11, No. 18 2129-2142
© 2002 Oxford University Press
Motifs within the CLN3 protein: modulation of cell growth rates and apoptosis
1University Program in Genetics, 2Departments of Pediatrics and Neurobiology and 3Department of Pharmacology, Duke University Medical Center, Durham, NC, 27710, USA
Received April 26, 2002; Revised June 27, 2002; Accepted June 28, 2002
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONCONCLUSIONSMATERIALS AND METHODSREFERENCES |
|---|
Juvenile Batten disease (JNCL) is an autosomal recessive disease that results from mutations in the CLN3 gene. The wild-type CLN3 gene coding sequence has 15 exons, and the translated protein consists of 438 amino acids. The most commonly observed mutation is a 1.02 kb deletion in the genomic DNA. This deletion results in a truncated protein due to the loss of amino acids 154–438, and the introduction of 28 novel amino acids at the c-terminus. We demonstrate that, compared to normal controls, CLN3-deficient immortalization of lymphoblasts homozygous for this deletion grow at a slower rate, and show increased sensitivity to etoposide-induced apoptosis, supporting the notion that CLN3 may negatively regulate apoptosis. Using immortalized JNCL lymphoblast cell lines as a model system, we assess the effects of specific CLN3 mutations on cell growth rates and protection from etoposide-induced apoptosis. Protection from etoposide-induced apoptosis occurs and the cell growth rate is restored following transfection of JNCL lymphoblasts with mutant CLN3 cDNA that includes exons 11 or 13. We show that deletion of the glycosylation sites 71NQSH74 and 310NTSL313, and also mutations within the highly conserved amino acid stretches 184WSSGTGGAGLLG195, 291VYFAE295 and 330VFASRSSL337, result in slowed growth and susceptibility to apoptosis.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONCONCLUSIONSMATERIALS AND METHODSREFERENCES |
|---|
Batten disease is a collective term for a group of diseases known as the neuronal ceroid lipofuscinoses (NCLs). Batten disease is one of the most common neurodegenerative disorders of childhood, with an incidence of 1 in 100 000 in the general population (1). JNCL, or the juvenile-onset form of NCL (OMIM 204200), is caused by mutations within the CLN3 gene, and is an autosomal recessive disease (2). The age of onset is usually 4–10 years, with the first manifestations being rapid loss of vision due to retinitis pigmentosa. Clinical hallmarks are seizures, psychomotor and cognitive decline, blindness, and death, commonly in the mid-20s to late 20s (2,3).
The pathology of JNCL is remarkable for massive neuronal death, photoreceptor loss and the accumulation of lipopigments in lysosomes (4,5). Autofluorescent bodies and fingerprint profiles within cells are widespread in JNCL brain tissue, especially within the surviving pyramidal neurons of the hippocampal sectors CA2–CA4 (5,6). These bodies consist mainly of subunit c or 9 of mitochondrial ATP synthase (5–8).
Other NCL genes include CLN1, CLN2, CLN5, CLN6 and CLN8. CLN1 and CLN2 code for palmitoyl-protein thioesterase 1 (PPT1) and for the lysosomal protease tripeptidyl peptidase 1 (TPP1), respectively (9,10). CLN3 (11), CLN5 (12), CLN6 (13) and CLN8 (14) most likely are transmembrane proteins. The hydropathy profile for CLN3 suggests that it has 5–11 transmembrane domains (11). There is some controversy concerning the location of the CLN3 protein. Different studies have localized the CLN3 protein to the cell membrane (15), the lysosomal membrane in non-neural tissues (16,17), the mitochondria in retinal cells (18) and the Golgi apparatus (19). CLN3 has been shown to be abundantly expressed in primary rat neurons, rat neuronal tissue and synaptosomes, with exclusion, however, from both lysosomes and synaptic vesicles (20,21). The different conclusions reached in these studies suggest that the true location or locations of CLN3p may be species and/or tissue specific.
We have previously claimed that CLN3 may have antiapoptotic properties (22,23). This claim has been supported by findings based on electron microscopy and TUNEL assay in CLN3-deficient human brain (24). Increased apoptosis in CLN3-deficient or JNCL patient leukocytes has been documented by others (25). Previous studies have demonstrated that overexpression of CLN3 enhances growth of NT2 cells, and protects from apoptosis induced by chemotherapeutic drugs and serum starvation (22). CLN3 is developmentally regulated (26). It is also upregulated in a number of human and mouse cancer cell lines and solid human colon cancer. Blocking of CLN3 expression in human cancer cells has led to inhibition of cancer cell growth and to increased apoptosis (27). Blocking of CLN3 expression in human differentiated, postmitotic hNT neurons results in spontaneous and unprovoked apoptosis (23). Yeast knockout models have suggested that CLN3 is involved in pH regulation (28,29). There is evidence to suggest that the onset of apoptosis is characterized by an initial alkalinization of the mitochondrial matrix, followed by cytosolic acidification and cell death (30).
The CLN3 gene is highly conserved across species. Homologs for CLN3 in the dog, mouse, Caenorhabiditis elegans and yeast bear 78%, 75%, 37% and 34% identity to the human CLN3 gene, respectively (22). This underscores the biological importance of CLN3. CLN3 is a novel protein with stretches of conserved amino acids along its length, including residues 184WSSGTGGAGLLG195, 291VYFAE295 and 330VFASRSSL337. We hypothesized that the apoptosis-modulating activity of CLN3 could be defined by stretches of amino acids within these conserved regions.
There are 31 mutations known to cause JNCL (31–33). Eighty five per cent of the cases are due to a 1.02 kb deletion in the genomic DNA (2), while the other 15% result from point or frameshift mutations. Lymphoblasts derived from patients homozygous for this deletion grow more slowly than normal controls, and show increased sensitivity to etoposide-induced apoptosis. We address the following questions: (1) Do specific mutations create a cell phenotype similar to that observed in JNCL cell lines derived from patients homozygous for the 1.02 kb deletion? (2) Do these mutations reside within highly conserved regions of CLN3? (3) Does the elimination of potentially important biological sites within the CLN3 protein affect cell growth rates and apoptosis? (4) Does altering the charge and/or hydrophobicity of specific amino acids within CLN3 affect growth and apoptosis of cells? We present data that support the notion that CLN3 may regulate apoptosis, and that many but not all of the elements responsible for this regulation reside at the c-terminus of the CLN3 protein.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONCONCLUSIONSMATERIALS AND METHODSREFERENCES |
|---|
Slowed growth rate of JNCL patient lymphoblasts
Our results indicate that untransfected CLN3-deficient cells derived from patients homozygous for the common deletion show reduced growth rates when compared to normal untransfected lymphoblasts, and the P-value is <0.05 (Fig. 1A). We also show that untransfected immortalized JNCL lymphoblast cell lines derived from four different patients harboring the 1.02 kb deletion show similar growth rates (P>0.05). The introduction of CLN3 by transfection into these JNCL patient lymphoblasts reverses this growth rate defect when compared to transfection with an empty vector control (Fig. 1B) (P<0.05). Notably, JNCL patient cells show an increased rate of cell death compared to normal lymphoblasts, with a P value of <0.05 (Fig. 1C). A reversal in the rate of death is accomplished following transfection with a CLN3-bearing plasmid (P>0.05). Maximum CLN3 expression is observed between time points 0 and 48 h, after which it declines.
|
The CLN3 protein does not affect DNA synthesis
Rates of thymidine incorporation into DNA by JNCL lymphoblasts were compared following transfection with an empty vector control or a CLN3 cDNA construct. Rates for JNCL lymphoblasts transfected with either an empty vector control or wild-type CLN3 cDNA are shown in Figure 1D. The rates of thymidine incorporation show no significant differences, and the P-value is >0.5. We also compared the rate of thymidine incorporation for deletion constructs containing CLN3 exons 1–4, 1–7, 1–8, 1–9, 1–10, 1–11, 1–12, 1–13, 1–14 and 1–15. The effect of the CLN3 deletion construct harboring exons 1–4 of the CLN3 cDNA on thymidine incorporation is shown in Figure 1D. There are no significant differences in the rates of thymidine incorporation among any of the generated deletion constructs, and the P-value for all points was >0.5. This suggests that the effect of CLN3 on the cell growth rate is not due to differences in the rate of DNA synthesis, but due to increased cell death in patient cell lines (Fig. 1C).
Increased etoposide-induced apoptosis in JNCL patient lymphoblasts
Immortalized JNCL lymphoblasts homozygous for the common deletion were transfected with either an empty vector control or the full CLN3 cDNA, treated with etoposide and then analyzed using the TUNEL assay. JNCL lymphoblasts transfected with the full CLN3 cDNA (Fig. 2C and D) show more protection from etoposide-induced apoptosis than is observed in JNCL lymphoblasts transfected with an empty vector (Fig. 2A and B), with a P-value of <0.05. Our results confirm that deletion of CLN3 results in increased cell death. Results were corroborated using the JC-1 mitochondrial membrane potential assay (Table 1). The degree of protection is determined by the percentage of live cells after treatment with etoposide divided by the percentage of live cells before treatment with etoposide. Compared to JNCL cells transfected with the empty vector control, JNCL patient lymphoblasts transfected with the full CLN3 cDNA show significant protection from etoposide-induced apoptosis (48±0.8%, P>0.5 versus 96±2.8%, P<0.005) (Table 1, Fig. 2). This supports the results of previous studies (20–22).
|
|
Effect of exons 11 and 13 of CLN3 on cell growth rates and apoptosis
Our next aim was to determine the regions of the CLN3 gene that may impact on cell growth rates. We used deletion constructs generated from the conserved 3' end (34), and established growth curves following transfection of the CLN3-deficient cells with each construct. The constructs containing exons 1–13 restore cell growth potential to normal levels as opposed to transfection with the empty vector, and the P value is <0.05 (Fig. 3). The effect of transfection of the construct containing exons 1–11 into JNCL patient cells is a significantly enhanced growth rate, as opposed to the effect of the empty vector, and the P-value is <0.05. Inclusion of exon 11 results in a rate of cell growth that is somewhat slower than that observed for JNCL patient cells transfected with the full CLN3 cDNA, and the P-value is <0.05. Both TUNEL and JC-1 analyses demonstrate that the presence of exons 11 and 13 provides transfected JNCL lymphoblasts with protection from etoposide-induced apoptosis, and the degrees of protection observed were 87% and 96%, respectively (raw data not shown; Table 1). CLN3 constructs lacking exons 11 and 13 fail to confer significant protective properties to JNCL patient lymphoblasts after transfection (Table 1).
|
Amino acid residues important for the impact of CLN3 on apoptosis
We have shown that inclusion of exons 11 and 13 is necessary for the normalizing effect of the CLN3 gene on cell growth rates and apoptosis. Three naturally occurring mutations within exon 11 and 13 are of interest: E295K (exon 11), V330C and R334D (both in exon 13) (31,32).
We introduced changes that could either alter charge and/or hydrophobicity at these specific amino acid sites or keep these parameters similar. The data (Fig. 4A) illustrate that transfection with CLN3 cDNA harboring mutations at R334 that result in a change of charge from basic to acidic (R334D, R334H, R334C) did not correct growth rate defects and had similar effects to transfection with the empty vector control (P<0.05). However, transfection with the mutant CLN3 cDNA construct resulting in R334K (changing basic to basic) did restore the cellular growth rate to within normal levels (P<0.05). Transfection with any type of CLN3 cDNA harboring mutations affecting residues E295 (E295K or E295D) and V330 (V330C or V330L) maintained the slowed JNCL cell growth rate phenotype (Fig. 4B). TUNEL (Fig. 5) and JC-1 (data not shown) staining demonstrate that transfection of immortalized JNCL lymphoblasts with mutant CLN3 cDNA harboring the mutations R334D, E295K and V330C did not reverse sensitivity to etoposide-induced apoptosis. The percentage protection from etoposide-induced apoptosis observed in cells after transfection with these constructs was similar to that of JNCL cells transfected with the empty vector (Table 1). The mutations E295K, E295D, V330C, V330L, R334D, R334H and R334C were not quantitatively distinguishable from each other (P>0.05).
|
|
In 85% of JNCL patients, the block of conserved amino acids 184WSSGTGGAGLLG195 is lost due to the 1.02 kb deletion in genomic DNA. This deletion results in a truncated protein with loss of more than 200 amino acids from the c-terminus (2). A growth phenotype consistent with the one observed in JNCL patient cells is observed following transfection with mutant CLN3 cDNA that corresponds to this deletion, and the P-value is >0.05 (Fig. 4C). The residues V330 and R334 form part of a highly conserved stretch of amino acids, 330VFASRSSL337. Mutations affecting this region maintain the deficient cell growth rate phenotype (Fig. 4A and B).
The motif 291VYFAE295 is highly conserved among CLN3 orthologs, and is homologous to amino acid sequences within the S5 domain of K+ channels (35), and the amyloid-ß precursor peptide 19VFFAE22 (36,37) (http://www.ncbi.nlm.nih.gov). We show that constructs resulting in deletion of residues 291VYFAE295 maintain the phenotype of diminished growth, similar to the one seen in JNCL patient lymphoblasts, and the P-value is >0.5. The mutation resulting in Y292W, located within this stretch of residues, also maintains a reduced growth rate phenotype (P>0.05). However, the CLN3 mutation resulting in Y292F, F being the corresponding residue in normal amyloid-ß precursor protein, restores the rate of cell growth to normal, and the P-value is >0.05 (Fig. 4D).
Impact of glycosylation residues 71–74 and 310–313 on cell growth rates
There are four potential glycosylation sites and 12 potential phosphorylation sites within CLN3 (2). It is suggested that the CLN3 protein undergoes post-translational modification by phoshorylation (38,39) and glycosylation (17). CLN3 also harbors potential mitochondrial targeting signal (11) and farnesylation sites (38).
Mutant CLN3 clones affecting each of these sites were individually transfected into JNCL patient lymphoblasts, and growth curves were established for each by the trypan blue dye exclusion assay. Deletion of each of the phoshorylation sites produces a growth curve similar to that established for JNCL patient lymphoblasts transfected with the full-length CLN3 cDNA (P>0.05, data not shown). Deletion of the mitochondrial targeting signal or the farnesylation site did not affect cell growth rate (Fig. 4C; P>0.05). Independent deletion of each of the glycosylation sites corresponding to residues 49NSFY52 and 85NSSS88 did not affect the growth rate (P>0.05, data not shown). Deletion of the DNA sequences corresponding to the glycosylation sites defined by amino acid residues 71NQSH74 results in a similar growth rate pattern to the one observed in JNCL patient lymphoblasts transfected with the empty vector control, and the P-value is >0.05 (Fig. 6A). Deletion of the DNA sequences corresponding to glycosylation sites defined by amino acid residues 310NTSL313 results in a decrease in the growth potential when compared to JNCL patient cells transfected with the full CLN3 cDNA. This decrease was less marked than that observed for the deletion of residues 71NQSH74, and the P-value is <0.05.
|
A hydropathy profile created at www.expasy.ch/tools/TMHMM indicates that the CLN3 protein could harbor 5–11 possible transmembrane domains with a putative pore at amino acid residues 299–343 (Fig. 6B). According to this profile, the glycosylation site defined by residues 71NQSH74 is located within the first putative external loop, and the glycosylation site defined by residues 310NTSL313 is within the putative pore of CLN3. Residues V330 and R334 also reside within this pore. The conserved stretch 291VYFAE295 is located in transmembrane domain 7, which lines the putative pore (Fig. 6B).
Levels of CLN3p expression are unaffected following mutation of the CLN3 cDNA
Mutagenesis of cDNA can lead to unstable and easily disrupted nascent protein and erroneous results. Here we show that this is not the case. Western blot analysis (Fig. 7) illustrates that CLN3p is expressed in higher quantities in JNCL lymphoblasts transfected with either wild-type or mutated CLN3 cDNA than in JNCL lymphoblasts transfected with the empty vector control. We also show that CLN3p is expressed at similar levels in JNCL lymphoblasts transfected with either wild-type CLN3 cDNA or mutated CLN3 cDNA. The results also confirm the transient effect of CLN3p 50 h post-transfection. We show by growth curve analysis and western blot that the CLN3 protein levels start to decline 48–50 h following transfection with CLN3 cDNA.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONCONCLUSIONSMATERIALS AND METHODSREFERENCES |
|---|
Immortalized CLN3-deficient lymphoblasts derived from patients homozygous for the common 1.02 kb deletion were used as the assay model system for testing the effect of various CLN3 cDNA constructs on cell numbers and apoptosis, for the following reasons. Patient leukocytes show inclusions, similar to affected neurons in the brain. Lymphoblasts, as well as neurons, are susceptible to apoptosis. CLN3-overexpressing lymphoblasts show enhanced rates of growth, similar to CLN3-overexpressing neuron precursors. Lymphoblasts are easy to grow and maintain. Last, but not least, lymphoblasts derived from patients are abundantly accessible, unlike naturally deficient human neurons.
Immortalized CLN3-deficient lymphoblasts have provided a simple, reproducible and reliable model system for assessing the differential effect of a number of mutant CLN3 constructs on cell growth rates and apoptosis. Growth curve analyses and JC-1 and TUNEL assays of JNCL patient lymphoblasts transfected with different CLN3 constructs provided confirmation for the expected phenotypes of diminished growth and sensitivity to apoptotis (22,23,27). Expression of the CLN3 protein, confirmed by western blot, was shown to be similar for all CLN3 cDNA constructs used and greater than expression with the empty vector control.
Previous studies have implicated CLN3 in cell growth and regulation pathways (28). Our results demonstrate that JNCL patient lymphoblasts have a slowed growth rate when compared to normal lymphoblasts. Transient transfection of intact CLN3 cDNA into immortalized JNCL patient cells, with ensuing production of CLN3 protein, completely reverses the growth defect. Western blot analysis shows that the mutated CLN3 protein's expression is not affected. The growth pattern is maintained while expression of the transfectant cDNA is present, but reverts back to deficient growth as expression of the CLN3p declines, as demonstrated by western blot. JNCL patient cells have a higher apoptotic rate than do normal cells. Again, introduction of the wild-type CLN3 cDNA blocks apoptosis in JNCL patient cells.
The rates of thymidine incorporation into both wild-type and JNCL patient lymphoblast DNA are similar. However, growth analyses show that the respective cellular growth rates are statistically different. Results suggest that the CLN3 protein does not affect the rate at which the cell synthesizes DNA. The more likely scenario is that absence or mutation of the nascent CLN3 protein impacts on cell death pathways.
Previous studies have placed CLN3 upstream of ceramide in the mitochondria-dependent apoptotic cascade (22). Etoposide or VP-16 is a chemotherapeutic agent that inhibits topoisomerase II and induces mitochondria-dependent apoptosis (40–42). We demonstrate that JNCL patient cells show more protection from etoposide-induced apoptosis when transfected with full-length CLN3 cDNA, as opposed to transfection with an empty vector control (96% versus 48%). This was determined by JC-1 and TUNEL assays. Expression of CLN3 protein, therefore, protects lymphoblasts from etoposide-induced apoptosis.
It is known that CLN3 has sequence-specific signals that can target the protein to individual organelles: a farnesylation site at residues 435CQLS438 (43) which anchors proteins to membranes, a mitochondrial targeting signal at residue 11 with its cleavage site at residue 19 (11), and a di-leucine lysosomal targeting motif at 424LL425 (17). The CLN3 protein is homologous across species for various motifs of potential biological relevance. These are the N-myristoylation site 2GGCAGS7, N-glycosylation sites 49NFSY52, 71NQSH74, 85NSSS88 and 310NTSL313, and the isoprenylation site 435CQLS438 (34). This indicates that CLN3 may be post-translationally modified in either the endoplasmic recticulum or the Golgi apparatus (17,19). Phosphorylation plays an important role in the targeting of proteins to membranes and can influence interactions with other proteins. Although there is evidence to suggest that CLN3 is phosphorylated (38), the independent deletion of each of the phosphorylation sites does not affect cell growth rates in our assay system. Deletion of the sequences corresponding to the CaaX box within the CLN3 cDNA at residues 435–438 and the potential mitochondrial targeting signal did not have an effect on cell growth rates either. These results suggest but do not prove that CLN3 may be neither phosphorylated nor farnesylated, and nor is it targeted to the mitochondria. An alternative and more likely scenario is that post-translational modifications at the CaaX box and the phosphorylation sites do not determine the impact of the CLN3 protein on cell growth rates.
A CLN3 cDNA deletion strategy was designed to find which regions of the CLN3 protein contribute to the maintenance of cell growth rates and prevention of apoptosis in lymphoblasts. Sequences within exons 11 and 13 correspond to amino acid residues that are necessary for the CLN3 protein to negatively regulate apoptosis, and positively impact on cell growth rates. JNCL patient cells transfected with CLN3 deletion constructs missing exons 11 and 13 showed a growth potential similar to the one observed in JNCL patient or CLN3-deficient cells transfected with the empty vector. JC-1 analyses demonstrated that deletion of these regions had a corresponding effect on apoptosis.
CLN3 is a highly hydrophobic membrane-bound protein (11). The modeling algorithm used shows that CLN3 could harbor up to 11 transmembrane segments and a pore region. The generated hydropathy profile suggests that exon 11 forms part of putative transmembrane domain 7, and exon 13 forms part of the putative pore region. There are three residues within these exons that are highly conserved; they are E295 in exon 11, and V330 and R334 in exon 13.
Residue E295 is part of a highly conserved stretch of amino acids (VYFAE) found in CLN3 orthologs, K+ channels (VYFAE) (35) and the amyloid-ß precursor peptide (VFFAE) (36,37). The specific mutation E22K in the amyloid-ß precursor protein has been shown to cause apoptosis in at least 30% of Alzheimer's patients (37). A similar mutation, E295K, is observed in JNCL patients (44). Decreased cell numbers and increased sensitivity to etoposide-induced apoptosis results when the DNA sequence for this five amino acid stretch is deleted or when E295 is mutated to K295.
The mutation of tyrosine to phenylalanine at residue 292 (Y292F) does not have a significant effect on growth, but the change to tryptophan (Y292W) creates a growth phenotype similar to the one observed in JNCL patient cells. Phenylalanine (F) and tryptophan (W) are highly hydrophobic aromatic residues, in contrast to tyrosine (Y), which contains a hydroxyl group. Tryptophan is larger than phenylalanine. Both charge and size of the residue at position 292 in CLN3 could modulate the effect of the CLN3 protein on apoptosis and cell growth rates.
The naturally occurring mutations R334D, R334C and R334H (31) result in alteration of the charge of the amino acid from basic to acidic and produce a JNCL phenotype. The mutation R334K creates a similar charge, and therefore does not cause any significant growth defect. This suggests that the basic charge at R334 is of significance and needs to be maintained for normal CLN3 impact on growth and apoptosis. The mutation at residue V330 has two major effects: it creates a growth phenotype similar to that observed for JNCL patient lymphoblasts, and also increases the sensitivity of cells to etoposide-induced apoptosis. Residues E295 and V330 are vital for preserving the effect of CLN3 on cell growth rates and apoptosis, but it is the charge and/or hydrophobicity at residue R334 that maintains these effects.
All these residues fall within the conserved amino acid stretches 291VYFAE295 and 330VFASRSSL337, suggesting these are two motifs within CLN3 are vital for safeguarding its normal modulation of cell growth rate and apoptosis.
CLN3 may be glycosylated, as two of the potential glycosylation sites at residues 71–74 and 310–313 are necessary for preservation of CLN3 impact on cell growth rates and apoptosis. According to the hydropathy profile, the glycosylation site at amino acid residues 71–74 resides within the putative first external loop (residues 61–98), and the glycosylation site at amino acid residues 310–313 is within the putative pore region (residues 299–343). These sites are in locations that could be important for protein interactions. These could involve proteins deficient in other Batten variants, proteins that could interact with them, or proteins important for preserving CLN3 impact on cell growth rates. The effect of deleting residues 71–74 on the cell growth rate is worse than that of deleting residues 310–313. A possible explanation for this is that residues 71–74 are located within the first external loop of the CLN3 protein, a position that may be essential for targeting protein–protein and/or protein–lipid interactions. According to the presumed hydropathy profile presented in Figure 6B, residues 310–313 are internal. Glycosylation at these residues may be of no consequence; however, the location of these residues in the transmembrane segment bordering the putative pore may be important.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONCONCLUSIONSMATERIALS AND METHODSREFERENCES |
|---|
Additional data that substantiate the claim that the CLN3 protein impacts on apoptosis have been presented. The motif 291VYFAE295 appears to contribute to the preservation of this regulation and the positive modulating effect that CLN3 has on cell growth rates. Residues E295 and V330 also seem necessary for a CLN3 impact on cell growth rates and apoptosis. At site 334, it could be the charge of the residue that modulates this function. The potential glycosylation sites at residues 71–74 and 310–313 also appear to be important for the impact of CLN3 protein on cell growth rates. This indirectly supports previous claims that the CLN3 protein may be glycosylated.
According to the hydropathy profile, the glycosylation site 71NQSH74 resides within external loop 1, and the motif 291VYFAE295 and the potential glycosylation site 310NTSL313 line part of a putative pore region within transmembrane domain 7. Residues V330 and R334 are within the putative pore region. Will CLN3 turn out to be an ion pump, ion channel or transporter? Uncovering these amino acid motifs and knowledge of their contribution to the effects of the CLN3 protein on cell growth rates and apoptosis has narrowed specific molecular regions within this protein. These will facilitate the search for molecular partners of CLN3.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONCONCLUSIONSMATERIALS AND METHODSREFERENCES |
|---|
Tissue culture
Lymphoblasts were grown in 24-well plates at 1–5x104 cells/ml per well at 37°C and in 5% CO2 in RPMI 1640 media (Sigma, St Louis, MO, USA)/10% fetal bovine serum/1% antibiotic–antimycotic/1x gentamicin. Immortalized lymphoblast cell lines, originally obtained from both JNCL patients and normal donors, were used for all experiments. These cell lines were established according to previously described methods (45). All JNCL patient lymphoblasts utilized are homozygous for the 1.02 kb deletion in the genomic DNA.
Trypan blue dye exclusion
Cells were harvested from 24-well plates by centrifugation at 1200 r.p.m. for 5 min and the supernatant was removed. Each pellet was resuspended in 1 : 1 mixture of 2% trypan blue dye (Gibco, Rockville, MD, USA) and 1x phosphate-buffered saline (PBS). Cells were loaded onto the hemocytometer and counted. Viable cells are white, and dead cells are blue. Live cell counts at each time point were performed in triplicate. The mean and standard deviation were calculated for each time point. Live cell counts were plotted for growth curves as live cell countx104/ml against time. Live cell counts at each time point were then compared to those of transfection controls containing the full CLN3 cDNA or an empty vector using Student's t-test to assess statistical significance.
Transfection of immortalized JNCL lymphoblasts
Cells, 8x105, were plated in 24-well plates and incubated for 4–16 h at 37°C and in 5% CO2 with CLN3–pGEM DNA–liposome complexes. The protocols according to those outlined for the Qiagen (Valencia, CA, USA) Effectine Transfection kit and the Lipofectamine 2000 (Invitrogen-Life Technologies, Carlsbad, CA, USA) were followed. The cells were then washed and plated for analysis. The transfection efficiency was determined using the Clontech (Palo Alto, CA, USA) Luminescent ß-galactosidase Detection Kit II. Transfected samples were compared to positive and negative controls. Light emission was then recorded by exposure to X-ray film or the Xenogen plate luminometer system. Transfection efficiencies in the range 90–95% were obtained (P<0.05).
Cell proliferation/DNA synthesis assayed by thymidine incorporation
Cells, 5x104, were plated as described previously and incubated with 0.2 µl of 0.1 µCi/µl stock 3H for 2 h at 37°C and in 5% CO2. The cells were then washed twice with ice-cold PBS (Gibco-Invitrogen) and precipitated with 0.5% trichloroacetic acid on ice for 30 min. The cells were harvested and resuspended in 0.2 ml of 0.25% NaOH. The counts per minute (c.p.m.) were then measured using the Pharmacia scintillation counter (Pharmacia Corp. Peapack, NJ, USA). Cell counts/min at different time points were carried out in triplicate and the standard deviation was calculated. Statistical analysis was carried out using the Student's t-test.
Treatment with etoposide
Immortalized JNCL patient cells were transfected individually with the pGEM/CLN3 construct or with an empty vector control. The cells were then harvested and incubated with 1 µg/ml of etoposide (VP-16, Sigma) at 37°C and in 5% CO2 for 18 h to induce apoptosis. The degree of protection from etoposide-induced apoptosis afforded by each construct was determined using the JC-1 assay. Comparisons were then made with transfected cells not treated with etoposide.
JC-1 analysis of mitochondrial membrane potential
JC-1 stain (5,5'-6'-tetrachloro-1,1,3,3-tetraethylbenzimidazolylcarbocyanine iodide; Molecular probes, Eugene, OR, USA) was used to assess the decrease in 
m following treatment with etoposide. JC-1 is a cationic dye that forms J-aggregates with the cytochrome c–APAF-1 complex. Cells undergoing apoptosis are visualized when a shift in the emission spectrum from green (525 nm) to red (590 nm) occurs. This shift signals the onset of apoptosis. Immortalized lymphoblasts were grown in T-25 flasks at 5x105 in 5 ml of media and treated with etoposide as described above. The cells were then harvested, washed with 1x PBS and incubated with 1 µg/ml of JC-1 stain for 15 min at 37°C and in 5% CO2. The cells were again washed in PBS and then mounted on a slide with 1 : 1 PBS/glycerol mix. The cells in two different fields of vision were counted (60–200 cells/field) and the average numbers obtained. The degree of protection from etoposide-induced apoptosis was determined by dividing the percentage of live cells observed following treatment with etoposide by the percentage of live cells observed when not treated with etoposide. The statistical significance of the difference in the numbers of apoptotic cells (cells containing J aggregates appear yellow–red) between repeat experiments was determined by the two-tailed Student t-test.
TUNEL assay (terminal deoxynucleotide end labeling)
Sterile slides were treated overnight with poly(D) lysine (Sigma) and bordered off with a Super ht Pap pen (Research Products International Corp., Baltimore, MD, USA). Cells were plated and transfected as described previously. Cells, 1x104, were placed onto the marked region of the treated slides and allowed to adhere. The TUNEL assay allows fragmented DNA, as is observed in apoptosis, to be end-labeled by a terminal transferase enzyme with digoxigenein. The APOALERT (Clontech) kit protocol was used. The cells were then viewed at 400x magnification with a LEICA fluorescent microscope. The cells in two different fields of vision were counted (60–200 cells/field) and the average numbers obtained. The degree of protection from etoposide-induced apoptosis was determined by the percentage of live cells treated with etoposide divided by the percentage of live cells not treated with etoposide. The statistical significance of the difference in the numbers of apoptotic cells between repeats was determined by the two-tailed Student t-test.
Deletion construct generation
The CLN3 cDNA was cloned into the pGEM (+)7f(-) plasmid (Promega, Madison, WI, USA). CLN3/pGEM (+)7f(-), 10 µg, was double digested with 10–20 U each of the restriction enzymes SacI (ExoIII susceptible) and BamHI (ExoIII resistant) (New England Biolabs, Beverly, MA, USA) overnight at 37°C. The appropriate bands were then cut out of a 1% agarose gel. The linearized plasmid was then digested with ExoIII nuclease according to the Promega Erase-a-Base Kit protocol. A constant temperature of 25°C was maintained to enable accurate digestion by the exonuclease (90 base/min). The desired fragments were then individually ligated and cloned into supercompetent bacterial cells (Stratagene, La Jolla, CA, USA). Each clone was then sequenced by automated sequencing to verify that the required exons are present. These were plated for cell transfections and growth curve analysis.
Western blot
Western blot analysis was performed as previously described (22). Cells were transfected with different CLN3 cDNA constructs (described above) and harvested in lysis buffer as previously described (27). Cells were harvested at maximum CLN3p expression, 48 h post-transfection.
Site-directed mutagenesis
The Stratagene Quikchange Site-Directed Mutagenesis kit protocol was used to introduce the desired mutations in the CLN3 cDNA. The PCR cycles used were as per the protocol: 12 cycles for point mutations; 16 cycles for single amino acid changes; and 18 cycles for multiple amino acid deletion/insertion. JNCL mutations include: V330C, V330L, R334D, R334K, R334C, R334H, E295K, E295D, 
V291–E295, Y292F, Y292W and 184WSSGTGGAGLLG195. Phosphorylation mutations include deletions of the DNA sequences for amino acid residues 7, 12, 14, 19, 64, 73, 86, 125, 232, 270, 336 and 400. Glycosylation mutations include deletion of residues 49–52, 71–74, 85–88 and 310–313. The mitochondrial targeting signal mutation involved the deletion of residue 11. The isoprenylation site mutation involved the deletion of residues 435–438. Primers were made to incorporate these changes, with the mutations residing in the middle of normal CLN3 flanking sequence. All primers are 24–30 nucleotides in length.
|
ACKNOWLEDGEMENTS |
|---|
The authors wish to thank the Batten Diseases Support and Research Association for their generous support. This work was carried out in part with support from 3 RO1 NS30170-08S1 (R-M.B.), and the Duke University Genetics Training Grant 5T32GM07754 (D.P.S.).
|
FOOTNOTES |
|---|
* To whom correspondence should be addressed at: Duke University Medical Center, MSRB Box 2604, Research Drive, Durham, NC, 27710, USA. Tel: +1 9196816220; Fax: +1 9196818090; Email: boust001{at}mc.duke.edu
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONCONCLUSIONSMATERIALS AND METHODSREFERENCES |
|---|
1 Zhong, N. (2000) Neuronal ceroid lipofuscinoses and possible pathogenic mechanism. Mol. Gen. Met., 71, 195–206.
2 The International Batten Consortium (1995) Isolation of a novel gene underlying Batten disease, CLN3. Cell, 82, 949–957.[Web of Science][Medline]
3 Boustany, R-M. (1996) Batten disease or neuronal ceroid lipofuscinosis. In Vinken, P.J., Bruyn, G.W. and Moser, H.W. (eds), Handbook of Clinical Neurology, Neurodystrophies and Neurolipidoses. Vol. 66, Elsevier Sciences, New York, pp. 671–900.
4 Goebel, H.H. (1997) Morphologic diagnosis in neuronal ceroid lipofiscinoses. Neuropediatrics, 28, 67–69.[Web of Science][Medline]
5 Schmechel, D.E. (1999) Apoptosis in neurodegenerative disorders. In Hannun, Y. and Boustany, R.-M. (eds), Apoptosis in Neurobiology. CRS Press, Boca Raton, FL, pp. 29–30.
6 Kida, E., Golabek, A. and Wisniewski, K.E. (2001) Cellular pathology and pathogenic aspects of neuronal ceroid lipofuscinses. Adv. Genet., 45, 35–68.[Medline]
7 Johnson, D.W., Speier, S., Qian, W.-H., Lane, S., Cook, A., Suzuki, K., Daniel, P. and Boustany R.-M. (1995) Role of subunit 9 of mitochondrial ATP synthase in Batten disease. Am. J. Med. Genet., 57, 350–360.[Web of Science][Medline]
8 Pullarkat, R.K., Kim, K.S., Sklower, S.L. and Patel, V.K. (1998) Oligosaccharyl diphosphodolichols in the ceroid-lipofuscinoses. Am. J. Med. Genet. suppl. 5, 243–251.
9 Sleat, D.E., Donnelly, R.J., Lackland, H., Liu, C.G., Sohar, I., Pullarkat, R.K. and Lobel, P.S. (1997) Association of mutations in a lysosomal protein with classical late-infantile neuronal ceroid lipofuscinosis. Science, 277, 5333, 1802–1805.
10 Vesa, J., Hellsten, E., Verkruyse, L.A., Camp, L.A., Rapola, J., Santavuori, P., Hofmann, S.L. and Peltonen, L. (1995) Mutations in the palmitoyl protein thioesterase gene causing infantile neuronal ceroid lipofuscinosis. Nature, 376, 584–587.[Medline]
11 Janes, R.W., Munroe, P.B., Mitchison, H.A., Gardiner, R.M., Mole, S.E. and Wallace, B.A. (1996) A model for Batten disease protein CLN3: functional implications from homology and mutations. FEBS Lett., 399, 75–77.[Web of Science][Medline]
12 Savukoski, M., Klockars, T., Holmberg, V., Santavuori, P., Lander, E.S. and Peltonen, L. (1998) CLN5, a novel gene encoding a putative transmembrane protein mutated in Finnish variant late infantile neuronal ceroid lipofuscinosis. Nat. Genet., 19, 286–288.[Web of Science][Medline]
13 Gao, H., Boustany, R.-M., Espinola, J.A., Cotman, S., Srinidhi, L., Antonellis, K.A., Gillis, T., Stout, D., Haines, J.L., Lerner, T.J. and MacDonald, M.E. (2002) Mutations in a novel CLN6 encoded transmembrane protein cause variant neuronal ceroid lipofuscinosis in mouse and man. Am. J. Hum. Genet., 70, 324–335.[Web of Science][Medline]
14 Ranta, S., Zhang, Y., Ross, B., Lonka, L., Takkunen, E., Messer, A., Sharp, J., Wheeler, R., Kusumi, K., Mole, S.E. et al. (1999) The neuronal ceroid lipofuscinoses in human EPMR and mnd mutant mice are associated with mutations in CLN8. Nat. Genet., 23, 233–236.[Web of Science][Medline]
15 Margraf, L.R., Boriack, R.L., Routheut, A.A., Cuppen, I., Alhilali, L., Bennett, C.J. and Bennett, M.J. (1999) Tissue expression and subcellular localization of CLN3, the Batten disease protein. Mol. Gen. Met., 66, 283–289.
16
Haskell, R.E., Carr, C.J., Pearce, D., Bennett, M.J. and Davidson B.L. (2000) Batten disease: evaluation of CLN3 mutations on protein localization and function. Hum. Mol. Genet., 9, 735–744.
17
Jarvela, I., Sainio, M., Rantmaki, T., Olkkonen, V.M., Carpen, O., Peltonen, L. and Janes, R.W. (1998) Biosynthesis and intracellular targeting of the CLN3 protein defective in Batten disease. Hum. Mol. Genet., 7, 265–271.
18
Katz, M.L., Gao, C.-L., Prabhakaram, M., Shibuya, H., Liu, P.-C. and Johnson, G.S. (1997) Immunochemical localization of the Batten disease (CLN3) protein in retina. Invest Ophthalmol Vis. Sci., 38, 2375–2386.
19
Kremmidiotis, G., Lensink, I., Bilton, R.L., Woolatt, E., Chataway, T.K., Sutherland, G.R. and Callen, D.F. (1999) The Batten disease gene product (CLN3p) is a Golgi integral membrane protein. Hum. Mol. Genet., 8, 523–531.
20
Jarvela, I., Lehtovirta, M., Tikkanen, R., Kyttala, A. and Jalanko, A. (1999) Defective intracellular transport of CLN3 is the molecular basis of Batten disease. Hum. Mol. Genet., 8, 1091–1098.
21
Luiro, K., Kopra, O., Lehtovirta, M. and Jalanko, A. (2001) CLN3 protein is targeted to neural synapses but excluded from synaptic vesicles: new clues to Batten disease. Hum. Mol. Genet., 10(19), 2123–2131.
22 Puranam, K.L., Guo, W.-X., Qian, W.-H., Nikbakht, K. and Boustany, R.-M. (1999) CLN3 defines a novel antiapoptotic pathway operative in neurodegeneration and mediated by ceramide. Mol. Gen. Met., 66, 294–308.
23 Dhar, S., Bitting, R.L., Rylova, S., Jansen, P.J., Lockhart, E., Koeberl, D.D., Amalfitano, A. and Boustany, R.-M. (2001) Flupirtine blocks apoptosis in Batten patient lymphoblasts and in human post-mitotic CLN3- and CLN2-deficient neurons. Ann. Neurol., 51, 448–466.
24 Lane, S.C., Jolly, R.D., Schmechel, D.E., Alroy, J. and Boustany, R.-M. (1996) Apoptosis as the mechanism of neurodegeneration in Batten's disease. J. Neurochem., 67, 677–683.[Web of Science][Medline]
25 Kieseier, B.C., Wisniewski, K.E., Park, E., Schiller-Levis, G., Mehta, P.D. and Goebel, H.H. (1997) Leukocytes in neuronal ceroid lipofuscinoses; function and apoptosis. Brain Dev., 19, 317–322.[Web of Science][Medline]
26 Pane, M.A., Puranam, L. and Boustany, R.-M. (1999) Expression of cln3 in human NT2 neuronal precursor cells and neonatal rat brain. Pediatr. Res., 46, 367–374.[Web of Science][Medline]
27
Rylova, S.N., Amalfitano, A., Persaud-Sawin, D.A., Guo, W.-X., Chang, J., Jansen, P.J., Proia, A.D. and Boustany, R.-M. (2002) The CLN3 gene is a novel molecular target for cancer drug discovery. Cancer Res., 62, 801–808.
28 Guo, W.-X., Mao, C., Obeid, L. and Boustany, R.-M. (1999) A disrupted homolog of the CLN3 or Juvenile Neuronal Ceroid Lipofuscinosis gene in Saccharomyces cerevisiae. Cell. Mol. Neurol., 19, 671–680.
29 Pearce, D.A., Ferea, T., Nosel, S.A., Das, B. and Sherman, F. (1998) Action of BTN1, the yeast ortholog of the gene mutated in Batten disease. Nat. Genet., 22, 55–58.
30 Matsuyama, S., Llopis, J., Devereaux, Q.L., Tsien, R.Y. and Reed, J.C. (2000) Changes in intramitochondrial and cytosolic pH: early events that modulate caspase activation during apoptosis. Nat. Cell Biol., 28, 18–20.
31 Mole, S.E., Mitchison, H.A. and Munroe, P.B. (1999) Molecular basis of the neuronal ceroid lipofuscinoses: mutations in CLN1, CLN2, CLN3 and CLN5. Hum. Mut., 14, 199–215.[Web of Science][Medline]
32 Mole, S.E., Zhong, N., Saprong, A., Logan, W.P., Hofmann, S.L., Zhong, N., Franken, P.F., Van Diggelen, O.P., Bruening, M.H., Moroziewicz, D.N. et al. (2001) New mutations in the neuronal ceroid lipofuscinosis genes. Eur. J. Pediatr. Neurol., 5, 7–10.
33 Munroe, P.B., Mitchison, H.A., O'Rawe, A.M., Anderson, J.W., Boustany, R.-M., Lerner, T.J., Taschner, P.E.M., de Vos, N., Bruening, M.H., Gardiner, R.M. and Mole, S.E. (1997) Spectrum of mutations in the Batten disease gene, CLN3. Am. J. Hum. Genet., 61, 310–316.[Web of Science][Medline]
34 Taschner, P.E.M., de Vos, N. and Bruening, M.H. (1997) Cross-species homology of the CLN3 gene. Neuropediatrics, 28, 18–20.[Web of Science][Medline]
35 Kerr, I.D., Ranatunga, M.K. and Sansom, M.S.P. (1999) The voltage gated potassium channel: sequence analysis and pore modeling of the pore domain. Perspect. Drug Disc. Des., 15/16, 187–214.
36 Gelbard, H.A., Boustany, R.-M. and Schor, N.F. (1997) Apoptosis in development and disease of the nervous system: II. Apoptosis in childhood neurologic disease. Pediatr. Neurol., 16, 93–97.[Web of Science][Medline]
37
Miravalle, L., Tokuda, T., Chiarle, R., Giaccone, G., Bugiani, O., Tagliavini, F., Frangione, B. and Ghiso, J. (2000) Substitution at codon 22 of the Alzheimer's A-beta peptide induces conformational changes and apoptotic effects in the human cerebral endothelial cells. J. Biol. Chem., 275, 27110–27116.
38 Michalewski, M.P., Kaczmarski, W., Golabek, A., Kida, E., Kaczmarski, A.L. and Wisniewski, K.E. (1998) Evidence for phosphorylation of CLN3 protein associated with Batten disease. Biochem. Biophys. Res. Commun., 253, 458–462.[Web of Science][Medline]
39 Michalewski, M.P., Kaczmarski, W., Golabek, A., Kida, E., Kaczmarski, A.L. and Wisniewski, K.E. (1999) Posttranslational modification of CLN3 protein and its possible functional implication. Mol. Gen. Met., 66, 272–276.
40 Facompre, M., Wattez, N., Kluza, J., Lansiaux, A. and Bailly, C. (2000) Relationship between cell cycle changes and variations of the mitochondrial membrane potential induced by etoposide. Mol. Cell Biol. Res. Commun., 4, 37–42.[Medline]
41 Sawada, M., Nakashima, S., Banno, Y., Kamakawa, H., Hayashi, K., Tanaka, K., Nishimura, Y., Saki, N. and Nozawa, Y. (2000) Ordering of ceramide activation and Bax/Bcl-2 expression during etoposide induced apoptosis in C6 glioma cells. Cell Death Diff., 7, 761–772.[Web of Science][Medline]
42 Cervinka, M., Bereiter-Hahn, J., Peychl, J., Rudolf, E. and Cervinkova, Z. (1999) The role of the mitochondria in apoptosis induced in vitro. Gen. Phys. Biophys., 18, 33–40.
43 Pullarkat, R.K. and Morris, G.N. (1997) Farnesylation of Batten disease CLN3 protein. Neuropediatrics, 28, 42–44.[Web of Science][Medline]
44 Zhong, N., Wisniewski, K.E., Kaczmarski, A.L., Ju, W., Xu, W.M., Xu, W.W., Melendon, L., Liu, B., Kaczmarski, W., Brooks, S.S. and Brown, W.T. (1998) Molecular screening of Batten disease: identification of a mis-sense mutation (E295K) in the CLN3 gene. Hum. Genet., 102, 57–62.[Web of Science][Medline]
45 Anderson, M.A. and Gusella, J.F. (1984) Use of cyclosporine A in establishing Epstein–Barr virus-transformed human lymphoblastoid cell lines. In Vitro, 20, 856–858.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Codlin and S. E. Mole S. pombe btn1, the orthologue of the Batten disease gene CLN3, is required for vacuole protein sorting of Cpy1p and Golgi exit of Vps10p J. Cell Sci., April 15, 2009; 122(8): 1163 - 1173. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Codlin, R. L. Haines, J. Jemima, E. Burden, and S. E. Mole btn1 affects cytokinesis and cell-wall deposition by independent mechanisms, one of which is linked to dysregulation of vacuole pH J. Cell Sci., September 1, 2008; 121(17): 2860 - 2870. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R.-M. Boustany, J. W. Britton, J. E. Parisi, and B. M. Keegan A 23-year-old man with seizures and visual deficit Neurology, January 1, 2008; 70(1): 73 - 78. [Full Text] [PDF]
|
|
|
|
 |
|
 J.-W. Chang, H. Choi, H.-J. Kim, D.-G. Jo, Y.-J. Jeon, J.-Y. Noh, W. J. Park, and Y.-K. Jung Neuronal vulnerability of CLN3 deletion to calcium-induced cytotoxicity is mediated by calsenilin Hum. Mol. Genet., February 1, 2007; 16(3): 317 - 326. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Ramirez-Montealegre, P. G Rothberg, and D. A Pearce Another disorder finds its gene Brain, June 1, 2006; 129(6): 1353 - 1356. [Full Text] [PDF]
|
|
|
|
 |
|
 S. Storch, S. Pohl, and T. Braulke A Dileucine Motif and a Cluster of Acidic Amino Acids in the Second Cytoplasmic Domain of the Batten Disease-related CLN3 Protein Are Required for Efficient Lysosomal Targeting J. Biol. Chem., December 17, 2004; 279(51): 53625 - 53634. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. B. Narayan, J. V. Pastor, H. M. Mitchison, and M. J. Bennett CLN3L, a novel protein related to the Batten disease protein, is overexpressed in Cln3-/- mice and in Batten disease Brain, August 1, 2004; 127(8): 1748 - 1754. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Kyttala, G. Ihrke, J. Vesa, M. J. Schell, and J. P. Luzio Two Motifs Target Batten Disease Protein CLN3 to Lysosomes in Transfected Nonneuronal and Neuronal Cells Mol. Biol. Cell, March 1, 2004; 15(3): 1313 - 1323. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||