Regulatory, Integrative and Comparative Physiology

High sodium chloride intake decreases betaine-homocysteine S-methyltransferase expression in guinea pig liver and kidney

Cassandra V. Delgado-Reyes, Timothy A. Garrow


Betaine-homocysteine S-methyltransferase (BHMT) is the only enzyme known to catabolize betaine. In addition to being a substrate for BHMT, betaine also functions as an osmoprotectant that accumulates in the kidney medulla under conditions of high extracellular osmolarity. The mechanisms that regulate the partitioning of betaine between its use as a methyl donor and its accumulation as an osmoprotectant are not completely understood. The aim of this study was to determine whether BHMT expression is regulated by salt intake. This report shows that guinea pigs express BHMT in the liver, kidney, and pancreas and that the steady-state levels of BHMT mRNA in kidney and liver decrease 68% and 93% in guinea pigs consuming tap water containing high levels of salt compared with animals provided untreated tap water. The animals consuming the salt water also had ∼50% less BHMT activity in the liver and kidney, and steady-state protein levels decreased ∼30% in both organs. Pancreatic BHMT activity and protein levels were unaffected by the high salt treatment. The complex mechanisms involved in the downregulation of hepatic and renal BHMT expression in guinea pigs drinking salt water remain to be clarified, but the physiological significance of this downregulation may be to expedite the transport and accumulation of betaine into the kidney medulla under conditions of high extracellular osmolarity.

  • choline
  • hypernatremia
  • osmolyte
  • tonicity

betaine-homocysteine s-methyltransferase (BHMT, EC catalyzes a methyl transfer from betaine (Bet) to homocysteine to form dimethylglycine and methionine (Met), respectively. BHMT has been shown to be expressed at high levels in the liver of all animals tested (26), but its expression outside this organ is species dependent and has not been exhaustively evaluated. It is known that rats only express BHMT in the liver, whereas other species, including guinea pigs (GPs) (7), monkeys (25), and pigs and humans (26), also express this enzyme in the kidney. BHMT expression in the kidney of humans and pigs has been shown to be limited to the proximal tubules of the cortex (6). Notable exceptions to the general rule that BHMT is a liver and kidney enzyme include the high expression of this protein in the lenses of rhesus monkeys (23) and in the pancreas of sheep (31).

There are two known metabolic roles for Bet: methyl donor and osmoprotectant. As a methyl donor for the BHMT-catalyzed reaction, Bet has a role in sulfur amino acid homeostasis. The methylation of homocysteine is physiologically important because Met is an essential dietary amino acid and is needed for protein synthesis and S-adenosylmethionine biosynthesis. The regeneration of Met from homocysteine (referred to as remethylation) precludes homocysteine from participating in the transsulfuration pathway, essentially sparing Met from catabolism. It is believed that the oxidation of Bet can be quantitatively important in human one-carbon metabolism. Elevating Bet consumption is an effective plasma homocysteine-lowering treatment for homocystinuria (29, 30), and supplementing choline in the diet of humans reduces the de novo synthesis of one-carbon units from other precursors (18, 19), presumably serine and glycine. A physiological role for BHMT as an enzyme that spares Met from catabolism is supported by our research in rats that showed that BHMT gene expression is dramatically upregulated during Met deficiency (20).

Besides being a methyl donor, Bet is also one of the major organic osmolytes that accumulates in a variety of cells under conditions of hypertonicity (13), the most investigated being the kidney medulla. Several mechanisms are available to the medulla to accumulate this compound. It has been shown that the expression of the γ-aminobutyrate-Bet transporter of medullary cells dramatically increases under conditions of hyperosmolarity (32). This suggests that the ability of the medulla to oxidize choline to Bet is not sufficient to meet the need for Bet as an osmolyte under these conditions, even if Bet synthesis in liver and/or kidney increases with renal hypertonicity. In fact, a study in which rats were injected with [14C]choline into the renal artery indicated that the majority of Bet accumulated in the medulla is synthesized in the cortex, which is then transported to the medulla by the tubular and/or vascular route (17). This work is supported by the observation that in hypernatraemic rats the activity of renal choline dehydrogenase activity increases, with induction levels observed in the cortical region that are higher than in the inner medulla (12). Rats do not express BHMT in the kidney, and in humans and pigs the expression of BHMT is strictly localized to the proximal tubules (6). These facts eliminate the possibility that Bet catabolism in the medulla is a mechanism to regulate Bet concentrations in this region of the kidney. In addition, Bet has been shown to rapidly efflux out of medullary cells when hyperosmotic conditions are replaced with hypoosmotic conditions (2). In summary, changes in the rate of Bet synthesis, transport, and efflux have all been proposed to have a role in the regulation of Bet concentrations in the medulla, but the extent each factor plays could vary across species and the degree of osmotic pressure in the medulla.

We thought that an elaboration of the regulatory mechanisms to enhance the ability of the renal medulla to accumulate Bet during hyperosmotic conditions might downregulate BHMT activity so as to decrease its use as a methyl donor, thus making it more available for transport. As a first approach to test this hypothesis, we investigated the effect of high salt consumption on BHMT expression. We chose to use the GP as an animal model because, like humans, it expresses BHMT in the kidney, an organ that has a central role in osmotic balance.



Male GPs were purchased from Harlan (Indianapolis, IN). Primers and dual-labeled fluorescent probes were purchased from Integrated DNA Technologies (Coralville, IA). TRIzol reagent, Superscript II RNase H-reverse transcriptase, pCR-XL-TOPO cloning vector, and One Shot Top10 competent cells were supplied by Invitrogen Life Technologies (Carlsbad, CA). Perfect RNA eukaryotic kits and Taq DNA polymerase were purchased from Eppendorf (Westbury, NY). The TaqMan Universal PCR master mix was obtained from Applied Biosystems (Foster City, CA). [methyl-14C]choline chloride (56 mCi/mmol) was obtained from Amersham (Piscataway, NJ). Anti-BHMT polyclonal antibodies (rabbit) were prepared as described previously (97), and peroxidase-labeled goat anti-rabbit IgG secondary antibody was supplied by Vector Laboratories (Burlingame, CA). Western Lightning Western blot chemiluminescence reagent plus was purchased from Perkin Elmer (Wellesley, MA). All other reagents were of the highest purity commercially available.


One male GP (chow fed) was euthanized by decapitation, and its liver, small intestine, kidneys, lungs, pancreas, and spleen were excised, fixed in 10% buffered formalin (12 h), and then stored in 70% ethanol until embedded in paraffin, sectioned into 3-mm slices, and mounted onto glass slides. We stained slices for BHMT using the peroxidase avidin-biotin system previously described (6). Diaminobenzidine served as the peroxidase substrate, and all samples were counterstained with hematoxylin.

Animal protocol.

Ten male GPs were housed individually in shoebox cages and exposed to standard lighting (12:12-h light-dark) and temperature (23°C) conditions. They had free access to tap water and standard chow during a 5-day acclimation period, after which the animals weighed an average of 401 g and were randomized into two treatment groups (n = 5). The experimental treatments were free access to either tap water (control) or NaCl (500 mM)-containing tap water, similar to hyperosmotic protocols used for rats (1, 22). Water consumption was recorded daily. The chow intake of the control GPs was limited to the average amount consumed by the treatment group. Treatment lasted 72 h, at which time the animals were euthanized by decapitation and their livers, kidneys, and pancreas were excised, flash-frozen in liquid nitrogen, and stored at −80°C until analyzed. The animal procedures used in this study were approved by the Institutional Animal Care and Use Committee of the University of Illinois (Urbana-Champaign, IL).

cDNA cloning of GP BHMT.

We isolated total liver RNA using the TRIzol reagent per the manufacturer's instructions and stored it in Formazol at −80°C. First-strand cDNA synthesis was performed by using total RNA, oligo(dT) primer, and Superscript II RNase H-reverse transcriptase according to the manufacturer's instructions. The liver cDNA library was then stored at −20°C.

The following oligonucleotide primers were designed from the cDNA sequence of human liver BHMT: 5′-TTTGTCTTTGCACTGGAGAAG-3′ (sense) and 5′-CAAAGAGCTCTTTCAGCTGCTGC-3′ (antisense). These primers were used to amplify a major portion of the open reading frame of GP BHMT by PCR, using the liver cDNA, prepared as described above, as template. A single PCR product of 1.1 kb was observed on an agarose gel and ligated into pCR-XL-TOPO and transformed into One Shot Top10 competent cells. Both strands of the resulting plasmid insert of pCR-XL-gpBHMT were sequenced.

Relative quantification of GP BHMT and actin mRNA by TaqMan real-time PCR.

We performed isolation of total GP liver and kidney cortex RNA using the Perfect RNA eukaryotic kit per the manufacturer's instructions. Synthesis of cDNA was performed as described above. Primers and dual-labeled probe combinations were designed using PrimerExpress software (version 2.0, ABI). The following oligonucleotide primers for PCR amplification of BHMT were synthesized: 5′-GGCAACTACGTGGCAGAGAAG-3′ (sense) and 5′-CCTGCCGGGCAATGTC-3′ (antisense). The following primers for β-actin PCR amplification were synthesized: 5′-TGTGGATCGGCGGCTCTA-3′ (sense) and 5′-GGGCCCGACTCATCGTACT-3′ (antisense). For real-time cDNA amplification, the following 5′-6-FAM/3′ TAMRA probes for BHMT and β-actin were used: 5′-/56-FAM/TATCTGGGCAAAAAGTCAACGAAGCTGC/36-TAMTph/-3′ (BHMT) and 5′-/56-FAM/CCTGGCCTCACTCTCCACCTTCCA/36-TAMTph/-3′ (β-actin). Real-time PCR was performed in duplicate on an ABI Prism 7700 sequence detection system. The relative abundance of amplified cDNA was calculated as 2−ΔCt, where ΔCt (change in cycle threshold) equals Ct of BHMT target − Ct of β-actin loading control. Results are expressed as mean relative BHMT mRNA/β-actin mRNA values.

Enzyme assays.

GP livers and pancreas were sectioned on dry ice, weighed, and homogenized in 5 vol of buffer containing 30 mM Tris (pH 7.9), 2 mM EDTA, and 5 mM β-mercaptoethanol using a glass and Teflon homogenizer. Renal cortical tissue was separated from medullary tissue and homogenized as described above. Homogenates were centrifuged for 1 h at 16,000 g (4°C). The supernatants were assayed for total protein and BHMT activity as described previously (9).

Choline dehydrogenase activity was assayed with the use of a modified protocol (14). In summary, we homogenized frozen tissue in 3 vol of buffer containing 10 mM Tris (pH 7.9), 250 mM sucrose, 1 mM EDTA, and 5 mM β-mercaptoethanol using a glass and Teflon homogenizer. The homogenates were centrifuged for 10 min at 1,000 g (4°C), and the recovered supernatants were again centrifuged for 15 min at 3,500 g (4°C) to produce a pellet containing heavy mitochondria. We used the Bradford method to determine total mitochondrial protein concentration. Choline dehydrogenase activities were measured with 0.1 mg of mitochondrial protein in buffer containing 3.5 mM Tris (pH 7.9), 350 nM EDTA, 350 nM β-mercaptoethanol, 500 μM [methyl-14C]choline (0.2 μCi), 0.3% (m/v) phenazine methosulfate, and 350 μM CaCl2 (50 μl final volume). The mixtures were incubated for 15 min at 37°C, and then the reactions were stopped by transferring the tubes to ice water. The product, Bet aldehyde, was converted to Bet by adding 20 μl of 1 N NaOH and 15 μl of 30% hydrogen peroxide and allowing the mixture to stand for 1 h at room temperature. Water was then added to a final volume of 500 μl, and the reaction mixtures were loaded onto minicolumns containing 1 ml of Dowex 50W-40X (H+) resin. The resin was washed three times with 833 μl of ice-cold water. All flow-through fractions were collected (3 ml); to this eluant, we added 17 ml of Scintisafe scintillation fluid. The degradations per minute were then measured by scintillation counting. Blanks contained all of the reaction components except enzyme, and their counts were subtracted from sample values. All samples were assayed in triplicate. Specific activity (U/mg) was expressed as nanomoles Bet produced per hour per milligram of mitochondrial protein.

Western analysis for BHMT.

Tissues were homogenized, and protein concentrations were determined as described above. Extracts containing 0.5 μg of liver protein or 25 μg of kidney protein were subjected to SDS-PAGE and transferred to a nitrocellulose membrane using a semi-dry transfer system. Separated protein was probed for BHMT as previously described (6). BHMT protein was visualized with chemiluminescence, and the optical densities of the bands were quantified with Image Station 440 from Eastman Kodak (Rochester, NY). The results for each group are expressed as the mean optical density. Results for the salt-treated animals are expressed relative to those for the control animals, whose average value was defined as 100%.

Statistical methods.

Data were analyzed by a Student's t-test using the Simple Interactive Statistical Analysis online software (SISA: Results are expressed as means ± SE. Mean values were considered significantly different at P < 0.05.


Immunohistochemical localization of GP BHMT.

Consistent with our prior study using human, rat, and pig liver samples (6), immunohistochemical staining of GP liver indicates that BHMT is abundantly expressed in the cytoplasm of the hepatocytes (not shown). Like humans and pigs, GP kidney BHMT is only expressed in the three segments of the cortical proximal tubules (Fig. 1A). No staining could be detected in the medulla. There is significant staining in the cytoplasm of the acinar cells of the exocrine pancreas (Fig. 1B). Humans, rats, and pigs do not express BHMT in this organ (6), but sheep have been reported to have high levels of BHMT activity in the pancreas (31). No BHMT is detected in GP lung, spleen, or any segment of the small intestine (not shown). BHMT cannot be detected in these tissues from human, rat, and pig as well (6, 26).

Fig. 1.

Immunohistochemical detection of betaine-homocysteine S-methyltransferase (BHMT) protein in guinea pig (GP) kidney cortex and pancreas (×400 magnification). With the use of polyclonal anti-BHMT antibodies, positive staining for BHMT protein (brown) in the GP kidney (A) demonstrates expression to be restricted to the proximal tubules (PT) of the renal cortex with no expression in the glomerulus (G), or distal tubules (DT). Positive staining for BHMT protein (brown) in the GP pancreas (B) demonstrates expression in the acinar cells (AC) but no expression in the islets of Langerhans (IL), artery (A), vein (V), ducts (D), or lumen (L). Staining protocols are described in experimental procedures.

Effect of salt intake on GP water and food intake and body weight.

Over the 3-day treatment period, the GPs that were given free access to salt water drank about one-half the volume of fluid consumed by the animals given free access to tap water. From calculations of the level of sodium in GP food and that added to the water, we found that the control and salt-loaded animals had an average daily sodium intake of 0.15 and 2.9 g/kg body wt, respectively. During the 3-day experiment, the control GPs maintained their body weight despite their intake being restricted to the levels consumed by the salt-loaded animals, the latter of which lost 11.0% of their initial body weight.

Cloning and sequencing of GP liver BHMT.

A near full-length open reading frame encoding a GP BHMT cDNA was sequenced.1 There is 83% nucleotide and 91% amino acid identity across the 1,198-nucleotide alignment of GP and human BHMT. All of the residues known to be essential for human BHMT activity [Gly 214 (3); Cys 217, 299, and 300 (4); and Tyr 77 (5, 10)] are conserved in the GP sequence. In addition, those residues shown to facilitate substrate and/or product binding to human BHMT are conserved in the GP enzyme, namely, Trp 44 (5), Glu 159 (5, 10), Tyr 160 (5, 11), and His 338 (27).

Effect of high salt intake on steady-state levels of BHMT mRNA.

Steady-state BHMT mRNA levels in liver and kidney cortex were measured with TaqMan real-time PCR. When corrected for β-actin mRNA, both liver and kidney cortex BHMT mRNA were dramatically lower in GPs consuming salt water compared with those consuming tap water (Fig. 2). Compared with the tap water-consuming control animals, steady-state mRNA levels decreased 93% in liver and 68% in kidney cortex with high salt intake.

Fig. 2.

Effect of high salt intake on BHMT steady-state mRNA levels in GP liver and kidney cortex. GPs were provided free access (72 h) to either tap water (control) or salt water (500 mM NaCl). At the end of the treatment period, total RNA from liver and kidney cortex were isolated and reverse transcribed to cDNA; GP BHMT and β-actin (internal control) cDNAs were then quantified by real-time PCR, as described in experimental procedures. Results are expressed as mean BHMT mRNA/β-actin mRNA (±SE), normalized relative to the control animals (mean equals 100%). *Significant difference at P < 0.05. **Significant difference at P < 0.01.

Effect of high salt intake on GP BHMT and choline dehydrogenase activity.

The GPs consuming the salt water had about one-half of the liver and kidney cortex BHMT activity of those animals consuming tap water (Fig. 3). Consistent with our immunohistochemistry results, GP pancreas expresses a considerable level of BHMT activity, but this activity was unaffected by salt consumption (15.6 ± 0.9 vs. 17.1 ± 0.9 U/mg). Consumption of salt water also had no effect on the level of choline dehydrogenase activity measured in liver (1.6 ± 0.2 vs. 1.3 ± 0.1 U/mg) or kidney cortex (32.4 ± 1.2 vs. 33.1 ± 1.5 U/mg).

Fig. 3.

Effect of high salt intake on hepatic and renal BHMT activities and steady-state protein levels. GPs were provided free access (72 h) to either tap water (control) or salt water (500 mM NaCl). At the end of treatment period, liver and kidney cortex were assayed for BHMT activity and BHMT protein content as described in experimental procedures. Results are expressed as means ± SE, normalized to the control animals (mean equals 100%). Control GP liver and kidney cortex BHMT activities were 42.5 ± 2.0 and 12.4 ± 0.6 U/mg, respectively. *Significant difference at P < 0.05. **Significant difference at P < 0.01.

Effect of high salt intake on steady-state levels of GP BHMT protein.

We performed Western blot analyses on liver and kidney BHMT homogenates to determine whether the salt-dependent reduction of BHMT activity in these organs was due to a decrease in BHMT protein content. In both organs, the high salt treatment decreased BHMT protein levels ∼30% compared with the control group (Fig. 3).


The kidney is the most common organ studied with respect to the accumulation of organic osmolytes. This organ has an obvious need for osmotic protection because its extracellular solute concentration can vary greatly as it functions to regulate whole body electrolyte and water balance. However, the liver is also susceptible to osmotic stress. Elevated salt intake has been shown to increase portal blood and extracellular hepatocyte osmolarity (21). BHMT is expressed in both the liver and kidney of humans and GPs but only in the liver of rats. Therefore, it may be that the GP is a better animal to model the role of BHMT in the osmotic regulation of Bet concentrations in human kidney.

In this report, we show that BHMT expression in GPs is affected by salt consumption. Compared with animals consuming tap water, those consuming salt water had a dramatic reduction in steady-state levels of BHMT mRNA in liver and kidney cortex, as well as significant reductions in the level of BHMT activity and protein content in both organs. The much larger decrease in mRNA compared with protein content suggests that the BHMT protein has a considerably longer half-life than the BHMT mRNA under these conditions.

Whether the reduction in liver and kidney BHMT mRNA was due to decreased transcription and/or increased degradation was not investigated. It is known that some genes involved in osmotic regulation are affected at the level of transcription. For example, the 5′-flanking regions of the γ-aminobutyrate-Bet transporter, aldose reductase and myo-inositol transporter 1 genes, have been shown to contain a transcription factor binding sequence called the tonicity responsive regulatory element (TonE) (2, 13). When the TonE binding protein binds to this sequence, transcription increases. Whether a functional TonE sequence regulates BHMT expression is unknown, but if it does it would be the first example of this system decreasing transcription. It may also be that BHMT transcription is regulated by hormones, e.g., antidiuretic hormone, in response to changes in fluid and solute intake. These possibilities warrant further investigation.

The larger reduction in BHMT activity compared with BHMT protein levels in the salt-loaded animals compared with controls could represent a change in the catalytic efficiency of the enzyme. To date, the only known soluble effectors of mammalian BHMT are its products, dimethylglycine and Met, although it is possible that a heretofore unknown effector is differentially influencing BHMT activity in salt-treated vs. control GPs. To determine whether a soluble effector was influencing BHMT differently in the two treatment groups, we mixed equal volumes of individual liver or kidney extracts prepared from GPs from the control (water) or high salt treatment groups and assayed for BHMT activity. We observed that the BHMT-specific activities of the mixed samples were equal to the arithmetic mean of the specific activities obtained from the individual extracts assayed alone. This observation indicates that the differences observed in BHMT activity between the experimental groups and the observation that activity levels decreased to a larger degree than the levels of immunodetectable protein in the salt-loaded GPs were not due to any soluble effector present in one extract and not the other. The possibility that the catalytic efficiency of BHMT is affected by some unknown covalent modification(s) or some other more complex mode of regulation was not tested and warrants further investigation. For example, hypertonicity has been shown to activate phosphorylation events (13). There are no reports indicating that the catalytic efficiency of BHMT is regulated by phosphorylation, but this possibility cannot be eliminated at this time. Furthermore, there is one report that used Western analysis to detect BHMT following two-dimensional electrophoresis of protein extract made from monkey lenses, which suggested that BHMT is subject to posttranslational modification (23). This report did not characterize the modification(s) to BHMT or whether the modified form(s) differed in terms of its catalytic efficiency.

In addition to kidney and liver expression, BHMT is also expressed at high levels in the acinar cells of the GP pancreas (Fig. 1B). We are not aware of any report that has assayed for BHMT activity and/or protein in human pancreas, but no BHMT mRNA could be detected by Northern analysis using a commercially prepared blot containing human pancreatic mRNA. It is known that sheep express very high levels of BHMT activity in the pancreas (31) but that rats and pigs have no measurable activity in this organ (6, 8). Hence, like kidney BHMT expression, pancreatic expression is species dependent. Despite the large reductions of BHMT expression in the liver and kidney cortex of salt-loaded GPs, there was no change in the BHMT activity measured in the pancreas. The lack of change in activity may be due to the absence of the necessary osmotic response factors for BHMT in the pancreas, or perhaps the pancreas is not subject to changes in tonicity due to high salt intake.

The level of liver and kidney choline dehydrogenase activity was also measured in this study. Choline dehydrogenase catalyzes the first and committed reaction of two enzymatic steps required to convert choline to Bet, but no change in activity was observed in the salt-loaded GPs compared with the control group. This is in contrast to a report that indicated that hypertonicity induces a 50% increase in choline dehydrogenase activity in rat kidney cortex (12). However, we cannot directly compare our results with those obtained in the aforementioned study because we did not measure blood osmolality. Furthermore, no statement can be made as to whether the flux from choline to Bet changes in GP liver and/or kidney in response to changes in salt intake because enzyme activity levels only measure reaction capacity. Similarly, it is not possible to know whether the reduction of BHMT expression measured in the salt-loaded GPs causes any change in flux through the BHMT-catalyzed reaction relative to the control animals.

The regulation of BHMT activity by osmolarity has precedence in lower organisms. It has been reported that the halophytic cyanobacteria, A. halophytica, accumulates Bet as a mechanism to survive in high salt media and that it regulates intracellular Bet concentrations by altering BHMT activity (15). In this organism, BHMT activity is low when cells are grown in high salt media, but it is rapidly elevated when the salt is removed from the media. The induction of BHMT activity is followed by a rapid reduction of intracellular Bet. It was not determined whether changes in BHMT activity in this organism were due to changes in transcription, translation, or catalytic efficiency. However, in a separate study, it was reported that the BHMT protein from A. halophytica is sensitive to salt concentrations (28). When the purified enzyme is assayed in the presence of increasing levels of NaCl, it progressively loses activity such that when 200 mM NaCl is reached it has lost all activity (8). It was not reported whether this is specific effect of one of the monovalent ions or whether it is a general effect of ionic strength. However, it is known that concentrations of sodium and chloride in A. halophytica can vary from 80–180 mM (24) and 35–150 mM (16), respectively, depending on culture conditions. Regardless of whether such large changes can occur in mammalian tissues, BHMT activity in crude GP liver extracts or the activity of the pure human liver enzyme is not sensitive to salt concentration. Addition of up to 500 mM NaCl or KCl to these BHMT solutions has no affect on enzyme activity. This points to a key difference in the way A. halophytica and mammals use BHMT for osmoregulation.

In summary, this report shows that the expression of liver and kidney BHMT is downregulated in salt-loaded GPs. All of the mechanistic features involved in this downregulation have yet to be determined, but we propose that the physiological significance of this downregulation might be to decrease the use of Bet as a methyl donor, making it more readily available for transport and accumulation in the kidney medulla as a means to combat high extracellular osmolarity. Confirmation of this hypothesis awaits in vivo measurements of Bet oxidation under varying osmotic conditions.


This research was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-52501 and the Illinois Agricultural Experiment Station project ILLU-698-352.


We thank Rod Johnson, Manabu Nakamura, and Matthew Wallig for many helpful discussions. We also thank Sandy Szegedi for help with the editing of this manuscript.

Portions of these data were presented at the national meeting of the Federation of American Societies for Experimental Biology in Washington, DC, 2004.


  • 1 The nucleotide sequence reported in this paper has been deposited to the GenBank/EBI Data Bank with accession number AY398694.

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract